Synthesis, characterization and cell viability test of six vanadyl complexes with acetylacetonate derivatives

Article abstract of DOI:10.1016/j.jinorgbio.2013.07.015

Vanadium compounds are known to display a number of therapeutic effects, namely insulin-mimetic and cardiovascular effects. Evidence of the antiproliferative and proapoptotic activity of a number of vanadyl complexes, together with their low toxicity, establishes these metal compounds as promising antitumoral therapeutic agents. In the present work, we describe the synthesis and full characterization of six new vanadyl complexes with acetylacetonate derivatives bearing asymmetric substitutions on the β-dicarbonyl moiety: the complexes were characterized in the solid state as well as in solution. Our results show that all complexes are in square pyramidal geometry; cis isomers in the equatorial plane are favored in the presence of strongly coordinating solvents. EPR evidence suggests that all complexes are in the bis-chelate form, although in two cases the mono-chelated complex seems to be present as well. Preliminary tests carried out on non-tumor and tumor cell lines show that these complexes are effective in suppressing cell viability and elicit a distinct response of tumor and non-tumor cells.

Full text of DOI:10.1016/j.jinorgbio.2013.07.015

Journal of Inorganic Biochemistry 128 (2013) 26–37
Contents lists available at SciVerse ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Synthesis, characterization and cell viability test of six vanadyl complexes
with acetylacetonate derivatives
a
a
a
a
a
Speranza Sgarbossa , Eliano Diana , Domenica Marabello , Annamaria Deagostino , Silvano Cadamuro ,
b
a
b
b
a,
Alessandro Barge , Enzo Laurenti , Margherita Gallicchio , Valentina Boscaro , Elena Ghibaudi ⁎
Dip.to di Chimica, University of Torino, Via Giuria 7, I-10125 Torino, Italy
a
b
Dip.to di Scienza e Tecnologia del Farmaco, University of Torino, Via Giuria 9, I-10125 Torino, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Vanadium compounds are known to display a number of therapeutic effects, namely insulin-mimetic and cardio-
vascular effects. Evidence of the antiproliferative and proapoptotic activity of a number of vanadyl complexes, to-
gether with their low toxicity, establishes these metal compounds as promising antitumoral therapeutic agents.
In the present work, we describe the synthesis and full characterization of six new vanadyl complexes with
acetylacetonate derivatives bearing asymmetric substitutions on the β-dicarbonyl moiety: the complexes were
characterized in the solid state as well as in solution. Our results show that all complexes are in square pyramidal
geometry; cis isomers in the equatorial plane are favored in the presence of strongly coordinating solvents. EPR
evidence suggests that all complexes are in the bis-chelate form, although in two cases the mono-chelated com-
plex seems to be present as well. Preliminary tests carried out on non-tumor and tumor cell lines show that these
complexes are effective in suppressing cell viability and elicit a distinct response of tumor and non-tumor cells.
Received 11 April 2013
Received in revised form 3 July 2013
Accepted 8 July 2013
Available online 18 July 2013
Keywords:
Vanadyl compounds
Antitumoral metal complexes
Acetylacetonate derivatives
X-ray crystallography
EPR spectroscopy
©
2013 Elsevier Inc. All rights reserved.
Vibrational spectroscopy
1
. Introduction
toxicity through oxidative stress mechanisms [13]. In addition,
VO(acac) ] was proposed as the structural basis for a new class of
[
2
Vanadium compounds have been shown to display several therapeu-
cancer-specific MRI contrast agents that are non-toxic and highly
sensitive to cancer metabolism [1,14].
A variety of vanadyl complexes with acac-derivatives (β-dicarbonyl
ligands) have been reported [15] and shown to display anticancer activ-
ity [16]: e.g. the [VO(curcumin)] complex was found able to inhibit
mouse lymphoma cell growth [17].
Due to these reasons, vanadyl complexes with ligands bearing sym-
metric and asymmetric substitutions on the β-dicarbonyl moiety have
raised a strong interest [15,18] and their behavior in terms of pharmaco-
logical and structural properties deserves to be investigated. In fact,
highlighting the structure–function relationships of this class of com-
pounds is crucial for designing new pharmacologically-active molecules
and unraveling the molecular targets of these species.
A crucial issue related with vanadyl complexes with acac-derivatives
concerns their behavior in solution, as these compounds may interact
with solvents, giving rise to a wide range of chemical species with dis-
tinct pharmacological effects [19].
tic effects, including insulin-mimetic, cardiovascular, anticarcinogenic,
antiproliferative and proapoptotic effects [1,2]. In particular, the anti-
tumoral properties of vanadium complexes raise attention due to the rel-
atively low toxicity of vanadium compounds [2].
A number of vanadyl complexes have been synthesized and investi-
gated relatively to their application in anticancer therapies [3,4]: they
include vanadocene derivatives [5] and many complexes of the VO2
moiety with organic ligands (Schiff bases, phenanthroline and quinox-
aline derivatives, etc.) that may interact with DNA and other cellular
targets [6].
+
The vanadyl complex with pentane-2,4-dionato (acetylacetonate,
acac) is one of the oldest vanadium therapeutic agent ever synthe-
sized [7]. Apart from its insulin-mimetic activity [8,9], more recent
studies [9,10] have highlighted its anticancer potential on human
hepatoma cell line. Although the molecular details of such effect
are not clear, [VO(acac)
it is an efficient DNA cleaving agent at submicromolar concentration
8]; it has been found more effective than vanadyl sulphate in stim-
2
] has been shown to display several actions:
Whereas these complexes in the solid state display a square pyra-
midal geometry with the oxo-ligand in the axial position and the two
β-dicarbonyl ligands in the equatorial plane [20,21], upon dissolu-
tion in organic solvents they may interact with solvent molecules
via hydrogen bonding or by direct coordination to the sixth vacant
[
ulating the activity of a cytosolic protein kinase [11,12], blocking
cell cycle progression at G1 phase [10] and inducing mitochondrial
site of the transition metal ion, leading to [VO(acac)
L = coordinating solvent molecule) [15,19]. Cis/trans isomers in
the equatorial plane may form in the presence of ligands bearing
2
L] adducts
(
⁎
Corresponding author. Tel.: +39 11 6707951; fax: +39 11 6707855.
E-mail address: elena.ghibaudi@unito.it (E. Ghibaudi).
0162-0134/$ – see front matter © 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jinorgbio.2013.07.015

