Instruments and materials

Instruments and materials
HSCCC (TBE-300) is from Shenzhen Tauto Biotech,
China, with three preparative coils connected in series
(diameter of tube = 2.6 mm, total volume = 300 ml)
and a 20ml sample loop. The HPLC system is from
Shimadzu, Japan with a 20 l sample loop. The column
used was Ultrasphere C18 column (150mm 亊 4.6mm
i.d., 5 m, Shimadzu, Japan). Ethanol and n-hexane were
analytical-grade chemicals from Atoz Fine Chemicals,
Tianjin, China. Reverse osmosis water (18M , Milli-Q,
Milipore, USA) was used for all solutions and dilutions.
Acetonitrile and aqueous trifluoroacetic acid (TFA) were
chromatographic-grade chemicals from Fisher Scientific,
UK.
2.2. Solvent system for HSCCC
Solvent system A: n-hexane.ethanol.water (10:5.5:4.5,
v/v). Solvent system B: n-hexane.ethanol.water (10:7:3,
v/v) [5]. Each mixture was equilibrated thoroughly in a funnel
at room temperature. The upper phase and lower phase
were separated before use.
2.3. Sample preparation
The powdered dried roots (20 g) of S. miltiorrhiza Bunge
from Hebei, Shandong and Jiangsu province were added
to 50 ml of n-hexane.ethanol (1:1, v/v). The mixture was
shaken for 45 min. It was then centrifuged at 10,000 亊 g
for 10 min, and the supernatant was saved. The same extraction
procedure was repeated once and the supernatants
were combined. The extract was diluted with water at a
ratio of 1:2 and equilibrated for 2 h. The organic phase
was separated and washed with 30% aqueous ethanol until
the water phase was almost colorless. The organic extracts
were dried by rotary vaporization at 40 .C to yield
the final crude samples [5]. 100 mg crude sample was dissolved
in the lower phase of solvent system A before HSCCC
separation.
2.4. HSCCC separation
Preparative separation was performed using a stepwise
elution with solvent systems A and B in sequence. First,
the coiled column was filled with the upper phase of solvent
system A. Then, the apparatus was rotated at 900 rpm,
and at the same time, the lower phase of solvent system
A was pumped through the column at a flow-rate of
2.0 ml/min. After the mobile phase emerged in the effluent
and hydrodynamic equilibrium was established in the
column, 6ml of the sample solution containing 100 mg
of the crude tanshinones was injected through the valve.
The effluent was monitored with a UV-Vis detector at
280 nm and the peak fractions collected separately. After
470 min of elution, the mobile phase was changed
to the lower phase of system B until the end of the
separation.
2.5. HPLC analysis
The column used was a Ultrasphere C18 column
(150mm 亊 4.6mm i.d., 5 m, Shimadzu, Japan). The
mobile phase was solvent A (0.1% aqueous TFA) and solvent
B (0.1% TFA + acetonitrile) in the gradient mode as
M. Gu et al. / Journal of Chromatography A, 1022 (2004) 139.144 141
follows: 0.5 min, 0% B; 5.25 min, 0.70% B; 25.40 min,
70% B; 40.41 min, 70.0% B. The flow-rate was 1.0 ml/min.
The effluent was monitored at 280 nm.
2.6. UV-Vis spectrophotometer scanning
Each peak fraction and standard sample was scanned by
a UV-Vis spectrophotometer at 900.200 nm.
3. Results and discussion
3.1. Fingerprint development of S. miltiorrhiza Bunge
by HSCCC
The main components in the crude sample of S. miltiorrhiza
Bunge were tanshinones [8]. The medicinal plants
from three different provinces were separated by HSCCC
as described in Section 2.4. Step-wise elution showed better
performance than one-step elution in previous studies
[9]. An optimized step-wise elution strategy was performed
in our studies: 0.470 min, in solvent system A; then in
solvent system B. HSCCC system was performed at a
speed of 900 rpm and at a flow-rate of 2 ml/min [8]. Retention
of the stationary phase was 78.8%, which assured
the resolution of separation. In our experiments, 12 distinct
peak fractions were eluted, respectively, from the
three crude samples within 13 h (Fig. 1). More peak fractions
were eluted in our studies than previous reports [8,9]
after an optimized elution strategy was applied in separation
of tanshinones from Salvia miltiorrhiza Bunge by
HSCCC.
The content of each peak fraction varied greatly in different
samples as shown in Fig. 1 and Table 1, which confirmed
that location and climate had great impact on the quality of
Table 1
Relative contents of peaks in HSCCC separation
Peak no. Hebei (%) Shandong (%) Jiangsu (%)
1 2.0 2.8 1.4
2 0.3 0.1 0.1
3 0.5 0.1 0.1
4 2.0 2.5 2.3
5 2.4 1.0 1.9
6 0.5 0.9 0.3
7 19.5 12.9 15.3
8 5.3 20.2 9.1
9 9.0 1.6 4.2
10 10.6 8.0 13.6
11 40.0 48.7 48.5
12 8.2 1.3 3.3
Conditions: (column) multilayer coil of 2.6mm i.d. tube with a total
capacity of 300 ml; rotary speed: 900 rpm; stationary phase: the upper
phase of solvent system A; mobile phase: 0.470 min, the lower phase of
solvent system A and after 470 min, the lower phase of solvent system B;
flow-rate: 2ml min.1; detection at 280 nm; sample size: 100 mg; retention
of stationary phase: 78.8%.
