Differentiation and Authentication of Panax ginseng, Panax

Jin-Gyeong Cho , Seo-Ji In , Ye-Jin Jung , Byeong-Ju Cha , Dae-Young Lee .... Re as an effective tool for the determination of ginsenosides in various...
0 downloads 0 Views 128KB Size
Anal. Chem. 2000, 72, 1281-1287

Differentiation and Authentication of Panax ginseng, Panax quinquefolius, and Ginseng Products by Using HPLC/MS T. W. D. Chan,†,‡ P. P. H. But,*,†,§ S. W. Cheng,† I. M. Y. Kwok,† F. W. Lau,† and H. X. Xu†

Chinese Medicinal Material Research Centre, Department of Chemistry, and Department of Biology, The Chinese University of Hong Kong, Shatin, NT, Hong Kong

An LC/MS-based method is established for the differentiation and authentication of specimens and commercial samples of Panax ginseng (Oriental ginseng) and Panax quinquefolius (American ginseng). This method is based on the separation of ginsenosides present in the ginseng methanolic extracts using high-performance liquid chromatography (HPLC), followed by detection with electrospray mass spectrometry. Differentiation of ginsenosides is achieved through simultaneous detection of intact ginsenoside molecular ions and the ions of their characteristic thermal degradation products. An important parameter used for differentiating P. ginseng and P. quinquefolius is the presence of ginsenoside Rf and 24(R)-pseudoginsenoside F11 in the RICs of Oriental and American ginsengs, respectively. It is important to stress that ginsenoside Rf and 24(R)-pseudoginsenoside F11, which possess the same molecular weight and were found to have similar retention times under most LC conditions, can be unambiguously distinguished in the present HPLC/ MS method. The method developed is robust, reliable, reproducible, and highly sensitive down to the nanogram level. The roots of Panax ginseng C. A. Meyer (Oriental ginseng) and Panax quinquefolius L. (American ginseng) are highly treasured medicinal herbs in traditional Chinese medicines. Research studies have shown that they possess antistress,1 CNSstimulating,2 antidiabetic,3 and anticancer4 properties. Ginseng products have become very popular and have attracted worldwide consumption. However, Oriental ginseng and American ginseng * Corresponding author. Tel: (852) 2609-6140. Fax: (852) 2603-5248. E-mail: [email protected]. † Chinese Medicinal Material Research Centre. ‡ Department of Chemistry. § Department of Biology. (1) (a) Grandhi, A.; Mujumdar, A. M.; Patwardhan, B. J. Ethnopharmacol. 1994, 44 (3), 131-135. (b) Kumar, R.; Grover, S. K.; Divekar, H. M.; Gupta, A. K.; Shyam, R.; Srivastava, K. K. Int. J. Biometeorol. 1996, 39 (4), 187-191. (2) (a) Kim, H. S. Planta Med. 1990, 56, 158-160. (b) Kim, H. S. Pharmacol. Biochem. Behav. 1996, 53, 185-190. (c) Tokuyama, S. Pharmacol. Biochem. Behav. 1996, 54, 671-676. (3) (a) Wu, J. Y. J. Immunol. 1992, 148, 1519-1525. (b) Newman, M. J. J. Immunol. 1992, 148, 2357-2362. (4) (a) Qian, B. C.; Zhang, X. X.; Li, B.; Xu, C. Y.; Deng, X. Y. Acta Pharmacol. Sin. 1987, 8, 277-280. (b) Cott, J. Psychopharmacol. Bull. 1995, 31, (4), 745-751. 10.1021/ac990819z CCC: $19.00 Published on Web 02/11/2000

