Development of next-generation photolabile copper cages with improved copper binding properties

Article abstract of DOI:10.1039/c0dt00770f

Seven new nitrogen-donor ligands that contain a photoactive nitrophenyl group within the ligand backbone have been prepared and evaluated for their binding affinity for copper(ii) and zinc(ii). Among this series, the ligand 3Gcage (pyridine-2-carboxylic a

Full text of DOI:10.1039/c0dt00770f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Development of next-generation photolabile copper cages with improved
copper binding properties
Katie L. Ciesienski, Kathryn L. Haas and Katherine J. Franz*
Received 1st July 2010, Accepted 21st July 2010
DOI: 10.1039/c0dt00770f
Seven new nitrogen-donor ligands that contain a photoactive nitrophenyl group within the ligand
backbone have been prepared and evaluated for their binding affinity for copper(II) and zinc(II). Among
this series, the ligand 3Gcage (pyridine-2-carboxylic acid {1-(2-nitro-phenyl)-3-[(pyridin-2-ylmethyl)-
amino]-propyl}-amide) has the best affinity for copper(II), with an apparent dissociation constant at
pH 7.4 of 0.18 fM. Exposure of buffered aqueous solutions of 3Gcage or Cu(II)-bound 3Gcage to UV
light induces bond cleavage in the ligand backbone, which reduces the denticity of the ligands. The
quantum yields of photolysis for 3Gcage in the absence and presence of Cu(II) are 0.66 and 0.43,
respectively. Prior to photolysis, the 3Gcage ligand inhibits copper from generating hydroxyl radicals in
the presence of hydrogen peroxide and ascorbic acid; however, hydroxyl radical formation increases by
more than 300% following light activation, showing that the reactivity of the copper center can be
triggered by light.
Toward these goals, we have developed photoactive caged
metal complexes by incorporating a nitrophenyl group into the
Introduction
backbone of metal-binding ligands.19–21 Our previously reported
first-generation chelator, nicknamed H2cage, cages Cu2+ in a
tetradentate binding site until exposure to UV light cleaves the
ligand backbone to release photoproducts with diminished affinity
for Cu2+.19 Although UV light can be damaging to cells, the
high quantum efficiency of uncaging for these compounds should
minimize such damage.21 In addition to the change in binding
affinity, photolysis also causes a change in reactivity of caged vs.
uncaged copper, as we have shown by the 160% increase in the
ability of copper to generate hydroxyl radicals under Fenton-
like conditions following light-induced uncaging. While these
are promising steps towards using light-activated compounds to
increase oxidative stress in a targeted manner, H2cage is not
without its limitations. One of its drawbacks is that it does not
completely suppress hydroxyl radical formation pre-photolysis.
Furthermore, its effective dissociation constant, Kd¢, at pH 7.4 of
16 pM is not strong enough to keep copper sequestered in the
presence of endogenous copper-binding proteins. For example,
copper circulates in serum predominantly bound to human serum
albumin, which has a Kd¢ of 1 pM for Cu2+ at physiological pH.22
To address these problems, we synthesized and characterized
seven new photolabile analogs of H2cage, the structures of which
are shown in Fig. 1. Our primary goals in this study were to
investigate the effects of changes in the ligand backbone of
H2cage on Cu2+ chelation and reactivity. Studies on several of the
second-generation analogs revealed that replacement of at least
one amide nitrogen in H2cage with a secondary amine improves
the ligand’s affinity for Cu2+. This information led to the design
of our best copper cage to date, the 3rd generation chelator
nicknamed 3Gcage. As shown in Scheme 1, 3Gcage chelates
copper with two pyridyl nitrogens, an amide, and a secondary
amine. Exposure to UV light cleaves the ligand backbone to
release the expected picolinamide (pic) and nitroso-containing
(nitroso) fragments with weaker copper affinity. 3Gcage not only
Copper is an essential element for life and plays an important
role in many physiological processes. Its redox activity makes it an
essential cofactor in enzymes such as Cu, Zn superoxide dismutase,
cytochrome oxidase, ceruloplasmin, tyrosinase, peptidyl-glycyl a-
amidating monooxidase, and many others.1 The same property
that makes copper essential renders it potentially toxic when there
is an imbalance in its cellular homeostasis, which can lead to the
formation of toxic reactive oxygen species (ROS).2 Metal-induced
oxidative stress is associated with several neurodegenerative dis-
eases including Menkes and Wilson diseases, Alzheimer’s disease,
familial amyotrophic lateral sclerosis and prion disease.3–6 Due
to the potentially toxic nature of copper, cells tightly regulate
its intracellular distribution.7–11 Understanding how cells traffic
and utilize copper while suppressing its toxicity has important
ramifications for health and disease.6–12
It has been shown that copper levels are elevated in numerous
malignancies and directly correlate to cancer progression.13 In
addition, cancer cells are known to have an altered redox balance
in which elevated levels of ROS signal adaptations that not only
enable cancer cell survival and proliferation, but in turn exhaust
cellular ROS-buffering capabilities.13–16 While scavenging ROS
may be one approach to depress tumor growth, an opposite
strategy is based on the hypothesis that further increasing ROS
beyond the buffering capacity of malignant cells could provide a
selective way to induce cancer cell death. In this scenario, copper’s
pro-oxidant ability is medicinally promising if it can be used to
induce oxidative stress as a cancer chemotherapy strategy.13,17,18
New methods for controlling the spatial and temporal release
of copper intracellularly would be useful for both the targeted
delivery of copper as a pro-oxidant to induce cell death and for
uncovering its complex involvement in healthy vs. disease states.
Department of Chemistry, Duke University, P.O. Box 90346, Durham, North
Carolina, 27708, USA. E-mail: Katherine.Franz@duke.edu
9538 | Dalton Trans., 2010, 39, 9538–9546
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
absorption at 280 nm. A linear gradient from 10% A in B to 60% A
in B was run from 2 to 32 min with a total run time of 42 min, where
A is MeCN/4% 10 mM pH 7 ammonium acetate buffer and B is
10 mM pH 7 ammonium acetate buffer/2% MeCN. UV-vis spectra
were recorded on either a Cary 50 UV-vis spectrophotometer
or an SI Photonics (Tucson, AZ) model 420 fiber optic CCD
array UV-vis spectrophotometer in a 1-cm pathlength quartz cell.
Photolysis experiments were performed using a screwtop quartz
cuvette illuminated in a Rayonet RPR-100 Photochemical Reactor
˚
containing 16 bulbs, each 3500 A. Reaction mixtures were purified
by preparative reverse-phase HPLC on a YMC C18 column
(250 ¥ 20 mm I.D.) with a linear 40-min gradient from 7–70%
acetonitrile in water or an XBridge semi-prep C18 Column (19 ¥
250 mm) with a linear 40-min gradient from 7–70% acetonitrile in
water.
