Low mass MS/MS fragments of protonated amino acids used for distinction of their 13C- isotopomers in metabolic studies

Article abstract of DOI:10.1007/s13361-012-0574-9

Glu, Gln, Pro, and Ala are the main amino acids involved in ammonia detoxification in mosquitoes. In order to develop a tandem mass spectrometry method (MS²) to monitor each carbon of the above isotopically-labeled ¹³C-amino acids for metabolic studies, the compositions and origins of atoms in fragments of the protonated amino acid should be first elucidated. Thus, various electrospray (ESI)-based MS² tools were employed to study the fragmentation of these unlabeled and isotopically-labeled amino acids and better understand their dissociation pathways. A broad range of fragments, including previously-undescribed low m/z fragments was revealed. The formulae of the fragments (from m/z 130 down to m/z 27) were confirmed by their accurate masses. The structures and conformations of the larger fragments of Glu were also explored by ion mobility mass spectrometry (IM-MS) and gas-phase hydrogen/deuterium exchange (HDX) experiments. It was found that some low m/z fragments (m/z 27-30) are common to Glu, Gln, Pro, and Ala. The origins of carbons in these small fragments are discussed and additional collision induced dissociation (CID) MS² fragmentation pathways are proposed for them. It was also found that small fragments (≤m/z 84) of protonated, methylated Glu, and methylated Gln are the same as those of the underivatized Glu and Gln. Taken together, the new approach of utilizing low m/z fragments can be applied to distinguish, identify, and quantify ¹³C-amino acids labeled at various positions, either in the backbone or side chain. [Figure not available: see fulltext.]

Full text of DOI:10.1007/s13361-012-0574-9

B American Society for Mass Spectrometry, 2013
J. Am. Soc. Mass Spectrom. (2013) 24:622Y631
DOI: 10.1007/s13361-012-0574-9
RESEARCH ARTICLE
Low Mass MS/MS Fragments of Protonated Amino Acids
13
Used for Distinction of Their C- Isotopomers in Metabolic
Studies
Xin Ma,1,2 Shai Dagan,1,3 Árpád Somogyi,1 Vicki H. Wysocki,1,2 Patricia Y. Scaraffia1
1Department of Chemistry and Biochemistry, The University of Arizona, Tucson, AZ 85721, USA
2Department of Chemistry and Biochemistry, The Ohio State University, Columbus, OH 43210, USA
3Israel Institute for Biological Research, Ness Ziona, 74100, Israel
Abstract. Glu, Gln, Pro, and Ala are the main amino acids involved in ammonia
detoxification in mosquitoes. In order to develop a tandem mass spectrometry
method (MS2) to monitor each carbon of the above isotopically-labeled 13C-amino
acids for metabolic studies, the compositions and origins of atoms in fragments of
the protonated amino acid should be first elucidated. Thus, various electrospray
(ESI)-based MS2 tools were employed to study the fragmentation of these
unlabeled and isotopically-labeled amino acids and better understand their
dissociation pathways. A broad range of fragments, including previously-
undescribed low m/z fragments was revealed. The formulae of the fragments
(from m/z 130 down to m/z 27) were confirmed by their accurate masses. The
structures and conformations of the larger fragments of Glu were also explored by ion mobility mass
spectrometry (IM-MS) and gas-phase hydrogen/deuterium exchange (HDX) experiments. It was found that
some low m/z fragments (m/z 27–30) are common to Glu, Gln, Pro, and Ala. The origins of carbons in these
small fragments are discussed and additional collision induced dissociation (CID) MS2 fragmentation
pathways are proposed for them. It was also found that small fragments (≤m/z 84) of protonated, methylated
Glu, and methylated Gln are the same as those of the underivatized Glu and Gln. Taken together, the new
approach of utilizing low m/z fragments can be applied to distinguish, identify, and quantify 13C-amino acids
labeled at various positions, either in the backbone or side chain.
Key words: Low m/z fragments, 13C-amino acids, ESI, CID MS2, FTICR, IM-MS, Q-TOF
Received: 10 October 2012/Revised: 10 December 2012/Accepted: 13 December 2012/Published online: 27 February 2013
direct infusion electrospray (ESI) MS2 [6–20]. Jeffrey et al.
Introduction
[16] compared the performance of NMR, GC-MS (Electron
Ionization-EI) and GC-MS2 in the analysis of isotopically-
table isotopes such as 15N and 13C are widely used as
labeled Glu in the citric acid cycle. The authors concluded
that MS2 and NMR are comparable, whereas the full-scan
tracers to monitor and quantify metabolic pathways in
S
humans as well as other organisms. Usually, isotopically-
labeled compounds are introduced into a biological system
and the fluxes of these stable isotopes can then be quantified
to provide accurate information about different metabolic
processes [1–5]. The common techniques to monitor and
quantify these labeled compounds in complex metabolic
networks are nuclear magnetic resonance (NMR) spectros-
copy, gas chromatography mass spectrometry (GC-MS), and
EI-MS is the least accurate technique [16]. Accurate mass
measurements by Fourier transform ion cyclotron resonance
(FTICR) MS also show potential to be a useful tool to
monitor 13C distribution in amino acids, based on the exact
mass shifts of 13C labeled precursor ions and fragments [10,
11, 20].
In our laboratory, we have applied ESI-MS2 for studying
the ammonia metabolism in Aedes aegypti mosquitoes, the
main vectors of dengue and yellow fever. By feeding A.
aegypti females with isotopically-labeled 15N-compounds
(such as 15NH4Cl), we discovered that mosquitoes have an
extraordinary biochemical machine to efficiently detoxify
ammonia. Indeed, they synthesize specific amino acids and
excrete several nitrogen wastes through multiple metabolic
Electronic supplementary material The online version of this article
(doi:10.1007/s13361-012-0574-9) contains supplementary material, which
is available to authorized users.
