The reaction of peroxynitrite with morpholine (secondary amines) revisited: The overlooked hydroxylamine formation

Article abstract of DOI:10.1002/hlca.200690222

The reaction of peroxynitrite/peroxynitrous acid with morpholine as a model compound for secondary amines is reinvestigated in the absence and presence of carbon dioxide. The concentration- and pH-dependent formation of N-nitrosomorpholine and N-nitromorpholine as reported in three previous papers ([25] [26] [14]) is basically confirmed. However, ¹³C-NMR spectroscopic product analysis shows that, in the absence of CO₂, N-hydroxymorpholine is, at pH≥7, the major product of this reaction, even under anaerobic conditions. The formation of N-hydroxymorpholine has been overlooked in the three cited papers. Additional (ring-opened) oxidation products of morpholine are also detected. The data account for radical pathways for the formation of these products via intermediate morpholine-derived aminyl and α-aminoalkyl radicals. This is further supported by EPR-spectrometric detection of morpholine-derived nitroxide radicals, i.e., morpholin-4-yloxy radicals. N-Nitrosomorpholine, however, is very likely formed by electrophilic attack of peroxynitrite-derived N₂O₄. ¹⁵N-CIDNP Experiments establish that, in the presence of CO₂, N-nitro- and C-nitromorpholine are generated by radical recombination. The present results are in full accord with a fractional (28±2% ) homolytic decay of peroxynitrite/peroxynitrous acid with release of free hydroxyl and nitrogen dioxide radicals.

Full text of DOI:10.1002/hlca.200690222

Helvetica Chimica Acta – Vol. 89 (2006)
2399
The Reaction of Peroxynitrite with Morpholine (Secondary Amines)
Revisited: The Overlooked Hydroxylamine Formation
by Michael Kirsch*a), Hans-Gert Korth*b), Angela Wensinga), Manfred Lehnigc), Reiner Sustmannb),
and Herbert de Groota)
a) Institut für Physiologische Chemie, Universitätsklinikum, Hufelandstrasse 55, D-45122 Essen
(e-mail: michael.kirsch@uni-duisburg-essen.de)
b) Institut für Organische Chemie, Universität Duisburg-Essen, Universitätsstrasse 5, D-45117 Essen
(e-mail: hans-gert.korth@uni-duisburg-essen.de)
c) Fachbereich Chemie, Universität Dortmund, Otto Hahn Strasse 6, D-44221 Dortmund
In memoriam Hanns Fischer
The reaction of peroxynitrite/peroxynitrous acid with morpholine as a model compound for secon-
dary amines is reinvestigated in the absence and presence of carbon dioxide. The concentration- and pH-
dependent formation of N-nitrosomorpholine and N-nitromorpholine as reported in three previous
papers ([25][26][14]) is basically confirmed. However, 13C-NMR spectroscopic product analysis shows
that, in the absence of CO2, N-hydroxymorpholine is, at pHꢀ7, the major product of this reaction,
even under anaerobic conditions. The formation of N-hydroxymorpholine has been overlooked in the
three cited papers. Additional (ring-opened) oxidation products of morpholine are also detected. The
data account for radical pathways for the formation of these products via intermediate morpholine-
derived aminyl and a-aminoalkyl radicals. This is further supported by EPR-spectrometric detection
of morpholine-derived nitroxide radicals, i.e., morpholin-4-yloxy radicals. N-Nitrosomorpholine, how-
ever, is very likely formed by electrophilic attack of peroxynitrite-derived N2O4. 15N-CIDNP Experi-
ments establish that, in the presence of CO2, N-nitro- and C-nitromorpholine are generated by radical
recombination. The present results are in full accord with a fractional (28ꢁ2%) homolytic decay of per-
oxynitrite/peroxynitrous acid with release of free hydroxyl and nitrogen dioxide radicals.
1. Introduction. – Since the seminal paper of Beckman and colleagues [1] in which
the formation of peroxynitrite/peroxynitrous acid (=oxoperoxonitrate(1ꢂ) (ONOOꢂ)/
hydrogen oxoperoxonitrate(1ꢂ) (ONOOH)) has been proposed to occur in living
organisms from the diffusion-controlled reaction of endogenous superoxide radical
anion (O2Cꢂ) and nitric oxide (=nitrogen monooxide; CNO), the chemical and pathophy-
siological characteristics of this putative toxic intermediate have intensely been inves-
tigated. To date, the essential chemical characteristics of peroxynitrite in aqueous sol-
ution have been firmly established, revealing that radical processes, initiated by homo-
lytic OꢂO bond cleavage to produce a fraction of 28ꢁ2% free (cage-escaped) hydroxyl
and nitrogen dioxide radicals (Eqn. 1), play a major role for its chemical reactivity [2–6].
In the absence of any substrates, nitrate is by far the major product from peroxyni-
trite at acidic pH, but at pH>6 substantial amounts of molecular oxygen and nitrite are
being produced in competition to nitrate. We have recently demonstrated that the for-
mation of these products and the underlying kinetics in the full pH range 1–14 is
ꢀ 2006 Verlag Helvetica Chimica Acta AG, Zürich

2400
Helvetica Chimica Acta – Vol. 89 (2006)
related to a complex reaction cascade following the initial release of the highly reactive
HOC and NO2C radicals [4]. In the presence of carbon dioxide, the reactivity of peroxy-
ꢂ1
A
M
sꢂ1) reaction with CO2 to
produce NO2C and the carbonate radical anion, CO3Cꢂ, at a fraction of ca. 30–33%
[7–9]. In accord with the release of strongly oxidizing radicals, peroxynitrite-mediated
oxidation of organic and inorganic substrates mainly follows radical pathways.
The pathophysiological relevance of endogenously produced peroxynitrite is still a
matter of debate. For the elucidation of possible biochemical pathways of peroxynitrite,
a variety of model studies have been performed in the past. Reaction with typical mod-
els for biological compounds preferably leads to oxygenation (hydroxylation), nitra-
tion, and nitrosation [6][10–16]. The yields of such products, however, are generally
rather low, very rarely as high as 30% relative to the amount of applied peroxynitrite.
Noteworthy, yields of ca. 8 and 13% (ꢁmagic yieldsꢂ) seem to be very characteristic for a
broad variety of reactants (see [4] for a collection of related references). Significantly
lower yields (ca. 1–2%) have been observed for the peroxynitrite-dependent nitrosa-
tion of thiols [17–19] and of indole derivatives [20][21].
The peroxynitrite-induced nitrosation of 1,2-phenylenediamine (=benzene-1,2-dia-
mine) has been contradictorily explained by an NO+ activity of peroxynitrous acid [22]
and by a homolytic pathway [23]. We have shown that peroxynitrite oxidizes tertiary
amines in an initial one-electron step to generate the corresponding amine radical cat-
ions, a process which strongly indicates the intermediacy of HOC radicals [24]. In three
recent studies, the reaction of peroxynitrite with secondary amines – by which cancero-
genic N-nitrosamines may be produced – has been probed with the model compound
morpholine (MorH) [14][25][26]. In these studies, N-nitrosomorpholine (MorNO)
and N-nitromorpholine (MorNO2) have been reported as the sole products, produced
in a pH-dependent manner, and with their yields (relative to peroxynitrite) in the range
of the above mentioned ꢁmagic yieldsꢂ of 8 and 13%. The formation of MorNO and
MorNO2 was uniformly interpreted in terms of radical processes, following attack of
HOC on morpholine to generate via NꢂH H-atom transfer the corresponding aminyl
radical (MorC).
However, as HOC-mediated H-abstraction from the methylene groups of MorH is
also a viable process (see Sect. 3), and as above pH 6 substantial amounts (up to
30% at pH 9–10) of O2 are being produced during decomposition of peroxynitrite,
one would also expect the formation of N- and/or C-oxygenated (e.g., hydroxylated)
products because the reaction of both aminyl and C-centered radicals with O2 is essen-
tially diffusion-controlled (see below). Surprisingly, in none of the three above-cited
studies, morpholine-based products other than MorNO and MorNO2, e.g., deriving
from C-centered morpholine radicals or from further transformation of initially pro-
duced MorNO/MorNO2, have been reported.

