Synthesis, anti-tubercular activity and 3D-QSAR study of coumarin-4-acetic acid benzylidene hydrazides

Article abstract of DOI:10.1016/j.ejmech.2008.01.016

A set of 25 coumarin-4-acetic acid benzylidene hydrazides were synthesized and characterized by NMR, IR and mass spectroscopic techniques. The compounds were evaluated for their anti-tubercular activity against Mycobacterium tuberculosis H₃₇Rv strain using the BACTEC 460 system to determine percentage inhibition. To understand the relationship between structure and activity, a 3D-QSAR analysis has been carried out by Comparative Molecular Field Analysis (CoMFA). Several statistically significant CoMFA models were generated. The CoMFA model generated with database alignment was the best in terms of overall statistics. The CoMFA contours provide a good insight into the structure activity relationships of the compounds reported herein.

Full text of DOI:10.1016/j.ejmech.2008.01.016

Available online at www.sciencedirect.com
European Journal of Medicinal Chemistry 43 (2008) 2395e2403
http://www.elsevier.com/locate/ejmech
Original article
Synthesis, anti-tubercular activity and 3D-QSAR study
of coumarin-4-acetic acid benzylidene hydrazides
Atul Manvar c, Alpeshkumar Malde b, Jitender Verma b, Vijay Virsodia c, Arun Mishra a,
c,
b,**
, Anamik Shah
*
Kuldip Upadhyay d, Hrishikesh Acharya a, Evans Coutinho
a Jubilant Organosys, 1A, Sector 16A, Noida 201 301, Uttar Pradesh, India
b Department of Pharmaceutical Chemistry, Bombay College of Pharmacy, Kalina, Santacruz (E), Mumbai 400 098, India
c Department of Chemistry, Saurashtra University, University Road, University Campus, Rajkot 360 005, Gujarat, India
d Torrent Research Centre, Bhat, Gandhinagar, India
Received 24 May 2007; received in revised form 3 January 2008; accepted 3 January 2008
Available online 29 January 2008
Abstract
A set of 25 coumarin-4-acetic acid benzylidene hydrazides were synthesized and characterized by NMR, IR and mass spectroscopic tech-
niques. The compounds were evaluated for their anti-tubercular activity against Mycobacterium tuberculosis H37Rv strain using the BACTEC
460 system to determine percentage inhibition. To understand the relationship between structure and activity, a 3D-QSAR analysis has been
carried out by Comparative Molecular Field Analysis (CoMFA). Several statistically significant CoMFA models were generated. The CoMFA
model generated with database alignment was the best in terms of overall statistics. The CoMFA contours provide a good insight into the structure
activity relationships of the compounds reported herein.
Ó 2008 Elsevier Masson SAS. All rights reserved.
Keywords: Coumarin-hydrazides; Anti-tubercular activity; CoMFA; 3D-QSAR study
1. Introduction
drug discovery technologies [5]. The current WHO-approved
treatment for TB, known as directly observed therapy short-
course (DOTS), involves an intensive phase with three or
four different drugs viz. isoniazid (INH), rifampin (RFP),
pyrazinamide (PZA), and ethambutol (EB) for a minimum
of 6 months [6]. INH, a well-known anti-tubercular drug, is
believed to kill mycobacteria by inhibiting the biosynthesis
of mycolic acids, critical components of the mycobacterial
cell wall. The catalase and peroxidase activities are thought
to participate in the drug sensitivity mechanism by converting
INH in vivo into its biologically active form, which then act on
its intracellular target [7]. In analogy to INH, pyridines
substituted with alkylated tetrazoles (designed as lipophilic
precursors of isosteres of isonicotinic acid) have been reported
to possess anti-tubercular activity against the H37Rv strain of
M. tuberculosis. These compounds, after penetration of the
mycobacterial cell wall, are very likely biotransformed by
esterases or peroxidaseecatalases. They are more active than
the unmodified polar isosteres of isonicotinic acid, which
Mycobacterium tuberculosis, a human pathogen causing
tuberculosis (TB), claims more human lives than any other
bacterial pathogens [1,2]. It is estimated that 9 million new
TB cases and approximately 2 million deaths occurred world-
wide in the year 2004, and more that 80% of those were in
developing countries [3]. According to WHO, from 2002 to
2020, there will be about 1 billion more people newly infected
with TB and approximately 36 million deaths, if the world-
wide ravage of tuberculosis is left unchecked [4]. As regards
new drug development, not a single new anti-tubercular agent
has been launched since the introduction of rifampicin in
1965, despite the great advances that have been made in
* Corresponding author. Tel.: þ91 281 2581013; fax: þ91 281 2576802.
** Corresponding author. Tel.: þ91 22 26670905.
E-mail addresses: evans@bcpindia.org, evans@bcp.edu.in (E. Coutinho),
anamik_shah@yahoo.com, anamik_shah@hotmail.com (A. Shah).
0223-5234/$ - see front matter Ó 2008 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2008.01.016

2396
A. Manvar et al. / European Journal of Medicinal Chemistry 43 (2008) 2395e2403
may be due to better penetration of these agents into the cell
wall of the mycobacteria [8].
Scheme 1). To probe the relationship between structure and
activity for this class of compounds, Comparative Molecular
Field Analysis (CoMFA) was also carried out.
