Reversible reactions of thiols and thiyl radicals with nitrone spin traps

Article abstract of DOI:10.1021/jp049026t

The reactions of the reversible addition of thiols and thiyl radicals to the nitrone spin traps DMPO (5,5dimethyl-1-pyrroline N-oxide) and DEPMPO (5-diethoxyphosphoryl-5-methyl-l-pyrroline N-oxide) are described. Addition of the thiols to the double C=N b

Full text of DOI:10.1021/jp049026t

J. Phys. Chem. B 2004, 108, 9315-9324
9315
Reversible Reactions of Thiols and Thiyl Radicals with Nitrone Spin Traps
Dmitrii I. Potapenko,† Elena G. Bagryanskaya,† Yuri P. Tsentalovich,† Vladimir A. Reznikov,‡
Thomas L. Clanton,| and Valery V. Khramtsov*,§,|
International Tomography Center, NoVosibirsk 630090, Russia, Institute of Organic Chemistry,
NoVosibirsk 630090, Russia, and Institute of Chemical Kinetics & Combustion, NoVosibirsk 630090, Russia,
and Dorothy M. DaVis Heart & Lung Research Institute, The Ohio State UniVersity, 201 HLRI,
473 W 12th AVe, Columbus, Ohio 43210
ReceiVed: March 3, 2004
The reactions of the reversible addition of thiols and thiyl radicals to the nitrone spin traps DMPO (5,5-
dimethyl-1-pyrroline N-oxide) and DEPMPO (5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide) are
described. Addition of the thiols to the double CdN bond of the nitrones results in the formation of the
corresponding hydroxylamines, measured using 31P NMR and the phosphorus-containing trap DEPMPO.
Subsequent mild oxidation of these hydroxylamines into the paramagnetic adducts may interfere with genuine
spin trapping of thiyl radicals representing the Forrester-Hepburn mechanism. The reverse decomposition of
hydroxylamines to the parent nitrone and thiol and of paramagnetic adducts to the nitrone and thiyl radical
were observed for the first time. The recycling of reduced thiols from thiyl radicals by nitrones may comprise
the mechanism of their effective antioxidant activity, in vivo. The release of thiyl radicals upon the breakdown
of the paramagnetic adduct may significantly affect not only the quantitative analysis of the spin trapping
data but even the conclusions regarding the origin of short-lived radical intermediates. The equilibrium constant
for the reactions of the formation of the product of DEPMPO with S-centered nucleophiles decreases in the
series: sulfite > thioglycolic acid > cysteine > glutathione. The rate constants for the reaction of the
monomolecular decomposition of the radical adducts back to the nitrone and glutathiyl radical were found to
be equal to 0.3 ( 0.1 s-1 and 0.02 s-1 for DMPO/GS• and DEPMPO/GS• radical adducts, correspondingly.
Introduction
diamagnetic product formation due to the decay of EPR-
detectable paramagnetic adducts of nitrones with thiyl radicals.
The chemistry of nitrones has significant interest due to their
wide applications as spin traps for EPR detection of short-lived
radical intermediates in biological systems1,2 and their growing
applications as therapeutic agents.3,4 Recently we observed new
reversible nonradical reactions of the nitrone spin traps 5,5-
dimethyl-1-pyrroline N-oxide (DMPO)5 and 5-diethoxyphos-
phoryl-5-methyl-1-pyrroline N-oxide (DEPMPO)6 with (bi)-
sulfite forming the corresponding hydroxylamine compounds.
Mild oxidations of these compounds lead to the formation of
the radical adducts, identical with those obtained by genuine
EPR spin trapping. This reaction might also be of importance
with respect to the therapeutic activity of nitrones, taking into
account sulfite toxicity,7,8 particularly for patients with sulfite
hypersensitivity.9
The ability of nitrones to undergo nucleophilic addition by
(bi)sulfite anions opens a critical question whether similar
reactions may proceed with other S-centered anions of biological
importance, e.g. anions of glutathione (GSH) or cysteine
(CySH). To answer this question, we have used the NMR-spin
trapping approach, which is ideally suited for studies of EPR-
invisible diamagnetic products formed in the reactions of
nitrones with nucleophiles5,6. In a complimentary way, EPR spin
trapping has been used to evaluate alternative pathways of the
Our data support the existence of reversible nucleophilic addition
of thiols to nitrone spin traps. Surprisingly, we also observed
the reversibility of thyil radical adition to the nitrones and, for
the first time, demonstrate monomolecular decomposition of the
corresponding adducts back to thyil radicals and the parent
nitrone. The findings provide a more complete understanding
of the possible biological activity of nitrones and provide
evidence for largerly overlooked reactions that necessitate a
reinterpretation of the many EPR spin trapping experiments.
Materials and Methods
Chemicals. DMPO, DEPMPO, diethylenetriaminepentaacetic
acid (DTPA), disodium and monopotassium phosphate, Chelex-
100, cysteine, GSH, oxidized glutathione (GSSG), sodium
nitrite, formic and thioglycolic acid were obtained from Sigma.
DMPO was distilled under vacuum. nitrosocysteine (CySNO)
was synthesized according to ref 10. Nitrosothioglycolic acid
(AcSNO) was synthesized by combining an equimolar concen-
tration of thioglycolic acid (AcSH) with sodium nitrite in 0.2
N HCl. After that up to 0.1 M of the sodium phosphate was
added, and the solution was titrated to pH 7.0. Nitrosoglutathione
was synthesized as described in the literature11 and stored in a
freezer. Sodium formate was obtained by neutralization of
formic acid by NaOH and dried afterward. For all measurements,
deionized water was used. Phosphate buffer was passed through
column with Chelex-100 to remove trace amounts of transition
metal ions.
* Corresponding author. Tel. (614) 688-3664. Fax (614) 293-4799.
E-mail: khramtsov-1@medctr.osu.edu.
† International Tomography Center.
‡ Institute of Organic Chemistry.
§ Institute of Chemical Kinetics & Combustion.
| The Ohio State University.
10.1021/jp049026t CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/29/2004

9316 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Potapenko et al.
NMR Measurements. NMR spectra were measured with a
Bruker AVANCE 200 spectrometer in 5 mm tubes. In 31P NMR
experiments, D2O and thioglycolic acid were added to 0.2 M
phosphate buffer, 0.5 mM DTPA. The mixture was titrated to
pH 7.0, DEPMPO was added afterward, and 31P NMR spectra
were measured. The final concentration of D2O in the sample
was about 10% to perform a lock standard π/2 pulse sequence
with 1H decoupling. The number of scans was from 100 to 512,
the sweep width was 50 ppm, LB ) 0.5, the acquisition time
was 8.074 s, and the delay between scans was 15 s.
1H NMR spectra were registered in D2O solution. Na2HPO4,
DTPA, and thioglycolic acid were dissolved in deuterium oxide,
and the solution was titrtrated to pHobs 6.58, which corresponds
Figure 1. 31P NMR spectrum of the reaction mixture of DEPMPO
and thioglycolic acid in 0.1 M phosphate buffer. First, 0.2 M Na2HPO4,
0.5 mM DTPA, and 576 mM AcSH were dissolved in D2O solution
and the mixture was titrated to pHobs ) 6.58. Afterward, 42.5 mM
DEPMPO was added and 31P NMR spectrum was measured. Number
of scans, four.
1
to pD ) 7.0.12 DEPMPO was added afterward, and H NMR
spectra were measured. A standard π/2 pulse sequence was used.
Typical settings were as following: sweep width, 10 ppm;
acquisition time, 8.179 s; number of scans, 32; delay between
scans, 5 s.
