A. Darque et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5962–5964
5963
Table 2
Ratio of various quinolinones
X
O
O
O
O
Y
X
Z
Y
Z
+
N
O
O
O
1
Linear ratio
Bent ratio
50
Compounds 2
Yield (%)
H
H2N
O
a
b
c
d
e
f
One possibility
50
One possibility
50 (2d)
2a
53
79
88
95
52
76
90
35
88
76
82
35
1 a-o
No separation
2c
2d (10%)
2e
2f
2g
2h
2i
Scheme 1. Synthesis of Meldrum’s acid derivatives 1. Reagents and conditions:
trimethylorthoformate 5 mL, MWI, 100 °C, 2 min.
50
100
100
100
g
h
i
j
k
l
m
n
o
nitroaniline as starting material.13 We optimized the protocol on a
milligram scale, and performed a successful MWI-scale-up to gram
quantities with comparable yields. We have applied this method to
various heterocycles (Table 1) successfully (compounds 1a–o) with
good yields (39–92%). None of our compounds has been previously
described using this method of synthesis.
The second step was a thermal cyclization usually done in di-
phenyl ether at 240 °C for 20 min. Using MWI requires a solvent
with a high dielectric constant to take advantage of MW heating.
Diphenyl ether can rise without degradation to 200 °C in 30 min
at full power MWI (300 W). Ionic liquids enhance MW heating,14
thus we added bmimPF6 (50 mg) into diphenyl ether (2g). It al-
lowed reaching 220 °C in 3 min with full power MWI (300 W)
(Scheme 2). The cyclization of our compounds occurred by precip-
itating the product in a mixture of diethyl ether/acetonitrile and
removing ionic liquid and diphenyl ether.
100
100
One possibility
50
50
2j
50
50
No separation
No separation
No cyclization
No cyclization
No cyclization
One possibility
tons H4 and H9 of the structure. In our case, MWI was a clean, sim-
ple and faster protocol compared to the conventional heating
method. We have obtained nine new quinolinone (2a–j) within
5 min. of reaction time.
Human therapies already use benzothiazole derivatives
although they have a potential toxicity due to the presence of a
benzyl moiety in their structure. Our goal was to synthesize similar
products with less toxicity, by suppressing an aromatic nucleus
while maintaining the antiparasitic activity. This synthesis was al-
ready effective with quinolones.15
Products obtained were bent or linear tricyclic quinolinones,
depending on the starting heterocycle. Structures and isomeric ra-
tio of bent and linear derivatives (Table 2) were determined with-
out ambiguity by 1H and 13C NMR spectroscopy. There was an AB
system for protons H4 and H5 (JAB = 9 Hz) for the bent structure.
For the linear one, two singlets appeared, that corresponded to pro-
Most of our products (Table 3) were not cytotoxic (IC50 THP1
>250 lM), and only five had low to moderate cytotoxicity. Com-
pounds 1a and 1d were even less toxic than their respective cy-
clized derivatives, compounds 2a and 2d.
Although some had a structural analogy with chloroquine, 16
compounds had no antimalarial activity while the others had a
moderate activity. Cyclization did not increase activity seeing that
intermediate compounds had antimalarial activities similar to
those of their respective cyclized derivatives. Cyclized compounds
are usually more active than their corresponding intermediate
structures. It was not the case with these series, moreover, inter-
mediate compounds were less toxic than their cyclized derivatives.
Therefore, it is relevant to measure activity of both intermediate and
cyclized compounds, rather than that of only the cyclized compounds.
None of the products had a significant antileishmanial activity
on either promastigote or amastigote forms, from low to moderate
concentrations. None had any antitrichomonal activity either.
Compounds 1m and 1k were active on HIV-1, while compounds
1c, 2j, and 3a were active on HIV-2. Globally, non-cyclized com-
pounds seemed more active than their respective cyclized deriva-
tives on HIV. Within our products, 1k was the only one
combining activity against both malaria and HIV, having also a
low cytotoxicity and the highest specificity index (see ‘Supplemen-
tary data’ for detailed results on HIV strains).16,17 Although higher
than that of chloroquine, IC50 of compound 1k was close to that of
doxycycline, another antimalarial drug of reference, both experi-
Table 1
MWI Meldrum acids preparation
Name
X
Y
Z
Yield (%)
92
1
p-Nitroaniline
m-Nitroaniline
o-Nitroaniline
Benzodioxane
Benzothiazole
Benzotriazole
Indazole
Benzoxazole
Benzothiazole
2-(phenyl-2-amino)benzothiazole
Benzodioxol
Indane
p-Nitroaniline
2-(Phenyl-3-amino)benzothiazole
Benzimidazolone
NO2
a
b
c
d
e
f
g
h
i
j
k
l
m
n
o
NO2 66
90
O
S
O
N
N–H
NH
N
CH2–CH2
C–CH3
N
71
83
79
N
N
C–H 83
O
S
C–N–Me2
C–H
39
83
81
39
41
96
67
N
O
CH2
NO2
CH2
CH2
O
CH2
N–H C@O
N–H 67
Yield (%): Average of three different experiments with 1, 10 and 20 mmol of amino
compounds.
Y
X
O
Z
mentally (6.5 lM) (Table 3) and as reported (8.5 l
M).19 Moreover,
compound 1k was less toxic than doxycycline and had a better
specificity index on THP1 (Table 3).18,19 We are studying structures
differing by a single moiety to explain the differences of activity
between a compound and its cyclized derivative. A concomitant
goal is to improve the structure of our most interesting drug can-
didate. All of our compounds are new chemical structures deriving
from acridine and heterocyclic structures. Suppressing an aromatic
nucleus from the acridine nucleus only reduced toxicity. Consider-
ing that this third aromatic nucleus is required for biological activ-
ity, we are trying to synthesize new products with a third benzyl
nucleus that would not be adjacent to the quinolinone structure.
bent
4
N
H
5
X
Z
O
O
Y
and/or
N
O
H
O
O
4
X
linear
Y
Z
N
H
9
Scheme 2. Thermal cyclization with MWI in full power irradiation.