Exploring the therapeutic potential of Cu(II)-complexes with ligands derived from pyridoxal

Article abstract of DOI:10.1016/j.ica.2020.119558

Three new copper(II) complexes formulated as [Cu(L)(X)], where X = H₂O or Cl and H₂L is a Schiff base (H₂L^1,2) or its reduced version (H₃L³Cl) derived from pyridoxal, are prepared, as well

Full text of DOI:10.1016/j.ica.2020.119558

Journal Pre-proofs
Research paper
Exploring the therapeutic potential of Cu(II)-complexes with ligands derived
from pyridoxal
Patrique Nunes, Fernanda Marques, Isabel Cavaco, João Costa Pessoa, Isabel
Correia
PII:
DOI:
Reference:
S0020-1693(19)31677-9
https://doi.org/10.1016/j.ica.2020.119558
ICA 119558
To appear in:
Inorganica Chimica Acta
Received Date:
Revised Date:
Accepted Date:
4 November 2019
22 January 2020
26 February 2020
Please cite this article as: P. Nunes, F. Marques, I. Cavaco, J. Costa Pessoa, I. Correia, Exploring the therapeutic
potential of Cu(II)-complexes with ligands derived from pyridoxal, Inorganica Chimica Acta (2020), doi: https://
doi.org/10.1016/j.ica.2020.119558
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier B.V.

Exploring the therapeutic potential of Cu(II)-complexes with ligands
derived from pyridoxal
Patrique Nunes,a Fernanda Marques,b Isabel Cavaco,a,c João Costa Pessoa,a Isabel
Correiaa,*
a Centro de Química Estrutural, Departamento de Engenharia Química, Instituto
Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisboa,
Portugal.
b Centro de Ciências e Tecnologias Nucleares, Instituto Superior Técnico,
Universidade de Lisboa, Estrada Nacional 10, 2695-066 Bobadela LRS, Portugal.
cDepartamento de Química Bioquímica e Farmácia, Universidade do Algarve,
Campus de Gambelas, 8005-139 Faro, Portugal
*To whom correspondence should be addressed:
Isabel Correia
e-mail: icorreia@tecnico.ulisboa.pt
Tel: +351 218 419 270
Fax: +351 218 419 239
1

Abstract
Three new copper(II) complexes formulated as [Cu(L)(X)], where X= H2O or Cl and
H2L is a Schiff base (H2L1,2) or its reduced version (H3L3Cl) derived from pyridoxal,
are prepared, as well as two ternary complexes [Cu(L)(phen)] also containing 1,10-
phenanthroline. All compounds are characterized by the usual techniques: elemental
analyses, ESI mass spectrometry, UV-Vis absorption, FTIR and EPR
spectroscopies. The ligands coordinate the Cu(II) center forming complexes with
square-planar based geometries. Their antioxidant properties are evaluated with a
radical scavenging activity assay, with one of the ligand precursors showing activity
higher than the positive control, ascorbic acid. The antiproliferative activity of all
compounds is evaluated against two cancer cell lines: ovarian (A2780) and breast
(MCF7). All complexes show moderate to excellent activity with the ternary Cu-
complexes showing IC50 values between 0.7-9.3 M after 24 h of incubation, values
much lower than those reported for cisplatin, the reference drug. The hydrolytic
stability of the complexes and their ability to bind albumin and DNA are evaluated by
spectroscopic techniques, showing that the compounds bind bovine serum albumin.
The [Cu(L)(phen)] complexes show ability to target DNA via intercalation.
Keywords: Pyridoxal, Schiff bases, copper, cytotoxicity, albumin
2

1. Introduction
Cancer is a health problem with very high impact in modern society and cancer
deaths are estimated to rise as life expectancy increases; therefore, the interest in
developing new anticancer metallodrugs with higher efficacy and presenting less
adverse side effects is a growing research area.
Vitamin B6 comprises a group of compounds that are involved in a multiplicity of
biochemical reactions. Most of these reactions are co-catalyzed by the B6 vitamer,
pyridoxal-5’-phosphate, making it a crucial and versatile co-factor in numerous cell-
metabolic processes [1, 2]. Additionally, vitamin B6 is an antioxidant and incubation
of cancer cells with high doses of pyridoxal or pyridoxine yielded significant reduction
in cell proliferation when compared to control cultures [3]. Interest in metal
coordination of B6-vitamers started with the discovery of non-enzymatic model-
systems containing pyridoxal and metal ions that could mimic enzymatic reactions
involving vitamin B6, by formation of Schiff base (SB) complexes [1]. Our group has
worked extensively on the speciation and therapeutic potential of metal complexes
derived from pyridoxal Schiff bases and reduced Schiff bases [4-8]. The V(IV) and
Zn(II) complexes of these ligands exhibited antidiabetic activity [4, 9]. Also, one of
the V(IV)-SB complexes previously reported by us [5], showed selective cytotoxicity
for human melanoma and lung carcinoma cells, with a mechanism involving
apoptosis triggered by increase in reactive oxygen species (ROS) formation [10].
Pyridoxal isonicotinoyl hydrazone derivatives and its iron complexes have shown
anti-neoplastic activity and/or protection against oxidative stress [11-13]. Pyridoxal
thiosemicarbazone Cu(II) complexes showed suppressive effect on erythroleukemia
proliferation [14] and Co(III) complexes exhibited significant activity against breast
(MCF7) and cervix (HeLa) cancer cells [15]. Therefore, exploration of pyridoxal-
containing complexes has high potential to successfully develop effective
chemotherapeutics.
Copper complexes are gaining importance as alternative anticancer drugs,
anticipating that their endogenous character may induce less systemic toxic side
effects [16-18]. Additionally, the use of aromatic N-heterocycles as co-ligands in the
metal ion coordination could allow targeting DNA, as this type of molecules have
shown high ability to bind to DNA double strands [19-23]. Inspired by these findings,
3

copper(II) complexes were prepared containing ligands derived from vitamin B6
vitamers, as well as ternary complexes also containing 1,10-phenanthroline. In this
report, we describe the synthesis, characterization and biological assessment of
these new Cu(II) complexes. The compounds were characterized in solution and in
the solid state and their stability and ability to bind albumin and DNA were evaluated.
Their cytotoxicity was determined in two human tumoral cell lines and the influence
of different structural elements on the biological performance of the compounds is
discussed.
2. Experimental
2.1. Materials and methods
Pyridoxamine dihydrochloride and pyridoxal hydrochloride were purchased from
Merck, 2-aminophenol from Reidel-de-Haen, salicylaldeyde from Acros, sodium
borohydride from Alfa Aesar and DPPH from Sigma-Aldrich. All chemicals were p.a.
grade and used without further purification, unless otherwise stated. The 1H and 13C
NMR spectra were obtained on Bruker Avance+ 400 MHz and 300 MHz
spectrometers. 1H chemical shifts (δ) are expressed in ppm relative to the deuterated
solvent residual peak. Electronic absorption spectra (UV-Vis) were recorded with a
Perkin Elmer Lambda 35 spectrophotometer. Circular Dichroism (CD) spectra were
recorded at 25 °C on a Jasco J-720 spectropolarimeter with an UV–Vis (180–800
nm) photomultiplier (EXEL-308). Infrared spectra (IR, 4000–400 cm–1) were recorded
either on a Nicolet Impact 400D or a BIO-RAD FTS 3000 MX spectrophotometers in
KBr pellets; wavenumbers are expressed in cm–1. Elemental analyses were carried
out at Laboratório de Análises of Instituto Superior Técnico, using an FISONS EA
1108 CHNS-O apparatus. A LCQ Fleet™ Ion Trap Mass Spectrometer from Thermo
Scientific was used to measure ESI-MS spectra of methanolic solutions of the
complexes for characterization, and of aqueous solutions (3% DMSO-NH4HCO3
aqueous buffer, pH 7.4) for stability evaluation. A DIONEX ICS-1500 ion
chromatography system, equipped with an IonPac® AS14A 4-mm analytical (4×250
mm) column and a suppressed conductivity at 10 μSFS ASRS®-ULTRA
AutoSuppression® recycle mode detection system was used for chloride
determination. Fluorescence measurements were carried out on a SPEX® Fluorolog
4

