Synthesis, characterization, and cytotoxicity studies of Cu(II), Zn(II), and Fe(III) complexes of N-derivatized 3-hydroxy-4-pyridiones

Article abstract of DOI:10.1016/j.jinorgbio.2013.12.003

The deleterious role of metal ions in Alzheimer's disease has inspired the study of various metal chelators. We previously showed the synthesis and in vitro activity of several bidentate hydroxypyridinone compounds, including 3-hydroxy-2-methyl-1-phenyl-4(1H)-pyridinone (1), 1-(4-aminophenyl)-3-hydroxy-2- methyl-4(1H)-pyridinone (2), and 1-(2-benzothiazolyl)-3-hydroxy-2-methyl-4(1H)- pyridinone (3). While the focus has been on the Cu(II) ion, the other biorelevant metals, Zn(II) and Fe(III) have been largely neglected. Herein, we report the synthesis of Zn(II) and Fe(III) complexes of ligands 1, 2, and 3, and their characterization by infrared (IR) spectroscopy, high resolution mass spectrometry (HR-MS), elemental analysis, and NMR, where applicable. Solid state structures of Zn(1)₂, Fe(1)₃, and Cu(3)₂ are analyzed with X-ray crystallography. The cytotoxicity of pro-ligands 1, 2, and 3, and the three metal complexes of 2 are examined in a neuronal cell line to determine the effect of metal chelation on toxicity of the compounds.

Full text of DOI:10.1016/j.jinorgbio.2013.12.003

Journal of Inorganic Biochemistry 132 (2014) 59–66
Contents lists available at ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Synthesis, characterization, and cytotoxicity studies of Cu(II), Zn(II), and
Fe(III) complexes of N-derivatized 3-hydroxy-4-pyridiones
Maria A. Telpoukhovskaia, Cristina Rodríguez-Rodríguez, Lauren E. Scott, Brent D.G. Page,
⁎
Brian O. Patrick, Chris Orvig
Medicinal Inorganic Chemistry Group, Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, BC V6T 1Z1, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 30 July 2013
Received in revised form 9 December 2013
Accepted 13 December 2013
Available online 22 December 2013
The deleterious role of metal ions in Alzheimer's disease has inspired the study of various metal chelators. We
previously showed the synthesis and in vitro activity of several bidentate hydroxypyridinone compounds, includ-
ing 3-hydroxy-2-methyl-1-phenyl-4(1H)-pyridinone (1), 1-(4-aminophenyl)-3-hydroxy-2-methyl-4(1H)-
pyridinone (2), and 1-(2-benzothiazolyl)-3-hydroxy-2-methyl-4(1H)-pyridinone (3). While the focus has
been on the Cu(II) ion, the other biorelevant metals, Zn(II) and Fe(III) have been largely neglected. Herein, we
report the synthesis of Zn(II) and Fe(III) complexes of ligands 1, 2, and 3, and their characterization by infrared
(IR) spectroscopy, high resolution mass spectrometry (HR-MS), elemental analysis, and NMR, where applicable.
Solid state structures of Zn(1)2, Fe(1)3, and Cu(3)2 are analyzed with X-ray crystallography. The cytotoxicity of
pro-ligands 1, 2, and 3, and the three metal complexes of 2 are examined in a neuronal cell line to determine
the effect of metal chelation on toxicity of the compounds.
Keywords:
Hydroxypyridinone
Copper
Zinc
Iron
Metal complex cytotoxicity
Alzheimer's disease
© 2013 Elsevier Inc. All rights reserved.
1. Introduction
The title ligands are also considered for other applications. For in-
stance, 2 was part of a library of chelators investigated for inhibition of
The class of 3-hydroxy-4-pyridinone (HPO) ligands has been widely
studied for its metal-chelating ability, with one of its members, 3-
hydroxy-1,2-dimethylpyridin-4-one (0) (Fig. 1) being used to treat an
iron overload disorder, β-thalassemia [1]. This class of pro-ligands is
being investigated in our laboratory for potential use in Alzheimer's dis-
ease (AD) through chelating to biorelevant metals Cu(II), Zn(II), and
Fe(III) [2]. We have previously reported coordination complexes of 1,
2, and 3 (Fig. 1) with Cu(II) and determined the positive effect that
the chelators have in various preliminary in vitro assays [3,4]. For in-
stance, these chelators were seen to disaggregate amyloid-beta1–40
(Aβ) aggregates due to their interaction with Cu(II) and Zn(II). As
these three metal ions are implicated in AD, in part through interaction
with plaques or its main constituent, Aβ [5,6], as well as inducing oxida-
tive stress [7], we set out to fully characterize the zinc and iron chelation
of these ligands. For more information, the reader is directed to two re-
cent review collections —Metal Ions in Neurodegenerative Diseases in
Coordination Chemistry Reviews (2012) and Metals in Neurodegenera-
tive Diseases in Metallomics (2011). Most recently, it was found that
Cu(II) increases Aβ accumulation in mouse brains, which leads to neu-
rotoxic effects [8]. As well, iron was found in increased concentrations
in hippocampi of living patients with AD over controls, and this correlat-
ed with increased tissue damage [9].
Zn(II)-dependent matrix metalloproteinases, although it did not show
significant activity [10], and it was studied as an extracting agent for
Fe(III) [11]. It should be noted that the Zn(II) and Fe(III) complexes
were not synthesized in the solid state in these studies.
The ultimate location for these potential drugs is the brain, where
they are expected to bind Cu(II), Zn(II), and Fe(III) ions. In cytotoxicity
studies, attention is generally paid to the pro-ligands, while the metal
complexes are left unexamined. Here, we look into the change in cyto-
toxicity of 2 upon chelation to Cu(II), Zn(II), and Fe(III). Although
there is indirect evidence that another chelator, PBT2 (5,7-dichloro-2-
((dimethylamino)methyl)-8-hydroxyquinoline) (currently in phase
IIb clinical trials) acts as a metal chaperone, transporting the metal ion
into the cell [12], it is valuable to look at toxicity of these species, even
if they may be transient.
