Synthesis, antiplatelet activity and comparative molecular field analysis of substituted 2-amino-4H-pyrido[1,2-a]pyrimidin-4-ones, their congeners and isosteric analogues

Article abstract of DOI:10.1016/S0968-0896(00)00010-9

2-(1-Piperazinyl)-4H-pyrido[1,2-a]pyrimidin-4-one (5a) is a recently described in vitro inhibitor of human platelet aggregation which specifically inhibits the activity of high affinity cAMP phosphodiesterase. A number of substitution derivatives, isosteres, and analogues of 5a were now synthesized and tested in vitro for their inhibitory activity on human platelet aggregation induced in platelet-rich plasma by ADP, collagen, or the Ca²⁺ ionophore A23187. Among the most effective compounds, the 6-methyl, 8-methyl and 6,8-dimethyl derivatives of 5a resulted nearly as active as the lead when platelet aggregation was induced by ADP or A23187, but less active when collagen was the inducer. On the basis of present results and those previously obtained by us in this and 2-aminochromone structural fields, we have developed a statistically significant 3-D QSAR model, using comparative molecular field analysis (CoMFA), describing the variation of the antiplatelet activity in terms of molecular steric and electrostatic potential changes. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(00)00010-9

Bioorganic & Medicinal Chemistry 8 (2000) 751±768
Synthesis, Antiplatelet Activity and Comparative Molecular Field
Analysis of Substituted 2-Amino-4H-pyrido[1,2-a]pyrimidin-4-
ones, their Congeners and Isosteric Analogues
a,
c
a
a
Giorgio Roma, * Nunzia Cinone, Mario Di Braccio, Giancarlo Grossi,
b
b
c,
Giuliana Leoncini, Maria Grazia Signorello and Angelo Carotti *
aDipartimento di Scienze Farmaceutiche, Universit aÁ di Genova, viale Benedetto XV, 16132 Genoa, Italy
bDipartimento di Medicina Sperimentale, Sezione Biochimica, Universit aÁ di Genova, viale Benedetto XV, 16132 Genoa, Italy
cDipartimento Farmacochimico, Universit aÁ di Bari, via Orabona 4, 70125 Bari, Italy
Received 27 May 1999; accepted 19 October 1999
AbstractÐ2-(1-Piperazinyl)-4H-pyrido[1,2-a]pyrimidin-4-one (5a) is a recently described in vitro inhibitor of human platelet
aggregation which speci®cally inhibits the activity of high anity cAMP phosphodiesterase. A number of substitution derivatives,
isosteres, and analogues of 5a were now synthesized and tested in vitro for their inhibitory activity on human platelet aggregation
induced in platelet-rich plasma by ADP, collagen, or the Ca2+ ionophore A23187. Among the most e€ective compounds, the 6-
methyl, 8-methyl and 6,8-dimethyl derivatives of 5a resulted nearly as active as the lead when platelet aggregation was induced by
ADP or A23187, but less active when collagen was the inducer. On the basis of present results and those previously obtained by us
in this and 2-aminochromone structural ®elds, we have developed a statistically signi®cant 3-D QSAR model, using comparative
molecular ®eld analysis (CoMFA), describing the variation of the antiplatelet activity in terms of molecular steric and electrostatic
potential changes. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
The interesting platelet anti-aggregating properties
shown in vitro by some of the 1,2-fused pyrimidine
derivatives 5±1013 subsequently prompted us to widen
the study also in the ®eld of fused pyran derivatives.
Thus, we synthesized a group of properly chosen com-
pounds 1±3 and 11±13 (Chart 3), structural analogues of
compounds 5±7 and 8±10, respectively, and evaluated
their in vitro antiplatelet activity. Also some of the
In the course of our previous chemical and pharma-
cological studies on 2-(dialkylamino)chromones 1 and
their benzo-fused derivatives 2, 3 and 4 (Chart 1), whose
±6
®
rst examples were described by us,1 we reported the
in vitro antiplatelet activity of some 7-substituted com-
7
±10
pounds 1.
1
4
compounds 1±3, 11±13 were signi®cantly active.
Considering that the platelet aggregation inhibitory
properties of some pyrido[1,2-a]pyrimidine derivatives
had also been reported,11 we have more recently syn-
thesized and tested for their in vitro antiplatelet activity
a number of N-substituted 2-amino-4H-pyrido[1,2-a]-
On the whole, if we consider the antiplatelet activity
data obtained in the above studies for all classes of
compounds tested (1±3, 5±13) towards all the platelet
aggregation inducers used (i.e. adenosine diphosphate
(ADP), collagen, and the Ca ionophore A23187 (cal-
cimycin)), the following remarks can be made.
2
3
12,13
2+
pyrimidin-4-ones 5 (R =R =H, X=O),
structu-
rally analogous to compounds 1, and of their angular
benzo-fused derivatives 6 and 7,13 as well as some
1
3
examples of compounds 8, 9 and 10, isomers of 5
2
. 1-Piperazinyl proved to be the most e€ective
amino substituent among all those used. Actu-
ally the (1-piperazinyl) substituted compound
3
(
R =R =H, X=O), 6 and 7, respectively (Chart 2).
(
) was the most active one in
each structural class, generally towards all platelet
aggregation inducers (in the case of compounds
*
5
3
Corresponding authors. AC. Tel.: +39-080-5442782; fax: +39-080-
442230; e-mail: carotti@farmchim.uniba.it GR. Tel.: +31-010-
538374; fax: +39-010-3538358.
8
±10 this derivative was not obtainable).
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00010-9

7
52
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
prepared and tested by us,7±10,12±14 with respect both
to the heterocyclic nucleus and the dialkylamino
substituent.
For this reason we have now been dealing with the
synthesis and biological evaluation of several new com-
pounds (5b±r, 17a,b, 19a±d and 20), whose structures
have been derived from that of 5a through proper
modi®cations of the pyridine ring and/or the 2-sub-
stituent, aiming at reaching higher antiplatelet activities
and widening the knowledge of structure±activity rela-
tionships in this ®eld. The chemical and biological
results of this study are reported in this paper.
Chart 1.
Furthermore, an extended antiplatelet activity data set,
obtained by adding the actual results to those previously
reported for compounds 1±3, 11±1314 and 5±10
12,13
has
.
.
.
The (1-piperazinyl) substituted compounds 5±7
were more active than the corresponding com-
pounds 1±3.
The activities of (1-piperazinyl) substituted cou-
marin derivatives 11±13 were very similar to those
of the corresponding chromone derivatives 1±3.
The 2-(1-piperazinyl) substituted compound 5a
was the most active of all compounds tested,
whereas 7-ethoxy-4-(1-piperazinyl)coumarin 11b
was the most active of the fused pyran derivatives
been submitted to a 3-D QSAR (CoMFA) study which
is also described here.
The platelet antiaggregating properties of compound 5a
derive from its ability to speci®cally inhibit the activity
of high anity cAMP phosphodiesterase and, conse-
1
5
quently, increase intracellular cAMP concentration.
Chemistry
1
±3, 11±13 (Table 1).
Syntheses of intermediates 14a±c, 15b,c,i,j and 18a, and
of ®nal compounds 5b±r, 17a,b, 19a±d and 20 are
depicted in Schemes 1 and 2, respectively.
For convenience, the IC50 values of the most active of
7±10,12±14
all the compounds tested in the above studies
are summarized in Table 1.
According to a synthetic route previously reported in
the literature,16 the reaction at room temperature of the
suitable acyl chlorides with heteroarylamines a€orded
the N-substituted ethyl malonamates 14a±c that were in
turn heated with a mixture of polyphosphoric acid and
phosphorus oxychloride to give the chloroderivatives
In the light of these results, we can plainly conclude that
-(1-piperazinyl)-4H-pyrido[1,2-a]pyrimidin-4-one (5a) is
the most interesting in vitro inhibitor of human platelet
2
aggregation among compounds 1±3, 5±13 previously
Chart 2.
Chart 3.

