Positive and negative electrospray ionisation travelling wave ion mobility mass spectrometry and low-energy collision-induced dissociation of sialic acid derivatives

Article abstract of DOI:10.1002/rcm.5217

Mono- or oligosaccharide-containing samples, whether they are derived from biological sources or products of chemical synthesis, are often mixtures of spatial or constitutional isomers. The possibility of characterising or performing quality control on su

Full text of DOI:10.1002/rcm.5217

Research Article
Received: 4 July 2011
Revised: 5 August 2011
Accepted: 8 August 2011
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244
wileyonlinelibrary.com) DOI: 10.1002/rcm.5217
(
Positive and negative electrospray ionisation travelling wave ion
mobility mass spectrometry and low-energy collision-induced
dissociation of sialic acid derivatives
1
2
2
1
Wolfgang Winkler , Wolfgang Huber , Reinhard Vlasak and Günter Allmaier *
1
Institute of Chemical Technologies and Analytics, Vienna University of Technology, Vienna, Austria
Department of Molecular Biology, University Salzburg, Salzburg, Austria
2
Mono- or oligosaccharide-containing samples, whether they are derived from biological sources or products of
chemical synthesis, are often mixtures of spatial or constitutional isomers. The possibility of characterising or
performing quality control on such samples by mass spectrometry is hampered because these isomers cannot be
separated by their mass-to-charge ratio alone. Therefore, the use of techniques to separate the isobaric sample
compounds prior to mass spectrometric characterisation is mandatory. Travelling wave ion mobility separation offers
the possibility of separating mixtures based on their compound’s collisional cross-sections in the gas phase and can
easily be combined with mass spectrometry for further characterisation. Here, we use 5-N-acetylneuraminic acid
and several derivatives as model compounds to evaluate the separation power of travelling wave ion mobility
spectrometry and present an approach to clearly identify constitutional isomers in mixtures in combination with
low-energy collision-induced dissociation (CID) in the negative ion mode even if they cannot be completely
separated by ion mobility. Copyright © 2011 John Wiley & Sons, Ltd.
For several decades, ion mobility spectrometry was mainly
used to analyse small volatile molecules most notably for detec-
performed on mono- and oligosaccharides or glycoconjugates.
Using an instrument equipped with a drift-tube-type ion
mobility spectrometer operated at ambient pressure, the
[1,2]
tion of explosives or chemical warfare agents.
By combining
[12,13]
ion mobility separation and mass spectrometry, analytical
instruments were developed, which allowed the characterisa-
tion of analyte ions based on their three-dimensional structure
or more exactly the collisional cross-section as well as their
mass/charge ratio. These instruments – many of them being
home-built or otherwise customised – were of very heteroge-
neous designs as they combined different types of ion sources,
separation of differently linked di- and trisaccharides
was shown, as well as the separation of anomeric species of
[14]
monosaccharides. Using a high-field asymmetric waveform
(FAIMS)-type ion mobility separator, the possibility of
separating linkage and position isomers of disaccharides was
[15]
shown. Since the introduction of a travelling wave (T-wave)
[16]
type ion mobility mass spectrometer by Waters,
several
[2–4]
ion mobility spectrometers and mass analysers.
Ion mobi-
studies using this instrument to analyse carbohydrates have
been presented. The possibility of distinguishing protein-
lity mass spectrometry has now been successfully applied to
study structural or conformational isomers as well as complex
samples, where it allows the partitioning of compounds by sub-
[
17]
derived glycan isomers,
and differently linked tri- and
[18]
hexasaccharides, and of significantly reducing the complex-
[5]
stance class or charge state. The samples that were studied
covered a broad range of molecular weights from small, even
ity of mass spectra when analysing glycoconjugate mixtures of
[19]
extreme heterogeneity was shown.
[
6]
[7,8]
chiral, molecules, complex peptide mixtures,
synthetic
This study focuses on the use of ion mobility mass
spectrometry (IM-MS) to analyse mixtures of constitutional
isomers of sialic acids, which are a large family of derivatives
[
9]
[10,11]
polymers, and proteins up to intact protein assemblies.
The characterisation of sugars, oligosaccharides or glycans is
especially challenging for mass spectrometry alone, due to the
sheer number of isobaric species arising from constitutional
and spatial isomerism, which often exist simultaneously and
in equilibrium. When characterising such samples, the option
to additionally separate the compounds based on their
collisional cross-section is quite valuable. Consequently, many
studies utilising different types of ion mobility separators were
[20]
of neuraminic acid. Most of these compounds are found as
[
21]
[22]
components of glycoproteins,
glycolipids,
or as part of
[
23,24]
bacterial cell walls. A common modification of sialic
acids is O-acetylation at carbons 4, 7, 8 or 9.
[25]
A number of
bacterial pathogens use sialic acids. While E. coli K1,
N. meningitides and C. jejuni obtain sialic acids by de novo
synthesis, others, like H. influenzae, N. gonorrhoeae and E. coli
K12, are able to take up free sialic acids from the environ-
[26]
ment by specific transporters.
To study the catabolism
of O-acetylated sialic acids in bacteria, pure preparations
of sialic acid derivatives are required. Using such
preparations, Steenbergen et al. recently identified the YihS
gene product of E. coli required for growth on 9-O-acetylated
*
Correspondence to: G. Allmaier, Institute of Chemical
Technologies and Analytics, Vienna University of Technology,
Getreidemarkt 9/164-IAC, A-1060 Vienna, Austria.
