Please cite this article in press as: J. Attoumbr, et al., J. Chromatogr. B (2012), http://dx.doi.org/10.1016/j.jchromb.2012.09.

025
ARTICLE IN PRESS
GModel
CHROMB-18104; No. of Pages 5
Journal of Chromatography B, xxx (2012) xxxxxx
Contents lists available at SciVerse ScienceDirect
Journal of Chromatography B
j our nal home page: www. el sevi er . com/ l ocat e/ chr omb
Short communication
Preparative separation of glycoalkaloids -solanine and -chaconine by
centrifugal partition chromatography
Jacques Attoumbr
a
, David Lesur
b
, Philippe Giordanengo
c,d
, Sylvie Baltora-Rosset
e,
a
S.I.P.R.E Comit Nord, rue des champs Potez, 62217 Achicourt, France
b
Laboratoire des Glucides FR3517, UPJV, 33 rue St Leu, 80039 Amiens cedex, France
c
Universit de Picardie Jules Verne, 33 rue Saint Leu, 80039 Amiens cedex 1, France
d
Institut Sophia Agrobiotech, CNRS 7254 INRA 1355, Universit de Nice Sophia Antipolis, 400 route des Chappes, BP167, 06903 Sophia Antipolis, France
e
EDYSAN EA 4698, Universit de Picardie Jules Verne, Facult de Pharmacie, 1 rue des Louvels, 80037 Amiens cedex, France
a r t i c l e i n f o
Article history:
Received 23 April 2012
Accepted 14 September 2012
Available online xxx
Keywords:
Glycoalkaloids
-Chaconine
-Solanine
Centrifugal partition chromatography (CPC)
High performance liquid chromatography
(HPLC)
a b s t r a c t
The main glycoalkaloids of a commercial potato cultivar, -chaconine and -solanine, were extracted
from sprouts of Solanum tuberosum cv. Pompadour by a mixture of MeOH/H
2
O/CH
3
COOH (400/100/50,
v/v/v). In these conditions, 2.8 0.62 g of crude extract were obtained from 50g of fresh sprouts and
the total glycoalkaloid content was determined by analytical HPLC at 216.5 mg/100 g. -Chaconine and
-solanine were separated in a preparative scale using centrifugal partition chromatography (CPC). In
a solvent system composed of a mixture of ethyl acetate/butanol/water (15/35/50, v/v/v), -chaconine
(54mg) and -solanine (15 mg) were successfully isolated from the crude extract in one step of puri-
cation. The purity of isolated compounds was determined to be higher than 92% by HPLC analysis.
2012 Elsevier B.V. All rights reserved.
1. Introduction
Steroidal glycoalkaloids constitute an important class of biolog-
ically active compounds in potato (Solanaceae) plant family which
is a very valuable crop that provides high-quality nutrition to bil-
lions of people around the world. These secondary metabolites
are involved in plant resistance to pests and predators and have
been shown to be toxic to a wide range of organisms from fungi
to humans [1,2]. Glycoalkaloids (GAs) are usually distributed in all
plant organs with a main localization in sprouts, owers and skin
[3]. Their structures and concentrations largely depend on potato
lines and environmental factors and it has been shown that GAs
can accumulate to high levels in greened, stored, damaged, and
irradiated tubers for example [4].
-Chaconine and -solanine are the major potato glycoalka-
loids (PGAs) in commercial cultivars and are intensively studied
because of their greatest contribution to the total GA content and
their bioactivity. The potential human poisoning effect of PGAs has
led to the implementation of safety regulations limiting their con-
tent in edible tubers to 20mg/100g fresh weight involving the
development of numerous reliable analytical methods [3]. More-
over, biological activity of PGAs is not limited to their toxicity and

Corresponding author. Tel.: +33 03 2282 7770; fax: +33 03 2282 7469.
E-mail address: [email protected] (S. Baltora-Rosset).
they have been reported to possess pharmacological properties [2].
