色谱  2015, Vol. 33 Issue (5): 514-521   PDF (931 KB)    
扩展功能
加入收藏夹
复制引文信息
加入引用管理器
Email Alert
RSS
本文作者相关文章
张翠英
陈士林
董梁
超高效液相色谱法结合化学计量学分析评价4种商品人参的质量
张翠英1,2, 陈士林1 , 董梁1    
1. 中国医学科学院药用植物研究所, 北京 100193;
2. 中国中医科学院广安门医院, 北京 100053
收稿日期：2015-01-14
基金项目：National Science and Technology Support Project (2006BAI09B02); National Natural Science Foundation (81202932).
通讯作者：陈士林,E-mail:slchen@implad.ac.cn
摘要：建立了快速、灵敏、准确的超高效液相色谱方法,用来分析4种商品人参(人参、红参、人参叶、人参须)中12种人参皂苷的含量,并用化学计量学方法评价了商品人参的质量。采用ACQUITY UPLCTM BEH C18色谱柱(50 mm×2.1 mm, 1.7 μm),以乙腈-水为流动相进行梯度洗脱。对所建立的测定12种人参皂苷的UPLC方法进行了线性方程、准确度、重复性、回收率等方法学考察。采用聚类分析和主成分分析的化学计量学方法对4种商品人参进行分析,评价了其质量。结果表明聚类分析和主成分分析2种化学计量学方法非常适合大样本、多成分的中药材质量分析。
关键词： 超高效液相色谱     聚类分析     主成分分析     人参皂苷     人参    
Analysis and assessment of four commercial products of Asian ginseng by ultra-performance liquid chromatography and chemometric analysis
ZHANG Cuiying1,2, CHEN Shilin1 , DONG Liang1    
1. Institute of Medicinal Plant Development, Chinese Academy of Medical Sciences, Beijing 100193, China;
2. Guang'anmen Hospital, China Academy of Chinese Medical Sciences, Beijing 100053, China
Received: 2015-01-14
Abstract: Ginseng Radix et Rhizoma (GRR, also named as white ginseng), Ginseng Radix et Rhizoma Rubra (GRR Rubra, also named as red ginseng), Ginseng Folium (GF) and Ginseng Rootlet (GR) products from Asian ginseng, one of the well-known Chinese traditional medicine for thousands of years, are now widely used around the world. Thus the comprehensive quality control is of paramount concern basing on the contents of the bioactive ginsenosides. A rapid, sensitive and reliable method of ultra-performance liquid chromatography coupled with a photodiode array detection (UPLC-PAD) was developed for the quantitative analysis of the 12 ginsenosides in the four commercial ginseng products of Asian ginseng. The chromatography was performed on an ACQUITY UPLCTM BEH C18 column using a gradient elution with acetonitrile/water as the mobile phases. Method validation including calibration curves, accuracies, precisions, repeatabilities and recoveries was investigated. The contents of the 12 ginsenosides were determined in 20 GRR, 4 GF, 4 GR and 11 GRR Rubra samples. To evaluate the sample quality, chemometric methods including hierarchical cluster analysis (HCA) and principal components analysis (PCA) were engaged in evaluating the GRR, GRR Rubra, GF and GR products from Asian ginseng. The results showed that HCA and PCA can be considered as the attractive chemometric techniques in situations where high sample throughput and multiple analytes are required.
