- Email: [email protected]
Metabolomic evaluation of ginsenosides distribution in Panax genus (Panax ginseng and Panax quinquefolius) using multivariate statistical analysis
Fitoterapia 101 (2015) 80–91
Contents lists available at ScienceDirect
Fitoterapia journal homepage: www.elsevier.com/locate/fitote
Metabolomic evaluation of ginsenosides distribution in Panax genus (Panax ginseng and Panax quinquefolius) using multivariate statistical analysis Roberto Pace ⁎, Ernesto Marco Martinelli, Nicola Sardone, Eric DE Combarieu a r t i c l e
i n f o
Article history: Received 8 October 2014 Accepted in revised form 23 December 2014 Accepted 26 December 2014 Available online 3 January 2015 Keywords: PLC Metabolomic Principal component analysis Ginseng Panax ginseng Panax quinquefolius
a b s t r a c t Ginseng is any one of the eleven species belonging to the genus Panax of the family Araliaceae and is found in North America and in eastern Asia. Ginseng is characterized by the presence of ginsenosides. Principally Panax ginseng and Panax quinquefolius are the adaptogenic herbs and are commonly distributed as health food markets. In the present study high performance liquid chromatography has been used to identify and quantify ginsenosides in the two subject species and the different parts of the plant (roots, neck, leaves, flowers, fruits). The power of this chromatographic technique to evaluate the identity of botanical material and to distinguishing different part of the plants has been investigated with metabolomic technique such as principal component analysis. Metabolomics provide a good opportunity for mining useful chemical information from the chromatographic data set resulting an important tool for quality evaluation of medicinal plants in the authenticity, consistency and efficacy. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Ginseng is an ancient Chinese medicine that possesses CNS-stimulating, cardiotonic and antistress effects [1]. The drug is obtained from roots, two main sources are used: Asian ginseng (Panax ginseng C. A. Meyer) and North American ginseng (Panax quinquefolius). Numerous studies suggest that its pharmacological properties are due to triterpene glycosides called ginsenosides [1]. Ginsenosides are dammarane glycosides possessing (20S)-protopanaxadiol and (20S)-protopanaxatriol aglycone moieties (Table 1) [2]. They are divided into neutral and acidic saponins. The latter are ginsenosides having a malonyl residue located on a glycosidic chain [3]. The malonyl ginsenosides are thermally unstable and readily demalonylated by heating [4]. Ginsenoside Ro is the only oleanolic acid-type saponin identified in the root of P. ginseng[1].
⁎ Corresponding author at: Indena S.p.A., Via Don Minzoni 6, 20090 Settala, Italy. Tel.: +39 0295413 606; fax: +39 0295413 676. E-mail address: [email protected] (R. Pace).
http://dx.doi.org/10.1016/j.fitote.2014.12.013 0367-326X/© 2015 Elsevier B.V. All rights reserved.
As ginseng derived products the extracts are multicomponent complex mixtures whose quality relies on the quality of the plant material and the manufacturing process from which they originate. On this regard the ginsenoside content and ratios, responsible for different bioactivities, depend strictly from the species, geographical origins and from the used part of the plant [5,6]. Since it is known that different ginseng plant parts may possess different content of actives compounds, regulatory agencies, such as the US FDA [7], EMA [8] and China SFDA [9], required that herbal substances are prepared from specific parts of the botanical raw material. Because of the different costs that P. ginseng and P. quinquefolius roots have on the market and particularly the lower cost of aerial parts with respect to the root, frequent cases of adulteration are reported [10,11]. From a botanical point of view P. ginseng and P. quinquefolius roots are difficult to be distinguished from each other. To address the identity and quality of botanicals one hurdle has been to develop analytical methods to adequately identify the source, i.e., different plant parts, of the raw material in order to ensure that the botanical drug substance and drug product
R. Pace et al. / Fitoterapia 101 (2015) 80–91
81
Table 1 Structures of main ginsenosides. Glc = β-D-glucose; Rha = α-L-rhamnose; Arap = α-L-arabinose (pyranose); Araf = α-L-arabinose (furanose); Xyl = β-D-xylose; Ma = malonyl. Ginsenoside
R1
R2
R3
Rh1 Rg1 Rf 20-Glc-Rf Rg2 Re
H H H H H H
H Glc H Glc H Glc
Glc Glc Glc2-Glc Glc2-Glc Glc2-Rha Glc2-Rha
Rh2 Rg3 Rd Malonyl-Rd Rb2 Malonyl-Rb2 Rc Malonyl-Rc Rb1 Malonyl-Rb1 Rb3 Ra1 Ra2 Ra3
Glc Glc2-Glc Glc2-Glc Glc2-Glc6-Ma Glc2-Glc Glc2-Glc6-Ma Glc2-Glc Glc2-Glc6-Ma Glc2-Glc Glc2-Glc6-Ma Glc2-Glc Glc2-Glc Glc2-Glc Glc2-Glc
H H Glc Glc Glc6-Arap Glc6-Arap Glc6-Araf Glc6-Araf Glc6-Glc Glc6-Glc Glc6-Xyl Glc6-Arap4-Xyl Glc6-Araf2-Xyl Glc6-Glc3-Xyl
– – – – – – – – – – – – – –
Ro
Glc2-Glc
Glc
–
can be reproducibly manufactured to provide the same safety and efficacy. Among the various analytical methods, ultra- and highperformance liquid chromatography with suitable detectors (e.g. MS/MS, ELSD) are used in comprehensive and reliable ginsenoside profiling and quantitation for various ginseng products [12–17]. Furthermore fingerprint technology can be applied as a powerful method, with metabolomics approach, for characterizing quality of ginseng products [14]. Metabolomics primarily focuses on comprehensive and quantitative profiling for small-molecule metabolites in a biological system. It has been applied to a variety of areas, such as plant toxicology, nutrition, and systems biology [18–20]. Multiple analytical methods, such us liquid and gas chromatography, nuclear magnetic resonance, have been applied in metabolic profiling in order to differentiate Panax species [21–23].
In the present study a metabolomics approach combining a HPLC-ELSD based analysis with pattern recognition method (principal component analysis, PCA), as illustrated in the workflow of study design shown in Fig. 1, has been used to compare and investigate the ginsenoside distribution in different parts of the P. ginseng and P. quinquefolius plant materials. 2. Materials and methods 2.1. Chemicals and materials Ginsenosides Rg1, Re, Ro, Rb1 and Rb2 were isolated and characterized at the Indena R&D Laboratories (Indena SpA, Milan, Italy). Rg2, Rc and Rd were kindly provided by the Institut de Pharmacognosie et Phytochimie, Université de Lausanne, (Lausanne, Switzerland). These pure compounds
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 1. Workflow of study design.
