Tunable heptamethine-azo dye conjugate as an NIR fluorescent probe for the selective detection of mitochondrial glutathione over cysteine and homocysteine

Article abstract of DOI:10.1021/ja500962u

Although a lot of mitochondria-targeting biothiol probes have been developed and applied to cellular imaging through thiol-induced disulfide cleavage or Michael addition reactions, relatively few probes assess mitochondrial GSH with high selectivity over Cys and Hcy and with NIR fluorescence capable of noninvasive imaging in biological samples. In order to monitor mitochondrial GSH with low background autofluorescence, we designed a heptamethine-azo conjugate as an NIR fluorescent probe by introducing a tunable lipophilic cation unit as the biomarker for mitochondria and a nitroazo group as the GSH-selective reaction site as well as the fluorescence quencher. The probe exhibited a dramatic off-on NIR fluorescence response toward GSH with high selectivity over other amino acids including Cys and Hcy. Further application to cellular imaging indicated that the probe was highly responsive to the changes of mitochondrial GSH in cells.

Full text of DOI:10.1021/ja500962u

Article
pubs.acs.org/JACS
Tunable Heptamethine−Azo Dye Conjugate as an NIR Fluorescent
Probe for the Selective Detection of Mitochondrial Glutathione over
Cysteine and Homocysteine
Soo-Yeon Lim,† Keum-Hee Hong,† Dae Il Kim,‡ Hyockman Kwon,‡ and Hae-Jo Kim*,†
†Department of Chemistry, Hankuk University of Foreign Studies, Yongin 449-791, Republic of Korea
‡Department of Bioscience and Biotechnology, Hankuk University of Foreign Studies, Yongin 449-791, Republic of Korea
S
* Supporting Information
ABSTRACT: Although a lot of mitochondria-targeting
biothiol probes have been developed and applied to cellular
imaging through thiol-induced disulfide cleavage or Michael
addition reactions, relatively few probes assess mitochondrial
GSH with high selectivity over Cys and Hcy and with NIR
fluorescence capable of noninvasive imaging in biological
samples. In order to monitor mitochondrial GSH with low
background autofluorescence, we designed a heptamethine−
azo conjugate as an NIR fluorescent probe by introducing a
tunable lipophilic cation unit as the biomarker for mitochon-
dria and a nitroazo group as the GSH-selective reaction site as
well as the fluorescence quencher. The probe exhibited a dramatic off−on NIR fluorescence response toward GSH with high
selectivity over other amino acids including Cys and Hcy. Further application to cellular imaging indicated that the probe was
highly responsive to the changes of mitochondrial GSH in cells.
rearrangement reaction to produce the thermodynamically
stable amino-substituted BODIPY complex, whereas the probe
formed a thioether type of BODIPY complex with GSH and
produced a strong fluorescence at 588 nm. Strongin and his co-
workers reported another interesting micelle-mediated GSH
probe, which displayed a GSH selective green fluorescence in
the presence of cationic micelles.4b Abliz and his co-workers
also devised a reversible fluorescent sensor, which was
successfully applied for selective intracellular imaging of
GSH.4c However, these GSH-selective probes were limited in
their ability to target mitochondria.
Until now, few fluorescent probes were available to target
mitochondrial GSH with NIR fluorescence, which is more
desirable for noninvasive in vivo imaging of mitochondrion-
related diseases without interfering with cellular autofluor-
escence. Herein we report a heptamethine−azo dye conjugate
(MitoGP), where the tunable lipophilic cation of a
heptamethine unit was utilized as the mitochondrial biomarker
and a nitroazo group was introduced as the GSH-selective
reaction unit as well as the fluorescence quencher. In the
presence of GSH, MitoGP exhibited a dramatic fluorescence
off−on response at λmax 810 nm, while the other amino acids
including Cys and Hcy displayed no significant NIR
fluorescence relative to GSH. Cellular imaging experiments
INTRODUCTION
■
Mitochondrion is an essential organelle within eukaryotic cells,
utilizing oxygen to digest carbohydrates and fats and releasing
reactive oxygen species (ROS) as well as biochemical energy in
the form of adenosine triphosphate. In addition, mitochondria
are also involved in the event of ROS-induced apoptosis. Due
to these crucial functions, oxidative damage to the mitochon-
dria is a significant factor on cell death which underlies many
neurodegenerative diseases and pathologies. In order to protect
cells from the oxidative stress, mitochondria are equipped with
a number of free radical scavengers and the mitochondrial
glutathione (γ-glutamylcysteinylglycine or GSH) pool is a
critical antioxidant reservoir within cells.1 Therefore, it is a
challenge to develop a mitochondrial GSH probe and to apply
it for the detection of the level of mitochondrial GSH in cells.
In spite of the remarkable advances in the design of
fluorescent probes for biothiols2 including Cys and Hcy3 over
the past decade, relatively few probes have been developed that
exhibit a selective fluorescence response to mitochondrial
GSH.4 Most of them, including the triphenylphosphonium-
appended disulfide fluorophore and the commercially available
monochlorobimane (mCB), cannot discriminate GSH from
Cys or Hcy due to the functional similarity of those biothiols.