TABLE 1: Values of pK ) -log KNuH for the Series of
S-centered Nucleophiles and Estimates of the Product Kp )
EPR Measurements. EPR measurements were carried out
using a Bruker EMX spectrometer with a quartz flat cell, 0.25
mm thickness, or 1 mm quartz capillaries. The spectrometer
settings were as following: modulation amplitude, 1.0 G; sweep
width, 80 G for DMPO or 120 G for DEPMPO experiments,
correspondingly; frequency, 10.05 GHz; modulation frequency,
100 kHz; microwave power, 5.19 mW; time constant, 81.92
ms; sweep time, 41.94 s. An eximer Xe-Cl laser with pulse
power of about 30 mJ per pulse, 308 nm wavelength, was used
as the light source in kinetic studies. Radical adduct concentra-
tion was determined relative to the integral signal intensity of
the stable nitroxyl radical, 3-carboxy-2,2,5,5-tetramethylpyrro-
lidine-1-yloxy, measured at the same settings. To study the
possible influence of the dissolved oxygen on the kinetics, the
measurements of the air-saturated and argon-bubbled (30 min)
samples were performed.
K
NuHKadKb for Their Corresponding Reactions of
Nucleophilic Addition to Nitrone Spin Traps DMPO and
DEPMPO, Generalized in Eqs 23-25
spin
trap
NuH
pKNuH
Kp, M-1
ref
-
DMPO
HSO3
HSO3
7.2
7.2
10.05
8.35
9.0
18
5
6
132,a 257b
0.14
-
DEPMPO
DEPMPO
DEPMPO
DEPMPO
HSCH2COO-
CysSH
this work
this work
this work
≈0.03
GSH
<0.003
a Trans-addition. b Cis-addition.
a LeCroy 9310A digitizer. The monitoring light, focused in a
rectangle of 3 mm height and 1 mm width passed through the
cell along the front (laser irradiated) window. Thus, in all
experiments the excitation optical length was 1 mm, and the
monitoring optical length was 8 mm. All solutions were bubbled
with argon starting 15 min prior to and during irradiation. The
experiments were carried out at room temperature in 0.1 M
phospate buffer, pH 7.0.
UV Lamp Photolysis. Photolysis was performed using a
high-pressure mercury lamp, 500 W, in a quartz cell with optical
path 1.0 cm. The wavelengths band from 500 to 600 nm was
selected using light filters. The samples containing 10 mM
nitrosothiol, 42.5 mM DEPMPO in 0.1 M phosphate buffer,
pH 7.0, 10% D2O, were irradiated with light intensity 15 mW/
cm2 during 1.5 h. The solution was bubbled by Ar for the period
of time starting 10 min before the irradiation and until the end
of photolysis. The spectrophotometric measurements showed
about 90% of light-induced decomposition of the nitrosothiols.
31P NMR spectra were measured after irradiation and compared
with those measured before irradiation.
Results
31P NMR Studies of Nucleophilic Addition of Thiols to
the Spin Trap DEPMPO. Figure 1 shows 31P NMR spectrum
of the reaction mixture of 288 µmol of thioglycolic acid and
21.3 µmol of DEPMPO in 0.5 mL of phosphate buffer in D2O,
pHobs ) 6.58. The appearance of the new line at 29.72 ppm
indicates the formation of a new diamagnetic product, P, in the
amount of 1.65 µmol with a corresponding decrease of the
DEPMPO signal down to 19.6 µmol.
The identical spectra were obtained in air-saturated and Ar-
bubbled buffer solutions, in the absence of the oxidants,
therefore favoring a nonradical mechanism for the reaction. The
dilution of the sample three times resulted in a proportional
decrease of the product P down to 0.55 µmol. The results are
also consistent with recycling of the DEPMPO, increasing to
20.7 µmol, from backward transformation of the product P to
the parent DEPMPO. Figure 2 shows that the concentration of
the product P is directly proportional to the concentrations of
DEPMPO and thioglycolic acid, supporting an equilibrium
between parent DEPMPO, AcSH, and DEPMPO/AcSH adduct
(see Figure 2 captions and Table 1 for the equilibrium constant).
The above observation is qualitatively identical to the
previously described reversible reaction of DEPMPO with (bi)-
sulfite, resulting in the formation of a hydroxylamine deriva-
Laser Photolysis. A Nd:YAG laser (wavelength 532 nm,
power 30 mJ per pulse, frequency 25 Hz, irradiated sample area,
1 cm2) was used as a light source. The 2 mL solution of the
GSNO and DEPMPO in 0.1 M phosphate buffer was irradiated
during 30 min in a quartz cell with optical path 1 cm. The
solution was bubbled by Ar for the period of time starting 10
min before the irradiation and until the end of photolysis. After
irradiation the solution was transferred to a NMR tube and the
31P NMR spectrum was measured.
Laser Flash Photolysis (LFP). A detailed description of the
LFP equipment has been published previously13,14. Solutions
in a rectangular cell (10 mm × 10 mm) were irradiated with a
Lambda Physik EMG 101 excimer Xe-Cl laser (308 nm, pulse
energy up to 100 mJ, pulse duration 15-20 ns). The dimensions
of the laser beam at the front of the cell were 3 mm × 8 mm.
The monitoring system includes a DKSh-150 xenon short-arc
lamp connected to a high current pulser, a homemade mono-
chromator, a 9794B photomultiplier (Electron Tubes Ltd.), and

Reaction of Thiols and Thiyl Radicals with Nitrones
J. Phys. Chem. B, Vol. 108, No. 26, 2004 9317
Figure 3. EPR spectra obtained under laser photolysis at 308 nm of
5 mM GSNO (a, b) or 5 mM AcSNO (c, d) in the presence of 88.5
mM spin trap (a, c, DMPO; b, d, DEPMPO) in 0.1 M phosphate buffer,
pH 7.0. Laser frequency was 4 Hz, and the number of scans four. The
Figure 2. Dependence of the concentration of diamagnetic product P
on the DEPMPO and thioglycolic acid concentrations. AcSH was added
to 0.2 M phosphate buffer solution, 0.5 mM DTPA, 10% D2O. The
mixture was titrated to pH 7.0 and then DEPMPO was added and the
31P NMR spectrum was measured. Number of scans was from 100 to
512, depending on the concentration of the product P. (O) Variation
of AcSH concentration at initial DEPMPO concentration, [DEPMPO]0
) 42.5 mM; (9) variation of DEPMPO concentration at initial AcSH
concentration, [AcSH]0 ) 288 mM. The excellent linear extrapolation
of the data is in agreement with eq 26, namely, [P]/[DEPMPO] )
hyperfine splittings determined by EPR spectra simulations are aN
)
15.12 G, aH ) 16.11 G (a); aP ) 45.81 G, aN ) 14.14 G, aH ) 14.94
G (b); aN ) 15.19 G, aH ) 16.92 G (c); aP ) 46.57 G, aN ) 14.26 G,
aH ) 15.53 G (d).
To test this hypothesis, we used the photolysis of the
nitrosothiols GSNO and AcSNO, both by UV lamp and laser
in the presence of spin traps DEPMPO or DMPO. The
photolysis of nitrosothiols leads to the formation of thiyl radicals
via homolitic cleavage of the S-N bond:16-19
K
NuHKadKb[AcSH], yielding KNuHKadKb ) 0.14 ( 0.1 M-1 (see Discus-
sion).
