Synthesis, biological evaluation, and molecular modeling of ribose-modified adenosine analogues as adenosine receptor agonists

Article abstract of DOI:10.1021/jm049408n

A number of 3′-C-methyl analogues of selective adenosine receptor agonists such as CPA, CHA, CCPA, 2′-Me-CCPA, NECA, and IB-MECA was synthesized to further investigate the subdomain of the receptor that binds the ribose moiety of the ligands. Affinity dat

Full text of DOI:10.1021/jm049408n

1550
J. Med. Chem. 2005, 48, 1550-1562
Synthesis, Biological Evaluation, and Molecular Modeling of Ribose-Modified
Adenosine Analogues as Adenosine Receptor Agonists
Loredana Cappellacci,† Palmarisa Franchetti,† Michela Pasqualini,† Riccardo Petrelli,† Patrizia Vita,†
Antonio Lavecchia, Ettore Novellino, Barbara Costa,‡ Claudia Martini,‡ Karl-Norbert Klotz,§ and
Mario Grifantini*,†
Dipartimento di Scienze Chimiche, Universita` di Camerino, 62032 Camerino, Italy, Dipartimento di Chimica
Farmaceutica e Tossicologica, Universita` di Napoli “Federico II”, 80131 Napoli, Italy, Dipartimento di Neurobiologia,
Farmacologia e Biotecnologia, Universita` di Pisa, 56126 Pisa, Italy, and Institut fu¨r Pharmakologie und Toxikologie,
Universita¨t Wu¨rzburg, D-97078 Wu¨rzburg, Germany
Received July 24, 2004
A number of 3′-C-methyl analogues of selective adenosine receptor agonists such as CPA, CHA,
CCPA, 2′-Me-CCPA, NECA, and IB-MECA was synthesized to further investigate the
subdomain of the receptor that binds the ribose moiety of the ligands. Affinity data at A1, A2A,
and A3 receptors in bovine brain membranes showed that the 3′-C-modification in adenosine
resulted in a decrease of the affinity at all three receptor subtypes. When this modification
was combined with N6-substitution with groups that induce high potency and selectivity at A1
receptor, the affinity and selectivity were increased. However, all 3′-C-methyl derivatives proved
to be very less active than the corresponding 2′-C-methyl analogues. The most active compound
was found to be 3′-Me-CPA which displayed a Ki value of 0.35 µM at A1 receptor and a selectivity
for A1 vs A2A and A3 receptors higher than 28-fold. 2′-Me-CCPA was confirmed to be the most
selective, high affinity agonist so far known also at human A1 receptor with a Ki value of 3.3
nM and 2903- and 341-fold selective vs human A2A and A3 receptors, respectively. In functional
assay, 3′-Me-CPA, 3′-Me-CCPA, and 2-Cl-3′-Me-IB-MECA inhibited forskolin-stimulated
adenylyl cyclase activity with IC50 values ranging from 0.3 to 4.9 µM, acting as full agonists.
A rhodopsin-based model of the bovine A1AR was built to rationalize the higher affinity and
selectivity of 2′-C-methyl derivatives of N6-substituted-adenosine compared to that of 3′-C-
methyl analogues. In the docking exploration, it was found that 2′-Me-CCPA was able to form
a number of interactions with several polar residues in the transmembrane helices TM-3, TM-
6, and TM-7 of bA1AR which were not preserved in the molecular dynamics simulation of 3′-
Me-CCPA/bA1AR complex.
Introduction
treated with agonists or antagonists of adenosine recep-
tors there are Parkinson’s disease, hypoxia/ischemia,
epilepsy, kidney disease, asthma and cancer.2 A bene-
ficial therapeutic effect of adenosine A1 receptor agonists
for type II diabetic patients was also suggested.3 Most
of the agonists of adenosine receptors so far known are
adenosine derivatives;1 only recently have A2B non-
adenosine agonists been reported.4 Structure-activity
relationship studies have pointed out that the ribose
recognition domain of adenosine and adenosine ana-
logues contributes strongly to the affinity for adenosine
receptors. Among ribose modifications, replacement of
hydrogen atoms of the ribose ring with a methyl group
in adenosine and adenosine analogues afforded com-
pounds with various affinity and selectivity.5,6 We have
reported that the introduction of a methyl group at the
C-2′ position in adenosine resulted in a decrease of
affinity, particularly at A2A and A3 ARs.5 However, when
this modification was combined with N6-substitution
with groups that induce high potency and selectivity at
the A1 receptor, the selectivity for A1 vs A2A and A3 ARs
was increased. So, 2′-Me-CCPA,5 the 2′-C-methyl ana-
logue of 2-chloro-N6-cyclopentyladenosine (CCPA),7
proved to be a potent and highly selective agonist at
bovine A1AR with A3/A1 selectivity higher than that of
CCPA (2′-Me-CCPA A3/A1 selectivity ) 2856, CCPA A3/
Adenosine modulates a great variety of physiological
functions mediated by different subtypes of G protein-
coupled receptors. Many efforts have been made to
develop therapeutic agents based on selective interac-
tions with one of the four adenosine receptor subtypes
A1, A2A, A2B, and A3 (ARs).1 Activation of these receptors
inhibits or stimulates the intracellular enzyme aden-
ylate cyclase (AC). A1 and A3 receptors are coupled to
the inhibitory G-protein Gi/Go. Recent studies have
pointed out that A1 receptors are also coupled in atrial
cardiac myocytes to other effectors such as calcium or
potassium ion channels and phospholipase C (PLC),
while A3 receptors are also coupled to both PLC and
phospholipase D (PLD). A2A and A2B receptors are
coupled to the stimulatory Gs protein, thus stimulating
AC activity. A2B is also coupled to PLC. In recent years
it has become more and more clear that adenosine
receptors may be targets for the development of new
drugs. Among the pathological conditions that might be
* To whom correspondence should be addressed. Phone: +39-0737-
402233. Fax +39-0737-637345. E-mail: mario.grifantini@unicam.it.
† Universita` di Camerino.
Universita` di Napoli “Federico II”.
‡ Universita` di Pisa.
§ Universita¨t Wu¨rzburg.
10.1021/jm049408n CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/05/2005

Ribose-Modified Adenosine Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5 1551
Chart 1. Chemical Structures of Reported 2′- and 3′-C-Methyl-adenosine Derivatives
Scheme 1a
A1 selectivity ) 87). On the other hand, we found that
the 1′-C-methyl modification in N6-substituted adeno-
sine analogues decreased the affinity, particularly at A1
and A2A ARs.6
The potential clinical applications of selective A1/A3
agonists as cardioprotective and neuroprotective agents
prompted us to further investigate the receptor sub-
domain that binds the ribose moiety by the study of 3′-
C-methyl derivatives of N6-substituted adenosine and
2-chloroadenosine. The present paper reports on the
synthesis and binding studies of these compounds as
well as of novel di- and trisubstituted analogues of 2′-
C-methyladenosine. Furthermore, 3′-C-methyl deriva-
tives of the A3AR selective agonist IB-MECA (N6-(3-
iodobenzyl)-adenosine-5′-N-methyluronamide), and of
its even more selective 2-chloro (2-Cl-IB-MECA) and
2-methylamino derivatives,8 were synthesized to inves-
tigate the role of the conformation of these nucleosides
in binding to adenosine receptors. The 3′-C-methyl
analogue of NECA (adenosine-5′-N-ethyluronamide)
was also synthesized (Chart 1).
Chemistry
3′-C-Methyl adenosine N6-substituted and N6/C-2
disubstituted derivatives and novel 2′-C-methyl ana-
logues were synthesized as reported in Scheme 1.
Nucleophilic displacement of the 6-chlorine atom in the
protected compounds 10, 115 and 12, 139 with cyclo-
pentylamine, cyclohexylamine, benzylamine, or 3-iodo-
benzylamine in anhydrous ethanol gave, after depro-
tection with methanolic ammonia, the corresponding N6-
substituted nucleosides 1-9. The assignment of the
â-anomeric structure of these compounds was performed
by proton NOE data.
3′-C-Methyl derivatives of NECA (14) and IB-MECA
(15) were also obtained (Scheme 2). 2′,3′-Isopropylidene
derivatives of compounds 16 and 21 (17 and 22, respec-
tively) were oxidized by the TEMPO (2,2,6,6-tetra-
methylpiperidin-1-yloxy) BAIB (diacetoxyiodo-benzene)
system10 in a 1:1 acetonitrile-water solvent to give the
corresponding 5′-carboxylic acids 18 and 23. We pre-
ferred to use the TEMPO-BAIB system because this
is a mild and efficient method that allows the synthesis
of 5′-carboxylic acid nucleosides in high yield with a
short reaction time. Compounds 18 and 23 were con-
verted into ethyl esters 19 and 24, respectively, by
a Reagents and conditions: (i) cyclopentylamine, cyclohexyl-
amine, or benzylamine, EtOH, ∆, or 3-iodobenzylamine hydro-
chloride, TEA, EtOH, ∆; (ii) NH3/MeOH, rt.
reaction with SOCl2 and anhydrous ethanol at room
temperature. Displacement of the 6-chlorine atom of 19
with 3-iodobenzylamine and treatment of intermediate
20 with methylamine followed by deprotection with 90%
formic acid gave 3′-Me-IB-MECA (15). In the same way,
compound 24 was converted into 3′-Me-NECA (14) by
reaction with ethylamine.
The synthesis of 2-Cl-2′-Me-IB-MECA (25), 2-Cl-3′-
Me-IB-MECA (26), 2-methylamino-2′-Me-IB-MECA (27),
and 2-methylamino-3′-Me-IB-MECA (28) was carried
out as described in Scheme 3. Compounds 3 and 7 were
protected as 2′,3′-O-isopropylidene derivatives (29 and
32, respectively) by reaction with 2,2-dimethoxypropane,
camphorsulfonic acid in acetone. Oxidation of 29 and
32 with the TEMPO-BAIB system gave the corre-
sponding 5′-carboxylic acids 30 and 33, respectively. The
ethyl esters 31 and 34, obtained from 30 and 33 as
described for 19 and 24, were converted into nucleosides
25 and 26, respectively, by treatment with methylamine

1552 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5
Cappellacci et al.
Scheme 2a
a Reagents and conditions: (i) 2,2-dimethoxypropane, camphorsulfonic acid, acetone, ∆; (ii) TEMPO-BAIB, MeCN/H2O (1:1), rt; (iii)
SOCl2, EtOH, rt; (iv) 3-iodobenzylamine hydrochloride, TEA, EtOH; (v) CH3NH2, -20 °C to rt; (vi) 90% HCOOH, ∆; (vii) C2H5NH2.
Scheme 3a
a Reagents and conditions: (i) 2,2-dimethoxypropane, camphorsulfonic acid, acetone, ∆; (ii) TEMPO-BAIB, MeCN/H2O (1:1), rt; (iii)
SOCl2, EtOH, rt; (iv) CH3NH2, -20 °C to rt; (v) 90% HCOOH, ∆; (vi) 2 N CH3NH2/THF, ∆.
and then deisopropylidenation. Displacement of 2-chlo-
rine atom of compounds 25 and 26 with methylamine
gave 27 and 28, respectively.
ments. When anomeric proton of compounds 4-9, 14, 15,
26, and 28 was irradiated, an enhancement of the H-8
signal was observed, indicating that a population of syn
conformers is present in these compounds. However,
because the H-8 enhancement in 14, 15, and 26 (about
6.3%) is lower than that observed in NECA (22%), the
conformation of 3′-Me-NECA (14), 3′-Me-IB-MECA (15),
and 2-Cl-3′-Me-IB-MECA (26) should diverge slightly
from those of the nonmethylated analogues. On the
basis of the correlation reported by Rosemeyer et al.,12
2′-Me-Ado, 3′-Me-Ado, 2-Cl-2′-Me-Ado, 2′-Me-CPA, 2′-
Me-CCPA, 2′-Me-IB-Ado, 2′-Me-NECA, and 2′-Me-IB-
MECA were prepared according to known procedures
as reference compounds for biological evaluation.5,11
Information concerning the predominant solution
conformation of synthesized 3′-C-methyl adenosine de-
1
rivatives was obtained via H NMR and NOE experi-

