J. Sep. Sci. 2007, 30, 1229 – 1234 E. B.

Walker 1229

Edward B. Walker Original Paper

Department of Chemistry, Weber
State University, Ogden, Utah,
USA
HPLC analysis of selected xanthones in
mangosteen fruit
Xanthones are unique chemical compounds found in nature, composed of a tricy-
clic aromatic system with a variety of phenolic, methoxy, and isoprene substituents,
giving rise to numerous derivatives. They dissolve to varying degrees in solvents
ranging from alcohol to hexane. An optimum solvent mixture of acetone/water
(80:20) selectively and effectively extracts a wide variety of xanthones. Subsequent
HPLC analysis using standard C-18 RP and a 30-min gradient of 65 – 90% MetOH in
0.1% formic acid detects and separates numerous different xanthones with UV
detection at 254 nm. The xanthones alpha-mangostin, 8-desoxygartanin, gartanin,
beta-mangostin, 3-mangostin, and 9-hydroxycalabaxanthone have been extracted,
identified, and quantitatively determined using this method. This analytical
method is applied to the analysis of these xanthones in the rind of the mangosteen
fruit, Garcinia mangostana.
Keywords: Garcinia mangostana / HPLC / Mangosteen / Mangostin / Xanthones /
Received: January 18, 2007; revised: February 16, 2007; accepted: February 18, 2007
DOI 10.1002/jssc.200700024

1 Introduction
Xanthones are naturally occurring compounds that con-
tain a distinctive chemical structural component,
namely a tricyclic aromatic ring system (Fig. 1). Most
often, these ring systems are substituted with a variety of
isoprene, phenolic, and methoxy groups that give rise to
a large variety of possible structures.
Discovery of biological activities exhibited by naturally
occurring xanthones has increased the interest and
demand for nutritional supplement products containing
these ingredients. Xanthones have been reported to have
antioxidant [1, 2], antimalarial [3], antimicrobial [4 – 6],
and antiacne activities [7]. Xanthones from Gutiferae
exhibit antiinflammatory properties and are reported to
show antiulcer activity as well [8]. Matsumoto et al. [9],
reported six individual xanthones from Garcinia mangos-
tana pericarp, all exhibited growth inhibitory effects on
human leukemia cells. Selected xanthones from this
same fruit inhibit lipoprotein oxidation [10] and inhibit
both histamine release and prostaglandin synthesis [11]. Figure 1. Chemical structures of xanthones.
Xanthones have been isolated from plants of the Clusia-
ceae (Gutiferae) family, including stems of Garcinia multi- isolated selected xanthones from this species and then
flora [12], heartwood of G. mangostana [13] as well as other demonstrated the chemical interconversion of xan-
anatomical parts of this species [14]. Jefferson et al. [15], thones from one to another via selective demethylation.
Although many researchers have successfully isolated
xanthones via exhaustive extractions and separation
Correspondence: Dr. Edward B. Walker, Department of Chemis-
try, Weber State University, Ogden, Utah 84403-2503, USA schemes, there are few quantitative analytical methods
E-mail: ewalker@weber.edu reported in the literature. In one case, Bo et al. [16] report
Fax: 801-626-7445 HPLC and CE separation methods using mixtures of nine

