- Email: [email protected]
Species differentiation and quality assessment of Radix Paeoniae Rubra (Chi-shao) by means of high-performance liquid chromatographic fingerprint
Journal of Chromatography A, 1216 (2009) 2163–2168
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Species differentiation and quality assessment of Radix Paeoniae Rubra (Chi-shao) by means of high-performance liquid chromatographic fingerprint Shunjun Xu a , Liu Yang b,∗ , Runtao Tian c , Zhengtao Wang d , Zhijun Liu a , Peishan Xie c,∗∗ , Qianru Feng b a
School of Renewable Natural Resources, LSU Agricultural Center, Louisiana State University, Baton Rouge 70803, USA Second Affiliated Hospital, Guangzhou University of Chinese Medicine, Guangzhou 510120, China c Key Laboratory of Standardization of Chinese Medicines of Ministry of Education, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China d Zhuhai Chromap Institute of Herbal Medicine Research, Zhuhai 519085, China b
a r t i c l e
i n f o
Article history: Available online 3 May 2008 Keywords: High-performance liquid chromatography Fingerprint analysis Radix Paeoniae Rubra Paeonia lactiflora Paeonia veitchii
a b s t r a c t There are two species under the monograph of Radix Paeoniae Rubrae (“Chi-shao” in Chinese) in Chinese Pharmacopoeia 2005 edition-Paeonia lactiflora Pallas and Paeonia veitchii Lynch. Due to different species and growing conditions, there are significant chemical differences between the two species, which may result in the improper clinical usage under the same name. Chemical pattern expressed by high performance liquid chromatographic (HPLC) fingerprint analysis can play an important role in species differentiation and quality control of Radix Paeoniae Rubra. In the present work, HPLC fingerprints of two kinds of Radix Paeoniae Rubra have been established and analysed with chemometric methods including similarity evaluation and principal component analysis. Both of the fingerprint common patterns of the two species comprise 13 characteristic peaks, nine of which were common peaks of the two species. However, significant differences between the roots of P. veitchii and P. lactiflora exist not only in the content of certain constituents, especially phenolic acids but also in peak-to-peak ratios expressed by the fingerprint patterns. According to the recent pharmacological studies on polyphenolic constituents, root originating from P. veitchii may possess better efficacy and quality than that from P. lactiflora. Our research reveals that further pharmacological investigation is very necessary to determine whether the two species should be embodied under the same monograph in Chinese Pharmacopoeia. © 2008 Elsevier B.V. All rights reserved.
1. Introduction The entity and content of bioactive components of traditional Chinese medicine (TCM) are always highly variable, which depend on the species, geographical origins, cultivation and harvesting methods and post-harvesting formulation processes [1,2]. Chromatographic fingerprint analysis adequately caters for the complex characteristics of TCM, so it has been used for species differentiation, evaluation of quality and ensuring the consistency and stability of TCM [3–6]. Radix Paeoniae Rubra (“Chi-shao” in Chinese) is one of the most widely used Chinese herbal drugs. According to Chinese Pharmacopoeia 2005 edition, the herbal drug originates from the dried roots of two species, Paeonia lactiflora Pallas and Paeonia veitchii Lynch [7]. P. lactiflora naturally grows in Inner Mongolia
∗ Corresponding author. Tel.: +86 20 81887233x30908; fax: +86 20 81867705. ∗∗ Corresponding author. Tel.: +86 756 3326296; fax: +86 756 3326961. E-mail addresses: [email protected] (L. Yang), [email protected] (P. Xie). 0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.04.064
autonomous region, Heilongjiang, Jilin and Liaoning provinces, and P. veitchii is mainly distributed in Sichuan, Shaanxi and Qinghai provinces. Different species and growing conditions may cause significant difference of chemical patterns between P. lactiflora Pallas and P. veitchii Lynch, however, they have been used under the same name in clinical prescriptions for ages. Moreover, although paeoniflorin is a major bioactive component in Chi-shao, it is contrary to TCM principle to designate paeoniflorin as one and only marker compound for quality evaluation of both species in Chinese Pharmacopoeia, because the therapeutic efficacy of herbal drug is always attributed to its all kinds of bioactive components but not to any single ingredient according to TCM’s principle. These contradictions reveal that a comprehensive study of the chemical patterns of the two species would be valuable to help understand their activity and usage. Unfortunately, previous investigations on quality control of Chi-shao mostly rested on determination of individual or several analytical compounds [8–11], and only a few studies focused on the discrimination between Chi-shao and its adulterants using genetic fingerprinting or infrared spectroscopy [12–14]. In the present study, by combining the chemometric methods such as similarity evaluation and principal component analysis
2164
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
(PCA), we used HPLC-photodiode array detection to develop a specific, practical and valid chromatographic fingerprint analysis approach for quality assessment and species differentiation of P. lactiflora Pallas and P. veitchii Lynch. 2. Materials and methods 2.1. Apparatus and reagents HPLC analyses were performed on a Waters 600 series HPLC system together with column compartment and photodiode array detector (PDA). HPLC-grade acetonitrile was purchased from Fisher Chemicals (Fair lawn, New Jersey, USA). HPLC-grade formic acid was purchased from Tedia Company, Inc. (Fairfield, OH, USA). The water used in the experiment was doubly distilled in the laboratory. Other chemicals and solvents were of analytical grade. 2.2. Plant materials Nineteen batches of Chi-shao from eleven provinces of China were collected (Table 1). These herbal samples were authenticated by Professor Mian Zhang (China Pharmaceutical University, Nanjing, China). Voucher specimens are stored at Guangdong provincial hospital of TCM. 2.3. Reference compounds Albiflorin and 1,2,3,4,6-penta-O-galloyl--d-glucose were provided by Shanghai University of Traditional Chinese Medicine, Shanghai, China. Oxypaeniflorin was provided by Chromap Institute of Herbal Medicine Research (Zhuhai, China). Gallic acid, (+)-catechin, paeoniflorin and paeonol were purchased from the National Institute for the Control of Pharmaceutical and Biological Products, Tiantanxili No. 2, Beijing, China. Benzoylpaeniflorin and methyl gallate were isolated and purified by our own laboratory. These reference compounds were dissolved in methanol and then injected into HPLC after filtration with a 0.22 m filter (PTFE). 2.4. Sample preparation All samples were dried at 60 ◦ C for 12 h before use. Each dried sample was ground to a fine powder (40 mesh, 450 m I.D.) using a pulverizer. An aliquot of 2 g was accurately weighed and macerated