spectrofluorometer (Horiba Jobin Yvon) in a FL3-11 configuration, equipped with a
Xenon lamp. The instrumental response was corrected by means of a correction
function provided by the manufacturer. The experiments were carried out at room
temperature and are all steady-state measurements.
2.2. Synthesis
4-(((2-hydroxybenzylidene)amino)methyl)-5-(hydroxymethyl)-2-methylpyridin-3-ol
(H2L1)
H2L1 was prepared by adding a solution of pyridoxamine dihydrochloride (0.482 g,
2.00 mmol) to 10 mL of methanol with KOH at pH ca. 9.0. After 30 minutes the white
solid (KCl) was filtered off and salicylaldehyde (0.213 mL, 2.00 mmol) dissolved in
6.5 mL of methanol:CHCl3 (8% v/v) was added to the solution. The reaction mixture
was refluxed at 50 ºC for 3 h, under stirring. After cooling the pH was adjusted to ca.
7.0 with methanolic HCl and the reaction volume was reduced under vacuum till
precipitation occurred. The mixture was kept at 4 ºC overnight and the yellow solid
that formed was filtered, washed with water, cold methanol and diethyl ether and
dried under vacuum. Yield: 91%, 0.412 g. MM (C15H16N2O3) = 272.3 g/mol.
Elemental Analysis: Calc. for C15H16N2O3.0.5H2O: C, 64.04; H, 6.09; N, 9.96; Found:
C, 63.8; H, 6.1; N, 9.9. ESI-MS: m/z [Found (Calcd)]: 271.9 (271.3) (25%) [L-H]–;
543.2 (543.6) (100%) [2L-H]–; 273.13 (273.3) (100%) [L+H]+. UV-Vis in DMSO,
1
λmax/nm (ε/M-1cm-1): 285 (7965), 315 (3782), 413 (183); H NMR (dmso-d6, δ/ppm):
2.38 (s, 3H, H19); 4.55 (s, 2H, H16); 4.88 (s, 2H, H9); 6.82-6.89 (m, 2H, H2 and H6);
7.30 (t, 1H, H1); 7.43 (d, 1H, H3); 7.96 (s, 1H, H12); 8.61 (s, 1H, H7); 5.20 (s, 1H, R-
OH); 9.05 (s, 1H, Ar-OH) and 13.30 (s, 1H, Ar-OH). 13C NMR (dmso-d6, δ/ppm):
19.87 (C19); 52.33 (C9); 58.95 (C16); 116.53 (C6); 118.53 (C2); 118.76 (C4); 130.20
(C11); 131.54 (C3); 132.40 (C1); 134.11 (C10); 139.68 (C12); 146.23 (C14); 149.46
(C15); 160.51 (C5) and 166.14 (C7). LogP = 1.29 [24].
5-(hydroxymethyl)-4-(((2-hydroxyphenyl)imino)methyl)-2-methylpyridin-3-ol (H2L2)
H2L2 was prepared by stirring 2-aminophenol (0.218 g, 2.00 mmol) dissolved in 6 mL
of methanol/KOH pH ca. 9.0 and pyridoxal hydrochloride (0.407 g, 2.00 mmol) in 8
mL methanol, at room temperature for 3 h. After cooling the mixture was kept at 4 ºC
overnight and the orange solid was filtered, washed with water, cold methanol and
diethyl ether and dried under vacuum. Yield: 85%, 0.390 g. MM (C14H14N2O3) =
5

258.3 g/mol. Elemental Analysis: Calc. for C14H14N2O3: C, 65.11; H, 5.46; N, 10.85;
Found: C, 64.9; H, 5.7; N, 10.7. ESI-MS: m/z [Found (Calcd)]: 259.15 (259.3) (100%)
[L+H]+; 257.11 (257.29) (100%) [L-H]–. UV-Vis in dmso, λmax/nm (ε/M-1cm-1): 295
(8282), 360 (11601), 463 (421); 1H NMR (dmso-d6, δ/ppm): 2.55 (s, 3H, H17); 4.89 (s,
2H, H18); 6.97 (t, 1H, H2); 7.07 (d, 1H, H6); 7.28 (t, 1H, H1); 7.69 (d, 1H, H3); 7.98 (s,
1H, H12); 9.30 (s, 1H, H9); 5.63 (s, 1H, R-OH) and 10.45 (s, 1H, Ar-OH). 13C NMR
(dmso-d6, δ/ppm): 16.15 (C17); 57.83 (C18); 116.82 (C6); 119.64 (C2); 119.73 (C3);
121.82 (C11); 129.43 (C12); 130.45 (C1); 131.30 (C4); 136.82 (C10); 147.13 (C14);
151.98 (C15); 156.28 (C9) and 157.82 (C5). LogP = 2.21 [24]
5-(hydroxymethyl)-4-(((2-hydroxyphenyl)amino)methyl)-2-methylpyridin-3-ol (H3L3Cl)
H3L3Cl was obtained by reducing H2L2 (0.250 g, 0.970 mmol) with an excess of
NaBH4 in the presence of KOH in MeOH at room temperature for 48 h. A beige
colored solid was obtained, washed with water and diethyl ether and dried under
vacuum. Yield: 40%, 0.100 g. MM (C14H17Cl2NaN2O3) = 355.1 g/mol. Elemental
Analysis: Calc. for C14H17N2O3NaCl2.2.5H2O: C, 42.02; H, 5.54; N, 7.00; Cl, 17.72.
Found: C, 42.2; H, 5.4; N, 6.8, Cl, 18.0. ESI MS: m/z [Found (Calcd)]: 259.61
(259.11) (10%) [L-H]–; 295.17 (295.1) (100%) [L+Cl]–; 297.02 (297.07) (25%) [L+K-
1
2H]–. UV Vis in dmso, λmax/nm (ε/M-1cm-1): 285 (744), 342 (159); H NMR (D2O,
δ/ppm): 2.67 (s, 3H, H17); 4.64 (s, 2H, H18); 4.79 (s, 2H, H9); 6.89 (t, 1H, H2); 6.97 (d,
1H, H6); 7.23-7.29 (m, 2H, H1 and H3) and 8.17 (s, 1H, H12). 13C NMR (D2O, δ/ppm):
15.4 (C17); 45.03 (C9); 58.49 (C18); 116.86 (C6); 120.74 (C2); 121.42 (C4); 123.89
(C3); 130.95 (C12); 131.55 (C1); 134.06 (C10); 138.90 (C11); 142.46 (C14); 149.62 (C5)
and 153.63 (C15). LogP = 1.10 [24]
General synthesis of the [Cu(L)(X)] complexes (X = H2O or Cl)
Copper(II) complexes 1-3 were obtained by drop-wise addition of 1 mL of CuCl2
(1.00 mmol) in methanol to a stirred solution of 7-10 mL of a methanolic solution of
the appropriate ligand precursor (1.00 mmol), followed by addition of a few drops of
KOH in methanol to adjust the pH at ca. 7.5-8.0. The mixture was stirred at room
temperature for ~4 h and kept in the fridge overnight. The solids formed were
separated by filtration, washed with water, a small amount of cold methanol or
acetone and diethyl ether, and dried under vacuum.
6

