Cage-Escape of Geminate Radical Pairs Can Produce Peroxynitrate from Peroxynitrite under a Wide Variety of Experimental Conditions

Article abstract of DOI:10.1021/ja991077u

The spontaneous and CO₂-catalyzed decomposition of peroxynitrite yields HO^? and ^-CO₃^? radicals, respectively, together with ^?NO₂. The geminate HOVNO2 and ^-CO₃^?/^?NO₂ pairs undergo competitive in-cage collapse to nitrate and diffusive separation. Free HO^? and ^-CO₃^? radicals react with H₂O₂ and, in the presence of O₂, suitable alcohols or formate to produce superoxide, which is trapped by the ^?NO₂ to form peroxynitrate. The formation of peroxynitrate may influence the rate of change in optical density at 302 nm, the wavelength normally employed to monitor peroxynitrite decay, leading to misleading kinetic traces. Tetranitromethane (TNM) was used as a colorimetric probe for superoxide to quantify the yield of free HO^? (27-28%) and free ^-CO₃^? (32-33%). The yields of both of these free radicals are in excellent agreement with other recent estimates. Superoxide was also detected in some oxygenated aldehyde-catalyzed peroxynitrite decompositions both by peroxynitrate formation and by its reaction with TNM. Superoxide yields, measured with TNM, were aldehyde (RCHO) dependent (R = ^-O₂CC₆H₄, CH₃, CH₃CH₂, (CH₃)₃C and HOCH₂CHOH; yields were 15, 9, 0.8, 0, and 30%, respectively).

Full text of DOI:10.1021/ja991077u

J. Am. Chem. Soc. 1999, 121, 10695-10701
10695
Cage-Escape of Geminate Radical Pairs Can Produce Peroxynitrate
from Peroxynitrite under a Wide Variety of Experimental Conditions1
,2
George R. Hodges* and K. U. Ingold*
Contribution from the Steacie Institute for Molecular Sciences, National Research Council of Canada,
Ottawa, Ontario, Canada K1A 0R6
ReceiVed April 5, 1999. ReVised Manuscript ReceiVed September 7, 1999
•
-
•
Abstract: The spontaneous and CO2-catalyzed decomposition of peroxynitrite yields HO and CO3 radicals,
•
• •
-
• •
respectively, together with NO2. The geminate HO / NO2 and CO3 / NO2 pairs undergo competitive in-cage
collapse to nitrate and diffusive separation. Free HO and CO3 radicals react with H2O2 and, in the presence
of O2, suitable alcohols or formate to produce superoxide, which is trapped by the NO2 to form peroxynitrate.
•
-
•
•
The formation of peroxynitrate may influence the rate of change in optical density at 302 nm, the wavelength
normally employed to monitor peroxynitrite decay, leading to misleading kinetic traces. Tetranitromethane
•
(
TNM) was used as a colorimetric probe for superoxide to quantify the yield of free HO (27-28%) and free
-
•
CO3 (32-33%). The yields of both of these free radicals are in excellent agreement with other recent estimates.
Superoxide was also detected in some oxygenated aldehyde-catalyzed peroxynitrite decompositions both by
peroxynitrate formation and by its reaction with TNM. Superoxide yields, measured with TNM, were aldehyde
-
(
RCHO) dependent (R ) O2CC6H4, CH3, CH3CH2, (CH3)3C and HOCH2CHOH; yields were 15, 9, 0.8, 0,
and 30%, respectively).
Introduction
mM, respectively),14 and hence most peroxynitrite will be
trapped by CO2. Therefore, the subsequent reactions of the
peroxynitrite/CO2 adducts are of key importance in understand-
ing the effects of peroxynitrite in vivo.
In 1990, Beckman and co-workers3 proposed that two
endogenous, but relatively unreactive, free radicals, superoxide
and nitrogen monoxide, could undergo a rapid combination
1
0
-1 -1 4
Alvarez and co-workers have reported that hydrogen peroxide,
mannitol, or ethanol decreases the rate of peroxynitrite decom-
position in a concentration-dependent manner, following a
reaction (k ) 1.9 × 10
M
s ) to yield, in vivo, the powerful
5
oxidant peroxynitrite. The peroxynitrite anion is thermally
stable; however, peroxynitrous acid (pKa ) 6.55 and 7.33 in 1
and 650 mM phosphate buffer, respectively) spontaneously
decays with a rate constant of 1.20 s at 25 °C. A highly
reactive oxidizing species is formed during the spontaneous
15,16
4
hyperbolic function.
Addition of 100 mM hydrogen perox-
-
1
4
ide, mannitol, or ethanol reduced the spontaneous rate of
peroxynitrite decomposition by ca. 50, 34, and 25%, respec-
tively. It was tentatively proposed that these reagents hydrogen-
bonded to peroxynitrite and thereby slowed its rate of decom-
position. The possibility of endogenous compounds (e.g.,
glucose) associating themselves with peroxynitrite and altering
its reactions and reactivity could have wide-ranging biological
implications and could influence the results of in vivo studies.
decay of the peroxynitrous acid, which is probably the hydroxyl
radical.6
-10
The peroxynitrite anion reacts rapidly with CO2 (k
4
-1 -1 11
-
)
3 × 10 M s ) to yield an adduct, O2COONO, which
can undergo subsequent reactions with added substrates and/or
spontaneously decay at a higher rate than the peroxynitrous
11-13
acid.
The concentrations of CO2 in vivo are relatively high
(bicarbonate in intracellular and interstitial fluids of 12 and 30
The abilities of water and simple primary alcohols to act as
H
hydrogen bond donors, as described by their a values of 0.353
(
(
(
1) Issued as NRCC No. 42195.
2) NSERC Postdoctoral Fellow, 1998-99.
3) Beckman, J. S.; Beckman, T. W.; Chen, J.; Marshall, P. A.; Freeman,
17
and 0.328, respectively, are very similar, with water being
the slightly better hydrogen bond donor. Analogously, the
abilities of water and ethanol to act as hydrogen bond acceptors,
B. A. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 1620-1624.
4) Kissner, R.; Nauser, T.; Bugnon, P.; Lye, P. G.; Koppenol, W. H.
