Content uploaded by Yuguang Wang
Author content
All content in this area was uploaded by Yuguang Wang on Mar 31, 2015
Content may be subject to copyright.
Available via license: CC BY-NC-ND 3.0
Content may be subject to copyright.
ORIGINAL ARTICLE
Chemical comparison of dried rehmannia root and prepared
rehmannia root by UPLC-TOF MS and HPLC-ELSD with
multivariate statistical analysis
Qiande Liang
a,y
, Jing Ma
b,y
, Zengchun Ma
a
, Yuguang Wang
a
, Hongling Tan
a
,
Chengrong Xiao
a
, Ming Liu
a
, Beibei Lu
a
, Boli Zhang
b
, Yue Gao
a,
n
a
Institute of Radiation Medicine, Academy of Military Medical Science, Beijing 100850, China
b
Tianjin University of Traditional Chinese Medicine, Tianjin 300193, China
Received 18 August 2012; revised 28 September 2012; accepted 22 October 2012
KEY WORDS
Rehmannia root;
Liquid chromatography;
Mass spectrometry;
Evaporated light
scattering detection;
Multivariate statistical
analysis
Abstract To identify the chemical differences which lead to the different therapeutic effects of
dried rehmannia root (DRR) and prepared rehmannia root (PRR), we compared the chemical
composition of decoctions of randomly purchased DRR and PRR using ultra performance liquid
chromatography (UPLC) coupled with time-of-flight mass spectrometry and high performance
liquid chromatography (HPLC) coupled with evaporative light scattering detection (ELSD) with
the aid of multivariate statistical analysis. Both approaches clearly revealed compositional and
quantitative differences between DRR and PRR. UPLC-MS data indicated stachyose, rehmaiono-
side A (or rehmaionoside B), acteoside (or forsythiaside, or isoacteoside), 6-O-coumaroylajugol
(or 6-O-E-feruloylajugol, or 6-O-Z-feruloylajugol) as important discriminators between DRR and
PRR decoctions. HPLC-ELSD analysis showed that the content of fructose in the decoctions of
PRR was about four times greater than that of DRR (Po10
5
), while sucrose content in the
Institute of Materia Medica, Chinese Academy of Medical Sciences
Chinese Pharmaceutical Association
www.elsevier.com/locate/apsb
www.sciencedirect.com
Acta Pharmaceutica Sinica B
2211-3835 &2012 Institute of Materia Medica, Chinese Academy of Medical Sciences and Chinese Pharmaceutical Association. Production and
hosting by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apsb.2012.11.001
n
Corresponding author.
E-mail address: gaoyue@bmi.ac.cn (Yue Gao).
y
These authors contributed equally to this work.
Peer review under responsibility of Institute of Materia Medica, Chinese Academy of Medical Sciences and Chinese Pharmaceutical Association.
Acta Pharmaceutica Sinica B 2013;3(1):55–64
decoctions of PRR was only about one seventh of that in DRR (Po0.01). Our results suggest that
some compounds, such as fructose, stachyose and rehmaionoside, may be responsible for the
differing therapeutic effects of DRR and PRR. Furthermore, improvements in quality control for
PRR, which is currently lacking in the Chinese Pharmacopoeia, are recommended.
&2012 Institute of Materia Medica, Chinese Academy of Medical Sciences and Chinese Pharmaceutical
Association. Production and hosting by Elsevier B.V. All rights reserved.
1. Introduction
Dried rehmannia root (DRR) and prepared rehmannia root
(PRR) are herbal drugs that are frequently used in traditional
Chinese medicine (TCM). Although they are both prepared
from rehmannia root (Rehmannia glutinosa Libosch) and PRR
is prepared by further braising or steaming DRR
1
, they have
different therapeutic effects and are used for different ther-
apeutic purposes. According to the theory and practice of
TCM and the ancient Chinese philosophy, ‘‘Yin’’ is a concept
which is a contrary to ‘‘Yang’’ and ‘‘Yin’’ refers to the inner,
lower, weak, passive, tranquil or material aspects of things
(including the human body) while ‘‘Yang’’ refers to the outer,
upper, strong, positive, active or functional aspects of things.
Based on this concept, in addition to their common effect of
‘‘nourishing the body’s Yin’’, DRR possesses a ‘‘cold’’
character and can ‘‘reduce the body’s heat’’, while PRR
possesses a ‘‘warm’’ character as a simple tonic and it does
not ‘‘reduce the body’s heat’’ but can ‘‘replenish the blood and
reinforce the body’s essence and marrow’’
1,2
. These ‘‘pharma-
cological’’ differences have been documented in the TCM
literature for at least three hundred years
3
. In clinical practice,
DRR is used mainly for diseases showing ‘‘hot’’ symptoms,
such as feeling hot, deep-red tongue, fidgets, thirst and
exanthemata, while PRR is used mainly for diseases showing
‘‘blood-lacking’’ or ‘‘Yin-lacking’’ symptoms, such as feeling
faint, weak legs, yellow complexion, palpitations and tinnitus.
The reason for the difference in therapeutic effects between
DRR and PRR is currently not clear but surely lies in chemical
differences. A few studies concerning the chemical differences have
beencarriedout.In2004,Wenetal.
4
investigated the changes in
saccharide content and composition in rehmannia root prepara-
tions using high performance liquid chromatography (HPLC)
with differential refractive index detection during processing. They
found that peaks of glucose and fructose became outstanding
when DRR was steamed to PRR and deduced that defructosyla-
tion of stachyose, a major component of DRR, may take place
during steaming. The increase of fructose and decrease of
stachyose was also reported by Bian et al.