hν8 RS• + NO•
(2)
tives.6 Therefore, in a similar way, we assigned the product P
to the corresponding hydroxylamine of DEPMPO, namely
[5-(diethoxyphosphoryl)-5-methyl-1-oxy-4,5-dihydro-3H-pyrrol-
2-ylthio]acetic acid, formed by addition of the sulfur group of
thioglycolic acid to the double bond of the nitrone by reaction
1 (RSH ) AcSH):
RSNO
9
In the presence of nitrone spin trap, ST, thiyl radical, RS•, is
effectively scavenged with the formation of the radical adduct,
ST/RS•:
kst
9
RS• + ST
8 ST/RS•
(3)
Figure 3 demonstrates the EPR spectra of the adducts of
DMPO (a, c) and DEPMPO (b, d) with GS• and AcS• radicals
generated via photolysis of the GSNO and AcSNO, correspond-
ingly. The spectral parameters of the adducts with GS• radical
(see Figure 3 caption) coincide with the literature data for
DMPO/GS• and DEPMPO/GS• adducts,16,20 while these param-
eters for the adduct with AcS• are described here for the first
time. The reduction of the adduct, ST/RS•, will result in the
formation of the corresponding hydroxylamine, HA-SR. The
same compound, HA-SR, as well as the nitrone, N-SR, are
formed due to a second-order disproportionation reaction:15
The value of the 31P NMR chemical shift of the compound
P is approximately 30 ppm, which is typical for DEPMPO
hydroxilamine derivatives.6,15
In similar experiments using DEPMPO and cysteine, an
appearance of several additional 31P NMR signals were observed
in the region 29-32 ppm, characteristic for DEPMPO hydroxy-
lamine derivatives,6,15 with total integral intensity being directly
proportional to the concentrations of DEPMPO and CySH (data
not shown, see Table 1 for the estimate of equilibrium
constants). However, their quantitative analysis was not possible,
due to low product concentrations. NMR spectra of the
DEPMPO mixture with GSH did not reveal any additional
diamagnetic products. One possible explanation of this fact could
be the lower equilibrium constant for the reaction of DEPMPO
with GSH and, therefore, a lower equilibrium concentration of
the adduct, DEPMPO/GSH, being below the 31P NMR threshold
of detection.
If the reversibility of the reaction 1 is correct, then the
hydroxylamine compounds, HA-SR, should not be stable and
would decompose to the parent nitrone, ST, and thiol, RSH. In
agreement with this hypothesis, the photolysis of the solutions
of 10 mM GSNO or 10 mM AcSNO in the presence of 42.5
mM DEPMPO, with a UV lamp (15 mW/cm2 intensity) at 500-
600 nm over a time interval up to 90 min did not lead to the
appearance of any new 31P NMR signals (data not shown). The
fraction of the decomposition of GSNO in this case was up to
90% (see Materials and Methods). Interestingly, 30 min of laser
photolysis of the same solution at 532 nm, with higher irradiation
EPR and 31P NMR Detection of the Adducts from
Reactions of Nitrone Spin Traps with Thiyl Radicals.
Reversibility of the reaction of nitrones with thiols provides a
unique detoxification mechanism of thiyl radicals by these spin
traps. This possibility is made more likely by the fact that the
main pathway of the decay of the paramagnetic adducts of the
nitrones with RS• in biological systems is their reduction to the
corresponding hydroxylamines. Therefore, following decom-
position of the hydroxylamine back to nitrone and RSH will
result in thiol recycling.

9318 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Potapenko et al.
Figure 4. 31P NMR spectrum of the mixture of 42.5 mM DMPO and
10 mM GSNO in 0.1 M phosphate buffer, pH 7.0, measured after
irradiation by a Nd:YAG laser at 532 nm with intensity 30 mJ/cm2 per
pulse. The laser frequency was 25 Hz, and the irradiation time 30 min.
Number of scans, 256; LB ) 0.3. The estimated concentration of the
product from the integral intensity of the new peaks at 23.80 and 23.76
ppm is equal to 2.2 ( 1 mM.
power of about 750 mW/cm2 (30 mJ/pulse, frequency 25 Hz,
irradiated area 1 cm2) results in the appearance of two new
products with ratio 1:1 at 23.80 and 23.76 ppm and a total
concentration of 2.2 ( 1 mM (Figure 4). These values of
chemical shift are characteristic for nitrone derivatives of
DEPMPO.15 Taking into account that both glutathione and
DEPMPO structures have chiral centers, these products may
be assigned to different diastereoisomers of the expected nitrone,
N-SG. The observation of the signals assigned to the nitrone
compounds, N-SG, exclusively at high irradiation power is
consistent with the prevailing bimolecular decay of the para-
magnetic adduct by reaction 4 at high initial concentration of
ST/RS•. Meanwhile, it also opens a question about the origin
of other pathways of the adduct degradation in our system that
compete with bimolecular decay (4) at lower irradiation power.
Note that kinetic analysis of the nitrone formation upon laser
photolysis (see kinetic analysis section below) predicts nitrone
accumulation of up to about 2.4 mM, which is in a good
agreement with this experimental data.
Figure 5. EPR kinetics decay measured after one, two, three, and four
pulses of laser irradiation of the 0.1 M phosphate buffer, pH 7.0, 5
mM GSNO, and 88.5 mM DMPO (a) or 85 mM DEPMPO (b). The
calculated kinetics is a nonlinear least-squares fit to the eq 6, yielding
parameters kI ) (3.1 ( 0.4) × 10-2 s-1 and kII ) (7.9 ( 2.0) ×102
M-1 s-1 (a) and kI ) (4.1 ( 0.2) × 10-3 s-1 and kII ) (1.0 ( 0.3)
×102 M-1 s-1 (b).
and second-order decays was proposed, namely
d[ST/RS•]
) -kI[ST/RS•] - kII[ST/RS•]2
(5)
dt
where kI and kII are the observed rate constants of first- and
second-order decay. The analytical solution of the eq 5 is given
by eq 6
-1
kII
kI
kII
1
[ST/RS•] )
+
ekIt
-
(6)
•
(
[
)
k
]
[ST/RS ]0
I
EPR Studies of the Kinetics and Mechanism of the Decay
of the Adducts of Nitrones with Thiyl Radicals. Figure 5
shows the kinetics of the decay of the glutathiyl radical with
nitrone spin traps DMPO and DEPMPO. The kinetics was
measured after various numbers of laser pulses of irradiation at
308 nm of the 0.1 M phosphate buffer solution, pH 7.0,
containing 5 mM GSNO and 88.5 mM DMPO or 85 mM
DEPMPO. As seen in Figure 5, the initial signal of the DMPO/
GS• and DEPMPO/GS• adducts is proportional to the number
of laser pulses and completely decays after 150 or 800 s,
correspondingly. The photolysis of CySNO in the presence of
the traps shows a similar qualitative picture, while quantitative
analysis was impossible due to the limited stability of CySNO.
The observed lifetime of DMPO thiyl radical adduct is
somewhat lower than previously observed 17,21. An explanation
of the lifetime differences in various systems will be given after
detailed analysis of the kinetic scheme.
where [ST/RS•]0 is the initial concentration of the radical adduct.
The fitting of the eq 6 to the experimental data provides an
excellent agreement of calculated and experimental kinetics, as
shown by the filled curves in Figure 5. The second-order decay
can be explained by the bimolecular disproportionation described
by reaction 4. The origin of the monomolecular decay is
somewhat more intriguing and was carefully elaborated on in
the present work. The simplest explanation can be done by
hypothesizing the reversibility of reaction 3, i.e., decomposition
of the spin adduct back to parent nitrone and glutathiyl radical.
The further analysis of the whole body of the kinetic studies is
consistent with this hypothesis and is presented in a separate
section. Nevertheless, the reversibility of the basic spin trapping
reaction 3 is never discussed in the literature and requires direct
experimental proof, particularly taking into account its far-
reaching implications.
Fitting DMPO/GS• and DEPMPO/GS• kinetics (Figure 5)
with the first- or second-order decay curves does not provide
satisfactory agreement with the experimental kinetics (results
of this analysis not shown). Therefore, a combination of first-
Direct Experimental Proof of the Reversibility of Spin
Trapping of Glutathiyl Radicals by Nitrone Spin Traps. 1.
