Molecular basis of the interaction of novel tributyltin(IV) 2/4-[(E)-2-(aryl)-1-diazenyl]benzoates endowed with an improved cytotoxic profile: Synthesis, structure, biological efficacy and QSAR studies

Article abstract of DOI:10.1016/j.jinorgbio.2010.05.001

A series of tributyltin(IV) complexes based on 2/4-[(E)-2-(aryl)-1-diazenyl]benzoate ligands was synthesized, wherein the position of the carboxylate and aryl substituents (methyl, tert-butyl and hydroxyl) varies. The complexes, Bu₃SnL^1-4<

Full text of DOI:10.1016/j.jinorgbio.2010.05.001

Journal of Inorganic Biochemistry 104 (2010) 950–966
Contents lists available at ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Molecular basis of the interaction of novel tributyltin(IV) 2/4-[(E)-2-(aryl)-1-diazenyl]
benzoates endowed with an improved cytotoxic profile: Synthesis, structure, biological
efficacy and QSAR studies
a
b
b
b
Tushar S. Basu Baul a, , Anup Paul , Lorenzo Pellerito , Michelangelo Scopelliti , Claudia Pellerito ,
⁎
Palwinder Singh c, Pooja Verma c, Andrew Duthie d, Dick de Vos e, Rajeshwar P. Verma f, Ulli Englert g
Department of Chemistry, North-Eastern Hill University, NEHU Permanent Campus, Umshing, Shillong 793 022, India
Dipartimento di Chimica Inorganica e Analitica “Stanislao Cannizzaro” Università degli Studi di Palermo, Viale delle Scienze, Parco D'Orleans II, Edificio 17, 90128 Palermo, Italy
Department of Chemistry, Guru Nanak Dev University, Amritsar 143005, India
School of Life and Environmental Science, Deakin University, Geelong Victoria 3217 Australia
Pharmachemie BV, PO Box 552, 2003 RN Haarlem, The Netherlands
Department of Chemistry, Pomona College, 645 North College Avenue, Claremont, California 91711, USA
Institut für Anorganische Chemie, RWTH Aachen University, 52056 Aachen, Germany
a
b
c
d
e
f
g
a r t i c l e i n f o
a b s t r a c t
Article history:
A
series of tributyltin(IV) complexes based on 2/4-[(E)-2-(aryl)-1-diazenyl]benzoate ligands was
Received 10 March 2010
Received in revised form 2 May 2010
Accepted 3 May 2010
synthesized, wherein the position of the carboxylate and aryl substituents (methyl, tert-butyl and hydroxyl)
varies. The complexes, Bu3SnL1–4H (1–4), have been structurally characterized by elemental analysis and IR,
NMR (1H, 13C, and 119Sn) and 119Sn Mössbauer spectroscopy. All have a tetrahedral geometry in solution and
a trigonal bipyramidal geometry in the solid-state, except for Bu3SnL4H (4) that was ascertained to have
tetrahedral coordination by X-ray crystallography. Cytotoxicity studies were carried out on human tumor
cell lines A498 (renal cancer), EVSA-T (mammary cancer), H226 (non-small-cell lung cancer), IGROV
(ovarian cancer), M19 MEL (melanoma), MCF-7 (mammary cancer) and WIDR (colon cancer). Compared to
cisplatin, test compounds 1–4 had remarkably good activity, despite the presence of substantial steric bulk
due to Sn–Bu ligands. The quantitative structure–activity relationship (QSAR) studies for the cytotoxicity of
organotin(IV) benzoates, along with some reference drug molecules, is also discussed against a panel of
human tumor cell lines. Molecular structures of the tributyltin(IV) complexes (1–4) were fully optimized
using the PM6 semi-empirical method and docking studies performed with key enzymes associated with the
propagation of cancer, namely ribonucleotide reductase, thymidylate synthase, thymidylate phosphorylase
and topoisomerase II. The theoretical results are discussed in relation to the mechanistic role of the cytotoxic
active test compounds (1–4).
Available online 10 May 2010
Keywords:
Tributyltin(IV) compounds
Arylazobenzoate
Spectroscopy
X-ray crystallography
Cell lines
Anti-cancer drugs
Hydrophobicity
QSAR
Docking studies
© 2010 Elsevier Inc. All rights reserved.
1. Introduction
and Pd) has emerged that might exhibit comparable cytotoxic
properties accompanied by a different pattern of antitumor specificities
Cisplatin [1–4] was the first inorganic cancer chemotherapeutic
agent, and remains a front-line treatment for testicular, ovarian and
other cancers. The clinical effectiveness of cisplatin is limited by
considerable side effects and the emergence of drug resistance.
Consequently new platinum complexes such as carboplatin and
oxaliplatin have been approved for clinical use [5].
These advances have spurred a surge of investigations to identify
newinorganic agents for use in chemotherapy with improved specificity
and decreased toxic side effects. As a result, a great deal of interest in
other platinum and non-platinum metallodrugs (e.g., Sn, Ti, Au, Cu, Ru,
and by a more favorable pharmacological and toxicological profile [6]. In
addition, non-platinum metal-based antitumor agents have been
developed where the activity does not rely on direct DNA damage and
may involve proteins and enzymes. Metal-based compounds exhibit a
wide range of coordination numbers and geometries, significantly
increasing the possible spatial arrangements of the ligands and hence
the possibility of creating molecules with superior modalities of attack
to specific biological targets. Moreover, the redox potential of the metal
can interact with the balanced cellular redox state, modifying cell
viability either directly or through conversion of a rather inert
compound to an activated one, thus tuning the inherent toxicity of the
drug.
Among the non-platinum metal compounds with antitumor
activity, particular interest has focused on gold and tin derivatives,
⁎
Corresponding author. Tel.: +91 364 2722626; fax: +91 364 2721000.
E-mail addresses: basubaul@nehu.ac.in, basubaul@hotmail.com (T.S. Basu Baul).
0162-0134/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jinorgbio.2010.05.001

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
951
which have a common activity on mitochondria and a strong affinity to
thiol groups of proteins and enzymes [7–9]. As a result, a large number
of organotin(IV) derivatives have been prepared and tested in vitro and
in vivo, firstly against murine leukemia cell lines and then against
different panels of human cancer cell lines [10–12]. Several diorgano-
tin(IV) and triorganotin(IV) compounds showed high antiproliferative
activity in vitro against a variety of solid and hematologic cancers
[12,13]. Diorganotin(IV) and triorganotin(IV) terebates and lithoco-
lates, tested against a panel of seven human cancer cell lines, were
found to be highly active and more potent than cisplatin. In general,
tributyltin(IV) derivatives demonstrated higher efficacy than triphe-
nyltin(IV) and dibutyltin(IV) derivatives [14]. The tributyltin(IV)-3,4-
diaminobenzoate, 3,5-diaminobenzoate and 2-[4-(dimethylamino)
phenylazo]benzoate were also found to be promising cytostatic agents
in vitro when tested against human cell line A549 (lung adenocarci-
noma) [15]. The cytotoxicity of a tributyltin(IV) lupinylsulfide
hydrogen fumarate has also been studied and proved to be extremely
active when tested in vitro against a panel of tumor cell lines (MCF7,
MDA-MB-231 (breast), A2780, OVCAR-3 (ovary), DBTRG-0.5MG, U87
MG, U373 MG, A-172 (glioma) and a mouse glioma cell line (GL261))
and in vivo against P388 (myelomonocytic leukemia) and the B16–F10
(melanoma) cell lines [16].
We recently investigated the cytotoxic potential of triphenyltin
(IV) 2-[(E)-2-(aryl)-1-diazenyl]benzoates and dibutylbis{2-[(E)-2-
(aryl)-1-diazenyl]benzoato}tin(IV) and found to exhibit high activity
when tested in vitro against human tumor cell lines [17,18]. Among
these, the triphenyltin(IV) compounds were found to be better
performers than dibutyltin(IV) compounds. Further, the molecular
docking studies of these compounds indicated that the azo group
nitrogen atoms and formyl, carbonyl, ester and hydroxyl oxygen
atoms in the ligand moiety exhibit hydrogen bonding interactions
with the active site of the amino acids of various enzymes, such as
ribonucleotide reductase, thymidylate synthase and thymidylate
phosphorylase [17,18]. The high activity was attributed to the
presence of an azo group in the organotin(IV) complexes molecules
[17,18]. In this paper, we present the synthesis of a new series of
complexes where the 2/4-[(E)-2-(aryl)-1-diazenyl]benzoate ligands
are substituted with tributyltin(IV) (Scheme 1) in the expectation
that they will improve dissolution properties and thereby influence
cytotoxicity. The tributyltin(IV) complexes, Bu3SnL1–4H (1–4) were
characterized by spectroscopic techniques and crystal structure
determination of tributyltin(IV) complex 4. The molecular docking
of 1–4 has been investigated with some selected enzymes and
preliminary in vitro cytotoxicity data are reported for a panel of
human tumor cell lines consisting of A498 (renal cancer), EVSA-T
(mammary cancer), H226 (non-small-cell lung cancer), IGROV
(ovarian cancer), M19 MEL (melanoma), MCF-7 (mammary cancer)
and WIDR (colon cancer).
2. Experimental
2.1. Materials and methods
(Bu3Sn)2O, 2-methylphenol, 4-methylphenol, 4-tert-butylphenol
(Merck), anthranilic acid (Spectrochem), and p-aminobenzoic acid
(Hi Media) were used without further purification. The solvents used
in the reactions were of AR grade and were dried using standard
procedures. Toluene was distilled from sodium benzophenone ketyl.
2.2. Synthesis and characterization of ligands
Ligands 2-[(E)-(2-hydroxy-5-methylphenyl)diazenyl]benzoic acid
(L1HH′) [19], 2-[(E)-(5-tert-butyl-2-hydroxyphenyl)diazenyl]benzoic
acid (L2HH′) [20], 4-[(E)-(5-tert-butyl-2-hydroxyphenyl)diazenyl]ben-
zoic acid (L3HH′) [18] and 4-[(E)-(4-hydroxy-3-methylphenyl)diaze-
nyl]benzoic acid (L4HH′) [18]were prepared by the method described in
Scheme 1. Syntheses of ligands L1HH′–L4HH′, numbering protocol and the structures of tributyltin(IV) complexes Bu3SnL1–4H (1–4).

