Purine and deazapurine nucleosides: Synthetic approaches, molecular modelling and biological activity

Article abstract of DOI:10.1016/S0014-827X(03)00019-3

A number of ligands for the adenosine binding sites has been obtained by using nucleoside convergent and divergent synthesis. Most of our nucleosides have been synthesized by coupling 2,6-dichloropurine (1), 2,6-dichloro-1-deazapurine (2), 2,6-dichloro-3-

Full text of DOI:10.1016/S0014-827X(03)00019-3

Il Farmaco 58 (2003) 193ꢀ204
/
www.elsevier.com/locate/farmac
Purine and deazapurine nucleosides: synthetic approaches, molecular
modelling and biological activity
Gloria Cristalli *, Stefano Costanzi, Catia Lambertucci, Sara Taffi, Sauro Vittori,
Rosaria Volpini
Dipartimento di Scienze Chimiche, Universita` di Camerino, via S. Agostino 1, 62032 Camerino, MC, Italy
Accepted 4 July 2002
Abstract
A number of ligands for the adenosine binding sites has been obtained by using nucleoside convergent and divergent synthesis.
Most of our nucleosides have been synthesized by coupling 2,6-dichloropurine (1), 2,6-dichloro-1-deazapurine (2), 2,6-dichloro-3-
deazapurine (3) with ribose, 2- and 3-deoxyribose and 2,3-dideoxyribose derivatives. The use of these versatile synthons allowed the
introduction of various substituents in 2- and/or 6-positions. The glycosylation site and the anomeric configuration of the obtained
nucleosides were assigned on the basis of spectroscopic studies and confirmed by molecular models. A series of potent adenosine
receptor ligands has been obtained by using divergent approaches, mostly starting from guanosine. Substitutions in 2, 6, 8, and 5?
position of adenosine molecule led to ligands selective for the different adenosine receptor subtypes. Furthermore, we investigated
the molecular bases of the different behavior of 2- and 8-alkynyl adenosines, by means of NMR experiments and molecular
modeling studies. With docking experiments, we demonstrated that the two class of molecules should have different binding modes
that explain their different degree of affinity and the shift of their activity from agonistic (2-substituted derivatives) to antagonistic
(8-substituted derivatives).
´
# 2003 Editions scientifiques et me´dicales Elsevier SAS. All rights reserved.
Keywords: Adenosine receptor ligands; Adenosine receptor molecular modeling; Glycosylation reactions; Nucleoside synthesis
1. Introduction
2. Convergent syntheses
Purine and deazapurine nucleosides are of great
interest both from chemical and pharmacological point
of view. In particular, compounds structurally related to
the intracellular modulator adenosine have been shown
to possess biological activity as adenosine receptor
ligands [1], antiviral and antitumor agents [2], and
enzyme inhibitors [3].
Our group has been involved since many years in this
research field and we have extensively described cou-
pling of substituted purines and deazapurines with a
variety of sugars. An attempt to resume such synthetic
work is shown in Fig. 1.
Most of our nucleosides have been synthesized by
coupling 2,6-dichloropurine (1), 2,6-dichloro-1-deaza-
purine (2), 2,6-dichloro-3-deazapurine (3) with ribose, 2-
and 3-deoxyribose and 2,3-dideoxyribose derivatives.
The use of these versatile synthons allowed the intro-
duction of various substituents in 2- and/or 6-positions:
in particular, in the case of the less reactive 1- and 3-
deazanucleosides, nucleophilic substitutions in 6-posi-
tion can be achieved under milder conditions when a
chlorine atom is present in 2-position [4ꢀ9].
/
For the synthesis of hydroxylamine derivatives of
purines and 1-deazapurines two additional synthons
have been used, 6-nitro-1-deazapurine (7-nitro-3H-imi-
dazo[4,5-b]pyridine, 4) and 2-chloro-6-nitro-1-deaza-
purine (5-chloro-7-nitro-3H-imidazo[4,5-b]pyridine, 5),
which were prepared starting from the common inter-
mediate 6-nitro-1-deazapurine-3-oxide (7-nitro-3H-imi-
dazo[4,5-b]pyridine-4-oxide) [10].
* Correspondence and reprints.
E-mail address: gloria.cristalli@unicam.it (G. Cristalli).
´
0014-827X/03/$ - see front matter # 2003 Editions scientifiques et me´dicales Elsevier SAS. All rights reserved.
doi:10.1016/S0014-827X(03)00019-3

194
G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ204
/
Fig. 1. Coupling of substituted purines and deazapurines with various sugars.
For the synthesis of 1,7-dideazapurine derivatives 6-
been performed to assign the correct configuration to
the resulting nucleosides.
nitro-1,7-dideazapurine (4-nitro-1H-pyrrolo [2,3-b]py-
ridine, 6) was utilized [11]. All these synthons are
reported in Fig. 2.
Depending upon coupling conditions as well as base
and sugar structure and reactivity, different anomeric
and isomeric mixtures have been obtained. Extensive
studies, utilizing chemical and physical methods, have
We report here on the coupling of 2,6-dichloropurine
and 1- and 3-deazapurine with 3-deoxyribose, which has
been recently and extensively investigated in our labora-
tory. These coupling reactions have been carried out
both by fusion method and by acidic catalysis. In the
fusion method, reported below, coupling was performed
Fig. 2. Substituted purines and deazapurines.