G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
753
Table 1. In vitro inhibitory activity of (1-piperazinyl) substituted compounds 1±3, 5±7, 11±13 (Charts 1±3) on human platelet aggregation induced
a
in PRP by ADP, collagen and A23187
IC50 (mM)‹SD
Compoundb
R2
R3
X
ADP (5.0 mM)
Collagen (10.0 mg/mL)
c
A23187 (20.0 mM)
1
2
3
5
6
7
1
1
1
1
a
H
Ð
Ð
H
Ð
Ð
Ð
H
Ð
Ð
Ð
O
56‹13
39‹3
60‹12
56‹16
117‹32
3.6‹1.2
15‹8
240‹90
201‹61
278‹15
19‹9
a
a
119‹40
6‹1.8
13‹4
a
a
Ð
Ð
H
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
28‹8
a
38‹10
49‹14
24‹14
53‹15
81‹23
21‹9
54‹17
250‹38
44‹18
245‹65
334‹89
1a
1b
2a
3a
36‹10
10‹3.6
49‹17
71‹27
2 5
7-OC H
Ð
Ð
a
PRP=platelet-rich plasma.
Compounds 5a, 6a, 7a: ref 13; compounds 1a, 2a, 3a, 11a,b, 12a, 13a: ref 14.
c5.0 mg/mL for compounds 5a, 6a, 7a.
b
1
5b,i and 18a, respectively. On the other hand, the
derivative 19d, respectively. The same treatment of 5-
chloroderivative 18a a€orded a lower yield of 2-methyl-
5-(1-piperazinyl)-7H-isoxazolo[2,3-a]pyrimidin-7-one
chloroderivatives 15c,j were prepared through the
reaction of 2-chloro-4H-pyrido[1,2-a]pyrimidin-4-one
(15a) with bromine at room temperature or the Law-
esson's reagent in re¯uxing toluene, respectively.
1
7
20
(19a). Moreover, also (chloromethyl)derivatives 16a,b
reacted with piperazine under similar conditions to
give the [(1-piperazinyl)methyl] derivatives 17a and 20,
respectively. Finally, the hydrogenation of compound
5a, using 5% palladium on activated charcoal as cata-
lyst, a€orded the tetrahydroderivative 19c.
The substituted 2-amino-4H-pyrido[1,2-a]pyrimidin-4-
ones 5c±j,l±r and 4-thione analogue 5k (see Table 2)
were generally obtained in high yields by treating the
corresponding 2-chloroderivatives 15a±j with excess
amines (ethanol at re¯ux), whereas 3-nitro-2-(1-piper-
azinyl)-4H-pyrido[1,2-a]pyrimidin-4-one (5b) (Table 2)
was prepared through the nitration of compound 5a13
with a mixture of concentrated nitric and sulfuric
acids.
The structures attributed to the compounds described in
this paper are supported by the results of elemental
1
analyses and IR and H NMR spectral data (see
Experimental and Table 3).
Thus, b-enaminonic compounds 5b±j,l±r a€orded char-
acteristic H NMR signals for protons H-3 (singlet) and
1
By heating a solution of 2-chloroderivative 15a in eth-
ylene glycol, in the presence of anhydrous potassium
carbonate, 2-(2-hydroxyethoxy)-4H-pyrido[1,2-a]pyrimi-
din-4-one (17b) was obtained. The treatment of 7-chloro-
H-6 at d 5.27±5.67 and 8.71±8.99, respectively, and the
±1
n CO IR band at 1656±1680 cm , as we previously
Only the H
observed in this structural ®eld.1
2,13,21±24
1
1
8
19
derivatives 18b
re¯uxing ethanol) gave high yields of 7-(1-piperazinyl)-
H-thiazolo[3,2-a]pyrimidin-5-one (19b) or its 2,3-dihydro-
or 18c
with excess piperazine
NMR H-3 singlet of 5p is shifted more down®eld (d
6.11), due to the lower shielding e€ect of 2-hydrazino
substituent on this proton.12
(
5

7
54
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
Scheme 1.
Also the spectra of compounds 19a±d show the 1
NMR singlet of the pyrimidine hydrogen at d 5.17±5.39
H
and propranolol were also tested under the same condi-
tions as reference compounds. The IC50 values obtained
are reported in Table 2 and suggest the following
remarks with reference to the three groups below of
compounds, which are structurally related to 5a, taken
as a lead (see Table 1).
�
1
and the n CO IR band at 1640±1663 cm , in accor-
dance with the above results and their structural simi-
2
larity with compounds 5 (X=O, R =H).
Finally, it is interesting to remark that the replacement
of the (1-piperazinyl) substituent of compounds 5a13
and 19b with a [(1-piperazinyl)methyl] group (com-
pounds 17a and 20, respectively) resulted in a decreased
shielding of the pyrimidine hydrogen of latter com-
Compounds 5l±r, 17a,b (deriving from modi®cation of
-piperazinyl substituent of 5a) and 20
1
The activities of these compounds are far lower than
that of 5a. In particular, these results seem to indicate
that the open chain 2-substituents (5q,r, 17b), the steric
hindrance at 1-piperazinyl NH group (5l), as well as the
lack of the direct link of 1-piperazinyl group to position
pounds and in the consequent appreciable down®eld
1
shift of its H NMR signal (CDCl ), due to the lack of
3
the enamino substituent and of its electron-releasing
resonance e€ect.
2
(17a, 20) are structural features unfavorable to anti-
platelet activity of these analogues of 5a. Even the 1-
homopiperazinyl substituent (5m) is far less e€ective
than 1-piperazinyl.
Results and Discussion
Structure±activity relationships (SAR)
Compounds 19a±d (deriving from modi®cation of
pyridine ring of 5a)
Compounds 5b±r, 17a,b, 19a±d and 20 were tested in
vitro for their inhibitory activity on the aggregation
of human platelets induced in platelet-rich plasma
When ADP or A23187 were used as platelet aggregation
inducers, isoxazolo[2,3-a]pyrimidine and thiazolo[3,2-a]-
pyrimidine derivatives 19a and 19b, respectively, resul-
ted in being moderately less e€ective than their isosteric
(
the Ca
PRP) by adenosine diphosphate (ADP), collagen, or
2
+
ionophore A23187 (calcimycin) (see Experi-
mental). Acetyl salicylic acid (ASA), tri¯uoperazine,

G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
755
Scheme 2.
analogue 5a, as well as the hydrogenated derivatives 19c
and 19d compared with their respective parent com-
pounds 5a and 19b. When collagen was used, the dif-
ference of activity between compounds 19a, 19b or 19c
and compound 5a proved to be rather higher.
in positions 3 and 9 of the 4H-pyrido[1,2-a]pyrimidine
nucleus appears especially unfavorable (compare the
IC50 values of methylderivatives 5c,h with those of
5e±g,i). Moreover, the introduction of electron with-
drawing 3-substituents into 5a a€ords derivatives (5b,d)
considerably less active than the parent compound.
Compounds 5b±j (deriving from introduction of further
substituents into 5a nucleus) and 5k
Three dimensional quantitative structure±activity
relationships (3-D QSAR)
All the compounds of this group are less active than
the parent, except for methylderivatives 5e±g,i and
chloroderivative 5j which are nearly equiactive to 5a
towards ADP and A23187. The presence of substituents
The in vitro antiplatelet activity data reported in Tables
4±9 were subjected to a 3-D QSAR study by means of
Comparative Molecular Field Analysis (CoMFA).
2
5

7
56
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
Table 2. In vitro inhibitory activity of compounds 5b±r, 17a,b, 19a±d, 20 on platelet aggregation induced in human PRP by ADP, collagen and
A23187
IC50 (mM)‹SD
Compounda
R2
R3
H
ADP (5.0 mM)
Collagen (5.0 mg/mL)
A23187 (20.0 mM)
5
b
NO
2
Ð
366‹108
789‹90
853‹38
5
5
5
5
5
5
5
5
5
5
c
d
e
f
CH
Br
H
3
H
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
28‹5
182‹38
5.6‹1.7
8.7‹1.2
5.8‹0.7
76.5‹2.8
5.1‹1.0
9.9‹2.7
75‹8
93‹26
230‹74
18.5‹3.6
18.5‹5.6
16.2‹3.8
93.5‹44.6
14.6‹3.3
20.7‹5.3
160‹32
>1000
76‹2
H
202‹90
17.4‹1
6-CH
7-CH
8-CH
9-CH
3
3
3
3
H
15.5‹5.8
15.8‹6.4
177‹17
19.6‹7.6
17.8‹5.9
142‹34
>1000
g
h
i
H
H
H
6,8-CH
7-Cl
H
3
j
H
k
l
H
H
H
970‹45
5
m
H
H
Ð
270‹39
564‹83
777‹156
5
5
n
o
H
H
H
H
Ð
Ð
>1000
>1000
>1000
958‹59
891‹13
731‹185
5
5
p
q
H
H
H
H
Ð
Ð
>1000
>1000
>1000
>1000
>1000
>1000
5
1
r
H
H
Ð
>1000
>1000
>1000
9a
Ð
Ð
Ð
15.6‹3.2
39.4‹10
49.0‹15
19b
19c
19d
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
Ð
19.5‹4.5
27‹4
46.4‹3.3
97.4‹13
197‹41
51.3‹10
53‹9
130‹20
265‹77
ASA
Tri¯uoperazine
Propranolol
>1000
180‹48
291‹76
159‹24
138‹41
131‹31
>1000
318‹61
269‹62
aThe IC50 values of compounds 17a,b and 20 (see Scheme 2) were >1000 mM towards all the three platelet aggregation inducers.

