Reactions of simple and peptidic alpha-carboxylate radical anions with dioxygen in the gas phase

Article abstract of DOI:10.1039/c1cp20784a

α-Carboxylate radical anions are potential reactive intermediates in the free radical oxidation of biological molecules (e.g., fatty acids, peptides and proteins). We have synthesised well-defined α-carboxylate radical anions in the gas phase by UV laser photolysis of halogenated precursors in an ion-trap mass spectrometer. Reactions of isolated acetate (CH₂CO ₂^-) and 1-carboxylatobutyl (CH₃CH ₂CH₂CHCO₂^-) radical anions with dioxygen yield carbonate (CO₃^-) radical anions and this chemistry is shown to be a hallmark of oxidation in simple and alkyl-substituted cross-conjugated species. Previous solution phase studies have shown that C_α-radicals in peptides, formed from free radical damage, combine with dioxygen to form peroxyl radicals that subsequently decompose into imine and keto acid products. Here, we demonstrate that a novel alternative pathway exists for two α-carboxylate C_α-radical anions: the acetylglycinate radical anion (CH₃C(O)NHCHCO₂ ^-) and the model peptide radical anion, YGGFG^-. Reaction of these radical anions with dioxygen results in concerted loss of carbon dioxide and hydroxyl radical. The reaction of the acetylglycinate radical anion with dioxygen reveals a two-stage process involving a slow, followed by a fast kinetic regime. Computational modelling suggests the reversible formation of the C_α peroxyl radical facilitates proton transfer from the amide to the carboxylate group, a process reminiscent of, but distinctive from, classical proton-transfer catalysis. Interestingly, inclusion of this isomerization step in the RRKM/ME modelling of a G3SX level potential energy surface enables recapitulation of the experimentally observed two-stage kinetics.

Full text of DOI:10.1039/c1cp20784a

Physical Chemistry Chemical Physics
www.rsc.org/pccp
Volume 13 | Number 36 | 28 September 2011 | Pages 16109–16456
ISSN 1463-9076
COVER ARTICLE
HOT ARTICLE
Auinger, Mayrhofer et al.
Near-surface ion distribution and
da Silva, Blanksby et al.
Reactions of simple and peptidic
alpha-carboxylate radical anions with buffer effects during electrochemical
dioxygen in the gas phase reactions
1463-9076(2011)13:36;1-R

Dynamic Article Links
PCCP
Cite this: Phys. Chem. Chem. Phys., 2011, 13, 16314–16323
www.rsc.org/pccp
PAPER
Reactions of simple and peptidic alpha-carboxylate radical anions
with dioxygen in the gas phasew
a
a
a
a
Tony Ly,z Benjamin B. Kirk, Pramesh I. Hettiarachchi, Berwyck L. J. Poad,y
a
b
ac
Adam J. Trevitt, Gabriel da Silva* and Stephen J. Blanksby*
Received 16th March 2011, Accepted 22nd June 2011
DOI: 10.1039/c1cp20784a
a-Carboxylate radical anions are potential reactive intermediates in the free radical oxidation of
biological molecules (e.g., fatty acids, peptides and proteins). We have synthesised well-defined
a-carboxylate radical anions in the gas phase by UV laser photolysis of halogenated precursors in
ꢀ
ꢁ
an ion-trap mass spectrometer. Reactions of isolated acetate ( CH CO ) and 1-carboxylatobutyl
2
2
ꢀ
2ꢁ
ꢀ
ꢁ
(
CH
and this chemistry is shown to be a hallmark of oxidation in simple and alkyl-substituted
cross-conjugated species. Previous solution phase studies have shown that C -radicals in peptides,
3
CH
2
CH
2
CHCO ) radical anions with dioxygen yield carbonate (CO
3
) radical anions
a
formed from free radical damage, combine with dioxygen to form peroxyl radicals that
subsequently decompose into imine and keto acid products. Here, we demonstrate that a novel
alternative pathway exists for two a-carboxylate C -radical anions: the acetylglycinate radical
a
ꢀ
ꢁ
ꢀ
ꢁ
anion (CH C(O)NH CHCO ) and the model peptide radical anion, YGGFG . Reaction of
3
2
these radical anions with dioxygen results in concerted loss of carbon dioxide and hydroxyl
radical. The reaction of the acetylglycinate radical anion with dioxygen reveals a two-stage
process involving a slow, followed by a fast kinetic regime. Computational modelling suggests
a
the reversible formation of the C peroxyl radical facilitates proton transfer from the amide to the
carboxylate group, a process reminiscent of, but distinctive from, classical proton-transfer
catalysis. Interestingly, inclusion of this isomerization step in the RRKM/ME modelling of a
G3SX level potential energy surface enables recapitulation of the experimentally observed
two-stage kinetics.
1
tropospheric processing of volatile organics, combustion of
Introduction
2
hydrocarbon fuels, and free radical damage in vivo. These
3
Carbon-centred radicals are key intermediates in numerous
important chemical systems, such as those formed during
radicals react rapidly with ambient molecular oxygen to
4
produce peroxyl radicals. Mass spectrometry has been widely
implemented to study peroxyl radical chemistry, although in
general it has been applied to the analysis of closed-shell
reaction products. In some instances however, mass spectro-
metry has been used to synthesize charge-tagged radicals that
are analogous to elusive neutral radical intermediates. These
analogues, which have isolated charged and radical sites,
a
School of Chemistry, University of Wollongong, NSW 2522,
Australia. E-mail: blanksby@uow.edu.au
Department of Chemical and Biomolecular Engineering,
The University of Melbourne, Victoria 3010, Australia.
E-mail: gdasilva@unimelb.edu.au
ARC Centre of Excellence for Free Radical Chemistry and
b
c
5
Biotechnology, University of Wollongong, NSW 2522, Australia
w Electronic supplementary information (ESI) available: Experimental
mass spectra showing the CID of bromoacetate, the PD of iodo-
2
pentanoate, the reaction of deuterium-exchanged 6 + O , and the
synthesis of YGGFG and its reaction with O
are termed ‘distonic’ radical ions. The distonic radical ion
approach, coupled with mass spectrometry, allows for the
isolation and interrogation of these reactive intermediates in
6
ꢀ
the gas phase.
2
, are provided in the
supporting information. Supporting theoretical evidence provided
includes the full potential energy diagram calculated for acetate radical
anion, concentrations of 6, 7, and 8 derived from the kinetic model,
and coordinates, vibrational frequencies, and moments of inertia for
active stationary points in the RRKM/ME calculations. See DOI:
Previous work has shown that the reaction of distonic
carboxylatocyclohexyl radical anions (1 and 2 in Scheme 1)
and dioxygen results in the formation of a chemically activated
peroxyl adduct that can be collisionally deactivated or can
ꢀ
ꢀ
10.1039/c1cp20784a
decompose through loss of HO
and HO , in good agreement
2
z Current Address: College of Life Sciences, University of Dundee,
7,8
with observations from studies of neutral cyclohexyl radical.
