Prefluorescent nitroxide probe for the highly sensitive determination of peroxyl and other radical oxidants

Article abstract of DOI:10.1021/ac901374m

Fluorescamine derivatized 3-amino-2,2,5,5,-tetramethyl-1-pyrrolidinyloxy (I) is shown to undergo an irreversible reaction with peroxyl radicals and other radical oxidants to generate a more highly fluorescent diamagnetic product (II) and thus can be used

Full text of DOI:10.1021/ac901374m

Anal. Chem. 2009, 81, 8033–8040
Prefluorescent Nitroxide Probe for the Highly
Sensitive Determination of Peroxyl and Other
Radical Oxidants
Min Jia,† Yu Tang,†,‡ Yiu-Fai Lam,† Sarah A. Green,*,‡ and Neil V. Blough*,†
Department of Chemistry and Biochemistry, University of Maryland, College Park, Maryland 20742, and Chemistry
Department, Michigan Technological University, Houghton, Michigan 49931
Fluorescamine derivatized 3-amino-2,2,5,5,-tetramethyl-
1-pyrrolidinyloxy (I) is shown to undergo an irreversible
reaction with peroxyl radicals and other radical oxidants
to generate a more highly fluorescent diamagnetic product
(II) and thus can be used as a highly sensitive and
versatile probe to determine oxidant production optically,
either by monitoring the changes in fluorescence intensity,
by HPLC analysis with fluorescence detection, or by a
combination of both approaches. By changing the [O2]/
[I] ratio, we show that peroxyl radicals can be detected
and quantified preferentially in the presence of other
radical oxidants. Detection of photochemically pro-
duced peroxyl radicals is achieved by employing
3-amino-2,2,5,5,-tetramethyl-1-pyrrolidinyloxy (3-ap)
alone, followed by derivatization with fluorescamine.
With employment of HPLC analysis, the detection limit
of II at a S/N of 2 is ∼3 nM for a 125 µL injection.
Preliminary applications include the detection of per-
oxyl radicals generated thermally in soybean phos-
phatidylcholine liposomes and produced photochem-
ically in tap water.
most of these methods suffer from one or more problems such
as lack of selectivity, low sensitivity, limited stability, and most
importantly, difficulties in quantification.
Since their initial introduction in the late 1980s and early
1990s,7-11 the use of prefluorescent nitroxide sensors12 to detect and
quantify radical formation has expanded to include a wide range of
applications.13-23 This approach utilizes stable nitroxide radicals as
optical switches. By covalently coupling a nitroxide at a short distance
from a chromophore, fluorescence emission from the chromophore
can be largely quenched.7,8,24 Upon reaction of the nitroxide moiety
with (usually carbon-centered) radicals to form diamagnetic prod-
ucts,25,26 the intramolecular quenching pathway is eliminated and
fluorescence emission is greatly enhanced,7,8,24 thereby allowing
radicals to be detected and quantified, either through changes in
(6) Heyne, B.; Beddie, C.; Scaiano, J. C. Org. Biomol. Chem. 2007, 5, 1454–
1458
(7) Blough, N. V.; Simpson, D. J. J. Am. Chem. Soc. 1988, 110, 1915–1917
(8) Green, S. A.; Simpson, D. J.; Zhou, G.; Ho, P. S.; Blough, N. V. J. Am. Chem.
Soc. 1990, 112, 7337–7346
(9) Kieber, D. J.; Blough, N. V. Anal. Chem. 1990, 62, 2275–2283
(10) Kieber, D. J.; Blough, N. V. Free Radical Res. Commun. 1990, 10, 109–
117
(11) Gerlock, J. L.; Zacmanidis, P. L.; Bauer, D. R.; Simpson, D. J.; Blough, N. V.;
Salmeen, I. T. Free Rad. Res. Commun. 1990, 10, 119–121
.
.
.
.
.
.
(12) A term originally coined by Scaiano and co-workers and employed
Peroxyl radicals, formed by the rapid addition of molecular
oxygen to carbon-centered radicals, play an important role in many
pathological processes, chronic diseases, and aging due to their
involvement in the oxidative degradation of DNA and proteins
and in the autoxidation of lipids within cell membranes.1 These
species are also key intermediates in thermal and photochemical
degradation of both natural and man-made materials.2
Although a variety of approaches have been employed to detect
peroxyl radicals, including spin trapping with electron paramag-
netic resonance (EPR) detection3 and fluorescence sensors,4-6
herein: Ballesteros, O. G.; Maretti, L.; Sastre, R.; Scaiano, J. C. Macromol-
ecules 2001, 34, 6184–6187
(13) Zhu, Q.; Lian, Y.; Thyagarajan, S.; Rokita, S. E.; Karlin, K. D.; Blough, N. V.
J. Am. Chem. Soc. 2008, 130, 6304–6305
(14) Gan, D.; Jia, M.; Vaughan, P. P.; Falvey, D. E.; Blough, N. V. J. Phys. Chem.
A 2008, 112, 2803–2812
(15) Alaghmand, M.; Blough, N. V. Environ. Sci. Technol. 2007, 41, 2364–2370
(16) Fairfull-Smith, K. E.; Blinco, J. P.; Keddie, D. J.; George, G. A.; Bottle, S. E.
