Isolation, identification and antioxidative capacity of water-soluble phenylpropanoid compounds from Rhodiola crenulata

Article abstract of DOI:10.1016/j.foodchem.2012.04.011

Six water-soluble phenylpropanoid compounds obtained from Rhodiola crenulata (R. crenulata) were fractionated by high-speed counter-current chromatography (HSCCC), and purified by semi-preparative high-performance liquid chromatography (Semi-prep HPLC). The purities of the six compounds were all above 98.0% and their structures were identified by spectroscopic methods. Among them, a new compound, 2-(4-hydroxyphenyl)-ethyl-O-β-d-glucopyranosyl-6-O- β-d-glucopyranoside (1), together with two known phenylpropanoids, p-hydroxyphenacyl-β-d-glucopyranoside (3) and picein (4) were isolated from R. crenulata for the first time. Meanwhile, the contents of six isolated ingredients from the crude extract of R. crenulata had been simultaneously detected, with satisfactory results. Furthermore, the antioxidant activities of the six compounds were accessed by measuring the radical scavenging activity against 2,2-diphenyl-1-picrylhydrazy (DPPH), and four compounds exhibited potent antioxidative activity.

Full text of DOI:10.1016/j.foodchem.2012.04.011

Food Chemistry 134 (2012) 2126–2133
Contents lists available at SciVerse ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Isolation, identification and antioxidative capacity of water-soluble
phenylpropanoid compounds from Rhodiola crenulata
⇑
Danjun Chen, Junting Fan, Peng Wang, Lanying Zhu, Yang Jin, Yan Peng, Shuhu Du
School of Pharmacy, Nanjing Medical University, Nanjing 210029, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Six water-soluble phenylpropanoid compounds obtained from Rhodiola crenulata (R. crenulata) were frac-
tionated by high-speed counter-current chromatography (HSCCC), and purified by semi-preparative
high-performance liquid chromatography (Semi-prep HPLC). The purities of the six compounds were
all above 98.0% and their structures were identified by spectroscopic methods. Among them, a new com-
Received 6 December 2011
Received in revised form 12 March 2012
Accepted 3 April 2012
Available online 13 April 2012
pound, 2-(4-hydroxyphenyl)-ethyl-O-b-
D
-glucopyranosyl-6-O-b-D-glucopyranoside (1), together with
two known phenylpropanoids, p-hydroxyphenacyl-b-
D
-glucopyranoside (3) and picein (4) were isolated
Keywords:
Rhodiola crenulata
High-speed counter-current
chromotography
Semi-preparative high-performance liquid
chromatography
from R. crenulata for the first time. Meanwhile, the contents of six isolated ingredients from the crude
extract of R. crenulata had been simultaneously detected, with satisfactory results. Furthermore, the anti-
oxidant activities of the six compounds were accessed by measuring the radical scavenging activity
against 2,2-diphenyl-1-picrylhydrazy (DPPH), and four compounds exhibited potent antioxidative
activity.
Identification
Antioxidative activity
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
Previous phytochemical studies have reported the isolation of
many compounds from this plant, including phenols, phenylprop-
Rhodiola crenulata (Hook. f. et Thoms) H. Ohba, as an important
member of the Crassulaceae family, is mainly distributed in the
high cold region of Yunnan and Sichuan province as well as the
Tibetan Autonomous Region. Its roots have been usually used as
a health food, antidepressant and antifatigue for a long time. Mod-
ern pharmacological studies have demonstrated that the genus
Rhodiola possessed the functions of reinforcing immunity (Mishra,
Padwad, Jain, Karan, Ganju, & Sawhney, 2006), improving memory
(Fan, Tezuka, Namba, & Kadota, 1999; Tezuka, Fan, Kasimu, &
Kadota, 1999), scavenging active-oxygen species (Furmanowa,
_
anoids, phenyethanoids, flavonoids, monoterpenoids, cyanogens
and their corresponding glycosides (Du & Xie, 1994; Nakamura,
Li, Matsuda, & Yoshikawa, 2008; Peng, Ma, & Ge, 1995; Wu, Guo,
Guo, Li, Wang, & Ma, 2008; Yang et al., 2012). Among the isolated
compounds, tyrosol, gallic acid and crenulatin had been proved to be
good radical scavengers, based on the scavenging activity of the sta-
ble 2,2-diphenyl-1-picrylhydrazy (DPPH) free radical (Domínguez,
Torres, & Núñez, 2001; Lee et al., 2000). The above investigation
suggests that it would also be a very interesting and plentiful
source of antioxidants. Therefore, further chemical research from
this herb is warranted for new pharmaceutical preparations or
functional products and DPPH radical scavenging capacity assay.
The classical methods of preparative separation and purification
of compounds from complex plant extracts usually require
multiple chromatographic steps on silica gel, polyamide, Sephadex
LH-20 column, and so on. These methods often lead to the loss of
active components (due to the dilution effects and/or decomposi-
tion of the components, especially antioxidants during the
isolation and purification processes) (Hostettmann, Wolfender, &
Terreaux, 2001). In order to avoid the problems mentioned above,
it is necessary to develop more efficient methods for separating the
bioactive compounds from natural products. High-speed counter-
current chromatography (HSCCC), being a continuous liquid–liquid
partition chromatographic method, eliminates complications such
as irreversible adsorption on the solid support, denaturation of
´
Skopinska-Rózewska, Rogala, & Hartwich, 1998; Ohsugi et al.,
1999), lowering blood glucose (Kwon, Jang, & Shetty, 2006; Yang,
Liu, Feng, Jiang, & Zhang, 2012), resisting tumour (Dement’eva &
Iaremenko, 1987; Udintsev & Schakhov, 1991), and so on. Espe-
cially, R. crenulata has the bi-directional function of accommodat-
ing central nervous system and endocritic system, and keeps the
body in excellent balance situation (You et al., 2003). Additionally,
extracts of R. crenulata are usually made into pharmaceutical prep-
arations or functional foods used by astronauts, pilots, divers and
sports who work under special circumstances (Shen, Jiang, Gu,
Qiu, Shen, & Yang, 2010; Zhao et al., 1998). Owing to the important
virtues of R. crenulata, its study is a subject of great interest.
