Neoceptor concept based on molecular complementarity in GPCRs: A mutant adenosine A3 receptor with selectively enhanced affinity for amine-modified nucleosides

Article abstract of DOI:10.1021/jm010232o

Adenosine A₃ receptors are of interest in the treatment of cardiac ischemia, inflammation, and neurodegenerative diseases. In an effort to create a unique receptor mutant that would be activated by tailor-made synthetic ligands, we mutated the

Full text of DOI:10.1021/jm010232o

J . Med. Chem. 2001, 44, 4125-4136
4125
Neocep tor Con cep t Ba sed on Molecu la r Com p lem en ta r ity in GP CRs: A Mu ta n t
Ad en osin e A3 Recep tor w ith Selectively En h a n ced Affin ity for Am in e-Mod ified
Nu cleosid es
Kenneth A. J acobson,*,† Zhan-Guo Gao,† Aishe Chen,† Dov Barak,†,‡ Soon-Ai Kim,† Kyeong Lee,† Andreas Link,§
Philippe Van Rompaey,| Serge van Calenbergh,| and Bruce T. Liang
Molecular Recognition Section, Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney
Diseases, National Institutes of Health, Bethesda, Maryland 20892, Department of Medicine, Cardiovascular Division, and
Department of Pharmacology, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania 19104, Laboratorium
voor Medicinale Chemie (FFW), Harelbekestraat 72, B-9000 Gent, Belgium, and Institut fu¨r Pharmazie, Fachbereich Chemie,
Universita¨t Hamburg, Bundesstrasse 45, 20146 Hamburg, Germany
Received May 24, 2001
Adenosine A3 receptors are of interest in the treatment of cardiac ischemia, inflammation,
and neurodegenerative diseases. In an effort to create a unique receptor mutant that would be
activated by tailor-made synthetic ligands, we mutated the human A3 receptor at the site of a
critical His residue in TM7, previously proposed to be involved in ligand recognition through
interaction with the ribose moiety. The H272E mutant receptor displayed reduced affinity for
most of the uncharged A3 receptor agonists and antagonists examined. For example, the
nonselective agonist 1a was 19-fold less potent at the mutant receptor than at the wild-type
receptor. The introduction of an amino group on the ribose moiety of adenosine resulted in
either equipotency or enhanced binding affinity at the H272E mutant relative to wild-type A3
receptors, depending on the position of the amino group. 3′-Amino-3′-deoxyadenosine proved
to be 7-fold more potent at the H272E mutant receptor than at the wild-type receptor, while
the corresponding 2′- and 5′-amino analogues did not display significantly enhanced affinities.
An 3′-amino-N6-iodobenzyl analogue showed only a small enhancement at the mutant (Ki )
320 nM) vs wild-type receptors. The 3′-amino group was intended for a direct electrostatic
interaction with the negatively charged ribose-binding region of the mutant receptor, yet
molecular modeling did not support this notion. This design approach is an example of
engineering the structure of mutant receptors to recognize synthetic ligands for which they
are selectively matched on the basis of molecular complementarity between the mutant receptor
and the ligand. We have termed such engineered receptors “neoceptors”, since the ligand
recognition profile of such mutant receptors need not correspond to the profile of the parent,
native receptor.
In tr od u ction
tion of receptor engineering, agonist design, and gene
therapy. Elements of this potentially general approach
include (a) engineering of a receptor protein to recognize
synthetic ligands, for which it was selectively modified
according to the molecular complementarity of the
respective binding elements, while retaining its capacity
for signal transduction (neoceptor); (b) synthesis of novel
agonists that are not effective at the native receptor but
do activate the engineered receptor (neoligand); and (c)
a delivery vector to provide for selective expression of
the neoceptor in the target area.
Therapeutic intervention using agonists is subject to
side effects related in part to the widespread occurrence
of the corresponding receptor throughout the body.1,2
Currently, the specificity of a given drug for a target
organ is usually achieved through manipulation of its
pharmacokinetic properties. Past attempts to overcome
this problem included generation of prodrugs that were
to be preferentially activated at the target organ, either
through enhanced metabolic processes or through a
unique enzymatic system characteristic to the organ.3
Our goal was to introduce and investigate a novel
concept for therapeutic intervention at a specific ana-
tomical and/or physiological locus, through a combina-
One of the numerous cases where agonist therapy
has been problematic due to the widespread occurrence
of receptors is the adenosine receptor family. For ex-
ample, the hypotensive and bradycardiac side effects of
adenosine agonists have been in part responsible for the
difficulty of developing adenosine-based therapeutics
for cardio- and cerebropotection.1,3-6 The only adenosine
agonist approved so far for clinical use has been adeno-
sine itself, based on its short duration of action in the
treatment of supraventricular tachycardia and in radio-
nuclide imaging.7 Thus, engineering of novel receptor-
* To whom correspondence should be addressed at Molecular
Recognition Section, Bldg. 8A, Rm. B1A-19, NIH, NIDDK, LBC,
Bethesda, MD 20892-0810. Telephone: (301) 496-9024. Fax: (301) 480-
8422. E-mail: kajacobs@helix.nih.gov.
† National Institutes of Health.
‡ On leave from Israel Institute for Biological Research, Ness Ziona,
Israel.
§ Universita¨t Hamburg.
| Laboratorium voor Medicinale Chemie (FFW).
University of Pennsylvania Medical Center.
10.1021/jm010232o CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/16/2001

4126 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
J acobson et al.
F igu r e 1. A3 adenosine receptor agonists and antagonists. Wild-type and mutant receptor affinities appear in Table 1.
ligand interactions in the adenosine receptor family,
through specific tailoring of both the receptors and the
ligands, would provide a suitable and relevant system
for investigation of the neoceptor-neoligand concept.
Adenosine is released in large amounts during is-
chemia and has been shown to be protective in the
heart,8 brain,5,9 and other organs. Adenosine, when
elevated prior to ischemia in cardiac tissue,10,11 “pre-
conditioned” the heart and protected it against injury
during a subsequent period of prolonged ischemia.
Synthetic agonists selective for either A1 or A3 adenosine
receptor simulated this preconditioning effect.12 The
beneficial effects were seen following acute activation
of A1 or chronic activation of A3 receptors. In addition
to cardioprotection, adenosine A3 receptor ligands have
been proposed for the treatment of stroke,13 inflamma-
tion,14 and glaucoma.15 In a model of global brain
ischemia in gerbils, both A1 and A3 adenosine receptor-
selective agonists had protective effects,5 as judged
by the histochemical and behavioral outcome follow-
ing recovery. Thus, there has been much interest in
the development of new therapeutic agents acting at
adenosine A3 receptors.
Ligand recognition in adenosine receptors, principally
A1 and A2A receptors, has been extensively investigated
using mutagenesis and molecular modeling.16-20 The
putative nucleoside binding site, which is highly ho-
mologous among subtypes of the adenosine receptors,
is proposed to involve transmembrane helices (TMs) 3,
5, 6, and 7. Two conserved His residues (6.52 and 7.43,
by the notation of van Rhee and J acobson21) in TMs 6
and 7 of A1 and A2A receptors are considered to be
among the most important amino acids involved in
binding to adenosine. An assembly of aromatic amino
acid side chains in the human A2A receptor, principally
in TMs 5 and 6, is proposed to recognize the adenine
moiety. The ribose moiety is likely coordinated to
hydrophilic residues in TMs 3 and 7.19,22,23 In fact,
several hydroxyl-containing residues, Thr88 (3.36) and
Ser277 (7.42) of the human A2A receptor, have been
shown to be associated exclusively with agonist, but not
antagonist, recognition.19,22
In the present study, we have utilized this hypotheti-
cal bound orientation of adenosine to identify a site on
TM7 thought to be in proximity to the ribose moiety and
amenable to introduction of a charged group intended
for electrostatic interaction with the ligand. The site
chosen was the conserved His (H272), present also in
the A3 receptor, which we have mutated to Glu. Accord-
ing to molecular modeling, this mutation would be
expected to decrease the affinity of simple adenosine
analogues,19,23,24 except when strategically modified
by the introduction of an amino group on the ribose
moiety. Since alteration of the ribose moiety substitution
pattern was also known to greatly diminish the affinity
at adenosine receptors,25 such substitution would ensure
that the neoligand would not activate endogenous
receptors at an effective concentration for the neoceptor.
Our results suggested both the viability of the neo-
ceptor-neoligand concept and the need for further
optimization of the A3 neoceptor-neoligand interactions.
Resu lts
Design of a Neocep tor : Cr ea tion of a Mu ta n t A3
Recep tor Th a t Is Activa ted Selectively by Novel
Agon ist Der iva tives. On the basis of previous studies
of ligand recognition in adenosine receptors,16,19,23,24
a
His residue in TM7 of the A3 adenosine receptor was
selected as the site for mutagenesis, i.e., the introduction
of a negative charge to be complementary to an amine-
derivatized ligand. We examined the recognition of both
known adenosine ligands (Figure 1) and synthetic
agonist analogues (Figure 2) designed as neoligands, for
selective recognition by the H272E mutant receptor (see
below).
Mu ta n t vs Wild -Typ e A3 Recep tor s Usin g Kn ow n
Ad en osin e Recep tor Liga n d s. The receptor binding
affinities of various adenosine agonist and antagonist
derivatives were measured in standard binding assays
using wild-type and H272E mutant human A3 receptors.
The high affinity of the radioligand [125I]I-AB-MECA
([125I]N6-(4-amino-3-iodobenzyl)-5′-N-methylcarbamoyl-
adenosine),26 2b, was retained in the mutant receptor,

