Synthesis, adenosine receptor binding and 3D-QSAR of 4-substituted 2-(2′-furyl)-1,2,4-triazolo[1,5-a]quinoxalines

Article abstract of DOI:10.1016/j.bmc.2007.10.103

A collection of 25 2-(2′-furyl)-1,2,4-triazolo[1,5-a]quinoxalines incorporating different substitution patterns at position 4 have been synthesized and their binding affinity towards human adenosine receptors (hA₁, hA_2A, hA_2B and hA₃) was determinated. The biological data show that several potent at hA₁, but lightly selective, adenosine ligands were identified. Moreover, these results confirmed the hypothesis that the structural modifications carried out on the 4-position of the tricyclic system produces a remarkable modification of the adenosine receptorial profile. A 3D-QSAR modelling study (GRIND/ALMOND methodology) performed on the hA₁ data gave further support to the pharmacological results, and it is presented as a useful tool for the future design of ligands with better pharmacological profiles.

Full text of DOI:10.1016/j.bmc.2007.10.103

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 2103–2113
Synthesis, adenosine receptor binding and 3D-QSAR
of 4-substituted 2-(20-furyl)-1,2,4-triazolo[1,5-a]quinoxalines
a
b
b
a,
*
´
´
´
´
Ana Martınez, Hugo Gutierrez-de-Teran, Jose Brea, Enrique Ravina,
˜
Maria Isabel Loza,b Maria Isabel Cadavid,b Ferran Sanz,c Bernat Vidal,d
Victor Segarrad and Eddy Soteloa,*
aDepartment of Organic Chemistry, Faculty of Pharmacy, University of Santiago de Compostela,
Santiago de Compostela E-15782, Spain
bDepartment of Pharmacology, Faculty of Pharmacy, University of Santiago de Compostela, Santiago de Compostela E-15782, Spain
cUnitat de Recerca en Informa´tica Me´dica (GRIB), IMIM, Dep. CEXS, Universitat Pompeu Fabra,
Dr. Aiguader 88, Barcelona E-08003, Spain
dMedicinal Chemistry Department, Almirall, Treball 2-4, St. Just Desvern, Barcelona E-08960, Spain
Received 14 June 2007; revised 2 October 2007; accepted 17 October 2007
Available online 24 January 2008
Abstract—A collection of 25 2-(20-furyl)-1,2,4-triazolo[1,5-a]quinoxalines incorporating different substitution patterns at position 4
have been synthesized and their binding affinity towards human adenosine receptors (hA1, hA2A, hA2B and hA3) was determinated.
The biological data show that several potent at hA1, but lightly selective, adenosine ligands were identified. Moreover, these results
confirmed the hypothesis that the structural modifications carried out on the 4-position of the tricyclic system produces a remark-
able modification of the adenosine receptorial profile. A 3D-QSAR modelling study (GRIND/ALMOND methodology) performed
on the hA1 data gave further support to the pharmacological results, and it is presented as a useful tool for the future design of
ligands with better pharmacological profiles.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
therapeutic agents, but also useful pharmacological tools.
As a result of this, a huge number of AR agonists and
antagonists have been synthesized.
Interaction of adenosine with a set of specific receptors,
classified as A1, A2A, A2B and A3, may regulate different
physiological functions.1–5 For this reason, adenosine
receptors (AR) are currently considered as attractive
targets in the treatment of diverse pathophysiologic
disorders,6–9 particularly, AR antagonists are sought as
renal protective,16–18 anti-Parkinson,19,20 antiasthmat-
ic,21,22 antidiabetic,23 or antiischemic agents.24,25 AR be-
long to the superfamily of the G-protein coupled
receptors (GPCRs) and therefore they exert their physio-
logical role by activation or inhibition of different second
messenger systems.1 Many efforts have been invested in
the search for novel ligands with remarked selectivity,10–15
either for a given AR subtype or for their tissue distribu-
tion. This research not only pursues obtaining potential
The naturally occurring xanthines (e.g., caffeine and the-
ophylline) were the initial prototypic AR antago-
nists.26,27 The many attempts to improve their potency
and selectivity have resulted in the preparation of a large
number of xanthine derivatives, and much is known
now in terms of their structure–activity and structure–
selectivity relationships, as well as about their pharma-
cological activity.28–30 However, most of these xanthine
derivatives showed poor water solubility and a high met-
abolic rate (specially due to their interaction with cyto-
chrome P450 family) which strongly limitate its
drugability profile. The extensive research in this topic
have also lead to the discovery of different classes of
non-xanthine AR antagonists, being most of them nitro-
gen-containing heterocyclic compounds (Fig. 1).31–49
Keywords: Adenosine receptor binding; 1,2,4-Triazolo[1,5-a]quinoxa-
lines; 3D-QSAR.
*
Some of the early described tricyclic adenosine antago-
nists were identified from collections of compounds
Corresponding authors. Tel.: +34 981 563100 15221; fax: +34 981
528093; e-mail addresses: qosotelo@usc.es; qofara@usc.es
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.10.103

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2104
R
N
R
R
N
N
N
N
HN
N
N
N
NHR
N
NHR
N
NHR
Triazolo[1,5-c]quinazolines
Imidazo[4,5-c]quinolines
Triazolo[4,3-a]quinoxalines
R
R
O
N
N
N
N
N
N
N
NHR
N
NH2
Triazolo[4,3-a]quinoxalin-1-ones
Triazolo[1,5-a]quinoxalines
Figure 1. Structure of some representative tricyclic adenosine antagonists.
initially designed as ligands of the benzodiazepine recep-
tors.35–38 Several elegant examples of the pharmaco-
modulation of these prototypes have allowed the
development of new selective adenosine antagonists.39–42
Within the huge number of tricyclic heteroaromatic
systems tested as adenosine antagonists, those contain-
ing a 1,2,4-triazolo[1,5-a]quinoxaline scaffold have been
extensively explored, showing to be an extremely versa-
tile motif during the identification of new valuable and
selective AR ligands. To the best of our knowledge,
most of the published SAR studies related to this chem-
otype have been devoted to investigate the effect of dif-
ferent substituents (H, COOR, Aryl, Het.) at position
239 and to study the consequences of the bioisosteric
replacement on the azole ring39 (triazole, imidazole)
while the introduction of diversity at position 4 re-
mained almost unexplored until recently.43 In this sce-
nario, and as a part of our interest in the search for
novel adenosine receptor antagonists,47,48 we aimed to
evaluate the adenosine antagonistic profile at the human
receptors of a series of 4-substituted 2-(20-furyl)-1,2,4-
triazolo[1,5-a]quinoxalines (Fig. 2). Further motivation
for this work was found in the preliminary result that
showed significant affinity differences between human
and rat receptors for the parent compound 11 (see Table
3 and Ref. 39). We herein report their synthesis, phar-
macological characterisation at the four human adeno-
sine receptors and 3D-QSAR studies for the case of
the hA1 receptor.
comprehensive procedures described by Cecchy et al.39
for this heterocyclic system. Scheme 1 describes the
synthetic pathway which led to the 1,2,4-triazolo-
[1,5-a]quinoxaline core, briefly, reaction of 2-nitro-
phenylhydrazine 1 with the 2-furoic anhydride 2 afford
the hydrazide 3 in excellent yield (96%). Deshydrochlo-
rination of 3 was performed at reflux in phosphorus oxy-
chloride for 45 min (70%) (Scheme 1). Treatment of the
iminochloride 4 with ammonium hydroxide in a sealed
tube yielded the amidine 5 (98%) as a bright red solid
which was cyclized to 1,2,4-triazole 6 (92%) by reaction
with ethoxyoxalyl chloride. The reflux conditions
required to reduce the nitro group in 6, employing iron
in glacial acetic acid, additionally promote ring closure,
yielding the tricyclic derivative 7 (98%). Reaction of
2-furyl-1,2,4-triazolo[1,5-a]quinoxalin-4-one 7 with phos-
phoryl oxychloride allows to isolate (80%) the key inter-
mediate 8 (Scheme 1).
The structural diversification performed at the position 4
of the heterocyclic system is shown in Scheme 2. Nucleo-
philic replacement of the chloro atom at the highly reac-
tive iminochloride 8 by ammonia, amines or alcohols
afforded amines 9, 11 and ethers 10, respectively (Tables
1 and 2). Finally, the synthesis of the amides 12 has been
performed by direct acylation allowing to react 11 with
acyl chlorides in pyridine (Scheme 2, Tables 1 and 2).
Most relevant analytical and spectroscopic data of ob-
tained compounds are described in Tables 1 and 2.
1.1. Chemistry
2. Pharmacology
The general method employed for the synthesis of the 4-
substituted-2-(20-furyl)-1,2,4-triazolo[1,5-a]quinoxalines
9–12 is illustrated in Schemes 1 and 2 and follow the
The binding activities of the obtained compounds at the
four human adenosine receptors subtypes were deter-
mined using radioligand binding assays procedures
(see Section 5). The affinity of those compounds that
did not fully displace radioligand binding was expressed
as the percentage of inhibition of specific binding at
10ꢀ6 M (Table 3).
O
N
N
X
N
3. Results and discussion
N
Affinities of 4-substituted 2-(20-furyl)-1,2,4-triazolo[1,5-
a]quinoxalines 9–12 in radioligand binding assays at
Figure 2. General structure of the obtained 4-substituted 2-(20-furyl)-
1,2,4-triazolo[1,5-a]quinoxalines.

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2105
O
H
N
NHNH2
O
O
i
N
H
+
O
O
NO2
NO2
O
O
1
2
3
ii
O
NH2
Cl
H
N
H
N
N
N
iii
iv
O
O
N
N
N
NO2
NO2
COOEt
NO2
6
5
4
v
O
O
N
N
N
N
N
N
vi
N
H
O
N
Cl
7
8
Scheme 1. Synthesis of the 2-(20-furyl)-1,2,4-triazolo[1,5-a]quinoxaline core. Reagents and conditions: (i) EtOH, reflux; (ii) POCl3, reflux; (iii) NH3/
100 °C; (iv) ClCOCOOEt, reflux; (v) Fe/AcOH, reflux; (vi) POCl3.
O
O
O
N
N
N
N
N
N
N
N
N
i
ii
N
NRR´
N
Cl
N
OR
9a-o
8
10a-e
iii
O
O
N
N
N
N
N
N
iv
N
NH2
N
NHCOR
11
12a-e
Scheme 2. Synthesis of 4-substituted 2-(20-furyl)-1,2,4-triazolo[1,5-a]quinoxalines. Reagents and conditions: (i) RNHR/EtOH, 100 °C; (ii) ROH/
THF, NaH; (iii) NH3; (iv) RCOCl/pyridine.
the four human adenosine receptors (A1, A2A, A2B and
A3) are reported in Table 3. It should first be pointed
out that, in our hands, the affinity elicited for the early
described amine 11 at the human A1 receptors was slight
superior to the values described in the literature39
(pKi = 7.52 vs pKi = 8.25), most probably due to either
the different species of the receptor employed during
the studies (rat vs human) or the experimental differ-
ences among both assays. Examination of the binding
results listed in Table 3 indicated that several potents,
but lightly selective, adenosine ligands were identified.
Moreover, the obtained data confirmed our initial
hypothesis that the structural modifications carried out
on the 4-position of the tricyclic system produces a
remarkable modification of the adenosine receptorial
profile in terms of affinity and/or selectivity.
The 4-amino derivatives (9a–o and 11) represent the
most populated set of compounds obtained during this
study (16 derivatives). As expected, replacement of

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2106
Table 1. Structure of the obtained 4-substituted 2-(20-furyl)-1,2,4-triazolo[1,5-a]-quinoxalines 9–12
O
N
N
N
N
X
Compound
X
Mp (°C)
Recr. solvent
Yield (%)
11
9a
NH2
NHPh
NHCH2Ph
276–277
214–216
149–151
177–179
241–243
202–204
185–187
141–143
113–114
108–110
121–123
144–146
105–106
210–212
156–157
272–274
174–176
163–165
206–208
191–193
176–178
224–226
228–230
189–190
266–267
234–236
EtOH
78
85
73
82
88
73
73
73
91
71
82
91
69
88
93
65
68
64
45
80
73
28
18
44
23
18
MeOH
MeOH
MeOH
MeOH
MeOH
i-PrOH
MeOH
i-PrOH
MeOH
i-PrOH
i-PrOH
EtOH
9b
9c
NHCH2-(2-Furyl)
NHCH2CH2Ph
NHCH2CH2OH
NHCH2CH(OH)CH3
NHCH2CH(OH)CH2NH2
NHCH2CH2OCH3
NHCH2CH2CH2OCH3
NHCH2CH(OCH3)2
N(CH3)2
9d
9e
9f
9g
9h
9i
9j
9k
9l
N(CH2CH3)2
9m
9n
N(CH2CH2OH)2
N(CH2CH2)2NH
N(CH2CH2)2N–CH2CH2OH
OCH3
MeOH
i-PrOH
MeOH
MeOH
MeOH
i-PrOH
MeOH
i-PrOH
MeOH
i-PrOH
i-PrOH
MeOH
i-PrOH
9o
10a
10b
10c
10d
10e
12a
12b
12c
12d
12e
OCH2CH3
OCH2Ph
OCH2-(2-Furyl)
OCH2-(3-Furyl)
NHCOCH3
NHCOCH2Ph
NHCOCH2CH2Ph
NHCO-(2-Furyl)
NHCO-(2-Thienyl)
hydrogen atoms in the amino group of the parent com-
pound of the series (11) to obtain different secondary
and tertiary amines had contrasting pharmacological
effects on the affinity and selectivity profiles. Thus,
phenylation of 11 to give 9a reduced the affinity to-
wards all adenosine receptor subtypes (Table 3). A sim-
ilar behaviour is observed for the N-side chain superior
homologues of 9a (9b and 9c). A comparison of the
biological data obtained for the 4-benzyl and 4-furfuryl
1,2,4-triazolo[1,5-a]quinazolines (9b and 9c) reveals that
the substitution of the phenyl ring by the furan pro-
duced a moderate (fivefold) increase in potency at the
hA1 receptor and, simultaneously, improve the selectiv-
ity towards the other adenosine receptor subtypes. In
tion (9h, 9i, 9j)] afforded derivatives with attenuated
biological activity. These data suggest the importance
in the binding site of possible moieties with hydrogen-
bond acceptor capabilities. A dramatic drop of affinity
was observed for tertiary or heterocyclic amines, (9k,
9l, 9n, 9o) a result which confirm the occurrence of a
strong interaction of the NHR group at 4 position with
a hydrogen-bond acceptor in the active binding site.46
Interestingly, in this series the most remarkable com-
pound is the bis-2-hydroxyethyl derivative (9m) which
elicited a moderate hA1 affinity (pKi = 7.23). Such a re-
sult suggests that the lack of the NH group could be
partially counterbalanced by the presence of two highly
flexible carbinol moieties of the alkyl chain, thus recov-
ering the capacity of hydrogen-bond interaction with
the receptor.
general, introduction of amines containing
a 2-
hydroxyethyl fragment (9e, 9f, 9g, 9m) positively influ-
ences the hA1 and the hA3 binding activity, while de-
creases hA2A and hA2B affinity. It is worth noting
that the 2-hydroxyethylamino derivative 9e possess
the highest hA1AR affinity (pKi = 8.50) among the
herein reported compounds. As observed in Table 3
the modifications designed to evaluate the role of the
alcoholic function on the lead compound 9e [e.g., intro-
duction of substituents in the carbon b of the alkyl
chain (methyl 9f or aminomethyl 9g) and ether forma-
Introduction of alcohoxy groups at 4 (10a, 10b, 10c)
produces an important decrease in the antagonistic
AR activity (Table 3). The weak hA3 antagonistic activ-
ity observed for the benzyl derivative 10c aimed us to
prepare the furyl analogues 10d and 10e. In these pair
of compounds, with diminished lipophilicity on the
substitution at 4, an increased A3 binding affinity is ob-
served. Moreover, derivative 10d showed to be not only

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2107
Table 2. Analytical and spectroscopic data of 4-substituted 2-(20-furyl)-1,2,4-triazolo[1,5-a]quinoxalines 9–12
Compound
MS (M+)
m/z (%)
1H NMR (CDCl3/TMS) d (ppm), J (Hz)
Analysis C, H, N (Calcd/Found)
11
9a
251 (100)
327 (83)
8.29 (d, J = 8.1, 1H), 7.78 (d, J = 8.0, 1H), 7.70–7.40 (m, 3H), 7.22
(d, J = 3.0, 1H), 6.76 (m, 1H), 5.82 (br s, 2H)
C (62.15/62.32), H, (3.61/3.68), N (27.87/
27.97)
8.12 (dd, J = 8.0, 1.1, 1H), 7.89–7.86 (m, 3H), 7.73 (dd, J = 8.1, 1.1,
1H), 7.52 (br s, 1H), 7.39–7.28 (m, 4H), 7.09 (d, J = 3.3, 1H), 7.02
(d, J = 7.3, 1H), 6.47 (dd, J = 3.3, 1.7, 1H)
C (69.71/70.02), H, (4.00/4.12), N (21.39/
21.54)
9b
9c
9d
341 (62)
331 (72)
355 (20)
8.25 (dd, J = 8.1, 1.1, 1H), 7.79 (d, J = 8.0, 1H), 7.60 (d, J = 0.7,
1H), 7.51–7.25 (m, 7H), 7.16 (d, J = 3.3, 1H), 6.56 (dd, J = 3.3, 1.7,
1H), 6.48 (t, J = 5.4, 1H), 4.87 (d, J = 5.4, 2H)
C (70.37/70.52), H, (4.43/4.65), N (20.52/
20.78)
8.25 (dd, J = 8.1, 1.3, 1H), 7.79 (dd, J = 8.1, 1.3, 1H), 7.61 (dd,
J = 1.7, 0.6, 1H), 7.53 (m, 3H), 7.17 (dd, J = 3.4, 0.6, 1H), 6.57 (dd,
J = 3.4, 1.7, 1H), 6.42–6.31 (m, 3H), 4.90 (d, J = 5.2, 2H)
8.26 (dd, J = 8.1, 1.1, 1H), 7.80 (dd, J = 8.1, 1.1, 1H), 7.62 (d,
J = 0.9, 1H), 7.52 (m, 1H), 7.47 (d, J = 1.3, 1H), 7.37–7.15 (m, 6H),
6.58 (dd, J = 3.4, 1.7, 1H), 6.28 (br s, 1H), 4.00 (m, 2H), 3.07 (m,
2H)
C (65.25/65.29), H, (3.95/4.10), N (21.14/
21.37)
C (70.97/71.21), H, (4.82/4.95), N (19.71/
19.79)
9e
9f
295 (30)
309 (20)
8.20 (dd, J = 8.1, 1.1, 1H), 7.71 (d, J = 8.1, 1H), 7.66 (d, J = 0.9,
1H), 7.52 (m, 1H), 7.41 (m, 1H), 7.18 (d, J = 3.4, 1H), 6.61 (dd,
J = 3.4, 1.7, 1H), 3.93 (t, J = 4.8, 2H), 3.84 (t, J = 4.8, 2H)
8.21 (dd, J = 7.9, 1.0, 1H), 7.68 (d, J = 1.1, 1H), 7.60 (d, J = 0.9,
1H), 7.48–7.36 (m, 2H), 7.34 (d, J = 0.9, 1H), 7.14 (d, J = 3.0, 1H),
6.68 (br s, 1H), 6.56 (dd, J = 3.3, 1.5, 1H), 4.72 (m, 1H), 4.22 (m,
1H), 3.88 (m, 1H), 3.69 (m, 1H), 1.32 (s, 3H)
C (61.01/61.22), H, (4.44/4.61), N (23.72/
24.01)
C (62.13/62.25), H, (4.89/4.97), N (22.64/
22.91)
9g
9h
9i
324 (10)
309 (25)
323 (35)
8.22 (d, J = 8.1, 1H), 7.71 (d, J = 8.1, 1H), 7.59 (m, 1H), 7.51–7.30
(m, 2H), 7.15 (d, J = 3.5, 1H), 6.70 (br s, 1H), 6.55 (d, J = 3.5, 1H),
3.93 (m, 2H), 3.74 (m, 1H), 2.91 (m, 2H), 2.72 (br s, 2H)
8.21 (d, J = 8.0, 1H), 7.71 (d, J = 8.0, 1H), 7.60 (m, 1H), 7.42 (m,
1H), 7.34 (m, 1H), 7.16 (d, J = 2.8, 1H), 6.55 (d, J = 1.3, 1H), 3.88
(m, 2H), 3.68 (m, 2H), 3.40 (s, 3H), 2.30 (br s, 1H)
C (59.25/59.43), H, (4.97/5.08), N (25.91/
26.25)
C (62.13/62.32), H, (4.89/4.99), N (22.64/
22.87)
8.22 (d, J = 8.1, 1H), 7.74 (d, J = 8.1, 1H), 7.62 (m, 1H), 7.60 (t,
J = 7.6, 1H), 7.35 (t, J = 7.6, 1H), 7.18 (d, J = 3.3, 1H), 6.58 (m,
1H), 6.40 (m, 1H), 3.80 (m, 2H), 3.49 (t, J = 5.9, 2H), 3.39 (s, 3H),
2.04 (m, 2H)
C (63.15/63.32), H, (5.30/5.41), N (21.66/
21.85)
9j
339 (10)
8.30 (dd, J = 8.0, 1.4, 1H), 7.70 (dd, J = 8.0, 1.4, 1H), 7.64 (dd,
J = 1.5, 0.6, 1H), 7.51–7.40 (m, 2H), 7.21 (dd, J = 3.5, 0.6, 1H),
6.58 (dd, J = 3.5, 1.5, 1H), 6.20 (br s, 1H), 4.67 (m, 1H) 3.90 (m,
2H), 3.47 (s, 6H)
C (60.17/60.32), H, (5.05/5.11), N (20.64/
20.75)
9k
9l
279 (51)
307 (45)
8.22 (dd, J = 8.1, 1.1, 1H), 7.68–7.61 (m, 2H), 7.41 (d, J = 8.1, 1H),
7.32 (dd, J = 3.3, 1.7, 1H), 7.15 (d J = 3.3, 1H), 6.56 (d J = 1.7,
1H), 3.59 (s, 6H)
C (64.51/64.59), H, (4.69/4.87), N (25.07/
25.21)
8.16 (dd, J = 8.1, 1.3, 1H), 7.59–7.52 (m, 2H), 7.33 (m, 1H), 7.07 (d,
J = 3.4, 1H), 6.47 (dd, J = 3.4, 1.8, 1H), 4.02 (q, J = 7.0, 4H), 1.25
(t, J = 7.0, 6H)
C (66.43/66.52), H, (5.58/5.67), N (22.79/
23.01)
9m
9n
339 (35)
320 (30)
8.24 (d, J = 6.4,1H), 7.62–7.54 (m, 2H), 7.47–7.40 (m, 2H), 7.15 (m,
1H), 6.56 (m, 1H), 4.20 (m, 4H), 4.04 (m, 4H)
C (60.17/60.24), H, (5.05/5.14), N (20.64/
20.80)
8.30 (dd, J = 8.1, 1.2, 1H), 7.71 (dd, J = 8.1, 1.2, 1H), 7.63 (m, 1H),
7.53 (m, 1H), 7.48 (m, 1H), 7.21 (m, 1H), 6.58 (dd, J = 3.3, 1.7,
1H), 4.34 (m, 4H), 3.10 (m, 4H), 2.85 (br s, 1H)
C (63.74/63.79), H, (5.03/5.21), N (26.23/
26.38)
9o
364 (60)
8.28 (dd, J = 8.0, 1.2, 1H), 7.72 (dd, J = 8.0, 1.2, 1H), 7.63–7.34 (m,
2H), 7.20 (d, J = 3.3, 1H), 6.58 (dd, J = 3.3, 1.7, 1H), 4.41 (m, 4H),
3.71 (m, 2H), 2.72 (m, 4H), 2.62 (m, 2H), 2.04 (br s, 1H)
8.46 (dd, J = 7.7, 1H), 8.39 (m, 1H), 8.10 (d, J = 7.7, 1H), 7.60–7.50
(m, 2H), 7.32 (m, 1H) 6.60 (s, 1H), 4.29 (s, 3H)
C (62.62/62.72), H, (5.53/5.68), N (23.06/
23.17)
10a
10b
266 (100)
280 (100)
C (63.15/63.31), H, (3.79/3.97), N (21.04/
21.26)
8.40 (m, 1H), 7.90 (m, 1H), 7.65 (dd, J = 1.7, 0.9, 1H), 7.61–7.57
(m, 2H), 7.32 (d, J = 3.4, 1H), 6.60 (dd, J = 3.4, 1.7, 1H), 4.79 (q,
J = 7.1, 2H), 1.60 (t, J = 7.1, 3H)
C (64.28/64.33), H, (4.32/4.45), N (19.99/
20.17)
10c
10d
342 (30)
332 (25)
8.35 (1H, m), 7.88 (1H, m), 7.62–7.53 (5H, m), 7.41–7.26 (4H, m),
6.57 (dd, J = 3.4, 1.7, 1H), 5.57 (s, 2H)
C (70.17/70.35), H, (4.12/4.22), N (16.37/
16.51)
8.40 (m, 1H), 7.94 (m, 1H), 7.63 (m, 3H), 7.45 (dd, J = 1.8, 0.7,
1H), 7.30 (dd, J = 3.3, 0.7, 1H), 6.62 (d, J = 3.3, 1H), 6.58 (dd,
J = 3.3, 1.8, 1H), 6.38 (dd, J = 3.3, 1.8, 1H)
C (65.06/65.22), H, (3.64/3.81), N (16.86/
17.01)
12a
293 (20)
8.78 (br s, 1H), 8.38 (m, 1H), 8.02 (m, 1H), 7.68–7.62 (m, 3H), 7.24
(dd, J = 3.4, 0.5, 1H), 6.62 (dd, J = 3.4, 1.7, 1H), 2.74 (s, 3H)
C (61.43/61.52), H, (3.78/4.00), N (23.88/
23.97)
(continued on next page)

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2108
Table 2 (continued)
Compound
MS (M+)
m/z (%)
1H NMR (CDCl3/TMS) d (ppm), J (Hz)
Analysis C, H, N (Calcd/Found)
12b
369 (15)
8.85 (br s, 1H), 8.37 (m, 1H), 8.05 (m, 1H), 7.66–7.62 (m, 3H),
7.45–7.26 (m, 6H), 7.20 (dd, J = 3.4, 0.6, 1H), 6.60 (dd, J = 3.4, 1.8,
1H), 4.37 (s, 2H)
C (68.28/68.41), H, (4.09/4.18), N (18.96/
19.02)
12c
383 (52)
8.77 (br s, 1H), 8.37 (m, 1H), 8.00 (m, 1H), 7.65–7.61 (m, 3H),
7.33–7.18 (m, 5H), 6.61 (dd, J = 3.5, 1.7, 1H), 3.39 (m, 2H), 3.16
(m, 2H)
C (68.92/69.07), H, (4.47/4.54), N (18.27/
18.35)
12d
12e
345 (25)
361 (26)
9.61 (br s, 1H), 8.45 (br s, 1H), 8.24 (m, 1H), 7.71–7.68 (m, 3H),
7.45 (m, 1H), 7.33 (d, J = 3.4, 1H), 6.60 (m, 1H)
11.60 (br s, 1H), 8.44 (m, 1H), 8.22–7.84 (m, 6H), 7.32 (m, 2H),
6.76 (m, 1H)
C (62.61/62.74), H, (3.21/3.40), N (20.28/
20.42)
C (59.82/60.00), H, (3.07/3.24), N (19.38/
19.53)
Table 3. Structure and affinity binding data of the obtained 4-Substituted 2-(20-furyl)-1,2,4-triazolo[1,5-a]-quinoxalines 9–12 at the human adenosine
receptors
O
N
N
N
N
X
Compound
X
Human adenosine receptors (pKi or % inhibition at 10ꢀ6 M)
A1
A2A
A2B
A3
11
9a
NH2
NHPh
NHCH2Ph
8.25 0.12
7.34 0.12
6.99 0.09
7.48 0.20
7.03 0.16
8.50 0.21
8.40 0.07
6.85 0.14
6.52 0.08
7.13 0.10
6.24 0.10
31.10 3%
6.87 0.23
7.23 0.10
12.30 2%
24.54 3%
6.48 0.14
5.43 0.13
13.83 4%
55.11 3%
6.27 0.2
7.95 0.10
7.19 0.06
6.52 0.05
6.31 0.15
7.29 0.10
7.20 0.13
7.32 0.10
57 4%
7.88 0.08
49 4%
24 4%
6.48 0.07
71.7 3%
32.6 4%
57.3 1%
6.43 0.15
7.52 0.15
7.64 0.19
43 4%
9b
9c
NHCH2-(2-Furyl)
NHCH2CH2Ph
NHCH2CH2OH
NHCH2CH(OH)CH3
NHCH2CH(OH)CH2NH2
NHCH2CH2OCH3
NHCH2CH2CH2OCH3
NHCH2CH(OCH3)2
N(CH3)2
5.76 0.10
46 5%
9d
9e
6.91 0.15
7.17 0.13
34 3%
9f
9g
9h
9i
5.51 0.10
6.43 0.18
20 2%
32 4%
6.36 0.10
82 3%
6.37 0.08
22 5%
17 2%
9j
55.3 4%
50.7 3%
10.4 2%
51.7 2%
26.7 3%
24 4%
9k
9l
6.51 0.09
6.40 0.08
6.96 0.21
N(CH2CH3)2
5.62 0.15
46 4%
9m
9n
N(CH2CH2OH)2
N(CH2CH2)2NH
N(CH2CH2)2N–CH2CH2OH
OCH3
7
2%
4
3
1%
2%
9o
12 4%
6.4 0.07
38 2%
38 3%
10a
10b
10c
10d
10e
12a
12b
12c
12d
12e
28 3%
34 4%
57.4 2%
69 1%
OCH2CH3
OCH2Ph
3
3
2%
2%
5.81 0.10
7.18 0.07
6.17 0.10
6.60 0.06
7.0 0.08
6.9 0.12
6.6 0.13
52 3%
OCH2-(2-Furyl)
OCH2-(3-Furyl)
NHCOCH3
8
1%
6.24 0.07
6.70 0.12
6.8 0.10
7.2 0.05
6.4 0.02
10 2%
6.84 0.1
6.6 0.14
6.6 0.05
6.3 0.06
5.6 0.08
21 3%
6.95 0.12
7.2 0.07
7.9 0.20
NHCOCH2Ph
NHCOCH2CH2Ph
NHCO-(2-Furyl)
NHCO-(2-Thienyl)
6.7 0.12
7.3 0.06
a relatively potent (pKi = 7.18) but also a highly selective
ligand of the A3 receptor subtype. The next exploration
consisted on the replacement of one of the hydrogen
atoms of the amino group at 11 by acyl or aroyl frag-
ments to afford amides 12. Such an exploration aimed
to validate previous findings reported in analogous het-
eroaromatic systems regarding the modification of the
affinity and selectivity profile as a consequence of the in-
creased acidity of the NH moiety.21,30,48 As observed in
Table 3, amidation of the amino group at 4 position of
the heteroaromatic system produces no improvements in
affinity at any receptor subtype. The only remarkable
information extracted from this series is the data ob-
tained for the 2-thienoyl derivative 12e, which is a rela-
tively potent (pKi = 7.3) and highly selective hA1
antagonist.

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2109
A 3D-QSAR modelling study was performed by means
of the GRIND/ALMOND methodology,49 in order to
gain insight into the structure–activity relationships de-
rived from the variability explored in position 4. Inter-
pretable models need to be constructed using a
minimum set of compounds that offers a large enough
window of activities. For this reason, only hA1 (n = 21
compounds, DpKi = 3.1 log units) and hA2A models
(n = 18 compounds, DpKi = 2.4 log units) were at-
tempted. Given the high planarity of the 1,2,4-triazol-
o[1,5-a]-quinoxaline scaffold, the only existing
conformational degree of freedom is the bond that links
the quinoxaline and furane rings. The torsional angle
(N4–C3–C2–O) was explored through a quantum
mechanical torsion scan performed on the parent com-
pound 11. As a result, two local minima were found with
the 2-furyl located co-planar with respect to the quinox-
aline scaffold, that is, with the oxygen pointing in the
same direction of N4 (conformation A) or in the direc-
tion of N2 (conformation B). The small energy differ-
ence between both conformations (0.9 kcal/mol)
justified the construction of two independent QSAR
models, considering either A or B conformation. How-
ever, this conformational variability had a minimum im-
pact on the performance of the models (see Table 4).
Given that the statistical performance is slightly better
in the model generated from conformation A in both
the A1 and A2A receptors, we will only discuss about
that model from now on.
due to the different substitutions explored in position 4,
does not influence the model. This means that the size
variability in the present series has little correlation with
the activities. Therefore, the most successful and explan-
atory model was built using only 3 correlograms (O–O,
N1–N1 auto-correlograms and O–N1 cross-correlo-
gram) containing a total number of 74 variables, after
one FFD selection cycle. The PLS regression consider-
ing 3 latent variables (LV) resulted in a statistically sat-
isfactory model (r2 = 0.82, qð2LOOÞ ¼ 0:64). As stated
before, the inclusion of the TIP probe not only compli-
cates the model, but also reduces slightly the predictabil-
ity (3 LV, r2 = 0.85, qð2LOOÞ ¼ 0:56). The interpretation of
the model is on the same trend as the SAR observations.
The presence of an H-bond donor in the variable region
(4-substitution) is favourable for the hA1 activity. This is
captured by the regions with highest PLS coefficient, lo-
cated in the O–N1 cross-correlogram, depicting a phar-
˚
macophoric distance of 10 A between H-bond donor in
position 4 and the H-bond acceptor region correspond-
ing to N2 in the heterocycle (Fig. 3A). Interestingly, the
only molecule not bearing a NH-R substituent that
shows high values for these variables is the active com-
pound 9m, which has a tertiary amine with two free hy-
droxyl groups. The presence of two H-bond donors
(e.g., compounds 9e, 9f, 9m and 11) strengths for hA1
activity.
To further investigate the shape tolerance of the hA1
binding site, the 5 molecules without Ki values were in-
cluded in a QSAR model for qualitative predictions.51
This was done using the technique of PLS combined
with a discriminant analysis (PLS-DA), by transforming
the activity values in discrete binary values: 0 if
pKi < 6.5 or absence of pKi, 1 for pKi > 5. This cutoff
was selected to equilibrate the number of actives and
inactives, but after a first inspection compound 9l
(pKi = 6.87) was identified as an outlier and left out
for further consideration, resulting in a final set of 15 ac-
tives and 10 inactives. Once more, we found that the best
The 3D-QSAR model for hA1AR activities had a satis-
factory statistical quality, both in terms of fitting and
internal validation. Due to its negligible impact on the
model, the DRY probe was removed from the initial
set of probes. This is reasonable, since only a few mole-
cules presented any favourable region for the interaction
with the DRY probe in the area corresponding to the
variable region (substituent in 4 of the quinoxaline scaf-
fold). More surprising is the fact that the inclusion of
TIP probe,50 which captures the differences in the shape
Table 4. Statistics of PLS models for the A1 and A2A human adenosine receptors
O
O
N
N
N
N
N
N
N
A
X
N
X
B
Receptor
hA1AR
n
Conformation
LV
r2
q2
SDEC
SDEP
21a
A
B
3
3
2
0.82
0.82
0.98
0.64
0.60
0.89
0.31
0.31
0.27
0.44
0.46
0.39
25b
18a
A
hA2AAR
A
B
3
2
0.92
0.88
0.70
0.62
0.15
0.19
0.30
0.33
All the models were built with single FFD run and LOO standard validation procedure.
a Only compounds with experimentally available pKi were included in the models.
b PLS-DA model. Compound 9l was left out as an outlier.

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2110
Figure 3. Active compound 9f is represented with the corresponding MIFs used to generate the 3D-QSAR models. The most relevant GRIND
descriptors (with high PLS coefficients) are represented by a red bar. (A) Most important distance between H-bond donor (red) and H-bond acceptor
(blue) regions for the hA1 model. The same region has also high relevance for the A2A model. (B) The most relevant descriptors for the O-TIP cross-
correlogram in the hA2A model. (C) Molecular shape, showing the optimum distance between the 2-furlyl substitution and the variable substituent in
position 4 as observed in the hA2A model.
model (2 LV, r2 = 0.98, q2 = 0.89) was obtained by only
taking into consideration the N1 and O MIFs, thus con-
firming that the present series do not explore the limits
of the shape or flexibility of the binding site of the
hA1 receptor. The interpretability of the model is analo-
gous to what is described above for the model developed
for qualitative predictions. It is worth to note the excel-
lent statistical parameters of this model, showing a clear
capacity to distinguish between active and inactive com-
pounds using the aforementioned description of the
molecules.
pound 11 (R = H) or compounds bearing relatively
bulky substitutions (e.g., 9f, 11c) all display good hA1
affinities. The results of this study clearly reveal that
the structural manipulation at position 4 and specially
the derivatization of the amino group could allow the
discovery of new potent and selective ligands for the
hA1 and hA3 receptors.
The 3D-QSAR modelling confirmed that the optimum
substitution in position 4 should include H-bond accep-
tor/donor capabilities. The moderate affinity of molecule
9m, where the lack of NH– in position 4 is counterbal-
anced with the presence of hydroxyl groups in the vari-
able region, suggests for the presence of a flexible
hydrogen-bonding counterpart of position 4 in the hA1
receptor. This interpretation is reinforced by the fact
that the inclusion of TIP probe, which has proven as a
very useful tool to correlate the shape of the molecules
with their activities,50–52 had a negligible effect in the
hA1 QSAR model, observed in both the PLS and
PLS-DA models. On the contrary, the inclusion of such
a description of the molecular shape is essential in order
to obtain a predictable model for the hA2A receptor. We
can conclude that the binding site of the hA1 receptor
tolerates well the variations within the shape limits ex-
plored by the present series, while the hA2A receptor is
more restrictive. Whether this is due to a flexible binding
pocket in the hA1 receptor or to a pocket size in this
receptor which is greater or equal to the maximum size
explored by the present series cannot be determined with
the present methodology. All together, the results here
reported should guide further chemical explorations of
the position 4 of the heterocycle and receptor-based
molecular modelling exercises.
Regarding the A2A receptor, the statistical quality of a
3D-QSAR model that only considers the O and N1 probes
has unacceptable statistical parameters (3LV, r2 = 0.83,
q2ðLOOÞ ¼ 0:35) while the inclusion of the shape probe
(TIP) significantly improves these values (3LV,
r2 = 0.92, qð2LOOÞ ¼ 0:70). Therefore, as opposed to the
hA1 receptor, the consideration of the molecular shape
is determinant for the predictability of a hA2A model.
The observation that the highest activity is obtained with
a H-bond donor in position 4, with some spatial con-
strain, is also true for the present case. Thus, highest
PLS pseudocoefficients for the O–N1 cross-correlogram
establish the same pharmacophoric distance for the H-
bond donor in 4 and the H-bond acceptor corresponding
to N2 (Fig. 3A). Additionally, in the A2A model the posi-
tion of this H-bond donor is conditioned by the shape due
to the variable region in 4 and the furyl at position 2 of the
quinoxaline (Fig. 3B). Finally, the highest PLS coefficient
for the TIP auto-correlogram identifies the optimal shape
˚
for the most active molecules (TIP–TIP 18 A, shown in
Fig. 3C), which can be related to the depth of the binding
pocket of the A2A receptor.
4. Conclusions
5. Experimental
5.1. Synthesis
In summary, the biological data shown in Table 2 con-
firm several previous theoretical and experimental stud-
ies regarding the importance of the presence of a NH
group at position 4 of the heterocyclic system for their
interaction with the active binding site in the receptor.
Further, our results point that the binding site allocating
the explored position of the heterocycle tolerates well
substituents with different sterical properties, as exem-
plified with the observation that even the parent com-
Melting points were measured on a Gallenkamp appara-
tus and are uncorrected. Infrared spectra were recorded
on a Perkin-Elmer 1640 FT-IR spectrophotometer. H
NMR spectra were obtained on Bruker WM250 and
AM300 spectrometers using tetramethylsilane as inter-
nal standard (chemical shifts are d values, J in Hz). Mass
spectra were determined on a Varian MAT-711 instru-
1

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2111
ment. Elemental analyses were performed on a Perkin-
5.4. General procedure for the preparation of amides
(12a–e)
Elmer 240B apparatus at the Microanalysis Service of
the University of Santiago de Compostela and the ob-
tained results for C, N and H were within 0.4% of
the theoretical values. The progress of the reactions
was monitored by thin layer chromatography with
2.5 mm Merck silica gel GF 254 strips, and the purified
compounds each showed a single spot; unless otherwise
stated iodine vapour and/or UV light were used for
detection. Chromatographic separations were per-
formed on silica gel columns by flash chromatography
(Kieselgel 40, 0.040–0.063 mm). The synthesis of the
precursors 8 and 11 has been performed following the
previously described procedure.36,38,39
To a solution of 4-amino-2-(20-furyl)-1,2,4-triazolo[1,5-
a]quinoxaline 11 (0.40 mmol) in 5 mL of dry pyridine
under argon atmosphere was added slowly the appropri-
ate acyl chloride (0.80 mmol). The reaction mixture was
stirred at room temperature during 8 h and then heated
at 40 °C until the starting material had disappeared. (1–
6 h). The reaction mixture was then treated with 10%
hydrochloric acid and extracted with methylene chlo-
ride. The organic extracts were dried (Na2SO4), concen-
trated to dryness under reduced pressure and the residue
was purified by column chromatography on silica gel to
obtain the expected amides 12. Further purification was
performed by recrystallization of the proper solvent (Ta-
bles 1 and 3).
5.2. General procedure for the preparation of amines (9
and 11)
A solution of 4-chloro-2-(20-furyl)-1,2,4-triazolo[1,5-
a]quinoxaline 8 (1.1 mmol) in ethanol (8 mL) was mixed
with the appropriate amine (2.4 mmol) and potassium
iodide (0.020 g) in a sealed tube under argon atmo-
sphere. The reaction mixture was then stirred and
heated at 100 °C (oil bath) until the starting material
had disappeared (1.5–16 h). After cooling, the solution
was concentrated to dryness under reduced pressure
and the residue was purified by column chromatography
on silica gel to give the corresponding amines 9 and 11.
Further purification was performed by recrystallization
of the proper solvent (Tables 1 and 3).
5.5. Pharmacology: radioligands binding studies
Radioligand binding competition assays were performed
in vitro as previously described53 using A1, A2A, A2B and
A3 human receptors expressed in transfected CHO (A1),
HeLa (A2A and A3) and HEK-293 (A2B) cells.
The experimental conditions used are summarized in
Table 5.
In each instance aliquots of membranes (15 lg for A1,
10 lg for A2A, 18 lg for A2B and 90 lg for A3) in buffer
A (see Table 5) were incubated for the specified period at
25 °C with the radioligand (2–35 nM) and 6 different
concentrations (ranging from 0.1 nM to 1 lM) of the
test molecule in a final volume of 200 ll. The binding
reaction was stopped by rapid filtration in a multiscreen
5.3. General procedure for the preparation of ethers
(10a–d)
To a solution of 4-chloro-2-(20-furyl)-1,2,4-triazolo[1,
5-a]quinoxaline 8 (0.30 mmol) in 4 mL of dry tetrahydro-
furane under argon atmosphere were added the appropri-
ate alcohol (0.90 mmol) and sodium hydride
(0.90 mmol). The reaction mixture was stirred at room
temperature until the starting material had disappeared.
(0.5–1 h) and then treated with 5 mL of methanol and
1 mL of water. The solution was concentrated to dryness
under reduced pressure and the residue was purified by
column chromatography on silica gel to obtain the ex-
pected ethers 10. Further purification was performed by
recrystallization of the proper solvent (Tables 1 and 3).
manifold system (Milipore Iberica, Madrid, Spain). Un-
´
bound radioligand was removed by washing four times
with 250 ll ice-cold buffer B for A1 and A2A receptors,
and six times 250 ll ice-cold buffer B for A2B and A3
receptors (see Table 4). Non-specific binding was deter-
mined using a 50 or 400 lM NECA solution for A2A
and A2B receptors and 10 or 100 lM R-PIA solution
for A1 and A3, respectively. Radioactivity retained on
filters was determined by liquid scintillation counting
using Universol (ICN Biochemicals, Inc.). The binding
affinities were determined using [3H]DPCPX (130 Ci/
mmol; Amersham Biosciences, Barcelona, Spain) as
Table 5. Conditions used for radioligand binding assays using A1, A2A, A2B and A3 human receptors
A1
A2A
A2B
A3
Buffer A
20 mM Hepes, 100 mM NaCl,
10 mM MgCl2, 2 U/mL
adenosine deaminase (pH 7.4)
50 mM Tris–HCl, 1 mM
EDTA, 10 mM MgCl2,
2 U/mL adenosine
50 mM Tris–HCl, 1 mM
EDTA, 10 mM MgCl2,
0.1 mM benzamidine,
2 U/mL adenosine
deaminase (pH 6.5)
50 mM Tris–HCl (pH
6.5)
50 mM Tris–HCl, 1 mM
EDTA, 5 mM MgCl2,
2 U/mL adenosine
deaminase (pH 7.4)
deaminase (pH 7.4)
Buffer B
20 mM Hepes, 100 mM NaCl,
10 mM MgCl2, (pH 7.4)
50 mM Tris–HCl, 1 mM
EDTA, 10 mM MgCl2
(pH 7.4)
50 mM Tris–HCl (pH
7.4)
Plate
GF/C
[3H]DPCPX 2 nM
10 lM (R)-PIA
GF/C
[3H]ZM241385 3 nM
50 lM NECA
GF/B
[3H]DPCPX 35 nM
400 lM NECA
GF/B
[3H]NECA 30 nM
100 lM (R)-PIA
Radioligand
Non-specific
binding
Incubation
25 °C/60 min
25 °C/30 min
25 °C/30 min
25 °C/180 min

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2112
the radioligand for A1 and A2B, [3H]ZM241385 (21 Ci/
mmol; Tocris, Madrid, Spain) for A2A and [3H]NECA
(15.3 Ci/mmol; NEN-Perkin-Elmer Life Sciences, Ma-
drid, Spain) for A3. The inhibition constant (Ki) of each
compound was calculated by the expression: Ki = IC50/
(1 + (C/KD)); where IC50 is the concentration of com-
pound that displaces the binding of radioligand by
50%, C is the free concentration of radioligand and
KD is the apparent dissociation constant of each radio-
ligand. The percentage of displacement of specific bind-
˚
spacing was equal to 0.5 A, the smoothing window of
the correlograms was set to 0.8, and the size of the cor-
relograms was automatically established by the pro-
gram. The optimum number of nodes and weight of
the field was set empirically to 80 nodes and 35% weight.
These numbers deviate from the default values of 100
nodes and 50% weight, which is due to the relatively
small size of the molecules. A soft variable selection
method consisting on one fractional factorial design
(FFD) was also applied in all models. Chemometric
analyses were performed with the tools included in Al-
mond 3.3 for principal component analysis, PCA,58
and partial least squares regression, PLS.59 The optimal
number of latent variables (LV) for the PLS regression
analyses was selected based on cross-validation proce-
dure (standard leave one our method, LOO).
ing was calculated by the expression:
%
of
displacement = ((BT ꢀ dpm) 100)/(BT ꢀ NSB); where
*
BT is the total binding of the radioligand in the assay,
NSB is the non-specific binding of the radioligand in
the assay and dpm are the radioactive measure obtained
by competing the radioligand binding with a given con-
centration of the test compound.
5.6. Molecular modelling
Acknowledgment
All calculations were run on an AMD-X2 processor run-
ning linux (FC 5) operative system. All the GRid inde-
pendent descriptors (GRIND)49 of this study were
computed and analyzed with the program Almond
3.3.0.54 This procedure for 3D-QSAR studies consists
on three steps. First, a set of molecular interaction fields
(MIFs) are computed on the basis of the Grid force
field,55 using selected chemical probes that account for
the receptor chemical groups. Then, a selection of the
most relevant ligand-probe interaction nodes is done
and the resulting filtered nodes are encoded via a
MACC2 transform into a set of GRIND variables.
These alignment independent descriptors represent geo-
metrical relationships between pairs of the selected MIF
nodes, which are usually represented with the so-called
‘correlograms’, where the product of the interaction en-
ergy is plotted against the euclidean distance for a given
pair of nodes. The correlogram is composed from ‘auto-
correlograms’ (i.e., the node pairs belong to one MIF)
and ‘cross-correlograms’ (i.e., the node pairs belong to
two different MIFs). Additional advantage of the meth-
od is that the points in the correlogram can be decoded
back into the original MIF descriptors, thus allowing
analysis and interpretation of the correlogram in terms
of molecular interaction capabilities.
Financial support from Almirall Laboratories (Barce-
lona) is gratefully acknowledged.
References and notes
1. Von Lubitz, D. K. J. E.; Jacobson, K. A. Behavioral
Effects of Adenosine Receptors Stimulation. In Adenosine
and Adenine Nucleotides: From Molecular Biology to
Integrative Physiology; Bellardinelli, L., Pelleg, A., Eds.;
Kluwer: Norwell, 1995; pp 489–498.
2. Ralevic, V.; Burnstock, G. Pharmacol. Rev. 1998, 50, 413.
3. Hess, S. Expert Opin. Ther. Pat. 2001, 11, 1533.
4. Fredholm, B. B.; Arslan, G.; Haldner, L.; Kull, G.;
Schulte, G.; Wasserman, W.; Nauyn-Schmied, A. Arch.
Pharmacol. 2000, 362, 364.
5. Fredholm, B. B.; Ijzerman, A. P.; Jacobson, K. A.; Klotz,
K. N.; Linden, J. Pharm. Rev. 2001, 53, 527.
6. Baraldi, P. G.; Cacciari, B.; Romagnoli, R.; Spalluto, G.
Expert Opin. Ther. Pat. 1999, 9, 515.
7. Muller, C. E. Curr. Top. Med. Chem. 2003, 3, 445.
8. Muller, C. E. Curr. Med. Chem. 2000, 7, 1269.
9. Merighi, S.; Mirandola, P.; Varani, K.; Gessi, S.; Leung,
E.; Baraldi, P. G.; Tabrizi, M. A.; Borea, P. A. Pharmacol.
Ther. 2003, 100, 31.
10. Baraldi, P. G.; Borea, P. A. Trends Pharmacol. Sci. 2000,
21, 456.
11. Baraldi, P. G.; Cacciari, B.; Romagnoli, R.; Meringhi, S.;
Verani, K.; Borea, P. A.; Spalluto, P. Med. Res. Rev. 2000,
20, 103.
12. Baraldi, P. G.; Bovero, A.; Fruttarolo, F.; Romagnoli, R.;
Tabrizi, M.; Aghazadeh, P.; Delia, V.; Katia, B.; Borea, P.
A.; Moorman, A. R. Bioorg. Med. Chem. 2003, 11,
4161.
13. Beukers, M. W.; Chang, L. C. W.; von Kuenzel, J. K.;
Mulder-Krieger, Th.; Spanjersberg, R. F.; Brussee, J.;
IJzerman, A. P. J. Med. Chem. 2004, 47, 3707.
14. Harada, H.; Asano, O.; Hoshino, Y.; Yoshikawa, S.;
Matsukura, M.; Kabasawa, Y.; Niijima, J.; Kotake, Y.;
Watanabe, N.; Kawata, T.; Inoue, T.; Horizoe, T.;
Yasuda, N.; Minami, H.; Nagata, K.; Murakami, M.;
Nagaoka, J.; Kobayashi, S.; Tanaka, I.; Abe, S. J. Med.
Chem. 2001, 44, 170.
The 3D structures needed for the GRIND computations
were automatically generated with the program CORI-
NA56 from the 2D SDfile of the molecules. Given that
the present series share a common rigid structure with
only one point of diversity (i.e., substituent in position
4), the conformations obtained from CORINA were
used directly for the calculation of the descriptors. To
account for the most probable conformation of the 2-
furyl ring, a torsion scan for the bond connecting the
triazol and furane rings was performed with GAUSS-
IAN9857 (keywords HF/6-31G* SCAN). Different com-
binations of the GRID probes O, N1 and DRY and the
recently reported TIP probe50 were initially considered
in different calculations, although not all the probes
were considered in the final models (see Section 3). Most
of the Almond parameters were set to default values, for
example, the ALMD directive was equal to 1, the grid
15. Ozola, V.; Thorand, M.; Diekmann, M.; Qurishi, R.;
Schumacher, B.; Jacobson, K. A.; Muller, C. E. Bioorg.
Med. Chem. 2003, 11, 347.

´
A. Martınez et al. / Bioorg. Med. Chem. 16 (2008) 2103–2113
2113
16. Sato, H.; Nagashima, K.; Nomura, H.; Karasawa, A.;
Nishikawa, S. J. Toxicol. Path. 1999, 12, 21.
17. Kozo, Y.; Hideaki, K.; Kiyoshi, S.; Akira, K. Jpn. J.
Pharmacol. 1994, 65, 167.
18. Hideaki, M.; Akira, K.; Kazuhiro, K. J. Pharmacol. Exp.
Ther. 1993, 266, 200.
42. Colotta, V.; Catarzi, D.; Varano, F.; Cechi, L.; Filacchi-
oni, G.; Martini, C.; Trincavelli, L.; Lucacchini, A.
J. Med. Chem. 2000, 43, 3118.
43. Catarzi, D.; Colotta, V.; Varano, C.; Filacchioni, G.;
Martini, C.; Trincavelli, L.; Lucacchini, A. Farmaco 2004,
59, 71.