Table 3. Spectral, physicochemical and reaction data of substituted 2-amino-4H-pyrido[1,2-a]pyrimidin-4-ones 5c±j, l±r and 2-(1-piperazinyl)-4H-pyrido[1,2-a]pyrimidine-4-thione 5k
Compounda Reaction Yield
mp
ꢀ
(solv.)
C
Molecular
formulac
IRd (cm
� 1
)
1
H NMR (d, ppm)
e
b
time (h)
(%)
82
86
82
62
76
70
f
5
5
5
5
5
5
c
d
e
f
3
1
1
1
1
1
146±147 (A)
150±151 (B)
140±141 (A)
148±149 (A)
143±144 (A)
169±170 (A)
C
13
H
16
N
4
O
3205 (NH), 1658 (CO),
628, 1552, 1527
3290 (NH), 1666 (CO),
627, 1552, 1528s
3270 (NH), 1656 (CO),
638sh, 1554br, 1500
3320 (NH), 1663 (CO),
1638, 1540sh, 1530, 1505
3290w (NH), 1663 (CO),
646, 1540sh, 1530, 1500
3295w (NH), 1660 (CO),
635, 1540, 1520sh,
493w
3305 (NH), 1666 (CO)
and 1650 (fused), 1550br,
1.77 (s, 1H, NH), 2.21(s, 3H, CH
7.23±7.84 (m, 2H, H-8, 9), 8.98 (m, 1H, H-6)
1.74 (s, 1H, NH), 3.02 (m, 4H, CH NH), 3.75 (m, 4H, NCH ), 7.02 (m, 1H, H-7), 7.41 (m, 1H, H-9),
7.71 (m, 1H, H-8), 8.99 (m, 1H, H-6)
1.90 (s, 1H, NH), 2.94 (m, 4H, CH NH), 3.01 (s, 3H, CH ), 3.64 (m, 4H, NCH
(dd, 1H, H-7), 7.17 (dd, 1H, H-9), 7.39 (dd, 1H, H-8)
2.00 (s, 1H, NH), 2.34 (s, 3H, CH ), 2.94 (m, 4H, CH NH), 3.66 (m, 4H, NCH
3 2 2
), 3.03 (m, 4H, CH NH), 3.47 (m, 4H, NCH ), 6.98 (m, 1H, H-7),
1
f
C
12
H13BrN
4
O
2
2
1
f
C
C
C
C
13
H
13
H
13
H
13
H
16
N
16
N
16
N
16
N
4
O
4
O
4
O
4
O
2
3
2
), 5.46 (s, 1H, H-3), 6.46
), 5.64 (s, 1H, H-3), 7.19
1
f
3
2
2
(near d, 1H, H-9), 7.48 (near d, 1H, H-8), 8.71 (near s, 1H, H-6)
1.88 (s, 1H, NH), 2.37(s, 3H, CH ), 2.92 (m, 4H, CH NH), 3.66 (m, 4H, NCH ), 5.58 (s, 1H, H-3), 6.70
(near d, 1H, H-7), 7.07 (near s, 1H, H-9), 8.79 (near d, 1H, H-6)
1.87 (s, 1H, NH), 2.45 (s, 3H, CH ), 2.96 (m, 4H, CH NH), 3.70 (m, 4H, NCH ), 5.66 (s, 1H, H-3), 6.79
(dd, 1H, H-7), 7.43 (dd, 1H, H-8), 8.85 (dd, 1H, H-6)
g
f
g
h
3
2
2
1
g
f
3
2
2
1
1
f
5
i
1
74
174±174.5 (A)
C
14
H
18
N
4
O
1.80 (s, 1H, NH), 2.24 (s, 3H, 8-CH
3 2 3 2
), 2.90 (m, 4H, CH NH), 2.96 (s, 3H, 6-CH ), 3.60 (m, 4H, NCH ),
5.39 (s, 1H, H-3), 6.29 (near s, 1H, H-7), 6.89 (near s, 1H, H-9)
1
500
3312 (NH), 1670 (CO),
624, 1550, 1532, 1508
3245 (NH), 1640, 1580sh,
568s, 1548, 1530sh,
112 (CS)
3255w (NH), 1672s (CO),
f
5
j
1
85
92
191±192 (A)
136±137 (A)
C
12
H
13ClN
4
O
1.85 (s, 1H, NH), 2.94 (m, 4H, CH
2 2
NH), 3.67 (m, 4H, NCH ), 5.61 (s, 1H, H-3), 7.22 (near d, 1H, H-9),
1
7.54 (dd, J8,9=9.6 Hz, J8,6=2.5 Hz, 1H, H-8), 8.93 (near d, 1H, H-6)
f
5
k
0.5
C
12
H
14
4
N S
1.90 (s, 1H, NH), 2.95(m, 4H, CH
2
NH), 3.77(m, 4H, NCH
2
), 6.62±7.99 (m, 3H, H-7, 8, 9), 7.22 (s, 1H,
1
H-3), 10.26 (m, 1H, H-6)
1
f
5
l
1
90
163±164 (A)
C
14
H
18
N
4
O
1.14 (d, 6H, CH
3
), 1.66 (s, 1H, NH), 2.48 (dd, 2H, axial NCH), 2.77±2.97 (m, 2H, CHCH
3
), 4.31 (dd,
1
628, 1556, 1534, 1505br
2H, equatorial NCH), 5.63 (s, 1H, H-3), 6.85 (m, 1H, H-7), 7.27 (m, 1H, H-9), 7.57 (m, 1H, H-8), 8.87
m, 1H, H-6)
), 2.21 (s, 1H, NH), 2.66±3.23 (m, 4H, 2 CH
5.58 (s, 1H, H-3), 6.87 (m, 1H, H-7), 7.27 (m, 1H, H-9), 7.59 (m, 1H, H-8), 8.91 (m, 1H, H-6)
1.47±1.67 (m, 2H, b-CH ), 1.85±2.05 (m, 3H, b-CH +OH; m, 2H, after treatment with D O), 3.32 (m,
2H, NCH ), 3.98 (m, 1H, CHOH), 4.06±4.22 (m, 2H, NCH ), 5.67 (s, 1H, H-3), 6.85 (m, 1H, H-7), 7.28
(m, 1H, H-9), 7.57 (m, 1H, H-8), 8.87 (m, 1H, H-6)
(
5
m
n
4
2
74
96
178±179 (C)
160±161 (A)
C
13
H
O
16
4.
N
4
O.
3310 (NH), 1660 (CO),
1631, 1554, 1535, 1510sh
3595w and 3355br (OH),
g
1.61±2.12 (m, 2H, CH
2
f
2 2
NH), 3.53±4.01 (m, 4H, 2 NCH ),
C
4
H
4
0.5H
2
O
5
C
13
H
15
N
3
O
2
2
2
2
1
658 (CO), 1631, 1554sh,
533, 1506sh
3345, 3260 w and 3170br
NH+NH ), 1664 (CO),
640, 1630sh, 1558br,
538, 1485
3305w (NH), 1668 (CO),
632, 1577, 1542, 1480
3335w (NH), 1665 (CO),
637, 1561, 1542, 1511w
3460-3210br (NH+NH ), 1.40 (s, 2H, NH
674 (CO), 1642, 1575,
545, 1480
2
2
1
f
5
o
2
87
200±201.5 (D)
C
14
13
H
18
N
4
O
1.04±1.38 (m, 4H, 2 CH
NH
2
), 1.70±1.98 (m, 5H, CH NH
f
2
+2CH
2
), 2.51 (m, 1H, NHCH), 3.30 (broad, 2H,
(
2
2
), 5.27 (s, 1H, H-3), 7.00 (m, 1H, H-7), 7.08 (d, 1H, NH), 7.26 (m, 1H, H-9), 7.75 (m, 1H, H-8), 8.75
1
(m, 1H, H-6)
1
f
5p
5q
5r
3
2
2
30
60
46
149±150 (E)
158±159 (F)
166.5±167 (F)
C
H
16
N
.
4
O
1.00±1.95 (m, 6H, b-CH
2
+g-CH
2
), 2.38±2.98 (m, 4H, NCH
2
), 5.73 (broad s, 1H, NH), 6.11 (s, 1H, H-3),
1
6.89 (m, 1H, H-7), 7.23 (m, 1H, H-9), 7.62 (m, 1H, H-8), 8.96 (m, 1H, H-6)
1.98 (s, 1H, NH), 2.48 (s, 3H, NHCH ), 2.85 (t, 2H, CH NHCH ), 3.08 (s, 3H, CH ), 3.76 (t, 2H,
NCH ), 5.53 (s, 1H, H-3), 6.82 (m, 1H, H-7), 7.27 (m, 1H, H-9), 7.54 (m, 1H, H-8), 8.86 (m, 1H, H-6)
f
C
12
10
H
16
N
4
O C
4
H
4
O
4
4
3
2
3
3
1
2
.
O C
f
f
C
H
12
N
4
4
H
4
O
2
2 2 2 2
), 2.87 (near t, 2H, CH NH ), 3.27 (m, 2H, NHCH ), 5.31 (broad s, 1H, NH), 5.41 (s,
1
1H, H-3), 6.79 (m, 1H, H-7), 7.17 (m, 1H, H-9), 7.51 (m, 1H, H-8), 8.83 (m, 1H, H-6)
1
a
1
For compounds 5m,q,r, yields, melting points (mp), crystallization solvents, molecular formulas and analyses refer to the corresponding maleates, whereas IR and H NMR spectral data of the free bases are reported.
Crystallization solvents: A=ethyl acetate, B=ethyl acetate/petroleum ether, C=methanol/petroleum ether, D=chloroform, E=isopropyl ether, F=dichloromethane/petroleum ether.
Analyses: C, H, N, Br for 5d, C, H, N, Cl for 5j, C, H, N, S for 5k, C, H, N for all other compounds.
b
c
d
In CHCl
3
solution: compounds 5g,h,m,n,p,q,r; in KBr pellet: compounds 5c,d,e,f,i,j,k,l,o.
for all compounds except 5o, for which (CD SO was used.
O.
From the spectrum of compound in the solid state (KBr).
e
Solvents: CDCl
3
3 2
)
f
Disappeared with D
2
g

7
58
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
Table 4. Chemical structures and platelet antiaggregating activity of compounds of class A
R1
R2
3
R , R4
5
R , R6
pIC50a
Ref
Compound
A
1
A
2
A
3
A
4
A
5
A
6
N(CH
3
)
2
H
H
H
H
H
H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
3.13
3.48
3.44
3.46
3.55
3.21
12
12
12
12
12
12
N(C
N(C
2
H
5
)
2
2
4 9
H )
N(iC
N(C
4
H
9
)
2
2
H
5
)C
6
H
5
A
A
A
A
7
H
H
H
H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
3.94
3.40
2.75
2.75
12
12
12
12
8
9
10
A
11
A
12
A
13
A
14
NHC
NHiC
NHCH
2
H
5
H
H
H
H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
3.04
3.11
3.12
3.23
12
12
12
12
4 9
H
2
C
6
H
5
A
15
16
H
H
H, H
H, H
H, H
H, H
3.01
3.21
12
12
A
A
17
18
H
H
H, H
H, H
H, H
H, H
3.34
3.43
12
12
A
A
19
A
20
A
21
A
22
H
H
H
H
H, H
H, H
H, H
H, H
H, H
H, H
3.15
5.44
3.62
4.82
12
13
13
13
H, H
(CH=CH)
2
A
23
A
24
A
25
A
26
A
27
N(CH
3
)
2
H
H
H
H
H
(CH=CH)
(CH=CH)
(CH=CH)
(CH=CH)
(CH=CH)
2
2
2
2
2
H, H
H, H
H, H
H, H
H, H
3.43
4.11
3.29
3.30
3.62
13
13
13
13
13
N(C
N(C
2
H
5
)
)
2
2
4
H
9
2 4 2
N(C H OH)
A
28
A
29
A
30
A
31
H
H
H
H
(CH=CH)
(CH=CH)
(CH=CH)
(CH=CH)
2
2
2
2
H, H
H, H
H, H
H, H
4.12
3.68
4.22
4.06
13
13
13
13