K[Cu(L1)Cl] (1) - A dark green solid was obtained. Yield: 24%, 0.042 g. Elemental
Analysis: Calc. [C15H14N2O3ClCu]K•0.5H2O: C, 43.16; H, 3.62; N, 6.71; Cl, 8.49;
Found: C, 43.1; H, 3.4; N, 6.5; Cl, 8.9. ESI-MS: m/z [Found (Calcd)]: 368.04 (368.00)
(20%) [M+Cl]–; 378.2 (378.0) (57%) [M+FA-H]–; considering M the neutral complex
without chloride and potassium ions; FA is formic acid. UV-Vis in dmso, λmax/nm
(ε/M-1cm-1): 266 (11447), 320 (4718), 370 (3498) and 670 (113). EPR parameters
(DMSO, 3mM): g = 2.071, g|| = 2.280, A = 16.5×10-4 cm-1 and A|| = 180.7×10-4 cm-1;
(MeOH, 3mM): g = 2.058, g|| = 2.278, A = 16.8×10-4 cm-1 and A|| = 176.9×10-4 cm-1
[Cu(L2)(H2O)] (2) - A dark red solid was obtained. Yield: 66%, 0.210 g. MM
(C14H14N2O4Cu) = 337.8 g/mol. Elemental Analysis: Calc. for C14H14CuN2O4: C,
49.78; H, 4.18; N, 8.29; Found: C, 49.5; H, 3.7; N, 8.1. ESI-MS: m/z [Found (Calcd)]:
179.74 (178.48) (45%) [M+H+K]2+, 397.77 (398.09) (10%) [M+H2O+IsoProp+H]+,
364.38 (364.0) (100%) [M+FA-H]–, 683.35 (683.0) (60%) [2M+FA-H]–, considering M
the neutral complex without H2O. UV-Vis in DMSO, λmax/nm (ε/M-1 cm-1): 281
(10251), 420 (9391, sh) and 457 (10818).
Na[Cu(L3)Cl] (3) - A dark red solid was obtained. Yield: 57%, 0.062 g. Elemental
Analysis: Calc. for [C14H14N2O3ClCu]Na•0.5H2O: C, 43.21; H, 3.88; N, 7.20; Cl, 9.11;
Found: C, 43.5; H, 3.4; N, 7.0; Cl, 9.5. ESI-MS [Found (Calcd)]: 356.54 (356.00)
(90%) [M+Cl]–, 366.41 (366.03) (90%) [M+FA-H]–, 397.82 (397.95) (100%) [M+2K-
H]+, considering M the neutral complex without the chloride and sodium ions. UV-Vis
in DMSO, λmax/nm (ε/M-1cm-1): 286 (11540), 422 (8804), 454 (10282).
General synthesis of the complexes [Cu(L)(phen)]
Copper(II) complexes 4 and 5 were prepared by drop-wise addition of 1 mL of CuCl2
(1.00 mmol) in methanol to a stirred solution of 7-10 mL of a methanolic solution of
the appropriate ligand precursor (1.00 mmol), followed by the addition of few drops
of KOH in methanol to adjust the pH at ca. 7.5-8.0. The mixture was stirred at room
temperature for 15 min and 1,10-phenanthroline (1.00 mmol) was added drop wise.
The mixture was stirred for another 4 h and kept in the fridge overnight. The resulting
solid was filtrated and washed with water, cold ethanol or methanol and diethyl ether,
dried under vacuum and recrystallized from methanol, when necessary.
[Cu(L1)(phen)] (4) - A green solid was obtained. Yield: 66%, 0.210 g. MM
(C27H22CuN4O3) = 514.07 g/mol. Elemental Analysis: Calc. [C27H22CuN4O3]•0.5H2O:
7

C, 62.00; H, 4.43; N, 10.71; Found: C, 62.1; H, 4.2; N, 10.7. ESI-MS: m/z [Found
(Calcd)]: 514.02 (514.1) (100%) [M+H]+, 178.72 (179.37) (23%) [M+2H+Na]3+. UV-
Vis in dmso, λmax/nm (ε/M-1cm-1): 275 (49367), 293 (21402), 371 (4833), 450 (744)
and 672 (159). EPR parameters (MeOH, 3mM): g = 2.080, g|| = 2.268, A = 36.9×10-4
cm-1 and A|| = 167.9×10-4 cm-1
[Cu(HL2)(phen)]Cl (5) - An orange solid was obtained. Yield: 55 %, 0.136 g; MM
(C26H20N4O3Cu)
=
500.02
g/mol.
Elemental
Analysis:
Calc.
for
[C26H21N4O3ClCu]•1.5H2O: C, 55.42; H, 4.29; N, 9.94; Found: C, 55.8; H, 4.2; N, 9.8.
ESI-MS: m/z [Found (Calcd)]: 500.08 (499.97) (75%) [M+H]+; 999.16 (1000.4) (28%)
[2M+H]+; 498.07 (497.27) (45%) [M-H]–, considering M the neutral complex without
the Cl- ion and not protonated. UV-Vis in dmso, λmax/nm (ε/M-1cm-1): 295 (17357),
400 (6811) and 467 (6620). EPR parameters (DMF, 3mM): g = 2.058, g|| = 2.247, A
= 32.0×10-4 cm-1 and A|| = 181.0×10-4 cm-1; (MeOH, 3mM): g = 2.077, g|| = 2.278, A =
46.6×10-4 cm-1 and A|| = 160.8×10-4 cm-1
2.3. Chloride determination by anion exchange chromatography
The compounds were stirred overnight at room temperature in an aqueous solution
containing 0.1 - 1.0 M of H2SO4, with a minimum amount of DMSO (used to dissolve
the complexes), to obtain a final concentration of ca. 4 mM. The resulting solutions
were diluted with Milli-Q water in order to obtain Cl− concentrations within the
calibration curve. The calibration curve was determined by preparing a concentrated
stock solution of 500 ppm of Cl− ion in Milli-Q water. This stock solution was diluted
to obtain five solutions of Cl− with concentrations ranging from 5.0 to 50 ppm. The
solutions of the samples were injected and the Cl− concentration determined.
2.4. Evaluation of the radical scavenging activity
Stock solutions (2.0 mM) of the compounds were diluted to obtain concentrations of
9.0, 19, 28, 37 and 74 M in methanol. To each solution 1.25 mL of DPPH (2,2-
diphenyl-1-picrylhydrazyl, 50 M) in methanol was added and the final volume (3.5
mL) was obtained by adding methanol. The solutions were kept in the dark for 30
min. The radical scavenging activity (RSA) [25] was evaluated by following the
decrease of DPPH absorbance at 517 nm. The data were collected between 300-
8