952
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
our earlier reports and purities were established from melting point,
elemental analysis and 1H NMR spectroscopy. Abbreviations used
for the reported 1H NMR spectra are as follows: s = singlet, d = doublet,
t = triplet, dd = doublet of doublets, dt = doublet of triplets, m =
multiplet.
crystals of the desired product. Yield: 28%. M.p.: 80–82 °C. Anal. calc.
for C26H38N2O3Sn: C, 57.27; H, 7.02; N, 5.14%. Found. C, 57.40; H, 6.99;
N, 5.25%. IR (cm−1) 1606 ν(OCO)asym 1H NMR (CDCl3): δH: ligand
.
skeleton: OH not detected, 8.18 [dd, 2.5 Hz, 8.0 Hz, 2H, H2/6], 7.87 [dd,
2.5 Hz, 8.0 Hz, 2H, H3/5], 7.79 [d, 2.5 Hz, 1H, H6′], 7.71 [dd, 2.5, 8 Hz,
1H, H2′], 6.91 [d, 8 Hz, 1H, H5′], 2.34 [s, 3H, CH3]; Sn–Bu skeleton: 1.68
[m, 6H, H1*], 1.38 [m, 12H, H2* and H3*], 0.92 [t, 9H, H4*] ppm. 13C
NMR (CDCl3); δ C: ligand skeleton: δ C: 171.9 [CO2], 16.4 [CH3], other
carbons: 158.4, 155.5, 147.2, 133.5, 131.5, 125.8, 125.5, 124.1, 122.5,
115.6; Sn–Bu skeleton (nJ(13C–119Sn, Hz)): 14.1 [C4*(nd)], 17.1 [C1*
(363)], 27.4 [C3*(70)], 28.2 [C2*(20)] ppm. 119Sn NMR (CDCl3):
121.7 ppm. 119Sn Mössbauer, mm s−1: δ=1.39, Δ=3.04, Γav =0.87;
ρ=2.18.
2.3. Synthesis and characterization of tributyltin(IV) complexes
2.3.1. Synthesis of Bu3SnL1H (1)
Compound 1 was prepared following the literature method of
reacting equimolar amounts of sodium 2-[(E)-(2-hydroxy-5-methyl-
phenyl)diazenyl]benzoate and Bu3SnCl in methanol. The microana-
lytical and NMR (1H and 13C, in CDCl3) data for the complex are in
agreement with previous results [19]. The data for 117Sn NMR
(CDCl3): 123.0 ppm and 119Sn Mössbauer, mm s− 1: δ = 1.50,
Δ=3.85 [19] are presented here for convenience of the discussion.
2.4. Physical measurements
Carbon, hydrogen and nitrogen analyses were performed with a
Perkin Elmer 2400 series II instrument. IR spectra in the range 4000–
400 cm−1 were obtained on a Perkin Elmer Spectrum BX series FT-IR
spectrophotometer with samples investigated as KBr discs. The 1H
and 13C spectra were recorded on a Bruker AMX 400 spectrometer and
measured at 400.13 and 100.62 MHz, respectively, while 119Sn NMR
spectra were recorded either on a Bruker AMX 400 or a Jeol GX 270
spectrometer and measured at 149.18 and 100.75 MHz, respectively.
The 1H, 13C and 119Sn chemical shifts were referenced to Me4Si, CDCl3
and Me4Sn set at 0.00, 77.0 and 0.00 ppm respectively. The Mössbauer
spectra were recorded with a conventional spectrometer operating in
the transmission mode. The source was Ca119SnO3 (Ritverc GmbH, St.
Petersburg, Russia; 10 mCi), moving at room temperature with
constant acceleration in a triangular waveform. The driving system,
multi-channel analyser, proportional counter gas detector and the
related electronics were purchased from Takes (Ponteranica, Bergamo,
Italy). The solid absorber samples, containing about 0.5 mg cm−2 of
119Sn, were held between aluminum foil windows at 77.3 K in a NRD-
1258-DMB liquid–nitrogen cryostat (Cryo Industries, USA). The
velocity was calibrated using a 10 mCi 57Co Mössbauer source and
4 μm thick α57Fe foil as the absorber (both Ritverc GmbH, St.
Petersburg, Russia). The isomer shifts are relative to room temperature
Ca119SnO3.
2.3.2. Synthesis of Bu3SnL2H (2)
Compound 2 was synthesized by mixing L2HH′ (0.40 g, 1.34 mmol)
and (Bu3Sn)2O (0.40 g, 0.67 mmol) in 40 ml of anhydrous toluene, in a
100 ml flask equipped with a Dean–Stark moisture trap and a water
cooled condenser. The reaction mixture was refluxed for 7 h and the
solvent was then removed by distillation. A thick viscous mass was
dissolved in small amount of anhydrous benzene, concentrated slowly
on a hot plate, cooled to room temperature and petroleum ether
added. The pasty mass obtained was triturated in ice-cold conditions
which afforded yellow-orange crystalline material. Yield: (0.35 g,
44%). M.p.: 46–48 °C. Anal. calc. for C29H44N2O3Sn: C, 59.30; H, 7.55; N,
4.77 %. Found. C, 58.30; H, 7.80; N, 5.08%. IR (cm−1) 1578 ν(OCO)asym
.
1H NMR (CDCl3): δH: Ligand skeleton: 12.6 [brs, 1H, OH], 7.96 [d, 8 Hz,
1H, H3], 7.82 [m, 2H, H6/H6′], 7.47 [t, 8 Hz, 1H, H5], 7.38 [t, 8 Hz, 1H,
H4], 7.27 [dd, 2.5, 8 Hz, 1H, H4′], 6.83 [d, 8 Hz, 1H, H3′], 1.30 [s, 9H,
CH3]; Sn–Bu skeleton: 1.61 [m, 6H, H1*], 1.30 [m, 12H, H2* and H3*],
0.82 [t, 9H, H4*] ppm. 13C NMR (CDCl3); δ C: Ligand skeleton: 170.3
[CO2], 149.5 [C2′], 148.2 [C1], 140.7 [C1′], 136.6 [C4′], 130.8 [C6′], 130.6
[C5], 129.7 [C3], 128.9 [C4], 128.8 [C5′], 128.7 [C2], 117.1 [C3′], 114.7
[C6], 32.9 [C-7′], 30.4 [CH3]; Sn–Bu skeleton (nJ(13C–119Sn, Hz)): 13.9
[C4*(nd)], 15.7 [C1*(363)], 26.0 [C3*(70)], 26.8 [C2*(20)] ppm. 119Sn
NMR (CDCl3): 124.9 ppm. 119Sn Mössbauer, mm s−1: δ=1.51,
Δ=3.86, Γav =0.94; ρ=2.44.
2.5. Experimental protocol and cytotoxicity tests
2.3.3. Synthesis of Bu3SnL3H (3)
Compound 3 was prepared analogously by following the method
and conditions described for 2 and using L3HH′ and (Bu3Sn)2O. The
pasty mass was dissolved in a benzene–hexane mixture and upon
cooling in a refrigerator furnished orange crystalline material of 3.
Yield: 25%. M.p.: 40–42 °C. Anal. calc. for C29H44N2O3Sn: C, 59.30; H,
7.55; N, 4.77 %. Found. C, 58.42; H, 7.70; N, 5.01%. IR (cm−1) 1570 ν
The in vitro cytotoxicity test of tributyltin(IV) compounds 1–4 was
performed using the SRB test for the estimation of cell viability. The
human cancer cell lines examined in the present study were WIDR,
M19 MEL, A498, IGROV and H226, belonging to the currently used
anticancer screening panel of the NCI, USA [21]. The MCF7 cell line is
estrogen receptor (ER)+/progesterone receptor (PgR)+and the cell
line EVSA-T is (ER)–/(Pgr)–. Prior to the experiments, a mycoplasma
test was carried out on all cell lines and found to be negative. All cell
lines were maintained in a continuous logarithmic culture in RPMI
1640 medium with Hepes and phenol red. The medium was
supplemented with 10% FCS, 100 μg/ml penicillin and 100 μg/ml
streptomycin. The cells were mildly trypsinized for passage and for
use in the experiments. RPMI and FCS were obtained from Gibco
(Invitrogen, Paisley, Scotland). SRB, DMSO, penicillin and streptomy-
cin were obtained from Sigma (St. Louis MO, USA), TCA and acetic acid
from Baker BV (Deventer, NL) and PBS from NPBI BV (Emmer-
Compascuum, NL).
(OCO)asym.
1H NMR (CDCl3): δH: ligand skeleton: δ H: 12.6 [brs, 1H,
OH], 8.16 [dd, 2.5 Hz, 8.0 Hz, 2H, H2/6], 7.89 [m, 3H, H3/5 and H6′],
7.37 [dd, 2.5, 8 Hz, 1H, H4′], 6.92 [d, 8 Hz, 1H, H3′], 1.38 [s, 9H, CH3];
Sn–Bu skeleton: 1.69 [m, 6H, H1*], 1.38 [m, 12H, H2* and H3*], 0.95 [t,
9H, H4*] ppm. 13C NMR (CDCl3); δ C: Ligand skeleton: 167 [CO2], 34.0
[C-7′], 31.4 [CH3], other carbons: 152.8, 150.8, 142.3, 137.1, 133.8,
131.4, 131.2, 129.9, 121.7, 117.9; Sn–Bu skeleton (nJ(13C–119Sn, Hz)):
13.6 [C4*(nd)], 16.6 [C1*(353)], 27.0 [C3*(70)], 27.9 [C2*(20)] ppm.
119Sn NMR (CDCl3): 120.0 ppm. 119Sn Mössbauer, mm s−1: δ=1.36,
Δ=3.60, Γav =1.00; ρ=2.65.
2.3.4. Synthesis of Bu3SnL4H (4)
The test compounds 1–4 and reference compounds were dissolved
at a concentration of 5 mg/ml in DMSO. The compounds were
subsequently diluted to a final concentration of 250,000 ng/ml in
full medium. Cytotoxicity was estimated by the microculture
sulforhodamine B (SRB) test [22].
Compound 4 was prepared analogously by following the method
and conditions described for 2 and using L4HH′ and (Bu3Sn)2O. A thick
viscous mass was obtained that was then dissolved in a small amount
of hexane and kept in a refrigerator for two weeks. The crystalline
material obtained was extracted in hot hexane and filtered. The
filtrate upon slow evaporation at room temperature yielded orange
The experiment was started on day 0. On day 0, 150 μl of
trypsinized tumor cells (1500–2000 cells/well) were plated in 96-

