Synthesis and structure-antituberculosis activity relationship of 1H-indole-2,3-dione derivatives

Article abstract of DOI:10.1016/j.bmc.2007.05.063

New series of 5-fluoro-1H-indole-2,3-dione-3-thiosemicarbazones 2a-k and 5-fluoro-1-morpholino/piperidinomethyl-1H-indole-2,3-dione-3-thiosemicarbazones 3a-r were synthesized. The structures of the synthesized compounds were confirmed by spectral data, elemental and single crystal X-ray diffraction analysis. The new 5-fluoro-1H-indole-2,3-dione derivatives, along with previously reported 5-nitro-1H-indole-2,3-dione-3-thiosemicarbazones 2l-v, 1-morpholino/piperidinomethyl-5-nitro-1H-indole-2,3-dione-3-thiosemicarbazones 4a-l, and 5-nitro-1H-indole-2,3-dione-3-[(4-oxo-1,3-thiazolidin-2-ylidene)hydrazones] 5a-s, were evaluated for in vitro antituberculosis activity against Mycobacterium tuberculosis H37Rv. Among the tested compounds, 5-nitro-1H-indole-2,3-dione-3-thiosemicarbazones (2p, 2r, and 2s) and its 1-morpholinomethyl derivatives (4a, 4e, 4g, and 4i) exhibited significant inhibitory activity in the primary screen. The antituberculosis activity of molecules with diverse skeletons was investigated by means of the Electronic-Topological Method (ETM). Ten pharmacophores and ten anti-pharmacophores that have been found by this form the basis of the system capable of predicting the structures of potentially active compounds. The forecasting ability of the system has been tested on structures that differ from those synthesized. The probability of correct identification for active compounds was found as equal to 93% in average. To obtain the algorithmic base for the activity prediction, Artificial Neural Networks were used after the ETM (the so-called combined ETM-ANN method). As the result, only 9 pharmacophores and anti-pharmacophores were chosen as the most important ones for the activity. By this, ANNs classified correctly 94.4%, or 67 compounds from 71.

Full text of DOI:10.1016/j.bmc.2007.05.063

Bioorganic & Medicinal Chemistry 15 (2007) 5888–5904
Synthesis and structure–antituberculosis activity relationship
of 1H-indole-2,3-dione derivatives
Nilgun Karalı,a,* Aysel Gursoy,a Fatma Kandemirli,b Nathaly Shvets,c F. Betul Kaynak,d
¨
¨
¨
d
e
c,f
¨
Suheyla Ozbey, Vasyl Kovalishyn and Anatholy Dimoglo
¨
aIstanbul University, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, 34116 Beyazıt, Istanbul, Turkey
bKocaeli University, Department of Chemistry, 41100 Kocaeli, Turkey
cGebze Institute of Technology, PK-141, 41400 Gebze, Kocaeli, Turkey
dHacettepe University, Department of Physics Engineering, 06532 Beytepe, Ankara, Turkey
eInstitute of Bioorganic Chemistry & Petrochemistry, 02660 Kyiv, Ukraine
fInstitute of Applied Physics, Academy of Sciences of Moldova, MD-2028 Kishinev, Republic of Moldova
Received 11 December 2006; revised 22 May 2007; accepted 30 May 2007
Available online 2 June 2007
Abstract—New series of 5-fluoro-1H-indole-2,3-dione-3-thiosemicarbazones 2a–k and 5-fluoro-1-morpholino/piperidinomethyl-1H-
indole-2,3-dione-3-thiosemicarbazones 3a–r were synthesized. The structures of the synthesized compounds were confirmed by spec-
tral data, elemental and single crystal X-ray diffraction analysis. The new 5-fluoro-1H-indole-2,3-dione derivatives, along with pre-
viously reported 5-nitro-1H-indole-2,3-dione-3-thiosemicarbazones 2l–v, 1-morpholino/piperidinomethyl-5-nitro-1H-indole-2,3-
dione-3-thiosemicarbazones 4a–l, and 5-nitro-1H-indole-2,3-dione-3-[(4-oxo-1,3-thiazolidin-2-ylidene)hydrazones] 5a–s, were evalu-
ated for in vitro antituberculosis activity against Mycobacterium tuberculosis H37Rv. Among the tested compounds, 5-nitro-1H-
indole-2,3-dione-3-thiosemicarbazones (2p, 2r, and 2s) and its 1-morpholinomethyl derivatives (4a, 4e, 4g, and 4i) exhibited signif-
icant inhibitory activity in the primary screen. The antituberculosis activity of molecules with diverse skeletons was investigated by
means of the Electronic-Topological Method (ETM). Ten pharmacophores and ten anti-pharmacophores that have been found by
this form the basis of the system capable of predicting the structures of potentially active compounds. The forecasting ability of the
system has been tested on structures that differ from those synthesized. The probability of correct identification for active com-
pounds was found as equal to 93% in average. To obtain the algorithmic base for the activity prediction, Artificial Neural Networks
were used after the ETM (the so-called combined ETM–ANN method). As the result, only 9 pharmacophores and anti-pharmaco-
phores were chosen as the most important ones for the activity. By this, ANNs classified correctly 94.4%, or 67 compounds from 71.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
regimens currently used to treat the disease, in conjunc-
tion with the advent of the AIDS epidemic and the
increased mobility of human populations, has led to
the emergence of numerous multidrug-resistant Myco-
bacterium tuberculosis strain.3 Resistance to frontline
therapeutics, most notably INH and rifampicin, results
in treatment of patients with second-line agents. Among
these second tier drugs for the treatment of multidrug-
resistant TB, thiacetazone (Fig. 1) has been widely used
in Africa and South America. It has been found to be a
very effective drug for the treatment of TB when admin-
istered in combination with INH.4
Tuberculosis (TB) is a contagious disease with high mor-
tality worldwide. The statistics indicate that 3 million
people worldwide die annually from complication of
TB1 and there are estimated 8 million of new cases each
year, 95% of which occur in developing countries.2
Drugs such as isoniazid (INH) and rifampicin have his-
torically been successful in the treatment of TB infec-
tions. In recent history, however, poor compliance
with the prolonged and complicated chemotherapeutic
Recently, a number of thiacetazone derivatives have
been synthesized and described for their activity
against M. tuberculosis, M. avium, and other myco-
bacterial species.5–7 Results indicate that some
Keywords: 1H-Indole-2,3-diones; Antituberculosis activity; Structure–
activity relationships; Electronic-topological method (ETM).
*
Corresponding author. Fax: +90 212 440 0252; e-mail: karalin@
istanbul.edu.tr
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.05.063

N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
5889
N
NH2
R
NH
NH2
Ar
N
CH3
N
S
S
O
NH
I
II
H2N
NH
NH2
S
N
N
NH
S
N
N
O
III
IV
H3C
Figure 1. Structures of thiacetazone (I), S-alkylisothiosemicarbazones (II), SRI-286 (III), and methisazone (IV).
S-alkylisothiosemicarbazones and SRI-286 (Fig. 1)
can be useful in the therapy and prophylaxis of
mycobacteria infections and can represent a template
for the development of novel antimycobacterial drugs.
As known, the better is the description of molecules in
terms of structural parameters related to the activity in
view, the better are results on pattern recognition and
separation of molecules by their activities. The ETM is
capable of taking into account any individual properties
of atoms and bonds, and this fact may be crucial for
revealing details of interactions between a biological
receptor and an active molecule. The ETM was used
here for the structure–anti-TB activity study in the series
of molecules with quite diverse skeletons.
1H-Indole-2,3-dione is an endogenous compound
identified in humans and its effect has been studied in
a variety of systems. Biological properties of 1H-in-
dole-2,3-dione include a range of actions in the brain
and offer protection against certain types of infections.
Methisazone (Fig. 1) plays an important role as prophy-
lactic agent against several viral diseases.8
2. Results and discussion
2.1. Chemistry
In recent years, Schiff and Mannich bases of 1H-indole-
2,3-diones are reported to exhibit broad-spectrum che-
motherapeutic properties such as antiviral,9–11 anti-
TB,12,13 antifungal, and antibacterial activities.14,15
Investigation of the structure–activity relationships in
1H-indole-2,3-dione derivatives revealed that 5-haloge-
nation,12–15 N-alkylation,16,17 N-Mannich base,13–15
and 3-thiosemicarbazone formation10,11,14 were effective
in causing a marked rise in activity against various bac-
teria, fungi, and virus. Moreover, cyclization of 1H-in-
dole-2,3-diones to 4-thiazoline,12 4-thiazolidinone,18
and pyridazinoindole17 was efficient in increasing anti-
microbial activity.
In this study, 5-fluoro-1H-indole-2,3-dione (1a) reacted
with different N-substituted thiosemicarbazides in
ethanol containing a catalytic amount of sulfuric acid
to give the corresponding 5-fluoro-1H-indole-2,
3-dione-3-thiosemicarbazones (2a–k). 5-Fluoro-1-mor-
pholino/piperidinomethyl-1H-indole-2,3-dione-3-thiose-
micarbazones (3a–r) were synthesized from the
consecutive treatment of 2a–k with formaldehyde
solution and morpholine or piperidine (Scheme 1).
The structures of the synthesized compounds were con-
firmed by elemental analyses and spectral (IR, 1H
NMR, 13C NMR-APT, HETCOR-2D, and LCMS-
APCI) data. The IR spectra of 2 and 3 showed absorp-
tion bands in the 3371–3171, 1694–1687, and 1154–
1130 cmÀ1, and 3309–3211, 1701–1686 and 1158–
1130 cmÀ1 regions resulting from the NH, C@O, and
C@S functions, respectively. 13C NMR-APT of 2d, 3g,
and 3h, and HETCOR spectra of 2b, 2e, 3c, and 3d sup-
ported the IR findings and displayed signals at d 162.06–
163.13, 138.87–140.34 and 177.06–177.31 ppm which
showed that there was no 13C– 1H connection attributed
Because of all the said and as a continuation of our
works on 1H-indole-2,3-dione derivatives, we have syn-
thesized 5-fluoro-1H-indole-2,3-dione-3-thiosemicarba-
zone derivatives, in order to obtain new and more
potent anti-TB compounds. These new 5-fluoro-1H-in-
dole-2,3-dione derivatives along with previously re-
ported 5-nitro-1H-indole-2,3-dione derivatives19,20 were
evaluated for in vitro anti-TB activity against M. tuber-
culosis H37Rv.
Previously, structure–activity relationships (SAR) were
studied for a series of novel 5-aryl-2-thio-1,3,4-oxa-
diazo21 and 2,5-disubstituted-1,3,4-thiadiazoles22 which
were synthesized and screened for in vitro anti-TB
activity against M. tuberculosis H37Rv. This system-
atic SAR study was performed through the applica-
tion of the ETM approach to the series of
compounds with experimentally measured anti-TB
activity.
1
to the C@O, C@N, and C@S functions. The H NMR
spectra of 2a–k displayed the NH protons of the thio-
semicarbazone moiety (d 8.87–11.09 and 12.45–
12.89 ppm) and the indole NH proton (d 11.09–
11.30 ppm) as three separate signals. The indole C7,
C6, and C4 protons appeared as double doublets, double
triplets, and double doublets with the 1H– 1H and
1H–19F connections at d 6.78–6.95, 7.03–7.24, and

