Adenosine A2A receptor-antagonist/dopamine D2 receptor-agonist bivalent ligands as pharmacological tools to detect A 2A-D2 receptor heteromers

Article abstract of DOI:10.1021/jm900298c

Adenosine A_2A (A_2AR) and dopamine D₂ (D₂R) receptors mediate the antagonism between adenosinergic and dopaminergic transmission in striatopallidal GABAergic neurons and are pharmacological targets for the treatment of Parkinson's disease.Here, a family of heterobivalent ligands containing a D₂R agonist and an A _2AR antagonist linked through a spacer of variable size was designed and synthesized to study A_2AR-D₂R heteromers. Bivalent ligands with shorter linkers bound to D₂R or A_2AR with higher affinity than the corresponding monovalent controls in membranes from brain striatum and from cells coexpressing both receptors. In contrast, no differences in affinity of bivalent versus monovalent ligands were detected in experiments using membranes from cells expressing only one receptor. These findings indicate the existence of A_2AR-D₂R heteromers and of a simultaneous interaction of heterobivalent ligands with both receptors. The cooperative effect derived from the simultaneous interaction suggests the occurrence of A_2AR-D₂R heteromers in cotransfected cells and in brain striatum. The dopamine/adenosine bivalent action could constitute a novel concept in Parkinson's disease pharmacotherapy.

Full text of DOI:10.1021/jm900298c

5590 J. Med. Chem. 2009, 52, 5590–5602
DOI: 10.1021/jm900298c
Adenosine A2A Receptor-Antagonist/Dopamine D2 Receptor-Agonist Bivalent Ligands as
Pharmacological Tools to Detect A2A-D2 Receptor Heteromers
‡
‡
†
§
ꢀ †
Aroa Soriano,†,3 Ruben Ventura,‡,3 Anabel Molero, Rob Hoen, Vicent Casado, Antoni Cortes, Francesca Fanelli,
ꢀ
Fernando Albericio, ,^,# Carmen Lluıs,† Rafael Franco,*,† and Miriam Royo*,‡,#
´
†Institut d’Investigacions Biomeꢁdiques August Pi i Sunyer (IDIBAPS), Centro de Investigacioꢀn Biomeꢀdica en Red sobre Enfermedades
Neurodegenerativas (CIBERNED), and Department of Biochemistry and Molecular Biology, University of Barcelona, Avenida Diagonal 645,
E-08028 Barcelona, Spain, ‡Combinatorial Chemistry Unit, Barcelona Science Park;University of Barcelona, Baldiri Reixac 10, E-08028
Barcelona, Spain, §Dulbecco Telethon Institute and Department of Chemistry and Advanced Scientific Computing Laboratory, University of
Modena and Reggio Emilia, via Campi 183, 41100 Modena, Italy, IRB Barcelona, Barcelona Science Park;University of Barcelona, Baldiri
^
´
ꢀ
Reixac 10, E-08028 Barcelona, Spain, Department of Organic Chemistry, University of Barcelona, Martı i Franques 1-11, E-08028 Barcelona,
Spain, and #CIBER-BBN, Networking Centre on Bioengineering, Biomaterials and Nanomedicine, Barcelona Science Park, Baldiri Reixac 10,
E-08028 Barcelona, Spain. 3 These authors have contributed equally to this paper.
Received March 10, 2009
Adenosine A2A (A2AR) and dopamine D2 (D2R) receptors mediate the antagonism between adenosinergic
and dopaminergic transmission in striatopallidal GABAergic neurons and are pharmacological targets for
the treatment of Parkinson’s disease. Here, a family of heterobivalent ligands containing a D2R agonist and
an A2AR antagonist linked through a spacer of variable size was designed and synthesized to study
A2AR-D2R heteromers. Bivalent ligands with shorter linkers bound to D2R or A2AR with higher affinity
than the corresponding monovalent controls in membranes from brain striatum and from cells coexpressing
both receptors. In contrast, no differences in affinity of bivalent versus monovalent ligands were detected in
experiments using membranes from cells expressing only one receptor. These findings indicate the existence
of A2AR-D2R heteromers and of a simultaneous interaction of heterobivalent ligands with both receptors.
The cooperative effect derived from the simultaneous interaction suggests the occurrence of A2AR-D2R
heteromers in cotransfected cells and in brain striatum. The dopamine/adenosine bivalent action could
constitute a novel concept in Parkinson’s disease pharmacotherapy.
Introduction
last 10 years, several studies have provided solid evidence
indicating that GPCRs are expressed on the plasma mem-
brane as homo-, heterodimers, and/or higher-order
oligomers.1-4 These GPCRs heteromers have differential
characteristics with respect to the constituting receptors and
thus play a significant role in the understanding of receptor
function and pharmacology.5-7 As a result, novel therapies
based on GPCR oligomerization have been recently pro-
posed, such as the development of bivalent ligands.8-10
A2A-D2 receptor heteromerization has been established in
coimmunoprecipitation studies on neuroblastoma cells11 and
by FRET and BRET analysis of living cells.12,13 A2A-D2
receptor heteromers were found to be constitutive because
they were also detected in the absence A2A and D2 receptor
agonist exposure. However, the stoichiometry of A2A-D2
receptor heteromers is unknown and they could represent
dimers or high-order oligomers. In addition antagonistic
interactions between adenosine A2AR and dopamine D2Rhave
been determined at the biochemical, functional, and behavioral
levels in the striatum, the main input structure of the basal
Adenosine A2A and dopamine D2 receptors belong to the
superfamily of G-protein-coupled-receptors (GPCRsa). Inthe
*To whom correspondence should be addressed: Phone: 0034
934021208. Fax: 0034 934021559. E-mail: rfranco@ub.edu. Phone:
0034 934037122. Fax: 0034 934020496. E-mail: mroyo@pcb.ub.cat.
a Abbreviations: A2AR, adenosine 2A receptor; D2R, dopamine 2
receptor; D2 LR, dopamine 2 receptor (long variant); cAMP, cyclic
adenosine monophosphate; GPCR, G-protein coupled receptor; GABA,
γ-aminobutyric acid L-DOPA, (S)-2-amino-3-(3,4-dihydroxyphenyl) pro-
panoic acid; PEG, polyethylene glycol; XCC, 2-(4-(2,6-dioxo-1,3-dipro-
pyl-2,3,4,5,6,7-hexahydro-1H-purin-8-yl)phenoxy)acetic acid; XAC,
8-(40-carboxymethoxyphenyl)-1,3-dipropylxanthine-2-aminoethylamide;
PPHT-NH2, 6-((4-aminophenethyl)(propyl)amino)-5,6,7,8-tetrahydro-
naphthalen-1-ol; PPHT, (()-2-(N-phenethyl-N-propyl)amino-5-hydroxy-
tetralin; AM-MBHA, 2-(4-(amino(2,4-dimethoxyphenyl)methyl)phenoxy)-
N-(phenyl(p-tolyl)methyl)acetamide; DIPCDI, N,N0-diisopropylcarbodii-
mide; PyBOP, benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluor-
ophosphate; DIEA, diisopropylethylamine; HOBt, hydroxybenzotriazole;
HOAt, 1-hydroxy-7-azabenzotriazole; Fmoc, fluorenylmethyloxycarbonyl;
TFA, trifluoroacetic acid; Ltk, leukocyte tyrosine kinase; DMSO, dimethyl
sulfoxide; RP-HPLC-MS, reversed phase-high performance liquid chroma-
tography-mass spectroscopy; Alloc, allyloxycarbonyl; DMF, dimethylfor-
mamide; DBU, diaza(1,3)bicyclo[5.4.0]undecane; DCM, dichloromethane;
cDNA, complementary deoxyribonucleic acid; ADA, adenosine deaminase;
Tris-HCl, tris(hydroxymethyl)aminomethane hydrochloric acid; Ac2O,
acetic anhydride; AcOH, acetic acid; EtOAc, ethyl acetate; DIAD, diisopro-
pyl azadicarboxylate; DCE, 1,2-dichloroethane; THF, tetrahydrofuran;
Et3N, triethylamine; PBS, phosphate buffered saline; HEK-293, human
embryonic kidney cells; [3H]-YM09151-2, tritium labeled nemonapride
(N-(1-benzyl-2-methylpyrrolidin-3-yl)-5-chloro-2-methoxy-4-(methylamino)-
benzamide); [3H]-ZM241358, tritium labeled 4-(2-[7-amino-2-(2-furyl)-
[1,2,4]triazolo[2,3-a][1,3,5]triazin-5-ylamino]ethyl)phenol. Abbrevia-
tions used for amino acids follow the IUPAC-IUB Comission of
Biochemical Nomenclature in Jones, J. H. J. Pept. Sci. 2003, 9, 1-8.
ganglia and a key component of the motor system.14-16 A2A
and D2R are colocalized in the dendritic spines of the striato-
pallidal GABAergic neurons,17-19 and the activation of A2A
R
R
reduces D2R recognition, G-protein-coupling, and signaling.20
In the early 1990s, it was proposed that A2AR antagonists
should be developed as an anti-Parkinson therapy through
their ability to enhance striatal D2R signaling.21 In fact, anti-
Parkinsonian therapies that combine A2AR selective antago-
nist and D2R agonists are currently in clinical trials (phase II).
r
pubs.acs.org/jmc
Published on Web 08/27/2009
2009 American Chemical Society

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18 5591
Adenosine receptor antagonists increase the therapeutic index
ratio between the therapeutic and unwanted side effects of
L-DOPA and other D2R agonists.22,23 In addition, preclinical
studies have raised the possibility that these therapies may
afford neuroprotective and antidyskinetic benefits.24
Despite consistent data providing evidence of heteromer-
ization in heterologous expression systems and indirect evi-
dence of heteromer occurrence in striatal neurons, there is not
yet any direct proof of the existence of A2A-D2 receptor
heteromers in native tissues. Molecules with the capability to
target A2A-D2 receptor heteromers can be functional phar-
macological tools to detect heteromers in brain striatum and
may also be useful to explore the potential of the A2A-D2
receptor heteromer as a target for Parkinson’s disease phar-
macotherapy.
Bivalent ligands are defined as molecules that contain two
pharmacophores linked through a spacer with the potential to
bind simultaneously to the two ligand binding sites present in
a GPCR heterodimer. This may result in very high affinity for
the receptor and it might allow the targeting of certain
heteromeric subtypes, increasing the selectivity of drug action.
This approach has been successfully used for some GPCRs
leading to interesting results. For instance, Jacobson et al.
described heterobivalent ligands composed of an A1 and A3
adenosine receptor agonists linked by a spacer that activate
both receptors.25 This coactivation resulted in a cardiopro-
tective effect that is significantly greater than that induced by
activation of either receptor individually. Furthermore,
opioid-induced tolerance and dependence in mice modulated
by the distance between pharmacophores in a bivalent ligand
series has been reported by Portoguese et al., a fact that
suggests a physical association between μ and δ opioid
receptors as heterodimers.26
Here we report the design and synthesis of a family of
A2AR-antagonist/D2R-agonist heterobivalent ligands as
pharmacological tools to study A2A-D2 receptor heteromers.