⇑
Corresponding author. Tel./fax: +86 25 86862762.
E-mail address: shuhudu@njmu.edu.cn (S. Du).
0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2012.04.011

D. Chen et al. / Food Chemistry 134 (2012) 2126–2133
2127
sample, tailing of the solute peaks, etc. (Ito, 2005). Based on the
2.2. Regents and materials
advantages of the HSCCC system, it has been successfully applied
to the separation of active constituents of various herbs (Du, Li, &
Ito, 2011; Han, Zhang, Wei, Cao, & Ito, 2002).
All solvents used for HSCCC were of analytical grade (Merck,
Darmstadt, Germany). Methanol used for analytical HPLC was of
chromatographic grade (Merck, Darmstadt, Germany). Deionized
and purified water was prepared using a water system (Millipore,
Bedford, USA). D-101 Macroporous resin was purchased from the
chemical plant of Nankai University (Tianjin, China). 2,2-Diphe-
The purpose of this work was to set up a very convenient and
efficient method for the rapid separation of water-soluble phenyl-
propanoid compounds from R. crenulata and for screening their
antioxidative activities. Firstly, water-soluble phenylpropanoid
compounds from R. crenulata were isolated by offline coupling of
HSCCC and semi-preparative high-performance liquid chromatog-
raphy (Semi-prep HPLC). Secondly, the chemical structures of the
isolated compounds were elucidated by ultraviolet spectra (UV),
infrared spectrum (IR), high resolution mass spectrum (HRMS)
and nuclear magnetic resonance spectrum (NMR). Moreover, the
method of simultaneous determination of isolated compounds
from the crude extract of R. crenulata was developed. Finally, the
antioxidant activities of the pure compounds were determined
by measuring the radical scavenging activity against DPPH, which
may give support for the traditional use of the herb.
nyl-1-picrylhydrazy (DPPH),
D-glucose and L-glucose were pur-
chased from Sigma–Aldrich (Steinheim, Germany). DPPH radical
solution was freshly prepared in methanol every day and kept pro-
tected from light. Rutin standard was supplied by National Insti-
tute of the Control of Pharmaceutical and Biological Products
(Beijing, China).
The dried roots of R. crenulata (Hook. f. et Thoms) H. Ohba were
collected from the Tibetan Autonomous Region of China in August,
2009. The botanical origin of the material was identified and the
voucher specimens were deposited at the School of Pharmacy,
Nanjing Medical University (Nanjing, China).
2.3. Preparation of crude extract
2. Materials and methods
The dried roots of R. crenulata (2.0 kg) were pulverised and ex-
tracted three times (1 h for each time), each time with 6000 ml of
80% ethanol. The extract solutions were combined and concen-
trated by rotary evaporator at 65 °C to form a brownish syrup
(about 400 g). This was dissolved in hot water by sonication,
statically cooled down and filtered. The filtrate was partitioned
three times with equal volume of ethyl acetate. The ethyl acetate
fractions (upper layer) were removed. Subsequently, a suitable
volume of the lower layer was added into a glass column of mac-
roporous resin (1.8 ꢀ 70 cm, contained 80 g of D-101 macroporous
resin). The resin was washed by water until the effluent almost
colorless, and sequentially eluted with 30% ethanol. The elution
fractions were collected and evaporated to dryness under reduced
pressure. Finally, 15 g of crude extract was obtained, which was
stored in the refrigerator for the further HSCCC isolation.
2.1. Apparatus
The HSCCC apparatus employed in the current study is
TBE-300A high-speed counter-current chromatography (Tauto
Biotechnique Company, Shanghai, China) with three multilayer coil
separation columns connected in series (i.d. of the tubing is
1.6 mm, total volume is 260 ml) and a 20 ml manual sample loop.
The revolution speed of the instrument could be regulated with a
speed controller in the range from 0 to 1000 rpm. An HX 1050
constant temperature circulator (Beijing Boyikang Lab Instrument
Company, Beijing, China) was used to control the separation
temperature. An ÄKTA prime system (GE Healthcare, Uppsala,
Sweden) was used to pump the two-phase solvent system and
performed the UV absorbance measurement. The detection wave-
length was 280 nm. An automatic fraction collector (GE Healthcare,
Uppsala, Sweden) was used to collect the fractions. The data was
collected with N2000 chromatography workstation (Hangzhou
Zhida Science Apparatus Company, Hangzhou, China).
2.4. Procedure of HSCCC pre-separation
The two-phase solvent system composed of ethyl acetate–n-
butanol–water (0.5:4.5:5, v/v/v) was used for HSCCC separation.
Each component of the solvent system was added to a separatory
funnel and thoroughly equilibrated by shaking at room tempera-
ture. The upper phase and lower phase were separated and de-
gassed by sonication for 30 min shortly before use.
The sample solution for the HSCCC pre-separation was prepared
by dissolving 150 mg of the dried powder of crude extract in 20 ml
of the lower phase of two-phase solvent system. The HSCCC pre-
separation was performed as follows: the coiled column was first
pumped with the upper phase of the solvent system and then
the lower phase was pumped into the column at a flow rate of
2.0 ml/min, meanwhile the apparatus was rotated at 850 rpm.
After a hydrodynamic equilibrium was reached, 20 ml of the sam-
ple solution were injected into the separation column. The effluent
was monitored with UV detector at 280 nm and the peak fractions
were automatically collected according to the chromatographic
profile.
