Intracellular Distribution of Fluorescent Copper and Zinc Bis(thiosemicarbazonato) Complexes Measured with Fluorescence Lifetime Spectroscopy

Article abstract of DOI:10.1021/acs.inorgchem.5b01599

The intracellular distribution of fluorescently labeled copper and zinc bis(thiosemicarbazonato) complexes was investigated in M17 neuroblastoma cells and primary cortical neurons with a view to providing insights into the neuroprotective activity of a copper bis(thiosemicarbazonato) complex known as Cu^II(atsm). Time-resolved fluorescence measurements allowed the identification of the Cu^II and Zn^II complexes as well as the free ligand inside the cells by virtue of the distinct fluorescence lifetime of each species. Confocal fluorescent microscopy of cells treated with the fluorescent copper(II)bis(thiosemicarbazonato) complex revealed significant fluorescence associated with cytoplasmic puncta that were identified to be lysosomes in primary cortical neurons and both lipid droplets and lysosomes in M17 neuroblastoma cells. Fluorescence lifetime imaging microscopy confirmed that the fluorescence signal emanating from the lipid droplets could be attributed to the copper(II) complex but also that some degree of loss of the metal ion led to diffuse cytosolic fluorescence that could be attributed to the metal-free ligand. The accumulation of the copper(II) complex in lipid droplets could be relevant to the neuroprotective activity of Cu^II(atsm) in models of amyotrophic lateral sclerosis and Parkinson's disease.

Full text of DOI:10.1021/acs.inorgchem.5b01599

Article
pubs.acs.org/IC
Intracellular Distribution of Fluorescent Copper and Zinc
Bis(thiosemicarbazonato) Complexes Measured with Fluorescence
Lifetime Spectroscopy
James L. Hickey,†,⊥ Janine L. James,‡,§,⊥ Clare A. Henderson,†,∥ Katherine A. Price,‡,§ Alexandra I. Mot,‡,§
Gojko Buncic,† Peter J. Crouch,‡,§ Jonathan M. White,† Anthony R. White,*,‡,§ Trevor A. Smith,*,†,∥
and Paul S. Donnelly*,†
‡
∥
†School of Chemistry and Bio21 Molecular Science and Biotechnology Institute, Department of Pathology, and Ultrafast and
Microspectroscopy Laboratories, School of Chemistry, University of Melbourne, Parkville, Melbourne, Victoria 3010, Australia
§The Florey Institute of Neuroscience and Mental Health, Parkville, Melbourne, Victoria 3052, Australia
S
* Supporting Information
ABSTRACT: The intracellular distribution of fluorescently
labeled copper and zinc bis(thiosemicarbazonato) complexes
was investigated in M17 neuroblastoma cells and primary
cortical neurons with a view to providing insights into the
neuroprotective activity of a copper bis(thiosemicarbazonato)
complex known as CuII(atsm). Time-resolved fluorescence
measurements allowed the identification of the CuII and ZnII
complexes as well as the free ligand inside the cells by virtue of
the distinct fluorescence lifetime of each species. Confocal
fluorescent microscopy of cells treated with the fluorescent copper(II)bis(thiosemicarbazonato) complex revealed significant
fluorescence associated with cytoplasmic puncta that were identified to be lysosomes in primary cortical neurons and both lipid
droplets and lysosomes in M17 neuroblastoma cells. Fluorescence lifetime imaging microscopy confirmed that the fluorescence
signal emanating from the lipid droplets could be attributed to the copper(II) complex but also that some degree of loss of the
metal ion led to diffuse cytosolic fluorescence that could be attributed to the metal-free ligand. The accumulation of the
copper(II) complex in lipid droplets could be relevant to the neuroprotective activity of CuII(atsm) in models of amyotrophic
lateral sclerosis and Parkinson’s disease.
INTRODUCTION
■
Bis(thiosemicarbazones) have a wide range of pharmacological
activity that is linked to their ability to coordinate copper(II)
and zinc(II). The resulting complexes are relatively stable,
neutral, lipophilic, and often capable of crossing cell
membranes. Small modifications to the ligand structure can
result in dramatic changes to the chemistry, allowing some
degree of control of biological activity. Investigations into the
antineoplastic activity of bis(thiosemicarbazones) began in the
1950s when glyoxalbis(thiosemicarbazone) (H2gtsm) was
Figure 1. CuII(gtsm), CuII(ptsm), and CuII(atsm).
shown to inhibit sarcoma 180 tumor growth in Swiss mice
when administered orally in the diet.1 It was recognized that
chelation of copper and zinc to form either CuII(gtsm) (Figure
substituent on the backbone, CuII(ptsm), has been investigated
as tracer of blood perfusion (Figure 1).4,11,12
The presence of two electron-donating methyl groups in
CuII(atsm) stabilizes the CuII complexes with respect to
reduction to CuI. The CuII/CuI reduction potential is E0′ =
−0.59 V for CuII(atsm), E0′ = −0.51 V for CuII(ptsm), and E0′
= −0.43 V for CuII(gtsm) (vs Ag/AgCl).13,14 It is thought that
1) or ZnII(gtsm) could be responsible for the antitumor
activity.2
Sustained interest in the biological activity of metal
complexes of bis(thiosemicarbazones) has led to them being
used as ligands to coordinate copper radionuclides in diagnostic
and therapeutic radiopharmaceuticals.3−5 A copper complex
with two methyl substituents on the backbone of the ligand,
CuII(atsm), where H2atsm = diacetylbis(4-methyl-3-thiosemi-
carbazone), is currently in clinical trials as a hypoxia imaging
agent,6−10 whereas the complex with a single methyl
Received: July 16, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 1. Synthesis of H2L1, CuIIL1, and ZnIIL1
the CuII/CuI reduction potential plays an important role in the
cellular biology, as the CuII complexes of bis-
(thiosemicarbazonato) ligands are stable (KA = 1018) but the
CuI complexes are less stable. It is likely that following
intracellular reduction to CuI the metal ion is susceptible to
transfer to copper metallochaperone proteins, consequently
becoming increasingly bioavailable.15
The intracellular redox-mediated copper-releasing properties
of CuII(gtsm) have been utilized as a unique way to increase
intracellular bioavailable copper in cell and animal models of
relevance to the pathology of Alzheimer’s disease.16−18
CuII(atsm) is also of interest as a potential treatment for
neurodegeneration, as treatment with CuII(atsm) resulted in
improved motor and cognitive function in four different animal
models of Parkinson’s disease19 and improved mouse survival
and locomotor function in animal models of amyotrophic
lateral sclerosis (ALS).20−22
The zinc(II) complexes of bis(thiosemicarbazones), such as
ZnII(atsm), are fluorescent, and this intrinsic fluorescence can
be used to monitor their cellular uptake and intracellular
distribution using fluorescence microscopy.23−27 The copper-
(II) complexes of bis(thisoemicarbazones) derived from
acenaphthenequinone are weakly fluorescent, and fluorescence
microscopy of HeLa cells treated with these complexes revealed
significant accumulation in the external cell membrane and
relatively slow internalization.28 CuII(atsm) is not fluorescent,
so to gain insight into the subcellular distribution of CuII(atsm),
a derivative with a fluorescent pyrene functional group was
prepared, CuII(atsm/a-pyrene). This derivative localized in
lysosomes in HeLa cancer epithelial cells.29 The subcellular
localization of this complex (CuII(atsm/a-pyrene)) was also
investigated in cells of neuronal origin (M17 neuroblastoma),
where the complex accumulated in distinct punctate structures
that partially colocalized with lysosomes and an endoplasmic
reticulum dye (ER Tracker). There was also some evidence that
CuII(atsm/a-pyrene) associated with autophagic structures.30
Pyrene-based probes, however, are not ideal for some
fluorescence imaging applications, particularly of biological
materials, requiring relatively high energy excitation and being
susceptible to excimer formation.
Conventional imaging using confocal fluorescence micros-
copy relies on spatial variation of emission intensity. Most
metal complexes with appended fluorophores have similar
excitation and emission profiles to the corresponding metal-free
ligand. This means that measurement of emission intensity
does not discriminate between the intact metal complex and
metal-free ligand. This contrasts with cellular imaging using
fluorescence lifetime imaging microscopy (FLIM), where
emission lifetimes are measured that are normally independent
of concentrations (in the absence of aggregation effects), and it
is also likely that the emission lifetimes of metal complexes with
appended fluorophores will be distinctly different from the
emission lifetimes of the metal-free ligands.31−34 In this Article
we describe the use of FLIM to elucidate the subcellular
distribution and speciation of a new derivative of H2atsm
tethered to a 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene (BOD-
IPY) fluorophore and its copper and zinc complexes. In general
BODIPY dyes possess reasonably sharp fluorescence emission
and a high fluorescence quantum yield, are reasonably
insensitive to pH and solvent polarity, and are sufficiently
stable to physiological conditions.35−37
During the preparation of this Article independent research
was published38 using an alternative synthetic approach to
prepare different BODIPY derivatives of CuII(atsm) and
CuII(gtsm) where FLIM was used to integrate metal release
properties. This work is differentiated from theirs in that the
BODIPY fluorophore is linked to the bis(thiosemicarbazone)
ligand through an aromatic group rather than a two-carbon
B
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 2. ORTEP (40% probability) representation of H2L. Solvent molecules and selected hydrogen atoms are omitted for clarity.
aliphatic chain but also by extending the study to cells of
relevance to models of neurodegeneration with a view to
providing some insight into the neuroprotective activity of
CuII(atsm) in mouse models of neurodegenerative dis-
eases.19−22 In this Article the cellular uptake and intracellular
distribution of the new BODIPY-labeled ligand and its
copper(II) and zinc(II) complexes was studied in BE(2)-M17
neuroblastoma cells (referred to as “M17” cells henceforth in
this Article), a secondary cell line cloned from the SK-N-BE(2)
cell line, originating from a metastatic neuroblastoma that has
the biochemical properties of neuronal cells when grown in
culture.39 Immortalized secondary cell cultures derived from
cancerous cells such as M17 neuronal-like neuroblastoma cells
may contain oncogenes or a loss of control of cell replication
that can compromise their suitability as models of non-
cancerous neurons; so additional studies in murine primary
cortical neurons grown in culture were also performed.
group to the amine (using H2, Pd/C) followed by reaction with
H2atsm/m2 (Scheme 1).