Macromolecules 2008, 41, 1577–1580
(17) Coenjarts, C.; Garcia, O.; Llauger, L.; Palfreyman, J.; Vinette, A. L.; Scaiano,
J. C. J. Am. Chem. Soc. 2003, 125, 620–621
(18) Aliaga, C.; Juarez-Ruiz, J. M.; Scaiano, J. C.; Aspee, A. Org. Lett. 2008, 10,
2147–2150
(19) Flicker, T. M.; Green, S. A. Environ. Health Perspect. 2001, 109, 765–771
(20) Chen, W.; Wang, X.; Tu, X.; Pei, D.; Zhao, Y.; Guo, X. Small 2008, 6, 759–
764
(21) Kalai, T.; Schindler, J.; Balog, M.; Fogassy, E.; Hideg, K. Tetrahedron 2008,
64, 1094–1100
(22) Borisenko, G. G.; Martin, I.; Zhao, Q.; Amoscato, A. A.; Kagan, V. E. J. Am.
Chem. Soc. 2004, 126, 9221–9232
(23) Fairfull-Smith, K. E.; Bottle, S. E. Eur. J. Org. Chem. 2008, 32, 5391–5400
(24) Herbelin, S. E.; Blough, N. V. J. Phys. Chem. B 1998, 102, 8170–8176
.
.
.
.
.
.
.
.
* To whom correspondence should be addressed. Phone, 301 405 0051; fax,
301 314 9121; e-mail, neilb@umd.edu (N. V. Blough). Phone, 906 487 2048; fax,
906 487 2061; e-mail, sgreen@mtu.edu (S. A. Green).
.
† University of Maryland.
.
‡ Michigan Technological University.
(1) Halliwell, B.; Gutteridge, J. M. C. Free Radicals in Biology and Medicine,
4th ed.; Oxford University Press: Oxford, U.K., 2007.
(2) Micallef, A. S.; Blinco, J. P.; George, G. A.; Reid, D. A.; Rizzardo, E.; Huang,
.
.
.
S. H.; Bottle, S. E. Polym. Degrad. Stab. 2005, 89, 427–435
(3) Jones, C. M.; Burkitt, M. J. J. Chem. Soc., Perkin Trans. 2 2002, 2044–
2051
(4) DeLange, R. J.; Glazer, A. N. Anal. Biochem. 1989, 177, 300–306
(5) Maulik, G.; Salgia, R.; Makrigiorgos, G. M. Methods Enzymol. 2002, 352,
80–91
.
(25) Ingold, K. U. In Landolt-Bornstein Numerical Data and Functional Relation-
ships in Science and Technology, Subvolume C, Radical Reaction Rates in
Liquids; Fischer, H., Ed.; Springer-Verlag: New York, 1983; Vol. 13, pp 166-
270.
.
.
(26) Beckwith, A. L. J.; Bowry, V. W.; Ingold, K. U. J. Am. Chem. Soc. 1992,
.
114, 4983–4992.
10.1021/ac901374m CCC: $40.75  2009 American Chemical Society
Published on Web 09/11/2009
Analytical Chemistry, Vol. 81, No. 19, October 1, 2009 8033

Scheme 1. Reaction of I with Carbon-Centered Radicals to Form the Alkoxylamine Product under Anaerobic
Conditions and with Peroxyl Radicals to Form II under Aerobic Conditions
fluorescence intensity7,11,12,17,18 or through product analysis by high-
performance liquid chromatography (HPLC)9,27-31 or liquid chro-
matography/mass spectrometry (LC/MS).32,33
EXPERIMENTAL SECTION
Chemicals. Nitroxides (3-ap, 3-aminomethyl-2,2,5,5-tetram-
ethyl-1-pyrrolydinyloxy (3-amp), 3-carbamoyl-2,2,5,5-tetramethyl-
1-pyrrolidineoxy (3-cp), and 4-amino-2,2,6,6-tetramethylpiperidi-
nyloxy (4-at)) and fluorescamine were purchased from Acros and
Sigma Aldrich, respectively. 2,2′-Azo-bis-(2-amidinopropane) di-
hydrochloride (AAPH) and azobisisobutyronitrile (AIBN) were
purchased from Cyman Chemical Co. Sodium phosphate (99.999%)
and sodium hydroxide were obtained from Aldrich. DMSO, HPLC
grade methanol, acetonitrile, and glacial acetic acid were all
purchased from Fisher. Refined lecithin was purchased from Alfa
Aesar. All chemicals were used without further purification. Pure
water for all experiments was obtained from a Millipore Milli-Q
system.
Here we demonstrate that the prefluorescent nitroxide probe,
I (Scheme 1), synthesized by reaction of 3-amino-2,2,5,5,-tetram-
ethyl-1-pyrrolidinyloxy (3-ap) with fluorescamine,27,28 undergoes
a unique, irreversible reaction with peroxyl radicals to form a more
highly fluorescing diamagnetic product (II) and thus can be
employed to detect and quantify peroxyl radicals optically.
Although other one-electron (radical) oxidants such as •NO2 and
CO3•- are also shown to react with I to form II, we demonstrate
that peroxyl radicals can be preferentially detected in their
presence by changing the [O2]/[I] ratio (Scheme 1).
In this article, the stoichiometry of the peroxyl radical reaction
with I is first established and is shown to differ from that of other
amino-nitroxides. The use of I to determine the formation rate of
peroxyl radicals by thermolysis and radiolysis of simple model
systems is shown, with this approach then employed to determine
the production of peroxyl radicals in an aqueous dispersion of
soybean phosphatidylcholine (PC) liposomes. Finally, 3-ap alone
is used to determine peroxyl radicals generated photochemically
from tap water following derivatization of its product with fluo-
rescamine and subsequent chromatographic analysis.
