Fluorescence ratiometry and fluorescence lifetime imaging: Using a single molecular sensor for dual mode imaging of cellular viscosity

Article abstract of DOI:10.1021/ja1104014

Intracellular viscosity strongly influences transportation of mass and signal, interactions between the biomacromolecules, and diffusion of reactive metabolites in live cells. Fluorescent molecular rotors are recently developed reagents used to determine the viscosity in solutions or biological fluid. Due to the complexity of live cells, it is important to carry out the viscosity determinations in multimode for high reliability and accuracy. The first molecular rotor (RY3) capable of dual mode fluorescence imaging (ratiometry imaging and fluorescence lifetime imaging) of intracellular viscosity is reported. RY3 is a pentamethine cyanine dye substituted at the central (meso-) position with an aldehyde group (CHO). In nonviscous media, rotation of the CHO group gives rise to internal conversion by a nonradiative process. The restraining of rotation in viscous or low-temperature media results in strong fluorescence (6-fold increase) and lengthens the fluorescence lifetime (from 200 to 1450 ps). The specially designed molecular sensor has two absorption maxima (λ_abs 400 and 613 nm in ethanol) and two emission maxima (in blue, λ_em 456 nm and red, 650 nm in ethanol). However it is only the red emission which is markedly sensitive to viscosity or temperature changes, providing a ratiometric response (12-fold) as well as a large pseudo-Stokes shift (250 nm). A mechanism is proposed, based on quantum chemical calculations and ¹H NMR spectra at low-temperature. Inside cells the viscosity changes, showing some regional differences, can be clearly observed by both ratiometry imaging and fluorescence lifetime imaging (FLIM). Although living cells are complex the correlation observed between the two imaging procedures offers the possibility of previously unavailable reliability and accuracy when determining intracellular viscosity.

Full text of DOI:10.1021/ja1104014

ARTICLE
pubs.acs.org/JACS
Fluorescence Ratiometry and Fluorescence Lifetime Imaging: Using a
Single Molecular Sensor for Dual Mode Imaging of Cellular Viscosity
Xiaojun Peng,*,† Zhigang Yang,† Jingyun Wang,‡ Jiangli Fan,† Yanxia He,† Fengling Song,† Bingshuai Wang,†
Shiguo Sun,† Junle Qu,§ Jing Qi,§ and Meng Yan§
†State Key Laboratory of Fine Chemicals, ‡School of Life Science and Biotechnology, Dalian University of Technology, 2 Linggong Road,
Hi-tech Zone, Dalian 116024, P.R. China
§Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education, Shenzhen University, Shenzhen 518060, P.R. China.
S Supporting Information
b
ABSTRACT: Intracellular viscosity strongly influences trans-
portation of mass and signal, interactions between the bioma-
cromolecules, and diffusion of reactive metabolites in live cells.
Fluorescent molecular rotors are recently developed reagents
used to determine the viscosity in solutions or biological fluid.
Due to the complexity of live cells, it is important to carry out
the viscosity determinations in multimode for high reliability
and accuracy. The first molecular rotor (RY3) capable of dual
mode fluorescence imaging (ratiometry imaging and fluores-
cence lifetime imaging) of intracellular viscosity is reported.
RY3 is a pentamethine cyanine dye substituted at the central (meso-) position with an aldehyde group (CHO). In nonviscous media, rotation
of the CHO group gives rise to internal conversion by a nonradiative process. The restraining of rotation in viscous or low-
temperature media results in strong fluorescence (6-fold increase) and lengthens the fluorescence lifetime (from 200 to 1450 ps).
The specially designed molecular sensor has two absorption maxima (λabs 400 and 613 nm in ethanol) and two emission maxima (in
blue, λem 456 nm and red, 650 nm in ethanol). However it is only the red emission which is markedly sensitive to viscosity or
temperature changes, providing a ratiometric response (12-fold) as well as a large pseudo-Stokes shift (250 nm). A mechanism is
1
proposed, based on quantum chemical calculations and H NMR spectra at low-temperature. Inside cells the viscosity changes,
showing some regional differences, can be clearly observed by both ratiometry imaging and fluorescence lifetime imaging (FLIM).
Although living cells are complex the correlation observed between the two imaging procedures offers the possibility of previously
unavailable reliability and accuracy when determining intracellular viscosity.
’ INTRODUCTION
rotation can be restrained by viscous environments, following
which the fluorescence intensity or lifetime of dyes will be seen to
increase.
At the intracellular level, viscosity strongly influences intracel-
lular transportation of mass and signal, interactions between
biomacromolecules, and diffusion of reactive metabolites such as
ROS and RNS. Changes in viscosity at cellular level relate to a
number of diseases and pathologies.1,2 The existing mechanical
methods—such as capillary viscometer, falling ball viscometer,
and rotational viscometer for viscosity determination3—cannot
provide measures of viscosity at a cellular level. In response to
this, Luby-Phelps4 reported a ratiometric system of fluorescent
Cy5 and Cy3 attached to macromolecular Ficoll 70 to measure
intracellular viscosity. However, small molecule fluorescent
sensors with cell-membrane permeability should be much more
effective in the investigation of intracellular microviscosity.
Recently developed molecular rotors are a group of fluores-
cent molecules that can be used to determine the viscosity in
solutions or biological fluids. Such molecular rotors typically
comprise a conjugated domain that can freely rotate in non-
viscous or low-viscosity solutions, providing relaxation of the
electronically excited dyes by nonradiative processes.5ꢀ10 Such
Fluorescence intensity-based measurements, however, are
influenced by fluid optical properties, dye concentration, and
other experimental or instrumental factors. A ratiometric ap-
proach, using probes that possess dual emission maxima from
which the intensity ratio can be obtained, avoids most of the
interferences11,12 and although influenced by viscosity the ratios
are independent of the dye concentration.6,7,9 The fluorescence
lifetimes of dyes are also independent of the interfering factors
noted above. Moreover they do not even require calculation of
the ratio, and offer easy system calibration together with ultra-
sensitive detection.8 So fluorescence ratiometry and FLIM are
ideal procedures for intracellular viscosity investigations.
