Absolute rate constants for the formation of nitrogen-centred radicals from chloramines/amides and their reactions with antioxidants

Article abstract of DOI:10.1039/b202526d

Pulse radiolysis techniques have been employed to investigate the one-electron reduction of a variety of chloramines and chloramides. These include models for the side-chain of Lys (6-aminohexanoic acid chloramine and α-N-acetyl-Lys chloramine), Gly chloramine, β-alanine chloramine and two models of protein backbone amides, the chloramides of cyclo-(Gly)₂ and cyclo-(Ala)₂. The molar absorption coefficients and stabilities of these chloramines/amides have been determined. The one-electron reduction of these chloramine/amide species by hydrated electrons occurs with second-order rate constants of the order of 10⁹-10¹⁰ M^-1 s^-1, and results in cleavage of the N-Cl bonds to yield nitrogen-centred radicals and chloride ions (as measured by high performance ion chromatography). The reactivities of the nitrogen-centred radicals have been investigated with the readily oxidisable quenchers, hydroquinone and Trolox. These quenchers were used as models of the in vivo antioxidants, ubiquinol-10 and α-tocopherol, and react with second-order rate constants between 2 × 10⁷ and 1 × 10⁸ M^-1 s^-1. No evidence was obtained in these pulse radiolysis experiments for a rapid rearrangement of the oxidising nitrogen-centred radicals to reducing carbon-centred radicals, though such reactions have been indicated in previous EPR studies.

Full text of DOI:10.1039/b202526d

View Article Online / Journal Homepage / Table of Contents for this issue
Absolute rate constants for the formation of nitrogen-centred
radicals from chloramines/amides and their reactions with
antioxidants
2
David I. Pattison,a Michael J. Davies*a and Klaus-Dieter Asmusb
a The Heart Research Institute, 145 Missenden Rd, Camperdown, NSW 2050, Australia
b Faculty of Chemistry, Adam-Mickiewicz-University, Grunwaldska 6, 60-780 Poznan, Poland
Received (in Cambridge, UK) 12th March 2002, Accepted 29th May 2002
First published as an Advance Article on the web 20th June 2002
Pulse radiolysis techniques have been employed to investigate the one-electron reduction of a variety of chloramines
and chloramides. These include models for the side-chain of Lys (6-aminohexanoic acid chloramine and α-N-acetyl-
Lys chloramine), Gly chloramine, β-alanine chloramine and two models of protein backbone amides, the chloramides
of cyclo-(Gly)2 and cyclo-(Ala)2. The molar absorption coefficients and stabilities of these chloramines/amides have
been determined. The one-electron reduction of these chloramine/amide species by hydrated electrons occurs with
second-order rate constants of the order of 109–1010
M
Ϫ1 sϪ1, and results in cleavage of the N–Cl bonds to yield
nitrogen-centred radicals and chloride ions (as measured by high performance ion chromatography). The reactivities
of the nitrogen-centred radicals have been investigated with the readily oxidisable quenchers, hydroquinone and
Trolox. These quenchers were used as models of the in vivo antioxidants, ubiquinol-10 and α-tocopherol, and
react with second-order rate constants between 2 × 107 and 1 × 108 MϪ1 sϪ1. No evidence was obtained in these
pulse radiolysis experiments for a rapid rearrangement of the oxidising nitrogen-centred radicals to reducing
carbon-centred radicals, though such reactions have been indicated in previous EPR studies.
resulted in formation of the aminyl radical of N-phenylglycine,
Introduction
with a second-order rate constant, k = (2.9 0.1) × 109 MϪ1 sϪ1
.
The oxidation of proteins by the powerful oxidant hypo-
chlorous acid (HOCl) has been studied extensively.1–5 This
interest is due to the detection of specific marker compounds
for HOCl-induced protein damage (e.g. 3-chlorotyrosine) in
many disease states, including atherosclerosis, arthritis and
other inflammatory diseases.4–7 HOCl is produced in vivo by
activated phagocytes as part of the body immune defense
system, but misplaced, or excessive production of HOCl
leads to damage to proteins, DNA and other biological
compounds.4,5,8,9
The subsequent reactivity of this radical was not investigated.