The 1H NMR spectra of H2L1 display the expected
resonances. The 13C NMR spectrum of H2L1 displays all
resonances as expected including those for the quaternary
thiocarbonyl carbon atoms at δ 178.5 and 176.5 ppm. The 19F
NMR spectrum displays a distorted quartet due to 19F−11B
coupling (spin = 3/2, 80.1% natural abundance) with a
broadened base caused by the overlap of signals associated with
19F−10B coupling (spin = 3, 19.9% natural abundance).
Crystals of H2L1 suitable for X-ray crystallography were
grown from a saturated solution of H2L1 in dimethyl sulfoxide
(Figure 2). In the solid state H2L1 adopts an (E,E)-
configuration about the imine double bonds and an s-trans
(antiperiplanar) conformation about the C(3)−C(4) bond
similar to H2atsm.47 The orthogonal dihedral angle (89.9(9)°)
to the plane of the BODIPY demonstrates the restricted
rotation of the phenyl group resulting from the steric hindrance
between the methyl groups in the 4-position on the pyrrole and
the arene ring. The C−S bond lengths of 1.679(4) and
1.670(3) Å indicate, as expected, more thione- than thiol-like
character in the “free” ligand.
RESULTS AND DISCUSSION
■
Chemical Synthesis. The BODIPY fluorophore was
prepared from 2,4-dimethylpyrrole followed by the addition
of a meso-aryl bridge. The steric hindrance of methyl
substituents restricts rotation of the BODIPY core plane
(typically at 90°) with respect to the aromatic ring. This
electronic decoupling of the π-systems suppresses nonradiative
processes, resulting in narrow absorption and emission bands.
2,4-Dimethylpyrrole was prepared according to literature
methods by nitrating ethyl acetoacetate with sodium nitrite
under acidic conditions, followed by reduction with zinc to
produce a primary amine and condensation with a second
equivalent of ethyl acetoacetate to give 2,4-dimethyl-3,5-
dicarbethoxypyrrole. A global base hydrolysis of the esters
gives 2,4-dimethylpyrrole, separated by steam distillation and
further purified by fractional distillation (Scheme 1).40,41
The atsm-BODIPY ligand, H2L1, was prepared by a high-
yielding transamination reaction where a primary amine
Orange ZnIIL and red-brown CuIIL1 could be prepared by
addition of the appropriate metal acetate salt to H2L in
acetonitrile or dimethyl sulfoxide (Scheme 1). Alternatively, the
Cu complex could be prepared by transmetalation of the Zn
1
complex with the addition of CuII(CH3CO2)2. The H NMR
spectrum of diamagnetic ZnIIL1 in d6-dmso is as expected, with
coordination of the metal ion and the presence of acetate
resulting in double deprotonation of the bis-
(thiosemicarbazone) ligand, as suggested by the absence of
the expected resonances for the imino thiosemicarbazone
protons that are present in the 1H NMR spectrum of the “free”
ligand (although it is not possible to rule out an increase in NH
exchange). Redistribution of electron density throughout the
conjugated chelate results in a significant upfield shift of both
resonances attributed to the SC−NH protons by 0.5 and 1
ppm and a moderate downfield shift of the bridging aromatic
ring resonances by ∼0.25 ppm. The 13C NMR spectrum is of
particular interest due to the absence of several expected
resonances for quaternary thiocarbonyl and imine carbons.
Interestingly, an HMBC experiment identified the imine carbon
atoms coupled to the adjacent methyl protons on the
bis(thiosemicarbazone) backbone. Both complexes revealed a
peak in the ESI mass spectrum corresponding to the expected
m/z value for [MII + H+] with the expected isotope pattern. An
RP-HPLC system coupled with UV/vis detection of CuIIL1 (tR
displaces a tertiary amine on H2atsm/m2 (Scheme 1).42,43
A
BODIPY precursor with a primary amine functional group was
prepared via condensation of 2,4-dimethylpyrrole with p-
nitrobenzaldehyde to give a p-nitrophenyl-substituted dipyrro-
methane followed by oxidation with 2,3-dichloro-5,6-dicyano-p-
benzoquinone (DDQ) and subsequent treatment with BF3·
(OEt)2 under basic conditions, which affords 4,4-difluoro-8-(4-
nitrophenyl)-1,3,5,7-tetramethyl-4-bora-3a,4a-diaza-s-inda-
cene.44,45 Ligand H2L was prepared in 60% yield, in a two-step,
one-pot procedure, by initial reduction of the p-nitro functional
C
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 3. ORTEP (40% probability) representation of ZnIIL1. Solvent molecule and selected hydrogen atoms are omitted for clarity.
Figure 4. ORTEP (40% probability) representation of CuIIL1. Solvent molecules and selected hydrogen atoms are omitted for clarity.
Table 1. Crystallographic Data for H2L1, ZnIIL1, and CuIIL1
H2L1
ZnIIL1
CuIIL1
chemical formula
C30H43BF2N8O2S4
724.77
C30H41BF2N8O2S4Zn
788.13
C32H42BF2N10O2S2Cu
775.23
fw
cryst syst
space group
a/Å
triclinic
triclinic
triclinic
P1
P1
̅
P1
̅
̅
9.6568(10)
10.9113(13)
17.452(2)
93.022(10)
105.926(10)
90.971(9)
1764.9(3)
130(2)
11.0973(17)
13.769(3)
13.847(2)
114.133(17)
111.069(15)
90.039(14)
1774.6(5)
130(2)
7.1471(7)
15.8888(11)
17.5263(15)
66.870(7)
80.516(8)
84.990(7)
1804.7(3)
130(2)
b/Å
c/Å
α/deg
β/deg
γ/deg
V/Å3
T/K
λ/Å
1.5418 (Cu K/α)
2
1.5418 (Cu K/α)
2
1.5418 (Cu K/α)
2
Z
Dc/g cm−3
abs coeff, μ/mm−1
F(000)
size/mm
Tmin/max
2θmax
1.364
1.475
1.427
2.905
3.598
2.404
764
820
808
0.4 × 0.2 × 0.05
0.85938/1
73.32
0.10 × 0.06 × 0.01
0.783/0.961
67.49
0.14 × 0.09 × 0.02
0.791/0.963
67.49
Nt
11 148
12 139
12 392
Nind
6776
6370
6501
Rint
0.0593
0.2313
0.0885
N0(I > 2σ(I))
R1
3904
1806
2609
0.0895
0.2187
0.1332
Rw
0.0475
0.0500
0.0509
GOF on F2
0.839
0.544
0.683
D
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 5. Absorption (a) and fluorescence spectra (λex = 490 nm) (b) of H2L1 in CH3CN when titrated with Zn(NO)3 to give ZnIIL1. Absorbance
(c) and fluorescence spectra (d) of a solution of H2L1 when titrated with Cu(NO)3 to give CuIIL1, and fluorescence spectrum of CuIIL1 acquired
with an increased applied PMT voltage (e).
= 16.49 min) confirmed the purity of the complex, whereas
ZnIIL1 eluted at the same retention as H2L1 with the peak
giving the same electronic spectrum as the ligand (tR = 16.08
min) due to instability in the acidic conditions of the mobile
phase (0.1% trifluoroacetic acid/acetonitrile/water).
(Figure 5a). The ligand is fluorescent (ϕf = 0.29) with an
emission peak at λem = 510 nm (λex = 490 nm) with a sharp
profile and small Stokes shift characteristic of a BODIPY
fluorophore (Figure 5b).
Upon titration of H2L1 with Zn2+, the intensity of the ligand-
based bis(thiosemicarbazone) absorbance collapses (ε = 3.0 ×
104 cm−1 M−1) and evidence of a broad absorption at 420 nm
(ε = 1.6 × 104 cm−1 M−1), characteristic of a bis-
(thiosemicarbazonato)ZnII MLCT band, appears. This is
accompanied by two isosbestic points, at λ = 279 nm and λ
= 369 nm. Addition of Zn2+ results in a decrease in the
fluorescence intensity by ca. 2.5-fold (Figure 5b) to give a
fluorescence quantum yield of ϕf = 0.01 for ZnIIL1.