5
in 1995 using thin
layer chromatogram scan analysis. However, due to low resolution
or limited sensitivity of the above-mentioned methods, differences
in sugar content between DRR and PRR warrants verification by
more advanced methods. In 2010, Li et al.
2
explored potential
chemical markers using ultra high performance liquid chromato-
graphy coupled with time-of-flight mass spectrometry (TOF MS)
and multivariate statistical analysis (MVA) for the discrimination
of DRR from PRR. Leonuride or its isomer and 5-(a-D-
glucopyranosyl-(1-6)-a-D-glucopyranosyloxymethyl)-2-furancar-
boxaldehyde were found as the most characteristic markers of
DRR and PRR. But their study did not report any differences in
sugar content or composition. Perhaps this was partly due to the
poor sensitivity of the electrospray ionization (ESI) mass spectro-
metric method for neutral carbohydrates
6,7
and the weak retention
of many neutral carbohydrates on the HSS T3
TM
column used in
this study. Another reason may be that their sample extraction
method was ultrasonic extraction with methanol–water under
room temperature, rather than hot water extraction as carried out
by Wen et al.
4
. However, we believe that hot water extraction is
more reasonable than room temperature methanol–water extrac-
tion, because water decoction is the routinely used procedure for
both DRR and PRR, and chemical differences in decoctions
prepared in this manner may correlate more closely with differ-
ences in therapeutic effects. Hence, in order to identify chemical
differences between DRR and PRR which may be responsible for
the differences in therapeutic effects, a more comprehensive
investigation is required between the decoctions of DRR and
PRR that includes both saccharide and non-saccharide
constituents.
Liquid chromatography-mass spectrometry (LC-MS) with
MVA is a feasible strategy for chemical analysis of TCM drug
samples which are complex mixtures
2
. MVA is helpful in analysis
of the huge data matrix of multiple variables generated by LC-
MS. It finds out whether there is difference between groups of
samples utilizing principal component analysis (PCA) and which
signals or variables contribute the grouping by orthogonal partial
least-squares to latent structures discriminate analysis (OPLS-
DA). MS (especially high-resolution MS) provides further useful
qualitative information on the chemical substances producing the
signals present in this analysis.
In this study, with the aid of MVA, we employed both ultra
performance liquid chromatography (UPLC) coupled with
TOF-MS and HPLC coupled with evaporative light scattering
detection (ELSD), which is a suitable method for sugar
analysis, to compare the chemical compositions of decoctions
of DRR and PRR. Our goal was to acquire a relatively
comprehensive analysis of their chemical differences, which
may explain their differing therapeutic effects.
2. Materials and methods
2.1. Drugs and reagents
PRR and DRR samples (six each) were purchased from six
different drug stores in China (Table 1). D-Fructose was
purchased from Sinopharm Chemical Reagent Co., Ltd.
(Shanghai, China). D-Glucose and sucrose were purchased
from Beijing Chemical Reagent Company (Beijing, China).
Acetonitrile was of HPLC grade (Fisher Scientific, Fair Lawn,
New Jersey, USA). Deionized water was prepared with a
Millipore water purification system.
Qiande Liang et al.56
2.2. Analytical apparatus
A Waters Acquity UPLC system consisting of a binary solvent
manager, a sample manager and an ELSD detector, a Waters
SYNAPT HDMS quadrupole time-of-flight mass spectro-
meter and an analytical workstation with Waters MassLynx
4.1 software were used.
2.3. Sample preparation
20 g of each sample was weighed, soaked into 160 mL water
for 60 min at room temperature and boiled for 30 min.
The suspension was filtered and 100 mL water was added
for the second decoction with a duration of 20 min. The
filtered and mixed suspension from the two decoctions was
adjusted to a volume of 200 mL. 20 mL of the decoction was
precipitated by adding 80 mL alcohol and filtered. The filtrates
were evaporated to less than 1 mL at 50 1Cin vacuo. The
evaporated residue was dissolved with 50% acetonitrile into a
volumetric flask. The final volume of the sample solution was
set to 25 mL. Three standard compounds, D-fructose, D-
glucose and sucrose, were weighed and dissolved in 50%
acetonitrile to give serial concentrations within the range of
0.80–10.08 mg/mL, 0.80–20.08 mg/mL and 0.81–20.24 mg/
mL, respectively. The solutions of standards and samples
were filtered through a 0.45 mm (HPLC)/0.22 mm (UPLC)
membrane before injection into the UPLC or HPLC system.
2.4. UPLC-TOF MS analysis
Separations were carried out with a Waters Acquity HSS T3
column (100 mm 2.1 mm I.D., 1.8 mm). The mobile phase
consisted of (A) water containing 0.1% formic acid and (B)
acetonitrile containing 0.1% formic acid. The eluting conditions
were: isocratic 2% B (0–1 min), linear gradient from 2% to 5% B
(1–2 min), 5% to 12% B (2–5 min), 12% to 20% B (5–10 min),
20% to 30% B (10–12 min), 30% to 50% B
(12–13 min), 50% to 100% B (13–15 min), isocratic 100% B for
1 min, then back to 2% B in 1 min and isocratic 2% B for 3 min
before next run. The flow rate was 0.5 mL/min.
The column was maintained at 24 1C. The injection volume was
2mL. Mass spectrometry was performed with ESI source operat-
ing in negative ion mode. The lock mass compound was leucine
enkephalin (m/z554.2615). The lock spray reference scan
frequency was 20 s with reference cone voltage of 30 V. The
source temperature was set at 100 1C and the desolvation
temperature was set at 450 1C with desolvation gas flow of
900 L/h. The capillary voltage and cone voltage were set to
3 kV and 40 V, respectively. The collision energies were set as
6 V (trap)/4 V (transfer) with 2.00 mL/min trap gas flow. The LC-
MS data acquisition was controlled by MassLynx 4.1 software.
