Design, synthesis and biological evaluation of bivalent ligands against A 1-D 1 receptor heteromers

Article abstract of DOI:10.1038/aps.2012.151

Aim:To design and synthesize bivalent ligands for adenosine A ₁-dopamine D ₁ receptor heteromers (A ₁-D ₁ R), and evaluate their pharmacological activities.Methods:Bivalent ligands and their corresponding A 1 R monovalent ligands were designed and synthesized. The affinities of the bivalent ligands for A 1 R and D 1 R in rat brain membrane preparation were examined using radiolabeled binding assays. To demonstrate the formation of A ₁-D ₁ R, fluorescence resonance energy transfer (FRET) was conducted in HEK293 cells transfected with D 1-CFP and A 1-YFP. Molecular modeling was used to analyze the possible mode of protein-protein and protein-ligand interactions.Results:Two bivalent ligands for A 1 R and D 1 R (20a, 20b), as well as the corresponding A 1 R monovalent ligands (21a, 21b) were synthesized. In radiolabeled binding assays, the bivalent ligands showed affinities for A 1 R 10-100 times higher than those of the corresponding monovalent ligands. In FRET experiments, the bivalent ligands significantly increased the heterodimerization of A1R and D 1 R compared with the corresponding monovalent ligands. A heterodimer model with the interface of helixes 3, 4, 5 of A1R and helixes 1, 6, 7 from D 1 R was established with molecular modeling. The distance between the two ligand binding sites in the heterodimer model was approximately 48.4 ?, which was shorter than the length of the bivalent ligands.Conclusion:This study demonstrates the existence of A ₁-D ₁ R in situ and a simultaneous interaction of bivalent ligands with both the receptors.

Full text of DOI:10.1038/aps.2012.151

Acta Pharmacologica Sinica (2013) 34: 441–452
© 2013 CPS and SIMM All rights reserved 1671-4083/13 $32.00
www.nature.com/aps
npg
Original Article
Design, synthesis and biological evaluation of
bivalent ligands against A1–D1 receptor heteromers
*
*
*
Jian SHEN, Lei ZHANG, Wan-ling SONG, Tao MENG, Xin WANG, Lin CHEN, Lin-yin FENG , Ye-chun XU , Jing-kang SHEN
The State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai 201203,
China
Aim: To design and synthesize bivalent ligands for adenosine A1–dopamine D1 receptor heteromers (A1–D1R), and evaluate their phar-
macological activities.
Methods: Bivalent ligands and their corresponding A1R monovalent ligands were designed and synthesized. The affinities of the biva-
lent ligands for A1R and D1R in rat brain membrane preparation were examined using radiolabeled binding assays. To demonstrate
the formation of A1–D1R, fluorescence resonance energy transfer (FRET) was conducted in HEK293 cells transfected with D1-CFP and
A1-YFP. Molecular modeling was used to analyze the possible mode of protein-protein and protein-ligand interactions.
Results: Two bivalent ligands for A1R and D1R (20a, 20b), as well as the corresponding A1R monovalent ligands (21a, 21b) were syn-
thesized. In radiolabeled binding assays, the bivalent ligands showed affinities for A1R 10–100 times higher than those of the corre-
sponding monovalent ligands. In FRET experiments, the bivalent ligands significantly increased the heterodimerization of A1R and D1R
compared with the corresponding monovalent ligands. A heterodimer model with the interface of helixes 3, 4, 5 of A1R and helixes 1,
6, 7 from D1R was established with molecular modeling. The distance between the two ligand binding sites in the heterodimer model
was approximately 48.4 Å, which was shorter than the length of the bivalent ligands.
Conclusion: This study demonstrates the existence of A1–D1R in situ and a simultaneous interaction of bivalent ligands with both the
receptors.
Keywords: G protein-coupled receptors; adenosine; dopamine; A1 receptor; D1 receptor; heterodimers; bivalent ligands; radiolabeled
binding assay; FRET; molecular modeling
Acta Pharmacologica Sinica (2013) 34: 441–452; doi: 10.1038/aps.2012.151; published online 21 Jan 2013
noprecipitation and double immunolabeling experiments[5].
Introduction
G protein-coupled receptors (GPCRs) constitute the largest There are two subtypes of striatal GABAergic efferent neu-
protein family of cell surface receptors, and they mediate a rons, projecting to the thalamus across two distinct pathways:
variety of signaling processes. Over the last 30 years, many the striato pallidal neurons (indirect pathway) and the stria-
researchers have identified that GPCRs can form dimeric and tonigrostriatoentopeduncular neurons (direct pathway). Evi-
higher-order oligomers in natural tissues[1, 2]. In comparison dence indicates that A1R and D1R are coexpressed in the basal
with the constituting receptors, GPCR heteromers, which have ganglia and prefrontal cortex, which are particularly involved
different characteristics, are important for the understanding in the direct pathway[6]. In rat models of Parkinson’s disease,
of receptor function and pharmacology. As a result, bivalent the motor activating effects of the D1 agonist, SKF38393, were
ligands are currently being developed as a promising strategy enhanced by the A1 receptor antagonist[7]. This synergistic
for drug discovery[3, 4]
.
mechanism may provide some insight into the design of novel
The adenosine A1 and dopamine D1 receptors belong to the agents for the treatment of Parkinson’s disease and neuropsy-
superfamily of GPCRs. Evidence for the existence of A1R and chiatric disorders based on the pharmacological properties of
D1R heteromers has been presented and is based on immu- the A1–D1 heteromeric complex[5].
Although heteromerization was reported in native tissues,
*
To whom correspondence should be addressed.
there is no direct evidence to prove the existence of A1–D1
receptor heterodimerization. A1–D1 bivalent ligands could
be important research tools to detect heterodimerization and
further explore the biological functions of A1–D1 receptor het-
E-mail: jkshen@mail.shcnc.ac.cn (Jing-kang SHEN);
lyfeng@mail.shcnc.ac.cn (Lin-ying FENG);
ycxu@mail.shcnc.ac.cn (Ye-chun XU)
Received 2012-05-31 Accepted 2012-10-08

www.nature.com/aps
npg
Shen J et al
442
erodimers.
Bivalent ligands are molecules that consist of two phar-
linked by 4 and 6 oligoethylene glycol units.
We also synthesized the corresponding monovalent ligand
macophores linked by a spacer[8]. Recently, synthesis and compounds for the A1 antagonist as a control to determine
biological investigations of bivalent ligands were described to whether the bivalent ligands showed improved affinity for the
target serotonin[9], muscarinic[10], opioid[11] and other receptors. A1 receptor compared to the monovalent ligand controls. In
Rafael Franco and Miriam Royo designed bivalent ligands 1991, Neumeyer et al synthesized the derivative of SKF38393
targeting A2AR–D2R heterodimerization, which contained a to identify effective fluorescent probes. However, coupling
D2R agonist, a A2AR antagonist and their corresponding mon- of the large group to the 4’ site of SKF38393 resulted in a con-
ovalent control compounds. Their results indicated that the siderable loss of affinity[20]. As a result, we did not synthesize
existence of A2AR–D2R heterodimerization and a simultaneous the corresponding monovalent ligand compounds for the D1
interaction of bivalent ligands with both receptors[12]
We designed and synthesized a group of bivalent ligands
.
antagonist as a control.
containing a D1R agonist and a A1R antagonist linked by dif- Preparation of the bivalent ligands and their monovalent ligands
ferent lengths of PEG linkers. The binding properties of these Bivalent ligands linked by PEG chains were synthesized by
compounds were determined by radioligand binding studies the route demonstrated in Scheme 1. Commercially available
in membrane preparations from brain striatum. FRET experi- polyethylene glycol was easily converted to compound 2a–2b
ments were performed to test if these bivalent ligands can pro- using acrylonitrile via the Michael addition reaction. Reduc-
mote the formation of A1R and D1R heterodimers (Figure 1).
tion of compound 2a–2b with borane produced the compound
3a–3b.
The synthesis of the common intermediate, compound 4,
has previously been described[13]. Treatment of compound 4
with ethyl 4-bromobutanoate produced compound 5. Then,
compound 5 was heated in 20% NaOH (aq), followed by ring
closure, to generate compound 6. Compound 6 was coupled
with 3a–3b in the presence of PyBop/DIPEA in DMF to yield
7a–7b.
Commercially available nitroacetophenone was transformed
into compound 8 using a catalytic amount of Lewis acid. The
reaction of 8 with NaBH4 produced compound 9. Compound
9 and commercially available 3,4-dimethoxy phenyleth-
ylamine were heated at reflux in THF to yield compound 10.
Compound 10 was heated in PPA, followed by ring closure,
to produce the cyclization compound 11. The reaction of 11
Figure 1. General structure of bivalent ligands.
Materials and methods
Chemistry
Compound design
A1–D1 bivalent ligands were designed by combining two phar- with iron powder generated compound 12. The addition of a
macophoric entities with different lengths of linkers. First, Nosyl group to 12 in the presence of basic conditions yielded
compound 1, a naphthylxanthine derivative, was chosen as compound 13. Compound 13 was converted to 14 via a Mit-
the A1 antagonist pharmacophore. Then, SKF38393, a 1-aryl- sunobu reaction. Deprotection of compound 14 with BBr3 at
3-benzazepine analog that displays high binding affinity at the -78 °C produced compound 15 without future purification.
D1 receptor, was selected as the D1 agonist pharmacophore. The addition of a MOM group to 15 under basic conditions
Structure-activity relationship indicated that the N-3 site of produced compound 16. Compound 17 was easily obtained
compound 1 and 4’ site of SKF38393 are not active sites[13–16]
.
from 16 via an ester hydrolysis reaction.
Thus, linkers were introduced at these sites to combine the A1
antagonist and D1 agonist pharmacophores.
Compounds 18a–18b were obtained from 17 and 7a–7b
using standard peptide synthesis procedures, with EDCI/
Polyethylene glycol (PEG) is widely used due to its low HOBt as the catalytic coupling agents. Removal of the Nosyl
toxicity, excellent aqueous solubility and low antigenicity. group in 18a–18b, in the presence of K2CO3 and PhSH, yielded
Recently, Kiessling et al found that an unexpected enhance- compound 19a–19b. Deprotection of compound 19a–19b with
ment in the biological activity of a GPCR ligand was induced BBr3 at -78oC gave the target compound 20a–20b (Scheme 1).
by a polyethylene glycol substitute[17]. Considering the fact
Monovalent ligands were synthesized, and the route is dem-
that PEG has a moderate ΔStor (the entropy associated with tor- onstrated in Scheme 2. Compounds 21a–21b were obtained
sional motions about a single bond)[18], PEG was chosen as the using hexanoic acid, and 7a–7b were obtained using standard
linker for the selected pharmacophores. Recently, Gmeiner peptide synthesis procedures, with EDCI/HOBt as the cata-
et al reported the synthesis and biological investigation of lytic coupling agents.
bivalent ligands for D2-like receptors. Bivalent ligands, linked
by 5–8 oligoethylene glycol units, showed up to a 70-fold Reagents and conditions
increase of D3 binding affinity compared to monovalent ligand a) acrylonitrile, THF, 60 oC, 1.5 h (99%); b) BH3, THF, 60 oC,
compounds[19]. As a result, we designed bivalent ligands 3.5 h (97%); c) ethyl 4-bromobutanoate, DBU, DMF, 60oC, 48 h
Acta Pharmacologica Sinica

