SAR of a series of anti-HSV-1 acridone derivatives, and a rational acridone-based design of a new anti-HSV-1 3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyridine series

Article abstract of DOI:10.1016/j.bmc.2007.09.032

Herpes Simplex Virus (HSV) infections are among the most common human diseases. In this work, we assess the structural features and electronic properties of a series of ten 1-hydroxyacridone derivatives (1a-j) recently described as a new class of non-nucl

Full text of DOI:10.1016/j.bmc.2007.09.032

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 313–321
SAR of a series of anti-HSV-1 acridone derivatives,
and a rational acridone-based design of a new anti-HSV-1
3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyridine series
b,
b
Alice M. R. Bernardino,a,* Helena C. Castro, Izabel C. P. P. Frugulhetti,
*
Natalia I. V. Loureiro, Alexandre R. Azevedo,a Luiz C. S. Pinheiro,a
b,c
´
Thiago M. L. Souza,b Viveca Giongo,b Fabiana Passamani,c Uiaran O. Magalhaes,c
˜
Magaly G. Albuquerque,d Lu´cio M. Cabralc and Carlos R. Rodriguesa,c,
*
a
ˆ
Universidade Federal Fluminense, Instituto de Quı´mica, Departamento de Quı´mica Organica,
ˆ
Programa de Po´s-Graduac¸a˜o em Quı´mica Organica, Campus do Valonguinho, 24210-150 Niteroi, RJ, Brazil
´
bUniversidade Federal Fluminense, Instituto de Biologia, Departamento de Biologia Celular e Molecular, 24020-150 Nitero´i, RJ, Brazil
cUniversidade Federal do Rio de Janeiro, Faculdade de Farma´cia, ModMolQSAR, 24020-150 Rio de Janeiro, RJ, Brazil
d
ˆ
Universidade Federal do Rio de Janeiro, Instituto de Quı´mica, Departamento de Quı´mica Organica,
21949-900 Rio de Janeiro, RJ, Brazil
Received 18 May 2007; revised 16 September 2007; accepted 19 September 2007
Available online 22 September 2007
Abstract—Herpes Simplex Virus (HSV) infections are among the most common human diseases. In this work, we assess the struc-
tural features and electronic properties of a series of ten 1-hydroxyacridone derivatives (1a–j) recently described as a new class of
non-nucleoside inhibitors of Herpes Simplex Virus-1 (HSV-1). Based on these molecules, we applied rigid analogue and isosteric
replacement approaches to design and synthesize nine new 3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyridine derivatives (2a–i). The
biological and computational results of these new molecules were compared with 1-hydroxyacridones. An inhibitory profile was
observed in 10-Cl substituted 3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyridine derivative (2f), which presents the same substituent
at the analogous position of 1-hydroxyacridone derivative (1b). The structure–activity relationship (SAR) studies pointed out the
10-position next to nitrogen atom as important for the anti-HSV-1 profile in the pyrazolo-naphthyridine derivatives tested, which
reinforced the promising profile for further experimental investigation. The most potent acridone and pyrazolo-naphthridine deriv-
atives were also submitted to an in silico ADMET screening in order to determine their overall drug-score, which confirmed their
potential antiviral profile.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
infections.2 Symptomatic or asymptomatic recurrent
infections are common and maternal transmission of
HSV-2 to the fetus or neonate may result in herpes infec-
tions, which are rare but severe and potentially fatal.1–3
Herpes Simplex Virus (HSV) infections are among the
most common human diseases.1 Currently, there are
two HSV serotypes (HSV-1 and HSV-2) that may cause
oral disease, keratoconjutivitis, encephalitis, and genital
Most of clinical anti-herpes virus compounds are nucle-
oside analogues, such as acyclovir (ACV), which is the
most common drug used on treatment of HSV infec-
tions.3,4 However, during long ACV-based treatment,
drug-resistant HSV strains frequently emerge. In addi-
tion, these strains often show cross-resistance with other
anti-HSV agents, such as pencyclovir and gancyclovir.5,6
In order to manage ACV-resistant strains, other non-
nucleoside analogue, such as foscarnet, have been used.7
Keywords: Structure–activity relationship (SAR); Acridone; Pyrazolo-
naphthyridine; Antiviral agent; HSV-1; In silico ADMET screening.
*
Corresponding authors. Tel.: +55 21 26292146; fax: +55 2126292135
(A.M.R.B.); tel.: +55 21 25626444 (C.R.R.); +55 21 26292294
(H.C.C.); e-mail addresses: alicerolim@globo.com; gqolice@vm.uff.