E-mail: guenter.allmaier@tuwien.ac.at
[27]
sialic acid.
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244
Copyright © 2011 John Wiley & Sons, Ltd.

W. Winkler et al.
There are several other chromatographic or spectroscopic
methods available, which can be used to separate such mix-
tures and characterise the components. Nuclear magnetic
resonance (NMR) is widely used for structural elucidation
or confirmation of synthesis products and would allow the
determination of the absolute configuration. However,
compared with MS-based methods, NMR is slow, needs
high amounts of sample and is not generally applicable to
off and the filtrate evaporated in vacuum to obtain 8,9-O-
isopropylidene-Neu5Ac (620 mg, 89%, used for MS analysis).
8,9-O-Isopropylidene-Neu5Ac (200 mg, 0.57 mmol) was
suspended in 8 mL methylene chloride. DMAP (167 mg,
1.37 mmol) and acetic anhydride (500 mg, 4.9 mmol) were
added and the suspension stirred at room temperature for
2 h. The reaction mixture was extracted with H O (4 Â 1 mL)
2
and the aqueous phase re-extracted once with methylene
[28]
even simple mixtures.
like high pH anion-exchange chromatography or gas
have already been used to separate sialic
Chromatography-based methods
chloride (1 mL). The combined H O phases containing 8,9-O-
2
[29]
isopropylidene-Neu4,5Ac were used in the next step without
2
[
30]
chromatography
isolation. To the aqueous solution of 8,9-O-isopropylidene-
acids. These methods either require sample derivatisation or
suffer from the need to use volatile solvent compounds when
directly coupled to mass spectrometers.
Neu4,5Ac 4 mL of concentrated acetic acid were added and
stirred for 3 h at 60 C. After evaporation of the solvent in
2
ꢀ
vacuum the resulting crude Neu4,5Ac was dissolved in 1 mL
2
of 1-propanol/H O (3:1, v/v), loaded onto a short silica gel
2
column and chromatographed by dry-column flash chroma-
tography. Elution was started with pure 1-propanol followed
EXPERIMENTAL
by 90%, 80% and 75% 1-propanol/H
column was sucked to dryness after each fraction of the
solvent. Fractions containing Neu4,5Ac were pooled and
evaporated to obtain Neu4,5Ac (71 mg, 35% over 2 steps).
2
O (v/v) as eluents. The
Materials
2
All chemicals were of analytical grade (p.a.). N,N-Dimethyl-
2
4
-pyridinamine (DMAP) (>98%), 2,2-dimethoxypropane
98%), p-toluenesulfonic acid monohydrate (98,5%) and acetic
anhydride (98%) were purchased from Sigma Aldrich
Deisenhofen, Germany). The cation-exchange (50W-X8 resin,
0–50 mesh, hydrogen form) and the anion-exchange mate-
rial (1-X8, 100–200 mesh, formate form) were from Bio-Rad
Laboratories (Hercules, CA, USA). Silica gel N (without
binder, for thin layer chromatography) was purchased from
Machery-Nagel (Düren, Germany). Sodium hydroxide,
acetone, methanol, 2-propanol, methylene chloride, formic
acid and acetic acid were from Fluka (Buchs, Switzerland).
Neu5Ac was obtained from Codexis (Jülich, Germany).
(
MS and MS/MS measurements
(
2
All mass spectrometric and ion mobility mass spectrometric
measurements were carried out using a SYNAPT HDMS
instrument of IM/QRTOF geometry (Waters, Manchester,
UK) equipped with a LockSpray dual electrospray source.
The source conditions were as follows: source temper-
ꢀ
ature: 80 C; desolvation gas: nitrogen (flow rate: 600 L/h,
ꢀ
temperature: 250 C); cone gas: nitrogen (flow rate: 50 L/h).
The acquisition parameters were: electrospray voltage:
2.0–2.5 kV; extraction cone voltage: 2.5 V, Ar flow rate:
1.5 mL/min (resulting in a pressure of 0.01 mbar in the trap/
transfer cell); trap and transfer CE (collision energy) adjusted
as needed to avoid unwanted fragmentation in MS mode
(typically 1–2 eV) and induce fragmentation in MS/MS mode
8,9-O-Isopropylidene-5-N-acetylneuraminic acid (Neu5Ac-8,9-
protected) was an intermediate product of the Neu4,5Ac
2
synthesis. 5-N-Acetyl-9-O-acetylneuraminic acid (Neu5,9Ac2)
was obtained from Applied Biotech (Salzburg, Austria).
(
typically 4–12 eV). The samples were dissolved at
.0025 mg/mL in methanol/water, 1: 1 (v/v) (negative mode)
and 0.01 mg/mL in methanol/water, 1: 1 (v/v) containing
.1% formic acid (positive mode) and infused via a syringe
0
Synthesis of 5-N-acetyl-4-O-acetylneuraminic acid and
5
-N-acetyl-8,9-O-isopropylideneneuraminic acid
0
The synthesis of 5-N-acetyl-4-O-acetylneuraminic acid was
pump at a flow rate of 5 mL/min. The reflectron TOF mass
analyser (allowing multiple ion reflections) was operated in
V-geometry for the acquisition of positive and negative ion
mode mass spectra, and calibration for both polarities in the
m/z range of 100–1000 was performed using sodium formate
solution (generated by mixing equal volumes of aqueous
solutions of sodium hydroxide (0.1 mol/L) and formic acid
(0.2 mol/L) and diluting 1:500 using 2-propanol/water, 8: 2 (v/v)
at a flow rate of 5 mL/min).