i 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

1230 E. B. Walker
purified standards. However, there is a definite lack of
rapid, dependable analytical methods for xanthones in
their naturally occurring states. A method for rapid and
efficient extraction coupled with HPLC analysis is
described that can assist in the quantitative determina-
tion of xanthones from their natural sources.
2 Experimental
2.1 Instrumentation
HPLC experiments were performed using a Waters Alli-
ance System equipped with a vacuum degasser, quater-
nary solvent mixing, autosampler, and a Waters 996
diode array detector. UV spectra were collected across
the range of 200 – 400 nm, extracting 254 nm for chro-
matograms. Waters Millennium32 software was utilized
for instrument control, data collection, and data process-
ing. The column was a Waters Nova-Pak C-18
(3.96150 mm2). The mobile phases consisted of A: 0.1%
formic acid in water and B: methanol; optimal separa-
tions were achieved using a gradient of 65 to 90% B over
0 – 30 min at the flow rate of 1.0 mL/min. Injection vol-
ume for all the samples and standards was 10 lL.
2.2 Chemicals
Synthetic xanthone, 97% was purchased from Aldrich
Chemical (Aldrich Catalog# X600, CAS# 90-47-1). Natural
xanthone standards such as alpha-mangostin, beta-man-
gostin, 3-mangostin, gartanin, 8-desoxygartanin, and 9-
hydroxycalabaxanthone were purchased from Chroma-
Dex (2952 S Daimler St., Santa Ana, CA, 92705). Dried,
ground whole mangosteen fruit rind was obtained from
Xango, LLC (3098 West Executive Parkway, Lehi, Utah
84043, USA). Methanol, acetone, and formic acid were all
HPLC grade. Purified water was prepared using a Milli-Q
system (Millipore, Bedford, MA).
2.3 Isolation of standards
Xanthone standards were isolated in our laboratory by
mixing 10 kg of dried, whole mangosteen fruit rind with
50 L of hexane and mixing at 608C for 8 h. After filtra-
tion, the hexane was removed by roto-evaporation under
reduced pressure. The resulting solids were redissolved
in methanol. Alpha-mangostin was recrystallized from
this solution after adding small portions of water (1:20)
to the warm methanol solution and allowing to cool.
More alpha-mangostin and the other five xanthones
were purified by preparative HPLC, utilizing the same
solvent system and UV detection as for the analytical
method. The preparative column (Waters Nova-Pak HR
C18 6 lm, 196300 mm2) provided excellent separations
when 1.0 mL volumes of the extract were injected and
i 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Sep. Sci. 2007, 30, 1229 – 1234
the gradient of 65 – 90%B was extended to 0 – 140 min at
the flow rate of 10.0 mL/min. Fractions from multiple
injections were collected with an automated Gilson frac-
tion collector and pooled. The individual xanthone solids
were recovered by roto-evaporation and recrystallized
from methanol – water solutions.
3 Results and discussion
3.1 Sample preparation
Naturally occurring xanthones are insoluble in water
but are soluble in a variety of other solvents, ranging in
polarity from methanol to hexane. An acetone – water
mixture of 80:20 (acetone/water) is an excellent solvent
for the extraction of xanthones from their natural sour-
ces such as fruits. This is due to both the excellent solubil-
ity of the xanthones in this solvent mixture and the abil-
ity of this solvent system to penetrate the sample matrix
during extraction. The 80:20 mixture of acetone/water
was also robust, showing no change in total xanthone
extraction yields when altered across the range of 70 –
90% acetone in water. Extraction of 0.1 g of dried, finely
ground mangosteen fruit rind in 50 mL of 80:20 acetone/
water for 30 min at room temperature resulted in quan-
titative removal of the xanthones.
To prepare samples for HPLC analysis, dried, ground
mangosteen fruit rind was quantitatively extracted by
mixing 0.1 g with 40 mL of 80:20 acetone/water in a
50 mL volumetric flask and mixing vigorously on a wrist-
shaker for 30 min. After final dilution to 50 mL with
additional 80:20 acetone/water, the sample was filtered
through a 0.45 lm PTFE filter and 10 lL of the resulting
solution was injected into the HPLC.
Extraction efficiency was validated by subjecting sam-
ples to the complete extraction scheme, then removing
50% of the extraction suspension from the extraction
flask. Fresh 80:20 acetone/water extraction medium was
added to replace the volume of medium removed, thus
diluting the extraction suspension by 50%. Following
another 30-min of vigorous shaking, the resulting sus-
pension was analyzed by the same procedure as the origi-
nal extraction. The difference in concentrations for each
of the analytes during triplicate dilution procedures was
48.4 – 53.6%, demonstrating that the extraction medium
was, indeed, able to quantitatively extract each of the six
xanthones from the sample matrix.
A further check on the extraction efficiency was con-
ducted by adding reference standards to the sample
matrix prior to extraction. Spike levels of 15 and 30% to
the original xanthone levels in the sample matrix were
quantitatively recovered within the range of 98.0 –
104.3% for each standard, further substantiating the
extraction procedure (Table 1).
www.jss-journal.com
J. Sep. Sci. 2007, 30, 1229 – 1234 Liquid Chromatography 1231

Table 1. Xanthone extraction efficiency data

Xanthone peak 50% Dilution 15% Spike 30% Spike Average recovery
(% change) recovery recovery (%)

3-Isomangostin 50.3 100.7 101.0 100.9

8-Desoxygartanin 50.7 99.0 98.9 98.9
Gartanin 48.4 98.1 96.1 97.1
Alpha-mangostin 51.2 98.4 98.0 98.2
9-Hydroxycalabaxanthone 53.6 99.2 99.7 99.4
Beta-mangostin 56.6 102.8 104.3 103.6

Figure 2. HPLC chromatogram of mangosteen fruit rind extract. The lower plot is expanded on the absorbance scale to reveal
the numerous minor xanthone peaks that are present in the extract.