Ribose-Modified Adenosine Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5 1553
Table 1. Affinity of 2′-, and 3′-C-Methyl-adenosine Derivatives in Radioligand Assays at Bovine Cortical Membranes (A1 and A3) and
Bovine Brain Striatum (A2a) Receptorsa-c
Ki (nM) or % displacement at 10-5
M
a
b
c
compd no.
compd label
A1
A2A
28%
40%
A3
A2A/A1
A3/A1
A2A/A3
1
2
2′-Me-CHA
2′-Me-B-Ado
2-CI-2′-Me-IB-Ado
3′-Me-CPA
3′-Me-CHA
3′-Me-B-Ado
3′-Me-IB-Ado
3′-Me-CCPA
2-CI-3′-Me-IB-Ado
3′-Me-NECA
3′-Me-IB-MECA
2-CI-2′-Me-IB-MECA
2-CI-3′-Me-IB-MECA
2-MeNH-2′-Me-IB-MECA
2-MeNH-3′-Me-IB-MECA
2′-Me-Ado
3′-Me-Ado
2-CI-2′-Me-Ado
2′-Me-CPA
2′-Me-CCPA
IB-Ado
2′-Me-IB-Ado
NECA
2′-Me-NECA
IB-MECA
2-Cl-IB-MECA
2′-Me-IB-MECA
30 ( 3.1
460 ( 51
85 ( 7.8
350 ( 38
22%
2560 ( 238
670 ( 65
163 ( 17
30%
>333
>21.7
5.5
85.3
>3.9
>14.0
2.68
1.4
3
470 ( 44
2%
1.9
4
>28.6
>28.6
5
0%
0%
6
23%
30%
21%
15%
7
0%
10%
8
5370 ( 611
4300 (407
6%
5620 ( 570
325 ( 37
4530 (452
8%
1310 ( 128
1120 ( 80
13%
350 ( 40
10 ( 1.5
3 ( 0.6
23:t 2
17%
20%
>1.9
0.74
9
3200 ( 345
0%
2700 ( 282
28%
3210 ( 306
480 ( 44
5780( 576
12%
0.62
1.18
5.79
14
15
25
26
27
28
5%
>1.78
8.55
>2.21
0.57
1.48
1.27
2780 ( 264
15%
0%
4000 ( 411
5800 ( 950
4%
4890 ( 850
7560 ( 735
4560 ( 850
121 + 10
680 ( 58
16 ( 3
220 ( 23
40 ( 4
401 ( 25
1260 ( 131
1200 ( 123
0%
3.05
5.2
0.92
>8.9
3.33
>0.58
0%
7%
13.9
756
1520
5.3
13.6
1.14
0.28
3.33
0.45
3.6
>28.5
22
2856
0.66
1.3
5.2
0.54
073
0.00025
0.68
<0.49
34.3
0.53
7.96
10.5
0.22
0.52
4.54
1822.7
5.25
220 ( 23
8570 ( 848
15.2 ( 2
65 ( 61
73 ( 5
420 ( 38
8.8 ( 0.9
0.22 ( 0.02
240 ( 25
50 ( 4.1
14 ( 4
770 ( 65
12 ( 2
890 ( 61
350 ( 38
a
Displacement of specific [3H]DPCPX binding in bovine cortical membranes expressed as Ki ( SEM in nM (n ) 3). b Displacement of
specific [3H]CGS21680 binding in bovine striatal membranes expressed as Ki ( SEM in nM (n ) 3). c Competition assay of [125I]-N6-(3-
iodo-4-aminobenzyl)-5′-N-methylcarboxamido-adenosine ([125I]AB-MECA) to bovine cortical membranes in the presence of the A1AR selective
antagonist DPCPX (20 nM); the residual binding which was not displaced by DPCPX represented binding to A3AR.
the observed H-1′ enhancement translates into a 60%
population of the syn conformer. Information concerning
the solution conformation of the ribosyl moiety of these
nucleosides was obtained by the coupling constant
values. The 3′-C-substitution precludes getting informa-
tion from J2′3′ and J3′4′ values, leaving only J1′2′ as a clue
to sugar puckering. It was found that the J1′2′ value of
compounds 4-9, 14, 15, 26, and 28 is in the range of
7.7-8.2 Hz, indicating that they are predominantly
S-puckered.
2′-methyl modification confirmed A1 selectivity vs A2A
and A3 receptors. We ascribed the different affinity of
2′-Me-Ado and 3′-Me-Ado to the different conformation
of the furanose ring of these nucleosides in solution
(North (3T2)-anti and South (2T3)-syn, respectively).5 The
3′-C-methyl-substituted adenosine analogues 4-9 showed
a marked preference for the South (2T3)-syn conforma-
1
tion as determined by H NMR data.
N6-Substitution with a cyclopentyl group in 3′-Me-
Ado increased the affinity at A1AR (Ki ) 0.35 µM),
restoring A1 selectivity. 3′-Me-CPA proved to be the
most active and selective compound at A1AR among the
3′-C-methyl-N6-substituted adenosine analogues re-
ported in this study. Introduction of a chlorine atom in
the 2-position of 3′-Me-CPA (3′-Me-CCPA, 8) induced a
15-fold decrease of the affinity at A1AR. This is a
surprising result because a similar substitution in 2′-
Me-CPA (2′-Me-CCPA) brings about an increase of both
affinity and selectivity at this receptor subtype. 2′-Me-
CCPA was confirmed to be a potent agonist and the
most selective one at bovine A1AR known so far. This
compound proved to be a potent A1 agonist also versus
human adenosine receptor (Table 2), with a 2903-fold
and 341-fold selectivity vs A2A and A3 ARs, respectively.
Biological Evaluation
Ribose-modified nucleosides 1-9 and NECA and IB-
MECA analogues (14, 15, and 25-28) were evaluated
in radioligand binding assays to determine their affinity
at A1, A2A, and A3 ARs (Table 1). In particular, affinities
for A1 and A2A receptors were determined in competition
assays in bovine cortical membranes (A1) and bovine
brain striatum (A2A) using, respectively, [3H]DPCPX
(1,3-dipropyl-8-cyclopentylxanthine) and [3H]CGS21680
(2-[4-(2-carboxyethyl)phenyl]ethyl-amino-5′-N-ethyl-
carboxamido-adenosine) as radioligands.13,14 Affinity for
A3AR was determined in competition assays of [125I]AB-
MECA ([125I]-N6-(3-iodo-4-aminobenzyl)-5′-N-methylcar-
boxamido-adenosine) to bovine cortical membranes in
the presence of the A1AR selective antagonist DPCPX
(20 nM);14 the residual binding which was not displaced
by DPCPX represented binding to A3AR. 2′-Me-Ado, 3′-
Me-Ado, 2-Cl-2′-Me-Ado, 2′-Me-CPA, 2′-Me-CCPA, IB-
Ado, 2′-Me-IB-Ado, NECA, 2′-Me-NECA, IB-MECA,
2-Cl-IB-MECA, and 2′-Me-IB-MECA were used as refer-
ence compounds. Introduction of a methyl group in the
adenosine 3′-position proved to be not tolerated, while
N6-Substitution of 3′-Me-Ado with a benzyl or 3-iodo-
benzyl group did not increase the affinity at bovine ARs;
however, the introduction of a chlorine in the 2-position
of 3′-Me-IB-Ado to give 2-Cl-3′-Me-IB-Ado (9) conferred
a better affinity at A1/ A2A/A3 receptors.
With regard to the novel 2′-C-methyl-N6-substituted
adenosine analogues (compounds 1-3), the introduction
of a cyclohexyl group (2′-Me-CHA, 1) resulted in a good
affinity at A1AR with a selectivity vs A2A and A3 ARs of

1554 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5
Cappellacci et al.
Table 2. Binding Affinity of 2′-Me-CCPA and CCPA at Human
A1, A2A, A2B, A3 Adenosine Receptor Subtypes Expressed in
CHO Cells
Table 3. Inhibition of Adenylyl Cyclase Activity in Rat
Cortical Membranes by Compounds 3, 4, 8, 25, and 26a
compd
IC50 (nM)
% maximal inhibition
Ki (nM)
CHA
3
1.41 ( 0.7
102 ( 9
16.2 ( 2.2
22.0 ( 1.3
18.1 ( 1.0
21.8 ( 1.3
19.3 ( 1.0
17.8 ( 1.0
a
compd
A1
A2A
A2B
A3
A
2A/A1 A3/A1
A
2A/A3
8.3
4
304 ( 20
2′-Me-CCPA 3.3 9580 37600 1150 2903
CCPA 0.8 2300 18800 42 2875
341
53
8
25
26
4010 ( 360
279 ( 24
4908 ( 470
55
a EC50 values (nM) are reported for the agonist-mediated
stimulation of adenylyl cyclase activity in a membrane prepara-
tion.
a IC50 values and the maximal inhibitory effects were obtained
from nonlinear curve fitting of data using GraphPad computer
program. All values are the mean ( SEM of three independent
experiments
>333- and 85.3-fold, respectively. N6-Benzyl substitu-
tion in 2′-Me-Ado (compound 2) resulted in moderate
affinity at A1 and A3 ARs but low selectivity, while the
introduction of a chlorine atom at the 2-position of 2′-
Me-IB-Ado (compound 3) reduced both A1 and A3
affinity.
Molecular Modeling. The above-discussed SARs do
not allow us to establish whether both 2′-C-methyl and
3′-C-methyl derivatives of N6-substituted adenosine
interact with the bA1AR recognition site with a similar
or different binding mode. So, a computational study
was performed to elucidate the hypothetical binding
mode of 2′-Me-CCPA and 3′-Me-CCPA, which show a
different activity profile, and to interpret our experi-
mental results.
The bovine A1AR (bA1AR) model used for this purpose
included the seven transmembrane helical domains
(TMs) and was built in homology with the recently
published X-ray structure of bovine rhodopsin.15 De-
tails of the model building are given in the Experi-
mental Section. To determine the most favorable bind-
ing locations and orientations for the above-mentioned
compounds, we used the automated docking program
AutoDock.16 The 50 independent docking runs carried
out for each ligand generally converged to a small
number of different positions (“clusters” of results
differing by less than 1.5 Å rmsd). Generally, the top
clusters (i.e. those with the most favorable ∆Gbind) were
also associated with the highest frequency of occurrence,
which suggests a good convergence behavior of the
search algorithm. The best results in terms of free
energy of binding were all located in a similar position
at the active site. The most important interactions found
for each compound are summarized in Table 4.
For each of the ribose-modified ligands (2′-Me-CCPA
and 3′-Me-CCPA), very clear preference for a single
position in the binding site could be obtained. Interest-
ingly, both results were located in the cavity between
TM3, TM5, TM6, and TM7 and involved essentially the
same residues, but playing different roles in agonist
binding in each case. For 2′-Me-CCPA, the result with
top binding energy (-12.0 kcal/mol) was found 19 times
out of 50. The N6 amino group of the ligand formed a
hydrogen bond with the CO oxygen of Asn254. The
ribose moiety was coordinated to several hydrophilic
residues in TM3 and TM7. In particular, the 2′-OH, 3′-
OH, and 5′-OH groups of the ribose ring were involved
in hydrogen bonding with the Nδ imidazole nitrogen of
His278 (TM7) and the OH oxygen of Thr91 (TM3),
respectively. Moreover, the ribose O-4′ oxygen was
engaged in a hydrogen bond with the OH hydrogen of
Thr91. A very similar binding position was also observed
for compound 3′-Me-CCPA (∆G of -11.9 kcal/mol, found
25 times out of 50). However, in this case, no hydrogen
bonds involving the ribose moiety were observed, and
only a single interaction was achieved, between the 5′-
OH and the Ser277 side chain. Moreover, it is worth
noting that the N6 amino group of the ligand was
In the case of the adenosine 5′-uronamide analogues
14, 15, 25-28, introduction of a methyl group in the 2′-
or 3′-position brought about a decrease of affinity at all
three receptor subtypes as compared to NECA and IB-
MECA. 2′-Me-NECA (Ki values in the range of 0.22-
0.77 µM) resulted 55-fold less active than NECA at A1,
13.7-fold at A2A, and 5.7-fold at A3AR. 2′-Me-IB-MECA
proved to be less potent than IB-MECA with Ki values
ranging from 0.24 to 1.2 µM, and a moderate selectivity
for A3AR. On the contrary, the 2′-C-methyl modification
in 2-Cl-IB-MECA (compound 25) resulted in increased
affinity at A1 receptor (2.7-fold) but reduced affinity at
A2A and A3 ARs (6.9- and 2180-fold, respectively).
A methyl group in the 3′-position of NECA (compound
14) is not tolerated at all three receptor subtypes. As
found in the case of rat and human adenosine receptors,
N6-substitution of MECA with a 3-iodobenzyl group (IB-
MECA) increased the affinity, in particular at A3 and
A1 receptors. However, when this modification was
combined with the introduction of a methyl group in the
3′-position of the ribose moiety (compound 15), both
affinity and selectivity were decreased. Finally, the
2-substitution in 3′-Me-IB-MECA with a chlorine atom
(compound 26) did not improve the affinity at A3AR,
while an analogous substitution in IB-MECA increased
the affinity of 40-fold at this receptor subtype. Introduc-
tion in 2-position of a methylamino group gave rise to a
different result. In fact, the 2′-C-methyl derivative 27
was devoid of affinity, while 3′-C-methyl isomer 28
showed low affinity at all three receptor subtypes.
Compounds 3, 4, 8, 25, and 26 were also tested in a
functional assay at A1 receptors in rat cortical mem-
branes for their ability to inhibit forskolin-stimulated
adenylyl cyclase. The efficacy of these compounds was
compared with that obtained for CHA, a selective A1
adenosine receptor agonist (Table 3, and Figure 1 of the
Supporting Information). CHA proved to be the most
potent inhibitor (IC50 1.41 nM) with a maximal inhibi-
tory effect of 16.2% similar to that of the other com-
pounds (values ranging from 17.8 to 22%). Compounds
3 (2-Cl-2′-Me-IB-Ado), 4 (3′-Me-CPA), and 25 (2-Cl-2′-
Me-IB-MECA) showed the highest IC50 values (102, 304,
and 279 nM, respectively) among the novel compounds.
Finally, although compounds 8 (3′-Me-CCPA) and 26 (2-
Cl-3′-Me-IB-MECA) showed lower IC50 values (4010 and
4908 nM, respectively), they achieved the maximal
efficacy. Thus, in the functional assay all compounds
tested behave as full agonists for A1AR.

