UC Riverside
UC Riverside Previously Published Works
Title
A membrane-activatable near-infrared fluorescent probe with ultra-photostability for
mitochondrial membrane potentials.
Permalink
https://escholarship.org/uc/item/7gg0k6sp
Journal
The Analyst, 141(12)
ISSN
0003-2654
Authors
Ren, Wei
Ji, Ao
Karmach, Omran
et al.
Publication Date
2016-06-01
DOI
10.1039/c5an01860a
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California
Analyst
PAPER
Cite this: Analyst, 2016, 141, 3679
Received 9th September 2015,
Accepted 28th October 2015
DOI: 10.1039/c5an01860a
www.rsc.o rg/analys t
A membrane-activatable near-infrared fluorescent
probe with ultra-photostability for mitochondrial
membrane potentials†
Wei Ren,‡
a
Ao Ji,‡
a
Omran Karmach,
b
David G. Carter,
c
Manuela M. Martins-Green*
b
and Hui-wang Ai*
a
Mitochondrial membrane potential (MMP) is a frequently used indicator for mitochondrial function.
Herein, we report a photostable near-infrared (NIR) fluorescent dye for monitoring MMP. This new probe,
named NIMAP, is non-fluorescent in aqueous solution and can be activated by cell membranes, providing
high fluorescence contrast and low background fluorescence. NIMAP has been validated for monitoring
MMP in living mammalian cells and in mice. Due to the large fluorescence response, low fluorescence
background, high photostability, and excellent tissue penetration resulting from red-shifted excitation and
emission in the “optical window” above 600 nm, broad applications of this new probe are expected.
Introduction
Mitochondrial membrane potential (MMP) is required for cel-
lular respiration and ATP synthesis.
1
It is also important for
intracellular calcium dynamics, the production of reactive
oxygen species, neural synapse, inflammation, cell prolifer-
ation, and cell death.
2–6
A direct monitoring of MMP is, there-
fore, of high interest. Previously, tetraphenylphosphonium
(TPP
+
)-sensitive microelectrodes have been used to measure
the membrane potentials of isolated mitochondria or mito-
chondria in permeabilized cells.
7
Measuring absolute MMP in
living cells is technically challenging.
8
There are only few
reports, which used end-point measurements based on radio-
isotope tracers to assay absolute MMP in suspensions of living
cells.
9,10
On the other hand, fluorescent MMP sensors are
popular for assessing MMP in living cells,
11
despite that only
qualitative or semi-quantitative information can typically be
derived from those imaging experiments. Fluorescent MMP
dyes are mostly lipophilic cationic compounds. When applied
to healthy cells, they tend to accumulate in the inner mito-
chondrial inner membrane, because healthy cells maintain
cytoplasmic and mitochondrial transmembrane potentials to
induce a negatively charged mitochondrial inner membrane.
12
The distribution of these probes is affected by the transmem-
brane electric fields, following the Nernst equation.
13
There-
fore, the observed mitochondrial fluorescence intensity can be
utilized as an indicator for MMP, although other factors, such
as the cytoplasmic transmembrane potential, the volume ratio
of the mitochondria and a whole cell, and the participation
coefficients of dyes in mitochondria and the cytosol, also have
an impact on the mitochondrial accumulation of these
cationic molecules.
Fluorescent probes with excitation and emission in the
600–900 nm spectral region are desirable for mammalian
studies, due to the weak absorbance of hemoglobin, myoglo-
bin, melanin, and water in this NIR optical window.
14,15
More-
over, compared to light at shorter wavelengths, light with
wavelengths matching this optical window overlaps less with
the excitation of endogenous chromophores, such as flavins
and NADH.
16
These lower-energy photons also cause less
photodamage to living cells and tissues.
17
Not surprisingly, a
current trend is to develop novel fluorescent dyes and sensors
that fluoresce in this NIR spectral region, such as recently
reported NIR dyes that are sensitive to cytoplasmic membrane
potentials or that can highlight the mitochondrial structure in
live cells.