Helvetica Chimica Acta – Vol. 89 (2006)
2401
In the course of our studies on the reactivity of peroxynitrite with ubiquitous bio-
molecules [12][27–29], we found evidence that from secondary amines also products
other than N-nitrosated or N-nitrated compounds were formed. We, therefore, rein-
vestigated the reaction of morpholine with peroxynitrite more closely by means of
13C- and 15N-NMR spectroscopy. In the present study, we show that in the presence
of atmospheric or peroxynitrite-released O2, N-hydroxymorpholine (MorOH) is a
major product of this reaction, accompanied by a variety of minor oxidation products
of morpholine. Detection of the intermediate morpholine-derived nitroxide radical
MorOC by EPR spectroscopy and 15N-CIDNP-NMR experiments further supports a
radical pathway of peroxynitrite-mediated oxidations of secondary amines in the
absence and presence of carbon dioxide.
2. Results. – Carbon Dioxide-Free Conditions. Formation of N-Nitrosomorpholine
and N-Nitromorpholine. In previous studies on peroxynitrite-mediated nitrosation
and nitration of morpholine (MorH), both reactants were either applied in equimolar
amounts or moderate concentration ratios not exceeding tenfold excesses of the one
compound over the other [14][25][26]. However, although the rates of reaction of
HOC with secondary amines and cyclic ethers are close to the diffusion-controlled
limit (k=109 –1010
M
ꢂ1 sꢂ1; see the numerous data in the NIST database [30]), effective
trapping of peroxynitrite-released HOC radicals would require higher concentrations of
MorH because HOC attack on its own precursor peroxynitrite is also a diffusion-control-
led process [k(HOC+ONOOꢂ)=4.8·109 ꢂ1 sꢂ1][31], i.e., may compete efficiently with
M
attack on MorH at low [MorH]/[peroxynitrite] ratios. Therefore, in our experiments, we
varied the [MorH]/[peroxynitrite] ratio up to a limiting value of 300 :1 (see below). N-
Nitrosomorpholine (MorNO) and N-nitromorpholine (MorNO2) were initially identi-
fied by GC/MS, capillary-zone electrophoresis (CE), and thin-layer chromatography
(TLC) by comparison with authentic material. Quantification was performed by CE.
The detection limits of our CE instrument were ca. 2.5 mM for MorNO and ca. 15 mM
for MorNO2.
At a fixed concentration of 1 mM peroxynitrite, MorNO formation at pH 7.5 and 378
could not be detected by CE at MorH concentrations lower than 5 mM (Fig. 1); about
25 mM MorNO was found in the presence of 10 mM MorH, the formation of MorNO
increased exponentially with increasing morpholine concentration, approaching a pla-
teau value of ca. 77 mM at 300 mM morpholine.
At higher concentrations of peroxynitrite (ꢀ2 mM), formation of MorNO2 was also
detected. At pH 7.5, production of both MorNO and MorNO2 is strictly proportional
(r2 =0.998) to the amount (1–10 mM) of added peroxynitrite (Fig. 2). The relative
yields of MorNO and MorNO2, as derived from the slope of the regression line, corre-
sponded to ca. 7.2 and 1.6 mol-%, respectively, of the initial amount of peroxynitrite.
The latter value agrees fully with the maximum yield of MorNO2 reported by Uppu,
Squadrito, and co-workers for 2 mM peroxynitrite and 5 mM MorH at pH 9 [25]. It
must be emphasized that under the applied conditions, MorNO2 is not produced
from further oxidation of MorNO by peroxynitrite, as revealed by blank experiments
where authentic MorNO (0.1–1 mM) was allowed to react with excess peroxynitrite
(1–10 mM) (data not shown).

2402
Helvetica Chimica Acta – Vol. 89 (2006)
Fig. 1. Dependence of N-nitrosomorpholine production from peroxynitrite (1 mM) on the concentration
ofmorpholine at pH 7.5 . [Phosphate buffer]=50 mM; T 378. The solid line is a fit of a single expo-
nential to the data points.
As reported previously, the production of both MorNO and MorNO2 is strongly
pH-dependent [14][25][26]. At pH 6, formation of MorNO from 300 mM MorH and
1 mM peroxynitrite was undetectable by CE (Fig. 3), i.e., ꢃ2.5 mM. With increasing
pH, MorNO formation increased to reach a maximum value of ca. 140 mM (14%) at
pH 10. Above this pH, the yield of MorNO decreased slightly. The relative maximum
yield at pH 10 is close to the maximum yield of 12% at 258 given by Uppu, Squadrito,
and co-workers [25] for a 2.5 :1 [MorH]/[peroxynitrite] ratio.
Differently, Ohshima and co-workers [26] reported only a MorNO yield of 5% at
pH 9 from an equimolar (1 mM) ratio of the reactants. Therefore, the dependence of
MorNO formation on the concentration of MorH was also determined at pH 11 (Fig.
4). At 1 mM equimolar concentrations, a yield of ca. 85 mM (8.5%) of MorNO was
found, which increased to a maximum value of ca. 138 mM (ca. 14%) at 50 mM morpho-
line. At the limiting concentration of 300 mM MorH, the yield of MorNO had slightly
decreased to 123ꢁ9 mM (ca. 12%).
Since at high MorH concentrations (ꢀ10 mM), our solutions cannot be considered
to be true phosphate buffer systems, but rather represent amine/ammonium buffers, we
performed exploratory experiments to check whether MorNO formation would be
affected by the concentration of the phosphate buffer (Fig. 5). Clearly, MorNO produc-
tion from 1 mM peroxynitrite and 200 mM MorH is virtually independent on phosphate
buffer concentration up to 250 mM, but drops to about one-half at 750 mM. This effect
yet has to be explored.
Ohshima and co-workers [26] have shown for a variety of secondary amines that N-
nitrosation by peroxynitrite seems to be a general process and that the yields of N-

Helvetica Chimica Acta – Vol. 89 (2006)
2403
Fig. 2. Dependence of N-nitrosomorpholine (upper and middle trace) and N-nitromorpholine produc-
tion (lower trace) from 300 mM morpholine on the concentration ofperoxynitrite . [Phosphate buf-
fer]=50 mM; T 378. The middle trace is the expansion of the MorNO data points at low peroxynitrite
concentrations. Open circles: solutions treated with DTPA (=diethylenetriaminepentaacetic acid=
N,N-bis{2-[bis(carboxymethyl)amino]ethylglycine); open triangles: solutions treated with Chelex.
Fig. 3. pH-Dependence of N-nitrosomorpholine production from 1 mM peroxynitrite and 300 mM mor-
pholine. [Phosphate buffer]=50 mM; T 378.

2404
Helvetica Chimica Acta – Vol. 89 (2006)
Fig. 4. Dependence of N-nitrosomorpholine production from 1 mM peroxynitrite on the concentration
ofmorpholine at pH 11. T 378.
Fig. 5. Effect of phosphate buffer concentration on N-nitrosomorpholine formation from 1 mM peroxy-
nitrite and 200 mM morpholine. At pH 7.5, T 378.
nitrosoamines and N-nitroamines depend on the pKa of the corresponding ammonium
ion. This was confirmed by reaction of buffered solutions of ꢁdiethanolamineꢂ (=2,2’-
iminobis[ethanol]), piperidine, and pyrrolidine (300 mM each) with 1 mM peroxynitrite.

Helvetica Chimica Acta – Vol. 89 (2006)
2405
At pH 7.5, out of these amines, only diethanolamine gave yields of the related N-nitro-
soamine (14.3ꢁ1.9 mM, 1.4%) exceeding our detection limit. At pH 10, where morpho-
line (pKa =8.36) gave ca. 13.5% MorNO, the N-nitrosoamines of diethanolamine,
A
piperidine, and pyrrolidine were produced in yields of 5.9, 0.5, and 0.5%, respectively,
from 300 mM of the amine and 10 mM peroxynitrite (Table 1). The yields of the N-nitro-
soamines correlate with the pKa values of the corresponding ammonium ions of the
amines (the plot of log[N-nitrosoamine] vs. pKa is reasonably linear), showing that
access to the unshared electron pair at the N-atom is essential for the nitrosation reac-
tion. Note that our yields of MorNO and N-nitrosopyrrolidine were about twice as high
as the yields reported by Ohshima and co-workers [26] for equimolar ratios of the reac-
tants.
Table 1. N-Nitrosoamine Production from Reaction (3 h) ofPeroxynitrite (10 mM) with Various Secondary
Amines (300 mM each) in Potassium Phosphate Buffer (50 mM; pH 10) at 378
+
Amine
N-Nitrosoamine [mM]a)
pKa
(R2
NH2
)
Morpholine
Diethanolamineb)
Piperidine
1345ꢁ38 (13.5)
588ꢁ15 (5.9)
52ꢁ2 (0.5)
8.36
8.88
11.22
11.27
Pyrrolidine
49ꢁ1 (0.5)
a) Percentage yields relative to applied peroxynitrite given in parentheses. b) Systematic name: 2,2’-imi-
nobis[ethanol].
13C-NMR Experiments. 13C-NMR Spectra were recorded after rapid mixing at 208 of
deoxygenated and CO2-free solutions of MorH (300 mM) in phosphate buffer pH 7.5
with alkaline solutions of peroxynitrite to give a 10 mM final concentration. (The
[MorH]/[peroxynitrite] ratio 30 :1 was chosen to avoid complete suppression of peroxy-
A
mixing revealed the formation of a variety of products having resonances exclusively
in the d 40–70 and 160–170 ppm range (Fig. 6,a).
In accord with the short lifetime of peroxynitrite at this pH (t1/2 ꢄ4 s), product for-
mation is essentially complete when the first spectrum can be recorded. By comparison
(spiking) with the mixture of authentic material (see Exper. Part), MorNO2, MorNO,
and MorOH were identified as products (Table 2), formed in an approximate
1:2.7:3.6 molar ratio. Hence, MorOH is in fact the major product from peroxynitrite
and MorH under these conditions. The MorNO/MorNO2 ratio corresponded reasona-
bly well with the 4.4 :1 ratio as determined by CE for a reaction temperature of 378. By-
products of similar signal intensity had resonances at d 40.0, 54.3, 62.2, 65.3, and 170,
minor signals were detected at d 39.6, 45.5, 48.5, 59.4, 66.1, 66.3, and 162.1. A DEPT
spectrum showed that all resonances in the d 30–70 range are due to CH2 groups, estab-
lishing that no significant amounts of stable mono-substituted oxidation products (e.g.
C-nitro, C-nitroso, or C-hydroxy compounds) were formed. On partial decoupling, the
signals at d 162 and 170 were both split into doublets, i.e., they were certainly due to
aldehyde groups.
From the relative intensity of the sum of the d 35–70 resonances, a MorH conver-
sion of roughly 8% (the ꢁmagic yieldꢂ) was estimated, in reasonable agreement with the

2406
Helvetica Chimica Acta – Vol. 89 (2006)
Fig. 6. 13C-NMR Spectra of products from reaction of peroxynitrite with morpholine in phosphate buf-
fer at a) pH 9.0, b) pH 7.4, and c) pH 4.8. [Peroxynitrite]=10 mM, [MorH]=300 mM, [phosphate buf-
fer]=50 mM; T 258. Spectra are sums of four spectra recorded in 15-min intervalls after mixing of the
reactants. Labels: a=MorOH, b=MorNO, c=MorNO2.