G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ
/
204
195
in the presence of a catalytic amount of p-toluensul-
phonic acid at 160 8C in vacuo for 10 min.
Coupling of 2,6-dichloropurine with 1,2-di-O-diace-
were obtained (51, 23, 3 and 13% yield, respectively)
after purification and the structure reported in Scheme 3
was assigned to compounds 16, 17, 18 and 19. Accord-
ingly, the four nucleosides show the same glycosylation
site and correspond to beta and alfa ribofuranose
derivatives and beta and alfa arabinofuranose deriva-
tives, respectively [7].
tyl-5-O-benzoyl-3-deoxy-b-
D-ribofuranose (7) gave a
mixture of beta (8) and alfa (9) anomers in one to one
ratio and in a 65% total yield. Deblocking with liquid
ammonia at room temperature both removed the
protections and introduced the amino group in 6-
position to yield compounds 10 and 11. Other amines
can be introduced in 6 position at this stage by using the
suitable reagent. However, it no possible to obtain
deprotection without substitution in 6-position (Scheme
1) [12].
The glycosylation site and the anomeric configuration
of these nucleosides were assigned on the basis of
spectroscopic studies. The UV spectra of the four
nucleosides are practically identical showing a max-
imum of absorption at 259 nm with an extinction
coefficient of about 13 000; these findings suggested
that the glycosylation site was the same for all the
obtained compounds.
Coupling of 2,6-dichloro-1-deazapurine (5,7-dichloro-
3H-imidazo[4,5-b]pyridine (2) with the same sugar 7
gave again a mixture of beta (12) and alfa (13) anomers
in 2 to 1 ratio and in a total yield of 81%. In this case,
deblocking with methanolic ammonia at room tempera-
ture led to the 2,6-dichloro deblocked nucleosides 14
and 15 in 85% total yield. In fact, differently from
purines, with 1-deazapurines it is possible to obtain
deprotection without substitution in 6-position (Scheme
2) [8].
1H NMR spectra and NOE studies, performed by
irradiating different protons of the sugars, demonstrated
unequivocally the structures of the four nucleosides. In
1
fact, studies of 1D H NOE differential spectroscopy
established unequivocally N(9) as glycosylation site in
the case of all compounds 16ꢀ
ingly, saturation of HÃC(1?) resulted in a NOE at the
HÃC(8) and HÃC(3) signal. These data excluded with-
/19 (Table 1a). Accord-
/
/
/
Coupling of 2,6-dichloro-3-deazapurine (4,6-dichloro-
1H-imidazo[4,5-c]pyridine, 3) with 1,2-di-O-diacetyl-5-
out any doubt the presence of a possible glycosylation
occurred at N(7). In this case, in fact, no NOE effect at
O-benzoyl-3-deoxy-b-
D-ribofuranose (7) gave a mixture
the HÃ
The epimerization at C(2?) of nucleosides 18 and 19
has been demonstrated by the NOE observed at the HÃ
C(3?b) signal upon saturation of both HÃC(2?) and HÃ
/
C(3) signal would have been found.
of four nucleosides in a 90% total yield. By reacting this
mixture with methanolic ammonia for 24 h at room
temperature the corresponding deprotected derivatives
/
/
/
Scheme 1.

