Reactivity and selectivity of charged phenyl radicals toward amino acids in a fourier transform ion cyclotron resonance mass spectrometer

Article abstract of DOI:10.1021/ja111280t

The reactivity of 10 charged phenyl radicals toward several amino acids was examined in the gas phase in a dual-cell Fourier transform ion cyclotron resonance mass spectrometer. All radicals abstract a hydrogen atom from the amino acids, as expected. The most electrophilic radicals (with the greatest calculated vertical electron affinities (EA) at the radical site) also react with these amino acids via NH₂ abstraction (a nonradical nucleophilic addition-elimination reaction). Both the radical (hydrogen atom abstraction) and nonradical (NH₂ abstraction) reaction efficiencies were found to increase with the electrophilicity (EA) of the radical. However, NH₂ abstraction is more strongly influenced by EA. In contrast to an earlier report, the ionization energies of the amino acids do not appear to play a general reactivity-controlling role. Studies using several partially deuterium-labeled amino acids revealed that abstraction of a hydrogen atom from the α-carbon is only preferred for glycine; for the other amino acids, a hydrogen atom is preferentially abstracted from the side chain. The electrophilicity of the radicals does not appear to have a major influence on the site from which the hydrogen atom is abstracted. Hence, the regioselectivity of hydrogen atom abstraction appears to be independent of the structure of the radical but dependent on the structure of the amino acid. Surprisingly, abstraction of two hydrogen atoms was observed for the N-(3-nitro-5-dehydrophenyl)pyridinium radical, indicating that substituents on the radical not only influence the EA of the radical but also can be involved in the reaction. In disagreement with an earlier report, proline was found to display several unprecedented reaction pathways that likely do not proceed via a radical mechanism but rather by a nucleophilic addition-elimination mechanism. Both NH₂ and ¹⁵NH₂ groups were abstracted from lysine labeled with ¹⁵N on the side chain, indicating that NH₂ abstraction occurs both from the amino terminus and from the side chain. Quantum chemical calculations were employed to obtain insights into some of the reaction mechanisms.

Full text of DOI:10.1021/ja111280t

ARTICLE
pubs.acs.org/JACS
Reactivity and Selectivity of Charged Phenyl Radicals toward
Amino Acids in a Fourier Transform Ion Cyclotron Resonance
Mass Spectrometer
George O. Pates, Leonard Guler, John J. Nash, and Hilkka I. Kentt€amaa*
Department of Chemistry, Purdue University, West Lafayette, Indiana 47907, United States
S Supporting Information
b
ABSTRACT: The reactivity of 10 charged phenyl radicals toward several amino
acids was examined in the gas phase in a dual-cell Fourier transform ion cyclotron
resonance mass spectrometer. All radicals abstract a hydrogen atom from the
amino acids, as expected. The most electrophilic radicals (with the greatest
calculated vertical electron affinities (EA) at the radical site) also react with these
amino acids via NH2 abstraction (a nonradical nucleophilic additionꢀelimina-
tion reaction). Both the radical (hydrogen atom abstraction) and nonradical
(NH2 abstraction) reaction efficiencies were found to increase with the electro-
philicity (EA) of the radical. However, NH2 abstraction is more strongly
influenced by EA. In contrast to an earlier report, the ionization energies of the amino acids do not appear to play a general
reactivity-controlling role. Studies using several partially deuterium-labeled amino acids revealed that abstraction of a hydrogen
atom from the R-carbon is only preferred for glycine; for the other amino acids, a hydrogen atom is preferentially abstracted from the
side chain. The electrophilicity of the radicals does not appear to have a major influence on the site from which the hydrogen atom is
abstracted. Hence, the regioselectivity of hydrogen atom abstraction appears to be independent of the structure of the radical but
dependent on the structure of the amino acid. Surprisingly, abstraction of two hydrogen atoms was observed for the N-(3-nitro-5-
dehydrophenyl)pyridinium radical, indicating that substituents on the radical not only influence the EA of the radical but also can be
involved in the reaction. In disagreement with an earlier report, proline was found to display several unprecedented reaction
pathways that likely do not proceed via a radical mechanism but rather by a nucleophilic additionꢀelimination mechanism. Both
NH2 and 15NH2 groups were abstracted from lysine labeled with 15N on the side chain, indicating that NH2 abstraction occurs both
from the amino terminus and from the side chain. Quantum chemical calculations were employed to obtain insights into some of the
reaction mechanisms.
’ INTRODUCTION
While the reactions of the OH radical and other oxygen-
centered radicals in solution have been extensively investigated,3
the reactivity of carbon-centered σ-type organic radicals, such as
the phenyl radical, toward amino acids and proteins is nearly
entirely unexplored.6a Tothe best ofour knowledge, onlyone such
study has appeared in the literature—and in this study, only
glycine was examined.6a This lack of studies is due to the difficulty
in cleanly generating carbon-centered organic radicals in solution.7ꢀ9
However, these reactions are of great interest since organic
radicals and biradicals released by some drugs and antitumor
antiobiotics, such as the enediynes, are known to attack proteins,
in addition to other biopolymers.10
Our laboratory has advanced the “distonic ion approach” for
the investigation of the reactivity of phenyl radicals in the gas
phase,11,12 that is, via the addition of a chemically inert charged
group (to form a distonic radical ion) that allows the manipula-
tion of the radical in mass spectrometers, such as a Fourier trans-
form ion cyclotron resonance (FT-ICR) mass spectrometer.11,12
Attack of radicals on proteins is known to cause proteins to
fragment, oxidize, denature, and lose enzymatic activity.1,2 The size
and complexity of proteins make it difficult to obtain knowledge on
these processes at the molecular level. Hence, many solution
studies aimed at improving the understanding of the reactivity of
radicals toward proteins have been carried out by using small
peptides or individual amino acids as the substrates.1,2
One of the most thoroughly studied radicals is the OH radical.
