Synthesis, biological evaluation, and pharmacophore generation of uracil, 4(3H)-pyrimidinone, and uridine derivatives as potent and selective inhibitors of parainfluenza 1 (Sendai) virus

Article abstract of DOI:10.1021/jm010938i

Several new 6-oxiranyl-, 6-oxiranylmethyluracils, and pyrimidinone derivatives, synthesized by lithiation-alkylation sequence of 1,3,6-trimethyluracil, 1,3-dimethyl-6-chloromethyluracil, and 2-alkoxy-6-methyl-4(3H)-pyrimidinones, showed a potent and selective antiviral activity against Sendai virus (SV) replication. To gain insight into the structural features required for SV inhibition activity, the new compounds were submitted to a pharmacophore generation procedure using the program Catalyst. The resulting pharmacophore model showed high correlation and predictive power. It also rationalized the relationships between structural properties and biological data of these inhibitors of SV replication.

Full text of DOI:10.1021/jm010938i

4554
J . Med. Chem. 2001, 44, 4554-4562
Syn th esis, Biologica l Eva lu a tion , a n d P h a r m a cop h or e Gen er a tion of Ur a cil,
4(3H)-P yr im id in on e, a n d Ur id in e Der iva tives a s P oten t a n d Selective In h ibitor s
of P a r a in flu en za 1 (Sen d a i) Vir u s
Raffaele Saladino,*,† Claudia Crestini,† Anna Teresa Palamara,‡,§ Maria Chiara Danti,| Fabrizio Manetti,|
Federico Corelli,| E. Garaci,§ and Maurizio Botta*,|
Dipartimento Agrochimico Agrobiologico, Universita` degli Studi di Viterbo “La Tuscia”, via San Camillo de Lellis,
01100 Viterbo, Italy, Dipartimento di Biologia e Patologia Cellulare e Molecolare L. Califano, Universita` di Napoli Federico II,
via Pansini 5, 80131 Napoli, Italy, Dipartimento di Medicina Sperimentale e Scienze Biochimiche, Universita` di Roma “Tor
Vergata”, via Tor Vergata 135, 00133 Roma, Italy, and Dipartimento Farmaco Chimico Tecnologico, Universita` degli Studi di
Siena, via Aldo Moro, 53100 Siena, Italy
Received May 29, 2001
Several new 6-oxiranyl-, 6-oxiranylmethyluracils, and pyrimidinone derivatives, synthesized
by lithiation-alkylation sequence of 1,3,6-trimethyluracil, 1,3-dimethyl-6-chloromethyluracil,
and 2-alkoxy-6-methyl-4(3H)-pyrimidinones, showed a potent and selective antiviral activity
against Sendai virus (SV) replication. To gain insight into the structural features required for
SV inhibition activity, the new compounds were submitted to a pharmacophore generation
procedure using the program Catalyst. The resulting pharmacophore model showed high
correlation and predictive power. It also rationalized the relationships between structural
properties and biological data of these inhibitors of SV replication.
In tr od u ction
antiviral agent, is effective against PIV in cell culture
and has been used for the treatment of lower respiratory
tract disease in immunocompromized hosts.9 Case re-
ports documenting decreased viral load and clinical
improvement in children with severe combined immu-
nodeficiency following multiple treatments with aero-
solized ribavirin therapy have been described.10 How-
ever, the efficacy of therapy with ribavirin is still under
debate11 and the potential environmental release of drug
has caused concern in hospital personnel because of the
potential teratogenicity in the rodent model.12
At present, there is a lack of effective monotherapies
for specific respiratory viruses, particularly for parain-
fluenza viruses.
Sendai virus has been extensively used in studies that
have defined many of the basic biochemical and molec-
ular biological properties of the paramyxoviruses. More-
over, it has been employed as an in vivo experimental
model of parainfluenza virus infection or for in vitro
studies of new antiviral agents.13
Even if low molecular weight compounds such as
thapsigargin (a specific inhibitor of Ca2+-ATPase),14
ambazone (1,4-benzoquinone-guanylhydrazone-thiosemi-
carbazone),15 dihydroheptaprenol,16 and methylpred-
nisolone acetate17 have been tested, selective and spe-
cific antiviral agents against SV have been reported for
the first time by our research group.18
In particular, some 6-(oxiranylmethyl)uracil deriva-
tives (compounds 2a -g in Scheme 1), prepared during
a general screening of anti-HIV compounds, showed
very interesting activity as inhibitors of Sendai virus
even if associated with a non-negligible toxicity (evalu-
ated as alteration of normal cell morphology, viability,
and count).18
Sendai virus (SV) is a murine subtype of parainflu-
enza 1 viruses (PIVs) belonging to the family of Paramyx-
oviridae. SV is an enveloped virus with nonsegmented
negative strand RNA and six major structural proteins.
Nucleoprotein (NP), large (L) protein, and P protein are
associated with the nucleocapside, while hemagglutinin-
neuraminidase (HN), fusion (F), and matrix (M) proteins
constitute the envelope. The surface glycoproteins F and
HN are the most important viral antigens that induce
humoral immune response in the hosts.1
PIV types 1, 2, and 3 were first discovered in 19562
and were recognized as important human pathogens
causing upper and lower respiratory tract infections. In
particular, they were identified as the major causes of
croup or laryngotracheobronchitis in children.3,4
In older children and adults PIVs cause frequent
reinfections that are generally mild in healthy persons
but may cause serious outcomes in immunocompro-
mized patients or patients with underlying cardiopul-
monary diseases.5,6 Immunocompromized hosts often
have upper respiratory tract symptoms similar to those
experienced by normal hosts, as well as a higher
incidence of lower respiratory tract symptoms. Lower
respiratory tract infection can lead to respiratory failure
and death in bone marrow transplantation recipients
of all ages, especially if the infection occurs in the period
soon after transplantation.7,8
To date, there is no licensed antiviral therapy for the
treatment of PIV infections. Ribavirin, a broad-spectrum
* To whom correspondence should be addressed. For R.S.: phone,
+39 0761 357284; fax, +39 0761 354272; e-mail: saladino@unitus.it.
For M.B.: phone, +39 0577 234306; fax, +39 0577 234333; e-mail:
botta@unisi.it.
† Universita` degli Studi di Viterbo “La Tuscia”.
‡ Universita` di Napoli Federico II.
The goal of this work was the synthesis of new uracil,
pyrimidinone, and uridine derivatives possibly charac-
terized by high affinity and selectivity toward SV. In
§ Universita` di Roma “Tor Vergata”.
| Universita` degli Studi di Siena.
