Identification of isomers of resveratrol dimer and their analogues from wine grapes by HPLC/MSn and HPLC/DAD-UV

Article abstract of DOI:10.1016/j.foodchem.2010.12.133

A facile method based on HPLC/(-)ESI-MSⁿ is established for the analysis of seven isomers of resveratrol dimers and three of their analogues in Xinjiang wine grapes. The structures of these compounds are positively or tentatively determined. Among them, three are tentatively identified as new compounds. MSⁿ experiments on the [M-H]^- ions provide abundant structural information, especially regarding the relative abundance of the key product ions, m/z 333 and 369 (385 in compound 3), which can be utilised to distinguish whether or not the compound identified contains the scaffold of the isomer of a resveratrol dimer. The relative abundance of key product ions remains unchanged as collision energy varies from 0.60 to 0.95 V. All the trans-, and cis-isomers could be identified by HPLC/DAD-UV spectra. The UV spectra of compounds 2 and 9 tentatively show cis and trans- configurations, respectively.

Full text of DOI:10.1016/j.foodchem.2010.12.133

Food Chemistry 127 (2011) 727–734
Contents lists available at ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
Identification of isomers of resveratrol dimer and their analogues from wine
grapes by HPLC/MSn and HPLC/DAD-UV
Q.J. Kong a,b, X.Y. Ren b, N. Hu a, C.R. Sun a, Y.J. Pan a,
⇑
a Department of Chemistry, Zhejiang University, Hangzhou 310027, China
b College of Life Sciences, Shi Hezi University, Shi Hezi 832003, China
a r t i c l e i n f o
a b s t r a c t
A facile method based on HPLC/(ꢀ)ESI-MSn is established for the analysis of seven isomers of resveratrol
dimers and three of their analogues in Xinjiang wine grapes. The structures of these compounds are pos-
itively or tentatively determined. Among them, three are tentatively identified as new compounds. MSn
experiments on the [MꢀH]ꢀ ions provide abundant structural information, especially regarding the rel-
ative abundance of the key product ions, m/z 333 and 369 (385 in compound 3), which can be utilised to
distinguish whether or not the compound identified contains the scaffold of the isomer of a resveratrol
dimer. The relative abundance of key product ions remains unchanged as collision energy varies from
0.60 to 0.95 V. All the trans-, and cis-isomers could be identified by HPLC/DAD-UV spectra. The UV spectra
of compounds 2 and 9 tentatively show cis and trans- configurations, respectively.
Article history:
Received 18 May 2010
Received in revised form 26 October 2010
Accepted 31 December 2010
Available online 8 January 2011
Keywords:
Isomers of resveratrol dimer
HPLC/(ꢀ)ESI-MSn
Wine grapes
Ó 2011 Elsevier Ltd. All rights reserved.
HPLC/DAD-UV
Relative abundance
1. Introduction
formulas. A combination of these two methods may facilitate the
identification of components of wine grape extracts.
Resveratrol dimers and their analogues belong to a group of
compounds called stilbenes, which, along with anthocyanins, cate-
chins, proanthocyanidins, and flavonols, are natural phenolic com-
pounds found in grapes and red wine. Stilbenes have a variety of
structural isomers, all of which exhibit various biological proper-
ties, such as potent antioxidative properties (He, Jiang, Wu, Pan,
& Sun, 2009; He, Wu, Pan, & Jiang, 2008; Kong, Ren, Jiang, Pan, &
Sun, 2009; Renaud & Delorgeril, 1992; Zern & Fernández, 2005),
platelet aggregation inhibition (Van Wauwe & Goossens, 1983),
and antimicrobial (Palma, Taylor, Varela, Cutler, & Cutler, 1999)
and anti-ageing activities (Shukitt-Hale, Carey, Simon, Mark, & Jo-
seph, 2006). Thus, the identification of bioactive resveratrol dimers
and their analogues from wine grapes is of great importance.
