The contribution of reactive oxygen species to the photobleaching of organic fluorophores

Article abstract of DOI:10.1111/php.12204

Photoexcitation of fluorophores commonly used for biological imaging applications generates reactive oxygen species (ROS) which can cause bleaching of the fluorophore and damage to the biological system under investigation. In this study, we show that sin

Full text of DOI:10.1111/php.12204

Photochemistry and Photobiology, 2014, 90: 448–454
The Contribution of Reactive Oxygen Species to the Photobleaching of
†
Organic Fluorophores
‡
1,2
‡3
1
1,2
Qinsi Zheng , Steffen Jockusch , Zhou Zhou and Scott C. Blanchard*
1
Department of Physiology and Biophysics, Weill Medical College of Cornell University, New York, NY
Tri-Institutional Training Program in Chemical Biology, New York, NY
Department of Chemistry, Columbia University, New York, NY
2
3
Received 2 September 2013, accepted 31 October 2013, DOI: 10.1111/php.12204
ABSTRACT
to molecular oxygen leading to the formation of energetically
excited singlet oxygen. Second, electron transfer can occur from
the triplet fluorophore to molecular oxygen leading to the forma-
tion of a superoxide radical (Scheme 1) (9,11). Singlet oxygen
and superoxide radical, along with other oxidizing species
formed subsequently (Scheme 1), are collectively termed reactive
oxygen species (ROS) which can degrade fluorophores (10–14)
and damage biomolecules in various contexts (6,7,15–17). How-
ever, the role of specific ROS in fluorophore photobleaching is
not well understood. In particular, while it is widely held that
singlet oxygen is a direct participant in the photobleaching path-
way (9–11,13,18,19), at least one early study provided evidence
suggesting that it does not directly participate in the photobleach-
ing of fluorescein isothiocyanate (FITC) (20). Here, we aim to
elucidate the contribution of singlet oxygen to the photobleaching
of Cy5 and ATTO 647N, two organic fluorophores commonly
used in single-molecule and super-resolution measurements.
To mitigate fluorophore photobleaching and phototoxicity,
molecular oxygen, normally at 0.3 mM in aqueous solution (21),
can be depleted by enzymatic oxygen scavenging system (18,22).
Moreover, small-molecule protective agents (PA), such as COT, ni-
trobenzyl alcohol (NBA) and Trolox (23,24), can be added to the
imaging solution to mitigate the transient dark state caused by the
triplet and radical state species (5,10). However, the use of oxygen
scavenging systems and PA in solution can be incompatible with, or
perturbing to, biological systems, particularly for measurements
involving lipid membranes (25) or live cells.
Photoexcitation of fluorophores commonly used for biological
imaging applications generates reactive oxygen species (ROS)
which can cause bleaching of the fluorophore and damage to
the biological system under investigation. In this study, we
show that singlet oxygen contributes relatively little to Cy5 and
ATTO 647N photobleaching at low concentrations in aqueous
solution. We also show that Cy5 generates significantly less
ROS when covalently linked to the protective agents, cyclooc-
tatetraene (COT), nitrobenzyl alcohol (NBA) or Trolox. Such
fluorophores exhibit enhanced photostability both in bulk solu-
tions and in single-molecule fluorescence measurements. While
the fluorophores ATTO 647N and ATTO 655 showed greater
photostability than Cy5 and the protective–agent-linked Cy5
derivatives investigated here, both of ATTO 647N and ATTO
6
55 generated singlet oxygen and hydroxyl radicals at rela-
tively rapid rates, suggesting that they may be substantially
more phototoxic than Cy5 and its derivatives.
INTRODUCTION
Fluorescence microscopy has seen revolutionary advancements in
both sensitivity and resolution over the past decade (1). Progress
in this field is highlighted by recent developments in single-mol-
ecule and super-resolution fluorescence imaging techniques that
have yielded new insights into a variety of biological systems in
unprecedented detail (2–4). However, photoinduced degradation
Recently, we demonstrated that the covalent linkage of a sin-
gle COT, NBA and Trolox to cyanine fluorophores substantially
reduces the frequency of transient dark state and the rate of
photobleaching, thereby circumventing the requirements of PA
and, in some cases, of oxygen scavenger system in solution
(
photobleaching) of the fluorophores used remains a key obstacle
that limits the temporal and spatial resolutions of imaging (5).