18,19
Common MMP dyes,
11
such as TMRM (tetra-
methylrhodamine methyl ester), TMRE (tetramethylrhodamine
ethyl ester), Rhodamine 123, and JC-1, and a recent TPE-indo
dye
20
all have fluorescence excitation and emission peaks less
than 600 nm. Other concerns of these existing MMP probes
include nonspecific background fluorescence, poor locali-
zation, high cytotoxicity, and unsatisfactory photostability.
11
† Electronic supplementary information (ESI) available: Procedures for chemical
synthesis, spectroscopic characterization, experimental optimization and com-
parison, and time-lapse movies. See DOI: 10.1039/c5an01860a
‡ These two authors contributed equally to this work.
a
Department of Chemistry, University of California, Riverside, CA 92521, USA.
E-mail: huiwang.ai@ucr.edu
b
Department of Cell Biology and Neuroscience, University of California, Riverside,
CA 92521, USA. E-mail: manuela.martins@ucr.edu
c
Institute for Integrative Genome Biology, University of California, Riverside,
CA 92521, USA
This journal is © The Royal Society of Chemistry 2016 Analyst,2016,141,3679–3685 | 3679
Published on 28 October 2015. Downloaded by University of California - Riverside on 12/9/2019 11:35:01 PM.
View Article Online
View Journal
| View Issue
Herein, we report the development of a NIR probe (designated
NIMAP) for monitoring MMP in living mammalian cells and
mice. NIMAP is highly attractive, because not only does it
command the above-mentioned advantages of a NIR probe,
but it also shows a large fluorescence response to MMP
changes, membrane-activated fluorescence and minimal
background signals, ultra-photostability, and excellent tissue
penetration in live animals.
Experimental
Materials and general methods
All chemicals and reagents were purchased from Fisher Scien-
tific (Pittsburgh, PA) or Sigma-Aldrich (St. Louis, Mo) unless
specified elsewhere. Plasmid DNA was purified using Syd Labo-
ratories Miniprep columns (Malden, MA). The scheme and
procedure to synthesize NIMAP are outlined in the ESI.† The
structure and purity of NIMAP were analyzed with ESI-MS,
1
H NMR, and
13
C NMR. ESI-MS was run on an Agilent LC-TOF
system by direct fusion. NMR spectra were recorded on a
Varian Inova 400 instrument with chemical shifts relative to
tetramethylsilane. Stock solutions of NIMAP were made in
DMSO (Sigma-Aldrich).
Spectroscopic characterization
A monochromator-based Synergy Mx Microplate Reader
(BioTek, Winooski, VT) was used to record all spectra. The
absorption spectrum of NIMAP was measured by preparing
NIMAP (50 μM) in a 1 : 1 (v/v) mixture of methanol and Tris-
HCl aqueous buffer (50 mM, pH 7.4). Absorbance values of
NIMAP at 600 nm was measured at various pH conditions by
preparing a series of 1 : 1 (v/v) mixtures of methanol and
aqueous solutions containing citric acid (200 mM) and phos-
phate (200 mM) with a pH range from 3 to 11. To measure the
fluorescence excitation and emission spectra of NIMAP, we
prepared NIMAP in liposomes as described elswhere.
21
L-α-Lecithin egg yolk and cholesterol were purchased from
EMD Millipore (Billerica, MA) and Alfa Aesar (Ward Hill, MA),
respectively. Cholesterol (2.5 mg) and
L-α-Lecithin egg yolk
(10 mg) in chloroform : methanol (v/v 3 : 2) were vacuumed on
a rotary evaporator for 3 h to form a thin, oily film on the glass
wall of the flask. Next, double distilled water (ddH
2
O, 1 mL)
was added and the rotation of the rotary evaporator continued
for another 15 min to form an oil/water mixture in the flask.
The resulting mixture was then sonicated 5 min on a probe
sonicator to yield a slightly hazy and transparent solution.
Formed liposomes were next diluted with equal volume
phosphate buffered saline (PBS, pH 7.4). NIMAP was added to
a final concentration of 100 μM. The mixture was incubated at
room temperature for 15 min. To record the emission
spectrum, the excitation wavelength was set at 630 nm, and
the emission scanned from 650 nm to 870 nm. To record the
excitation spectrum, the emission wavelength was set at
730 nm, and the excitation scanned from 500 nm to 700 nm.