with 20 ml 50% aqueous methanol (v/v) at room temperature for 30 min. Sample was thrice extracted under reflux for 45 min each (50% methanol volume: 20, 15 and 15 ml). The sample solution was filtered through Whatman 1 filter paper. The filtrate was combined and decanted into a 50 ml volumetric flask, and then 50% aqueous methanol was added to make up to the scale. The final solution was passed through a 0.22 m membrane prior to use. An aliquot of 20 l of each sample solution was injected into the HPLC system for analysis. 2.5. Chromatographic condition The HPLC fingerprinting analysis was carried out on a Waters Symmetry C18 column (250 mm × 4.6 mm I.D., 5 m). A binary gradient elution system composed of acetonitrile as solvent A and 0.1% phosphoric acid in water as solvent B was applied for the fingerprint analysis with the gradient elution as follows: 0–5 min, 10–15% A; 5–25 min, 15–22% A; 25–45 min, 22–70% A; 45–46 min, 70–80% A; 46–50 min, 80% A. The mobile phase flow rate was 0.8 ml/min, and column temperature was maintained at 25 ◦ C. The PDA detector was set at 254 nm, and the on-line UV spectra were recorded in the range of 195–400 nm. 2.6. Data analysis of chromatogram Data analysis was performed by a pattern recognition program recommended by the Chinese Pharmacopoeia committee and the complementary software developed by Zhuhai Chromap Institute of Herbal Medicine Research that allows for a statistical evaluation of chromatographic patterns. In the present work, similarity evaluation and principal component analysis were used to assess HPLC fingerprints of all samples. 3. Results and discussion 3.1. Optimization of extraction conditions Considering a variety of components with relatively high polarity in Chi-shao, aqueous methanol was the preferred choice of as extraction solvent to effectively extract overall compounds in the herb. Related extraction conditions were examined and evaluated, which involved the following experimental factors and corresponding levels: extraction method (ultrasonication or reflux), extraction
Table 1 A summary of tested Chi-shao samples No.
Collection location
Collection time
Species
Voucher specimens nrs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Ganzi, Sichuan, China Kunming, Yunnan, China Nanning, Guangxi, China Shanghai, China Guangzhou, Guangdong, China Duolun, Neimeng, China Bozhou, Anhui, China Jiling, China Changchun, Jiling, China Hongkong, China Hongkong, China Taibei, Taiwan, China Chendu, Sichuan, China Zhongjiang, Sichuan, China Xichang, Sichuan, China Nanchang, Jiangxi, China Bozhou, Anhui, China Bozhou, Anhui, China Ganzi, Sichuan, China
November, 2004 April, 2004 March, 2003 March, 2003 April, 2005 April, 2005 March, 2005 April, 2003 March, 2003 December, 2004 May, 2005 April, 2004 December, 2004 December, 2004 November, 2004 April, 2005 October, 2005 November, 2005 October, 2005
P. veitchii P. veitchii P. veitchii P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. veitchii P. veitchii P. veitchii P. lactiflora P. lactiflora P. lactiflora P. veitchii
PV20041111 N/Aa N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A PL20051013 PL20051107 PV20051022
a
Not available.
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
2165
Fig. 1. Average artificial HPLC fingerprint common pattern of P. lactiflora based on 12 samples.
repetitions (1, 2, 3 or 4 times), methanol concentration (0, 50, 100%, v/v), and extraction time (0.5, 0.75, 1 h). The number and areas of characteristic peaks of HPLC fingerprint were regard as a criterion for the selection of optimal conditions. As a result, the extraction method of Chi-shao is presented as shown in Section 2.4. 3.2. Optimization of HPLC conditions It is very difficult to separate chemical constituents of Radix Paeoniae Rubra because of their similar physicochemical properties, so chromatographic parameters were optimized to achieve a higher separation quality of the fingerprint and if possible in a reduced analysis time. Waters Symmetry C18, Agilent XDB-C18 , Agilent SB-C18 , Agilent XDB C8, Capcell pak C18 MG, Hypersil ODS and Kromasil ODS columns were all tested, and a Waters Symmetry C18 column finally was selected for the analysis. Methanol–water and acetonitrile–water as mobile phase were compared, and acetonitrile–water could achieve better resolutions. The retention behavior of phenolic acids and galloyl glucoses in Radix Paeoniae Rubra on the reversed-phase HPLC column was significantly affected by the pH of the mobile phase, so different acids including formic acid, acetic acid, phosphoric acid, and amount of acid added into mobile phase were also tested. As a result, the addition of 0.1% (v/v) phosphoric acid could achieve a good baseline and satisfactory resolution of these major peaks. The detection wavelength was set at 230 nm because most of the characteristic components in Chishao have satisfactory sensitivity at this UV wavelength. Column temperature, flow rate and the slope of the gradient are also individually examined at several levels. The best condition for a given factor is selected based on the peak capacity observed. All best factor levels are then combined. Finally, the optimal HPLC condition is shown in Section 2.5. 3.3. Validation of methodology Method precision and reproducibility were evaluated by the analysis of five injections of the same sample solution and five replicates of the same sample, respectively. The relative standard deviations (RSD) of retention time (tR ) and peak area (PA) of characteristic peaks in the precision test were found in the range of 0.4–3.1%, whereas in the reproducibility test the RSDs of tR and PA were also below 0.7 and 3.3%, respectively. Meanwhile, the analyses of the same sample solution at intraday different time point (0, 2, 4, 8, 24 and 48 h) were evaluated as the stability test of sample solution, and the RSDs of tR and PA were less than 0.4 and 3.2%, respectively. All results indicated that the method of HPLC fingerprint analysis was valid and satisfactory.
