Analysis of the Correlation between Commodity Grade and Quality of Angelica sinensis by Determination of Active Compounds Using Ultraperformance Liquid Chromatography Coupled with Chemometrics

The contents of ferulic acid, senkyunolide A, butylidenephthalide, ligustilide, and n-butylphthalide were determined by UPLC analytical method; the correlation among the grade, average weight, and content was explored by correlation analysis and analysis of variance (ANOVA); the different commercial grades with average weight and content were revealed by principal component analysis (PCA) and then rationality analysis grade classification of A. sinensis. The results showed that various commercial grades can be distinguished by PCA analysis. And there was significant negative correlation between the commodity grades and average weight, commodity, and the content of bioactive compounds, while the content of senkyunolide A had significant negative correlation with commodity grades (P < 0.01). Average weight had no correlation with chemicals compounds. Additionally, there was significant positive correlation among the bioactive compounds (content of ferulic acid and phthalides) of different grades of A. sinensis. The content of senkyunolide A, butylidenephthalide, and ligustilide had significant positive correlation with the content of ferulic acid. The content of ligustilide and butylidenephthalide had significant positive correlation with the content of senkyunolide A. The content of ligustilide had significant positive correlation with the content of butylidenephthalide. The basis of grades classification is related with the difference levels of the bioactive compounds.


Introduction
Commercial specifications and grades of Chinese medicinal materials, the summary of practical experience (such as shape, size, texture, color, odor, and taste are based primarily on the human senses) for quality evaluation of traditional Chinese medicine, were the reflect of the quality of traditional Chinese medicine, which had played an important role in market circulation and ensuring clinical safety because it is easy, fast, and effective. However, most of these evaluation methods were still in the stage of experiential description which made it difficult to be inherited and applied. So far, the SFDA (State Food and Drug Administration) only issued 3 times of the commercial specification, involving 76 varieties, compared with the amount of commonly used trading of medicinal materials (more than 600 kinds); it is difficulty to satisfy the market demand [1].
Angelica sinensis (Danggui in Chinese) is one of the main exports of traditional Chinese medicine, derived from root of Angelica sinensis (Oliv.) Diels (Umbelliferae) [2], a well-known Chinese herbal medicine, first documented in Shennong Bencao Jing (Shennong's Classic of Materia Medica; 200-300 AD) [3,4]. It has been used historically as a tonic, hematopoietic, and anti-inflammatory agent for the treatment of gynecological diseases such as menstrual disorders, amenorrhea, and dysmenorrheal for thousands of years in traditional Chinese Medicinal prescriptions [5,6]. It has also been widely marketed as health food for women's care in Asia [7] and as a dietary supplement in Europe and America [8,9]. However, quality assessment in the markets is extremely difficult and impractical when considering large number of export of this product.
Therefore, it is necessary to investigate the grades and quality assessment of A. sinensis [10][11][12]. This research took 2 Evidence-Based Complementary and Alternative Medicine A. sinensis as the study subject. We collected different commodity grades of A. sinensis in the main producing areas and determined the content of ferulic acid and ligustilides, which are thought to be the biologically active components [13][14][15][16][17] by UPLC method for studies of the correlation between grades and quality. The results are helpful to provide the basis for the establishment of medicinal industry standards.

Preparation of Standard and Sample
Solutions. The five reference compounds were accurately weighed: 1 mg dissolved in 10 mL volumetric flask with 70% methanol to produce standard stock solutions. The stock solution was diluted to yield a series of standard solution in the concentration range of 32-1722 g/mL, 30-415 g/mL, 156-722 g/mL, 332-6786 g/mL, and 33-2543 g/mL for ferulic acid, senkyunolide, n-butylphthalide, ligustilide, and butylidenephthalide, respectively. Samples of herbal materials were ground into fine powder then passed through a 20 mesh (0.9 mm) sieve. Sample powder (0.2 g) was accurately weighed and transferred into a 60 mL round bottom flask. 70% methanol (20 mL) was added and refluxed for 30 min. When cool, the methanol was added to compensate for weight loss. After filtering through a 0.22 m filter membrane, the filtrate was ready to be used.

