Article. Jiliang Cao 1, Jinchao Wei 1, Cheng Xiang 2, Mi Zhang 2, Baocai Li 2, Jianbo Wan 1, Huanxing Su 1, and Peng Li 1, * Abstract.

Transcription

1 Journal of Chromatographic Science, 2016, Vol. 54, No. 8, doi: /chromsci/bmw059 Advance Access Publication Date: 2 May 2016 Article Separation and Determination of Four Tanshinones in Danshen and Related Medicinal Plants by Micellar Electrokinetic Chromatography Using Ionic Liquids as Modifier Jiliang Cao 1, Jinchao Wei 1, Cheng Xiang 2, Mi Zhang 2, Baocai Li 2, Jianbo Wan 1, Huanxing Su 1, and Peng Li 1, * 1 State Key Laboratory of Quality Research in Chinese Medicine, Institute of Chinese Medical Sciences, University of Macau, Macau , China, and 2 Faculty of Life Science and Technology, Kunming University of Science and Technology, Kunming , China *Author to whom correspondence should be addressed. pli1978@hotmail.com Received 7 July 2015; Revised 18 February 2016 Abstract A simple and fast micellar electrokinetic chromatography (MEKC) method using ionic liquids as modifier was established for simultaneous determination of four hydrophobic tanshinones, including dihydrotanshinone I (1), cryptotanshinone (2), tanshinone I (3) and tanshinone IIA (4), in Danshen and related medicinal plants. In normal MEKC using sodium dodecyl sulfate (SDS) as surfactant and organic solvents as additives, the four tanshinones, especially cryptotanshinone and tanshinone I, could not be well separated. Fortunately, further addition of ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate ([bmim]bf 4 ) resulted in a baseline separation of these four analytes. After an optimization study, 10 mm borax 10 mm SDS 10 mm [bmim]bf 4 containing 15% acetonitrile (v/v) at ph 9.6 was adopted as the running buffer to complete the separation within 16 min at the voltage of 25 kv, temperature of 25 C and detection wavelength of 254 nm. The relative standard deviations of migration time and peak area were in the range of and %, respectively, indicating the good repeatability of the developed method. This method was extensively validated by evaluating the linearity (R ), limits of detection ( μg ml 1 ), limits of quantification ( μg ml 1 ) and recovery ( %). Under the optimum conditions, samples of Danshen and related medicinal plants were well analyzed with high separation efficiency. Introduction Danshen, the dried root and rhizome of Salvia miltiorrhiza, is one of the most widely used traditional Chinese herbs in China with the properties of promoting blood circulation and removing blood stasis (1). Modern pharmacological studies have revealed its beneficial effects on cardio-cerebrovascular diseases, hepatitis, hepatocirrhosis, chronic renal failure, dysmenorrhea, neuroasthenic insomnia, and antitumor activity (2 5). Besides S. miltiorrhiza, the unique and official origin of Danshen in Chinese Pharmacopoeia (6), several other species of Salvia genus such as S. przewalskii, S. castanea and S. yunnanensis are also widely used as Danshen in some areas of China (1, 7). As a major type of bioactive components in Danshen and some related medicinal plants, tanshinones, especially dihydrotanshinone I (8, 9), cryptotanshinone (10, 11), tanshinone I (12, 13) and tanshinone IIA (14 16) (structures shown in Figure 1) are usually chosen as the chemical markers for the quality control of Danshen and its preparations. Some analytical techniques including HPLC (17 20), UPLC (21, 22) or LC MS (23, 24) have been applied to the qualitative and quantitative analysis of tanshinones in Danshen. Although these The Author Published by Oxford University Press. All rights reserved. For Permissions, please journals.permissions@oup.com 1435

2 1436 Cao et al. Figure 1. The chemical structures of four investigated tanshinones: (1) dihydrotanshinone I, (2) cryptotanshinone, (3) tanshinone I and (4) tanshinone IIA. methods could easily obtain satisfactory repeatability and high sensitivity, they have suffered from a variety of disadvantages, especially the large amount of organic solvents consuming. Recently, as an effective and economical technique with the advantages of short analysis time, high resolution and separation efficiency, minimal consumption of sample and solvents (25), capillary electrophoresis (CE) has been successfully developed for the determination of tanshinones in Danshen. Due to the high hydrophobicity of tanshinones, only a few of CE modes have been employed for the determination of these compounds, including nonaqueous CE (26, 27), microemulsion electrokinetic chromatography (28) and capillary electrochromatography (29). However, when analyzing the tanshinones, these methods were still associated with the problems of inefficient separation, relative instability of separation system or extreme fragility of packed capillaries. Micellar electrokinetic chromatography (MEKC), an important branch of CE, has attracted much attention in the analysis of phytochemical components in recent years because of its special ability for