The failure of patients to complete the therapy has led to
the emergence of multidrug-resistant TB (MDR-TB), and
multidrug-resistant (MDR) strains of M. tuberculosis that are
resistant to the two major drugs isonicotinic acid hydrazide
and rifampicin has further complicated the situation. The
AIDS pandemic has led to an explosion of HIV/TB co-infec-
tion for patients living with HIV/AIDS. TB is a leading cause
of death among people who are HIV-positive (13% of AIDS
deaths worldwide). Consequently, there is a pressing need
for the development of novel drugs for treating AIDS related
TB that are effective against both drug-sensitive and drug-
resistant MTB strains. The development of potent new anti-
tubercular agents with low-toxicity profiles, effective against
both drug-susceptible and drug-resistant strains of M. tubercu-
losis and capable of shortening the current duration of therapy
are the need of the hour [9]. Recognizing these serious facts,
we initiated a program to synthesize and screen diverse hetero-
cyclic entities like quinolines and coumarins as anti-tubercular
agents [10,11]. Among the heterocyclic group of compounds,
naturally occurring derivatives of coumarin like Calanolide A
show potent anti-tubercular activity with a MIC value of
3.13 mg/ml, and Calanolide B which is readily available
from the Calophyllum seed, is claimed to have a comparable
spectrum of activity as Calanolide A [12]. Inspired by this
fact, we set upon a program of making anti-tubercular agents
using the central coumarin ring in Calanolide A (structure A in
Scheme 1) as the template and adding substituents as we
deemed necessary to impart activity, on the various positions
on the coumarin ring. As the first part of a series of coumarin
analogs, we report here work on coumarin-4-acetic acid ben-
zylidene hydrazides as anti-tubercular agents (structure B in
2. Results and discussion
2.1. Chemistry
The synthetic route employed to produce the series of 25
compounds 5(aey) are portrayed in Scheme 2. Citric acid in
the presence of sulphuric acid at low temperature (10 ꢀC)
yields acetone dicarboxylic acid, which was not isolated, but
directly reacted with phenol (1) to produce coumarin-4-acetic
acid (2) under Pechmann reaction conditions [13]. In the sec-
ond step, esterification of coumarin-4-acetic acid was carried
out with methanol in presence of an acid catalyst (3). Couma-
rin-4-acetic acid hydrazide (4) was synthesized by refluxing
coumarin-4-acetic acid methyl ester with hydrazine hydrate.
Final reflux with aldehyde in alcohol leads to the coumarin-
4-acetic acid benzylidene hydrazides (5aey).
2.2. Biology
All synthesized compounds were screened for anti-tubercular
activity against M. tuberculosis H37Rv using rifampicin as the
standard. MIC values were determined by the protocol discussed
in Section 4.2, and the percentage inhibition was found to lie in
a large range of 6e93%. It is important to note that the 7-
hydroxycoumarin derivative with R6]NO2 (Table 2) shows
the highest anti-TB activity, while introduction of methyl groups
at R1 and R4 gives the second most potent compound. The
dimethyl derivatives with nitro (5f) and chloro (5c) groups at
R6 show anti-TB activity with 65 and 62% inhibition, respec-
tively. The effect of the electron donating methyl group on the
coumarin skeleton along with electron withdrawing groups on
the phenyl hydrazide moiety, on the activity, makes an interest-
ing observation.
A
O
2.3. CoMFA studies
The geometry of the groups at the 4-position of the couma-
rin ring was adopted from the X-ray structure of a related mol-
ecule (vide supra). Various CoMFA14 models were generated
for two alignments, namely database and field fit, by varying
the grid size and orientation of the aligned molecules in the
grid. Only the two best models are reported here; model 1 ob-
tained by database alignment and model 2 derived from field fit
alignment. The statistics of these models are shown in Table 3.
The activities of molecules 5d and 5j in the test set were poorly
predicted by both models. The inclusion of these molecules in
the test set decreased the value of r2cv significantly. Hence, these
molecules were considered as outliers and not included in the
calculation of the predictive correlation coefficient (r2pred).
The two best models have parameters, which are statistically
significant. Model 1 has overall better statistical qualities e
a conventional correlation coefficient (r2) of 0.983, a cross-
validated correlation coefficient (q2) of 0.636, and a predictive
O
O
O
HO
B
R7
R8
R6
O
R5
N-HN
O
R1
R2
R3
O
R4
Scheme 1.

A. Manvar et al. / European Journal of Medicinal Chemistry 43 (2008) 2395e2403
2397
O
O
R1
HO
R1
MeO
R1
R2
R3
R2
R3
R2
R3
MeOH/H+
Reflux
Citric acid
H2SO4/10°C
O
HO
O
O
O
R4
R4
R4
1
2
3
Hydrazine
98 Reflux
CHO
R4
O
R5
R6
O
O
R3
H2NHN
R1
R8
R2
R2
R7
Ethanol/80°C
R1
NHN
O
R5
O
O
R3
R6
R4
R8
R7
5a-y
4
Scheme 2.
correlation coefficient (r2pred) of 0.338. Although the model
obtained by field fit alignment exhibits higher r2pred (0.464),
the values of r2 and rc2v are relatively lower. Plots of the pre-
dicted vs. experimental activity of the training set molecules
for models 1 and 2 are shown in Fig. 2.
