A targeted strategy for ingredients analysis and enrichment by mass-based preparative LC method: application to three isomeric C₂₁ steroids from Marsdenia tenacissima

Abstract

This paper proposed a strategy to acquire specified ingredients in small amounts from a complicated herbal extract by combination of simple Sephadex LH-20 column chromatography and a mass-based preparative liquid chromatography (LC) method. The developed method was applied to the analysis and enrichment of tenacigenin A, tenacigenin B, and 17β-tenacigenin B, which lack any significant chromophores from Marsdenia tenacissima. A chloroform extract of M. tenacissima was initially subjected to a Sephadex LH-20 column to yield two fractions. From 20 mg of the smaller molecular weight fraction, 3.0 mg tenacigenin A, 2.0 mg tenacigenin B and 3.0 mg 17β-tenacigenin B were successfully separated via preparative-LC employing mass-based fraction collection. In contrast to the active splitter system previously applied, the modified purification system via a tee was economic with a high convenience. The purities of the three obtained compounds were all above 90% as determined by LC coupled with evaporative light scattering detection (ELSD), and their structures were identified by 1D NMR spectra. The results demonstrated that the proposed strategy may well be an efficient and environmentally friendly technique for systematic isolation of bioactive components from traditional medicinal herbs.

---

Introduction

In most cases, separation and purification of natural products from plants are conducted by classic methods such as repeated manual column chromatography,¹ which, however, is tedious and may cause adsorbing effects in the stationary phase. In this instance, it is difficult to obtain minor and/or specified compounds. Furthermore, the application of chlorinated solvents in conventional purification steps is considered to contaminate the environment. Given that liquid chromatography hyphenated to mass spectrometry (LC-MS) is a rapid, sensitive and powerful tool for natural compound discovery,² preparative (pre-) LC employing a mass-based fraction collection method has been a specific alternative approach for compounds acquisition, especially for those compounds that lack any significant chromophores when compared with fraction triggering with a less specific detector such as a UV detector. The key advantage of mass-based fraction collection was that only the target compound is collected by the mass selective detector (MSD), and no redundant fractions need to be sorted out which saves time and resources. However, one-injection LC-MS analysis for plant extracts could characterize many compounds, in which minor and/or specified constituents are usually overlapping in the MS spectrum, and repetitive loading of crude samples onto the column significantly reduces column lifetime³ and contaminates the MS detector. Therefore, sample preparation before pre-LC analysis was suggested, which is important for LC separation and good for the MS detector.

Marsdenia tenacissima (Chinese name: Tong-guan-teng), the stem of M. tenacissima (Roxb.) Wight et Arn. (Family Asclepiadaceae), is widely used in China for the treatment of asthma, cancer, trachitis, tonsillitis, pharyngitis, cystitis, and pneumonia.⁴ C₂₁ steroids and their glycosides (polyoxypregnanes) are rich in the stems of M. tenacissim.^5–11 Preliminary studies in China have shown that C₂₁ steroids show cytotoxicity against the KB-V1 cell line,⁷ can reverse multidrug resistance in P-glycoprotein-overexpressing multidrug-resistant cancer cells,¹² and possess anticancer efficacy.^13,14 The standard extract of M. tenacissima (Trade name: Xiao-ai-ping Injection) was reported to be clinically effective for various types of cancer including; esophagus, liver, colorectal cancer, cervical cancer, leukemias and other malignancies, alone or when combined with chemotherapeutic agents.^11,15–20 Compared with C₂₁ steroidal glycosides, which are the major components in M. tenacissima, tenacigenin A, tenacigenin B, and 17β-tenacigenin B as the main C₂₁ steroids are minor components (Fig. 1). Those three compounds were constitutional isomers (functional group isomers) without suitable chromophores for LC UV detection. Therefore, preparation of these C₂₁ steroids is necessary for further pharmacological and clinical researches, both in vitro and in vivo.

Fig. 1 Chemical structures of three isomeric C₂₁ steroids.