the separation of neutral and hydrophobic compounds with high selectivity and micellar system stability. In MEKC, a surfactant is added to the buffer solution above the critical micelle concentration to act as pseudo-stationary phase (PSP). Analytes, both charged and neutral, can be separated from each other on the basis of their differential partitioning between the PSP and the aqueous buffer phase. Sodium dodecyl sulfate (SDS) is the most commonly used surfactant in MEKC and the inner core of the SDS micelles is strongly hydrophobic. Therefore, when using SDS as the surfactant, the separation of highly hydrophobic analytes usually lacks of selectivity because all the components tend to be completely absorbed into the micelles and migrate with the velocity of the micelles (30). In such case, the addition of cyclodextrins (31), organic solvents, such as MeOH (32), acetonitrile (ACN) (33) or EtOH (34), as well ionic liquids (ILs) (35 37) as buffer modifiers would be useful to offer different selectivity. Especially, ILs provide an interesting alternative or supplement to organic modifiers in MEKC to improve the separation selectivity of hydrophobic compounds, as they can be used at high concentrations without current breakdowns or destruction of the micellar system. Recently, we have reported a MEKC method with the buffer containing sodium tetraborate, SDS, β-cyclodextrins (β-cd), IL 1-butyl- 3-methylimidazolium tetrafluoroborate ([bmim]bf 4 ) and ACN for the separation of seven hydrophilic and four hydrophobic active components in the root of three Salvia species (38). Yet, the buffer system was a little complex resulting to its relative instability and difficult preparation. Moreover, the mechanisms of improved selectivity for hydrophobic tanshinones were still necessary to be deeply studied after the addition of ILs in MEKC. Considering these reasons, in this work, we developed a simple and rapid MEKC method using modifier of [bmim]bf 4 to separate the four hydrophobic tanshinones, including dihydrotanshinone I (1), cryptotanshinone (2), tanshinone I (3) and tanshinone IIA (4). More importantly, the possible separation mechanisms of IL [bmim]bf 4 in MEKC were studied and discussed in this article. Experimental Chemicals and materials Cryptotanshinone and tanshinone IIA (98%) were purchased from National Institutes for Food and Drug Control (China). Tanshinone

3 Separation and Determination of Four Tanshinones in Danshen 1437 I and dihydrotanshinone I ( 96%) were obtained from Shanghai Tauto Biotech Co., Ltd (Shanghai, China). Borax was supplied by Tianjin Damao Chemical Reagent Factory (Tianjin, China) and SDS was purchased from Beyotime Institute of Biotechnology (Jiangsu, China). The IL [bmim]bf 4 was the product of Aladdin Reagent Co., Ltd (Shanghai, China). HPLC-grade ACN was obtained from Merck (Darmstadt, Germany) and deionized water was prepared by a Milli-Q water purification system (MA, USA). All other reagents and chemicals were of analytical grade. Three Salvia species, including S. przewalskii (Sample 1), S. castanea (Sample 2) and S. miltiorrhiza (Samples 3 7), were collected from different regions of China. All the samples were authenticated by Prof. Shimin Guo at Yunnan Institute of Traditional Chinese Medicine and Medical Materials, Kunming, China. Apparatus A Beckman P/ACE MDQ CE system (Beckman Coulter, Fullerton, CA, USA) equipped with an autosampler and a photodiode array detector, and a fused silica capillary (70 cm 75 μm i.d., 50 cm effective length) were used throughout this study. A SevenEasy ph-meter (Mettler Toledo, Greifensee, Switzerland) was used to adjust the ph value of background electrolyte (BGE). Both sample extraction and solution degassing were performed using an ultrasonic cleaning bath (Branson Ultrasonic Corp., Danbury, CT, USA). The hydrodynamic size and ζ potential of SDS micelles were determined by dynamic light scattering using a Nano ZSP Zetasizer (Malvern Instruments, Malvern, UK). Pyrene fluorescence spectra were recorded on a Lumina Fluorescence Spectrometer (Thermo Scientific, USA). Electrophoretic procedure The MEKC experiment was carried out at the separation voltage of 25 kv using a 10 mm borax 10 mm SDS 10 mm [bmim]bf 4 containing 15% ACN (v/v) (ph 9.6) as running buffer. The temperature of the capillary column was set at 25 C and samples were injected under the pressure of 0.3 psi for 2 s. The detection was performed at 254 nm. Before each experiment, the capillary was sequentially rinsed with 0.1 M NaOH for 2 min, purified water for 2 min and finally running buffer for 5 min. All the BGE solutions were filtered through a 0.45 μm membrane and degassed by ultrasonication before use. Sample preparation An aliquot of 1.00 g root powder of Salvia plants (through 50 mesh screen) was accurately weighed and extracted with 25 ml of MeOH for 30 min by an ultrasonic cleaning bath. After adjusted to its original weight, the extract was centrifuged at 4,500 r.p.m. for 15 min and the supernatant