10.1021/jm010938i CCC: $20.00 © 2001 American Chemical Society
Published on Web 11/22/2001

Inhibitors of Parainfluenza 1 Virus
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4555
Sch em e 1
Ta ble 1. Measured and Calculated Activity against
Parainfluenza 1 (Sendai) Virus and Cytotoxicity of the New
Compounds
In addition, to investigate the structure-activity
relationships of these compounds, we have applied the
program Catalyst19 to develop a pharmacophore model
for SV inhibitors. The calculated model, able to rational-
ize the relationships between the chemical features of
the new structures and their biological data, consists
of a hydrophobic region and three hydrogen bond
acceptor features. It shows a good statistical significance
(r ) 0.85, rmsd ) 1.40) and successfully predicts the
affinities of the molecules of, and external to, the
training set.
measured calculated stereochemistry
a
d
ED50
,
ED50
,
assigned by
MTC,c CC50
µM µM
,
compd
µM
µM
the programb
2a
0.24
0.18
0.59
>2
>2
0.42
>2
338
1.11
116
176
0.82
0.13
0.57
2.2
3
0.97
0.72
320
0.75
60
170
24
R
S
S
0.48
4.0
3.7
7.9
89
89
8.4
2b
2c
0.37
>0.79
>2
>2
>0.84
>2
446
>223
>333
>352
330
294
342
735
383
2d e
2ee
2f
S,Rf
S,Sf
R,Rf
S,Sf
S
2ge
3a e
3b
3c
40
1218
1380
547
>352
>330
>294
>342
>360
130
S
S
Ch em istr y
5a
5b
5c
5d
5e
6a
R,Rf
S
Apart from 1-[(2-hydroxyethoxy)methyl]-6-(phenyl-
thio)thymine (HEPT) and 3,4-dihydro-2-alkoxy-6-benz-
yl-4-oxypyrimidine (DABO) derivatives,20 only little
scattered data are available concerning the antiviral
activity of 6-substituted pyrimidines and uracils prob-
ably because the preparation of such compounds is not
a trivial task.21
Our previous work on the synthesis of bipyrimidino-
nes and bipyrimidinylmethane derivatives showed that
the metalation of 6-methyl-4(3H)-pyrimidinones takes
place at the methyl in the 6-position in an essentially
regiospecific manner.22 This metalation-alkylation se-
quence was used for the preparation of several types of
6-substituted uracils.23
To fully exploit the synthetic potential of this proce-
dure, we started to investigate the reaction of the
lithium derivative of 1,3,6-trimethyluracil 1 with several
acyclic and cyclic R-halogenoketones as electrophiles.
The reaction of 1a , prepared from 1 with lithium
diisopropylamide (LDA, 1.2 equiv) in THF at -78 °C,
with chloroacetone, 2-chloroacetophenone, 1-chloropi-
nacolone, 3-chloro-2-butanone, 2-chlorocyclopentanone,
and 2-chlorocyclohexanone, afforded 6-oxiranylmethyl
uracil derivatives 2a -g (Scheme 1) in acceptable yields
(Table 2).
16.6
147
171
42
320
S
R,Sf
S
0.37
14 (4.4)h
76
166
152
146
893
187
164
186
>446
300
195
267
S
S
S
S
S
S
S
S
6b
6ce
6d
6e
130
110
6.9
305
292
893
>376
326
373
446
397
376
>305
170
90
1620
860
816
1082
1218
1016
579
220
220
230
300
290
110
270
1
6fe
6g
6h e
6i
S
S
S
6j
8
10a
10b
10c
375
0.47
0.38
193
R,Rg
R,Sg
R,Rg
>0.81
7.3
7.7
>380
0.57
160
>0.80
380
a
Inhibitory concentration required to reduce virus yield by 50%.
b Assigned to each compound by the program with the aim of fitting
the pharmacophore model (see Experimental Section). c Minimun
toxic concentration required to cause a microscopically detectable
alteration of normal cell morphology. The results listed are the
d
mean values of two or three independent determinations. Con-
centration required to inhibit cell proliferation by 50%. e Com-
pounds constituting the test set. f Refers to the chiral center at
g
the 2′- and 3′-positions, respectively. Refers to the chiral center
h
at the 1′- and 2′-positions, respectively. In parentheses is the
calculated activity for an alternative orientation of 5e in the
pharmacophore model (see text).
In the case of 3-chloro-2-butanone, two possible dia-
stereoisomers, compounds 2d and 2e, were obtained
after chromatographic purification (ratio value 2d /2e )
1:5), and their stereochemistry was determined by a
series of 1D-NMR nuclear Overhauser effect (NOE)
measurements. In particular, the proximity of the 2′-
Me and 3′-Me protons in 2e was revealed by their
mutual NOE effect (8.1%), suggesting the E stereo-
chemistry. On the other hand, the Z stereochemistry of
this paper we report the synthesis and biological evalu-
ation of compounds 3a -c, 5e, 6a -j, 8 (Schemes 1-3
and Table 1), and 12a ,b and 13-15 (Scheme 5) showing
good anti-SV activity associated with very low toxicity.
We also provide a full description of the synthetic
preparation and biological evaluation of compounds 2a -
g, 5a -d , and 10a -c.

4556 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
Saladino et al.
Sch em e 2
Ta ble 2. Chemical and Physical Data of the New Compounds
Sch em e 3
compd
formula
mp, °C
oil
oil
85.8
oil
oil
oil
155-157
73-75
82-84
oil
oil
oil
oil
oil
oil
oil
oil
oil
oil
oil
oil
oil
oil
oil
oil
oil
yield, %
2a
2b
2c
2d
2e
2f
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
10H14N2O3
17H18N2O3
13H20N2O3
11H16N2O3
11H16N2O3
12H16N2O3
13H18N2O3
11H16N2O3
12H18N2O3
17H20N2O3
15H22N2O3
17H20N2O3
20H24N2O3
16H24N2O3
15H16N2O3
17H20N2O3
20H26N2O3
48
65
67
20
47
68
83
21
23
18
40
65
69
73
81
15
36
63
85
35
41
46
33
23
38
78
56
43
35
80
69
47
28
23
2g
3a
3b
3c
5a
5b
5c
5d
5e
6a
6b
6c
6d
6e
6f
6g
6h
6i
6j
8
10a
10b
10c
12a
12b
13
14
15
4b, and 4c, respectively), in the presence of 2-chloro-
cyclohexanone, 2-chloroacetophenone, and 3-chloro-2-
butanone, the reaction sequence afforded 6-oxiranyl-
methyl-4(3H)-pyrimidinone derivatives 5a -e (Scheme
2) in yields ranging from 40% to 81% (Table 2). The
proximity of the 2′-Me and 3′-H protons of 5d was
revealed by their mutual NOE effect (9%), leading to
the conclusion that this compound represents the Z
isomer.
C20H26N2O3
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
13H20N2O3
14H22N2O3
17H26N2O3
15H24N2O3
11H16N2O3
13H20N2O3
14H22N2O3
14H22N2O3
14H16N2O3
14H14N2O3
14H14N2O3
35H32N2O3
40H34N2O3
37H36N2O10
37H36N2O10
42H42N2O10
Compound 4a was prepared by regiospecific N-3
methylation of 2-methoxy-6-methyl-4(3H)-pyrimidinone
with dimethyl sulfate (Me2SO4) in NaOH (30% aqueous
solution) at room temperature in 90% yield. 4b-d were
prepared by the transalkoxylation reaction of the same
substrate with appropriate sodium alkoxides,25 followed
by methylation of the new obtained 2-alkoxy derivatives.