3.4. HPLC fingerprint analysis of Chi-shao As shown in Figs. 1 and 2, Table 2, there exist a total of 16 characteristic peaks in the two HPLC fingerprint common patterns of the two Paeonia species, of which nine characteristic peaks were assigned by comparing the UV spectra and their retention time with those of the reference compounds, i.e. peak 2, 3, 4, 5, 6, 7, 12, 15 and 16 were identified as gallic acid, oxypaeoniflorin, (+)-catechin, methyl gallate, albiflorin, paeoniflorin, 1,2,3,4,6-pentagalloyl glucose, benzoylpaeoniflorin and paeonol, respectively. Peak 10 and 11 were tentatively identified as tetragalloyl glucose and galloyl paeoniflorin based on their MS fragmentation behaviors obtained by LC–MS/MS [15]. By visually comparing the HPLC fingerprints of roots of P. lactiflora and P. veitchii, the peaks 6 (albiflorin), 9 (unknown) and 16 (paeonol) could not be detected in the HPLC fingerprints of P. lactiflora, whereas the peaks 3 (oxypaeoniflorin), 4 ((+)-catechin) and 14 (unknown) could not be observed in the HPLC fingerprints of P. veitchii. Moreover, the peak 7 (paeoniflorin) was the highest in the chromatograms of P. lactiflora, while 1,2,3,4,6-pentagalloyl glucose (peak 12) was the most in the roots of P. veitchii. Besides, other gallic acid derivatives in roots of P. veitchii such as gallic acid (peak 2), methyl gallate (peak 5), tetragalloyl glucose (peak 10) and galloylpaeoniflorin (peak 11) were also more abundant than those in roots of P. lactiflora. As shown in Fig. 3, two similarity indices of each chromatogram could be obtained with test solution using the HPLC fingerprint Table 2 The retention time (tR ) and relative peak area of 16 characteristic peaks of HPLC fingerprint common pattern of P. lactiflora and P. veitchii Peak no.
1 2a 3a 4a 5a 6a 7(R)a 8 9 10 11 12a 13 14 15a 16a a
tR (min)
4.35 5.14 12.09 13.25 14.12 17.47 19.82 21.94 23.25 24.15 26.99 28.34 33.52 34.90 40.22 44.70
Relative area
Compound
P. lactiflora
P. veitchii
0.02 0.03 0.04 0.11 0.05 N/A 1.00 0.04 N/A 0.04 0.03 0.17 0.15 0.02 0.04 N/A
0.06 0.14 N/A N/A 0.48 0.04 1.00 0.05 0.04 0.10 0.06 1.20 0.35 N/A 0.06 0.05
Compared with reference compounds.
Unknown Gallic acid Oxypaeoniflorin (+)-Catechin Methyl gallate Albiflorin Paeoniflorin Unknown Unknown 1,2,3,6-Tetragalloyl glucose Galloyl paeoniflorin 1,2,3,4,6-Pentagalloyl glucose Unknown Unknown Benzoylpaeoniflorin Paeonol
2166
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
Fig. 2. Average artificial HPLC fingerprint common pattern of P. veitchii based on seven samples.
common patterns of P. lactiflora and P. veitchii. By comparison with the fingerprint common pattern of P. lactiflora, the similarity index of every fingerprint of P. lactiflora was not less than 0.97 except those of samples 4 and 5 (see Fig. 3a); and the similarity indices of the P. veitchii’s fingerprints were more than 0.90 aside from sample 15’s by comparison with the fingerprint common pattern of P. veitchii (see Fig. 3b). The result of similarity evaluation showed the developed HPLC fingerprint common patterns, as quality assessment models for P. lactiflora and P. veitchii are practical and valid. Furthermore, based on the similarity evaluation, any Chi-shao sample can be easy classified. 3.5. Quality assessment of Chi-shao by principal component analysis (PCA) For quality evaluation of Chi-shao, PCA, a method for feature extraction and dimensionality reduction, was carried out. The PCA computation was implemented by performing singular value decomposition on the data array of the fingerprints, which consisted of a total of 19 × 16 data matrix, each row represented a plant sample and each column contained the values of 16 characteristic peak areas. In this study, in order to amplify the difference between P. lactiflora and P. veitchii and easily discriminate the outliers from normal samples, matrices based on not only the raw data but also standardized (auto-scaling [16]) data were all submitted to PCA analysis. The first two loadings of PCA results were used to generate the projection plot that provides a visual determination of the similarity among the sample fingerprints.
In the PCA result based on raw data, the loading of 1,2,3,4,6pentagalloyl glucose was 0.95, the highest in the first principal component (PC1 ), while the loading of paeoniflorin was 0.87 in the second principal component (PC2 ). In the PCA result based on standardized data, the loadings of 1,2,3,4,6-pentagalloyl glucose and paeoniflorin were 0.72 and −0.52, respectively, in PC1 , and the loadings of gallic acid and galloylpaeoniflorin were −0.59 and 0.61, respectively, in PC2 . As shown Fig. 4a and b, both the two kinds of PCA showed the similar result that 15 batches of Chi-shao samples were classified into two groups: P. lactiflora and P. veitchii except for samples 4, 5, 14 and 15. Moreover, the result of PCA was at largely consistent with that of similarity evaluation. Both of samples 4 and 5 from P. lactiflora displayed a significant difference from other P. lactiflora samples as the results of above two chemometric analyses, so it was reasonable to judge the two samples as outliers. Fig. 5a, for example, clearly showed that these differences between HPLC fingerprint of sample 4 and the common pattern of P. lactiflora consisted in the integrated area values of characteristic peaks as well as peak-to-peak ratios. Samples 14 and 15 were regarded as outliers of P. veitchii according to the result of PCA based on raw data, however, sample 14 was determined as a normal sample in similarity evaluation in view of its high similarity index (over 0.9), and the result of PCA based on standardized data also indicated that the difference of sample 14 was not significant (see Fig. 4b). Thus, only sample 15 was determined as an outlier of P. veitchii. As shown in Fig. 5b, although all of the characteristic peaks of P. veitchii could be observed in the HPLC fingerprint of sample 15, due to very low content of characteristic
Fig. 3. Similarity evaluation of HPLC fingerprint for Chishao. (a) Refer to the fingerprint common pattern of P. lactiflora and (b) refer to the fingerprint common pattern of P. veitchii.