was filtered through a 0.22 μm membrane (PVDF Millex-GV, 13 mm, Millipore) prior to injection into the CE system. Micellar inner polarity measurement Pyrene was used as the fluorescent probe to investigate the micellar inner polarity (39). Micellar solutions with different concentrations of [bmim]bf 4 were prepared by adding appropriate contents of borax, SDS, ACN and/or [bmim]bf 4, and then diluted with saturated pyrene solution. The mixtures were ultrasonicated for 15 min and kept overnight. Pyrene fluorescence was measured at excitation wavelength of 335 nm with excitation slit width set at 10 nm and emission at 5 nm. The intensity ratio of the first emission peak (I 1 = 373 nm) to the third emission peak (I 3 = 384 nm) was used to measure the inner polarity of SDS micelles (40). The larger ratio (I 1 /I 3 ) could reflect the stronger environment polarity in the inner core of SDS micelles. Results Initially, ordinary capillary zone electrophoresis mode was tested using different buffer compositions such as borate buffer and phosphate buffer. However, due to the four tanshinones could not be charged in the buffers, all the analytes had no electrophoretic mobilities and migrated with the electroosmotic flow (EOF) (Figure 2A). Subsequently, MEKC mode was taken into account since it could always achieve the separation of both charged and neutral analytes. Under normal MEKC conditions without any modifiers (concentration of SDS mm, concentration of borax mm and ph ), no successful separation of the four analytes was obtained (Figure 2B). The poor resolution was possibly attributed to the high hydrophobicity of the investigated analytes, which tended to be absorbed virtually completely into the inner core of SDS micelles. So, modifiers were further added to improve the selectivity of the MEKC system, including β-cd and different organic solvents. After added β-cd, the four analytes just exhibited two partially overlapping peaks and no better separation could be achieved even with varied concentrations of β-cd (as shown in Figure 2C). Organic solvents with various types (MeOH, EtOH, ACN and isopropanol) and different ratios (5 20%) were also tested and similar results were obtained. As indicated in Figure 2D, cryptotanshinone and tanshinone I were co-eluted as one peak. As is known, the separation mechanism of analytes in MEKC is based on differential partitioning between the micellar stationary phase and the aqueous phase. Compared with dihydrotanshinone I and tanshinone IIA, cryptotanshinone and tanshinone I might have more similar partitioning coefficients in the micellar phase, resulting in the co-elution peak of the two analytes. Then, the IL [bmim]bf 4 was added into the BGE to enhance the separation selectivity. Fortunately, the four hydrophobic analytes with similar structures were completely resolved with baseline separation. Therefore, [bmim]bf 4 was selected as a modifier of MEKC and further optimization was needed in the following study. Discussion Method optimization In order to achieve the baseline separation of the four tanshinones and to meet the requirement of accurate quantification in real samples, the MEKC method was further optimized on the basis of the preliminary results described in Results. The effects of SDS, organic solvents, [bmim]bf 4, buffer concentration and buffer ph have been investigated. Effect of SDS Different concentrations of SDS (0, 5, 10, 15, 20 and 25 mm) were tested to separate the mixed standards. As shown in Figure 3A, the four analytes could not be separated from each other in the absence of SDS. When SDS was added to the buffer within the range of 5 15 mm, a remarkable separation of the analytes with high resolution could be obtained, which suggested that the SDS micelles had formed and interacted with the target components. However, once the concentration of SDS was higher than 15 mm, the peaks of cryptotanshinone and tanshinone I were partially covered and their resolution was getting worse with increasing migration times, revealing the decreased selectivity of SDS for the highly hydrophobic compounds. Although both 5 and 10 mm with high peak resolutions seemed to be suitable for this study, higher peak signal response was observed when using 10 mm SDS, which could improve the method sensitivity. So, 10 mm SDS was chosen for the subsequent experiment.

4 1438 Cao et al. Figure 2. The typical electropherograms of the mixed standard solution using different running buffers. Conditions: (A) 20 mm borax; (B) 20 mm borax 20 mm SDS; (C) 20 mm borax 20 mm SDS 10 mm β-cd; (D) 20 mm borax 10 mm SDS 15% MeOH. Peaks: (1) dihydrotanshinone I, (2) cryptotanshinone, (3) tanshinone I and (4) tanshinone IIA. Figure 3. Effect of SDS and ACN concentration on migration behavior of the target analytes. Conditions: (A) 10 mm borax 10 mm [bmim]bf 4 containing 15% ACN (v/v) at ph 9.6 with different concentrations of SDS; (B) 10 mm borax 10 mm SDS 10 mm [bmim]bf 4 at ph 9.6 with different percentage of ACN.

