Hybrid ortho/allosteric ligands for the adenosine A1 receptor

Article abstract of DOI:10.1021/jm901252a

Many G protein-coupled receptors (GPCRs), including the adenosine A ₁ receptor (A₁AR), have been shown to be allosterically modulated by small molecule ligands. So far, in the absence of structural information, the exact location of

Full text of DOI:10.1021/jm901252a

3028 J. Med. Chem. 2010, 53, 3028–3037
DOI: 10.1021/jm901252a
Hybrid Ortho/Allosteric Ligands for the Adenosine A1 Receptor
Rajeshwar Narlawar,† J. Robert Lane,† Munikumar Doddareddy, Judy Lin, Johannes Brussee, and Adriaan P. IJzerman*
Division of Medicinal Chemistry, Leiden/Amsterdam Center for Drug Research, Leiden University, P.O. Box 9502, 2300 RA Leiden,
The Netherlands. †These authors contributed equally to this work.
Received August 20, 2009
Many G protein-coupled receptors (GPCRs), including the adenosine A1 receptor (A1AR), have been
shown to be allosterically modulated by small molecule ligands. So far, in the absence of structural
information, the exact location of the allosteric site on the A1AR is not known. We synthesized a series of
bivalent ligands (4) with an increasing linker length between the orthosteric and allosteric pharmaco-
phores and used these as tools to search for the allosteric site on the A1AR. The compounds were tested
in both equilibrium radioligand displacement and functional assays in the absence and presence of a
reference allosteric enhancer, (2-amino-4,5-dimethyl-3-thienyl)-[3-(trifluoromethyl)phenyl]methanone,
PD81,723 (1). Bivalent ligand N6-[2-amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno[2,3-c]-
pyridin-6-yl-9-nonyloxy-4-phenyl]-adenosine 4h (LUF6258) with a 9 carbon atom spacer did not show
significant changes in affinity or potency in the presence of 1, indicating that this ligand bridged both
sites on the receptor. Furthermore, 4h displayed an increase in efficacy, but not potency, compared to
the parent, monovalent agonist 2. From molecular modeling studies, we speculate that the allosteric site
of the A1AR is located in the proximity of the orthosteric site, possibly within the boundaries of the
second extracellular loop of the receptor.
Introduction
changes of the receptor which in turn affect the binding of the
orthosteric ligand and/or receptor activity.5 An allosteric
modulator may have no activity by itself but may potentiate
or inhibit activation of the receptor by its natural ligand.
In terms of ligand binding affinity, different scenarios are
feasible: allosteric modulators are capable of enhancing
the binding of the orthosteric ligands, termed as positive
cooperativity, of diminishing the ligand binding (negative
cooperativity), or of leaving the binding unaffected (neutral
cooperativity). Moreover, the nature of allosteric modulation
can be greatly dependent on the orthosteric ligand, so-called
probe dependency.6 Despite these complexities, the concept of
allosteric modulation of established drug targets such as
GPCRs has fueled an intensive search for new classes of lead
compounds which are different in their mode of action
compared to the classical agonists and antagonists.5,7,8
A bivalent ligand is a hybrid molecule in which two
pharmacophores are covalently linked with each other via a
spacer.9 The two pharmacophores are the same in the case of
homobivalent ligands and different in the case of heterobiva-
lent ligands. Furthermore, heterobivalent ligands may have
pharmacophores that bind to different molecular targets or to
two distinct sites on the same molecular target, for example,
an orthosteric and allosteric binding site, so-called “bitopic”
ligands.9-13 Bivalent ligands have been reported to display
enhanced receptor subtype selectivity and/or enhanced affi-
nity for the receptor target.14-16 In addition, such compounds
have proven to be useful tools in efforts to estimate the
distance between two receptors or indeed two binding sites
within the same target.14,17-19 Bivalent ligands have been
shown to enhance biological activity via the activation of
different receptors thatmediatethe same effect and, in thecase
of an agonist/antagonist bivalent ligand for the same receptor,
G protein-coupled receptors (GPCRsa) represent the lar-
gest family of cell-surface proteins mediating cellular commu-
nication.1 Consequently, they represent attractive drug
targets. Indeed, it is estimated that approximately 30% of
the marketed medicines act through members of this protein
class.2 Classically, such drugs are agonists, antagonists, or
inverse agonists and interact with the so-called orthosteric
ligand binding site. This orthosteric binding site is the natural
binding site for endogenous ligands such as hormones, neuro-
transmitters, and autocrine factors. The nucleoside adenosine
is such an endogenous agonist, for which there are four
characterized GPCRs (A1, A2A, A2B, and A3 receptors).3
The adenosine A1 receptor (A1AR) is responsible for mediat-
ing the physiological effects of adenosine in various tissues
and cell types such as the brain, heart, kidney, and adipocytes,
and signals mainly via G proteins of the Gi/o family to inhibit
adenylyl cyclase and thus reduce intracellular levels of cyclic
adenosine monophosphate (cAMP).4
Many GPCRs, including the adenosine A1 receptor, have
been shown to be allosterically modulated by both small-
molecule ligands and ions. The term “allosteric” refers to
binding sites that are topographically distinct from the orthos-
teric ligand binding site. The binding of small-molecule
allosteric modulators is believed to result in conformational
*To whom correspondence should be addressed. Phone: þ31- (0)71
527 4651. Fax: þ31- (0)71 527 4565. E-mail address: ijzerman@lacdr.
leidenuniv.nl.
a Abbreviations: cAMP, cyclic adenosine-50-monophosphate; CCPA,
2-chloro-N6-cyclopentyladenosine; CPA, N6-cyclopentyladenosine; CHO-
hA1, Chinese hamster ovary cells expressing human A1 receptor; DMF,
dimethylformamide; DPCPX, 8-cyclopentyl-1,3-dipropylxanthine; GPCR,
G protein-coupled receptor; SEM, standard error of the mean.
r
pubs.acs.org/jmc
Published on Web 03/26/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8 3029
the ability to produce partial agonism.15,20 Finally, the rela-
tively novel concept of bitopic ligands-those targeting an
orthosteric andan allosteric site withinthe same receptor-has
yielded ligands that can have improved receptor selectivity or
affinity. Moreover, such ligands can engender functional
selectivity in the actions of the orthosteric pharmacophore,
highlighting a viable means of further sculpting GPCR ligand
responses.9-13
Bhattacharya et al. have provided evidence that the recog-
nition site for the reference allosteric enhancer 1, (2-amino-
4,5-dimethyl-3-thienyl)-[3-(trifluoromethyl)phenyl]methanone,
PD81,723 (Scheme 1), is on the A1AR and that this enhancer
acts to directly stabilize the receptor in a conformational state
capable of coupling with Gi or Go.21 However, so far, in the
absence of structural information, the exact location of the
allosteric site on the A1AR is not known. Several mutation
studies have been performed in order to identify the amino
acids involved in the binding site of allosteric modulators on
the A1AR. However, none of these studies to date have
provided credible evidence regarding the probable location
of the allosteric site and the amino acids involved in the
binding of the allosteric ligand.22-24 It is often speculated
that the allosteric site on the A1AR is in close proximity to
the orthosteric site, most recently by Ferguson et al. who
synthesized a series of dihydrothieno[3,4-d]pyridazines as
allosteric modulators.25
We hypothesized that, when a heterobivalent ligand with a
correct spacer length occupies both the allosteric and orthos-
teric sites of the receptor, it may not show an increase in
activity in the further presence of an “external” allosteric
enhancer. Therefore, to test this hypothesis we designed and
synthesized bivalent ligands for A1AR linking an allosteric
and orthosteric pharmacophore. We decided to couple an
orthosteric adenosine-derived agonist 2 (Scheme 2) to an
allosteric enhancer using aliphatic chains of varying lengths
as spacers. It is important that the agonist and the allosteric
enhancer are functionalized at positions not involved in
binding to the recognition site. It is known from literature
that the introduction of bulky substituents at the N6 position
of adenosine is often well-tolerated and does not influence the
agonistic activity significantly.26,27 Consequently, we decided
to introduce the linkers at an N6 substituent, as in 2. In the
early 1990s, Bruns and co-workers reported on the allosteric
modulator PD81,723, capable of enhancing the binding and
activity of A1 receptor agonists.28,29 They observed that the
amino group at the 1-position and carbonyl group at the
2-position on the thiophene ring are vital for allosteric activity
and any modification of the amino functionality resulted in
the complete loss of activity. Previously, we reported piper-
idine-derived analogues of PD81,723 (e.g., 3, see Scheme 2)
with improved allosteric enhancing activity.30 From that
series, we observed that the introduction of bulkier substitu-
ents at the 4 and 5 positions of the thiophene ring was
well-tolerated and did not affect the allosteric activity. Hence,
we used that scaffold to link with adenosine-derived agonist 2.