S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
27
asymmetric substitutions on the β-dicarbonyl moiety (we will refer
2.2. Ligand b) 1-(4-methoxyphenyl)-4,4,4-trifluorobutane-1,3-dione
to as cis-planar and trans-planar isomers); in addition, depending
on the nature of the solvent, the L ligand may lie either in the equa-
torial plane (cis to the VO moiety) or occupy the 6th coordination po-
sition (trans to the VO moiety) [22]. The concerted use of different
spectroscopical techniques may help to discriminate between the
species found in solution.
In the present work, we present the synthesis and characterization
of six new vanadyl complexes with acac-ligand derivatives. The species
in the solid state were characterized by X-ray crystallography and
Raman spectroscopy; a combination of EPR and optical spectroscopy
was employed to characterize the species in solution. Preliminary
in vitro tests performed on different cell lines (HCT 116, HT-29,
hTERT-HME1 and human immortalized podocytes) have shown specif-
ic cell toxicity of some complexes and establish these compounds as po-
tential antitumoral agents.
Following the above-described procedure, p-methoxyacetophenone
(4.95 g, 33.0 mmol) was treated with ethyl trifluoroacetate and
EtONa in EtOH affording a pale yellow solid (4.79 g, 59% — petroleum
ether/ethyl ether 90/10 v/v, 10% formic acid). Found C, 53.69;
H, 3.70, F, 22.99%. Calc. for C11
H
9
F
3
O
3
: C, 53.67; H, 3.68, F, 23.15%.
(200 MHz; CDCl , Me Si) (2H, d, C_CH),
), 6.52 (1H, s, C_CH), 7.00 (2H, d, J = 8.0 Hz, Ar),
7.94 (2H, d, J = 8.0 Hz, Ar); δ (50.2 MHz; CDCl , Me Si) 55.1 (1 × q),
−
1
νmax(neat)/cm
1600. δ
H
3
4
3.91 (3H, s, OCH
3
C
3
4
91.1 (1 × d), 114.1 (2 × d), 117.1 (q, J (CF) = 300.5 Hz), 124.9
(1 × s), 129.7 (2 × d), 164.5 (1 × d), 175.3 (q, J (CF) = 35.5 Hz),
+
186.0 (1 × s). MS (EI, 70 eV): m/z (%) = 246 (97, M ), 177 (100), 69
(54); m.p. 60.0 °C–60.1 °C.
2.3. Ligand d) 1-(6-methoxy-2-naphtyl)-4,4,4-trifluorobutane-1,3-dione
2
. Experimental
Following the above-described procedure, 6-methoxy-2-
acetonaphtone (6.60 g, 33.0 mmol) was treated with ethyl
All commercially obtained reagents and solvents were used as
trifluoroacetate and EtONa in EtOH affording a brilliant yellow solid
received. Products were purified by a preparative column chroma-
tography on Macherey Nagel silica-gel for flash chromatography,
(8.50 g, 87% — CH
2
Cl
2
/MeOH 95/5 v/v). Found C, 61.00; H, 3.70, F,
: C, 60.82; H, 3.74, F, 19.24%. νmax(neat)/
(200 MHz; CDCl , Me Si) (2H, d, C_CH), 3.97 (3H, s,
), 6.70 (1H, s, C_CH), 7.32 (2H, m, Ar), 7.89 (3H, m, Ar) 8.44
(1H, bs, Ar); δ (50.2 MHz; CD COCD , Me Si) 54.8 (1 × q), 92.3
19.20%. Calc. for C15
11 3 3
H F O
−
1
0
.04–0.063 mm/230–400 mesh. Reactions were monitored by TLC
cm
OCH
1628. δ
H
3
4
using silica-gel on TLC-PET foils (Fluka), 2–25 μm, layer thickness
3
0
.2 mm, medium pore diameter 60 Å.
C
3
3
4
The synthesis of the ligands was performed according to the protocol
described by Li and coworkers [23] and modified by the authors.
In a 250 mL round bottom flask the ethyl trifluoroacetate (10.37 g,
3.0 mmol) and the suitable ketone (33.0 mmol) were dissolved in a
1% solution of EtONa (73.0 mmol) in EtOH. The reaction mixture was
(1 × d), 105.8 (1 × d), 117.4 (q, J (CF) = 280.0 Hz), 119.8 (1 × d),
123.3 (1 × d), 127.4 (1 × d), 127.8 (2 × s), 130.6 (1 × d), 131.2
(1 × d), 138.0 (1 × s), 160.4 (1 × s), 174.8 (q, J (CF) = 35.0 Hz),
187.2 (1 × s). MS (EI, 70 eV): m/z (%) = 296 (100, M ), 227 (45), 69
(38); m.p. 92.5 °C–920.7 °C.
+
7
2
stirred at 70 °C until disappearance (3–4 h) of the ketone spot at TLC
CH Cl /MeOH 95/5 v/v). Then the solvent was evaporated, and the
solid residue was dissolved with H O, acidified with HCl 6 N until
pH 3–4. Then it was extracted with CH Cl (2 × 20 mL), dried over
Na SO , filtered and evaporated under reduced pressure, to give the
(
2
2
2
2.4. Ligand e) 1-(N-methyl-3-indolyl)-4,4,4-trifluorobutane-1,3-dione
2
2
2
4
Following the above-described procedure, N-methyl-3-acetylindole
(5.71 g, 33.0 mmol) was treated with ethyl trifluoroacetate and EtONa
2 2
crude reaction product that was purified by chromatography (CH Cl /
MeOH 95/5 v/v).
in EtOH affording a brilliant yellow solid (6.30 g, 71% — CH
2 2
Cl /MeOH
N-methyl-3-acetylindole was synthesized following the procedure
reported in literature and the spectral data corresponded to those report-
ed in the literature [24]. Ligands 1-(2-naphtyl)-4,4,4-trifluorobutane-1,3-
dione (ligand c) and 1-(3-thienyl)-4,4,4-trifluorobutane-1,3-dione (li-
gand f) were purchased from Sigma Aldrich.
95/5 v/v). Found C, 58.09; H, 3.81, F, 21.24, N 5.28 %. Calc. for
−
1
C
13
H
10
F
3
NO
(200 MHz; CDCl
6.38 (1H, s, C_CH), 7.38 (3H, m, Ar), 7.89 (1H, s, N\CH) 8.25 (1H, m,
Ar); δ (50.2 MHz; CD COCD , Me Si) 33.1 (1 × q), 93.2 (1 × d), 110.7
2
: C, 58.00; H, 3.74, F, 21.17, N 5.20 %. νmax(neat)/cm
1628. δ
H
3
, Me Si) (2H, d, C_CH), 3.92 (3H, s, NCH
4
3
),
C
3
3
4
The synthesis of the vanadyl complexes was performed according to
the protocol described by Li and coworkers [23] and modified by the
authors.
(1 × d), 111.8 (1 × s), 118.0 (q, J (CF) = 300.5 Hz), 121.8 (1 × d),
122.8 (1 × d), 123.7 (1 × d), 125.6 (1 × s), 138.1 (1 × s), 138.4
(1 × s), 168.0 (q, J (CF) = 35.0 Hz), 186.1 (1 × s). MS (EI, 70 eV): m/z
+
A 2-fold molar excess of the ligand dissolved in 2 mL EtOH was
(%) = 269 (100, M ), 200 (61), 132 (53); m.p. 125.9 °C–126.2 °C.
1
mixed with vanadyl sulphate (60 mg) dissolved in 1.5 mL H
lution was kept under stirring and basified with diluted ammonia (NH
O 1/10 v/v) until a precipitate appeared. After centrifugation, the pre-
cipitate was recrystallized in EtOH at 4 °C.
2
O. The so-
H NMR spectra were recorded at 200 MHz or 300 MHz on a Bruker
spectrometer, and 13C NMR spectra at 50.2 MHz, in CDCl
3
/
3
.
H
2
4
Data were reported as follows: chemical shifts in ppm from Me Si as
an internal standard, integration, multiplicity, coupling constants (Hz),
and assignments. 13C NMR spectra were measured with complete pro-
ton decoupling. Chemical shifts were reported in ppm from the residual
solvent peak as an internal standard. GC-MS spectra were obtained on a
mass selective detector HP 5970 B instrument operating at an ionizing
voltage of 70 eV connected to a HP 5890 GC with a cross linked methyl
silicone capillary column (25 m × 0.2 mm × 0.33 μm film thickness).
ESI-MS spectra were recorded on a Waters Micromass ZQ instru-
ment equipped with ESCi source. MS analyses of the complexes were
carried out in the ESI+ modality. Each vanadyl complex was dissolved
2
.1. Ligand a) 1-phenyl-4,4,4-trifluorobutane-1,3-dione
Following the above-described procedure, acetophenone (4.0 g,
3
3.0 mmol) was treated with ethyl trifluoroacetate and EtONa in EtOH
affording a pale yellow solid (4.28 g, 60% — petroleum ether/ethyl
ether 90/10 v/v, 10% formic acid). Found C, 55.60; H, 3.27, F, 26.33%.
−
1
Calc. for C10
602. δ (200 MHz; CDCl
Ar), 8.10 (2H, m, Ar); δ (50.2 MHz; CDCl
q, J(CF) = 281 Hz), 126.9 (2 × d), 128.7 (2 × d), 132.5 (1 × s), 133.2
H F
7
3
O
2
: C, 55.56; H, 3.26, F, 26.37%. νmax(neat)/cm
, Me Si) 6.59 (1H, s, C_CH), 7.60 (3H, m,
, Me Si) 91.9 (1 × d), 117.1
6
1
H
3
4
in a d -DMSO solution, diluted in methanol/water (9:1 v/v) + formic
C
3
4
acid 0.1% v/v, and injected in the ESI source.
(
(
(
(
IR spectra of ligands and complexes (in the solid state) were
recorded on a PerkinElmer BX FT-IR and a Bruker Vertex 70 spectropho-
tometer, equipped with RAM-II module. FT-IR was measured by using
an anvil ATR cell. Laser wavelength of the Raman module is 1064 nm.
1 × d), 176.9 (q, J(CF) = 36.5 Hz), 186.0 (1 × s). Mass spectrometry
MS) (Electronic impact (EI), 70 eV): m/z (%) = 216 (76, M ), 147
100), 69 (94); m.p. 37.8 °C–39 °C.
+