One example of a nitrogen-centred radical that has been
studied in aqueous solutions is the succinimidyl radical. The
reduction of N-chlorosuccinimide by eϪaq in H2O occurs rapidly
(k = 1.6 × 1010
M
Ϫ1 sϪ1) to yield chloride ions, and the nitrogen-
centred succinimidyl radical.22 The succinimidyl radical is
capable of oxidising tris(2,2Ј-bipyridyl)-Ru() with a second-
order rate constant, k = 1.3 × 109 MϪ1 sϪ1 22
and abstracts
,
hydrogen atoms from acetonitrile and methanol with second-
order rate constants of k = 1.2 × 105 MϪ1 sϪ1 and k = 1 × 107
Some of the major targets for HOCl-mediated protein oxid-
ation are the free amino groups of the Lys side-chains, and
those of the N-terminus, to yield moderately stable chloramine
species.10–12 Chloramines are capable of oxidising further
substrates (e.g. thiols, thioethers etc.) but are less reactive than
HOCl.13–15 It has been shown by EPR spectroscopic studies that
nitrogen-centred radicals are formed during decomposition
of chloramines via cleavage of the N–Cl bond.1,2,12 The rates of
formation of nitrogen-centred radicals from chloramines
were enhanced by addition of one-electron reductants, such
M
Ϫ1 sϪ1, respectively.23 In addition to intermolecular oxidation
reactions, the succinimidyl radical undergoes a first-order ring-
4
ؒ
opening reaction to yield CH2CH2C(O)NCO (k = 7.8 × 10
s
Ϫ1).22 Similar studies with N-chloroglutarimide have shown
that it reacts rapidly with eϪaq (k = 1.9 × 1010
M
Ϫ1 sϪ1) to form
the glutarimidyl radical, which undergoes further oxidation and
H-abstraction reactions approximately ten times faster than the
succinimidyl radical.23 No evidence for ring-opening reactions
was obtained for the glutarimidyl radical.23
Despite the scarcity of data for the reactivity of aminyl and
amidyl radicals in aqueous solutions, several studies have been
undertaken in organic solvents.24–27 However, it is well estab-
lished that the reactivities of analogous alkoxyl radicals vary
dramatically between organic and aqueous/alcoholic solu-
tions,26,27 and it might be expected that a similar solvent
dependency exists for aminyl and amidyl radicals. In particular,
it has been demonstrated that alkoxyl radicals do not undergo
1,2-shifts in organic solvents, but these reactions are facilitated
in aqueous solutions due to the role of H2O in the reaction.28 It
might be expected that analogous aminyl radicals would behave
in a similar manner.
as reducing metal ions (e.g. Fe2ϩ
,
Cuϩ) or superoxide
radicals.1,2,12,16 Once formed, nitrogen-centred radicals can
undergo a variety of rearrangement reactions to give carbon-
centred species, via intramolecular 1,2 or 1,5 H-shifts, β-scission
or decarboxylation processes or intermolecular
H atom
abstraction reactions.12,17–20
Thus, the identities of radicals that are formed via decom-
position of chloramines are well characterised. However, the
reactivities of nitrogen-centred radicals, and the rearranged
carbon-centred radicals, with potential biological targets have
not been studied in detail, particularly in aqueous solutions.
A recent report21 showed that the pulse radiolysis technique
could be used to investigate the rate of reduction of
N-chloro,N-phenylglycine by hydrated electrons (eϪaq). This
In this paper, the rates of reduction of a series of chloramines
and chloramides (Schemes 1 and 2) by eϪ have been invest-
aq
igated by pulse radiolysis techniques. Among the chloramines
DOI: 10.1039/b202526d
J. Chem. Soc., Perkin Trans. 2, 2002, 1461–1467
This journal is © The Royal Society of Chemistry 2002
1461

View Article Online
Fig. 1 Determination of the molar absorption coefficient for β-AlaCl
at 252 nm. β-AlaCl concentrations were determined by the TNB assay,
and the absorbance at 252 nm was determined relative to a β-Ala
baseline. The molar absorption coefficients are given in Table 1
Scheme 1 Structures of the chloramines/amides used in these studies:
6-aminohexanoic acid chloramine, CANCl; α-N-acetyl-Lys
chloramine, N-Ac-LysCl; Gly chloramine, GlyCl; β-Ala chloramine,
β-AlaCl; cyclo-(Gly)2 chloramide, (Gly)2Cl; cyclo-(Ala)2 chloramide,
(Ala)2Cl.
5-mercapto-2-nitrobenzoic acid (5-thio-2-nitrobenzoic acid)
to the corresponding disulfide; the TNB assay) as shown for
β-AlaCl (Fig. 1). The calculated values are very similar to molar
absorption coefficients that have been determined for other
chloramines (ε = 340–370 MϪ1 cmϪ1) using different experi-
mental methods.30,31 The molar absorption coefficients for the
two cyclic chloramide compounds could not be determined
due to interfering absorbances from the parent compounds
themselves, and other reaction products.
The reactions of HOCl with 6-aminohexanoic acid and
Gly, to yield CANCl and GlyCl respectively, gave almost
100% yields of the chloramines. Stability studies carried out
with the TNB assay showed that neither chloramine decayed by
more than 5% at 22 ЊC over a period of 3 h (the maximum
period for carrying out the pulse radiolysis experiments
described below). Negligible loss was observed at 4 ЊC. The
reactions of HOCl with (Gly)2 and (Ala)2 to form (Gly)2Cl
(66% conversion) and (Ala)2Cl (45% conversion) were less
efficient under the conditions used to prepare the stock
solutions for pulse radiolysis. However, the chloramides formed
were reasonably stable with < 20% decay of either chloramide
at 4 ЊC after 4 h, and ca. 50% loss of both chloramides at 22 ЊC
after 4 h.
Scheme 2 Proposed reaction scheme for the reduction of chloramines/
amides by eϪ and the subsequent radical reactions with oxidisable
(QH2 and Troaloq x) and reducible (MV2ϩ) quenchers.
selected for study were α-N-acetyllysine chloramine (N-Ac-
LysCl) and 6-aminohexanoic acid chloramine (CANCl), which
have been used to model the reactivity of Lys side-chain
chloramines generated during HOCl-mediated protein oxid-
ation. Two other chloramines, namely glycine chloramine
(GlyCl) and β-alanine chloramine (β-AlaCl), have also been
investigated as models for α-amino chloramines, and to deter-
mine the effects of neighbouring groups on the reactivity of the
resulting nitrogen-centred radicals. The cyclic chloramides,
(Gly)2Cl and (Ala)2Cl, are suitable models for protein backbone
chloramides, and are readily generated from reaction of cyclo-
(Gly)2 (glycine anhydride) or cyclo-(Ala)2 (alanine anhydride)
with HOCl.29 The subsequent reactivity of the oxidising
nitrogen-centred radicals (Scheme 2) has been investigated with
hydroquinone (QH2; a simple model for the antioxidant,
ubiquinol-10) and with Trolox (a water-soluble analogue of
Vitamin E). Further rearrangement reactions (Scheme 2) of
the nitrogen-centred radicals to give reducing carbon-centred
radicals have been investigated with methyl viologen (MV2ϩ).