2. Experimental
2.1. Materials and methods
All chemicals and solvents were purchased from commercial suppliers
(Aldrich, Alfa Aesar, Eastman Organic Chemicals, Mallinckrodt) and used
without further purification. Water was purified to 18.2 MΩ·cm by Elga
Ultra Pure Lab System. BEnd.3 cell line was provided by Dr. Kelly
McNagny's laboratory at the University of British Columbia. Supplies for
cytotoxicity studies were purchased from Invitrogen Life Technologies.
Pro-ligands 3-hydroxy-2-methyl-1-phenyl-4(1H)-pyridinone (1) and its
⁎
Corresponding author. Tel.: +1 604 822 4449.
E-mail address: orvig@chem.ubc.ca (C. Orvig).
0162-0134/$ – see front matter © 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jinorgbio.2013.12.003

60
M.A. Telpoukhovskaia et al. / Journal of Inorganic Biochemistry 132 (2014) 59–66
removed in vacuo and H2O (~5 mL) was added to precipitate the com-
plex after an overnight reaction. The resultant red solid Fe(1)3
(0.0171 g, 41% (first crop)), was collected by a fine frit and washed
with H2O and diethyl ether, and dried in vacuo. X-ray quality crystals
were obtained from a saturated solution in an acetonitrile/water
FeN3O6 (M + H+) calcd (found):
56
mixture. HR-ESIMS m/z for C
H
36 31
657.1562 (657.1554). IR (cm−1, total reflectance): 1588, 1538, 1504,
1468, 768.
2.3.4. Bis(1-(4-aminophenyl)-3-hydroxy-2-methyl-4(1H)-pyridinato)
Fig. 1. Pro-ligands 0, 1, 2, and 3.
copper(II), Cu(2)2
63
H ) calcd (found):
CuN4O4 (M + H+
23
HR-ESIMS m/z for C24
494.1015 (494.1009).
2.3.5. Bis(1-(4-aminophenyl)-3-hydroxy-2-methyl-4(1H)-pyridinato)
zinc (II), Zn(2)2
Cu(II) complex [3], 1-(4-aminophenyl)-3-hydroxy-2-methyl-4(1H)-
pyridinone (2) and its Cu(II) complex [4], and 1-(2-benzothiazolyl)-3-
hydroxy-2-methyl-4(1H)-pyridinone (3) and its Cu(II) complex [4]
were synthesized according to literature procedures.
Compound 2 (0.0516 g, 0.239 mmol) in min. amount of 50/50
MeOH:H2O solution (~15 mL) was added dropwise to a 50 mL round
bottom flask containing zinc perchlorate hexahydrate (0.0484 g,
0.130 mmol) dissolved in min. amount of the same solvent (~5 mL).
The pH of the mixture was raised to 7 with NaOH and it was left stirring
overnight. The resultant beige solid Zn(2)2 (0.0342 g, 58% (first crop))
was collected by a fine frit and washed with acetone, and dried
in vacuo. 1H NMR (300 MHz, DMSO-d6): δ = 2.03 (s, 3H), 5.46 (s,
2H), 6.29 (d, J = 6.85 Hz, 1H), 6.63 (d, J = 8.45 Hz, 2H), 6.99 (d,
J = 8.68 Hz, 2H), 7.27 (d, J = 6.85 Hz, 1H). 13C NMR (75 MHz,
DMSO-d6): δ = 13.89, 106.78, 113.59, 126.80, 128.98, 131.29, 132.57,
149.19, 155.39, 171.32. Anal. calcd (found) for C24H22N4O4Zn·2H2O: C,
54.20 (54.57); H, 4.93 (4.82); N, 10.53 (10.30%). HR-ESIMS m/z for
2.2. Instrumentation
1H NMR and 13C NMR spectra were recorded with a Bruker AV-300
spectrometer at 300.13 MHz and 75.48 MHz, respectively, at 24.7 °C
and calibrated using residual solvent peaks. High-resolution mass
spectra were obtained with Waters/Micromass LCT, and elemental
analyses (CHN) were done using a Carlo Erba Elemental Analyzer
EA 1108 by Mr. David Wong, Mr. Derek Smith, Mr. Marco Yeung,
and Mr. Marshall Lapawa. Infrared spectra were collected neat on a
Perkin Elmer FTIR spectrometer. For solid state structure analysis,
Bruker APEX DUO and Bruker X8 APEX II diffractometers with graph-
ite monochromated Mo-Kα radiation were used. Cytotoxicity stud-
ies were done with the aid of a DTX880 plate (UV–vis) reader.
C
24H23N4O644Zn (M + H+) calcd (found): 495.1011 (495.1007). IR
(cm−1, total reflectance): 3360, 3328, 1608, 1588, 1540, 1512, 1464,
1292, 764.
2.3.6. Tris(1-(4-aminophenyl)-3-hydroxy-2-methyl-4(1H)-pyridinato)
iron (III), Fe(2)3
2.3. Synthesis and characterization of metal complexes
Compound 2 (0.117 g, 0.542 mmol) in min. amount of 50/50
MeOH/H2O (~25 mL) was added dropwise to a 50 mL round bottom
flask containing iron perchlorate hexahydrate (0.0134 g, 0.029 mmol)
dissolved in the same solvent (~5 mL). The pH was increased to 7
with NaOH and the reaction was stirred overnight. The red precipitate
Fe(2)3 (0.0858 g, 68% yield) was collected by a fine frit. The solid was
recrystallized from an acetonitrile/water solution to yield single
crystals that resulted in partially solved solid-state structure as ana-
lyzed by X-ray crystallography and they were used in the following
analyses after being dried in vacuo. Anal. calcd (found) C36H33FeN6
2.3.1. Bis(3-hydroxy-2-methyl-1-phenyl-4(1H)-pyridinato)
copper(II), Cu(1)2
High resolution electrospray ionization mass spectrometry (HR-
63
ESIMS) m/z for C
H
CuN2O4 (M + H+) calcd (found): 494.0797
24 21
(494.0791).