5 Separation and Determination of Four Tanshinones in Danshen 1439 Effect of organic solvents In order to improve the separation performance, different kinds of organic solvents were tested as additives, including MeOH, EtOH, ACN and isopropanol. The addition of ACN provided the shortest analysis time in the case of ensuring the complete separation of the four analytes. So, different concentrations of ACN (0, 5, 10, 15, 20 and 25%) on the separation of these analytes were further investigated. As shown in Figure 3B, the resolution was notably improved and the baseline separation was obtained with the increase of ACN concentration in running buffer. While the migration times of analytes also increased in the range of 5 10% because an increasing ACN concentration could decrease the EOF rate, resulting in the increasing migration time. However, when the concentrations of ACN increased from 15 to 25%, there was a decreasing trend for the migration times of all compounds, as well as the deteriorating resolution, especially for cryptotanshinone and tanshinone I. One possible reason was that, when the ACN concentration was higher enough, it could dramatically enhance the hydrophobicity of aqueous phase in MEKC and led to the increasing partitioning of the analytes into the aqueous phase. Considering the highest resolutions for the four analytes, 15% ACN was selected in this study. Effect of [bmim]bf 4 According to preliminary experiments, an IL [bmim]bf 4 was used as another modifier to further enhance the selectivity of the MEKC system toward the investigated analytes, especially for cryptotanshinone and tanshinone I. Different concentrations of [bmim]bf 4 (0, 5, 10, 15 and 20 mm) were explored and the corresponding results were shown in Figure 4. It can be found that both the selectivity and the resolution of these analytes were significantly improved with the increasing [bmim]bf 4 concentration. The two analytes (i.e., cryptotanshinone and tanshinone I) that represented as a single peak in Figure 2D could now achieve a baseline separation when the concentration of [bmim]bf 4 increased to 10 mm. Although a further increase of the [bmim]bf 4 concentration achieved higher resolution between cryptotanshinone and tanshinone I, the greater Joule heat that generated would make the peak shapes of the target compounds broader. Finally, 10 mm was selected as the optimum concentration of [bmim]bf 4. The complete separation of cryptotanshinone and tanshinone I after the addition of [bmim]bf 4 could be explained by the variation of the inner polarity of SDS micelle. The inner polarity of SDS micelle could be evaluated by the fluorescent intensity ratio of the first emission peak to the third one (I 1 /I 3 ). The results in Table I showed that, with the increased addition of [bmim]bf 4,thevalueofI 1 /I 3 also increased, indicating an increasing inner polarity of SDS micelle. Thus, the difference of polarity between the micellar phase and the aqueous phase became smaller, and the partitioning of the hydrophobic analytes into the aqueous phase got greater, resulting in the Figure 4. Effect of [bmim]bf 4 concentration on migration behavior of the investigated analytes. Conditions: 10 mm borax 10 mm SDS 10 containing 15% ACN (v/v) at ph 9.6 with different concentrations of [bmim]bf 4. Peaks: (1) dihydrotanshinone I, (2) cryptotanshinone, (3) tanshinone I and (4) tanshinone IIA.

6 1440 Cao et al. Table I. Effect of [bmim]bf 4 on the Microstructure of SDS Micelle 10 mm borax 10 mm SDS 15% ACN with different concentrations of [bmim]bf 4 0 mm 5 mm 10 mm 15 mm 20 mm I 1 /I ζ potential (mv) Hydrodynamic size (d, nm) I 1 /I 3 : the fluorescence intensity ratio of the first emission peak (I 1 ) to the third emission peak (I 3 ) of pyrene. Figure 5. Effect of buffer ph and buffer concentration on migration behavior of the investigated analytes. Conditions: (A) 10 mm borax 10 mm SDS 10 mm [bmim] BF 4 containing 15% ACN (v/v) at different ph values; (B) 10 mm SDS 10 mm [bmim]bf 4 containing 15% ACN (v/v) with different concentrations of borax. Peaks: (1) dihydrotanshinone I, (2) cryptotanshinone, (3) tanshinone I and (4) tanshinone IIA. enhanced selectivity toward cryptotanshinone and tanshinone I. In addition, comparing with tanshinone I, cryptotanshinone had a stronger polarity and then a larger partitioning into the aqueous phase. Therefore, as shown in Figure 4, the peak of cryptotanshinone appeared prior to that of tanshinone I. Moreover, the migration times of all the analytes became shorter after the addition of [bmim]bf 4 to running buffer, which