All MS and MS/MS spectra were recorded as accurate
mass data with Neu5Ac for lockmass correction in positive
and negative ion mode (positive mode: 0.01 mg/mL in
methanol/water, 1:1 (v/v) containing 0.1% formic acid,
negative mode: 0.001 mg/mL in methanol/water, 1:1 (v/v),
flow rate 5 mL/min for both modes) using the protonated
and deprotonated molecule, respectively. In cases where the
signal intensity of the analyte or lockmass exceeded the
recommended value for accurate mass measurement of 200
counts per second, the transmission rate was reduced by
setting the sample cone voltage to sufficiently low values.
Accurate mass values were derived from centroid mode mass
[31]
performed according to Ogura et al.
with a modification
in the acetylation step, where DMAP was used as catalyst
instead of pyridine. Briefly, Neu5Ac (1.00 g, 3.23 mmol) and
+
1
.0 g Dowex 50W X-8 (H ) cation-exchange resin were stirred
in 30 mL methanol for 16 h at room temperature. The ion-
exchange resin was removed by filtration and the solution
evaporated in vacuum to obtain Neu5Ac1Me (780 mg, 75 %).
Neu5Ac1Me (780 mg, 2.41 mmol) was dissolved in 15 mL
acetone. After addition of 2,2-dimethoxypropane (1.20 g,
1
1.5 mmol) and 1.5 mg p-toluenesulfonic acid the suspension
was stirred for 1 h at room temperature. Then, 0.2 g of Dowex
-X8 (formate) anion-exchange resin was added and stirred
1
for 15 min at room temperature. The ion-exchange resin
was removed by filtration and the solution was evaporated
in vacuum to obtain 8,9-O-isopropylidene-Neu5Ac1Me
(
2
(
850 mg, 97%). 8,9-O-Isopropylidene-Neu5Ac1Me (730 mg,
.00 mmol) was suspended in 10.0 mL 0.2 M NaOH
2.00 mmol) and stirred for 4 h at room temperature. The
+
solution was neutralised with 0.5 g of Dowex 50 W X-8 (H )
cation-exchange resin. The ion-exchange resin was filtered
wileyonlinelibrary.com/journal/rcm
Copyright © 2011 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244

ESI travelling wave IM-MS and low-energy CID of sialic acid derivatives
spectra and every assignment of an ion species to a peak
was verified by calculation of the elemental composition
from accurate mass data using the Elemental Composition
module (version 4.0) included in MassLynx (version 4.1
SCN 704, Waters).
Ion mobility-MS and MS/MS measurements
All the T-wave ion mobility measurements were actually
carried out as MS/MS mobility experiments, i.e. the quadru-
pole was operated in resolving mode to ensure that only the
ion species of interest (deprotonated, protonated or sodiated
molecule) was transferred into the ion mobility cell. In case
of mixtures of non-isobaric analytes, the LM (low-mass)
resolution parameter controlling the width of the quadrupole
isolation window was changed to allow transmission of all spe-
cies of interest. The acquisition parameters were: electrospray
voltage: 2.5 kV; sample cone voltage: 15 V; extraction cone volt-
age: 3 V; Ar flow rate: 3 mL/min (resulting in a pressure of
0.016 mbar in the trap/transfer cell); trap DC BIAS: 14 V; trap
and transfer CE: 6 eV (negative mode) and 5 eV (positive
mode); ion mobility gas flow rate: 32 mL/min nitrogen (result-
ing in a pressure of 0.79 mbar in the ion mobility cell); IMS
wave height: 8 V (negative mode) and 7.5 V (positive mode);
IMS wave velocity: 300 m/s; scan time: 3 s. The samples were
dissolved at 0.001 mg/mL in methanol/water, 1:1 (v/v) (nega-
tive ion mode) or 0.005 mg/mL in methanol/water, 1:1 (v/v)
containing 0.1% formic acid (positive ion mode) and infused
via syringe pump at flow rates of 10 mL/min. The ion mobility
data were loaded into Driftscope software (version 2.1,
Waters) and exported to MassLynx for display and further
processing. To determine the centroids of the arrival time peaks,
the data were imported into Origin (version 8.0, Originlabs,
Northampton, MA, USA) to fit Gaussian functions.
Actual T-wave IM-MS/MS experiments (with fragmentation
taking place in the transfer cell, i.e. after mobility separation of
the precursor ions) were carried out by applying a transfer CE
of 20 eV.
RESULTS
To evaluate the resolving capability of the ion mobility separa-
tion on isobaric and non-isobaric derivatives of Neu5Ac, a set of
four substances was chosen: Neu5Ac, the constitutional
isomers Neu4,5Ac and Neu5,9Ac and Neu5Ac-8,9protected,
2
2
an intermediate product obtained during the synthesis of
Neu4,5Ac (see Table 1 for detailed information on these sub-
2
stances). The purities and elemental compositions of all synth-
esis products were assured by accurate mass measurement in
positive and negative ion mode (data not shown). Figure 1
shows the arrival time distributions of the deprotonated (left
column), the protonated (center column) and sodiated (right
column) molecules of each individual substance. The arrival
time for each ion was determined and is listed in Table 1. For
consecutive measurements of the same sample, variations of
Æ0.01 ms in arrival time were typically observed (data not
shown). Not surprisingly, the arrival times for Neu5Ac in posi-
tive and negative ion mode are clearly lower than those
determined for its derivatives, which exhibit 40 Da and 42 Da
higher molecular weights. These three Neu5Ac derivatives
clearly show different arrival times in negative ion mode, but
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244
Copyright © 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