Correspondence to: Patricia Y. Scaraffia; e-mail: scaraffi@email.arizona.edu

X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
623
pathways [6–9]. By using this strategy, female mosquitoes
are able to survive toxic ammonia concentrations released
during the digestion of a blood meal. To better understand
this complex metabolic process and to identify possible
targets that can be utilized for controlling mosquito
populations, we are now interested in investigating the
carbon metabolism of Glu, Gln, Pro, and Ala, the main
amino acids involved in ammonia detoxification in A.
aegypti. In order to monitor the carbons of isotopically-
labeled amino acids by ESI-MS2, there must be fragments
from one or more fragmentation events that can be used to
separately identify each carbon. Towards this goal, we first
studied the compositions and origins of the atoms in
fragments of the protonated amino acids. It has been
reported that the carboxylic acid groups dissociate from
protonated amino acids at low collision energies (CE) in
collision induced dissociation (CID) [21]. For example,
protonated Glu (m/z 148) is reported to lose two CH2O2
units to produce the m/z 56 fragment [21–24], and this
fragment is used for quantification of [1-13C]- Glu and [5-
13C] – Glu [14, 15]. The fragmentation of protonated Gln
(m/z 147) is reported to be similar to that of protonated Glu
[21–24]. It was also reported that protonated Pro (m/z 116)
and Ala (m/z 90) lose CH2O2 to form m/z 70 (Pro fragment)
and 44 (Ala fragment) peaks [21]. However, these relatively
large fragments can only be used to distinguish, identify and
quantify the carbons of the carboxylic acid groups. In order
to monitor each carbon in Glu, as well as the other amino
acids, understanding of the compositions and origins of the
carbons in smaller fragments is required.
Here, we further explore and extend the range of
fragments of protonated Glu (m/z 148), Gln (m/z 147), Pro
(m/z 116), and Ala (m/z 90) all the way down to m/z 27. Glu
is studied in detail and presented as a representative example
for the other amino acids. FTICR and quadrupole time of
flight (Q-TOF) mass spectrometry methods were employed
to measure the accurate masses of the fragments and reveal
their chemical formulae. Also, the origins of the carbons in
the fragments were elucidated based on the mass shifts of the
fragments in the isotopically-labeled amino acids. Gas-phase
hydrogen/deuterium exchange (HDX) and ion mobility mass
spectrometry (IM-MS) were also used to explore the
possible configurations of some of the fragments. Based on
these results, a CID MS2 pathway map from the precursor,
all the way to the m/z 27 fragment of protonated Glu, is
proposed. Some of the low m/z fragments (m/z 27–30),
found in protonated Glu, are also observed in the protonated
Gln, Pro and Ala. The origins of the carbons of these small
fragments are discussed.
purchased from Sigma-Aldrich (St. Louis, MO, USA). [5-
13C] – Glu was obtained from Santa Cruz Biotechnology,
Inc. (San Diego, CA, USA). [4-13C]-Gln and [4-13C] – Glu
were generously provided by Dr. Kanamatsu (Soka University,
Tokyo, Japan) and Dr. Takatori (Meiji Pharmaceutical Univer-
sity, Tokyo, Japan) respectively. [15N] – Glu, [1,2-13C] – Glu,
[13C515N] – Glu, [2-15N] – Gln, [5-15N] – Gln, [2,5-15N2] – Gln,
[1,2-13C] – Gln, [3-13C] – Gln, [13C515N2] – Gln, [15N] – Pro,
[1-13C] – Pro, [13C515N] – Pro, [15N] – Ala, [1-13C] – Ala, [2-
13C] – Ala, and [3-13C] – Ala were purchased from Cambridge
Isotope Laboratories (Andover, MA, USA). LC/MS grade
water was purchased from Mallinckrodt Baker Inc. (Phillips-
burg, NJ, USA). HPLC grade methanol and acetonitrile were
purchased from EMD Chemicals Inc. (Gibbstown, NJ, USA).
Amino Acid Derivatization
For some experiments, the amino acids were methylated by
a method described previously with some modifications
[25]. Briefly, 50 nmol of each amino acid was dissolved in
3 mL acetone/methanol (70/30 vol/vol). Ten μL of 2 M
trimethylsilyldiazomethane was added into the solution
dropwise and stirred for 4 h. The solvents were removed
under a stream of N2 gas.
Low Resolution Tandem Mass Spectrometry
Low resolution MS2 was performed using ESI/triple quad-
rupole (QqQ) mass spectrometers (AB/SCIEX 4000 QTRAP
and AB/SCIEX 3000 QqQ; Foster City, CA, USA) in the
positive ion mode. The spray solvent utilized was a
methanol-water-formic acid mixture (70:30:0.1 vol/vol/vol).
Several unlabeled and isotopically-labeled amino acid
solutions at 0.5∼10 μM were introduced into the ESI source
by infusion at a flow rate of 7 μL/min. The ESI capillary
voltage utilized was 5 kV and the source temperature was
250 °C. Nitrogen served as the collision gas (8 mTorr). A
laboratory collision energy of 15–55 eV was applied to
dissociate the ions to various degrees. To perform MS3
experiments, the declustering potential was set to 80 V to in-
source dissociate the precursor, while the precursor fragment
was selected in the first quadrupole and CID was used to
further dissociate it.
Fourier Transform Ion Cyclotron Resonance
(FTICR) Mass Spectrometry
High resolution MS2 and gas-phase HDX experiments were
performed in an ESI/FTICR (Bruker APEX Qh 9.4T,
Billerica, MA, USA) mass spectrometer in the positive ion
mode. The FTICR instrument was externally calibrated for
the low m/z region and the measured masses are reported
with a four decimal accuracy and compared with theoretical
(calculated) values using the Bruker data processing soft-
ware. The spray solvent utilized was a water-acetonitrile-
formic acid mixture (50:50:0.1 vol/vol/vol). Unlabeled and
Experimental
Chemicals and Reagents
Unlabeled amino acids, Glu-dimethyl-ester, trimethylsilyl-
diazomethane, formic acid, ND3, and [3-13C] – Glu were

624
X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
isotopically-labeled amino acid solutions (10 μM) were
infused into the ESI source with a flow rate of 2.5 μL/min.