can be illustrated by the variations of the size and ζ potential of SDS micelle. As shown in Table I, both the reduction of the negative charge on the surface of SDS micelle and the increase of SDS micelle size were significant after the addition of 5 mm [bmim]bf 4. Then, the negative charge reduced slowly as the concentration of [bmim]bf 4 increased from 5 to 20 mm as well as the increased SDS micelle size. It can be preliminary speculated that imidazolium cations on [bmim]bf 4 were electrostatically attracted to the negatively charged surface of SDS micelles and then the negative charge was neutralized, resulting in the reduction of the negative charge. At the same time, some [bmim]bf 4 molecules might insert into the inner phase of SDS micelle and made the micelle size larger. Therefore, the charge-to-mass ratio of SDS micelles decreased with the increase of the concentration of [bmim]bf 4, resulting in the decrease of electrophoresis speed of negatively charged SDS micelle in the opposite direction to EOF and the consequent reduction of the migration time of SDS micelle. Effect of buffer ph In order to study the effect of buffer ph on the migration behavior of the target analytes, different ph values ranging from 8.7 to 10.0 were investigated to analyze the mixed standards and the result was presented in Figure 5A. Despite the complete separation of these analytes could be observed at all investigated ph values, the target peaks showed very poor shapes with low signal responses, which would decrease the sensitivity of the method, except at ph 9.6. In terms of

7 Separation and Determination of Four Tanshinones in Danshen 1441 analysis time, it decreased when the ph value increased from 8.7 to 9.6, whereas an increasing trend was observed when the ph ranged from 9.6 to So, with regard to the high sensitivity and the short analysis time, ph 9.6 was chosen as the optimal ph condition for the buffer solution. Effect of buffer concentration The effect of buffer concentration was examined using different concentrations of borax, including 0, 5, 10, 15 and 20 mm. Theoretically, an increasing buffer concentration can raise the buffer viscosity and ionic strength, which could decrease the EOF rate as reflected by longer migration time and better resolution. As demonstrated in Figure 5B, the migration times of all analytes increased and better resolutions were obtained when the borax concentration increased from 5 to 20 mm. Meanwhile, as indicated in Figure 5B, a higher buffer concentration could produce more Joule heat, which resulted in peak broadening. Thus, 5 mm borax seemed to be the best option; however, the investigated compounds could not be resolved from other interferences in real samples. At last, 10 mm borax was used (with the current lower than 65 μa). Method validation The developed MEKC method was validated by studying the linearity, limit of detection (LOD), limit of quantification (LOQ), precision and accuracy. Linearity, LOD and LOQ For construction of calibration curves of the four analytes, the mixed standard solution was diluted with MeOH to seven different concentrations; then triplicate injections were performed for each concentration. The calibration curves were constructed by plotting the peak area (y) versus the concentration of each analyte (x). The standard curves, linear ranges and correlation coefficients (R 2 ) were summarized in Table II, which displayed good linearity for all analytes with R 2 values of at least Under the present chromatographic conditions, the LODs and LOQs of the developed method were determined by analyzing a decreasing series of the mixed standard solution based on the signal-to-noise ratio of 3 : 1 for LOD and 10 : 1 for LOQ, with the results listed in Table II. Precision Intra- and interday variations were chosen to determine the precision of the established method. For intraday variability test, the standard solution was examined six replicates within 1 day, while for interday variability test, the standard solution was assayed in duplicates for consecutive 6 days. As shown in Table III, the relative standard deviation (RSD) values of migration times and peak areas for intraday were lower than 1.07 and 3.36%. And for interday variation, the RSD values of migration times and peak areas were no more than 2.21 and 3.90%, respectively. Table II. Calibration Curves, LODs and LOQs of the Investigated Analytes Analyte Migration time (min) Standard curve R 2 Linear range (μg ml 1 ) LOD (μgml 1 ) LOQ (μgml 1 ) Dihydrotanshinone I y = 294x Cryptotanshinone y = 185x Tanshinone I y = 206x Tanshinone IIA y = 209x Table III. Intraday and Interday Precision of the Investigated Tanshinones (n =6) Analyte Concentration (μg ml 1 ) Intraday RSD (%) Interday RSD (%) Migration times Peak areas Migration times Peak areas Dihydrotanshinone I Cryptotanshinone Tanshinone I Tanshinone IIA Table IV. Recoveries of the Four Tanshinones (n =3) Analyte Original (μg ml 1 ) Added (μgml 1 ) Found (mean, μg ml 1 ) Recovery (mean, %) RSD (%) Dihydrotanshinone I Cryptotanshinone Tanshinone I Tanshinone IIA