Thermolysis. Thermolysis of AAPH was carried out in a 4
mL quartz cuvette in pH 7.4, 50 mM sodium phosphate buffer in
a thermostatted water bath at 38.5 0.5 °C. The stock AAPH
solution (80 mM) was prepared daily and stored in an ice bath
prior to thermolysis. Reactions were initiated by immersion of the
reaction vessel in the water bath. Reactions were terminated by
placing aliquots of the reaction mixture in a -20 °C freezer for
minimally 10 min. For anaerobic experiments, samples were
purged for 20 min with N2, with the headspace of the cuvette
continuously purged with N2 during thermolysis. The liposome
preparation procedure has been reported previously.34
Briefly, 1 mg of AIBN and 375 mg of lecithin were dissolved in 5
-36
(27) Li, B.; Gutierrez, P. L.; Blough, N. V. Anal. Chem. 1997, 69, 4295–4302
(28) Li, B.; Gutierrez, P. L.; Blough, N. V. Methods Enzymol. 1999, 300, 202–
216
(29) Li, B.; Gutierrez, P. L.; Amstad, P.; Blough, N. V. Chem. Res. Toxicol. 1999,
12, 1042–1049
(30) Thomas-Smith, T. E.; Blough, N. V. Environ. Sci. Technol. 2001, 35, 2721–
2726
.
.
.
(34) Yamamoto, Y.; Niki, E.; Kamiya, Y. Lipids 1982, 17, 870–877
(35) Yamamoto, Y.; Niki, E.; Kamiya, Y.; Shimasaki, H. Biochim. Biophys. Acta
1984, 332–340
(36) Takahashi, M.; Tsuchiya, J.; Niki, E. J. Am. Chem. Soc. 1989, 111, 6350–
6353
.
.
.
(31) Vaughan, P. P.; Blough, N. V. Environ. Sci. Technol. 1998, 32, 2947–2953
.
(32) Johnson, C. G.; Caron, S.; Blough, N. V. Anal. Chem. 1996, 68, 867–872
(33) Kieber, D. J.; Johnson, C. G.; Blough, N. V. Free Radical Res. Commun.
.
.
(37) Notre Dame Radiation Chemistry Data Center. http://www.rcdc.nd.edu/
Solnkin2/.
1992, 16, 35–39
.
8034 Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

mL of chloroform, with the solvent then evaporated to produce a
thin film. I (28.4 µM) dissolved in 5 mL of 0.1 M NaCl solution
was added to obtain a milky liposome suspension. This suspension
RCM 8 cm × 10 cm Waters radial compression module was
employed for separation. In most cases, the mobile phase was
65% sodium acetate buffer (50 mM, pH 4.0) and 35% methanol
(v/v) with a flow rate of 1.0 mL/min. The temperature of the
column was maintained at 29 °C. The fluorescence detector was
a Hitachi model L4870 set at 390 nm (excitation) and 490 nm
(emission). A 125 µL sample loop was employed.
Fluorometer. Emission spectra were collected with an SLM-
Aminco AB2 spectrofluorometer. Wavelengths were set at 390
nm (excitation) and 400-700 nm (emission) with 4 nm
bandpasses on both the excitation and emission monochroma-
tors. Fluorescence quantum yields were measured as described
previously.24
Electron Paramagnetic Resonance. A Bruker ESP300E EPR
spectrometer was used to follow the nitroxide spin loss. Samples
were drawn into 50 µL capillary tubes, sealed top and bottom with
Critoseal, and placed within standard 3 mm i.d. quartz EPR tubes.
Standard instrument settings were frequency 9.8 GHz; microwave
power 9.4 mW; modulation amplitude 1.0 G. Spin levels were
determined from areas obtained by double integration of the EPR
spectra, as compared with known concentrations of 3-cp.
Cyclic Voltammetry. Cyclic voltammetry of the nitroxides was
performed using a homemade three-electrode cell on a CHI660A
workstation. The scan range was 0-1.0 V vs Ag/AgCl reference
electrode, and the scan rate was 100 mV/s. Both the working
and counter electrodes were platinum wire. Nitroxides were
dissolved in pH 7.4, 50 mM sodium phosphate buffer saturated
with K2SO4.
was promptly placed into a 50
1 °C water bath under an
atmospheric of oxygen to initiate the reaction. The reaction was
terminated in the same way as in AAPH experiments. The
suspensions were filtered through a 0.2 µm Puradisc filter before
injection into an HPLC column.
Steady State Radiolysis. Steady state radiolysis was per-
formed using a 60Co source at room temperature. The dose rate
was determined in each experiment using a Fricke dosimeter.