The molecular structure of ZnIIL1 was confirmed by X-ray
crystallography (Figure 3). The ZnII is five-coordinate square-
pyramidal, with the dianionic tetradentate bis-
(thiosemicarbazone) ligand coordinating through two thio-
late-like sulfur atoms reflected in the C−S bond lengths of
1.783(7) and 1.740(7) Å compared with 1.679(4) and 1.670(3)
Å for H2L1 and two azamethinic nitrogen atoms to give a 5−5−
5 (S, N, N, S) chelate ring system. The ZnII is 0.43(1) Å out of
the plane of the bis(thiosemicarbazonato) donor atoms, with
the fifth coordination site occupied by oxygen from a molecule
of the crystallization solvent, dimethyl sulfoxide. The
orientation of the Zn−N2S2 system is essentially coplanar,
with the arene ring is close to perpendicular to the hindered
BODIPY fluorophore with a dihedral angle of 80.1(9)°, similar
to H21.679(4) and 1.670(3) Å for H2L1.
Titration of Cu2+ into a solution of H2L1 resulted in a
collapse in the absorbance attributed to the bis-
(thiosemicarbazone) fragment and a hypsochromic shift to
309 nm (ε = 3.2 × 104 cm−1 M−1) (Figure 5c). A broad
absorption centered at λ = 365 nm (ε = 2.2 × 104 cm−1 M−1) is
characteristic of MLCT behavior. The copper complex is still
fluorescent, λem = 510 nm (λex = 490 nm), but a significant
quenching of fluorescence is observed (Figure 5d). The
comparatively strong fluorescence of H2L1 allows the use of a
low photomultiplier voltage on the spectrometer, but the weak
fluorescence of CuIIL1 (ϕf = 0.005) is best observed by using a
higher voltage gain (Figure 5e).
Crystals of CuIIL1 were grown from a solution of CuIIL1 in
dimethylformamide. A representation of the molecular
structure of CuIIL1 (Figure 4) reveals the copper binding to
the tetradentate ligand in a similar manner to ZnIIL1, but the
copper(II) is in a four-coordinate distorted square planar
geometry. As seen for ZnIIL1, coordination results in two
thiolate-like sulfur atom C−S bond lengths of 1.764(5) and
1.769(5) Å, which are similar to the C−S bond distances in
CuII(atsm), as are the Cu−N and Cu−S bond lengths.46,47
Electronic Spectroscopy. The absorption and emission
spectroscopic properties of H2L1 in an acetonitrile solution
were evaluated, and two absorption bands are prominent: the
first at λ = 337 nm is attributed to the bis(thiosemicarbazone)
fragment (ε = 6.7 × 104 cm−1 M−1), and the second at λ = 497
nm (ε = 7.3 × 104 cm−1 M−1), with a shoulder at λ = 480 nm (ε
= 2.0 × 104 M), is attributed to the BODIPY fluorophore
Electrochemistry. The electrochemistry of H2L1, ZnIIL1,
and CuIIL1 was investigated by cyclic voltammetry in
dimethylformamide, with a glassy carbon working electrode
against a Ag+/Ag reference electrode (potentials are quoted
versus an internal ferrocinium/ferrocene couple where
E0′1/2(Fc/Fc+) = 0.54 V vs SCE; E0′1/2 = midpoint between a
reversible reductive (Epc) and oxidative (Epa) couple, so E0′1/2
=
(Epc + Epa)/2). All compounds exhibit a one-electron quasi-
reversible process between −1.05 and −1.20 V that can be
attributed to the reduction of the BODIPY fragment to the
radical anion (BODIPY•−).36 The quasi-reversible process at
E
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 6. Cyclic voltammograms (scan rate = 100 mV s−1) for 0.1 mM solutions of compounds of (a) H2L1, (b) ZnIIL1, and (c) CuIIL1 in
dimethylformamide (10 mM tetrabutylammonium tetrafluoroborate). Potentials are quoted relative to ferrocene (Em(Fc/Fc+) = 0.54 V).
E0′1/2 = −0.6 V is attributed to the CuII/I process and is at a
similar potential to CuII(atsm) that occurs at E0′1/2 = −0.62 V
under the same conditions (Figure 6).13−15 This is important
given the correlations between CuII/I reduction potentials and
the biological activity of copper complexes of bis-
(thiosemicarbazones). Oxidative processes close to E = 0.75
V and E = 0.90 V in ZnIIL1 and CuIIL1, respectively, are
possibly due to the formation of the BODIPY radical cation
(BODIPY•+),36 which overlaps with another oxidation process
in the case of CuIIL1 that is due to either a CuIII/II process or
oxidation associated with a ligand-based orbital with π-character
(the analogous oxidative process in CuII(atsm) occurs at E0′1/2
= 0.69 V).48,49
Cellular Uptake of Complexes and Colocalization.
Treatment of both primary cortical neurons and M17
neuroblastoma cells with CuIIL1 (25 μM) at 37 °C was
monitored by live cell scanning confocal fluorescence
microscopy and transmitted light/bright field imaging. There
was no evidence of changes to cell morphology after incubation
for several hours with respect to the control, suggesting that the
complex did not cause observable cell death or any other visual
cell stress. This behavior is consistent with treatments with
comparable concentrations with CuII(atsm). The fluorescence
from CuIIL1 is distributed throughout the cytosol in live cell
M17 neuroblastoma, but significant puncta with high
fluorescence intensity are also present (Figure 7). Colocaliza-
tion studies with Lysotracker showed that some of the punctate
regions of high fluorescence intensity associated with CuIIL1 are
within lysosomes (Figure 7d and e). Most of the bright
punctate staining is colocalized with HCS lipidTOX neutral
deep red lipid stain (shown in blue, Figure 7c), and these
punctate regions could also be visualized through transmitted
light or bright field imaging, with features characteristic of
neutral lipid droplets.50−53 In cell culture, various amounts of
neutral lipids are known to accumulate in lysosomes, and it has
been shown that basal levels of lipids are found within
lysosomes in neurons.54,55
Figure 7. (a) Live cell confocal image of M17 cells treated with CuIIL1
(green, λex = 488 nm), Lysotracker (red, λex = 568 nm), and neutral
lipid stain (blue, λex = 640 nm). Scale bar = 10 μm. (b) Magnified
section (indicated by a white box in part a) showing CuIIL1. (c) CuIIL1
overlay with neutral lipid stain, showing partial colocalization with
CuIIL1 and lipid droplets. (d) CuIIL1 overlay with Lysotracker, showing
partial colocalization with CuIIL1 and lysosomes. (e) Overlay of CuIIL1
(green), Lysotracker (red), and neutral lipid stain (blue). Scale bar = 3
μm.
Investigation of primary cortical neurons revealed apparent
colocalization between Lysotracker and the neutral lipid stain,
with a majority of lysosomes in our primary cortical neuron
culture also staining positive for neutral lipids (Figure 8). These
punctate regions, positive for Lysotracker and neutral lipid
stain, also accumulated CuIIL1 (Figures 8b−e). As with the
M17 cells, CuIIL1 is absent from the cell nucleus, resulting in a
diffuse cytosolic distribution but also distinct bright puncta. In
the primary cortical neurons, all punctate regions were found to
stain positive for Lysotracker (shown in red) (Figure 8c).
These lysosomes also stained positive for the neutral lipid stain
(blue, Figure 8d), and together with the green BODIPY
fluorophore from CuIIL1, the overlay of the three primary
colors shows as overall white puncta (Figure 8e).
F
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Time-Resolved Fluorescence. Fluorescence lifetime
imaging can provide valuable information complementary to
that obtained using conventional confocal fluorescence
intensity-based imaging particularly relating to the molecular
nature of the fluorescent species. The free ligand H2L1, CuIIL1,
and ZnIIL1 emit at similar wavelengths (λem ≈ 510 nm) (Figure
5) despite their differences in quantum yield, so it is not
possible to distinguish between them inside cells by confocal
microscopy. However, both the ZnII and CuII complexes have
dramatically different fluorescence lifetimes when compared to
the free ligand. Time-resolved emission measurements were
carried out, using time-correlated single-photon counting, on
the ligand and the complexes in bulk solution in order to
characterize their fluorescence lifetimes (Figure S1, Supporting
Information). The ligand H2L1 displays a fluorescence decay
profile adequately fitted by a single-exponential decay function,
giving a fluorescence lifetime of 2.83 ns. In comparison, fitting
the fluorescence decay curves of the Zn and Cu complexes
requires double-exponential decay functions. CuIIL1 has a
signature dominant (76%) initial short decay component of τ1
≈ 250 ps along with a relatively minor contribution from a
longer lived decay component of τ2 ≈ 2.79 ns. Similarly, ZnIIL1
has a dominant (83%) short decay component of ∼170 ps and
a minor (17%) long decay component of τ2 ≈ 2.55 ns. These
findings are in general agreement with those of Dilworth et al.
on related but structurally different bis(thiosemicarbazone)
complexes conjugated to a BODPY fluorophore through a two-
carbon aliphatic linker except that for the complexes presented
here, ZnIIL1 and CuIIL1, the longer lifetime component does
not exceed that of the free ligand.38 The dominant short decay
component of the fluorescence of both complexes is attributed
to the quenched emission of the ligand, as its magnitude
relative to the free ligand correlates well with the degree of
Figure 8. (a) Live cell confocal image of primary cortical neurons,
isolated from E14 mice, treated with CuIIL1 (green, λex = 488 nm),
Lysotracker (red, λex = 568 nm), and neutral lipid stain (blue, λex = 640
nm). Scale bar = 10 μm. (b) Magnified section indicated by a white
box in part a showing CuIIL1. (c) CuIIL1 overlay with Lysotracker,
showing partial colocalization with CuIIL1 and lysosomes. (d) CuIIL1
overlay with neutral lipid stain, showing partial colocalization with
CuIIL1 and lipid droplets. (e) Overlay of CuIIL1 (green), Lysotracker
(red), and neutral lipid stain (blue). Scale bar = 3 μm.