Helvetica Chimica Acta – Vol. 89 (2006)
2407
Table 2. 13C-NMR Chemical-Shift Data (rel. to SiMe4 (=0 ppm))
Position
d [ppm]
Acetate buffer Phosphate buffer Phosphate buffer CD3
CNc)
G
pH 4.5a)
pH 7.5b)
pH 9.0b)
MorH (R=H)
C(3), C(5) 43.1
C(2), C(6) 63.4
43.1
63.5
49.9
66.2
64.8
41.0
46.6
63.3
56.9
65.0
43.1
65.0
49.1
65.4
64.6
40.0
46.8
63.6
56.2
64.2
44.2
64.5
50.6
66.8
66.2
41.0
49.3
66.9
59.6
65.7
MorNO (R=NO)
C(3)d)
C(2)d)
C(6)e)
C(5)e)
50.0
66.5
65.1
41.6
MorNO2 (R=NO2) C(2), C(6) 46.3
C(3), C(5) n.r.f)
C(3), C(5) 57.7
C(2), C(6) 64.8
MorOH (R=OH)
a) From the NaNO2 +H2O2 reaction. b) From the reaction with preformed peroxynitrite. c) 15N-labelled
ꢂ
compounds, from electrolysis of 15NO2 . d) anti to N=O; J(C,N)=5.6. e) syn to N=O. f ) Not resolved, cov-
ered by MorH resonance.
yields of MorNO/MorNO2 determined by CE (cf. Fig. 2). The relative intensity of all
NMR signals did not vary noticeably within 2 h, thus, further transformation of the
products, e.g., by hydrolysis, is slow. No further attempts were made to elucidate the
structures of the by-products, but the absence of additional 15N-NMR resonances
ꢂ
other than those of Mor15NO, Mor15NO2, 15NO2ꢂ , and 15NO3 , when this experiment
was performed with (15N)peroxynitrite (data not shown), ruled out significant incorpo-
ration of peroxynitrite-derived N-atoms.
At pH 9, product formation was complete after 15 min, where the overall conver-
sion of MorH had increased to ca. 12%, and the relative yields of MorOH, MorNO,
and MorNO2 had changed in favor of MorOH to an approximate ratio of 4.4 :2.2 :1
(Fig. 6,b). The enhanced production of MorOH compared to MorNO, and MorNO2
is in line with the fact that at this pH substantial amounts of O2 are produced from
peroxynitrite. Noticeably, less by-products than at pH 7.5 were formed, major peaks
A
were detected at d 39.2, 44.8, 64.0, and 169.6. In addition to the d 162 and 170 aldehyde
resonances already observed at pH 7.5, further carbonyl resonances were found at
d 163.3 and 166.3.
Because at pH<6 peroxynitrite decomposes completely within a fꢂew seconds, per-
oxynitrite was generated at pH 4.8 in situ from the reaction of 15NO2 with hydrogen
peroxide in acetate buffer in the presence of 300 mM MorH. The 13C-NMR spectra
recorded immediately after mixing of the reactants revealed a low conversion of
MorH. A plethora of new resonances was observed (Fig. 6,c). From the relative signal
intensities of the CH2 groups in the d 30–70 range, a total conversion of MorH of ca.
3% after 1 h was estimated, with MorNO being the major product (ca. 1.2%).
MorOH was formed to a smaller extent (about one-fourth of [MorNO]), in agreement
with the low production of oxygen at this pH [4]. MorNO2 could not be detected. This
was confirmed by the 15N-NMR spectrum, in which only MorH (in natural abundance
of 15N), 15NO3 , and Mor15NO were detected (data not shown). The Mor15NO signal
ꢂ

2408
Helvetica Chimica Acta – Vol. 89 (2006)
grew slowly with time, indicating additional nitrosation of MorH via the common elec-
trophilic HNO2/N2O3 and/or HOONO2 pathways.
Effect of Hydroxyl-Radical Scavengers. To substantiate the involvement of HOC rad-
icals in the N-nitrosation process, the influence of compounds commonly regarded as
typical hydroxyl-radical scavengers, namely benzoic acid, mannitol, ethanol, and
DMSO (200 mM each) on reaction of peroxynitrite (1 mM) with morpholine (50 mM)
was studied (Table 3). As can be seen, formation of MorNO was partially reduced in
the presence of these compounds, with DMSO being the most effective trap.
Table 3. Effect of Hydroxyl-Radical Scavengers (200 mM each) on the Production of N-Nitrosomorpho-
line from Reaction of Peroxynitrite (1 mM) with Morpholine (50 mM) in Potassium Phosphate Buffer (50
mM, pH 7.5) at 378
ꢂ1
Scavenger
N-Nitrosoamine [mM]
DYield [%]
k(HOC+scavenger)a) [109
M
sꢂ1
]
None
64.9ꢁ3.9
55.1ꢁ2.7
47.2ꢁ4.3
40.5ꢁ3.9
38.3ꢁ2.3
–
Benzoic acid
Ethanol
Mannitol
DMSO
ꢂ15
ꢂ27
ꢂ38
ꢂ41
3.30
1.80
1.85
7.0
a) Data from [30].
Influence of Azide on N-Nitrosomorpholine Formation. The influence of the azide
ion on the formation of nitrosated products is commonly used to underline the action of
N2O3 as nitrosating species ([32] and ref. cit. therein). In our system (1 mM peroxy-
nitrite, 50 mM MorH), azide concentrations in the millimolar range had to be applied
to reduce the peroxynitrite-dependent MorNO production significantly (Fig. 7). The
yield of MorNO was diminished roughly linearly (r2 =0.989) with increasing azide con-
centration and was completely abolished at ꢀ5 mM N3ꢂ. Goldstein and Czapski [33]
reported that azide reacts with peroxynitrite in the presence of CO2. To verify whether
a similar reaction would occur also in the absence of CO2, the yields of nitriteꢂand ni-
trate under the above conditions were determined. In view of the fact that N3 reactꢂs
with both N2O3 and N2O4 (Eqns. 2 and 3) [32] and assuming that production of NO2
from peroxynitrite proceeds ꢂvia hydrolysis of intermediately formed N2O3 (Eqn. 4)
[4], the decrease of the NO2 yield was expected to correlate with the consumption
ꢂ
ꢂ
of N3ꢂ. However, contrary to this expectation, formation of NO3 but not of NO2
was found to be diminished by added azide (Fig. 8). Rather, at lower azide concentra-
tions, nitrite production was even slightly enhanced. Noteworthy, the combined
ꢂ
NO2ꢂ +NO3 yield was continuously diminished to a limiting percentage of 17.5% at
high concentration of N3ꢂ. This observation strongly suggests a partial conversion of
peroxynitrite to nonionic/nonhydrolyzable N-products, most reasonably N2 and/or
N2O. This result favors the view that the HOC radicals produced from peroxynitrite
had reacted with N3ꢂ (k(HOC+N3 )=1.2·1010
M
sꢂ1) [30] (see Sect. 3).
ꢂ
ꢂ1
ꢂ
N3ꢂ +N2O3 ! N2 +N2O+NO2
(2)
(3)
N3ꢂ +N2O4 ! N2 +N2O+NO3
ꢂ

Helvetica Chimica Acta – Vol. 89 (2006)
2409
Fig. 7. Effect of azide on N-nitrosomorpholine formation from 1 mM peroxynitrite and 300 mM mor-
pholine at pH 7.5. T 378.
N2O3 +H2O ! 2 NO2ꢂ +2 H+
(4)
Formation ofHydrogen Peroxide . Nitrosation of secondary amines by peroxynitrite
might, in principle, be expected to proceed also via a direct, nonradical (electrophilic)
pathway. In such a case, production of H2O2 is expected, which should correlate with
the production of MorNO in a 1:1 stoichiometry (Eqn. 5).
R
A
G
(5)
The base level of H2O2 in our peroxynitrite stock solution was found to be 4.2ꢁ0.3
mM (0.58 mol-%) at all selected pH values (data not shown). After reaction of peroxy-
nitrite (1 mM) with morpholine (300 mM) in the pH range 5–10, a significant increase of
the H2O2 level between pH 5 and 8 was found, with a maximum yield of ca. 40 mM (4%
relative to peroxynitrite, corrected for nitrite base level) at pH 6 (Fig. 9). From inspec-
tion of Fig. 9, it becomes obvious that the yields of MorNO and H2O2 are uncorrelated,
which, in combination with the concentration dependence of MorNO formation (see
above), largely rules out the contribution of an electrophilic nitrosation process. Note-
worthy, at pH 10, peroxynitrite-mediated formation of H2O2 was completely abolished.
Formation ofNitric Oxide . By means of our FNOCT methodology [34–36]
(FNOCT=fluorescent nitric oxide cheletropic trap), an accumulated production of
20.4ꢁ3.3 mM NOC was detected from decay of 5 mM peroxynitrite at pH 7.5, corre-
sponding to a yield of 0.4%. It should be noted, however, that at this pH ca. 17% of
O2 is released from peroxynitrite [4], which might have masked a higher production
of NOC due to competitive autoxidation to CNO2.