2.5. HPLC-ELSD analysis
Separations were carried out with a SHISEIDO carbohydrate
column (150 mm 2.0 mm I.D., 5 mm). The mobile phase was
an isocratic mixture of acetonitrile and water (80:20, v/v).
The column temperature was kept constant at 24 1C and the
flow rate was 0.6 mL/min. The drift tube temperature, gas
pressure and gain were 40 1C, 40 psi and 25, respectively. The
sample injection volume was 2 mL. The compounds were
identified by comparing their retention time values with those
of the standards (D-fructose, D-glucose and sucrose). The
HPLC-ELSD data acquisition was controlled by MassLynx
4.1 software.
2.6. Principal component analysis (PCA)
The PCA was carried out with the data of all samples. The
UPLC-MS data were processed by MarkerLynx XS and
EZinfo, sub modules of the MassLynx 4.1 software, to
accomplish signal deconvolution and MVA. The intensity
threshold was set to 100 counts. The HPLC-ELSD chromato-
grams were integrated by MassLynx software to give peak
area values of D-fructose, D-glucose and sucrose. PCA were
performed with UPLC-MS data (peak heights) and HPLC-
ELSD data (peak areas of D-fructose, D-glucose and sucrose)
separately to acquire an overview of trends or patterns.
The data scaling method was Pareto.
2.7. Orthogonal partial least-squares to latent structures
discriminate analysis (OPLS-DA)
Using the EZinfo module, OPLS-DA was carried out with
UPLC-MS data of 11 samples, DRR1–6 and PRR2–5, to
discover relatively important chemical markers. The PRR1
was excluded as an abnormal sample according to PCA
results. The data scaling method was Pareto.
Table 1 DRR and PRR samples analyzed in the present study.
Sample Location of purchase Time of purchase
DRR1 Kang Shou Tang drug store, Beijing 2009-10-8
DRR2 Xi Dan Jin Xiang drug store, Beijing 2009-10-8
DRR3 Gong Zhu Fen Tong Ren Tang drug store, Beijing 2009-10-8
DRR4 Kang Yu drug store, Beijing 2009-10-8
DRR5 Qian Men Tong Ren Tang drug store, Beijing 2009-10-8
DRR6 Bao Ding Tong Ren Tang drug store, Baoding City, Hebei Province 2009-10-2
PRR1 Kang Shou Tang drug store, Beijing 2009-10-8
PRR2 Xi Dan Jin Xiang drug store, Beijing 2009-10-8
PRR3 Gong Zhu Fen Tong Ren Tang drug store, Beijing 2009-10-8
PRR4 Kang Yu drug store, Beijing 2009-10-8
PRR5 Qian Men Tong Ren Tang drug store, Beijing 2009-10-8
PRR6 Bao Ding Tong Ren Tang drug store, Baoding City, Hebei Province 2009-10-2
Chemical comparison of dried rehmannia root and prepared rehmannia root by UPLC-TOF MS and HPLC-ELSD 57
2.8. Establishment of in-house library and preliminary
identification of chemical markers
By searching from Zhou et al.’s handbook
8
, PubMed and the
Chinese National Knowledge Infrastructure (CNKI) of Tsin-
ghua University, all compounds reported in the literature on
rehmannia root were summarized in a Microsoft Excel table to
establish an in-house library, which includes the name, mole-
cular formula and chemical structure of each published known
compound. The exact m/zvalues of putative deprotonated
quasi-molecular ions [M–H]
, the formic acid group adductive
ions [MþHCOO]
and the deprotonated dimeric ions [2M–
H]
were calculated to compare with the observed exact m/z
values of those relatively important chemical makers discov-
ered by OPLS-DA. The possible identities of the markers were
deduced at mass accuracy of less than 3 ppm.
2.9. Determination of D-fructose, D-glucose and sucrose
Log/log calibration curves of D-fructose, D-glucose and
sucrose were obtained from the log values of the peak areas
over the log values of the concentrations of the standards.
Concentrations of D-fructose, D-glucose and sucrose in the
decoctions of the samples were calculated from this regression
analysis, and expressed as mean7standard deviation. The
significance of difference between two groups was analyzed
statistically by Student’s t-test.
3. Results and discussion
3.1. Data overview with PCA score plots
Figs. 1 and 2 show typical UPLC-MS and HPLC-ELSD
chromatograms. Figs. 3 and 4 show PCA score plots of
UPLC-MS data and HPLC-ELSD sugar data, respectively.
Distinct pattern of discrimination by the first principal
component scores (t[1]) between DRR and PRR samples can
be observed in both figures, i.e., the DRR and PRR samples
clustered towards opposite ends of the t[1] axis. The presence
of sample PRR1 markedly near the DRR cluster but far from
the PRR cluster in both figures suggests that PRR1 is likely to
be a substandard product or a product incorrectly handled in
some step during the course from factory to market. There-
fore, we have excluded sample PRR1 from further analysis.
3.2. Marker discovery by OPLS-DA and preliminary
assignment of identities
Fig. 5 shows the S-plots produced by the OPLS-DA of UPLC-
MS data with PRR1 excluded. Each green dot represents a
variable in the OPLS-DA model, i.e., a LC-MS peak with a
certain m/zvalue and a certain retention time (t
R
). The S-Plot
visualizes the variable influence in the model where the X-axis
represents contribution (covariance) and the Y-axis (spans
between 1 and 1) represents correlation (reliability)
9
.Potential
Figure 1 Typical UPLC-MS BPI chromatograms of (a) DRR and (b) PRR. BPI chromatogram: base peak intensity chromatogram.
