Stable dioxetane precursors as selective trap-and-trigger chemiluminescent probes for singlet oxygen

Article abstract of DOI:10.1021/ac048293s

A set of highly selective chemiluminescent probes has been developed for the detection and quantitation of singlet oxygen (¹O₂), a reactive oxygen species that is known to transform organic pollutants in the aquatic environment and e

Full text of DOI:10.1021/ac048293s

Anal. Chem. 2005, 77, 1200-1205
Stable Dioxetane Precursors as Selective
Trap-and-Trigger Chemiluminescent Probes for
Singlet Oxygen
Laura A. MacManus-Spencer, Douglas E. Latch, Kim M. Kroncke, and Kristopher McNeill*
Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455
•-
chemistry, from the reaction of O2 and H2O2,9 a controversial
A set of highly selective chemiluminescent probes has
been developed for the detection and quantitation of
singlet oxygen (1O2), a reactive oxygen species that is
known to transform organic pollutants in the aquatic
environment and elicits cytotoxic effects in biological
systems. In this study, a trap-and-trigger detection method
is employed, based on the reaction of 1O2 with a spiroada-
mantyl-substituted vinyl ether probe to form the corre-
sponding thermally stable dioxetane, which undergoes
chemiluminescent decomposition upon addition of a
chemical trigger. The detection method is highly selective
for 1O2 relative to superoxide anion and hydrogen perox-
ide. The sensitivity of this method allows for the accurate
measurement of environmentally relevant (picomolar)
steady-state 1O2 concentrations in relatively short expo-
sure times. The detection method was used to detect and
quantify 1O2 production in the reaction of dibenzoyl
peroxide with superoxide anion.
observation that has been met with skepticism.10,11 Evidence for
the formation of O2 has also been obtained in the reactions of
diacyl peroxides with O2•-.6 Establishing the presence of O2 in
systems involving other ROS, such as O2•- and peroxides, requires
a detection method that is both sensitive to and selective for 1O2.
Interest in the detection of 1O2 has led to the development and
application of different types of molecular probes, including those
based on absorbance,12-14 fluorescence,15,16 and chemilumi-
nescence.17-25 Absorbance-based probes, such as 2,5-dimethyl-
furan (DMF)12 and furfuryl alcohol (FFA),13 rely on monitoring
the loss of the probe molecule through reaction with O2. Any
processes that lead to the destruction of the probe molecule, such
as self-photooxidation, will result in an overestimation of the O2
concentration. More recently, absorbance-based probes for O2
based on 3-substituted pyrroles, such as the commercially available
tert-butyl-3,4,5-trimethylpyrrolecarboxylate (BTMPC), have been
applied; such probes address the issue of overestimated 1O2
concentrations, as self-photooxidation is significantly less than for
the furan-based probes.14 Still, absorbance-based detection is
inherently less sensitive than luminescence detection.
1
1
1
1
1
Singlet oxygen, 1∆g, (1O2) is a reactive oxygen species (ROS)
that plays key roles in both environmental and biological systems.
In the environment, 1O2 formed in sunlit natural waters is
responsible for oxidative transformations of certain classes of
organic pollutants.1,2 In living cells, 1O2 is proposed to be involved
in changes in the mitochondrial membrane pore transition3 and
has recently been proposed as being key to the bactericidal
response of certain antibodies.4,5 Singlet oxygen’s formation and
chemistry are intimately related to another transient species,
superoxide anion (O2•-),6-8 complicating its detection. Singlet
oxygen has also been suggested as a product of Haber-Weiss
(9) Khan, A. U.; Kasha, M. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 12365-
12367.
(10) Koppenol, W. H. Redox Rep. 2001, 6, 229-234.
(11) Koppenol, W. H. Redox Rep. 2002, 7, 59-60.
(12) Zepp, R. G.; Wolfe, N. L.; Baughman, G. L.; Hollis, R. C. Nature 1977,
267, 421-423.
(13) Haag, W. R.; Hoigne, J. Environ. Sci. Technol. 1986, 20, 341-348.
(14) Denham, K.; Milofsky, R. E. Anal. Chem. 1998, 70, 4081-4085.
(15) Tanaka, K.; Miura, T.; Umezawa, N.; Urano, Y.; Kikuchi, K.; Higuchi, T.;
Nagano, T. J. Am. Chem. Soc. 2001, 123, 2530-2536.
(16) Singlet oxygen sensor green reagent. http://www.probes.com/servlets/
product?item)36002 (December 13, 2004), Molecular Probes online product
information.
* Corresponding author. Phone: (612) 625-0781. Fax: (612) 626-7541.
E-mail: mcneill@chem.umn.edu.
(1) Larson, R. A.; Marley, K. A. Handb. Environ. Chem. 1999, 2, 123-137.