Neoceptor from Adenosine A3 Receptor
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4127
F igu r e 2. Derivatives of adenosine tested as novel receptor agonists. Wild-type and mutant receptor affinities appear in Table
2.
Ta ble 1. Binding Affinity of Known Adenosine Agonists and Antagonists at Wild-type and Mutant (H272E) Human A3 Adenosine
Receptorsa,b
Ki (nM)
Ki (nM)
agonists
WT
H272E
antagonists
WT
H272E
2b [125I]AB-MECA (Kd)
1.2 ( 0.2
169 ( 32
650 ( 130
4.3 ( 1.6
1.6 ( 0.3
2.1 ( 0.3
4
5
6
253 ( 87
61.7 ( 23.4
110 ( 45
3260 ( 994
1690 ( 343
20 100 ( 4500
1a
1b
2a
3
3200 ( 770
11 600 ( 4200
20.1 ( 3.6
9.9 ( 2.6
a
b
Structures provided in Figure 1. Results are from three independent experiments performed in duplicate, using [125I]AB-MECA at
a concentration of 1 nM.
thus enabling the determination of the affinities of a
wide range of competing ligands. Ki values for these
ligands are shown in Table 1, and representative
binding curves are shown in Figure 3. The closest mimic
of the affinity of adenosine itself27 was 2-chloroadeno-
sine, 1b, which was 18-fold less potent in binding at the
H272E mutant receptor than at the wild-type receptor.
In most other cases, the affinity of competing ligands,
1-7, was significantly reduced in the mutant receptor.
For example, compound 1a was 19-fold less potent at
the mutant receptor than the wild-type receptor. The
potent antagonist, xanthine amine congener 6,19 which
contains a distal amino group, was 180-fold less potent
at the mutant than the wild-type receptor. Other
ligands, such as potent A3 receptor agonists 2a and the
rigid analogue 3,13,28 were shifted to lower affinity in
binding to the mutant receptor by smaller factors, i.e.,
5- and 6-fold, respectively.
Syn t h esis of Neoliga n d s. The amino derivatives
of adenosine (7-9), a N6-cyclopentyladenosine (11), and
a 5′-uronamidoadenosine (14) were prepared as de-
scribed.29-31 A 3′-guanidino derivative of adenosine (10)
was obtained in the protected form (15) from the
corresponding amine using di-Boc-triflylguanidine,32
followed by deprotection (Figure 4a). A sugar-modified
analogue of 2, i.e. 13, was available from a previous
study.25 A 3′-amino derivative (12), which also contained
an N6-substituent favorable for A3 receptor affinity and
selectivity, was prepared as shown in Figure 4b. 3-Azido-
1,2-di-O-acetyl-5-O-(4-methylbenzoyl)-3-deoxy-â-D-ribo-
furanose, 16, was prepared in seven steps from xylose
following a slightly altered literature procedure.33 Cou-
pling with N6-(3-iodobenzyl)adenine by the method of
Vorbru¨ggen et al.,34 followed by alkaline deprotection,
yielded 18% of the â-anomer 17. Reduction of the azido
moiety with triphenylphosphine led to smooth conver-
sion of 17 to the desired amine 12 without detectable
loss of the benzylic N6-substituent.
Bin d in g to Mu ta n t vs Wild -Typ e A3 Recep tor s.
The receptor binding affinities of amine-functionalized
adenosine derivatives were measured in standard bind-
ing assays using wild-type and H272E mutant human
A3 receptors expressed in COS-7 cells, rat A3 receptors
expressed endogenously in RBL-2H3 cells, and rat brain
A1 and A2A receptors.6,26,35,36 Ki values for these ligands
are shown in Table 2, and representative binding curves
are shown in Figure 5.
The introduction of an amino group on the ribose
moiety of adenosine resulted in either equipotency or
enhanced binding affinity at the H272E mutant relative
to wild-type A3 receptors, depending on the position of
the amino group. 3′-Amino-3′-deoxyadenosine, 8, proved

4128 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
J acobson et al.
F igu r e 3. Binding of known A3 adenosine receptor agonists and antagonists to wild type and mutant receptors (9), wild type;
(2), H272E mutant receptor).
to be 7-fold more potent at the H272E mutant than
the wild-type receptor. Two other isomers, 2′-amino-
2′-deoxyadenosine, 7, and 5′-amino-5′-deoxyadenosine,
9, an inhibitor of adenosine kinase,37 did not display
significantly enhanced affinity. The affinity of a 3′-
guanidino analogue, 10, was enhanced by 4-fold in the
mutant versus wild-type receptors. A 2-aminoethyl
analogue of 1a , i.e., 14,38 was only 2-fold more potent
at the mutant receptor. A 3′-amino-N6-iodobenzyl ana-
logue, 12, showed a 3-fold enhancement at the mutant
vs wild-type receptors, with a Ki value of 320 nM.
The ability of both mutant and wild-type receptors
to activate second messengers was demonstrated. Both
compounds 1b and 11 at 10 µM inhibited cyclic AMP
production39 stimulated by forskolin in COS-7 cells
expressing either the wild-type or H272E mutant recep-
tor (Figure 6). Furthermore, both the wild-type and
H272E mutant receptors in the presence of 10 µM of
compound 2a were shown to fully activate phospholi-
pase C (data not shown), as determined by the method
reported.40
Molecu la r Mod elin g. A model of the human A3
receptor was built in homology to the recently published
X-ray structure of bovine rhodopsin.41 The model in-
cluded the seven TMs and the second extracellular loop
(EL2). In this model, residue His272 was within inter-
action distance from Glu19 (1.39) in TM1 (Nπ-Oꢀ
distance, 2.40 Å; Figure 7a). An analogous interaction

