Iron(III) salicylates of dipicolylamine bases showing photo-induced anticancer activity and cytosolic localization

Article abstract of DOI:10.1016/j.poly.2015.10.026

An iron(III) salicylate having a dipicolylamine base (andpa) with a photoactive anthracenyl moiety is prepared, characterized, and studied for its photo-induced anticancer activity and cellular localization in HeLa and MCF-7 cells. Its phenyl analogue is

Full text of DOI:10.1016/j.poly.2015.10.026

Polyhedron 102 (2015) 668–676
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Iron(III) salicylates of dipicolylamine bases showing photo-induced
anticancer activity and cytosolic localization
a,
Aditya Garai a, Ila Pant b, Paturu Kondaiah b, , Akhil R. Chakravarty
⇑
⇑
a Department of Inorganic and Physical Chemistry, Indian Institute of Science, Sir C.V. Raman Avenue, Bangalore 560012, India
b Department of Molecular Reproduction and Development Genetics, Indian Institute of Science, Sir C.V. Raman Avenue, Bangalore 560012, India
a r t i c l e i n f o
a b s t r a c t
Article history:
An iron(III) salicylate having a dipicolylamine base (andpa) with a photoactive anthracenyl moiety is
prepared, characterized, and studied for its photo-induced anticancer activity and cellular localization
in HeLa and MCF-7 cells. Its phenyl analogue is structurally characterized by X-ray crystallography.
The complex has a ternary structure in which the dipicolylamine ligand and salicylic acid in dianionic
form (sal) display respective tridentate and bidentate mode of coordination in [Fe(sal)(phdpa)Cl] (1).
Complex [Fe(sal)(andpa)Cl] (2) having a pendant anthracenyl moiety shows significant photocytotoxicity
Received 20 August 2015
Accepted 17 October 2015
Available online 31 October 2015
Keywords:
Iron
Salicylic acid
Crystal structure
Photocytotoxicity
Cellular imaging
in visible light (400–700 nm) giving IC50 values of 8.6 0.7 and 3.4 0.9
l
M in HeLa and MCF-7 cells,
while being essentially nontoxic in the dark (IC50 > 100
lM). The complex shows cytosolic localization
in the cancer cells. Formation of hydroxyl radicals (ÅOH) as the reactive oxygen species is evidenced from
the pUC19 DNA photocleavage studies.
Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction
complexes of other metal ions are also known to show photo-
chemotherapeutic activity [16–25]. Six-coordinate platinum(IV)
Iron–bleomycins (Fe–BLMs) as metalloglycopeptides are the
naturally occurring anticancer antibiotics [1,2]. Their successful
use in cancer therapy has led to the development of the chemistry
of synthetic iron complexes as structural models and potent
chemotherapeutic agents [3–8]. Iron–bleomycins act as chemical
nucleases in which the metal in its high-spin +2 oxidation state
activates molecular oxygen to form reactive hydroxyl species that
oxidizes the deoxyribose sugar moiety causing damage to DNA [9].
There are only few synthetic models of iron–bleomycins reported
in the literature using N-donor multidentate ligand systems stabi-
lizing the metal in its +2 and +3 oxidation states [6–8]. Unlike the
natural iron–bleomycins, these model complexes lack any selectiv-
ity towards cancer cells over normal cells. The currently available
synthetic models are thus likely to damage the healthy cells due
to their low IC50 values in the dark [6]. A convenient way to cir-
cumvent this predicament is to design iron complexes that are
nontoxic in the dark under aerobic conditions but gains activity
only when exposed to visible light of suitable wavelength. We have
earlier shown that iron complexes could be suitably designed to
show photo-induced anticancer activity [10–15]. Besides,
complexes having two trans-azide ligands are known to generate
reactive platinum(II) species on photoactivation showing signifi-
cant cytotoxicity even in cisplatin resistant cells [16]. Photocyto-
toxic nitrosyl ruthenium complexes are used to deliver toxic
nitric oxide (NO) on exposure to visible light [17]. Dirhodium(II)
complexes are reported to show cytotoxicity in visible light [18].
Copper(II) and oxovanadium(IV) complexes having photoactive
moieties are known to show excellent PDT effect in visible light
[19–22]. Photodynamic therapy (PDT) has evolved as a non-
invasive treatment modality of cancer with promising results
[23–25]. Porphyrin derivatives with the hematoporphyrin species
PhotofrinÒ is currently used as a PDT drug. With the advent of this
new methodology, metal-based photochemotherapeutic agents
have gained considerable importance as substitutes of the
porphyrin-based PDT agents [26,27].
The present work stems from our interest to design new iron
complexes which can show significant anticancer activity on
photo-irradiation while remaining essentially inactive in the dark.
We have chosen iron for being an essential metal for cellular
growth including that for cancer cells [28]. The presence of large
number iron receptors on cancer cells can be used to successfully
transport iron-based anticancer agents into the cell [29]. Besides,
there is no report on any major adverse impact of the metal in
iron–bleomycins except pulmonary fibrosis [30]. A stable iron
⇑
Corresponding authors. Tel.: +91 80 2293 3259; fax: +91 80 23600999
(P. Kondaiah). Tel.: +91 80 2293 2533; fax: +91 80 23600683 (A.R. Chakravarty).
E-mail addresses: paturu@mrdg.iisc.ernet.in (P. Kondaiah), arc@ipc.iisc.ernet.in
(A.R. Chakravarty).
http://dx.doi.org/10.1016/j.poly.2015.10.026
0277-5387/Ó 2015 Elsevier Ltd. All rights reserved.

A. Garai et al. / Polyhedron 102 (2015) 668–676
669
complex under physiological conditions can overcome the chal-
lenges associated with this metal. We have used dipicolylamine
base having a pendant photoactive moiety for achieving the
desired PDT activity. Salicylic acid in its dianionic form is used
for its bio-essential nature and for its affinity to bind iron. Using
these ligand systems we have been able to synthesize ternary
complexes that show reduced chemical nuclease activity in the
presence of reducing cellular thiols. The dipicolylamine base
(andpa) having a pendant anthracenyl moiety as a fluorophore is
used as a photosensitizer and to study cellular localization of the
complex. Herein, we report the synthesis, crystal structure and vis-
ible light-induced cytotoxicity of two iron(III) salicylates, viz. [Fe
(sal)(phdpa/andpa)Cl] (1, 2), where phdpa is (phenyl)dipicoly-
lamine (in 1) and andpa is (anthracenyl)dipicolylamine (in 2)
(Fig. 1). Significant results of this study are the observation of
remarkable PDT effect of the andpa complex 2 in visible light
(400–700 nm) in cervical cancer HeLa and breast cancer MCF-7
cells along with cytosolic localization of the complex from the
fluorescence microscopy.