To determine the responsiveness of NIMAP fluorescence to pH
changes, we mixed NIMAP (50 μM) with a series of 9 : 1 (v/v)
mixtures of glycerol and the aforementioned aqueous solu-
tions with a pH range from 3 to 11. Fluorescence at 710 nm
was recorded with the excitation wavelength set at 650 nm.
Mammalian cell culture and dye loading
Human embryonic kidney (HEK) 293T cells or cervical cancer
HeLa cells were maintained in T25 flasks containing 5 mL
Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented
with 10% fetal bovine serum (FBS) and incubated at 37 °C
with 5% CO2 in humidified air. Cells at 80% confluence were
passaged into 35 mm culture dishes in a ratio of 1 : 5 or 1 : 20
for transfection or direct dye loading on the following day.
To stain cells, fluorescent dyes were added into the cell culture
medium, and the incubation condition was typically 45 min
at 37 °C, unless stated otherwise.
Mammalian cell transfection
Transfection complexes were prepared by mixing DNA and PEI
(polyethylenimine, linear, 25 kD; DNA : PEI [w/w] = 1 : 2.5) in
Opti-MEM at room temperature for 20 min. For every 35 mm
culture dish, 10 μL PEI (1 μg μL
−1
) was used to prepare a
500 μL transfection mixture. We used a pcDNA3 plasmid
harboring a mitochondria-localized mW asabi (Mito-mWasabi).
22
The prepared transfection mixture was then added to cells and
incubated for 2 h at 37 °C. Next, pre-warmed, fresh DMEM
containing 10% FBS was used to replace the transfection
medium. On the following day, cells were stained with NIMAP
before imaging.
Fluorescence imaging
Fluorescence microscopy was performed with a Leica SP5
inverted confocal fluorescence microscope with the spectral
imaging capability (Leica, Boston, MA), unless otherwise
stated. The instrument was equipped with a 488 nm laser
(1.5 mW at full power), a 514 nm laser (2 mW at full power),
and a 633 nm laser (1 mW at full power). A 40× water lens was
used for all imaging studies. Unless otherwise noted, NIMAP
was imaged using a 633 nm laser line at 15% laser power for
excitation, and the gain was set to 800 mV. The emission was
set from 650 nm to 800 nm. To record the emission spectrum
for NIMAP loaded into live-cell mitochondria, the emission
bandwidth was set at 10 nm and images were taken from
650 nm to 790 nm. To image mWasabi GFP, the excitation was
set at 488 nm with 10% laser power, and the gain was set to
900 mV. The emission channel was set from 495 nm to
550 nm. To image Rhodamine 123, the excitation was set at
514 nm with 15% laser power, and the gain was set to 900 mV.
The emission was set from 525 nm to 570 nm. Time-lapse
images were acquired every minute. After the initial few
frames, oligomycin (MP Biomedicals, Solon, OH) was added to
a final concentration of 1 μgmL
−1
. Several additional consecu-
tive images were acquired before adding FCCP (Enzo Life
Sciences, Farmingdale, NY) at a final concentration of 100 nM.
Additional frames were recorded. Similarly, we added H
2
O
2
(0.2%) to fresh HEK 293T cells loaded with NIMAP. Images
Paper Analyst
3680
| Analy st,2016,141,3679–3685 This journal is © The Royal Society of Chemistry 2016
Published on 28 October 2015. Downloaded by University of California - Riverside on 12/9/2019 11:35:01 PM.
View Article Online
were captured at 1 min intervals to observe the changes in
mitochondrial membrane potential and morphology. Cell
images were processed with Leica Application Suite (LAS) or
ImageJ.
Mouse imaging
Human prostate cancer PC3 cells were cultured as previously
described.
23
At 100% confluency, 4 × 10
6
cells per group were
stained with NIMAP (10 μM) or Rhodamine 123 (10 μM) for 1 h
at 37 °C. After staining with NIMAP, cells were washed once
with 5 ml of PBS. For collection, cells were suspended in the
culture medium and centrifuged at 3000 rpm for 2 min. Cells
were then resuspended in PBS and centrifuged at 3000 rpm for
2 min. The staining of cells with Rhodamine 123 was similar,
except two additional washes were carried out prior to the col-
lection of cells in order to remove excess Rhodamine 123.