Photoinduced fluorophore toxicity (phototoxicity) may also lead
to unwanted perturbations to the biological system (6–8) that can
obscure the signal of interest. Further advancements in fluores-
cence imaging would therefore be greatly aided by a deeper
understanding of the mechanisms of fluorophore photobleaching
and the development of fluorophores exhibiting slow photoble-
aching and low-toxicity.
Molecular oxygen plays a critical role in fluorophore photo-
bleaching and phototoxicity (5,7,9,10) as it can participate in two
probable reactions with an excited fluorophore. First, energy
transfer can occur from a fluorophore in the triplet excited state
(26,27). The protective effect for Cy5-COT results from its
capacity to reduce the lifetime of the triplet state by up to 50-
fold through energy-transfer quenching, whereas the distinct
mechanism for Cy5-NBA and Cy5-Trolox was unknown (28).
Here, we show that COT-, NBA- and Trolox-linked Cy5 deriva-
tives (Cy5-PA) (Scheme 2) exhibit reduced rates of ROS genera-
tion compared with the parent Cy5 molecule. These findings
imply that PA-conjugated fluorophores may exhibit reduced
phototoxicity compared with the fluorophores commonly used
for fluorescence imaging, and suggest that the Cy5-NBA and
Cy5-Trolox in aqueous solution are also protected by triplet state
quenching.
†
This paper is part of the Special Issue honoring the memory of Nicholas J. Turro.
‡
These authors contributed equally.
Corresponding author email: scb2005@med.cornell.edu
2013 The American Society of Photobiology
*
©
4
48

Photochemistry and Photobiology, 2014, 90 449
MATERIALS AND METHODS
Materials. N-hydroxysuccinimide (NHS) ester activated fluorophores Cy5
(
(
GE Healthcare), ATTO 647N (ATTO-TEC GmbH) and ATTO 655
ATTO-TEC-GmbH) were used as received. Cy5-COT, Cy5-NBA and
Cy5-Trolox were synthesized as previously described (28).
Fluorophore bleaching experiments. Fluorophores (5 lM or 26 lM for
Cy5; 5 lM for Cy5-COT, Cy5-NBA and Cy5-Trolox; 10 lM for ATTO
6
55 and ATTO 647N) were dissolved in solvent (H
CDCl , acetontrile or 100 mM Tris-acetate aqueous solution at pH 7.5).
Samples (2 mL) were examined in a 1 cm cuvette, illuminated with a
00 W Tungsten halogen lamp, using an RG570 longpass filter to block
2 2 3
O, D O, CHCl ,
3
3
short-wavelength light. Absorbance spectra of each fluorophore were
measured using a UV–vis spectrometer (Agilent 8453). The decrease in
the absorption was used to estimate the relative rate of fluorophore photo-
bleaching over the first 2 min using a linear fit.
Singlet oxygen measurements. Singlet Oxygen Sensor Green (SOSG;
Invitrogen) experiments were performed in aqueous buffer as previously
described (29) in a 1 cm cuvette containing 2 mL of 100 mM Tris-acetate
(
pH 7.5), SOSG (2 lM) fluorophore (5 or 10 lM). Samples were illumi-
nated for fixed periods of time with a 300 W Tungsten halogen lamp in
conjugation with a RG570 longpass filter. After each photolysis period,
the fluorescence of SOSG was recorded over a range of 510–630 nm
using 504 nm for excitation.
Due to the incompatibility of SOSG with organic solvent, 9,10-diphe-
nylanthracene (DPA) was used to detect singlet oxygen in acetonitrile.
DPA exhibits strong absorbance at 373 and 393 nm, and forms 9,10-
endoperoxide, a compound that has no absorbance above 350 nm (Figure
S4a). The absorption of DPA was measured with a UV–vis spectrometer
(Agilent 8453).
Single-molecule fluorescence measurements. Single-molecule fluores-
cence measurements were performed using a wide-field, prism-based total
internal reflection fluorescence microscope as previously described
Scheme 1. Some probable steps to reactive oxygen species via the reac-
3
tions between molecular oxygen ( O
2 0
) and the fluorophore. S : fluoro-
phore in ground state; S : fluorophore in the singlet excited state; T :
1
1
+
fluorophore in the triplet state; R : fluorophore in the radical cationic
state.
(
26,28).