To study and validate their potential health-promoting effects in
humans, effective extraction procedures and preparative separa-
tion methods are needed. Various analytical methods have been
reported for the determination of GAs. The simplest methods, such
as gravimetric and colorimetric ones, lack the required specicity
and suffer fromcontamination by other potato components. Other
techniques such as immunoassays rely on the specicity of anti-
bodies to get better sensitivity and to eliminate tedious sample
preparation process but are not suitable to recover isolated prod-
ucts [5]. The main methods used to detect, quantify and isolate
GAs are chromatographic techniques such as gas chromatogra-
phy (GC), high-performance liquid chromatography (HPLC) or
liquid chromatography electrospray ionization-mass spectrometry
(LCESI-MS) [69]. Because of their complex chemical structures
(hydrophobic 27-carbon skeleton of steroidal alkamine attached
to a hydrophilic trisaccharide), serious technical difculties are
associated with PGAs quantitative analysis and isolation. GC anal-
ysis requires chemical derivatization. As GAs show no suitable
chromophore for HPLCUV methods, their detection has to be
made at around 200210nmwhere many compounds absorb light
leading to tedious sample preparations to overcome background
noise. Solid-phase extraction (SPE) currently used to purify GAs
from plant matrixes before their analysis by HPLC often leads to
extremely variable results for recoveries [7,8]. Moreover, evenafter
SPE clean-up, the analysis of GAs extracts by HPLC requires a strict
1570-0232/$ see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jchromb.2012.09.025
Please cite this article in press as: J. Attoumbr, et al., J. Chromatogr. B (2012), http://dx.doi.org/10.1016/j.jchromb.2012.09.025
ARTICLE IN PRESS
GModel
CHROMB-18104; No. of Pages 5
2 J. Attoumbr et al. / J. Chromatogr. B xxx (2012) xxxxxx
pH control to observe efcient detection and separation and the
composition of the mobile phase is also an important factor to
ensure good solubility of GAs. The need to use buffers to set a
suitable pH for the HPLC technique limits its use for the develop-
ment of easy preparative procedures since a nal extraction step is
necessary to remove salts and to recover the isolated product.
Only few protocols for GAs preparative isolation are described
in the literature. -Chaconine and -solanine were isolated at a
preparativescalebySoulet al. byMPLC. 86.7mg and66.5mg were,
respectively, obtained from 1kg of potato peel [10]. The authors
mentioned the simplicity but the low yield obtained in the MPLC
separation. Moreover, the purity of solanine (85%) is not sufcient
for further use. Recently, Paul et al. developed an efcient LCMS
protocol for preparative scale isolation and quantication of two
steroidal alkaloids solamargine and solasonine fromSolanum xan-
thocarpum [11]. However, in addition to the equipment which is
not readily available, this method requires the use of formic acid
in the mobile phase. This may involve additional processing of the
puried compounds according to their future use because GAs can
be hydrolyzed in the presence of acid when they are evaporated to
dryness [7].
Countercurrent chromatography (CCC) is a chromatographic
technique which benets from some advantages when compared
with LC techniques: (i) no non-specic adsorption to a solid sup-
port, (ii) higher selectivity, (iii) higher sample loading capacity,
(iv) reduction of solvent quantity and (v) shorter separation time.
Fukuhara et al. described an efcient semi-preparative scale iso-
lation of GAs from Solanum incanum by the sequential use of
rotationlocular countercurrent chromatography and droplet coun-
tercurrent chromatography (RLCC and DCC) [12]. 170mg of pure
compounds were obtained from 550g of fresh ripe fruits of S.
incanum. A new GA, arudonine, was also isolated by CCC by
bioassay-guided fractionation of a root bark extract of Solanum
arundo [13]. A pH-zone-rening CPC was successfully imple-
mented for the preparative isolation of two GAs, solamargine
and solasonine [14]. Only solamargine was recovered in one
step by CCC while a further step using MPLC was necessary to
yield pure solasonine. In addition when using this technique,
removal of acid or base from the separated products may be
inconvenient [15].
The aimof this study was to develop an efcient method for the
isolationandpuricationof -chaconine and-solanine fromfresh
sprouts of Solanum tuberosum (cv. Pompadour) without sample
clean-up and in the shortest time. The objective was fully achieved
by the use of CPC which led to the successful separation of the
two GAs of high purity used for further chemical modulations and
biological studies.
2. Experimental
2.1. Chemical reagents
-Chaconine and -solanine used as reference standards were
purchased from Extrasynthese (France). All organic solvents were
analytical grade and purchased fromProlabo (France).
2.2. Apparatus
The CPC instrument used in this study is a SPOT CPC 100 Light
(Armen Instrument) tted with a rotor of 10 circular partition
disks (1000 partition cells: 0.1ml per cell; total column capacity
of 100ml). Rotation speed can be chosen from 0 to 4000rpm. The
efuent was monitored by a Lash 06 DAD detector (ECOM, Prague)
equipped with a preparative owcell operating at 202 and 210nm
and collected by a LS 5600 (Armen) fraction collector.
