Isomer differentiation via collision-induced dissociation: The case of protonated a-, β2- and β3-phenylalanines and their derivatives

Article abstract of DOI:10.1002/rcm.4576

A combination of electrospray ionisation (ESI), multistage and high-resolution mass spectrometry experiments is used to examine the gas-phase fragmentation reactions of the three isomeric phenylalanine derivatives, a-phenylalanine, b²-phenylalanine and b³-phenylalanine. Under collision-induced dissociation (CID) conditions, each of the protonated phenylalanine isomers fragmented differently, allowing for differentiation. For example, protonated b³-phenylalanine fragments almost exclusively via the loss of NH₃, only b²-phenylalanine via the loss of H₂O, while a- and b²- phenylalanine fragment mainly via the combined losses of H₂ORCO. Density functional theory (DFT) calculations were performed to examine the competition between NH₃ loss and the combined losses of H₂O and CO for each of the protonated phenylalanine isomers. Three potential NH₃ loss pathways were studied: (i) an aryl-assisted neighbouring group; (ii) 1,2 hydride migration; and (iii) neighbouring group participation by the carboxyl group. Finally, we have shown that isomer differentiation is also possible when CID is performed on the protonated methyl ester and methyl amide derivatives of a-, b²- and b³-phenylalanines.

Full text of DOI:10.1002/rcm.4576

RAPID COMMUNICATIONS IN MASS SPECTROMETRY
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/rcm.4576
Isomer differentiation via collision-induced dissociation:
The case of protonated a-, b2- and b3-phenylalanines and
their derivativesy
Adrian K. Y. Lam1–3 and Richard A. J. O’Hair1–3
*
1School of Chemistry, University of Melbourne, Victoria 3010, Australia
2Bio21 Institute of Molecular Science and Biotechnology, The University of Melbourne, Victoria 3010, Australia
3ARC Centre of Excellence for Free Radical Chemistry and Biotechnology
Received 17 March 2010; Revised 11 April 2010; Accepted 12 April 2010
A combination of electrospray ionisation (ESI), multistage and high-resolution mass spectrometry
experiments is used to examine the gas-phase fragmentation reactions of the three isomeric phenyl-
alanine derivatives, a-phenylalanine, b2-phenylalanine and b3-phenylalanine. Under collision-
induced dissociation (CID) conditions, each of the protonated phenylalanine isomers fragmented
differently, allowing for differentiation. For example, protonated b3-phenylalanine fragments almost
exclusively via the loss of NH3, only b2-phenylalanine via the loss of H2O, while a- and b2-
phenylalanine fragment mainly via the combined losses of H2O R CO. Density functional theory
(DFT) calculations were performed to examine the competition between NH3 loss and the combined
losses of H2O and CO for each of the protonated phenylalanine isomers. Three potential NH3 loss
pathways were studied: (i) an aryl-assisted neighbouring group; (ii) 1,2 hydride migration; and
(iii) neighbouring group participation by the carboxyl group. Finally, we have shown that isomer
differentiation is also possible when CID is performed on the protonated methyl ester and methyl
amide derivatives of a-, b2- and b3-phenylalanines. Copyright # 2010 John Wiley & Sons, Ltd.
The low-energy fragmentation reactions of protonated a-
amino acids and their peptides are fairly well understood
through the use of a range of tandem mass spectrometry
(MS/MS) techniques and molecular modelling. This has led
to the development of the concept of the mobile proton and
its refinements.1–7 In contrast, there has been considerably
less work done on the gas-phase chemistry of the closely
related b-amino acids, which are: involved in mammalian
metabolism;8–11 components of peptidic and non-peptidic
natural products;12 precursors in the biosynthesis of the
cancer chemotherapy compound taxol.13 With the discovery
that b-amino acid containing peptides are ideal peptidomi-
metic candidates exhibiting a higher protease resistance than
their a-amino acid analogues,14 there has been growing
interest in the biological use of various forms of b-amino
acids. b-Amino acids differ from their a-counterparts in the
presence of an extra methylene moiety between the amino
and carboxyl termini, allowing for two types of b-amino
acids: b2 and b3 (Scheme 1).
of hybrid peptides;21 interaction between various cations and
b3-amino acid containing peptides;22 and differentiation of
positional isomers containing a- and b3-amino acids.23–25
The differentiation of isomeric b2- and b3-amino acids via
low-energy CID does not appear to have been addressed
to date. Here we use a combination of multistage mass
spectrometry experiments, CID, and density functional
theory (DFT) calculations to compare the low-energy
fragmentation reactions of b2-phenylalanine (2) (R2 ¼ Phe)
and b3-phenylalanine (3) (R3 ¼ Phe); for structures, see
Scheme 2. We also compare the fragmentation reactions of
these b-amino acids with those of a-phenylalanine (4,
Scheme 2), where we have previously shown that the main
loss of the combined elements of water and CO proceeds via
sequential bond-cleavage reactions involving ion-molecule
complexes (path A of Scheme 3) and that the minor loss of
ammonia proceeds via a neighbouring group reaction
involving the phenyl ring to give a phenonium ion (path
Apart from a number of studies focused on the gas-phase
chemistry of b-alanine (1, Scheme 2) and its peptides,15–19
only a few more complicated b-amino acids systems have
been examined via MS-based methods. There have been
reports on: the use of MS/MS to sequence synthetic b-
peptides;20 low-energy collision-induced dissociation (CID)
*Correspondence to: R. A. J. O’Hair, School of Chemistry, The
University of Melbourne, Victoria 3010, Australia.
E-mail: rohair@unimelb.edu.au
yPart 71 of the series ‘Gas-Phase Ion Chemistry of Biomolecules’.
Scheme 1. The two types of b-amino acids: b2 and b3.
Copyright # 2010 John Wiley & Sons, Ltd.

