2.3. HPLC analysesSeparation of xan

2.3. HPLC analyses
Separation of xanthones was performed using an Agilent HPLC
series 1200 system (Agilent, Waldbronn, Germany) with ChemStation
software, a model G1379 degasser, a model G1312B binary
gradient pump, a model G1367D thermoautosampler, a model
G1316B column oven and a model G1315C diode array detector.
The column used was a 50  4,6 Zorbax Eclipse XDB from Agilent
(Waldbronn, Germany) operated at a temperature of 25 C.
The mobile phase consisted of 2% acetic acid in water (eluent A)
and 0.5% acetic acid in acetonitrile (eluent B) using a gradient program
as follows: 50–60% B (20 min), 60–70% B (35 min), 70–
100% B (5 min), 100% B isocratic (3 min), 100–0% B (2 min) at a
flow rate of 0.6 mL/min. Total run time was 65 min. The injection
volume was 8 lL for all samples. Xanthones were monitored at254 nm and UV/vis spectra were recorded in the range of 200–
600 nm.
Xanthones were identified by UV/vis and mass spectra and
quantified using a calibration curve of the a-mangostin standard.
The quantification of further, structurally related xanthones was
performed applying a molecular weight correction factor (Chandra,
Rana, & Li, 2001). All determinations were carried out in triplicate.
2.4. LC–MS analyses
LC–MS analyses were performed with an Agilent HPLC series
1100 system equipped with ChemStation software, a model
G1322A degasser, a model G1312A binary gradient pump, a model
G1329/G1330A thermoautosampler, a model G1316A column oven
and a model G1315A diode array detector. The column, eluents and
HPLC parameters were the same as described above. The HPLC-system
was coupled on-line with a Bruker (Bremen, Germany) model
esquire 3000 + ion trap mass spectrometer fitted with an ESI
source. Data acquisition and processing were performed using Esquire
Control Software. Negative ion mass spectra of the column
eluate were recorded in the range of m/z 50–1000. Nitrogen was
used both as drying gas at a flow rate of 10 L/min and as nebulising
gas at a pressure of 60 psi. The nebuliser temperature was set at
365 C. Collision-induced dissociation spectra were obtained with
a fragmentation amplitude of 1.2 V. Helium was used as the collision
gas.
3. Results and discussion
As prenylated xanthones are the major secondary plant metabolites
in mangosteens, they are supposed to be responsible for the
positive health effects of the fruits, thus being the basis for the
claimed effectiveness of functional products from mangosteens.
For most commercial mangosteen products including the analysed
functional beverage, processing of mangosteens is already realised
in their cultivation areas, commonly using the entire fully ripened
fruits including the edible aril segments and the pericarp (Bin Osman
& Milan, 2006; Kanchanapoom & Kanchanapoom, 1998; Morton,
1987). Since mangosteen pericarp has only been consumed in
its country of origin so far and has not been used for human consumption
to a significant extent within the European Community
before May 1997, it is still considered a novel food despite its
long-lasting utilisation as a food and in traditional medicine. This
implies that products containing mangosteen pericarp need to be
authorised according to the Novel Food Directive No. 258/97.
Due to its strict interpretation, thus creating economic trade barriers,
revision of the legislation is currently under discussion (European
Commission, 2011). According to an updated draft version,
exotic fruits with a significant history of use outside the EU will
be permitted to undergo a simplified application procedure based
on a safety assessment considering their history of safe use.
Therefore, one objective of the present study was to provide evidence
of essential equivalence of mangosteen pericarp composition
with particular emphasis on its xanthone profile as
compared to the edible aril segments. Besides the characterisation
and quantification of these constituents in both plant parts, a functional
beverage, produced from purees made from the whole mangosteen
fruits, was also included in the xanthone profiling.