Ribose-Modified Adenosine Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5 1555
Table 4. Result of 50 Independent Docking Runs for Each Liganda
ligand
Ntot
focc
∆Gbind
-12.3
surrounding residues
2′-Me-CCPA
10
19
Ala84 (TM3), Val87 (TM3), Leu88 (TM3), Ilel89 (TM3), Thr91 (TM3),
Gln92 (TM3), Met180 (TM5), Val181 (TM5), Asn184 (TM5), Phe243 (TM6),
Ser246 (TM6), Trp247 (TM6), Leu250 (TM6), His251 (TM6), Asn254 (TM6),
Thr257 (TM6), Ile270 (TM7), Ala273 (TM7), Ile274 (TM7), Phe275 (TM7),
Ser277 (TM7), His278 (TM7), Asn280 (TM7)
Ala84 (TM3), Val87 (TM3), Leu88 (TM3), Leu90 (TM3), Thr91 (TM3),
Gln92 (TM3), Met180 (TM5), Val181 (TM5), Asn184 (TM5), Val189 (TM5),
Ser246 (TM6), Trp247 (TM6), Leu250 (TM6), His251 (TM6), Asn254 (TM6),
Ile270 (TM7), Ala273 (TM7), Ile274 (TM7), Phe275 (TM7), Ser277 (TM7),
His278 (TM7), Asn280 (TM7)
3′-Me-CCPA
8
25
-11.9
a
Ntot is the total number of clusters; the number of results in the top cluster is given by the frequency of occurrence, focc; ∆Gbind is the
estimated free energy of binding for the top cluster results and is given in kcal/mol. The last column shows the contacting residues for the
binding mode of the top cluster. Only residues with at least five van der Waals contacts to the ligand are shown. Residues that form
hydrogen bonds with the ligand are highlighted in bold.
our ligands. Moreover, from the MD trajectories (Figure
2b-e of the Supporting Information, on the left), it can
be deduced that the ribose moiety of 2′-Me-CCPA/bA1AR
complex maintained a rather stable network of polar
interactions during the MD simulation. After 50 ps of
equilibration, in which the ligand merely underwent
small amplitude fluctuations in the binding site, it
rapidly achieved stable interactions with the receptor
key residues throughout the rest of the trajectory. In
particular, the 2′-OH, 3′-OH, and 5′-OH groups of the
ribose ring were hydrogen-bonded to the Nδ imidazole
nitrogen of His278 and the OH oxygen of Thr91,
respectively, whereas the O-4′ oxygen interacted with
the side chain oxygen of Thr91.
Surprisingly, 3′-Me-CCPA produced an instable com-
plex with bA1AR. In fact, with the exception of the
hydrogen bond formed between the N6 amino group of
the ligand and the CO oxygen of Asn254 side chain,
which remained stable along the 150 ps MD trajectory
(Figure 2a of the Supporting Information, on the right),
the remaining polar interactions resulted not strong
enough to be preserved throughout the MD simulation,
Figure 1. Side view of the 2′-Me-CCPA/bA1AR complex
giving average distances longer than that of an ideal
hydrogen bond (Figure 2b-e of the Supporting Informa-
tion, on the right).
model. The side chains of the important residues in proximity
(5 Å) to the docked 2′-Me-CCPA molecule are highlighted and
labeled.
Figure 2 depicts a poor superimposition between the
2′-Me-CCPA and 3′-Me-CCPA inside the receptor bind-
ing domain. In particular, the adenine and the ribose
moieties of the 3′-Me-CCPA derivative are shifted out
of position with respect to those presented by the 2′-
Me-CCPA analogue, decreasing the stability of the 3′-
Me-CCPA/bA1AR complex. A possible reason for these
differences in the binding modes could be given by the
fact that the ribose ring of 3′-Me-CCPA adopts a South
(2T3)-syn conformation, in which the methyl group at
position 3′ points toward the His278, thus preventing
the hydrogen bonding formation of the OH groups at 2′
and 3′ positions with the Nδ imidazole nitrogen of
His278. This fact might explain the different biological
activity observed for the two examined compounds.
oriented in a such a way that no hydrogen bond
formation with the Asn254 side chain was possible.
To assess the dynamic stability of the 2′-Me-CCPA/
bA1AR and 3′-Me-CCPA/bA1AR complexes and to ana-
lyze the potential ligand-receptor interactions, a mo-
lecular dynamics (MD) simulation of 150 ps at a
constant temperature of 300 K was run. The distances
between the ligands and the key receptor residues
(Figure 2 of the Supporting Information) were moni-
tored along the complete 150 ps MD trajectory. Exami-
nation of the average structure of the 2′-Me-CCPA/
bA1AR complex showed that the exocyclic amino group
at the 6-position was located within hydrogen bonding
distance from the CO oxygen of the Asn254 (TM6) side
chain (Figure 1).
The trajectory plot of the analyzed complex (Figure
2a of the Supporting Information, on the left) shows that
this hydrogen bond was quite stable throughout the MD
simulation. Asn254, conserved among all adenosine
receptor subtypes, was found to be important for ligand
binding. In fact, the inability of the N250A mutant
A3AR17 or the corresponding mutant A2AAR18 to bind
either radiolabeled agonist or antagonist was consistent
with a proposed direct interaction of this residue with
Our present model is consistent with several experi-
mental results concerning recognition of the ribose or
ribose-like moiety common to various adenosine ago-
nists: (i) the hydrophilic interaction at His278 was
required for high-affinity binding of both A1 agonists
and antagonists.19 In the A2AAR, this site was mutated
to Ala with the loss of high-affinity binding of both
agonists but not antagonists.20 At the A3AR, His at this
site was proposed as the basis for enhanced affinity of

1556 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5
Cappellacci et al.
receptor.17 In an UV absorption study,25 it has been
suggested that Trp265 in bovine rhodopsin (correspond-
ing to Trp247 in bA1AR) tilts toward the membrane
plane during conversion of the inactive to the active
state (MI f MII) of the receptor.
Conclusions
A series of 3′-C-methyl derivatives of N6-substituted
adenosine and 2-chloroadenosine, as well as of N6-
substituted adenosine-5′-N-uronamides were studied to
determine their affinity toward bovine A1, A2A, and A3
adenosine receptors. We found that the 3′-C-methyl
modification in adenosine is not tolerated at all evalu-
ated ARs subtypes. When this modification was com-
bined with N6-substitution with groups that induce high
potency and selectivity at A1AR such as the cyclopentyl
one, the activity was in part restored and the selectivity
increased. The 3′-C-methyl modification in adenosine-
5′-N-uronamides was also not tolerated at all AR
subtypes, even when this modification was combined
with N6-substitution with groups which induce high
potency and selectivity at A3AR, such as the 3-iodo-
benzyl one (3′-Me-IB-MECA).
In general it may be concluded that the introduction
of a methyl group at C-3′ position in adenosine and
adenosine derivatives is less tolerated than that in C-1′,
or C-2′ or C-4′ positions of the ribose ring. This behavior
might be explained by the marked preference of 3′-C-
methyl analogues of adenosine to adopt in solution a
South (2T3)-syn conformation around the glycosidic
bond, which is not adequate for the binding at ARs. The
2′-C-methyl modification is better tolerated, in particu-
lar at A1 receptor subtype. 2′-Me-CCPA was confirmed
to be a potent and highly selective agonist at bovine
A1AR. Its potency and selectivity was maintained also
at human A1 AR. Thus, 2′-Me-CCPA proved to be the
most selective agonist for human A1 adenosine receptor
versus human A3 receptor so far known, and could be a
useful pharmacological tool for investigation of A1
adenosine receptor-mediated events.
Figure 2. Superimposition of the docked structures of 2′-Me-
CCPA (orange) and 3′-Me-CCPA (green). The van der Waals
surface of the 3′-C-methyl of compound 3′-Me-CCPA, pointing
toward His278 imidazole ring, is shown in white. The side
chains of the important residues in proximity to the docked
molecules are highlighted and labeled.
xanthine-7-ribosides relative to the parent xanthines.21
(ii) Mutation of Thr91 to alanine in the A1 and A2A
receptors, respectively, was shown to substantially
decrease agonist affinity.20,22
As depicted in Figure 1, the cyclopentyl group of 2′-
Me-CCPA appears to be surrounded by a hydrophobic
pocket formed by Leu250 (TM6), Ile270 (TM7), Ile274
(TM7), and, to a lesser extent, by Thr257 (TM6), a
pocket that contributes to an increase in the affinity and
selectivity of 2′-C-methyl derivatives, in agreement with
experimental data (see Table 1). Several items of
mutational studies point toward these residues in the
ligand-binding process. In fact, chimera and site-
directed mutagenesis experiments identified Ile270 as
being primarily responsible for species differences in the
binding of N6-adenine-substituted compounds.23 The
size of this hydrophobic pocket is large enough to
accommodate bulkier N6 substituents, such as the
cyclohexyl group (compound 1), determining an en-
hancement in A1AR potency and a parallel gain in
selectivity vs A2A and A3 ARs. In derivative 2, the
presence of the N6-benzyl group, characterized by a
higher degree of conformational flexibility, increases the
steric hindrance inside the binding cavity, decreasing
the relative stability of the complex. This may be the
reason for the low A1AR affinity of 2. The chlorine atom
of 2′-Me-CCPA favors the anchoring of the ligand into
the binding site, making hydrophobic interactions with
the Trp247 indole ring as well as slightly favorable
electrostatic interactions with the Asn234 and Gln92
NH2 groups (Figure 1). The interactions observed here
could justify the enhanced potency of 2′-Me-CCPA (Ki
) 3 nM) with respect to 2′-Me-CPA (Ki ) 10 nM),5 which
does not have the chlorine atom at position 2. Mutation
of Gln92 of the human A2A adenosine receptor has been
shown to affect agonist and antagonist binding.22 Trp247,
highly conserved in the G protein-coupled receptors
(GPCRs) superfamily, has been postulated as playing a
key role in the activation of GPCRs.24 In fact, for hA3AR,
it has been found that the corresponding mutant W243A
receptor is able to bind agonists but fails to activate the
Finally, a model of the bovine A1AR was built to
rationalize the higher affinity and selectivity of 2′-C-
methyl derivatives of N6-substituted-adenosine com-
pared to that of 3′-C-methyl analogues. In the docking
exploration, it was found that 2′-Me-CCPA was able to
form a number of interactions with several polar
residues in the transmembrane helices TM-3, TM-6, and
TM-7 of bA1AR which were not preserved in the molec-
ular dynamics simulation of 3′-Me-CCPA/ bA1AR com-
plex.
Experimental Section
Chemistry. Melting points were determined on a Bu¨chi
apparatus and are uncorrected. Elemental analyses were
determined on an EA 1108 CHNS-O (Fisons Instruments)
analyzer. Thin-layer chromatography (TLC) was run on silica
gel 60 F254 plates (Merck); silica gel 60 (70-230 mesh, Merck)
for column chromatography was used. Nuclear magnetic
resonance 1H NMR spectra were determined with a Varian
VXR-300 spectrometer at 300 MHz. The chemical shift values
are expressed in δ values (parts per million) relative to
tetramethylsilane as an internal standard. All exchangeable
protons were confirmed by addition of D2O. Stationary NOE
experiments were run on degassed solutions at 25 °C. A
presaturation delay of 1 s was used, during which the

