Design of a practical fluorescent probe for superoxide based on protection-deprotection chemistry of fluoresceins with benzenesulfonyl protecting groups

Article abstract of DOI:10.1002/chem.200600522

A strategy for designing probes based on protection-deprotection chemistry involving fluoresceins and their benzenesulfonyl (BES) derivatives has led to the development of a much more practical superoxide (O₂^-.) probe than the previously reported bis(2,4-dinitro-BES) tetrafluorofluorescein (6a). Examination of various BES derivatives, developed from the starting point of the prototype probe 6a. yielded 4,5-dimethoxy-2-nitro-BES tetrafluorofluorescein (BESSo; 7j) as the optimal reagent. A microtiter plate assay with BESSo showed a tenfold improved detection limit for O ₂^-. compared with such an assay based on 6a. BESSo showed markedly better specificity for O₂^-. than for GSH or other reactive oxygen species, and this specificity was significantly higher than that of Fe²⁺ and some reducing enzymes. These features have resulted in the development of an assay based on BESSo that is capable of providing more unambiguous results for O₂^. release from neutrophils, with or without stimulation by phorbol myristate acetate, as compared with an assay based on 6a. Intracellular generation of O₂^. in human Jurkat T cells stimulated by butyric acid has been measured by using flow cytometry and fluorescence microscopy utilizing the acetoxymethyl derivative of BESSo.

Full text of DOI:10.1002/chem.200600522

DOI: 10.1002/chem.200600522
Design of a Practical Fluorescent Probe for Superoxide Based on Protection–
Deprotection Chemistry of Fluoresceins with Benzenesulfonyl Protecting
Groups
Hatsuo Maeda,*[a] Kayoko Yamamoto,[a] Iho Kohno,[a] Leila Hafsi,[a] Norio Itoh,[a]
Shinsaku Nakagawa,[a] Naoko Kanagawa,[a] Keiichiro Suzuki,[b] and Tadayuki Uno[a]
Abstract:
A
strategy for designing
assay with BESSo showed a tenfold im-
proved detection limit for O2 com-
pared with such an assay based on 6a.
These features have resulted in the de-
velopment of an assay based on BESSo
that is capable of providing more un-
ꢀC
probes based on protection–deprotec-
tion chemistry involving fluoresceins
and their benzenesulfonyl (BES) deriv-
atives has led to the development of a
ꢀC
ꢀC
BESSo showed markedly better specif-
ambiguous results for O2 release from
ꢀC
icity for O2 than for GSH or other re-
neutrophils, with or without stimulation
by phorbol myristate acetate, as com-
pared with an assay based on 6a. Intra-
much more practical superoxide (O2
)
active oxygen species, and this specific-
ity was significantly higher than that of
Fe2+ and some reducing enzymes.
probe than the previously reported
bis(2,4-dinitro-BES) tetrafluorofluores-
cein (6a). Examination of various BES
derivatives, developed from the starting
point of the prototype probe 6a, yield-
ed 4,5-dimethoxy-2-nitro-BES tetra-
fluorofluorescein (BESSo; 7j) as the
optimal reagent. A microtiter plate
ꢀC
cellular generation of O2 in human
Jurkat T cells stimulated by butyric
acid has been measured by using flow
cytometry and fluorescence microscopy
utilizing the acetoxymethyl derivative
of BESSo.
Keywords: fluoresceins
cence spectroscopy
probes molecular design
superoxide
·
fluores-
·
fluorescent
·
·
Introduction
py, microplate readers, and cell sorters, which can supply
spatial, temporal, or quantitative information. Hydroethi-
dine (dihydroethidine, HE) has been widely used as a fluo-
rescent probe.[2–4] The major drawback of HE is its poor se-
ꢀC
Reactive oxygen species (ROS) such as superoxide (O2 ),
hydrogen peroxide (H2O2), and hydroxyl radicals (HO ) are
C
ꢀC
important mediators in pathological conditions caused by
some diseases.[1] Chemiluminescence- and fluorescence-
based assays have been widely used to measure cell-derived
ꢀC [2]
lectivity for O2 . Oxidation of HE to fluorescent ethidium is
ꢀ
[3d]
C
also induced by peroxynitrite (ONOO ), HO , tBuOOH,
H2O2 in the presence of peroxidase,[3b] and cytochrome c.[4]
In addition, HE appears to catalyze the dismutation of
ꢀC [4]
O2
.
Although less sensitive than chemiluminescent meth-
ꢀC
ods, by detecting O2 with fluorescent probes one may ad-
vantageously exploit the benefits of fluorescence microsco-
O2
.
These reactions limit the applicability of HE as a
ꢀC
probe for the quantitative measurement of O2 , which is
mainly achieved through an oxidation-based fluorescing
mechanism. O2 acts not only as an oxidant but also as a re-
ꢀC
[a]Prof. Dr. H. Maeda, K. Yamamoto, I. Kohno, L. Hafsi, Dr. N. Itoh,
Prof. Dr. S. Nakagawa, N. Kanagawa, Prof. Dr. T. Uno
Graduate School of Pharmaceutical Sciences, Osaka University
1–6 Yamada-oka, Suita, Osaka, 565-0871 (Japan)
Fax : (+81)6-6879-8206
ductant, and such a dual ability is not observed with other
ROS. This suggests that a reduction-based mechanism could
offer higher specificity in fluorescent probes for the detec-
ꢀC
tion of O2 as compared with an oxidation-based mecha-
E-mail: h-maeda@phs.osaka-u.ac.jp
nism. However, no fluorescent probe for detecting cell-de-
ꢀC
[b]Prof. Dr. K. Suzuki
rived O2 has been developed due to difficulties encoun-
Department of Biochemistry, Hyogo College of Medicine
1-1 Mukogawa-cho, Nishinomiya, 663-8501 (Japan)
tered when using spectrophotometric probes such as cyto-
chrome c and nitro blue tetrazolium (NBT), the absorbances
of which are bathochromically shifted as a result of reduc-
tion by ascorbic acid, glutathione (GSH), and various reduc-
Supporting information for this article, including details of the experi-
mental procedures, is available on the WWW under http://www.che-
meurj.org/ or from the author.