Chem. Res. Toxicol. 1997, 10, 1285-1292.
5) As has become convention, the term “peroxynitrite” is used here to
mean the sum of both the peroxynitrous acid and the peroxynitrite anion.
(
H
18
as described by their â values of 0.38 and 0.41, respectively,
are also very similar, with ethanol being the slightly better
(
(
6) Mahoney, L. R. J. Am. Chem. Soc. 1970, 92, 5262-5263.
(12) Lemercier, J.-N.; Padmaja, S.; Cueto, R.; Squadrito, G. L.; Uppu,
R. M.; Pryor, W. A. Arch. Biochem. Biophys. 1997, 345, 160-170.
(13) Denicola, A.; Freeman, B. A.; Trujillo, M.; Radi, R. Arch. Biochem.
Biophys. 1996, 333, 49-58.
(14) Carola, R.; Harely, J. P.; Noback, C. R. Human Anatomy and
Physiology; McGraw-Hill: New York, 1990; p 803.
(15) Alvarez, B.; Denicola, A.; Radi, R. Chem. Res. Toxicol. 1995, 8,
859-864.
(16) Alvarez, B.; Ferrer-Sueta, G.; Radi, R. Free Radical Biol. Med. 1998,
24, 1331-1337.
(17) Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.; Morris,
J. J.; Taylor, P. J. J. Chem. Soc., Perkin Trans. 2 1989, 699-711.
(18) Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor,
P. J. J. Chem. Soc., Perkin Trans. 2 1990, 521-529.
(7) Richeson, C. E.; Mulder, P.; Bowry, V. W.; Ingold, K. U. J. Am.
Chem. Soc. 1998, 120, 7211-7219.
8) Gatti, R. M.; Alvarez, B.; Vasquez-Vivar, J.; Radi, R.; Augusto, O.
Arch. Biochem. Biophys. 1998, 349, 36-46.
9) (a) Merenyi, G.; Lind, J. Chem. Res. Toxicol. 1998, 11, 243-246.
b) Merenyi, G.; Lind, J.; Goldstein, S.; Czapski, G. Chem. Res. Toxicol.
(
(
(
1
1
998, 11, 712-713. (c) Merenyi, G.; Lind, J. Chem. Res. Toxicol. 1997,
0, 1216-1220.
(
10) Lymar, S. V.; Hurst, J. K. Chem. Res. Toxicol. 1998, 11, 714-715.
Coddington, J. W.; Hurst, J. K.; Lymar, S. V. J. Am. Chem. Soc. 1999,
1
21, 2438-2443.
11) Lymar, S. V.; Hurst, J. K. J. Am. Chem. Soc. 1995, 117, 8867-
868.
(
8
1
0.1021/ja991077u CCC: $18.00 Published 1999 by the American Chemical Society
Published on Web 11/05/1999

10696 J. Am. Chem. Soc., Vol. 121, No. 46, 1999
Hodges and Ingold
hydrogen bond acceptor. It therefore appeared unlikely to us
that 100 mM alcohol could compete with 55 M water as a
hydrogen bond acceptor and/or donor with peroxynitrite. The
possibility that molecular products of formal reactions involving
the peroxynitrite anion were responsible for the observations
1
5,16
of Alvarez et al.
was virtually eliminated using spectroscopic
1
9
methods. We report herein our observations on the effects of
alcohols, hydrogen peroxide, and some other additives which,
under certain conditions, can induce complexities in the
measured kinetics of peroxynitrite decay.
Results
Kinetic experiments were carried out in a stopped-flow
apparatus at 24 °C using a 1.4 mM aqueous stock solution of
argon-purged (and therefore dioxygen-free) peroxynitrite (pH
12) in one syringe and buffer (purged or otherwise) in the other
syringe to reduce the pH to 6.3 or 8.5 after mixing. All
concentrations given below are those obtained after mixing equal
volumes from the two syringes; i.e., in the kinetic experiments
the peroxynitrite concentration was 0.7 mM. The rigorous
exclusion of adventitious CO2 was achieved as described
Figure 1. Change in OD at 302 nm on mixing equal volumes of
peroxynitrite (1.4 mM, argon-purged and CO -free) with phosphate
2
buffer (200 mM, dioxygen-purged) containing either no added reagents
(trace a) or 2 mM sodium bicarbonate (added 2-5 min prior to use)
and 0, 10, 20, 50, 100, 200, 300, 400, or 600 mM 2-propanol (traces
b-j, respectively). Final pH 6.3 and 24 °C. Inset: Alternative plot of
the data from trace j.
7
previously. Optical density (OD) refers to ODs measured at
3
02 nm unless otherwise stated.
In 1998, Alvarez et al.16 reported that, “when contaminating
was simple to demonstrate. The addition of sodium bicarbonate
to deoxygenated solutions allowed the rate-retarding effect of
added hydrogen peroxide to be observed, but ethanol did not
retard the rate unless the solutions contained dioxygen. Rate
retardation was also induced by methanol and 2-propanol in
the presence of both CO2 and O2. No rate retardation was
induced by tert-butyl alcohol, 2,2,2-trifluoroethanol, pure tert-
carbon dioxide was eliminated through bubbling freshly prepared
(
(
air-equilibrated) solutions with argon for 3 h, the slowing effect
produced by mannitol and ethanol on the rate of peroxynitrite
decay) was observed”. In our initial attempts to duplicate the
results of Alvarez et al.,1 we therefore employed rigorously
CO2-free conditions. Under such conditions (pH 6.3, dioxygen
free), neither ethanol nor hydrogen peroxide (both up to 150
mM) had any readily observable effect on the rate of perox-
ynitrite decay, and certainly not effects comparable to those
which had been reported.1 A later, very careful reexamination
of these additives under exactly the same conditions confirmed
that 150 mM ethanol had no effect on the measured rate constant
for peroxynitrite decay but revealed that 150 mM hydrogen
peroxide actually did reduce this rate constant by about 10%
5,16
21
butyl hydroperoxide, tetrahydrofuran, benzoate, or p-toluene-
sulfonate under any experimental conditions explored. The
effects of aldehydes are described below.