HPLC/ESI-MSn features high sensitivity, low levels of sample
consumption, relatively short analysis times, and results of rich
structural information. It has the ability to rapidly screen minor
constituents in plant extracts that are difficult to obtain through
conventional phytochemical means (Chen, Li, Sun, Pan, & Schlu-
negger, 2008; Huang, Song, Liu, & Liu, 2007; Tong et al., 2008;
Wu, Wang, & Simon, 2003). It, however, cannot provide accurate
masses of precursor and fragment, Fourier transform ion cyclotron
ion resonance tandem mass spectrometry (FTICR-MS/MS) is used
to identify compounds and provide information on the molecular
To the best of our knowledge, no study has yet been conducted
regarding the identification of the isomers of resveratrol dimers
and their analogues from Xinjiang wine grapes. This study thus
aims to identify these compounds using HPLC/ESI-MSn, FTICR-
MS/MS and HPLC/DAD-UV.
2. Materials and methods
2.1. Materials
The stems of wine grapes (Vitis Vinifera) were collected in Au-
gust 2008 in Shihezi, Xinjiang Province, China. The material was
identified by Dr. Yongxiang Cheng (Shihezi University, Shihezi, Chi-
na). Methanol (chromatographic grade) used for analytical HPLC
was purchased from Merck (Darmstadt, Germany). Deionized
water (18 MX) was obtained from a Milli-Q water purification sys-
tem (Millipore, Bedford, MA, USA).
2.2. Sample preparation
The dried stems (4 kg) of wine grapes (Vitis Vinifera) were ex-
tracted three times with MeOH (15.0 L ꢁ 3) at room temperature.
The solvent was evaporated in vacuo to afford a concentrated
MeOH extract (350 g), which was then diluted with H2O (2.0 L)
to give the aqueous solution (2.0 L). The aqueous solution was ex-
tracted with ethyl acetate (4.0 L ꢁ 3). The combined ethyl acetate
⇑
Corresponding author. Tel.: +86 571 87953000; fax: +86 571 87951629.
E-mail address: cheyjpan@zju.edu.cn (Y.J. Pan).
0308-8146/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2010.12.133

728
Q.J. Kong et al. / Food Chemistry 127 (2011) 727–734
layers were concentrated to dryness in vacuo to provide an ethyl
acetate extract (92 g), which was subjected to silica gel CC
(1000 g, 5 cm diameter) eluted with light petroleum–ethyl acetate
mixtures (100:1–1:10) yielding 21 fractions. Fraction 11 (1.3 g)
was separated by RP-18 CC (MeOH–H2O, 50:50%) to give com-
which provides extensive structural information, the negative ion
mode of ESI was selected. Fig. 2 shows the structures of the resve-
ratrol dimer and its analogous compounds. The details of each
compounds are further described in the following section.
pounds scirpusin A (3) (18 mg), trans-
e
-viniferin (5) (28.1 mg)
3.1. Identification of standards compounds 3–5, 8 and 10
and trans-d-viniferin (9) (10 mg). Compounds (3, 5, 8) were identi-
fied by direct comparison of their 1H NMR, 13C NMR with those re-
ported in the literatures, respectively (Lin, Li, Li, Yu, & Liang, 1992;
Compounds 4, 5, 8 and 10, all of which are isomers of resvera-
trol dimers, were identified as cis-e-viniferin, trans-e-viniferin,
Pezet et al., 2003; Shao et al., 2007). Cis-
in(10) were obtained after sunlight exposition of methanolic solu-
tions of trans- -viniferin, trans-d-viniferin, respectively, and used
e
-viniferin(4), cis-d-vinifer-
trans-d-viniferin and cis-d-viniferin, respectively. Compound 3, an
analogues of a resveratrol dimer, was identified as scirpusin A.
Two key product ions (m/z 333 and 369) were observed and con-
sidered as characteristics for the structural identification of resve-
ratrol dimers and their analogues. The fragmentations of these
reference standard compounds are discussed below in detail.
In the MS/MS spectra of the [MꢀH]ꢀ ion at m/z 469 of compound
3, the losses of 94 Da (assigned to the phenol moiety), 58 Da
(C2H2O2), 84 Da (two C2H2O), 110 Da (assigned to the pyrocatechol
moiety), 18 Da (H2O), and 136 Da (benzo[c][1,2]dioxete-5-carbal-
dehyde) were readily observed to yield the product ions at m/z
375, 411, 385, 359, 451 and 333, respectively. In the MS3 experi-
ment of the fragment ion at m/z 375 several interesting fragment
ions were generated as eliminations of 42 Da (C2H2O), 110 Da (as-
signed to the pyrocatechol moiety) and 18 Da (H2O). These assign-
ments were supported by an off-line ESI-FTICR-MS2 experiment
e
as cis-standards. Fraction 10 was dissolved in methanol, mem-
brane-filtered (0.45 mm), and analysed by LC/(ꢀ)ESI-MSn.