5890
N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
O
3c, and 3d were observed ethyl CH2 protons at
3.66 ppm, ethyl CH2 carbons were observed together
R1
with DMSO in the 13C NMR spectra. HETCOR spectra
O
1
of 2b, 3c, and 3d supported the H and 13C NMR find-
ings. Protons of the cyclohexyl moiety in the H NMR
1
N
H
spectra of 2e and 3i showed as the axial and equivatorial
protons on the basis of HETCOR. The indole carbons
in the spectra of 2 and 3 which explicitly showed the
13C–19F connections were observed as separate doublets
with evident connection constant. LCMS-APCI of 2b,
2d, 2f, 2g, 2h, 2j, 3g, 3m, 3n, and 3q chosen as prototypes
displayed molecular ions with different intensities. The
major fragmentation pathway involved concomitant
breaking of the NH–CS bond yielding fragments [m/z
178 (À) and 180 (+)] characteristic for the 5-fluoro-
1H-indole-2,3-dione-3-hydrazone moiety.
1a
1b
R1= F
R1= NO2
NH2
HN
i
S
R2 NH
NH
R2
S
NH
N
2.2. X-ray analysis
R1
The X-ray structures of 3e and 3f were determined in or-
der to confirm the assigned structures and to establish
conformations of the molecules. Relevant crystal data
and details of the structure determinations are given in
Table 1, and selected geometric parameters are given
in Table 2. ORTEP drawings of structures with atomic
numbering are shown in Figures 2 and 3. As expected,
the indole moiety is nearly planar in both molecules.
The maximum displacements from the least-squares
plane for all nine atoms contained in the ring are
O
N
H
2a-k
2l-v
R1= F
R1= NO2
O
ii
R3
iii
R2
N
S
NH
˚
R2
À0.028(3) for C2 (3e) and 0.007(3) A for C8 (3f). In
S
3e, the morpholine ring is taken in a chair shape having
spherical
N
N
˚
Q = 0.640(3) A,
polar
set
values23
NH
N
R1
h = 175.6(3)ꢁ, and u = 3(4)ꢁ. Atoms O2 and N2 are dis-
placed from the C10/C11/C12/C13 mean plane by
R1
O
˚
O
0.701(2) and À0.788(2) A, respectively. Similarly, in 3f
N
H
the piperidine (N2/C10/C11/C12/C13/C14) ring adopts
the chair conformation. The puckering parameters of
N
5a-s
R1= NO2 R3= H, CH3
˚
this ring are Q = 0.581(4)A, h = 0.0(4)ꢁ, and
N
u = 63(12)ꢁ. N2 and C12 atoms are displaced from the
C10/C11/C13/C14 mean plane by À0.672(2) and
X
˚
0.677(4) A, respectively.
3a-r
4a-l
R1= F
R1= NO2
As can be seen from Figures 2 and 3, the molecules are
not planar. The dihedral angles between the morpholine
or piperidine rings and indole–thiosemicarbazone sys-
tem in 3e and 3f are 86.2(1) and 82.1(1)ꢁ, respectively.
The approximate planarity of indole–thiosemicarbazone
system is due to the intramolecular hydrogen bonds,
existence in the molecules. In both molecules, the intra-
and intermolecular N–HÁ Á ÁO, C–HÁ Á ÁO, and N–HÁ Á ÁN
hydrogen bonds are observed (Table 3). In 3e, dimeric
units are formed through the C–HÁ Á ÁF interactions of
the phenyl rings and dimers of molecules linked through
the C–HÁ Á ÁO hydrogen bonds which extend along the c
axis (Fig. 4). In 3f, N–HÁ Á ÁS intermolecular contacts
also generate infinite chains along c axis (Fig. 5).
R2= CH3, C2H5, CH2-CH=CH2, n-C4H9, cyclo-C6H11, C6H5CH2,
C6H5, 4-CH3C6H4, 4-ClC6H4, 4-FC6H4, 4-NO2C6H4
X= CH2, O
Scheme 1. Reagents and conditions: (i) EtOH, reflux, 5 h; (ii)
morpholine or piperidine, 37% HCHO, absolute EtOH, stir, rt, 10 h
or reflux, 4 h; (iii) ethyl bromoacetate or ethyl 2-bromopropionate,
anhydrous sodium acetate, absolute EtOH, reflux, 4 h.
7.44–7.65 ppm on the basis of HETCOR data. Observa-
tion of only two NH signals assigned to the thiosemicarb-
azone moiety (d 8.92–10.93 and 12.36–12.65 ppm) and
of a singlet due to N–CH2–N function (d 4.46–
4.51 ppm and d 61.81–62.51 ppm) and of signals attrib-
In addition, there are three intermolecular C–HÁ Á Áp
interactions involving the indole heterocycle that is pres-
ent in both molecules. In molecule 3e, one of the C–
HÁ Á Áp interactions happens between pyrrole ring and
N5 atom, and two others are related with the phenyl
ring of the indole heterocycle and the hydrogen atoms
uted to morpholine or piperidine ring in the 1H and 13
C
NMR spectra of 3 provided support for N-Mannich
base formation. While in the H NMR spectra of 2b,
1

N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
Table 1. Crystal data and structure refinement parameters for 3e and 3f
5891
Compound
Formula
3e
C
3f
17H20FN5O2S
C18H22FN5OS
375.47
Model molecular weight
Crystal system
Space group
377.44
Triclinic
P 1
Monoclinic
P21/c
Unit cell dimensions
˚
a (A)
˚
6.3291(8)
10.0080(13)
9.1286(5)
19.0680(9)
11.4206(5)
—
b (A)
˚
c (A)
14.4884(19)
82.498(11)
a (ꢁ)
b (ꢁ)
c (ꢁ)
85.569(11)
90.742(11)
106.955(5)
—
3
˚
V (A )
906.9(2)
1.382
1901.51(16)
1.312
Dcalc (g cmÀ3
)
Z
2
293(2)
4
293(2)
Temperature (ꢁK)
Crystal size (mm)
Crystal color
0.04 · 0.10 · 0.10
Yellow
0.10 · 0.10 · 0.60
Orange
Prism
Crystal habit
˚
Prism
0.71073
k (A)
0.71073
1.197
69.4
l (mmÀ1
)
0.21
74.7
2hmax (ꢁ)
Index ranges
À9 6 h 6 7; À11 6 k 6 14; À20 6 l 6 21
À14 6 h 6 14; À30 6 k 6 30; À15 6 l 6 17
Reflections collected/used in refinement
No. of refined parameters
R/Rw values
GOF on F2
8742/5613
243
74670/7340
243
0.0646/0.1581
0.884
0.000
0.0776/0.2202
1.1
0.000
Final shift
À3
˚
(Dq)min , (Dq)max (e A
)
0.288/À0.275
0.493/À0.329
of C12. The shortest distances between the mentioned
atoms and centroids of the benzene or pyrrole rings of
˚
indole range from 3.0 to 3.31 A, with centroid angles
˚
Table 2. Selected bond lengths and angles (ꢁ, A) for 3e and 3f
3e
3f
ranging from 86.0 to 151.5ꢁ. In molecule 3f, there are
also two intermolecular C–HÁ Á Áp interactions between
the pyrrole and either benzene rings of the indole hetero-
cycle or C18 atom of allyl group of symmetry related
molecule. The other C–HÁ Á Áp interaction is between
C16 atom and the benzene ring of indole. The shortest
distances between the mentioned atoms and centroids
of benzene and pyrrole rings of indole range from 3.08
to 3.18 A, with centroid angles ranging from 129.7ꢁ
to 148.5ꢁ, both ranges being observed for classical
C–HÁ Á Áp interactions.24
C1–O1
C1–N1
1.396(3)
1.313(3)
1.612(4)
1.163(3)
1.575(4)
1.567(4)
1.630(4)
1.253(3)
1.478(3)
1.423(3)
1.302(4)
1.481(4)
1.357(4)
1.311(4)
1.532(4)
1.245(3)
1.451(4)
1.864(3)
1.645(5)
1.327(3)
1.238(5)
1.492(3)
116.6(2)
111.9(2)
116.8(2)
116.7(3)
119.8(3)
160.0(2)
118.3(3)
À1.0(3)
1.1(4)
C1–O1
C1–N1
1.213(3)
1.367(3)
1.509(4)
1.291(3)
1.448(3)
1.355(3)
1.409(3)
1.436(3)
1.459(3)
1.463(3)
1.486(4)
1.503(5)
1.516(5)
1.500(4)
1.471(3)
1.323(3)
1.369(3)
1.669(3)
1.440(4)
1.456(3)
1.201(5)
1.341(3)
110.7(2)
124.6(2)
121.6(2)
124.9(2)
126.7(2)
166.1(2)
106.3(3)
À1.0(3)
À1.4(4)
C1–C2
C2–N3
C1–C2
C2–N3
C2–C3
C5–F1
C2–C3
C5–F1
C8–N1
C9–N2
C8–N1
C9–N2
C9–N1
C10–N2
C9–N1
C10–N2
˚
C10–C11
C11–O2
C12–O2
C10–C11
C11–C12
C12–C13
C13–C14
C14–N2
2.3. Primary antituberculosis activity
C12–C13
C13–N2
C14–N4
C15–N5
C15–N4
The new 5-fluoro-1H-indole-2,3-diones 2a–k and
3a–r, along with the previously synthesized19,20
5-nitro-1H-indole-2,3-dione-3-thiosemicarbazones 2l–v,
1-morpholino/piperidinomethyl-5-nitro-1H-indole-2,3-
dione-3-thiosemicarbazones 4a–l, and 5-nitro-1H-indole-
2,3-dione-3-[(4-oxo-1,3-thiazolidin-2-ylidene)hydrazones]
5a–s, were evaluated at 6.25 lg/mL for in vitro anti-
TB activity against Mycobacterium tuberculosis
H37Rv (ATCC 27294) in BACTEC 12B medium
using a broth microdilution assay, the Microplate Ala-
mar Blue Assay (MABA). The fluorescence exhibiting
compounds 2a, 2f, 2l, 3h, and 4a were tested in the
BACTEC 460 radiometric system in BACTEC 12B
medium. The primary anti-TB screening was
performed in accordance with the protocol of the
C14–N5
C14–S1
C15–S1
C16–C17
C16–N5
C15–C16
C15–N5
C16–C17
N3–N4
C17–C18
N3–N4
C1–N1–C8
C1–N1–C9
C14–N4–N3
C15–N5–C14
N3–C2–C1
N1–C9–N2–C10
N2–C9–N1–C1
C1–C2–N3–N4
S1–C14–N5–C15
C1–N1–C8
C1–N1–C9
N3–N4–C15
C15–N5–C16
N3–C2–C1
N1–C9–N2–C10
N2–C9–N1–C1
C1–C2–N3–N4
S1–C15–N5–C16

5892
N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
Figure 2. Molecular structure of 3e showing the atom labeling scheme. Displacement ellipsoids are drawn as 30% probability level.
Figure 3. Molecular structure of 3f showing the atom labeling scheme. Displacement ellipsoids are drawn as 30% probability level.
˚
Table 3. Hydrogen bonding and short contact geometry for 3e and 3f (A, ꢁ)
Compound
D–HÁ Á ÁA
D–H
HÁ Á ÁA
1.88(3)
DÁ Á ÁA
D–HÁ Á ÁA
130(2)
3e
N4–H4NÁ Á ÁO1
C4–H4Á Á ÁF1i
C6-H6Á Á ÁO2ii
C9-H9BÁ Á ÁO1
0.98(3)
0.93
0.93
2.610(3)
3.355(4)
2.932(4)
2.778(3)
2.43
2.32
177
123
0.97
2.27
111
3f
N4–H4Á Á ÁO1
N5–H5Á Á ÁN3
N5–H5Á Á ÁS1iii
C16–H16BÁ Á ÁS1
0.97(2)
1.02(3)
1.02(3)
0.97
1.92(2)
2.28(3)
2.49(3)
2.71
2.727(3)
2.648(3)
3.433(2)
3.139(3)
139(2)
100(2)
154(2)
107
Symmetry codes: (i) 2 À x,1 À y,1 À z; (ii) 3 À x,1 À y, À z; (iii) x,1/2 À y,1/2 + z.
Tuberculosis Antimicrobial Acquisition and Coordi-
nating Facility (TAACF) Southern Research Insti-
tute.25 Rifampin was used as control drug in the
tests. Activity of the compound was estimated as the
percent inhibition (Table 4). Compounds demonstrat-
ing at least 90% inhibition in the primary screen were
retested against M. tuberculosis H37Rv to determine
the actual minimum inhibitory concentration (MIC)
in the MABA. The MIC was defined as the lowest
concentration effecting a reduction in fluorescence of
90% relative to controls. Concurrent with the determi-
nation of MICs, compounds were tested for cytotoxic-
ity (IC50) in VERO cells at concentrations 10· the
MIC for M. tuberculosis H37Rv. However, as can
be seen from Table 5, the whole of the tested com-
pounds had high cytotoxicity.
When these data are examined, it is observed that in 1H-
indole-2,3-dione-3-thiosemicarbazones and its N-Man-
nich bases, most of R1-nitro substituted derivatives were
more active than R1-fluoro substituted derivatives. In
fact, 5-nitro-1H-indole-2,3-dione-3-thiosemicarbazones
(2p, 2r, 2s, and 2t) and 1-morpholinomethyl-5-nitro-
1H-indole-2,3-dione-3-thiosemicarbazones (4a, 4e–g,
4i, and 4j) exhibited significant inhibitory activity
(MIC P 75%), whereas 5-fluoro-1H-indole-2,3-dione-
3-thiosemicarbazones (2e and 2g–k) and its N-Mannich
bases (3g, 3i, and 3l–r) showed average activity