This paper describes the synthesis and pharmacological
characterization of a series of bivalent ligands with an
Scheme 1. Chemical Structure of PD81,723
Scheme 2. Synthetic Strategy for the Synthesis of Bivalent Ligands 4

3030 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8
Scheme 3. Synthetic Route to Bivalent Ligands 4a
Narlawar et al.
a Reagents and conditions: (a) dibromoalkane, K2CO3, acetone, 60-70 °C, 4-8 h, 36-72%; (b) 4-piperidinone hydrochloride monohydrate, K2CO3,
DMF, RT, 24-48 h, 68-93%; (c) 7, S, Et2NH, EtOH, heat, 50 °C, 2-4 h, 17-40%; (d) anhydrous SnCl2, abs. EtOH, 70 °C, 6-12 h, 33-58%;
(e) 6-chloropurine 9-β-D-ribofuranoside, Et3N, EtOH, MW, 120 °C, 2.5 h, 43-65%.
increasing linker length between the orthosteric and allosteric
pharmacophores.
Table 1. Emax and EC50 Values for the Ability of Allosteric Modulator
PD81,723 and Analogue Compounds to Enhance the Binding of the
Orthosteric Agonist Radioligand [3H]CCPA at the Human Adenosine
A1 Receptor
Results and Discussion
Chemistry. A series of allosteric/orthosteric bivalent ligands
was synthesized as depicted in Scheme 3. The adenosine
component was introduced later in the reaction sequence in
order to minimize the handling of polar products. The synth-
esis commenced from 4-nitrophenol 5. Dibromoalkane was
monosubstituted with the phenolic -OH of 5 using anhydrous
K2CO3 in acetone at 60-70 °C. The N-alkylation of 6 was
achieved using piperidinone monohydrate hydrochloride and
anhydrous K2CO3 in DMF to afford ketone 7. The inter-
mediate benzoylacetonitrile 8 was synthesized as reported
previously in a two-step synthesis from 3,4-dichlorobenzoic
acid.30 The multicomponent Gewald reaction of ketone 7,
benzoylacetonitrile 8, and sulfur in absolute ethanol gave the
aminothiophene 9. The chemoselective nitro group reduction
of 9 was accomplished using anhydrous SnCl2 in absolute
ethanol at 70 °C to get amine 10. The subsequent incorpora-
tion of the adenosine component was achieved by the reaction
of the amine moiety of compound 10 with commercially
available 6-chloropurine riboside using absolute ethanol and
triethyl amine under microwave conditions at 120 °C to obtain
the bivalent ligands 4a-i with spacer lengths n = 2-10
(Scheme 1, Table 2).
compd.
9a
9b
chain length (n)
pEC50 (EC50, μM)
Emax (%)a
2
3
5.07 ( 0.1 (8.5)
5.43 ( 0.2 (3.7)
5.28 ( 0.1 (5.2)
5.76 ( 0.1 (1.7)
250 ( 6
204 ( 6
185 ( 2
140 ( 12
9d
PD81,723
5
-
a Specific [3H]CCPA binding in the absence of any modulator set at
100%.
of the compounds for enhancing orthosteric agonist bind-
ing are in the low micromolar range, similar to that of
PD81,723. This provides further proof that introduction of
bulky substituents at the C-4 and C-5 positions of the
thiophene ring did not negatively affect the allosteric
enhancing ability.
Next, the affinities of the bivalent compounds 4 and the
parent compound 2 were determined for the human A1AR
by competitive displacement of the radiolabeled inverse
agonist [3H]DPCPX. In the control condition, the biphasic
curve displayed by parent compound 2 was best fit with a
Pharmacology and Structure-activity Relationships. The
novel PD81,723 analogues 9a, 9b, and 9d, as representatives
of the whole series of 9, were first tested for their ability to
allosterically enhance the binding of the orthosteric radiola-
beled agonist [3H]CCPA (2-chloro-N6-cyclopentyladenosine).
All three compounds were more efficacious in their ability
to enhance [3H]CCPA binding than PD81,723 (Table 1).
Furthermore, as shown in Table 1, the potencies (EC50 values)

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8 3031
Table 2. Affinity of Bivalent Ligands 4 (Ki Values) at the Human Adenosine A1 Receptor Determined by Their Ability to Displace the Radioligand
[3H]DPCPXa
increase in
affinity (fold)
max. displacement, %
specific [3H]DPCPX binding
compd.
spacer length (n)
pKi (Ki, nM)
pKi,þPD 81 (Ki, nM)
4a
4b
2
3
6.13 ( 0.06 (751)
6.38 ( 0.08 (439)
6.21 ( 0.04 (618)
6.67 ( 0.06 (219)
6.32 ( 0.06 (419)
6.35 ( 0.06 (450)
6.11 ( 0.12 (836)
6.18 ( 0.08 (679)
5.90 ( 0.20 (1470)
6.08 ( 0.06 (842)
7.37 ( 0.3 (71)
6.33 ( 0.03 (469)*
6.61 ( 0.11 (268)
6.51 ( 0.03 (313)*
6.96 ( 0.08 (114)*
6.56 ( 0.07 (288)*
6.46 ( 0.01 (347)
6.34 ( 0.20 (591)
6.19 ( 0.05 (651)
5.89 ( 0.11 (1360)
6.72 ( 0.06 (195)**
1.6
1.9
1.9
1.9
1.7
1.3
1.4
1.07
1.07
4.3
100 ( 5
100 ( 4
100 ( 5
100 ( 5
103 ( 2
91 ( 2
65 ( 2
83 ( 1
55 ( 3
100 ( 6
4c
4d
4
5
4e
6
4f
4g
7
8
4h
4i
9
10
-
2
2Ki,high
2Ki,low
5.73 ( 0.1 (1990)
a Values of pKi were compared between control and þ 10 μM PD81,723 conditions using a Student’s t-test. Significant differences were annotated as
follows: p < 0.05 - *, p < 0.01 - **, p < 0.001 - ***. Values of pKi obtained for the compounds in the control condition were compared using a one-way
ANOVA with a Bonferroni post-hoc test. Significant differences were observed as follows: 4d > 4a, 4c, 4g, 4h, 4i, p < 0.05 - *,.
Figure 1. Displacement of specific [3H]DPCPX binding by increasing concentrations of compounds 2, 4a-i (A). Displacement of specific
[3H]DPCPX binding by increasing concentrations of compounds 2 (B), 4d (C), and 4h (D) in the absence or presence of 10 μM PD81,723.
Ki,high of 71 ( 48 nM and a Ki,low of 1.99 ( 0.5 μM. In the
presence of the allosteric enhancer PD81,723 (10 μM), the
curve was monophasic (Ki,þPD81,723 = 195 nM) (Figure 1B,
Table 2). Conversely, the displacement of [3H] DPCPX by all
bivalent ligands (4) was best characterized by monophasic
isotherms. The pentyl-linked bivalent ligand 4d (Figure 1C,
Table 2) was found to have the highest affinity for the A1AR.
When tested in the presence of allosteric enhancer PD81,723
(10 μM), bivalent ligands with a spacer length of 6 or less
showed a significant increase in affinity (2-fold), although
this was less than that observed for the orthosteric ligand 2
(4-fold, Table 2). Smaller, nonsignificant 1.3-1.4-fold shifts
were observed in the cases of 4f and 4g, respectively. For
compounds 4h (see also Figure 1D) and 4i, the increase in
affinity was negligible.