In contrast with the distonic carboxylatocyclohexyl radical
anions 1 and 2, the isomeric a-carboxylate radical ion
Dow St, Dundee DD1 5EH, UK.
y Current Address: Department of Chemistry and Biochemistry,
University of California, San Diego, USA.
1
6314 Phys. Chem. Chem. Phys., 2011, 13, 16314–16323
This journal is c the Owner Societies 2011

Laboratories (Andover, MA). Bromoacetic acid, chloramine
T trihydrate, sodium iodide (anhydrous), leucine enkephalin
peptide (YGGFL), 2-bromopentanoic acid, and sodium meta-
bisulfite were purchased from Sigma-Aldrich (St. Louis, MO).
L-Tyrosine was purchased from BDH Lab Supplies (now part
of VWR International, West Chester, PA). All commercial
compounds were used without further purification. 2-Iodo-
pentanoic acid was prepared from 2-bromopentanoic acid by a
1
2,13
standard procedure.
Water was purified to 18.2 MO
resistivity using a Millipore (Billerica, MA) Synergy UV water
Scheme 1
purification system.
(
structure 3 in Scheme 1) differs significantly in chemistry,
ꢁ
Mass spectrometry
ꢀ
reacting with oxygen to give carbonate radical anion (CO
3
)
Methanolic solutions of analytes were prepared at 10–50 mM
concentrations and infused at a rate of 3–5 mL/min into
as the exclusive product ion, presumably via decomposition of
a fleeting peroxyl radical intermediate through a lower energy
pathway than those available to 1 and 2.
the electrospray ionisation source of
a ThermoFisher
(
San Jose, CA) LTQ linear quadrupole ion-trap mass spectro-
Reaction of 3 with nitric oxide yields displacement of CO
2
meter. The LTQ was modified to perform both photodissocia-
tion and ion-molecule reactions in the ion trap (see below for
details). Aqueous ammonia was added to raise the pH when
required to facilitate the dianion formation. Typical source
parameters were spray voltage ꢁ2.5 and +3.5 kV (for negative
and positive ion mode, respectively), capillary temperature
by NO, which is a hallmark of cross-conjugated radical anions
9
where the radical and the charge interact directly. Given the
unique chemistry of 3 towards oxygen, we were interested in
using a combined experimental and theoretical approach to
further study the chemistry of a-carboxylate radicals and to
explore whether model biomolecular systems, based on this
framework, also react with dioxygen in a similar fashion. In
this study, the archetypal a-carboxylate radical anion along
with model systems for a-carboxylate radicals derived from
fatty acids and peptides (structures 4, 5 and 6 respectively,
Scheme 1) were synthesized by UV photodissociation of
halogenated precursors in a linear ion-trap mass spectro-
2
00–250 1C, sheath gas flow between 10 and 30 (arbitrary
units), and sweep and auxiliary gas flow set at between 0 and
0 (arbitrary units). For collision experiments, ions were mass-
1
selected with a window of 1–4 m/z and fragmented using
default instrument parameters, i.e. 0.250 activation q-parameter,
30 ms activation time (unless otherwise noted) and normalized
1
0
collision energies of 20–40 (arbitrary units).
meter. Once synthesized, the radical anions were isolated
in the presence of dioxygen allowing direct observation of the
1
1
Photodissociation
ionic reaction products and measurement of reaction rates.
We found that carbonate radical anion formation is a general
feature of reactions between simple and alkyl-substituted
a-carboxylate radicals (4, 5) and dioxygen. Furthermore this
was shown to be consistent with statistical reaction rate theory
calculations based on potential energy surface derived from
ab initio calculations. In the peptidic systems however, the
amide functionality modifies the ion-molecule chemistry
The linear ion-trap mass spectrometer is modified in a similar
1
0,14
fashion to previously described.
A 2.75 in. quartz viewport
(MDC Vacuum Products, Hayward, CA) is affixed to the
posterior plate of the spectrometer vacuum housing with a
CF flange to allow transmission of 266 nm laser pulses from
a flashlamp-pumped Nd : YAG laser (Continuum, Santa
Clara, CA). The laser beam is directed into the trap via two
right-angle prisms, which are adjusted to optimize overlap
with the ion cloud. The laser energy is ca. 2 mJ/pulse,
unfocussed, contained within a 3 mm spot diameter and 5 ns
pulse duration. A single laser pulse irradiates trapped ions at a
time synchronized to the beginning of the activation step of a
significantly. The reaction of dioxygen and the simple acetyl-
ꢀ
2
glycinate radical (6) produces concerted loss of CO and HO
from the peroxyl adduct. Computational modelling and
kinetic analysis were performed on the smaller acetylglycinate
radical anion, allowing us to propose a mechanism consistent
with all experimental observations. Both theory and experi-
ment indicate that reaction occurs in two stages, namely via
n
typical MS experiment by feeding a differential signal from
the instrument to the laser flashlamp trigger via a digital delay
generator (Berkeley Nucleonics, San Rafael, CA). The MS
stage that is synchronized with the pulsing of the laser can be
chosen using the Diagnostics menu of the default instrument
oxygen-assisted isomerization followed by further oxidation to
ꢀ
lose CO
2
and the highly reactive HO radical, ultimately
yielding a smaller diamide. Interestingly, the same dissociation
chemistry is also observed for a-carboxylate radicals derived
tuning software. The photodissociation efficiencies achieved
10,15
ꢀ
ꢁ
from a larger peptide, i.e., YGGFG
.
are comparable to previously reported,
e.g., the yield of
iodine atom loss is B20% for iodoYGGFL (see Supporting
Information Fig. S1w).
Experimental methods
Materials
Ion-molecule reactions
Methanol (HPLC grade), acetic anhydride, and ammonia
solution (28%) were obtained from Ajax (Sydney, Australia).
Modifications to the linear ion-trap mass spectrometer to
allow the introduction of neutral gases into the ion trap
11
1
8
O2 gas (95%) was obtained from Cambridge Isotope
region of the instrument have been described previously.