www.chinaphar.com
Shen J et al
npg
443
Scheme 1. Synthesis of the bivalent ligands linked by PEG.
Acta Pharmacologica Sinica

www.nature.com/aps
npg
Shen J et al
444
(44%); d) 20% NaOH (aq), 60oC, 3 h (61%); e) PyBOP, DIPEA, nonspecific binding tube.
DMF, rt, overnight (60%); f) Br2, AlCl3 (Cat.), Et2O, 0 oC, rt,
In the A1 receptor competitive binding assay, compounds of
3 h (95%); g) NaBH4, 2 mol/L NaOH (aq), THF-MeOH, 0oC, the A1 receptor antagonist, [3H]-DPCPX, and 200 g rat cortex
rt, 35 min (69%); h) 3,4-dimethoxyphenylethylamine, THF, membrane protein mixture were added to each tube. The final
reflux overnight (56%); i) PPA 100oC, 1 h (75%); j) Fe powder, volume was adjusted to 200 μL with Tris-HCl buffer (Tris-
HCl (aq), EtOH-H2O, reflux for 1 h (40%); k) Nosyl-Cl, Et3N, HCl 50 mmol/L, NaCl 120 mmol/L, KCl 5 mmol/L, MgCl2
DCM, 0oC, rt, overnight (80%); l) Ph3P, DIAD, hex-4-yn-1-ol, 1 mmol/L, CaCl2 2 mmol/L, pH 7.4). The tubes were kept
rt, overnight (80%); m) BBr3,CH2Cl2, -78 oC, rt, overnight; n) in a 37°C water bath for 30 min. Twenty microliters of Tris-
Bromomethyl methyl ether, DIPEA, rt, overnight; o) 2 mol/L HCl buffer was used to replace the sample in the total binding
NaOH(aq), THF, CH3OH, rt, 1 h (70%); p) EDCI, HOBt, tube, and 20 μL of the specific ligand, DPCPX (10 μmol/L) was
DIPEA, DMF, rt, overnight (40%); q) PhSH, K2CO3, 80oC, over- used to replace the sample in the nonspecific binding tube.
night (60%); r) TFA, DCM, rt, 1 h (60%).
When the water bath was finished, the reaction buffer was
immediately filtered using a Brandel 24-well cell collector and
GF/B filter to stop the reaction. The filter was washed three
Receptor binding assay
A P2 membrane fraction was prepared from whole brains of times with ice-cold 50 mmol/L Tris-HCl buffer (pH=7.4).
adult Wistar rats (200–300 g). The tissue was homogenized After the filter was dry and transferred to a 0.5 mL EP, flash-
in 10× (w/v) 10% sucrose solution, and the suspension was ing liquid was added and a MicroBeta2 porous plate was used
centrifuged at 1000×g for 10 min (4oC). The supernatant was to measure the radiation activity. The equation for calculating
decanted and retained on ice. The pellet was resuspended in the inhibition rate was as follows:
50 mmol/L Tris-HCl buffer (pH=7.4) at 1:5 (w/v). Then, the
Lowry method was employed to determine the protein con-
Inhibition rate (%)=100–(cpmsample–cpmnon-specific binding)/
(cpmtotal binding–cpmnon-specific binding)×100
centration. All the solutions were diluted to 2 mg/mL and Non-linear regression implemented using GraphPad Prism 4.0
stored at -80oC.
(GraphPad Software Inc) was used to calculate the IC50.
In the D1 receptor competitive assay, compound samples,
the D1 receptor antagonist, [3H]-SCH23390 (0.5 nmol/L) and Measurement of FRET efficiency in HEK293 cells
200 μg of rat cortex membrane protein mixture were added FRET technology detects protein-protein interactions in live
to each tube. All of the above operations were performed in cells with high sensitivity[21]. FRET occurs when a donor
an ice bath. The sequence in which the samples were added chromophore, initially in its electronic excited state, transfers
and their volumes are listed in Table 1. The final volume was energy to a nearby acceptor chromophore (10–100 Ǻ) through
adjusted to 200 μL with Tris-HCl buffer (Tris-HCl 50 mmol/L, non-radiative dipole-dipole coupling. The adenosine A1 recep-
NaCl 120 mmol/L, KCl 5 mmol/L, MgCl2 1 mmol/L, CaCl2 2 tor and dopaminergic D1 receptor form heterodimers when
mmol/L, pH 7.4). The tubes were kept in a 37°C water bath co-localized. This experiment employs FRET technology to
for 30 min. Twenty microliters of Tris-HCl buffer was used screen for compounds that promote the formation of A1R and
to replace the sample in the total binding tube, and 20 μL of D1R heterodimers.
SCH23390 (0.1 mg/mL) was used to replace the sample in the
HEK293 cells were transiently transfected with plasmid
DNA corresponding to the D1-CFP (acceptor) and A1-YFP
(donor), with a ratio of donor to acceptor DNA of 1:1. GFP-
YFP plasmids were used as a control. A mixture of cells
separately transfected with either A1 YFP or D1 CFP (A1YFP+
D1CFP) were used as negative controls. HEK293 cells were
transfected with the plasmids using the Trans Messenger
Transfection Reagent (QIAGEN) for 48 h before collection with
cold D-Hanks’ buffer. The cells were washed twice, resus-
pended and seeded into 96-well plates at a density of 1×105
cells/well for fluorescence detection. Fluorescence intensity
Table 1. The protocol of competitive binding assay.
Serial [3H]Ligand Ligand Drug Tris-HCl Protein
number
(μL)
(μL) (μL)
(μL)
(μL)
Total binding
Nonspecific binding
Sample
1, 2, 3
4, 5, 6
7, 8, 9
…
20
20
20
20
–
20
-
–
–
20
20
20
–
–
160
160
160
160
…
-
–
Scheme 2. Synthesis of the monovalent ligand compounds. Reagents and conditions: a) EDCI, HOBt, DIPEA, DMF, rt, overnight (70%).
Acta Pharmacologica Sinica