Key words: ultra-performance liquid chromatography (UPLC)     hierarchical cluster analysis (HCA)     principal components analysis (PCA)     ginsenosides     Asian ginseng    
Asian ginseng (or Panax ginseng),derived from Panax ginseng C. A. Mey. (Family Araliaceae),has been used as one of the well-known Chinese traditional medicines in China,Korea and Japan for thousands of years,and is now widely used around the world. There are various types of commercial ginseng products such as Ginseng Radix et Rhizoma (GRR,also named as white ginseng,fresh ginseng dried by sun),Ginseng Radix et Rhizoma Rubra (GRR Rubra,also named as red ginseng,fresh ginseng steamed for 2-3 h at 95-100 ℃ and then dried),Ginseng Folium (GF) and Ginseng Rootlet (GR). Panax ginseng roots such as GRR and GRR Rubra are considered to be the main ginseng commercial products used for medicinal and food additive purposes. There are many reports about the chemical components and pharmacological activities of GRR and GRR Rubra [1, 2, 3, 4],however,there are only a few publications on GF and GR [5, 6, 7]. Ginsenosides,the major components,are responsible for the various pharmacological activities of Panax ginseng. More than 30 ginsenosides are isolated and characterized in every part of Panax ginseng [4]. Ginsenosides (G)-Rg1,-Re,-Rb1,-Rb2,-Rc and -Rd have been recognized as the main active ingredients in Panax ginseng [8, 9, 10]. However,only G-Rg1,-Re,-Rb1 have been designated as quality-control markers for GRR,GRR Rubra,and GF in the Chinese pharmacopoeia [11]. Consistent quality cannot be guaranteed based on the contents of only a few of the ginsenosides. For example,the contents of G-Re or -Rb1 in GR are higher than those in GRR [9, 12]. GRR and GR can mediate anti-diabetic effects whose mechanisms are different. GRR can improve hyperglycemia in KKAy mice to low blood glucose levels while GR can upregulate adipocytic PPAR- γ protein expression [7]. Therefore,the synergistic action among multiple active ingredients plays an important role in ginseng bioactivities. On the other hand,the holistic quality assessment of multiple constituents is credible and reasonable. Traditional high-performance liquid chromatography (HPLC) coupled with various detectors has been employed as the most common quantitative analysis technique for ginsenosides [13]. However,the quantitative analysis of the ginsenosides by traditional HPLC generally requires more than one hour per sample run because of their similar structures [14, 15]. The main advantage of ultra-performance liquid chromatography (UPLC) was particularly a significant reduction of analysis time with improved sensitivity and resolution compared with HPLC. From this point of view,UPLC could play a significant role in the future of liquid chromatography and be more convenient for complex analytical determination of pharmaceutical preparations such as ginsenosides in Panax ginseng,Panax notoginseng or Panax quinquefolius [16]. Multivariate statistical analysis such as hierarchical cluster analysis (HCA) and principal components analysis (PCA) are engaged in sorting medicine herbals into groups and evaluating their holistic quality [17]. In this study,12 ginsenosides were investigated by rapid,high-throughput UPLC-PDA method in four commercial ginseng products including GRR,GRR Rubra,GF and GR. At the same time,chemometric methods such as HCA and PCA were applied to the holistic evaluation of different commercial ginseng products.

1 Experimental
1.1 Plant materials
Sixteen GRR samples (S01-S16) were collected in ginseng GLP planting bases from Liaoning,Jilin and Heilongjiang provinces in China. Panax ginseng samples from Tonghua,Jilin were detached into three parts including GRR samples (S17-S20),GR samples (S21-S24) and GF samples (S25-S28). Samples mentioned above were sun-dried. Eleven GRR Rubra samples (S29-S39) were purchased from the drug stores in China and Korea. All the ginseng samples (S01-S39) were identified as Panax ginseng by Professor LIN Yulin (Institute of Medicinal Plant Development,Chinese Academy of Medical Sciences and Peking Union Medical College) and the detail information are listed in Table 1. The voucher specimens were deposited in the Herbarium of Institute of Medicinal Plant Development,Chinese Academy of Medical Sciences and Peking Union Medical College.
Table 1 Raw herbs used in this work
1.2 Reagents and standards
G-Rg1,-Re,-Rf,-Rg2,-Rg3,-Rb1,-Rb2,-Rb3,-Rc and -Rd were provided by National Institute for the Control of Pharmaceutical and Biological Products (Beijing,China),whereas G-Rh1 and G-Rh2 were purchased from Shanghai Tau Biotech Co.,Ltd (Shanghai,China). Acetonitrile (HPLC grade) was obtained from Merck Company (Darmstadt,Germany). Water was prepared from Millipore water purification system (Millipore,USA). Solvents were filtered through 0.22 μ m membranes (WATERS Co.,USA) before use. Other solvents were of analytical grade (Beijing Chemical Factory,Beijing,China).
1.3 Apparatus
All the experiments were performed on an ACQUITY UPLCTM system (WATERS Co.,USA),which was equipped with a binary solvent delivery system,a sample manager,a column oven,and a photodiode array detector (PDA). The data were analyzed with WATERS Empower Chemistation.