© 2000 American Chemical Society

are known to have different properties and medicinal values. The red form of Oriental ginseng, produced by steaming, is “warm” and known to replenish the “vital energy”, whereas American ginseng is “cool” and is mainly used for reducing the “internal heat” and promoting the secretion of body fluids.5 In the ginseng markets worldwide, American ginseng usually commands a much higher price than the sun-dried Oriental ginseng. Since the roots of these ginsengs are similar in appearance and many commercial ginseng products are in form of powder or shredded slices, identification of the origins of the ginseng products is not an easy task. Authentication of the two sources of ginseng and ginseng products based on chemical profiling has aroused much interest. Ginsenosides (ginseng saponins) are known to be the bioactive components of ginsengs. According to the difference in the aglycone in these saponins, ginsenosides are classified into three types: the 20(S)-protopanaxadiol type (e.g., ginsenosides Rb1, Rc, Rb2, and Rd), the 20(S)-protopanaxatriol type (e.g., ginsenosides Rg1, Rf, and Re), and the oleanolic acid type (e.g., ginsenoside Ro) (see Figure 1a-c). Ginsenosides Rg1, Re, Ro, Rb1, Rc, Rb2, and Rd are present in both Oriental and American ginsengs in different proportions, whereas ginsenoside Rf can only be found in Oriental ginseng.6-8 In contrast, American ginseng contains 24(R)-pseudoginsenoside F11, an ocotillol type triterpene (Figure 1d), which is absent in Oriental ginseng.7-9 Many analytical approaches have been used to identify ginsenosides in ginseng extracts. Among these methods, the use of high-performance liquid chromatography (HPLC) in conjunction with mass spectrometry (MS or MS/MS) appears to be most promising. In comparison with the use of gas chromatography (GC) coupled with electron impact mass spectrometry (EI-MS), the HPLC/MS approach can overcome the problems associated with ginsenoside derivatization and the low abundance of molec(5) (a) Shanghai College of Traditional Chinese Medicine. Zhongcaoyaoxue (Chinese Materia Medica); Commerce Press: Hong Kong, 1975; pp 511515. (b) Shanghai College of Traditional Chinese Medicine. Zhongyao Linchuang Shouce (Clinical Manual of Chinese Medicine); Shanghai People’s Publishing House: Shanghai, 1977; pp 357-359. (6) (a) Lang, W. S.; Lou, Z. C.; But, P. P. H. J. Chin. Pharm. Sci. 1993, 2 (2), 133-143. (b) Chuang, W. C.; Wu, H. K.; Sheu, S. J.; Chiou, S. H.; Chang, H. C.; Chen, Y. P. Planta Med. 1995, 61, 459-465. (7) Dou, D.-Q.; Hou, W.-B.; Chen, Y.-J. Planta Med. 1998, 64, 585-586. (8) Tanaka, O.; Kasai, R.; Morita, T. Abstr. Chin. Med. 1986, 1 (1), 130-152. (9) Chen, S. E.; Staba, E. J.; Taniyasu, S.; Kasai, R.; Tanaka, O. Planta Med. 1981, 42 (4), 406-411.

Analytical Chemistry, Vol. 72, No. 6, March 15, 2000 1281

Figure 1. Structures of ginsenosides in forms of (a) 20(S)-protopanaxadiol, (b) 20(S)-protopanaxatriol, and (c) oleanolic acid and of pseudoginsenosides in form of (d) 24(R)-ocotillol type triterpene.

ular ions.10 It has also been reported that HPLC systems equipped with analytical columns containing materials ranging from reversedphase silica11 and ion exchange packings12 to specialized carbohydrate-containing aminopropyl functional groups13 could separate intact, underivatized ginsenosides effectively from ginseng ex(10) (a) Kim, B. Y.; Lee, M. Y.; Cho, K. H.; Park, J. H.; Park, M. K. Arch. Pharm. Res. 1992, 328-332. (b) Cui, J.-F.; Bjo¨rkhem, I.; Eneroth, P. J. Chromatogr., B 1997, 689, 349-355. (11) (a) Kanazawa, H.; Nagata, Y.; Kurosaki, E.; Matsushima, Y.; Takai, N. J Chromatogr. 1993, 632, 79-85. (b) Court, W. A.; Hendel, J. G.; Elmi, J. J. Chromatogr., A 1996, 755, 11-17.

1282 Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

tracts. The use of an on-line MS detector offers superiority over conventional UV absorbance detection in terms of specificity and sensitivity.14,15 Among the various mass spectrometry approaches, (12) (a) Parr, M. K.; Parr, J. H.; Lee, M. Y.; Kim, S. J.; Park, I. J. J. Liq. Chromatogr. 1994, 17, 1171-1182. (b) Yamaguchi, H.; Kasai, R.; Matsura, H.; Tanaka, O.; Fuwa, T. Chem. Pharm. Bull. 1988, 36, 3468-3473. (13) (a) Paik, N. H.; Park, M. K.; Choi, K. J.; Cho, Y. H. Arch. Pharm. Res. 1982, 5, 7-12. (b) Yamaguchi, H.; Matsuura, H.; Kasai, R.; Tanaka, O.; Satake, M.; Kohda, H.; Izumi, H.; Nuno, M.; Katsuki, S.; Isoda, S.; Shoji, J.; Goto, K. Chem. Pharm. Bull. 1988, 36, 4177-4181. (c) Park, M. K.; Park, J. H.; Han, S. B.; Shin, Y. G.; Park, I. H. J. Chromatogr., A 1996, 736, 77-81.