Analytical high-performance liquid chromatography (HPLC)
was performed on a Shimadzu LC-10A system equipped with a
UV detector (monitoring at 275 nm) and Shimadzu LC-20A system
equipped with SPD-M20A detector. The semi-preparative high-
performance liquid chromatography (Semi-prep HPLC) equipment
used consisted of two LC-6A pumps with a UV–visible detector
(monitoring at 275 nm) and a Phenomenex C18 column (250 ꢀ
10.0 mm, 8 lm).
The IR spectrum was obtained on a Tensor 27 FT-IR spectro-
photometer with KBr pellet (Bruker, Ettlingen, Germany). The
UV spectrum was obtained on a UV-2501 spectrometer (Shima-
dzu, Toyko, Janpan). HRMS was recorded on a 6520 Accurate-Mass
Q-TOF LC/MS (Agilent Technologies, Santa Clara, USA). The NMR
spectrometer used was AV-300 MHZ NMR (Bruker, Fällanden,
Switzerland). The reference compound tetramethylsilane (TMS)
was used as an internal standard for the determination of chem-
ical shifts. Gas chromatograph (GC) analysis was performed on a
HP5890 gas chromatograph (Agilent Technologies, Massy, France)
equipped with
0.32 mm, 0.25
a
30QC2/AC-5 quartz capillary column (30 ꢀ
2.5. HPLC analysis of HSCCC peak fractions
l
m) and a flame ionisation detection (FID). The
optical rotation was determined on a p-1020 polarimeter (Jasco,
Tokyo, Japan). 96 Well microtiter plates (Costar, New York, USA)
and a Spectra Max M2 microplate reader (Molecular Devices, Sun-
nyvale, USA) were used in the radical scavenging activity
experiments.
Each fraction isolated from HSCCC was analysed by analytical
HPLC, which was performed with Phenomenex C18 column
(250 ꢀ 4.6 mm, 5
nol–water (12:88, v/v) and the flow rate was 1.0 ml/min. All sam-
ple solutions were filtered through 0.45 microporous
lm). The mobile phase was composed of metha-
a
lm

2128
D. Chen et al. / Food Chemistry 134 (2012) 2126–2133
membrane before injection. The injection volume was 20
the detection was carried out at 275 nm.
l
l and
day, while for the inter-day variability test, the solutions were de-
tected in duplicates for consecutive three days. Variations were ex-
pressed in terms of relative standard deviation (RSD).
2.6. HPLC purification of HSCCC peak fractions
The recovery was used to evaluate the accuracy of the method.
A known amount of six reference compounds (compounds 1–6)
were added into a certain amount of crude extract solution. Three
replicates were performed in the test. The precision was expressed
in terms of RSD.
To confirm the repeatability, six replicates of the same sample
(crude extract) were analysed as mentioned above. The RSD value
was calculated as a measurement of the method repeatability.
Each fraction isolated from HSCCC was purified by Semi-prep
HPLC, which was performed with a Phenomenex C18 column
(250 ꢀ 10.0 mm, 8
lm). The mobile phase (ratio of methanol to
water) was determined based on the analytical HPLC results and
the flow rate was 4.0 ml/min, the column temperature was at
25 °C. All sample solutions were filtered through 0.45 lm filter be-
fore delivering into the system. The injection volume was 200
ll
and the chromatograms recorded at 275 nm.
2.10. Analysis of crude extract
2.7. Identification of the purified compounds
The crude extract (0.10 g) was transferred into 50 ml volumetric
flask which was made up to its volume with 30% methanol and fil-
tered through a 0.45 lm nylon membrane filter prior to injection
A Tensor 27 FT-IR spectrometer with a resolution of 2 cmꢁ1 and
a spectral range of 4000–500 cmꢁ1 was employed to examine
infrared spectra of samples by a pressed tablet (sample:KBr = 1:14
in mass). UV spectrum was recorded on a UV 2501 spectrometer.
HRMS data was measured on a 6520 Accurate-Mass Q-TOF LC/
MS with an ESI source, a negative ion mode was employed. The
1H NMR, 13C NMR and 2D NMR experiments were performed on
a AV-300MHZ NMR spectrometer using DMSO-d6 as solvent.
into the HPLC system. Then the contents of the isolated compounds
(compounds 1–6) from the crude extract were determined by HPLC,
which was performed with Phenomenex C18 column (250 ꢀ
4.6 mm, 5 lm). The mobile phase was composed of methanol–
water (15:85, v/v) and the flow rate was 1.0 ml/min. The injection
volume was 20 l and the detection was carried out at 275 nm.
l
2.11. Evaluation of antioxidative activity
2.8. Acid hydrolysis of compound 1
The antioxidative activities of compounds 1–6 were assessed on
the basis of the scavenging activity of the stable DPPH free radical
(Brand-Williams, Cuvelier, & Berset, 1995; Shi, Zhao, Zhou, Zhang,
Jiang, & Huang, 2008). Typically, isolated compounds and reference
compound were dissolved in methanol to form various concentra-
The isolated compound 1 (4.0 mg) was hydrolysed with 2 M HCl
in 1,4-dioxane (1:1, 4 ml) under reflux at 80 °C for 6 h. Each reac-
tion mixture was extracted with CHCl3 three times (2 ml ꢀ 3).
The aqueous layer was neutralised with 2 M NaOH and then dried
to give one or more monosaccharides. The dried powders were dis-
tions of test samples, respectively. 100
ll of test samples mixed
solved in pyridine (2 ml), L-cysteine methyl ester hydrochloride
with 100 l DPPH methanol solution (100
l
lg/ml) in 96 well micro-
(1.5 mg) was added, and the mixture was heated at 60 °C for 1 h.