The CoMFA model, with its hundreds or thousands of
terms, is generally represented as a 3D ‘coefficient contour’.
Colored contours in the map represent those areas in 3D space
where changes in the steric and electrostatic field values of
a compound correlate strongly with concomitant change in
its biological activity. The CoMFA steric and electrostatic
contour plots of the best models are shown in Figs. 3 and 4,
respectively.
Analysis of the steric contours of model 1 (Fig. 3A) reveals
large green colored contours surrounding almost all the posi-
tions of the phenyl ring of the benzylidene hydrazide moiety.
No yellow colored contours are visible, neither are any signif-
icant steric contours seen around the coumarin ring. The steric
contours of model 2 (Fig. 3B) also lead to the same interpre-
tation. The molecules with medium size substituents, for e.g.
methoxy group, on the phenyl ring exhibit good inhibitory
activity. For this very reason, molecules 5h and 5u with
a trisemethoxyphenyl, molecule 5e with o-methoxyphenyl,
molecule 5r with p-methoxyphenyl and molecule 5v with
m,p-dimethoxyphenyl moieties have high inhibitory activities.
Analysis of the electrostatic contours of model 1 (Fig. 4A)
reveals blue colored contours near the meta- and para-positions
of the phenyl ring of the benzylidene hydrazide moiety. Also
there are two medium-sized blue colored contours away from
the ortho-position of the phenyl ring. No electrostatic contours
are seen around the coumarin ring. The electrostatic contours
of model 2 (Fig. 4B) advocate almost similar requirements
for activity as far as the benzylidene hydrazide moiety is
concerned. In addition, there is also a big blue colored contour
in the vicinity of the C6 and C7 positions, and a red colored
contour close to the C7 and C8 positions of the coumarin
ring. The highly active molecule (5y) in the series possesses
an electronegative eOH substituent at the C7 position of the
coumarin ring, which is near the favored red colored contour,
leading to good activity. Molecules with a C6 methyl substitu-
ent (5ne5x) exhibit moderate to good inhibitory activity as the
methyl group is near the electropositive favorable blue colored
contour. Molecules 5je5n also have a C6 methyl substituent,
but these have low activity as the additional C7 methyl substit-
uent is near the disfavored red colored contour, which contrib-
utes to the reduction in their activity. Molecules with a methoxy
substituent (5e, 5h, 5r, 5u, 5v) in the phenyl ring of the benzy-
lidene hydrazide moiety have their methyl portion buried
inside the favorable blue colored contour and thus exhibit
good inhibitory activity. These blue colored contours partially
overlap with the sterically favorable green colored contour
around the phenyl ring.
3. Conclusions
Coumarin-4-acetic acid benzylidene hydrazides were syn-
thesized in four convenient steps in ca. 60e70% yields. All
compounds were tested for activity against M. tuberculosis
H37Rv strain using rifampicin as the standard with the BAC-
TEC 460 system to determine the percentage inhibition. A
3D-QSAR analysis of these compounds was carried out with
CoMFA to map the structural features contributing to the
inhibitory activity of these molecules. A number of models
were generated based on database alignment as well as field

2398
A. Manvar et al. / European Journal of Medicinal Chemistry 43 (2008) 2395e2403
Table 1
Physical data of coumarin-4-acetic acid benzylidene hydrazides (5aey)
R4
R3
R2
O
O
R5
H
H
N
R6
R1
N
O
R7
R8
No.
R1
R2
R3
R4
R5
R6
R7
R8
Mp ꢀC
Yield %
5a
5b
5c
5d
5e
5f
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
H
H
H
H
H
H
H
H
H
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
H
Cl
H
H
H
H
H
H
H
H
H
H
174e176
168e170
157e159
163e165
161e163
175e177
181e183
185e187
135e137
178e179
145e147
165e167
201e203
189e191
201e203
195e197
198e200
150e152
168e170
168e170
178e180
185e187
184e186
160e162
175e177
62
68
67
59
55
69
56
61
67
60
61
69
71
59
60
64
60
55
62
58
57
61
61
59
60
H
H
Cl
H
H
H
H
NMe2
H
H
H
OMe
NO2
H
H
H
H
H
5g
5h
5i
H
NO2
OMe
H
H
H
H
H
OMe
H
OMe
H
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
CH3
CH3
CH3
CH3
CH3
H
Cl
H
5j
H
H
H
Cl
H
5k
5l
H
H
H
H
SMe
H
H
H
H
NO2
H
H
H
5m
5n
5o
5p
5q
5r
5s
H
H
NO2
H
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
NO2
H
H
H
H
H
H
H
OMe
F
H
H
H
H
H
H
H
5t
H
H
H
OH
H
H
H
H
5u
5v
5w
5x
5y
H
H
H
OMe
OMe
OMe
Cl
OMe
OMe
H
OMe
H
H
H
H
H
H
H
H
H
H
H
H
OH
H
H
H
H
H
H
H
NO2
H
H
fit alignment methodologies. Of the various models evaluated,
the CoMFA model derived from database alignment was the
best with a good correlation and predictive power. Analysis
of the CoMFA contours provides details on the fine relation-
ship linking structure and activity, and provides clues for struc-
tural modifications that can improve the activity.