br; hcastrorangel@vm.uff. br; hcastrorangel@yahoo.com.br; rangel
@pharma.ufrj.br
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.09.032

314
A. M. R. Bernardino et al. / Bioorg. Med. Chem. 16 (2008) 313–321
Nonetheless, resistance to foscarnet is also emerging,
which demands new antiviral agents for HSV
treatment.6,7
compounds. By using traditional medicinal chemistry
principles, such as isosteric replacement and rigid ana-
logue approaches, we designed a new potential antiviral
heterocycle scaffold, namely 3H-benzo[b]pyrazolo[3,4-
h]-1,6-naphthyridine (Scheme 1), based on two heterocy-
cle systems, 1H-pyrazolo[3,4-b]pyridine-5-carboxylic
acid, presenting a diverse antiviral profile,13–15 and the
former synthetic acridone derivatives. To investigate
the impact of the isosteric aromatic ring replacement
and the influence of the restricted flexibility imposed
by the new naphthyridine system on the antiviral activ-
ity, nine derivatives (2a–i) from this new series (Table 2)
were synthesized, tested for anti-HSV-1 effects, and eval-
uated in SAR studies where they were compared with
Natural and synthetic acridone derivatives have been
identified as new heterocyclic systems with anti-HSV
activity.20–25 In 1989, Yamamoto and co-workers8 re-
ported a new acridone alkaloid, citrusinine-I (Fig. 1),
which exhibited potent activity against HSV-1 and
HSV-2.8 More recently, Lowden and Bastow reported a
new synthetic series of 1-hydroxyacridones with two
1,3-dihydroxyacridone derivatives showing potent
HSV-1 inhibitory activity, namely, 5-methoxy-1,3-dihydr-
oxyacridone (1a, ED50 = 2.2 lM) and 5-chloro-1,3-
dihydroxyacridone (1b, ED50 = 4 1 lM) (Fig. 1) .10–12
R
H
N
In this work, we used a molecular modeling approach
for developing structure–activity relationship (SAR)
studies on ten of these acridones (Table 1) thus we iden-
tified molecular properties important to the anti-HSV-1
activity, which may be used for designing new antiviral
R1
N
H
N
OH
N
Ph
R2
N
HO2C
O
OH
(
)
1a-i
(
)
3a-i
OH
OMe
R
Me
N
H
N
Rigid
analog
approach
Isosteric
replacement
approach
OMe
OH
R1
R1
H
O
O
N
N
OH
OH
N
N
1a, R=OMe
1b, R=Cl
N
N
Ph
(
)
Ph
Citrusinine-I
R2
R2
N
N
Figure 1. Structures of some acridone derivatives with anti-herpes
virus activity: citrusinine-I, a natural acridone alkaloid, and two
synthetic acridone derivatives 1a (5-methoxy-1,3-dihydroxyacridone)
and 1b (5-chloro-1,3-dihydroxyacridone).
R1=H, Me
R2=H, Me, OMe, Cl
Cl
O
(
)
(
)
2c-i
2a-b
Scheme 1. Isosteric replacement and rigid analogue approaches
applied into the design of the 3H-benzo[b]pyrazolo[3,4-h]-1,6-naph-
thyridine derivatives.
Table 1. Anti-HSV-1 activities (ED50, lM) and the calculated HOMO
(EHOMO, Ev) and LUMO (ELUMO, Ev) energies and molecular dipole
moments (l, Debye) of ten 1-hydroxyacridones
Table 2. Anti-HSV-1 activities (HSV-1
% inhibition at 50 lM;
Cl
H
H
mean SEM) and the calculated HOMO (EHOMO, Ev) and LUMO
(ELUMO, Ev) energies and molecular dipole moments (l, Debye) of
nine 3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyridines
5
4
4
N
R3
N
6
7
6
7
10
10
R
9
2
9
1
1
8
8
R1
R1
N
N
N
H
O
OH
O
1a-1i
OH
1
2 N
10
7
1j
N
N
Ph
Ph
3
9
11
11
R2
R2
a
Compound
R
R3
ED50
EHOMO ELUMO
l
N
N
6
6
8
4
5
1a
1b
1c
1d
1e
1f
5-OMe OH
2.2
4
À7.75
À8.16
À7.89
À8.22
À8.06
À8.06
À7.98
2.24
1.86
2.19
1.88
1.95
1.86
2.14
1.90
1.82
1.13
6.69
4.49
6.53
5.84
7.56
7.62
6.30
4.95
4.27
2.80
O
Cl
(
)
(2c-i)
2a-b
5-Cl
5-Me
6-Cl
8-Cl
7-Cl
H
OH
OH
OH
OH
OH
OH
1
1
8
Compound R1 R2
HSV-1
EHOMO ELUMO
l
10
10
16
45
1
2
4
2a
2b
2c
2d
2e
2f
H
H
74
92
2
4
À8.15
À8.20
1.52
1.63
5.60
5.39
5.26
5.45
4.85
5.37
2.47
4.60
5.44
H
8-Me
8-Me
8-OMe
H
1g
1h
1i
H
82 1.3 À7.83
0.06
0.01
5-Cl
5-Cl
OMe >50(6) À8.08
H
Me
68
70
2
6
À7.69
À7.84
Me
b
NA
NA
À7.97
À7.39
0.04
b
1j
Me 10-Cl
Me 8-Cl
100 1.1 À7.91
39 3.1 À7.93
66 0.8 À7.81
À0.21
À0.19
0.07
2g
2h
2i
NA, not active.
Me 10-Me
Me 8-Me
a ED50 values were taken from Lowden & Bastow [22]. Parenthetical
value = % inhibition at stated concentration.
b Structure shown above.
NA
À7.75
0.08
NA, not active.

A. M. R. Bernardino et al. / Bioorg. Med. Chem. 16 (2008) 313–321
315
those from the 1-hydroxyacridone derivatives. More-
R2
over, the most potent acridone and 2a–i were submitted
to an in silico ADMET screening in order to identify
their promising biological potential.
H
Cl
N
N
N
R1
N
R2
R1
N
CO2Et
CO2Et
NH2
N
N
2h
80-56%
2. Computational and experimental methods
Ph
(
)
4a-i
Ph
(
)
5a-b
2.1. Molecular modeling, SAR studies, and in silico
ADMET screening
Alkaline
hydrolysis
98-88%
R1
R2
The molecular modeling study was performed using
SPARTAN’04 software package (Wavefunction Inc., Ir-
vine, CA, 2000). The minimum energy conformation of
ten acridone derivatives and 2a–i compounds was
obtained by the AM1 semiempirical Hamiltonian. In
order to better evaluate the electronic properties of the
AM1 minimum energy conformations, they were sub-
mitted to a single-point energy ab initio calculation at
the 6-311G* level.