electrospray mass spectrometry (ESI-MS) is considered to be the best to couple with HPLC. Similar to other mass spectrometric techniques, ESI-MS was found to produce intense protonated or cationizated ginsenoside ions.14,15 Application of electrospray tandem mass spectrometry (ESI-MS/MS) involving collisioninduced dissociation (CID) of the selected ions has been shown to provide product ion spectra which are useful for structure determination. At present, only a few papers have reported the use of HPLC/ ESI-MS and HPLC/ESI-MS/MS for differentiating Oriental and American ginsengs. Though the techniques were proved to be very powerful, further efforts devoted to method development are required. For instance, van Breeman et al.14 pointed out that certain ginsenosides were detected only in the methanolic extract of Oriental ginseng. The assignment of the constituents, however, was based purely on the correction of a mass shift of m/z 138 corresponding to the column bleed-off adducts. The results had not been confirmed by using authentic ginsenoside standards. Recently, Wang and coauthors15 differentiated Oriental and American ginsengs by contrasting the relative abundances of ginsenosides present in both species. It was reported that the ginsenoside ratios Rg1/Rf and Rc/Rb2 are higher for American ginseng than for Oriental ginseng. Nevertheless, it was well realized that American ginseng does not contain ginsenoside Rf6-8 but contains 24(R)-pseudoginsenoside F11.7-9 The problem is that ginsenoside Rf and 24(R)-pseudoginsenoside F11 share the same molecular weight and are found to elute at similar retention times under most LC conditions; false identification of Rf in American ginsengs is thus possible. The objective of the present investigation is to develop a simple LC/ESI-MS technique for differentiation of Oriental and American ginsengs. The approach relies on the detection of indicative constituents besides comparing the differences in relative abundance of the ginsenosides common to both species. Positive identification of ginsenoside constituents was accompanied by the simultaneous detection of the sodiated ginsenoside molecular ions and the ions derived from in-source thermal degradation of the protonated ginsenoside molecular ions. The method developed was applied to the analysis of ginseng samples cultivated in China, Korea, Canada, and the United States. In addition, this method was also used to analyze commercial ginseng slices and ginseng teas. EXPERIMENTAL SECTION Standards of ginsenosides including Rg1, Re, Ro, Rf, Rb1, Rb2, Rc, Rd, and 24(R)-pseudoginsenoside F11 were kindly provided by Prof. Y. J. Chen of Shenyang Pharmaceutical University, Shenyang, China, and the Tobacco and Ginseng Monopoly, South Korea. Structures of these ginsenosides and the pseudoginsenoside are shown in Figure 1. Samples of Oriental ginseng were obtained from China and Korea. Roots of American ginseng derived from cultivated samples in the United States, Canada, and China were gifts from the Wisconsin Ginseng Board. The solvents, acetonitrile and methanol, were of HPLC grade (Mallinckrodt, Paris, KY). Other reagents were of analytical grade. Deionized (14) van Breemen, R. B.; Huang, C.-R.; Lu, Z.-Z.; Rimando, A.; Fong, H. H. S.; Fitzloff, J. F. Anal. Chem. 1995, 67, 3985-3989. (15) Wang, X.; Sakuma, T.; Asafu-Adjaye, E.; Shiu, G. K. Anal. Chem. 1999, 71, 1579-1584.