Theodorakis6,7 and Suhling8ꢀ10,13 have reported on some fluor-
escent molecular rotors and have made good progress regarding
Received: November 28, 2010
Published: April 08, 2011
r
2011 American Chemical Society
6626
dx.doi.org/10.1021/ja1104014 J. Am. Chem. Soc. 2011, 133, 6626–6635
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Structures of RY3 and Related Compounds
where Φ represents quantum yield; F stands for integrated area under
the corrected emission spectrum; A is absorbance at the excitation
wavelength; λex is the excitation wavelength; n is the refractive index of
the solution (because of the low concentrations of the solutions
(10ꢀ7ꢀ10ꢀ8 mol/L), the refractive indices of the solutions were
replaced with those of the solvents); and the subscripts x and s refer
to the unknown and the standard, respectively.
Viscosity Determination, Fluorescence Spectral Measure-
ment, and Fluorescence Lifetime Detection. The solvents were
obtained by mixing ethanol-glycerol, methanol-glycerol, and deionized
water-glycerol systems in different proportions. Measurements were
carried out with a NDJ-7 rotational viscometer, and each viscosity value
was recorded. The solutions of RY1-RY5 of different viscosity were
prepared by adding the stock solution (1.0 mM) to 10 mL of solvent
mixture (ethanol-glycerol, water-glycerol, and methanol-glycerol sol-
vent systems) to obtain the final concentration of the dye (1.0 μM).
These solutions were sonicated for 5 min to eliminate air bubbles. After
standing for 1 h at a constant temperature, the solutions were measured
in a UV spectrophotometer and a fluorescence spectrophotometer.
A fluorescence lifetime measuring equipment (Horiba Jobin Yvon
Fluoromax-4p) was used to obtain the fluorescence lifetimes of RY3,
with the excitation wavelength at 376 nm and detection at 650 nm.
Low Temperature 1H NMR and Fluorescence Spectra. The
temperatures of solutions were slowly lowered from 25 °C by liquid
nitrogen to 0 °C, ꢀ25 °C, and ꢀ50 °C. The 1H NMR spectra of RY1 and
RY3 in CDCl3 were obtained using a nuclear magnetic resonance spectro-
meter (Varian INOVA 400 MHz). Fluorescence spectra of RY1-RY4 were
measured on a FP-6500 spectrophotometer (Jasco, Japan), with the
excitation wavelength fixed at 600 nm. RY3 (1.0μM) in 70% glycerolꢀwater
mixture was measured on the fluorescence spectrophotometer at ꢀ5 °C,
0 °C, 5 °C...40 °C with excitation wavelength of 400 nm, and the
fluorescence lifetimes of RY3 at the same temperature were measured.
Cell Incubation and Imaging. PC12, MCF-7, and NIH-3T3 cells
were cultured in DEME (Invitrogen) supplemented with 10% FCS
(Invitrogen). One day before imaging, cells were seeded into 24-well flat-
bottomed plates. The next day, the cells were incubated with 5.0 μM dye for
0.5 h at 37 °C under 5% CO2 and washed with phosphate-buffered saline
(PBS) three times. Fluorescence imaging of PC12 cells was carried out using
a Nikon Eclipse TE2000ꢀ5 inverted fluorescence microscope with a 10ꢁ
objective lens (excited with green light). MCF-7 cells were incubated with
RY3 and observed under a Leica TCS-SP2 confocal fluorescence micro-
scope, 100ꢁ objective lens. Then the fluorescence intensity and fluorescence
lifetime of RY3 in NIH-3T3 cells were observed under a Leica TCS-SP2
confocal fluorescence microscope and TCSPC FLIM equipment (SPC150)
at 4 °C, 20 °C, and 37 °C.
the investigation of intracellular viscosity. In practical applica-
tions, however it is important that intracellular viscosity detec-
tion can use both fluorescence ratiometry and lifetime detection,
to provide high reliability and accuracy. Up to now, however,
there is no such dual-mode molecular sensor could be used to
detect viscosity and map intracellular viscosity by both ratio-
metric imaging and FLIM.
Here we report a fluorescent molecular rotor, RY3 (Scheme 1,
X = CHO). This provides two emission peaks: one in the blue
region (λem 456 nm, in EtOH), and one in the red (λem 650 nm,
in EtOH). Fluorescence is very weak in nonviscous environ-
ments. However on increasing viscosity, the red emission is
greatly enhanced while the blue emission remains insensitive,
providing the basis for a ratiometric procedure. In parallel to this,
the fluorescence lifetime for the red emission is also markedly
prolonged in viscous media, from 200 ps to more than 1400 ps.
RY3 is the first fluorescent molecule which can be used as dual-
mode molecular fluorescence sensor in both fluorescence ratio-
metry and lifetime fluorescence detection.
’ EXPERIMENTAL SECTION
Sample Preparation. 2, 3, 3-trimethyl-3H-indole, indolenium
quaternized salts14 and other intermediates15,16 were prepared accord-
ing to literature methods. The preparation procedures of RY1-RY5 and
related intermediates are given in the Supporting Information (SI). All
solvents and reagents used were reagent grade. All reactions were carried
out under a nitrogen atmosphere with dry, freshly distilled solvents
under anhydrous conditions. Silica gel (100ꢀ200 mesh) was used for
flash column chromatography for purifications. Bovine Serum Albumin
(BSA) was purchased from Shanghai Sangon Biotech Co., Ltd. Water
used in all experiments was doubly purified by Milli-Q Systems equip-
ment. The solutions of RY1ꢀRY5 were typically prepared from 1.0 mM
stock solutions in DMSO.
Titration by Bovine Serum Albumin (BSA) and Cysteine.
The stock solution of RY3 (1.0 mM) was diluted by 20.0 mM phosphate
buffer solution (PBS, pH 7.4) to obtain a solution with concentration of 2.0
μM. BSA was dissolved in doubly purified water with a concentration of 50.0
mg/mL. The BSA solution was gradually added to the RY3 solution, over the
range 0ꢀ2.0 mg/mL. The solution obtained was left to stand to eliminate air
bubbles, and measured in the UV spectrophotometer and fluorescence
spectrophotometer. When finishing the titration with a BSA concentration of
2.0 mg/mL, the solution was stood at 37 °C for 120 min, then the absorption
and fluorescence spectra were measured on the spectrometer every 10 min to
examine the spectra changes of RY3 upon addition of BSA.