Hence, this paper proposed a strategy to acquire these C₂₁ steroids in small amounts from M. tenacissima extract by combination of simple Sephadex LH-20 column chromatography and a mass-based pre-LC method. The results also demonstrated that the developed mass-based pre-LC method could be an efficient technique for systematic isolation of specific or minor bioactive components from traditional medicinal herbs.

Experimental

Reagents and chemicals

The dried chloroform extract of M. tenacissima (CMTE) was kindly provided by Nanjing Sanhome Pharmaceutical Co., Ltd. (Nanjing, Jiangsu, China). Methanol was of HPLC grade (Merck, Darmstadt, Germany). Formic acid was of HPLC grade (Newark, DE 19711, USA). Ultrapure water was prepared using a Milli-Q Water purification system (Millipore, MA, USA). All samples were filtered through a 0.45 μm Millipore membrane before sample injection.

Macroporous adsorbing resin and Sephadex LH-20 column chromatography

CMTE (229.8 mg) was chromatographed to a macroporous adsorbing resin (D-101, ZTC-1, 0.3–1.2 mm, Tianjin Zhentiancheng Science & Technology Co., Ltd, China) column (1.5 × 10 cm), eluted with MeOH/H₂O (0 : 100 to 100 : 0, stepwise, v/v), yielding three fractions. All fractions were dried under vacuum (MAR1-3, 28.1, 14.5, and 20.5 mg, respectively). Each fraction was dissolved in 2 mL methanol (HPLC grade).

CMTE (229.4 mg) was chromatographed to Sephadex LH-20 gel (40–70 μm, Amersham Pharmacia Biotech AB, Uppsala, Sweden) column (1.8 × 150 cm), eluted with MeOH/H₂O (50 : 50, v/v), yielding two fractions. All fractions were dried under vacuum (26.0 mg S1 and 34.7 mg S2). Each fraction was dissolved in 3 mL methanol (HPLC grade).

For analytical HPLC-ESI-MS analysis, unprocessed CMTE sample (0.25 mg) was dissolved in 1 mL methanol (HPLC grade).

All solvents used were of analytical grade (Shanghai Chemical Reagents Company, Ltd.).

Analytical LC condition

Analytical LC-ESI-MS analyses of CMTE and fractions MAR1, MAR2, MAR3, S1, and S2 were performed on an Agilent 1200 Series LC and interfaced to an Agilent 6110 single-quadrupole mass spectrometer (Agilent Corporation, MA, USA). All samples were ionized in the electrospray ionization (ESI) and operated in positive mode. An Agilent Eclipse Plus C₁₈ column (4.6 × 100 mm, 3.5 μm) was applied for the analysis at a low flow rate of 0.4 mL min⁻¹, to avoid using a tee to split. The column oven was set at 30 °C. The mobile phase system consisted of 0.1% aqueous formic acid (phase A) and methanol (phase B). Samples were eluted by a linear gradient increasing from 50% to 100% phase B over 30 min. The re-equilibration time was 15 min. The injection volume was 5 μL. The system was controlled using the Agilent ChemStation software (rev. B.01.01).

Analytical LC-ELSD analyses for a purity check of the obtained compounds were performed on an Agilent 1260 Series LC (Agilent Corporation, MA, USA) coupled with a Sedex 75 ELSD system (Sedere, Alfortville, France), which was connected via the Agilent 35900E A/D converter. An Agilent Eclipse XDB C₁₈ column (4.6 × 250 mm, 5 μm) at temperature of 30 °C was applied for all the analyses, based on a flow rate of 0.8 mL min⁻¹. Samples were eluted by a linear gradient increasing from 50% to 100% phase B over 30 min. The re-equilibration time was 15 min. Sample injection volume was 10 μL. The usual ELSD settings were as follows: gain 7, drift tube temperature 40 °C, nebulizer gas pressure 3.6 bar. The system was controlled using the HP Chemstation software (Agilent Technologies, Palo Alto, CA, USA).