In these compounds, the D2R agonist and the A2AR antago-
nist are linked through a spacer of variable length. The spacers
were based on trifunctional amino acids in combination with
PEG-polyamide unit repeats, and their size was varied (from
26 to 118 atoms) to look for the best interaction with the
A2A-D2 receptor heteromer. The binding properties of these
compounds were determined by radioligand binding studies
in membrane preparations from brain striatum. Furthermore,
intracellular cAMP production assays were performed to test
the agonistic D2R and antagonistic A2AR nature of these
heterobivalent molecules. Experiments to evaluate the speci-
ficity of heterobivalent ligands interaction with A2A-D2
receptor heteromer were performed in cells expressing adeno-
sine and dopamine receptors.
PEG-polyamide oligomer. Spacer length could be varied
by changing the number of repetitive PEG-based units
(n=0-7) in the central part.
In prior studies, we explored libraries of 1 and 2, structu-
rally modified with diverse trifunctional amino acids that
acted as linkers to optimize the linkage of the pharmaco-
phores to the PEG polyamide based spacer. The most
appropriate derivatization for each pharmacophore was
chosen on the basis of binding assays and structure-activity
relationships.30 A lysine and a glutamic acid moiety were
chosen as linkers for 1 and 2, respectively.
In addition to the heterobivalent ligand library
(compounds 3-10, HXPn n=0-7), two libraries of mono-
valent ligands (compounds 11-18, MXn n=0-7) for A2A
R
and for D2R (compounds 19-26, MPn n=0-7), containing
capped linkers, were synthesized as controls in order to
factor out possible contributions of the linker to the binding
affinity (Figure 1).
Molecular Docking Studies. Docking experiments were
performed to confirm that the shortest designed ligand could
bind to an A2A-D2 receptor heterodimer. Tridimensional
models of the compound and our previously reported
A2A-D2 receptor heterodimer model12 were used in these
experiments. In this model, the heterodimer interface is
mainly formed by the second intracellular loop (I2), helices
4, 3, and 5 from D2R, and I2, helices 5, 3, and 4 from A2AR.
The estimated maximum inter-CR distance between the two
binding pockets of the receptors that participate in the dimer
is in the range of 40-45 A. These distances did not account
for the presence of additional lipids or domain swapping.
The linkers synthesized varied from 26 atoms (3) to 118
atoms (10) and were characterized for being hydrophilic and
highly flexible (Figure 1). Compound 3 with the linker (Lys-
Lys-[PEG/polyamide]0-Lys-Glu) was manually docked into
the putative binding sites of the two receptors by probing
slightly different conformations of the ligand and the recep-
tor side chains. The complex in Figure 1 in the Supporting
Information (SI) shows tight binding for compound 3 with
all three positively charged groups of the ligand interacting
with anionic amino acids and with most of the heteroatoms
of the ligand involved in H-bonds with the receptor. In this
complex, different from the XCC adenosinic pharmaco-
phore, the PPHT head performs specific interactions with
D2R. Such interactions, which consist of the H-bonds be-
tween the protonated nitrogen atom and the hydroxy group
of the ligand and, respectively, D114 and S197 of the
receptor, are consistent with the agonistic nature of PPHT.
The docking study predicts that even when the shortest
bivalent ligand adopts a partially extended conformation,
it can be enough to allow the two pharmacophoric parts to
bind to an A2A-D2 receptor heterodimer. However, due to
the flexibility of the linker part, we decided to explore other
ligand lengths.
Results
Design of the Libraries. The designed ligands consist of a
D2R agonist and an A2AR antagonist, which are connected by
a linker to a variable length spacer. Pharmacophores 1 (XCC),
a precursor of xanthine amine congener (XAC),27 and 2 ((()-
PPHT-NH2), a derivative of PPHT,28,29 were chosen as an
A2AR antagonist and a D2R agonist, respectively.
The spacer features a central PEG-polyamide oligomer
flanked by adjacent trifunctional amino acid moieties.
Trifunctional amino acids were incorporated in the design
of the bivalent ligands as linkage moieties to have an easy
synthetic connection between the pharmacophore and the
Synthesis of the Different Building Blocks of the Library.
The synthesis of the pharmacophores 1 and 2 is described in
Schemes 1 and 2, respectively. Compound 1 could be ob-
tained over five steps starting from the commercially avail-
able 1,3-dipropyl-urea (27) and cyano acetic acid (28).
Condensation of urea 27 with cyano acetic acid followed
by ring closure under basic conditions gave 1,3-dipropylpyr-
imidine-2,4(1H,3H)-dione (29) in 60% yield over two
steps. Uracil 31 was obtained in 38% yield over two steps
by nitrite addition to 29 and subsequently reduction of
the nitrite functionality to an amine. Finally, the two-step

5592 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18
Soriano et al.
Figure 1. Designed mono- and heterobivalent ligands.
Scheme 1. Synthesis of 1a
a (a) Ac2O, 80 ꢀC, 90%; (b) 70% NaOH(aq), pH 10-11, 67%; (c) NaNO2, 42% AcOH(aq), 52%; (d) Na2S2O4, EtOAc/H2O, 74%; (e) 2-(4-
formylphenoxy)acetic acid (32), EtOH, AcOH, Δ, then DIAD, toluene, Δ, 54%.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18 5593
Scheme 2. Synthesis of 2a
hydrolysis. After the synthesis was completed, the ligands
were cleaved from the solid support with TFA/H2O (95:5).
Purification of the ligands was done by preparative HPLC,
yielding high purity ligands (g95%).
Binding Properties of Bivalent and Monovalent Ligands.
The binding properties of heterobivalent ligands 3-10 and
monovalent ligands 11-26 were assayed via competitive
radioligand experiments using brain striatal membranes given
the fact that D2R and A2AR are naturally coexpressed in
striatal neurons.32 The A2AR selective antagonist [3H]-
ZM241385 was used to evaluate the binding properties of
heterobivalent ligands 3-10 and monovalent ligands 11-18
to this receptor (Figure 2a). XAC was tested as a reference
compound showing that the introduction of linkers to the
pharmacophoreleads to a reduction inthe displacementof the
radioligand binding to A2AR. As predicted, the binding of
monovalent ligands decreased when the ligand spacer length
increased. To illustrate this, the shortest ligand 11 exhibited a
radioligand binding decrease of 58%, compared to the longest
ligand 18 that showed a radioligand binding decrease of 17%.
All heterobivalent ligands showed greater capacity to displace
specific [3H]-ZM241385 binding than the corresponding mon-
ovalent ligands. Heterobivalent ligands 3-6, with linkers
from 26 to 66 atoms, competed with the [3H]-ZM241385)
with similar results, showing approximately a radioligand
binding decrease of 70% and higher radioligand binding
displacement than longer bivalent compounds.
A similar competition experiment was performed using
the D2R antagonist [3H]-YM09151-2 (Figure 2b) as radio-
ligand to evaluate the binding properties of these ligands for
D2R. In these assays, heterobivalent ligands 3-10 exhibited
higher displacement of specific [3H]-YM09151-2 binding
than the corresponding monovalent compounds 19-26
without any clear influence of the spacer length. A reduction
in the binding of these ligands for D2R was detected when
compared with the reference control (()PPHT binding,
however no significant differences in the binding properties
were found between members of the same library
(monovalent or heterobivalent ligands).
Binding Affinities of the Selected Ligands. After screening,
shorter heterobivalent ligands 3-6, that exhibited higher
displacement of specific A2AR radioligand binding and the
corresponding monovalent compounds 11-14 and 19-22
were selected to determine their pharmacological profiles.
Competition experiments in brain striatal membranes were
performed using 2.1 nM [3H]-ZM241385 for the A2AR binding
or 0.9 nM [3H]-YM09151-2 for the D2R binding and increasing
concentrations of heterobivalent and monovalent compounds
and for the control competition experiments increasing concen-
trations of ZM-241385 for A2AR or YM-09151-2 for D2R.
Representative competition curves for compounds 6, 14, and
22 are shown in Figure 3a. Considering the existence of homo-
dimers of A2AR33 and homodimers of D2R,34 binding data were
fitted to the two-state dimer receptor model35-37 to calculate
macroscopic equilibrium dissociation constants (KDB1 and
KDB2) and the binding cooperativity evaluated as the coopera-
tive index DC (Table 1). Competition curves were biphasic for
the D2R binding showing negative cooperativity (negative DC
values) arising from molecular communication between the two
subunits of the receptor homodimer (Figure 3b). Monophasic
competition curves were obtained for A2AR binding, indicating
cooperativity 0; that means compounds bind with the same
affinity to the two subunits in the A2AR homodimer and there-
fore KDB1 is enough to characterize the binding (Figure 3a).38
a (a) Me2SO4, K2CO3, acetone, Δ, 94% (b) EtOH, Na, 50 ꢀC to Δ,
then conc. HCl(aq), Δ 88% (c) DCE, NaBH(OAc)3, 67% (d) 2-(4-nitrophe-
nyl)acetyl chloride (36), Et3N, DCM, RT, 87% (e) 1M BH3 THF in
3
THF, 0 ꢀC to Δ, 67% (f) H2NNH2, Raney-Ni, EtOH, Δ, 74% (g) 1M BBr3
in DCM, -78 ꢀC to RT, 85%.
condensation of uracil 31 with 2-(4-formylphenoxy)acetic
acid (32) was performed to give the final product 1 in 54%
yield. Most of the products were easy to purify because they
were insoluble in the reaction solvents. Simple filtration and
subsequent drying of the solid gave the desired products in
good to excellent yields with high purities.
Compound 2 was synthesized in a slightly different route
compared to an earlier published synthesis.28 Etherification of
the commercially available 1,6-dihydroxynaphthalene (33)
followed by a Birch reduction gave 5-methoxy-2-tetralone
(34) inan83% yieldover two steps. Reductiveaminationof 34
with propyl amine and NaBH(OAc)3 as reductive agent
yielded the secondary amine 35 in a good yield. Reaction of
35 with acid chloride 36 gave amide 37 as a 1:1 mixture of two
1
rotamers as could be observed by H and 13C NMR. Sub-
sequent reduction of the amide functionality with a 1 M
solution of BH3 THF complex in THF yielded 38 in 58%
3
over two steps. Reduction of the NO2 group with hydrazine
and Raney-Ni followed by deprotection of the methyl ether by
a BBr3 solution in DCM gave the desired PPHT-NH2 (2) in a
63% yield. All products were characterized by standard
techniques such as 1H and 13C NMR and HPLC-MS.