Quantum Calculations. All the quantum-chemical calculations
were done with the Gaussian 09 suite.17 The parameter referred to the
work of Han.18 The geometry optimizations of the dyes were performed
using density functional theory (DFT)19 with Becke’s three-parameter
hybrid exchange function with LeeꢀYangꢀParr gradient-corrected
correlation functional (B3-LYP functional) and 6-31G ** basis set. No
constraints to bonds/angles/dihedral angles were applied in the calcula-
tions and all atoms were free to optimize. The electronic transition
energies and corresponding oscillator strengths were calculated with
time-dependent density functional theory (TD-DFT)20,21 at the
B3LYP/6-31G ** level.
The similar titrations of RY3 (2.0 μM) by cysteine (25.0 mM) was
carried out under the same conditions as that using BSA, the final
concentration of cysteine being 200 μM. After the titration, the solution
was stood at 37 °C for 120 min, then the absorption and fluorescence
spectra were measured on the spectrometer every 10 min to check for
any spectral changes of RY3 upon addition of cysteine.
Absorption and Fluorescence Quantum Yields Measure-
ments. Absorption spectra were measured on a Lamda LS35 spectro-
photometer. Fluorescence spectra were obtained with a FP-6500
spectrophotometer (Jasco, Japan). The relative fluorescence quantum
yields were determined with Rhodamine B as a standard22 and calculated
using the following equation:
Photostability. RY1-RY4 were dissolved in DMSO-water (5:5 v/v)
at a concentration of 10.0 μM, respectively. The solutions were irradiated
2
Φx ¼ ΦsðFx=FsÞðAs=AxÞðλexs=λexxÞðnx=nsÞ
6627
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635

Journal of the American Chemical Society
ARTICLE
Table 1. Spectral Data of the Dye RY1ꢀRY4 in Different Solvents
dyes
solvents
dielectric constant
ηa (cP)
λabsb,c (nm)
λemb,d (nm)
εb,e
2.27
Φfb,f
RY1
THF
7.43
0.53
0.43
1.20
0.59
2.00
1.01
653
661
652
649
655
648
662
671
663
659
668
656
0.37
DCM
8.93
3.06
3.02
2.59
2.72
2.27
0.29
0.27
0.22
0.33
0.13
ethanol
methanol
DMSO
H2O
24.85
32.61
46.83
78.36
RY2
THF
7.43
0.53
0.43
1.20
0.59
2.00
1.01
657
663
654
649
656
648
661
669
659
656
663
654
1.86
2.00
2.00
1.97
1.76
1.36
0.12
0.11
0.10
0.09
0.25
0.05
DCM
8.93
ethanol
methanol
DMSO
H2O
24.85
32.61
46.83
78.36
RY3
THF
7.43
0.53
0.43
1.20
0.59
2.00
1.01
(399) 617
(399) 624
(400) 613
(400) 609
(400) 613
(413) 600
(400)606
(399)604
(398)608
(455) 651
(454) 649
(456) 650
(450) 648
(451) 630
(460) 638
(453)654
(454)652
(461)650
(0.16) 1.02
(0.11) 1.05
(0.15) 1.10
(0.13) 1.11
(0.14) 1.10
(0.15) 0.77
(ꢀ)1.04
(0.006) 0.03
(0.006) 0.04
(0.005) 0.03
(0.004) 0.02
(0.010) 0.07
(0.002) 0.02
(ꢀ) 0.067
DCM
ethanol
methanol
DMSO
H2O
A1 g
A2 g
8.93
24.85
32.61
46.83
78.36
(ꢀ)1.02
(ꢀ) 0.059
A3 g
45.8
950
(ꢀ)0.83
(ꢀ) 0.181
RY4
THF
7.43
0.53
0.43
1.20
0.59
2.00
1.01
690
695
684
681
688
680
706
708
704
702
710
697
1.2
2.4
2.1
2.2
2.1
1.5
0.09
0.11
0.07
0.05
0.17
0.04
DCM
8.93
ethanol
methanol
DMSO
H2O
24.85
32.61
46.83
78.36
a The viscosity data of the solvents were measured at 25 °C. b There are two group of absorption and emission peaks for RY3; and the data in the
parentheses are the absorption, emission wavelengths, molar extinction coefficient, and fluorescence quantum yields of RY3 at the second peak. c The
e
absorption peaks of dyes (nm). d RY3 was excited at 400 nm, and other dyes were excited at their maximum absorption wavelengths. ꢁ105
molꢀ1cmꢀ1L. f Rhodamine B was used as a standard reference with a fluorescence quantum yield of 0.97 in ethanol, errors were estimated to be 10% for
fluorescence quantum yield. g The fluorescence quantum yields of RY3 at 650 nm in the mixture solvents and 99% glycerol were just considered; A1,
ethanol-glycerol (1: 1 v/v); A2, methanolꢀglycerol (1: 1 v/v); A3, 99% glycerol.
under a 500 W iodineꢀtungsten lamp for 7 h at a distance of 250 mm away.
An aqueous solution of sodium nitrite (50.0 g/L) was placed between the
samples and the lamp as a light filter (to cut off the light shorter than 400 nm)
and heat filter. The photostabilities were expressed in the terms of remaining
absorption (%) calculated from the changes of absorbance at the absorption
maximum before and after irradiation by iodineꢀtungsten lamp.
chain.23,24 As seen in Table 1, three dyes (RY1, RY2, and RY4)
possess only one group of absorption (λabs) and fluorescence (λem)
peaks: RY1 (λabs 652 nm and λem 663 nm in ethanol), RY2 (λabs 654
and 659 nm in ethanol), and RY4 (λabs 684 nm and λem 704 nm in
ethanol), respectively. However, RY3, with an aldehyde substituent
at the meso-position of the pentamethine chain, has two absorption
peaks (λabs 400 and 613 nm in ethanol) and two emission peaks
(λem 456 and 650 nm in ethanol). Compared with RY1 and RY5
(Figure 1), the spectrum of RY3 has a short wavelength peak similar
to RY5 (λabs 415 and λem 466 nm), and a wavelength peak similar to
RY1. For the red emission RY4, with a conjugated electron-donating
vinyl group, has the longest wavelength; whereas RY3, conjugated
with an electron-withdrawing CHO group, has the shortest wave-
length. The molar extinction coefficient (ε) of RY3 is also the
smallest (about 1.0 ꢁ 105 molꢀ1cmꢀ1L),onlyhalfthatofotherdyes.