W. Winkler et al.
Figure 1. Arrival time distributions for Neu5Ac (A–C) and derivatives (Neu5Ac-8,9protected: D–F, Neu4,5Ac2:
G–I, Neu5,9Ac : J–L) in positive (B, C, E, F, H, I, K, L) and negative (A, D, G, J) ion mode. The bold solid lines
2
show the total arrival time distribution for all ions across the whole m/z range, the dotted lines represent the
arrival time of the respective intact deprotonated, protonated or sodiated molecule. Arrival time labels represent
the centroid of the peak. Additional peaks are caused by product ions generated prior to ion mobility separation
and are labelled with their arrival time and plotted using dashed and dashed-dotted lines, in case their peaks
overlap the peak of the intact precursor ion. Peaks labelled “doubly charged” arise from doubly charged species.
the arrival times found for the protonated and sodiated
molecules are nearly identical. The observed peak widths at
half height are also different in positive and negative ion mode,
ranging from 0.87–1.00 ms for the deprotonated molecules to
The findings described above clearly show that an IM-
based separation of these substances can only be successful
when carried out in negative ion mode, as the substances
show inadequate arrival time differences and significant
post-source fragmentation in positive ion mode. To evaluate
the applicability of IM-MS on more complex samples, binary
mixtures of Neu5Ac and its derivatives were analysed.
Figure 2(A) shows the arrival time distribution of a mixture
of equal amounts of Neu5Ac and Neu5Ac-8,9protected. This
mixture contains the two compounds which exhibit the
highest possible difference in ion mobility and can clearly be
separated into two major components with arrival times of
4.27 and 5.58 ms, which perfectly match those values, that
have been determined for Neu5Ac (4.29 ms) and Neu5Ac-8,
9protected (5.62 ms) when analysed individually. Figure 2(B)
shows the arrival time distribution obtained for the mixture of
0.71–0.80 and 0.66–0.77 ms for the protonated and sodiated
molecules. Although the acquisition conditions have been
optimised to avoid fragmentation before and after ion mobility
separation, some arrival time distributions show the presence
of product ions, which have been generated between the
quadrupole and the ion mobility separation cell. In particular,
the protonated molecules of all species (Fig. 1, middle column)
are susceptible to fragmentation by neutral loss of water and/or
acetic acid. The arrival time distributions obtained from sodiated
molecules show the presence of considerable amounts of doubly
charged ions, with m/z values very close to those of the sodiated
molecules, which therefore could pass the quadrupole.
wileyonlinelibrary.com/journal/rcm
Copyright © 2011 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244

ESI travelling wave IM-MS and low-energy CID of sialic acid derivatives
the isobaric compounds Neu4,5Ac
2
2
and Neu5,9Ac . Unfortu-
nately, the arrival time difference and resolution are not suffi-
cient to achieve a level of separation comparable with that of
the other mixture. The arrival time distribution shows only a
single peak with a centroid at an arrival time of 5.13 ms. It is
not distorted or irregularly shaped and does not exhibit a
significantly higher width at half maximum height than the
compounds when analysed individually (1.06 ms vs. 0.87
and 0.96 ms). Figure 2(C) shows that the curve obtained by
summing up the arrival time distributions of Neu4,5Ac and
2
Neu5,9Ac when recorded individually is nearly identical in
2
terms of centroid, intensity and shape to the peak obtained by
analysing the corresponding mixture.
The results achieved so far suggest that the resolving power
of the ion mobility separator used in these experiments is not
sufficient to achieve even partially separated peaks in the
arrival time distributions if the compounds’ arrival times
do not differ by more than about 1 ms. Therefore, a mixture
of isobaric compounds like Neu4,5Ac
be identified by recording mass and ion mobility spectra alone.
As a consequence, MS/MS spectra of Neu4,5Ac and
Neu5,9Ac were recorded in positive and negative ion mode
2 2
and Neu5,9Ac cannot
2
2
to identify product ions, which allowed the unambiguous
assignment of the position of the O-acetylation. Figure 3
shows MS/MS spectra of protonated and deprotonated
2 2
Neu4,5Ac and Neu5,9Ac . The product ions that are generated
from the protonated molecules (Figs. 3(A) and 3(B)) are nearly
exclusively formed by neutral losses of acetic acid (from
O-acetylation), ketene (from N-acetylation) and water (from
hydroxy groups). Naturally, there are differences in the positive
ion MS/MS spectra that would allow us to discriminate
between Neu4,5Ac and Neu5,9Ac , but these are mainly
2 2
differences in relative peak intensities rather than in the types
of product ions that are observed. This makes the decision as
to whether a sample contains a mixture of Neu4,5Ac
and Neu5,9Ac or only a single compound very difficult. The
2
2
MS/MS spectra generated from sodiated Neu4,5Ac2 and
Neu5,9Ac (data not shown) are equally unsuitable for dis-
2
criminating between Neu4,5Ac and Neu5,9Ac , as they show
2
2
a similar amount of intense peaks corresponding to neutral
losses alongside some minor peaks from cross-ring cleavages.