1946
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1946 – 1954

FULL PAPER
tases such as cytochrome P450 (CYP) reductase/NADPH as
ꢀC [2b,c,5]
This strategy has been successfully applied in the design
of novel fluorescent probes for H2O2,[9] thiols,[10] and sele-
nols[11] merely by selecting the appropriate 1 and BES chlor-
ides from pools of compounds with a wide variety of reactiv-
ities (Scheme 2). Perhydrolysis smoothly induced deprotec-
tion of 4 prepared from 2’,7’-difluorofluorescein (1a) and
pentafluoro-BES chloride. HOOꢀ behaves as a stronger nu-
cleophile than HOꢀ (the so-called a effect), to an extent
that depends on the type of electrophile.[12–15] Among the
BES derivatives examined, 4 proved to be the best substrate
for perhydrolysis, although the reason for this is not well un-
well as O2
.
Thus, the focus in developing a fluoromet-
ꢀC
ric detection method for cell-derived O2 is to design a fluo-
rescing mechanism for probes that is not reliant on a redox
reaction. The solution of this issue is of great importance. A
suitable system would be expected to impart a high degree
ꢀC
of specificity to fluorescent probes for detecting O2 in com-
plex biological systems.
Recently, we proposed a novel strategy for the design of
fluorescent probes based on protection–deprotection
chemistry involving fluoresceins (1) and their benzenesul-
fonyl (BES) derivatives (Scheme 1). Our strategy has stem-
med from the following observations and suppositions. Com-
pound 1 is highly fluorescent in aqueous solution, whereas 2
and 3 (protected forms of 1 with BES groups) exhibit little
and no fluorescence, respectively. Thus, deprotection of 2 or
3 to yield 1, induced by a chemical reaction highly character-
istic of a certain target molecule, would allow the BES de-
rivatives to function as fluorescent probes for the target
molecule. Our strategy would thus have scope for modifica-
tion of the specificity and sensitivity of the BES derivatives
to be used as fluorescent probes based on a deprotection
mechanism through rational molecular design. In the con-
text of this strategy, the combination of 1 and the BES
groups is likely to be the most important factor in the
design of fluorescent probes, primarily because 1 and the
BES groups function as leaving and electrophilic groups, re-
spectively, upon deprotection of 2 or 3. A matched pair of
leaving and electrophilic groups is required in order to ach-
ieve specific and effective deprotection by a target molecule
(i.e., to impart a high degree of specificity and sensitivity to
2 or 3 as probes for the target molecule). As general meth-
ods for the synthesis of 1[6,7] and BES chlorides[8] have been
established, a pool of compounds with diverse reactivities is
available for the design of a wide variety of analogues of 2
or 3. Our strategy to develop and tune 2 or 3 as a fluores-
cent probe for a target molecule takes full advantage of this
fact. The choice of protection mode, that is, whether 2 or 3
is used, influences the sensitivity of the BES derivatives as
fluorescent probes because mono-protected compound 2 is
more susceptible to deprotection than the corresponding
bis-protected compound 3. In addition, 2 is preferred to 3
with regard to the miscibility of probe solutions in organic
solvents with aqueous systems, because the mono-protected
compound is less hydrophobic.
Scheme 2. Fluorescent probes for hydrogen peroxide, thiols, selenols, and
superoxide developed by the described optimization strategy, with indica-
tions of the initial reactions responsible for the fluorescing mechanism.
Scheme 1. Proposed strategy for the design of fluorescent probes based on protection–deprotection chemistry involving fluoresceins (1) and their BES
derivatives (2 or 3).
Chem. Eur. J. 2007, 13, 1946 – 1954
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1947

H. Maeda et al.
derstood. By exploiting perhydrolysis rather than oxidation
as a fluorescing reaction, 4 is capable of functioning as a flu-
orescent probe for H2O2, showing relatively high specificity
and sensitivity.[9] Our strategy was also applied to the design
of a red fluorescent probe for H2O2, namely bis(4-methyl-
BES) naphthofluorescein.[16] As an alternative nonoxidative
fluorescing mechanism to perhydrolysis of BES groups, it
has recently been demonstrated that deprotective hydroxyl-
ation of boronates is useful for the design of H2O2 fluores-
cent probes.[17,18]
When a combination of 2’,7’-dimethylfluorescein (1b) and
a 2,4-dinitro-BES group was selected, nucleophilic aromatic
substitution[19,20] by thiols predominated over other reactions
in the deprotection of the corresponding BES derivative
5.[10] Compared with existing fluorescent[20] or chemilumines-
cent[21] thiol probes, the miscibility of an organic solution of
5 with an aqueous system was much better; hence 5 proved
useful as a thiol probe in thiol quantification-based homoge-
neous assays for cholinesterase at pH 7.4. The assay with 5
worked well as a simple method for determining the inhibi-
tory constants of inhibitors such as donepezil towards
acetyl- and butyrylcholinesterases. For the design of a thiol
probe based on this strategy, 1b was preferred to fluores-
ceins bearing electron-withdrawing groups, such as 1a. This
is consistent with the trend in nucleophilic aromatic substi-
tution that reaction proceeds more effectively with de-
creased acidity of the leaving group.[19,20] The lower acidity
of 1b relative to 1a also helped to prevent the correspond-
ing 2,4-dinitro-BES derivative from undergoing hydrolysis
(i.e., providing large blank responses).
enzyme systems such as xanthine oxidase (XO)/hypoxan-
thine (HPX) in the presence of superoxide dismutase
(SOD), CYP reductase/NADPH, and diaphorase/NADH.