Qiande Liang et al.58
markers were indicated by those dots marked with red square,
which were automatically selected by the ‘‘select default’’ func-
tion of the software. These dots are the variables that differ-
entiate the most between the two groups and are therefore
relatively important. The t
R
and m/zvalue of these selected
variables and their identities tentatively assigned are listed in
Table 2. The identities of markers were assigned by searching
and matching the reported constituents of rehmannia root in the
in-house library at the mass accuracy of less than 3 ppm using
measured and theoretical exact m/zvalues. Among these selected
markers, stachyose (C
24
H
42
O
21
,MW¼666, peak 9)
10
,rehmaio-
noside A (or rehmaionoside B, C
19
H
34
O
8
,MW¼390, peak 11)
11
,
acteoside (or forsythiaside or isoacteoside, C
29
H
36
O
15
,
MW¼624, peak 12)
8,12,13
,6-O-coumaroylajugol (or 6-O-E-
feruloylajugol or 6-O-Z-feruloylajugol, C
25
H
32
O
12
,MW¼524,
peak 13)
8
were tentatively identified. The measured exact m/z
value of peak 2 and peak 6 approximate the theoretical exact m/z
value of [M–H]
of D-mannitol (C
6
H
14
O
6
,MW¼182)
14
and
double trisaccharide complex (C
36
H
64
O
32
,MW¼1008)
14
,but
with mass accuracy of 9.94 and 7.64 ppm, respectively. Fig. 6
shows the selected ion intensity trend plots of four tentatively
identified markers (peaks 9, 11, 12 and 13).
3.3. Contents of D-fructose, D-glucose and sucrose in
decoctions of DRR and PRR
Log/log calibration curves showed good linear relations
between the log values of the peak areas by HPLC-ELSD
and the log values of the concentrations of the standard
solutions. The regression equations were y¼0.6835x2.5027
(r
2
¼0.9974) with linear range 0.80–10.08 mg/mL for D-fruc-
tose, y¼0.6228x2.0732 (r
2
¼0.9993) with linear range 0.80–
20.08 mg/mL for D-glucose, and y¼0.6443x2.2590
(r
2
¼0.9993) with linear range 0.81–20.24 mg/mL for sucrose,
where xrepresents log value of peak area and yrepresents log
value of concentration. Concentrations of D-fructose, D-glu-
cose and sucrose in DRR and PRR with PRR1 excluded are
summarized in Fig. 7. The fructose content in the decoctions
of PRR (8.75 mg/mL) was about 4 times that in DRR
(2.29 mg/mL) (Po10
5
), while the sucrose content in the
decoctions of PRR (0.52 mg/mL) was only about one seventh
of that in DRR (3.55 mg/mL) (Po0.01). The glucose content
in the decoctions of PRR (9.08 mg/mL) increased approxi-
mately 57% as compared with DRR (5.80 mg/mL) (Po0.05).
These results show that the increase in fructose and decrease in
sucrose when DRR is converted to PRR is significant and
remarkable. The change in the content of these sugars does
not contradict the findings of Wen et al.
4
and Bian et al.
5
and
is much clearer now from our results. The preparation process
increases the amount of fructose in the PRR decoction while
eliminating most of the sucrose in the DRR decoction,
indicating that the saccharide changes caused by preparation
are substantial and distinctive.
3.4. Physiological and pharmaceutical indications
It is interesting that our findings on chemical markers in this
study differ largely from study of Li et al.
2
reported in 2010
which also used negative ion mode UPLC-TOF MS with MVA.
In their study, six markers were selected from the OPLS-DA S-
Plot, including a(t
R
2.34 min, m/z449.1283, matching 5-(a-D-
glucopyranosyl-(1-6)-a-D-glucopyranosyloxymethyl)-2-furancar-
boxaldehyde), b(t
R
4.11 min, m/z236.0553, unidentified), c(t
R
2.41 min, m/z465.1236, unidentified), d(t
R
3.66 min, m/z
393.1360, matching leonuride or its isomer), e(t
R
3.64 min, m/
z393.1361, matching leonuride or its isomer) and f(t
R
6.51 min,
m/z345.1549, matching rehmapicroside). The major difference
between our research methods is that they used ultrasonic
extraction with methanol–water under room temperature, while
we used hot water extraction which we deemed to be relatively
closer to clinical practice than their method.
Our study shows distinct changes in saccharide content in the
decoctions after DRR is converted to PRR, i.e., the contents of
monosaccharide, especially fructose, increased while the contents
of some oligosaccharides decreased. The content of fructose and
glucose increased by 282% and 57%, respectively, while the
content of sucrose decreased by 85%. Peak 9 (C
24
H
42
O
21
,
Figure 2 Typical HPLC-ELSD chromatograms of (a) standard
mixture of D-fructose, D-glucose and sucrose, (b) DRR and
(c) PRR.
Chemical comparison of dried rehmannia root and prepared rehmannia root by UPLC-TOF MS and HPLC-ELSD 59
MW¼666) matches stachyose, whose signal intensity dropped
from tens (DRR) to zero (PRR). Stachyose is a tetrasaccharide
reported as one of the most abundant components in DRR with
the content of 30–60%
4,14,15
. The peak matches verbascose (peak
14, C
30
H
52
O
26
,MW¼828, mass accuracy ¼1.69 ppm, t
R
¼0.59
min), a pentasaccharide in DRR
14
, and also shows a descending
trend (Fig. 8). Although it was not automatically selected, it has a
correlation of 0.9056 and is very near those selected peaks in the
S-Plot (Fig. 5, dot marked with black square). The increase in
fructose and glucose as well as the decrease in stachyose after
DRR conversion to PRR has been reported in literature
4,5
,andis
confirmed by our results. However the decrease of sucrose and
verbascose has not been reported.
Our previous study on Si-Wu-Tang (SWT) showed that
fructose is a hematopoiesis-promoting compound
16
.Fructoseis
a major constituent in SWT, a traditional Chinese formula
consisting of PRR, angelica root, chuanxiong rhizoma and
paeonia root, which possesses hematopoiesis-promoting activity
and is used mainly for ‘‘blood-enrichment’’. Similarly to SWT
administration, the oral administration of fructose increased the
number of peripheral leukocytes and four types of progenitor
cells in bone marrow of irradiated mice (anemia model),
including colony-forming unit-granulocyte-macrophages
(CFU-GM), colony-forming unit-mature erythroid cells (CFU-
E), colony-forming unit-immature erythroid cells (BFU-E), and
colony-forming unit-multipotential cells (CFU-mix)
16
. Our study
further showed that the free fructose and glucose in SWT
decoction is almost completely provided by PRR
17
. According
to the conventional theory of TCM which uses ‘‘Jun’’
(‘‘emperor’’, the most important ingredient or the ingredient
plays a central role in a formula), ‘‘Chen’’ (‘‘minister’’), ‘‘Zuo’’
(‘‘assistor’’), and ‘‘Shi’’ (‘‘emissary’’) to identify the importance
and roles of different ingredient drugs in a formula, PRR is the
‘‘Jun’’ in the SWT formula
18
. Therefore, an interesting coin-
cidence appeared that the ‘‘Jun’’ drug (PRR) in a hematopoiesis-
promoting formula (SWT) provides most of the hematopoiesis-
promoting constituent (fructose). The present study showed a
tremendous increase of fructose content in the decoction after
DRR is converted to PRR. According to the principles and
practice of TCM, PRR is basically used for ‘‘blood-enrichment’’
Figure 3 PCA score plot of UPLC-MS data.
Figure 4 PCA score plot of HPLC-ELSD sugar data.
Figure 5 OPLS-DA S-plot of UPLC-MS data excluded PRR1. Peak 9 matches stachyose. Peak 11 matches rehmaionoside A or B. Peak
12 matches acteoside or forsythiaside or isoacteoside. Peak 13 matches 6-O-Coumaroylajugol or 6-O-E-Feruloylajugol or 6-O-Z-
Feruloylajugol. Peak 14 matches verbascose. Identities of other peaks were not assigned.
Qiande Liang et al.60
not ‘‘heat-reduction’’, while DRR is used for ‘‘heat-reduction’’
not ‘‘blood-enrichment’’. Therefore, another interesting coinci-
dence appeared that a hematopoiesis-promoting constituent
(fructose) increases tremendously when a ‘‘non-blood-enrich-
ment’’ drug (DRR) is prepared to become a ‘‘blood-enrichment’’
drug (PRR). These two coincidences support the speculation that
fructose is an important hematopoiesis-promoting constituent in
SWT as well as in PRR, and at the same time, provide a possible
explanation for why PRR is a ‘‘blood-enriching’’ drug while
DRR is not. The possible mechanism for the hematopoiesis-
Figure 6 Selected ion intensity trend plots of tentatively identified markers automatically selected from OPLS-DA S-plots. (a) Peak 9, m/z
665.2135; (b) Peak 11, m/z435.2227; (c) Peak 12, m/z623.1983; (d) Peak 13, m/z523.1813.
Table 2 Automatically selected variables indicating potential makers in OPLS-DA S-plots and their preliminarily assigned
identities.
Peak
No.
t
R
(min)
Measured
m/z
Assigned identity Molecular
formula
Theoretical exact m/zMass
accuracy
(ppm)
Reference
[M–H]
[MþHCOO]
[2M–H]
1 0.50 195.0495 NA
2 0.51 181.0694 NA
3 0.57 105.0179 NA
4 0.57 447.1347 NA
5 0.60 579.1773 NA
6 0.60 1007.3380 NA
7 0.66 799.2383 NA
8 0.81 755.2392 NA
9 0.84 665.2135 Stachyose C
24
H
42
O
21
665.2140
a
711.2195 1331.4359 0.75 8,15
10 7.66 183.0995 NA
11 8.81 435.2227 Rehmaionoside A or B C
19
H
34
O
8
389.2175 435.2230
a
779.4429 0.69 8,12
12 10.84 623.1983 Acteoside or
forsythiaside
or isoacteoside
C
29
H
36
O
15
623.1976
a
669.2031 1247.4030 1.12 8,13,14
13 11.84 523.1813 6-O-coumaroylajugol or
6-O-E-feruloylajugol or
6-O-Z-feruloylajugol
C
25
H
32
O
12
523.1816
a
569.1870 1047.3709 0.57 8
a
The underlined theoretical exact m/zvalue is the one matching the measured value. NA, not assigned.
Chemical comparison of dried rehmannia root and prepared rehmannia root by UPLC-TOF MS and HPLC-ELSD 61
promoting effect of fructose was discussed in our previous
report
19
.
The molecular structures of stachyose, verbascose, raffinose
(also a constituent of rehmannia root
14
) and sucrose all contain a
terminal fructose group. Our results indicate decreases in
stachyose, verbascose and sucrose after DRR is converted to
PRR. It does not contradict Wen et al.’s viewpoint
4
that
defructosylation may take place during the preparation process.