82
R. Pace et al. / Fitoterapia 101 (2015) 80–91
were used for the characterization of a purified P. ginseng dry extract, which was used as working standard. P. ginseng plant materials were collected in China, Korea and United States. Plant materials were divided in main roots (4 samples; about: diameter 10–20 mm, length 50–90 mm), secondary roots (9 samples; about: diameter 0.2–4.0 mm, length 1.0–9.0 cm), neck and aerial parts (leaves (12 samples), stem (6 samples), flowers (1 sample)). One sample of each part of P. quinquefolius was analyzed: main roots (about: diameter 10–20 mm, length 50–90 mm), secondary roots (about: diameter 0.2–4.0 mm, length 1.0– 9.0 cm), neck, leaves, stem, green fruits and red fruits. Samples of P. ginseng and P. quinquefolius used for this study have been identified by Indena botanical R&D laboratory. 2.2. Sample preparation About 100 g of representative dried samples of each type was ground with a mill (b0.2 mm). In order to preserve the malonyl ginsenosides that are thermally unstable, the plant materials have been extracted with 40% aqueous ethanol by sonication for 15 min and shaking for 4 h at room temperature. The optimum sample weight is between 4 and 8 g extracted with 100 mL of 40% aqueous ethanol. The sonication in the extraction procedure is necessary to reduce the time and to improve the efficiency of the extraction. As most of the commercial sonication baths heat the solution during the sonication procedure it is necessary to keep under control the bath temperature (≤30 °C) in order to avoid the malonylginsenoside degradation [15]. 2.3. HPLC-ELSD analysis The HPLC system consisted of a Waters 2690 Alliance (Bedford, MA, USA). ELS detector was a SEDEX 55 from S.E.D.E.R.E. (Altfortville, France) working at a temperature of the nebulizer of 70 °C, nitrogen as nebulizing gas at a pressure of 2.5 bar. The column temperature was set at 25 °C and controlled with a Thermostatted Column Compartment 1100 Series from Hewlett Packard (Waldbronn, Germany). The chromatographic data were recorded and processed by the Waters Empower software. Analyses were carried out at 25 °C on a Hypersil BDS C18 column (250X4.6 mm I.D., 5 μm, Hypersil Astmoor Runcorn, UK). Chromatographic separation was carried out using 8 mM ammonium acetate adjusted at pH 7 with ammonium hydroxide (A) and acetonitrile (B) in a linear gradient program according the following profile with flow rate set at 0.7 mL/min: 0–26 min, 84–72% A, 16–28% B; 26–46 min, 72–68% A, 28–32% B; 46–66 min, 68–60% A, 32–40% B; 66–71 min, isocratic 60% A, 40% B; 71–75 min, 60–84% A, 40–16% B; 75–85 min, isocratic 84% A, 16% B. The injection volume was 15 μL. Identification of ginsenosides and malonylginsenosides was based on HPLC-ESI/MS analysis [16]; their quantitation was achieved through log–log linear calibration [15]. 2.4. Metabolomic data analysis The obtained data have been analyzed using multivariate metabolomic method (principal component analysis—PCA) in order to enhance the comparison.
83
Principal component analysis (PCA) is a standard tool in modern data analysis in different fields because it is a simple, non-parametric method for extracting relevant information from complex data sets allowing to project simultaneously all the variables (e.g. peaks, absorbances, etc.) into a bidimensional space, preserving most of the variance, with a minimal loss of information. PCA condenses the information contained in a large number of original variables into a smaller set of new composite dimensions with a minimal loss of information (or noise effects). It reduces the original dimension of a data set (corresponding to the different variables, e.g. number of wavelengths of spectra, or number of pixels into the images) into fewer dimensions where each new dimension is defined as a linear combination of the original variables. PCA provides a roadmap for how to reduce a complex data set to a lower dimension to reveal the sometimes hidden, simplified structures that often underlie its discarding noise. The dataset has been exported and evaluated making use of PCA (MathWorks, Natick, USA). The original data have been previously subjected to normalization, centering and auto-scaling. 3. Results and discussion 3.1. HPLC analysis of ginseng genus and part of the plant. The HPLC-ELSD analysis performed as described in the paragraph 2.3 allowed to identify and quantify sixteen ginsenosides. As an example, in Fig. 2, the typical HPLC profiles of P. ginseng roots and leaves have been reported. The obtained results outline that several malonyl-ginsenosides characterize the ginseng roots; on the contrary the content of these constituents is negligible in the aerial parts. Complete results are reported in Supplementary material (Fig. S1). Some of the features seem to be related with the investigated genus and part of the plant. Nevertheless due to the amount of data the metabolomic evaluation is needed in order to enhance the comparison and tentatively identify the markers. 3.2. Discrimination of ginseng genus and part of the plant The explained variance of the model shows that the first two principal components (PC1 and PC2) explain the majority of the observed variance (more than 60%, Supplementary material, Fig. S2). Then, the analysis has been limited on the components PC1 and PC2 also because these components account for the separation of the classes as explained afterwards. The data points of the multidimensional space are projected down onto the max variance bi-dimensional space: i.e. PC1 vs. PC2 (Fig. 3). As outlined the Panax genus and the parts of the plant are identified in specific clusters: the Principal Component 1 justifies the differences between the P. ginseng roots and aerial parts; the principal component 2 justifies the differences between the P. quinquefolius roots, the aerial parts and the ginseng species. The principal components PC1 and PC2 are the linear combination of the original variables (ginsenosides content) with the calculated coefficients of correlation (loading 1 and loading 2) and represent the contribution of each ginsenoside content to the separation of the sample points (Figs. 4 and 5,
mV
150.00 140.00 130.00 120.00 110.00
90.00
100.00
80.00 70.00 60.00 50.00
Rd Rd
40.00
Ra
a
b
10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 40.00 42.00 44.00 46.00 48.00 50.00 52.00 54.00 56.00 58.00 60.00 62.00 64.00 66.00 68.00 70.00 72.00 74.00 76.00 78.00 80.00 Minutes
Malonil-Rd Isomer
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 2. HPLC profile of Panax ginseng root (a) and Panax ginseng leaves (b).