The extent of these reactions was 6–18% of that observed
ꢀC
for O2 . These results probably preclude the use of 6a in
ꢀC
the specific measurement of O2 in biological systems more
complex than the experimental cell system. The fluorescent
response towards GSH at a non-negligible level is likely to
ꢀC
compromise the high specificity of 6a for O2 among other
ꢀC
ROS in intracellular measurement of O2 using this probe,
because GSH is ubiquitous in cells at mm levels.[24] Thus, fur-
ther tuning of the probe performance of 6a through struc-
tural modification of its BES group has been conducted
ꢀC
with a view to developing a practical O2 probe that not
only shows high specificity over GSH as well as other ROS,
but also shows improved specificity over Fe2+, CYP reduc-
tase/NADPH, and diaphorase/NADH. Based on our strat-
egy presented herein, various BES derivatives have been de-
signed in a stepwise manner from 6a, and their performan-
ꢀC
ces as O2 fluorescent probes have been examined. Our
strategy based on protection–deprotection chemistry has
again worked well for the rational design of fluorescent
ꢀC
probes, and a more practical O2 probe than the prototype
probe 6a, namely 2-nitro-4,5-dimethoxy-BES tetrafluoro-
fluorescein (7j; BESSo), has been obtained through screen-
ing of these carefully designed candidates (Scheme 3).
Interestingly, 5 has Janus-faced ability to probe selenols as
well as thiols. These dual abilities of 5 can be selected
merely by changing the pH of the medium. Compound 5
functions as a selenol probe at pH 5.8, facilitating not only
the assay of selenocysteine, generated through in situ reduc-
tion of selenocystine by dithiothreitol, but also the measure-
ment of selenocysteine residues in selenoproteins such as
glutathione peroxidase and mammalian thioredoxin reduc-
tase after denaturing with guanidine.[11]
ꢀC
A fluorescent probe for O2 has also been developed
based on this strategy. It was found that a combination of
2’,4’,5’,7’-tetrafluorofluorescein (1c) and 2,4-dinitro-BES
groups could be favorably employed for the deprotection of
BES derivatives according to a nonredox mechanism based
ꢀC
on nucleophilic substitution by O2 . In this case, bis(2,4-di-
nitro-BES) tetrafluorofluorescein (6a) showed better probe
performance than the corresponding mono-protected deriva-
tive, because the latter was prone to hydrolysis, resulting in
significantly high blank responses. The transformation of 6a
into 1c represents a fluorescing reaction that is highly spe-
ꢀC
cific and sensitive towards O2 as opposed to other ROS
(Scheme 2).[23] As a result of this high specificity, 6a can
ꢀC
serve as a fluorescent probe for measuring extracellular O2
Scheme 3. Probe candidates examined in this study.
released from neutrophils stimulated by phorbol myristate
acetate (PMA). However, the specificity of this prototype
probe had yet to be optimized. Fluorescence augmentation
with 6a also occurred through reaction with GSH, Fe2+, or
1948
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1946 – 1954

A Practical Fluorescent Probe for Superoxide
FULL PAPER
Results and Discussion
than is a 4-nitro-BES group. These implications were con-
firmed by similar experiments using the mono-protected
compounds 7b and 7c. These compounds were examined to
provide higher fluorescence responses toward O2 than the
bis-protected compounds 6b and 6c. Both of the mono-pro-
2,4-Dinitro-BES versus 2- or 4-nitro-BES derivatives: In
order to solve the selectivity problem of 6a, a synthetic
study employing 2,4-dinitro-, 2-nitro-, and 4-nitro-BES func-
tionalities as protecting groups for amines was informa-
tive,[25–27] as it suggested that bis-protected derivatives of 1c
with a 2- or 4-nitro-BES group would exhibit a higher spe-
ꢀC
tected compounds did indeed show increased fluorescence
ꢀC
upon reaction with O2 relative to 6b and 6c. In addition,
ꢀC
7b and 7c provided FI values against O2 more than five-
ꢀC
cificity for O2 over GSH. To confirm this, fluorescence in-
fold greater than in the case of 6a, although the RFI was
lower than that of 6a (entries 4 and 5). However, 7c was su-
ꢀC
tensities after reaction of 6b or 6c with O2 and GSH (final
ꢀC
concentration 50 mm) at 378C for 10 min were compared
with the intensity in the case of 6a. In addition, the effect of
SOD (final concentration 50 UmLꢀ1) on the fluorescence in-
perior to 7b as a fluorescent probe for O2 . Compared with
ꢀC
7b, 7c exhibited higher sensitivity toward O2 , less suscepti-
bility to deprotection by GSH as well as by the reduced
form of XO, and a lower control response. In fact, 7c
showed no response to GSH, while providing a satisfactorily
ꢀC
tensities after reaction with O2 was also examined. All
measurements were made on solutions in 96-well microtiter
plates using a fluorescent plate reader, with excitation and
emission wavelengths set at 505 and 544 nm, respectively. A
working probe solution was prepared by diluting each probe
solution (5 mm in DMSO) 200 times with pH 7.4 HEPES
ꢀC
high response to O2 . This observation gives a strong indica-
tion of the potential utility of a 2-nitro-BES group as a pro-
ꢀC
tecting group in designing a practical O2 probe. However,
problems were encountered with 7c compared with proto-
type probe 6b: 1) a relatively higher control response, 2) a
ꢀC
buffer. For the generation of O2 , an XO (final concentra-
ꢀC
tion 13 mUmLꢀ1)/HPX (final concentration 50 mm) system
lower RFI on reaction with O2 , and 3) a higher response to
O2 in the presence of SOD.
ꢀC
was used. The results are summarized in Table 1, in which
2-Nitro-BES versus 4-methyl- or 4-alkoxy-2-nitro-BES de-
rivatives: Incubating 7b and 7c in pH 7.4 HEPES buffer at
378C for 60 min resulted in 9% and 6% decomposition, re-
spectively, to 1c. Under similar conditions, 6a underwent
only 0.5% decomposition. These results indicate that the
high control responses observed for 7b and 7c can be attrib-
uted to their hydrolysis to form 1c. In general, the suscepti-
bility of benzenesulfonates to alkaline hydrolysis decreases
with increasing electron density of the BES benzene rings.[29]
Therefore, modification of the 2-nitro-BES group with alkyl
or alkoxy groups should prevent the high control response
of 7c. This modification can also be expected to reduce the
Table 1. Relative quantum efficiencies of 6a–6c and 7b–7k and relative
fluorescent intensities (RFI) upon reaction with O2 with and without
SOD and GSH.
