In-source CID of guanosine: gas phase ion-molecule reactions.

Abstract

In-source collision induced dissociation was applied to access second generation ions of protonated guanosine. The in-source gas-phase behavior of [BH2]+-NH3 (m/z 135, C5H3N4O+) was investigated. Adduct formation and reactions with available solvent molecules (H2O and CH3OH) were demonstrated. Several addition/elimination sequences were observed for this particular ion and solvent molecules. Dissociation pathways for the newly formed ions were developed using a QqTOF mass spectrometer, permitting the assignment of elemental compositions of all product ions produced. Reaction schemes were suggested arising from the ring-opened intermediate of the protonated base moiety [BH2]+, obtained from fragmentation of guanosine. The mass spectral data revealed that the in-source CH3OH-reaction product underwent more complex fragmentations than the comparable ion following reaction with H2O. A rearrangement and a parallel radical dissociation pathway were discerned. Apart from the mass spectrometric evidence, the fragmentation schemes are supported by density functional theory calculations, in which the reaction of the ring-opened protonated guanine intermediate with CH3OH and a number of subsequent fragmentations were elaborated. Additionally, an in-source transition from the ring-opened intermediate of protonated guanine to the ring-opened intermediate of protonated xanthine was suggested. For comparison, a low-energy collision induced dissociation study of xanthosine was performed. Its dissociation pathways agreed with our assumption.

Full text

Page 1

In-Source CID of Guanosine: Gas Phase
Ion-Molecule Reactions
Robin Tuytten, Filip Lemière, and Eddy L. Esmans
Department of Chemistry, Nucleoside Research and Mass Spectrometry Unit and Center for Proteomics
and Mass Spectrometry, University of Antwerp, Antwerp, Belgium
Wouter A. Herrebout and Benjamin J. van der Veken
Department of Chemistry, Cryospectroscopy, University of Antwerp, Antwerp, Belgium
Ed Dudley and Russell P. Newton
Biochemistry Group, School of Biological Sciences and Biomolecular Analysis Mass Spectrometry Facility,
University of Wales, Swansea, United Kingdom
Erwin Witters
Department of Biology, Laboratory for Plant Biochemistry and Center for Mass Spectrometry
and Proteomics, University of Antwerp, Antwerp, Belgium
In-source collision induced dissociation was applied to access second generation ions of
protonated guanosine. The in-source gas-phase behavior of [BH2]?-NH3(m/z 135, C5H3N4O?)
was investigated. Adduct formation and reactions with available solvent molecules (H2O and
CH3OH) were demonstrated. Several addition/elimination sequences were observed for this
particular ion and solvent molecules. Dissociation pathways for the newly formed ions were
developed using a QqTOF mass spectrometer, permitting the assignment of elemental
compositions of all product ions produced. Reaction schemes were suggested arising from the
ring-opened intermediate of the protonated base moiety [BH2]?, obtained from fragmentation
of guanosine. The mass spectral data revealed that the in-source CH3OH-reaction product
underwent more complex fragmentations than the comparable ion following reaction with
H2O. A rearrangement and a parallel radical dissociation pathway were discerned. Apart from
the mass spectrometric evidence, the fragmentation schemes are supported by density
functional theory calculations, in which the reaction of the ring-opened protonated guanine
intermediate with CH3OH and a number of subsequent fragmentations were elaborated.
Additionally, an in-source transition from the ring-opened intermediate of protonated guanine
to the ring-opened intermediate of protonated xanthine was suggested. For comparison, a
low-energy collision induced dissociation study of xanthosine was performed. Its dissociation
pathways agreed with our assumption. (J Am Soc Mass Spectrom 2006, 17, 1050–1062) © 2006
American Society for Mass Spectrometry
D
(QIT) instruments. The latter’s capabilities of accessing
MSnproduct ions by applying repeated stages of mass
selection and fragmentation proved to be very useful
[1]. The maturation of the time of flight (TOF) technol-
ogy and, more importantly, its use in hybrid-arrange-
ments such as the popular orthogonal QqTOF mass
spectrometer made accurate mass data of product ions
more readily available [2]. In many cases, the elemental
compositions of the product ions can unambiguously be
uring the last decade, most fragmentation path-
ways were elaborated on triple quadrupole
(QqQ) and particularly quadrupole ion-trap
assigned when using a QqTOF mass spectrometer.
Product ions lower in the genealogical ladder are also
accessible by applying relatively high collision energies
in the collision cell. Yet disentanglement of different
dissociation pathways originating from a communal
progenitor is —in general—still confined to QIT set ups,
although the interpretation of energy resolved spectra
also has proven its usefulness [3].
Recently we and others showed [4 – 6] that by mak-
ing optimal use of in-source CID, combined with the
inherent tandem MS capabilities of the QqTOF, what is
in effect up to MS5data can be acquired. Compared to
QIT data the QqTOF data obtained is superior in respect
of higher mass accuracy. On the other hand QIT-CID
experiments, in general, access almost exclusively the
dissociation reaction channels of the lowest energy of
activation [7]. This energy-dependent information can
Published online June 5, 2006
Address reprint requests to Professor F. Lemière, Department of Chemistry,
Nucleoside Research and Mass Spectrometry Unit and Center for Proteom-
ics and Mass Spectrometry, University of Antwerp, Groenenborgelaan 171,
B-2020 Antwerp, Belgium. E-mail: filip.lemiere@ua.ac.be
© 2006 American Society for Mass Spectrometry. Published by Elsevier Inc.
1044-0305/06/$32.00
doi:10.1016/j.jasms.2006.03.012
Received October 11, 2005
Revised March 24, 2006
Accepted March 24, 2006