In aqueous solution, the OH radical reacts with most amino acids
predominantly via hydrogen atom abstraction and with aromatic
amino acids via addition to the aromatic ring.3,4 The hydrogen
atom abstraction occurs from the R-carbon or from the side chain
of the amino acid (abstraction of a hydrogen atom from the side
chain is preferred over abstraction from the R-carbon for amino
acids with side chains), and it corresponds to the major reaction
pathway for amino acids with aliphatic non-sulfur-containing side
chains.5 For the sulfur-containing amino acids methionine and
cysteine, the OH radical attacks the sulfur atom. Abstraction of a
hydrogen atom from the SH group in cysteine by the OH radical
has also been observed.3,4
Received: December 14, 2010
Published: May 25, 2011
r
2011 American Chemical Society
9331
dx.doi.org/10.1021/ja111280t J. Am. Chem. Soc. 2011, 133, 9331–9342
|

Journal of the American Chemical Society
ARTICLE
Chart 1. Structures of the Amino Acids Studied and Their Ionization Energies13
This approach was justified by results demonstrating that gaseous
positively charged phenyl radicals undergo the same reactions
(e.g., hydrogen atom abstraction from thiophenol and addition
to phenol, aniline, and toluene) and show similar reactivity
trends compared to neutral phenyl radicals in solution.11,12
The advantages associated with examining radical reactions by
using the distonic ion approach in a mass spectrometer include
the ability to manipulate the radicals, carry out MSn experiments,
and trap the radicals for variable periods of time to measure their
second-order reaction rate constants.11,12 Furthermore, it is
possible to carry out experiments under clean conditions where
only the desired radicals and reagent molecules are present.
Finally, because of the lack of solvent effects, intrinsic radical
reactivity can be explored.
Using the above approach, reactions of five phenyl radicals
with glycine and glycine-2,2-d2, as well as with alanine, valine,
proline, cysteine, and methionine, were examined in the gas
phase.12b These studies demonstrated that the R-carbon is not
the only site in glycine and labeled glycine from which a
hydrogen atom is abstracted.12b The less electrophilic radicals
were found to react with the amino acids with no sulfur atoms
only by hydrogen atom abstraction. However, the more electro-
philic radicals also react via NH2 group abstraction (the occur-
rence of this reaction demonstrates that the amino acids are not
in zwitterionic forms in these experiments). NH2 abstraction has
been proposed to occur via a nucleophilic additionꢀelimination
mechanism.12b The electrophilicity of the radicals (quantified by
calculated electron affinities (EAs) of the radical sites12c) and the
ionization energies of the amino acids were found to be the major
reactivity-controlling parameters.12b Various other possible re-
activity-controlling factors, such as the reaction enthalpy, were
not found to play an important role here, in agreement with the
literature on polar radicals.13,14
(Chart 1), many of which are partially deuterium- or 15N-labeled,
and a more diverse group of phenyl radicals (Chart 2).
’ EXPERIMENTAL SECTION
All experiments were carried out in a Finnigan FTMS 2001 dual-cell
FT-ICR mass spectrometer. The amino acids were introduced into the
mass spectrometer by using a thermal solids probe. Depending on the
identity of the amino acid, the manual probe was heated to 140ꢀ200 °C.
Observation of an abundant protonated amino acid upon reaction with
protonated acetone (proton affinity15 (PA) 194 kcal/mol) confirmed
that the amino acids were introduced into the instrument without
thermal decomposition. The proton affinities of the amino acids are
211.9 kcal/mol for L-glycine, 238 kcal/mol for L-lysine, 220 kcal/mol for
L-proline, 218.6 kcal/mol for L-leucine, and 219.3 kcal/mol for L-
isoleucine.15 Only minor fragmentation was observed in the reactions
of the amino acids with protonated acetone.
The radicals were generated as described previously.11,12 For example,
N-(2,3,5,6-tetrafluoro-4-dehydrophenyl)pyridinium ion (radical a;
Chart 2) was formed by introducing pyridine and 1,4-diiodotetrafluo-
robenzene into the same cell through two variable leak valves. An
electron beam of 20ꢀ25 eV kinetic energy was used to ionize both
species; the filament current was 7 μA and the ionization time 1 s. This
ionization yielded molecular ions as well as various fragment ions from
both species. The 1,4-diiodotetrafluorobenzene radical cation was
isolated by ejecting all other ions via the application of a stored-
waveform inverse Fourier transform16 (SWIFT) excitation pulse that
was applied to the plates of the cell. The 1,4-diiodotetrafluorobenzene
radical cation was then allowed to react for 1ꢀ20 s with pyridine to form
the N-(2,3,5,6-tetrafluoro-4-iodophenyl)pyridinium cation by substitu-
tion of one of the iodine atoms in 1,4-diiodotetrafluorobenzene with
pyridine. The N-(2,3,5,6-tetrafluoro-4-iodophenyl)pyridinium cation
was transferred into the other cell by grounding the conductance limit
plate for about 140 μs. The technique of sustained off-resonance
irradiated collision-activated dissociation17 (SORI-CAD) with argon
target gas was used to homolytically cleave the remaining car-
bonꢀiodine bond. In SORI-CAD, the peak pressure of argon in the
To probe the generality of the above findings, and to further
explore the regioselectivity of the radicals, we have now extend-
ed this study to several previously unexamined amino acids
9332
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
Chart 2. Structures and Calculated Vertical EAs20 of the Positively Charged Phenyl Radicals Studied
cell was about 1 ꢁ 10ꢀ5 Torr. The ions were collisionally activated with
argon for 0.5ꢀ1 s at a frequency 1000 Hz above the cyclotron frequency
of the ions. All other radicals were generated using a similar procedure
except that different precursors were used: 3-iodopyridine for b,
3-iodopyridine and bromobenzene for c, pyridine and 3,5-dinitroiodo-
benzene for d, pyridine and 3,5-dichloroiodobenzene for e, 3-fluoropyr-
idine and 1,3-diiodobenzene for f, pyridine and 1,3-diiodobenzene for g,
pyridine and 1,4-diiodobenzene for h, quinoline and 1,4-diiodobenzene
for i, and pyridine and 4,40-diiodobiphenyl for j.