3-(2-Amino-acetylamino)-3-(2-nitro-phenyl)-N-pyridin-2-
ylmethyl-propionamide (Amcage)
Carbobenzyloxyglycine (0.400 g, 1.912 mmol) and NMM
(0.420 mL, 3.824 mmol) in CH2Cl2 (6 mL) were added to a
150-mL round-bottom flask equipped with a stir bar. Then Py-
BOP (benzotriazol-1-yl-oxytripyrrolidinophosphonium hexaflu-
orophosphate) (0.996 g, 1.912 mmol) in CH2Cl2 (4 mL) was
added and the reaction mixture was allowed to stir for 18 h at
room temperature. After 18 h, one equivalent of 3-amino-3-(2-
nitrophenyl)propionic acid (0.400 g, 1.912 mmol) dissolved in
hot DMF (30 mL) was added and the reaction mixture was
stirred for an additional 18 h. NMM (N-methylmorpholine)
(0.420 mL, 3.824 mmol) and 2-(aminomethyl)pyridine (0.196 mL,
1.912 mmol), both in CH2Cl2 (6 mL) were added to the reaction
mixture. PyBOP (0.996 g, 1.912 mmol) in CH2Cl2 (4 mL) was
then added, and the reaction mixture stirred for 18 h at room
temperature. The solvent was removed and the resulting oil was
taken up in CH2Cl2 (25 mL), filtered, and washed with 1 N HCl
(3 ¥ 10 mL). The aqueous layer was then adjusted to pH 12
and washed with CH2Cl2 (4 ¥ 10 mL). The organic layers were
combined, dried over MgSO4, filtered, and solvent removal gave
an oil that was purified by preparative reverse-phase HPLC on
a YMC C18 column (250 ¥ 20 mm I.D.) with a linear 40-min
gradient from 7–70% acetonitrile in water, to give a white solid
after solvent evaporation. The white solid was then dissolved in
CH2Cl2 (15 mL) and hydrobromic acid (≥ 33% in glacial acetic
acid) (15 mL) was added and the reaction mixture was allowed to
stir for 2 h at room temperature. After 2 h, the reaction mixture
was washed with H2O (4 ¥ 10 mL). The aqueous layer was then
adjusted to pH 12 and washed with CH2Cl2 (4 ¥ 10 mL). The
organic layers were combined, dried over MgSO4, filtered, and
Fig. 1 H2cage and derivatives.
Scheme 1 Synthesis and photolysis of the copper complex of 3Gcage,
[Cu(3Gcage)]+.
has a significantly improved Cu2+ binding affinity, it also retains
a high quantum efficiency of photolysis and provides greater
protection against copper-promoted hydroxyl radical formation
pre-photolysis, while maintaining the increased reactivity post-
photolysis.
Experimental section
General
Unless otherwise noted, chemicals were purchased commercially
and used without further purification. ZinCleav-1 was kindly
provided by S. Burdette.23 All reactions were carried out in the
dark and glassware was covered with foil. The ligand H2cage19
was synthesized and purified as described previously. Spectra for
1H NMR were collected on a Varian Inova 400 spectrometer
with chemical shifts reported in ppm and J values in Hz. Liquid
chromatography/mass spectrometry (LC/MS) was performed on
an Agilent 1100 Series apparatus with an LC/MSD trap and a
Daly conversion dynode detector. A Varian Polaris C18 (150 ¥
1.0 mm) column was used and peaks were detected by UV
1
solvent removal gave a white solid (0.2164 g, 31.7%). H NMR
((CD3)2CO): d 8.45 (1H, d, J = 3.97), 7.97 (1H, d, J = 8.02),
7.72 (1H, d, J = 7.62), 7.64 (2H, m), 7.53 (1H, t, J = 7.62, J =
7.62), 7.21 (1H, d, J = 7.64), 5.91 (1H, t, J = 5.37, J = 5.37), 4.43
(2H, dd, J = 15.98, J = 40.76), 3.74 (2H, m), 3.06 (1H, dd, J =
6.62, J = 14.52), 2.94 (1H, dd, J = 4.79, J = 14.76). 13C NMR
((CD3)2CO): d 169.86, 169.75, 158.24, 148.83, 148.59, 137.67,
136.41, 133.29, 129.20, 128.18, 124.31, 121.90, 120.97, 54.00,
46.47, 44.25, 39.94; m/z (EI) 358.1 [M+H]+, C17H20N5O4 requires
358.15.
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9538–9546 | 9539
©

View Article Online
2-Methyleneamino-but-2-enoic acid [2-[2-(1H-imidazol-4-yl)-
ethylcarbamoyl]-1-(2-nitro-phenyl)-ethyl]-amide (Imcage)
18 h. NMM (0.089 mL, 0.813 mmol) and di-(2-picolyl)amine
(0.015 mL, 0.081 mmol), both in CH2Cl2 (3 mL) were added to
the reaction mixture. PyBOP (0.042 g, 0.081 mmol) in CH2Cl2
(2 mL) was then added, and the reaction mixture stirred for 18 h
at room temperature. The solvent was removed and resulting oil
was purified by HPLC, giving a yellow oil (0.036 g, 44%). 1H NMR
(CD3OD): d 8.76 (dd, J = 10.9, J = 5.2, 2H), 8.66 (d, J = 4.5, 1H),
8.37 (dd, J = 11.1, J = 4.6, 1H), 8.07–7.88 (m, 5H), 7.81 (dd, J =
13.8, J = 7.3, 1H), 7.62–7.53 (m, 5H), 7.53–7.47 (m, 1H), 5.92 (dd,
J = 7.8, J = 4.5, 1H), 5.17 (dd, J = 55.6, J = 18.2, 2H), 5.06–4.96 (m,
2H), 3.39–3.36 (m, 1H), 3.14–3.08 (m, 1H). 13C NMR (CD3OD):
d 172.33, 164.49, 155.59, 154.92, 148.41, 147.69, 145.05, 141.74,
139.79, 137.57, 136.48, 133.32, 128.61, 128.43, 126.74, 125.61,
124.87, 124.47, 123.86, 123.02, 121.87, 53.23, 51.14, 37.34; m/z
(EI) 497.1 [M+H]+, C27H25N6O4 requires 497.19.