In the latter case, the alkylation of 2-n-propyloxy-6-
methyl-4(3H)-pirimidinone with Me2SO4 in NaOH (30%
aqueous solution) at 25 °C gave the desired N-3 deriva-
tive 4b as the only recovered product. The same reaction
performed with CH2N2 in MeOH at 25 °C afforded both
the N-3 methyl derivative 4b and the O-4 methyl
derivative 7. Compound 7 was also used as starting
material for the metalation-alkylation procedure to
give compound 8 (Scheme 3).
oil
129-131
162
oil
oil
oil
oil
oil
2d was confirmed by the lack of any detectable NOE
effect between 2′-Me and 3′-Me protons.
In a similar way, starting from 1, by application of
the same metalation reaction described above, followed
by quenching with 4-chlorobutyraldehyde,24 5-chloro-
2-pentanone, and 1-phenyl-4-chloro-1-butanone, the
1,3-dimethyl-6-(tetrahydrofuranylmethyl)uracils 3a -c
(Scheme 1) were obtained in yields ranging between
18% and 23% (Table 2).
The same metalation-alkylation sequence performed
on 2-alkyloxy-3,6-dimethyl-4(3H)-pyrimidinones 4a -d ,
using various electrophiles, afforded compounds 5a -e
and 6a -j (Scheme 2). It is interesting to note that the
possible formation of bipyrimidinylmethane derivates
or C-2 transalkoxylation reactions (see below for the
preparation of compounds 4b-d ) were not operative
side processes under these experimental conditions. In
particular, by using 2-methoxy-, 2-n-propyloxy-, and
2-cyclohexyloxy-3,6-dimethyl-4(3H)-pyrimidinones (4a ,
With the aim of applying this procedure to the
synthesis of 6-oxiranyluracil derivatives, we studied the
reaction of the lithium derivative of 1,3-dimethyl-6-
chloromethyluracil 9 with various carbonyl compounds
as electrophiles. The reaction of 9a , prepared from 9
with LDA (1.2 equiv) in THF at -78 °C with acetophe-
none and benzaldehyde afforded compounds 10a -c
(Scheme 4) in acceptable yields (Table 2). In compound
10a , obtained as the only recovered product, the prox-
imity of the 2′-Phe and 1′-H protons was revealed by
their mutual NOE effects (7.5%), suggesting the E
stereochemistry. In the case of diastereoisomers 10b
and 10c (ratio value 10b/10c ) 1:1.2) the stereochem-
istry was assigned on the basis of the value of the
coupling constant between vicinal 1′-H and 2′-H protons
(8.3, and 1.9 Hz, respectively), suggesting the Z stereo-
chemistry for 10b and the E stereochemistry for 10c.

Inhibitors of Parainfluenza 1 Virus
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4557
Sch em e 4
Sch em e 5
Finally, to evaluate the influence on the anti-SV
activity of a sugar substituent at the N-1 of the uracil
ring, 2′,3′,5′-tri-O-benzoyl-3,6-dimethyluridine 1126 was
treated with chloroacetone and 2-chloroacetophenone in
the presence of LDA in dry THF at -78 °C (Scheme 5).
Under these experimental conditions the corresponding
6-(oxiranylmethyl)uridine derivatives 12a ,b were ob-
tained as a chromatographically unseparable mixture
of diastereoisomers (ratio value 1:1) in 80% and 69%
yield, respectively (Table 2). Even if the synthesis of
5-oxiranyluridines is described in the literature,27 to the
best of our knowledge, this is the first report dealing
with the preparation of uridine derivatives bearing the
oxiranyl moiety in the C-6 position. The reaction was
then performed, under similar experimental conditions,
with 5-chloro-2-pentanone as the electrophile. The
6-(tetrahydrofuranylmethyl)uridine derivative 13 and
the unexpected 5-(tetrahydrofuranyl)uridine derivative
14 (Scheme 5) were obtained in 47% and 28% yield,
respectively. The structure of compound 14 was con-
firmed by the presence of an absorption signal of the
more sterically hindered R-halogenoketone, the racemic
3-bromocamphor, was used as an electrophile. In this
case, the corresponding C5-alkylated uridine derivative
15 was obtained as the only recovered product in low
yield (23%) besides the unreacted substrate.
Bioa ssa y Resu lts
All new products have been evaluated for their
antiviral activity against parainfluenza 1 replication in
Madin Darby canine kidney cells (MDCK cells).
To test viral production, supernatants from infected
cells (see Experimental Section) were harvested at
different time points after virus challenge and were
tested by measuring the hemagglutinin units (HAU),
as described by Garaci.28 In confluent MDCK cells,
cytotoxicity was evaluated on the basis of microscopic
examination of cell morphology, evaluation of cell vi-
ability after tripan blue staining, and cell count. The
viability of proliferating mouse myeloma cells (NSO cells
from American Tissue Culture Collection) and normal
human lymphocytes stimulated with PHA (5 µg/mL)
was measured by using tritiated thymidine incorpora-
tion tests. Statistical evaluation of the results allowed
us to calculate CC50 for both proliferating mouse my-
eloma cells and normal human lymphocytes.
1
C-6 methyl moiety in the H NMR, while the integrity
of the C5-C6 double bond was verified by its IR
spectrum. We have not studied in detail the mechanism
of this reaction, but it is reasonable to suggest that the
N-1 sugar moiety in the syn orientation of the nucleoside
may affect the access of a very sterically demanding
electrophile, such as 5-chloro-2-pentanone, to the 6-posi-
tion. Thus, the C-5 alkylation may occur, through a
mesomeric form, as a side reaction. This hypothesis is
further confirmed by a probe experiment where an ever
Most of the analyzed compounds showed interesting
inhibitory activity. In particular, 2a -c, 2f, 5e, and
10a ,b presented an ED50 (the inhibitory concentration
required to reduce virus yield by 50%) lower than
micromolar, while most of the remaining compounds

4558 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
Saladino et al.
were about 3 orders of magnitude less potent (Table 1).
Moreover, 3b presented a micromolar ED50 (1.1 µM)
with a selectivity index of about 200.
Some of the tested compounds showed an associated
toxic effect that, in the case of 2c, 2f, 3b-d , 5a , 6e, and
8, was not microscopically detectable at the concentra-
tion range in which compounds have been found active.
In particular, we report a study that applies a ligand-
based drug design (pharmacophore development) method
to rationalize the relationships between structures and
biological data. Accordingly, experimentally determined
affinities are used to derive a pharmacophore model that
describes the three-dimensional structural properties
required to have profitable interactions with the corre-
sponding binding site on SV.
Resu lts a n d Discu ssion
This computational approach has been applied to a
set of 22 Sendai virus inhibitors (the training set, Table
1) chosen according to the Catalyst guidelines.30 These
compounds are characterized by activity values span-
ning ∼3.7 orders of magnitude, the minimum value of
0.18 µM being associated with compound 2b and the
maximum value of 893 µM being associated with
compound 6d .