Validation of the UPLC Method
2.6.1. Calibration Curves. Methanol stock solutions of 5 reference compounds were prepared and diluted to appropriate concentrations for the construction of calibration curves. Six concentrations of the solution were analysed in triplicate; the calibration curves were constructed by plotting the peak areas against the concentrations of the analytes. The stock solutions mentioned above were diluted to a series of appropriate concentrations with methanol, and an aliquot of the diluted solutions were injected into UPLC-DAD for analysis. The stock solutions mentioned above were diluted to a series of appropriate concentrations with methanol, and an aliquot of the diluted solutions were injected into UPLC-DAD for analysis. The limits of determinations (LODs) and limits of quantifications (LOQs) under the present conditions were determined at signal to noise (S/N) ratio of about 3 and 10, respectively.
2.6.2. Precision, Accuracy, and Stability. Intra-and interday variations were chosen to determine the precision of the UPLC method. 0.2 g of A. sinensis sample was extracted and analysed as described in Sections 2.3 to 2.5. The intraday variability test was performed by triplicate extraction and analysed during a single day. The interday variability test was carried out on three different days. Variations were expressed by the relative standard deviations (R.S.D.) for intra-and interday. The recovery test was used to evaluate the accuracy of the method. Accurate amounts of ferulic acid, senkyunolide, n-butylphthalide, ligustilide, and butylidenephthalide were added to approximately 0.25 g of A. sinensis sample and then extracted and analysed as described in Section 2.3. The average recoveries were calculated by the following formula: recovery(%) = (amount found − regional amount)/amount spiked × 100%. For the stability test, A. sinensis sample was accurately weighed (approximately 0.5 g) and ultrasonic-extracted with 2.0 mL of methanol for 30 min at room temperature. The samples were then analysed at 0, 2, 4, 6, 8, 12, 24, and 36 h with the established method. The relative standard deviations (R.S.D.) of five data were used to evaluate the stability.

Statistical
Methods. Spearman rank correlation analysis was carried out by SPSS system for windows release version 19.0 (SPSS Institute, Cary, NC, USA), the normality, mean, standard deviation (SD), median, and the outlier range of the A. sinensis; the values of the standard quality characteristics were acquired on the basis of mean ± SD.

Method Optimization.
In sample preparation procedure, multifarious solvents, such as different concentrations (10%, 30%, 50%, 70%, and 90%) of ethanol and methanol, were tested and 70% methanol solution was selected because of its excellent dissolving capacity for A. sinensis sample. For UPLC analysis, two mobile phase systems, including acetonitrilewater and methanol-water, in various proportions were compared and different mobile phase additives, such as phosphate buffer, formic acid, and acetic acid, were also investigated, and finally, 0.1% formic acid aqueous solution and acetonitrile were used as mobile phases which could provide satisfactory separation and peak shapes of investigated compounds. In addition, ferulic acid and phthalides substances absorb different wavelengths under the UV conditions. Ten times difference of response value in the same substance under different wavelengths. 281 nm was employed as the detection wavelength since the difference of response value among ferulic acid, senkyunolide A, n-butylphthalide, and ligustilide compared with the response values under the maximum absorption wavelength was not obvious, while response values under 281 nm are only 1/10 compared with 261 nm, which is difficult to observe in the chromatogram. Therefore, 261 nm was chosen as detection of butylidenephthalide wavelength.

Method
Validation. The linearity, regression, and linear ranges of five analytes were determined using the developed UPLC-PDA method. The data indicated a good relationship between concentrations and peak areas of the analytes within the test ranges ( 2 ≥ 0.9990). The LOQs and LODs of all analytes were less than 4.15 and 11.06 mg/mL. The overall RSDs of intra-and interday variations for analytes were not more than 4.49% and 6.54%, respectively. The established method also had acceptable accuracy with spike recovery of 98.33-102.22% for all analytes. As to stability test, the RSDs of the peak areas for analytes detected within 24 h were lower than 4.72% ( Table 2). All these results demonstrated that the developed UPLC method was linear, sensitive, precise, accurate, and stable enough for simultaneous quantification of the five investigated compounds in A. sinensis.