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 16314–16323 16315

ꢁ
1
Briefly, gas reagents in a syringe are introduced into a flow of
Ultra High Purity (UHP) helium (3–5 psi) via a heated septum
inlet (25–250 1C). The gas mixture is introduced into the ion
trap via a variable leak valve, which is adjusted until the ion
with a constant DEdown value of 600 cm . A grain size of
1
10 cm is used for the energy-grained component of the
ꢁ
ꢁ
1
hybrid master equation, for energies up to 30 000 cm . The
continuum component of the master equation is then solved
ꢁ
5
ꢁ1
gauge reads B0.9 ꢂ 10 Torr, which indicates an ion trap
pressure of B2.5 mTorr. The temperature of the vacuum
manifold surrounding the ion trap was measured at 307 ꢃ 1 K,
which is taken as being the effective temperature for ion-
up to 200 000 cm across a further 6000 grains. Barrierless
reactions are described using the hindered Gorin transition
2
3
state model, with the high-pressure limit rate constant set to
that calculated by collision theory. Reported RRKM/ME
simulations represent ten million trials of 200 collisions each.
Calculations were performed at 307 K and 2.5 mTorr of He.
Coordinates for optimized structures, vibrational frequencies,
and moments of inertia for all active species and transition
states involved in the RRKM/ME calculations are provided as
Supporting Information.
1
6
molecule reactions observed therein. Reaction times of
0–10 000 ms were set using the activation time parameter
3
within the control software using no collision energy applied.
All spectra presented are an average of at least 50 scans. The
concentration of dioxygen in the ion trap was measured using
a calibrant reaction where the true second order rate constant
is known (carboxylatoadamantyl radical anion + O ): thus,
2
the [O ] is simply kpseudo first order/ksecond order. Rate coefficients
2
Results and discussion
were determined as previously described with an estimated
1
1
measurement accuracy of ꢃ25%. Collision rates for the
The acetate radical anion (4) is an archetypal cross-conjugated
9
radical anion, which is conveniently synthesized in vacuo
ion-molecule reactions were estimated by the parameterized
1
7
trajectory collision theory of Su and Chesnavich
and branching ratios were calculated using the method of
by UV photolysis of the C–Br bond in bromoacetate
ꢁ
(BrCH
2
CO
2
). Electrospray ionisation (ESI) of a methanolic
1
8
Grabowski and Zhang.
solution of bromoacetic acid yields the two naturally occurring
isotopologues of the bromoacetate anion at m/z 137 and 139.
Isolation of these ions in the ion trap and photodissociation
Theoretical methods
(
PD) via irradiation with a single pulse of 266 nm photons
from an Nd : YAG laser yields the target radical anion 4 at m/z
8 along with bromide anions at m/z 79 and 81 (Fig. 1a). In
Quantum chemical calculations
5
Electronic structure theory calculations are performed in
contrast, conventional collision induced dissociation (CID)
of bromoacetate yields bromide as the exclusive product ion
1
9
Gaussian 09. The B3LYP/6-31G(2df,p) model chemistry
is used to optimize structures and calculate vibrational
frequencies. The high-level G3SX composite theoretical
(
see Supporting Information Fig. S2). As observed previously
for the a-carboxylate cyclohexyl radical (3), reaction of 4 and
dioxygen yields carbonate radical anion at m/z 60, which
2
0
method is used for molecular energies. The G3SX method
uses B3LYP/6-31G(2df,p) optimized geometries and zero
point energies along with higher-level wavefunction theory
single point energies from HF through QCISD(T) theory, with
basis sets of incrementally decreasing size. These energies are
combined with empirical scaling corrections to arrive at the
final G3SX energy. The G3SX method is particularly accurate
increases in abundance with increasing reaction time, as shown
18
in Fig. 1b. Reaction of 4 with
2
O produces a 2 Th mass shift
in the product ion, indicating incorporation of one atom from
dioxygen, as shown in Fig. 1c. A plot of the natural log of the
fractional precursor intensity versus time is well-described by a
2
linear fit (R = 0.992), indicating that the reaction conditions
2
1
for barrier heights (relative to computational expense), and
1
are pseudo first-order (see Supporting Information Fig. S3). A
ꢁ11
9
also performs well for a range of thermochemical properties.
The G3SX enthalpies reported in this study are expected to be
3
ꢁ1 ꢁ1
second order rate constant of 7.5 ꢂ 10
is obtained using the pseudo first-order rate constant and the
cm molecule s
2
0,21
accurate to ꢃ1.5 kcal/mol (95% confidence intervals).
9
ꢁ3
11
concentration of O
2
in the trap (7.1 ꢂ 10 molecule cm ),
which is determined using a calibrant ion-molecule reaction.
Reaction rate calculations
A calculated reaction efficiency of B12% is derived from the
measured second order rate constant and the predicted
Reaction yields and apparent rate constants, k(T,P), for
chemically activated reaction of both the acetate (4) and
ꢁ
10
3
cm molecule s .
ꢁ1 ꢁ1
collision frequency, which is 6.5 ꢂ 10
acetylglycinate (6) radical anions with O are obtained from
2
The reaction efficiency is relatively high, and is comparable to
rates determined for the reactions of distonic radicals with
7,11
stochastic solutions of the 1D Master Equation (ME),
2
2
using MultiWell-2010.1. Densities and sums of states are
determined from Stein-Rabinovitch-Beyer-Swinehart counts,
based on B3LYP/6-31G(2df,p) vibrational frequencies and
moments of inertia. Internal degrees of freedom are treated
as harmonic-oscillators, with external degrees of freedom
modeled as an active 1D and an inactive 2D rotor, following
the conventional assumption of a symmetric top. Rice-
Ramsperger-Kassel-Marcus (RRKM) theory is used for
microcanonical k(E), with the inclusion of tunneling corrections
for H-shift reactions via an Eckart barrier. Collisional energy
transfer is described using a single exponential-down model
molecular oxygen.
While PD of 2-bromopentanoate produced a very low
abundance of the desired radical anion, the 2-iodopentanoate
anion (m/z 227) gave rise to a moderate abundance of 5 at m/z
100 (see Supporting Information Fig. S4). By analogy with
bromoacetate (Fig. 1a), PD of the 2-iodopentanoate anion
also yields the halide anion (in this case iodide at m/z 127) as a
major competing pathway. Interestingly, the extension of the
alkyl chain presents two additional competitive channels via
ꢀ
b-scission of the nascent radical yielding m/z 99 (–H ) and
ꢀ
m/z 71 (–CH CH ). These processes presumably occur over a
3
2
1
6316 Phys. Chem. Chem. Phys., 2011, 13, 16314–16323
This journal is c the Owner Societies 2011

Fig. 2 Calculated energy surface for reaction of acetate radical ion
4) with O . Energies represent relative 298 K enthalpies calculated at
the G3SX level of theory, in kcal mol
(
2
ꢁ1
.