The stability of CANCl has been investigated in the presence
of both QH2 and Trolox to ensure that thermal reactions do not
interfere with the pulse radiolysis quenching experiments over
the timescales employed. With a 2.5-fold Trolox excess over
CANCl (320 µM CANCl, 800 µM Trolox), the concentration of
CANCl (measured by the TNB assay) dropped to 95% of the
initial value over a period of 3 h at 22 ЊC (Fig. 2a). These con-
ditions are comparable to the concentrations and time period
used for the pulse radiolysis experiments described below, and
show that thermal reactions between Trolox and CANCl do
not affect the pulse radiolysis data. A similar approach was
employed to investigate the reaction between CANCl and QH2,
but preliminary experiments using aerated solutions showed
that any QϪ interfered with the TNB assay. Thus, the reaction
ؒ
was monitored directly by UV/vis spectroscopy. Under the
conditions used for the pulse radiolysis experiments (where
solutions are purged with N2) there was less than 10% change in
absorbance of the QH2 peak at 285 nm over 1 h (280 µM
CANCl, 350 µM QH2; Fig. 2b), showing that significant
reaction between the chloramines and QH2 does not occur. In
contrast, in aerated solutions, rapid absorbance changes were
detected (280 µM CANCl, 350 µM QH2), and over 30 min the
QH2 peak at 285 nm was diminished to ca. 50% of the initial
intensity.
Results and discussion
Determination of molar absorption coefficients and chloramine/
amide stability
The chloramine/amide solutions were prepared by adding
a small molar excess (1.1–1.2) of the desired amine or amide
to HOCl to ensure complete consumption of the HOCl
without the formation of dichlorinated products.10 The molar
absorption coefficients of the chloramines were determined
by plotting the direct UV absorbance at 252 nm against the
chloramine concentration (measured by the oxidation of
Quenching of hydrated electrons by chloramines/amides
The reduction of various chloramines/amides (Scheme 1) by
eϪ has been studied in 10% tert-butyl alcohol degassed with
aq
1462
J. Chem. Soc., Perkin Trans. 2, 2002, 1461–1467

View Article Online
which are able to undergo further rearrangement reactions to
12,17
ؒ
yield carbon-centred radicals, >C –NH–,
although the
yields of these rearrangement reactions are not known.2,12,16
The stable products of CANCl (210 µM) reduction by eϪ
aq
(generated by steady-state γ-radiolysis) were investigated by ion
chromatography. The yield of chloride ions released relative to
the concentration of hydrated electrons generated within the
solution was approximately 200% (Fig. 4). This shows that
Fig. 4 Plot showing the increased release of chloride ions from a
solution of CANCl (210 µM in 10% tert-butyl alcohol under N2)
following exposure to increasing concentrations of hydrated electrons
(generated by steady state radiolysis with a 60Co source with a dose rate
of 37.5 Gy minϪ1).
Fig. 2 Investigation of the thermal stability of CANCl in the presence
of radical quenchers: a) CANCl stability in the absence of Trolox (᭹),
and in the presence of 800 µM Trolox (ꢀ) with concentrations
determined by the TNB assay; b) UV/vis spectra measured every 6 min
for the anaerobic (under N2) reaction of 280 µM CANCl with 340 µM
QH2.
electron attachment to the N–Cl bond to give ClϪ and, by infer-
ence, nitrogen-centred radicals is the major process under these
conditions, but there are also secondary reactions leading to
further release of ClϪ. The mechanisms of these secondary
reactions are the subject of further studies. The reactivity of the
nitrogen-centred radicals generated by electron attachment to
the N–Cl bonds was investigated further by pulse radiolysis
experiments.
N2. The rate of decay of eϪ was monitored at 720 nm, and
aq
increased linearly with the concentration of chloramine/amide
present (Fig. 3). The second-order rate constants derived from
Quenching of oxidising radicals
The nitrogen-centred radicals generated by reduction of chlor-
amines are expected to be strongly oxidising, by analogy with
the nitrogen-centred radicals generated from deoxynucleosides
and the succinimidyl and glutarimidyl species.22,32,33 The reac-
tivities of these oxidising radicals with two readily oxidisable
quenchers, hydroquinone (QH2) and Trolox, were investigated
[eqn. (1)]. Pseudo-first order reactions of the nitrogen-centred
(1)
Fig. 3 Plot of the pseudo-first order rate of decay of eϪ (monitored
aq
at 720 nm in 10% tert-butyl alcohol degassed with N2) in the presence of
increasing concentrations of GlyCl.