5,16
The rate of peroxynitrite decay is enhanced by the addition
of CO2, the increase in rate being directly proportional to the
1
1
CO2 concentration. At low CO2 concentrations (e.g., 0.05-
.1 mM), the addition of XOH (where XOH ) methanol,
0
ethanol, 2-propanol, or hydrogen peroxide) slowed the loss in
OD as peroxynitrite decayed in a concentration-dependent
manner, provided O2 was also present for the three alcohols.
These systems become “saturated” in XOH before the accelerat-
ing effect of the CO2 was completely suppressed. At these low
CO2 concentrations, the decrease in OD always appeared to
follow first-order kinetics, regardless of the XOH concentration
-
1
(
from 0.76 to 0.68 s , see Note Added in Proof). Alvarez et
al. used “air-equilibrated” solutions in their experiments with
ethanol,16 and, although the possible role of CO2 was not
15
addressed in their earlier experiments with hydrogen peroxide,
we assume that these solutions were also air-equilibrated. We
therefore turned to “air-equilibrated” buffer solutions while
keeping the alkaline peroxynitrite solutions free of dioxygen
and CO2. Under these conditions, both ethanol and hydrogen
peroxide reduced the rate of decrease in OD as peroxynitrite
decayed, just as had been reported.15,16
(up to 1 M). However, with 1 mM CO2, although there was an
analogous slowing of the decrease in OD on addition of XOH,
there was a deviation from first-order kinetics (Figure 1), and,
at high XOH, the results could be fitted to second-order kinetics
It appeared likely that the rate-retarding effects of hydrogen
peroxide and ethanol on decay of the 302-nm absorption were
(see inset in Figure 1). The observed kinetic effects were the
same whether the XOH was premixed in the peroxynitrite anion
syringe or in the phosphate buffer syringe.
The apparent shift to second-order kinetics shown in the
Figure 1 inset is very misleading, since the rate of OD change
at a given OD is dependent on the starting peroxynitrite
concentration (see Figure 2). The kinetic traces shown in Figure
dependent on the presence of CO2, which is known to catalyze
the decomposition of peroxynitrite.1
1-13
The critical role of CO2
(
19) (i) The UV-Vis spectrum of the peroxynitrite anion (0.7 mM, pH
)
12) was unaltered upon the addition of ethanol (up to 750 mM). (ii) The
Raman spectrum of an aqueous solution (pH 12) containing peroxynitrite
anion (1 M) and ethanol (4 M) was identical to the sum of the spectra for
the same concentrations of these two species measured separately: the bands
were not shifted, nor were any new bands formed. (The Raman spectrum
of the peroxynitrite anion (1 M, pH 12) matched that previously reported
2
would be consistent with either the loss of a reagent or the
22,23
build-up of an unstable intermediate.
To distinguish between
these two possibilities, the CO2 concentration was increased to
increase the rate of peroxynitrite decay and hence increase the
rate of formation of any intermediate. At relatively high CO2
concentrations, it became immediately apparent that the unusual
2
0
14
by Tsai et al. ) (iii) The N NMR spectrum of the peroxynitrite anion (1
M, pH 12) showed a very broad band (590 Hz) at 173 ppm (relative to
nitromethane at zero ppm), which was not shifted, nor were any new bands
observed between -60 and 280 ppm (scan width 9800 Hz) on addition of
ethanol (up to 1 M).
(
20) Tsai, J.-H. M.; Harrison, J. G.; Martin, J. C.; Hamilton, T. P.; van
der Woerd, M.; Jablonsky, M. J.; Beckman, J. S. J. Am. Chem. Soc. 1994,
16, 4115-4116.
(21) Commercial tert-butyl hydroperoxide showed a rate-retarding effect
on the CO2-catalyzed decomposition of peroxynitrite, but this was found
to be due to the presence of the impurity, H2O2.
1

Cage-Escape of Geminate Radical Pairs
J. Am. Chem. Soc., Vol. 121, No. 46, 1999 10697
observed in the absence of XOH. However, the nitroform anion
was readily identified upon the addition of all (active) XOH. It
is, of course, produced by the one-electron reduction of TNM,
a compound which has been widely employed as a probe for
the superoxide anion radical, reaction 1, in various biomimetic
2
7
and chemical systems (to which we can now add the
2
8
peroxynitrite/CO2/XOH(O2) systems).
We believe that the unstable intermediate formed during the
decay of peroxynitrite in the presence of CO2 and active XOH
(
+O2) is peroxynitrate, as was first proposed by Goldstein and
3
0
Czapski (see below). Their mechanism for its formation from
active alcohols and dioxygen is shown in reactions 2-8.
Figure 2. Change in OD at 302 nm on mixing equal volumes of
peroxynitrite (3.4, 1.7, and 0.85 mM; ×, O, and b respectively) with
phosphate buffer (200 mM, dioxygen-purged) containing ethanol (700
mM) and bicarbonate (2 mM, added 2-5 min prior to use). Final pH
6.3 and 24 °C. For the two lower peroxynitrite concentrations, the times
have been offset so that the differences between the decay traces for
equal concentration of “fresh” peroxynitrite and “partially decayed”
peroxynitrite can be clearly seen. Inset: Change in OD at 302 nm on
mixing equal volumes of peroxynitrite (1.4 mM, argon-purged and CO
2
-
free) with phosphate buffer (200 mM, air-purged) containing ethanol
(
700 mM) and sodium bicarbonate (10 mM, added 2-5 min prior to
use). Final pH 6.3 and 24 °C.
It seemed likely to us that peroxynitrate might also be
produced when CO2-free peroxynitrite was decomposed in the
presence of aliphatic aldehydes since Pryor and co-workers31
have shown that these compounds react rapidly with the
peroxynitrite anion and, like CO2, catalyze its decomposition.