ꢀC
Entry Compound Ffl
FI (au) for
control[a]
RFI (times control)
O2 +SOD
ꢀC
ꢀC
O2
GSH
1
2
3
4
5
6
7
8
6a
6b
6c
7b
7c
7d
7e
7 f
7g
7h
7i
0
0
0
112
51
41
73.7
1.8
1.8
8.3
1.6
1.0
13.5
8.7
14.5
23.7
12.6
6.6
5.5
4.8
9.0
2.7
12.8
1.0
1.0
1.4
1.0
1.0
1.1
1.1
1.1
1.1
1.2
1.1
1.1
0.0015
0.0004
0.0002
0.0003
0.0003
0.0003
0.0002
0.0002
0.0006
0.0003
2652
1104
629
172
157
229
330
230
239
297
15.8
47.9
65.3
144.6
73.0
53.9
33.3
9.7
ꢀC
9
response of 7c to O2 in the presence of SOD, because fluo-
10
11
12
13
rescent responses of the BES derivatives to the reduced
form of XO probably involve electron transfer between the
BES groups and the enzyme. Such electron-transfer reac-
tions to the BES groups will become unfavorable with an in-
crease in the electron density of the BES benzene rings. The
molecular design for imparting resistance to hydrolysis and
electron transfer reactions may also decrease the fluorescent
7j
7k
103.4
38.0
[a]Fluorescent intensity (FI) (control responses) observed on incubating
each compound alone in pH 7.4 HEPES buffer.
ꢀC
all fluorescent responses are shown as relative fluorescent
intensities (RFI), that is, relative to the control, rather than
as absolute values (FI). The control responses were obtained
after merely incubating each BES derivative (final concen-
tration 21.3 mm) in pH 7.4 HEPES buffer at 378C for 10 min.
Quantum efficiencies (Ffl) estimated relative to the Ffl
(0.85) of fluorescein in 0.1m NaOH as a standard[28] are also
included in the table. Compared with 6a, each of the reac-
tions led only to a small increase in fluorescence from 6b
and 6c (entries 1–3). Although the response levels were too
low to evaluate unequivocally, the results implied the fol-
lowing: 1) 2- and 4-nitro-BES groups are deprotected not by
responses of BES derivatives toward O2 . However, achiev-
ing the highest RFI with the lowest blank response rather
than the highest FI with a relatively high blank response
was believed to be a more important criterion for develop-
ing a practical probe.
Methyl, methoxy, ethoxy, and isopropyloxy groups were
chosen as electron-donating groups and introduced at the 4-
position of the 2-nitro-BES group. The performances of 7d–
ꢀC
7g as O2 probes were then evaluated as above. The results
are also included in Table 1. As expected, these compounds
proved to be more stable to simple hydrolysis, providing
smaller control responses as compared with 7c (entries 6–9).
The alkoxy groups caused greater attenuation of the blank
responses as compared with the methyl group. Compound
ꢀC
GSH, but by O2 . 2) A 2-nitro-BES group is less susceptible
to deprotection by reaction with the reduced form of XO
Chem. Eur. J. 2007, 13, 1946 – 1954
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1949

H. Maeda et al.
ꢀC
7e exhibited the best RFI values toward O2 (entry 7).
encountered with 7e caused by reactivity toward the re-
duced form of XO, allowing 7j to provide superior probe
performance. The O2 -induced RFI with 7j was approxi-
Comparing the probe performance using 7e with that in the
case of 7c, the methoxy group on the 2-nitro-BES group re-
duced the blank response by 84%, while reducing the abso-
lute FI value toward O2 by only 47%. The effect of the
methoxy group on FI toward O2 was smaller compared
with that of any other substitution, and the overall effect of
the substituents on 7e resulted in a favorable RFI value
toward O2 that was threefold greater than in the case of
ꢀC
mately 70% of that with 7e, while 7j allowed SOD to scav-
ꢀC
ꢀC
enge O2 twice as effectively as 7e. It should be mentioned
ꢀC
here that incubating 7j in pH 7.4 HEPES buffer at 378C for
1 h led to only 2% decomposition to 1c. We concluded that
7j (BESSo) represents the optimal fluorescent probe for
ꢀC
ꢀC
O2 based on protection–deprotection chemistry.
7c. The reactivities of 7d–7g toward the reduced form of
XO were not attenuated as much as the control responses.
The ratios of the RFI values from reactions of 7d–7g with
an XO/HPX system in the presence and absence of SOD
were similar to that observed for 7c.
Detailed probe performance of BESSo: A 96-well microtiter
plate assay with BESSo (final concentration 21.3 mm) provid-
ꢀC
ed a highly sensitive method for measuring O2 generated
by an XO (final concentration 26 mUmLꢀ1)/HPX system in
pH 7.4 HEPES buffer at 378C for 10 min. The detection
ꢀC
4-Methoxy- versus 4,5-dimethoxy-2-nitro-BES derivatives:
The problems encountered with 7c regarding its control re-
sponse and the relatively low RFI on reaction with O2
limit corresponded to the amount of O2 generated from
HPX at 0.1 pmol/well (relative standard deviation (RSD),
n=8; 3.5%), estimated as the lowest concentration afford-
ing fluorescent augmentation threefold greater than the
standard deviation of blank responses. The use of BESSo in-
stead of 6a improved the detection limit of the fluorescent
assay for O2 by a factor of ten. A linear calibration curve
for O2 was obtained over the range from 1.0 to 1000 pmol
HPX/well, with a correlation coefficient of 0.997 and a slope
of 2.69 aupmolꢀ1. Although the linear concentration range
produced by BESSo was similar to that obtained with 6a,
ꢀC
were solved by introducing a 4-methoxy substituent on the
2-nitro-BES group. A further attempt to attenuate the rela-
tively high fluorescence augmentation observed on reaction
ꢀC
ꢀC
of 7e with O2 in the presence of SOD was made by intro-
ꢀC
ducing another methoxy substituent on the 4-methoxy-2-
nitro-BES group. Considering practical access to BES chlor-
ides, a favorable position to introduce the second methoxy
group was identified by comparing probe performances of
dimethyl-2-nitro-BES derivatives 7h and 7i. As also shown
in Table 1, the performance of 7i was markedly poorer than
that of 7h (entries 10 and 11). It was considered likely that
the presence of another methyl group at the 6-position of
the 4-methyl-2-nitro-BES group would result in steric hin-
ꢀC
the sensitivity of BESSo toward O2 , as expressed by the
slope of the calibration curve, was approximately threefold
greater than that of 6a.
