Mechanistic studies of fluorescent sensors for the detection of reactive oxygen species

Article abstract of DOI:10.1039/b713575k

Two new sensors for the detection of reactive oxygen species have been synthesized and characterized; they contain the 4-amino-7-nitrobenzofurazan (ABF) moiety covalently tethered to a phenol. Comparison of sensors ABFhd and dABFhd (the latter containing two ortho-methyl groups) shows that introduction of steric bulk leads to an improvement of fluorescent sensor performance, thus confirming our previous predictions based on computational chemistry. ABFhd and dABFhd are new tools for biological applications involving reactive oxygen species, in particular oxygen-centered radicals. This journal is The Royal Society of Chemistry.

Full text of DOI:10.1039/b713575k

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Mechanistic studies of fluorescent sensors for the detection of reactive oxygen
species†
Belinda Heyne, Sara Ahmed and J. C. Scaiano*
Received 4th September 2007, Accepted 15th November 2007
First published as an Advance Article on the web 3rd December 2007
DOI: 10.1039/b713575k
Two new sensors for the detection of reactive oxygen species have been synthesized and characterized;
they contain the 4-amino-7-nitrobenzofurazan (ABF) moiety covalently tethered to a phenol.
Comparison of sensors ABFhd and dABFhd (the latter containing two ortho-methyl groups) shows that
introduction of steric bulk leads to an improvement of fluorescent sensor performance, thus confirming
our previous predictions based on computational chemistry. ABFhd and dABFhd are new tools for
biological applications involving reactive oxygen species, in particular oxygen-centered radicals.
Introduction
During the past two decades, there has been a growing interest
into the investigation of reactive oxygen species (ROS), which
play a central, significant role in oxidative damage of biological
systems.1 ROS are derived from the metabolism of molecular
oxygen and have been implicated in the aging process and
in numerous diseases such as atherosclerosis, neurodegenerative
diseases and cancer.2–7 ROS are known to react with a large variety
Chart 1 Structure of the different sensors and the fluorescent molecules
of molecules including proteins, lipids, carbohydrates and nucleic
they release upon exposure to ROS.
acids. Targeted compounds, which are more reactive towards ROS,
include unsaturated lipids, specific amino acids (sulfydryl group)
and aromatic compounds.2 On the other hand, it is believed that
peroxyl radical at the para-position was involved in the release
of the fluorescent moiety NBF.8 A computational chemistry
approach proved that favorable energetics for the formation of
the ortho-coupling product served as a side reaction that does not
involve NBF release. On the basis of computational studies it was
suggested that introducing steric bulk, such as methyl groups at the
ortho-positions, could lead to more para-coupling, and as a result
may help to increase the amount of fluorescent moiety released.
Due to severe oxidation problems, the synthesis of a dimethyl
ortho-substituted analogue of NBFhd proved to be extremely
difficult. Therefore, in order to assess the prediction based on
computational chemistry, a modified version of NBFhd, ABFhd
1 and its dimethyl analogue, dABFhd 4 were synthesized. As
observed in Chart 1, the tertiary amine of the fluorescent sensor
was substituted by a hydrogen in ABFhd 1, instead of a methyl
group.
different ROS can have characteristic roles and reactivities in vivo.2
It is therefore essential to develop accurate and sensitive
methods for their detection. Fluorescent sensors have received
much attention because of their sensitivity and their non-invasive
nature. Good fluorescent sensors should be easy to synthesize, have
a dark initial state (i.e., non-fluorescent) and be fully compatible
with the target media. For biological sensors, such compatibility
must include absorbance in a convenient spectral region and
minimal overlap of their fluorescence with cell autofluorescence.
In a recent study, we proposed a new sensor for the detection of
ROS: NBFhd, which contains a fluorescent moiety, N-methyl-4-
amino-7-nitrobenzofurazan (NBF), tethered to a phenolic struc-
ture that acts as a quencher and also as a hydrogen donor
(Chart 1).8 While this new sensor proved to be efficient in the
detection of a variety of ROS, it is also limited in the amount
of fluorescent molecule that can be released (15% to 20% of
the starting material). This result was linked to the mechanism
of reaction of the sensor. It was proposed in the case of the
interaction of NBFhd with peroxyl radical that two radicals were
required. The first radical was initiating the formation of an
NBFhd phenoxyl radical intermediate, which in turn reacted with
a second peroxyl radical and water. Only the attack of this second
Experimental
All chemicals were purchased from Aldrich and used as received.