The Formation of DEPMPO/GS• Adduct during Decomposition
of DMPO/GS•. The reversibility of the spin trapping reaction 3

Reaction of Thiols and Thiyl Radicals with Nitrones
J. Phys. Chem. B, Vol. 108, No. 26, 2004 9319
Figure 6. (a) EPR spectrum obtained during laser irradiation of the
solution of DMPO and GSSG. First, 0.1 M GSSG and 88.5 mM DMPO
were added to 0.1 M phosphate buffer. Then the solution was titrated
to pH 7.0, bubbled by Ar during 15 min, and transferred to the EPR
cell. The spectrum has been measured during laser irradiation at 308
nm. Number of scans, four. (b) A 2 mL sample of the same mixture as
in part a was placed into the 1 cm quartz cell and irradiated by laser
light with frequency 10 Hz during 20 s (∼30 mJ per pulse). Immediately
after that, 85 mM DEPMPO was added and the solution was transferred
to the EPR flat cell. The spectrum was measured about 2 min after
DEPMPO addition. The spectrometer settings were as follows: sweep
width, 150 G; microwave power, 5 mW; modulation amplitude, 1 G;
time constant, 81.92 ms; sweep time, 40 s; number of scans, four. (c).
Simulated EPR spectrum with parameters aP ) 45.81 G, aN ) 14.14
G, and aH ) 14.94 G.
Figure 7. EPR spectra of the solution of DMPO, GSNO, and
HCO2Na in phosphate buffer measured after laser irradiation. First, 2
M HCO2Na was added to 0.1 M phosphate buffer and the solution
was titrated to pH 7.0. After that 5 mM GSNO and 88.5 mM DMPO
were added. The spectrometer settings were as follows: sweep width,
80 G; modulation amplitude, 1 G; time constant, 81.92 ms; microwave
power, 5.19 mW, sweep time, 41.94 s. (a) Spectrum measured
immediately after laser irradiation (four pulses × 30 mJ), one scan; an
arrow indicates the moment of irradiation; the peaks of DMPO/GS•
adduct are marked by an asterisk. (b) Spectrum measured 2 min later,
eight scans. The hyperfine constants determined by simulation of the
spectrum are aN ) 15.7 G and aH ) 19.0 G, in good agreement with
proposes release of glutathiyl radicals from the DMPO/GS•
adduct. Therefore, subsequent addition of the DEPMPO spin
trap should result in the appearance of the DEPMPO/GS• adduct.
To test this hypothesis, we generated GS• by the photolysis of
disulfide GSSG. Figure 6a shows an EPR spectrum registered
during laser photolysis of 0.1 M GSSG in the presence of 88.5
mM DMPO, which was assigned to DMPO/GS• adduct based
on the values of its hyperfine constants, aN ) 15.12 G and aH
) 16.11 G. In parallel experiments after irradiation was stopped,
42.5 mM DEPMPO was added to the solution, and the EPR
spectrum was measured afterward (Figure 6b) with the param-
eters characteristic for DEPMPO/GS• adduct (aP ) 45.81 G,
aN ) 14.14 G, aH ) 14.94 G). The formation of DEPMPO/
GS• adduct after addition of DEPMPO to DMPO/GS• indicates
the presence of GS• radical in the solution, although GS• radicals
are fully scavenged by 88.5 mM DMPO spin trap (rate constant,
2.6 × 108 M-1 s-1 22) in a less than 100 ns. Taking into account
that DEPMPO/GS• was not formed in the absence of photolysis
(data not shown), it should be concluded that GS• radical is
generated directly from DMPO/GS• radical adduct.
published data for the DMPO/CO2 adduct.24
•-
Figure 8. Time evolution of the DMPO spin adducts concentration,
[SA], measured from EPR signal intensity changes after GSNO
photolysis in the presence of sodium formate. First 2 M HCO2Na was
added to the 0.1 M phosphate buffer, and the solution was titrated to
pH 7.0. After that 5 mM GSNO and 88.5 mM DMPO were added.
The mixture was bubbled by Ar during 30 min and transferred to the
EPR cavity. The kinetics of DMPO/GS• and DMPO/CO2•- adducts were
registered using corresponding high field spectral components (see
Figure 7). The calculated kinetics of DMPO/GS• decay is a nonlinear
least-squares fit to eq 6, yielding parameters kI ) (3.2 ( 0.4) × 10-2
2. The Formation of CO2 Radical from DMPO/GS• in the
•-
Precence of Formate. Glutathiyl radical is able to abstract the
hydrogen atom from formate ion with the rate constant k7 ) 8
s-1 and kII ) (1.2 ( 0.3) × 103 M-1 s-1
.
× 105M-1 s-1 23
:
•-
stable DMPO/CO2 radical adduct. In agreement with this
hypothesis, the photolysis of 5 mM GSNO in Ar-bubbled 0.1
M phosphate buffer in the presence of 88.5 mM DMPO and 2
M sodium formate (Figures 7 and 8) resulted in the decay of
DMPO/GS• adduct with corresponding accumulation of DMPO/
CO2•- adduct. EPR spectrum shown in Figure 7a was registered
immediately after laser irradiation (four pulses × 30 mJ) and
k7
GS• + HCO2- 98GSH + CO2
(7)
•-
On the basis of reaction 7, release of glutathiyl radicals from
the DMPO/GS• adduct in the presence of formate, HCOO-,
•-
should lead to the formation of the carbonyl radical, CO2
,
and, as a consequence, to the accumulation of comparatively

9320 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Potapenko et al.
consists of signals from the both radical adducts. The DMPO/
GS• adduct completely disappeared within 2 min with parallel
accumulation of the EPR signal of DMPO/CO2•- adduct (Figure
7b). Figure 8 demonstrates the experimental kinetics measured
immediately after laser irradiation. It should be noted that under
the experimental concentrations of the spin trap and formate,
GS• radical initially formed under photolysis is predominantly
scavenged by DMPO (rate constant 2.6 × 108 M-1 s-1 22) rather
than reacting with formate by reaction 7 (k7 ) 8 × 105 M-1
s-1 23), in agreement with experimental data (Figure 7a). The
release of GS• radical from DMPO/GS• adduct results in
establishment of quasistationary level of GS•, which reacts with
formate, forming carbonyl radical, with subsequent accumulation
of DMPO/CO2•- adduct (Figure 7b and 8). Fitting the kinetics
of the DMPO/GS• decay shown in Figure 8 allows good
agreement between calculated and experimental dependences,
with the rate constant of monomolecular decomposition being
equal to 0.03 s-1 (see next section for details).
Figure 9. Dependence of the kI value on the term k11[GSNO]/(k9a-
[DMPO] + k11[GSNO]). The experiments were done in air-saturated
0.1 M phosphate buffer, pH 7.0, containing GSNO and DMPO. Kinetic
decay was registered after irradiation of the solution by one, two, three,
and four laser pulses. Obtained kinetics with different radical adduct
concentrations (as in Figure 5) were fitted using eq 6 with fitting
parameters kI and kII. (O) Variation of GSNO concentration at [DMPO]
) 88.5 mM; (4) variation of DMPO concentration at [GSNO] ) 5
mM. Note that the value of k11 ) (6.1 ( 1.5) ×108 M-1 s-1 measured
in this work (see laser flash photolysis section) was used in calculations.
Kinetic Analysis of Laser Photolysis of GSNO Solution
in the Presence of the Nitrones. 1. Kinetic Scheme and Decay
of ST/GS• Adduct. To quantitatively describe the kinetics of the
decay of the nitrone adducts with glutathiyl radical, ST/GS•,
formed due to photolysis of GSNO (Figure 5), the following
reactions should be considered:
decay measured upon GSNO photolysis in air-saturated and
argon-bubbled solutions (data not shown, see Materials and
Methods) is consistent with the negligible role of reaction 14
upon experimentally used high concentrations of the spin trap.
The performed numerical calculations showed that the contribu-
tion of the reactions 12 and 13, being of second-order on the
concentration of generated radical species, is negligible even
for diffusion-controlled reactions and may become significant
only at bimolecular rate constants being larger than 1012 M-1
s-1. In agreement with this statement, NO significantly decreased
the lifetime of the radical adduct only at concentrations one
order larger compared with those formed in our experiments
(data not shown).