www.chinaphar.com
Shen J et al
npg
445
was measured with a double-channel Novo star fluorescent was added to the heterodimer and minimized using Sybyl 7.5.
spectrophotometer (channel1: excitation wavelength 440 nm,
emission wavelength 520 nm; channel 2: excitation wavelength
Results
440 nm, emission wavelength 470 nm). The energy transfer Chemistry
efficiency (FRET efficiency) was calculated using the following The bivalent ligands, 20a–20b, and monovalent ligands, 21a–
formula:
21b, were designed and synthesized through the route out-
FRET efficiency=F440/520/F440/470
.
lined in Scheme 1. All the compounds were purified by silica
F440/520 was the fluorescence intensity detected through channel gel thin-layer chromatography (>95%), and their structures
1, and F440/470 was the fluorescence intensity detected through were determined by nuclear magnetic resonance spectra and
channel 2. The excitation wavelength for CFP was 440 nm and low-resolution mass spectra. Details of their structures are
for YFP was 520 nm, and 470 nm was the emission wavelength shown in Table 2.
for CFP.
Receptor binding assay
Ligand docking studies
The binding affinity assay was performed using the bivalent
Homology models of the adenosine A1 receptor and dopamine ligands, 20a–20b, with 4 and 6 polyethylene glycol units,
D1 receptor were built using Modeller 9.10. The model of the respectively, along with their monovalent ligands, 21a–21b.
A1 receptor was constructed using the crystal structure of the As expected, the bivalent ligands, 20a–20b, showed moder-
adenosine A2A receptor (PDB code: 3EML) as a template[22]
.
ate binding for the A1 receptor and D1 receptor, while the
The model of the D1 receptor was constructed using multiple monovalent control ligands, 21a–21b, showed moderate bind-
templates from the crystal structures of Rhodopsin (PDB code: ing for the A1 receptor. However, there was no correlation
1U19)[23], the β2 adrenergic receptor (PDB code: 2RH1)[24], the between linker length and binding affinity for the bivalent
histamine H1 receptor (PDB code: 3RZE)[25], the S1P1 recep- ligands, 20a–20b. It is possible that the high flexibility of the
tor (PDB code: 3V2Y)[26], and the dopamine D3 receptor (PDB PEG linker could explain this phenomenon. The affinities for
code: 3PBL)[27]. All the alignments were generated using Mod- A1R was 10–100 times higher for bivalent ligands compared
eller 9.10, using default settings that were manually modified with their monovalent controls. These results are summarized
afterwards with careful attention to the alignment of all the in Table 2.
conserved residues. The ‘Biopolymer structure preparation’
protocol of Sybyl 7.5 was used to prepare the models, and Measurement of FRET efficiency in HEK293 cells
the minimization tool of Sybyl 7.5 was used to minimize all FRET technology detects protein-protein interactions in live
the models stepwise from hydrogen side-chains and main- cells with high sensitivity. The adenosine A1 receptor and
chains to whole molecule minimization. The dopamine D1 dopaminergic D1 receptor form heterodimers when co-local-
receptor models were later subjected to molecular dynamics ized. This experiment employed FRET technology to screen
(MD) simulation using Gromacs 4.5 for further refinement of for compounds that promote the formation of A1R and D1R
the long ECL2. The MD simulation was performed for 40 ns heterodimers.
using the CHARMM27 force field, and all the heavy atoms,
As shown in Figure 2, A1YFP+D1CFP represents the cells
except those on ECL2, were restrained by a constant force of transfected with either A1YFP or D1CFP, set as the negative
100000 (kJ·mol-1·nm-2). The production runs were conducted control. CFPYFP represents the positive control. A1YFP/
using NTP ensemble at 300K, with a 1 fs time-step integration. D1CFP indicates the cells co-transfected with both D1-CFP and
The Particle Mesh Ewald method (PME) was used to calculate
the electrostatic contribution to non-bonded interactions, with
a cutoff of 9 Å and a time step of 1 fs. The cutoff distance of
the van der Waals interaction and coulomb interaction were
14 Å and 9 Å, respectively. The LINCS algorithm was applied
to the system. All the models were evaluated and validated
using PROCHECK and the Pro-SA web server.
Autodock 4.2 was used to search for the binding conforma-
tion of compound 1 and SKF38393 in the A1 receptor and D1
receptor, respectively. A binding box of 19×19×22 Å, centered
at the binding cavity, was used, and 100 conformations were
generated for each ligand using genetic algorithm. The bind-
ing conformation was chosen from the largest cluster (cluster
tolerance was determined using default settings) and vali-
dated with the available mutation experiment data.
The ZDOCK protocol in Discovery studio 2.5 suite was used
Figure 2. Fluorescence resonance energy transfer (FRET) experiment con-
to search the heterodimer interface. Default parameters were
used, and the results were selected manually. The PEG linker
firmed bivalent ligands promoting the formation of heterodimers of A1R
and D1R. bP<0.05, cP<0.01 vs negative control group.
Acta Pharmacologica Sinica

www.nature.com/aps
npg
Shen J et al
446
Table 2. Binding affinity of bivalent and monovalent ligands to A1R and D1R in brain striatum membranes.
A1
D1
Name
Structure
IC50 (µmol/L)
IC50 (µmol/L)
1
0.092
NA
NA
SKF38393
0.47
3.13
20a
0.29
20b
21a
21b
0.44
3.65
1.98
NA
48.87
NA
A1-YFP. Plates were read 48 h after the transfection was fin- heterodimer interface was searched by ZDOCK and checked
ished. Parameters of the spectrophotometer were set accord- manually based on the ZDOCK score and associated bioin-
ing to the manual.
formatics data. Tarakanov and Fuxe deduced a set of triplets
FRET efficiency values were obtained from the positive that may be responsible for receptor-receptor interactions[28, 29]
.
control, in which the cells co-transfected with A1YFP/D1CFP Leucine-rich motifs, one type of these triplets, are charged
differed significantly compared to the negative control group residues that may serve as an ‘adhesive guide’ for the forma-
(cells transfected with either A1YFP or D1CFP, P<0.01). Similar tion of strong bonds between the amino acids of two receptors.
significant differences in the FRET efficiency values were also As seen in Figure 3, the homology triplets, ILS and AAV, exist
detected in cells treated with 100 nmol/L 20a, 1 μmol/L 20a or in helix 4 of A1 and helix 1 of D1. Combined with the docking
100 nmol/L 20b compared to the negative control cells. These rank, we established a heterodimer model with the interface
results indicate that the bivalent ligands, 20a–20b, increased of helixes 3, 4, and 5 of the A1 receptor and helixes 1, 6, and
the heterodimerization of the A1 receptor and D1 receptor to a 7 from the D1 receptor (Figure 4). The distance between the
higher extent than the monovalent control ligands, 21a–21b. It two ligand binding sites in this heterodimer model is approxi-
was consistent with that of the receptor binding assay.
mately 48.4 Å, which is shorter than the length of the bivalent
ligands. The linker of 20b was manually added to the docked
ligands in the heterodimer model and minimized with Sybyl
Ligand docking studies
A docking experiment was performed to confirm the length of tools. As shown in Figure 4, the linker tightly interacts with
the linker and the bivalent ligand binding mode. The A1–D1 ECL2 and ECL1 of the A1 receptor and D1 receptor. Moreover,
Acta Pharmacologica Sinica