Briefly, cells were allowed to incubate with the fresh culture
medium at 37 °C for 1 h, and this process was repeated before
collection of cells. Next, 4 × 10
6
cells were resuspended in
100 μL PBS and injected subcutaneously into the back of
immunocompromised NOD scid gamma mice (Jackson Labora-
tory, Stock # 005557) after air removal. Mice were imaged
several times over a week, using a Luminescence Dark Box
equipped with light-emitting diodes (LED) for excitation. To
record the fluorescence of NIMAP, we used a 630 nm LED
array, a 650/40 nm bandpass excitation filter, and a 695/30 nm
bandpass emission filter. To record the fluorescence of Rhoda-
mine 123, we used a 490 nm LED array, a 500/40 nm bandpass
excitation filter, and a 550/30 nm bandpass emission filter.
Images were processed with ImageJ.
Results and discussion
Designing and preparation of NIMAP
Our development of the new NIR MMP probes originated from
a Förster resonance energy transfer (FRET) experiment in our
laboratory, which used a QSY-21 core structure as a dark
quencher.
24
QSY-21 (1 in Fig. 1) is essentially nonfluorescent
(zero quantum yield) and has a high extinction coefficient
(90 000 M
−1
cm
−1
at 661 nm).
25
It has been widely utilized as
an efficient nonfluorescent FRET acceptor for far-red and NIR
fluorophores. We prepared a QSY-21 labeled biotin derivative
(2), and observed mitochondria-like organellar fluorescence in
cultured mammalian cells stained with 100 μM of the dye (ESI
Fig. S1A†). In contrast, incubating HEK 293T cells with a
neutral zwitterionic sulfonic acid form of QSY-21 (3) did not
induce significant NIR mitochondrial fluorescence (Fig. S1B†).
Unstained HEK 293T cells also had no autofluorescence at this
spectral region (Fig. S1C and S1D†). Considering the facts that
QSY-21 labeled biotin has some degrees of lipophilicity and a
positive charge, we reasoned that it was partially sequestered
into the lipid bilayers of cells, followed by migration into mito-
chondria and fluorescence activation in this hydrophobic,
viscous microenvironment.
In light of the potential use of QSY-21 derivatives as mito-
chondrial dyes, we designed a new molecule (4) containing a
hexanoamide connected to the QSY-21 core through a pipera-
zine sulfonamide linker. A short C6 fatty acid chain was
chosen to enhance the lipophilicity while retaining adequate
solubility of the resultant molecule in neutral aqueous solu-
tions for dye loading. This molecule was designated “NIMAP”
for its NIR, mitochondria-activatable property. A synthetic
route was developed to prepare NIMAP in 7 steps with 18%
overall yield from commercially available, inexpensive starting
materials (ESI Scheme S1†).
Optimization of NIMAP loading into mammalian cells
We next loaded various concentrations of NIMAP into HEK
293T cells. No obvious cell death was observed upon 24 h
incubation with up to 100 μM NIMAP. We also studied the
impacts of dye concentrations and loading time on the
accumulation of NIMAP in mitochondria at 37 °C. To keep the
loading period as short as 45 min, 1–5 μM NIMAP in the cell
culture medium was needed to induce strong mitochondrial
fluorescence detectable by confocal microscopy with reason-
able instrumental settings (Fig. S2A†). In comparison, 100 μM
2 was needed to gain similar images (Fig. S1A†). These results
indicate that the fusion of the QSY-21 core with a C6 fatty acid
chain increases its partition into the lipid phase and its
accumulation in mammalian mitochondria.