GSNO
GS• + ST T ST/GS•
9
hν8 GS• + NO•
(8)
(9)
GS• + DMPO T DMPO/GS• k9a ) 2.6 × 108 M-1 s-1 22
k-9a ) 0.3 s-1 (this work) (9a)
GS• + DEPMPO T DEPMPO/GS•
k9b ) 1.5 × 108 M-1 s-1
The consideration of reactions 9-11 using the approximation
of the quasistationary concentration for [GS•] gives the following
solution for spin adduct decay:
k-9b ) 0.02 s-1 (this work) (9b)
2(ST/GS•) f bimolecular decay
2(DMPO/GS•) f bimolecular decay
k10a ) 4 × 102 M-1 s-1 (this work) (10a)
2(DEPMPO/GS•) f bimolecular decay
(10)
d[ST/GS•]
k11[GSNO]
-9k9[ST] + k11[GSNO]
) -k
[ST/GS•] -
dt
2k10[ST/GS•]2 (15)
Equation 15 coinsides with eq 5, which describes the experi-
mental kinetics, assuming
k10b ) 0.5 × 102 M-1 s-1 (this work) (10b)
[ST/RS•] ) [ST/GS•], kI )
k11[GSNO]
GS• + GSNO f GSSG + NO• k11 ) 1.7 × 109 M-1 s-1 18
k11 ) (6.1 ( 1.5) × 108 M-1 s-1 (this work) (11)
k
, and kII ) 2k10 (16)
-9k9[ST] + k11[GSNO]
2GS• f GSSG k12 ) 1.5 × 109 M-1 s-1 23
GS• + NO• f GSNO ref 20
(12)
(13)
where ki and k-i are the rate constant of forward and reverse
reaction i, correspondingly.
Figure 9 demonstrates that the dependence of kI on the term
k11[GSNO]/(k9a[DMPO] + k11[GSNO]) for DMPO allows
excellent linear approximation, namely
GS• + O2 f GSOO• k15 ) 2 × 109 M-1 s-1 25 (14)
Reaction 9 is shown as reversible to account for the experi-
mentally observed GS• radical release from the ST/GS• adduct.
This release results in establishment of a quasistationary
concentration of GS• radical, which decays also in reactions
11-14. Among the latter reactions, only reaction 11 can
significantly contribute to GS• decay by mechanisms other than
spin-trapping reaction under our experimental conditions. The
insignificant difference between the kinetics of DMPO/GS•
k11[GSNO]
kI(s-1) ) (0.30 ( 0.01)
( 0.0027
k9a[DMPO] + k11[GSNO]
(17)
Therefore, according to eqs 16 and 17 the rate constant for the
decomposition of the DMPO/GS• adduct, k-9a, is equal to (0.30
( 0.01) s-1. As for bimolecular decay, we did not observe any
dependence of the observed rate constant kII on the concentra-

Reaction of Thiols and Thiyl Radicals with Nitrones
J. Phys. Chem. B, Vol. 108, No. 26, 2004 9321
tions of [GSNO] and [DMPO], in agreement with eq 16. The
The analytical solution of this equation can be obtained after
substitution of the [DMPO/GS•] (eq 6 with parameters in eq
19), namely
k
10a value was found to be equal to (4.0 ( 1.0) × 102 M-1 s-1
.
As we mentioned above, the degradation of the DEPMPO/
GS• adduct (Figure 5b) also is described by first- and second-
order decay with kI ) (4.1 ( 0.2) × 10-3 s-1 and kII ) (1.0 (
0.3) × 102 M-1 s-1 (see Figure 5 caption). Applying similar
arguments to the analysis of these kinetics, based on the
reactions 9-11, and therefore using eq 16, we obtain the rate
constant for the monomolecular decomposition of DEPMPO/
GS• adduct, k-9b ) (2.1 ( 0.1) × 10-2 s-1, and the rate constant
for its bimolecular decay, k10b ) (50 ( 15) M-1 s-1. The
calculation of the value of k-9b requires knowledge of the rate
constant k9b for the direct reaction of GS• trapping by DEPMPO.
In the present paper we have measured this constant by laser
flash photolysis and found it to be equal to 1.5 × 108 M-1 s-1
(see laser flash photolysis section).
kIII
kII
kII
kI
[DMPO/CO2•-] )
-kI + ln
[DMPO/GS•]0 + 1 ×
(
((
)
kII
exp(kIt) - [DMPO/GS•]0 (21)
))
kI
where [DMPO/GS•]0 is the initial concentration of the DMPO/
GS• adduct. Equation 21 qualitatively describes the experimen-
tally observed accumulation of DMPO/CO2 adduct. The
following expression can be obtained for the maximal value of
the concentration of the DMPO/CO2 adduct that can be
•-
•-
accumulated in the absence of its decay:
•-
2. Kinetics of Accumulation of DMPO/CO2 Adduct in the
k
kII
•-
Presence of Formate. In the presence of formate, an additional
pathway of GS• decay by reaction 7 must be considered. Again,
using a quasistationary approximation for the concentration of
GS• radical, we obtain the following equations for DMPO/GS•
adduct decay:
[DMPO/CO2
]
)
III ln [DMPO/GS•]0 + 1
(22)
max
(
)
kII
kI
Using the obtained values of k-9a and k10a and published values
of the rate constants k7, k9a, and k11 we estimated expression 22
to be equal to 1.1 × 10-5 M, in reasonable agreement with the
experimental value 0.72 × 10-5 M. Note that quantitative
d[DMPO/GS•]
•-
description of the kinetics of accumulation of DMPO/CO2
) -
dt
adduct requires knowledge of the mechanism of its decay, which
was not a subject of this study.
Therefore, the kinetic analysis of the experimental data
support the existence of monomolecular decomposition of the
adduct of nitrones with glutathiyl radicals, ST/GS•, with
recycling of the parent nitrone and GS• radical.
-
k11[GSNO] + k7[HCO2 ]
k-9a
[DMPO/GS•] -
-
(
)
k9a[ST] + k11[GSNO] + k7[HCO2 ]
2k10a[DMPO/GS•]2 (18)
3. Accumulation of DEPMPO-SG Nitrone upon GSNO
Photolysis. Bimolecular disproportionation of the DEPMPO/
GS• adduct results in formation of the nitrone DEPMPO-SG,
according to reaction 4. The efficiency of this process is
determined by the ratio of radical adduct degradation in the first-
and second-order reactions, namely by the factor, kII[DEPMPO/
GS•]/kI (see eqs 5 and 16), and therefore by the rate, I, of GS•
generation upon continuous photolysis. The value of I that will
result in equal efficiency of mono- and bimolecular decay was
estimated as 4 µmol/L/s for the experimentally used concentra-
tions of DEPMPO and GSNO (42.5 and 10 mM, correspond-
ingly), assuming steady-state approximation for [DEPMPO/GS•]
and [GS•] and using the rate constants obtained in the previous
sections. The initial value of I in the case of UV lamp photolysis
was estimated to be significantly lower, about 0.2 µmol/L/s
(average extinction coefficient, ꢀGSNO ≈ 7 M-1 s-1, in the range
500-600 nm, and quantum yield about 0.02617). In the case of
laser photolysis (see Materials and Methods), this value was
estimated to be about 6.7 µmol/L/s, which should result in 2.4
mM accumulation of the nitrone DEPMPO-SG, in fairly good
agreement with the experimentally observed value (Figure 4).