28
S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
DFT calculations were performed by using the hybrid B3LYP func-
tional and the 6-311g(d,p) basis set for all elements, by means of Gauss-
ian 09 package [25].
Cells were seeded in 100 μL complete medium at appropriate densi-
ty (1500–2000 cells/well) in 96-well plastic culture plates in triplicate.
The following day, after serial dilutions, 100 μL of each compound, li-
Optical spectra of the complexes dissolved in acetone, methanol and
DMSO were recorded on a double-beam UVICAM300 spectrophotome-
ter (ThermoSpectronic). All solvent were bubbled with argon for
gand and VOSO
tichannel pipette. Vehicle and medium-only containing wells were
added as controls. Plates were incubated at 37 °C in 5% CO for 6 days,
4
in serum-free medium were added to cells with a mul-
2
1
5 min, in order to remove oxygen.
7 K EPR spectra of saturated solution of each complex in acetone
after which cell viability was assessed by ATP content using the
CellTiter-Glo Luminescent Assay (Promega). All luminescence measure-
ments (indicated as relative light units) were recorded on a Victor X5
multimode plate reader (PerkinElmer).
7
were recorded on a CW-EPR spectrometer ESP300E (Bruker) equipped
with a cylindrical cavity. Experimental settings were as follows: micro-
wave frequency ~9.5 GHz; modulation frequency 100 KHz; modulation
amplitude 4 G; microwave power 2 mW; and time constant 163 ms. All
spectra were simulated with the EPRSim32.03 software [26] whose spin
Hamiltonian takes into account second order effects typical of vanadyl
systems.
3. Results
Fig. 1 reports the structural formulas of six new vanadyl com-
plexes that were synthesized either using commercially available
acac-derivatives or ligands synthesized by the authors.
X-ray crystal structures were determined according to the following
procedure.
For all compounds the intensity data were collected at 153 K on
an Oxford Diffraction Gemini R-Ultra diffractometer equipped with
nitrogen low temperature device and Enhanced Ultra (Cu) X-ray
Source (Agilent Technologies). The intensities were corrected for ab-
sorption with the numerical correction based on Gaussian integra-
tion over a multifaceted crystal model. Software used: CrysAlisPro
3
.1. Synthesis and NMR/MS characterization of the ligands and the
complexes
Ligands for complexes a, b, d, e were synthesized in good yields
a: 60%; b: 59%; d: 87%; e: 71%) through a Claisen condensation applied
(
to ethyl trifluoroacetate and phenyl-, indolyl-, or 2-naphtyl etanone
derivatives. The methoxy electron donor substituent was introduced
both on the phenyl and naphthyl rings at positions 4 and 6, respectively.
All the ligands were purified and fully characterized (Supplementary
material, Figs. 1S–5S).
These ligands were subsequently employed for the synthesis of
vanadyl complexes a–f according to the protocol described in the
Experimental section. The formation of each complex was monitored
by MS and NMR spectroscopy.
(
Agilent Technologies, Version 1.171.36.20 (release 27-06-2012
CrysAlis171.NET, compiled Jul 11 2012,15:38:31) for data collection,
data reduction and absorption correction; SHELXTL(Sheldrick M.
(
1997), SHELXTL, Version 5.1, Bruker AXS inc., Madison) for data so-
lution, structure analysis and drawing preparation; SHELXL-2012
Sheldrick M. (2012), SHELXL-2012) for refinement. Details of crys-
(
tal data, data collection and refinement parameters are given in
Table 1S. Crystallographic data (without structure factors) for the
structure(s) reported in this paper have been deposited with the
Cambridge Crystallographic Data Centre (CCDC) as supplementary
publication nr. CCDC 930935 (complex e-ACETONE), CCDC 930936
1
2+
H NMR spectra of complexes a–f highlight the VO paramagnetic
effect on the ligand resonances that results in line broadening (due to
enhanced spin relaxation) without affecting their chemical shifts to a
significant extent as already reported for the [VO(acac) ] complex and
2
its derivatives [16,27,28]. Line broadening is taken as an evidence of
complexation, although the invariancy of chemical shift prevents the
extrapolation of any information on the geometry of the complex
(
complex e-DMSO), CCDC 930937 (complex f), CCDC 930938 (com-
plex b). Copies of the data can be obtained free of charge from the
CCDC (12 Union Road, Cambridge CB2 1EZ, UK; Tel.: +44-1223-
3
36408; Fax: +44-1223-336003; e-mail: deposit@ccdc.cam.ac.uk;
Web site: http://www.ccdc.cam.ac.uk).
(Fig. 2, upper and lower spectra).
Crystals suitable for X-ray analysis were obtained at room tempera-
ture, by slow evaporation of the solvent, dimethylformamide for com-
plex b, dimethylsulfoxide for complexes f and e-DMSO, and acetone for
complex e-ACETONE. The intensity data were collected using graphite-
monochromatized Mo-Kα radiation (λ = 0.71073 Å) for compound b
and mirror monochromatized Cu-Kα radiation (λ = 1.5418 Å) for f, e-
DMSO and e-ACETONE. All non-hydrogen atoms were anisotropically re-
fined, except for the DMSO-free molecule of complex f, the C and O
atoms of the free disordered DMSO molecules of complex e-DMSO and
+
The pseudomolecular ions (M + H ) and the complex adducts with
+
+
sodium and potassium (M + Na , M + K ), where M refers to the
VO
analyses of complexes a–f in the ESI+ modality. In addition, signals as-
signable to VO2 associated with a single ligand molecule and one or
two DMSO molecules were found (Supplementary material, Figs. 6S–
2
+
complex with two acac-derived ligands, were detected by MS
+
1
1S). The replacement of ligands a–f with DMSO can be explained by
taking into account the protonation effect of formic acid onto the
enolate ion that suppresses the coordination ability of the ligand and al-
lows ligand exchange. This process certainly occurs inside the MS
source, although it cannot be excluded that similar events occur in solu-
3
the disordered CF groups in complex e-ACETONE. Hydrogen atoms
were calculated and refined riding with Uiso = 1.2 or 1.5 Ueq. or of the
carbon atom connected.
1
tion or inside cells. H NMR spectra, recorded in acetone or DMSO, do
In vitro cell tests were performed on the following cell lines: HCT
not highlight the presence of significant amounts of free ligands,
which would be expected as a result of the dissociation process al-
though, this evidence does not rule out the possibility of dissociation
processes occurring on small amounts of the complexes.
1
16, HT-29, and hTERT-HME1.
HCT 116, HT-29, and hTERT-HME1 were purchased from the Amer-
ican Type Culture Collection; immortalized human podocytes were
kindly provided by Prof. Gianluca Miglio (University of Torino, Dip.to
Scienza e Tecnologia del Farmaco).
3
.2. X-ray crystal structure analysis
HCT 116, HT-29 and podocytes were cultured in DMEM medium
(
Sigma-Aldrich) supplemented with 10% fetal bovine serum (PAA Labo-
ratories GmbH). hTERT-HME1 cells were grown in DMEM-F12 medium
Sigma-Aldrich), supplemented with 2% fetal bovine serum, 20 ng/mL
In order to characterize the structure of complexes a–f in the solid
(
state, several crystallization attempts of these compounds were made
in different solvents. Finally, crystals suitable for X-ray analysis were
obtained by slow evaporation of the solvent at room temperature, for
complex b (in DMF), complex f (in DMSO) and complex e (in DMSO
and acetone that we refer to as e-DMSO and e-ACETONE, respectively).
Their molecular structures with atom labeling are reported in Fig. 3.
EGF, 10 g/mL insulin, and 100 g/mL hydrocortisone. All cell culture
media were supplemented with 50 units/mL penicillin and 50 mg/mL
streptomycin.
All vanadyl compounds and ligands were dissolved in DMSO and
4
VOSO in the appropriate medium.

S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
29
Fig. 1. Structural formulas of complexes a–f: a) 1-phenyl-4,4,4-trifluorobutane-1,3-dione; b) 1-(4-methoxyphenyl)-4,4,4-trifluorobutane-1,3-dione; c) 1-(2-naphtyl)-4,4,4-trifluorobutane-
,3-dione; d) 1-(6-methoxy-2-naphtyl)-4,4,4-trifluorobutane-1,3-dione; e) 1-(N-methyl-3-indolyl)-4,4,4-trifluorobutane-1,3-dione; f) 1-(3-thienyl)-4,4,4-trifluorobutane-1,3-dione.
1
In all cases, except complex e-ACETONE, a strongly distorted octahe-
C(4A)\O(3A)] and 0.009 for ring B [V(1)\O(2B)\C(2B)\C(3B)\
C(4B)\O(3B)]). The carbonyl units are bent towards the coordinat-
ed DMF molecule (the angles between the V(1)\O(1) bond and
rings A and B are 98.3° and 91.8°, respectively), likely due to the
great repulsion exerted by the vanadyl–oxygen electron couplets
on the four equatorial carbonyl oxygens. In the β-dicarbonyl ligands,
a slight rotation of the phenyl rings is observed (the angle between
ring A and the phenyl ring C(5A)\C(10A) is 7.8°, whereas the
angle between ring B and the phenyl ring C(5B)\C(10B) is 3.2°),
probably due to crystal packing; this rotation is too small to prevent
delocalization of charge density through the ligands, as witnessed by
the lower C(4)\C(5) distance (1.478(4) Å av.) with respect to a typ-
ical localized single bond.
dral geometry was found, with two 1,3-carbonyl units coordinated to
the metal center in the equatorial positions and the two axial coordina-
tion sites occupied by the vanadyl oxygen and a solvent molecule, re-
spectively. Conversely, complex e-ACETONE does not coordinate the
solvent and showed the typical square pyramidal geometry already
found in similar complexes [20,21], with the VO moiety in the apical po-
sition and the two β-dicarbonyl ligands in the square base.
The relative orientation of the CF moieties in the ligands shows that
3
complexes b, f and e-DMSO are cis-planar isomers while the e-ACETONE
complex is trans-planar.
The asymmetric unit of complex b contains one compound mole-
cule (Fig. 3) with one solvent molecule coordinated to the vanadium
atom. The two rings formed by complexation of the metal atom by
the 1,3-carbonyl ligands are almost planar (the mean deviation
from planarity is 0.030 for ring A [V(1)\O(2A)\C(2A)\C(3A)\
In the case of complex f, the asymmetric unit contains one com-
pound molecule with a solvent molecule coordinated to the vanadium
atom and one free DMSO molecule. Once again, the two rings formed
by complexation of the metal atom by the 1,3-carbonyl ligands are al-
most planar (mean deviation from planarity is 0.013 for ring A and
0.012 for ring B) and the carbonyl units are more bent towards the co-
ordinated DMSO molecule (the angles between the V(1)\O(1) bond
and the rings A and B are 100.9° and 100.8° respectively) with respect
to complex b. The thiophene moieties are slightly rotated with respect
to the 1,3-carbonyl rings (the angle between ring A and the thiophene
moiety C(5A)\S(9A) is 4.8° whereas the angle between ring B and
the thiophene ring C(5B)\S(9B) is 9.8°), although the C(4)\C(5)
bond lengths (1.46(1) Å av.) (Table 1) suggest charge density delocali-
zation on the whole ligand.
The asymmetric unit of complex e-DMSO contains two independent
compound molecules, each one bearing a solvent molecule coordinated
to the metal center, and two disordered free solvent molecules. In anal-
ogy to complexes b and f, the two rings formed by complexation of the
metal atom by the 1,3-carbonyl ligands are almost planar (mean devia-
tions from planarity are 0.069 for ring A(V1), 0.0196 for ring B(V1),
0.0842 for ring D(V2) and 0.0402 for ring E(V2)) and the 1,3-carbonyl
units are bent towards the bound DMSO (the angles between the
V(1)\O(1) and rings A and B are 94.4° and 101.9°, respectively and
1
4
Fig. 2. Magnified aromatic region of the H NMR spectra in methanol d of ligand c (upper
spectrum), freshly dissolved complex c (lower spectrum) and complex c after 5 day incu-
bation at room temperature (middle spectrum).

