Direct stereochemical assignmenet of hexose and pentose residues in flavonoid O-glycosides by fast atom bombardment and electrospray ionization mass spectrometry

Article abstract of DOI:10.1002/jms.402

Mass spectrometric methods have been developed which allow the direct stereochemical assignment of terminal monosaccharide residues in flavonoid O-glycosides without the need for chemical hydrolysis. Standards containing a glucose, galactose, mannose, xylose, arabinose or apiose residue were examined because these monosaccharides are by far the most commonly encountered in flavonoid glycosides. Following acetylation, the major peracetylated sugar related fragments, generated by fast atom bombardment (FAB) or electrospray ionization (ESI), were selected for collisional activation employing a broad range of collision energies. Both FAB and ESI proved to be useful as ionization techniques. Stereoselective fragmentation was achieved and allowed us clearly to differentiate and characterize isomeric monosaccharide residues. The method developed was successfully applied to an unknown flavonoid containing a terminal pentose and hexose residue which was isolated from Farsetia aegyptia. Copyright

Full text of DOI:10.1002/jms.402

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2002; 37: 1272–1279
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jms.402
Direct stereochemical assignment of hexose and
pentose residues in flavonoid O-glycosides by fast atom
bombardment and electrospray ionization mass
spectrometry†
Filip Cuyckens, Abdelaaty A. Shahat, Luc Pieters and Magda Claeys∗
Department of Pharmaceutical Sciences, University of Antwerp (UIA), Universiteitsplein 1, B-2610 Antwerp, Belgium
Received 21 May 2002; Accepted 24 October 2002
Mass spectrometric methods have been developed which allow the direct stereochemical assignment of
terminal monosaccharide residues in flavonoid O-glycosides without the need for chemical hydrolysis.
Standards containing a glucose, galactose, mannose, xylose, arabinose or apiose residue were examined
because these monosaccharides are by far the most commonly encountered in flavonoid glycosides.
Following acetylation, the major peracetylated sugar related fragments, generated by fast atom
bombardment (FAB) or electrospray ionization (ESI), were selected for collisional activation employing
a broad range of collision energies. Both FAB and ESI proved to be useful as ionization techniques.
Stereoselective fragmentation was achieved and allowed us clearly to differentiate and characterize
isomeric monosaccharide residues. The method developed was successfully applied to an unknown
flavonoid containing a terminal pentose and hexose residue which was isolated from Farsetia aegyptia.
Copyright  2002 John Wiley & Sons, Ltd.
KEYWORDS: flavonoid glycosides; sugar residues; electrospray ionization; fast atom bombardment
carbohydrates present (mono-, di-, tri- or tetrasaccharides
and hexoses, deoxyhexoses or pentoses)9–11 and the type
of interglycosidic linkages.12–15 Early investigations of car-
bohydrates by means of MS analysis were performed by
resorting to acetyl and methyl derivatives.16,17 Derivatization
of the compounds was required in order to obtain sufficient
volatility necessary for electron ionization (EI) and chemical
ionization (CI). Although more modern soft ionization tech-
niques such as fast atom bombardment (FAB)18–20 allow the
detection of molecular ion species (i.e. protonated, depro-
tonated or sodiated species) of underivatized saccharides,
Richter and co-workers21–24 showed that acetylation can still
be useful in the structural characterization of oligosaccha-
rides and glycoside residues. It was demonstrated that the
differentiation between a terminal glucose, galactose and
mannose residue of naturally occurring glycosides could be
achieved by low-energy collision-induced dissociation (CID)
using a triple-quadrupole mass spectrometer equipped with
a CI source. In this study, we followed the same approach for
developing methodology to differentiate and characterize the
terminal monosaccharide residue of flavonoid O-glycosides.