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
953
wells flat-bottomed micro-titer plates (Cellstar, Greiner Bio-one). The
plates were pre-incubated for 48 h at 37 °C, 5% CO2 to allow the cells
to adhere to the bottom. On day 2, a three-fold dilution sequence of
ten steps was made in full medium, starting with the 250,000 ng/ml
stock solution. Every dilution was used in quadruplicate by adding
50 μl to a column of four wells. This procedure results in the highest
concentration of 625,000 ng/ml being present in column 12. Column 2
was used for the blank. Medium was added to column 1 to diminish
interfering evaporation.
On day 7, the incubation was terminated. Subsequently, the cells
were fixed with 10% trichloroacetic acid in PBS and stored at 4 °C for
an hour. After three washings with tap water, the cells were stained
for at least 15 min with 0.4% SRB dissolved in 1% acetic acid. After
staining, the cells were washed with 1% acetic acid to remove the
unbound stain. The plates were air-dried and the bound stain was
dissolved in 150 μl 10 mM tris-base. The absorbance was measured at
540 nm using an automated microplate reader (Labsystems Multiskan
MS). The data were used for construction of concentration–response
curves and determination of the ID50 values using Deltasoft 3
software.
deviation) [29]. Each QSAR model includes 95% confidence limits for
each term in parentheses. Statistical diagnostics and internal
validation (cross-validation and Y-randomization) tests have validat-
ed all the QSAR models. The external validation test was not
performed due to the small data sets.
2.7. X-ray crystallography
Single crystals of Bu3SnL4H (4) suitable for an X-ray crystal
structure determination were obtained by slow evaporation of a
hexane solution. Crystal data for 4: C26H38N2O3Sn, M=545.27,
monoclinic, P21/n, a=9.7101(8), b=18.6054(16), c=14.8477(13)
Å, β=102.630(2)°, V=2617.5(4) Å3, Z=4, Dx =1.384, F(000)=
1128, μ=1.004 mm−1. An orange platelet (0.25×0.13×0.04 mm)
was mounted on a cryo loop and placed directly in the cold stream
(130 K) of dinitrogen from an Oxford Cryosystems Cryostream 700
cooler. Intensity data were collected with a Bruker Smart APEX CCD
(Mo-Kα radiation, λ=0.71073 Å, graphite monochromator) area
detector on a D8 goniometer in the ω scan mode. A total of 33,405
reflections were collected within θmax =27.3° and a multi-scan
absorption correction [30] was applied. Merging of symmetry
equivalent reflections (Rint =0.064) resulted in 5866 unique data,
4351 of these with IN2σ(I). The structure was solved by a Patterson
synthesis [31] and refined on F2 [31]. Carbon atoms C24–C26
associated with one of the n-butyl groups were disordered, so two
alternative orientations of this moiety were populated in the ratio
0.716(8):0.284(8) and refined with geometry restraints. Isotropic
displacement parameters were assigned to the carbon atoms in this
disordered part of the structure and anisotropic displacement
parameters to all other non-hydrogen atoms. Hydrogen atoms were
treated as riding in idealized positions. Convergence was reached for
R=0.0569, wR2=0.0844 and GOF=1.050 for all 5866 data and 292
variables. The residual electron density in a Fourier difference
synthesis amounted to 1.2 e·Å−3 in the disordered n-butyl group. A
displacement ellipsoid plot at 50% probability of a molecule in the
crystal of 4 is shown in Fig. 2 [30].
The variability of the in vitro cytotoxicity test depends on the cell
lines used and the serum applied. With the same batch of cell lines
and the same batch of serum the inter-experimental coefficient of
variation (CV) is 1–11% depending on the cell line and the intra-
experimental CV is 2–4%. These values may be higher with other
batches of cell lines and/or serum.
2.6. Quantitative structure–activity relationship (QSAR) methods
The C-QSAR program [23] was used to derive all QSAR models by
using multiregression analyses (MRA). In this program, the selection of
descriptors has been made on the basis of permutation and correlation
matrices among the descriptors in order to avoid collinearity
problems. Details of the C-QSAR program, the search engine, the
choice of parameters, and their use in the development of QSAR
models, have been discussed in previous publications [24,25]. The in
vitro cytotoxicity data (ID50; ng/ml) of test compounds 1–4 along with
some related organotin(IV) azo-compounds (5–9) and some standard
drug molecules against the seven human tumor cell lines are listed in
Table 2. It is preferred in the QSAR analysis that the more effective
compounds should have a higher ‘activity’ and not a lower. Thus, it is
very common to transform the concentration of a desired effect ‘C’ to
an activity (negative logarithm of the concentration) by the following
equation: A=−log C=log 1/C, where C is in molar concentration. This
is the reason, the cytotoxicity data (Table 2) was first converted into
molar concentration (C1–C7), where C1–C7 represent the ID50 (M)
values of the organotin(IV) compounds 1–9 along with five anticancer
drugs (doxorubicin: DOX, cisplatin: CDDP, 5-fluorouracil: 5-FU,
methotrexate: MTX and etoposide: ETO) against the seven human
tumor cell lines (A498, EVSA-T, H226, IGROV, M19MEL, MCF-7 and
WIDR). Biological parameters for the development of QSAR models are
defined by their subsequent dependent variables i.e. activities (log 1/
C1, log 1/C2, log 1/C3, log 1/C4, log 1/C5, log 1/C6, and log 1/C7) (see
Tables 3 and 4). log P is the calculated partition coefficient of a
compound in n-octanol/water and a measure of its hydrophobicity.
HBD is the number of hydrogen bond donors. The value of log P and
HBD were calculated using the ACD PhysChem Suite [26].
2.8. Computational methods
The molecular structures and geometries of the tributyltin(IV)
compounds (1–4) were fully optimized using the PM6 semiempirical
quantum chemistry method [32–36] using the Program CP2K 2.1.7
(Development Version) working with quickstep method [37]. Dock-
ings of compounds (1–4) in the active sites of various enzymes are
performed using ArgusLab 4.0.1. [32,38–40]. This program was also
applied for visualization and molecular modeling of the compounds.
The three dimensional coordinates of key enzymes, such as ribonu-
cleotide reductase (pdb ID: 4R1R), thymidylate synthase (pdb ID:
2G8D), thymidylate phosphorylase (pdb ID: 1BRW) and topoisome-
rase II (pdb ID: 1QZR) were obtained from the Internet at the Research
Collaboratory for Structural Bioinformatics (RCSB) Protein Data Bank.
The docking program implements an efficient grid-based docking
algorithm that approximates an exhaustive search within the free
volume of the binding site cavity. The conformational space was
explored by the geometry optimization of the flexible ligand (rings
are treated as rigid) in combination with the incremental construction
of the ligand torsions. Thus, docking occurs between the flexible
ligands part of the compounds and enzymes. The ligand orientation is
determined by a shape scoring function based on Ascore and the final
positions are ranked by lowest interaction energy values. Prior to
docking, the ground state was optimized using the PM6QM
implemented in the geometry optimization module of program
package to confirm that there was no significant divergence in the
complex conformations due to crystal packing effects.
In QSAR models, n is the number of data points, r2 is the square of
the correlation coefficient and represents the goodness of fit, q2 is the
cross-validated r2, s is the standard deviation, Q is the quality factor,
and F is the Fischer ratio. The cross-validated r2 (q2) is obtained by
using the leave-one-out (LOO) procedure [27]. In each QSAR equation,
the value of F in parenthesis refers to their literature value at 95% level
[28]. Compound was deemed to be an outlier on the basis of the
deviation between observed and predicted activities from the
respective QSAR model (obsd–predN2 s, where s is the standard

954
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
Table 1
3. Results and discussion
3.1. Synthesis and spectroscopy
Selected bond lengths (Å) and angles (°)a obtained from energy minimized structures
of the tributyltin(IV) complexes 1–4 and from X-ray crystallography of tributyltin(IV)
complex 4.
Ligands L1HH′–L4HH′ were prepared by reacting the appropriate
ortho/para-carboxybenzenediazonium chloride with 2-methylphenol,
4-methylphenol or 4-tert-butylphenol in alkaline solution under cold
conditions [18–20]. Reaction of L1HH′–L4HH′ with (Bu3Sn)2O at a 2:1
molar ratio in anhydrous toluene gave tributyltin(IV) complexes
Bu3SnL1H (1), Bu3SnL2H (2), Bu3SnL2H (3) and Bu3SnL3H (4),
respectively (Scheme 1).
In solution (CDCl3), complexes 1–4 exhibited a single sharp
119Sn NMR chemical shift in the range 120–125 ppm that is indicative
of a four-coordinate environment at the Sn atom [19,41–43]. The
13C NMR spectra displayed the expected carbon signals due to
ligand and Sn–Bu skeletons. Assignment of the tri-n-butyl 13C reso-
nances was based on the nJ(13C–119/117Sn) coupling constants. It is
known that nJ(13C–119/117Sn) coupling constants decrease in the order
1J(13C–119/117Sn)NN3J(13C–119/117Sn)N2J(13C–119/117Sn), depending on
ligand properties, and increase with increasing of coordination number
at the tin atom [44,45]. The 1J(13C–119/117Sn) coupling constants for
complexes 1–4 in CDCl3 solution are in the range 336–362 Hz,
confirming that in solution these compounds have a tetrahedral Sn
atom. The 1H NMR integration values were completely consistent with
the formulation of the products. Furthermore, the signals between 12
and 13 ppm observed for CO2H of the ligands [18–20] were found to be
absent in complexes 1–4, confirming the bonding of the carboxylato
group to the Sn atom.
Bond lengths (Å)
and angles (°)
1
2
3
4
X-ray
Sn(1)–O(1)
Sn(1)–O(2)
Sn(1)–C(15)
Sn(1)–C(19)
Sn(1)–C(23)
O(1)–C(1)
O(2)–C(1)
N(1)–N(2)
O(1)–Sn(1)–O(2)
O(1)–Sn(1)–C(15)
O(1)–Sn(1)–C(19)
O(1)–Sn(1)–C(23)
O(2)–Sn(1)–C(15)
O(2)–Sn(1)–C(19)
O(2)–Sn(1)–C(23)
C(15)–Sn(1)–C(19)
C(19)–Sn(1)–C(23)
C(15)–Sn(1)–C(23)
Sn(1)–O(1)–C(1)
Sn(1)–O(2)–C(1)
O(1)–C(1)–O(2)
2.161
2.952
2.150
2.154
2.149
1.321
1.234
1.260
2.166
2.939
2.150
2.154
2.149
1.321
1.235
1.260
2.149
3.028
2.151
2.154
2.150
1.321
1.228
1.260
2.142
3.064
2.151
2.155
2.149
1.322
1.227
1.256
2.083(2)
2.982(2)
2.129(3)
2.129(3)
2.143(4)
1.295(4)
1.222(4)
1.225(4)
47.8(7)
104.3(11)
97.5(12)
105.2(12)
84.3(11)
144.4(11)
76.8(11)
117.2(14)
112.0(15)
117.3(14)
115.6(2)
73.8(17)
122.8(3)
47.7
47.8
102.1
94.8
103.2
82.7
142.3
74.8
115.5
117.3
118.1
114.9
79.6
46.8
101.8
94.2
104.9
86.7
139.7
73.3
116.1
117.1
117.4
116.9
76.4
46.2
101.6
95.9
103.7
86.1
140.9
73.1
115.7
117.1
117.6
118.0
75.4
102.1
95.2
103.3
84.0
142.3
73.9
115.3
117.1
118.3
115.4
79.2
117.7
117.6
119.9
120.3
a
For energy minimized structures of compounds 1 and 4, refer to Figs. 4 and 5 and for
compounds 2 and 3, refer to Figs. S1 and S2. Refer to Fig. 1 for atom numbering schemes.
and Ph; X=Me and Br) [42]. The other carboxylate O atom of the
benzoate ligand also coordinates weakly to the Sn atom with the Sn
(1)⋯O(2) distance being 2.982(2) Å. This interaction is the cause of the
distortion of the tetrahedral primary coordination sphere, but the Sn
(1)⋯O(2) distance is considered to be too long for the Sn atom to be
described as truly five-coordinate. In addition, the bond angles around
the Sn atom in 4 are more consistent with a tetrahedral environment
than with a trigonal bipyramidal five-coordinate environment.
Compound 4 exhibits the usual distorted tetrahedral structural
motif in which the phenoxy hydrogen atom of the carboxylate ligand
forms an intermolecular hydrogen bond with the non-coordinating
carbonyl O atom of the carboxylate group of a neighbouring molecule;
the O⋯O donor⋯acceptor distance in this interaction amounts to 2.778
(3)Å. Fig. 2 (see Scheme 1 for schematic representation) illustrates
how this translation results in infinite chains extending along [010].
Mössbauer spectroscopy was used to obtain insight into the solid-
state structures of complexes 2 and 3, for which no suitable crystals
could be obtained for X-ray crystallography. Typically, one doublet
spectra with narrow average full width at half maximum of the
resonant peaks (Γav, 0.87–1.00 mm s−1), characteristic of occurrence
of unique tin coordination site, were obtained for all four compounds.
From the deconvolution of the obtained spectra, isomer shifts (δ) of
1.51 (2), 1.36 (3), and 1.39 (4) mm s−1, typical of organotin(IV)
The crystal structure of Bu3SnL1H (1) has been reported previously
and is thus available for comparison [19]. In 1, the carboxylate O
atoms of a single aryl ligand bridge two Sn atoms and the pattern then
repeats itself to give a continuous single-stranded polymeric
structure, as illustrated in Scheme 1. The Sn atoms have a slightly
distorted trans-R3SnO2 trigonal bipyramidal coordination geometry
with equatorial butyl groups and two axial carboxylate O atoms, one
being from each of two aryl ligands. In contrast, the molecular
structure of Bu3SnL4H (4) is best described as four-coordinate with a
distorted C3O tetrahedral geometry involving one of the carboxylate O
atoms and three C atoms from the butyl ligands. The molecular
structure of 4 is shown in Fig. 1, with selected geometric parameters
collected in Table 1). Similar tetrahedral geometry about the Sn atom
was observed in (Ph3Sn[O2CC6H4(N=NC6H3-2-OH-5-Me)-o]) and its
acetone solvated complex [46,47], (Ph3Sn{O2CC6H3-p-OH[N=N
(C6H4-X)]} (X=H, 2-Me, 3-Me, 4-Me, 4-OMe, 4-Cl) [41,48,49] and
(R3Sn[O2CC6H4{N=N(C6H3-4-OH(C(H)=NC6H4X-4))}-o]) (R=Bu
derivatives [50], were extracted and a similar value (δ=1.51 mm s−1
was observed earlier for 1 [19].
)
The tributyltin(IV) complexes 2 and 3 exhibited very similar 119Sn
Mössbauer spectra characterized by |Δexp| values of 3.86 and
3.60 mm s− 1, respectively, which are characteristic of trigonal
bipyramidal structures with butyl groups in equatorial position and
axial electronegative ligands. A similar value (Δ̣
=3.85 mm s−1) was
found earlier for the complex 1 which was characterized crystallo-
graphically [19] and displayed a distorted trans-R3SnO2 trigonal
bipyramidal coordination geometry in the polymer. The |Δexp| values
for the complexes 1–3 are in the same order of magnitude, suggesting
that they assume the same structural motif as depicted in Scheme 1.
The crystallographic data presented above for complex 4 show that the
caboxylate moiety acts as a monodentate ligand, giving rise to a
distorted C3O tetrahedral geometry involving one of the carboxylate O
atoms and the C atoms from the three butyl ligands. On this basis, a
Fig. 1. Displacement ellipsoid plot [30] (50% probability) of a molecule in Bu3SnL4H (4)
showing the numbering scheme. H3 has been drawn with arbitrary radius, other
hydrogen atoms and the minority conformer of the disordered n-butyl group C23–C26
have been omitted.