Laser Flash Photolysis Studies of the Reaction of GS•
Radical with DEPMPO and GSNO. The rate constants of GS•
addition to GSNO, k11, and to DEPMPO, k9b, were calculated
from the measurements of the decay of GS• radical monitored
at 350 nm using the laser flash photolysis setup described in
the Materials and Methods. The absorption signal decayed
exponentially upon the photolysis of 0.4-1.6 mM GSNO
solutions. The observed pseudo-first-order rate constant linearly
depends on the initial GSNO concentration, in agreement with
Equation 18 formally coincides with the eq 5. Therefore, the
decay of DMPO/GS• adduct is described by eq 6 with the
parameters
[ST/RS•] ) [DMPO/GS•], kI )
-
k11[GSNO] + k7[HCO2 ]
k-9a
, and kII ) 2k10a
-
k9a[ST] + k11[GSNO] + k7[HCO2 ]
(19)
A good agreement between calculated and experimental kinetics
(Figure 8) has been observed for the decay of DMPO/GS• adduct
using known values of the rate constants and yielding fitting
parameters k-9a ) (0.2 ( 0.05) s-1 and k10a ) (600 ( 150)
M-1 s-1. These values of k-9a and k10a are in a reasonable
agreement with those obtained in the absence of formate (Figure
5a), particularly taking into account some uncertainty in the
16,23
value of k7
and appearance of an additional pathway of
bimolecular decomposition between DMPO/GS• and DMPO/
CO2 adducts.
An accumulation of DMPO/CO2 adduct, in its turn, is
•-
•-
described by the following equation upon the assumptions that
(i) CO2•- radical generated by reaction 7 is completely trapped
•-
by DMPO and (ii) the formed DMPO/CO2 adduct insignifi-
cantly decays for the experimental time:
•-
d[DMPO/CO2
]
)
dt
-
k7[HCO2 ]
k9a[ST] + k11[GSNO] + k7[HCO2 ]
kIII[DMPO/GS•] (20)
reaction 11, giving k11 ) (6.1 ( 1.5) × 108 M-1 s-1
.
k-9a
[DMPO/GS•] )
-
The measurement of the rate constant k9b of the reaction of
GS• with DEPMPO was performed by irradiation of 1 mM
GSNO solution in the presence of DEPMPO; the concentration

9322 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Potapenko et al.
of DEPMPO varied from 2.84 to 11.35 mM. The absorption
signal decayed exponentially, and the value k9b was calculated
from the obtained linear dependence of the observed pseudo-
first-order rate constant on the term k9b[DEPMPO] +
K
NuH, Kad, and Kb of the reactions 23-25 and by the equilibrium
concentrations of the NuH at given pH, namely
[ST(H)Nu]/[ST] ) KNuHKadKb[NuH]
(26)
k11[GSNO], yielding k9b ) (1.5 ( 0.3) × 108 M-1 s-1
.
Table 1 represents several S-centered nucleophiles, their pK )
-log KNuH values, and estimates of the product Kp ) KNuHKadKb.
The strong pH dependence of the equilibrium hydroxylamine
concentration has been observed for (bi)sulfite addition toward
DMPO and DEPMPO, in agreement with eqs 23 and 26.5,6 For
the other nucleophiles shown in Table 1, the pH dependence is
leveled at neutral pH (pKNuH . pH, resulting in [NuH] .
[Nu-]). A thousand times difference in Kp value for the reaction
of DEPMPO with nucleophiles of the same charge, SO32- and
-SCH2COO-, correlates with the difference in their pKNuH (∆pK
≈ 3). Also the lower value of Kp for the reaction of DEPMPO
with GSH (undetectable in our experiments) compared with that
for CySH may be related to higher pK value of GSH.
The further oxidation of the hydroxylamines results in the
formation of the corresponding radical adduct with sulfite radical
anion or with the thiyl radicals, exemplifying nonconventional
formation of the paramagnetic adduct by the Forrester-
Hepburn32 mechanism. The hydroxylamines have low redox
potentials, their anodic half-wave redox potentials, E1/2, being
in the range of -(0.4-0.5) V vs SCE in aqueous solution for
a series of alkyl and arylhydroxylamines in their anionic form
and 0.5-0.8 V in acetonitrile in their neutral form.39,42-44
Therefore, they can be oxidized even by weak oxidants such as
dioxygen (E° (O2/O2•-)) -0.4 V in water) or by a wide range
of redox proteins in biological systems. Taking into account
Discussion
The chemistry of nitrones is of significant interest, due to
their wide applications as spin traps for EPR detection of short-
lived radical intermediates in biological systems1,2 and growing
applications as therapeutic agents.3,4 An observation of the new
reactions of nitrones with (bi)sulfite and thiols may be of critical
importance for the understanding of their biological activities.
Indeed, the reverse reaction of nucleophilic addition of thiols
to the nitrones provides the unique mechanism of the detoxi-
fication of thiyl radicals. Thiols, being primary antioxidants in
vivo, form thiyl radicals during their antioxidant activity. In
turn, nitrones may recycle thiol antioxidants by trapping the
thiyl radicals and releasing the thiols after the reduction of the
paramagnetic adduct by biologically relevant reductants, such
as ascorbate. The knowledge of these new reactions of nitrones
can assist investigators in avoiding erroneous interpretation of
spin-trapping experiments. The reversible nucleophilic addition
of both thiols and thiyl radicals to the nitrones may significantly
affect not only the quantitative analysis of the data but even
the conclusion about the origin of the paramagnetic adduct.
The potential for nucleophilic addition to nitrones has been
known for a long time.26 For example, Bonnet at al.27 described
the irreversible addition of C-centered anions (Grignard reagents,
hydrogen cyanide) to a variety of nitrones with formation of
isolatable hydroxylamine derivatives. Nucleophilic addition
reactions of O-centered anions (-OH, -OCH3)28,29 to DMPO
and of the anion of H2O2 to 4-POBN30 are also described. Alberti
et al.31 showed the nucleophilic addition of a range of
heterocyclic N-H bases toward PBN and DMPO. Forrester and
Hepburn32 discussed the nucleophilic addition of a numbers of
anions (of nitromethane, phthalamide, hydrogen and benzyl
cyanide, and acetate) to PBN with formation of the hydroxy-
lamine derivatives followed by their oxidation to EPR-detectable
nitroxides. This artificial pathway of paramagnetic adduct
formation was named after this work, i.e., the Forrester-
Hepburn mechanism, and must be a matter of great concern in
interpretation of spin-trapping experiments. There are numerous
reports on the involvement of the Forrester-Hepburn mecha-
nism in the formation of EPR-detectable adducts of spin
traps33-38 (see also refs 39 and 40 for reviews).
that redox potentials of sulfite and thiols are normally in the
•-
same range or higher (0.4 V for SO3•-/SO32-, 0.6 V for SO3
/
- 45
HSO3
,
and 0.7 V for CyS•/CySH46), one has to unambigu-
ously consider the possibility of paramagnetic adduct formation
from nitrones in the presence of thiols and oxidants via the
Forrester-Hepburn mechanism, i.e., reactions 23-25 followed
by hydroxylamine oxidation. The oxidation of hydroxylamine
in the presence of oxygen will result in superoxide generation,
followed by its own chemistry, therefore adding complexity to
the system. Note that equilibrium concentrations of hydroxy-
lamine formed are sufficient to explain a significant rate of
accumulation of the paramagnetic adduct in the presence of the
spin trap, nucleophile, and oxidant. At 1 mM concentration of
nucleophile and 50 mM of spin trap, it corresponds, according
to eq 26 with Kp given in Table 1, to concentrations of
hydroxylamine derivative ST(H)Nu in amounts about 0.9 mM
(NuH ) HSO3-), 7 µM (NuH ) HSCH2COO-), 1.5 µM (NuH
) CySH), and <0.1 µM (NuH ) GSH).