Fluorescence quantum yield measurements. Fluorescence spectra of
solutions containing 0.7 lM Cy5, Cy5-COT, Cy5-NBA or Cy5-Trolox
were recorded with a spectrofluorometer (Fluorolog 3; HORIBA Jobin
Yvon), using an excitation wavelength of either 649 (aqueous Tris buffer,
pH 7.5) or 654 nm (acetonitrile). The fluorescence quantum yield was
calculated using Cy5 in aqueous solution (/ = 0.20 (30)) as a standard.
Fluorescence lifetime measurements. Fluorescence lifetimes were mea-
sured by time correlated single photon counting (TC-SPC) with an
OB920 spectrometer (Edinburgh Analytical Instruments) in conjunction
with a pulsed diode laser emitting at 659 nm. Fluorescence decay traces
were monitored at 670 nm.
RESULTS AND DISCUSSIONS
The organic fluorophores Cy5, ATTO 647N and ATTO 655 used
for the present investigation represent the three main categories
of fluorophore structures (cyanines for Cy5, carbopyronines for
ATTO 647N, and oxazines for ATTO 655) that are widely used
for single-molecule and super-resolution fluorescence imaging
(31–33). Removal of molecular oxygen from the imaging solu-
tion substantially reduces the photobleaching rate of each fluoro-
phore (20,31), implying that photobleaching involves molecular
oxygen. Here, we show that the photobleaching rates of these
dyes follow the order: Cy5 > ATTO 647N > ATTO 655
(Fig. 1), inversely correlating with the oxidation potential of the
fluorophores (EOX,Cy5 = 0.97 V, EOX,ATTO647N = 1.11 V, EOX,
ATTO655 = 1.31 V (32,34)), where an increased oxidation poten-
tial implies a decreased propensity to be oxidized. Given the dis-
tinct structures and properties of these fluorophores (Scheme 2),
this inverse correlation strongly supports that photobleaching is
dominated by photoinduced oxidation reactions in the presence
of oxygen.
To test the involvement of singlet oxygen in fluorophore
photobleaching, we examined whether fluorophore photobleach-
ing exhibits a solvent isotope effect (9). Given that the lifetime
Scheme 2. Structures of the compounds used in this study.

4
50 Qinsi Zheng et al.
led to its generation, a significant solvent isotope was observed
for Cy5 at 5 lM in chloroform (CHCl ) and deuterated chloro-
form (CDCl ) (Figure S1), where singlet oxygen has much
longer lifetime (~230 ls in CHCl and 7 ms in CDCl (35)) than
3
3
3
3
it has in water. The significant solvent isotope effects observed
in water with high Cy5 concentration, and in chloroform, demon-
strate that singlet oxygen can contribute to fluorophore photoble-
aching under certain conditions. Consistent with previous results
(20), the absence of any significant of solvent isotope effect at
low fluorophore concentration in aqueous solution, suggests that
singlet oxygen contributes negligibly to fluorophore photobleach-
ing under the conditions of most biological studies at the single-
molecule scale.
To examine whether the PA-conjugated fluorophores produce
less singlet oxygen, we attempted to spectroscopically measure
singlet oxygen via its phosphorescence emission band at
Figure 1. Photobleaching of Cy5 (5 lM), ATTO 647N (10 lM) and
ATTO 655 (10 lM) in air-saturated H
2
O (red) or D
2
O (blue), where k
B
is
the photobleaching rate.
1
270 nm (36) upon fluorophore illumination. However, under
the conditions of the experiment (5 lM fluorophore at room tem-
perature), the yield of singlet oxygen was too low to be reliably
detected using this approach (data not shown). We therefore used
SOSG as an indirect probe of singlet oxygen generation. SOSG
itself is weakly fluorescent due to intramolecular electron-transfer
quenching and becomes highly fluorescent following cycloaddi-
tion of singlet oxygen (Fig. 3a) (29).