Recently, a pioneering work by Yang and his co-workers
showed that a BODIPY-based probe could produce a highly
selective and ratiometric fluorescence change with GSH relative
to Cys and Hcy.4a The probe, if treated with Cys or Hcy,
formed an initial thioether and subsequently underwent a
Received: January 28, 2014
Published: April 23, 2014
© 2014 American Chemical Society
7018
dx.doi.org/10.1021/ja500962u | J. Am. Chem. Soc. 2014, 136, 7018−7025

Journal of the American Chemical Society
Article
clearly indicated that the fluorescence was well reflected by the
changes of GSH concentration in mitochondria.
introduced as the GSH-selective reaction unit as well as the
fluorescence quencher unit. Through a plausible substitution
reaction of the azo group with GSH, the probe was expected to
significantly enhance the fluorescence of MitoGP. The
mitochondrial GSH probe was effectively prepared from
heptamethine chloride (1) and nitroazo phenol (2) (Scheme
2). The precursor, heptamethine chloride (1) was readily
available according to a modified procedure previously
described in the literature.5 The analogues of MitoGP (3, 4)
were synthesized through hydrolysis of MitoGP and sub-
sequent amide coupling, respectively, in order to investigate the
effect of the tunable functionality of MitoGP on quantum
yields and cellular distribution.
RESULTS AND DISCUSSION
■
Design and Synthesis of MitoGP. To target mitochon-
drial GSH with a high signal-to-noise (S/N) ratio, we designed
an NIR fluorophore (MitoGP) by employing both the
mitochondria-targetable site with a tunable cationic heptame-
thine unit and the GSH-activatable site with a labile nitroazo
aryl ether group (Scheme 1). The nitroazo ether group was
Scheme 1. Design of an NIR-Based Mitochondrial GSH
Probe
Reaction Kinetics. The addition of GSH (10 mM) to
MitoGP (10 μM in HEPES buffer, 0.10 M, pH 7.4) induced a
typical ratiometric change with two isosbestic points at 600 and
824 nm in UV−vis spectra, whose kinetic analysis gave the
second-order rate constant of k 0.062 ( 0.005) M−1 s−1 at 25
°C. The fluorescence spectra of MitoGP displayed a
remarkable red-shift (λem 764 to 810 nm) in its maximum
wavelength with a dramatic increase in the intensity at λmax 810
nm upon addition of GSH (Figure 1). The quantum yield of
MitoGP was measured in the absence and presence of GSH,
using IR-820 as a reference compound,6 resulted in Φ 0.0011
and 0.187 for MitoGP and MitoGP−GSH, respectively. As
expected, the initial quantum yield of MitoGP was quite low
due to the quenching effect of a nitroazo group of MitoGP
through a probable photoinduced electron transfer (PeT) from
cyanine to the nitroazo group. However, reaction of MitoGP
with GSH resulted in a significantly larger value that was as high
Scheme 2. Synthesis of a Mitochondrial GSH Probe (MitoGP) and Its Analogues
7019
dx.doi.org/10.1021/ja500962u | J. Am. Chem. Soc. 2014, 136, 7018−7025

Journal of the American Chemical Society
Article
Figure 1. Time-dependent UV−vis (A) and fluorescence (B) spectra of MitoGP (10 μM, ex 600 nm) upon addition of GSH (10 mM) in HEPES
buffer (0.10 M, pH 7.4). Inset: UV−vis kinetics.
as Φ 0.187 (MitoGP−GSH). This dramatic fluorescence turn-
on effect is one of the advantages of MitoGP compared to the
other cyanine-based NIR dye such as control compound 1 (Φ
0.022), where the inherent fluorescence was relatively strong
compared to MitoGP (Table 1).7 The analogues 3 and 4 also
Table 1. Spectroscopic Properties of MitoGP and Its
a
Analogues
max. λabs
/
ε/105 M−1 max. λem
/
entry
cmpd
nm
cm−1
nm
Φ (%)
1
2
3
4
5
1
780
790
755
770
787
0.20
0.43
0.61
0.69
1.01
790
764
776
775
810
2.16
0.11
5.59
4.06
18.7
MitoGP
3
4
b
MitoGP−GSH
a
Spectral measurements were performed in HEPES buffer (0.10 M,
b
pH 7.4) by using IR-820 as a reference compound. Spectral data of
the complex between MitoGP and GSH was measured 12 h after
addition of 10 mM of GSH in HEPES buffer (0.10 M, pH 7.4).
exhibited strong intrinsic fluorescence with Φ 0.056 and 0.041,
respectively, at wavelengths similar to that of MitoGP. The
strong fluorescence of the analogues and compound 1 was not
likely to increase the S/N ratio for cellular imaging. We,
therefore, expected that MitoGP would be a good candidate as
an imaging probe for mitochondrial GSH in cells.