Prep-LC separation coupled online with ESI-MS analysis conditions

Mass-directed fraction collection was performed on an Agilent 1200 Series LC and interfaced to an Agilent 6110 single-quadrupole mass spectrometer equipped with an ESI source and active splitter, coupled with an analytical fraction collection system (Agilent Corporation, MA, USA). An Agilent Eclipse-XDB C₁₈ column (9.4 × 250 mm, 5 μm) at a temperature of 30 °C was applied for the separation, based on a flow rate of 3.0 mL min⁻¹. The mobile phase system consisted of 0.05% aqueous formic acid and methanol at a 53 : 47 (v/v) ratio. The injection volume was 25 μL and run time was 22 min. The solvents for the make-up flow were composed of 75% water, 25% methanol, and 0.1% formic acid, at a flow rate of 0.4 mL min⁻¹. When an active splitter was applied, the split-ratio was set at 100 : 1 (Fig. 2A1, B1 and B3); when it was not applied, it was replaced by a tee (Fig. 2A2). In the positive ESI mode, the MS parameters were set as follows: the capillary voltage was 3.0 kV; nitrogen was used to dry (350 °C, 12 L min⁻¹); and the nebulizer gas was 35 psig. The target fragment ion at m/z 387 for the three isomers was monitored in the selected ion monitoring (SIM) mode. The fragmentor voltage was 150 V. Parameters for mass-based fraction collection were optimized as follows: the delay time 0.08 min; threshold 50 000 counts; peak slope 500 counts; max. peak width 0.80 min. The system was controlled using the Agilent ChemStation software (rev. B.01.01).

Fig. 2 Instrumental set-up of the Agilent 1200 Series purification system for mass-based fraction collection (standard: A1 and tee-applied: A2); for (A1) and (A2), DAD could be removed or not. (B1) The active splitter with the plumbing connections on the front. SET buttons were indicated in the red circle; (B2) New (upper left) and worn (down left) rotor seals, and new (upper right) and worn (down right) stator face. (B3) Amplified image of the red circle in (B1).

NMR analysis

¹H and ¹³C NMR spectra for the obtained compounds were acquired on a Bruker Avance III (600, 150 MHz, CD₃OD, ppm relative to TMS) spectrometer.

Results and discussion

Selection of different preparation conditions

CMTE compromised a multitude of compounds which could be detected by MS, and accordingly, pre-purification of CMTE was strongly suggested. In light of the fact that purification by silica gel results in rather low overall yields, column chromatography methods using macroporous adsorbing resin D-101 and Sephadex LH-20 were envisioned as preferred purification steps. The established LC-ESI-MS method (0.1% aqueous formic acid and methanol gradient in 30 min) was employed for the analyses of the obtained fractions MAR1, MAR2, MAR3, S1, and S2, in order to check which fraction enriched the three isomeric steroids. Total ion chromatogram (TIC) results showed the complexity of CMTE (Fig. 3A). Three isomeric steroids (MW = 364) were obviously detected, based on an extracted ion of 387, because all compounds showed a strong sodium adduct ([M + Na]⁺). Extracted ion chromatogram (EIC) analysis for CMTE showed that the three steroids were eluted in the first 15 min. Indicated by the EICs at m/z 387 for CMTE, MAR1, MAR2, MAR3, S1, and S2, respectively (Fig. S1†), fraction S2 (34.7 mg, 11.57 mg mL⁻¹) was the optimal fraction which enriched the target compounds. Thus, the initial pre-purification step by Sephadex LH-20 column chromatography was an aid to LC-MS analysis, despite its extra use of time and solvent, because a fraction enriched with the three target compounds was yielded. As the increase in injection volume caused the base peak width in TIC to expand gradually,²¹ the optimal injection volume was set at 25 μL to avoid mass overloading.

Fig. 3 (A) TICs of CMTE, MAR1, MAR2, MAR3, S1, and S2. (B) TICs of S2 using methanol–0.05% aqueous formic acid mixture with different proportions (v/v) as the mobile phase: 55 : 45 (M1), 53 : 47 (M2), 51 : 49 (M3), 49 : 51 (M4), 47 : 53 (M5), 45 : 55 (M6). The targeted compounds (1, 2, and 3) were marked in the overlapped spectra. (C) MSD signal settings for fraction collection and TIC of S2 with fraction collection information. (D) Overlapped TICs of the re-analyzed obtained compounds (1, 2, and 3) by LC-ESI.