The monomeric PEG-based unit (39) was synthesized
according to a previously described procedure.31
Synthesis of the Libraries. The synthesis of the libraries of
heterobivalent and monovalent ligands was carried out by an
Fmoc-based solid-phase strategy using the AM-MBHA
resin as a polymeric support (Scheme 3). The amino acids
were coupled under standard peptide coupling conditions
using DIPCDI or PyBOP/DIEA as coupling reagents and
HOBt or HOAt as additives. Initial coupling of a Fmoc-
Lys(Alloc)-OH to the resin and subsequent selective removal
of the Fmoc-protecting group was followed by coupling of a
monomeric Fmoc-PEG-based unit (39). The process of
Fmoc elimination and coupling of a new monomeric
Fmoc-PEG-polyamide unit was repeated to construct di-
verse length spacers (n=0-7). Further construction of the
ligands is depicted in Scheme 3. Whereas the 1 and its
corresponding lysine linker were introduced separately into
the ligands, 2 and its glutamic acid linker were introduced
as one building block, previously synthesized by coupling of
2 to Fmoc-Glu(t-Bu)-OH with a subsequent t-Bu ester

5594 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18
Soriano et al.
Scheme 3. Synthetic Procedure for Heterobivalent Ligands (3-10) and Their Corresponding Monovalent Ligands (11-18, 19-26)a
a Conditions: (a) 1:4 or 1:1 piperidine in DMF; (b) Fmoc-Lys(Alloc)-OH, DIPCDI/HOBt; (c) Fmoc-HN(PEG)-OH, PyBOP, HOAt, DIEA; (d) Ac-
Lys(Fmoc)-OH, PyBOP, HOAt, DIEA; (e) Boc-Lys(Fmoc)-OH, PyBOP, HOAt, DIEA; (f) 1, PyBOP, HOAt, DIEA; (g) Pd(PPh3)4-PhSiH3 in DCManh
(h)Ac2O, DIEA; (i) TFA-H2O (95:5); (j) Fmoc-Glu(PPHT)-OH, PyBOP, HOAt, DIEA.
;
In agreement with the results obtained in the screening, the
A2AR binding affinity of xanthine-derived monovalent li-
gands decreased with increasing ligand length (from 160 to
364 nM). In contrast, irrespective of linker length, hetero-
bivalent ligands displayed similar affinities for A2AR (50 to
80 nM). The affinity was 3-7 times higher for heterobivalent
compounds compared to their monovalent counterparts
(Table 1, column 3). Taking into account the values of
KDB1 and KDB2 for D2R, compounds 5 and 6 showed the
highest binding enhancement (7-18 fold increase). For D2R
binding, no correlation was observed between linker length
and binding affinity of the compounds.
heterobivalent ligands with A2A-D2 receptor heteromers and
consequently validate their capacity to detect these heteromers
in native tissue, competition experiments were performed using
membrane preparations from Ltk cells expressing human
A2AR, human D2R, or both receptors. Representative screen-
ing of heterobivalent compounds 4 and 6 and their correspond-
ing monovalent ligands 12, 14, 20, and22 are shown in Figure 4.
In membranes from cells that coexpressed D2R and
A2AR, heterobivalent ligands displayed higher displace-
ment of the radioligand [3H]ZM-241385 binding than that
of the monovalent counterparts (MXn). Nevertheless,
when only A2AR was expressed the same radioligand dis-
placement was observed for heterobivalent and mono-
valent ligands alike (Figure 4a,b). In the case of D2R,
Heterobivalent Ligand Capacity to Detect A2A-D2 Recep-
tor Heteromers. To confirm the specific interaction of

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18 5595
Figure 2. Screening of heterobivalent (black bars) and monovalent (gray bars) compounds using striatal membranes. (a) Displacement of
A
2AR selective antagonist [3H]ZM-241385 (2.3 nM) binding by XAC and by heterobivalent 3-10 and monovalent 11-18 compounds.
(b) Displacement of D2R antagonist [3H]YM09151-2 (1.2 nM) binding by (()PPHT and by heterobivalent 3-10 and monovalent 19-26
compounds. The concentration of the different molecules was 1 μM. Values are given as % of control (specific radioligand binding determined
in the absence of displacer). Data are mean ( SD of three independent experiments. Heterobivalent ligand significantly different (*p < 0.05,
**p < 0.005, ***p < 0.0005) vs the corresponding monovalent ligand by Student’s t test for unpaired samples.
Figure 3. Representative competition curves for compounds 6 (b), 14 (0), and 22 (O) in striatum membranes. (a) Compounds 6 (b) and 14 (0)
or ZM 241358 (2) competed with 2.1 nM [3H] ZM241358. (b) Compounds 6 (b) and 22 (O) or YM09151-2 (1) competed with 0.9 nM
[3H]-YM09151-2. Data are means ( SD from representative experiments performed in triplicate.
Table 1. Binding Affinity of Heterobivalent and Monovalent Ligands to A2AR and D2R in Brain Striatum Membranesa
A2A
DB1 (nM)
55 ( 9
R
D2R
linker and length
ligand
K
DC
ligand
KDB1 (nM)
KDB2 (μM)
DC
(Lys-Lys-[PEG/polyamide]0-Lys-Glu)
26 atoms
3
11
0
0
3
19
1.0 ( 0.2
4 ( 1
9 ( 2
9 ( 2
-3.3
-2.8
160 ( 17
(Lys-Lys-[PEG/polyamide]1-Lys-Glu)
40 atoms
4
12
59 ( 11
246 ( 37
0
0
4
20
1.1 ( 0.3
11 ( 2
7 ( 3
15 ( 5
-3.2
-2.5
(Lys-Lys-[PEG/polyamide]2-Lys-Glu)
53 atoms
5
13
80 ( 10
268 ( 40
0
0
5
21
1.1 ( 0.3
11 ( 2
1.4 ( 0.5
12 ( 3
-2.5
-2.5
(Lys-Lys-[PEG/polyamide]3-Lys-Glu)
66 atoms
6
14
50 ( 8
364 ( 45
0
0
6
22
1.8 ( 0.4
13 ( 3
0.5 ( 0.1
9 ( 2
-1.8
-2.2
a Results are given as parameter values ( SEM of three independent experiments. KDB1 and KDB2 are, respectively, the equilibrium dissociation
constants of the first (high affinity) and second (low affinity) binding of ligand to the dimeric A2AR or D2R. Because A2AR is non-cooperative, in this
case, KDB2 is 4KDB1. DC is the dimer cooperativity index for the binding of ligand.37
similar results were observed; heterobivalent ligands
showed higher displacement of the radioligand binding than
that of the corresponding monovalent compounds when D2R
and A2AR are coexpressed, but when only D2R was expressed
were no differences found in the displacement of radioli-
gand [3H]-YM09151-2 binding (Figure 4c,d). The binding of
[3H]-YM09151-2 or [3H]ZM-241385 was not affected by MXn
or MPn monovalent ligands, respectively, in cells expressing
D2R or A2AR, respectively (data not shown).
In conclusion, heterobivalent ligands (4 and 6) showed a
different behavior in both receptors when A2A-D2 receptor
heteromers are present increasing their capacity to displace
the radioligands compared to their corresponding monova-
lent ligands (12, 14, 20, and 22). As we explained before, this

5596 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18
Soriano et al.
Figure 5. cAMP assays. (a) HEK-293 cells expressing the human
A
2AR were treated with 200 nM of the A2AR specific agonist CGS
21680 (CGS) and/or 10 μM of the A2AR antagonist ZM 241385
(ZM), 10 μM 4, or 10 μM 6. Results are presented as % of cAMP
accumulation achieved after treatment with the agonist, CGS
21680. (b) HEK-293 cells expressing the human D2R were treated
with 10 μM forskolin in the presence of: 1 μM of the D2R agonist
(()PPHT, 30 μM of the D2R antagonist raclopride (Racl) plus 1 μM
(()PPHT, 10 μM 4, or 10 μM 6. Results are presented as % of
cAMP accumulation achieved after treatment with forskolin. “Ba-
sal” represents cAMP levels in nonstimulated cells. Results are
mean ( SD of three independent experiments. (Student’s t-test
showed significant cAMP decreases in relation to treatment with
CGS 21680 in (a) or in relation to forskolin treatment in (b): **p <
0.001 and ***p < 0.0001).
Figure 4. Interaction of heterobivalent ligands with A2A-D2 re-
ceptor heteromers. Competition experiments of A2AR selective
antagonist [3H]ZM-241385 (2.1 nM) vs heterobivalent 4 and 6
and monovalent 12 and 14 compounds using membrane prepara-
tions from mouse fibroblast Ltk cells coexpressing A2AR and D2R
(a) or expressing A2AR (b). Competition experiments of D2R
antagonist [3H]YM09151-2 (1.9 nM) vs heterobivalent 4 and 6
and monovalent 20 and 22 compounds using membrane prepara-
tions from mouse fibroblast Ltk cells coexpressing A2AR and D2R
(c) or expressing D2R (d). The concentration of the different
molecules was 1 μM. Values are given as % of control (specific
radioligand binding determined in the absence of displacer). Data
are mean ( SD of three independent experiments. Heterobivalent
ligand was significantly different (*p < 0.05, **p < 0.005, ***p <
0.0005) compared to the corresponding monovalent ligand by
Student’s t test for unpaired samples.
and 6 antagonized CGS21680-induced cAMP production
and therefore they behaved as A2AR antagonists. As antici-
pated, given that the selective antagonist ZM241385 used as
a control had a higher affinity than xanthine congeners and
that the introduction of the linker to the pharmacophore
most likely decreased the affinity, heterobivalent ligands
were less potent than ZM241385. Because D2R couple to
Gi proteins, the efficacy of 10 μM of heterobivalent ligands
to decrease cAMP levels was tested in forskolin-treated cells
(Figure 5b). Compounds 4 and 6 conserved the (()PPHT
agonistic character, causing, respectively, an 87% and a
78%, decrease in the forskolin-induced cAMP levels .
effect was also detected in native tissue that coexpressed both
receptors as brain striatal membranes. However, in the
absence of A2A-D2 receptor heteromers and when both
receptors were in their monomeric or homodimeric form,
no differences were detected between heterobivalent and
monovalent ligands. This fact demonstrates the specific
interaction of these heterobivalent ligands with A2A-D2
receptor heteromers and makes these ligands useful as
pharmacological tools to detect receptor heteromers in
native tissue.
A2AR Antagonist and D2R Agonist Heterobivalent Ligand
Behavior. Bivalent drugs useful for Parkinson’s disease
would act as agonists for D2R and as antagonists for
A2AR. Activated A2AR couples to Gs protein, which leads
to the activation of adenylate cyclase and therefore to an
increase in cAMP levels. On the contrary, D2R activation
results in coupling to Gi proteins, which leads to the inhibi-
tion of adenylate cyclase and as a result to a decrease in
cAMP levels. To find out the functional properties of
heterobivalent ligands, cAMP determination assays were
performed. In these assays, it was not possible to evaluate
Discussion
GPCRs dimer/oligomer formation influences receptor
function and pharmacology and has an important impact
on GPCR drug design.8,10,39 The development of bivalent
ligands that target GPCR dimers is an approach for the study
of GPCR oligomerization and for the design of novel thera-
pies. Several reports have described the application of the
“bivalent ligand” approach for the oligomerization study of a
variety of GPCRs including δ/κ opioid,40,41 5-HT(1B/1D),
and 5-HT(4) homodimers serotonin,42,43 as well as muscari-
nic-acetylcholine44 and melanocortin-4/δ-opioid45 receptors.
