Fragmentation of deprotonated cyclic dipeptides by electrospray ionization mass spectrometry

Article abstract of DOI:10.1002/jms.1595

The fragmentation pathways of deprotonated cyclic dipeptides have been studied by electrospray ionization multi-stage mass spectrometry (ESI-MSn) in negative mode. The results showed that the fragmentation pathways of deprotonated cyclic dipeptides depend

Full text of DOI:10.1002/jms.1595

Research Article
Received: 27 February 2009
Accepted: 14 April 2009
Published online in Wiley Interscience: 21 May 2009
(www.interscience.com) DOI 10.1002/jms.1595
Fragmentation of deprotonated cyclic
dipeptides by electrospray ionization mass
spectrometry
Yanchun Guo,a Shuxia Cao,a∗ Donghui Wei,a Xiangkun Zong,a
Xingbo Yuan,a Mingsheng Tang,a and Yufen Zhaoa,b∗
The fragmentation pathways of deprotonated cyclic dipeptides have been studied by electrospray ionization multi-stage
mass spectrometry (ESI-MSn) in negative mode. The results showed that the fragmentation pathways of deprotonated cyclic
dipeptides depended significantly on the different substituents, the side chains of amino acid residues at the diketopiperazine
ring. In the spectra of deprotonated cyclic dipeptides, the ion [M–H–substituent radical]− was firstly observed in the ESI
mode. The characteristic fragment ions [M–H–substituent radical]− and [M–H–(substituent–H)]− could be used as the symbols
of particular cyclic dipeptides. The hydrogen/deuterium (H/D) exchange experiment, the high-resolution mass spectrometry
(Q-TOF)andtheoreticalcalculationswereusedtorationalizetheproposedfragmentationpathwaysandtoverifythedifferences
between the fragmentation pathways. The relative Gibbs free energies (ꢀG) of the product ions and possible fragmentation
pathways were estimated using the B3LYP/6–31++G(d, p) model. The results have some potential applications in the structural
elucidation and interpretation of the mass spectra of homologous compounds and will enrich the gas-phase ESI-MS ion
c
chemistry of cyclic dipeptides. Copyright ꢀ 2009 John Wiley & Sons, Ltd.
Keywords: cyclic dipeptides; electrospray ionization; fragmentation pathway; theoretical calculation; radical anion
Introduction
chicken essence and verified the molecular structures of the
DKPs using electrospray ionization tandem mass spectrometry
(ESI-MS/MS). Moreover, they found that the protonated molecule
often produces [M C H ꢁ CO ꢁ NH3]C ion accompanied by
losing CO and NH3.[5] A. E. M. Crotti’s group analyzed a series of
five-membered lactones (both saturated and unsaturated) and
lactam by ESI-MS/MS and electrospray ionization multi-stage
mass spectrometry (ESI-MSn).[10] It was observed that the main
fragment ions of their protonated molecules are derived from
neutral losses of CO and/or H2O (NH3 for the lactam). Recently,
they used ESI-MS/MS to investigate the fragmentation of a series
of DKPs previously isolated from Aspergillus fumigatus and found
that losing CO directly from the protonated molecule was a
fragmentation process common to all the compounds analyzed.
Furthermore, their results revealed a series of ions that are
diagnostic of the substituents at the DKPs’ ring.[11] With regard
to deprotonated ions, there are just some elucidations about
the collision-induced spectra of a variety of peptides.[12] Many
studies[13–15] of small deprotonated peptides suggested that the
fragmentation modes of the [M ꢁ H]ꢁ ions depended significantly
on the different substituents at the side chain of amino acids.
Cyclic dipeptides [(piperazine 2,5-diones or 2,5-diketopiperazines
(DKPs)] are relatively simple compounds and, therefore, are
among the most common peptide derivatives found in nature.
In comparison to linear peptides, cyclic peptides are more
bioavailable and more stable to degradative peptidases, and
their research has been fundamental to many aspects of peptide
chemistry. It was found that naturally occurring and synthesized
cyclic dipeptides exhibit a multitude of interesting biological
activities as antiviral, antibiotic, antimicrobial and antitumor
agents.[1–4] There are many reports in the literature that cyclic
dipeptides were identified in several beverages and foods,
especially fermented and thermally treated ones, which include
aged sake, beer, cocoa, dried squid,[5] hydrolyzed vegetable
protein,[6] roasted coffee,[7,8] roasted malt and others. However,
identification of the structures of cyclic dipeptides in foods
and beverages still has low throughput because of the limited
availability of standards.[5] With the advent of mild ionization
techniques, electrospray ionization mass spectrometry (ESI-MS)
using low-energy collision-induced dissociation (CID) has become
an increasingly important tool for the analysis of various kinds
of compounds with excellent sensitivity. But, for full structure
elucidation based on their mass spectrometry data, it is necessary
to understand all the gas-phase chemistry involved in each
fragmentation step and to clarify the relationships between all the
different possible pathways. Researchers have synthesized their
own cyclic dipeptides of interest to achieve some identification
by mass spectrometry for each instance.[7–9] Up to now, some
reports have been published concerning the systemic studies
of the fragmentation of protonated cyclic dipeptides by ESI-
MS in positive mode. Y. H. Chen et al. have identified DKPs in
Ł
Correspondenceto:Shuxia CaoandYufen Zhao,KeyLaboratoryofChemicalBi-
ologyandOrganicChemistry,DepartmentofChemistry,ZhengzhouUniversity,
Zhengzhou,P.R.China.E-mail: csx@zzu.edu.cn;nmy05@mails.tsinghua.edu.cn
a
Key Laboratory of Chemical Biology and Organic Chemistry, Department of
Chemistry, Zhengzhou University, Zhengzhou, P.R. China
b
Key Laboratory for Bioorganic Phosphorus Chemistry of Ministry of Education,
Department of Chemistry School of Life Sciences and Engineering, Tsinghua
University, Beijing, P.R. China
c
J. Mass. Spectrom. 2009, 44, 1188–1194
Copyright ꢀ 2009 John Wiley & Sons, Ltd.