196
G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ204
/
Scheme 2.
Scheme 3.

G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ
/
204
197
Table 1
Comp.
Irradiated proton Observed NOE (%)
(a) NOE (%) data of compounds 16, 17, 18 and 19 in (D6 ꢀ
/
DMSO) at 238
C(7) (8.9), HÃC(2?ꢂ
C(7) (6.7), HÃC(2?) (6.8), HÃ
C(7) (4.0), HÃC(2?) (3.0)
C(1?) (4.4), HÃC(3?b) (1.8)
C(3?b) (2.8)
C(2) (4.2), HÃ
C(2) (3.0), HÃ
C(2) (1.7), HÃ
16
17
18
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
16
/
C(1?)
C(1?)
C(1?)
C(2?)
C(4?)
C(1?)
C(2?)
C(4?)
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
/
C(2) (4.6), HÃ
C(2) (4.2), HÃ
C(2) (3.2), HÃ
/
/
/
4?) (6.0)
/
/
/
/
/
C(3?a) (1.0)
/
/
/
/
/
/
/
/
/
19
/
/
/
C(7) (6.5), HÃ
C(7) (3.0), HÃ
C(7) (1.4), HÃ
/
C(2?) (2.0), HÃ
/C(3?a) (0.9)
/
/
/
/
C(1?) (2.1), HÃ
/C(3?b) (3.1)
/
/
/
/C(3?b) (2.9)
Comp.
17
18
19
(b) Distances of protons from HÃ
/
C(1?) and relative NOE effect upon HÃC(1?) saturation
/
˚
A
˚
˚
NOE
a
6.0
A
2.359 6.8
NOE
A
NOE Ang NOE
3.057 2
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
HÃ
/
C(2?)
3.049
4.152
3.535
3.486
2.351
4.169
3.809
3.627
3.979
2.308
3
/
C(3?a)
C(3?b)
C(4?)
C(2) (anti) 3.995
C(7) 2.320
C(2) (syn) 2.611
C(7) 4.206
3.453
4.153
3.856
3.973
1
3.436 0.9
4.167
3.844
/
a
/
6.0
/
4.006
/
8.9
4.6
2.337 6.7
2.589 4.2
4.222
4
2.281 6.5
2.634 4.2
4.181
/
2.581 3.2
4.212
/
a
HÃ
/
C(2?)ꢂ
/
HÃC(4?).
/
C(4?). This result confirmed that the three protons are
located on the same side of the tetrahydrofuryl ring.
in Table 1b, the distances from the HÃ
/
C(1?) calculated
for the protons of the four nucleosides are in very good
agreement with the experimental data.
Irradiation of HÃ
the HÃC(2?) and the HÃ
configuration. On the other hand, saturation of HÃ
of compound 18, resulted in a NOE at the HÃ
signal but not at the one of HÃ
effect, even though is rather strange for a beta nucleo-
side, could be explained by the distortion that the
arabinose introduces in the nucleoside because of the
/
C(1?) of compound 19 gave NOE at
The other convergent approach was the coupling
reaction carried out in acetonitrile and using stannic
chloride or TMS triflate as acidic catalyst. In this case,
only the beta anomer was obtained with all the three
bases.
Couplings of the same bases with (cyclo)alkyl halides
have been also carried out, aimed at obtaining adeno-
sine receptor antagonists [14,15] and enzyme inhibitors
[3].
/
/
C(3?a) signals, proving the alfa
/
C(1?)
/
C(2?)
/
C(4?). The lack of this
position of the OHÃC(2?).
/
On the bases of 1D 1H NOE obtained, molecular
models were constructed by utilizing the program
HYPERCHEM [13]. The structures were optimized by
using the semiempirical method PM3 with the mini-
mization algorithm Eigenvector Following until a gra-
3. Synthesis and biological evaluation of adenosine
receptor agonists obtained by using divergent approaches
dient RMS of 10ꢁ3 kcal/A mol was reached. As shown
˚
Other series of adenosine receptor ligands have been
obtained by using divergent approaches. Alkynyl chains
were introduced in 2- and 8-position of adenosine (20)
and 5?-N-ethylcarboxamidoadenosine (NECA, 21) by
multi-step charts starting from guanosine (22) or 8-
bromoadenosine (23), respectively [16] Fig. 3.
Combination with substitutions in 6-position yielded
agonists potent and selective for the various adenosine
receptor subtypes. Adenosine is in fact a signaling
molecule that, among its various biological functions,
can interact with four specific membrane receptors
belonging to the purinergic P1 family and named A1,
A2A, A2B and A3 [17]. All these subtypes belong to the
super-family of 7 trans-membrane domains G-protein
coupled receptors (7TMs GPCR) and are classified by
Fig. 3. Adenosine and NECA.

198
G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ204
/
the GPCR database [18] in the family of rhodopsine like
receptors (Class A of GPCR). Among human adenosine
receptors, the A3 subtype was the last one to be
identified, and, unlike what happened for A1, A2A and
A2B, it was discovered by means of molecular biology
techniques followed by pharmacological experiments.
The biological functions of the A3 subtype are still not
very well understood. This is mainly due to the lack of
truly selective ligands for in vivo studies and to the poor
structural characterization of the receptor itself. In fact
just few mutagenesis studies on this subtype have been
Starting from commercially available guanosine very
versatile synthons functionalized in three different posi-
tions can be obtained. In Scheme 4 the approach to
obtain 2,6-disubstituted adenosines is reported. Guano-
sine (22) is converted in three steps in 6-chloro-2-iodo
acetylated adenosine (24) [30], from which the two
substitutions can be carried out.
Other disubstituted adenosines can be obtained ac-
cording to the procedure depicted in Scheme 5. In this
case the starting material 2-iodoadenosine (25) comes
from the previous method and is transformed in four
steps in 2-iodo-5?-ethylesteradenosine 33, from which 2-
carried out [19ꢀ21].
/
There are reports that agonists acting via the adenosine
A3 receptor have cardioprotective effects [22,23] although
it was proven that the A3 receptor serves a different
cardioprotective function as compared with the adeno-
sine A1 receptor [24]. Chronic administration of adeno-
sine A3 receptor agonists has also been shown to have
protective actions in the brain, in contrast to acute high-
dose administration, which can cause toxicity [25,26].
In recent years we demonstrated that C-2-substituted
NECA are potent and selective ligands for adenosine
receptor subtypes [27,28]. In particular compounds
bearing a hexynyl, phenylethynyl, and phenylhydroxy-
propynyl chains at the 2-position of NECA possess high
affinity for A3 receptors combined, in same cases, with
good selectivity [15]. In order to investigate the role of
the carboxamido group at the 5?-position and to
simplify the structure of these molecules, adenosine
derivatives bearing the above mentioned alkynyl chains
at the C-2-position were synthesized [29].
alkynylNECAs 35ꢀ37 have been obtained [27].
/
Trisubstituted adenosines can be obtained by using 2-
iodoNECA (34) as the starting material. Diazotation of
the 6-amino group, followed by nucleophilic substitu-
tion of the 6-iodine atom gave compounds of general
structure 39, which can be additionally modified in 2-
position by using the cross coupling reaction or nu-
cleophylic substitutions to obtain compounds of general
formula 40 (Scheme 6) [31,32].
Recently, the four human receptor subtypes have been
transfected and overexpressed on CHO cells. Whereas
radioligand binding studies have been utilized for A1,
A2A and A3 receptors, the relative potencies of agonists
for the A2B adenosine receptor have been tested by
measuring the activity of receptor-stimulated adenylyl
cyclase [33,34]. We are currently utilizing these tran-
fected CHO cells for the development of new ligands
and the affinity and potency at the four subtypes of the
Scheme 4.