The radicals of interest were isolated by ejecting all other ions via the
application of a series of SWIFT excitation pulses to the plates of the cell.
The isolated charged phenyl radicals were allowed to react with each
amino acid for a variable period of time (typically 0.5ꢀ1000 s).
Detection was performed by using “chirp” excitation of 124 V amplitude,
2.7 MHz bandwidth, and 3.2 kHz/ms sweep rate. All of the spectra are
the average of 10 transients, which were recorded as 64k data points and
subjected to one zero fill prior to Fourier transformation. Each reaction
spectrum was background corrected by using a procedure described
previously.18
were calculated using a parametrized trajectory theory.21 The efficiency
of each reaction (the fraction of collisions that leads to reaction) is given
by kexp/kcoll. The primary products’ relative abundances are reported as
branching ratios, which are given as the ratio of a primary product ion’s
abundance to the sum of all primary product ions’ abundances.
All amino acids (purity g98.5%), except for DL-lysine-ε-15N and DL-
glycine-2,2-d2, were obtained from Fluka Biochemika. DL-Lysine-ε-15N
(purity g98.5%, isotopic purity (15N) 98þ%) was obtained from
Cambridge Isotope Laboratories, Inc. and DL-glycine-2,2-d2 (purity
g98.5%, isotopic purity (D) 98%) from Sigma-Aldrich. DL-Proline-1-
d1 was synthesized from L-proline by using racemization via a Schiff base
procedure.22 This procedure involves selective exchange of the R-carbon
hydrogen atom of L-proline with deuterium by racemization of
this stereogenic center (R-carbon) in monodeuterated acetic acid
(CH3CO2D). A vial with 500 mg of L-proline was refluxed with
0.04 mL of salicylaldehyde and 9.4 mL of monodeuterated acetic acid at
100 °C for 1 h. The deuterated amino acid was recrystallized. The
1
product’s structure was verified by H NMR and mass spectrometry
(isotopic purity 95 atom % D).
In the FT-ICR mass spectrometer, the concentration of ions (in this
case, charged monoradicals) inside the cell is much smaller than that of
neutral molecules (amino acid molecules). Hence, the concentration of
the amino acid can be assumed to be constant. Indeed, all the reactions
studied followed pseudo-first-order kinetics, which allowed for the
derivation of the second-order reaction rate constant (kexp) from a
semilogarithmic plot of the relative abundance of the reactant ion versus
reaction time and the concentration of the amino acid. The pressure
readings inside the cell were measured by two ionization gauges located
on each side of the dual cell. The ion gauge pressure readings were
corrected for the sensitivity of the ion gauges toward each amino acid
and for the pressure gradient between the ion gauge and the cell. The
correction factors were obtained by measuring the reaction rate of an
exothermic proton transfer reaction from protonated acetone or proto-
nated methanol to the given amino acid.19 Such reactions can be
expected to occur at the collision rate.20 The accuracy of the measured
rate constants is estimated to be about 50%, with the precision estimated
Quantum chemical calculations were performed with the Gaussian
9823 electronic structure program suite. Molecular geometries for the
radicals were calculated as previously described.12c Charge and spin
densities for the radicals were calculated at the UBPW91/cc-pVDZ level
of theory. Relative transition-state energies for addition of ammonia to
different positions in the 3-dehydropyridinium ion were calculated at the
UHF/6-31G(d,p)//UHF/6-31G(d,p) level of theory.
’ RESULTS AND DISCUSSION
The 10 positively charged phenyl radicals used in this study
(aꢀj) are shown in Chart 2. These radicals were generated by
using literature methods.11,12 The isolated charged phenyl radi-
cals were allowed to react with L-glycine, DL-glycine-2,2-d2, L-
lysine, 15N-labeled DL-lysine, L-isoleucine, L-leucine, DL-leucine-
1-d1, L-proline, and DL-proline-1-d1 (Chart 1) for variable periods
of time to determine the reaction efficiencies and product
branching ratios (Tables 1ꢀ5).
to be better than 20%. The theoretical collision rate constants (kcoll
)
9333
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
Table 1. Reaction Efficiencies (Eff.) and Product Branching Ratios for Reactions of Monoradicals aꢀj with Glycine and
Glycine-2,2-d2
a Calculated vertical EAs from ref 12c. b abs. = abstraction. c Data taken from refs 12b and 12c. d The efficiency (%) is actually 0.00003.
Table 2. Reaction Efficiencies (Eff.) and Product Branching Ratios for Reactions of Monoradicals aꢀg with L-Leucine,
DL-Leucine-1-d1, and L-Isoleucine
a Calculated vertical EAs from ref 12c. b abs. = abstraction.
The ionic avoided curve crossing model14b,c predicts a strong
relationship between the vertical EA (the energy released upon
attachment of an electron to the radical site with no geometry
change) of an electrophilic radical and its gas-phase radical
reactivity.24 The higher the EA of the radical, the faster the
reactions are predicted to be. This has been found to be true
for many reactions. Hence, favorable polarization of the transi-
tion state appears to be a crucial controlling factor in these
reactions.24 Consequently, radicals with widely different EA
values, ranging from 3.31 to 6.18 eV (Chart 2), were chosen
9334
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
Table 3. Reaction Efficiencies (Eff.) and Product Branching Ratios for Reactions of Monoradicals aꢀh with L-Lysine and
DL-Lysine-ε-15N
a Calculated vertical EAs from ref 12c. b abs. = abstraction.