Standards containing the major hexoses and pentoses found
attached to flavonoids were analyzed utilizing FAB and elec-
trospray ionization (ESI). Deoxyhexoses were not included in
this study because, to our knowledge, only rhamnose occurs
in association with flavonoids and can easily be distinguished
from pentoses and hexoses by its different mass. Finally, the
monosaccharide residues of an unknown flavonoid, isolated
INTRODUCTION
Flavonoid glycosides are predominant forms of naturally
occurring flavonoids in plants, representing a large group of
secondary plant metabolites. They all contain a C15 skeleton
as an aglycone, are usually divided into O- and C-glycosyl
flavonoids and are of interest because they have various
biological activities, are useful for chemotaxonomy and are
used as tracers in medicinal plant preparations.1,2 Flavonoids
generally occur as glycosides in plants because the effect of
glycosylation renders the flavonoid less reactive and more
water soluble, permitting storage of the flavonoids in the
cell vacuole where they are mostly found. Glucose is by far
the monosaccharide most commonly bound to flavonoids,
whereas galactose, rhamnose, xylose and arabinose are not
uncommon and mannose and apiose are also occasionally
encountered.1 Although 1H and 13C NMR techniques are
preferred for stereochemical assignment, mass spectrometry
(MS) can make a valuable contribution because it can be
applied to much smaller quantities of isolated compounds.
It provides the molecular mass, in addition to structural
information on the flavonoid skeleton,3–15 attachment points
of carbohydrate residues to the aglycone,6–8 the types of
ŁCorrespondence to: Magda Claeys, Department of Pharmaceutical
Sciences, University of Antwerp (UIA), Universiteitsplein 1, B-2610
Antwerp, Belgium. E-mail: claeys@uia.ua.ac.be
†Paper presented at the 20th Informal Meeting on Mass
Spectrometry, Fiera di Primiero, Italy, 12–16 May 2002.
Copyright  2002 John Wiley & Sons, Ltd.

Hexose and pentose assignment in flavonoids 1273
from Farsetia aegyptia, containing a terminal pentose and
hexose, were characterized using the method developed.
400 or 800 V and helium, methane or xenon was used as the
collision gas until the precursor ion beam was reduced to
50% of its original value. Using this approach, a wide range
of collision energies could be used.
EXPERIMENTAL
ESI spectra were acquired on the same instrument.
The ESI source was operated at 4 kV. The acetylated
compounds were dissolved in water–methanol (1 : 1, v/v) at
a concentration of 100 µM. They were introduced into the ESI
source by a syringe pump (Model 22 syringe infusion pump,
Harvard Apparatus, South Natick, MA, USA) employing a
500 µl syringe (Hamilton, Reno, NV, USA) at a constant flow-
Materials
Quercetin-3-O-ˇ-D-glucopyranoside (isoquercitrin), quer-
cetin-3-O-ˇ-D-galactopyranoside (hyperoside), quercetin-40-
O-ˇ-D-glucopyranoside (spiraeoside), syringetin-3-O-ˇ-D-
glucopyranoside,
syringetin-3-O-ˇ-D-galactopyranoside,
quercetin-3-O-˛-L-arabinopyranosyl-(1 ! 6)-ˇ-D-glucopyra-
noside (peltatoside) and apigenin-7-O-ˇ-D-apiofuranosyl-
(1 ! 2)-ˇ-D-glucopyranoside (apiin) were obtained from
Extrasynthe`se (Genay, France). Because flavonoids con-
taining a xylose or mannose residue are commercially
unavailable, 4-O-ˇ-D-xylopyranosyl-ˇ-D-xylopyranose (xy-
lobiose) and 3-O-˛-D-mannopyranosyl-˛-D-mannopyranose
(mannobiose) were used instead. They were both purchased
from Sigma (St. Louis, MO, USA). The structures of the
monosaccharide residues examined are shown in Fig. 1. All
compounds were acetylated overnight at room temperature
with pyridine–acetic anhydride (1 : 1, v/v). The unknown
flavonoid was isolated from the leaves of Farsetia aegyptia
(full details will be reported elsewhere) and its complete
structure was elucidated by the combined use of NMR and
MS methods.
rate of 5 µl minꢀ1. Nitrogen was used as both bath gas (80 C)
°
and nebulizing gas. The effect of the cone voltage on the
ionization was examined. Product ion tandem mass spectra
using low- and high-energy CID were recorded as done in the
case of ionization by FAB. Data acquisition and processing
were performed using OPUS V3.1X software. All scans
were acquired in the continuum mode. The nomenclature
proposed by Domon and Costello25 for glycoconjugates was
adopted to denote the fragment ions. If the charge is retained
on the carbohydrate residue, fragments are designated as
k,lAi, Bi and Ci, where i represents the number of the
glycosidic bond cleaved counting from the non-reducing
terminus and the superscripts k and l indicate the cleavages
within the carbohydrate rings.