The HPLC used was a Shimadzu HPLC System including a LC-
20AT pump and a SPD-M20A diode-array detector.
LCMS spectra were performed on a Waters 2695 Alliance cou-
pled with a quadrupole mass spectrometer ZQ (Water-Micromass,
Manchester, UK) equipped with an electrospray ion source (ESI-
MS). LCESIMS were recorded in the positive and negative ion
mode. The capillary voltage was 3.5kV and a cone voltage range
from20 to 60V was used. Data acquisition and processing were
performed with MassLynx V4.0 software.
2.3. Preparation of crude extract
50g of freshsprouts of S. tuberosumcv. Pompadour were ground
and extracted by 550ml of CH
3
OH/H
2
O/CH
3
COOH (400/100/50,
v/v/v) under agitationfor 12hat roomtemperature. After ltration,
the mixture was concentrated under reduced pressure to obtain a
syrup consistence. This residue was dissolved in 75ml of CH
3
OH
and75ml of NH
4
OH(commercial solutionat 30%) was addedunder
cooling. The precipitate was recovered from the mixture by cen-
trifugation at 5000rpmfor 30min. 2.8g of residue were obtained,
re-dissolved in CH
3
OH and stored at 4

C for subsequent analysis

and separation. The operation was repeated 3 times.
2.4. HPLC and LCMS analyses
HPLC analyses of the crude extract and of the CPC peak
fractions were conducted on a 250mm4.6mm, 5-m, Prevail
reverse-phase C18 column (Grace) using a linear binary gradient
of H
2
OH
2
KPO
4
0.1M (solvent A) and CH
3
CN (solvent B) with a
ow rate of 1ml/min as follows: 2040% B (015min), 4080% B
(1530min), 80% B (3035min), 8020% B (3540min). The HPLC
eluate was monitored at 202nm. 20l were used for injection
which was repeated 3 times.
LCMS analyses of the samples were conducted on a
250mm4.6mm, 5-m, Prevail reverse-phase C18 column
(Grace) using a linear binary gradient of H
2
O(solvent A) andCH
3
CN
(solvent B) bothcontaining0.1%(v/v) formic acid, withaowrateof
1ml/min as follows: 2040% B (015min), 4080% B (1530min),
80% B (3045min), 8020% B (3540min). The HPLC eluate was
monitored at 202nm.
The calibrationcurves were preparedusingsixdifferent concen-
trations of the two glycoalkaloids in CH
3
OH. 20l of each solutions
ranging from 1 to 0.03125mg/ml (2-fold serial dilutions) were
injectedintriplicate inthe column. Calibrationgraphs were plotted
based on linear regression analysis of the peak area vs. concentra-
tion, the curves showed good linearity (r
2
=0.983 for chaconine and
0.988 for solanine).
2.5. CPC separation
2.5.1. Selection of the two-phase solvent system
The solvent system was selected according to the distribution
coefcient K
C
of -solanine and -chaconine. The K
C
value was
determined by HPLC analysis. Suitable amount of crude extract was
dissolved in the tested solvent systemand vortexed for 30s. After
separation and evaporation under reduced pressure, the residue
of each layer was dissolved in 500l of methanol for HPLC analy-
sis of its GA content. The K
C
was calculated according to the ratio:
concentration in the stationary phase/concentration in the mobile
phase [16].
2.5.2. Collection and analysis of fractions
The solvent system used for separation was ethyl
acetate/butanol/water in the ratio 15/35/50 (v/v/v). Biphasic
system was prepared just before use by thoroughly mixing vol-
umes of solvent in the above ratio. After the equilibration was
Please cite this article in press as: J. Attoumbr, et al., J. Chromatogr. B (2012), http://dx.doi.org/10.1016/j.jchromb.2012.09.025
ARTICLE IN PRESS
GModel
CHROMB-18104; No. of Pages 5
J. Attoumbr et al. / J. Chromatogr. B xxx (2012) xxxxxx 3
Fig. 1. Chromatogramof the crude extract of S. tuberosum(cv. Pompadour) sprouts by HPLC.
established, the stationary phase (lower phase in the ascending
mode) was pumped into the column at a ow-rate of 30ml/min
while the apparatus was run at 500rpm according to the equil-
ibration mode of the apparatus. After injection of the sample
(0.8g of crude extract in 10ml of the stationary phase) the
mobile phase was perfused at 2500rpmat a ow-rate of 8ml/min
under 2628b during the runs. The eluent was monitored at
202nm and fractions of 15ml were collected and analyzed by
HPLC. Fractions collected before the fraction number 8 contained
no GAs.