2.3. Analytical and preparative high-performance liquid
chromatography
An Agilent 1100 analytical HPLC system with a G1312 Binpump,
G1314A variable-wavelength detector (VWD), model 7725 injector
fitted with a 20 lL sample loop, along with an Agilent ChemStation
data system and Waters 600 separations module equipped with a
Waters model 2996 diode-array detector were used. A Phase Agi-
lent Extend C8 column (4.6 mm ꢁ 150 mm, 5
lm) was used for
separation; The mobile phase consisted of methanol (30%, v/v)
and water (70%, v/v) to start, the percentage of methanol changed
from 30% to 70% over the 40 min with the flow rate of 0.8 mL min–1
at room temperature. The detection wavelength was 280 nm.
with the compositions C28H21Oꢀ (measured m/z 469.1294, error
7
ꢀ0.2 ppm), C26H19Oꢀ (measured m/z 411.1249, error ꢀ2.61 ppm),
5
C
24H17Oꢀ (measured m/z 385.1079, error ꢀ0.52 ppm), C28H19Oꢀ
5
6
(measured m/z 451.1181, error ꢀ0.2 ppm), C22H15Oꢀ (measured
6
m/z 375.0875, error ꢀ0.26 ppm), and C22H15Oꢀ (measured m/z
5
2.4. Mass spectrometry
359.0926, error ꢀ0.28 ppm). It was found that the key product
ion at m/z 333 was observed in the MS/MS spectra of compound
3, the relative abundance of which was 4% (the most abundant
ion at m/z 375 is considered to be as 100%). The relative abundance
of the characteristic ion from the loss of benzo[c][1,2]dioxete-5-
carbaldehyde (136 Da) of m/z 469, is possibly due to hydroxy link-
ages to the benzene group of 2,3-dihydrobenzofuran of R1 as shown
in Fig. 3. The key product ion at m/z 385 (the loss of 84 Da (two
C2H2O) from m/z 469) was observed in the MS/MS spectra of com-
pound 3, the relative abundance of which was 28% (the most abun-
dant ion m/z at 375 is considered to be 100%), which corresponds to
the loss of C2H2O (42 Da) from resorcinol (R5) and the loss of other
C2H2O (42 Da) from pyrocatechol (R6), as shown in Fig. 3.
Compounds 8 and 10 were a pair of trans- and cis-isomers,
respectively, which have the same MSn spectra. In the MS/MS spec-
tra of the [M–H]ꢀ ion at m/z 453, compound 8 produced five frag-
ment ions at m/z 435, 411, 369, 347, and 333, which were
generated by the losses of H2O (18 Da), C2H2O (42 Da), two
C2H2O (84 Da), C7H6O (106 Da), C7H4O2 (120 Da), respectively.
The assignments were supported by their composition of
HPLC/(ꢀ)ESI-MSn experiments were performed using the Agi-
lent HPLC system described above combined with a Bruker Esquire
3000plus ion trap mass spectrometer (Bruker-Franzen Analytik
GmbH, Bremen, Germany) equipped with an electrospray ionisa-
tion (ESI) source. The ion source was set to 250 °C, while the needle
voltage was always set to 4.0 kV. Instrument control and data
acquisition were performed using Esquire 5.0 software. To improve
the trapping efficiency and act as a collision gas for the MSn proce-
dure, helium was introduced into the trap at an evaluated pressure
6 ꢁ 10ꢀ6 mbar. Nitrogen was used as both the drying and nebuliz-
ing gas at a back-pressure of 30 psi and a flow rate of 10 L minꢀ1
.
The mass spectrometer was optimised in the collision energy range
of 0.38–1.0 V to maximise the ion current in the spectra.
Off-line FTICR-MS experiments were performed using an Apex
III FTICR mass spectrometer with a 7.0 T actively shielded super-
conducting magnet (Bruker Daltonics, Billerica, MA, USA) com-
bined with an Apollo ESI source operated in the negative ion
mode. The solutions were infused at a rate of 3.0 mL minꢀ1 with
a Cole-Parmer syringe pump. Accurate mass measurements were
performed using CF3COONa as an external calibration compound.