Fragmentation of deprotonated cyclic dipeptides
For example, the study on the fragmentation reactions of a
variety of deprotonated dipeptides and tripeptides containing
phenylalanine revealed that the benzyl-group has a substantial
effect on the fragmentation reactions observed.[13] Although
there are many reports about the fragmentation pathways of
deprotonated peptides and protonated cyclic dipeptides, to the
best of our knowledge, much less is known so far concerning
the investigation of structure–fragmentation relationships of
deprotonated cyclic dipeptides. J. H. Bowie’s group has studied
the mass spectra of some deprotonated cyclic dipeptides in the
negative chemical ionization (CI) mode.[16] It was found that there
were two major fragmentations of [M ꢁ H]ꢁ ions for symmetrical
and unsymmetrical cyclic dipeptides. One fragmentation pathway
was the characteristic side-chain loss that could be used to identify
the particular cyclic dipeptide, and the other fragmentation
pathway was an unusual loss of RCHO (R is the substituent) that
involved initial 1,2 migration of R to the carbon of the adjacent
carbonyl group. Nevertheless, there is not report describing
studies of the fragmentation of deprotonated cyclic dipeptides
in ESI negative mode. It is known that the fragmentation
pathways of protonated and deprotonated species generated
by ESI are quite different from those generated by other
ionization techniques, and now, ESI-MS is used extensively, so an
understandingofthefragmentationmechanismsofdeprotonated
cyclic dipeptides in ESI mode has critical significance. It can be
applied to the fragmentation elucidation of larger more complex
natural products and compounds of biological significance that
may contain cyclic dipeptides as part of their structures. At
the same time, this study can also enrich the gas-phase ion
chemistry of ESI-MS under low-energy CID condition. In this study,
eight symmetrical cyclic dipeptides were synthesized and their
fragmentation pathways in ESI negative mode were studied
to evaluate the role of the substituents at the DKPs ring on
the fragmentation of the deprotonated cyclic dipeptides. The
fragmentationmechanismswereinvestigatedbymulti-stagemass
spectrometry incorporating hydrogen/deuterium (H/D) exchange
experiments. Furthermore, the results were validated by the ESI-
MS/MS data of high-resolution accurate-mass quadrupole-time of
flight (Q-TOF) and theoretical calculations. The relative Gibbs free
energies (ꢀG) of the product ions and possible fragment ions were
estimated using the B3LYP/6–31CCG (d, p) model.
Scheme 1. Synthesis of cyclic dipeptides.
are cyclo(Ala-Ala)(1), cyclo(Leu-Leu)(2), cyclo(Ile-Ile)(3), cyclo(Val-
Val)(4), cyclo(Phe-Phe)(5), cyclo(Trp-Trp)(6), cyclo(Tyr-Tyr)(7) and
cyclo(Met-Met)(8). Methanol (HPLC grade) was obtained from
TediaCompanyanddeionizedwater(Milli-Q)wasusedthroughout
the study. The reaction and the structures of cyclic dipeptides are
shown as Scheme 1.
Mass spectrometry
Mass spectrometric analysis was performed by a Bruker Esquire
3000 (Bruker Dalton, Germany) ion trap mass spectrometer
interfaced with an ESI source. Ionization of analytes was carried
out using the following setting of ESI: nebulizer gas flow 7 psi,
dry gas 4 l/min, dry temperature 300 ŽC and capillary voltage
4000 V. Calibration of m/z was performed using a standard ESI-
tuning-mix. Scan range was 15–500 m/z and scan resolution
was normal (13 000 m/z/s). ESI-MSn spectra were obtained by
CID experiments with helium after isolation of the appropriate
precursor ions. High-resolution mass spectra were performed on
an ESI-Q-TOF-MS spectrometer (Micromass, England). Solutions
of the compounds (0.1 mg/ml) were prepared by dissolving the
compounds in methanol.
Computational methods
Experimental
Using density functional theory, the geometries of the reactant, in-
termediates and products in all reaction channels were optimized
at the B3LYP/6–31CCG (d, p) level. All theoretical calculations
were performed using the GAUSSIAN03[17] suite of software.
Chemicals
Compounds 1–8 were synthesized from L-amino acids and
phosphorus trichloride. Each L-amino acid (20 mM) was dissolved
in tetrahydrofuran (THF) (30 ml), and then PCl3 (10 mM) was
added in batches. The mixture was stirred and refluxed for
2 h. After evaporation of the solvent, H2O was added and
the solution was adjusted to pH 7–8 with saturated aqueous
NaHCO3. Then, the reaction mixture was filtered and the solid
was washed with water (3 ð 10 ml). The crude cyclic dipeptide
was obtained. Analytical sample was obtained by column
chromatography (silica gel, methanol/ethyl acetate D 1 : 19 to
1 : 4) or recrystallization from acetone and petroleum ether as
white solid. The yield of cyclic dipeptide was 20–60%. The
structures were characterized by 1H NMR, 13C NMR, ESI-MS,
high resolution mass spectrometer (HRMS) and IR methods.