Page 2

pass unrecognized in QqTOF tandem MS experiments
due to a less controlled energy deposit on the parent
ions. However, in QqTOF experiments fragmentation
pathways with higher activation energies become
within reach due to higher energy-transfer in a single
collision at high CE settings [6, 7].
During an in-depth study of the fragmentation be-
havior of the nucleoside guanosine with the optimized
in-source CID-QqTOF MS approach it became clear that
some of its product ions have a high affinity for H2O.
Additionally, some other product ions arose, which
suggested even more sophisticated ion-molecule reac-
tions [5]. It has to be noted that in-source addition of
solvent molecules has been observed before by Gabelica
et al. for benzylpyridinium cations [8].
This work focuses on these gas-phase ion-molecule
interactions and reactions of one particular product ion
of protonated guanosine. By in-source CID of the pro-
tonated guanosine the second generation product ion
C5H3N4O? at m/z 135 (Figure 1), was produced [5, 9]. Its
in-source formed adducts with H2
subsequently submitted to tandem MS analysis by
QqTOF MS and accurate mass data of the according
product ions was acquired. Different neutral gain/
neutral loss sequences involving H2O and CH3OH were
distinguished.
Based on the product ions, an in-source formation of
ring-opened xanthine was predicted, emanating from
the reaction with H2O of the ring opened protonated
guanine. To support this assumption the complete
low-energy CID fragmentation study of xanthosine was
elaborated.
The product ion data of the CH3OH adduct indicated
that a reaction of the same ring opened protonated
guanine with CH3OH occurred. Product ions were
observed resulting from the transfer of the methyl
group from the methanol to the heterocyclic base. The
tentative reaction- and fragmentation schemes deduced
18O and CH3OH were
H+
NH
C
N
C
C
C
HC
N
O
NH2
N
CC
C
O
C
OHOH
HH
H
CH2
H
HO
H+
C
N
C
C
C
HC
N
O
+NH2
N
NH
C
C
C
HC
N
O
NH2
NH
CH
NH
C
C
N
C
NH2
NH
C
NH2
C10H14N5O5+
Exact Mass: 284.0995
C5H6N5O+
Exact Mass: 152.0572
C5H3N4O+
Exact Mass: 135.0307
C4H4N3O+
Exact Mass: 110.0354
C4H5N4+
Exact Mass: 109.0514
CH3N2+
Exact Mass: 43.0296
-NH3
-NHCNH
-HNCO
C6H6N5O+
exact mass: 152.0572
Ring-opened intermediate
Interactions
with CH3OH
Interactions and/or reactions
with H218O
C5H5N4
Exact Mass: 155.0455
16O18O+
C6H7N4O2
Exact Mass: 167.0569
+
+CH3OH
-NH3
+H218O
-NH3
N(1)
C(2)
N(3)
C(4)
C(5)
C(6)
C(8)
N(7)
O(6)
N(2)
N(9)
H
H
H
H
H
NH2
C
N
C
C
C
HC
N
O
NH2
N
H
Figure 1.
product ions, together with the ions at m/z 155 and 167, respectively, formed upon interaction/
reaction with H2
adjusted CV-values.
The main dissociation pathways of protonated guanosine up to the second generation
18O or CH3OH. For this study the latter ions were generated in-source by applying
1051
J Am Soc Mass Spectrom 2006, 17, 1050–1062IN-SOURCE CID OF GUANOSINE

Page 3

from the mass spectral data were confirmed by quan-
tumchemical calculations in which theoretical informa-
tion on the anticipated gas-phase reaction and the
ensuing fragmentation pathways were obtained at the
DFT level. To this end, equilibrium geometries were
obtained starting from the tentative product ion struc-
tures, and the reaction energies ?Erbetween the re-
agents/precursor ions and products/fragments were
determined. In addition, for all reactions studied, the
transition states were localized and the corresponding
activation energies were calculated. The results illus-
trated that for all reactions studied, the theoretical data
compared favorably with the tentative schemes. In
addition, the DFT calculations provided additional in-
sights with respect to proton positions and the proton
transfers.
Experimental
Chemicals
Guanosine, xanthosine, and N7-methylguanosine were
purchased from Sigma (Bornem, Belgium).18O-labeled
water, H2
Isotope Laboratories, Inc. (Adover, MA). Methanol,
CH3OH, (HPLC grade), and phosphoric acid, H3PO4,
were acquired from Acros Organics (Geel, Belgium).
H2O [when the isotope label for oxygen is not men-
tioned, it is adopted to be
purchased from Fisher Scientific (Loughborough, Lei-
cestershire, UK). Formic acid, HCOOH, (99–100%; ul-
trapure) was obtained from VWR International (Leu-
ven, Belgium).
1 mL stock solutions of 10?3M of each nucleoside
were prepared in 50/50 (vol/vol) CH3OH/H2O or
CH3OH/H2
were obtained by diluting the stocks further to 10?5M in
50/49.9/0.1 (vol/vol) CH3OH/H2O/HCOOH or 50/
49.9/0.1 (vol/vol) CH3OH/H2
purpose, a 10% formic acid solution in H2O or H2
was used.
18O (18O, 95%) was obtained from Cambridge
16O] (HPLC grade) was
18O (only guanosine). Working solutions
18O/HCOOH. For this
18O
Instrumentation and MS Conditions
A Q-TOF II mass spectrometer (Waters, Manchester,
UK) was equipped with a standard pneumatically-
assisted electrospray probe and a Z-spray source (Wa-
ters). Electrospray mass spectra were recorded in the
positive ion mode. The applied experimental settings
were: spray voltage 3.5 kV, source temperature 80 °C,
and desolvation temperature 100 °C. A cone gas flow-
rate of ca. 60 L/h and a desolvation gas flow-rate of ca.
150 L/h were applied. In MS/MS experiments, the
precursor ions were selected with an isolation window
of ?0.5 Da, adequately removing isotopic (13C) interfer-
ences from the spectrum. Scan time and inter-scan time
were, respectively 1.0 s and 0.1 s. For each MS/MS
experiment the nucleoside solution was infused for 50
min at a flow rate of 5 ?L/min via a syringe pump
(Harvard Apparatus, Natick, MA). During the MS/MS
experiment the collision energy was varied in a cyclic
manner (5 eV intervals, ranges cf. Table 1) permitting
the construction of collision energy profiles. Before the
MS/MS experiments of the in-source formed product
ions the cone voltage was optimized for each ion of
interest. Therefore, a nucleoside solution was infused at
5 ?L/min and full scan spectra were recorded (m/z
100–600) with a cone voltage varying between 15 and
90 V (5 V intervals, each 30 s). The optimal cone voltage
(CV) for each ion of interest was selected upon the
evaluation of the extracted ion chromatograms and the
corresponding spectra (cf. isobaric interferences).
For high mass accuracy the QqTOF MS was cali-
brated using 0.1% phosphoric acid in 50:50 (vol/vol)
MeOH/H2O. Any instrument drift was compensated
by applying a lock mass correction. Depending on the
experiment the protonated nucleoside [M ? H]?, the
protonated base [BH2]?, or other product ions were
used as lock mass. In the latter case only these frag-
ments of which the elemental composition was con-
firmed by us and others were utilized. The TOF ana-
lyzer was programmed to record the m/z range 50–400
for the protonated nucleosides and the m/z range 50–
250 for [BH2]?and the other ionic species studied. Only
nominal masses are given in the discussions, accurate
mass data is available in the Supplementary Informa-
tion, (which can be found in the electronic version of
this article).
The data were processed using MassLynx version
3.5. The following criteria were applied to distinguish
“true” product ions from interferences in the spectra: (1)
only ions with intensity above 1% of the base peaks
intensity were included; (2) the product ions must be
present in the spectra for at least two collision energy
values applied to any given precursor ion; (3) the
product ions must yield credible empirical formulas in
respect of the parent ion from which they arose (taking
into account possible adduct formation in the collision
cell as described previously) [5, 10].
Calculations
Density functional theory (DFT) calculations were per-
formed using Gaussian 03 [11], as installed on the
computing cluster CalcUA. Equilibrium geometries and
transitions states were obtained at the B3LYP/6-31G(d)
Table 1.
ionic species subjected to MS/MS
Applied experimental conditions for the different
m/z CV (V)CE-range (eV)
Guo[BH2]?-NH3
[BH2]?-NH3?H2
[BH2]?-NH3?CH3OH
[M?H]?
[BH2]?
[BH2]?-NH3
[BH2]?-HNCO
135
155
167
285
153
136
110
65
65
65
20
30
50
65
5–40
5–40
5–50
5–60
5–50
5–40
5–40
18O
Xan
1052
TUYTTEN ET AL.J Am Soc Mass Spectrom 2006, 17, 1050–1062