Zeiss LSM5 10 apochromat confocal laser scanning microscope.
Magnetic measurements at 298 K were done using Sherwood Sci-
entific, Cambridge (U.K.), magnetic susceptibility balance.
2.2. Preparation of [Fe(sal)(phdpa/andpa)Cl] (phdpa, 1; andpa, 2)
To a methanol solution of ferric chloride (0.16 g, 1.0 mmol) was
added the dipicolylamine base (0.29 g phdpa for 1 and 0.39 g
andpa for 2, 1.0 mmol) dissolved in methanol (20 mL). The solution
was stirred for 30 min to get a precipitate of the precursor complex
[Fe (phdpa/andpa)Cl3] to which was added drop-wise a solution of
salicylic acid (H2sal, 0.14 g, 1.0 mmol) and triethylamine (0.10 g,
1.0 mmol) in 10 mL methanol to obtain a deep purple colored solu-
tion which on slow evaporation of the solvent gave a solid that was
isolated, washed with diethyl ether and finally dried in vacuum
over P4O10
.
[Fe(sal)(phdpa)Cl] (1): Yield: 0.42 g (75%). Anal. Calc. for C26H23
-
ClFeN3O3 (MW: 516.78): C, 60.43; H, 4.49; N, 8.13. Found: C, 60.69;
H, 4.27; N, 7.80. ESI-MS in MeOH: m/z 481.1038 [M ꢀ Clꢀ]+. IR (in
solid phase, cmꢀ1): 2940 (m), 2650 (m), 2370 (m), 1600 (s), 1500
(s), 1390 (s), 1280 (m), 1150 (s), 1030 (s), 975 (s), 815 (w), 770
(m), 730 (m), 550 (w), 450 (m) cmꢀ1 (s, strong; m, medium; w,
2. Experimental
weak). UV–Vis (DMF) kmax, nm (
e
, Mꢀ1 cmꢀ1) = 450 (1300), 290
2.1. Materials and measurements
(3360). Molar conductance in DMF at 25 °C: KM = 69 S m2 Mꢀ1
.
l
eff = 5.85 lB at 298 K.
[Fe(sal)(andpa)Cl] (2): Yield: 0.54 g (79%). Anal. Calc. for C34H27
The chemicals and reagents were procured from the commercial
sources (s.d. Fine Chemicals, India; Aldrich–Sigma, USA; Invitrogen
Bio Services, India). They were used without further purification.
Supercoiled plasmid pUC19 DNA (CsCl purified) was from Bangalore
Genie (India). Tris-(hydroxymethyl)aminomethane–HCl (Tris–HCl)
buffer was prepared using deionised and sonicated triple distilled
water. Solvents used for this work were purified by standard
procedures. Dipicolylamine bases, viz. N-benzyl-1-(pyridin-2-yl)-
N-[(pyridin-2-yl)methyl]-methanamine (phdpa) and 1-(anthra-
cen-9-yl)-N,N-bis(pyridin-2-ylmethyl)methanamine (andpa) were
prepared following literature methods [31,32].
-
ClFeN3O3 (MW: 616.90): C 66.20, H 4.41, N 6.81. Found: C 65.50, H
4.67, N 6.61. ESI-MS in MeOH: m/z = 580.99 [M ꢀ Clꢀ]+. IR (in solid
phase, cmꢀ1): 2950 (br), 2600 (m), 2490 (m), 2350 (s), 1660 (m),
1580 (s), 1450 (s), 1360 (m),1270 (m), 1230 (m), 1150 (m), 1000
(m), 815 (s), 750 (s), 650 (s), 570 (w), 530 (w), 470 (w), 430 (w)
cmꢀ1 (br, broad). UV–Vis (DMF) kmax, nm ( , Mꢀ1 cmꢀ1): 450
e
(1256), 390 (4450), 370 (4470), 350 (3200), 333 (2440), 300
(5350). Molar conductance in DMF at 25 °C: KM = 72 S m2 Mꢀ1
eff = 5.80 lB at 298 K.
.
l
The elemental analyses were done with a Thermo Finnigan
FLASH EA 1112 CHNS analyzer. The infrared, absorption and emis-
sion spectral measurements were done using Perkin-Elmer make
model Lambda 35, Lambda 650 and LS 55 spectrophotometer,
respectively, at 25 °C. Molar conductivity measurements were
made using a Control Dynamics (India) conductivity meter. Elec-
trochemical studies were done at 25 °C with an EG&G PAR model
253 VersaStat potentiostat/galvanostat with electrochemical anal-
ysis software 270 using a three-electrode setup that consists of a
glassy carbon working, platinum wire auxiliary and a saturated
calomel reference electrode (SCE). Tetrabutylammonium perchlo-
rate (TBAP, 0.1 M) was used as a supporting electrolyte. ESI-MS
measurements were made with a Bruker Daltonics make Esquire
300Plus ESI model. Flow cytometric analysis was performed using
a FACS Verse (Becton Dickinson (BD)) cell analyzer at FL2 channel
(595 nm). Fluorescence microscopy images were obtained from
2.3. X-ray crystallographic procedure
The crystal structure of [Fe(sal)(phdpa)Cl] H2O was obtained by
X-ray diffraction method. Crystals of 1 as a monohydrate were
obtained from a methanol solution of the complex on slow evapo-
ration of the solvent. A suitable crystal was mounted on a glass
fiber with epoxy cement and the geometric and intensity data were
collected at room temperature using an automated Bruker SMART
APEX CCD diffractometer equipped with a fine-focus 1.75 kW
sealed-tube Mo K
(width of 0.3° per frame) at a scan speed of 5 s per frame. Inten-
sity data were collected using -2h scan mode and corrected for
a X-ray source (k = 0.71073 Å) with increasing
x
x
Lorentz-polarization effects and for absorption [33]. The structure
solution was done by a combination of Patterson and Fourier tech-
niques and full-matrix least-squares refinement was done using
Fig. 1. Schematic drawings of the complexes 1 and 2.