RESULTS AND DISCUSSION
Mass spectrometry
Saccharidic fragment ions are rarely seen in first-order spec-
tra of underivatized flavonoids because the charge generally
stays associated with the aglycone part.9–15 By acetylating
flavonoid O-glycosides, peracetylated sugar related frag-
ment ions are also detected with a high relative abundance
and can readily be selected for further CID experiments. The
fragmentation pathways of the peracetylated hexose and
pentose residues are shown in Schemes 1 and 2, respectively.
They relate mainly to the loss of acetic acid (60 u) and of
ketene (42 u). The latter process seems to be facilitated if pre-
ceded by loss of acetic acid and if the acetyl groups involved
Mass spectrometric data were obtained on an Autospec-
oa-ToF mass spectrometer (Micromass, Manchester, UK)
with an EBE-oa-time-of-flight configuration equipped with a
cesium ion source. The acronym FAB is used throughout to
refer to cesium ion bombardment. The acceleration voltage
in the source was 8 kV. The samples were dissolved in
methanol (10 µg µlꢀ1) and 1 µl of the solution was mixed
with 2 µl of the liquid matrix on the stainless-steel probe.
Glycerol and 3-nitrobenzyl alcohol were tested as matrices.
Glycerol resulted in a better signal-to-noise ratio and was
employed in subsequent experiments. High-energy CID
spectra were obtained using linked scanning at constant
B/E with helium as collision gas until 50% attenuation of
the precursor ion beam. In the tandem mass spectrometric
(MS/MS) mode precursor ions were selected by MS1 (EBE
configuration) and, after passing the collision cell, product
ions were recorded on the microchannel plate detector of the
time-of-flight analyzer. The collision cell was floated at 200,
are in a 1,2- or a 1,3-relationship.16,26,27
.
AcOCH2
O
AcO
OAc
AcO
m/z 331
- CH3COOH
- CH2CO
CH OH
CH OH
2
CH OH
2
2
O
O
OH
O
- CH2CO
- CH2CO
m/z 271
m/z 229
m/z 187
m/z 145
m/z 99
OH
OH
OH
OH
OR
OR
OR
OH
OH
- CH3COOH
- CH2CO
- CH3COOH
- CH2CO
- CH3COOH
- CO
OH
OH
D-glucopyranosyl
D-mannopyranosyl
D-galactopyranosyl
m/z 211
m/z 169
m/z 127
O
OH
O
O
OR
- CH3COOH
- CO
CH OH
2
OH
OR
OH
OR
OH
m/z 109
m/z 81
OH
OH
OH
OH
D-xylopyranosyl
D-apiofuranosyl
L-arabinopyranosyl
C
Scheme 1. Fragmentation pathways of the B1 ion of per-
Figure 1. Structures of the monosaccharides studied.
Copyright  2002 John Wiley & Sons, Ltd.
acetylated hexose-containing compounds.
J. Mass Spectrom. 2002; 37: 1272–1279

1274
F. Cuyckens et al.
O
AcO
AcO
OAc
m/z 259
- CH3COOH
- CH2CO
- CH2CO
m/z 199
m/z 157
m/z 115
m/z 69
- CH3COOH
- CH3COOH
- CH2CO
- CO
- CO
m/z 139
m/z 41
m/z 97
C
Scheme 2. Fragmentation pathways of the B1 ion of per-
acetylated pentose-containing compounds.
Hexoses
Fast atom bombardment
The first-order FAB spectra of the peracetylated flavonoid-
glycosides examined reveal, along with the peracetylated
sugar fragment ion (B1C), nearly all major sugar fragments
listed in Scheme 1. Although most of them have a high
relative abundance, we found it almost impossible to
distinguish the different hexose residues in a reproducible
way. Therefore, CID experiments were performed on all the
major sugar fragments to reveal differences in fragmentation
behavior between the three hexose residues investigated.