N
N
H
N
H
N
N
R1
N
Ph
CO2H
R2
POCl3
N
1h
70-69%
N
O
(
)
3a-i
Ph
(
)
2a-b
POCl3
3h
73-60%
R1=H, Me
R2=H, Me, OMe, Cl
R1
In order to perform structure–activity relationship
(SAR) studies, some electronic properties, such as
HOMO (Highest Occupied Molecular Orbital) and
LUMO (Lowest Unoccupied Molecular Orbital) en-
ergy values, HOMO and LUMO orbital coefficients
distribution, molecular dipole moment (l), and molec-
ular electrostatic potential (MEP), were calculated.
MEP isoenergy surface maps were generated in the
range from À25.0 (deepest red color) to +30.0 (deep-
est blue color) kcal/mol and superimposed onto a
molecular surface of constant electron density of
0.002 e/au3. Each point of the three-dimensional
molecular surface map expresses the electrostatic inter-
action energy value evaluated with a probe atom of
positive unitary charge providing an indication of
the overall molecular size and location of attractive
(negative) or repulsive (positive) electrostatic
potentials. The most potent acridone and the new
derivatives (2a–i) also were submitted to an in silico
ADMET screening available in http://www.acteliion.
com/ or http://www.organic-chemistry.org/ to deter-
mine their overall drug-score.
N
N
N
Ph
R2
N
Cl
(
)
2c-i
Scheme 2. Synthetic approach used to obtain the 3H-benzo[b]pyraz-
olo[3,4-h]-1,6-naphthyridine derivatives.
in solvents such as N,N-dimethylformamide.22 Subse-
quently, 4a–i were hydrolyzed to the corresponding
acids (3a–i) (88–98%) and then cyclized to the corre-
sponding tetracyclic heteroaromatic system (2a–i, 60–
73%) (Scheme 2). Both ‘aniline ester’ (4a–i) and ‘aniline
acids’ (3a–i) intermediates were characterized through
1
infrared (IR) and H nuclear magnetic resonance (1H
NMR) spectroscopic methods.
3. Results and discussion
2.2. Chemistry
3.1. Molecular modeling and SAR studies of 1-hydroxy-
acridone derivatives
A known synthetic approach16–18 was used for prepar-
ing the new nine 3H-benzo[b]pyrazolo[3,4-b]-1,6-naph-
thyridine derivatives (2a–i), starting from two ethyl
4-chloro-1H-pyrazolo[3,4-b]pyridine-5-carboxylates (5a–
b) as precursors (Scheme 2). Briefly, 5a–b were prepared
from 5-aminopyrazoles, through condensation with
diethyl ethoxymethylenemalonate and cyclization, fol-
lowed by reaction with anilines, hydrolys, and a key step
of ‘chlorocyclization’ using phosphorus oxychloride.18–21
5a–b act as good precursors for the synthesis of the present
ring system due to the presence of the readily reactive
4-chloro and 5-carboxylate groups.
The minimum energy conformations of 1a–j derivatives,
calculated by the AM1 semiempirical Hamiltonian,
showed that, as expected, all rings of these compounds
are co-planar. Subsequently, a single-point energy ab ini-
tio calculation was performed at the 6-311G* level in or-
der to derive electronic properties, such as HOMO and
LUMO energy values, HOMO and LUMO orbital coef-
ficients distribution, molecular dipole moment (l), and
molecular electrostatic potential (MEP), which could
be related to the variation of the anti-HSV-1 activity
of these compounds. The SAR results showed that,
although the substitution pattern into the 1-hydroxyac-
ridone structures lead to different antiviral profile,
HOMO and LUMO energy values and molecular dipole
moment values did not present any direct correlation
with it, as shown in Table 1.
5a–b on fusion with several anilines led to the required
ethyl 4-anilino-1H-pyrazolo[3,4-b]pyridine-5-carboxyl-
ates (4a–i) in good yields (56–80%). However, better re-
sults were obtained when hese reactions were carried out

316
A. M. R. Bernardino et al. / Bioorg. Med. Chem. 16 (2008) 313–321
As receptors recognize stereo-electronic effects and not
atom per se, studies of molecular electronic properties
could be very effective in interpreting the electronic
structure in a comprehensive way.21,22 Therefore, the
MEP is a useful approach for understanding the electro-
static contribution for the receptor-ligand binding pro-
cess.21 The analysis of R3 position (Table 1) on MEPs
generated for these compounds (Fig. 3A) revealed a neg-
ative potential (red color, i.e., negative charge) at this re-
gion when hydroxyl (1a–1g) or methoxyl (1h) are the
substituents. However, differently of the methoxyl group
(1h), the hydroxyl (1a–1g) may act as both donor and
acceptor of hydrogen bond. In addition, methyl substi-
tuent (1i) or an extra aromatic ring (1j) has shown a re-
duced or absence of the negative potential in this region.
This result suggested that the 3-hydroxyl group or other
electron rich substituent is an important feature for
reaching an anti-HVS-1 leading compound. Interest-
ingly, presence of the 3-hydroxyl group but absence of
substituents in other positions (R@H, 1g) dramatically
decreased the antiviral activity. Thus, this result points
to other functional groups at C5, C6, C7, and 8 posi-
tions (R substituent, Table 1) as important for conserv-
ing a significant biological profile. Chlorine substituent
at C5 (1b), C6 (1d), C7 (1f), and C8 (1e) position is
somehow important for the biological profile (Table
1). However, the presence of methoxyl group (1a) in-
creased the electronic density of the MEP represented
by intense red color in the ring system (Fig. 3A). This
feature could be involved in the 2- to 8-fold increase
of 1a derivative antiviral activity when compared to 1b
and 1c (Fig. 3A, Table 1).