2410
Helvetica Chimica Acta – Vol. 89 (2006)
Fig. 8. Effect of azide on the formation of nitrite and nitrate from 1.2 mM peroxynitrite at pH 7.5.
T 378.
&
*
Fig. 9. pH Dependence ofhydrogen peroxide
( ) and MorNO ( ) formation from decomposition of
1 mM peroxynitrite in the presence of300 m M morpholine. Values corrected for 4.2 mM base level of
H2O2. T 378.
EPR Spectroscopic Experiments. The foregoing results confirm the view
[14][25][26] that product formation is initiated by attack of peroxynitrite-released

Helvetica Chimica Acta – Vol. 89 (2006)
2411
HOC radicals on MorH to produce the corresponding N- and/or C-centered radicals as
intermediates (see Sect. 3). Direct EPR-spectroscopic detection of such radicals is ham-
pered by their high reactivity (their bimolecular self-decay can safely be assumed to be
diffusion-controlled [37]) in relation to the relatively slow rate of formation. Most pre-
vious attempts to detect radical intermediates from decaying peroxynitrite, therefore,
utilized EPR spin-trapping techniques [38–40]. Although radical production from per-
oxynitrite was in fact indicated in such experiments, ambiguity generally remained with
regard to the nature of the initially released radicals. As aminyl, a-aminoalkyl, and a-
alkoxyalkyl radicals all react rapidly (kꢀ109
M
ꢂ1 sꢂ1) [41] with O2, we expected that the
putative morpholine-derived aminyl radical (MorC) would finally produce the corre-
sponding morpholine-derived nitroxide radical (MorOC), similarly to what has been
observed for other secondary amines [42][43]. In fact, after rapid mixing of a solution
of MorH (300 mM) with peroxynitrite (5 mM final concentration) in phosphate buffer at
pH 7.5, the growth of the known EPR spectrum of MorOC [44][45] could be monitored
(Fig. 10,a).
This EPR spectrum shows the interesting feature of an alternating linewidth effect;
i.e., at the applied temperature (208), six of the 15 hyperfine lines expected if the four Hb
atoms are assumed to be equivalent are broadened beyond detection due to a confor-
mational chair-twist interconversion of the six-membered ring at an ꢁintermediateꢂ rate
(k=1.2·108 sꢂ1, according to simulation).
The EPR spectrum could also be recorded, with a reduced signal intensity, under
anaerobic conditions, confirming some release of O2 from peroxynitrite at this pH.
In accord with the increased production of O2 at alkaline pH, a more intense EPR spec-
trum was recorded at pH 10.5. At pH 4.5, where O2 production is low, signals of MorOC
could no longer be detected. When octadeuterated morpholine (D8MorH; 95% D) was
employed in a similar experiment, a spectrum corresponding to an approximate ꢁrapid
exchangeꢂ of the axial and equatorial deuterium atoms on the spectrometerꢂs time scale
was observed, due to the smaller difference of the related hyperfine splittings (Fig.
10,b). The EPR signal intensity of the MorOC radical increased strongly within the
first minute after mixing of the reactants, followed by a slower decay with a half-life
of ca. 2.5 min at 208.
Presence ofCarbon Dioxide. Influence ofHydrogen Carbonateꢂon Product Forma-
tion. The decomposition of peroxynitrite in the presence of HCO3 /CO2 at 2.5 and equi-
molar [MorH]/[peroxynitrite] ratios has been studied by Uppu, Squadrito, and co-ꢂwork-
ers [25], and Ohshima and co-workers [26], respectively, showing that HCO3 /CO2
enhances MorH nitration over nitrosation at pHꢀ7. Therefore, we studied the peroxy-
nitrite-mediated formation of MorNO and MorNO2 at pH 7.5 and 378 only at the limit-
ꢂ
ing value of 300 mM morpholine (Fig. 11). At 1 mM HCO3 , the yield of MorNO was
ꢂ
essentially the same (78 mM) as in the absence of HCOꢂ3 , but formation of MorNO2
had increased to ca. 32 mM. With increasing HCO3 concentration, the yield of
MorNO was slightly (ca. 29%) decreased to a constant level of 55 mM, whereas produc-
tion of MorNO2 was strongly enhanced to reach a plateau value of ca. 103 mM at 25 mM
HCO3ꢂ. Control experiments confirmed that MorNO2 was not formed from further
reaction of MorNO (0.1–1 mM) with peroxynitrite (1–10 mM) in the presence of
ꢂ
HCO3ꢂ (25–50 mM) (data not shown), similar to what was found in the absence of
HCO3 /CO2.

2412
Helvetica Chimica Acta – Vol. 89 (2006)
Fig. 10. EPR Spectra ofthe (D )morpholin-N-yloxy radical (D8MorOC) from decomposition of peroxy-
8
nitrite in phosphate buffer in the presence of (D8)morpholine: a) Experimental spectrum ofmorpholin-
N-yloxy (MorOC) (pH 7.4). b) Simulated exchange-broadened spectrum for a ring-inversion rate of
1.2·108 sꢂ1 (hyperfine splittings: a(N)=1.805, a(Hb,ax)=2.46 and 0.0 (2 H), a(Hb,eq)=0.0 and 2.46 (2
H), and a(Hg)=0.03 (4 H) mT; g=2.00576). c) Experimental spectrum ofthe (D )morpholin-N-yloxy
8
radical (D8MorOC) (pH 10.6). d) Simulated exchange-broadened spectrum for a ring-inversion rate of
8.8·108 sꢂ1 (hyperfine splittings: a(N)=1.801, a(Db,ax)=0.367 and 0.0 (2 D), a(Db,eq)=0.0 and 0.367 (2
D), a(Dg)=0.02 (4 D) mT; g=2.00576). [MorH] or [D8MorH]=300 mM, [peroxynitrite]=5 mM; T 208
ꢂ
Morpholine-Dependent Product Formation. In the presence of 25 mM HCO3 , for-
mation of MorNO2 from the reaction of 1 mM peroxynitrite with morpholine at pH
7.5 could only be detected above 2 mM MorH. MorNO2 production increased exponen-
tially with [MorH] to level off at ca. 95 mM at morpholine concentrations ꢀ100 mM (Fig.
12). MorNO formation was evident only at [MorH] ꢀ10 mM, reaching a maximum
value of ca. 63 mM at 300 mM morpholine. The yields of MorNO2/MorNO were some-
what affected when the solutions were additionally bubbled with air/CO2 95 :5% (Fig.

Helvetica Chimica Acta – Vol. 89 (2006)
2413
Fig. 11. N-Nitroso- and N-nitromorpholine formation from 1 mM peroxynitrite and 300 mM morpholine
at pH 7.5 as a function of added HCO3ꢂ/CO2. T 378.
12). Now, MorNO2 was already detected at 1 mM morpholine, and its maximum yield at
300 mM [MorH] had raised to ca. 150 mM. The yields of MorNO were less affected by
air/CO2 bubbling, levelling out at ca. 55 mM. These differences can be attributed to
the reaction of morpholine with CO2 to give morpholine carbamate [12], by this
means depleting the CO2 level in the solution if no extra CO2 is supplied.
Peroxynitrite-Induced CIDNP Effects. To provide further support for the radical
pathway of peroxynitrite-mediated formation of MorNO and MorNO2, 15N-CIDNP
studies were performed. 15N-CIDNP Signals, i.e., 15N-NMR resonances of reaction
products showing emission or enhanced absorption, unequivocally establish that
these products are formed via radical pairs. Because the kinetic characteristics of per-
oxynitrite decay makes addition of preformed peroxynitrite to morpholine solutions
unsuitable for CIDNP experiments, peroxynitrite was generated in situ at pH 4.5 by
reaction of 15N-labelled nitrite with H2O2. After addition of H2O2 (1M) to a mixture
of Na15NO2 (0.25M), NaHCO3 (0.05M), and morpholine (5M) in phosphate buffer
(0.3M), four new major and one minor transient CIDNP signals were detected in the
15N-NMR spectrum (Fig. 13,a). The signal at d ꢂ29 appeared in emission and has pre-
viously been identified to be due to 15N-peroxynitric acid (O215NOOH) [46–48]. A sec-
ond emission signal at d +11.5 is very characteristic for C-nitroso compounds, hence
tentatively attributed to a C-(15N)nitrated morpholine product. The signals of 15NO3
ꢂ
(d ꢂ1.5) and Mor15NO2 (d ꢂ21) showed enhanced absorption. Analogous effects of
ꢂ
15NO3 , O215NOOH, and C-15NO2 have been observed during the reaction of in situ gen-
erated peroxynitrite with both tyrosine [11] and bovine albumine [49].
On the contrary, 15NO2ꢂ (d 232) and Mor15NO (d 146.7) were CIDNP-silent (signals
not shown), hence, are not products of radical reactions. After completion of the reac-