water (18 MΩ) was prepared by passing distilled water through an ultrapure water system (Nanopure, Barnstead, Germany). The ginseng sample was first homogenized. The powdered sample (300 mg) was purified by extraction with 10 mL of chloroform in a microextractor under reflux for 3 h. The sample was then filtered, and the chloroform extract was discarded. The solid residue was further extracted with 10 mL of methanol under the same conditions. The methanolic extract was evaporated to dryness under vacuum. The residue of ginsenosides formed was redissolved in 3.0 mL of deionized water. The aqueous solution was applied to an SEP-PAK C18 cartridge (Waters Associates). The ginsenosides were washed with 10 mL of water and then eluted with 10 mL of 50% acetonitrile. The solution collected was filtered through a 0.2 µm filter (Millipore) prior to LC/MS analysis. For direct flow injection experiments, solutions of ginsenoside standards were prepared at concentrations of 1 µg/µL. In each experiment, 10 µL of the sample solution was used. A Hewlett-Packard 1090 liquid chromatograph system was interfaced to the mass spectrometer. An analytical column of 46 mm × 150 mm packed with a 3 µm hydrophobic bonded C18 phase (Adsorbosphere HS, Alltech Associates Inc.) was utilized. Gradient elution generated by the proportional mixing of two solvents was used. Solvent A consisted of 50% acetonitrile and solvent B contained 10 mM ammonium acetate, both in deionized water. Solvent gradient conditions were a linear change of solvent B from 70 to 50% in the first 30 min and to 10% in the next 30 min. This solvent composition was maintained for another 10 min before returning to 70% solvent B in 2 min. The flow rate was 1.0 mL/ min with the column kept at ambient temperature. For flow injection experiments, the HPLC instrument delivered mobile phases of A and B in a 1:1 ratio at a flow rate of 100 µL/min. Electrospray mass spectra were acquired using an MAT TSQ 7000 triple-stage quadrupole mass spectrometer (Finnigan, San Jose, CA). The needle potential was set at 5 kV. A heated stainless steel capillary was used as a bridge between the atmospheric pressure side and the first pumping stage. Desolvation occurred in this capillary. Nitrogen was used as the nebulizing gas at a pressure of 50 psi. Full-scan mass spectra were acquired by scanning the first quadrupole (Q1) over the range m/z 300-2300 at a scan time of 1.5 s. Q2 and Q3 were set at the rf-only mode for transmission of all ions of the selected mass range. An electron multiplier was employed as the detector and was set at a potential of -1300 V. RESULTS AND DISCUSSION Positive-Ion Mass Spectra of Ginsenoside Standards. In an attempt to obtain mass spectrometric-based information for differentiating different types of ginsenosides, direct flow injection experiments of ginsenoside standards were initially performed with a capillary temperature of 175 °C. In contrast to the results of earlier studies using 252Cf-plasma desorption,16 field desorption,17 laser desorption,18 and liquid secondary ion mass spectrometry,19 (16) Elkin, Y. N.; Makhankov, V. V.; Uvarova, N. L.; Bondarenko, P. V.; Zubarev, R. A.; Knysh, A. N. Acta Pharmacol. Sin. 1993, 14, 97-100. (17) Schulten, H. R.; Soldati, F. J. Chromatogr. 1981, 212, 37-49. (18) Zhou, Y.; Song, F. R.; Liu, S. Y.; Li, X. G. Acta Chim. Sin. 1998, 56, 298301. (19) Yamamoto, M.; Sugiyama, K.; Ichio, Y.; Yokota, M.; Maeda, Y.; Senda, N.; Shizukuishi, K. Shoyakugaku Zasshi 1992, 46, 394-396.

Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

1283

Figure 2. Positive-ion ESI-MS spectra of ginsenoside Rg1 at capillary temperatures of (a) 175 °C, (b) 200 °C, (c) 250 °C, and (d) 300 °C. Table 1. Major Ions (m/z) Observed in the Positive-Ion ESI-MS Spectra of the Ginsenoside Standardsa ginsenoside

type

[M + H]+

[M + Na]+

other major ions

Rb1 Rb2 Rc Ro Rd Re Rf Rg1 F11

20(S)-protopanaxadiol 20(S)-protopanaxadiol 20(S)-protopanaxadiol oleanolic acid 20(S)-protopanaxadiol 20(S)-protopanaxatriol 20(S)-protopanaxatriol 20(S)-protopanaxatriol 24(R)-pseudoginsenoside