Subsequently, trimethylsilylimidazole (1.5 ml) was added to the
above mixture in ice-cold water and kept at 60 °C for another
titer plates, respectively. Meanwhile, the methanol solution of
DPPH was served as a control, which was prepared freshly every
day. The absorbance was measured at 517 nm after the mixture
was kept in the dark for 30 min at room temperature. The antiox-
idant activity was expressed as radical scavenging percentage
(RSP) calculated according to the following equation:
30 min. An aliquot of the supernatant (4 ll) was directly subjected
to GC analysis under the following conditions: column tempera-
ture, 180–280 °C; programmed increase, 3 °C/min; carrier gas, N2
(1.0 ml/min); injector and detector temperature, 250 °C; split ratio,
A1 ꢁ A2
1:50. The configuration of
mined by comparing the retention times with the derivatives of
references ( -glucose and -glucose).
D-glucose for compound 1 was deter-
RSP ¼
ꢀ 100%
A1
D
L
where A1 is the absorbance of the DPPH solution and A2 is the absor-
bance of the DPPH solution after the addition of samples. The sam-
ple concentration providing 50% inhibition (IC50) was calculated
from the graph plotting inhibition percentage. All tests were oper-
ated in triplicate.
2.9. Validation study of analytical HPLC
2.9.1. Calibration curves
Stock solution containing six reference compounds (compounds
1–6) was prepared and diluted to appropriate concentrations for
construction of calibration curves. Each concentration of the mixed
standard solution was injected in duplicate, and then the calibra-
tion curves were constructed by plotting the peak area versus
the concentration of each analyte.
3. Results and discussion
3.1. Optimisation of suitable HSCCC solvent system
The optimum of the separation conditions for HSCCC is related
to various parameters including a two-phase solvent system, revo-
lution speed, flow rate, column temperature, etc. In order to select
the optimum two-phase solvent system, a series of tests were per-
formed in this study. According to the rules in selecting the optimal
conditions introduced by Ito (Ito, 2005), the solvent system of ethyl
acetate–n-butanol–water would be suitable for water-soluble
compounds (i.e., salidroside, tyrosol, picein, etc.). Also, several
types of the solvent systems composed of ethyl acetate–n-buta-
nol–water at different volume ratios were chosen and assessed
for K-values. The K-values of compounds 1–6 for the solvent sys-
tem under the volume ratios of 3:2:5, 4:1:5, 1:4:5 and 0.5:4.5:5
were summarised in Table 1, respectively. It could be seen that
2.9.2. LOD and LOQ
The stock solution containing six reference compounds (com-
pounds 1–6) was diluted to a series of appropriate concentrations,
and an aliquot of the diluted solution was injected into HPLC for
analysis. The limits of detection (LOD) and quantification (LOQ) un-
der the present chromatographic conditions were determined at a
signal-to-noise ratio (S/N) of about 3 and 10, respectively.
2.9.3. Precision, accuracy and repeatability
Intra- and inter-day variations were selected to determine the
precision of the method. For intra-day variability test, the mixed
standard solutions were analysed for six replicates within one

D. Chen et al. / Food Chemistry 134 (2012) 2126–2133
2129
Table 1
The K values of the target components in ethyl acetate–n-butanol–water.
Solvent system (v/v/v)
K
Compound 1
Compound 2
Compound 3
Compound 4
Compound 5
Compound 6
3:2:5
4:1:5
1:4:5
0.5:4.5:5
0.19
0.15
0.26
0.29
0.39
0.37
0.63
0.68
0.33
0.30
0.48
0.49
0.33
0.32
0.51
0.52
0.73
0.72
0.84
0.85
0.68
0.66
0.81
0.83
Fig. 1. HSCCC chromatogram of the crude extract from R. crenulata. Experimental conditions: two-phase solvent system, ethyl acetate–n-butanol–water (0.5:4.5:5, v/v/v);
stationary phase, upper phase; mobile phase, lower phase; elution mode, head to tail; flow rate, 2.0 ml/min; revolution speed, 850 rpm; separation temperature, 30 °C;
detection wavelength, 280 nm; sample loading, 150 mg/20 ml; retention of stationary phase: 68.9%.
Fig. 2. Semi-prep HPLC chromatogram of compound 1 (A), compound 3, compound 4 and compound 5 (B), and compound 6 (C). Experimental conditions: column,
Phenomenex C18 column (250 ꢀ 10.0 mm, 8
lm); mobile phase, methanol–water (12:88, v/v) (A), methanol–water (11:89, v/v) (B) and methanol–water (10:90, v/v) (C),
respectively; flow rate, 4.0 ml/min; detection wavelength, 275 nm; column temperature, 25 °C.
3:2:5, 4:1:5 had minor K-value and when performed on HSCCC, the
In the 0.5:4.5:5 solvent system, several kinds of flow rates (1.0,
peak was overlapped. The other 1:4:5 and 0.5:4.5:5 had appropri-
ate K-values. Further HSCCC separation was performed and the re-
sults showed that the solvent system composed of ethyl acetate–n-
butanol–water (0.5:4.5:5, v/v/v) could achieve a better retention of
stationary phase (60.9%) for preparative HSCCC, and the four frac-
tions (1–4) could be separated along with acceptable separation
time. So, it was selected as the solvent system of HSCCC in the fol-
lowing studies.
1.5, 2.0 and 2.5 ml/min) were tested. The results showed that high
flow rate (2.5 ml/min) was not good for the retention of the sta-
tionary phase and reducing the flow rate (1.0 and 1.5 ml/min)
could improve the retention of the stationary phase to some de-
gree, but the separation time would be extended. Thus, the flow
rate was set at 2.0 ml/min in the experiment.