The capillary voltage was 4.8 kV, the spray shield was
3.3 kV, and the source temperature was 200 °C. Sustained
off-resonance irradiation (SORI)-CID was used to dissociate
the ions. The argon pressure measured at the FTICR cell was
∼3×10–8mbar. The SORI power was typically 1 %–2 %
with a pulse duration of 100 ms and frequency offset of
500 Hz. Gas-phase HDX experiments were performed with
ND3. The ND3 pressure was ∼0.8 mbar and the reaction time
in the FTICR cell varied between 0.1 and 10 s.
and [5-13C] – Glu + H+ (Figure 1g, 1i, and 1k, respectively).
The m/z 103 peaks in these spectra show that C3, C4, and
C5 carbons remain in this fragment, with C1 lost. In the
[13C515N] – Glu (Figure 1m), the neutral loss is 47 Da,
confirming that only one carbon is in the neutral lost.
One of the major fragments observed at 15 eV collision
energy is the m/z 84 peak (Figure 1a), which is the result of
the 64 Da neutral loss (two water molecules and one carbon
monoxide) from Glu + H+ [10, 11, 21–24]. In the spectrum
of [15N] – Glu + H+ (Figure 1c), the m/z 84 peak shifts up by
one Da, indicating that the nitrogen is part of this fragment.
The carbon monoxide loss to form the m/z 84 peak also
contains C1. As a result, in the spectrum of [1,2-13C] –
Glu + H+ (Figure 1e), the m/z 84 peak shifts up by only one
Da, which is also true for the spectra of [3-13C] – Glu + H+,
[4-13C] – Glu + H+ and [5-13C] – Glu + H+ (Figure 1g, 1i,
and 1k, respectively). In the spectrum of [13C515N] – Glu +
H+ (Figure 1m), the m/z 84 peak, containing one nitrogen
and four carbons, shifts up by five Da, in agreement with
previous studies that show that C1 dissociates from
protonated Glu first, followed by dissociation of C5 [10,
11, 21–24].
ESI fragments smaller than m/z 56 were not reported
previously. The compositions of smaller fragments may
provide specific and useful information to separately
monitor each 13C-carbon in Glu and other amino acids. To
further dissociate and explore smaller products of Glu + H+,
a higher CID collision energy (55 eV, red spectra in
Figure 1) was applied. High energy MS2 experiments
provided low m/z fragments with higher intensity than
various MS3 experiments. At this higher CE, the dominant
peak is m/z 56, and the neutral loss from the precursor is
92 Da, which results from two water and two carbon
monoxide molecules. The two carbon monoxide molecules
contain the C1 and C5 carbons of the carboxyl groups,
which are easier to dissociate from Glu + H+ [10, 11]. The
m/z 56 peak shifts up by one Da in the MS2 spectrum of
[15N] – Glu + H+ (Figure 1c), indicating that the nitrogen is
in this fragment. In the spectra of [1,2-13C] – Glu + H+,
[3-13C] – Glu + H+, [4-13C] – Glu + H+ (Figure 1e, 1 g, and
1i, respectively), this peak also shifts up by one Da,
suggesting that C1 is involved in the neutral loss. The other
carbon monoxide contains C5 as the m/z 56 peak does not
shift in the spectrum of [5-13C] – Glu + H+ (Figure 1k).
A m/z 41 peak was also observed at 55 eV CID
(Figure 1a). Our data suggest that this peak results from
CONH loss (43 Da loss) from the m/z 84 peak, and not from
the m/z 56 fragment, as is discussed below. The MS2
spectrum of [15N] – Glu + H+ (Figure 1c) indicates that the
nitrogen is in the neutral loss group (CO15NH) because the
m/z 41 fragment does not shift. In correlation with these
data, the spectrum of [13C515N] – Glu + H+ (Figure 1m) gives a
neutral loss from m/z 89 of 45 Da, corresponding to 13CO15NH,
whereas in the [1,2-13C] – Glu + H+, [3-13C] – Glu + H+ and
[4-13C] – Glu + H+ (Figure 1e, 1 g, and 1i, respectively) this
neutral loss is 43 Da (CONH). In the [5-13C] – Glu + H+
Electrospray Quadrupole-Time of Flight (Q-TOF)
Ion Mobility Mass Spectrometry (ESI-IM-MS)
IM-MS and accurate mass measurements were performed
using an ESI/quadrupole/IM/TOF mass spectrometer
(Waters, SYNAPT G2, Manchester, UK) in the positive
ion mode. The calibration was performed for the low mass
range. The concentration of amino acid solutions was
0.1 mM and the sample flow rate was set to be 5 μL/min.
The ESI capillary voltage was 3 kV and the source
temperature was 260 °C. Argon (3×10–2mbar) was used as
the collision gas and the CE was 10–15 eV. The drift times
of amino acid fragments were measured in the “ion
mobility” mode. The pressure of nitrogen in the ion mobility
chamber was ∼2.2 mbar. The ion mobility spectrometry
(IMS) wave height and velocity were optimized at 16.0 V
and 1500 ms–1. Accurate mass measurements were per-
formed in the TOF “resolution” mode.
Results and Discussion
Low and High Energy Collision Induced
Dissociation (CID) of Protonated Glu
and Isotopically-Labeled Glu
In order to develop a method to monitor each possible 13C
carbon in Glu, several different Glu isotopomers were used.
Figure 1 shows the QqQ MS2 spectra of Glu + H+ (m/z 148
precursor), [15N] – Glu + H+ (m/z 149), [1,2-13C] – Glu + H+
(m/z 150), [3-13C] – Glu + H+ (m/z 149), [4-13C] – Glu + H+
(m/z 149), [5-13C] – Glu + H+ (m/z 149), and [13C515N] –
Glu + H+ (m/z 154) (refer to Scheme 1 for numbering of
atoms). The MS2 spectra exhibit m/z 130, 102, and 84 peaks
or their analogs in the labeled compounds at 15 eV collision
energy (blue spectra in Figure 1). This is in agreement with
previous studies performed with unlabeled Glu [21–24]. The
m/z 130 peak results from a loss of neutral water, whereas
m/z 102 peak results from a 46 Da neutral loss (CH2O2) [22–
24]. In the MS2 spectrum of [15N] – Glu + H+ (Figure 1c),
the neutral loss is also 46 Da indicating that the nitrogen is
not lost. Because carbon number 1 (C1) is lost, the m/z 102
peak shifts up by only one Da in the [1,2-13C] – Glu + H+
spectrum (Figure 1e). This conclusion is further confirmed
by the spectra of [3-13C] – Glu + H+, [4-13C] – Glu + H+,

X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
625
Figure 1. MS2 of protonated Glu and isotopically-labeled Glu performed on a QqQ instrument in the product ion scan mode.