8 1442 Cao et al. prepared in regard to the method described in Sample preparation and then analyzed using the established MEKC procedure mentioned in Electrophoretic procedure. Three replicates were performed for the assay. As summarized in Table IV, the average recoveries for all the analytes were in the range of % with RSDs <4.38%, indicating the good accuracy of the developed MEKC method for quantitative determination of the four tanshinones in real samples. Quantification of four tanshinones in Salvia plants The MEKC method was then applied to the quantitative analysis of four tanshinones in the roots of Salvia plants from different origins. The typical electropherograms of the mixed standards and samples are shown in Figure 6. The identification of the target analytes was carried out by comparing their migration times and UV absorption spectra with those of corresponding standards (Supplementary Figure S1), or by spiking individual standards into sample solutions. From Figure 6, it can be seen that the four target tanshinones were all well separated from other interfering components in three different kinds of Salvia species and could be accurately quantified. The contents of the analytes in samples were calculated and listed in Table V. Among the tested samples, all the four analytes could be detected in all samples except that dihydrotanshinone I was not detected in one S. miltiorrhiza sample from Sichuan, indicating the good quality of the analyzed samples. And the content of tanshinone IIA ranged from 0.62 to 3.55 mg g 1 was the most abundant of target compounds in all samples, which was used as an indicator for the quality control of Danshen in Chinese Pharmacopoeia (2 mg g 1 ) (6). The contents of cryptotanshinone and tanshinone I were relatively close in the range of and mg g 1, but much lower than tanshinone IIA. Of these four analytes, the content of dihydrotanshinone I was the lowest with no detection or under LOD in some individual samples. In terms of species difference, it was clearly discovered that the overall contents of the four tanshinones were much higher (2 folds at least) in S. przewalskii and S. castanea than those in S. miltiorrhiza, which proved the rationality of the application of these Salvia plants as alternatives of Danshen resource. Moreover, some phenolic acids, such as salvianolic acid A, salvianolic acid B, danshensu, protocatechuic aldehyde and protocatechuic acid, are also bioactive components in Danshen and related medicinal plants. Apart from tanshinones, these bioactive phenolic acids should also be chosen as chemical markers to compare the therapeutic potential of Salvia species with that of Danshen. Conclusion Figure 6. Typical electropherograms of the mixed standard solution and real samples under optimum conditions. (A) Mixed standards; (B) Sample no. 4 S. miltiorrhiza; (C) Sample no.1 S. przewalskii; (D) Sample no. 2 S. castanea. Peaks: (1) dihydrotanshinone I, (2) cryptotanshinone, (3) tanshinone I and (4) tanshinone IIA. Accuracy In order to assess the accuracy of the method, the recovery study was carried out by spiking accurate amounts of the four standards at three different levels (low, medium and high) into a certain amount (0.5 g) of S. miltiorrhiza material (Sample 4). The fortified samples were In this work, a simple and rapid MEKC method has been developed to analyze four tanshinones in the roots of three Salvia species. The addition of [bmim]bf 4 was demonstrated to be effective enough for the simultaneous separation and determination of these highly hydrophobic compounds with high resolution and relatively short migration time. Meanwhile, experimental study revealed that the addition of IL [bmim]bf 4 in this established MEKC method had effects on the inner polarity, size and ζ potential of SDS micelle to change the selectivity for hydrophobic tanshinones. Supplementary Material Supplementary materials are available at Journal of Chromatographic Science (