The 1H NMR spectra were recorded on a Bruker-Avance-400
spectrometer at 400 MHz. Samples were prepared in methyl
sulfoxide-d6 (99%) purchased from Aldrich. The chemical shifts
are quoted using the d scale and the coupling constants are
expressed in Hz. Mass spectra (see ESI†) were recorded on a
Kratos-Concept-II instrument.
University of Ottawa, Department of Chemistry, 10 Marie Curie, Ottawa,
ON, K1N 6N5, Canada. E-mail: tito@photo.chem.uottawa.ca
† Electronic supplementary information available: NMR spectra, Stern–
Volmer plot, HPLC chromatogram, estimation of number of radicals,
synthesis of quinoneimine. See DOI: 10.1039/b713575k
354 | Org. Biomol. Chem., 2008, 6, 354–358
This journal is
The Royal Society of Chemistry 2008
©

View Online
Synthesis of 4-(4-hydroxyphenylamino)-7-nitrobenzofurazan
(ABFhd), 1
9.73 (s, 1H), 10.7 (br s, 1H). The accurate mass was measured to
be 300.08479 m/z for a structure C14H12N4O4.
The synthesis of ABFhd was based on a procedure outlined in a
previous study.8 Briefly, 610 mg of 4-chloro-7-nitrobenzofurazan
(3.05 mmol) were dissolved in 5 ml methanol. Then, 500 mg
of 4-aminophenol (4.58 mmol) were added. The mixture was
refluxed at 75 ◦C under N2 atmosphere. After 1 h, 769 mg of
sodium bicarbonate (9.15 mmol) dissolved in 5 ml of Millipore
water were added dropwise to the stirring solution. Upon cooling,
dark brown–reddish crystals of the desired compound precipitated
(760 mg, 2.79 mmol, 91.5% yield). 1H NMR (DMSO-d6, 298 K):
d = 6.45 (d, J = 9.12 Hz, 1H), 6.85 (d, J = 8.8 Hz, 2H), 7.23 (d,
J = 8.8 Hz, 2H), 8.44 (d, J = 9.12 Hz, 1H), 9.73 (s, 1H), 10.7 (br s,
1H). The accurate mass was measured to be 272.0597 m/z for a
structure C12H8N4O4.
Synthesis of 4-amino-7-nitrobenzofurazan (ABF), 5
The fluorescent compound 5 was synthesized according to the
literature.11 The crude product was purified by column chromatog-
raphy with 3 : 2 ethyl acetate : hexane elution, to afford pure 5
(98 mg, 0.544 mmol, 46.9% yield). 1H NMR (DMSO-d6, 298 K):
d = 6.35 (d, J = 9.0 Hz, 1H), 8.47 (d, J = 9.0 Hz, 1H), 8.85 (s,
2H). EI found: 180 m/z.
Spectroscopic measurements
All the steady state emission spectra were recorded using a Pho-
ton Technology International spectrofluorimeter. The excitation
wavelength for both ROS sensors was set up at 460 nm.
Lifetime fluorescence measurements were performed on a
Easylife LS (Photon Technology International). The excitation
was performed with a 440 nm pulsed LED and the broadband
emission was measured at 540 nm.
All absorption spectra were recorded on a Varian CARY-50
UV–vis spectrophotometer. All samples were placed in a spectro
quality quartz cuvette, 1 cm path length.