In the present study, we synthesized a library of hetero-
bivalent ligands constituted by two pharmacophoric moieties,
XCC (1) as A2AR antagonist and (()-PPHT-NH2 (2) as D2R
agonist, that are connected via Lys-Lys-[PEG/polyamide]n-
Lys-Glu(n=0-7)linkers. These compoundsshoweda greater
capacity to displace specific A2AR and D2R radioligand
binding than their corresponding monovalent controls when
both receptors were present in brain striatum membranes.
Monovalent ligands, containing one pharmacophore with
matched spacers, were tested in brain striatum tissue to
determine whether a change in spacer length influenced ligand
whether the designed bivalent ligands behaved as A2A
R
antagonists and D2R agonists at the same time. For this
reason, human D2R and human A2AR were expressed
separately in HEK cells. Because receptors were singly
expressed, heterobivalent ligands bound to A2AR or to
D2R in a monovalent mode. These ligands did not act as
A2AR agonists because they did not increase basal cAMP
levels (Figure 5a). In these experiments, bivalent ligands 4

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18 5597
binding. In this regard, the incorporation of the linker to the
pharmacophore led to a decrease in the displacement of the
radioligand binding. The capacity of xanthine-derived mono-
valent ligands to displace [3H]-ZM241385 in A2AR decreased
with increasing ligand molecular weight. In contrast, no
correlation was observed between linker length and binding
to D2R of (()PPHT-derived monovalent compounds.
account dissociation constant values (KDB1 and KDB2), com-
pounds 5 and 6 showed the highest binding affinity enhance-
ment (7-18 fold increase) for D2R.
To bear out the specific interaction of the two pharmaco-
phoric parts of the bivalent ligand with the A2A-D2 receptor
heteromer, competition binding assays were performed using
membrane preparation from Ltk cells expressing human
A2AR or human D2R or both receptors at the same time.
The results revealed that it was only when both receptors were
coexpressed that bivalent ligands showed higher displacement
of specific radioligand bindingthanmonovalent ligands. Thus
the high affinity of heterobivalent ligand versus monovalent
ligands is only observed in cells coexpressing both receptors,
indicating a simultaneous interaction of heterobivalent li-
gands with D2R and A2AR. This high affinity is difficult to
be explained without considering a significant concentration
factor exerted by heterobivalent ligands that can be consid-
ered an indicator of A2A-D2 receptor heteromers existence.
This effect was also observed when the binding experiments
were carried out with brain striatum tissue as described above
and could be indicative of the existence of A2A-D2 receptor
heteromers in native brain membranes.
Shorter heterobivalent ligands with linkers Lys-Lys-[PEG/
polyamide]n-Lys-Glu n=0-3 exhibited higher displacement
of specific A2AR radioligand binding. As predicted by pre-
liminary docking experiments (see SI Figure 1), the shortest
linker (26 atoms) is long enough to allow the two pharmaco-
phoric moieties to bind to an A2A-D2 receptor heterodimer.
The estimated maximum inter-CR distance between the two
binding pocketsofthe receptorsthat participate in the dimer is
in the range of 40-45 A. This range is similar to the one
proposed in recent studies for other receptor dimers, including
opioid receptors,40 serotonin 5-HT4 receptor,43 and human
melanocortin receptor 4.46 Our estimations, which had been
based on the complex between compound 3 and a
D2R-A2AR heteromer model previously built,12 found sup-
port in a couple of alternative complexes involving two
alternative heterodimer models (see SI Figures 1-3). The
latter are made by two variances of a novel D2R model based
upon the crystal structure of the β2-adrenergic receptor47 (β2-
AR) (PDB code: 2RH1), on one hand, and the very recently
released crystal structure of A2AR48 (PDB code: 3EML), on
the other one. Although the intermonomer interfaces vary in
the three different heterodimers, the two pharmacophoric
moieties of 3 dock into the same receptor sites. This is
particularly true for the D2R moiety that experiences the H-
bonds with D114 in helix 3 and with either S193 or S197 in
helix 5 (see SI Figures 1-3).
The pharmacological characterization of heterobivalent
ligands with Lys-Lys-[PEG/polyamide]n-Lys-Glu (n= 0-3)
linkers was performed in brain striatum tissue by means of
competition experiments using radiolabeled A2AR or D2R
antagonists and increasing concentrations of heterobivalent
ligands. To calculate receptor affinity for heterobivalent
ligands, the previously described two-state dimer receptor
model was used to fit binding data.37 The KDB1 and KDB2
values reported here are the macroscopic dissociation con-
stants describing the binding of the first and the second ligand
molecule to the homodimeric receptor. While competition
curves of [3H]-YM09151-2 vs heterobivalent ligands 3-6 or
monovalent ligands 19-22 were biphasic, those correspond-
ing to [3H]-ZM241385 vs heterobivalent ligands 3-6 or
monovalent ligands 11-14 were monophasic. The coopera-
tivity index values obtained indicated negative cooperativity
in ligand binding for D2R but not for A2AR.
Taken all together, these results indicate that a heterobiva-
lent ligand bound to D2R may subsequently bind to A2A
R
located in the vicinity with a higher affinity than the corre-
sponding A2AR monovalent ligand. Similarly, a heterobiva-
lent compound bound to A2AR subsequently binds to D2R
located in closer proximity with a higher affinity than its D2R
monovalent counterpart. This in turn indicates a significant
concentration factor exerted by heterobivalent ligands that
can be considered only if both receptors are in a very close
proximity. Thus, dopamine-adenosine heterobivalent li-
gands are useful tools to show the occurrence of these hetero-
mers in brain striatum and the increased affinity of these
ligands detected when both receptors are present is a strong
indication of binding to A2A-D2 receptor heteromers.
Although powerful biophysical techniques such as fluores-
cence resonance energy transfer (FRET) and bioluminescence
resonance energy transfer (BRET) have been decisive in
demonstrating heteromer formation in heterologous expres-
sion systems, the demonstration of occurrence of dimers in
natural tissues is still a matter of controversy. Coimmunopre-
cipitation of two receptors is not considered proof of hetero-
merization due to the fact that the result may be a detergent-
induced membrane protein aggregation process.49 Heterobi-
valent ligands are valuable pharmacological tools that allow
demonstration of the existence of A2A-D2 receptor hetero-
mers even in native tissue and can be used to study a specific
GPCR dimer behavior without any receptor modification.
Furthermore, heterobivalent ligands strategies can be devel-
oped to identify heteromers that involve other neurotrans-
mitter and/or neuromodulator receptors, which may lead to
novel therapeutic approaches to combat a variety of neuro-
logical and neuropsychiatric diseases.
Heterobivalent ligands 3-6 with linker lengths ranging
from 26 to 66 atoms displayed similar binding affinities for
A2AR (between 50 ( 8 and 80 ( 10 nM), which were 3-7
times higher compared to those of their monovalent counter-
parts 11-14. Recent computational modeling has revealed
that similar PEG/polyamide based linkers could assume
multiple conformations, from linear to curved, with a wide
range of distances.46 The high flexibility of the Lys-Lys-[PEG/
polyamide]n-Lys-Glu linker likely explains the lack of correla-
tion between linker length and binding affinity. For D2R,
heterobivalent ligands 3-6 exhibited higher binding affinities
(1.0 ( 0.2 to 1.8 ( 0.4 nM for KDB1) compared to that of
their corresponding monovalent ligands 19-22 but again
showing no correlation with linker length. Taking into
In summary, in the present work, we designed and synthe-
sized heterobivalent ligands that serve as a probe for A2AD2
receptor heteromers in native tissues and could open new
avenues for the design of heteromer-selective drugs for treat-
ment of Parkinson’s disease.
Experimental Section
Materials and Equipment. All chemical reagents were ob-
tained from commercial suppliers and used without further

5598 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18
Soriano et al.
purification. Melting points were determined on a Buchi melting
point B-549 apparatus. 1H and 13C NMR spectra were recorded
at 298 K on a Varian Mercury-400 Fourier transform spectro-
meter. Chemical shifts are reported in δ units (ppm) relative to
the residual deuterated solvent signals of CHCl3 (1H NMR:
δ 7.26; 13C NMR: δ 77.0), DMSO (1H NMR: δ 2.49; 13C NMR:
δ 39.5). The splitting patterns are designated as follows: s
(singlet), d (doublet), t (triplet), q (quartet), b (broad). The
RP-HPLC analyses were performed on a Waters Alliance
instrument and RP-HPLC-MS on a Waters Alliance instrument
coupled to a Micromass ZQ spectrometer with an electrospray
(ES) probe. The purifications by preparative RP-HPLC were
performed on a Waters HPLC autopurification FractionLynx
UV/MS system with an electrospray (ES) probe. Analytical
thin-layer chromatography (TLC) was performed on precoated
plates (Merck silica gel 60ACC, F254). Visualization of the
developed chromatogram was achieved with UV light. Manual
flash column chromatography was performed using silica
(Merck, 70-230 mesh). Automated flash chromatography
was performed on a Teledyne Isco module companion with
photodiode array detector using Silica-RediSep columns.
Dulbecco’s modified Eagle’s medium (DMEM), sodium pyr-
uvate, L-glutamine, penicillin/streptomycin, FBS, and Lipofec-
tamine 2000 were purchased from Invitrogen (Grand Island,
NY). Adenosine deaminase (ADA) EC 3.5.4.4 was supplied by
Roche (Basel, Switzerland), and bicinchoninic acid method was
supplied by Pierce Chemical Co (Rockford, IL). Protease
inhibitor cocktail, raclopride, (()-PPHT, and forskolin were
purchased from Sigma (St. Louis, MO). [3H]- ZM 241385 was
purchased from American Radiolabeled Chemicals (St. Louis,
MO). ZM241385, CGS 21680, and zardaverine were supplied by
Tocris Biosciences (Avonmouth, UK). [3H]-YM 09151-2 was
supplied by PerkinElmer (Boston, MA). Ecoscint H scintillation
cocktail was purchased from National Diagnostics (Atlanta,
GA) and Cyclic AMP (3H) Assay system from Amersham
Biosciences (Uppsala, Sweden). Membranes were homogenized
with Polytron homogenizer (PTA 20 TS rotor, setting 3; Kine-
matica, Basel, Switzerland). Radioligand binding experiments
were performed using a Brandel cell harvester (Gaithersburg,
MD) and a Packard 1600 Tri-Carb scintillation counter. GRA-
FIT was supplied by Erithacus Software (Surrey, UK).
column, gradient 5% to 60% B in 7 min; A: H2O-HCOOH
(99.9:0.1); B: CH3CN-HCOOH (99.3:0.7). RP-HPLC conditions
(methodB):X-TerraMSC18, (3.5μm, 4.6 mm ꢀ 100 mm) column,
gradient 0% to 70% B in 10 min; A: H2O-TFA (99.9:0.1); B:
CH3CN-TFA (99.9:0.1). Method A has been used to determine
the purity of final compounds. HPLC data and purities of
compounds 3-26 are included in Table 1 (SI).
