2490 Xiaoxue Wu1,2 Jiao He1,2 Huarong Xu1,2 Kaishun Bi1,2 Qing Li1,2 1 School
of Pharmacy, Shenyang Pharmaceutical University, Shenyang, China 2 National and Local United Engineering Laboratory for Key Technology of Chinese Material Medica Quality Control, Shenyang Pharmaceutical University, Shenyang, China
Received May 4, 2014 Revised June 12, 2014 Accepted June 15, 2014
J. Sep. Sci. 2014, 37, 2490–2498
Research Article
Quality assessment of Cinnamomi Ramulus by the simultaneous analysis of multiple active components using high-performance thin-layer chromatography and high-performance liquid chromatography A novel and improved method for the quality assessment of Cinnamomi Ramulus was developed and completely validated. The method was established using fingerprint technology and simultaneous quantitative determination of six main marker compounds including coumarin, cinnamic alcohol, cinnamic acid, 2-methoxy cinnamic acid, cinnamaldehyde, and 2-methoxy cinnamaldehyde in the herbal medicine for the first time. A newly developed high-performance thin-layer chromatography method, which achieved simultaneous definition of five marker components by comparing the colors and retardation factor values of the bands in high-performance thin-layer chromatography, was first used for the authentication of Cinnamomi Ramulus. The fingerprints of 26 batches of herbal samples from different regions of China showed very similar chromatographic patterns that were evaluated by similarity analysis and hierarchical clustering analysis. In addition, six marker compounds were simultaneously determined using single standard to determine multiple components by the relative response factors. Compared with the external standard method, the new quantitative method was validated to determine multiple compounds in 26 batches of Cinnamomi Ramulus samples. All results demonstrated that the simple and rapid method could be effectively utilized for the quality control of Cinnamomi Ramulus. Keywords: Cinnamomi Ramulus / Fingerprint / Quality assessment / Relative response factors DOI 10.1002/jssc.201400494
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Cinnamomi Ramulus (CR) is the dried young stem of Cinnamomum cassia Presl (Lauraceae), widely distributed in Vietnam, Myanmar, Laos, and the southern part of mainland China (Guangxi, Guangdong, and Yunnan Province) [1]. CR contains abundant volatile oil with a special pungent taste and scent on account of the volatile, cinnamaldehyde. It also contains phenols, organic acids, polysaccharide, glycosides, coumarin, and tannins. As one of the most important TraCorrespondence: Dr. Qing Li, School of Pharmacy, Shenyang Pharmaceutical University, Shenyang 110016, China E-mail: [email protected] Fax: +86-24-23984392
Abbreviations: CR, Cinnamomi Ramulus; ESM, external standard method; HCA, hierarchical clustering analysis; HPTLC, high-performance thin-layer chromatography; RPA, relative peak area; RRF, relative response factor; RRT, relative retention time; SSDMC, single standard to determine multicomponent C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ditional Chinese Medicines, CR is widely used for treating dyspepsia, gastritis, blood circulation issues, and inflammatory disease [2–5]. In Chinese Pharmacopoeia (2010 edition), the authentication and quantitation of one marker component (cinnamaldehyde) are set as the parameter for the QC of CR [6]. However, other compounds such as coumarin, cinnamic alcohol, and cinnamic acid also exhibit specific biological effects [7, 8] and are relatively abundant in CR herbs, indicating multiple components analysis is necessary for CR QC. At present, only a few papers [9–11] have focused on the quantitative analysis of one or several constituents that were far from enough to control the overall quality of CR. Chromatographic fingerprint analysis with multipeak information content of a chromatogram has been introduced as a rational strategy for assessing herbal samples. Compared with the traditional approaches that select one or more compounds as active markers for identification and quality assessment, fingerprint approaches rely on the inherent
Colour Online: See the article online to view Fig. 1–3 in colour. www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 2490–2498
relationships between multiple compounds to display the chemical pattern of herbal sources [12–18]. Thus, chromatographic fingerprint is highly recommended for application to QC of herbal medicines. Furthermore, in order to comprehensively control the quality of Traditional Chinese Medicines, the quantitative analysis of multicomponents is requisite. But the traditional external standard method (ESM) needs relevant multiple reference standards to calculate the concentrations of components quantified. These reference standards require sufficient chemical purity. In addition, they are mostly expensive and insufficient, especially when the component is of low content and hard to be purified from the plant [19, 20], such as cinnamic alcohol and 2-methoxy cinnamic acid in CR. Moreover, some constituents of herbal medicines become unstable when they are purified from a complicated matrix, like cinnamaldehyde in this study. Cinnamaldehyde is light sensitive and easy to be oxidized because of the structure of olefine aldehyde. To overcome this problem, a simple quantitative method has been successfully applied for the QC of botanicals with a single standard to determine multicomponents (SSDMCs) [20–22]. The SSDMC method uses a single internal standard to simultaneously quantify other components by relative response factors. The method has also been successfully applied in QC of botanical extracts and dietary supplements according to USP, European Pharmacopoeia, and Chinese Pharmacopoeia [21, 23–25]. In the present study, an improved quality assessment of CR based on multiple components was established using high-performance thin-layer chromatography (HPTLC) and HPLC methods. Compared with TLC, HPTLC provides better separation of samples and higher performances, which as a versatile, high-throughput, and cost-effective technology is uniquely suited to assessing the identity and quality of botanical materials. The chromatograms obtained by HPTLC and HPLC methods could show adequate qualitative information for the authentication and identification of CR. In addition, six active components in CR were successfully quantitatively determined by SSDMC method for the first time. Cinnamic acid was chosen as an internal reference substance to determine five other components by their UV relative response factors at specific wavelength. In order to ensure the response factors reliability, some environmental and operational factors were investigated in a one-variableat-a-time procedure. The SSDMC method was successfully applied for the quantitative analysis of 26 batches of CR samples. In other words, the quantitative analysis combined with qualitative evaluation offered comprehensive and reliable information for the QC of CR.
2 Materials and methods 2.1 Chemicals and materials Acetonitrile (Fisher Scientific, USA, A452–4), methanol (Shandong Yuwang Chem, China), and phosphoric acid C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2491
(Damao Chem Tech, China) were of HPLC grade; cyclohexane, ethyl acetate, and acetic acid (Damao Chem Tech) were of analytical grade; water was redistilled. Coumarin, cinnamic acid, and protocatechuic acid were bought from Chengdu MUST Bio-technology; cinnamic alcohol, cinnamaldehyde, and 2-methoxy cinnamic acid were bought from Aladdin; 2-methoxy cinnamaldehyde was bought from Shanghai Siyu Chemical Technology. The identity and purity of all seven reference standards were confirmed by HPLC. Twenty-six batches of CR samples meeting the requirements of the Pharmacopoeia of the People’s Republic of China (2010 edition) were collected from different provinces in China. All samples were identified by Professor Ying Jia and deposited in the Centre of Chinese Material Medica, Shenyang Pharmaceutical University. They were labeled as Nos. 1–18 from Guangxi province, Nos. 19–24 from Guangdong province, and Nos. 25, 26 from Yunnan province. Nos. 1, 8, 19, and 21–24 were collected from Guangdong province. Nos. 3, 10–13, and 15 were collected from Guangxi province. Nos. 4 and 20 were collected from Anhui province. Nos. 6 and 26 were collected from Shandong province. Nos. 14, 17, and 18 were collected from Liaoning province. Nos. 2, 5, 7, 9, 16, and 25 were collected from Hebei, Fujian, Yunnan, Heilongjiang, Jiangxi, and Hubei province, respectively.
2.2 Apparatus Three kinds of TLC plates were investigated, which were TLC plate silica gel GF254, HPTLC plate silica gel GF254 (20 cm × 10 cm, Qingdao Haiyang Chemical Industrial), and HPTLC plate silica gel 60 F254 (20 cm × 10 cm, Merck, 34271 Darmstadt, Germany). The Shimadzu 20A Prominance UFLC XR LC System (Shimadzu Corporation, Japan) comprised a binary solvent delivery system, an online degasser, an autosampler, a column temperature controller, and photodiode array detector coupled with lab solution software. Additional comparative analysis was done on an Agilent 1260 HPLC System (Agilent Technologies, USA) comprising a quaternary solvent delivery system, an online degasser, an autosampler, a column temperature controller, and photodiode array detector coupled with an analytical workstation (Chemstation For LC 3D Systems A10.02).
2.3 HPTLC method for identification One gram of powdered herbal sample was extracted with 10 mL 70% methanol for 20 min by sonication. After centrifugation, the supernatant was prepared for HPTLC analysis. Test solutions (10 L) were applied to the HPTLC silica gel 60 F254 plate. The plate was then developed in a suitable double-trough chamber with a glass lid. The chamber was pre-equilibrated with the mobile phase for 20 min. After developing the plate in a chamber containing developing www.jss-journal.com
2492
X. Wu et al.
J. Sep. Sci. 2014, 37, 2490–2498
Figure 1. HPTLC result of Cinnamomi Ramulus visualized under 254 and 365 nm. (1) Cinnamic acid (Rf = 0.15); (2) cinnamic alcohol (Rf = 0.28); (3) coumarin (Rf = 0.39); (4–7) sample solutions of three batches; (8) 2-methoxy cinnamaldehyde (Rf = 0.51); (9) cinnamaldehyde (Rf = 0.56); 18⬚C; 30% (HPTLC, Yantai).
solvent system 1 (cyclohexane/ethyl acetate, 7:2, v/v) over a path of 8 cm, the plate was air-dried, developed the same plate in a chamber containing developing solvent system 2 (cyclohexane/ethyl acetate/acetic acid, 7:2:0.1, v/v/v) over a path of 3 cm. The plate was dried and examined under UV light at wavelengths of 254 and 365 nm (Fig. 1).