The picture is completely different for the negative ion mode
MS/MS spectra (Figs. 3(C) and 3(D)) of deprotonated
Neu4,5Ac and Neu5,9Ac , which are less complex and mainly
2
2
show peaks corresponding to product ions from cross-ring
cleavages. In addition, the deprotonated Neu4,5Ac2 and
Figure 2. Arrival time distributions for binary
mixtures. (A) Mixture of Neu5Ac and Neu5Ac-8,9-
protected (0.001 mg/mL each). The bold solid line
shows the total arrival time distribution for all ions
across the whole m/z range, the dotted and dashed line
represent each deprotonated molecule. (B) Mixture of
Neu5,9Ac give rise to different sets of product ions, which
2
easily allow discrimination of the two species or their presence
in a mixture. Figure 4 shows the putative fragmentation
pathway. For the deprotonated Neu4,5Ac , the neutral loss of
2
acetic acid seems to be the preferred fragmentation. The
resulting product ion (m/z 290.0877) can further fragment by
0,4-cross-ring cleavage to form a product ion at m/z 170.0447;
the formation of the expected second product ion at m/z
2 2
Neu4,5Ac and Neu5,9Ac (0.001 mg/mL each). The
bold solid line shows the total arrival time distribution
for all ions, the dotted line represents the total inten-
sity for both deprotonated molecules, the dashed-
dotted line represents a product ion overlapping the
peak of the precursor ion. (C) The bold solid line
shows the total arrival time distribution for all ions
119.0344 could not be confirmed experimentally. A similar
fragmentation process has already been observed for unsatu-
rated glycosides in negative mode fast atom bombardment
spectra with high-energy collisions and proposed to be result
of the same Neu4,5Ac /Neu5,9Ac – mixture as used
2
2
[32]
of a retro-Diels-Alder (RDA) reaction.
The product ion at
in (B); the dotted and dashed lines represent arrival
time distributions of Neu4,5Ac and Neu5,9Ac when
m/z 170.0447 gives rise to ions at m/z 126.0556 and 98.0607 by
neutral loss of CO , and of CO followed by CO, respectively.
The deprotonated Neu5,9Ac can fragment using the same
2
2
recorded individually (shown in Figs. 1(G) and 1(J)).
The dashed-dotted line is the sum of both.
2
2
2
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244
Copyright © 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

W. Winkler et al.
Figure 3. Centroid low-energy CID spectra of Neu4,5Ac
2 2
in positive (A) and negative (C) mode and Neu5,9Ac in
positive (B) and negative (D) mode. Nomenclature of cross-ring cleavage fragments is according to Domon
[34]
and Costello.
The elemental composition of the suggested ion structures was confirmed by accurate
mass measurement.
pathway: neutral loss of water from carbon #4 (according to
(m/z 262.0935 and 87.0083), which could not have taken place
if bond #2 is a double bond.
With these results it is possible to decide whether a sample
[
33]
nomenclature ) generating the ion at m/z 332.0983 followed
[34]
by cleavage of bonds 0 and 4 (according to nomenclature
by a RDA reaction to form product ions at m/z 170.0448 and
61.0449. The product ions derived from m/z 170.0448 are the
same as those described for Neu4,5Ac . The elimination of the
substituent at carbon #4 seems to be less favourable for
deprotonated Neu5,9Ac ; therefore, product ions can be
)
2 2
contains Neu4,5Ac , Neu5,9Ac or a mixture of both by the
1
absence or presence of peaks from characteristic product ions
for each isomer. Of course, this approach is only possible if
pure samples of all compounds are available at the beginning
of investigations, which is often not the case. In addition, the
acquisition and thorough analysis of MS/MS spectra can be
2
2
observed that are generated by a 0,2-cross-ring cleavage
wileyonlinelibrary.com/journal/rcm
Copyright © 2011 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244

ESI travelling wave IM-MS and low-energy CID of sialic acid derivatives
Figure 4. Fragmentation pathways of deprotonated Neu4,5Ac and Neu5,9Ac . Nomenclature of cross-ring cleavage fragments
2
2
[34]
is according to Domon and Costello.
For each product ion, the exact mass (calculated using the molecular mass calculator
application included in MassLynx) and the observed m/z (set in italic) are provided.
2 2
Figure 5. (A) Arrival time distribution of a mobility MS/MS experiment on equimolar Neu4,5Ac /Neu5,9Ac – mixture. The
bold line represents the total arrival time for all ion species; dotted, dashed and dashed-dotted lines represent arrival time
distributions for product ions at m/z 290.0888, 87.0078 and 170.0460. Labels are arrival time peak centroids. (B) Negative ion
2 2
MS/MS spectrum of Neu4,5Ac /Neu5,9Ac – mixture generated by combining across whole arrival time range.
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244
Copyright © 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

W. Winkler et al.
rather time-consuming. Therefore, an IM-MS/MS experiment
was set up where the deprotonated molecules of Neu4,5Ac
and Neu5,9Ac were first separated as well as possible
by ion mobility and subsequently fragmented by low-energy
CID. Figures 5(A) and 5(B) show the results of such an IM-
MS/MS experiment for an equimolar mixture of Neu4,5Ac2
molecule. Unfortunately, there are no data available on the
gas-phase structure of cationised Neu5Ac or derivatives. Cal-
culations for hexoses have shown that sodium adducts are
multidentate complexes and that the interaction with sodium
2
2
[35]
can promote a certain conformation of the pyranose ring.