So dye RY3 must have a unique emission mechanism. Note that a
low quantum yield corresponds to a low nonsignal background,
when fluorescence enhancement in the target will provide a good
signal-to-noise ratio.
Redox Potential Measurements. The redox potential (cyclic
voltammetry, CV; and differential pulse voltammetry, DPV) was
measured using a BAS 100B electrochemical analyzer. A three-electrode
cell was composed of a glass-carbon as working electrode, a platinum
wire as counter electrode, and Ag/AgCl as reference electrode (10.0 mM
AgNO3). The supporting electrolyte was TBAPF6 (0.1 M tetra n-
butylammonium hexafluorophosphate) and the concentrations of
RY1-RY4 were 1.0 mM in acetonitrile.
’ RESULTS AND DISCUSSION
Spectral Properties. The four pentamethine dyes (RY1-
RY4) have different substitutents (H, Cl, CHO, and CHdCH2
respectively) at the central (meso-) position of the pentamethine
6628
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635

Journal of the American Chemical Society
ARTICLE
Figure 1. The normalized absorption (Left) and emission (Right) spectra of dyes RY1, RY3, and RY5 in 99% glycerol.
Figure 2. The fluorescence spectra of RY3 (A), RY1 (B), and the comparison of fluorescence increase fold of different dyes with the increase of solution
viscosity (C), RY3 was excited at 600 nm and RY1, RY2, and RY4 were excited at the maximum absorption peaks.
Generally solvents with different polarity always influence the
wavelengths and quantum yields to some extent. However for
RY3, the influence was very limited. In a set of dielectric constant
values ranging from 7.43 (THF) to 78.36 (H2O), the quantum
yields of RY3 showed little change (0.03 in THF and 0.02 in
H2O). This means that, in nonviscous solvents, the polarity of
the solvents has little effect on fluorescence quantum yields. This
characteristic is crucial for a molecular rotor to sense environ-
mental viscosity, while excluding the influence of polarity.25
Fluorescence Response to Solvent Viscosity. Unlike its
insensitivity to polarity, the fluorescence of RY3 was more
Figure 3. The linearity of log If versus log η plot of RY3 (1.0 μM) in
ethanol-glycerol mixed solvents, excited at 600 nm.
sensitive to solution viscosity than the other three dyes. By
adding glycerol to the ethanol solution of RY3, the viscosity of
the solution was increased from 1.2 cP (ethanol) to 950 cP (99%
glycerol). As the viscosity increased, fluorescence of RY3 was
greatly enhanced by 6 fold in quantum yield (0.03 in ethanol and
Scheme 2. Hypothetical Rotation of CHO Group
0.181 in 99% glycerol in Table 1), although its absorbance
showed almost no change. With increasing proportions of
glycerol in the mixture solvent, the fluorescence intensity of
RY3 at 650 nm increased continuously (Figure 2A), until close to
the intensity of RY1 (Figure 2B). Other dyes did not show such
fluorescence enhancement under the same conditions: RY1
(0.25 fold), RY2 (1.5 fold), and RY4 (1.9 fold) (Figure 2C).
The quantitative relationship between the fluorescence inten-
sity If of a molecular rotor and the viscosity η of the solvent is well
x = 0.30, λem 650 nm) (Figure 3). RY1, with hydrogen atom at
expressed by the F€orsterꢀHoffmann eq 1:26
the meso-position, has little response (x = 0.04).
The difference in sensitivity may be attributed to the different
log If ¼ C þ xlog η
ð1Þ
substituents (X) at the central position of the polymethine-chain
of the dyes. In the case of RY1 (X = H) the dye has high quantum
yield in most solvents with different viscosity, so this kind of dye
is good for fluorescent labels but not for good viscosity sens-
ing. Where bulky substitutents (RY3 (X = CHO) and RY4
where C is a concentration- and temperature-dependent con-
stant and x is a dye- and temperature-dependent constant. RY3
fits the linear relationship between log If and log η (R2 = 0.991,
6629
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635

Journal of the American Chemical Society
ARTICLE
Figure 4. NMR evidence for restricted rotation of the aldehyde group at low temperature, and the interaction between the oxygen and adjacent
hydrogen atom at pentamethine chain (above), note the split of the H-chemical shift of conjugated chain of RY3 in CDCl3 at different temperature
(0 °C, ꢀ25 °C, and ꢀ50 °C) (below).
0
(X = CHdCH2) are present, then these groups may rotate5ꢀ9
(Scheme 2) or vibrate to change the geometry of the molecule in
the excited state. In nonviscous solution, these changes occur
readily, and permit the excited state to relax by a nonradiative
process, reducing the quantum yield. In viscous solution, how-
ever, this change is restrained, so the fluorescence quantum yield
or intensity should increase. It is possible that the CHO group in
RY3 rotates around the C—C bond more easily than does the
vinyl group in RY4. Alternatively the electron-withdrawing
character of CHO results in ICT (internal charge transfer), so
RY3 has a lower fluorescence background and better enhance-
ment than RY4.
intense than H2/H2 . At higher temperature (0 °C), the free
rotation of the CHO group generates no such split. RY1, even at
low temperature (ꢀ25 °C and ꢀ50 °C), has no split or shift in
1H NMR (SI Figure S2) and little fluorescence change.