The symmetrical cyclic dipeptides investigated in this study
Results and Discussion
Negative ESI-MS/MS spectra of cyclic-dipeptides
ESI-MSn spectraofcompounds1–8innegativemodewerestudied
in detail and the data are shown in Tables 1 and 2. Comparison
among the structures of compounds 1–8, which differ only in the
side-chain substituent at C(3) and C(6) of DKP ring, revealed that
the fragmentation patterns are correlative with the types of these
substituents at C(3) and C(6). Therefore, in accordance with the
different substituent at DKP ring, the cyclic dipeptides are divided
into three groups (1) compounds 1–4 (R is an alkyl substituent),
(2) compounds 5–7 (R contains an aromatic substituent) and
c
J. Mass. Spectrom. 2009, 44, 1188–1194
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

Y. Guo et al.
(3) compound 8 (R is an alkyl substituent containing a hetero-
atom). A significant difference was observed in the cleavage of
deprotonated cyclic dipeptides in low-energy CID process due
to the effect of the side-chain substituents on fragmentation.
The ESI-MSn spectral data showed that there are three main
fragmentation pathways (path a, b and c), as shown in Scheme
2. The fragmentations from ions A occur at the side chain with
consecutive elimination of a CO and a saturated substituent
molecule (R C H), elimination of an unsaturated substituent
molecule (R ꢁ H) or elimination of a substituent radical to form
ions D, B1 or C1, respectively. The first pathway (a) occurs only for
compounds 1–4. The second pathway (b) occurs for compounds
6 and 7 and the third pathway (c) occurs for compounds 1–5.
It seems strange that for compounds 1–5, the formation of the
product ion C1 involves an elimination of an odd-electron loss
and a substituent radical might be lost in the fragmentation
process.
The electron impact ionization (EI) mass spectra of small organic
compounds, recorded at 70 eV, are often performed to identify
unknown compounds because of its excellent reproductibility and
structure-specific nature of the dissociation products. In EI spectra,
the loss of free radical is very common in the fragmentation
process. However, for the need of volatility of the sample, the
application of EI is limited in peptides analysis. Now, soft ionization
methodssuchasESIandmatrixassistedlaserdesorptionionization
(MALDI) have been widely applied to produce protonated and
deprotonated proteins or peptides, and these ions yielded under
CIDconditionshavebeenusedmoreandmoretodetermineamino
acid sequences.[18–20] Comparison with EI, CID process provides
lower energy, and the unusual losses, such as the loss of a radical,
are rarely observed. However, there are still some reports about
the observation of the radical loss under CID conditions.[11,21–25]
K. Whitehead’s study about seven mycosporine-like amino acids
(MAAs) showed that the small loss of 15 Da was the result of the
loss of a group radical from the protonated MAA molecule.[24]
Subsequently, K.H.M. Cardozo et al. confirmed the small loss
was a methyl radical by (H/D)exchange, ESI-MS/MS and high-
resolution mass spectrometry (Q-TOF).[25] Recently, by theoretical
calculations, they further demonstrated that losing a radical in
the CID process for the protonated MAA molecule was reasonable
and the elimination of mass 15 Da involved a methyl radical.[26]
However, most of the observations of radical loss under CID
condition come from cations in positive mode. To the best of
our knowledge, only a few reports have been published so far
describing studies of the fragmentation of odd-electron peptide
radical anions through low-energy CID in ESI negative mode.[27]
Considering the ability of CID to produce odd-electron radical
cation and anion, it is supposed that the odd-electron loss from
ions A to form ions C1 in the fragmentation reaction of cyclic
dipeptides might be a radical loss.
tion of another benzyl radical from ion C1, the product ionC2
at m/z 112 is formed that still includes the active hydrogen. The
fragment ion at m/z 146 is produced by a loss of amino acid
residue from deprotonated molecule ion A, accompanied by the
active hydrogen lost. Corresponding to the deuterated sample,
the non-deuterated deprotonated molecule A is observed at m/z
293 [shown in Fig. 1(5b)], and it can successively eliminate two
benzyl radicals to produce the ions C1 and C2 at m/z 202 and
111, respectively. Therefore, the deuterium-labeled experiment
of compound 5 can prove that the processes of the formation
of ions C1 and C2 do not relate to the loss of active hydrogen.
Moreover, the base peak at m/z 146 is formed by losing an amino
acid residue that includes the active hydrogen. However, different
from compound 5, compounds 6 and 7 often eliminate an unsat-
urated substituent molecule to form ions B1 and B2 as shown in
Scheme 2. Furthermore, the process of the formation of ionsB1
and B2 does not relate to the loss of active hydrogen either. In
Fig. 1(6a), the deuterium-labeled deprotonated molecule at m/z
372 of compound 6 gives a base peak ion [M(D) ꢁ H ꢁ 129]ꢁ
at m/z 243, corresponding to ion B1. It is formed by a neutral
loss of the unsaturated substituent molecule, 3-methylene-indole,
which has a bigger conjugated structure. The fragment ion at m/z
114 corresponding to ion B2 is produced by continuously losing
another unsaturated substituent molecule. The non-deuterated
deprotonated molecule at m/z 371 likewise produces ions B1 and
B2 at m/z 242 and 113, respectively as shown in Fig. 1(6b). There-
fore, the deuterium-labeled experiment of compound 6 proves
that the ions B1 and B2 still contain the active hydrogen after
elimination of the unsaturated substituent molecules. Apparently,
there is not any fragment ion produced by a radical lose for com-
pound 6. So the deuterium-labeled results of compounds 5 and 6
indirectly testify to the fragmentation pathway b and c in Scheme
2. Moreover, the results also indicate that the deprotonation has
occurred directly at the nitrogen of the DKP ring, not at the side
chain of compounds 6 and 7.