670
A. Garai et al. / Polyhedron 102 (2015) 668–676
SHELX system of programs [34]. Hydrogen atoms of the complex that
were placed in their calculated positions were refined using a
riding model. Positional disorder of few atoms of the dianionic
salicylate ligand was observed during the structure solution and
refinement. The concerned atoms in two groups, viz. labeled-A
and labeled-B, were refined with a site occupancy of 0.5. The disor-
der treatment was done using the provisions available in the SHELX
program. The salicylate aromatic ring carbon atoms C30, C31 and
C32 and the non-coordinated oxygen atom O(3) were found to
have positional disorder and they were grouped to two parts
[A and B] each corresponding to one set. This disorder did not
have any apparent effect on the chemistry of the complex.
The non-hydrogen atoms bearing the disordered ones were
refined anisotropically. The molecular view was obtained by using
ORTEP [35].
base-pairs) was studied by agarose gel electrophoresis using the
complexes in 50 mM Tris–HCl buffer (pH 7.2) and 50 mM NaCl
containing 10% DMF in light of 532 nm using Ar–Kr mixed gas
ion laser of Spectra Physics with a laser power 100 mW using a
reported procedure [12]. Mechanistic studies were carried out
using different additives. The extent of SC DNA cleavage was mea-
sured from the intensities of the bands using the UVITEC Gel Doc-
umentation System. The observed error in measuring the band
intensities was ca. 5%.
3. Results and discussion
3.1. Synthesis and general aspects
The iron(III) salicylates [Fe(sal)(phdpa/andpa)Cl] (1, 2) were
prepared by reacting a methanol solution of the dipicolylamine
derivative with FeCl3 to obtain a precursor complex which on fur-
ther treatment with salicylic acid (H2sal) and triethylamine in
methanol formed the complexes in high yield. The complexes were
characterized from the analytical, spectral and other physicochem-
ical data (Table 1). The ESI-MS of the complexes in methanol dis-
played a prominent peak corresponding to the [M ꢀ Clꢀ]+ species.
The complexes are found to be 1:1 electrolyte in DMF giving a
molar conductivity (KM) value of ꢁ70 S m2 Mꢀ1 due to dissociation
of the chloride ligand as evidenced from the ESI-MS spectra. The
electronic absorption spectra of the complexes in DMF showed a
broad but intense peak near 450 nm with a tail extending up to
600 nm (Fig. 2). The spectral peaks in the UV region are assignable
2.4. Cellular experiments
MTT assay was done to assess the photocytotoxicity of the com-
plexes [36]. Details of the experimental procedures are given as
Supplementary information. Photo-irradiation of the cells was
done in visible light of 400–700 nm using a Luzchem photoreactor
(model LZC-1, Ontario, Canada; light fluence rate = 2.4 mW cmꢀ2
;
light dose = 10 J cmꢀ2). Molecular Devices Spectra Max M5 plate
reader was used for measuring the absorbance of formazan at
540 nm. The cytotoxicity of the test compounds was measured as
the percentage ratio of the absorbance of the treated cells over
the untreated controls. The IC50 values were determined by nonlin-
ear regression analysis (GraphPad Prism 5). Flow cytometric anal-
ysis was performed to study the rate of uptake of complex 2 into
HeLa and MCF-7 cancer cells and effect on the cell cycle by com-
plex 2. Detailed procedure is given in Supporting information.
DCFDA and annexin-V/FITC assays were used to detect generation
of cellular reactive oxygen species (ROS) and apoptotic process by
flow cytometric analysis [37]. Cell permeable DCFDA on oxidation
by ROS generates fluorescent DCF having an emission maximum at
528 nm. The distribution of DCFDA stained HeLa cells was deter-
mined by flow cytometry. Annexin-V/FITC assay was used to study
the apoptotic cell death. Detailed procedures are given as Supple-
mentary information. The fluorescence of the cells was determined
with a flow cytometer. Cells involved in the early apoptotic process
were stained with the annexin V-FITC alone. Live cells showed no
staining by either propidium iodide or annexin V-FITC and necrotic
cells were stained by both propidium iodide and annexin V-FITC.
Fluorescence microscopy was used to study uptake of the fluores-
cent complex 2 into the HeLa cells using a procedure that is given
as Supplementary material.
to the p–
p⁄ transitions involving the ligands [40]. The low energy
visible band is assignable to the ligand-to-metal charge transfer
(LMCT) transition as observed for an analogous hydroxamate com-
plex [12]. Complex 2 showed an emission band at 420 nm upon
excitation at 370 nm giving a quantum yield (u) value of 0.04 in
DMSO (Fig. 2). The u value of the andpa ligand is 0.13 under similar
experimental conditions. The IR spectra of the complexes displayed
a strong peak at 1600 cmꢀ1 corresponding to the C@O stretching
vibrations of the salicylate ligand [41]. The iron(III) complexes
are paramagnetic with five unpaired electrons giving room tem-
perature magnetic moment value of ꢁ5.75
lB. Cyclic voltammetry
of the complexes showed a quasi-reversible Fe(III)–Fe(II) redox
couple with an E value of ꢁ0.05 V versus SCE in DMF-0.1 M TBAP
½
with large
DEp value indicating sluggish electron transfer process
with the possibility of a structural change on reduction of the
metal (Fig. 3(a)). The presence of Epc near ꢀ0.5 V indicates the
redox stability of +3 oxidation state of the metal. This is likely to
Table 1
Selected physicochemical and ct-DNA binding data of the complexes 1 and 2.