Figure 2. (A) Fluorescence response of MitoGP (10 μM, λex 600 nm)
with various AAs, GSH, Cys and Hcy (10 mM, 3 h) in HEPES buffer
(0.10 M, pH 7.4). (B) The competitive assay of GSH against MitoGP
+ AA: Black columns for MitoGP + AA, cross bar columns for
MitoGP + AA + GSH. F0 is the fluorescence intensity of MitoGP
alone at λ 810 nm.
GSH Selectivity. The GSH selectivity of MitoGP in
HEPES buffer (0.10 M, pH 7.4) was measured for various
amino acids (AAs) including Cys and Hcy. MitoGP did not
produce any detectable fluorescence without GSH (F0 = 0.138
at λmax 810 nm) but the addition of GSH induced a dramatic
fluorescence increase with F/F0 460 at λmax 810 nm (Figure
2A). Interestingly, Cys and Hcy showed weak and blue-shifted
fluorescence responses at λmax 747 nm8 but their fluorescence
enhancement was very weak compared to GSH: F/F0 10 and
15 for Cys and Hcy, respectively. In addition, competitive
fluorescence assay showed that most AAs did not exhibit any
interference with GSH unlike Cys and Hcy (Figure 2B). Cys or
Hcy might form stable amino-substituted covalent complexes
with MitoGP through a sequential thiol substitution and
rearrangement reaction as observed in Yang’s BODIPY-based
probe.4a We observed that the initial thioether underwent a
subsequent intramolecular amination reaction (Figure S15 in
the Supporting Information [SI]), which induced a blue shift in
the fluorescence spectra of low intensity associated with Cys/
Hcy (Figure S16 in the SI). The competitive experiments
indicated that MitoGP was so selective toward GSH that its
fluorescence was dramatically turned on by GSH.
NMR Spectral Analysis. To gain insight into the reaction
1
mechanism of MitoGP with GSH, we performed H NMR
spectral analysis of MitoGP in the presence of 2-mercaptoe-
thanol (ME), an organic soluble analogue of GSH. Upon
addition of ME in methanol-d4, the spectrum of MitoGP was
slowly converted into another discrete set of MitoGP−ME. In
7020
dx.doi.org/10.1021/ja500962u | J. Am. Chem. Soc. 2014, 136, 7018−7025

Journal of the American Chemical Society
Article
Figure 3. Partial 1H NMR spectra of MitoGP (10 mM) in the presence of ME (1.2 equiv) in CD3OD. (A) MitoGP, (B) 0.5 h after addition of ME,
(C) 2 h, (D) 12 h, (E) MitoGP−ME.
Figure 4. (A) Fluorescence changes of MitoGP (10 μM, λex 600 nm) on the incremental addition of GSH in HEPES buffer (0.10 M, pH 7.4);
(inset) linear plot of the fluorescence intensity of MitoGP against GSH. (B) Job’s plot between MitoGP (λex/λem 600/810 nm) and GSH. Total
concentrations of MitoGP and GSH were kept constant at 10 mM during incubation (3 h) and then diluted with HEPES buffer (0.10 M, pH 7.4) to
measure their fluorescence intensities.
1
the presence of 1 equiv ME, the vinylic protons of MitoGP
were dramatically shifted to downfield regions (Ha 8.01 to 8.98
ppm, Hb 6.33 to 6.40 ppm) while the nitroazo aryl ether proton
of MitoGP was shifted to an upfield region (Hc 7.40 ppm to
6.97 ppm) owing to the release of free nitroazophenol. The
product structure was confirmed by an authentic MitoGP−ME
compound (Figure 3). The spectral changes indicated that the
azophenol of MitoGP was easily substituted with a thiol group,
which produced a dramatic change in the fluorescence. When
treated with 0.50 equiv of Cys, however, MitoGP was
converted from the initial thioether into a stable amine
structure through a rearrangement reaction that resulted in a
stable 2:1 complex between MitoGP and Cys visible in the H
NMR and mass spectra (Figures S15 and S13 in the SI).9
Limit of Detection and Reaction Stoichiometry. The
limit of detection of GSH was measured with MitoGP (10 μM)
by incremental addition of GSH in HEPES buffer (0.1 M, pH
7.4) (Figure 4A). The fluorescence intensity of MitoGP
gradually increased according to the amount of GSH added and
eventually saturated at 10 mM GSH (Figure S14 in the SI).
From the linear range of the titration plot, the detection limit of
GSH was determined to be as low as 26 nM at 3σ/m, where σ
is the standard deviation of blank measurements of MitoGP
and m is the slope obtained from the linear plot of MitoGP
fluorescence against GSH (Figure 4A, inset).
7021
dx.doi.org/10.1021/ja500962u | J. Am. Chem. Soc. 2014, 136, 7018−7025

Journal of the American Chemical Society
Article
Figure 5. Proposed reaction mechanism of MitoGP with GSH vs Cys/Hcy.
Figure 6. Reaction profile of MitoGP with ME (both 10 mM) in different solvents (A) and in the presence of radical/electron scavengers (B). The
1
reaction was measured by H NMR spectral analysis.
The reaction ratio of MitoGP was precisely determined using
the Job’s plot. The fluorescence intensity of MitoGP at 810 nm
reached the maximum value when the mole fraction of GSH
was 0.5, indicating that one-to-one reaction took place between
MitoGP and GSH (Figure 4B).