www.chinaphar.com
Shen J et al
npg
447
that the bivalent ligand binds to both binding sites in the het-
erodimer.
Discussion
According to the binding assay results shown in Table 2, biva-
lent ligands 20a–20b showed moderate binding for the A1 and
D1 receptors, while monovalent control ligands 21a–21b only
showed moderate binding for the A1 receptor. The affinity for
A1R was 10–100 times higher for bivalent ligands compared
with their monovalent controls. One possible reason for this
phenomenon is conformational crosstalk within one recep-
tor protomer or modulation of the interaction between two
protomers of a receptor dimer induced by positive cooperativ-
ity. Positive cooperativity can also be induced when bivalent
ligands bind two adjacent binding sites of A1–D1 receptor
heteromers because binding of the second pharmacophore
is significantly accelerated due to the vicinity of ligand, thus
facilitating local concentration enrichment. As a result, a
bivalent ligand that binds to D1R may subsequently have a
higher affinity with A1R located in the near vicinity compared
to the corresponding A1R monovalent ligand[9, 30]. These data
indicate the specific interaction of these bivalent ligands with
A1–D1 receptor heteromers and suggest that these ligands
could be useful as pharmacological tools to detect receptor
heteromers in native tissue.
Figure 3. Analysis of sequence and alignment of A1 receptor and D1 re-
ceptor in silico receptor.
FRET technology has been decisive for demonstrating het-
eromer formation in heterologous expression systems. This
study employed FRET technology to screen for bivalent
ligands that promote the formation of A1R and D1R heterodim-
ers. FRET efficiency values obtained from cells co-transfected
with A1YFP/D1CFP differed significantly compared to those
of the negative control group. Similar significant differences
were also observed in the FRET efficiency values of cells
treated with bivalent ligands compared to the negative control
cells. These results confirm that bivalent ligands can detect
the presence of heterodimers of A1R and D1R.
The docking experiments showed that the linker of the
bivalent ligands is long enough to allow for the two pharma-
cophoric moieties to bind an A1–D1 heterodimer. The mini-
mum distance between the two ligand binding sites in this
heterodimer model was 48.4 Å. This length is similar to the
one proposed for other GPCR dimers, such as A2A–D2 recep-
tor heterodimers. As shown in Figure 4, the linker has a tight
interaction with ECL2 and ECL1 of the A1 and D1 receptors.
Moreover, the oxygen atoms on the PEG linker form polar
interactions with the polar residues on the extracellular loops
of the heterodimer, which enhances the binding ability of the
bivalent ligand to the heterodimer. Ligand docking studies
verify that bivalent ligands bind to both binding sites in the
heterodimer.
Figure 4. Energy minimized heterodimer model with compound 20b bind-
ing. The left is D1 receptor and the right A1 receptor. (A) The side view of
the compound 20b binding in the heterodimer. Ligand colored grey and
the residues interacting with it are shown in lines. (B) The top view of the
compound 20b binding in the heterodimer. Residues that form hydrogen
bonds with PEG linker are shown in sticks and colored magentas. (C) Biva-
lent ligand is shown in sticks and colored white. Residues that form hydro-
gen bonds with PEG linker are shown in sticks and colored black. These
residues include: Glu153, Lys173 from A1 receptor and Asn311, Ser310
from D1 receptor.
Conclusion
This report provides evidence for the existence of A1R and D1R
heteromers. However, in native tissues, there is no direct evi-
dence to prove the existence of A1–D1 receptor heterodimeriza-
tion. In the present study, we designed and synthesized a
the oxygen atoms on the PEG linker form polar interactions
with the polar residues on the extracellular loops of the het-
erodimer, which enhances the binding ability of the bivalent
ligand to the heterodimer. The ligand docking studies verify
Acta Pharmacologica Sinica

www.nature.com/aps
npg
Shen J et al
448
described for the preparation of compound 3a; LRMS (ESI) m/z
set of the bivalent ligands, differing in length, that link two
different pharmacophores. The binding affinities for bivalent
ligands for A1R were 10–100 times higher compared with
their monovalent controls. The results of the binding assays
revealed the specific interaction of these bivalent ligands with
A1–D1 receptor heteromers. FRET efficiency values obtained
from cells co-transfected with 100 nmol/L 20a, 1 μmol/L 20a
or 100 nmol/L 20b differed significantly compared to those
of the negative control group. These results confirmed that
bivalent ligands can detect the presence of A1R and D1R het-
erodimers. The docking experiments also verified that biva-
lent ligands act on both binding sites of the heterodimer and
indicated the possible length of the linker. In summary, we
designed and synthesized bivalent ligands that act as molecu-
lar probes for A1–D1 receptor heteromers. A1–D1 bivalent
ligands might be an important research tool to detect heterodi-
merization and explore further biological functions of A1–D1
receptor heterodimers. Further studies are in progress.
397 [M+H]+.
N-(6-amino-1-(3-hydroxypropyl)-2,4-dioxo-3-propyl-1,2,3,4-
tetrahydropyrimidin-5-yl)-2-naphthamide (5)
To a solution of Compound 4 (1.2 g, 3.55 mmol) in dry DMF (50
mL) stirred under N2 at rt, were added ethyl 4-bromobutanoate
(2.03 mL, 14.2 mmol), DBU (0.72 mL, 7.1 mmol). The reaction
mixture was stirred at 60oC for 48 h. After cooling reaction, the
solvent was removed under vacuo, the residue was dissolved
in EtOAc (50 mL), washed with water and brine, dried over
Na2SO4, filtered and concentrated. The crude product was puri-
fied by chromatography (2:1 petroleum ether-EtOAc) to give
1
the title compound as an oil (700 mg, 44%); H-NMR (300 MHz,
DMSO-d6): δ 9.08 (s, 1H), 8.62 (s, 1H), 8.07–7.98 (m, 4H), 7.63–7.60
(m, 2H), 6.79 (s, 2H), 4.10–4.05 (m, 2H), 3.92 (t, J=5.7 Hz, 2H), 3.73
(t, J=5.1 Hz, 2H), 2.38 (t, J=5.7 Hz, 2H), 1.84–1.80 (m, 2H), 1.55–1.49
(m, 2H), 1.19 (t, J=5.7 Hz, 3H), 0.84 (t, J=5.7 Hz, 3H).
3-(3-hydroxypropyl)-8-(naphthalen-2-yl)-1-propyl-1H-purine-
Appendix
The chemical reagents (chemicals) were purchased from Lan- 2,6(3H,7H)-dione (6)
caster, Acros, and Shanghai Chemical Reagent Company which To a solution of Compound 5 (600 mg, 1.32 mmol) in methanol
were used without further purification. 1H- and 13C-NMR spectra (120 mL) stirred at rt, was added 20% NaOH (11 mL). The reac-
were recorded in DMSO-d6 or CDCl3 on Varian Mercury-300 or tion mixture was stirred at 60oC for 3 h. After cooling the reac-
Varian Mercury-400 instruments. The ESI-MS were carried out on tion, the solution was acidified by 2 mol/L HCl to adjust pH<3,
Thermo Finnigan LCQDECAXP. TLC was carried out with glass the solid was collected by filter, dried to give the title compound
1
pre-coated silica gel GF254 plates.
as white solid (350 mg, 61%); H-NMR (400 MHz, DMSO-d6): δ
8.69 (s, 1H), 8.22 (d, J=8.8 Hz, 1H), 8.02–7.93 (m, 3H), 7.58–7.55 (m,
2H), 4.12 (t, J=6.4 Hz, 2H), 3.85 (t, J=7.2 Hz, 2H), 2.32–2.29 (t, J=7.2,
4,7,10,13,16-pentaoxanonadecane-1,19-dinitrile (2a)
In a RB flask, was added acrylonitrile (50 mL, 0.76 mol), 2,2’-(2,2’- 2H), 2.20–1.59 (m, 2H), 1.61–1.55 (m, 2H), 0.87 (t, J=7.6 Hz, 3H);
oxybis(ethane-2,1-diyl)bis(oxy))diethanol (6.53 g, 33.67 mmol) and LRMS (ESI) m/z 379 [M+H]+; mp 190–194oC.
KOH (28.8 mg, 0.51 mmol), the reaction mixture was stirred at rt
for 1.5 h. The reaction was quenched with several drops of HCl (6 N-(19-amino-4,7,10,13,16-pentaoxanonadecyl)-4-(8-(naphthalen-2-
mol/L), the solvent was evaporated under vacuo, the residue was yl)-2,6-dioxo-1-propyl-1H-purin-3(2H,6H,7H)-yl)butanamide (7a)
dissolved in DCM, filter through celite, the filtrate was evaporated To a solution of compound 6 (100 mg, 0.25 mmol) and 3a (60 mg,
to dryness to give the title compound as yellow oil without further 0.28 mmol) in dry DMF (10 mL) stirred at rt, were added PyBOP
purification (7.75 g, 99%); LRMS (ESI) m/z 301 [M+H]+.
(143 mg, 0.28 mmol), DIPEA (88 μL, 0.5 mmol), the reaction mix-
ture was stirred at rt overnight. The solvent was evaporated, the
residue was purified by chromatography (10: 1 DCM-MeOH) to
4,7,10,13,16,19,22-Heptaoxapentacosanedinitrile (2b)
Compound 2b (yield 98 %) was prepared according to the method afford the title compound as a white solid (96 mg, 60%). 1H-NMR
described for the preparation of compound 2a; LRMS (ESI) m/z (400 MHz, DMSO-d6) δ 8.70 (s, 1H), 8.22–8.21 (m, 1H), 8.05-7.95
389 [M+H]+.
(m, 3H), 7.79 (m, 1H), 7.60–7.57 (m, 1H), 4.10 (t, J=8 Hz, 2H), 3.87
(t, J=9.1 Hz, 2H), 3.48–3.43 (m, 18H), 3.36–3.32 (m, 2H), 3.06–3.04
(m, 2H), 2.83 (t, J=9.6 Hz, 2H), 2.15 (m, 2H), 2.01–1.99 (m, 2H),
4,7,10,13,16-pentaoxanonadecane-1,19-diamine (3a)
To a solution of compound 3a (4 g, 13.3 mmol) in dry THF (30 1.78–1.74 (m, 2H), 1.61–1.55 (m, 4H), 0.89 (t, J=9.6 Hz, 3H); LRMS
mL) stirred under N2 at 0oC, was added 1 mol/L BH3-THF (100 (ESI) m/z 697 [M+H]+.
mL, 100 mmol) dropwise during 30 min. The reaction mixture
was heated to reflux for 3.5 h. Then the reaction was quenched N-(25-amino-4,7,10,13,16,19,22-heptaoxapentacosyl)-4-(8-
with MeOH (30 mL) and HCl (7 mL) at 0oC. After removing of (naphthalen-2-yl)-2,6-dioxo-1-propyl-1H-purin-3(2H,6H,7H)-yl)
the solvent, the residue was dissolved in 2 mol/L NaOH (50 mL), butanamide (7b)
extracted with DCM (25 mL*15), dried over Na2SO4, filtered and Compound 7b (yield 50%) was prepared according to the meth-
concentrated to give the title compound as colorless oil without odology described for the preparation of compound 7a. 1H-NMR
further purification (4.01 g, 97%); LRMS (ESI) m/z 309 [M+H]+.
(400 MHz, DMSO-d6): δ 8.70 (s, 1H), 8.04–8.01 (m, 1H), 7.97–7.96
(m, 3H), 7.60–7.59 (m, 1H), 7.68–7.57 (m, 1H), 4.10 (t, J=8 Hz, 2H),
3.87 (t, J=9.1 Hz, 2H), 3.50–3.42 (m, 26 H), 3.34 (t, J=8.4 Hz, 2H),
4,7,10,13,16,19,22-heptaoxapentacosane-1,25-diamine (3b)
Compound 3b (yield 98%) was prepared according to the method 3.06–3.04 (m, 2H), 2.82 (t, J=10.4 Hz, 2H), 2.16 (m, 2H), 2.03–1.96
Acta Pharmacologica Sinica