According to Richter and co-workers,21 the ease of
differentiation between hexose residues depends on the
collision energy employed and an optimum for distinction
is readily attained at lower collision energies. Product ion
tandem mass spectra were recorded at different collision
energies by changing the collision gas and the voltage of the
collision cell. Selection of the peracetylated hexose fragment
(B1C) at m/z 331 did not result in the expected difference in
[m/z 127]/[m/z 109] ratios as described by Richter and co-
workers.21 However, a [B1 ꢀ CH3COOH]C fragment ion at
m/z 271 was visible in all the spectra of peracetylated glucose
and mannose derivatives independently of the collision
energy used, whereas it was hardly seen, if at all, in the
spectra of peracetylated galactose derivatives. The intensity
of the peak at m/z 271 exceeded that at m/z 229 for glucose
and mannose, whereas the peak at m/z 229 was more intense
in the case of a galactose residue. The differentiation based on
the relative abundance of the m/z 271 fragment ion is even
more straightforward using B/E linked scanning, where
[B1 ꢀ CH3COOH]C at m/z 271 is the most abundant product
ion for glucose, whereas it is a minor product for galactose
and has an intermediate relative abundance for mannose
(Fig. 2).
C
Figure 2. High-energy CID spectra of the peracetylated B1
ion of (a) glucose-, (b) galactose- and (c) mannose-containing
compounds (isoquercitrin, hyperoside and mannobiose,
respectively) generated by FAB using linked scanning at
constant B/E and helium as collision gas.
ratios are obtained for the whole range of collision energies
applied, indicating that this method is robust.
Another way to differentiate clearly between hex-
ose residues can easily be made by CID of the [B1 ꢀ
2CH3COOH ꢀ 2CH2CO]C ion at m/z 169. Figure 3 illustrates
the low-energy CID spectra obtained for the monosaccha-
ride fragment at m/z 169 using helium as collision gas. The
[B1 ꢀ 2CH3COOH ꢀ 2CH2CO]C ion at m/z 127 has a higher
relative abundance than the [B1 ꢀ 3CH3COOH ꢀ CH2CO]C
ion at m/z 109 for galactose, whereas the opposite holds for
a mannose and glucose residue. The same peak intensity
The spectra obtained for the BC1 and [B1 ꢀ 2CH3COOH ꢀ
2CH2CO]C ions of peracetylated quercetin-40-O-ˇ-D-glu-
copyranoside, syringetin-3-O-ˇ-D-glucopyranoside and
syringetin-3-O-ˇ-D-galactopyranoside were similar to those
illustrated for their glucose and galactose analogues,
quercetin-3-O-ˇ-D-glucopyranoside and quercetin-3-O-ˇ-D-
galactopyranoside (Figs 2 and 3).
Copyright  2002 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 1272–1279

Hexose and pentose assignment in flavonoids 1275
comparison with galactose (Fig. 4(b)) are consistent with the
behavior discussed previously for FAB. The ESI spectrum
illustrated in Fig. 4(c) for peracetylated mannose shows an
intermediate behavior that is closer to that of galactose.
This technique should be used with caution because the
relative abundance of the ions can change considerably
with increasing cone voltage. In our experience, higher cone
voltages (200–250 V) are to be preferred because they result
in the clearest differentiation and the highest sensitivity and
reproducibility.
Figure 5 illustrates the low-energy CID spectra of per-
acetylated B1C. The loss of one acetic acid residue is relatively
Figure 3. Low-energy CID spectra obtained for the
[B1 ꢀ 2CH3COOH ꢀ 2CH2CO]C fragment at m/z 169 of (a)
glucose-, (b) galactose- and (c) mannose-containing
compounds (isoquercitrin, hyperoside and mannobiose,
respectively) generated by FAB.
Electrospray ionization
ESI mass spectra were examined over the full range ofCcone
voltages (0–250 V). Above a cone voltage of 75 V, a B1 ion
at m/z 331 becomes clearly visible. On further increasing
the cone voltage, CID-like processes occur in the ion source,
giving rise to all major sugar related fragments and are as
such providing structure information. Figure 4 shows the
lower mass range of the first-order ESI spectra acquired at
a cone voltage of 200 V. The higher relative abundance of
the ion at m/z 109 compared with that at m/z 127 and the
more abundant fragment at m/z 271 for glucose (Fig. 4(a)) in
Figure 4. Lower mass range of the first-order ESI spectra of
peracetylated (a) isoquercitrin, (b) hyperoside and (c)
mannobiose acquired at higher cone voltage (200 V).