30
S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
Fig. 3. Asymmetric unit content for complexes b, f, e-DMSO and e-ACETONE with atom labeling. Displacement ellipsoids for all but hydrogen atoms are drawn at 50% probability.
the angles between the V(2)\O(2) and the rings D and E are 91.3° and
07.5°, respectively). The great difference in these angles found within a
V(1A)\O(1A) and the rings A and B are 96.6° and 93.3°, respectively).
The agreement of these values with those found in complexes b and
e-DMSO, whose metal center is six-coordinated, leads to the conclusion
that solvent coordination does not influence the square pyramidal ge-
1
same molecule suggests that ligands bending towards DMSO is due to
the combined influence of electronic effects of the vanadyl oxygen and
crystal packing. As in previous complexes, a deviation from planarity
of the carbonyl ligands is evident in complex e-DMSO. This is due to
the slight rotation of the indole moieties with respect to the 1,3-carbonyl
planes (the angle between ring A and the indole moiety C(5A)\C(13A)
is 5.7°, whereas the angle between ring B and the indole moiety C(5B)\
C(13B) is 9.8°). In analogy to complexes b and f, the C(4)\C(5) bond
lengths (1.433(5) Å av.) suggest charge density delocalization on the
whole ligands. Due to the cis-planar arrangement, the two indole moie-
ties are very close as the distances between the closest H atoms of the
phenyl rings range between 2.41 and 2.64 Å: this suggests that steric in-
teractions between the two indole moieties concur to their rotation, to-
gether with crystal packing.
As compared to the previous complexes, the species e-ACETONE
represents an exception as it exhibits a square pyramidal geometry
and does not coordinate the solvent. As in previous cases, both ligands
lie in the equatorial plane, but their relative orientation is typical of a
trans-planar isomer. The asymmetric unit contains one independent
molecule of compound and the structure shows a great crystallographic
disorder: the vanadyl group occupies two opposite positions with re-
spect to the equatorial plane defined by the 1,3-carbonyl units, with dis-
4
ometry of the VO(O ) moiety. This is also supported by the similar
bond lengths found in e-DMSO and e-ACETONE. In contrast to complex
e-DMSO, the indole moieties of e-ACETONE are almost coplanar with
the 1,3-carbonyl planes (the angle between ring A and the indole moi-
ety C(5A)\C(13A) is 2.4°, whereas the angle between ring B and the in-
dole moiety C(5B)\C(13B) is 2.0°), due probably to the trans-planar
arrangement of the ligands that minimize the steric interaction between
the indole rings. The extended structural disorder prevents a more de-
tailed molecular description of complex e-ACETONE.
Table 1 reports some relevant bond lengths and angles for com-
plexes b, f, e-DMSO and e-ACETONE. Taking into account the e.s.d's of
the reported bond distances, no significant differences on the V\O(1)
bonds are found between the complexes. As far as the 1,3-carbonyl
ligands is concerned, no significant differences are evident between
the C(2)\O(2) and C(4)\O(3) bonds. V\O(2) bond distances in
complex b are slightly longer than V\O(3), whereas they are similar
in complex f; an inverted trend is found in complexes e-DMSO and
e-ACETONE. As for the C(4)\C(5) bonds, they follow this trend:
b N f N e. This can likely be related with the electronic properties of
the substituents on the 1,3-carbonyl units: the higher electron-donor
character of the indole moiety in complex e is responsible for the
shorter C(4)\C(5) lengths and the longer V\O(3) bond. A comparison
tinct refined occupation factors of 79% and 21%. The CF
carbonyl ligand lies as well in two distinct positions, with different re-
fined occupation factors (65% and 35% for CF (A) and 58% and 42% for
CF (B)). The 1,3-carbonyl units are bent downwards with respect to
3
group of each
3
3 3 3
with complex [Sc(CF COCHCOCH ) ] that bears a non-electron donor
3
methyl substituent on the β-dicarbonyl ligand [29] supports this state-
ment (Table 1). As expected, the C(4)\C(5) length in the Sc(III) com-
plex is longer as compared to complexes b, f, and e, whereas the
the vanadyl oxygen (the angles between the V(1)\O(1) and rings A
and B are 96.2° and 93.1°, respectively and the angles between the

S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
31
Table 1
Selected bond distances (Å) of the 1,3-carbonyl units from X-ray data for complexes b, f, e-DMSO, e-ACETONE and Sc(III)(acac)
VO(acac)
3
and from DFT calculations for complexes a–f and
2
.
Complex
Bond lengths
V\O(1)
V\O(2)
V\O(3)
C(2)\O(2)
C(4)\O(3)
C(1)\C(2)
C(2)\C(3)
C(3)\C(4)
C(4)\C(5)
V\Osolv
X-ray diffraction data
b
1.596(2)
1.603(5)
1.598(3)
2.005(2)
1.998(2)
1.997(2)
2.010(5)
2.015(5)
2.011(2)
2.007(2)
2.022(2)
2.010(2)
2.024(9)
1.996(9)
1.978(6)
1.986(6)
2.089(2)
2.091(2)
1.274(4)
1.270(4)
1.268(9)
1.267(9)
1.293(4)
1.287(4)
1.283(4)
1.289(4)
1.31(1)
1.263(3)
1.268(3)
1.276(9)
1.274(8)
1.283(4)
1.272(4)
1.279(4)
1.276(4)
1.278(9)
1.278(9)
1.520(4)
1.531(4)
1.53(1)
1.52(1)
1.524(5)
1.520(5)
1.519(5)
1.516(5)
1.50(2)
1.52(2)
1.375(4)
1.372(4)
1.37(1)
1.36(1)
1.348(5)
1.351(5)
1.358(5)
1.362(5)
1.37(1)
1.38(1)
1.426(4)
1.424(4)
1.40(1)
1.41(1)
1.422(5)
1.433(5)
1.431(5)
1.429(5)
1.42(1)
1.41(1)
1.478(4)
1.478(4)
1.46(1)
2.204(2)
2.174(5)
2.190(3)
2.190(3)
2
.006(2)
2.008(5)
.021(5)
1.998(2)
.993(2)
1.996(2)
.976(2)
f
2
1.46(1)
e-DMSO
1.431(5)
1.429(5)
1.432(5)
1.441(5)
1.43(1)
1
1
.600(3)
1
e-ACETONE
1.56(2)
.58(1)
1.909(9)
1.901(9)
1
1.288(9)
1.42(1)
1
1
.944(6)
.946(6)
[Sc(CH
3 3 3
COCHCOCF ) ]
2.072(2)
2.080(2)
1.277(3)
1.275(3)
1.260(3)
1.257(3)
1.529(3)
1.535(3)
1.373(3)
1.370(3)
1.413(3)
1.414(3)
1.503(3)
1.498(3)
[29]
DFT calculation results
a
b
c
d
1.559
1.560
1.560
1.560
1.562
1.569
1.563
1.559
1.568
1.564
1.991
1.990
1.992
1.990
1.993
2.029
2.009
1.991
2.035
1.993
1.998
1.997
1.998
1.994
1.987
2.031
2.003
1.999
2.030
1.993
1.270
1.271
1.270
1.271
1.278
1.273
1.272
1.272
1.271
1.275
1.274
1.276
1.276
1.276
1.278
1.269
1.273
1.277
1.267
1.275
1.537
1.537
1.537
1.537
1.533
1.534
1.534
1.536
1.537
1.507
1.384
1.382
1.384
1.383
1.376
1.377
1.379
1.382
1.382
1.401
1.415
1.418
1.416
1.417
1.427
1.429
1.427
1.417
1.423
1.401
1.485
1.476
1.483
1.479
1.446
1.455
1.451
1.456
1.462
1.507
e
e-DMSO
e-ACETONE
f
2.294
2.572
f-DMSO
2.266
[VO(acac)
2
]
[31]
V\O(2) and V\O(3) lengths are similar. Furthermore, in all complexes,
the C(2)\C(3) bond lengths are shorter than the C(3)\C(4), by a same
extent. As the effect is the same in all compounds, it cannot be related
with the substituents on C(4). It is rather due to the CF group, although
3
it does not reflect on the C(1)\C(2) bond lengths that are all close to
the typical C\C single bond distance.
geometry with an average O(1)\V\O angle of 106°, the same found in
the crystalline [VO(acac) ] complex [31]. The angle between V\O and
2
the mean plane formed by the OC\C\CO fragment is close to 100° in
all complexes, with the only exception of complex a (103°). The six-
atom rings defined by the ketone carbonyls are not planar; the angle be-
tween the dihedral planes O(2)\V\O(3) and O(2)\C(2)\C(3)\
C(4)\O(3) is ~13° in all cases, a value very close to that found in
Finally, the distance between the oxygen of the solvent molecule co-
ordinated to the vanadium atom (V\Osolv, Table 1) is larger for DMF
2 .
[VO(acac) ] (12°) Modeling of complexes e and f with an axially coor-
(
complex b) than for DMSO (complexes f and e-DMSO) and this sug-
dinated DMSO molecule induces a deformation of the square-based pyr-
amid that results in O(1)\V\O angles of 97° instead of 106° found in
the unsolvated complexes; this value is consistent with an almost octa-
hedral arrangement.
gests a stronger coordination ability of the DMSO solvent with respect
to DMF.
3
.3. DFT calculations
Unlike the X-ray diffraction data, an influence of solvent coordination
is found on the vanadyl oxygen: solvent coordination weakens the V\O
bond that is longer as compared to the unsolvated complex (1.569 Å
in e-DMSO vs. 1.562 Å in e; 1.568 Å in f-DMSO vs. 1.559 Å in f); accord-
Optimized structures of complexes a–f show geometrical parame-
ters in reasonable agreement with those measured experimentally
Table 1).
The structures were optimized without constraints in the gas phase.
−
1
(
ingly, the computed ν(V\O) frequency is lowered (1096 cm
vs.
−
1
1127 cm , respectively). Acetone, which is a weakly coordinating
solvent, induces smaller geometry variations with respect to the
unsolvated complex: e-ACETONE exhibits O(1)\V\O angles of
101° and a slightly lengthened V\O(1) bond (1.562 Å in e vs.
1.563 Å in e-ACETONE). Distinct solvent coordination effects are
also highlighted by the Osolvent–vanadium bond length: V\Osolv distance
in e-DMSO and f-DMSO is 2.294 Å and 2.266 Å respectively, whereas it
is 2.572 Å in e-ACETONE. Comparison between cis- and trans-planar ar-
rangements of β-dicarbonyl ligands in complex e highlights inter-ligand
The discrepancies between the calculated geometries and the experi-
mental ones (determined in the crystalline state) can be easily
explained by considering the absence of intermolecular interactions
and crystal forces that make the gas-phase optimized structures more
similar to those found in solution. This is supported by previous studies
on vanadyl complexes in solution that show the good match between
computed structures and geometries inferred from ENDOR measure-
4
ments [30]. In all compounds, the VO(O ) moiety has a square pyramid