2414
Helvetica Chimica Acta – Vol. 89 (2006)
Fig. 12. N-Nitroso- and N-nitromorpholine production from 1 mM peroxynitrite in the presence of25
mM HCO3ꢂ/CO2 as a function of the morpholine concentration: with additional flushing with air/CO2
(95 :5%; black symbols) and without additional flushing with air/CO2 (white symbols). T 378.
tion, only 15NO3ꢂ and Mor15NO could be detected at normal gain. At higher gain, small
signals of Mor15NO2 (d ꢂ29) and a C-15NO2 compound (d +11.5) became visible (Fig.
13,b), both corresponding to ca. 1% of all 15N-labelled products. The 15N-NMR reso-
nance of O215NOOH was no longer observed, in accord with its limited lifetime (t1/2
30 min [46]) at this pH. It should be noted, however, that O215NOOH is not a unique
intermediate in the presence of MorH, but rather is formed from further oxidation
of peroxynitrous acid by excess H2O2 [46].
ꢂ
Formation ofNO C and H2O2. In the presence of 25 mM HCO3 /CO2, we were unable
to detect at pH 7.5 any NOC released from 5 mM peroxynitrite, in accord with the effec-
tive trapping of peroxynitrite by CO2. In the absence of morpholine, 4.3ꢁ0.3 mM H2O2
were found after decomposition of 1 mM peroxynitrite at pH 7.5, i.e., the same amount
that was quantified in the absence of CO2. After decomposition in the presence of 300
mM morpholine, the final H2O2 level had slightly increased to 9.3ꢁ0.4 mM. Thus, H2O2
formation was suppressed by ca. 85% compared to the CO2-free situation (see above).
3. Discussion. – Reaction in the Absence ofCarbon Dioxide . The present investiga-
tion confirms the findings from three earlier reports [14][25][26] that both N-nitroso-
(MorNO) and N-nitromorpholine (MorNO2) are produced from the reaction of
peroxy
Nitrosoamine formation was also determined for some other secondary amines, show-
ing that the yields of these products are dependent on pH, i.e., related to the pKas of the
nitrite with morpholine (MorH) as a model compound for secondary amines.
AHCTREUNG
corresponding ammonium ions. However, in marked contrast to the above reports, we
here demonstrate that N-hydroxymorpholine (MorOH) is in fact a major, at pHꢀ7 the
major product from the peroxynitrite–morpholine reaction over a wide range of pH

Helvetica Chimica Acta – Vol. 89 (2006)
2415
Fig. 13. 15N-NMR Spectra from reaction of morpholine with in situ generated peroxynitrite at pH 4.5
in the presence ofHCO ꢂ/CO2, a) recorded 1 min after mixing of the reactants, and b) recorded 1 h
3
after mixing. [MorH]=1.0M, [Na15NO2]=0.25M, [H2O2]=1.0M, [NaHCO3]=0.05M, [phosphate buf-
fer]=300 mM; T 208.
values. With regard to the fact that in the above cited papers product formation has
been discussed in terms of a radical mechanism, with the intermediacy of the N-mor-
pholinyl radical (MorC), it is highly surprising that formation of MorOH has been over-
looked or not even suspected1), because substantial O2 production from peroxynitrite
at pH>6 has been shown before, and the very fast reaction of N- and C-centered rad-
1
)
This statement seems to be true for [26] and [14], whereas Uppu, Squadrito, and co-workers [25] in
fact noted in a footnote that ꢁThere might be minor products (arising from b-scission or other pro-
A

2416
Helvetica Chimica Acta – Vol. 89 (2006)
icals with O2 is common knowledge. Generation of the morpholine-derived nitroxide
radical (MorOC) in this system also has not been reported before. Hence, formation
of oxygenated products must always be expected from reaction of peroxynitrite with
organic substrates, even in the absence of atmospheric O2, and, in fact, has frequently
been observed.
MorOH formation is established here by 13C-NMR spectroscopy (Figs. 6,a–c),
which also indicates that MorOH is a ꢁdirectꢂ product from attack of peroxynitrite on
MorH, i.e., does not derive from further conversion of initially formed MorNO or
MorNO2.
Formation of MorOH/MorOC as well as the detection of a variety of morpholine-
based by-products provides valuable insight into the mechanism by which peroxynitrite
reacts with secondary amines. As noted in Sect. 1, there is overwhelming evidence for
partial (ca. 28%) release of free (cage-escaped) HOC and CNO2 radicals during peroxy-
nitrite decomposition (Eqn. 1). In the present system, the rather low yields of the
detected morpholine-derived products, approaching constant values at high MorH con-
centrations, are fully explained by the combined action of these radicals. At the limiting
MorH concentration of ca. 300 mM, the released HOC radicals should react exclusively
with morpholine. Due to their high reactivity, HOC radicals may attack MorH at various
positions, i.e., by CꢂH and/or NꢂH H-atom abstraction (HAT) and/or by stepwise one-
electron oxidation/deprotonation or proton-coupled electron transfer (PCET) at NH.
To see what kind of intermediate(s) might preferably be produced, we estimated the
thermochemistry of HOC attack on unprotonated and protonated morpholine (Eqns.
6 and 7) by CBS-QB3 calculations. For the estimation of the free energies of reaction
in aqueous solution, free energies of hydration were computed by utilizing the PCM-
UAHF solvation model (Table 4).
C4H8NO+HOC ! C4H7NOC+H2O
(6)
(7)
+
C4
H9
U
The data of Table 4 show that the differences in the reaction energies for the four
homolytic processes fall in a narrow range of just 5 kcal molꢂ1 in the gas phase as
well as in aqueous solution. Since the driving forces are very similar, no reasonable con-
clusion can thus be drawn concerning a preferred pathway. The rates of HOC attack are
Table 4. CBS-QB3-Calculated Thermochemical Data [kcal molꢂ1] for Hydrogen Abstraction from the
Morpholine/Morpholinium Ion by Hydroxyl Radicals (Eqns. 6 and 7).
Cleaved bond
DrH8(g)a)
DrG8(g)a)
DrG8(aq)b)
BDEc)
HꢂN
ꢂ24.9
ꢂ26.6
ꢂ22.1
ꢂ24.1
ꢂ26.1
ꢂ27.9
ꢂ23.4
ꢂ25.2
ꢂ26.9
ꢂ29.0
ꢂ24.9
ꢂ25.8
94.3d)
92.5e)
97.0
HꢂC(3)
HꢂC(2)
HꢂN+(H)
94.9
a) Gas phase. b) Aqueous phase, including hydration free energies calculated with the PCM-UAHF sol-
vation model. c) Gas-phase bond-dissociation enthalpy (DH8). d) Exper.: 92.9 kcal molꢂ1 [73]. e) Exper.:
93.0 kcal molꢂ1 [73].

Helvetica Chimica Acta – Vol. 89 (2006)
2417
kinetically controlled, i.e., may be governed by dissimilar intrinsic activation barriers
for H-transfer, especially for the CH vs. NH moieties. (This issue has been addressed
in a recent paper [50]). Nevertheless, since the largest energy difference is found for
H-abstraction from the N- and O-bound CH2 groups, the 13C-NMR-detected by-prod-
ucts may thus be assumed to derive mainly from attack at the a-position with respect to
the N-atom (i.e., at C(3)) to give an a-aminoalkyl radical. With regard to the NH-
derived products, we note that there is clear evidence that HOC reacts with tertiary
amines by one-electron transfer [24], and this likely happens as well with secondary
amines, i.e., with MorH [51]. The thus formed amine radical cation (MorHC+) would
undergo rapid deprotonation, leading to the same intermediate N-morpholinyl radical
(MorC) as would be produced in a direct HAT process. Noteworthy, the computational
data predict that NH H-abstraction from the morpholinium ion should also be feasible.
However, there are many examples [30] establishing that HOC reacts slower with pro-
tonated amines than it reacts with unprotonated ones (which in most cases react at the
ꢂ1
diffusion-controlled limit, k=109 –1010
M
sꢂ1). Especially, abstraction of a H-atom
from the ammonium group, e.g., as shown in Eqn. 8, is a relatively slow reaction
ꢂ1
(k8 ꢄ1·106
M
sꢂ1) [30].
+
(CH3)2
G
NHC +H2O
(8)
In agreement with this, overall conversion of MorH as well as production of
MorOH is strongly diminished at pH 4.8 and enhanced at pH 9 compared to neutral
conditions (Fig. 6). The lower rates for H-abstraction from protonated amines reflect
the fact that the kinetics of HOC reactions are influenced by pronounced polar effects
(charge separation in the transition state), a consequence of the high electrophilicity
of the HOC radical. In conclusion, bond-dissociation enthalpies alone cannot be used
to predict the (relative) rates of H-abstraction by such kind of radicals [50].
With consideration of the foregiong, the formation of MorOH, MorOC, and the
minor by-products can be fully explained by the steps depicted in the Scheme. Attack
of HOC radicals on MorH either by HAT, PCET, or sequential one-electron transfer/
deprotonation generates the morpholin-N-yl and the morpholin-2-yl radicals. Both rad-
ꢂ1
icals can safely be assumed to react diffusion-controlled (k>109
M
sꢂ1) with O2 (see
[30][41] for leading reports) to yield the corresponding dioxy radicals. Along the
upper pathway of the Scheme, the aminyldioxy radical then converts to the morpho-
lin-N-yloxy radical MorOC. A related aminyldioxy radical from tetramethylpiperidine
has recently been observed, and a mechanism for its conversion to the tetramethylpi-
peridine derived-nitroxide radical has been suggested [43]. It has been proposed that
O2 addition to the aminyl radical is a reversible process so that the aminyl radical
may recombine with the aminyldioxy radical to give a highly unstable peroxide,
which further breaks into two nitroxide radicals. Alternatively, the aminyldioxy radical
may dimerize to an unstable tetroxide, which decomposes into two MorOC and O2. Con-
sidering the known decay mechanisms of nitroxide radicals, MorOH and the corre-
sponding nitrone (=imine N-oxide) then should preferably be formed by disproportio-
nation of MorOC. However, MorOH formation by H-transfer from an activated CꢂH
bond cannot be excluded, since the decay of the EPR signals of MorOC neither follows
a clean second-order nor a clean first-order rate law (data not shown). This reaction