Further pH-dependent fluorescence measurement showed
that MitoGP−GSH complex displayed a strong fluorescence at
810 nm with a stable pH profile between 6 and 9, whereas
MitoGP exhibited negligible fluorescence over a wide range of
pHs (Figure S17 in the SI).
Mechanistic Study. From the combined results of UV−vis,
fluorescence, and NMR spectral data, we propose the reaction
mechanism of MitoGP with GSH in water. The initial MitoGP
is nonfluorescent due to a possible photoinduced electron
transfer (PeT) from indocyanine to a nitroazo group10 and,
furthermore, due to the soluble aggregate formation such as
vesicles or micelles in an aqueous buffer.11 The labile nitroazo
group was rapidly replaced by the 1,6-conjugate addition of an
alkyl thiol group of anionic GSH to the vinyl analogue of a
cationic aza Michael acceptor.12 The subsequent elimination
reaction produced strong fluorescence on the probe (Figure 5).
A radical nucleophilic substitution (SRN1) mechanism
through an initial single-electron transfer from the nucleophilic
thiol is also another possible explanation for the substitution in
organic solvents.13 However, we observed that the reaction rate
of MitoGP with GSH was severely affected by the solvents used
(Figure 6A), which would be implausible with a neutral radical
mechanism. For example, the rate was accelerated more than 10
times in protic solvent (methanol) than aprotic solvent
(DMSO). Furthermore, introduction of molecular oxygen
(O2), a radical scavenger, or nitrobenzene (PhNO2), an
electron scavenger, did not significantly alter the reaction
rates of MitoGP to ME: the second-order reaction kinetics
1
measured at 25 °C in H NMR spectroscopy gave the rate
constants (k) as 0.0189( 0.0042), 0.0208( 0.0022), and
0.0189( 0.0024) M−1 s−1 for MitoGP alone, MitoGP+O2,
7022
dx.doi.org/10.1021/ja500962u | J. Am. Chem. Soc. 2014, 136, 7018−7025

Journal of the American Chemical Society
Article
and MitoGP+PhNO2, respectively (Figure 6B). Consequently,
the thiol substitution reaction was likely to proceed through the
nucleophilic addition and elimination mechanism instead of the
radical substitution mechanism, at least in the protic solvent.
It is remarkable that neutral Cys and Hcy underwent
relatively a slow addition reaction with cationic MitoGP and
further rearrangement reaction with the neighboring amino
group created the amino substituted cyanine dyes, which were
thermodynamically stable and resistant to the nucleophilic
substitution reaction of the second GSH as observed in the
competitive reaction of MitoGP−Cys and MitoGP−Hcy
against GSH. The resulting amino-substituted cyanine dyes
provoked smaller changes in the fluorescence intensity than the
thiol-substituted probe (MitoGP−GSH), which was the key
contributor for discrimination of GSH from Cys and Hcy.
Role of Tunable Functional Group. We found that the
butyl ester functional group of MitoGP is very important for
the selective recognition and reaction kinetics of GSH in
mitochondria compared to that of carboxylic acid or amide.
While the cationic MitoGP was a good fluorescent probe for
mitochondrial GSH with a high S/N ratio as well as a
reasonable reaction rate, its acid analogue 3 was a poor probe
with a low S/N ratio for GSH and exhibited a very sluggish
reaction with GSH due to the electrostatic repulsion between
GSH and the negatively charged 3 (Figure S18 in the SI). The
amide analogue (4) showed almost a similar reaction rate as
MitoGP with a biothiol but its S/N ratio was not as good as
that of MitoGP. Furthermore, the analogues 3 and 4, unlike
MitoGP, were designed to target the biothiols in cytosol and
lysosome, respectively. We anticipate that the probe will be
extendable as organelle-specific dyes by changing the tunable
functionality at the position of the ester group of the probe,
which is currently in progress.
Figure 7. Confocal laser scanning microscopic images of MitoGP and
rhodamine 123 in living HeLa cells. (A) Rhodamine 123 (1.0 μM, 15
min, λex 488 nm). (B) MitoGP (1.0 μM, 30 min, λex 555 nm). (C)
Colocalization of the cells with the mitochondrial tracker and MitoGP.
Colocalization coefficient between rhodamine 123 and MitoGP was
analyzed by the implemented ZEN software to be 0.93 (0: no
colocalization, 1: colocalization in all pixels).
Cellular Imaging. Owing to the favorable biocompatible
attributes of MitoGP, we performed a cellular imaging
experiment of MitoGP under the confocal laser scanning
microscope. Colocalization study of MitoGP with a green
fluorescent mitochondrial tracker (rhodamine 123) clearly
showed that cellular mitochondria was effectively stainable with
1.0 μM of MitoGP and displayed a strong fluorescence whose
efficiency was comparable with that of the commercial
MitoTracker (Figure 7).
