Quality Evaluation of Apocyni Veneti Folium from Different Habitats and Commercial Herbs Based on Simultaneous Determination of Multiple Bioactive Constituents Combined with Multivariate Statistical Analysis

Apocyni Veneti Folium (AVF) is a kind of staple traditional Chinese medicine with vast clinical consumption because of its positive effects. However, due to the habitats and adulterants, its quality is uneven. To control the quality of this medicinal herb, in this study, the quality of AVF was evaluated based on simultaneous determination of multiple bioactive constituents combined with multivariate statistical analysis. A reliable method based on ultra-fast liquid chromatography tandem triple quadrupole mass spectrometry (UFLC-QTRAP-MS/MS) was developed for the simultaneous determination of a total of 43 constituents, including 15 flavonoids, 6 organic acids, 13 amino acids, and 9 nucleosides in 41 Luobumaye samples from different habitats and commercial herbs. Furthermore, according to the contents of these 43 constituents, principal component analysis (PCA) was employed to classify and distinguish between AVF and its adulterants, leaves of Poacynum hendersonii (PHF), and gray relational analysis (GRA) was performed to evaluate the quality of the samples. The proposed method was successfully applied to the comprehensive quality evaluation of AVF, and all results demonstrated that the quality of AVF was higher than the PHF. This study will provide comprehensive information necessary for the quality control of AVF.


Introduction
Apocyni Veneti Folium (AVF, Luobumaye in Chinese), derived from the dried leaves of Apocynum venetum L. (Apocynaceae), is one of the most popular traditional Chinese medicines (TCMs) and is officially documented in the Chinese Pharmacopoeia. It has been applied to treat cardiac disease, hypertension, nephritis, and neurasthenia for hundreds of years in China [1]. Modern studies in pharmacology and therapeutics have demonstrated that AVF has antihypertensive [2,3], antidepressant [4], hepatoprotective [5], anxiolytic [6], cardiotonic [7], antioxidative [8], and diuretic [9] effects. In Japan and Northern China, tea preparations from AVF have recently become popular because of their health benefits, specifically their ability to improve immunity and delay aging.

Optimization of UFLC Conditions
Notably, aglycones have the property of hydrophobicity, and the glycosides are hydrophilic in flavonoids. Organic acids, amino acids, and nucleosides have high hydrophilicity. Hydrophilic chromatography columns, such as an Acquity UPLC BEH Amide (100 mm × 2.1 mm, 1.7 µm) or a Waters Atlantis T3 (100 mm × 2.1 mm, 5 µm) column, have extremely weak or no retention ability to polar-weak flavonoids. Consequently, taking into account the hydrophilic and hydrophobic radical group, C 18 chromatography columns including a Thermo Acclaim TM RSLC 120 C 18 (150 mm × 2.1 mm, 2.2 µm) and an Agilent ZORBAX SB-C 18 column (250 mm × 4.6 mm, 5 µm) were both compared to test samples. UFLC results showed that the latter was better than the former in separating effect. Different mobile phases (for example, acetonitrile/water, methanol/water, acetonitrile/0.1% aqueous formic acid, methanol/0.1% aqueous formic acid, acetonitrile containing 0.2% formic acid solution/0.2% aqueous formic acid, and methanol containing 0.2% formic acid solution/0.2% aqueous formic acid), flow rates (0.4 mL/min, 0.8 mL/min, and 1.0 mL/min) and column temperatures (25,30, and 35 • C) were examined and compared. Finally, it was determined that a gradient elution using acetonitrile containing 0.2% formic acid as Eluent A and 0.2% formic acid solution as Eluent B at a flow rate of 1.0 mL/min under the column temperature of 30 • C resulted in the desired separation in a short analysis time.