Experimental data and calculations for linear polymers and
oligosaccharides have shown that the linear molecules adapt
and Neu5,9Ac . The MS/MS spectrum (Fig. 5(B)) contains
2
[36,37]
all those peaks corresponding to product ions that are
generated from each precursor ion – indicating that the
applied conditions do not selectively fragment only one
compound. The arrival time distribution (Fig. 5(A)) shows
the traces for selected product ions: m/z 87.0078 (exclusively
cyclic structures upon interaction with cations.
There-
fore, it has to be assumed that Neu5Ac and derivatives form
very similar gas-phase structures upon cationisation, whereas
the abstraction of a proton does not induce the generation of
structures with similar collisional cross-sections.
formed from Neu5,9Ac ), 290.0888 (exclusively formed from
As a result of the low resolution in ion mobility mode,
species with arrival time differences of about 0.4 ms are
detected as single peaks in the arrival time distribution.
Analysing a mixture of compounds with 1.3 ms arrival time
difference revealed two distinct peaks in the arrival time
distribution, but these compounds differed by 40 Da in their
molecular weight. Probably such arrival time differences
cannot be achieved for most small isobaric molecules but,
by using CID to fragment the mobility-separated precursors,
their arrival time differences are retained by their product
ions and can be detected. This approach could be successfully
applied to a mixture of Neu5Ac carrying O-acetylations at
different positions, but only because of the existence of
product ions, that allowed the discrimination of the constitu-
tional isomers. Unfortunately, it is very unlikely to find such
product ions that would allow the discrimination of standard
monosaccharides (e.g. glucose or mannose), but linkage and
sugar composition-specific fragmentation has been shown
2
Neu4,5Ac ), and 170.0460 (formed from both Neu4,5Ac2
2
and Neu5,9Ac ). Those product ions, that are formed exclu-
2
sively from one compound, are detected at different arrival
times (4.94 ms for m/z 290.0888 and 5.29 ms for m/z
8
7.0078), indicating that they were derived from partially
mobility separated precursors. The arrival times of these
product ions are in perfect agreement with those
previously determined for their respective precursor ions
(
Neu4,5Ac
the arrival time peak of the product ion m/z 170.0460
5.03 ms), which is derived from both precursors, is located
2 2
: 4.92 ms; Neu5,9Ac : 5.29 ms). The centroid of
(
between the arrival times of its precursors. These results
clearly show that IM-MS/MS experiments allow the
detection of mixtures of isobaric compounds if they give rise
to unique product ions.
[38]
for oligosaccharides.
DISCUSSION
In this study, we investigated the applicability of T-wave
ion mobility mass spectrometry for the separation of a
mixture of 5-N-acetylneuraminic acids O-acetylated at
different positions and for the unambiguous identification
of the compounds. Although the constitutional isomers
showed arrival time differences of nearly 0.4 ms, not even
partially separated peaks could be observed in the arrival
time distributions of equimolar mixtures. This can either be
caused by the lower resolving power of the travelling
wave type ion mobility spectrometer as it is used in the
SYNAPT instrument in comparison with other ion mobility
CONCLUSIONS
The approach presented in this paper of using ion mobility
mass spectrometry to detect and separate conformational iso-
mers of acetylated sialic acids has not been completely suc-
cessful. The resolution (ion mobility resolution Ω (collision
cross-sections)/ΔΩ (full width of mobility peak at half
height) of the T-wave ion mobility spectrometer as implemen-
ted in the applied instrument was not sufficient (Ω/ΔΩ = 10)
to clearly (base-line) separate the isobaric species by their dif-
ferent ion mobilities. Nevertheless, the absence or presence of
each of the conformational isomers could be confirmed when
ion mobility and negative ion mode MS/MS were used in
combination. This experimental setup could also be used to
analyse any mixture of isobaric species where pure com-
pounds are not available to analyse them individually.
Undoubtedly, this procedure would enormously benefit from
instrumental improvement with regard to the performance of
the mobility separation (e.g. the recently released SYNAPT
G2 or G2S with an improved ion mobility resolving
[
17]
spectrometer types
or peak broadening as a consequence
of the superposition of different gas-phase structures,
conformers and anomers. Unfortunately, there are no data
available on the occupancy rates of the possible conformers
and anomers for neuraminic acid or derivatives in the used
solvent mixture. As the separation of monosaccharide
anomers by ion mobility mass spectrometry has already been
[14]
shown with another instrument type,
it has to be assumed
that the contribution of different gas-phase structures or
conformers to peak broadening is rather marginal. An
unexpected finding was that the arrival times of protonated
and sodiated molecules were nearly identical for the com-
pounds with similar or identical molecular weights, whereas
the arrival times of the deprotonated molecules in negative
mode showed far higher differences. However, not only
the substances’ arrival time differences changed, but also
their order: Neu5Ac-8,9protected showed the highest
arrival time as a deprotonated and the lowest as a sodiated
[39]
power (Ω/ΔΩ = 45) ). In addition, independence from
liquid-phase separation methods to separate conforma-
tional isomers by using ion mobility separation could
dramatically reduce analysis time and sample consumption
as well as add an additional degree of peak capacity to the
commonly applied HPLC/ESI-MS approach. Our approach,
despite there being room for improvement, is helpful in
monitoring and quality control in chemical synthesis of sialic
wileyonlinelibrary.com/journal/rcm
Copyright © 2011 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244

ESI travelling wave IM-MS and low-energy CID of sialic acid derivatives
acid derivatives and might be applied in the characterisation
of complex hydrolysates of glycans moieties derived
from glycoproteins as well as in metabolomics of e.g.
bacterial pathogens.