It was important to characterize the bivalent ligands in
terms of their ability to modulate agonist binding. Bound
[3H]CCPA was completely displaced by the orthosteric
agonist 2 with a Ki value of 48 nM (Figure 2, Table 3). This
value is similar to the Ki,high of compound 2 observed in the
[3H]DPCPX displacement assay, consistent with displacement
of [3H]CCPA bound to the high affinity G protein-coupled

3032 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8
Narlawar et al.
with the addition of 10 μM or 100 μM of PD81,723, 4d, or 4h
caused an increase in measured specific [3H]CCPA binding
compared to the control (100 μM CPA only) condition
(Figure 3). Therefore, the allosteric pharmacophore of both
bivalent ligands was functional.
To demonstrate that the bivalent ligands retained their
ability as agonists to activate the receptor, we tested two
selected bivalent ligands, 4d and 4h, using a [35S]GTPγS
binding assay. Both compounds 4d and 4h acted as agonists
and compared to the monovalent agonist compound 2
showed
a significant increase in efficacy (Table 4,
Figure 4A). It is interesting to note that the binding affinities
and rank order of affinity observed for these three com-
pounds in the [3H]CCPA competition assays were similar to
the potencies obtained in this assay. Indeed, 4d and 4h were
2-fold and 3-fold less potent than the monovalent compound
2. A differential effect on the potencies of the three com-
pounds was observed with the addition of 10 μM PD81,723.
In the case of both the monovalent compound 2 and the
bivalent compound 4d, a 5-fold increase in potency was
observed with the addition of PD81,723. In contrast, a
smaller 2-fold shift in potency was observed for compound
4h. Similarly, a significant decrease in the maximal effect of
both compound 2 and 4d but not compound 4h was observed
with the addition of PD81,723. Intriguingly, the Hill slopes
for the bivalent ligands 4d and, in particular, 4h were
significantly higher than unity (1.3 and 1.5, respectively),
whereas the Hill slope of the monovalent agonist 2 was close
to unity (0.8). Hill slopes higher than unity are signs of
positive cooperativity, suggesting a simultaneous sensing
and synergy of the orthosteric and allosteric site on the
receptor. With addition of 10 μM PD81,723 the Hill slopes
of 4d and 4h no longer differed significantly from unity.
It was of further interest to determine the functional effect
of the bivalent ligands using a whole-cell assay, in this case
the measurement of ERK 1/2 phosphorylation. In this assay,
the monovalent compound 2 had the highest potency fol-
lowed by 4d and 4h, the same rank order observed for the
[35S]GTPγS data (Figure 4B, Table 5). When compared to
values obtained in the [35S]GTPγS assay, compounds 2 and
4d showed a 2-fold and 7-fold increase in potency for ERK
1/2 phosphorylation, while a 2.5-fold decrease in potency
was observed for compound 4h. The presence of 10 μM
PD81,723 caused a 4-fold increase in potency for compound
2, a 2-fold increase in potency for compound 4d, and no
significant increase in potency for compound 4h. Upon
addition of 10 μM PD81,723 a significant 30% increase in
maximal efficacy (Emax) was observed for compound 2.
Compounds 4d and 4h had a significantly higher Emax as
compared to compound 2 in the control condition. However,
only the Emax of compound 4h was unaffected by the addi-
tion of PD81,723.
Figure 2. Ability of increasing concentrations of compounds
PD81,723, 9d, 2, 4d, and 4h to displace or enhance the binding of
the radiolabeled agonist [3H]CCPA to CHO membranes expressing
the A1AR.
Figure 3. Ability of compounds 4d and 4h to decrease the rate of
dissociation in a single-point kinetic assay. At time = 0, dissociation
was initiated with the addition of 100 μM CPA with or without
modulator compounds PD81,723, 4h, or 4d or a combination.
Reactions were stopped after 30 min and bound [3H]CCPA was
determined. Results are presented as a percentage of specific
[3H]CCPA bound at time = 0.
Table 3. Affinity of Bivalent Ligands 4 (Ki values) at the Human
Adenosine A1 Receptor Determined by Their Ability to Displace the
Radioligand [3H]CCPAa
maximal displacement, %
of specific [3H]CCPA
binding
spacer
length (n)
compd.
pKi (Ki, nM)
4d
4h
2
5
9
6.99 ( 0.05 (103)
6.16 ( 0.12 (738)
7.32 ( 0.06 (48)
97 ( 0.9
79 ( 0.3
100 ( 0.9
-
a Values of pKi obtained for the compounds were compared using a
one-way ANOVA with a Bonferroni post-hoc test. Significant differ-
ences were observed as follows: 2 > 4d > 4h. p < 0.05.
The linking of the parent A1AR agonist 2 to an A1AR
allosteric modulator has resulted in the generation of biva-
lent ligands with novel properties. Initial characterization of
the bivalent ligands using competition experiments with
[3H]DPCPX suggested that the ligands had a similar (4h)
or slightly improved (4d) affinity for the receptor. However,
subsequent experiments demonstrated that these ligands had
a lower affinity for receptors labeled with the agonist
[3H]CCPA and a lower potency in functional assays. Linking
such an agonist to an allosteric enhancer might be expected
to result in an increase in affinity or potency. Indeed, one
such study demonstrated an increased affinity of a bitopic for
population of A1AR. The values of Ki observed for com-
pounds 4d and 4h for the displacement of [3H]CCPA (103 and
738 nM, respectively) are within the same order of magnitude
to values of Ki obtained for the displacement of [3H]DPCPX.
However, while compound 2and compound 4d show complete
displacement of [3H]CCPA, compound 4h displaces bound
[3H]CCPAto a maximal value of approximately 80% (Table 3,
Figure 2). Finally, it was important to show that the function-
ality of the allosteric pharmacophore was retained in the
bivalent ligands. A single-point kinetic assay revealed that

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8 3033
Table 4. Ability of 2, 4d, and 4h to Activate the Human Adenosine A1 Receptor Stably Expressed in CHO Cell Membranes Measured Using a
35S]GTPγS Binding Assaya
[
pEC50 (EC50, nM)
Emax, %
Hill slope
compd.
control
þ PD81,723
fold shift
control
þ PD81,723
control
þ PD81,723
4d
4h
2
6.57 ( 0.04 (273)
6.41 ( 0.03 (390)
6.91 ( 0.05 (142)
7.29 ( 0.06** (52)
6.75 ( 0.04 ** (181)
7.52 ( 0.03*** (29)
5.2
2.2
4.9
110 ( 3
110 ( 2
100 ( 2
97 ( 2*
105 ( 3
91 ( 3*
1.29 ( 0.09
1.45 ( 0.03
0.84 ( 0.04
0.89 ( 0.07 *
0.88 ( 0.09 **
0.98 ( 0.13
a Values are displayed for the control condition or in the presence of 10 μM PD81,723. Emax values are displayed as a percentage of maximal
stimulation of compound 2. For all compounds, values of pEC50, Emax, and Hill slope were compared between control and þ PD81,723 conditions using
a Students t-test. Significant differences were annotated as follows: p < 0.05 - *, p < 0.01 - **, p < 0.001 - ***. Values of pEC50 and Emax obtained for the
compounds were compared using a one-way ANOVA with a Bonferroni post-hoc test. Rank orders were as follows: Emax 4d = 4h > 2, p < 0.05; pEC50
2 > 4d = 4h, p < 0.05.
A1ARs to G proteins rather than to increase the affinity of
the receptor to agonists; indeed, PD81,723 has been shown to
be an allosteric agonist.33-35 Therefore, both the leftward
shift observed upon addition of PD81,723 in [3H]DPCPX
displacement experiments and the increase in potency and
efficacy observed in functional assays can be attributed to an
increase in receptor coupling rather than an increase in
affinity for the G protein-coupled population of receptors.
Consequently, the linking of an agonist to a PD81,723
analogue pharmacophore should not necessarily be expected
to increase ligand affinity. Conversely, an increase in ligand
efficacy compared to the parent compound 2 might be
expected, and indeed, this is observed for both compounds
4d and 4h.