2418
Helvetica Chimica Acta – Vol. 89 (2006)
Scheme
sequence is supported by the observation that the EPR signal of MorOC as well as the
yield of MorOH are both enhanced by raising the pH from 7.5 to 9 (Fig. 6). The puta-
tive nitrone, or, a further reaction product of it, might be responsible for some of the
observed minor CH2 13C-NMR resonances. As peroxides are generally prone to nucle-
ophilic attack by amines, leading to OꢂO bond cleavage, one might suspect that in the
present system MorOH would be formed by such a nonradical process (Eqn. 9). How-
ever, this reaction can be excluded to play a significant role because of the low yield of
MorOH, particularly at alkaline pH where the lifetime of peroxynitrite is strongly
enhanced. Further, at pH 4.8, no additional formation of 15NO2 could be detected
ꢂ
by 15N-NMR-spectrometry from 15N-labelled peroxynitrite. It cannot be excluded, how-
ever, that a fraction of MorOH might be formed by attack of MorC on peroxynitrite
(Eqn. 10).
MorH+ONOOꢂ/ONOOH ! MorOH+NO2 /HNO2
MorC+ONOOꢂ/ONOOH ! MorO(ꢂ)H+NO2C
(9)
ꢂ
(10)
Along the lower reaction path in the Scheme, rapid elimination of superoxide (O2Cꢂ)
from the a-aminodioxy radical is feasible (see [41] for leading reports) to give the cor-
responding cyclic iminium ion/imine, which further might be hydrolyzed to a ring-
opened aldehyde. This prototypal reaction would explain part of the aldehyde-type sig-
nals as well as some of the CH2 resonances observed in the 13C-NMR spectrum (Fig. 6).
The formation of hydrogen peroxide (Fig. 9) via dismutation of superoxide (Eqn. 11) is
also in line with such a reaction.

Helvetica Chimica Acta – Vol. 89 (2006)
2419
2 O2Cꢂ +2 H+ ! H2O2 +O2
(11)
Of course, in an oxidizing environment such as the present one, the reactions out-
lined in the Scheme cannot fully describe all peroxynitrite-dependent alterations on
morpholine, and a variety of side reactions, e.g., starting from the depicted radical inter-
mediates, are expected. The several observed carbonyl 13C-NMR resonances strongly
indicate that oxidative opening of the morpholine ring proceeds at other positions,
too. Additional, ring-opened products may also be expected from oxidative attack on
the ꢁearlyꢂ products, e.g., MorNO [52]. Noteworthy, no monosubstituted morpholine-
derived products were detected in the 13C-NMR spectra, indicating that such com-
pounds are rapidly further oxidized and/or easily decomposed by hydrolysis [52].
With regard to the proposed intermediacy of MorC, the formation of MorNO has
been explained by Uppu, Squadrito, and co-workers [25], and Ohshima and co-workers
[26] by the recombination of peroxynitrite-released nitric oxide with MorC. However,
the fact that the detectable production of CNO from peroxynitrite (in the absence of
any other substrate) is only 0.4% renders the formation of MorNO by this recombina-
tion reaction highly unlikely. Alternatively, one might, in analogy to Eqn. 9, propose
that peroxynitrite would act as a direct, electrophilic nitrosating agent (Eqn. 5).
In fact, van der Vliet and co-workers [19] speculated that the peroxynitrite-depen-
dent formation of S-nitrosoglutathione occurs by such a process, although no satisfying
evidence has been presented. However, the limiting, low yields of MorNO (Figs. 1 and
4) at high MorH concentrations in conjunction with the observation that there is no cor-
relation between formation of MorNO and of H2O2 (Fig. 9) safely rules out that
MorNO is produced via such a process. Rather, the intermediacy of a peroxynitrite-
derived nitrosating species, which can be completely trapped at high morpholine conꢂ-
centrations, is indicated. Of course, at pH 4.5, either N2O3 or HNO2, formed from NO2
are expected to be the major nitrosating intermediates. At pHꢀ7, where the concen-
tration of NO2ꢂ (ca. 150 mM) is given by the contamination (15 mol-%) of the peroxy-
nitrite stock solution, the formation of these nitrosating species can be ignored. Fur-
thermore, the inhibitory effect of azide on the peroxynitrite-mediated formation of
MorNO (Fig. 7) cannot solely be related to the scavenging of N2O3, because low
azide concentrations even slightly increase nitrite production from peroxynitrite (Fig.
8). In addition, the significant nitrosation of morpholine at pH 11 (Fig. 4) disproves
both HNO2 and N2O3 as major nitrosating species, because the latter entity is rapidly
terminated by OHꢂ (k(N2O3 +OHꢂ)=2·103 +108 [OHꢂ] sꢂ1 [53]). The inhibitory effect
of azide can be reasonably explained by the reaction sequence of Eqns. 12–16.
ONOOH ! COH+CNO2
N3ꢂ +COH ! OHꢂ +CN3
(12)
(13)
(14)
(15)
(16)
CN3 +ONOOꢂ ! N2 +N2O+O2Cꢂ
O2Cꢂ +CNO2 ! O2
A
ꢂ
O
2NOOꢂ ! O2 +NO2

2420
Helvetica Chimica Acta – Vol. 89 (2006)
According to this sequence, azide is expected to down-regulate the concentration of
nitrogen dioxide which exists in equilibrium with its dimer N2O4, a potent nitrosating
species [32][54][55]. The radical terminating capabilities of azide in the peroxyni-
trite/CO2/MorH system has previously been emphasized by Uppu, Squadrito, and co-
ꢂ
workers [25], who stated that ꢁ…N3 acts as a general free-radical quencher rather
than as a specific scavenger of NO+ carriersꢂ. N2O4 is well-known to N-nitrosate secon-
dary aliphatic amines even at alkaline pH [56]. As peroxynitrous acid in part fragmen-
tizes to hydroxyl radical and nitrogen dioxide (Eqn. 1), N2O4 will be produced. Thus,
N2O4 should be the major nitrosating entity in our system. This conclusion is strongly
supported by the effect of azide on the yield of nitrate and nitrite (Fig. 8) as well as
by the fact that the [MorNO]/[MorNO2] product ratio of 4–5 is typical for the action
of N2O4 [57]. Thus, formation of carcinogenic N-nitrosoamines must generally be
expected when peroxynitrite decays in the presence of secondary amines.
In the absence of CO2, the yield of MorNO2 is low (1.6 mol-% relative to applied
peroxynitrite). We found no evidence for its formation via recombination of MorC with
NO2C. Instead, it was observed that the yield of MorNO2 increases at alkaline pH (Fig.
6) at the expense of both MorNO (Fig. 3) and H2O2 (Fig. 9). Therefore, we propose that
part of MorNO2 is generated from an electrophilic attack of N2O4 on morpholine in
addition to oxidation of MorNO with H2O2 [58].
Reaction in the Presence ofCarbon Dioxide . In the presence of CO2, peroxynitrite
generates carbonate radicals (CO3Cꢂ) with a yield of ca. 30–33% at the expense of COH
(Eqn. 17) [7–9]. As a consequence of this reaction, the production of both NO2ꢂ and O2
from peroxynitrite is strongly decreased because the HOC can no longer attack ONOOꢂ
effectively. The CO3Cꢂ radical (E8(CO3Cꢂ/CO32ꢂ =1.5 V) [59] is, as compared to HOC
(E8(HOC/OHꢂ =1.9 V, E8(H+,COH/OH2 =2.72 V) [59], a weaker oxidant but is never-
theless expected to oxidize morpholine (Eqn. 18) (e.g., CO3Cꢂ reacts moderately fast
ꢂ1
with piperazine (k=3·106
M
sꢂ1) [30]), thereby also producing the aminyl and the
a-aminoalkyl radicals similarly to the paths shown in the Scheme for HOC. The decisive
difference to the CO2-free situation lies in the fact that peroxynitrite does not release
significant amounts of O2 in the presence of CO2 [4], so that now NO2C can effectively
compete with the dissolved atmospheric O2 (if any) for the morpholine-derived radi-
cals. The latter assumption is supported by the fact that the yield of H2O2 from the per-
ꢂ
oxynitrite–morpholine reaction is reduced by ca. 85% on adding HCO3 /CO2 to the
reaction mixture and, of course, by the emission characteristic of the resonance of 3-
nitromorpholine in the 15N-NMR CIDNP spectrum (Fig. 13). Recombination of the
C-centered morpholinyl radical with freely diffusing NO2C builds up a so-called F-
pair which collapses to 3-nitromorpholine producing an emission (E) signal (Eqn. 19).
CO3Cꢂ +MorH ! MorHC+ +CO3
(18)
2ꢂ