700 nm, and the values were converted into percentage of RSA, according to the
following formula:
(1)
where Abssample is the absorption of DPPH solutions containing the test compound
and Absstandart is the absorption of DPPH in methanol. All ligands and complexes
have negligible absorbance under the applied concentrations at 517 nm (thus, no
interference between absorption of the tested compounds and DPPH). Ascorbic acid
was chosen as the positive control.
2.5. Cytotoxicity assays
The cytotoxicity of the complexes was evaluated by the MTT assay (MTT = 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) against A2780 ovarian and
MCF7 breast cancer cells in 96-well plates (2-3x106 cells/well). The compounds
were first dissolved in DMSO to prepare a 10 mM stock solution and then in medium
through serial dilutions in the range 0.1-100 μM. The final DMSO concentration was
1% (v/v) for the highest complex concentration; this amount showed no cytotoxic
effect. After treatment with the complexes, the cell medium was replaced by 200 µL
of a MTT solution in PBS (0.5 mg/mL) and further incubated for 24 h at 37 °C. The
resulting purple formazan crystals from the tetrazolium reduction were dissolved in
200 µL DMSO. The absorbance of controls and cell-treated wells was measured at
575 nm using a microplate reader (PowerWave Xs, Biotek Instruments, USA). The
cytotoxicity of the complexes was expressed as the IC50 values calculated from
dose-response curves using the GraphPad Prism software (vs. 5.0).
2.6 Stability studies
2.6.1. UV-Vis
Stability studies were performed in organic and aqueous medium by measuring
absorption spectra with increasing time. Stock solutions of complexes and ligands
were prepared in DMSO with concentrations ranging from 0.5 to 1.0 mM. Sample
solutions were prepared by dilution of the stock solutions in DMSO, NH4HCO3 (25
mM, pH~7.4) or PBS (10 mM, pH = 7.4, 2.7mM KCl and 0.137 M NaCl), to obtain
9

final concentrations between 23 and 143 M, with <5% DMSO (v/v) in the aqueous
solution samples.
2.6.2. 1H NMR
The stability of ligand precursors H2L1, H2L2 and H3L3Cl were monitored at room
1
temperature and at various time intervals for up to 24 h by H NMR spectroscopy.
Solutions of each ligand (ca. 20 mM) were prepared in DMSO-d6 and diluted in either
DMSO-d6 or D2O containing PBS buffer (10 mM, pH=7.4) to obtain final
concentrationd of ca. 1.0 mM.
2.7. BSA binding studies
The binding of metal complexes to BSA (bovine serum albumin) was evaluated by
UV-Visible and Circular Dichroism (CD) spectral measurements. Stock solutions of
BSA were obtained by dissolving the protein in PBS buffer (10 mM, pH~7.4), which
was kept at 4 ºC for at least 3 h to fully hydrate. The albumin concentration was
determined by measuring the absorbance at 279 nm, considering the molar
extinction coefficient 44309 M-1cm-1 [26]. Concentrated stock solutions of CuII-
complexes were prepared in DMSO and diluted with PBS buffer, due to their limited
solubility in aqueous media; the amount of DMSO in the samples was less than 5 %.
2.7.1. UV-Vis
Electronic absorption spectroscopy was used for monitoring (up to 24 h) the
compounds stability and also interactions with BSA. Samples containing BSA (ca. 25
M) and each complex in 1:1 molar ratio (<5 % DMSO v/v) were prepared. All
spectra were recorded in 10 mm optical path quartz cells (240-700 nm). The same
amount of BSA was added to the reference cell to cancel the contribution from the
albumin spectrum.
2.7.2. Circular dichroism
BSA solution in PBS buffer (ca. 385 µM) was titrated with ca. 3.4 mM of 1 or 2
dissolved in DMSO to obtain Cu(II)/BSA molar ratios ranging from 0.02 to 0.5 (<5 %
DMSO). The CD spectra of the solutions were measured and after reaching the
Cu(II)/BSA = 0.5 molar ratio for 1 or 0.2 molar ratio for 2, the CD spectra of the
solutions were periodically recorded up to 22 h. Blank assays were measured for
10

each CD spectrum with the same concentration of both complexes and DMSO
without the protein.
2.7.3. Fluorescence
Successive additions of stock solutions of the complexes in DMSO were directly
introduced into the cuvette containing BSA (ca. 0.68 M) in HEPES buffer (pH 7.4,
10 mM) to obtain a Cu(II)/BSA molar ratios ranging from 0.3 to 22 (<4% v/v DMSO).
The fluorescence emission spectrum was recorded for each ratio between 303 and
550 nm using the following parameters: λex = 295 nm, excitation and emission slit
band-widths of 5.0 nm. Blank assays were measured for each sample. The data
were corrected for absorption and inner-filter-effects.[27, 28]
The fluorescence quenching data were plotted according to the Stern-Volmer
equation [29]:
(2)
where I0 and I are the relative fluorescence intensities of the BSA protein in the
absence and presence of the quencher, respectively. [Q] is the quencher
concentration and Ksv is the Stern-Volmer quenching constant.
The KSV constants were obtained from the linear regression of the plot of vs. [Q].
The binding constants were determined using the Scatchard equation [30]:
(3)
where Ksch represents the binding constant obtained from the slope of the plot of
and n is the number of binding sites per protein molecule.
2.8. DNA binding
2.8.1. ctDNA solutions preparation
Stock solutions of calf thymus DNA (ctDNA) were prepared by dissolving the solid in
PBS buffer (10 mM, pH=7.4), and kept at 4 ºC for at least 48 h for hydration. The
absorbance ratio (A260/A280) of ca. 1.9 indicated that the ctDNA was sufficiently
11

protein-free. The ctDNA stock solutions concentration was calculated based on their
absorbance at 260 nm, using the per nucleotide extinction coefficient ε260 = 6600 M1
cm1 [31].
2.8.2. Electronic absorbance spectroscopy
Successive aliquots of ctDNA stock solution (ca. 3 mM) in PBS were added to
solutions of the complexes (15.9-35.5 M) in PBS (< 2.5% DMSO, v/v) to obtain final
concentrations of ctDNA ranging from 2.3 – 111 M. The mixtures were equilibrated
for 5 min prior to UV-Vis spectroscopy measurements. To compensate the
contribution of ctDNA and DMSO absorption exactly equal amounts of ctDNA and
DMSO were added to the sample and reference cells in each measurement (250-
600 nm). 10 mm optical path length quartz cells were used. Based on the UV-Vis
absorption variation the binding constants were determined by fitting the data to the
Kaminoh equation [32].
(4)
Where ASat is the absorbance value under saturation conditions, A0 is the
absorbance of the complex alone; K is the Kaminoh constant determined from the
plot of Abs vs. [ctDNA].
The binding constants (KBC) of the complex to ctDNA were also determined by fitting
the data with the Scatchard equation [33]:
(5)
where ap = [a-f]; = [b-f]; a is the ratio between the absorption measured and
the complex concentration (Abs/[complex]); b and f is the extinction coefficient of
the compound bound to ctDNA and free in solution, respectively; [ctDNA] is
concentration of ctDNA (in base pairs) and KBC is the intrinsic binding constant
calculated by the representation of [ctDNA]/ap vs. [ctDNA].
2.8.3. Circular Dichroism
12