32
S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
Table 3
Absorption bands in the electronic spectra of complexes a–f in acetone, methanol and
DMSO.
Complex
Solvent
λmax (nm)
a
Acetone
MeOH
Acetone
MeOH
DMSO
Acetone
MeOH
Acetone
Acetone
MeOH
DMSO
Acetone
MeOH
743
756
743
756
779
736
751
736
736
737
790
744
756
771
606
605
601
603
620
600w
n.d.
606w
590w
585
470sh
481
473sh
480
357sh
357sh
361sh
363sh
364sh
367sh
369sh
376
382
377
375
371sh
371sh
374sh
336
346
350
350
343
347
b
487
c
495sh
487sh
495sh
480sh
n.d.
488sh
496
496
443sh
444sh
d
e
368
625
f
605w
602sh
623
416sh
415sh
424sh
350
352
358
DMSO
505
Fig. 4. Vibrational spectra (IR (top) and Raman (bottom)) of complex a. Asterisks indicate
the modes of (COCHCO) VO fragment.
2
sh: shoulder; w: weak; n.d.: not detected.
steric effects: a cis-planar configuration induces a rotation of the aro-
matic moiety around the C(4)\C(5) bond (the angles between the ar-
omatic plane and the β-dicarbonyl plane are 12° for the cis-planar
complex and 5° for the trans-planar one), in agreement with the trend
observed in the X-ray structures. This confirms the influence of steric ef-
fects over the geometry of these complexes that can be modified by
crystal packing. This also explains the geometrical discrepancies be-
tween the solid-state structures and the structures computed in the
gas phase.
Tiny differences are observed between the two C\O and V\O
bonds of the β-dicarbonyl ligands, unlike shown by X-ray data. A
more significant variation of the C\C lengths within the six-atom
ring is found, both in the free and the coordinated ligand: the
mean length of the C(3)\C(4) bond conjugated to the aromatic
fragment is 1.417 Å, whereas the C(2)\C(3) bond adjacent to the
Spectral assignments were done by comparing the spectral patterns of
the complexes and the free ligands, and were based on previous assign-
ments of vibrational spectra of vanadyl acetylacetonate complexes
[33–35]. Assignments were checked through DFT computational calcu-
lations. A comparison of the vibrational patterns of complexes a–f with
the uncoordinated ligands highlights that metal coordination does not
affect significantly the principal internal modes of the CF
ments; consequently, we focused mainly on the (COCHCO)
3
and R frag-
VO vibra-
2
tions that usually are well detectable in the spectra (Fig. 4). Table 2
reports the most relevant vibrational modes. An inspection of the fre-
quencies shows the close similarity of the vibrational pattern of the
2 2
(COCHCO) VO unit in complexes a–f as compared to [VO(acac) ]
[36,37]: this suggests that complexes a–f share a square-pyramidal ge-
ometry and a similar force-field for the C\O, C\C and V\O bonds. The
carbonyl modes are substantially similar in all complexes, too; this sup-
ports the presence of analogous coupling effects between the distinct
CF
ed in the solid state. This difference is explained by the electron-
withdrawing effect of the CF group that shifts electron density
3
group is 1.383 Å long, in good agreement with the data collect-
aromatic moieties and the COCHCO vibrational modes. The ν
mode in complexes a–f falls at higher frequency as compared to
[VO(acac) ]; the C\O bond length in complexes a–f is shorter, on aver-
age. The V_O stretching mode falls at 930–940 cm , much lower than
in crystalline [VO(acac) ]. The dependence of this vibrational mode on
the effects of neighboring species is well known [37]. The few samples
recorded in CCl solution (complexes a and c, as the other are insoluble
s
(C\O)
3
from the C\C bond of the enolic form of the ligand toward the
C(1)\C(2) simple bond [32].
2
−
1
2
3
.4. IR and Raman spectroscopies
4
Infrared and Raman spectra were recorded from samples synthe-
in this solvent) do not show any variation of the ν(V_O) stretching as
compared to the solid state. Well-reproducible spectra of e-DMSO (to
be compared with e) could not been recorded, probably due to the
strong interaction of DMSO with the vanadyl ion that leads to the for-
mation of differing species: this makes doubtful the assignment of the
V_O stretching.
sized and crystallized from ethanol solutions, a weakly coordinating sol-
vent, so to minimize the possibility of hexa-coordinated vanadium
center. The vibrational spectra of complexes a–f can be interpreted as
derived from the contributions of three molecular fragments: the
2 3
dicarbonyl (COCHCO) VO unit, the CF group and the aromatic moieties.
Table 2
IR and Raman absorptions of complexes a–f. The vibrational modes diagnostic of vanadyl coordination are reported.
[
VO(acac)
2
]
Complex a
IR
Complex b
IR
Complex c
IR
Complex d
IR
Complex e
IR
Complex f
IR
IR
Raman
Raman
1543w
Raman
1569w
Raman
Raman
Raman
s
ν (C\O)
1558s
1597vs
1597s
1576vs
1543vs
1589vs
1569s
1535s
1596vs
1572s
1535s
1589s
1576s
1544s
1538s
1597s
1574vs
1543vs
1
572vs
1548s
538s
ν
as(C\C)
as(C\O)
1530vs
1535vw
1341m
1
ν
ν
1358s
1287m-s
998s
6
5
1338sh
1293vs
947s
601s
556s
1352s
1297s-sh
939m-s
598s
1356m
1292vvs
938m
597s
527
1356m
1292m
938m-w
1341m
1292vvs
945m
600s
1352s
1295s
939s
597s
551
1352m
1295vw
939m
s
(C\C)
1293m
947m
1297m
942s
1278vs
930(sh)
604s
1280w
930m
ν(V_O)
940m(sh)
10m-w
61vvw
550
530
569
ν(V\O)
483s
524m
448w
424w
373s
525
458
414
352s
517
448
424
356s
523
531s
527
4
4
63m
24m
408
360s
416
374s
δ
O
\
V
\
O
366m
363s

S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
33
3
.5. Optical spectra
Table 4
EPR parameters for vanadyl complexes a–f in acetone at 77 K. The spin Hamiltonian pa-
rameters were determined by spectral simulation carried out with the Sim32 software
Table 3 reports the λmax of the electronic spectra of complexes a–f in
three different solvents.
All spectra were characterized by three bands in the 800–400 nm
range, with the exception of complexes c and f that exhibited a fourth
absorption around 443 nm and 416 respectively.
[
26].
Complex Solvent
Spectral
contribution
g⊥
g∥
A⊥
A∥
(cm− ·104) (cm−1·104)
1
(%)
a
b
Acetone
Acetone
DMSO
Acetone
Acetone
100
100
100
100
58.6
41.4
44.5
1.9821 1.9429
1.9815 1.9419
1.9811 1.9452
1.9818 1.9434
1.9806 1.9467
1.9827 1.9411
1.9826 1.9482
1.9817 1.9385
1.9793 1.9434
1.9815 1.9428
1.9852 1.9406
64.81
64.99
63.49
64.4
62.14
66.46
60.5
66.1
62.78
64.6
173.7
173.7
171.5
172.6
170.2
174.3
170.6
173.7
171.4
173.9
–
Depending on the solvent, the [VO(acac)
2
] absorptions reported
in the literature [38–40] fall respectively around 750–780 nm,
5
80–600 nm and 390–400 nm and the three bands are assignable
c
d
⁎
⁎
to the following transitions: b
2
→ e
π
(band I); b
2
→ b
1
(band II)
⁎
and b
2
→ a
1
(band III) according to the MO scheme proposed by
e
f
Acetone
Ballhausen [39]. These data are in substantial agreement with our ex-
perimental findings on complexes a–f, although assignments are com-
plicated by the contribution of ligands' chromophores to the spectra,
as well as by the presence of vanadyl charge-transfer bands that may
overlap to band III. Due to this fact, band III is often detected as a shoul-
der of absorption bands at lower wavelength. In addition, the band po-
sition is affected by the polarity and the coordination ability of the
solvent. In fact, replacement of acetone with methanol brought about
a slight blue-shift of band I for all complexes a–f. The shift was more
evident in the presence of DMSO, a strongly coordinating solvent.
Band II was almost insensitive to solvents and fell around 600 nm in
all conditions, except in the presence of DMSO that shifted the absorp-
tion at λ ~620 nm. A red-shift effect by DMSO was also noticed on the
band at λ ~485 nm.
55.5
DMSO
Acetone
DMSO
100
100
100
66.52
3.6. Complex stability in solution assessed by NMR and optical spectroscopy
1
The complex stability in solution was assessed by H NMR and opti-
cal spectroscopy.
1
4
H NMR spectra of complexes a–d were recorded in d -methanol,
d -acetone and d -DMSO after 5-day, 2-week and 4-week incubation
6
6
1
at RT. Based on the invariance of the H NMR pattern vs. time, complex
c dissolved in acetone or DMSO was found stable over the full time
range and over the 20–80 °C temperature range. Conversely, methanol
affected its stability, as a 5-day incubation at RT resulted in the appear-
ance of 1H NMR signals assignable to the free ligand (Fig. 2, middle
trace). This might be related with the lower coordination strength of
methanol as compared to the other two solvents. The other complexes
exhibited a similar behavior: a rough quantitative analysis of NMR
data shows that the vanadyl complexes in methanol solution exhibit
the following stability order: a, b b d b c.
Complexes c and f exhibited a fourth absorption band above
00 nm. According to Garribba et al. [19,41] this might arise from sym-
4
metry distortions, although the superposition of contributions from the
ligands and the vanadyl moiety in this spectral region make the inter-
pretation of these data not straightforward. Slight differences in the po-
sition of band at λ b 400 nm in the electronic spectra of complexes a–b
and c–d were likely due to the inductive effect of the methoxyl substit-
uent on the aromatic ring. Assignment of these bands to the ligands is
justified by their invariance upon solvent change.
In agreement with the 1H NMR evidence, the optical spectrum of
complex c in acetone did not highlight changes even after 26-day incu-
bation at room temperature, whereas significant spectral changes were
found in methanol after 5 days. Conversely, complex a was affected by
both solvents, although by methanol to a higher extent. Optical data
thus confirm the NMR experimental evidence.
The solubility of all complexes was higher in acetone as compared to
methanol.
3.7. EPR spectra
Fig. 5 reports the experimental and calculated EPR spectra of
complexes a–f dissolved in acetone recorded at 77 K. Spectral simu-
lations show that the spectral pattern of complexes a, b, c, and f
stems from a single species whereas two distinct spectral contribu-
tions, assignable to distinct species, are found in complexes d and
e. The spin Hamiltonian parameters reported in Table 4 suggest
that the single species found in complexes a, b, c, and f coincides
with one of the two species found in complexes d and e. Spectral pa-
rameters of complexes b, e and f dissolved in DMSO show that this
solvent, which is reported to coordinate to the sixth vacant site of
the vanadium center, lowers A as expected.
77 K EPR spectra of complexes a–f treated according to the protocol
employed for biological tests (i.e. dissolved in a small volume of DMSO
and diluted with the biological medium employed for cell tests) are sim-
ilar to those recorded in DMSO and have not been reported.
3.8. Cell tests
The effect of complexes a–f on the viability of two non-tumor cell
lines, hTERT-HME1 and podocytes, and two colorectal cancer cell
lines, HCT 116 and HT-29, was evaluated. Complexes a–f inhibited
hTERT-HME1 viability in a dose-dependent manner (Fig. 6A): the ef-
Fig. 5. A and B — 77 K EPR spectra of complexes a–f dissolved in acetone.
4
fect of complexes a–d and f was similar to that induced by VOSO .