1.4 Chromatographic conditions
The chromatographic separation was carried out on an ACQUITY UPLCTM BEH C18 column (50 mm×2.1 mm,1.7 μ m) (WATERS Co.,USA). The column was maintained at 30 ℃,and the detection wavelength was set at 203 nm. The mobile phases consisted of acetonitrile as A and water as B. The separation was performed by gradient elution: 0-3 min,19%A; 3-4 min,19%A-21%A; 4-5 min,21%A-26%A; 5-9 min,26%A-27%A; 9-12 min,27%A-32%A; 12-15 min,32%A-43%A; 15-18 min,43%A-60%A; 18-20 min,60%A-100%A. The flow rate was 0.3 mL/min. The injection volume of the test samples and standard solutions was 2 μ L. Each washing cycle after the analysis consisted of 200 μ L strong solvent (acetonitrile-water,8 ∶ 2,v/v) and 600 μ L weak solvent (acetonitrile-water,2 ∶ 8,v/v).
1.5 Sample preparation
The powdered GRR,GR and GRR Rubra samples (1.0 g) were accurately weighed and macerated with 50 mL methanol in a measuring flask for 10 min,then extracted in an ultrasonic bath (500 W) for 30 min. The methanol extracts (25 mL),after removed the solvent in vacuo,were dissolved in 10 mL methanol and filtered through a membrane filter (with pore size of 0.22 μ m) prior to injection. The powdered GF samples (1.0 g) were accurately weighed and macerated with 50 mL methanol in a measuring flask for 10 min,and were extracted in an ultrasonic bath (500 W) for 30 min. The methanol extracts (15 mL),after removed the solvent in vacuo,were dissolved in 10 mL water and loaded on a D101 resin column (15 cm×1.5 cm). Firstly,50 mL water and 50 mL 20% (v/v) ethanol washing solution were discarded. Secondly,80 mL 80% (v/v) ethanol washing solution were collected and the solvent removed in vacuo. Finally,the residues were dissolved in 10 mL methanol and filtered through a membrane filter (with pore size of 0.22 μ m) prior to injection.
1.6 Multivariate statistical analysis
HCA is one of the multivariate analysis methods used for finding relatively homogeneous clusters to classify samples into groups. PCA is a sophisticated technique widely used for reducing the dimensions of multivariate problems. In this study,HCA and PCA of samples were performed with SPSS 16.0 software (SPSS,USA).
2 Results and discussion
2.1 Chromatographic conditions
To achieve better separation of the adjacent peaks in a shorter analysis time,the chromatographic conditions were optimized first. The chromatographic condition was transferred from the previous research by HPLC-UPLC method conversion calculator software [18]. Further investigation was taken to obtain the baseline separation of the target ginsenosides,especially for G-Rg1 and -Re,-Rb2 and -Rb3. Under the optimized conditions [19],12 ginsenosides and other components in the four commercial ginseng samples were well separated,as shown in the representative chromatograms (Fig. 1). The use of the UPLC-PDA method produced a significant increase in peak capacity,which resulted in a better chromatographic resolution and increased sensitivity.
Fig.1 UPLC chromatograms of (a) mixed standards, (b) GRR,(c) GR,(d) GF and (e) GRR Rubra Peaks: 1. G-Rg1; 2. G-Re; 3. G-Rf; 4. G-Rh1; 5. G-Rg2; 6. G-Rb1; 7. G-Rc; 8. G-Rb2; 9. G-Rb3; 10. G-Rd; 11. G-Rg3; 12. G-Rh2.
2.2 Method validation
Twelve ginsenosides from ginseng samples were identified by comparing the retention times and online UV spectra with the corresponding standard data. The linearity was evaluated by plotting the peak area of each ginsenoside against its concentration (six levels). The equations and correlation coefficients (R2) corresponding to the linear regression curves were listed in Table 2. The LODs and LOQs were defined as the concentrations with signals three and ten times higher than the noise level,respectively.