Page 4

level. Additional evidence supporting the nature of the
stationary was obtained by calculating the correspond-
ing Hessian, and by carefully analyzing the number of
imaginary vibrational frequencies derived from them.
The Cartesian coordinates, charge distributions, and
vibrational frequencies for the geometries obtained are
available from the authors upon request.
It should be noted that for the reagent/parent ions
and the product ions, computational evidence of donor-
acceptor interactions between the most electrophilic
region of the ions with an electron-rich region of the
incoming (CH3OH) or leaving (H2O, NH3, HOCN, or
CH3NH2) species was found. The appearance of these
intermolecular interactions, evidently, influences the
calculated energies. In this study, the activation ener-
gies and the reaction energies were calculated by com-
paring the energies of the initial and final adducts,
without correcting for the complexation energies
present.
Results and Discussion
H2O- and CH3OH-Adduct Formation of In-Source
CID Product ions
As a starting point of this study a cone voltage (CV)
optimization was performed: the CV was gradually
increased while infusing an acidified guanosine solu-
tion into the ESI-source. Increasing the CV resulted
firstly in a well known break up of the anomeric bond,
leading to [BH2]?. Further CV increase gave rise to the
in-source formation of the initial product ions of the
known [BH2]?-NH3(m/z 135; C5H3N4O?) and [BH2]?-
NHCNH pathways (m/z 110; C4H4N3O?) and the minor
[BH2]?-HNCO (m/z 109; C4H5N4
Earlier experiments, augmented with DFT calculations,
showed that both the m/z 110 and m/z 135 product ions
interacted strongly with H2O present in the source and
with traces of H2O present in the collision cell [5]. In the
latter study, the occurrence of a non-covalent (hydrogen
bonded) adduct m/z 128 (C4H4N3O?·H2O), was derived
from the MS/MS-spectra and supported by the DFT
calculations. However further characterization of the
m/z 153 due to C5H3N4O?·H2O with MS/MS revealed
several new product ions in addition to the known
fragments of m/z 135 [5].
Inspection of the full scan spectra clearly illustrated
that adduct formation with methanol also occurred for
the product ions m/z 110 and m/z 135. In agreement with
earlier observations suggesting that CH3OH is a much
stronger Lewis base than H2O, the formation of molec-
ular complexes with CH3OH is much more pronounced
than with H2O (Figure 2). Thus a significant fraction of
these, by in-source CID produced, second generation
product ions of guanosine was promptly transformed
to ions with higher m/z ratios. The correlation between
the in-source formed product ion at m/z 135 and its
H2
and 167) is clear from Figure 2: with increasing CV
?) pathway (Figure 1).
18O- and CH3OH-adducts (respectively, at m/z 155
parallel appearances of the m/z ratios of [BH2]?-NH3,
[BH2]?-NH3? H2
Moreover, also in the product ion spectra of the in-
source formed ion at m/z 167 (C5H3N4O?? CH3OH),
product ions are seen which cannot be explained solely
by a noncovalent ion-neutral interaction: gas-phase
reactions have to be assumed. In contrast the product
ion spectra of m/z 142, (C4H4N3O?? CH3OH), shows
no other product ions than these corresponding to the
initial ions from m/z 110.
18O, [BH2]?-NH3? CH3OH occur.
C5H5N4
Formation and H2
16O18?: A Result of H2
18O-Gas-Phase Reactions
18O-Adduct
To investigate the phenomena arising from the solvent-
adducts of m/z 135 (C5H3N4O?) it is favorable to replace
H2
at m/z 153 nominally coincides with the natural
isotopic signal of the [BH2]?of guanosine. However, it
has to be recognized that considerable amounts of
H2
ment with the quoted commercial purity of the H2
(18O, 95%) and the atmospheric conditions in the ESI-
source.
Application of a cone voltage of 65 V generated
in-source the ion at m/z 135 together with its H2
adduct at m/z 155 (C5H5N4
135 and 155 were monoisotopically selected by the
quadrupole filter and accurate mass energy-resolved
spectra were collected in the range of 5–40 eV. Exem-
plary product ion spectra of both ions at CE 15 eV and
CE 30 eV are shown in Figure 3. The accurate mass data
confirmed that m/z 155 (C5H5N4
with the addition of H2
The prominent presence of the m/z 135 (C5H3N4
the spectra of m/z 155 at low CE agrees with the
non-covalentnature of
C5H3N4
demonstrated the stability of the H2
155 at CE 15 eV little product ions, apart from m/z 135,
were produced, whereas a fairly extensive fragmenta-
tion was apparent at CE 15 eV in the product ion
spectrum of m/z 135 itself. The typical product ions of
m/z 135 only appeared in the product ion spectra of m/z
155 at higher CEs. The product ions related with the
[BH2]?-NH3 pathway have already been discussed [5,
9], but a variety of new product ions were also observed
in the MS/MS spectra of m/z 155 in the current study.
All the product ions associated with the dissociation of
m/z 155 can be found in the Supplementary Materials
section, which can be found in the online version of this
article.
The product ions of m/z 155 at m/z 138 and 137 could
not be explained by a simple ion-molecule cluster. The
unexpected yet abundant ion at m/z 137 (C5H3N4
is found upon loss of H2
composition of m/z 137 requires that the16O originally
present in the carbonyl group (C6) of the guanine is
16O by H2
18O in the sample, since the H2
16O -adduct
13C-
16O are still present during in-source CID in agree-
18O
18O-
16O18O?). Both the ions at m/z
16O18O?) corresponded
18O to m/z 135 (C5H3N4
16O?).
16O?) in
the interaction between
16O?and H2
18O. At the same time these spectra
18O-adduct: for m/z
18O?)
16O from m/z 155. The elemental
1053
J Am Soc Mass Spectrom 2006, 17, 1050–1062IN-SOURCE CID OF GUANOSINE