Circular dichroism titrations of 4 and 5, which contain phenanthroline as co-ligand,
with ctDNA were done keeping the concentration of ctDNA constant while varying
the complexes’ concentration. ctDNA solutions were titrated by adding small
amounts of the complexes dissolved in DMSO (1.0-1.9 mM) to obtain complex/DNA
molar ratios ranging from 0.02 to 1.5. The assays were performed directly in the 10
mm quartz cuvette and the CD spectra were collected from 230 to 600 nm with 1 nm
bandwidth using a UV photomultiplier. Blank spectra were measured without ctDNA
and subtracted from the corresponding sample.
3. Results and discussion
3.1. Synthesis and characterization
Schiff base ligands H2L1 and H2L2 were obtained in good yields by condensation
between either pyridoxamine and salicylaldehyde or pyridoxal and 2-aminophenol
(see Scheme 1). The reduced Schiff base H3L3Cl was obtained in moderate yield
(40%) by reduction of H2L2 with NaBH4. All compounds were characterized by
elemental analysis and the usual spectroscopic techniques [see Experimental part
and Supplementary Material (SM) for details], which confirmed the proposed
formulations.
Scheme 1. Reaction scheme for the synthesis of the Schiff bases (H2L1-2) and
reduced Schiff base (H3L3Cl).
13

The ligand precursors reacted with CuCl2 in methanol, yielding Cu(II)-complexes 1-3,
in moderate yields. In the reactions in which phenanthroline was also added, ternary
[Cu(L)(phen)] complexes (4 and 5) were obtained also in moderate yields. All
complexes are new and were characterized in the solid state and in solution by the
usual spectroscopic techniques, which will be discussed in detail in the following
sections. The formulation proposed for the complexes is depicted in Figure 1.
Fig. 1. Molecular formulas proposed for the prepared CuII complexes.
All compounds were characterized by ESI-MS spectrometry and the spectra are
collected in SM (see also the Experimental part). In most cases it was possible to
identify the molecular ion peaks in the positive, [L+H]+ or [M+H]+, or negative mode,
[L-H]− or [M-H]−, as well as the copper isotropic pattern in the complexes’ peaks.
Adducts with formic acid or isopropanol were also found in the MS spectra of some
metal complexes. The presence of chloride ions in complexes and ligands was
confirmed by anion exchange chromatography. For complexes 1, 3 and 5, only one
chloride ion was detected, while two were determined for H3L3Cl. For the remaining
14

complexes and ligand precursors it was not possible to carry out the chloride
determination due to their low solubility, and thus, formulations were proposed based
on elemental analysis and mass spectrometry data.
3.1.1 NMR spectroscopy
The 1D and 2D 1H NMR and 13C NMR spectra of the ligands were measured at room
temperature in DMSO-d6 or D2O (Table 1). The azomethine proton signal for H2L1
appears as a singlet at 8.61 while for H2L2 this proton is much more deshielded
appearing at 9.30 ppm. However, the corresponding azomethine carbon is more
shielded in H2L1 than in H2L2, possibly due to additional conjugation with the
aromatic system of the pyridoxal moiety in H2L2. For H3L3Cl the azomethine proton
signal is not detected, as expected. Also, the appearance of a signal at 4.79 ppm,
attributed to a CH2 moiety, resulting from the azomethine reduction, was difficult to
distinguish in the 1H NMR spectrum due to overlap with the proton signal of residual
water. However, this signal was undoubtedly identified in the HSCQ NMR
experiment (see Fig S1.1.17), which showed a cross peak at 4.79 ppm (1H NMR)
and 45.03 ppm (13C NMR). The broad signals observed downfield for the samples
dissolved in DMSO-d6 (H2L1 and H2L2) correspond to the hydroxyl protons. These
exchangeable protons were identified by its disappearance after addition of small
amounts of D2O (data not shown). For H2L1 the two deshielded signals (~ 13.30 and
9.05 ppm) are assigned to the aromatic hydroxyl protons, while the peak at 5.20 ppm
is assigned to the aliphatic hydroxyl proton. However, for H2L2 the resonances of the
exchangeable protons disappear after the addition of a drop of D2O, followed by
hydrolysis (see section 3.4). The chemical shift of all protons and carbons are listed
in Experimental section and the respective spectra are collected in SM.
15

Table 1
Selected peaks and assignment for protons and carbon atoms of the NMR spectra
(chemical shift / ppm).
Compound /
solvent
CH=N
CH2N
CHpyr
CH3
Haromatic
H2L1 / DMSO-d6
1H
13C
1H
13C
1H
8.61
166.14
9.30
156.28
-
-
7.96
139.68
7.98
2.38
19.87
2.55
6.82-7.45
116.53-149.46
6.95-7.71
-
H2L2 / DMSO-d6
H3L3Cl / D2O
-
-
129.43
8.17
16.15
2.67
116.82-151.98
6.86-7.29
4.79
45.03
13C
-
130.95
15.42
116.86-153.63
3.1.2 FTIR
Selected IR bands of the prepared compounds are collected in Table 2. Assignments
were made based on the literature [34] and by comparison of the spectra of
complexes and respective ligand precursors to determine groups involved in
coordination, as shifts and/or intensity changes are expected to occur upon
coordination of the ligand to the metal ion.
All compounds present broad bands in the region between 2700-3500 cm1,
corresponding to absorption bands of the OH groups involved in hydrogen bonding
(symmetrical and asymmetrical) and overtone N-H bands. The IR spectra of the
Schiff base ligands show the characteristic strong (C=N) band at 1643 and 1597
cm1, for H2L1 and H2L2, respectively (Fig S1.1.21). In the complexes’ spectra, the
(C=N) absorption appears at slightly lower energy indicating coordination to the
copper(II) ion. A similar behavior was observed for the bands assigned to the (C-
O)phenolate (ca. 1250 cm1), which is compatible with the involvement of this group in
the coordination to the Cu(II) ion [35]. The phenolic O-H stretching vibration is
observed in the free ligands between 3140 and 3427 cm1. Upon complexation the
disappearance of the O-H stretching absorption is expected, due to the
deprotonation and coordination to the metal center. Nevertheless, the presence of
the aliphatic O-H group from the pyridoxal moiety, together with the presence of
water molecules in the complexes adds some difficulty to establish by FTIR the
16