Principle of mass-based pre-LC operation and comparison of two purification systems

The standard purification system was set up as illustrated in Fig. 2A1. Since the MSD was a destructive detector, and to prevent it from overloading, a flow splitter was contained in a mass-based fraction collection system, by means of which the flow coming from the column was divided to the MSD and the fraction collector. The main flow (3 mL min⁻¹) led from the LC system to the active splitter, which was connected to the two flow paths: one to the detector and the other finally to the analytical fraction collector. Since the flow rate of the main flow (3 mL min⁻¹) was too high to route it directly into the ESI source, and as the flow from active splitter to the MSD was greatly reduced, a make-up flow (0.4 mL min⁻¹) led from the auxiliary pump (isocratic pump) directly to the MSD.²² The split ratio (100 : 1) was set according to the table provided with the equipment in which split factors corresponding to a particular LC flow rate and desired split ratio were listed,²³ in order to prevent the MSD from overloading but still ensure that it obtained signals with a desired signal to noise.³ Thus, 0.1% formic acid was added to the make-up flow to give a good sensitivity for MSD analysis.

The tee-applied purification system was set up as illustrated in Fig. 2A2. The main flow (3 mL min⁻¹) led from the LC system to a simple tee, which was also connected to the two flow paths: one to the detector (0.4 mL min⁻¹, the volume of the flow divided by the tested time gave the splitting flow rate); and the other (2.6 mL min⁻¹ accordingly) finally to the analytical fraction collector. The lengths of tubing connected with the tee and the ESI source (Fig. 2A2, red flow, Agilent red PEEK tubing, i.d. 0.13 mm, 40 cm) and tubing connected with the tee and the fraction collector (Fig. 2A2, blue flow, Agilent white tubing for fraction collection, i.d. 1.02 mm, 94 cm) were masdde suitable by experiment, in order to make sure that the desired compounds in the red flow arrived at the MSD before reaching the fraction collector in the blue flow. Furthermore, the restrictive length of the tubing was integral to avoid excessive backpressure in this configuration.

The key difference between the two systems above is the application of the active splitter (Fig. 2B1). Maintenance for the splitter consists of replacing the rotor seal and stator face assembly every 1 million cycles.²⁴ When the operating life of the assembly reaches its end, the active splitter cannot direct any flow to the MSD. Moreover, the price of the assembly is high. Comparison of new and worn assemblies showed a sharp contrast (Fig. 2B2). Even though the calculated split ratio of the tee was 6.5 : 1 (2.6/0.4), resulting in a larger sample loss than application of an active splitter, the tee-applied purification system was economic for compound purification. On the other hand, the operator was required to turn-on or turn-off manually the active splitter (Fig. 2B3), which highlights the convenience of that the tee-applied purification system. Finally, the tee-applied purification system was adopted for purification studies.

Optimization of LC-ESI-MS conditions

The optimum LC-ESI-MS conditions allow separation of the mixed components of sample S2 and provide useful information about the purity of the target compounds. Due to the complexity of traditional medicinal herbs, it is still a challenge to accomplish the isolation of three isomers on a C₁₈ column. Application of a chiral column was avoided, as the enantioselective chromatographic method required a two-step purification process and chiral columns were generally costly.³ Although mass-directed fractionation has been known to provide additional selectivity compared to analog detectors, as it could collect a relatively pure compound without baseline resolution (which is achieved by threshold settings on compound of interest and/or the close-eluting impurities),²² good resolution for the target compounds was a great concern in this study.

The TIC results showed that the ionization of the mixture components of CMTE was more efficient in the positive-ion mode than that in the negative-ion mode when the mobile phase was composed of a mixture of acetonitrile–0.1% aqueous formic acid,²⁵ so methanol–0.05% aqueous formic acid was initially applied as the mobile phase. It was found that isocratic elution resulted in better separation without baseline fluctuation than gradient elution, thus the mobile phase consisting of a methanol–0.05% aqueous formic acid mixture in different proportions (v/v) was tested (Fig. 3B). Finally, the separation of the target compounds (1, 2, and 3) from S2 was achieved when the proportion was 47 : 53 (Fig. 3C).