G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
759
Table 4 (continued)
R1
R2
H
R , R4
3
R , R6
5
pIC50a
4.67
Ref
13
Compound
A
32
H, H
(CH=CH)
2
A
A
33
2
N(C H
5
)
2
H
H
H, H
H, H
(CH=CH)
(CH=CH)
2
3.05
3.92
13
13
34
2
A
35
A
36
A
37
H
H
H
H, H
H, H
H, H
(CH=CH)
(CH=CH)
(CH=CH)
2
2
2
3.54
3.15
3.79
13
13
13
A38 (5r)
A39 (5q)
A40 (17b)
A41 (5n)
NHCH
2
CH
2
NH
NHCH
OH
2
H
H
H
H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
2.75
2.75
2.75
2.75
(b)b
(b)
(b)
(b)
NCH
3
CH
2
CH
CH
2
3
OCH
2
2
A42 (5o)
A43 (5e)
A44 (5i)
A45 (5j)
A46 (17a)
A47 (5b)
A48 (5c)
A49 (19c)
A50 (5d)
A51 (5k)
A52 (5l)
H
H
H, H
H, H
H, H
3.05
4.73
4.83
4.68
2.75
3.10
4.03
4.01
3.64
3.79
2.75
(b)
(b)
(b)
(b)
(b)
(b)
(b)
(b)
(b)
(b)
(b)
CH
3
3
, H
, H
H
CH
3
CH , H
H
H, Cl
H, H
H, H
H, H
2H, 2H
H, H
H, H
H, H
H, H
H, H
H, H
H, H
2H, 2H
H, H
H, H
H, H
H
NO
2
CH
3
H
Br
H
H
A53 (5p)
A54 (5h)
A55 (5g)
A56 (5f)
H
H
H
H
H, H
H, H
H, H
H, H
2.75
4.03
4.79
4.73
(b)
(b)
(b)
(b)
H, CH
3
3
CH , H
H, CH
3
H, H
A
A
57
2
N(C H
5
)
2
C(ˆO)H
C(ˆO)H
H, H
H, H
H, H
H, H
3.23
3.23
12
12
58
A59 (5m)
H
H, H
H, H
3.25
(b)
a
The IC50 values were determined in the assay with collagen 5.0 mg/mL.
Described in the present paper.
b

7
60
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
Table 5. Chemical structures and platelet antiaggregating activity of compounds of class B
R1
R , R3
2
R , R5
4
pIC50a
4.22
Ref
14
Compound
B
1
B
2
B
3
H, H
H, H
H, H
H, H
OC
2
H
5
, H
4.13
2.75
14
14
H, H
B
B
4
HNCH
2
CH
2
NH
2
H, H
(CHˆCH)
H, H
H, H
3.19
4.25
14
14
5
2
2
2
2
2
B
6
B
7
B
8
(CHˆCH)
(CHˆCH)
(CHˆCH)
H, H
H, H
H, H
3.19
2.75
3.54
14
14
14
B
B
9
N(C
2
H
5
)
2
(CHˆCH)
H, H
(CHˆCH)
3.71
3.93
14
14
10
H, H
2
2
2
2
B
11
B
12
B
13
H, H
H, H
H, H
(CHˆCH)
(CHˆCH)
(CHˆCH)
3.18
2.75
3.22
14
14
14
2 5 2
N(C H )
aThe IC50 values were determined in the assay with collagen 10.0 mg/mL.
CoMFA relates the biological activities of a series of
molecules with their steric and electrostatic ®elds calcu-
lated at several grid points around the molecule by
of a three-dimensional model, able to interpret the
changes in the antiplatelet activity, in terms of varia-
tions of steric and electrostatic characteristics of the
analyzed compounds. Hopefully, such a model could
suggest some structural modi®cations aimed at sig-
ni®cantly improving the target activity of the examined
compounds.
3
means of a suitable probe, usually an sp carbon atom
26
with a charge of +1. Partial least squares (PLS ) is
used as the regression method to develop the relation-
ship between steric and electrostatic potentials and bio-
logical activity. The graphical representation of
CoMFA model, in the form of coecient isocontour
maps, eciently located the regions where the variation
in steric and electrostatic properties of di€erent mole-
cules in a data set is correlated with the variation of
biological activity. CoMFA isocontour maps may fur-
nish useful indications to prioritize future synthesis and
to develop sound working hypotheses on the nature of
putative ligand±macromolecule interactions. Therefore,
the main goal of our CoMFA study was the development
For the sake of clarity and an easier discussion of the
results from the 3-D QSAR study, the analyzed com-
pounds were grouped in six congeneric seriesÐcalled A,
B, C, D, E and FÐand progressively numbered within
each class (see Tables 4±9) (Chart 4).
In the ®rst column of Tables 4±9, the diverse compound
numbering used in the Chemistry section is also repor-
ted. Classes A±F comprised a quite diverse number of

G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
761
Table 6. Chemical structures and platelet antiaggregating activity of
compounds of class C
Table 8. Chemical structures and platelet antiaggregating activity of
compounds of class E
Compound
R1
R2
H
R3
H
pIC50a
4.33
Ref
(b)
R1
2
R , R3
4
R , R5
pIC50a Ref
Compound
E
1
(19b)
C
1
C
2
C
3
C
4
H, H
H, H
4.44
5.00
4.31
4.15
14
14
14
14
E
2
(20)
H
H
2.75
3.70
(b)
(b)
H, H
(CHˆCH)
H, H
2 5
OC H , H
E
3
(19d)
2H
2H
2
H, H
a
(CHˆCH)
2
The IC50 values were determined in the assay with collagen 5.0 mg/mL.
Described in the present paper.
b
aThe IC50 values were determined in the assay with collagen 10.0 mg/mL.
Table 9. Chemical structure and platelet antiaggregating activity of
compound 19a (class F)
Table 7. Chemical structures and platelet antiaggregating activity of
compounds of class D
Compound
pIC50a
4.40
Ref
(b)
R1
R2, R3
R4, R5
pIC50a Ref
1
F (19a)
Compound
a
The IC50 value was determined in the assay with collagen 5.0 mg/mL.
Described in the present paper.
D
D
1
N(C
2
H
5
)
)
2
2
(CHˆCH)
2
H, H
H, H
3.41
3.26
13
13
b
2
(CH=CH)
2
D
D
3
N(C
2
H
5
H, H
H, H
(CH=CH)
(CH=CH)
2
3.44
3.42
13
13
The pIC50 values of the 84 compounds tested for their in
vitro antiplatelet activity cover a relatively wide range
4
2
(from 2.75 to 5.44) and are acceptably well distributed
along it.
aThe IC50 values were determined in the assay with collagen 5.0 mg/mL.
The CoMFA study was performed within the QSAR
module of SYBYL rel 6.4.27
compounds: 59 for class A (39 already reported and 20
described herein for the ®rst time), 13 for class B, 4 for
both class C and D, 3 for class E and ®nally only 1 for
class F. The (3,5-dimethyl-1-piperazinyl) derivative A52
In a ®rst step the whole data set was divided in two
parts, the training set (TS) and the prediction set (PS),
composed by 73 and 10 molecules, respectively. The PS
was chosen as to have representative molecules, with
pIC50 values spanning in a regular way the full range of
activities (see Table 11).
(5l), carrying a peculiar and unique structural variation
at the piperazinyl ring, resulted a strong outlier and was
not included in the CoMFA study.
The in vitro antiplatelet activity data shown in Tables
The molecular alignment chosen for the CoMFA study
followed a conformational analysis on the amine
substituents linked to the di€erent rigid heterocyclic
skeletons.
4±9 as pIC50 derive from the assay with collagen (5 mg/
mL; 10 mg/mL for data from ref 14), since in this case a
higher number of de®nite pIC50 could be obtained.
However, in the case of IC >1000 mM, a truncated
50
pIC50 value of 2.75 was arbitrarily used to avoid the loss
of relevant structural information for the modeling
study.
Most of the compounds bearing a cyclic amine pre-
sented a minimum energy conformer in which the amine
and the heterocyclic planes were almost perpendicular.