Fig. 3 Optimized structure for the transition state corresponding to
2ꢁ
2
CO ) cleaves the
isomerization of the peroxyl radical formed in the acetate radical ion
Fig. 1 (a) Irradiation of bromoacetate (BrCH
carbon-bromine bond to produce acetate radical anion (1) at m/z 58.
b) Reaction of isolated 1 with molecular oxygen for the following
reaction times (in ms), bottom to top: 30, 200, 400, 600, 800, 1000,
(
2
4) + O reaction (B3LYP/6-31G(2df,p) level). The transition state
structure illustrates concerted C–C bond scission and C–O bond
formation.
(
1
6
200, 1500, 1700, and 2000 yields the carbonate radical anion at m/z
18
0. (c) The corresponding reaction with
2
O yields a carbonate radical
entrance channel. The observed carbonate radical is consistent
with the mechanism presented in Fig. 2, which also accurately
predicts that one of the oxygen atoms in carbonate radical
originates from molecular oxygen. Interestingly, hydrogen
carbonate and formyl radical are predicted to be the most
thermodynamically favorable products on the C H O energy
anion that is 2 mass-units heavier, indicating incorporation of one
oxygen atom from molecular oxygen. (d) Isolation of the 1-carboxyl-
atobutyl radical anion (5) at m/z 100 in the presence of dioxygen and
allowed to react for 4000 ms. The ion at m/z 99 forms from m/z 100
2
during isolation of the latter and is unreactive toward O (see text).
2
2
4
ꢁ
1
low barrier and as a consequence it was found that the target
radical anion 5 at m/z 100 could not be isolated in the absence
of m/z 99 (Fig. 1d). Nevertheless the ion abundance of the
closed-shell species m/z 99 was found to be unaffected by
increasing the reaction time in the presence of dioxygen and as
such can be considered a spectator to its radical companion at
m/z 100. The radical anion 5 at m/z 100 was found to yield the
carbonate radical anion at m/z 60 as the exclusive ionic
product of the reaction with dioxygen (Fig. 1d) consistent
with the archetypal a-carbonate radical 4 (Fig. 1b) and the
surface (i.e., 89.6 kcal mol below 4 + O at the G3SX level;
2
see Supporting Information Figure S5). It may be possible for
these products to form via a roaming radical mechanism in the
formaldehyde–carbonate radical anion complex before disso-
ciation to the observed products, although the experiments
reported here provide no evidence for this taking place.
2
The 4 + O reaction has been simulated with RRKM/ME
theory, according to the mechanism depicted in Fig. 2.
Because the barrier for b-scission of the peroxide intermediate
is predicted to fall below the radical energy, this step is
neglected in these calculations, and we assume instead that
the ipso rearrangement transition state leads directly to
formaldehyde + carbonate radical. The RRKM/ME simula-
tions (see Methods: Reaction Rate Calculations) predict that
the formaldehyde + carbonate radical products account for
8.6% of the reaction yield, with the remainder corresponding
7
previously investigated 1-carboxylatocyclohexyl radical 3.
The calculated energy surface for the reaction of 4 and
ꢁ1
dioxygen is shown in Fig. 2. Addition of O
exothermic, which is typical of O association with an alkyl
radical. The 4 + O reaction yields a peroxyl radical that
2
is around 40 kcal mol
2
2
undergoes an ipso rearrangement to form methyl carbono-
peroxoate radical anion, via the transition state structure
depicted in Fig. 3. b-Scission from the a-peroxide radical takes
place with a negligible barrier, and leads to formaldehyde and
carbonate radical anion. The barriers to these unimolecular
rearrangements are modest, and remain lower than the
2
to reverse reaction back to 4 + O . This compares favorably
to the experimentally determined efficiency of 12%, given the
expected uncertainties in the theoretical calculations, the
collision theory rate constant, and difficulties in quantifying
the accuracy of the experimental measurements.
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 16314–16323 16317

anion at m/z 60 but rather a neutral loss of 29 Da (i.e., m/z 86
1
8
in Fig. 4b). The analogous reaction with
2
O shifts the
product mass-to-charge ratio by 2 Th (Fig. 4c) indicating
the transfer of a single oxygen atom. While substitution of
the amide hydrogen with a deuterium does not change the
product mass, suggesting that the amide hydrogen is lost to
form m/z 86 (see Supporting Information Fig. S6). It is
2
proposed that this product ion arises from addition of O to 6,
ꢀ
followed by concerted loss of CO
2
and HO and this
Scheme 2
interpretation is supported by theoretical modelling of the
potential energy surface as discussed below. Unlike the 4 + O
2
We sought to investigate whether similar chemistry is
reaction, a plot of the natural log of the fractional precursor
intensities versus time is notably non-linear, indicating more
complex reaction dynamics, as explained below.
observed for peptidic a-carboxylate radical anions. Potential
peptidic sites for a-carboxylate radicals include the C
aspartate, the C of glutamate, and the C of any non-
amidated C-terminal residue. Radical acetylglycinate (6) was
chosen as a model system for a peptidic C-terminal C -centred
b
of
g
a
The energy surface developed for the 6 + O reaction is
2
illustrated in Fig. 5. Initial association of the radical anion
a
with O forms a peroxyl radical intermediate, although in this
2
radical. As in 4 and 5, the radical in 6 interacts with the charge.
However, the amide group provides an additional lone pair
that can conjugate with the radical and thus modify the
ion-molecule chemistry. Previous work has suggested that
protonated glycine radical can be prepared from UV photo-
case the exothermicity is reduced considerably from that
ꢁ
1
calculated for 4 + O
2
(ꢁ32.9 kcal mol compared with
ꢁ
1
40.1 kcal mol for the latter). This is due to conjugation
ꢁ
between the amide group and the radical site that is lost upon
2
4
O addition. Despite this, we still identify four transition states
2
dissociation of tyrosine via neutral loss of 107 Da.
resulting in transformation of the peroxyl radical that occur
We predicted that similar UV photodissociation chemistry
could be leveraged to synthesize 6, as shown in Scheme 2. The
mass spectrum resulting from irradiation of deprotonated
N-acetyltyrosine at 266 nm (Fig. 4a) shows numerous photo-
dissociation products including neutral loss of 107 Da,
with energy below the entrance channel. The highest-energy
ꢀ
pathway is for HO2 elimination, which forms a hydrogen-
ꢀ
bonded complex involving the imine and the HO radical.
