Search tips
Search criteria 


Logo of chromsciLink to Publisher's site
J Chromatogr Sci. 2016 April; 54(4): 664–669.
Published online 2016 February 9. doi:  10.1093/chromsci/bmw007
PMCID: PMC4885409

Determination of Five Major 8-Prenylflavones in Leaves of Epimedium by Solid-Phase Extraction Coupled with Capillary Electrophoresis


A simple, accurate and reproducible method which is based on the capillary electrophoresis, coupled with solid-phase extraction, has been developed for simultaneous determination of multiple 8-prenylflavones from Chinese Herba Epimedii. In this study, the author has mainly illustrated the experimental process and research results of five major components including epimedin C, icariin, diphylloside A, epimedoside A and icarisoside A that have been extracted and identified from Herba Epimedii for the first time. Experimental conditions have been optimized to achieve the best separation efficiency for the following factors: the buffer pH, buffer concentration and applied voltage. The experiment can be conducted through two separable stages: the first stage is to obtain the crude extracts through the solid-phase extraction; and the second stage is to further separate five major components by using the capillary electrophoresis. The separation of the five components and the analysis of the experiment are relatively fast and can be completed within 20 min. The concentration ranges of the construction of standard curves of five major 8-prenylflavones are 32.0–395.0, 23.4–292.0, 42.1–526.0, 18.8–233.5 and 29.7–371.0 µg mL−1 respectively, which have showed acceptable linearity with a correlation coefficient, r ≥ 0.999. The coefficient varies within 2.0% for both intra- and inter-days tests. The recoveries of five components range from 92.3 to 104.1%. The relative standard deviations of recoveries of five components range from 1.2 and 2.8%. This new method will facilitate the extraction and expedite the determination of medical components from Herba Epimedii.


Herba Epimedii (family Berberidaceae, Yin-Yang-Huo), used for medical practice in China over 2000 years, is capable of enhancing the kidney functions, which is expressed in traditional Chinese medicine theory as strengthening the Yang of human body (1, 2). More recent research studies suggested that flavonoids, as the major active components of Herba Epimedii, have shown multiple beneficial therapeutic effects in vitro, such as the anti-inflammatory effects on the mouse' ear edema and the protection of human umbilical vein endothelial cells against H2O2 damage (39). Among over 15 flavonoids identified from Herba Epimedii and other related species (1013), diphylloside A, icariin, epimedin C, epimedoside A and icarisoside A have been identified as the main flavonoids (14). One challenge in the processing and preservation of Herba Epimedii is the poor stability of flavonoids, presumably due to their 8-isopentenyl structure (Figure 1). A reliable method is in great demand for the rapid, facile and sensitive identification and quantification of flavonoids, particularly for the development of Herba Epimedii-related medicines. Previous reports of the identification of flavonoids from Herba Epimedii focused on the methods based on HPLC (1518). Unlike HPLC, the methods based on high performance capillary electrophoresis (HPCE) require less sample volume and low maintenance of the electrophoresis column and produce much lesssolvent waste (1922). Although HPCE has been reported for the determination of icariin, epimedin A, epimedin B and epimedin C (20, 21), simultaneous analysis of the multiple prenylflavones (epimedin C, icariin, diphylloside A, epimedoside A and icarisoside A) has not been documented. In this article, we propose a HPCE method for simultaneous determination of five major 8-isopentenyl-flavones in the leaves of ten varieties of Herba Epimedii.

Figure 1.
Structures (peak number) of the five flavonoids: 1: epimedin C, 2: icariin, 3: diphylloside A, 4: epimedoside A and 5: icarisoside A.

It is difficult to accurately control and rapidly determine the extraction with capillary electrophoresis (CE). Usually, pretreatment is often necessary for herbs extraction. Liquid–liquid extraction, high-speed counter-current and solid-phase extraction are common pretreatment methods. Because solid-phase extraction has the characteristics of high selectivity, high automation and less organic solvent consumption and it is much more simple and efficient than other methods, it is very often used as a sample pretreatment method. In order to effectively eliminate the impurities interference of the samples on the CE and achieve a higher purity of crude extracts, we used solid-phase ultrasonic extraction to enrich the chemical components of flavonoids as a sample pretreatment for further HPCE analysis.