7
62
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
Chart 4.
On the other hand, compounds carrying secondary and
tertiary amine substituents presented minimum energy
conformers with the heterocyclic ring and the plane
containing the three N±C bonds almost coplanar.
Unfortunately, the lack of parametrization of several
fragments in structures A±F did not allow the calcula-
tion of molecular lipophilicity potential (MLP)2 for 15
molecules, so that in the case of models based on neu-
tral molecules (N ±N ) only 58 molecules remained in
8
4
10
For the molecular overlay, the minimum energy con-
former of the piperazinyl derivative 5a was selected as
the template because of its high activity and relatively
low conformational freedom.
the TS; the situation was still worse when the proto-
nated molecules were taken into account. As a con-
sequence, no attempt to introduce the MLP as a third
®eld was made.
All the molecules in the TS were superimposed on the
template 5a by a pair-wise ®t of the corresponding
heavy atoms of the heterocyclic rings. To reach a better
A careful analysis of the statistical ®gures of the PLS
models reported in Table 10 pointed out a similar ®tting
power for the two-®eld (steric plus electrostatic) models
developed for neutral (N ) and protonated (P ) mole-
®
t, the geometry of the minimum energy conformers
1
1
was slightly modi®ed by manual rotation around selec-
ted torsion angles of the amine substituents, allowing,
occasionally, an energy penalty up to 5 kcal/mol.
cules. In terms of predictive ability, model P is slightly
1
better than model N , both showing improved statistics
1
compared to single-®eld models (P , P and N , N ,
2
3
2
3
respectively).
The statistical parameters of the models derived for
both neutral (N ±N ) and protonated (at the nitrogen
of amine substituents) (P ±P ) molecules are listed in
The examination of PLS models coming from the
reduced set of 58 compounds revealed that the lipophilic
®eld alone gave a very poor model (N ) whereas statis-
1
10
1
6
Table 10. For congeners with a further basic group on
the amine substituents (i.e. A , A , A and A ) only the
unconjugated nitrogen was considered protonated.
7
tically improved models were obtained with the two-
and three-®eld combinations (N ±N ). These ®ndings
indicated that, besides the steric and electrostatic
7
8
20
22
8
10
Table 10. Statistical data of PLS models
q2
r2
Table 11. Observed versus predicted platelet antiaggregating activity
of compounds of the PS
Models
Fields
n
ONC
s
N
N
N
N
N
N
N
N
N
N
1
2
3
4
5
6
7
8
9
10
ste+ele
ste
ele
ste+ele
ste
ele
73
73
73
58
58
58
58
58
58
58
73
73
73
73
73
73
6
6
5
6
6
6
4
4
4
4
6
6
5
6
5
5
0.580
0.508
0.420
0.619
0.474
0.508
0.290
0.603
0.546
0.631
0.688
0.495
0.431
0.682
0.392
0.431
0.822
0.764
0.717
0.871
0.799
0.830
0.699
0.794
0.812
0.839
0.858
0.751
0.655
0.874
0.723
0.722
0.286
0.330
0.358
0.258
0.323
0.297
0.388
0.321
0.306
0.284
0.256
0.339
0.396
0.242
0.357
0.358
Compound
pIC50 (obs)a
pIC50 (pred)b
A
A
A
A
A
A
A
A
A
A
38
30
24
15
50
7
2.75
4.22
4.11
3.01
3.64
3.94
3.48
4.68
3.05
3.25
3.09
3.66
3.46
3.24
4.30
3.48
3.26
4.47
3.32
3.65
lipo
lipo+ste
lipo+ele
lipo+ste+ele
ste+ele
ste
2
P
P
P
P
P
P
1
45
33
59
2
3
4
5
6
ele
ste+ele
ste
ele
aThe IC50 values were determined in the assay with collagen 5.0 mg/
mL.
b
4
Predicted from model P .

G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
763
properties, also the lipophilic characteristics of the tes-
ted compounds may play a role in the modulation of the
antiplatelet activity.
The better statistics obtained for charged molecules
(
mod. P ) prompted us to analyze for protonated com-
1
pounds 8±13 also a di€erent ®tting criterion on the
template 5a. The two diverse ®tting topologies explored
(a) and (b) are reported in Chart 5.
Interestingly, the statistical ®gures of the two-®eld
model developed from topology (b) (model P , Table
4
1
0) were comparable, or even slightly better in terms of
tting power, to those from topology (a) (model P1).
Figure 1. Isocontour steric map for PLS model P4 expressed as
STDEVÂCOEFF. (contour levels: 0.039 green; � 0.050 red. Color
®
The good predictive ability of model P can be appre-
ciated by comparing the observed and predicted values
reported in Table 11 for the PS.
code: green increased activity, red decreased activity). Compounds A20
,
4
8 7
A44, A and B are shown to help interpretation.
Model P was also selected to develop the coecient
4
4H-pyrido[1,2-a]pyrimidin-4-one 5a13 (Table 1), that is,
the most active compound synthesized by us in this
structural ®eld, we can draw the following conclusions.
(a) The presence of the 2-(1-piperazinyl) substituent as
enamino group of the b-enaminonic moiety seems to be
essential, or at least very favorable, to the activity of 5a.
(b) The introduction of properly chosen substituents in
the free positions 6,7, and/or 8 (but not 3 and/or 9) of
compound 5a, as well as the replacement of the fused
pyridine ring of 5a with a suitable ®ve- or six-membered
heteroaromatic ring, can likely a€ord very active inhi-
bitors of human platelet aggregation.
isocontour maps reported in Figures 1 and 2. The steric
isocontour map reported in Figure 1 showed sterically
accessible regions (in green) occupied by 1-piperazinyl
substituents of very active compounds and sterically
hindered zones (in red) reached by substituents on
piperazinyl nitrogen N of molecules with low activity.
4
In the electrostatic isocontour map depicted in Figure 2,
zones in which a high electron density favors or dis-
favors the biological activity are indicated by the
magenta and white colors, respectively. Highly active
molecules place their aromatic rings on the magenta
zone, whereas low activity molecules occupy the white
zone with electron-rich groups.
The 3-D QSAR study of the antiplatelet activity of
more than eighty diverse compounds, structurally rela-
ted to 5a, gave further insight on the main physico-
chemical features governing the antiaggregating
inhibitory properties in this chemical ®eld.
Conclusions
Considering the IC50 values obtained for the in vitro
antiplatelet activity of compounds described here (Table
In particular, both steric and electrostatic interactions
seem to take place in well de®nite spatial locations,
suggesting possible proper structural modi®cations for a
signi®cant improvement of the antiplatelet activity. An
important in¯uence of lipophilic interactions on the
antiplatelet activity cannot be excluded. Unfortunately,
2), with reference to those a€orded by 2-(1-piperazinyl)-
4
Figure 2. Isocontour electrostatic maps for PLS model P expressed as
STDEVÂCOEFF. (contour levels: 0.042 white; � 0.025 magenta.
Color code: white decreased anity for negative charge; magenta
increased anity for negative charge). Compounds A31, A53 and C
are shown to help the interpretation.
2
Chart 5.

7
64
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
the lack of parametrization of diverse fragments for the
MLP calculations has not allowed the introduction of
MLP as third ®eld in our CoMFA study. The synthesis
of new inhibitors carrying fully parametrized hetero-
aromatic fragments could help clarifying this important
aspect of the structure±activity relationships.
Preparation of ethyl N-(5-chloro-2-pyridyl)malonamate
(14b) and ethyl N-(5-methyl-3-isoxazolyl)malonamate
(14c). The solution of 40.0 mmol of 5-chloropyridin-2-
amine (5.10 g) or 5-methylisoxazol-3-amine (3.92 g) in
30 mL of dichloromethane and 10 mL of dry pyridine
was added dropwise at room temperature to a stirred
solution of ethyl 3-chloro-3-oxopropanoate (7.98 g,
5
3.0 mmol) in 30 mL of dichloromethane (an exothermic
reaction with emission of white fumes occurred during
the addition). The resulting warm mixture was stirred at
room temperature for 30 min then poured into 300 mL
of cold water; an excess of sodium carbonate was care-
fully added with stirring and the mixture was further
stirred at room temperature for 1 h. The organic layer
was then collected and the aqueous phase extracted
several more times with dichloromethane. The com-
bined extracts were washed with water, dried over
anhydrous sodium sulfate, then evaporated to dryness
in vacuo. By treating the nearly solid residue with a little
ethyl ether, pure compound 14b (8.64 g, 89%) or 14c
(5.52 g, 65%), respectively, separated out as a whitish
solid which was then crystallized from the suitable solvent.
Experimental
Chemistry
Melting points (mp) were determined using a Fisher±Johns
apparatus and are uncorrected. IR spectra were recor-
ded on a Perkin±Elmer 398 spectrophotometer (abbre-
viations relative to IR bands: br=broad, s=strong,
1
w=weak, sh=shoulder). H NMR spectra were recor-
ded on a Hitachi Perkin±Elmer R 600 (60 MHz) spec-
trometer for compounds 5b±k,m,p, 14b, 15b,c,j, 17a,
19b±d, 20 and on a Varian Gemini 200 (200 MHz)
spectrometer for 5l,n,o,q,r, 14a,c, 15i, 17b, 18a, 19a and
chemical shifts (d) are reported in ppm using (CH ) Si
3
4
as an internal reference (d=0). Spin multiplicities are
given as follows: s (singlet), d (doublet), dd (double
doublet), t (triplet), q (quartet), m (multiplet). Analyses
of all new compounds were performed, for the indicated
elements, by the Laboratorio di Microanalisi del
Dipartimento di Scienze Farmaceutiche, University of
Genoa, and the results were within ‹0.4% of the theo-
retical values.
ꢀ
Compound 14b. Whitish solid, mp 109±110 C (isopro-
1
pyl ether). H NMR (CDCl ) d 1.31 (t, 3H, CH CH ),
3
2
3
3.57 (s, 2H, CH ), 4.29 (q, 2H,CH CH ), 7.71 (dd, 1H,
0
2
2
3
0
0
H-4 ), 8.09±8.46 (m, 2H, H-3 ,6 ), 9.80 (broad s, 1H,
NH; disappeared with D O). IR (CHCl ) 3410 and 3310
br (NH), 1728 (ester CO), 1698 (CO), 1594 w, 1579,
2
3
�
1
1522 br cm . Anal. (C H ClN O ) C, H, N, Cl.
11
10
2
3
ꢀ
Thin-layer chromatograms were run on Merck silica
gel 60 F254 precoated plastic sheets (layer thickness
Compound 14c. White crystals, mp 130.5±131 C (ethyl
1
acetate). H NMR (CDCl ) d 1.31 (t, 3H, CH CH ),
3
2
3
0
2.41 (s, 3H, 5 -CH ), 3.49 (s, 2H, CH ), 4.26 (q,
3 2
0.2 mM). Column chromatography was performed
0
2H,CH CH ), 6.67 (s, 1H, H-4 ), 9.88 (broad s, 1H, NH;
2 3
using Carlo Erba silica gel (0.05±0.20 mM) or Carlo
Erba neutral aluminium oxide (Brockmann activity I).
disappeared with D O). IR (CHCl ) 3400 and 3280 br
2
(NH), 1725 (ester CO), 1705 (CO), 1620, 1548 br cm
Anal. (C H N O ) C, H, N.
3
�
1
.
Ethyl 2-methyl-N-(2-pyridyl)malonamate (14a). Phos-
phorus pentachloride (17.39 g, 83.5 mmol) was slowly
added to an ice-cooled stirred solution of ethyl hydrogen
9
12
2
4
Preparation of chloroderivatives 15b,i and 18a. Follow-
ing a procedure previously used16 in the same chemical
®eld, a mixture of 20.0 mmol of the appropriate com-
pound 14, 9.20 g (60.0 mmol) of phosphorus oxychlor-
ide and 1.50 g of polyphosphoric acid was heated with
stirring at 130 C for 3 h (110 C for 1 h in the case of
compound 18a). After cooling, anhydrous ethanol
(20 mL) was added and the mixture was further heated
at re¯ux and stirred for 30 min, then allowed to cool.
The chloroderivative was then recovered from the mix-
ture as described below.
2-methylmalonate (11.69 g, 80.0 mmol) in 80 mL of
dichloromethane. The resulting solution was stirred at
room temperature for 3 h, then 2-aminopyridine (4.99 g,
5
4
3.0 mmol), dissolved in 50 mL of dichloromethane and
0 mL of dry pyridine, was added dropwise so that the
ꢀ
ꢀ
exothermic acylation of amine occurred. The mixture
was further stirred at room temperature for 30 min, then
poured into cold water (500 mL) and vigorously stirred
overnight at room temperature. After careful addition
of excess sodium carbonate with stirring, the organic
layer was collected and the aqueous phase exhaustively
extracted with dichloromethane. The combined extracts
were washed with water, dried over anhydrous sodium
sulfate, and solvents removed under reduced pressure.
The resulting oil was distilled in vacuo and pure com-
pound 14a collected (6.01 g, 51%) as colorless thick oil,
2-Chloro-3-methyl-4H-pyrido[1,2-a]pyrimidin-4-one (15b).
The solution obtained from the reaction of 14a (4.44 g)
was poured into water (400 mL) and the whitish com-
pound 15b that separated out was collected by ®ltration,
washed with water and dried (1.02 g, 26%); white crys-
ꢀ
1
ꢀ
1
bp 114±115 C (0.08 mM Hg). H NMR (CDCl ) d 1.28
tals, mp 167±168 C (ethyl acetate). H NMR (CDCl ) d
3
3
(
CH), 4.22 (q, 2H,CH CH ), 7.06 (m, 1H, H-5 ), 7.70 (m,
1
t, 3H, CH CH ), 1.51 (d, 3H, 2-CH ), 3.48 (q, 1H,
2.35 (s, 3H, CH ), 7.22 (m, 1H, H-7), 7.51±8.00 (m, 2H,
H-8,9), 9.05 (m, 1H, H-6). IR (CHCl ) 1668 br (CO),
2
3
3
3
0
H, H-4 ), 8.18 (m, 1H, H-3 ), 8.28 (m, 1H, H-6 ), 9.20
2
3
3
0
broad s, 1H, NH; disappeared with D O). IR (®lm)
0
0
� 1
1637, 1568, 1521 cm . Anal. (C H ClN O) C, H, N, Cl.
9
7
2
(
2
3
1
320 br (NH), 1745 (ester CO), 1700 (CO), 1602 w,
�
2,7-Dichloro-4H-pyrido[1,2-a]pyrimidin-4-one (15i).
From the reaction carried out with 14b (4.85 g) a sus-
1
583, 1530 br cm . Anal. (C H N O ) C, H, N.
14
11
2
3