2.5. DNA binding and cleavage experiments
1
2
IRa/cmꢀ1
Electronicb: kmax/nm
[
t
(C@O)]
1600
1580
ꢀ
DNA binding experiments were done in Tris–HCl buffer (5 mM,
pH = 7.2) using DMF solution of the complexes 1 and 2 using
reported procedures [12]. Calf thymus DNA (ca. 250 lM NP) was
(e
/Mꢀ1 cmꢀ1
)
450 (1300)
–
5.85
69
0.071 (640)
(2.2 0.5) ꢂ 105
1
466 (1250)
420 (0.04)
5.80
Emissionc [kem/nm] (u)
d
leff
used for the binding studies. The intrinsic equilibrium binding con-
stant (Kb) of 1 and 2 to ct-DNA were obtained by the McGhee–von
Hippel (MvH) method using the expression of Bard et al. by mon-
itoring the change of the absorption intensity of the spectral band
with increasing concentration of ct-DNA [38,39]. The viscosity
measurements were done using a Schott Gerate AVS 310 auto-
mated viscometer attached to a constant temperature bath at
37 °C. DNA thermal denaturation studies were done by monitoring
KMe/S m2 Mꢀ1
Ep/mV)
Kbg/Mꢀ1
Tmh/°C
72
Eff/V(
D
0.067 (650)
(4.4 0.8) ꢂ 105
3
D
a
b
c
d
e
f
In KBr phase.
Visible electronic spectral band in DMF.
Emission peak of 2 in DMSO with kex of 370 nm u, the quantum yield value.
Magnetic moment ( eff) in lB
l
.
KM, molar conductance in DMF at 25 °C.
the absorption intensity of ct-DNA (150 lM) at 260 nm using UV-
Fe(III)–Fe(II) redox couple in DMF/0.1 M TBAP at 50 mV sꢀ1 scan rate.
Equilibrium DNA binding constant determined from the UV–visible absorption
2600 Shimadzu model CPS-240A and Peltier CPS-240 A spectrom-
g
eter with a controller at an increase rate of 1 °C minꢀ1 of the
titration.
h
DNA melting temperature.
solution. The cleavage of SC pUC19 DNA (0.2 lg, 30 lM, 2686

A. Garai et al. / Polyhedron 102 (2015) 668–676
671
Table 2
Selected crystallographic data for [Fe(sal)(phdpa)Cl] H2O (1 H2O).
Empirical formula
Formula weight (g Mꢀ1
Crystal system
Space group
a (Å)
b (Å)
c (Å)
a = c (°)
C26H23ClFeN3O3
516.77
)
monoclinic
P21/n
10.2616(9)
15.6897(9)
15.6060(15)
90.0
b (°)
106.379(10)
2410.6(4)
4
V (Å3)
Z
T (K)
293 (2)
1.424
qcalc (g cmꢀ3
)
k (Å) (Mo K
a)
0.71073
0.770
l
(cmꢀ1
)
Data/restraints/parameters
F (000)
6533/0/333
1068
Goodness-of-fit (GOF)
1.048
R(Fo)a, I > 2
r
(I)[wR (Fo)]b
0.0804[0.1879]
0.1424 [0.2253]
0.896, ꢀ0.301
R (all data) [wR(all data)]
Largest diff. peak and hole (e Åꢀ3
)
Fig. 2. UV–Visible spectrum of complex 2 in DMF with the inset showing its
emission spectrum in DMSO (kex = 370 nm).
a
R =
wR =
R
||Fo| ꢀ |Fc||/
R|Fo|.
b
R
[w(Fo2 ꢀ Fc2)2]/ [w(Fo)2]}1/2. w = [r2F2o+(AP)2 + BP]ꢀ1, where P = (Fo2 + 2F2c)/
R
3, A = 0.0433; B = 0.7893.
reduce the chemical nuclease activity of the complexes in presence
of reducing agents. Addition of 3-mercaptopropionic acid to the
solution of complex 2 showed gradual decrease in the UV–visible
spectral peak absorbance of the LMCT band (Fig. 3(b)). This could
be due to reduction of the metal to its +2 oxidation state [42].
Table 3
Selected bond distances (Å) and bond angles (°) for [Fe(sal)(phdpa)Cl] H2O (1 H2O).
Fe(1)–O(1)
Fe(1)–O(2)
Fe(1)–N(1)
Fe(1)–N(2)
Fe(1)–N(3)
Fe(1)–Cl(1)
O(1)–Fe(1)–O(2)
O(1)–Fe(1)–N(1)
O(1)–Fe(1)–N(2)
O(1)–Fe(1)–N(3)
O(2)–Fe(1)–N(1)
1.907(3)
1.920(3)
2.194(3)
2.271(4)
O(2)–Fe(1)–N(2)
O(2)–Fe(1)–N(3)
O(1)–Fe(1)–Cl(1)
O(2)–Fe(1)–Cl(1)
N(1)–Fe(1)–N(2)
N(1)–Fe(1)–N(3)
N(2)–Fe(1)–N(3)
N(1)–Fe(1)–Cl(1)
N(2)–Fe(1)–Cl(1)
N(3)–Fe(1)–Cl(1)
99.27(13)
85.39(15)
102.60(11)
95.98(11)
73.61(13)
94.28(15)
76.14(17)
91.30(10)
162.48(10)
96.66(13)
3.2. Crystal structure
Complex 1, as its monohydrate, was structurally characterized
by single-crystal X-ray crystallography. It crystallized in P21/n
space group of the monoclinic crystal system. Selected crystallo-
graphic data and important bond distance and bond angle param-
eters are given in Tables 2 and 3, respectively. The structure shows
the iron in a six coordinate {FeIIIN3O2Cl} geometry in which the
metal is bonded to the tridentate N,N,N-donor (phenyl)dipicoly-
lamine ligand in a facial mode of bonding, bidentate O,O-donor sal-
icylate dianion and a chloride ligand. The coordination geometry is
significantly distorted from an octahedral structure. During struc-
ture solution, few atoms of the salicylate ligand were found to have
positional disorders. This disorder, however, had no impact on the
overall structure of this complex. The disordered atoms were
refined with same site-occupancy and were labeled as (A) and
(B). The ORTEP diagram of the complex in Fig. 4 shows both orien-
tations of the salicylate ligand. The Fe(1)–N(2) bond with the N(2)
atom in a sp3 hybridization is found to be longer than the other
two Fe–N distances. The Fe–O distances are ꢁ2.0 Å. The Fe–Cl bond
2.146(4)
2.3135(14)
90.09(14)
87.83(14)
86.00(15)
160.58(16)
172.70(14)
distance is 2.313(8) Å. The N(1)–Fe(1)–N(3) bond angle is 94.28
(16)°. The O(1)–Fe(1)–O(2) angle is 90.03(14)°. The crystal struc-
ture is found to get stabilized by non-covalent
involving two pyridyl rings at a distance of 3.8 Å between two cen-
troids of the rings. Complex 1 shows different bonding features
compared to its benzhydroxamate analogue [12]. The Fe–Cl bond
in the benzhydroxamate complex is trans to the Fe–O bond, while
complex 1 has the Fe–Cl bond trans to the Fe(1)–N(2) bond. The
p–p interaction
Fig. 3. (a) Cyclic voltammogram of complex 2 showing quasi-reversible Fe(III)–Fe(II) redox couple in DMF–0.1 M TBAP. (b) Addition of 3-mercaptopropionic acid to the
solution of complex 2 showing a gradual decrease in the UV–visible spectral peak absorbance.