Optimization of fraction collection conditions

Fraction collection directed by mass is critical and necessary for the compound purification. The threshold, peak slope and maximum peak width were three important parameters to be investigated in the experiment. The full mass spectra for S2 revealed the predominant peaks at m/z 387 as sodiated molecular ions ([M + Na]⁺), and fraction collection was triggered on the MSD signal at m/z 387 with a threshold of 50 000 counts, a peak slope of 500 counts, and a maximum peak width of 0.80 min. To avoid collecting the non-target compounds which could be present, molecular ions at m/z 387 during the initial 10 min, the fraction collection time was set at 10 min. The green and red vertical lines indicted the beginning and end of fraction collection, respectively (Fig. 3C). Additionally, both positions of the collected fraction (vial numbers) in the fraction collector and the target mass were given in the spectrum. Exclusive collection of valid peaks was triggered when all conditions for fraction collection were fulfilled. At last, compounds 1 (3.0 mg), 2 (2.0 mg), and compound 3 (3.0 mg) were obtained from sample S2 (20.0 mg).

In addition, the silica gel column chromatography method was also employed. As described, the separation and purification of compounds by conventional manual column chromatography is not only time-consuming and tedious, but also uses chlorinated organic solvents which are environmental pollutants. Moreover, the yield with the silica gel column chromatography method was less than that of our method established in this paper, because the former technique caused adsorbing effects in the stationary phase, and three compounds were dispersed across all the obtained fractions.

Purity analysis and structural identification

Peak purity verification by reinjection of the enriched compounds to analytical LC-ESI-MS conditions was analyzed (Fig. 3D). The mass spectra indicated that purification was successful (Fig. S1, A1–A3†). Besides, LC-ELSD was ideal for analysis of these isomers since it can detect compounds with no chromophoric group. The purity of each compound was assessed to be greater than 90% by the established LC-ELSD method (Fig. S2, B1–B3†). Compared the obtained 1D NMR spectra data with those from the literature, compounds 1–3 were identified as 17β-tenacigenin B,²⁶ tenacigenin B,²⁶ and tenacigenin A,²⁷ respectively (NMR data S1).

Conclusions

Being an ideal method for purification of compounds in small amounts, especially for the compounds with no UV-visible chromophore, mass-based fraction collection eliminates identification the fraction of interest by another technique, such as previously used off-line mass spectrometry. When applied to gain three isomeric C₂₁ steroids from M. tenacissima, Sephadex LH-20 column chromatography was initially used as a supplementary technique to enrich the target compounds, and the mass-based pre-LC method was further applied to separate these compounds. Moreover, compared to the previous applied active splitter, the modified purification system via a tee was economic and convenient. These results showed that a combination of Sephadex LH-20 and the mass-based pre-LC method successfully purified three isomeric C₂₁ steroids. This feasible and reliable approach should be a valuable and potentially environment-friendly method for systematic isolation of bioactive and/or specified components from traditional medicinal herbs.

Acknowledgements

This work was supported by grants from National Natural Science Foundation of China (81202866). The authors greatly appreciate the valuable guidance and support from Customer Service Engineer Hu Zheng (Agilent Technologies Co., Ltd., Shanghai, China) for concerning the mass-based fraction collection method and the tee-application. The authors also would like to thank and acknowledge Xiao Wan and Chen Bo (Agilent, Shanghai) for their kind assistance.