Equimolar quantities of picolinic acid (0.100 g, 0.813 mmol) and
NMM (0.089 mL, 0.813 mmol) in CH2Cl2 (3 mL) were added to a
50-mL round-bottom flask equipped with a stir bar. Then PyBOP
(0.423 g, 0.813 mmol) in CH2Cl2 (2 mL) was added and the reaction
mixture was allowed to stir for 18 h at room temperature. After
18 h, one equivalent of 3-amino-3-(2-nitrophenyl)propionic acid
(0.171 g, 0.813 mmol) dissolved in hot DMF (15 mL) was added
and the reaction mixture was stirred for an additional 18 h. NMM
(0.089 mL, 0.813 mmol) and histamine (0.091 g, 0.815 mmol),
both in CH2Cl2 (3 mL) were added to the reaction mixture. PyBOP
(0.423 g, 0.815 mmol) in CH2Cl2 (2 mL) was then added, and the
reaction mixture stirred for 18 h at room temperature. The solvent
was removed and the resulting oil was purified by HPLC, giving a
yellow oil (0.053 g, 16%). 1H NMR ((CD3)2CO): d 9.86 (1H, d, J =
7.49), 8.73 (1H, s), 8.67 (1H, d, J = 4.33), 8.00 (1H, d, J = 7.58),
7.95 (2H, d, J = 7.42), 7.78 (1H, d, J = 7.86), 7.65 (1H, t, J = 7.60,
J = 7.60), 7.57 (1H, m), 7.50 (1H, t, J = 7.74, J = 7.74), 7.32 (1H,
s), 5.96 (1H, m), 3.52 (2H, t, J = 6.41, J = 6.41), 3.06 (1H, dd, J =
6.88, J = 14.87), 2.93 (3H, m). 13C NMR ((CD3)2CO): d: 170.40,
163.59, 149.82, 148.62, 137.55, 137.47, 133.58, 133.43, 131.78,
128.93, 128.43, 126.62, 124.35, 121.97, 116.36, 46.89, 39.89, 37.99,
24.52; m/z (EI) 409.2 [M+H]+, C20H21N6O4 requires 409.16.
3-[2-(Bis-pyridin-2-ylmethyl-amino)-acetylamino]-3-(2-nitro-
phenyl)-N-pyridin-2-ylmethyl-propionamide (3arm-1)
Amcage (0.029 g, 0.081 mmol), bromoacetic acid (0.013 mL,
0.178 mmol) and K2CO3 (0.045 g, 0.324 mmol) in DMF (10 mL)
were combined in a 50-mL round-bottom flask equipped with a
stir bar and the reaction mixture was allowed to stir for 18 h at
room temperature. The solvent was removed and the resulting oil
was taken up in CH2Cl2 (20 mL) and washed with H2O (4 ¥ 10 mL).
The organic layers were combined, the solvent was removed and
the resulting oil was purified by HPLC, giving a yellow oil (0.022 g,
49.6%). Analysis by LC-MS and 1H NMR showed ≥ 98% purity.
1H NMR ((CD3OD): d 8.76–8.71 (m, 3H), 8.67–8.65 (m, 1H),
8.59–8.54 (m, 1H), 8.26–8.22 (m, 1H), 8.18 (td, J = 7.8, J = 1.6,
2H), 8.12–8.07 (m, 1H), 8.02 (d, J = 8.0, 1H), 7.95 (dd, J = 9.0, J =
7.7, 2H), 7.73–7.60 (m, 8H), 7.53 (ddd, J = 15.5, J = 7.1, J = 1.4,
2H), 5.82 (t, J = 6.8, 1H), 5.03 (s, 2H), 4.60 (d, J = 5.9, 2H), 4.27
(s, 2H), 3.60 (d, J = 7.3, 2H), 2.96 (d, J = 6.9, 2H). m/z (HR-FAB)
540.2366 [M+H]+,C29H30N7O4 requires 540.2359.
7,14-Bis-(2-nitro-phenyl)-1,4,8,11tetraaza-cyclotetradecane-
2,5,9,12-tetraone (Macrocage)
Macrocage was synthesized on a Protein Technologies PS3
automated peptide synthesizer using Fmoc-Gly-Wang resin (Ad-
vanced ChemTech) (0.048 g, 0.033 mmol). Fmoc was removed
using 20% piperdine in DMF. Couplings of Fmoc-3-amino-3-(2-
nitrophenyl)propionic acid (0.028 g, 0.066 mmol) and Fmoc-Gly
(0.029 g, 0.099 mmol) were carried out using O-benzotriazole-
N,N,N¢,N¢-tetramethyl-uronium-hexafluoro-phosphate (HBTU)
(0.038 g, 0.099 mmol) and NMM (5 mL, 0.045 mmol) in DMF
(10 mL) for 30 min. The uncyclized peptide was obtained by
treating the resin-bound molecule with a 95% trifluoroacetic acid
(TFA) solution for 2 h. TFA was removed and the resulting oil
was washed (3 ¥ 5 mL) with diethyl ether. The resulting oil and
NMM (5 mL, 0.045 mmol) in DMF (200 mL) were added to a
500-mL round-bottom flask equipped with a stir bar. Then HBTU
(0.038 g, 0.099 mmol) in DMF (5 mL) was added and the reaction
mixture was allowed to stir for 18 h at 100 ◦C. Solvent removal gave
an oil that was purified by HPLC, giving a white solid (0.003 g,
16.4%). Analysis by LC-MS and HR-FAB showed > 98% purity.
m/z (HR-FAB) 499.1578 [M+H]+, C22H23N6O8 requires 499.1577.
Pyridine-2-carboxylic acid {1-(2-nitro-phenyl)-3-[(pyridin-2-
ylmethyl)-amino]-propyl}-amide (3Gcage)
Equimolar quantities of picolinic acid (0.05 g, 0.40 mmol) and
NMM (0.04 mL, 0.40 mmol) in CH2Cl2 (2 mL) were added
to a 25-mL round-bottom flask equipped with a stir bar. The
reaction mixture was cooled over an ice bath for ten min then
PyBOP (0.21 g, 0.40 mmol) in CH2Cl2 (2 mL) was added. The
reaction mixture was allowed to warm to room temperature with
stirring for 18 h. After 18 h, one equivalent of 3-amino-3-(2-
nitrophenyl)propionic acid (0.08 g, 0.40 mmol) dissolved in hot
DMF (5 mL) was added and the reaction mixture was stirred
for an additional 18 h. NMM (0.04 mL, 0.40 mmol) and PyBOP
(0.21 g, 0.40 mmol), both in CH2Cl2 (2 mL) were added to the
reaction mixture. The resulting solution was cooled over an ice
bath for ten min then the reaction mixture stirred for 18 h at room
temperature. The reaction mixture was again cooled over an ice
bath for 30 min then NaBH4 (1.51 g, 4.0 mmol) was added. The
reaction mixture was allowed to warm to room temperature and
continued stirring for 10 min. The solvent was removed and the
resulting oil was taken up in CH2Cl2 (15 mL) and washed with
saturated sodium bicarbonate (3 ¥ 10 mL). The organic layers
were combined, dried over MgSO4, filtered, and solvent removal
Pyridine-2-carboxylic acid [2-(bis-pyridin-2-ylmethyl-
carbamoyl)-1-(2-nitro-phenyl)-ethyl]-amide (3 arm-3)
Equimolar quantities of picolinic acid (0.010 g, 0.081 mmol) and
NMM (0.009 mL, 0.081 mmol) in CH2Cl2 (3 mL) were added to a
50-mL round-bottom flask equipped with a stir bar. Then PyBOP
(0.042 g, 0.0813 mmol) in CH2Cl2 (2 mL) was added and the
reaction mixture was allowed to stir for 18 h at room temperature.