Table 2 Linear regressions,linear ranges,R2,LODs and LOQs of the 12 ginsenosides

The intra-day and inter-day (three days) accuracies showed precisions with relative standard deviations (RSDs) ranging from 1.01% to 2.48% . The repeatability was measured with S06 for 6 parallel samples. The stability was tested at 0,4,12,16,20 and 24 h,respectively. The RSDs for the repeatability and stability were within 3.70% and 3.75% ,respectively. The sample S05 spiked with the 12 standards was extracted,processed and quantified in accordance with the sample preparation and chromatographic analysis methods described above. The recoveries of the analytes were determined at three spiked levels (low,medium and high). The average recoveries of the 12 ginsenosides in the spiked samples were found within the range of 93.3% to 102.6% ,with RSDs between 0.29% and 3.62% .

2.3 Comparison of four different commodities of Asian ginseng
The contents of 12 ginsenosides were investigated and varied widely in the four ginseng commercial products (Table 3). G-Rh2 have not been detected in the 39 samples of four ginseng commodities. G-Rg3 and -Rh1 only existed as two characteristic components in GRR Rubra. From Table 3,it is concluded that the total contents of ginsenosides in GF and GR were higher than those in GRR and GRR Rubra. It should be pointed out that the content of G-Re in GR was about four or five times higher than that of G-Rg1. G-Rc,-Rb1 and -Rb2,also the main components,were more than 5.5 mg/g in GR. The content of G-Re in GF was about two times higher than that of G-Rg1. G-Rb1,-Rb2 and -Rd were the main components besides G-Re and -Rg1. Therefore,G-Re,-Rg1,-Rb1,-Rb2,-Rc and -Rd were the major components in Asian ginseng for quality evaluation.
Table 3 Contents of the 12 ginsenosides in four commercial products of Asian ginseng (n=3)
2.4 Quality assessment and classification of four commercial ginseng products based on HCA and PCA
The contents of the 11 ginsenosides in Table 3 were imported into SPSS 16.0 software for analysis. In this analysis,the contents of the 11 ginsenosides determined for the 39 ginseng samples formed an 11×39 matrix. The similarities and classification of HCA and PCA among the 39 samples were calculated with the SPSS software,which revealed the relationships among the samples. HCA is a statistical method for finding relatively homogeneous clusters in cases based on measured characteristics. The results of the HCA were shown in Fig. 2. The between-groups linkage method as the amalgamation rule and squared euclidean distance as metric were used to establish clusters. The samples could be divided into two levels and three clusters: 20 GRR samples and 11 GRR Rubra samples in Cluster A,belonging to Level Ⅰ; four GR samples and four GF samples in Clusters B and C respectively,both belonging to Level Ⅱ. It could be seen that the rescaled distance between Clusters B and A (or Clusters C and A) was 25 (arbitrary units),whereas the rescaled distance between Clusters B and C was 6.7. These distances indicated that the qualities of Clusters B and C were more similar to each other than with those of Cluster A. From the plot,all the samples from GRR and GRR Rubra were clustered in Level Ⅰ (Cluster A),showing that the qualities of these samples were homogenous. GR samples (Cluster B) and GF (Cluster C) samples were clustered in Level Ⅱ,which indicated that their ginsenosides varied only a little.
Fig.2 UPLC chromatograms of (a) mixed standards, (b) GRR,(c) GR,(d) GF and (e) GRR Rubra

Unsupervised PCA was also employed for classification with the contents of ginsenosides as input data. On the basis of eigenvalues＞1,the first three principal components PC1,PC2 and PC3 were often used to provide a visual aid for identifying inhomogeneity in the data sets. The score plot of the first three principal components showed that the samples could be classified into four groups,indicating that GRR,GR,GF and GRR Rubra samples were indeed different in the levels or occurrence of their components. Group Ⅰ-Ⅳ were samples from GRR,GF,GR and GRR Rubra,respectively (Fig. 3). It was interesting to find that the samples of Group Ⅰ and Group Ⅱ were far apart from Group Ⅲ and Ⅳ,while the samples of Group Ⅲ were near to Group Ⅳ. It suggested that the commodities of GRR and GRR Rubra should have a closer relationship and similar quality among samples from Group Ⅲ and Ⅳ. The results were in accord with the above-mentioned HCA. The results of HCA and PCA could validate each other and provide more references for the quality evaluation of ginseng samples. From the above results of these samples,unsupervised PCA classification with original data obtained from the 12 ginsenosides alone could only discriminate the samples according to different commodities.