MS/MS calibration was performed using a reference standard and
its fragment peaks. Through the isolation of the desired precursor
ion MS/MS analysis was performed using a correlated sweep. Ar-
gon was used as the collision gas and pulsed into the ICR cell. Each
spectrum obtained is the average of eight transients, each com-
posed of 512 K points, and acquired using a workstation operating
XMASS version 6.1.1.
C
28H21Oꢀ (measured m/z 453.1345, error ꢀ0.2 ppm), C28H19Oꢀ
6
5
(measured m/z 435.1236, error ꢀ0.46 ppm), C26H19Oꢀ (measured
5
m/z 411.1237 error ꢀ0.24 ppm),
369.1131, error ꢀ0.27 ppm), and
C
24H17Oꢀ (measured m/z
4
C
22H15Oꢀ (measured m/z
5
359.0927, error ꢀ0.56 ppm) obtained by off-line FTICR-MS2. In
the MS3 spectrum fragment ion at m/z 369, the neutral losses of
two C2H2O (84 Da) and C2H2O (42 Da) gave corresponding frag-
ment ions at m/z 285 and 327. The key product ions at m/z 333
were also observed in the MS2 of compounds 8 and 10, the relative
abundance of which was 36% (the most abundant ion of at m/z 369
is considered to be 100%). The loss of C7H4O2 (120 Da) from m/z
453 in compounds 8 and 10 occurs relatively easily, possibly be-
cause no hydroxy groups link to the benzene moiety of 2,3-dihy-
drobenzofuran of R3, as shown in Fig. 3. The key product ion at
m/z 369 was also observed in the MS/MS of compounds 8 and
10, the relative abundance of which was 100%, this corresponds
3. Results and discussion
The HPLC/UV chromatogram and total ion chromatogram (TIC)
of the wine grape extracts are shown in Fig. 1. The seven resvera-
trol dimers and three analogous compounds investigated were
numbered as compounds 1–10 by retention time. For MSn analysis,

Q.J. Kong et al. / Food Chemistry 127 (2011) 727–734
729
Fig. 1. Hplc-UV separation (280 nm) (I) and TIC (II) of resveratrol dimer and their analogues compounds from ethyl acetate extract of the stems of wine grapes. For
chromatography and mass spectrometry conditions, see the experimental section.
to the loss of two C2H2O (84 Da) from two molecules of resorcinol
(R5) from compounds 8 and 10, as shown in Fig. 3.
tural similarities. First, the key product ion at m/z 333 was ob-
served in MS/MS spectra of compounds 3, 4, 5, 8 and 10, but the
relative abundances were different (in Fig. 3). These differences
are attributed to hydroxyl whether or not links to benzene groups
of 2,3-dihydrobenzofuran of compounds 3, 4, 5, 8 and 10. When a
hydroxyl group is linked to the benzene group of 2, 3-dihydroben-
zofuran, the relative abundance of compounds 3, 4 and 5 at m/z
333 in the MS2 spectra was less than 10%; when no hydroxy links
were present, as in compounds 8 and 10, the relative abundance at
m/z 333 Da was more than 30%. The key product ion at m/z 369
(385 in compound 3) was from the neutral loss of two C2H2O
(84 Da) in the MS2 of compounds 3, 4, 5, 8, and 10, but the relative
abundance of the fragment ion at m/z 369 (m/z 385 in compound 3)
compared to the most abundant ion was different (see Fig. 3). Since
the structure of compounds 8 and 10 contained two resorcinols
(R5), the relative abundance at m/z 369 in their MS2 spectra was
more than 90% (the most abundant ion at m/z 435 is considered
to be 100%). The structure of compound 3 contained resorcinol
(R5) and pyrocatechol (R6). Thus, its relative abundance at m/z
369 in MS2 was more than 20%. The structure of compounds 4
and 5, on the other hand, contained resorcinol (R5) and phenol
(R7), thus, their relative abundance at m/z 369 in MS2 was less than
10%.