G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ
/
204
199
Scheme 5.
selected 2-alkynyl derivatives of adenosine and NECA
are reported in Table 2.
high affinity and selectivity for A3 receptors, whereas the
2-phenylhydroxypropynyl derivative 29 is more potent
but less selective for all this three subtypes.
Modification of the sugar by introducing a N-
ethylcarboxyamido group enhanced the affinity of all
compounds and the selectivity of the phenethynyl
derivative 36 at the A3 subtype.
From these data it is possible to note that the
introduction of alkynyl chains on the adenosine structure
led to compounds with high affinity and different degree
of selectivity on A1, A2A, and A3 subtypes. In particular,
compound 28, bearing a phenylethynyl group, presents
Scheme 6.

200
G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ204
/
Table 2
Adenosine derivatives affinities in radioligand binding assays at human A1, A2A and A3 receptors, and in adenylyl cyclase assays at human A2B
subtype
Comp.
R
R1
Ki or EC50 (nM)
a
b
c
d
Ki (A1)
KI (A2A
)
EC50 (A2B
100 000
)
KI (A3)
27 2-HEAdo
28 2-PEAdo
29 2-PHPAdo
30
CH3Ã
/
(CH2)3
H
18
5.7
4.7
16
C6H5
H
391
0.67
327
1690
8.4
363
7.0
ꢀ/
100 000
C6H5Ã
CH3Ã(CH2)3
C6H5
C6H5Ã
/CH(OH)
H
2400
3.3
1.1
/
CH3
CH3
CH3
1230
8530
273
6.4
100 000
100 000
2400
31
3.4
32
/CH(OH)
0.76
2.4
35 HENECA
36 PENECA
37 PHPNECA
41 8-HEAdo
42 8-PEAdo
43 8-PKPAdo
60
6100
560
2.7
620
3.1
ꢀ
1100
/
100 000
6.2
0.42
650
790
9830
CH3Ã
C6H5Ã
C6H5Ã
/
(CH2)3Ã
/
Å
/
Ã
/
ꢀ
/
100 000
100 000
100 000
ꢀ/100 000
ꢀ
/
100 000
100 000
100 000
/
Å
/
Ã
/
ꢀ
/
ꢀ/100 000
ꢀ
/
/
C(O)Ã
/
Ä
/
Ã
/
ꢀ
/
46 600
ꢀ/
a
Displacement of specific [3H]CCPA binding in CHO cells, stably transfected with human recombinant A1 adenosine receptor, expressed as Ki
(nM).
b
Displacement of specific [3H]NECA binding in CHO cells, stably transfected with human recombinant A2A adenosine receptor, expressed as Ki
(nM).
c
Measurement of receptor-stimulated adenylyl cyclase activity in CHO cells, stably transfected with human recombinant A2B adenosine receptor,
expressed as EC50 (nM).
d
Displacement of specific [3H]NECA binding in CHO cells, stably transfected with human recombinant A3 adenosine receptor, expressed as Ki
(nM).
But the most relevant results have been obtained by
introducing on the adenosine structure of a small
substituent on the 6-amino group. In fact in this case
both selectivity and affinity of the phenylethynyl deri-
vative 31 for the A3 subtype has been improved. On this
basis, an application of the trisubstitution chemical
approach, described in Scheme 6, can be depicted. In
fact, the data reported in Table 2 demonstrated that the
2-alkynyl derivatives of adenosine possess high affinity
for all AdoR subtypes.
The presence of a 5?-ethylcarboxamido group en-
hanced A3 affinity of these molecules.
The presence of a N6-methyl group in 2-alkynyl
adenosines greatly improved A3 selectivity.
Hence, combination of these three substitutions should
be carried out to obtain molecules which should possess
higher affinity and selectivity for the A3 receptor subtype.
Furthermore, we wanted to investigate the effect of
the same alkynyl chains in a different position of the
purine ring, and so we moved them from the 2- to the 8-
position of the molecule, also because this kind of
substitution was not too much investigated in the past.
Hence, 8-substituted adenosines 41ꢀ43 were synthesized
/
by cross-coupling reaction starting from 8-Br-adenosine
(Scheme 7) [35,36].
Unfortunately, the reaction with the phenylhydrox-
ypropynyl chain did not provide the desired compound
but a product deriving from the tautomeric rearrange-
ment of the side-chain which ended up being a
phenylketopropene (43).
The results of the binding studies carried out with the
8-substituted compounds (Table 2) show that the
introduction of an alkynyl chain in C-8 position is
very disadvantageous for the affinity and potency at A1,
A
2A and A2B receptors, while is more tolerated by the A3
subtype. The new 8-alkynyl derivatives, in fact, are
endowed with an affinity for A3 receptors in the high
nanomolar range.