G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
765
pension of yellowish crystalline solid was obtained. This
solid was collected by ®ltration then treated with aqu-
eous sodium carbonate, and the mixture was exhaus-
tively extracted with dichloromethane. The combined
extracts were washed with water and dried (anhydrous
sodium sulfate). Removal of the solvent gave a solid
residue from which, after treatment with a little ethyl
stirred ice-cooled suspension of compound 5a (1.04 g,
4.5 mmol) in 20 mL of 96% sulfuric acid. The resulting
yellow solution was stirred at 0 C for 15 min, then
ꢀ
poured onto crushed ice. The solution obtained was
made alkaline by adding excess 28% aq ammonia solu-
tion, then exhaustively extracted with chloroform. The
combined extracts were dried (anhyd sodium sulfate)
and evaporated to dryness in vacuo to yield a yellow
solid residue, which was taken up in a little ethyl ether
and ®ltered to a€ord 0.86 g (69%) of pure 5b; yellow
acetate, pure compound 15i was recovered (2.63 g,
ꢀ
6
1%); pale-yellow crystals, mp 222±222.5 C (chloro-
1
form:ethanol 1:1). H NMR (CDCl ) d 6.52 (s, 1H, H-
3
ꢀ
1
3), 7.62 (d, J9,8=9 Hz, 1H, H-9), 7.80 (dd, J8,9=9 Hz,
J8,6=2.2 Hz, 1H, H-8), 9.07 (d, J6,8=2.2 Hz, 1H, H-6).
crystals, mp 174±176 C dec (ethyl acetate). H NMR
(CDCl ) d 1.78 (s, 1H, NH; disappeared with D O), 2.96
3
2
�
1
IR (CHCl ) 1700 (CO), 1685 sh, 1630, 1558, 1509 cm
3
.
(m, 4H, CH NH), 3.65 (m, 4H, NCH ), 7.01 (m, 1H, H-
7), 7.32 (m, 1H, H-9), 7.79 (m, 1H, H-8), 8.90 (m, 1H,
H-6). IR (KBr) 3312 w (NH), 1680 (CO), 1630, 1543
2 2
Anal. (C H Cl N O) C, H, N, Cl.
8
4
2
2
�
s,br, 1519 sh cm . Anal. (C H N O ) C, H, N.
12 13 5 3
1
5
(
-Chloro-2-methyl-7H-isoxazolo[2,3-a]pyrimidin-7-one
18a). The reaction performed with 4.24 g of 14c a€or-
ded a ®nal suspension that was worked up exactly as
described above for the corresponding suspension in the
case of 15i. Pure compound 18a (1.26 g, 34 %) was so
General procedure for synthesis of substituted 2-amino-
4H-pyrido[1,2-a]pyrimidin-4-ones 5c±j,l±r and 2-(1-piper-
azinyl)-4H-pyrido[1,2-a]pyrimidine-4-thione (5k). A mix-
ture of the suitable amine (40.0 mmol) and 2-
chloroderivative 15 (4.0 mmol), and 50 mL of ethanol
was re¯uxed for 1±4 h, while stirring, then evaporated to
dryness under reduced pressure. The residue obtained
was partitioned between water and chloroform, and the
aqueous phase extracted several more times with
chloroform. The combined extracts (dried over anhyd
sodium sulfate), after removal of solvents, a€orded an
oily or solid residue from which compound 5 was
recovered as described below.
ꢀ
obtained; whitish solid, mp 214±215 C (ethyl acetate).
1
H NMR (CDCl ) d 2.60 (s, 3H, CH ), 6.34 and 6.35
3
3
(
s+s, 2H, H-3+H-6). IR (CHCl ) 1710 sh, 1686 (CO),
3
�
618, 1560, 1500 cm . Anal. (C H ClN O ) C, H, N,
7 5 2 2
1
1
Cl.
3-Bromo-2-chloro-4H-pyrido[1,2-a]pyrimidin-4-one (15c).
The solution of 0.39 mL (7.5 mmol) of bromine in 5 mL
of dichloromethane was added dropwise at room tem-
perature to a stirred solution of 0.90 g (5.0 mmol) of
1
7
compound 15a in 10 mL of dry pyridine. The resulting
mixture was stirred at room temperature for 15 min.
The orange solid that separated out was collected by
2-(1-Piperazinyl) substituted compounds 5c±k. The thick
oily residues obtained from the reactions of piperazine
with 2-chloroderivatives 15b,c, 15d,16 15e±g, 15h,
29
16
®
ltration then treated with 10% aq sodium carbonate,
and the mixture was extracted several times with
dichloromethane. The combined extracts were washed
twice with 10% aq sodium carbonate then with water,
and dried over anhyd sodium sulfate. After removal of
solvents, the solid residue was taken up in a little ethyl
ether and ®ltered to give 1.11 g (85%) of pure 15c;
15i, or 15j were treated with a little ethyl acetate±ethyl
ether. After standing, compounds 5c±k, respectively,
separated out as white or whitish solids (yellow solid in
the case of thione 5k) which were recovered by ®ltration,
dried and crystallized from the suitable solvents.
ꢀ
whitish crystals, mp 224 C (ethyl acetate with char-
1
Compounds 5l,n,o. The solid residues obtained from the
17
coal). H NMR (CDCl ) d 7.30 (m, 1H, H-7), 7.53±8.03
reactions of 15a with 2,6-cis-dimethylpiperazine, 4-
3
(
1
m, 2H, H-8,9), 9.08 (m, 1H, H-6). IR (CHCl ) 1688 (CO),
hydroxypiperidine, or 1,4-trans-diaminocyclohexane,
were taken up in a little ethyl acetate and ®ltered; thus
the pure compounds 5l, 5n or 5o, respectively, were
obtained as white solids which were then crystallized
from the suitable solvents.
3
�
633, 1551 w, 1510cm . Anal. (C H BrClN O) C, H, N.
8 4 2
1
2
-Chloro-4H-pyrido[1,2-a]pyrimidine-4-thione (15j). A
17
mixture of 0.90 g (5.0 mmol) of compound 15a, 2.43 g
6.0 mmol) of Lawesson's reagent and 50 mL of anhy-
(
drous toluene was heated at re¯ux for 3 h, while stirring.
The solvent was then removed under reduced pressure
and the residue was taken up in a little dichloromethane
and chromatographed on a neutral aluminium oxide
column, eluting with the same solvent. The eluate col-
lected, after removal of solvent, a€orded pure com-
pound 15j as yellow crystals (0.79 g, 80%); mp 182±
2-[(1-Piperidinyl)amino]-4H-pyrido[1,2-a]pyrimidin-4-one
1
7
(5p). The thick oil obtained from the reaction of 15a
with 1-aminopiperidine was chromatographed on a
silica gel column, eluting with the mixture dichlor-
omethane:ethyl acetate (1:1). The eluate collected, after
removal of solvents, gave an oily residue which was
treated with ethyl ether and allowed to stand so that
pure compound 5p separated out as a white solid which
was then crystallized from isopropyl ether.
ꢀ
1
1
83 C (ethyl acetate). H NMR (CDCl ) d 7.30±8.26
3
(
IR (CHCl ) 1637, 1565 s, 1526, 1151 (CS) cm . Anal.
m, 3H, H-7,8,9), 7.74 (s, 1H, H-3), 10.29 (m, 1H, H-6).
1
�
3
(
C H ClN S) C, H, N, Cl, S.
8
Compounds 5m,q,r. The thick oily residues derived from
5
2
the reactions of 15a17 with homopiperazine, N,N -dime-
thylethylenediamine, or ethylenediamine were dissolved
in a little anhyd ethanol and treated with an ethanolic
solution of maleic acid (0.50 g) (in the case of com-
0
3
(
2
-Nitro-2-(1-piperazinyl)-4H-pyrido[1,2-a]pyrimidin-4-one
5b). A cold solution of 0.40 mL of 99% nitric acid in
mL of 96% sulfuric acid was added dropwise to a