300 MHz spectrometer in DMSO-d6. Chemical shifts (d) are
given in ppm relative to TMS, coupling constants (J ) in Hz.
FAB-Mass spectra were recorded on Jeol SX 102/DA-6000
and IR spectra were recorded on a Shimadzu FTIR-8400 using
KBr optics.
4.1.1. Synthesis of substituted coumarin-4-acetic acid (2)
It was synthesized according to the procedure reported in
the literature [13]. Accordingly, a mixture of citric acid
(1 mol) and conc. sulphuric acid (32 ml) was stirred for half
an hour, then the temperature was slowly raised during an
interval of 10e15 min and as soon as the evolution of gas
slackened, the flask was removed form the bath, allowed to
stand for 15 min till the reaction mixture became clear and
free from carbon monoxide bubbles; this was then cooled to
10 ꢀC. To this solution, substituted phenol (1 mol) was added
at 10 ꢀC, drop wise. After the addition of phenol, the reaction
4. Experimental
4.1. Chemistry
All the chemicals and solvents were purchased from Spec-
trochem, S D Fine Chemicals and Loba Chemie (India) and
used without further purification. Melting points were deter-
mined in open capillary tubes with an Electrothermal-9200
melting point apparatus and are reported here as uncorrected
1
values. H NMR spectra were recorded on Bruker Avance II

A. Manvar et al. / European Journal of Medicinal Chemistry 43 (2008) 2395e2403
2399
Table 2
Anti-tubercular activity of coumarin-4-acetic acid benzylidene hydrazides
(5aey)
Table 3
A summary of the statistics of the 3D-QSAR models
Parameter
Database alignment
Field fit alignment
R4
CoMFA model 1
CoMFA model 2
N
6
5
R3
R2
O
O
r2
0.983
0.636
0.522
116
0.951
0.403
0.473
51
r2cv
LGO (5)
F-value
r2pred
r2bs
R5
H
H
N
0.338
0.992
0.052
0.464
0.991
0.056
R6
R1
SD
N
Contributions (%)
Steric
Electrostatic
O
59.5
40.5
53.1
46.9
R7
N ¼ optimum number of components, r2 ¼ conventional (non-cross-validated)
correlation coefficient, r2cv ¼ cross-validated correlation coefficient using
SAMPLS, LGO (5) ¼ cross-validation correlation coefficient by Leave-
Group-Out (in groups of 5), F-value ¼ Fisher statistics, r2pred ¼ predictive
(test molecules) correlation coefficient, r2bs ¼ correlation coefficient after 100
runs of bootstrapping analysis, and SD ¼ standard deviation from 100 boot-
strapping runs.
R8
No.
Conc. (mg/ml)
% Inhibition
logit Values
5a
5b
5c
5d
5e
5f
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.25
70
48
62
7
4.80
4.44
4.68
3.36
4.53
4.75
4.29
4.57
3.78
4.14
3.29
3.57
4.13
4.42
4.47
4.12
4.14
4.44
4.28
4.85
4.53
4.40
4.12
4.18
5.89
54
65
39
52
17
32
6
4.1.3. Synthesis of coumarin-4-acetic acid hydrazides (4)
A mixture of coumarin-4-acetic acid methyl ester (0.1 mol)
and hydrazine hydrate 98% (0.5 mol) was refluxed for an
appropriate time. The reaction was monitored by TLC (metha-
nol: chloroform:0.5:9.5). After completion of the reaction as
indicated by TLC, the reaction mixture was poured into water
and the solid which separated filtered off. It was the recrystal-
lized from ethanol to give the title compounds (yield 65e72%).
5g
5h
5i
5j
5k
5l
11
31
34
34
19
19
33
26
55
34
29
19
21
93
5m
5n
5o
5p
5q
5r
5s
4.1.4. General synthesis of 4-(substituted
benzylidenacetohydrazide)-coumarin (5aey)
A mixture of coumarin-4-acetic acid hydrazides (0.01 mol)
and aromatic aldehydes (0.01 mol) were refluxed in absolute
alcohol for an appropriate time. The reaction was monitored
by TLC (ethylacetate:hexane:8:2). After completion of the
reaction as indicated by TLC, ethanol was evaporated, to
give the crude title compounds. Finally purification was
achieved on silica gel (100e200 mesh) column chromatogra-
phy using ethyl acetate/hexane as the eluents.
5t
5u
5v
5w
5x
5y
mixture was stirred at room temperature for 48 hr. The reac-
tion mixture was then poured onto crushed ice; the separated
solid was filtered and dissolved in saturated sodium bicarbon-
ate solution which on acidification gave the title compounds
(yield 55e70%).
4.1.4.1. 4-(20-methoxy benzylidenacetohydrazide)-5,8-dime-
thylcoumarin 5e. 1H NMR (DMSO-d6) d ppm: 2.21 (s, 3H,
eCH3), 2.24 (s, 3H, -CH3), 3.24 (s, 3H, eOCH3), 3.39e
4.09 (m, 2H, eCH2), 6.35 (d, 1H, eCH), 6.79 (q, J ¼ 9.7,
1H, eCH), 7.22 (d, J ¼ 8.9, 1H, AreH), 7.45 (d, J ¼ 8.8,
1H, AreH), 7.49e7.70 (m, 3H, AreH); 13C NMR (d):
171.5, 161.5, 159.2, 156.4, 150.9, 144.2, 134.5, 133.9,
133.1, 130.2, 129.3, 128.6, 125.1, 122.3, 117.5, 114.9,
112.9, 56.2, 44.9, 19.1, 14.8; IR (KBr): cmꢁ1 3079 (]CeH),
2970 (CeH as), 2865 (CeH sy), 1695 (C]N), 1687 (C]O,
amide), 1717 (C]O, lactone), 1644 (C]C), 1254 (CeO);
FAB Mass: m/z 364 (Mþ).