Next, we compared the reactivities of BESSo (final con-
ꢀC
centration 21.3 mm) toward O2 , other ROS (final concen-
ꢀC
drance of the reaction with O2 , and that this effect would
trations 50 mm), GSH (final concentration 50 mm), Fe2+
(250 mm), and reductases such as CYP reductase (final con-
centration 68 mUmLꢀ1)/NADPH (final concentration
outweigh the electron-donating effect disfavoring reaction
with the reduced form of XO. In contrast to the 6-position,
additional introduction of a methyl group at the 5-position
would reduce the response on reaction with the XO/HPX
50 mm) and diaphorase (65 mUmLꢀ1)/NADH (final concen-
1
C
C
tration 50 mm) in pH 7.4 HEPES buffer. HO , O2, NO , and
ꢀC
system in the presence of SOD. The response of 7h to O2
ONOOꢀ were generated in situ by reaction of H2O2 (final
was decreased both in the absence and in the presence of
concentration 50 mm) and Fe(ClO4)2 (final concentration
A
SOD. However, 7h provided a higher ratio between the RFI
values for reaction with O2 in the absence and presence of
250 mm), H2O2 (final concentration 50 mm) and NaOCl (final
ꢀC
concentration 50 mm),[30] 3-(3-aminopropyl)-1-hydroxy-3-iso-
SOD as compared with 7d (entries 6 and 10). These results
suggest that the 4,5-positions are preferable to the 4,6-posi-
tions for the introduction of two methoxy substituents on
the 2-nitro-BES group of 7c. Accordingly, the probe perfor-
mance of 7j with a 4,5-dimethoxy-2-nitro-BES protecting
group was evaluated. Its diethoxy counterpart 7k was sub-
jected to a similar evaluation, to confirm that a methoxy
group was preferable to an ethoxy group for probe perfor-
mance of 4,5-disubstituted 2-nitro-BES derivatives, as in the
case of 4-substituted 2-nitro-BES compounds. Indeed, 4,5-di-
methoxy substitution imparted a greater degree of probe
performance compared with 4,5-diethoxy substitution. Com-
propyl-2-oxo-1-triazene
(NOC-5; final concentration
50 mm),[31] and 3-morpholinosydnonimine (SIN-1; final con-
centration 50 mm),[32] respectively. The results are summar-
ized in Figure 1, in which the observed fluorescent responses
are shown as percentages of the fluorescence intensity in re-
ꢀC
sponse to O2 produced by BESSo (24700 au). Note that a
value of 1% indicates no difference in fluorescent intensity
from the blank response. The previously reported reactivi-
ties[23] of 6a toward these compounds and enzyme systems,
similarly expressed with respect to the fluorescence intensity
ꢀC
(9000 au) toward O2 , are also shown in Figure 1. It can be
seen that BESSo exhibited a highly specific response to
ꢀC
ꢀC
pound 7j showed a higher RFI toward O2 than did 7k,
O2 , not only over GSH, but also over H2O2, tBuOOH,
1
ꢀ
C
while these compounds produced similar ratios between the
NaOCl, O2, NO , and ONOO . The specificity of BESSo
toward these ROS was even better than that of 6a. Our pre-
vious work raised the possibility that Fe2+, CYP reductase/
NADPH, and diaphorase/NADH might affect the highly
ꢀC
RFI values toward O2 in the absence and presence of SOD
(entries 12 and 13). In addition, modification of 7e by a fur-
ther methoxy substituent partially alleviated the problem
1950
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1946 – 1954

A Practical Fluorescent Probe for Superoxide
FULL PAPER
Figure 1. Comparison of fluorescence augmentations observed on reac-
tion of BESSo or 6a with ROS, reductases, and GSH. All data are shown
as percentages of the fluorescence intensity (FISO) produced by BESSo
ꢀC
or 6a in response to O2
.
Figure 2. HPLC chromatograms of 1c (25.2 mm) (a), 8 (24.9 mm) (b),
BESSo (25.6 mm) (c), and products generated by the reaction of BESSo
(25.0 mm) with KO2 (d).
ꢀC
specific determination of O2 by 6a.[23] In particular, 6a
showed fluorescence augmentation in response to Fe2+ and
CYP reductase/NADPH amounting to more than 15% of
ꢀC
the augmentation observed in response to O2 . The prob-
lems associated with this reducing reagent and these en-
zymes were alleviated by the use of BESSo instead of 6a.
The fluorescence responses of BESSo to Fe2+, CYP reduc-
tase/NADPH, and diaphorase/NADH were less than half of
those seen with 6a. The levels of fluorescence augmentation
for the reactions of BESSo and 6a with a Fenton (H2O2/
Fe2+) system were similar. The response of 6a to a Fenton
dimethoxy-2-nitrobenzenesulfonic acid (8) in HPLC yields
of 91% and 96%, respectively. This result suggests that
BESSo is deprotected through nucleophilic substitution by
O2 , leading to the formation of 1c and 8, as in the case of
6a (Scheme 4). Thus, BESSo is transformed to 1c by reac-
ꢀC
system was attributed to reaction with Fe2+ rather than
2+
C
HO , because Fe alone induced greater fluorescent aug-
mentation than the Fenton system. In contrast, BESSo pro-
vided greater fluorescent intensity in response to a Fenton
system than to Fe2+, suggesting that in this case HO is re-
C
sponsible for the fluorescence augmentation. These results
ꢀC
demonstrate that BESSo is a more practical O2 probe than
ꢀC
Scheme 4. Proposed mechanism of deprotection of BESSo by O2 to
the prototype probe 6a in terms of specificity and sensitivity.