DMSO (20 mM) in 10 mM pH 7.4 phosphate buffer was
saturated either with N2O alone or with 80% N2O/20% O2 in a
Wheaton 5 mL serum vial. The vials were immediately sealed
(while still purging) with Wheaton rubber plug stoppers, which
•-
were then fastened by an aluminum crimping tool. CO3 and
•NO2 were produced by γ irradiation of N2O saturated, 100
mM Na2CO3 solution, or 10 mM NaNO2 solution in 5 mM
phosphate buffer (pH 7.4), respectively. The generation of these
radicals and their possible subsequent reactions are provided
below:
•
-
•
H2O f OH + eaq + H3O+ + H + OH- + H2 + H2O2
(1)
k2 )9.1×109 M-1 s-1
-
eaq + N2O + H2O
•OH + OH- + N2
(2)
Synthesis and Purification of Products. Synthesis of I and
of fluorescamine-derivatized 3-amp and 4-at have been described
previously.24,27,28 Bulk electrolysis of ∼2 mM 3-ap, 3-amp, and 4-at
was performed with a three-electrode system on a BAS CV-50W
voltammetric analyzer in neutral pH buffer saturated with K2SO4.
The voltage was 750 mV vs Ag/AgCl, and the electrolysis time
was 90 min. A 10 mM fluorescamine solution in acetonitrile
was added to the reaction mixture in a 1 mL centrifuge tube
and vortexed for 1 min at room temperature. The pH of the
solution was adjusted to 8.1, and fluorescamine was added in
5-10-fold molar excess relative to the product. II was then
separated and purified by HPLC and then extracted in chlo-
roform. II was dried under N2 flow and stored in the freezer
at -20 °C until use. Reaction of the 3-ap oxidation product with
benzoyl chloride was carried out in aqueous solution. Briefly,
18 mg of 3-ap was oxidized either electrochemically or by
t-APOO radical in 10 mL of aqueous solution in a 25 mL conical
flask. A total of 800 mg of sodium bicarbonate was added while
stirring. Benzoyl chloride (24 mg) was added into the solution,
and the mixture was stirred for 20 min. A volume of 20 mL of
chloroform was used to extract the product. The solvent was
evaporated, and the dry product was redissoved in 3 mL of
50%/50% by volume methanol/water. The product was purified
by RP HPLC (mobile phase, 75% H2O/25% methanol), extracted
into chloroform, and dried under N2. The dry product was
stored at -20 °C until further use.
k4 )6.0×109 M-1 s-1
k5 )3.9×108 M-1 s-1
k6 )1.6×109 M-1 s-1
-
•OH + NO2
•NO2 + OH-
(4)
(5)
(6)
2-
•-
•OH + CO3
CO3 + OH-
•CH3 + O2
CH3OO•
k8 )106-107 M-1 s-1 (ref 39)
CH3OO• + I
•NO2 + I
II
(8)
(9)
k9 )108 M-1 s-1 (ref 40)
II
k10 )108 M-1 s-1 (ref 44)
•-
CO3 + I
II
(10)
Photolysis. Samples were placed in a 1 cm cuvette and
irradiated with a 300 W xenon lamp. The actinic light was first
filtered through 22 cm of Milli Q water and then through a 275
Product Analysis by NMR and Mass Spectrometry. 1H,
13C NMR and COSY, HMBC, HSQC, NOESY, and H2BC were
performed on Bruker 500 and 600 MHz spectrometers. Samples
were prepared in deuterated chloroform at room temperature.
nm long pass filter to provide a light intensity of 150 mW/cm2
.
HPLC. The reversed-phase HPLC system has been described
previously.27,28 A 4 µm RP C18 packing Nova-Pack column in a
Analytical Chemistry, Vol. 81, No. 19, October 1, 2009 8035

A linear quadrupole-ion trap (LTQ) Fourier-transform ion
cyclotron resonance (FTICR) mass spectrometer at the Woods
Hole Oceanographic Institution was employed for high-resolu-
tion mass spectral measurements. The ESI spray voltage was
4.3 kV for positive mode and 4.0 kV for negative mode. The
capillary temperature was 250 °C. A JEOL AccuTOF-CS ESI-
time-of flight (TOF) mass spectrometer at the University of
Maryland was also employed in product analysis.
RESULTS
Reaction Stoichiometry. Alkyl radicals (t-AP•) were pro-
duced by thermolysis of AAPH under anaerobic conditions
(Scheme 1). Under aerobic conditions, these alkyl radicals were
converted quantitatively to peroxyl radicals (t-APOO•; ∼95%),
owing to the high [O2]/[I] ratios employed (varying from ∼5
to 20), as well as to the higher reaction rates of carbon-centered
radicals with O2 (∼2-4 × 109 M-1 s-1
)
38 as compared with the
nitroxides (∼1-8 × 108 M-1
s
-1).25,26 EPR was then used to
follow the spin loss of I under anaerobic and aerobic conditions
to compare the reactivity of I with t-AP• and t-APOO•, respec-
tively (Scheme 1; a representative example of EPR spectra is
provided in Figure S1 in the Supporting Information).
Identical rates of spin loss were observed for I under N2 and
air during the thermolysis of AAPH (Figure 1A), demonstrating
that t-APOO• reacted with I to the same extent as t-AP•.
Further, the rates were independent of [I], indicating that both
t-AP• (under N2) and t-APOO• (under air) were reacting
quantitatively with I. Identical rates of loss were also observed
in the reaction of I with CH3• and CH3OO• radicals generated
by steady-state radiolysis in the absence and presence of 20%
O2, respectively (Figure 1B), showing that this reaction is not
unique to t-APOO•. At the temperatures employed in this study,
nitroxides are known to react irreversibly with carbon-centered
radicals to produce stable alkoxylamine products.7-11,25,26 Thus,
the identical rates of loss in the presence and absence of O2 and
with increasing [I] demonstrates that I also reacts irreversibly
with peroxyl radicals.