Figure 9. Fluorescence lifetime images with color scales mapping the mean fluorescence lifetime for (a) H2L, (b) CuIIL1, and (c) ZnIIL1, respectively,
in M17 cells. (d, e, and f) Corresponding fluorescence lifetime distributions/histograms for the average fluorescence lifetime, τav, τ1, and τ2.
G
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
quenching of fluorescence observed in steady-state spectra
when the coordination complex is formed. The similarity of the
long decay times of the complexes to that of the free ligand’s
single decay process suggests that this component can be
broadly attributed to residual unquenched ligand-based
processes.
Treatment of cells with CuIIL1 results in images with
significantly different behavior compared to cells treated with
H2L1; the τav distribution is far narrower and is dominated by
emission with a lifetime characteristic of the free ligand (∼2.8
ns). This suggests that the CuIIL1 experiences a more uniform
range of cellular environments than H2L1 alone and perhaps
that a proportion of the copper(II) is being released from the
complex upon incubation with the cells. Treatment of M17
cells with ZnIIL1 results in images where the lifetime
distribution is, somewhat surprisingly, dominated by a narrow
peak centered at around 1.4 ns, which is not apparent for ZnIIL1
in solution. A broad shoulder on the distribution is apparent at
around the lifetime reported above for this complex in solution
(2−2.5 ns), again suggesting that a proportion of the Zn is
being lost from the complex during incubation.
Fluorescence Lifetime Imaging Microscopy. Fluores-
cence lifetime imaging microscopy was performed on M17 cells
separately incubated with H2L1, CuIIL1, or ZnIIL1, with the
corresponding emission intensity and lifetime maps shown in
Figure 9a−c, respectively. The experimental acquisition times
required for the FLIM images of the cells containing the metal
complexes were substantially longer than in the case where just
the ligand was present due to the quenching induced by the
metal ion. The images shown represent fluorescence lifetime
maps resulting from double-exponential fitting in each pixel
across the entire image, with the color indicative of the
amplitude-weighted mean fluorescence lifetime, τav, in each
pixel (see color legend). This is in contrast with conventional
confocal images, in which emission intensity is plotted, which
can be influenced by fluorophore concentration and environ-
ment/speciation. We do not associate the discrete fluorescence
lifetimes with two specific emitting species, but an interpreta-
tion of the double-exponential analysis used to generate the
FLIM images is still informative.
Along with each FLIM image is shown the corresponding
lifetime distribution plots for the average fluorescence lifetime,
τav (red traces), and the two individual lifetime components, τ1
and τ2, recovered from the biexponential analysis of the
fluorescence lifetime images of H2L1, CuIIL1, and ZnIIL1
(Figure 9d−f, respectively). The distribution of fluorescence
lifetimes throughout the M17 cells incubated with the
complexes displays the same diffuse cytosolic distribution
characteristic of partial lysosomal/partial lipid droplet local-
ization, as shown using live cell confocal imaging with
Lysotracker and a neutral lipid stain (Figure 7), and these
punctate regions show a higher propensity for the short-lived
copper and zinc complexes, shown as yellow/orange puncta
inside the cells (Figure 9). The diffuse cytosolic fluorescence
tends to be due to emission with a longer lifetime, shown as
diffuse blue/green regions within the cytosol (Figure 9). As
with the confocal images (Figure 7) there is no detectable
fluorescence within the cell nucleus (Figure 9).
For H2L1 alone (Figure 9d) the τav distribution indicates that
any association with metal ions taken up from the cellular
environment must be minimal, as the measurements in bulk
solution show that the emission is quenched significantly (τf ≈
200 ps) when the ligand is fully complexed with Cu or Zn,
whereas the τav distribution does not indicate any subnano-
second lifetime components. However, the τav distributions are
somewhat misleading, being weighted toward the longer lived
components by the generally minor contribution (in terms of
the pre-exponential factors of the biexponential analysis) of
very broad τ2 distributions, generally maximized at ∼3−4 ns,
that again might indicate some loss of the fluorophore from the
ligand.
The most information is gained by concentrating on just the
shorter lived (τ1) components in the fluorescence lifetime
distributions (blue traces in Figure 9d−f). Two peaks are
apparent at ∼1400 and ∼500 ps in the distributions of all three
molecules, and the relative abundance of these distribution
bands is dependent on the complex. The peak at ∼1400 ps
dominates the τ1 distribution in the case of CuIIL1 (Figure 9e)
but is the minor feature in the ZnIIL1 case (Figure 9f), which is
instead dominated by the shorter lifetime contribution (∼500
ps). In the case where the cells have been treated with H2L1
(Figure 9d), the 1400 and ∼500 ps components are roughly
equivalent in their contribution. This short-lived (τ1 ≤ 550 ps)
emission is a characteristic of CuIIL1 and ZnIIL1, so the peaks at
around 500 and 1400 ps are likely to be a consequence of the
free ligand, as expected, coordinating to CuII/ZnII in the cell
media or intracellular environment.
Treatment of the cells with H2L1 (Figure 9a) shows a broad
average lifetime distribution with components that exhibit both
shorter and longer fluorescence lifetimes (Figure 9d) compared
with that characteristic of H2L1 in solution reported above
(∼2.83 ns). There is some evidence of a slight shoulder on the
distribution at around this lifetime. The broadness of this
distribution most likely reflects a wide range of local
environments experienced by H2L1, in agreement with the
findings of Dilworth et al.38 The distribution toward short-lived
decay components is presumably representative of the emission
of the ligand being quenched to some degree probably due to
the presence of trace copper and zinc in cell media and cells
leading to the formation of copper or zinc complexes of L1. The
longer-lived components of the distribution may correspond to
an enhanced emission quantum yield due to a slightly more
rigid environment experienced by the fluorophore or may be
indicative of some degree of metabolic cleavage of the BODIPY
fragment from the rest of the ligand, as dyes of this family
typically have fluorescence lifetimes of several (up to 5)
nanoseconds.
For the cells treated with CuIIL1 the τ1 distribution has a
larger peak around 1400 ps due to the presence of CuIIL1, but
there is also evidence of a peak at around 500 ps consistent with
the presence of ZnIIL1, suggesting that, at least to some degree,
CuIIL1 is releasing copper, most likely through reduction of the
metal to CuI, which is then sequestered by intracellular copper-
binding proteins, and the liberated ligand is now capable of
binding to zinc, leading to the formation of ZnIIL1. Intracellular
zinc concentrations are higher than intracellular copper
concentrations, and CuI is largely bound to high-affinity
chaperone proteins.15
The τ1 component for the ZnIIL1-treated cells shows a large
peak around 500 ps, with a smaller peak around 1400 ps, once
again consistent with metal exchange and that the ZnIIL1
complex zinc can be transmetalated to give CuIIL1. The rapid
exchange of ZnII for CuII in bis(thiosemicarbazone) complexes
has been used in radiolabel bis(thiosemicarbazones) with
radioactive copper isotopes48,56 and has been shown to occur in
cell culture16 and in vivo in a SOD1G37R mouse model of
H
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
familial ALS.22 The formation of copper(II) bis-
(thiosemicarbazonato) complexes by transmetalation of their
ZnII complexes has also been implicated in their antitumor
activity.57
the parent complexes, CuII(atsm) and ZnII(atsm). The
biological activity of CuII(atsm) is at least partially dependent
on the CuII/I reduction potential, and electrochemical measure-
ments using cyclic voltammetry revealed that CuIIL1 possesses a
quasi-reversible reduction at a similar potential to CuII(atsm).
Ligand H2L1 is fluorescent with a narrow emission band
characteristic of BODIPY (λem = 510 nm, λex = 490 nm, ϕf =
0.29). Coordination of either ZnII or CuII to the ligand results
in a significant quenching of fluorescence, but both ZnIIL1 and
CuIIL1 are still fluorescent. The free ligand H2L1, CuIIL1, and
ZnIIL1 all emit at similar wavelengths (λem ≈ 510 nm), so it is
not possible to distinguish between them inside cells by
conventional confocal microscopy. However, ZnIIL1 and CuIIL1
have shorter fluorescence lifetimes than the free ligand H2L1, so
FLIM enables important information on the nature of the
emitting species visible with fluorescent microscopy.