The homolytic portion of the proposed mechanism31 is given
by reactions 9 and 10. Using aldehydes in place of CO2 offered
kinetics arose from the build-up of an unstable intermediate
which absorbs at the 302 nm wavelength used to monitor
peroxynitrite decay. With 5 mM CO2 and, for example, 350
mM ethanol (using an air-saturated buffer solution), the time
scales for peroxynitrite decay and for decay of the intermediate
became well separated, both following first-order kinetics (see
inset in Figure 2). Another important observation was that the
rate of decay of the intermediate (i.e., the slow, second stage
of reaction) was identical, within the accuracy of our kinetic
measurements, for all (active) XOH reagents. This implies that
the same intermediate is formed in all active XOH, peroxynitrite,
CO2 (and O2), systems.
As a further probe into the origin of the unstable intermediate,
we added tetranitromethane (TNM, 0.2-2.5 mM) to peroxyni-
trite/bicarbonate (buffer dioxygen-saturated) systems. The ni-
the potential advantage of working at higher pH’s, where the
fraction of peroxynitrite present as peroxynitrous acid would
be small, and consequently the spontaneous decay of perox-
-
1
-1 26
troform anion (λmax 350 nm, ꢀ ) 15 000 M cm ) was not
3
2
(
22) Since [XOH] . [CO2], the traces shown in Figure 1 could also be
ynitrite would be slow. The reactions described below were
performed at pH 8.5 using dioxygen-saturated, CO2-free pyro-
consistent with a reduction in the free CO2 concentration because of complex
formation between XOH and CO2, i.e.,
33
phosphate buffer (100 mM). Trimethylacetaldehyde increased
(
27) See, e.g.: McCord, J. M.; Fridovich, I. J. Biol. Chem. 1969, 244,
049-6055. Ingold, K. U.; Paul, T.; Young, M. J.; Doiron, L. J. Am. Chem.
Soc. 1997, 119, 12364-12365.
28) Although strongly reducing radicals (such as MeC (OH)2, see below)
also react rapidly with TNM to form the nitroform anion, this reaction
can be ignored because our experiments were done in dioxygen-saturated
(1.4 mM) solutions (both buffer and peroxynitrite). Most of the carbon-
centered radicals formed will react with dioxygen rather than with (0.2 mM)
TNM.
6
•
(
29
However, such an interaction can be ruled out because the strong,
asymmetric stretch IR band of (CO2)aq centered at 2344 cm-1 had neither
its intensity nor its position affected by the addition of ethanol (up to 200
mM).
23) At [CO2] < [ OONO], the addition of tyrosine or nitrite25 does
-
24
(
not affect the initial rate of decay of peroxynitrite, but the first-order plots
are curved because the reaction between CO2 and peroxynitrite is catalytic
in CO2 and the tyrosine or nitrite reduces the efficiency of re-formation of
CO2, which, of course, lowers the CO2 concentration as the reactions
(29) Schuchmann, M. N.; von Sonntag, C. J. Am. Chem. Soc. 1988, 110,
5698-5701.
(30) Goldstein, S.; Czapski, G. J. Am. Chem. Soc. 1998, 120, 3458-
3463.
2
4,25
-
proceed.
for our results can be ruled out.
24) Zhang, H.; Squadrito, G. L.; Pryor, W. A. Nitric Oxide: Biol. Chem.
997, 1, 301-307.
25) Pryor, W. A.; Lemercier, J.-N.; Zhang, H.; Uppu, R. M.; Squadrito,
G. L. Free Radical Biol. Med. 1997, 23, 331-338.
26) Bielski, B. H. J.; Allen, A. O. J. Phys. Chem. 1967, 71, 4544-
549.
Since we worked at [CO2] > [ OONO], a similar explanation
(31) Uppu, R. M.; Winston, G. W.; Pryor, W. A. Chem. Res. Toxicol.
1997, 10, 1331-1337.
•
(
(32) In fact, the anion dissociates slowly but reversibly to give NO and
-
• 9a
1
O2 . The rate of anion decay can be accelerated by adding TNM to remove
(
the superoxide.9a
(33) Buffers made from borate are known to accelerate the decay of
,34
(
peroxynitrite.4
4
(34) Hughes, M. N.; Nicklin, H. G. J. Chem. Soc. (A) 1968, 450-452.

10698 J. Am. Chem. Soc., Vol. 121, No. 46, 1999
Hodges and Ingold
and to extend the conditions under which peroxynitrate forma-
tion can be observed.
Briefly, under appropriate conditions (viz., high CO2 con-
30
centrations), Goldstein and Czapski observed that peroxynitrite
decay monitored at 302 nm showed two consecutive first-order
processes (similar to the decay trace shown in the inset in our
Figure 2). They clearly identified the intermediate responsible
for the slow, second stage of the decay as peroxynitrate formed
in a maximum yield of 30-33%. (The peroxynitrate anion
3
0,37,38
absorbs at 302 nm,
whereas peroxynitric acid, pKa )
3
0,37,38
30,37,39
5.9,
does not absorb above 280 nm
). Their proposed
mechanism for the formation of peroxynitrate in the presence
of methanol and 2-propanol (and ethanol) is shown in reactions
2
-8. The critical steps in the generation of peroxynitrate include
hydrogen atom abstraction from the alcohol by the carbonate
radical anion (reaction 5), subsequent formation of the
superoxide radical anion (via reactions 6 and 7), and its trapping
4
0
Figure 3. Change in OD at 280 nm on mixing equal volumes of
peroxynitrite (1.4 mM, argon-purged and CO -free) with pyrophosphate
free) containing acetalde-
•
by the NO2 radical (reaction 8) which is, itself, formed, together
2
with the carbonate radical anion, in the initial homolysis of
O2COONO (reaction 4).
buffer (100 mM, dioxygen-purged and CO
2
-
hyde (350 mM). Final pH 8.5 and 24 °C.