2.4 HPLC method for fingerprint and quantitative determination
2.4.2 Preparation of standard solution The standards were accurately weighed and dissolved in methanol to obtain a mixed standard stock solution (No. 1) containing coumarin (0.02900 mg/mL), cinnamic alcohol (0.00925 mg/mL), cinnamic acid (0.05088 mg/mL), 2methoxy cinnamic acid (0.004000 mg/mL), cinnamaldehyde (0.5057 mg/mL), and 2-methoxy cinnamaldehyde (0.07056 mg/mL). Samples (4.0, 3.0, 2.5, 2.0, 1.0, 0.5 mL) of standard stock solution were quantitatively transferred to six 5 mL amber volumetric flasks (Nos. 2–7), diluted with methanol to volume, and mixed. All standard solutions were kept at −20⬚C.
2.4.1 Preparation of sample solution 2.4.3 Chromatographic conditions The powdered herbal sample (0.50 g) was extracted with 25 mL 70% methanol for 30 min by sonication, allowed to cool, and adjusted to the initial weight by adding a mixture of 70% methanol as needed. After centrifugation, the supernatant was filtered through a 0.22 m membrane into an HPLC amber sample vial for HPLC analysis. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The separation was performed on a Luna C18 column (4.6 × 250 mm, 5 m, Phenomenex, CA, USA) protected by a Security Guard C18 guard column (4.0 × 3.0 mm, 5 m, Phenomenex) with a sample injection volume of 10 L. And the separation was conducted at 25⬚C with a flow rate of www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 2490–2498
1.0 mL/min. The mobile phase consisted of solvent A, acetonitrile and solvent B, 0.05% aqueous phosphoric acid. The detection wavelength was set at 265 nm. The linear gradient profile for qualitative assessment was 10–20% B for 10 min, 20–50% B for 25 min, 50% B for 5 min. For quantitative analysis, the linear gradient profile was further optimized to be 25% B for 1 min, 25–38% B for 20 min, 38–40% B for 9 min, 40–48% B for 5 min. 2.4.4 Evaluation methods All samples were analyzed by Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine (Version 2009A). A similarity index of each sample was calculated based on correlation coefficient value against simulative reference standard fingerprint. For further study, hierarchical clustering analysis (HCA) was employed by software SPSS 19.0. Cinnamic acid was chosen the internal standard reference component as it was stable, available, and had appropriate peak area and retention time in chromatogram. 2.4.5 Data analysis RRF is the relative response factor of component to be measured, which is a mean value calculated as reporter previously based on the linearity data [21]. RRF =
Ax × Ci Ai × Cx
(1)
Ax and Cx are the peak area and concentration of the internal reference standard, cinnamic acid in this study. Ai and Ci are the peak area and concentration of other reference standards, coumarin, cinnamic alcohol, 2-methoxy cinnamic acid, cinnamaldehyde, and 2-methoxy cinnamaldehyde. The relative retention time (RRT) is the ratio of retention time of the reference standard (Tx ) and the internal reference standard (Ti ). RRT =
Tx Ti
(2)
2.5 Validation of the quality assessment The HPTLC method was validated on its specificity, stability, and robustness. The methodology of HPLC fingerprint was validated on its repeatability, precision, and stability. To ensure the validity of this new method (SSDMC), validation tests for linearity, LOD, LOQ, accuracy, precision (intraday and interday variability), and stability were performed. Validation was assessed by calculating the RSDs. The results obtained by SSDMC were compared to the results obtained by ESM. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2493
3 Results and discussion 3.1 Optimization of the chromatographic conditions for HPTLC analysis In the study, different extraction solvents including absolute ethyl alcohol, ethyl ether, methanol, 70% methanol, and 50% methanol were investigated. Five dark components of sample solution extracted by 70% methanol showed clearest bands in HPTLC. Different volumes of application (5–15 L) and bandwidths (2–10 mm) were tested. Results showed that increasing the volume or decreasing bandwidths of sample application, the separation between two adjacent bands as cinnamaldehyde and 2-methoxy cinnamaldehyde on the top of the chromatogram would be reduced. Finally, an appropriate chromatogram was obtained by the application of 10 L of sample solution at 10 mm. The developing method including the developing solvent system and the developing procedure was investigated. Different types of developing solvent system such as petroleum ether (60–90⬚C)/ethyl acetate, cyclohexane/ethyl acetate, cyclohexane/acetone, cyclohexane/methanol, cyclohexane/ethyl acetate/methanol with different ratios were investigated. Results showed that developed plates obtained by developing solvent system of cyclohexane/ethyl acetate (7:2, v/v) over a path of 8 cm could move and separate most compounds well except a polar compound cinnamic acid with obvious trail and band close to the baseline. Some molecules of cinnamic acid with the carboxyl in CR might be partially dissociated in the developing procedure. Adding acetic acid in the developing solvent system such as cyclohexane/ethyl acetate/acetic acid (7:2:0.1, v/v/v) could help to move cinnamic acid off the baseline without trail, whereas two nonpolar compounds cinnamaldehyde and 2-methoxy cinnamaldehyde with similar polarity on the top of the chromatogram could not be well separated. The Rf values of most nonpolar compounds were reduced simultaneously. Finally, a second developing method was taken in to account and all components in CR showed a good separation (Section 2.3).
3.2 Optimization of the chromatographic conditions for HPLC analysis The detection wavelength was crucial for developing a reliable fingerprint and for accurate quantitative analysis of marker compounds in the herb. The solution prepared was scanned in the entire UV range (200–400 nm). The maximum absorption wavelength of cinnamic alcohol is 202 nm, which was very different from the others (coumarin 277 nm, cinnamic acid 277 nm, 2-methoxy cinnamic acid 276 nm, cinnamaldehyde 290 nm, 2-methoxy cinnamaldehyde 287 nm). But all components showed an appropriate absorption at 265 nm simultaneously with smooth baseline. At this wavelength, more characteristic peaks in the fingerprint chromatogram were observed, and the marker compounds were eligible for being detected in the HPLC quantitation. www.jss-journal.com
2494
X. Wu et al.
For good resolution within a short analysis time, the HPLC separation conditions, such as mobile phase system, pH of mobile phase, program of mobile phase, flow rate, column temperature, and injection volume were further optimized. Obviously, acetonitrile could accelerate the elution with a suitable separation of all active components. Adding phosphoric acid helped to improve peak symmetry, especially for cinnamaldehyde. There was no sharp effect upon changing the column temperature (20–30⬚C), flow rate (0.8– 1.2 mL/min), and injection volume (5–15 L) on either the resolution of the eluted peaks and the peak symmetry. In consideration of short analysis time and most peaks at acceptable resolution, the column temperature, the flow rate, and injection volume were set at 25⬚C, 1.0 mL/min, and 10 L, respectively. A number of mobile phases with different gradients were screened in order to obtain a reliable chromatogram with most peaks at acceptable resolution and balance for the HPLC fingerprint and to obtain baseline separation of each marker compound (resolution factor > 1.5) in a relatively short analytical time for the HPLC quantitation. Finally, two gradient programs (described in Section 2.4.3) were established for the HPLC fingerprint and quantitative analyses of the herb. 3.3 Method validation 3.3.1 Method validation of HPTLC analysis To ensure good reliability and reproducibility, the HPTLC method was validated on its specificity, stability, and robustness. For specificity, five compounds were identified by matching the colors and Rf values of bands with marker standards under 254 and 365 nm (Fig. 1). For stability, four sample solutions placed at room temperature (0, 1, 2, and 3 days) were tested. For robustness, different particle size of silica gel (TLC plates and HPTLC plates), manufacturers of the precoated plates (Yantai, China; Merck, USA), different temperature (10, 20, and 30⬚C), and different humidity (30, 60, and 90%) were investigated. Results showed that the sample solution was stable for three days and different size of silica gel, different manufacturers, temperature, and humidity among test range had no significant influence on the chromatograms of standard and sample solutions. But the best chromatography could be conducted by HPTLC plates (Merck) at low temperature and humidity about 10⬚C, 40%. 3.3.2 Method validation of HPLC fingerprint The standardized HPLC fingerprint of CR was generated through the average method from the general comparison of all 26 CR samples shown in Fig. 2A. Then, 14 common chromatograph peaks were obtained by the aforementioned software for similarity analysis for identification and peak 11 (cinnamic acid) as the interreference peak for calculation of the RRT and relative peak area (RPA) of other common characteristic peaks. There were seven peaks identified by C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Sep. Sci. 2014, 37, 2490–2498