10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 40.00 42.00 44.00 46.00 48.00 50.00 52.00 54.00 56.00 58.00 60.00 62.00 64.00 66.00 68.00 70.00 72.00 74.00 76.00 78.00 80.00 Minutes
Rb2 Rb3
30.00
Rc Rc
20.00
8.00
Rb1 Rg2 Rb1 Rg2
0.00
6.00
Rf Malonyl-Rd Malonyl-Rd
10.00
4.00
8.00
Malonyl-Rb2 orRb3orRc
-10.00 2.00
6.00
Rf
0.00
4.00
Malonyl-Rb1 Malonyl-Rb2 or Rb3 or Rc Malonyl-Ra Malonyl-Rb2 orRb3orRc Malonyl-Rb1 Malonyl-Rb2 or Rb3 or Rc
280.00
2.00
Ro Ro
260.00 240.00 220.00 200.00 180.00 160.00 140.00 120.00
80.00
100.00
60.00 40.00 20.00 0.00
0.00
-20.00
Rb2 Rb3
Re Rg1 Re Rg1
84
mV
R. Pace et al. / Fitoterapia 101 (2015) 80–91 Fig. 3. Principal component 1 (PC1) vs principal component 2 (PC2). Green dots and sky-blue dots represent P. ginseng aerial parts (leaves and steam respectively), blue and red dots represent P. ginseng root samples (main and secondary roots respectively), pink dots represent the P. quinquefolius aerial parts (leaves and fruits) and the yellow dots represent P. quinquefolius roots (main, secondary roots and neck). The ellipses account for the variability of each cluster. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
85
86
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 4. Loadings for PC1.
respectively). The positive loading 1 values justify the cluster distribution in the positive part of the PC1 axis; on the contrary the negative loading 1 values explain the data distribution in the negative part of the PC1 axis (Supplementary material
Fig. S3). PC1 justifies the differences among the part clusters: roots differ from the leaves and steam for a higher content of malonyl ginsenosides, Rc, Rb1, Ra, Rb2; on the contrary Rg1, Re, Rd and malonyl-Rd characterize the aerial parts.
Fig. 5. Loadings for PC2.
R. Pace et al. / Fitoterapia 101 (2015) 80–91
In the same manner, separations among PC2 axis are justified by the loadings 2 (Supplementary material Fig. S4). PC2 justifies the differences between P.ginseng and P.quinquefolius clusters. P.quinquefolius root cluster differs from the aerial parts cluster mainly for a higher content of malonyl ginsenosides, Rb3 and Rf; on the contrary Rg1, Re, Rg2, Rc, Ra, Rb1, Rd and malonyl-Ra and malonyl-Rd characterize the P. quinquefolius aerial parts. The same markers justify the separation of the P. quinquefolius and P. ginseng species. 3.3. Assignment of tentative makers of ginseng parts of the plant (within genus) To investigate the power discrimination of the model to identify the different parts of the Panax plants the P.ginseng and P.quinquefolius species have been investigated separately. Additionally, to evaluate the detectability of ginseng roots extract adulteration with leaves extracts mixtures of aerial part and roots in different proportions have been added to the model for both the genus. P. ginseng. In Supplementary material (Fig. S5) the matrix of investigated samples with the ginsenoside content (%) of P. ginseng roots, P. ginseng aerial part and their mixtures thereof is reported. The PCA evaluation of the data set outlines that the principal components PC1and PC2 explain about 75% of the observed variance. Then, the analysis has been limited on the first two principal components (PC1 to PC2) as sole PC1 and PC2 account for the separation of the classes. In Fig. 6, the data distribution in the PC1 and PC2 space is reported. The parts of
87
P. ginseng plants are well identified in specific clusters and the mixture of leaves and roots lay down between the two clusters: the principal component 1 justifies the differences between the clusters allowing the identification of potential adulteration of the roots extracts with leaves extracts. The loadings of principal component PC1 represent the contribution of each ginsenoside content to the separation of the sample points. In Fig. 7, the loadings for PC1 vs variables are reported. As reported in the previous evaluation (Figs. 3 and 4) PC1 justifies the differences among the P. ginseng clusters: roots differ from the aerial parts for a higher content of malonyl ginsenosides, Rc, Rb1, Ra, Rb2 and Rb3; on the contrary Rg1, Re, Rd and malonyl-Rd characterize the P. ginseng aerial parts. P. quinquefolius. In Supplementary material (Fig. S6) the matrix of investigated samples of P.quinquefolius roots, P.quinquefolius aerial parts and mixtures thereof are reported. The PCA evaluation of the data set outlines that the principal components PC1 and PC2 explain more than 70% of the observed variance. Then, the analysis has been limited on the first two principal components (PC1 to PC2) as sole PC1 and PC2 account for the separation of the classes. In Fig. 8 the data distribution in the PC1 and PC2 space is reported. The parts of P. quinquefolius plant are grouped in clusters and the mixture of leaves and roots lay down between the two clusters: the principal component 1 justifies the differences between the clusters allowing the identification of potential adulteration of the roots extracts with leaves ones. The principal component PC1 represents the contribution of each ginsenoside content to the separation of the samples points. In Fig. 9 the loadings vs
Fig. 6. Principal component 1 (PC1) vs principal component 2 (PC2). Green dots represent P. ginseng aerial parts (leaves and steam), blue dots represent P. ginseng roots samples (main and secondary roots), and white dots represent the mixtures in different proportions between underground and aerial part of P. ginseng (leaves 80%/ roots 20%, leaves 50%/roots 50% and leaves 20%/roots 80%; W/W). The ellipses account for the variability of each cluster. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 7. Loadings for PC1.
88
R. Pace et al. / Fitoterapia 101 (2015) 80–91
89
Fig. 8. Principal component 1 (PC1) vs principal component 2 (PC2). Green dots represent P. quinquefolius aerial parts (leaves and steam), blue dots represent P. quinquefolius roots samples (main and secondary roots), and white dots represent the mixtures in different proportions between underground and aerial part of P. ginseng (leaves 70%/roots 30%, leaves 50%/roots 50% and leaves 30%/roots 70%; W/W). The ellipses account for the variability of each cluster. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 9. Loadings for PC1.
90
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 10. Principal component 1 (PC1) vs principal component 2 (PC2). Green dots represent P. ginseng roots batches, and blue dots represent the P. quinquefolius roots batches. White dots are the mixtures in different proportions between underground part of P. ginseng and P. quinquefolius (P. ginseng 70%/P. quinquefolius 30% and P. ginseng 30%/P. quinquefolius 70%; W/W). The ellipses account for the variability of each cluster. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
variables for PC1 are reported. As reported in the previous evaluation (Figs. 3 and 5) PC1 justifies the differences among the clusters: roots differ from the aerial parts for an higher
content of Rg1, Rb1, Rc, Rd and malonyl Rb1; on the contrary Re, Rf, Rb2, Rb3 and malonyl-Rb2/Rb3/Rc characterize the aerial parts.
Fig. 11. Loadings for PC1.