Reaction of I with t-APOO• and CH3OO• also led to a
dramatic increase in the steady-state fluorescence intensity
(Figure 2 and Figure S2 in the Supporting Information), which
exhibited a linear relationship with the loss of I as determined
by EPR (Figure 2 inset and Figure S2 inset in the Supporting
Information). Fluorescence quantum yields of HPLC-purified II
(see below) in H2O (0.045) and in pH 7.4, 50 mM phosphate
buffer (0.041) were about 90-fold higher than that obtained
previously for I (0.00046)24 and only about 2-fold lower than
that obtained previously for an alkoxyamine adduct of I
(0.097).24
Figure 1. Reaction of nitroxides with peroxyl radicals. (Panel A) Time
course for the reaction of I with t-AP• (under N2) and t-APOO• (under
air) generated by thermolysis of AAPH (1 mM) in 50 mM phosphate
buffer, pH 7.4 at 38.5 ( 0.5 °C. [O2] ) 210 µM at this temperature.
•
(Panel B) Time course for the reaction of I with CH3 (under N2O)
and CH3OO• (80% N2O/20% O2) generated by γ radiolysis in 10 mM
phosphate buffer, pH 7.4 containing [DMSO] ) 20 mM, [O2] ) 250
µM. (Panel C) Time course for the reaction of fluorescamine deriva-
tized 4-at (4-atf) and fluorescamine derivatized 3-amp (3-ampf) with
t-AP• (under N2) and t-APOO• (under air) in 50 mM phosphate buffer,
pH 7.4. [4-atf] or [3-ampf] ) 10 µM. Other reaction conditions are
the same as in Panel A.
HPLC analysis employing fluorescence detection showed that
reaction of either t-APOO• or CH3OO• with I produced a single
product (II) that was not formed under anaerobic conditions
(Figure 3 and Figure S3 in the Supporting Information). The
fluorescence peak area of II increased linearly with the loss of I
as determined by EPR (Figure 3, inset). This same product was
also generated either by the electrochemical oxidation of 3-ap (see
below) or by 3-ap reaction with peroxyl radicals when followed
immediately by derivatization with fluorescamine at pH 8.1 (Figure
S4 in the Supporting Information). II was also formed by the
reaction of I with •NO2 and CO3•- radicals generated radiolyti-
cally (Figure S5 in the Supporting Information). Prior work has
40
44
shown that nitroxides react with peroxyl,39 •NO2, and CO3
•-
(39) Goldstein, S.; Samuni, A. J. Phys. Chem. A 2007, 111, 1066–1072
(40) Goldstein, S.; Samuni, A.; Russo, A. J. Am. Chem. Soc. 2003, 125, 8364–
8370
(41) Goldstein, S.; Merinyi, G.; Russo, A.; Samuni, A. J. Am. Chem. Soc. 2003,
125, 789–795
(42) Samuni, A.; Goldstein, S.; Russo, A.; Mitchell, J.; Krishna, M.; Neta, P. J. Am.
Chem. Soc. 2002, 124, 8719–8724
.
.
.
(38) Maillard, B.; Ingold, K. U.; Scaiano, J. C. J. Am. Chem. Soc. 1983, 105,
5095–5099
.
.
8036 Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

(Figure 4), clearly revealing that the oxoammonium cation of 3-ap
is far less stable than that of 4-at and 3-amp and reacts rapidly to
form a secondary product. Increasing the scan rate from 100 to
500 mV/s did not lead to the appearance of a reductive wave in
the cyclic voltammogram, further indicating that the lifetime of
this oxoammonium cation must be very short, under roughly one
second, and thus that it is much less likely to be intercepted by
solution reductants as is demonstrated by the results in Figure 1.
Product Analysis. Analysis by AccuTOF ESI-MS of the
electrochemical oxidation product of 3-ap revealed two major
peaks at 157.134 and 118.083 after extraction with chloroform at
pH 9 but only one major peak (m/z 118.083) after extraction at
pH 12 (Figure S6 in the Supporting Information). We further
observed that the initial 3-ap oxidation product, as detected
chromatographically following fluorescamine derivatization to form
II, was not completely stable in aqueous solution but very slowly
converted to a more polar, less fluorescent species as the time
between formation and derivatization was increased (Figure S7
in the Supporting Information). Ultrahigh resolution FTICR mass
spectrometry of the products of 3-ap extracted at pH 9 provided
the molecular formula of these two species, C8H17N2O (M + H,
157.133 47) and C5H12NO2 (M + H, 118.086 20). These results
suggest that the oxoammonium cation of 3-ap first undergoes
rapid ring cleavage and proton loss to form an initial product,
either IIIa or IIIb (M + H, 157.133 47) (Scheme 2), which
subsequently converts at higher pH to a secondary product in
the absence of derivatization with fluorescamine; this product was
tentatively identified as IV (M + H: 118.086 20, C5H12NO2; see
Scheme 2 and below).
Figure 2. Fluorescence emission spectra of II formed by reaction
of I with t-APOO• produced by thermolysis of AAPH (1 mM) at 38.5
( 0.5 °C in 50 mM phosphate buffer, pH 7.4. [O2] ) 210 µM and [I]
) 10 µM. Excitation wavelength was at 386 nm. Inset: Dependence
of fluorescence area on the loss of I as determined by EPR.