(a) Glu + H+ (m/z 148); (c) [15N] – Glu + H+ (m/z 149); (e) [1,2-13C] – Glu + H+ (m/z 150); (g) [3-13C] – Glu + H+ (m/z 149); (i) [4-13C]
– Glu + H+ (m/z 149); (k) [5-13C] – Glu + H+ (m/z 149); (m) [13C515N] – Glu H+ (m/z 154). On the left (a, c, e, g, i, k, and m),
superimposed spectra are shown, obtained at a CE of 15 eV (blue) and 55 eV (red). On the right (b, d, f, h, j, l, and n), blow-ups
of the low m/z region (m/z 26–35, CE 55 eV) are shown by magnifying 20 to 100 times. The spectra are normalized to the highest
peak in each m/z range

626
X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
Scheme 1. Proposed CID MS2 fragmentation pathways, ion structures* and origins of carbons of Glu. The accurate masses of
the large fragments are measured by FTICR and those of the small fragments (≤m/z 30) are measured by Q-TOF. The thinner
arrow from m/z 130 to m/z 102 represents a minor pathway. The fragments in the circle on the right are from the m/z 56 and m/z
84 fragments. The m/z 41 and m/z 39 fragments may also form the m/z 27 fragment, although via high energy reactions. The
fragments in each red box make a pair of charge-competing products. Structures of fragments ≥m/z 84 were proposed
previously [10, 11, 22–24]. (*For simplicity not all possible isomeric ion structures are indicated)
(Figure 1k), the neutral loss is 44 Da (13CONH) showing that
the carbon in the leaving group is C5.
Glu and isotopically-labeled Glu (Figure 2 and Table 1b).
The spectra of Glu + H+ (Figure 2a) and [5-13C] – Glu + H+
(Figure 2e) in this m/z range are exactly the same, indicating
the fragments in this m/z range do not contain C5. In the
spectrum of Glu + H+, the m/z 28.019 peak is dominant
(Figure 2a) and its accurate mass corresponds to CH2N. In
the spectrum of [1,2-13C] – Glu + H+ (Figure 2b), the peak
shifts to m/z 29.021 (13CH2N), indicating this fragment
contains either C1 or C2. Because C1 is in the carboxylic
acid group, it dissociated from Glu at a low CE [10, 11, 22–
24]. The carbon in the m/z 28.019 peak is therefore most
likely C2. The m/z 30.033 (CH4N) fragment also contains
nitrogen and carbon. However, only about half of the m/z
30.033 peak population shifts to m/z 31.037 (13CH4N) in the
spectrum of [1,2-13C] – Glu + H+ (Figure 2b). The other half
shifted to m/z 31.037 (13CH4N) in the spectrum of [4-13C] –
Glu + H+ (Figure 2d), but there was no shift in the spectrum
of [3-13C] – Glu + H+ (Figure 2c), indicating that the CH4N
peak at m/z 30.033 peak is a mixture implies either C2 or
C4. This difference between the m/z 28.019 and m/z 30.033
labeled fragments indicates that there are separate pathways
to form each of these two fragments. Thus, there is most
likely more than one pathway involving different carbons to
form CH4N (Scheme 1). One of these pathways presumably
involves a rearrangement to combine C4 and the nitrogen.
Another major peak in the m/z 27–32 range is at m/z
29.040, which corresponds to C2H5. This peak does not shift
in the spectrum of [1,2-13C] – Glu + H+ (Figure 2b),
indicating that this fragment contains neither C1 nor C2. As
discussed previously, the low m/z fragments do not contain
C5, thus it is likely that the carbons in this m/z 29 fragment
are C3 and C4. The spectra of [3-13C] – Glu + H+ and
[4-13C] – Glu + H+ (Figure 2c, d) confirm this hypothesis.
In the m/z 27–32 range, several small peaks were
observed in the MS2 spectrum of Glu at a collision energy
of 55 eV (Figure 1b). The m/z 28 and 30 fragments contain
nitrogen as these two peaks shift up by one Da in the
spectrum of [15N] – Glu + H+ (Figure 1d). In contrast, the
m/z 29 peak does not shift in the spectrum of [15N] – Glu +
H+ (Figure 1d), indicating that this fragment does not
contain nitrogen. The m/z 28 and 30 fragments also contain
C2 because they shift up by one Da in the spectrum of [1,2-
13C] – Glu + H+ (Figure 1f). It is known from the literature
that C1 is the first to leave [21–24], and therefore m/z 28
must contain C2 and not C1. In the spectrum of [3-13C] –
Glu + H+, [4-13C] – Glu + H+ and [5-13C] – Glu + H+
(Figure 1h, j, and l, respectively), the m/z 28 and m/z 30
peaks do not shift, confirming that these fragments contain
C2. Based on the information above, the formula of the m/z
28 peak corresponds to CH2N whereas the formula of the
m/z 30 peak corresponds to CH4N. In [13C515N] – Glu + H+
(Figure 1n), all three peaks, including m/z 29, shift up by
two Da, suggesting that the m/z 29 peak contains two
carbons. Therefore, the proposed formula for this peak is
C2H5. Because shifted peaks may overlap with other peaks,
high resolution, accurate mass measurements of these frag-
ments are required to further explore and discuss the origins
of the carbons.