Our finding that neither tert-butyl alcohol, 2,2,2-trifluoroet-
hanol, nor tert-butyl hydroperoxide influence the observed decay
kinetics of peroxynitrite (in the presence of CO2 and O2) further
supports the Goldstein and Czapski mechanism. That is, tert-
butyl alcohol has no R-hydrogen atom available for abstraction
by the carbonate radical anion. Although trifluoroethanol has
the rate of peroxynitrite decay in a concentration-dependent
manner. These reactions followed pseudo-first-order kinetics for
up to seven half-lives, with dramatic increases in the perox-
ynitrite decay rate; e.g., with 100 mM trimethylacetaldehyde
(from a near saturation 200 mM in the buffer), the half-life was
•
R-hydrogen atoms which might be abstracted, any CF3CH(OO )-
reduced from ∼35 to 0.06 s. Peroxynitrate formation was not
OH radicals formed would not undergo heterolytic fragmentation
to superoxide and a carbocation (reaction 7) because the
carbocation would be destabilized by the electron-withdrawing
CF3 group. Radical attack on tert-butyl hydroperoxide would
not yield superoxide.
observed, and when the reaction was repeated with added
3
5
TNM (0.1 or 2.5 mM) the nitroform anion was not detected.
More interestingly, while acetaldehyde also accelerated the
decomposition of peroxynitrite (in a concentration-dependent
manner), there was a build-up and subsequent decay of an
unstable intermediate (Figure 3). The decay of this intermediate
Dioxygen is required for formate to exhibit its effect on the
30
36
apparent kinetics of peroxynitrite decay (reaction 11), a result
we have confirmed. However, dioxygen is not required with
H2O2 because the superoxide-forming reactions 5, 6, and 7 are
followed essentially the same kinetics as that of the intermedi-
ate formed in the peroxynitrite/CO2/(active) XOH(O2) systems,
from which we conclude that it is peroxynitrate (see Figure 3
and compare with inset in Figure 2). The addition of 0.05 mM
TNM reduced the yield of peroxynitrate, while 0.2 mM TNM
prevented any build-up of peroxynitrate and gave, instead, the
3
0
replaced by reactions 12 and 13.
2
8
nitroform anion.
In the presence of 4-carboxybenzaldehyde, the decay of
peroxynitrite and possible build-up and decay of peroxynitrate
could not be monitored reliably because of a strong background
absorbance below 320 nm. However, upon the addition of TNM
It will be obvious that any powerful one-electron oxidant
produced during the decay of peroxynitrite should yield super-
oxide (and hence peroxynitrate, reaction 8) from suitable
substrates. There is now strong evidence that peroxynitrous acid
decomposes to give the hydroxyl radical in significant yields
under conditions where CO2 has been rigorously excluded
(
8
0.2 mM) to carboxybenzaldehyde/peroxynitrite systems at pH
.5, the absorbance due to the nitroform anion grew in at a rate
directly proportional to the aldehyde concentration (25, 50, and
00 mM). In the presence of TNM, the propionaldehyde- and
1
glyceraldehyde-catalyzed decomposition of peroxynitrite also
gave the nitroform anion.
6-14
(reaction 14).
Peroxynitrate should therefore be formed from
peroxynitrite in the presence of suitable substrates in the
complete absence of CO2. This raises the following question:
Why has peroxynitrate been observed only (or, probably only,
Discussion
16
in view of the argon-bubbling experiments of Alvarez et al. )
in the presence of CO2? The answer lies in the general usage
After this work had been initiated, we became aware of
Goldstein and Czapski’s30 very thorough and careful study of
the effect of XOH on the kinetics of decay of peroxynitrite in
the presence of CO2 for XOH ) methanol, 2-propanol, hydrogen
peroxide, and formate. Our present results serve to further
confirm both their observations and mechanistic conclusions
(37) Logager, T.; Sehested, K. J. Phys. Chem. 1993, 97, 10047-10052.
(38) Goldstein, S.; Czapski, G. Inorg. Chem. 1997, 36, 4156-4162.
(
39) Appelman, E. H.; Gosztola, D. J. Inorg. Chem. 1995, 34, 787-
7
91.
(40) There is now overwhelming evidence that the peroxynitrite/CO2
41-43
reaction yields the carbonate radical anion.
(
35) There is a direct reaction between aldehydes and TNM, but it is
insignificant on the 1 s time scale of our experiments.
36) The decay is actually slightly faster at pH 8.5 than at pH 6.3. This
(41) Bonini, M. G.; Radi, R.; Ferrer-Sueta, G.; Ferreira, A. M. Da C.;
Augusto, O. J. Biol. Chem. 1999, 274, 10802-10806.
(42) Goldstein, S.; Czapski, G. J. Am. Chem. Soc. 1999, 121, 2444-
2447.
(43) Meli, R.; Nauser, T.; Koppenol., W. HelV. Chim. Acta 1999, 82,
722-725.
(
increase in rate can be quantitatively accounted for by the decreased ratio
of (relatively stable) peroxynitric acid (pKa 5.9) to (short-lived) peroxynitrate
anion.

Cage-Escape of Geminate Radical Pairs
J. Am. Chem. Soc., Vol. 121, No. 46, 1999 10699
whereas the peroxynitric acid would decay much more slowly.44
In the second stage, this solution was rapidly mixed with an
equal volume of aqueous sodium hydroxide (50 or 100 mM) to
increase the pH to 7.6 or 9.3, respectively. The OD at 280 nm
increased instantly and then decayed with first-order kinetics.46
When the ethanol experiment was repeated using a peroxynitrite
solution which had been allowed to decompose by standing at
room temperature for 72 h, there was no transient peroxynitrate
absorbance at 280 nm. However, peroxynitrate was detected
when hydrogen peroxide was used, which can be attributed to
the in situ formation of peroxynitrite via reaction 17.
It is obvious that CO2 is not required in order to form
peroxynitrate from peroxynitrite in the presence of active XOH
(
and dioxygen). However, except under certain carefully chosen
Figure 4. Change in OD at 280 nm on mixing equal volumes of
conditions, peroxynitrate will not be observed because it decays
at a rate similar to or even faster than that of the peroxynitrite.
The CO2 aids in the “visualization” of peroxynitrate because it
increases the rate of peroxynitrite decay without affecting the
decay of peroxynitrate, which thus becomes easier to observe.