After 18 h, one equivalent of 3-amino-3-(2-nitrophenyl)propionic
acid (0.017 g, 0.081 mmol) dissolved in hot DMF (15 mL) was
added and the reaction mixture was stirred for an additional
9540 | Dalton Trans., 2010, 39, 9538–9546
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
gave a yellow oil. To the resulting oil equimolar quantities of p-
toluenesulfonyl chloride (0.076 g, 0.40 mmol) and triethylamine
(0.056 mL, 0.40 mmol) in CH2Cl2 (2 mL) were added with
stirring for 2 h. After 2 h, 2-(aminomethyl)pyridine (0.04 mL,
0.40 mmol) was added and reaction mixture stirred for 18 h at
room temperature. The solvent was removed and the resulting oil
was taken up in CH2Cl2 (15 mL) and washed with 1 N HCl (3
¥ 5 mL). 1 N NaOH (20 mL) was added to the aqueous layer
and washed with CH2Cl2 (4 ¥ 10 mL). The organic layers were
combined, dried over MgSO4, filtered, and solvent removal gave
an oil that was purified by HPLC, giving a white solid (0.04 g,
25%). 1H NMR (CD3)2CO): d 10.38 (1H, d, J = 5.73), 90.2 (1H, d,
J = 4.65) 8.95 (1H, d, J = 4.40), 8.47 (1H, d, J = 7.75), 8.41 (2H, dd,
J = 4.74, J = 12.32), 8.27 (1H, d, J = 7.26), 8.16 (2H, m), 8.03 (2H,
m), 7.96 (1H, t, J = 7.24, J = 7.24), 7.67 (1H, dd, J = 5.17, J = 6.92),
6.23 (1H, dt, J = 4.52, J = 7.82, J = 7.85), 4.44 (1H, d, J = 14.34),
4.35 (1H, d, J = 14.34), 2.76 (1H, m), 2.67 (1H, m), 2.52 (1H, td,
J = 2.14, J = 2.14, J = 4.25). 13C NMR (CD3OD): d 150.69, 149.32,
148.85, 139.16, 137.86, 136.59, 133.57, 129.29, 128.34, 126.78,
124.60, 122.43, 122.32, 122.12, 55.52, 49.32, 46.58, 34.88; m/z
(HR-FAB) 392.1726 [M+H]+, C21H22N5O3 requires 392.1723.
100 Photochemical Reactor at 350 nm for 15 s. Photodegradation
of the caged compounds was monitored by LC-MS analysis.
Aliquots of 3.0 mL of each sample before and after photolysis were
injected by autoinjector and run in triplicate. The experiment was
repeated to ensure reproducibility. Comparison of the integrated
peak areas indicated that after 15 s of UV exposure, 43.3% of the
caged Pi, 37.9% of 3Gcage, and 43.3% of [Cu(3Gcage)]+ had been
photolyzed. The quantum yield for each sample (Usample) was
calculated by using the following equation (eqn (1)):
%Dsample
%Dcp
A350cp
Fsample = Fcp×
×
(1)
A350sample
where Ucp is 0.54, the reference quantum yield of photolysis for
caged Pi, %Dsample and %Dcp are the percent change in integrated
peak area after photolysis for the sample and caged Pi, respectively,
and A350 is the absorbance at 350 nm in a 1-mm cuvette for
caged Pi (cp) and for the sample. The calculated quantum yields
of photolysis for 3Gcage and [Cu(3Gcage)]+ are 0.66 and 0.43,
respectively.
Deoxyribose assay for hydroxyl radical production
The 2-deoxyribose assay was used to measure hydroxyl radical
formation.26 A mixture of copper, ascorbic acid, and hydrogen
peroxide generates hydroxyl radicals by Fenton-like chemistry.
Hydroxyl radicals attack 2-deoxyribose to form malondialdehyde
(MDA), which upon heating with TBA under acidic conditions
produces a pink chromophore. Chelators that prevent copper from
reacting with ascorbic acid and hydrogen peroxide result in less
chromophore formation. Assays were performed in a 96-well plate
using 20 mM NaH2PO4 buffered to pH 7.4. The following reagents
were added sequentially to obtain a 100 mL buffered solution with
these final concentrations: chelator (10 mM), CuSO4 (10 mM), 2-
deoxyribose (15 mM), H2O2 (100 mM), and ascorbic acid (2 mM).
For photolyzed samples, 100 mM solutions of [Cu(OH2)(cage)] or
[Cu(3Gcage)]+ in 20 mM pH 7.4 NaH2PO4 buffer were photolyzed
in screwtop quartz cuvettes for 4 min in the photoreactor, then
immediately diluted into the deoxyribose reaction mixtures to
obtain final concentrations of 10 mM in Cu. Stock solutions of
CuSO4, ascorbic acid, and H2O2 were prepared fresh daily, other
solutions werepreparedweekly. Thereactionmixtures wereshaken
in a plate reader at 37 ◦C for 1 h, then 100 mL of TBA (1% w/v in
50 mM NaOH) and 100 mL of TCA (trichloroacetic acid) (2.8%
w/v◦in water) were added. The 96-well plate was placed over a
100 C water bath for 20 min, then cooled to room temperature
and the absorbance at 490 nm was recorded. Values are reported
as A/A0 where A0 is the absorbance without chelator present and
A is the absorbance with chelator added. The value for CuSO4
alone is A/A0 = 1. Error bars represent standard deviations from
measurements done in at least triplicate.
2-(2-Nitro-phenyl)-N-pyridin-2-ylmethyl-2-[(pyridin-2-ylmethyl)-
amino]-acetamide (1Gcage)
2-bromo-2-(2-nitrophenyl)acetic acid (BNPA) was synthesized
using a procedure from Bayley et al.24 Equimolar quantities
of BNPA (0.100 g, 0.385 mmol) and 2-(aminomethyl)pyridine
(0.04 mL, 0.385 mmol) in CH2Cl2 (3 mL) were added to a 50-
mL round-bottom flask equipped with a stir bar and the reaction
mixture was allowed to stir for 2 h at room temperature. After
2 h, NMM (0.085 mL, 0.770 mmol) and 2-(aminomethyl)pyridine
(0.04 mL, 0.385 mmol), both in CH2Cl2 (6 mL) were added to the
reaction mixture. PyBOP (0.200 g, 0.385 mmol) in CH2Cl2 (4 mL)
was then added, and the reaction mixture stirred for 18 h at room
temperature. The solvent was removed and the resulting oil was
taken up in CH2Cl2 (25 mL), filtered, and washed with saturated
aqueous NaCl solution (3 ¥ 10 mL). The organic layers were
combined, dried over MgSO4, filtered, and solvent removal gave an
oil that was purified by chromatography (silica, EtOAc:hexanes,
3 : 1 with 1% triethylamine, Rf 0.30), giving a white crystalline solid
(0.06 g, 42%). 1H NMR (CD3OD): d 8.51 (1H, m), 8.45 (1H, m),
7.96 (1H, dd, J = 1.27, J = 8.16), 7.58 (4H, m), 7.42 (1H, ddd,
J = 1.68, J = 7.26, J = 8.20), 7.33 (1H, d, J = 7.83), 7.26 (1H, d,
J = 7.85), 7.14 (2H, m), 4.72 (1H, s), 4.58 (2H, dq, J = 5.45, J =
15.95, J = 15.97, J = 15.97), 3.92 (2H, q, J = 14.01). 13C NMR:
d 170.67, 158.63, 156.89, 149.18, 149.03, 148.88, 136.78, 136.64,
133.83, 133.43, 132.02, 128.79, 125.31, 122.36, 122.30, 122.25,
122.00, 63.54, 53.27, 44.67; m/z (EI) 378.1 [M+H]+, C20H20N5O3
requires 378.16.