radicals with the quenchers were achieved by adding increasing
quencher concentrations to chloramine/amide solutions (in
10% tert-butyl alcohol degassed with N2). A typical trace
(Fig. 5a) for quenching with QH2 (monitored at 427 nm)
showed a small instantaneous absorbance due to the absorb-
these experiments are shown in Table 1.† Second-order rate con-
stants for the reactions of eϪaq with N-chloro,N-phenylglycine21
and N-chloroglutarimide23 have been determined previously as
ance of eϪaq. This absorbance decayed as eϪ reacted with the
k = (2.9
0.1) × 109 MϪ1 Ϫ1, and k = 1.9 × 1010 MϪ1 sϪ1
s ,
aq
chloramine/amide to generate nitrogen-centred radicals (inset,
Fig. 5a). Aminyl/amidyl radicals are known to absorb weakly
(typically ε < 1500 MϪ1 cmϪ1) in the visible region of the spec-
trum,21,22 but no absorbances due to nitrogen-centred radicals
were observed in this study, probably due to the relatively poor
signal-to-noise ratio. Immediately following the decay of eϪaq a
slow growth due to oxidation of the quencher was observed
(inset, Fig. 5a). The rate of growth became faster as the added
quencher concentration increased, and the second-order rate
constants were obtained from linear plots of kobs versus the
quencher concentration (Fig. 5b, Table 1). At longer times a
respectively. These values are in close agreement with the rates
determined here (Table 1).
EPR spectroscopy has shown that chloramine reduction by
Ϫ
metal ions (Cuϩ, Fe2ϩ, Ti3ϩ) or superoxide radicals (O2 ) gen-
ؒ
ؒ
erates nitrogen-centred radicals of the general type >CH–N –,
† The errors in the second-order rate constants are formally expressed
as 95% confidence limits. However, due to partial spontaneous decay
of the chloramines/amides these errors only indicate the reproducibility
of the kinetic determinations, and do not include systematic errors
due to loss of chloramines/amides.
J. Chem. Soc., Perkin Trans. 2, 2002, 1461–1467
1463

View Article Online
Table 1 Second-order rate constants determined for the reduction of chloramines/amides by hydrated electrons, and for the reactions of the
resulting oxidising nitrogen-centred radicals with the radical scavengers, QH2 and Troloxa
Substrate
ε/MϪ1 cmϪ1
k2(eϪaq ϩ substrate)/MϪ1 sϪ1
k2(QH2)/MϪ1 sϪ1
k2(Trolox)/MϪ1 sϪ1
CANCl
N-Ac-LysCl
GlyCl
β-AlaCl
(Gly)2Cl
(Ala)2Cl
342 16
323 45
351 33
348 27
Nd
(9.3 0.6) × 109
(6.3 0.7) × 109
(6.5 0.5) × 109
(6.1 0.7) × 109
(1.5 0.1) × 1010
(1.4 0.1) × 1010
(1.1 0.1) × 108
(8.2 0.7) × 107
(7.7 0.7) × 107
(4.8 0.7) × 107
(5.4 0.9) × 107
(2.9 1.2) × 107
(7.0 0.3) × 107
Nd
(4.6 0.4) × 107
(4.5 0.3) × 107
(2.1 0.3) × 107
Nd
Nd
a All errors are formally expressed as 95% confidence limits. However, due to partial spontaneous decay of the chloramines/amides these errors
only indicate the reproducibility of the kinetic determinations, and do not include systematic errors due to loss of chloramines/amides. Nd, not
determined.
protonation state of the semiquinone. Unfortunately, the
resolution achievable under the experimental conditions did
not allow any definite conclusions to be drawn in this direction.
In any case, leaving this mechanistic question open has no
bearing on the interpretation of all other data presented in this
study.
The second-order rate constants determined for the reactions
of nitrogen-centred radicals are similar for both QH2 and
Trolox (Table 1). The nitrogen-centred radicals obtained from
the models of the Lys side-chain chloramine exhibit the fastest
kinetics with both species. The rates for the GlyCl and β-AlaCl
derived nitrogen-centred radicals are slightly slower. This effect
is probably due to the electron delocalising properties of the
nearby carboxylic acid substituent. The rate determined for the
GlyCl-derived nitrogen-centred radical with QH2 is in close
agreement with data reported for the aminyl radical generated
7
Ϫ1
ؒ
by HO reaction with Gly in basic solutions, k = 7.4 × 10 M
sϪ1 18
.
The nitrogen-centred radicals derived from the cyclic
chloramides reacted more slowly, probably due to increased
radical stability induced via electron delocalisation into the
neighbouring amide bond, and possibly steric interactions. It
should be noted that the data obtained for (Gly)2Cl with QH2
deviated from linearity at higher [QH2], thus the given second-
order rate constant is a maximal estimate, determined by
fitting only the linear portion of the curve (Fig. 5c). The reason
for this deviation from linearity is not known and will not be
speculated upon at this point.