Compounds 4 and 5 were a pair of cis- and trans-isomers,
respectively, which have the same MSn spectra. In the MS/MS spec-
tra of the [MꢀH]ꢀ ion at m/z 453, the losses of 94 Da (assigned to
the phenol moiety), 106 Da (C7H6O), 120 Da (C7H4O2) and 84 Da
(two C2H2O) were observed to yield the product ions at m/z 359,
347, 333 and 369, respectively. The fragmentation pathways were
supported by ESI-FTICR-MSn data, and all errors were less than
ꢀ2.0 ppm. In the MS3 experiment of the fragment ion at m/z 359,
several interesting fragment ions were generated from the elimi-
nations of 70 Da (C2H2O and CO), 94 Da (phenol), 42 Da (C2H2O),
thus producing the ions at m/z 289, 265 and 317. The key product
ions at m/z 333 Da was also observed in the MS/MS spectra of com-
pound 4 and 5, the relative abundance of which was 10% (the most
abundant ion of at m/z 359 is considered to be 100%). The loss of
C7H4O2 (120 Da) from compounds 4 and 5, possibly is relatively
difficult, possibly because of the hydroxy group linked to the ben-
zene moiety of 2,3-dihydrobenzofuran of R2, as shown in Fig. 3. The
key product ion at m/z 369 was also observed in the MS/MS spectra
of compounds 4 and 5, the relative abundance of which was less
than 10% (the most abundant ion at m/z 369 is considered to be
100%), which corresponds to the loss of C2H2O (42 Da) from the
resorcinol (R5) and the loss of other C2H2O (42 Da) from the phenol
(R7) group of compounds 4 and 5 in 3.
3.2. Identification of the unknown compounds 1, 2, 6, 7 and 9
Based on the above analyses of five standard compounds, we
can conclude that the fragmentation patterns were similar in
MSn experiments, thus pointing to the consistency of their struc-
The full-scan mass spectrum of compound 9 showed a deproto-
nated ion at m/z 453 which had similar MS2 fragmentation behav-

730
Q.J. Kong et al. / Food Chemistry 127 (2011) 727–734
Fig. 2. Structures assigned to seven resveratrol dimers and three analogues compounds (1–10) studied in the stems of wine grape. 1a and 1b, 2a and 2b, 9a and 9b: proposed
structures for compounds 1, 2 and 9, respectively.
Fig. 3. Summary of key product ions (m/z 333 and 369) 10 compounds. m/z of the key product ions are given with relative abundance in parentheses (100 for the most
abundant ion). R3:4-(2,3-dihydrobenzofuran-2-yl)phenol.
iours with compounds 8 and 10. Compound 9 was supported by an
ylenecyclobexa (106 Da), C7H4O2 (120 Da), which corresponds to
the product ions at m/z 435, 411, 369, 347, and 333, respectively.
The relative abundance of key fragment ion at m/z 333 was 45%,
indicating 2,3-dihydrobenzofuran of R3 was included in the struc-
ture of compound 9. It is possible that no hydroxyl groups were
off-line ESI-FTICR-MS experiment with a composition of C28H21Oꢀ
6
(measured m/z 453.1330, error ꢀ3.1 ppm). In the MS/MS spectra of
the [M–H]ꢀ ion at m/z 453, compound 9 generated the losses of
H2O (18 Da), C2H2O (42 Da), two C2H2O (84 Da), 4-meth-

Q.J. Kong et al. / Food Chemistry 127 (2011) 727–734
731
linked to the benzene ring of 2,3-dihydrobenzofuran. The relative
abundance of key fragment ion at m/z 369 was 90%, indicating
two resorcinols (R5) contained in the structure of compound 9.
Compound 9 is tentatively proposed to be a new compound, as
shown in Fig. 2.
structures of R2 were probably included in structures of com-
pounds 6 and 7. Hydroxy links to the benzene group of 2,3-dihy-
drobenzofuran appear to exist for both compounds. The relative
abundance of key fragment ions at m/z 369 was less than 10%,
implying that the structure of resorcinol (R5) and phenol (R7) were
included in the structures of compounds 6 and 7. According to the
above data and those in the literature (Kato, Tokunaga, & Sakan,
2009), compound 7 is (E)-5-(6-(4-hydroxystyryl)-4-hydroxy-2-(4-
The [M–H]ꢀ ions for compounds 6 and 7 were observed at m/z
453, both of which have the same MSn spectra. According to the
UV spectrum in Fig. 4, compounds 6 and 7 appear to be a pair of
cis- and trans-isomers. Their MS2 spectra exhibited similar frag-
mentation behaviours compared to compounds 4 and 5. In the
MS/MS spectra of the [MꢀH]ꢀ ion at m/z 453, the loss of 94 Da ,
106 Da, 120 Da and 84 Da were observed to yield product ions at
m/z 359, 347, 333, and 369, respectively. The relative abundance
of key product ions at m/z 333 was less than 10%. Therefore, similar
hydroxyphenyl)-2,3-dihydrobenzofuran-3-yl)
benzene-1,3-diol,
also known as Gnetin C. compound 6, on the other hand, is tenta-
tively proposed to be a new compound.