The volume of the stationary phase displaced by the mobile
phase was measured and used to determine V
M
. The station-
ary phase volume (V
S
) was calculated according to the relation
V
C
=V
M
+V
S
since the column volume (V
C
) is known. The stationary
phase retention was expressed as V
S
/V
C
. The general chromato-
graphic retentionequationwhichdirectlylinks retentionvolume of
a solute (V
R
) with its partition coefcient was used: V
R
=V
M
+K
C
V
S
to estimate the K
C
observed during the separation. The selectivity
(separation factor) and the number of theoretical plates were cal-
culated as follows: =K
C(solanine)
/K
C(chaconine)
and N=(4V
R
/W)
2
(W:
peak width at the baseline) [16].
3. Results and discussion
The 2.80.62g of crude extract produced from the treat-
ment of 50g of fresh sprouts were analyzed by HPLC (Fig. 1).
Two main peaks with retention times (Rt) of 14.96min and
15.18min were detected and were consistent to those of authen-
tic samples of -solanine and -chaconine, respectively. Fromthe
calibration curves, the amounts of -chaconine and -solanine
that could be recovered from the crude extract were estimated
to74.1311.62mg and to 34.1310.75mg, respectively. Consid-
ering that these two compounds mainly contribute to the total GA
content of commercial cultivar, we estimated that the average con-
tent of GAs alkaloids in sprouts of S. tuberosumcv. Pompadour was
216.5mg/100g FW. These results are consistent with the litera-
ture data (200700mg/100g FW in sprouts depending on potato
species [3]) and also conrmed that -chaconine contributed more
largely than -solanine (68% vs. 32%) to the total potato GA
content [2].
A rst test of separation of the two GAs by preparative HPLC
was performed but the results were not satisfactory. Only 0.35% of
-chaconine (95% purity) was obtained after two sequential steps
of HPLC separation and subsequent extraction with chloroformto
remove salts, while the amount of -solanine was too lowand was
not quantied. Given this poor result that was not really surprising
in the light of the literature, we decided to implement a separation
by CPC.
Separation of natural products using CPC is based on the parti-
tion behaviors of target compounds between immiscible solvents
used as a mobile and stationary phases. Key points in performing
CPCseparationareagoodsolubilityof thesampleinthesolvent sys-
tem as well as the determination of a suitable two-phase system.
As shown in Table 1, the mixture of ethyl acetate/butanol/water
(15/35/50, v/v/v) led to the best results with the compounds of
interest almost equally distributed between the two phases (K
C
value around 1).
Fig. 2 shows a typical CPC chromatogram obtained during the
separation of 0.8g of crude alkaloid extract. Each fraction collected
during the separationprocess was analyzedby HPLC(Fig. 2(A)(C)).
It should be noted that the elution order of the GAs was reversed
during the two methods used in this study: -chaconine was rst
eluted by CPC and presented a higher retention time by HPLC
demonstrating that -chaconine is more polar than -solanine. For
Table 1
The K
C
(distribution coefcient) values of -chaconine and -solanine in different
solvent systems.
Solvent systemcomposition K
C
value K
C
value
Ethyl acetate/butanol/water -Chaconine -Solanine
10/40/50 0.37 0.59
15/35/50 0.77 1.25
20/30/50 1.25 2.00
Please cite this article in press as: J. Attoumbr, et al., J. Chromatogr. B (2012), http://dx.doi.org/10.1016/j.jchromb.2012.09.025
ARTICLE IN PRESS
GModel
CHROMB-18104; No. of Pages 5
4 J. Attoumbr et al. / J. Chromatogr. B xxx (2012) xxxxxx
Fig. 2. CPC chromatogram and HPLC control [(A) -chaconine, (B) -solanine and -chaconine, (C) -solanine] for the separation of crude extract from S. tuberosum (cv.
Pompadour) sprouts.
each fraction, analysis of HPLC calibration curves allowed to deter-
mine the amounts of each of the two glycoalkaloids (Fig. 3) as well
as their purity (Table 2). Table 2 also presents for all the fractions,
the yield of each compound calculated on the mass of the crude
extract as well as the value of the cumulative yield throughout the
purication process.