R. Pace et al. / Fitoterapia 101 (2015) 80–91
3.4. Assignment of tentative makers of ginseng genus In Supplementary material (Fig. S7) the matrix of investigated samples of P.quinquefolius, P.ginseng roots and mixtures thereof is reported. With the purpose of identify the ginseng roots species, and to verify the multivariate method suitability, batches of P.ginseng roots and P.quinquefolius roots and mixtures thereof have been evaluated. The PCA evaluation of the data set outlines that the principal components PC1and PC2 explain about 85% of the observed variance. Then, the analysis has been limited on the first two principal components (PC1 to PC2) as sole PC1 and PC2 account for the separation of the classes. In Fig. 10 the data distribution in the PC1 and PC2 space is reported. The Panax genus are grouped in two clusters and the mixtures lay down in between: the principal component 1 justifies the differences between the clusters allowing the identification of the roots blending of the genus Panax. The loadings of principal component PC1 represent the contribution of each ginsenoside content to the separation of the samples points. In Fig. 11 the loadings for PC1 are reported. The positive signals of the loading 1, ascribable to ginsenosides Rg1, Rg2, Ra, Rb2, Rb3 and malonyl-Rd, malonyl Ra, malonyl-Rb2/ Rb3/Rc, characterize the P. ginseng root cluster. Likewise the loading 1 negative signals, ascribable to ginsenosides Re, Rd and malonyl-Rb2/Rb3/Rc, malonyl-Rb1, malonyl-Rd, characterize the P.quinquefolius root cluster. 4. Conclusions Optimized chromatographic profiling making use of the Evaporative Light Scattering detection combined with conventional multivariate analysis resulted to be a powerful technique to distinguish among the roots and aerial parts of the model species, P. ginseng and P. quinquefolius. In a quality control setting, having a suitable and reliable methodology to identify botanical raw materials, will help to support and guarantee the quality of herbal products providing assurance of their safety and efficacy. In this regard, this approach could be used to monitor the quality of the ginseng herbal preparations for pharmaceutical market and of the food supplements in order to detect and prevent accidental and economic adulterations. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.fitote.2014.12.013. References [1] Hostettmann K, Marston A. Saponins-Chemistry and Pharmacology of Natural ProductsCambridge: Cambridge University Press; 1995. [2] Tanaka O, Kasai R. In: Herz W, Grisebach H, Kirby GW, Tamm C, editors. Progress in the Chemistry of Organic Natural Products, vol. 46. New York, NY: Springer; 1984.
91
[3] Kitagawa I, Taniyama T, Yoshikawa M, Ikenishi Y, Nakagawa Y. Chemical studies on crude drug processing. VI.: chemical structures of malonylginsenosides Rb1, Rb2, Rc, and Rd isolated from the root of P. ginseng C. A. MEYER. Chem Pharm Bull 1989;37:2961–70. [4] Yamaguchi H, Kasai R, Matsuura H, Tanaka O, Fuwa T. High-performance liquid chromatographic analysis of acidic saponins of ginseng and related plants. Chem Pharm Bull 1988;36:3468–73. [5] Sengupta S, Toh SA, Sellers LA, Skepper JN, Koolwijk P, Leung HW, Yeung HW, Wong RN, Sasisekharan R, Fan FP. Modulating angiogenesis: the yin and the yang in ginseng. Circulation 2004;110:1219–25. [6] Sievenpiper JL, Arnason JT, Leiter LA, Vuksan V. Decreasing, null and increasing effects of eight popular types of ginseng on acute postprandial glycemic indices in healthy humans: the role of ginsenosides. J Am Coll Nutr 2004;23:248–58. [7] US Department Of Health And Human Services Food and Drug Administration, Center for Drug Evaluation and Research. Guidance for industry: botanical drug products. Silver: Spring; 2004. [8] Medicines Agency European. Guidelines on specifications: test procedures and acceptance criteria for herbal substances. Herbal preparations and herbal medicinal products/traditional herbal medical products; 2011 London. http://www.emea.europa.eu. [9] Food State, Drug Administration of China. Technical requirements for the development of fingerprints of TCM injections SFDA; 2000 Beijing. http://www.cmjournal.org/content/6/1/9. [10] Unpublished data. [11] Harnly J, Chen P, Harrington PB. Probability of identification: adulteration of American ginseng with Asian ginseng. J AOAC Int 2013:961258–65. [12] Liu Y, Li J, He J, Abliz Z, Qu J, Yu S. Identification of new trace triterpenoid samponins from the roots of Panax notoginseng by high-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry. Rapid Commun Mass Spectrom 2009;23:667–79. [13] Sun BS, Xu MY, Li Z, Wang YB, Sung CK. UPLC-Q-TOF-MS/MS analysis for steaming times-dependent profiling of steamed Panax quiquefolius and its ginsenosides transformations induced by repetitious steaming. J Ginseng Res 2012;36:277–90. [14] Kim N, Kim K, Choi BY, Lee D, Shin YS, Bang KH, Cha SW, Lee JW, Choi HK, Jang DS, Lee D. Metabolomic approach for age discrimination of P. ginseng using UPLC-Q-Tof MS. J Agric Food Chem 2011;59:10435–41. [15] Fuzzati N, Gabetta B, Jayakar K, Pace R, Ramaschi G, Villa F. Determination of ginsenosides in P. ginseng roots by liquid chromatography with evaporative light-scattering detection. J AOAC Int 2000;83:820–9. [16] Fuzzati N, Gabetta B, Jayakar K, Pace R, Peterlongo F. Liquid chromatography–electrospray mass spectrometric identification of ginsenosides in P. ginseng roots. J Chrom A 1999;854:69–79. [17] Fuzzati N. Analysis methods of ginsenosides. J Chrom B 2004;812:119–33. [18] Watkins SM, German JB. Toward the implementation of metabolomic assessments of human health and nutrition. Curr Opin Biotech 2002;13: 512–6. [19] Weckwerth W. Metabolomics in systems biology. Annu Rev Plant Biol 2003;54:669–89. [20] Robertson DG. Metabonomics in toxicology: a review. Toxicol Sci 2005;85: 809–22. [21] Shellie RA, Marriott PJ, Hule CW. Comprehensive two-dimensional gas chromatography (GC × GC) and GC × GC-quadrupole MS analysis of Asian and American ginseng. J Sep Sci 2003;26:1185–92. [22] Xie GX, Ni Y, Su MM, Zhang YY, Zhao AH, Gao XF, Liu Z, Xiao PG, Jia W. Application of ultra- performance LC-TOF MS metabolite profiling techniques to the analysis of medicinal Panax herbs. Metabolomics 2008;4:248–60. [23] Lee AR, Gautam M, Kim J, Shin WJ, Choi MS, Bang YS. A multi analytical approach for determining the geographical origin of ginseng using strontium isotopes, multielements, and 1H NMR analysis. J Agric Food Chem 2011;59:8560–7.