The yields‡ of QH2 and Trolox oxidation were also deter-
mined from the kinetic traces. The experimental conditions
were selected to give eϪaq as the major primary reducing species
Fig. 5 Graphs showing the data obtained following pulse radiolysis of
chloramines/amides in the presence of the readily oxidised quencher,
QH2: a) a kinetic trace (λ, 427 nm; [CANCl], 310 µM; [QH2], 450 µM;
ؒ
(G = 2.7), with a small contribution from H (G = 0.6). The
10% tert-butyl alcohol) showing a rapid absorption increase due to eϪ
aq
ؒ
absorption, followed by growth and decay of QϪ . Raw data are shown
ؒ
oxidising contribution of HO was removed by addition of 10%
ؒ
tert-butyl alcohol, which converted HO to the mildly reducing
as open circles, with the calculated fit shown by a solid line. Inset shows
the absorbance changes occurring due to eϪaq decay and QϪ formation
ؒ
ؒ
CH2C(CH3)2OH radicals (G = 2.7). Typically, the maximum
in the first 8 µs following the pulse. Absorbance is given in terms
of G, which represents the approximate micromolar concentration per
10 J absorbed energy. b) Plot showing the pseudo-first order rate of
yields of QH2 oxidation initiated by chloramine reduction were
G = 2.7–3.2 (Fig. 6). These results are consistent with a prac-
tically quantitative conversion of eϪ to QϪ , suggesting that
ؒ
formation of QϪ following pulse radiolysis of 10% tert-butyl alcohol
ؒ
aq
the primary product of chloramine reduction by eϪ is the
aq
solutions (under N2) containing 310 µM CANCl in the presence
oxidising nitrogen-centred radical. The yield of QϪ is slightly
ؒ
of increasing [QH2]. c) Plot showing the pseudo-first order rate of
higher (∆G ∼ 0.5) than can be accounted for by eϪaq alone, and
formation of QϪ following pulse radiolysis of 10% tert-butyl alcohol
ؒ
ؒ
solutions (under N2) containing 300 µM (Gly)2Cl in the presence of
increasing [QH2]. Linear fit provides a maximal estimate of the second-
order rate constant for radical quenching by QH2.
is probably due to H -mediated reduction of the chloramines.
An identical experiment was carried out with an N2O (instead
of N2) degassed solution of CANCl to rapidly quench eϪaq, and
ؒ
leave H as the major reactive species. In this case, the
maximum yield of QH2 oxidation was G ∼ 0.7, consistent with
reduction of CANCl by H to yield species (presumably
nitrogen-centred radicals) that are capable of oxidising QH2.
The maximum QH2 oxidation yields, G = 1.7–2.0, obtained
by reduction of the cyclic chloramides were lower than with
decay in absorbance was observed (Fig. 5a) due to radical–
radical termination reactions between the quencher radicals
and other radicals present, and these decays were allowed for
in the determination of individual kobs values (as described
previously18). The oxidation of QH2 and Trolox by the
nitrogen-centred radicals is assumed to occur via electron trans-
fer and subsequent deprotonation as described previously.34,35
However, hydrogen abstraction is an equally acceptable
mechanistic alternative and cannot be discounted on the basis
of these results. One possibility to shed further light on this
question would be an unambiguous identification of the initial
ؒ
chloramines (Fig. 6), and show that there is incomplete con-
Ϫ
version of eϪ to QϪ . The reaction of e with parent (Gly)2
ؒ
aq
aq
was investigated at 720 nm, and yielded a second-order rate
‡ Yields are quoted as the number of species formed per 100 eV of
absorbed energy. G = 1 corresponds to 0.1036 µM JϪ1 in SI units.
1464
J. Chem. Soc., Perkin Trans. 2, 2002, 1461–1467

View Article Online
the small growth is not due to intermolecular reactions of the
nitrogen-centred radicals to yield reducing radicals. The process
responsible for this small growth could not be determined, but
may be due to one (or more) of the previously characterised
intramolecular reactions that yield carbon-centred radical
species.12
Similar experiments with (Gly)2Cl yielded larger signals and
the observed rate of growth due to reaction of MV2ϩ with
secondary radicals varied linearly with the MV2ϩ concen-
tration, to yield a second-order rate constant, k2 = (8.5 0.9) ×
108 MϪ1 Ϫ1. The maximum yield due to secondary radical
s
reactions was determined as G = 0.45. However, experiments
with parent (Gly)2 yielded similar data, thus it is likely that the
species responsible for MV2ϩ reduction with (Gly)2Cl are
derived from the reactions of eϪ with excess (Gly)2 or the
Fig. 6 Plot showing the yields of QϪ generated by pulse radiolysis of
ؒ
aq
non-chlorinated amide groups of (Gly)2Cl.
310 µM CANCl (᭹) or 300 µM (Gly)2Cl (᭿) in 10% tert-butyl alcohol
solutions (under N2) in the presence of increasing [QH2].
Thus, despite conclusive evidence for the formation of
reducing radicals during chloramine decomposition by EPR
spectroscopy and other techniques,12 it has not been possible to
determine the kinetics of these rearrangement reactions in these
pulse radiolysis studies.
constant, k2 = (1.1 0.1) × 109 MϪ1 sϪ1, in good agreement with
previous studies, where k = 1.7 × 109 MϪ1 sϪ1 at pH 9.2.36 The
addition of QH2 to these solutions did not result in QH2 oxid-
ation, which is consistent with the formation of non-oxidising,
cyclic and ring-opened, carbon-centred radicals (Scheme 3) as
Conclusions
These studies have shown that chloramines/amides are rapidly
reduced by hydrated electrons (k, 6 × 109 – 1.5 × 1010
M
Ϫ1 sϪ1) to
yield oxidising radicals and chloride ions. The oxidising radicals
undergo rapid reactions (k, 2 × 107 – 1 × 108 MϪ1 sϪ1) with QH2
and Trolox, which are models of the common biological anti-
oxidants, ubiquinol-10 and α-tocopherol. These results suggest
that ubiquinol-10 and α-tocopherol may provide protection in
biological systems against damage induced by nitrogen-centred
radicals derived from chloramine/amide degradation.