comparing the retention times, UV spectra, and by spiking with chemical reference substances as follows: peaks 3, 9, 10, 11, 12, 13, 14 were protocatechuic acid, coumarin, cinnamic alcohol, cinnamic acid, 2-methoxy cinnamic acid, cinnamaldehyde, and 2-methoxy cinnamaldehyde, respectively. The repeatability was determined by determining five parallel solutions of No. 13, and their RSDs were 0.9–2.1% for RPA and 0.0–0.1% for RRT, respectively. The precision was examined by the analysis of five injections of No. 13 consecutively, and their RSDs were 0.2–3.9% for RPA and 0.0–0.1% for RRT. The stability was evaluated by analysis of sample solution of different time points, i.e. 0, 2, 4, 8, 10, 12, and 24 h. The stability RSDs were 0.6–4.9% (n = 1) for RPA and 0.0– 0.2% (n = 1) for RRT. These results indicated that the above fingerprint method was feasible and the sample solution was stable within 24 h. 3.3.3 Method validation of quantitative analysis by SSDMC method Six marker compounds that included almost all compounds of the relatively high content and main active components in CR were simultaneously determined using SSDMC method (Fig. 2B). The linearity was evaluated by the standard solutions (Nos. 1–7). The calibration curves calculated by plotting the peak area Y of each compound against the concentration x (g/mL) of each compound were Y = 3.675 × 104 x + 4.735 × 103 , Y = 3.606 × 104 x + 2.395 × 103 , Y = 7.717 × 104 x + 1.171 × 103 , Y = 4.242 × 104 x + 3.526 × 103 , Y = 5.193 × 104 x + 215.4 × 103 , and Y = 2.867 × 104 x + 9.95 × 103 for coumarin, cinnamic alcohol, cinnamic acid, 2-methoxy cinnamic acid, cinnamaldehyde, and 2-methoxy cinnamaldehyde, respectively. Within test ranges, all calibration curves showed well linear regressions (R2 ࣙ 0.9998). The LOQ measured based on S/N of 10 was 0.017–0.06 ng/mL for the six compounds, which might be due to the high column efficiency and sensitive detector selectivity under the testing conditions. Accuracy was calculated as the formula: Recovery (%) = 100 × (amount found − original amount) / amount spiked. Prepared sample solutions in three different amount levels (75, 100, 125%) and triplicate experiments at each level. The results were in the range of 93.6–106.1% with RSD < 2.3%. For precision, the intraday variability of this method was assessed using nine test solutions of the same sample (No. 13) covering three different concentration levels (50, 100, 150%) and triplicate experiments at each level. Interday variability was still assessed using nine test solutions prepared by different analysts in one day and one analyst in successive three days. All RSDs of the six components calculated by ESM and SSDMC method were arranged in 0.4–1.8%. The stability of sample solution was carried out by comparing the peak areas of six components in the chromatograph of the same sample solution, after storing at room temperature for different times (0, 2, 4, 6, 8, 12, and 24 h). The RSDs of stability were ࣘ1.8%. www.jss-journal.com
J. Sep. Sci. 2014, 37, 2490–2498
Liquid Chromatography
2495
Figure 2. The standard reference chromatograms of HPLC fingerprint (A) and quantitative analysis (B) for Cinnamomi Ramulus. The mobile phase (acetonitrile/0.05 aqueous phosphoric acid) programs: (A) 0/10/35/40 min, 10/20/50/50 acetonitrile%; (B) 0/1/21/30/35 min, 25/25/38/40/48 acetonitrile%. Seven components identified (3) protocatechuic acid; (9) coumarin; (10) cinnamic alcohol; (11) cinnamic acid; (12) 2-methoxy cinnmic acid; (13) cinnamaldehyde; (14) 2-methoxy cinnamaldehyde.
Compared all contents of six marker components obtained by SSDMC method to those by ESM with t-test, and the result (P = 0.152, P > 0.05) illustrated that there was no significant difference between two methods and the SSDMC method was appropriate for simultaneous quantitative determination of six components in CR.
3.3.4 The robustness and ruggedness of the RRF and RRT of SSDMC In order to apply this SSDMC method to quantitatively analyze CR herb samples in different laboratories, some important factors were slightly varied to investigate the robustness, including flow rate (±0.1 mL/min), injection volume (±5 L), wavelength (±1 nm), column temperature (±5⬚C), acid concentration (±0.03% H3 PO4 ), ratio of organic phase (±1% B), time of gradient (±1 min). The affection was assessed by the RSD of RRT, RRF, and the total content of six components in CR calculated by ESM and SSDMC method. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The results showed that (i) there were no great influence on RRT, RRF, and the total content when slightly changed flow rate, injection volume, column temperature, the ratio of organic phase, and time of gradient in the test; (ii) but a slight variation of the wavelength affected significantly the RRF of cinnamic alcohol (RSD = 18.0%), cinnamaldehyde (RSD = 4.0%), and 2-methoxy cinnamaldehyde (RSD = 5.3%). The RRF of SSDMC method was very sensitive to wavelength variation when the UV spectrum of analytes was not detected at its maximum UV absorption as reported [21]. Thus, the wavelength of UV detector should be controlled strictly in SSDMC method. Through five concentration standard solutions, the ruggedness of RRT and RRF were compared with different equipment shown in Table 1. The results showed that the RRT was little different in two equipment and three columns. But the RRF varied a little in different equipment, which was fluctuated in a relative narrow range. Especially for the RRF of 2-methoxy cinnamaldehyde, its total RSD% of different equipment and columns was about 5.0%, relatively more than others. www.jss-journal.com
2496
J. Sep. Sci. 2014, 37, 2490–2498
X. Wu et al.
Table 1. Ruggedness of the RRT and RRF in Cinnamomi Ramulus, n = 5
Equipment
Column
SHIMADZU 20A
RSD (%) Agilent 1260
1 2 3 1 2 3
RSD (%) Total RSD (%)
Cinnamic alcohol
Coumarin
2-Methoxy cinnamic acid
Cinnamaldehyde
2-Methoxy cinnamaldehyde
RRT
RRF
RRT
RRF
RRT
RRF
RRT
RRF
RRT
RRF
0.82 0.83 0.82 0.7 0.81 0.83 0.81 1.5 1.1
2.11 2.10 2.11 0.3 2.12 2.09 2.11 0.8 0.5
0.90 0.90 0.90 0.0 0.90 0.90 0.90 0.0 0.0
2.10 2.06 2.07 1.0 1.99 1.96 1.97 0.8 3.0
1.14 1.16 1.16 1.0 1.14 1.16 1.16 1.0 0.9
1.82 1.82 1.82 0.0 1.82 1.83 1.83 0.4 0.3
1.26 1.31 1.31 2.3 1.27 1.32 1.32 2.3 2.1
1.48 1.48 1.49 0.4 1.57 1.57 1.58 0.4 3.3
1.52 1.61 1.59 3.0 1.52 1.63 1.60 3.6 3.0
2.68 2.70 2.69 0.4 2.91 3.00 2.91 1.8 5.0
Column: 1, Phenomenex Luna C18(2) 100A (4.6 × 250 mm, 5 m); 2, Agilent SB C18 (4.6 × 250 mm, 5 m); 3, Agilent XDB C18 (4.6 × 250 mm, 5 m). Table 2. The content and the similarity of 26 batches of Cinnamomi Ramulus
No.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Content by SSDMC (%)
Total content (%)
1
2
3 (ESM)
4
5
6
ESM
SSDMC
0.1006 0.0865 0.0531 0.1133 0.0488 0.0408 0.0566 0.0403 0.0874 0.0402 0.0697 0.0611 0.0581 0.0744 0.1305 0.0412 0.0725 0.0781 0.0410 0.0787 0.0983 0.0425 0.0556 0.0524 0.0415 0.0478
0.0137 0.0057 0.0059 0.0032 0.0091 0.0081 0.0235 0.0089 0.0136 0.0040 0.0332 0.0096 0.0204 0.0244 0.0055 0.0048 0.0192 0.0132 0.0179 0.0130 0.0077 0.0310 0.0080 0.0043 0.0068 0.0098
0.1257 0.0864 0.0739 0.1432 0.1009 0.0695 0.0673 0.0616 0.1243 0.0555 0.1150 0.0840 0.0633 0.1346 0.1347 0.0434 0.0892 0.1135 0.0405 0.1102 0.1277 0.0523 0.0441 0.0900 0.1369 0.0877
0.0078 0.0097 0.0051 0.0113 0.0058 0.0100 0.0063 0.0082 0.018 0.0062 0.0079 0.0055 0.0073 0.0124 0.0075 0.0049 0.0107 0.0102 0.0026 0.0119 0.0191 0.0032 0.0049 0.0097 0.0101 0.009
1.880 1.680 1.089 1.403 1.142 0.7965 1.014 0.8170 1.253 0.9382 1.642 1.462 0.9153 1.704 2.738 1.069 0.8615 1.129 1.185 0.9509 2.555 1.175 0.9876 1.024 0.8151 0.7047
0.2370 0.2389 0.1464 0.2927 0.1363 0.1383 0.1821 0.1398 0.2343 0.1380 0.2005 0.1692 0.1698 0.2105 0.3126 0.1002 0.1727 0.2212 0.1144 0.1720 0.6858 0.1216 0.1593 0.1438 0.1295 0.1540
2.380 2.121 1.382 1.979 1.452 1.070 1.357 1.082 1.741 1.190 2.081 1.803 1.241 2.174 3.351 1.272 1.233 1.574 1.411 1.345 3.514 1.435 1.267 1.333 1.146 1.019
2.364 2.107 1.373 1.967 1.443 1.063 1.349 1.076 1.730 1.182 2.068 1.792 1.234 2.160 3.329 1.264 1.226 1.565 1.402 1.337 3.494 1.426 1.259 1.324 1.140 1.013
RSD (%)
Similarity
0.5 0.5 0.5 0.4 0.5 0.4 0.4 0.4 0.4 0.5 0.4 0.5 0.4 0.4 0.5 0.5 0.4 0.4 0.5 0.4 0.4 0.5 0.5 0.4 0.4 0.4
0.990 0.991 0.992 0.997 0.975 0.994 0.973 0.997 0.996 0.995 0.989 0.996 0.982 0.987 0.992 0.993 0.992 0.996 0.936 0.988 0.907 0.954 0.956 0.983 0.934 0.991
1, Coumarin; 2, cinnamic alcohol; 3, cinnamic acid; 4, 2-methoxy cinnamic acid; 5, cinnamaldehyde; 6, 2-methoxy cinnamaldehyde.
