Preparative Separation of Alkaloids from Stem of Euchresta tubulosa Dunn. by High-Speed Counter-Current Chromatography Using Stepwise Elution

Euchresta tubulosa Dunn. is a Chinese herbal medicine with biological activity, but there are few studies on its components at present. Alkaloids in the stem of Euchresta tubulosa Dunn. were isolated and purified by high-speed counter-current chromatography (HSCCC) using stepwise elution. First of all, liquid-liquid extraction (methylene chloride-methanol-water, 5:1:4, v/v) was used for the preliminary enrichment. According to the partition coefficient (K) of a target compound in a series of different two-phase solvents, the final result was that carbon tetrachloride-methylene chloride-methanol-water (2:3:3:2, v/v) (1) and methylene chloride-methanol-water (5:3:2, v/v) (2) were suitable for the HSCCC using stepwise elution. As a result, the purity was all higher than 93% and matrine (1), oxymatrine (2), N-formyl cytisine (3), and N-acetyl cytisine (4) can be eluted at one time by this mode. Cytisine-type alkaloids were isolated for the first time in this plant. Finally, the applicability of the mode was verified.

to the earlier work of fractions A and B by silica gel column chromatography. The purity and structure of the monomer were analyzed, and the applicability of the model was verified.

Results
This model is based on a previous study of silica gel column chromatography, combined with the current HSCCC separation technology, which can be used to separate and purify alkaloids with a similar structure. The model is shown in Figure 1.

Results
This model is based on a previous study of silica gel column chromatography, combined with the current HSCCC separation technology, which can be used to separate and purify alkaloids with a similar structure. The model is shown in Figure 1.

Selection of Optimum Solvent Systems of HSCCC Separation
The K of the target compound was calculated according to Formula (1) and HPLC-UV. Table 1 shows the K of the four target compounds.
The two-phase solvent system of fraction A with a bigger Rf was screened by the Ito method [18]. Firstly, this experiment began with CH2Cl2-MeOH-H2O (2:1:1, v/v). The proper K can be obtained by changing the volume ratio of the solvent and adding carbon tetrachloride in the system. Table 1 shows the K values. According to the actual situation and HSCCC operation, CCl4-CH2Cl2-MeOH-H2O (2:3:3:2, v/v) was determined as the optimum solvent system of fraction A.
The two-phase solvent system of fraction B with a smaller Rf was screened by the Ito method, too. Table 1 shows that K3 and K4 are suitable in CH2Cl2-MeOH-H2O (2:1:1, v/v). Minor adjustments were made according to the actual situation and HSCCC operation. Finally, CH2Cl2-MeOH-H2O (5:3:2, v/v) was determined to be the optimal solvent system of fraction B.
It was found that the polarities of the two fractions were quite different. However, the stationary phase of the two solvent systems was the same. Therefore, preparative separation of four alkaloids by HSCCC using stepwise elution was considered.

Selection of Optimum Solvent Systems of HSCCC Separation
The K of the target compound was calculated according to Formula (1) and HPLC-UV. Table 1 shows the K of the four target compounds.
The two-phase solvent system of fraction A with a bigger R f was screened by the Ito method [18]. Firstly, this experiment began with CH 2 Cl 2 -MeOH-H 2 O (2:1:1, v/v). The proper K can be obtained by changing the volume ratio of the solvent and adding carbon tetrachloride in the system. Table 1 shows the K values. According to the actual situation and HSCCC operation, CCl 4 -CH 2 Cl 2 -MeOH-H 2 O (2:3:3:2, v/v) was determined as the optimum solvent system of fraction A.
The two-phase solvent system of fraction B with a smaller R f was screened by the Ito method, too. Table 1 shows that K 3 and K 4 are suitable in CH 2 Cl 2 -MeOH-H 2 O (2:1:1, v/v). Minor adjustments were made according to the actual situation and HSCCC operation. Finally, CH 2 Cl 2 -MeOH-H 2 O (5:3:2, v/v) was determined to be the optimal solvent system of fraction B.
It was found that the polarities of the two fractions were quite different. However, the stationary phase of the two solvent systems was the same. Therefore, preparative separation of four alkaloids by HSCCC using stepwise elution was considered.

Sample Pretreatment of HSCCC Separation
The results showed that four alkaloids were easily dissolved in dichloromethane. It was found that CH 2 Cl 2 -MeOH-H 2 O (5:1:4, v/v) can retain most of the four alkaloids in the lower phase and can remove the compounds with high water and methanol solubility. The crude alkaloid extract (1.5 g) was extracted until the dichloromethane layer almost had no precipitation reaction with bismuth potassium iodide. Finally, 0.8 g of the sample was obtained by combining the dichloromethane layer.

HSCCC Separation
Thus as to shorten the preparation time and reduce the workload, HSCCC using stepwise elution was adopted. Finally, both CCl 4 -CH 2 Cl 2 -MeOH-H 2 O (2:3:3:2, v/v) and CH 2 Cl 2 -MeOH-H 2 O (5:3:2, v/v) were selected for the isolation of four target alkaloids. The two alkaloids (matrine-type) were eluted by the lower phase of CCl 4 -CH 2 Cl 2 -MeOH-H 2 O (2:3:3:2, v/v), and then the two alkaloids (cytisine-type) were eluted by the lower phase of CH 2 Cl 2 -MeOH-H 2 O (5:3:2, v/v). Figure 2 shows the results of HSCCC with stepwise elution; peak 1 (matrine), peak 2 (oxymatrine), peak 3 (N-formyl cytisine), and peak 4 (N-acetyl cytisine). Cytisine-type alkaloids were isolated in this plant for the first time. The results showed that four alkaloids were easily dissolved in dichloromethane. It was found that CH2Cl2-MeOH-H2O (5:1:4, v/v) can retain most of the four alkaloids in the lower phase and can remove the compounds with high water and methanol solubility. The crude alkaloid extract (1.5 g) was extracted until the dichloromethane layer almost had no precipitation reaction with bismuth potassium iodide. Finally, 0.8 g of the sample was obtained by combining the dichloromethane layer.

Identification of Purified Alkaloids
The structure elucidations of compounds 1-4 were finally achieved by analyzing the EI-MS and NMR spectroscopic data and comparing them with data from the literature [19][20][21]. Their spectral data are shown in Figure S1-S4. Each group of 13 C-NMR of compound 3 had two sets of similar overlapping signals. Because the C-N linked to formyl groups on N is linked by a single bond, it is easy to change the configuration, and the NMR spectra show splitting [22,23]. The NMR spectra of compound 4 were almost the same as those of compound 3. They were paired signals. However, the n-aldehyde hydrogen signal disappeared, and there was an additional methyl proton signal associated with the acyl group and a methyl carbon signal associated with the acyl group in 13 C-NMR. Finally, compounds 1-4 were identified as Matrine (1), Oxymatrine (2), N-formyl cytisine (3), N-acetyl cytisine (4), respectively. Their structures are shown in Figure 4.

Identification of Purified Alkaloids
The structure elucidations of compounds 1-4 were finally achieved by analyzing the EI-MS and NMR spectroscopic data and comparing them with data from the literature [19][20][21]. Their spectral data are shown in Figure S1-S4. Each group of 13 C-NMR of compound 3 had two sets of similar overlapping signals. Because the C-N linked to formyl groups on N is linked by a single bond, it is easy to change the configuration, and the NMR spectra show splitting [22,23]. The NMR spectra of compound 4 were almost the same as those of compound 3. They were paired signals. However, the n-aldehyde hydrogen signal disappeared, and there was an additional methyl proton signal associated with the acyl group and a methyl carbon signal associated with the acyl group in 13 C-NMR. Finally, compounds 1-4 were identified as Matrine (1)

Precision and Repeatability Test
The Rlative Sandard Deviation (RSD) of the four alkaloid peak areas was 1.81%, 1.59%, 2.36%, 2.17%, respectively. It shows that the precision of the instrument was good, and the experiment was reproducible.

Applicability Test
The higher the HSCCC temperature was, the easier the two-phase solvent system was to emulsify to get a higher Sf. Considering that dichloromethane has a boiling point of only 39 °C, it was easy to cause solvent vaporization due to a too high temperature, which will reduce the separation effect. In combination with the use of the instrument, conditions were set to the speed of 700-900 rpm/ min, a flow rate of 2-4 mL/min, and a temperature of 25 °C.
The Sf value was calculated according to Formula (2) and the results are shown in the Table 2. It can be found that this method has wide applicability (Sf > 50%).