4.1.2. Synthesis of coumarin-4-methylacetate (3)
The substituted coumarin-4-acetic acid (1 mol) was dis-
solved in methanol (100 ml), and a few drops of sulphuric
acid were added. The resulting reaction mixture was refluxed
for an appropriate time. After the completion of the reaction
as indicated by TLC (ethylacetate:hexane:4:6), methanol was
evaporated and the resulting reaction mixture extracted with
ethyl acetate, washed with sodium bicarbonate, then with brine
and dried over anhydrous sodium sulphate. The solvent was
removed in vacuo to give the title compounds (yield 75e79%).
4.1.4.2. 4-(40-thiomethoxy benzylidenacetohydrazide)-6,7-di-
methylcoumarin 5k. 1H NMR (DMSO-d6) d ppm: 2.29 (s,
3H, eCH3), 2.32 (s, 3H, eCH3), 2.47 (s, 3H, eSCH3)

2400
A. Manvar et al. / European Journal of Medicinal Chemistry 43 (2008) 2395e2403
Fig. 1. A view of the molecules aligned using (A) database and (B) field fit alignment strategies.
3.41e4.13 (m, 2H, eCH2), 6.29 (d, 1H, eCH), 6.79 (q,
J ¼ 9.2, 1H, eCH), 7.75 (s, 1H, AreH), 7.29 (s, 1H, AreH),
7.67 (m, 3H, AreH); 13C NMR (d): 171.9, 161.7, 155.4,
147.7, 144.2, 139.2, 137.1, 134.2, 130.7, 129.9, 127.3,
127.0, 120.9, 118.5, 113.2, 45.9, 18.8, 17.9, 15.2; IR (KBr):
cmꢁ1 3087 (]CeH), 2960 (CeH as), 2872 (CeH sy), 1690
(C]N), 1680 (C]O, amide), 1719 (C]O, lactone), 1652
(C]C), 795 (CeS); FAB Mass: m/z 380 (Mþ).
J ¼ 10.8, 1H, eCH), 7.81e7.12 (m, 6H, AreH); 13C NMR
(d): 172.1, 161.5, 155.7, 147.6, 144.2, 136.9, 135.6, 131.9,
131.2, 129.4, 129.1, 127.6, 122.0, 121.4, 112.8, 45.8, 24.8; IR
(KBr): cmꢁ1 3071 (]CeH), 2985 (CeH as), 2885 (CeH
sy), 1684 (C]N), 1688 (C]O, amide), 1722 (C]O, lactone),
1639 (C]C), 790 (AreCl); FAB Mass: m/z 355 (Mþ).
4.1.4.5. 4-(30,40-dimethoxy benzylidenacetohydrazide)-6-meth-
ylcoumarin 5v. 1H NMR (DMSO-d6) d ppm: 2.30 (s,
3H, eCH3), 3.19 (s, 3H, eOCH3), 3.22 (s, 3H, eOCH3)
3.41e4.11 (m, 2H, eCH2), 6.39 (d, 1H, eCH), 6.81 (q,
J ¼ 9.3, 1H, eCH), 7.65e711 (m, 5H, AreH); 13C NMR
(d): 171.8, 161.4, 155.3, 152.5, 150.2, 147.7, 143.9, 135.5,
129.0, 127.6, 127.2, 122.9, 121.9, 121.4, 115.9, 115.0, 112.7,
56.5, 56.2, 45.8, 25.0; IR (KBr): cmꢁ1 3069 (]CeH), 2980
(CeH as), 2869 (CeH sy), 1685 (C]N), 1680 (C]O,
amide), 1720 (C], lactone), 1645 (C]C), 1260 (CeO);
FAB Mass: m/z 379 (Mþ).
4.1.4.3. 4-(benzylidenacetohydrazide)-6-methylcoumarin 5n.
1H NMR (DMSO-d6) d ppm: 2.35 (s, 3H, eCH3), 3.49e4.09
(m, 2H, eCH2), 6.31 (d, 1H, eCH), 6.79 (q, J ¼ 10.1,
1H, eCH), 7.81e7.21 (m, 8H, AreH); 13C NMR (d):
171.7, 161.3, 155.8, 147.5, 143.4, 135.3, 134.3, 131.5,
129.9, 129.4, 129.1, 127.5, 121.7, 121.0, 113.2, 46.2,
25.2; IR (KBr): cmꢁ1 3070 (]CeH), 2977 (CeH as),
2875 (CeH sy), 1679 (C]N), 1681 (C]O, amide), 1724
(C]O, lactone), 1649 (C]C); FAB Mass: m/z 320 (Mþ).