7
66
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
pounds 5q,r further addition of some ethyl ether and
ꢀ
IR (CHCl ) 3360 br (OH), 1676 br (CO), 1636, 1568,
3
� 1
standing at 4 C were then necessary). This way pure
1518 cm . Anal. (C H N O C, H, N.
3)
10
10
2
maleates of compounds 5m, 5q or 5r, respectively,
separated out as white crystalline solids which were
recrystallized from the suitable solvents.
Preparation of (1-piperazinyl)derivatives 19a,b,d. A mix-
ture of 4.0 mmol of the proper chloroderivative 18,
3
.44 g (40.0 mmol) of piperazine, and 50 mL of ethanol
1
IR and H NMR spectra of compounds 5m,q,r were
was re¯uxed for 1 h, while stirring. Following the
procedure above described for the preparation of com-
pounds 5c±k, compounds 19a, 19b or 19d were then
obtained.
recorded using pure samples of the free amines,
obtained from analytical samples of the corresponding
maleates.
Data of compounds 5c±r are reported in Table 3.
2-Methyl-5-(1-piperazinyl)-7H-isoxazolo[2,3-a]pyrimidin-
7
-one (19a). The reaction carried out with 0.74 g of 18a
.
Preparation of 2-[(1-piperazinyl)methyl]-4H-pyrido[1,2-
a]pyrimidin-4-one (17a) and 7-[(1-piperazinyl)methyl]-
yielded 0.41 g (42%) of 19a 0.5 H O; whitish crystals,
2
ꢀ
(s, 1H, NH; disappeared with D O), 2.49 (s, 3H, CH ),
1
mp 146±147 C (ethyl acetate). H NMR (CDCl ) d 1.81
3
5
5
1
8
H-thiazolo[3,2-a]pyrimidin-5-one (20). A mixture of
.0 mmol of (chloromethyl)derivatives 16a (0.97 g) or
6b (1.00 g), 50.0 mmol (4.31 g) of piperazine and
0 mL of ethanol was re¯uxed for 30 min, while stirring.
2
3
20
2.93 (m, 4H, CH NH), 3.54 (m, 4H, NCH ), 5.33 (s, 1H,
2 2
2
0
H-6), 6.08 (s, 1H, H-3). IR (KBr) 3390 br (OH), 3270
�
1
(NH), 1663 br (CO), 1621, 1583, 1527 br cm . Anal.
.
(C H N O 0.5H O) C, H, N.
11 14 4 2 2
After removal of solvent in vacuo the residue was
partitioned between water and chloroform, then the
aqueous phase was extracted several more times with
chloroform. The combined extracts (dried over anhyd
sodium sulfate) were evaporated to dryness under
reduced pressure to give an oily residue from which,
after treatment with a little ethyl ether, compound 17a
7-(1-Piperazinyl)-5H-thiazolo[3,2-a]pyrimidin-5-one (19b).
Starting from 0.75 g of 18b,18 0.75 g (79%) of 19b was
ꢀ
obtained; white crystals, mp 180±180.5 C (ethyl ace-
1
tate). H NMR (CDCl ) d 1.81 (s, 1H, NH; disappeared
3
with D O), 2.90 (m, 4H, CH NH), 3.60 (m, 4H, NCH ),
5.36 (s, 1H, H-6), 6.74 (d,J =5Hz, 1H, H-2), 7.85 (d,
2,3
2
2
2
(0.55 g, 45%) or 20 (0.60 g, 48%), respectively, sepa-
rated out as a whitish solid.
J =5Hz, 1H, H-3). IR (KBr) 3275 (NH), 1650 br (CO),
3
,2
�
568 sh, 1552, 1513 cm . Anal. (C H N OS) C, H, N, S.
10 12 4
1
1
ꢀ
crystallization from ethyl acetate±petroleum ether with
Compound 17a. Whitish crystals, mp 139±140 C, after
2,3-Dihydro-7-(1-piperazinyl)-5H-thiazolo[3,2-a]pyrimidin-
19
1
charcoal. H NMR (CDCl ) d 1.63 (s, 1H, NH; dis-
5-one (19d). Compound 18c (0.75 g) a€orded 0.65 g
ꢀ
3
appeared with D O), 2.55 (m, 4H, CH NH), 2.98 (m,
4H, NCH ), 3.57 (s, 2H, 2-CH N), 6.73 (s, 1H, H-3),
(68%) of 19d; white crystalline solid, mp 200.5±201.5 C
1
2
2
(ethyl acetate). H NMR (CDCl ) d 1.87 (s, 1H, NH;
2
2
3
6
.99±7.43 (m, 1H, H-7), 7.55±7.90 (m, 2H, H-8,9), 9.10
disappeared with D O), 2.89 (m, 4H, CH NH), 3.20±
3.70 (m, 6H, NCH +2-CH ), 4.43 (t, 2H, 3-CH ), 5.17
2 2 2
2
2
(
m, 1H, H-6). IR (KBr) 3325 (NH), 1683 sh, 1665 br
1
�
(
H, N.
CO), 1624, 1560 w, 1525 cm . Anal. (C H N O) C,
16
(s, 1H, H-6). IR (KBr) 3280 w (NH), 1653 br (CO), 1635
13
4
�
sh, 1580, 1527 cm . Anal. (C H N OS) C, H, N, S.
10 14 4
1
ꢀ
after crystallization from ethyl acetate±isopropyl ether.
Compound 20. Ivory-white crystals, mp 157±159 C,
6,7,8,9-Tetrahydro-2-(1-piperazinyl)-4H-pyrido[1,2-a]-
pyrimidin-4-one (19c). A solution of 5a13 (1.15 g,
5.0 mmol) in 80 mL of ethanol, in the presence of 0.5 g
of 5% palladium on activated charcoal, was subjected
to hydrogenation under 30 psi hydrogen pressure using
a Parr apparatus. The mixture was shaken at room
temperature until the hydrogen absorption ceased (3 h).
The catalyst was then removed by ®ltration and the
solution obtained was evaporated to dryness under
reduced pressure. The resulting colorless oil was treated
with ethyl ether:isopropyl ether (1:1) to give 0.52 g
(44%) of pure compound 19c as white solid; mp 94.5±
1
H NMR (CDCl ) d 1.63 (s, 1H, NH; disappeared with
3
D O), 2.51 (m, 4H, CH NH), 2.97 (m, 4H, NCH ), 3.48
s, 2H, 7-CH N), 6.52 (s, 1H, H-6), 7.10 (d, J2,3=5 Hz,
2
2
2
2
(
1
H, H-2), 8.04 (d, J3,2=5 Hz, 1H, H-3). IR (KBr) 3295
�
1
w (NH), 1663 br (CO), 1636 sh, 1560 sh, 1544 cm
Anal. (C H N OS) C, H, N, S.
.
1
1
14
4
2
(
-(2-Hydroxyethoxy)-4H-pyrido[1,2-a]pyrimidin-4-one
17b). A mixture of 15a17 (0.90 g, 5.0 mmol), anhyd
potassium carbonate (1.0 g) and ethylene glycol (15 mL)
ꢀ
ꢀ
1
was heated with stirring at 160 C for 1 h. After cooling,
95.5 C (ethyl ether). H NMR (CDCl ) d 1.65±2.20 (m,
3
the mixture was poured into ice-water and the resulting
solution was thoroughly extracted with dichloro-
methane. The combined extracts (dried over anhyd
sodium sulfate), after removal of solvent, gave a solid
residue which was taken up in ethyl ether and ®ltered.
Pure compound 17b (0.86 g, 83%) was so obtained;
4H, 7-CH +8-CH ), 1.82 (s, 1H, NH; disappeared with
D O), 2.62±3.20 (m, 6H, CH NH+9-CH ), 3.52 (m,
2
2
2
2
2
4H, NCH ), 3.90 (m, 2H, 6-CH ), 5.39 (s, 1H, H-3). IR
2
2
�
1
(CHCl ) 1655 sh, 1640 s (CO), 1577, 1534 cm . Anal.
3
(C H N O) C,H,N.
18
12
4
ꢀ
1
white crystals, mp 143±144 C (ethyl acetate). H NMR
Biology
(
D O), 3.97 (m, 2H, OCH CH OH), 4.49 (m, 2H,
CDCl ) d 3.15 (broad s, 1H, OH; disappeared with
3
Platelet aggregation. Human blood of healthy volun-
teers was added to a 130 mM trisodium citrate aqueous
solution (volume ratio 9:1), then centrifuged at 100 g for
2
2
2
OCH CH OH), 5.85 (s, 1H, H-3), 7.14 (m, 1H, H-7),
2
2
7.59 (m, 1H, H-9), 7.79 (m, 1H, H-8), 9.06 (m, 1H, H-6).