In contrast, II, characterized by FTICR MS following HPLC
purification of the fluorescamine-derivatized electrochemical oxi-
dation product of 3-ap, showed no evidence of either of these two
structures. Instead, dominant peaks at m/z 395.160 13 and
393.145 64 were observed in positive and negative ion mode,
respectively (Figure S8 in the Supporting Information), corre-
sponding to a molecular formula of C22H22N2O5 (predicted
masses: 395.160 15 in positive ion mode and 393.145 60 in
negative ion mode). Because the absorption and fluorescence
spectra of II were identical to that of the alkoxylamine adducts
of I (Scheme 1, Figure 2), we concluded that the fluorescamine
moiety within II must be structurally intact. On this basis, II was
tentatively identified as the structure shown in Scheme 2. Because
II was also formed following reaction of I with peroxyl radicals
(Figure 3 and Figure S3 in the Supporting Information) and NO2
Figure 3. Reversed phase HPLC chromatograms showing the
formation of II upon reaction of I with t-APOO• produced by
thermolysis of AAPH (1 mM) at 38.5 ( 0.5 °C in 50 mM phosphate
buffer, pH 7.4. [O2] ) 210 µM and [I] ) 10 µM. The HPLC mobile
phase was 65% sodium acetate buffer/35% MeOH v/v. Excitation
wavelength 390 nm, emission wavelength 490 nm. Injection volume
was 125 µL. Inset: Dependence of fluorescence peak area on the
loss of I as determined by EPR.
radicals to initially form the oxoammonium cation (Scheme 2).
However, II was stable in aqueous solution at room temperature
for at least 12 h, both in the absence and presence of NADH,39-42
illustrating that II was not the oxoammonium cation. Further, the
1:1 stoichiometric reaction of this amino-nitroxide (3-ap) with
peroxyl radicals (Figure 1) was unique; other fluorescamine-
derivatized amino-nitroxides (3-amp and 4-at) exhibited much
lower rates of spin loss in the presence of air than in its absence
(Figure 1C). For these nitroxides, the lower rates of spin loss in
air presumably result from the reversible regeneration of the
nitroxyl moiety via the reduction by solution constituents of the
oxoammonium cation.39-42 Unlike both 4-at and 3-amp, 3-ap
exhibited an irreversible oxidative wave in its cyclic voltammogram
-
and CO3 radicals (Figure S5 in the Supporting Information),
III′ (a or b) is also a likely intermediate in the formation of II
(Scheme 2). Unfortunately, II was not sufficiently stable in organic
solvents over the time frame needed to obtain an unequivocal
identification by NMR.
In an attempt to resolve this issue, the electrochemical
oxidation product of 3-ap was reacted with benzoyl chloride at
pH 8.5 to form the stable amide (Scheme 2). Following HPLC
purification, this product was identified as the cyclic peroxide V
(Scheme 2) as established by FTICR MS (Figure S9 in the
Supporting Information) and extensive NMR analyses (Figure
S10-S23 in the Supporting Information). Although this assign-
ment is consistent with the secondary product IV observed by
MS following oxidation of 3-ap (Figure S6 in the Supporting
(43) Krishina, M. C.; Grahame, D. A.; Samuni, A.; Mitchell, J. B.; Russo, A. Proc.
Natl. Acad. Sci. U.S.A. 1992, 89, 5537–5541
(44) Goldstein, S.; Samuni, A.; Hideg, K.; Merinyi, G. J. Phys. Chem. A 2006,
110, 3679–3685.
.
Analytical Chemistry, Vol. 81, No. 19, October 1, 2009 8037

Scheme 2. Proposed Reaction Pathways for the Oxoammonium Cation of 3-ap and I
Information and Scheme 2), the subsequent reactions by which
this product forms is unclear. Similarly, the secondary reactions
by which II forms is not obvious, although this product clearly
differs from IV, suggesting that the presence of the fluorescamine
moiety alters the reaction pathway from III (and III′) to produce
II (Scheme 2 and Figure S4A,C,D in the Supporting Information).
Regardless, the observation that I (and 3-ap) undergoes irrevers-
ible reaction with peroxyl radicals while most other nitroxides
do not39 can be attributed to the amine moiety at the ꢀ-position,
which we suggest facilitates the rapid initial ring cleavage and
proton loss (Scheme 2).
Selectivity. Our results indicate that other radicals capable
of oxidizing I to the oxoammonium cation will also generate II
(Figure S5 in the Supporting Information) and thus will also be
detected. In addition to •NO2 and CO3•-, such radicals include
the hydroxyl (•OH),,42 alkoxyl (RO•), and possibly superoxide
(O2•-).43,44 Radiolytically produced O2•-, however, did not react
with I to form II even under high fluxes of this radical (8 µM/
min), consistent with our prior results27,28 and the fact that the
rate constants for reaction of superoxide with pyrrolidinyl nitrox-
ides are 2-3 orders of magnitude lower43,44 than those for peroxyl
radicals39-41 at neutral to alkaline pH.