1780 A. K. Y. Lam and R. A. J. O’Hair
Scheme 2. Structures of b-alanine (1), b2-phenylalanine (2) (R2 ¼ Phe), b3-
phenylalanine (3) (R3 ¼ Phe) and a-phenylalanine (4).
Scheme 3. Pathways for losses of H2O þ CO and NH3 from a-phenylalanine.
B of Scheme 3) rather than via a 1,2-hydride pathway to yield
the benzyl cation (path C of Scheme 3).26
Synthesis of N-terminal methyl amide
derivatives
Following a variation of the method used by Feenstra et al.,27
2 mg of the unpurified amino acid methyl ester derivative
was dissolved in 1 mL of 30% CH3NH2/H2O and allowed to
stand for 30 min at room temperature. The methyl amide was
then used without further purification.
EXPERIMENTAL
Materials
All purchased materials were used without further purifi-
cation: a-phenylalanine and b3-phenylalanine were obtained
from Sigma-Aldrich (St. Louis, MO, USA) and b2-phenyl-
alanine was purchased from Rare Chemicals GmbH (Kiel,
Germany). Methanol (HPLC grade) and acetonitrile (HPLC
grade) were obtained from Burdick & Jackson (Muskegon,
MI, USA). Acetic acid (glacial) was obtained from BDH
(Poole, UK).
Mass spectrometry experiments
All mass spectrometric experiments were conducted on a LTQ
FT hybrid mass spectrometer (Thermo Scientific, Bremen,
Germany) consisting of a linear ion trap (LTQ) coupled to a
Fourier transform ion cyclotron resonance (FT-ICR) mass
spectrometer. Methanolic samples were introduced into the
electrospray ionisation (ESI) source of the mass spectrometer
at a flow rate of 5 mL/min. The sheath gas flow rate, capillary
voltage and temperature were adjusted to ca. 10 arbitrary
units, 3.0 kV and 2508C, respectively. Low-energy CID studies
using helium as the collision gas were performed within the
LTQ using the standard procedure of mass selection of the
desired precursor ion and subjection of these ions to collisional
activation, in which the activation time for these experiments
was 30ms with the collision energy varied to ensure that 10–
15% of the precursor ion remained.
Synthesis of N-terminal methyl ester
derivatives
Lyophilised amino acids (10 mg) were dissolved in 1 mL of
methyl esterification reagent (prepared by the dropwise
addition of 800 mL of acetyl chloride to 5 mL of anhydrous
methanol with stirring) and allowed to stand for 2 h at room
temperature. The samples were dried by lyophilisation and
used without further purification.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

CID of protonated a-, b2- and b3-phenylalanines 1781
All high-resolution mass spectrometry (HRMS) experiments
were conducted using the FT-ICR mass analyser component of
the LTQ FT instrument. The ions of interest were mass-selected
in the linear ion trap using standard procedures and subjected
to CID. The resulting product ions were transferred to the FT-
ICR cell to generate a high-resolution mass spectrum. Positive
mode calibration was performed via the automatic calibration
function using the recommended LTQ calibration solution,
consisting of caffeine, the small tetrapeptide MRFA, and
Ultramark 1621 solution.
b3A1a ¼ Structure involved in the b3-phenylalanine PES,
Involved in dissociation pathway A, Structure number 1a;
aC2,b3B3 ¼ Structure involved in the a- and b3-phenyl-
alanine PES, Involved in dissociation pathways C and B,
respectively, Structure numbers 2 and 3, respectively.
RESULTS AND DISCUSSION
Gas-phase fragmentation of a-, b2- and b3-
phenylalanine
The MS/MS product ion spectra of protonated a-phenyl-
alanine (4), b2-phenylalanine (2) and b3-phenylalanine (3) are
shown in Figs. 1(A)–1(C). An examination of these spectra
shows that the isomers can be distinguished from each other
based on the different abundances of small molecule losses (e.g.
Theoretical methods
Due to the large number of dihedral angles for these
phenylalanine isomers, it was not feasible to systematically
examine every possible conformer. Thus, a range of possible
conformers for each isomer was constructed through rotation in
908 increments of the principle dihedrals of an initial conformer.
Following geometry optimisation the lowest energy structure
from this conformer search was used to construct the potential
energy surfaces (PES) for the various fragmentation reactions.
Geometry optimisations and electronic energy calculations
were performed using the Gaussian 03 molecular modelling
package.28 The structures of minima and transition states were
optimised at the B3-LYP level of theory29,30 with the 6-
31 þ G(d,p) basis set. All optimised structures were subjected
to vibrational frequency analysis to ensure that they
corresponded to either true minima (no imaginary frequen-
cies) or transition states (one imaginary frequency). Intrinsic
reaction coordinate (IRC) runs were performed on each
transition state, followed by geometry optimisations to ensure
that the states connected to the appropriate reactant and
product(s). The final energies used to calculate the PES were
corrected with the B3-LYP/6-31 þ G(d,p) zero-point
vibrational energies, (ZPVE) (Ereported ¼ Eelectronic þ Ezpve).
Complete structural details can be found in the Supporting
Information. In cases where combined losses of molecules
were modelled (i.e. loss of H2O and CO) for both b2- and b3-
phenylalanine, the initial departing neutral molecule was not
removed from the intermediate ion-molecule complex as it
was found to significantly lower the energies of both this
intermediate and the subsequent transition-state structure(s).
This is in line with the approach adopted in previous work.26
Due to the large number of possible structures involved
in the PES of the three phenylalanine isomers, we have
constructed a nomenclature for each structure based on its
origin and role in the PES. All structures are prefixed with the
identity of the phenylalanine isomer involved (i.e. a-, b2- and
b3). This is followed by a set of letters denoting the role that
the structure plays. Conformers of the phenylalanine isomers
are denoted by use of a lower-case letter. Structures involved
in a fragmentation reaction begin with an upper-case letter
denoting the dissociation pathway, followed by the structure
number. Finally, in cases where an individual ion can
originate from two separate reactions by different phenyl-
alanine isomers, two labels separated by a comma are used.
Examples of the nomenclature used in this text are
summarised below:
Figure 1. LTQ CID MS/MS spectra of the [M þ H]þ ions of:
(A) a-phenylalanine; (B) b2-phenylalanine; and (C) b3-phenyl-
alanine. An asterisk refers to the mass-selected precursor ion.
The z symbol indicates that the assignment of the molecular
formula of the neutral(s) lost has been confirmed via HRMS.
b2a ¼ Structure involved in the b2-phenylalanine PES, Pro-
tonated b2-phenylalanine, conformer a;
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