2.3.2. Bis(3-hydroxy-2-methyl-1-phenyl-4(1H)-pyridinato) zinc (II), Zn(1)2
Compound 1 (0.0338 g, 0.168 mmol) in min. amount of MeOH
(~3 mL) and one drop of EtN3 were added dropwise to a 10 mL round
bottom flask containing zinc perchlorate hexahydrate (0.0317 g,
0.085 mmol) dissolved in min. amount of MeOH (~1 mL). The reaction
mixture was stirred overnight; solvent was removed in vacuo and H2O
(~5 mL) was added to precipitate the complex. The resultant white
solid Zn(1)2 (0.015 g, 39% (first crop)) was collected by a fine frit and
washed with H2O and diethyl ether, and dried in vacuo. X-ray quality
crystals were obtained from a saturated solution in MeOH by slow evap-
oration at room temperature. 1H NMR (300 MHz, MeOD-d4): δ = 2.19
(s, 3H), 6.64 (d, J = 6.85 Hz, 1H), 7.35–7.42 (m, 2H), 7.48 (d,
J = 7.08 Hz, 1H), 7.55–7.64 (m, 3H). 13C NMR (75 MHz, MeOD-d4):
δ = 14.81, 109.82, 127.78, 130.86, 131.11, 133.73, 135.24, 144.17,
155.04, 172.77. HR-ESIMS m/z for C24H20N2O644Zn (M+) calcd (found):
464.0715 (464.0719). IR (cm−1, total reflectance): 3204, 3537, 1608,
1590, 1538, 1506, 1468, 1290, 760.
O6·1.5H2O: C, 59.35 (59.09); H, 4.98 (4.82); N, 11.54 (10.94%). HR-
56
ESIMS m/z for C
H
36 34
FeN6O6 (M + H+) calcd (found): 702.1889
(702.1912). IR (cm−1, total reflectance): 3356, 3236, 1608, 1594,
1538, 1510, 1468, 1292, 766.
2.3.7. Bis(1-(2-benzothiazolyl)-2-methyl-3-oxy-4-pyridonato)
copper(II), Cu(3)2
63
H
CuN4O4S2 (M + H+) calcd (found):
19
HR-ESIMS m/z for C26
578.0144 (578.0140). X-ray quality crystals were obtained from a lay-
ered solution of CHCl3 under MeOH.
2.3.8. Bis(1-(2-benzothiazolyl)-2-methyl-3-oxy-4-pyridonato)
zinc (II), Zn(3)2
Compound
3 (0.0327 g, 0.127 mmol) and EtN3 (17.7 μL,
127 mmol) dissolved in min. amount of 9:1 MeOH:DCM (~10 mL)
were added dropwise to a 25 mL round bottom flask containing
zinc perchlorate hexahydrate (0.0237 g, 0.064 mmol) dissolved in
min. amount of the same solvent (~3 mL). The solution was stirred
overnight, and the volume of the solvent was reduced to one third
in vacuo. The flask was cooled in the freezer and the resultant beige
solid Zn(3)2 (0.0208 g, 55% (first crop)) was collected by a fine frit
2.3.3. Tris(3-hydroxy-2-methyl-1-phenyl-4(1H)-pyridinato)
iron (III), Fe(1)3
Compound 1 (0.0374 g, 0.186 mmol) in min. amount of MeOH
(~3 mL) and one drop of EtN3 were added dropwise to a 10 mL round
bottom flask containing iron perchlorate hexahydrate (0.0287 g,
0.062 mmol) dissolved in min. amount of MeOH (~1 mL). Solvent was

M.A. Telpoukhovskaia et al. / Journal of Inorganic Biochemistry 132 (2014) 59–66
61
and dried in vacuo. 1H NMR (300 MHz, DMSO-d6): δ = 2.33 (s, 3H),
3. Results and discussion
6.51 (d, J = 7.08 Hz, 1H), 7.57–7.70 (m, 2H), 7.88 (d, J = 6.85, 1H),
8.12 (d, J = 7.99, 1H), 8.25 (d, J = 7.77, 1H). 13C NMR (75 MHz,
DMSO-d6): δ = 13.60, 108.21, 122.82, 123.58, 126.63, 126.94,
127.31, 133.11, 134.98, 149.18, 155.05, 160.69, 173.90. Anal. calcd
(found) for C26H18N4O4S2Zn·H2O: C, 52.22 (52.47); H, 3.37 (3.28);
3.1. Synthesis
The synthesis of zinc and iron complexes was achieved by combining
metal ions with the appropriate ligand under various conditions. For pro-
ligands 1 and 3, it was performed in methanol and dichloromethane/
methanol solutions, respectively, with triethylamine as a base [4],
while for 2, a water–methanol solution with sodium hydroxide as a
base was used. It was noted that using the same procedure to synthesize
Cu(II) complexes with 1 and 3 resulted in precipitation of the desired
species. Conversely, with Zn(II) and Fe(III), the complexes were soluble
in the reaction solution and had to be isolated by reducing volume or
using more polar solvents. The relative insolubility of the Cu(II) com-
plexes presented challenges in characterization by mass spectrometry
as well as cytotoxicity, in which cases solutions containing compounds
are required to be at a certain concentration.
N, 9.37 (8.93%). HR-ESIMS m/z for C26H19N4O4S624Zn (M + H+
)
calcd (found): 579.0145 (579.0139). IR (cm−1, total reflectance):
1604, 1552, 1512, 1468, 1268, 694.