In general, the revolution speed of the apparatus also has an
influence on the retention percentage of the stationary phase and

2130
D. Chen et al. / Food Chemistry 134 (2012) 2126–2133
Table 2
Finally, when the solvent system of ethyl acetate–n-butanol–
water at a volume ratio of 0.5:4.5:5, flow rates of 2.0 ml/min, rev-
olution speed of 850 rpm and column temperature of 30 °C were
used, respectively, four fractions (1–4) were obtained by one-step
elution and the whole separation time was less than 3 h (see
Fig. 1). As shown in Fig. 1, the HPLC analysis of HSCCC fraction
‘‘1’’ (peak I, collected during 80.5–92.5 min) revealed that it con-
tained compound 1 and a minor unknown compound. Fraction
‘‘2’’ (peak II, collected during 95–118 min) contained pure com-
pound 2 (26.5 mg), the purity of which was 98.3%. Fraction ‘‘3’’
(peak III, collected during 122.5–137 min) included compounds 3,
4 and 5. Fraction ‘‘4’’ (peak IV, collected during 138–160.5 min)
consisted of compounds 5 and 6.
NMR data (300 MHz, DMSO-d6) and HMBC correlations of compound 1.
Position
dC
Mult.
dH (J in Hz)
HMBCa
1
127.6
129.6
C
2, 6
3, 5
4
7
8
CH
CH
C
CH2
CH2
6.98, d (8.4)
6.59, d (8.4)
3, 4, 5, 7
1, 4
115.2, 115.3b
156.7
34.8
69.9
2.70, t (7.5)
3.86, dd (7.5, 17.1)
3.58, overlap
4.16, d (7.8)
2.95, overlap
3.14, m
1, 2, 6, 8
10, 1
10
20
30
4’
50
60
102.7
73.3
76.7
70.0
75.7
68.4
CH
CH
CH
CH
CH
CH2
8
10, 30
20, 40
30
3.05, overlap
3.30, m
3.97, brd (10.5)
3.58, overlap
4.25, d (7.8)
2.95, overlap
3.10, overlap
3.10, overlap
3.05, overlap
3.66, brd (11.1)
3.43, m
100
3.2. HPLC purification of HSCCC peak fractions
100
200
300
400
500
600
103.3
73.5
76.6
70.1
76.8
61.0
CH
CH
CH
CH
CH
CH2
60
100, 300
In order to obtain pure compounds, the samples of fraction ‘‘1’’,
fraction ‘‘3’’ and fraction ‘‘4’’ were subjected to Semi-prep HPLC to
be purified further (Fig. 2). The conditions of Semi-prep HPLC
experiments were directly transferred from the analytical HPLC re-
sults. The effluent of each target compound was collected and ana-
lysed by analysis HPLC-DAD. The results indicated that the purities
of compound 1 (20.9 mg), 3 (11.1 mg), 4 (10.2 mg), 5 (120.2 mg)
and 6 (16.2 mg) were 99.6%, 98.2%, 98.7%, 99.1% and 98.9%,
respectively.
200, 400, 500
500
300
400
a
HMBC correlations are from proton (s) stated to the indicated carbon.
The signals of C-3 and C-5 could be changed.
b
3.3. Structure elucidation of the purified compounds
Compound 1 was obtained as a white powder with a negative
specific rotation ð½a D25
ꢂ
ꢁ 34:4ꢃÞ. Its molecular formula was estab-
lished as C20H30O12 on the basis of negative-mode HRMS (m/z
461.1669 [MꢁH]ꢁ, calcd 461.1664), indicating 6 degrees of unsatu-
ration. Its UV spectrum showed the existence of a phenyl group
based on the absorptions at 228, 278 nm. The IR absorption bands
indicated the presence of hydroxyl (3385 cmꢁ1) and aromatic
(1614, 1516 cmꢁ1) groups. In the 1H NMR spectrum (Table 2), two
aromatic proton signals resonated at [dH 6.98 (2H, d, J = 8.4 Hz, H-
2, 6), dH 6.59 (2H, d, J = 8.4 Hz, H-3, 5)] indicating a para-substituted
benzene ring. The 13C NMR and DEPT spectrum of 1 displayed 20
carbon signals corresponding to six aromatic carbons, two methy-
lene carbons, and twelve carbons arising from two glucopyranosyl
units. The 1H NMR and 13C NMR showed spectral features similar
Fig. 3. Selected 1H–1H COSY and HMBC corrections of compound 1.
to those of 30-O-b-
D-glucopyranosylsalidroside (Machida, Ohkawa,
Ohsawa, & Kikuchi, 2009). In the heteronuclear multiple quantum
correlation (HMBC) spectrum, the methylene signal at dH 2.70
(H-7) was attached to C-1, as deduced from HMBC correlations of
H-7 with C-1, C-2, and C-6. Then, the 1H–1H COSY correlations
of H-8 (dH 3.58, 3.86) with H-7, as well as the HMBC correlation
of H-8 with C-1, suggested the linkage of C-8 and C-7. As for the glu-
copyranosyl moieties of 1, HMBC correlations of the anomeric pro-
ton (dH 4.16, d, J = 7.8 Hz, H-10) (b-form) with C-8, and of H-8 with
C-10 indicated the glucose unit was attached to C-8. Moreover, the
H-100 signal at dH 4.25 (d, J = 7.8 Hz) (b-form) showed a HMBC cor-
relation with C-60 (dC 68.4), which implied a (1 ? 6) linkage of the
two glucose units (Fig. 3). Furthermore, based on the molecular
mass or fragment mass of 1 together with the apparent downfield
shift of C-4 (dC 156.7), a hydroxy group was assigned to be located
Fig. 4. The chemical structures of six water-soluble phenylpropanoid compounds.
2-(4-hydroxyphenyl)-ethyl-O-b- -glucopyranosyl-6-O-b- -glucopyranoside (1),
-glucopyranoside (3), picein (4), salidroside
D
D
icariside D2 (2), p-hydroxyphenacyl-b-
(5), tyrosol (6).