Contents lists available at ScienceDirect
Fitoterapia journal homepage: www.elsevier.com/locate/fitote
Metabolomic evaluation of ginsenosides distribution in Panax genus (Panax ginseng and Panax quinquefolius) using multivariate statistical analysis Roberto Pace ⁎, Ernesto Marco Martinelli, Nicola Sardone, Eric DE Combarieu a r t i c l e
i n f o
Article history: Received 8 October 2014 Accepted in revised form 23 December 2014 Accepted 26 December 2014 Available online 3 January 2015 Keywords: PLC Metabolomic Principal component analysis Ginseng Panax ginseng Panax quinquefolius
a b s t r a c t Ginseng is any one of the eleven species belonging to the genus Panax of the family Araliaceae and is found in North America and in eastern Asia. Ginseng is characterized by the presence of ginsenosides. Principally Panax ginseng and Panax quinquefolius are the adaptogenic herbs and are commonly distributed as health food markets. In the present study high performance liquid chromatography has been used to identify and quantify ginsenosides in the two subject species and the different parts of the plant (roots, neck, leaves, flowers, fruits). The power of this chromatographic technique to evaluate the identity of botanical material and to distinguishing different part of the plants has been investigated with metabolomic technique such as principal component analysis. Metabolomics provide a good opportunity for mining useful chemical information from the chromatographic data set resulting an important tool for quality evaluation of medicinal plants in the authenticity, consistency and efficacy. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Ginseng is an ancient Chinese medicine that possesses CNS-stimulating, cardiotonic and antistress effects [1]. The drug is obtained from roots, two main sources are used: Asian ginseng (Panax ginseng C. A. Meyer) and North American ginseng (Panax quinquefolius). Numerous studies suggest that its pharmacological properties are due to triterpene glycosides called ginsenosides [1]. Ginsenosides are dammarane glycosides possessing (20S)-protopanaxadiol and (20S)-protopanaxatriol aglycone moieties (Table 1) [2]. They are divided into neutral and acidic saponins. The latter are ginsenosides having a malonyl residue located on a glycosidic chain [3]. The malonyl ginsenosides are thermally unstable and readily demalonylated by heating [4]. Ginsenoside Ro is the only oleanolic acid-type saponin identified in the root of P. ginseng[1].
⁎ Corresponding author at: Indena S.p.A., Via Don Minzoni 6, 20090 Settala, Italy. Tel.: +39 0295413 606; fax: +39 0295413 676. E-mail address: [email protected] (R. Pace).
http://dx.doi.org/10.1016/j.fitote.2014.12.013 0367-326X/© 2015 Elsevier B.V. All rights reserved.
As ginseng derived products the extracts are multicomponent complex mixtures whose quality relies on the quality of the plant material and the manufacturing process from which they originate. On this regard the ginsenoside content and ratios, responsible for different bioactivities, depend strictly from the species, geographical origins and from the used part of the plant [5,6]. Since it is known that different ginseng plant parts may possess different content of actives compounds, regulatory agencies, such as the US FDA [7], EMA [8] and China SFDA [9], required that herbal substances are prepared from specific parts of the botanical raw material. Because of the different costs that P. ginseng and P. quinquefolius roots have on the market and particularly the lower cost of aerial parts with respect to the root, frequent cases of adulteration are reported [10,11]. From a botanical point of view P. ginseng and P. quinquefolius roots are difficult to be distinguished from each other. To address the identity and quality of botanicals one hurdle has been to develop analytical methods to adequately identify the source, i.e., different plant parts, of the raw material in order to ensure that the botanical drug substance and drug product
R. Pace et al. / Fitoterapia 101 (2015) 80–91
81
Table 1 Structures of main ginsenosides. Glc = β-D-glucose; Rha = α-L-rhamnose; Arap = α-L-arabinose (pyranose); Araf = α-L-arabinose (furanose); Xyl = β-D-xylose; Ma = malonyl. Ginsenoside
R1
R2
R3
Rh1 Rg1 Rf 20-Glc-Rf Rg2 Re
H H H H H H
H Glc H Glc H Glc
Glc Glc Glc2-Glc Glc2-Glc Glc2-Rha Glc2-Rha
Rh2 Rg3 Rd Malonyl-Rd Rb2 Malonyl-Rb2 Rc Malonyl-Rc Rb1 Malonyl-Rb1 Rb3 Ra1 Ra2 Ra3
Glc Glc2-Glc Glc2-Glc Glc2-Glc6-Ma Glc2-Glc Glc2-Glc6-Ma Glc2-Glc Glc2-Glc6-Ma Glc2-Glc Glc2-Glc6-Ma Glc2-Glc Glc2-Glc Glc2-Glc Glc2-Glc
H H Glc Glc Glc6-Arap Glc6-Arap Glc6-Araf Glc6-Araf Glc6-Glc Glc6-Glc Glc6-Xyl Glc6-Arap4-Xyl Glc6-Araf2-Xyl Glc6-Glc3-Xyl
– – – – – – – – – – – – – –
Ro
Glc2-Glc
Glc
–
can be reproducibly manufactured to provide the same safety and efficacy. Among the various analytical methods, ultra- and highperformance liquid chromatography with suitable detectors (e.g. MS/MS, ELSD) are used in comprehensive and reliable ginsenoside profiling and quantitation for various ginseng products [12–17]. Furthermore fingerprint technology can be applied as a powerful method, with metabolomics approach, for characterizing quality of ginseng products [14]. Metabolomics primarily focuses on comprehensive and quantitative profiling for small-molecule metabolites in a biological system. It has been applied to a variety of areas, such as plant toxicology, nutrition, and systems biology [18–20]. Multiple analytical methods, such us liquid and gas chromatography, nuclear magnetic resonance, have been applied in metabolic profiling in order to differentiate Panax species [21–23].
In the present study a metabolomics approach combining a HPLC-ELSD based analysis with pattern recognition method (principal component analysis, PCA), as illustrated in the workflow of study design shown in Fig. 1, has been used to compare and investigate the ginsenoside distribution in different parts of the P. ginseng and P. quinquefolius plant materials. 2. Materials and methods 2.1. Chemicals and materials Ginsenosides Rg1, Re, Ro, Rb1 and Rb2 were isolated and characterized at the Indena R&D Laboratories (Indena SpA, Milan, Italy). Rg2, Rc and Rd were kindly provided by the Institut de Pharmacognosie et Phytochimie, Université de Lausanne, (Lausanne, Switzerland). These pure compounds
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 1. Workflow of study design.