4.1.4.4. 4-(40-chloro benzylidenacetohydrazide)-6-methylcou-
marin 5p. 1H NMR (DMSO-d6) d ppm: 2.30 (s, 3H, eCH3),
3.47e4.19 (m, 2H, eCH2), 6.37 (d, 1H, eCH), 6.85 (q,
4.1.4.6. 4-(30-methoxy benzylidenacetohydrazide)-6-methylcou-
marin 5w. 1H NMR (DMSO-d6) d ppm: 2.25 (s, 3H, eCH3),
3.21 (s, 3H, eOCH3), 3.39e4.19 (m, 2H, eCH2), 6.50 (d, 1H,
Fig. 2. Predicted vs. experimental activity for molecules in the training set based on models 1 and 2.

A. Manvar et al. / European Journal of Medicinal Chemistry 43 (2008) 2395e2403
2401
Fig. 3. CoMFA contours showing steric fields around molecule 5y for (A) model 1 and (B) model 2. The green colored contour favors steric bulk, while sites where
steric bulk is disfavored are shown in yellow [for interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article].
eCH), 6.77 (q, J ¼ 10.6, 1H, eCH), 7.79e7.25 (m, 6H, Are
H); 13C NMR (d): 171.3, 160.9, 161.4, 155.5, 147.8, 144.1,
135.7, 135.2, 130.4, 129.0, 127.5, 121.9, 121.7, 121.0,
117.2, 113.9, 112.8, 56.3, 45.7, 24.5; IR (KBr): cmꢁ1 3070
(]CeH), 2987 (CeH as), 2875 (CeH sy), 1677 (C]N),
1689 (C]O, amide), 1712 (C]O, lactone), 1641 (C]C),
1269 (CeO); FAB Mass: m/z 350 (Mþ).
concentration (MIC, lowest concentration inhibiting 99% of
the inoculums, MIC value of rifampicin is 0.25 g/ml at 95%
inhibition of H37Rv strain). M. tuberculosis H37Rv strain
(ACTT 27294; American type culture collection, Rockville,
MD) was cultured at 37 ꢀC on a rotary shaker in middle brook
7H9 broth (Difco Laboratories, Detroit, MI) supplemented
with 0.2 v/v glycerol and 0.05% v/v Tween 80, until the
culture turbidity achieved an optical density of 0.45e0.55 at
550 nm. Bacteria were pelleted by centrifugation, washed
twice and resuspended in one fifth of the original volume in
Dulbecoo’s phosphate buffered saline [PBS, Irvine Scientific,
Santa Ana (A)]. Large bacterial clumps were removed by pas-
sage through an 8 mm filter (Malgene, Rochester, NY) and al-
iquots were frozen at ꢁ80 ꢀC. Cultures were prepared and an
appropriate dilution performed such that a BACTEC-12B vial
inoculated with 0.1 ml would reach a growth index (GI) of 999
in 5 days. One tenth of the diluted inoculum was used to inoc-
ulate 4 ml of fresh BACTEC-12B broth containing the test
compounds. An additional control vial was included which
received a further 1:100 diluted inoculum (as well as 50 mL
DMSO) for use in calculating the MIC of rifampicin, respec-
tively, by established procedures. Cultures were incubated at
37 ꢀC and the Growth of Inhibition (GI) determined daily until
4.2. Anti-tubercular activity
Anti-tubercular activity was determined using the modified
BACTEC 460 system in which stock solutions as test com-
pounds were prepared in dimethylsulfoxide (DMSO) at
1 mg/ml and sterilized by passage through 0.22 mm PFTE
filters (Millex-FG, Millepore, Bedford, MA) 50 ml was added
to 4 ml radiometric 7H12 broth (BACTEC-12B; Bectron Dick-
inson Diagnostic Instrument system, Sparks, MD) to achieve
a final concentration of 12.5 mg/ml (5aem) and 6.25 mg/ml
(5ney) (Table 2). Controls received 50 mL DMSO. Rifampin
(Sigma Chemicals Co., St. Louis, MO) was included as a pos-
itive drug control. Rifampicin was solubilized and diluted in
DMSO and added to BACTEC-12 broth to achieve a range
of concentration for determination of minimum inhibitory
Fig. 4. CoMFA contours showing electrostatic fields around molecule 5y for (A) model 1 and (B) model 2. The red contour shows regions where electronegative
substituents are favored, while the blue contour is associated with positions where electropositive substituents improve activity [for interpretation of the references
to color in this figure legend, the reader is referred to the web version of this article].

2402
A. Manvar et al. / European Journal of Medicinal Chemistry 43 (2008) 2395e2403
control cultures achieved a GI of 999. Assays were usually
completed in 5e8 days. Percent inhibition was defined as 1-
(GI of test sample/GI of control) ꢂ 100. Minimum inhibitory
concentration of compound effecting a reduction in daily
change in GI, which was less than that observed with
a 1:100 diluted control culture on day the later reached a GI
of at least 30.
4.3.4. Alignment
The most crucial input for the CoMFA is the alignment of
the molecules. Molecule 5y with the highest activity was
chosen as the template and all other molecules were aligned
to it using the database alignment method in Sybyl
(Fig. 1A). The molecules were aligned with reference to the
coumarin ring. A second alignment was also carried out using
the field fit method, where the steric and electrostatic fields
around the molecules were superimposed over the same fields
of the template molecule (Fig. 1B).