2
ꢁ1
The isomerization occurs at 1.7 kcal mol below the energy
of the reactants and the dissociated products are actually
ꢁ
presumably through cleavage of the C –C bond to form 6.
a
b
1
.8 kcal mol above the reactants. Formation of an imine
3
In contrast to both 4 and 5, reaction of isolated 6 with
and the hydroperoxyl radical is considered the dominant
5
dioxygen does not result in formation of the carbonate radical
2
pathway in the oxidation of peptide radicals, and we do
detect evidence for this reaction as a very minor channel in the
6
2
+ O process (i.e., m/z 114 in Fig. 4b). The pathway with the
next lowest energy is for formation of carbonate radical anion
plus a neutral carbonyl compound (N-formyl-acetamide),
ꢁ
1
which occurs at 6.4 kcal mol below the entrance channel.
This reaction takes place via a transition state of similar
structure to that depicted in Fig. 3 for the analogous process
in acetate radical ion oxidation. The barrier for isomerization
ꢁ1
of the peroxyl radical (26.5 kcal mol with respect to the
peroxyl radical) is also similar to that found for the corres-
ꢁ
1
ponding process in the 4 + O mechanism (24.7 kcal mol ).
2
Despite this pathway requiring a somewhat smaller barrier
ꢀ
than the imine + HO
2
channel, no experimental evidence is
detected for carbonate radical formation (i.e., m/z 60 is not
observed in Fig. 4b).
The 6 + O mechanism reveals a transition state lower yet
2
in energy, leading directly from the peroxyl radical to the
hydroxyl radical, CO , and an anionic carbonyl species
2
(
N-formylacetamide anion, 8). The transition state structure
for this dissociation reaction is depicted in Fig. 6, where we
observe that it involves a concerted intramolecular hydrogen
Fig. 4 (a) PD spectrum of acetyltyrosinate anion. (b) Reaction of the
acetylglycinate radical anion (6) with molecular oxygen for 1 s yields
shift and C–CO
2
bond scission. Intrinsic reaction coordinate
scans demonstrate that initial fragmentation to CO
ꢁ
2
and
ꢀ
the product ion at m/z 86, which is due to addition of molecular
ꢀ
ꢀ
CH C(O)N C (H)OOH is followed by b-scission to HO
3
oxygen, followed by concerted loss of CO
2
and HO . A very minor
and 8, without any further barrier. This transition state lies
ꢁ
ꢁ
1 Da product observed at m/z 114, which is not present during
ꢀ
1
15.1 kcal mol below the reactant energies, with barrier of
isolation, is likely due to +O
2
, followed by –HO . (c) Reaction with
2
18
ꢁ1
O
2
produces a 2 Th shift to form m/z 88 with the low abundance
only 17.8 kcal mol relative to the peroxyl radical. Loss of
ꢀ
16
m/z 86 arising from residual
O
2
.
HO and CO from the anion 6 upon O addition via this
2
2
1
6318 Phys. Chem. Chem. Phys., 2011, 13, 16314–16323
This journal is c the Owner Societies 2011

2
Fig. 5 Calculated energy surface for reaction of acetylglycinate radical ion (6) with O . Energies represent relative 298 K enthalpies calculated at
ꢁ1
the G3SX level of theory, in kcal mol
.
Fig. 6 Optimized structure for the transition state corresponding to
ꢀ
CO
2
, HO loss from the peroxyl radical formed in the acetylglycinate
2
radical anion (6) + O reaction (B3LYP/6-31G(2df,p) level). This
transition state structure illustrates concerted hydrogen shift and C–C
bond scission.
mechanism is responsible for the dominant product ion signal
observed at m/z 86 (Fig. 4b). The mechanism identified for the
formation of these products is also consistent with the isotope
labeling experiments, which indicated loss of the amide
hydrogen and incorporation of one oxygen atom from O in
2
the product ion. The fact that the carbonate radical is not
detected in acetylglycinate radical anion oxidation, whereas it
is the dominant acetate radical anion oxidation product,
supports our theoretical prediction that the former radical
possesses much lower-energy reaction pathways in its reaction
with molecular oxygen.
The results described above explain the observed reaction
2
products from 6 + O , but do not explain the multi-step
reaction kinetics measured in these experiments. We find that
this behavior can be explained by a novel oxygen-assisted
isomerization. Fig. 5 illustrates that the 6 + O reaction forms
2
a peroxyl radical for which the lowest-energy transition state
corresponds to proton transfer from the amide nitrogen to a
carboxylate oxygen resulting in the slightly more energetic
Fig. 7 (a) Optimized structure for the transition state corresponding
to isomerization of the peroxyl radical formed in the acetylglycinate
ꢁ
1
(
+7.6 kcal mol ) amide-centred anion. The structure of this
radical anion (6) + O
intramolecular proton transfer connecting radical anions 6 and 7
B3LYP/6-31G(2df,p) level). (c) Reaction energetics for direct proton
2
reaction. (b) Optimized structure for direct
transition state is shown in Fig. 7(a) and reveals that the
˚
migrating proton is closer to the carboxylate oxygen (1.18 A)
˚
than the amide nitrogen (1.33 A). For comparison, Fig. 7(b)
(
transfer compared with that facilitated by reversible oxygen addition.
shows the equivalent transition state for direct, unassisted
Energies represent relative 298 K enthalpies calculated at the G3SX
ꢁ1
unimolecular conversion of 6 to 7 and reveals the migrating
level of theory, in kcal mol .
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 16314–16323 16319

Table 1 Rate constants for reaction of the radical ions 4, 6, and 7
with dioxygen
involvement of the oxygen-assisted isomerization pathway
provides a more satisfactory explanation for the measured
reaction kinetics (see below).
3
k (cm molecule
ꢁ1 ꢁ1
s )
Reaction rate simulations have been performed for both
a
Reaction
Experimental
Theoretical
6
+ O and 7 + O , in both cases incorporating all pathways
2 2
ꢀ
ꢁ
ꢀ
ꢁ11
ꢁ11
ꢁ10
ꢁ10
ꢁ10
4
6
6
7
+ O
+ O
+ O
+ O
2
2
2
2
- HCHO + CO
- 7 + O (k
3
(k
1
)
7.51 ꢂ 10
ꢁb
5.54 ꢂ 10
4.13 ꢂ 10
1.76 ꢂ 10
4.25 ꢂ 10
depicted in Fig. 5. Rate constants for the radical ion + O
2
ꢁ10b
ꢁ10
3
cm molecule s
ꢁ1 ꢁ1
2
2
)
1.52 ꢂ 10
steps were determined to be 5.89 ꢂ 10
- 8 + CO
- 8 + CO
2
+ HO (k
3
)
)
from trajectory collision theory, whereas dissociation of the
ꢀ
ꢀ
1.52 ꢂ 10ꢁ
10b
2
+ HO (k
4
imine–HO
2
complex was treated with the reverse reaction
ꢁ11
a
b
From RRKM/ME simulations. Determined from a least-squares
fit assuming a two-step (k , k ) reaction mechanism.