82
R. Pace et al. / Fitoterapia 101 (2015) 80–91
were used for the characterization of a purified P. ginseng dry extract, which was used as working standard. P. ginseng plant materials were collected in China, Korea and United States. Plant materials were divided in main roots (4 samples; about: diameter 10–20 mm, length 50–90 mm), secondary roots (9 samples; about: diameter 0.2–4.0 mm, length 1.0–9.0 cm), neck and aerial parts (leaves (12 samples), stem (6 samples), flowers (1 sample)). One sample of each part of P. quinquefolius was analyzed: main roots (about: diameter 10–20 mm, length 50–90 mm), secondary roots (about: diameter 0.2–4.0 mm, length 1.0– 9.0 cm), neck, leaves, stem, green fruits and red fruits. Samples of P. ginseng and P. quinquefolius used for this study have been identified by Indena botanical R&D laboratory. 2.2. Sample preparation About 100 g of representative dried samples of each type was ground with a mill (b0.2 mm). In order to preserve the malonyl ginsenosides that are thermally unstable, the plant materials have been extracted with 40% aqueous ethanol by sonication for 15 min and shaking for 4 h at room temperature. The optimum sample weight is between 4 and 8 g extracted with 100 mL of 40% aqueous ethanol. The sonication in the extraction procedure is necessary to reduce the time and to improve the efficiency of the extraction. As most of the commercial sonication baths heat the solution during the sonication procedure it is necessary to keep under control the bath temperature (≤30 °C) in order to avoid the malonylginsenoside degradation [15]. 2.3. HPLC-ELSD analysis The HPLC system consisted of a Waters 2690 Alliance (Bedford, MA, USA). ELS detector was a SEDEX 55 from S.E.D.E.R.E. (Altfortville, France) working at a temperature of the nebulizer of 70 °C, nitrogen as nebulizing gas at a pressure of 2.5 bar. The column temperature was set at 25 °C and controlled with a Thermostatted Column Compartment 1100 Series from Hewlett Packard (Waldbronn, Germany). The chromatographic data were recorded and processed by the Waters Empower software. Analyses were carried out at 25 °C on a Hypersil BDS C18 column (250X4.6 mm I.D., 5 μm, Hypersil Astmoor Runcorn, UK). Chromatographic separation was carried out using 8 mM ammonium acetate adjusted at pH 7 with ammonium hydroxide (A) and acetonitrile (B) in a linear gradient program according the following profile with flow rate set at 0.7 mL/min: 0–26 min, 84–72% A, 16–28% B; 26–46 min, 72–68% A, 28–32% B; 46–66 min, 68–60% A, 32–40% B; 66–71 min, isocratic 60% A, 40% B; 71–75 min, 60–84% A, 40–16% B; 75–85 min, isocratic 84% A, 16% B. The injection volume was 15 μL. Identification of ginsenosides and malonylginsenosides was based on HPLC-ESI/MS analysis [16]; their quantitation was achieved through log–log linear calibration [15]. 2.4. Metabolomic data analysis The obtained data have been analyzed using multivariate metabolomic method (principal component analysis—PCA) in order to enhance the comparison.
83
Principal component analysis (PCA) is a standard tool in modern data analysis in different fields because it is a simple, non-parametric method for extracting relevant information from complex data sets allowing to project simultaneously all the variables (e.g. peaks, absorbances, etc.) into a bidimensional space, preserving most of the variance, with a minimal loss of information. PCA condenses the information contained in a large number of original variables into a smaller set of new composite dimensions with a minimal loss of information (or noise effects). It reduces the original dimension of a data set (corresponding to the different variables, e.g. number of wavelengths of spectra, or number of pixels into the images) into fewer dimensions where each new dimension is defined as a linear combination of the original variables. PCA provides a roadmap for how to reduce a complex data set to a lower dimension to reveal the sometimes hidden, simplified structures that often underlie its discarding noise. The dataset has been exported and evaluated making use of PCA (MathWorks, Natick, USA). The original data have been previously subjected to normalization, centering and auto-scaling. 3. Results and discussion 3.1. HPLC analysis of ginseng genus and part of the plant. The HPLC-ELSD analysis performed as described in the paragraph 2.3 allowed to identify and quantify sixteen ginsenosides. As an example, in Fig. 2, the typical HPLC profiles of P. ginseng roots and leaves have been reported. The obtained results outline that several malonyl-ginsenosides characterize the ginseng roots; on the contrary the content of these constituents is negligible in the aerial parts. Complete results are reported in Supplementary material (Fig. S1). Some of the features seem to be related with the investigated genus and part of the plant. Nevertheless due to the amount of data the metabolomic evaluation is needed in order to enhance the comparison and tentatively identify the markers. 3.2. Discrimination of ginseng genus and part of the plant The explained variance of the model shows that the first two principal components (PC1 and PC2) explain the majority of the observed variance (more than 60%, Supplementary material, Fig. S2). Then, the analysis has been limited on the components PC1 and PC2 also because these components account for the separation of the classes as explained afterwards. The data points of the multidimensional space are projected down onto the max variance bi-dimensional space: i.e. PC1 vs. PC2 (Fig. 3). As outlined the Panax genus and the parts of the plant are identified in specific clusters: the Principal Component 1 justifies the differences between the P. ginseng roots and aerial parts; the principal component 2 justifies the differences between the P. quinquefolius roots, the aerial parts and the ginseng species. The principal components PC1 and PC2 are the linear combination of the original variables (ginsenosides content) with the calculated coefficients of correlation (loading 1 and loading 2) and represent the contribution of each ginsenoside content to the separation of the sample points (Figs. 4 and 5,
mV
150.00 140.00 130.00 120.00 110.00
90.00
100.00
80.00 70.00 60.00 50.00
Rd Rd
40.00
Ra
a
b
10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 40.00 42.00 44.00 46.00 48.00 50.00 52.00 54.00 56.00 58.00 60.00 62.00 64.00 66.00 68.00 70.00 72.00 74.00 76.00 78.00 80.00 Minutes
Malonil-Rd Isomer
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 2. HPLC profile of Panax ginseng root (a) and Panax ginseng leaves (b).