4.3. Computational details
4.3.1. Data set
A set of 25 molecules (Table 1) was used in the 3D-QSAR
study. The data set was divided into a training set (19 mole-
cules) and a test set (6 molecules: i.d.’s 5b, 5d, 5j, 5q, 5u,
5w) by means of chemical as well as biological diversity. Day-
light fingerprints of the molecules along with the pIC50 data
were used to separate the molecules into training and test
sets based on the Tanimoto similarity coefficient [15].
4.3.5. CoMFA interaction energy calculation
The steric and electrostatic fields in CoMFA were calcu-
lated using an sp3 carbon atom with þ1.0 charge as the probe.
The van der Waals potential and Coulombic energy between
the probe and the molecule were calculated using the standard
Tripos force field. A distance-dependent dielectric constant of
1.0r was used in the calculation of the electrostatics. The steric
field was truncated at points where the value exceeded
ꢄ30.0 kcal/mol, and the electrostatic fields were ignored at
those lattice points where the steric interactions were high.
In order to investigate the effect of grid spacing, initially the
CoMFA models were developed at varying grid spacing values
4.3.2. Biological data
For the QSAR study, the activity values were transformed
as follows [16]:
Activity ¼ ꢁlog c þ logit
˚
(i.e. 0.5, 1.0, 1.5 and 2.0 A). In both database as well as field
fit alignment of molecules, the best q2 values were obtained
where c is the molar concentration ¼ concentration (mg/ml) ꢂ
˚
when the grid spacing was set to 2.0 A which is the default
grid spacing in Sybyl. Thus, further model development was
0.001/(molecular weight).
˚
done at a grid spacing of 2.0 A.
logit ¼ log½% inhibition=ð100 ꢁ % inhibitionÞꢃ
Further, the influence of the overall orientation of the
aligned molecules in the grid was examined by a protocol sim-
ilar to that of Cho and Tropsha [20]. For this, starting from
a random orientation, the entire aggregate of aligned mole-
cules was rotated systematically in steps of 90ꢀ around the
x, y and z-axes using the STATIC ROTATE command in Sybyl.
For each orientation, the optimum number of components and
q2 values were obtained. The q2 values varied from 0.165 to
0.636 for database alignment and from 0.108 to 0.403 for field
fit alignment. The orientation of the aligned molecules with
the best q2 values was selected for further model development.
4.3.3. Molecular modeling
The CoMFA [14] studies were carried with Sybyl 7.1 [17]
installed on a Pentium 2.8 GHz PC running under the Linux
OS (Red Hat Enterprise WS 2.3.1).
The coumarin ring, common to all molecules, is a rigid
moiety. The orientation of the substituent at the 4th position
of the coumarin ring was derived from a related crystal struc-
ture cr1165 (Scheme 3) [18]. The geometry of the eCON-
HN]CHePh group was derived from the crystal structure
of bt2126 (Scheme 2). The structures of the molecules were
built with the Sketcher module in Sybyl and energy minimized
by Powell’s method using the MMFF94 force field [19] with
a distance-dependent dielectric term.
4.3.6. Partial Least Squares (PLS) analysis
The PLS method was used to set up a correlation between
the molecular fields and the inhibitory activity of the mole-
cules. The optimal number of components was determined
with SAMPLS (Samples-distance Partial Least Square) [21]
and cross-validation was carried out by the leave-one-out
method. The model with the optimum number of components
(highest q2) and with the lowest standard error of prediction
(SDEP) was considered for further analysis. Equal weights
were assigned to the steric and electrostatic fields by the
COMFA_STD scaling option. To speed up the analysis and
reduce noise, columns with an s value below 2.0 kcal/mol
were filtered off and the conventional r2 was calculated using
the optimum number of components. The final models were
then refined by the Region Focusing [20] option in Sybyl,
which concentrates only on those lattice points which are
most pertinent to the model and thus enhances its resolution
O
O
Cl
Cl
N
H
cr1165
O
N
N
N
N
H
bt2126
H
Scheme 3.

A. Manvar et al. / European Journal of Medicinal Chemistry 43 (2008) 2395e2403
2403
and predictive power. To further assess the robustness and sta-
tistical confidence of the derived models, cross-validation by
Leave-Group-Out22 (LGO in groups of 5) and bootstrapping
[22] analysis for 100 runs was performed. Bootstrapping in-
volves the generation of many new data sets from the original
data set and is obtained by randomly choosing samples from
the original data set. The statistical calculation is performed
on each of these bootstrap samplings. The difference between
the parameters calculated from the original data set and the
average of the parameters calculated from the many bootstrap
samplings is a measure of the bias of the original calculations.
Models with a cross-validation (q2) value above 0.3 were
sought, since at this value the probability of chance correlation
is less than 5% [23].
and Technology (DST), India for the financial assistance.
Alpeshkumar Malde and Jitender Verma thank the Council of
Scientific and Industrial Research (CSIR), India for the award
of a Senior Research Fellowship. The computational facilities
at BCP were provided by the All India Council for Technical Ed-
ucation through grant F. No. 8022/RID/NPROJ/RPS-5/2003-04.
References
[1] P.J. Dolin, M.D. Raviglione, Bull. WorldHealthOrgan. 72(1994) 213e220.