3.2 Chromatographic results

Each of the xanthone peaks is well resolved from the
neighboring peaks and displays excellent peak symmetry
and separation efficiency, as seen in Fig. 2 and Table 2.
The unique chromophoric nature of xanthones makes
them easy to identify from their UV diode-array absorp-
tion spectra. All the unique spectra have a strong, split
absorption peak maximum range of 240 – 300 nm. In
addition, a secondary single peak of slightly lower extinc-
tion is observed at 310 – 370 nm (Fig. 3).

i 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

1232 E. B. Walker J. Sep. Sci. 2007, 30, 1229 – 1234

Table 2. Chromatography system suitability results

Xanthone peak Retention time Efficiency Resolution Peak asymmetry

(min) (plates)

3-Isomangostin 12.7 19 204 N/A 0.87

8-Desoxygartanin 16.2 33 907 10.27 0.94
Gartanin 19.9 55 225 10.07 0.84
Alpha-mangostin 21.0 58 431 2.88 0.79
9-Hydroxycalabaxanthone 26.7 85 010 14.73 0.92
Beta-mangostin 28.2 110 964 2.50 0.95

Figure 3. Spectrum index plot. Characteristic UV absorption spectra at 200 – 400 nm clearly identify components with xanthone-
like properties.

Alpha-mangostin is the most prominent peak in the
chromatogram, followed by 8-desoxygartanin and beta-
mangostin. The exact identification of these xanthones
as well as 3-mangostin, gartanin, and 9-H-C-xanthone
was accomplished by the analysis of their purified stand-
ards. All six of these xanthones were isolated by prepara-
tive chromatography, collected, and recrystallized from
alcohol – water solutions. The retention times and UV
spectra of these isolated compounds were used to iden-
tify and quantify them. In addition, independent com-
mercial standards were purchased and used to compare
identities and individual detector response factors.

i 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

J. Sep. Sci. 2007, 30, 1229 – 1234 Liquid Chromatography 1233

Table 3. Xanthone retention times and response factors

Xanthone standard Retention time Slope of Number of R2 Response

(min) calibration plot calibration levels factor

Synthetic xanthone 4.4 2.80E + 04 7 1.0000 1.00

3-Mangostin 12.7 3.64E + 04 6 0.9996 1.30
8-Desoxygartanin 16.2 2.90E + 04 5 0.9969 1.03
Gartanin 19.9 2.59E + 04 7 0.9937 0.92
Alpha-mangostin 21.0 3.90E + 04 6 0.9988 1.39
9-Hydroxycalabaxanthone 26.7 1.31E + 04 6 0.9981 0.47
Beta-mangostin 28.2 3.73E + 04 5 0.9989 1.33

Detection: Integrated peak areas at UV-254 nm; concentration units: lg/mL.

Table 4. Example experimental data: Mangosteen dried fruit rind

Xanthone Retention time Peak area Response Xanthone concen-

(min) at 254 nm factor tration (lg/g)

3-Mangostin 12.7 76 864 1.30 1.05E + 03

8-Desoxygartanin 16.2 874 970 1.03 1.50E + 04
Gartanin 19.9 108 529 0.92 2.1E + 03
Alpha-mangostin 21.0 4 332 678 1.39 5.51E + 04
9-Hydroxycalabaxanthone 26.7 23 988 0.47 9.1E + 02
Beta-mangostin 28.2 127 656 1.33 1.70E + 03

Sample preparation: 0.1007 g of dry, ground rind extracted with 50.00 mL of 80:20 acetone/water.