Accurate masses of the fragments, from m/z 130 down to
m/z 30, were measured by FTICR-MS and their formulae
were confirmed (Table 1a). Because of the difficulty in
detecting fragments ≤m/z 30 in FTICR, the Q-TOF was
employed to explore the composition and origin of the
carbons of the smaller fragments (m/z 27–32) of protonated

X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
627
Table 1. Accurate masses of Glu + H+ fragments measured by (a) FTICR; (b) Q-TOF
(a)
Proposed formula
C5H8NO3
C4H8NO2
C4H6NO
C3H6N
Theoretical mass (Da)
130.0499
102.0550
84.0444
Experimental mass (Da)
130.0507
Error (ppm)
+6.1
102.0555
84.0447
56.0495
+4.8
+3.5
+0.0
56.0495
C3H5
41.0386
41.0385
–2.4
C3H3
39.0229
39.0228
–2.5
CH4N
30.0338
30.0337
–3.3
(b)
Proposed formula
CH4N
C2H5
Theoretical mass (Da)
30.034
Experimental mass (Da)
30.033
Δ (Da)
–0.001
+0.001
+0.001
+0.000
+0.000
+0.001
+0.000
+0.000
29.039
28.018
27.023
31.037
30.042
29.021
28.026
29.040
28.019
27.023
31.037
30.043
29.021
28.026
CH2N
C2H3
13CH4N
13CCH5
13CH2N
13CCH3
The m/z 29.040 (C2H5) peak completely shifts to m/z 30.043
(13CCH5) in both of these spectra. This C2H5 and the CH2N
fragments could both originate from the m/z 56 (C3H6N)
fragment.
Unlike the C2H5 fragment, about half of the m/z 27.023
(C2H3) peak population shifts to 28.026 (13CCH3) in the
spectra of [1,2-13C] – Glu + H+ and [4-13C] – Glu + H+
(Figure 2b and d). This m/z 27.023 (C2H3) peak shifts
completely to 28.026 (13CCH3) in the spectrum of [3-13C] –
Glu + H+ (Figure 2c). Therefore, the carbons in this
fragment mixture are C2, C3, and C3, C4. The 3,4-C2H3
and the 2-CH4N fragments could originate from the m/z 56
(C3H6N) or m/z 84 (C4H6NO) fragments, whereas the 2,3-
C2H3 and the 4-CH4N fragments could be another pair of
competing products through a rearrangement pathway. The
compositions of these small fragments provide information
about the origins of the carbons, and pathways to form them,
as discussed below.
Figure 2. Low m/z fragments in MS2 of Glu and isotopically-labeled Glu obtained on the Q-TOF instrument at 15 eV CE. (a)
Glu + H+ (m/z 148); (b) [1,2-13C] – Glu + H+ (m/z 150); (c) [3-13C] – Glu + H+ (m/z 149); (d) [4-13C] – Glu + H+ (m/z 149); (e) [5-13C] –
Glu + H+ (m/z 149). The arrows show shifts of the peaks with the labeled compounds (blue for [1,2-13C] – Glu + H+, red for
[3-13C] – Glu + H+ and grey for [4-13C] – Glu + H+). Dash-line arrows indicate that only part of the peak population shifts.
Actually measured m/z values are indicated

628
X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
further dissociates to m/z 56 and 41 fragments at ∼30 eV and to
smaller fragments at ∼40 eV. The pathways leading to large
fragments, in particular from m/z 148 to 102 and from m/z
130 to 102 were previously reported and discussed by
Harrison [22, 23].
The m/z 102 fragment has two proposed structures [22].
One is linear and the other is a ring (Scheme 1). The MS2
spectrum of protonated Glu dimethyl ester provides a clue
about the structure of the Glu m/z 102 fragment (Figure S-1).
A m/z 116 peak which is the methylated analog of the m/z
102 fragment was observed. Because the carboxylic acid
group is still methylated, no ring is likely to be formed.
Therefore, the m/z 116 is linear and, thus, a linear structure
may prevail also with the Glu m/z 102 peak.
To further explore the structures of the larger fragments
of Glu + H+, gas-phase ammonia HDX and IM-MS were
employed. The gas-phase HDX experiment shows that the
m/z 84 fragment has at least three exchangeable hydrogens
(Figure 4a, b), which does not fit any of the structures
previously proposed [10, 11, 22–24]. This result could be
explained with some linear, rearranged structures for the m/z
84 fragment. The IM drift times of the m/z 148, 130, and 84
peaks obtained in MS2 analysis of Glu + H+ at CE of 15 eV
(Figure 4c–e) show that the m/z 84 peak is much broader and
tailing than the other two, suggesting more than one
structure in the m/z 84 fragment as proposed by Harrison
[22, 23]. Further study is required to reveal the structures
and pathways involving the m/z 84 fragment.
Dissociation Pathways and Substructures
of Protonated Glu
Energy resolved mass spectrometry (ERMS) curves of the
precursor Glu+H+ (m/z 148 MS2) and its m/z 130 and 102
fragments (MS3) are shown in Figure 3 to reveal the relative
energetics for dissociation pathways of Glu + H+. Both the
m/z 130 and 102 ions are formed directly from Glu + H+ ion
at m/z 148 (Figure 3a). The curve of the m/z 102 fragment
runs almost parallel with the m/z 130 fragment, albeit with a
slightly higher energy. ERMS plots of the m/z 130 fragment
(Figure 3c), confirmed that dissociation from m/z 130 to
form the m/z 102 fragment is only a minor pathway. The
comparison between the dissociation curves of [Glu + H]+,
the 130 Da and 102 Da fragments, (Figure 3a, c, and e)
indicates that the major pathway to form the 102 Da
fragment is from the molecular ion. In general, the
dissociation pattern of the 102 Da fragment is similar to
that of the Glu + H+ and 130 Da (Figure 3b, d, and f). The
m/z 84 fragment is the major fragmentation product of the
m/z 130 and 102 ions (Figure 3c and e). This m/z 84 fragment
The dissociation pathway map of Glu + H+ is shown in
Scheme 1, now extended to low m/z fragments. The
pathways to form fragments ≥m/z 84 were discussed
previously. At higher CE (40 eV), the m/z 84 fragment
dissociates to m/z 56 [10, 11, 21] and m/z 41 fragments.
These fragments further dissociate to low abundance,
smaller fragments (m/z 39, 30, 29, 28, and 27) at a CE of
55 eV. The m/z 39 fragment can be formed from the m/z 41
or 56 fragments. The m/z 29 fragment does not originate
from either of the fragments, m/z 41 or 39, because the m/z
29 fragment is pure 3,4-C2H5, where 2,3-C2H5 is not
observed. Because the m/z 41 and m/z 39 fragments are
symmetrical, if they dissociate to the m/z 29 fragment, the
2,3-C2H5 should exist too. In summary, the m/z 29 fragment
originated from the m/z 56 fragment, or possibly directly
dissociated from the m/z 84 fragment.