[2] A.M. Rouchi, Chem. Eng. News 77 (1999) 52e69.
[3] Global Tuberculosis Control: Surveillance, Planning, Financing WHO
Report, World Health Organization, Geneva, 2006, (WHO/HTM/TB/
2006.362).
[4] WHO fact sheet Available from: http://www.who.int/mediacentre/
factsheets/who104/en/print.html.
[5] W.J. Burman, B.E. Jones, Am. J. Respir. Crit. Care Med. 164 (2001) 7e12.
[6] R.V. Somu, H. Boshoff, C. Qiao, E.M. Bennett, C.E. Barry, C.C. Aldrich,
J. Med. Chem. 49 (2006) 31e34.
4.3.7. Predictive correlation coefficient
The predictive ability of each 3D-QSAR model was deter-
mined from a test set of 6 molecules not included in the model
generation. The predictive correlation coefficient (r2pred), based
on the test set molecules, is defined as
[7] K. Johnsson, P.G. Schultz, J. Am. Chem. Soc. 116 (1994) 7425e7426.
[8] G.A. Wachter, M.C. Davis, A.R. Martin, S.G. Franzblau, J. Med. Chem.
¨
41 (1998) 2436e2438.
[9] (a) S. Kaufman, J. Van Embden, Trends. Microbiol. 1 (1993) 2e5;
(b) B.R. Bloom, C.J. Murray, Science 257 (1992) 1055e1064;
(c) M.A. Fischl, G.L. Daikos, R.B. Uttamchandani, R.B. Poblete,
J.N. Moreno, R.R. Reyes, A.M. Boota, L.M. Thompson, T.J. Cleary,
S.A. Oldham, Ann. Intern. Med. 117 (1992) 184e190.
[10] N. Dodia, A. Shah, Ind. J. Heterocycl. Chem. 9 (1999) 139e142.
[11] N. Dodia, A. Shah, Ind. J. Heterocycl. Chem. 10 (2000) 69e70.
[12] A. Nayyar, R. Jain, Curr. Med. Chem. 12 (2005) 1873e1886.
[13] (a) V. Pechmann, Duisberg, Ber. Dtsch. Chem. Ges. 16 (1883) 2127;
(b) V. Pechmann, Kraft, Ber. Dtsch. Chem. Ges. 34 (1901) 421;
(c) B.B. Dey, J. Chem. Soc., Perkin Trans. 1 107 (1915) 1606e1651;
(d) D.J. Limaye, J. Indian Chem. Soc. 4 (1927) 159e161.
[14] R.D. Cramer, D.E. Patterson, J.D. Bunce, J. Am. Chem. Soc. 110 (1988)
5959e5967.
rp2red ¼ ðSD ꢁ PRESSÞ=SD
where SD is the sum of squared deviations between the biolog-
ical activity of the test set and the mean activity of the training
set molecules and the PRESS is the sum of squared deviations
between predicted and actual activity values for every mole-
cule in the test set.
4.3.8. CoMFA contour maps
Contour maps were generated as a scalar product of coeffi-
cients and standard deviation (SD ꢂ Coeff) associated with
each column. Favored and disfavored levels, fixed at 80 and
20%, respectively, were used to display the steric and the elec-
trostatic fields. The contours for steric fields are shown in
green (more bulk favored) and yellow (less bulk favored),
while the electrostatic field contours are displayed in red (elec-
tronegative substituents favored) and blue (electropositive sub-
stituents favored) colors.
[15] E.J. Martin, J.M. Blaney, M.A. Siani, D.C. Spellmeyer, A.K. Wong,
W.H. Moos, J. Med. Chem. 39 (1995) 1431e1436.
[16] CoMFA and QSAR Manual, Sybyl 7.1, Associates Inc., 1699 S Hanley
Road, St. Louis, MO 631444, USA.
[17] Sybyl 7.1, Tripos Associates Inc., 1699 S Hanley Road, St. Louis, MO
631444, USA.
[18] (a) A. Jain, M.V. Kulkarni, N.C. Shivaprakash, Acta Crystallogr. C62
(1995) 2585e2586;
(b) Y. Cui-Cui, W. Li, X. Zhi-Yong, S. Xue-Fang, Acta Crystallogr. E62
(2006) o3274eo3275.
Acknowledgements
[19] T.A. Halgren, J. Am. Chem. Soc. 112 (1990) 4710e4723.
[20] Sung Jin Cho, Alexander Tropsha, J. Med. Chem. 38 (1995) 1060e
1066.
The authors are grateful to Department of Chemistry, Sau-
rashtra University, Rajkot, India for providing necessary infra-
structural facilities for carrying out synthetic work, and Dr.
Cecil Kwong, TAACF, USA for anti-tubercular activity of the
compounds. Atul Manvar is thankful to Department of Science
[21] B.L. Bush, R.B. Nachbar, J. Comput.-Aided. Mol. Des. 7 (1993) 587e619.
[22] R.D. Cramer III, J.D. Bunce, D.E. Patterson, I.E. Frank, Quant. Struct.
Act. Relat. 7 (1988) 18e25.
[23] M. Clark, R.D. Cramer, D.M. Jones, D.E. Patterson, P.E. Simeroth,
Tetrahedron Comput. Methodol. 3 (1990) 47e59.