Copyright  2002 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 1272–1279

1276
F. Cuyckens et al.
more pronounced for glucose, resulting in an ion with higher
relative abundance at m/z 271 compared with that at m/z
229, whereas the opposite holds for galactose and mannose.
Hence, as noted in the first-order ESI spectra, the fragmen-
tation of mannose follows more closely that of galactose
and is in contrast with observations in FAB. However, the
peak intensity ratios between [m/z 271]/[m/z 229] and [m/z
271]/[m/z 211] differ sufficiently to distinguish both hexose
residues. The values were 4.5 š 0.6 ([m/z 229]/[m/z 271])
and 10.5 š 0.9 ([m/z 211]/[m/z 271]) for galactose, whereas
they were 1.5 š 0.3 ([m/z 229]/[m/z 271]) and 3.1 š 0.3 ([m/z
211]/[m/z 271]) for mannose.
Product ion tandem mass spectra of the [B1 ꢀ 2CH3
COOH ꢀ 2CH2CO]C ion at m/z 169 were also obtained
(results not shown). Whereas in FAB the m/z 109 ion was the
most abundant fragment for glucose and mannose (Fig. 3),
the m/z 127 ion was always much more abundant than the
m/z 109 ion in the product ion tandem mass spectra of the
m/z 169 ion generated in ESI. Different [m/z 127]/[m/z 109]
peak intensity ratios were also obtained in ESI for the three
epimers studied, but these differences are rather small; the
relative abundance of the m/z 127 ion increases slightly with
increasing cone voltage, making the determination of the
hexose residue uncertain.
Figure 5. Low-energy CID spectra obtained for the
C
peracetylated B1 ion of (a) isoquercitrin, (b) hyperoside and (c)
Figure 6. Lower mass range of the first-order FAB spectra of
mannobiose generated by ESI using helium as collision gas.
peracetylated (a) xylobiose, (b) peltatoside and (c) apiin.
Copyright  2002 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 1272–1279

Hexose and pentose assignment in flavonoids 1277
abundance than the m/z 199 ion in the spectrum of arabinose
compared with xylose where the m/z 139 peak has a
negligible relative abundance.
The first-order FAB spectra of peltatoside and apiin
show additional fragments at m/z 109, 127, 169 and 271
(Fig. 6). These hexose-related fragments originate from the
internal glucose present in the disaccharide moiety. The m/z
169 product ion spectrum (not shown) revealed an [m/z
127]/[m/z 109] ratio identical with that shown in Fig. 3(a)
for the m/z 169 fragment of peracetylated isoquercitrin.
Pentoses
Fast atom bombardment
Figure 6 illustrates the lower mass range of the first-
order FAB spectra of the peracetylated pentose-containing
compounds, xylobiose, peltatoside and apiin. The relative
abundance of the [B1 ꢀ 2CH3COOH]C ion at m/z 139 can
be compared with that of the m/z 97, 157, 199 and 259 ions
to differentiate and characterize the three pentose residues
(Table 1). Especially the [m/z 139]/[m/z 199] peak intensity
ratios are strikingly different. The relative abundance of the
m/z 199 ion [B1 ꢀ CH3COOH]C follows the order xylose >
arabinose > apiose. The more abundant [B1 ꢀ CH3COOH]C
ion for xylobiose compared with peltatoside is in full
agreement with observations made for the hexoses (m/z
271) and as expected because the structure of xylose is
stereochemically similar to that of glucose whereas arabinose
is similar to galactose (Fig. 1). It has been reported by Blok-
Tip et al.28 that the decomposition of the [B1 ꢀ CH3COOH]C
ion proceeds to a greater extent for ribofuranose than
ribopyranose. This could also partly explain the different
ratios observed because apiose is in the furanose form
whereas xylose and arabinose are in the pyranose form.