Stachyose, verbascose and raffinose belong to the galactosido-
sucrose series of sugars which consist of a1,6-linked chains of
D-galactose attached to the 6-glucosyl position of sucrose. These
are non-digestible by human digestive juices and may promotes
flatulence, but can be preferentially consumed by potentially
remedial intestinal bacteria in the colon
20,21
. A US patent of 2007
reported that orally administered stachyose has an antiallergenic
action, particularly in suppressing atopic dermatitis
22
. Hypogly-
cemic activity was also reported for stachyose and Rehmannia
glutinosa oligosaccharides, the majority of which is sta-
chyose
21,23,24
. Whether the special ‘‘cold’’ character and ther-
apeutic effect of ‘‘reducing the body’s heat’’ of DRR is involved
with antiallergenic or hypoglycemic activities requires further
study.
Peak 11 matching rehmaionoside A or B decreased after
preparation of PRR from DRR. The pharmacological effect
of rehmaionoside B was reported by Nakase et al.
11
in 1991.
They investigated the effects of rehmaionoside B and some
other constituents of rehmannia root on bladder and urethral
smooth muscle strips of mice, and rehmaionoside B most
effectively inhibited the contraction induced by noradrenaline
in urethral smooth muscle. Rehmaionoside B not only
inhibited the spontaneous contraction but also slightly relaxed
the normal tone of the bladder smooth muscle. They con-
cluded that the effect of rehmaionoside B on the bladder was
an indomethacin-like action. As is known, indomethacin is
outstanding for potent anti-inflammatory, analgesic and anti-
pyretic effects
25–33
. Interestingly, it has been reported that the
aqueous extract of DRR showed markedly hypotensive,
sedative and anti-inflammatory effects
34–37
but the extract of
PRR was ineffective in preventing the development of edema
in arthritic rats and in acute or chronic inflammation
38
.
Although Nakase et al.’s report
11
did not mention whether
rehmaionoside B has hypotensive, sedative or anti-
inflammatory effects, the close relationship between smooth
muscle and blood pressure, and the outstanding anti-
inflammatory effect of indomethacin prompted us to surmise
that rehmaionoside B may contribute to DRR’s hypotensive
or anti-inflammatory activity. It is necessary and worthwhile
to test this hypothesis in future studies. If this hypothesis is
true, we can further surmise that the special ‘‘cold’’ character
and therapeutic effect of ‘‘reducing the body’s heat’’ is, to
some extent, the results of the activity exerted by rehmaiono-
side as an indomethacin-like hypotensive or anti-inflammatory
drug. Further studies are required to verify these hypotheses.
Peak 12 showed an increasing trend after DDR is converted to
PRR, but it is more difficult to surmise its pharmacological
indication because it corresponds to three isomeric compounds in
rehmannia root, i.e., acteoside, forsythiaside and isoacteoside.
Acteoside was reported to exhibit anti-inflammatory, antinephri-
tic, anti-hepatotoxic, antiproliferative, cytotoxic against various
tumor cells, antimetastastic, immunomodulatory and antioxidant
activities
39–44
. Isoacteoside also showed an antiproliferative
Figure 8 Selected ion intensity trend plots of a peak (t
R
¼0.59 min, m/z¼827.2683) identified tentatively as verbascose.
Figure 7 Concentrations of D-fructose, D-glucose and sucrose in
the decoctions of DRR and PRR excluding PRR1. Data
expressed as mean7S.D., n¼6 for DRR, n¼5 for PRR.
Qiande Liang et al.62
effect
45
. It is reported that forsythiaside has strong antioxidant,
antibacterial and antiviral activities, and also exhibits neuropro-
tective properties and a slow relaxation effect against
norepinephrine-induced contraction of rat aorta
46–50
.Peak13
also corresponds to three isomeric compounds reported in
rehmannia root, i.e., 6-O-coumaroylajugol, 6-O-E-feruloylajugol
and 6-O-Z-feruloylajugol, which showed a decreasing trend after
DDR is converted to PRR. Unfortunately, we have not found
any pharmacological reports concerning these compounds.
Another notable point is that the quality standard of PRR in
the Chinese Pharmacopoeia which controls the quality of the
preparation process as well as the quality of the final product
currently is lacking a quality control compound for this prepara-
tion
1
. The quality of PRR is routinely assessed based on subjective
indices such as color, smell and glutinosity. The findings of this
study suggest that fructose, stachyose and rehmaionoside A or B
could be considered as potential quality control compounds due to
their characteristic contents in DRR and PRR.
4. Conclusions
In the present study, a chemical comparison between decoctions
of DRR and PRR by UPLC-TOF MS and HPLC-ELSD with
MVA was explored. Our results show characteristic changes in
the content of major monosaccharides and oligosaccharides as
DRR is converted to PRR, indicate a possible special role for
fructose, stachyose and rehmaionoside in the differing therapeu-
tic effects of DRR and PRR, and suggest potential quality
control markers for assessing PRR and DRR.
Acknowledgment
The authors are grateful for financial support from the
National Nature Science Foundation of China (Grant Nos.
81073161, 81130067 and 30730112), the National Basic
Research Program of China (Grant No. 2011CB505304) and
the Natural Science Foundation of Beijing (Grant No.
7112110), and for technical support from Mr. Yong Wang
and other technologists of Waters China Ltd.
References
1. Chinese Pharmacopeia Commission. The pharmacopeia of the
People’s Republic of China. Beijing: Chemical and Industrial
Publisher; 2005. p. 82–3.
2. Li SL, Song JZ, Qiao CF, Zhou Y, Qian K, Lee KH, et al. A
novel strategy to rapidly explore potential chemical markers for
the discrimination between raw and processed Radix Rehmanniae
by UHPLC–TOFMS with multivariate statistical analysis. J
Pharm Biomed Anal 2010;51:812–23.