Scheme
3
Structures of the carbon-centred radicals previously
proposed to be generated by eϪaq reduction of (Gly)2.36,37
reported previously.36,37 Thus, the lower yield of QH2 oxidation
for the cyclic chloramides, compared to the chloramines, can be
explained by reaction of eϪaq with any excess (Gly)2 (or (Ala)2)
remaining in solution, and with the non-chlorinated amide site
within (Gly)2Cl (or (Ala)2Cl).
For example, it has been shown that oxidation of low density
lipoproteins (LDL) by HOCl (300 fold excess) results primarily
in protein oxidation (> 80% of HOCl consumption) to form
predominantly Lys chloramines, with some lipid peroxidation.38
Extensive loss of ubiquinol-10 was also observed.38 The results
presented here suggest that some ubiquinol-10 may be con-
sumed by reaction with nitrogen-centred radicals from Lys
chloramines, as well as by lipid peroxide radicals. Similarly,
the extent of lipid peroxidation induced by HOCl in phos-
phatidylcholine liposomes is attenuated by the presence of
α-tocopherol.39 The authors commented that the peroxidation
is probably mediated by free radical formation, via an unknown
mechanism. One possible explanation for this phenomenon is
that chloramine formation occurs at the choline head group40
and α-tocopherol acts as an antioxidant against the nitrogen-
centred radicals formed upon chloramine degradation.
Furthermore, α-tocopherol and ubiquinol-10 might be
expected to play a role in protection against chloramine-
induced damage in cellular systems, particularly as a role for
nitrogen-centred radicals has recently been demonstrated in
chloramine-mediated lysis of red blood cells.41
The yield data obtained from the Trolox quenching experi-
ments follow a similar pattern to that for QH2, but the max-
imum yield for the chloramines is G = 3.1–3.4, and for (Gly)2Cl
it is slightly lower, with G = 2.9. These yields are slightly higher
than the QH2 data, but can still be attributed to chloramine
reduction primarily by eϪaq, with a small contribution from
ؒ
ؒ
H -mediated reduction and possibly CH2C(CH3)2OH radicals.
Quenching of reducing radicals
It is well established that nitrogen-centred radicals can
undergo further reactions, such as inter- or intra-molecular
hydrogen atom abstraction, β-scission and decarboxylation
processes.12,17,18,20 The radicals produced by these reactions are
typically carbon-centred radicals, and have reducing properties
(Scheme 2).12,17,18,20 Quenching experiments were carried out in
the presence of methyl viologen (MV2ϩ), which reacts rapidly
with strongly reducing radicals [eqn. (2)], to determine whether
Reducing radicals ϩ MV2ϩ
MVϩ ϩ product (2)
ؒ
Finally, evidence for rearrangement reactions of these
nitrogen-centred radicals to form reducing carbon-centred
radicals was not observed in these experiments, despite the
well-documented formation of these species in EPR spin-
trapping studies.12
reducing radicals could be detected following the reduction
of chloramines/amides by eϪaq. The chloramine/amide concen-
tration was increased to 10–20 mM for these experiments, in
order to maximise reduction of the chloramines, and minimise
direct reduction of MV2ϩ by eϪaq. Due to the high chloramine/
amide concentration the solutions were more basic (pH 9–10)
for these studies than for the quenching of oxidising radicals
described above.
Experimental
Materials
Studies with CANCl and GlyCl yielded predominantly direct
All chemicals were of the highest purity available from Aldrich
or Sigma and were used without further purification. Solutions
were prepared with water purified by Serv-A-Pure Co. or
Micropore Milli-Q systems. The pH was adjusted by addition
of appropriate quantities of NaOH or HClO4 for the pulse
radiolysis and ion chromatography experiments, or was con-
MV2ϩ reduction by eϪaq. However, a slow growth due to form-
ation of MVϩ was also observed, with a maximum yield of
ؒ
G = 0.15, but this signal was too small to allow accurate kinetic
data to be determined. Addition of excess 6-aminohexanoic
acid to the solutions did not alter the data. This suggests that
J. Chem. Soc., Perkin Trans. 2, 2002, 1461–1467
1465

View Article Online
trolled with Chelex-treated phosphate buffer (100 mM, pH 7.4)
for all other experiments.
shielded conductivity detector. Standard curves were deter-
mined with ClϪ concentrations from 0–150 µM. Chloramine
decomposition was achieved by steady state radiolysis experi-
ments with a 60Co source (dose rate, 37.5 Gy minϪ1). Samples
were prepared in 10% tert-butyl alcohol solutions and degassed
with N2 immediately before irradiation. The HOCl used
for chloramine formation in these experiments was purified
to reduce the initial ClϪ concentration by the method of
Henderson et al.44
Pulse radiolysis
Pulse radiolysis experiments were carried out at the University
of Notre Dame using an 8 MeV Titan Beta model TBS-8/16-1S
linear accelerator. Pulse lengths were typically 2–10 ns with
absorbed doses of 2–8 Gy (J kgϪ1). Dosimetry was carried out
using N2O-saturated KSCN solutions as described previously.17
A full description of the pulse radiolysis setup and the data
acquisition system has been detailed elsewhere.17–19
Determination of chloramine molar absorption coefficients and
stability studies
All solutions were prepared with 10% tert-butyl alcohol and
thoroughly degassed with N2. Under these conditions the
The UV absorbance of chloramine solutions at 252 nm
was determined on a Perkin Elmer Lambda 40 UV/visible
spectrometer. The chloramine concentrations were determined
by the use of 5-mercapto-2-nitrobenzoic acid (TNB) as
described previously.2,10 Experiments to investigate the stability
of the chloramides showed 65% conversion of HOCl to
chloramide for (Gly)2, or 45% for (Ala)2; these values were used
to calculate the chloramide concentrations in the stock solu-
tions used for pulse radiolysis. The molar absorption coef-
ficients could not be determined accurately due to interfering
absorbances from the amides themselves, and other products.