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
2167
Fig. 4. PCA analysis for the fingerprints of Chishao. (a) Projection plot of 1–2 principal component based on the raw data and (b) projection plot of 1–2 principal component based on the standardized data.
Fig. 5. Comparison between chromatograms of outlier samples 4, 15 and common patterns for P. lactiflora and P. veitchii, respectively. (a) Sample 4 (upper) and fingerprint common pattern of P. lactiflora (lower) and (b) sample 15 (upper) and fingerprint common pattern of P. veitchii (lower).
components, especially paeoniflorin, this sample was picked out as an outlier of P. veitchii. 4. Conclusions In this paper, a HPLC fingerprint analysis with chemometric methods was used for the purpose of species differentiation, quality evaluation and the consistency check of Chi-shao collected from different sources. The result showed that the developed HPLC fingerprint analytical method was practical and reliable for above purpose. Obviously, the outliers picked out by means of developed fingerprint analysis method should be rejected to ensure curative effect of the herb and its finished product. Furthermore, 11 of 16 characteristic peaks in the common pattern were identified by comparing with the reference compounds based on their UV spectra and retention time or the data obtained by LC–MS/MS to further characterize the chromatographic fingerprint and contribute to the quality control of Chi-shao. There are significant differences between the roots of P. veitchii and P. lactiflora not only in the content of major common constituents, especially phenolic acids, but also in peak-to-peak ratios expressed by the fingerprint patterns, which raises doubt about the ‘bio-equivalence’ of the two species with the same dose. It is worth noting that root of P. veitchii possesses much more gallic acid derivatives than that of P. lactiflora such as galloyl glucoses, gallic acid, methyl gallate and galloylpaeoniflorin. In addition, content of paeoniflorin is close in the two species as emphasized in
Chinese Pharmacopoeia. Recently, galloyl glucoses isolated from Chishao were found to significantly decrease the plasma endotoxin level [17], which was consistent with the therapeutic efficacy of this herb. Methyl gallate, gallic acid and paonol were also reported as major potent constituents of Chi-shao [18]. Considering these pharmacological investigations, it seems to suggest that roots originating from P. veitchii possess better efficacy and quality than those from P. lactiflora. Hence, further investigation is necessary to determine whether the two herbs should be recorded under the same monograph in Chinese Pharmacopoeia. Acknowledgements The research is funded by Science and Technology Department of Guangdong Province and State Administration of Traditional Chinese Medicine of China. References [1] P. Zou, Y. Hong, H.L. Koh, J. Pharm. Biomed. Anal. 38 (2005) 514. [2] L.W. Yang, D.H. Wu, X. Tang, W. Peng, X.R. Wang, Y. Ma, W.W. Su, J. Chromatogr. A 1070 (2005) 35. [3] P.S. Xie, S.B. Chen, Y.Z. Liang, X.H. Wang, R.T. Tian, R. Upton, J. Chromatogr. A 1112 (2006) 171. [4] W. Jin, R.L. Ge, Q.J. Wei, T.Y. Bao, H.M. Shi, P.F. Tu, J. Chromatogr. A 1132 (2006) 320. [5] S.B. Chen, H.P. Liu, R.T. Tian, D.J. Yang, S.L. Chen, H.X. Xu, A.S.C. Chan, P.S. Xie, J. Chromatogr. A 1121 (2006) 114. [6] G.H. Lu, K. Chan, Y.Z. Liang, K. Leung, C.L. Chan, Z.H. Jiang, Z.Z. Zhao, J. Chromatogr. A 1073 (2005) 383.
2168
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
[7] The State Pharmacopoeia Commission of People’s Republic of China, Pharmacopoeia of People’s Republic of China, vol. 1, Chemical Industry Press, Beijing, 2005. p. 109. [8] Y. He, H.D. Zhao, L.Y. Tang, Z.J. Wang, Q.W. Zhang, Chin J. Chinese Materia Med. 31 (2006) 1070. [9] S.L. Hu, G.L. Fu, X.F. Feng, X.J. Tang, X.R. He, Chin J. Chinese Materia Med. 25 (2000) 10. [10] C.H. Chen, J.Q. Han, Y.Q. Yao, Acta Academiae Medicinae Zhejiang 2 (2006) 30. [11] X.M. Xie, C.Z. Yu, H. Xu, S. Wang, D.L. Wang, L. Zhang, G.X. Chou, Z.T. Wang, Chin J. Chin. Materia Med. 29 (2004) 46.
[12] H.T. Zhou, S.L. Hu, B.L. Guo, X.F. Feng, Y.N. Yan, J.S. Li, Acta Pharm. Sin. 37 (2002) 383. [13] B. Dong, S.Q. Sun, H.T. Zhong, S.L. Hu, Spectrosc. Spectral Anal. 22 (2002) 232. [14] C.G. Cheng, D.T. Li, J.H. Chen, L.C. Kong, Z.F. Cheng, Chin. Pharm. J. 41 (2006) 580. [15] S.J. Xu, L. Yang, X. Zeng, M. Zhang, Z.T. Wang, Rapid Commun. Mass Spectrom. 20 (2006) 3275. [16] D.L. Massart, L. Kaufman, The Interpretation of Analytical Data by the Use of Cluster Analysis, Wiley, New York, 1983, 20. [17] G. Lv, J. Zheng, H. Zhou, Y.M. Zheng, L.X. Wang, G. Wei, M. Huang, D.L. Jiang, L.Z. Wei, Int. Immunopharmacol. 5 (2005) 1007. [18] H.L. Huang, G.M. Wang, B. Xu, Food Sci. 28 (2007) 76.