As compound 6, compound 7 can also eliminate its stable
neutral substituent molecule, 4-methylenecyclohexa-2,5-dienone
O
C6H4 CH2, to form ions B1 and B2 at m/z 219 and 113,
respectively. However, according to the compound 5, cyclo(Phe-
Phe) with benzyl substituent, its base peak in the deprotonated
ESI-MS/MS spectrum is ion F formed by the elimination of a
phenylalanine residue molecule, since it cannot eliminate a stable
substituent molecule with a bigger conjugative structure as those
of compounds 6 and 7 (data in Table 1). Since the produced
fragment ion F of compound 5 contains a bigger conjugative
structure, it is observed as the base peak. The substituent of
compound8containsasulfuratom,sothefragmentationreactions
are always driven by the sulfur atom. Just like their protonated
molecule ion, the heterolytic cleavage takes place near the sulfur
atom.Then,thedeprotonatedcyclo(Met-Met)ispronetoeliminate
one or two CH3SH molecule to form ionsH1 ([M ꢁ H ꢁ CH3SH]ꢁ)
or H2 ([M ꢁ H ꢁ 2CH3SH]ꢁ) at m/z 213 or 165, respectively. The
ion H1 is observed as the base peak in the ESI-MS2 spectra of
the deprotonated compound 8. To test the mechanism proposed
above, ESI-MS/MS of the deprotonated cyclic dipeptides was
carried out by high-resolution mass spectra. High-resolution mass
data allowed us to determine the elemental composition for each
ionicspeciesdetectedinESI-MS/MSspectra.Thedataarealsolisted
in Table 1, and the results are in agreement with the proposed
fragmentation pathways.
To validate the mechanisms, the deuterium-labeled experi-
ments were first used to investigate the fragmentation pathways
of the deprotonated cyclic dipeptide. For example, the ESI-MS2
spectra of deuterium-labeled and non-deuterium-labeled com-
pounds 5 and 6 are shown in Fig. 1. Figure 1(5a, 6a) shows the
spectra of deuterated sample and Fig. 1(5b, 6b) shows those of
non-deuterated sample. As shown in Fig. 1, compound 5, as com-
pounds 1–4, always in the same way eliminates the substituent
radical from ion A to form ion C1. In Fig. 1(5a), the deuterium-
labeled deprotonated molecule at m/z 294 of compound 5 gives
a product ion C1 [M(D) ꢁ H ꢁ 91]ꢁ at m/z 203, which is formed
by the elimination of a benzyl radical. After successive elimina-
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1188–1194

Fragmentation of deprotonated cyclic dipeptides
Table 1. ESI-MS2 data of deprotonated cyclic-dipeptides
Type of ions (relative abundance %)
Compounds
A
B1
B2
C1
C2
D
F
H1
H2
Cyclo(Ala-Ala)(1)
141.0(3)
125.8(19)
126.0429
126.0398
167.7(52)
168.0899
168.0925
167.7(100)
168.0899
168.0941
153.7(100)
154.0748
154.0757
201.7(47)
202.0748
202.0610
96.7(100)
97.0528
70.1(2)
a
141.0664
141.0664
225.0(5)
70.0293
70.0277
b
97.0402
Cyclo(Leu-Leu)(2)
Cyclo(Ile-Ile)(3)
110.7(11)
111.0195
111.0184
110.9(9)
138.7(100)
139.0634
139.0881
138.7(56)
139.0634
139.0954
124.9(17)
125.0483
125.0737
a
225.1603
225.1603
225.0(60)
225.1603
225.1603
197(37)
b
a
111.0195
111.0234
110.6(26)
111.0200
111.0207
110.9(13)
111.0200
111.0145
b
Cyclo(Val-Val)(4)
Cyclo(Phe-Phe)(5)
Cyclo(Trp-Trp)(6)
Cyclo(Tyr-Tyr)(7)
Cyclo(Met-Met)(8)
a
197.1296
197.1290
293.0(17)
293.1296
293.1290
371.0(20)
371.1513
371.1508
325.0(4)
b
145.8(100)
146.0612
146.0523
a
b
241.6(100)
242.0935
242.0984
218.6(100)
219.0775
219.0808
184.6(2)
112.7(22)
113.0357
113.0422
112.7(33)
113.0356
113.0412
a
b
a
325.1194
325.1189
261.0(18)
261.0731
261.0731
b
212.7(100)
213.0697
213.0748
164.8(6)
165.0663
165.0734
a
185.0384
185.0510
b
a, the exact mass of ions by theoretical calculation; b, The actual mass of ions determined by high-resolution ESI-Q-TOF-MS; R, substituent. A, [M ꢁ
H]ꢁ; B1, [A ꢁ (R ꢁ H)]ꢁ; B2, [B1 ꢁ (R ꢁ H)]ꢁ; C1, [A ꢁ R]ꢁ; C2, [C1 ꢁ R]ꢁ; D, [A ꢁ CO ꢁ (R C H)]ꢁ; F, [A ꢁ residue]ꢁ; H1, [A ꢁ part of sbustituent]ꢁ; H2,
[H1 ꢁ part of sbustituent]ꢁ.