The original hypothesis of this research was that if a
bivalent ligand was generated in which the allosteric and
orthosteric pharmacophores occupy the receptor simulta-
neously then this compound would be insensitive to the
addition of an additional allosteric enhancer. Indeed, for
compounds with a spacer length of greater than 7 carbon
atoms no significant affinity or potency shift was observed
with the addition of 10 μM PD81,723. Furthermore, no
increase in efficacy was observed upon the addition of excess
PD81,723. The bivalent ligands have a number of different
modes of binding. Therefore, interpretation of this result is
complex due to the reversibility of binding and the distribu-
tion of the ligands across a number of receptor species. These
would include species in which the ligand is bound only to the
allosteric site, only to the orthosteric site, or if the linker is
long enough, simultaneously to both sites. Furthermore, one
receptor can be occupied by two bivalent ligands, one
binding allosterically and one binding orthosterically. Upon
addition of extra monovalent allosteric modulator, two
scenarios are possible: first, that the bivalent ligand will fail
to bind resulting in a loss of potency, or second, that
the bivalent ligand still binds orthosterically. In this case,
the change in potency will depend on the interaction with the
allosteric ligand and the nature and efficacy of this allosteric
ligand compared to that of the allosteric pharmacophore of
the bivalent ligand. In the present study, the allosteric
pharmacophore is structurally related to and has a similar
efficacy to that of the added monovalent modulator
PD81,723. Therefore, if the linker is of sufficient length to
allow simultaneous occupation of the allosteric and orthos-
teric binding sites, the addition of excess PD81,723 will have
a minimal effect on the potency of the bivalent compounds as
was observed for compounds 4f-i. Following the same
hypothesis, if the linker length is insufficient to occupy both
sites, a bivalent ligand can occupy either the orthosteric or
allosteric site. In this situation, the occupancy of the allos-
teric site will be related to the concentration of the bivalent
Figure 4. Ability of increasing concentrations of compounds 2, 4d,
and 4h, both in the absence and in the presence of PD81,723, to (A)
stimulate [35S]GTPγS binding in CHO cell membranes expressing
the A1AR or (B) to stimulate A1AR-mediated ERK1/2 phosphor-
ylation in intact CHO cells. Value of 100% is equivalent to a
stimulation of 5-fold and 7.2-fold over basal of [35S]GTPγS and
ERK1/2 phosphorylation data, respectively.
the target receptor compared to the binding affinities of the
relevant individual pharmacophores.11 However, it has been
demonstrated that the combination of a negative allosteric
modulator pharmacophore with an agonist pharmacophore
resulted in a decrease in affinity of the bitopic ligand as
compared to the parent orthosteric compound.9,12 There-
fore, gain or loss of affinity of a bitopic ligand compared to
the parent ligand is greatly dependent on the allosteric
pharmacophore partner. Several reports have demonstrated
that PD81,723 acts to increase the fractional coupling of

3034 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8
Narlawar et al.
Table 5. Ability of 2, 4d, and 4h to Stimulate ERK1/2 Phosphorylation in Intact CHO Cells Stably Expressing the Human Adenosine A1 Receptora
pEC50 (EC50, nM)
Emax, %
compd.
control
þ PD81,723
fold shift
control
þ PD81,723
4d
4h
2
6.82 ( 0.07 (156)
6.02 ( 0.12 (1071)
7.68. ( 0.06 (22)
7.05 ( 0.00* (89)
6.16 ( 0.14 (813)
8.34 ( 0.11** (5)
1.8
1.3
4.4
114 ( 5
123 ( 6
100 ( 1
131 ( 2*
122 ( 8
131 ( 3***
a The data shown are the mean ( SEM of at least 3 separate experiments. Values are displayed for the control condition or in the presence of 10 μM
PD81,723. Emax values are displayed as a percentage of maximal stimulation of compound 2. For all compounds, values of pEC50, Emax were compared
between control and þ PD81,723 conditions using a Student’s t-test. Significant differences were annotated as follows: p < 0.05 -*, p < 0.01 - **, p <
0.001 - ***. Values of pEC50 and Emax obtained for the compounds were compared using a one-way ANOVA with a Bonferroni post-hoc test. Rank
orders were as follows: Emax 4h > 2 = 4d, p < 0.05; pEC50 2 > 4d > 4h, p < 0.05.
Figure 5. Molecular docking studies to demonstrate potential binding orientation of 4h, the bitopic ligand with a 9 carbon atom linker. A
homology model of the human A1 receptor was generated (B-D) using the adenosine A2A receptor crystal structure as a template (A). Using
this model, docking analysis was performed for compounds 2 (B) and 4h (C,D). The potential orientations of the allosteric pharmacophore are
demonstrated (C), including one orientation in which the allosteric pharmacophore makes significant interactions with extracellular loop 2 (D).
ligand added and the affinity of the bivalent ligand for the
allosteric site. In the case of compound 4d, the affinity of the
ligand for the orthosteric binding site as indicated by the
[3H]DPCPX displacement assay is approximately 10-fold
higher than the potency of the allosteric pharmacophore
(Table 1). Therefore, the allosteric site will not be fully
occupied. Hence, for this compound a shift in affinity is
observed upon addition of 10 μM PD81,723 as the occu-
pancy of the allosteric site increases.
The relatively short length of linker (g7 carbon atoms)
required to observe potential bitopic ligand binding suggests
that the allosteric site is located in close proximity of the
orthosteric site. We therefore performed homology model-
ing and docking studies using the A2A adenosine receptor
crystal structure as a template (Figure 5A), yielding
a structure for the A1AR complexed with agonist 2
(Figure 5B). As shown by the 3-D model (Figure 5C), the
best solutions for compound 4h (9 carbon atom linker)
demonstrate that the linker and allosteric pharmacophore
extend out of the binding pocket into the extracellular space.
Once such solution (Figure 5D) shows the allosteric phar-
macophore interacting with extracellular loop 2, a region
that has been suggested as the binding site for allosteric
modulators at other GPCRs.36 However, due to the inherent
flexibility of both the linker region of the molecule and the
extracellular loop it is inappropriate to implicate specific
amino acid residues in this interaction without validation
with mutagenesis studies.37 In agreement with the extracel-
lular orientation of the linker and allosteric pharmacophore,
it has been observed that large substituents are tolerated at
the N6 position. Indeed, bivalent ligands with adenosine A1
receptor and β2-adrenergic receptor pharmacophores have
been successfully generated using this strategy.31
It is interesting to note that bivalent ligands with a linker
length of seven or greater failed to completely displace bound
[3H]DPCPX or [3H]CCPA as compared to the maximal
displacement observed for the monovalent agonist 2. This
inability of bivalent compounds with a linker length greater
than seven to displace the competing radioligand fully also
coincided with their observed insensitivity to PD81,723 in
both binding and functional assays. Therefore, this incom-
plete displacement could be associated with their ability to
simultaneously bind to both the allosteric and orthosteric
binding sites. Incomplete displacement has been observed in
a previous study characterizing a bitopic ligand generated
by linking orthosteric ligand and allosteric enhancer phar-
macophores.12 Such behavior was reconciled with the idea
that such a bitopic ligand would produce an effect that
was intermediate between that of a positive allosteric mod-
ulator and that of a simple competitive orthosteric inhibi-
tor.12 One might argue that incomplete displacement of
radioligand by compounds 4h and 4i is explained by insolu-
bility of these compounds. However, the curves have a Hill
slope of unity and begin to plateau at low concentrations of
approximately 1 μM. This is not consistent with dissolution
issues.
The data from the present study show that linking orthos-
teric and allosteric pharmacophores for the A1AR into a
single ligand provides clues about the location of the allosteric
site. This advocates that a minimum chain length is needed to
connect both sites on the receptor. Similarly, work by Por-
toghese et al. demonstrated that an optimal linker length was
required for the generation of bivalent ligands that specifically
targeted δ-κ opioid receptor heterodimers.17,32 In this case, a

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8 3035
˚
spacer length of 21 atoms (maximal molecule length 90 A)
was required to allow the simultaneous binding of the
κ-opioid selective pharmacophore and the δ-opioid selective
pharmacophore to their respective orthosteric sites on each
monomer of the δ-κ opioid receptor heterodimer. This is
much greater than the 9 carbon atom spacer length (maximal
(MeOH/CH2Cl2, 1:9) to afford the desired product as yellow
colored solid.