Initially, singlet oxygen and photobleaching resulting Cy5
illumination (>570 nm) were tracked by monitoring fluorophore
absorption and SOSG emission spectra as a function of illumina-
tion time (647 and 528 nm, respectively) (Fig. 3b,d). As
expected, the observed increase in SOSG fluorescence emission
paralleled the decrease in Cy5 absorption. The increase in SOSG
fluorescence was confirmed to arise from its specific reaction
with singlet oxygen by performing the same measurement in
90% deuterated water, where the lifetime of singlet oxygen is
significantly extended (see discussion above). In contrast to the
negligible solvent isotope effect observed for photobleaching of
Cy5 at low concentrations (Fig. 1), the rate of SOSG fluores-
cence increase was enhanced by approximately eight-fold in
of singlet oxygen is extended by more than 20-fold in D
68 ls vs. 3.1 ls in H
principally mediated by singlet oxygen, a substantial enhance-
2
O
(
2
0) (35), if fluorophore photobleaching is
ment of the rate of photobleaching should be observed in D O
2
(
9). However, as observed previously for the organic fluorophore
fluorescein isothiocyanate (FITC) (20), in an initial test the rates
of Cy5 (5 lM) and ATTO 647N (10 lM) photobleaching were
only slightly higher in D O than in H O (Fig. 1). In contrast, a
2 2
significant solvent isotope effect was observed in aqueous condi-
tions when the Cy5 concentration was increased five-fold to
2
6 lM (Fig. 2), consistent with the involvement of singlet oxy-
gen in the photobleaching reaction. Experiments at high concen-
tration were not possible for ATTO 647N due to aggregation of
the fluorophore. The absence of a solvent isotope effect at low
concentration, and its presence at high Cy5 concentration, can be
explained by the fact that encounter rate between singlet oxygen
and fluorophore increases with increasing fluorophore concentra-
tion. At low fluorophore concentrations, singlet oxygen decays
much faster than it encounters and reacts with fluorophores.
Under such conditions singlet oxygen contributes negligibly to
Cy5 photobleaching compared with other photobleaching reac-
tions (Scheme 1). At high fluorophore concentration, singlet oxy-
gen can react with neighboring fluorophores to contribute to
photobleaching to a greater extent. Consistent with the model
that singlet oxygen decays too rapidly in aqueous solutions to
contribute meaningfully to the bleaching of the fluorophore that
solution in 90% D O (Figure S2).
2
In parallel, the rates of singlet oxygen generation were also
measured for Cy5-COT, Cy5-NBA, Cy5-Trolox, ATTO 655 and
ATTO 647N. As observed for the experiment of Cy5, SOSG flu-
orescence increased upon illumination of each fluorophore
(Fig. 3c,e). All three PA-conjugated fluorophores exhibited lower
rates of singlet oxygen generation, which paralleled their
increased photostability compared with Cy5. Specifically, Cy5-
Trolox, Cy5-COT and Cy5-NBA exhibited 2.5-fold, 5-fold and
2
0-fold reductions in the rate of singlet oxygen generation,
respectively (Fig. 3c; Table 1). In contrast, ATTO 655 exhibited
rates of singlet oxygen generation that were similar to those
observed for Cy5, whereas ATTO 647N generated singlet oxy-
gen at a rate that was approximately five-fold greater than Cy5
(Fig. 3c). These data imply that Cy5-PA fluorophores may be
significantly less phototoxic than Cy5, ATTO 647N and ATTO
6
55 in biological contexts.
To further explore the generation of ROS, we examined the
rates of hydroxyl radical (HOÁ) formation using the probe
Aminophenyl fluorescein (APF) (37). Upon reaction with
hydroxyl radicals, APF converts to fluorescein, a highly fluo-
rescent species (37) (Figure S3a). Aqueous solutions containing
APF and individual fluorophores showed marked increases in
fluorescein fluorescence (kmax = 518 nm) as a function of illu-
Figure 2. Photobleaching of Cy5 at 26 lM in air-saturated H
or D O (open), where k is the photobleaching rate.
2
O (solid)
2
B

Photochemistry and Photobiology, 2014, 90 451
1
Figure 3. (a) The mechanism of singlet oxygen ( O
2
) induced fluorescence of Singlet Oxygen Sensor Green (SOSG). (b) The emission spectra of SOSG
1
upon illumination of Cy5. (c) O
2
generation, reported by the increase in SOSG fluorescence, for distinct fluorophores at different illumination time. The
relative rates of O generation compared with the Cy5 fluorophore are shown in parentheses. (d) The absorption spectrum of Cy5 observed at different
2
1
illumination time. (e) The decrease in Cy5, Cy5-COT, Cy5-NBA, Cy5-Trolox, ATTO 647N and ATTO 655 absorption at 647 nm as a function of illu-
mination time. Experiments were performed in 100 mM Tris-acetate-buffered aqueous solution, pH 7.5.
mination time (Figure S3b). Consistent with their reduced rates
of singlet oxygen generation, the PA-conjugated fluorophores
To directly examine the performance of fluorophores in sin-
gle-molecule fluorescence measurement in ambient oxygen con-
ditions, each fluorophore was conjugated to a double-stranded
DNA oligonucleotide and imaged under a total internal reflec-
tion fluorescence microscope (26). Visual inspection of fluores-
cence traces from individual molecules revealed that most of
ATTO 647N and ATTO 655 molecules showed substantial tran-
(
Cy5-COT, Cy5-NBA and Cy5-Trolox) exhibited significantly
reduced rates of hydroxyl radical generation (Figure S3c).