G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ
/
204
201
Scheme 7.
Furthermore, the maximal induction of guanosine 5?-
It is well known, in fact, that nucleosides are endowed
with free rotation around the glycosidic bond and that
they generally have two minima of energy, one in syn
and the other in anti conformation. The conformational
preference of these molecules, obviously, is affected by
the surroundings, so different conformations can be
found in solid state, in solution, in vacuo or upon
binding to a protein.
It has been demonstrated by Bruns, on the bases of a
detailed SAR analysis of adenosine receptor agonists
that the binding of adenosine must occur in anti
conformation [37] and this data was later confirmed
by a number of mutagenesis and molecular modeling
experiments.
(g-thio)triphosphate ([35S]GTPgS) binding to G proteins
and the inhibition of NECA-stimulated binding, in
membranes of CHO cells, which express the human A3
receptor, were used to determine the intrinsic activity of
these nucleosides at the human A3 adenosine receptor.
The results showed that these new adenosine derivatives
are very selective ligands for the A3 receptor subtype
and behave as adenosine antagonists, since they do not
stimulate basal [35S]GTPgS binding, but inhibit NECA-
stimulated binding, as shown in Fig. 4. This is the first
report that adenosine derivatives, with unmodified
ribose moiety, could behave as adenosine receptor
antagonists.
On the other hand, Stolarski and Dudycz demon-
strated that the chemical shift of H2? in purine nucleo-
sides can be used as an indicator of the sugar-base
orientation. Typical D values for H2? in DMSO, in fact,
are 5.2 ppm for compounds restricted in syn conforma-
tion and 4.2 ppm for those in anti [38,39]. This big
difference is explained by the influence that the nitrogen
in 3 position can exerts on H2? when the nucleoside is in
syn conformation.
4. NMR and molecular modelling studies
Since 2- and 8- alkynyl adenosines behave in a so
different way we wanted to investigate their conforma-
tion about the glycosidic bond.
Since the syn ꢀanti equilibrium is very rapid, the H2?
/
chemical shift of a given compound will have an average
value determined by the statistical distribution of the
various conformations. Adenosine, for example, shows
an absolutely intermediate value of 4.62 ppm.
The 2-alkynyl derivatives 27ꢀ29 presented slightly
/
lower chemical shifts, which indicate a moderate pre-
Table 3
H2?chemical shift of 2- and 8-substituted adenosines in DMSO
solution
Comp.
H2? chemical shift
Adenosine (20)
2-HEAdo (27)
2-PEAdo (28)
4.62
4.57
4.54
4.52
5.04
5.03
4.89
2-(R,S)-PHPAdo (29)
8-HEAdo (41)
8-PEAdo (42)
Fig. 4. Percentage of stimulation of [35S]GTPgS binding (gray bars)
and percentage of inhibition of NECA-stimulated [35S]GTPgS binding
8-PKPAdo (43)
(black bars) by compounds 41ꢀ/43 in membranes of CHO cells which
express the human A3 receptor.

202
G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ204
/
Fig. 5. Total energy (AM1) of 2- and 8-PEAdo in dependence of the O4?Ã
/
C4?Ã
/
C5?ÃO5? (x) torsional angle.
/
ference for the anti conformation. On the other hand,
the 8-derivatives 41ꢀ43 showed a chemical shift of about
5 ppm, which reveals a clear preference for the syn
NMR and theoretical data are in very good agreement
[40].
/
Furthermore, to investigate how this structural differ-
ences can affect the interactions with the adenosine
receptors, we tried to get an insight into the molecular
structure of the A3 subtype and its binding site using
molecular modeling techniques.
For this purpose, using the new crystal structure of
bovine rhodopsine as template, after the alignment of
the amino acid sequences, the model of the seven trans-
membrane domains of the receptor with the homology
modeling technique was constructed [41,42]. The agonist
2-phenylethynyladenosine (2-PEAdo, 28) was docked in
its anti conformation between the helices 3, 5, 6 and 7,
as showed in the Fig. 6.
conformation (Table 3).
After the NMR studies, some theoretical calculation
utilizing the semi-empirical method AM1 was per-
formed. The energetic differences, in kcal/mol, between
the syn and the anti conformation of 2- and 8-alkynyl
adenosine derivatives are reported in Fig. 5.
It is clear that, from the energy point of view, the 2-
substituted compounds prefer the anti conformation,
while the 8-substituted prefer the syn one. It is worth
to note that the torsional angle found in syn conforma-
tions is always about ꢁ
/
1008, and so the nitrogen in 3
position results to be very close to the H2?. Hence, the
Fig. 6. The agonist 2-PEAdo docked into the seven trans-membrane domain of the human A3 receptor. View from outside the cell (left) and details
of the active site (right).