Purification by Preparative RP-HPLC-MS. Monovalent and
heterobivalent ligands were purified by preparative RP-HPLC-
MS using different linear gradients of H2O (containing
0.1% HCOOH) and CH3CN (containing 0.07% HCOOH) at a
flow rate of 25 mL/min. Column: symmetry C18, 5 μm, 30 mm ꢀ
100 mm.
General Protocol for the Synthesis of Monovalent Ligands
(11-26) and Heterobivalent (3-10) Ligands. Fmoc-NH-AM-
MBHA resin (5 g, f=0.61 mmol/g) was previously swelled with
DCM (1 ꢀ 1 min, 2 ꢀ 10 min) and DMF (5 ꢀ 1 min, 1 ꢀ 15 min)
before use and treated with piperidine as described before. After
washings, the coupling mixture of Fmoc-L-Lys(Alloc)-OH
(3 equiv), HOBt (3 equiv), and DIPCDI (3 equiv) in DMF
was added to the resin and the resulting mixture was stirred for 2
h. The resin was washed and the Fmoc group removed yielding
NH2-L-Lys(Alloc)-AM-MBHA. At this point, a portion of resin
(450 mg) was separated and divided on three aliquots of 150 mg
to continue the synthesis of monovalent (11, 19) and hetero-
bivalent (3) ligands corresponding to spacer length n=0.
The linker Fmoc-PEG-based unit (2.5 equiv, 39) was intro-
duced using PyBOP (2.5 equiv) and DIEA (5 equiv) and HOAt
(2.5 equiv) as coupling reagents in DMF for 5 h at 25 ꢀC. An
acylation step was performed when the coupling was not
completed. The three-step cycle (Fmoc elimination, Fmoc-
PEG-based unit coupling, acylation) was repeated seven times,
adding a repeating monomeric unit to the oligomeric chain.
After each cycle, a portion of oligomeric resin (150-200 mg) was
separated to continue the synthesis of the different lengths of
monovalent and heterobivalent ligands (from n=1 to n=7).
Each aliquot of resin corresponding to each monovalent
or heterobivalent precursor was washed and a mixture of
Ac-L-Lys(Fmoc)-OH (3 equiv), PyBOP (3 equiv), HOAt
(3equiv), and DIEA (6 equiv) in DMF was added. The coupling
mixtures were then stirred for 4 h at 25 ꢀC. The resins were
washed and the ε-NH-Fmoc group was then removed by
following the procedure described previously. From this point
onward, monovalent and heterobivalent ligand synthesis was
continued separately.
Synthesis of XCC Monovalent Ligands (11-18). Boc-L-Lys-
(Fmoc)-OH (2.1 g, 3 equiv), PyBOP (3 equiv), HOAt (3 equiv),
and DIEA (6 equiv) were dissolved in DMF and added to each
XCC monovalent ligand precursor resin (n =0-7). Once the
coupling was finished, the Fmoc group was removed in each case
as described before and 1 (3equiv), PyBOP (3 equiv), HOAt
(3 equiv), and DIEA (6 equiv) in DMF were added. The mixture
was then stirred for 4 h. The ε-NH-Alloc group was then
eliminated as described previously followed by acetylation by
treatment with Ac2O (5 equiv) and DIEA (10 equiv) in DCM for
30 min. Finally, ligands 11-18 were cleaved from their corre-
sponding resins with TFA:H2O (95:5, v/v, 2 mL) for 1 h. The
solvent was evaporated to dryness. Compounds 11-18 were
purified by preparative RP-HPLC-MS using a linear gradient
10-50% CH3CN for 30 min. Purities of all final compounds
were g95% (see SI, Table 1).
Solid-Phase Synthesis. All solid-phase syntheses were carried
out manually in a polypropylene syringe fitted with a polyethy-
lene porous disk. Solvents and soluble reagents were removed by
suction. Peptide synthesis for this work employed a combined
Fmoc/Alloc/tBu solid phase strategy on a Fmoc-AM-MBHA
resin. Washings between deprotection, coupling, and subse-
quent deprotection steps were carried out with DMF (5 ꢀ
1 min) and DCM (5 ꢀ 1 min) using 10 mL of solvent/g of resin
each time. All the couplings and Fmoc removal were monitored
using the Kaiser test.
Fmoc Group Removal. Fmoc group removal involved the
following sequence: (i) DMF (5 ꢀ 1 min), (ii) piperidine-DMF
(1:1) (1 ꢀ 1 min þ 2 ꢀ 15 min), (iii) DMF (5 ꢀ 1 min). In the cases
where the Kaiser test was not a clear positive, an additional
treatment with DBU:piperidine:toluene:DMF (5:5:20:70) (1 ꢀ
10 min) was performed.
Alloc Group Removal. Alloc group removal involved the
following sequence: (i) Pd(PPh3)4 (0.1 equiv) and PhSiH3
(10 equiv) in anhydrous DCM (3 ꢀ 15 min), (ii) anhydrous
DCM (5 ꢀ 1 min), (iii) DCM (5 ꢀ 1 min), (iv) DMF (5 ꢀ 1 min),
and (v) 0.02 M solution of sodium diethyldithiocarbamate in
DMF (3 ꢀ 15 min), DMF (5 ꢀ 1 min), DCM (5 ꢀ 1 min), and
DMF (5 ꢀ 1 min).
(11). Amount of crude product: 21.4 mg. Purification gave
4.1 mg of pure 11. (ESþ) calcd for C41H63N11O9, 853.48; found
[M þ H]=854.7, [M þ 2H/2]=427.5.
(12). Amount of crude product: 31.5 mg. Purification gave
17 mg of pure 12. (ESþ) calcd for C56H91N13O14, 1083.6; found
[M þ H]=1084.8, [M þ 2H/2]=543.1.
Analysis by RP-HPLC and RP-HPLC-MS. Crudes were
analyzed by RP-HPLC-MS. The purified compounds were
characterized by analytical RP-HPLC and RP-HPLC-MS using
two different reverse-phase C18 columns and flow 1 mL/min in
two separate elution systems at 254 nm. RP-HPLC-MS condi-
tions (method A): symmetry C18 (3.5 μM, 4.6 mm ꢀ 75 mm,)
(13). Amount of crude product: 40 mg. Purification gave
3.0 mg of pure 13. (ESþ) calcd for C71H119N15O19, 1313.7; found
[M þ H]=1315.0, [M þ 2H/2]=658.2.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18 5599
(14). Amount of crude product: 23.6 mg. Purification gave
(24). Amount of crude product: 95 mg. Purification gave
0.4 mg of pure 24. (ESþ) calcd for C92H154N18O27, 1943.12;
found [M þ H]=1945.5, [M þ 2H/2]=973.5, [M þ 3H/3]=
649.4, [M þ 4H/4]=487.4.
2.1 mg of pure 14. (ESþ) calcd for C86H147N17O24, 1544.79;
found [M þ H]=1546.2, [M þ 2H/2]=773.5, [M þ 2H/2]=
516.0.
(15). Amount of crude product: 48 mg. Purification gave
17.0 mg of pure 15. (ESþ) calcd for C102H177N19O28, 1775.1;
found [M þ H]=1777.0, [M þ 2H/2]=888.7, [M þ 3H/3]=592.9.
(16). Amount of crude product: 54 mg. Purification gave 29.3
mg of pure 16. (ESþ) calcd for C116H203N21O34, 2005.31; found
[M þ H]=2007.1, [M þ 2H/2]=1003.9, [M þ 3H/3]=669.6.
(17). Amount of crude product: 63 mg. Purification gave pure
23.7 mg of 17. (ESþ) calcd for C131H231N23O39, 2236.06; found
[M þ 2H/2]=1119.2, [M þ 3H/3]=746.5.
(25). Amount of crude product: 103 mg. Purification gave
0.4 mg of pure 25. (ESþ) calcd for C102H172N20O31, 2173.25;
found [M þ 2H/2]=1088.4, [M þ 3H/3]=725.9, [M þ 4H/4]=
544.8, [M þ 4H/4]=436.1.
(26). Amount of crude product: 69 mg. Purification gave
2.7 mg of pure 26. (ESþ) calcd for C112H190N22O35, 2403.38;
found [M þ 2H/2]=1203.0, [M þ 3H/3]=802.4, [M þ 4H/4]=
602.3, [M þ 4H/4]=482.2.
Synthesis of XCC-PPHT Heterobivalent Ligands (3-10).
Boc-L-Lys(Fmoc)-OH (3 equiv), HOBt (3 equiv), and DIPCDI
(3 equiv) in DMF were added to each heterobivalent ligand
precursor resin (n=0-7), and the resulting mixture was stirred
for 2 h. After washings, the Fmoc group was removed in each
case, 1 (3 equiv), PyBOP (3 equiv), HOAt (3 equiv), and DIEA
(6 equiv) were added, and the mixture stirred for 4 h. After
elimination of ε-NH-Alloc, a mixture of Fmoc-L-Glu(γ-(()-
PPHT-NH2)-OH (3 equiv), PyBOP (3 equiv), HOAt (3 equiv),
and DIEA (6 equiv) in DMF was added to each heterobivalent
ligand precursor resin. Once the reaction was finished, the Fmoc
group was eliminated and subsequently followed by cleavage of
ligands 3-10 from their resins with a mixture of TFA:H2O
(95:5, v/v, 2 mL) for 1 h. The solvent was evaporated to dryness.
Products 3-10 were purified by preparative HPLC-MS using a
linear gradient 0-50% CH3CN in 30 min. Purities of all final
compounds were g95% (see SI, Table 1).
(18). Amount of crude product: 27.5 mg. Purification gave
pure 1.2 mg of 18. (ESþ) calcd for C146H259N25O44, 2465.83;
found [M þ 2H/2]=1234.04, [M þ 3H/3]=823.05, [M þ 4H/4]=
617.5.
Synthesis of Fmoc-L-Glu(γ-(()-PPHT-NH2)-OH (40). HOBt
(1 equiv) and DIPCDI (1 equiv) were added to a solution of
Fmoc-L-Glu-OtBu (642 mg) in DMF (10 mL). The reaction
mixture was stirred for 5 min. Next, (()-PPHT-NH2 (2) (1equiv)
was added and the reaction mixture was stirred for an additional
18 h. The solvent was concentrated in vacuo and the obtained
solid was dissolved in DCM (25 mL) and washed with 5%
NaHCO3(aq) (3 ꢀ 50 mL). The organic layer was dried over
MgSO4 and evaporated to dryness to yield Fmoc-L-Glu(γ-(()-
PPHT-NH2)-OtBu (1.145 g, yield = 95%, purity 92% (RP-
HPLC-MS conditions). (ESþ) calcd for C45H53N3O6, 731.39;
found [M þ H]=732.45.
(3). Amount of crude product: 171 mg. Purification gave
20.8 mg of pure 3. (ESþ) calcd for C65H94N14O11, 1246.72; found
[M þ H]=1248.7, [M þ 2H/2]=624.9, [M þ 3H/3]=417.0.