All data indicated that RRT was more insensitive. The six characteristic chromatographic peaks could be correctly identified by combining RRT, peak shapes, and their UV absorption characteristics. When wavelength and equipment were determined, the SSDMC method was more steady applied to quantitative analyze all of the six components in CR.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.4 Sample analysis 3.4.1 Similarity analysis of HPLC fingerprint Twenty-six batches of samples from different traditional locations of China were investigated. The common pattern of
www.jss-journal.com
J. Sep. Sci. 2014, 37, 2490–2498
Liquid Chromatography
2497
Figure 3. Dendrogram of clustering of 26 Cinnamomi Ramulus samples. The average between-groups linkage method as the amalgamation rule and the cosine as metric were used to establish clusters.
the HPLC fingerprint of CR was established by the software mentioned in Section 2.4.4. Fourteen peaks because of their acceptable heights and good resolution were selected as characteristic peaks for the identification. Peak 11 (cinnamic acid) was selected as the marker peak because of its abundance in content and appropriate position in the chromatographic profile. Comparing with the reference fingerprint, all similarity values of the chromatograms of the 26 samples were more than 0.998, which could not distinct them well. After carefully analyzing the fingerprint profiles of these samples, the peak area of cinnamaldehyde in CR occupied the largest portion, about 70–80% in sum area of all common peaks, which might impact sharply the calculation of similarity values. Thus peak 13 was overlooked in evaluation of similarity. The similarity values of the chromatograms of the 26 samples without peak 13 were shown in Table 2. All samples from Guangxi province had the bigger correlation coefficients of similarities (ࣙ0.973), while the similarities of most samples from Guangdong and Yunnan province were relatively lower (0.907–0.956) except S20, S24, S26. The results indicated that C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
there was a little difference in internal quality of CR between Guangxi and the other two provinces (Guangdong and Yunnan), but no remarkable difference was showed in Guangxi. 3.4.2 HCA of HPLC fingerprint The HCA shown in Fig. 3 was performed based on the RPAs of the 14 common characteristic chromatographic peaks selected. Supposing an appropriate distance level chosen, the samples could be classified into three quality clusters. Cluster I was formed by S4, S9, S17, S18, S20, S26, in which S26 from Yunnan also was separated alone. Only S25 from Yunnan collected from Shandong province belonged to cluster III, and others were divided into cluster II. The HCA could not distinguish CR based on the producing areas very well, but after analyzing the collecting origins of 26 samples, all samples produced from the traditional producing areas and collected from the south of China belonged to cluster II. CR is harvested in spring and summer and contains abundant volatile oil, which needs to be stored in sealed containers out of heat, www.jss-journal.com
2498
J. Sep. Sci. 2014, 37, 2490–2498
X. Wu et al.
light, and humidity. It is rainy and hot in south of China, while dry and cold in north of China. The northern cities are far from main traditional producing regions. In the process of transportation in long distance, the herbal materials might be inevitably exposed to air or sunlight. It is important that because of being far from producing regions, large reserves are indispensable to cost saving. Supporting Information Table S1 shows that most material might be stored for longer than one year in northern China. The results of HCA might suggest that the different climates of various districts could affect the internal quality of herb medicines somehow, like temperature and humility. CR in China was primly introduced from Vietnam to Guangxi and Guangdong provinces, which were the main traditional producing regions with relatively appropriate environment and rich experience in cultivation, therefore CR produced from Guangxi and Guangdong provinces is more popular with higher quality in the market, especially from Guangxi province. Comprehensively, there was somehow no significant difference in CR from different origins in China. 3.4.3 HPLC quantitative analysis To validate the SSDMC method, 26 batches of CR samples were analyzed, and six active components were calculated (Table 2). There were no remarkable differences between ESM and SSDMC method using RSDs (all ࣘ0.5), F-test (P = 0.954, P ࣙ 0.05, n = 26) and t-test (P = 0.978, P ࣙ 0.05, n = 26).
4 Conclusions In this study, an ameliorative and comprehensive quality assessment method for CR was established. Comparing with the conventional quality assessment standard of CR [6], the present improved method using HPTLC and HPLC fingerprint provides more chemical information for identification. The HPTLC method has the advantages of simplicity, rapidity, and can be visualized, whereas the HPLC fingerprint method bears the advantages of specificity, powerful separation ability, and ability to derive detailed chemical information. Six active marker compounds in CR were simultaneously analyzed quantitatively by the SSDMC method, which helped more for the QC of CR. The SSDMC method was reasonable, reliable, and appropriate to quantitatively control the quality of CR, especially when the reference standards were difficult to obtain. In conclusion, simultaneously qualitative and quantitative analysis of multiple active components using HPTLC and HPLC methods could be more comprehensive and reliable to assess and control the quality of CR. This study was financially supported by Liaoning Innovative Research Team in University (LNIRT, Grant No. LT2013022). The authors have declared no conflict of interest.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
5 References [1] Loi, D. T., Vietnamese Medicinal Plants and Ingredients, Medical Publishing House, Hanoi 2004, p. 862. [2] Han, Y., Jung, H. W., Bae, H. S., Kang, J. S., Park, Y. K., Pharm. Biol. 2013, 51, 961–967. [3] He, K., Li, X., Chen, X., Ye, X., Huang, J., Jin, Y., Li, P., Deng, Y., Jin, Q., Shi, Q., Shu, H., J. Ethnopharmacol. 2011, 137, 1135–1142. [4] Xu, S., Shen, J., Xie, Y., Tradit. Chin. Drug Res. Clin. Pharmcol. 2007, 3, 186–189. [5] Unlu, M., Ergene, E., Unlu, G. V., Zeytinoglu, H. S., Vural, N., Food Chem. Toxicol. 2010, 48, 3274–3280. [6] The State Pharmacopoeia Commission of P. R. China, Pharmacopoeia of P. R. China, Chemical Industry Press, Beijing 2010, p. 259. [7] Zhang, H., Zhou, Q., Cao, J., Wang, Y., Spectrochim. Acta A 2013, 116, 251–257. [8] Belsito, D., Bickers, D., Bruze, M., Calow, P., Greim, H., Hanifin, J. M., Rogers, A. E., Saurat, J. H., Sipes, I. G., Tagami, H., Food Chem. Toxicol. 2007, 45, S1–S23. [9] Liu, S. G., Wang, Y., Peng, G. Y., Chin. Med. Mat. 2007, 2, 174–175. [10] Zhu, X. W., Cai, G., Pan, G. X., Chin. Tradit. Herbal Drugs 2004, 35, 695–696. [11] Yang, S., Jia, Y., Bi, K. S., J. Chin. Pharm. Anal. 2004, 24, 143–146. [12] Springfield, E. P., Eagles, P. K. F., Scott, G., J. Ethnopharmacol. 2005, 101, 75–83. [13] Xie, P., Chen, S., Liang, Y. Z., Wang, X., Tian, R., Upton, R., J. Chromatogr. A 2006, 1112, 171–180. [14] Li, N., Che, Y., Zhang, L., Zhang, J., Zhou, Y., Jiang, Y., Tu, P., J. Chin. Pharm. Sci. 2013, 22, 55–63. [15] Zhang, R., Chen, J., Shi, Q., Li, Z., Peng, Z., Zheng, L., Wang, X., J. Sep. Sci. 2014, 37, 45–52. [16] Kong, W. J., Zhao, Y. L., Xiao, X. H., Jin, C., Li, Z. L., Phytomedicine 2009, 16, 950–959. [17] Liao, H. J., Lai, Z. Q., Su, J. Y., Yi, Y. Y., Li, Y. C., Lai, X. P., Su, Z. R., Lin, Z. X., J. Sep. Sci. 2012, 35, 2193–2202. [18] Funari, C. S., Carneiro, R. L., Andrade, A. M., Hilder, E. F., Cavalheiro, A. J., J. Sep. Sci. 2014, 37, 37–44. [19] Liang, Y. Z., Xie, P., Chan, K., J. Chromatogr. B 2004, 812, 53–70. [20] Zhu, J. J., Wang, Z. M., Ma, X. Y., Feng, W. H., Zhang, Q. W., Chin. Herbal Med. 2012, 4, 157–163. [21] Hou, J. J., Wu, W. Y., Da, J., Yao, S., Long, H. L., Yang, Z., Cai, L. Y., Yang, M., Liu, X., Guo, B. H., Jiang, D. A., J. Chromatogr. A 2011, 1218, 5618–5627. [22] Wang, Z. M., Qian, Z. Z., Zhang, Q. W., Zhu, J. J., Gao, H. M., Wang, Z. T., Chin. J. Chin. Mat. Med. 2011, 36, 657. [23] Li, Y. B., Tanui, E. K., Jin, J., Zhang, Y. J., Anal. Methods 2013, 5, 3741–3746. [24] Gao, X. Y., Jiang, Y., Lu, J. Q., Tu, P. F., J. Chromatogr. A 2009, 1216, 2118–2123. [25] Luo, Z. L., Kong, W. J., Qiu, F., Yang, M. H., Li, Q., Wei, R. W., Yang, X. L., Qin, J. P., J. Sep. Sci. 2013, 36, 699– 705.