[14] P. Dwivedi, B. Bendiak, B. Clowers, H. Hill. Rapid
resolution of carbohydrate isomers by electrospray ioniza-
tion ambient pressure ion mobility spectrometry-time-of-
flight mass spectrometry (ESI-APIMS-TOFMS). J. Am. Soc.
Mass Spectrom. 2007, 18, 1163.
[
[
15] W. Gabryelski, K. Froese. Rapid and sensitive
differentiation of anomers, linkage, and position isomers of
disaccharides using high-field asymmetric waveform ion
mobility spectrometry (FAIMS). J. Am. Soc. Mass Spectrom.
Acknowledgements
Financial support from the Austrian Science Fund (FWF,
Project No. L608-B03) and Salzburg Research Fellowship
Project No. P144001-01) is gratefully acknowledged. The
Synapt HDMS instrument was made available by a grant (to
G.A.) from the program UniInfra IV (Austrian federal ministry
of Science). The authors thank Martin Zehl (University of
Vienna), Ernst Pittenauer (Vienna University of Technology)
and Mathias Hofmann (Waters) for helpful discussions and
technical support.
2
003, 14, 265.
16] S. D. Pringle, K. Giles, J. L. Wildgoose, J. P. Williams,
S. E. Slade, K. Thalassinos, R. H. Bateman, M. T. Bowers,
J. H. Scrivens. An investigation of the mobility separation
of some peptide and protein ions using a new hybrid
quadrupole/travelling wave IMS/oa-ToF instrument.
Int. J. Mass Spectrom. 2007, 261, 1.
(
[17] J. P. Williams, M. Grabenauer, R. J. Holland, C. J. Carpenter,
M. R. Wormald, K. Giles, D. J. Harvey, R. H. Bateman,
J. H. Scrivens, M. T. Bowers. Characterization of simple
isomeric oligosaccharides and the rapid separation of
glycan mixtures by ion mobility mass spectrometry. Int. J.
Mass Spectrom. 2010, 298, 119.
REFERENCES
[
18] T. Yamagaki, A. Sato. Isomeric oligosaccharides analyses
using negative-ion electrospray ionization ion mobility
spectrometry combined with collision-induced dissociation
MS/MS. Anal. Sci. 2009, 25, 985.
[
[
[
[
1] M. A. Makinen, O. A. Anttalainen, M. E. T. Sillanpää. Ion
mobility spectrometry and its applications in detection of
chemical warfare agents. Anal. Chem. 2010, 82, 9594.
2] D. C. Collins, M. L. Lee. Developments in ion mobility
spectrometry-mass spectrometry. Anal. Bioanal. Chem.
[
19] S. Y. Vakhrushev, J. Langridge, I. Campuzano, C. Hughes,
J. Peter-Katalinic. Ion mobility mass spectrometry
analysis of human glycourinome. Anal. Chem. 2008, 80,
2
002, 372, 66.
2
506.
3] A. B. Kanu, P. Dwivedi, M. Tam, L. Matz, H. H. Hill.
Ion mobility-mass spectrometry. J. Mass Spectrom. 2008,
[20] T. Miyagi, K. Yamaguchi. Sialic acids, in Comprehensive
Glycoscience – From Chemistry to Systems Biology, (Eds:
J. P. Kamerling, G.-J. Boons, Y. C. Lee, A. Suzuki, N. Taniguchi,
A. G. J. Voragen), Elsevier, Oxford, 2007, pp. 297–323.
4
3, 1.
4] C. S. Creaser, J. R. Griffiths, C. J. Bramwell, S. Noreen,
C. A. Hill, C. L. P. Thomas. Ion mobility spectrometry: a
review. Part 1. Structural analysis by mobility measurement.
Analyst 2004, 129, 984.
5] L. Fenn, M. Kliman, A. Mahsut, S. Zhao, J. McLean. Charac-
terizing ion mobility-mass spectrometry conformation
space for the analysis of complex biological samples. Anal.
Bioanal. Chem. 2009, 394, 235.
6] J. R. Enders, J. A. McLean. Chiral and structural
analysis of biomolecules using mass spectrometry and
ion mobility-mass spectrometry. Chirality 2009, 21, E253.
7] J. A. Taraszka, A. E. Counterman, D. E. Clemmer. Gas-phase
separations of complex tryptic peptide mixtures. Fresenius’ J.
Anal. Chem. 2001, 369, 234.
8] S. J. Valentine, A. E. Counterman, D. E. Clemmer. A data-
base of 660 peptide ion cross sections: use of intrinsic size
parameters for bona fide predictions of cross sections.
J. Am. Soc. Mass Spectrom. 1999, 10, 1188.
[
[
[
21] A.
Kobata.
Glycoprotein
glycan
structures,
in
Comprehensive Glycoscience – From Chemistry to Systems
Biology, (Eds: J. P. Kamerling, G.-J. Boons, Y. C. Lee,
A. Suzuki, N. Taniguchi, A. G. J. Voragen), Elsevier, Oxford,
2007, pp. 39–72.
[
22] R. K. Yu, M. Yanagisawa, T. Ariga. Glycoshingolipid
structures, in Comprehensive Glycoscience – From Chemistry
to Systems Biology, (Eds: J. P. Kamerling, G.-J. Boons,
Y. C. Lee, A. Suzuki, N. Taniguchi, A. G. J. Voragen),
Elsevier, Oxford, 2007, pp. 73–122.