The full-scan mass spectrum of compound 1 gives an [M–H]ꢀ
ion at m/z 485, which is supported by an off-line ESI-FTICR-MS
experiment with the composition of C29H25Oꢀ (measured m/z
7
Fig. 4. Hplc/Dad-UV spectra of three pairs of trans- and cis-isomers (4 and 5, 8 and 10, 6 and 7) and HPLC/DAD-UV spectra of compounds 2, 9.

732
Q.J. Kong et al. / Food Chemistry 127 (2011) 727–734
485.1616, error ꢀ2.1 ppm). The mass difference between com-
pound 1 and the dimer of resveratrol is 32 Da, which is consistent
with the molecular mass of methanol. In the MS/MS spectrum, the
fragmentation of the [M–H]ꢀ ion at m/z 485 yielded two series of
fragments. The series containing fragment ions at m/z 453 and
467, which were generated by the losses of methanol and H2O
(18 Da), indicating the presence of methoxyl group in the structure
of compound 1. The MS3 spectra showed similar fragmentation
behaviours compared to compound 9. The fragment ion was ob-
served at m/z 333, 347, 359, 369, 411, and 435 in the MS3 spectrum
of compound 1. The fragment originating from the losses of 120 Da,
106 Da, 94 Da, 84 Da, 42 Da, and18 Da were obtained in the MS3
spectra, indicating the loss of C7H4O2, the phenol moiety, two
C2H2O groups, C2H2O, C6H6O, and H2O, respectively. The relative
abundance of the key fragment ion at m/z 333 was more than
30%. Thus, the structure of R3 was probably included in compound
1. The relative abundance of the key product ion at m/z 369 was
23%, indicating that resorcinol (R5) and pyrocatechol (R6) were pos-
sibly included in the structure of compound 1. Compound 1 is ten-
tatively proposed to be a new compound, the structure of which is
shown in Fig. 2. The fragmentation of compound 1 is shown in
Fig. 5.
respectively, indicating the presence of methoxyl groups in the
structure of compound 2. Fragment ions were observed at m/z
211, 185, 141, and 123 in the MS3 spectrum of the ion at m/z
229, indicating the loss of H2O (18 Da), CO2 (44 Da), CO2 (44 Da),
and C7H6O (106 Da), respectively. These fragmentation behaviours
are similar to that of 3,40,50-trihydroxy-5-methoxystilbene, 3,5,50-
trihydroxy-40-methoxystilbene, 3,40,5-trihydroxy-50-methoxystil-
bene, as reported in the literature (Jerkovic, Nguyen, Nizet, & Col-
lin, 2007). Based on the above data, HPLC/DAD-UV spectra, and
supplementary information from the literature (Jerkovic et al.,
2007; Cardona, Fernandez, Garcia, & Pedro, 1986), compound 2
was tentatively identified to be either (Z)-5-(4-hydroxy-3-meth-
oxystyryl) benzene-1,3-diolor or (Z)-5-(3-hydroxy-4-methoxysty-
ryl) benzene-1, 3-diol. Its structure is shown in Fig. 2.The
orientation of the methoxy substitute needs to be identified in
the future work.
3.3. Test the relative abundance of the key product ions in collision
energy
Diastereoisomer could be differentiated by the breakdown
curves of the collision energy (Chen, He, Mao, Sun, & Pan, 2009).
Fig. 6 shows how the fractional abundances of the key product ions
of compounds 1, 3, 5, 7, 8 and 9 changed as the collision energy
varied from 0.60 to 0.95 V, along with the collision energy in-
creased, the relative abundance of deprotonated ion at m/z 453
wear off, and the relative abundance of key product ions increase
The full-scan mass spectrum of compound 2 gave [MꢀH]ꢀ ion at
m/z 257. In the MS/MS spectrum, the fragmentation of the [MꢀH]ꢀ
ion at m/z 257 yielded two series of fragments. The first series con-
tained fragment ions at m/z 229 and 225, which were generated by
the loss of CO (28 Da) and the consecutive loss of methanol (32 Da),
Fig. 5. Fragmentation pathways proposed for [MꢀH]ꢀ ions of compound 1.