The m/z 28 fragment is 2-CH2N, as discussed before in
Figure 2. This fragment and the m/z 29 fragment (3,4-C2H5)
may be charge-competing products from the m/z 56
fragment, depending on where the proton is attached. The
possibility that this fragment is directly dissociated from the
m/z 84 fragment cannot be excluded. In contrast to these two
fragments, the m/z 27 and 30 fragments are mixtures and
there is more than one pathway to form them. Both
fragments could originate from the m/z 84 or m/z 56
fragments. Because m/z 27 is a mixture of 2,3-C2H3 and
3,4-C2H3, it may also originate from the m/z 41 and 39
fragments.
Figure 3. ERMS of dissociation of Glu + H+ (m/z 148, MS2),
the m/z 130 fragment (MS3), and the m/z 102 fragment (MS3)
in a QqQ. (a) ERMS of Glu + H+. (b) Normalized ERMS of low
m/z fragments (≤m/z 84) of Glu + H+. Each fragment is
normalized to its own maximum; (c) ERMS of m/z 130
fragment; (d) Normalized ERMS of low m/z fragments (≤m/z
84) of m/z 130 fragment. (e) ERMS of the m/z 102 fragment; (f)
Normalized ERMS of low m/z fragments (≤m/z 84) of the m/z
102 fragment. Each fragment is normalized to its own
maximum

X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
629
Figure 5. MS2 of protonated Glu, Gln, Pro, and Ala
performed on a QqQ in the product ion scan mode. (a)
Glu + H+ (m/z 148); (c) Gln + H+ (m/z 147); (e) Pro + H+ (m/z
116); (g) Ala + H+ (m/z 90). On the left (a, c, e, and g)
superimposed spectra are shown; a, c, and e are obtained at
a CE of 15 eV (blue) and 55 eV (red); g is obtained at a CE of
13 eV (blue) and 65 eV (red). On the right (b, d, f, and h),
blow-ups of the low m/z region (m/z 26–35, region in
rectangle, CE 55 eV for b, d, and f, CE 65 eV for h) are
shown by magnifying 30 to 100 times. The spectra are
normalized to the highest peak in each m/z range
Figure 4. FTICR gas-phase HDX and IM-MS measure-
ments. (a,b) FTICR gas-phase HDX of the m/z 84 fragment
at 0 and 10 s (c–e) ion mobility drift times of m/z 148, 130,
and 84 peaks in the MS2 spectrum of Glu + H+ at a 10 eV CE
performed on Synapt G2 Q-TOF-IM-MS. The drift time peak
of m/z 84 is tailing and much wider than those of the m/z 148
and 130, suggesting more than one isomer exists in the m/z
84 fragment
distinguish, identify, and eventually quantify each labeled
carbon in protonated Glu and Gln, specific fragments should
be monitored. Table 2 summarizes the origins of the carbons
in the low m/z fragments. Carbon identification is as follows;
by comparing the shifts of the precursor and m/z 84 fragment
in isotopically-labeled protonated Glu and Gln, C1 can be
monitored. Similarly, the shifts of the m/z 84 and 56
fragments can be utilized to track C5. The shift of the m/z
28 fragment can be used to monitor C2 and the m/z 31 peak
(shifted from 30) can be used to track C2 and C4. Because
C2 can be monitored based on the m/z 28 fragment, C4 can
also be tracked based on this m/z 31 peak. If C1, C2, C4, and
C5 are monitored, C3 can also be deduced based on the
information of the precursor and other carbons. This
knowledge of the low m/z fragments is necessary to monitor
each isotopically-labeled carbon in protonated Glu and Gln.
Low m/z Fragments of Other Amino Acids
MS2 fragmention patterns of protonated Gln, Pro and Ala
were explored by QqQ and Q-TOF and compared with those
of protonated Glu to elucidate general rules for the low m/z
fragments of similar amino acids. Protonated Glu and Gln
have the same medium-to-low mass (≤m/z 84) fragmentation
pattern (Figure 5a–d, Table 2, and Figure S-2). To

630
X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
Table 2. Origins of carbons in low m/z fragments of protonated Glu, Gln, Pro, and Ala
m/z
Proposed formula
Glu
2,3,4-C
-
-
Gln
2,3,4-C
-
-
Pro
Ala
56
53
44
43
42
41
39
30
29
29
28
27
C3H6N
C4H5
C2H6N
C3H7
C2H4N
C3H5
C3H3
CH4N
C2H5
CH3N
CH2N
C2H3
-
-
-
2,3,4,5-C
2,3-C and 3,4-C 4,5-Ca
2,3,4-Ca and 3,4,5-Ca
2,3-C
2,3-C
-
-
-
-
-
2,3-C
2,3,4-C
2,3,4-C
2-C and 4-C
3,4-C
2,3,4-C
2,3,4-C
2-C and 4-C
3,4-C
2,3,4-Ca and 3,4,5-Ca
-
-
-
-
-
2-C and 5-C 4-C 3-Ca
3,4-Ca
-
-
-
2-C
2-Cb
2,3-C
2-C
2-C
2-C and 5-C 4-C 3-Ca
2,3-C and 3,4-C 4,5-Ca
2,3-C and 3,4-C
2,3-C and 3,4-C
a
Origins of carbons 3,4,5 are not validated.
Not confirmed whether it is a pure origin.
b
In protonated Pro, the m/z 41 (C3H5), 30 (CH4N), and 28
Methylation of Glu and Gln
(CH2N) fragments common to Glu and Gln, are also
observed (Figure 5e, f, Table 2, and Figure S-3). The
formulae of these fragments are similarly confirmed by
isotopically-labeled samples (Figure S-3) and the accurate
masses measured by Q-TOF (data not shown). However, the
m/z 29 (C2H5) peak of Pro is very low in intensity relative to
that of the m/z 28 and 30 fragments, supporting our previous
suggestion that this ion is not originating from the m/z 41
fragment. In the m/z 84 fragment of Glu, there is an oxygen
on the C5, therefore C5 is likely to dissociate as 5-CO.