These data provide supporting evidence that the Cy5-PA flu-
orophores may be less phototoxic than the Cy5, ATTO 647N
and ATTO 655.

4
52 Qinsi Zheng et al.
Table 1. A summary of the data. 100 mM Tris-acetate buffer (pH = 7.5) was used in all experiments in aqueous buffer. All of the rates were normalized
to the corresponding numbers for Cy5.
Solvent
Cy5
Cy5-COT
Cy5-NBA
Cy5-Trolox
Aqueous
Relative rate of single-molecule photobleaching
1 Æ 0.2
1 Æ 0.2
0.29 Æ 0.02
0.45 Æ 0.03
0.18 Æ 0.01
0.43 Æ 0.01
0.95 Æ 0.02
0.20
0.50 Æ 0.01
0.57 Æ 0.06
1.18 Æ 0.05
0.22
0.27 Æ 0.02
0.24 Æ 0.04
0.05 Æ 0.01
0.235 Æ 0.004
0.59 Æ 0.02
0.07
1.0 Æ 0.1
0.92 Æ 0.06
0.93 Æ 0.04
0.18
0.42 Æ 0.04
0.78 Æ 0.06
0.41 Æ 0.01
0.367 Æ 0.009
0.76 Æ 0.03
0.11
0.9 Æ 0.1
0.86 Æ 0.06
1.02 Æ 0.03
0.19
Relative rate of ensemble photobleaching
1
Relative rate of O
2
generation
1 Æ 0.05
1 Æ 0.03
0.92 Æ 0.02
0.20 (30)
1 Æ 0.08
1 Æ 0.08
1.11 Æ 0.03
0.22
Relative rate of hydroxyl radical generation
Fluorescence lifetime (ns)
Fluorescence quantum yield
Acetonitrile
Relative rate of ensemble photobleaching
1
Relative rate of O
2
generation
Fluorescence lifetime (ns)
Fluorescence quantum yield
Triplet state lifetime (ls)
63 Æ 3
1.1 Æ 0.1
62 Æ 3
60 Æ 4
Figure 4. Representative single-molecule fluorescence traces for Cy5, Cy5-COT, Cy5-NBA, Cy5-Trolox, ATTO 647N and ATTO 655 covalently linked
to DNA oligonucleotides and imaged in air-saturated aqueous solution using a total internal reflection microscope under continuous laser excitation
(
641 nm). l is the average number of detected photons before fluorophore photobleaching.
sient dark states and intensity fluctuations, whereas Cy5, Cy5-
COT, Cy5-NBA and Cy5-Trolox rarely blinked or fluctuated
conjugates investigated here were substantially dimmer than
Cy5 (~30% and 50%, respectively; Fig. 4), likely due to the
electron transfer between the PA and Cy5 singlet excited state.
The slow-photobleaching, nonblinking and nonfluctuating prop-
erties of the Cy5 derivatives investigated here render them suit-
able for single-molecule quantitative measurements, such as
fluorescence resonance energy transfer, anisotropy, and stoichi-
ometry. In contrast, the fluctuating emission behaviors of ATTO
647N and ATTO 655 make them more suitable for single-mole-
cule tracking.