Helvetica Chimica Acta – Vol. 89 (2006)
2421
C15NO2 +CC(3)Mor ! [O215NC;CC(3)Mor]F ! 3-15NO2MorH (E)
(19)
Most interestingly, the resonance of MorNO2 exhibits enhanced absorption (A)
during reaction, and this CIDNP effect cannot be produced by recombination of freely
diffusing NO2C radicals with MorC radicals. Since the pKa value of the morpholinium ion
is 8.5 at 258 [60] and because the pKa value of nonaromatic aminium radical ions
+
(R2
G
+
ions (R
2NH2 ) [61][62], a pKa value in the range of 5.5–6.5 is expected for MorHC+.
Since the CIDNP experiments had to be performed with in situ generated peroxynitrite
at pH 4.5, MorHC+ rather than MorC should be the free N-centered radical that reacts
with NO2C. As aliphatic aminium ions are highly reactive in H-transfer processes
[62], their lifetime should be extremely short at the applied high morpholine concen-
+
tration of 300 mM. Thus, recombination of MorHC with nitrogen dioxide is strongly dis-
favored. Even though, such a reaction would lead to an emission (E) signal in the 15N-
NMR spectrum of MorNO2. The observed enhanced absorption (A) signal of MorNO2
thus can be only explained by the occurrence of a cage-substitution reaction, thereby
building up radical S-pairs (Eqns. 20 and 21).
[O215NC;CꢂO3C]S +MorH ! [MorH+C;C15NO2]S +CO3
(20)
(21)
2ꢂ
+
[MorHC ;C15NO2]S ! MorNO2 (A)+H+ (cage reaction)
At pHꢀ7, however, it is expected that the morpholine-derived aminyl radical
recombines with nitrogen dioxide to yield MorNO2. From the yields of MorNO2
and MorNO at such pH values, i.e., 15 and 5.5% relative to peroxynitrite (Fig. 12),
we can conclude that the peroxynitrite-derived yield of free CNO2 must be in the
range of ca. 26%, in good agreement with earlier reports [7–9].
We are grateful to Heinz Bandmann (Universität Duisburg-Essen) for the NMR measurements.
Experimental Part
Materials. Peroxynitrite (ONOOꢂ) stock solns. (710ꢁ20 mM) were prepared as described previously
[24]. MorOH and MorNO2 were prepared by standard procedures [63][58]. All other compounds and
solvents were commercially available. Na15NO2 labelled with 99.3% 15N (Isotec Inc.) was employed.
Care was taken to exclude possible contamination by transition metals by treating the buffer solns.
with the heavy-metal-scavenger resin Chelex-100 [24]. For easy 13C- and 15N-NMR identification of
the peroxynitrite–morpholine reaction products, a mixture of Mor15NO, Mor15NO2, and MorOH was
prepared by electrolysis (Pt electrode, 2 cm2; current density 25–30 mA cmꢂ2; 4 h) of a mixture of
Na15NO2 (1M) and morpholine (1M) in a rapidly stirred biphasic aq. KOH soln. (pH 13)/CH2Cl2 1:1
(v/v) (30 ml) system according to Evtyugin et al. [64]. After separation of the org. layers, the aq. phase
was extracted with CH2Cl2 (3”), the org. phases were combined, and the solvent was evaporated at 08.
13C-NMR Analysis (CD3
(CD
C
G
The spectral assignments are in accord with literature data [65–67].
CE Measurements. MorOH, MorNO, MorNO2, and the nitrosamines of ꢁdiethanolamineꢂ, piperi-
dine, and pyrrolidine were quantified by capillary-zone electrophoresis (CE) on a Beckman-P/ACE-

2422
Helvetica Chimica Acta – Vol. 89 (2006)
5000 apparatus under the following conditions: fused silica capillary (effective length 50 cm, internal
diameter 75 mm), hydrodynamic injection for 5 s, T 308, voltage 20 kV, normal polarity, UV detection
at 254 nm. As electrolyte system, a mixture of 20 mM sodium phosphate and 100 mM sodium dodecyl sul-
fate (pH 6.45) was used. To each sample, 1 mM N-nitrosodimethylamine was added as internal standard.
Nitrate and nitrite were quantified as described previously [4].
NMR Measurements. 13C- and 15N-NMR Spectra: Bruker-DRX-500 spectrometer; at 125.7 and 50.7
MHz, resp. 13C-NMR: acquisition by collecting 16 transients; d(C) rel. to SiMe4 (=0 ppm). 15N-NMR: rel.
to (15N)nitromethane (=0 ppm) as external standard. Reaction of peroxynitrous acid with morpholine
was carried out by adding aliquots of the alkaline peroxynitrite stock soln. under vortexing to 300 mM
solns. of MorH in the buffer soln. to give a final concentration of 10 mM peroxynitrite. D2
as internal lock (H2O/D2O 9 :1), and the mixture was transferred to 5-mm NMR tubes.
A
AHCTREUNG
EPR Measurements. EPR Spectra: Bruker-ESP-300E-X-band spectrometer (Bruker, Rheinstetten,
Germany), equipped with a TM110 wide-bore cavity; at 208. Solns. were prepared from 2 ml of the phos-
phate buffer soln. (pH 7.4) containing morpholine (300 mM). Aliquots of the alkaline (pH 14) stock soln.
of peroxynitrite were added to the morpholine soln. by vortexing under aerobic and anaerobic conditions
to give the desired concentrations. The mixtures were quickly transferred to a 0.4-mm aq.-soln. quartz cell
(Willmad, Buena, N.J., USA). The first spectra were run as fast as possible, i.e., within 1 min after mixing.
Instrument settings: microwave frequency 9.8 GHz, microwave power 20 mW, sweep range 100 G, sweep
time 4 min, and modulation amplitude 1 G, unless otherwise indicated. Standard spectral simulations
were carried out with the WinSim program [68][69], exchanged-broadenend spectra were simulated
with the EPREXN program [70][71].
CIDNP Measurements. The 15N-CIDNP experiments were performed as reported before
[11][49][47]. The mixtures were prepared in 10-mm NMR tubes by adding 1M H2O2 to solns. of
Na15NO2 (0.05M) in H2O/D2
O 9 :1 containing phosphate buffer (0.3M) and NaHCO3 (0.05M). The pH
U
was adjusted with H2SO4 and NaOH. During reaction, the pH decreased by about half a pH unit because
of the formation of nitric acid. The pH values given in Fig. 13 refers to the initial pH after mixing of the
reactants. The tubes were quickly transferred into the probe head of the 15N-NMR spectrometer (Bruker
DPX-300) and locked within 1 min after mixing of the reactants (internal lock: D
2O). The 15N-NMR
G
spectra were then taken by using single pulses with pulse angles of 90o. Chemical shifts are given in d val-
ues relative to (15N)nitromethane dissolved in MeCN as an external reference.
Quantum-Chemical Calculations. Complete Basis Set (CBS-QB3) computations were carried out
with the Gaussian 03 suite of programs [72]. Gibbs free energies of solvation for water were estimated
for the optimized gas-phase geometries with the PCM-UAHF procedure incorporated in Gaussian 03.
Both the PCM/(U)HF/6-31+G(d) and the CBS-QB3 methodology are known to provide estimates
within ꢁchemical accuracyꢂ (ꢁ1 kcal molꢂ1).
REFERENCES
[1] J. S. Beckman, T. W. Beckman, J. Chen, P. A. Marshall, B. A. Freeman, Proc. Natl. Acad. Sci. U.S.A.
1990, 87, 1620.
[2] S. V. Lymar, J. K. Hurst, Chem. Res. Toxicol. 1998, 11, 714.
[3] C. E. Richeson, P. Mulder, V. W. Bowry, K. U. Ingold, J. Am. Chem. Soc. 1998, 120, 7211.
[4] M. Kirsch, H.-G. Korth, A. Wensing, R. Sustmann, H. de Groot, Arch. Biochem. Biophys. 2003, 418,
133.
[5] Y. Zhao, K. N. Houk, L. P. Olson, J. Phys. Chem. A 2004, 108, 5864.
[6] V. P. Tretꢂyakov, L. A. Minꢂko, E. S. Rudakov, Theor. Exp. Chem. 2004, 40, 309.
[7] S. V. Lymar, J. K. Hurst, J. Am. Chem. Soc. 1995, 117, 8867.
[8] S. Goldstein, G. Czapski, J. Lind, G. Merenyi, Chem. Res. Toxicol. 1999, 12, 132.
[9] M. G. Bonini, R. Radi, G. Ferrer-Sueta, A. M. D. C. Ferreira, O. Augusto, J. Biol. Chem. 1999, 274,
10802.
[10] H. Ischiropoulos, L. Zhu, J. Chen, M. Tsai, J. C. Martin, C. D. Smith, J. S. Beckman, Arch. Biochem.
Biophys. 1992, 298, 431.
[11] M. Lehnig, Arch. Biochem. Biophys. 1999, 368, 303.