Addition of CO2 also slightly increases (vide infra) the yield of
the one-electron oxidizing agent needed to generate the super-
oxide form XOH (viz., CO3 vs OH). It is important to realize
that peroxynitrate will almost certainly also be formed in any
CO2-free or CO2-containing systems in which peroxynitrite is
generated in situ from a source of superoxide and nitric oxide
peroxynitrite (1.4 mM, argon-purged and CO
buffer (200 mM, dioxygen-purged and CO -free) containing either no
s) or 700 mM (- - -) ethanol. Final pH 6.3 and 24 °C.
2
-free) with phosphate
2
(
-
•
•
of 302 nm to monitor peroxynitrite decay and the kinetics of
decay of peroxynitrite and peroxynitrate at the pH’s commonly
employed in such studies (e.g., 6.3 in the present work). That
is, the peroxynitrite anion (λmax ) 302 nm) is remarkably stable,
•
(because NO2 will be produced during peroxynitrite decay and
will compete with the NO for the available superoxide).
•
4
but peroxynitrous acid (pKa ) 6.5-7.3, depending on buffer,
TNM (2 mM) was used to measure the yields of superoxide
formed during the decomposition of peroxynitrite/XOH (24 °C,
pH 6.3) in the absence and presence of CO2. In the absence of
CO2, the superoxide was generated from H2O2 and EtOH/O2
but not from formate/O2 because CO2 would also be produced
(reaction 18). In the presence of CO2, all three superoxide-
see Introduction) decays with a half-life, τ1/2 ) 0.58 s at 25
4
30,37,38
°
C. In contrast, peroxynitric acid (pKa ∼ 5.9)
is more
37
37
stable (τ1/2 ) 16.5 min) than its anion (τ1/2 ∼ 0.7s).
Moreover, although the peroxynitrate anion has a broad absorp-
tion band centered at 280 nm, peroxynitric acid does not absorb
appreciably above 280 nm,30,37,39 making it “invisible” at 302
nm.
With the above information, it becomes obvious that perox-
ynitrate formation should become kinetically observable at 280
nm if the pH is reduced slightly from 6.3. This will increase
the rate of decay of peroxynitrite and decrease the rate of decay
of peroxynitrate. However, this kinetic advantage has to be
balanced against the reduction in detection sensitivity due to
protonation of the peroxynitrate anion. We therefore chose a
pH of 5.5. As expected, the 280 nm trace for peroxynitrite decay
followed perfect first-order kinetics for CO2-free, O2-saturated
solutions in the absence of ethanol, but in the presence of ethanol
the rate of loss of the 280 nm absorbance decreased with time
generating systems were employed. The measured yields of
superoxide decreased as the peroxynitrite concentration was
increased (Figure 5 and Supporting Information), presumably
because peroxynitrite and its nitrite contaminant compete with
the substrate for the available oxidizing radicals. This was most
47
noticeable with H2O2 because only 50 mM could be employed
in order to limit its background reaction with TNM), whereas
00 mM EtOH and formate were used. The yields of superoxide,
(
5
measured as the nitroform anion, when extrapolated to zero
peroxynitrite concentration can be equated to the yields of cage-
(see Figure 4 and compare with Figure 1). The slower, second
•
-
•
escaped HO and CO3 radicals (see Table 1). The yields for
these two radicals are independent of the superoxide-generating
system employed (although a small correction is necessary for
stage of these reactions can be attributed to decay of peroxyni-
trate formed during peroxynitrite decay. In these CO2-free
systems, reactions 2, 4, and 5 are replaced by reactions 14 and
•
the HO /EtOH system, see footnote a in Table 1) and are in
1
6.
outstanding agreement with what appear to be the best literature
(
44) Under the conditions used, the peroxynitric acid had a half-life of
37
about 1 min (rather than 16.5 min) because of its reaction with nitrous
acid (pKa 3.4) present from the peroxynitrite synthesis.45
Peroxynitrate formation was also demonstrated by studying
the decay of CO2-free peroxynitrite in two stages. In the first
stage, under CO2-free conditions, the peroxynitrite anion (pH
(
45) Pryor, W. A.; Cueto, R.; Jin, X.; Koppenol, W. H.; Ngu-Schwemlein,
12) was rapidly mixed with an equal volume of oxygen-saturated
M.; Squadrito, G. L.; Uppu, P. L.; Uppu, R. M. Free Radical Biol. Med.
1995, 18, 75-83.
phosphate buffer (200 mM)/HCl (100 mM) containing either
hydrogen peroxide (80 mM) or ethanol (300 mM) to give a
solution with a final pH of 3.3. Under these conditions, the
peroxynitrous acid would decay completely in a few seconds,
(46) For reasons we did not explore, the rate constant for peroxynitrate
decay was about 50% larger than the value found in all other experiments.
47) Nitroform anion yields were, indeed, reduced by the addition of
nitrite or decomposed peroxynitrite to H2O2-containing systems.
(

10700 J. Am. Chem. Soc., Vol. 121, No. 46, 1999
Hodges and Ingold
The formation of superoxide (and, hence, of peroxynitrate)
in the reaction of oxygenated peroxynitrite with all the aldehydes
examined except trimethylacetaldehyde is intriguing.56 These
different results can be understood by a consideration of the
likely reactions of the cage-escaped hydroxyalkoxyl radical,
•
RCH(OH)O , formed in reaction 10a. This radical will undergo
two principal competing reactions, â-scission (reaction 19) and
•
a rearrangement to give the carbon-centered radical, RC (OH)2,
via the well-known, water-mediated, 1,2-hydrogen atom shift5
8-60
(reaction 20). In water, these electron-rich, carbon-centered
radicals are known to react with dioxygen to form superoxide
2
9,61
and a carbocation
(reaction 21).