1782 A. K. Y. Lam and R. A. J. O’Hair
NH3, H2O and H2O, CO). The fragmentation reactions of each
of the three isomers will be discussed in further detail below.
fragmentation channels observed include an abundant loss
of H2O to yield the ion at m/z 148 (Eqn. (4)), together with a
minor loss of NH3 to form m/z 149 (Eqn. (5)). Water loss
appears to be a diagnostic reaction channel that allows
differentiation between the three isomers. It is worth noting
that water loss has also been observed for b-alanine,18,19
which contrasts with the virtual absence of water loss for a-
alanine.35 Possible mechanisms for these small molecule
losses are shown in Scheme 4. By analogy with a-
phenylalanine, the combined loss of H2O and CO may
involve stepwise bond cleavage (Scheme 4, path A). A
difference is that cleavage of the C–O bond gives rise to a
lactam, which is stable and can be observed at m/z 148
(Eqn. (4)). Loss of ammonia can proceed via three different
mechanisms (Scheme 4, paths B–D). The first is a neighbour-
ing group mechanism involving the phenyl group (Scheme 4,
path B), thereby forming the same phenonium ion as for
a-phenylanine (Scheme 3, path B). The second is a 1,2-
hydride shift to give a benzyl cation (Scheme 4, path C). The
final mechanism involves the carboxyl group acting as a
neighbouring group (Scheme 4, path D), to give a lactone.
Each of these pathways was investigated via DFT calcu-
lations and they are discussed in detail below.
a-Phenylalanine (Fig. 1(A))
Under low-energy CID conditions, the linear ion trap MS/
MS spectrum for protonated a-phenylalanine (4) is almost
identical to that observed previously in the 3D ion-trap.26
Thus loss of H2O þ CO (Eqn. (1)) to form the ion at m/z 120
(Fig. 1(A), path A of Scheme 3) dominates the spectrum, with
a minor loss of NH3 also being observed. The latter loss has
been shown experimentally and via DFT calculations to
occur via a neighbouring group process (Eqn. (2)) involving
attack of the aryl ring onto the alpha carbon to yield a
phenonium ion26,31–33 (Scheme 3, path B) rather than via a
1,2-hydride migration mechanism (Scheme 3, path C).34
½H2NCHðCH2PhÞCOOH þ Hꢀþ
! ½H2NCHðCH2PhÞꢀþ þ H2O þ CO
½H2NCHðCH2PhÞCOOH þ Hꢀþ
! ½CHðCH2PhÞCOOHꢀþ þ NH3
(1)
(2)
b2-Phenylalanine (Fig. 1(B))
In a similar fashion to a-phenylalanine26,31 (Fig. 1(A)), the
main product ion for b2-phenylalanine (Fig. 1(B)) also arises
via the loss of H2O and CO to form m/z 120 (Eqn. (3)). Other
½H2NCH2CHðPhÞCOOH þ Hꢀþ
! ½NH2CH2CHðPhÞꢀþ þ H2O þ CO
(3)
Scheme 4. Pathways for losses of H2O, H2O þ CO and NH3 from b2-phenylalanine.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

CID of protonated a-, b2- and b3-phenylalanines 1783
½H2NCH2CHðPhÞCOOH þ Hꢀþ
! ½NH2CH2CHðPhÞCOꢀþ þ H2O
ammonia via formation of a benzyl cation (Scheme 5, path B,
species aC2,b3B3). NH3 loss promoted via 1,2-hydride migration
(Scheme 5, path C) and carboxyl neighbouring group participa-
tion (Scheme 5, path D) are also mechanistic possibilities.
(4)
(5)
½H2NCH2CHðPhÞCOOH þ Hꢀþ
! ½CH2CHðPhÞCOOHꢀþ þ NH3
½H2NCHðPhÞCH2COOH þ Hꢀþ
! ½CHðPhÞCH2COOHꢀþ þ NH3
(6)
(7)
b3-Phenylalanine (Fig. 1(C))
½H2NCHðPhÞCH2COOH þ Hꢀþ
The low-energy CID spectrum of protonated b3-phenylalanine
(Fig. 1(C)) is quite different from those of a- and b2-
phenylalanine, which readily allows for isomer distinction.
The dominant fragmentation channel is loss of NH3 (to form m/z
! ½CHðPhÞCHCOꢀþ þ NH3 þ H2O
149, Eqn. (6)), while
a
minor combined loss of NH3
DFT calculations of possible structures and
relative energies of the product ion resulting
from NH3 loss (C9H9OR2 )
and H2O is observed to yield a product ion at m/z 131
(Eqn. (7)). It is interesting to note that the ion arising via the
combined loss of H2O and CO (Scheme 5, path A) is absent for
b3-phenylalanine, which is in stark contrast to the CID spectra of
the a- and b2-phenylalanine isomers where it is the dominant
reaction channel. Once again, there are three possible mechan-
isms for NH3 loss (Scheme 5, paths B–D). In this case, however,
the phenonium ion formed via a neighbouring group pathway
for a- and b2-phenylalanine isomers (Schemes 3 and 4, path B) is
not possible. Instead, the phenyl group can promote loss of
As there are a variety of potential structures of the C9H9O2þ
product ion that results from the loss of NH3 from the MS/MS
of protonated a-, b2- and b3-phenylalanine (Figs. 1(A)–1(C)), we
have carried out a theoretical survey into all the possible
structures that arise from the: (i) aryl-assisted neighbouring
group (path B); (ii) 1,2 hydride migration (path C); and
(iii) neighbouring group participation by the carboxyl group
(path D). In total, nine different isomers were considered
Scheme 5. Pathways for losses of H2O þ CO and NH3 from b3-phenylalanine.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