Table 2. ESI-MSn data of deprotonated cyclic-dipeptides
Precursor ions
Fragment ions
(relative abundance %)
Compounds
[M ꢁ H]ꢁ
Cyclo(Ala-Ala)(1)
Cyclo(Leu-Leu)(2)
141.0(3)
225.0(5)
96.7(100),125.8(19),70.1(2)
167.7(52), 138.7(100),
110.7(11)
Cyclo(Ile-Ile)(3)
225.0(60)
167.7(100), 138.7(56),
110.9(9)
168.0(9)
197(37)
110.8(100)
Cyclo(Val-Val)(4)
Cyclo(Phe-Phe)(5)
153.7(100), 124.9(17),
110.6(26)
293.0(17)
274.7(26), 201.7(47),
183.6(4), 145.8(100),
110.9(13)
275.0(16)
202.0(4)
183.7(100)
110.9(100)
146.0(100)
371.0(20)
242.0(37)
325.0(4)
42.8(69)
Cyclo(Trp-Trp)(6)
Cyclo(Tyr-Tyr)(7)
Cyclo(Met-Met)(8)
241.6(100), 112.7(22)
112.6(100)
218.6(100), 112.7(33)
112.7(100)
219.0(13)
261.0(18)
213.0(12)
212.7(100), 184.6(2),164.8(6)
184.7(13), 164.7(100),
147.7(19), 136.8(26),
120.7(6), 108.6(3)
Scheme 2. Proposed fragmentation pathways of deprotonated cyclic
dipeptides. The values above the arrows (in kcal/M) represent the relative
Gibbs free energies (ꢀG) for each individual fragmentation reaction.
c
J. Mass. Spectrom. 2009, 44, 1188–1194
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

Y. Guo et al.
Table 3. Bond lengths (Å) for neutral [M] and deprotonated [M ꢁ H]
compound 1
O8
CH3
10
5
HN
4
3
6
1
N
9
H3C
2
7
O
M
[M ꢁ H]ꢁ
D-valuea
C(2)–N(1)
C(2)–O(7)
N(1)–C(6)
C(5)–C(6)
C(5)–N(4)
C(5)–O(8)
N(4)–C(3)
C(3)–C(2)
C(3)–C(9)
C(6)–C(10)
1.361
1.224
1.462
1.535
1.361
1.224
1.462
1.535
1.528
1.528
1.320
1.258
1.452
1.535
1.357
1.237
1.462
1.577
1.526
1.533
ꢁ0.041
C0.034
ꢁ0.010
0.000
Figure 1. Deuterium-labeled (a) and non-deuterium-labeled (b) ESI-MS2
spectra of compounds 5 and 6 in negative mode.
ꢁ0.004
C0.013
0.000
C0.042
ꢁ0.002
C0.005
Theoretical calculations of deprotonated cyclic dipeptides
The fragmentation pathways of deprotonated cyclic dipeptides
revealed that the different fragmentation pathways are closely
related to their substituents at the DKP ring. In some cases, the
deprotonation is found to occur directly at the nitrogen of the
lactam group. Furthermore, using density functional theory, the
geometries of the deprotonated molecule were optimized at the
B3LYP/6–31CCG (d, p) level. The results indicated that nitrogen
atom of cyclic dipeptide is the most favorite deprotonation site
than carbon atom. Since the cyclic dipeptides studied are all
symmetrical molecules, N(1) site is completely equal to N(4)
site as shown in Table 3. Thus, the deprotonation is assumed
directly at N(1) atom of cyclic dipeptides instead of C(3) or
C(6). Because compound 1 is the simplest compound studied,
compound 1 is used as an example for theoretical calculations.
The geometries of the neutral and the deprotonated compound
1 were fully optimized and the bond lengths were calculated
using B3LYP/6–31CCG (d, p) model. The bond lengths data for
neutralanddeprotonatedcompound1areshowninTable 3.Some
differences can be obviously observed that the bond lengths of
C(2)–C(3), C(2)–O(7), C(5)–O(8) and C(6)–C(10) in deprotonated
species become longer comparing with those in neutral molecule
of compound1. The C(2)–C(3) bond length in the deprotonated
molecule is approximately 0.042 Å longer than the corresponding
one in the neutral molecule, accordingly that of C(2)–O(7) bond
is 0.034 Å longer. The C(5)–O(8) bond length in the deprotonated
molecule is approximately 0.013 Å longer than the corresponding
one in the neutral molecule, accordingly that of C(6)–C(10) bond
is 0.005 Å longer. In the same manner, if the deprotonation takes
place at N(4), the bond lengths of C(5)–C(6), C(5)–O(8), C(3)–C(9)
and C(2)–O(7) in deprotonated molecule will similarly be longer
than those in neutral molecule. The bond lengths of C(2)–O(7)
and C(5)–O(8) in the anion are much longer than those in the
neutral because deprotonation at atom N(1) results in a decrease
of the C(2)–N(1) distance by the negative charge migrating to
the oxygen atom. Therefore, the change of the C(2)–C(3) and
C(6)–C(10) bond after deprotonation may be an initial evidence
for the direction of further fragmentation reactions.
a Difference between bond lengths of neutral and deprotonated
compound 1.