For this purpose, extracts of the mangosteen pericarp, the edible
aril segments and the functional beverage were analysed with
HPLC–DAD and LC–MS. Due to the limited availability of xanthone
references and because UV/vis spectral data do not allow unambiguous
identification of individual components, HPLC coupled to
mass spectrometry proved to be an indispensable tool for peak
assignment. Since limited fragmentation facilitated the tentative
identification of the separated xanthones (Zarena & Sankar,
2009), ESI techniques have been confirmed to be perfectly suitable
for xanthone characterisation due to their soft ionisation yielding
mainly integral molecular ions. The LC–MS data of all identified
xanthones are summarised in Table 1.
3.1. Identification of xanthones from pericarp, aril segments and the
functional beverage with LC–MSn
Fig. 1 shows the HPLC profiles of xanthones extracted from
mangosteen pericarp, aril segments and the functional beverage.
Polarity and hence, the retention time of the compounds was affected
by the position of hydroxyl and prenyl groups varying the
xanthen-9-one core. Similar findings have already been reported
in a previous study by Chaivisuthangkura et al. (2009) evaluating
a method for the determination of xanthones in mangosteen fruits
with HPLC–DAD. In agreement with our study, 254 nm was chosen
as detection wavelength for the fingerprint chromatogram and
quantification of the xanthones.
For ESI-MS analysis, xanthones were recorded in the negative
ion mode. Contrary to the positive mode, negative ionisation resulted
in increased fragmentation of the xanthones in the MS/MS
spectra, thus facilitating their identification and structure elucidation.
In agreement with Zhou, Han, Song, Qiao, and Xu (2008) as
well as Da Costa et al. (2000) fragmentation pathways of the
polyprenylated xanthones were found to follow general trends.
All compounds showed abundant [MH] ions and yielded successive
loss of prenyl residues C4H8 (56 Da) in the MS/MS spectra.
3.1.1. Identification of xanthones from pericarp and aril segments
Astonishingly, the pericarp and aril segments of mangosteens
exhibited similar fingerprints each consisting of two major and five
minor components. Based on molecular weights, order of elution,
comparison with a-mangostin standard, and fragmentation, all 7
compounds belonging to the xanthones were tentatively identified.
Concordant with literature data, the predominant peak could be
assigned to a-mangostin (compound 6) (Chaivisuthangkura et al.,
2009; Jung et al., 2006; Walker, 2007). Fig. 2 shows the mass spectrum
and the suggested fragmentation pathway for a-mangostin.
Fragmentation of m/z 409 for a-mangostin yielded a de-methylated
product ion at m/z 394. In contrast to most other detected
fragments, the product ion at m/z 394 exhibited an even mass.
Since xanthones do not contain three bonding atoms such as nitrogen
or phosphorus, the even mass leads to the assumption, that the
dissociation of the C–O-bond followed a homolytic mechanism.
The resulting radical ion showed a relatively high intensity that
may be due to resonance stabilisation. Similar fragmentation patterns
with a homolytic cleavage and radical fragments of even
masses were also observed for the other xanthones. The predominant
peak in the MS spectrum of a-mangostin is at m/z 351, resulting
from a following ring closure under loss of a C3H7 radical
(43 Da). As a competing reaction, the de-methylated radical ion
reacted under loss of a C4H7 radical (55 Da) from a prenyl group,
yielding a product ion at m/z 339. Under the applied mass spectrometric
conditions, the formation of a de-methoxylated product ion
(32 Da) at m/z 377 was also observed. In addition, the loss of the
second prenyl group in allylic position (56 Da) could be observed
in the MS3-experiments. The fragmentation reactions of a-mangostin
are comparable to those reported by Zhou et al. (2008), who
analysed polyprenylated xanthones from Garcinia xipshuanbannaensis
Y.H. Li. extracts by HPLC–ESI-QTOF-MSn. The second major
peak of mangosteen pericarp (compound 2) was assigned to cmangostin.
Compared to a-mangostin, the c-form is characterised
by the presence of a hydroxyl group at position 7 instead of a
methyl group. Therefore, the elimination of prenyl moieties in
allylic position may be the preferred reaction. Mass spectrometric
J. Wittenauer et al. / Food Chemistry 134 (2012) 445–452 447