Table 4. Reaction Efficiencies (Eff.) and Product Branching Ratios for Reactions of Monoradicals aꢀh with L-Proline and
DL-Proline-1-d1
a Calculated vertical EAs from ref 12c. b abs. = abstraction. c Distinguishing between D atom abstraction (2.014102 amu) and 2 H atom abstraction
(2.01588 amu) is difficult due to the closeness of their masses and the detection limits of the mass spectrometer.
for this study. The vertical EAs of the radicals were calculated
earlier at the (U)B3LYP/6-31þG(d) level of theory.12c
while the more electrophilic radicals also show NH2 abstraction.
The relative reactivity of the radicals was found to depend on
their EA, and was independent of the types of reactions observed.
This is intriguing as these reactions occur via drastically different
mechanisms. While hydrogen atom transfer is a radical reaction,
the mechanism for NH2 transfer is thought to be a nucleophilic
The amino acids L-glycine, L-lysine, L-isoleucine, and L-leucine
react with the radicals predominantly by hydrogen atom transfer
and additionꢀelimination pathways (e.g., NH2 transfer). The
less electrophilic radicals only exhibit hydrogen atom abstraction,
9335
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
Table 5. Reaction Efficiencies (Eff.) and Product Branching Ratios for Reactions of Monoradical c with L-Proline and
DL-Proline-1-d1
a Calculated vertical electron affinity from ref 12c. b abs. = abstraction.
additionꢀelimination reaction rather than a radical reaction
(illustrated for lysine in Scheme 1 and discussed in detail
below).12b In general, for the radicals and amino acids studied
here, more NH2 abstraction and less hydrogen atom abstraction
were observed as the EA of the radical increased. This finding is in
agreement with a literature report that, as the EA of a radical
increases, its addition reactions become faster than hydrogen
atom abstraction.11 While NH2 abstraction is the most common
group abstraction observed, other group abstraction products
were observed for reactions of L-proline and DL-proline-1-d1 with
radical c. The reactivities observed for the different amino acids
are discussed in detail below.
The hydrogen atom abstraction efficiency of radicals dꢀi
increases with increasing EA of the radical.
No reactions were observed for radical j. This lack of reactivity
is rationalized by its very low EA of 3.31 eV (the other radicals
have EAs ranging from 4.50 to 6.18 eV). In agreement with the
present findings, previous studies12c have shown that radical j
abstracts a hydrogen atom from cyclohexane (IE = 10.32 eV) and
isopropyl alcohol (IE = 10.44 eV) only at very low reaction
efficiencies of 0.0034 and 0.0029, respectively.
Previous examination of the reactions of OH radicals with
glycine in solution has shown that hydrogen atoms can be
abstracted from either the NH2 or OH group in glycine, in
addition to the R-carbon, which is thermodynamically favored.25
This occurs in spite of the fact that the homolytic CꢀH bond
dissociation energy in glycine (79.2 kcal/mol) is lower than that
of the NꢀH bond (102.6 kcal/mol), which is substantially lower
than that of the OꢀH bond (112.9 kcal/mol).26
On the basis of studies on polar effects in radical reactions,12c
amino acids with lower ionization energies (IEs) are expected to
react faster via hydrogen atom transfer than those with higher IE
values. The preliminary study on a smaller group of amino acids
supported this expectation.12b However, in the present study,
only some of the radicals were found to follow this trend. For
example, although radical e abstracts a hydrogen atom fastest
from proline (IE = 8.3 eV) and slowest from glycine (IE = 8.9 eV),
radical a reacts at similar efficiencies with glycine and lysine
(IE = 8.6 eV), in spite of their quite different IE values.
L-Glycine and DL-Glycine-2,2-d2. The more electrophilic
radicals aꢀc display two pathways upon reaction with glycine,
namely, hydrogen atom abstraction and NH2 abstraction
(Table 1). The greater the EA of the radical, the more the
NH2 abstraction is favored and the greater the total reaction
efficiency. The less electrophilic radicals dꢀi react with L-glycine
solely by hydrogen atom abstraction (Table 1). This finding is in
agreement with previous studies12b that have shown that radicals
d and e only react by hydrogen atom abstraction with glycine.
Glycine-2,2-d2 was allowed to react with radicals aꢀj to obtain
a better understanding of which position is preferred for hydro-
gen atom abstraction from glycine. All these reactions, except
those of radicals aꢀc, occur via both hydrogen and deuterium
atom abstraction (NH2 abstraction was also observed for radicals
aꢀc, Table 1), and hydrogen atom abstraction dominates over
deuterium atom abstraction in most cases. Hence, hydrogen
atom abstraction also occurs from other sites besides the R-
carbon. Since abstraction of a deuterium atom is likely to be
slower than of a hydrogen atom from the same site due to the
primary kinetic isotope effect,24b abstraction from the R-carbon
is seemingly less favored than it actually is.