672
A. Garai et al. / Polyhedron 102 (2015) 668–676
Fig. 4. An ORTEP view of [Fe(sal)(phdpa)Cl] (1) showing atom numbering scheme for the metal and hetero-atoms. The hydrogen atoms are omitted for clarity. The positional
disordered arrangements (1:1) are shown by labeling A and B.
Fig. 5. (a) Photocytotoxicity of complexes 1 and 2 in HeLa cancer cells on 4 h incubation in the dark (black circle and solid line) and after irradiation with visible light of 400–
700 nm (10 J cmꢀ2, open circle and dash line). (b) Photocytotoxicity of complexes 1 and 2 in MCF-7cancer cells on 4 h incubation in the dark (black circle and solid line) and
after irradiation with visible light of 400–700 nm (10 J cmꢀ2, open circle dash line).
UV–visible spectroscopy in aqueous DMSO (1:1 v/v). There was
Table 4
no spectral change observed even after 48 h. This suggests that
the complexes are stable in an aqueous medium thus making them
suitable for cellular studies.
IC50 values of 1, 2 and related compounds in HeLa cancer cells.
Complex
IC50
(l
M) in
IC50 (lM) in
the dark
light (400–700 nm)
1a
>100
8.6 0.7
–
4.3 0.2
–
6.2 0.1
14.6 0.7
>100
>100
7.7 0.1
>41
20.5 0.2
>100
2a
3.4. Cellular studies
[Fe(phqpy)]2+b
Photofrinc
Cisplatind
MTT assay was done to assess the cytotoxicity of the complexes
in visible light (400–700 nm, Luzchem photoreactor, light dose of
10 J cm–2) in human cervical HeLa cancer cells and breast cancer
MCF-7 cells. The complexes were found to be essentially non-toxic
[FeL0(cat)NO3]f
[Fe(BHA)(pydpa)Cl]Clg
77.0
a
The IC50 in light is on visible light irradiation (400–700 nm, 10 J cmꢀ2). The
in the dark giving IC50 values of >100
showed significant photo-induced cytotoxicity in visible light of
400–700 nm with IC50 values of 8.6 0.7 and 3.4 0.9 M in HeLa
lM. Complex 2, however,
values are >100 M for 1 and 3.4 0.9 M for 2 in MCF-7 cells in light (400–
l
l
700 nm). The complexes are inactive in darkness in MCF-7 cancer cells.
b
Value taken from Ref. [6]. Ligand phqpy is 2,20:60, 200:600,2000:6000,20000
-
l
phenylquinquepyridine.
and MCF-7 cells, respectively (Fig. 5). The control complex 1 which
lacks any photoactive moiety did not show any significant photo-
cytotoxicity. The andpa ligand was inactive with an IC50 value of
>100 lM in light. A comparison of the IC50 values of the present
and related complexes is made in Table 4 [6,12,43].
c
Values taken from Ref. [43].
Value taken from Ref. [6].
Ref. [10]. L0 is 1-(anthracen-9-yl)-N,N-bis(pyridin-2-ylmethyl)methanamine.
d
f
g
Reference [12]. BHA is benzhydroxamate anion. Ligand pydpa is (pyrenyl)
dipicolylamine.
Cellular uptake of complex 2 was studied for 2 and 4 h incuba-
tion in MCF 7 and HeLa cells (Fig. 6). After 2 h incubation the
uptake of the complex in HeLa cells was found to be ꢁ77% whereas
in MCF-7 cell line it was ꢁ95%. Even after 4 h of incubation the
uptake in HeLa was ꢁ85% whereas that of MCF-7 cells was ꢁ96%.
Therefore, complex 2 displayed better uptake in MCF-7 cells as
compared to HeLa. To investigate the role of complex 2 in cell cycle
progression, the DNA content of the cells treated with complex 2
was measured using propidium iodide (PI) in a fluorocytometer.
The experiments were carried out in light as well as in darkness.
Complex 2 in the dark did not show any significant cell death
(Sub G1 population) but upon light irradiation it showed sub G1
hydroxamate and salicylate iron(III) complexes thus differ in their
molecular structures.
3.3. Solubility and stability
The complexes were soluble in common organic solvents like
methanol, ethanol, DMF, DMSO, acetonitrile with moderate
solubility in water. They were, however, insoluble in hydrocarbon
solvents. The solution stability of the complexes was studied by

A. Garai et al. / Polyhedron 102 (2015) 668–676
673
Fig. 6. (a) FACS study for the uptake of the complex 2 in HeLa and MCF-7 cell lines. (b) Bar diagram showing cell cycle distribution (sub-G1, G1, S, and G2/M phases) of the
HeLa and MCF-7 cells treated with complex 2 and irradiated with visible light of 400–700 nm (10 J cmꢀ2), along with the dark controls.
Fig. 7. (a) The shift in the fluorescence band position compared to the untreated cells in DCFDA assay measured by FACS analysis: (i) only cells, (ii) cells + DCFDA, (iii) cells
+ DCFDA + complex 2 (in the dark), (iv) cells + DCFDA + complex 2 (on light activation) and (v) cells + DCFDA + H2O2. (b) Cells treated with propidium iodide (PI), annexin-V/
FITC conjugate and the complex. The data in the panels are for the% of healthy cells (LL), early apoptotic cells (LR), late apoptotic cells (UR) and necrotic cells (UL) in the
different quadrants.
population of 22% and 40% in HeLa and MCF-7 cells respectively.