Q.J. Kong et al. / Food Chemistry 127 (2011) 727–734
733
Fig. 6. Breakdown curves of selected product ions for compounds 1, 3, 5, 7, 8 and 9.
gradually. However, the relative abundance of the key product ions
at m/z 333 and 369 (m/z 385 in compound 3) compared with most
abundant ion kept unchanged as the collision energy varied from
0.60 to 0.95 V. The relative abundance of the key product ions of
compounds 1, 3, 5, 7, 8 and 9, alonging with the collision energy
variety, were delineated in detail below. The relative abundance
of the key product ions at m/z 333 compared with most abundant
ion at m/z 359 (375 in compound 3) was less than 10% in Fig. 6 (3,
5, 7), the relative abundance of the key product ion at m/z 333 com-
pared with most abundant ion at m/z 359 (m/z 435 in compound 9
and m/z 369 in compound 8) was more than 30% in Fig. 6 (1, 8, 9);
in addition the relative abundance of the key product ion at m/z
369 compared with most abundant ion at m/z 359 (435 in com-
pound 9) was more than 90% in Fig. 6 (8, 9), the relative abundance
of the key product ions at m/z 369 (385 in compound 3) compared
with most abundant ion at m/z 347 (375 in compound 3) was more
than 20% in Fig. 6 (1, 3), the relative abundance of the key product
ions at m/z 369 compared with most abundant ion at m/z 359 was
less than 10% in Fig. 6 (5, 7). Therefore, isomers of resveratrol dimer
were differentiated by the relative abundance of key product ions.
3.4. Differentiation of the trans- and cis-isomers by HPLC/DAD-UV
No differences between the trans- and cis-isomers among the
relative abundances of key product ions were observed. The
HPLC/DAD-UV spectra between the trans- and cis-isomers, how-
ever, were quite different (Chen et al., 2009). As a result, the UV
spectra between the trans- and cis-isomers was quite different
(Fig. 4). The trans-isomers (compounds 5, 7 and 8) had different
adsorption wavelengths and stronger intensities than the cis-iso-
mers (compounds 4, 6 and 10). This difference may be attributed
to the more stable the conjugations of the trans-isomers compared
to that of the cis-isomers. Based on the above HPLC/DAD-UV spec-
tra, compounds 2 and 9 were tentatively assigned to as cis- and
trans-configuration, respectively, as shown in Fig. 4.

734
Q.J. Kong et al. / Food Chemistry 127 (2011) 727–734
chromatography/electrospray ionization multiple-stage tandem mass
4. Conclusions
spectrometry. Journal of Mass Spectrometry, 42, 1148–1161.
Jerkovic, V., Nguyen, F., Nizet, S., & Collin, S. (2007). Combinatorial synthesis,
reversed-phase and normalphase high-performance liquid chromatography
elution data and liquid chromatography/positive atmospheric pressure
chemical ionization tandem mass spectra of methoxylated and glycosylated
resveratrol analogues. Rapid Communications in Mass Spectrometry, 21,
2456–2466.
Kato, E., Tokunaga, Y., & Sakan, F. (2009). Stilbenoids isolated from the seeds of
Melinjo (Gnetum gnemon L.) and their biological activity. Journal of Agricultural
and Food Chemistry, 57, 2544–2549.
Kong, Q. J., Ren, X. Y., Jiang, L. Y., Pan, Y. J., & Sun, C. R. (2009). Scirpusin A, a
hydroxystilbene dimer from Xinjiang wine grape, acts as an effective singlet
oxygen quencher and DNA damage protector. Journal of the Science of Food and
Agriculture, 90, 823–828.
Lin, M., Li, J. B., Li, S. Z., Yu, D. Q., & Liang, X. T. (1992). A dimeric stilbene from
Gnetum parvifolium. Phytochemistry, 32, 633–638.