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
955
value and C–Sn–O bond angle [51], the calculated C–Sn–O angle in 4
was found to be 109°, in reasonable conformity with that derived from
X-ray crystallography.
The solid-state IR spectra of the complexes display a strong, sharp
band due to the [νasym(OCO)] stretching vibration at ∼1575 cm−1 for 1
and 2 and ∼1605 cm−1 for 3 and 4 as a result of carboxylatecoordination
to the Sn atom [41,42,48,49]. This compares to bands at ∼1700 cm−1
(L1–2HH′) [20] and ∼1685 cm−1 (L3–4HH′) [18] for the free ligands. The
assignment of the symmetric [νsym(OCO)] stretching vibration band
could not be made owing to the complex pattern of the spectra.
3.2. Quantum chemical calculations
The geometries of the tributyltin(IV) complexes 1–4 were fully
optimized using the semi-empirical method (PM6). Harmonic
frequency calculations were performed for all the stationary points
to characterize their nature and to ensure that the optimized
structures correspond to global minima. The molecular structures of
1 and 4 obtained after full geometry optimization at PM6 levels are
shown in Figs. 4 and 5 (for the structures of 2 and 3 refer to S1 and S2),
respectively, while the optimized geometric parameters are collected
in Table 1. The X-ray experimental geometrical parameters for 4 are
also included in Table 1, which match closely with that of calculated
values for 4. Most of the geometric parameters for 1–4 are found to be
insensitive to the nature of the substituents R (see Scheme 1). Since
the basic structures and coordination geometry of tributyltin(IV)
complexes 1–4 are similar despite the differently substituted ligands,
it may be expected that the ligand properties have a direct influence
on the stability of the corresponding tributyltin(IV) complexes, as
well as on their cytotoxic activity (see below).
Fig. 2. A segment of the chain formed by hydrogen bonds in Bu3SnL4H (4). The view
direction is slightly inclined to [001] for clarity. Symmetry operator i=−1/2−x, 1/2+y,
1/2−z.
point charge model formalism [51] has been applied for complex 4,
using {Bu}tet =−1.13 mm s−1 [52], {COO−}tet bridging=0.1145 mm s−1
calculated from [COO−]tba [53] (tet = tetrahedral, tba = trigonal
bipyramidal axial), considering a tetrahedral structure of the complex
where three positions are occupied by the butyl groups and the fourth
by an oxygen donor of a bridging carboxylate (see Fig. 3) and the Δcalcd
3.3. Cytotoxicity studies
The in vitro cytotoxic results for organotin(IV) compounds of
formulations Bu3SnL1–4H (1–4), Ph3SnL5–6H (5–6) [17] and Bu2Sn
(L1–3H)2 (7–9) [18] were investigated across the panel of human
tumor cell lines A498, EVSA-T, H226, IGROV, M19 MEL, MCF-7 and
WIDR, as per NCI protocol (Table 2). The cytotoxicity data (ID50) for
the test compounds (1–4) are of the same order of magnitude and
the change of ligand substitution does not improve the cytotoxic
activity significantly. Upon closer inspection of the ID50 values of 1–4,
one can see that Bu3SnL1H (1) was more cytotoxic than the other test
compounds (2–4). A marginal decrease in activity was seen on
increasing the steric bulk further at the coupling site of the ligand
framework by the addition of a tert-butyl group, as in compound 2.
The change of tributyltin ester from ortho- (2) to para-position (3)
did not yield any change in cytotoxic activity except for the EVSA-T
(mammary cancer) cell line, which showed substantial increase in
activity (ID50 =27 ng/ml). Changing substituents i.e. tributyltin ester
and hydroxyl group from ortho- (1) to para-position (4) have
demonstrated lesser activity. Thus, it can be inferred that the
substituents in the ligand skeleton play a vital role in determining
the cytotoxic potentials of a compound. In general, the cytotoxic
results of compounds 1–4 are undoubtedly far superior to CDDP, 5-
FU and ETO and related dibutyltin(IV) compounds 7–9. While test
compounds 1–4 in this investigation displayed comparable activity to
triphenyltin(IV) compounds 5 and 6 [17] across a panel of seven
human tumor cell lines, compounds 1–4 have the advantage of
higher solubility. The activity has been attributed to the tetrahedral
geometry of the complexes in solution, as well as the presence of an
azo functionality in the ligand framework, and was subsequently
confirmed by docking results [17,18]. Ruthenium complexes contain-
ing azo-ligands have also shown higher activity in comparison to the
complexes with non-azo-ligands [54,55]. Promising cytotoxic activity
was also noted for the related tributyltin(IV) 2-[4-(dimethylamino)
phenylazo]benzoate, Bu3Sn[O2CC6H4(N = NC6H4N(CH3)2-4)-o],
value was found to be −2.97 mm s−1
.
The |Δexp| value of 3.04 mm s− 1 observed for complex 4
corresponds well with that of the Δcalcd value and is within the
range of 2.3–3.0 mm s−1, characteristic for four-coordinate tetrahe-
dral geometry [50]. Further, the |Δexp| value is comparable with that of
(Bu3Sn[O2CC6H4{N=N(C6H3-4-OH(C(H)=NC6H4Br-4))}-p]), which
has also been characterized crystallographically [42]. The ratio of
the quadrupole splitting to isomer shift value (ρ=|Δexp|/δ) can be
used to distinguish between the different coordination states of the
central tin atom [50]. Tributyltin(IV) complexes 1, 2 and 3 have ρ
values in the range 2.44–2.65, suggestive of a five-coordinated tin
geometry, while complex 4 has a ρ value of 2.18, indicative of a four-
coordinated tin environment, as reflected in the crystal structure, and
in agreement with reported ρ values for similar tetrahedral tributyltin
(IV) complexes [35]. Using the Parish relationship between the |Δexp
|
Fig. 3. A line diagram showing chain structure of Bu3SnL4H (4).

956
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
Fig. 4. The structure of Bu3SnL1H (1) obtained after full geometry optimization.
which also contains an azo group, when tested in vitro against Human
cell line A549 (lung adenocarcinoma) [15]. Thus, the cytotoxicity data
of compound 1, together with its better solubility, suggest it might be a
promising candidate for further in vitro and in vivo studies after
appropriate modifications. The cytotoxicity data of compounds 1–4
indicate that structural variation of the Sn–R and L–R skeletons
influence the activity. However, the reason behind factors influencing
the activity of these compounds is yet to be fully understood and more
in vitro and in vivo studies are needed in order to derive structure–
activity relationships for these kinds of complexes.
QSAR for the cytotoxicity of compounds 1–9 against EVSA-T cancer
cell line (Table 3)
log 1= C2 = –0:13ðꢀ0:09Þ log P + 7:71ðꢀ0:66Þ
ð2Þ
n = 7; r2 = 0:729; s = 0:155; q2
F1;5 = 13:450ð6:608Þ
= 0:557; Q = 5:510;
Outlier: compounds 3 and 7
QSAR for the cytotoxicity of compounds 1–9 against H226 cancer
cell line (Table 3)
3.4. QSAR Studies
log 1= C3 = –0:16ðꢀ0:05Þ log P + 7:63ðꢀ0:41Þ
n = 9; r2 = 0:882; s = 0:120; q2 = 0:831; Q = 7:825;
F1;7 = 52:322ð5:591Þ
ð3Þ
From the data in Table 3, the following QSAR models 1–7 were
developed:
QSAR for the cytotoxicity of compounds 1–9 against A498 cancer
cell line (Table 3)
QSAR for the cytotoxicity of compounds 1–9 against IGROV cancer
cell line (Table 3)
log 1= C1 = –0:17ðꢀ0:06Þlog P + 7:65ðꢀ0:46Þ
log 1= C4 = –0:18ðꢀ0:07Þ log P + 7:65ðꢀ0:52Þ
n = 8; r2 = 0:864; s = 0:129; q2 = 0:787; Q = 7:209;
F1;6 = 38:118ð5:987Þ
ð4Þ
n = 8; r2 = 0:876; s = 0:113; q2 = 0:797; Q = 8:283;
ð1Þ
F1;6 = 42:387ð5:987Þ
Outlier: compound 7
Outlier: compound 7
Fig. 5. The structure of Bu3SnL4H (4) obtained after full geometry optimization.

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
957
Table 2
In vitro ID50 values (ng/ml) of test compounds 1–4, related organotin(IV) azo-compounds 5–9 and some standard drugs against seven human tumor cell lines.a
Test compoundb
Cell lines
A498
EVSA-T
H226
IGROV
M19 MEL
MCF-7
WIDR
Bu3SnL1H (1)
Bu3SnL2H (2)
Bu3SnL3H (3)
Bu3SnL4H (4)
Ph3SnL5H (5)c [17]
Ph3SnL6H (6)c [17]
Bu2Sn(L2H)2 (7)c [18]
Bu2Sn(L3H)2 (8)c [18]
Bu2Sn(L4H)2 (9)c [18]
DOX
162
176
177
182
101
103
429
1169
382
90
1503
143
37
1314
b3.2
97
100
27
101
41
148
165
167
163
104
101
816
1140
497
199
645
340
2287
3934
214
253
269
239
109
101
617
1489
449
60
229
297
7
580
b3.2
118
126
127
125
103
104
374
901
294
16
711
442
23
505
b3.2
113
120
112
118
92
106
105
105
106
104
95
896
2125
551
11
576
225
b3.2
150
b3.2
49
78
134
371
134
8
493
475
5
236
416
162
10
653
750
18
CDDP
5-FU
MTX
ETO
317
b3.2
2594
b3.2
TAX
b3.2
a
Abbreviation: DOX, doxorubicin; CDDP, cisplatin; 5-FU, 5-fluorouracil; MTX, methotrexate; ETO, etoposide and TAX, paclitaxel.
Standard drug reference values are cited immediately after the test compounds under identical conditions. For compounds 5 and 6, the reference values for CCDP were 2253, 422,
b
3269, 169, 558, 699 and 967 for A498, EVSA-T, H226, IGROV, M19 MEL, MCF-7 and WIDR cell lines, respectively.
c
Reported triphenyltin(IV) and dibutyltin(IV) compounds (5–9) have been included for comparison; see refs. 17,18: LH is a carboxylate residue where L5H, 2-[(E)-2-(4-hydroxy-
5-methylphenyl)-1-diazenyl]benzoate and L6H, 2-[(E)-2-(3-formyl-4-hydroxyphenyl)-1-diazenyl]benzoate while all other LH are described in Scheme 1.
QSAR for the cytotoxicity of compounds 1–9 against M19 MEL
QSAR for the cytotoxicity of compounds 1–9 against MCF-7 cancer
cancer cell line (Table 3)
cell line (Table 3)
log 1= C5 = –0:15ðꢀ0:06Þ log P + 7:64ðꢀ0:42Þ
n = 8; r2 = 0:874; s = 0:104; q2 = 0:664; Q = 8:990;
F1;6 = 41:619ð5:987Þ
ð5Þ
log 1= C6 = –0:10ðꢀ0:05Þ log P + 7:36ðꢀ0:32Þ
n = 8; r2 = 0:823; s = 0:080; q2 = 0:587; Q = 11:338;
F1;6 = 27:898ð5:987Þ
ð6Þ
Outlier: compound 7
Outlier: compound 7
Table 3
Biological and hydrophobicity (log P) parameters of compounds 1–9 used to derive QSARs 1–7.
No.
Compound
log1/C1 (Eq. (1))
log1/C2 (Eq. (2))
log1/C3 (Eq. (3))
Obsd.
Pred.
Δ
Obsd.
Pred.
Δ
Obsd.
Pred.
Δ
1
1
2
3
4
5
6
7
8
9
6.53
6.52
6.52
6.48
6.78
6.78
6.29
5.85
6.29
6.65
6.42
6.40
6.63
6.70
6.72
5.92
5.88
6.35
−0.12
0.10
0.12
−0.15
0.08
0.06
0.37
−0.03
−0.06
6.75
6.77
7.34
6.73
7.17
7.10
6.79
6.35
6.74
6.93
6.76
6.74
6.92
6.97
6.99
6.37
6.34
6.71
−0.18
0.01
0.60
−0.19
0.20
0.11
0.42
0.01
0.03
6.57
6.55
6.55
6.52
6.76
6.79
6.01
5.86
6.17
6.65
6.43
6.41
6.63
6.70
6.73
5.95
5.91
6.37
−0.08
0.12
0.14
−0.11
0.06
0.06
0.06
−0.05
−0.20
2
3b
4
5
6
7a,b
8
9
No.
Compound
log1/C1 (Eq. (4))
log1/C2 (Eq. (5))
log1/C3 (Eq. (6))
Obsd.
Pred.
Δ
Obsd.
Pred.
Δ
Obsd.
Pred.
Δ
1
2
3
4
5
6
7a
8
9
1
2
3
4
5
6
7
8
9
6.41
6.37
6.34
6.36
6.74
6.79
6.13
5.74
6.22
6.56
6.32
6.30
6.54
6.62
6.64
5.78
5.73
6.25
−0.15
0.05
0.04
−0.18
0.12
0.15
0.35
0.01
−0.03
6.66
6.67
6.67
6.64
6.77
6.77
6.34
5.96
6.40
6.73
6.52
6.51
6.71
6.78
6.80
6.07
6.03
6.46
−0.07
0.15
0.16
−0.07
−0.01
−0.03
0.27
−0.07
−0.06
6.68
6.69
6.72
6.66
6.82
6.90
6.54
6.30
6.66
6.78
6.65
6.64
6.77
6.81
6.82
6.36
6.34
6.61
−0.10
0.04
0.08
−0.11
0.01
0.08
0.18
−0.04
0.05
No.
Compound
log1/C7 (Eq. (7))
Obsd.
log P
Pred.
Δ
1
2
3
4
5
6
7
8
9
1
2
3
4
5
6
7
8
9
6.71
6.75
6.75
6.71
6.76
6.81
5.97
5.59
6.13
6.77
6.48
6.46
6.75
6.84
6.87
5.84
5.79
6.40
−0.06
0.27
0.29
−0.04
−0.08
−0.06
0.13
−0.20
−0.27
5.94
7.27
7.38
6.05
5.63
5.49
10.23
10.46
7.66
a
Not included in the derivation of QSAR 1, 4, 5, 6.
Not included in the derivation of QSAR 2.
b

958
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
QSAR for the cytotoxicity of compounds 1–9 against WIDR cancer
QSAR for the cytotoxicity of compounds 1–9 along with five
cell line (Table 3)
anticancer drugs (DOX, CDDP, 5-FU, MTX, and ETO) against H226
cancer cell line (Table 4)
log 1 = C3 = 0:38ðꢀ0:12Þlog P – 0:03ðꢀ0:01Þlog P2
+
5:50ðꢀ0:28Þ
ð10Þ
log 1= C7 = –0:22ðꢀ0:09Þ log P + 8:06ðꢀ0:69Þ
n = 9; r2 = 0:818; s = 0:205; q2 = 0:700; Q = 4:415;
F1;7 = 31:462ð5:591Þ
ð7Þ
n = 12; r2 = 0:866; s = 0:229; q2 = 0:745; Q = 4:061;
F2;9 = 29:082ð4:257Þ
Optimum log P=5.78(5.11–6.88).Outliers: DOX and CDDP
QSAR for the cytotoxicity of compounds 1–9 along with five
anticancer drugs (DOX, CDDP, 5-FU, MTX, and ETO) against IGROV
cancer cell line (Table 4)
Molecular hydrophobicity of organotin(IV) compounds 1–9 is
found to be the single most important parameter for all the QSAR
models 1–7. Based on deviations (obsd–predN2×s), compound 7
behaves as a common outlier for five of the seven QSAR models
(QSARs 1, 2, 4, 5 and 6). While the exact reason for 7 being an outlier is
not clear, comparison with dibutyltin(IV) compounds 8 and 9, which
have the same skew trapezoidal geometry, indicates that steric factors
may play a critical role for the unusual activity of 7. To assess the
effects of excluding outliers, QSAR models were examined before and
after the removal of compound. The linear log P models 1–7 with
negative coefficients suggest that the cytotoxicity of compounds 1–9
decreases with increasing hydrophobicity against all seven cancer cell
lines (A498, EVSA-T, H226, IGROV, M19MEL, MCF-7 and WIDR).
Although there are high statistics associated with QSARs 1–7
(r2 =0.729–0.882), the existence of linear correlations between
activity and hydrophobicity of the compounds is not great enough
to establish the upper limit of the activity. Since log P of compounds
1–9 are very high (5.49–10.46), these compounds may have more
than the optimum log P value. This suggests that QSARs 1–7 may
represent only the second half of the parabolic/bilinear model in
terms of log P and may be the cause of the negative log P term in all
QSARs. Thus, more compounds with lower log P (log Pb5.49) values
will be needed to establish the upper log P limit either by the
development of a parabolic or bilinear QSAR model. To solve this
problem, QSAR models 8–14 were developed using the same
cytotoxicity data of compounds 1–9 along with that of anticancer
drugs DOX, CDDP, 5-FU, MTX, and ETO. These drug molecules have log
P values ranging from −2.53 to 0.28, which may represent the first half
of the parabolic/bilinear model and may then be helpful in establishing
the optimum log P value for the series of compounds (1–9).
log 1 = C4 = 0:30ðꢀ0:13Þlog P – 0:03ðꢀ0:01Þlog P2
+
0:35ðꢀ0:14ÞHBD + 5:46ðꢀ0:54Þ
ð11Þ
n = 12; r2 = 0:854; s = 0:251; q2 = 0:629; Q = 3:681;
F3;8 = 15:598ð4:067Þ
Optimum log P=4.75(3.87–5.73)Outliers: DOX and ETO
QSAR for the cytotoxicity of compounds 1–9 along with five
anticancer drugs (DOX, CDDP, 5-FU, MTX, and ETO) against M19MEL
cancer cell line (Table 4)
log 1 = C5 = 0:36ðꢀ0:09Þlog P – 0:03ðꢀ0:01Þlog P2
+
0:30ðꢀ0:09ÞHBD + 5:34ðꢀ0:38Þ
ð12Þ
n = 14; r2 = 0:898; s = 0:209; q2 = 0:767; Q = 4:536;
F3;10 = 29:346ð3:708Þ
Optimum log P=5.51(4.90–6.34)
QSAR for the cytotoxicity of compounds 1–9 along with five
anticancer drugs (DOX, CDDP, 5-FU, MTX, and ETO) against MCF-7
cancer cell line (Table 4)
log 1 = C6 = 0:39ðꢀ0:10Þlog P – 0:03ðꢀ0:01Þlog P2
+
0:35ðꢀ0:09ÞHBD + 5:20ðꢀ0:40Þ
ð13Þ
n = 13; r2 = 0:921; s = 0:209; q2 = 0:746; Q = 4:589;
QSAR for the cytotoxicity of compounds 1–9 along with five
anticancer drugs (DOX, CDDP, 5-FU, MTX and ETO) against A498
cancer cell line (Table 4)
F3;9 = 34:975ð3:863Þ
Optimum log P=5.98(5.27–7.01)Outlier: ETO
QSAR for the cytotoxicity of compounds 1–9 along with five
anticancer drugs (DOX, CDDP, 5-FU, MTX, and ETO) against WIDR
cancer cell line (Table 4)
log 1= C1 = 0:23ðꢀ0:08Þ log P – 0:02ðꢀ0:01Þ log P2
+
5:95ðꢀ0:22Þ
ð8Þ
n = 12; r2 = 0:844; s = 0:201; q2 = 0:692; Q = 4:572;
log 1 = C7 = 0:38ðꢀ0:07Þlog P– 0:04ðꢀ0:01Þlog P2
F2;9 = 24:346ð4:257Þ
+
0:28ðꢀ0:08ÞHBD + 5:66ðꢀ0:30Þ
ð14Þ
n = 12; r2 = 0:955; s = 0:152; q2 = 0:811; Q = 6:428;
Optimum log P=5.41(4.59–6.87)Outliers: DOX and MTX
QSAR for the cytotoxicity of compounds 1–9 along with five
anticancer drugs (DOX, CDDP, 5-FU, MTX and ETO) against EVSA-T
cancer cell line (Table 4)
F3;8 = 56:593ð4:067Þ
Optimum log P=4.72(4.35–5.12)outlier: Compound 9
All the above QSAR models (8–14) are parabolic correlations in
terms of log P, which suggest that the cytotoxic activity of these
compounds first increases with an increase in their hydrophobicity up
to an optimum log P and then decreases, as defined by the equations
for respective cancer cell lines. The optimum log P for QSAR models 8–
14 range from 4.72 to 5.99. HBD is present in five QSAR models (Eqs.
(9),(11)–(14)) and its positive coefficient suggest that the cytotoxic
activity of these compounds, against five respective cancer cell lines,
increases with increase in the number of hydrogen bond donor. On
the other hand, it has been suggested by Lipinski et al. that a HBD≤5
log 1= C2 = 0:41ðꢀ0:14Þ log P – 0:03ðꢀ0:02Þ log P2
ð9Þ
+
0:37ðꢀ0:13ÞHBD + 5:28ðꢀ0:58Þ
n = 14; r2 = 0:839; s = 0:318; q2 = 0:658; Q = 2:881;
F3;10 = 17:371ð3:708Þ
Optimum log P=5.99(5.06–7.63).

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
959
Table 4
Biological, hydrophobicity (log P), and HBD parameters of compounds 1-9 along with five anticancer drugs (DOX, CDDP, 5-FU, MTX, and ETO) used to derive QSARs 8–14.
No.
Compound
log1/C1 (Eq. (8))
log1/C2 (Eq. (9))
log1/C3 (Eq. (10))
Obsd.
Pred.
Δ
Obsd.
Pred.
Δ
Obsd.
Pred.
Δ
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
6.53
6.52
6.52
6.48
6.78
6.78
6.29
5.85
6.29
6.78
5.30
5.96
7.09
5.65
6.57
6.51
6.50
6.57
6.58
6.58
6.08
6.04
6.47
6.01
5.23
5.79
5.84
6.02
−0.04
0.01
0.02
−0.09
0.20
0.20
6.75
6.77
7.34
6.73
7.17
7.10
6.79
6.35
6.74
7.83
5.78
5.44
7.96
6.27
6.88
6.83
6.82
6.88
6.88
6.87
6.64
6.57
7.16
8.00
5.52
5.75
7.71
6.52
−0.13
−0.06
0.52
−0.15
0.29
6.57
6.55
6.55
6.52
6.76
6.79
6.01
5.86
6.17
6.44
5.67
5.58
5.30
5.17
6.59
6.52
6.51
6.59
6.59
6.59
5.94
5.88
6.47
5.59
4.34
5.24
5.33
5.60
−0.02
0.03
0.04
−0.07
0.17
0.20
0.23
0.15
0.21
0.07
−0.19
−0.18
0.77
0.07
0.17
1.25
−0.37
−0.22
−0.42
−0.17
0.26
−0.31
0.25
−0.02
−0.30
0.85
1.33
0.34
−0.03
−0.43
9
9
10a,b
11b
12
13a
14
DOX
CDDP
5-FU
MTX
ETO
−0.25
No.
Compound
log1/C4 (Eq. (11))
log1/C5 (Eq. (12))
log1/C6 (Eq. (13))
Obsd.
Pred.
Δ
Obsd.
Pred.
Δ
Obsd.
Pred.
Δ
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
6.41
6.37
6.34
6.36
6.74
6.79
6.13
5.74
6.22
6.96
6.12
5.64
7.81
6.01
6.47
6.32
6.30
6.47
6.49
6.50
5.94
5.86
6.61
8.01
5.93
5.97
7.80
6.60
−0.06
0.05
0.04
−0.11
0.25
0.29
6.66
6.67
6.67
6.64
6.77
6.77
6.34
5.96
6.40
7.53
5.63
5.47
7.30
6.07
6.63
6.54
6.52
6.63
6.64
6.64
6.20
6.13
6.78
7.50
5.40
5.69
7.25
6.33
0.03
0.13
0.15
0.01
0.13
0.13
0.14
−0.17
−0.38
0.03
6.68
6.69
6.72
6.66
6.82
6.90
6.54
6.30
6.66
7.74
5.66
5.24
7.40
5.36
6.69
6.64
6.63
6.69
6.69
6.69
6.46
6.39
6.95
7.71
5.40
5.63
7.44
6.34
−0.01
0.05
0.09
−0.03
0.13
0.21
0.19
0.08
−0.12
−0.39
−0.05
0.19
−0.33
0.01
−0.09
−0.29
0.03
0.26
−0.39
0.04
9
9
10c
11
12
13
14c,d
DOX
CDDP
5-FU
MTX
ETO
0.23
−0.22
0.05
−0.59
−0.26
−0.98
No.
Compd.
log1/C7 (Eq.(14))
log P
HBD
Obsd.
Pred.
Δ
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
6.71
6.75
6.75
6.71
6.76
6.81
5.97
5.59
6.13
7.69
5.72
5.76
–
6.78
6.58
6.55
6.77
6.81
6.82
5.90
5.79
6.77
7.72
5.56
5.95
7.45
6.60
−0.07
0.17
0.20
5.94
7.27
7.38
6.05
5.63
5.49
10.23
10.46
7.66
1
1
1
1
1
1
2
2
2
7
4
2
7
3
−0.06
−0.05
−0.01
0.07
−0.20
−0.64
−0.03
0.16
9e
10
11
12
13
14
9
DOX
CDDP
5-FU
MTX
ETO
0.24
−2.53
−0.65
−0.45
0.28
−0.19
–
6.59
−0.01
a
Not included in the derivation of QSAR 8.
Not included in the derivation of QSAR 10.
Not included in the derivation of QSAR 11.
Not included in the derivation of QSAR 13.
Not included in the derivation of QSAR 14.
b
c
d
e
improves oral bioavailability [56]. The organotin(IV) compounds of
the present investigation 1–9 have HBD values of either 1 or 2. This
means that we have room to increase the HBD and improve the
activity of these compounds. Although the HBD term is present in five
QSARs, log P is the most important indicator of compound activity and
explains the major part of the variance in the data. Thus most
attention should be paid to the hydrophobicity of the compounds.
validation was carried out using cross-validation (q2N0.5) and Y-
randomization tests [59,62] (see Table 5). In the Y-randomization tests,
the poor values of r2 and q2 ensure the robustness of QSAR models 1–14
and also the lack of over fitting. Due to the small data sets, the external
validation test was not considered.
3.4.2. Comparative QSAR study on the cytotoxic activities of compounds
1–9 against seven cancer cell lines (A498, EVSA-T, H226, IGROV, M19
MEL, MCF-7 and WIDR)
3.4.1. Validation of QSAR models 1–14
All QSAR models (1–14) were validated in two steps involving
statistical diagnostics and internal validation. In statistical diagnostics,
QSAR models 1–14 were filtered through the following seven necessary
conditions: (i) n/p≥4 (ii) r2N0.6 (iii) q2N0.5 (iv) r2−q2b0.3 (v) FNF(lit)
at 95% level (vi) high Q value, and (vii) low s value [28,57–61]. Internal
On comparison among QSAR models 1–7, one can suggest that
the mechanism for the cytotoxic activity of compounds 1–9 against
A498, EVSA-T, H226, IGROV, M19 MEL, MCF-7 and WIDR cancer cell
lines is almost the same and directly dependent on the hydropho-
bicity of the compound. It must be noted here that QSARs 1 and 3–7

960
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
Table 5
Y-Randomization test data for QSARs 1–14.
QSAR
No.
NOR-1a
NOR-2
NOR-3
NOR-4
NOR-5
r2
q2
r2
q2
r2
q2
r2
q2
r2
q2
1
2
3
4
5
6
7
8
0.699
0.567
0.488
0.676
0.620
0.501
0.419
0.435
0.415
0.236
0.778
0.393
0.435
0.466
0.047
−0.440
0.083
0.003
0.081
0.010
0.033
0.010
0.002
0.068
0.445
0.341
0.557
0.490
0.390
0.358
0.250
−0.679
−0.621
−0.544
−0.635
−0.636
−0.646
−0.417
0.189
−0.106
0.312
−1.268
−0.021
−0.153
−2.121
0.392
0.397
0.486
0.309
0.424
0.324
0.517
0.430
0.466
0.121
0.596
0.404
0.368
0.423
−0.210
−0.193
0.090
−0.306
−0.226
−0.309
0.111
−0.028
−0.139
−0.406
−2.079
−0.329
−0.810
−2.015
0.011
0.048
0.120
0.003
0.095
0.002
0.273
0.240
0.564
0.376
0.538
0.552
0.266
0.525
−0.763
−0.675
−0.393
−0.791
−0.614
−0.335
−0.142
−0.391
0.154
0.037
−1.395
0.137
−0.988
−1.147
0.008
0.021
0.037
0.092
0.321
0.089
0.513
0.031
0.244
0.041
0.334
0.177
0.211
0.714
−0.374
−0.441
−0.293
−0.782
−0.377
−0.850
0.226
−0.781
−0.353
−0.475
−1.960
−0.463
−0.411
−0.179
−0.086
−0.025
−0.597
−0.005
−0.096
−0.150
−0.070
−0.491
−0.213
−0.251
−1.401
9
10
11
12
13
14
a
NOR = number of Y-randomization.
Table 6
Correlations among the cytotoxic activities of compounds 1–9 against seven cancer cell lines e.g. A498, EVSA-T, H226, IGROV, M19 MEL, MCF-7, and WIDR.
QSAR no.
QSAR models (n=9)
r2
q2
s
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
log 1/C1=0.76 log 1/C2+1.23
log 1/C1=0.82 log 1/C3+1.17
log 1/C1=0.89 log 1/C4+0.80
log 1/C1=1.05 log 1/C5−0.39
log 1/C1=1.62 log 1/C6−4.33
log 1/C1=0.58 log 1/C7+2.69
log 1/C2=0.65 log 1/C3+2.70
log 1/C2=0.70 log 1/C4+2.39
log 1/C2=0.85 log 1/C5+1.28
log 1/C2=1.39 log 1/C6−2.42
log 1/C2=0.46 log 1/C7+3.86
log 1/C3=0.98 log 1/C4+0.21
log 1/C3=1.19 log 1/C5−1.37
log 1/C3=1.76 log 1/C6−5.32
log 1/C3=0.71 log 1/C7+1.85
log 1/C4=1.10 log 1/C5−0.86
log 1/C4=1.79 log 1/C6−5.59
log 1/C4=0.60 log 1/C7+2.44
log 1/C5=1.46 log 1/C6−3.19
log 1/C5=0.58 log 1/C7+2.82
log 1/C6=0.33 log 1/C7+4.52
0.628
0.904
0.968
0.947
0.940
0.844
0.518
0.557
0.581
0.642
0.495
0.875
0.920
0.834
0.935
0.862
0.945
0.747
0.884
0.956
0.766
0.085
0.768
0.927
0.921
0.914
0.712
0.245
0.383
0.427
0.517
0.204
0.819
0.655
0.713
0.886
0.779
0.900
0.597
0.784
0.883
0.555
0.185
0.094
0.055
0.070
0.074
0.120
0.220
0.211
0.205
0.189
0.225
0.124
0.099
0.143
0.090
0.124
0.079
0.169
0.096
0.060
0.088
explain 82.3–88.2% of the variance in the data sets either without or
with one outlier. On the other hand, QSAR 2 explains only 72.9% of
the variance in the data set with two outliers. The poor statistics of
QSAR 2 compared to the other QSAR models (Eqs. (1) and (3)–(7))
suggest that it may have an additional mechanism to the common one.
Similar/dissimilar mechanisms involved in the cytotoxic activity of
compounds 1–9 against seven different cancer cell lines (A498, EVSA-
T, H226, IGROV, M19 MEL, MCF-7 and WIDR) can also be understood
by direct comparison among their activities, known as activity–
activity relationships. Correlations among the cytotoxic activities of
compounds 1–9 against seven different cancer cell lines are given in
Table 6 (Eq. (15)–(35)). The high mutual correlations among the
cytotoxicities against six cancer cells in QSARs 16–20 (r2/q2 =0.904/
0.768, 0.968/0.927, 0.947/0.921, 0.940/0.914, and 0.844/0.712) sug-
gest that compounds 1–9 may obey a similar mechanism for their
cytotoxicities against six cancer cells (A498, H226, IGROV, M19MEL,
MCF-7 and WIDR), as shown in Fig. 6. On the other hand, the invalid
QSARs 15, 21, 22, 23, and 25 (r2/q2 =0.628/0.085, 0.518/0.245, 0.557/
0.383, 0.581/0.427 and 0.495/0.204) suggest that the mechanism for
the cytotoxicities of compounds 1–9 against EVSA-T cancer cell line is
different to that of the other five cancer cell lines (A498, H226, IGROV,
M19 MEL and WIDR). The poor statistics, but valid QSAR 24 (r2/
q2 =0.642/0.517), suggest that there is some similarity in the
mechanism for the cytotoxicities of compounds 1–9 against EVSA-T
Fig. 6. Plot of observed log 1/C1 versus predicted log 1/C1 from QSARs 16–20 (Table 6).

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
961
Fig. 9. Plot of observed log 1/C5 versus predicted log 1/C5 from QSARs 33 and 34
(Table 6).
Fig. 7. Plot of observed log 1/C3 versus predicted log 1/C3 from QSARs 26–29 (Table 6).
and MCF-7 cell lines. This is not surprising because both are breast
cancer cell lines. EVSA-T and MCF-7 are the ER−/PR+and ER+/PR+
breast cancer cell lines, respectively. Although the similarity between
these two breast cancer cell lines is due to the progesterone receptor-
positive (PR+), the dissimilarity is due to the estrogen receptor-
negative/positive (ER−/ER+). The similar cytotoxic mechanism for
compounds 1–9 among the five cancer cell lines H226, IGROV, M19
MEL, MCF-7 and WIDR has been further supported by their high mutual
correlations, as shown by QSARs 26–35 (Table 6) and Figs. 7–10. From
the above discussions, it can be suggested that the cytotoxicities of
compounds 1–9 against six cancer cell lines (A498, H226, IGROV,
M19MEL, MCF-7 and WIDR) are due to a similar mechanism, but the
same is not true against EVSA-T cancer cell.
Studies on the effect of iron, copper, and zinc ions on RNR in freshly
isolated normal and leukemic human lymphocytes found that zinc
acts to inhibit the enzyme in both cell types, while iron and copper
had stimulatory effects [63,64]. Metal complexes of carbothioamides
[65–67] and thiosemicarbazones [68] have been found to inhibit RNR
and possess anticancer activity. On the other hand, anthracyclines and
some V(IV) and Mo(IV) compounds have been shown to exert various
effects on DNA or inhibit topoII [69,70]. Recently, information has
been acquired from the enzymes docking studies of compounds 5 and
6 [17] and 7–9 [18]. Combined with the encouraging cytotoxic activity
obtained for test compounds 1–4, this prompted us to carry out
molecular docking studies of 1–4 to understand and hopefully
overcome some of the challenges occurring by formation of strong
covalent attachments to target enzymes.
3.5. Docking study
The results of docking studies of 1 and 4 with enzymes are shown
in Figs. 11–18 while results for 2 and 3 can be seen in S3–S10. The
docking program is validated by docking ADP in the active site of
enzyme RNR, with a close overlap between the docked ligand and the
native ligand being observed [17]. The docking studies revealed that
tributyltin(IV) compounds 1–4 interact with enzymes RNR (4R1R), TS
(2G8D), TP (1BRW) and topoII (1QZR) at various sites and their
hydrogen bonding interactions are given in Table 7. All compounds
exhibited hydrogen bonding interactions through azo nitrogen atom
(s) with amino acid residues of the enzymes RNR, TS and TP. Both azo
group N(1) and N(2) atoms of compounds 1 and 2 interacted with
RNR, while there was interaction with either N(1) or N(2) for the
Dockings of compounds 1–4 in the active sites of the key enzymes
RNR, TS, TP and topoII have been determined since these enzymes are
promising targets for cancer therapy. For example, RNR performs a
key, rate limiting step in DNA synthesis, controls the balance of the
deoxyribonucleotide pools, and changes in its activity can alter the
spontaneous mutation rate of cells [63]. Increased RNR activity has
been associated with disease states including cancer [63,64] and
inhibition of this enzyme is an attractive target for cancer therapy.
Fig. 8. Plot of observed log 1/C4 versus predicted log 1/C4 from QSARs 30–32 (Table 6).
Fig. 10. Plot of observed log 1/C6 versus predicted log 1/C6 from QSAR 35 (Table 6).

962
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
Fig. 11. Bu3SnL1H (1) docked into the binding site of the enzyme ribonucleotide reductase (4R1R). Hydrogen bonding interactions between the various groups of 1 and amino acid
residues are shown. Hydrogen atoms are omitted for clarity.
other enzymes. Hydroxyl group oxygen atom(s) of the compounds
show more hydrogen bonding interactions with active sites of the
enzymes than other functional groups of the compounds (Table 7, see
Figs. 11–18 and S3–S10 for specific amino acid residues). It should be
mentioned that tributyltin(IV) compounds 1, 2 and 4 exhibited
hydrogen bonding interactions through Sn–oxygen atom(s) with
Fig. 12. Bu3SnL1H (1) docked into the binding site of the enzyme thymidylate synthase (2G8D). Hydrogen bonding interactions between the various groups of 1 and amino acid
residues are shown. Hydrogen atoms are omitted for clarity.

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
963
Fig. 15. Bu3SnL4H (4) docked into the binding site of the enzyme ribonucleotide
reductase (4R1R). Hydrogen bonding interactions between the various groups of 4 and
amino acid residues are shown. Hydrogen atoms are omitted for clarity.
Fig. 13. Bu3SnL1H (1) docked into the binding site of the enzyme thymidylate
phosphorylase (1BRW). Hydrogen bonding interactions between the various groups of
1 and amino acid residues are shown. Hydrogen atoms are omitted for clarity.
enzymes TS, TP and topoII and that such interactions were not
observed earlier for triphenyltin(IV) compounds (5 and 6) [17] and
dibutyltin(IV) compounds (7–9) [18].
Docking studies also revealed that compounds 1–4 are masked
inside the active site of topoII and exhibited hydrogen bonding
interactions through various atoms of the molecule with amino acid
residues of the enzyme, for instance R77 (in compound 2), K102 (in
compound 3), S127, G143, T195 and S127 (in compound 4). Such
interactions were not observed at all earlier for compounds 5–9 [17,18].
Therefore, on the basis of docking studies, it is inferred that the
anticancer activities of compounds 1–4 might be emanating from
their interactions with enzymes RNR, TS, TP and topoII. The docking
studies also indicate that the azo group nitrogen atoms, hydroxyl,
carbonyl oxygen atoms and Sn–oxygen atom in the ligand moiety play
an important role during the dockings of the tributyltin(IV)
compounds into the active sites of various enzymes. Nevertheless,
the possibility of coordination through tin beyond the active site of
the enzymes cannot be ruled out completely. It is very difficult to
envisage the role of such atoms in binding proteins in relation to
improved cytotoxic activity.
4. Conclusions and outlook
The present series of tributyltin(IV) 2/4-[(E)-2-(aryl)-1-diazenyl]
benzoates (1–4) was designed to provide a new skeletal framework
that incorporates a tributyltin(IV) ester and a ligand that varies in
both the position of the carboxylate and OH, Me and tert-Bu
substituents. It was anticipated that this variation would determine
the number of possible hydrogen bonds between a complex and
enzyme amino acid residues and hence influence the activity of the
compounds. In solution, the organotin(IV) benzoates exist as four-
coordinate tetrahedral tin species, while in the solid-state they have
either a distorted tetrahedral (4) or distorted trigonal bipyramidal
geometry (1–3). Promising cytotoxic effects were noted for the
complexes, with Bu3SnL1H (1) being more cytotoxic than the other
test compounds (2–4). In general, the cytotoxic results of compounds
Fig. 14. Bu3SnL1H (1) docked into the binding site of the enzyme topoisomerase II
(1QZR). Hydrogen bonding interactions between the various groups of 1 and amino
acid residues are shown. Hydrogen atoms are omitted for clarity.

964
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
Fig. 16. Bu3SnL4H (4) docked into the binding site of the enzyme thymidylate synthase (2G8D). Hydrogen bonding interactions between the various groups of 4 and amino acid
residues are shown. Hydrogen atoms are omitted for clarity.
1–4 are far superior to CDDP, 5-FU, ETO and related dibutyltin(IV)
compounds 7–9, and comparable to their triphenyltin(IV) analogues 5
and 6 across a panel of seven human tumor cell lines. Docking studies
of compounds 1–4 indicated that the cytotoxic activity may be
associated with the tetrahedral geometry of the complexes in solution
and the presence of an azo functionality in the ligand framework. The
Fig. 17. Bu3SnL4H (4) docked into the binding site of the enzyme thymidylate phosphorylase (1BRW). Hydrogen bonding interactions between the various groups of 4 and amino
acid residues are shown. Hydrogen atoms are omitted for clarity.

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
965
Fig. 18. Bu3SnL4H (4) docked into the binding site of the enzyme topoisomerase II (1QZR). Hydrogen bonding interactions between the various groups of 4 and amino acid residues
are shown. Hydrogen atoms are omitted for clarity.
potential of these tributyltin(IV) complexes as enzyme inhibitors
was shown by the exquisite binding and highly specific target
enzymes recognition. Careful characterization of kinetics and
enzyme/inhibitor complex structures is necessary to truly demon-
strate specificity. Thus, the docking studies are very important since
disease pathology often correlates with abnormal enzyme activity
and therefore enzyme inhibition can be a powerful and versatile tool
in the treatment of disease. Data from the present study suggests that
the tributyltin(IV) complex 1 merits further investigation as a new
drug and may be a suitable candidate for modification in order to
improve cytotoxic and dissolution properties. Further work in this
area is underway.
Acknowledgements
The financial support of the Department of Science and Technol-
ogy, New Delhi, India (grant nos. SR/S1/IC-03/2005,TSBB and SR/S1/
OC-11A/2006, PS), the Università degli Studi di Palermo, Italy (grants
ORPA079E5M and ORPA0737W2, LP) and the University Grants
Commission, New Delhi, India through SAP-DSA, Phase-III, are
gratefully acknowledged. The in vitro cytotoxicity experiments were
carried out by Ms. P. F. van Cuijk in the Laboratory of Translational
Pharmacology, Department of Medical Oncology, Erasmus Medical
Centre, Rotterdam, The Netherlands, under the supervision of Dr. E. A.
C. Wiemer and Prof. Dr. G. Stoter.
Table 7
Hydrogen bonding interactions of tributyltin(IV) compounds 1–4 with ribonucleotide reductase (4R1R), thymidylate synthase (2G8D), thymidylate phosphorylase (1BRW) and
topoisomerase II (1QZR).
Compound
Amino acids involved in hydrogen bonding
Compd
no.
Groups
4R1R
2G8D
1BRW
1QZR
Amino acid residue
Amino acid residue
R23–N1 (2.94)
Amino acid residue
Amino acid residue
1
2
Azo group nitrogen atom(s)
T624–N1 (2.62)
T624–N2 (2.92)
E623 (2.31)
Carbonyl oxygen atom(s)
Hydroxyl group oxygen atom(s)
Sn–oxygen atom(s)
R23 (2.75 and 2.82)
R23 (2.71 and 2.99)
S219 (2.86), H259 (2.97)
R23–N1 (2.96)
T209 (2.90), T624 (2.44)
E1022 (2.90)
Q1064 (2.23)
R1026–N2 (2.99)
Azo group nitrogen atom(s)
C439–N1 (2.64)
C439–N2 (2.07)
Hydroxyl group oxygen atom(s)
Sn–oxygen atom(s)
Q1064 (2.39)
R23 (2.98 and 2.64)
R77 (2.29)
3
4
Azo group nitrogen atom(s)
Carbonyl oxygen atom(s)
Hydroxyl group oxygen atom(s)
Azo group nitrogen atom(s)
Carbonyl oxygen atom(s)
Hydroxyl group oxygen atom(s)
Sn–oxygen atom(s)
R1026–N2 (2.97)
G1066 (2.75), S1065 (2.81)
C225 (2.14)
K102 (2.30)
Y730 (2.90), C439 (2.47)
T209–N1 (2.40)
S622 (2.37)
H259 (2.85)
R23–N2 (2.31)
S127 (2.96), G143 (2.23)
T195 (2.33)
S127 (2.88)
R251 (2.56)
E1019 (2.75)
H259 (2.08)