(
(
Fig. 4). By tracking the fluorescence of many single molecules
>1000) over time (Fig. 4), the fluorescence intensity and the
photobleaching rates were quantified. To compare each fluoro-
phore’s photobleaching rate, l (the average numbers of photons
detected from individual fluorophore prior to photobleaching)
was calculated. As anticipated from prior investigations (26), the
Cy5-COT, Cy5-NBA and Cy5-Trolox fluorophores exhibited
significantly reduced rates of photobleaching (Table 1), emitting
up to four-fold more photons than Cy5 prior to photobleaching
In previous studies (28), we showed that Cy5-COT, but not
Cy5-NBA and Cy5-Trolox, exhibits shortened triplet state life-
time compared with Cy5 (Table 1). However, our previous mea-
surements were exclusively performed in acetonitrile. Here, we
(
Fig. 4). As previously reported (26), the COT-conjugated flu-
orophore displayed a roughly 30% increase in brightness relative
to Cy5 (Fig. 4). In contrast, both Cy5-NBA and Cy5-Trolox

Photochemistry and Photobiology, 2014, 90 453
have shown that each of the Cy5-PA fluorophores exhibit reduc-
tions in photobleaching rates and in ROS generation, implying
that in aqueous solution Cy5-COT, Cy5-NBA and Cy5-Trolox
possess substantially shorter triplet state lifetime than Cy5. These
data suggest that the effects of NBA and Trolox are strongly sol-
vent-dependent: in aqueous solution they effectively quench the
triplet state of Cy5, likely through an electron-transfer mecha-
nism (38,39), whereas in acetonitrile they do not quench the trip-
let state due to the solvent-dependent nature of the electron
transfer where radical states form (9). The rate of electron trans-
fer is known to be affected by solvent polarity (40). In contrast,
COT can quench the Cy5 triplet state in both solvents because
of the solvent-independent nature of triplet–triplet energy transfer
with APF. (b) The emission spectra of APF at different time
points of the photoillumination of Cy5. (c) The generation of
hydroxyl radical, reported by the increase in APF fluorescence,
upon illumination of Cy5, Cy5-COT, Cy5-NBA, Cy5-Trolox,
ATTO 655 and ATTO 647N. In the parenthesis are the relative
rates of hydroxyl radical generation.
1
Figure S4. (a) The reaction of O
2
with 9,10-diphenylanthra-
cene. (b) The spectrum of 9,10-diphenylanthracene absorption
during photobleaching of Cy5 (5 lM) in acetonitrile. (c) In the
parentheses are the relative rates of singlet oxygen generation
reported by the decrease in absorption at 393 nm.
(
9,28). This mechanism is supported by a further comparison of
fluorophore properties in aqueous solution and in acetonitrile: the
fluorescent lifetime, the fluorescent quantum yield, the photoble-
aching rate and the singlet oxygen generation rate for Cy5-NBA
and Cy5-Trolox were strongly solvent-dependent, whereas the
properties for Cy5 and Cy5-COT were only weakly solvent-
dependent (Table 1; Figure S4).
REFERENCES
1. Lakowicz, J. (2006) Principles of Fluorescence Spectroscopy.
Springer Science + Business Media LLC, New York, NY.
. Moerner, W. E. (2002) A Dozen Years of Single-Molecule Spectros-
2
copy in Physics, Chemistry, and Biophysics. J. Phys. Chem. B 106,
9
10–927.
3
. Joo, C., H. Balci, Y. Ishitsuka, C. Buranachai and T. Ha (2008)
Advances in Single-Molecule Fluorescence Methods for Molecular
Biology. Annu. Rev. Biochem. 77, 51–76.
CONCLUSION
4
. Huang, B., M. Bates and X. Zhuang (2009) Super-Resolution Fluo-
rescence Microscopy. Annu. Rev. Biochem. 78, 993–1016.
. Ha, T. and P. Tinnefeld (2012) Photophysics of Fluorescent Probes
for Single-Molecule Biophysics and Super-Resolution Imaging.
Annu. Rev. Phys. Chem. 63, 595–617.
The results presented suggest that the contribution of singlet oxy-
gen to fluorophore photobleaching is negligible at low fluoro-
phore concentrations (<5 lM) in aqueous buffers. Singlet oxygen
does, however, play a significant role under solvent where its
lifetime is significantly extended (e.g. chloroform) or at higher
fluorophore concentrations (>20 lM) where the collision fre-
quency with neighboring fluorophore molecules is greatly
increased. These observations explain discrepancies in the litera-
ture regarding the role of singlet oxygen in fluorophore photoble-
aching (13,20) and suggest that singlet oxygen is unlikely to
contribute to fluorophore photobleaching under most in vitro sin-
gle-molecule imaging conditions. Cy5 derivatives covalently
linked to a protective agent exhibit enhanced photostability as a
direct consequence of distinct mechanisms of triplet state
quenching that result in lower rates of singlet oxygen and hydro-
xyl radical generation. Correspondingly, these fluorophores may
also be less perturbing and exhibit reduced phototoxicity when
used for live cell imaging applications.
5
6
. Sheetz, M. P. and D. E. Koppel (1979) Membrane damage caused
by irradiation of fluorescent concanavalin A. Proc. Natl Acad. Sci.