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-2-ethyloxy-4-phenyl]-adenosine (4a). Yield: 45
mg, 47%. Mp 158-160 °C. 1H NMR (400 MHz, CDCl3): δ =
8.35 (s, 1H), 8.17 (s, 1H), 7.65 (d, J = 8.8 Hz, 2H), 7.57-7.51 (m,
2H), 7.31 (dd, J = 8.0 Hz, J = 1.0 Hz, 1H), 6.96 (d, J = 8.8 Hz,
2H), 5.93 (d, J = 6.8 Hz, 1H), 4.81 (t, J = 6.0 Hz, 1H), 4.37 (d,
J = 4.4 Hz, 1H), 4.21 (t, J = 5.6 Hz, 2H), 3.96 (d, J = 12.4 Hz,
1H), 3.77 (d, J = 12.4 Hz, 1H), 3.70 (s, 2H), 3.06 (t, J = 6.0 Hz,
2H), 2.78 (t, J = 6.0 Hz, 2H), 2.02 (t, J = 6.4 Hz, 2H) ppm.
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-3-propyloxy-4-phenyl]-adenosine (4b). Yield:
˚
molecule length 49 A) of compound 4h in this study.
It is interesting to note that other bitopic ligands, simulta-
neously binding to orthosteric and allosteric sites within the
same receptor, have similar spacer lengths. Three separate
studies characterized distinct bitopic ligands for the mus-
carinic M2 receptor, with molecule lengths between 17 and
1
10-12
˚
135 mg, 65%. Mp 143-145 °C. H NMR (400 MHz, CDCl3):
45 A.
Therefore, the length of spacer used for these
δ = 8.35 (s, 1H), 8.17 (s, 1H), 7.62 (d, J = 8.8 Hz, 2H), 7.56-7.51
(m, 2H), 7.31 (dd, J = 8.4 Hz, J = 2.8 Hz, 1H), 6.93 (d, J = 8.8
Hz, 2H), 5.93 (d, J = 6.8 Hz, 1H), 4.80 (t, J = 6.4 Hz, 1H), 4.37
(dd, J = 5.2 Hz, J =1.6 Hz, 1H), 4.27(d, J = 1.6 Hz, 1H), 4.05 (t,
J = 6.0 Hz, 2H), 3.96 (dd, J = 12.6 Hz, J = 1.8 Hz, 1H), 3.77(dd,
J = 12.6 Hz, J = 1.4 Hz, 1H), 3.52 (s, 2H), 2.74 (t, J = 8.0 Hz,
2H), 2.62 (t, J = 5.4 Hz, 2H), 2.09-1.96 (m, 4H) ppm.
bitopic ligands, and indeed the bitopic ligands described in
this study, is considerably smaller than that required for the
simultaneous occupation of two orthosteric sites within
a receptor heterodimer as determined by Portoghese and
co-workers.17,32
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-4-butyloxy-4-phenyl]-adenosine (4c). Yield: 40
mg, 52%. Mp 140-142 °C. 1H NMR (400 MHz, CDCl3): δ =
8.34(s, 1H), 8.21 (s, 1H), 7.62-7.52 (m, 4H), 7.32 (dd, J =8.4 Hz,
J = 2.8 Hz, 1H), 6.93 (d, J = 8.8 Hz, 2H), 5.94 (d, J = 6.4 Hz,
1H), 4.80 (t, J = 6.8 Hz, 1H), 4.37 (d, J = 2.8 Hz, 1H), 4.26 (s,
1H), 4.02 (t, J = 6.0 Hz, 2H), 3.96 (d, J = 12.4 Hz, 1H), 3.77 (d,
J = 12.4Hz, 1H), 3.56(s, 2H), 2.71-2.62 (m, 4H), 2.02(t, J = 6.4
Hz, 2H), 1.80-1.65 (m, 4H) ppm.
Conclusion
A series of bivalent ligands linking both the agonist and
allosteric pharmacophores of the A1AR was synthesized and
used as a tool to locate the allosteric site on the A1AR. These
ligands were tested in radioligand competition assays in the
absence and presence of an allosteric enhancer. Bivalent
ligands 4f-i did not show a significant affinity change in the
presence of an allosteric enhancer, indicating that they bridge
both sites on the receptor. The data suggest that the allosteric
site of A1AR is in close proximity to the orthosteric site and
that both sites can be simultaneously occupied by one bivalent
ligand. Homology modeling and molecular docking studies
using the bivalent ligands suggest that the allosteric pharma-
cophore may interact with extracellular loop 2 of the same
A1AR. Thus, bivalent ligands may be used as a tool to
determinethe distance between two sites and a strategysimilar
to ours could be used to determine the location of allosteric
sites on other GPCRs.
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-5-pentyloxy-4-phenyl]-adenosine (4d). Yield:
120 mg, 61%. 123-125 °C. 1H NMR (400 MHz, CDCl3): δ =
8.36 (s, 1H), 8.09 (s, 1H), 7.63 (td, J = 8.8 Hz, J = 2.8 Hz, 2H),
7.54 (d, J = 2.0 Hz, 1H), 7.51 (d, J = 8.4 Hz, 1H), 7.30 (dd, J =
8.4 Hz, J = 2.0 Hz, 1H), 6.93 (td, J = 8.8 Hz, J = 2.8 Hz, 2H),
5.89 (d, J = 6.8 Hz, 1H), 4.82 (t, J = 6.8 Hz, 1H), 4.37 (dd, J =
5.2 Hz, J = 1.6 Hz, 1H), 4.28 (d, J = 1.6 Hz, 1H), 4.03-3.95 (m,
3H), 3.77 (dd, J = 12.6 Hz, J = 1.8 Hz, 1H), 3.48 (s, 2H),
2.63-2.52 (m, 4H), 1.84 (t, J = 6.8 Hz, 2H), 1.74-1.58 (m, 2H),
1.67-1.60 (m, 2H), 1.55-1.47 (m, 2H) ppm.
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-6-hexyloxy-4-phenyl]-adenosine (4e). Yield:
1
Experimental Section
70 mg, 43%. Mp 133-135 °C. H NMR (400 MHz, CDCl3):
δ = 8.34 (s, 1H), 8.16 (s, 1H), 7.60 (dd, J = 9.2 Hz, J = 3.2 Hz,
2H), 7.56-7.51 (m, 2H), 7.30 (dd, J = 8.4 Hz, J = 2.0 Hz, 1H),
6.93 (td, J = 8.8 Hz, J = 2.8 Hz, 2H), 5.92 (d, J = 6.8 Hz, 1H),
4.81 (t, J = 6.8 Hz, 1H), 4.37 (dd, J = 6.0 Hz, J = 2.0 Hz, 1H),
4.27 (d, J = 2.0 Hz, 1H), 4.02-3.94 (m, 3H), 3.77 (dd, J = 12.8
Hz, J = 2.0 Hz, 1H), 3.43 (s, 2H), 2.57-2.47 (m, 4H), 1.94 (t, J =
6.8 Hz, 2H), 1.85-1.72 (m, 2H), 1.63-1.48 (m, 4H), 1.46-1.37
(m, 2H) ppm.
Chemistry: Materials and Methods. All reagents were obtained
from commercial sources and all solvents were of analytical grade.
1H and 13C NMR spectra were recorded on a Bruker AC 400 (1H
NMR, 400 MHz; 13C NMR, 100 MHz) with tetramethylsilane as
an internal standard. Chemical shifts are reported in δ (ppm).
Reactions were routinely monitored by TLC using Merck silica
€
gel F254 plates. Melting points were measured on a Buchi melting
point apparatus and are uncorrected. Microwave reactions were
performed on an Emrys Optimizer (Biotage AB). Wattage was
automatically adjusted to maintain the desired temperature.