Quantum yield of photolysis
General spectroscopic methods for determining apparent binding
constants
The quantum yields of photolysis for 3Gcage and its Cu(II)
complex were determined by comparison to the quantum yield of
1-(o-nitrophenyl)ethyl phosphate (caged Pi) as previously reported
by Ellis-Davis and Kaplan.25 Samples of 3Gcage (500 mM) with
and without one equiv of CuClO4, or caged Pi (1 mM) in
NaH2PO4 buffered to pH 7.4 were irradiated in a Rayonet RPR-
Solutions were prepared in 50 mM HEPES (N-(2-hydroxyethyl)-
piperazine-N¢-2-ethanesulfonic acid) buffered to either pH 7 or
7.4. The reaction vessel was a 3-mL quartz cuvette and initial
solution volumes were 1 mL. All titrations were carried out at
25 ◦C. Aliquots were pipetted into the reaction mixture and
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9538–9546 | 9541
©

View Article Online
monitored spectrophotometrically. After each addition, solutions
were manually mixed and equilibrated for 10 min before data were
collected. Titrations were repeated in at least triplicate to ensure
reproducibility. Association constants of copper or zinc ion with
Cage compounds (Cage) were calculated assuming 1 : 1 binding of
Cage and metal, M, in the presence of a competitive chelator, L.
The apparent Cu2+ and Zn2+ binding constants of Cage com-
pounds were determined from competition experiments between
each Cage ligand and a ligand (L) of known affinity for either
Cu2+ or Zn2+ at a selected pH. The equilibrium expressions for
the chelation of M by either Cage or L, shown in eqn (2) and
(3), give rise to the formation constants for each ligand (eqn
(4) and eqn (5)). An equilibrium constant for the exchange
reaction (Kex, eqn (6), (7)) of metal between Cage and L can be
calculated based on the known concentrations of total M, Cage,
and L added to solution and from the concentration of metal
cage complex, [M(Cage)], and competitive ligand metal complex,
[M(L)], determined from the absorbance spectra. Values for Kex
were calculated manually by using eqn (7) and data obtained
from at least 10 points along a titration where an equilibrium
between [M(Cage)] and [M(L)] was established and concentrations
of both species could be accurately determined. Titrations were
done in triplicate and errors are reported as standard deviation.
The apparent binding constant is then obtained by solving eqn
(8) for K¢[M(Cage)], where K¢[M(L)n] is the known conditional binding
constant of L at the selected pH.
Competition study of ethylenediaminetetraacetic acid (EDTA) vs.
3Gcage for Cu2+
Solutions of [Cu(3Gcage)]+ with an initial concentration of
0.2 mM were prepared in 10 mM HEPES buffered to pH 7.4.
Aliquots of 4 mL of the competitive chelator EDTA (5 mM) were
pipetted into [Cu(3Gcage)]+ solutions and monitored spectropho-
tometrically. The apparent copper binding affinity for 3Gcage,
K¢[Cu(3Gcage)], was calculated using eqn (2)–(8) and a log K¢[Cu(EDTA)]
15.81 at pH 7.40.
=
Competition study of 4-(2-pyridylazo)resorcinol (PAR) vs. Cage
for Zn2+
The Zn2+ apparent binding constants of Cage ligands was deter-
mined by competition with PAR.23 All titrations were performed
in 10 mM HEPES at pH 7. The spectrum of PAR was recorded as
the background. Upon the addition of Zn2+ a new peak forms at
500 nm due to the formation of the [Zn(PAR)2] complex. To ensure
that the 2 : 1 PAR:Zn2+ complex is formed over the 1 : 1 complex an
excess (>20 equivalents) of PAR was used. Cage ligands were then
titrated into solution and the absorbance spectra were recorded
after each addition. The apparent binding constant was obtained
by solving eqn (7) and (8) for K¢[Zn(Cage)] where log K[Zn(PAR)2] = 12.34.
Results and discussion
Cu2+ binding studies
M + nL ꢀ [M(L)n]
M + Cage ꢀ [M(Cage)]
[M(L)n ]
(2)
A limitation of H2cage, our first generation copper cage, is its 16
pM dissociation constant for Cu2+ at physiological pH,19 which is
likely not tight enough to keep copper sequestered in the presence
of endogenous copper binding proteins. One of the primary goals
of the current study was therefore to develop a higher affinity
compound while maintaining all of the beneficial properties of
H2cage, like its high quantum efficiency and post-photolysis
reactivity. The copper complex of H2cage, [Cu(OH2)(cage)], has
a distorted trigonal bipyramidal geometry. This distortion in
geometry is attributed to a steric interaction between the a
hydrogens on the pyridyl rings. We hypothesized that alleviating
this steric clash to allow a favorable square planar arrangement
might improve the binding affinity for Cu2+. Therefore, one of the
pyridyl rings (labeled 1 and 4 in Fig. 2) was replaced with either a
primary amine to give Amcage or an imidazole to give Imcage.
(3)
K[M(L)
=
(4)
]
[M][L]n
n
[M(Cage)]
[M][Cage]
K[M(cage)]
=
(5)
(6)
[M(Cage)] + nL ꢀ [M(L)n] + Cage
[M(L)n ][Cage]
Kex
=
(7)
(8)
[L]n[M(Cage)]
′
K[M(L)
]
n
Kex
=
′
K[M(Cage)]
Competition study of nitrilotriacetic acid (NTA) vs. caged
compounds for Cu2+
Fig. 2 Nitrogen numbering scheme for H2cage.
Titrations to determine the apparent Cage copper affinity,
K¢[Cu(Cage)], were performed with NTA (log K[Cu(NTA)2] = 10.68 at
pH 7.4) as a competitive chelator for Cu2+. Solutions of 0.5–1 mM
[Cu(Cage)] were titrated with at least two equivalents of NTA
using a 100 mM NTA stock solution.