1784 A. K. Y. Lam and R. A. J. O’Hair
(Table 1), with the relative energies of each isomer expressed in
two ways: (i) relative to the most stable C9H9Oþ2 isomer,
aC2,b3B3 (shown in italics); (ii) with respect to the specific
isomer global minimum (shown in brackets).
loss of NH3 from each phenylalanine species, they do not
give any information about the relative barrier heights for
the transition states associated with the various competing
fragmentation reactions. Hence additional calculations were
carried out on each of the three isomeric amino acids to
determine the potential energy surfaces (PES) of each
fragmentation pathway in order to establish the mechanisms
by which they fragment under low-energy CID conditions.
Of particular interest was the competition between NH3 loss
and the combined loss of H2O and CO, which appear to be
sensitive to the structure of the amino acid (Fig. 1). The
possible pathways that we considered are those shown in
Schemes 3–5. On a final note, the PES presented below all
show the relative enthalpies for various processes at 0 K and
thus are uncorrected for temperature and entropy. These
relative entropies may underestimate processes where
entropy is important. For example, the combined loss
of H2O and CO is expected to be entropically favoured over
loss of water or ammonia alone.
The data shown in Table 1 demonstrates that 1,2-hydride
migration (path C) results in the most stable products for
both a-phenylalanine (aC2,b3B3, 26.8 kcal.molꢁ1) and b2-
phenylalanine (b2C2, 32.8 kcal.molꢁ1). In contrast, the
bicyclic ion from path B of these two amino acids, aB2,b2B2,
is considerably less stable (>9 kcal.molꢁ1). This difference
may be attributed to the presence of the strained three-
membered ring and loss of aromaticity from the phenyl
group. Structure aC2,b3B3 (path B) is the only stable product
ion arising from the loss of NH3 from b3-phenylalanine, with
the ions resulting from paths C and D isomerising into
aC2,b3B3 following optimisation. A key finding is that loss
of ammonia via path D is uncompetitive with the other
possible mechanistic paths for all Phe isomers. Thus, product
ions formed via path D are either the least stable C9H9O2þ
isomer (aD and b2D2) or appear to be unstable as they
isomerise into an ion formed via another pathway (b3D).
Potential mechanisms for the fragmentation
reactions of a-phenylalanine
DFT calculations of potential mechanisms for
the decomposition reactions of the amino acids
of a-, b2- and b3-phenylalanine
Although the results of the DFT calculations shown in Table 1
allow us to predict the most stable C9H9Oþ2 isomer following
Supplementary Fig. S1 (see Supporting Information) shows
the PES for the loss H2O þ CO (path A) and loss of NH3 via an
aryl-based neighbouring group attack (path B) and via
hydride transfer (path C). The PES for paths A–C have been
calculated in a previous study26 and have been recalculated
Table 1. Potential isomers of formula [C9,H9,O2]þ corresponding to the ion arising from loss of NH3 from the three isomeric
phenylalanine derivatives, a-phenylalanine, b2-phenylalanine and b3-phenylalanine. The relative energies listed below each
isomer are from DFT calculations carried out at the B3-LYP/6-31G(d,p) level of theory. The relative energies of each isomer
(in kcal.molꢁ1) are expressed in two ways: (i) relative to the most stable C9H9Oþ2 isomer (shown in italics); (ii) with respect to the
specific isomer global minima (shown in brackets)
Path
B
C
D
a-phenylalanine
aC2,b3B3
0.0
aB2,b2B2
12.4
aD
43.5
(39.2)
(26.8)
(70.3)
b2-phenylalanine
b2D2
b2C2
6.0
aB2,b2B2
15.8
22.9
(42.6)
(32.8)
(49.7)
b3-phenylalanine
aC2,b3B3
6.7
b3C
b3D
(Isomerises into
aC2,b3B3)
(Isomerises into
aC2,b3B3)
(33.5)
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