Optimization of MS Conditions
In order to obtain a sensitive and accurate quantitative method, individual solutions of all standard compounds (about 100 ng/mL) were injected separately into the electrospray ionization (ESI) source by a full-scan mass spectrometry (MS) method in direct infusion mode. Each compound was examined in both positive and negative modes to optimize the parameters of fragmentor voltage (FV) and collision energy (CE) with the highest sensitivity. After trial and error inspection, the vast majority of flavonoids and organic acids have a good condition in the negative ion mode. However, amino acids, nucleosides, and five flavonoids (trifolin, kaempferol-3-O-rutinoside, quercetin-3-O-sophoroside, avicularin, and amentoflavone) responded strongly in the positive ion mode. MRM (multiple reaction monitoring) transition from MS/MS spectrum was chosen when the most abundant, specific, and stable fragment ions appeared. Representative extract ion chromatograms of 43 analytes in MRM mode were presented in Figure 1, the detailed information of retention time (t R ), MS information ([M − H] − or [M + H] + ), FV and CE for each analyte was listed in Table 1, and total ion chromatograms (TIC) of AVF and PHF are shown in Figure S1. As shown, catechin-epicatechin, gallocatechin-epigallocatechin, chlorogenic acid-neochlorogenic acid-cryptochlorogenic acid, hyperoside-isoquercetin, and leucine-isoleucine were isomers with the same precursor ion-product ion pairs; therefore, single reference substance was injected to UFLC-QTRAP-MS/MS to determine the compound with the help of t R . As hyperoside had the same t R and precursor ion-product ion pairs with isoquercitrin and both of them were main flavonoids in AVF [22,23], these two analytes were identified as hyperoside/isoquercitrin together. Under the present condition of chromatographic and MS, constituents were identified and calculated by searching the database and the literature.

Optimization of MS Conditions
In order to obtain a sensitive and accurate quantitative method, individual solutions of all standard compounds (about 100 ng/mL) were injected separately into the electrospray ionization (ESI) source by a full-scan mass spectrometry (MS) method in direct infusion mode. Each compound was examined in both positive and negative modes to optimize the parameters of fragmentor voltage (FV) and collision energy (CE) with the highest sensitivity. After trial and error inspection, the vast majority of flavonoids and organic acids have a good condition in the negative ion mode. However, amino acids, nucleosides, and five flavonoids (trifolin, kaempferol-3-O-rutinoside, quercetin-3-Osophoroside, avicularin, and amentoflavone) responded strongly in the positive ion mode. MRM (multiple reaction monitoring) transition from MS/MS spectrum was chosen when the most abundant, specific, and stable fragment ions appeared. Representative extract ion chromatograms of 43 analytes in MRM mode were presented in Figure 1 Table 1, and total ion chromatograms (TIC) of AVF and PHF are shown in Figure S1. As shown, catechin-epicatechin, gallocatechin-epigallocatechin, chlorogenic acid-neochlorogenic acid-cryptochlorogenic acid, hyperoside-isoquercetin, and leucine-isoleucine were isomers with the same precursor ion-product ion pairs; therefore, single reference substance was injected to UFLC-QTRAP-MS/MS to determine the compound with the help of tR. As hyperoside had the same tR and precursor ion-product ion pairs with isoquercitrin and both of them were main flavonoids in AVF [22,23], these two analytes were identified as hyperoside/isoquercitrin together. Under the present condition of chromatographic and MS, constituents were identified and calculated by searching the database and the literature.

Method Validation
All method validations for quantification were performed using the UFLC-QTRAP-MS/MS technique. The calibration curve was obtained by injecting each analyte three times and more than six appropriate concentrations. The data are summarized in Table 2. The determination coefficients (r 2 > 0.9990) indicate that all calibration curves had good linearity within the test ranges. The limits of detection (LODs) and limits of quantitation (LOQs) were measured in the range of 0.31-41.14 ng/mL and 1.02-137.13 ng/mL, respectively. The intra-day, inter-day, repeatability, stability test of the 43 analytes are presented with an RSD less than 5%, and the overall recoveries lay between 95.12% and 104.3%, with an RSD of no more than 4.76%. The approach of a calibration curve using standard samples with known analyte concentration was built to account for matrix effects. The slope ratio values of the matrix curve to pure solution curve are between 0.95 and 1.06, indicating that the matrix effect on the ionization of analytes was not obvious under optimized conditions. The results are shown in Table 2.