On the basis of the biological results reported above,
the following structure-activity relationships can be
tentatively reported. (i) The N,N-dimethyluracil scaffold,
very unusual for antiviral compounds, along with the
6-substitution on the uracil ring, seems to be an
important feature for active compounds. (ii) The sub-
stitution pattern and the stereochemistry of the oxirane
ring play an important role in modulating both the
activity and the toxic effect of the products. In particu-
lar, the 6-oxiranyl derivatives 10a and 10b, and the
6-oxiranylmethyl derivatives 2a -c, 2f, and 5e showed
very interesting activity depending on the size of the
hydrophobic substituent at the 2′-positions, with the
phenyl moiety associated with the best activity values
(2b, 5e, 10a ,b). Moreover, the influence of the stereo-
chemistry on the antiviral activity can be highlighted
in 10b,c, with the former about 3 orders of magnitude
more active. (iii) The substitution of the oxiranyl moiety
with a tetrahydrofuranyl ring resulted in less active
compounds, with the exception of 3b, which is only 1
order of magnitude less active with respect to 2b.
Compound 3b showed an interesting selectivity index
(about 200) as a consequence of decreased toxic effects.
In addition, substituents at the 2′-position on the
tetrahydrofuranyl ring play an important role in modu-
lating both the activity and the toxic effect. In fact,
derivatives bearing a phenyl ring at the 2′-position
showed the highest activities, with 6c being the most
active compound of this series. The presence of a methyl
group at the same position gave rise to less active
products associated with low (6e) or no (6d ) toxicity.
Unsubstituted derivatives are characterized by low
values of inhibitory activity and relevant toxicity. (iv)
Finally, nucleosides 12 and 13 showed an appreciable
antiviral activity.29
From the biological data reported in Table 1, it can
be seen that most of the cytotoxicity values associated
with the compounds are comparable to the values of the
antiviral activity. In fact, anti-SV activity for each
compound is very similar to the MTC and CC50 values.
Few exceptions to this trend are worthy of consideration.
In particular, compound 5b is characterized by a cyto-
toxic concentration about 20-fold higher than the anti-
viral dose (330 vs 16.6 µM, respectively). Moreover,
compound 5e shows a very interesting biological profile
with cytotoxic concentration at least 3 orders of mag-
nitude higher than the antiviral dose (>360 vs 0.37 µM,
respectively). Finally, 3b is characterized by a selectivity
index of about 200, with an antiviral activity 1 order of
magnitude higher than the most active compound 2b.
Com p u ta tion a l Stu d ies. To obtain a better under-
standing of the SARs of the above-described inhibitors
of Sendai virus replication, we performed molecular
modeling and pharmacophore generation experiments
using Catalyst software.
Because no experimental data on the biologically
relevant conformations of the selected compounds (for
example, atomic coordinates derived from X-ray crystal-
lographic studies of their complexes with the putative
receptor) are available, we resorted to a molecular
mechanics approach to build the conformational model
to be used for pharmacophore generation. The molecules
were built within Catalyst and submitted to conforma-
tional analysis. All the conformers of each compound,
within a range of 20 kcal/mol with respect to the global
minimum, have been employed to derive a set of
pharmacophore hypotheses.
The resulting pharmacophore hypotheses use chemi-
cal functions and their spatial location to explain the
differences in inhibitory activity within the training set.
The measured and calculated activity values for both
the training set and the test set are listed in Table 1.
All but 1 of the 10 generated hypotheses have in
common the presence of three hydrogen bond acceptor
groups (HBA) and one hydrophobic region (HY). Only
one selected hypothesis possesses an additional hydro-
phobe as a common chemical function instead of a
hydrogen bond acceptor.
Hypothesis 1, characterized by the highest scoring
and the best statistical parameters (correlation coef-
ficient and root-mean-square deviation (rmsd)), is the
most likely to yield relevant information about the
pharmacophore elements of the studied compounds.
Accordingly, hypothesis 1 has been chosen to represent
“the pharmacophore model”.
The regression analysis of “true” versus “estimated”
SV inhibition activity values for the training set exhib-
ited a correlation coefficient r of 0.85 and an rmsd of
1.40. Moreover, because of the relatively small range of
bits between hypothesis 1 and the null hypothesis (45
bits) corresponding to about a 75% chance of true
correlation, special care was taken to evaluate the
statistical significance of the pharmacophore model. In
particular, a randomized trial procedure derived from
the Fischer method31 was applied to the training set.
Experimental activities were scrambled 19 times to
obtain spreadsheets with randomized data. Nineteen
pharmacophore generation runs were performed using
the scrambled training sets. As a result, among the 190
resulting pharmacophore hypotheses, none was found
with a lower cost than hypothesis 1, suggesting that
there was at least a 95% chance of true correlation in

Inhibitors of Parainfluenza 1 Virus
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4559
F igu r e 2. Compound 10b (Z enantiomer, in red) and 10c
(E enantiomer, in black) mapped to the four-feature best
hypothesis (the pharmacophore model) for Sendai virus inhibi-
tors. The aromatic ring of 10c cannot reach the hydrophobic
region (HY, blue) of the hypothesis. Color codes are as in
Figure 1.
F igu r e 1. Compound 2b mapped to the four-feature best
hypothesis (the pharmacophore model) for Sendai virus inhibi-
tors. Chemical functionalities are color-coded: blue for a
hydrophobic region (HY) and green for hydrogen bond accep-
tors (HBA1, HBA2, and HBA3).
more active than 5a and 5d (24 and 42 µM versus 170
and 320 µM, respectively) because of their phenyl group
that can function as a hydrophobic moiety fulfilling the
hydrophobic site HY of the hypothesis. On the other
hand, the methyl substituent at the 2′-position and the
condensed cyclohexyl ring of compounds 5a and 5d ,
respectively, match HY less well than 5b or 5c. These
considerations suggest that the mapping to the hydro-
phobic feature of the hypothesis could explain the order
of magnitude difference in measured activity for com-
pounds 5b and 5c with respect to 5a and 5d .
the biological data. These findings indicate a reliable
ability to estimate the affinities of the training set.
As a representative example, Figure 1 shows the best-
fitted conformer of 2b (the most active compound of the
training set) into the pharmacophore model character-
ized by three hydrogen bond acceptors and a hydropho-
bic feature. The chemical functionalities of this hypoth-
esis are matched by the chemical groups of 2b. The
phenyl ring fits within the region of the hydrophobic
group (HY), the carbonyl groups at the 2- and 4-position
occupy two hydrogen bond regions (HBA2 and HBA3,
respectively), and finally, the oxirane oxygen is located
in the third hydrogen bond acceptor region (HBA1) of
the hypothesis. The conformer shown has a calculated
energy of 1.1 kcal/mol above the calculated lowest
energy conformer of compound 2b, and its activity is
properly estimated by the pharmacophore hypothesis
(estimated activity of 0.13 µM versus an experimental
value of 0.18 µM).
Compounds 2a , 2c, and 2f showed very similar
orientation with the respect to 2b, the methyl, tert-butyl,
and a portion of the condensed cyclohexyl substituent
at the 2′-position, respectively, lying within the hydro-
phobic region of the hypothesis.