Table 2
Retention times of peaks of HSCCC separation in HPLC analysis
Peak no. Retention time (min)
Heibei Shandong Jiangsu Average
1 27.67 27.69 27.65 27.67
2 27.60 27.64 27.63 27.62
3 29.00 29.00 29.01 29.00
4 28.29 28.28 28.27 28.28
5 29.13 29.14 29.13 29.13
6 30.91 30.94 30.90 30.92
7 30.79 30.77 30.79 30.78
8 29.46 29.42 29.43 29.44
9 31.07 31.21 31.21 31.16
10 32.29 32.29 32.31 31.30
11 34.69 34.64 34.66 34.66
12 35.54 35.53 35.52 35.53
Conditions: (column) reversed-phase Ultrasphere C18 column (150mm亊
4.6mm i.d., 5 m); mobile phase: solvent A (0.1% aqueous TFA) and
solvent B (0.1% TFA + acetonitrile) in the gradient mode; flow-rate:
1.0 ml min.1; detection at 280 nm.
CTM. A feasible quality control system is therefore necessary
for CTM.
The 12 peak fractions from each crude sample were collected,
respectively. Their retention times in HPLC analysis
and absorption spectrum in UV-Vis scanning were identified
individually (Tables 2 and 3). Although retention times
and absorption spectrums of peak fraction 1 were similar
to those of peak fraction 2, there was not enough evidence
to argue that the two peak fractions were the same constituent.
Therefore, we consider peak fractions 1 and 2 were
different. Retention time and absorption spectrum of peak
fractions 1.12 from three crude samples were in good correspondence
so they were consistent in all samples [12].
The good correspondence of the peaks of all crude samples
made it an applicable fingerprint for S. miltiorrhiza
Bunge. Thus, the peak profile of the 12 components made
up the fingerprint of S. miltiorrhiza Bunge with universal
features.
All 12 components can be defined as common peaks in
the fingerprint, which means non-common peak area is 0,
less than the national standard (10%). Retention time was
an important parameter in the fingerprint [1], so precision
of HSCCC could be defined by relative standard deviation
(R.S.D.) of retention time. R.S.D. values of the 12 peaks
were less than 3% (maximum 2.9% and average 2.7%) as
shown in Table 4, which met the demands of the national
standard.
If semi-preparative HSCCC with 300 ml capacity was replaced
by analytical HSCCC with 50 ml capacity, precision
could be increased greatly and run time could be decreased
evidently [5]. Temperature was not the most important parameter
in HSCCC separation [13]. So it was not discussed
in this study.
HSCCC was proven to be feasible as a new method to develop
the fingerprint of CTM in our studies. When applying
142 M. Gu et al. / Journal of Chromatography A, 1022 (2004) 139.144
Fig. 1. (A) Chromatograms of crude samples of Salvia miltiorrhiza Bunge from three different provinces by HSCCC separation; (A1) sample from
Hebei; (A2) sample from Shandong; (A3) sample from Jiangsu. Conditions: (column) multilayer coil of 2.6mm i.d. tube with a total capacity of 300 ml;
rotary speed: 900 rpm; stationary phase: the upper phase of solvent system A; mobile phase: 0.470 min, the lower phase of solvent system A and after
470 min, the lower phase of solvent system B; flow-rate: 2ml min.1; detection at 280 nm; sample size: 100 mg; retention of stationary phase:78.8%. (B)
Chromatograms of crude samples of Salvia miltiorrhiza Bunge from three different provinces by HPLC analysis; (B1) sample from Hebei; (B2) sample
from Shandong; (B3) sample from Jiangsu. Conditions: (column) reversed-phase Ultrasphere C18 column (150mm 亊 4.6mm i.d., 5 m); mobile phase:
solvent A (0.1% aqueous TFA) and solvent B (0.1% TFA + acetonitrile) in the gradient mode; flow-rate: 1.0 ml min.1; detection at 280 nm.
HSCCC in practice to develop a fingerprint, some factors
need to be considered carefully, including the sample stability
(10 collections of samples should be included at least),
the precision and repeatability of apparatus.
3.2. Fingerprint development of S. miltiorrhiza
Bunge by HPLC
HPLC, being conventionally applied in analysis and separation
of natural products, is recommended by Chinese
Pharmacopoeia in development of the fingerprint. As a
comparison to HSCCC, HPLC was used to develop fingerprint
of S. miltiorrhiza Bunge in our studies as described in
Section 2.5 [1].
There were 11 common peak fractions shown, respectively,
in the three samples by HPLC (Fig. 1). It is known that
the correspondence of peaks in a group of chromatograms
can be preliminarily determined by the retention time in
HPLC analysis [1]. There is professional software for further
analysis of similarity of a group of chromatograms. In
this study, we deterred preliminary correspondence based on
the retention time.
Conditions: scanned by a UV-Vis spectrophotometer at 900.200 nm.
The fingerprint developed by HPLC was composed of the
11 common peaks eluted in 45 min. R.S.D. values of retention
times of the corresponding peaks in HPLC analysis
were very small (maximum 0.66% and average 0.10%) as
shown in Table 5. HPLC showed great advantages in precision
and running time over semi-preparative HSCCC. Total
peak area of non-common peak was less than 10%, which
met the standards.
There were 12 peaks in HSCCC separation and 11 peaks
in HPLC analysis due to some differences between two systems.
There is solid matrix in HPLC column, which possibly
lead to adsorption of some constituents. Therefore, it
was hard to find correlations between peak areas in HSCCC
and in HPLC. Fingerprints of Salvia miltiorrhiza Bunge
could be developed by either