Ratiometric Fluorescence of RY3. As illustrated in Figure 1,
RY3 possesses a unique spectral character, with two emission
peaks (such as λem 456 and 650 nm in ethanol). Most impor-
tantly only the red emission of RY3 responded significantly to the
viscosity of the solvent. As show in Figure 5 and SI Figure S3, in
different viscous solvent systems (ethanol-glycerol (Figure 5A),
water-glycerol (Figure 5C) and methanol-glycerol (SI Figure
S3)), the red fluorescence intensity at 650 nm increased sensi-
tively with viscosity of the solvents. The blue emission at 456 nm,
however, gave only very small responses. This permits ratio-
Fluorescence Increase at Low Temperature. Similar to the
restricted rotation of the CHO of RY3 in high-viscous solutions,
lowering the temperature of nonviscous solutions could also
retard the rotation of the CHO group. Chloroform was chosen as
the solvent to avoid interference from hydrogen bonding be-
tween solvent molecules and the CHO group. As shown in SI
Figure S1A, from 25 to ꢀ50 °C, enhancement of fluorescence
intensity of RY3 in chloroform increased nearly 4 fold. Little
enhancements for RY1, RY2, and RY4 were found under the
same conditions (SI Figure S1B).
metric changes with a linear relationship between log(I650/I456
)
and log η. This is fitted by the F€orsterꢀHoffmann equation, R2 of
the linear relation increases in the ethanol-glycerol system
(Figure 5B) and water-glycerol system (Figure 5D) were 0.989
and 0.99, and the slope x are 0.30 and 0.38, respectively. This
demonstrated that RY3 could be employed to ratiometrically
detect the solution viscosity.
When the temperature was reduced from 40 °C to ꢀ5 °C, the
fluorescence intensity at 650 nm was enhanced many fold, while
the blue emission at 456 nm showed little increase (Figure 6A).
Similarly the ratio linearity between log(I650/I456) and tempera-
ture was obtained (Figure 6B). This shows that cooling has the
similar effect as viscosity in restraining the rotation of the
CHO group.
Theoretical Calculation and Mechanism. The structural
optimization of RY3 by DFT/TDDFT in B3LYP/6-31G** level
of Gaussian 09 (the atomic standard orientations after geometry
optimization of ground state and excited state have been given in
The rotation of CHO group can be inferred from the 1H NMR
at low temperature. At ꢀ50 °C, the rotation of the CHO group of
RY3 slows down. This results in a different interaction with its
adjacent hydrogen atom at the pentamethine chain, such as
possible hydrogen bond interaction between the oxygen atom of
0
aldehyde group with H3 shown in Figure 4, where H3 and H3
0
shift to 7.95 and 6.45 from 7.28 ppm (at ꢀ50 °C); H2 and H2 also
shift to08.31 and 8.03 from 8.20 ppm (at ꢀ50 °C). As the shifts of
0
H3/H3 are much larger than H2/H2 , the action of the oxygen
0
atom of the CHO group on spatially fitted H3/H3 is more
6630
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635

Journal of the American Chemical Society
ARTICLE
Figure 5. The emission spectra changes of RY3 (1.0 μM) with the increase of the viscosity of solvent and the linear response between log(I650/I456) and
logη: in the solvent of ethanol-glycerol (A and B) and water-glycerol (C and D), excited at 400 nm.
Figure 6. The fluorescence spectra and ratio of RY3 (1.0 μM) at different temperatures in water-glycerol (30/70, v/v), excited at 400 nm (A); and the
linearity of the plot of log(I650/I456) versus temperature (B).
Supporting Information), is shown in Figure 7. Two main
electronic transitions (S0 f S1, oscillator strength fab = 1.75;
and S0 f S2, oscillator strength fab = 0.15) were obtained,
corresponding to the two absorption peaks (λabs 613 and
400 nm in ethanol). The two main electronic transitions from
excited state (S1 f S0, oscillator strength fem = 2.08; and S2 f S0,
oscillator strength fem = 0.17), corresponded to two emission
(λem 650 and 456 nm in ethanol) peaks of RY3. The short
absorption and emission (λabs 400 nm and λem 456 nm) have
very small oscillator strength f (fab 0.15 and fem 0.17 respectively),
corresponding to the low molar extinction coefficient (ε) and
fluorescence quantum yields (Figure 1 and Table 1).
When RY3 was excited at 400 nm, the dye molecules under-
went a transition to the second excited state (S2), accompanied
byintramolecularcharge transfers (ICT). Thereare two competing
decay processes from the excited state S2. The first is fluorescence
emission where the electron returns to ground state (S0) directly
from S2 with a blue fluorescence emission (λem 456 nm). The
second is internal conversion where the molecule is relaxed to the
first excited state (S1). This second decay mechanism, internal
conversion, is a rapid process (femto- or pico-second) compared
to fluorescence emission (nanosecond). Consequently the most
likely outcome is to transfer to the first excited state by internal
conversion (S2 f S1). According to the classic “energy gap law”
(the smaller the energy gap, the greater the possibility of internal
conversion), internal conversion S2 f S1 is much easier than that
of S2 f S0.
Similarly with S2, the S1 state also decays to S0 via two com-
peting mechanisms, fluorescence emission and internal conver-
sion. In nonviscous solution (Figure 8A), the rotation of the
CHO group raises the vibrational energy, consequently internal
conversion is the favored mechanism, and it dominates the decay.
However in viscous or cooled media (Figure 8B), the rotation is
restricted, so fluorescence emission dominates. This emission
(S1 f S0) contributes to the strong red fluorescence of λem
650 nm occurring in viscous media or at low temperature.
Therefore, a ratio increase (I650/I456) with viscosity or tem-
perature results from the different internal conversions con-
trolled by the energy gap law and influenced by CHO rotation.
The energy gap law allows internal conversion to predominate
6631
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635

Journal of the American Chemical Society
ARTICLE
Figure 7. Frontier molecular orbital (MO) of RY3 calculated with time-dependent density functional theory (TDDFT) at the B3LYP/6-31G** level
(number of states: 6) using Gaussian 09: A, the dihedral angle of O1C1C2C3 is ca. 0°; B, the proposed diagram of the energy levels of RY3; C, the
dihedral angle of O1C1C2C3 is ca. 90°; and D, the dihedral angle of O1C1C2C3 is figured in the structure of RY3.
emission) are prohibited and only one set of absorption and
emission peaks with long wavelength has been observed in the
spectra of RY4.