The pharmacophore hypothesis is also able to distin-
guish between 10b (Z stereochemistry, ED50 ) 0.38 µM)
and 10c (E stereochemistry, ED50 ) 193 µM). Figure 2
shows the mapping of these two isomers on hypothesis
1. While compound 10b shows the same alignment on
the hypothesis as compound 2b, the phenyl ring of
compound 10c cannot interact with the hydrophobic
region, resulting in a lower estimated activity (the
activities of compounds 10b and 10c were correctly
estimated in 0.57 and 160 µM, respectively).
Finally, the proposed model is unable to distinguish
between compounds 5b and 5c probably because the
hydrophobic group at the 2′-position, representing the
sole structural difference between the two compounds
(cyclohexyl versus n-propyl, respectively), lies in a region
of space defined by a hydrogen bond acceptor site.
Compounds 5e and 5b have the same alignment on
the hypothesis as 2b, and their lower estimated activity
(14 and 24 µM, respectively) is explained by a missing
hydrogen bond acceptor group interacting with HBA2.
There is also an alternative mapping for 5e (estimated
activity 4.4 µM) that places the alkoxy oxygen atom
much closer to the center of the HBA function, resulting
in the proper interaction with the feature itself. These
findings led to the suggestion that the lack of interaction
involving the HBA feature at the 2′-position of the
inhibitors seems to be important for rationalizing the
trend of the activity values associated with compounds
2b, 5e, 5b, and 5c. In addition, the substituent at the
2′-position, which can allow for interaction with the
HBA feature of the hypothesis, is also the structural
key element in the case of the subset consisting of 3b,
6d , 6e, 6f, and 6g. While compounds 3b and 6d , which
are able to make contact with the HBA2 feature, show
an estimated activity of 0.75 and 6.9 µM, the activities
of 6e-g are predicted to be about 2 orders of magnitude
less (220, 220, and 230 µM, respectively), in good
agreement with experimental data. Moreover, the ad-
ditional lack of activity associated with compounds 3a ,
6h , 6i, and 6j could be attributed to the susbstitution
of the 2′-methyl group with a hydrogen atom and
consequent decreased fitting of superimposition onto the
HY pharmacophore feature.
Compounds 5a -e of the training set are characterized
by structural properties similar to those of compounds
2a -g and, as expected, present similar orientations.
The key difference in the alignment of compounds 2 and
5 to the pharmacophore hypothesis is the consequence
of the alkoxy substitution at the 2′-position. In fact,
compounds 5 cannot reach one (the hydrogen bond
acceptor site HBA2) of the functional regions of the
hypothesis, resulting in a lower estimated activity.
Moreover, compounds 5b and 5c are predicted to be
In light of these considerations, the HBA2 described
above could play a very important role in defining the

4560 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
Saladino et al.
(1 mL), and allowed to warm to room temperature. The organic
layer diluted with EtOAc (60 mL) was then separated, dried
with Na2SO4, and evaporated under reduced pressure. The
residue was purified by flash chromatography using CHCl3/
MeOH ) 9.5:0.5 as the mobile phase to give compound 3a in
15% yield (Table 1): mp 73-75 °C (EtOAc/n-hexane), νmax
(cm-1) 1690, 1665, 1520; δH (CDCl3) 1.80 (2H, m, CH2), 2.01
(2H, m, CH2), 2.46 (2H, m, CH2), 3.34 (3H, s, N-CH3), 3.40 (3H,
s, N-CH3), 3.50 (2H, m, O-CH2), 3.93 (1H, m, CH), 5.60 (1H, s,
CH); δC (CDCl3) 26.13 (CH2), 27.78 (CH3), 30.39 (CH2), 33.31
(CH2), 65.62 (CH), 68.12 (CH2), 97.04 (CH), 153.81 (C), 161.94
(C), 170.0 (C). Anal. (C11H16N2O3) C, H, N. MS: m/e ) 224
(M+).
activity for all the compounds of the training set. In fact,
it is remarkable that the check performed by the
program for surface accessibility of the molecular fea-
tures32 (see Experimental Section) has found that all
the weakly active molecules are lacking a chemical
moiety that could function as a hydrogen bond acceptor
group corresponding to HBA2.
Con clu sion s
A qualitative structure-activity relationship analysis
suggested that the N,N-dimethyluracil scaffold associ-
ated with a 2- and/or 6-substitution seems to show an
important role in determining the activity against SV.
On the other hand, substituents at both the N-1 and
C-5 positions are still suitable in the search for more
active and less toxic SV inhibitors.
More accurate and quantitative SAR considerations
have been made by means of a pharmacophore model
for SV inhibitors. This computationally built model is
able to account for the major SARs associated with the
new compounds, mainly on the basis of a pharmacoph-
ore hydrophobic region and a hydrogen bond acceptor
group. In fact, both these features are common struc-
tural elements of compounds characterized by high
activity toward the Sendai virus.
Gen er a l P r oced u r e for th e Syn th esis of Com p ou n d s
10a -c. An E xa m p le: (E)-1,3-Dim et h yl-6-(2′-m et h yl-2′-
p h en yl-1′-oxir a n yl)u r a cil (10a ). To a solution of 1,3-di-
methyl-6-chloromethyl uracil 9 in dry THF (6 mL) cooled to
-78 °C, freshly prepared lithium diisopropylamide (LDA, 1.5
mmol) was added dropwise under a nitrogen atmosphere. After
the mixture was stirred for 30 min, acetophenone (1.3 mmol)
was added while maintaining the temperature below -70 °C.
The mixture was stirred for 6 h, quenched with a saturated
NH4Cl solution (1 mL), and allowed to warm to room temper-
ature. The organic layer diluted with EtOAc (60 mL) was then
separated, dried with Na2SO4, and evaporated under reduced
pressure. The residue was purified by flash chromatography
using CHCl3/MeOH ) 9.5:0.5 as the mobile phase to give
compound 10a in 56% yield (Table 1): oil, νmax (cm-1) 1700,
1655; δH (CDCl3) 1.86 (3H, s, CH3), 3.17 (3H, s, N-CH3), 3.37
(3H, s, N-CH3), 3.86 (1H, s, CH), 5.69 (1H, s, CH), 7.24 (5H,
m, Ph-H); δC (CDCl3) 23.83 (CH3), 27.71 (CH3), 31.42 (CH3),
61.15 (CH), 65.53 (C), 100.56 (CH), 126.04, 128.38, 128.44 (Ph),
135.48 (C), 148.17 (C), 152.15 (C), 162.10 (C). Anal. (C14H16N2O3)
C, H, N. MS: m/e ) 272 (M+).