www.chinaphar.com
Shen J et al
npg
449
(m, 2H), 1.78 (m, 2H), 1.61–1.55 (m, 4H), 0.89 (t, J=9.6 Hz, 3H); C2H5OH (30 mL) and H2O (10 mL), followed by Fe (185 mg, 3.3
LRMS (ESI) m/z 785 [M+H]+.
mmol) and several drops of HCl (aq). The reaction mixture was
heated to reflux for 1 h. After cooling the reaction, saturated
NaHCO3 was added to adjust the pH>8, filtered through celite,
2-bromo-1-(4-nitrophenyl)ethanone (8)
To a solution of 4-nitroacetophenone (30 g, 0.18 mol) in dry ether washed with EtOAc, the combined filtrate was washed with
(500 mL) stirred at 0 oC, was added AlCl3 (cat), followed by Br2 brine, dried over Na2SO4, filtered and concentrated to give the
1
(10.3 mL, 0.20 mol). The reaction mixture was stirred at rt fot 3 title compound as a red solid (170 mg, 40%); H-NMR (300 MHz,
h. Then the reaction was quenched with Na2S2O3-NaHCO3 (aq), DMSO-d6): δ 6.80–6.76 (m, 3H), 6.55–6.53 (m, 2H), 6.36 (s, 1H), 4.95
extracted with DCM (400 mL*2), the combined organic phase was (brs, 2H), 4.19–4.16 (m, 1H), 3.71 (s, 3H), 3.50 (s, 3H), 3.20–3.17 (m,
washed with brine, dried over Na2SO4, filtered and concentrated 1H), 3.05–2.84 (m, 3H), 2.83–2.76 (m, 2H), 2.48–2.40 (m, 1H); LRMS
1
to give the title compound as yellow solid (42 g, 95%); H-NMR (ESI) m/z 229 [M+H]+; mp 148–150oC.
(300 MHz, CDCl3): δ 8.36–8.33 (m, 2H), 8.17–8.14 (m, 2H), 4.46 (s,
2H); LRMS (ESI) m/z 244 [M+H]+; mp 76–79oC.
N-(4-(7,8-dimethoxy-3-(2-nitrophenylsulfonyl)-2,3,4,5-tetrahydro-1H-
benzo[d]azepin-1-yl)phenyl)-2-nitrobenzenesulfonamide (13)
To a solution of 12 (4.5 g, 15.1 mmol), triethylamine (4.62 mL, 33.2
2-(4-nitrophenyl)oxirane (9)
To a solution of compound 8 (43 g, 0.18 mol) in 150 mL CH3OH- mmol) in dry DCM (150 mL) stirred under N2 at 0oC, was added
THF (V/V=1/1.2) stirred at 0 oC, was added NaBH4 (6.8 g, 0.18 2-nitrobenzenesulfonyl chloride (7.03 g, 31.7 mmol) dropwise dur-
mol), followed by 2 mol/L NaOH (90 mL). The reaction mixture ing 10 min, the reaction mixture was stirred at rt overnight. The
was stirred at rt for 35 min. The solvent was evaporated under mixture was washed with H2O (20 mL) and brine (15 mL), dried
vacuum, the residue was treated with HOAc to adjust pH=4, over Na2SO4, filtered and concentrated. The crude product was
extracted with DCM (100 mL), the organic phase was washed with then purified by chromatography (200:1 DCM-MeOH) to give
saturated NaHCO3, water and brine, dried over Na2SO4, filtered the title compound as an amorphous yellow solid (4.5 g, 80%);
and concentrated. The crude product was then purified by chro- 1H-NMR (300 MHz, CDCl3): δ 7.96–7.71 (m, 8H), 7.06–7.02 (m, 4H),
matography (10:1 petroleum ether-EtOAc) to give the title com- 6.78 (s, 1H), 6.34 (s, 1H), 4.40–4.27 (m, 1H), 4.00–3.85(m, 1H), 3.70
pound as yellow solid (20 g, 69%); 1H-NMR (300 MHz, DMSO-d6): (s, 3H), 3.68–3.61 (m, 1H), 3.47 (s, 3H), 3.40–3.31 (m, 2H), 2.81–2.96
δ 8.20 (d, J=8.7 Hz, 2H), 7.55 (d, J=8.7 Hz, 2H), 4.12–4.10 (m, 1H), (m, 2H); LRMS (ESI) m/z 669 [M+H]+.
3.20 (dd, J=5.4, 4.5 Hz, 1H), 2.87 (dd, J=5.7, 2.7 Hz, 1H); LRMS
(ESI) m/z 166 [M+H]+; mp 78–80oC.
N-(4-(7,8-dimethoxy-3-(2-nitrophenylsulfonyl)-2,3,4,5-tetrahydro-
1H-benzo[d]azepin-1-yl)phenyl)-2-nitro-N-(pent-4-ynyl)benzene-
sulfonamide (14)
2-(3,4-dimethoxyphenethylamino)-1-(4-nitrophenyl)ethanol (10)
To a solution of 9 (5 g, 0.03 mol) in dry THF (50 mL) stirred under To a solution of 13 (5 g, 7.48 mmol), triphenyl phosphine (4.56 g,
N2 at rt, was added 3,4-dimethoxyphenethylamine (5.6 mL, 0.033 14.9 mmol) and hex-5-yn-1-ol (1.65 mL, 14.9 mmol) in dry DCM
mol), the reaction mixture was heated to reflux overnight. The (30 mL) stirred under N2 at 0oC, was added DIAD (2.95 mL, 14.9
solvent was evaporated under vacuum, and the residue was mmol) dropwise during 10 min, the reaction mixture was stirred
purified by chromatography (10:1 DCM-MeOH) to give the title at rt overnight. The mixture was washed with H2O (20 mL) and
1
compound as an a yellow solid (18 g, 56%); H-NMR (300 MHz, brine (15 mL), dried over Na2SO4, filtered and concentrated. The
CDCl3): δ 8.16–8.14 (m, 2H), 7.60–7.57 (m, 2H), 6.82–6.68 (m, 3H), crude product was then purified by chromatography (200:1 DCM-
5.58 (s, 1H), 4.77 (s, 1H), 3.70 (s, 3H), 3.68 (s, 3H), 2.74–2.60 (m, MeOH) to give the title compound as an white solid (4.5 g, 80%);
6H), 1.51 (s, 1H); LRMS (ESI) m/z 347 [M+H]+; mp 90–93oC.
1H-NMR (400 MHz, DMSO-d6): δ 7.96–7.92 (m, 3H), 7.90–7.83 (m,
2H), 7.65–7.63 (m, 2H), 7.60–7.57 (m, 1H), 7.14 (s, 4H), 6.82 (s, 1H),
7,8-dimethoxy-1-(4-nitrophenyl)-2,3,4,5-tetrahydro-1H-benzo[d] 6.44 (s, 1H), 4.48–4.42 (m, 1H), 4.02–4.00 (m, 3H), 3.81–3.78 (m,
azepine (11)
1H), 3.72 (s, 3H), 3.61–3.58 (m, 2H), 3.56 (s, 3H), 3.40–3.35 (m, 1H),
Compound 10 (10 g, 0.029 mmol) was treated with PPA (80 mL), 3.36–3.33 (m, 1H), 2.96–2.81 (m, 2H), 2.21–2.11 (m, 2H), 1.37 (s,
the reaction was heated to 100oC for 1.5 h. After cooling the reac- 2H), 1.35 (s, 2H), 1.20–1.16 (m, 1H), 1.14–1.11 (m, 3H); LRMS (ESI)
tion, the solution was poured into ice water, basidified by ammo- m/z 749 [M+H]+.
nia to adjust the pH>9, extracted with DCM (200 mL*2). The com-
bined organic phase was washed with water and brine, dried over Ethyl6-(N-(4-(7,8-dihydroxy-3-(2-nitrophenylsulfonyl)-2,3,4,5-tetra-
Na2SO4, filtered and concentrated to give the title compound as hydro-1H-benzo[d]azepin-1-yl)phenyl)-2-nitrophenylsulfonamido)
1
a red solid (11.3 g, 75%); H-NMR (300 MHz, CDCl3): δ 8.16–8.14 hexanoate (15)
(m, 2H), 7.38–7.35 (m, 2H), 6.79 (s, 1H), 6.58 (s, 1H), 4.34–4.20 (m, To a solution of 14 (160 mg, 0.20 mmol) in dry DCM (10 mL)
1H), 3.73 (s, 3H), 3.60 (s, 3H), 3.60–3.51 (m, 1H), 3.10–3.00 (m, 1H), stirred under N2 at -78oC, was added 1 mol/L BBr3 (590 μL, 0.59
2.90–2.80 (m, 1H), 2.71–2.58 (m, 3H); LRMS (ESI) m/z 329 [M+H]+; mmol) dropwise during 2 min. The reaction mixture was warmed
mp 131–134oC.
to rt and stirred overnight. The reaction was quenched with
MeOH (2 mL) when the solution was cooled to -78oC. Then the
solvent was removed under vacuo to give the title compound
4-(7,8-dimethoxy-2,3,4,5-tetrahydro-1H-benzo[d]azepin-1-yl)aniline (12)
Compound 11 (470 mg, 1.1 mmol) was added to a mixture of as brown solid without further purification; LRMS (ESI) m/z 783
Acta Pharmacologica Sinica