Quality of Five Analytes in A. sinensis Samples.
The developed UPLC method was successfully employed for simultaneous determination of the five major active components in 19 A. sinensis samples collected from different localities. Typical chromatograms of reference compounds (a) and A. sinensis samples (b) were shown in Figure 3. The chemical structure of the reference compounds were shown in Figure 2. The identification of the investigated compounds was carried out by comparison of their retention time and UV spectra with reference chemicals. The contents of five investigated compounds in 19 A. sinensis samples were summarized in Table 3. 19 batches of different region of A. sinensis were determined, as shown in Figure 4. The sequence of ferulic acid content is Yunnan (0.14%∼0.15%) > Gansu (0.04%∼0.12%) > Sichuan (0.01%∼0.02%) > Hubei (0.005%∼0.006%). The ferulic acid contents in Yunnan and Gansu were accorded with the quality level request in Chinese Pharmacopoeia 2010 (≥0.05%), of which the highest content of ferulic acid was Yunnan sample. While the ferulic acid content of A. sinensis in Sichuan and Hubei was far below the standard of Chinese Pharmacopeias, they were substandard samples.

The Correlation among the Main Chemical Compounds,
Average Weight, and Grades. The grade of whole A. sinensis was divided into five by using grading measurement method. From the data on the correlation coefficients between the main chemical compounds, average weight versus grades in Table 4, respectively, we can see that significantly negative correlation exists between average dry weight versus grades and the content of senkyunolide A versus grades ( < 0.01), which indicates that the higher the grade of A. sinensis, the heavier the weight. From the level of 1, 2, 3, 4, to 5, the average weight decreased gradually; the result coincident with the fact. And the content of senkyunolide A decreased with the increase of grade. Average weight had no correlation with the main chemicals compounds. Table 5 showed that 1-5 grades were analyzed by ANOVA; there were significant differences in content of ferulic acid and phthalides of different grades of A. sinensis ( Table 5). As shown in Figure 5, the content of senkyunolide A, butylidenephthalide, and ligustilide had significant positive correlation with the content of ferulic acid. The content of ligustilide and butylidenephthalide had significant positive correlation with the content of senkyunolide A. The content of ligustilide had significant positive correlation with the content of butylidenephthalide. We found the content of ligustilide somehow related to the content of ferulic acid, senkyunolide A, and butylidenephthalide, which indicates that ligustilide and senkyunolide A serious possibility be one of the pharmacological biomarkers compounds in A. sinensis [18][19][20][21][22].
Evidence-Based Complementary and Alternative Medicine 5        1st  1st  1st  1st  1st  1st  1st  2nd  2nd  2nd  3rd  3rd  3rd  3rd  3rd  4th  4th  4th  4th Butylidenephthalide Ligustilide n-Butylphthalide Senkyunolide A Ferulic acid Grade Content (mg/g) (%) Figure 4: The content of ferulic acid, senkyunolide A, n-butylphthalide, ligustilide, and butylidenephthalide for different commercial grades of A. sinensis samples. Figure 6 shows that A. sinensis samples were divided into 5 main clusters observed in the PCA scores plot. In Figure 6(a), such division indicated that different grades could significantly distinguish from 1st to 5th by the different average weights of A. sinensis, while Figure 6(b) indicated that grades were distinctly separated by different contents of the chemical compounds. This result reflected rationality of grades classification.

Conclusion
In summary, the contents of ferulic acid, senkyunolide A, butylidenephthalide, ligustilide, and n-butylphthalide were determined by UPLC analytical method. The correlation among the grade, average weight, and content was explored by correlation analysis and analysis of variance (ANOVA); the different commercial grades with average weight and content were revealed by principal component analysis (PCA) and then rationality analysis classification of grade and quality of A. sinensis. The results showed that various commercial grades can be distinguished by PCA analysis. And there were significant negative correlation between the commodity grades and average weight, commodity, and the content of bioactive compounds. While the content of senkyunolide A had significant negative correlation with commodity grades ( < 0.01), average weight had no correlation with chemicals compounds. Additionally, there was significant positive correlation among the bioactive compounds (content of ferulic acid and phthalides) of different grades of A. sinensis. The content of senkyunolide A, ligustilide, and butylidenephthalide had significant positive correlation with the content of ferulic acid. The content of ligustilide and butylidenephthalide had significant positive correlation with the content of senkyunolide A. The content of ligustilide had significant positive correlation with the content of butylidenephthalide.
Quality standards for A. sinensis components are described in the Chinese Pharmacopoeia, such as percentage of total ash, acid-insoluble ash, and alcoholic extract content; ferulic acid content is also described. Our results suggest Evidence-Based Complementary and Alternative Medicine 9 that senkyunolide A is a candidate for an efficient marker to define a radix quality standard. This work will provide useful references for the quantification and standardization of quality control for A. sinensis.