NDb ND ND ND ND ND ND ND 801.5

1132.1 1102.0 1101.9 979.6 969.9 969.7 823.9 823.8 823.7

407.4, 425.4, 443.3, 569.5, 587.6, 605.5, 749.7, 784.8 330.1, 407.4, 425.5, 443.5, 587.8, 605.6, 786.0 407.4, 425.4, 443.5, 587.5, 605.7, 790.5 439.2 407.4, 425.5, 443.6, 569.6, 605.7, 749.7, 784.9 405.5, 423.4, 441.4, 459.6, 587.7, 603.7, 621.9, 749.8, 767.9 405.3, 423.3, 441.3, 459.3, 587.5 405.2, 423.2, 441.2, 459.2, 603.3, 621.5 308.9, 421.2, 439.2, 457.2, 475.2, 603.7, 655.4, 783.5

a

The temperature of the heated capillary was set at 300 °C. b ND, not detected.

positive-ion mass spectra of all ginsenosides and 24(R)-pseudoginsenoside F11 featured signals indicating the production of [M + H]+ and [M + NH4]+ molecular ions. No significant signals corresponding to the sodiated molecular ions were observed. Figure 2a shows the positive-ion flow injection mass spectrum of Rg1. Apart from the abundant [M + H]+ and [M + NH4]+ molecular ions, some low-mass ions corresponding to the thermal degradation products of the ginsenoside Rg1 (as proved subsequently by using LC/MS analysis) were also observed. Initial attempts to utilize these low-mass signals for positive identification of the ginsenosides were hampered by their relatively low intensities and high variability. Figure 2b-d shows the positiveion flow injection mass spectra of Rg1 at other capillary temperatures. The intensities of the low-mass ions were substantially increased by elevating the temperature of the heated capillary. It is interesting to note that the [M + H]+ and [M + NH4]+ ions were completely degraded at capillary temperatures above 250 °C. In addition, signals corresponding to the [M + Na]+ ions appeared to form at high capillary temperatures (>250 °C). The production of ions of thermal degradation products simulta1284 Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

neously with cationized molecular ions appears to be advantageous. Because the mass spectra obtained reveal common features in the fragmentation behavior of the ginsenosides, they are useful for the characterization and identification of these species. For example: for 20(S)-protopanaxadiol type ginsenosides, fragment ions at m/z 443, 425, 407 were observed; for 20(S)-protopanaxatriol type ginsenosides, the corresponding fragment ions were observed m/z 441, 423, 405 with an additional fragment at m/z 459; and for ginsenoside Ro, the oleanolic acid type ginsenoside, a single fragment at m/z 439 was exhibited. For 24(R)-pseudoginsenoside F11, fragment ions at m/z 475, 457, 439, 421 were observed. The spectral information is summarized in Table 1. Figure 3 shows the mass spectra of ginsenosides Rb2, Rf, and Ro and 24(R)-pseudoginsenoside F11. One striking spectral feature of 24(R)-pseudoginsenoside F11 was the presence of abundant protonated molecular ions, [M + H]+, even at high capillary temperatures. All major thermal degradation product ions appeared to arise from the loss of H2O molecules from the deglycoslyated ginsenoside core moieties. For 20(S)-protopanaxadiol type ginsenosides,

Figure 3. Positive-ion ESI-MS spectra of (a) ginsenoside Rb2, (b) ginsenoside Rf, (c) 24(R)-pseudoginsenoside F11, and (d) ginsenoside Ro. The temperature of the heated capillary was set at 300 °C.

the simultaneous loss of R1 and R2 groups yielded initially the core moiety at m/z 461. As a result of combinational loss of H2O molecules derived from the three hydroxyl groups of the core moiety, fragment ions at m/z 443, 425, 407 were produced. For 20(S)-protopanaxatriol type ginsenosides, the loss of R1 and R2 groups gave rise to a core moiety with four hydroxyl functional groups at m/z 477. The mass spectra therefore featured four characteristic ions, i.e., those at m/z 459, 441, 423, 405. For the oleanolic acid type ginsenosides, the corresponding core moiety arising from the loss of R1 and R2 contained only one hydroxyl group. Loss of a water molecule from this hydroxyl group produced a product ion at m/z 439. For the 24(R)-pseudoginsenoside F11, the core moiety with four hydroxyl groups at m/z 493 was first derived from the loss of the R group. Successive elimination of four H2O molecules generated four product ions at m/z 475, 457, 439, 421. With in-source thermal degradation ESI-MS, ginsenosides of different types possessing identical molecular weight, e.g. Rg1/ Rf vs F11 (MW ) 800) and Re vs Rd (MW ) 946), could be easily recognized by their diagnostic product ions. However, for two pairs of ginsenosides, Rg1/Rf (MW ) 800) and Rc/Rb2 (MW ) 1078), unambiguous identification could only rely on the observation of other minor product ions such as those at m/z 603 in Rg1 and m/z 330 in Rb2. However, with the use of on-line liquid chromatographic separation, their identifications could be achieved from their characteristic retention times. In summary, the present ESI-MS method provided a reliable means to distinguish different types of ginsenosides. Both the molecular weight and structural information could be revealed in a single mass spectrum. The fragmentation patterns generated were characteristic of the ginsenoside types. Our results also suggested that the positive ESI-MS approach can be applied to

identify the type and molecular weight of an unknown ginsenoside even if an authentic standard is not available. LC/MS Analysis of a Mixture of Ginsenoside Standards. To demonstrate the identification of individual ginsenosides in a complex mixture, the ginsenoside standards were mixed and analyzed by HPLC/MS. Figure 4 shows the positive-ion LC/MS traces of ginsenoside standards collected in 60 min. A distinct difference between the present LC method and those that were previously reported6a,15 was the use of a rather gentle change of mobile phase composition, thus resulting in a longer time of elution. By sacrificing the time for analysis, we achieved a complete resolution of ginsenosides Rg1/Re and Rf/F11. Complete separation of 24(R)-pseudoginsenoside F11 and ginsenoside Rf was especially important, since the former component is the characteristic constituent of American ginseng while the latter is the characteristic constituent of Oriental ginseng. As shown in Figure 4, all ginsenosides were well resolved under the present LC conditions. Their respective mass spectra are similar to those obtained by direct flow injection analysis. Unlike the results obtained with the use of an analytical column containing aminopropyl functional groups on the silica-based stationary phase, no unexpected adduct ions from the column materials were observed.14 Ginsenosides labeled on the reconstructed ion chromatogram were identified on the basis of their retention times and the m/z values of the intact molecular ions and the thermal degradation product ions. Table 2 summarizes the retention times and the set of ions for positive identification of individual ginsenosides. Wang et al.15 had previously reported the detection of ginsenoside Rf in the American ginsengs cultivated in both the United States and Canada by using LC/MS/MS with selected reaction monitoring (SRM) (i.e., m/z 801 f 423). However, it is well-known Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

1285

Figure 4. The LC/MS traces of the mixture of ginsenoside standards Rg1, Re, Ro, Rf, Rb1, Rc, Rb2, Rd, and 24(R)-pseudoginsenoside F11. Table 2. Retention Times and the Set of Ions for Positive Identification of Individual Ginsenosides m/z ginsenoside

retention time (min)

intact molecular ion

characteristic thermal degradation ions

Rb1 Rb2 Rc Ro Rd Re Rf Rg1 F11

42.1 44.4 43.2 26.6 47.1 23.4 38.5 22.5 37.4

1132 1102 1102 979 969 969 823 823 801

407, 425, 443 407, 425, 443 407, 425, 443 439 407, 425, 443 405, 423, 441, 459 405, 423, 441, 459 405, 423, 441, 459 421, 439, 457, 475

that American ginsengs do not contain ginsenoside Rf but pseudoginsenoside F11. The false-positive identification of Rf in their samples is intriguing. Although 24(R)-pseudoginsenoside F11 and ginsenoside Rf share the same molecular weight and are very often coeluted under most LC conditions, misrecognition of these components might happen if the assignments are based purely on the molecular masses and retention times, especially when authentic standards are not available for reference. However, under SRM conditions (i.e., m/z 801 f 423), such misrecognition would not be possible because of the absence of the CID fragment of F11 at m/z 423 (data not included). Anyway, the present LC/ MS method provides two independent types of parameters, i.e., retention time and mass spectrometric information, for the positive identification of Rf, F11, and other ginsenoside constituents. 1286 Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