coordination of ligands by the phenolic donor group to the metal center. On the other
hand, the two absorption bands at 3220 and 3376 cm1 of the reduced Schiff base
ligand (H3L3Cl) may originate from the vibration of the asymmetrical and symmetrical
stretching modes of the (N-H) group [36]. At 3429 cm1 (appearing as a shoulder)
the band might be attributed to (OH) vibrations. Moreover, the H3L3Cl ligand
precursor presents two weak bands around 2700 and 2600 cm1 associated with the
amine salt functional group of the reduced Schiff base [34], which corroborates the
proposed formulation. These peaks disappear when the ligand is coordinated to the
Cu(II) center.
Table 2
Selected FTIR bands and assignments for the ligand precursors and Cu(II)
complexes (cm−1).
Compound
(OH) (C=N) (N-H) (C-O) (C-O)
imine
phenolate
alcohol
3422
3142
1035
H2L1
1642
---
---
1246
1023
1032
1010
1039
1020
1040
1049
1019
1047
1017
H2L2
3362
3429
1597
1236
3376
3220
---
H3L3Cl
---
1279
1241
1201
1
2
3340
3341
3218
3429
3350
3428
3164
3447
3277
1628
1587
---
---
---
---
3
4
5
---
1260
1243
1206
1635
1583
1031
1040
1014
3.1.3. Electronic absorption spectra
17

The UV-Vis spectroscopic characterization was done in DMSO by measuring spectra
at different concentrations (see SM). Figure 2 shows the spectra measured for all
compounds. For the Schiff bases H2L1-2, a strong band is observed between 315 and
360 nm, which can be assigned to the n transition, associated with the
azomethine bond. This band is not detected in the spectrum of the reduced Schiff
base H3L3Cl and the molar absorptivity is much higher for H2L2 than H2L1. Probably
this difference is due to the extended conjugated  system that is present in H2L2,
since the transition dipolar moment is higher in -conjugated systems and
consequently the transition probability increases [37]. Closer observation shows that
this band is rather broad in H2L2 and must be, in fact, two bands: the n and

system suffer strong
bathochromic shifts, when compared to their position in H2L1. Bands centered at ca.
285 nm, present in the spectra of all compounds are due to  intraligand
transitions. Complexes 4 and 5 show very strong bands in the UV region due to the
presence of the aromatic N-heterocyclic 1,10-phenanthroline ligand. The spectra of
the Cu-complexes show the presence of ligand to metal charge transfer (LMCT)
bands at ca. 450 nm, due to transitions from the phenolate to the metal center.
Bands in the visible were observed for solutions of complexes 1 and 4 at ca. 670 nm
(see Fig S1.2.2 and S.1.2.11) due to d-d transitions [38, 39]. Although present in the
spectra of the other Cu-complexes they are not easily observed.
a
18

b
Fig. 2. Electronic absorption spectra of ligand precursors (a) and complexes (b)
measured in DMSO.
3.1.4. Electronic Paramagnetic Resonance
The X-band EPR spectra of the Cu(II) complexes were recorded at 77 K or 100K.
Complexes 1-4 were dissolved in DMSO, DMF or MeOH and complex 5 in DMF. All
complexes show spectra with axial symmetry and g|| > g⊥ > 2.0, indicating the
presence of a dx2-y2 ground state with copper(II) located in square-based
geometries, but only 1, 4 and 5 show well resolved spectra with four hyperfine
coupling bands in the high field region. No super hyperfine coupling was observed in
any complex spectra. Figure 3 shows the spectrum measured for complex 1 in
DMSO and the one obtained by simulation [40] which yielded the spin Hamiltonian
parameters included in Table 3. The empirical ratio gz/Az calculated for complexes 1,
4 and 5 is 134 and 130 cm, respectively, within the interval 100 - 135 cm, which
according to Addison et. al. [41] indicates the presence of small distortions in square-
based coordination geometries. The remaining complexes (2 and 3 and 4) display
broad bands and poor resolution spectra with g~2.1 (Fig. S1.2.16). This can be due
attributed to lower solubility but it is probably due to or formation of aggregates and
oligomeric species. It is reasonable to assume that the molecular geometry of
complexes 1, 2 and 3 are square planar, while complexes 4 and 5 assume a
pyramidal square-based geometry which is in agreement with similar structures
proposed for related compounds.[42-45]
19

Fig. 3. First derivative X-band EPR spectrum measured for 1 in DMSO (ca. 3 mM)
as well as the simulated spectrum, which yielded the parameters listed in Table 3.
Table 3
EPR spin Hamiltonian parameters of Cu(II) complexes 1, 4 and 5.
Complex Solvent
g||
A||(x104
cm-1)
g||/A|| (cm)
g
A( x104

1
cm )
1
1
4
5
DMSO
MeOH
MeOH
DMF
2.280
2.278
2.268
2.247
2.071
2.058
2.080
2.058
180.7
176.9
167.9
181.0
16.5
16.8
36.9
32.0
134
129
135
130
3.2. Radical scavenging activity
1,1-Diphenyl-2-picryl-hydrazyl (DPPH) is a stable free radical that has one
delocalized unpaired electron. The radical forms deep purple coloured solutions in
organic solvents, which upon reaction with other radicals, electrons, or hydrogen
atoms, leads to the formation of a colourless species. Following the absorbance at
517 nm of solutions prepared with different ratios of compound/DPPH allows the
assessment of their radical scavenging activity. Therefore, the DPPH assay was
20

used to evaluate the antioxidant potential of the complexes and their respective
ligand precursors (see experimental part). Correlations have been found for the
radical scavenging activity and redox potentials of compounds [46-48]. Table 4
shows the effective concentration (IC50) obtained for all compounds. Complexes 1, 4
and 5 show similar and moderate activity, bellow 100 M. Unexpectedly, H2L2 shows
very high radical scavenging capacity, even better than comparable to the positive
control, ascorbic acid. However, reduction of the C=N bond (in H3L3Cl) resulted in
total loss of the radical scavenging activity. While for H2L1 coordination to Cu(II)
increases considerably its activity (compare IC50 values for H2L1 and 4) for H2L2 and
H3L3Cl this did not happen, as both corresponding Cu-compounds show very poor
antioxidant activity. Complexes 4 and 5, containing phen, show equal or higher
antioxidant potential than 1 and 2, which do not have the aromatic N-base.
The very high radical scavenging activity observed for H2L2 is probably related to its
extended conjugated system, which stabilizes the phenolic radical, which is formed
upon transfer of a hydrogen atom from its hydroxyl group.
Table 4
IC50 values (μM) of the scavenging activity determined for the synthetized
compounds by the DPPH assay. The standard deviation of 3 replicates is included.
Compound
IC50 /M
704 ± 172
14 ± 4
H2L1
H2L2
H3L3Cl
417 ± 36
1
83 ± 4
172 ± 26
443 ± 17
85 ± 14
83 ± 11
11.6
2
3
4
5
Ascorbic acid
3.3 Cytotoxicity assays
The cytotoxic activity of the complexes and ligand precursors (Table 5) was
evaluated on ovarian (A2780) and breast (MCF7) human cancer cell lines, after 24 h
21