D
at C-4. Acid hydrolysis of 1 yielded tyrosol (6) and
absolute configuration of glucose was determined by GC analysis
of its corresponding trimethylsilylated -cysteine adduct. Therefore,
D-glucose. The
expediting the rotary can increase the retention of the stationary
phase. Although the revolution speed of the apparatus could be
L
performed with
a
speed controller in
a
range from
0
to
from these structural elements we conclude that compound 1 is a
1000 rpm, complete separation was invariably achieved at
850 rpm in this test. Meanwhile, the influence of column temper-
ature was investigated. After tested at 20, 25, 30 and 35 °C, it could
be seen that good results would be obtained when the column
temperature was controlled at 30 °C.
phenylethanoid disaccharide and defined as 2-(4-hydroxy-
phenyl)-ethyl-O-b-D-glucopyranosyl-6-O-b-D-glucopyranoside.
Compounds 2–6 (Fig. 4) were identified by UV, IR, MS, 1H NMR,
13C NMR and specific rotation measurements (figures of 1H NMR of
compounds 2–6 in supplementary information). Their spectral and

D. Chen et al. / Food Chemistry 134 (2012) 2126–2133
2131
Table 3
Linear regression data, limit of detection (LOD), limit of quantification (LOQ) of the isolated compounds.
Compounds
Linearity range (
lg/ml)
Linear regression equation
Correlation coefficient (r2)
LOD (l
g/ml)
LOQ (lg/ml)
1
2
3
4
5
6
1.50–31.12
2.00–56.03
0.50–16.81
0.50–12.56
12.00–180.21
1.50–36.76
y = 24.452x + 1.302
y = 49.790x + 2.265
y = 34.785x ꢁ 0.043
y = 30.623x ꢁ 0.051
y = 81.012x + 3.923
y = 48.603x + 1.589
0.9997
0.9998
0.9997
0.9996
0.9999
0.9998
0.04
0.05
0.01
0.01
0.10
0.05
0.25
0.30
0.10
0.08
0.80
0.28
Table 4
Recoveries for the assay of six isolated compounds from crude extract.
Analytes
Originals (
lg)
Spiked (
lg)
Founda
(
lg)
Recoveryb (%)
RSD (%)
60.1
75.1
90.1
118.6
148.0
173.4
33.1
42.1
50.8
25.4
30.2
134.0
149.9
163.8
255.5
286.8
311.6
72.1
80.6
89.6
54.6
59.4
98.2
99.7
98.6
98.8
100.3
99.9
99.7
98.6
99.4
99.6
99.7
99.2
99.7
100.1
99.4
101.6
99.3
98.7
2.2
2.0
1.9
2.3
1.8
2.0
2.7
2.9
2.4
3.0
2.5
2.4
1.9
1.5
1.1
2.1
2.4
2.7
Compound 1
75.0
Compound 2
Compound 3
Compound 4
Compound 5
Compound 6
138.3
39.1
29.3
38.1
67.1
376.4
470.0
564.3
80.2
100.1
121.4
841.9
937.2
1027.6
180.2
198.1
218.5
466.7
98.7
a
The data was average of three determinations.
Recovery (%) = (amount found ꢁ original amount)/amount spiked ꢀ 100%.
b
Fig. 5. Typical HPLC chromatograms of crude extract of R. crenulata (A) and mixed reference compounds (compounds 1–6) (B). 2-(4-hydroxyphenyl)-ethyl-O-b-
glucopyranosyl-6-O-b- -glucopyranoside (1), icariside D2 (2), p-hydroxyphenacyl-b- -glucopyranoside (3), picein (4), salidroside (5), tyrosol (6). Experimental conditions:
m); mobile phase, methanol–water (15:85, v/v); flow rate, 1.0 ml/min; detection wavelength, 275 nm; column
D-
D
D
column, Phenomenex C18 column (250 ꢀ 4.6 mm, 5
temperature, 25 °C.
l
physical data are in agreement with previous literature data: icar-
3.4. Validation of method
iside D2 (Peng, Luo, Lu, Chen, Xie, Chen, 2009), p-hydroxyphenacyl-
b-
D
-glucopyranoside (Peng et al., 2009), picein (Peng et al., 2009),
The linearities, regressions and linear ranges of six target
analytes were performed using the developed HPLC method
(Table 3). The correlation coefficient values (r2 > 0.9996) indicated
salidroside (Lalonde, Wong, & Tsai, 1976), tyrosol (Lalonde et al.,
1976), respectively.

2132
D. Chen et al. / Food Chemistry 134 (2012) 2126–2133
Table 5
The contents of isolated compounds 1–6 from crude extract of R. crenulata.
Batch no.a
Contents (mg/g) (RSD%, n = 3)
Compound 1
Compound 2
Compound 3
Compound 4
Compound 5
Compound 6
1
2
3
1.46 (2.2)
1.51 (1.9)
1.39 (2.8)
2.69 (2.1)
2.78 (1.1)
2.89 (1.6)
0.76 (2.6)
0.82 (2.0)
0.79 (2.1)
0.57 (2.9)
0.64 (2.5)
0.61 (2.8)
9.08 (1.6)
9.12 (1.3)
9.57 (1.1)
1.92 (1.3)
1.85 (1.8)
1.78 (2.4)
a
Three batches of crude extract of R. crenulata, which was collected from the same region (Tibetan, China).
(Lee et al., 2000; Ohsugi et al., 1999), which have been reported
to have better antioxidant activity. Among the six isolated com-
Table 6
Antioxidant activity of isolated compounds 1–6 by DPPH assay radical scavenging
assay.
pounds, compound 6 (IC50 8.10
tive activity. Compound 5 (IC50 10.97
(IC50 13.45 M) showed moderate antioxidative activity. These
lM) showed the highest antioxida-
l
M), 1 (IC50 12.13 M) and 3
l
Compounds
IC50 (lM)
7.49 0.52a
12.13 0.54
l
Crude extract
compounds might be contributed to the strong antioxidative activ-
ity of R. crenulata, which can be explained by the presence of
hydroxybenzene.