3. Gao XM. Zhong yao xue. Beijing: People’s Medical Publishing
House; 2000, p. 1740.
4. Wen XS, Yang SL, Xiao JM, Jun HZ. HPLC chromatogram
changes with processing for roots of Radix Rehmanniae. Chin
Tradit Herb Drugs 2004;35:153–6.
5. Bian BL, Wang HJ, Ni MY. Determination of total saccharide
and several main saccharides in Rehmannia glutinosa Libosh and
its processed products. China J Chin Mater Med 1995;20:469–71.
6. Wan ECH, Yu JZ. Analysis of sugars and sugar polyols in
atmospheric aerosols by chloride attachment in liquid chromato-
graphy/negative ion electrospray mass spectrometry. Environ Sci
Technol 2007;41:2459–66.
7. Bruggink C, Maurer R, Herrmann H, Cavalli S, Hoefler F.
Analysis of carbohydrates by anion exchange chromatography
and mass spectrometry. J Chromatogr A 2005;1085:104–9.
8. Zhou JJ, Xie GR, Yan XJ. Traditional Chinese medicines mole-
cular structures. Natural sources and applications. Beijing: Chemi-
cal Industry Press; 2004.
9. Wiklund S, Johansson E, Sjostrom L, Mellerowicz EJ, Edlund
U, Shockcor JP, et al. Visualization of GC/TOF-MS-based
metabolomics data for identification of biochemically interest-
ing compounds using OPLS class models. Anal Chem 2008;80:
115–122.
10. Chang WT, Choi YH, van der Heijden R, Lee MS, Lin MK,
Kong H, et al. Traditional processing strongly affects metabolite
composition by hydrolysis in Rehmannia glutinosa roots. Chem
Pharm Bull 2011;59:546–52.
11. Nakase K, Kimura I, Kimura M, Kitagawa I. Inhibitory effect of
rehmaionoside B, a constituent of rehmaniae radix, on isolated
bladder and urethral smooth muscle of mice. Phytother Res
1991;5:67–71.
12. Son YO, Lee SA, Kim SS, Jang YS, Chun JC, Lee JC. Acteoside
inhibits melanogenesis in B16F10 cells through ERK activation
and tyrosinase down-regulation. J Pharm Pharmacol
2011;63:1309–19.
13. Sasaki H, Nishimura H, Morota T, Chin M, Mitsuhashi H,
Komatsu Y, et al. Immunosuppressive principles of Rehmannia
glutinosa var. hueichingensis. Planta Med 1989;55:458–62.
14. Tomoda M, Kato S, Onuma M. Water-soluble constituents of
Rehmanniae Radix. I. carbohydrates and acids of Rehmannia
glutinosa f. hueichingensis. Chem Pharm Bull 1971;19:1455–60.
15. Zhang RX, Li MX, Jia ZP. Rehmannia glutinosa: review of
botany, chemistry and pharmacology. J Ethnopharmacol
2008;117:199–214.
16. Liang QD, Gao Y, Tan HL, Guo P, Li YF, Zhou Z, et al. Effects
of four Si-Wu-Tang’s constituents and their combination on
irradiated mice. Biol Pharm Bull 2006;29:1378–82.
17. Ma J, Liang QD, Ma ZC, Wang YG, Liu M, Lu BB, et al.
Rehmanniae Radix provides most of the free fructose and glucose
in Si-Wu-Tang decoction. Drug Discov Ther 2010;4:179–83.
18. Li F. Fang Ji Xue. Beijing: People’s Medical Publishing House;
2002, p. 844.
19. Liang QD, Xiao CR, Ma ZC, Wang YG, Lu BB, Tan HL, et al.
Serum fructose concentration in rats after single dose oral
administration of Si-Wu-Tang. Drug Discov Ther 2010;4:175–8.
20. Peterbauer T, Mucha J, Mach L, Richter A. Chain elongation of
raffinose in pea seeds. J Biol Chem 2002;277:194–200.
21. Zheng NX, Ruan JX, Zhang YX, Ge ZH, Gu GM. Pharmaco-
kinetic study on absorption of stachyose. Chin J Integr Tradit
Chin West Med 2000;20:444–6.
22. Toshiro S, Shuichi K, Yutaka O, Yasushi U, Inventors. Anti-
allergic composition. US Patent US 7166582 B2; 2007 January 23.
23. Wu WH. Oligosaccharide preparation of Rehmannia glutinosa and
their bioactive assay [Dissertation]. Jinan: Shandong University;
2006, p. 70–3.
24. Zhang RX, Gu GM, Zhang YX, Zhou JH. The modulatory
effects of Rehmannia glutinosa oligosaccharides (ROS) on glucose
metabolism in experimental diabetic and hyperglycemic rats.
Pharmacol Clin Chin Mat Med 1996;12:14–7.
25. Ackerstaff E, Gimi B, Artemov D, Bhujwalla ZM. Anti-
inflammatory agent indomethacin reduces invasion and alters meta-
bolism in a human breast cancer cell line. Neoplasia 2007;9:222–35.
26. Nishikawa M, Onuki Y, Isowa K, Takayama K. Formulation
optimization of an indomethacin-containing photocrosslinked
polyacrylic acid hydrogel as an anti-inflammatory patch. AAPS
PharmSciTech 2008;9:1038–45.
27. Skalka N, Krol A, Schlesinger H, Altstein M. Monitoring of the
non-steroid anti-inflammatory drug indomethacin: development
of immunochemical methods for its purification and detection.
Anal Bioanal Chem 2011;400:3491–504.