Neoceptor from Adenosine A3 Receptor
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4129
F igu r e 4. Synthesis of novel adenosine derivatives 10 and 12 containing 3′-guanidino (a, top) and 3′-amino (b, bottom) groups.
Reagents: (a) (i) di-Boc-triflylguanidine, DMF (ii) TFA; (b) (i-1) N6-(3-iodobenzyl)adenine, HMDS, TMSCl, TMSOTf, pyridine,
(i-2) 0.1 N NaOMe, MeOH (18% combined yield); (ii) Ph3P, pyridine, NH4OH.
Ta ble 2. Binding Affinity of Amine-Derivatized Adenosine Analogues at Wild-Type and Mutant (H272E) Human A3 and Rat
Adenosine Receptorsa,b
b
Ki (µM)
compda
WT
H272E
rat A3
rat A1
rat A2A
f
7
8
9
10
11
12
13
14
c
301 ( 142
75 ( 32
d (10-4
)
d (10-4
)
)
)
)
)
d (10-4
d (10-4
d (10-4
d (10-4
d (10-4
)
)
)
)
)
f
f
442 ( 121
84 ( 26
d (10-4
d (10-4
d (10-4
d (10-4
c
425 ( 217
33.3 ( 7.1
0.19 ( 0.08
0.32 ( 0.10
2.3 ( 0.57
9.6 ( 3.5
d (10-4
ND
)
)
130 ( 34
0.54 ( 0.13
0.87 ( 0.18
4.6 ( 1.3
19.6 ( 3.1
f
d (10-4
ND
8.1 ( 0.47
28 ( 7.5
3.40 ( 0.79e
6.69 ( 0.74e
d (10-4
d (10-4
)
)
e
e
e
14.7 ( 2.5e
d (10-4
)
a
b
Structures provided in Figure 2. Results are from three independent experiments performed in duplicate, using [125I]AB-MECA at
a concentration of 1 nM at human A3 receptors expressed in COS-7 cells or at rat A3 receptors in RBL-2H3 cells, unless noted. Rat A1 and
A
2A receptors were in rat brain. c IC50 values of compounds 7 and 9 at the wild type receptor were estimated to be approximately 1 mM.
d
The inhibition of [125I]AB-MECA binding by 7 and 9 at the highest concentration examined (1 mM) was 51 and 54%, respectively. <10%
displacement of binding at the indicated concentration (M). e From Gallo-Rodriguez et al.38 and J acobson et al.25 using recombinant rat
A3 receptors expressed in CHO cells. f IC50 value of 9 in the absence of adenosine deaminase estimated to be approximately 100 µM.
Compounds 7, 8, and 11 gave identical results in the absence of adenosine deaminase.
had already been proposed for the A2A receptor,23,24 in
order to explain the observed involvement of both
residues in agonist binding. In the present model, the
carboxylate group of Glu19 appeared to interact also
with Tyr265 (7.36) and Ser73 (2.65), resulting in a
relatively rigid juxtaposition of the imidazole moiety of
His272 relative to other elements of the ligand binding
environment (Figure 7b).
Examination of the optimized model of the 1b-A3
receptor complex showed that the N6-amine nitrogen
was located within H-bonding distances of the amide
oxygen of Trp243 (6.48) and Oγ of Ser247 (6.52) of 2.76
and 2.51 Å, respectively (Figure 7b). The 2′-hydroxy
substituent of the ribose ring was adjacent to both the
Oꢀ of Gln167 and Nú of Lys152 (3.16 and 3.75 Å,
respectively). The corresponding 3′-hydroxy substituent
was within H-bonding distance from His272, and the
terminal oxygen interacted with Ser271 (7.42) and
Asn274 (7.45) in TM7. Thus, 1b seemed to be accom-
modated by interactions with residues of TMs 6 and 7
and of EL2, with both 2′- and 3′-ribose hydroxy substit-
uents proximal to basic residues.
wedged between TM5 and TM6, interacting with resi-
dues Phe182 (5.43), Ile186 (5.47), and Phe187 (5.48).
Consequently, the whole ligand was displaced away
from TM7, compared to the corresponding complex of
1b. The N6 was still within interaction distance from
Ser247 but was over 5 Å away from the amide oxygen
of Trp243. Furthermore, the 3′-hydroxy substituent did
not seem to interact with His272 (the 3′-O-Nτ distance
was >4.5 Å). Reorientation of the latter was prevented
by interaction with Glu19 (see above).
We used modeling to test the hypothesis that an
electrostatic interaction between the positively charged
ligand and the now negatively charged ribose-binding
region of the receptor led to the affinity enhancement
of 8 at the H272E mutant receptor. Replacement of
His272 by glutamate resulted in a structure with two
adjacent carboxylates, one of which was most likely
protonated. Since the mobility of Glu19 was restricted
by interactions with Tyr265 and Ser73, the carboxylate
group of Glu272 was held in an orientation that pre-
vented a direct H-bond interaction with the 3′-hydroxy
ribose substituent (Figure 7a). This was consistent with
the 20-fold lower affinity of the H272E mutant toward
agonists, such as 1a or 1b, while the corresponding
In the optimized model of the 2a -A3 receptor complex
(Figure 7c) the N6-benzyl substituent appeared to be

4130 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
J acobson et al.
F igu r e 5. Binding of novel adenosine derivatives at wild-type and H272E mutant adenosine A3 receptors (9), wild type; (2),
H272E mutant receptor).
effect on the agonists related to 2a was much smaller.
The notion that, due to the H272E mutation, a polar
interaction was lost was also consistent with the 15-
20-fold affinity decrease of the mutant receptor toward
the three antagonists examined (Table 1). In all of those
cases, modeling suggested that Glu272 could not replace
His in accommodating the A3 ligands through H-
bonding interactions.
Toward this goal we have both mutated the A3 receptor
and chemically modified the corresponding agonists,
aiming to preserve the structural complementarity
required for agonist function. On one hand, adenosine
analogues carrying modifications of the ribose 2′- and
3′-substituents were known to be mostly inactive as
agonists.25 Thus, we concentrated on aminoadenosines
in order to avoid activation of the wild-type A3 receptors.
On the other hand, simple aminoadenosines were nearly
isosteric with the corresponding adenosines, implying
that the former could act as agonists, provided a proper
juxtaposition of its binding elements with those of an
appropriate receptor could be achieved. For instance,
Discu ssion
We have investigated an approach to target agonist
therapy to a specific organ or tissue based on selective
activation of mutant receptors by synthetic ligands.

Neoceptor from Adenosine A3 Receptor
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4131
F igu r e 6. Cyclic AMP production by COS-7 cells expressing
wild-type and H272E mutant human A3 receptors after addi-
tion of compound 1b (10 µM) or compound 11 (10 µM). Values
are means ( SEM of three independent experiments per-
formed in duplicate. For values marked with an asterisk, P <
0.05 compared with control.
introduction by mutagenesis of an acidic residue within
the ligand binding site could have resulted in an
electrostatic interaction with the amine substituent of
the ligand.
Position 272 on TM7 of the human A3 receptor (7.43),
selected for such mutagenesis, since it appeared to play
an important role across the GPCR family. In rhodopsin,
this is the location of the Lys residue that forms a
Schiff base with retinal. In all four of the adenosine
receptors this site is occupied by His, which has been
proposed to be critical for recognition of the ribose or
ribose-like moiety common to all adenosine agonists
thus far reported. At the A3 receptor, His at this site
was proposed as the basis for enhanced affinity of
xanthine-7-ribosides relative to the parent xanthines.27
In the A1 receptor, mutation of this His to Ala resulted
in decreased affinity of both agonists and antagonists.
In the A2A receptor, this site has been mutated to Ala
with the loss of high-affinity binding of both agonists
and antagonists,19 while mutation to Tyr preserved the
ability to bind ligands.24 Thus, substitution that pre-
served H-bonding capability was allowed at this critical
site. Substitution of His272 with Glu was the first
example of a nonaromatic residue at this position in
adenosine receptors that still allowed ligand recognition.
It was especially surprising in light of the proposal that
in the human A2A receptor this His appeared to be
coupled spatially to a Glu in TM1 through the formation
of a H-bond.24
The affinities of the aminoadenosine analogues 7-14,
for the H272E mutant A3 receptor were higher than for
the corresponding wild type, demonstrating that in
principle A3 receptors could be engineered for selective
interaction with synthetic agonists. In particular this
was evident from the 7-fold affinity enhancement of 8
toward the mutated receptor. Thus, the notion that the
H272E mutant A3 receptor could be selectively activated
in the presence of wild-type A3 receptors appeared
feasible.
The role of residue His272 in accommodating A3
agonists and antagonists, as well as the consequences
of its replacement by glutamate were investigated by
molecular modeling19,23 of the ligand-receptor com-
plexes. The 3′-amino group was intended for a direct
electrostatic interaction with the negatively charged
F igu r e 7. (a, top) Details of the putative H-bonding inter-
actions among residues in proximity to the binding site of
the human A3 receptor either in the native receptor (His in
blue and white) or for the H272E mutant (Glu in yellow).
H-bonding distances to the carboxylate group of Glu19 are
shown. (b, middle) Docked conformation of the nonselective
adenosine agonist 1b showing its position with respect to
critical residues in the putative binding site of the wild-type
human A3 receptor. (c, bottom) Docked conformation of the A3
selective adenosine agonist 2a showing its position with
respect to critical residues in the putative binding site of the
wild-type human A3 receptor.