In addition, the solubility of BESSo in buffer is higher than
that of 6a: when each of the probe solutions in DMSO was
diluted 200 times with pH 7.4 HEPES buffer, 6a was almost
saturated at a final concentration of around 25 mm, while the
maximum final concentration of BESSo was found to be
200 mm. Although we cannot rule out the possibility that the
yield 1c.
ꢀC
tion with O2 as a nucleophile rather than as a reducing
agent, which requires one equivalent of O2 for the forma-
tion of one molecule of 1c. Transformation of 6a to 1c re-
quires two equivalents of O2 based on a similar mecha-
nism. Thus, BESSo is also preferred to 6a as an O2 probe
with regard to stoichiometry.
ꢀC
ꢀC
ꢀC
coexistence of HO , Fe2+, and CYP reductase/NADPH
C
ꢀC
might impair the specific determination of O2 by BESSo,
this possibility is believed to be relatively low and should
The rate constant, kobsd, for the conversion of BESSo to
ꢀC
ꢀC
not limit the application of BESSo as an O2 probe. These
1c by O2 was estimated by competitive kinetic analy-
ꢀC
detrimental compounds are not always present where O2 is
ses,[33–35] as are generally applied to determine rate constants
ꢀC
generated and the advantages of BESSo are greater than
those of HE.
for O2 probes or scavengers. As a competitor to BESSo,
SOD was used. A saturated final concentration of BESSo of
0.3 mm was used. The deployment of BESSo at this high
ꢀC
Analysis of mechanism and kinetics: BESSo was subjected
to reaction with KO2 (2.0 equiv) in pH 7.4 HEPES buffer.
Immediately after initiation of the reaction at room temper-
ature, the mixture turned fluorescent green and BESSo was
completely consumed, as shown by HPLC analysis of the re-
action mixture (Figure 2). The reaction yielded 1c and 4,5-
concentration allowed negligible dismutation of O2 , and
ꢀC
produced a rate of 1c formation equivalent to that of O2
formation in the absence of SOD. Based on such competi-
tive kinetic analyses, a steady-state approximation under
this assumption allowed definition of the ratio of the rates
ꢀC
of formation of 1c from reaction of BESSo and O2 in the
Chem. Eur. J. 2007, 13, 1946 – 1954
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1951

H. Maeda et al.
absence and the presence of SOD (V0 and V, respectively)
as shown in Equation (1):
response to PMA-stimulated neutrophils than in response to
unstimulated cells 10 min after incubation. The responses
observed to the stimulated cells with both probes were re-
duced upon addition of SOD. These results indicate that the
fluorescence responses observed with 6a and BESSo to
V0
V
kSOD½SODŠ
ð1Þ
¼ 1 þ
kobsd½BESSoŠ
ꢀC
PMA-stimulated neutrophils result from O2 release. How-
ever, the differences between fluorescence augmentation
from stimulated and unstimulated cells, and from stimulated
cells in the absence and presence of SOD, were larger with
BESSo than with 6a. Thus, BESSo showed improved specif-
icity and sensitivity and hence represents a more practical
ꢀC
in which kSOD is the rate constant for the reaction of SOD
with O2 , and [SOD]and [BESSo]are the concentrations of
SOD and BESSo, respectively. According to this equation,
ꢀC
kobsd was estimated from the slope of a linear plot of V0/V
ꢀC
versus [SOD]. Thus, reactions of BESSo and O2 in the
probe than 6a for fluorescent assays of extracellular O2
.
presence of SOD at various concentrations were followed
fluorometrically, affording V0 and V values. A typical plot
of V0/V versus [SOD]is shown in Figure 3. When the report-
ꢀC
Detection of intracellular O2 : We also applied BESSo in
ꢀC
the detection of intracellular O2 generation. For this pur-
pose, its acetoxymethyl derivative (BESSo-AM) was synthe-
Figure 3. A typical plot of V0/V versus concentration of SOD for estimat-
ing kobsd
.
sized, which, it was envisaged, would be better able to per-
meate into cells than BESSo itself, wherein it would be
transformed into BESSo by the action of intracellular ester-
ase. The maximum final concentration of a BESSo-AM
working solution was about 50 mm when prepared by diluting
a stock solution in DMSO 150 or 200 times with buffer.
Jurkat T cells were shown to undergo apoptosis upon treat-
ment with butyric acid, which induces the production of
ROS as well as ceramide in the cytosol.[37] Flow cytometry
and fluorescence microscopy with BESSo-AM clearly indi-
ed value (3.1”109 mꢀ1 sꢀ1)[36] of kSOD was used, kobsd was esti-
mated to be 4.0ꢁ0.2”103 mꢀ1 sꢀ1. The rate constant for HE
has been reported to be 2.6ꢁ0.6”105 mꢀ1 sꢀ1.[34] Although
the rate constant for BESSo was thus 65 times smaller than
that for HE, the specificity and sensitivity of BESSo as an
ꢀC
O2 probe are better than those of HE.