Page 5

removed by formation of the carbonyl hydrate (H2
see Scheme 1, and 2,) of m/z 135 in-source (m/z 155) and
dehydration (H2
port for this16O/18O-exchange at C6 was found in the
ion at m/z 95 C4HN2
the product ion m/z 93 (C4HN2
ized fragment of m/z 135 [9], which conserves the initial
oxygen of m/z 135 (18O: m/z 137 and m/z 95). All the
other product ions of m/z 137 coincided with the prod-
uct ions in the original [BH2]?-NH3(m/z 135) pathway
since none of them retains the oxygen atom.
The accurate mass data for m/z 138 pointed to the
elemental composition C5H2N3
with a loss of NH3. Its interrelationship with m/z 155
requires a covalently bounded18O for C5H5N4
and not solely a noncovalent ion-molecule interaction.
Most likely an addition/elimination reaction occurs at
C(2) of the open-ring intermediate of protonated gua-
18O,
16O) in the collision cell. Further sup-
18O?. This ion corresponded with
16O?), a well character-
16O18O?, which agreed
16O18?,
nine structure proposed by Gregson and McCloskey [9].
A nucleophilic attack by H2
cient C(2) can result in a subsequent elimination of
NH3, leading to the ion at m/z 155 with the observed
elemental composition. A second loss of NH3from the
selected ion m/z 155 will result in the product ion at m/z
138.
Alternatively, a loss of HNCO was easily envisioned
from the newly formed ion at m/z 155: product ions at
m/z 110 and 112, depending on the position of the
18O-label, were anticipated. Indeed, an ion (C4H4N3O?)
was detected as a low abundant ion at m/z 112 (18O).
The16O analogue at m/z 110 was not seen since it was
obscured by a new abundant ion as explained below.
The structure resulting from the loss of HNCO starting
from m/z 155 was thought to be the same as that
emanating from [BH2]?-NHCNH in the fragmentation
of guanine [9]. This was confirmed by the presence of
18O onto the electron defi-
Figure 2.
([BH2]?-NH3) and the latter’s H2
(CV; CV increased 5 V every 30 s). Bottom: The accumulated spectrum of the 4.5–5.5 min time span.
Top: The extracted ion profiles corresponding to the [M ? H]?, [BH2]?, the m/z 135
18O and CH3OH-adducted ions in function of increasing cone voltage
1054
TUYTTEN ET AL. J Am Soc Mass Spectrom 2006, 17, 1050–1062

Page 6

product ions known from the [BH2]?-NHCNH frag-
mentation pathway resulting from two consecutive
losses of HCN, at m/z 85 C3H3N2
C2H2N18O?. The corresponding16O ions were not de-
tected either due to overlapping signals or low signal
intensity. Loss of carbon monoxide (C18/16O) from m/z
112/110 results in the known ion at m/z 82 which
further loses HCN.
Clearly by in-source addition of H2O (H2
[BH2]?-NH3ion and the subsequent loss of HNCO from
m/z 153 (m/z 155), a blending in of the two major known
fragmentation pathways ([BH2]?-NH3(m/z 135) and the
[BH2]?-HNCNH (m/z 110)) occurs, that are normally
separate for the fragmentation of protonated Gua [5, 9].
In Scheme 2 the other product ions arising from m/z
138 are given. The expulsion of H16/18OCN leads to the
pair of product ions at m/z 93/95, with the same
elemental compositions as the product ions derived
from the ions at m/z 135 and m/z 137. An additional loss
of carbon monoxide (C16/18O) from the ions at m/z
93/95 also produces an ion at m/z 65 (C3HN2
again was common to a product ion appearing in the
[BH2]?-NH3pathway. Evidence to support these con-
vergent pathways will follow further. At the same time
a second dissociation pathway developed from the ion
at m/z 138 (C5H2N3
monoxide (18/16O) led to the ions at m/z 110 and 108
(very low abundance). From these ions a subsequent
18O?and m/z 58
18O) to the
?) which
16O18O?). An initial loss of carbon
loss of HCN was also observed in the energy resolved
spectra giving rise to the unique ions at m/z 83 and at
m/z 81.
C6H7N4O2
Formation and CH3OH-Gas-Phase Reactions
?: A Result of CH3OH-Adduct
As mentioned previously and indicated in Figure 2 a
significant fraction of the in-source generated ion at m/z
135, ensuing in-source CID of protonated guanosine,
was converted into its CH3OH-adduct (m/z 167). It was
found interesting to verify if neutral gain/neutral loss
sequences similar to those observed for the H2O-adduct
would occur as well for this CH3OH-adduct. Infusing a
Guo solution under the application of a CV of 65V
produced high levels of the ion at m/z 167 in-source of
which energy resolved spectra were recorded. As there
was no isobaric interference there was no need to use
isotopic labeled CH3OH. In Figure 4 the accumulated
product ion spectra of m/z 167 at CE 15 eV and 30 eV are
shown; additionally the accurate mass data and elemen-
tal compositions of all product ions are available in the
Supplementary Information. The accurate mass mea-
surement of m/z 167 (C6H7N4O2
tion of CH3OH to C5H3N4O?.
Noncovalent interaction of methanol with the in-
source formed fragment at m/z 135 was indicated by the
?) confirmed the addi-
Figure 3.
protonated guanosine (CV 65 V and CE 15 eV (upper) and 30 eV (lower); no lock mass correction,
uncentered). Right: Product ion spectra of the in-source generated ion at m/z 155 (C5H4N3O?? H2
outgoing protonated guanosine (CV 65 V and CE 15 eV (upper) and 30 eV (lower); no lock mass
correction, uncentered).
Left: Product ion spectra of the in-source generated ion at m/z 135 (C5H4N3O?) outgoing
18O)
1055
J Am Soc Mass Spectrom 2006, 17, 1050–1062IN-SOURCE CID OF GUANOSINE