N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
5893
at a MIC value of 6.25 lg/mL, compounds 2p, 2s, 4a,
4g, and 4i exhibited anti-TB activity showing 90%,
90%, 96%, 93%, and 92% inhibition, respectively, at a
MIC value of >6.25 lg/mL.
Comparison of the thiosemicarbazones (2a–v), its N-
Mannich bases (3a–r and 4a–l) and thiazolidinones
(5a–s) revealed that the anti-TB activity was increased
due to elongation of the alkyl chain. Generally,
replacement of the alkyl in the R2 position with cyclo-
hexyl or (non)substituted phenyl has been found to
yield more active compounds (except for 4a). More-
over, the 1-morpholinomethyl derivatives were more
active than the corresponding the 1-piperidinomethyl
derivatives. The presence of the morpholine ring in
some R1-nitro substituted N-Mannich bases also seems
to have a significant impact on the resultant anti-TB
activity. Compound 4a, a morpholine derivative incor-
porating a methyl group at R2, was found to be one of
the most active inhibitors, whereas ligand 2l was inac-
tive. It is interesting to note that, for the R2-phenyl
substituted N-Mannich derivatives 4g and 4h, the mor-
pholine derivative 4g was substantially more active
than both the ligand 2q and the piperidine derivative
4h, whereas the activity of 4h was virtually unchanged
when compared with the ligand 2q. There were only
minor variations in activities of other N-Mannich
bases.
In comparison with compounds 2l–v, none of the 4-thia-
zolidinone derivatives (5) were active against M. tuber-
culosis. These modifications of parent molecules led to
a complete loss of the anti-TB activity. Further modifi-
cations of the structure of 2 may indeed lead to an im-
proved antimycobacterial compound with more
potency and less cytotoxicity.
Figure 4. Crystal packing of 3e showing the dimers and the dimer
interactions. Dotted lines show the intermolecular interactions.
(31% 6 MIC 6 70%). 5-Nitro-1H-indole-2,3-dione-3-thio-
semicarbazones (2p, 2r, and 2s) and 1-morpholinometh-
yl-5-nitro-1H-indole-2,3-dione-3-thiosemicarbazones (4a,
4e, 4g, and 4i), demonstrating at least 90% inhibition
in the primary screen, are actually the most potent
inhibitors of M. tuberculosis growth described in this
study. While compounds 2r and 4e displayed anti-TB
activity showing 93% and 90% inhibition, respectively,
Thus, the complex of pharmacophores and anti-phar-
macophores, taken as a whole, plays an important role
not only in the prediction of activities, but in the search
for new drugs as well. The set of activity/inactivity
Figure 5. Crystal packing of 3f. Dotted lines show the intermolecular interactions.

5894
N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
Table 4. Primary in vitro antituberculosis activity screening results of 2–5 against M. tuberculosis H37Rv at 6.25 lg/mL
Compound
R1
R2
R3
X
GI (%)a
Compound
R1
R2
R3
X
GI (%)a
2a
2b
2c
2d
2e
2f
F
CH3
C2H5
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
0
8
3o
3p
3q
3r
4a
4b
4c
4d
4e
4f
F
4-ClC6H4
4-ClC6H4
4-FC6H4
4-FC6H4
CH3
—
—
O
58
63
48
38
96
19
22
30
90
87
93
31
92
88
65
54
0
F
F
CH2
O
F
CH2–CH@CH2
n-C4H9
—
—
30
16
39
0
F
—
—
F
F
CH2
O
F
cycl-C6H11
C6H5CH2
C6H5
—
—
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
—
—
F
CH2–CH@CH2
CH2–CH@CH2
n-C4H9
O
2g
2h
2i
F
—
—
61
36
37
58
31
0
—
—
CH2
O
F
4-CH3C6H4
4-ClC6H4
4-FC6H4
4-NO2C6H4
CH3
F
—
—
cycl-C6H11
cycl-C6H11
C6H5
—
—
O
2j
F
CH2
O
2k
2l
F
—
—
4g
4h
4i
—
—
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
NO2
F
C6H5
CH2
O
2m
2n
2o
2p
2q
2r
2s
2t
C2H5
CH2–CH@CH2
n-C4H9
—
—
0
4-CH3C6H4
4-CH3C6H4
4-ClC6H4
4-ClC6H4
CH3
—
—
25
42
90
33
93
90
75
24
2
4j
CH2
O
—
—
4k
4l
—
—
cycl-C6H11
C6H5
CH2
—
—
—
5a
5b
5c
5d
5e
5f
H
4-CH3C6H4
4-BrC6H4
4-ClC6H4
4-FC6H4
4-NO2C6H4
CH3
C2H5
CH2–CH@CH2
n-C4H9
H
—
—
0
—
—
H
2
H
—
—
0
2u
2v
3a
3b
3c
3d
3e
3f
—
—
cycl-C6H11
C6H5
H
0
H
—
—
0
O
12
20
1
5g
5h
5i
4-CH3C6H4
4-BrC6H4
4-ClC6H4
4-FC6H4
CH2–CH@CH2
n-C4H9
H
17
0
F
CH3
C2H5
C2H5
CH2
O
H
—
—
F
H
0
F
CH2
O
2
5j
H
—
—
0
F
CH2–CH@CH2
CH2–CH@CH2
n-C4H9
24
20
34
0
5k
5l
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
3
F
CH2
O
—
—
10
28
9
3g
3h
3i
F
5m
5n
5o
5p
5q
5r
5s
cycl-C6H11
C6H5
F
n-C4H9
CH2
O
—
—
F
cycl-C6H11
C6H5CH2
C6H5CH2
C6H5
33
0
4-CH3C6H4
4-BrC6H4
4-ClC6H4
4-FC6H4
4-NO2C6H4
0
3j
F
O
—
—
3
3k
3l
F
CH2
O
3
0
F
56
45
51
—
—
5
3m
3n
F
C6H5
4-CH3C6H4
CH2
CH2
9
F
a Growth inhibition of virulent H37Rv strain of M. tuberculosis.
Table 5. The GI, MIC, and IC50values of 2p, 2r, 2s, 4a, 4e, 4g, and 4i
against M. tuberculosis H37Rv
Table 4. A preliminary analysis of the molecules in-
cluded optimization of their structures. Further the opti-
mized structures were used in quantum-chemical
calculations (AM1-method)26 and their results provided
various input data in the course of the ETM application.
With the aim of more detailed SAR analysis, all com-
pounds were separated into four groups. The first series
included 10 compounds with high antituberculosis activ-
ity (MIC P 75%). The second series contained 20 mole-
cules with average activity (31% 6 MIC 6 70%). The
third and fourth sets of low-active (5% 6 MIC 6 30%)
and inactive compounds (MIC < 4%) included 19 and
22 molecules, respectively. The ETM has been applied
to each set separately to identify the pharmacophores
and anti-pharmacophores.
c
Compound
GIa (%)
MICb (lg/mL)
IC50 (lg/mL)
2p
2r
90
93
90
96
90
93
92
97
>6.25
6.25
4.74
2.45
2.59
2s
4a
>6.25
>6.25
6.25
7.61
d
4e
4g
>6.25
>6.25
0.125
2.8
17
4i
Rifampin
>100
a Growth Inhibition of virulent H37Rv strain of M. tuberculosis.
b The actual minimum inhibitory concentration.
c The cytotoxicity.
d Insoluble in DMSO at 1 mg/mL. Unable to determine cytotoxicity.
fragments obtained in this study constitutes a basis for
the development of a system for the anti-TB activity
prediction.
2.5. The search for pharmacophores (Ph) and anti-
pharmacophores (APh) by using ETM
The ETM analyzes molecules represented by matrices
which are named electronic-topological matrices of con-
junction (ETMC), because they are formed of electronic
and 3D-topology data.27–30 Since details of the ETM
can be found in literature,31,32 we give here only the
2.4. Data sets
Compounds under study (71 in total, see Scheme) are
shown along with their common structural skeletons in

N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
5895
most distinguished properties of the ETM relative to
other methods used in SAR studies. When the system
for the antituberculosis activity prognostication has
been formed, compounds that represent each series
and possess the highest levels of inhibitory activity and
selectivity in the given group have been taken as tem-
plates to be compared with the rest of compounds as
ETMCs.
revealed as well. They are structural fragments that can
be found in inactive compounds only. As an example,
skeletons of template compounds and submatrices of
their ETMCs (named as ETSCs, for short) are shown
in Figures 6 and 7, which correspond to some of the
pharmacophores and anti-pharmacophores.
As the first step, the compounds from the 1st and 4th
groups were compared (i.e., high active and inactive
ones). The ETM-calculations detected an activity fea-
ture 1 (named as a Ph1 pharmacophore), that is com-
mon for all high-active compounds. Ph1 and the
corresponding ETSC (ETSCPh1) which has been calcu-
lated relative to the template compound 4a are shown
in Figure 6a. The ETSCPh1 is of the order 7 · 7, and only
From the ETM application to the series mentioned, a
number of pharmacophores (Phi) have been calculated,
which form the base of the system for the antituberculo-
sis activity prognostication. Besides the pharmaco-
phores, molecular fragments responsible for the
activity loss (or anti-pharmacophores—APhi) have been
Figure 6. Sub-matrices and corresponding structures of the Ph1–Ph3 pharmacophores, obtained relative to the template (active) compounds 4a, 4k,
and 4d.

5896
N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
specific electronic-topological parameters (here, atomic
charges and inter-atomic 3D-distances) that are con-
tained in the ETSCPh1. As seen from Figure 6a, Ph1 in-
cludes properties of three main parts of active
compounds that is indole, thiosemicarbazone, and
morpholine.
20
26
5
Analogous ETSCs were obtained when compounds
from other groups had been compared. Thus, the phar-
macophore Ph2 (Fig. 6b) was found for the template 4k
from the comparison of compounds belonging to the
2nd and 4th groups. This pharmacophore is present in
17 active compounds (n2 = 17) and one inactive com-
pound (n4 = 1). Thus, the probability of the Ph2 pres-
ence in the set of active compounds, Pa, was estimated
as 0.90. Less informative pharmacophore was calculated
from the comparison of the 3rd and 4th groups of com-
pounds (Fig. 6c, template compound 4d).
12
17
3
16
1
5l
15
23
a
b
C1
N5
N15
C16 H23
O26
0.07 1.00 5.49 5.77 6.27 3.51
-0.31 6.51 6.55 7.36 2.27
-0.30 1.11 0.88 8.76
0.14 2.05 8.76
A ‘break of activity’, or antipharmacophore, was
calculated relative to the inactive template compound
5l. This antipharmacophore (APh1) was detected in 21
inactive compound (n4 = 21) and was not found at all
in the structures of high active compounds (n1 = 0).
The APh1 and its corresponding ETSCAPh1 calculated
relative to the template compound 5l are shown in
Figure 7a.
n4/n1= 21/0
Pα = 0.96
0.25 9.60
-0.30
C1
C12 C16 H23 O26
3.43 5.77 6.27 3.51
0.07
-
2.34 3.23 6.53
0.02
0.14 2.05 8.76
0.25 9.60
-0.29
n4/n2= 18/1
Pα = 0.90
The probability Pia of the APh1 presence in the set of
inactive compounds was estimated as 0.96. In Figures
7b and c, those antipharmacophores are shown that
have been found from the 2nd and 3rd groups of com-
pounds comparison.
c
C3
N5
N15
C17
C20
-0.35 2.40 8.00 9.66 10.01
-0.31 6.51 7.74 7.90
-0.30 2.51 3.72
2.6. Combined ETM–artificial neural networks approach
(ETM–ANN)
n4/n3= 12/1
-0.20 1.44
The ETM data (initial matrices and the fragments
found) can be used in the NN applications with the
aim of obtaining the algorithmic base for the activity
prediction. Here, some important observations as to
their joint use are presented.
0.00
Pα = 0.87
Figure 7. Sub-matrices and corresponding structures of the APh1–
APh3 anti-pharmacophores, obtained relative to the template (inac-
tive) compound 5l.
its upper triangle is given because of the symmetry of
bonds. There are two chemically bonded atoms N9–
C11 in Ph1, and the Wiberg’s index (electronic popula-
tion on bond) is given for this bond as the WN9–C11
element of the submatrix (WN9–C11 = 0.90). All other
off-diagonal elements of the ETSCPh1 are 3D-distances
for the corresponding pairs of atoms in the Ph1. This
pharmacophore was found in 9 active compounds
(n1 = 9), while for inactive compounds it was not found
(n4 = 0). In this way, the probability Pa of the Ph1 pres-
ence in the set of high-active compounds was estimated
as 0.91: Pa (Phi) = (n1+1)/(n1 + n4+2), where n1, n4 are
numbers of high active/inactive compounds, respec-
tively, that contain the Ph. In the case of an anti-phar-
macophore (APhi), the corresponding probability Pia is
calculated in the same way, but (n4 + 1) is to be used
at place of (n1 + 1) in the equation.
As known, Artificial Neural Networks (ANNs) can be
divided into two main categories related to learning
algorithms (i.e., supervised and unsupervised learning).
First, the Kohonen neural network33 (SOM, self-orga-
nized map) using an unsupervised learning algorithm
has been applied to realize the projection of initial data
(ETMCs) on the SOM units. Next, the supervised algo-
rithm has been used for an feed forward neural network
(FFNN) with backpropagation34 to calculate dependen-
cies between input and output variables being weights of
the molecular fragments and activity values, corre-
spondingly. The supervised learning was performed
using a variant of FFNN known as the Associative Neu-
ral Network (ASNN).35 Below, we briefly summarize the
principles of this approach, while the detailed descrip-
tion of the algorithm can be found in the literature.36
It is clear that ETMCs cannot be used in a straightfor-
ward manner as the input for ANNs. Consequently,
the information contained in ETMCs should be rear-
Seven atoms of the Ph1 pharmacophore belong to differ-
ent parts of the template molecule. All these atoms have