G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ
/
204
203
Fig. 7. Superimposition of 2- and 8-PEAdo (left) with the C2 of a molecule matching the C8 of the other one and vice versa, and relative docking of
8-PEAdo into the seven trans-membrane domain of the human A3 receptor (right).
The alkynyl chain is accommodated in a hydrophobic
region between the helices 3, 4 and 5. The ribose moiety,
which is fundamental for the agonist activity, is in strict
contact with the helices 3 and 7. In particular the 5?-
hydroxyl group is within hydrogen-bond distance re-
spect to Asn274 (TM7), while the 3?-hydroxy, which has
been largely demonstrated to be required for full
agonistic activity [9,43], is closely linked to the crucial
His272 (TM7). The adenine region is found between the
helices 3 and 6, with Asn250 (TM6) hydrogen-bonded to
the amine in 6 position of the purine ring and Thr94
Acknowledgements
This work was supported by grants from the Ministry
of Research (COFIN, no 2001052834, 2001 and no
200061553, 2002) and from the Italian Research Council
(CNR, no C001044ꢀ
001).
/
References
[1] G. Cristalli, R. Volpini (ed.), Adenosine Receptors: Medicinal
Chemistry, Pharmacology and Therapeutic Applications, Current
very close to N3. The receptorꢀligand complex results to
/
Topics in Medicinal Chemistry, 3 (2003) 355ꢀ469.
/
be more stable with respect to the separated molecules of
about 60 kcal/mol.
[2] D. Salvatori, R. Volpini, S. Vincenzetti, A. Vita, S. Costanzi, C.
Lambertucci, G. Cristalli, S. Vittori, Adenine and deazaadenine
nucleoside and deoxy-nucleoside analogues: inhibition of viral
replication of sheep MVV (in vitro model for HIV) and bovine
On the contrary, 8-phenylethynyladenosine (8-
PEAdo, 42) cannot bind to the A3 receptor in the
same way of compound 28 because the alkynyl chain in
C-8 position would overlap with the helix 6. The binding
mode of this molecule, then, should be different.
We propose that the purine moieties of 2-PEAdo (28)
and 8-PEAdo (42) can be superimposed in a way that
the C2 of a molecule matches the C8 of the other one as
showed in Fig. 7. In this way there is a very good steric
and electrostatic correspondence at the level of the
purine moieties and the phenylethynyl chains, even
though there is not overlapping of the ribose moieties.
As a consequence, 8-PEAdo is not able to stimulate the
receptor response anymore, as the biological data in
Table 2 clearly revealed.
BHV-1, Bioorg. Med. Chem., 10 (2002) 2973ꢀ2980.
/
[3] G. Cristalli, S. Costanzi, C. Lambertucci, G. Lupidi, S. Vittori, R.
Volpini, E. Camaioni, Adenosine deaminase: functional implica-
tion and different classis of inhibitors, Med. Res. Rev. 21 (2001)
105ꢀ128.
/
[4] G. Cristalli, S. Vittori, A. Eleuteri, M. Grifantini, R. Volpini, G.
Lupidi, L. Capolongo, E. Pesenti, Purine and 1-deazapurine
ribonucleosides and deoxyribonucleosides: synthesis and biologi-
cal activity, J. Med. Chem. 33 (1991) 2226ꢀ2230.
/
[5] G. Cristalli, S. Vittori, A. Eleuteri, R. Volpini, E. Camaioni, G.
Lupidi, N. Mahmood, F. Bevilacqua, G. Palu, Synthesis and
biological evaluation of N6-cycloalkyl derivatives of 1-deazapur-
ine nucleosides: a new class of anti-immunodeficiency virus
agents, J. Med. Chem. 38 (1995) 4019ꢀ4025.
[6] G. Cristalli, E. Camaioni, E. Di Francesco, A. Eleuteri, S. Vittori,
R. Volpini, Potent and selective ligand for adenosine binding site,
/
Nucleosides Nucleotides 16 (1997) 1379ꢀ1388.
[7] R. Volpini, E. Camaioni, S. Costanzi, S. Vittori, G. Cristalli,
Synthesis of new 3?-deoxy ribonucleosides employing the acid-
catalyzed fusion method, Helv. Chim. Acta 81 (1998) 2326ꢀ2331.
[8] R. Volpini, E. Camaioni, S. Costanzi, S. Vittori, G. Lupidi, G.
Cristalli, Synthesis and adenosine deaminase inhibitory activity of
/
These are probably the molecular bases that make the
8-alkynyladenosines the first example of adenosine
receptor antagonist with nucleosidic structure and with
an intact ribose moiety [35].
/