Figure 5. Yields of nitroform anion upon mixing equal volumes of
TNM (4 mM, phosphate buffer, dioxygen saturated) containing H
b/O, 50 mM), EtOH (9/0, 500 mM), or -HCO
(4, 500 mM) with
various concentrations of peroxynitrite (dioxygen-saturated). The filled
symbols are for data acquired under strictly CO -free conditions, and
the open symbols are for data acquired with 12 mM bicarbonate added
to the TNM solution 2-5 min prior to mixing (under these conditions,
2 2
O
(
2
The 1,2-H atom shift for analogous rearrangements of
•
RCH2O radicals is believed to occur more rapidly when R is
2
5
9
60
allyl or phenyl than when R is an alkyl group because of
the increased thermodynamic driving force provided by reso-
•
nance stabilization of the product RC H(OH) radical. Reaction
>
99% of the peroxynitrite will react with the CO
4 °C.
2
). Final pH 6.3 and
2
0 will therefore be particularly fast in the peroxynitrite/
2
carboxybenzaldehyde system. Moreover, in this system the
â-scission with loss of the carboxyphenyl radical will be
extremely slow for thermodynamic reasons (to judge from the
extensive data available for structurally analogous alkoxyl
Table 1. Estimated Yields of Cage-Escaped HO and -CO
•
•
Free
3
Radicals Produced in the Spontaneous and CO
2
-Catalyzed
a
Decomposition of Peroxynitrite, Respectively
62
radicals). Carboxybenzaldhyde and other benzaldehydes would
therefore be expected to give a molecule of superoxide for every
•
RCH(OH)O radical which escapes from the solvent cage. In
contrast, reaction 20 is expected to be somewhat slower when
carboxybenzaldehyde is replaced by trimethylacetaldehyde,
whereas â-scission, with loss of the tert-butyl radical, is expected
to be extremely fast. Thus, the oxygenated peroxynitrite/
trimethylacetaldehyde system was not expected to give super-
oxide in significant yield, and, in fact, none could be detected
as peroxynitrate or with TNM.
Superoxide yields were quantified with TNM (2 mM) in
aldehyde/peroxynitrite/dioxygen systems at pH 8.5 under CO2-
free conditions and with aldehyde concentrations high enough
that >98% of the peroxynitrite underwent aldehyde-catalyzed
decay. These superoxide yields (which were independent of the
peroxynitrite concentration, 0.02-1.0 mM) were as follows:
15%, 4-carboxylbenzaldehyde (100 mM); 9%, acetaldehyde (35
b
6
2
c
a
These yields are based on superoxide formation in the presence of
appropriate substrates, as quantified by nitroform anion formation from
TNM and extrapolated to zero peroxynitrite concentration (see Figure
). Final pH 6.3 and 24 °C. Abstraction of hydrogen atoms from EtOH
3 2
(13.2%), CH (84.3%), and OH (2.5%)
positions; see: Asmus, K. D.; Mockel, H.; Henglein, A. J. Phys. Chem.
973, 77, 1218-1221. Superoxide is formed from CH
requires abstraction from the CH or OH group (86.8%). The nitroform
anion yield extrapolated to zero peroxynitrite concentration (24.4%)
was therefore divided by 0.868. Not measured because of CO
formation, see text.
b
5
•
by HO occurs from the CH
•
1
3
C HOH, which
2
c
2
(
56) The reaction of peroxynitrite with pyruvate shows many parallels
57
with its reactions with aldehydes, viz.,
values for yields of cage-escaped HO• 6,48,49 and CO3
-
• 30,53,54
radicals measured by quite different procedures.55
(
57) V a´ squez-Vivar, J.; Denicola, A.; Radi, R.; Augusto, O. Chem. Res.
(
48) Gerasimov, O. V.; Lymar, S. V. Inorg. Chem. 1999, 38, 4317-
321.
49) A highly reactive oxidant which was said not to be the HO radical
Toxicol. 1997, 10, 786-794.
4
(58) Gilbert, B. C.; Laue, H. A. H.; Norman, R. O. C. J. Chem. Soc.,
Perkin Trans. 2 1976, 1040-1044. Dobbs, A. J.; Gilbert, B. C.; Laue, H.
A. H.; Norman, R. O. C. J. Chem. Soc., Perkin Trans. 2 1976, 1044-
1047. Gilbert, B. C.; Holmes, R. A. G.; Laue, H. A. H.; Norman, R. O. C.
J. Chem. Soc., Perkin Trans. 2 1976, 1047-1052.
•
(
5
0
was found to be formed in a yield of 40%, which was later revised to
2%. Still later it was admitted that this oxidant was the HO radical
though its yield was still given as 40%). The yield of free HO from these
experiments
5
1
9b
•
3
•
(
5
0,51
52
should be considered to be 32%.
(59) Elford, P. E.; Roberts, B. P. J. Chem. Soc., Perkin Trans. 2 1996,
2247-2256.
(60) Konya, K. G.; Paul, T.; Lin, S.; Lusztyk, J.; Ingold, K. U. J. Am.
Chem. Soc., submitted.
(
50) Goldstein, S.; Czapski, G. Inorg. Chem. 1995, 34, 4041-4048.
(51) Goldstein, S.; Czapski, G. Nitric Oxide: Biol. Chem. 1997, 1, 417-
4
22.
(
(
(
(
52) Goldstein, S. Private communication.
(61) Bothe, E.; Schuchmann, M. N.; Schulte-Frohlinde, D.; von Sonntag,
C. Photochem. Photobiol. 1978, 28, 639-644. Bothe, E.; Schulte-Frohlinde,
D.; von Sonntag, C. J. Chem. Soc., Perkin Trans. 2 1978, 416-420.
(62) Howard, J. A.; Scaiano, J. C. In Radical Reaction Rates in Liquids;
Fischer, H., Ed.; Landolt-B o¨ rnstein, New Series; Springer-Verlag: Berlin,
1984; Vol. 13d. Lusztyk, J. In Radical Reaction Rates in Liquids; Fischer,
H., Ed.; Landolt-B o¨ rnstein, New Series; Springer-Verlag: Berlin, 1997;
Vol. 18d-1.
53) Goldstein, S.; Czapski, G. Inorg. Chem. 1997, 36, 5113-5117.