N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
5897
ranged somehow in order to serve as input vectors of
120
100
80
60
40
20
0
equal dimensionality for an ANN. The training of the
Kohonen’s SOM has the property of mapping input vec-
tors from the n-dimensional space (n >> 2), which pos-
sess similar properties to the same or nearby neurons
in the two-dimensional space. Therefore, by considering
all input vectors projected to the same output neuron, it
is possible to determine clusters of vectors that have sim-
ilar properties in the n-dimensional space.
0
20
40
60
80
100
120
In short, the principal idea of the combined approach is
to determine the weights of fragments represented by
ETSCs and, afterwards, to use these weights as descrip-
tors (WDs) for the ASNNs training. To do this, the frag-
ments are being projected on the Kohonen’s maps that
correspond to their initial ETMCs. In such way the de-
gree of each fragment’s presence in a molecule can be
determined.
Observed
Figure 8. Plot of predicted antituberculosis activity versus actual
observed values.
The last step includes application of the pruning meth-
ods39,40 that aim in a set of the most relevant ETMC
fragments selection.
The first step is a procedure that can be called as ‘tri-
ples calculation’. That is, data elements in the initial
set are triples (d1, d2, d3), where d1 and d2 are
charges for a pair of atoms and d3 is a connection
between them. The values of di, i = 1, 2, 3, are taken
from the ETMCs. The total number of triples equals
to the amount of all two-atomic connections taken
from all ETMCs. Second step is the Kohonen’s net-
work (SOM) initialization and training (detailed
description of a Kohonen’s network can be found
in37). The approximate number of elements in our
2.7. ETM–ANN approach application
To reflect the realistic internal structure of the data, all
compounds were separated into two main series. The
first one included 49 active compounds, and the second
series included 22 inactive compounds. For the data, 339
fragments were selected. The first step consisted of using
the LOO cross-validation procedure for total set of com-
pound. The ASNNs recognized correctly 88.7%, or 63
from 71 compounds. Next, the importance of the de-
tected fragments for the observed activity was evaluated
by using pruning methods. The most part of the ETMC
fragments (as descriptors) were detected as non-signifi-
cant and removed by the pruning algorithms. As the re-
sult, only 9 ETMC-fragments were chosen from 339
fragments in total as the most important ones. By this,
ASNN classified correctly 94.4%, or 67 compounds
from 71.
Kohonen’s map is calculated as S = k
SETM, where
*
k varies in the range of, [1.0, 2.0] and SETM is the
size of the largest ETM matrix. The third step in-
volves the calculations of pharmacophores as subma-
trices of the template ETMCs. At the fourth step, the
weight of each fragment (that is either a pharmaco-
phore or antipharmacophore) is estimated versus each
compound as the fragment’s projection error, Eij, rel-
ative to those nodes of the Kohonen’s map that were
found for its comprising ETMC. Then its weight is
taken as the inverse of its error Eij: Wij = 1 À Eij/
Emax,j. Here i is the molecule’s number, j is the frag-
ment’s number, and Emax,j is the maximal error, for
all j. A new table containing the calculated fragment
weights (descriptors) is being formed for further
ASNN training.35
The main goal of the second stage in the data analysis
was to evaluate an efficacy of ASNN in the SAR model
generation for the real values of the activity. The first
step was the LOO cross-validation procedure applied
to the total set of compounds. It has been found that
a cross-validated q2 coefficient of the ASNN predictions
was 0.68 0.01. After applying the pruning methods 8
fragments were selected from 339 in total. The cross-val-
idated q2 value was 0.77 0.01.
Briefly, the number of neurons in the input layer corre-
sponds to the number of descriptors. The hidden layer
contains five neurons. The bias neuron is presented both
on the input and hidden layers. An ensemble of M = 100
neural networks was trained. By this, the activity values
of each compound were calculated for each ASNN and
averaged over all M networks. This value was used to
calculate cross-validation coefficients.38
Thus, the obtained results indicate that application of
pruning methods provides higher prognosing ability
compared to the case when all descriptors are used for
the activity prediction. As it is seen, the approach pre-
sented in this study has shown quite satisfactory results
(Fig. 8). This fact tells in favor of workability of the
both models found. These models can be applied to
the design of new potent antituberculosis drugs.
The quality of each final model was assessed by the
leave-one-out method (LOO). By the method, each
molecule is removed from the training set, and the
remaining set is used to separate molecules into clas-
ses of activity, thereby predicting the activity of this
molecule and evaluating the quality of the decision
rule.
3. Conclusions
A new series of 5-fluoro-1H-indole-2,3-dione-3-thiose-
micarbazones and 5-fluoro-1-morpholino/piperidino-
methyl-1H-indole-2,3-dione-3-thiosemicarbazones was

5898
N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
synthesized. The structures of the synthesized com-
pounds were confirmed by spectral data, elemental and
single crystal X-ray diffraction analysis. These new
5-fluoroindol-2,3-dione derivatives and 5-nitro-1H-in-
dole-2,3-dione-3-thiosemicarbazones, as well as previ-
ously synthesized 1-morpholino/piperidinomethyl-5-
acid, the mixture was refluxed on a water bath for 5 h. The
product formed after cooling was filtered and washed
with ethanol or recrystallized from ethanol.
4.2.1. 5-Fluoro-1H-indole-2,3-dione-3-(N-methylthiose-
micarbazone) (2a). Orange crystals (97%): mp 273–
274 ꢁC; IR (KBr): t 3251 (NH), 1692 (C@O), 1148
(C@S); 1H NMR (DMSO-d6/400 MHz): d 3.08 (d,
J = 4.58 Hz, 3H, CH3), 6.92 (dd, J = 8.58, 4.18 Hz,
1H, indole C7–H), 7.18 (dt, J = 8.97, 2.97 Hz, 1H, indole
C6–H), 7.44 (dd, J = 8.13, 2.67 Hz, 1H, indole C4–H),
9.32 (q, J = 4.56 Hz, 1H, N4–H), 11.21 (s, 1H, indole
NH), 12.47 (s, 1H, N2–H); Anal. Calcd for
C10H9FN4OS (252.26): C, 47.61; H, 3.60; N, 22.21.
Found: C, 47.77; H, 3.25; N, 22.14.
nitro-1H-indole-2,3-dione-3-thiosemicarbazones
and
5-nitro-1H-indole-2,3-dione-3-[(4-oxo-1,3-thiazolidin-2-
ylidene)hydrazones], have been evaluated for the anti-
TB activity. Among the tested compounds, 5-nitro-1H-
indole-2,3-dione-3-thiosemicarbazones and 1-morpholi-
nomethyl-5-nitro-1H-indole-2,3-dione-3-thiosemicarba-
zones showed good inhibitory activity in the primary
screen. The anti-TB activity of molecules with diverse
skeletons was investigated by means of the ETM. The
structural fragments of activity and ‘breaks of activity’,
as well as their numerical parameters found by the
ETM, provide some rules, which are to be obeyed by ac-
tive compounds. The more important is that these con-
ditions can be directly used to design new active
compounds with required properties.
4.2.2. 5-Fluoro-1H-indole-2,3-dione-3-(N-ethylthiosemi-
carbazone) (2b). Dark yellow crystals (92%): mp 245–
246 ꢁC; IR (KBr): t 3325, 3173 (NH), 1687 (C@O),
1
1144 (C@S); H NMR (DMSO-d6/500 MHz): d 1.21 (t,
J = 7.17 Hz, 3H, ethyl CH3), 3.66 (p, J = 6.56 Hz, 2H,
ethyl CH2), 6.93 (dd, J = 8.70, 4.12 Hz, 1H, indole C7–
H), 7.19 (dt, J = 8.54, 2.74 Hz, 1H, indole C6–H), 7.50
(dd, J = 8.08, 2.59 Hz, 1H, indole C4–H), 9.33 (t,
J = 5.64 Hz, 1H, N4–H), 11.20 (s, 1H, indole NH),
12.45 (s, 1H, N2–H); 13C NMR (HETCOR DMSO-d6/
100 MHz): d 14.43 (ethyl CH3), 39.33–40.58 (m, ethyl
CH2, DMSO), 108.17 (d, J = 25.8 Hz, indole C4),
112.50 (d, J = 8.10 Hz, indole C7), 117.71 (d,
J = 24.1 Hz, indole C6), 121.86 (d, J = 9.20 Hz, indole
C3a), 131.41 (d, J = 3.10 Hz, indole C7a), 138.90 (indole
C3), 158.68 (d, J = 237.7 Hz, indole C5), 163.13 (indole
C@O), 177.11 (C@S); LCMS-APCI (À/+): m/z (%) 265
(MHÀ, 100). Anal. Calcd for C11H11FN4OS (266.29):
C, 49.61; H, 4.16; N, 21.04. Found: C, 49.41; H, 3.88;
N, 21.19.
All activity features detected by the ETM are character-
ized by the probabilities Pa that lie in the limits of 0.84–
0.91. Comparison of the activity features with the
‘breaks of activity’ provides a detailed explanation of
why the activity varies so much in all four series
investigated.
4. Experimental
Melting points were estimated with a Buchi 540 melting
point apparatus in open capillaries and are uncorrected.
Elemental analyses were performed on a Thermo Finn-
igan Flash EA 1112 elemental analyzer. IR spectra were
recorded on KBr disks, using a Perkin-Elmer Model
1
1600 FT-IR spectrometer. H NMR, 13C NMR APT,
4.2.3. 5-Fluoro-1H-indole-2,3-dione-3-(N-allylthiosemi-
carbazone) (2c). Shining dark yellow crystals (74%):
mp 219–220 ꢁC; IR (KBr): t 3371, 3175 (NH), 1691
and HETCOR-2D spectra were obtained on Bruker
Avance DPX 400 and VarianUNITY INOVA 500 spec-
trophotometers using DMSO-d6. Mass spectra were
determined on Finnigan TM LCQ (Advantage Max
LC/MS) or Mass-AILENT 1100 MSD instruments.
All chemicals and solvents were purchased from
Merck-Schuchardt, Aldrich, and Fluka.
1
(C@O), 1152 (C@S); H NMR (DMSO-d6/400 MHz):
d 4.27 (t, J = 5.38 Hz, 2H, allyl C1–H2), 5.16 (d,
J = 10.28 Hz, 1H, allyl C3–Hcis), 5.21 (d, J: 17.20 Hz,
1H, allyl C3–Htrans), 5.89–5.96 (m, 1H, allyl C2–H),
6.93 (dd, J = 8.55, 4.10 Hz, 1H, indole C7–H), 7.19 (dt,
J = 8.91, 2.59 Hz, 1H, indole C6–H), 7.51 (dd,
J = 8.14, 2.57 Hz, 1H, indole C4–H), 9.52 (t,
J = 5.79 Hz, 1H, N4–H), 11.22 (s, 1H, indole NH),
12.50 (s, 1H, N2–H); Anal. Calcd for C12H11FN4OS
(278.30): C, 51.79; H, 3.98; N, 20.13. Found: C, 51.64;
H, 3.61; N, 20.25.
4.1. General method for the synthesis of N-substituted
thiosemicarbazides
To a solution of hydrazine hydrate (5 mmol) in ethanol,
a suspension of an appropriate isothiocyanate (5 mmol)
in ethanol was added dropwise with vigorous stirring
and cooling in an ice bath. The mixture was allowed
to stand overnight. The crystals formed were recrystal-
lized from ethanol.
4.2.4. 5-Fluoro-1H-indole-2,3-dione-3-(N-n-butylthiosemi-
carbazone) (2d). Yellow powder (81%): mp 211–212 ꢁC;
1
IR (KBr): t 3246 (NH), 1694 (C@O), 1140 (C@S); H
NMR (DMSO-d6/500 MHz): d 0.93 (t, J = 7.32 Hz,
3H, butyl CH3), 1.33–1.37 (m, 2H, butyl C3–H2), 1.63
(p, J = 7.39 Hz, 2H, butyl C2–H2), 3.62 (q,
J = 6.91 Hz, 2H, butyl C1–H2), 6.92 (dd, J = 8.54,
4.27 Hz, 1H, indole C7–H), 7.19 (dt, J = 8.42, 2.75 Hz,
1H, indole C6–H), 7.51 (dd, J = 8.08, 2.59 Hz, 1H, in-
dole C4–H), 9.30 (t, J = 5.79 Hz, 1H, N4–H), 11.19 (s,
1H, indole NH), 12.45 (s, 1H, N2–H); 13C NMR (APT
4.2. General method for the synthesis of 5-fluoro/nitro-
1H-indole-2,3-dione-3-thiosemicarbazones (2a–v)19
A solution of N-substituted thiosemicarbazides (3.5
mmol)in ethanol (10 mL)was addedtoa solution of 5-fluo-
ro/nitro-1H-indole-2,3-dione 1a/1b (3.5 mmol) in ethanol
(20 mL). After addition of a drop of concentrated sulfuric