acridone ring are maintained in derivatives 2a–b as
N11-H and C6-carbonyl groups of naphthyridinone
ring, respectively. Meanwhile for 2c–i derivatives, we
changed the hydrogen-bonding donor (N10-H of acri-
done) to an acceptor (N11 of naphthyridine) character
at the former position, and from C9-carbonyl group of
acridone ring to C6-chlorine groups of naphthyridine
ring. The C1-hydroxyl group of acridone ring has no
counterpart in the naphthyridinone (2a–b) or naphthyri-
dine (2c–i) rings, although C1-hydroxyl is close to N4 of
naphthyridine ring, allowing the same hydrogen bond-
ing acceptor capability (Scheme 1). However, according
to Lowden and Bastow, removal and/or replacement of
the C9-carbonyl and/or the C1-hydroxyl groups is a
desirable approach, since these motifs may be responsi-
ble for cell toxicity.9
Therefore, we designed 2a–i derivatives with R1 and R2
substituents (Scheme 1 and Table 2) to identify the role
of some structural features, such as the electronic prop-
erties, to the antiviral profile. The R1 position of 2a–i
derivatives has no counterpart in the acridone system,
so hydrogen or methyl group could be used as substitu-
ents. On the other hand, the R2 position corresponds to
the C5-, C6-, C7- or C8-substituent of acridone ring, and
it could be hydrogen, methyl, methoxyl, or chlorine at
C10 or C8 positions of the 2a–i derivatives.
3.2.2. Chemistry. The nine derivatives (2a–2i) were syn-
thesized in good yields (60–73%) using a brief scalable
and easy procedure (Scheme 2).22 The substitution pat-
tern is similar to that of N-phenylpyrazole and 1H-pyr-
azolo[3,4-b]pyridine and indazole. Ring system, where
the free position of the pyrazole is the condensed ring,
is reactive toward electrophilic substitution reaction.
The products 2a and 2b have been represented here in
the ‘oxo’ rather than the ‘hydroxy’ tautomer in accor-
dance to that reported for such related system.17 The
infrared spectra of the compounds obtained after cycli-
zation reaction showed a broad absorption band in the
region assigned to NH frequency of 3400 cmÀ1 for
2a–b and 3430 cmÀ1 for 2c–i derivatives due to the
N11 protonation. The carbonyl absorption was ob-
served between 1620 and 1720 cmÀ1, which also indi-
cated the predominance of ‘oxo’ tautomer (2a–b).
Therefore, the C9-carbonyl group of the acridone sys-
tem corresponds to the C6-carbonyl group of the 2a–b
compounds and to the C6-chlorine of the 2c–i
derivatives.
Importantly, the positive potential (blue color, i.e., posi-
tive charge) observed on MEP maps corresponding to
the C4 and N10 hydrogen atom positions decreased
for 1h–j derivatives. Therefore, the lower or loss of the
antiviral activity for 1h–1j derivatives could be probably
due to a clear reduction of the positive potential of these
hydrogen atoms (Fig. 3A). Then, based on these SAR
results and the structural features observed, we designed
the nine 3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyridine
derivatives (2a–i).
3.2. New series: 3H-benzo[b]pyrazolo[3,4-h]-1,6-
naphthyridines
3.2.1. Design. 2a–i derivatives were designed combining
isosteric replacement and rigid analogue approaches
based on two heterocycle systems, 1H-pyrazolo[3,4-
b]pyridine-5-carboxylic acid presenting anti-HSV-1
activity among a diverse antiviral profile13,14,21 and syn-
thetic 1-hydroxyacridone derivatives with anti-HSV-1
activity. Scheme 1 shows that the acridone motif is a
classical isostere of the 3H-benzo[b]pyrazolo[3,4-h]-1,6-
naphthyridine aromatic system, while the ring closure
of the 1H-pyrazolo[3,4-b]pyridine-5-carboxylic acid sys-
tem leads to the 3H-benzo[b]pyrazolo[3,4-h]-1,6-naph-
thyridine motif corresponding to the rigid analogue
approach.
3.2.3. Antiviral and cytotoxicity assays. In order to eval-
uate the biological profile of 2a–2i derivatives, we tested
them as potential inhibitors of HSV-1 replication in cell
culture as described on Experimental section (Table 2).
Among all compounds evaluated, derivative (2f) with
chlorine atom at 10-position (R2@10-Cl, Table 2), struc-
turally analogous to acridone 1b, exhibited a high anti-
HSV-1 activity (100%). In contrast, the derivative 2i
(Fig. 3B) with methyl group at 8-position did not show
any inhibitory activity. Neither substituents in R1 (H or
CH3) nor in C6- (chlorine or carbonyl) positions com-
pletely abolished the antiviral activity, which reveals
R2 position as the most important determinant for these
As a result, considering the unchanged scaffold of acri-
done system, the N10-H and C9-carbonyl groups of

A. M. R. Bernardino et al. / Bioorg. Med. Chem. 16 (2008) 313–321
317
compounds’ antiviral profile (Table 2). Apparently, po-
sition in the ring (10 or 8) is one of the most important
structural features to be considered in R2 as 2f and 2h
(10-substituents) presented higher activity than its anal-
ogous 2g and 2i (8-substituents) (Table 2).
might be important for the activity observed due to its
hydrophobic character rather than its ability as a hydro-
gen-bonding acceptor. Moreover, since Lowden and
Bastow have proposed the removal and/or replacement
of the C9-carbonyl and/or the C1-hydroxyl groups of
acridone system in order to decrease cell toxicity,9 we
designed a new series that does not contain these motifs
simultaneously.