54) Lymar, S. V.; Hurst, J. K. Inorg. Chem. 1998, 37, 294-301.
•
55) We estimated previously that HO radicals were formed in a yield
7
of ca. 10%, which was based on their trapping with dimethyl sulfoxide.
Although we proved by product identification that HO radicals were
undoubtedly produced, their yield was clearly underestimated for reasons
we do not yet understand.
•

Cage-Escape of Geminate Radical Pairs
J. Am. Chem. Soc., Vol. 121, No. 46, 1999 10701
mM); 0.8%, propionaldehyde (50 mM); and 0%, trimethylac-
etaldehyde (30 mM). We attribute the decrease in the yield of
superoxide along this series of aldehydes to the undoubted
increase in the importance of â-scission62 (reaction 19) relative
to the 1,2-H atom shift (reaction 20). That â-scission plays a
significant role even in the acetaldehyde/peroxynitrite reaction
has been unequivocally demonstrated by Nakao, Ouchi, and
seem probable that aldehyde/peroxynitrite reactions should give
a ca. 30% yield of cage-escaped radicals and that our lower
values for 4-carboxybenzaldehyde (15%) and acetaldehyde
(19%) are a consequence of as-yet unrecognized “other”
reactions of RCH(OH)O radicals. If (i) is correct, then there
are some fascinating structure/activity relations waiting to be
explored. Further work is planned.
•
6
3
Augusto. These workers spin-trapped the methyl radical in
yields of about 10% (based on peroxynitrite), which paralleled
Conclusions
(
as they should) the yields of formate (8% at pH 6.7 and 13%
We have provided additional evidence in support of the
3
0
at pH 9.1). Comparing our measured yield of superoxide (9%
at pH 8.5) with Augusto and co-workers’ 63 yields of methyl
radicals and formate indicates that the rates of reactions 19 and
Goldstein and Czapski mechanism(s) by which the decom-
position of peroxynitrite in the presence of CO2 and H2O2 or
dioxygen plus formate or a suitable alcohol yields peroxynitrate,
the formation of which can affect apparent peroxynitrite/CO2
decay kinetics monitored at 302 nm. We have also shown that
peroxynitrate can be a confounding factor during the decay of
peroxynitrite in the presence of H2O2 or a suitable alcohol plus
dioxygen in the complete absence of CO2. Peroxynitrate is also
formed in the presence of certain aldehydes and dioxygen. The
formation of peroxynitrate under so many different conditions
certainly raises the question of its biological significance.
Finally, we have used tetranitromethane to determine the yields
of superoxide formed from peroxynitrite under a variety of
conditions. These superoxide yields can be used to quantify the
2
0 are roughly equal when R ) CH3. Furthermore, the total
yield of cage-escaped radicals in the acetaldehyde/peroxynitrite
system must be ca. 9% + (8-13%), i.e., ca. 19%. Acetate is,
of course, formed concomitantly with superoxide, i.e., reaction
2
2. However, our 9% yield of superoxide is lower than the
6
3
reported yields of acetate for this system, viz., 17% (pH 6.7),
1
9% (pH 7.0),31 and 20% (pH 9.1), which implies that there
63
must be another, previously unrecognized, route to acetate. We
suggest that this is an in-cage disproportionation (or electron
transfer) involving the initial geminate radical pair (reaction 23).
That is, this caged radical pair undergoes three competing
forward reactions: reactions 10a (R ) CH3), 10b (R ) CH3),
and 23.
•
yields of freely diffusing NO2 and oxyl radicals for peroxyni-
•
trous acid, 27-28% HO , and for the peroxynitrite/CO2 adduct,
-
•
32-33% CO3 , both percentages being in excellent agreement
6
,30,48,49,53,54
with recent literature data.
Not all peroxynitrite/
aldehyde adducts homolyze to hydroxyalkoxyl radicals which
quantitatively rearrange and react with dioxygen to form
•
superoxide. However, our data suggest that freely diffusing NO2
and oxyl radicals are formed in yields of 15% from carboxy-
benzaldehyde, 30% from glyceraldehyde, and (when combined
with data from the literature) ca. 19% from acetaldehyde.
The reaction of glyceraldehyde (50 mM) with oxygenated
peroxynitrite and TNM was studied because it appeared likely
that the hydroxyalkoxyl radical formed in reaction 10 would
be able to yield superoxide both via the “normal” 1,2-H atom
shift and from the hydroxyalkyl radical formed by â-scission
6
3
Acknowledgment. We sincerely thank Drs. O. Augusto, G.
Czapski, S. Goldstein, W. H. Koppenol, S. V. Lymar, and G.
Merenyi and an anonymous reviewer for their comments,
criticisms, and suggestions regarding an earlier version of this
paper. We also thank Drs. T. Paul and J. Lusztyk for helpful
discussions, Dr. C. A. Talk for the Raman spectra, D. Leek for
(reactions 24-27). The yield of superoxide (measured as the
nitroform anion) with glyceraldehyde was 29.7%, a result which
is very comparable to the yields of superoxide (cage-escaped
-
free radicals) from HOONO (∼28%) and O2COONO (∼33%).
14
the N NMR spectra, and the National Foundation for Cancer
The result with glyceraldehyde raises two intriguing pos-
Research for partial support of this work.
Note Added in Proof: Dr. Sara Goldstein has demonstrated
by chemical kinetic simulation that in the absence of nitrite (a
contaminant in our peroxynitrite) the reduction is the peroxy-
nitrite’s apparent rate of decay induced by 150 mM H2O2 in
the absence of CO2 would be ca. 210%, rather than the 10%
we measured.
Supporting Information Available: Full experimental
details and a table of nitroform anion yields in the absence and
presence of CO2 using various superoxide generating systems
sibilities: (i) different aldehydes give very different yields of
cage-escaped radicals and (ii) different aldehydes give similar
yields of cage-escaped radicals. If (ii) is correct, then it would
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
(63) Nakao, L. S.; Ouchi, D.; Augusto, O. Chem. Res. Toxicol. 1999,
1
2, 1010-1018.
JA991077U