10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 40.00 42.00 44.00 46.00 48.00 50.00 52.00 54.00 56.00 58.00 60.00 62.00 64.00 66.00 68.00 70.00 72.00 74.00 76.00 78.00 80.00 Minutes
Rb2 Rb3
30.00
Rc Rc
20.00
8.00
Rb1 Rg2 Rb1 Rg2
0.00
6.00
Rf Malonyl-Rd Malonyl-Rd
10.00
4.00
8.00
Malonyl-Rb2 orRb3orRc
-10.00 2.00
6.00
Rf
0.00
4.00
Malonyl-Rb1 Malonyl-Rb2 or Rb3 or Rc Malonyl-Ra Malonyl-Rb2 orRb3orRc Malonyl-Rb1 Malonyl-Rb2 or Rb3 or Rc
280.00
2.00
Ro Ro
260.00 240.00 220.00 200.00 180.00 160.00 140.00 120.00
80.00
100.00
60.00 40.00 20.00 0.00
0.00
-20.00
Rb2 Rb3
Re Rg1 Re Rg1
84
mV
R. Pace et al. / Fitoterapia 101 (2015) 80–91 Fig. 3. Principal component 1 (PC1) vs principal component 2 (PC2). Green dots and sky-blue dots represent P. ginseng aerial parts (leaves and steam respectively), blue and red dots represent P. ginseng root samples (main and secondary roots respectively), pink dots represent the P. quinquefolius aerial parts (leaves and fruits) and the yellow dots represent P. quinquefolius roots (main, secondary roots and neck). The ellipses account for the variability of each cluster. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
85
86
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 4. Loadings for PC1.
respectively). The positive loading 1 values justify the cluster distribution in the positive part of the PC1 axis; on the contrary the negative loading 1 values explain the data distribution in the negative part of the PC1 axis (Supplementary material
Fig. S3). PC1 justifies the differences among the part clusters: roots differ from the leaves and steam for a higher content of malonyl ginsenosides, Rc, Rb1, Ra, Rb2; on the contrary Rg1, Re, Rd and malonyl-Rd characterize the aerial parts.
Fig. 5. Loadings for PC2.
R. Pace et al. / Fitoterapia 101 (2015) 80–91
In the same manner, separations among PC2 axis are justified by the loadings 2 (Supplementary material Fig. S4). PC2 justifies the differences between P.ginseng and P.quinquefolius clusters. P.quinquefolius root cluster differs from the aerial parts cluster mainly for a higher content of malonyl ginsenosides, Rb3 and Rf; on the contrary Rg1, Re, Rg2, Rc, Ra, Rb1, Rd and malonyl-Ra and malonyl-Rd characterize the P. quinquefolius aerial parts. The same markers justify the separation of the P. quinquefolius and P. ginseng species. 3.3. Assignment of tentative makers of ginseng parts of the plant (within genus) To investigate the power discrimination of the model to identify the different parts of the Panax plants the P.ginseng and P.quinquefolius species have been investigated separately. Additionally, to evaluate the detectability of ginseng roots extract adulteration with leaves extracts mixtures of aerial part and roots in different proportions have been added to the model for both the genus. P. ginseng. In Supplementary material (Fig. S5) the matrix of investigated samples with the ginsenoside content (%) of P. ginseng roots, P. ginseng aerial part and their mixtures thereof is reported. The PCA evaluation of the data set outlines that the principal components PC1and PC2 explain about 75% of the observed variance. Then, the analysis has been limited on the first two principal components (PC1 to PC2) as sole PC1 and PC2 account for the separation of the classes. In Fig. 6, the data distribution in the PC1 and PC2 space is reported. The parts of
87
P. ginseng plants are well identified in specific clusters and the mixture of leaves and roots lay down between the two clusters: the principal component 1 justifies the differences between the clusters allowing the identification of potential adulteration of the roots extracts with leaves extracts. The loadings of principal component PC1 represent the contribution of each ginsenoside content to the separation of the sample points. In Fig. 7, the loadings for PC1 vs variables are reported. As reported in the previous evaluation (Figs. 3 and 4) PC1 justifies the differences among the P. ginseng clusters: roots differ from the aerial parts for a higher content of malonyl ginsenosides, Rc, Rb1, Ra, Rb2 and Rb3; on the contrary Rg1, Re, Rd and malonyl-Rd characterize the P. ginseng aerial parts. P. quinquefolius. In Supplementary material (Fig. S6) the matrix of investigated samples of P.quinquefolius roots, P.quinquefolius aerial parts and mixtures thereof are reported. The PCA evaluation of the data set outlines that the principal components PC1 and PC2 explain more than 70% of the observed variance. Then, the analysis has been limited on the first two principal components (PC1 to PC2) as sole PC1 and PC2 account for the separation of the classes. In Fig. 8 the data distribution in the PC1 and PC2 space is reported. The parts of P. quinquefolius plant are grouped in clusters and the mixture of leaves and roots lay down between the two clusters: the principal component 1 justifies the differences between the clusters allowing the identification of potential adulteration of the roots extracts with leaves ones. The principal component PC1 represents the contribution of each ginsenoside content to the separation of the samples points. In Fig. 9 the loadings vs
Fig. 6. Principal component 1 (PC1) vs principal component 2 (PC2). Green dots represent P. ginseng aerial parts (leaves and steam), blue dots represent P. ginseng roots samples (main and secondary roots), and white dots represent the mixtures in different proportions between underground and aerial part of P. ginseng (leaves 80%/ roots 20%, leaves 50%/roots 50% and leaves 20%/roots 80%; W/W). The ellipses account for the variability of each cluster. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 7. Loadings for PC1.
88
R. Pace et al. / Fitoterapia 101 (2015) 80–91
89
Fig. 8. Principal component 1 (PC1) vs principal component 2 (PC2). Green dots represent P. quinquefolius aerial parts (leaves and steam), blue dots represent P. quinquefolius roots samples (main and secondary roots), and white dots represent the mixtures in different proportions between underground and aerial part of P. ginseng (leaves 70%/roots 30%, leaves 50%/roots 50% and leaves 30%/roots 70%; W/W). The ellipses account for the variability of each cluster. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 9. Loadings for PC1.
90
R. Pace et al. / Fitoterapia 101 (2015) 80–91
Fig. 10. Principal component 1 (PC1) vs principal component 2 (PC2). Green dots represent P. ginseng roots batches, and blue dots represent the P. quinquefolius roots batches. White dots are the mixtures in different proportions between underground part of P. ginseng and P. quinquefolius (P. ginseng 70%/P. quinquefolius 30% and P. ginseng 30%/P. quinquefolius 70%; W/W). The ellipses account for the variability of each cluster. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
variables for PC1 are reported. As reported in the previous evaluation (Figs. 3 and 5) PC1 justifies the differences among the clusters: roots differ from the aerial parts for an higher
content of Rg1, Rb1, Rc, Rd and malonyl Rb1; on the contrary Re, Rf, Rb2, Rb3 and malonyl-Rb2/Rb3/Rc characterize the aerial parts.
Fig. 11. Loadings for PC1.