www.jss-journal.com
of Pharmacy, Shenyang Pharmaceutical University, Shenyang, China 2 National and Local United Engineering Laboratory for Key Technology of Chinese Material Medica Quality Control, Shenyang Pharmaceutical University, Shenyang, China
Received May 4, 2014 Revised June 12, 2014 Accepted June 15, 2014
J. Sep. Sci. 2014, 37, 2490–2498
Research Article
Quality assessment of Cinnamomi Ramulus by the simultaneous analysis of multiple active components using high-performance thin-layer chromatography and high-performance liquid chromatography A novel and improved method for the quality assessment of Cinnamomi Ramulus was developed and completely validated. The method was established using fingerprint technology and simultaneous quantitative determination of six main marker compounds including coumarin, cinnamic alcohol, cinnamic acid, 2-methoxy cinnamic acid, cinnamaldehyde, and 2-methoxy cinnamaldehyde in the herbal medicine for the first time. A newly developed high-performance thin-layer chromatography method, which achieved simultaneous definition of five marker components by comparing the colors and retardation factor values of the bands in high-performance thin-layer chromatography, was first used for the authentication of Cinnamomi Ramulus. The fingerprints of 26 batches of herbal samples from different regions of China showed very similar chromatographic patterns that were evaluated by similarity analysis and hierarchical clustering analysis. In addition, six marker compounds were simultaneously determined using single standard to determine multiple components by the relative response factors. Compared with the external standard method, the new quantitative method was validated to determine multiple compounds in 26 batches of Cinnamomi Ramulus samples. All results demonstrated that the simple and rapid method could be effectively utilized for the quality control of Cinnamomi Ramulus. Keywords: Cinnamomi Ramulus / Fingerprint / Quality assessment / Relative response factors DOI 10.1002/jssc.201400494
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Cinnamomi Ramulus (CR) is the dried young stem of Cinnamomum cassia Presl (Lauraceae), widely distributed in Vietnam, Myanmar, Laos, and the southern part of mainland China (Guangxi, Guangdong, and Yunnan Province) [1]. CR contains abundant volatile oil with a special pungent taste and scent on account of the volatile, cinnamaldehyde. It also contains phenols, organic acids, polysaccharide, glycosides, coumarin, and tannins. As one of the most important TraCorrespondence: Dr. Qing Li, School of Pharmacy, Shenyang Pharmaceutical University, Shenyang 110016, China E-mail: [email protected] Fax: +86-24-23984392
Abbreviations: CR, Cinnamomi Ramulus; ESM, external standard method; HCA, hierarchical clustering analysis; HPTLC, high-performance thin-layer chromatography; RPA, relative peak area; RRF, relative response factor; RRT, relative retention time; SSDMC, single standard to determine multicomponent C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ditional Chinese Medicines, CR is widely used for treating dyspepsia, gastritis, blood circulation issues, and inflammatory disease [2–5]. In Chinese Pharmacopoeia (2010 edition), the authentication and quantitation of one marker component (cinnamaldehyde) are set as the parameter for the QC of CR [6]. However, other compounds such as coumarin, cinnamic alcohol, and cinnamic acid also exhibit specific biological effects [7, 8] and are relatively abundant in CR herbs, indicating multiple components analysis is necessary for CR QC. At present, only a few papers [9–11] have focused on the quantitative analysis of one or several constituents that were far from enough to control the overall quality of CR. Chromatographic fingerprint analysis with multipeak information content of a chromatogram has been introduced as a rational strategy for assessing herbal samples. Compared with the traditional approaches that select one or more compounds as active markers for identification and quality assessment, fingerprint approaches rely on the inherent
Colour Online: See the article online to view Fig. 1–3 in colour. www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 2490–2498
relationships between multiple compounds to display the chemical pattern of herbal sources [12–18]. Thus, chromatographic fingerprint is highly recommended for application to QC of herbal medicines. Furthermore, in order to comprehensively control the quality of Traditional Chinese Medicines, the quantitative analysis of multicomponents is requisite. But the traditional external standard method (ESM) needs relevant multiple reference standards to calculate the concentrations of components quantified. These reference standards require sufficient chemical purity. In addition, they are mostly expensive and insufficient, especially when the component is of low content and hard to be purified from the plant [19, 20], such as cinnamic alcohol and 2-methoxy cinnamic acid in CR. Moreover, some constituents of herbal medicines become unstable when they are purified from a complicated matrix, like cinnamaldehyde in this study. Cinnamaldehyde is light sensitive and easy to be oxidized because of the structure of olefine aldehyde. To overcome this problem, a simple quantitative method has been successfully applied for the QC of botanicals with a single standard to determine multicomponents (SSDMCs) [20–22]. The SSDMC method uses a single internal standard to simultaneously quantify other components by relative response factors. The method has also been successfully applied in QC of botanical extracts and dietary supplements according to USP, European Pharmacopoeia, and Chinese Pharmacopoeia [21, 23–25]. In the present study, an improved quality assessment of CR based on multiple components was established using high-performance thin-layer chromatography (HPTLC) and HPLC methods. Compared with TLC, HPTLC provides better separation of samples and higher performances, which as a versatile, high-throughput, and cost-effective technology is uniquely suited to assessing the identity and quality of botanical materials. The chromatograms obtained by HPTLC and HPLC methods could show adequate qualitative information for the authentication and identification of CR. In addition, six active components in CR were successfully quantitatively determined by SSDMC method for the first time. Cinnamic acid was chosen as an internal reference substance to determine five other components by their UV relative response factors at specific wavelength. In order to ensure the response factors reliability, some environmental and operational factors were investigated in a one-variableat-a-time procedure. The SSDMC method was successfully applied for the quantitative analysis of 26 batches of CR samples. In other words, the quantitative analysis combined with qualitative evaluation offered comprehensive and reliable information for the QC of CR.
2 Materials and methods 2.1 Chemicals and materials Acetonitrile (Fisher Scientific, USA, A452–4), methanol (Shandong Yuwang Chem, China), and phosphoric acid C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2491
(Damao Chem Tech, China) were of HPLC grade; cyclohexane, ethyl acetate, and acetic acid (Damao Chem Tech) were of analytical grade; water was redistilled. Coumarin, cinnamic acid, and protocatechuic acid were bought from Chengdu MUST Bio-technology; cinnamic alcohol, cinnamaldehyde, and 2-methoxy cinnamic acid were bought from Aladdin; 2-methoxy cinnamaldehyde was bought from Shanghai Siyu Chemical Technology. The identity and purity of all seven reference standards were confirmed by HPLC. Twenty-six batches of CR samples meeting the requirements of the Pharmacopoeia of the People’s Republic of China (2010 edition) were collected from different provinces in China. All samples were identified by Professor Ying Jia and deposited in the Centre of Chinese Material Medica, Shenyang Pharmaceutical University. They were labeled as Nos. 1–18 from Guangxi province, Nos. 19–24 from Guangdong province, and Nos. 25, 26 from Yunnan province. Nos. 1, 8, 19, and 21–24 were collected from Guangdong province. Nos. 3, 10–13, and 15 were collected from Guangxi province. Nos. 4 and 20 were collected from Anhui province. Nos. 6 and 26 were collected from Shandong province. Nos. 14, 17, and 18 were collected from Liaoning province. Nos. 2, 5, 7, 9, 16, and 25 were collected from Hebei, Fujian, Yunnan, Heilongjiang, Jiangxi, and Hubei province, respectively.
2.2 Apparatus Three kinds of TLC plates were investigated, which were TLC plate silica gel GF254, HPTLC plate silica gel GF254 (20 cm × 10 cm, Qingdao Haiyang Chemical Industrial), and HPTLC plate silica gel 60 F254 (20 cm × 10 cm, Merck, 34271 Darmstadt, Germany). The Shimadzu 20A Prominance UFLC XR LC System (Shimadzu Corporation, Japan) comprised a binary solvent delivery system, an online degasser, an autosampler, a column temperature controller, and photodiode array detector coupled with lab solution software. Additional comparative analysis was done on an Agilent 1260 HPLC System (Agilent Technologies, USA) comprising a quaternary solvent delivery system, an online degasser, an autosampler, a column temperature controller, and photodiode array detector coupled with an analytical workstation (Chemstation For LC 3D Systems A10.02).
2.3 HPTLC method for identification One gram of powdered herbal sample was extracted with 10 mL 70% methanol for 20 min by sonication. After centrifugation, the supernatant was prepared for HPTLC analysis. Test solutions (10 L) were applied to the HPTLC silica gel 60 F254 plate. The plate was then developed in a suitable double-trough chamber with a glass lid. The chamber was pre-equilibrated with the mobile phase for 20 min. After developing the plate in a chamber containing developing www.jss-journal.com
2492
X. Wu et al.
J. Sep. Sci. 2014, 37, 2490–2498
Figure 1. HPTLC result of Cinnamomi Ramulus visualized under 254 and 365 nm. (1) Cinnamic acid (Rf = 0.15); (2) cinnamic alcohol (Rf = 0.28); (3) coumarin (Rf = 0.39); (4–7) sample solutions of three batches; (8) 2-methoxy cinnamaldehyde (Rf = 0.51); (9) cinnamaldehyde (Rf = 0.56); 18⬚C; 30% (HPTLC, Yantai).
solvent system 1 (cyclohexane/ethyl acetate, 7:2, v/v) over a path of 8 cm, the plate was air-dried, developed the same plate in a chamber containing developing solvent system 2 (cyclohexane/ethyl acetate/acetic acid, 7:2:0.1, v/v/v) over a path of 3 cm. The plate was dried and examined under UV light at wavelengths of 254 and 365 nm (Fig. 1).