If nucleophilic addition to nitrones, in our opinion, were not
sufficiently taken into consideration, particularly in spin-trapping
experiments, there would be even less noticed about its
reversibility. Meanwhile, the reverse reaction is worth consider-
ing for a variety of reasons. First, it shows that the product of
nucleophilic addition to the nitrones may be important in a
reaction mixture while it would not be isolated as an individual
substance. The drastic example is sulfite addition to nitrones
with mM (!) product formation. Second, the reversibility of
reaction 24 provides a unique mechanism for the detoxification
of reactive radicals, Nu•, by nitrones:
In a common case, addition of negatively charged nucleophile,
Nu-, to a nitrone, ST, can be described by the following
generalized equations:
NuH 7K 8 Nu- + H+
Nu- + ST 7Kad8 ST- Nu
(23)
(24)
(25)
NuH
Kb
ST- Nu + H+ 798 ST(H)Nu
where ST(H)Nu and ST- Nu are the corresponding hydroxy-
lamine and its deprotonated form. The acidity of hydrohylamines
in water is supposed to be similar to that for aliphatic alcohols,
pKa ≈ 16-18; e.g. the acidity of N-phenylhydroxylamine, PBN,
is similar to that for isopropyl alcohol, pK ) 17.1.41 Therefore,
at neutral pH at equilibrium, reaction 25 is strongly shifted to
the right, and the ratio of finally formed hydroxylamine to
initially available nitrone is determined by equilibrium constants
Nu• + ST
9
8 ST•Nu
(27)
ST•Nu Reduction8 ST(H)Nu
(28)
(29)
ST(H)Nu a ST + NuH

Reaction of Thiols and Thiyl Radicals with Nitrones
J. Phys. Chem. B, Vol. 108, No. 26, 2004 9323
This mechanism seems to be particularly important for thiols,
NuH ) RSH, widely considered as a primary antioxidant.
Reaction 29, being thermodynamically equivalent to the sum
of reactions 24 and 25, essentially explains our failure to
accumulate ST(H)-SG adduct, even at very high rates of GS•
generation by laser photolysis of nitrosoglutathione. The only
adduct we were able to detect was assigned to DEPMPO-SG
nitrone on the basis of the 31P NMR shift. Interestingly, also
only one diamagnetic product of degradation of the DMPO/
GS• adduct was measured using HPLC by Stoyanovsky et al.47
and assigned to the DMPO-SG nitrone. These data are in
agreement with our report on reverse reaction 29 for DEPMPO/
OH and DEPMPO/OOH hydroxylamines that prevented us from
accumulating these diamagnetic products upon generation of
reaction (7) of formate with GS•. The simple kinetic estimates
show that in the presence of 20 mM DEPMPO and 400 mM
formate, the latter reaction can hardly compete with GS•
scavenging by DEPMPO (the ratio k7[HCOO-]/k9b[DEPMPO]
≈ 1.3 × 10-3 if k7 ) 104 M-1 s-1 16
or it is ≈0.1 if k7 ) 8 ×
,
105 M-1 s-1 23). These data are in agreement with our
observation (Figure 8) and can be easily explained by establish-
ment of a quasistationary concentration of GS• due to the
reversibility of reaction 9. The reversibility of the spin-trapping
reactions with the radicals beyond GS•, if they exist, may
significantly influence the lifetime of the radical adducts and
even facilitate the formation of new products. The probable role
of the reverse spin trapping reaction in degradation of nitrone
adducts with peroxyl and superoxide radicals is worth consid-
eration. For example, the conversion of DMPO/-OOH to
DMPO/-OH54 with formation of -OH• radical,55 the mecha-
nism for which has been debated for many years, and the
accumulation of carbon-centered adducts of DEPMPO upon
superoxide generation51 may be well related to this intrinsic
instability of the superoxide adduct. Note also the difference
of about 15 times in the lifetimes of superoxide adducts with
DEPMPO and DMPO49-51,56 This is in striking coincidence with
15 times lower rate constant of the reverse reaction 9 for
DEPMPO/GS• adduct compared with that for DMPO/GS•.
Moreover, the theoretically calculated energy difference between
spin trap, ST, and its adduct with superoxide, ST/O2•-, for the
series of nitrones correlates well with the lifetime of the
adducts,50 further supporting a possible contribution of their
breakdown through the reverse spin-trapping reaction.
•-
OH• and O2 radicals.15,48
The attempts to accumulate hydroxylamine derivatives of
DEPMPO adduct during photochemical generation of RS•
encouraged us to look into the details of the DEPMPO/GS• and
DMPO/GS• decays. The kinetics showed superposition of first-
and second-order decays that is quite typical for the spin adduct
degradation.49,50 Normally second-order decay is observed only
at high concentrations of generated adduct in comparatively pure
chemical systems,15,49 while first-order decay dominates at low
adduct concentrations in various biological systems.51 The
mechanism of second-order decay for most cases is believed to
be bimolecular disproportionation, as shown by eq 4 for thiyl
radical adduct. The mechanisms of first-order decay vary
depending on the sample composition and the presence of
reducing agents. Nevertheless, it is a recognized intrinsic
instability of the spin adducts of nitrones with the most
biologically relevant radicals, e.g. superoxide, thiyl, and peroxyl
radicals. The mechanisms for the self-degradation of these
adducts may include several pathways but are far from
understood. In the present work we have demonstrated degrada-
tion of the glutathiyl radical adducts with the nitrones DMPO
and DEPMPO via the reverse spin-trapping reaction 9. The
release of GS• radical from the DMPO/GS• adduct has been
unambiguously proved by its further competitive reactions with
another nitrone or with formate (Figures 6-8). The proposed
kinetic scheme (reactions 7-11) allows quantitative description
of all the observed kinetics. The reciprocal proportionality of
the pseudo-first-order decay rate constant, kI, to the concentration
In conclusion, we provide strong evidence for the importance
of the reversible reaction of thiols and thiyl radicals with nitrone
spin traps. We propose that the observations described here
represent a fundamental chemical behavior that applies to a
variety of redox reactions with nitrone spin traps that must be
considered relevant.
Acknowledgment. This work was supported by grants from
NHLB 53333, the Russian Foundation for Basic Research grants
02-04-48374 and 03-04-06189, and Ministry of Education of
RF grant A03-2.11-822.
References and Notes
(1) Mason, R. P.; Hanna, P. M.; Burkitt, M. J.; Kadiiska, M. B. EnViron.
Health Perspect. 1994, 102, 33-36.
of spin trap and its direct proportionality to the concentration
-
of the competitive reagent for GS• (GSNO, eq 16, or HCO2
,
(2) Khan, N.; Swartz, H. Mol. Cell. Biochem. 2002, 234-235, 341-
357.
eq 18) are in excellent agreement with this scheme (see Figure
9). The rate constant of the GS• release from DMPO and
(3) Floyd, R. A.; Hensley, K.; Forster, M. J.; Kelleher-Andersson, J.
A.; Wood, P. L. Mech. Aging DeV. 2002, 123, 1021-1031.
(4) Kotake, Y. Antioxid. Redox Signal. 1999, 1, 481-499.
(5) Potapenko, D. I.; Bagryanskaya, E. G.; Reznikov, V. V.; Clanton,
T. L.; Khramtsov, V. V. Magn. Reson. Chem. 2003, 41, 603-608.
(6) Potapenko, D. I.; Clanton, T. L.; Bagryanskaya, E. G.; Gritsan, N.
P.; Reznikov, V. A.; Khramtsov, V. V. Free Radic. Biol. Med. 2003, 34,
196-206.
(7) Bryson, P. D. Sulfites and monosodium glutamate. In Comprehen-
siVe ReView in Toxicology for Emergency Clinicians; Taylor & Francis:
Washington, D.C., 1996; pp 697-703.
(8) Bush, R. K.; Taylor, S. L.; Busse, W. J. Allergy Clin. Immunol.
1986, 78, 191-202.
(9) Gunnison, A. F.; Jacobsen, D. W. Sulfite hypersensitivity. A critical
review. In CRC Critical ReView in Toxicology; CRC Press: Boca Raton,
FL, 1987; pp 185-214.
(10) Pietraforte, D.; Mallozzi, C.; Scorza, G.; Minetti, M. Biochemistry
1995, 34, 7177-7185.
(11) Hart, T. W. Tetrahedron Lett. 1985, 26, 2013-2016.
(12) Glasoe, P. K.; Long, F. A. J. Phys. Chem. 1960, 64, 188-190.