To investigate the binding affinity of the cage analogs, a
competition experiment was performed between the cage ligand
and the common chelator nitrilotriacetic acid (NTA), which at
pH 7.4 has a Kd of 23 pM. Since our goal was to identify
9542 | Dalton Trans., 2010, 39, 9538–9546
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
compounds with better affinity than H2cage, binding constants
were only calculated for ligands that could at least compete with
NTA for Cu2+. Imcage falls into the category of ligands unable to
compete with NTA for Cu2+. On the other hand, Amcage is able
to compete with NTA for copper, but has a Kd of 200 pM that is
still weaker than H2cage.
The lackluster binding of Amcage and Imcage is likely due to
both electronic and steric effects. The conversion of nitrogen 1 to
a primary amine may alleviate the steric hindrance between the a
hydrogens on the pyridyl rings of H2cage. However, an aliphatic
amine has a higher pKa than a pyridyl nitrogen, which will decrease
its effective metal-binding ability at acidic and neutral pH values,27
therefore this conversion may contribute to Amcage’s decreased
binding affinity as compared to H2cage at pH 7.4. In H2cage,
nitrogens 1 and 2 form a 5 membered ring, 2 and 3 form a 6
membered ring, and 3 and 4 form a 5 membered ring with copper.
This 5-6-5 ring system has been shown to favor Cu2+ binding.28,29
Imcage has a 5-6-6 ring system that is less favorable than the 5-6-5
ring system for Cu2+ binding.28,29
Manipulations of the pyridyl nitrogens to form Imcage and
Amcage were found to be unfavorable for Cu2+ binding. Since
[Cu(OH2)(cage)] has a water molecule providing a fifth donor
atom, the incorporation of additional donor atoms to the ligand
backbone may be beneficial for copper binding. Therefore, we
decided to investigate the effect of adding an additional pyridyl
ring to nitrogen 3 (Fig. 2) to form 3arm-3 and two pyridyl rings to
the primary amine on Amcage to give 3arm-1. In H2cage, deproto-
nation of amide 3 is needed to stabilize copper in its tetradentate,
dianionic binding pocket. Substitution at this site occupies the
electrons that were once donated to copper, and also provides
a monoanionic binding pocket. As a result, 3arm-3 is unable to
compete with NTA for copper, establishing that substitution at
this site is not favorable for copper binding. The incorporation
of two picolyl arms to Amcage produces a potentially dianionic
ligand with six donor atoms: three pyridyl nitrogens, two amides,
and a tertiary amine. Unfortunately, 3arm-1 was also found to be a
weaker copper chelator than NTA. Therefore, we determined that
all of the modifications we made to our ring system hinder copper
binding more than the steric interaction of the a hydrogens on the
pyridyl rings of H2cage.
based on carboxylate-rich binding domains that are not ideal for
our system since copper in this coordination environment can
act as a pro-oxidant and photolysis releases CO2 and carbon
centered radicals.31 While there are few examples of photocaged
d-metal complexes, Burdette and coworkers recently published
ZinCleav-1 as a photolabile Zn2+ chelator.23 This tetradentate
ligand contains two tertiary amines and two pyridyl amines, as
shown in Fig. 1. We found that ZinCleav-1 binds copper rather
tightly since NTA could not compete with ZinCleav-1 for copper.
ZinCleav-1 is built on the framework of EBAP (ethylene-bis-a,a¢-
(2-aminomethyl)pyridine), which has a reported absolute Kd for
Cu2+ of 0.05 fM.27 These examples suggest that replacing the
amides in H2cage with amines could improve copper binding. Any
improvement in copper binding, however, needs to be evaluated
along with other implications these modifications have on other
properties of the complex. For example, the copper complex of
ZinCleav-1, which only contains neutral donor atoms, has a 2+
charge overall that may limit its ability to diffuse across cell
membranes. The photolytic efficiency, as discussed further below,
is another important parameter that may be altered by ligand
modification.
Zn2+ binding studies
While most of the cage analogs show feeble binding properties for
Cu2+, they were also screened for their ability to coordinate Zn2+.
The Zn2+ apparent binding constants were measured by titrating
solutions of [Zn(PAR)2], where PAR is 4-(2-pyridylazo)resorcinol,
with various cage derivatives. The Zn2+ complexes of the cages
lack a spectroscopic signal, therefore by monitoring the decrease
at 500 nm of the [Zn(PAR)2] complex upon the addition of the
cage chelators, apparent binding constants can be obtained.
Table 1 shows the apparent binding constants calculated using
the PAR titration data. While none of our derivates bind zinc
as strongly as ZinCleav-1, which has a Kd¢ of 1.5 pM,23 3arm-
1 has the highest affinity among the series for Zn2+ with a Kd¢ of
0.92 nM. However, to keep zinc sequestered in a biological system,
a binding constant of at least pM is required.32
Photochemistry
Due to the fact that macrocyclic ligands are known to be more
stable than their linear counterparts, we synthesized macrocage in
an attempt at improving the Cu2+ binding affinity of the ligand
backbone. Macrocage and its copper complex are unfortunately
only soluble in DMF; very minimal amounts of water can be
added before precipitation occurs. For this reason, competition
experiments with NTA were not performed. A similar compound
to macrocage that lacks the nitrophenyl groups has been reported
by Margerum and coworkers.30 A pH higher than 13 is needed to
form the macrocyclic copper complex, but EPR analysis showed
that copper coordinates to four deprotonated amides and is
stable down to pH 8. However, at physiological pH this same
coordination is not observed, causing a decrease in binding affinity
and a higher probability of ligand exchange reactions. Due to the
lack of water solubility and incompatibility with physiological
pH, macrocage will not be an affective copper chelator for our
purposes.
Ortho-nitrophenyl groups have been used in countless examples
for caging biological molecules by covalent modification of a
key functional group required for biological activity. The caged
molecule is thereby biologically inactive, but exposure to UV light
Table 1 Apparent binding constants (Kd¢) for Cu2+ and Zn2+ and
photolysis quantum yields for apo ligands (Ucage) and their metal complexes
(U[M(cage)]) of H2cage analogs, where M = Cu2+ for H2cage and 3Gcage
and M = Zn2+ for ZinCleav-1. Values that are not available are labeled
“na”
Kd¢ for Cu2+ at
pH 7.4 fM
Kd¢ for Zn2+ at
pH 7 nM
Ucage
U[M(cage)]
H2cage19
Imcage
Amcage
3arm-3
1.60 ¥ 104
na
20.93 1.76
13.69 1.58
21.57 2.02
12.55 0.81
0.92 0.16
na
0.73
na
na
na
na
0.66
0.024
0.32
na
na
na
na
0.43
0.0055
2.00 ¥ 105
na
3arm-1
na
3Gcage
0.18 0.03
0.05
The most well known ligands for photocaging metal ions are
those used to cage calcium; however, these chelating structures are
ZinCleav-123
0.0015
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9538–9546 | 9543
©

View Article Online
releases the photolabile protecting group to restore biological
activity. Like Ca2+ cages, our Cu2+ cages are designed so that a
nitrophenyl group is incorporated into the ligand backbone of
a high affinity copper chelator such that bond cleavage induced
by UV light releases bidentate compounds with a low affinity for
copper, as shown in Scheme 1. In this case the nitrophenyl group
is incorporated to control the spatial and temporal release of the
metal cargo.