2.4 HPLC method for fingerprint and quantitative determination
2.4.2 Preparation of standard solution The standards were accurately weighed and dissolved in methanol to obtain a mixed standard stock solution (No. 1) containing coumarin (0.02900 mg/mL), cinnamic alcohol (0.00925 mg/mL), cinnamic acid (0.05088 mg/mL), 2methoxy cinnamic acid (0.004000 mg/mL), cinnamaldehyde (0.5057 mg/mL), and 2-methoxy cinnamaldehyde (0.07056 mg/mL). Samples (4.0, 3.0, 2.5, 2.0, 1.0, 0.5 mL) of standard stock solution were quantitatively transferred to six 5 mL amber volumetric flasks (Nos. 2–7), diluted with methanol to volume, and mixed. All standard solutions were kept at −20⬚C.
2.4.1 Preparation of sample solution 2.4.3 Chromatographic conditions The powdered herbal sample (0.50 g) was extracted with 25 mL 70% methanol for 30 min by sonication, allowed to cool, and adjusted to the initial weight by adding a mixture of 70% methanol as needed. After centrifugation, the supernatant was filtered through a 0.22 m membrane into an HPLC amber sample vial for HPLC analysis. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The separation was performed on a Luna C18 column (4.6 × 250 mm, 5 m, Phenomenex, CA, USA) protected by a Security Guard C18 guard column (4.0 × 3.0 mm, 5 m, Phenomenex) with a sample injection volume of 10 L. And the separation was conducted at 25⬚C with a flow rate of www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2014, 37, 2490–2498
1.0 mL/min. The mobile phase consisted of solvent A, acetonitrile and solvent B, 0.05% aqueous phosphoric acid. The detection wavelength was set at 265 nm. The linear gradient profile for qualitative assessment was 10–20% B for 10 min, 20–50% B for 25 min, 50% B for 5 min. For quantitative analysis, the linear gradient profile was further optimized to be 25% B for 1 min, 25–38% B for 20 min, 38–40% B for 9 min, 40–48% B for 5 min. 2.4.4 Evaluation methods All samples were analyzed by Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine (Version 2009A). A similarity index of each sample was calculated based on correlation coefficient value against simulative reference standard fingerprint. For further study, hierarchical clustering analysis (HCA) was employed by software SPSS 19.0. Cinnamic acid was chosen the internal standard reference component as it was stable, available, and had appropriate peak area and retention time in chromatogram. 2.4.5 Data analysis RRF is the relative response factor of component to be measured, which is a mean value calculated as reporter previously based on the linearity data [21]. RRF =
Ax × Ci Ai × Cx
(1)
Ax and Cx are the peak area and concentration of the internal reference standard, cinnamic acid in this study. Ai and Ci are the peak area and concentration of other reference standards, coumarin, cinnamic alcohol, 2-methoxy cinnamic acid, cinnamaldehyde, and 2-methoxy cinnamaldehyde. The relative retention time (RRT) is the ratio of retention time of the reference standard (Tx ) and the internal reference standard (Ti ). RRT =
Tx Ti
(2)
2.5 Validation of the quality assessment The HPTLC method was validated on its specificity, stability, and robustness. The methodology of HPLC fingerprint was validated on its repeatability, precision, and stability. To ensure the validity of this new method (SSDMC), validation tests for linearity, LOD, LOQ, accuracy, precision (intraday and interday variability), and stability were performed. Validation was assessed by calculating the RSDs. The results obtained by SSDMC were compared to the results obtained by ESM. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2493
3 Results and discussion 3.1 Optimization of the chromatographic conditions for HPTLC analysis In the study, different extraction solvents including absolute ethyl alcohol, ethyl ether, methanol, 70% methanol, and 50% methanol were investigated. Five dark components of sample solution extracted by 70% methanol showed clearest bands in HPTLC. Different volumes of application (5–15 L) and bandwidths (2–10 mm) were tested. Results showed that increasing the volume or decreasing bandwidths of sample application, the separation between two adjacent bands as cinnamaldehyde and 2-methoxy cinnamaldehyde on the top of the chromatogram would be reduced. Finally, an appropriate chromatogram was obtained by the application of 10 L of sample solution at 10 mm. The developing method including the developing solvent system and the developing procedure was investigated. Different types of developing solvent system such as petroleum ether (60–90⬚C)/ethyl acetate, cyclohexane/ethyl acetate, cyclohexane/acetone, cyclohexane/methanol, cyclohexane/ethyl acetate/methanol with different ratios were investigated. Results showed that developed plates obtained by developing solvent system of cyclohexane/ethyl acetate (7:2, v/v) over a path of 8 cm could move and separate most compounds well except a polar compound cinnamic acid with obvious trail and band close to the baseline. Some molecules of cinnamic acid with the carboxyl in CR might be partially dissociated in the developing procedure. Adding acetic acid in the developing solvent system such as cyclohexane/ethyl acetate/acetic acid (7:2:0.1, v/v/v) could help to move cinnamic acid off the baseline without trail, whereas two nonpolar compounds cinnamaldehyde and 2-methoxy cinnamaldehyde with similar polarity on the top of the chromatogram could not be well separated. The Rf values of most nonpolar compounds were reduced simultaneously. Finally, a second developing method was taken in to account and all components in CR showed a good separation (Section 2.3).
3.2 Optimization of the chromatographic conditions for HPLC analysis The detection wavelength was crucial for developing a reliable fingerprint and for accurate quantitative analysis of marker compounds in the herb. The solution prepared was scanned in the entire UV range (200–400 nm). The maximum absorption wavelength of cinnamic alcohol is 202 nm, which was very different from the others (coumarin 277 nm, cinnamic acid 277 nm, 2-methoxy cinnamic acid 276 nm, cinnamaldehyde 290 nm, 2-methoxy cinnamaldehyde 287 nm). But all components showed an appropriate absorption at 265 nm simultaneously with smooth baseline. At this wavelength, more characteristic peaks in the fingerprint chromatogram were observed, and the marker compounds were eligible for being detected in the HPLC quantitation. www.jss-journal.com
2494
X. Wu et al.
For good resolution within a short analysis time, the HPLC separation conditions, such as mobile phase system, pH of mobile phase, program of mobile phase, flow rate, column temperature, and injection volume were further optimized. Obviously, acetonitrile could accelerate the elution with a suitable separation of all active components. Adding phosphoric acid helped to improve peak symmetry, especially for cinnamaldehyde. There was no sharp effect upon changing the column temperature (20–30⬚C), flow rate (0.8– 1.2 mL/min), and injection volume (5–15 L) on either the resolution of the eluted peaks and the peak symmetry. In consideration of short analysis time and most peaks at acceptable resolution, the column temperature, the flow rate, and injection volume were set at 25⬚C, 1.0 mL/min, and 10 L, respectively. A number of mobile phases with different gradients were screened in order to obtain a reliable chromatogram with most peaks at acceptable resolution and balance for the HPLC fingerprint and to obtain baseline separation of each marker compound (resolution factor > 1.5) in a relatively short analytical time for the HPLC quantitation. Finally, two gradient programs (described in Section 2.4.3) were established for the HPLC fingerprint and quantitative analyses of the herb. 3.3 Method validation 3.3.1 Method validation of HPTLC analysis To ensure good reliability and reproducibility, the HPTLC method was validated on its specificity, stability, and robustness. For specificity, five compounds were identified by matching the colors and Rf values of bands with marker standards under 254 and 365 nm (Fig. 1). For stability, four sample solutions placed at room temperature (0, 1, 2, and 3 days) were tested. For robustness, different particle size of silica gel (TLC plates and HPTLC plates), manufacturers of the precoated plates (Yantai, China; Merck, USA), different temperature (10, 20, and 30⬚C), and different humidity (30, 60, and 90%) were investigated. Results showed that the sample solution was stable for three days and different size of silica gel, different manufacturers, temperature, and humidity among test range had no significant influence on the chromatograms of standard and sample solutions. But the best chromatography could be conducted by HPTLC plates (Merck) at low temperature and humidity about 10⬚C, 40%. 3.3.2 Method validation of HPLC fingerprint The standardized HPLC fingerprint of CR was generated through the average method from the general comparison of all 26 CR samples shown in Fig. 2A. Then, 14 common chromatograph peaks were obtained by the aforementioned software for similarity analysis for identification and peak 11 (cinnamic acid) as the interreference peak for calculation of the RRT and relative peak area (RPA) of other common characteristic peaks. There were seven peaks identified by C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Sep. Sci. 2014, 37, 2490–2498