Another characteristic of the apiose derivative apiin is a
[B1 ꢀ CH3COOH ꢀ CH2CO ꢀ CO]C ion at m/z 129 that was
absent for the xylose- and arabinose-containing compounds,
xylobiose and peltatoside.
Electrospray ionization
At cone voltages exceeding 100 V the B1 ion at m/z 259
C
appears, accompanied by other sugar-related fragments.
A further increase in the cone voltage reveals all major
fragments. Apiose can easily be recognized because only the
[B1 ꢀ 2CH3COOH]C fragment at m/z 139 can be detected. For
peracetylated xylose and arabinose also fragments at m/z 157
and 199 can be found but their peak intensity ratios fluctuate
considerably when the cone voltage increases, making the
distinction between the two epimers hard to accomplish.
Figure 7 shows the low-energy CID spectra obtained
C
for the peracetylated B1 ion at m/z 259 using helium
as collision gas. Again, apiose can clearly be identified
because the fragment at m/z 199 is virtually absent and
the signal intensity of m/z 157 low ([m/z 139]/[m/z 157]
D 13.3 š 0.5). Another characteristic is that an additional
[B1 ꢀ CH3COOH ꢀ CH2CO ꢀ CO]C fragment at m/z 129
can only be seen in the spectrum of peracetylated apiose.
The relative abundances of the fragments at m/z 139, 157
and 199 can be used to differentiate between an arabinose
([m/z 139]/[m/z 199] D 10.4 š 2.0; [m/z 139]/[m/z 157]
D 3.7 š 0.8) and a xylose ([m/z 139]/[m/z 199] D 3.5 š 0.3;
[m/z 139]/[m/z 157] D 1.6 š 0.1) residue. The values given
here were obtained using helium as collision gas, but similar
results were also found with xenon.
The same observations were made for the product ion
tandem mass spectra of the peracetylated B1 ion at m/z
259 (Table 1). The product ion tandem mass spectra are
very similar to the spectra for the B1 ion generated by ESI
C
C
(Fig. 7). Both low- and high-energy CID can be used but the
collision energy should be kept constant because the relative
abundances and peak intensity ratios change with increasing
collision energy.
Similar tendencies can also be noted using linked
scanning at constant B/E value (results not shown). The [B1 ꢀ
CH3COOH]C fragment is absent for the apiose derivative
apiin, whereas it is the major fragment for the xylose- and
arabinose-containing compounds xylobiose and peltatoside.
The fragments at m/z 139 and 157 have a larger relative
Application to an unknown flavonoid isolated
from Farsetia aegyptia
An unknown flavonoid, isolated from the leaves of Farse-
tia aegyptia, was characterized by mass spectrometric
Table 1. Mass spectral data obtained by FAB: relative abundances were obtained by averaging the values from three different
analyses on three different days
m/z
m/z
m/z
m/z
Spectra Compound m/z 259
m/z 199
m/z 157
m/z 139
m/z 97
139/259
139/199
139/157
139/97
FABa
Heb
Xec
Xylose
Arabinose
Apiose
Xylose
Arabinose
Apiose
Xylose
Arabinose
apiose
55.3 š 6.8 40.2 š 5.6 76.7 š 5.3
81.7 š 7.6 18.1 š 0.9 58.1 š 1.1
93.3 š 7.0
100
99.6 š 0.7 1.7 š 0.1
91.9 š 6.5 1.2 š 0.1
2.3 š 0.2
5.6 š 0.3
1.2 š 0.1
1.7 š 0.1
7.0 š 1.1
2.1 š 0.1
3.9 š 0.8
9.1 š 1.8
2.1 š 0.2
2.8 š 0.5
0.9 š 0.1
1.1 š 0.1
2.7 š 0.5
0.7 š 0.1
1.1 š 0.4
2.5 š 0.4
1.2 š 0.2
1.7 š 0.1
4.9 š 0.2
35.7 š 8.1
100
5.8 š 2.6 14.6 š 2.3
16.5 š 4.6 25.2 š 2.4
12.1 š 4.1 17.9 š 0.9
100
38.2 š 7.9 2.9 š 0.6 20.8 š 12.3
51.5 š 4.2
69.2 š 5.0 0.5 š 0.0
3.2 š 0.6
6.0 š 1.1
100
69.8 š 13.8 65.6 š 10.7 0.7 š 0.1
100
100
100
100
0.8 š 0.2
6.7 š 1.2
59.2 š 3.4
21.0 š 0.3
19.7 š 1.6
26.1 š 1.5
24.0 š 3.5 0.6 š 0.0 79.5 š 15.8
12.3 š 2.1 10.3 š 1.0
18.2 š 2.9 0.2 š 0.0
12.1 š 0.9 0.2 š 0.0
1.6 š 0.4
3.4 š 0.8
6.1 š 1.3
7.2 š 1.3
1.3 š 0.7
1.4 š 0.4
5.4 š 0.2 0.3 š 0.0 26.1 š 16.2 19.7 š 6.8
a FAB: first-order FAB spectra of the peracetylated flavonoid O-glycosides.