Indeed, because published rate constants for reaction of
pyrrolidinyl nitroxides with peroxyl radicals are reasonably large
(∼106-107 M-1
s
-1),39 the use of low concentrations of I
(<40-50 µM) should allow quantitative reaction with peroxyl
radicals (e.g., Figures 1A,C, 2, and 3 and Figure S3 in the
Supporting Information), while largely precluding interferences
from more highly reactive oxidants such as •OH and RO•, which
under most circumstances will react preferentially with other
solution constituents at higher concentrations. Nevertheless,
this may not be the case when more selective oxidants such
as •NO2 and CO3 are also present. Because these other
•-
oxidants are not O2-reactive, while the formation of peroxyl
radicals is O2-dependent, the presence of oxidants other than
peroxyl radicals should be readily identified by decreasing the
[O2]/[I] ratio (Scheme 1); a decrease in the yield of II, combined
with a concomitant increase in alkoxylamine products, would be
Figure 4. Cyclic voltammetry of nitroxides in K2SO4-saturated
phosphate buffer (50 mM, pH 7.4). Nitroxide concentrations were
2.5-5 mM. Electrodes: platinum (working and auxiliary) and Ag/AgCl
(reference). The scan rate was 100 mV/s, and the scan range was
0-1000 mV.
8038 Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

5B, and Scheme 1); under complete deoxygenation, II decreases
by 10%, in good agreement with the calculated percentage of
methyl radicals generated under these conditions (8%, see above).
As shown in the Supporting Information, the decrease in the yield
of II and the concomitant increase in the yield of VI with
decreasing [O2] are in good agreement with those calculated
based on known rate constants (reactions 1-9 and Scheme 1).
Although the [O2]/[I] ratio was altered by decreasing the
[O2] in the above example, an increase in [I] could also have
been employed. It should be further noted that this example
is particularly challenging in that a small amount of peroxyl
•
radical was detected over a very high background of NO2.
Thermal and Photochemical Stability of I and II. Both I
and II were stable for 12 or more hours in aqueous solution over
a broad range of conditions, including pH ranging from 4 to 10
and temperatures up to 50 °C, as determined by UV-vis,
fluorescence, and EPR spectroscopies and by chromatographic
analysis. However, irradiation of I or II with high-intensity
polychromatic light led to observable photodegradation (15%) of
I and significant photodegradation (60%) of II over the course of
60 min, as determined by UV-vis. These results indicate that
while I can be employed in thermal systems, it will not be suitable
generally for photochemical studies. In contrast, 3-ap and the 3-ap
oxidation product were stable at pH 6-8 under either thermal
(up to 38.5 °C) or irradiation conditions over time scales extending
to 90 min as established by UV-vis and HPLC. Thus, I was
employed to determine peroxyl radicals in thermal systems, while
3-ap, followed by derivatization of the product to form II, was used
in photochemical studies (see below).
Figure 5. Determination of peroxyl radicals in the presence of •NO2
by γ radiolysis of a solution containing 30 µM I, 1.5 mM DMSO, and
20 mM NaNO2. (Panel A) Decrease of II formation with decreasing
[O2]/[I] ratio. HPLC conditions were as in Figure 3. (Panel B) Increase
of methoxylamine product (VI) with decreasing [O2]/[I] ratio. Mobile
phase was 60% MeOH and 40% pH 4.0 acetate buffer. Other HPLC
conditions are as in Figure 3.
Linearity, Precision, and Detection Limit. Employing the
AAPH system, the concentration of II was calibrated by comparing
the loss of nitroxide as measured by EPR with the increase in
the fluorescence peak area in HPLC chromatograms (Figure 3
inset). Once product concentrations in the micromolar range were
established, a calibration curve in the nanomolar range, where
EPR determinations were not possible, was generated by serial
dilution of II. Steady-state emission spectra were calibrated in the
same fashion (Figure 3 inset). Chromatographic peak areas
increased linearly over the concentration range of 5-7500 nM (n
) 9) with an R2 value of 0.998, and the y intercept was not
significantly different from zero. Employing direct fluorometric
detection, the linear range was from 0.66 to 6.2 µM with an R2
value of 0.997. The precision of the HPLC analysis method was
determined from the reproducibility of five HPLC injections
of identical samples under the same experimental conditions
over a period of 7 days at concentrations in the nanomolar
range (10-50 nM). The maximum RSD was 10.6% for HPLC
separation with fluorescence detection. In a similar fashion, the
maximum RSD for direct fluorometric detection method was
9.2%. Defining the detection limit as twice the standard error
of the intercept obtained from linear regressions of calibration
curves of II, with concentrations ranging from 5 to 240 nM (6
points; HPLC separation with fluorescence detection) and from
0.66 to 6.2 µM (6 points; direct fluorometry), we estimate values
of ∼3 nM for the HPLC analysis (125 µL injection volume)
and ∼0.35 µM for direct fluorometric detection.
predicted when peroxyl radicals contribute significantly to the total
oxidant pool (Scheme 1).
To test this idea, solutions containing 1.5 mM DMSO, 20 mM
NaNO2, and N2O/O2 ratios ranging from pure N2O to 80% N2O/
20% O2 were subjected to γ radiolysis in the presence of 30
µM I (Figure 5, see reactions 1-9). Because of the high N2
O
concentrations employed, all e-(aq) is converted quantitatively
to •OH under all conditions (reactions 1 and 2). •OH then reacts
-
•
•
competitively with DMSO and NO2 to form CH3 and NO2
radicals, respectively (reactions 3 and 4). On the basis of the
relative concentrations of DMSO and NO2- and their published
rate constants for the •OH reaction,37 the yield of •NO2 and
•CH3 is calculated to be 92% and 8%, respectively, independent
of [O2]. Under 20% O2, the methyl radicals are converted to
methyl peroxy radicals (∼95%, reaction 6), as calculated from
the relative concentrations of O2 and I and their published rate
constants for reaction with methyl radicals;37 II formation then
•
•
represents the sum of NO2 and CH3 formation (reactions 8
and 9; Figure 5A). With decreasing [O2], II formation decreases,
while the methoxylamine (VI) formation increases due to
greater reaction of I with the methyl radical (reaction 7, Figure
Preliminary Applications. Employing I, peroxyl radical
formation was detected in an aqueous dispersion of soybean PC
Analytical Chemistry, Vol. 81, No. 19, October 1, 2009 8039

at 90 min. This significantly lower yield of the alkoxylamine
products relative to II suggests the presence of an additional
oxidant or oxidants.