3
cm molecule
ꢁ1 ꢁ1
having a rate constant of 1 ꢂ 10
s ,
2
4
consistent with variational transition state theory calculations
30,31
ꢀ
on dissociation of similar HO
2
radical complexes.
The
˚
proton closer to the carboxylate oxygen (1.22 A) than the
˚
amide nitrogen (1.32 A). This geometry suggests an earlier
6 + O
2
reaction is predicted to yield 70.1% 7 + O and 29.8%
2
ꢀ
8 + CO + HO , with 0.1% 6 + O via the reverse reaction.
2
2
ꢀ
transition state consistent with this slightly exothermic
The carbonate radical anion and HO2 radical pathways were
predicted to take place to only a minor extent (respective yields
ꢁ
1
isomerization (ꢁ4.7 kcal mol ). A comparison of the energetics
of direct proton transfer to that facilitated by formation of
intermediate peroxyl radicals is shown in Fig. 7(c) and illus-
trates the net reduction in the reaction barrier from +3.4 to
of 0.004% and 0.00001%). For the 7 + O2 reaction, the
ꢀ
dominant products are 8 + CO
2
+ HO (72.1%), with a
significant null reaction channel reforming 7 + O
2
(27.9%).
ꢁ
1
ꢁ
25.3 kcal mol afforded by participation of the dioxygen
Minor amounts of 6 + O (0.005%) and carbonate radical +
2
molecule. This dramatic lowering of the barrier is reminiscent
of classical ‘‘proton transfer catalysis’’ where the transition
state energies for intramolecular proton transfer within an ion
are lowered by partial solvation of the proton by a neutral
N-formyl-acetamide (0.001%) are also predicted to form,
ꢀ
whereas the branching ratio to the HO
to calculate.
2
channel is too low
Calculated rate constants for the formation of important
2
6,27
atom or molecule.
There are however three distinctive
products in reactions of 6 and 7 with O are listed in Table 1.
2
features of the effect observed here: (i) the neutral molecule
is covalently attached to the ion in the transition state, i.e., the
dioxygen in this instance is present as a formal peroxyl moiety
in the transition state; (ii) the proton transfer is remote from
These rate constants have been used to model the oxidation
kinetics of acetylglycinate in the ion trap. Consistent with the
experimental settings, we assumed that all concentrations were
homogeneous and that O
total constant [O
2
was in significant excess, with the
9
ꢁ3
˚
the peroxyl moiety by some 3.19 A and; (iii) while there is a net
2
] concentration at 7.1 ꢂ 10 molecule cm
.
lowering in transition state energy due to addition of the
The set of simultaneous ordinary differential equations
oxygen, the barrier for the proton transfer relative to the
ꢁ
(ODEs) describing [6], [7], and [8] were then solved numeri-
32
cally, using previously-developed code. The resultant concen-
1 28
)
peroxyl radical is actually higher (+7.8 kcal mol
than
ꢁ1
that for the unassisted pathway (+3.4 kcal mol ). The
absence of even a partial bonding or solvation interaction
between the neutral ‘‘catalyst’’ and the transferring proton in
the transition state is distinct from classical proton-transfer
catalysis. It appears that the catalytic role of the oxygen in the
trations, normalized to unity initial reactant concentration, are
provided as Supporting Information (Fig. S7), compared to
experimental fractional intensities for m/z = 115 (6 + 7) and
86 (8). The calculated rate constants result in an apparent
reaction rate faster than that observed experimentally,
although they do manage to describe the two-stage kinetics.
Here, the slow initial reaction rate can be attributed to oxygen-
assisted isomerization of 6 to 7, with the apparent rate of
present system is simply to provide the necessary energy
1
(
ꢁ32.9 kcal molꢁ ) to fuel the proton transfer reaction via
reversible formation of the peroxyl radical. While the rever-
sible addition of dioxygen to carbon-centred radicals is
reaction subsequently increasing due to the faster 7 + O
ꢀ
2
2
9
well-known from solution phase studies, the indirect proton
transfer catalysis observed here is likely to be unique to the gas
phase where the excess energy from formation of the new
C–OO bond remains with the system and can be used to power
the isomerization.
reaction to form 8 + CO
2
+ HO .
In order to further investigate the complex kinetics observed
for the 6 + O reaction in the ion trap we attempted to
numerically fit experimental values to k –k . Initial trials
2
2
4
revealed that the results could be adequately described using
Calculations reveal that isomerization of the intermediate
peroxyl radical anion can result in formation of a rearranged
only k and k , implying the mechanism 6 + O - 7 + O -
2
4
2
2
ꢀ
8 + CO + HO . The closed-form solution to the reduced
2
radical anion 7 by loss of O but, in addition, the intermediate
2
system of ODEs describing [6] and [7] (where [8]=1–[6]–[7]) as
ꢀ
can undergo competitive loss of CO
2
+ HO (Fig. 5).
a function of time was obtained using the inverse Laplace
3
3
The latter pathway is actually somewhat lower in energy
transform method (with unity initial reactant concentration).
A least-squares procedure was then used to obtain best fits to
and k , which for both rate constants provided values of
1
(
ꢁ6.7 kcal molꢁ below the entrance channel) but will be less
favored in terms of entropy than the direct dissociation.
Because 6 and 7 are isobaric, the mass spectrometry experi-
ments do not allow for direct differentiation of these products,
nor can they exclude the possibility of direct rearrangement of
k
2
4
ꢁ
10
3
ꢁ1 ꢁ1
1.52 ꢂ 10
cm molecule s . Fig. 8 demonstrates that the
fitted two-step mechanism accurately reproduces the observed
kinetics. Reasonable fits can still be obtained using small
ꢁ
1
.
ꢁ11
3 ꢁ1 ꢁ1
cm molecule s ),
6
to 7 over the relatively modest barrier of +3.4 kcal mol
values of k3 (on the order of 10
Through analysis of the kinetics however, we show that the
although no value for this rate constant improved the agreement
1
6320 Phys. Chem. Chem. Phys., 2011, 13, 16314–16323
This journal is c the Owner Societies 2011

Fig. 8 Kinetics of the acetylglycinate radical ion (6) + O
2
reaction. Dot points represent experimental results (relative ion intensities). Solid lines
ꢀ
represent least-squares fit of the experimental data assuming the mechanism 6 + O
2
- 7 + O
2
- 8 + CO
2
+ HO .
between model and experiment. The experimental rate
constants for 6 + O and 7 + O correspond to a reaction
efficiency of 26%, comparable to the value determined for
oxygen is the production of hydroxyl radicals with concomi-
tant expulsion of carbon dioxide.