[
[
[
23] M. R. King, S. M. Steenbergen, E. R. Vimr. Going for
baroque at the Escherichia coli K1 cell surface. Trends
Microbiol. 2007, 15, 196.
[24] S. M. Steenbergen, E. R. Vimr. Biosynthesis of the
Escherichia coli K1 group polysialic acid capsule
2
occurs within a protected cytoplasmic compartment. Mol.
Microbiol. 2008, 68, 1252.
[
9] A. P. Gies, M. Kliman, J. A. McLean, D. M. Hercules. Char-
acterization of branching in aramid polymers studied by
MALDIÀion mobility/mass spectrometry. Macromolecules
[
[
25] C. Mandal, R. Schwartz-Albiez, R. Vlasak. Functions and
biosynthesis of O-acetylated sialic acids. Top. Curr. Chem.
2011, in press.
26] E. Severi, D. W. Hood, G. H. Thomas. Sialic acid utilization
by bacterial pathogens. Microbiology 2007, 153, 2817.
2
008, 41, 8299.
[
[
[
10] C. Uetrecht, R. J. Rose, E. van Duijn, K. Lorenzen, A. J. R. Heck.
Ion mobility mass spectrometry of proteins and protein
assemblies. Chem. Soc. Rev. 2010, 39, 1633.
11] B. T. Ruotolo, J. L. P. Benesch, A. M. Sandercock, S.-J. Hyung,
C. V. Robinson. Ion mobility-mass spectrometry analysis of
large protein complexes. Nat. Protoc 2008, 3, 1139.
12] B. Clowers, P. Dwivedi, W. Steiner, H. Hill, B. Bendiak.
Separation of sodiated isobaric disaccharides and trisac-
charides using electrospray ionization-atmospheric pressure
ion mobility-time of flight mass spectrometry. J. Am. Soc.
Mass Spectrom. 2005, 16, 660.
13] M. L. Zhu, B. Bendiak, B. Clowers, H. H. Hill. Ion mobility-
mass spectrometry analysis of isomeric carbohydrate
precursor ions. Anal. Bioanal. Chem. 2009, 394, 1853.
[27] S. M. Steenbergen, J. L. Jirik, E. R. Vimr. YjhS (NanS)
is required for Escherichia coli to grow on 9-O-acetylated
N-acetylneuraminic acid. J. Bacteriol. 2009, 191, 7134.
[28] A. J. Simpson, G. Woods, O. Mehrzad. Spectral editing
of organic mixtures into pure components using NMR
spectroscopy and ultraviscous solvents. Anal. Chem. 2007,
80, 186.
[29] J. S. Rohrer, J. Thayer, M. Weitzhandler, N. Avdalovic. Analy-
sis of the N-acetylneuraminic acid and N-glycolylneuraminic
acid contents of glycoproteins by high-pH anion-exchange
chromatography with pulsed amperometric detection
(HPAEC/PAD). Glycobiology 1998, 8, 35.
[
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244
Copyright © 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm

W. Winkler et al.
[
[
[
30] G. Reuter, R. Schauer. Comparison of electron and chemical
[36] S. Lee, T. Wyttenbach, M. T. Bowers. Gas phase structures
of sodiated oligosaccharides by ion mobility/ion chroma-
tography methods. Int. J. Mass Spectrom. Ion Processes 1997,
167/168, 605.
[37] T. Wyttenbach, G. von Helden, M. T. Bowers. Conformations
of alkali ion cationized polyethers in the gas phase:
polyethylene glycol and bis[(benzo-15-crown-5)-15-ylmethyl]
pimelate. Int. J. Mass Spectrom. Ion Processes 1997, 165/166,
377.
ionization mass spectrometry of sialic acids. Anal. Biochem.
1
986, 157, 39.
31] H. Ogura, K. Furuhata, S. Sato, K. Anazawa, M. Itoh,
Y. Shitori. Synthesis of 9-O-acyl- and 4-O-acetylsialic acids.
Carbohydr. Res. 1987, 167, 77.
32] M. Brakta, D. Sinou, M. Becchi, J. Banoub. Negative ion mass
spectra of some unsaturated C-glycosides following keV ion
bombardment. Org. Mass Spectrom. 1992, 27, 621.
[
[
33] G. Reuter, R. Schauer. Suggestions on the nomenclature of
sialic acids. Glycoconjugate J. 1988, 5, 133.
34] B. Domon, C. E. Costello. A systematic nomenclature for
carbohydrate fragmentations in FAB-MS/MS spectra of
glycoconjugates. Glycoconjugate J. 1988, 5, 397.
[38] G. Verardo, I. Duse, A. Callea. Analysis of underivatized
oligosaccharides by liquid chromatography/electrospray
ionization tandem mass spectrometry with post-column
addition of formic acid. Rapid Commun. Mass Spectrom. 2009,
23, 1607.
[
35] B. A. Cerda, C. Wesdemiotis. Thermochemistry and
[39] K. Giles, J. P. Williams, I. Campuzano. Enhancements in
traveling wave ion mobility resolution. Rapid Commun. Mass
Spectrom. 2011, 25, 1590.
+
structures of Na coordinated mono- and disaccharide
stereoisomers. Int. J. Mass Spectrom. 1999, 189, 189.
wileyonlinelibrary.com/journal/rcm
Copyright © 2011 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2011, 25, 3235–3244