7
6, 3314–3317.
7. Dixit, R. and R. Cyr (2003) Cell damage and reactive oxygen spe-
cies production induced by fluorescence microscopy: Effect on mito-
sis and guidelines for non-invasive fluorescence microscopy. Plant J.
3
6, 280–290.
8
. Dobrucki, J. W., D. Feret and A. Noatynska (2007) Scattering of
Exciting Light by Live Cells in Fluorescence Confocal Imaging:
Phototoxic Effects and Relevance for FRAP Studies. Biophys. J. 93,
1
778–1786.
9
. Turro, N. J., V. Ramamurthy and J. C. Scaiano (2010) Modern
Molecular Photochemistry of Organic Molecules. University Science
Books, Sausalito, CA.
10. Levitus, M. and S. Ranjit (2011) Cyanine dyes in biophysical
research: The photophysics of polymethine fluorescent dyes in bio-
molecular environments. Q. Rev. Biophys. 44, 123–151.
1
́
1. Bonnett, R. and G., Martõnez (2001) Photobleaching of sensitisers
used in photodynamic therapy. Tetrahedron 57, 9513–9547.
2. Chen, P., J. Li, Z. Qian, D. Zheng, T. Okasaki and M. Hayami
Acknowledgements—This work was supported by the National Institutes
of Health (GM098859-01A1 to SCB) and the National Science
Foundation (CHE 11-11392 to SJ). We are grateful to Roger B. Altman
and Daniel S. Terry for critical discussions and comments on the
manuscript.
1
(
1998) Study on the photooxidation of a near-infrared-absorbing ben-
zothiazolone cyanine dye. Dyes Pigm. 37, 213–222.
13. Byers, G. W., S. Gross and P. M. Henrichs (1976) Direct and Sensi-
tized Photooxidation of Cyanine Dyes. Photochem. Photobiol. 23,
3
7–43.
1
1
4. Kuramoto, N. and T. Kitao (1982) The contribution of singlet oxy-
gen to the photofading of triphenylmethane and related dyes. Dyes
Pigm. 3, 49–58.
5. Vigers, G. P., M. Coue and J. R. McIntosh (1988) Fluorescent
microtubules break up under illumination. The Journal of Cell Biol-
ogy 107, 1011–1024.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online
version of this article:
Figure S1. Photobleaching rates of Cy5 in air-saturated
CHCl and CDCl , where k is the photobleaching rate.
16. Davies, M. J. (2004) Reactive species formed on proteins exposed to
singlet oxygen. Photochem. Photobiol. Sci. 3, 17–25.
7. Sies, H. and C. F. M. Menck (1992) Singlet oxygen induced DNA
damage. Mutation Research/DNAging 275, 367–375.
8. Aitken, C. E., R. A. Marshall and J. D. Puglisi (2008) An oxygen
scavenging system for improvement of dye stability in single-mole-
cule fluorescence experiments. Biophys. J. 94, 1826–1835.
3
3
B
1
1
Figure S2. Relative rate of singlet oxygen generation, in H2O
solution and in 90% D2O solution, reported by the increase in
SOSG fluorescence integrated from 518 to 538 nm. Experiments
were performed with 0.7 lM Cy5 in 100 mM Tris-acetate-buf-
fered solution (pH 7.5).
1
9. Toutchkine, A., D.-V. Nguyen and K. M. Hahn (2007) Merocyanine
Dyes with Improved Photostability. Org. Lett. 9, 2775–2777.
Figure S3. (a) The mechanism of hydroxyl radical detection

4
54 Qinsi Zheng et al.
2
0. Johnson, G. D., R. S. Davidson, K. C. McNamee, G. Russell, D.
Goodwin and E. J. Holborow (1982) Fading of immunofluorescence
during microscopy: A study of the phenomenon and its remedy. J.
Immunol. Methods 55, 231–242.
31. Roy, R., S. Hohng and T. Ha (2008) A practical guide to single-mol-
ecule FRET. Nat. Methods 5, 507–516.
32. Vogelsang, J., T. Cordes, C. Forthmann, C. Steinhauer and P. Tinne-
feld (2009) Controlling the fluorescence of ordinary oxazine dyes for
single-molecule switching and superresolution microscopy. Proc Natl
Acad Sci 106, 8107–8112.