CO and a saturated substituent molecule (R C H) to form ions D,
(path b) elimination of unsaturated substituent molecule (R ꢁ H)
to form ions B1 and B2 and (path c) elimination of substituent
radical to form ions C1 and C2. Path a is observed only for
compounds 1–4; path b is observed for compounds 6 and 7 and
path c is observed for compounds 1–5. Furthermore, the ions
C1 are observed as base peaks for compounds 3 and 4 and as
secondary intense peaks for compounds 1, 2 and 5. To rationalize
the relative abundances of some product ions in the spectra of
cyclic dipeptides, the B3LYP/6–31CCG (d, p) model was used
to calculate the relative Gibbs free energies for the reactant and
product ions. It is considered that in CID-ESI-MS process, the
energy that is transferred to the precursor ion due to the collision
activation can be even higher than the activation barrier for some
specific fragmentation reactions, thus the difference between the
initial and final energy states can be used to understand the
mechanism, and it is not necessary to consider the energies of
any transition state species.[26] Therefore, the relative Gibbs free
energies of compounds 1, 4 and 6, which represent three different
pathways a, c and b, respectively, were calculated and one of the
A isomers was assumed as the reference (0.0 kcal/M). The data of
calculation are shown in Scheme 2.
Most of the species of product ions in the ESI-MS spectra of
compounds 1–5 (listed in Tables 1 and 2) are similar, except
their different base peaks. This means that they follow similar
fragmentation pathways, and their different substituents play a
crucial role in the intensity of some peaks. For compounds 1 and
4, the formation of ions E from the deprotonated molecule A (via
path a) is an endergonic process, as evidenced by the relative
energies of A (0.0 kcal/M) and E (30.21 kcal/M for 1, 48.01 kcal/M
for 4) as shown in Scheme 2. The process can be interpreted in
terms of the energy that is required for the heterolytic cleavage of
the C(2)–C(3) bond, which length in the deprotonated molecule
of compound 1 is approximately 0.042 Å longer than that in the
neutral molecule, and for successive losing a CO molecule. This
energycontentisconsideredtobetransferredtothedeprotonated
As shown in Scheme 2, the MSn data suggested that the
fragmentation of the deprotonated cyclic dipeptides follows
three major competitive pathways starting from the deprotonated
molecule ions A: (path a) consecutive elimination of a molecule
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1188–1194

Fragmentation of deprotonated cyclic dipeptides
molecule owing to the CID process. The analysis of the relative
ꢀG revealed that the formation of the ions E from ions A is
a thermodynamically disfavored process. However, the product
ions E can spontaneously transform to ions D (1: ꢁ10.55 kcal/M, 4:
ꢁ27.90 kcal/M) by losing a molecule (R C H). Therefore, the ion D
is observed as the base peak for compound 1, while the relative
abundances of ion D is relatively lower (17%) for compound 4.
Another reason will be interpreted later.
product ions B1 and B2 are not formed in the spectra of com-
pounds 1–5. For compound 6, the process from the deprotonated
molecule A to form product ion B1 (via path b, ꢀG D 27.09 kcal/M)
is a thermodynamically more favored process in comparison with
the formation of ion D (path a, ꢀG D 44.11 kcal/M) and C1 (path
c, ꢀG D 40.38 kcal/M). Moreover, the lost unsaturated substituent
molecule (R–H), 3-methylene-indole, is very stable considering
its special conjugative structure. Therefore, for compound 6,
cyclo(Trp-Trp), it is obvious that the formation of the product ion
B1 is relatively easier than that of the product ions D and C1.
So in the ESI-MS/MS spectrum of compound 6, the base peak is
ion B1 and the secondary intense peak is the ion B2 formed by
the elimination of another unsaturated substituent molecule. For
compound1, however, thedeprotonatedformofionAisexpected
to produce predominantly the product ion D (30.21 kcal/M) or C1
(55.51 kcal/M) instead of B1 (not observed, 114.29 kcal/M), as the
former is a thermodynamically more favored process.
It is reported that the transfer of energy content on the CID
process is high enough to induce the homolytic cleavage of the
chemical bond.[11,26] Apparently, the fragmentation reaction to
produce a radical by homolytic cleavage is a thermodynamically
disfavored process, since the production and the lostfragment are
all radicals. However, for compounds 3 and 4, their base peaks are
allionsC1producedbylosingasubstituteradical(viapathc), while
ions C1 are secondary intense peak in the ESI-MS/MS spectrum for
compounds 1, 2 and 5 (as shown in Tables 1 and 2). In addition, the
activation barrier from ions A to C1 is a little higher than the other
fragmentation pathways as shown in Scheme 2, except comparing
to the path b for compound 1. In the previous study of our group,
the amides and esters were found to undergo hydrolysis reaction
catalyzed by amino-group during ion trap mass analysis in ESI
conditions.[28] Other studies have also reported in-trap hydrolysis,
forexample,thethioesteranalogsofthealkenyldiarylmethanes,[29]
the esters of some alkenyldiarylmethanes and benzophenones[30]
and the trimethylsilyl esters of long-chain fatty acids.[31] Prompted
by the mechanism of the hydrolysis reactions in ion trap and
the mechanism of CI, which produces ions by gas-phase ion-
molecule reactions, we speculated the lost substitute radical RÐ
might collide with ions C1, involving elimination of molecule
R–R. First, the deprotonated molecule (ion A) gains enough
energy from the CID process and eliminates a substitute radical
R. Then, the substitute radical R reacts with ions C1 to eliminate
dual-substitute molecule (R-R) to form ions C2. This continuous
process is a radical-driven process. By theoretical calculations, the
first step is a thermodynamically disfavored process (ꢀG > 0,
1: 55.51 kcal/M; 4: 50.18 kcal/M). However, the second step
is a thermodynamically more favored process (ꢀG < 0, 1:
ꢁ43.01 kcal/M; 4: ꢁ36.48 kcal/M). It means that the lost substitute
radical R can drive product ion C1 losing dual-substitute molecule
(R-R) to form product ion C2, which has a more stable conjugative
structure.