9 Separation and Determination of Four Tanshinones in Danshen 1443 Table V. Contents (mg g 1 ) of Four Tanshinones in Salvia Plants from Different Origins (n =3) No. Samples Origin Dihydrotanshinone I Cryptotanshinone Tanshinone I Tanshinone IIA 1 S. przewalskii Yunnan 0.48 ± 0.10 a 2.03 ± ± ± S. castanea Yunnan 0.17 ± ± ± ± S. miltiorrhiza Shandong 0.21 ± ± ± ± S. miltiorrhiza Hebei 0.26 ± ± ± ± S. miltiorrhiza Hebei 0.10 ± ± ± ± S. miltiorrhiza Sichuan <LOD 0.29 ± ± ± S. miltiorrhiza Sichuan ND 0.40 ± ± ± 0.17 ND, not detected. a Mean ± SD (n = 3). Acknowledgments This research was supported by the National Natural Science Foundation of China ( ), the Macao Science and Technology Development Fund (No. 052/2012/A2) and the Research Committee of the University of Macau (No. MYRG109-ICMS13-LP). References 1. Li, M.H., Chen, J.M., Peng, Y., Wu, Q., Xiao, P.G.; Investigation of Danshen and related medicinal plants in China; Journal of Ethnopharmacology, (2008); 120: Lu, Y., Foo, L.Y.; Polyphenolics of Salvia a review; Phytochemistry, (2002); 59: Lin, H.C., Chang, W.L.; Diterpenoids from Salvia miltiorrhiza; Phytochemistry, (2000); 53: Wu, C.T., Mulabagal, V., Nalawade, S.M., Chen, C.L., Yang, T.F., Tsay, H.S.; Isolation and quantitative analysis of cryptotanshinone, an active quinoid diterpene formed in callus of Salvia miltiorrhiza BUNGE; Biological & Pharmaceutical Bulletin, (2003); 26: Liu, A.L., Wang, Y.T., Du, G.H., Yang, X.Y., Zhang, J.T., Tang, M.K., et al.; Active constituents in Danshen and their pharmacological actions. In Dan Shen (Salvia miltiorrhiza) in medicine: volume 2. Pharmacology and quality control. Springer, New York, (2014), pp Pharmacopoeia Committee of P.R. China. Pharmacopoeia of People s Republic of China. Part I. China Medical Science Press, Beijing, (2010). 7. Lu, L.Y., Liu, Y., Zhang, Z.F., Zhang, H.; Analysis of Danshen and twelve related Salvia species; Natural Product Communications, (2012); 7: Lee, P., Hur, J., Lee, J., Kim, J., Jeong, J., Kang, I., et al.; 15,16-Dihydrotanshinone I suppresses the activation of BV-2 cell, a murine microglia cell line, by lipopolysaccharide; Neurochemistry International, (2006); 48: Lam, F.F., Yeung, J.H., Chan, K.M., Or, P.M.; Dihydrotanshinone, a lipophilic component of Salvia miltiorrhiza (danshen), relaxes rat coronary artery by inhibition of calcium channels; Journal of Ethnopharmacology, (2008); 119: Zhou, Z., Wang, S.Q., Liu, Y., Miao, A.D.; Cryptotanshinone inhibits endothelin-1 expression and stimulates nitric oxide production in human vascular endothelial cells; Biochimica et Biophysica Acta, (2006); 1760: Hur, J.M., Shim, J.S., Jung, H.J., Kwon, H.J.; Cryptotanshinone but not tanshinone IIA inhibits angiogenesis in vitro; Experimental and Molecular Medicine, (2005); 37: Kim, S.Y., Moon, T.C., Chang, H.W., Son, K.H., Kang, S.S., Kim, H.P.; Effects of tanshinone I isolated from Salvia miltiorrhiza bunge on arachidonic acid metabolism and in vivo inflammatory responses; Phytotherapy Research, (2002); 16: Lee, C.Y., Sher, H.F., Chen, H.W., Liu, C.C., Chen, C.H., Lin, C.S., et al.; Anticancer effects of tanshinone I in human non-small cell lung cancer; Molecular Cancer Therapeutics, (2008); 7: Wang, X., Wei, Y., Yuan, S., Liu, G., Lu, Y., Zhang, J., et al.; Potential anticancer activity of tanshinone IIA against human breast cancer; International Journal of Cancer, (2005); 116: Fu, J., Huang, H., Liu, J., Pi, R., Chen, J., Liu, P.; Tanshinone IIA protects cardiac myocytes against oxidative stress-triggered damage and apoptosis; European Journal of Pharmacology, (2007); 568: Yang, R.L., Lu, Y., Liu, J.H.; Identification of tanshinone IIA as a natural monoacylglycerol lipase inhibitor by combined in silico and in vitro approach; Medicinal Chemical Communications, (2014); 5: Pan, X., Niu, G., Liu, H.; Microwave-assisted extraction of tanshinones from Salvia miltiorrhiza bunge with analysis by high-performance liquid chromatography; Journal of Chromatography A, (2001); 922: Shi, Z., He, J., Yao, T., Chang, W., Zhao, M.; Simultaneous determination of cryptotanshinone, tanshinone I and tanshinone IIA in traditional Chinese medicinal preparations containing Radix salvia miltiorrhiza by HPLC; Journal of Pharmaceutical and Biomedical Analysis, (2005); 37: Chen, Y., Duan, G., Xie, M., Chen, B., Li, Y.; Infrared-assisted extraction coupled with high-performance liquid chromatography for simultaneous determination of eight active compounds in Radix Salviae miltiorrhizae; Journal of Separation Science, (2010); 33: Zhou, L., Chow, M., Zuo, Z.; Improved quality control method for Danshen products consideration of both hydrophilic and lipophilic active