comparing the retention times, UV spectra, and by spiking with chemical reference substances as follows: peaks 3, 9, 10, 11, 12, 13, 14 were protocatechuic acid, coumarin, cinnamic alcohol, cinnamic acid, 2-methoxy cinnamic acid, cinnamaldehyde, and 2-methoxy cinnamaldehyde, respectively. The repeatability was determined by determining five parallel solutions of No. 13, and their RSDs were 0.9–2.1% for RPA and 0.0–0.1% for RRT, respectively. The precision was examined by the analysis of five injections of No. 13 consecutively, and their RSDs were 0.2–3.9% for RPA and 0.0–0.1% for RRT. The stability was evaluated by analysis of sample solution of different time points, i.e. 0, 2, 4, 8, 10, 12, and 24 h. The stability RSDs were 0.6–4.9% (n = 1) for RPA and 0.0– 0.2% (n = 1) for RRT. These results indicated that the above fingerprint method was feasible and the sample solution was stable within 24 h. 3.3.3 Method validation of quantitative analysis by SSDMC method Six marker compounds that included almost all compounds of the relatively high content and main active components in CR were simultaneously determined using SSDMC method (Fig. 2B). The linearity was evaluated by the standard solutions (Nos. 1–7). The calibration curves calculated by plotting the peak area Y of each compound against the concentration x (g/mL) of each compound were Y = 3.675 × 104 x + 4.735 × 103 , Y = 3.606 × 104 x + 2.395 × 103 , Y = 7.717 × 104 x + 1.171 × 103 , Y = 4.242 × 104 x + 3.526 × 103 , Y = 5.193 × 104 x + 215.4 × 103 , and Y = 2.867 × 104 x + 9.95 × 103 for coumarin, cinnamic alcohol, cinnamic acid, 2-methoxy cinnamic acid, cinnamaldehyde, and 2-methoxy cinnamaldehyde, respectively. Within test ranges, all calibration curves showed well linear regressions (R2 ࣙ 0.9998). The LOQ measured based on S/N of 10 was 0.017–0.06 ng/mL for the six compounds, which might be due to the high column efficiency and sensitive detector selectivity under the testing conditions. Accuracy was calculated as the formula: Recovery (%) = 100 × (amount found − original amount) / amount spiked. Prepared sample solutions in three different amount levels (75, 100, 125%) and triplicate experiments at each level. The results were in the range of 93.6–106.1% with RSD < 2.3%. For precision, the intraday variability of this method was assessed using nine test solutions of the same sample (No. 13) covering three different concentration levels (50, 100, 150%) and triplicate experiments at each level. Interday variability was still assessed using nine test solutions prepared by different analysts in one day and one analyst in successive three days. All RSDs of the six components calculated by ESM and SSDMC method were arranged in 0.4–1.8%. The stability of sample solution was carried out by comparing the peak areas of six components in the chromatograph of the same sample solution, after storing at room temperature for different times (0, 2, 4, 6, 8, 12, and 24 h). The RSDs of stability were ࣘ1.8%. www.jss-journal.com
J. Sep. Sci. 2014, 37, 2490–2498
Liquid Chromatography
2495
Figure 2. The standard reference chromatograms of HPLC fingerprint (A) and quantitative analysis (B) for Cinnamomi Ramulus. The mobile phase (acetonitrile/0.05 aqueous phosphoric acid) programs: (A) 0/10/35/40 min, 10/20/50/50 acetonitrile%; (B) 0/1/21/30/35 min, 25/25/38/40/48 acetonitrile%. Seven components identified (3) protocatechuic acid; (9) coumarin; (10) cinnamic alcohol; (11) cinnamic acid; (12) 2-methoxy cinnmic acid; (13) cinnamaldehyde; (14) 2-methoxy cinnamaldehyde.
Compared all contents of six marker components obtained by SSDMC method to those by ESM with t-test, and the result (P = 0.152, P > 0.05) illustrated that there was no significant difference between two methods and the SSDMC method was appropriate for simultaneous quantitative determination of six components in CR.
3.3.4 The robustness and ruggedness of the RRF and RRT of SSDMC In order to apply this SSDMC method to quantitatively analyze CR herb samples in different laboratories, some important factors were slightly varied to investigate the robustness, including flow rate (±0.1 mL/min), injection volume (±5 L), wavelength (±1 nm), column temperature (±5⬚C), acid concentration (±0.03% H3 PO4 ), ratio of organic phase (±1% B), time of gradient (±1 min). The affection was assessed by the RSD of RRT, RRF, and the total content of six components in CR calculated by ESM and SSDMC method. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The results showed that (i) there were no great influence on RRT, RRF, and the total content when slightly changed flow rate, injection volume, column temperature, the ratio of organic phase, and time of gradient in the test; (ii) but a slight variation of the wavelength affected significantly the RRF of cinnamic alcohol (RSD = 18.0%), cinnamaldehyde (RSD = 4.0%), and 2-methoxy cinnamaldehyde (RSD = 5.3%). The RRF of SSDMC method was very sensitive to wavelength variation when the UV spectrum of analytes was not detected at its maximum UV absorption as reported [21]. Thus, the wavelength of UV detector should be controlled strictly in SSDMC method. Through five concentration standard solutions, the ruggedness of RRT and RRF were compared with different equipment shown in Table 1. The results showed that the RRT was little different in two equipment and three columns. But the RRF varied a little in different equipment, which was fluctuated in a relative narrow range. Especially for the RRF of 2-methoxy cinnamaldehyde, its total RSD% of different equipment and columns was about 5.0%, relatively more than others. www.jss-journal.com
2496
J. Sep. Sci. 2014, 37, 2490–2498
X. Wu et al.
Table 1. Ruggedness of the RRT and RRF in Cinnamomi Ramulus, n = 5
Equipment
Column
SHIMADZU 20A
RSD (%) Agilent 1260
1 2 3 1 2 3
RSD (%) Total RSD (%)
Cinnamic alcohol
Coumarin
2-Methoxy cinnamic acid
Cinnamaldehyde
2-Methoxy cinnamaldehyde
RRT
RRF
RRT
RRF
RRT
RRF
RRT
RRF
RRT
RRF
0.82 0.83 0.82 0.7 0.81 0.83 0.81 1.5 1.1
2.11 2.10 2.11 0.3 2.12 2.09 2.11 0.8 0.5
0.90 0.90 0.90 0.0 0.90 0.90 0.90 0.0 0.0
2.10 2.06 2.07 1.0 1.99 1.96 1.97 0.8 3.0
1.14 1.16 1.16 1.0 1.14 1.16 1.16 1.0 0.9
1.82 1.82 1.82 0.0 1.82 1.83 1.83 0.4 0.3
1.26 1.31 1.31 2.3 1.27 1.32 1.32 2.3 2.1
1.48 1.48 1.49 0.4 1.57 1.57 1.58 0.4 3.3
1.52 1.61 1.59 3.0 1.52 1.63 1.60 3.6 3.0
2.68 2.70 2.69 0.4 2.91 3.00 2.91 1.8 5.0
Column: 1, Phenomenex Luna C18(2) 100A (4.6 × 250 mm, 5 m); 2, Agilent SB C18 (4.6 × 250 mm, 5 m); 3, Agilent XDB C18 (4.6 × 250 mm, 5 m). Table 2. The content and the similarity of 26 batches of Cinnamomi Ramulus
No.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Content by SSDMC (%)
Total content (%)
1
2
3 (ESM)
4
5
6
ESM
SSDMC
0.1006 0.0865 0.0531 0.1133 0.0488 0.0408 0.0566 0.0403 0.0874 0.0402 0.0697 0.0611 0.0581 0.0744 0.1305 0.0412 0.0725 0.0781 0.0410 0.0787 0.0983 0.0425 0.0556 0.0524 0.0415 0.0478
0.0137 0.0057 0.0059 0.0032 0.0091 0.0081 0.0235 0.0089 0.0136 0.0040 0.0332 0.0096 0.0204 0.0244 0.0055 0.0048 0.0192 0.0132 0.0179 0.0130 0.0077 0.0310 0.0080 0.0043 0.0068 0.0098
0.1257 0.0864 0.0739 0.1432 0.1009 0.0695 0.0673 0.0616 0.1243 0.0555 0.1150 0.0840 0.0633 0.1346 0.1347 0.0434 0.0892 0.1135 0.0405 0.1102 0.1277 0.0523 0.0441 0.0900 0.1369 0.0877
0.0078 0.0097 0.0051 0.0113 0.0058 0.0100 0.0063 0.0082 0.018 0.0062 0.0079 0.0055 0.0073 0.0124 0.0075 0.0049 0.0107 0.0102 0.0026 0.0119 0.0191 0.0032 0.0049 0.0097 0.0101 0.009
1.880 1.680 1.089 1.403 1.142 0.7965 1.014 0.8170 1.253 0.9382 1.642 1.462 0.9153 1.704 2.738 1.069 0.8615 1.129 1.185 0.9509 2.555 1.175 0.9876 1.024 0.8151 0.7047
0.2370 0.2389 0.1464 0.2927 0.1363 0.1383 0.1821 0.1398 0.2343 0.1380 0.2005 0.1692 0.1698 0.2105 0.3126 0.1002 0.1727 0.2212 0.1144 0.1720 0.6858 0.1216 0.1593 0.1438 0.1295 0.1540
2.380 2.121 1.382 1.979 1.452 1.070 1.357 1.082 1.741 1.190 2.081 1.803 1.241 2.174 3.351 1.272 1.233 1.574 1.411 1.345 3.514 1.435 1.267 1.333 1.146 1.019
2.364 2.107 1.373 1.967 1.443 1.063 1.349 1.076 1.730 1.182 2.068 1.792 1.234 2.160 3.329 1.264 1.226 1.565 1.402 1.337 3.494 1.426 1.259 1.324 1.140 1.013
RSD (%)
Similarity
0.5 0.5 0.5 0.4 0.5 0.4 0.4 0.4 0.4 0.5 0.4 0.5 0.4 0.4 0.5 0.5 0.4 0.4 0.5 0.4 0.4 0.5 0.5 0.4 0.4 0.4
0.990 0.991 0.992 0.997 0.975 0.994 0.973 0.997 0.996 0.995 0.989 0.996 0.982 0.987 0.992 0.993 0.992 0.996 0.936 0.988 0.907 0.954 0.956 0.983 0.934 0.991
1, Coumarin; 2, cinnamic alcohol; 3, cinnamic acid; 4, 2-methoxy cinnamic acid; 5, cinnamaldehyde; 6, 2-methoxy cinnamaldehyde.