204
G. Cristalli et al. / Il Farmaco 58 (2003) 193ꢀ204
/
3?-deoxy-1-deazaadenosines, Helv. Chim. Acta 82 (1999) 2212ꢀ
/
[28] G. Cristalli, E. Camaioni, S. Vittori, R. Volpini, P.A. Borea, A.
2218.
Conti, S. Dionisotti, E. Ongini, A. Monopoli, 2-Aralkynyl and 2-
heteroalkynyl derivatives of adenosine-5?-N-ethyluronamide as
selective A2a adenosine receptor agonists, J. Med. Chem. 38
[9] S. Vittori, A. Lorenzen, C. Stannek, S. Costanzi, R. Volpini, A.P.
IJerman, J.K. von Frijtag Drabble Kunzel, G. Cristalli, A1
adenosine receptor partial agonists: N-6-substituted-2?-deoxyade-
(1995) 1462ꢀ1472.
/
nosines, J. Med. Chem. 43 (2000) 250ꢀ
/
260.
[10] G. Cristalli, M. Grifantini, S. Vittori, W. Balduini, F. Cattabeni,
[29] R. Volpini, S. Costanzi, C. Lambertucci, S. Taffi, S. Vittori, K.-N.
Klotz, G. Cristalli, N6-Alkyl-2-alkynyl derivatives of adenosine as
potent and selective agonists at human adenosine A3 receptor and
a starting point searching A2B ligands, J. Med. Chem., 45 (2002)
Adenosine and 2-chloroadenosine deaza-analogues as adenosine
receptor agonists, Nucleosides Nucleotides 4 (1985) 625ꢀ639.
/
[11] G. Cristalli, S. Vittori, A. Eleuteri, R. Volpini, D. Cola, E.
Camaioni, P.V. Gariboldi, G. Lupidi, Synthesis of 1,7-dideaza-
purine ribonucleosides and deoxyribonucleosides, Nucleosides
3271ꢀ3279.
/
[30] V. Nair, S.G. Richardson, Modification of nucleic acid bases via
radical intermediates: synthesis of dihalogenated purine nucleo-
Nucleotides 12 (1993) 39ꢀ
/53.
sides, Synthesis (1982) 670ꢀ672.
/
[12] S. Vittori, S. Costanzi, C. Lambertucci, S. Taffi, R. Volpini, G.
Cristalli, Deaza and deoxy adenosine derivatives: synthesis and
inhibition of animal viruses as human infection models, Nucleos
Nucleot Nucl (in press).
[31] P.G. Baraldi, B. Cacciari, M.J. Pineda de Las Infantas, R.
Romagnoli, G. Spalluto, R. Volpini, S. Costanzi, S. Vittori, G.
Cristalli, N. Melman, K. Park, X. Ji, K.A. Jacobson, Synthesis
and biological activity of a new series of N6-arylcarbamoyl, 2-
(ar)alkynyl-N6-arylcarbamoyl, and N6-carboxamido derivatives
of adenosine-5?-N-ethyluronamide as A1 and A3 adenosine
[13] HYPERCHEM 5.02, Hypercube Inc., USA, 1997.
[14] E. Camaioni, S. Costanzi, S. Vittori, R. Volpini, K.-N. Klotz, G.
Cristalli, New substituted 9-alkylpurines as adenosine receptor
receptor agonists, J. Med. Chem. 41 (1998) 3174ꢀ3185.
/
ligands, Bioorg. Med. Chem. 6 (1998) 523ꢀ
[15] G. Cristalli, E. Camaioni, S. Costanzi, S. Vittori, R. Volpini, K.N.
Klotz, Characterization of potent ligands at human adenosine
/
533.
[32] R. Volpini, E. Camaioni, S. Costanzi, S. Vittori, K.-N. Klotz, G.
Cristalli, Synthesis of di- and tri-substituted adenosine derivatives
and their affinity at human adenosine receptor subtypes, Nucleo-
receptor, Drug Dev. Res. 45 (1998) 176ꢀ
/181.
sides Nucleotides 18 (1999) 2511ꢀ2520.
/
[16] R. Volpini, S. Costanzi, C. Lambertucci, S. Vittori, G. Cristalli,
Purine nuclosides bearing 1-alkynyl chains as adenosine receptor
[33] K.-N. Klotz, J. Hessling, J. Hegler, B. Owman, B. Kull, B.B.
Fredholm, M.J. Lohse, Comparative pharmacology of human
adenosine subtypes-characterization of stably transfected recep-
agonists, Curr. Pharm. Des., 8 (2002) 2285ꢀ
[17] V. Ralevic, G. Burnstock, Receptors for purines and pyrimidines,
Pharmacol. Rev. 50 (1998) 413ꢀ492.
/
2298.
tors in CHO cells, Naunynꢀ/Schmiedeberg’s Arch. Pharmacol.
/
357 (1998) 1ꢀ9.
/
[18] F. Horn, G. Vriend, F.E. Cohen, Collecting and harvesting
[34] K.-N. Klotz, E. Camaioni, R. Volpini, S. Kachler, S. Vittori, G.
Cristalli, 2-Alkynyl derivatives of adenosine and adenosine-5?-N-
ethyluronamide as selective agonists at A3 adenosine receptors,
biological data: the GPCRDB and NucleaRDB databases,
Nucleic Acids Res. 29 (2001) 346ꢀ349.
/
[19] T.M. Palmer, G.L. Stiles, Identification of threonine residues
controlling the agonist-dependent phosphorylation and desensiti-