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Species differentiation and quality assessment of Radix Paeoniae Rubra (Chi-shao) by means of high-performance liquid chromatographic fingerprint Shunjun Xu a , Liu Yang b,∗ , Runtao Tian c , Zhengtao Wang d , Zhijun Liu a , Peishan Xie c,∗∗ , Qianru Feng b a
School of Renewable Natural Resources, LSU Agricultural Center, Louisiana State University, Baton Rouge 70803, USA Second Affiliated Hospital, Guangzhou University of Chinese Medicine, Guangzhou 510120, China c Key Laboratory of Standardization of Chinese Medicines of Ministry of Education, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China d Zhuhai Chromap Institute of Herbal Medicine Research, Zhuhai 519085, China b
a r t i c l e
i n f o
Article history: Available online 3 May 2008 Keywords: High-performance liquid chromatography Fingerprint analysis Radix Paeoniae Rubra Paeonia lactiflora Paeonia veitchii
a b s t r a c t There are two species under the monograph of Radix Paeoniae Rubrae (“Chi-shao” in Chinese) in Chinese Pharmacopoeia 2005 edition-Paeonia lactiflora Pallas and Paeonia veitchii Lynch. Due to different species and growing conditions, there are significant chemical differences between the two species, which may result in the improper clinical usage under the same name. Chemical pattern expressed by high performance liquid chromatographic (HPLC) fingerprint analysis can play an important role in species differentiation and quality control of Radix Paeoniae Rubra. In the present work, HPLC fingerprints of two kinds of Radix Paeoniae Rubra have been established and analysed with chemometric methods including similarity evaluation and principal component analysis. Both of the fingerprint common patterns of the two species comprise 13 characteristic peaks, nine of which were common peaks of the two species. However, significant differences between the roots of P. veitchii and P. lactiflora exist not only in the content of certain constituents, especially phenolic acids but also in peak-to-peak ratios expressed by the fingerprint patterns. According to the recent pharmacological studies on polyphenolic constituents, root originating from P. veitchii may possess better efficacy and quality than that from P. lactiflora. Our research reveals that further pharmacological investigation is very necessary to determine whether the two species should be embodied under the same monograph in Chinese Pharmacopoeia. © 2008 Elsevier B.V. All rights reserved.
1. Introduction The entity and content of bioactive components of traditional Chinese medicine (TCM) are always highly variable, which depend on the species, geographical origins, cultivation and harvesting methods and post-harvesting formulation processes [1,2]. Chromatographic fingerprint analysis adequately caters for the complex characteristics of TCM, so it has been used for species differentiation, evaluation of quality and ensuring the consistency and stability of TCM [3–6]. Radix Paeoniae Rubra (“Chi-shao” in Chinese) is one of the most widely used Chinese herbal drugs. According to Chinese Pharmacopoeia 2005 edition, the herbal drug originates from the dried roots of two species, Paeonia lactiflora Pallas and Paeonia veitchii Lynch [7]. P. lactiflora naturally grows in Inner Mongolia
∗ Corresponding author. Tel.: +86 20 81887233x30908; fax: +86 20 81867705. ∗∗ Corresponding author. Tel.: +86 756 3326296; fax: +86 756 3326961. E-mail addresses: [email protected] (L. Yang), [email protected] (P. Xie). 0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.04.064
autonomous region, Heilongjiang, Jilin and Liaoning provinces, and P. veitchii is mainly distributed in Sichuan, Shaanxi and Qinghai provinces. Different species and growing conditions may cause significant difference of chemical patterns between P. lactiflora Pallas and P. veitchii Lynch, however, they have been used under the same name in clinical prescriptions for ages. Moreover, although paeoniflorin is a major bioactive component in Chi-shao, it is contrary to TCM principle to designate paeoniflorin as one and only marker compound for quality evaluation of both species in Chinese Pharmacopoeia, because the therapeutic efficacy of herbal drug is always attributed to its all kinds of bioactive components but not to any single ingredient according to TCM’s principle. These contradictions reveal that a comprehensive study of the chemical patterns of the two species would be valuable to help understand their activity and usage. Unfortunately, previous investigations on quality control of Chi-shao mostly rested on determination of individual or several analytical compounds [8–11], and only a few studies focused on the discrimination between Chi-shao and its adulterants using genetic fingerprinting or infrared spectroscopy [12–14]. In the present study, by combining the chemometric methods such as similarity evaluation and principal component analysis
2164
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
(PCA), we used HPLC-photodiode array detection to develop a specific, practical and valid chromatographic fingerprint analysis approach for quality assessment and species differentiation of P. lactiflora Pallas and P. veitchii Lynch. 2. Materials and methods 2.1. Apparatus and reagents HPLC analyses were performed on a Waters 600 series HPLC system together with column compartment and photodiode array detector (PDA). HPLC-grade acetonitrile was purchased from Fisher Chemicals (Fair lawn, New Jersey, USA). HPLC-grade formic acid was purchased from Tedia Company, Inc. (Fairfield, OH, USA). The water used in the experiment was doubly distilled in the laboratory. Other chemicals and solvents were of analytical grade. 2.2. Plant materials Nineteen batches of Chi-shao from eleven provinces of China were collected (Table 1). These herbal samples were authenticated by Professor Mian Zhang (China Pharmaceutical University, Nanjing, China). Voucher specimens are stored at Guangdong provincial hospital of TCM. 2.3. Reference compounds Albiflorin and 1,2,3,4,6-penta-O-galloyl--d-glucose were provided by Shanghai University of Traditional Chinese Medicine, Shanghai, China. Oxypaeniflorin was provided by Chromap Institute of Herbal Medicine Research (Zhuhai, China). Gallic acid, (+)-catechin, paeoniflorin and paeonol were purchased from the National Institute for the Control of Pharmaceutical and Biological Products, Tiantanxili No. 2, Beijing, China. Benzoylpaeniflorin and methyl gallate were isolated and purified by our own laboratory. These reference compounds were dissolved in methanol and then injected into HPLC after filtration with a 0.22 m filter (PTFE). 2.4. Sample preparation All samples were dried at 60 ◦ C for 12 h before use. Each dried sample was ground to a fine powder (40 mesh, 450 m I.D.) using a pulverizer. An aliquot of 2 g was accurately weighed and macerated