Experimental section

Materials and reagents

Purified flavonids used as standard chemicals

Five flavonoids (epimedin C, icariin, diphylloside A, epimedoside A and icarisoside A) which are isolated from the roots of Epimedium wushanense T. S. Ying were used as standards in this work. The structure and purity (>98%) of these five flavonoids were characterized using UV-Vis, 1H and 13C NMR, MS and HPLC.

Acquisition of Herba Epimedii and sample preparation

Ten batches of Herba Epimedii leaves were purchased from six cities of China between October 2008 and November 2009. Epimedium wushanense T. S. Ying. was collected from Zhenping county, Ankang city, Shannxi province. The fresh leaves were crushed and dried at 60°C immediately after acquisition. A Lichrolut® RP-18 solid-phase extraction column SPE (Merck, made in Germany) was used for the sample preparation.


HPLC grade methanol was purchased from Tianjin Chemical Factory (Tianjin, China). Water used in this experiment was purified through a Milli-Q plus system (Millipore, Madrid, Spain) and further filtered through a 0.22 µm filter membrane. Analytical grade sodium tetraborate was purchased from Damao Chemical Factory (Tianjing, China). The pH of the running buffer was adjusted by using 0.2 M HCl or 1 M NaOH.

HPCE system

The analysis was carried out on a Beckman MDQ system which is equipped with a diode array detector, a fluid-cooled column cartridge, an automatic injector and an uncoated fused-silica capillary (Beckman) of 65.0 cm × 50.0 µm I.D. (effective length 50 cm). The experiment was completed under the condition of applied voltage 18 kV by using normal polarity. Sample injection was carried out in a hydrodynamic mode for 10 s with a pressure of 20 psi. The capillary temperature was maintained at 25°C and the detector was set at 270 nm. Borate buffer (30 mM, pH 9.5) containing 40% methanol was used as a running buffer. When a new capillary was used, the capillary was rinsed with 1 M hydrochloric acid for 5 min, water for 5 min, then 0.1 M sodium hydroxide for 10 min, water for 5 min and buffer for 10 min prior to the first analysis. Between each run, the capillary was rinsed with 0.1 M sodium hydroxide for 3 min, water for 2 min and buffer for 3 min successively. The Gold software for system control and data processing was used in this experiment.

Preparation of standard solutions

For each flavonoid, six standard solutions were prepared by using methanol–water (50 : 50, v/v) within the concentration range as specified below: 32.0–395, 23.4–292, 42.1–526, 18.8–233.5, 29.7–371 µg mL−1 for epimedin C, icariin, diphylloside A, epimedoside A and icarisoside A, respectively. To each solution, an internal standard (IS) (rutin) was added to achieve a final concentration of 55 µg mL−1 of rutin. All solutions were found to be stable for at least 1 month when stored at 5°C.

Sample preparation

The crushed Herba Epimedii leaves were purified through No. 2 sieve. About 0.3 g of commercial samples were weighed into a 100-mL flask and extracted by ultrasonic treatment with methanol–water (50 : 50, v/v, 30 mL) for 40 min, and then methanol was recovered from the sample. In order to achieve a higher purity of crude extracts with less impurities that may interfere the HPCE analysis, we used solid-phase extraction to enrich the chemical components of flavonoids as pretreatment before further HPCE analysis.

The solid-phase extraction column (Lichrolut®RP-18) was activated before using with 3.0 mL methanol and 2.0 mL water at a rate of 0.5 mL/min. The sample was added to this column, followed by washing with 2.0 mL water and 3.0 mL methanol. The water eluent was discarded and the follow-up methanol eluate was collected, filtered and diluted to a final volume of 30 mL with methanol–water (50 : 50, v/v) as the sample. Then, 5 mL of the sample was taken and 5.00 mL of the IS solution (rutin) was accurately added and adjusted to 10 mL. The mixture was shaken up, centrifuged and filtered through a 0.22-µm filter before injection, respectively.