components; Journal of Pharmaceutical and Biomedical Analysis, (2006); 41: Li, P., Xu, G., Li, S.P., Wang, Y.T., Fan, T.P., Zhao, Q.S., et al.; Optimizing ultraperformance liquid chromatographic analysis of 10 diterpenoid compounds in Salvia miltiorrhiza using central composite design; Journal of Agricultural and Food Chemistry, (2008); 56: Liu, M., Li, Y., Chou, G., Cheng, X., Zhang, M., Wang, Z.; Extraction and ultra-performance liquid chromatography of hydrophilic and lipophilic bioactive components in a Chinese herb Radix Salviae Miltiorrhizae; Journal of Chromatography A, (2007); 1157: Hu, P., Luo, G.A., Zhao, Z.Z., Jiang, Z.H.; Quantitative determination of four diterpenoids in Radix Salviae Miltiorrhizae using LC-MS-MS; Chemical & Pharmaceutical Bulletin, (2005); 53: Wu, J.L., Yee, L.P., Jiang, Z.H., Cai, Z.; One single LC-MS/MS analysis for both phenolic components and tanshinones in Radix Salviae Miltiorrihizae and its medicinal products; Talanta, (2007); 73: Ganzera, M.; Quality control of herbal medicines by capillary electrophoresis: potential, requirements and applications; Electrophoresis, (2008); 29: Chen, A.J., Li, C.H., Gao, W.H., Hu, Z.D., Chen, X.G.; Application of nonaqueous micellar electrokinetic chromatography to the analysis of active components in radix Salviae miltiorrhizae and its medicinal preparations; Journal of Pharmaceutical and Biomedical Analysis, (2005); 37: Che, A.J., Zhang, J.Y., Li, C.H., Chen, X.F., Hu, Z.D., Chen, X.G.; Separation and determination of active components in Radix Salviae miltiorrhizae and its medicinal preparations by nonaqueous capillary electrophoresis; Journal of Separation Science, (2004); 27: Cao, J., Chen, J., Yi, L., Li, P., Qi, L.W.; Comparison of oil-in-water and water-in-oil microemulsion electrokinetic chromatography as methods for the analysis of eight phenolic acids and five diterpenoids; Electrophoresis, (2008); 29:

10 1444 Cao et al. 29. Li, P., Li, S.P., Yang, F.Q., Wang, Y.T.; Simultaneous determination of four tanshinones in Salvia miltiorrhiza by pressurized liquid extraction and capillary electrochromatography; Journal of Separation Science, (2007); 30: Yashima, T., Tsuchiya, A., Morita, O., Terabe, S.; Separation of closely related large peptides by micellar electrokinetic chromatography with organic modifiers; Analytical Chemistry, (1992); 64: Nishi, H., Fukuyama, T., Terabe, S.; Chiral separation by cyclodextrin-modified micellar electrokinetic chromatography; Journal of Chromatography A, (1991); 553: Takeda, S., Wakida, S., Yamane, M., Kawahara, A., Higashi, K.; Migration behavior of phthalate-esters in micellar electrokinetic chromatography with or without added methanol; Analytical Chemistry, (1993); 65: Butehorn, U., Pyell, U.; Simultaneous use of urea and acetonitrile as organic modifiers for optimization of resolution in micellar electrokinetic chromatography; Journal of Chromatography A, (1997); 792: Huang, H.Y., Chiu, C.W., Chen, Y.C., Yeh, J.M.; Comparison of microemulsion electrokinetic chromatography and micellar electrokinetic chromatography as methods for the analysis of ten benzophenones; Electrophoresis, (2005); 26: Mwongela, S.M., Numan, A., Gill, N.L., Agbaria, R.A., Warner, I.M.; Separation of achiral and chiral analytes using polymeric surfactants with ionic liquids as modifiers in micellar electrokinetic chromatography; Analytical Chemistry, (2003); 75: Rageh, A.H., Pyell, U.; Imidazolium-based ionic liquid-type surfactant as pseudostationary phase in micellar electrokinetic chromatography of highly hydrophilic urinary nucleosides; Journal of Chromatography A, (2013); 1316: Tian, K., Qi, S.D., Cheng, Y.Q., Chen, X.G., Hu, Z.D.; Separation and determination of lignans from seeds of Schisandra species by micellar electrokinetic capillary chromatography using ionic liquid as modifier; Journal of Chromatography A, (2005); 1078: Cao, J.L., Hu, J., Wei, J.C., Li, B.C., Zhang, M., Xiang, C., et al.; Optimization of micellar electrokinetic chromatography method for the simultaneous determination of seven hydrophilic and four lipophilic bioactive components in three Salvia species; Molecules, (2015); 20: Ray, G.B., Chakraborty, I., Moulik, S.P.; Pyrene absorption can be a convenient method for probing critical micellar concentration (cmc) and indexing micellar polarity; Journal of Colloid and Interface Science, (2006); 294: Ni, X.J., Yu, M.J., Cao, Y.H., Cao, G.Q.; Microstructure of microemulsion modified with ionic liquids in microemulsion electrokinetic chromatography and analysis of seven corticosteroids; Electrophoresis, (2013); 34: