1-Substituted pyrazolo[1,5-c]quinazolines as novel Gly/NMDA receptor antagonists: Synthesis, biological evaluation, and molecular modeling study

Article abstract of DOI:10.1016/j.bmc.2005.07.010

A new set of 5,6-dihydro-pyrazolo[1,5-c]quinazoline-2-carboxylates (2-18), bearing different substituents (COOEt, Cl, Br, CH₃, and COOH) at position-1, were synthesized in order to investigate the influence of various groups at this specific position on Gly/NMDA receptor affinity and/or selectivity. All the herein reported compounds were evaluated for their binding at the Gly/NMDA, AMPA, and KA receptors. Some selected compounds were also tested for their functional antagonistic activity at both the AMPA and NMDA receptor-ion channels. The results obtained in this study have highlighted that a C-1 lipophilic substituent on the pyrazolo[1,5-c]quinazoline-2-carboxylate core shifts selectivity toward the Gly/NMDA receptor, while a C-1 anionic carboxylate residue is able to increase affinity toward this receptor subtype. In particular, the 2-carboxylic acids 15 and 16, bearing a chlorine atom at position-1, are not only potent (K_i = 0.18 and 0.16 μM, respectively), but also highly Gly/NMDA versus AMPA selective (selectivity ratio > 500). Furthermore, the 1,2-dicarboxylic acids 13 and 14 are endowed with the highest Gly/NMDA receptor binding activity (K_i = 0.09 and 0.059 μM, respectively), among the pyrazoloquinazoline series of derivatives. A molecular modeling study has been carried out to better understand receptor affinity and selectivity of these new pyrazoloquinazoline derivatives.

Full text of DOI:10.1016/j.bmc.2005.07.010

Bioorganic & Medicinal Chemistry 13 (2005) 5536–5549
1-Substituted pyrazolo[1,5-c]quinazolines as novel
Gly/NMDA receptor antagonists: Synthesis, biological
evaluation, and molecular modeling study
Flavia Varano,a,* Daniela Catarzi,a Vittoria Colotta,a Francesca Romana Calabri,a
Ombretta Lenzi,a Guido Filacchioni,a Alessandro Galli,b Chiara Costagli,b
Francesca Deflorianc and Stefano Moroc
a
`
Dipartimento di Scienze Farmaceutiche, Universita degli Studi di Firenze, Polo Scientifico, via Ugo Schiff,
6, 50019 Sesto Fiorentino (FI), Italy
b
`
Dipartimento di Farmacologia Preclinica e Clinica, Universita degli Studi di Firenze, viale G. Pieraccini, 6, 50134 Firenze, Italy
Molecular Modeling Section, Dipartimento di Scienze Farmaceutiche, Universita di Padova, via Marzolo 5, I-35131 Padova, Italy
c
`
Received 18 February 2005; revised 29 June 2005; accepted 1 July 2005
Abstract—A new set of 5,6-dihydro-pyrazolo[1,5-c]quinazoline-2-carboxylates (2–18), bearing different substituents (COOEt, Cl, Br,
CH3, and COOH) at position-1, were synthesized in order to investigate the influence of various groups at this specific position on
Gly/NMDA receptor affinity and/or selectivity. All the herein reported compounds were evaluated for their binding at the Gly/
NMDA, AMPA, and KA receptors. Some selected compounds were also tested for their functional antagonistic activity at both
the AMPA and NMDA receptor-ion channels.
The results obtained in this study have highlighted that a C-1 lipophilic substituent on the pyrazolo[1,5-c]quinazoline-2-carboxylate
core shifts selectivity toward the Gly/NMDA receptor, while a C-1 anionic carboxylate residue is able to increase affinity toward this
receptor subtype. In particular, the 2-carboxylic acids 15 and 16, bearing a chlorine atom at position-1, are not only potent
(Ki = 0.18 and 0.16 lM, respectively), but also highly Gly/NMDA versus AMPA selective (selectivity ratio > 500). Furthermore,
the 1,2-dicarboxylic acids 13 and 14 are endowed with the highest Gly/NMDA receptor binding activity (Ki = 0.09 and
0.059 lM, respectively), among the pyrazoloquinazoline series of derivatives.
A molecular modeling study has been carried out to better understand receptor affinity and selectivity of these new pyrazoloquinaz-
oline derivatives.
Ó 2005 Elsevier Ltd. All rights reserved.
1. Introduction
cell damage and death. This excitotoxic mechanism is
thought to be involved in several neurodegenerative dis-
orders, such as ParkinsonÕs, HuntingtonÕs, and Alzhei-
merÕs diseases, as well as in brain ischemia and
epilepsy. Therefore, a pharmacological intervention
with iGluRs antagonists has been suggested to have a
potential therapeutic benefit.1,3,4 In particular, the
non-competitive NMDA antagonists acting at the gluta-
mate coagonist glycine binding site on the NMDA
receptor complex (Gly/NMDA)5 have been recognized
as potential therapeutic agents for the above mentioned
disorders.6–10 In fact, selective Gly/NMDA receptor
antagonists are endowed with a better side-effect profile
than competitive and other non-competitive NMDA
receptor antagonists.1,10 Furthermore, it is worth men-
tioning that increasing evidence supports the hypothesis
that the NMDA receptor plays a pivotal role in the
Glutamate, the most important excitatory neurotrans-
mitter in mammalian brain, acts through the activation
of either ionotropic (iGluRs) (NMDA, AMPA, and
Kainate subtypes) or metabotropic (mGluRs)
receptors.1,2
It is now widely accepted that over-activation of iGluRs,
especially of the NMDA receptor, is one of the major
causes of Ca2+ overload in cells, potentially leading to
Keywords: Ionotropic glutamate receptors; Gly/NMDA receptor ant-
agonists; AMPA receptor antagonists; Pyrazoloquinazoline
derivatives.
*
Corresponding author. Tel.: +39 55 457 3721; fax: +39 55 457
3780; e-mail: flavia.varano@unifi.it
0968-0896/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.07.010

F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
5537
development and maintenance of pain hypersensitivity.
R1
COOR2
N
Following such a hypothesis, Gly/NMDA receptor
antagonists should also be effective for the treatment
of chronic pain.6
R9
R8
N
N
O
In recent years, some research in our laboratory has
been directed toward the synthesis of tricyclic heteroaro-
matic systems as Gly/NMDA and/or AMPA receptor
antagonists.11–18 Structure–activity relationship (SAR)
studies on some of these tricyclic derivatives have point-
ed out that introduction of suitable substituents on pre-
cise positions of the benzofused moiety shifts selectivity
toward AMPA or Gly/NMDA receptors.13,14,16–18
H
2-18
R1 = COOEt, Cl, Br, CH3, COOH
R2 = Et, H
R8, R9 = H, Cl
Chart 2. C-1 substituted pyrazolo[1,5-c]quinazoline-2-carboxylates.
As a part of this research project, we reported the syn-
thesis and binding activity at Gly/NMDA, AMPA,
and kainate (KA) receptors of some 5,6-dihydro-pyraz-
olo[1,5-c]quinazoline-2-carboxylates (1 PQZ) (Chart 1)
bearing various substituents on the benzofused
moiety.17
2. Chemistry
Compounds 2–18 were synthesized as depicted in
Schemes 1–4. The diethyl 1,2-dicarboxylate derivatives
2–4 were obtained starting from the commercially avail-
able isatine 19 and its 5-chloro- and 5,6-dichloro-ana-
logues 20–21,19,20 respectively, which when treated
with an equimolar amount of p-toluenesulfonohydraz-
ide, yielded the corresponding sulfonylhydrazones 22–
2421,22 (Scheme 1). The latter, by hydrolysis in alkaline
medium, led to the 3-diazo-1,3-dihydro-indol-2-one
derivatives 25–27,21,22 which yielded the tricyclic diethyl
pyrazolo[1,5-c]quinazoline-1,2-dicarboxylates 2–4 after
The binding results showed that some of these com-
pounds are potent and/or selective AMPA or Gly/
NMDA receptor antagonists, depending on the benzo-
fused substitutions. In particular, in accordance with
our previously reported data,13,14,16 the presence of both
a chlorine atom at position-8 and a N3-nitrogen-con-
taining heterocycle at position-9 is required for potent
and highly selective AMPA receptor antagonists. More-
over, some Gly/NMDA receptor selectivity is obtained
when the benzofused moiety of the PQZ derivatives is
unsubstituted or when it bears chlorine atom(s).
reaction
with
acetylenedicarboxylate.
an
excess
of
diethyl
Partial hydrolysis of esters 2–4 at room temperature in
the presence of a hydro-methanolic solution of KOH
gave the corresponding 1-ethoxycarbonyl-pyrazoloqui-
nazoline-2-carboxylic acid derivatives 9–11. On the
contrary, by refluxing esters 2–4 in the presence of a
hydro-ethanolic solution of KOH, the corresponding
pyrazoloquinazoline-1,2-dicarboxylic acid derivatives
12–1423 were obtained. The 1-ethoxycarbonyl-2 carbox-
ylic acid structure of compounds 9–11 were assigned by
transforming them into their corresponding decarboxyl-
ate analogues. Thus, upon heating of 9–11 at a temper-
ature just over their melting points, the corresponding
ethyl 5,6-dihydro-5-oxo-pyrazolo[1,5-c]quinazoline-1-
carboxylates 28–30 were isolated. The 1-ethyl carboxyl-
ate structure of 28–29 was unambiguously assigned,
since the corresponding ethyl pyrazolo[1,5-c]quinazo-
line-2-carboxylate regio-isomers had already been
obtained by an univocal synthesis.17 The 1-ethyl carbox-
ylate structure of 30 was attributed on the basis of
the 1H NMR signal of the pyrazole proton: in fact,
the chemical shift value of this proton (8.45 ppm) is very
close to that of the H-2 proton of 28 and 29 (8.43,
8.44 ppm, respectively).
Continuing with the study on PQZ derivatives, in the
present paper we report the synthesis of a new set of
1-substituted pyrazolo[1,5-c]quinazoline-2-carboxylates
2–18 (Chart 2), in order to investigate the influence of
different groups (COOEt, Cl, Br, CH3, and COOH) at
position-1 of the PQZ tricyclic system on Gly/NMDA
receptor affinity and/or selectivity.
Since our goal was to obtain potent and/or selective Gly/
NMDA receptor antagonists, the herein reported
pyrazoloquinazoline-2-carboxylates are either unsubsti-
tuted on the benzofused moiety or bear chlorine
atom(s). Moreover, a molecular modeling study has
been carried out to better understand receptor affinity
and selectivity of these new pyrazoloquinazoline
derivatives.
COOR2
R9
R8
N
O
N
Reaction of ethyl 8-chloro-4,5-dihydro-pyrazolo[1,5-
c]quinazoline-2-carboxylates17 with sulfuryl chloride in
glacial acetic acid gave the 1-chloro derivative 5, which
was easily transformed into its corresponding acid 15
(Scheme 2).
N
H
1 ( PQZ )
R2 = Et, Me, H
R8 = H, Cl
R9 = H, Cl,
N
N
N
The ethyl 8,9-dichloro-5,6-dihydro-pyrazolo[1,5-c]qui-
nazoline-2-carboxylates 6 and 7, 1-chloro and 1-bromo
substituted, respectively (Scheme 3) were prepared
Chart 1. Previously reported pyrazolo[1,5-c]quinazoline-2-carboxy-
lates.

5538
F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
O O
N NH S
CH3
O
R5
R6
R5
R6
a
O
O
N
N
H
H
19 R5=R6=H
20 R5=H R6=Cl
21 R5=R6=Cl
22 R5=R6=H
23 R5=H R6=Cl
24 R5=R6=Cl
b
COOEt
EtOOC
N2
N
O
R5
R6
R9
R8
c
N
O
N
H
N
H
25 R5=R6=H
26 R5=H R6=Cl
27 R5=R6=Cl
2-4
e
d
COOH
EtOOC
COOH
HOOC
N
R9
N
R9
R8
N
N
R8
N
O
N
O
H
H
12-14
9-11
f
R8
R9
H
H
2, 9, 12, 28
3, 10, 13, 29
4, 11, 14, 30
H
Cl
Cl
EtOOC
Cl
N
R9
R8
N
N
O
H
28-30
Scheme 1. Reagents and conditions: (a) p-toluenesulfonohydrazide, MeOH; (b) 0.35 M NaOH; (c) diethyl acetylenedicarboxylate, anhydrous
toluene; (d) 1.5 M KOH, MeOH, room temperature; (e) 1.5 M KOH, EtOH, reflux; (f) heating over the melting points.
and 30 was obtained. We were unable to separate 31 from
30 with satisfactory yield; thus their combined mixture
was used for the next step. Reaction of this mixture with
either sulfuryl chloride or bromine in glacial acetic acid
gave a solid in which only the 1-chloro- or 1-bromo deriv-
atives 6 and 7 were present, respectively; in both, cases the
less reactive ethyl pyrazolo[1,5-c]quinazoline-1-carboxyl-
ate 30 was recovered unmodified in the mother liquor. The
esters 6 and 7 were then hydrolyzed in alkaline medium to
their corresponding acids 16 and 17.
COOEt
COOR
Cl
N
N
a
N
N
Cl
N
O
Cl
N
O
H
H
5 R=Et
15 R=H
b
Scheme 2. Reagents and conditions: (a) SO2Cl2, glacial acetic acid; (b)
1.5 M KOH, MeOH, reflux.
The tricyclic ethyl 1-methyl-pyrazoloquinazoline-2-car-
boxylate derivative 824,25 (Scheme 4) was obtained in
mixture (1:1, H NMR analysis) with its regio-isomer
1
starting from the 5,6-dichloro-3-diazoindole derivative
27 obtained as above and depicted in Scheme 1.
ethyl 2-methyl-pyrazoloquinazoline-1-carboxylate 32,
by reacting the 3-diazoindol-2-one intermediate 2521
with an excess of ethyl 2-butinoate. Pure 8 and 32 were
then obtained by column chromatography.
By reacting 27 with an excess of ethyl propiolate, a mix-
ture (1:1, 1H NMR analysis) of the two regio-isomers 31

F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
COOEt
5539
EtOOC
N2
N
O
N
Cl
Cl
Cl
Cl
Cl
Cl
a
N
N
O
N
N
H
N
H
O
H
27
31
30
b, c
COOR2
R1
R1
Cl
Br
Cl
Br
R2
Et
Et
H
6
7
16
17
N
Cl
Cl
N
N
O
H
H
6-7
16-17
d
Scheme 3. Reagents and conditions: (a) ethyl propiolate, anhydrous toluene; (b) SO2Cl2, glacial acetic acid; (c) Br2, glacial acetic acid; (d) 1.5 M
KOH, MeOH, reflux.
COOR
CH3
H3C
EtOOC
10
N2
10
N
N
a
N
N
O
N
N
O
N
H
O
H
H
25
32
8 R=Et
b
18 R=H
Scheme 4. Reagents and conditions: (a) ethyl 2-butinoate, toluene; (b) 1.5 M KOH, MeOH, reflux.
The structures of the two isomers 8 and 32 were assigned
on the basis of their 1H NMR spectra, using as key tool the
signal of the proton at position-10, which is easily identi-
fied in this class of tricyclic derivatives. In fact, the H-10 is,
in general, the most deshielded aromatic proton. In both
1H NMR spectra of isomers 8 and 32, the H-10 signal ap-
pears as a doublet, but with a different chemical shift:
while the H-10 signal of 8 appears at 8.08 ppm, that of
32 is at 9.01 ppm, due to the paramagnetic effect of the
carbonyl function of the ester group at position-1.
tricyclic system is important for shifting selectivity to-
ward the Gly/NMDA receptor. In fact, derivatives 15
and 16, bearing a chlorine atom at position-1, are the
most Gly/NMDA versus AMPA selective compounds
among the PQZ series, showing a selectivity ratio com-
parable to that of DCKA (selectivity ratio > 500). More-
over, 15 and 16 also show good Gly/NMDA receptor
affinity (Ki = 0.18 and 0.16 lM, respectively). In con-
trast, derivatives 13 and 14, bearing a carboxylic acid
group at position-1, are scarcely Gly/NMDA selective
(selectivity ratio ffi 12), but they are the most potent
Gly/NMDA receptor antagonists among the PQZ series
(Ki = 0.09 and 0.059 lM, respectively).
Finally, by treatment of 8 in alkaline medium, the corre-
sponding acid 1825 was obtained.
This study confirms that the presence of an anionic car-
boxylate residue at position-2 of the pyrazoloquinazo-
line tricyclic system, being able to bind with a proton
donor site of the receptors, is an important feature for
anchoring at both AMPA and Gly/NMDA recep-
tors.13,14,16–18 In fact, the 2-carboxylic acids 9–11, 15–18
are more active than their corresponding 2-carboxylate
esters 2–8 at both receptors, with the exception of the
1-chloro-2-carboxylic acid derivatives 15 and 16, which
like their corresponding esters 5 and 6, are inactive at
the AMPA receptor, and the 1-bromo-2-carboxylic acid
17, which is less AMPA active than its ethyl ester 7.
3. Results and discussion
The pyrazoloquinazolines 2–18 were tested for their
ability to displace tritiated glycine, AMPA, and KA
from their specific binding sites in rat cortical mem-
branes. The binding data are shown in Table 1, together
with those of the previously reported C-1 unsubstituted
derivatives 1a–e,17 included as reference compounds.
Moreover, the binding data of the well-known Gly/
NMDA antagonist 5,7-dichlorokynurenic acid (DCKA)
are also reported.
In general, the binding results suggest that the nature of
the substituent at position-1 of the pyrazoloquinazoline
As regards the benzofused substitutions, it is important
to note that the benzo-unsubstituted derivatives 2, 8, 9,

5540
F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
Table 1. Displacement of [3H]glycine, [3H]AMPA, and [3H]Kainate bindinga
R2
N
R1
R9
R8
N
N
O
H
1a-e; 2-18
Compound
R1
R2
R8
R9
Ki (lM)b or I%c
IC50 (lM)d or I%c
[3H]glycine
[3H]AMPA
[3H]KA
1ae
1be
1ce
1de
1ee
2
H
COOEt
COOEt
COOH
COOH
COOH
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOH
COOH
COOH
COOH
COOH
COOH
COOH
COOH
COOH
COOH
H
H
H
H
H
Cl
H
H
Cl
H
Cl
Cl
H
H
H
Cl
H
H
Cl
H
Cl
Cl
H
33.3 7.0
26.5 4.4
1.41 0.3
0.48 0.04
0.16 0.04
2.95 0.25
3.4 0.4
42 6.8
72 2.3
12.4 2.5
2.3 0.4
2.4 0.8
50 6.0
30%
12%
15%
H
Cl
H
H
26%
H
Cl
Cl
H
60 3.0
41 3.0
27%
H
COOEt
COOEt
COOEt
Cl
3
Cl
Cl
Cl
Cl
Cl
H
10%
0%
4
50 13
25%
15%
5
25%
22%
7.1 0.7
30%
1.15 0.2
6%
9%
6
Cl
Br
10%
39
7
5
8%
13%
8
CH3
COOEt
COOEt
COOEt
COOH
COOH
COOH
Cl
19%
9
H
15.2 1.3
1.6 0.3
2.5 0.1
34.8 4.0
1.1 0.2
0.71 0.08
16%
48%
10
11
12
13
14
15
16
17
18
DCKA
Cl
Cl
H
0.86 0.2
0.22 0.04
0.54 0.12
0.09 0.01
0.059 0.01
0.18 0.04
0.16 0.01
4.0 0.6
78 7.0
43 3.0
47%
Cl
Cl
Cl
Cl
Cl
H
11.5 1.4
19.5 11
100
8
Cl
Br
47%
91 15
92 8.0
100 15
47.5 4.2
5%
CH3
0.78 0.13
0.09 0.02
33%
8%
a The tested compounds were dissolved in 50% DMSO and then diluted with the appropriate buffer.
b Inhibition constant (Ki) values were means SEM of three or four separate determinations in triplicate.
c Percentage of inhibition (I%) of specific binding at 100 lM concentration.
d Concentrations necessary for 50% inhibition (IC50). The IC50 values were means SEM of three or four separate determinations in triplicate.
e Ref. 17.
12, and 18 are endowed with micromolar or submicrom-
olar Gly/NMDA receptor binding affinity, the only
exception being the 1-methyl 2 ethyl ester 8, which is
inactive. Based on these data, the presence of a substitu-
ent on the benzofused moiety is not essential for the
anchoring of PQZ derivatives to the Gly/NMDA bind-
ing site. However, in accordance with our previous
data,13,17 an 8-chloro-, or better still an, 8,9-dichloro
substitution is profitable for Gly/NMDA receptor affin-
ity (compare 10 and 11 with 9, and 12 and 13 with 14),
the only exception being the ethyl 8,9-dichloro 1,2-
dicarboxylate 4, which shows a markedly reduced Gly/
NMDA binding activity compared to those of its corre-
sponding unsubstituted and 8-chloro substituted deriva-
tives 2 and 3, respectively.
NH lactam acidity, thus reinforcing the essential hydro-
gen-bond receptor–ligand interaction at this level.26–28
Furthermore, the 8,9-dichloro derivatives 11 and 14,
are equi-active compared to the 8-chloro substituted
10 and 13, respectively, at the AMPA receptor.
The presence of a C-1 ethyl carboxylate group on the
PQZ core has contrasting effects. The ethyl 1,2-dicarb-
oxylate derivatives 2 and 3, compared to their C-1
unsubstituted parent compounds 1a and 1b, respective-
ly, are 8- and 11-fold more active at the Gly/NMDA
receptor and also much more Gly/NMDA versus
AMPA selective (selectivity ratio = 17 and 30, respec-
tively). In contrast, the presence of a C-1 ethyl carboxyl-
ate group in the 2-carboxylic acid derivatives 9–11 does
not increase the affinity for either receptor subtype; in
fact, the Gly/NMDA and AMPA receptor binding
activities of 9–11 are comparable to those of their parent
C-1 unsubstituted derivatives 1c–e, respectively. The
behavior of 9–11, especially toward the Gly/NMDA
receptor, was unexpected. In fact, the micromolar Gly/
NMDA binding activities of the ethyl 1,2-dicarboxylates
2 and 3 had led us to hypothesize that a C-1 ethyl car-
boxylate group, able to act as hydrogen bond acceptor,
On the contrary, the presence of a chlorine atom at
position-8 is necessary for an effective AMPA recep-
tor–ligand interaction. Indeed, the benzo-unsubstituted
2-carboxylic acids 9 and 12 are about 10-fold and
30-fold less active than their corresponding 8-chloro-
substituted 2-carboxylic acids 10 and 13, respectively.
These data confirm that the 8-chloro substituent, due
to its electron-withdrawing effect, could increase the

F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
5541
could have engaged a new hydrogen bond interaction
with the Gly/NMDA receptor site, thus reinforcing the
receptor–ligand interaction. Trying to clarify these con-
trasting results, we have introduced at position-1 of our
pyrazoloquinazoline scaffold, a more powerful and less
hindered hydrogen bond acceptor group than the ethyl
carboxylate. Thus we synthesized compounds 12–14
bearing a carboxylic acid function at position-1. These
compounds are from 3- to 5-fold more active at the
Gly/NMDA receptor than the corresponding C-1
unsubstituted derivatives 1c–e, respectively, and from
2- to 9-fold more active than their corresponding C-1 es-
ters 9–11. In particular, the 1,2-dicarboxylic acid deriv-
atives 13 and 14 (Ki = 0.09 and 0.059 lM,
respectively), are the most potent Gly/NMDA receptor
antagonists among the PQZ series, showing an inhibi-
tion constant value comparable to that of DCKA
(Ki = 0.09 lM). These data indicate that a C-1 hydrogen
bond acceptor substituent, the best being a carboxylic
acid function, positively affects the anchoring of PQZ
derivatives to the binding site of the Gly/NMDA receptor,
probably through a new hydrogen bond interaction.
However, C-1 hydrogen bond acceptor groups, like ester
or carboxylic acid, are unable to distinguish the Gly/
NMDA receptor from the AMPA one. In fact, with the
exception of the benzo-unsubstituted 12, which is 70-fold
Gly/NMDA versus AMPA selective, compounds 9–11
and 13–14 show a selectivity toward the Gly/NMDA
receptor comparable to that of their corresponding C-1
unsubstituted derivatives 1c–e, respectively.
NMDA versus AMPA selective (selectivity ratio = 60).
In fact, 18 shows a comparable Gly/NMDA receptor
binding activity with respect to 1c and, due to the pres-
ence of a C-1 lipophilic substituent, it is 4-fold less active
at the AMPA receptor than 1c. Unfortunately, no data
about the presence of both the methyl moiety at posi-
tion-1 and chlorine atom(s) at position-8 or 8,9 are
available. In fact, all attempts to synthesize the 8-chloro
and 8,9-dichloro analogues of 18 by reacting the corre-
sponding 3-diazoindole 26–27 with ethyl 2-butinoate
were unsuccessful.
The pyrazoloquinazoline 2–18 are in general, inactive or
active in the high micromolar range at the KA receptor.
As for the AMPA receptor–ligand interaction, in com-
pounds 2–18 the contemporary presence of a chlorine
atom at position-8 and a 2-carboxylic acid group on
the pyrazoloquinazoline core is required for the KA
receptor–ligand interaction. Furthermore, the presence
of a 1-carboxylic acid group (i.e., compounds 13 and
14) led to an increase in KA receptor binding activity
with respect to their C-1 unsubstituted parent com-
pounds 1d and 1e, respectively. These data confirm the
high similarity of the structural requirements of the
binding pockets of AMPA and KA receptors.27,29
Functional antagonism at the NMDA receptor-ion chan-
nel complex was evaluated by assessing the ability of
some selected pyrazoloquinazolines (2, 3, 8, 10, 12–16,
and 18) to inhibit the binding of the channel-blocking
agent [3H]-(+)-MK801 ((+)-5-methyl-10,11-dihydro-
5H-benzo[a,d]cyclohepten-5,10-iminemaleate) in rat cor-
tical membranes incubated with 10 lM glutamate and
0.1 lM glycine.10,30,31 The results, shown in Table 2, indi-
cate that, as in the case of [3H]glycine binding assays,
compound 14 is the most potent tested compound with
an IC50 value of 0.15 0.01 lM higher than that of
DCKA (IC50 = 0.74 0.09 lM).
For this reason we decided to evaluate the effect on Gly/
NMDA receptor selectivity of a lipophilic substituent at
position-1 of the pyrazoloquinazoline tricyclic system.
Introduction of a C-1 chlorine atom positively affects
the Gly/NMDA selectivity. Indeed, while the 1-chloro-
2-carboxylic acids 15 and 16 are completely inactive to-
ward the AMPA receptor, they show Gly/NMDA bind-
ing activities comparable to those of their parent
compounds 1d and 1e. Thus, compounds 15 and 16
are endowed with the highest Gly/NMDA versus
AMPA selectivity (selectivity ratio > 500) among the
PQZ series. Therefore, we hypothesized that a lipophilic
substituent at position-1 on the pyrazoloquinazoline tri-
cyclic system could be important for shifting selectivity
toward the Gly/NMDA receptor. This encouraging re-
sult prompted us to further investigate in this direction.
Thus we synthesized the 1-bromo-2-carboxylic acid 17,
which unexpectedly, is 25-fold less active than 1e at
the Gly/NMDA receptor. Moreover, 17 is also 25-fold
less active than the corresponding 1-chloro derivative
16, probably due to the increased steric hindrance of
the bromine atom with respect to the chlorine one. How-
ever, 17 is 45-fold less active at the AMPA receptor
when compared to 1e, thus confirming the negative ef-
fect of a C-1 lipophilic substituent for the anchoring of
PQZ derivatives at the AMPA receptor. Indeed, the 1-
bromo derivative 17 still shows a good Gly/NMDA ver-
sus AMPA selectivity (selectivity ratio = 25).
Furthermore, compounds 13–15 were also evaluated by
testing their ability to inhibit depolarization induced by
5 lM AMPA or 5 lM NMDA in mouse cortical wedge
Table 2. Inhibition of stimulated [3H]-(+)-MK-801 binding
Compd
IC50 (lM)a or I%b
2
3.2 0.1
4.0 0.4
43%
3
8
10
12
13
14
15
16
18
DCKA
1.1 0.06
0.85 0.008
0.7 0.08
0.15 0.01
1.2 0.15
1.0 0.1
5.6 0.2
0.74 0.09
a Concentration giving 50% inhibition of stimulated [3H]-(+)-MK-801
binding. All assays were carried out in the presence of 10 lM glu-
tamate and 0.1 lM glycine. IC50 values were means SEM of three
or four separate determinations in triplicate.
b Percentage of inhibition (I%) of specific binding at 100 lM
concentration.
Introduction of 1c at position-1 of the lipophilic methyl
moiety, yields compound 18, which is highly Gly/

5542
F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
Table 3. Functional antagonism at NMDA and AMPA sites
explain the experimental affinity data and the selectivity
of PQZ ligands between AMPA and Gly/NMDA
receptors.
Compd
Mouse cortical wedge preparation:
IC50 (lM) versus agonist-induced
depolarizationa
So far, several computational approaches have been
used in order to understand the binding modes of li-
gands on Gly/NMDA receptor. Before the crystal struc-
tures of glycine binding core in different states were
available, pharmacophore34–36 and three-dimensional
homology models37,38 of Gly/NMDA binding site were
used for explaining and rationalizing known structure–
activity relationships of diverse classes of compounds.
In 2003, Baskin et al. proposed CoMFA models of a
set of pyrazolo[1,5-c]quinazolines derivatives for evalua-
tion of binding selectivity between Gly/NMDA and
AMPA receptors.39 Our results showed to be in agree-
ment with most of the previous models, even if the
Gly/NMDA crystallographic structure availability al-
lowed us to perform a more accurate comparison be-
tween the two different receptor binding sites.
NMDA
AMPA
13
14
0.7 0.1
0.5 0.1
1.6 0.1
4.7 0.9
7.1 1.2
6.0 1.0
38 7.0
52 11
15
DCKA
a Concentration that inhibits 50% depolarization (IC50) induced by
5 lM NMDA or AMPA. The IC50 values were means SEM of four
separate determinations.
preparations (Table 3). Among the PQZ series presented
in this article, compounds 13 and 14 were chosen due to
their high Gly/NMDA receptor binding affinity; and
compound 15 for its high selectivity.
The results obtained in these electrophysiological assays
indicate that the inhibitory actions of 13–15 on depolar-
ization induced by NMDA are much higher than those
on AMPA-evoked response, confirming that these deriv-
atives are more potent antagonists at the Gly/NMDA
receptor. Moreover, compounds 13–15 are more potent
Gly/NMDA receptor antagonists than the reference
DCKA.
To simplify the discussion of the computational results
and the comparisons between the two receptors, we pre-
fer to divide the receptor binding sites into key regions,
from sub-site I to sub-site IV (Fig. 1).
As represented in Figure 1, the binding pocket of Gly/
NMDA and AMPA receptors shows some conserved
features important for the binding of the ligands, and
some crucial structural dissimilarities, required for the
selectivity of the ligands between the two receptors.
Sub-site I (ss-I) seems to be the most conserved part of
the receptor binding cleft and the most important in
determining ligand binding affinity on both Gly/NMDA
and AMPA receptors. Sub-sites II–IV seem to affect
receptor selectivity the most. In fact, these regions of
the receptor binding pocket interact with the substitu-
ents R1, R2, R8, and R9 of the PQZ derivatives.
4. Molecular modeling studies
A molecular modeling investigation has been carried out
to support the observed SAR. Molecular docking tech-
niques have been used to explore the binding conforma-
tion of all new synthesized PQZ derivatives. The
published crystallographic structures of glycine-binding
domain S1S2 of NR1 subunit in complex with the antag-
onist DCKA, and of glutamate-binding domain S1S2 of
GluR2 subunit in complex with the antagonist DNQX
(6,7-dinitroquinoxalin-2,3-dione) were used to perform
molecular docking studies of Gly/NMDA and AMPA
receptor, respectively (see Experimental section for
details).32,33 From a structural point of view, the ligand–
binding segments S1 and S2 of NR1 present in the glycine
binding recombinant NR1 fusion protein display a
sequence identity of 22% with the glutamate-binding
domains (S1–S2) of the AMPA receptor. As in bacterial
amino acid binding proteins, segments S1 and S2 form
the globular lobes 1 and 2, which are connected by two
hinge regions. The lobes are separated by a central cleft
that provides the ligand binding pocket , and which is
lined by loop regions. These loops harbor determinants
of glycine site pharmacology. Computational docking
of the glycine site antagonist DCKA and the glutamate-
binding domain antagonist DNQX have been carried
out to verify the quality of our computational docking
protocol. Our docking protocol accurately reproduced
the crystallographic ligand conformations solved in both
receptor binding sites (data not shown).
Sub-site I (ss-I) contains the most important and highly
conserved triad of residues crucial for binding the li-
gands on both receptors: an arginine (R131 in the Gly/
NMDA receptor and R96 in the AMPA receptor), a
threonine (T126 and T91, respectively), and a proline
residue (P124 and P89, respectively). Two H-bond
acceptor groups in the antagonist structure are essential
for the interaction with the above mentioned arginine
and threonine. Molecular docking clearly shows a good
electrostatic complementarity between the structure of
common pyrazolo[1,5-c]quinazoline moiety of PQZ
derivatives and the three amino acids R131/R96, T126/
T91, and P124/P89 in both the ss-I sub-sites (Fig. 2).
Moreover, an aromatic residue is present in both recep-
tors (F92 in Gly/NMDA receptor and Y61 in AMPA
receptor), which can bind through p–p interaction with
the aromatic scaffold of quinazoline ring.
Sub-site II (ss-II) seems to regulate the accommodation
of R8 substituent of the PQZ derivatives. In the Gly/
NMDA receptor, ss-II is formed by the presence of five
highly hydrophobic residues: F16, F92, F250, A226, and
V227. Considering also the small size of this cavity,
small and hydrophobic substituents, such as chlorine
The new series of PQZ derivativeshas been analyzed using
the same reference docking protocol on both receptor
binding sites. Docking information has been used to

F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
5543
Figure 1. Schematic representation of the ligand binding pocket of the Gly/NMDA receptor (panel A) and AMPA receptor (panel B). Crucial amino
acids are highlighted. Sub-sites I–IV are classified. General binding motif of PQZ derivatives is also reported.
atom, can be efficiently allocated in it. On the other
hand, ss-II of the AMPA receptor is a less hydrophobic
pocket formed by Y16, Y61, Y220, T195, and M196 res-
idues. However, the presence of a chlorine atom at the 8-
position guarantees that the antagonist assumes the
right conformation for an efficient interaction with the
ss-I triad RTP (compare derivatives 12 and 13).
E193, and M196 delimit the ss-III cavity of the AMPA
receptor. In comparison, Q13, W223, D224, and V227
delimit the ss-III cavity of the Gly/NMDA receptor.
The presence of a chlorine atom at the 9-position of
PQZ derivatives plays an important role in ligand
binding by stabilizing the conformation that allows the
ligand to interact with the ss-I triad RTP. Consequently,
the double 8,9-dichloro substitution of PQZ derivatives
is very well tolerated at both Gly/NMDA and
AMPA receptors. However, considering the peculiar
shape and hydrophobicity properties of the two ss-III
cavities, the increase of ligand affinity versus Gly/
Sub-site III (ss-III) is deputed to accommodate R9 sub-
stituent. In the AMPA receptor, ss-III is roomier and
more hydrophilic than the corresponding pocket in
Gly/NMDA receptor. Residues L138, T174, L192,

5544
F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
Figure 2. Best docking conformation of PQZ derivatives 14 and 16 obtained by using Gly/NMDA (on the left) and AMPA (on the right)
˚
ligand–receptor binding sites. The side chains of some very important residues in proximity (6 5 A) to the docked PQZ derivatives are highlighted
and labeled. Hydrogen bonding interactions are indicated (dash line).
NMDA receptor is more significant than that versus
AMPA receptor (compare derivatives 13 and 14).
charged carboxylate group and the guanidine moiety of
the crucial arginine of the ss-I pocket, and the strength
of this stabilization strongly depends on the final accom-
modation of each PQZ derivative inside the whole recep-
tor binding cleft (from ss-I to ss-IV).
Finally, sub-site IV (ss-IV) accommodates R1 and R2
substituents. Residues L138, S142, G141, and T143
delimit the ss-IV cavity of the AMPA receptor; V176,
S 179, S180, and V181 delimit the ss-IV cavity of the
Gly/NMDA receptor. As shown by docking studies,
R1 and R2 substituents are positioned very close to a he-
lix domain (helix F) of the S2 unit of both receptors.
Moreover, R2 substituent is also in the proximity of
the crucial arginine residue of the ss-I pocket. Accord-
ingly, the activity of PQZ derivatives increases when
the carboxyl group at R2 position is near to the crucial
arginine residue of the ss-I pocket. This stabilization is
due to an electrostatic interaction between the negatively
Considering R1 substitutions and comparing Gly/
NMDA and AMPA ligand binding pockets, a different
steric control seems to operate. Valine 176 (V176) of
Gly/NMDA receptor is substituted with the bulky leu-
cine 138 (L138) of AMPA receptor. Analyzing our
docking studies, V176 and L138 are positioned in prox-
imity to R1 position. This could explain the reason why
bulky substituents at R1 position of PQZ moiety are not
well tolerated by AMPA receptor binding pocket (see
compound 7 as an example). H-bond acceptor groups,

F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
5545
like ester or carboxylic acid, can interact with S180 of
Gly/NMDA receptor, stabilizing the antagonist–recep-
tor complex. Moreover, the environment created by he-
lix F around R1 and R2 positions is more polar in
AMPA receptor with respect to Gly/NMDA receptor.
In fact, the valine residue (V181) present on Gly/NMDA
receptor is replaced by a threonine (T143) on AMPA
receptor. Accordingly, compounds 15 and 16, bearing
chlorine atom at R1 position, are scarcely effective ver-
sus the more polar environment of AMPA receptor.
Conversely, compounds 13 and 14, bearing a carboxylic
acid group at R1 position, are also effective versus
AMPA receptor, because they form an extra electrostat-
ic interaction between the carboxylic moiety and the NH
group of serine 142 (S142) (Fig. 2).
additional hydrogen bond interaction with a proton
donor site of the receptor. Indeed, the 1,2-dicarboxylic
acids 13 and 14 are the most active compounds at the
Gly/NMDA receptor among the pyrazoloquinazoline
series.
Based on our experimental and computational results,
we have proposed that the interaction of the ligand with
sub-site I is essential for receptor recognition. In addi-
tion, the increase of ligand affinity and selectivity neces-
sarily involves a delicate mechanism of balancing
between steric and electrostatic control, operated by
the interaction of the ligand with sub-sites II–IV. Fur-
ther investigations are in progress to confirm the robust-
ness of this scenario.
5. Conclusions
6. Experimental section
6.1. Chemistry
The synthesis of the herein reported new C-1 substituted
pyrazoloquinazoline 2-carboxylates has allowed us to
further investigate the SAR of tricyclic heteroaromatic
systems as Gly/NMDA receptor antagonists.
Silica gel plates (Merck F254) and silica gel 60 (Merck;
70–230 mesh) were used for analytical and column chro-
matography, respectively. All melting points were deter-
mined on a Gallenkamp melting point apparatus.
Microanalyses were performed with a Perkin–Elmer
260 elemental analyzer for C, H, N, and the results were
within 0.4% of the theoretical values, unless otherwise
stated (Table 4). The IR spectra were recorded with a
Perkin–Elmer Spectrum RX I spectrometer in Nujol
The results of this study show that a C-1 lipophilic
substituent on the pyrazolo[1,5-c]quinazoline-2-carbox-
ylate core shifts selectivity toward the Gly/NMDA
receptor. In fact, the 2-carboxylic acids 15 and 16,
bearing a chlorine atom at position-1, are the most
Gly/NMDA versus AMPA selective compounds
among the pyrazoloquinazoline series (selectivity ra-
tio > 500). Furthermore, this study has highlighted that
a C-1 anionic carboxylate residue leads to an increase
in Gly/NMDA receptor affinity, probably through an
1
mulls and are expressed in cmÀ1. The H NMR spectra
were obtained with a Varian Gemini 200 instruments at
200 MHz. The chemical shifts are reported in d (ppm),
and are relative to the central peak of the solvent, which
Table 4. Analytical data of the newly synthesized compounds
Compd
Formula
C calcd–found
H calcd–found
N calcd–found
2
3
C
16H15N3O5
58.35–58.54
52.82–52.67
48.25–48.45
47.87–48.06
43.30–43.17
38.55–38.43
61.98–62.20
55.81–55.93
50.08–50.18
45.42–45.21
52.75–52.61
46.84–47.02
42.13–42.26
44.32–44.21
39.73–39.84
35.04–35.12
59.25–59.47
44.48–44.53
57.12–56.99
51.50–51.66
46.88–47.02
60.69–60.60
53.52–53.46
47.87–47.93
47.87–47.98
61.98–61.81
4.60–4.72
3.89–3.96
3.30–3.38
2.79–2.91
2.24–2.17
1.99–2.11
4.84–4.88
3.69–3.76
3.01–3.08
2.46–2.48
2.59–2.64
1.97–1.90
1.48–1.54
1.69–1.76
1.21–1.28
1.07–1.10
3.74–3.70
1.40–1.42
4.16–4.21
3.46–3.50
2.89–2.93
4.32–4.27
3.46–3.51
2.79–2.83
2.79–2.84
4.84–4.76
12.76–12.80
11.55–11.60
10.55–10.53
12.89–12.93
11.66–11.71
10.38–10.42
15.49–15.62
13.95–13.79
12.52–12.58
11.35–11.40
15.38–15.30
13.66–13.60
12.29–12.34
14.10–14.18
12.64–12.70
11.15–11.23
17.28–17.18
6.49–6.51
C16H14ClN3O5
C16H13Cl2N3O5
4
5
C
C
13H9Cl2N3O3
13H8Cl3N3O3
6
7
C13H8BrCl2N3O3
C14H13N3O3
C14H11N3O5
8
9
10
11
12
13
14
15
16
17
18
21
22
23
24
28
29
30
31
32
C14H10ClN3O5
C14H9Cl2N3O5
C
12H7N3O5
C12H6ClN3O5
C12H5Cl2N3O5
C11H5Cl2N3O3
C
11H4Cl3N3O3
C11H4BrCl2N3O3
C12H9N3O3
C8H3Cl2NO2
C
C15H12ClN3O3S
15H13N3O3S
13.33–13.40
12.02–11.94
10.94–10.90
16.34–16.38
14.41–14.36
12.89–12.94
12.89–12.94
15.49–15.41
C
C
C
15H11Cl2N3O3S
13H11N3O3
13H10ClN3O3
C13H9Cl2N3O3
C
C
13H9Cl2N3O3
14H13N3O3

5546
F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
is always DMSO-d6. All the exchangeable protons were
confirmed by addition of D2O.
6.4.2. 3-Diazo-6-chloro-1,3-dihydro-indol-2-one (26). Yield
90%; 1H NMR 6.91 (s, 1H, ar) 7.03 (d, 1H, ar, J = 8.0 Hz)
7.43 (d, 1H, ar, J = 8.0 Hz) 10.81 (br s, 1H, NH).
6.2. Synthesis of 5,6-dichloro-1H-indole-2,3-dione (21)20
6.4.3. 3-Diazo-5,6-dichloro-1,3-dihydro-indol-2-one (27).
Yield 85%; H NMR 7.04 (s, 1H, ar) 7.45 (s, 1H, ar)
10.90 (br s, 1H, NH).
1
Sulfuryl chloride (11.6 mmol) and a crystal of I2 as cat-
alyst were added to a suspension of 6-chloro-1H-in-
dole-2,3-dione 2019 (5.5 mmol) in glacial acetic acid
(20 mL). The mixture was heated at 50 °C for 8 h
and then cooled at room temperature. The resulting
solid was collected and washed with petroleum ether.
Yield 70%; mp 268–270 °C (AcOH) (lit. mp 266–
272 °C). 1H NMR 7.11 (s, 1H, ar), 7.78 (s, 1H, ar),
11.25 (br s, 1H, NH).
6.5. General procedure to prepare diethyl 5,6-dihydro-5-
oxo-pyrazolo[1,5-c]quinazoline-1,2-dicarboxylates 2–4
An excess of diethyl acetylenedicarboxylate (5.16 mmol)
was added to a suspension of diazoindole 25–27
(2.58 mmol) in anhydrous toluene (50 mL). The resulting
mixture was refluxed, under nitrogen atmosphere, until
disappearance of the starting material [2–9 h, TLC mon-
itoring, eluting system CHCl3/MeOH (9:1)] and then
cooled at room temperature. The resulting solid was col-
lected by filtration and washed with diethyl ether.
6.3. General procedure to prepare 4-methyl-N0-(2-oxo-
1,2-dihydro-3H-indol-3-ylidene)benzenesulfonohydrazides
22,21 23,22 and 24
To a warmed (70 °C) solution of isatin 19–2119,20
(16.2 mmol) in methanol (80 mL), an equimolar amount
of p-toluenesulfonylhydrazide (16.2 mmol) was added.
After all of the latter was dissolved, the solution was al-
lowed to stand at room temperature for 2 h. The yellow
platelets were filtered and washed with a little cold
methanol.
6.5.1. Diethyl 5,6-dihydro-5-oxo-pyrazolo[1,5-c]quinazo-
line-1,2-dicarboxylate (2). Yield 75%; mp 192–194 °C
(EtOH) H NMR 1.28–1.37 (m, 6H, 2CH3) 4.33–4.44
(m, 4H, 2CH2) 7.34–7.42 (m, 2H, ar) 7.59–7.63 (m,
1H, ar) 8.48 (d, 1H, ar, J = 8.0 Hz) 12.36 (br s, 1H,
NH); IR 3420, 1750, 1720.
1
6.3.1. 4-Methyl-N0-(2-oxo-1,2-dihydro-3H-indol-3-ylidene)-
benzenesulfonohydrazide (22). Yield 88%; mp 193–200 °C
dec (PhH) (lit. mp 190–200 °C from methanol). 1H
NMR 2.36 (s, 3H, CH3) 6.84–6.91 (m, 1H, ar) 7.03–7.15
(m, 1H, ar) 7.35–7.51 (m, 3H, ar) 7.82–7.93 (m, 3H, ar)
10.73 (br s, 1H, NH).
6.5.2. Diethyl 8-chloro-5,6-dihydro-5-oxo-pyrazolo[1,5-
c]quinazoline-1,2-dicarboxylate (3). Yield 70%; mp 242–
245 °C (EtOH) 1H NMR 1.26–1.36 (m, 6H, 2CH3)
4.30–4.39 (m, 4H, 2CH2) 7.38–7.43 (m, 2H, ar) 8.56
(d, 1H, ar, J = 8.4 Hz) 12.45 (br s, 1H, NH).
6.5.3. Diethyl 8,9-dichloro-5,6-dihydro-5-oxo-pyrazol-
o[1,5-c]quinazoline-1,2-dicarboxylate (4). Yield 75%; mp
265–268 °C (AcOH) 1H NMR 1.26–1.36 (m, 6H,
2CH3) 4.33–4.39 (m, 4H, 2CH2) 7.51 (s, 1H, ar) 8.95
(s, 1H, ar) 12.58 (br s, 1H, NH); IR 3464, 1735, 1712.
6.3.2. 4-Methyl-N0-(6-chloro-2-oxo-1,2-dihydro-3H-indol-3-
ylidene)benzenesulfonohydrazide (23). Yield 77%; mp
206–208 °C (AcOH). H NMR 2.38 (s, 3H, CH3), 6.86
(d, 1H, ar, J = 1.8 Hz), 7.13 (dd, 1H, ar, J = 8.4,
1.8 Hz), 7.43 (d, 2H, ar, J = 8.4 Hz), 7.79–7.88 (m, 3H,
ar), 10.88 (br s, 1H, NH).
1
6.6. General procedure to prepare 1-(ethoxycarbonyl)-5,6-
dihydro-5-oxo-pyrazolo[1,5-c]quinazoline-2-carboxylic acids
9–11
6.3.3. 4-Methyl-N0-(5,6-dichloro-2-oxo-1,2-dihydro-3H-
indol-3-ylidene)benzenesulfonohydrazide (24). Yield 90%;
1
mp 212–218 °C (AcOH) H NMR 2.37 (s, 3H, CH3)
A solution of KOH (10% w/w, 20 mL) was added to a
suspension of 2–4 (2.75 mmol) in methanol (20 mL).
The mixture was left at room temperature under stir-
ring, until disappearance of the starting material [1–
7 h, TLC monitoring, eluting system CHCl3/MeOH
(9:1)]. Addition of water (40 mL) and acidification with
concentrated HCl afforded a suspension that was left at
room temperature for 30 min. The resulting solid was
collected by filtration and washed with water.
7.09 (s, 1H, ar) 7.41 (d, 2H, ar, J = 7.7 Hz) 7.61 (s, 1H,
ar) 7.85 (d, 2H, ar, J = 7.7 Hz) 11.20 (br s, 1H, NH).
6.4. General procedure to prepare 3-diazo-1,3-dihydro-indol-
2-ones 25,21 26,22 and 27
A mixture of 22–24 (4.29 mmol) and an aqueous solu-
tion of NaOH (0.35 M, 31.5 mL) was heated at 60 °C
for 1 h, and then allowed to stand at room temperature
for 2 h. Addition of ice (40 g) and acidification with
acetic acid afforded a solid that was filtered and
washed well with water. The residue was unstable upon
recrystallization. Nevertheless, the crude product was
pure enough and thus used without further
purification.
6.6.1. 1-(Ethoxycarbonyl)-5,6-dihydro-5-oxo-pyrazolo-
[1,5-c]quinazoline-2-carboxylic acid (9). Yield 80%; mp
226–229 °C (EtOH) 1H NMR 1.30 (t, 3H, CH3,
J = 7.3 Hz) 4.36 (q, 2H, CH2, J = 7.3 Hz) 7.30–7.41
(m, 2H, ar) 7.62 (t, 1H, ar, J = 7.7 Hz) 8.38 (d, 1H, ar,
J = 7.7 Hz) 12.30 (br s, 1H, NH); IR 3430, 1759, 1713.
6.4.1. 3-Diazo-1,3-dihydro-indol-2-one (25). Yield 90%;
1H NMR 6.89–7.70 (m, 3H, ar) 7.41 (d, 1H, ar,
J = 7.3 Hz) 10.67 (br s, 1H, NH).
6.6.2. 8-Chloro-1-(ethoxycarbonyl)-5,6-dihydro-5-oxo-pyr-
azolo[1, 5-c]quinazoline-2-carboxylic acid (10). Yield 90%;
mp 240–243 °C (AcOH) 1H NMR 1.29 (t, 3H, CH3,

F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
5547
J = 7.3 Hz) 4.35 (q, 2H, CH2, J = 7.3 Hz) 7.39–7.43
(m, 2H, ar) 8.48 (d, 1H, ar, J = 8.3 Hz) 12.39 (br s 1H,
NH).
6.9. Synthesis of ethyl 1,8-dichloro-5,6-dihydro-5-oxo-pyr-
azolo[1,5-c]quinazoline-2-carboxylate (5)
To a suspension of ethyl 8-chloro-5,6-dihydro-5-oxo-pyr-
azolo[1,5-c]quinazolini-2-carboxylate17 (2.40 mmol) in
glacial acetic acid (15 mL), sulfuryl chloride (5.10 mmol)
was added. The reaction mixture was heated at 60 °C for
2 h, and then cooled at room temperature. The solid was
collected by filtration and washed with water. Yield 90%;
mp 300–302 °C (DMF/H2O) 1H NMR 1.36 (t, 3H, CH3,
J = 7.0 Hz) 4.40 (q, 2H, CH2, J = 7.0 Hz) 7.36–7.46 (m,
2H, ar) 8.44 (d, 1H, ar, J = 8.8 Hz) 12.34 (br s, 1H,
NH); IR 3220, 1770, 1700.
6.6.3. 8,9-Dichloro-1-(ethoxycarbonyl)-5,6-dihydro-5-oxo-
pyrazolo[1,5-c]quinazoline-2-carboxylic acid (11). Yield 85%;
mp 230–232 °C (EtOH) H NMR 1.28 (t, 3H, CH3, J =
7.3 Hz) 4.35 (q, 2H, CH2, J = 7.3 Hz) 7.49 (s, 1H, ar) 8.85
(s, 1H, ar) 12.52 (br s, 1H, NH); IR 3467, 3339, 1744, 1620.
1
6.7. General procedure to prepare 5,6-dihydro-5-oxo-pyraz-
olo[1,5-c]quinazoline-1,2-dicarboxylic acids 12,23 13, and 14
A solution of KOH (10% w/w, 7 mL) was added to a
suspension of 2–4 (0.825 mmol) in ethanol (7 mL). The
mixture was heated at 100 °C for 1 h. Addition of water
(15 mL) and acidification with concentrated HCl affor-
ded an unfilterable suspension that, after heating at
100 °C for 30 min, yielded a solid, which was collected
by filtration and washed well with water.
6.10. Synthesis of 1,8-dichloro-5,6-dihydro-5-oxo-pyrazol-
o[1,5-c]quinazoline-2-carboxylic acid (15)
To a suspension of ester 5 (0.40 mmol) in methanol
(3.25 mL), a solution of KOH (10% w/w, 3.25 mL)
was added, and then the mixture was refluxed for
10 min. After cooling at room temperature, addition of
water (15 mL) and acidification with concentrated HCl
afforded a solid that was collected by filtration and
washed with water. Yield 70%; mp >300 °C (DMF/
6.7.1. 5,6-Dihydro-5-oxo-pyrazolo[1,5-c]quinazoline-1,2-
dicarboxylic acid (12). Yield 90%; mp >300 °C
(EtOH)1H NMR 7.30–7.41 (m, 2H, ar) 7.58–7.66 (m,
1H, ar) 8.62 (d, 1H, ar, J = 8.0 Hz) 12.68 (br s, 1H,
NH); IR 3450, 3300, 1750, 1720.
1
H2O) H NMR 7.38–7.42 (m, 2H, ar) 8.44 (d, 1H, ar,
J = 8.1 Hz) 12.33 (br s, 1H, NH).
6.7.2. 8-Chloro-5,6-dihydro-5-oxo-pyrazolo[1,5-c]quinaz-
oline-1,2-dicarboxylic acid (13). Yield 75%; mp >300 °C
(EtOH) H NMR 7.38–7.43 (m, 2H, ar) 8.74 (d, 1H, ar,
6.11. Synthesis of ethyl 1,8,9-trichloro-5,6-dihydro-5-oxo-
pyrazolo[1,5-c]quinazoline-2-carboxylate (6)
1
J = 8.7 Hz) 12.36 (br s, 1H, NH).
To a suspension of 3-diazoindole 27 (6.58 mmol) in anhy-
drous toluene (150 mL), an excess of ethyl propiolate
(13.2 mmol) was added. The reaction mixture was re-
fluxed, under nitrogen atmosphere and under stirring,
for an overall time of 20 h. During this time, an equimolar
amount of ethyl propiolate (13.2 mmol) was added three
times at intervals of 4 h. The suspension was cooled at
room temperature, and the solid, made up of a mixture
of isomers 30 and 31 (overall yield 80%), was collected
by filtration and washed with diethyl ether. From
100 mg of the mixture of 30 and 31, only a few milligrams
of pure ethyl 8,9-dichloro-5,6-dihydro-5-oxo-pyrazol-
o[1,5-c]quinazoline-2-carboxylate (31) was obtained after
repeated crystallizations from acetic acid [mp >300 °C 1H
NMR 1.34 (t, 3H, CH3, J = 7.0 Hz) 4.37 (q, 2H, CH2,
J = 7.0 Hz) 7.48 (s, 1H, ar) 7.81 (s, 1H, H-1) 8.55 (s, 1H,
ar) 12.27 (br s, 1H, NH)]. Thus, the mixture of 30 and
31 was used without purification for the next step. Sulfu-
ryl chloride (1.23 mmol) and a crystal of I2 as catalyst
were added to a suspension of the mixture of 30 and 31
in glacial acetic acid (15 mL). The reaction mixture was
heated at 60 °C for 1 h and then, without cooling, the solid
(compound 6) was collected by filtration and washed with
petroleum ether. Yield 25% (from 27); mp >300 °C
6.7.3. 8,9-Dichloro-5,6-dihydro-5-oxo-pyrazolo[1,5-c]qui-
nazoline-1,2-dicarboxylic acid (14). Yield 75%; mp
>300 °C (AcOH) 1H NMR 7.52 (s, 1H, ar) 9.14 (s,
1H, ar) 12.45 (br s, 1H, NH); IR 3444, 3257, 1737,
1712, 1672.
6.8. General procedure to prepare ethyl 5,6-dihydro-5-oxo-
pyrazolo[1,5-c]quinazoline-1-carboxylates 28–30
Compounds 9–11 (0.54 mmol) were heated in an oil bath,
over their melting points for 1 h. The cooled, fused mass
was worked-up with ethanol, collected, and purified.
6.8.1. Ethyl 5,6-dihydro-5-oxo-pyrazolo[1,5-c]quinazo-
line-1-carboxylate (28). Yield 45%; mp >300 °C (DMF)
1H NMR 1.36 (t, 3H, CH3, J = 7.3 Hz) 4.37 (q, 2H,
CH2, J = 7.3 Hz) 7.35–7.40 (m, 2H, ar) 7.63 (t, 1H, ar,
j = 7.8 Hz) 8.43 (s, 1H, ar, H-2) 9.15 (d, 1H, ar,
J = 7.8 Hz) 12.40 (br s, 1H, NH).
6.8.2. Ethyl 8-chloro-5,6-dihydro-5-oxo-pyrazolo[1,5-c]qui-
nazoline-1-carboxylate (29). Yield 50%; mp >300 °C
(DMF) H NMR 1.35 (t, 3H, CH3, J = 7.0 Hz) 4.33 (q,
1
1
(AcOH); H NMR 1.35 (t, 3H, CH3, J = 7.1 Hz) 4.39
2H, CH2, J = 7.0 Hz) 7.38–7.43 (m, 2H, ar) 8.44 (s, 1H,
ar, H-2) 9.28 (d, 1H, ar, J = 8.4 Hz) 12.38 (br s, 1H, NH).
(q, 2H, CH2, J = 7.1 Hz) 7.52 (s, 1H, ar) 8.49 (s, 1H, ar)
12.48 (br s, 1H, NH); IR 3216, 1777, 1696.
6.8.3. Ethyl 8,9-dichloro-5,6-dihydro-5-oxo-pyrazolo[1,5-
c]quinazoline-1-carboxylate (30). Yield 60%; mp >300 °C
(AcOH) 1H NMR 1.34 (t, 3H, CH3, J = 7.3 Hz) 4.34 (q,
2H, CH2, J = 7.3 Hz) 7.51 (s, 1H, ar) 8.45 (s, 1H, ar, H-
2) 9.51 (s, 1H, ar) 12.43 (br s, 1H, NH).
6.12. Synthesis of ethyl 1-bromo-8,9-dichloro-5,6-dihydro-5-
oxo-pyrazolo[1,5-c]quinazoline-2-carboxylate (7)
The mixture of 30 and 31, obtained starting from 27
(6.58 mmol) as above described for the preparation of

5548
F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
6, was suspended in glacial acetic acid (15 mL). The
reaction mixture was heated at 70 °C, and then was add-
ed drop-wise to a solution of Br2 (0.46 mmol) in glacial
acetic acid (2 mL). Heating (70 °C) was continued for an
overall time of 20 h. During this time, an equimolar
amount of Br2 (0.46 mmol) was added three times at
intervals of 4 h. The resulting solid (compound 7), with-
out cooling, was collected by filtration and washed with
petroleum ether. Yield 20% (from 27); mp >300 °C
7.37 (m, 2H, ar) 7.59–7.61 (m, 1H, ar) 9.01 (d, 1H, ar
H-10, J = 8.4 Hz) 12.13 (br s, 1H, NH); IR 3140, 1735,
1710.
6.15. Synthesis of 1-methyl-5,6-dihydro-5-oxo-pyrazol-
o[1,5-c]quinazoline-2-carboxylic acid (18)
The title compound was obtained by alkaline hydrolysis
of 8, following the same experimental procedure de-
scribed above to prepare 15. Yield 80%; mp 281–
1
(AcOH); H NMR 1.35 (t, 3H, CH3, J = 7.0 Hz) 4.39
1
(q, 2H, CH2, J = 7.0 Hz) 7.54 (s, 1H, ar) 8.78 (s, 1H,
ar) 12.45 (br s, 1H, NH); IR 3223, 1776, 1702.
283 °C (EtOH) H NMR 2.68 (s, 3H, CH3) 7.33–7.52
(m, 3H, ar) 8.07 (d, 1H, ar H-10, J = 8.0 Hz) 11.97 (br
s, 1H, NH); IR 3450, 3200, 1740.
6.13. General procedure to prepare 1,8,9-trichloro-5,6-dihy-
dro-5-oxo-pyrazolo[1,5-c]quinazoline-2-carboxylic acid (16)
and 1-bromo-8,9-dichloro-5,6-dihydro-5-oxo-pyrazolo[1,5-
c]quinazoline-2-carboxylic acid (17)
7. Biochemistry
7.1. Binding assay
The title compounds were obtained by alkaline hydroly-
sis of 6 and 7, respectively, following the same experi-
mental procedure described above to prepare 15.
Rat cortical synaptic membrane preparation and
[3H]glycine, [3H]AMPA, [3H]-(+)-MK-801, and high-af-
finity [3H]kainate binding experiments were performed,
following the procedures described in Refs. 15, 17, 35,
and 40, respectively.
6.13.1. 1,8,9-Trichloro-5,6-dihydro-5-oxo-pyrazolo[1,5-c]-
quinazoline-2-carboxylic acid (16). Yield 85%; mp
1
>300 °C (EtOH); H NMR 7.50 (s, 1H, ar) 8.47 (s, 1H,
ar) 12.41 (br s, 1H, NH); IR 3650, 3451, 1747, 1684.
7.2. Electrophysiological assay
6.13.2. 1-Bromo-8,9-dichloro-5,6-dihydro-5-oxo-pyrazol-
o[1,5-c]quinazoline-2-carboxylic acid (17). Yield 90%; mp
>300 °C (AcOH); H NMR 7.50 (s, 1H, ar) 8.75 (s, 1H,
ar) 12.41 (br s, 1H, NH); IR 3454, 3245, 1743, 1685.
The cortical wedge preparation described by Mannaioni
et al.41 was used, while the electrophysiolocical assays
were performed following the procedures described in
Ref. 17.
1
6.14. Synthesis of ethyl 1-methyl-5,6-dihydro-5-oxo-pyr-
azolo[1,5-c]quinazoline-2-carboxylate (8)
7.3. Sample preparation and results calculation
A stock 1 mM solution of the test compound was pre-
pared in 50% DMSO. Subsequent dilutions were accom-
plished in buffer. The IC50 values were calculated from
three or four displacement curves based on four to six
scalar concentrations of the test compound using the
ALLFIT computer program42 and, in the case of tritiated
glycine and AMPA binding, converted to Ki values, by
application of the Cheng–Prusoff equation.43 Under
our experimental conditions, the dissociation constants
(KD) for [3H]glycine (10 nM) and [3H]-DL-AMPA
(8 nM) were 75 6 and 28 3 nM, respectively.
To a suspension of 3-diazoindole 25 (5.99 mmol) in
anhydrous toluene (40 mL), an excess of ethyl 2-butino-
ate (8.99 mmol) was added. The reaction mixture was re-
fluxed, under nitrogen atmosphere and under stirring,
for an overall time of 72 h. After 30 h, ethyl 2-butinoate
(2.99 mmol) was added. Another addition of ethyl 2-
butinoate (5.98 mmol) was made after 56 h. After the
reaction was cooled at room temperature, the solid
was collected by filtration. The residue, made up of a
mixture of 8 and of its regio-isomer ethyl 2-methyl-5,6-
dihydro-5-oxo-pyrazolo[1,5-c]quinazoline-1-carboxylate
(32), was purified on silica gel column, eluting with
CHCl3/AcOEt/AcOH (8.5/1/0.5). Evaporation at re-
duced pressure of the solvent, of the first and the last
elutes afforded compounds 8 and 32, respectively.
8. Computational methodologies
All molecular modeling studies were carried out on a
six-CPU (PIV 2.0–3.0 GHz) Linux cluster running un-
der openMosix architecture.44
6.14.1. Ethyl 1-methyl-5,6-dihydro-5-oxo-pyrazolo[1,5-
c]quinazoline-2-carboxylate (8). Yield 35%; mp 293–
1
295 °C (EtOH) H NMR 1.35 (t, 3H, CH3, J = 7.1 Hz)
Three-dimensional structures of iGluR extracellular
binding domains were obtained from the Brookhaven
Protein Data Bank (PDB). For docking, the structure
of glycine-binding domain S1S2 of NR1 subunit in com-
plex with the antagonist DCKA (PDB code: 1PBQ)32,
and the structure of glutamate-binding domain S1S2
of GluR2 subunit in complex with the antagonist
DNQX (PDB code: 1FTL)33 were used. Since all X-
ray crystallographic files do not contain hydrogen
atoms, they need to be added to the protein by using
2.68 (s, 3H, CH3) 4.36 (q, 2H, CH2, J = 7.1 Hz) 7.34–
7.38 (m, 2H, ar) 7.50–7.53 (m, 1H, ar) 8.08 (d, 1H, ar
H-10, J = 7.7 Hz) 12.04 (br s, 1H, NH); IR 3160, 1750,
1730.
6.14.2. Ethyl 2-methyl-5,6-dihydro-5-oxo-pyrazolo[1,5-
c]quinazoline-1-carboxylate (32). Yield 20%; mp 249–
1
251 °C (EtOH) H NMR 1.36 (t, 3H, CH3, J = 7.0 Hz)
2.51 (s, 3H, CH3) 4.36 (q, 2H, CH2, J = 7.0 Hz) 7.30–

F. Varano et al. / Bioorg. Med. Chem. 13 (2005) 5536–5549
5549
17. Varano, F.; Catarzi, D.; Colotta, V.; Filacchioni, G.;
`
Galli, A.; Costagli, C.; Carla, V. J. Med. Chem. 2002, 45,
1035.
18. Catarzi, D.; Colotta, V.; Varano, F.; Calabri, F. R.;
Molecular Operating Environment (MOE, version
2003.03) modeling suite45, before carrying out docking
studies. To minimize contacts between hydrogen atoms,
the structures were subjected to Amber94 energy mini-
`
Filacchioni, G.; Galli, A.; Costagli, C.; Carla, V. J. Med.
Chem. 2004, 47, 262.
mization until the rms of conjugate gradient was
À1, keeping the heavy atoms fixed
˚
A
<0.15 kcal molÀ1
19. Rapport, M. M.; Senear, A. E.; Mead, J. F.; Koepfli, J. B.
J. Am. Chem. Soc. 1946, 68, 2697.
at their crystallographic positions.
20. Senear, A. E.; Sargent, H.; Mead, J. F.; Koepfli, J. B. J.
Am. Chem. Soc. 1946, 68, 2695.
21. Cava, M. P.; Litle, R. L.; Napier, D. R. J. Am. Chem. Soc.
1958, 80, 2257.
22. Vogt, B. R. US 4112098, 1978.
23. Yamazaki, T.; Shechter, H. Tetrahedron Lett. 1973, 14,
1417.
24. Augusti, R.; Kascheres, C. J. Org. Chem. 1993, 58, 7079.
25. Vogt, B. R. US 4112096, 1978.
PQZ derivatives were generated by using the Molecular
Operating Environment suite. Atomic charges were as-
signed using the PEOE, implemented by MOE.45 Each li-
gand was placed into the binding pockets and the
semiflexible docking was achieved with the program
SCHRo¨DINGER/GLIDE.46,47 The final docked complexes
of ligand–receptor were selected according to the criteria
of interaction energy combined with geometrical match-
ing quality.
26. Bigge, C. F.; Nikam, S. S. Exp. Opin. Ther. Patents 1997,
7, 1099.
27. Bigge, C. F.; Boxer, P. A.; Ortwaine, D. F. Curr. Pharm.
Des. 1996, 6, 397.
28. Ohmori, J.; Shimizu-Sasamata, M.; Okada, M.; Sakam-
oto, S. J. Med. Chem. 1997, 40, 2053.
Acknowledgments
The molecular modeling work coordinated by S.Moro
was carried out with financial support of Associazione
Italiana per la Ricerca sul Cancro (AIRC), Milan, and
the Italian Ministry for University and Research
(MIUR), Rome, Italy. S. Moro is also grateful to Chem-
ical Computing Group for their scientific and technical
partnership.
29. Madsen, U.; Stensbøl, T. B.; Krogsgaard-Larsen, P. Curr.
Med. Chem. 2001, 8, 1291.
30. Wong, E. H. F.; Kemp, J. A.; Priestley, T.; Knight, A. R.;
Woodruff, G. N.; Iverson, L. L. Proc. Natl. Acad. Sci.
U.S.A. 1986, 83, 7104.
31. Wong, E. H. F.; Knight, A. R.; Ranson, R. Eur. J.
Pharmacol. 1987, 142, 487.
32. Furukawa, H.; Gouaux, E. Embo J. 2003, 22, 2873.
33. Armstrong, N.; Gouaux, E. Neuron 2000, 28, 165.
34. Leeson, P. D.; Carling, R. W.; Moore, K. W.; Moseley, A.
M.; Smith, J. D.; Stevenson, G.; Chan, T.; Baker, R.;
Foster, A. C.; Grimwood, S.; Kemp, J. A.; Marshall, G.
R.; Hoogsteen, K. J. Med. Chem. 1992, 35, 1954.
35. Colotta, V.; Catarzi, D.; Varano, F.; Cecchi, L.; Filacchi-
oni, G.; Galli, A.; Costagli, C. Arch. Pharm. Pharm. Med.
Chem. 1997, 330, 129.
36. Micheli, F.; Di Fabio, R.; Capelli, A. M.; Cugola, A.;
Curcuruto, O.; Feriani, A.; Gastaldi, P.; Gaviraghi, G.;
Marchioro, C.; Orlandi, A.; Pozzan, A.; Quaglia, A. M.;
Reggiani, A.; van Amsterdam, F. Arch. Pharm. Weinheim
1999, 332, 73.
References and notes
1. Bra¨uner-Osborne, H.; Egebjerg, J.; Nielsen, E. Ø.; Mad-
sen, U.; Krogsgaard-Larsen, P. J. Med. Chem. 2000, 43,
2609.
2. Dingledine, R.; Borges, K.; Bowie, D.; Traynelis, S. F.
Pharmacol. Rev. 1999, 51, 7.
3. Lees, G. J. Drugs 2000, 59, 33.
4. Auberson, Y. P. Drugs Future 2001, 26, 463.
5. Johnson, J. W.; Ascher, P. Nature 1987, 325, 529.
6. Danysz, W.; Parsons, C. G. Pharmacol. Rev. 1998, 50, 597.
7. Kulagowski, J. J.; Leeson, P. D. Exp. Opin. Ther. Patents
1995, 5, 1061.
37. Tikhonova, I. G.; Baskin, I. I.; Palyulin, V. A.; Zefirov, N.
S. J. Med. Chem. 2003, 46, 1609.
8. Kulagowski, J. J. Exp. Opin. Ther. Patents 1996, 6, 1069.
9. Donati, D.; Micheli, F. Exp. Opin. Ther. Patents. 2000, 10,
667.
10. Leeson, P. D.; Iversen, L. L. J. Med. Chem. 1994, 37, 4053.
11. Varano, F.; Catarzi, D.; Colotta, V.; Cecchi, L.; Filacchi-
oni, G.; Galli, A.; Costagli, C. Il Farmaco 1997, 52, 179.
12. Varano, F.; Catarzi, D.; Colotta, V.; Cecchi, L.; Filacchi-
oni, G.; Galli, A.; Costagli, C. Arch. Pharm. Pharm. Med.
Chem. 1999, 332, 201.
38. Moretti, L.; Pentikainen, O. T.; Settimo, L.; Johnson, M.
S. J. Struct. Biol. 2004, 145, 205.
39. Baskin, I. I.; Tikhonova, I. G.; Palyulin, V. A.; Zefirov, N.
S. J. Med. Chem. 2003, 46, 4063.
40. Nielsen, E. O.; Madsen, U.; Schaumburg, K.; Brehm, L.;
Krogsgaard-Larsen, P. Eur. J. Chem.-Chim. Ther. 1986,
21, 433.
`
41. Mannaioni, G.; Carla, V.; Moroni, F. Br. J. Pharmacol.
1996, 118, 1530.
13. Catarzi, D.; Colotta, V.; Varano, F.; Cecchi, L.; Filacchioni,
G.; Galli, A.; Costagli, C. J. Med. Chem. 1999, 42, 2478.
14. Catarzi, D.; Colotta, V.; Varano, F.; Cecchi, L.; Filacchi-
42. De Lean, A.; Munson, P. J.; Rodbard, D. Am. J. Physiol.
1978, 235, E97.
43. Cheng, Y.-C.; Prusoff, W. H. Biochem. Pharmacol. 1973,
22, 3099.
44. OpenMosix, http://www.openMosix.org.
45. Molecular Operating Environment (MOE 2003.02), Chem-
ical Computing Group, Inc., 1255 University St., Suite
1600, Montreal, Quebec, Canada H3B 3X3.
46. Friesner, R. A. J. Med. Chem. 2004, 47, 1739.
47. Halgren, T. A. J. Med. Chem. 2004, 47, 1750.
`
oni, G.; Galli, A.; Costagli, C.; Carla, V. J. Med. Chem.
2000, 43, 3824.
15. Varano, F.; Catarzi, D.; Colotta, V.; Cecchi, L.; Filacchi-
oni, G.; Galli, A.; Costagli, C. Eur. J. Med. Chem. 2001,
36, 203.
16. Catarzi, D.; Colotta, V.; Varano, F.; Filacchioni, G.; Galli,
`
A.; Costagli, C.; Carla, V. J. Med. Chem. 2001, 44, 3157.