Sample Determination
Forty-three batches of samples from different habitats and commercial herbs are listed in Table 3. It can be seen that the origin of AVF was Tianjin, Xinjiang, Jilin, and Jiangsu, respectively, and the origin of PHF was Xinjiang, Tianjin, Guangxi, Hebei, and Nei Menggol, respectively. The developed UFLC-QTRAP-MS/MS method was subsequently applied to the comprehensive quality evaluation of samples of AVF and PHF. The results of the quantitative determination of the 43 analytes from these samples summarized in Table S2 indicated that significant differences were shown in bioactive constituents between these two species. AVF mainly contained hyperoside/isoquercitrin, quercetin-3-O-sophoroside, epicatechin, gallocatechin, chlorogenic acid, neochlorogenic acid, cryptochlorogenic acid, asparagine, and glutamine. The representative composition of AVF was hyperoside/isoquercitrin, with its content up to 5118.49 µg/g, much higher than that in PHF. While the main constituents of PHF were hyperoside/isoquercetin, quercetin-3-O-sophoroside (as high as 12,745.06 µg/g, about twice as much as that in AVF), epicatechin, epigallocatechin, and chlorogenic acid. Great differences were displayed between these two species in the contents of amenoflavone, chlorogenic acid, trifolin, neochlorogenic acid, astragalin, 2 -deoxyguanosine, glutamic acid, asparagine, isoleucine, and quercetin-3-O-sophoroside. Moreover, in addition to quercetin-3-O-sophoroside, the contents of other significant variation ingredients in AVF were much higher than that in PHF.
The total contents of flavonoids, organic acids, amino acids and nucleosides in AVF and PHF range from 25,473.72 to 21,385.70 µg/g and from 21,003.32 to 17,761.11 µg/g, respectively; 15 flavonoids range from 13,592.62 to 10,396.30 µg/g and from 17,111.8 to 14,053.08 µg/g, respectively; 6 organic acids range from 11,175.56 to 8486.02 µg/g and from 3097.35 to 1705.86 µg/g, respectively; 13 aminos acids range from 2444.02 to 1375.43 µg/g and from 994.09 to 414.47 µg/g, respectively; and 9 nucleosides range from 177.95 to 84.43 µg/g and from 129.52 to 29.90 µg/g, respectively. As can be seen in the above, significant differences were exhibited based on these bioactive constituents between AVF and PHF.

PCA of Samples
PCA based on the content of multiple bioactive constituents was used to classify and differentiate AVF and PHF. In the PCA scores plot ( Figure 1A), each coordinate represents a sample, and it could be observed that the determined samples were clearly divided into two clusters: AVF was mostly located in a PC 1 positive axis; however, PHF was mostly distributed in a PC 1 negative axis. The first two principal components (PC 1 and PC 2) with >80% of the total variance were extracted for analysis. Among them, PC 1 and PC 2 accounted for 75.0% and 4.67% of the total variance, respectively. In the PCA loading plot (Figure 1B), the ions detected with large loading values can be considered as those that strongly contribute to the classification of those samples. The constituents of apigenin, amenoflavone, asparagine, trifolin, epigallocatechin, quercetin-3-O-sophoroside, and kaempferol-3-O-rutinoside were suggested to be diagnostic chemical markers for differentiating Luobumaye samples. As can be seen, PCA might be an effective method of evaluating the quality of AVF and PHF by virtue of the distribution on the PC axis.

GRA of Samples
GRA provides a reliable guarantee for the quality evaluation of traditional Chinese medicine. GRA was performed to evaluate the variation of AVF and PHF on the basis of the contents of 43 constituents.
The gray comprehensive evaluation values (r i ) and quality-rankings are listed in Table 4. Obviously, from a quality point of view, AVF is better than PHF, with r i values ranging from 0.5895 to 0.4653 and from 0.3780 to 0.2937, respectively, and there was no quality crossover between these two species. In general, the quality level of AVF from the main producing area of Tianjin was quite different, and the same was true for PHF from the main producing area of Xinjiang. The quality of PHF from Guangxi, compared to that from other regions, does not appear to be as high. Perhaps, these imply that geographical location and growth environment have a certain impact on the quality of AVF and PHF. In summary, the quality of Luobumaye samples can be successfully evaluated by GRA based on the contents of their multiple constituents.  Table 3.

Plant Materials
Forty-one samples of AVF (S1-S22) and PHF (S23-S41) were obtained in different pharmacies or local Chinese herbal medicine market. Detailed information is shown in Table 3. The botanical origins of the materials were identified by Professor Xunhong Liu (Department for Authentication of Chinese Medicines, Nanjing University of Chinese Medicine, Nanjing, China), and the voucher specimens, named the same as those in Table 3, were deposited at the Herbarium in School of Pharmacy, Nanjing University of Chinese Medicine, China.