An important quality of a photolabile group is its quantum
efficiency. Because nitrophenyl groups are activated by near-UV
light (350 nm), it is important to have high quantum efficiency to
keep exposure times short and minimize cellular damage. One way
found to improve quantum efficiency in calcium cages is to increase
the extinction coefficient at 350 nm by incorporating dimethoxy
substituents at the four and five position of the ortho-nitrophenyl
group.33 Increasing the photons absorbed at the photolysis wave-
length maximizes photolytic efficiency, meaning less light is needed
to photolyze an equivalent amount of compound.
Our findings above show that replacing the amides with amines
may be beneficial with respect to Cu(II) affinity. However, a
comparison of the quantum efficiencies of H2cage (0.73 and 0.32
for apo and Cu-bound, respectively) vs. ZinCleav-1 (0.024 and
0.0055 for apo and Zn-bound, respectively) suggests that such an
alteration has a negative impact on photolysis efficiency. H2cage is
thirty times more efficient than ZinCleav-1 in the absence of metal,
and nearly sixty times more efficient in the metal-bound form, even
though ZinCleav-1 has the more favorable dimethoxy-substituted
ortho-nitrophenyl group (Fig. 1). This difference in efficiency may
be due to the fact that the amide at position 2 in H2cage (see Fig. 2)
is a better leaving group than the secondary amine of ZinCleav-
1. To test this hypothesis, we used 1Gcage, a ligand similar to
H2cage, but with a primary amine as the leaving group at position
2 instead of an amide. We found that bond cleavage does not occur
with 1Gcage even after hours of UV exposure whereas H2cage is
completely photolyzed in 4 min (data not shown). This result
further supports our hypothesis that the leaving group plays a
major role in the cleavage efficiency of the ligand and indicates
that converting both amide nitrogens to amines is not an ideal
strategy for improving copper cages.
Fig. 3 Titration of 4 mL aliquots of 5 mM EDTA into 0.200 mM solution
of [Cu(3Gcage)]+ in HEPES buffer, pH 7.4.
Kd¢ for 3Gcage of 0.18 0.03 fM at pH 7.4, an improvement
over H2cage’s Kd¢ of 16 pM by more than 4 orders of magnitude.
In addition, 3Gcage shows a selectivity for Cu2+ over Zn2+, as
revealed by a lack of any spectral change in the copper d-d band
of [Cu(3Gcage)]+ when one equiv of Cu2+ is incubated with 3Gcage
in the presence of 50-fold excess Zn2+ for over 2 h (data not shown).
To analyze the photolysis of our 3rd generation cage a 100 mM
[Cu(3Gcage)]+ solution in pH 7.4 phosphate buffer was irradiated
with UV light centered at 350 nm. As shown in Fig. 4, changes
in the UV-vis spectra are apparent within seconds and a shift
in the d-d band centered at 570 nm indicates a reorganization
of the coordination sphere. Cleavage of the ligand backbone is
complete within 3 min, as confirmed by LC-MS analysis (data
not shown). The chromatographic peak for the intact Cu complex
disappears and is replaced by peaks corresponding to the expected
photoproducts (see Scheme 1) and their copper-bound forms, as
confirmed by their mass spectra. The quantum yield of photolysis
for 3Gcage and its copper complex, 0.66 and 0.43 respectively,
is similar to that of H2cage (0.73) and [Cu(OH2)(cage)] (0.32).
Combined, these results confirm that changing the amide that is
not involved in the photolysis mechanism (position 3 in Fig. 2) to
a secondary amine does not hinder photolysis efficiency but does
significantly improve copper binding.
In order to assess how the reactivity of [Cu(3Gcage)]+ is affected,
we used the deoxyribose assay to monitor the ability of the
complex pre- and post-photolysis to generate hydroxyl radicals
in the presence of H2O2 and a reductant. Fenton-like conditions
propagate hydroxyl radical formation when Cu2+ is reduced by
ascorbic acid and reacts with H2O2 to degrade deoxyribose. The
degraded deoxyribose then reacts with thiobarbituric acid to af-
ford a pink chromaphore that is measured spectrophotometrically.
Ligands added to the reaction mixture that prevent the reduction
of Cu2+ attenuate the formation of hydroxyl radicals and in turn
chromophore formation. As shown in Fig. 5, 3Gcage provides
70% protection of hydroxyl radical formation as compared to
Cu2+ alone. This outcome is an improvement over H2cage, which
3rd generation cage
While it seems unfortunate that none of the derivatives described
above bind copper more tightly than the parent H2cage at
physiological pH, several key results were obtained: replacing one
of the pyridyl nitrogens, altering the 5-6-5 chelate ring system, or
substituting at nitrogen 3 (Fig. 2) are all negative design factors
with respect to copper binding. The most significant improvement
for Cu2+ binding is to replace the amides for amines, but this
change dramatically decreases photolysis efficiency. In order to
improve copper binding while maintaining photolysis efficiency,
we therefore designed the third-generation chelator, 3Gcage, with
an amide as the photolysis leaving group at position 2 and a
secondary amine at position 3.
Initial assessment of Cu2+ binding by 3Gcage with NTA
indicated that NTA was too weak of a competitive inhibitor;
therefore, the apparent binding constant was obtained by titration
with EDTA. Analysis of the spectral changes shown in Fig. 3
as EDTA is titrated into a solution of [Cu(3Gcage)]+ provides a
9544 | Dalton Trans., 2010, 39, 9538–9546
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Fig. 4 UV-vis spectra of 100 mM [Cu(3Gcage)]+ irradiated with UV light in 10 s intervals for a total of 180 s. a) An increase in absorbance at 320 nm
is consistent with the formation of a nitroso photoproduct. Inset: A plot of the absorbance at 320 nm vs. UV exposure indicates that the photolysis of
[Cu(3Gcage)]+ is complete in approximately 60 s. b) Changes in the d-d band centered at 570 nm indicate a change in the copper coordination sphere
upon UV exposure.
indicating that the bidentate chelators produced post-photolysis
improve the catalytic properties of the metal with respect to
Fenton-like chemistry.