To verify the cytotoxicity profile of this naphthyridine
series, we also performed cytotoxicity assays using Vero
cells. Our results showed that active and inactive deriv-
atives present
a
cytotoxic concentration higher
The comparison of the theoretical results of (2f) and (2g)
derivatives revealed that electronic density significantly
decreases in the last compound when chlorine atom is
in 8-position (Table 2, Fig. 3A). Probably loss of this
negative region may be deleterious for 2g binding to
receptor/target of the virus, in contrast to 2f (10-substi-
tuted derivative) as observed in the biological assays.
(CC50 = 640–4258 lM) than acyclovir (CC50 = 126 lM)
(Fig. 2). Thus these data indicated the low cytotoxicity
profile of this naphthyridine series whereas reinforced
their antiviral activity.
3.2.4. Molecular modeling and SAR studies of 2a–i
derivatives. The same molecular modeling procedure
used for acridones was applied for the nine 3H-
benzo[b]pyrazolo[3,4-h]-1,6-naphthyridine derivatives.
The minimum energy conformation for all compounds
shows that the N3-phenyl ring is co-planar to the pyraz-
olo-naphthyridine system probably due to the electro-
static attraction between the ortho-hydrogen atoms of
N3-phenyl ring and the N2 and N4 nitrogen atoms of
the pyrazolo-naphthyridine system (Scheme 1, Table 2,
and Fig. 3B). The overall analysis of naphthyridines’
calculated electronic properties, such as HOMO and
LUMO energies and molecular dipole moment values,
showed no clear or direct correlation with anti-HSV-1
activity, similar to that observed for acridones (Table 2).
The comparison of the MEP energy isosurface maps of
the acridone (1a–g) and naphthyridine (2a–i) derivatives
(Fig. 3A) showed the highest negative electrostatic
potentials close to R3@OH substituent in acridone ana-
logues and to the N2 in naphthyridine derivatives. In
fact, the three-dimensional MEP maps in both com-
pound series reveal that the center of the most negative
region (darkest red) lies near to these substituents. The
structural alignment of the most active (2f) and the inac-
tive (2i) compounds with acridone 1b infers a restrictive
role for the positions analyzed (Fig. 3B).
Despite the interesting results, it is important to consider
the limitation of ligand-based SAR design when using
cell-based assay data, especially when the target/mecha-
nism of action are unknown. In case of acridone series
(1), it is known that the mechanism of two regioisomers
(1f and 1h) is different, which suggests they act on differ-
ent targets.22 This may partially explain the absence of
direct correlation between some calculated electronic
parameters and molecular dipole moments, and the bio-
logical activity (Table 1). Thus the fact that the target/
mechanism of the naphthyridine compounds studied
herein is also not fully understood limits the interpreta-
tion of the computational approach used and this
should be considered.
Analysis of the MEP maps of 2a–i derivatives (Fig. 3A)
showed significant changes in C6-position of naphthyri-
dine moiety in 2a and 2b compounds (carbonyl group)
in contrast with compounds 2c–i (chlorine substituent).
However, since there are compounds with almost equiv-
alent activity with carbonyl (2b, 100%) or chlorine (2f,
92%) at C6-position (Table 2), the electrostatic potential
is not the determinant factor for the activity at this posi-
tion, which is related to the C9-carbonyl group in acri-
done derivatives. This result suggested that this region
5000
4000
3000
2000
3.2.5. In silico ADMET screening. Currently, there are
many approaches to assess drug-likeness and/or lead-
likeness of drug candidates.23,24 One of such approaches
is the Actelion Property Explorer (http://www.acteli-
ion.com/ or http://www.organic-chemistry.org/), a web-
based system named Osiris, which is able to calculate
indices and/or properties such as toxicity risk assess-
ment, logP prediction, solubility (logs) prediction,
molecular weight, fragment-based drug-likeness predic-
tion, and overall drug-score. The overall drug-score
combines drug-likeness, cLogP, clogs, molecular weight,
and toxicity risk indices in one single value, where the
occurrence frequency of each fragment is determined
within the collection of traded drugs and within the sup-
posedly non-drug like collection of Fluka compounds.
1000
0
2a
2b
2c
2d
2g
2h
Acyclov
In this work, we used the Osiris Property Explorer for
calculating the overall drug-score of the most active
acridone (1a and 1b) and 2b, 2c, and 2f compounds
Figure 2. Comparison of the cytotoxicity concentration (CC50) of
active and inactive naphthyridine compounds and acyclovir (Acyclov).

318
A. M. R. Bernardino et al. / Bioorg. Med. Chem. 16 (2008) 313–321
Figure 3. Theoretical studies of acridones and naphthyridines. (A) Comparison of the Molecular Electrostatic Potential (MEP) energy isosurfaces of
1-hydroxyacridones (up) and 3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyridines (down) superimposed onto total electron density of 0.002 e/a.u.3. The
color code is in the range of À25 (deepest red) to +30 (deepest blue) kcal/mol. (B) Structural alignment of 5-chloro-1,3-dihydroxyacridone (carbon
atoms are in orange color) (1b) with the most active (2f) and the inactive (2i) 3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyridine derivatives (carbon atoms
are in gray color). (C) Calculated side effects (i.e., reproductive, irritant, and tumorigenic theoretical toxicity risks) and (D) calculated DrugScore
values for the most potent compounds (1a, 2b–c, and 2f) and acyclovir (ACV) using Osiris program (http://www.organic-chemistry.org/prog/peo/
druglikeness.html).
using acyclovir (ACV) as positive control as described
elsewhere.23,24 Potential toxicity risks including repro-
ductive, irritant, and tumorigenic effects are shown in
Figure 3C while the overall drug-score is shown in
Figure 3D. Interestingly, 2b, 2c, and 2f derivatives
presented a drug-score similar to acyclovir, while the
acridone compounds presented a greater drug-score
(Fig. 3D). Importantly, only 2b presented the worst pre-
diction of one of these toxicity risks (Fig. 3C). The pre-
dicted irritant profile of 2b is a direct indication of its
non-potential drug-score. The low theoretical reproduc-
tive, irritant, and tumorigenic effects of the other com-
pounds (1a, 1b, 2c, and 2f) indicated a low risk drug
like profile; similar to acyclovir that in the analyzed as-
pects was also safe. It is important to notice that the tox-
icity predicted herein neither is a fully reliable toxicity
prediction, nor guarantee that these compounds are
completely free of any toxic effect. However it reinforced
the promising profile of these compounds for further
experimental investigation.