www.nature.com/aps
npg
Shen J et al
450
[M+H]+.
26-azatriacontyl)hexanamide (18b)
Compound 18b (yield 38%) was prepared according to the meth-
Ethyl6-(N-(4-(7,8-bis(methoxymethoxy)-3-(2-nitrophenylsulfonyl)- odology described for the preparation of compound 18a. 1H-NMR
2,3,4,5-tetrahydro-1H-benzo[d]azepin-1-yl)phenyl)-2-nitrophenyl- (400 MHz, DMSO-d6): δ 8.85 (s, 1H), 8.36–8.34 (m, 1H), 7.93–7.87
sulfonamido) hexanoate (16)
(m, 4H), 7.67–7.51 (m, 11H), 7.25–7.06 (m, 4H), 7.74–6.63 (m, 2H),
To a solution of compound 15 (100 mg, 0.13 mmol) in dry DCM 5.19–5.10 (m, 4H), 4.38–4.24 (m, 3H), 3.75–3.72 (m, 1H), 3.68–3.62
(10 mL) stirred at 0 oC, was added MOMBr (22 μL, 0.27 mmol), (m, 10H), 3.61–3.52 (m, 12H), 3.36–3.34 (m, 15H), 3.27–3.21 (m,
DIPEA (48 μL, 0.28 mmol), the reaction mixture was warmed to 5H), 3.19–3.11 (m, 5H), 2.28–2.19 (m, 4H), 2.07–1.99 (m, 5H), 1.75–
rt and stirred overnight. The solution was diluted with DCM (20 1.65 (m, 9H), 1.55–1.53 (m, 3H); LRMS (ESI) m/z 1609 [M+H]+.
mL), washed with water and brine, dried over Na2SO4, filtered
and concentrated. The residue was purified by chromatography 6-(4-(7,8-bis(methoxymethoxy)-2,3,4,5-tetrahydro-1H-benzo[d]
with CH2Cl2 to afford the title compound as colorless oil without azepin-1-yl)phenylamino)-N-(24-(8-(naphthalen-2-yl)-2,6-dioxo-1-
further purification; LRMS (ESI) m/z 888 [M+NH4]+.
propyl-1H-purin-3(2H,6H,7H)-yl)-21-oxo-4,7,10,13,16-pentaoxa-20-
azatetracosyl)hexanamide (19a)
6-(N-(4-(7,8-bis(methoxymethoxy)-3-(2-nitrophenylsulfonyl)-2,3,4,5- To a solution of compound 18a (45 mg, 0.03 mmol) and K2CO3 (40
tetrahydro-1H-benzo[d]azepin-1-yl)phenyl)-2-nitrophenylsulfonamido) mg, 0.3 mmol) in dry DMF (10 mL) stirred at rt, was added PhSH
hexanoic acid (17)
(30 μL, 0.3 mmol), the reaction mixture was stirred at rt overnight.
To a solution of compound 16 (30 mg, 0.03 mmol) was dissolved The solvent was evaporated under vacuo, the residue was puri-
in a mixture of THF/CH3OH (1:1) (10 mL) stirred at rt, was added fied by silica gel thin-layer chromatography (9:1 DCM-MeOH) to
2 mol/L NaOH (0.5 mL), the reaction mixture was stirred at rt afford the title compound as a yellow solid (20 mg, 60%); 1H-NMR
for 1 h. The solvent was evaporated, 2 mol/L HCl was added to (400 MHz, DMSO-d6): δ 8.79 (s, 1H), 8.32 (d, J=8.8 Hz, 1H), 8.02–
the residue to adjust pH<3, extracted with DCM (30 mL), washed 8.00 (m, 1H), 7.92–7.90 (m, 1H), 7.88–7.83 (m, 1H), 7.70–7.69 (m,
with brine, dried over Na2SO4, filtered and concentrated. The 1H), 7.52–7.50 (m, 2H), 7.21–7.19 (m, 1H), 6.99 (s, 1H), 6.83–6.81
residue was purified by chromatography (20: 1 DCM-MeOH) to (m, 2H), 6.74 (d, J=7.6 Hz, 2H), 6.29 (d, J=8 Hz, 2H), 5.24–5.23 (m,
1
afford the title compound as yellow oil (20 mg, 70%); H-NMR 2H), 5.12–5.11 (m, 2H), 4.28–4.24 (m, 3H), 3.94–3.89 (m, 1H), 3.79–
(300 MHz, CDCl3): δ 7.95–7.90 (m, 1H), 7.75–7.46 (m, 7H), 7.16–7.08 3.75 (m, 1H), 3.61–3.60 (m, 12H), 3.56–3.52 (m, 14H), 3.44 (s, 3H),
(m, 4H), 6.96 (s, 1H), 6.70 (s, 1H), 5.22 (s, 2H), 5.13–5.07 (m, 2H), 3.35–3.32 (m, 5H), 3.14–3.08 (m, 2H), 2.89–2.86 (m, 4H), 2.32–2.30
4.38–4.35 (m, 1H), 4.14–4.04 (m, 1H), 3.85–3.70 (m, 3H), 3.53 (s, (m, 2H), 2.20–2.16 (m, 5H), 1.88–1.73 (m, 5H), 1.54–1.51 (m, 4H),
3H), 3.51–3.49 (m, 1H), 3.45 (s, 3H), 3.40–3.34 (m, 1H), 2.96–2.90 0.88–0.84 (m, 3H); LRMS (ESI) m/z 1151 [M+H]+.
(m, 2H), 2.32 (t, J=7.2 Hz, 2H), 1.70–1.58 (m, 2H), 1.55–1.38 (m,
4H); LRMS (ESI) m/z 843 [M+H]+.
6-(4-(7,8-bis(methoxymethoxy)-2,3,4,5-tetrahydro-1H-benzo[d]
azepin-1-yl)phenylamino)-N-(30-(8-(naphthalen-2-yl)-2,6-dioxo-
6-(N-(4-(7,8-bis(methoxymethoxy)-3-(2-nitrophenylsulfonyl)-2,3,4,5- 1-propyl-1H-purin-3(2H,6H,7H)-yl)-27-oxo-4,7,10,13,16,19,22-
tetrahydro-1H-benzo[d]azepin-1-yl)phenyl)-2-nitrophenylsulfo- heptaoxa-26-azatriacontyl)hexanamide (19b)
namido)-N-(24-(8-(naphthalen-2-yl)-2,6-dioxo-1-propyl-1H-purin- Compound 19b (yield 55%) was prepared according to the meth-
3(2H,6H,7H)-yl)-21-oxo-4,7,10,13,16-pentaoxa-20-azatetracosyl) odology described for the preparation of compound 19a; 1H-NMR
hexanamide (18a)
(400 MHz, DMSO-d6): δ 8.85 (s, 1H), 8.36–8.34 (m, 1H), 8.06–8.03
To a solution of compound 17 (110 mg, 0.13 mmol) and 7a (91 mg, (m, 1H), 7.95–7.92 (m, 1H), 7.87–7.85 (m, 1H), 7.55–7.52 (m, 2H),
0.13 mmol) in dry DMF (10 mL) stirred at rt, was added EDCI 7.22–7.21 (m, 1H), 7.00–6.99 (m, 1H), 6.80–6.77 (m, 4H), 6.39–6.36
(31 mg, 0.16 mmol), HOBt (22 mg, 0.16 mmol) and DIPEA (68 μL, (m, 2H), 5.25–5.24 (m, 2H), 5.11–5.09 (m, 2H), 4.32–4.30 (m, 1H),
0.38 mmol), the reaction mixture was stirred at rt overnight. The 4.28 (t, J=16Hz, 2H), 3.85–3.77 (m, 2H), 3.63–3.62 (m, 1H), 3.61–3.52
solvent was evaporated, and the residue was purified by chro- (m, 32H), 3.45 (s, 3H), 3.41–3.31 (m, 5H), 3.08–3.05 (m, 2H), 2.95–
matography (10:1 DCM-MeOH) to afford the title compound as 2.90 (m, 3H), 2.20–2.18 (m, 5H), 1.78–1.72 (m, 5H), 1.69–1.52 (m,
an oil (80 mg, 40%); 1H-NMR (400 MHz, DMSO-d6): δ 8.66 (s, 1H), 8H), 0.88–0.86 (m, 3H); LRMS (ESI) m/z 1239 [M+H]+.
8.22–8.21 (m, 1H), 7.96–7.81 (m, 4H), 7.60–7.51 (m, 11H), 7.10–7.06
(m, 2H), 7.05–7.02 (m, 2H), 6.65–6.58 (m, 2H), 5.14–5.10 (m, 2H), 6-(4-(7,8-dihydroxy-2,3,4,5-tetrahydro-1H-benzo[d]azepin-1-yl)
5.02–5.01 (m, 2H), 4.28–4.27 (m, 3H), 4.16–4.15 (m, 1H), 3.98–3.94 phenylamino)-N-(24-(8-(naphthalen-2-yl)-2,6-dioxo-1-propyl-1H-purin-
(m, 1H), 3.72–3.65 (m, 12H), 3.62–3.55 (m, 14H), 3.45–3.41 (m, 3H), 3(2H,6H,7H)-yl)-21-oxo-4,7,10,13,16-pentaoxa-20-azatetracosyl)
3.32–3.25 (m, 5H), 3.15–3.07 (m, 6H), 2.22–2.12 (m, 2H), 2.02–1.93 hexanamide (20a)
(m, 5H), 1.77–1.65 (m, 9H), 0.95–0.93 (m, 3H); LRMS (ESI) m/z 1521 To a solution of compound 19a (12 mg, 0.01 mmol) in DCM (4
[M+H]+.
mL) stirred at rt, was added TFA (1 mL), the reaction mixture was
stirred at rt for 30 min. The solvent was evaporated under vacuo,
6-(N-(4-(7,8-bis(methoxymethoxy)-3-(2-nitrophenylsulfonyl)- the residue was purified by silica gel thin-layer chromatography
2,3,4,5-tetrahydro-1H-benzo[d]azepin-1-yl)phenyl)-2-nitrophenyl- (9:1 DCM-MeOH) to afford the title compound as an amorphous
sulfonamido)-N-(30-(8-(naphthalen-2-yl)-2,6-dioxo-1-propyl-1H- white solid (6.5 mg, 60%). 1H-NMR (400 MHz, DMSO-d6): δ 8.81
purin-3(2H,6H,7H)-yl)-27-oxo-4,7,10,13,16,19,22-heptaoxa- (brs, 1H), 8.71 (s, 1H), 8.66 (brs, 1H), 8.26–8.23 (m, 1H), 8.04–7.93
Acta Pharmacologica Sinica