CuIIL1 Accumulates in Lipid Droplets. The FLIM
measurements reveal that the strong fluorescence observed in
the large puncta in cells treated with CuIIL1 are due to the
copper complex (Figure 9b). These puncta appear to be lipid
droplets based on their colocalization with HCS lipidTOX
neutral deep red lipid stain (Figure 7). Lipid droplets are
“cytoplasmic lipid inclusions consisting of a core of neutral
lipids such as triacylglycerols and cholesteryl esters surrounded
by a monolayer of phospholipids and associated proteins”.50
The lipids stored in lipid droplets are used for metabolism,
membrane synthesis, and steroid synthesis. Lipid droplets are
thought to play a fundamental role in intracellular lipid
homeostasis and in storing cholesterol in the form of
cholesterol esters. Assuming that the subcellular localization
of this fluorescently labeled derivative of CuII(atsm) reflects the
cellular distribution of the parent compound, the accumulation
of CuIIL1 within lipid droplets is of potential interest to the
neuroprotective activity of CuII(atsm) in models of ALS and
Parkinson’s disease. Mitochondrial defects in neurons lead to
elevated levels of reactive oxygen species that can elevate lipid
synthesis in neurons and result in the formation of lipid
droplets in glial cells. It is also worth noting that impaired
mitochondrial electron transfer leads to increased cellular
retention of copper in cells treated with CuII(atsm).58,59 The
lipid droplets in glia contribute to neurodegeneration through
elevated levels of lipid peroxidation. In Drosophila and mice
lipid droplets form in glia before or at the onset of
degeneration, and disruption to lipid droplet metabolism can
contribute to neurodegeneration.60 The formation and
metabolism of lipid droplets have also been identified as
playing a role in a subtype of ALS caused by mutations in
human VAMP (vesicle-associated membrane protein).61 The
CuII(atsm) complex possesses both a reductive CuII/I couple
and an oxidative process that has been attributed to either
oxidation of the metal (CuIII/II) or ligand-based electro-
chemistry. The accumulation of redox-active CuII(atsm) in
lipid droplets could partially explain the fact that treatment with
CuII(atsm) resulted in lower levels of lipid peroxidation in a
model of ischemic reperfusion injury62 and could also lead to
less peroxidation of lipids in glia partially contributing to the
neuroprotective activity of the complex in mouse models of
ALS.20 It is also possible that the presence of CuII(atsm) in
lipid droplets could protect these vital lipids from peroxynitrite
(ONOO−)-mediated nitration, but further studies are required
to substantiate such speculation.19 Radiolabeled CuII(atsm) is
currently under evaluation as a hypoxia tracer, and lipid
droplets become an important source of energy for cell
proliferation and may serve a protective role in hypoxic
environments. Therefore, it is interesting to speculate that
changes to lipid droplet metabolism could contribute to
increased accumulation of CuII(atsm) in hypoxic cells.60
Concluding Remarks. A new fluorescent derivative of
H2atsm tethered to a BODIPY fluorophore (H2L1) has been
prepared using a transamination reaction where a primary
amine on a BODIPY fluorophore displaces a tertiary amine on
H2atsm/m2. The ligand and its neutral CuII and ZnII complexes
have been characterized by X-ray crystallography and reveal
that addition of the fluorescent tag does not change the
coordination environment of the metal ion when compared to
Confocal microscopy of neuronal cells treated with CuIIL1
revealed significant fluorescence signal associated with distinct
cytosolic puncta, which were found to be lysosomes that also
stained positive for lipid in primary cortical neurons or a
combination of lysosomes and lipid droplets in secondary M17
neuroblastoma cells. FLIM measurements were able to identify
that treatment of cells with CuIIL1 led to the complex
accumulating in large cytoplasmic inclusions that were
consistent with the lipid droplets identified by confocal
microscopy but also, at least to some extent, release of the
metal ion to give H2L1 and ZnIIL1 presumably due to the
liberated ligand coordinating to bioavailable zinc. Treatment
with ZnIIL1 also resulted in the accumulation of the zinc
complex in cytoplasmic inclusions but also some degree of
transmetalation to form the copper complex. This work adds to
the growing examples where FLIM has provided valuable
complementary information to conventional confocal fluores-
cent microscopy.31−34,38 It would be of interest to apply FLIM
to the characterization of sensors designed to probe the cellular
metabolism of copper as a complementary approach to
ratiometric fluorescent probes.63−66 The biological activity of
CuII(atsm) can be related to the stability of the complex and its
CuII/I reduction potential, and in these respects the fluorescent
derivative, CuIIL1, retains these properties. However, it is
acknowledged that the addition of the BODIPY fluorophore
could result in a copper complex with a different intracellular
distribution from the parent compound. Providing the
intracellular distribution of CuIIL1 accurately reflects the
behavior of CuII(atsm), the colocalization within lipid droplets
could be relevant to the neuroprotective activity of CuII(atsm),
as disruption to lipid droplet metabolism has been identified as
contributing to neurodegeneration. These results also suggest it
is of interest to investigate the significance of the oxidative
redox couple displayed by CuII(atsm) to its biological activity.
Further studies that probe the interplay between treatment of
cells with CuII(atsm), lipid droplet metabolism, and neuro-
protection are warranted.
EXPERIMENTAL SECTION
■
Crystallography. Crystals of H2L1, ZnIIL1, and CuIIL1, respec-
tively, were mounted in low-temperature oil, then flash cooled to 130
K using an Oxford low-temperature device. Intensity data were
collected at 130 K with an Oxford XCalibur X-ray diffractometer with
a Sapphire CCD detector using Cu Kα radiation (graphite crystal
monochromator, λ = 1.541 84 Å). Data were reduced and corrected
for absorption.The structures were solved by direct methods and
difference Fourier synthesis using the SHELX suite of programs as
implemented within the WINGX software.67,68 Thermal ellipsoid plots
I
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
were generated using the program ORTEP-3 integrated within the
WINGX suite of programs.68 CCDC deposition numbers: 1402571,
1402572, and 1402573.
CS), 7.85−7.84 (m, AA′BB′, 2H, ArH), 7.37−7.34 (m, AA′BB′, 2H,
ArH), 6.19 (s, 2H, ArH), 3.04−3.03 (m, 3H, NH-CH3), 2.45 (s, 6H,
ArCH3), 2.30 (s, 3H, NC-CH3), 2.28 (s, 3H, NC-CH3), 1.45 (s,
6H, ArCH3). 13C{1H} NMR (125.7 MHz; DMSO-d6): δ/ppm 178.5
(CS), 176.7 (CS), 154.8 (PyC), 149.6 (CN-N), 147.7 (CN-
N), 142.8 (Ar-C-Py), 141.8 (PyC), 139.9 (ArC), 130.8 (PyC), 130.6
(ArC), 127.7 (ArCH), 125.5 (ArCH), 121.4 (PyCH), 31.2 (NH-CH3),
14.3 (Py-CH3), 14.2 (Py-CH3), 12.2 (NC-CH3), 11.9 (NC-CH3).
19F NMR (471 MHz; DMSO-d6): δ/ppm −141.46 (dq, 1JB,F = 32 Hz,
General Procedures. All reagents and solvents were obtained
from commercial sources (Sigma-Aldrich) and used as received unless
otherwise stated. 2,4-Dimethylpyrrole,40,41 diacetyl-4-methyl-4′-di-
methyl-3-bis(thiosemicarbazone) (atsmm2),42 4,4-difluoro-8-(4-nitro-
phenyl)-1,3,5,7-tetramethyl-4-bora-3a,4a-diaza-s-indacene (p-
NO2BODIPY), and 8-(4-aminophenyl)-4,4-difluoro-1,3,5,7-tetrameth-
yl-4-bora-3a,4a-diaza-s-indacene (p-NH2BODIPY)44,45 were prepared
following published procedures. Elemental analyses for C, H, and N
were carried out by Chemical & MicroAnalytical Services Pty. Ltd.,
Vic. Nuclear magnetic resonance spectra were recorded on a Varian
FT-NMR 500 spectrometer (1H NMR at 499.9 MHz, 13C NMR at
125.7 MHz, and 19F NMR at 471 MHz) at 298 K and referenced to
the internal solvent residue for 1H and 13C and external
hexafluorobenzene (δ −164.9 ppm) for 19F. Mass spectra were
recorded on an Agilent 6510-Q-TOF LC/MS mass spectrometer and
calibrated to internal references.
UV/Visible Spectroscopy. UV/vis spectra were recorded on a Cary
300 Bio UV−vis spectrophotometer, from 800 to 200 at 0.5 nm data
intervals with a 600 nm/min scan rate. Solutions of H2L1 in
acetonitrile (10 μM) were titrated with 6 μL aliquots of standardized
Cu(NO3)2 or Zn(NO3)2 solutions (1 mM in H2O). Successive scans
were performed measuring absorbance every 2 min (800−250 nm).
Fluorescence Spectroscopy. Fluorescence emission spectra were
measured on a Varian Cary Eclipse fluorescence spectrophotometer.
Solutions of H2L1 in acetonitrile (300 nM) were titrated with 2 μL
aliquots of standardized Cu(NO3)2 or Zn(NO3)2 solutions (1 mM in
2F, BF2). MS(ES+): m/z (calcd) 569.2247 (569.2174) {M + H+}.