A. M. R. Bernardino et al. / Bioorg. Med. Chem. 16 (2008) 313–321
319
4. Conclusions
5.1.3. 4-Anilino-1-phenyl-1H-pyrazolo[3,4-b]-pyridine-5-
carboxylic acids (3a–i). A mixture of 3 mmoles of the es-
ter (4a–i), 10 ml of 20% sodium hydroxide solution, and
10 ml of ethanol was heated under reflux for 1–3 h. On
cooling mixture was acidified with diluted hydrochloric
acid (1:3), and the precipitated acids were filtered and
recrystallized from a mixture of DMF and water.
Overall in both acridones and 2a–i compounds, the po-
sition of functional groups has impact on the activity as
observed in the SAR study. The structure activity-rela-
tionship of acridone derivatives shows that the 5-chloro
position is interesting for displaying an anti-HSV-1 ro-
file. Similar feature was observed in 2a–i compounds,
in which the highest inhibitory activity is noticed when
the chloride atom is in the 10-position next to nitrogen
atom of central ring.
5.1.4. 3-Phenyl-3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyri-
dines (2a–i). A mixture of the ‘aniline acids’ (3a–i) and
phosphorus oxychloride (5 mL) was heated under reflux
for 3 h. The reaction mixture was inverted over crushed
ice. In some cases the excess of phosphorus oxychloride
was removed under reduced pressure before inverting
over crushed ice and neutralized with water. The result-
ing precipitate was collected and purified by flash col-
umn chromatography (FC, silica gel).
Our results with acridones and 2a–i compounds pointed
their electrostatic features and other electronic proper-
ties as important in the antiviral profile. Although the
Osiris risk alerts are not a fully reliable toxicity predic-
tion, the theoretical low-toxicity profile of these com-
pounds reinforces the significant activity of 2c and 2f.
This theoretical result is reinforced by the experimental
cytotoxicity assay, which confirmed these compounds’
low toxicity profile. Therefore, the analyses of other
molecular properties could be useful for designing new
anti-HSV-1 drugs using some of these derivatives such
as 2f as a leading compound.
(2a) 3-Phenyl-3H-benzo[b]pyrazolo[3,4-h]-1,6-naphthyri-
din- 6(11H)one. Yield: 70%; mp 258 ꢁC; IR (KBr, cmÀ1):
(mNH3400, mC@O 1650–1640); 1H NMR: 6.01(s); 8.94(s);
7.84–7.47(m).
(2b) 3-Phenyl-8-methyl-3H-benzo[b]pyrazolo[3,4-h]-1,6-
naphthyridin-6(11H)one. Yield: 69%; mp >300 ꢁC; IR
1
(KBr, cmÀ1): (mNH3400, mC@O 1650–1640); H NMR:
6.22(s); 9.08(s); 7.51–7.28(m).
5. Experimental
5.1. Chemistry
(2c) 6-Chloro-3-phenyl-8-methyl-3H-benzo[b]pyrazolo-
[3,4-h]-1,6-naphthyridine. Yield: 68%; mp 231 ꢁC; IR
1
All reagents and solvents used were of analytical grade.
TLC was carried out using silica gel F-254 Glass Plate
(20 · 20 cm). The solid samples were measured using
potassium bromide pellets. Melting points (mp) were
(KBr, cmÀ1): (mNH3430, mC@C 1593, mC@N 1502); H
NMR: 6.22(s); 9.19(s); 7.77(s); 7.45(dd. 7.8; 7.8); 8.11(d.
7.5); 8.33(d. 7.8); 7.28(d. 7.2); 7.02(d, 7.5); 2.51(s).
1
determined with a Fisher–Johns apparatus. The H Nu-
(2d) 6-Chloro-3-phenyl-8-methyl-3H-benzo[b]pyrazol-
o[3,4-h]-1,6-naphthyridine. Yield: 70%; mp 245 ꢁC; IR
clear agnetic Resonance (1H NMR) spectra were ob-
tained from Varian model Unity Plus spectrometer
operating at 300.00 MHz, using tetramethylsilane as
internal standard. The chemical shifts (d) are reported
in ppm and the coupling constants (J) in Hertz. Fourier
transform infrared (FT-IR) absorption spectra were re-
corded in a Perkin-Elmer model Spectrum One FT-IR
spectrophotometer.
1
(KBr, cmÀ1): (mNH3430, mC@C 1593, mC@N 1502); H
NMR: 6.70(s); 9.62(s); 7.59–7.50(m); 8.12(d, 9.3);
8.22(d, 7.8); 7.37(d, 7.5); 4.01(s).
(2e) 6-Chloro-3-phenyl-1-methyl-3H-benzo[b]pyrazolo-
[3,4-h]-1,6-naphthyridine. Yield: 60% mp 195–96 ꢁC; IR
1
(KBr, cmÀ1): (mNH3430, mC@C 1596, mC@N 1502); H
NMR: 9.58(s); 8.20(d, 9.0); 7.88–7.52(m); 8.17(d, 9.0).
3H-Benzo[b]pyrazolo[3,4-h]-1,6-naphthyridines and the
intermediate ‘aniline esters’ and ‘aniline acids’ were pre-
pared as follows.