(2) Blough, N. V.; Zepp, R. G. In Active Oxygen: Reactive Oxygen Species in
Chemistry; Foote, C. S., Valentine, J. S., Greenberg, A., Liebman, S. F., Eds.;
Blackie Academic and Professional: London, 1995.
(17) Goto, T.; Takagi, T. Bull. Chem. Soc. Jpn. 1980, 53, 833-834.
(18) Sugioka, K.; Nakano, M.; Kurashige, S.; Akuzawa, Y.; Goto, T. FEBS Lett.
1986, 197, 27-30.
(19) Schaap, A. P.; Handley, R. S.; Giri, B. P. Tetrahedron Lett. 1987, 28, 935-
938.
(20) Schaap, A. P.; Chen, T. S.; Handley, R. S.; DeSilva, R.; Giri, P. P. Tetrahedron
(3) Beghetto, C.; Renken, C.; Eriksson, O.; Jori, G.; Bernardi, P.; Ricchelli, F.
Eur. J. Biochem. 2000, 267, 5585-5592.
(4) Wentworth, P., Jr.; Jones, L. H.; Wentworth, A. D.; Zhu, X.; Larsen, N. A.;
Wilson, L. A.; Xu, X.; Goddard, W. A.; Janda, K. D.; Eschenmoser, A.; Lerner,
R. A. Science 2001, 293, 1806-1811.
(5) Wentworth, P., Jr.; McDunn, J. E.; Wentworth, A. D.; Takeuchi, C.; Nieva,
J.; Jones, T.; Bautista, C.; Ruedi, J. M.; Gutierrez, A.; Janda, K. D.; Babior,
B. M.; Eschenmoser, A.; Lerner, R. A. Science 2002, 298, 2195-2199.
(6) Danen, W. C.; Arudi, R. L. J. Am. Chem. Soc. 1978, 100, 3944-3945.
(7) Saito, I.; Matsuura, T.; Inoue, K. J. Am. Chem. Soc. 1983, 105, 3200-3206.
(8) Kanofsky, J. R. J. Am. Chem. Soc. 1986, 108, 2977-2979.
Lett. 1987, 28, 1155-1158.
(21) Schaap, A. P.; Sandison, M. D.; Handley, R. S. Tetrahedron Lett. 1987, 28,
1159-1162.
(22) Adam, W.; Fell, R.; Schulz, M. H. Tetrahedron 1993, 49, 2227-2238.
(23) Adam, W.; Bronstein, I.; Edwards, B.; Engel, T.; Reinhardt, D.; Schneider,
F. W.; Trofimov, A. V.; Vasil’ev, R. F. J. Am. Chem. Soc. 1996, 118, 10400-
10407.
(24) Matsumoto, M.; Watanabe, N.; Shiono, T.; Suganuma, H.; Matsubara, J.
Tetrahedron Lett. 1997, 38, 5825-5828.
(25) Watanabe, N.; Suganuma, H.; Kobayashi, H.; Mutoh, H.; Katao, Y.;
Matsumoto, M. Tetrahedron 1999, 55, 4287-4298.
1200 Analytical Chemistry, Vol. 77, No. 4, February 15, 2005
10.1021/ac048293s CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/15/2005

More sensitive fluorescent probes for 1O2 have also been
developed, such as the fluorescein derivative 9-[2-(3-carboxy-
9,10-dimethyl)anthryl]-6-hydroxy-3H-xanthen-3-one (DMAX), which
is characterized by only weak fluorescence in its native state but
which becomes highly fluorescent upon reaction with 1O2.15
Molecular Probes also offers a Singlet Oxygen Sensor Green,
which is a sensitive fluorescent probe for 1O2 that does not yield
a significant signal from hydroxyl radical (•OH) or O2•-. However,
this probe is intended for use primarily in aqueous solutions of
neutral or acidic pH, with basic conditions and several organic
solvents (acetonitrile, dimethyl sulfoxide, N,N-dimethylformamide,
acetone) leading to a fluorescence signal that is unrelated to 1O2.16
Among the most sensitive 1O2 probes are those in which
detection is based on chemiluminescence (CL), as such probes
benefit from the low background in CL measurements, where no
light source is needed to stimulate luminescence. CL measure-
ments also require less sophisticated equipment. A set of widely
Scheme 1. Chemiluminescent Detection
Schemea
a 1O2 is trapped with probe 1 to form a thermally stable
dioxetane (2), which emits visible light (470 nm) upon addition of
a chemical trigger (F-).
1
studied CL probes for O2 are the Cypridina luciferin analogues
(CLAs),17,18 an example of which is methyl Cypridina luciferin
analogue (MCLA). For the detection of 1O2, however, MCLA has
1
1
Due to our interest in photochemically generated O2 and in
several disadvantages. First, CLAs are not selective for O2; in
fact, they are commonly used for the detection of O2 .
•- 26 Second,
kinetic studies of its formation, we sought a detection method
that allows for a time delay between 1O2 trapping and CL detection.