Conclusions
In summary, ESI-MSn spectra of eight cyclic dipeptides in negative
mode were investigated and the fragmentation pathways were
elucidated. Under low-energy CID conditions, the deprotonated
cyclic dipeptides tend to eliminate an unsaturated substituent
molecule (R ꢁ H), a substituent radical, successively eliminate
a CO and a saturated substituent molecule (R C H). Using
(H/D) exchange experiments and combining with the high-
resolution MS, it is first found that in negative-ion ESI mode,
the cyclic dipeptides with alkyl substituents tend to eliminate their
substituent radical to form radical ions C1 and C2. At the same
time, the cyclic dipeptides with aromatic substituents (Trp, Tyr)
tend to eliminate the substituent molecule (R ꢁ H) to form ions B1
and B2. Therefore, the characteristic side-chain losses can be used
for the identification of the composing amino acids in the cyclic
dipeptides. The results of theoretical calculations successfully
demonstrated that their formation can be rationalized on the
basis of the relative Gibbs free energies. These observations may
have some potential applications in the structural elucidation and
interpretation of mass spectra of homologous compounds.
Acknowledgements
This study was supported by the National Natural Science
Foundation of China (No. 20672104; No. 20732004).
As shown in Scheme 2, the analysis of the relative ꢀG reveals
that, in the case of compounds 1 and 4, the total energy needed
for ion A to produce ion C1 (path c) is lower than that of ion D
(path a). However, the base peak of compound 1 is not ion C1 but
ion D, which is different to compound 4. One of the reasons is that
the lost radical (Rž) of compound4 is a secondary radical and more
stable than the corresponding primary radical of compound 1, so
the relative abundance of the product ion C1 in the spectrum of
compound4(100%)isapparentlyhigherthanthatinthespectrum
of compound 1 (19%). Another reason is that for compound 1,
either producing ion D (via path a) or producing ion C1 (path c)
all needs to go over an activation barrier and then spontaneously
transform to product ion. The energy barrier involved in path
a (ꢀG D 30.21 kcal/M) is apparently lower than that in path c
(ꢀG D 55.51 kcal/M). Therefore, it is not difficult to imagine that
ionAofcompound1isapttoproduceion Dalongpatha, resulting
in ion D as the base peak instead of ion C1, and ion A of compound
4 is apt to produce ion C1 as the base peak along path c.
References
[1] C. Prasad. Bioactive cyclic dipeptides. Peptides 1995, 16, 151.
[2] W. A. Loughlin, R. L. Marshall, A. Carreiro, K. E. Elson. Solution-phase
combinatorial synthesis and evaluation of piperazine-2,5-dione
derivatives. Bioorganic and Medicinal Chemistry Letters 2000, 10, 91.
[3] K. H. Rhee. Cyclic dipeptides exhibit synergistic, broad spectrum
antimicrobial effects and have anti-mutagenic properties.
International Journal of Antimicrobial Agents 2004, 24, 423.
[4] R. Musetti, R. Polizzotto, A. Vecchione, S. Borselli, L. Zulini,
M. D’Ambrosio, L. Sanita` di Toppi, I. Pertot. Antifungal activity
of diketopiperazines extracted from Alternaria alternata against
Plasmopara viticola: An ultrastructural study. Micron 2007, 38, 643.
[5] Y. H. Chen, S. E. Liou, C. C. Chen. Two-step mass spectrometric
approach for the identification of diketopiperazines in chicken
essence. European Food Research and Technology 2004, 218, 589.
[6] S. A. Eriksen, I. S. Fagerson. Nonvolatile nitrogen compounds in
hydrolyzed vegetable protein. In Cereals for Food and Beverages:
Recent Progress in Cereal Chemistry and Technology, G. E. Inglett,
L. Munck (eds). Academic Press: New York, 1980.
For the spectra of compounds 6 and 7, the product ions D and
C1 are not observed in the ESI-MS/MS spectrum. Conversely, the
c
J. Mass. Spectrom. 2009, 44, 1188–1194
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/jms

Y. Guo et al.
[7] M. Ginz, U. H. Engelhardt. Identification of proline-based
diketopiperazines in roasted coffee. Journal of Agricultural and
Food Chemistry 2000, 48, 3528.
low-energy collision-induced dissociation in ESI- and MALDI mass
spectrometry. Journal of the American Society for Mass Spectrometry
2007, 18, 807.
[8] M. Ginz, U. H. Engelhardt. Identification of new diketopiperazines
in roasted coffee. European Food Research and Technology 2001,
213, 8.
[19] Z. Li, T. Yalcin, C. J. Cassady. C-terminal amino acid residue loss for
deprotonated peptide ions containing glutamic acid, aspartic acid,
or serine residues at the C-terminus. Journal of Mass Spectrometry
2006, 41, 939.
[20] E. Pittenauer, M. Zehl, O. Belgacem, E. Raptakis, R. Mistrik,
G. Allmaier. Comparison of CID spectra of singly charged
polypeptide antibiotic precursor ions obtained by positive-ion
vacuum MALDI IT/RTOF and TOF/RTOF, AP-MALDI-IT and ESI-IT
mass spectrometry. Journal of Mass Spectrometry 2006, 41, 421.
[21] J. F. Zhao, T. Shoeib, K. W. M. Siu, A. C. Hopkinson. The fragmen-
tation of protonated tyrosine and iodotyrosines: The effect of
substituents on the losses of NH3 and of H2O and CO. International
Journal of Mass Spectrometry 2006, 255–256, 265.