2.3.9. Tris(1-(2-benzothiazolyl)-2-methyl-3-oxy-4-pyridonato)
iron (III), Fe(3)3
Compound 3 (0.0268 g, 0.104 mmol) and EtN3 (15 μl, 108 mol)
in min. amount of 9:1 MeOH:DCM (~10 mL) were added dropwise
to a 25 mL round bottom flask containing iron perchlorate hexahy-
drate (0.0159 g, 0.034 mmol) dissolved in min. amount of the same
solvent (~3 mL). After the reaction stirred overnight, solvent was re-
duced to a third of the volume in vacuo. After the solution was cooled
in the freezer, the resultant red solid Fe(3)3 (0.0099 g, 35% (first
crop)) was collected by a fine frit and dried in vacuo. Anal. calcd
With NMR spectroscopy, it was confirmed that Zn(II) complexation
goes to completion. While it is expected that all reactions herein go to
completion, the isolated yield is lower due to the high solubility of the
complexes. It should be noted that the zinc coordination sphere is not
saturated when the 2:1 ligand to metal stoichiometry is used, thus
bridging ligands and water molecules are observed. As well, previous
metal binding studies reveal a binding constant for the 3:1 ligand to
metal species, and it is present at high pH in physiological-like condi-
tions [17,18]. When this ratio was used during reactions, a 3:1 complex
was not present in high yield and could not be isolated, although its
presence is confirmed with mass spectrometry (data not shown).
When prepared and isolated in the absence of water, the complexes
of 1 were found to be deliquescent and were challenging to manage.
Their tendency to hydrate is evident in the solid state analysis, as 1 crys-
tallized with water chelated to Zn(1)2, while each Fe(III) center has six
waters in the solid state (see Section 3.2). The presence of solvent spe-
cies and/or perchlorate ions was evident in the elemental analyses for
these two complexes, and the elemental composition could not be con-
firmed for the bulk Zn(1)2 and Fe(1)3. On the other hand, elemental
analyses for the other four title complexes confirmed the purity of the
bulk material, revealing one to three waters present per metal center.
In order to further study the metal complexes, various analytical
techniques were used. For example, comparing IR spectra of ligands to
their corresponding complexes, several observations were made
(Table 1). In metal complexes, save Zn(1)2, it was found that the peak
due to the OH stretch is absent, as is expected due to deprotonation
upon chelation. Peaks in the fingerprint region that are assigned to the
C_O and ring stretching, and CH bending modes had a bathochromic
shift upon complexation, consistent with previous observations [4]. As
well, new peaks were observed in the 600–800 cm−1 region that are
assigned to new metal–oxygen vibrations. The two peaks in the OH re-
gion for Zn(1)2 might be due to water or residual uncoordinated com-
pound, as suggested to exist in the solid state (Section 3.2).
(found) for C39H27FeN6O6S3·3H2O: C, 53.12 (53.17); H, 3.77 (3.52);
56
N, 9.53 (9.50%). HR-ESIMS m/z for C
H
39 28
FeN6O6S3 (M + H+) calcd
(found): 828.0582 (828.0583). IR (cm−1, total reflectance): 1590,
1544, 1504, 1464, 1276, 662.
2.4. X-ray data collection
The crystals were mounted on a fiber glass and the measurements
were taken with the Bruker APEX DUO diffractometer for the Zn and
the Fe, and the Bruker X8 APEX II diffractometer for the Cu complexes.
Data sets were collected and integrated using the Bruker SAINT [13]
software package. Data were corrected for absorption effects using the
multi-scan technique (SADABS) [14]. The structures were solved by di-
rect methods [15] and all non-hydrogen atoms were refined anisotrop-
ically. All O–H hydrogen atoms were located in difference maps and
refined isotropically, while all other hydrogen atoms were placed in cal-
culated positions.
2.5. Biological studies
Cytotoxicity studies were performed according to previously
established procedure [16]. Briefly, brain endothelial cancerous
mouse cell line (bEnd.3) was maintained in culture flasks with
Dulbecco's Modified Eagle's Medium supplemented with 10% fetal
bovine serum, 100 units/mL penicillin, and 100 μg/mL streptomycin,
under 5% CO2 atmosphere at 37 °C. Into each of the 96 wells,
10,000 cells were seeded and incubated for 24 h. After this point,
compounds, whose stock solutions were prepared in DMSO and
which were diluted into the media, were added to wells. Six concen-
trations were analyzed in triplicate for each compound. As a control,
media without the compounds was added. Following a 72 h incuba-
tion, the MTT reagent (3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-
2H-tetrazolium bromide) was added. After a 3 h incubation period,
the liquid was removed from each well, and the remaining solid
was dissolved in DMSO. Absorbance reading at 595 nm was per-
formed with a plate reader, and the data were fitted using GraphPad
Prism (version 4.00 for Windows, GraphPad Software, San Diego
California USA, www.graphpad.com). The concentrations were
transformed into log[C], and a sigmoidal dose–response with vari-
able Hill Slope fit was run according to Eq. (1), with no restrictions.
Solubility limits of Cu(2)2 and Fe(2)3 resulted in incomplete data
sets for fitting.
Table 1
Selected IR bands and their assignments [19].
Signature IR peaks (cm−1
)
ν NH
ν OH
ν C_O, ring, δas CH
ν CO ν MO
1
3200
1626, 1599, 1579, 1537, 1485 1296
Zn(1)2
Fe(1)3
2
3204, 3537 1608, 1590, 1538, 1506, 1468 1290 760
*, 1588, 1538, 1504, 1468
1622, *, 1555, 1512, 1487
1608, 1588, 1540, 1512, 1464 1292 764
1608, 1594, 1538, 1510, 1468 1292 766
1624, 1576, 1538, 1508, 1483 1302
*
768
*, 3313
3204
3091
1292
Zn(2)2 3360, 3328
Fe(2)3
3
3356, 3236
Zn(3)2
Fe(3)3
1604, 1552, *, 1512, 1468
1590, 1544, *, 1504, 1464
1268 694
1276 662
ðtop−bottomÞ
Symbols denote the following: *, obscured by an adjacent peak, ν, stretching, δas
asymmetric in plane bend.
,
Y ¼ bottom þ
:
ð1Þ
ꢀ
50−xÞ Hill Slope
1 þ 10ð logEC

62
M.A. Telpoukhovskaia et al. / Journal of Inorganic Biochemistry 132 (2014) 59–66
Furthermore, NMR spectroscopy was useful in studying Zn com-
octahedral geometry with six oxygen atoms from the ligands. As can
be seen in Fig. 3, the ligand bridges between the two Zn centers, with
Zn\Zn distance 3.0326(4) Å (Table 3). Zn\O distances range from
2.00 to 2.11 Å, with an exception of an oxygen atom that belongs to a
ligand that is bridging between the two centers, with distance equal
to 2.37 Å (e.g. Zn2\O4, Fig. 3, Table 3).