N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
5899
DMSO-d6/100 MHz): d 14.20 (butyl CH3), 20.06 (butyl
C3), 30.96 (butyl C2), 44.37 (butyl C1), 108.21 (d,
J = 25.7 Hz, indole C4), 112.48 (d, J = 8.09 Hz, indole
C7), 117.69 (d, J = 24.1 Hz, indole C6), 121.86 (d,
J = 9.26 Hz, indole C3a), 131.37 (d, J = 3.09 Hz, indole
C7a), 138.87 (indole C3), 158.68 (d, J = 237.6 Hz, indole
C5), 163.13 (indole C@O), 177.31 (C@S); LCMS-APCI
(À/+): m/z (%) 293 (MHÀ, 100). Anal. Calcd for
C13H15FN4OS (294.34): C, 53.05; H, 5.14; N, 19.03.
Found: C, 52.97; H, 5.01; N, 19.02.
C15H11FN4OS (314.33): C, 57.31; H, 3.53; N, 17.82.
Found: C, 57.19; H, 3.16; N, 17.73.
4.2.8. 5-Fluoro-1H-indole-2,3-dione-3-[N-(4-methylphenyl)-
thiosemicarbazone] (2h). Orange powder (89%): mp 244–
246 ꢁC; IR (KBr): t 3315, 3242 (NH), 1692 (C@O), 1133
1
(C@S). H NMR (DMSO-d6/400 MHz): d 2.17 (s, 3H,
CH3), 6.78 (dd, J = 8.55, 4.16 Hz, 1H, indole C7–H), 7.04
(dt, J = 9.30, 2.64 Hz, 1H, indole C6–H), 7.08 (d,
J = 8.34 Hz, 2H, phenyl C3–H, C5–H), 7.31 (d,
J = 8.25 Hz, 2H, phenyl C2–H, C6–H), 7.48 (dd, J = 8.15,
2.56 Hz, 1H, indole C4–H), 10.63 (s, 1H, N4–H), 11.09 (s,
1H, indole NH), 12.49 (s, 1H, N2–H); LCMS-APCI (À/+):
m/z (%) 329 (MH+, 39), 180 (97), 152 (100). Anal. Calcd
for C16H13FN4OS (328.36): C, 58.52; H, 3.99; N, 17.06.
Found: C, 58.69; H, 4.02; N, 17.04.
4.2.5. 5-Fluoro-1H-indole-2,3-dione-3-(N-cyclohexylthio-
semicarbazone) (2e). Shining yellow crystals (89%): mp
249–250 ꢁC; IR (KBr): t 3360, 3171 (NH), 1692
1
(C@O), 1130 (C@S); H NMR (DMSO-d6/400 MHz):
d 1.15 (t, J = 12.54 Hz, 1H, cycl. C4–Hax), 1.28–1.35
(m, 2H, cycl. C3–Hax, C5–Hax), 1.42–1.51 (m, 2H, cycl.
C2–Hax, C6–Hax), 1.64 (d, J = 12.14 Hz, 1H, cycl.
4.2.9. 5-Fluoro-1H-indole-2,3-dione-3-[N-(4-chlorophenyl)-
thiosemicarbazone] (2i). Orange powder (97%): mp
260 ꢁC; IR (KBr): t 3319, 3195 (NH), 1689 (C@O),
1149 (C@S); 1H NMR (DMSO-d6/400 MHz): d 6.78
(dd, J = 8.56, 4.16 Hz, 1H, indole C7–H), 7.05 (dt,
J = 9.05, 2.70 Hz, 1H, indole C6–H), 7.34 (d,
J = 8.75 Hz, 2H, phenyl C3–H, C5–H), 7.45 (dd,
J = 8.13, 2.65 Hz, 1H, indole C4–H), 7.51 (d,
J = 8.76 Hz, 2H, phenyl C2–H, C6–H), 10.72 (s, 1H,
N4–H), 11.11 (s, 1H, indole NH), 12.51 (s, 1H, N2–H).
Anal. Calcd for C15H10ClFN4OS (348.78): C, 51.65;
H, 2.89; N, 16.06. Found: C, 52.13; H, 2.74; N, 16.06.
C4–Hequiv), 1.77 (d, J = 12.83 Hz, 2H, cycl. C3–Hequiv
C5–Hequiv), 1.92 (d, J = 9.78 Hz, 2H, cycl. C2–Hequiv
,
,
C6–Hequiv), 4.18–4.21 (m, 1H, cycl C1–H), 6.92 (dd,
J = 8.55, 4.17 Hz, 1H, indole C7–H), 7.19 (dt, J = 9.08,
2.73 Hz, 1H, indole C6–H), 7.62 (dd, J = 8.19, 2.71 Hz,
1H, indole C4–H), 8.87 (d, J = 8.49 Hz, 1H, N4–H),
11.21 (s, 1H, indole NH), 12.47 (s, 1H, N2–H); 13C
NMR (HETCOR DMSO-d6/100 MHz): d 25.36 (cycl.
C3, C5), 25.57 (cycl. C4), 31.96 (cycl. C2, C6), 54.27 (cycl.
C1), 108.60 (d, J = 25.7 Hz, indole C4), 112.46 (d,
J = 8.10 Hz, indole C7), 117.76 (d, J = 24.1 Hz, indole
C6), 121.83 (d, J = 9.41 Hz, indole C3a), 131.57 (d, J =
3.23 Hz, indole C7a), 138.91 (indole C3), 158.70 (d, J =
237.4 Hz, indole C5), 163.17 (indole C@O),
176.17 (C@S). Anal. Calcd for C15H17 FN4OS (320.38):
C, 56.23; H, 5.35; N, 17.49. Found: C, 55.96; H, 4.81; N,
17.52.
4.2.10. 5-Fluoro-1H-indole-2,3-dione-3-[N-(4-fluorophenyl)-
thiosemicarbazone] (2j). Orange powder (89%): mp 246–
248 ꢁC; IR (KBr): t 3331, 3194 (NH), 1692 (C@O),
1146 (C@S); 1H NMR (DMSO-d6/400 MHz): d 6.78
(dd, J = 8.56, 4.15 Hz, 1H, indole C7–H), 7.03 (dt,
J = 8.76, 2.52 Hz, 1H, indole C6–H), 7.12 (t,
J = 8.79 Hz, 2H, phenyl C3–H, C5–H), 7.43–7.46 (m,
3H, indole C4–H, phenyl C2–H, C6–H), 10.63 (s, 1H,
N4–H), 11.10 (s, 1H, indole NH), 12.52 (s, 1H, N2–H);
LCMS-APCI (À/+): m/z (%) 331 (MHÀ, 21), 178
(100), 333 (MH+, 82), 180 (100). Anal. Calcd for
C15H10F2N4OS (332.32): C, 54.21; H, 3.03; N, 16.86.
Found: C, 53.84; H, 3.00; N, 16.82.
4.2.6. 5-Fluoro-1H-indole-2,3-dione-3-(N-benzylthiosemi-
carbazone) (2f). Shining orange crystals (92%): mp 246–
247 ꢁC; IR (KBr): t 3312, 3207 (NH), 1690 (C@O),
1140 (C@S); 1H NMR (DMSO-d6/500 MHz): d 4.90
(d, J = 6.41 Hz, 2H, benzyl CH2), 6.93 (dd, J = 8.54,
4.27 Hz, 1H, indole C7–H), 7.20 (dt, J = 9.15, 2.75 Hz,
1H, indole C6–H), 7.28 (t, J = 6.86 Hz, 1H, benzyl C4–
H), 7.34–7.39 (m, 4H, benzyl C2–H, C3–H, C5–H, C6–
H), 7.49 (dd, J = 8.24, 2.74 Hz, 1H, indole C4–H), 9.87
(t, J = 6.10 Hz, 1H, N4–H), 11.24 (s, 1H, indole NH),
12.57 (s, 1H, N2–H); LCMS-APCI (À/+): m/z (%) 327
(MHÀ, 100). Anal. Calcd for C16H13FN4OS (328.36):
C, 58.52; H, 3.99; N, 17.06. Found: C, 58.52; H, 3.77;
N, 17.23.
4.2.11. 5-Fluoro-1H-indole-2,3-dione-3-[N-(4-nitrophenyl)-
thiosemicarbazone] (2k). Orange powder (91%): mp
267 ꢁC; IR (KBr): t 3298, 3190 (NH), 1694 (C@O),
1154 (C@S); 1H NMR (DMSO-d6/500 MHz): d 6.95
(dd, J = 8.54, 4.14 Hz, 1H, indole C7–H), 7.24 (dt,
J = 9.02, 2.28 Hz, 1H, indole C6–H), 7.63 (dd,
J = 7.81, 2.44 Hz, 1H, indole C4–H), 8.09 (d,
J = 9.27 Hz, 2H, phenyl C2–H, C6–H), 8.31 (d,
J = 8.78 Hz, 2H, phenyl C3–H, C5–H), 11.09 (s, 1H,
N4–H), 11.30 (s, 1H, indole NH), 12.89 (s, 1H, N2–H).
Anal. Calcd for C15H10FN5O3S (359.33): C, 50.14; H,
2.81; N, 19.49. Found: C, 49.81; H, 2.36; N, 19.88.
4.2.7. 5-Fluoro-1H-indole-2,3-dione-3-(N-phenylthiosemi-
carbazone) (2g). Orange powder (79%): mp 218 ꢁC; IR
(KBr): t 3363, 3177 (NH), 1690 (C@O), 1142 (C@S);
1H NMR (DMSO-d6/400 MHz): d 6.93 (dd, J = 8.56,
4.17 Hz, 1H, indole C7–H), 7.20 (dt, J = 8.37, 2.72 Hz,
1H, indole C6–H), 7.29 (t, J = 7.16 Hz, 1H, phenyl C4–
H), 7.44 (t, J = 7.01 Hz, 2H, phenyl C3–H, C5–H), 7.62
(d, J = 7.44 Hz, 2H, phenyl C2–H, C6–H), 7.65 (dd,
J = 8.02, 2.68 Hz, 1H, indole C4–H), 10.81 (s, 1H, N4–
H), 11.22 (s, 1H, indole NH), 12.67 (s, 1H, N2–H);
LCMS-APCI (À/+): m/z (%) 313 (MHÀ, 12), 178
(100), 315 (MH+, 72), 180 (100). Anal. Calcd for
4.3. General method for the synthesis of 5-fluoro-1-
morpholino/piperidinomethyl-1H-indole-2,3-dione-3-
thiosemicarbazones (3a–r)
To a suspension of 2a–k (2 mmol) in absolute ethanol
(20 mL), 37% formaldehyde solution (0.5 mL) and