Figure 8. Confocal laser scanning microscopic images of MitoGP (2.0
μM, 30 min, λex 555 nm) in the presence of a mitochondrial GSH
scavenger (3-HP). (A) MitoGP only, (B) MitoGP + 3-HP (1 mM, 5
min, then 2 h incubation), (C) MitoGP + 3-HP (1 mM, 5 min, then 4
h incubation), (D) 3-HP concentration-dependent fluorescence of
MitoGP (3-HP for 5 min, then 2 h incubation).
If a mitochondrial specific enzyme (3-hydroxybutanoate
dehydrogenase)-related GSH scavenger such as 3-hydroxy-4-
pentenoate (3-HP, 1 mM, 5 min)14 or 3-oxo-4-pentenoate (3-
OP, 1 mM, 10 min)15 was applied prior to the administration of
MitoGP to HeLa cells, the fluorescence intensity was
significantly decreased. The degree of the fluorescence decrease
was dependent on the incubation time and the amount of the
scavenger (Figure 8). While the fluorescence intensity of
MitoGP was significantly reduced as much as 52% of the
original intensity by the pretreatment of 3-HP (1 mM, 5 min
and then incubation for 2 h), the cellular pretreatment with α-
lipoic acid (LPA, a GSH enhancer, 1.8 mM, 24 h)16 induced an
approximately 1.5-fold increase in the fluorescence intensity
relative to MitoGP alone (Figure 9). These cellular experi-
ments indicated that the fluorescence intensity of MitoGP was
well reflected on the levels of GSH in mitochondria and that
MitoGP is a good fluorescent indicator for mitochondrial GSH
in cells.
CONCLUSION
■
We designed a novel off−on mitochondrial GSH probe
(MitoGP) by introduction of both the mitochondria-targeting
site and the GSH reaction site to the backbone of a tunable
NIR heptamethine probe. The initial nitroazo probe was
nonfluorescent, but the reaction of MitoGP with GSH resulted
in a dramatic increase in the NIR fluorescence with a high
selectivity toward GSH over Cys/Hcy. The heptamethine−azo
conjugate was successfully applied to the direct cellular imaging
of mitochondrial GSH with a good response to varying
amounts of mitochondrial GSH in cells, which is useful as a
mitochondrial GSH tracker and much more superior to the
commercially available thiol probe (mCB) or GSH-silent
(rhodamine 123) MitoTracker. We anticipate the probe can
be further served as a therapeutic reagent for mitochondrial
GSH-related pathology.
7023
dx.doi.org/10.1021/ja500962u | J. Am. Chem. Soc. 2014, 136, 7018−7025

Journal of the American Chemical Society
Article
(48 mg, 0.500 mmol) was added to the solution and stirred for
20 h at room temperature. Trifluoroacetic acid (0.076 mL, 1.00
mmol) was slowly added to the mixture, which was then stirred
at room temperature for 6 h to produce a dark-green solution
from the original red color. After the reaction was complete, all
solvents were evaporated under reduced pressure. The crude
compound was purified by silica gel column chromatography
(CH2Cl2/CH3OH = 10/1) to obtain compound 3 as a green
solid. (63 mg, 78%).
3
1H NMR (CD3OD, 400 MHz): δ 8.40 (d, J = 9.2 Hz, 2H),
8.11 (d, 3J = 9.2 Hz, 2H), 8.05 (d, 3J = 9.2 Hz, 2H), 7.97 (d, 3J
3
= 14 Hz, 2H), 7.37−7.16 (m, 10H), 6.40 (d, J = 14 Hz, 2H),
4.38 (t, 3J = 7.2 Hz, 4H), 2.85 (t, 3J = 5.6 Hz, 4H), 2.62 (t, 3J =
7.2 Hz, 4H), 2.09 (q, 3J = 5.6 Hz, 2H), 1.37 (s, 12H). 13C NMR
(CD3OD, 100 MHz): δ 176.56, 172.21, 162.95, 162.56, 155.72,
148.66, 147.78, 141.96, 141.41, 141.15, 128.35, 125.86, 124.80,
124.47, 123.03, 121.87, 121.72, 115.50, 110.83, 100.45, 48.87,
41.63, 34.91, 26.82, 23.89, 21.04. (26 carbon peaks) HRMS
(MALDI+, DHB): [M]+ calcd for C48H48N5O7, 806.3554;
found, 806.3550.
Figure 9. Confocal laser scanning microscopic images of HeLa cells by
MitoGP (2.0 μM, 30 min, λex 555 nm) in the presence of LPA or 3-
HP. (A) MitoGP only, (B) MitoGP + LPA (1.8 mM, 24 h), (C)
MitoGP + 3-HP (1 mM, 5 min then 2 h incubation), (D) their
histograms of fluorescence intensities.
Synthesis of Compound 4. Compound 3 (11 mg, 14
μmol) was dissolved in anhydrous DCM (1 mL), and DCC (15
mg, 70 μmol) was added to the solution and stirred for 10 min
at 0 °C. N-Hydroxysuccinimide (NHS, 8 mg, 70 μmol) was
added to the mixture, which was stirred for 1 h at 0 °C. After
monitoring the reaction by TLC, 4-(2-aminoethyl)morpholine
(4 μL, 30 μmol) was added, and the solution stirred for an
additional 20 min at 0 °C. After the reaction wa complete, all
solvents were evaporated under reduced pressure. The crude
compound was purified by silica gel column chromatography
(CH2Cl2/CH3OH = 10/1) to obtain compound 4 as a green
solid. (5 mg, 33%).