For the more electrophilic monoradicals aꢀc, a trend was
observed between the EA of the radical and hydrogen/deuterium
9336
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
Scheme 1. Possible Mechanisms for the NH2 Abstraction Reactions of Radical c with Lysine-ε-15N
atom vs NH2 abstractions (as the EA of the radical decreases, the
extent of both deuterium and hydrogen atom abstractions
increases relative to that of NH2 abstraction; however, the total
reaction efficiency goes down). This competition between
hydrogen/deuterium abstraction and NH2 abstraction is not
present for radicals dꢀj, where hydrogen and deuterium atom
abstractions are the only observed reaction pathways. When NH2
group abstraction is not observed, there is no correlation
between EA and deuterium vs hydrogen atom abstraction. A
similar finding was reported earlier for ethanol.24a However, a
correlation has been reported in the literature for 2-propanol
(less electrophilic radicals are more selective for hydrogen atoms
bound to the R-carbon).24b
For a given radical, abstraction of a hydrogen atom from the R-
carbon rather than from the NH2 group is not only thermodyna-
mically but also kinetically favored. Previously calculated transition-
state energies for radicals aꢀc, e, and g show that the transition
state for hydrogen atom abstraction from the R-carbon in glycine is
2ꢀ3 kcal/mol12b lower than from the NH2 group (calculations
were not performed for the OH group). However, both the present
and previous results show that the R-carbon in glycine is not the
only site from which a hydrogen atom is abstracted.12b
The site at which the initial attack occurs in the NH2
abstraction reaction pathway is likely the radical site since this
readily leads to a homolytic CꢀN bond cleavage in the amino
acid (Scheme 1), and charged aromatic compounds with no
radical sites do not undergo this reaction.12b This is in spite of the
fact that, for different radicals, the radical site may or may not be
the most electrophilic site. This was explored by calculating the
charge densities for 2-, 3-, and 4-dehydropyridinium cations at
the UBPW91/cc-pVDZ//UBPW91/cc-pVDZ level of theory
(Figure 1). For 2-dehydropyridinium cation (radical site charge
density 0.162) and 4-dehydropyridinium cation (radical site
charge density 0.126), the radical sites are the most electrophilic
sites. Thus, for these two isomers, the radical site is the preferred
site for nucleophilic attack. For 3-dehydropyridinium cation, the
radical site (charge density 0.033) is less electrophilic than the
2-position (charge density 0.076) or the 4-position (charge
density 0.087; in spite of a smaller positive charge, the 2-position
may be favored over the 4-position for nucleophiles that can form
a stabilizing hydrogen bond with the NHþ group). Although the
2- and 4-positions are calculated to be the most electrophilic sites,
addition also must occur at the 3-position since a homolytic
CꢀN bond cleavage was observed to occur for the adduct.
9337
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
the hydrogen atoms bound to the nitrogen atom, and the odd spin
is delocalized over the pyridinium ring, in agreement with the
Lewis structure presented for an analogous adduct in Scheme 1.
These findings suggest that the nucleophilic attack of amino acids
to the charged phenyl radicals, leading to NH2 group abstraction,
occurs at the radical site and yields an adduct with the odd spin
delocalized over the pyridine ring.
L-Leucine, DL-Leucine-1-d1, and L-Isoleucine. The difference
in the alkyl side chain for isomeric L-leucine and L-isoleucine does
not substantially influence their reactivity (Table 2). Just as for
glycine, both NH2 abstraction and hydrogen atom abstraction
reactions were observed when this group of amino acids was
allowed to react with the highly electrophilic charged radicals
a and c (Table 2). Radical a shows more NH2 abstraction than
radical c. Radical b is an exception to the above behavior.
This radical reacts by proton transfer with these amino acids
(L-leucine, PA = 218.6 kcal/mol;15 L-isoleucine, PA = 219.3 kcal/
mol15) because the amino acids are more basic than the
conjugate base of radical b.12b The proton transfer is so facile
that it suppresses all radical reactions. Radicals dꢀg, with
relatively low EA, react exclusively by hydrogen atom abstrac-
tion (Table 2). The lower the EA of the radical, the slower the
hydrogen atom abstraction.
Figure 1. Calculated (UBPW91/cc-pVDZ//UBPW91/cc-pVDZ)
charges for heavy atoms of 2-, 3-, and 4-dehydropyridinium cations.
DL-Leucine-1-d1 was allowed to react with radicals aꢀg to
obtain a better understanding of the selectivity of hydrogen atom
abstraction from leucine. The radicals (with the exception of
radical c) did not abstract a deuterium atom from the R-carbon of
the partially labeled leucine, as expected, but instead a hydrogen
atom from somewhere else in the molecule. Although abstraction
of the deuterium atom is slowed by the primary kinetic isotope
effect, this finding nevertheless suggests that the R-carbon is not
the kinetically preferred site for hydrogen atom abstraction from
this amino acid. This finding is in sharp contrast to the behavior
of partially labeled glycine, where extensive deuterium atom
abstraction from the R-carbon was observed. Apparently, the
presence of several hydrogen atoms in the side chain of leucine,
as opposed to only one at the R-carbon, makes hydrogen atom
abstraction from the side chain kinetically favored. Furthermore,
the side chain may sterically hinder abstraction of the hydrogen
atom from the R-carbon.
Reactions of Phenyl Radicals with L-Lysine and DL-Lysine-
ε-15N. As observed for leucine and isoleucine, radical b (PA = 216
kcal/mol), the only Br€onsted acid among the radicals studied,
reacts exclusively by proton transfer with L-lysine (PA = 238 kcal/
mol15). Furthermore, as reported for glycine, leucine, and
isoleucine above, NH2 and hydrogen atom abstractions were
observed for the most electrophilic radicals a and c with lysine
(Table 3), while the less reactive radicals dꢀh exhibit only
hydrogen atom abstraction.
Figure 2. Calculated (UBPW91/cc-pVDZ//UBPW91/cc-pVDZ)
charges (top) and spin density (bottom) for the NH3 adduct of
3-dehydropyridinium cation.