The results illustrated the potential advantage of the method of
HPLC/(-)ESI-MSn and HPLC/DAD-UV for identification isomers of
resveratrol dimers and their analogues from wine grapes. Based
on the fragmentation behaviours, ten compounds were analysed,
among them, three compounds are tentatively identified as new
compounds. The relative abundance of two diagnostic product ions
at m/z 333 and 369 (385 in compound 3) were utilised for the key
structure identification of a resveratrol dimer. Three pairs of the
trans- and cis-isomers could be identified by their HPLC/DAD-UV
spectra. Based on their UV spectra, compounds 2 and 9 were tenta-
tively identified as cis and trans- configuration, respectively.
Palma, M., Taylor, L. T., Varela, R. M., Cutler, S. J., & Cutler, H. G. (1999). Fractional
extraction of compounds from grape seeds by supercritical fluid extraction and
analysis for antimicrobial and agrochemical activities. Journal of Agricultural and
Food Chemistry, 47, 5044–5048.
Acknowledgments
Pezet, R., Perret, C., Jean-denis, J. B., Tabacchi, R., Gindro, K., & Viret, O. (2003). d-
Viniferin, a resveratrol dehydrodimer: One of the major stilbenes synthesized
by stressed grapevine leaves. Journal of Agricultural and Food Chemistry, 51,
5488–5492.
Financial support from NSFC of China (20942010) is gratefully
acknowledged.
Renaud, S., & Delorgeril, M. (1992). Wine, alcohol, platelets and the French Paradox
for coronary heart-disease. Lancet, 339, 523–1526.
References
Shao, B., Guo, H. Z., Cui, Y. J., Liu, A. H., Yu, H. L., Guo, H., et al. (2007). Simultaneous
determination of six major stilbenes and flavonoids in Smilax china by high
performance liquid chromatography. Journal of Pharmaceutical and Biomedical
Analysis, 44, 737–742.
Shukitt-Hale, B., Carey, A., Simon, L., Mark, D. A., & Joseph, J. A. (2006). Effects of
Concord grape juice on cognitive and motor deficits in aging. Nutrition, 22,
295–302.
Tong, L., Wang, Y. Z., Xiong, J. F., Cui, Y., Zhou, Y. G., & Yi, L. C. (2008). Selection and
fingerprints of the control substances for plant drug Eucommia ulmodies oliver
by HPLC and LC–MS. Talanta, 76, 80–84.
Van Wauwe, J., & Goossens, J. (1983). Effects of antioxidants on cyclooxygenase and
lipoxygenase activities in intact human platelets: comparison with
indomethacin and etya. Prostaglandins, 26, 725–730.
Cardona, M. L., Fernandez, M. I., Garcia, M. B., & Pedro, J. R. (1986). Synthesis of
natural polyhydroxy stilbenes. Tetrahedron, 42, 2725–2730.
Chen, J. J., He, S., Mao, H., Sun, C. R., & Pan, Y. J. (2009). Characterization of
polyphenol compounds from the roots and stems of Parthenocissus laetevirens
by high-performance liquid chromatography/tandem mass spectrometry. Rapid
Communications in Mass Spectrometry, 23, 737–744.
Chen, J. J., Li, X. Y., Sun, C. R., Pan, Y. J., & Schlunegger, U. P. (2008). Identification of
polyoxypregnane glycosides from the stems of Marsdenia tenacissima by high-
performance liquid chromatography/tandem mass spectrometry. Talanta, 77,
152–159.
He, S., Jiang, L. Y., Wu, B., Pan, Y. J., & Sun, C. R. (2009). Pallidol, a resveratrol dimer
from red wine, is
a selective singlet oxygen quencher. Biochemical and
Biophysical Research Communications, 379, 283–287.
Wu, Q. L., Wang, M. F., & Simon, J. E. (2003). Determination of isoflavones in red
clover and related species by high-performance liquid chromatography
combined with ultraviolet and mass spectrometric detection. Journal of
Chromatography A, 1016, 195–209.
He, S., Wu, B., Pan, Y. J., & Jiang, L. Y. (2008). Stilbene oligomers from Parthenocissus
laeteWirens: Isolation, biomimetic synthesis, absolute configuration, and
implication of antioxidative defense system in the plant. The Journal of
Organic Chemistry, 73, 5233–5241.
Huang, X., Song, F. R., Liu, Z. Q., & Liu, S. Y. (2007). Studies on lignan constituents
from Schisandra chinensis (Turcz.) Baill. fruits using high-performance liquid
Zern, T. L.,
& Fernández, M. L. (2005). Cardioprotective effect of dietary of
polyphenols. Journal of Nutrition, 135, 2291–2294.