From these results, it appears that -chaconine was easily iso-
lated under the conditions used with a global yield of 1.92%.
Fig. 3. -Chaconine and -solanine contents of the 12 CPC fractions as determined
by HPLCanalysis. The values represent the means S.D. of three independent exper-
iments.
Moreover, four of the ve fractions collected showed very high
purities. On the other hand, the isolation of -solanine was less
effective, the amount recovered was less important (yield 0.53%)
and only three fractions had a purity greater than 90%. To sup-
port this difference observed for the two compounds, the efciency
of the purication process can be calculated on the basis of the
values obtained by analytical HPLC on the crude extract. 54mg
of -chaconine (corresponding to F8 to F13 excepted F10) were
obtained while an amount of 74mg was expected resulting in an
Table 2
-Chaconine and -solanine amounts isolated from the fractionation of the crude
extract (2.8g) of Solanumtuberosum(cv. Pompadour) sprouts by CPC.
Fraction Purity
a
(%) Yield
b, c
(%) Cumulated yield (%)
F8 100.0 0.20
b

F9 91.9 0.15
b
0.35
F10 58.8
F11 100.0 0.23
b
0.58
F12 100.0 0.66
b
1.24
F13 100.0 0.68
b
1.92
b
F14 66.3
F15 87.0
F16 95.8 0.23
c

F17 98.4 0.24

c
0.47
F18 93.0 0.06
c
0.53
c
F19 78.3
a
Purity based on HPLC results.
b
Yields of -chaconine based on the weight of the crude extract for the fractions
of purity above 90%.
c
-Solanine based on the weight of the crude extract for the fractions of purity
above 90%.
Please cite this article in press as: J. Attoumbr, et al., J. Chromatogr. B (2012), http://dx.doi.org/10.1016/j.jchromb.2012.09.025
ARTICLE IN PRESS
GModel
CHROMB-18104; No. of Pages 5
J. Attoumbr et al. / J. Chromatogr. B xxx (2012) xxxxxx 5
Table 3
Experimental results obtained in CPC separation estimated on the basis of HPLC
results of Fig. 3 (see Section 2).
-Chaconine -Solanine
VR (HPLC) (ml) 240 285
Peak widths W(ml) 90 75
K
C
calculated preliminary in vials 0.77 1.25
K
C
observed during the run 3.54 4.36
isolation efciency of 73%. On the other hand, 15mg of -solanine
were isolated on an expected quantity of 34mg giving an efciency
of only44%. We sawbeforeintheintroductionthat studies toisolate
GAs fromplant material are often unique studies and it is therefore
difcult to have comparison points, however our results are quite
satisfactory within the context of the isolation of natural products.
Indeed, this result may berelatedtotheworkof Soulet al. [10] who
have isolated only about 86mg -chaconine from 1kg of potato
peel.
Under the conditions of CPC purication used, we never
observed usual chromatographic peaks but chromatograms similar
to those obtained during the implementation of pH-zone-rening
CPC [14]. To evaluate the parameters of the GAs purication pro-
cess, we estimated that the HPLC contents shown in Fig. 3 had
a shape that could be considered in a rst approximation as a
Gaussian peak. Fromthis hypothesis, we were able to evaluate the
retention volumes as well as the peak width at the baseline for
eachof thetwoGAs. Basedontheseresults, someparameters which
allowto get a rst overviewof the performance of the process were
calculated (Table 3).
For efcient CPC separation, the K
C
of the target compounds
shouldlie inthe range 0.5<K
C
<2.0. Moreover, the separationfactor
between any two components should be greater than 1.5 [17]. In
our case, the value calculated fromthe K
C
preliminary extraction
data set is 1.62, whereas the value obtained by the estimated
values of Table 3 is 1.23.
Given the values of the parameters Sf =0.55 and =162 a much
higher resolution of our peaks was expected [16]. Several explana-
tions can be proposed to explain the shape of the chromatogram.