EXPERIMENTAL SECTION
■
General Methods. All chemicals and solvents for synthesis
were purchased from commercial suppliers (Sigma-Aldrich,
Fluka, Acros) and were used without further purification, unless
otherwise stated. The composition of mixed solvents is given by
the volume ratio (v/v). Thin layer chromatography (TLC) was
performed using Merck silica gel 60 F254 alumina plates with a
thickness of 0.25 mm. Flash column chromatography was
performed on silica gel 60 (230−400 mesh ASTM; Merck). 1H
NMR and 13C NMR spectra were recorded using Bruker 400
1
(400 MHz for H, 100 MHz for 13C) with chemical shifts (δ)
3
1H NMR (CDCl3, 400 MHz): δ 8.39 (d, J = 9.2 Hz, 2H),
reported in ppm relative to the solvent residual signals of
8.08 (d, 3J = 9.2 Hz, 2H), 8.03 (d, 3J = 9.2 Hz, 2H), 7.92 (t, 3J =
1
CDCl3 (7.16 ppm for H, 77.16 ppm for 13C), CD3OD (3.31
3
ppm for 1H, 49.00 ppm for 13C) and coupling constants
reported in Hz. High resolution mass spectra (HRMS) were
obtained using a MALDI-TOF or FAB mass spectroscopy.
Synthesis of MitoGP. NaH (62 mg, 2.6 mmol) was added
slowly to a solution of compound 2 (0.42 g, 1.73 mmol) in
anhydrous DMF (3 mL). The mixture was stirred for 10 min,
and compound 1 (1.37 g, 1.73 mmol) in DMF (2 mL) was
added dropwise to the mixture. The resulting reaction mixture
was stirred at room temperature for 12 h. After the reaction was
complete, the solvent was removed under reduced pressure.
The crude product was purified by column chromatography on
silica gel using dichloromethane and methanol (10:1, v/v, Rf
0.35) as eluent, to give MitoGP as a dark-blue solid (1.30 g) in
75% yield.
5.2 Hz, 1H), 7.88 (d, J = 14 Hz, 2H), 7.38−7.15 (m, 10H),
6.34 (d, 3J = 14 Hz, 2H), 4.47 (t, 3J = 6.8 Hz, 4H), 3.69 (t, 3J =
4.4 Hz, 8H), 3.40 (dt, 3J = 6.4, 5.2 Hz, 4H), 2.87 (m, 8H), 2.54
(t, 3J = 6.4 Hz, 4H), 2.50 (t, 3J = 4.4 Hz, 8H), 2.10 (t, 3J = 5.6
Hz, 2H), 1.37 (s, 12H). 13C NMR (CDCl3, 100 MHz): δ
171.17, 169.43, 162.64, 155.57, 148.61, 147.67, 145.43, 141.84,
141.42, 140.73, 128.90, 126.15, 125.22, 124.75, 123.92, 122.92,
121.88, 115.43, 111.21, 110.02, 66.85. 57.09, 48.95, 41.17,
36.26, 33.86, 29.70, 27.95, 24.57, 21.11. HRMS (MALDI+,
DHB): [M]+ calcd for C60H72N9O7, 1030.5555; found,
1030.5530.
UV−Vis and Fluorescence Spectral Measurement. A
stock solution (10 mM) of MitoGP in DMSO was prepared.
For UV−vis spectral measurement, a sample solution was
prepared by mixing an appropriate amount of the stock solution
of MitoGP with an appropriate amount of each AA and finally
diluting with HEPES buffer (0.10 M, pH 7.4) to obtain the
desired concentration of MitoGP and AA. The fluorescence
intensities of the NIR dyes were similarly measured with a slit
width of 10 nm × 10 nm in medium or low mode of intensity.
Quantum Yield Measurement. IR-820 was chosen as a
reference compound for determination of the quantum yields.
The compound exhibits spectral properties similar to those of
MitoGP in both absorbance and fluorescence with a maximum
wavelength at 835 and 836 nm, respectively. The measurement
was carried out in HEPES buffer (0.10 M, pH 7.4). Quantum
yields were calculated by the comparison of the ratio between
the areas under the fluorescence curve of the dyes and the
reference compound. The measurements were taken at the
1H NMR (400 MHz, CDCl3) δ 8.37 (d, J = 9.2 Hz, 2H),
3
8.06 (d, 3J = 9.2 Hz, 2H), 8.01 (d, 3J = 9.2 Hz, 2H), 7.85 (d, 3J
3
= 14 Hz, 2H), 7.36−7.13 (m, 10H), 6.38 (d, J = 14 Hz, 2H),
4.61 (t, 3J = 6.4 Hz, 4H), 4.03 (t, 3J = 6.8 Hz, 4H), 2.92 (t, 3J =
3
3
6.4 Hz, 4H), 2.88 (t, J = 6.0 Hz, 4H), 2.11 (q, J = 6.0 Hz,
3
3
2H), 1.55 (q, J = 6.8 Hz, 4H), 1.34 (s, 12H), 1.31(q, J = 7.2
3
Hz, 4H), 0.89 (t, J = 7.2 Hz, 6H). 13C NMR (100 MHz,
CDCl3) δ 171.85, 170.91, 162.67, 162.65, 155.61, 148.58,
147.69, 141.81, 141.38, 140.79, 128.45, 126.16, 125.24, 124.73,
123.67, 123.12, 122.03, 115.43, 111.10, 101.45, 65.17, 48.95,
40.62, 32.26, 30.43, 27.93, 24.63, 21.12, 19.00, 13.64 (30 carbon
peaks). HRMS (MALDI+, DHB): [M]+ calcd for C56H64N5O7,
918.4803; found, 918.4808.