Analytical conditions

The detection wavelength is set to 270 nm because it is the maximum absorption band of epimedin C, icariin, diphylloside A, epimedoside A and icarisoside A. Rutin is used as an IS because its HPCE signal does not overlap with any of the five flavonoids. The first separation parameter is the buffer pH, which plays a dual role in CE because it influences the generation of the magnitude of the electro-osmotic flow (EOF) and the amount of ionizable species. Flavonoids and boric acid anions could form negatively charged complexes when the pH value of the buffer solution is higher than 7.0, and the formation of such complexes is favored at higher pH. High pH, however, will cause longer separation time and larger EOF, and the peak resolution changed worse with the change of pH. Therefore, the optimal pH conditions need to be optimized in the experiment. A borate system with different pH values (8.6–11.6) is tested (Figure 2A). Although a reasonable separation between icariin, diphylloside A, epimedoside A and icarisoside A has been achieved in the pH range of 8.6–9.0, epimedin C and icariin have shown severe overlap in the electropherogram. Good separation is discovered between five major prenylflavones in the pH range 10.1–11.6 with the migration time of icarisoside A extended to 33.3–46.8 min. In this investigation, pH 9.5 is selected.

Figure 2.
Effect of pH (A), buffer concentration (B) and separation voltage (C) on migration time.

Buffer concentration is another key factor for CE separation. Higher buffer concentrations provide an enhanced ionic strength for better column efficiency and resolution. The influence of buffer concentration (10–50 mM) on migration time and resolution was investigated (Figure 2B). The migration time and resolution get an increase when the concentration of borate is increased. Epimedin C and icariin are not separated in 10 mM buffer, and the run time is prolonged in 50 mM buffer. Finally, the concentration of 30 mM buffer is adopted which results in good migration time and resolution.

Method validation

Extraction recoveries with solid-phase extraction

A testing solution was prepared by mixing the standard solutions in methanol–water (50 : 50, v/v) to reach the final concentrations as specified as follows: epimedin C 160 µg mL−1, icariin 117 µg mL−1, diphylloside A 210.5 µg mL−1, epimedoside A 94 µg mL−1 and icarisoside A 148.5 µg mL−1. The final concentration of IS (rutin) in this solution was 55 µg mL−1. The sample was treated with activated solid-phase column before determination, and we got the peaks of five prenylflavones. Additionally, the same standard solution was analyzed by CE directly. Each sample which had been treated by solid-phase extraction was repeatedly measured five times and calculated as follows: the recovery = A1 (peak area of extraction)/A2 (peak area of no extraction). Five reference standard recoveries of solid-phase extraction were calculated as shown in Table I. The results have shown that solid-phase extraction recoveries are 84.6–92.1%, and the relative standard deviation is <3.0%. The SPE conditions are appropriate to the determination of real samples.

Table I.
Extraction Recoveries of Five Prenylflavones in Epimedium (n= 5)

Calibrate graphs

The peak area ratio, y, and quality ratio, x were obtained in the range of 32.0–395.0, 23.4–292.0, 42.1–526.0, 18.8–233.5, 29.7–371.0 µg mL−1 for epimedin C, icariin, diphylloside A, epimedoside A and icarisoside A, respectively. The regression equations of the five curves and their correlation coefficients were calculated as follows: epimedin C, y = 0.2172x + 0.3542 (r = 0.9996); icariin, y = 0.2066x + 0.8994 (r = 0.9991); diphylloside A, y = 0.2639x + 1.2661 (r = 0.9992); epimedoside A, y = 0.4793x + 0.7953 (r = 0.9990); and icarisoside A, y = 0.6878x + 2.0742 (r = 0.9993). The signal which is three times higher than the peak noise height is taken as the detection limit (limit of detection, LOD).

The detection limits of five prenylflavones are 3, 2, 4, 2 and 3 µg mL−1 for epimedin C, icariin, diphylloside A, epimedoside A and icarisoside A, respectively. The precision of these methods was evaluated for five times on the same day, and a 5-day period analysis was carried out by injected standard solutions containing epimedin C, icariin, diphylloside A, epimedoside A and icarisoside A, The coefficient variations of intra- and inter-days were both <2.0% (n = 5), indicating satisfactory precision of the experimental data. The extraction recovery was tested by adding known amounts of five components. The recoveries of five components ranged from 92.3 to 104.1%. The relative standard deviation (RSD) of recoveries of five components ranged from 1.2 and 2.8%.