All data indicated that RRT was more insensitive. The six characteristic chromatographic peaks could be correctly identified by combining RRT, peak shapes, and their UV absorption characteristics. When wavelength and equipment were determined, the SSDMC method was more steady applied to quantitative analyze all of the six components in CR.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.4 Sample analysis 3.4.1 Similarity analysis of HPLC fingerprint Twenty-six batches of samples from different traditional locations of China were investigated. The common pattern of
www.jss-journal.com
J. Sep. Sci. 2014, 37, 2490–2498
Liquid Chromatography
2497
Figure 3. Dendrogram of clustering of 26 Cinnamomi Ramulus samples. The average between-groups linkage method as the amalgamation rule and the cosine as metric were used to establish clusters.
the HPLC fingerprint of CR was established by the software mentioned in Section 2.4.4. Fourteen peaks because of their acceptable heights and good resolution were selected as characteristic peaks for the identification. Peak 11 (cinnamic acid) was selected as the marker peak because of its abundance in content and appropriate position in the chromatographic profile. Comparing with the reference fingerprint, all similarity values of the chromatograms of the 26 samples were more than 0.998, which could not distinct them well. After carefully analyzing the fingerprint profiles of these samples, the peak area of cinnamaldehyde in CR occupied the largest portion, about 70–80% in sum area of all common peaks, which might impact sharply the calculation of similarity values. Thus peak 13 was overlooked in evaluation of similarity. The similarity values of the chromatograms of the 26 samples without peak 13 were shown in Table 2. All samples from Guangxi province had the bigger correlation coefficients of similarities (ࣙ0.973), while the similarities of most samples from Guangdong and Yunnan province were relatively lower (0.907–0.956) except S20, S24, S26. The results indicated that C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
there was a little difference in internal quality of CR between Guangxi and the other two provinces (Guangdong and Yunnan), but no remarkable difference was showed in Guangxi. 3.4.2 HCA of HPLC fingerprint The HCA shown in Fig. 3 was performed based on the RPAs of the 14 common characteristic chromatographic peaks selected. Supposing an appropriate distance level chosen, the samples could be classified into three quality clusters. Cluster I was formed by S4, S9, S17, S18, S20, S26, in which S26 from Yunnan also was separated alone. Only S25 from Yunnan collected from Shandong province belonged to cluster III, and others were divided into cluster II. The HCA could not distinguish CR based on the producing areas very well, but after analyzing the collecting origins of 26 samples, all samples produced from the traditional producing areas and collected from the south of China belonged to cluster II. CR is harvested in spring and summer and contains abundant volatile oil, which needs to be stored in sealed containers out of heat, www.jss-journal.com
2498
J. Sep. Sci. 2014, 37, 2490–2498
X. Wu et al.
light, and humidity. It is rainy and hot in south of China, while dry and cold in north of China. The northern cities are far from main traditional producing regions. In the process of transportation in long distance, the herbal materials might be inevitably exposed to air or sunlight. It is important that because of being far from producing regions, large reserves are indispensable to cost saving. Supporting Information Table S1 shows that most material might be stored for longer than one year in northern China. The results of HCA might suggest that the different climates of various districts could affect the internal quality of herb medicines somehow, like temperature and humility. CR in China was primly introduced from Vietnam to Guangxi and Guangdong provinces, which were the main traditional producing regions with relatively appropriate environment and rich experience in cultivation, therefore CR produced from Guangxi and Guangdong provinces is more popular with higher quality in the market, especially from Guangxi province. Comprehensively, there was somehow no significant difference in CR from different origins in China. 3.4.3 HPLC quantitative analysis To validate the SSDMC method, 26 batches of CR samples were analyzed, and six active components were calculated (Table 2). There were no remarkable differences between ESM and SSDMC method using RSDs (all ࣘ0.5), F-test (P = 0.954, P ࣙ 0.05, n = 26) and t-test (P = 0.978, P ࣙ 0.05, n = 26).
4 Conclusions In this study, an ameliorative and comprehensive quality assessment method for CR was established. Comparing with the conventional quality assessment standard of CR [6], the present improved method using HPTLC and HPLC fingerprint provides more chemical information for identification. The HPTLC method has the advantages of simplicity, rapidity, and can be visualized, whereas the HPLC fingerprint method bears the advantages of specificity, powerful separation ability, and ability to derive detailed chemical information. Six active marker compounds in CR were simultaneously analyzed quantitatively by the SSDMC method, which helped more for the QC of CR. The SSDMC method was reasonable, reliable, and appropriate to quantitatively control the quality of CR, especially when the reference standards were difficult to obtain. In conclusion, simultaneously qualitative and quantitative analysis of multiple active components using HPTLC and HPLC methods could be more comprehensive and reliable to assess and control the quality of CR. This study was financially supported by Liaoning Innovative Research Team in University (LNIRT, Grant No. LT2013022). The authors have declared no conflict of interest.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
5 References [1] Loi, D. T., Vietnamese Medicinal Plants and Ingredients, Medical Publishing House, Hanoi 2004, p. 862. [2] Han, Y., Jung, H. W., Bae, H. S., Kang, J. S., Park, Y. K., Pharm. Biol. 2013, 51, 961–967. [3] He, K., Li, X., Chen, X., Ye, X., Huang, J., Jin, Y., Li, P., Deng, Y., Jin, Q., Shi, Q., Shu, H., J. Ethnopharmacol. 2011, 137, 1135–1142. [4] Xu, S., Shen, J., Xie, Y., Tradit. Chin. Drug Res. Clin. Pharmcol. 2007, 3, 186–189. [5] Unlu, M., Ergene, E., Unlu, G. V., Zeytinoglu, H. S., Vural, N., Food Chem. Toxicol. 2010, 48, 3274–3280. [6] The State Pharmacopoeia Commission of P. R. China, Pharmacopoeia of P. R. China, Chemical Industry Press, Beijing 2010, p. 259. [7] Zhang, H., Zhou, Q., Cao, J., Wang, Y., Spectrochim. Acta A 2013, 116, 251–257. [8] Belsito, D., Bickers, D., Bruze, M., Calow, P., Greim, H., Hanifin, J. M., Rogers, A. E., Saurat, J. H., Sipes, I. G., Tagami, H., Food Chem. Toxicol. 2007, 45, S1–S23. [9] Liu, S. G., Wang, Y., Peng, G. Y., Chin. Med. Mat. 2007, 2, 174–175. [10] Zhu, X. W., Cai, G., Pan, G. X., Chin. Tradit. Herbal Drugs 2004, 35, 695–696. [11] Yang, S., Jia, Y., Bi, K. S., J. Chin. Pharm. Anal. 2004, 24, 143–146. [12] Springfield, E. P., Eagles, P. K. F., Scott, G., J. Ethnopharmacol. 2005, 101, 75–83. [13] Xie, P., Chen, S., Liang, Y. Z., Wang, X., Tian, R., Upton, R., J. Chromatogr. A 2006, 1112, 171–180. [14] Li, N., Che, Y., Zhang, L., Zhang, J., Zhou, Y., Jiang, Y., Tu, P., J. Chin. Pharm. Sci. 2013, 22, 55–63. [15] Zhang, R., Chen, J., Shi, Q., Li, Z., Peng, Z., Zheng, L., Wang, X., J. Sep. Sci. 2014, 37, 45–52. [16] Kong, W. J., Zhao, Y. L., Xiao, X. H., Jin, C., Li, Z. L., Phytomedicine 2009, 16, 950–959. [17] Liao, H. J., Lai, Z. Q., Su, J. Y., Yi, Y. Y., Li, Y. C., Lai, X. P., Su, Z. R., Lin, Z. X., J. Sep. Sci. 2012, 35, 2193–2202. [18] Funari, C. S., Carneiro, R. L., Andrade, A. M., Hilder, E. F., Cavalheiro, A. J., J. Sep. Sci. 2014, 37, 37–44. [19] Liang, Y. Z., Xie, P., Chan, K., J. Chromatogr. B 2004, 812, 53–70. [20] Zhu, J. J., Wang, Z. M., Ma, X. Y., Feng, W. H., Zhang, Q. W., Chin. Herbal Med. 2012, 4, 157–163. [21] Hou, J. J., Wu, W. Y., Da, J., Yao, S., Long, H. L., Yang, Z., Cai, L. Y., Yang, M., Liu, X., Guo, B. H., Jiang, D. A., J. Chromatogr. A 2011, 1218, 5618–5627. [22] Wang, Z. M., Qian, Z. Z., Zhang, Q. W., Zhu, J. J., Gao, H. M., Wang, Z. T., Chin. J. Chin. Mat. Med. 2011, 36, 657. [23] Li, Y. B., Tanui, E. K., Jin, J., Zhang, Y. J., Anal. Methods 2013, 5, 3741–3746. [24] Gao, X. Y., Jiang, Y., Lu, J. Q., Tu, P. F., J. Chromatogr. A 2009, 1216, 2118–2123. [25] Luo, Z. L., Kong, W. J., Qiu, F., Yang, M. H., Li, Q., Wei, R. W., Yang, X. L., Qin, J. P., J. Sep. Sci. 2013, 36, 699– 705.
www.jss-journal.com