Page 7

high abundance of the latter in the product ion spectra
of m/z 167 at low CE. With increased CE, the typical
product ions of the [BH2]?-NH3dissociation pathway
also appeared. The high affinity of m/z 135 for H2O is
seen in the product ion observed at m/z 153, the result of
scavenging traces of H2
by the product ion at m/z 135 as reported earlier [5].
In analogy with the elimination of NH3and HNCO
from the H2
m/z 167 showed the loss of NH2CH3(m/z 136) and
CH3OCN (m/z 110), and corresponding product ions
derived thereof (Scheme 3). The presence of these ions
suggested that also for the adducts with methanol, an
addition (CH3OH)/elimination (NH3) reaction depart-
ing from the ring-opened protonated guanine occurred
similar to that observed for H2
obtained from DFT calculations, in which the reagents
and the reactions products are studied, confirmed this
reaction pathway. The equilibrium geometry of the ring
opened intermediate of protonated guanine (m/z 152) of
the product ion (m/z 167) and of typical product ions
resulting from further fragmentation of m/z 167 are
summarized in Scheme 4. The transition states involved
and the corresponding values for Eaand ?Erare col-
lected in Table 2.
16O available in the collision cell
18O-adduct m/z 135, the methanol adduct
18O. Theoretical data
The geometry optimizations of protonated ring-
opened guanine suggested that the positive charge
(proton) is preferentially located at the cyanamide moi-
ety rather than at the carbonyl at the (6) position [9]. The
DFT results revealed that the addition/elimination re-
actions are preceded by the formation of a donor:
acceptor complex in which an incoming methanol mol-
ecule interactswiththe
Subsequently, a concerted reaction takes place: covalent
bonding of the oxygen of CH3OH at C(2) proton trans-
fer from CH3O*H to N(1)/(2)N and formation of NH3
proceeds in a single step. The transition-state for the
concerted reaction, TS1, is shown in Table 2. The
corresponding values for ?Erand Eaare ?1.7 and 223.6
kJ mol?1. The activation (Ea) and reaction energies (?Er)
(Table 2) are similar to those obtained for proton
transfers [12] and collision induced dissociations [13] in
the product ions formation of other nucleosides. The
experimental settings, i.e., the collision energy, allow
for the calculated transitions.
As the methanol-product was formed with a higher
abundance, the product ions m/z 136 and 110 as well as
their fragmentation products were rather easily identi-
fied. The loss of methylamine leading to m/z 136 re-
quires a rearrangement of the ion, in which the methyl
ring opened guanine.
C10H14N516O5+
Exact mass:
284.0995
[M+H]+
C6H6N516O+
Exact mass:
152.0572
[BH2]+
C6H6N516O+
Exact mass: 152.0572
Ring-opened intermediate
-Ribose
N
C
N
C
C
C
HC
N
16O
N
-NH3
C5H3N416O+
Exact Mass: 135.0307
N
C
N
C
C
C
HC
N
O
N
HH
HH
18O
H
H
C5H5N416O18O+
Exact Mass: 155.0455
Non-covalent H218O-adduct
N
C
N
C
C
C
HC
N
18O
N
HH
18O
C
N
H
C
C
C
HC
N
16O
NH2
N
H
Interaction with H218O
+H218O
-H216O
Reaction
Interaction with H216O
N
C
N
C
C
C
HC
N
18O
N
HH
16O
H
H
C5H5N416O18O+
Exact Mass: 155.0455
Non-covalent H216O-adduct
and O(6) exchange
+H218O
-NH3
Reaction
C5H5N416O18O+
Exact Mass: 155.0455
Addition H218O / elimination NH3
at C(2) of intermediate
+H216O
-NH3
16O
C
N
H
C
C
C
HC
N
16O
NH2
N
H
Reaction
+H218O
-H216O
Reaction
16O
C
N
H
C
C
C
HC
N
18O
NH2
N
H
C5H5N416O18O+
Exact Mass: 155.0455
Addition H216O / elimination NH3
at C(2) and O(6) exchange
m/z 155
A
B
C
D
NH2
C
N
C
C
C
HC
N
16O
NH2
N
H
Scheme 1
1056
TUYTTEN ET AL. J Am Soc Mass Spectrom 2006, 17, 1050–1062

Page 8

Scheme 2
1057
J Am Soc Mass Spectrom 2006, 17, 1050–1062IN-SOURCE CID OF GUANOSINE