5900
N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
morpholine or piperidine (2 mmol) were added dropwise
with vigorous stirring. After combining all reagents, the
reaction mixture was stirred at room temperature for
10 h. The solid product was filtered and washed with
petroleum ether.
N), 7.27–7.28 (m, 2H, indole C7–H, C6–H), 7.55 (dd,
J = 7.62, 1.83 Hz, 1H, indole C4–H), 9.38 (t,
J = 5.64 Hz, 1H, N4–H), 12.38 (s, 1H, N2–H); 13C
NMR (HETCOR DMSO-d6/100 MHz): d 14.41 (ethyl
CH3), 24.01 (pip. C4), 25.81 (pip. C3, C5), 39.36–40.61
(m, pip. C2, C6, DMSO), 51.78 (ethyl CH2), 62.51 (N–
CH2–N), 107.81 (d, J = 25.7 Hz, indole C4), 112.85 (d,
J = 7.96 Hz, indole C7), 117.48 (d, J = 24.1 Hz, indole
C6), 121.20 (d, J = 9.38 Hz, indole C3a), 130.25 (d,
J = 3.22 Hz, indole C7a), 140.34 (indole C3), 159.05 (d,
J = 238.48 Hz, indole C5), 162.07 (indole C@O),
177.07 (C@S). Anal. Calcd for C17H22FN5OS (363.45):
C, 56.18; H, 6.10; N, 19.27. Found: C, 55.95; H, 6.40;
N, 19.01.
4.3.1. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-(N-methylthiosemicarbazone) (3a). Yellow pow-
der (98%): mp 209 ꢁC; IR (KBr): t 3239 (NH), 1694
1
(C@O), 1139 (C@S); H NMR (DMSO-d6, 400 MHz):
d 2.55 (t, J = 4.19 Hz, 4H, morph. C3–H2, C5–H2),
3.10 (d, J = 4.40 Hz, 3H, CH3), 3.54 (t, J = 4.34 Hz,
4H, morph. C2–H2, C6–H2), 4.49 (s, 2H, N–CH2–N),
7.28–7.30 (m, 2H, indole C7–H, C6–H), 7.51 (dd,
J = 6.99, 2.13 Hz, 1H, indole C4–H), 9.38 (br s, 1H,
N4–H), 12.39 (s, 1H, N2–H). Anal. Calcd for
C15H18FN5O2S (351.39): C, 51.27; H, 5.16; N, 19.93.
Found: C, 51.14; H, 5.31; N, 19.78.
4.3.5. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-(N-allylthiosemicarbazone) (3e). Dark yellow
powder (81%): mp 153–154 ꢁC; IR (KBr): t 3235
1
(NH), 1686 (C@O), 1155 (C@S); H NMR (DMSO-d6/
4.3.2. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-(N-methylthiosemicarbazone) (3b). Yellow pow-
der (85%): mp 186–187 ꢁC; IR (KBr): t 3242 (NH),
1695 (C@O), 1155 (C@S); 1H NMR (DMSO-d6/
400 MHz): d 1.33 (br s, 2H, pip. C4–H2), 1.45 (br s,
4H, pip. C3–H2, C5–H2), 2.51 (s, pip. C2–H2, C6–H2,
DMSO), 3.09 (d, J = 4.52 Hz, 3H, CH3), 4.46 (s, 2H,
N–CH2–N), 7.25–7.27 (m, 2H, indole C7–H, C6–H),
7.49 (dd, J = 7.98, 1.16 Hz, 1H, indole C4–H), 9.36 (br
s, 1H, N4–H), 12.40 (s, 1H, N2–H). Anal. Calcd for
C16H20FN5OS (349.42): C, 55.00; H, 5.77; N, 20.04.
Found: C, 54.42; H, 5.88; N, 20.04.
400 MHz): d 2.55 (t, J = 3.97 Hz, 4H, morph. C3–H2,
C5–H2), 3.54 (t, J = 4.29 Hz, 4H, morph. C2–H2, C6–
H2), 4.27 (t, J = 5.54 Hz, 2H, allyl C1–H2), 4.49 (s, 2H,
N–CH2–N), 5.16 (dd, J = 10.23, 1.53 Hz, 1H, allyl C3–
Hcis), 5.21 (dd, J = 17.24, 1.58 Hz, 1H, allyl C3–Htrans),
5.89–5.96 (m, 1H, allyl C2–H), 7.28–7.30 (m, 2H, indole
C7–H, C6–H), 7.58 (dd, J = 7.72, 2.16 Hz, 1H, indole
C4–H), 9.58 (t, J = 5.86 Hz, 1H, N4–H), 12.41 (s, 1H,
N2–H). Anal. Calcd for C17H20FN5O2S (377.43): C,
54.10; H, 5.34; N, 18.56. Found: C, 54.10; H, 4.98; N,
18.65.
4.3.6. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-(N-allylthiosemicarbazone) (3f). Orange powder
(87%): mp 145–147 ꢁC; IR (KBr): t 3211 (NH), 1696
4.3.3. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-(N-ethylthiosemicarbazone) (3c). Yellow powder
(97%): mp 184 ꢁC; IR (KBr): t 3279 (NH), 1698
1
(C@O), 1145 (C@S); H NMR (DMSO-d6/400 MHz):
1
(C@O), 1139 (C@S); H NMR (DMSO-d6, 500 MHz):
d 1.33 (br s, 2H, pip. C4–H2), 1.46 (br s, 4H, pip. C3–
H2, C5–H2), 2.51 (s, DMSO, pip. C2–H2, C6–H2), 4.27
(t, J = 5.43 Hz, 2H, allyl C1–H2), 4.46 (s, 2H, N–CH2–
N), 5.16 (dd, J = 10.19, 1.04 Hz, 1H, allyl C3–Hcis),
5.24 (dd, J = 17.23, 1.23 Hz, 1H, allyl C3–Htrans), 5.88–
5.96 (m, 1H, allyl C2–H), 7.25 (d, J = 5.87 Hz, 2H, in-
dole C7–H, C6–H), 7.55 (d, J = 7.95 Hz, 1H, indole
C4–H), 9.55 (t, J = 5.84 Hz, 1H, N4–H), 12.41 (s, 1H,
N2–H). Anal. Calcd for C18H22FN5OS (375.46): C,
57.58; H, 5.91; N, 18.65. Found: C, 57.71; H, 5.80; N,
18.60.
d 1.21 (t, J = 7.17 Hz, 3H, ethyl CH3), 2.56 (t,
J = 3.66 Hz, 4H, morph. C3–H2, C5–H2), 3.55 (t,
J = 4.27 Hz, 4H, morph. C2–H2, C6–H2), 3.66 (p,
J = 6.56 Hz, 2H, ethyl CH2), 4.49 (s, 2H, N–CH2–N),
7.27–7.29 (m, 2H, indole C7–H, C6–H), 7.55 (dd,
J = 8.54, 1.52 Hz, 1H, indole C4–H), 9.39 (t,
J = 5.79 Hz, 1H, N4–H), 12.36 (s, 1H, N2–H); 13C
NMR (HETCOR DMSO-d6/100 MHz): d 14.42 (ethyl
CH3), 39.32–40.57 (m, ethyl CH2, DMSO), 50.96
(morph. C3, C5), 61.81 (N–CH2–N), 66.44 (morph. C2,
C6), 107.78 (d, J = 26.0 Hz, indole C4), 112.81 (d,
J = 8.01 Hz, indole C7), 117.50 (d, J = 24.0 Hz, indole
C6), 121.28 (d, J = 9.44 Hz, indole C3a), 130.30 (d,
J = 3.43 Hz, indole C7a), 140.07 (d, J = 1.19 Hz, indole
C3), 159.12 (d, J = 238.36 Hz, indole C5), 162.07 (indole
C@O), 177.06 (C@S). Anal. Calcd for C16H20FN5O2S
(365.42): C, 52.59; H, 5.52; N, 19.16. Found: C, 52.52;
H, 5.40; N, 19.16.
4.3.7. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-(N-butylthiosemicarbazone) (3g). Dark yellow
powder (75%): mp 140–141 ꢁC; IR (KBr): t 3216
1
(NH), 1701 (C@O), 1130 (C@S); H NMR (DMSO-d6/
500 MHz): d 0.94 (t, J = 7.25 Hz, 3H, butyl CH3),
1.35–1.38 (m, 2H, butyl C3–H2), 1.63 (p, J = 7.05 Hz,
2H, butyl C2–H2), 2.56 (t, J = 4.03 Hz, 4H, morph.
C3–H2, C5–H2), 3.55 (t, J = 4.23 Hz, 4H, morph. C2–
H2, C6–H2), 3.63 (q, J = 6.84 Hz, 2H, butyl C1–H2),
4.49 (s, 2H, N–CH2–N), 7.27–7.29 (m, 2H, indole C7–
H, C6–H), 7.57 (dd, J = 7.45, 2.21 Hz, 1H, indole C4–
H), 9.36 (t, J = 5.84 Hz, 1H, N4–H), 12.36 (s, 1H, N2–
H); 13C NMR (APT DMSO-d6/100 MHz): d 14.21 (bu-
tyl CH3), 20.06 (butyl C3), 30.96 (butyl C2), 44.42 (butyl
C1), 50.96 (morph. C3, C5), 61.81 (N–CH2–N), 66.45
(morph. C2, C6), 107.90 (d, J = 25.8 Hz, indole C4),
4.3.4. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-(N-ethylthiosemicarbazone) (3d). Orange powder
(92%): mp 176–177 ꢁC; IR (KBr): t 3251 (NH), 1690
1
(C@O), 1142 (C@S); H NMR (DMSO-d6/500 MHz):
d 1.21 (t, J = 7.17 Hz, 3H, ethyl CH3), 1.34 (br s, 2H,
pip. C4–H2), 1.47 (p, J = 5.34 Hz, 4H, pip. C3–H2, C5–
H2), 2.50–2.52 (m, DMSO, pip. C2–H2, C6–H2), 3.66
(p, J = 6.56 Hz, 2H, ethyl CH2), 4.48 (s, 2H, N–CH2–