Purity of all final products was determined by analytical HPLC
to be g95%. HPLC purity of compounds was measured with a
reverse-phase column (RP18, 125 ꢀ 4 mm, 5 μm, 254 nm and
diode array, 210-360 nm) with two diverse solvent systems. A:
Water, 10% acetonitrile, 10 mM HOAc, 5 mM SDS. B: Water,
90% acetonitrile, 10 mM HOAc, 5 mM SDS. Samples were eluted
by a gradient between 100% A and 100% B at a flow rate of
0.6 mL/min.
General Procedure for Microwave-Assisted Amination of
6-Chloropurine 9-β-D-Ribofuranoside. To a solution of aniline
10 (1 equiv) in ethanol, 6-chloropurine 9-β-D-ribofuranoside
(1 equiv), triethylamine (1.5 equiv) was added and the resulting
mixture was heated in microwave at 120 °C for 2.5 h. Upon
completion (TLC), the reaction mixture was diluted with
chloroform (20 mL) and evaporated in vacuo. The crude
compound was purified by flash column chromatography
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-7-heptyloxy-4-phenyl]-adenosine (4f). Yield:
1
60 mg, 54%. Mp 125-126 °C. H NMR (400 MHz, CDCl3):
δ = 8.35 (s, 1H), 8.15 (s, 1H), 7.61 (d, J = 8.4 Hz, 2H), 7.52 (d,
J = 9.6 Hz, 2H), 7.30 (d, J = 8.0 Hz, 1H), 6.93 (d, J = 8.8 Hz,
2H), 5.92 (d, J = 6.4 Hz, 1H), 4.81 (t, J = 5.8 Hz, 1H), 4.37 (d,
J = 6.0 Hz, 1H), 4.27 (d, J = 2.0 Hz, 1H), 4.02-3.94 (m, 3H),
3.77(d, J = 12.4 Hz, 1H), 3.43 (s, 2H), 2.55-2.46(m, 4H), 1.94 (t,
J = 5.8 Hz, 2H), 1.85-1.74 (m, 2H), 1.61-1.47 (m, 4H),
1.44-1.33 (m, 4H) ppm.
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-8-octyloxy-4-phenyl]-adenosine (4g). Yield: 35
mg, 47%. Mp 128-129 °C. 1H NMR (400 MHz, CDCl3): δ =
8.36(s, 1H), 8.06 (s, 1H), 7.61 (d, J = 8.8 Hz, 2H), 7.52 (d, J = 9.2
Hz, 2H), 7.30 (d, J = 8.4 Hz, 1H), 6.94 (d, J = 8.8 Hz, 2H), 5.88
(d, J = 6.8 Hz, 1H), 4.83 (t, J = 6.0 Hz, 1H), 4.35 (d, J = 5.2 Hz,
1H), 4.27 (d, J = 1.2 Hz, 1H), 4.01-3.94 (m, 3H), 3.77 (d, J =
12.4 Hz, 1H), 3.43 (s, 2H), 2.53-2.44 (m, 4H), 1.94 (t, J = 5.8 Hz,

3036 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8
Narlawar et al.
2H), 1.82-1.75 (m, 2H), 1.58-1.44 (m, 4H), 1.42-1.29 (m,
6H) ppm.
Briefly, Chinese hamster ovary (CHO) cells stably transfected
with the human A1AR were grown to 90% confluence and
maintained in Dulbecco’s modified eagle serum (DMEM) con-
taining 10% newborn calf serum (NCBS), streptomycin (50 mg/
mL), penicillin (50 IU/ml), and G418(0.2 mg/mL) at 37 °C in 5%
CO2. Cells were then harvested using phosphate-buffered saline
(PBS) þ 0.48 mM EDTA followed by centrifugation (300 g,
5 min). Cells were then seeded into 96 well plates at a density of
40 000 cells/well in DMEM/NCBS. After 6 h, plates were
washed twice with PBS and then maintained in DMEM for
14 h. Prior to agonist stimulation, cells were pretreated for
30 min with 0.8 IU/mL adenosine deaminase (ADA) in phe-
nol-red-free DMEM. Stimulation was then initiated with the
addition of test compound with or without PD81,723. Stimula-
tion was allowed to proceed for 5 min before the reaction was
terminated by the removal of media and the addition of SureFire
lysis buffer (100 μL to each well). The plate was then agitated for
2 min before 4 μL of lysate was added to a white opaque 384-well
proxiplate. Seven microliters of a 1:1:10:60 mix of Alphascreen
acceptor beads, Alphascreen donor beads, activation buffer,
and reaction buffer was added to each well. The plate was then
incubated in the dark at room temperature for at least 3 h before
reading on an EnVision plate reader (Perkin-Elmer, Groningen,
NL) using standard alphascreen settings. All data was expressed
as a percentage of the maximal stimulation observed for com-
pound 2 minus the basal signal. Typically, an 8-fold stimulation
above basal was observed.
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-9-nonyloxy-4-phenyl]-adenosine (4h). Yield: 22
mg, 43%. Mp 124-126 °C. 1H NMR (400 MHz, CDCl3): δ =
8.35 (s, 1H), 8.09 (s, 1H), 7.62 (d, J = 8.4 Hz, 2H), 7.53 (d, J = 9.6
Hz, 2H), 7.31 (d, J = 8.4 Hz, 1H), 6.93 (d, J = 8.8 Hz, 2H), 5.89
(d, J = 6.8 Hz, 1H), 4.82 (t, J = 6.0 Hz, 1H), 4.36 (d, J = 4.4 Hz,
1H), 4.27 (d, J = 1.6 Hz, 1H), 4.02-3.93 (m, 3H), 3.77 (d, J =
12.8 Hz, 1H), 3.66 (s, 2H), 2.84-2.67 (m, 4H), 2.03 (t, J = 5.8 Hz,
2H), 1.84-1.74 (m, 2H), 1.69-1.59 (m, 2H), 1.51-1.42 (m, 2H),
1.41-1.31 (m, 8H) ppm.
N6-[2-Amino-3-(3,4-dichlorobenzoyl)-4,5,6,7-tetrahydrothieno-
[2,3-c]pyridin-6-yl-10-decyloxy-4-phenyl]-adenosine (4i). Yield: 75
mg, 52%. Mp 119 121 °C. 1H NMR (400 MHz, CDCl3): δ = 8.35
(s, 1H), 8.11 (s, 1H), 7.62 (d, J = 7.2 Hz, 2H), 7.53 (d, J = 9.6 Hz,
2H), 7.30 (d, J = 7.2 Hz, 1H), 6.94 (d, J = 7.6 Hz, 2H), 5.91 (d,
J = 5.2 Hz, 1H), 4.82 (t, J = 6.0 Hz, 1H), 4.36 (d, J = 4.8 Hz, 1H),
4.28 (d, J = 1.6 Hz, 1H), 4.03-3.93 (m, 3H), 3.77 (d, J = 12.4 Hz,
1H), 3.42 (s, 2H), 2.58-2.41 (m, 4H), 1.93 (t, J = 5.4 Hz, 2H),
1.85-1.73 (m, 2H), 1.61-1.41 (m, 4H), 1.41-1.31 (m, 10H) ppm.
Biology. Binding Studies. [3H]DPCPX was purchased from
Amersham Biosciences (NL), [3H]CCPA was purchased
from Perkin-Elmer (NL). CHO cells expressing the human
adenosine A1 receptor (CHO-hA1) were provided by Dr. Andrea
Townsend-Nicholson, University College London, U.K., and
Dr. Stephen Hill, University of Nottingham, U.K. For [3H]D-
PCPX displacement experiments, reactions were performed in a
Data Analysis. Data were analyzed using nonlinear regression
analysis software available in GraphPad Prism 5.0. Differences
in obtained Ki, IC50, and EC50 values were assessed for statistical
significance using a paired Student’s t-test or one-way ANOVA
with a Bonferroni posthoc test where appropriate.
total volume of 100 μL, 50 mM Tris HCl pH 7.4, adenosine
3
deaminase 0.8 IU/mL, 1.6 nM [3H]DPCPX, and increasing con-
centrations of ligand with and without the presence of 10 μM
PD81,723. CHO-hA1 membranes containing 10 μg of protein
(Bmax = 12.8 pmol.mg-1 determined by [3H]DPCPX binding)
were added and the reaction incubated for 1 h at 25 °C in a shaking
water bath before termination by fast-flow filtration over GF/B
filters under reduced pressure with a Brandel harvester. Nonspecific
binding was determined with the addition of 10 μM CPA. Filters
Homology Modeling and Docking Studies. A homology model
of the A1AR was generated by using the “protein modeling”
module provided in MOE.38 The crystal structure of the adeno-
sine A2A receptor (PDB code: 3EML) was used as the template.