Page 9

group is transferred from the methanol oxygen to the
nitrogen N(1)/(2)N (cf) randomization of N(2)/(2)N
subsequent ring-opening [9] (Scheme 3). The occurrence
of this transfer is confirmed by the DFT calculations
(Scheme 4 and Table 2).
Apart from the features described above, new, less
abundant product ions appeared at m/z 154 C5H4N3O3
m/z 152 C5H4N4O2
C6H5N4O?and m/z 124 C5H6N3O?. In agreement with
the results reported earlier [5], the ion m/z 154
C5H4N3O3
collision cell to m/z 136. The ions m/z 149 and 150 were
assigned to species produced after elimination of H2O,
NH3, and HOCN from the m/z 167 ion.
The latter experimental data are easily rationalized
by the obtained equilibrium geometries and transition
states derived from the DFT calculations. The elimina-
tion of H2O, NH3, and HOCN develops via a unique
intermediate (VI), which emanates from V by a methyl
transfer from the methanol oxygen to the N(9)position
of the original guanine (TS4), followed by a proton
transfer (TS5) from VI to VII. Literature and our own
(unpublished) mass spectral data further supported the
presence of the methyl group on the imidazole ring:
product ions at m/z 149 and 124 showed to be distinctive
for a methyl group positioned on a nitrogen of the
imidazole part of a purine (e.g., N7- and N9-methyl-
guanine) [14]. The further eliminations of H2O and NH3
?,
?·, m/z 150 C6H5N3O2
?,m/z 149
?resulted from the addition of H2O in the
Figure 4.
m/z 167 (C5H4N3O?? CH3OH) outgoing protonated guanosine
(CV 65 V and CE 15 eV (upper) and 30 eV (lower); no lock mass
correction, uncentered).
Product ion spectra of the in-source generated ion at
NH2
C
N
C
C
C
HC
N
O
Ring opened intermediate of [BH2]+
NH2
N
H
+CH3OH
-NH3
H2N
C
N
C
C
C
HC
N
O
O
N
H
CH3
C5H6N5O+
Exact Mass: 152.0572
Reaction
C6H7N4O2
Exact Mass: 167.0569
+
HN
C
N
C
C
C
HC
N
O
O
N
CH3
H
H
-HNCOCH3
N
C
C
C
HC
N
O
N
H
H
H+
C4H4N3O+
Exact Mass: 110.0354
NH2
C
N
C
C
C
HC
N
O
O
N
H
H3C
H2N
C
N
C
C
C
HC
N
O
O
N
H
methyl transfer
H3C
-H2O
-NH3
NH
C
N
C
C
C
HC
N
O
N
H3C
C6H5N4O+
Exact Mass: 149.0463
C
O+
N
C
C
C
HC
N
O
N
H3C
C6H4N3O2
Exact Mass: 150.0304
+
-HOCN
NH2
C5H6N3O+
+
C
C
C
HC
N
O
N
H3C
H+
Exact Mass: 124.0511
-CH3
C5H4N4O2
Exact Mass: 152.0334
•+
-NH2CH3
C
N
C
C
C
HC
N
O
+O
N
H
C5H2N3O2
Exact Mass: 136.0147
+
O
C
N
C
C
C
HC
N
O
NH2
NH+
H3C
proton transfer
C4H6N3
+
Exact Mass: 96,0562
-CO
Scheme 3
1058
TUYTTEN ET AL.J Am Soc Mass Spectrom 2006, 17, 1050–1062

Page 10

from VII proceed via a single step, involving either TS7
or TS6. In contrast, the elimination of HOCN requires
for two different subsequent proton transfers: a first
transfer from VII to X through TS8 and a second
transfer from X to XI through TS9 resulting finally in
the expulsion of HOCN. An additional product ion at
m/z 96 (C4H6N3
the ion at m/z 124.
Interestingly there also appeared a parallel dissocia-
tion pathway comprising only radical product ions in
the energy resolved spectra of m/z 167. This series of
radical cations starts off with the loss of a methyl radical
from the ion at m/z 167 to m/z 152 C5H4N4O2
different pathways evolved from there, starting with
the loss of H2O yielding the ion at m/z 134 (C5H2N4O?·)
or the loss of HNCO leading to the ion at m/z 109
(C4H3N3O?·), with the latter being the most abundant.
From these two product ions multiple losses of HCN
and/or CO showed, as commonly seen in the dissocia-
tion pathways of nucleosides [5, 6, 9, 15, 16]. It has to be
noted that once more an addition of H2O in the collision
cell was observed; the radical ion at m/z 109 formed an
adduct with H2O yielding the radical ion at m/z 127
(C4H5N3O2
nominal masses that coincide with ions of other disso-
ciation pathways (e.g., C4H3N4
?) corresponded an extra loss of CO from
?. Two
?·). Many of these radical product ions have
?calc. m/z 107.0358,
+CH3OH
-NH3
methyl transfer-H2O
-NH3
-HOCN
proton transfer
TS1
-NH2CH3
proton transfer
TS2
TS3
TS4
TS5
TS6
TS8
TS9
proton transfer
C5H6N5O+
Exact Mass: 152.0572
C6H7N4O2
Exact Mass: 167.0569
+
C6H7N4O2
Exact Mass: 167.0569
+
C6H7N4O2+
Exact Mass: 167.0569
C6H7N4O2
Exact Mass: 167.0569
+
C6H7N4O2
Exact Mass: 167.0569
+
C6H7N4O2
Exact Mass: 167.0569
+
C5H2N3O2
Exact Mass: 136.0147
+
C6H5N4O+
Exact Mass: 149.0463
C6H4N3O2
Exact Mass: 150.0304
+
C5H6N3O+
Exact Mass: 124.0511
methyl transfer
TS7
V
VII
VI
II
I
IV
III
IX
X
VIII
XI
Scheme 4
Table 2.
the fragmentation of m/z 167 as presented in Scheme 4. Energies
are given in kJ mol?1
The optimized B3LYP/6-31G(d) transition states of
1059
J Am Soc Mass Spectrom 2006, 17, 1050–1062IN-SOURCE CID OF GUANOSINE