of incubation. The IC50 values were determined using the MTT assay, which
measures the mitochondrial dehydrogenase activities in living cells.
All compounds exhibited moderate to very high in vitro cytotoxicity against both
cancer cell lines (IC50 values ranging from 0.7 and 86 μM) while the ligand
precursors are not active (IC50 > 100 μM), at least within the time frame chosen, 24
h. Comparison between the two different cell types showed that compounds
containing the phen ligand exhibit higher cytotoxicity against the ovarian cancer cells
than the breast cells, while the other complexes showed more efficacy against the
breast cells. The complex containing the reduced Schiff base (3) revealed lower
activity than the other one with the Schiff base (2) in both cell lines. Taking into
account the similarity of the ligands and the lipophilicity of the compounds, H2L2 is
expected to be much more lipophilic than H3L3Cl, the logP order being H2L2 > H2L1 >
H3L3Cl [24], and thus, a similar order is expected for their copper complexes.
Therefore, for 2 and 3 lipophilicity can be correlated with cytotoxicity. However, other
properties must also be at play since IC50 (1) > IC50 (3) and LogP (1) > LogP (3).
Most complexes are comparable to, or much better than cisplatin, which showed
moderate activity for these cells at 24 h incubation. For both tumor cells the ternary
complexes containing phen showed IC50 values much lower than cisplatin. Overall,
data shows that there is a relevant enhancement of the anti-proliferative activity due
to the synergetic effect of combining Cu(II) and a Schiff base, which is further
enhanced by the presence of phen coordinated to the metal center, increasing
significantly the cytotoxicity, particularly in the ovarian cancer cell line, with a
decrease in IC50 of more than 30 times.
22

Table 5
Cytotoxicity (IC50 in μM ± SD) of the ligand precursors and copper(II) complexes on
two human cancer cells (A2780 and MCF7) after 24 h of incubation.
Compound
A2780
86 ± 38
23 ± 8.5
49 ± 18
1.8 ± 0.6
0.7 ± 0.1
>100
MCF7
34 ± 11
6.8 ± 3.5
19 ± 5.7
9.3 ± 3.0
2.6 ± 1.0
>100
1
2
3
4
5
phen, H2L1, H2L2 and H3L3Cl
[Cu(phen)2]
0.4 ± 0.2
36 ± 8.0
1.5 ± 0.4
59 ± 12
cisplatin [49]
3.4 Stability studies
1
The stability of the Schiff bases H2L1,2 was evaluated in aqueous buffer by H NMR
and UV-Vis absorption spectroscopy, as this type of ligand typically hydrolyzes in the
presence of water containing solvents. The studies shown in SM confirm that the
Schiff bases H2L1,2 indeed hydrolyze in the presence of water. The same assays
were carried out for the reduced Schiff base H3L3Cl and, as expected,[50] no signs
of hydrolysis were observed (Fig. S3.4).
The stability of the complexes in organic solvents was followed by UV-visible
spectroscopy in DMSO. All Cu(II)-complexes are stable, maintaining their original
spectra with minor changes after ca. 24 h (Fig. S3.7a-S3.11a). In order to mimic
physiological conditions, the stability was also evaluated in pH 7.4 buffered media
(PBS or NH4HCO3) containing a minimum amount of organic solvent. The behavior
of the compounds in aqueous solvent over time is very similar for both buffer
solutions. Complexes 2 and 3 start to precipitate after 30 minutes at concentrations ≥
25 M. The same solubility issue occurs for 5 (after 2 h), with lower absorbance
values being recorded with time, but no evident changes in the absorption bands’
position. Complex 1 is stable for at least 24 h and 4 shows only very small spectral
changes over time (less than 10 % of absorbance decrease at  = 356 nm after 24
h) (Fig. S3.10b). The small changes observed in most spectra may simply be due to
solvent exchange (H2O and DMSO).
23

As general conclusion, the Schiff base ligands suffer hydrolysis in aqueous solution
under conditions mimicking physiological media, although when coordinated to Cu(II)
their stability increases considerably. The complexes maintain their structure during
the time frame usually necessary to carry out interactions studies with biomolecules.
3.5 Interaction with BSA
Serum albumins are the most abundant proteins in the blood vascular system.
Human serum albumin (HSA) is ca. 60 % in the plasma and it is mainly involved in
transport to cells [51]. It is widely accepted that the binding strength between a drug
and albumin may influence its ADME (absorption, distribution, metabolism, and
excretion) properties, which may lead to loss or enhancement of the drug’s efficacy
[52, 53]. In the following studies, BSA was chosen as model protein, due to its
structural similarity to the homolog HSA [54] and to its relative importance in cell
culture media assays, since it is present in relatively high concentrations. The cells
ability to uptake the complexes may also depend on the binding strength between
the compound and BSA or/and other complex species formed in solution that can
have different affinities to this protein. Furthermore, several specific cell receptors for
albumin have been reported in the literature [55], which may facilitate the complexes
entry into cells. Therefore, understanding the interaction between this protein and the
new complexes is extremely relevant to properly evaluate their therapeutic potential.
Besides one coordination position occupied by a labile molecule, complexes 1-3
have other coordinating positions that may be used to form ternary species of the
type [Cu(Ln)(BSA)]. In complexes 4 and 5, BSA coordination, if it occurs, may take
place in the apical free position, with formation of an octahedral species,
[Cu(Ln)(phen)(BSA)], or by loss of one of the ligands forming [Cu(Ln)(BSA)] or
[Cu(phen)(BSA)] species. Other possibilities, for all complexes, is the loss of all
ligands and binding as [Cu(BSA)] and [Cu2(BSA)], since Cu(II) has high affinity for
BSA with albumin paying an important role in copper transport in the bloodstream
[56, 57]. Therefore, different spectroscopic studies were performed aiming to
properly understand the binding of the complexes to BSA.
3.5.1. UV-visible spectra
24