Gen er a l P r oced u r e for th e Syn th esis of Com p ou n d s
12a ,b a n d 13-15. An Exa m p le: 3-Meth yl-6-(2′′-m eth yl-2′′-
oxir a n ylm eth yl)-2′,3′,5′-tr i-O-ben zoylu r id in e (12a ). To a
solution of 3,6-dimethyl-2′,3′,5′-tri-O-benzoyluridine 11 in dry
THF (6 mL) cooled to -78 °C, freshly prepared lithium
diisopropylamide (LDA, 1.5 mmol) was added dropwise under
a nitrogen atmosphere. After the mixture was stirred for 30
min, chloroacetone (1.3 mmol) was added while maintaining
the temperature below -70 °C. The mixture was stirred for 6
h, quenched with a saturated NH4Cl solution (1 mL), and
allowed to warm to room temperature. The organic layer
diluted with EtOAc (60 mL) was then separated, dried with
Na2SO4, and evaporated under reduced pressure. The residue
was purified by flash chromatography using CHCl3/MeOH )
9.5:0.5 as the mobile-phase to give compounds 12a in 80% yield
(Table 1): oil, νmax (cm-1) 2988, 1720, 1680, 1540, 1420; δH
(CDCl3) 1.30 (3H, m, CH3), 2.10 (2H, m, CH2), 3.38 (3H, s,
N-CH3), 4.0-4.20 (2H, m, CH), 4.50-4.80 (3H, m, CH + CH2),
5.61 (1H, s, CH), 6.15 (2H, m, CH), 6.64 (1H, m, CH), 7.30-
8.10 (15H, m, Ph-H); δC (CDCl3) 20.80 (CH3), 27.78 (CH3),
38.44 (CH2), 48.39 (CH2), 58.56 (C), 63.59 (CH2), 70.75 (CH),
73.31 (CH), 81.0 (CH), 81.43 (CH), 98.46 (CH), 128.20 (CH),
128.71 (CH), 128.90 (CH), 129.42 (CH), 130.96 (C), 132.60
(CH), 132.92 (C), 135.59 (CH), 146.85 (C), 159.80 (C), 160.41
(C), 162.31 (C), 166.84 (C), 170.0. Anal. (C35H32N2O10) C, H,
N. MS: m/e ) 640 (M+).
Exp er im en ta l Section
Ch em istr y. NMR spectra were recorded on a Bruker (200
MHz) spectrometer and are reported in δ values. Mass spectra
were recorded on a VG 70/250S spectrometer with an electron
beam of 70 eV. Elemental analyses were performed by a Carlo
Erba 1106 analyzer. Infrared spectra were recorded on a
Perkin-Elmer 298 spectrophotometer using NaCl plates. All
solvents are ACS reagent grade and were redistilled and dried
according to standard procedures. Melting points were re-
corded on a Mettler FP-80 apparatus. Chromatographic pu-
rifications were performed on columns packed with Merck
silica gel 60, 230-400 mesh for the flash technique. Thin-layer
chromatography was carried out using Merck platten Kieselgel
60 F254.
Gen er a l P r oced u r e for th e Syn th esis of Com p ou n d s
2a -g a n d 5a -e. An Exa m p le: 1,3-Dim eth yl-6-(2′-m eth yl-
2′-oxir a n ylm eth yl)u r a cil (2a ). To a solution of compound 1
(1 mmol) in dry THF (6 mL) cooled to -78 °C, freshly prepared
lithium diisopropylamide (LDA, 1.2 mmol) was added dropwise
under nitrogen atmosphere. After the mixture was stirred for
30 min, chloroacetone (1.3 mmol) was added while maintaining
the temperature below -70 °C. The mixture was stirred for 6
h, quenched with a saturated NH4Cl solution (1 mL), and
allowed to warm to room temperature. The organic layer
diluted with EtOAc (60 mL) was then separated, dried with
Na2SO4, and evaporated under reduced pressure. The residue
was purified by flash chromatography using CHCl3/MeOH )
9.5:0.5 as the mobile phase to give compound 2a in 48% yield
(Table 1): oil, νmax (cm-1) 1690, 1660, 1490, 1370; δH (CDCl3)
1.34 (3H, s, CH3), 2.75 (2H, dd, J ) 14.44 Hz, CH2), 3.21 (3H,
s, N-CH3), 3.43 (3H, s, N-CH3), 3.49 (2H, dd, J ) 14.44 Hz,
CH2), 5.62 (1H, s, C5-H); δC (CDCl3) 25.48 (CH3), 27.92 (CH3),
33.12 (CH3), 40.00 (CH2), 52.51 (CH2), 72.41 (C), 102.94 (CH),
151.46 (C), 152.59 (C), 162.36 (C). Anal. (C10H14N2O3) C, H,
N. MS: m/e ) 210 (M+).
Biology. SV Gr ow th In h ibition Assa y in MDCK Cells.
MDCK cells were grown in RPMI 1640 supplemented with 5%
fetal calf serum (FCS) and antibiotics in a 5% CO2 atmosphere.
Confluent cell monolayers were infected with SV (10 HAU/
105 cells) prepared by allantoic inoculation of embrionated
eggs. After incubation for 1 h at 37 °C (adsorption period), virus
inocula were removed, monolayers were washed three times
with PBS and incubated with fresh medium containing 2%
FCS. All the products were added, at the appropriate dilution,
after the adsorption period and maintained in the culture
media until the end of the experiments.
Gen er a l P r oced u r e for th e Syn th esis of Com p ou n d s
3a -c, 6a -j, a n d 8. An Exa m p le: 1,3-Dim eth yl-6-(2′-tet-
r a h yd r ofu r ylm eth yl)u r a cil (3a ). To a solution of compound
4a (1 mmol) in dry THF (6 mL) cooled to -78 °C, freshly
prepared lithium diisopropylamide (LDA, 1.2 mmol) was added
dropwise under a nitrogen atmosphere. After the mixture was
stirred for 30 min, 5-chloro-2-pentanone (1.3 mmol) was added
while maintaining the temperature below -70 °C. The mixture
was stirred for 6 h, quenched with a saturated NH4Cl solution
Molecu la r Mod elin g a n d P h a r m a cop h or e Gen er a tion
Stu d ies. A training set of 22 derivatives with biological data

Inhibitors of Parainfluenza 1 Virus
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4561
ranging from 0.18 to 893 µM was selected from the original
set and used in this study.
The ideal and null hypotheses have a cost of 90.76 and
163.01 bits, respectively, while the 10 resulting hypotheses
possessed costs from 118.30 to 133.45 bits. Although the
program assigns a general score to each hypothesis generated
(based on the correlation between the observed and estimated
activities for each compound in the training set), a thorough
analysis of the structural and functional fitting of compounds
listed in Table 1 was carried out in order to select the
hypothesis that would represent better the most active com-
pounds. We expected that the best hypothesis selected (hy-
pothesis 1) corresponds to the most probable common arrange-
ment of chemical functionalities for this set of Sendai virus
inhibitors. The best hypothesis, which was confirmed by the
Fischer test to have at least a 95% chance of true correlation,
yielded a low rmsd of 1.40 and an r of 0.85, indicating a good
correlation among the data in the training set.