Radical d deviates from the other radicals in that it displays an
additional reaction pathway—abstraction of two hydrogen
atoms from L-lysine and DL-lysine-ε-15N (Table 3). This reaction
is as fast as a single hydrogen atom abstraction. Apparently, after
hydrogen atom abstraction by the radical site, the NO2 group can
abstract another hydrogen atom from the lysine radical. The
same reaction was observed for radical d upon interaction with
proline (as discussed below) but not for other amino acids.
A possible mechanism for proline is shown in Scheme 2.
Radical c forms a stable adduct with L-lysine and DL-lysine-
ε-15N, with branching ratios of 32% and 23%, respectively
(Table 3). The formation of a stable adduct is most likely due
to (1) the presence of two amino groups in lysine (since only this
The relative transition-state energies calculated at the UHF/6-
31G(d,p)//UHF/6-31G(d,p) level of theory for addition of
ammonia (instead of an amino acid to simplify the calculations)
to 3-dehydropyridinium cation are 0.5 kcal/mol (C-2), 0.0 kcal/
mol (C-3), 1.4 kcal/mol (C-4), 19.5 kcal/mol (C-5), and
2.5 kcal/mol (C-6). Thus, nucleophilic attack at the 3-position
has the lowest transition-state energy, making addition of
ammonia to the 3-position kinetically favored for radical b,
although additions to the other sites may also occur. Figure 2
shows the calculated charge and spin densities for the ammonia
adduct of 3-dehydropyridinium cation. On the basis of these
results, it seems likely that most of the positive charge resides on
9338
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
Scheme 2. Possible Mechanism for the Abstraction of Two Hydrogen Atoms by Radical d from Proline
Scheme 3. Possible Mechanism for the Formation of a Stable Adduct between Radical c and Lysine
amino acid yields an abundant stable adduct) and 2) the intrinsic
properties of radical c (the only radical that forms a stable
adduct). With the exception of radical b, which only undergoes
proton transfer reactions, radical c is the only radical studied here
wherein the radical site is in the same ring as the charge site.
Hence, lysine can interact with both these sites of the radical.
9339
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
A possible mechanism for the formation of a stable adduct (m/z
301) for radical c with lysine only is shown in Scheme 3. The
initial addition is proposed to occur as in Scheme 1 and involve
the amino group in the lysine side chain since this group should
be more nucleophilic than that in the amino acid moiety (the
adjacent electron-withdrawing carbonyl group reduces its
nucleophilicity). However, addition of either amino group
should be followed by the same reaction sequence. After addi-
tion, the second amino group may abstract the proton from the
protonated amino group attached to the aromatic ring. This
reversible step would explain why this reaction was only observed
for lysine. An irreversible hydrogen atom transfer to the radical
site (now delocalized over the ring) would generate a resonance-
stabilized radical cation. Some support for this mechanism was
obtained when the stable adduct ion (m/z 301) was subjected to
collision-activated dissociation. It does not fragment to yield back
the reactant monoradical c via loss of intact lysine, which
demonstrates that the adduct has not retained the initially
formed structure (shown at the left in the middle row in
Scheme 3). It also does not fragment via loss of NH3, which
indicates that it does not contain a protonated amino group
formed upon proton transfer to the second amino functionality
(shown at the right in the middle row in Scheme 3). Instead, the
stable adduct (m/z 301) fragments by loss of an NH2 group (to
yield an ion of m/z 285) and loss of CO2 (to yield an ion of m/z
257), in support of the final structure proposed for the adduct in
Scheme 3.
Scheme 4. Possible Mechanism of C2H4N Abstraction upon
the Reaction of Radical d with Proline
DL-Lysine-ε-15N was studied to provide support for the
mechanism shown in Scheme 3. Indeed, radical c was found to
abstract either one of the two NH2 groups in DL-lysine-ε-15N,
with branching ratios for 15NH2 and NH2 group abstractions of
8% and 21%, respectively. Radical a also abstracts 15NH2 and
NH2 groups, with branching ratios of 18% and 53%, respectively.
On the basis of these four values, it can be concluded that the
electrophilicity of the radical does not have a major influence on
the radicals’ preference toward side chain NH2 abstraction versus
NH2 abstraction from the amino acid moiety. Scheme 1 shows
possible mechanisms for NH2 and 15NH2 abstraction from DL-
lysine-ε-15N. Abstraction of the NH2 group from the amino acid
moiety is more facile than the abstraction of the 15NH2 group
(Table 3), likely due to the different stabilities of the radicals
produced in these reactions. The radical produced upon the
abstraction of the NH2 group of the amino acid moiety is
resonance stabilized and hence more stable than the radical
produced in the abstraction of the NH2 group from the
side chain.
L-Proline and DL-Proline-1-d1. Radical b again reacts only by
proton transfer when allowed to react with proline (Table 4). In
contrast, radicals a and eꢀh react with proline via exclusive
hydrogen atom abstraction (Table 4). NH2 abstraction from
proline does not occur because the nitrogen atom is a part of a
five-membered ring. DL-Proline-1-d1 was allowed to react with
radicals aꢀh to obtain a better understanding of the selectivity of
hydrogen atom abstraction from proline. Only a minor amount
of deuterium atom abstraction was observed (Tables 4 and 5).