Page 11

meas. m/z 107.0362 and C4HN3O·?calc. m/z 107.0120,
meas. m/z 107.0127). They could only be discerned
because of the sufficient resolving power and accuracy
of the QqTOF MS.
In-Source Generation of the Ring-Opened
Protonated Xanthine
The hydrolytic deamination of guanine to xanthine is
well documented: an enzyme guanine deaminase catal-
yses the addition of H2O and elimination of NH3 [17].
There is a clear similarity with the above described
phenomena in the product ion spectra of guanosine. To
probe this similarity, an in-depth CID study of the
nucleoside xanthosine was performed by means of
combined in-source and in-collision cell CID by using
the QqTOF mass spectrometer. To the best of our
knowledge, no extensive ES-MS fragmentation study of
xanthosine has been carried out before.
Initially the energy resolved spectra of protonated
xanthosine [M ? H]?were collected between CE 5–60
eV and based on the accurate mass data elemental
compositions assigned (cf. Supplementary Data). Akin
to the majority of nucleosides the initial energy deposit
caused a break of the anomeric bond. Though unlike
other nucleosides an important part of the positive
charge was channelled into the sugar-moiety leading to
the product ion S?at m/z 133 (C5H9O4
ciation of this sugar-moiety led to a variety of ions. A
plethora of losses of different small neutrals (i.e., water,
carbon monoxide, formaldehyde, acetaldehyde) was
displayed.
?). Further disso-
More important to this study was the simultaneous
generation of the [BH2]?ion at m/z 153, essentially
being protonated xanthine (C5H5N4O2
CV of 30 V the protonated base moiety (m/z 153) was
formed in-source and its product ion spectra recorded.
In Figure 5, the product ion spectra of the protonated
xanthine base at CE 15 eV and 30 eV are shown. The
protonated xanthine primarily dissociated into the ions
at m/z 136 (C5H2N3O2
minor fragment at m/z 135 (C5H3N4O?). These frag-
ments could be rationalized via a ring-opened interme-
diate of protonated xanthine in analogy to the ring
opened protonated guanine (Scheme 2). This ring
opened intermediate (m/z 153) is identical to the struc-
ture following the H2O-addition/NH3-elimination on
the ring-opened intermediate of protonated guanine.
In-source generation of the MS3-like product ions at m/z
110 and 136 (respective CVs: 55 V and 50 V) and their
subsequent product ions in collision cell CID study
supported this conclusion. The plots of the energy
resolved spectra of the product ions related to m/z 110
from protonated xanthine matched exactly the plots of
the m/z 110 originating from protonated guanine (not
shown). The further dissociation of the xanthine-frag-
ment at m/z 136 coincided with the represented path-
way (Scheme 2).
In Table 3 a summary of all relevant product ions of
the discussed spectra is given, which stresses once more
the interrelations between the product ions of [BH2]?,
[BH2]?-NH3, and [BH2]?-HNCO of protonated xantho-
sine and the [BH2]?-NH3, [BH2]?-NH3?H2
[BH2]?-NH3?CH3OH of protonated guanosine. The
absence of some of the predicted product ions in the
corresponding product ion spectra of m/z 135 ? H2
(m/z 155, C5H5N4
dances. Clearly, the m/z 155 ion population was rather
limited. Moreover, the signal of m/z 155 was composed
of several ionic structures and adducts, all fragmenting
differently.
?). By applying a
?) and m/z 110 (C4H4N3O?) and a
(18)O and
18O
16O18?) is attributed to too low abun-
Conclusions
This study demonstrated that in-source CID and prod-
uct ion spectra thereof must be handled with care, due
to the creation of various solvent dependent ionic
species. Initial fragment information might be lost as
some of the in-source generated product ions might—to
greater or lesser extent—be channelled directly into
solvent-adducts before their actual MS/MS analysis.
The reaction of the product ions with methanol and
water occurs both in the electrospray source and, espe-
cially for water, also in the collision cell. Depending on
the experimental settings, product ions resulting from
reaction of guanosine derived ions with H2O and
CH3OH could yield false positive results for the pres-
ence of xanthosine and/or methylated guanosines. Sim-
ilar interactions with solvents might be expected for
example for the nucleoside cytidine yielding uridine
Figure 5.
ion of xanthosine at m/z 153 (C5H5N4O2
(upper) and 30 eV (lower); no lock mass correction, uncentered).
Product ion spectra of the in-source generated [BH2]?
?) (CV 20 V and CE 15 eV
1060
TUYTTEN ET AL.J Am Soc Mass Spectrom 2006, 17, 1050–1062

Page 12

after hydrolytic deamination and methylated uridine
after reaction with methanol.
The unexpected presence of radical cations in the
spectra of these protonated molecules further compli-
cates the obtained product ion pattern. High-resolution
product ion spectra are required to resolve these radical
cations from isobaric even-electron product ions and
correct interpretation of the obtained data (Supplemen-
tary Material Tables 4, 5, and 6).
Acknowledgments
Research funded by a Ph.D. grant of the Institute for the Promo-
tion of Innovation through Science and Technology in Flanders
(IWT-Vlaanderen). This research was also funded by the Flemish
Foundation for Scientific Research (FWO-Vlaanderen), the Flem-
ish Government (GOA), and the University of Antwerp (RAFO).
Financial support allowing the purchase of the computer cluster
CalcUA was obtained through the “Impulsfinanciering” provided
by the Flemish Government.
Table 3.
110) of protonated xanthosine and the product ions of the [BH2]?- NH3(m/z 135), [BH2]?- NH3? H2
CH3OH (m/z 167) of protonated guanosine
Summary of the product ions following the dissociation of the [BH2]?(m/z 153), [BH2]?-NH3(m/z 136), [BH2]?-HNCO (m/z
18O (m/z 155), [BH2]?-NH3?
Exact m/z
Elemental
Compositions
Guo m/z
(135?H2
18O)
Guo m/z
(135?CH3OH)b
Guo
m/z
135
Xan
m/z
110
Xan
m/z
136
Xan
m/z
153
Elemental
Compositions
Nucleoside
Exact m/z
XC6H7N4O2
?
167.0569
156.0296C5H4N3
16O2
18O?
X
XXXXC5H4N3O3
?
154.0253
154.0253
155.0455
C5H4N3
C5H5N4
16O3
16O18O?
?
X
XXXC5H5N4O2
?
153.0413
153.0413
138.0189
C5H5N4
C5H2N3
16O2
16O18O?
?
X
X
XXXXC5H2N3O2
?
136.0147
136.0159
137.0349
C5H2N3
C5H3N4
16O2
18O?
?
Xa
X
XX(X)C5H3N4O?
135.0307
135.0307
112.0397
C5H3N4
C4H4N3
16O?
18O?
X
X
XXXC4H4N3O?
110.0354
110.0354
110.0240
C4H4N3
C4H2N3
16O?
18O?
X
XXXC4H2N3O?
108.0198
108.0198
107.0358
95.0131
C4H2N3
C4H3N4
C4HN2
16O?
?
X
X
X
XXXC4H3N4
?
107.0358
18O?
XXXXC4HN2O?
93.0089
93.0089
85.0288
C4HN2
C3H3N2
16O?
18O?
X
X
XXXC3H3N2O?
83.0245
83.0245
82.0405
83.0131
C3H3N2
C3H4N3
C3HN2
16O?
?
X
X
XXXC3H4N3
?
82.0405
18O?
XXXC3HN2O?
81.0089
81.0089
80.0249
69.0089
68.0136
68.0253
67.0296
65.0140
58.0197
C3HN2
C3H2N3
C2HN2
C3H2N16O?
C2H2N3
C3H3N2
C3HN2
C2H2N18O?
16O?
?
X
XX
X
X
XX
X
X
C3H2N3
C2HN2O?
C3H2NO?
C3H2N?
C3H3N2
C3HN2
?
80.0249
69.0089
68.0136
68.0253
67.0296
65.0140
16O?
X
XXX
X*
?
?
X
X
XX
X
?
?
X
X
XX
?
XXXXC2H2NO?
56.0136
56.0136
55.0296
53.9980
53.0140
C2H2N16O?
C2H3N2
C2NO?
C2HN2
?
X
X
X
XXXX
X
X
C2H3N2
C2NO?
C2HN2
?
55.0296
53.9980
53.0140
X
X
?
XXX*
?
(X) Following a transition between different fragmentation pathways
aMinor abundance
bOnly product ions shown which are common with the protonated xanthosine- or the m/z 135 out of protonated guanosine pathways
1061
J Am Soc Mass Spectrom 2006, 17, 1050–1062IN-SOURCE CID OF GUANOSINE