Precision and Repeatability Test
The Rlative Sandard Deviation (RSD) of the four alkaloid peak areas was 1.81%, 1.59%, 2.36%, 2.17%, respectively. It shows that the precision of the instrument was good, and the experiment was reproducible.

Applicability Test
The higher the HSCCC temperature was, the easier the two-phase solvent system was to emulsify to get a higher S f . Considering that dichloromethane has a boiling point of only 39 • C, it was easy to cause solvent vaporization due to a too high temperature, which will reduce the separation effect. In combination with the use of the instrument, conditions were set to the speed of 700-900 rpm/min, a flow rate of 2-4 mL/min, and a temperature of 25 • C.
The S f value was calculated according to Formula (2) and the results are shown in the Table 2. It can be found that this method has wide applicability (S f > 50%).

Reagents and Materials
Whole plants of E. tubulosa Dunn. were collected from the Tianping Mountain Reserve of Badagongshan National Forest Park in Zhangjiajie, Hunan Province, China, identified by Ai-Wen Dong, research fellow, Hunan Key Laboratory of Forest Chemical Engineering, Jishou University. The voucher deposit number: YSF2017-11694. All solvents used for experiments were of analytical grade (Concord Technology Co., Ltd., Tianjin, China). Methanol used for HPLC-UV was of chromatographic grade (Concord Technology Co., Ltd., Tianjin, China). Silica gels (60-100, 100-200 mesh) used for sample preparation were purchased from the Qingdao Ocean Chemical Plant (Shanghai, China); water used was laboratory self-control distilled water. Thin-layer chromatography (TLC) was used with GF254 plates (Qingdao Marine Chemical, Inc., Shandong, China).

Apparatus
The instrument of HSCCC using stepwise elution was a TBE-300B high-speed counter-current chromatograph (Tauto Biotechnique, Shanghai, China) assembled by a polytetrafluoroethylene multilayer coil (i.d. of the tubing = 1.5 mm, total volume = 285 mL) and a 20 mL sample loop.
The β values of the multilayer coil varied from 0.5 at the internal terminal to 0.8 at the external terminal (β = r/R, where r was the rotation radius from the coil to the holder shaft, and R was the revolution radius between the holder axis and central axis of the centrifuge or the distance). The system was also equipped with a TBP1002ST constant pump (Tauto Biotechnique, Shanghai, China), a model TBD2000-UV monitor (Tauto Biotechnique, Shanghai, China). The datum was recorded by the WH500-USB chromatography workstation (Tauto Biotechnique, Shanghai, China).
NMR experiments were carried out using a Bruker AVANCE III NMR spectrometer (600 Hz, Bruker Biospin, Germany) with CHCl 3 (CDCl 3 ) or MeOH (MeOD) as the solvent. MS analysis of purified alkaloids was performed by the EI mass spectrometer (DSQ, Thermo, USA).

Preparation of the Crude Sample
The dried powder (60 mesh) from the stem of Euchresta tubulosa Dunn. (3 kg) were extracted under reflux with 10 L of 95% industrial alcohol 3 times, at 3 hours each time, at 80 • C. Then, the extract was filtrated, combined, and vacuum evaporated at 50 • C. The residue of the mixed extract was dissolved in 0.8 L acid aqueous solution with a pH of 2.0 adjusted by 2% hydrochloric acid. The acid water solution was extracted with petroleum ether to degrease, and then the pH of the acidic aqueous solution was slowly increased to 9.5 using aqueous ammonia. Finally, the total alkaloid extract was extracted with equivalent volume trichloromethane 4 times. The trichloromethane extract was combined and vacuum evaporated at 50 • C. The crude alkaloid extract (15.2 g) was preserved at −5 • C for further separation and purification.