[9] M. Gautschi, J. P. Schmid, T. L. Peppard, T. P. Ryan, R. M. Tuorto,
X. Yang. Chemical characterization of diketopiperazines in beer.
Journal of Agricultural and Food Chemistry 1997, 45, 3183.
[10] A. E. M. Crotti, T. Fonseca, H. Hong, J. Staunton, S. E. Galembeck,
N. P. Lopes, P. J. Gates. The fragmentation mechanism of five-
membered lactones by electrospray ionisation tandem mass
spectrometry. International Journal of Mass Spectrometry 2004,
232, 271.
[11] N. A. J. C. Furtado, R. Vessecchi, J. C. Tomaz, S. E. Galembeck,
J. K. Bastos, N. P. Lopes, A. E. M. Crotti. Fragmentation of
diketopiperazines from Aspergillus fumigatus by electrospray
ionization tandem mass spectrometry (ESI-MS/MS). Journal of Mass
Spectrometry 2007, 42, 1279.
[12] C. S. Brinkworth, S. Dua, A. M. McAnoy, J. H. Bowie. Negative ion
fragmentations of deprotonated peptides: backbone cleavages
directed through both Asp and Glu. Rapid Communications in Mass
Spectrometry 2001, 15, 1965.
[13] A. G. Harrison. Effect of phenylalanine on the fragmentation of
deprotonated peptides. Journal of the American Society for Mass
Spectrometry 2002, 13, 1242.
[14] A. M. Bradford, R. J. Waugh, J. H. Bowie. Characterization of
underivatized tetrapeptides by negative-ion fast-atom
bombardment mass spectrometry. Rapid Communications in Mass
Spectrometry 1995, 9, 677.
[15] J. H. Bowie, C. S. Brinkworth, S. Dua. Collision-induced fragmenta-
tions of the (M-H)ꢁ parent anions of underivatized peptides: an
aid to structure determination and some unusual negative ion
cleavages. Mass Spectrometry Reviews 2002, 21, 87.
[16] P. A. Wabnitz, R. J. Waugh, M. A. Eckersley, S. Dua, T. Blumenthal,
J. H. Bowie. The negative ion mass spectra of deprotonated 2,5-
diketopiperazines. International Journal of Mass Spectrometry and
Ion Processes 1996, 154, 193.
[17] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, J. A. Montgomery, T. Vreven Jr, K. N. Kudin,
J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone,
B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson,
H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene,
X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Vot,
P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich,
A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, AG. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-
Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill,
B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, J. A. Pople. Gaussian
03, Revision C. 02. Gaussian, Inc.: Wallingford, CT, 2004.
[22] A. H. El, G. Orlova, A. C. Hopkinson, K. W. M. Siu. Gas-phase
fragmentation reactions of protonated aromatic amino acids:
concomitant and consecutive neutral eliminations and radical
cation formations. Journal of Physical Chemistry. A 2004, 108, 3844.
[23] P. Hvelplund,
S. B. Nielsen,
M. Sørensen,
J. U. Andersen,
T. J. D. Jørgensen. Electron loss from multiply protonated
lysozyme ions in high energy collisions with molecular oxygen.
Journal of the American Society for Mass Spectrometry 2001, 12, 889.
[24] K. Whitehead, J. I. Hedges. Electrospray ionization tandem
mass spectrometric and electron impact mass spectrometric
characterization of mycosporine-like amino acids. Rapid
Communications in Mass Spectrometry 2003, 17, 2133.
[25] K. H. M. Cardozo, V. M. Carvalho, E. Pinto, P. Colepicolo. Fragmen-
tation of mycosporine-like amino acids by hydrogen/deuterium
exchange and electrospray ionisation tandem mass spectrometry.
Rapid Communications in Mass Spectrometry 2006, 20, 253.
[26] K. H. M. Cardozo, R. Vessecchi, V. M. Carvalho, E. Pinto, P. J. Gates,
P. Colepicolo, S. E. Galembeck, N. P. Lopes. A theoretical and mass
spectrometrystudyofthefragmentationofmycosporine-likeamino
acids. International Journal of Mass Spectrometry 2008, 273, 11.
[27] C. N. W. Lam, I. K. Chu. Formation of anionic peptide radicals In
vacuo. Journal of the American Society for Mass Spectrometry 2006,
17, 1249.
[28] X. L. Yang, C. J. Zhu, S. X. Cao, X. C. Liao, Y. F. Jiang, Y. F. Zhao.
Amino-catalyzed hydrolysis of amides and esters in the
fragmentation by electrospray ionization tandem mass
spectrometry using an ion trap. Rapid Communications in Mass
Spectrometry 2003, 17, 1927.
[29] M. Cullen, D. Miles, A. Rothwell, C. Bonham, K. V. Wood,
M. Cushman. Hydrolysis of thioesters in an ion trap. Rapid
Communications in Mass Spectrometry 2008, 22, 1094.
[30] M. A. Silvestri, D. Miles, A. P. Rothwell, K. V. Wood, M. Cushman. The
‘apparent’ hydrolysis of alkyl esters during electrospray ionization.
Rapid Communications in Mass Spectrometry 2003, 17, 1703.
[31] K. V. Wood, C. C. Bonham, M. A. Jenks. The effect of water on the ion
trap analysis of trimethylsilyl derivatives of long-chain fatty acids
and alcohols. Rapid Communications in Mass Spectrometry 2001, 15,
873.
[18] H. Y. Yin, A. Chacon, N. A. Porter, D. S. Masterson. Free radi-
cal–induced site-specific peptide cleavage in the gas phase:
c
www.interscience.wiley.com/journal/jms
Copyright ꢀ 2009 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 1188–1194