zation of the rat A(3) adenosine receptor, Mol. Pharmacol. 57
Naunynꢀ
/
Schmiedeberg’s Arch. Pharmacol. 360 (1999) 103ꢀ108.
/
[35] R. Volpini, S. Costanzi, C. Lambertucci, S. Vittori, A. Lorenzen,
K.-N. Klotz, G. Cristalli, Introduction of alkynyl chains on C-8
of adenosine led to very selective antagonists of the A3 adenosine
(2000) 539ꢀ545.
/
[20] A. Chen, Z.G. Gao, D. Barak, B.T. Liang, K.A. Jacobson,
Constitutive activation of A(3) adenosine receptors by site-
directed mutagenesis, Biochem. Biophys. Res. Commun. 284
receptor, Bioorg. Med. Chem. Lett. 11 (2001) 1931ꢀ1934.
/
[36] C. Lambertucci, S. Costanzi, S. Vittori, R. Volpini, G. Cristalli,
Synthesis and adenosine receptor affinity and potency of 8-
alkynyl derivatives of adenosine, Nucleosides Nucleotides Nucleic
(2001) 596ꢀ601.
/
[21] K.A. Jacobson, Z.G. Gao, A. Chen, D. Barak, S.A. Kim, K. Lee,
A. Link, P.V. Rompaey, S. van Calenbergh, B.T. Liang,
Neoceptor concept based on molecular complementarity in
Acids 20 (2001) 1153ꢀ1157.
[37] R.F. Bruns, Adenosine receptor activation in human fibroblasts:
/
nucleoside agonists and antagonists, Can. J. Physiol. Pharmacol.
GPCRs:
enhanced affinity for amine-modified nucleosides, J. Med.
Chem. 44 (2001) 4125ꢀ4136.
[22] B.T. Liang, K.A. Jacobson, Adenosine and ischemic precondi-
tioning, Curr. Pharm. Des. 5 (1999) 1029ꢀ1041.
[23] K. Stambaugh, K.A. Jacobson, J.L. Jiang, B.T. Liang, A novel
cardioprotective function of adenosine A1 and A3 receptors
during prolonged simulated ischemia, Am. J. Physiol. 273
a
mutant adenosine A(3) receptor with selectively
58 (1980) 673ꢀ
[38] R. Stolarski, L. Dudycz, D. Shugar, NMR studies in the syn ꢀ
dynamic equilibrium in purine nucleosides and nucleotides, Eur.
J. Biochem. 108 (1980) 111ꢀ121.
[39] L. Dudycz, R. Stolarski, R. Pless, D. Shugar, A 1H NMR study of
the syn ꢀanti dynamic equilibrium in adenine nucleosides and
nucleotides with the aid of some synthetic model analogues with
fixed conformations, Z. Naturforsch., Teil. C 34C (1979) 359ꢀ373.
/
691.
/
anti
/
/
/
/
/
(1997) H501ꢀ
/
H505.
[40] S, Costanzi, C. Lambertucci, S. Vittori, R. Volpini, G. Cristalli, 2-
and 8-alkynyladenosines: conformational studies and docking tp
human adenosine A3 receptor can explain their different biologi-
[24] B.T. Liang, K.A. Jacobson, A physiological role of the adenosine
A3 receptor: sustained cardioprotection, Proc. Natl. Acad. Sci.
USA 95 (1998) 6995ꢀ/6999.
cal behavior, J. Molec. Graphics Model., 21 (2003) 253ꢀ262.
/
[25] D.K. Von Lubitz, R.C. Lin, M. Boyd, N. Bischofberger, K.A.
Jacobson, Chronic administration of adenosine A3 receptor
agonist and cerebral ischemia: neuronal and glial effects, Eur. J.
[41] K. Palczewski, T. Kumasaka, T. Hori, C.A. Behnke, H. Mo-
toshima, B.A. Fox, I. Le Trong, D.C. Teller, T. Okada, R.E.
Stenkamp, M. Yamamoto, M. Miyano, Crystal structure of
rhodopsin: AG protein-coupled receptor, Science 289 (2000)
Pharmacol. 367 (1999) 157ꢀ163.
/
[26] K.A. Jacobson, D.K. Von Lubitz, J.W. Daly, B.B. Fredholm,
Adenosine receptor ligands: differences with acute versus chronic
739ꢀ745.
/
[42] H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat, H.
treatment, Trends Pharmacol. Sci. 17 (1996) 108ꢀ
/113.
Weissig, I.N. Shindyalov, P.E. Bourne, The protein data bank,
[27] G. Cristalli, R. Volpini, S. Vittori, E. Camaioni, A. Monopoli, A.
Conti, S. Dionisotti, C. Zocchi, E. Ongini, 2-Alkynyl derivatives
of adenosine-5?-N-ethyluronamide (NECA) as selective agonists
at A2 adenosine receptor agonists with potent inhibitory activity
Nucleic Acids Res. 28 (2000) 235ꢀ242.
/
[43] E.M. van der Wenden, J.K. Von Frijtag Drabbe Kunzel, R.A.
Mathot, M. Danhof, A.P. IJzerman, W. Soudijn, Ribose-modified
adenosine analogues as potential partial agonists for the adeno-
on platelet aggregation, J. Med. Chem. 37 (1994) 1720ꢀ
/
1726.
sine receptor, J. Med. Chem. 38 (1995) 4000ꢀ4006.
/