The sequence of the A1AR (accession code: p30542) was aligned
on the A2A receptor using “sequence editor”, which allows
organization and modification of amino acids at the chain and
residue levels. Amber99,39 parametrized particularly for pro-
teins and nucleic acids, was used as the force field. By default, ten
independent intermediate models were built based on permuta-
tional selection of different loop candidates and side chain
rotamers. The best scoring model according to the chosen
scoring function (GB/VI)40 was selected for further analysis
and was subjected to energy minimization. Finally, the mini-
mized model was inspected for unusual or geometrically incor-
rect stereochemistry by using an updated Ramachandran j-ψ
dihedral plot with refined reference data set.41 The outliers were
subsequently subjected to fine grain minimization to obtain
more typical j-ψ angles. The final model was used for docking
analysis of compounds 2 and 4h using the “docking module” in
MOE. The homology model of A1AR was prepared for docking
by adding hydrogens and partial charges using the protonate 3D
application given in the program. The active site was defined by
selecting amino acids, namely, Phe171, Asn254, Lys265, Thr270
and His278, corresponding to amino acids of the active site of
the A2A receptor for the co-crystallized ZM241385 in 3EML.
Ligands were placed in the active site by using the “Triangle
matcher” algorithm where poses are generated by superposition
of ligand atom triplets and triplets of receptor site points in a
systematic way. The docked poses were first energy minimized in
the active site and than ranked according to MM/GBVI binding
free energy estimation.
were washed three times with ice-cold 50 mM Tris HCl pH 7.4,
3
placed in vials, and counted on a liquid scintillation counter (Tri-
Carb 2900 TR, Perkin-Elmer). [3H]CCPA binding experiments
were performed as described for [3H]DPCPX binding experiments,
but with the following differences. Twenty micrograms of CHO-
hA1 membrane protein was added to a final reaction volume of
100 μL containing 2 nM [3H]CCPA and increasing concentrations
of modulator and incubated for 2 h at 25 °C before filtration and
scintillation counting as described above. For single point kinetics
assays, experiments were performed as above, but after two hours
incubation at 25 °C, dissociation was initiated by the addition of
100 μM CPA with or without the presence of modulator com-
pounds 2, 4d, and 4h. Reactions were stopped after 30 min by fast-
flow filtration as described above.
A1 Adenosine Receptor-Mediated [35S]GTPγS Binding. CHO-
hA1 membrane homogenates (5 μg, Bmax = 4.1 pmol.mg-1 deter-
mined by [3H]DPCPX binding) were equilibrated in
a
90 μL total volume of assay buffer (50 mM Tris, 100 mM NaCl,
10 mM MgCl, pH 7.4) containing 3 μM GDP and a range of
concentrations of either the monovalent agonist 2 or the bivalent
ligands 4d and 4h at 25 °C for 30 min. After this, 10 μL of
[35S]GTPγS (final concentration 0.3 nM) were added and incuba-
tion continued for 45 min at 25 °C. Incubation was terminated by
rapid filtration through a 96 well Unifilter (PerkinElmer, NL)
using a Filtermate Unifilter 96-well harvester (Perkin-Elmer).
Filters were washed three times with ice-cold assay buffer before
drying. Twenty-five microliters of Microscint scintillation cocktail
was added to each well, and plates were counted in a 1450 Trilux
Microbeta liquid scintillation and luminescence counter (Perkin-
Elmer). Typically a 5-fold stimulation above basal was observed.
A1 Receptor Mediated ERK1/2 Phosphorylation. Experi-
ments were performed using an ERK Surefire Alphascreen kit
(Perkin-Elmer, Groningen, NL) as per the vendor’s instructions.
Acknowledgment. This study was performed within the fra-
mework of Top Institute Pharma, project number D1-105. We
appreciate the help of our colleagues at the Amsterdam branch of
the LACDR in setting up the [35S]GTPγS binding assay.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8 3037
Supporting Information Available: Experimental details of the
increases chimaeric A1/A2A adenosine receptor coupling with Gs.
Biochem. J. 2006, 396, 139–146.
(22) de Ligt, R. A.; Rivkees, S. A.; Lorenzen, A.; Leurs, R.; IJzerman,
A. P. A “locked-on,” constitutively active mutant of the adenosine
A1 receptor. Eur. J. Pharmacol. 2005, 510, 1–8.
(23) Kourounakis, A.; Visser, C.; de Groote, M.; IJzerman, A. P.
Differential effects of the allosteric enhancer (2-amino-4,5-dimeth-
yl-trienyl)[3-trifluoromethyl) phenyl]methanone (PD81,723) on
agonist and antagonist binding and function at the human wild-
type and a mutant (T277A) adenosine A1 receptor. Biochem.
Pharmacol. 2001, 61, 137–144.
1
synthesis of the compounds described in this paper, their H
NMR and 13C NMR spectroscopic data, HRMS data and
HPLC data. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) Ma, P.; Zemmel, R. Value of novelty? Nat. Rev. Drug Discovery
2002, 1, 571–572.
(2) Overington, J. P.; Al-Lazikani, B.; Hopkins, A. L. How many drug
targets are there? Nat. Rev. Drug Discovery 2006, 5, 993–996.
(3) Ralevic, V.; Burnstock, G. Receptors for purines and pyrimidines.
Pharmacol. Rev. 1998, 50, 413–492.
(4) Fredholm, B. B.; IJzerman, A. P.; Jacobson, K. A.; Klotz, K. N.;
Linden, J. International Union of Pharmacology. XXV. Nomen-
clature and classification of adenosine receptors. Pharmacol. Rev.
2001, 53, 527–552.
(5) May, L. T.; Leach, K.; Sexton, P. M.; Christopoulos, A. Allosteric
modulation of G protein-coupled receptors. Annu. Rev. Pharmacol.
Toxicol. 2007, 47, 1–51.
(6) Christopoulos, A. Allosteric binding sites on cell-surface receptors: novel
targets for drug discovery. Nat. Rev. Drug Discovery 2002, 1, 198–210.
(7) Soudijn, W.; Van Wijngaarden, I.; IJzerman, A. P. Allosteric
modulation of G protein-coupled receptors: perspectives and
recent developments. Drug Discovery Today 2004, 9, 752–758.
(8) Gao, Z. G.; Jacobson, K. A. Allosterism in membrane receptors.
Drug Discovery Today 2006, 11, 191–202.
(9) Valant, C.; Sexton, P. M.; Christopoulos, A. Orthosteric/allosteric
bitopic ligands: going hybrid at GPCRs. Mol. Interv. 2009, 9, 125–135.
(10) Valant, C.; Gregory, K. J.; Hall, N. E.; Scammells, P. J.; Lew, M. J.;
Sexton, P. M.; Christopoulos, A. A novel mechanism of G protein-
coupled receptor functional selectivity. Muscarinic partial agonist
McN-A-343 as a bitopic orthosteric/allosteric ligand. J. Biol.
Chem. 2008, 283, 29312–29321.
(11) Steinfeld, T.; Mammen, M.; Smith, J. A.; Wilson, R. D.; Jasper,
J. R. A novel multivalent ligand that bridges the allosteric and
orthosteric binding sites of the M2 muscarinic receptor. Mol.
Pharmacol. 2007, 72, 291–302.