Sample analysis

The contents of the five 8-prenylflavones in samples of Herba Epimedii were determined by using above-mentioned the optimized CE condition. The electropherograms of the standards and the extract of Herba Epimedii. (e.g., in E. wushanense Ying) are shown in Figure 3A and B. The contents of five components in different samples are shown in Table II, which show that the contents of five components in all commercial samples of Herba Epimedii vary significantly. The dominant flavonoids were found to be epimediin C and diphylloside A in some samples (e.g., in E. wushanense Ying), and icariin in others samples. Because icariin is the key component used for the evaluation of Herba Epimedii quality (2), our experimental results suggest that this traditional method exclusively basing on icariin could cause highly biased results.

Table II.
Result of contents determination of different samples (%)
Figure 3.
Electropherograms and the migration time of a mixture of five 8-8-prenylflavones (A) and extract of Epimedii (B): 1: epimedin C, 2: icariin, 3: diphylloside A, 4: epimedoside A, 5: icarisoside A; IS: rutin.


In capillary electrophoresis, the increase of separation voltage can shorten the separation time, but high-voltage induced Joule heat can intensify the diffusion of solutes and further reduce the separation efficiency. In this experiment, the separation efficiency was investigated with the voltage ranging from 13 to 20 kV. As shown in Figure 2C, shorter migration time was observed when separation voltage was increased. The separation voltage of 18 kV was selected as the operation voltage for this experiment.

The influences of different buffers [citric acid–sodium citrate buffer, phosphate buffer (with surfactants SDS)] were also tested at different concentrations (10–50 mM), but the resolution of epimedin C and icariin was not improved and the migration time of icarisoside A was extended.

The EOF is affected by the amount of the organic solvent in the mobile phase. Methanol was used as the organic solvent in this work due to the excellent solubility of prenylflavones in methanol. The methanol concentrations were optimized to study its influence on the separation resolution and the HPCE current. The results showed that low concentrations (10%) of methanol reduced the current very slightly, but high concentrations (50%) caused the increased elution time. The optimized condition was found to be 40% methanol with the current of ~90 mA, which resulted in good resolution (>1.5) of epimedin C and icariin.


In summary, we presented a HPCE method for facile, accurate and repeatable analysis of flavonoids in Herba Epimedii. The pretreatment of samples by using solid-phase extraction helped to minimize the negative influences caused by the undesired impurities during HPCE analysis. This method was demonstrated by the separation of five flavonoids from the leaves of Epimedium with reasonable analysis time and excellent separation resolution.

Compared with HPLC, this method offers higher separation efficiencies, quicker analysis and lower buffer cost for the rapid and simultaneous evaluation of the multiple prenylflavones in Herba Epimedium.


This research was supported by the Biomedicine Key Laboratory of Shaanxi Province, Northwest University (Xi'an, Shaan'xi, China).


We really appreciate the support from the Drug Enforcement Administration in Ankang (Shaanxi, China) and Luoyang Land-flora Institute of Natural Plants (Henan, China).