Fig.3 PCA score plot of commercial Asian ginseng samples
2.5 Quality assessment and classification of GRR Rubra based on HCA and PCA
GRR Rubra is processed by steaming fresh ginseng at 95-100 ℃ for 2-3 h and then drying. The HCA results of the GRR Rubra samples are shown in Fig. 4,and the samples could be divided into two levels and three clusters: two samples (S29 and S31) were in Cluster C,belonging to Level Ⅰ; one sample (S33) and eight samples in Clusters A and B respectively,both belonging to Level Ⅱ. Three samples including two samples (S29 and S31) with many rootlet and one sample (S33) with blackish brown-color were considerably dispersed from their expected clusters,indicating that the holistic qualities of these samples were different from many GRR Rubra samples. Unsupervised PCA results of GRR Rubra samples are shown in Fig. 5. It was evident that the 11 commercial GRR Rubra samples were clearly clustered into three groups,and the eight samples clustered except for the three samples (S29,31 and S33). The results of PCA were consistent with those of HCA.
Fig.4 Dendrogram of clustering the GRR Rubra samples

Fig.5 PCA score plot of GRR Rubra samples
3 Conclusions
A rapid UPLC-PDA method was developed to shorten the analytical time for both the quantitative determination of the 12 bioactive ginsenosides and the quality evaluation of four commercial ginseng products,including GRR,GRR Rubra,GF and GR. Furthermore,by two chemometric methods,the samples could be efficiently classified according to their different medicinal parts and processing methods. With the chemometric methods,UPLC-PDA is rapid,simple and sensitive for the quantitative analysis and the quality evaluation,and well suitable for the quality evaluation of commercial ginseng products.
参考文献
[1]Kim K, Kim H Y. J Ethnopharmacol, 2008, 120(2): 190  
[2]Qi L W, Wang C Z, Yuan C S. Nat Prod Rep, 2011, 28(3): 467  
[3]Wang X M, Sakuma T, Asafu-Adjaye E, et al. Anal Chem, 1999, 71(8): 1579  
[4]Jia L, Zhao Y Q. Curr Med Chem, 2009, 16(19): 2475  
[5]Dou D Q, Chen Y J, Ma Z Z, et al. Chinese Journal of Medicinal Chemistry, 1996, 6(1): 54
[6]Wu Q, Chen X C, Du S Y. Journal of Beijing University of Traditional Chinese Medicine, 2006, 29(5): 344
[7]Chung S H, Choi C G, Park S H. Arch Pharm Res, 2001, 24(3): 214  
[8]Kim S N, Ha Y W, Shin H, et al. J Pharm Biomed Anal, 2007, 45(1): 164  
[9]Shi W, Wang Y T, Li J, et al. Food Chem, 2007, 102: 664  
[10]Zhang H M, Li S L, Zhang H, et al. J Pharm Biomed Anal, 2012, 62: 258  
[11]China Pharmacopoeia Committee. Pharmacopoeia of the People's Republic of China. Beijing: China Medical Science Press, 2010: 8, 9, 143
[12]Yan H, Zhao S Y, Liu Y J, et al. China Journal of Experimental Traditional Medical Formulae, 2013, 19(6): 105
[13]Qian Z M, Lu J, Gao Q P, et al. J Chromatogr A, 2009, 1216: 3825  
[14]Zhang K, Wang X, Ding L, et al. Chemical Research in Chinese Universities (English edition), 2008, 24: 707
[15]Wang A, Wang C Z, Wu J A, et al. Phytochem Anal, 2005, 16(4): 272  
[16]Li L, Luo G A, Liang Q L, et al. J Pharm Biomed Anal, 2010, 52(1): 66  
[17]Pereira A C, Reis M S, Saraiva P M, et al. Anal Chim Acta, 2010, 660(1): 8
[18]Dong L, Zhang C Y, Chen S L. Acta Pharmaceutica Sinica, 2011, 46(2): 198
[19]Zhang C Y, Dong L, Chen S L, et al. Acta Pharmaceutica Sinica, 2010, 45(10): 1296