This is in accordance with the MTT assay data wherein MCF7 cells
(breast cancer cell line) were observed to be more sensitive to
complex 2 mediating phototoxicity than the HeLa cells.
The mechanistic aspects of the cell death in HeLa cells was stud-
ied from the 20,70-dichlorodihydrofluorescein diacetate (DCFDA)
assay (Fig. 7(a)). DCFDA is a cell permeable fluorogenic probe. This
dye on oxidation by ROS forms 20,70-dichlorofluorescein (DCF) giv-
ing an emission maximum at 528 nm. The detection of DCF fluores-
cence and quantification from the fluorescence-activated cell
sorting (FACS) analysis gave the evidence of ROS generation. HeLa
cells were treated with complex 2 (20 lM) and DCFDA. The cells
kept in dark did not show any enhancement in green fluorescence
upon treatment with complex 2. However, when they were
exposed to visible light (400–700 nm), a significant shift in green
fluorescence was observed indicating ROS generation. The annexin
V-FITC/PI assay was performed with HeLa cells to explore if the cell
death is via apoptotic pathway. Propidium iodide (PI) emits in the
red region. The annexin V-FITC dye showed green fluorescence.
This methodology allowed us for quantification of the apoptotic
cells (early – annexin high, PI low; late – annexin and PI high),
viable cell (unstained, only showing auto-fluorescence) and dead

674
A. Garai et al. / Polyhedron 102 (2015) 668–676
Fig. 8. Fluorescence microscopy images of complex 2 in MCF-7 and HeLa cells recorded after 4 h incubation using propidium iodide (PI): panels (a) bright field image
(differential interference contrast); (b) correspond to fluorescence of complex 2; panels (c) correspond to PI and panel (d) correspond to the merge images. Scale bar
corresponds to 0.8 lm.
Fig. 9. (a) Plots showing the effect of addition of an increasing quantity of 1 or 2, ethidium bromide (EB) and Hoechst dye on the relative viscosity of ct-DNA (150
l
M) at
37.0 °C (EB, ethidium bromide). (b) DNA melting plot showing the increase in the DNA melting temperature upon binding to complexes giving
DTm value 1 and 3 for 1 and 2
respectively.
cell populations (stained by PI). The complex had apoptosis induc-
ing potential as evidenced from the dot plots shown in Fig. 7(b).
The percentage population of the viable cells decreased while that
in the fourth quadrant Q3 and Q4 showed an increase, indicating
the ligand-centred band of the complexes near 370 nm for 2 and
270 nm for 1. A significant hypochromicity along with minor bath-
ochromic shift of the band suggests groove binding nature of the
complexes to ct-DNA. The Kb values of 2.2 ꢂ 105 and
4.4 ꢂ 105 Mꢀ1 for 1 and 2 indicate good DNA binding strengths of
the complexes. The relative specific viscosity of ct-DNA on binding
to the complexes was obtained from viscometric titration experi-
ments. The relative specific viscosity (g/g0) of DNA gave a measure
of the increase in the contour length associated with the separation
of DNA base pairs caused by intercalation of the complex between
DNA base pairs (g and g0 are the specific viscosities of DNA in the
that complex
2 triggered apoptosis only on light exposure.
Complex 2 having a pendant anthracene moiety showing fluores-
cence was used to study the cellular uptake and localization of this
complex in the cancer cells. The cells were incubated with complex
2 in the dark for 4 h. A significant uptake of the complex in the
cytosol was evidenced from the fluorescence microscopy images
along with that of nuclear staining dye propidium iodide (PI)
(Fig. 8). The complex was found to be retained in the cytosol even
after 4 h of incubation.
presence and absence of the complex respectively). A DNA interca-
lator like ethidium bromide (EB) showed significant increase in vis-
cosity of the ct-DNA solution. In contrast, the groove or surface
binding dye induced only minor changes in the effective length
3.5. Binding and photo-cleavage of DNA
of DNA resulting smaller changes in the relative viscosity of the
1/3
ct-DNA solution. The viscometric titration plot of (
g/g0
)
versus
The binding propensity of the complexes to calf-thymus (ct)
DNA was studied by UV–visible titration and viscometric methods.
The equilibrium ct-DNA binding constant (Kb) values of the
complexes were obtained from the UV–visible absorption experi-
ments by monitoring the change in the absorption intensity of
[complex]/[ct-DNA] ratio indicated partial intercalative mode of
binding of complex 2 to ct-DNA (Fig. 9(a)). The DNA binding nature
of the complexes to ct-DNA was studied from DNA denaturation
experiments in a phosphate buffer of pH 6.8. The observed small

A. Garai et al. / Polyhedron 102 (2015) 668–676
675
diagram is shown in Fig. 10(b). The hydroxyl radical scavengers
showed significant inhibition, while singlet oxygen quenchers
showed no apparent effect on the DNA cleavage activity. The
photo-induced DNA cleavage activity involving the absorption
band in the visible region may follow a metal-assisted mechanistic
pathway forming the hydroxyl radicals. The photo-redox pathway
is a possibility in which reduction of Fe(III) on photo-activation
+
could form a charge separated Fe2+ꢀLÅ (ligand radical cation) spe-
ꢀ
cies that reduces O2 to OÅ2 by the reactive Fe2+ ion with subsequent
formation of hydroxyl radicals (HOÅ) in the reaction: 3O2Åꢀ + 2H+ ?
HOÅ + HOꢀ + 2O2 [4]. This pathway is facilitated in the presence of a
quasi-reversible Fe(III)–Fe(II) redox couple. The ROS involved in
the cellular damage could be reactive hydroxyl species as evi-
denced from the plasmid DNA photocleavage study (see Fig. 10).