3.3 Detector response factors
Detector response factors for individual xanthones were
determined preparing multiple-level calibration solu-
tions of purified xanthones and subjecting them to chro-
matographic analysis under the same conditions as used
for the separation of more complicated xanthone
extract. Excellent agreement was observed between the
isolated standards and the commercial standards (see
Table 3.).
Synthetic (unsubstituted) xanthone was also evaluated
as potential standard. This xanthone elutes at substan-
tially shorter elution times, well ahead of most of the nat-
ural substituted xanthones. No xanthone was observed
in the mangosteen fruit extracts, nominating it as an
excellent internal standard. Furthermore, its response
factor at 254 nm is similar to many of the naturally
occurring xanthones.
present at extremely low concentrations. Each of these
possible xanthones is well-resolved from the known xan-
thones identified in this method and offers further
opportunity for future identification and subsequent
analysis.
4 Concluding remarks
This HPLC analytical method demonstrates excellent
selectivity and resolution for six different xanthone com-
pounds in extracts of mangosteen fruit, in a relatively
short run of approximately 30 min. Extraction with a
mixture of 80:20 acetone/water is an excellent, robust
method for the preferential extraction of xanthones
from their natural sources. Six specific xanthones have
been identified and their relative response factors deter-
mined.

3.4 Results
5 References
Alpha-mangostin was present at the highest concentra-
tion in the dried mangosteen fruit rind at a concentra-
tion of 5510 lg/g. The other xanthones in order of
descending concentration were 8-desoxygartanin, garta-
nin, beta-mangostin, 3-mangostin, and 9-hydroxycala-
baxanthone (results listed in Table 4).
UV absorption spectra of other smaller, resolved peaks
suggest the presence of more xanthone-like compounds
[1] Chiang, Y., Kuo, Y., Oota, S., Fukuyama, Y., J. Nat. Prod. 2003, 66,
1070 – 1073.
[2] Jung, H., Su, B., Keller, W. J., Mehta, R. G., Kinghorn, A. D., J. Agric.
Food Chem. 2006, 54, 2077 – 2082.
[3] Likhitwitayawuid, K., Phadungcharoen, T., Krungkrai, J., Planta
Med. 1998, 64, 70 – 72.
[4] Rukachaisirikul, V., Kamkaew, M., Sukavisit, D., Phongpaichit,
S., et al., J. Nat. Prod. 2003, 66, 1531 – 1535.
[5] Mahabusarakam, W., Wiriyachitra, P., Phongpaichit, S., J. Sci. Soc.
Thailand 1986, 12, 239 – 242.

i 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

1234 E. B. Walker
[6] Sundaram, B. M., Gopalakrishnan, C., Subramanian, S., Shankar-
anarayanan, D., Kameswaran, L., Planta Med. 1983, 48, 59 – 60.
[7] Chomnawang, M. T., Surassmo, S., Nukoolkarn, V. S., Gritsana-
pan, W., J. Ethnopharmacol. 2005, 101, 330 – 333.
[8] Bennett, G. J., Lee, H., Phytochemistry 1989, 28, 967 – 998.
[9] Matsumoto, K., Akao, Y., Kobayashi, E., Ohguchi, K., et al., J. Nat.
Prod. 2003, 66, 1124 – 1127.
[10] Mahabusarakam, W., Proudfoot, J., Taylor, W., Croft, K., Free Rad.
Res. 2000, 33, 643 – 659.
[11] Nakatani, K., Atsumi, M., Arakawa, T., Oosawa, K., et al., Biol.
Pharm. Bull. 2002, 25, 1137 – 1141.
i 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Sep. Sci. 2007, 30, 1229 – 1234
[12] Chiang, Y., Kuo, Y., Oota, S., Fukuyama, Y., J. Nat. Prod. 2003, 66,
1070 – 1363.
[13] Nilar, H. L. J., Phytochemistry 2002, 60, 541 – 548.
[14] Govindachari, T. R., Kalyanaraman, P. S., Muthukumaraswamy,
N., Pai, B. R., Tetrahedron 1971, 27, 3919 – 3926.
[15] Jefferson, A., Quillinan, A. J., Scheinmann, F., Sim, K. Y., Aust. J.
Chem. 1970, 23, 2539 – 2543.
[16] Bo, T., Liu, F., Li, K. A., Liu, H., J. Liq. Chrom. Rel. Technol. 2003, 26,
993 – 998.
www.jss-journal.com