www.chinaphar.com
Shen J et al
npg
451
(m, 3H), 7.96–7.80 (m, 2H), 7.59–7.56 (m, 2H), 6.84 (d, J=10.4 Hz, 8.25–8.22 (m, 1H), 8.05–7.96 (m, 3H), 7.83–7.76 (m, 2H), 7.60–7.57
2H), 6.61–6.54 (m, 3H), 6.05 (s, 1H), 5.61–5.54 (m, 1H), 4.33–4.31 (m, 2H), 4.11 (t, J=8.4 Hz, 2H), 3.87 (t, J=8.8 Hz, 2H), 3.46–3.42 (m,
(m, 1H), 4.10 (t, J=9.6 Hz, 2H), 3.87 (t, J=8.8 Hz, 2H), 3.46–3.41 (m, 24H), 3.40–3.35 (m, 8H), 3.07–3.01 (m, 4H), 2.15 (t, J=9.6 Hz, 2H),
12H), 3.35–3.33 (m, 16H), 3.07–3.01 (m, 4H), 2.96–2.91 (m, 2H), 2.03–1.98 (m, 4H), 1.60–1.54 (m, 6H), 1.46–1.45 (m, 2H), 0.86 (t, J=8
2.18–2.13 (m, 2H), 2.06–1.95 (m, 5H), 1.60–1.48 (m, 9H), 0.87 (t, J=8 Hz, 3H), 0.83–0.84 (m, 3H); LRMS (ESI) m/z 883 [M+H]+.
Hz, 3H); 13C-NMR (100 MHz, DMSO-d6) δ 179.67, 172.45, 171.76,
154.73, 151.15, 150.52, 148.31, 143.64, 143.62, 134.24, 133.88, 133.11,
130.21, 130.10, 129.65, 129.09, 128.97, 128.24, 128.12, 127.72, 127.43,
127.03, 126.42, 124.11, 118.07, 117.12, 112.53, 70.19, 70.15, 69.97,
69.93, 68.55, 68.52, 51.01, 45.95, 44.81, 43.28, 43.18, 42.63, 36.26,
36.15, 35.88, 33.15, 29.85, 29.77, 29.02, 26.90, 25.66, 24.39, 21.35,
14.8, 11.70; LRMS (ESI) m/z 1063 [M+H]+.
Acknowledgements
This study was supported by the ‘100 Talents Project’ of
CAS to Ye-chun XU and National S&T Major Project (No
2012ZX09301-001-005).
Author contribution
Jing-kang SHEN, Lin-yin FENG, and Ye-chun XU designed the
research; Jian SHEN, Tao MENG, Lei ZHANG, and Wan-ling
SONG performed the research; Xin WANG and Lin CHEN
contributed analytical tools and reagents; Jian SHEN and Tao
MENG analyzed data; and Jian SHEN wrote the paper.
6-(4-(7,8-dihydroxy-2,3,4,5-tetrahydro-1H-benzo[d]azepin-1-yl)
phenylamino)-N-(30-(8-(naphthalen-2-yl)-2,6-dioxo-1-propyl-1H-
purin-3(2H,6H,7H)-yl)-27-oxo-4,7,10,13,16,19,22-heptaoxa-26-
azatriacontyl)hexanamide (20b)
Compound 20b was prepared according to the methodology
described for the preparation of compound 20a. Yield: 55%.
1H-NMR (400 MHz, DMSO-d6): δ 8.78 (brs, 1H), 8.71 (s, 1H), 8.67
(brs, 1H), 8.26–8.23 (m, 1H), 8.04–7.93 (m, 3H), 7.96–7.80 (m, 2H),
7.60–7.58 (m, 2H), 6.85 (d, J=8.4 Hz, 2H), 6.59–6.55 (m, 3H), 6.06 (s,
1H), 5.56–5.54 (m, 1H), 5.32–5.39 (m, 1H), 4.33–4.31 (m, 1H), 4.11 (t,
J=9.6 Hz, 2H), 3.88 (t, J=8.8 Hz, 2H), 3.45–3.38 (m, 30H), 3.06–3.04
(m, 5H), 2.98–2.95 (m, 2H), 2.85–2.65 (m, 3H), 2.15–2.12 (m, 2H),
2.15–1.96 (m, 8H), 2.08–1.96 (m, 8H), 1.60–1.49 (m, 9H), 0.86 (t, J=8
Hz, 3H); 13C-NMR (100 MHz, d-DMSO) δ 174.77, 172.42, 171.73,
151.14, 148.33, 143.67, 133.92, 133.09, 130.11, 129.56, 129.08, 128.98,
128.25, 127.79, 127.47, 126.49, 124.07, 118.03, 112.53, 70.20, 69.98,
69.93, 68.55, 68.52, 50.88, 45.92, 44.77, 43.27, 43.17, 42.66, 36.27,
36.16, 35.89, 33.14, 29.86, 29.77, 29.02, 26.90, 25.65, 24.38, 22.55,
21.34, 14.42, 11.69; LRMS (ESI) m/z 1151 [M+H]+.
References
1
Milligan G. G protein-coupled receptor dimerization: function and
ligand pharmacology. Mol Pharmacol 2004; 66: 1–7.
2
Carriba P, Navarro G, Ciruela F, Ferre S, Casado V, Agnati L, et al.
Detection of heteromerization of more than two proteins by sequential
BRET-FRET. Nat Methods 2008; 5: 727–33.
3
4
George SR, O’Dowd BF, Lee SP. G-protein-coupled receptor oligomeri-
zation and its potential for drug discovery. Nat Rev Drug Discov 2002;
1: 808–20.
Milligan G. G-protein-coupled receptor heterodimers: pharmacology,
function and relevance to drug discovery. Drug Discov Today 2006;
11: 541–9.
5
6
Franco R. Dopamine D1 and adenosine A1 receptors form functionally
interacting heteromeric complexes. PNAS 2000; 97: 8606–11.
Franco R, Lluis C, Canela EI, Mallol J, Agnati L, Casadó V, et al.
Receptor-receptor interactions involving adenosine A1 or dopamine
D1 receptors and accessory proteins. J Neural Transm 2007; 114:
93–104.
N-(24-(8-(naphthalen-2-yl)-2,6-dioxo-1-propyl-1H-purin-3(2H,6H,7H)-
yl)-21-oxo-4,7,10,13,16-pentaoxa-20-azatetracosyl)hexanamide
(21a)
7
Popoli P, Gimenez-Llort L, Pezzola A, Reggio R, Martinez E, Fuxe K.
Adenosine A1 receptor blockade selectively potentiates the motor
effects induced by dopamine D1 receptor stimulation in rodents.
Neurosci Lett 1996; 218: 209–13.
To a solution of compound 7a (28 mg, 0.04 mmol) and hexanoic
acid (5.88 μL, 0.044 mmol) in dry DMF (2 mL) stirred at rt, was
added EDCI (9.2 mg, 0.048 mmol), HOBt (6.5 mg, 0.048 mmol) and
DIPEA (14 μL, 0.08 mmol), the reaction mixture was stirred at rt
overnight. The solvent was evaporated under vacuo, the residue
was purified by chromatography with CH2Cl2/CH3OH (9:1) to
afford the title compound as an amorphous white solid (21 mg,
70 %); 1H-NMR (400 MHz, DMSO-d6): δ 14.01 (s, 1H), 8.72 (s, 1H),
8.23–8.22 (m, 1H), 8.05–7.98 (m, 3H), 7.83–7.76 (m, 2H), 7.60–7.57
(m, 2H), 4.10 (t, J=8.4 Hz, 2H), 3.87 (t, J=8.8 Hz, 2H), 3.46–3.42 (m,
16H), 3.38–3.36 (m, 8H), 3.07–3.00 (m, 4H), 2.13 (t, J=9.6 Hz, 2H),
2.03–1.98 (m, 4H), 1.63–1.55 (m, 6H), 1.49–1.41 (m, 2H), 0.86 (t, J=8
Hz, 3H), 0.83–0.85 (m, 3H); LRMS (ESI) m/z 697 [M+H]+.
8
Liu Z, Zhang J, Zhang A. Design of multivalent ligand targeting G-pro-
tein-coupled receptors. Curr Pharm Des 2009; 15: 682–718.
Russo O, Berthouze M, Giner M, Soulier JL, Rivail L, Sicsic S, et al.
Synthesis of specific bivalent probes that functionally interact with
5-HT(4) receptor dimers. J Med Chem 2007; 50: 4482–92.
9
10 Christopoulos A, Grant MK, Ayoubzadeh N, Kim ON, Sauerberg P,
Jeppesen L, et al. Synthesis and pharmacological evaluation of
dimeric muscarinic acetylcholine receptor agonists. J Pharmacol Exp
Ther 2001; 298: 1260–8.
11 Peng X, Neumeyer JL. Kappa receptor bivalent ligands. Curr Top Med
Chem 2007; 7: 363–73.
12 Soriano A, Ventura R, Molero A, Hoen A, Casado V, Cortes A, et al.
Adenosine A2A receptor-antagonist/dopamine D2 receptor-agonist
bivalent ligands as pharmacological tools to detect A2A-D2 receptor
heteromers. J Med Chem 2009; 52: 5590–602.
13 Miiller C, Schobert U, Hipp J, Geis U, Frobenius W, Pawlowski M.
Configurationally stable analogs of styrylxanthines as A2A adenosine
receptor antagonists. Eur J Med Chem 1997; 32: 709–19.
14 González MP, Terán C, Teijeira M. Search for new antagonist ligands
N-(30-(8-(naphthalen-2-yl)-2,6-dioxo-1-propyl-1H-purin-3(2H,6H,7H)-
yl)-27-oxo-4,7,10,13,16,19,22-heptaoxa-26-azatriacontyl)
hexanamide (21b)
Compound 21b was prepared according to the methodology
described for the preparation of compound 21a. Yield: 71%;
1H-NMR (400 MHz, DMSO-d6): δ 14.01 (s, 1H), 8.71 (s, 1H),
Acta Pharmacologica Sinica