The above results arise from the key action of CHO group in
RY3. Because of the electron-withdrawing effect, CHO induces
an internal charge transfer process (ICT). The LUMOþ1 of
RY3 (S2, Figure 7A) clearly shows the unequal distribution of
charge. In particular when CHO group rotates, for example twists
to 90 °C where the electron cloud density increases markedly
on the CHO group at the LUMOþ1 level (Figure 7C), twisted
intramolecular charge transfer (TICT) might occur. This would
enlarge the fluctuation in vibration energy level and increase the
possibility of internal conversion which is sensitive to the
viscosity and the temperature of media.5 The low molar extinc-
tion coefficients of RY3 (Table 1) is the further evidence for the
ICT processes.
Figure 8. The schematic illustration of the decay processes of RY3 in
nonviscous (A) and viscous or cooled media (B). IC1 and IC2 are two
internal conversion processes. F2 is the emission of λem 650 nm.
when there is a small energy gap between two energy levels. The
CHO-rotation enhances the internal conversion by raising the
vibration energy. The blue emission (λem 456 nm) of RY3 is from
the electron transition S2 f S0 which is only slightly influenced
by CHO rotation as the energy gap (S2ꢀS0) is relatively large. By
contrast, the red emission (λem 650 nm) of RY3 is from the
electron transition S1 f S0 which is strongly influenced by CHO
rotation as the energy gap (S1ꢀS0) is rather small and sensitive to
the vibrational change. The fast internal conversion of S2 f S1
provides the major transient species of S1 and increases the
probability of red emission. These processes are illustrated in
Figure 8.
The effect of CHO group also can be seen in the case of RY5.
The fluorescence of RY5 shows a 10-fold enhancement from
ethanol to 99% glycerol (SI Figure S4), whereas RY1 has no such
response under the same conditions.
Fluorescence Lifetime. As reported by Suhling et al.,10ꢀ13
FLIM is another good way to investigate the intracellular
viscosity and is also independent of the concentration of the
dye. The quantitative relationship between the fluorescence
lifetime (τf) of molecular rotor and viscosity (η) of the solvent
is well depicted by F€orsterꢀHoffmann eq 2:26
With similar calculation, by DFT/TDDFT in B3LYP/6-31G**
level of Gaussian 09, compared with RY3, RY4 shows different
results. The oscillator strength for the electron transition from
HOMO (S0) to LUMOþ1 (S2) is only 0.0075, compared to the
normal electron jump from HOMO (S0) to LUMO (S1) with an
oscillator strength of 1.90. So, short wavelengths (absorption and
log τf ¼ C þ xlog η
ð2Þ
where C is a concentration- and temperature-dependent con-
stant and x is a dye- and temperature-dependent constant.
As shown in Figure 9A, when the proportion of glycerol in the
solvent mixture increases, and especially when the glycerol
6632
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635

Journal of the American Chemical Society
ARTICLE
Figure 9. The fluorescence lifetime of RY3 (1.0 μM) changes with viscosity of solvents at 650 nm. (A) Fluorescence decays as a function of viscosity by
increasing the proportions of glycerol in ethanol-glycerol solutions (with a Horiba Jobin Yvon Fluoromax-4p system excited at 376 nm); and (B) the
linearity between log τf and log η.
proportion is over 70%, the lifetime (τf) gradually becomes longer.
From pure ethanol to 99% glycerol, the total fluorescence
lifetimes of RY3 increased from 200 to 1450 ps, with a linear
relationship between log τf and log η (Figure 9B) (R2 =
0.996) and large slope (x = 0.49), giving a good fit with the
F€orsterꢀHoffmann eq 2.
Dual Mode Fluorescence Imaging. Cyanine dyes have been
widely used as fluorescence sensors34ꢀ36 and some have been
used for imaging in vivo.37ꢀ39 However, the ratiometric system is
rather rare.40,41 RY3 can penetrate live cell membranes, and
image viscosity in cytoplasmic structures of living cells; such as
PC12 cells shown in SI Figure S10, breast cancer cells MCF-7 in
SI Figures S11, S12 and NIH-3T3 cells in SI Figure S13. After
MCF-7 cells were incubated with RY3 (5.0 μM), clear fluores-
cence images at 650 ( 20 nm map the cellular viscosity
(SI Figure S12B). In the blue channel (λem 456 ( 20), however,
the fluorescence image is very weak (SI Figure S12A). The ratio
image (I650(20/I456(20) displays the different viscosity areas in
MCF-7 cells and is independent of the concentration or dis-
tribution of the dyes (Figure 11). In most imaging areas of MCF-
7 cells are blue, with small ratio values, and low viscosity (less
than 100 cP); some regions are green, indicating an intermediate
ratio value of 100ꢀ400 cP of viscosity; and the small intracellular
red zone indicates larger ratio values and higher viscosity, of more
than 900 cP, which can be estimated by the calibration linearity
between log(I650(20/I456(20) and logη.
Using a time-correlated single photon counting (TCSPC)
system,42,43 and RY3, a fluorescence lifetime image (Figure 12)
for cellular viscosity was obtained. It can be seen that some of
cellular regions are colored cyan, indicating that the fluorescence
lifetime of RY3 in those areas is less than 500 ps (or 100 cP). The
main part (green region) represents the fluorescence lifetime of
RY3 inside cells around 500ꢀ800 ps, corresponding to the
viscosity value of 100ꢀ400 cP. In some small orange areas of
the cells, the lifetime is as long as ca. 1800 ps (ca. 900 cP), close
to 99% glycerol solution (1450 ps, or 950 cP).
In a 70/30 glycerolꢀethanol solution, as the temperature is
increased from 5 to 40 °C, the fluorescence decay becomes faster
(Figure 10). The fluorescence lifetime of RY3 was shortened
from 711.2 to 183.6 ps in a clearly linear inverse relationship
between log τf and temperature with R2 0.996. This is expected
because the viscosity of a solution is inversely proportional to
temperature.
Possible Influences from Proteins and Photodecomposi-
tion. Proteins and mercapto-containing amino acids such as
cysteine sometimes react with aldehydes.27ꢀ29 Regarding the
importance of CHO in RY3, it is important to establish whether
RY3 reacts with such biochemicals intracellularly. When RY3
(2.0 μM) in PBS solution (pH 7.4) is titrated with 0ꢀ2.0 mg/mL
BSA (SI Figure S5) or cysteine (0ꢀ200 μM, SI Figure S6), the
changes in absorbance (except for a new absorption peak of BSA
at 280 nm), fluorescence intensity, and fluorescence ratio are all
negligible, even after the mixture is stirred at 37 °C for more than
120 min. Comparing the results with and without BSA, under
similar reaction conditions, the ratio values (I650/I456) (SI Figure S7)
and the fluorescence lifetimes (SI Figure S8) of RY3 have similar
linear relationships.