(2f) 6,10-Dichloro-3-phenyl-1-methyl-3H-benzo[b]pyr-
azolo[3,4-h]-1,6-naphthyridine. Yield: 71%; mp 224 ꢁC;
IR (KBr, cmÀ1): (mNH3430, mC@C 1596, mC@N 1502);
1H NMR: 9.06(s); 7.81–7.48(m); 8.35(d, 8.1); 7.73(d,
8.1); 7.53(d, 7.2); 1.39(s).
5.1.1. Ethyl 4-chloro-1-phenyl-1H-pyrazolo[3,4-b]-pyri-
dine-5-carboxylate (5a) and Ethyl 4-chloro-3-methyl-1-
phenyl-1H-pyrazolo[3,4-b]pyridine-5-carboxylate
These were prepared according to literature method:
(5b).
(2g) 6,8-Dichloro-3-phenyl-1-methyl-3H-benzo[b]pyraz-
olo[3,4-h]-1,6-naphthyridine. Yield: 68%; mp 243 ꢁC;
IR (KBr, cmÀ1): (mNH3430, mC@C 1596, mC@N 1502);
1H NMR:9.43(s); 7.93(s); 7.85(d, 7.8); 8.05(d, 7.8);
8.36(d, 7.5); 7.74(d, 7.8); 7.52(d, 7.8); 1.38(s).
(5a), mp 98 ꢁC (mp 97 ꢁC); (5b), mp 110 ꢁC (mp
16
110 ꢁC).
5.1.2. Ethyl 4-anilino-1-phenyl-1H-pyrazolo[3,4-b]-pyri-
dine-5-carboxylate (4a–i). Method A: An equimolar mix-
ture of 5a or 5b and an aniline (5 mmoles) in 10 ml N,N-
dimethylformamide was heated under reflux for 4 h. The
reaction mixture, after cooling, was poured into 50 ml of
ice cold water. The precipitated ‘aniline esters’ were fil-
tered, dried, and recrystallized from a mixture of ethanol
and water.
(2h) 6-Chloro-3-phenyl-1,10-dimethyl-3H-benzo[b]pyr-
azolo[3,4-h]-1,6-naphthyridine. Yield: 67%; mp >300 ꢁC;
IR (KBr, cmÀ1): (mNH3430, mC@C 1596, mC@N 1502);
1H NMR: 9.18(s); 7.76(s); 8.11(d, 8.0); 7.45(dd, 8.0);
8.33(d, 7.8); 7.52(dd, 7.5; 7.5); 7.28(dd, 7.5; 7.5); 2.51(s);
1.37(s).

320
A. M. R. Bernardino et al. / Bioorg. Med. Chem. 16 (2008) 313–321
(2i) 6-Chloro-3-phenyl-1,8-dimethyl-3H-benzo[b]pyraz-
olo[3,4-h]-1,6-naphthyridine. Yield: 73%; mp 201 ꢁC; IR
Acknowledgments
1
(KBr, cmÀ1): (mNH3430, mC@C 1592, mC@N 1503); H
This study was supported by the following Brazilian
agencies and governmental institutions: Fundac¸a˜o Uni-
NMR: 9.54(s); 8.05(d, 4.5); 7.63(dd, 9.0); 8.03(d, 9.0);
8.20(d, 7.5); 7.55(d, 7.8); 7.36(d, 7.5); 3.07(s); 1.30(s).
´
versitaria Jose Bonifacio/Universidade Federal do Rio
de Janeiro (FUJB/UFRJ), Coordenac¸a˜o de Aperfeic¸oa-
´
´
´
mento de Pessoal de Nıvel Superior (CAPES), Conselho
5.2. Antiviral assay
´
Nacional de Desenvolvimento Cientıfico e Tecnologico
(CNPq), Universidade Federal Fluminense (UFF), and
´
African green monkey kidney cells (Vero) were cul-
tured in Dulbecco’s modified Eagle’s medium
(DMEM; GIBCO/BRL) supplemented with 5% fetal
calf serum (FCS), 100 U/mL penicillin, 100 lg/mL
streptomycin, and 250 lg/mL amphotericin-b. The cells
were incubated at 37 ꢁC in a 5% CO2 atmosphere. As
the cells became confluent, they were washed with
PBS-EDTA, dissociated with trypsin 0.25%, and then
resuspended in culture medium. Monolayers in
25 cm2 bottles were infected by acyclovir-resistant
strain of HSV-1 (AR-29) at 0.1 multiplicity of infection
(m.o.i.). Next, cells were lysed 24 h after infection by
three cycles of freezing and thawing, centrifuged at
3000 rpm at 4 ꢁC for 20 min, and the supernatant
was stored at À70 ꢁC for use.
`
Fundac¸a˜o de Amparo a Pesquisa do Estado do Rio de
Janeiro (FAPERJ). Loureiro NI and Rodrigues CR
have CNPq fellowships.
References and notes
1. Khan, M. T. H.; Ather, A.; Thompson, K. D.; Gambari,
R. Antiviral Res. 2005, 67, 107–119.
2. Efstathiou, S.; Preston, CM. Virus Res. 2005, 111,
108–119.
3. D. A. Baker, Management of herpes in pregnancy ACOG:
Washington, DC, 1999.
4. De Clercq, E.; Naessens, L.; De Bolle, L.; Schols, D.;
Zhang, Y.; Neyts, J. Rev. Med. Virol. 2001, 11,
381–395.