This provides a fourth example of a Zn-pyridinone complex charac-
terized by X-ray crystallography. Ligand 0 was found in a Zn(0)2 com-
plex with distorted square pyramidal geometry [21] and incorporated
into a mixed hydrotris(3,5-phenylmethylpyrazolyl)borate ligand sys-
tem, showing distorted trigonal bipyramidal geometry [22]. This latter
coordination environment and geometry models catalytic sites in ma-
trix metalloproteinases [22]. As well, HPO ligand 1-(benzo[d]thiazol-
2-ylmethyl)-3-hydroxy-2-methylpyridin-4(1H)-one was found in a
4:2 coordination with Zn(II), with both centers having distorted trigonal
bipyramidal geometry [18]. Interestingly, the 3-hydroxy-2-methyl-
pyran-4-one analog of the Zn(II) complex crystallizes with two metal
centers—one in octahedral and one in square-pyramidal geometry, but
with no bridging ligands, with the rest of the sites occupied by water
molecules [21].
The complex between ligand 1 and Fe(III) revealed two hexadentate
iron centers and twelve non-coordinating solvent molecules in the
asymmetric unit. Both Fe(III) centers have three ligands coordinating
via the oxygen atoms (e.g. Fig. 4), each with slightly distorted octahedral
geometry (Table 3). Other HPO ligands are seen to crystallize as fac
Fe(III) isomers, such as 1-ethyl-3-hydroxy-2-methylpyridin-4-one
[23] and 1-(2′-methoxyethyl)2-methyl-3-hydroxy4-pyridinone [24].
In the present case, the Fe1 center is a fac isomer, while the Fe2 center
is a mer isomer, with the Fe\O distances that range from 1.97 to
2.07 Å (Table 3). The water molecules within the unit cell make a hy-
drogen bonded network (Fig. 5), similar to fac Fe(0)3 crystallized with
12 waters [25].
The dihydrate Fe(1)3 complex has been previously synthesized
and studied by extended X-ray absorption fine structure (EXAFS)
spectroscopy, as the authors predicted that single crystals are “not
likely to be obtained” [26]. By this method, bond lengths were
established to be 1.98 and 2.06 Å for the Fe\O bond, which closely
matches the X-ray analysis (Table 3). Single crystal analysis also con-
firms the distance of carbon and nitrogen atoms in relation to the
metal center. This EXAFS model did not take into account water
molecules.
plexes, as they are diamagnetic; shifts in proton and carbon peaks
were seen upon chelation to 1, 2, and 3. The most drastic shifts were ob-
served for atoms around the pyridinone ring. For instance, as can be
seen in Fig. 2, the biggest shift upon Zn coordination with 1 is downfield
with H5 (marked by symbol +) and upfield with H6 (symbol #), consis-
tent with other NMR studies of coordination of HPOs to diamagnetic
metals [20].
Finally, mass spectrometry confirms 2:1 and 3:1 ligand to metal ra-
tios for Zn(II) and Fe(III) complexes, respectively. These species are con-
firmed by the isotopic pattern in the low resolution spectra (data not
shown) and with the accurate mass measurements by high resolution
spectrometry. Interestingly, low resolution spectra for complexes with
divalent species indicated that there are [M2L3]+ (where M is the
metal and L is the ligand) species present as well, likely an effect in-
duced by the ionization techniques. These species were observed for
Zn(1)2, Zn(2)2, and Zn(3)2 with m/z masses of 728, 775, and 901 for
the tallest peaks in ESI-MS(+). When the same analyses were done
on the analogous copper species, the same pattern was seen, with 3:2
species for Cu(1)2, Cu(2)2, and Cu(3)2 with the tallest peaks at 728,
773, and 897 m/z.
3.2. Solid state structures
Single crystals suitable for X-ray crystallography for three complexes
—Zn(1)2, Fe(1)3, and Cu(3)2 were obtained by slow evaporation of sat-
urated solutions, as detailed in the experimental section, above; crystal
and refinement data is summarized in Table 2.
The complex between Zn(II) and compound 1 crystallized as a
tetramer—with four Zn(II) molecules per eight ligands (Fig. 3),
consistent with the 1:2 stoichiometry seen in HRMS. With each tetra-
mer there are also two coordinating water molecules, two non-
coordinating methanol solvent molecules, and a distorted perchlorate
ion that was modeled using mild SAME, ISOR and DELU restraints. As
the ligand on each end retains its neutral form, the water molecule re-
places the oxygen in coordination to Zn, and the perchlorate ion bal-
ances out the charge. Each tetramer contains two dimers that are
related to each other by inversion symmetry. Interestingly, each dimer
contains two Zn centers with different geometries. The Zn1 center has
distorted square pyramidal geometry with 4 oxygen atoms from ligands
and one from a water molecule, while the Zn2 center has distorted
Fig. 2. 1H NMR spectra comparing 1 to Zn(1)2 (300 MHz, 25 °C, MeOD-d4).

M.A. Telpoukhovskaia et al. / Journal of Inorganic Biochemistry 132 (2014) 59–66
63
Table 2
Crystal and refinement data for Zn(1)2, Fe(1)3, and Cu(3)2.