CID of protonated a-, b2- and b3-phenylalanines 1785
and included as a reference to the b-amino acids studied here.
With the exception of slight differences in energies for the final
products arising from paths B and C, the energies of the
structures are identical to those published previously.26 The
PES corresponding to path D, involving a carbonyl-based
neighbouring group attack, could not be found. However, as
the final endothermicity of the product ion, aD, is consider-
ably higher (Table 1) than both the energies of the transition
states of the competing pathways and those of their final
products, this pathway would not be expected to operate.
Supplementary Fig. S2. Briefly the proposed reaction occurs
via b2b, which is 8.6 kcal.molꢁ1 higher in energy than b2a
(Supplementary Fig. S2, see Supporting Information). The
dissociation occurs via a concerted mechanism in which the
ionising proton is transferred from the N-terminal nitrogen
onto the C-terminal hydroxyl group of the carboxylic acid,
poised to lose H2O and CO. This process occurs via transition
state TS(b2b-b2A3b) which has been calculated to have a barrier
of 49.9 kcal.molꢁ1 above b2a. This yields the ion-molecule
complex b2A3b, which upon separation results in b2A4.
The stepwise loss of H2O þ CO shown in Fig. 2 (Eqn. (3),
path A) from protonated b2-phenylalanine occurs via a
neighbouring group attack by the amino nitrogen to form the
ion-molecule complex b2A1. This pathway is predicted to
occur via conformer b2b which lies approximately
8.6 kcal.molꢁ1 above the local minimum, b2a. The neigh-
bouring group attack by the amino nitrogen onto the
carbonyl carbon occurs via transition state TS(b2b-b2A1)
with an energy of 27.6 kcal.molꢁ1 higher than b2b.
Separation of the resultant ion-molecule complex, b2A1,
yields the lactam ion b2A2, with an endothermicity of
40.2 kcal.molꢁ1. The subsequent loss of CO from the ion-
molecule complex b2A1 occurs via TS(b2A1-b2A3a), which,
following separation of the ion-molecule complex b2A3a,
yields the ion b2A4. The barrier for this pathway is
45.6 kcal.molꢁ1 relative to the lowest energy structure b2a.
As the barrier against this stepwise pathway is considerably
lower than that for the concerted mechanism H2O, CO loss
would be expected to occur via this pathway.
Loss of H2O and CO from protonated
a-phenylalanine
The loss of H2O and CO to yield the ion aA4 is predicted
to occur stepwise beginning with the loss of H2O via the
transition state TS(aAb-aA1), which has
a barrier of
39.1kcal.molꢁ1 above the local minima aa. Separation of the
resultant ion-molecule complex aA1 gives the ion aA2 which
is 49.2 kcal.molꢁ1 higher in energy than aa. The loss of CO
from the complex aA1 is proposed to occur via TS(aA1-aA3),
which lies 38.6 kcal.molꢁ1 higher in energy than aa. Separation
of the resultant ion-molecule complex aA3 yields the ion aA4
with an overall endothermicity of 25.3 kcal.molꢁ1
.
Loss of NH3 from protonated a-phenylalanine
As reported previously, the loss of NH3 from protonated a-
phenylalanine is predicted to occur via a neighbouring group
attack by the aryl group to form a phenonium cation (path B)
rather than by 1,2-hydride migration (path C).26,31 Briefly,
the thermodynamically less favoured hydride migration of
path C is predicted to operate via TS(aa-aC1), which lies
43.6kcal.molꢁ1 higher in energy than aa to yield the product
ion aC2,b3B3. The more favourable neighbouring group path
B is predicted to occur via TS(ac-aB1) which is 6.6 kcal.molꢁ1
lower in energy than the competing H2O loss transition state
TS(ab-aA1). Optimisation of the ion that arises from the
following separation of the ion-molecule complex aB1 leads to
the separated product ion aB2,b2B2 and NH3, with an overall
endothermicity of 39.2 kcal.molꢁ1. Although the barrier against
this pathway is smaller than the competing H2Oþ CO loss
from path A, as noted previously26 the minor abundance of the
NH3 loss ion compared with that of H2O þ CO is because
the H2O þ CO pathway is kinetically favoured.
Loss of NH3 from protonated b2-phenylalanine
Each of the three mechanistic pathways for the loss of NH3
from b2-phenylalanine (Scheme 4) was modelled using DFT
calculations and the pathways are shown in Supplementary
Fig. S3 (see Supporting Information). From this figure it can
be observed that path C which involves a 1,2-hydride
migration via the key transition state, TS(b2a-b2C1), has a
barrier height of 49.8 kcal.molꢁ1 and is the highest energy
pathway. Interestingly, the combined energy of the final
products from this pathway also has the lowest endother-
micity of all the products, with an endothermicity of
32.8 kcal.molꢁ1 (b2C2). NH3 loss via a neighbouring group
reaction via the carbonyl oxygen (path D) occurs first via a
gauche to anti conformational change, TS(b2a-b2c), which
has an activation energy of 9.0 kcal.molꢁ1, followed by the
energy required for the neighbouring group transition state,
TS(b2c-b2D1), of 41.8 kcal.molꢁ1. Finally, the energetically
favoured path B is predicted to arise via an aryl-assisted
neighbouring group reaction, resulting in the formation of
the phenonium ion aB2,b2B2, which is identical to the
product ion formed following loss of NH3 from a-
phenylalanine via the related phenyl neighbouring group
processes (Supplementary Fig. S1, see Supporting Infor-
mation). This process involves the transition state TS(b2b-
b2B1) with an activation energy of 37.7 kcal.molꢁ1. The low
abundance of the ion for this pathway in B is due to this
comparatively higher activation energy and the endother-
micity of aB2,b2B2 (42.6 kcal.molꢁ1 relative to b2a) than for
the competing H2O loss pathway (36.2 kcal.molꢁ1 relative to
b2a).
Potential mechanisms for the fragmentation
reactions of b2-phenylalanine
As shown in Fig. 1(B), the CID spectrum of protonated
b2-phenylalanine shows that it fragments mainly via the loss
of H2O and CO. The spectrum also shows the loss of NH3 and
the loss of H2O. The latter loss is not observed in the CID
spectra of the [M þ H]þ ions of the other isomers. The PES of
all these losses are discussed below. In addition to the three
possible paths for the loss of NH3, two potential mechanistic
pathways corresponding to the concerted and stepwise loss
of H2O þ CO were found.
Loss of H2O and CO from protonated
b2-phenylalanine
The PES for the concerted loss of H2O and CO via an aryl-
assisted neighbouring group reaction is shown in
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