(12) Antony, J.; Kellershohn, K.; Mohr-Andra, M.; Kebig, A.; Prilla,
S.; Muth, M.; Heller, E.; Disingrini, T.; Dallanoce, C.; Bertoni, S.;
Schrobang, J.; Trankle, C.; Kostenis, E.; Christopoulos, A.; Holtje,
H. D.; Barocelli, E.; De Amici, M.; Holzgrabe, U.; Mohr, K.
Dualsteric GPCR targeting: a novel route to binding and signaling
pathway selectivity. FASEB J. 2009, 23, 442–450.
(13) Disingrini, T.; Muth, M.; Dallanoce, C.; Barocelli, E.; Bertoni, S.;
Kellershohn, K.; Mohr, K.; De Amici, M.; Holzgrabe, U. Design,
synthesis, and action of oxotremorine-related hybrid-type allos-
teric modulators of muscarinic acetylcholine receptors. J. Med.
Chem. 2006, 49, 366–372.
(24) Barbhaiya, H.; McClain, R.; IJzerman, A.; Rivkees, S. A. Site-
directed mutagenesis of the human A1 adenosine receptor: influ-
ences of acidic and hydroxy residues in the first four transmem-
brane domains on ligand binding. Mol. Pharmacol. 1996, 50, 1635–
1642.
(25) Ferguson, G. N.; Valant, C.; Horne, J.; Figler, H.; Flynn, B. L.;
Linden, J.; Chalmers, D. K.; Sexton, P. M.; Christopoulos, A.;
Scammells, P. J. 2-Aminothienopyridazines as novel adenosine A1
receptor allosteric modulators and antagonists. J. Med. Chem.
2008, 51, 6165–6172.
(26) Middleton, R. J.; Briddon, S. J.; Cordeaux, Y.; Yates, A. S.; Dale,
C. L.; George, M. W.; Baker, J. G.; Hill, S. J.; Kellam, B. New
fluorescent adenosine A1-receptor agonists that allow quantifica-
tion of ligand-receptor interactions in microdomains of single
living cells. J. Med. Chem. 2007, 50, 782–793.
(27) Thompson, R. D.; Secunda, S.; Daly, J. W.; Olsson, R. A. N6,
9-Disubstituted adenines: potent, selective antagonists at the A1
adenosine receptor. J. Med. Chem. 1991, 34, 2877–2882.
(28) Bruns, R. F.; Fergus, J. H. Allosteric enhancement of adenosine A1
receptor binding and function by 2-amino-3-benzoylthiophenes.
Mol. Pharmacol. 1990, 38, 939–949.
(29) Bruns, R. F.; Fergus, J. H.; Coughenour, L. L.; Courtland, G. G.;
Pugsley, T. A.; Dodd, J. H.; Tinney, F. J. Structure-activity relation-
ships for enhancement of adenosine A1 receptor binding by 2-amino-
3-benzoylthiophenes. Mol. Pharmacol. 1990, 38, 950–958.
(30) van der Klein, P. A.; Kourounakis, A. P.; IJzerman, A. P. Allosteric
modulation of the adenosine A1 receptor. Synthesis and biological
evaluation of novel 2-amino-3-benzoylthiophenes as allosteric
enhancers of agonist binding. J. Med. Chem. 1999, 42, 3629–3635.
(31) Karellas, P.; McNaughton, M.; Baker, S. P.; Scammells, P. J.
Synthesis of bivalent beta2-adrenergic and adenosine A1 receptor
ligands. J. Med. Chem. 2008, 51, 6128–6137.
(32) Xie, Z.; Bhushan, R. G.; Daniels, D. J.; Portoghese, P. S. Interac-
tion of bivalent ligand KDN21 with heterodimeric delta-kappa
opioid receptors in human embryonic kidney 293 cells. Mol.
Pharmacol. 2005, 68, 1079–1086.
(33) Childers, S. R.; Li, X.; Xiao, R.; Eisenach, J. C. Allosteric
modulation of adenosine A1 receptor coupling to G-proteins in
brain. J. Neurochem. 2005, 93, 715–723.
(34) Figler, H.; Olsson, R. A.; Linden, J. Allosteric enhancers of A1
adenosine receptors increase receptor-G protein coupling and
counteract Guanine nucleotide effects on agonist binding. Mol.
Pharmacol. 2003, 64, 1557–1564.
(35) Kollias-Baker, C. A.; Ruble, J.; Jacobson, M.; Harrison, J. K.;
Ozeck, M.; Shryock, J. C.; Belardinelli, L. Agonist-independent
effect of an allosteric enhancer of the A1 adenosine receptor in
CHO cells stably expressing the recombinant human A1 receptor.
J. Pharmacol. Exp. Ther. 1997, 281, 761–768.
(36) Avlani, V. A.; Gregory, K. J.; Morton, C. J.; Parker, M. W.;
Sexton, P. M.; Christopoulos, A. Critical role for the second
extracellular loop in the binding of both orthosteric and allosteric
G protein-coupled receptor ligands. J. Biol. Chem. 2007, 282,
25677–25686.
(14) Portoghese, P. S.; Larson, D. L.; Sayre, L. M.; Yim, C. B.;
Ronsisvalle, G.; Tam, S. W.; Takemori, A. E. Opioid agonist and
antagonist bivalent ligands. The relationship between spacer length
and selectivity at multiple opioid receptors. J. Med. Chem. 1986, 29,
1855–1861.
(15) Perez, M.; Jorand-Lebrun, C.; Pauwels, P. J.; Pallard, I.; Halazy, S.
Dimers of 5HT1 ligands preferentially bind to 5HT1B/1D receptor
subtypes. Bioorg. Med. Chem. Lett. 1998, 8, 1407–1412.
(16) Halazy, S.; Perez, M.; Fourrier, C.; Pallard, I.; Pauwels, P. J.;
Palmier, C.; John, G. W.; Valentin, J. P.; Bonnafous, R.; Martinez,
J. Serotonin dimers: application of the bivalent ligand approach to
the design of new potent and selective 5-HT1B/1D agonists. J. Med.
Chem. 1996, 39, 4920–4927.
(17) 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 delta-
kappa heterodimers. J. Med. Chem. 2004, 47, 2969–2972.
(18) Portoghese, P. S.; Larson, D. L.; Yim, C. B.; Sayre, L. M.;
Ronsisvalle, G.; Lipkowski, A. W.; Takemori, A. E.; Rice, K. C.;
Tam, S. W. Stereostructure-activity relationship of opioid agonist
and antagonist bivalent ligands. Evidence for bridging between
vicinal opioid receptors. J. Med. Chem. 1985, 28, 1140–1141.
(19) Kramer, R. H.; Karpen, J. W. Spanning binding sites on allosteric
proteins with polymer-linked ligand dimers. Nature 1998, 395, 710–713.
(20) 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.
(37) Jaakola, V. P.; Griffith, M. T.; Hanson, M. A.; Cherezov, V.;
Chien, E. Y.; Lane, J. R.; IJzerman, A. P.; Stevens, R. C. The 2.6
angstrom crystal structure of a human A2A adenosine receptor
bound to an antagonist. Science 2008, 322, 1211–1217.
(38) Molecular Operating Environment (MOE 2008.10); Chemical Com-
puting Group, Inc.; Quebec, Canada.
(39) Wang, J. M.; Cieplak, P.; Kollman, P. A. How well does a
restrained electrostatic potential (RESP) model perform in calcu-
lating conformational energies of organic and biological mole-
cules? J. Comput. Chem. 2000, 21, 1049–1074.
(40) Labute, P. The generalized Born/volume integral implicit solvent
model: Estimation of the free energy of hydration using London
dispersion instead of atomic surface area. J. Comput. Chem. 2008,
29, 1693–1698.
(41) Lovell, S. C.; Davis, I. W.; Adrendall, W. B.; de Bakker, P. I. W.;
Word, J. M.; Prisant, M. G.; Richardson, J. S.; Richardson, D. C.
Structure validation by C alpha geometry: phi,psi and C beta
deviation. Proteins: Struct. Funct. Bioinf. 2003, 50, 437–450.
(21) Bhattacharya, S.; Youkey, R. L.; Ghartey, K.; Leonard, M.;
Linden, J.; Tucker, A. L. The allosteric enhancer PD81,723