2
2
a
Peptidic C radicals are predicted to form as a result of free
4
+ O (12%) and, as noted, comparable to literature values
2
radical damage in vivo. These radicals are highly stabilized by
7
,11
for other radical ion reactions with O2.
The experimental
captodative interactions from the neighbouring amide and
5,36
3
rate constants are in relatively good agreement with the purely
theoretical RRKM/ME values, which are too large by around
carbonyl groups.
However, pulse radiolysis experiments
on model amino acids have shown that these radicals react
quickly with oxygen to produce peroxyl radical intermediates
that undergo backbone scission to form an amide and an a
a factor of two. The reported acetylglycinate radical ion + O
2
experiments, and accompanying theoretical and kinetic
analysis, therefore provide evidence for two novel low-energy
reactions, one that achieves oxygen-assisted isomerization of
the reactant and another that results in formation of the highly
reactive hydroxyl free radical.
2
5
keto acid. Further experiments demonstrated that other
species such as pyruvic acid, acetaldehyde, acetic acid, and
3
7
carbon dioxide are also produced in significant yields. Inter-
molecular recombination of two amino acid peroxyl radicals
was proposed as a possible explanation for these side-
products. In this manuscript, we show that a novel uni-
molecular dissociation pathway is available to a-carboxylate
peroxyl radicals isolated in the gas phase, namely loss of
A peptide was examined next to determine whether the ion-
molecule chemistry observed for 6 would extrapolate to larger
systems. It has been previously shown that radical peptides can
3
4
be synthesized from iodinated peptide precursors. Here, the
tyrosine residue of YGGFL was iodinated using published
procedures and then acetylated to block deprotonation of the
tyrosine sidechain, thus favouring a charged C-terminus. PD
of the peptide derivative cleaves the carbon-iodine bond,
yielding a radical at the tyrosine sidechain (see Supporting
Information Fig. S8a). CID of the peptide radical produces
sidechain losses, predominantly from leucine, and backbone
carbon dioxide and generation of hydroxyl radical. This
ꢀ
8 + CO
2
+ HO pathway is, however, likely to be competitive
with the imine + keto acid pathway (cf. Scheme 4) even in
aqueous solution, due to the low barrier heights relative to the
peroxyl radical (Fig. 5). Indeed, the observation of carbon
dioxide in gamma-irradiated amino acids and peptides is
entirely consistent with the novel pathway described here.
1
5
fragmentation. Previous studies have shown that sidechain
losses and backbone fragmentation are due to radical migra-
tion (via hydrogen atom abstraction), followed by b-scission.
In the case of the 56 Da loss, b-scission occurs from a Leu Cg
Conclusions
We have explored the gas phase ion-molecule reactions of
oxygen and cross-conjugated radical anions, including model
and larger biomolecular systems. The archetypal cross-
conjugated radical anion, the acetate radical anion (4), reacts
with oxygen to produce carbonate radical anion exclusively.
The reactivity observed is completely consistent with the
radical, effectively transforming Leu to a C -radical Gly, i.e.,
a
an alpha carboxylate radical, as shown in Scheme 3. Reaction
ꢀ
ꢁ
of YGGFG with dioxygen yields loss of 29 Da (Scheme 3),
which is the same loss observed for 6 (see Supporting
Information Fig. S8b). The loss of 29 Da is not observed in
ꢀ
ꢁ
7
the CID of YGGFG (see Supporting Information Fig. S8c),
confirming that the product is a result of specific ion-molecule
chemistry and not simply dissociation of a metastable
precursor. These results indicate that a general feature of
reactions between peptidic a-carboxylate radical anions and
1-carboxylatocyclohexyl radical (3) examined previously.
The reactivity of the 1-carboxylatobutyl radical anion (5) is
diminished somewhat, but carbonate radical remains a
significant product. Combined, these results indicate that the
carbonate radical anion formation is a general feature in the
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 16314–16323 16321

Scheme 3
Scheme 4
reactions of simple and alkyl-substituted a-carboxylate radical
Acknowledgements
anions with oxygen. Furthermore, the efficiency and selectivity
of this reaction suggests that this product ion at m/z 60 may be
used as a diagnostic ion for the presence of cross-conjugated
radical anions. Indeed, the clear distinction in reactivity
between the radical anion 3 and its distonic isomers 1 and 2
suggests that the term distonoid radical ions—coined by
S.J.B., B.L.J.P., T.L., G.d.S and A.J.T acknowledge funding
through the Australian Research Council (DP0986738,
DP110103889 and DP1094135) and B.B.K. is supported by
an Australian Postgraduate Award. S.J.B. and A.J.T.
acknowledge the generous provision of computing resources
from the NCI National Facility (Canberra) through the Merit
Allocation Scheme and G.d.S. is thankful to the Victorian
Partnership for Advanced Computing (VPAC) for providing
computational resources. B.B.K. and S.J.B. thank Dr Shuji
Kato (University of Colorado, Boulder) for helpful
discussions and providing a SigmaPlot routine for calculation
of collision rates.
3
8
Eberlin and co-workers —may be usefully applied to the
former system as well as the other a-carboxlyate radical anions
studied here (4, 5 and 6).
We demonstrate that the ion-molecule chemistry changes
significantly when the a-carboxylate radical anion motif is
incorporated into amino acids and peptides. In such instances,
the presence of the neighbouring amide functionality enables
access to new pathways that are significantly lower in energy
than carbonate radical anion formation. In the archetypal
peptidic system, acetylglycinate radical anion (6), the
concerted loss of carbon dioxide and hydroxyl radical is the
dominant pathway observed. Similar ion-molecule chemistry
is observed in a larger peptidic a-carboxylate radical anion.
Interestingly, the kinetics measured for the acetylglycinate +
References
1
P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau,
G. D. Hayman, M. E. Jenkin, G. K. Moortgat and F. Zabel,
Atmos. Environ., 1992, 26A, 1805.
2 J. Zador, C. A. Taatjes and R. X. Fernandes, Prog. Energy
Combust. Sci., 2011, 37, 371, DOI: 10.1016/j.pecs.2010.06.006.
3
B. Halliwell and J. M. C. Gutteridge, Free radicals in biology and
medicine, OUP, Oxford, UK, 3rd edn, 1999.
B. J. Finlayson-Pitts and J. N. Pitts, Science, 1997, 276, 1045.
2
O reaction had significant curvature, consisting of a ‘‘slow’’
followed by a ‘‘fast’’ component. Theoretical calculations
point to a novel oxygen-assisted isomerization whereby
reversible formation of a peroxyl moiety facilitates proton
transfer from the amide to the carboxylate group. Indeed, the
experimental reaction kinetics were fully recapitulated by
including the isomerization step in the RRKM/ME modelling.