Synthesis of Compound 3. MitoGP (100 mg, 0.100
mmol) was dissolved in anhydrous THF (3 mL), and t-BuONa
7024
dx.doi.org/10.1021/ja500962u | J. Am. Chem. Soc. 2014, 136, 7018−7025

Journal of the American Chemical Society
Article
(6) IR-820 as a reference compound, ε at 835 nm = 1.2 × 105 M−1
cm−1, Φ 0.034 in DMSO, ex 720 nm. Berezin, M. Y.; Lee, H.; Akers,
W.; Achilefu, S. Biophys. J. 2007, 93, 2892.
(7) (a) Ogawa, M.; Kosaka, N.; Choyke, P. L.; Kobayashi, H. Cancer
Res. 2009, 69, 1268. (b) Tan, X.; Luo, S.; Wang, D.; Su, Y.; Cheng, T.;
Shi, C. Biomaterials 2012, 33, 2230.
(8) A complex between MitoGP and ethanolamine showed a similar
blue shift fluorescence maximum at 756 nm (Figure S21 in the SI),
which was generally observable by the fluorophores with amino-
substituted heptamethine dye: (a) Yang, Z.; Lee, J. H.; Jeon, H. M.;
Han, J. H.; Park, N.; He, Y.; Lee, H.; Hong, K. S.; Kang, C.; Kim, J. S. J.
Am. Chem. Soc. 2013, 135, 11657. (b) Kiyose, K.; Aizawa, S.; Sasaki, E.;
Kojima, H.; Hanaoka, K.; Terai, T.; Urano, Y.; Nagano, T. Chem.
Eur. J. 2009, 15, 9191.
(9) It is noteworthy that 1:1 complex formation between MitoGP
and Cys or Hcy in the fluorescence spectral conditions is assumed due
to the very low concentration of probes at micromolar concentration
and excess amount of Cys/Hcy (Figure S12 in the SI), as compared to
2:1 complex condition (Figure S13 in the SI).
same absorbance intensity for both of the dyes and the
reference compound at the indicated excitation wavelength.
Cellular Imaging Experiment. For the detection of GSH
in live cells, HeLa cells were cultured in Dulbecco’s modified
Eagle’s medium (DMEM) supplemented with 100 units/mL
penicillin, 100 μg/mL streptomycin, and 10% heat-inactivated
fetal bovine serum. The cells were seeded on a Ø 35 mm glass-
bottomed dish at a density of 1 × 105 cells in a culture medium
and incubated overnight in preparation for live-cell imaging by
confocal laser scanning microscope (CLSM). The HeLa cells
were treated with 1 or 2 μM of MitoGP in a serum free
medium for 30 min and washed twice with prewarmed 1 × PBS
before imaging by CLSM. To enhance the concentration of
cellular GSH, HeLa cells were treated with α-lipoic acid (LPA,
1.8 mM) for 24 h, washed twice with prewarmed 1 × PBS
buffer, and followed by MitoGP for 30 min before the cellular
imaging experiment was performed for the live cells. In order to
reduce the concentration of mitochondrial GSH, HeLa cells
were treated with mitochondrial enzyme substrates, 3-hydroxy-
4-pentenoate (3-HP) or 3-oxo-4-pentenoate (3-OP),17 washed
twice with prewarmed 1 × PBS buffer, and followed by
MitoGP for 30 min before the cellular images were taken under
a CLSM using the excitation channel (λex 555 nm).
(10) The nitro group usually acts as a fluorescence quencher. We
adopted the nitroazo group as a quencher because its extinction
coefficient is high enough to cause effective quenching of the NIR
fluorescence of MitoGP. (a) Wang, R.; Yu, F.; Chen, L.; Wang, L.;
Zhang, W. Chem. Commun. 2012, 48, 11757. (b) Han, M.; Hara, M. J.
Am. Chem. Soc. 2005, 127, 10951. (c) Rau, H. Angew. Chem., Int. Ed.
1972, 12, 224.