(4). Amount of crude product: 64.0 mg. Purification gave
12.1 mg of pure 4. Purity: 99%. (ESþ) calcd for C75H112N16O15,
1476.85; found [M þ H]=1479.4, [M þ 2H/2]=740.0, [M þ
3H/3]=493.8.
(5). Amount of crude product: 84.0 mg. Purification gave
4.9 mg of pure 5. (ESþ) calcd for C85H130N18O19, 1706.98; found
[M þ H]=1708.7, [M þ 2H/2]=855.0, [M þ 3H/3]=570.4, [M þ
4H/4]=428.2.
Fmoc-L-Glu(γ-(()-PPHT-NH2)-OtBu (1.145 g) was sus-
pended in 4 M HCl in dioxane (15 mL) and stirred for 4 h.
The reaction mixture was concentrated in vacuo and coevapo-
rated three times with dioxane and three times with H2O:ACN
(1:1, v/v) to yield Fmoc-L-Glu(γ-(()-PPHT-NH2)-OH as a
brown solid (1.031 g, yield=98%). The purity of 40 was 95%
(RP-HPLC-MS conditions). (ESþ) calcd for C41H45N3O6,
675.33; found [M þ H]=676.33. The product was used without
any further purification and characterization.
Synthesis of PPHT Monovalent Ligands (19-26). Each ali-
quot of resin that corresponded to each PPHT monovalent
ligand (n = 0-7) were acetylated by treatment with Ac2O
(10 equiv) and DIEA (10 equiv) in DCM for 30 min. After that,
the ε-NH-Alloc group was eliminated and a mixture of Fmoc-L-
Glu(γ-(()-PPHT-NH2)-OH (3 equiv), HOBt (3 equiv), and
DIPCDI (3 equiv) in DMF was added to each PPHT monova-
lent ligand precursor resin (n=0-7). Finally, after Fmoc group
removal, ligands 19-26 were cleaved from their corresponding
resins with a mixture of TFA:H2O (95:5, v/v, 2 mL) for 1 h. The
solvent was evaporated to dryness. Products 19-26 were pur-
ified by preparative C18 RP-HPLC-MS using a linear gradient
0-50% CH3CN in 30 min. Purities of all final compounds were
g95% (see SI, Table 1) with the exception of compounds 21
(90%) and 26 (91%).
(6). Amount of crude product: 63.1 mg. Purification gave
5.8 mg of pure 6. (ESþ) calcd for C95H148N20O23, 1937.10; found
[M þ H]=1940.0, [M þ 2H/2]=970.2, [M þ 3H/3]=647.1, [M þ
4H/4]=485.8.
(7). Amount of crude product: 65.9 mg. Purification gave
4.8 mg of pure 7. (ESþ) calcd for C105H166N22O27, 2167.23;
found [M þ H]=2169.6, [M þ 2H/2]=1085.2, [M þ 3H/3]=
723.9, [M þ 4H/4]=543.2.
(8). Amount of crude product: 103.2 mg. Purification gave
2.5 mg of pure 8. (ESþ) calcd for C115H184N24O31, 2397.36;
found [M þ 2H/2]=1200.4, [M þ 3H/3]=800.7, [M þ 4H/4]=
600.8, [M þ 5H/5]=481.0.
(9). Amount of crude product: 106.0 mg. Purification gave
2.6 mg of pure 9. (ESþ) calcd for C125H202N26O35, 2627.48;
found [M þ 2H/2]=1315.6, [M þ 3H/3]=877.4, [M þ 4H/4]=
658.3, [M þ 5H/5]=527.1, [M þ 6H/6]=439.4.
(19). Amount of crude product: 76 mg. Purification RP-
HPLC-MS gave 2.3 mg of pure 19. (ESþ) calcd for
C42H64N8O7, 792.49; found [M þ H] = 793.8, [M þ 2H/2] =
397.5.
(10). Amount of crude product: 125.3 mg. Purification gave
1.1 mg of pure 10. (ESþ) calcd for C135H220N28O39, 2857.61;
found [M þ 2H/2]=1431.2, [M þ 3H/3]=954.2, [M þ 4H/4]=
715.9, [M þ 5H/5]=573.1, [M þ 6H/6]=477.9.
(20). Amount of crude product: 55 mg. Purification gave
3.6 mg of pure 20. (ESþ) calcd for C52H82N10O11, 1022.62;
found [M þ H]=1024.0, [M þ 2H/2]=512.8.
Docking Experiments. The D2-A2A receptor heterodimer in
the present computational experiments belongs to the cluster of
docking solutions achieved in previous studies.12 It resembles
the semiempirical model of dimeric rhodopsin.50 In this model,
the heterodimer interface is mainly formed by the second
intracellular loop (I2), helices 4, 3, and 5 from D2R, and I2,
helices 5, 3, and 4 from A2AR. Compound 3 was manually
docked into the putative binding sites of the two receptors
by probing slightly different conformations of the ligand
and the receptor side chains. The docking criteria were aimed
(21). Amount of crude product: 85 mg. Purification gave
3.2 mg pure 21. (ESþ) calcd for C62H100N12O15, 1252.74; found
[M þ H]=1254.1, [M þ 2H/2]=627.9, [M þ 3H/3]=418.9.
(22). Amount of crude product: 47 mg. Purification gave
4.9 mg of pure 22. (ESþ) calcd for C72H118N14O19, 1482.87;
found [M þ H]=1484.7, [M þ 2H/2]=743.1, [M þ 3H/3]=
495.8, [M þ 4H/4]=372.1.
(23). Amount of crude product: 60 mg. Purification gave
0.7 mg of pure 23. (ESþ) calcd for C82H136N16O23, 1713.00; found
[M þ H]=1715.5, [M þ 2H/2]=858.2, [M þ 3H/3]=572.5.

5600 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18
Soriano et al.
at: (a) allowing the formation of salt bridge interactions between
all the three protonated amino groups of the ligand and anionic
sites of the dimeric receptor, (b) causing all the polar atoms of
the ligand to be engaged in H-bonding interactions, and (c)
allowing for the accomplishment of a few critical interactions
between the dopamine part of the ligand and the D2R. These
interactions included: the charge-reinforced H-bond between
(a) the protonated nitrogen atom of PPTH moiety and D114 in
helix 3 of D2R, and (b) an H-bond between the hydroxyl group
of PPTH moiety and S197 in helix 5 of D2R. The probed
complexes between the receptor heterodimer and the bivalent
ligand were subjected to energy refinement through 100 ps of
balanced Molecular Dynamics simulations, following the same
computational protocol employed to refine monomeric D2R
and A2AR models.12 For each complex, the structures averaged
over the 200 structures collected during the last 100 ps of the
balanced MD trajectory were minimized and subjected to
structural analysis. As a last step, the complex characterized
by the best ligand-receptor complementarity was finally se-
lected. We considered also two alternative heterodimers made
by a β2-AR-based model of D2R (PDB code: 2RH1)47 and the
crystal structure of A2AR48 (PDB code: 3EML) (see SI Figures 2
and 3). Predictions of the two heterodimers were based on a well-
established protocol.51,52 Refinement of the D2R and of the
heteromer receptors-3 complexes was carried out by employing
the GBSW implicit water/membrane water implemented in
CHARMM.53
Screenings were performed in the absence or presence of non-
labeled compounds at a unique concentration of 1 μM and for
displacement curves in the absence or presence of increasing
concentrations of nonlabeled compounds (triplicates of 11
different displacer concentrations from 0.1 nM to 50 μM).
Nonspecific binding was determined in the presence of 50 μM
ZM-241385 for A2AR or 50 μM YM-09151-2 for D2R and, in
competition experiments, it was confirmed that the value was
the same as calculated by extrapolation of the competition
curves. Free and membrane-bound ligand were separated by
rapid filtration of 500 μL aliquots in a Brandel cell harvester
trough Wathman GF/C filters embedded in 0.3% polyethyleni-
mine. Filters were washed in 5 mL of ice-cold Tris-HCl buffer
(50 mM, pH 7.4) and transferred in vials containing 10 mL of
Ecoscint H scintillation cocktail. After overnight shaking,
radioactivity counts were determined using a Packard 1600
TRI-CARB scintillation counter with an efficiency of 62%.54
Binding data from competition experiments were analyzed by
nonlinear regression using the commercial Grafit curve-fitting
software by fitting the binding data to the two-state dimer
receptor model.35,36 To calculate the macroscopic equilibrium
dissociation constants involved in the binding of the compounds
ꢀ
to the dimer as a whole, the new equations deduced by Casado et
al. were employed.37 The macroscopic equilibrium dissociation
constants involved in the binding of the agonists were calculated
following eq 137 below:
Abound ¼ ðKDA2A þ 2A2þKDA2AB=KDABÞRT=ðKDA1KDA2
þKDA2A þ A2þKDA2AB=KDAB þ KDA1KDA2B=KDB1
Cell Culture and Transfection. Human embryonic kidney-293
(HEK-293) cell line and mouse fibroblast Ltk cell line were
cultured in Dulbecco’s modified Eagle’s medium (DMEM) with
4.5 mg/mL glucose and 0.11 mg/mL sodium pyruvate supple-
mented with 10% fetal bovine serum, 100 units/mL penicillin,
100 μg/mL streptomycin, and 2 mM L-glutamine at 37 ꢀC in a
humidified atmosphere of 5% CO2. For binding experiments,
mouse fibroblast Ltk cells were seeded in 150 mm dishes and
transiently transfected with cDNA corresponding to human
þ KDA1KDA2B2=ðKDB1KDB2ÞÞ
ð1Þ
where A represents the radioligand (the dopamine D2 receptor
antagonist [3H]YM-09151-2 or the adenosine A2A receptor
antagonist [3H]ZM 241385) concentration, RT is the total
amount of receptor dimers, and KDA1 and KDA2 are the macro-
scopic dissociation constants describing the binding of the first
and the second radioligand molecule (A) to the dimeric receptor.
B represents the assayed competing compound concentration,
and KDB1 and KDB2 are, respectively, the equilibrium dissocia-
tion constants of the first and second binding of B; KDAB can be
described as a hybrid equilibrium radioligand/competitor dis-
sociation constant, which is the dissociation constant of B
binding to a receptor dimer semioccupied by A.
A
2AR or human D2R (long isoform) or both cDNAs at the same
time. For cAMP experiments, HEK-293 cells were grown in
25 cm2 flasks and transiently transfected with the human cDNA
encoding for A2AR or D2R. Tranfections were performed using
the Lipofectamine 2000 reagent kit according to the manufac-
turer’s instructions. HEK-293 cells transiently expressing A2AR
or D2 LR were incubated in serum-free medium in the presence
of 1.5 U/mL ADA for cells expressing A2AR, 16 h before the
experiment.