The systems examined in this manuscript are direct models
of free radical damaged peptides and fatty acids. Due to the
prevalence of oxygen in the atmosphere and in the bio-
molecular environment, it is conceivable that some of the
mechanisms shown also occur in aqueous solution. Formation
of carbonate radical anion is higher in energy than simple
decomposition of the peroxyl adduct in aqueous solution to
dioxygen and the fatty acid radical. Thus, the formation of
carbonate radical anion is only expected to be favourable in
the gas phase. In contrast, the novel decomposition pathway
observed in the peptidic systems leading to loss of CO2
4
5 B. F. Yates, W. J. Bouma and L. Radom, Tetrahedron, 1986,
2, 6225.
H. Kentta
4
6
¨
maa, Ion-Molecule Reactions of Distonic Radical
Cations, in Encyclopedia of Mass Spectrometry, ed. N. M. M.
Nibbering, Elsevier, Amsterdam, 2005, pp. 160–164.
7 B. B. Kirk, D. G. Harman and S. J. Blanksby, J. Phys. Chem. A,
010, 114, 1446–1456.
2
8
A. M. Knepp, G. Meloni, L. E. Jusinski, C. A. Taatjes,
C. Cavallotti and S. J. Klippenstein, Phys. Chem. Chem. Phys.,
2007, 9, 4315.
9
P. G. Wenthold and R. R. Squires, J. Am. Chem. Soc., 1994, 116,
1
1890–11897.
1
0 T. Ly and R. R. Julian, J. Am. Chem. Soc., 2008, 130, 351–358.
1 D. G. Harman and S. Blanksby, Org. Biomol. Chem., 2007, 5, 3495.
1
12 C. A. Horiuchi and J. Y. Satoh, Chem. Lett., 1984, 1509–1510.
1
1
3 J. L. Belletire and D. F. Fry, J. Org. Chem., 1987, 52, 2549–2555.
4 T. Y. Kim, M. S. Thompson and J. P. Reilly, Rapid Commun. Mass
Spectrom., 2005, 19, 1657–1665.
15 Q. Sun, H. Nelson, T. Ly, B. M. Stoltz and R. R. Julian,
J. Proteome Res., 2009, 8, 958–966.
16 S. Gronert, J. Am. Soc. Mass Spectrom., 1998, 9, 845.
ꢀ
and generation of HO is likely a competitive pathway in
1
1
7 T. Su and W. J. Chesnavich, J. Chem. Phys., 1982, 76, 5183.
8 J. J. Grabowski and L. J. Zhang, J. Am. Chem. Soc., 1989,
111, 1193.
oxygenated aqueous solutions and is consistent with the
products observed from gamma-irradiated peptides.
1
6322 Phys. Chem. Chem. Phys., 2011, 13, 16314–16323
This journal is c the Owner Societies 2011

1
9 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li,
H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin,
V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari,
A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,
N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross,
V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, R. L. Martin, K. Morokuma, V. G.
Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
24 T. Tabarin, R. Antoine, M. Broyer and P. Dugourd, Rapid
Commun. Mass Spectrom., 2005, 19, 2883–2892.
25 W. M. Garrison, Chem. Rev., 1987, 87, 381–398.
26 A. J. Chalk and L. Radom, J. Am. Chem. Soc., 1997, 119,
7573–7578.
27 D. K. Bohme, Int. J. Mass Spectrom. Ion Processes, 1992, 115,
95–110.
28 Interestingly this barrier is comparable to that calculated for the
analogous proton transfer in the even-electron acetylglycinate
ꢁ1
anion itself (B8 kcal mol , data not shown). This further
supports the conclusion that in the oxygen-assisted pathway the
radical does not play a direct role in the observed catalysis.
29 P. J. Barker, A. L. J. Beckwith and Y. Fung, Tetrahedron Lett.,
1983, 24, 97–100.
30 G. da Silva, J. W. Bozzelli, L. Liang and J. T. Farrell, J. Phys.
Chem. A, 2009, 113, 8923.
¨
S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V.
Ortiz, J. Cioslowski and D. J. Fox, GAUSSIAN 09 (Revision A.02),
Gaussian, Inc., Wallingford CT, 2009.
0 L. A. Curtiss, P. C. Redfern, K. Raghavachari and J. A. Pople,
J. Chem. Phys., 2001, 114, 108.
1 J. Zheng, Y. Zhao and D. G. Truhlar, J. Chem. Theory Comput.,
31 G. da Silva, J. Phys. Chem. A, 2011, 115, 291.
32 G. da Silva, B. Z. Dlugogorski and E. M. Kennedy, Int. J. Chem.
Kinet., 2007, 39, 645.
2
2
2
33 C
6
(t) = exp(ꢁt/t
1
); C
= t
7
(t) = K
1
exp(ꢁt/t
1
) + K
[O ]); t
2
exp(ꢁt/t
2
);
]).
K
1
= t /(t ); K
2
1
ꢁ t
2
2
2
/(t ); t
2
ꢁ t
1
1
= 1/(k
3
2
2
= 1/(k [O
5
2
2009, 5, 808.
34 T. Ly and R. R. Julian, Angew. Chem., Int. Ed., 2009, 48,
7130–7137.
2 (a) MultiWell-2010.1 Software, 2010, designed and maintained by
John R. Barker with contributors Nicholas F. Ortiz, Jack M. Preses,
Lawrence L. Lohr, Andrea Maranzana, Philip J. Stimac, T. Lam
Nguyen, and T. J. Dhilip Kumar, University of Michigan, Ann
Arbor, MI; http://aoss.engin.umich.edu/multiwell/; (b) J. R. Barker,
Int. J. Chem. Kinet., 2001, 33, 232; (c) J. R. Barker, Int. J. Chem.
Kinet., 2009, 41, 748; (d) T. L. Nguyen and J. R. Barker, J. Phys.
Chem. A, 2010, 114, 3718.
35 H. G. Viehe, Z. Janousek, R. Merenyi and L. Stella, Acc. Chem.
Res., 1985, 18, 148–154.
36 G. P. F. Wood, D. Moran, R. Jacob and L. Radom, J. Phys.
Chem. A, 2005, 109, 6318–6325.
37 W. M. Garrison, M. Kland-English, H. A. Sokol and M. E. Jayko,
J. Phys. Chem., 1970, 74, 4506.
38 D. M. Tomazela, A. A. Sabino, R. Sparrapan, F. C. Gozzo and
M. N. Eberlin, J. Am. Soc. Mass Spectrom., 2006, 17, 1014–1022.
2
3 G. P. Smith and D. M. Golden, Int. J. Chem. Kinet., 1978, 10, 489–501.
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 16314–16323 16323