34
S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
4
Fig. 6. Effect of complexes a–f on cell viability. Each cell line was treated with complexes a–f, VOSO and the respective ligands. Cell viability was estimated by determining ATP content in
three replicate wells. Results are normalized to the growth of cells treated with DMSO and are represented as mean ± SEM of at least three independent experiments.
Conversely complex e exhibited an IC50 significantly greater as compared
to VOSO (17.40 ± 10.18 vs 9.85 ± 1.20, respectively; Table 5). All
reported in Table 5 highlight that the behavior of the complexes was not
statistically different from VOSO . Once again, the ligands alone (in par-
4
4
ligands, but ligand f, influenced cell viability although their effect was
significantly lower as compared to the corresponding complex. The inhi-
bition of podocyte viability by complexes a–f was dose-dependent
ticular ligands e and f) affected cell viability, although their IC50 values
were almost one order of magnitude lower as compared to the corre-
sponding compounds.
(
Fig. 6B); a similar result was found in the presence of VOSO
4
: IC50 values
Colocancer cell lines provided a distinct response to complexes a–f.
4
Both cell lines were insensitive to VOSO . Complexes a–d at the highest
concentration (50 μM) completely inhibited HCT 116 viability, whereas
their respective ligands were inactive (Fig. 6C; Table 5). Complex e was
as effective as ligand e, whereas complex f was able to inhibit cell prolif-
eration by a 40% factor. Ligand f was ineffective. HT-29 cells were more
sensitive to complexes a–d as compared to complexes e and f. In fact,
complex e exhibited only a 40% of viability inhibition. The ligands
were completely ineffective (Fig. 6D; Table 5).
In summary, complexes a–f affected the viability of all the tested cell
lines to a different extent. Non-tumor cell lines were more sensitive
than tumor cell lines, and podocytes turned out to be the most sensitive
Table 5
IC50 values for vanadyl complexes a–f tested on 4 distinct cell lines.
Complex
IC50 (mM)a
hTERT-HME1
Podocytes
HCT 116
HT-29
29.23 ± 3.28b
32.90 ± 2.97b
31.70 ± 3.28
31.00 ± 3.28
26.03 ± 1.75
23.10 ± 1.20
N50.00
b
b
b
b
a
b
c
d
e
f
5.83 ± 1.21
4.74 ± 1.21
6.68 ± 1.26
5.64 ± 1.20
17.40 ± 1.20 b
11.04 ± 1.30
9.85 ± 1.20
1.53 ± 1.23
1.82 ± 1.26
1.97 ± 1.24
2.51 ± 1.29
2.98 ± 1.29
3.13 ± 1.38
2.12 ± 1.35
b
24.39 ± 4.28
28.70 ± 2.30b
33.40 ± 3.30b
N50.00
N50.00
N50.00
N50.00
VOSO
4
cell line (Fig. 6 and supplementary Fig. 12S). Interestingly, VOSO
4
uniquely affected normal cells and was ineffective towards tumor cell
lines (supplementary Fig. 12S).
a
Values are means ± SE; n ≥ 3.
P b 0.05 versus VOSO .
4
b

S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
35
4
. Discussion
is unclear whether an influence of the coordinated solvent plays a role
in the distortion, due to the disagreement between the X-ray and DFT-
computed structures.
Complexes a–f are characterized by ligands with increasing steric
hindrance and distinct electron density distributions. The latter is
modulated by the presence or absence of a methoxyl moiety or of
electron-rich heteroaromatic rings.
Complexation of the vanadyl moiety by ligands a–f is demonstrated
by ESI-MS and NMR experimental findings. ESI data show that, in all
cases, a 2:1 ligand:vanadyl ratio is found, although partial disruption
of the complex may occur within the ionization chamber of the ESI
instrument.
The characterization of the newly synthesized complexes was carried
out both in the solid state and in solution.
Complexes a–f in the solid state were investigated by X-ray diffrac-
tion, IR and Raman spectroscopy and the assignments were supported
by DFT calculations.
Both DFT calculations and X-ray data highlight the electron-
withdrawing effect of the CF moiety that results in shortened C(2)\
3
C(3) bond lengths as compared to the C(3)\C(4) bond, within the
1,3-dicarbonyl unit. X-ray data also highlight a significant electronic in-
fluence of the aromatic substituent on the 1,3 carbonyl moiety that re-
sults in shortened C(4)\C(5) and longer V(1)\O(3) bond lengths:
this is due to the stronger electron-donating character of the indole
moiety (in complex e) as compared to the methoxyphenyl and the thio-
phene rings of complexes b and f, respectively.
A steric intermolecular interaction is also observed between the two
more encumbering indole substituents in the cis-planar complexes that
cause a slight deviation of planarity of the 1,3-carbonyl ligands.
Finally the interaction of DMSO with the vanadium center results in
shorter V\Osolv distances with respect to the complexes with other co-
ordinated solvents or with the uncoordinated ones, in agreement with
the stronger coordinating properties of the former. DFT calculations
have also highlighted the trans-effect exerted by the solvent coordinated
in the sixth position that reflects in an elongation of the V_O bond in the
presence of the strongly coordinating solvent DMSO.
The solution state of complexes a–f was characterized by UV–Vis
and EPR spectroscopies.
A comparison between the two states is important, as vanadyl com-
plexes are known to undergo speciation equilibria that are strongly
dependent on the solvent, the pH and the ligands. These processes
may include solvent coordination and isomerization [19,20] and may
influence the pharmacological potential of the species.
4.2. Characterization of complexes a–f in solution
4
.1. Characterization of complexes a–f in the solid state
A characterization of complexes a–f in solution was carried out by
Both crystallographic and vibrational data show that, in the solid
UV–Vis and EPR spectroscopy.
state, complexes a–f exist either in square pyramidal or in the distorted
octahedral geometry, with the two acac-derived ligands lying in the
equatorial plane and thus defining a trans isomer with respect to the
A comparison between the solid and the solution state is necessary,
as vanadyl compounds are reported to undergo complex equilibria that
may result in ligand replacements by solvent molecules, isomerization,
establishment of labile or stable interactions between the metal complex
and the solvent. These phenomena are strongly dependent on the kinetic
and thermodynamic stability of the possible isomers, the coordinating
properties of solvents, pH effects (when applicable), the ligand basicity,
etc. Discriminating between isomers or mono- and bis-chelated com-
plexes is relevant, as distinct molecular arrangements may affect the
mode of interaction of vanadyl complexes with biological targets.
The optical spectra of vanadyl complexes may help to assign the co-
ordination geometry in solution and to establish whether a solvent is
coordinated or not to the vanadyl center, as they are sensitive to solvent
polarity and coordination ability [19,38,40,41].
6
th coordination site [42]. The square-planar geometry was found
when crystals were separated from non-coordinating or weakly-
coordinating solvents, such as ethanol or acetone, whereas the distorted
octahedral geometry was typical of crystals separated from stronger
coordinating solvents such as DMF or DMSO. In fact, the V_O stretching
frequencies recorded on complexes crystallized from ethanol suggest
assignments to non-solvent-coordinated species.
This is clearly highlighted by the X-ray structures of complex e, crys-
tallized from acetone or DMSO, and f, crystallized from DMSO. Based on
the crystallographic structure, DMSO coordinates the vanadium center
through the oxygen atom of the sulfoxide moiety and it occupies the
6
th coordination position that is vacant. This arrangement was con-
Optical spectra of vanadyl complexes show three bands associated
with the vanadyl moiety. Band I (around 780 nm) is known to be sensi-
tive to coordination changes along the VO2 axis. Garribba et al. [19,41]
firmed by DFT calculations on complex f and agrees with the structure
of other vanadyl complexes reported in the literature [19]. Conversely,
the 6th axial coordination site in complex e-ACETONE is free.
As ligands a–f bear asymmetric substitutions on the β-dicarbonyl
moiety, an additional cis/trans isomerism was possible in the plane.
+
report that the coordination of a solvent molecule on the axial site re-
⁎
sults in a stabilization of the e
π
level and brings about a red-shift with
respect to the square pyramidal complex. This criterion may help to dis-
criminate between 5- and 6-coordinated vanadyl complexes. In addi-
tion, band I is strongly sensitive to symmetry: a trigonal distortion
3
Two isomers could form, either with the CF group of each ligand in
cis-planar or trans-planar arrangements with respect to each other.
Crystallographic data on 4 complexes show that the cis-planar isomer
is always found in the presence of strongly coordinating solvents,
whereas complex e crystallized from acetone was a trans-planar isomer.
As the formation of the cis-planar isomer is generally unfavored for
sterical reasons, a stabilization effect exerted by the solvent molecule
coordinated to the metal center may be posited. As an alternative expla-
nation, this experimental behavior might be related with polarity: the
cis-planar and trans-planar isomers exhibit distinct dipolar momentum.
Thus, a selective crystallization might have occurred, depending on the
solvent.
that drives the system from the C
4
v symmetry, typical of square pyrami-
level [41], giving rise
dal complexes, towards C
2
v symmetry splits the e⁎
π
to two absorptions that may fall at λ N 800 nm and λ b 600 nm. Band II
(around 580 nm) is relatively weak and is sensitive to equatorial contri-
⁎
butions, as level b
1
derives from the contributions of dx2–y2 orbitals of
vanadium and orbitals from the ligands: it is almost unaffected by the
solvent. Band III falls around 390 nm and it is often found as a shoulder,
due to the superposition with the vanadyl charge-transfer bands or the
absorptions from the ligands. According to Bernal [38], a slight influence
⁎
of solvent coordination is expected on this absorption too, as level a
1
Both crystallographic data and DFT calculation show that the pyra-
midal geometry defined by the V(O)O moiety is distorted and the
4
contains contributions from the dz2 orbital of vanadium and from
ligands; nevertheless, the solvent effect is a minor one, as compared to
band I. In summary, the presence of 3 bands between 400 and 850 nm
rings defined by the carbonyl units are bent towards the 6th coordina-
tion site, both in solvent-coordinated and uncoordinated complexes.
Such distortion is likely due to the repulsion effect exerted by the
vanadyl oxygen towards the donor oxygens of the carbonyl moiety, al-
though some influence of the crystal packing has also to be invoked. It
4
is taken as an evidence of C v symmetry (square pyramidal complex)
[19,41], whereas 4 absorptions suggest a trigonal distortion. Depending
on the extent of distortion, a corresponding splitting of g values in the
EPR spectra may be found [43].