R. Pace et al. / Fitoterapia 101 (2015) 80–91
3.4. Assignment of tentative makers of ginseng genus In Supplementary material (Fig. S7) the matrix of investigated samples of P.quinquefolius, P.ginseng roots and mixtures thereof is reported. With the purpose of identify the ginseng roots species, and to verify the multivariate method suitability, batches of P.ginseng roots and P.quinquefolius roots and mixtures thereof have been evaluated. The PCA evaluation of the data set outlines that the principal components PC1and PC2 explain about 85% of the observed variance. Then, the analysis has been limited on the first two principal components (PC1 to PC2) as sole PC1 and PC2 account for the separation of the classes. In Fig. 10 the data distribution in the PC1 and PC2 space is reported. The Panax genus are grouped in two clusters and the mixtures lay down in between: the principal component 1 justifies the differences between the clusters allowing the identification of the roots blending of the genus Panax. The loadings of principal component PC1 represent the contribution of each ginsenoside content to the separation of the samples points. In Fig. 11 the loadings for PC1 are reported. The positive signals of the loading 1, ascribable to ginsenosides Rg1, Rg2, Ra, Rb2, Rb3 and malonyl-Rd, malonyl Ra, malonyl-Rb2/ Rb3/Rc, characterize the P. ginseng root cluster. Likewise the loading 1 negative signals, ascribable to ginsenosides Re, Rd and malonyl-Rb2/Rb3/Rc, malonyl-Rb1, malonyl-Rd, characterize the P.quinquefolius root cluster. 4. Conclusions Optimized chromatographic profiling making use of the Evaporative Light Scattering detection combined with conventional multivariate analysis resulted to be a powerful technique to distinguish among the roots and aerial parts of the model species, P. ginseng and P. quinquefolius. In a quality control setting, having a suitable and reliable methodology to identify botanical raw materials, will help to support and guarantee the quality of herbal products providing assurance of their safety and efficacy. In this regard, this approach could be used to monitor the quality of the ginseng herbal preparations for pharmaceutical market and of the food supplements in order to detect and prevent accidental and economic adulterations. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.fitote.2014.12.013. References [1] Hostettmann K, Marston A. Saponins-Chemistry and Pharmacology of Natural ProductsCambridge: Cambridge University Press; 1995. [2] Tanaka O, Kasai R. In: Herz W, Grisebach H, Kirby GW, Tamm C, editors. Progress in the Chemistry of Organic Natural Products, vol. 46. New York, NY: Springer; 1984.
91
[3] Kitagawa I, Taniyama T, Yoshikawa M, Ikenishi Y, Nakagawa Y. Chemical studies on crude drug processing. VI.: chemical structures of malonylginsenosides Rb1, Rb2, Rc, and Rd isolated from the root of P. ginseng C. A. MEYER. Chem Pharm Bull 1989;37:2961–70. [4] Yamaguchi H, Kasai R, Matsuura H, Tanaka O, Fuwa T. High-performance liquid chromatographic analysis of acidic saponins of ginseng and related plants. Chem Pharm Bull 1988;36:3468–73. [5] Sengupta S, Toh SA, Sellers LA, Skepper JN, Koolwijk P, Leung HW, Yeung HW, Wong RN, Sasisekharan R, Fan FP. Modulating angiogenesis: the yin and the yang in ginseng. Circulation 2004;110:1219–25. [6] Sievenpiper JL, Arnason JT, Leiter LA, Vuksan V. Decreasing, null and increasing effects of eight popular types of ginseng on acute postprandial glycemic indices in healthy humans: the role of ginsenosides. J Am Coll Nutr 2004;23:248–58. [7] US Department Of Health And Human Services Food and Drug Administration, Center for Drug Evaluation and Research. Guidance for industry: botanical drug products. Silver: Spring; 2004. [8] Medicines Agency European. Guidelines on specifications: test procedures and acceptance criteria for herbal substances. Herbal preparations and herbal medicinal products/traditional herbal medical products; 2011 London. http://www.emea.europa.eu. [9] Food State, Drug Administration of China. Technical requirements for the development of fingerprints of TCM injections SFDA; 2000 Beijing. http://www.cmjournal.org/content/6/1/9. [10] Unpublished data. [11] Harnly J, Chen P, Harrington PB. Probability of identification: adulteration of American ginseng with Asian ginseng. J AOAC Int 2013:961258–65. [12] Liu Y, Li J, He J, Abliz Z, Qu J, Yu S. Identification of new trace triterpenoid samponins from the roots of Panax notoginseng by high-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry. Rapid Commun Mass Spectrom 2009;23:667–79. [13] Sun BS, Xu MY, Li Z, Wang YB, Sung CK. UPLC-Q-TOF-MS/MS analysis for steaming times-dependent profiling of steamed Panax quiquefolius and its ginsenosides transformations induced by repetitious steaming. J Ginseng Res 2012;36:277–90. [14] Kim N, Kim K, Choi BY, Lee D, Shin YS, Bang KH, Cha SW, Lee JW, Choi HK, Jang DS, Lee D. Metabolomic approach for age discrimination of P. ginseng using UPLC-Q-Tof MS. J Agric Food Chem 2011;59:10435–41. [15] Fuzzati N, Gabetta B, Jayakar K, Pace R, Ramaschi G, Villa F. Determination of ginsenosides in P. ginseng roots by liquid chromatography with evaporative light-scattering detection. J AOAC Int 2000;83:820–9. [16] Fuzzati N, Gabetta B, Jayakar K, Pace R, Peterlongo F. Liquid chromatography–electrospray mass spectrometric identification of ginsenosides in P. ginseng roots. J Chrom A 1999;854:69–79. [17] Fuzzati N. Analysis methods of ginsenosides. J Chrom B 2004;812:119–33. [18] Watkins SM, German JB. Toward the implementation of metabolomic assessments of human health and nutrition. Curr Opin Biotech 2002;13: 512–6. [19] Weckwerth W. Metabolomics in systems biology. Annu Rev Plant Biol 2003;54:669–89. [20] Robertson DG. Metabonomics in toxicology: a review. Toxicol Sci 2005;85: 809–22. [21] Shellie RA, Marriott PJ, Hule CW. Comprehensive two-dimensional gas chromatography (GC × GC) and GC × GC-quadrupole MS analysis of Asian and American ginseng. J Sep Sci 2003;26:1185–92. [22] Xie GX, Ni Y, Su MM, Zhang YY, Zhao AH, Gao XF, Liu Z, Xiao PG, Jia W. Application of ultra- performance LC-TOF MS metabolite profiling techniques to the analysis of medicinal Panax herbs. Metabolomics 2008;4:248–60. [23] Lee AR, Gautam M, Kim J, Shin WJ, Choi MS, Bang YS. A multi analytical approach for determining the geographical origin of ginseng using strontium isotopes, multielements, and 1H NMR analysis. J Agric Food Chem 2011;59:8560–7.