Page 13

References
1. March, R. E. An introduction to quadrupole ion trap mass spectrometry.
J. Mass Spectrom. 1997, 32, 351–369.
2. Chernushevich, I. V.; Loboda, A. V.; Thomson, B. A. An introduction to
quadrupole-time-of-flight mass spectrometry. J. Mass Spectrom. 2001, 36,
849–865.
3. Lemiere, F.; Vanhoutte, K.; Jonckers, T.; Marek, R.; Esmans, E. L.;
Claeys, M.; Van den Eeckhout, E.; Van Onckelen, H. Differentiation
between isomeric phenylglycidyl ether adducts of 2=-deoxyguanosine
and 2=-deoxyguanosine-5=-monophosphate using liquid chromatogra-
phy/electrospray tandem mass spectrometry. J. Mass Spectrom. 1999, 34,
820–834.
4. Nunez, O.; Moyano, E.; Galceran, M. T. High mass accuracy in-source
collision-induced dissociation tandem mass spectrometry and multi-
step mass spectrometry as complementary tools for fragmentation
studies of quaternary ammonium herbicides. J. Mass Spectrom. 2004, 39,
873–883.
5. Tuytten, R.; Lemière, F.; Van Dongen, W.; Esmans, E. L.; Witters, E.;
Herrebout, W.; Van, D.V; Dudley, E.; Newton, R. P. Intriguing mass
spectrometric behavior of guanosine under low energy collision-in-
duced dissociation: H2O adduct formation and gas-phase reactions in
the collision cell. J. Am. Soc. Mass Spectrom.2005, 16, 1291–1304.
6. Dudley,E.;Tuytten,R.;Bond,A.;Lemière,F.;Brenton,A.G.;Esmans,E.L.;
Newton, R. P. Study of the mass spectrometric fragmentation of
pseudouridine: Comparison of fragmentation data obtained by matrix-
assisted laser desorption/ionization post-source decay, electrospray ion
trap multistage mass spectrometry, and by a method utilizing electrospray
quadrupole time-of-flight tandem mass spectrometry and in-source
fragmentation. Rapid Commun. Mass Spectrom. 2005, 19, 3075–3085.
7. Sleno, L.; Volmer, D. A. Ion activation methods for tandem mass
spectrometry. J. Mass Spectrom. 2004, 39, 1091–1112.
8. Gabelica, V.; Lemaire, D.; Laprevote, O.; De Pauw, E. Kinetics of solvent
addition on electrosprayed ions in an electrospray source and in a
quadrupole ion trap. Int. J. Mass Spectrom. 2001, 210, 113–119.
9. Gregson, J. M.; McCloskey, J. A. Collision-induced dissociation of
protonated guanine. Int. J. Mass Spectrom. 1997, 165, 475–485.
10. Frycak, P.; Huskova, R.; Adam, T.; Lemr, K. Atmospheric pressure
ionization mass spectrometry of purine and pyrimidine markers of
inherited metabolic disorders. J. Mass Spectrom. 2002, 37, 1242–1248.
12. Turecek, F.; Chen, X. H. Protonated adenine: Tautomers, solvated
clusters, and dissociation mechanisms. J. Am. Soc. Mass Spectrom. 2005,
16, 1713–1726.
13. Yao, C. X.; Cuadrado-Peinado, M. L.; Polasek, M.; Turecek, F. Gas-phase
tautomers of protonated 1-methylcytosine. Preparation, energetics, and
dissociation mechanisms. J. Mass Spectrom. 2005, 40, 1417–1428.
14. Curcuruto, O.; Tarzia, G.; Hamdan, M. Differentiation between isomeric
methylated purine derivatives by means of chemical ionization and
collision-induced dissociation. Rapid Commun. Mass Spectrom. 1992, 6,
596–600.
15. Nelson, C. C.; McCloskey, J. A. Collision-induced dissociation of uracil
and its derivatives. J. Am. Soc. Mass Spectrom. 1994, 5, 339–349.
16. Boos, K. S.; Grimm, C. H. High-performance liquid chromatography
integrated solid-phase extraction in bioanalysis using restricted access
pre-column packings. TRAC-Trends Anal. Chem. 1999, 18, 175–180.
17. Roberts, E. L. L.; Newton, R. P. Estimation of guanine deaminase using
guanosine as a “prosubstrate”. Anal. Biochem. 2004, 324, 250–257.
1062
TUYTTEN ET AL.J Am Soc Mass Spectrom 2006, 17, 1050–1062