Firstly, the column loading is an important factor. In our separa-
tion conditions, we introduced 800mg of sample on a column of
100ml, i.e. a loadingvalueof 8mg/ml whichis highcomparedtothe
average loading of 2.2mg/ml observed in the literature and could
contribute to produce peak broadening and overlapping [18]. This
peak broadening as well as the differences between the values
could also be explained by the particular nature of the compounds
that we try to separate. According to Pauli et al. [18], highly concen-
trated or problematic samples can cause disruptions to the solvent
system equilibrium inside a running CC instrument and result in
trouble during the separation. The amphiphilic complex nature of
our sample can cause a signicant stationary phase loss during the
runand/or candisturb the mass transfer. Insupport of this explana-
tioninrelationto the chemical structure of the studiedcompounds,
Berthod et al. recently showed a discrepancy between the mea-
sured and calculated separation factors during the separation of
benzoic acid that could be explained by the specic nature of the
compound [19].
4. Conclusion
In this study, CPC was successfully implemented for the iso-
lation of polar and close Rt steroidal alkaloids -chaconine and
-solanine whereas their separation was almost impossible by
semi-preparative HPLC. Despite an unusual formof chromatogram
and disruptions related to the nature of their structures, we
obtained the alkaloids of interest with good yields and high purity
(higher than 93%). These results demonstrate the high effective-
ness of this methodology for providing signicant amounts of these
bioactive glycoalkaloids for further chemical modulations and bio-
logical studies, we werethus abletoincreasetheyieldfrom0.35%to
1.92% for -chaconine using CPC rather than HPLC. The conditions
developed in this study allow to isolate large amounts of complex
compounds within a few hours, without any preliminary cleanup
or concentrations steps and avoid the problem of poor detection
of alkaloids in UV. In addition, a scaling can easily be considered
which is quite crucial in our case to achieve our future research
objectives.
Acknowledgments
This work was supported by the Conseil Regional de Picardie,
the Fond Sucre, the S.I.P.R.E Comit Nord and was labeled by
the Ple de Comptitivit IAR.
We gratefully acknowledged Dr. Serge Pilard, the Director of the
Plate-FormeAnalytique from UPJV for providing facilities, help
and encouragement.
References
[1] M. Friedman, G.M. McDonald, Naturally Occurring Glycosides: Chemistry, Dis-
tribution and Biological Properties, Wiley, NewYork, 1999.
[2] M. Friedman, J. Agric. Food Chem. 54 (2006) 8655.
[3] S.E. Milner, N.P. Brunton, P.W. Jones, N.M. OBrien, S.G. Collins, A.R. Maguire, J.
Agric. Food Chem. 59 (2011) 3454.
[4] M. Stobiecki, I. Matysiak-Kata, R. Franski, J. Skala, J. Szopa, Phytochemistry 62
(2003) 959.
[5] M. Friedman, F.F. Bautista, L.H. Stanker, K.A. Larkin, J. Agric. Food Chem. 46
(1998) 5097.
[6] M. Distl, M. Wink, Potato Res. 52 (2009) 79.
[7] E.J. Edwards, A. Cobb, J. Agric. Food Chem. 44 (1996) 2705.
[8] H. Kodamatani, K. Saito, N. Niina, S. Yamazaki, Y. Tanaka, J. Chromatogr. A 1100
(2005) 26.
[9] A. Sotelo, B. Serrano, J. Agric. Food Chem. 48 (2000) 2472.
[10] S. Soul, A. Vazquez, G. Gonzalez, P. Moyna, F. Ferreira, Potato Res. 40 (1997)
413.
[11] A.T. Paul, S. Vir, K.K. Bhutani, J. Chromatogr. A 1208 (2008) 141.
[12] K. Fukuhara, I. Kubo, Phytochemistry 30 (1991) 685.
[13] K. Fukuhara, K. Shimizu, I. Kubo, Phytochemistry 65 (2004) 1283.
[14] A. Maurya, S. Gupta, S. Negi, S.K. Srivastava, J. Sep. Sci. 32 (2009) 3126.
[15] A. Marston, K. Hostettmann, J. Chromatogr. A 1112 (2006) 181.
[16] W.D. Conway, J. Chromatogr. A 1218 (2011) 6015.
[17] V.G.P. Severino, C.de M. Cazal, M.R. Forim, M.F.d.G.F. da Silva, E. Rodrigues-Filho,
J.B. Fernandes, P.C. Vieira, J. Chromatogr. A 1216 (2009) 4275.
[18] G.F. Pauli, S.M. Pro, J.B. Friesen, J. Nat. Prod. 71 (2008) 1489.
[19] A. Berthod, N. Mekaoui, J. Chromatogr. A 1218 (2011) 6024.