36
S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
Unambiguous spectral assignment of vanadyl bands may be done
Hypothesis iii) has never been reported to result in meaningful
changes of the A parameters and it is thus discarded. In fact, a cis/trans
isomerism in the equatorial plane keeps the first coordination sphere
of the metal unchanged: no changes of the coupling constant should
be expected in this case, according to the additivity relationship [48,49].
Finally, species 1 and 2 in complexes d and e might arise from an
equilibrium between the solvent-coordinated and non-coordinated
complex on the axial coordination site. According to Gorelsky et al.
[50] the presence of a solvent molecule in the axial position should con-
sistently reduce A values, whereas A in species 1 is only slightly lower
than in species 2. In addition, a comparison with the X-ray structure of
complex e-ACETONE highlights the absence of solvent molecules coor-
dinated to the metal center in such position. By analogy, the axial coor-
dination of acetone seems unlikely to occur in the complex in solution
and may be ruled out.
based on the relationship between the optical transition energy and
the g values, as reported by Bernal [38]. According to this equation, the
absorption at ~730–790 nm corresponds to g ~1.98 and is assignable
to transition I; whereas the absorption at ~590–620 nm corresponds to
g ~1.94 and is assignable to transition II. The latter assignment is sup-
ported by the negligible solvent-dependence expected for this transition.
As band I of complexes a–f is not splitted and the EPR data are consistent
4
with an axial rather than a rhombic symmetry, a C v symmetry is pre-
dicted for all complexes a–f in acetone solution.
Assignments of bands between 500 and 430 nm is more difficult, as
the absorption maxima do not correspond to those expected for band III.
In fact, Rangel et al. [43] report that band III is often hidden by other ab-
sorptions, i.e. charge-transfer bands and absorption from the ligands.
Absorptions at λ b 400 nm are likely to arise mainly from the
ligands: the almost coincidence of the absorption maxima in similar
ligands (e.g. ligands a and b, or c and d) and the absence of solvent ef-
fects support this conclusion and rule out their assignment to transition
III.
Finally, the marked red-shift of band I in complexes dissolved
in DMSO suggests that a molecule of this strongly coordinating sol-
vent enters the first coordination sphere of vanadium, in agreement
with X-ray data available on complex e-DMSO. The increasing red-
shift of band I with increasing coordination strength of the solvent
As a conclusion, the overall optical and EPR evidence suggests that
complexes a–f are in C v symmetry: the absence of rhombicity in the
4
EPR spectra rules out dramatic symmetry distortions in the system. The
analysis of the spin Hamiltonian parameters suggests that – according
to similar complexes reported in the literature [19,42] – complexes a–f
bear both ligands in the equatorial plane, although cis-planar and
trans-planar isomers cannot be distinguished based on the spectral pa-
rameters. Conversely, the presence of mono-chelate complexes may be
tentatively postulated for complexes d and e. A conclusive assignment
of these species requires further investigations by ENDOR and pulsed
EPR spectroscopy.
(
acetone b MeOH b DMSO) agrees with the expectations [19].
UV–Vis assignments rule out dramatic symmetry distortions to-
wards the trigonal bipyramidal geometry that might be expected in 5-
coordinated complexes.
4.3. Cell tests
The analysis of the EPR data at 77 K (in acetone and DMSO) shows
spin-Hamiltonian parameters that are consistent with those reported
in the literature for vanadyl complexes with acac-ligand derivatives
Tests on cell viability were carried out on non-tumor and tumor
cells. Two colon cancer cell lines were selected as they are known to
be sensitive to platinum-based chemotherapy (e.g., FOLFOX that con-
tains oxaliplatin) and, by extension, are expected to respond to other
metal complexes, namely vanadyl compounds.
Our experimental findings showed that normal cell lines are more
sensitive to complexes a–f as compared to tumor cells. They are also re-
sponsive to vanadyl sulphate: this suggests that the effect of complexes
is mainly due to the vanadyl moiety and the influence of the ligands is
negligible in the case of non-tumor cells.
Interestingly, tumor cell lines are sensitive to complexes a–f but re-
sistant to vanadyl sulphate as well as to ligands. A specific effect related
with the whole complex, and not solely to the vanadyl moiety, may thus
be speculated. Ligands undoubtedly play a modulating role, as the ex-
perimental data show that cells are less responsive to complexes e
and f as compared to the other vanadyl compounds.
[19]. Spectral simulation highlights that all complexes, but d and e, are
characterized by a single species. Hyperfine coupling constant A and g
values of complexes a, b, c and f correspond to those of species 2 in com-
plexes d and e.
The literature data available on the equilibria undergone by the
vanadyl systems in solution brings forth four distinct hypothesis with
regard to these species: i) the presence of cis/trans isomerism in
space; ii) the presence of a mixture of bis- and mono-chelate com-
plexes; iii) the presence of cis/trans isomerism in the equatorial plane;
and iv) the presence of solvent coordination equilibria.
Hypothesis i) is unlikely, as cis/trans isomers in space should result
in rhombic spectral patterns [3,6,19,43–45] that are never associated
to complexes a–f.
A comparison with a salicylaldehyde-derivative of vanadyl com-
plexes reported by Costa-Pessoa [44] shows that the A and g values of
complexes a–f might be consistent with the bis-chelate complex (spe-
cies 2, higher A value) and the mono-chelate complex (species 1,
lower A value). According to this assignment, complexes a, b, c and f
would be the bis-chelated form, whereas the mono-chelated form
would exists only for complexes d and e, a hypothesis in agreement
with the stability data. As the EPR spectra are recorded in acetone, this
would imply the coordination of at least two acetone molecules in the
equatorial plane. Acetone is reported as a very weakly-coordinating sol-
vent [42,46]; as a consequence, its coordination in equatorial position
seems unlikely. Although, evidence reported by Costes [47] shows an
acetone molecule coordinated to the vanadium center of a Gd-vanadyl
complex. This assignment is also supported by decreased A values
with increased in plane π-bonding in 1,3-butandionate complexes
In fact, the possibility of fine-tuning the effects of vanadium (mini-
mizing adverse effects and preserving the benefits) by modulating the
structure of organic ligands is demonstrated by a number of studies
on insulin-mimetic vanadium complexes [51]. The ligands may influ-
ence the absorption, tissue uptake and intracellular mobility of vanadi-
um. More specifically, they are crucial in determining the ability of a
complex to cross biological membranes, a requirement that is mostly
important for those metals whose absorption relies on passive diffusion,
such is the case of vanadium. A good balance between lipophilicity and
hydrophilicity of the complexes is crucial in modulating the uptake. The
responsiveness of tumor cells to complexes a–f as compared to the VO2
ion might be explained in the light of these considerations. Further, the
distinct response to complexes by the two tumor cell lines is not unex-
pected, as these lines display different histological origin and, conse-
quently, different protein and enzyme expression. Typically, HT-29
cells derived from a colorectal adenocarcinoma show a basal over-
expression of COX-2, whereas HCT116 cells (isolated from a patient
with a colorectal carcinoma) do not express COX-2 constitutively.
The experimental data suggest that the effectiveness of complexes
a–f on tumor cell lines may be improved by modulating the ligand
structure, in order to get compounds that are selectively active towards
+
[44] that is reflected in the A values of species 1 vs. species 2 in com-
plexes d and e. On the other hand, ENDOR studies on [VO(acac) ] in ac-
2
etone [42,46] show that this solvent may, at the utmost, establish weak
interactions with the vanadyl oxygen or the ligands. Most probably, the
possibility for acetone to interact with the vanadium center is strongly
dependent on the electron-acceptor properties of the metal center
that are modulated by the electronic properties of ligands.