1. Li S.Z.; Compendium of materia medica, Vol. 12 Foreign Language Press, Beijing, (2003), pp. 467.
2. National Pharmacopoeia Committee. Pharmacopoeia of People's Republic of China, part 1. Chemical Industry Press, Beijing, (2010), pp. 306.
3. Zhang Y.F., Yu Q.H.; The anti-inflammatory effects of total flavonoids of Epimedium; Journal of Shenyang Pharmaceutical University, (1999); 2: 122–124.
4. Zhang X.Y., Li Y.J., Yang X.D., Wang K., Ni J.Z., Qu X.G.; Inhibitory effect of Epimedium extracts on the activity of S-adenosyl-l-homocysteine hydrolase and biomethylation; Life Sciences, (2005); 78: 180–186. [PubMed]
5. Meng F.H., Li Y.B., Xiong Z.L., Jiang Z.M., Li F.M.; Osteoblastic proliferative activity of Epimedium brevicornum Maxim; Phytomedicine, (2005); 12: 189–193. [PubMed]
6. Zhang G., Qin L., Hung W.Y., Shi Y.Y., Leung P.C., Yeung H.Y. et al. ; Epimedium brevicornum Maxim prevent OVX-induce dosteoporo-sisinratsin dependent of its enhancement in intestinal calciumab-sorption; Bone, (2006); 38: 818–825. [PubMed]
7. Wang Y.K., Huang Z.Q.; Protective effects of icariin on human umbilical vein endothelial cell injury induced by H2O2 in vitro; Pharmacological Research, (2005); 2: 174–182. [PubMed]
8. Pan Y., Kong L.D., Xia X., Zhang W.Y., Xia Z.H., Jiang F.X.; Antidepressant-like effect of icariin and its possible mechanism in mice; Pharmacology Biochemistry and Behavior, (2005); 4: 686–694. [PubMed]
9. Liu T.Z., Chen C.Y., Yiin S.J., Chen C.H., Cheng J.T., Shih M.K. et al. ; Molecular mechanism of cell cycle blockage of hepatoma SK-Hep-1 cells by Epimedin C through suppression of mitogen-activated protein kinase activation and increased expression of CDK inhibitors p21(Cip1) and p27(Kip1); Food and Chemical Toxicology, (2006); 44: 227–235. [PubMed]
10. Li W.K., Zhuang R.Y., Xiao P.G.; Flavonoids from Epimedium wanshanense; Phytochemistry, (1996); 2: 527–530. [PubMed]
11. Guo B.L., Li W.K.; Brevicornin, a flavonol from Epimedium brevicornum; Phytochemistry, (1995); 3: 991–992.
12. Yao C.S., Zhuang X.H.; New flavonol glycosides from Epimedium sagittatum; Natural Product Research and Development, (2004); 2: 101–103.
13. Huang Y.L., Qu J.C., Chen C.C.; Three new 2-phenoxychromones from the leaves of Epimedium sagittatum; Journal of Natural Products, (1993); 2: 275–278.
14. Wang J.X., Fan X.W., Niu J.F.; Chemical constituents of Epimedium platyetalum; Acta Botanica Sinica, (2002); 10: 1258–1260.
15. Zhou Y.X., Zhang H.R., Fan D.D.; HPLC analysis of different parts in Epimedium wushanense from different regions; Chemical Engineering (China), (2013); 2: 6–8.
16. Guo B.L., Xiao P.G.; Flavonoid composition analysis from Epimedium of Sichuan produce; West China Journal of Pharmaceutical Sciences, (1996); 2: 75–77.
17. Chen X.J., Guo B.L., Li S.P., Zhang Q.W., Tu P.F., Wang Y.T.; Simultaneous determination of 15 flavonoids in Epimedium using pressurized liquid extraction and high-performance liquid chromatography; Journal of Chromatography A, (2007); 1163: 96–104. [PubMed]
18. Pozharitskaya O.N., Kosman V.M., Shikov A.N., Demchenko D.V.; Comparison between HPLC and HPTLC densitometry for the determination of icariin from Epimedium koreanum extracts; Journal of Separation Science, (2007); 30: 708–712. [PubMed]
19. Zhao H.Y., Sun J.H., Fan M.X., Fan L., Zhou L., Li Z. et al. ; Analysis of phenolic compounds in Epimedium plants using liquid chromatography coupled with electrospray ionization mass spectrometry; Journal of Chromatography A, (2008); 1190: 157–181. [PubMed]
20. de la Iglesia P., Gago-Martineza A., Yasumoto T.; Advanced studies for the application of high-performance capillary electrophoresis for the analysis of yessotoxin and 45-hydroxyyessotoxin; Journal of Chromatography A, (2007); 1156: 160–166. [PubMed]
21. Akiyama T., Yamada T., Maitani T.; Analysis of enzymatically glucosylated flavonoids by capillary electrophoresis; Journal of Chromatography A, (2000); 895: 279–283. [PubMed]
22. Maccari F., Tripodi F., Volpi N.; High-performance capillary electrophoresis separation of hyaluronan oligosaccharides produced by Streptomyces hyalurolyticus hyaluronate lyase; Carbohydrate Polymers, (2004); 56: 55–63.

Articles from Journal of Chromatographic Science are provided here courtesy of Oxford University Press