www.nature.com/aps
npg
Shen J et al
452
for adenosine receptors from QSAR point of view. How close are we?
Med Res Rev 2008; 28: 329–71.
15 Neumeyerc JL, Kula NS, Bergman J, Baldessarini RJ. Receptor
affinities of dopamine D1 receptor-selective novel phenylbenzazepines.
Eur J Pharmacol 2003; 474: 137–40
16 Zhang J, Xiong B, Zhen XC, Zhang A. Dopamine D1 receptor ligands:
Where are we now and where are we going? Med Res Rev 2009; 29:
272–94.
receptor bound to an antagonist. Science 2008; 322: 1211–7.
23 Okada T, Sugihara M, Bondar AN, Elstner M, Entel P, Buss V. The
retinal conformation and its environment in rhodopsin in light of a
new 2.2 A crystal structure. J Mol Biol 2004; 342: 571–83.
24 Cherezov V, Rosenbaum DM, Hanson MA, Rasmussen SG, Thian FS,
Kobilka TS, et al. High-resolution crystal structure of an engineered
human beta(2)-adrenergic G protein-coupled receptor. Science 2007;
318: 1258–65.
17 Jiarpinitnun C, Kiessling L. Unexpected enhancement in biological 25 Shimamura T, Shiroishi M, Weyand S, Tsujimoto H, Winter G, Katritch
activity of a GPCR ligand induced by an oligoethylene glycol sub-
stituent. J Am Chem Soc 2010; 132: 8844–5.
V, et al. Structure of the human histamine H1 receptor complex with
doxepin. Nature 2011; 475: 65–70.
18 Mammen M, Shakhnovich E, Whitesides G. Using a convenient, 26 Hanson MA, Roth CB, Jo E, Griffith MT, Scott FL, Reinhart G, et al.
quantitative model for torsional entropy to establish qualitative trends
for molecular processes that restrict conformational freedom. J Org
Chem 1998; 63: 3168–75.
Crystal structure of a lipid G protein-coupled receptor. Science 2012;
335: 851–5.
27 Chien EY, Liu W, Zhao Q, Katritch V, Han GW, Hanson MA, et al.
Structure of the human dopamine D3 receptor in complex with a D2/D3
selective antagonist. Science 2010; 330: 1091–5.
19 Huber D, Lober S, Hubner H, Gmeiner P. Bivalent molecular probes
for dopamine D2-like receptors. Bioorg Med Chem 2012; 20: 455–
66.
28 Tarakanov AO, Fuxe KG. Triplet puzzle: homologies of receptor
heteromers. J Mol Neurosci 2010; 41: 294–303.
20 Bakthavachalam V, Baindur N, Madras BK, Neumeyer JL. Fluorescent
probes for dopamine receptors: synthesis and characterization of 29 Tarakanov AO, Fuxe KG. The triplet puzzle of homologies in receptor
fluorescein and 7-nitrobenz-2-oxa-l,3-diazol-4-y conjugates of D-1 and
D-2 receptor ligands. J Med Chem 1991; 34: 3235–41.
heteromers exists also in other types of protein-protein interactions. J
Mol Neurosci 2011; 44: 173–7.
21 Wu P, Brand L. Resonance energy transfer: methods and applications. 30 Kuhhorn J, Hubner H, Gmeiner P. Bivalent dopamine D2 receptor
Anal Biochem 1994; 218: 1–13.
ligands: synthesis and binding properties. J Med Chem 2011; 54:
22 Jaakola VP, Griffith MT, Hanson MA, Cherezov V, Chien EY, Lane JR,
et al. The 2.6 angstrom crystal structure of a human A2A adenosine
4896–903.
Acta Pharmacologica Sinica