Membrane Preparation and Protein Determination. Mem-
brane suspensions from lamb brain striatum were prepared as
described previously.54 Tissue was disrupted with a Polytron
homogenizer for three 10 s periods in 10 volumes of ice-cold
Tris-HCl buffer (50 mM, pH 7.4), containing a protease inhi-
bitor cocktail. Membranes were obtained by centrifugation at
105000g for 40 min at 4 ꢀC, and the pellet was resuspended and
recentrifuged under the same conditions. The resulting pellet
was stored at -80 ꢀC and was washed once more as described
above and resuspended in Tris-HCl buffer (50 mM, pH 7.4) for
immediate use. Protein was quantified by the bicinchoninic acid
method using bovine serum albumin dilutions as the standard.
Mouse fibroblast Ltk cells were lifted from dishes with a cell
scraper 48 h after transfection and harvested by centrifugation
at 1500g for 5 min. Cell pellet was washed twice with PBS and
resuspended in 10 volumes of ice-cold Tris-HCl buffer (50 mM,
pH 7.4). Cell suspension was disrupted with a Polytron homo-
genizer for three 10 s periods, and the homogenates were
processed as described above for striatum membranes.
Because the radioligand A (the dopamine D2 receptor
antagonist [3H]YM-09151-2 or the adenosine A2A receptor
antagonist [3H]ZM 241385) showed noncooperative behavior,
37
eq 1 was simplified to eq 2 due to the fact that KDA2=4KDA1
:
Abound ¼ ð4KDA1A þ 2A2 þ 4KDA1AB=KDABÞRT=ð4KDA1
2
þ 4KDA1A þ A2 þ 4KDA1AB=KDAB þ 4KDA12B=KDB1
þ 4KDA12B2=ðKDB1KDB2ÞÞ
The dimer cooperativity index for the competing ligand B37
ð2Þ
was calculated following eq 3:37
DC ¼ logð4KDB1=KDB2
Þ
ð3Þ
DC would be a measure of ligand homotropic cooperativity. It
measures the orthosteric dissociation equilibrium constant va-
lue modifications occurring when a protomer senses the binding
of the same ligand molecule to the partner protomer (in a dimer).
The way the index is defined is such that its value is “0” when the
binding of the first molecule of ligand to one protomer of the
empty dimer does not affect the binding of the second molecule
to the empty protomer in the dimer. Positive or negative values
of Dc indicate that the presence of the first bound molecule
increases or decreases respectively the affinity for binding of the
second molecule to the empty protomer in the dimer.37
Radioligand Binding Experiments. Competition experiments
were performed by incubating (2 h) membranes (0.3-0.5 mg of
protein/mL) at 25 ꢀC in Tris-HCl buffer (50 mM, pH 7.4)
containing 10 mM MgCl2 and 2 U/mL adenosine deaminase
(ADA) with the indicated concentrations of the A2AR antago-
nist [3H]ZM-241385 or the D2R or antagonist [3H]YM-09151-2.
Goodness of fit was tested according to a reduced χ2
value given by the nonlinear regression program. The test of

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18 5601
significance for two different population variance models was
based upon the F distribution.54 Using this F test, a probability
greater than 95% (p < 0.05) was considered the criterion to
select a more complex model (cooperativity) over the simplest
one (noncooperativity). In all cases, a probability of less than
70% (p > 0.30) resulted when one model was not significantly
better than the other.
(10) Milligan, G. G-protein-coupled receptor heterodimers: pharma-
cology, function and relevance to drug discovery. Drug Discovery
Today 2006, 11, 541–549.
ꢀ
(11) Hillion, J.; Canals, M.; Torvinen, M.; Casado, V.; Scott, R.;
Terasmaa, A.; Hansson, A.; Watson, S.; Olah, M. E.; Mallol, J.;
Canela, E. I.; Zoli, M.; Agnati, L. F.; Ibanez, C. F.; Lluıs, C.;
´
ꢀ
Franco, R.; Ferre, S.; Fuxe, K. Coaggregation, cointernalization,
and codesensitization of adenosine A2A receptors and dopamine
D2 receptors. J. Biol. Chem. 2002, 277, 18091–18097.
cAMP Determination. HEK-293 cells, 48 h after transfection,
were preincubated with 50 μM zardaverine as phosphodiester-
ase inhibitor for 10 min at 37 ꢀC in serum-free medium contain-
ing 10 mM MgCl2 and 1.5 U/mL ADA (for A2AR). The ligands
were added sequentially at the concentrations indicated: 10 min
antagonists, 10-15 min agonists, and 15 min forskolin.
To stop the reaction, the cells were placed on ice, detached,
and washed twice in ice-cold PBS. After centrifugation at 2500g
for 5 min at 4 ꢀC, the pellet was resuspended with 200 μL ice-cold
HClO4 (4%) for 30 min and 1.5 M KOH was added to reach
neutral pH. Samples were centrifuged at 15000g for 30 min at
4 ꢀC and the supernatant was frozen at -20 ꢀC. The accumula-
tion of cAMP in the samples was measured by a [3H] cAMP
assay system as described in the manual from the manufacturer.
(12) Canals, M.; Marcellino, D.; Fanelli, F.; Ciruela, F.; de Benedetti,
P.; Goldberg, S. R.; Neve, K.; Fuxe, K.; Agnati, L. F.; Woods, A.
ꢀ
S.; Ferre, S.; Lluıs, C.; Bouvier, M.; Franco, R. Adenosine
´
A2A-dopamine D2 receptor-receptor heteromerization: quali-
tative and quantitative assessment by fluorescence and bio-
luminescence energy transfer. J. Biol. Chem. 2003, 278, 46741–
46749.
(13) Kamiya, T.; Saitoh, O.; Yoshioka, K.; Nakata, H. Oligomerization
of adenosine A2A and dopamine D2 receptors in living cells.
Biochem. Biophys. Res. Commun. 2003, 306, 544–549.
ꢀ
(14) Agnati, L. F.; Ferre, S.; Lluıs, C.; Franco, R.; Fuxe, K. Pharmacol.
´
Rev. 2003, 55, 509–550.
ꢀ
(15) Ferre, S.; Fredholm, B. B.; Morelli, M.; Popoli, P.; Fuxe, K.
Adenosine-dopamine receptor-receptor interactions as an inte-
grative mechanism in the basal ganglia. Trends Neurosci. 1997, 20,
482–487.
ꢀ
~
(16) Ferre, S; Ciruela, F; Woods, A. S.; Canals, M; Burgueno, J;
Marcellino, D; Karcz-Kubicha, M; Hope, B. T.; Morales, M;
Popoli, P; Goldberg, S. R.; Fuxe, K; Lluıs, C; Franco, R; Agnati,
´
L. F. Glutamate mGlu5-adenosine A2A-dopamine D2 receptor
interactions in the striatum. Implications for drug therapy in
neuropsychiatric disorders and drug abuse. Curr. Med. Chem.
2003, 3, 1–26.
Acknowledgment. We acknowledge the technical help
ꢀ
Laboratory, Barcelona University). This work was supported
by Grants from Spanish Ministery of Science and Innovation
(SAF2005-00170 and SAF2006-05481 (R.F.) and CTQ2005-
00315/BQU and CTQ2008-00177/BQU (M.R.)), grant
obtained from Jasmina Jimenez (Molecular Neurobiology
(17) Schiffmann, S. N.; Jacobs, O.; Vanderhaeghen, J. J. Striatal
restricted adenosine A2 receptor (RDC8) is expressed by enkepha-
lin but not by substance P neurons: an in situ hybridization
histochemistry study. J. Neurochem. 1991, 57, 1062–1067.
(18) Fink, J. S.; Weaver, D. R.; Rivkees, S. A.; Peterfreund, R. A.;
Pollack, A. E.; Adler, E. M.; Reppert, S. M. Molecular cloning of
the rat A2 adenosine receptor: selective co-expression with D2
dopamine receptors in rat striatum. Brain Res. Mol. Brain Res.
1992, 14, 186–195.
ꢀ
ꢀ
060110 from Fundacio La Marato de TV3 (R.F.) and CI-
BERBBN (F.A.) and CIBERNED (R.F.) from Instituto de
Salud Carlos III. R.H. thanks The Netherlands Organization
for Scientific Research for their financial support.
Supporting Information Available: Molecular docking stu-
dies, experimental description of the synthesis and characteriza-
tion of compounds 1, 2, and 39 and its intermediates, table
containing 3-26 compound purities determined by HPLC, 1H,
and 13C NMR spectra and HPLC chromatograms of com-
pounds 1, 2, and 39 and its intermediates, and 3-26 compound
HPLC chromatograms. This material is available free of charge
via the Internet at http://pubs.acs.org.
ꢀ
~
(19) Ferre, S.; Ciruela, F.; Canals, M.; Marcellino, D.; Burgueno, J.;
ꢀ
Casado, V.; Hillion, J.; Torvinen, M.; Fanelli, F.; de Benedetti, P.;
Goldberg, S. R.; Bouvier, M.; Fuxe, K.; Agnati, L. F.; Lluıs, C.;
´
Franco, R.; Woods, A. Adenosine A2A-dopamine D2 receptor-
receptor heteromers. Targets for neuropsychiatric disorders. Par-
kinsonism Relat. Disord. 2004, 10, 265–271.
ꢀ
(20) Fuxe, K.; Ferre, S.; Genedani, S.; Franco, R.; Agnati, L. F.
Adenosine receptor-dopamine receptor interactions in the basal
ganglia and their relevance for brain function. Physiol. Behav. 2007,
92, 210–217.
ꢀ
(21) Ferre, S.; Fuxe, K.; von, E. G.; Johansson, B.; Fredholm, B. B.
References
Adenosine-dopamine interactions in the brain. Neuroscience
1992, 51, 501–512.
(1) Bouvier, M. Oligomerization of G-protein-coupled transmitter
receptors. Nat. Rev. Neurosci. 2001, 2, 274–286.
(2) Park, P. S.; Filipek, S.; Wells, J. W.; Palczewski, K. Oligomeriza-
tion of G protein-coupled receptors: past, present, and future.
Biochemistry 2004, 43, 15643–15656.
(22) Bara-Jimenez, W.; Sherzai, A.; Dimitrova, T.; Favit, A.; Bibbiani,
F.; Gillespie, M.; Morris, M. J.; Mouradian, M. M.; Chase, T. N.
Adenosine A(2A) receptor antagonist treatment of Parkinson’s
disease. Neurology 2003, 61, 293–296.
(23) Hauser, R. A.; Hubble, J. P.; Truong, D. D. Randomized trial of
the adenosine A2A receptor antagonist istradefylline in advanced
PD. Neurology 2003, 61, 297–303.
(3) Fuxe, K.; Canals, M.; Torvinen, M.; Marcellino, D.; Terasmaa, A.;
Genedani, S.; Leo, G.; Guidolin, D.; az-Cabiale, Z.; Rivera, A.;
Lundstrom, L.; Langel, U.; Narvaez, J.; Tanganelli, S.; Lluıs, C.;
´
(24) Schwarzschild, M. A.; Agnati, L.; Fuxe, K.; Chen, J. F.; Morelli,
M. Targeting adenosine A2A receptors in Parkinson’s disease.
Trends Neurosci. 2006, 29, 647–654.