966
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 104 (2010) 950–966
[25] R.P. Verma, C. Hansch, Development of QSAR models using C-QSAR program: a
Appendix A. Supplementary data
regression program that has dual databases of over 21,000 QSAR models, Nat.
Protoc. (2007), doi:10.1038/nprot.2007.125 Freely available at: http://www.
natureprotocols.com/2007/03/05/development_of_qsar_models_usi_1.php.
[26] ACD PhysChem Suite, 110 Yonge Street, 14th floor, Toronto, Ontario, M5C 1 T4,
Canada, www.acdlabs.com.
[27] R.D. Cramer III, J.D. Bunce, D.E. Patterson, I.E. Frank, Quant. Struct.-Act. Relat. 7
(1988) 18–25.
[28] C.A. Bennett, N.L. Franklin, Statistical Analysis in Chemistry and the Chemical
Industry, John Wiley, New York, 1967, pp. 708–709.
Crystallographic data (without structure factors) for the structure
of 4 have been deposited with the Cambridge Crystallographic Data
Centre (CCDC) as supplementary publication no. CCDC 757617. Copies
of the data can be obtained free of charge from the CCDC (12 Union
Road, Cambridge CB2 1EZ, UK; Tel.: +44-1223-336408; Fax: +44-
1223-336003; e-mail: deposit@ccdc.cam.ac.uk; Web site: http://
www.ccdc.cam.ac.uk).
[29] C.D. Selassie, S. Kapur, R.P. Verma, M. Rosario, J. Med. Chem. 48 (2005) 7234–7242.
[30] A.L. Spek, Acta Crystallogr. D 65 (2009) 148–155.
[31] G.M. Sheldrick, Acta Crystallogr. A 64 (2008) 112–122.
[32] J.J.P. Stewart, J. Comput. Chem. 10 (1989) 209–220.
[33] J.J.P. Stewart, J. Comput. Chem. 10 (1989) 221–264.
[34] J.J.P. Stewart, J. Comput. Chem. 12 (1991) 320–341.
[35] J.J.P. Stewart, J. Mol. Model. 10 (2004) 155–164.
[36] J.J.P. Stewart, J. Mol. Model. 13 (2007) 1173–1213.
[37] J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing, J. Hutter,
Comp. Phys. Comm. 167 (2005) 103–128.
[38] ArgusLab 4.0.1: Thompson MA, Planaria Software LLC, Seattle, WA, http://www.
argusLab.com
[39] Protein Data Bank, bhttp://www.rcsb.org/pdb/&gt.
[40] R. Wang, L. Lai, S. Wang, J. Comput. Aided Mol. Des. 16 (2002) 11–26.
[41] T.S. Basu Baul, S. Dhar, N. Kharbani, S.M. Pyke, R. Butcher, F.E. Smith, Main Group
Met. Chem. 22 (1999) 413–421.
[42] T.S. Basu Baul, K.S. Singh, A. Lycka, A. linden, X. Song, A. Zapata, G. Eng, Appl.
Organometal. Chem. 20 (2006) 788–797.
[43] T.S. Basu Baul, K.S. Singh, X. Song, A. Zapata, G. Eng, A. Lyčka, A. Linden, J.
Organomet. Chem. 689 (2004) 4702–4711.
[44] M. Nádvorník, J. Holeček, K. Handlir, A. Lyčka, J. Organomet. Chem. 275 (1984)
43–51.
[45] A. Lyčka, J. Holeček, M. Nádvorník, K. Handlir, J. Organomet. Chem. 280 (1985)
323–329.
[46] P.J. Harrison, K. Lambert, T.J. King, B. Majee, J. Chem. Soc., Dalton Trans. (1983)
363–369.
The following information (Figs. S1–S10) are available as Supple-
mentary Materials. Fig. S1: The structure of Bu3SnL2H (2) obtained
after full geometry optimization; Fig. S2: The structure of Bu3SnL3H
(3) obtained after full geometry optimization; Fig. S3: Bu3SnL2H (2)
docked into the binding site of the enzyme ribonucleotide reductase
(4R1R); Fig. S4: Bu3SnL2H (2) docked into the binding site of the
enzyme thymidylate synthase (2G8D); Fig. S5: Bu3SnL2H (2) docked
into the binding site of the enzyme thymidylate phosphorylase
(1BRW); Fig. S6: Bu3SnL2H (2) docked into the binding site of the
enzyme topoisomerase II (1QZR); Fig. S7: Bu3SnL3H (3) docked into
the binding site of the enzyme ribonucleotide reductase (4R1R); Fig.
S8: Bu3SnL3H (3) docked into the binding site of the enzyme
thymidylate synthase (2G8D); Fig. S9: Bu3SnL3H (3) docked into the
binding site of the enzyme thymidylate phosphorylase (1BRW); Fig.
S10: Bu3SnL3H (3) docked into the binding site of the enzyme
topoisomerase II (1QZR). Supplementary data associated with this
article can be found, in the online version, at doi: 10.1016/j.
jinorgbio.2010.05.001.
[47] T.S. Basu Baul, E.R.T. Tiekink, Z. Kristallogr. (NCS) 211 (1996) 489–490.
[48] T.S. Basu Baul, S. Dhar, S.M. Pyke, E.R.T. Tiekink, E. Rivarola, R. Butcher, F.E. Smith, J.
Organomet. Chem. 633 (2001) 7–17.
References
[49] T.S. Basu Baul, W. Rynjah, E. Rivarola, A. Linden, J. Organomet. Chem. 690 (2005)
613–621.
[1] A.S. Abu-Surrah, M. Kettunen, Curr. Med. Chem. 13 (2006) 1337–1357.
[2] K.R. Barnes, S.J. Lippard, Met. Ions Biol. Syst. 42 (2004) 143–177.
[3] V. Cepeda, M.A. Fuertes, J. Castilla, C. Alonso, C. Quevedo, J.M. Perez, Med. Chem. 7
(2007) 3–18.
[50] R. Barbieri, F. Huber, L. Pellerito, G. Ruisi, A. Silvestri, Chemistry of Tin: 119Sn
Mössbauer Studies on Tin Compounds, in: P.J. Smith (Ed.), Blackie, London, 1998,
pp. 496–540.
[4] P. Yang, M. Guo, Coord. Chem. Rev. 185–186 (1999) 189–211.
[5] B. Kelland, Nat. Rev. 7 (2007) 573–584.
[51] R.V. Parish, in: G.J. Long (Ed.), Mössbauer Spectroscopy Applied to Inorganic
Chemistry, vol. 1, Plenum Press, New York, 1984, pp. 528–544.
[52] G.M. Bancroft, R.H. Platt, Adv. Inorg. Chem. Rad. 15 (1972) 59–258.
[53] G.M. Bancroft, V.G. Kumar, T.K. Sham, M.G. Clark, J. Chem. Soc., Dalton Trans.
(1976) 643–654.
[6] P. Kopf-Maier, Eur. J. Clin. Pharmacol. 47 (1994) 1–16 and references therein.
[7] A.Y. Louie, T.J. Meade, Chem. Rev. 99 (1999) 2711–2734.
[8] P.J. Barnard, S.J. Berners-Price, Coord. Chem. Rev. 251 (2007) 1889–1902.
[9] J.D. Robertson, S. Orrenius, Toxicology 181–182 (2002) 491–496.
[10] A.J. Crowe, Drugs. Fut. 12 (1987) 255–275.
[54] E. Corral, A.C.G. Hotze, H. den Dulk, A. Leczkowska, A. Rodger, M.J. Hannon, J.
Reedijk, J. Biol. Inorg. Chem. 14 (2009) 439–448.
[11] M. Gielen (Ed.), Tin-based Antitumor Drugs, Springer Verlag, Berlin, 1990.
[12] M. Gielen, E.R.T. Tiekink (Eds.), Metallotherapeutic Drug and Metal-based
Diagnostic Agents, Wiley, Chichester, England, 2005.
[13] S.K. Hadjikakou, N. Hadjiliadis, Coord. Chem. Rev. 253 (2008) 235–249.
[14] M. Gielen, M. Biesemans, D. de Vos, R. Willem, J. Inorg. Biochem. 79 (2000)
139–145.
[15] F.P. Pruchnik, M. Banbula, Z. Ciunik, M. Latocha, B. Skop, T. Wilczok, Inorg. Chim.
Acta 356 (2003) 62–68.
[16] A. Alama, M. Viale, M. Cilli, C. Bruzzo, F. Novelli, B. Tasso, F. Sparatore, Invest. New
Drugs 27 (2009) 124–130.
[55] A.C.G. Hotze, M. Bacac, A.H. Velders, B.A.J. Jansen, H. Kooijman, A.L. Spek, J.G.
Haasnoot, J. Reedijk, J. Med. Chem. 46 (2003) 1743–1750.
[56] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Adv. Drug Delivery Rev. 23
(1997) 3–25.
[57] C. Hansch, R.P. Verma, ChemMedChem 2 (2007) 1807–1813.
[58] A. Golbraikh, A. Tropsha, J. Mol. Graph. Model. 20 (2002) 269–276.
[59] A. Tropsha, P. Gramatica, V.K. Gombar, QSAR Comb. Sci. 22 (2003) 69–77.
[60] L. Eriksson, J. Jaworska, A.P. Worth, M.T.D. Cronin, R.M. McDowell, P. Gramatica,
Environ. Health Prospect. 111 (2003) 1361–1375.
[61] L. Pogliani, Chem. Rev. 100 (2000) 3827–3858.
[17] T.S. Basu Baul, A. Paul, L. Pellerito, M. Scopelliti, P. Singh, P. Verma, D. de Vos,
Investig. New Drugs (2009) 10.1007-s10637-009-9293-x.
[18] T.S. Basu Baul, A. Paul, L. Pellerito, M. Scopelliti, P. Singh, P. Verma, A. Duthie, D. de
Vos, E.R.T. Tiekink, Invest. New Drugs (2009) 10.1007/s10637-009-9360-3.
[19] R. Willem, I. Verbruggen, M. Gielen, M. Bieseman, B. Mahieu, T.S. Basu Baul, E.R.T.
Tiekink, Organometallics 17 (1998) 5758–5766.
[62] S. Wold, L. Eriksson, Statistical validation of QSAR results. Validation tools,
chemometrics methods in molecular design, in: H. van de Waterbeemd (Ed.),
Methods and Principles in Medicinal Chemistry, vol 2, VCH, Weinheim, Germany,
1995, pp. 309–318.
[63] J.A. Wright, A.K. Chan, B.K. Choy, R.A.R. Hurta, G.A. McClarty, A.Y. Tagger, Biochem.
Cell Biol. 68 (1990) 1364–1371.
[20] T.S. Basu Baul, A. Paul, H.D. Arman, E.R.T. Tiekink, Acta Crystallogr. E 64 (2008)
o2125.
[21] M.R. Boyd, Status of the NCI preclinical antitumor drug discovery screen, Principles
and Practice of Oncology, Lippincott, Philadelphia, vol. 3, 1989, pp. 1–12.
[22] Y.P. Keepers, P.R. Pizao, G.J. Peters, J.V. Ark-Otte, B. Winograd, H.M. Pinedo, Eur. J.
Cancer 27 (1991) 897–900.
[64] M. Oblender, U. Carpentieri, Anticancer Res. 10 (1990) 123–128.
[65] G. Nocentini, F. Federici, M. Grifantini, A. Barzi, Pharmacol. Res. 25 (1992) 312–313.
[66] G. Nocentini, A. Barzi, Biochem. Pharm. 52 (1996) 65–71.
[67] G. Nocentini, A. Barzi, Gen. Pharmacol. 29 (1997) 701–706.
[68] D.X. West, A.E. Liberta, S.B. Padhye, R.C. Chikate, P.B. Sonawane, A.S. Kumbhar, R.G.
Yerande, Coord. Chem. Rev. 123 (1993) 49–71.
[23] C-QSAR Program, BioByte Corp., 201 W. 4th st., Suit 204, Claremont, CA 91711,
USA, www.biobyte.com.
[69] J. Alsner, B.S. Sorensen, V.K. Schmidt, O. Westergard, Bone Marrow Transplant. 12
(1993) S154–S155.
[24] C. Hansch, D. Hoekman, A. Leo, D. Weininger, C.D. Selassie, Chem. Rev. 102 (2002)
783–812.
[70] L.Y. Kuo, A.H. Liu, T.J. Marks, in: A. Sigel, H. Sigel (Eds.), Metal Ions in Biological
Systems, vol. 33, Marcel Dekker, New York, USA, 1996.