Crystal data
Zn4(1)8(H2O)2
Fe(1)3(H2O)6
Cu(3)2
(ClO4)2(MeOH)2
Formula
Fw
C100H102N8O30Zn4Cl2
2228.27
C36H42FeN3O6
764.58
C26H18N4O4S2Cu
578.10
Crystal system
Space group
a (Å)
b (Å)
c (Å)
α (deg)
β (deg)
γ (deg)
V [Å3]
Monoclinic
P 21/n
9.4434(4)
32.509(2)
15.6077(7)
90
95.599(2)
90
4768.6(4)
90.0(1)
Monoclinic
P 21/n
12.4671(2)
16.8835(3)
34.9820(5)
90
98.786(1)
90
7276.9(2)
90.0(1)
Monoclinic
P 21/c
10.2041(3)
8.0140(2)
13.8272(4)
90
95.809(2)
90
1124.92(5)
100.0(2)
2
Temp (K)
Z
2
8
D
calcd [g/cm3]
1.550
11.38
1.396
4.81
1.707
12.02
μ (Mo Kα) [cm−1
]
F000
2300
3208.00
590.00
θ
min–θmax (deg)
2.62–27.86
−13/11, −45/43, −21/20
47,359
2.22–26.36
−15/15, −21/20, −43/43
67,610
2.94–29.98
−14/14, −11/11, −19/19
24,693
hkl range
Reflections collected
Independent reflections (Rint
Observed reflection [I N 2 s(I)]
Agreement indexes [I N 2 s(I)]
)
12,687 (0.046)
9123
14,856 (0.039)
11,883
3282 (0.025)
2982
R(F) = 0.049
R(F) = 0.040
Rw(F2) = 0.085
R(F) = 0.056
Rw(F2) = 0.092
1.03
R(F) = 0.027
Rw(F2) = 0.070
R(F) = 0.031
Rw(F2) = 0.072
1.05
R
w(F2) = 0.099
Agreement indexes (all data)
GOF (all data)
R(F) = 0.082
w(F2) = 0.110
1.03
R
As well, a partial solid state structure was obtained for Fe(2)3, and it
revealed a fac isomer with 10 water molecules per metal center. The ge-
ometry is distorted octahedral (Fig. S1), with Fe\O distances ranging
from 1.98 to 2.06 Å. Similar to Fe(1)3, there is a hydrogen bonded
water network within the crystal, which in this case is more extensive,
involving water molecules from multiple unit cells (Fig. S2).
The Cu complex with ligand 3 was found to be in a distorted square
planar geometry with an inversion center at the metal ion (Fig. 6), with
the Cu\O distances 1.93 and 1.95 Å (Table 3). There was disorder at
one benzoxazole group along the N1\C7 bond. This geometry matches
other Cu-pyridinone complexes, which show trans chelation with near-
ly square planar geometries [3,4,27], including the case of Cu(0)2 [28].
3.3. Cytotoxicity
To study the cytotoxicity of these compounds, the bEnd.3 [29] cell
line was chosen, as it is a neuronal cell line, and the brain is the target
organ of these compounds. Cisplatin [30], a cancer drug with
established toxicity, was used as a reference. Fig. 7 illustrates cell viabil-
ity in relation to increasing concentrations of compounds; as can be
Fig. 3. Solid state structure of Zn(1)2, represented as an ellipsoid plot (50% probability) of the asymmetric unit. Hydrogen atoms, solvent molecules, and the perchlorate ion are omitted for
clarity.

64
M.A. Telpoukhovskaia et al. / Journal of Inorganic Biochemistry 132 (2014) 59–66
Table 3
Selected bond lengths (Å) and angles (°) for Zn(1)2, Fe(1)3, and Cu(3)2.
Distance (Å)
Zn(1)\O(1)
Zn(1)\O(3)
Zn(1)\O(4)
Zn(1)\O(6)
Zn(1)\O(9)
Zn(2)\O(4)
Zn(2)\O(5)
Zn(2)\O(6)
Zn(2)\O(8)
Zn(1)\Zn(2)
2.0015(19)
2.0567(18)
2.0507(18)
2.0656(17)
2.0046(19)
2.3671(18)
2.0244(17)
2.0947(17)
2.0917(18)
3.0326(4)
Fe(1)\O(1)
Fe(1)\O(2)
Fe(1)\O(3)
Fe(1)\O(4)
Fe(1)\O(5)
Fe(1)\O(6)
Fe(2)\O(7)
Fe(2)\O(8)
Fe(2)\O(9)
Fe(2)\O(10)
Fe(2)\O(11)
Fe(2)\O(12)
2.0726(14)
1.9867(15)
2.0278(14)
2.0068(14)
2.0262(15)
2.0152(14)
2.0414(15)
1.9739(14)
2.0308(14)
2.0421(14)
2.0487(14)
1.9800(15)
Cu(1)\O(1)
Cu(1)\O(2)
1.9456(10)
1.9270(10)
Angle (°)
O(1)\Zn(1)\O(9)
O(1)\Zn(1)\O(4)
O(9)\Zn(1)\O(4)
O(4)\Zn(1)\O(6)
O(8)\Zn(2)\O(4)
O(6)\Zn(2)\O(4)
108.89(8)
98.32(7)
152.15(8)
82.60(7)
157.92(6)
74.74(7)
O(2)\Fe(1)\O(4)
O(2)\Fe(1)\O(5)
O(4)\Fe(1)\O(5)
O(6)\Fe(1)\O(5)
90.74(6)
165.91(6)
102.38(6)
80.12(6)
O(2)\Cu(1)\O(1)
O(1)\Cu(1)\O(1)#1
O(2)#1\Cu(1)\O(1)
85.96(4)
180.00(6)
94.04(4)
seen, ligands 1, 2, and 3 have a cytotoxic effect comparative to that of
cisplatin across the concentration range. Fig. S3 illustrates cytotoxicity
profiles for the three ligands with respect to log of concentration.
With the normalized data sets, it is clear that the EC50 values (concen-
tration at which half of the maximal effect is observed) for 1, 2, and 3
are lower than that for cisplatin, with the numerical data summarized
in Table S1. The EC50 values for pro-ligands 1 and 2 were previously
established to be 17 [3] and 92 [4] μM, respectively, in a human liver
cancer cell line (HepG2), compared to an average value of 12 μM for cis-
platin. While in this cell line, the pro-ligands show a slight increase in
toxicity compared to cisplatin, the trend of increased lipophilicity lead-
ing to increased toxicity holds true.