The length of the loading period also affected the locali-
zation of NIMAP. With 1 μ M NIMAP, the mitochondrial fluo-
rescence continuously increased in a 2 h test window
(Fig. S2B†), indicating that an equilibrium state for dye distri-
bution was not reached under these conditions. At 2 h, NIMAP
started to saturate mitochondria to also stain cytoplasmic
membrane (Fig. S2C†). Due to the common practice favoring
short loading time, we used 1 μM NIMAP and a 45 min dye
incubation time in most of our following cell studies. No wash
step was needed to remove excess NIMAP in these experiments
because NIMAP was nonfluorescent in aqueous solutions. By
comparison, the background fluorescence of HEK 293T cells
stained with a conventional MMP dye, Rhodamine 123, was
strong before washing off
excess dye molecules (Fig. S3†). We
also noticed a drastic difference of NIMAP and Rhodamine
123 in terms of their photostability (Fig. 2). Almost no photo-
Fig. 1 Chemical structures of the dark quencher QSY-21 (1), a QSY-21
labeled biotin (2), a sulfonic acid form of QSY-21 (3), and NIMAP (4).
Analyst Paper
This journal is © The Royal Society of Chemistry 2016 Analyst,2016,141,3679–3685 | 3681
Published on 28 October 2015. Downloaded by University of California - Riverside on 12/9/2019 11:35:01 PM.
View Article Online
bleaching was observed for NIMAP under the highest laser
power (100% of 1 mW at 633 nm), whereas Rhodamine 123-
stained cells photobleached quickly under a similar condition
(50% of 2 mW at 514 nm). The excellent photostability of
NIMAP makes it suitable for long-term time-lapse imaging.
Mitochondrial localization of NIMAP
To further confirm the mitochondrial localization of NIMAP,
we loaded the molecule into HEK 293T cells simultaneously
expressing a green fluorescent mWasabi protein fused to a
mitochondrial localization tag (Mito-mWasabi).
22
Dual colour
imaging verified the colocalization of mWasabi and NIMAP in
the mitochondria (Fig. 3). The ability to use NIMAP and a
green fluorescent probe for dual color imaging is important
because many existing, popular fluorescent sensors are based
on GFP, whereas the fluorescence excitation or emission of
common MMP probes, such as Rhodamine 123 and JC-1,
11
is
partially overlapped with that of GFP.
Fluorescence activation of NIMAP by biolipids
To determine whether the observed mitochondrial fluo-
rescence was due to the activation of NIMAP by the lipid mem-
brane, we prepared NIMAP-containing liposomes, following a
previously reported liposome-synthetic procedure.
21
The fluo-
rescence of NIMAP-containing liposomes was high (Fig. 4).
The excitation maximum is 650 nm and the emission
maximum is 710 nm. In comparison, NIMAP alone was essen-
tially non-fluorescent in aqueous solutions. Moreover, the
emission profile of NIMAP in liposomes was similar to that of
NIMAP in live-cell mitochondria (Fig. S4A†). A previous mole-
cular dynamics (MD) study suggests that QSY-21 fluorescence
is quenched by ring rotations and electron transfer in the
excited state.
26
It is possible that the packed, viscous lipid
microenvironment may restrict the rotations and electron
transfer of NIMAP to enhance its fluorescence. We also
measured the absorbance of NIMAP in an aqueous solution,
which was similar to the profile of NIMAP fluorescence exci-
tation in liposomes (Fig. 4). Moreover, the absorbance was
insensitive to pH changes from pH 4 to 10 (Fig. S4B†). We
further characterized the in vitro fluorescence of NIMAP in a
series of aqueous buffers containing 90% (v/v) glycerol, and
the fluorescence intensities were essentially unchanged from
pH 4 to 10. These results are aligned with the conventional use
of QSY-21 as a dark quencher and our new finding that the
fluorescence of QSY-21 is greatly enhanced upon being inte-
grated into the mitochondrial membrane.
Fig. 2 Photostability of NIMAP (red) and rhodamine 123 (yellow) in HEK
293T cells under a confocal microscope.
Fig. 3 Colocalization of mitochondrial mWasabi (green channel) and
NIMAP (NIR channel) in HEK 293T cells (scale bar: 30 μm).
Fig. 4 Fluorescence excitation (open circle) and emission (closed
circle) spectra of NIMAP in liposomes (red) or an aqueous solution
(blue). The normalized absorbance of NIMAP in the aqueous solution is
also shown (gray line).
Paper Analyst
3682
| Analy st,2016,141,3679–3685 This journal is © The Royal Society of Chemistry 2016
Published on 28 October 2015. Downloaded by University of California - Riverside on 12/9/2019 11:35:01 PM.
View Article Online