Conclusion
In conclusion, an analysis of seven new analogs of nitrophenyl-
containing ligands reveals that 3Gcage is the most promising
photoactive ligand for Cu2+ to date. 3Gcage coordinates copper
in a tetradentate binding site that includes two pyridyl nitrogens,
one amide nitrogen, and one amine. The placement of the amide
as the leaving group for nitrophenyl photoactivation was found
to be a key feature for achieving a high quantum efficiency
for photolytic activity. The ability of copper to participate in
Fenton-like chemistry and catalyze the production of hydroxyl
radicals from H2O2 and ascorbic acid was analyzed by using
the deoxyribose assay. In the dark, 3Gcage inhibits hydroxyl
radical formation, on the other hand, after exposure to UV light
hydroxyl radical formation increases by more than 300%. The
0.18 fM copper binding affinity of our 3rd generation chelator
Fig. 5 Results of the deoxyribose assay for OH∑ production, where A and
is a great improvement from our first-generation H2cage chelator
A0 are the absorbance with and without ligand, so A/A0 = 1 for CuSO4
and should be strong enough to keep copper sequestered in the
alone; lower values indicate an inhibitory effect and higher values indicate a
presence of endogenous copper binding proteins. Our new caged
promotional effect of the ligand with respect to copper’s reactivity for OH∑
copper complex has a low charge (+1) and a molecular weight
less than 500 g mol-1, which may be useful for cellular application
and cell permeability. This is a promising step toward developing
photolabile compounds that can deliver metal ions in a spatial
and temporal manner to induce oxidative stress as a potential
chemotherapy strategy or to study metal trafficking pathways.
Future work will be aimed at expanding our 3rd generation
chelator library as well as investigating the biological application
of these compounds.
production. Prior to photolysis, [Cu(3Gcage)]+ inhibits OH∑ production
better than [Cu(cage)], whereas both compounds promote OH∑ production
at similar levels after 4 min of UV photolysis (indicated by *).
only provides 50% protection. The photoproducts of 3Gcage, on
the other hand, increase hydroxyl radical formation above the
level observed for Cu2+ alone. The reactivity of the photoproducts
of [Cu(3Gcage)]+ matches that of photolyzed [Cu(OH2)(cage)],
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9538–9546 | 9545
©

View Article Online
15 J. P. Fruehauf and F. L. Meyskens, Jr, Clin. Cancer Res., 2007, 13,
789–794.
Acknowledgements
16 S. Pervaiz, Curr. Pharm. Des., 2006, 12, 4469–4477.
17 A. Gupte, S. Wadhwa and R. J. Mumper, Bioconjugate Chem., 2008,
19, 1382–1388.
18 D. Z. Cen, D. Brayton, B. Shahandeh, F. L. Meyskens and P. J. Farmer,
J. Med. Chem., 2004, 47, 6914–6920.
19 K. L. Ciesienski, K. L. Haas, M. G. Dickens, Y. T. Tesema and K. J.
Franz, J. Am. Chem. Soc., 2008, 130, 12246–12247.
20 K. L. Ciesienski, L. M. Hyman, S. Derisavifard and K. J. Franz, Inorg.
Chem., 2010, 49, 6808–6810.
21 K. L. Ciesienski, L. M. Hyman, D. T. Yang, K. L. Haas, M. G. Dickens,
R. J. Holbrook and K. J. Franz, Eur. J. Inorg. Chem., 2010, 2224–2228.
22 M. Rozga, M. Sokolowska, A. M. Protas and W. Bal, J. Biol. Inorg.
Chem., 2007, 12, 913–918.
We thank the US National Science Foundation (CAREER
0449699), the Sloan Foundation, and the Camille and Henry
Dreyfus Foundation for funding various aspects of this work. KLC
acknowledges Kathleen Zielek and Burroughs Welcome Award
through the Department of Chemistry at Duke.
References
1 M. M. O. Pen˜a, J. Lee and D. J. Thiele, J. Nutrition, 1999, 1251–1260.
2 B.-E. Kim, T. Nevitt and D. J. Thiele, Nat. Chem. Biol., 2008, 4, 176–
185.
23 H. M. D. Bandara, D. P. Kennedy, E. Akin, C. D. Incarvito and S. C.
Burdette, Inorg. Chem., 2009, 48, 8445–8455.
24 C. Y. Chang, B. Niblack, B. Walker and H. Bayley, Chem. Biol., 1995,
2, 391–400.
25 G. C. R. Ellis-Davies and J. H. Kaplan, Proc. Natl. Acad. Sci. U. S. A.,
1994, 91, 187–191.
26 B. Halliwell, J. M. C. Gutteridge and O. I. Aruoma, Anal. Biochem.,
1987, 165, 215–219.
27 R. G. Lacoste and A. E. Martell, Inorg. Chem., 1964, 3, 881–884.
28 A. E. Martell, R. D. Hancock and R. J. Motekaitis, Coord. Chem. Rev.,
1994, 133, 39–65.
29 A. E. Martell and R. D. Hancock, Metal Complexes in Aqueous
Solution, Plenum Press, New York, 1996.
30 J. S. Rybka and D. W. Margerum, Inorg. Chem., 1980, 19, 2784–2790.
31 L. Sun, C. H. Wu and B. C. Faust, J. Phys. Chem. A, 1998, 102, 8664–
8672.
3 K. J. Barnham, C. L. Masters and A. I. Bush, Nat. Rev. Drug Discovery,
2004, 3, 205–214.
4 D. R. Brown and H. Kozlowski, Dalton Trans., 2004, 1907–1917.
5 E. Gaggelli, H. Kozlowski, D. Valensin and G. Valensin, Chem. Rev.,
2006, 106, 1995–2044.
6 E. Madsen and J. D. Gitlin, Annu. Rev. Neurosci., 2007, 30, 317–337.
7 F. Arnesano, L. Banci, I. Bertini and S. Ciofi-Baffoni, Eur. J. Inorg.
Chem., 2004, 1583–1593.
8 D. L. Huffman and T. V. O’Halloran, Annu. Rev. Biochem., 2001, 70,
677–701.
9 T. V. O’Halloran and V. C. Culotta, J. Biol. Chem., 2000, 275, 25057–
25060.
10 S. Puig and D. J. Thiele, Curr. Opin. Chem. Biol., 2002, 6, 171–180.
11 A. C. Rosenzweig and T. V. O’Halloran, Curr. Opin. Chem. Biol., 2000,
4, 140–147.
12 M. L. Schlief and J. D. Gitlin, Mol. Neurobiol., 2006, 33, 81–90.
13 A. Gupte and R. J. Mumper, Cancer Treat. Rev., 2009, 35, 32–46.
14 H. Pelicano, D. Carney and P. Huang, Drug Resist. Updates, 2004, 7,
97–110.
32 E. H. Cox and G. L. McLendon, Curr. Opin. Chem. Biol., 2000, 4,
162–165.
33 G. C. R. Ellis-Davies and R. J. Barsotti, Cell Calcium, 2006, 39, 75–83.
9546 | Dalton Trans., 2010, 39, 9538–9546
This journal is
The Royal Society of Chemistry 2010
©