Ribose-Modified Adenosine Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5 1557
decoupler low power was set at 20 dB attenuation. Mass
spectroscopy was carried out on an HP 1100 series instrument.
All measurements were performed in the positive ion mode
using an atmospheric pressure electrospray ionization (API-
ESI).
General Procedure for the Amination of 10, 12, and
13 into Compounds 1, 2, 4-6, 8. To a solution of 10, 12, or
139 (1.0 mmol) in anhydrous EtOH (15 mL) under nitrogen
atmosphere, the appropriate amine (molar ratio 1:6) was
added. The reaction mixture was heated at reflux for the time
reported below, concentrated in vacuo, and treated with
methanolic ammonia (saturated at 0 °C) (30 mL) at room-
temperature overnight to obtain the deblocked compounds
which were purified by chromatography.
General Procedure for the Amination of 11, 12 and
13 into Compounds 3, 7 and 9. A stirred solution of 11, 12
or 13 (1.0 mmol) in anhydrous EtOH (15 mL) was treated with
3-iodobenzylamine hydrochloride (1.1 mmol) and TEA (3.1
mmol), and the mixture was refluxed for the time reported
below. The solvent was evaporated to dryness and the residue
was dissolved in H2O (20 mL) and extracted with CH2Cl2 (3 ×
10 mL). The organic layers were dried over anhydrous Na2SO4,
and evaporated to dryness. The oily residue was treated with
methanolic ammonia (40 mL, saturated at 0 °C) and the
solution was stirred at room-temperature overnight. After
concentration in vacuo the residue was purified by chroma-
tography.
N6-(3-Iodobenzyl)-2-chloro-9H-(2-C-methyl-â-D-ribo-
furanosyl)adenine (3). Compound 3 was obtained starting
from 11 (reaction time 2 h). Chromatography on a silica gel
column (CHCl3-MeOH, 96:4) gave 3 as a white solid (87%
yield). 1H NMR (DMSO-d6): δ 0.80 (s, 1H, CH3), 3.65-4.10
(m, 4H, H-3′, H-4′, H-5′), 4.60 (d, J ) 5.5 Hz, 2H, CH2Ph,),
5.20 (m, 2H, OH), 5.32 (s, 1H, OH), 5.85 (s, 1H, H-1′), 7.15 (t,
J ) 7.7 Hz, 1H, Ph), 7.35 (d, J ) 7.7 Hz, 1H, Ph), 7.60 (d, J )
8.1 Hz, 1H, Ph), 7.75 (s, 1H, Ph), 8.55 (s, 1H, H-8), 8.90 (t, J
) 5.5 Hz, 1H, NH). MS: m/z 532.74 [M + H]+. Anal. (C18H19-
ClIN5O4) C, H, N.
N6-Cyclohexyl-9H-(2-C-methyl-â-D-ribofuranosyl)ade-
nine (1). Reaction of 10 with cyclohexylamine for 2 h followed
by chromatography on a silica gel column (CHCl3-MeOH, 95:
5) gave 1 as a white solid (82% yield). 1H NMR (DMSO-d6): δ
0.78 (s, 3H, CH3), 1.35 (m, 4H, cyclohexyl), 1.60-1.90 (2m, 6H,
cyclohexyl), 3.70, 3.80 (2m, 2H, H-5′), 3.90 (m, 1H, H-4′), 4.05
(m, 1H, CHNH), 4.10 (m, 1H, H-3′), 5.20 (m, 3H, OH), 5.95 (s,
1H, H-1′), 7.60 (d, J ) 8.8 Hz, 1H, NH), 8.20 (s, 1H, H-2), 8.45
(s, 1H, H-8). MS: m/z 364.42 [M + H]+. Anal. (C17H25N5O4) C,
H, N.
N6-(3-Iodobenzyl)-9H-(3-C-methyl-â-D-ribofuranosyl)-
adenine (7). The title compound was obtained starting from
12 (reaction time 4 h). Chromatography on a silica gel column
N6-Benzyl-9H-(2-C-methyl-â-D-ribofuranosyl)adenine
(2). Reaction of 10 with benzylamine for 2 h followed by
chromatography on a silica gel column (CHCl3-MeOH, 94:6)
1
1
(CHCl3-MeOH, 93:7) gave 7 as a white solid (80% yield). H
gave 2 as a white solid (83% yield). H NMR (DMSO-d6): δ
NMR (DMSO-d6): δ 1.30 (s, 3H, CH3), 3.60 (m, 2H, H-5′), 3.90
(br s, 1H, H-4′), 4.45 (t, J ) 7.3 Hz, 1H, H-2′), 4.70 (br s, 2H,
CH2 Ph), 4.85 (s, 1H, OH), 5.40 (d, J ) 6.6 Hz, 1H, OH), 5.85
(d and dd, J ) 8.0 Hz after exchange with D2O, 2H, H-1′, OH),
7.10 (t, J ) 7.9 Hz, 1H, Ph), 7.35 (d, J ) 7.7 Hz, 1H, Ph), 7.60
(d, J ) 7.3 Hz, 1H, Ph), 7.70 (s, 1H, Ph), 8.20 (s, 1H, H-2),
8.40 (s, 1H, H-8), 8.55 (br s, 1H, NH). MS: m/z 498.29 [M +
H]+. Anal. (C18H20IN5O4) C, H, N.
N6-(3-Iodobenzyl)-2-chloro-9H-(3-C-methyl-â-D-ribo-
furanosyl)adenine (9). Compound 9 was obtained starting
from 13 (reaction time 4 h). Chromatography on a silica gel
column (CHCl3-MeOH, 96:4) gave 9 as a white solid (85%
yield). 1H NMR (DMSO-d6): δ 1.30 (s, 3H, CH3), 3.55 (m, 2H,
H-5′), 3.90 (br s, 1H, H-4′), 4.35 (t, J ) 7.3 Hz, 1H, H-2′), 4.60
(d, J ) 5.5 Hz, 2H, CH2Ph), 4.90 (s, 1H, OH), 5.17 (t, J ) 5.3
Hz, 1H, OH), 5.43 (d, J ) 7.0 Hz, 1H, OH), 5.80 (d, J ) 8.1
Hz, 1H H-1′), 7.15 (t, J ) 7.7 Hz, 1H, Ph), 7.30 (d, J ) 7.0 Hz,
1H, Ph), 7.60 (d, J ) 8.8 Hz, 1H, Ph), 7.74 (s, 1H, Ph), 8.45 (s,
1H, H-8), 9.00 (t, J ) 6.0 Hz, 1H, NH). MS: m/z 532.74 [M +
H]+. Anal. (C18H19ClIN5O4) C, H, N.
6-Chloro-9H-(3-C-methyl-â-D-ribofuranosyl)purine (16).
A mixture of 12 (1.0 g, 2.0 mmol) and methanolic ammonia
(saturated at 0 °C, 50 mL) was stirred at room temperature
for 4 h. Compound 16 precipitated from the reaction mixture,
and the solid was filtered, washed with MeOH, and dried. The
filtrate was evaporated to dryness, and the residue was
purified by silica gel column chromatography (CHCl3-MeOH
94:6) to give 16 as a white solid. Overall yield: 55%. 1H NMR
(DMSO-d6): δ 1.30 (s, 3H, CH3), 3.60 (m, 2H, H-5′), 3.90 (m,
1H, H-4′), 4.45 (t, J ) 7.0 Hz, 1H, H-2′), 4.97 (s, 1H, OH), 5.20
(t, J ) 4.9 Hz, 1H, OH), 5.50 (d, J ) 6.6 Hz, 1H, OH), 6.04 (d,
J ) 7.7 Hz, 1H, H-1′), 8.82 (s, 1H, H-2), 8.95 (s, 1H, H-8). MS:
m/z 301.70 [M + H]+. Anal. (C11H13ClN4O4) C, H, N.
General Procedure for the Synthesis of 2′,3′-O-Isopro-
pylidene Derivatives 17, 22, 29, and 32. A mixture of 3, 7,
16, or 21 (1.0 mmol), 2,2-dimethoxypropane (18.1 mmol),
camphorsulfonic acid (1.0 mmol) in anhydrous acetone (10 mL)
was stirred at 55 °C for the time reported below. The solvent
was removed in vacuo, and the residue was purified by
chromatography.
0.80 (s, 3H, CH3), 3.70 (m, 2H, H-5′), 3.80 (m, 1H, H-4′), 4.10
(dd, J ) 6.8, 9.0 Hz, 1H, H-3′), 4.70 (br s, 2H, CH2Ph), 5.20
(m, 3H, OH), 5.95 (s, 1H, H-1′), 7.30 (m, 5H, Ph), 8.20 (s, 1H,
H-2), 8.40 (m, 1H, NH), 8.50 (s, 1H, H-8). MS: m/z 372.40 [M
+ H]+. Anal. (C18H21N5O4) C, H, N.
N6-Cyclopentyl-9H-(3-C-methyl-â-D-ribofuranosyl)ade-
nine (4). Reaction of 12 with cyclopentylamine for 2 h followed
by chromatography on a silica gel column (CHCl3-MeOH, 97:
3) gave 4 as a white solid (83% yield). 1H NMR (DMSO-d6): δ
1.30 (s, 3H, CH3), 1.50-2.00 (2m, 8H, cyclopentyl), 3.60 (m,
2H, H-5′), 3.88 (br s, 1H, H-4′), 4.45 (t, J ) 7.3 Hz, 1H, H-2′),
4.80 (s, 1H, OH), 5.40 (d, J ) 7.0 Hz, 1H, OH), 5.80 (d, J ) 7.7
Hz, 1H, H-1′), 5.93 (dd, J ) 3.3, 8.4 Hz, 1H, OH), 7.82 (d, J )
7.0 Hz, 1H, NH), 8.15 (s, 1H, H-2), 8.32 (s, 1H, H-8). MS: m/z
350.39 [M + H]+. Anal. (C16H23N5O4) C, H, N.
N6-Cyclohexyl-9H-(3-C-methyl-â-D-ribofuranosyl)ade-
nine (5). Reaction of 12 with cyclohexylamine for 3 h followed
by chromatography on a silica gel column (CHCl3-MeOH, 98:
2) gave 5 as a white solid (89% yield). 1H NMR (DMSO-d6): δ
1.10-1.40 (m, 5H, cyclohexyl), 1.32 (s, 3H, CH3), 1.60-1.96
(m, 5H, cyclohexyl), 3.60 (m, 2H, H-5′), 3.90 (t, J ) 3.1 Hz,
1H, H-4′), 4.45 (t, J ) 7.0 Hz, 1H, H-2′), 4.80 (s, 1H, OH), 5.40
(d, J ) 5.9 Hz, 1H, OH), 5.82 (d, J ) 8.1 Hz, 1H, H-1′), 5.90
(dd, J ) 3.3, 8.4 Hz, 1H, OH), 7.70 (d, J ) 8.4 Hz, 1H, NH),
8.15 (s, 1H, H-2), 8.35 (s, 1H, H-8), MS: m/z 364.42 [M + H]+.
Anal. (C17H25N5O4) C, H, N.
N6-Benzyl-9H-(3-C-methyl-â-D-ribofuranosyl)adenine
(6). Reaction of 12 with benzylamine for 2 h followed by
chromatography on a silica gel column (CHCl3-MeOH, 96:4)
1
gave 6 as a white solid (79% yield). H NMR (DMSO-d6): δ
1.30 (s, 3H, CH3), 3.60 (m, 2H, H-5′), 3.90 (br s, 1H, H-4′), 4.45
(t, J ) 7.5 Hz, 1H, H-2′), 4.70 (m, 2H, CH2Ph), 4.80 (s, 1H,
OH), 5.4 (d, J ) 7.0 Hz, 1H, OH), 5.80 (d and t, J ) 8.1 Hz
after exchange with D2O, 2H, H-1′, OH), 7.30 (m, 5H, Ph), 8.18
(s, 1H, H-2), 8.35 (s, 1H, H-8), 8.50 (br s, 1H, NH). MS: m/z
372.40 [M + H]+. Anal. (C18H21N5O4) C, H, N.
N6-Cyclopentyl-2-chloro-9H-(3-C-methyl-â-D-ribofur-
anosyl)adenine (8). Reaction of 13 with cyclopentylamine for
3 h followed by chromatography on a silica gel column (CHCl3-
MeOH, 97:3) gave 8 as a white solid (68% yield). 1H NMR
(DMSO-d6): δ 1.30 (s, 3H, CH3), 1.50-2.00 (2m, 8H, cyclo-
pentyl), 3.55 (m, 2H, H-5′), 3.85 (br s, 1H, H-4′), 4.40 (m, 2H,
H-2′, CHNH), 4.87 (s, 1H, OH), 5.20 (t, J ) 5.3 Hz, 1H, OH),
5.42 (d, J ) 7.0 Hz, 1H, OH). 5.80 (d, J ) 8.0 Hz, 1H, H-1′),
8.35 (s and d, 2H, NH, H-8). MS: m/z 384.84 [M + H]+. Anal.
(C16H22ClN5O4) C, H, N.
6-Chloro-(3-C-methyl-2,3-O-isopropylidene-â-D-ribo-
furanosyl)purine (17). The title compound was obtained
starting from 16 (reaction time 2 h). Chromatography on a
silica gel column (CHCl3-MeOH, 96:4) gave 17 as a white foam
(72% yield). 1H NMR (DMSO-d6): δ 1.40 (s, 3H, CH3), 1.53 (s,
6H, CH3), 3.60 (m, 2H, H-5′), 4.12 (t, J ) 5.8 Hz, 1H, H-4′),