G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
767
3
0 min to give platelet-rich plasma (PRP). Platelet
The predictive ability of the model was measured as q2
which is de®ned as:
aggregation, performed in an Aggrecoder PA-3210
aggregometer (A. Menarini, Florence, Italy), was mea-
sured following the Born's turbidimetric method and
quanti®ed by the light transmission reached after 3 min.
30
2
q ˆ ꢀSD � PRESS†=SD
ꢀ
PRP (500 mL) was preincubated at 37 C for 2 min with
2
solvent (dimethylsulfoxide, 5 mL) or drug solution
before the addition of the platelet aggregation agent.
PRP aggregation was induced by 5.0 mM ADP (Sigma),
collagen from bovine tendon (Mascia Brunelli) at the
where PRESS ˆ ÆꢁYpred � Yactual†
and SD ˆ
2
ÆꢁYactual � Ymean† .
SD is the sum of squares of deviations of the observed
values from their mean and PRESS is the prediction
error sum of squares.
®
(
nal concentration of 5.0 mg/mL, or 20.0 mM A 23187
Sigma). Before each experiment the stock solutions of
ADP (saline), collagen (saline), and A 23187 (DMSO)
were diluted in saline.
CoMFA was carried out on the TS in two subsequent
steps. Firstly, starting with a number of latent variables
equal to 8 and a number of cross-validation groups
equal to the number of examined compounds, the opti-
mal number of components (ONC) was determined. In
the second step, the ®nal models were developed by a
conventional (non-crossvalidated) regression analysis
with the ONC equal to that yielding the highest q2
value.
Calculation of inhibition. In order to calculate the per-
centage of inhibition, the extent of aggregation mea-
sured in the presence of the compounds tested was
always compared with that measured for a control
sample containing the solvent, in an experiment carried
out under the same conditions. From each series of
experiments, in which the inhibitors were tested in at
least ®ve concentrations, a percentage inhibition±con-
centration curve was derived. From this curve the IC50
value was calculated as the concentration of inhibitor
causing a 50% inhibition of the aggregation. The IC50
values reported in Table 2 are averages (‹standard
deviation) of those obtained from at least four di€erent
batches of platelets (usually 5±8 batches).
Since the ®nal equations are not very useful to represent
eciently the CoMFA models, three-dimensional
graphics or isocontour maps were developed. They
represent areas in the space where steric and electro-
static interactions are responsible for the observed
variations of the biological activity. The graphical
results from our ®nal PLS model P are represented in
4
Figures 1 and 2.
Molecular modeling and CoMFA
Molecular models of the neutral and protonated mole-
cules were constructed with standard bond distances
and angles using the molecular modeling software Sybyl
version 6.4 running on a Silicon Graphics Indigo2
R4400 workstation.
Acknowledgements
2
7
The ®nancial support from MURST (Rome), Uni-
versity of Genova and University of Bari is gratefully
acknowledged.
The conformational analysis was carried out with the
SYSTEMATIC SEARCH option of SYBYL, screening
the conformational space of the selected torsion angles
References and Notes
ꢀ
at 30 increments.
1
. Ermili, A.; Mazzei, M.; Roma, G.; Cacciatore, C. Il Farm-
aco-Ed. Sci. 1977, 32, 375.
. Ermili, A.; Balbi, A.; Di Braccio, M.; Roma, G. Il Farmaco-
Ed. Sci. 1977, 32, 713.
The geometries of all conformers generated in the con-
formational analysis were optimized by molecular
mechanics (Tripos Force Field), and the conformers
having an energy up 2 kcal/mol above the global mini-
mum reoptimized by AM1 Hamiltonian using the
parameter set reported in the MOPAC suite of pro-
2
3
4
6
5
. Ermili, A.; Roma, G. Gazz. Chim. Ital. 1971, 101, 269.
. Ermili, A.; Roma, G.; Balbi, A. Gazz. Chim. Ital. 1971, 101,
51.
. Ermili, A.; Balbi, A.; Roma, G.; Ambrosini, A.; Passerini,
3
1
grams. The partial Mulliken atomic charges from
AM1 calculations were used in the CoMFA study.
N. Il Farmaco-Ed. Sci. 1976, 31, 627.
. Roma, G.; Vigevani, E.; Mazzei, M.; Ermili, A.; Ambrosini,
A.; Passerini, N. Il Farmaco-Ed. Sci. 1978, 33, 822.
6
Ê
In CoMFA a three-dimensional cubic lattice, with a 2 A
7. Mazzei, M.; Balbi, A.; Roma, G.; Di Braccio, M.; Leoncini,
G.; Buzzi, E.; Maresca, M. Eur. J. Med. Chem. 1988, 23, 237.
grid spacing, was generated around the aligned com-
pounds based on the molecular volume of the structures
Ê
8
. Mazzei, M.; Sottofattori, E.; Di Braccio, M.; Balbi, A.;
Leoncini, G.; Buzzi, E.; Maresca, M. Eur. J. Med. Chem. 1990,
5, 617.
(
directions). Steric and electrostatic potentials were gen-
grid beyond the molecules extended by 4.0 A in all
2
3
9. Leoncini, G.; Maresca, M.; Colao, C.; Buzzi, E.; Mazzei,
M.; Balbi, A. Pharmacol. Res. 1991, 23, 139.
erated by a sp carbon atom probe with a charge of +1.
The option `drop electrostatic' was set to NO.
1
0. Leoncini, G.; Maresca, M.; Colao, C.; Buzzi, E.; Mazzei,
M. Cell. Biochem. Funct. 1991, 9, 79.
1. Hermecz, I.; Meszaros, Z. In Advances in Heterocyclic
The CoMFA models were derived by the partial least
squares (PLS) method, implemented in Sybyl, with the
leave-one-out cross validation procedure.32
1
Chemistry; Katritzky, A. R., Ed.; Academic: New York, 1983;
Vol. 33, pp 241±330.

7
68
G. Roma et al. / Bioorg. Med. Chem. 8 (2000) 751±768
23. Di Braccio, M.; Roma, G.; Mazzei, M.; Balbi, A.; Schian-
1
2. Roma, G.; Di Braccio, M.; Leoncini, G.; Aprile, B. Il
Farmaco 1993, 48, 1225.
3. Di Braccio, M.; Roma, G.; Leoncini, G. Eur. J. Med.
Chem. 1995, 30, 27.
4. Di Braccio, M.; Roma, G.; Leoncini, G.; Poggi, M. Il
Farmaco 1995, 50, 703.
5. Leoncini, G.; Signorello, M. G.; Roma, G.; Di Braccio, M.
Biochem. Pharmacol. 1997, 53, 1667.
6. Hermecz, I.; Horvoth, A.; Vasva
zaros, Z. Synthesis 1984, 152
tarelli, P.; Cadel, S.; Bongrani, S. Il Farmaco-Ed. Sci. 1988, 43,
705.
24. Di Braccio, M.; Roma, G.; Grossi, G. C.; Ciarallo, G. J.
Heterocyclic Chem. 1992, 29, 25.
25. Cramer, R. D.; Patterson, D. E.; Bunce, J. D. J. Am.
Chem. Soc. 1988, 110, 5959.
1
1
1
26. Dunn, Jr III, W. J.; Wold, S.; Edlund, U.; Helberg, S.
Quant. Struct.-Act. Relat. 1994, 3, 131.
1
Â
ri-Debreczy, L.; Mes-
27. Sybyl: Molecular Modelling Package; Tripos Ass., St
Louis, MO, USA.
28. Gaillard, P.; Carrupt, P.-A.; Testa, B.; Boundon, A. J.
Comput-Aided Mol. Design 1994, 8, 83.
1
7
1
7. Snyder, H. R.; Robison, M. M. J. Am. Chem. Soc. 1952,
4, 4910.
8. Di Braccio, M.; Roma, G.; Mazzei, M.; Balbi, A.; Testa,
R. Il Farmaco-Ed. Sci. 1986, 41, 183.
9. Glennon, R. A.; Bass, R. G.; Schubert, E. J. Heterocyclic
Chem. 1979, 16, 903.
29. Ingalls, E. A.; Popp, F. D. J. Heterocyclic Chem. 1967, 4,
523.
30. Born, G. V. R. Nature (London) 1962, 194, 927.
31. AM1 quantomechanical semiempirical method can be found
on the package MOPAC 6.0 which is available from QCPE
(506).
1
2
2
0. B'hme, H.; Weisel, K. H. Arch. Pharm. 1977, 310, 26.
1. Roma, G.; Ermili, A.; Di Braccio, M.; Mazzei, M. Il
Farmaco-Ed. Sci. 1982, 37, 747.
2. Roma, G.; Di Braccio, M.; Balbi, A.; Mazzei, M.; Ermili,
A. J. Heterocyclic Chem. 1987, 24, 329.
2
32. Cramer, Jr III, R. D.; Bunce, J. D.; Patterson, D. E.
Quant. Struct.-Act. Relat. 1988, 7, 18.