(11) We observed that the UV−vis spectra of MitoGP in a HEPES
buffer displayed ratiometric changes together with a bathochromic
shift (Δλ = + 10 nm) at its maximum wavelength probably due to J-
aggregation. The original spectra of MitoGP were rapidly recovered in
the presence of a surfactant such as cetyltrimethylammonium bromide,
indicating that a reversible reaction of vesicle or micelle-to-monomer
transition took place with MitoGP. Refer to the examples of J-
aggregation of cyanine dyes: (a) Vaidyanathan, S.; Patterson, L. K.;
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details including synthesis, NMR and MS spectra.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Mobius, D.; Gruniger, H. R. J. Phys. Chem. 1985, 89, 491.
̈
■
(b) Rotermund, F.; Weigand, R.; Holzer, W.; Wittmann, M.;
Penzkofer, A. J. Photochem. Photobiol. A: Chem. 1997, 110, 75.
(12) (a) Zhang, C.; Spokoyny, A. M.; Zou, Y.; Simon, M. D.;
Pentelute, B. L. Angew. Chem., Int. Ed. 2013, 52, 14001. (b) Gragg, J. L.
M.S. Thesis, Georgia State University, 2010, GA.
Corresponding Author
haejkim@hufs.ac.kr
Notes
The authors declare no competing financial interest.
(13) Strekowski, L.; Lipowska, M.; Patonay, G. J. Org. Chem. 1992,
57, 4578.
ACKNOWLEDGMENTS
■
(14) 3-HP is expected to be metabolized by the mitochondrial
specific enzyme (3-hydroxybutanoate dehydrogenase) to 3-OP, which
subsequently reacts with GSH and thereby depletes mitochondrial
GSH in cells. (a) Shan, X.; Jones, D. P.; Hashmi, M.; Anders, M. W.
Chem. Res. Toxicol. 1993, 6, 75. (b) Lluis, J. M.; Morales, A.; Blasco,
C.; Colell, A.; Mari, M.; Garcia-Ruiz, C.; Fernandez-Checa, J. C. J. Biol.
Chem. 2005, 280, 3224.
(15) Due to the cytotoxic effect of 3-OP (Figure S20 in the SI), we
prefer 3-HP as a mitochondrial GSH scavenger in cells.
(16) (a) Packer, L. Drug. Metab. Rev. 1998, 30, 245. (b) Packer, L.;
Tritschler, H. J.; Wessel, L. Free Radical Biol. Med. 1997, 22, 359.
(17) Zibuck, R.; Streber, J. M. J. Org. Chem. 1986, 54, 4717.
This work was supported by the National Research Foundation
of Korea (NRF 2011-0028456).
REFERENCES
■
(1) (a) Jocelyn, P. C. Biochim. Biophys. Acta 1975, 396, 427.
(b) Wahllander, A.; Soboll, S.; Sies, H. FEBS Lett. 1979, 97, 138.
(2) (a) Guo, Z.; Park, S.; Yoon, J.; Shin, I. Chem. Soc. Rev. 2014, 43,
16. (b) Lee, M. H.; Yang, Z.; Lim, C. W.; Lee, Y. H.; Dongbang, S.;
Kang, C.; Kim, J. S. Chem. Rev. 2013, 113, 5071. (c) Lim, C. S.;
Masanta, G.; Kim, H. J.; Han, J. H.; Kim, M. H.; Cho, B. R. J. Am.
Chem. Soc. 2011, 133, 11132. (d) Lee, M. H.; Han, J. H.; Lee, J.-H.;
Choi, H. G.; Kang, C.; Kim, J. S. J. Am. Chem. Soc. 2012, 134, 17314.
(3) (a) Yang, X.; Guo, Y.; Strongin, R. M. Angew. Chem., Int. Ed.
2011, 50, 10690. (b) Wang, H.; Zhou, G.; Gai, H.; Chen, X. Chem.
Commun. 2012, 48, 8341. (c) Guo, Z.; Nam, S. W.; Park, S.; Yoon, J.
Chem. Sci. 2012, 3, 2760. (d) Chen, H.; Zhao, Q.; Wu, Y.; Li, F.; Yang,
H.; Yi, T.; Huang, C. Inorg. Chem. 2007, 46, 11075. (e) Lee, K.-S.;
Kim, T.-K.; Lee, J. H.; Kim, H.-J.; Hong, J.-I. Chem. Commun. 2008,
6173.
(4) (a) Niu, L.-Y.; Guan, Y.-S.; Chen, Y.-Z.; Wu, L.-Z.; Tung, C.-H.;
Yang, Q.-Z. J. Am. Chem. Soc. 2012, 134, 18928. (b) Guo, Y.; Yang, X.;
Hakuna, L.; Barve, A.; Escobedo, J. O.; Lowry, M.; Strongin, R. M.
Sensors 2012, 12, 5940. (c) Shao, N.; Jin, H.; Wang, H.; Zheng, J.;
Yang, R.; Chan, W.; Abliz. J. Am. Chem. Soc. 2010, 132, 725.
(5) (a) Zhang, Z.; Achilefu, S. Org. Lett. 2004, 6, 2067.
(b) Narayanan, N.; Patonay, G. J. Org. Chem. 1995, 60, 2391.
7025
dx.doi.org/10.1021/ja500962u | J. Am. Chem. Soc. 2014, 136, 7018−7025