CONCLUDING REMARKS
The irreversible reaction of I and 3-ap with peroxyl radicals
and other radical oxidants has two major consequences. First,
because the reaction is stoichiometric (Figure 1A,B) and the
product of these reactions is reasonably stable (II), highly
sensitive, quantitative measurements of the rates of oxidant
formation can be obtained (Figures 1-5 and Figures S2 and S3
in the Supporting Information). Second, because the oxoammo-
nium cation is not stable and rapidly decays to another species
(Figures 3 and 4), possible secondary oxidation reactions by the
cation should be eliminated, thus precluding the formation of
spurious secondary oxidation products.39,45
One disadvantage of this approach is that the peroxy radicals,
unlike most carbon-centered radicals, do not form stable adducts
with nitroxides (Scheme 1, Figures 1-3, and Figure S3 in the
Supporting Information),39-42 and thus information on the struc-
ture of the peroxyl radical is lost. However, as shown here, lower
concentrations of O2 (or higher concentrations of I) can be
employed to allow interception of the carbon-centered radicals
by I prior to O2 reaction, with the stable alkoxyamine products
then generated providing information on structure (Scheme 1
and Figure 5).27,28,32,33 Further, changing the [O2]/[I] ratio can
allow discrimination between peroxyl radicals and other one-
electron oxidants (Figure 5).
Another potential problem is the reduction of I to the
hydroxylamine by certain compounds18,27,28 and by cells27-29 that
also results in enhanced fluorescence. Although this problem
should be alleviated to some extent by the use of low concentra-
tions of I, it will likely limit the direct fluorometric detection in
some cases. However, even in these cases, HPLC analysis will
still allow a clear discrimination among products (Figure 5)9 and
thus the selective detection of these radicals.
Figure 6. Application of I and 3-ap. (Panel A) Formation of II in
soybean PC liposome system incubated at 50 °C in the presence of
1 mM AIBN. [O2] ) 180 µM, [I] ) 28.4 µM, and [PC] ) 100 mM.
HPLC conditions were the same as in Figure 3. (Panel B) Peroxyl
radical detection in irradiated tap water using 3-ap as the probe
followed by derivatization with fluorescamine after irradiation. [I] )
10 µM. Irradiation conditions are as described in the Experimental
Section.
ACKNOWLEDGMENT
lipsomes, a common model for cell membranes.34-36 A water-
insoluble azo compound, 2,2′-azo-bis-isobutyronitrile (AIBN), was
used to initiate autoxidation of the polyunsaturated fatty acids
within the PC liposomes via incubation at 50 °C under air. Analysis
of the reaction mixture by HPLC showed that II increased with
increasing incubation time (Figure 6A), from 2.9 µM at 20 min to
7.4 µM by 50 min. In the absence of the radical initiator (AIBN),
II was not detected (data not shown).
This work was supported by a Graduate Research Board Award
from the University of Maryland to N.V.B. and in part by grants
from the ONR and the NSF (N.V.B.). We thank Prof. M.
Al-Sheikhly and Ms. Alia Weaver of the University of Maryland
for access to and assistance with the 60Co source. We thank Dr.
Elizabeth Kujawinski and Melissa Soule for the FTICR mass
spectra measurements.
SUPPORTING INFORMATION AVAILABLE
Photochemical formation of peroxyl radicals was observed
upon irradiation of tap water with polychromatic light; in this case,
3-ap was used with the product immediately derivatized with
fluorescamine at pH 8.1 after irradiation (Figure 6B). Formation
of II increased with increasing irradiation time from 180 nM at
60 min to 250 nM at 90 min. II was not formed when Milli-Q water
was irradiated in the presence of 3-ap nor when tap water was
irradiated in the absence of 3-ap. To test for the presence of other
oxidants, tap water was irradiated under anaerobic conditions and
the generation of alkoxylamine products was determined chro-
matographically. Consistent with prior work,9 acetyl and methyl
alkoxylamines were identified as the major products, with their
total concentration increasing from 130 nM at 60 min to 200 nM
The product formation of the I reaction with methyl peroxyl
radical; the structure analysis of II and V by FTICR MS and NMR,
and the product formation of •NO2 and CO3 radical reaction
•-
with I. This material is available free of charge via the Internet
at http://pubs.acs.org.
Received for review June 24, 2009. Accepted August 27,
2009.
AC901374M
(45) Lardinois, O.; Maltby, D.; Medzihradszky, K.; Montellano, P. R.; Tomer,
K.; Mason, R.; Deterding, L. Chem. Res. Toxicol. 2009, 22 (6), 1034–1049
.
8040 Analytical Chemistry, Vol. 81, No. 19, October 1, 2009