(25) Jacobson, K. A.; Xie, R.; Young, L.; Chang, L.; Liang, B. T. A
novel pharmacological approach to treating cardiac ischemia.
Binary conjugates of A1 and A3 adenosine receptor agonists. J.
Biol. Chem. 2000, 275, 30272–30279.
(26) Daniels, D. J.; Lenard, N. R.; Etienne, C. L.; Law, P. Y.; Roerig, S.
C.; Portoghese, P. S. Opioid-induced tolerance and dependence in
mice is modulated by the distance between pharmacophores in a
bivalent ligand series. Proc. Natl. Acad. Sci. U.S.A. 2005, 102,
19208–19213.
ꢀ
Ferre, S.; Woods, A.; Franco, R.; Agnati, L. F. Intramembrane
receptor-receptor interactions: a novel principle in molecular
medicine. J. Neural Transm. 2007, 114, 49–75.
ꢀ
ꢀ
(4) Carriba, P.; Navarro, G.; Ciruela, F.; Ferre, S.; Casado, V.; Agnati,
ꢀ
L.; Cortes, A.; Mallol, J.; Fuxe, K.; Canela, E. I.; Lluıs, C.; Franco,
´
R. Detection of heteromerization of more than two proteins by
sequential BRET-FRET. Nat. Methods 2008, 5, 727–733.
(5) Jordan, B. A.; Devi, L. A. G-protein-coupled receptor heterodi-
merization modulates receptor function. Nature 1999, 399, 697–
700.
(6) Terrillon, S.; Bouvier, M. Roles of G-protein-coupled receptor
dimerization. EMBO Rep. 2004, 5, 30–34.
(27) Jacobson, K. A.; Kirk, K. L.; Padgett, W. L.; Daly, J. W.
Functionalized Congeners of 1,3-Dialkylxanthines;Preparation
of Analogs with High-Affinity for Adenosine Receptors. J. Med.
Chem. 1985, 28, 1334–1340.
(28) Bakthavachalam, V.; Baindur, N.; Madras, B. K.; Neumeyer, J. L.
Fluorescent-probes for dopamine receptors;Synthesis and char-
acterization of fluorescein and 7-nitrobenz-2-oxa-1,3-diazol-4yl
conjugates of D1 and D2 receptor ligands. J. Med. Chem. 1991,
34, 3235–3241.
(7) Prinster, S. C.; Hague, C.; Hall, R. A. Heterodimerization of G
protein-coupled receptors: specificity and functional significance.
Pharmacol. Rev. 2005, 57, 289–298.
(8) George, S. R.; O’Dowd, B. F.; Lee, S. P. G-protein-coupled
receptor oligomerization and its potential for drug discovery.
Nat. Rev. Drug Discovery 2002, 1, 808–820.
(9) Maggio, R.; Novi, F.; Scarselli, M.; Corsini, G. U. The impact of
G-protein-coupled receptor hetero-oligomerization on function
and pharmacology. FEBS J. 2005, 272, 2939–2946.

5602 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 18
Soriano et al.
(29) Horn, A. S.; Tepper, P.; Kebabian, J. W.; Beart, P. M. N-0434, a
very potent and specific new D2 dopamine receptor agonist. Eur. J.
Pharmacol. 1984, 99, 125–126.
J. Serotonin dimers: application of the bivalent ligand approach to
the design of new potent and selective 5-HT(1B/1D) agonists. J.
Med. Chem. 1996, 39, 4920–4927.
(30) Unpublished results.
(43) Russo, O.; Berthouze, M.; Giner, M.; Soulier, J. L.; Rivail, L.;
Sicsic, S.; Lezoualc’h, F.; Jockers, R.; Berque-Bestel, I. Synthesis of
specific bivalent probes that functionally interact with 5-HT(4)
receptor dimers. J. Med. Chem. 2007, 50, 4482–4492.
(31) Song, A.; Zhang, J.; Lebrilla, C. B.; Lam, K. S. A novel and rapid
encoding method based on mass spectrometry for “one-bead-one-
compound” small molecule combinatorial libraries. J. Am. Chem.
Soc. 2003, 125, 6180–6188.
(44) Christopoulos, A.; Grant, M. K.; Ayoubzadeh, N.; Kim, O. N.;
Sauerberg, P.; Jeppesen, L.; El-Fakahany, E. E. Synthesis and
pharmacological evaluation of dimeric muscarinic acetylcholine
receptor agonists. J. Pharmacol. Exp. Ther. 2001, 298, 1260–1268.
(45) Vagner, J.; Xu, L.; Handl, H. L.; Josan, J. S.; Morse, D. L.; Mash,
E. A.; Gillies, R. J.; Hruby, V. J. Heterobivalent ligands crosslink
multiple cell-surface receptors: the human melanocortin-4 and
delta-opioid receptors. Angew. Chem., Int. Ed. 2008, 47, 1685–1688.
(46) Handl, H. L.; Sankaranarayanan, R.; Josan, J. S.; Vagner, J.;
Mash, E. A.; Gillies, R. J.; Hruby, V. J. Synthesis and evaluation of
bivalent NDP-alpha-MSH(7) peptide ligands for binding to the
human melanocortin receptor 4 (hMC4R). Bioconjugate Chem.
2007, 18, 1101–1109.
(32) Fuxe, K.; Marcellino, D.; Genedani, S.; Agnati, L. Adenosine A2A
receptors, dopamine D2 receptors and their interactions in Parkin-
son’s disease. Movement Disord. 2007, 22, 1990–2017.
~
(33) Canals, M.; Burgueno, J.; Marcellino, D.; Cabello, N.; Canela, E.
ꢀ
I.; Mallol, J.; Agnati, L.; Ferre, S.; Bouvier, M.; Fuxe, K.; Ciruela,
F.; Lluıs, C.; Franco, R. Homodimerization of adenosine A2A
´
receptors: qualitative and quantitative assessment by fluorescence
and bioluminescence energy transfer. J. Neurochem. 2004, 88, 726–
734.
(34) Lee, S. P.; O’Dowd, B. F.; Rajaram, R. D.; Nguyen, T.; George, S.
R. D2 dopamine receptor homodimerization is mediated by multi-
ple sites of interaction, including an intermolecular interaction
involving transmembrane domain 4. Biochemistry 2003, 42, 11023–
11031.
(47) Rosenbaum, D. M.; Cherezov, V.; Hanson, M. A.; Rasmussen, S.
G. F.; Thian, F. S.; Kobilka, T. S.; Choi, H.-J.; Yao, X.-J.; Weis, W.
I.; Stevens, R. C.; Kobilka, B. K. GPCR engineering yields high-
resolution structural insights into β2-adrenergic receptor function.
Science 2007, 318, 1266–1273.
ꢀ
ꢀ
ꢀ
(35) Franco, R.; Casado, V.; Mallol, J.; Ferre, S.; Fuxe, K.; Cortes, A.;
Ciruela, F.; Lluıs, C.; Canela, E. I. Dimer-based model for heptas-
´
panning membrane receptors. Trends Biochem. Sci. 2005, 30, 360–
366.
(36) Franco, R.; Casado, V.; Mallol, J.; Ferrada, C.; Ferre, S.; Fuxe, K.;
(48) Jaaloka, V.-P.; Griffith, M. T.; Hanson, M. A.; Cherezov, V.;
Chien, E. Y. T.; Lane, J. R.; Ijzerman, A. P.; Stevens, R. C. The 2.6
Angstrom crystal structure of a human A2A receptor bound to an
agonist. Science 2008, 322, 1212–1217.
(49) Milligan, G.; Bouvier, M. Methods to monitor the quaternary
structure of G-protein-coupled receptors. FEBS J. 2005, 272, 2914–
2925.
(50) Liang, Y.; Fotiadis, D.; Filipek, S.; Saperstein, D. A.; Palczewski,
K.; Engel, A. Organization of the G-protein-coupled receptors
rhodopsinand opsin in native membranes. J. Biol. Chem. 2003, 278,
21655–21662.
ꢀ
ꢀ
Cortes, A.; Ciruela, F.; Lluıs, C.; Canela, E. I. The two-state dimer
´
ꢀ
receptor model: a general model for receptor dimers. Mol. Phar-
macol. 2006, 69, 1905–1912.
ꢀ
ꢀ
ꢀ
(37) Casado, V.; Cortes, A.; Ciruela, F.; Mallol, J.; Ferre, S.; Lluıs, C.;
´
Canela, E. I.; Franco, R. Old and new ways to calculate the affinity
of agonists and antagonists interacting with G-protein-coupled
monomeric and dimeric receptors: the receptor-dimer coopera-
tivity index. Pharmacol. Ther. 2007, 116, 343–354.
ꢀ
ꢀ
ꢀ
(38) Franco, R.; Casado, V.; Cortes, A.; Mallol, J.; Ciruela, F.; Ferre,
S.; Lluıs, C.; Canela, E. I. G-protein-coupled receptor heteromers:
´
(51) Casciari, D.; Seeber, M.; Fanelli, F. Quaternary structure predic-
tions of transmembrane proteins starting from the monomer: a
docking-base approach. BMC Bioinformatics 2006, 7, 340–355.
(52) Casciari, D.; Dell’Orco, D.; Fanelli, F. Homodimerization in
Neurotensin 1 receptor involves helices 1, 2 and 4: insights form
quaternary structure predictions and dimerization free energy
estimations. J. Chem. Inf. Model. 2008, 48, 1669–1678.
(53) Im, W.; Feig, M.; Brooks, C. W., III An implicit membrane
generalizaed born theory for the study of the structure, stability,
and interactions of membrane proteins. Biophys. J. 2003, 85, 2900–
2918.
function and ligand pharmacology. Br. J. Pharmacol. 2008, 153
(Suppl 1), S90–S98.
(39) Maggio, R.; Innamorati, G.; Parenti, M. G protein-coupled re-
ceptor oligomerization provides the framework for signal discri-
mination. J. Neurochem. 2007, 103, 1741–1752.
(40) Portoghese, P. S. From models to molecules: opioid receptor
dimers, bivalent ligands, and selective opioid receptor probes. J.
Med. Chem. 2001, 44, 2259–2269.
(41) Bhushan, R. G.; Sharma, S. K.; Xie, Z.; Daniels, D. J.; Portoghese,
P. S. A bivalent ligand (KDN-21) reveals spinal delta and kappa
opioid receptors are organized as heterodimers that give rise to
delta(1) and kappa(2) phenotypes. Selective targeting of del-
ta-kappa heterodimers. J. Med. Chem. 2004, 47, 2969–2972.
(42) Halazy, S.; Perez, M.; Fourrier, C.; Pallard, I.; Pauwels, P. J.;
Palmier, C.; John, G. W.; Valentin, J. P.; Bonnafous, R.; Martinez,
ꢀ
(54) Casado, V.; Canti, C.; Mallol, J.; Canela, E. I.; Lluıs, C.; Franco, R.
´
Solubilization of A1 adenosine receptor from pig brain: character-
ization and evidence of the role of the cell membrane on the
coexistence of high- and low-affinity states. J. Neurosci. Res.
1990, 26, 461–473.