C
b He: low-energy CID spectra of peracetylated B1 C using helium as collision gas.
c Xe: high-energy CID spectra of peracetylated B1 using xenon as collision gas.
Copyright  2002 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 1272–1279

1278
F. Cuyckens et al.
an arabinose residue identity could be inferred. The values
were 12.0 š 6.6 ([m/z 139]/[m/z 199]) and 3.3 š 0.3 ([m/z
139]/[m/z 157]) using helium as collision gas and 6.8 š 1.7
([m/z 139]/[m/z 199]) and 2.7 š 1.2 ([m/z 139]/[m/z 157])
using xenon as collision gas. All values correspond within the
standard deviation of those obtained for arabinose. Taking
these results into account, the structure of the unknown
flavonoid was defined as isorhamnetin-3-O-arabinoside-
7-O-glucosylrhamnoside. Subsequently, the structure was
confirmed and completed (i.e. with a 1,2-linkage between
the rhamnose and glucose residues) by NMR spectroscopy.
CONCLUSIONS
Both FAB and ESI in combination with CID have been
shown to be useful for the direct stereochemical assignment
of hexose and pentose residues in acetylated flavonoid
O-glycosides. The differentiation between a glucose and
a galactose residue can easily be made by employing
the [m/z 127]/[m/z 109] peak intensity ratios and the
relative abundance of the [B1 ꢀ CH3COOH]C fragment at
m/z 271. Mannose shows intermediate relative abundances
and behaves in FAB more like glucose, whereas in ESI the
behavior is more similar to that observed for galactose. For
the three pentoses considered in this study, stereoselective
fragmentation was observed in the first-order FAB spectra
anCd in the product ion CID spectra of the peracetylated
B1 ion at m/z 259 generated by FAB or ESI. The
[B1 ꢀ 2CH3COOH]C fragment at m/z 139 revealed a different
relative abundance to those of the other sugar fragments
for the pentoses examined. Especially the [m/z 139]/[m/z
199] peak intensity ratios are strikingly different, showing
the order apiose > arabinose > xylose. Apiose can also
be recognized by the presence of an additional [B1 ꢀ
CH3COOH ꢀ CH2CO ꢀ CO]C fragment at m/z 129.
The methodology developed in this study is not restricted
to flavonoid O-glycosides and has also been successfully
applied to non-flavonoids (i.e. phytosterol O-glycosides), but
it should be mentioned that other monosaccharide residues
not included in this study can be encountered.
Acknowledgements
We acknowledge the financial support by the Fund for Scientific
Research (Belgium–Flanders; FWO; grant No. 6.0082.98) and the
Concerted Actions of the Regional Government of Flanders (grant
No. 99/3/34). F. Cuyckens thanks the Vlaams Instituut voor de
Bevordering van het Wetenschappelijk-Technologisch Onderzoek
in de Industrie (IWT) for a doctoral fellowship.
Figure 7. Low-energy CID spectra obtained for the
peracetylated B1 ion of (a) xylobiose, (b) peltatoside and (c)
C
apiin generated by ESI using helium as collision gas.
approaches as isorhamnetin-3-O-pentoside-7-O-hexosyl-
rhamnoside. The compound was peracetylated and the
methods developed in this work were applied to define the
pentose and hexose residues. The peracetylated monosaccha-
ride ions (B1C) at m/z 259 for the pentose and at m/z 331 for
the hexose were clearly visible in the first-order ESI spectrum.