LC/MS of Oriental and American Ginsengs. Twelve Oriental and American ginsengs, including some special reference ginseng standards purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China), were analyzed by the present LC/MS method. Figure 5 shows some representative ion chromatograms (RICs). Consistent with the general findings,7 ginsenoside Rf was found only in Oriental ginseng, whereas 24(R)-pseudoginsenoside F11 was found exclusively in American ginseng. In comparison with the commonly used thin-layer chromatography (TLC) methods, the LC/ MS method showed much superior quality in terms of sensitivity and specificity. It is also worthwhile to point out that 24(R)pseudoginsenoside F11 does not contain any suitable chromophore identifiable by a UV absorbance detector and could not be detected by normal LC analysis. Inspection of Figure 5 and RICs of other ginseng samples showed that the ratios Rg1/Re and Rb2/Rc were also found to be specific for these two species of ginsengs. The Rg1/Re ratio was always found to be much greater than 1 in Oriental ginseng, whereas the same ratio was found to be smaller than 1 in American ginseng. A more subtle difference was found in the Rb2/ Rc ratios, which approach 0.8 and 0.2 in Oriental ginseng and American ginseng, respectively. In addition, there was consistently a prominent peak eluted behind Rd (retention time: 48.9 min) for American ginseng. This signal peak was only present in extremely weak intensity for Oriental ginseng. Though the authentic standard was not available, some information on this component could be extracted from its characteristic thermal degradation products and its molecular weight. It was found to be a 20(S)-protopanaxadiol type ginsenoside having a molecular weight of 946; further identification of this component is currently under investigation. One of the major disadvantages of using relative peak height/area ratios for differentiating American and Oriental ginsengs is the dependence of the relative ginsenoside contents on the age of the ginseng and the physical parts of the ginseng being used. LC/MS of Commercial Ginseng Products. Twelve commercial ginseng products, including eight ginseng slices and four samples of ginseng tea granules, were analyzed by using the present LC/MS method. Among these products, two brands of ginseng slices and two brands of ginseng tea granules were found to contain ginsengs different from those described by the manufacturers. One brand of ginseng slices that claimed to be made from American ginseng actually contained Oriental ginseng. The opposite was true for the other brand of ginseng slices. For the ginseng tea granules, one brand of American ginseng tea granules was derived from Oriental ginseng, whereas another brand of American Ginseng tea granules was actually made from a mixture of both American and Oriental ginsengs. Figure 6 shows the positive-ion LC/MS traces of ginsenosides isolated. Although the RICs show mainly contributions from the Oriental ginseng, SIM at m/z 439 and 475 has positively identified the presence of American ginseng. Similar studies using conventional TLC methods would be very difficult, if not impossible. Close inspection of the RICs of Figure 6 reveals another important limitation in methods that utilize relative peak height/area ratios for differentiation and authentication of ginseng products. Mixing of different ginsengs could actually alter the relative abundance of

Figure 5. Reconstructed ion chromatograms for the LC/MS analyses of Oriental and American ginsengs extracts.

CONCLUSION The present study demonstrated a simple HPLC/MS method for differentiation and authentication of Oriental ginseng, American ginseng, and commercial ginseng products. For ginseng specimens, the differentiation hinged on the detection of indicative ginsenosides (i.e., Rf and F11) in addition to contrasting the differences in relative abundance of various ginsenosides. For commercial ginseng products, such as ginseng slices and ginseng tea granules, authentication can only be performed by detection of ginsenoside Rf (for Oriental ginseng) and 24(R)-pseudoginsenoside F11 (for American ginseng). The limits of detection for the present HPLC/MS method were determined. All ginsenosides could be detected at the nanogram on-column level. In contrast to the employment of UV absorbance detection, where the detection limit is only 40 µg,6a the LC/MS approach enormously increased detection sensitivity by over 1000 times. The method developed is also straightforward and convenient and requires no expensive tandem equipment. ACKNOWLEDGMENT

Figure 6. The LC/MS traces of the extract of a brand of American ginseng tea granules.

various ginsenosides in the final products and thus render the differentiation and authentication impossible. Results reported here indicate that Rf or F11 markers would be a better index for differentiating the origins and purities of ginseng products.

Partial support from the Industrial Support Fund (Grants AF/ 181/97 and AF/154/98) is gratefully acknowledged. We thank Prof. Y. J. Chen of Shenyang Pharmaceutical University, Shenyang, China, and the Tobacco and Ginseng Monopoly, South Korea for the gifts of the ginsenosides. Thanks are also extended to the Wisconsin Ginseng Board for providing the ginseng specimens.

Received for review July 23, 1999. Accepted December 14, 1999. AC990819Z Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

1287