Fig. 10. (a) Gel diagram showing the cleavage of SC pUC19 DNA (0.2
lg, 30 lM) by
[Fe(sal)(phdpa/andpa)Cl] (50 M, L = phdpa, 1; andpa, 2) and ligands in 50 mM
l
Tris–HCl/NaCl buffer (pH = 7.2) containing 10% DMF on addition of reducing agent
in dark (lanes 7–10) and on irradiation with light of 532 nm (2 h exposure) (lanes
1–6) with controls: lane 1, DNA control; lane 2, DNA + FeCl3; lane 3, DNA + phdpa;
lane 4, DNA + andpa; lane 5, DNA + 1; lane 6, DNA + 2; lane 7, DNA + MPA; lane 8,
DNA + MPA + 2; lane 9, DNA + H2O2; lane 10, DNA + H2O2 + 2. (b) The gel diagram
showing the photocleavage of pUC19 DNA in the presence of different additives.
The lanes are: lane1, DNA control; lane 2, DNA + 2; lane 3, DNA + 2 + DABCO
(0.5 mM); lane 4, DNA + 2 + TEMP (0.5 mM); Lane 5, DNA + 2 + NaN3 (0.5 mM); lane
4. Conclusions
The iron(III) salicylate of a dipicolylamine base with a pendant
photoactive anthracenyl moiety shows remarkable PDT effect with
an IC50 value of ꢁ3 and ꢁ8
lM in MCF-7 and HeLa cancer cells
while being essentially non-toxic in the dark. The complex also
shows cytosolic localization with the possibility of mitochondrial
DNA being the target. The complexes show excellent ct-DNA bind-
ing strengths. Photocleavage of plasmid DNA involves formation of
hydroxyl radicals. The iron centre in a ternary structure plays an
important role as the ligands alone are inactive. A Fenton-type
mechanism seems to be operative in a similar way as is known
for iron–bleomycins. The complexes show only moderate chemical
nuclease activity when compared to the iron-based model com-
plexes of bleomycins. The results are of significance considering
that the iron-based anticancer agents showing cytosolic localiza-
tion and remarkable PDT activity in visible light (400–700 nm)
are potentially suitable for photo-chemotherapeutic applications.
6, DNA + 2 + L-Histidine (0.5 mM); lane 7, DNA + 2 + DMSO (0.5 mM); lane 8, DNA
+ 2 + KI (0.5 mM); lane 9, DNA + 2 + catalase (4 units); lane 10, DNA + 2 + SOD (5
units).
shift in the DNA melting temperature suggests groove binding nat-
ure of the complexes when compared to that of ethidium bromide
that gave a large value of
calator (Fig. 9(b)).
DTm which is well known as DNA inter-
The chemical nuclease activity of the complexes was studied in
the presence of hydrogen peroxide (H2O2, 200 M) as an oxidizing
agent and 3-mercaptopropionic acid (MPA, 1.0 mM) as a reducing
agent using supercoiled (SC) pUC19 DNA (0.2 g, 30 M) in
l
l
l
50 mM Tris–HCl/50 mM NaCl buffer (pH 7.2). The extent of DNA
cleavage was estimated from the gel electrophoresis data shown
in Fig. 10(a). The iron(III) complexes showed only moderate chem-
ical nuclease activity in the presence of MPA due to less possibility
of the formation of the reactive iron(II) species in the presence of
high negative Epc value of ca. ꢀ0.5 V versus SCE. Iron(II) species
formed due to one-electron reduction of iron(III) complex can
readily generate hydroxyl radicals by activating molecular oxygen
as is known for the activity of iron–bleomycins [9]. Control exper-
iments using the dipicolylamine ligand, salicylic acid, H2O2 and
MPA alone did not show any apparent cleavage of DNA under sim-
ilar experimental conditions. The photo-induced DNA cleavage
activity of the complexes was studied using green light of
532 nm from a CW Ar–Kr mixed gas ion laser and the gel diagram
is shown in Fig. 10(a) along with the controls. Complex 2 is an effi-
Acknowledgements
A.R.C. is thankful to the Department of Science and Technology
(DST), Government of India, for funding (SR/S5/MBD-02/2007) and
for J.C. Bose National fellowship. Thanks are due to the Alexander
von Humboldt Foundation (Germany) for donation of an electro-
chemical system and DST for a CCD diffractometer facility. We
thank D.K. Swain for his help in solving the crystal structure.
Appendix A. Supplementary data
CCDC 1407971 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge via
http://www.ccdc.cam.ac.uk/conts/retrieving.html, or from the
Cambridge Crystallographic Data Centre, 12 Union Road, Cam-
bridge CB2 1EZ, UK; fax: (+44) 1223 336 033; or e-mail:
deposit@ccdc.cam.ac.uk. Supplementary data associated with this
article can be found, in the online version, at http://dx.doi.org/10.
1016/j.poly.2015.10.026.
cient photocleaver of SC DNA. A 50 lM solution of this complex
showed essentially complete cleavage of the SC DNA to its nicked
circular (NC) form upon irradiation with green light of 532 nm
for 2 h. The complexes did not show any DNA cleavage activity in
the dark. Control experiments with the metal salt or the ligands
alone showed no apparent photocleavage of DNA. The complexes
bind to the major groove of DNA. Addition of DNA minor groove
blocker distamycin did not inhibit the DNA cleavage activity. Addi-
tion of methyl green as the DNA major groove blocker significantly
reduced the cleavage activity of the complexes.
References
[1] J. Chen, J. Stubbe, Nat. Rev. Cancer 5 (2005) 102.
[2] H. Umezawa, Prog. Biochem. Pharmacol. 11 (1976) 18.
[3] R.P. Hertzberg, P.B. Dervan, J. Am. Chem. Soc. 104 (1982) 313.
[4] T.A. Berg, B.L. Feringa, G. Roelfes, Chem. Commun. (2007) 180.
[5] A. Mukherjee, S. Dhar, M. Nethaji, A.R. Chakravarty, Dalton Trans. (2005) 349.
[6] E.L.M. Wong, G.S. Fang, C.M. Che, N. Zhu, Chem. Commun. (2005) 4578.
[7] D.Y. Dawson, S.E. Hudson, P.K. Mascharak, J. Inorg. Biochem. 47 (1992) 109.
[8] M. Otsuka, T. Masuda, A. Haupt, M. Ohno, T. Shiraki, Y. Sugiura, K. Maeda, J. Am.
Chem. Soc. 112 (1990) 838.