However, the m/z 70 fragment in protonated Pro (Figure 5e)
contains no oxygen and C5 is not as easy to dissociate as in
the m/z 84 in Glu. This is one of the reasons that it is more
difficult to form the m/z 29 (C2H5) fragment containing C3
and C4. Owing to the lack of appropriate isotopically-
labeled Pro standards, the origins of C3, C4 and C5 in Pro
fragments were not confirmed in this study.
Depending on the amount of labeling, the nominal
molecular weight (MW) of protonated, isotopically-
labeled Gln may be the same as that of unlabeled
protonated Glu or protonated, isotopically-labeled Glu.
For example, the MW of protonated, doubly 13C labeled
Gln, is 149 which is the same as the MW of protonated,
singly 13C carbon labeled Glu. This mass overlap should
be addressed when samples from mosquitoes or other
organisms are analyzed for quantification of each carbon
in isotopically-labeled amino acids. It is noted that liquid
chromatography separation of Glu and Gln is also
challenging and usually involves ion pairing. Derivatiza-
tion of the carboxylic acid groups of Glu and Gln
eliminates this interference because these two amino
acids have a different number of acidic sites. Therefore,
with derivatized Gln and Glu, the mass of protonated,
isotopically-labeled Gln will not overlap with protonated
Glu or protonated, isotopically-labeled Glu. However, the
derivatizing group may form additional low m/z frag-
ments in CID MS2 and interfere with the analysis of the
low m/z fragments of the amino acids. Also, derivatiza-
tion may change the dissociation pathways of the
“native” amino acids.
We found that methylation of the carboxylic acids is a
specifically suitable way to derivatize Glu and Gln to
avoid m/z overlapping. Trimethylsilyldiazomethane is
used to methylate carboxylic acid groups under mild
conditions [25] in this research. The m/z value of the
protonated, carboxylic acids-methylated Glu is 176 and
that of the protonated, carboxylic acid-methylated Gln is
161. Figure S-1 shows the fragments of the protonated
dimethylated Glu. The fragments ≤m/z 84 are the same
as those of the underivatized protonated Glu. As a result,
the derivatization does not interfere with the analysis of
the low m/z fragments of the amino acids. Thus,
methylation can be used to eliminate cross m/z interfer-
ence of isotopically-labeled Gln and Glu or isotopically-
labeled Glu, in biological samples.
Protonated Ala has two dominant fragments in CID MS2
(Figure 5g, h) – m/z 44 and 29 fragments. The m/z 44 peak
shifts up by one Da in the [15N] – Ala, [2-13C] – Ala and
[3-13C] – Ala (Figure S-4 c, g, and i) indicating that this
peak contains nitrogen, C2 and C3, respectively. The
formula of this m/z 44 peak is therefore 2,3-C2H6N, which
is the result of losing the carboxylic acid group from the
precursor. Unlike the other three amino acids, the m/z 29
peak in protonated Ala is not C2H5. This peak shifts up by
only one Da in [15N] – Ala and [2-13C] – Ala (Figure S-4d
and S-4h) suggesting the formula to be 2-CH3N. This peak
is the only radical ion fragment observed in this study
among all nitrogen containing fragments. The m/z 27 peak
shifts to m/z 28 in [2-13C] – Ala, [3-13C] – Ala (Figure S-4h
and S-4j). Thus the m/z 27 peak is the same as that observed
in the other amino acids (2,3-C2H3).
In summary, backbone fragments containing nitrogen and
C2 are the same in the four amino acids studied in this research.
The m/z 28 fragment in protonated Glu, Gln and Pro and the
m/z 29 fragment in Ala can be used to monitor metabolism of
13C carbons at position 2 in these amino acids (Table 2).

X. Ma et al.: Fragmentation of 13C-Labeled Amino Acids
631
2. Zamboni, N.: 13C metabolic flux analysis in complex systems. Curr.
Opin. Biotechnol. 22, 103–108 (2011)
Conclusions
3. Wiechert, W., Mollney, M., Petersen, S., de Graaf, A.A.: A universal
framework for 13C metabolic flux analysis. Metab. Eng. 3, 265–283 (2001)
4. Weitzel, M., Wiechert, W., Noh, K.: The topology of metabolic isotope
labeling networks. BMC Bioinf. 8, 315 (2007)
5. Rantanen, A., Rousu, J., Jouhten, P., Zamboni, N., Maaheimo, H., Ukkonen,
E.: An analytic and systematic framework for estimating metabolic flux
ratios from 13C tracer experiments. BMC Bioinf. 9, 266 (2008)
The characterization of the fragmentation pattern of proton-
ated Glu, Gln, Pro and Ala is extended here to fragments as
low as m/z 27. The low m/z fragments are identified and
dissociation pathways to form them are proposed. This
knowledge is used to develop a new approach to distinguish
each 13C-labeled carbon in the above amino acids, and
potentially for more amino acids, by CID MS2. The m/z 28
fragment containing nitrogen and C2 is common to
protonated Glu, Gln and Pro. A similar m/z 29 fragment
with C2 and nitrogen is also observed in protonated Ala.
These are potential fragments to monitor C2 in the amino
acids by multiple reaction monitoring (MRM). The m/z 30
fragment in protonated Glu and Gln is a mixture of 2-CH4N
and 4-CH4N, suggesting more than one pathway to form it
and one of these pathways may involve rearrangement. This
fragment can be used to monitor C2 and C4. The m/z 29 and
27 fragments in protonated Glu and Gln have different
carbon origins, indicating they originate from different
pathways. The above carbon origins may also be fit for
other amino acids.
Methylation of carboxylic acids is a specifically
suitable way to further separate the MWs of Glu and
Gln to avoid MW overlapping, without changing or
interfering with the low m/z fragments. Although the
intensity of the low m/z peaks is much weaker than the
abundant high m/z fragments, by using MRM, the
sensitivity can be improved to be compatible with typical
levels in biological samples. The suggested MRM
approach will be tested in our laboratory to monitor the
incorporation of 13C from 13C-glucose into Glu, Gln, Pro
and Ala in A. aegypti mosquitoes.
6. Scaraffia, P.Y., Zhang, Q.F., Wysocki, V.H., Isoe, J., Wells, M.A.:
Analysis of whole body ammonia metabolism in Aedes aegypti using
[
15N]-labeled compounds and mass spectrometry. Insect Biochem. Mol.