HPLC: tR 16.08 min. Crystals suitable for single-crystal X-ray
diffraction were grown from dimethyl sulfoxide by slow vapor
diffusion of atmospheric water at room temperature.
Diacetyl-4-methyl-4′-(8-(4-aminophenyl)-4,4-difluoro-
1,3,5,7-tetramethyl-4-bora-3a,4a-diaza-s-indacene)-3-bis-
(thiosemicarbazonato)zinc(II), ZnIIL1. H2L1 (80 mg, 0.14 mmol)
was suspended in acetonitrile (20 mL) and heated to reflux. On
addition of zinc(II) acetate (34 mg, 0.15 mmol), the suspension
rapidly dissolved, and 15−20 min later, an orange precipitate began to
appear. On stirring for a further 2 h the reaction was cooled to room
temperature and the precipitate collected, washed repeatedly with
diethyl ether, and dried in vacuo to give a bright red/orange solid (60
1
mg, 68%). H NMR (500 MHz; DMSO-d6): δ/ppm 9.60 (s, 1H, Ar-
NH-CS), 8.02−8.00 (m, AA′BB′, 2H, ArH), 7.40 (m, 1H, CH3-NH-
CS), 7.21−7.19 (m, AA′BB′, 2H, ArH), 6.17 (s, 2H, ArH), 2.87 (m,
3H, NH-CH3), 2.44 (s, 6H, ArCH3), 2.32 (s, 3H, NC−CH3), 2.25
(s, 3H, NC-CH3), 1.44 (s, 6H, ArCH3). 13C{1H} NMR (125.7
MHz; DMSO-d6): δ/ppm 154.4 (ArC), 149.8 (CN−N), 144.6
(CN-N), 142.7 (ArC), 142.5 (C), 142.0 (ArC), 131.1 (ArC), 127.9
(ArCH), 126.3 (ArC), 121.2 (ArCH), 119.9 (ArCH), 14.9 (NC-
CH3), 14.2 (Ar-CH3), 14.2 (Ar-CH3), 13.8 (NC-CH3). MS(ES+):
m/z (calcd) 631.1391 (631.1309) {M + H+}. HPLC: tR 16.08 min.
Crystals suitable for single-crystal X-ray diffraction were grown from
dimethyl sulfoxide by slow vapor diffusion of atmospheric water at
room temperature.
Diacetyl-4-methyl-4′-(8-(4-aminophenyl)-4,4-difluoro-
1,3,5,7-tetramethyl-4-bora-3a,4a-diaza-s-indacene)-3-bis-
(thiosemicarbazonato)copper(II), CuIIL1. H2L1 (80 mg, 0.14
mmol) was suspended in acetonitrile (20 mL) and heated to reflux.
On addition of zinc(II) acetate (34 mg, 0.17 mmol), the suspension
rapidly dissolved, and 10 min later, copper(II) acetate (37 mg, 0.17
mmol) was charged into the reaction vessel, resulting in rapid
darkening of the solution. On stirring for a further 2 h the reaction was
cooled to room temperature and the precipitate collected, washed
repeatedly with diethyl ether, and dried in vacuo to give a dark red/
brown solid (72 mg, 82%). MS(ES+): m/z (calcd) 630.1399
(630.1314) {M + H+}. HPLC: tR 16.49 min. Crystals suitable for
single-crystal X-ray diffraction were grown from dimethylformamide
by slow vapor diffusion of atmospheric water at room temperature.
Fluorescence Decay Characterization Studies. The excitation
beam (490 nm) was selected, using a prism/slit arrangement from a
Fianium SC450-PP-HE Supercontinuum laser source producing pulses
of ∼100 ps duration and at 2 MHz repetition rate.69 The polarization
of this excitation beam (output power ∼0.5 mW) was tidied up by
passing it through a calcite rhomb polarizer to excite the sample with
vertically polarized light. A monochromator (Jobin Yvon H-10) was
used to isolate the emission, which was detected using a microchannel
plate photomultiplier tube (Eldy model EM1-132-1) coupled to a PC
card-based TCSPC electronics system (Edinburgh Instruments
TCC900).
H2O). Successive scans were performed measuring fluorescence (λex
=
490 nm) emission between 510 and 700 nm.
High-Performance Liquid Chromatography. Analytical RP-HPLC
traces were acquired using an Agilent 1200 series HPLC system
equipped with an Agilent Zorbax Eclipse XDB-C18 column (4.6 × 150
mm, 5 mm) with a 1 mL/min flow rate and UV spectroscopic
detection at 214, 220, and 270 nm. Retention times (tR/min) were
recorded using a gradient elution method of 0−100% B over 25 min;
solution A consisted of water (buffered with 0.1% trifluoroacetic acid),
and solution B consisted of acetonitrile (buffered with 0.1%
trifluoroacetic acid).
Electrochemistry. Cyclic voltammograms were recorded using an
AUTOLAB PGSTAT100 equipped with GPES V4.9 software.
Measurements of the complexes were carried out at approximately 1
× 10−4 M in dimethylformamide with tetrabutylammonium
tetrafluoroborate (1 × 10−2 M) as electrolyte using a glassy carbon
disk (d, 3 mm) working electrode, a Pt wire counter/auxiliary
electrode, and a Ag/Ag+ pseudo reference electrode (silver wire in
H2O (KCl (0.1 M)/AgNO3 (0.01 M))). Ferrocene was used as an
internal reference (Em(Fc/Fc+) = 0.54 V vs SCE), where Em refers to
the midpoint between a reversible reductive (Epc) and oxidative (Epa)
couple, given by Em = (Epc + Epa)/2. Irreversible systems are given only
reductive (Epc) and oxidative (Epa) values, respectively.
Diacetyl-4-methyl-4′-(8-(4-aminophenyl)-4,4-difluoro-
1,3,5,7-tetramethyl-4-bora-3a,4a-diaza-s-indacene)-3-bis-
(thiosemicarbazone), H2L1. p-NO2-BODIPY (550 mg, 1.31 mmol)
was dissolved in EtOH/CH2Cl2 (50:50, 50 mL) in the presence of
10% Pd/C (50 mg) and sparged with N2(g) and then H2(g). The
reaction was stirred at room temperature under H2(g) (1 atm) for 16 h,
after which reduction was complete (TLC: CH2Cl2). The Pd/C was
removed via filtration through Celite into a RB flask, and the Celite
was repeatedly washed with EtOH (3 × 5 mL). Atsmm2 (340 mg, 1.31
mmol) was added to the filtrate, which was degassed and then refluxed
for 16 h. On cooling to room temperature, an orange precipitate was
collected, washed with ether, and air-dried (185 mg). The filtrate was
concentrated by half and refluxed for a further 24 h, forming a second
batch of orange precipitate, which was isolated in the same manner as
the first (254 mg). The second batch was analytically identical to the
FLIM. Using the same laser source as the bulk solution
measurements, the beam was delivered via a single-mode optical
fiber to a modified Olympus confocal scanning microscope (FV300/
IX71) coupled with a Becker & Hickl SPC-830 FLIM module.
Confocally isolated emission was passed through a broad bandpass
filter to block scattered excitation light and detected using a PMC-100
single photon counting photomultiplier, and data were acquired by an
SPC-830 TCSPC card controlled using the SPCM operating software.
The fluorescence decay data were analyzed using “SPCImage” FLIM
data analysis software. Other details are as reported.70
1
first and combined to give a final yield of 60%. H NMR (500 MHz;
DMSO-d6): δ/ppm 10.69 (s, 1H, N-NH-CS), 10.31 (s, 1H, N-NH-
CS), 10.12 (s, 1H, Ar-NH-CS), 8.44−8.41 (m, 1H, CH3-NH-
J
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Cell Culture. All reagents were obtained from Life Technologies
and used as received unless otherwise stated. M17 neuroblastoma cells
were grown in culture, in Opti Mem medium supplemented with fetal
bovine serum (10%), Gibco MEM nonessential amino acids (1%), and
sodium pyruvate (100 μM). Cells were grown in a 37 °C, 5% CO2
incubator to confluency prior to passaging. Cells were detached using
trypsin/EDTA. Cells were seeded at 5 × 104 cells/cm2 into a 35 mm
Ibidi live cell imaging microdish and grown in culture for 24 h prior to
treatment. Primary cortical neurons were harvested from C57 Black 6
mice. Prior to the use of primary murine cell cultures, ethics approval
was obtained from the Biomedical Sciences Animal Ethics Committee,
Ethics ID: 1011753, The University of Melbourne. Cortices were
isolated from mouse fetuses at embryonic day 14 in Krebs solution,
with added glucose (14 mM).
Cortices were subsequently incubated for 20 min at 37 °C in Krebs
solution containing added trypsin (110 μM, Sigma). The solution was
then centrifuged for 3 min at 259 RCF. After supernatant was
removed, Krebs solution (10 mL) with DNase (25.8 nM) and STBI
(129.38 nM Sigma) was added to the pellet and incubated for 2 min.