Selection of the Two-Phase Solvent Systems
Total alkaloids (2.5 g) were dissolved in petroleum ether with 2.8 g 60-100 mesh silica gel and separated preliminaries with 100-200 mesh column chromatography silica-gel using a silica-gel column (26 mm × 457 mm). Gradient elution was carried out with a CH 2 Cl 2 : MeOH (1:0, 50:1, 25:1, 10:1, 0:1, v/v) solvent system (all adding triethylamine to adjust pH 7-8). Each 80 mL eluent was a fraction. TLC was used to track and detect each fraction. According to the R f of the eluent, 2 fractions A (0.7 g) and B (1.1 g) were found by merging their similar products. Other fractions were of less quality and more complex compounds, thus this study did not consider them.
Various types of two-phase solvent systems composed of CCl 4 -CH 2 Cl 2 -MeOH-H 2 O with different volume ratios were equipped with a small volume (8 mL) and completely stratified by ultrasound. An equivalent volume (2 mL) of the upper and lower phases of the 7 2-phase solvent systems was absorbed to make a small solvent system, and then 1 mg of crude alkaloid extracts was dissolved in the small solvent system. The K of the single alkaloid was calculated by HPLC-UV and the formula (1): A s : Peak area of a single alkaloid in the stationary phase; A m : Peak area of a single alkaloid in the mobile phase.

Preparation of HSCCC Two-Phase Solvent Systems and Sample Solution
The

HSCCC Separation Procedures
Before the beginning of the experiment, the multilayer helical column was completely filled with the upper phase of carbon tetrachloride-methylene chloride-methanol-water (2:3:3:2, v/v) at a flow rate of 15 mL/min. Then, the ultraviolet detector was turned on at 254 nm. The instrument was run at a revolution speed of 900 rpm and the constant temperature tank was opened at 25 • C. At the same time, the lower phase of the carbon tetrachloride-methylene chloride-methanol-water (2:3:3:2, v/v) was pumped into the chromatographic column at a speed of 1.8 ml/min of the head-tail elution mode until the liquid effluent formed a stable 2-phase system. The sample (300 mg) was injected into the column through the injection valve when hydrodynamic equilibrium was established. Fractions were manually collected per 10 mL according to the spectrum of the ultraviolet detector. The pump was stopped at about 170 min and the mobile phase was changed to the lower phase of methylene chloride-methanol-water (5:3:2, v/v) at 1.8 mL/min. The compounds were collected according to the UV detector. Then, samples were stored at −5 • C and merged by subsequent HPLC-UV analysis. Finally, all the instruments were closed at about 300 min and the liquid in the column was blown clean with nitrogen gas to recover the sample. The S f was 72%.

Precision and Repeatability Test
The HSCCC sample obtained by liquid-liquid extraction of total alkaloids was repeatedly injected for 6 times according to the method in 3.6., with 300 mg each time. The peak areas of target compounds were recorded separately.

Applicability Test
The S f was a very important parameter in HSCCC, and the feasibility of the operating parameters depended on S f . In this experiment, the effects of speed, temperature, and flow rate on S f were investigated. A higher S f value can obtain a better separation effect of HSCCC.
In this verification experiment, the upper phase of the 2-phase system (1) was filled with the whole pipeline at 15mL/min (the outlet of the pipeline was connected to a 250 mL measuring cylinder). Then, the lower phase of the system (1) was pumped into the instrument. The volume (V 1 ) of the upper phase in the measuring cylinder was recorded until the liquid outflow at the outlet of the pipe formed a 2-phase and the baseline of the UV detector was balanced. Then the mobile phase was replaced with the lower phase of the system (2). The volume (V 2 ) of the upper phase in the measuring cylinder was recorded until the liquid outflow at the outlet of the pipe formed a 2-phase and the baseline of the UV detector was balanced. Finally, all the liquid in the instrument was blown out to the measuring cylinder, and the volume (V 3 ) of the upper phase retained in the instrument was recorded.
V 1 : Loss volume of the first phase separation; V 2 : Loss volume of the second phase separation; V 3 : Volume of upper phase retained in the instrument.

Conclusions
Four alkaloids matrine (1), oxymatrine (2), N-formyl cytisine (3), and N-acetyl cytisine (4) were isolated and purified systematically by HSCCC using stepwise elution. Cytisine-type alkaloids were isolated for the first time from this plant by this mode. The results of the research provided a successful pattern for isolation and purification of alkaloids with different polarities. In this mode, only the liquid-liquid extraction and once the HSCCC using stepwise elution were used to obtain four alkaloids. It can make compounds separated quickly and effectively. This mode also solves the problem that acid or ionic solution needs to be added to obtain suitable K, which is difficult to deal with in the later stage, adsorption on the solid carriers, and separation of alkaloids with different polarities. The verification test (S f > 50%) of this mode shows that it has wide applicability.