19. Ochi, M.; Shiozaki, S.; Kase, H. Neuroscience 2004, 127,
223.
20. Kim, J. S.; Sohn, Y. H. Adv. Behav. Biol. 2002, 51, 365.
21. Jacobson, K. A.; Ijzerman, A. P.; Linden, J. Drug Dev.
Res. 1999, 47, 45.
22. Fozard, J. R. Curr. Opin. Pharmacol. 2003, 3, 264.
23. Feostikov, I.; Baggioni, I. Pharmacol. Rev. 1997, 49, 381.
24. Martinez-Tica, J. F.; Zornow, M. H. Brain Res. 2000, 872,
110.
25. Di Napoli, P.; Contegiacomo, G.; Taccardi, A. A.; Maggi,
A.; Di Muzio, M.; Statile, D.; Giuliani, P.; Di Iorio, P.;
Barsotti, A. J. Clin. Basic Cardiol. 1999, 2, 99.
26. Hamilton, H. W.; Ortwine, D. F.; Worth, D. F.; Badger,
E. W.; Bristol, J. A.; Bruns, R. F.; Haleen, S. J.; Steffen, R.
P. J. Med. Chem. 1985, 28, 1071.
44. Colotta, V.; Catarzi, D.; Varano, C.; Romana, F.; Lenzi, O.;
Filacchioni, G.; Trincavelli, L.; Martini, C.; Trincavelli, F.;
Deflorian, F.; Moro, S. J. Med. Chem. 2004, 47, 3580.
45. Catarzi, D.; Colotta, V.; Varano, C.; Lenzi, O.; Filacchi-
oni, G.; Trincavelli, L.; Martini, C.; Montopoli, C.; Moro,
S. J. Med. Chem. 2005, 48, 7932.
46. Lenzi, O.; Colotta, V.; Catarzi, D.; Varano, C.; Filacchi-
oni, G.; Martini, C.; Trincavelli, L.; Ciampi, O.; Varani,
K.; Marighetti, F.; Morizzo, E.; Moro, S. J. Med. Chem.
2006, 49, 3916.
47. Carotti, A.; Stefanacchi, A.; Ravin˜a, E.; Sotelo, E.; Loza,
M. I.; Cadavid, M. I.; Centeno, N. B.; Nicolotti, O. Eur. J.
Med. Chem. 2004, 39, 879.
48. Carotti, A.; Cadavid, M. I.; Centeno, N. B.; Esteve, C.;
´
Loza, M. I.; Martınez, A.; Nieto, R.; Ravin˜a, E.; Sanz, F.;
Segarra, V.; Sotelo, E.; Stefanacchi, A.; Vidal, B. J. Med.
Chem. 2006, 49, 282.
27. Doichinova, I. A.; Natcheva, R. N.; Mihailova, D. N. Eur.
J. Med. Chem. 1994, 29, 133.
´
28. Hayallah, A. M.; Sandoval-Ramırez, J.; Reith, U.; Scho-
bert, U.; Preiss, B.; Schumacher, B.; Daly, J. W.; Muller,
¨
49. Pastor, M.; Cruciani, G.; McLay, I.; Pickett, S.; Clementi,
S. J. Med. Chem. 2000, 43, 3233.
C. E. J. Med. Chem. 2002, 45, 1500.
50. Fontaine, F.; Pastor, M.; Sanz, F. J. Med. Chem. 2004, 47,
2805.
29. Baraldi, P. G.; Tabrizi, M. A.; Preti, D.; Bovero, A.;
Romagnoli, R.; Fruttarolo, F.; Zaid, N. A.; Moorman, A.
R.; Varani, K.; Gessi, S.; Merighi, S.; Borea, P. A. J. Med.
Chem. 2004, 47, 1434.
51. Afzelius, L.; Masimirembwa, C. M.; Karlen, A.; Andersson,
T. B.; Zamora, I. J. Comput. Aided Mol. Des. 2002, 16, 443.
52. Carosati, E.; Lemoine, H.; Spogli, R.; Grittner, D.;
Mannhold, R.; Tabarrini, O.; Sabatini, S.; Cecchetti, V.
Bioorg. Med. Chem. 2005, 13, 5581.
53. Bosch, M. P.; Campos, F.; Niubo, I.; Rosell, G.; Diaz, J.
L.; Brea, J.; Loza, M. I.; Guerrero, A. J. Med. Chem. 2004,
47, 4041.
54. Cruciani, G., Fontaine, F., Pastor, M. Almond 3.3.0.
Molecular Discovery, Ltd: Perugia, Italy.
55. Goodford, P. J. J. Med. Chem. 1985, 28, 849.
56. Gasteiger, J. C.; Rudolph, C.; Sadowski, J. Tetrahedron
Comp. Method 1990, 3, 537.
57. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G.
E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.;
Montgomery, Jr., J. A.; Stratmann, R. E.; Burant, J. C.;
Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.;
Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.;
Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.;
Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko,
A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R.
L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P.
M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J.
L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.;
Pople, J.A. Gaussian 98, Revision A.7, Gaussian: Pitts-
burgh PA, 1998.
30. Kim, S. A.; Marshall, M. A.; Melman, N.; Kim, H. S.;
Muller, C. E.; Linden, J.; Jacobson, K. A. J. Med. Chem.
¨
2002, 45, 2131.
31. Pastorin, G.; Das Ros, T.; Spatullo, G.; Deflorian, F.;
Moro, S.; Cacciari, B.; Saraldi, P. G.; Gessi, S.; Varani,
K.; Borea, P. A. J. Med. Chem. 2003, 46, 4278.
32. Maconi, A.; Pastorin, G.; Da Ros, T.; Spatullo, G.; Gao,
Z. G.; Jacobson, K. A.; Baraldi, P. G.; Varani, K.; Moro,
S.; Borea, P. A. J. Med. Chem. 2002, 45, 3579.
33. Colotta, V.; Catarzi, D.; Varano, F.; Cecchi, L.; Filacchi-
oni, G.; Martini, C. Arch. Pharm. Pharm. Med. Chem.
1999, 332, 39.
34. Muller, C. E.; Dieckman, M.; Thorand, M.; Ozola, V.
Bioorg. Med. Chem. Lett. 2002, 12, 501.
35. Bertelli, L.; Biagi, G.; Giorgi, I.; Manera, C.; Livi, O.;
Scartoni, V.; Betti, L.; Giannaccini, G.; Trincavelli, L.;
Barili, P. L. Eur. J. Med. Chem. 1998, 33, 113.
36. Colotta, V.; Cecchi, L.; Catarzi, D.; Conti, G.; Filacchi-
oni, G. Farmaco 1993, 48, 1051.
37. Calotta, V.; Catarzi, D.; Varano, F.; Cecchi, L.; Filacchioni,
G.; Galli, A.; Costagli, C. Arch. Pharm. 1997, 330, 387.
38. Catarzi, D.; Cecchi, L.; Colotta, V.; Melani, F.; Filacchi-
oni, G. Farmaco 1993, 48, 1065.
39. Colotta, V.; Cecchi, L.; Filacchioni, G.; Martini, C.; Tacchi,
P.; Lucacchini, A. Eur. J. Med. Chem. 1995, 30, 133.
40. Colotta, V.; Catarzi, D.; Varano, F.; Cecchi, L.; Filacchi-
oni, G.; Martini, C.; Trincavelli, L.; Lucacchini, A.
J. Med. Chem. 2000, 43, 1158.
58. Carey, R. N.; Wold, S.; Westgard, J. O. Anal. Chem. 1975,
47, 1824.
59. Hoskuldsson, A. J. Chemom. 1988, 2, 211.
41. Catarzi, D.; Cecchi, L.; Colotta, V.; Filacchioni, G.;
Martini, C.; Tacchi, P. J. Med. Chem. 1995, 38, 1330.