Cells were then resuspended. The suspension was further centrifuged
at 259 RCF for 3 min, supernatant was removed, and cells were
resuspended in cortical plating medium comprising 10% 10× MEM, L-
glutamine (1 μM), NaHCO3 (50 mM), fetal bovine serum (10%), and
horse serum (5%). Cells were plated at a density of 6 × 104 cells/cm2
in an Ibidi 35 mm live cell imaging dish coated with 20 μg/mL poly-D-
lysine. Cells were incubated at 37 °C for 24 h; then plating medium
was replaced with Neurobasal medium supplemented with 2% B-27
supplement and L-glutamine (500 μM). After 5 days in culture at 37
°C, 5% CO2, half of the medium was replaced. Cells were used for
experiments after 8 days in culture.
ACKNOWLEDGMENTS
■
Prof. Kevin Barnham (University of Melbourne) is acknowl-
edged for his pivotal role in initiating our research in this area
and the insight he continues to provide. We acknowledge Prof.
Xiaotao Hao and Dr. Lachlan McKimmie for assistance with
some of the FLIM measurements, and Prof. Andrew F. Hill (La
Trobe University, Australia) for guidance on live-cell imaging
studies. We acknowledge the Biological Optical Microscopy
Platform (BOMP), University of Melbourne, for use of the
confocal live cell microscopes.
REFERENCES
■
(1) French, F. A.; Freelander, B. L. Cancer Res. 1958, 18, 1290.
(2) Paterson, B. M.; Donnelly, P. S. Chem. Soc. Rev. 2011, 40, 3005.
(3) Green, M. A.; Klippenstein, D. L.; Tennison, J. R. J. Nucl. Med.
1988, 29, 1549.
(4) Green, M. A.; Mathias, C. J.; Welch, M. J.; McGuire, A. H.; Perry,
D.; Fernandez-Rubio, F.; Perlmutter, J. S.; Raichle, M. E.; Bergmann,
S. R. J. Nucl. Med. 1990, 31, 1989.
(5) Donnelly, P. S. Dalton Trans. 2011, 40, 999.
(6) Fujibayashi, Y.; Taniuchi, H.; Yonekura, Y.; Ohtani, H.; Konishi,
J.; Yokoyama, A. J. Nucl. Med. 1997, 38, 1155.
(7) Tanaka, T.; Furukawa, T.; Fujieda, S.; Kasamatsu, S.; Yonekura,
Y.; Fujibayashi, Y. Nucl. Med. Biol. 2006, 33, 743.
(8) Lewis, J. S.; Laforest, R.; Dehdashti, F.; Grigsby, P. W.; Welch, M.
J.; Siegel, B. A. J. Nucl. Med. 2008, 49, 1177.
(9) Vavere, A. L.; Lewis, J. S. Dalton Trans. 2007, 4893.
(10) Fujibayashi, Y.; Cutler, C. S.; Anderson, C. J.; McCarthy, D. W.;
Jones, L. A.; Sharp, T.; Yonekura, Y.; Welch, M. J. Nucl. Med. Biol.
1999, 26, 117.
(11) Basken, N. E.; Green, M. A. Nucl. Med. Biol. 2009, 36, 495.
(12) Haynes, N. G.; Lacy, J. L.; Nayak, N.; Martin, C. S.; Dai, D.;
Mathias, C. J.; Green, M. A. J. Nucl. Med. 2000, 41, 309.
(13) Dearling, J. L. J.; Lewis, J. S.; McCarthy, D. W.; Welch, M. J.;
Blower, P. J. Chem. Commun. 1998, 2531.
Cell Treatments. Cells were treated with compounds (25 μM)
and cotreated with combinations of molecular probes (Life
Technologies). Concentrations of the molecular probes were as
follows: LysoTracker red (1 μM), MitoTracker far red (2 μM), and
HCS LipidTOX deep red neutral lipid stain (2 μM). After incubation
for 1 h, the medium was removed and replaced with HEPES live cell
imaging buffer. Confocal images were attained using a Leica SP5
confocal microscope, with LAS AF (Leica) software.
(14) Dearling, J. L. J.; Lewis, J. S.; Mullen, G. D.; Welch, M. J.;
Blower, P. J. JBIC, J. Biol. Inorg. Chem. 2002, 7, 249.
(15) Xiao, Z.; Donnelly, P. S.; Zimmermann, M.; Wedd, A. G. Inorg.
Chem. 2008, 47, 4338.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.5b01599.
■
S
(16) Donnelly, P. S.; Caragounis, A.; Du, T.; Laughton, K. M.;
Volitakis, I.; Cherny, R. A.; Sharples, R. A.; Hill, A. F.; Li, Q.-X.;
Masters, C. L.; Barnham, K. J.; White, A. R. J. Biol. Chem. 2008, 283,
4568.
(17) Crouch, P. J.; Hung, L. W.; Adlard, P. A.; Cortes, M.; Lal, V.;
Filiz, G.; Perez, K. A.; Nurjono, M.; Caragounis, A.; Du, T.; Laughton,
K.; Volitakis, I.; Bush, A. I.; Li, Q.-X.; Masters, C. L.; Cappai, R.;
Cherny, R. A.; Donnelly, P. S.; White, A. R.; Barnham, K. J. Proc. Natl.
Acad. Sci. U. S. A. 2009, 106, 381.
Crystallographic data for H2L1 (CIF)
Crystallographic data for ZnIIL1 (CIF)
Crystallographic data for CuIIL1 (CIF)
Fluorescence decay data and curve fits for L, CuIIL1, and
ZnIIL1 (PDF)
(18) Fodero-Tavoletti, M. T.; Villemagne, V. L.; Paterson, B. M.;
White, A. R.; Li, Q.-X.; Camakaris, J.; O’Keefe, G.; Cappai, R.;
Barnham, K. J.; Donnelly, P. S. J. Alzheimer’s Dis. 2010, 20, 49.
(19) Hung, L. W.; Villemagne, V. L.; Cheng, L.; Sherratt, N. A.;
Ayton, S.; White, A. R.; Crouch, P. J.; Lim, S.; Leong, S. L.; Wilkins, S.;
George, J.; Roberts, B. R.; Pham, C. L. L.; Liu, X.; Chiu, F. C. K.;
Shackleford, D. M.; Powell, A. K.; Masters, C. L.; Bush, A. I.; O’Keefe,
G.; Culvenor, J. G.; Cappai, R.; Cherny, R. A.; Donnelly, P. S.; Hill, A.
F.; Finkelstein, D. I.; Barnham, K. J. J. Exp. Med. 2012, 209, 837.
(20) Roberts, B. R.; Turner, B. J.; Bush, A. I.; Masters, C. L.; Li, Q.-
X.; Lim, N. K. H.; McAllum, E. J.; Price, K. A.; Kanninen, K. M.;
Liddell, J. R.; Grubman, A.; Donnelly, P. S.; Lim, S.; Paterson, B. M.;
Hickey, J. L.; Hare, D. J.; Doble, P. A.; Rhoads, T. W.; Williams, J. R.;
Hung Lin, W.; Monty, J.-F.; Llanos, R. M.; Mercer, J. F. B.; Kramer, D.
R.; Duce, J. A.; Beckman, J. S.; Barnham, K. J.; White, A. R.; Crouch, P.
J. J. Neurosci. 2014, 34, 8021.
AUTHOR INFORMATION
Corresponding Authors
*E-mail (A. R. White): arwhite@unimelb.edu.au. Fax: +61 3
8344 4004. Tel: +61 3 8344 1805.
*E-mail (T. A. Smith): trevoras@unimmelb.edu.au. Tel: +61 3
8344 6272.
*E-mail (P. S. Donnelly): pauld@unimelb.edu.au. Tel: +61 3
8344 2399.
■
Author Contributions
⊥J. L. Hickey and J. L. James contributed equally.
Funding
Australian Research Council, National and Health and Medical
Research Council (Australia). P.S.D. is an ARC Future Fellow.
(21) Soon, C. P. W.; Donnelly, P. S.; Turner, B. J.; Hung, L. W.;
Crouch, P. J.; Sherratt, N. A.; Tan, J. L.; Lim, N. K. H.; Lam, L.; Bica,
L.; Lim, S. C.; Hickey, J. L.; Morizzi, J.; Powell, A.; Finkelstein, D. I.;
Notes
The authors declare no competing financial interest.
K
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Culvenor, J. G.; Masters, C. L.; Duce, J.; White, A. R.; Barnham, K. J.;
Li, Q. X. J. Biol. Chem. 2011, 286, 44035.
(22) McAllum, E. J.; Lim, N. K. H.; Hickey, J. L.; Paterson, B. M.;
Donnelly, P. S.; Li, Q.-X.; Liddell, J. R.; Barnham, K. J.; White, A. R.;
Crouch, P. J. Amyotrophic Lateral Scler. Frontotemporal Degener. 2013,
14, 586.
(23) Cowley, A. R.; Davis, J.; Dilworth, J. R.; Donnelly, P. S.;
Dobson, R.; Nightingale, A.; Peach, J. M.; Shore, B.; Kerr, D.;
Seymour, L. Chem. Commun. 2005, 845.
(24) Dayal, D.; Palanimuthu, D.; Shinde, S. V.; Somasundaram, K.;
Samuelson, A. G. JBIC, J. Biol. Inorg. Chem. 2011, 16, 621.