Fig. 7 also demonstrates the effect of complexation on toxicity. In the
cases of the copper and the iron complexes, the bases of the cytotoxicity
curves were not obtained with the concentration ranges achievable
(Fig. S4), thus their EC50 values are not determined with curve fitting,
but are presumed to be larger than the highest concentration in the
assay (Table S2). It should be noted that the concentrations in Figs. 7,
S3, and S4 are listed according to the complex, and each Cu(II) and
Zn(II) complex contains two ligands, while each Fe(III) complex con-
tains three of them. When the cell viability data are graphed according
to the concentration of 2 (Fig. S5), it is apparent that while the estimated
EC50 values (Table S2) vary between 2, Cu(2)2, and Zn(2)2, the 50% cell
viability is reached within a very narrow concentration range. On the
other hand, the iron complex is less toxic to the cells.
In the literature, metal complexes have been shown to either increase,
decrease, or not have a significant effect on cytotoxicity when compared
to the unchelated ligand. For example, in cell lines K-562 (leukemia),
MCF-7 (breast cancer), G-361 (melanoma), and HOS (osteogenic sarco-
ma) Cu(II) and Fe(III) complexes of bidentate N6-benzylaminopurine de-
rivatives were found to be more toxic than the pro-ligands [31], while
Zn(II) and Fe(III) monodentate 6-benzylamino-9-isopropylpurine deriv-
atives had EC50 values on the same order of magnitude as the pro-ligands
[32,33]. In other cases, copper chelation decreased EC50 by about an order
Fig. 4. Solid state structure of one Fe center (Fe1) of complex Fe(1)3, represented as an
ellipsoid plot (50% probability). Hydrogen atoms and solvent molecules are omitted for
clarity.
Fig. 5. Hydrogen bonded water network within the Fe(1)3 unit cell. Hydrogen atoms are
omitted for clarity.

M.A. Telpoukhovskaia et al. / Journal of Inorganic Biochemistry 132 (2014) 59–66
65
Fig. 6. Solid state structure of Cu(3)2, represented as an ellipsoid plot (50% probability). Hydrogen atoms are omitted for clarity.
of magnitude of monodentate 4-nitropyridine N-oxides in several cell
Acknowledgments
lines [34], while quinone thiosemicarbazone derivatives displayed higher
EC50 for Cu(II) complexes over the pro-ligands [35] and quercetin was
found to have a similar toxicity compared to its Zn complex, with the
EC50 for the complex lower in some cell lines [36]. This survey illustrates
that metal chelation may or may not have an effect on cytotoxicity, and it
is important to study each case individually.
It is important to note that the speciation of the metal complexes
may change upon introduction of the cell media, which contain amino
acids and albumin. Most literature reports cited herein do not comment
on stability of the complexes in the media. Present at higher concentra-
tions, ligands in the media are likely to compete for metals within com-
plexes, and thus complex concentration may not be accurately
estimated. In our case, especially the Cu(II) and Zn(II) complexes are
likely to be significantly perturbed by the species in the media. This
may explain why some results, including ours, show that cytotoxicity
is on the same order of magnitude upon chelation to certain metal
ions. This should be taken into consideration when interpreting cyto-
toxicity results of complexes, especially of those complexes that have
low denticity and moderate binding constants.
We are grateful to the Alzheimer Society of Canada for a doctoral
award (M.A.T.), the Natural Sciences and Engineering Research Council
of Canada (NSERC) for a Discovery Grant (C.O.), a PGSD-2 award
(M.A.T.) and a summer USRA scholarship (B.D.G.P.), and the Canadian
Institutes of Health Research (CIHR) Proof of Principle program (C.O.).
C.O. acknowledges the Canada Council for the Arts for a Killam Research
Fellowship (2011–2013) and the Alexander von Humboldt Foundation
for a Research Award, as well as Prof. P. Comba and his group in Heidel-
berg for their hospitality and interesting discussions. Additionally, we
thank Agència de Gestió d'Ajuts Universitaris i de Recerca (AGAUR)
from Generalitat de Catalunya (Beatriu de Pinós (BP-DGR-2009)) for a
postdoctoral fellowship funding (C.R.R.), and the Province of British Co-
lumbia for a Pacific Century Scholarship (L.E.S.). Lastly, we thank
Dr. Kelly McNagny for providing the bEnd.3 cell line, Jessie Chen from
the UBC Bioservices Laboratory for the help with cell studies, and the
aforementioned staff of the UBC Chemistry analytical services for their
help with characterization of complexes.
Appendix A. Supplementary Data
4. Conclusions
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jinorgbio.2013.12.003.
HPO ligands 1, 2, and 3 that are considered for application in
Alzheimer's disease [3,4] were chelated to Zn(II) and Fe(III) to complete
characterization with metals implicated in AD pathology [5]. The com-
plexes were studied by IR and NMR spectroscopies, MS, and EA to estab-
lish the stoichiometry of chelation to be 2 ligands to one metal for the
zinc complexes and 3 to 1 for the iron ones. Additionally, Zn(1)2,
Fe(1)3, and Cu(3)2 were characterized by X-ray crystallography. Cyto-
toxicity profiles of pro-ligands were established in a neuronal cell line
and the effect of chelation to metal ions was studied.
References
[1] K.H. Thompson, C. Barta, C. Orvig, Chem. Soc. Rev. 35 (2006) 545.
[2] H. Schugar, D.E. Green, M.L. Bowen, L.E. Scott, T. Storr, K. Böhmerle, F. Thomas, D.D.
Allen, P.R. Lockman, M. Merkel, K.H. Thompson, C. Orvig, Angew. Chem. Int. Ed. 119
(2007) 1746.
[3] L.E. Scott, B.D.G. Page, B.O. Patrick, C. Orvig, Dalton Trans. (2008) 6364.
[4] L.E. Scott, M. Telpoukhovskaia, C. Rodríguez-Rodríguez, M. Merkel, M.L. Bowen,
B.D.G. Page, D.E. Green, T. Storr, F. Thomas, D.D. Allen, P.R. Lockman, B.O. Patrick,
M.J. Adam, C. Orvig, Chem. Sci. 2 (2011) 642.
[5] A.I. Bush, R.E. Tanzi, Neurotherapeutics 5 (2008) 421.
[6] A.K. Sharma, S.T. Pavlova, J. Kim, J. Kim, L.M. Mirica, Metallomics 5 (2013) 1529.