with 20 ml 50% aqueous methanol (v/v) at room temperature for 30 min. Sample was thrice extracted under reflux for 45 min each (50% methanol volume: 20, 15 and 15 ml). The sample solution was filtered through Whatman 1 filter paper. The filtrate was combined and decanted into a 50 ml volumetric flask, and then 50% aqueous methanol was added to make up to the scale. The final solution was passed through a 0.22 m membrane prior to use. An aliquot of 20 l of each sample solution was injected into the HPLC system for analysis. 2.5. Chromatographic condition The HPLC fingerprinting analysis was carried out on a Waters Symmetry C18 column (250 mm × 4.6 mm I.D., 5 m). A binary gradient elution system composed of acetonitrile as solvent A and 0.1% phosphoric acid in water as solvent B was applied for the fingerprint analysis with the gradient elution as follows: 0–5 min, 10–15% A; 5–25 min, 15–22% A; 25–45 min, 22–70% A; 45–46 min, 70–80% A; 46–50 min, 80% A. The mobile phase flow rate was 0.8 ml/min, and column temperature was maintained at 25 ◦ C. The PDA detector was set at 254 nm, and the on-line UV spectra were recorded in the range of 195–400 nm. 2.6. Data analysis of chromatogram Data analysis was performed by a pattern recognition program recommended by the Chinese Pharmacopoeia committee and the complementary software developed by Zhuhai Chromap Institute of Herbal Medicine Research that allows for a statistical evaluation of chromatographic patterns. In the present work, similarity evaluation and principal component analysis were used to assess HPLC fingerprints of all samples. 3. Results and discussion 3.1. Optimization of extraction conditions Considering a variety of components with relatively high polarity in Chi-shao, aqueous methanol was the preferred choice of as extraction solvent to effectively extract overall compounds in the herb. Related extraction conditions were examined and evaluated, which involved the following experimental factors and corresponding levels: extraction method (ultrasonication or reflux), extraction
Table 1 A summary of tested Chi-shao samples No.
Collection location
Collection time
Species
Voucher specimens nrs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Ganzi, Sichuan, China Kunming, Yunnan, China Nanning, Guangxi, China Shanghai, China Guangzhou, Guangdong, China Duolun, Neimeng, China Bozhou, Anhui, China Jiling, China Changchun, Jiling, China Hongkong, China Hongkong, China Taibei, Taiwan, China Chendu, Sichuan, China Zhongjiang, Sichuan, China Xichang, Sichuan, China Nanchang, Jiangxi, China Bozhou, Anhui, China Bozhou, Anhui, China Ganzi, Sichuan, China
November, 2004 April, 2004 March, 2003 March, 2003 April, 2005 April, 2005 March, 2005 April, 2003 March, 2003 December, 2004 May, 2005 April, 2004 December, 2004 December, 2004 November, 2004 April, 2005 October, 2005 November, 2005 October, 2005
P. veitchii P. veitchii P. veitchii P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. lactiflora P. veitchii P. veitchii P. veitchii P. lactiflora P. lactiflora P. lactiflora P. veitchii
PV20041111 N/Aa N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A PL20051013 PL20051107 PV20051022
a
Not available.
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
2165
Fig. 1. Average artificial HPLC fingerprint common pattern of P. lactiflora based on 12 samples.
repetitions (1, 2, 3 or 4 times), methanol concentration (0, 50, 100%, v/v), and extraction time (0.5, 0.75, 1 h). The number and areas of characteristic peaks of HPLC fingerprint were regard as a criterion for the selection of optimal conditions. As a result, the extraction method of Chi-shao is presented as shown in Section 2.4. 3.2. Optimization of HPLC conditions It is very difficult to separate chemical constituents of Radix Paeoniae Rubra because of their similar physicochemical properties, so chromatographic parameters were optimized to achieve a higher separation quality of the fingerprint and if possible in a reduced analysis time. Waters Symmetry C18, Agilent XDB-C18 , Agilent SB-C18 , Agilent XDB C8, Capcell pak C18 MG, Hypersil ODS and Kromasil ODS columns were all tested, and a Waters Symmetry C18 column finally was selected for the analysis. Methanol–water and acetonitrile–water as mobile phase were compared, and acetonitrile–water could achieve better resolutions. The retention behavior of phenolic acids and galloyl glucoses in Radix Paeoniae Rubra on the reversed-phase HPLC column was significantly affected by the pH of the mobile phase, so different acids including formic acid, acetic acid, phosphoric acid, and amount of acid added into mobile phase were also tested. As a result, the addition of 0.1% (v/v) phosphoric acid could achieve a good baseline and satisfactory resolution of these major peaks. The detection wavelength was set at 230 nm because most of the characteristic components in Chishao have satisfactory sensitivity at this UV wavelength. Column temperature, flow rate and the slope of the gradient are also individually examined at several levels. The best condition for a given factor is selected based on the peak capacity observed. All best factor levels are then combined. Finally, the optimal HPLC condition is shown in Section 2.5. 3.3. Validation of methodology Method precision and reproducibility were evaluated by the analysis of five injections of the same sample solution and five replicates of the same sample, respectively. The relative standard deviations (RSD) of retention time (tR ) and peak area (PA) of characteristic peaks in the precision test were found in the range of 0.4–3.1%, whereas in the reproducibility test the RSDs of tR and PA were also below 0.7 and 3.3%, respectively. Meanwhile, the analyses of the same sample solution at intraday different time point (0, 2, 4, 8, 24 and 48 h) were evaluated as the stability test of sample solution, and the RSDs of tR and PA were less than 0.4 and 3.2%, respectively. All results indicated that the method of HPLC fingerprint analysis was valid and satisfactory.