Biol. 36, 614–622 (2006)
7. Scaraffia, P.Y., Tan, G., Isoe, J., Wysocki, V.H., Wells, M.A., Miesfeld,
R.L.: Discovery of an alternate metabolic pathway for urea synthesis in
adult Aedes aegypti mosquitoes. Proc. Natl. Acad. Sci. U.S.A. 105, 518–
523 (2008)
8. Scaraffia, P.Y., Zhang, Q.F., Thorson, K., Wysocki, V.H., Miesfeld,
R.L.: Differential ammonia metabolism in Aedes aegypti fat body and
midgut tissues. J. Insect Physiol. 56, 1040–1049 (2010)
9. Bush, D.R., Wysocki, V.H., Scaraffia, P.Y.: Study of fragmentation of
arginine isobutyl ester applied to arginine quantification of Aedes
aegypti mosquito excreta. J. Mass Spectrom. 47, 1364–1371 (2012)
10. Pingitore, F., Tang, Y., Kruppa, G.H., Keasling, J.D.: Analysis of amino
acid isotopomers using FT-ICR MS. Anal. Chem. 79, 2483–2490 (2007)
11. Tang, Y., Pingitore, F., Mukhopadhyay, A., Phan, R., Hazen, T.C.,
Keasling, J.D.: Pathway confirmation and flux analysis of central
metabolic pathways in Desulfovibrio vulgaris Hildenborough using gas
chromatography-mass spectrometry and Fourier transform-ion cyclotron
resonance mass spectrometry. J. Bacteriol. 189, 940–949 (2007)
12. Jiang, W., Wysocki, V.H., Dodds, E.D., Miesfeld, R.L., Scaraffia, P.Y.:
Differentiation and quantification of C1 and C2 13C-labeled glucose by
tandem mass spectrometry. Anal. Biochem. 404, 40–44 (2010)
13. Szyperski, T., Glaser, R.W., Hochuli, M., Fiaux, J., Sauer, U., Bailey,
J.E., Wuthrich, K.: Bioreaction network topology and metabolic flux
ratio analysis by biosynthetic fractional 13C labeling and two-dimen-
sional NMR spectroscopy. Metab. Eng. 1, 189–197 (1999)
14. de Graaf, A.A., Mahle, M., Mollney, M., Wiechert, W., Stahmann, P.,
Sahm, H.: Determination of full 13C isotopomer distributions for
metabolic flux analysis using heteronuclear spin echo difference NMR
spectroscopy. J. Biotechnol. 77, 25–35 (2000)
15. Wittmann, C.: Metabolic flux analysis using mass spectrometry. Adv.
Biochem. Eng. Biotechnol. 74, 39–64 (2002)
16. Jeffrey, F.M.H., Roach, J.S., Storey, C.J., Sherry, A.D., Malloy, C.R.:
13C isotopomer analysis of glutamate by tandem mass spectrometry.
Anal. Biochem. 300, 192–205 (2002)
17. Peng, L.F., Arauzo-Bravo, M.J., Shimizu, K.: Metabolic flux analysis
for a ppc mutant Escherichia coli based on 13C-labeling experiments
together with enzyme activity assays and intracellular metabolite
measurements. Fems. Microbiol. Lett. 235, 17–23 (2004)
Acknowledgments
The authors thank Yang Song for her contribution in the early
stages of this work, Mowei Zhou for his help on G2 Q-TOF,
and George Tsaprailis and Yelena Feinstein for access to the
AB/SCIEX 3000 QqQ and AB/SCIEX 4000 QTRAP mass
spectrometers at the Arizona Proteomics Consortium, The
University of Arizona. The AB/SCIEX 4000 QTRAP mass
spectrometer was provided by NIH/NCRR Grant
1S10RR022384-01. The authors are also grateful to Dr. Yayoi
Hongo (RIKEN, Wako, Japan) and David R. Bush (The
University of Arizona) for fruitful discussions, and Dr.
Kanamatsu (Soka University, Tokyo, Japan) and Dr. Takatori
(Meiji Pharmaceutical University, Tokyo, Japan) for kindly
providing [4-13C] – Gln and [4-13C] – Glu, respectively, which
are not commercially available. This work was financially
supported by NIH/NIAID Grant R01AI088092 (to PYS).
18. Shimizu, K.: Metabolic flux analysis based on 13C-labeling experiments
and integration of the information with gene and protein expression
patterns. Adv. Biochem. Eng. Biotechnol. 91, 1–49 (2004)
19. Wittmann, C.: Fluxome analysis using GC-MS. Microb. Cell Fact. 6, 6
(2007)
20. Kamleh, M.A., Dow, J.A., Watson, D.G.: Applications of mass
spectrometry in metabolomic studies of animal model and invertebrate
systems. Brief. Funct. Genomics Proteomics. 8, 28–48 (2009)
21. Piraud, M., Vianey-Saban, C., Petritis, K., Elfakir, C., Steghens, J.P.,
Morla, A., Bouchu, D.: ESI-MS/MS analysis of underivatised amino
acids: a new tool for the diagnosis of inherited disorders of amino acid
metabolism. Fragmentation study of 79 molecules of biological interest
in positive and negative ionisation mode. Rapid Commun. Mass
Spectrom. 17, 1297–1311 (2003)
22. Harrison, A.G.: Ion chemistry of protonated glutamic acid derivatives.
Int. J. Mass Spectrom. 210, 361–370 (2001)
23. Harrison, A.G.: Fragmentation reactions of protonated peptides con-
taining glutamine or glutamic acid. J. Mass Spectrom. 38, 174–187
(2003)
24. Harrison, A.G.: To b or not to b: the ongoing saga of peptide b ions.
Mass Spectrom. Rev. 28, 640–654 (2009)
25. Presser, A., Hufner, A.: Trimethylsilyldiazomethane - A mild and
efficient reagent for the methylation of carboxylic acids and alcohols in
natural products. Monatsh. Chem. 135, 1015–1022 (2004)
References
1. Wiechert, W.: 13C metabolic flux analysis. Metab. Eng. 3, 195–206
(2001)