4132 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
J acobson et al.
ribose-binding region of the mutant receptor, yet mo-
lecular modeling did not support this notion. In brief,
the higher affinity of 3′-amino-3′-deoxyadenosine at the
engineered A3 receptor appeared to result from the lack
of a repulsion (positively charged ligand and a positively
charged His272 side chain in the WT receptor), rather
than the attractive force of opposite charges. It is also
possible that a water molecule fills the space between
the 3′-amino group and Glu272, thus allowing water-
mediated H-bonding.
Previous models of an interaction between His278 of
A2A (position equivalent to 272 in the A3 receptor)23 and
adenosine suggested the involvement of both of the
ribose hydroxy substituents on binding. Yet substitution
of glutamate at position 272 had a larger effect on 8
than on 7, indicating that His/Glu272 interacted pre-
dominantly with the 3′-ribose substituent. Conversely,
for aminoadenosines where N6 was substituted by a
cycloalkyl (11) or iodobenzyl (12, 13), the effect of
Glu272 was small, irrespective of the position of the
amino ribose substituent (see Table 2). In addition,
while the affinity of 8 toward the rat A3 receptor
resembled that of the human H272E A3 receptor, the
corresponding affinity of 13 was similar to that of the
human wild-type A3 receptor. Rationalization of these
differences by means of molecular modeling of the
respective complexes has provided a more detailed
insight into the specific ligand-A3 receptor interactions,
allowing for further suggested modification of the A3
receptor binding environment.
From molecular models of the human A3 receptor
complexed with 8 (Figure 8) it appears that the nearly
1000-fold lower affinity, relative to 1b, results from a
loss of a H-bond interaction of the 3′-hydroxy substitu-
ent and the electrostatic repulsion of the 3′-amine
substituent and His272. Replacement at position 272
by glutamate did not restore binding with the 3′-
substituent. As already mentioned Glu272 was not
available for direct contact with the ligand due to an
interaction with Glu19. The gain of affinity for the
aminoadenosines seemed therefore to depend on relief
of electrostatic repulsion, being more pronounced for 8
than for 11 and 12, where the ligands would be
displaced toward TM5 (including the 3′-amine substitu-
ent). In this respect, the finding that affinity of the wild-
type rat A3 receptor toward 8 was similar to that of the
human H272E mutant A3 receptor was particularly
interesting. As already mentioned, the model of the
human A3 receptor indicated that the mobility of Glu19
was restricted by H-bond interactions with Tyr265 and
Ser73, which immobilized the Glu19 carboxylate be-
tween TM7 and TM2. In the rat A3 receptor, cysteine
occurs at the position corresponding to 265, thus prob-
ably allowing for motion of the His272-Glu19 assembly
in the direction of TM2 and away from the 3′-amino
substituent (model not shown). This interpretation
was consistent with the notion that replacement of
His272 resulted mostly in removal of electrostatic
repulsion but did not restore binding interaction with
the 3′-substituent of the adenosine derivatives. In the
case of 13 where such repulsive interaction was less
pronounced (see above), the affinity at the rat A3
receptor resembled those of both the human wild-type
and the H272E A3 receptors.
F igu r e 8. Docked conformation of the neoligand 8 in relation
to the helical bundle and EL2 of the H272E mutant human
A3 receptor. Views from the plane of the membrane (a, top:
TMs are indicated and EL2 is highlighted) or from the
extracellular region (b, bottom: showing TM side chains
mentioned in the text) are shown.
The notion that Glu272 did not interact directly with
the ligands was consistent with the finding that the A3
receptor affinity toward N-benzyl-substituted analogues
was insensitive to the nature of the residue at position
272. The corresponding affinity toward 1b was 20-fold
lower (see Table 1). Furthermore, the affinity of the
H272E mutant A3 receptor toward 1b was only 7-fold
higher than that toward the 3′-amino derivative 8,
suggesting that the binding environments for the two
ligands were similar. In fact, the lower affinity toward
8 was likely due to repulsive interactions with Lys152,
which, according to our model, was vicinal to the
2′-hydroxy substituent. This proximity may be respon-
sible for the low affinity of both human and rat A3
receptors toward the aminoadenosine 7. As already
suggested, the lack of Glu272 participation in ligand
accommodation, and in particular in interaction with
3-aminoadenosines, was due to its interaction with