(13) Molokov, I. F.; Tsentalovich, Yu. P.; Yurkovskaya, A. V.; Sagdeev,
R. Z. Photochem. Photobiol. A: Chem. 1997, 110, 159-165.
(14) Tsentalovich, Yu. P.; Kulik, L. V.; Gritsan, N. P.; Yurkovskaya,
A. V. J. Phys. Chem. A. 1998, 102, 7975-7980.
DEPMPO, k-9, was found to be equal to 0.3 and 0.02 s-1
,
correspondingly. Note that the characteristic lifetime of the
adducts strongly depends on the competitive reagent (1/τ ) kI,
eqs 16 and 18), explaining its significant variations in the
different experimental conditions (refs 17 and 21 and this work).
On the other hand, the ratio of the constants, showing about 15
times greater stability of the DEPMPO/GS• adduct compared
with that for DMPO/GS• toward this mechanism of degradation,
reflects intrinsic properties of the trap and does not depend on
the sample composition.
The reversibility of spin trapping reaction, to our knowledge,
has never been discussed in the literature. Nevertheless, the
analogous breakdown of the series of 3-imidazoline-3-oxide
nitroxides with formation of dinitrone 4H-imidazole di-N-oxides
was previously described.52,53 In another report, Karoui et al.16
explained a time-dependent decrease in DEPMPO/GS• adduct
•
with a parallel increase in DEPMPO/CO2 adduct upon the
•-
photolysis of GSNO, by the formation of CO2 radical in the

9324 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Potapenko et al.
(15) Khramtsov, V.; Berliner, L. J.; Clanton, T. L. Magn. Reson. Med.
1999, 42, 228-234.
(36) Eberson, L.; McCullough, J. J. J. Chem. Soc., Perkin Trans. 2 1998,
49-58.
(16) Karoui, H.; Hogg, N.; Frejaville, C.; Tordo, P.; Kalyanaraman, B.
J. Biol. Chem. 1996, 271, 6000-6009.
(37) Eberson, L.; Persson, O. J. Chem. Soc., Perkin Trans. 2 1997,
1689-1696.
(17) Singh, R. J.; Hogg, N.; Joseph, J.; Kalyanaraman, B. FEBS Lett.
1995, 360, 47-51.
(38) Sang, H.; Janzen, E. G.; Lewis, B. H. J. Org. Chem. 1996, 61,
2358-2363.
(18) Wood, P. D.; Mutus, B.; Redmond, R. D. Photochem. Photobiol.
1996, 64, 518-524.
(39) Eberson, L. Spin trapping: Problems and artifacts. In Toxicology
of the Human EnVironment. The Critical Role of Free Radicals; Rhodes,
C. J., Ed.; Taylor and Francis: New York, 2000; pp 25-47.
(40) Eberson, L. AdV. Phys. Org. Chem. 1998, 31, 91-141.
(41) Bordwell, F. G.; Liu, W.-Z. J. Am. Chem. Soc. 2003, 118, 8777-
8781.
(19) Barrett, J.; Debenham, D. F.; Glauser, J. Chem. Commun. 1965,
248-249.
(20) Pou, S.; Rosen, G. M. J. Chem. Soc., Perkin Trans. 2 1998, 1507-
1512.
(21) Shi, X.; Dong, Z.; Dalal, N. S.; Gannett, P. M. Biochim. Biophys.
Acta 1994, 1226, 65-72.
(42) Iversen, P. E.; Lund, H. Acta Chem. Scand. 1965, 19, 2303-2308.
(43) Sayo, H.; Ozaki, S.; Masui, M. Chem. Pharm. Bull. (Tokyo) 1973,
21, 1988-1995.
(44) Ozaki, S.; Masui, M. Chem. Pharm. Bull. (Tokyo) 1978, 26, 1364-
1369.
(22) Davies, M. J.; Forni, L. G.; Shuter, S. L. Chem.-Biol. Interact. 1987,
61, 177-188.
(23) Wardman, P. Reactions of Thiyl Radicals. In Biothiols in Health
and Disease; Packer L, Cadenas E, Eds.; Marcel Dekker: New York, Basel,
Hong Kong, 1995; pp 1-19.
(45) Huie, R. E.; Neta, P. J. Phys. Chem. 1984, 88, 5665-5669.
(46) Buettner, G. R. Arch. Biochem. Biophys. 1993, 300, 535-543.
(47) Stoyanovsky, D. A.; Goldman, R.; Jonnalagadda, S. S.; Day, B.
W.; Claycamp, H. G.; Kagan, V. E. Arch. Biochem. Biophys. 1996, 330,
3-11.
(48) Khramtsov, V. V.; Reznikov, V. A.; Berliner, L. J.; Litkin, A. K.;
Grigor’ev, I. A.; Clanton, T. L. Free Radic. Biol. Med. 2001, 30, 1099-
1107.
(49) Tuccio, B.; Lauricella, R.; Frejaville, C.; Bouteiller, J.-C.; Tordo,
P. J. Chem. Soc., Perkin Trans. 2 1995, 295-298.
(50) Villamena, F. A.; Zweier, J. L. J. Chem. Soc., Perkin Trans. 2 2002,
1340-1344.
(51) Roubaud, V.; Sankarapandi, S.; Kuppusamy, P.; Tordo, P.; Zweier,
J. L. Anal. Biochem. 1997, 247, 404-411.
(24) Mottley, C.; Mason, R. P. Arch. Biochem. Biophys. 1988, 267, 681-
689.
(25) Tamba, M.; Simone, G.; Quintiliani, M. J. Radiat. Biol. 1986, 150,
595-600.
(26) Breuer, E. Nitrones and nitronic acid derivatives: Their structure
and their role in synthesis. In Nitrones, Nitronates and Nitroxides; Patai S,
Rappoport Z, Eds.; John Wiley & Sons: New York, 1989; pp 139-244.
(27) Bonnet, R.; Brown, R. F. C.; Clark, V. M.; Sutherland, I. O.; Todd,
A. J. Chem. Soc. 1959, 2094-2102.
(28) Makino, K.; Hagiwara, T.; Hagi, A.; Nishi, M.; Murakami, A.
Biochem. Biophys. Res. Commun. 1990, 172, 1073-1080.
(29) Hanna, P. M.; Chamulitrat, W.; Mason, R. P. Arch. Biochem.
Biophys. 1992, 296, 640-644.
(52) Khramtsov, V. V.; Weiner, L. M.; Gogolev, A. Z.; Grigor’ev, I.
A.; Starichenko, V. F.; Volodarsky, L. B. Magn. Reson. Chem. 1986, 24,
199-207.
(30) Janzen, E. G.; Wang, Y. Y.; Shetty, R. V. J. Am. Chem. Soc. 1978,
100, 2923-2925.
(31) Alberti, A.; Carloni, P.; Eberson, L.; Greci, L.; Stipa, P. J. Chem.
Soc., Perkin Trans. 2 1997, 887-892.
(53) Grigor’ev, I. A.; Shchukin, G. I.; Khramtsov, V. V.; Vainer, L.
M.; Starichenko, V. F.; Volodarsky, L. B. Bull. Acad. Sci. USSR, DiV. Chem.
Sci. 1986, 34, 2169-2177.
(32) Forrester, A. R.; Hepburn, S. P. J. Chem. Soc. (C) 1971, 4, 701-
703.
(54) Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. Arch. Biochem.
Biophys. 1980, 200, 1-16.
(33) Carloni, P.; Eberson, L.; Greci, L.; Sgarabotto, P.; Stipa, P. J. Chem.
Soc., Perkin Trans. 2 1996, 1297-1305.
(55) Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. Mol. Pharmacol.
1982, 21, 262-265.
(34) Janzen, E. G.; Lin, C.-R.; Hinton, R. D. J. Org. Chem. 1992, 57,
1633-1635.
(35) Eberson, L.; McCullough, J. J.; Persson, O. J. Chem. Soc., Perkin
Trans. 2 1997, 133-134.
(56) Karoui, H.; Rockenbauer, A.; Pietri, S.; Tordo, P. Chem. Commun.
2002, 3030-3031.