To find a pharmacophore able to describe the ligand-
receptor interactions, instead of a superimposition approach
based on atoms or groups of the chemical structures, the
following steps have been carried out by means of the Catalyst
software.
i. En er gy Min im iza tion , Geom etr y Op tim iza tion , a n d
Con for m a tion a l An a lysis of a ll th e Com p ou n d s of th e
Tr a in in g Set. Because no experimental data on the biologi-
cally relevant conformations of these compounds are available
(for example, atomic coordinates derived from X-ray crystal-
lographic studies of the complexes between the receptor with
an embedded inhibitor), there was a need to build their
conformational models to be used in the hypothesis generation
step. Moreover, it is well-known that the biological activity of
chiral compounds is usually due to one of the diastereoisomers
or enantiomers. Because of this fact and taking into account
the presence of at least one asymmetric carbon atom common
to all the compounds of the training set, it was arbitrarily
decided to model the chiral compounds with undefined chiral-
ity, allowing the pharmacophore model generation procedure
to choose which configuration of the asymmetric carbon atoms
was the most appropriate. Compounds 5d and 10a -c were
modeled with the appropriate Z,E stereochemistry and with
undefined R,S chirality. The enantiomer or diastereoisomer
of each compound chosen by the program to fit the proposed
pharmacophore model has been reported in Table 1. To the
best of our knowledge, there are no experimental data sup-
porting the hypothesis that specific enantiomers or diastere-
oisomers are more active than the corresponding mirror-image
enantiomers or diastereoisomers because of the fact that any
enantiopure compound has been evaluated for its inhibitory
properties against the Sendai virus.
The validity and predictive power of hypothesis 1 were
further assessed using anti-SV compounds outside the training
set. The predicted activities of such compounds, calculated by
Catalyst on the basis of the pharmacophore model, are
reported in Table 1.
Ack n ow led gm en t. Financial support from Italian
MURST and Italian Ministry of Healt is acknowledged.
Italian Research National Council of Italy (CNR Target
Project on “Biotechnology”) is also acknowledged. M.
Botta expresses thanks for a grant from the Merck
Research Laboratories (Academic Development Program
in Chemistry).
Su p p or tin g In for m a tion Ava ila ble: NMR and MS data.
This material is available free of charge via the Internet at
http://pubs.acs.org.
The compounds used in this study were built using the
Catalyst 2D-3D sketcher, and a representative family of
conformations were generated for each molecule using the
Poling algorithm33 and the “best conformational analysis”
method. This approach has been selected because it represents
the method of choice when the conformations are to be used
for hypothesis generation, as in this case.
Refer en ces
(1) Lamb R. A.; Kolakofsky D. Paramyxoviridae and Their Replica-
tion. In Fields Virology, 3rd ed.; Fields B. N., Knipe D. M.,
Howley, P. M., Eds.; Lippincott-Raven Publishers: Philadelphia,
1996; pp 1177-1194
(2) Morris, J . A.; Blount, R. E., J r.; Savage, R. E. Recovery of
Cytopathogenic Agent from Chimpanzees with Coryza. Proc. Soc.
Exp. Biol. Med. 1956, 192, 544-549.
(3) Chanock, R. M. Association of a New Type of Cytopathogenic
Myxovirus with Infantile Croup. J . Exp. Med. 1956, 104, 555-
563.
(4) Chanock, R. M.; Parrott, R. H.; Cook, K. Newly Recognized
Myxoviruses from Children with Respiratory Disease. N. Engl.
J . Med. 1958, 258, 207-213.
(5) Welliver, R.; Wong, D. T.; Choi, T. S.; Ogra, P. L. Natural History
of Parainfluenza Virus Infection in Childhood. J . Pediatr. 1982,
101, 180-187.
(6) Todd Faulks, J .; Drinka, P. J .; Shult, P. A Serious Outbreak of
Parainfluenza Type 3 on a Nursing Unit. J . Am. Geriatr. Soc.
2000, 48, 1216-1218.
(7) Wendt, C. H.; Weisdorf, D. J .; J ordan, M. C.; Balfour, H. H., J r.;
Hertz, M. I. Parainfluenza Virus Respiratory Infection after
Bone Marrow Transplantation. N. Engl. J . Med. 1992, 326, 921-
926.
(8) Hohenthal, U.; Nikoskelainen, J .; Vainionpaa, R.; Peltonen, R.;
Routamaa, M.; Itala, M.; Kotilainen, P. Parainfluenza Virus
Type 3 Infections in a Hematology Unit. Bone Marrow Trans-
plant. 2001, 27, 295-300.
Conformations were selected that fell within 20 kcal/mol
range above the lowest energy conformation found.
ii. Gen er a tion of Bin d in g Hyp oth eses for Com p ou n d s
of t h e Tr a in in g Set a n d An a lysis of t h e Gen er a t ed
Hyp oth eses. The training set of molecules, with their associ-
ated conformational models, was submitted to hypothesis
generation. The chemical functions used in this generation step
included “hydrogen bond acceptor lipid” (HBA), “hydrophobic”
(HY), and “ring aromatic” (RA).34 The hydrogen bond acceptor
lipid includes the basic nitrogen not considered in the hydrogen
bond feature. A ring aromatic function was used in addition
to the generic hydrophobic one in order to emphasize that an
aromatic interaction is likely to be an important feature
because some of the compounds present an aromatic ring
within their structure. Moreover, the ring aromatic function
allows consideration of directionality of the interactions with
the receptor showing the ring centroid and a projected point
along the normal to the ring plane.
Because of the small molecular size of compounds of the
training set, the minimum permitted interfeature spacing was
reduced to 1 Å. The default value of about 3 Å works well for
most medium to large molecules, but it is not good for small
molecules that do not have many features or for hypotheses
where features are close together. In addition, the hypothesis
generator was constrained to report only hypotheses with four
features.
(9) Gelfand, E. W.; McCurdy, D.; Rao, C. P.; Middleton, P. J .
Ribavirin Treatment of Viral Pneumonitis in Severe Combined
Immunodeficiency. Lancet 1983, 2, 732-733.
(10) McIntosh, K.; Kurachek, S. C.; Cairns, L. M.; Burns, J . C.;
Goodspeed, B. Treatment of Respiratory Viral Infection in an
Immunodeficient Infant with Ribavirin Aerosol. Am. J . Dis.
Child. 1984, 138, 305-308.
(11) Elizaga, J .; Olavarria, E.; Apperley, J .; Goldman, J .; Ward, K.
Parainfluenza Virus 3 Infection after Stem Cell Transplant:
Relevance to Outcome of Rapid Diagnosis and Ribavirin Treat-
ment. Clin. Infect. Dis. 2001, 32, 413-418.
(12) Canonico, P. G.; Kende, M.; Huggins, J . W. Toxicology and
Pharmacology of Ribavirin in Experimental Animals. In Clinical
Applications of Ribavirin; Smith R. A., Knight V., Smith J . A.
D., Eds.; Academic Press: Orlando, FL, 1984; pp 65-98.
To be considered available, a chemical function must be
surface accessible and thus able to interact with the receptor.32
On the basis of these considerations, while the carbonyl and
methoxy groups are classified as HBA by the program (see
compounds 2b and 5e), oxygen atoms of substituents larger
than methoxy groups are buried and thus unable to interact
with the hypothetical receptor.

4562 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
Saladino et al.
(13) (a) Ishida, N.; Homma, M. Sendai Virus. Adv. Virus Res. 1978,
23, 349-383. (b) Amici, C.; Palamara, A. T.; Garaci, E.; Santoro,
M. G. Inhibition of Sendai Virus Replication by ∆12-Prostag-
landin J 2: Induction of Heat Shock Protein Synthesis and
Alteration of Protein Glycosylation. Antiviral Res. 1992, 19, 129-
138.