Hence, hydrogen atom abstraction from the unlabeled proline
mostly occurs from positions other than the R-carbon.
likely the same for proline and lysine (a possible mechanism is
shown in Scheme 2). Steric hindrance due to the methyl branch
may slow this reaction for leucine and isoleucine. As opposed to
an earlier study,12b wherein only hydrogen atom abstraction was
reported for the reaction of radical c with proline, additional
reaction pathways were observed in the present study. Although
hydrogen atom abstraction was found to dominate, HO abstrac-
tion, C2H4N abstraction, C5H7NO abstraction, and stable ad-
duct formation were also observed. This discrepancy between the
two studies may arise from the greater sensitivity of the more
recent measurements. The different reactivity of radical c, when
compared to the other radicals studied, may again be explained
by the fact that radical c is the only radical studied (with the
exception of acidic radical b) that has the radical site and the
positive charge in the same ring, hence allowing interactions of
proline with both. For example, C2H4N abstraction from proline
by radical c may be initiated by nucleophilic addition of the
proline nitrogen to the radical carbon, as shown in Scheme 1 for
lysine. For lysine, NꢀC bond cleavage leads to the NH2
abstraction product. For proline, this is not possible. Instead,
the radical (now delocalized over the ring) may abstract a
hydrogen atom from the carboxylic acid group of proline, leading
to the elimination of CO2 (Scheme 4). Finally, C2H4 may be
eliminated from the ring in proline, resulting in a resonance-
stabilized radical cation (the C2H4N abstraction product).
’ CONCLUSIONS
Examination of the gas-phase reactions of 10 different phenyl
radicals with several amino acids, many partially isotope labeled,
revealed the commonly observed hydrogen atom abstraction, but
also a number of other reactions. Results obtained for partially
deuterium-labeled leucine and proline demonstrate that only a
very small amount of hydrogen atoms, if any, are abstracted from
Radicals c and d (and b, as discussed above) are the only
radicals that undergo another reaction with proline besides
hydrogen atom abstraction. Radical d abstracts two hydrogen
atoms from L-proline. The abstraction of two hydrogen atoms by
d was also observed for lysine. The mechanism for this reaction is
9340
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
the R-carbon and a large amount from elsewhere (likely the alkyl
side chain). This is in sharp contrast to glycine whose R-carbon is
the predominant hydrogen atom donor site. The regioselectivity
of hydrogen atom abstraction appears to be independent of the
structure of the radical but dependent on the structure of the
amino acid.
(4) Stefanic, I.; Bonifacic, M.; Asmus, K. D.; Armstrong, D. A. J. Phys.
Chem. A 2001, 105, 8681–8690.
(5) (a) Hawkins, C. L.; Davies, M. J. Biochim. Biophys. Acta 2001,
1504, 196–219. (b) Hawkins, C. L.; Davies, M. J. J. Chem. Soc., Perkin
Trans. 1998, 2, 2617–2622.
(6) (a) Braslau, R.; Anderson, M. O. Tetrahedron Lett. 1998,
39, 4227–4230. (b) Jalkanen, K. J.; Elstner, M.; Suhai, S. J. Mol. Struct.
2004, 675, 61–77.
(7) Smith, G.; Leary, J. A. J. Am. Chem. Soc. 1998, 120, 13046–
13056.
(8) Tao, A. W.; Zhang, D.; Feng, W.; Thomas, P. D.; Cooks, R. G.
Anal. Chem. 1999, 71, 4427–4429.
(9) (a) Fales, H. M.; Wright, G. J. J. Am. Chem. Soc. 1977,
99, 2339–2440. (b) Piccirillo, S.; Bosman, C.; Toja, D.; Giardini-
Guidoni, A.; Pierini, M.; Trolani, A.; Speranza, M. Angew. Chem., Int.
Ed. Engl. 1997, 36, 1729–1731. (c) Smith, G.; Leary, J. A. J. Am. Chem.
Soc. 1998, 120, 13046–13056.
(10) (a) Zein, N.; Casazza, A. M.; Doyle, T. W.; Nadler, S. G. Proc.
Natl. Acad. Sci. U.S.A. 1993, 90, 8009–8012. (b) Zein, N.; Reiss, P.;
Bernatowicz, M.; Bolgar, M. Chem. Biol. 1995, 2, 451–455. (c) Zein, N.;
Schroeder, D. R. In Advances in DNA Sequence Specific Agents; Jones,
G. B., Ed.; JAI Press: Greenwich, CT, 1998; Vol. 3, p 201. (d) Zein, N.;
Solomon, W.; Casazza, A. M.; Kadow, J. F.; Krishnan, B. S.; Tun, M. M.;
Vyas, D. M.; Doyle, T. W. Bioorg. Med. Chem. Lett. 1993, 3, 1351–1356.
(e) Jones, G. B.; Plourde, G. W.; Wright, J. M. Org. Lett. 2000,
2, 811–813.
(11) (a) Ramirez-Arizmendi, L. E.; Guler, L.; Ferra, J. J.; Thoen,
K. K.; Kentt€amaa, H. I. Int. J. Mass Spectrom. 2001, 210/211, 511–520.
(b) Heidbrink, J. L.; Ramirez-Arizmendi, L. E.; Thoen, K. K.; Guler, L.;
Kentt€amaa, H. I. J. Phys. Chem. A 2001, 105, 7875–7884.
(12) (a) Smith, R. L.; Kentt€amaa, H. I. J. Am. Chem. Soc. 1995,
117, 1393. (b) Huang, Y.; Guler, L.; Heidbrink, J.; Kentt€amaa, H. I. J. Am.
Chem. Soc. 2005, 127, 3973–3978. (c) Jing, L.; Nash, J. J.; Kentt€amaa,
H. I. J. Am. Chem. Soc. 2008, 130, 17697–17709.
The mechanism of the previously reported NH2 abstraction
(proposed to occur via the nucleophilic additionꢀelimination
mechanism12b) was examined computationally. The relative
transition-state energies for addition of ammonia to different
positions in 3-dehydropyridinium cation revealed that addition
to the radical carbon is kinetically favored and hence the most
likely addition site for amino acids in charged phenyl radicals.
The calculated charge and spin densities in the ammonia adduct
of 3-dehydropyridinium cation support the Lewis structure
chosen to illustrate the amino acid adducts of the radicals.
Several unprecedented reaction pathways were also observed.