Such a delay allows for low-background CL measurements without
interference from the irradiation source, as well as a simple
method to measure the time-dependent formation of 1O2, and
would represent a significant improvement over other CL 1O2
probes, including CLAs17,18 and the thiafulvalene-substituted an-
thracene probe,28,29 which luminesce spontaneously upon reaction
MCLA is susceptible to autooxidation in the absence of ROS,27
leading to a high background signal. Finally, probes such as CLAs
emit light spontaneously upon reaction with 1O2, which limits their
utility in detecting photochemically generated 1O2. A selective
chemiluminescent probe for 1O2 based on a thiafulvalene-
substituted anthracene trap has also been reported recently.28,29
The chemistry of 1,2-dioxetanes offers a promising scheme
for the selective and sensitive chemiluminescent detection of 1O2.
The most general preparation of 1,2-dioxetanes is through the
formal [2+2] cycloaddition of 1O2 to electron-rich alkenes.30,31
Other routes to dioxetanes have been described,32-35 and these
reactions may cause interferences. Nevertheless, dioxetane forma-
tion can be taken as strong evidence for the presence of 1O2.
Detection of dioxetanes is straightforward because their decom-
position is accompanied by luminescence. Such strategies for the
selective detection of 1O2 using vinyl ether probes have been
employed by Posner (trans-1-(2-methoxyvinyl)pyrene)36 and Nied-
erlander (1,2-diethoxyethylene).37,38
1
with O2. Schaap,19,20 Adam,22,23 and others24,25 have reported a
series of spiroadamantylidene-substituted dioxetanes that are
unusually stable and require a chemical trigger to initiate their
chemiluminescent decomposition. These dioxetanes exhibit an
enhanced luminescence efficiency afforded by the incorporation
of an aryloxy moiety, which acts as an electron donor and an
emitter after triggered dioxetane decomposition,39 and they have
been used in extremely sensitive bioassays in which a specific
enzymatic reaction is used as the trigger.19,23,40-42 For our trap-
and-trigger 1O2 detection method, we chose to use a precursor to
such a stable dioxetane, a spiroadamantylidene- and aryloxy-
substituted vinyl ether (1, Scheme 1). We have synthesized and
tested the methyl-substituted vinyl ether previously prepared by
Sabelle et al.,43 2-[1-(3-tert-butyldimethylsiloxy)phenyl)-1-methoxy-
methylene]tricyclo[3.3.1.1]decane (1a), as well as the new tetra-
glycol-substituted derivative, 2-[1-(3-tert-butyldimethylsilyloxy)-
phenyl)-3,6,9,12-tetraoxa-1-tridecyl-13-hydroxy-methylene]tricyclo-
[3.3.1.1]decane (1b). The dioxetane of 1a (2a, Scheme 1),
previously prepared and used by Schaap et al.,20 has also been
synthesized and found to be stable in organic solvent and neutral
water at room temperature for up to 1 day. At 0 °C, the stability
of 2a in solution is increased to months; acidic and basic aqueous
conditions accelerate the decomposition of 2a.
(26) Osman, A. M.; Laane, C.; Hilhorst, R. Luminescence 2001, 16, 45-50.
(27) Shimomura, O.; Wu, C.; Murai, A.; Nakamura, H. Anal. Biochem. 1998,
258, 230-235.
(28) Li, X.; Zhang, G.; Ma, H.; Zhang, D.; Li, J.; Zhu, D. J. Am. Chem. Soc. 2004,
126, 11543-11548.
(29) Zhang, G.; Li, X.; Ma, H.; Zhang, D.; Li, J.; Zhu, D. Chem. Commun. 2004,
2072-2073.
(30) Burns, P. A.; Foote, C. S. J. Am. Chem. Soc. 1974, 96, 4339-4340.
(31) Frimer, A. A. Singlet Oxygen, Volume 2: Reaction Modes and Products, Part
I; CRC Press: Boca Raton, 1985.
(32) Schaap, A. P.; Zaklika, K. A.; Kaskar, B.; Fung, L. W. M. J. Am. Chem. Soc.
1980, 102, 389-391.
(33) Nanni, E. J., Jr.; Angelis, C. T.; Dickson, J.; Sawyer, D. T. J. Am. Chem.
Soc. 1981, 103, 4268-4270.
(34) Liochev, S. I.; Fridovich, I. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 2891-
2896.
(35) Kotani, H.; Ohkubo, K.; Fukuzumi, S. J. Am. Chem. Soc. 2004, 126, 15999-
16006.
(36) Thompson, A.; Seliger, H. H.; Posner, G. H. Methods Enzymol. 1986, 133,
569-584.
(37) Niederlander, H. A. G.; Van Assema, W.; Engelaer, F. W.; Gooijer, C.;
Velthorst, N. H. Anal. Chim. Acta 1991, 255, 395-401.
(38) Niederlaender, H. A. G.; Nuijens, M. J.; Dozy, E. M.; Gooijer, C.; Velthorst,
N. H. Anal. Chim. Acta 1994, 297, 349-368.
(39) Matsumoto, M. J. Photochem. Photobiol. C 2004, 5, 27-53.
(40) Sabelle, S.; Renard, P.-Y.; Pecorella, K.; de Suzzoni-Dezard, S.; Creminon,
C.; Grassi, J.; Mioskowski, C. J. Am. Chem. Soc. 2002, 124, 4874-4880.
(41) Edwards, B.; Sparks, A.; Voyta, J. C.; Strong, R.; Murphy, O.; Bronstein, I.
J. Org. Chem. 1990, 55, 6225-6229.
(42) Beck, S.; Koster, H. Anal. Chem. 1990, 62, 2258-2270.
(43) Sabelle, S.; Hydrio, J.; Leclerc, E.; Mioskowski, C.; Renard, P.-Y. Tetrahedron
Lett. 2002, 43, 3645-3648.
Analytical Chemistry, Vol. 77, No. 4, February 15, 2005 1201

Scheme 2. Decomposition of MNPO2, a
Thermal Singlet Oxygen Generation. The method was
adapted from Pierlot et al.44 A solution of MCLA (5 µM) or 1a
(10 µM) was prepared in pH 7.8 50 mM phosphate buffer. An
aliquot of a stock solution of MNPO2 in pH 7.8 50 mM phosphate
buffer was added to the probe solution to obtain a concentration
of 2.8 mM MNPO2 and a total reaction solution volume of 500
µL. The stock solution of MNPO2 was kept frozen when not in
use. Immediately after the addition of MNPO2, the solution was
incubated at 34 °C. For MCLA, the reaction was carried out in
the luminometer sample chamber, which maintained a tempera-
ture of 34 ((1) °C throughout the measurements, and the
resulting CL signal was measured for 1000 s. For 1a, the
incubation with MNPO2 was carried out in a water bath at 34 °C.
Time points were obtained by sampling aliquots of the reaction
mixture and quenching the reaction by 1:1000 dilution in
CH3CN. The CL signal was then measured for each time point as
described above. The integrated signal intensity was converted
to concentration of 2a by comparison to a calibration curve
generated from authentic samples of 2a (independently synthe-
sized, see Supporting Information). The average integrated CL
signal for a 40-min exposure of 1a to MNPO2 was measured in
the same manner. A competition experiment was also carried out
in which both 1a (10 µM) and FFA (4 mM) were exposed to
MNPO2 (2.8 mM initial concentration) under the conditions
described above. For 1a, time points were taken and analyzed as
described above. For FFA, time points were obtained by sampling
aliquots of the reaction mixture, and the reaction was immediately
quenched by cooling to 5 °C. This was followed by HPLC analysis
using a Supelco Discovery RP Amide C16, 150 × 4.6 mm, 5-µm
particle size column with a mobile phase of 10% CH3CN and 90%
pH 5 ammonium acetate buffer, a 1.0 mL/min. flow rate, a 10-µL
injection volume, and absorbance detection at 219 nm. The rate
constant for reaction of 1a with 1O2, krxn, in water was calculated
by comparison to FFA (krxn ) 8.3 × 107 M-1 s-1, in water)45 from
the observed time-dependent degradation of FFA and production
of 2a.
Naphthalene Endoperoxide, at 34 °C, with
1
Release of O2 (with Yield r) and the
Naphthalene MNP
In this work, we examine these precursors to stable dioxetanes
1
as O2 probes by characterizing the sensitivity and selectivity of
our trap-and-trigger chemiluminescence detection scheme using
well-defined 1O2-generating systems. We demonstrate the selectiv-
ity of this detection scheme for 1O2 with respect to O2•- and H2O2.
As an illustration of the selectivity, we report evidence for the
1
•-
formation of O2 in the reaction of O2 with dibenzoyl peroxide;
we demonstrate that the reaction follows second-order kinetics,
which is in agreement with the proposed reaction stoichiometry.6
EXPERIMENTAL SECTION
Materials. Perinaphthenone, furfuryl alcohol (distilled before
use), tetra-n-butylammonium fluoride (TBAF, 1 M in THF),
hydrogen peroxide (30%), sodium molybdate dihydrate (Na2MoO4‚
2H2O), MCLA, potassium superoxide (KO2), dibenzoyl peroxide
(purum, moistened (25%) with water), diethylenetriaminepen-
taacetic acid (DTPA), and 1,4-diazabicyclo[2.2.2]octane (DABCO)
were purchased from commercial suppliers. Toluene and aceto-
nitrile were used as received. Xanthine oxidase was purchased
as a suspension in (NH4)2SO4 solution and was dialyzed into pH
7.8 50 mM phosphate buffer. Probes 1a and 1b and dioxetane
2a (Scheme 1), as well as 3-(1,4-epidoxy-4-methyl-1,4-dihydro-1-
naphthyl)propanoate (MNPO2, Scheme 2), are not commercially
available and were synthesized for this study (see Supporting
Information for experimental details).
Chemiluminescence Measurements. Chemiluminescence
measurements were made using a Turner Designs 20/20 lumi-
nometer with 12 × 50 mm polypropylene sample tubes. For all
assays using MCLA, the luminometer sensitivity was set to 40.2%
(default setting), and the CL signal was recorded from the addition
of the most labile component in the reaction until the signal
decayed to the baseline. For all assays using 1a, CL samples were
prepared by diluting a 10-µL aliquot of the reaction solution to 10
mL in CH3CN. Chemiluminescence measurements were made in
99.9% CH3CN/0.1% water solutions, as this mixture was found to
optimally balance quenching of the luminescent signal by water
and signal attenuation due to dilution. The luminometer sensitivity
was increased to 60.1% for all 1a measurements to compensate
for the effects of the 1000-fold dilution in CH3CN. The CL signal
from 2a was triggered by adding 20 µL of 10 mM TBAF, via
syringe through a light-tight manual injection port, to 300 µL of
the diluted reaction solution in CH3CN, and the CL signal was
recorded for 800 s.
Photochemical Singlet Oxygen Generation. Samples of 1a,
1b, and FFA were exposed to 1O2 generated photochemically
using perinaphthenone (quantum yield for the generation of 1O2,
Φ, ) 0.95)46 as a sensitizer. The steady-state photochemistry
experiments were conducted using one Pyrex-filtered 175-W
medium-pressure Hg vapor lamp (Luma-Pro lamps, GE HR175A39/
CP bulb) above a turntable apparatus to irradiate the samples in
quartz tubes. The sample solutions contained perinaphthenone
(100 µM) and 1a (10 µM), 1b (10 µM), or FFA (100 µM) in
deionized water. At several time points, sample aliquots were taken
from the irradiated solutions. For 1a and 1b, a 10-µL portion of
each aliquot was then diluted to 10 mL in CH3CN for the CL
measurements. For FFA, time point aliquots were analyzed by
HPLC as described above. Rate constants for the reactions of 1a
1
and 1b with O2, krxn, in water were calculated by comparison to
FFA.
(44) Pierlot, C.; Hajjam, S.; Barthelemy, C.; Aubry, J. M. J. Photochem. Photobiol.,
B 1996, 36, 31-39.
(45) Latch, D. E.; Stender, B. L.; Packer, J. L.; Arnold, W. A.; McNeill, K. Environ.
Sci. Technol. 2003, 37, 3342-3350.
(46) Schmidt, R.; Tanielian, C.; Dunsbach, R.; Wolff, C. J. Photochem. Photobiol.,
A 1994, 79, 11-17.
1202 Analytical Chemistry, Vol. 77, No. 4, February 15, 2005

Cytochrome c Assay. The general method was adapted from
Messner and Imlay.47 Solutions (1.8 mL) were prepared in pH
7.8 50 mM phosphate buffer containing cytochrome c (50 µM),
xanthine (1 mM), xanthine oxidase (0.05 unit/mL), and 1a
(varying concentrations) or superoxide dismutase (SOD) (varying
concentrations). The xanthine oxidase was added last. The rate
•-
of reduction of cytochrome c by O2 was calculated from the
increase in absorbance of reduced cytochrome c at 550 nm as a
function of time, using a ∆ꢀ(red-ox) for cytochrome c of 21 000 M-1
•-
cm-1. The rate of O2 production was assumed to be equivalent
to the rate of reduction of cytochrome c.
Enzymatic Superoxide Anion Generation. The method was
adapted from Zhao et al.48 A solution (500 µL) was prepared in
pH 7.8 50 mM phosphate buffer containing MCLA xanthine (1
mM), xanthine oxidase (0.02 unit/mL) and either MCLA (5 µM)
or 1a (10 µM). The xanthine oxidase was added last and marked
time zero for the exposure time. The rate of O2•- production under
identical conditions, but in the absence of MCLA or 1a, was
determined using the cytochrome c assay as described above. For
MCLA, the reaction was carried out in the luminometer sample
chamber, and the resulting CL signal was measured for 1200 s.
For 1a, a 10-µL aliquot of the reaction solution was taken after
1200 s and diluted to 10 mL in CH3CN for the CL measurement.
Hydrogen Peroxide Exposure. A solution (1.0 mL) was
prepared in pH 7.8 50 mM phosphate buffer containing MCLA (5
µM) or 1a (10 µM) and H2O2 (50 mM). The H2O2 was added last
and marked time zero for the exposure time. For MCLA, the
reaction was carried out in the luminometer sample chamber, and
the resulting CL signal was measured for 2500 s. For 1a, a 10-µL
aliquot of the reaction solution was taken after 2500 s and diluted
to 10 mL in CH3CN for the CL measurement.
Figure 1. Production of dioxetane 2a during exposure of 1a (10
µM) to 1O2 generated by the thermal decomposition of MNPO2 (2.8
mM) at 34 °C in pH 7.8 50 mM phosphate buffer. The solid line
corresponds to a nonlinear fit of the first-order production of 2a. The
rate constant extracted from the fit is 2.8 ((0.4) × 10-4 s-1, and the
asymptote (dashed line), corresponding to the final yield of 2a, is 1.6
× 10-7 M.
DABCO solution (520 µM final concentration). In the trial without
DABCO, the 10-µL aliquot of stock DABCO solution was replaced
with 10 µL of toluene.
RESULTS AND DISCUSSION
Probes 1a and 1b were prepared using McMurry coupling
conditions (described in the Supporting Information) from ada-
mantanone and the silyl-protected benzoate esters.40 Our ability
to vary the vinyl ether substituent allows us to alter the probe’s
solubility for use in a wide range of solvent environments. The
trap-and-trigger 1O2 detection scheme using these vinyl ether
probes is illustrated in Scheme 1.
Reaction of Superoxide Anion with Dibenzoyl Peroxide.
All glassware was acid washed, rinsed thoroughly with Milli-Q
water, and dried in an oven (110 °C) prior to use. Saturated
solutions of KO2 containing 18-crown-6 (47 mM) and DTPA (4
mM) were prepared in toluene and 99% toluene-d8. The solutions
were vortexed for 1 min, followed by centrifugation at 3300 rpm
for 2 min to allow settling of any undissolved KO2. Solutions of
dissolved KO2 were used immediately upon preparation. Stock
solutions of dibenzoyl peroxide (DBP, 0.035 M) and DABCO
(0.026 M) were prepared in toluene. A stock solution of 1a (430
µM) was prepared in a 50% toluene/50% THF mixture. For each
trial, the following were combined in an amber 2-mL autosampler
vial containing a micro stir bar: 373 µL of KO2 solution, 117 µL
of stock solution of 1a (100 µM final concentration), and 10 µL of
stock DBP solution (0.71 mM final concentration). The total
reaction volume was 500 µL. For all trials, the DBP was added
last and marked time zero for the reaction. Reaction vials were
thermostated at 25 °C for the duration of the kinetic assay. Time
points were obtained and CL signals were recorded as described
above for the MNPO2 reaction. In the deuterated toluene experi-
ment, the final composition of the reaction was 85% toluene-d8. In
the DABCO quenching experiment, the reaction contents were
changed as follows: 363 µL of KO2 solution (toluene), 117 µL of
stock solution of 1a (100 µM final concentration), 10 µL of stock
DBP solution (0.71 mM final concentration), and 10 µL of stock
For the present study, we focused primarily on one probe, 1a,
to characterize the sensitivity and selectivity of our detection
method. Using an aqueous solution of 1a, we trapped and detected
1O2 generated by thermal decomposition of a water-soluble
naphthalene endoperoxide, MNPO244,49 (Scheme 2). MNPO2, like
1,4-naphthalenedipropanoate,50,51 is a water-soluble 1O2 carrier used
1
in the chemical generation of O2. In this experiment, MNPO2
was allowed to thermally decompose (34 °C) in pH 7.8 buffered
water in the presence of 1a. Kinetic time points were obtained
by sampling aliquots of the reaction mixture and quenching the
reaction by dilution in CH3CN. The luminescent decomposition
of the dioxetane (2a) was triggered by TBAF-induced desilylation,
and the integrated signal intensity was converted to concentration
2a as described above. The production of 2a displayed first-order
kinetics (k34°C ) 2.8 ((0.4) × 10-4
s
-1) (Figure 1). This rate
constant agrees well with the reported MNPO2 decomposition rate
constant (k37°C ) 5 × 10-4 s-1).44 The same experiment was carried
out in the presence of a competitive probe with a known trapping
rate, FFA (krxn ) 8.3 × 107 M-1
s
-1).45 From the rates of FFA
degradation, monitored by HPLC, and production of 2a, we
determined a rate constant for the reaction of 1a with 1O2, krxn, in
water of 2.2 ((0.7) × 106 M-1
s
-1. Through comparison to a
competitive probe with a well-known rate constant for reaction
(49) Aubry, J. M.; Cazin, B. Inorg. Chem. 1988, 27, 2013-2014.
(50) Di Mascio, P.; Sies, H. J. Am. Chem. Soc. 1989, 111, 2909-2914.
(51) Muller-Breitkreutz, K.; Mohr, H.; Briviba, K.; Sies, H. J. Photochem. Photobiol.
B 1995, 30, 63-70.
(47) Messner, K. R.; Imlay, J. A. Methods Enzymol. 2002, 349, 354-361.
(48) Zhao, H.; Kalivendi, S.; Zhang, H.; Joseph, J.; Nithipatikom, K.; Vasquez-
Vivar, J.; Kalyanaraman, B. Free Radical Biol. Med. 2003, 34, 1359-1368.
Analytical Chemistry, Vol. 77, No. 4, February 15, 2005 1203

Figure 2. Comparison of the CL signals elicited from 5 µM MCLA
and 10 µM 1a when exposed to three ROS: 1O2, O2•-, and H2O2.
Filled columns with error bars represent the average and one standard
deviation for three measurements, and the unfilled columns represent
the blank signal in each system.
Figure 3. Rate of O2•- production (measured by the cytochrome c
assay) in the presence of varying concentrations of 1a (squares) or
SOD (filled circles). Note the difference in concentration scales
between 1a and SOD.
determined using the cytochrome c assay,54 remained constant
even in the presence of 1 mM 1a (Figure 3).
with 1O2, the rate constant for reaction of 1a with 1O2 is calculated
with greater certainty. The alternative is to use the reported yield
of 1O2 from the thermal decomposition of MNPO2, which is
temperature-dependent and known with less certainty.
1
The trap-and-trigger methodology of O2 detection is particu-
larly useful in the quantitation of photochemically generated 1O2;
the ability to trap 1O2 for later quantitation allows for low-
background CL measurements that would be more difficult for
probes (e.g., MCLA18 and the thiafulvalene-substituted anthracene
trap28,29) that emit spontaneously upon reaction with 1O2. Aqueous
1O2, generated photochemically using perinaphthenone, was
trapped and detected by 1a and 1b in separate experiments. The
rates of formation of 2a and 2b were compared to the degradation
of FFA under the same conditions to yield reaction rate constants
1
The somewhat lower reactivity of 1a toward O2 (compared
to rate constants on the order of 107-108 M-1 s-1 for FFA,45
DMF,52 and BTMPC14) is a disadvantage and a factor that may
be addressed by replacing the vinyl ether methyl substituent with
a more electron-donating group in order to increase the probe’s
1
reactivity toward O2. For 1a, the luminescence efficiency of the
aryloxy emitter and the sensitivity of CL detection compensate
for its lower reactivity. The sensitivity of 1a for the detection of
1
of 1.7 ((0.2) × 106 and 3.5 ((0.4) × 106 M-1
s
-1, respectively.
This rate constant for 1a agrees well with that determined using
aqueous O2 was estimated by combining the limit of detection
for 2a (5 pM) with the krxn value for 1a in water. With 10 µM 1a,
1
we estimate that a steady-state O2 concentration, [1O2]SS, of 5 ×
MNPO2 as a source of 1O2. Rate constants for the total quenching
10-13 M can be accurately measured in a 10-min exposure time.
The [1O2]SS by definition is a balance of 1O2 formation and
deactivation pathways and reflects the short lifetime of 1O2 in water
(∼4 µs).53
1
of O2 by 1a and 1b, ktot, in CH2Cl2, determined by laser flash
1
photolysis (monitoring the 1270-nm O2 luminescence, see Sup-
porting Information for experimental details), were lower but also
indicate a 2-fold reactivity difference (1a, 4.6 ((0.3) × 105 M-1
The selectivity of our detection method for 1O2 was examined
through a direct comparison of 1a to MCLA. Each probe was
exposed to three ROS, 1O2, O2•-, and H2O2, under identical
conditions in water (pH 7.8, 50 mM phosphate buffer), and the
resulting CL signals were compared. For this selectivity assay,
s
-1; 1b, 8.7 ((0.2) × 105 M-1
s
-1).
As a rigorous test of the selectivity of our detection method
1
1
for O2 in an ROS-rich environment, the production of O2 from
the reaction of dibenzoyl peroxide (DBP) and O2 was investi-
•-
gated using probe 1a. Danen and Arudi previously investigated
the reaction of DBP and O2•-, quantifying the 1O2 produced using
the 1O2 traps 1,3-diphenylisobenzofuran and 1,2-dimethylcyclo-
1O2 was generated by the thermal dissociation of MNPO2. The
•-
xanthine/xanthine oxidase system was used to generate O2
.
The results of the selectivity comparison are summarized in
Figure 2, and they illustrate the superior selectivity of 1a for 1O2
compared to MCLA. The CL signals measured for 1a after
•-
hexene.6 In this study, DBP was reacted with O2 in toluene in
the presence of 1a and DTPA, added to bind trace metal
impurities). Time points were obtained as in the reaction with
MNPO2, and the formation of 2a over time was monitored by
measuring its CL signal, as described above. Danen and Arudi
proposed the following stoichiometry for the reaction of O2•- with
DBP (eq 1).6
•-
exposure to O2 and H2O2 are indistinguishable from the
background signals and are 93-97% lower than the signal from
1O2. In contrast, under identical conditions, all three ROS yield a
significant signal from MCLA. In all cases, the signals observed
for MCLA after ROS exposure are significantly larger than the
1
•-
background signals. The selectivity of 1a for O2 relative to O2
-
2O2•- + PhC(dO)OOC(dO)Ph f 2O2 + 2PhCO2 (1)
•-
was also verified through the observation that O2 production,
(52) Scully, F. E., Jr.; Hoigne, J. Chemosphere 1987, 16, 681-694.
(53) Schmidt, R. J. Am. Chem. Soc. 1989, 111, 6983-6987.
(54) McCord, J. M.; Fridovich, I. J. Biol. Chem. 1968, 243, 5753-5760.
1204 Analytical Chemistry, Vol. 77, No. 4, February 15, 2005

The production of 2a was fit well by a second-order kinetic
model incorporating this stoichiometry (eq 2).55 Here [2a]t is the
[DBP]0 - ∆0[DBP]0
[2a]t ) Rf
(2)
[O2•-]0e∆ kt - 2[DBP]0
0
(
)
concentration of dioxetane 2a at time t, ∆0 ) [O2•-]0 - 2[DBP]0,
R is the yield of O2 from the reaction of DBP and O2•-, and f is
1
1
the fraction of the O2 that is trapped by probe 1a (eq 3), where
krxn[1a]
f )
(3)
ksolv + kDTPA[DTPA] + krxn[1a]
1
Figure 4. Production of dioxetane 2a during exposure of 1a (100
µM) to KO2 (saturated solution, solubilized with 18-crown-6) and DBP
(0.71 mM) at 25 °C in toluene (circles) or 85% toluene-d8 (squares).
Data are fit to eq 2 with [DBP]0 ) 0.71 mM and assuming [O2•-]0 )
10 mM. Open and closed symbols represent duplicate trials under
each set of conditions. Inset shows data from duplicate experiments
under similar conditions (toluene), with (triangles) and without (circles)
the addition of DABCO (520 µM), a 1O2 quencher.
krxn is the rate constant for reaction of 1a with O2 and ksolv and
1
kDTPA are the rate constants for quenching of O2 by the solvent
and DTPA, respectively. Under the reaction conditions for this
experiment, only a small fraction of the O2 produced is trapped
(0.02%). The major O2 quenching processes in this system are
1
1
relaxation by the solvent (toluene) and quenching by DTPA.
1
The identity of O2 as the ROS responsible for the observed
formation of 2a in this reaction is supported by the observation
that neither KO2 nor DBP alone elicits a response from 1a. In
addition, performing the reaction in 85% deuterated toluene
resulted in a 1.8-fold enhancement in the yield of 2a (Figure 4).
Based upon the 1O2 quenching rate constants for the solvent
components and DTPA (ktot,DTPA ) 1.0 ((0.1) × 107 M-1 s-1, see
Supporting Information for experimental details) and the longer
1O2 lifetime in toluene-d8, an enhancement of 1.5 was expected.
Thus, the observed enhancement is in good agreement with
prediction. Also, performing the reaction with DABCO, a known
1O2 quencher, completely inhibited the formation of 2a (Figure
4, inset).
1O2 from the reactions of O2•- with DBP and other peroxides, and
studies in which probes 1a and 1b are used to quantify O2 in
sunlit natural waters are ongoing.
1
ACKNOWLEDGMENT
We thank Scott Brown for his contributions to the experimental
work described here, and the University of Minnesota for funding.
D.E.L. thanks the University of Minnesota Graduate School for a
Dissertation Fellowship award.
The results reported here demonstrate the sensitivity and
selectivity of probe 1a for 1O2 with respect to other ROS and give
SUPPORTING INFORMATION AVAILABLE
1
further evidence for the production of O2 in the reaction of acyl
The syntheses of 1a, 1b (and their precursors), 2a, and 3-(1,4-
epidoxy-4-methyl-1,4-dihydro-1-naphthyl)propanoate (MNPO2); ex-
perimental details for laser flash photolysis experiments to
measure rate constants for the total quenching of 1O2 by 1a, 1b,
and diethylenetriaminepentaacetic acid (DTPA). This material is
available free of charge via the Internet at http://pubs.acs.org.
peroxides with O2•-. The sensitivity of this CL-based detection
method enables its use in environmental systems where steady-
state concentrations of photochemically generated 1O2 are typically
subpicomolar, and the high degree of selectivity permits the
unambiguous detection of 1O2 in complex systems involving other
ROS. The trap-and-trigger nature of the detection method allows
for kinetic analyses and further investigation of reaction mecha-
nisms. Further studies are underway to determine the yield of
Received for review November 17, 2004. Accepted
January 2, 2005.
(55) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms; McGraw-Hill:
New York, 1995.
AC048293S
Analytical Chemistry, Vol. 77, No. 4, February 15, 2005 1205