3.4. HPLC fingerprint analysis of Chi-shao As shown in Figs. 1 and 2, Table 2, there exist a total of 16 characteristic peaks in the two HPLC fingerprint common patterns of the two Paeonia species, of which nine characteristic peaks were assigned by comparing the UV spectra and their retention time with those of the reference compounds, i.e. peak 2, 3, 4, 5, 6, 7, 12, 15 and 16 were identified as gallic acid, oxypaeoniflorin, (+)-catechin, methyl gallate, albiflorin, paeoniflorin, 1,2,3,4,6-pentagalloyl glucose, benzoylpaeoniflorin and paeonol, respectively. Peak 10 and 11 were tentatively identified as tetragalloyl glucose and galloyl paeoniflorin based on their MS fragmentation behaviors obtained by LC–MS/MS [15]. By visually comparing the HPLC fingerprints of roots of P. lactiflora and P. veitchii, the peaks 6 (albiflorin), 9 (unknown) and 16 (paeonol) could not be detected in the HPLC fingerprints of P. lactiflora, whereas the peaks 3 (oxypaeoniflorin), 4 ((+)-catechin) and 14 (unknown) could not be observed in the HPLC fingerprints of P. veitchii. Moreover, the peak 7 (paeoniflorin) was the highest in the chromatograms of P. lactiflora, while 1,2,3,4,6-pentagalloyl glucose (peak 12) was the most in the roots of P. veitchii. Besides, other gallic acid derivatives in roots of P. veitchii such as gallic acid (peak 2), methyl gallate (peak 5), tetragalloyl glucose (peak 10) and galloylpaeoniflorin (peak 11) were also more abundant than those in roots of P. lactiflora. As shown in Fig. 3, two similarity indices of each chromatogram could be obtained with test solution using the HPLC fingerprint Table 2 The retention time (tR ) and relative peak area of 16 characteristic peaks of HPLC fingerprint common pattern of P. lactiflora and P. veitchii Peak no.
1 2a 3a 4a 5a 6a 7(R)a 8 9 10 11 12a 13 14 15a 16a a
tR (min)
4.35 5.14 12.09 13.25 14.12 17.47 19.82 21.94 23.25 24.15 26.99 28.34 33.52 34.90 40.22 44.70
Relative area
Compound
P. lactiflora
P. veitchii
0.02 0.03 0.04 0.11 0.05 N/A 1.00 0.04 N/A 0.04 0.03 0.17 0.15 0.02 0.04 N/A
0.06 0.14 N/A N/A 0.48 0.04 1.00 0.05 0.04 0.10 0.06 1.20 0.35 N/A 0.06 0.05
Compared with reference compounds.
Unknown Gallic acid Oxypaeoniflorin (+)-Catechin Methyl gallate Albiflorin Paeoniflorin Unknown Unknown 1,2,3,6-Tetragalloyl glucose Galloyl paeoniflorin 1,2,3,4,6-Pentagalloyl glucose Unknown Unknown Benzoylpaeoniflorin Paeonol
2166
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
Fig. 2. Average artificial HPLC fingerprint common pattern of P. veitchii based on seven samples.
common patterns of P. lactiflora and P. veitchii. By comparison with the fingerprint common pattern of P. lactiflora, the similarity index of every fingerprint of P. lactiflora was not less than 0.97 except those of samples 4 and 5 (see Fig. 3a); and the similarity indices of the P. veitchii’s fingerprints were more than 0.90 aside from sample 15’s by comparison with the fingerprint common pattern of P. veitchii (see Fig. 3b). The result of similarity evaluation showed the developed HPLC fingerprint common patterns, as quality assessment models for P. lactiflora and P. veitchii are practical and valid. Furthermore, based on the similarity evaluation, any Chi-shao sample can be easy classified. 3.5. Quality assessment of Chi-shao by principal component analysis (PCA) For quality evaluation of Chi-shao, PCA, a method for feature extraction and dimensionality reduction, was carried out. The PCA computation was implemented by performing singular value decomposition on the data array of the fingerprints, which consisted of a total of 19 × 16 data matrix, each row represented a plant sample and each column contained the values of 16 characteristic peak areas. In this study, in order to amplify the difference between P. lactiflora and P. veitchii and easily discriminate the outliers from normal samples, matrices based on not only the raw data but also standardized (auto-scaling [16]) data were all submitted to PCA analysis. The first two loadings of PCA results were used to generate the projection plot that provides a visual determination of the similarity among the sample fingerprints.
In the PCA result based on raw data, the loading of 1,2,3,4,6pentagalloyl glucose was 0.95, the highest in the first principal component (PC1 ), while the loading of paeoniflorin was 0.87 in the second principal component (PC2 ). In the PCA result based on standardized data, the loadings of 1,2,3,4,6-pentagalloyl glucose and paeoniflorin were 0.72 and −0.52, respectively, in PC1 , and the loadings of gallic acid and galloylpaeoniflorin were −0.59 and 0.61, respectively, in PC2 . As shown Fig. 4a and b, both the two kinds of PCA showed the similar result that 15 batches of Chi-shao samples were classified into two groups: P. lactiflora and P. veitchii except for samples 4, 5, 14 and 15. Moreover, the result of PCA was at largely consistent with that of similarity evaluation. Both of samples 4 and 5 from P. lactiflora displayed a significant difference from other P. lactiflora samples as the results of above two chemometric analyses, so it was reasonable to judge the two samples as outliers. Fig. 5a, for example, clearly showed that these differences between HPLC fingerprint of sample 4 and the common pattern of P. lactiflora consisted in the integrated area values of characteristic peaks as well as peak-to-peak ratios. Samples 14 and 15 were regarded as outliers of P. veitchii according to the result of PCA based on raw data, however, sample 14 was determined as a normal sample in similarity evaluation in view of its high similarity index (over 0.9), and the result of PCA based on standardized data also indicated that the difference of sample 14 was not significant (see Fig. 4b). Thus, only sample 15 was determined as an outlier of P. veitchii. As shown in Fig. 5b, although all of the characteristic peaks of P. veitchii could be observed in the HPLC fingerprint of sample 15, due to very low content of characteristic
Fig. 3. Similarity evaluation of HPLC fingerprint for Chishao. (a) Refer to the fingerprint common pattern of P. lactiflora and (b) refer to the fingerprint common pattern of P. veitchii.