ꢀC
Detection of extracellular O2 : The usefulness of BESSo as
ꢀC
a probe for fluorescence-based assays of extracellular O2
ꢀC
was compared with that of 6a in experiments using neutro-
phils stimulated with PMA. A cell suspension (1.0”105 cells
per well) was incubated at 378C with BESSo or 6a in the
presence or absence of PMA. As shown in Figure 4, an
assay with 6a or BESSo produced greater fluorescence in
cated that the intracellular O2 levels in Jurkat T cells
became high when the cells were stimulated with butyric
acid. Jurkat T cells (2”106 cellsmLꢀ1, 500 mL) were incubat-
ed with BESSo-AM (final concentration 33 mm) at 378C for
1 h. The probe-loaded cells were further incubated in the
presence or absence of butyric acid (final concentration
5 mm) at 378C for 1 h, and were then subjected to the fluo-
ꢀC
rescence-based analysis for O2 . Flow cytometric measure-
ments yielded 793 and 331 au as the mean fluorescence in-
tensity values observed for stimulated and unstimulated
cells, respectively. When cells were loaded with 4,5-dihy-
droxy-1,3-benzenedisulfonic acid (Tiron, a cell-permeable
ꢀC
O2 scavenger) as well as BESSo-AM, and then subjected
to stimulation with butyric acid, the mean fluorescence in-
tensity value was reduced to 482 au. The phenomena ob-
served by flow cytometry of Jurkat T cells could be clearly
visualized by means of fluorescence microscopy. Represen-
tative phase contrast and fluorescence images obtained are
shown in Figure 5. Stimulation with butyric acid increased
the number of cells stained by the fluorescent product 1c.
Figure 4. Temporal changes in fluorescence intensities (FI) observed with
6a (a) and BESSo (b) for PMA-stimulated or unstimulated human neu-
trophils (1”105 cells per well). Data are expressed as mean ꢁ standard
deviations (n=8).
1952
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1946 – 1954

A Practical Fluorescent Probe for Superoxide
FULL PAPER
orescent probes useful for extra- and intracellularly generat-
ꢀC
ed O2 , respectively.
This study has also confirmed the utility of a strategy
based on protection–deprotection chemistry for the design
of fluorescent probes, which would be difficult to achieve by
other available methodologies. This strategy can provide
novel fluorescent probes with tunable sensitivities and spe-
cificities by the appropriate selection of fluoresceins and
BES chlorides, rather than by using totally different chemi-
cal structures. The pools of available fluoresceins and BES
chlorides from which one may choose are diverse. This
raises the possibility that this strategy may be more general-
ly applied for the design of fluorescent probes for target
molecules that show specific reactivity in inducing the de-
protection of BES derivatives. Additional studies can be ex-
pected to result in the development of novel fluorescent
probes for other target molecules.
Acknowledgements
This work was supported in part by a Grant-in-Aid for Scientific Re-
search (B) (15390012) from the Ministry of Education, Culture, Sports,
Science and Technology, Japan, and by research grants from the Suntory
Institute for Bioorganic Research, and the Mochida Memorial Founda-
tion for Medical and Pharmaceutical Research.
Figure 5. Phase contrast (b, d, f) and fluorescence images (a, c, e) ob-
tained after incubating human Jurkat T cells loaded with BESSo-AM at
378C for 1 h in the absence (a, b) or the presence (c–f) of 5 mm butyric
acid. (e, f) Cells loaded with Tiron as well as BESSo-AM were used.
[1]B. Halliwell, J. Gutteridge, Free Radicals in Biology and Medicine,
3rd ed., Clarendon Press, Oxford, 1999.
Tiron markedly inhibited cell staining induced by stimula-
tion with butyric acid. Since Tiron clearly functioned as an
intracellular scavenger of O2 in these experiments, the re-
[2]Selected recent reviews: a) M. Nakano, Cell. Mol. Neurobiol. 1998,
18, 565–579; b) C. L. Murrant, M. B. Reid, Microsc. Res. Tech. 2001,
55, 236–248; c) M. M. Tarpey, I. Fridovich, Circ. Res. 2001, 89, 224–
236; d) T. Münzel, I. B. Afanasꢁev, A. L. Klescchyov, D. G. Harrison,
Arterioscler. Thromb. Vasc. Biol. 2002, 22, 1761–1768; e) M. D. Es-
posti, Methods 2002, 26, 335–340.
ꢀC
sults demonstrate that BESSo-AM is capable of serving as a
ꢀC
fluorescent probe for the detection of intracellular O2 . This
intracellular assay with BESSo-AM thus revealed that ROS
[3]Selected references: a) G. Rothe, G. Valet, J. Leukocyte Biol. 1990,
47, 440–448; b) W. O. Carter, P. K. Narayanan, J. P. Robinson, J.
Leukocyte Biol. 1994, 55, 253–258; c) V. P. Bindokas, J. Jordµn,
C. C. Lee, R. J. Miller, J. Neurosci. 1996, 16, 1324–1336; d) A. B. Al-
Mehdi, H. Shuman, A. B. Fisher, Am. J. Physiol. 1997, 272, L294–
L300.
production in the cytosol of Jurkat T cells originates from
ꢀC
the generation of O2
.
Conclusion
[4]L. Benov, L. Sztejnberg, I. Fridovich, Free Radical Biol. Med. 1998,
25, 826–831.
[5]C. Beauchamp, I. Fridovich, Anal. Biochem. 1971, 44, 276–287.
[6]W.-C. Sun, K. R. Gee, D. H. Klaubert, R. P. Haugland, J. Org.
Chem. 1997, 62, 6469–6475.
[7]Y. Urano, M. Kamiya, K. Kanda, T. Ueno, K. Hirose, T. Nagano, J.
Am. Chem. Soc. 2005, 127, 4888–4894.
[8]R. V. Hoffman, Organic Synthesis, Vol. 60, Wiley, New York, 1981,
pp. 121–126, and references therein.
[9]H. Maeda, Y. Fukuyasu, S. Yoshida, M. Fukuda, K. Saeki, H. Matsu-
no, Y. Yamauchi, K. Yoshida, K. Hirata, K. Miyamoto, Angew.
Chem. 2004, 116, 2443–2445; Angew. Chem. Int. Ed. 2004, 43, 2389–
2391.
[10]H. Maeda, H. Matsuno, M. Ushida, K. Katayama, K. Saeki, N. Itoh,
Angew. Chem. 2005, 117, 2982–2985; Angew. Chem. Int. Ed. 2005,
44, 2922–2925.
[11]H. Maeda, K. Katayama, M. Matsuno, T. Uno, Angew. Chem. 2006,
118, 1842–1845; Angew. Chem. Int. Ed. 2006, 45, 1810–1813.