Since the hydrolytic stability of the complexes may increase by binding to albumin,
the stability of the complexes was evaluated in the presence of BSA, under the same
experimental conditions used in its absence. In all cases, although no precipitation
was observed by naked eye, the spectra measured for 2 and 3 show some
evidences of precipitation, such as, lower absorbance intensity values recorded with
time (Fig. S4.2 and S4.3).
For all complexes, changes were observed. For 1 (Fig. 4) it took ca. 1 h for the
system to reach equilibrium; some bands shifted to lower energy, new bands
appeared and isosbestic points were observed at 310, 340 and 395 nm. All these
observations are consistent with the transformation of the original [Cu(L1)Cl]−
complex into a new species, possibly [Cu(L1)(BSA)]. Comparison of the spectra (at t
> 300 min, Fig. S4.1) of solutions containing BSA and either complex 1, H2L1 or
[Cu2+ + H2L1] shows that it is quite difficult to see any differences between these
spectra in the UV range, and therefore we cannot state without doubt that the
observed time changes are due to formation of the ternary [Cu(L1)(BSA)] complex.
Complex 3 maintains the original UV-Visible spectrum in solution for at least 3 h, in
the presence of BSA, with no signs of interaction, hydrolysis or precipitation, after
which absorbance decreases (Fig. S4.3). Complex 4 shows a hyposochromic shift of
~3 nm of the band with maximum at 267 nm, after 24 hours, with two isosbestic
points appearing at 337 and 383 nm, indicating that the complex interacts with BSA
by forming a new species (Figure S4.4). Interestingly, there are quite a few
similarities in the electronic absorption spectra obtained for complex 1 and 4 with
time in the presence of 1 mole equivalent of BSA, such as the appearance of new
bands around 320 and 410 nm. This probably indicates the formation of the same
type of species, (p.e. [Cu(L1)(BSA)]), although we cannot state it with certainty based
only in UV-Vis experiments.
For complex 5 only minor changes are observed in the first 100 min, p.e. the band
centered at 440 nm decreases its intensity and an isosbestic point is present at 336
nm. These are consistent with formation of a new species but the phen ligand should
be in the coordination sphere of the metal ion as no changes are seen in the 267 nm
band.
25

Fig. 4. Electronic absorbance of complex 1 (24.9 M, 4.6 % DMSO v/v) with ca. 25
M of BSA measured in PBS (10 mM, pH~7.4) with time. Inset: absorbance variation
with time at = 403 nm (■); 363 nm (○); 321 nm (▲); 287 nm (●); 270 nm (□). The
arrows show absorbance variations with time.
3.5.2. Circular dichroism
The prepared Cu(II)-complexes 1-3 are not chiral compounds (3 and 4 may be
considered racemic mixtures with chirality at the metal), and therefore their solutions
do not show circular dichroism spectra. Upon binding to a chiral molecule, such as
BSA, chirality transfer may happen between the biomolecule and the Cu-complex.
Particularly interesting to observe are the copper d-d transitions as these may
appear in the CD spectra and provide useful information concerning the coordination
environment of the metal ion.
The CD spectra of BSA (388 µM) upon addition of increasing amounts of complex 1
are shown in Figure 5. Complex 1 is not optically active and BSA does not present
absorption bands in the visible range studied, so, the induced CD (ICD) bands
observed in Fig. 5 are clear proof of binding of the complex close to chiral groups of
BSA. The positive band at ca. 490 nm and the negative band at ca. 560 nm
somehow resemble the CD spectra of BSA when free Cu2+ is added (Fig. S4.10),
however the presence of a new distinct negative band in this system at ca. 400 nm,
confirms that this is a new species, probably [Cu(L1)(BSA)]. At higher complex/BSA
26

molar ratios precipitation occurs, which precludes the observation of new Cu-BSA
species for higher ratios. No relevant CD spectral changes were observed with time
(up to 22 h) for a solution containing a complex/BSA molar ratio of ca. 0.5 (Fig S4.6).
Very similar CD spectra, with the same shape and similar intensities were observed
when complex 4 was added to BSA (Fig. S4.8-4.9), indicating that the formed BSA-4
species probably shows the same type of binding, when compared with complex 1
and suggests that both CD-active species are analogous.
For complex 2 only a positive band at 420 nm developed upon addition of small
amounts of the complex (see Fig. S4.7), however, precipitation for ratios higher than
0.2, did not allow performing a complete study.
1/BSA
Fig. 5. CD spectra of a BSA titration (388 µM) with a solution of complex 1 (ca. 3.4
mM) in PBS buffer (pH 7.4, 10 mM, < 5 % DMSO). Complex 1/BSA molar ratios are
indicated in the figure.
3.5.3. Fluorescence
Fluorescence spectroscopy is one of the most commonly used techniques to
evaluate the role of BSA in the transport of metallodrug candidates. BSA has intrinsic
fluorescence due to the presence of fluorescent amino acids, the main contribution
being related to two tryptophans located at positions 134 and 212 [58]. Trp-134 is on
27

the albumin surface in subdomain IB and Trp-212 is positioned within a hydrophobic
pocket in subdomain IIA [54].
When BSA is excited at 295 nm, an emission band appears with maximum at ca.
340 nm. Under the same conditions complex 1 has intrinsic fluorescence and
exhibits a band with a maximum centered at ca. 401 nm. Hence, the data obtained
from the fluorescence quenching experiments were corrected by subtracting the
spectra of the complex at each concentration. The emission intensity changes upon
addition of complex 1 to BSA (Figure 6), this indicating a protein conformational
change or binding between the protein and the complex, which may be covalent or
by hydrophobic interactions ( stacking) [58]. Increasing the concentration of the
metal complex leads to progressive quenching of the fluorescence intensity at 340
nm, up to 64%, simultaneous to the appearance of the complex emission at 401 nm,
forming an isoemissive point at 379 nm. Additionally, a small blue shift of about 3 nm
is observed. This energy change may indicate that the polarity at the binding site of
the tryptophan environment in the protein decreases [59] indicating that the
fluorescent residues become located in a more hydrophobic environment [60]. A
similar blue shift in the fluorescence of BSA emission maximum is observed in the
presence of the other Cu-complexes as well as for 2,4-dinitrophenol (DNP). For DNP
binding to the protein was attributed to site I, in subdomain IIA [61].
Determination of binding constants is crucial for the development of a drug, since
most of the drugs in the market can reversibly bind albumin with moderate affinity
[binding constants in the range (1-15) × 104 M1] [62]. In the particular case of
complex 1, constants were calculated by applying the Stern-volmer (eq. 2) and
Scatchard (eq. 3) models. A linear Stern-Volmer plot was obtained suggesting a
single mechanism involving static quenching [63]. The binding constant shows that
the binding between complex 1 and BSA is moderate: KSV = 9.9 × 104 M1 and Ksch
=
9.2 × 104 M1 (see SM). When comparing the binding strength between BSA and the
complexes of the type [CuII(Ln)] no significant differences were found (Table 6).
However, when the co-ligand phen is coordinated to the metal center, the binding
affinity increases 3 to 4 times. It is known that the planarity of the ligands
coordinating the metal center plays an important role in terms of the binding affinity
to BSA [64] due to the importance of  stacking interactions with aromatic
28

residues. The higher binding strength of complexes 4 and 5 to albumin could be
explained by the intrinsic planarity of the phen co-ligand.
Table 6
BSA binding constants obtained by fitting the fluorescence quenching data with the
Scatchard (eq. 3) and Stern-Volmer (eq. 2) models.
Scatchard
KSch (× 104
M1)
Stern-volmer
KSV (× 104 M1)
Complex
1
2
3
4
5
9.9
7.9
9.3
9.7
8.0
8.1
13.5
34.1
37.1
30.2
Since BSA is one of the major components in the cellular medium used in
cytotoxicity assays and knowing that, in blood, albumin is the second more important
carrier of Cu(II) ions, these results indicate that a reversible complex/BSA adduct is
formed and the compound is able to be stored/protected and carried by the protein.
29