1786 A. K. Y. Lam and R. A. J. O’Hair
Figure 2. B3-LYP/6-31 þ G(d,p) calculated reaction pathways and associated structures for key structures involved in
the fragmentation of protonated b2-phenylalanine: Pathway A: H2O and CO loss (Major); and pathway B: NH3 loss
(Minor). All italicised numbers (in kcal.molꢁ1) are relative to b2a.
Potential mechanisms for the fragmentation
reactions of b3-phenylalanine
resulting in the b3A1a complex (Fig. 3). This transition
state has
barrier 34.6 kcal.molꢁ1 higher than the
a
In contrast to a- and b2-phenylalanine, CID of protonated b3-
phenylalanine almost exclusively involves the loss of NH3,
with virtually no loss of H2O þ CO being observed. The PES
for this latter loss occurring via a concerted mechanism is
discussed below. In the case of the loss of NH3, although
there are three possible paths as shown in Scheme 5, only the
transition state corresponding to path B involving the aryl
neighbouring group could be found, which is consistent with
the final product instability of the ions from paths C and D
(Table 1).
preceding ion, b3b. This ion-molecule complex b3A1a
may separate to yield the ion b3A2 which has an
endothermicity of 40.7 kcal.molꢁ1. Further dissociation
of the conformer b3A1b via TS(b3A1b-b3A3) requires
49.6 kcal.molꢁ1 and results in an ion-molecule complex
b3A4 which has an endothermicity of 48.4 kcal.molꢁ1 and
consists of a protonated lactam surrounded by a CO
and a H2O molecule.
Loss of NH3 from protonated b3-phenylalanine
Formation of the benzyl cation (aC2,b3B3) via NH3
loss is predicted to first occur via a conformational
change from an anti to a gauche conformation, TS(b3a-
b3c), resulting in b3c (Fig. 3). This transition state has
a barrier 10.3 kcal.molꢁ1 higher than the most stable
Loss of H2O and CO from protonated
b3-phenylalanine
The neighbouring group reaction required for the loss
of H2O from b3-phenylalanine occurs via TS(b3b-b3A1a)
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

CID of protonated a-, b2- and b3-phenylalanines 1787
Figure 3. B3-LYP/6-31 þ G(d,p) calculated reaction pathways and associated structures for key structures involved
in the fragmentation of protonated b3-phenylalanine: Pathway A: H2O and CO loss (Minor); and pathway B: NH3 loss
(Major). All italicised numbers (in kcal.molꢁ1) are relative to b3a.
conformer of protonated b3-phenylalanine b3a. The
transition state for the loss of NH3, TS(b3d-b3B1), involves
an aryl-assisted neighbouring group attack, a process
that lies 26.5 kcal.molꢁ1 higher in energy than b3a.
The resultant ion-molecule complex separates to give
the benzyl cation aC2,b3B3 which has an endothermicity
of 33.5 kcal.molꢁ1. As the energy of TS(b3d-b3B1) is
substantially lower than the barriers against the com-
Key features and overall trends from the DFT
calculations
To aid in the comparison of the PES of the three phenylalanine
amino acids, the energies associated with the key barriers
against the four pathways are summarised in Table 2. For path
A, the loss of H2O þ CO, two columns corresponding to the
barriers and endothermicities associated with H2O and the
subsequent CO losses have been included.
peting H2O loss, TS(b3b-b3A1a), 26.5 vs. 41.1 kcal.molꢁ1
,
From the DFT-calculated PES values presented above, the
following observations can be made:
only NH3 loss is observed in C. In addition this ion,
aC2,b3B3, is the same ion as is formed following the
loss of NH3 from a-phenylalanine following 1,2-hydride
migration (Scheme 3).
(i) The endothermicity of the product ions does not give a
good indication of the preferred mechanistic pathway.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

1788 A. K. Y. Lam and R. A. J. O’Hair
Table 2. Summary of the barriers for the key transition states involved in paths A–D. The numbers in parentheses corresponds to
the final endothermicities for the products following separation of the ion-molecule complex. All numbers (in kcal.molꢁ1) are
relative to the most stable protonated conformer of the respective amino acid: aa, b2a and b3a
Path
A (H2O loss)
A (CO loss)
B
C
D
a-phenylalanine
b2-phenylalanine
b3-phenylalanine
39.1 (49.2)
36.2 (40.2)
41.1 (40.7)
38.6 (25.3)
45.6 (50.9)
49.6 (48.4)
32.5 (39.2)
37.7 (42.6)
26.5 (33.5)
43.6 (26.8)
49.8 (32.8)
(70.3)
41.8 (49.7)
(Isomerises into aC2,b3B3)
Examples of this are the product ions for a- and b2-
phenylalanine: although the final energies of the ions
that result from 1,2-hydride migration (path C) are con-
siderably lower than those obtained from the other
products, 1,2-hydride migration is unlikely to occur
due to the considerably higher activation energy than
for the other competing pathways;
having little effect on their a-amino acid counterparts. A key
example of this is the increased relative abundance of NH3
loss with the introduction of the methyl ester and the amide
nitrogen for b2-phenylalanine (Figs. 1(B), 4(C), and 4(D)).
Each of these amino acid derivatives is now discussed in
further detail.
(ii) The almost exclusive loss of NH3 from b3-phenylalanine
(Fig. 1(C)) can be rationalised by the low barrier of this
loss (26.5 kcal.molꢁ1) which is considerably lower than
those of the competing H2O (41.1 kcal.molꢁ1) and CO
losses (49.6 kcal.molꢁ1);
(iii) Comparison of paths B and D of b2-phenylalanine gives
us an indication of which neighbouring group is pre-
ferred. As the barrier for aryl neighbouring group
participation is considerably lower than the correspond-
ing carbonyl-based attack (37.7 vs. 41.8 kcal.molꢁ1), this
suggests that the phenyl group is the better neighbour-
ing group. As far as we are aware,36 this is the first
comparison of the intrinsic neighbouring group abilities
of Ph and COOH;
Gas-phase fragmentation of a-phenylalanine-OCH3
and a-phenylalanine-NHCH3
Collisional activation of the [M þ H]þ ion of a-phenyl-
alanine-OCH3 (Fig. 4(A)) results in a spectrum that is directly
related to that observed for a-phenylalanine (Fig. 1(A)) with
the a1 ion (m/z 120) being the major product ion, and a minor
loss of NH3 also being observed. Likewise, the introduction
of a methyl amide moiety (Fig. 4(B)) does not change the
types of product ions formed, with the losses of
NH2CH3 þ CO (to yield the ion at m/z 120) and NH3 (to
yield the ion at m/z 162) the only observable fragmentation
pathways.
Gas-phasefragmentationofb2-phenylalanine-OCH3
and b2-phenylalanine-NHCH3
(iv) For all ions that contain a benzyl cation, the bond length
between the benzylic carbon and the phenyl ring
decreases, indicating charge delocalisation onto the
ring. This is demonstrated by examining the ion that
arises following the loss of CO from the protonated
lactam b2A1 from b2-phenylalanine. The position of the
phenyl group allows for neighbouring group participa-
tion by the aryl group to form an a-aminomethyl-
substituted benzyl cation (b2A4). Due to the position
of the phenyl group for b3-phenylalanine, this pathway
is not accessible and the aziridine ion b3A4 is formed
instead.
The low-energy CID of the methyl ester of b2-phenylalanine
shown in Fig. 4(C) yields a more complex spectrum than
those observed for the parent amino acid (Fig. 1(B)).
Although the major losses of CH3OH and CO to form m/z
120 and CH3OH (m/z 148) are analogous to the H2O and CO,
and H2O losses in Fig. 1(B), the introduction of the methyl
ester also promotes the fragmentation of b2-phenylalanine
via NH3 loss (m/z 163). In addition, it introduces a variety of
minor abundance ions not observed before, including a
water adduct at m/z 166 and NHCH2 loss (to yield the ion
at m/z 151), probably occurring via a retro-Mannich
mechanism discussed previously.19,20,25 Collisional acti-
vation of protonated b2-phenylalanine-NHCH3 alters the
preferred fragmentation pathways dramatically, with the
dominant channel now being loss of NH3 at m/z 162, with
only a minor amount of NHCH2 loss (m/z 150) being
observed.
Gas-phase fragmentation of C-terminal methyl
ester and methyl amide derivatives of a-, b2-
and b3-phenylalanine
The C-terminally protected methyl ester and methyl amide
derivatives of each of the three phenylalanine isomers were
subjected to positive ion mode ESI-MS and their [M þ H]þ
ions were mass-selected and subjected to low-energy CID,
resulting in the spectra shown in Fig. 4. A brief comparison of
the spectra for the unmodified amino acids in Fig. 1 and their
corresponding methyl esters and methyl amides in Fig. 4
indicates that introduction of a C-terminal protecting group
can have a profound effect on the fragmentation chemistry
that is observed for b-phenylalanine amino acids, whilst
Gas-phasefragmentationofb3-phenylalanine-OCH3
and b3-phenylalanine-NHCH3
Figures 4(E) and 4(F) reveal that the CID spectra of b3-
phenylalanine-OCH3 and b3-phenylalanine-NHCH3 are
almost identical to that of the unprotected amino acid
(Fig. 1(C)). Thus, each spectrum is dominated by NH3 loss (to
give the ions at m/z 163 and 162, respectively). Alteration of
the protecting group appears to influence the minor peaks
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