The antiviral activities of the 3H-benzo[b]pyrazolo-
[3,4-h]-1,6-naphthyridine derivatives (50 lM) were
investigated according to Reed and Munch 50% end
point method.25 Acyclovir was used as positive con-
trol at 50 lM (HSV = 100%). Briefly, Vero cells grown
to confluence in 96-well plate were infected with HSV-
1 at various m.o.i. (1, 0.1, 0.01, 0.001) for 1 h at 37 ꢁC.
After that, infected cells were washed with PBS to re-
move residual viruses and complete culture medium,
containing or not the derivatives, was added. At
72 h after infection virus titer of each sample was
determined in terms of the 50% tissue culture dose
(TCID 50/mL) by end point dilution. In addition we
also used Mock-treated HSV1-infected cells as posi-
tive control. All experiments were performed in dupli-
cate at least three times. Thus we presented the
compounds results by using the formula y = (100.T/
M) À 100 where y is the inhibition (%), T is the titer
of the HSV-1-infected cells treated with the naph-
thyridine compound, and M is the titer of the
Mock-treated HSV1-infected cells.
5. Hayden, F. G. In The Pharmacological Basis of Thera-
peutics; Hardman, J. G., Limbird, L. E., Eds.; McGraw
Hill Publisher: NY, 2001; pp 1295–1312.
6. Morfin, F.; Thouvenot, D. J. Clin. Virol. 2003, 26,
29–37.
7. Chilukuri, S.; Rosen, T. Dermatol. Clin. 2003, 21,
11–20.
8. Yamamoto, N.; Furukawa, H.; Ito, Y.; Yoshida, S.;
Maeno, K.; Nishiyama, Y. Antiviral Res. 1989, 12,
21–36.
9. Lowden, C. T.; Bastow, K. F. Antiviral Res. 2003, 59,
143–154.
10. Lowden, C. T.; Bastow, K. F. J. Med. Chem. 2003, 46,
5015–5020.
11. Akanitapichat, P.; Bastow, K. F. Antiviral Res. 2002, 53,
113–126.
12. Akanitapichat, P.; Lowden, C. T.; Bastow, K. F. Antiviral
Res. 2000, 45, 123–134.
13. Mello, H.; Echevarria, A.; Bernardino, A. M.; Canto-
Cavalheiro, M.; Leon, L. L. J. Med. Chem. 2004, 47, 5427–
5432.
14. Azevedo, A.; Ferreira, V. F.; Mello, H.; Lea˜o-Ferreira, L.
R.; Jabor, A. V.; Frugulhetti, I. C. P. P.; Pereira, H. S.;
´
Moussatche, N.; Bernardino, A. M. Heterocycl. Commun.
2002, 8, 427–432.
5.3. Cytotoxicity assay
15. Bare, T. M.; MacLaren, C. D.; Campbell, J. W.; Firor,
J. W.; Resch, J. F.; Walters, C. P.; Salama, A. I.;
Meiners, B. A.; Patel, J. B. J. Med. Chem. 1989, 32,
2561–2573.
16. Azevedo, A. R.; Frugulhetti, I. C. C. P.; Khan, M. A.;
Khakwani, S.; Bernardino, A. M. R. Heterocyclic Com-
munications 2002, 8, 47–54.
17. Khan, M. A.; Rolim, A. M. C. Monatsh. Chem. 1983, 114,
1079.
18. Lynch, B. M.; Khan, M. A.; C Teo, H.; Pedrotti, F. Can.
J. Chem. 1988, 66, 420–426.
19. Kuo, Y. C.; Chen, C. C.; Tsai, W. J.; Ho, Y. H. Antiviral
Res. 2001, 51, 95–109.
20. Bhattacharjee, A. K.; Majundar, D.; Guha, S. J. J. Chem.
Soc. Perkin Trans. 1992, 2, 805.
21. Politzer, P.; Laurence, P. R.; Jayasuriya, K. Environ
Health Perspect. 1985, 61, 191–202.
Vero cells were cultured in Dulbecco’s modified Ea-
gle’s medium (DMEM) supplemented with 5% fetal
bovine serum (FBS; HyClone, Logan, Utah), 100 U/
mL penicillin, and 100 lg/mL streptomycin, at 37 ꢁC
in 5% CO2. Monolayers of about 104 Vero cells in
96-multiwell plates were treated with several concen-
trations of the compounds for 72 h. Then, 50 ll of a
1 mg/mL solution of 3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyl tetrazolium bromide (MTT; Sigma) was
added to evaluate cell viability according to proce-
dures described elsewhere.26 The 50% cytotoxic con-
centration (CC50) was calculated by linear
regression analysis of the dose-response curves. All
experiments were performed in duplicate at least three
times.

A. M. R. Bernardino et al. / Bioorg. Med. Chem. 16 (2008) 313–321
321
22. Bastow, K. F. Curr. Drug Targets: Infect. Disorders 2004,
4323–4330.
24. Costa, M. S.; Boechat, N.; Rangel, E. A.; da Silva, F. D.;
de Souza, A. M.; Rodrigues, C. R.; Castro, H. C.; Junior,
I. N.; Lourenco, M. C.; Wardell, S. M.; Ferreira, V. F.
Bioorg. Med. Chem. 2006, 8644–8653.
25. Reed, B.; Munch, M. Am. J. Virol. 1938, 27, 492–
504.
26. Mosman, T. J. Immunol. Methods 1983, 65, 55–63.
23. Bernardino, A. M. R.; Pinheiro, L. C. S.; Rodrigues, C. R.;
Loureiro, N. I. V.; Castro, H. C.; Lanfredi-Rangel, A.;
Sabatini-Lopes, J.; Borges, J. C.; Carvalho, J. M.; Romeiro,
G. A.; Ferreira, V. F.; Frugulhetti, I. C. P. P.; Vannier-
Santos, M. A. Bioorg. Med. Chem. 2006, 14, 5765–5770.