[12]J. O. Edwards, R. G. Pearson, J. Am. Chem. Soc. 1962, 84, 16–24.
[13]G. Klopman, K. Tsuda, J. B. Louis, R. E. Davis, Tetrahedron 1970,
26, 4549–4554.
We have tested bis- and mono-protected derivatives of 1c
with a number of BES groups, which were selected not only
to eliminate or significantly reduce undesired reactivity of
ꢀC
the prototype O2 probe 6a, but also to validate our strat-
egy based on protection–deprotection chemistry as a novel
concept for probe design. Of the BES derivatives of 1c that
were examined, the 4,5-dimethoxy-2-nitro-BES derivative
ꢀC
(BESSo) proved to be the best O2 probe. Thus, compared
ꢀC
with 6a, BESSo allowed the measurement of O2 with
greater sensitivity, and exhibited greater specificity toward
ꢀC
O2 relative to GSH and to ROS such as H2O2, NaOCl,
1
ꢀ
C
tBuOOH, O2, NO , and ONOO . The use of BESSo also
ꢀC
significantly improved specificity toward O2 over Fe2+ as
well as over the reduced forms of CYP reductase and dia-
phorase. In particular, BESSo exhibited no fluorescent re-
sponse to GSH, which is ubiquitous in cells at mm levels.
These features allow BESSo and BESSo-AM to serve as flu-
[14]A. P. Grekov, V. Ya. Veselov, Russ. Chem. Rev. 1978, 47, 631–648.
[15]S. Hoz, E. Buncel, Isr. J. Chem. 1985, 26, 313–319.
Chem. Eur. J. 2007, 13, 1946 – 1954
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1953

H. Maeda et al.
[16]K. Xu, B. Tang, H. Huang, G. Yang, Z. Chen, P. Li, L. An, Chem.
Commun. 2005, 5974–5976.
[29]R. V. Vizgert, E. K. Savchuk, Zh. Obshch. Khim. 1956, 26, 2268–
2273; [Chem. Abstr. 1957, 25341].
[30]See, for example: a) T. Kajiwara, D. R. Kearns, J. Am. Chem. Soc.
1973, 95, 5886–5890; b) A. M. Held, D. J. Halko, J. K. Hurst, J. Am.
Chem. Soc. 1978, 100, 5732–5740.
[31]M. Feelisch, J. Cardiovasc. Pharmacol. 1991, 17 (Suppl. 3), 25–33.
[32]J. H. Hrabie, J. R. Klose, D. A. Wink, L. K. Keefer, J. Org. Chem.
1993, 58, 1472–1476.
[33]J. Vµsquez-Vivar, J. Whitsett, P. Martµsek, N. Hog, B. Kalyanara-
man, Free Radical Biol. Med. 2001, 31, 975–985.
[34]H. Zhao, S. Kalivendi, H. Zhang, J. Joseph, K. Nithipatikom, J.
Vµsquez-Vivar, B. Kalyanaraman, Free Radical Biol. Med. 2003, 34,
1359–1368.
[35]J. Weaver, P. Tsai, S. Pou, G. M. Rosen, J. Org. Chem. 2004, 69,
8424–8428.
[36]D.-K. Roth, J. Rabani, J. Phys. Chem. 1976, 80, 588–591.
[37]T. Kurita-Ochiai, S. Amano, K. Fukushima, K. Ochiai, J. Immunol.
2003, 171, 3576–3584.
[38]L. K. Mehta, J. Parrick, F. Payne, J. Chem. Soc. Perkin Trans. 1 1993,
1261–1267.
[39]D. S. Wulfman, C. F. Cooper, Synthesis 1978, 924–925.
[40]K. Morita, T. Nakamae, Japanese Patent Application JP 1997-
318984, 1997; [Chem. Abstr. 1997, 131, 25714].
[17]M. C. Y. Chang, A. Pralle, E. Y. Isacoff, C. J. Chang, J. Am. Chem.
Soc. 2004, 126, 15392–15393.
[18]E. W. Miller, A. E. Albers, A. Pralle, E. Y. Isacoff, C. J. Chang, J.
Am. Chem. Soc. 2005, 127, 16652–16659.
[19]G. P. Briner, J. Miller, M. Liveris, P. G. Lutz, J. Chem. Soc. 1954,
1265–1266.
[20]G. Bartoli, P. E. Todesco, Acc. Chem. Res. 1977, 10, 125–132.
[21]R. P. Haugland, Handbook of Fluorescent Probes and Research
Products, 9th ed., Molecular Probe, Eugene, 2002, pp. 79–98.
[22]S. Sabelle, P.-Y. Renard, K. Pecorella, S. de Suzzoni-Dꢂzard, C. Crꢂ-
minon, J. Grassi, C. Mioskowski, J. Am. Chem. Soc. 2002, 124,
4874–4880.
[23]H. Maeda, K. Yamamoto, Y. Nomura, I. Kohno, L. Hafsi, N. Ueda,
S. Yoshida, M. Fukuda, Y. Fukuyasu, Y. Yamauchi, N. Itoh, J. Am.
Chem. Soc. 2005, 127, 68–69.
[24]S. M. Deneke, Curr. Top. Cell. Regul. 2000, 36, 151–180.
[25]T. Fukuyama, C.-K. Jow, M. Cheung, Tetrahedron Lett. 1995, 36,
6373–6374.
[26]T. Fukuyama, M. Cheung, C.-K. Jow, Y. Hidai, T. Kan, Tetrahedron
Lett. 1997, 38, 5831–5834.
[27]T. Kan, T. Fukuyama, Chem. Commun. 2004, 353–359.
[28]H. Kojima, Y. Urano, K. Kikuchi, T. Higuchi, Y. Hirata, T. Nagano,
Angew. Chem. 1999, 111, 3419–3422; Angew. Chem. Int. Ed. 1999,
38, 3209–3212.
Received: April 12, 2006
Revised: October 17, 2006
Published online: November 30, 2006
1954
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1946 – 1954