2-(4-hydroxyphenyl)-ethyl-O-b-
glucopyranoside (1)
Icariside D2 (2)
p-hydroxyphenacyl-b-
Picein (4)
D-glucopyranosyl-6-O-b-D-
>40
13.45 0.79
>40
D
-glucopyranoside (3)
Salidroside (5)
Tyrosol (6)
10.97 0.89
8.10 0.98
4.12 0.54
4. Conclusion
Rutinb
We have demonstrated that HSCCC combined with Semi-prep
HPLC is an efficient and superior separation strategy for water-
soluble phenylpropanoids over other chromatographic methods
for complementary action between the two methods. In this work,
initial preparation by HSCCC enriched some minor water-soluble
phenylpropanoids from R. crenulata, followed by further purifica-
tion using high resolution Semi-prep HPLC. Six high purity phenyl-
Each value is the mean of three independent experiments standard deviation (SD).
a
The unit of IC50 of crude extract is
Used as positive control.
lg/ml.
b
good correlations between the tested compound concentrations
and their peak areas within the test ranges. The LOD and LOQ were
less than 0.10 and 0.80 lg/ml, and the overall intra- and inter-day
variations (RSD) of the six analytes were less than 1.6% and 2.9%,
respectively. The recoveries were between 98.2% and 101.6%
(Table 4), and the repeatabilities were presented as RSD (n = 6)
were between 1.2% and 3.0%. These results showed that the devel-
oped method had good accuracy and repeatability.
propanoid compounds, namely 2-(4-hydroxyphenyl)-ethyl-O-b-
glucopyranosyl-6-O-b- -glucopyranoside (1), icariside D2 (2), p-
hydroxyphenacyl-b- -glucopyranoside (3), picein (4), salidroside
(5) and tyrosol (6) were obtained from the crude extract of R. cre-
nulata. 2-(4-Hydroxyphenyl)-ethyl-O-b- -glucopyranosyl-6-O-b-
glucopyranoside (1) is a new compound, and p-hydroxyphenacyl-
b- -glucopyranoside (3) and picein (4) are isolated from the plant
D-
D
D
D
D-
D
of R. crenulata for the first time. Meanwhile, a simple and rapid
HPLC method for the simultaneous determination of the six phe-
nylpropanoid ingredients from the crude extract of R. crenulata
has also been developed for the first time. Furthermore, the antiox-
idant capacities of the six phenylpropanoid compounds were eval-
uated by using an established biochemical assay in vitro, the DPPH
free radicals scavenging assay. It was found that 2-(4-hydroxy-
3.5. Analysis of the crude extract
The developed method was applied to simultaneous analysis of
compounds 1–6 from the crude extract of R. crenulata. The typical
HPLC chromatogram of crude extract was shown in Fig. 5, and the
data was listed in Table 5. Among these investigated compounds,
the content of compound 5 (salidroside) in the three batches of
crude extract was the highest and the content of compound 4
(picein) in the three batches of crude extract was the lowest. The
results indicated that the proposed HPLC method could be used
to monitor the quality of the crude extract of R. crenulata.
phenyl)-ethyl-O-b-
D
-glucopyranosyl-6-O-b-D-glucopyranoside (1),
p-hydroxyphenacyl-b-
D-glucopyranoside (3), salidroside (5) and
tyrosol (6) possess significant inhibitory effects on the DPPH free
radicals. Hence, these results suggested that R. crenulata could be
a promising source of natural antioxidants.
3.6. Radical scavenging activity of isolated compounds
Acknowledgements
The effect of antioxidants on DPPH radical scavenging is gener-
ally based on their hydrogen-donating ability. The antioxidant
activities of the target isolated compounds and crude extract from
R. crenulata were determined spectrophotometrically by the DPPH
radical scavenging activity assay and expressed in terms of the
This work is supported by the Specialized Research Fund for the
Doctoral Program of Higher Education of China (No.
20113234110001) and the National Natural Science Foundation
of China (No. 21075066). The authors gratefully appreciate the
assistance of Professor Dongjun Chen for his excellent technical
support.
antioxidant concentration (lM) required to reduce the initial DPPH
concentration. Rutin (Zhang, Shi, Wang, & Huang, 2011) was used
as reference antioxidant. As shown in Table 6, the crude extract
References
(IC50 7.49
parison with the reference antioxidant rutin (IC50 4.12
is equal to 2.52 g/ml). Also, the antioxidant activity of the crude
lg/ml) showed potent antioxidative activities in com-
l
M, which
Brand-Williams, W., Cuvelier, M. E., & Berset, C. (1995). Use of a free radical method
to evaluate antioxidant activity. Lebensmittel Wissenschaft and Technologie, 28,
25–30.
Dement’eva, L. A., & Iaremenko, K. V. (1987). Effect of a Rhodiola extract on the
tumor process in an experiment. Voprosy Onkologii, 33(7), 57–60.
Domínguez, H., Torres, J., & Núñez, M. J. (2001). Antioxidant phenolics as food
additives from agricultural wastes. Polyphenols Actualites, 21, 25–30.
l
extract was higher than that of the six isolated compounds. One
main reason maybe is a synergistic effect of the phenolics. The
other reason is likely that the most active component is not one
of the isolated compounds, such as organic acids and flavonoids

D. Chen et al. / Food Chemistry 134 (2012) 2126–2133
2133
Du, Q., Li, Z., & Ito, Y. (2011). Preparative separation of isoflavone components in
soybeans using high-speed counter-current chromatography. Journal of
Chromatography A, 923, 271–274.
Du, M., & Xie, J. M. (1994). Studies on the chemical constituents of Rhodiola
crenulata. Acta Chimica Sinica, 52(9), 927–931.
Ohsugi, M., Fan, W., Hase, K., Xiong, Q., Tezuka, Y., Komatsu, K., et al. (1999). Active-
oxygen scavenging activity of traditional nourishing-tonic herbal medicines and
active constituents of Rhodiola sacra. Journal of Ethnopharmacology, 67(1),
111–119.