It is not surprising, as most of the aldehydes reactive with
proteins or mercapto-containing acids are aliphatic aldehydes.
However, RY3 is an aromatic aldehyde, and such species are
rather stable in neutral aqueous media.
To check the reliability, we changed the testing temperature from
4 to 37 °C. RY3 shows similar changes in ratio values and
fluorescence lifetimes. As shown in Figures 6 and 10, when the
Additionally, poor photostability is a general drawback of
polymethine dyes. It has been suggested that singlet oxygen
(1O2) and super oxide anion (O2ꢀ) oxidize excited dye mol-
ecules under light irradiation.30ꢀ33 As SI Figure S9 shows,
however, RY3 displayed excellent photostability. After irradia-
tion by a 500 W iodineꢀtungsten lamp for 7 h, the absorption of
RY3 remained as high as 95.9%, while other dyes were less stable:
RY1 (50.4%), RY2 (85.1%), and RY4 (4.1%), respectively. This
might result from the electron-withdrawing effect of the CHO
group which decreases the electron density in the pentamethine
chain of RY3. The order of first oxidization potential values
(SI Table S1) of these dyes—0.71 V (RY3) > 0.55 V (RY2) >
0.41 V (RY1) > 0.20 V (RY4)—supports the above viewpoint.
So RY3, with excellent photostability, is an ideal candidate for
laser confocal imaging.
temperatures of solutions are raised from 0 to 40 °C, the log(I650
/
I456) value was reduced from 1.15 to ca. 0.65, and the fluorescence
lifetime of RY3 at 650 nm was shortened from 711 to 183 ps.
Essentially the same results were obtained in living NIH-3T3 cells, as
demonstrated in SI Figures S13 and S14. The ratio values (SI Figure
S13B) are clearly lower and fluorescence lifetimes of RY3 (SI Figure
S14B) are similarly shortened. Therefore, the two modes of imaging
give similar viscosity results in living cells.
Although the live cells are very complicated, with the spatial
resolution of a confocal microscope via a simple calibration, we
can correlate the two-mode results and let them work together to
give reliable and accurate intracellular viscosity images previously
unavailable.
6633
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635

Journal of the American Chemical Society
ARTICLE
Figure 10. Fluorescence decays of RY3 (1.0 μM) in 70/30 glycerolꢀethanol solution with the increase of temperature at 650 nm (with a Horiba Jobin
Yvon Fluoromax-4p system excited at 376 nm) (A); and Inverse linearity between log τf and temperature (B).
in NIH-3T3 cells. To our best knowledge, RY3 is the first
molecule providing the dual mode fluorescence responses, which
will arouse more attention to discover more sensors to investi-
gate cellular viscosity related to diseases and pathology.
’ ASSOCIATED CONTENT
S
Supporting Information. Synthesis, experimental de-
b
tails, and additional spectroscopic data and images. This material
is available free of charge via the Internet at http://pubs.acs.org.
Figure 11. The fluorescence ratio image of MCF-7 cells stained by RY3
(5.0 μM) observed under a Leica TCS-SP2 confocal fluorescence
’ AUTHOR INFORMATION
microscope (100ꢁ objective lens) at 18 °C and excited at 800 nm;
the ratio image was obtained by image analysis software of Image
Corresponding Author
pengxj@dlut.edu.cn
Pro-plus 6.0.
’ ACKNOWLEDGMENT
This work was supported by NSF of China (20725621,
20876024, 21076032, 21072024, and 21006009), National Basic
Research Program of China (2009CB724706), Ministry of
Education of China, Cultivation Fund of the Key Scientific and
Technical Innovation Project (707016).
’ REFERENCES
(1) Stutts, M. J.; Canessa, C. M.; Olsen, J. C.; Hamrick, M.; Cohn,
J. A.; Rossier, B. C.; Boucher, R. C. Science 1995, 269, 847–850.
(2) Luby-Phelps, K. Int. Rev. Cytol. 2000, 192, 189–221.
(3) Kaliviotis, E.; Yianneskis, M. Proc. Inst. Mech. Eng., Part H 2007,
221, 887.
Figure 12. Fluorescence lifetime image of RY3 (5.0 μM) in MCF-7
cells under TCSPC FLIM equipment: SPC150, excited at 800 nm,
at 18 °C.
’ CONCLUSIONS
(4) Phelps, K. L.; Mujumdar, S.; Mujumdar, R. B.; Ernst, L. A.;
Galbraith, W.; Waggoner, A. S. Biophys. J. 1993, 65, 236–242.
(5) Haidekker, M. A.; Theodorakis, E. A. Org. Biomol. Chem. 2007,
5, 1669–1678.
(6) Haidekker, M. A.; Brady, T. P.; Lichlyter, D.; Theodorakis, E. A.
J. Am. Chem. Soc. 2006, 128, 398–399.
(7) Fischer, D. E.; Theodorakis, A.; Haidekker, M. A. Nature
Protocols 2007, 227–236.
(8) Kuimova, M. K.; Yahioglu, G.; Levitt, J. A.; Suhling, K. J. Am.
Chem. Soc. 2008, 130, 6672–6673.
(9) Kuimova, M. K.; Botchway, S. W.; Parker, A. W.; Balaz, M.;
Collins, H. A.; Anderson, H. L.; Suhling, K.; Ogilby, P. R. Nature Chem.
2009, 1, 69–73.
(10) Suhling, K.; French, P. M. W.; Phillips, D. Photochem. Photobiol
Sci. 2005, 4, 13–22.
(11) Zhang, J. F.; Zhou, Y.; Yoon, J.; Kim, Y.; Kim, S. J.; Kim, J. S.
Org. Lett. 2010, 12, 3852–3855.