ؒ
oxidising HO radicals were converted to relatively unreactive
ؒ
CH2C(CH3)2OH radicals (G = 2.7), leaving the strongly
reducing hydrated electrons (eϪaq) as the major primary product
(G = 2.7). In addition to eϪaq, about 20% of the primary species
ؒ
comprised reducing H atoms (G = 0.6). The contributions of
eϪaq and H atoms to the observed kinetics were investigated by
ؒ
degassing with N2O instead of N2, as this converts eϪ to
aq
ؒ
ؒ
the unreactive CH2C(CH3)2OH radicals (via HO formation),
leaving H atoms as the primary reactive species.
ؒ
Stock chloramine solutions were prepared in H2O and the
concentrations were determined by UV/visible spectroscopy
using the molar absorption coefficients given in Table 1.
Typically a small molar excess (1.1–1.2) of the substrate (amine
or amide) was added over the HOCl to ensure complete con-
sumption of the HOCl without the formation of dichlorinated
products (which are known to be formed at higher HOCl :
amine ratios).10 Stock solutions were kept in the dark at 4 ЊC
and aliquots were added via syringe to the bulk 10% tert-butyl
alcohol solution. The quenchers (hydroquinone (QH2) and
methyl viologen (MV2ϩ)) used to study the reactivity of the
secondary radicals were also prepared as concentrated stock
solutions and kept in the dark at 4 ЊC (under N2 or Ar) before
stepwise addition to the bulk sample. When Trolox was used
as the quencher, individual samples (in 10% tert-butyl
alcohol) were prepared for each Trolox concentration, as the
low solubility of Trolox prevented preparation of a con-
centrated stock solution. A fixed concentration of chloramine
was added to each Trolox-containing solution via syringe, and
the samples were purged with N2.
Acknowledgements
This work was supported by the Australian Research Council,
Grants A00001441 and F00001444. David Pattison would like
to thank the Sydney Free Radical Group for providing a travel
grant to undertake this research. He is also grateful to the
Radiation Laboratory of the University of Notre Dame and
the U.S. Department of Energy for the opportunity to use their
pulse radiolysis, and other laboratory, equipment. The authors
would also like to acknowledge Mr Kurt Belting and Dr Igor
ˇ
Stefanic for technical assistance. K.-D. Asmus is currently
on sabbatical leave from the Department of Chemistry and
Biochemistry of the University of Notre Dame.
References
1 C. L. Hawkins and M. J. Davies, Biochem. J., 1998, 332, 617.
2 C. L. Hawkins and M. J. Davies, Biochem. J., 1999, 340, 539.
3 J. Arnhold, S. Hammerschmidt, M. Wagner, S. Mueller, K. Arnold
and E. Grimm, Biomed. Biochim. Acta, 1990, 49, 991.
4 A. J. Kettle, FEBS Lett., 1996, 379, 103.
5 N. M. Domigan, T. S. Charlton, M. W. Duncan, C. C. Winterbourn
and A. J. Kettle, J. Biol. Chem., 1995, 270, 16542.
6 S. Fu, H. Wang, M. J. Davies and R. T. Dean, J. Biol. Chem., 2000,
275, 10851.
7 S. L. Hazen and J. W. Heinecke, J. Clin. Invest., 1997, 99,
2075.
8 C. L. Hawkins and M. J. Davies, Free Radical Biol. Med., 1998, 24,
1396.
9 W. A. Prutz, Arch. Biochem. Biophys., 1996, 332, 110.
10 E. L. Thomas, M. B. Grisham and M. M. Jefferson, Methods
Enzymol., 1986, 132, 569.
11 E. L. Thomas, Infect. Immun., 1979, 23, 522.
12 C. L. Hawkins and M. J. Davies, J. Chem. Soc., Perkin Trans. 2, 1998,
1937.
The reactions were monitored by optical absorbance at the
absorbance maxima of the species of interest: for eϪaq, λ = 720
nm, ε = 19000 MϪ1 cmϪ1; QϪ , λ = 427 nm, ε = 7200 MϪ1 cmϪ1
;
ؒ
Trolox-O , λ = 440 nm, ε = 5400 MϪ1 cmϪ1; MVϪ , λ = 600 nm,
ؒ
ؒ
ε = 12820 MϪ1 cmϪ1 19,42,43
.
Analysis of the kinetic traces was carried out using Origin 4.1
(Microcal). A single exponential fitting procedure was used to
analyse pseudo-first-order eϪ decay curves. However, for
aq
secondary quencher growth curves, a fitting expression that
corrected for long-term decays due to radical–radical reactions
was used. Absolute errors for kinetic radiation chemical
experiments are generally considered to be about 10%. These
errors also apply to the current data unless otherwise stated.