Peng, Y., Luo, J., Lu, Q., Chen, X. M., Xie, Y., Chen, L. N., et al. (2009). HPLC analysis,
semi-preparative HPLC preparation and identification of three impurities in
salidroside bulk drug. Journal of Pharmaceutical and Biomedical Analysis, 49,
828–832.
Peng, J. N., Ma, C. Y., & Ge, Y. C. (1995). Studies on the chemical constituents of
Rhodiola crenulata. Chinese Traditional and Herbal Drugs, 26(4), 177–179.
Shen, C., Jiang, L. Q., Gu, X. Y., Qiu, S. S., Shen, X. F., & Yang, Q. Q. (2010). Antioxidant
compounds and antioxidant activity of Rhodiola drink. Academic Periodical of
Farm Products Processing, 12, 32–34.
Shi, S. Y., Zhao, Y., Zhou, H. H., Zhang, Y. P., Jiang, X. Y., & Huang, K. L. (2008).
Identification of antioxidants from Taraxacum mongolicum by high-performance
liquid chromatography-diode array detection-radical-scavenging detection-
electrospray ionization mass spectrometry and nuclear magnetic resonance
experiments. Journal of Chromatography A, 1209, 145–152.
Fan, W., Tezuka, Y., Komatsu, K., Namba, T.,
& Kadota, S. (1999). Prolyl
endopeptidase inhibitors from the underground part of Rhodiola sacra S. H. Fu.
Biological Pharmaceutical Bulletin, 22(2), 157–161.
_
´
Furmanowa, M., Skopinska-Rózewska, E., Rogala, E., & Hartwich, M. (1998). Rhodiola
rosea in vitro culture-phytochemical analysis and antioxidant action. Acta
Societatis Botanicorum Poloniae, 67(1), 69–73.
Han, X., Zhang, T. Y., Wei, Y., Cao, X. L., & Ito, Y. (2002). Separation of salidroside
from Rhodiola crenulata by high-speed counter-current chromatography. Journal
of Chromatography A, 971, 237–241.
Hostettmann, K., Wolfender, J. L.,
& Terreaux, C. (2001). Modern screening
techniques for plant extracts. Pharmaceutical Biology, 39, 18–32.
Ito, Y. (2005). Golden rules and pitfalls in selecting optimum conditions for high-
speed counter-current chromatography. Journal of Chromatography A, 1065,
145–168.
Kwon, Y. I., Jang, H. D., & Shetty, K. (2006). Evaluation of Rhodiola crenulata and
Rhodiola rosea for management of type II diabetes and hypertension. Asia Pacific
Journal of Clinical Nutrition, 15(3), 425–432.
Lalonde, R. T., Wong, C., & Tsai, A. I. M. (1976). Polyglucosidic metabolites of
oleaceae. The chain sequence of oleoside aglucon, tyrosol, and glucose units in
three metabolites from Fraxinus Americana. Journal of the American Society,
98(10), 3007–3013.
Tezuka, Y., Fan, W., Kasimu, R., & Kadota, S. (1999). Screening of crude drug extracts
for prolyl endopeptidase inhibitory activity. Phytomedicine, 6(3), 197–203.
Udintsev, S. N.,
& Schakhov, V. P. (1991). Decrease of cyclophosphamide
haematotoxicity by Rhodiola rosea root extract in mice with Ehrlich and Lewis
transplantable tumors. European Journal of Cancer, 27(9), 1182.
Wu, S. X., Guo, Y. D., Guo, S. Y., Li, L., Wang, B. X., & Ma, T. (2008). Study of the
chemical constituents of ethanol extracts of Rhodiola crenulata H. Modern Food
Science and Technology, 24(4), 322–326.
Lee, M. W., Lee, Y. A., Park, H. M., Toh, S. H., Lee, E. J., Jang, H. D., et al. (2000).
Antioxidative phenolic compounds from the roots of Rhodiola sachalinensis A.
Bor. Archives of Pharmacal Research, 23, 455–458.
Yang, Y. N., Liu, Z. Z., Feng, Z. M., Jiang, J. S., & Zhang, P. C. (2012). Lignans from the
root of Rhodiola crenulata. Journal of Agricultural and Food Chemistry, 60(4),
964–972.
Machida, K., Ohkawa, N., Ohsawa, A., & Kikuchi, M. (2009). Two new phenolic
glycosides from Syringa reticulata. Journal of Natural Medicines, 63, 192–194.
Mishra, K. P., Padwad, Y. S., Jain, M., Karan, D., Ganju, L., & Sawhney, R. C. (2006).
Aqueous extract of Rhodiola imbricata rhizome stimulates pro-inflammatory
mediators via phosphorylated IkappaB and transcription factor nuclear factor-
kappaB. Immunopharmacology and Immunotoxicology, 28(2), 201–212.
Nakamura, S., Li, X. Z., Matsuda, H., & Yoshikawa, M. (2008). Bioactive constituents
from Chinese natural medicines. XXVIII. Chemical structures of acyclic alcohol
glycosides from the roots of Rhodiola crenulata. Chemical and Pharmaceutical
Bulletin, 56(4), 536–540.
You, D. Q., Zou, L. F., Gao, F., Ding, X. G., Yang, P., Han, L., et al. (2003). Effects of
Rhodiola crenulata compound on mental health condition of seamen during
voyage. Academic Journal of Second Military Medical University, 24(11),
1187–1189.
Zhang, Y. P., Shi, S. Y., Wang, Y. X., & Huang, K. L. (2011). Target-guided isolation and
purification of antioxidants from Selaginella sinensis by offline coupling of
DPPH-HPLC and HSCCC experiments. Journal of Chromatography B, 879,
191–196.
Zhao, Y. H., Liu, C. H., & Wu, T. H. (1998). Advances in studies and applications on
Rhodiola Rosea L. Quarterly of Forest By-product and Speciality in China, 3, 44–45.