Neoceptor from Adenosine A3 Receptor
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4133
deoxyadenosine (8),29 9-(3-amino-3-deoxy-â-D-xylofuranosyl)-
N6-cyclopentyladenine (11),31 and 1438 were prepared according
to known procedures.
Full-length cDNA encoding the human adenosine A3
receptor was kindly provided by M. Atkinson, A. Townsend-
Nicholson, and P. R. Schofield (Garvan Medical Institute,
Sydney, Australia) and was subcloned in pcDNA3 as pcDNA3/
hA1R, and pcDNA3/hA3R. The vector pcDNA3 was obtained
from Invitrogen (Carlsbad, CA).
Glu19. Consequently, replacement of Glu19 by a residue
that would not restrict the mobility of Glu272 may
result in a receptor with considerably higher affinity
toward aminoadenosines such as 8, and also a more
wild-type-like affinity toward the endogenous ligand
adenosine. Alternatively, the replacement of Ser73 by
a nonpolar residue such as Ala may allow enhanced
mobility of the Glu272-Glu19 assembly toward TM7 (see
above), bringing the Glu272 carboxylate within interac-
tion distance of the ligands. Thus, the doubly mutated
H272E/E19X and H272E/S73X mutant A3 receptors
may be more suitable as neoceptors for the aminoad-
enosine neoligands.
The notion that GPCRs can be engineered to accom-
modate unnatural ligands has been investigated in
other cases. For example, Schwartz and co-workers42
have engineered GPCRs to have the ability to bind
zinc ions through complexation with multiple His
residues. In those cases, the zinc ion bound as an
antagonist or partial agonist.43 More recently Conklin
and co-workers44 engineered the κ opioid receptor to
respond exclusively to synthetic small molecule ligands
and not to the receptor’s natural ligand. In this elegant
study, impressive selectivity toward bremazocine was
achieved, demonstrating the extent to which receptor
properties could be modified without compromising
functionality. Strader and co-workers45 also studied the
microscopic complementarity of functionality in receptor
binding. However, no attempt was made to modify both
the receptors and the ligands in the manner proposed
in the present study.
The feasibility of a tailor-made agonist (neoligand) to
interact selectively a mutant receptor (neoceptor) indi-
cates that this novel therapeutic approach, in which
such a neoceptor would be introduced specifically into
a target organ through gene transfer,46,47 may be
possible. Gene transfer to the heart has been demon-
strated,48-50 and gene therapy toward the goal of
cardioprotection has already been proposed.4 The trans-
fection of the adenosine A1 or A3 into a cardiac myocyte
culture enhances the protective effect of either endog-
enous adenosine or an exogenously added, synthetic
agonist of the appropriate receptor. By this approach,
a neoceptor could be genetically delivered to a target
organ, followed by the administration of a selective,
tailor-made ligand, as needed. This agent could be ad-
ministered in a dose range in which only the neoceptor,
not the endogenous parent receptor, would be activated.
P r ep a r a tion of Mu ta n t Recep tor s. Procedures for the
construction of the mutant receptor are provided elsewhere.40
The plasmids expressing mutant A3 adenosine receptor were
constructed as described in the instruction manual of the
QuikChange TM Site-Directed Mutagenesis Kit (La J olla, CA).
The plasmid pcDNA3/hA3-HA40 with cDNA of wild-type A3
adenosine receptor was used as a template for PCR. The
pair of primers that contained the desired mutation (histamine
to glutamic acid) was from Biosources Co. (Laurel, MD). Their
sequences were 5′-atcctgctgtccgaggccaactccatg-3′; 5′-catggagt-
tggcctcggacagcaggat-3′. Following PCR, the PCR products were
digested with Dpn I and transformed into E. coli. The mutant
plasmid was identified by sequencing, transfected in Cos-7
cells and expressed transiently.
1
Syn th esis. H NMR spectra were obtained with a Bruker
DRX 500 spectrometer. The solvent signal of DMSO-d6 was
used as a secondary reference. All signals assigned to amino
and hydroxyl groups were exchangeable with D2O. Exact mass
measurements were performed on a quadrupole/orthogonal-
acceleration time-of-flight (Q/oaTOF) tandem mass spectrom-
eter (qTOF 2, Micromass, Manchester, U.K.) equipped with a
standard electrospray ionization (ESI) interface. Samples were
infused in a 2-propanol:water (1:1) mixture at 3 µL/min.
P r ep a r a tion of 3′-Deoxy-3′-gu a n id in oa d en osin e, Tr i-
flu or oa cetic Acid Sa lt (10). 3′-Amino-3′-deoxyadenosine (8,
5 mg, 0.019 mmol) was added to a solution of di-Boc-
triflylguanidine32 (10 mg, 0.026 mmol) and triethylamine (10
µL, 0.07 mmol) in DMF (1 mL), and the mixture was heated
to 60 °C for 12 h. The solvent was removed by nitrogen stream,
and the residue was purified by preparative thin-layer chro-
matography (silica gel, chloroform: methanol ) 3:1) to give
di-Boc-protected 3′-guanidino-3′-deoxyadenosine (15, 5 mg,
53%) as a white solid: 1H NMR (CDCl3) δ 1.48 (s, 9H), 1.52
(s, 9H), 3.86(d, 1H, J ) 12.4 Hz), 4.07 (d, 1H, J ) 13.5 Hz),
4.55-4.66 (m, 2H), 5.77 (s, 2H), 6.09 (d, 1H, J ) 1.4 Hz), 8.33
(s, 1H), 8.47 (s, 1H), 9.05 (d, 1H, J ) 6.6 Hz), 11.36 (s, 1H);
high-resolution MS (positive-ion FAB) calcd for C21H33N8O7
[M + H+]+: 509.2472, found 509.2475.
A solution of 15 (3 mg, 0.0059 mmol) in CH2Cl2 /10%
trifluoroacetic acid (1 mL) was stirred overnight at 4 °C. The
solvent was removed under a nitrogen stream, and the residue
was purified by washing with ethyl ether (1 mL × 3) to give
3′-guanidino-3′-deoxyadenosine (10, 2.2 mg, 88%), as a tri-
fluoroacetic acid salt: 1H NMR (CD3OD) δ 3.75 (dd, 1H, J )
2.5, 12.6 Hz), 4.06 (dd, 1H, J ) 2.2, 12.6 Hz), 4.18-4.24 (m,
1H), 4.46-4.53 (m,1H), 4.66 (d, 1H, J ) 5.2 Hz), 6.15 (d, 1H,
J ) 1.4 Hz), 8.38 (s, 1H), 8.64 (s, 1H); high-resolution MS
(positive-ion FAB) calcd for C11H17N8O3 [M + H+]+: 309.1424,
found 309.1427.
Exp er im en ta l Section
9-(3-Azid o-3-d eoxy-â-D-r ib ofu r a n osyl)-N6-(3-iod ob en -
zyl)a d en in e (17). A 725 mg (1.92 mmol) amount of 16 in 35
mL of dry 1,2-dichloroethane was added to 810 mg (2.31 mmol)
of N6-(3-iodobenzyl)adenine silylated residue24 (417 µL, 2.31
mmol). (CH3)3SiOSO2CF3 was added dropwise under N2 and
continuous stirring to give a clear solution after approximately
30 min. The temperature was kept at 83 °C, and after 6 h 50
mL of CH2Cl2 and 100 mL of a 7% NaHCO3 solution were
added to the reaction mixture. The organic phase was ex-
tracted, dried with MgSO4, filtered, and the filtrate evaporated
to dryness. The residue was dissolved in 100 mL of 0.1 N
NaOCH3 in CH3OH, stirred for 2 h, neutralized with a 9:1
H2O-CH3COOH solution, evaporated in vacuo, purified by
column chromatography (CH2Cl2, then 97:3 CH2Cl2-MeOH)
and crystallized from CH3OH to yield 176 mg (18%) of the title
compound as a white solid: 1H NMR (DMSO-d6) δ 3.58 (app
Ma t er ia ls. Compounds 1a (5′-N-ethylcarboxamidoadeno-
sine), 1b (2-chloroadenosine), 2 (N6-(3-iodobenzyl)-2-chloro-
adenosine-5′-N-methyluronamide), 4 (N-[9-chloro-2-(2-furan-
yl)[1,2,4]triazolo[1,5-c]quinazolin-5-amine), 6 (8-[4-[[[[(2-amino-
ethyl)amino]carbonyl]methyl]oxy]phenyl]-1,3-dipropylxanthine),
and 9 (5′-amino-5′-deoxyadenosine) were from Sigma Chemical
Co. (St. Louis, MO). Compound 3 ((1′R,2′R,3′S,4′R,5′S)-4-{2-
chloro-6-[(3-iodophenylmethyl)amino]purin-9-yl}-1-(methyl-
aminocarbonyl)bicyclo[3.1.0]hexane-2,3-diol) was synthesized
as reported.28 Compound 6 (4-methoxy-N-[3-(2-pyridinyl)-1-
isoquinolinyl]benzamide)51 was the gift of Prof. Ad IJ zerman
of the LACDR, Leiden, The Netherlands. CHN analyses were
carried out by Prof. W. Pfleiderer (Konstanz, Germany).
5′-Amino-5′-deoxyadenosine tosylate was obtained from
Sigma (St. Louis, MO). 2′-Amino-2′-deoxyadenosine (7) was
prepared as reported.30 Other amino derivatives, 3′-amino-3′-

4134 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
J acobson et al.
dd, J ) 3.6 Hz and J 5A′,5B′ ) -12.3 Hz, H-5B′), 3.69 (app dd, J
) 3.5 Hz, H-5A′), 4.00 (app q, J ) 3.5 and 7.0 Hz, H-4′), 4.33
(q, J ) 3.8 and 5.2 Hz, H-3′), 4.67 (s, benzylic H), 5.02 (t, J )
5.6 Hz, H-2′), 5.54 (br s, 5′-OH), 5.92 (d, J ) 6.0 Hz, H-1′),
6.24 (br s, 2′-OH), 7.11 (t, J ) 7.8 Hz), 7.36 (d, J ) 7.6 Hz),
7.59 (d, J ) 7.8 Hz), 7.73 (s, aromatic H), 8.23 (s, H-2), 8.42
(s, H-8), 8.55 (br s, H-N6); exact mass (ESI-MS) calcd for
K2HPO4/EDTA buffer (K2HPO4, 150 mM; EDTA, 10 mM), and
either a mixture of 20 µL of the cell lysate and 30 µL of 0.1 M
HCl or 50 µL of cyclic AMP solution (0-16 pmol/200 mL for
standard curve). Bound radioactivity was separated by rapid
filtration through Whatman GF/C filters, which were washed
once with cold buffer. Bound radioactivity was measured by
liquid scintillation spectrometry.
C
17H17I1N8O3 [M+H]+: 509.0548, found 509.0547.
Molecu la r Mod elin g. A molecular model of the human
A3 receptor was built and optimized using the Sybyl 6.653
modeling package, following the homology modeling approach
described in our previous study54 using the recently reported
X-ray structure of bovine rhodopsin41 as a structural template.
All calculations were performed on a Silicon Graphics Octane
R12000 workstation. Briefly, sequences of the A3 transmem-
brane domains, identified in our previously published model,55
were amended by comparison to the corresponding domains
of rhodopsin, according to a published sequence alignment.56
Transmembrane A3 helices were built from these sequences,
in homology to the corresponding helices of rhodopsin, and
minimized individually. The minimized helices were then
grouped by adding one at a time until a helical bundle (TM
region), matching the overall characteristics of the crystal-
lographic structure of rhodopsin, had been obtained. The TM
region was further modified by the addition of residues 148-
173, comprising the second extracellular loop (EL2). The
conformation of EL2 was initially modeled according to the
corresponding domain in rhodopsin including the Cys88-
Cys166 disulfide bond. At each step the structures were
minimized using the Tripos force field with Amber57 all-atom
force parameters until the root mean square value of the
conjugate gradient (CG) was <0.1 kcal/mol/Å. A fixed dielectric
constant ) 4.0 was used throughout these calculations.
Models of adenosine and other A3 receptor ligands used in
this study were constructed using the “Sketch Molecule”
module of Sybyl. The ligands were minimized in Sybyl (using
MOPAC calculated partial atomic charges) and were rigidly
docked into the helical bundle using graphical manipulation
coupled to continuous energy monitoring (Dock module of
Sybyl). Whenever a final position was reached, consistent with
a local energy minimum, the complexes of receptor and ligand
were subjected to an additional CG minimization run of up to
1500 steps.
9-(3-Am in o-3-d eoxy-â-D-r ibofu r a n osyl)-N6-(3-iod oben -
zyl)a d en in e (12). A 100 mg (0.197 mmol) amount of 17 was
dissolved in 10 mL of pyridine, and 100 mg (0.381 mmol) of
Ph3P was added to the solution. After stirring for 1 h at room
temperature, 5 mL of concentrated NH4OH was added. The
reaction mixture was stirred for another 2 h, evaporated to
dryness,andpurifiedbycolumn chromatography(9:1CH2Cl2-MeOH)
to yield 67 mg (70%) of 12 as a white solid: 1H NMR (DMSO-
d6) δ 1.69 (br s, 3′-NH2), 3.47 (app t, J ) 6.1 Hz, H-3′), 3.57
(app dd, J ) 4.3 Hz and -12.6 Hz, H-5B′), 3.73 (app d, J )
7.9 Hz, H-5A′ and 4′), 4.29 (m, H-2′), 4.65 (s, benzylic H), 5.14
(br s, 5′-OH), 5.77 (br s, 2′-OH), 5.93 (d, J ) 2.8 Hz, H-1′),
7.11 (t, J ) 7.8 Hz), 7.35 (d, J ) 7.6 Hz), 7.58 (d, J ) 7.8 Hz),
7.71 (s, aromatic H), 8.21 (s, H-2), 8.44 (s, H-8), 8.48 (br s,
H-N6); exact mass (ESI-MS) calcd for C17H19I1N6O3 [M+H]+:
483.0643, found 483.0631. CHN analysis.
R a d ioliga n d Bin d in g St u d ies. Binding of [3H]R-N6-
phenylisopropyladenosine ([3H]R-PIA) to A1 receptors from rat
cerebral cortex membranes and of [3H]-2-[4-[(2-carboxyethyl)-
phenyl]ethylamino]-5′-N-ethylcarbamoyladenosine ([3H]CGS
21680) to A2A receptors from rat striatal membranes was
performed as previously described.6,35 Adenosine deaminase
(3 units/mL) was present during the preparation of the brain
membranes, in a preincubation of 30 min at 30 °C, and during
the incubation with the radioligands. Binding of [125I]AB-
MECA, 2b (Amersham, Chicago, IL), to membranes prepared
from CHO cells stably expressing the human A3 receptor was
performed as described.26 The assay medium consisted of a
buffer containing 10 mM Mg2+, 50 mM Tris, and 1 mM EDTA,
at pH 8.0. The glass incubation tubes contained 100 µL of
the membrane suspension (0.3 mg of protein/mL, stored at
-80 °C in the same buffer), 50 µL of [125I]AB-MECA (final
concentration 0.3 nM), and 50 µL of a solution of the pro-
posed antagonist. Nonspecific binding was determined in the
presence of 100 µM N6-phenylisopropyladenosine (R-PIA).
All nonradioactive compounds were initially dissolved in
DMSO and diluted with buffer to the final concentration,
where the amount of DMSO never exceeded 2%.
Incubations were terminated by rapid filtration over What-
man GF/B filters, using a Brandell cell harvester (Brandell,
Gaithersburg, MD). The tubes were rinsed three times with 3
mL of buffer each.
Ack n ow led gm en t. This work was supported by
RO1-HL48225 and an Established Investigatorship
Award (to B.T.L.). Z.-G.G. and D.B. thank Gilead
Sciences (Foster City, CA) for support. We thank Dr.
Neli Melman for carrying out binding assays.
At least five different concentrations of competitor, spanning
3 orders of magnitude adjusted appropriately for the IC50 of
each compound, were used. IC50 values, calculated with the
nonlinear regression method implemented in the InPlot pro-
gram (Graph-PAD, San Diego, CA), were converted to apparent
Ki values using the Cheng-Prusoff equation,52 and Ki values
were 1.0 nM ([3H]R-PIA), 14 nM ([3H]CGS 21680), and 0.59
and 1.46 nM ([125I]AB-MECA at human and rat A3 receptors,
respectively.
Cyclic AMP Assa y. COS-7 cells expressing either the
mutant or wold-type receptor were harvested by trypsinization.
After centrifugation and resupension in medium, cells were
plated in 24-well plates in 400 µL of medium. After 24 h, the
medium was removed and cells were washed three times with
500 µL of DMEM, containing 50 mM HEPES, pH 7.4. Cells
were then treated with rolipram (10 µM) and adenosine
deaminase (3 U/mL) and the agonist to be tested. After 45 min,
forskolin (10 µM) was added to the medium and incubation
was continued for an additional 15 min. The reaction was
terminated by removing the supernatant, and cells were lysed
upon the addition of 200 µL of 0.1 M cold HCl. The cell lysate
was resuspended and stored at -20 °C. For determination of
cyclic AMP production,39 protein kinase A (PKA, 50 µg of
protein/well) was incubated with [3H]cyclic AMP (2 nM) in
Refer en ces
(1) Kowaluk, E. A.; J arvis, M. F. Therapeutic potential of adenosine
kinase inhibitors. Expert Opin. Invest. Drugs 2000, 9, 551-564.
(2) Saxena, P. R.; De Vries, P.; Wang, W.; Heiligers, J . P.; Maassen
vandenBrink, A.; Bax, W. A.; Yocca, F. D. Effects of avitriptan,
a new 5-HT 1B/1D receptor agonist, in experimental models
predictive of antimigraine activity and coronary side-effect
potential. Naunyn-Schmiedebergs Arch. Pharmakol. 1997, 355,
295-302.
(3) Maillard, M.; Nikodijevic, O.; LaNoue, K. F.; Berkich, D. A.; J i,
X.-D.; Bartus, R.; J acobson, K. A. Adenosine receptor prodrugs:
Synthesis and biological activity of derivatives of potent, A1-
selective agonists. J . Pharm. Pharmacol. 1994, 83, 46-53.
(4) Liang, B. T.; J acobson, K. A. A physiological role of the adenosine
A3 receptor: Sustained cardioprotection. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 6995-6999.
(5) von Lubitz, D. K. J . E.; Lin, R. C.-S.; Popik, P.; Carter, M. F.;
J acobson, K. A. Adenosine A3 receptor stimulation and cerebral
ischemia. Eur. J . Pharmacol. 1994, 263, 59-67.
(6) J arvis, M. F.; Schutz, R.; Hutchison, A. J .; Do, E.; Sills, M. A.;
Williams, M. [3H]CGS 21680, an A2 selective adenosine receptor
agonist directly labels A2 receptors in rat brain tissue. J .
Pharmacol. Exp. Therap. 1989, 251, 888-893.
(7) Lerman, B. B.; Wesley, R. C.; Belardinelli, L. Electrophysiologic
effects of dipyridamole on atrioventricular nodal conduction and
supraventricular tachycardia. Role of endogenous adenosine.
Circulation 1989, 80, 1536-1543.

Neoceptor from Adenosine A3 Receptor
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24 4135
(8) Ely, S. W.; Berne, R. M. Protective effects of adenosine in
myocardial ischemia. Circulation 1992, 85, 893-904.
(9) Abbracchio, M. P.; Cattabeni, F. Brain adenosine receptors as
targets for therapeutic intervention in neurodegenerative dis-
eases. Ann. N. Y. Acad. Sci. 1999, 890, 79-92.
(10) Murry, C. E.; J ennings, R. B.; Reimer, K. A. Preconditioning with
ischemia: A delay of lethal cell injury in ischemic myocardium.
Circulation 1986, 74, 1124-1136.
(11) Downey, J . M. Ischemic preconditioning. Nature’s own cardio-
protective intervention. Trends Cardiovasc. Med. 1992, 2, 170-
176.
(12) Strickler, J .; J acobson, K. A.; Liang, B. T. Direct preconditioning
of cultured chick ventricular myocytes: Novel functions of
cardiac adenosine A2A and A3 receptors. J . Clin. Invest. 1996,
98, 1773-1779.
(13) J acobson, K. A. A3 adenosine receptors: Novel ligands and
paradoxical effects. Trends Pharmacol. Sci. 1998, 19, 184-191.
(14) Singer, B. F.; Fotheringham, J . A.; Mayne, M. B.; et al. Chronic
activation of A2A and A3 adenosine receptor subtypes inhibits
the antiinflammatory action of A1, A2A, and A3 receptors. FASEB
J . 2001, 15, A566-A566.
(15) Mitchell, C. H.; Peterson-Yantorno, K.; Carre, D. A.; McGlinn,
A. M.; Coca-Prados, M.; Stone, R. A.; Civan, M. M. A3 adenosine
receptors regulate Cl- channels of nonpigmented ciliary epithe-
lial cells. Am. J . Physiol. 1999, 276, C659-666.
(16) Olah, M. E.; Ren, H.; Ostrowski, J .; J acobson, K. A.; Stiles, G.
L. Cloning, expression, and characterization of the unique bovine
A1-adenosine receptor: Studies on the ligand binding site by site
directed mutagenesis. J . Biol. Chem. 1992, 267, 10764-10770.
(17) Townsend-Nicholson, A.; Schofield, P. R. A threonine residue
in the seventh transmembrane domain of the human A1 ad-
enosine receptor mediates specific agonist binding. J . Biol. Chem.
1994, 269, 2373-2376.
(18) Tucker, A. L.; Robeva, A. S.; Taylor, H. E.; Holeton, D.; Bockner,
M.; Lynch, K. R.; Linden, J . A1 adenosine receptors. Two amino
acids are responsible for species differences in ligand recognition.
J . Biol. Chem. 1994, 269, 27900-27906.
(31) van Calenbergh, S.; von Frijtag Drabbe Ku¨nzel, J . K.; Blaton,
N. M.; Peeters, O. M.; Rozenski, J .; Van Aerschot, A.; De Bruyn,
A.; De Keukeleire, D.; IJ zerman, A. P.; Herdewijn, P. N6-
Cyclopentyl-3′-substituted-xylofuranosyladenosines: A New Class
of Non-Xanthine Adenosine A1 Receptor Antagonists. J . Med.
Chem. 1997, 40, 3765-3772.
(32) Feichtinger, K.; Zapf, C.; Sings, H. L.; Goodman, M. Diprotected
Triflylguanidines: A New Class of Guanidinylation Reagents.
J . Org. Chem. 1998, 63, 3804-3805.
(33) Ozols, A. M.; Azhayev, A. V.; Dyatkina, N. B.; Krayevski, A. A.
Aminoglucosides and Their Derivates; IV. A New Synthesis of
1,2,5-tri-O-acyl-3-azido-3-deoxy-â-D-ribofuranose. Synthesis 1980,
7, 557-559.
(34) Vorbru¨ggen, H.; Krolikiewicz, K.; Bennua, B. Nucleoside Syn-
thesis with Trimethylsilyl Triflate and Perchlorate as Catalysts.
Chem. Ber. 1981, 114, 1234-1255.
(35) Schwabe, U.; Trost T. Characterization of adenosine receptors
in rat brain by (-)[3H]N6-phenylisopropyladenosine. Naunyn-
Schmiedebergs Arch. Pharmakol. 1980, 313, 179-187.
(36) Gao, Z.; Li, B. S.; Day, Y. J .; Linden, J . A3 adenosine receptor
activation triggers phosphorylation of protein kinase B and
protects rat basophilic leukemia 2H3 mast cells from apoptosis.
Mol. Pharmacol. 2001, 59, 76-82.
(37) Wiesner, J . B.; Ugarkar, B. G.; Castellino, A. J .; Barankiewicz,
J .; Dumas, D. P.; Gruber, H. E.; Foster, A. C.; Erion, M. D.
Adenosine kinase inhibitors as a novel approach to anticonvul-
sant therapy. J . Pharmacol. Exp. Ther. 1999, 289, 1669-1677.
(38) Gallo-Rodriguez, C.; J i, X.-D.; Melman, N.; Siegman, B. D.;
Sanders, L. H.; Orlina, J .; Fischer, B.; Pu, Q.-L.; Olah. M. E.;
van Galen, P. J . M.; Stiles, G. L.; J acobson, K. A. Structure-
activity relationships of N6-benzyladenosine-5′-uronamides as
A3-selective adenosine agonists. J . Med. Chem. 1994, 37, 636-
646.
(39) van Tilburg, E. W.; von Frijtag Drabbe Ku¨nzel, J .; de Groote,
M.; Vollinga, R. C.; Lorenzen, A., IJ zerman, A. P. N6,5′-Di-
substituted adenosine derivatives as partial agonists for the
human adenosine A3 receptor. J . Med. Chem. 1999, 42, 1393-
1400.
(19) Kim, J .; Wess, J .; van Rhee, A. M.; Scho¨neberg, T.; J acobson, K.
A. Site-directed mutagenesis identifies residues involved in
ligand recognition in the human A2a adenosine receptor. J . Biol.
Chem. 1995, 270, 13987-13997.
(20) Rivkees, S. A.; Barbhaiya, H.; IJ zerman, A. P. Identification of
the adenine binding site of the human A1 adenosine receptor.
J . Biol. Chem. 1999, 274, 3617-3621.
(21) van Rhee, A. M.; J acobson, K. A. Molecular architecture of G
protein-coupled receptors. Drug Dev. Res. 1996, 37, 1-38.
(22) J iang, Q.; van Rhee, A. M.; Kim, J .; Yehle, S.; Wess, J .; J acobson,
K. A. Hydrophilic side chains in the third and seventh trans-
membrane helical domains of human A2a adenosine receptors
are required for ligand recognition. Mol. Pharmacol. 1996, 50,
512-521.
(40) Chen, A.; Gao, Z. G.; Barak, D.; Liang, B. T.; J acobson, K. A.
Constitutive activation of A3 adenosine receptors by site-directed
mutagenesis. Biochem. Biophys. Res. Comm. 2001, 284, 596-
601.
(41) Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Moto-
shima, H.; Fox, B. A.; Trong, I. L.; Teller, D. C.; Okada, T.;
Stenkamp, R. E.; Yamamoto, M.; Miyano, M. Crystal structure
of rhodopsin: A G protein-coupled receptor. Science 2000, 289,
739-745.
(42) Elling, C. E.; Nielsen, S. M.; Schwartz, T. W. Conversion of
antagonist-binding site to metal-ion site in the tachykinin NK-1
receptor. Nature 1995, 374, 74-77.
(43) Holst, B.; Elling, C. E.; Schwartz, T. W. Partial agonism through
a zinc-ion switch constructed between transmembrane domains
III and VII in the tachykinin NK1 receptor. Mol. Pharmacol.
2000, 58, 263-270.
(44) Coward, P.; Wada, H. G.; Falk, M. S.; Chan, S. D.; Meng, F.;
Akil, H.; Conklin, B. R. Controlling signaling with a specifically
designed Gi-coupled receptor. Proc. Natl. Acad. Sci. U.S.A. 1998,
95, 352-357.
(45) Fong, T. M.; Yu, H.; Cascieri, M. A.; Underwood, D.; Swain, C.
J .; Strader, C. D. Interaction of glutamine 165 in the fourth
transmembrane segment of the human neurokinin-1 receptor
with quinuclidine antagonists. J . Biol. Chem. 1994, 269, 14957-
14961.
(46) Mack, C. A.; Patel, S. R.; Schwarz, E. A.; Zanzonico, P.; Hahn,
R. T.; Ilercil, A.; Devereux, R. B.; Goldsmith, S. J .; Christian,
T. F.; Sanborn, T. A.; Kovesdi, I.; Hackett, N.; Isom, O. W.;
Crystal, R. G.; Rosengart, T. K. Biologic bypass with the use
of adenovirus-mediated gene transfer of the complementary
deoxyribonucleic acid for vascular endothelial growth factor 121
improves myocardial perfusion and function in the ischemic
porcine heart. J . Thorac. Cardiovasc. Surg. 1998, 115, 168-177.
(47) Harvey, B. G.; Hackett, N. R.; El-Sawy, T.; Rosengart, T. K.;
Hirschowitz, E. A.; Lieberman, M. D.; Lesser, M. L.; Crystal, R.
G. Variability of human systemic humoral immune responses
to adenovirus gene transfer vectors administered to different
organs. J Virol. 1999, 73, 6729-6742.
(23) IJ zerman, A. P.; von Frijtag Drabbe Ku¨nzel, J . K.; Kim, J .; J iang,
Q.; J acobson, K. A. Site-directed mutagenesis of the human A2a
adenosine receptor. Critical involvement of Glu13 in agonist
recognition. Eur. J . Pharmacol. 1996, 310, 269-272.
(24) Gao, Z.-G, J iang, Q.; J acobson, K. A.; IJ zerman, A. P., Site-
directed mutagenesis studies of human A2A adenosine receptors.
Involvement of Glu13 and His278 in ligand binding and sodium
modulation. Biochem. Pharmacol. 2000, 60, 661-668.
(25) J acobson, K. A.; Siddiqi, S. M.; Olah, M. E.; J i, X.-d.; Melman,
N.; Bellamkonda, K.; Meshulam, Y.; Stiles, G. L.; Kim, H. O.
Structure-activity relationships of 9-alkyladenine and ribose-
modified adenosine derivatives at rat A3 adenosine receptors.
J . Med. Chem. 1995, 38, 1720-1735.
(26) Olah, M. E.; Gallo-Rodriguez, C.; J acobson, K. A.; Stiles, G. L.
[
125I]AB-MECA,
a high affinity radioligand for the rat A3
adenosine receptor. Mol. Pharmacol. 1994, 45, 978-982.
(27) J acobson, K. A.; Kim, H. O.; Siddiqi, S. M.; Olah, M. E.; Stiles,
G. L.; von Lubitz, D. K. J . E. A3 adenosine receptors: design of
selective ligands and therapeutic prospects. Drugs Future 1995,
20, 689-699.
(28) Lee, K.; Ravi, R. G.; J i, X.-d.; Marquez, V. E.; J acobson, K. A.
Ring-constrained (N)methanocarba-nucleosides as adenosine
receptor agonists: Independent 5′-uronamide and 2′-deoxy
modifications. Biorg. Med. Chem. Lett. 2001, 11, 1333-1337.
(29) Azhayev, A. V.; Ozols, A. M.; Bushnev, A. S.; Dyatkina, N.
B.; Kochetkova, S. V.; Victorova, L. S.; Kukhanova, M. K.;
Krayevski, A. A.; Gottikh, P. B. Aminonucleosides and their
derivatives. IV. Synthesis of the 3′-amino-3′-deoxynucleoside
5′-phosphates. Nucleic Acids Res. 1979, 6, 625-643.
(48) Logeart, D.; Hatem, S. N.; Rucker-Martin, C.; Chossat, N.; Nevo,
N.; Haddada, H.; Heimburger, M.; Perricaudet, M.; Mercadier,
J .-J . High efficiency adenovirus-mediated gene transfer to
cardiac myocytes after single-pass coronary delivery. Human
Gene Ther. 2000, 11, 1015-1022.
(49) Svensson, E. C.; Marshall, D. J .; Woodard, K.; Lin, H.; J iang,
F.; Chu, L.; Leiden, J . M. Efficient and stable transduction of
cardiomyocytes after intramyocardial injection or intracoronary
perfusion with recombinant adeno-associated virus vectors.
Circulation 1999, 99, 201-205.
(30) Robins, M. J .; Hawrelak, S. D.; Herna´ndez, A. E.; Wnuk, S. F.
Nucleic acid related compounds. 71. Efficient general synthesis
of purine (amino, azido, and triflate)-sugar nucleosides. Nucleo-
sides Nucleotides 1992, 11, 821-834.

4136 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 24
J acobson et al.
(50) Shah, A. S.; Lilly, R. E.; Kypson, A. P.; Tai, O.; Hata, J . A.;
Pippen, A.; Silvestry, S. C.; Lefkowitz, R. J .; Glower, D. D.; Koch,
W. J . Intracoronary adenovirus-mediated delivery and over-
expression of the â2-adrenergic receptor in the heart: prospects
for molecular ventricular assistance. Circulation 2000, 101, 408-
414.
(51) van Muijlwijk-Koezen, J . E.; Timmerman, H.; van der Goot, H.;
Mengge, W. M. P. B.; Frijtag von Drabbe-Ku¨nzel, J .; de Groote,
M.; IJ zerman, A. P. Isoquinoline and quinazoline urea analogues
as antagonists for the human A3 adenosine receptor. J . Med.
Chem. 2000, 43, 2227-2238.
(52) Cheng, Y. C.; Prusoff, W. H. Relationship between the inhibition
constant (Ki) and the concentration of inhibition which causes
50% inhibition (IC50) of an enzyme reaction. Biochem. Pharma-
col. 1973, 22, 3099-3108.
(53) The program SYBYL 6.3 is available from TRIPOS Associates,
St. Louis, MO, 1993.
(54) Kim, H. S.; Barak, D.; Harden, T. K.; Boyer, J . L.; J acobson, K.
A. Acyclic and cyclopropyl anlaogues of adenosine bisphosphate
antagonists of the P2Y1 receptor: Structure-activity relation-
ships and receptor docking. J . Med. Chem. 2001, 44, 3092-3108.
(55) Moro, S.; Li, A. H.; J acobson, K. A. Molecular modeling studies
of human A3 adenosine receptor antagonists: Structural homol-
ogy and receptor docking. J . Chem. Inf. Comput. Sci. 1998, 38,
1239-1248.
(56) Horn, F.; Weare, J .; Beukers, M. W.; Ho¨rsch, S.; Bairoch, A.;
Chen, W.; Edvardsen, Ø.; Campagne, F.; Vriend, G. GPCRDB:
An information system for G protein-coupled receptors. Nucleic
Acids Res. 1998, 26, 277-281.
(57) Weiner, S. J .; Kollman, P. A.; Nguyen, D. T.; Case, D. A. An
all-atom force field for simulation of protein and nucleic acids.
J . Comput. Chem. 1986, 7, 230-252.
J M010232O