N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
5901
112.76 (d, J = 7.95 Hz, indole C7), 117.45 (d,
J = 24.1 Hz, indole C6), 121.30 (d, J = 9.44 Hz, indole
C3a), 130.22 (d, J = 3.37 Hz, indole C7a), 140.05 (d,
J = 1.28 Hz, indole C3), 159.13 (d, J = 238.50 Hz, indole
C5), 162.06 (indole C@O), 177.30 (C@S). LCMS-APCI
(À/+): m/z (%) 392 (MHÀ, 100), 394 (MH+, 100). Anal.
Calcd for C18H24FN5O2S (393.47): C, 54.94; H, 6.15; N,
17.80. Found: C, 54.87; H, 6.04; N, 17.84.
4.3.11. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-(N-benzylthiosemicarbazone) (3k). Orange pow-
der (98%): mp 177 ꢁC; IR (KBr): t 3270 (NH), 1690
1
(C@O), 1158 (C@S); H NMR (DMSO-d6/500 MHz):
d 1.34 (br s, 2H, pip. C4–H2), 1.47 (br p, J = 5.18 Hz,
4H, pip. C3–H2, C5–H2), 2.51–2.53 (m, DMSO, pip.
C2–H2, C6–H2), 4.49 (s, 2H, N–CH2–N), 4.91 (d,
J = 6.10 Hz, 2H, benzyl CH2), 7.27–7.39 (m, 7H, indole
C7–H, C6–H, phenyl), 7.55 (dd, J = 7.32, 2.13 Hz, 1H,
indole C4–H), 9.92 (t, J = 6.30 Hz, 1H, N4–H), 12.49
(s, 1H, N2–H). Anal. Calcd for C22H24FN5OS
(425.52): C, 62.10; H, 5.68; N, 16.46. Found: C, 61.79;
H, 5.47; N, 16.33.
4.3.8. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-(N-butylthiosemicarbazone) (3h). Dark yellow
powder (88%): mp 145–146 ꢁC; IR (KBr): t 3214
1
(NH), 1697 (C@O), 1143 (C@S); H NMR (DMSO-d6/
500 MHz): d 0.93 (t, J = 7.35 Hz, 3H, butyl CH3),
1.33–1.38 (m, 4H, pip. C4–H2, butyl C3–H2), 1.46 (br
s, 4H, pip. C3–H2, C5–H2), 1.63 (p, J = 7.35 Hz, 2H, bu-
tyl C2–H2), 2.51–2.52 (m, DMSO, pip. C2–H, C6–H),
3.63 (q, J = 6.92 Hz, 2H, butyl C1–H2), 4.48 (s, 2H,
N–CH2–N), 7.26–7.27 (m, 2H, indole C7–H, C6–H),
7.56 (dd, J = 7.83, 1.11 Hz, 1H, indole C4–H), 9.35 (t,
J = 5.75 Hz, 1H, N4–H), 12.38 (s, 1H, N2–H);13C
4.3.12. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-(N-phenylthiosemicarbazone) (3l). Orange pow-
der (92%): mp 179 ꢁC; IR (KBr): t 3308, 3218
(NH), 1689 (C@O), 1155 (C@S); 1H NMR
(DMSO-d6/500 MHz):
d 2.58 (t, J = 4.42 Hz, 4H,
morph. C3–H2, C5–H2), 3.55 (t, J = 4.57 Hz, 4H,
morph. C2–H2, C6–H2), 4.51 (s, 2H, N–CH2–N),
7.28–7.31(m, 3H, indole C7–H, C6–H, phenyl C4–
H), 7.45 (dd, J = 7.93, 1.83 Hz, 2H, phenyl C3–H,
C5–H), 7.62 (d, J = 8.53 Hz, 2H, phenyl C2–H, C6–
H), 7.71 (dd, J = 8.69, 1.62 Hz, 1H, indole C4–H),
10.88 (s, 1H, N4–H), 12.58 (s, 1H, N2–H). Anal.
Calcd for C20H20FN5O2SÆ1/2H2O (422.47): C, 56.85;
H, 5.01; N, 16.57. Found: C, 56.27; H, 5.14; N,
16.05.
NMR (APT DMSO-d6/100 MHz):
d 14.20 (butyl
CH3), 20.04 (butyl C3), 24.00 (pip. C4), 25.80 (pip. C3,
C5), 30.94 (butyl C2), 44.41 (butyl C1), 51.78 (pip. C2,
C6), 62.50 (N–CH2–N), 107.87 (d, J = 25.7 Hz, indole
C4), 112.86 (d, J = 7.73 Hz, indole C7), 117.49 (d,
J = 23.9 Hz, indole C6), 121.20 (d, J = 9.30 Hz, indole
C3a), 130.28 (indole C7a), 140.34 (indole C3), 159.06 (d,
J = 238.4 Hz, indole C5), 162.09 (indole C@O), 177.29
(C@S). Anal. Calcd for C19H26FN5OS (391.50): C,
58.29; H, 6.69; N, 17.89. Found: C, 57.85; H, 6.80; N,
17.82.
4.3.13. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-(N-phenylthiosemicarbazone) (3m). Dark yellow
powder (98%): mp 174–175 ꢁC; IR (KBr): t 3285,
3215 (NH), 1692 (C@O), 1135 (C@S); 1H NMR
(DMSO-d6/500 MHz): d 1.34 (br s, 2H, pip. C4–H2),
1.47 (br p, J = 5.41 Hz, 4H, pip. C3–H2, C5–H2),
2.54 (br t, J = 4.88 Hz, 4H, pip. C2–H2, C6–H2), 4.49
(s, 2H, N–CH2–N), 7.27–7.31 (m, 3H, indole C7–H,
C6–H, phenyl C4–H), 7.44 (t, J = 7.93 Hz, 2H, phenyl
C3–H, C5–H), 7.62 (d, J = 8.07 Hz, 2H, phenyl C2–H,
C6–H), 7.70 (d, J = 7.62 Hz, 1H, indole C4–H), 10.87
(s, 1H, N4–H), 12.60 (s, 1H, N2–H); LCMS-APCI
(À/+): m/z (%) 410 (MHÀ, 10), 178 (100), 412
(MH+, 12), 180 (100). Anal. Calcd for C21H22FN5OS
(411.49): C, 61.29; H, 5.38; N, 17.01. Found: C,
61.32; H, 5.61; N, 16.95.
4.3.9. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-(N-cyclohexylthiosemicarbazone) (3i). Yellow
powder (93%): mp 176–177 ꢁC; IR (KBr): t 3220
1
(NH),1695 (C@O), 1135 (C@S); H NMR (DMSO-d6/
400 MHz): d 1.16 (t, J = 12.58 Hz, 1H, cycl. C4–Hax),
1.26–1.35 (m, 2H, cycl. C3–Hax, C5–Hax), 1.42–1.52
(m, 2H, cycl. C2–Hax, C6–Hax), 1.64 (d, J = 12.43 Hz,
1H, cycl. C4–Hequiv), 1.78 (d, J = 12.98 Hz, 2H, cycl.
C3–Hequiv, C5–Hequiv), 1.93 (d, J = 9.94 Hz, 2H, cycl.
C2–Hequiv, C6–Hequiv), 2.55 (t, J = 4.01 Hz, 4H, morph.
C3–H2, C5–H2), 3.54 (t, J = 4.22 Hz, 4H, morph. C2–
H2, C6–H2), 4.18–4.22 (m, 1H, cycl. C1–H), 4.49 (s,
2H, N–CH2–N), 7.26–7.29 (m, 2H, indole C7–H, C6–
H), 7.69 (dd, J = 7.42, 1.38 Hz, 1H, indole C4–H), 8.92
(d, J = 8.52 Hz, 1H, N4–H), 12.39 (s, 1H, N2–H). Anal.
Calcd for C20H26FN5O2S (419.51): C, 57.26; H, 6.25; N,
16.69. Found: C, 57.56; H, 6.15; N, 16.76.
4.3.14. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-[N-(4-methylphenyl)thiosemicarbazone] (3n). Or-
ange powder (89%): mp 161–162 ꢁC; IR (KBr): t
3291, 3215 (NH), 1689 (C@O), 1145 (C@S); 1H
NMR (DMSO-d6/500 MHz): d 1.35 (br s, 2H, pip.
C4–H2), 1.47 (br p, J = 5.40 Hz, 4H, pip. C3–H2,
C5–H2), 2.34 (s, 3H, CH3), 2.54 (br t, J = 4.88 Hz,
4H, pip. C2–H2, C6–H2), 4.49 (s, 2H, N–CH2–N),
7.24 (d, J = 8.54 Hz, 2H, phenyl C3–H, C5–H), 7.27–
7.29 (m, 2H, indole C7–H, C6–H), 7.48 (d,
J = 8.24 Hz, 2H, phenyl C2–H, C6–H), 7.69 (dd,
J = 8.24, 1.22 Hz, 1H, indole C4–H), 10.81 (s, 1H,
N4–H), 12.58 (s, 1H, N2–H); LCMS-APCI (À/+):
m/z (%) 426 (MH+, 10), 329 (100). Anal. Calcd for
C22H24FN5OS (425.52): C, 62.10; H, 5.68; N, 16.46.
Found: C, 61.98; H, 6.00; N, 16.48.
4.3.10. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-(N-benzylthiosemicarbazone) (3j). Orange pow-
der (96%): mp 171–172 ꢁC; IR (KBr): t 3247 (NH),
1691 (C@O), 1139 (C@S); 1H NMR (DMSO-d6/
500 MHz): d 2.57 (t, J = 4.42 Hz, 4H, morph. C3–H2,
C5–H2), 3.55 (t, J = 4.42 Hz, 4H, morph. C2–H2, C6–
H2), 4.50 (s, 2H, N–CH2–N), 4.91 (d, J = 6.10 Hz, 2H,
benzyl CH2), 7.27–7.39 (m, 7H, indole C7–H, C6–H,
phenyl), 7.56 (dd, J = 7.94, 1.97 Hz, 1H, indole C4–H),
9.93 (t, J = 6.30 Hz, 1H, N4–H), 12.47 (s, 1H, N2–H).
Anal. Calcd for C21H22FN5O2S (427.49): C, 59.00; H,
5.19; N, 16.38. Found: C, 59.05; H, 5.09; N, 16.45.

5902
N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
4.3.15. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-[N-(4-chlorophenyl)thiosemicarbazone] (3o). Dark
yellow powder (98%): mp 185–187 ꢁC; IR (KBr): t 3308,
3220 (NH), 1690 (C@O), 1140 (C@S); 1H NMR
(DMSO-d6/400 MHz): d 2.57 (br s, 4H, morph. C3–H2,
C5–H2), 3.55 (t, J = 4.17 Hz, 4H, morph. C2–H2, C6–H2),
4.51 (s, 2H, N–CH2–N), 7.31 (d, J = 6.81 Hz, 2H, indole
C7–H, C6–H), 7.51 (d, J = 8.69 Hz, 2H, phenyl C3–H,
C5–H), 7.67 (d, J = 8.59 Hz, 2H, phenyl C2–H, C6–H),
7.68 (d, J = 6.50 Hz, 1H, indole C4–H), 10.93 (s, 1H, N4–
H), 12.62(s, 1H, N2–H). Anal. Calcdfor C20H19ClFN5O2S
(447.91): C, 53.63; H, 4.28; N, 15.64. Found: C, 53.19; H,
4.57; N, 15.47.
pholine or piperidine (2 mmol) were added dropwise
with vigorous stirring. After combining all reagents,
the reaction mixture was refluxed on a water bath for
4 h. The product formed was filtered and washed with
petroleum ether.
4.5. General method for the synthesis of 5-nitro-1H-
indole-2,3-dione-3-[(4-oxo-1,3-thiazolidin-2-ylidene)-
hydrazones] (5a–s)20
To a suspension of 2l–v (2.5 mmol) in absolute ethanol
(25 mL), anhydrous sodium acetate (1 mmol) and ethyl
bromoacetate or ethyl 2-bromopropionate (2.5 mmol)
were added. The reaction mixture was refluxed for 4 h,
cooled, diluted with water, and allowed to stand over-
night. The yellow crystals obtained were filtered and
purified by recrystallization from ethanol.
4.3.16. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-[N-(4-chlorophenyl)thiosemicarbazone] (3p). Yel-
low powder (93%): mp 178–179 ꢁC; IR (KBr): t 3301,
3237 (NH), 1698 (C@O), 1140 (C@S); 1H NMR
(DMSO-d6/400 MHz): d 1.34 (br s, 2H, pip. C4–H2),
1.47 (br s, 4H, pip. C3–H2, C5–H2), 2.53 (br s, 4H,
pip. C2–H2, C6–H2), 4.50 (s, 2H, N–CH2–N), 7.30 (dd,
J = 7.37, 1.70 Hz, 2H, indole C7–H, C6–H), 7.51 (d,
J = 8.79 Hz, 2H, phenyl C3–H, C5–H), 7.65–7.68 (m,
3H, phenyl C2–H, C6–H, indole C4–H), 10.92 (s, 1H,
N4–H), 12.65 (s, 1H, N2–H). Anal. Calcd for
C21H21ClFN5OS. 1/2H2O (454.95): C, 55.44; H, 4.87;
N, 15.39. Found: C, 55.38; H, 4.89; N, 15.48.
4.6. Single crystal X-ray structure determinations
Specimens of 3e and 3f of suitable quality and size were
mounted on the ends of glass and used for the intensity
data collection. The X-ray diffraction data on 3e and 3f
were collected using an automatic X-ray four-circle
EXCALIBUR single crystal diffractometer with CCD
area detectors. Graphite monochromated Mo Ka radia-
˚
tion (k = 0.71073 A) was generated at 50 kV and 40 mA.
The intensities of the reflections were recorded in 104
frames (each frame consisting of 1024 · 1024 pixels with
2 · 2 pixels bining) for both crystals. The lattice param-
eters were calculated from refinement of positions of all
measured reflections. The data sets were corrected for
Lorentz and polarization effect. No absorption correc-
tion was applied for both crystals.
4.3.17. 5-Fluoro-1-(morpholin-4-ylmethyl)-1H-indole-2,3-
dione-3-[N-(4-fluorophenyl)thiosemicarbazone] (3q). Dark
yellow powder (67%). mp 167–168 ꢁC; IR (KBr): t
1
3309, 3288 (NH), 1689 (C@O), 1138 (C@S); H NMR
(DMSO-d6/500 MHz): d 2.57 (t, J = 4.39 Hz, 4H, morph.
C3–H2, C5–H2), 3.55 (t, J = 4.39 Hz, 4H, morph. C2–H2,
C6–H2), 4.51 (s, 2H, N–CH2–N), 7.26–7.31 (m, 4H, in-
dole C7–H, C6–H, phenyl C3–H, C5–H), 7.61 (dd,
J = 8.79, 4.88 Hz, 2H, phenyl C2–H, C6–H), 7.68 (dd,
J = 7.32, 1.46 Hz, 1H, indole C4–H), 10.88 (s, 1H, N4–
H), 12.59 (s, 1H, N2–H); LCMS-APCI (À/+): m/z (%)
430 (MHÀ, 2), 331 (100). Anal. Calcd for
C20H19F2N5O2S (431.45): C, 55.67; H, 4.44; N, 16.23.
Found: C, 55.31; H, 4.00; N, 15.90.
Both structures were solved by direct methods program
SHELXS-9741 and refined by the full-matrix least-
squares method based on F2 using SHELXL-97.42 All
non-hydrogen atoms were refined with anisotropic dis-
placement factors. N4 and N5 hydrogen atoms were
positioned from difference map, and redundant hydro-
gen atoms were positioned geometrically and refined
using a riding model. The geometric calculations were
carried out with the Platon43 program.
4.3.18. 5-Fluoro-1-(piperidin-1-ylmethyl)-1H-indole-2,3-
dione-3-[N-(4-fluorophenyl)thiosemicarbazone] (3r). Dark
yellow powder (74%): mp 149–151 ꢁC; IR (KBr): t
4.7. In vitro evaluation of antituberculosis activity
1
3305, 3230 (NH), 1697 (C@O), 1137 (C@S); H NMR
(DMSO-d6/500 MHz): d 1.35 (br s, 2H, pip. C4–H2),
1.48 (br p, J = 5.25 Hz, 4H, pip. C3–H2, C5–H2), 2.54
(t, J = 4.69 Hz, 4H, pip. C2–H, C6–H), 4.50 (s, 2H, N–
CH2–N), 7.27–7.30 (m, 4H, indole C7–H, C6–H, phenyl
C3–H, C5–H), 7.62 (dd, J = 9.06, 5.03 Hz, 2H, phenyl
C2–H, C6–H), 7.67 (d, J = 8.05 Hz, 1H, indole C4–H),
10.88 (s, 1H, N4–H), 12.62 (s, 1H, N2–H). Anal. Calcd
for C21H21F2N5OS (429.48): C, 58.73; H, 4.93; N,
16.31. Found: C, 58.40; H, 4.70; N, 16.74.
The primary screen was conducted at 6.25 lg/mL (or
molar equivalent of highest molecular weight compound
in a series of congeners) against Mycobacterium tubercu-
losis H37Rv (ATCC 27294) in BACTEC 12B medium
using a broth microdilution assay, the Microplate Ala-
mar Blue Assay (MABA). Compounds exhibiting fluo-
rescence were tested in the BACTEC 460 radiometric
system.25 Compounds effecting <90% inhibition in the
primary screen (MIC > 6.25 lg/mL) were not generally
evaluated further. Compounds demonstrating at least
90% inhibition in the primary screen were re-tested at
lower concentrations against M. tuberculosis H37Rv to
determine the actual minimum inhibitory concentration
(MIC) in the MABA. The MIC was defined as the low-
est concentration effecting the reduction in fluorescence
of 90% relative to controls. Concurrent with the deter-
4.4. General method for the synthesis of 5-nitro-1-
morpholino/piperidinomethyl-1H-indole-2,3-dione-3-
thiosemicarbazones (4a–l)19
To a suspension of 2l–v (2 mmol) in absolute ethanol
(20 mL), 37% formaldehyde solution (0.5 mL) and mor-