(12) Zhou, Y.; Wang, F.; Kim, Y.; Kim, S.; Yoon, J. Org. Lett. 2009,
11, 4442–4445.
In conclusion, RY3, with a CHO group at the meso-position of
the pentamethine chain, has low fluorescence quantum yields in
nonviscous solutions. It is argued that the reason for this is that
rotation of the CHO group releases the energy of the excited dye
molecule nonradiatively via internal conversion including an ICT
process. In viscous media, however, the rotation is restrained and
large fluorescence enhancement was observed. A similar effect
was obtained by lowering the temperature. The mechanism has
been interpreted by 1H NMR at low temperature and Theoretical
calculation (Gaussian 09 software). Both the increase of viscosity
and fall of temperature determine the rotation rates, resulting in
ratio performance and lifetime increase. RY3 can enter into living
cells and image intracellular viscosity by two modes of fluores-
cence ratiometry imaging and fluorescence lifetime imaging
(FLIM). The accuracy and reliability of this application have
been illustrated by comparison and correlation of the two images
6634
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635

Journal of the American Chemical Society
ARTICLE
(13) Levitt, J. A.; Kuimova, M. K.; Yahioglu, G.; Chung, P. H.;
Suhling, K.; Phillip, D. J. Phys. Chem. C 2009, 113, 11634–11642.
(14) Ock, K.; Jang, G.; Roh, Y.; Kim, S.; Kim, J.; Koh., K. Microchem.
J. 2001, 70, 301–305.
(15) Zoppellaro, G.; Geies, A.; Enkelmann, V.; Baumgarten, M. Eur.
J. Org. Chem. 2004, 2367–2374.
(16) Shao, F. W.; Weissleder, R.; Hilderbrand, S. A. Bioconj. Chem.
2008, 19, 2487–2491.
(17) Frisch, M. J. et al. Gaussian 09, Revision A.02, Gaussian, Inc.,
Wallingford CT, 2009.
(18) Zhou, L. C.; Zhao, G. J.; Liu, J. F.; Han, K. L.; Wu, Y. K.; Peng,
X. J.; Sun, M. T. J. Photochem. Photobiol. A: Chemistry 2007,
187, 305–310.
(19) Dreizler, M. R.; Gross, E. K. U. Density Functional Theory;
Springer Verlag: Heidelberg, 1990.
(20) Gross, E. K. U.; Kohn, W. Phys. Rev. Lett. 1985, 55, 2850–2852.
(21) Stratmann, R. E.; Scuseria, G. E.; Frisch, M. J. J. Chem. Phys.
1998, 109, 8218–8224.
(22) Velapoldi, R. A.; T€onnesen, H. H. J. Fluoresc. 2004,
14, 465–472.
(23) Reichardt, C.; Grahn., W. Tetrahedron 1976, 32, 125–134.
(24) Reichardt, C.; Mormann, W. Chem. Ber. 1972, 105, 1815–1839.
(25) Haidekker, M. A.; Brady, T. P.; Lichlyter, D.; Theodorakis, E. A.
Bioorg. Chem. 2005, 33, 415–425.
(26) F€orster, Th.; Hoffmann, G. Z. Phys. Chem. 1971, 75, 63–76.
(27) Li, H. L.; Fan, J. L.; Wang, J. Y.; Tian, M. Z.; Du, J. J.; Sun, S. G.;
Sun, P. P.; Peng, X. J. Chem. Commun. 2009, 5904–5906.
(28) Lee, K. S.; Kim, T. K.; Lee, J. H.; Kim, H. J.; Hong, J. I. Chem.
Commun. 2008, 6173.
(29) Chen, X. Q.; Zhou, Y.; Peng, X. J.; Yoon, J. Chem. Soc. Rev. 2010,
39 (6), 1861–2336.
(30) Toutchkine, A.; Kraynov, V.; Hahn, K. J. Am. Chem. Soc. 2003,
125, 4132–4145.
(31) Renikuntla, B. R.; Rose, H. C.; Eldo, J.; Waggoner, A. S.;
Armitage, B. A. Org. Lett. 2004, 6, 909–912.
(32) Song, B.; Zhang, Q.; Ma, W. H.; Peng, X. J.; Fu, X. M.; Wang,
B. S. Dyes Pigm. 2009, 82, 396–400.
(33) Song, F. L.; Peng, X. J.; Lu, E. H; Zhang, R.; Chen, X. Y.; Song,
B. J. Photochem. Photobiol., A 2004, 168, 53–75.
(34) Fabian, J. Chem. Rev. 1992, 92, 1197–1226.
(35) Leevy, W. M.; Gammon, S. T.; Smith, B. D. J. Am. Chem. Soc.
2006, 128, 16476.
(36) Zhu, M.; Yuan, M.; Liu, X.; Xu, J.; Lv, J.; Huang, C.; Liu, H.; Li,
Y.; Wang, S.; Zhu, D. Org. Lett. 2008, 10, 1481–1484.
(37) Tang, B.; Yu, F. B.; Li, P.; Tong, L. L.; Duan, X.; Xie, T.; Wang,
X. J. Am. Chem. Soc. 2009, 131, 3016–3023.
(38) Gruber, H. J.; Hahn, C. D.; Kada, G.; Riener, C. K.; Harms,
G. S.; Ahrer, W.; Dax, T. G.; Knaus, H. G. Bioconj. Chem. 2000,
11, 696–704.
(39) Weissleder, R.; Ntziachristos, V. Nat. Med. 2003, 9, 123–131.
(40) Peng, X. J.; Song, F. L.; Lu, E. H.; Wang, Y. N.; Zhou, W.; Fan,
J. L.; Gao, Y. L. J. Am. Chem. Soc. 2005, 127, 4170–4171.
(41) Kiyose, K.; Kojima, H.; Nagano, T. J. Am. Chem. Soc. 2006,
128, 6548–6549.
(42) O’Connor, D. V.; Phillips, D. Time-Correlated Single-Photon
Counting; Academic Press: London, 1984.
(43) Becker, W. Advanced Time-Correlated Single Photon Counting
Techniques; Springer: Berlin, 2005.
6635
dx.doi.org/10.1021/ja1104014 |J. Am. Chem. Soc. 2011, 133, 6626–6635