Helvetica Chimica Acta – Vol. 89 (2006)
2423
[12] M. Kirsch, H.-G. Korth, R. Sustmann, H. de Groot, Chem. Res. Toxicol. 2000, 13, 451.
[13] B. Alvarez, H. Rubbo, M. Kirk, S. Barnes, B. A. Freeman, R. Radi, Chem. Res. Toxicol. 1996, 9, 390.
[14] K. Kikugawa, K. Hiramoto, T. Ohkawa, Biol. Pharm. Bull. 2004, 27, 17.
[15] N. Nonoyama, H. Oshima, C. Shoda, H. Suzuki, Bull. Chem. Soc. Jpn. 2001, 74, 2385.
[16] M. Kirsch, M. Lehnig, Org. Biomol. Chem. 2005, 3, 2085.
[17] M. A. Moro, V. M. Darley-Usmar, D. A. Goodwin, N. G. Read, R. Zamora-Pino, M. Feelisch, M. W.
Radomski, S. Moncada, Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 6702.
[18] B. Mayer, A. Schrammel, P. Klatt, D. Koesling, K. Schmidt, J. Biol. Chem. 1995, 270, 17355.
[19] A. van der Vliet, P. A. Chr. ꢂt Hoen, P. S.-Y. Wong, A. Bast, C. E. Cross, J. Biol. Chem. 1998, 273,
30255.
[20] T. Suzuki, H. F. Mower, M. D. Friesen, I. Gilbert, T. Sawa, H. Ohshima, Free Radic. Biol. Med. 2004,
37, 671.
[21] M. Kirsch, H. de Groot, J. Pineal Res. 2006, 40, 10.
[22] N. Nonoyama, K. Hisatome, C. Shoda, H. Suzuki, Tetrahedron Lett. 1999, 40, 6939.
[23] R. M. Uppu, W. A. Pryor, J. Am. Chem. Soc. 1999, 121, 9738.
[24] M. Kirsch, E. E. Lomonosova, H. G. Korth, R. Sustmann, H. de Groot, J. Biol. Chem. 1998, 273,
12716.
[25] R. M. Uppu, G. L. Squadrito, R. M. Bolzan, W. A. Pryor, J. Am. Chem. Soc. 2000, 122, 6911.
[26] M. Masuda, H. F. Mower, B. Pignatelli, I. Celan, M. D. Friesen, H. Nishino, H. Ohshima, Chem. Res.
Toxicol. 2000, 13, 301.
[27] M. Kirsch, H. de Groot, J. Biol. Chem. 2000, 275, 16702.
[28] M. Kirsch, M. Lehnig, H.-G. Korth, R. Sustmann, H. de Groot, Chem.–Eur. J. 2000, 7, 3313.
[29] M. Kirsch, H. de Groot, FASEB J. 2001, 15, 1569.
[30] A. B. Ross, W. G. Mallard, W. P. Helman, G. Buxton, V. R. E. Huie, P. Neta, ꢁNDRL/NIST Solution
Kinetics Database 3.0ꢂ, NDRL/NIST, Gaithersburg, MD, 1998.
[31] S. Goldstein, A. Saha, S. V. Lymar, G. Czapski, J. Am. Chem. Soc. 1998, 120, 5549.
[32] D. L. H. Williams, ꢁNitrosation Reactions and the Chemistry of Nitric Oxideꢂ, Elsevier, Amsterdam,
2004.
[33] S. Goldstein, G. Czapski, J. Am. Chem. Soc. 1998, 120, 3458.
[34] M. Bätz, H.-G. Korth, P. Meineke, R. Sustmann, Methods Enzymol. 1998, 301, 532.
[35] P. Meineke, U. Rauen, H. de Groot, H.-G. Korth, R. Sustmann, Chem.–Eur. J. 1999, 5, 1738.
[36] P. Meineke, U. Rauen, H. de Groot, H.-G. Korth, R. Sustmann, Biol. Chem. 2000, 381, 575.
[37] ꢁLandolt-Bçrnstein, New Series. Radical Reaction Rates in Liquidsꢂ, Ed. H. Fischer, Springer, Ber-
lin, 1983, 1984, Vol. II/13a,b.
[38] O. Augusto, R. M. Gatti, R. Radi, Arch. Biochem. Biophys. 1994, 310, 118.
[39] S. Dikalov, M. Skatchkov, E. Bassenge, Biochem. Biophys. Res. Commun. 1997, 231, 701.
[40] H. Karoui, N. Hogg, C. Frejaville, P. Tordo, B. Kalyanaraman, J. Biol. Chem. 1996, 271, 6000.
[41] C. von Sonntag, H. P. Schuchmann, Angew. Chem., Int. Ed. Engl. 1991, 30, 1229.
[42] K. U. Ingold, J. R. Roberts, J. Am. Chem. Soc. 1973, 95, 3288.
[43] H. A. Gçttinger, V. E. Zubarev, O. Brede, J. Chem. Soc., Perkin Trans. 2 1997, 2167.
[44] J. J. Windle, J. A. Kuhnle, B. H. Beck, J. Chem. Phys. 1969, 50, 2630.
[45] A. Hudson, H. A. Hussain, J. Chem. Soc. B 1968, 251.
[46] E. H. Appelman, D. J. Gosztola, Inorg. Chem. 1995, 34, 787.
[47] M. Lehnig, M. Kirsch, H.-G. Korth, Inorg. Chem. 2003, 42, 4275.
[48] M. Lehnig, M. Kirsch, Org. Biomol. Chem. 2006, 4, 721.
[49] M. Lehnig, K. Jakobi, J. Chem. Soc., Perkin Trans. 2 2000, 2016.
[50] P. Mulder, H. G. Korth, K. U. Ingold, Helv. Chim. Acta 2005, 88, 370.
[51] O. Brede, D. Beckert, C. Windolph, H. A. Gçttinger, J. Phys. Chem. A 1998, 102, 1457.
[52] H.-J. Kim, J. C. Fishbein, Chem. Res. Toxicol. 2003, 16, 715.
[53] A. Treinin, E. Hayon, J. Am. Chem. Soc. 1970, 92, 5821.
[54] D. L. H. Williams, NITRIC OXIDE: Biol. Chem. 1997, 1, 522.
[55] M. Kirsch, H.-G. Korth, R. Sustmann, H. de Groot, Biol. Chem. 2002, 383, 389.
[56] B. C. Challis, S. A. Kyrtopoulos, J. Chem. Soc., Perkin Trans. 1 1978, 299.

2424
Helvetica Chimica Acta – Vol. 89 (2006)
[57] B. C. Challis, S. A. Kyrtopoulos, J. Chem. Soc., Perkin Trans. 1 1978, 1296.
[58] W. D. Emmons, A. F. Ferris, J. Am. Chem. Soc. 1953, 75, 4623.
[59] D. M. Stanbury, Adv. Inorg. Chem. 1989, 33, 69.
[60] H. B. Hetzer, R. G. Bates, R. A. Robinson, J. Phys. Chem. 1966, 70, 2869.
[61] R. W. Fessenden, P. Neta, J. Phys. Chem. 1972, 76, 2857.
[62] Y. L. Chow, W. C. Dauen, S. F. Nelsen, D. H. Rosenblatt, Chem. Rev. 1978, 78, 243.
[63] E. R. Blout, S. G. Cohen, M. Green, U.S. Pat. 2843481, 1958, p. 1–4.
[64] G. A. Evtyugin, V. U. Golꢂfrid, D. A. Semanov, V. Z. Latypova, Y. M. Kargin, L. N. Punegova, G. A.
Marchenko, J. Gen. Chem. USSR (Engl. Transl.) 1989, 59, 2497.
[65] B. C. Challis, T. I. Yousaf, J. Chem. Soc., Chem. Commun. 1990, 1598.
[66] Y.-D. Park, H.-K. Kim, J.-J. Kim, S.-D. Cho, S.-K. Kim, M. Shiro, Y.-J. Yoon,J. Org. Chem. 2003, 68,
9113.
[67] M. Witanowski, L. Stefaniak, G. A. Webb, Annu. Rep. NMR Spectrosc. 1993, 25, 468.
[68] D. R. Duling, ꢁWinSimꢂ, Laboratory of Molecular Biophysics, NIEHS, NIH, Research Triangle
Park, North Carolina, 1994.
[69] D. R. Duling, J. Magn. Reson. B 1994, 104, 105.
[70] J. Heinzer, ꢁEPREXNꢂ, ETH Zürich, 1971.
[71] J. Heinzer, Mol. Phys. 1971, 22, 167.
[72] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Mont-
gomery Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B.
Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene,
X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A.
Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain,
O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A.
G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi,
R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P.
M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, J. A. Pople, ꢁGaussian 03, Revision
C.02ꢂ, Gaussian Inc., Pittsburgh PA, 2003.
[73] D. D. M. Wayner, K. B. Clark, A. Rauk, D. Yu, D. A. Armstrong,J. Am. Chem Soc. 1997, 119, 8925.
Received April 28, 2006