These include the abstraction of two hydrogen atoms from
proline and lysine by the nitro-substituted radical d. Further-
more, abstractions of C2H4N, C5H7NO, and OH groups by
radical c were observed for proline. These reactions are likely to
occur by nucleophilic additionꢀelimination pathways, similar to
that leading to NH2 abstraction from all the other amino acids
but proline, where the nitrogen atom is part of a ring structure.
Finally, both NH2 and 15NH2 groups were abstracted from lysine
labeled with 15N on the side chain, indicating that NH2 abstrac-
tion occurs both from the amino terminus and from the side
chain of lysine.
The electron affinity of the radical appears to be the major
factor in controlling the radical’s reaction rates with the amino
acids. Both the radical (hydrogen atom abstraction) and non-
radical (NH2 abstraction) reaction efficiencies were found to
depend on the electrophilicity of the radical, although the
nonradical reactions are influenced more strongly. However, in
contrast to an earlier report,12b the ionization energies of the
amino acids do not appear to have a general reactivity-
controlling role.
(13) Pross, A. Theoretical and Physical Principles of Organic Reactivity;
John Wiley & Sons: New York, 1995.
(14) (a) Heberger, K.; Lopata, A. J. Org. Chem. 1998, 63, 8646.
(b) Donahue, N. M. J. Phys. Chem. A 2001, 105, 1489–1497. (c)
Donahue, N. M.; Clarke, J. S.; Anderson, J. G. J. Phys. Chem. A 1998,
102, 3923–3933. (d) Fossey, J.; Lefort, D.; Sorba, J. Free Radicals in
Organic Chemistry; Wiley: New York, 1997. (e) Jing, L.; Nash, J. J.;
Kentt€amaa, H. I. J. Am. Chem. Soc. 2008, 130, 17697. (f) Adeuya, A.;
Price, J.; Jankiewicz, B. B. J.; Nash, J. J.; Kentt€amaa, H. I. J. Phys. Chem. A
2009, 113, 13663. (g) Taylor, M. S.; Ivanic, S. A.; Wood, G. P. F.; Easton,
C. J.; Bacskay, G. B.; Radom, L. J. Phys. Chem. A 2009, 113, 11817.
(h) Beare, K. D.; Coote, M. L. J. Phys. Chem. A 2004, 108, 7211.
(15) Linstrom, P. J., Mallard, W. G., Eds. NIST Chemistry WebBook,
NIST Standard Reference Database Number 69; National Istitute of
Standards and Technology: Gaithersburg, MD, March 2003; http://
webbook.nist.gov.
’ ASSOCIATED CONTENT
S
Supporting Information. Absolute energies and coordi-
b
nates of atoms for all optimized structures and complete ref 23.
This material is available free of charge via the Internet at http://
pubs.acs.org.
’ AUTHOR INFORMATION
(16) Chen, L.; Wang, T.-C. L.; Ricca, T. L.; Marshall, A. G. Anal.
Chem. 1987, 59, 449–454.
(17) Gauthier, J. W.; Trautman, T. R.; Jacobson, D. B. Anal. Chem.
1991, 246, 211–225.
Corresponding Author
hilkka@purdue.edu
(18) Leeck, D. T.; Stirk, K. M.; Zeller, L. C.; Kiminkinen, L. K. M.;
Castro, L. M.; Vainiotalo, P.; Kentt€amaa, H. I. J. Am. Chem. Soc. 1994,
116, 3028–3030.
’ ACKNOWLEDGMENT
(19) Huang, Y.; Kentt€amaa, H. I. J. Am. Chem. Soc. 2005, 127,
7952–7960.
(20) Thoen, K. K.; Smith, R. L.; Nousiainen, J. J.; Nelson, E. D.;
Kentt€amaa, H. I. J. Am. Chem. Soc. 1996, 118, 8669–8676.
(21) Su, T.; Chesnavich, W. J. J. Chem. Phys. 1982, 76, 5183–
5185.
(22) Mitulovi, G.; Lammerhofer, M.; Maier, N. M.; Linder, W.
J. Labelled Compd. Radiopharm. 2000, 43, 449–461.
(23) Frisch, M. J.; et al. Gaussian 98, revision A.11.3; Gaussian, Inc.:
Pittsburgh, PA, 2002 (for complete reference, see the Supporting
Information).
Financial support for this work provided by Purdue University
and the National Institutes of Health is gratefully acknowledged.
’ REFERENCES
(1) Bonifacic, M.; Stefanic, I.; Hug, G. L.; Armstrong, D. A.; Ausmus,
K. D. J. Am. Chem. Soc. 1998, 120, 9930–9940.
(2) Maleknia, S. D.; Brenowitz, M; Chance, M. R. Anal. Chem. 1999,
71, 3965–3973.
(3) Nukuna, B. N.; Goshe, M. B.; Anderson, V. E. J. Am. Chem. Soc.
2001, 123, 1208–1214.
9341
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342

Journal of the American Chemical Society
ARTICLE
(24) (a) Jing, L.; Guler, L.; Nash, J. J.; Kentt€amaa, H. I. J. Am. Soc.
Mass Spectrom. 2004, 15, 913–919. (b) Jing, L.; Guler, L. P.; Pates, G.;
Kentt€amaa, H. I. J. Phys. Chem. A 2008, 112, 9708–9715.
(25) Goshe, B. M.; Chen, H. Y.; Anderson, E. V. Biochemistry 2000,
39, 1761–1770.
(26) Yu, D.; Rauk, A.; Armstrong, D. A. J. Am. Chem. Soc. 1995,
117, 1789–1796.
9342
dx.doi.org/10.1021/ja111280t |J. Am. Chem. Soc. 2011, 133, 9331–9342