The mechanistic aspects of the DNA photocleavage reactions
were studied using different additives, viz. 1,4-diazabicyclo[2.2.2]
octane (DABCO, 1.0 mM), 2,2,6,6-tetramethylpiperidine (TEMP,
1.0 mM), sodium azide (NaN3, 1.0 mM) and
as singlet oxygen quenchers and DMSO (4
L
-histidine (1.0 mM)
l
L), KI (1.0 mM) and
[9] R.M. Burger, Chem. Rev. 98 (1998) 1153.
[10] U. Basu, I. Khan, A. Hussain, P. Kondaiah, A.R. Chakravarty, Angew. Chem., Int.
Ed. 51 (2012) 2658.
catalase (4 units) as hydroxyl radical scavengers, and superoxide
dismutase (SOD, 4 units) as superoxide radical scavenger. The gel

676
A. Garai et al. / Polyhedron 102 (2015) 668–676
[11] U. Basu, I. Pant, A. Hussain, P. Kondaiah, A.R. Chakravarty, Inorg. Chem. 54
(2015) 3748.
[12] A. Garai, U. Basu, I. Khan, I. Pant, A. Hussain, P. Kondaiah, A.R. Chakravarty,
Polyhedron 73 (2014) 124.
[13] U. Basu, I. Khan, A. Hussain, B. Gole, P. Kondaiah, A.R. Chakravarty, Inorg. Chem.
53 (2014) 2152.
[14] A.Garai,U. Basu,I.Pant,P.Kondaiah,A.R.Chakravarty,J.Chem.Sci.127(2015)609.
[15] M. Roy, S. Saha, A.K. Patra, M. Nethaji, A.R. Chakravarty, Inorg. Chem 46 (2007)
4368.
[16] F.S. Mackay, J.A. Woods, P. Heringová, J. Kašpárková, A.M. Pizarro, S.A.
Moggach, S. Parsons, V. Brabec, P.J. Sadler, Proc. Nat. Acad. Sci. USA 104
(2007) 20743.
[17] N.L. Fry, P.K. Mascharak, Acc. Chem. Res. 44 (2011) 289.
[18] H.T. Chifotides, K.R. Dunbar, Acc. Chem. Res. 38 (2005) 146.
[19] (a) S. Banerjee, A. Dixit, R.N. Shridharan, A.A. Karande, A.R. Chakravarty, Chem.
Commun. 50 (2014) 5590;
(b) S. Banerjee, A. Hussain, P. Prasad, I. Khan, B. Banik, P. Kondaiah, A.R.
Chakravarty, Eur. J. Inorg. Chem. (2012) 3899.
[20] S. Banerjee, A. Dixit, A.A. Karande, A.R. Chakravarty, Eur. J. Inorg. Chem. (2015)
447.
[26] (a) N.A. Smith, P.J. Sadler, Phil. Trans. Royal Soc. A 371 (2013) 20120519;
(b) S. Banerjee, A.R. Chakravarty, Acc. Chem. Res. 48 (2015) 2075.
[27] U. Schatzschneider, Eur. J. Inorg. Chem. (2010) 1451.
[28] G. Cair, G. Cairo, F. Bernuzzi, S. Recalcati, Genes Nutr. 1 (2006) 25.
[29] J.L. Heath, J.M. Weiss, C.P. Lavau, D.S. Wechsler, Nutrients 5 (2013) 2836.
[30] J. Hay, S. Shahzeidi, G. Laurent, Arch. Toxicol. 65 (1991) 81.
[31] D. Rojas, A.M. García, A. Vega, Y. Moreno, D. Venegas-Yazigi, M.T. Garland, J.
Manzur, Inorg. Chem. 43 (2004) 6324.
[32] L.A. Mullice, R.H. Laye, L.P. Harding, N.J. Buurma, S.J.A. Pope, New J. Chem. 32
(2008) 2140.
[33] N. Walker, D. Stuart, Acta Crystallogr., Sect. A 39 (1983) 158.
[34] G.M. Sheldrick, SHELX-97, Programs for crystal structure solution and
refinement, University of Göttingen, Göttingen, Germany, 1997.
[35] C.K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak
Ridge, TN, 1976.
[36] L. Zhang, W. Liu, K.K. Szumlinski, J. Lew, Proc. Natl. Acad. Sci. USA 109 (2012)
20041.
[37] T. Takanashi, Y. Ogura, H. Taguchi, M. Hashizoe, Y. Honda, Invest. Ophthalmol.
Vis. Sci. 38 (1997) 2721.
[38] J.D. McGhee, P.H. Von Hippel, J. Mol. Biol. 86 (1974) 469.
[39] M.T. Carter, M. Rodriguez, A.J. Bard, J. Am. Chem. Soc. 111 (1989) 8901.
[40] T.A. Sherstneva, K.P. Birin, V.V. Arslanov, Prot. Met. 49 (2013) 72.
[41] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination
Compounds, John Wiley & Sons, New York, 1997.
[21] T.K. Goswami, B.V.S.K. Chakravarthi, M. Roy, A.A. Karande, A.R. Chakravarty,
Inorg. Chem. 50 (2011) 8452.
[22] A. Bhattacharyya, A. Dixit, K. Mitra, S. Banerjee, A.A. Karande, A.R. Chakravarty,
Med. Chem. Commun. 6 (2015) 846.
[23] R. Bonnett, Chemical Aspects of Photodynamic Therapy, Gordon & Breach,
London, UK, 2000.
[42] A.W. Addison, S. Burman, C.G. Wahlgren, O.A. Rajan, T.M. Rowe, E. Sinn, J.
Chem. Soc. Dalton Trans. (1987) 2621.
[24] J.F. Lovell, T.W.B. Liu, J. Chen, G. Zheng, Chem. Rev. 110 (2010) 2839.
[25] J.P. Celli, B.Q. Spring, I. Rizvi, C.L. Evans, K.S. Samkoe, S. Verma, B.W. Pogue, T.
Hasan, Chem. Rev. 110 (2010) 2795.
[43] E. Delaey, F. van Laar, D. De Vos, A. Kamuhabwa, P. Jacobs, P. de Witte, J.
Photochem. Photobiol. B 55 (2000) 27.