CID of protonated a-, b2- and b3-phenylalanines 1789
Figure 4. LTQ CID MS/MS spectra of the [M þ H]þ ions of the terminal methyl ester derivatives of: (A) a-phenyl-
alanine; (C) b2-phenylalanine; and (E) b3-phenylalanine. MS/MS spectra of the [M þ H]þ ions of the terminal methyl
amide derivatives of: (B) a-phenylalanine; (D) b2-phenylalanine; and (F) b3-phenylalanine. An asterisk refers to the
mass-selected precursor ion. The z symbol indicates that the assignment of the molecular formula of the neutral(s)
lost has been confirmed via HRMS.
that are observed, with the amino acid (Fig. 1(C)) losing a
combination of NH3 and H2O (to give an ion at m/z 131),
while the methyl ester (Fig. 4(E)) loses CH3OH and HCN (to
give an ion at m/z 121). Finally, the methyl amide (Fig. 4(F))
has unique losses to form ions at m/z 105 and 106 that
correspond to the losses of C8H9 and C7H8N, respectively.
Finally, the minor ion at m/z 120 is proposed to arise via the
combined loss of NH2CH3 and CO.
we have shown that the location of the phenyl group in
isomeric phenylalanine residues can have a significant
effect on the fragmentation chemistry that is observed. By
moving the phenylalanine group from the a- to the b2- or
b3-position, different types of fragmentation pathways
become favoured. Thus, different types of small molecule
losses dominate the low-energy CID spectra for a-, b2-
and b3-phenylalanine, allowing for isomer distinction.
Further work needs to be carried out to investigate the
fragmentation chemistry of b-amino acids and their
peptides that contain a wide range of different R groups
at the b2 and b3 positions. It will be particularly interesting
to examine the behaviour of side chains that are known to
be reactive (e.g. basic, acidic and sulfhydryl) and play roles
in promoting the fragmentation reactions of a-amino acids.
CONCLUSIONS
Previous work has highlighted a number of the key
differences in the fragmentation reactions of a- versus b-
amino acids. For example, extension of the carbon backbone
of b-amino acids allows b1 ions to be observed. In this study
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm

1790 A. K. Y. Lam and R. A. J. O’Hair
Key questions that need to be addressed are: (i) how
do these reactive side chains promote fragmentation of
b-amino acids and their peptides and (ii) does placement
of these substituents on either b2 or b3 influence the
fragmentation chemistry?
17. Parker CD, Hercules DM. Anal. Chem. 1985; 57: 698.
18. Cheng MKJ, Chan OYO, Abirami S, Ma NL, Tsang CW.
Abstracts of Papers, 229th ACS National Meeting, San Diego,
CA, USA, March 13–17, 2005; COMP-183.
19. Lam AKY, Ramarathinam SH, Purcell AW, O’Hair RAJ.
J. Am. Soc. Mass Spectrom. 2008; 19: 1743.
20. Schreiber JV, Quadroni M, Seebach D. Chimia 1999; 53:
621.
SUPPORTING INFORMATION
21. Ramesh V, Ramesh M, Srinivas R, Sharma GVM, Manohar V.
Int. J. Mass Spectrom. 2009; 282: 64.
Additional supporting information may be found in the
online version of this article.
22. Bandala Y, Avina J, Gonzalez T, Rivero IA, Juaristi E. J. Phys.
Org. Chem. 2008; 21: 349.
23. Reddy PN, Srikanth R, Swamy NS, Srinivas R, Sharma GVM,
Nagendar P, Krishna PR. J. Mass Spectrom. 2005; 40: 1429.
24. Reddy PN, Srinivas R, Kumar MR, Sharma GVM, Jadhav VB.
J. Am. Soc. Mass Spectrom. 2007; 18: 651.
Acknowledgements
We thank the ARC for financial support via the ARC Centre
of Excellence in Free Radical Chemistry and Biotechnology.
The authors gratefully acknowledge the generous allocation
of computing time from both the Victorian Partnership for
Advanced Computing and the National Computational
Infrastructure National Facility. The authors would also like
to thank the ARC and VICS for funding of the LTQ FT hybrid
mass spectrometer instrument. AKYL acknowledges the
award of a Melbourne Research Scholarship from The
University of Melbourne.
25. Ramesh V, Srinivas R, Sharma GVM, Jayaprakash P, Kunwar
AC. J. Mass Spectrom. 2008; 43: 1201.
26. Lioe H, O’Hair RAJ. Org. Biomol. Chem. 2005; 3: 3618.
27. Feenstra RW, Stokkingreef EHM, Reichwein AM,
Lousberg WBH, Ottenheijm HCJ, Kamphuis J, Boesten
WHJ, Schoemaker HE, Meijer EM. Tetrahedron 1990; 46:
1745.
28. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
Cheeseman JR, Montgomery JA Jr, Vreven T, Kudin KN,
Burant JC, Millam JM, Iyengar SS, Tomasi J, Barone V,
Mennucci B, Cossi M, Scalmani G, Rega N, Petersson GA,
Nakatsuji H, Hada M, Ehara M, Toyota K, Fukuda R,
Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O,
Nakai H, Klene M, Li X, Knox JE, Hratchian HP, Cross JB,
Bakken V, Adamo C, Jaramillo J, Gomperts R, Stratmann RE,
Yazyev O, Austin AJ, Cammi R, Pomelli C, Ochterski JW,
Ayala PY, Morokuma K, Voth GA, Salvador P,
Dannenberg JJ, Zakrzewski VG, Dapprich S, Daniels AD,
Strain MC, Farkas O, Malick DK, Rabuck AD,
Raghavachari K, Foresman JB, Ortiz JV, Cui Q, Baboul
AG, Clifford S, Cioslowski J, Stefanov BB, Liu G,
Liashenko A, Piskorz P, Komaromi I, Martin RL, Fox DJ,
Keith T, Al-Laham MA, Peng CY, Nanayakkara A,
Challacombe M, Gill PMW, Johnson B, Chen W, Wong
MW, Gonzalez C, Pople JA. Gaussian 03, Revision E.01,
Gaussian Inc.: Wallingford CT, 2004.
REFERENCES
1. Burlet O, Orkiszewski RS, Ballard KD, Gaskell SJ. Rapid
Commun. Mass Spectrom. 1992; 6: 658.
2. Jones JL, Dongre AR, Somogyi A, Wysocki VH. J. Am. Chem.
Soc. 1994; 116: 8368.
3. Dongre AR, Jones JL, Somogyi A, Wysocki VH. J. Am. Chem.
Soc. 1996; 118: 8365.
4. Tsaprailis G, Nair H, Somogyi A, Wysocki VH, Zhong W,
Futrell JH, Summerfield SG, Gaskell SJ. J. Am. Chem. Soc.
1999; 121: 5142.
5. Wysocki VH, Tsaprailis G, Smith LL, Breci LA. J. Mass
Spectrom. 2000; 35: 1399.
6. O’Hair RAJ. J. Mass Spectrom. 2000; 35: 1377.
7. Paizs B, Suhai S. Mass Spectrom. Rev. 2005; 24: 508.
8. Fink K, Henderson RB, Fink RM. J. Biol. Chem. 1952; 197: 441.
9. Roberts E, Bregoff HM. J. Biol. Chem. 1953; 201: 393.
10. Canellakis ES. J. Biol. Chem. 1956; 221: 315.
11. Pollitt RJ, Green A, Smith R. J. Inherited Metab. Dis. 1985;
8: 75.
29. Becke AD. J. Chem. Phys. 1993; 98: 5648.
30. Lee C, Yang W, Parr RG. Phys. Rev. B: Condens. Matter 1988;
37: 785.
31. Dookeran NN, Yalcin T, Harrison AG. J. Mass Spectrom. 1996;
31: 500.
32. Rogalewicz F, Hoppilliard Y, Ohanessian G. Int. J. Mass
Spectrom. 2000; 195/196: 565.
12. Lelais G, Seebach D. Biopolymers 2004; 76: 206.
13. Fleming PE, Mocek U, Floss HG. J. Am. Chem. Soc. 1993;
115: 805.
14. Aguilar M-I, Purcell AW, Devi R, Lew R, Rossjohn J, Smith
AI, Perlmutter P. Org. Biomol. Chem. 2007; 5: 2884.
15. Tsang CW, Harrison AG. J. Am. Chem. Soc. 1976; 98: 130.
16. Wysocki VH, Burinsky DJ, Cooks RG. J. Org. Chem. 1985; 50:
1287.
33. El Aribi, H, Orlova G, Hopkinson AC, Siu KWM. J. Phys.
Chem. A 2004; 108: 3844.
34. Shoeib T, Cunje A, Hopkinson AC, Siu KWM. J. Am. Soc.
Mass Spectrom. 2002; 13: 408.
35. O’Hair RAJ, Reid GE. Rapid Commun. Mass Spectrom. 2000;
14: 1220.
36. Capon B, McManus SP. Neighboring Group Participation,
vol. 1. Springer-Verlag: New York, 1976.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1779–1790
DOI: 10.1002/rcm