[7] X. Huang, R.D. Moir, R.E. Tanzi, A.I. Bush, J.T. Rogers, Ann. N. Y. Acad. Sci. 1012 (2004)
153.
[8] I. Singh, A.P. Sagare, M. Coma, D. Perlmutter, R. Gelein, R.D. Bell, R.J. Deane, E. Zhong,
M. Parisi, J. Ciszewski, R.T. Kasper, R. Deane, Proc. Natl. Acad. Sci. U. S. A. 110 (2013)
14771.
[9] E.P. Raven, P.H. Lu, T. Tishler, P. Heydari, G. Bartzokis, J. Alzheimers Dis. 37 (2013)
127.
[10] A. Agrawal, S.L. Johnson, J.A. Jacobsen, M.T. Miller, L.-H. Chen, M. Pellecchia, S.M.
Cohen, ChemMedChem 5 (2010) 195.
[11] A. Gojmerac Ivšić, V. Tomišić, Ž. Car, B. Prugovečki, S. Tomić, J. Mol. Struct. 990
(2011) 237.
[12] P.J. Crouch, M.S. Savva, L.W. Hung, P.S. Donnelly, A.I. Mot, S.J. Parker, M. Greenough,
I. Volitakis, P. Adlard, R. Cherny, C.L. Masters, A.I. Bush, K.J. Barnham, A.R. White,
J. Neurochem. 119 (2011) 220.
[13] SAINT, Version 7.68A, Bruker AXS Inc., Madison, Wisconsin, USA, 1997–2010.
[14] G.M. Sheldrick, Acta Crystallogr. Sect. A: Found. Crystallogr. 64 (2008) 112.
[15] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, C. Giacovazzo, A. Guagliardi, A.G.
Moliterni, G. Polidori, R. Spagna, J. Appl. Crystallogr. 32 (1999) 115.
[16] M.A. Telpoukhovskaia, B.O. Patrick, C. Rodríguez-Rodríguez, C. Orvig, Mol. Biosyst. 9
(2013) 792.
[17] M.A. Telpoukhovskaia, C. Orvig, Chem. Soc. Rev. 42 (2013) 1836.
[18] M.A. Telpoukhovskaia, C. Rodríguez-Rodríguez, J.F. Cawthray, L.E. Scott, B.D.G. Page,
J. Alí-Torres, M. Sodupe, G.A. Bailey, B.O. Patrick, C. Orvig, Metallomics (2014),
http://dx.doi.org/10.1039/c3mt00135k.
[19] W.O. Nelson, T.B. Karpishin, S.J. Rettig, C. Orvig, Can. J. Chem. 66 (1988) 123.
[20] D.E. Green, C.L. Ferreira, R.V. Stick, B.O. Patrick, M.J. Adam, C. Orvig, Bioconjug. Chem.
16 (2005) 1597.
Fig. 7. BEnd.3 cell line viability in the presence of compounds. Values are an average of
three trials with standard relative deviation. Individual data for each compound are
graphed in Figs. S3 and S4, and EC50 values are tabulated in Tables S1 and S2.

66
M.A. Telpoukhovskaia et al. / Journal of Inorganic Biochemistry 132 (2014) 59–66
[21] S.I. Ahmed, J. Burgess, J. Fawcett, S.A. Parsons, D.R. Russell, S.H. Laurie, Polyhedron 19
(2000) 129.
[29] Y. Omidi, L. Campbell, J. Barar, D. Connell, S. Akhtar, M. Gumbleton, Brain Res. 990
(2003) 95.
[22] D.T. Puerta, S.M. Cohen, Inorg. Chem. 42 (2003) 3423.
[23] G. Xiao, D. van der Helm, R.C. Hider, P.S. Dobbin, Dalton Trans. (1992) 3265.
[24] P.S. Dobbin, R.C. Hider, A.D. Hall, P.D. Taylor, P. Sarpong, J.B. Porter, G. Xiao, D. van
der Helm, J. Med. Chem. 36 (1993) 2448.
[30] National Cancer Institute, Cisplatin, 2013. (accessed 14 July 2013, http://www.
cancer.gov/cancertopics/druginfo/cisplatin).
[31] M. Maloň, Z. Trávnı
́ček, M. Maryško, R. Zbořil, M. Mašláň, J. Marek, K. Doležal, J.
Rolčık, V. Kryštof, M. Strnad, Inorg. Chim. Acta 323 (2001) 119.
́
[25] E.T. Clarke, A.E. Martell, J. Reibenspies, Inorg. Chim. Acta 196 (1992) 177.
[26] W. Schlindwein, E. Waltham, J. Burgess, N. Binsted, A. Nunes, A. Leite, M. Rangel,
Dalton Trans. (2006) 1313.
[32] Z. Trávnícek, V. Krystof, M. Sipl, J. Inorg. Biochem. 100 (2006) 214.
[33] Z. Trávnícek, I. Popa, M. Cajan, R. Zboril, V. Krystof, J. Mikulík, J. Inorg. Biochem. 104
(2010) 405.
[27] D.E. Green, M.L. Bowen, L.E. Scott, T. Storr, M. Merkel, K. Böhmerle, K.H.
Thompson, B.O. Patrick, H.J. Schugar, C. Orvig, Dalton Trans. 39 (2010)
1604.
[28] A. El-Jammal, P.L. Howell, M.A. Turner, N. Li, D.M. Templeton, J. Med. Chem. 37
(1994) 461.
[34] A. Puszko, A. Brzuszkiewicz, J. Jezierska, A. Adach, J. Wietrzyk, B. Filip, M. Pełczynska,
M. Cieslak-Golonka, J. Inorg. Biochem. 105 (2011) 1109.
[35] Z. Afrasiabi, P. Stovall, K. Finley, A. Choudhury, C. Barnes, A. Ahmad, F. Sarkar, A.
Vyas, S. Padhye, Spectrochim. Acta A 114C (2013) 114.
[36] J. Tan, B. Wang, L. Zhu, Bioorg. Med. Chem. Lett. 17 (2009) 614.