S. Sgarbossa et al. / Journal of Inorganic Biochemistry 128 (2013) 26–37
37
[
10] Y. Fu, Q. Wang, X.-G. Yang, X.-D. Yang, K. Wang, J. Biol. Inorg. Chem. 13 (2008)
001–1009.
11] J. Li, G. Elberg, D.C. Crans, Y. Shechter, Biochemistry 35 (1996) 8314–8318.
tumor cells and ineffective or little effective towards non-tumor cell
lines.
1
[
These very preliminary data highlight the need for a thorough study
[12] K.H. Thompson, C. Orvig, Coord. Chem. Rev. 219–221 (2001) 1033–1053.
[
[
[
[
[
[
13] Y. Zhao, J. Ye, H. Liu, Q. Xia, Y. Zhang, X. Yang, K. Wang, J. Inorg. Biochem. 104 (2010)
371–378.
14] D. Mustafi, B. Peng, S. Foxley, M. Makinen, G. Karczmar, M. Zamora, J. Ejnik, H.
Martin, J. Biol. Inorg. Chem. 14 (2009) 1187–1197.
15] S.S. Amin, K. Cryer, B. Zhang, S.K. Dutta, S.S. Eaton, O.P. Anderson, S.M. Miller, B.A.
Reul, S.M. Brichard, D.C. Crans, Inorg. Chem. 39 (2000) 406–416.
16] B.P. Baranwal, K. Tripathi, A.K. Singh, S. Tripathi, Spectrochim. Acta A 91 (2012)
365–369.
17] K.H. Thompson, K. Bohmerle, E. Polishchuk, C. Martins, P. Toleikis, J. Tse, V. Yuen, J.H.
McNeill, C. Orvig, J. Inorg. Biochem. 98 (2004) 2063–2070.
of the mechanism that lies behind the distinct response of the selected
cell lines to complexes and the VO2+ ion.
5
. Conclusions
In summary, vanadyl complexes a–f have been synthesized and
characterized in the solid state as well as in solution. According to pre-
vious results on vanadyl complexes with acac-derivatives, compounds
a–f display a square pyramidal geometry with two ligands in the equa-
torial plane. In the presence of strongly coordinating solvents, cis-planar
isomers seem to be favored.
18] P.-K. Hon, R.L. Belford, C.E. Pfluger, J. Chem. Phys. 43 (1965) 1323–1333.
[19] E. Garribba, G. Micera, D. Sanna, Inorg. Chim. Acta 359 (2006) 4470–4476.
[
[
20] G.R. Hanson, Y.S. Sun, C. Orvig, Inorg. Chem. 35 (1996) 6507–6512.
21] B. Nagarajan, B. Muller, O. Storcheva, K. Kohler, A. Poppl, Res. Chem. Intermed. 33
(2007) 705–724.
All complexes exhibit a good stability in solution, although EPR data
suggest the presence of both the bis- and the mono-chelate forms of
complexes d and e. Further investigation is needed to confirm this
assignment.
Preliminary test of complexes a–f on normal and tumor cells shows
that these complexes are effective in inhibiting cell viability. In normal
cell, this effect seems mainly related with the vanadyl moiety, whereas
the responsiveness of tumor cells seems to be mediated by the ligands'
properties.
[22] B. Kirste, H. van Willigen, J. Phys. Chem. 86 (1982) 2743–2749.
[
[
23] Y. Li, Y. Zhao, J. Fluoresc. 19 (2009) 641–647.
24] P. Diana, A. Carbone, P. Barraja, A. Montalbano, B. Parrino, A. Lopergolo, M. Pennati,
N. Zaffaroni, G. Cirrincione, ChemMedChem 6 (2011) 1300–1309.
[25] Gaussian 09, Revision C.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M.
A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H.
Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L.
Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R.
Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J.
Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken,
C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R.
Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G.
A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B.
Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2009
[26] T. Spalek, P. Pietrzik, J. Sojca, J. Chem. Inf. Model. 45 (2005) 18–29.
Further investigations are needed to clarify this point and conse-
quently improve the complex selectivity towards tumor cells.
Abbreviations
[
[
27] L.A. Bruno, B.M. Su, Appl. Spectrosc. 51 (1997) 755–758.
28] D.G. Pobedimskii, Sh.A. Nasybullin, P.A. Kirpichnikov, R.B. Svitych, O.P. Yablonskii,
A.L. Buchachenko, Organic Magnetic Resonance 9 (1977) 61–63.
COX-2
DMF
DFT
EI
cyclooxygenase 2
dimethylformamide
density functional theory
Electronic Impact
[29] D.W. Bennett, T.A. Siddiquee, D.T. Haworth, S.V. Lindeman, J. Chem. Cryst. 37 (2007)
207–212.
[
30] D.M. Murphy, I.A. Fallis, R.D. Farley, R.J. Tucker, K.L. Avery, D.J. Willock, Phys. Chem.
Chem. Phys. 4 (2002) 4937–4943.
ESI-MS Electron Spray Ionization Mass Spectrometry
e.s.d. estimated standard deviation
[
[
31] M. Hoshino, A. Sekine, H. Uekusa, Y. Ohashi, Chem. Lett. 34 (2005) 1228–1229.
32] S.F. Tayyari, M. Vakili, A. Nekoei, H. Rahemib, Y.A. Wang, Spectrochimica Acta Part A
66 (2007) 626–636.
[
33] K. Nakamoto, Y. Morimoto, A.E. Martell, J. Am. Chem. Soc. 83 (1961) 4533–4536.
Appendix A. Supplementary data
[34] R.C. Fay, T.J. Pinnavaia, Inorg. Chem. 7 (1968) 508–514.
[
35] I. Diaz-Acosta, J. Baker, J.F. Hinton, P. Pulay, Spectrochimica Acta Part A 59 (2003)
63–377.
36] B. Vlčková, B. Strauch, M. Horák, Coll. Czech. Chem. Comm. 52 (1987) 686–695.
3
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jinorgbio.2013.07.015.
[
[37] W. Linert, E. Herlinger, P. Margl, R. Boča, J. Coord. Chem. 28 (1993) 1–16.
[
[
38] I. Bernal, P.H. Rieger, Inorg. Chem. 2 (1963) 256–259.
39] C.J. Ballhausen, H.B. Gray, Inorg. Chem. 1 (1962) 111–122.
References
[40] J. Selbin, Chem. Rev. 65 (1965) 153–175.
41] E. Garribba, G. Micera, A. Panzanelli, Inorg. Chem. 42 (2003) 3981–3987.
[
[
[
1] R.K.B. Devi, S.P. Devi, R.K.H. Singh, Spectrosc. Lett. 45 (2012) 93–103.
2] J. Korbecki, I. Baranowska-Bosiacka, I. Gutowska, D. Chlubek, Acta Biochim. Pol. 59
[42] D. Mustafi, M.W. Makinen, Inorg. Chem. 44 (2005) 5580–5590.
[43] M. Rangel, A. Leite, M.J. Amorim, E. Garribba, G. Micera, E. Lodyga-Chruscinska,
Inorg. Chem. 45 (2006) 8086–8097.
(2012) 195–200.
[3] E.G. Ferrer, M.V. Salinas, M.J. Correa, L. Naso, D.A. Barrio, S.B. Etcheverry, L. Lezama, T.
Rojo, P.A.M. Williams, Biol. Inorg. Chem. 11 (2006) 791–801.
[44] J. Costa Pessoa, I. Cavaco, I. Correia, I. Tomaz, T. Duarte, P.M. Matias, J. Inorg. Biochem.
80 (2000) 35–39.
[
4] M.S. Molinuevo, A.M. Cortizo, S.B. Etcheverry, Cancer Chemother. Pharmacol. 61
[45] B.D. Howes, C. Kuhlmeyer, R. Pogni, R. Basosi, Magn. Reson. Chem. 37 (1999) 538–544.
[46] M.W. Makinen, M.J. Brady, J. Biol. Chem. 277 (2002) 12215–12220.
[47] J.P. Costes, F. Dahan, B. Donnadieu, J. Garcia-Tojal, J.P. Laurent, Eur. J. Inorg. Chem.
(2001) 363–365.
[48] N.D. Chasteen, in: L.J. Berliner, J. Reuben (Eds.), Biological magnetic resonance, vol. 3,
Plenum Press, New York, 1981, pp. 53–119.
[49] T.S. Smith II, R. LoBrutto, V.L. Pecoraro, Coord. Chem. Rev. 228 (2002) 1–18.
[50] S. Gorelsky, G. Micera, E. Garribba, Chem. Eur. J. 15 (2010) 8167–8180.
[51] K.H. Thompson, C. Orvig, Dalton Trans. (2000) 2885–2892.
(2008) 767–773.
[
[
5] A.M. Evangelou, Crit. Rev. Oncol. Hematol. 42 (2002) 249–265.
6] J. Benítez, L. Guggeri, I. Tomaz, J. Costa Pessoa, V. Moreno, J. Lorenzo, F.X. Avilés, B.
Garat, D. Gambino, J. Inorg. Biochem. 103 (2009) 1386–1394.
7] G.T. Morgan, H.W. Moss, J. Chem. Soc. 103 (1914) 78.
8] D.C. Crans, J. Inorg. Biochem. 80 (2000) 123–131.
[
[
[
9] N. Butenko, A.I. Tomaz, O. Nouri, E. Escribano, V. Moreno, S. Gama, V. Ribeiro, J. Paulo
Telo, J. Costa Pessoa, I. Cavaco, J. Inorg. Biochem. 103 (2009) 622–632.