Notes and references

1. O. Sticher, Natural product isolation, Nat. Prod. Rep., 2008, 25, 517–554.
2. W. Z. Yang, M. Ye, X. Qiao, C. F. Liu, W. J. Miao, T. Bo, H. Y. Tao and D. A. Guo, Anal. Chim. Acta, 2012, 739, 56–66.
3. L. Zeng, R. Xu, Y. Zhang and D. B. Kassel, J. Chromatogr., A, 2011, 1218, 3080–3088.
4. Jiangsu New College of Medicine, A Dictionary of Traditional Chinese Drugs, Shanghai Science and Technology Press, Shanghai, 1977, p. 1976.
5. S. Shobha, P. K. Maheshwari and K. Anakshi, Phytochemistry, 1980, 19, 2427–2430.
6. M. Shuji, Y. Kimiko, H. Koji, K. Koh and M. Hiroshi, Phytochemistry, 1986, 25, 2861–2865.
7. S. Q. Luo, L. Z. Lin, G. A. Cordell, L. Xue and M. E. Johnson, Phytochemistry, 1993, 34, 1615–1620.
8. Z. H. Xia, W. X. Xing, S. L. Mao, A. N. Lao, J. Uzawa, S. Yoshida and Y. Fujimoto, J. Asian Nat. Prod. Res., 2004, 6, 79–85.
9. J. Deng, Z. X. Liao and D. F. Chen, Phytochemistry, 2005, 66, 1040–1051.
10. X. L. Wang, Q. F. Li, K. B. Yu, S. L. Peng, Y. Zhou and L. S. Ding, Helv. Chim. Acta, 2006, 89, 2738–2744.
11. H. Zhang, A. M. Tan, A. Y. Zhang, R. Chen, S. B. Yang and X. Huang, Steroids, 2010, 75, 176–183.
12. Y. J. Hu, X. L. Shen, H. L. Lu, Y. H. Zhang, X. A. Huang, L. C. Fu and W. F. Fong, J. Nat. Prod., 2008, 71, 1049–1051.
13. Y. Q. Fang and X. M. Sun, Chin. J. Biochem. Pharm., 2011, 32, 165–167.
14. S. Y. Han, H. Y. Zhao, N. Zhou, F. Zhou and P. P. Li, J. Ethnopharmacol., 2014, 151, 210–217.
15. Z. Q. Huang, H. Tan, C. Y. Wang, H. Z. Zhang, D. Liu, C. Z. Zhou and X. Liu, Chin. J. Clin. Oncol., 2007, 12, 97–99.
16. W. Y. Wang, Y. Zhou, X. J. Zhang, T. H. Gao, Z. F. Luo and M. Y. Liu, Chin. J. Clin. Oncol., 2009, 14, 936–938.
17. K. Wang and Z. X. Guo, J. Basic Clin. Oncol., 2009, 22, 47–48.
18. S. M. Wang and H. P. Zhao, Chin. J Clin. Oncol., 2011, 38, 1135–1138.
19. S. Keyoumu, L. Y. Ma and Y. Tang, J. Basic Clin. Oncol., 2012, 25, 397–399.
20. Z. R. Huang, H. Lin, Y. Wang, Z. Cao, W. Lin and Q. Chen, Oncol. Lett., 2013, 5, 917–922.
21. X. Zhang, W. Picariello, N. Hosein, M. Towle and W. Goetzinger, J. Chromatogr., A, 2006, 1119, 147–155.
22. R. Moritz and N. O'Reilly, J. Biomol. Tech., 2003, 14, 136–142.
23. G1968E Kit Installation Instruction. Agilent 6100 Series Single Quad LC/MS Systems, p.6.
24. G1968E Kit Installation Instruction. Agilent 6100 Series Single Quad LC/MS Systems, p. 15.
25. S. Y. Han, M. B. Zhao, G. B. Zhuang and P. P. Li, Lung Cancer, 2012, 75, 30–37.
26. J. Deng, Z. X. Liao and D. F. Chen, Chin. Chem. Lett., 2005, 16, 487–490.
27. S. Q. Luo, L. Z. Lin, G. A. Cordell, L. Xue and M. E. Johnson, Magn. Reson. Chem., 1993, 31, 215–221.

---

Footnote

† Electronic supplementary information (ESI) available: Delay volume calibration, extraction and isolation, NMR data for compounds 1–3, Fig. S1 and Fig. S2 were available online. See DOI: 10.1039/c3ra45723k

---