All experiments were carried out at approximately 22 ЊC.
13 A. V. Peskin and C. C. Winterbourn, Free Radical Biol. Med., 2001,
30, 572.
Determination of Cl؊ concentrations
14 W. A. Prutz, Arch. Biochem. Biophys., 1999, 371, 107.
15 W. A. Prutz, R. Kissner and W. H. Koppenol, Arch. Biochem.
Biophys., 2001, 393, 297.
16 C. L. Hawkins, M. D. Rees and M. J. Davies, FEBS Lett., 2002, 510,
41.
17 M. Bonifacic, I. Stefanic, G. L. Hug, D. A. Armstrong
and K.-D. Asmus, J. Am. Chem. Soc., 1998, 120, 9930.
18 M. Bonifacic, D. A. Armstrong, I. Carmichael and K.-D. Asmus,
J. Phys. Chem. B, 2000, 104, 643.
19 I. Stefanic, M. Bonifacic, K.-D. Asmus and D. A. Armstrong,
J. Phys. Chem. A, 2001, 105, 8681.
Chloride ion concentrations were determined by means of
high performance ion chromatography using a Dionex LC20
chromatography enclosure coupled with a Dionex GP40
gradient pump and a Dionex ED40 electrochemical detector.
The ClϪ ions were separated by isocratic elution (retention time,
3.6 min) with a buffer containing 2.7 mM Na2CO3 and 0.3 mM
NaHCO3 (flow rate, 1.5 mL minϪ1) on an IonPac AS12A
(4 mm) analytical column with an IonPac AG12A (4 mm)
guard column. An anion self-regenerating suppressor (ASRS-
ULTRA (4 mm)) was used to minimise the background
conductivity (<15 µS). Ion peaks were detected with a CD25
20 G. L. Hug, M. Bonifacic, K.-D. Asmus and D. A. Armstrong,
J. Phys. Chem. B, 2000, 104, 6674.
1466
J. Chem. Soc., Perkin Trans. 2, 2002, 1461–1467

View Article Online
21 M. L. Canle, J. A. Santaballa and S. Steenken, Chem. Eur. J., 1999, 5,
1192.
22 J. Lind, M. Jonsson, T. E. Eriksen, G. Merenyi and L. Eberson,
J. Phys. Chem., 1993, 97, 1610.
34 G. S. Nahor, P. Neta and Z. B. Alfassi, J. Phys. Chem., 1991, 95,
4419.
35 P. Neta, R. E. Huie, P. Maruthamuthu and S. Steenken, J. Phys.
Chem., 1989, 93, 7654.
23 G. Merenyi, J. Lind and L. Eberson, Acta Chem. Scand., 1998, 52, 62.
24 J. A. Howard, Adv. Free Radical Chem., 1972, 4, 49.
25 R. S. Neale, Synthesis, 1971, 1, 1.
26 B. C. Gilbert, R. G. G. Holmes, H. A. H. Laue and R. O. C.
Norman, J. Chem. Soc., Perkin Trans. 2, 1976, 1047.
27 B. C. Gilbert, P. D. R. Marshall, R. O. C. Norman, N. Pineda and
P. S. Williams, J. Chem. Soc., Perkin Trans. 2, 1981, 1392.
28 P. E. Elford and B. P. Roberts, J. Chem. Soc., Perkin Trans. 2, 1996,
2247.
36 E. Hayon and M. Simic, J. Am. Chem. Soc., 1971, 93, 6781.
37 M. D. Sevilla and R. Failor-Koszykowski, J. Phys. Chem., 1977, 81,
1198.
38 L. J. Hazell and R. Stocker, Biochem. J., 1993, 290, 165.
39 O. M. Panasenko, J. Arnhold, J. Schiller, K. Arnold and
V. I. Sergienko, Biochim. Biophys. Acta, 1994, 1215, 259.
40 D. I. Pattison and M. J. Davies, unpublished work.
41 C. L. Hawkins, B. E. Brown and M. J. Davies, Arch. Biochem.
Biophys., 2001, 395, 137.
29 D. I. Pattison and M. J. Davies, Chem. Res. Toxicol., 2001, 14, 1453.
30 J. M. Antelo, F. Arce and M. Parajo, Int. J. Chem. Kinet., 1995, 27,
637.
31 L. Abia, X. L. Armesto, M. Canle, M. V. Garcia, M. Losada and
J. A. Santaballa, Int. J. Chem. Kinet., 1994, 26, 1041.
32 P. O’Neill and S. E. Davies, Int. J. Radiat. Biol., 1987, 52, 577.
33 A. J. S. C. Vieira and S. Steenken, J. Am. Chem. Soc., 1990, 112, 6986.
42 G. L. Hug, Optical spectra of nonmetallic inorganic transient species
in aqueous solution, National Bureau of Standards (United States),
1981, Vol. 69.
43 R. H. Bisby, S. Ahmed and R. B. Cundall, Biochem. Biophys. Res.
Commun., 1984, 119, 245.
44 J. P. Henderson, J. Byun and J. W. Heinecke, J. Biol. Chem., 1999,
274, 33440.
J. Chem. Soc., Perkin Trans. 2, 2002, 1461–1467
1467