Synthesis of 2,6-dimethyl-4-nitrosophenol, 2
The synthesis of 2 was based on a procedure outlined by Um et al.9
Briefly, 10 g of 2,6-dimethylphenol (81.9 mmol) were dissolved
in a solution of 85 ml of 99% ethanol containing 2.54 ml of
concentrated sulfuric acid, while kept below 0 ◦C. 6.42 g of
sodium nitrite (93 mmol) dissolved in 15 ml of Millipore water
were added dropwise to the stirring solution. A dark yellow
precipitate formed immediately. The precipitate, 2,6-dimethyl-4-
nitrosophenol 2, was collected by filtration (9.0 g, 59.6 mmol,
72.8% yield). 1H NMR(DMSO-d6, 298 K); d = 1.89 (d, J =
9.96 Hz, 6H), 7.09 (s, 1H), 7.51 (s, 1H). EI found: 151 m/z.
Results and discussion
In phosphate buffer, 4-amino-7-nitrobenzofurazan (ABF 5,
Chart 1) and N-methyl-4-amino-7-nitrobenzofurazan (NBF,
chart 1),12 present similar photophysical properties with a relative
quantum yield of 0.03 (compared to Rhodamine G6 at 22 ◦C). The
fluorescent molecule ABF 5 fulfils our requirement for developing
a fluorescent sensor. Indeed, with a fluorescence maximum around
540 nm, ABF will not interfere with cell autofluorescence.13 The
synthesis of ABFhd 1 and its dimethyl ortho-substituted form
(dABFhd 4) is illustrated in Scheme 1. 1H NMR spectra of
ABFhd and dABFhd are consistent with the proposed structure
and confirm the purity of the samples.
Synthesis of 2,6-dimethyl-4-aminophenol hydrochloride, 3
Following a literature procedure,10 33 mg of 10% palladium
on activated carbon were suspended in 4 ml of water. To the
suspension was added 252 mg of NaBH4 (6.6 mmol) dissolved in
5 ml of Millipore water. Then, a solution of 500 mg of 2 (3.3 mmol)
dissolved in 20 ml of 2 M NaOH was added dropwise under N2
to the stirring mixture. The mixture was refluxed under N2 for
2 h. 3 ml of pure hydrochloric acid were added in order to destroy
the excess NaBH4 and to form the hydrochloride salt since 2,6-
dimethyl-4-aminophenol is easily oxidized. The suspension was
filtered and the filtrate was evaporated under vacuum to obtain
a tan powder. The crude was used as obtained for the next step,
synthesis of compound 4. 1H NMR (DMSO-d6, 298 K): d = 2.16
(s, 6H), 6.87 (s, 2H), 8.56 (s, 1H), 10.12 (s, 3H).
Synthesis of 4-(2,6-dimethyl-4-hydroxyphenylamino)-
7-nitrobenzofurazan (dAFBhd), 4
The synthesis of dABFhd 4 has not been reported before and was
based on a procedure outlined in a previous study.8 Briefly, 372 mg
of 4-chloro-7-nitrobenzofurazan (1.87 mmol) were dissolved in
20 ml methanol. Then, 500 mg of 3 (3.62 mmol) were added.
The mixture was refluxed at 75 ◦C under N2 atmosphere. After
2 h, 471 mg of sodium bicarbonate (5.6 mmol) dissolved in 5 ml
of Millipore water were added dropwise to the stirring solution.
Upon cooling, dark brown crystals of the desired compound
precipitated (284.5 mg, 0.94 mmol, 50.3% yield). 1H NMR
(DMSO-d6, 298 K): d = 6.45 (d, J = 9.12 Hz, 1H), 6.85 (d, J =
8.8 Hz, 2H), 7.23 (d, J = 8.8 Hz, 2H), 8.44 (d, J = 9.12 Hz, 1H),
Scheme 1 Synthesis of ABFhd (1; eqn 1) and dABFhd (4; eqn 2 and 3).
Org. Biomol. Chem., 2008, 6, 354–358 | 355
This journal is
The Royal Society of Chemistry 2008
©

View Online
The maximum absorption of ABFhd and dABFhd is occurring
at 490 nm while the maximum of ABF is around 460 nm (Fig. 1). In
addition, neither ABFhd nor dABFhd exhibit fluorescence within
our detection limits. Both probes offer excellent contrast for the
detection of ROS. As already shown for other fluorescent sensors,8
the absence of fluorescence for ABFhd and dABFhd can be
rationalized on the basis of efficient electron transfer between the
phenolic moiety and the excited chromophore. The rate constant
for the fluorescence quenching of ABF by either phenol or 2,6-
dimethylphenol was obtained from Stern–Volmer experiments (see
ESI†). The fluorescence lifetime of ABF was measured to be 5.5 ns,
which gives a measured rate constant of 4 × 109 M−1 s−1 for
quenching of ABF fluorescence by phenol and 5.6 × 109 M−1 s−1
for quenching of ABF fluorescence by 2,6-dimethylphenol. These
values are close to those found in a previous study.8 One should
note that the methyl substitution of phenol at the ortho-position
does not lead to a difference in the fluorescence quenching rate
constant of excited ABF.
than 1% of dimethylformamide; both fluorescence sensors obey
the Beer–Lambert law, indicating the absence of self-association of
both sensors in the ground state. The molar extinction coefficients
were estimated at 490 nm to be 16 420 M−1 cm−1 for ABFhd and
17 300 M−1 cm−1 for dABFhd.
The ability of the two new fluorescence sensors to react with
peroxyl radical was analyzed in PBS by means of absorption
and fluorescence spectroscopy. In both cases, peroxyl radical was
derived from the thermal decomposition of the water soluble
2,2ꢀ-azobis(2-amidinopropane) dihydrochloride (AAPH). AAPH
is well known to decompose at 40 ◦C yielding two carbon-centered
radicals, which in turn react with molecular oxygen present in
the solution to give two peroxyl radicals.14,15 As observed in
Fig. 2, the absorption peak corresponding to the probes decreases
upon exposure to peroxyl radical while the fluorescence spectra
show the opposite behavior. In both cases, the increase of the
fluorescence intensity over the reaction time was attributed to
the release of the fluorescence compound ABF. The nature of
the fluorescence compound release was attributed by HPLC with
fluorescence detection (see ESI†). The decrease in absorption at
ca. 490 nm probably incorporates some attack of peroxyl radical
at the chromophore.
Upon comparison of the change in fluorescence over time
(Fig. 3b), both ABFhd and dABFhd experience the same small
change during the first 30 minutes. After this point, dABFhd be-
gins to evolve at an increasingly greater rate, and after 150 minutes
(Fig. 3b), the fluorescence observed is almost double in the case of
dABFhd compared to ABFhd. On the other hand, Fig. 3a shows
a contrasting behavior for the evolution of the absorption peak
of both sensors. Indeed, during the first 30 minutes the decrease
in absorbance in the case of dABFhd is much more prominent,
while after that period the decrease in absorbance occurs at similar
rate for both sensors (Fig. 3a). The results obtained in Fig. 3b
corroborate the prediction made by computational methods,8 that
dABFhd should release a greater amount of ABF than ABFhd.
Fig. 1 Absorption (left) and fluorescence (right) spectra of ABFhd (5 ×
10−6 M, thick line), dABFhd (5 × 10−6 M, thin line) and ABF (5 ×
10−6 M, dashed line) in phosphate buffer pH 7.4. The excitation was set
at 460 nm. Note: fluorescence spectra for ABFhd and dABFhd cannot be
distinguished from the wavelength axis.
For concentrations ranging from 5 × 10−7 M to 2 × 10−5 M in
phosphate buffer (PBS, 15 mM NaH2PH4–K2HPO4) with no more
Fig. 2 Absorption and fluorescence emission spectra of a phosphate buffer solution pH 7.4 containing 10−2M AAPH and 5 × 10−6 M ABFhd (a,c) or
5 × 10−6 M dABFhd (b,d), at 40 ◦C under air.
356 | Org. Biomol. Chem., 2008, 6, 354–358
This journal is
The Royal Society of Chemistry 2008
©

View Online
Scheme 2 Mechanism proposed for the interaction of the new fluorescent
sensors ABFhd and dABFhd with peroxyl radical. The mechanism is
similar for both sensors and only the case of dABFhd is illustrated here.
ABFhd and dABFhd can act also as a potential site of reactivity
towards peroxyl radicals. However, it seems unlikely that peroxyl
radicals will react initially with N–H rather than O–H. Indeed, it
has been shown that for 4-hydroxydiphenylamine,18 a compound
related to our sensors, that the bond dissociation energy of the
N–H group is higher (354.4 kJ mol−1) than that of the OH moiety
(339.9 kJ mol−1). Scheme 2 suggests that the phenoxyl radical
derived from the sensor can react with a second peroxyl radical
(a pathway not available to NBFhd) to form the quinoneimine
intermediate, which will then further slowly hydrolyze to give rise
to the fluorescent moiety ABF and a quinone.
Further experiments allow us to confirm these hypotheses.
Indeed, Fig. 4 shows that the bleaching of the sensor is almost
totally prevented by Trolox, a water soluble analogue of vitamin
E. This supports the occurrence of a process mediated by free
peroxyl radicals. In addition, the efficiency of ABFhd and dABFhd
consumption was evaluated by comparing the consumption rate
of the two sensors with the rate of the free radical production19
(see ESI†). From our experiments, it can be concluded that, on
average, 2 peroxyl radicals are scavenged by each sensor molecule
consumed. Further, the fact that an equimolar concentration of
Trolox almost totally suppresses the release of ABF suggests that
the rate constant for radical attack on Trolox is significantly faster
than that for sensors ABFhd and dABFhd.
Fig. 3 Comparison of the maximum of (a) absorbance at 490 nm and
(b) fluorescence emission at 540 nm of a phosphate buffer solution pH 7.4
containing 10−2M AAPH and 5 × 10−6 M ABFhd (circle) or 5 × 10−6
M
dABFhd (square), at 40 ◦C under air.
However, upon closer investigation it appears there are two phases
inherent in the interaction of peroxyl radical with both sensors
before and after ∼30 min. In the first phase there is an indication
of a much faster consumption of the starting material in the case of
dABFhd than ABFhd. This result is not surprising if one considers
that the first step in the reaction mechanism involves the oxidation
of the phenolic moiety. It is known that the rate constant for the
reaction of 2,6-dimethylphenol with peroxyl radical is higher than
that for phenol.16 In the second phase, the consumption of the two
sensors is similar with respect to the observed slow decrease in
absorbance. On the other hand, the fluorescence intensity increases
more rapidly in the case of the ortho-substituted sensor. Overall,
these observations suggest that there is an intermediate step, which
slows the release of ABF inherent in the mechanism of action for
both sensors.
The characterization of the products resulting from the in-
teraction of both sensors with peroxyl radical was investigated
by gas chromatography. The only products observed were 1,4-
benzoquinone in the case of the ABFhd sensor and 2,6-dimethyl-
1,4-benzoquinone in the case of dABFhd sensor (data not shown).
Even though the interaction of peroxyl radical with the two new
sensors ABFhd and dABFhd results in the release of a quinone
and the fluorescence moiety ABF, our results suggest a mechanism
of reaction different from that proposed for previously reported
sensors of the same family.8,17 Indeed, as emphasized above, an
intermediate step slowing the release of the fluorescent moiety
is observed. Since the tertiary amine of ABFhd and dABFhd
is substituted by a hydrogen instead of a methyl, a plausible
mechanism could implicate the formation of a quinoneimine as
an intermediate product in the reaction of both sensors with
peroxyl radical. As illustrated in Scheme 2, the first step of the
reaction is probably a hydrogen atom transfer likely mediated by
an electron transfer,8,17 from the phenolic moiety to the peroxyl
radical, leading to the formation of the corresponding phenoxyl
radical. One might argue that the presence of the N–H bond in
Fig. 4 Effect of the addition of 5 × 10−5 M Trolox on a phosphate buffer
solution pH 7.4 containing 10−2 M AAPH and 5 × 10−5 M dABFhd under
air at 40 ◦C. The circles represent the maximum absorption of the sensor
in the absence of Trolox and the squares in the presence of Trolox.
The nature of a quinoneimine as an intermediate product was
indicated by the results shown in Fig. 5. Indeed, when Trolox
is added to the solution after 30 minutes of reaction between
peroxyl radical and the sensors, the release of ABF by the sensor is
quenched as illustrated in Fig. 5b, while the absorbance is restored
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 354–358 | 357
©

View Online
shows that by introducing steric bulk such as ortho-methyl groups,
the fluorescent sensor performance is improved since dABFhd
is able to release more of the reporting fluorescent compound,
ABF, than ABFhd, thus confirming our previous hypothesis based
on computational chemistry. Thus, ABFhd and dABFhd are
new tools to explore the involvement of reactive oxygen species,
in particular oxygen-centered radicals, including superoxide, for
biological applications.
Acknowledgements
This work was generously supported by funding from the Natural
Sciences and Engineering Research Council of Canada, by the
Canadian Foundation for Innovation and by the Government of
Ontario. We would like also to thank Dr Keith U. Ingold for many
helpful suggestions.
References
1 T. Kalai, E. Hideg, I. Vass and K. Hideg, Free Radical Biol. Med., 1998,
24, 649.
2 R. Colavitti and T. Finkel, IUBMB Life, 2006, 57, 277.
3 B. Halliwell and J. M. C. Gutteridge, Free Radicals in Biology and
Medicine, Oxford University Press, 1989.
4 D. Harman, Age, 1983, 6, 86.
5 C. Hazelwood and M. J. Davies, J. Chem. Soc., Perkin Trans. 2, 1995,
895.
Fig. 5 Effect of the addition of 10−5 M Trolox after 35 min of reaction
between 5 × 10−6 M ABFhd and peroxyl radical produced by the thermal
decomposition of 10−2 M AAPH in the presence of oxygen (air) at 40 ◦C:
(a) absorption spectrum (t = 0 min in a; t = 6 min in b; t = 20 min in c,
t = 35 min in d; t = 35 min + Trolox in e; t = 50 min in f; t = 70 min in g),
(b) fluorescence intensity at 540 nm.
6 C. P. Schulze and R. T. Lee, Curr. Atheroscl. Rep., 2005, 7, 242.
7 G. Waris and H. Ahsan, J. Carcinog., 2006, 5, 14.
8 B. Heyne, C. Beddie and J. C. Scaiano, Org. Biomol. Chem., 2007, 5,
1454.
9 S. I. Um, J. K. Lee, Y. Kang and D. J. Baek, Dyes Pigm., 2005, 65, 93.
10 A. Vogel, Vogel’s Practical Organic Chemistry, 4th edn, Longman
Scientific & Technical, Essex, UK, 1978.
11 M. Boiocchi, L. Del Boca, D. Esteban-Gomez, L. Fabbrizzi, M.
Licchelli and E. Monzani, Chem.–Eur. J., 2005, 11, 3097.
12 S. Uchiyama, T. Santa and K. Imai, J. Chem. Soc., Perkin Trans. 2,
1999, 2525.
(Fig. 5a). The quinoneimine intermediate formed during the first
30 minutes is reduced by the Trolox added to the solution giving
back the starting material, which can account for the restoration
of the absorbance while the fluorescence increase is blocked. To
analyze the ability of the quinoneimine to slowly hydrolyze in
buffer solution, we synthesized the molecule in organic solvent
by reaction with lead dioxide (see ESI†). When dissolved in
phosphate buffer solution pH 7.4, the quinoneimine hydrolyzed
slowly to release the fluorescent compound ABF (data not shown).
These results confirmed the mechanism of reaction proposed in
Scheme 2.
13 J. E. Aubin, J. Histochem. Cytochem., 1979, 27, 36.
14 M. A. Cubillos, E. A. Lissi and E. B. Abuin, Chem. Phys. Lipids, 2000,
104, 49.
15 G. S. Hammond and R. C. Neuman, Jr., J. Am. Chem. Soc., 1963, 85,
1501.
16 J. A. Howard and K. U. Ingold, Can. J. Chem., 1963, 41, 2800.
17 B. Heyne, V. Maurel and J. C. Scaiano, Org. Biomol. Chem., 2006, 4,
802.
18 V. T. Varlamov, Russ. Chem. Bull., Int. Ed., 2004, 53, 306.
19 E. Pino, A. M. Campos and E. Lissi, Int. J. Chem. Kinet., 2003, 35, 525.
Conclusion
In summary, we have developed two new sensors for the detection
of ROS: ABFhd and dABFhd. Comparison of the two structures
358 | Org. Biomol. Chem., 2008, 6, 354–358
This journal is
The Royal Society of Chemistry 2008
©