(14) Ono, A.; Kawakita, M. Transport of Envelope Proteins of Sendai
Virus, HN and FO, Is Blocked at Different Steps by Thapsigargin
and Other Perturbants to Intracellular Ca2+. J . Biochem. (Tokyo)
1994, 116, 649-656.
(15) Iliescu, R.; Calinoiu, A.; Repanovici, R.; Popa, L. M.; Loeber, G.
Effect of Ambazone on the Multiplication and Biological Activi-
ties of Sendai Virus. Stud. Biophys. 1987, 117, 111-116.
(16) Iida, J .; Ishihara, C.; Mizukoshi, N.; Kitoh, K.; Tsukidate, K.;
Katsu, K.; Toyosawa, T.; Azuma, I. Prophylactic Activity of
Dihydroheptaprenol, a Synthetic Polyprenol Derivative, against
Sendai Virus Infection in Mice. Vaccine 1990, 8, 376-380.
(17) Kimsey, P. B.; Goad, M. E.; Zhao, Z. B.; Brackee, G.; Fox, J . G.
Methyl Prednisolone Acetate Modulation of Infection and Sub-
sequent Pulmonary Pathology in Hamsters Exposed to Parain-
fluenza-1 Virus (Sendai). Am. Rev. Respir. Dis. 1989, 140, 1704-
1711.
(18) Saladino, R.; Danti, M. C.; Mincione, E.; Crestini, C.; Palamara,
A. T.; Savini, P.; Marini, S.; Botta, M. A Potent and Selective
Inhibition of Parainfluenza 1 (Sendai) Virus by New 6-Oxiranyl-,
6-Methyloxiranyluracils, and 4(3H)-Pyrimidinone Derivatives.
Bioorg. Med. Chem. Lett. 1998, 8, 1833-1838.
(23) Botta, M.; Saladino, R.; Delle Monache, G.; Gentile, G.; Nicoletti,
R. Research on Antiviral Agents. 5. Lithiation of 6-Methyluracil
as
a New and Efficient Entry to C(6)-Substituted Uracils.
Heterocycles 1996, 43, 1687-1696.
(24) 4-Chlorobutyraldeide was obtained by oxidation of 4-chloro-1-
butanol with pyridinium chlorocromate (PCC) in CH2Cl2 at 25
°C.
(25) Botta, M.; De Angelis, F.; Finizia, G.; Gambacorta, A.; Nicoletti,
R. 6-Alkyl- and 5,6-Dialkyl-2-methoxy-4(3H)-pyrimidinones in
the Transformation of Pyrimidines. Conversion into 2-Substi-
tuted Amino- and 4-Chloro-pyrimidine Derivatives. Synth. Com-
mun. 1985, 15, 27-34.
(26) Compound 11 was prepared in a regiospecific manner starting
from commercially available 1-O-acetyl-2,3,5-tri-O-benzoyl-â-D-
ribofuranose and 2,4-dihydroxy-6-methylpyrimidine as described
in ref 35 followed by subsequent N-3 methylation with CH2N2
in MeOH at 25 °C.
(27) Botta, M.; Saladino, R.; Occhionero, F.; Nicoletti, R. Recent
Advances in the Chemistry of 6-Substituted Pyrimidines and
Uracil Derivatives. Trends Org. Chem. 1995, 5, 57-81.
(28) Garaci, E.; Palamara, A. T.; Di Francesco, P.; Favalli, C.; Ciriolo,
M. R.; Rotilio, G. Glutathione Inhibits Replication and Expres-
sion of Viral Proteins and Cultured Cells Infected with Sendai
Virus. Biochem. Biophys. Res. Commun. 1992, 188, 1090-1096.
(29) ED50, MTC, and CC50 values for compounds 12-15 (expressed
as µM concentrations) are reported below. 12a : 97, 156, 220.
12b: >142, 142, 369. 13: 151, 151, 386. 14: 90, 151, 432. 15:
77, 136, 313.
(19) Catalyst 4.5 is from MSI Software, Scranton Road, MA.
(20) (a) Tanaka, H.; Takashima, H.; Ubasawa, M.; Sekiya, K.; Nitta,
I.; Baba, M.; Shigeta, C. S.; Walzer, R. T.; De Clercq, E.;
Miyasaka, T. Synthesis and Antiviral Activity of Deoxy Analogs
of 1-[(2-Hydroxyethoxy)methyl]-6-(phenylthio)thymine (HEPT)
as Potent and Selective Anti-HIV-1 Agents. J . Med. Chem. 1992,
35, 4713-4719 and references therein. (b) Artico, M.; Massa,
S.; Mai, A.; Marongiu, M. E.; Piras, G.; Tramontano, E.; La Colla,
P. 3,4-Dihydro-2-alkoxy-6-benzyl-4-oxopyrimidines (DABOs): A
New Class of Specific Inhibitors of Human Immunodeficiency
Virus Type 1. Antiviral Chem. Chemother. 1993, 4, 361-368.
(21) Tanaka, H.; Baba, M.; Hayakawa, H.; Sakamaki, T.; Miyasaka,
T.; Ubasawa, M.; Takashima, H.; Sekiya, K.; Shigeta, S.; Walzer,
R. Y.; Balzarini, J .; De Clercq, E. A New Class of HIV-1 Specific
6-Substituted Acyclouridine Derivatives: Synthesis and Anti-
HIV-1 Activity of 5- or 6-Substituted Analogues of 1-[(2-Hy-
droxyethoxy)methyl]-6-(phenylthio)thymine (HEPT). J . Med.
Chem. 1991, 34, 349-357.
(30) The rules for picking training sets are reported at http://
www.msi.com/support/catalyst/hypogen.html.
(31) Fischer, R. The Principle of Experimentation, Illustrated by a
Psycho-Physical Experiment. In The Design of Experiments;
Hafner Publishing Co.: New York, 1966; Chapter II.
(32) For example, a hydrogen bond acceptor, such as a carbonyl group
or an alkoxy group, is not perceived unless the environment
around the group itself is sufficiently uncrowded.
(33) Smellie, A.; Teig, S.; Towbin, P. Poling: Promoting Conforma-
tional Coverage. J . Comput. Chem. 1995, 16, 171-187.
(34) Greene, J .; Kahn, S.; Savoj, H.; Sprague, P.; Teig, S. Chemical
Function Queries for 3D Database Search. J . Chem. Inf. Comput.
Sci. 1994, 34, 1297-1308.
(35) Botta, M.; De Angelis, F.; Finizia, G.; Nicoletti, R.; Delfini, M.
6-Alkyl- and 5,6-Dialkyl-2-methoxy-4(3H)-pyrimidinones in the
Transformations of Pyrimidines. Regiospecific 1-N-Acylation of
Pyrimidines. Tetrahedron Lett. 1985, 26, 3345-3348 and refer-
ences therein.
(22) Botta, M.; Saladino, R.; Gambacorta, A.; Nicoletti, R. An Unusual
Condensation of Pyrimidinones: Synthesis of Bipyrimidinones
and Bipyrimidinylmethane. Heterocycles 1991, 32, 1537-1545.
J M010938I