33. Dempsey, G. T., J. C. Vaughan, K. H. Chen, M. Bates and X. Zhu-
ang (2011) Evaluation of fluorophores for optimal performance in
localization-based super-resolution imaging. Nat. Methods 8,
1027–1036.
34. Vogelsang, J., R. Kasper, C. Steinhauer, B. Person, M. Heilemann,
M. Sauer and P. Tinnefeld (2008) A reducing and oxidizing system
minimizes photobleaching and blinking of fluorescent dyes. Angew.
Chem. Int. Ed. Engl. 47, 5465–5469.
35. Schweitzer, C. and R. Schmidt (2003) Physical Mechanisms of
Generation and Deactivation of Singlet Oxygen. Chem. Rev. 103,
1685–1758.
36. Khan, A. U. and M. Kasha (1979) Direct spectroscopic observation
of singlet oxygen emission at 1268 nm excited by sensitizing dyes
of biological interest in liquid solution. Proc. Natl Acad. Sci. 76,
6047–6049.
37. Setsukinai, K.-I., Y. Urano, K. Kakinuma, H. J. Majima and T. Nag-
ano (2003) Development of Novel Fluorescence Probes That Can
Reliably Detect Reactive Oxygen Species and Distinguish Specific
Species. J. Biol. Chem. 278, 3170–3175.
38. Tinnefeld, P. and T. Cordes (2012) ‘Self-healing’ dyes: Intra-
molecular stabilization of organic fluorophores. Nat Meth 9,
426–427.
21. Montalti, M., A. Credi, L. Prodi and M. T. Gandolfi (2006) Hand-
book of Photochemistry. CRC/Taylor & Francis, Boca Raton.
22. Benesch, R. E. and R. Benesch (1953) Enzymatic Removal of Oxygen
for Polarography and Related Methods. Science 118, 447–448.
23. Rasnik, I., S. A. McKinney and T. Ha (2006) Nonblinking and long-
lasting single-molecule fluorescence imaging. Nat Meth 3, 891–893.
24. Dave, R., D. S. Terry, J. B. Munro and S. C. Blanchard (2009) Miti-
gating Unwanted Photophysical Processes for Improved Single-Mol-
ecule Fluorescence Imaging. Biophys. J. 96, 2371–2381.
2
5. Alejo, J. L., S. C. Blanchard and O. S. Andersen (2013) Small-Mol-
ecule Photostabilizing Agents are Modifiers of Lipid Bilayer Proper-
ties. Biophys. J. 104, 2410–2418.
6. Altman, R. B., D. S. Terry, Z. Zhou, Q. Zheng, P. Geggier, R. A.
Kolster, Y. Zhao, J. A. Javitch, J. D. Warren and S. C. Blanchard
2
(
2012) Cyanine fluorophore derivatives with enhanced photostability.
Nat. Methods 9, 68–71.
2
2
7. Altman, R. B., Q. Zheng, Z. Zhou, D. S. Terry, J. D. Warren and S.
C. Blanchard (2012) Enhanced photostability of cyanine fluorophores
across the visible spectrum. Nat. Methods 9, 428–429.
8. Zheng, Q., S. Jockusch, Z. Zhou, R. B. Altman, J. D. Warren, N. J.
Turro and S. C. Blanchard (2012) On the Mechanisms of Cyanine
Fluorophore Photostabilization. The Journal of Physical Chemistry
Letters 3, 2200–2203.
2
3
9. Gollmer, A., J. Arnbjerg, F. H. Blaikie, B. W. Pedersen, T. Breiten-
bach, K. Daasbjerg, M. Glasius and P. R. Ogilby (2011) Singlet
39. Blanchard, S. C. (2012) Reply to “‘Self-healing’ dyes: Intramolec-
ular stabilization of organic fluorophores”. Nat Meth 9, 427–428.
40. Chandler, D. (1998) Electron transfer in water and other polar envi-
ronments, how it happens. In Classical and Quantum Dynamics in
Condensed Phase Simulations. (Edited by B. J. Berne, G. Ciccotti
and D. F. Coker), pp. 25–49. World Scientific, Singapore.
â
Oxygen Sensor Green : Photochemical Behavior in Solution and in
a Mammalian Cell. Photochem. Photobiol. 87, 671–679.
0. Mujumdar, R. B., L. A. Ernst, S. R. Mujumdar, C. J. Lewis and A.
S. Waggoner (1993) Cyanine dye labeling reagents: Sulfoindocya-
nine succinimidyl esters. Bioconjugate Chem. 4, 105–111.