N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
5903
mination of MICs, compounds were tested for cytotox-
icity (IC50) in VERO cells at concentrations 10· the
MIC for M. tuberculosis H37Rv.
vent with bacterial inoculum and solvent with a 1:100
dilution of bacterial inoculum (1:100 controls). Vials
were incubated at 37 ꢁC, and the GI was determined
in a BACTEC 460 instrument (Becton Dickinson) until
the GI of the 1:100 controls reached at least 30. All vials
were read the following day, and the GI and daily
change in GI (DGI) were recorded for each drug dilu-
tion. The MIC was defined as the lowest concentration
for which the DGI was less than the DGI of the 1:100
control. If the GI of the test sample was greater than
100, the sample was scored as resistant even if the DGI
was less than the DGI of the 1:100 control.
4.7.1. Microplate Alamar Blue Assay (MABA). Antimi-
crobial susceptibility testing was performed in black,
clear-bottomed, 96-well microplates (black view plates;
Packard Instrument Company, Meriden, Conn.) in or-
der to minimize background fluorescence. Outer perim-
eter wells were filled with sterile water to prevent
dehydration in experimental wells. Initial drug dilutions
were prepared in either dimethylsulfoxide or distilled
deionized water, and subsequent twofold dilutions were
performed in 0.1 mL of 7H9GC (no Tween 80) in the
microplates. BACTEC 12B-passaged inocula were ini-
tially diluted 1:2 in 7H9GC, and 0.1 mL was added to
wells. The determination of bacterial titer yielded
1 · 106 CFU/ml in plate well for H37Rv. Frozen inocula
were initially diluted 1:20 in BACTEC 12B medium fol-
lowed by a 1:50 dilution in 7H9GC. Addition of
1/10 mL to wells resulted in final bacterial titer of
20 · 105 CFU/ml for H37Rv. Wells containing drug only
were used to detect autofluorescence of compounds.
Additional control wells consisted of bacteria only (B)
and medium only (M). Plates were incubated at 37 ꢁC.
Starting at day 4 of incubation, 20 lL of 10· Alamar
Blue solution (Alamar Biosciences/Accumed, Westlake,
Ohio) and 12.5 lL of 20% Tween 80 were added to
one B well and one M well, and plates were reincubated
at 37 ꢁC. Wells were observed 12 and 24 h later for a col-
or change from blue to pink and for a reading of
P50,000 fluorescence units (FU). Fluorescence was
measured in a Cytofluor II microplate fluorometer
(PerSeptive Biosystems, Framingham, Mass.) in bot-
tom-reading mode with excitation at 530 nm and emis-
sion at 590 nm. If the B wells become pink after 24 h,
the reagent is being added to the entire plate. If the well
remains blue or P50,000 FU is measured, additional M
and B wells are tested daily until a color change oc-
curred, at which time reagents are added to all remain-
ing wells. Plates were then incubated at 37 ꢁC, and
results were recorded at 24 h post-reagent addition. Vi-
sual MICs were defined as the lowest concentration of
drug that prevented a color change. For fluorometric
MICs, a background subtraction was performed on all
wells with a mean of triplicate M wells. Percent inhibi-
tion was defined as (1 À (test well FU/mean FU of trip-
licate B wells) · 100). The lowest drug concentration
effecting an inhibition of P90% was considered the
MIC.
Acknowledgments
We thank Dr. Joseph A. Maddry from the Tuberculosis
Antimicrobial Acquisition and Coordinating Facility
(TAACF), National Institute of Allergy and Infectious
Diseases Southern Reserch Institute, Alabama, USA,
for the evaluation of anti-TB activity. This work was
supported by Istanbul University Research Fund, Pro-
ject Numbers: 167/15012004 and BYP-484/30092004.
The authors (FBK and SO) wish to acknowledge the
support of Hacettepe University Research Fund (Project
Number: 0302602001).
References and notes
1. Shinnick, T. M.; King, C. H.; Quinn, F. D. Am. J. Med.
Sci. 1995, 309, 92.
2. Snider, D. E., Jr.; La Montagne, J. R. J. Infect. Dis. 1994,
169, 1189.
3. Pablos-Mendez, A.; Raviglione, M. C.; Laszlo, A.; Binkin,
N.; Rieder, H. L.; Bustreo, F.; Cohn, D. L.; Lambregts-
van Weezenbeek, C. S.; Kim, S. J.; Cahulet, P.; Nunn, P.
N. Eng. J. Med. 1998, 338, 1641.
4. Houston, S.; Fanning, A. Drugs 1994, 48, 689.
5. Cocco, M. T.; Congiu, C.; Onnis, V.; Pellerano, M. L.; De
Logu, A. Bioorg. Med. Chem. 2002, 10, 501.
6. Bermudez, L. E.; Reynolds, R.; Kolonoski, P.; Aralar, P.;
Inderlied, C. B.; Young, L. S. Antimicrob. Agents Chemo-
ther. 2003, 47, 2685.
7. Logu, A. D.; Saddi, M.; Onnis, V.; Sanna, C.; Congiu, C.;
Borgna, R.; Cocco, M. T. Int. J. Antimicrob. Agents 2005,
26, 28.
8. Sethi, M. L. Antiviral agents and protease inhibitors. In
Principles of Medicinal Chemistry; Foye, W. O., Lemke, T.
L., Williams, D. A., Eds.; Williams and Wilkins: Balti-
more, 2002; p 952.
9. Sriram, D.; Perumal, Y. Curr. Med. Chem. 2003, 10, 1689.
10. Pirrung, M. C.; Pansare, S. V.; Das Sarma, K.; Keith, K.
A.; Kern, E. R. J. Med. Chem. 2005, 48, 3045.
11. Bal, T. R.; Anand, B.; Yogeeswari, P.; Sriram, D. Bioorg.
Med. Chem. Lett. 2005, 15, 4451.
4.7.2. BACTEC radiometric assay. A total of 1/10 mL of
BACTEC 12B-passaged inoculum was delivered with-
out prior dilution into 4 mL of test medium. The deter-
mination of bacterial titer yielded average titer of
1 · 105 CFU/ml of BACTEC 12B medium for H37Rv.
Frozen inocula were initially diluted 1:20 in BACTEC
12B medium, and then 0.1 mL was delivered to the test
medium. This yielded 5.0 · 105 CFU per BACTEC vial
for H37Rv. Twofold drug dilutions were prepared in
either dimethylsulfoxide (DMSO) or distilled deionized
water and delivered via a 0.5-mL insulin syringe in a
50-lL volume. Drug-free control vials consisted of sol-
12. Karalı, N.; Terziog˘lu, N.; Gursoy, A. Arzneim.-Forsch./
Drug Res. 1998, 48, 758.
¨
13. Sriram, D.; Yogeeswari, P.; Gopal, G. Eur. J. Med. Chem.
2005, 40, 1373.
14. Pandeya, S. N.; Sriram, D.; Nath, G.; DeClercq, E. Eur. J.
Pharm. Sci. 1999, 9, 25.
15. Pandeya, S. N.; Sriram, D.; Nath, G.; DeClercq, E.
Arzneim.-Forsch./ Drug Res. 2000, 50.
16. Webber, S. E.; Tikhe, J.; Worland, S. T.; Fuhrman, S. A.;
Hendrickson, T. F.; Matthews, D. A.; Love, R. A.; Patick,

5904
N. Karalı et al. / Bioorg. Med. Chem. 15 (2007) 5888–5904
A. K.; Meador, J. W.; Ferre, R. A.; Brown, E. L.; DeLisle,
D. M.; Ford, C. E.; Binford, S. L. J. Med. Chem. 1996, 39,
5072.
tational Chemistry; Lipkowitz, K. B.; Boyd, D. B., Eds.;
VCH: New-York, 1991; Chapter 10.
30. Dimoglo, A. S.; Shvets, N. M.; Tetko, I. V.; Livingstone,
D. J. Quant. Struct.-Act. Relat. 2001, 20, 31.
31. Shvets, N. M.; Dimoglo, A. S. Nahrung 1998, 42, 364.
32. Kandemirli, F.; Sarac¸og˘lu, M.; Kovalishyn, V. Mini Rev.
Med. Chem. 2005, 5, 479.
17. Velezheva, V. S.; Brennan, P. J.; Marshakov, V. Y.;
Gusev, D. V.; Lisichkina, I. N.; Peregudov, A. S.;
Tchernousova, L. N.; Smirnova, T. G.; Andreevskaya,
S. N.; Medvedev, A. E. J. Med. Chem. 2004, 47, 3455.
18. Badawy, M. A.; Abdel-Hady, S. A. Arch. Pharm. (Wein-
heim) 1991, 324, 349.
33. Kohonen, T. Self-organisation Maps; Springer: Berlin,
1995.
19. Karalı, N. Eur. J. Med. Chem. 2002, 37, 909.
20. Terziog˘lu, N.; Karalı, N.; Gursoy, A.; Pannecouque, C.;
¨
34. Bishop, M. Neural Networks for Pattern Recognition;
Oxford University Press: Oxford, 1995.
35. Tetko, I. V. www.vcclab.org, 2005.
36. Dimoglo, A.; Kovalishyn, V.; Shvets, N.; Ahsen, V. Mini
Rev. Med. Chem. 2005, 5, 879.
37. Kohonen, T.; Hinninen, J.; Kangas, J.; Laaksonen, J.
SOM_PAK: The self-organizing map program package is
obtainable via anonymous ftp from the internet address
‘‘cochlea.hut.fi’’ (130.233.168.48).
38. Tetko, I. V.; Livingstone, D. J.; Luik, A. I. J. Chem. Inf.
Comput. Sci. 1995, 35, 826.
Leysen, P.; Paeshuyse, J.; Neyts, J.; De Clercq, E.
ARKIVOC 2006, i, 109.
21. Macaev, F.; Rusu, G.; Pogrebnoi, S.; Gudima, A.;
Stingaci, E.; Vlad, L.; Shvets, N.; Kandemirli, F.;
Dimoglo, A.; Reynolds, R. Bioorg. Med. Chem. 2005,
13, 4842.
22. Oruc, E. E.; Rollas, S.; Kandemirli, F.; Shvets, N.;
Dimoglo, A. J. Med. Chem. 2004, 47, 6760.
23. Cremer, D.; Pople, J. A. J. Am. Chem. Soc. 1975, 97, 1354.
24. Umezawa, Y.; Tsuboyama, S.; Takahashi, H.; Uzawa, J.;
Nishio, M. Tetrahedron 1999, 55, 10047.
39. Tetko, I. V.; Villa, A. E. P.; Livingstone, D. J. J. Chem.
Inf. Comput. Sci. 1996, 36, 794.
25. Collins, L. A.; Franzblau, S. G. Antimicrob. Agents
Chemother. 1997, 41, 1004.
26. MOPAC: A General Molecular Orbital Package (Version
6.0), Steward J.J.P. QCPE #455.
27. Dimoglo, A. S. Khimiko-pharm. Zhurnal 1985, 4, 438.
28. Shvets, N. M. J. Comput. Sci. Moldova 1993, 1, 101.
29. Bersuker, I. B.; Dimoglo, A. S. The Electron- Topological
Approach to the QSAR Problem. In Reviews in Compu-
40. Kovalishyn, V. V.; Tetko, I. V.; Luik, A. I.; Kholodovych,
V. V.; Villa, A. E. P.; Livingstone, D. J. J. Chem. Inf.
Comput. Sci. 1998, 38, 651.
41. Sheldrick, G .M. SHELXS97. Program for crystal struc-
ture solution. University of Go¨ttingen, Germany, 1997.
42. Sheldrick, G. M. SHELXL97. Program for crystal struc-
ture refinement. University of Go¨ttingen, Germany, 1997.
43. Spek, A. L. Acta Cryst. 1990, A46, C34.