1558 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5
Cappellacci et al.
4.95 (t, J ) 4.9 Hz, 1H, OH), 5.03 (d, J ) 1.8 Hz, 1H, H-2′),
6.24 (d, J ) 2.2 Hz, 1H, H-1′), 8.82 (s, 1H, H-2) 8.87 (s, 1H,
H-8). Anal. (C14H17ClN4O4) C, H, N.
4.60 (br s, 3H, CH2Ph, H-4′), 5.02 (s, 1H, H-2′), 6.20 (s, 1H,
H-1′), 7.15 (t, J ) 7.7 Hz, 1H, Ph), 7.38 (d, J ) 7.3 Hz, 1H, Ph)
7.60 (d, J ) 7.7 Hz, 1H, Ph), 7.72 (s, 1H, Ph), 8.42 (s, 1H, H-8),
9.0 (br s, 1H, NH), 13.30 (br s, 1H, COOH). Anal. (C21H21-
ClIN5O5) C, H, N.
General Procedure for the Synthesis of Ethyl Esters
19, 24, 31, and 34. To a solution of 18, 23, 30, or 33 (0.65
mmol) in anhydrous EtOH (28 mL) at 0 °C was added dropwise
0.23 mL of SOCl2 and the solution was stirred under nitrogen
atmosphere at room temperature for the time reported below.
After evaporation in vacuo the residue was purified by chro-
matography.
9H-(3-C-Methyl-2,3-O-isopropylidene-â-D-ribofuran-
osyl)adenine (22). The title compound was obtained starting
from 21 (reaction time 11 h). Chromatography on a silica gel
column (CHCl3-MeOH, 93:7) gave 22 as a white solid (58%
yield). 1H NMR (DMSO-d6): δ 1.38 (s, 3H, CH3), 1.53 (2s, 6H,
CH3), 3.60 (m, 2H, H-5′), 4.05 (m, 1H, H-4′), 4.96 (m, 2H, H-2′,
OH), 6.08 (d, J ) 2.2 Hz, 1H, H-1′), 7.35 (br s, 2H, NH2), 8.15
(s, 1H, H-2), 8.32 (s, 1H, H-8). Anal. (C14H19N5O4 ) C, H, N.
N6-(3-Iodobenzyl)-2-chloro-9H-(2-C-methyl-2,3-O-iso-
propylidene-â-D-ribofuranosyl)adenine (29). The title com-
pound was obtained starting from 3 (reaction time 6 h).
Chromatography on a silica gel column (CHCl3-EtOAc, 99:1)
Ethyl 1-Deoxy-1-(6-chloro-9H-purin-9-yl)-3-C-methyl-
2,3-O-isopropylidene-â-D-ribofuranuroate (19). The title
compound was obtained starting from 18 (reaction time 3 h).
Chromatography on a silica gel column (CHCl3-MeOH, 99:1)
gave 19 as a foam containing a small amount of the insepa-
rable ethyl 1-deoxy-1-(6-ethoxy-9H-purin-9-yl)-3-C-methyl-2,3-
O-isopropylidene-â-D-ribofuranuroate (overall yield 48%).
Ethyl 1-Deoxy-1-(6-amino-9H-purin-9-yl)-3-C-methyl-
2,3-O-isopropylidene-â-D-ribofuranuroate (24). The title
compound was obtained starting from 23 (reaction time 7 h).
Chromatography on a silica gel column (CHCl3-MeOH, 92:8)
1
gave 29 as a white solid (89% yield). H NMR (DMSO-d6): δ
1.15 (s, 3H, CH3), 1.38 (s, 3H, CH3), 1.55 (s, 3H, CH3), 3.70
(m, 2H, H-5′), 4.25 (m, 1H, H-4′), 4.60 (m, 3H, H-3′, CH2Ph),
5.26 (t, J ) 5.5 Hz, 1H, OH), 6.15 (s, 1H, H-1′), 7.15 (t, J ) 7.7
Hz, 1H, Ph), 7.35 (d, J ) 8.1 Hz, 1H, Ph), 7.60 (d, J ) 8.1 Hz,
1H, Ph) 7.75 (s, 1H, Ph), 8.40 (s, 1H, H-8), 8.95 (t, J ) 6.0 Hz,
1H, NH). Anal. (C21H23ClIN5O4) C, H, N.
N6-(3-Iodobenzyl)-2-chloro-9H-(3-C-methyl-2,3-O-iso-
propylidene-â-D-ribofuranosyl)adenine (32). The title com-
pound was obtained starting from 7 (reaction time 4 h).
Chromatography on a silica gel column (CHCl3-MeOH, 99:1)
1
gave 24 as a white solid (62% yield). H NMR (DMSO-d6): δ
1.20 (t, J ) 7.0 Hz, 3H, CH3CH2), 1.40 (s, 3H, CH3), 1.50 (s,
3H, CH3), 1.53 (s, 3H, CH3), 4.15 (pseudo t, 2H, CH2CH3), 4.70
(s, 1H, H-4′), 5.15 (s, 1H, H-2′), 6.25 (s, 1H, H-1′), 7.40 (br s,
2H, NH2), 8.18 (s, 1H, H-2), 8.38 (s, 1H, H-8). Anal. (C16H21N5O5)
C, H, N.
1
gave 32 as a white solid (74% yield). H NMR (DMSO-d6): δ
1.30 (s, 3H, CH3), 1.52 (2s, 6H, CH3), 3.55 (m, 2H, OH), 4.10
(m, 1H, H-4′), 4.60 (d, J ) 5.8 Hz, 2H, CH2Ph), 4.90 (d, J )
2.6 Hz, 1H, H-2′), 6.0 (d, J ) 2.6 Hz, 1H, H-1′), 7.15 (t, J ) 7.7
Hz, 1H, Ph), 7.35 (d, J ) 8.1 Hz, 1H, Ph), 7.60 (d, J ) 8.1 Hz,
1H, Ph), 7.75 (s, 1H, Ph), 8.40 (s, 1H, H-8), 9.0 (pseudo t, 1H,
NH). Anal. (C21H23ClIN5O4) C, H, N.
Ethyl 1-Deoxy-1-[N6-(3-iodobenzyl)-2-chloro-9H-purin-
9-yl]-2-C-methyl-2,3-O-isopropylidene-â-D-ribofur-
anuroate (31). The title compound was obtained starting from
30 (reaction time 7 h). Chromatography on a silica gel column
(CHCl3-EtOAc, 95:5) gave 31 as a white solid (57% yield). 1H
NMR (DMSO-d6): δ 1.06 (t, J ) 7.1 Hz, 3H, CH3CH2), 1.18 (s,
3H, CH3), 1.44 (s, 3H, CH3), 1.52 (s, 3H, CH3), 4.0 (m, 2H,
CH2CH3), 4.60 (d, J ) 6.2 Hz, 2H, CH2Ph), 4.82 (d, J ) 2.9
Hz, 1H, H-4′), 5.25 (d, J ) 2.6 Hz, 1H, H-3′), 6.30 (s 1H, H-1′),
7.15 (t, J ) 7.7 Hz, 1H, Ph), 7.35 (d, J ) 8.1 Hz, 1H, Ph), 7.60
(d, J ) 8.1 Hz, 1H, Ph), 7.75 (s, 1H, Ph), 8.37 (s, 1H, H-8), 9.0
(t, J ) 6.0 Hz, 1H, NH). Anal. (C23H25ClIN5O5) C, H, N.
Ethyl 1-Deoxy-1-[N6-(3-iodobenzyl)-2-chloro-9H-purin-
9-yl]-3-C-methyl-2,3-O-isopropylidene-â-D-ribofur-
anuroate (34). The title compound was obtained starting from
33 (reaction time 6 h at 40 °C). Chromatography on a silica
gel column (CHCl3-EtOAc, 95:5) gave 34 as a white solid (60%
yield). 1H NMR (DMSO-d6): δ 1.12 (t, J ) 7.1 Hz, 3H,
CH2CH3), 1.40 (s, 3H, CH3), 1.55 (2s, 6H, CH3), 4.10 (m, 2H,
CH2CH3), 4.60 (d, J ) 5.9 Hz, 2H, CH2Ph), 4.70 (s, 1H, H-4′),
5.10 (d, J ) 1.5 Hz, 1H, H-2′), 6.20 (d, J ) 1.5 Hz, 1H, H-1′),
7.12 (t, J ) 7.7 Hz, 1H, Ph), 7.35 (d, J ) 7.7 Hz, 1H, Ph), 7.60
(d, J ) 8.0 Hz, 1H, Ph), 7.73 (s, 1H, Ph), 8.40 (s, 1H, H-8), 9.0
(t, J ) 6.0 Hz, 1H, NH). Anal. (C23H25ClIN5O5) C, H, N.
Ethyl 1-Deoxy-1-[N6-(3-iodobenzyl)-9H-purin-9-yl]-3-C-
methyl-2,3-O-isopropylidene-â-D-ribofuranuroate (20). A
stirred solution of 19 (1.0 mmol) in anhydrous EtOH (15 mL)
was treated with 3-iodobenzylamine hydrochloride (1.1 mmol)
and TEA (3.1 mmol), and the mixture was refluxed for 6 h.
The solvent was evaporated to dryness, and the residue was
dissolved in H2O (20 mL) and extracted with CH2Cl2 (3 × 10
mL). The organic layers were dried over anhydrous Na2SO4
and concentrated under reduced pressure. The oily residue was
purified by chromatography on a silica gel column (CHCl3-
EtOAc, 96:4) to give 20 as a white solid (60% yield). 1H NMR
(DMSO-d6): δ 1.15 (t, J ) 7.1 Hz, 3H, CH2CH3), 1.37 (s, 3H,
CH3), 1.50 (s, 3H, CH3), 1.55 (s, 3H, CH3) 4.15 (m, 2H,
CH2CH3), 4.65 (br s, 2H, CH2Ph), 4.72 (s, 1H, H-4′), 5.15 (d, J
) 1.8 Hz, 1H, H-2′), 6.25 (d, J ) 1.8 Hz, 1H, H-1′), 7.10 (t, J
) 7.9 Hz, 1H, Ph), 7.35 (d, J ) 7.3 Hz, 1H, Ph), 7.56 (d, J )
7.3 Hz, 1H, Ph), 7.75 (s, 1H, Ph), 8.22 (s, 1H, H-2), 8.40 (s,
1H, H-8), 8.53 (br s, 1H, NH). Anal. (C23H26IN5O5) C, H, N.
General Procedure for the Synthesis of Uronamides
14, 15, 25, and 26. A mixture of 20, 24, 31, or 34 (0.14 mmol)
and the suitable amine (1.1 mL) was stirred at -20 °C for 3 h
General Procedure for the Synthesis of Uronic Acids
18, 23, 30 and 33. A mixture of 17, 22, 29, or 32 (1.0 mmol),
BAIB (2.2 mmol), TEMPO (0.2 mmol) in 2 mL of 1:1 CH3CN-
H2O solution was stirred at room temperature for the time
reported below. A precipitate was obtained for compounds 23
and 33 which was filtered, triturated sequentially with diethyl
ether and acetone, and dried in vacuo. In the case of com-
pounds 18 and 30 the solvent was evaporated to dryness and
the resulting residue was triturated with diethyl ether, filtered
and dried in vacuo.
1-Deoxy-1-(6-chloro-9H-purin-9-yl)-3-C-methyl-2,3-O-
isopropylidene-â-D-ribofuranoic acid (18). The title com-
pound was obtained starting from 17 (reaction time 2 h) as a
1
solid (80% yield). H NMR (DMSO-d6 ): δ 1.40 (s, 6H, CH3),
1.52 (s, 3H, CH3), 4.50 (s, 1H, CH3), 4.78 (d, J ) 1.8 Hz, 1H,
H-2′), 6.38 (d, J ) 1.8 Hz, 1H, H-1′), 8.57 (s, 1H, H-2), 8.80 (s,
1H, H-8), 13.0 (br s, 1H, COOH). Anal. (C14H15ClN4O5) C, H,
N.
1-Deoxy-1-(6-amino-9H-purin-9-yl)-3-C-methyl-2,3-O-
isopropylidene-â-D-ribofuranoic acid (23). Compound 23
was obtained starting from 22 (reaction time 3 h) as a solid
1
(54% yield). H NMR (DMSO-d6): δ 1.38 (s, 3H, CH3), 1.52,
1.56 (2s, 6H, CH3), 4.60 (s, 1H, H-4′), 5.05 (d, J ) 1.8 Hz, 1H,
H-2′), 6.20 (d, J ) 1.8 Hz, 1H, H-1′), 7.40 (br s, 2H, NH2), 8.18
(s, 1H, H-2), 8.38 (s, 1H, H-8) 13.40 (br s, 1H, COOH). Anal.
(C14H17N5O5) C, H, N.
1-Deoxy-1-[N6-(3-iodobenzyl)-2-chloro-9H-purin-9-yl]-
2-C-methyl-2,3-O-isopropylidene-â-D-ribofuranoic acid
(30). The title compound was obtained starting from 29
1
(reaction time 3 h), as a solid (90% yield). H NMR (DMSO-
d6): δ 1.08 (s, 3H, CH3), 1.40 (s, 3H, CH3), 1.52 (s, 3H, CH3),
4.57 (d, J ) 1.9 Hz, 2H, CH2Ph), 4.77 (s, 1H, H-4′), 5.03 (s,
1H, H-3′), 6.28 (s, 1H, H-1′), 7.15 (q, J ) 7.9 Hz, 1H, Ph), 7.36
(d, J ) 7.0 Hz, 1H, Ph), 7.60 (d, J ) 7.6 Hz, 1H, Ph), 7.76 (s,
1H, Ph), 8.50 (s, 1H, H-8), 8.96 (t, J ) 6.0 Hz, 1H, NH), 13.30
(br s, 1H, COOH). Anal. (C21H21ClIN5O5) C, H, N.
1-Deoxy-1-[N6-(3-iodobenzyl)-2-chloro-9H-purin-9-yl]-
3-C-methyl-2,3-O-isopropylidene-â-D-ribofuranoic acid
(33). The title compound was obtained starting from 32
1
(reaction time 2 h), as a solid (74% yield). H NMR (DMSO-
d6): δ 1.30 (s, 1H, CH3), 1.46 (s, 3H, CH3), 1.48 (s, 3H, CH3),