REFERENCES
1. Markham KR. Techniques of Flavonoid Identification. Academic
Press: London, 1982.
2. Harborne JB. The Flavonoids: Advances in Research Since 1986.
Chapman and Hall: London, 1994.
3. Ma YL, Li QM, Van den Heuvel H, Claeys M. Rapid Commun.
Mass Spectrom. 1997; 11: 1357.
C
The low-energy CID spectrum of peracetylated B1 at m/z
4. Wolfender JL, Waridel P, Ndjoko K, Hobby KR, Major HJ,
Hostettmann K. Analusis 2000; 28: 895.
5. Fabre N, Rustan I, de Hoffmann E, Quetin-Leclercq J. J. Am. Soc.
Mass Spectrom. 2001; 12: 707.
6. Sakushima A, Nishibe S, Takeda T, Ogihara Y. Mass Spectroscopy
1988; 36: 71.
331 revealed a glucose structure for the hexose because the
peak intensity of the [B1 ꢀ CH3COOH]C ion at m/z 271 was
significantly higher than that of the ion at m/z 229 (see Fig. 5).
From the peak intensity ratios observed in the low- and high-
C
energy CID spectra of the peracetylated B1 ion at m/z 259,
Copyright  2002 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 1272–1279

Hexose and pentose assignment in flavonoids 1279
7. Becchi M, Fraisse D. Biomed. Mass Spectrom. 1989; 18: 122.
8. Li QM, Van den Heuvel H, Dillen L, Claeys M. Biol. Mass
Spectrom. 1992; 21: 213.
9. Sakushima A, Nishibe S. Phytochemistry 1988; 27: 915.
10. Sa¨gesser M, Deinzer M. J. Am. Soc. Brew. Chem. 1996; 54: 129.
11. Ha¨kkinen S, Auriola S. J. Chromatogr. A 1998; 829: 91.
12. Careri M, Elviri L, Mangia A. Rapid Commun. Mass Spectrom.
1999; 13: 2399.
19. Garozzo D, Giuffrida M, Impallomeni G, Ballistreri A, Mon-
taudo G. Anal. Chem. 1990; 62: 279.
20. Dallinga JW, Heerma W. Biol. Mass Spectrom. 1991; 20:
215.
21. Mu¨ller DR, Domon B, Blum W, Raschdorf F, Richter WJ. Biomed.
Environ. Mass Spectrom. 1988; 15: 441.
22. Domon B, Mu¨ller DR, Richter WJ. Org. Mass Spectrom. 1989; 24:
357.
13. Ma YL, Cuyckens F, Van den Heuvel H, Claeys M. Phytochem.
Anal. 2001; 12: 159.
23. Domon B, Mu¨ller DR, Richter WJ. Biomed. Environ. Mass
Spectrom. 1990; 19: 390.
14. Satterfield M, Brodbelt JS. J. Am. Soc. Mass Spectrom. 2001; 12:
537.
15. Cuyckens F, Rozenberg R, de Hoffmann E, Claeys M. J. Mass
Spectrom. 2001; 36: 1203.
16. Biemann K, De Jongh DC, Schnoes KH. J. Am. Chem. Soc. 1963;
85: 1763.
17. De Jongh DC, Biemann K. J. Am. Chem. Soc. 1963; 85: 2289.
18. Dallinga JW, Heerma W. Biomed. Environ. Mass Spectrom. 1989;
18: 363.
24. Vouros P, Mu¨ller DR, Richter WJ. J. Mass Spectrom. 1999; 34: 346.
25. Domon B, Costello CE. Glycoconj. J. 1988; 5: 397.
26. Radford T, DeJongh DC. In Biochemical Applications in Mass
Spectrometry, Waller GR (ed.). Wiley: New York, 1972;
313.
27. Kova´cˇik V, Petra´kova´ E, Miha´lov V, Tvarosˇka I, Heerma W.
Biomed. Mass Spectrom. 1985; 2: 49.
28. Blok-Tip L, Heerma W, Haverkamp J. Org. Mass Spectrom. 1993;
28: 139.
Copyright  2002 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2002; 37: 1272–1279