(25) Palanimuthu, D.; Shinde, S. V.; Dayal, D.; Somasundaram, K.;
Samuelson, A. G. Eur. J. Inorg. Chem. 2013, 2013, 3542.
(26) Pascu, S. I.; Waghorn, P. A.; Conry, T. D.; Betts, H. M.;
Dilworth, J. R.; Churchill, G. C.; Pokrovska, T.; Christlieb, M.;
Aigbirhio, F. I.; Warren, J. E. Dalton Trans. 2007, 4988.
(27) Pascu, S. I.; Waghorn, P. A.; Conry, T. D.; Lin, B.; Betts, H. M.;
Dilworth, J. R.; Sim, R. B.; Churchill, G. C.; Aigbirhio, F. I.; Warren, J.
E. Dalton Trans. 2008, 2107.
(28) Pascu, S. I.; Waghorn, P. A.; Kennedy, B. W. C.; Arrowsmith, R.
L.; Bayly, S. R.; Dilworth, J. R.; Christlieb, M.; Tyrrell, R. M.; Zhong,
J.; Kowalczyk, R. M.; Collison, D.; Aley, P. K.; Churchill, G. C.;
Aigbirhio, F. I. Chem. - Asian J. 2010, 5, 506.
(29) Lim, S.; Price, K. A.; Chong, S.-F.; Paterson, B. M.; Caragounis,
A.; Barnham, K. J.; Crouch, P. J.; Peach, J. M.; Dilworth, J. R.; White,
A. R.; Donnelly, P. S. JBIC, J. Biol. Inorg. Chem. 2010, 15, 225.
(30) Price, K. A.; Crouch, P. J.; Lim, S.; Paterson, B. M.; Liddell, J. R.;
Donnelly, P. S.; White, A. R. Metallomics 2011, 3, 1280.
(31) Baggaley, E.; Botchway, S. W.; Haycock, J. W.; Morris, H.;
Sazanovich, I. V.; Williams, J. A. G.; Weinstein, J. A. Chem. Sci. 2014, 5,
879.
R.; Donnelly, P. S.; Green, J. C.; Peach, J. M.; Vasudevan, S. R.;
Warren, J. E. Inorg. Chem. 2007, 46, 465.
(49) Holland, J. P.; Green, J. C.; Dilworth, J. R. Dalton Trans. 2006,
783.
(50) Martin, S.; Parton, R. G. Nat. Rev. Mol. Cell Biol. 2006, 7, 373.
(51) Farese, R. V., Jr.; Walther, T. C. Cell 2009, 139, 855.
(52) Guo, Y.; Cordes, K. R.; Farese, R. V., Jr.; Walther, T. C. J. Cell
Sci. 2009, 122, 749.
(53) Granneman, J. G.; Moore, H.-P. H.; Mottillo, E. P.; Zhu, Z. J.
Biol. Chem. 2009, 284, 3049.
(54) Singh, R.; Kaushik, S.; Wang, Y.; Xiang, Y.; Novak, I.; Komatsu,
M.; Tanaka, K.; Cuervo, A. M.; Czaja, M. J. Nature 2009, 458, 1131.
(55) Brotherus, J.; Renkonen, O. J. Lipid Res. 1977, 18, 191.
(56) Matsumoto, K.; Fujibayashi, Y.; Arano, Y.; Wada, K.; Yokoyama,
A. Nucl. Med. Biol. 1992, 19, 33.
(57) Petering, D. H. Biochem. Pharmacol. 1974, 23, 567.
(58) Donnelly, P. S.; Liddell, J. R.; Lim, S.; Paterson, B. M.; Cater, M.
A.; Savva, M. S.; Mot, A. I.; James, J. L.; Trounce, I. A.; White, A. R.;
Crouch, P. J. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 47.
(59) Yoshii, Y.; Yoneda, M.; Ikawa, M.; Furukawa, T.; Kiyono, Y.;
Mori, T.; Yoshii, H.; Oyama, N.; Okazawa, H.; Saga, T.; Fujibayashi, Y.
Nucl. Med. Biol. 2012, 39, 177.
(60) Liu, L.; Zhang, K.; Sandoval, H.; Yamamoto, S.; Jaiswal, M.;
Sanz, E.; Li, Z.; Hui, J.; Graham, B. H.; Quintana, A.; Bellen, H. J. Cell
2015, 160, 177.
(61) Sanhueza, M.; Chai, A.; Smith, C.; McCray, B. A.; Simpson, T.
I.; Taylor, J. P.; Pennetta, G. PLoS Genet. 2015, 11, e1005107.
(62) Wada, K.; Yasuhisa; Tajima, N.; Naoyuki; Yokoyama, A.; Akira;
Fujibayashi, Y. Biol. Pharm. Bull. 1994, 17, 701.
(63) Fahrni, C. J. Curr. Opin. Chem. Biol. 2013, 17, 656.
(64) Cotruvo, J. A., Jr.; Aron, A. T.; Ramos-Torres, K. M.; Chang, C.
J. Chem. Soc. Rev. 2015, 44, 4400.
(32) Baggaley, E.; Gill, M. R.; Green, N. H.; Turton, D.; Sazanovich,
I. V.; Botchway, S. W.; Smythe, C.; Haycock, J. W.; Weinstein, J. A.;
Thomas, J. A. Angew. Chem., Int. Ed. 2014, 53, 3367.
(33) Renfrew, A. K.; Bryce, N. S.; Hambley, T. W. Chem. Sci. 2013, 4,
3731.
(65) Shen, C.; New, E. J. Metallomics 2015, 7, 56.
(66) Price, K. A.; Hickey, J. L.; Xiao, Z.; Wedd, A. G.; James, S. A.;
Liddell, J. R.; Crouch, P. J.; White, A. R.; Donnelly, P. S. Chem. Sci.
2012, 3, 2748.
(67) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr.
(34) Arrowsmith, R. L.; Waghorn, P. A.; Jones, M. W.; Bauman, A.;
Brayshaw, S. K.; Hu, Z.; Kociok-Koehn, G.; Mindt, T. L.; Tyrrell, R.
M.; Botchway, S. W.; Dilworth, J. R.; Pascu, S. I. Dalton Trans. 2011,
40, 6238.
2008, 64, 112.
(68) Farrugia, L. J. J. Appl. Crystallogr. 2012, 45, 849.
(69) Wildanger, D.; Rittweger, E.; Kastrup, L.; Hell, S. W. Opt.
Express 2008, 16, 9614.
(70) Hao, X.-T.; McKimmie, L. J.; Smith, T. A. J. Phys. Chem. Lett.
(35) Ulrich, G.; Ziessel, R.; Harriman, A. Angew. Chem., Int. Ed. 2008,
2011, 2, 1520.
47, 1184.
(36) Ziessel, R.; Bonardi, L.; Ulrich, G. Dalton Trans. 2006, 2913.
(37) Loudet, A.; Burgess, K. Chem. Rev. 2007, 107, 4891.
(38) Waghorn, P. A.; Jones, M. W.; Theobald, M. B. M.; Arrowsmith,
R. L.; Pascu, S. I.; Botchway, S. W.; Faulkner, S.; Dilworth, J. R. Chem.
Sci. 2013, 4, 1430.
(39) Biedler, J. L.; Roffler-Tarlov, S.; Schachner, M.; Freedman, L. S.
Cancer Res. 1978, 38, 3751.
(40) Fischer, H. Organic Syntheses; Wiley: New York, 1935; Vol. XV,
p 20
(41) Fischer, H. Organic Syntheses; Wiley: New York, 1935; Vol. XV,
p 17
(42) Paterson, B. M.; Karas, J. A.; Scanlon, D. B.; White, J. M.;
Donnelly, P. S. Inorg. Chem. 2010, 49, 1884.
(43) Buncic, G.; Donnelly, P. S.; Paterson, B. M.; White, J. M.;
Zimmermann, M.; Xiao, Z.; Wedd, A. G. Inorg. Chem. 2010, 49, 3071.
(44) Azov, V. A.; Diederich, F.; Lill, Y.; Hecht, B. Helv. Chim. Acta
2003, 86, 2149.
(45) Azov, V. A.; Skinner, P. J.; Yamakoshi, Y.; Seiler, P.; Gramlich,
V.; Diederich, F. Helv. Chim. Acta 2003, 86, 3648.
(46) Cowley, A. R.; Dilworth, J. R.; Donnelly, P. S.; Labisbal, E.;
Sousa, A. J. Am. Chem. Soc. 2002, 124, 5270.
(47) Blower, P. J.; Castle, T. C.; Cowley, A. R.; Dilworth, J. R.;
Donnelly, P. S.; Labisbal, E.; Sowrey, F. E.; Teat, S. J.; Went, M. J.
Dalton Trans. 2003, 4416.
(48) Holland, J. P.; Aigbirhio, F. I.; Betts, H. M.; Bonnitcha, P. D.;
Burke, P.; Christlieb, M.; Churchill, G. C.; Cowley, A. R.; Dilworth, J.
L
DOI: 10.1021/acs.inorgchem.5b01599
Inorg. Chem. XXXX, XXX, XXX−XXX