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
2167
Fig. 4. PCA analysis for the fingerprints of Chishao. (a) Projection plot of 1–2 principal component based on the raw data and (b) projection plot of 1–2 principal component based on the standardized data.
Fig. 5. Comparison between chromatograms of outlier samples 4, 15 and common patterns for P. lactiflora and P. veitchii, respectively. (a) Sample 4 (upper) and fingerprint common pattern of P. lactiflora (lower) and (b) sample 15 (upper) and fingerprint common pattern of P. veitchii (lower).
components, especially paeoniflorin, this sample was picked out as an outlier of P. veitchii. 4. Conclusions In this paper, a HPLC fingerprint analysis with chemometric methods was used for the purpose of species differentiation, quality evaluation and the consistency check of Chi-shao collected from different sources. The result showed that the developed HPLC fingerprint analytical method was practical and reliable for above purpose. Obviously, the outliers picked out by means of developed fingerprint analysis method should be rejected to ensure curative effect of the herb and its finished product. Furthermore, 11 of 16 characteristic peaks in the common pattern were identified by comparing with the reference compounds based on their UV spectra and retention time or the data obtained by LC–MS/MS to further characterize the chromatographic fingerprint and contribute to the quality control of Chi-shao. There are significant differences between the roots of P. veitchii and P. lactiflora not only in the content of major common constituents, especially phenolic acids, but also in peak-to-peak ratios expressed by the fingerprint patterns, which raises doubt about the ‘bio-equivalence’ of the two species with the same dose. It is worth noting that root of P. veitchii possesses much more gallic acid derivatives than that of P. lactiflora such as galloyl glucoses, gallic acid, methyl gallate and galloylpaeoniflorin. In addition, content of paeoniflorin is close in the two species as emphasized in
Chinese Pharmacopoeia. Recently, galloyl glucoses isolated from Chishao were found to significantly decrease the plasma endotoxin level [17], which was consistent with the therapeutic efficacy of this herb. Methyl gallate, gallic acid and paonol were also reported as major potent constituents of Chi-shao [18]. Considering these pharmacological investigations, it seems to suggest that roots originating from P. veitchii possess better efficacy and quality than those from P. lactiflora. Hence, further investigation is necessary to determine whether the two herbs should be recorded under the same monograph in Chinese Pharmacopoeia. Acknowledgements The research is funded by Science and Technology Department of Guangdong Province and State Administration of Traditional Chinese Medicine of China. References [1] P. Zou, Y. Hong, H.L. Koh, J. Pharm. Biomed. Anal. 38 (2005) 514. [2] L.W. Yang, D.H. Wu, X. Tang, W. Peng, X.R. Wang, Y. Ma, W.W. Su, J. Chromatogr. A 1070 (2005) 35. [3] P.S. Xie, S.B. Chen, Y.Z. Liang, X.H. Wang, R.T. Tian, R. Upton, J. Chromatogr. A 1112 (2006) 171. [4] W. Jin, R.L. Ge, Q.J. Wei, T.Y. Bao, H.M. Shi, P.F. Tu, J. Chromatogr. A 1132 (2006) 320. [5] S.B. Chen, H.P. Liu, R.T. Tian, D.J. Yang, S.L. Chen, H.X. Xu, A.S.C. Chan, P.S. Xie, J. Chromatogr. A 1121 (2006) 114. [6] G.H. Lu, K. Chan, Y.Z. Liang, K. Leung, C.L. Chan, Z.H. Jiang, Z.Z. Zhao, J. Chromatogr. A 1073 (2005) 383.
2168
S. Xu et al. / J. Chromatogr. A 1216 (2009) 2163–2168
[7] The State Pharmacopoeia Commission of People’s Republic of China, Pharmacopoeia of People’s Republic of China, vol. 1, Chemical Industry Press, Beijing, 2005. p. 109. [8] Y. He, H.D. Zhao, L.Y. Tang, Z.J. Wang, Q.W. Zhang, Chin J. Chinese Materia Med. 31 (2006) 1070. [9] S.L. Hu, G.L. Fu, X.F. Feng, X.J. Tang, X.R. He, Chin J. Chinese Materia Med. 25 (2000) 10. [10] C.H. Chen, J.Q. Han, Y.Q. Yao, Acta Academiae Medicinae Zhejiang 2 (2006) 30. [11] X.M. Xie, C.Z. Yu, H. Xu, S. Wang, D.L. Wang, L. Zhang, G.X. Chou, Z.T. Wang, Chin J. Chin. Materia Med. 29 (2004) 46.
[12] H.T. Zhou, S.L. Hu, B.L. Guo, X.F. Feng, Y.N. Yan, J.S. Li, Acta Pharm. Sin. 37 (2002) 383. [13] B. Dong, S.Q. Sun, H.T. Zhong, S.L. Hu, Spectrosc. Spectral Anal. 22 (2002) 232. [14] C.G. Cheng, D.T. Li, J.H. Chen, L.C. Kong, Z.F. Cheng, Chin. Pharm. J. 41 (2006) 580. [15] S.J. Xu, L. Yang, X. Zeng, M. Zhang, Z.T. Wang, Rapid Commun. Mass Spectrom. 20 (2006) 3275. [16] D.L. Massart, L. Kaufman, The Interpretation of Analytical Data by the Use of Cluster Analysis, Wiley, New York, 1983, 20. [17] G. Lv, J. Zheng, H. Zhou, Y.M. Zheng, L.X. Wang, G. Wei, M. Huang, D.L. Jiang, L.Z. Wei, Int. Immunopharmacol. 5 (2005) 1007. [18] H.L. Huang, G.M. Wang, B. Xu, Food Sci. 28 (2007) 76.