Ribose-Modified Adenosine Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5 1559
1
and then at room-temperature overnight. The solution was
evaporated to dryness, and then 90% HCOOH was added and
stirred at 45 °C for 7 h. The solvent was removed in vacuo,
and the residue was coevaporated with MeOH (4 × 10 mL),
and then purified by chromatography.
gave 28 as a white solid (51% yield). H NMR (DMSO-d6): δ
1.15 (s, 3H, CH3), 2.68 (d, J ) 4.4 Hz, 3H, NHCH3), 2.77 (d, J
) 4.4 Hz, 3H, NHCH3), 4.20 (s, 1H, H-4′), 4.60 (br s, 3H,
CH2Ph, H-2′), 5.22 (s, 1H, OH), 5.55 (d, J ) 6.1 Hz, 1H, OH),
5.82 (d, J ) 8.0 Hz, 1H, H-1′), 6.28 (d, J ) 4.4 Hz, 1H, NHCH3),
7.10 (t, J ) 7.7 Hz, 1H, Ph), 7.38 (d, J ) 7.7 Hz, 1H, Ph), 7.58
(d, J ) 7.3 Hz, 1H, Ph), 7.76 (s, 1H, Ph), 8.10 (br s, 1H,
CH3NHCO), 8.15 (s, 1H, H-8), 8.35 (br s, 1H, NHCH2). MS:
m/z 554.36 [M + H]+. Anal. (C20H24 IN7O4) C, H, N.
Computational Procedures. All model building, energy
minimizations, and molecular dynamics calculations were
carried out using SYBYL 6.926 and AMBER 4.127,28 modeling
packages, respectively. All manipulations were performed on
a Silicon Graphics R12000 workstation.
N-Ethyl-1-deoxy-1-(6-amino-9H-purin-9-yl)-3-C-meth-
yl-â-D-ribofuranuronamide (14). The title compound was
obtained starting from 24 and anhydrous ethylamine. Chro-
matography on a silica gel column (CHCl3-MeOH, 85:15) gave
1
14 as a white solid (60% yield). H NMR (DMSO-d6): δ 1.10
(t, J ) 7.3 Hz, 3H, CH3CH2), 1.18 (s, 3H, CH3), 3.20 (m, 2H,
CH2CH3), 4.22 (s, 1H, H-4′), 4.40 (t, J ) 7.0 Hz, 1H, H-2′),
5.30 (s, 1H, OH), 5.60 (d, J ) 7.0 Hz, 1H, OH), 5.90 (d, J ) 8.1
Hz, 1H, H-1′), 7.45 (br s, 2H, NH2), 8.20 (s, 1H, H-2), 8.45 (s,
1H, H-8), 9.18 (t, J ) 5.5 Hz, 1H, NH). MS: m/z 323.33 [M +
H]+. Anal. (C13H18N6O4) C, H, N.
Adenosine A1 Receptor Model Building. The structural
model of the human A1AR was built using the recently
reported 2.8 Å crystal structure of bovine rhodopsin15 (PDB
entry code: 1F88) as a structural template. Briefly, sequences
of the human A1AR transmembrane domains were amended
by comparison to the corresponding domains of rhodopsin,
according to a published sequence alignment.29 Individual TM
helical segments were built as ideal helices (using φ-ψ angles
of -63.0° and -41.6°) with side chains placed in prevalent
rotamers and representative proline kink geometries. Each
model helix was capped with an acetyl group at the N-terminus
and an N-methyl group at the C-terminus. These structures
were then grouped by adding one at a time until a helical
bundle (TM region), matching the overall characteristics of the
crystallographic structure of rhodopsin, had been obtained. The
relative orientations and interactions between the helices were
adjusted based on incorporated structural inferences from
available experimental data, such as mutation and ligand
binding studies,30 cysteine scanning data,31 and site-directed
mutation experiments.32 Because earlier work showed that
polarity conserved positions cluster together in the cores of
proteins to create conserved hydrogen-bonding interactions,33
we refined the model by applying the additional hydrogen-
bonding constraints between the conserved polar residues
Asn27, Asp55, and Asn284 in accordance with data from site-
directed mutagenesis.32,33 The helical bundle was subjected to
energy-minimization using the SANDER module of the AM-
BER suite of programs27,28 until the rms value of the coniugate
gradient was 0.001 kcal/mol per Å. An energy penalty force
constant of 5 kcal/Å2/mol on the protein backbone atoms was
applied throughout these calculations.
For the conformational refinement of the bA1AR, the
minimized structure was then used as the starting point for
subsequent 200 ps of molecular dynamics (MD), during which
the positional constraints on the protein backbone atoms were
gradually released from 5 to 0.05 kcal/Å2/mol. The options of
MD at 300 K with 0.2 ps coupling constant were a time step
of 1 fs and a nonbonded update every 25 fs. The lengths of
bonds with hydrogen atoms were constrained according to the
SHAKE algorithm.34 The average structure from the last 50
ps trajectory of MD was reminimized with backbone con-
straints in the secondary structure. The conformational valid-
ity of main chain and side chain torsions in each residue within
the protein models was analyzed using the PROCHECK
program.35 Also, all ω angles for the peptide planarity were
measured. The chirality of all CR atoms, which in naturally
occurring amino acids is of the l-configuration, was checked.
RMS deviations between backbone atoms in all helices were
compared to the X-ray structure of rhodopsin as a template.
N-Methyl-1-deoxy-1-[N6-(3-iodobenzyl)-9H-purin-9-yl]-
3-C-methyl-â-D-ribofuranuronamide (15). The title com-
pound was obtained starting from 20 and anhydrous methyl-
amine. Chromatography on a silica gel column (CHCl3-MeOH,
1
94:6) gave 15 as a white solid (66% yield). H NMR (DMSO-
d6): δ 1.15 (s, 3H, CH3), 2.70 (d, J ) 4.0 Hz, 3H, CH3NH),
4.23 (s, 1H, H-4′), 4.40 (m, 1H, H-2′), 4.68 (m, 2H, CH2Ph),
5.32 (s, 1H, OH), 5.60 (d, J ) 6.8 Hz, 1H, OH), 5.94 (d, J ) 8.1
Hz, 1H, H-1′), 7.10 (t, J ) 7.5 Hz, 1H, Ph), 7.35 (d, J ) 7.0 Hz,
1H, Ph), 7.58 (d, J ) 8.8 Hz, 1H, Ph), 7.72 (s, 1H, Ph), 8.32 (s,
1H, H-2), 8.52 (s, 1H, H-8), 8.60 (br s, 1H, NHCH3), 9.14 (s,
1H, NH). MS: m/z 525.32 [M + H]+. Anal. (C19H21IN6O4) C,
H, N.
N-Methyl-1-deoxy-1-[(N6-3-iodobenzyl)-2-chloro-9H-
purin-9-yl]-2-C-methyl-â-D-ribofuranuronamide (25). The
title compound was obtained starting from 31. Chromatogra-
phy on a silica gel column (CHCl3-MeOH, 97:3) gave 25 as a
1
white solid (70% yield). H NMR (DMSO-d6): δ 0.80 (s, 3H,
CH3), 2.70 (d, J ) 4.4 Hz, 3H, CH3NH), 4.0 (pseudo t, 1H, H-3′),
4.26 (d, J ) 8.1 Hz, 1H, H-4′), 4.55 (d, J ) 5.5 Hz, 2H, CH2Ph),
5.46 (s, 1H, OH), 5.60 (d, J ) 5.5 Hz, 1H, OH), 5.88 (s, 1H,
H-1′), 7.10 (t, J ) 7.7 Hz, 1H, Ph), 7.32 (d, J ) 8.1 Hz, 1H,
Ph), 7.58 (d, J ) 7.3 Hz, 1H, Ph), 7.70 (s, 1H, Ph), 8.40 (br s,
1H, NHCH3), 8.98 (pseudo t, 1H, NHCH2). MS: m/z 559.76
[M + H]+. Anal. (C19H20ClIN6O4) C, H, N.
N-Methyl-1-deoxy-1-[N6-(3-iodobenzyl)-2-chloro-9H-
purin-9-yl]-3-C-methyl-â-D-ribofuranuronamide (26). The
title compound was obtained starting from 34. Chromatogra-
phy on a silica gel column (CHCl3-MeOH, 97:3) gave 26 as a
1
white solid (86% yield). H NMR (DMSO-d6): δ 1.12 (s, 3H,
CH3), 2.70 (d, J ) 4.6 Hz, 3H, CH3NH), 4.22 (s, 1H, H-4′), 4.30
(d, J ) 6.7 Hz, 1H, H-2′), 4.60 (d, J ) 5.5 Hz, 2H, CH2Ph),
5.35 (s, 1H, OH), 5.60 (d, J ) 6.1 Hz, 1H, OH), 5.85 (d, J ) 8.2
Hz, 1H, H-1′), 7.10 (t, J ) 7.8 Hz, 1H, Ph), 7.34 (d, J ) 7.9 Hz,
1H, Ph), 7.58 (d, J ) 7.9 Hz, 1H, Ph), 7.72 (s, 1H, Ph), 8.48 (d,
J ) 4.3 Hz, 1H, NHCO), 8.58 (s, 1H, H-8), 9.05 (t, J ) 5.7 Hz,
1H, NHCH2). MS: m/z 559.76 [M + H]+. Anal. (C19H20ClIN6O4)
C, H, N.
N-Methyl-1-deoxy-1-[(N6-3-iodobenzyl)-2-methylamine-
9H-purin-9-yl]-2-C-methyl-â-D-ribofuranuronamide (27).
A mixture of 25 (90 mg, 0.16 mmol) and 2 N CH3NH2/THF
(13.4 mL) was heated in Parr bomb for 20 h at 70 °C. The
solution was evaporated to dryness, and the residue was
purified by chromatography on a silica gel column (CHCl3-
MeOH, 93:7) to obtain 27 as a white solid (45% yield). 1H NMR
(DMSO-d6): δ 0.85 (s, 3H, CH3), 2.65 (d, J ) 4.8 Hz, 3H,
NHCH3), 2.75 (d, J ) 4.8 Hz, NHCH3), 4.25 (m, 2H, H-4′, H-3′),
4.58 (br s, 2H, CH2Ph), 5.25 (s, 1H, OH), 5.54 (d, J ) 5.8 Hz,
1H, OH), 5.81 (s, 1H, H-1′), 6.33 (d, J ) 5.4 Hz, 1H, NHCH3),
7.10 (t, J ) 7.7 Hz, 1H, Ph), 7.40 (d, J ) 8.0 Hz, Ph), 7.56 (d,
J ) 8.0 Hz, 1H, Ph), 7.75 (s, 1H, Ph), 7.97 (br s, 1H,
CH3NHCO), 8.25 (s, 1H, NHCH2), 8.30 (s, 1H, H-8). MS: m/z
554.36 [M + H]+. Anal. (C20H24IN7O4) C, H, N.
Docking Simulations. Docking was performed with ver-
sion 3.05 of the program AutoDock.16 It combines a rapid
energy evaluation through precalculated grids of affinity
potentials with a variety of search algorithms to find suitable
binding positions for a ligand on a given protein. While the
protein is required to be rigid, the program allows torsional
flexibility in the ligand. Docking to bA1AR was carried out
using the empirical free energy function and the Lamarckian
genetic algorithm, applying a standard protocol, with an initial
population of 50 randomly placed individuals, a maximum
number of 1.5 × 106 energy evaluations, a mutation rate of
N-Methyl-1-deoxy-1-[N6-(3-iodobenzyl)-2-methylamine-
9H-purin-9-yl]-3-C-methyl-â-D-ribofuranuronamide (28).
Compound 28 was obtained starting from 26 (90 mg, 0.16
mmol) as described for 27 (reaction time 3 days at 90 °C).
Chromatography on a silica gel column (CHCl3-MeOH, 93:7)

1560 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5
Cappellacci et al.
0.02, a crossover rate of 0.80, and an elitism value of 1.
Proportional selection was used, where the average of the
worst energy was calculated over a window of the previous 10
generations. For the local search, the so-called pseudo-Solis
and Wets algorithm was applied using a maximum of 300
iterations per local search. The probability of performing local
search on an individual in the population was 0.06, and the
maximum number of consecutive successes or failures before
doubling or halving the local search step size was 4.50
independent docking runs were carried out for each ligand.
Results differing by less than 1.5 Å in positional root mean-
square deviation (rmsd) were clustered together and repre-
sented by the result with the most favorable free energy of
binding.
Biological Methods. Materials. [3H]1,3-dipropyl-8-cyclo-
pentylxanthine ([3H]DPCPX) (specific activity 108 Ci/mmol),
[3H](2-[4-(2-carboxyethyl)phenyl]ethyl-amino-5′-N-ethylcar-
boxyamido-adenosine ([3H]CGS21680) specific activity 42.5 Ci/
mmol), [125I]N6-(3-iodo-4-aminobenzyl)-5′-N-methylcarboxamido-
adenosine ([125I]AB-MECA) (specific activity 2000 Ci/mmol),
and [R-32P]ATP were purchased from PerkinElmer Life Sci-
ences. CHA and DPCPX were purchased from Research
Biochemical Incorporated (RBI, Natick, MA). Adenosine de-
aminase, forskolin, and GTP were from Sigma-Aldrich. Myo-
kinase and creatine kinase were purchased from Boehringer-
Mannheim (Mannheim, Germany). All other reagents were
from standard commercial sources and of the highest grade
commercially available.
Receptor Binding Assay and Adenylyl Cyclase Assay.
Displacement of [3H]DPCPX from A1 adenosine receptor in
bovine cortical membranes was performed as described;13
displacement of [3H]CGS-21680 from A2AAR in bovine striatal
membranes and [125I]AB-MECA from A3AR in bovine cortical
membranes were performed as described elsewhere.14 Com-
pounds were dissolved in DMSO and diluted with assay buffer
to the final concentration where the amount of DMSO never
exceeded 2%. At least six different concentration of each
compound were used. The experiments (n ) 4), carried out in
triplicate, were analyzed by an iterative curve fitting pro-
cedure (GraphPad, Prism program, San Diego, CA),
which provided IC50 and SEM values for tested com-
pounds. IC50 values were converted to Ki values using the
Cheng and Prusoff equation.41 The dissociation constant (Kd)
of [3H]DPCPX, [3H]CGS-21680, and [125I]AB-MECA was 0.5,
14, and 1.02 nM, respectively. Adenylyl cyclase assay was
performed as previously described.5
Binding Assay and Adenylyl Cyclase Assay at Cloned
Human Adenosine Receptors. Ki values were determined
in competition experiments with membranes from CHO cells
stably transfected with the individual human adenosine recep-
tor subtypes.42 For A1 receptors, 1 nM [3H]CCPA was used as
a radioligand, and [3H]NECA was used for the A2A (30 nM)
and A3 subtypes (10 nM). The relative potency (EC50 values)
at A2B adenosine receptors was determined measuring the
activation of adenylyl cyclase in a membrane preparation of
CHO cells stably transfected with the human A2B subtype
following the procedure described earlier.42
(1) Ligand Setup. The core structures of 2′-Me-CCPA and
3′-Me-CCPA were retrieved from the Cambridge Structural
Database (CSD)36 and modified using standard bond lengths
and bond angles of the SYBYL fragment library. The CSD
refcodes of the ligands are GIDZIC and BOSGEV, respectively.
Geometry optimizations were realized with the SYBYL/Maxi-
min2 minimizer by applying the BFGS (Broyden, Fletcher,
Goldfarb, and Shannon) algorithm37 and setting a rms gradient
of the forces acting on each atom of 0.05 kcal/mol Å as the
convergence criterion. Atomic charges were assigned using the
Gasteiger-Marsili formalism,38 which is the type of atomic
charges used in calibrating the AutoDock empirical free energy
function. Finally, the two compounds were setup for docking
with the help of AutoTors, the main purpose of which is to
define the torsional degrees of freedom to be considered during
the docking process. The number of flexible torsions defined
for each ligand is five.
Superimposition of the geometry-optimized ligand struc-
tures was carried out using the “Fit Atoms” method imple-
mented in SYBYL. The quality of the fit is represented by the
rms value computed for the matched atoms.
(2) Protein Setup. The energy-minimized structure of
bA1AR model was setup for docking as follows: polar hydro-
gens were added using the biopolymers module of the SYBYL
program, (Arg, Lys, Glu, and Asp residues were considered
ionized while all His were considered neutral by default), and
Kollman united-atom partial charges were assigned. Solvation
parameters were added to the final protein file using the
Addsol utility of AutoDock. The grid maps representing the
proteins in the actual docking process were calculated with
AutoGrid. The grids (one for each atom type in the ligand, plus
one for electrostatic interactions) were chosen to be sufficiently
large to include not only the active site but also significant
portions of the surrounding surface. The dimensions of the
grids were thus 50 Å × 40 Å × 40 Å, with a spacing of 0.375
Å between the grid points.
Molecular Dynamics Simulations. Refinement of the
ligand/receptor bound complex was achieved by in vacuo
energy minimization with the SANDER module of AMBER
4.1 (50 000 steps; distance dependent dielectric function of ꢀ
) 4r), applying an energy penalty force constant of 5 kcal/mol
on the protein backbone atoms. The geometry-optimized
complexes were then used as the starting point for subsequent
150 ps MD simulation, during which the protein backbone
atoms were constrained as done in the previous step. The
simulations employed the Cornell force field,39 as implemented
in the AMBER 4.1 suite of programs. The additional param-
eters required for the ligands were derived by analogy to
existing parameters. Partial atomic charges for the ligands
were imported from the output files of AM1 full geometry
optimizations as implemented in the MOPAC 6.0 program.40
A time step of 1 fs and a nonbonded pairlist updated every 25
fs were used for the MD simulations. The temperature was
mantained at 300 K using the Berendsen algorithm73 with a
0.2 ps coupling constant. An average structure was calculated
from the last 100 ps trajectory and energy-minimized using
the steepest descent and conjugate gradient methods as
specified above. The MD trajectories were analyzed by means
of the CARNAL module of AMBER package.
Acknowledgment. This project was supported by
a grant from University of Camerino and by the Italian
MIUR funds (PRIN 2002). The work was presented in
part at the 2003 meeting “Purinergic Receptors-Medici-
nal Chemistry and Pharmacology of Purinergic Recep-
tors” (Camerino, Italy). We thank M. Brandi, F. Lupidi,
G. Rafaiani, and M. Ricciutelli for technical assistance.
Supporting Information Available: Dose-response
curves for the inhibition of forskolin-stimulated adenylyl
cyclase in rat cortical membranes induced by compounds 3, 4,
8, 25 and 26 (Figure 1), plots of the monitored distance
between the key residues of 2′-Me-CCPA and 3′-Me-CCPA and
the amino acids Asn254, His278 and Thr91 along the complete
MD trajectory (Figure 2), and elemental analytical data for
all new compounds. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) 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.
(2) Fredholm, B. B. Adenosine receptors as targets for drug develop-
ment. Drug News Perspect. 2003, 16, 283-289.
(3) Saloranta, C.; Franssila, K.; Ekstrand, A.; Taskinen, M. R.;
Groop, L. Modulation of hepatic glucose production by non-
esterified fatty acids in type 2 (noninsulin-dependent) diabetes
mellitus. Diabetologia 1991, 34, 409-415.

Ribose-Modified Adenosine Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5 1561
(4) Beukers, M. W.: Chang, L. C.; von Frijtag Drabbe Kunzel, J.
K.; Mulder-Krieger, T.; Spanjersberg, R. F.; Brussee, J.; IJzer-
man, A. P. New, nonadenosine, high-potency agonists for the
human adenosine A2B receptor with an improved selectivity
profile compared to the reference agonist N-ethylcarboxamido-
adenosine. J. Med. Chem. 2004, 47, 3707-3709.
(5) Franchetti, P.; Cappellacci, L.; Marchetti, S.; Trincavelli, L.;
Martini, C.; Mazzoni, M. R.; Lucacchini, A.; Grifantini, M. 2′-
C-Methyl analogues of selective adenosine receptor agonists:
synthesis and binding studies. J. Med. Chem. 1998, 41, 1708-
1715.
(6) Cappellacci, L.; Barboni, G.;Palmieri, M.; Pasqualini, M.; Gri-
fantini, M.; Costa, B.; Martini, C.; Franchetti, P. Ribose-modified
nucleosides as ligands for adenosine receptors: synthesis, con-
formational analysis and biological evaluation of 1′-C-methyl
adenosine analogues. J. Med. Chem. 2002, 45, 1196-1202.
(7) Lohse, M. J.; Klotz, K.-N., Schwabe, U.; Cristalli, G.; Vittori, S.;
Grifantini, M. 2-Chloro-N6-cyclopentyladenosine: a highly selec-
tive agonist at A1 adenosine receptors. Naunyn Schmiedeberg’s
Arch. Pharmacol. 1988, 337, 687-689.
(8) (a) Ralevic, V.; Burnstock, G. Receptors for purines and pyrim-
idines. Pharmacol. Rev. 1998, 50, 413-492. (b) Kim, H. O.; Ji,
X.-D.; Siddiqi, S. M.; Olah, M. E.; Stiles, G. L.; Jacobson, K. A.
2-Substitution of N6-benzyladenosine-5′-uronamides enhances
selectivity for A3 adenosine receptors. J. Med. Chem. 1994, 37,
3614-3621.
(9) Franchetti, P.; Cappellacci, L,; Pasqualini, M.; Petrelli, R.; Vita,
P.; Jayaram, H. N.; Horvath, Z.; Szekeres, T.; Grifantini, M.
Manuscript in preparation.
(10) Epp, J. B. and Widlanski, T. S. Facile preparation of nucleoside-
5′-carboxylic acids. J. Org. Chem. 1999, 64, 293-295.
(11) Nutt, R. F.; Dickinson, M. J.; Holly, F. W.; Walton, E. Branched-
chain sugar nucleosides. III. 3′-C-Methyladenosine. J. Org.
Chem. 1968, 33, 1789-1795.
(12) Rosemeyer, H.; Toth, G.; Golankiewicz, B.; Kazimierczuk, Z.;
Bourgeois, W.; Kretschmer, U.; Muth, H.-P.; Seela, F. Syn-anti
conformational analysis of regular and modified nucleosides by
1D 1H NOE difference spectroscopy: A simple graphical method
based on conformationally rigid molecules. J. Org. Chem. 1990,
55, 5784-5790.
(13) Dalpiaz, A.; Townsend-Nicholson, A.; Beukers, M. W.; Schofield,
P. R.; IJzerman, A. P. Thermodynamics of full agonist, partial
agonist, and antagonist binding to wild-type and mutant ad-
enosine A1 receptors. Biochem. Pharmacol. 1998, 56, 1437-1445.
(14) Da Settimo, F.; Primofiore, G.; Taliani, S.; Marini A. M., La
Motta, C.; Novellino, E.; Greco, G.; Lavecchia, A.; Trincavelli,
L.; Martini, C. 3-Aryl[1,2,4]triazino[4,3-a]benzimidazol-4(10H)-
ones: a new class of selective A1 adenosine receptor antagonists.
J. Med. Chem. 2001, 44, 316-327.
(15) Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Mo-
toshima, 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.
(16) (a) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart,
W. E.; Belew, R. K.; Olson, A. J. Automated docking using a
lamarckian genetic algorithm and an empirical binding free
energy function. J. Comput. Chem. 1998, 19, 1639-1662. (b)
Goodsell, D. S.; Morris, G. M.; Olson, A. J. Automated docking
of flexible ligands: applications of AutoDock. J. Mol. Recognit.
1996, 9, 1-5.
(17) Gao, Z.-G.; Chen, A.; Barak, D.; Kim, S.-K.; Mu¨ller, C. E.;
Jacobson, K. A. Identification by site-directed mutagenesis of
residues involved in ligand recognition and activation of the
human A3 adenosine receptor. J. Biol. Chem. 2002, 277, 19056-
19063.
(18) Kim, J.; Wess, J.; van Rhee, A. M.; Scho¨neberg, T.; Jacobson, K.
A. Site-directed mutagenesis identifies residues involved in
ligand recognition in the human A2a adenosine receptor. J. Biol.
Chem. 1995, 270, 13987-13997.
(19) (a) Klotz, K.-N.; Lohse, M. J.; Schwabe, U. Chemical modification
of A1 adenosine receptors in rat brain membranes: evidence for
histidine in different domains of the ligand binding site. J. Biol.
Chem. 1988, 263, 17522-17526. (b) Olah, M. E.; Ren H.;
Ostrowski, J.; Jacobson, 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.
(20) Jiang, Q.; van Rhee, M.; Kim. J.; Yehle, S.; Wess, J.; Jacobson,
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.
(22) Rivkees, S. A.; Barbhaiya, H.; IJzerman A. P. Identification of
the adenine binding site of the human A1 adenosine receptor.
J. Biol. Chem. 1999, 274, 3617-3621.
(23) 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.
(24) (a) Javitch, J. A.; Ballesteros, J. A.; Weinstein, H.; Chen, J. A
cluster of aromatic residues in the sixth membrane spanning
segment of the dopamine D2 receptor is accessible in the binding-
site crevice. Biochemistry 1998, 37, 998-1006. (b) Marie, J.;
Richard, E.; Pruneau, D.; Paquet, J.-L.; Siatka, C.; Larguier, R.;
Ponce, C.; Vassault, P.; Groblewski, T.; Maigret, B.; Bonnafous,
J. C. Control of conformational equilibria in the human B2
bradykinin receptor: modeling of nonpeptidic ligand action and
comparison to the rhodopsin structure. J. Biol. Chem. 2001, 276,
41100-41111.
(25) Lin, S. W.; Sakmar, T. P. Specific tryptophan UV-absorbance
changes are probes of the transition of rhodopsin to its active
state. Biochemistry 1996, 35, 11149-11159.
(26) Sybyl Molecular Modelling System (version 6.9), Tripos Inc., St.
Louis, MO.
(27) Pearlman, D. A.; Case, D. A.; Caldwell, J. W.; Ross, W. S.;
Cheatham T. E., III; Debolt, S.; Ferguson, D. M.; Seibel, G. L.;
Kollman, P. A. AMBER, a package of computer programs for
applying molecular mechanics, normal-mode analysis, molecular
dynamics and free energy calculations to simulate the structural
and energetic properties of molecules. Comput. Phys. Commun.
1995, 91, 1-41.
(28) Pearlman, D. A.; Case, D. A.; Caldwell, J. W.; Ross, W. S.;
Cheatham T. E., III; Ferguson, D. M.; Seibel, G.; Singh, U. C.;
Weiner, P. K.; Kollman, P. A. AMBER 4.1; Department of
Pharmaceutical Chemistry, University of California: San Fran-
cisco, CA, 1995.
(29) Horn, F.; Weare, J.; Beukers, M. W.; Horsch, S.; Bairoch, A.;
Chen, W.; Edvardsen, O.; Campagne, F. G., V. GPCRDB: an
information system for G protein-coupled receptors. Nucleic
Acids Res. 1998, 26, 275-279.
(30) (a) Baldwin, J. M. Structure and function of receptors coupled
to G proteins. Curr. Opin. Cell. Biol. 1994, 6, 180-190. (b)
Schwartz, T. W. Locating ligand-binding sites in 7TM receptors
by protein engineering. Curr. Opin. Biotechnol. 1994, 5, 434-
444. (c) van Rhee, A. M.; Jacobson, K. A. Molecular archi-
tecture of G protein-coupled receptors. Drug Dev. Res. 1996, 37,
1-38.
(31) (a) Xu, W.; Li, J.; Chen, C.; Huang, P.; Weinstein, H.; Javitch,
J. A.; Shi, L.; de Riel, J. K.; Liu-Chen, L.-Y. Comparison of the
amino acid residues in the sixth transmembrane domains
accessible in the binding-site crevices of mu, delta, and kappa
opioid receptors. Biochemistry 2001, 40, 8018-8029. (b) Shi, L.;
Simpson, M. M.; Ballesteros, J. A.; Javitch, J. A. The first
transmembrane segment of the dopamine D2 receptor: acces-
sibility in the binding site crevice and position in the trans-
membrane bundle. Biochemistry 2001, 40, 12339-12348.
(32) (a) Perlman, J. H.; Colson, A. O.; Wang, W.; Bence, K.; Osman
R.; Gershengorn, M. C. Interactions between conserved residues
in transmembrane helices 1, 2, and 7 of the thyrotropin-releasing
hormone receptor. J. Biol. Chem. 1997, 272, 11937-11942. (b)
Sealfon, S. C.; Chi, L.; Eversole, B. J.; Rodic, V.; Zhang, D.;
Ballesteros, J. A.; Weinstein, H. Related contribution of specific
helix 2 and 7 residues to conformational activation of the
serotonin 5-HT2a receptor. J. Biol. Chem. 1995, 270, 16683-
16688. (c) Zhou, W.; Flanagan, C.; Ballesteros, J.; Konvicka, K.;
Davidson, J. S.; Weinstein, H.; Millar, R. P.; Sealfon, S. C. A
reciprocal mutation supports helix 2 and helix 7 proximity in
the gonadotropin-releasing hormone receptor. Mol. Pharmacol.
1994, 45, 165-170.
(33) Zhang, D.; Weinstein, H. Polarity conserved positions in trans-
membrane domains of G-protein coupled receptors and bacte-
riorhodopsin. FEBS Lett. 1994, 337, 207-212.
(34) Ryckaert, J. P.; Ciccotti, G.; Berendsen, H. J. C. Numerical
integration of the Cartesian equations of motion for a system
with constraints: molecular dynamics of n-alkanes. J. Comput.
Phys. 1977, 23, 327-333.
(35) Laskowski, R. A.; McArthur, M. W.; Moss, D. S.; Thornton, J.
M. PROCHECK: a program to check the stereochemical quality
of protein structures. J. Appl. Crystallogr. 1993, 26, 283-291.
(36) Allen, F. H.; Bellard, S.; Brice, M. D.; Cartwright, B. A.;
Doubleday, A.; Higgs, H.; Hummelink, T.; Hummelink-Peters,
B. G.; Kennard, O.; Motherwell, W. D. S. The cambridge
crystallographic data center: computer-based search, retrieval,
analysis and display of information. Acta Crystallogr. 1979, B35,
2331-2339.
(21) Jacobson, 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.
(37) Head, J.; Zerner, M. C. A Broyden-Fletcher-Goldfarb-Shannon
optimization procedure for molecular geometries. Chem. Phys.
Lett. 1985, 122, 264-274.

1562 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 5
Cappellacci et al.
(38) Gasteiger, J.; Marsili, M. Iterative partial equilization of orbital
electronegativity - a rapid access to atomic charges. Tetrahedron
1980, 36, 3219-3228.
(39) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K.
M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.;
Kollman, P. A. A second generation force field for the simulation
of proteins, nucleic acids, and organic molecules. J. Am. Chem.
Soc. 1995, 117, 5179-5197.
(40) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P.
AM1: a new general purpose quantum mechanical molecular
model. J. Am. Chem. Soc. 1985, 107, 3902-3909. (b) MOPAC
(version 6.0) is available from Quantum Chemistry Program
Exchange, No. 455.
(41) Cheng, Y.; Prusoff, W. H. Relationship between the inhibition
constant (Ki) and the concentration of inhibitor which causes
50% inhibition (IC50) of an enzymatic reaction. Biochem. Phar-
macol. 1973, 22, 3099-3108.
(42) Klotz, K.-N.; Hessling, J.; Hegler, J.; Owman, B.; Kull, B.;
Fredholm, B. B.; Lohse, M. J. Comparative pharmacology of
human adenosine subtypes characterization of stably transfected
receptors in CHO cells. Naunyn-Schmiedeberg’s Arch. Pharma-
col. 1998, 357, 1-7.
JM049408N