Chemical comparison of dried rehmannia root and prepared rehmannia root by UPLC-TOF MS and HPLC-ELSD 63
28. Guieu R, Blin O, Pouget J, Serratrice G. Analgesic effect of
indomethacin shown using the nociceptive flexion reflex in
humans. Ann Rheum Dis 1992;51:391–3.
29. Segstro R, Morley-Forster P, Lu G. Indomethacin as a post-
operative analgesic for total hip arthroplasty. Can J Anesth
1991;38:578–81.
30. Shrestha M, Morgan DL, Moreden JM, Singh R, Nelson M,
Hayes JE. Randomized double-blind comparison of the analgesic
efficacy of intramuscular ketorolac and oral indomethacin in the
treatment of acute gouty arthritis. Ann Emerg Med 1995;26:682–6.
31. Parrott RF, Vellucci SV. Comparison of the antipyretic actions of
indomethacin and L-745,337, a selective cyclooxygenase-2 inhibitor,
in endotoxin-treated prepubertal pigs. Gen Pharm 1998;30:65–9.
32. Wilairatana P, Looareesuwan S. Antipyretic efficacy of indo-
methacin and acetaminophen in uncomplicated falciparum
malaria. Ann Trop Med Parasitol 1994;88:359–63.
33. Parrott R, Vellucci S. Cyclooxygenase-independent mechanisms
are not involved in the antipyretic action of indomethacin in the
prepubertal pig. Methods Find Exp Clin Pharmacol 1996;18:321–6.
34. Chang DM, Yue SH, Liu HX, Yang YT, Niu GP. Influence of
Rehmannia glutinosa on blood pressure of mice. J Xinxiang Med
Coll 1989;6:9–11.
35. Chang JM, Liu XY, Chang JH. Experimental study of Rehmannia
glutinosa affecting blood pressure. Lishizhen Med Mater Med Res
1998;9:416–7.
36. Liu HX, Cao ZL, Chang DM, Wu SK. Effection of HuaiQing
Radix Rehmanniae on hypotensive and sedative and anti-
inflammatory actions and analysis of its effective part. J Xinxiang
Med Coll 1998;15:219–21.
37. Wu SK, Liu HX, Liu CX, Xiao ZM. Depressant effect of
Rehmannia glutinosa on central nervous system. J Xinxiang Med
Coll 1989;6:12–4.
38. Michinori K, Toshiki A, Hidemi S, Hideaki M. Studies on
Rehmanniae Radix. I. Effect of 50% ethanolic extract from
steamed and dried Rehmanniae Radix on hemorheology in
arthritic and thrombosic rats. Biol Pharm Bull 1994;17:1282–6.
39. Papoutsi Z, Kassi E, Mitakou S, Aligiannis N, Tsiapara A, Chrousos
GP, et al. Acteoside and martynoside exhibit estrogenic/antiestro-
genic properties. J Steroid Biochem Mol Biol 2006;98:63–71.
40. Lee KW, Kim HJ, Lee YS, Park HJ, Choi JW, Ha J, et al.
Acteoside inhibits human promyelocytic HL-60 leukemia cell
proliferation via inducing cell cycle arrest at G0/G1 phase and
differentiation into monocyte. Carcinogenesis 2007;28:1928–36.
41. Lee KY, Jeong EJ, Lee HS, Kim YC. Acteoside of Callicarpa
dichotoma attenuates scopolamine-induced memory impairments.
Biol Pharm Bull 2006;29:71–4.
42. Hayashi K, Nagamatsu T, Ito M, Hattori T, Suzuki Y. Acetoside,
a component of Stachys sieboldii MIQ, may be a promising
antinephritic agent: effect of acteoside on crescentic-type anti-
GBM nephritis in rats. Jpn J Pharmacol 1994;65:143–51.
43. Ohno T, Inoue M, Ogihara Y, Saracoglu I. Antimetastatic activity of
acteoside, a phenylethanoid glycoside. Biol Pharm Bull 2002;25:666–8.
44. Hausmann M, Obermeier F, Paper DH, Balan K, Dunger N,
Menzel K, et al. In vivo treatment with the herbal phenylethanoid
acteoside ameliorates intestinal inflammation in dextran sulphate
sodium-induced colitis. Clin Exp Immunol 2007;148:373–81.
45. Abe F, Nagao T, Okabe H. Antiproliferative constituents in
plants 9. aerial parts of Lippia dulcis and Lippia canescens. Biol
Pharm Bull 2002;25:920–2.
46. Wang GN, Pan RL, Liao YH, Chen Y, Tang JT, Chang Q. An
LC-MS/MS method for determination of forsythiaside in rat
plasma and application to a pharmacokinetic study. J Chromatogr
B2010;878:102–6.
47. Zhou W, Di LQ, Shan JJ, Bi XL, Chen LT, Wang LC. Intestinal
absorption of forsythoside A in different compositions of Shuang–
Huang–Lian. Fitoterapia 2011;82:375–82.
48. Kim JM, Kim S, Kim DH, Lee CH, Park SJ, Jung JW, et al.
Neuroprotective effect of forsythiaside against transient cerebral
global ischemia in gerbil. Eur J Pharmacol 2011;660:326–33.
49. Lu T, Piao XL, Zhang Q, Wang D, Piao XS, Kim SW. Protective
effects of Forsythia suspensa extract against oxidative stress
induced by diquat in rats. Food Chem Toxicol 2010;48:764–70.
50. Li H, Wu J, Zhang Z, Ma Y, Liao F, Zhang Y, et al. Forsythoside
a inhibits the avian infectious bronchitis virus in cell culture.
Phytother Res 2011;25:338–42.
Qiande Liang et al.64