Preparation of Standard Solutions
A mixed standard stock solution containing 43 reference substances of 1-43 was prepared by dissolving them in water and their concentrations were as follows: (1)

Preparation of Sample Solutions
The dried sample powders of AVF and PHF of about 0.5 g, after they were passed through a 40 mesh sieve, were weighed accurately, ultrasonically extracted with 50 mL of water for 45 min, cooled at room temperature, and then supplemented with water to compensate for the lost weight. The extract solution was subsequently centrifuged at 12,000 rpm for 10 min. The supernatant was stored at 4 • C and filtered through a 0.22 µm membrane (Jinteng laboratory equipment Co., Ltd., Tianjin, China) prior to analysis.

Mass Spectrometric Conditions
Mass spectrometry detection was performed using an API5500 triple quadrupole mass (AB SCIEX, Framingham, MA, USA) equipped with an electrospray ionization (ESI) source operating in both positive and negative ion modes, and all data were acquired and analyzed by the Analyst 1.6.2 software. The ESI-MS spectra were acquired in the MRM. The parameters in the source were set as follows: GS1 flow: 55 L/min, GS2 flow: 55 L/min, and CUR flow: 40 L/min; gas temperature: 550 • C; pressure of the nebulizer of MS: 4500 V for the positive ion mode and −4500 V for the negative ion mode.

Linearity, LOD, and LOQ
The standard solution containing 43 reference substances was prepared and diluted with water to appropriate concentrations for the construction of calibration curves. Calibration curves were developed by plotting the peak areas versus the corresponding concentrations of each analyte in working standard solution. The LOD and LOQ of 43 constituents were measured at signal-to-noise (S/N) ratios of 3 and 10, respectively.

Precision, Repeatability, Stability, Accuracy, and Matrix Effect
The intra-and inter-day variations were used to evaluate the precision of the established method with the relative standard deviation (RSD) of the peak area as a measure. To confirm the repeatability, six different analytical sample solutions prepared from the same sample were analyzed and expressed by RSD. A stability test was further performed to analyze the variations in the sample solutions at 0, 2, 4, 8, 12, and 24 h, respectively.
A recovery test was used to evaluate the accuracy of the established method. The test was performed in triplicate by adding the corresponding constituents of a known amount at different levels (low, medium, and high, respectively) to 0.5 g of S19. The mixture was extracted and analyzed using the aforementioned method.
Matrix refers to the components of a sample other than the analyte of interest. The matrix can have a considerable effect on the way the analysis is conducted and the quality of the results obtained, for example, ion suppression or enhancement; such effects are called matrix effects. In this study, the matrix effects were evaluated by the slope comparison method. The standard addition method was used by analyzing the sample extracts spiked with appropriate amounts of standards. The slopes of the standard addition calibration curves were then compared with the slopes of the pure water standards. The slope ratio (slope matrix/slope solvent) was calculated to study the matrix effect. When the slope ratio was 1, the matrix does not suppress or enhance the response of the MS; otherwise, ionization suppression or enhancement was indicated [31,32].

Multivariate Statistical Analysis
Multivariate statistical analysis was introduced to analyze the multi-index, multivariate, and other related data that appeared in the study of quality control of traditional Chinese medicine, and the relationship between the laws and data hidden in it can be determined to achieve the effective quality of traditional Chinese medicine evaluation. The combination of the UFLC-QTRAP-MS/MS along with the multivariate statistical analysis details of the samples proved able to identify and select the important markers in samples, even at low concentration levels [33]. PCA is a statistical procedure that uses an orthogonal transformation to convert a set of observations of possibly correlated variables into a set of values of linearly uncorrelated variables called principal components. The normalized data of 41 samples of AVF and PHF tested by UFLC-QTRAP-MS/MS were subjected to PCA to evaluate the differences in chemical composition. Samples were classified on the basis of the content of 43 target constituents using SIMCA-P 13.0 software. GRA, a multivariate statistical analysis method, was introduced to evaluate the quality of the samples of AVF and PHF based on the concentrations of their bioactive constituents using Microsoft Excel. Through the establishment of sample dataset and normalization treatment of raw data, the optimal and the worst reference sequences were conducted. Then, correlation coefficient and correlation degree were calculated, followed by the relative correlation degree (r i ).

Conclusions
In this research, an effective and sensitive method was developed and validated for quality evaluation of AVF based on the simultaneous determination of multiple bioactive constituents combined with multivariate statistical analysis. The optimized UFLC-QTRAP-MS/MS technology was successfully applied to quantify 43 bioactive constituents including 15 flavonoids, 6 organic acids, 13 amino acids, and 9 nucleosides in 41 Luobumaye samples (AVF and PHF) from different habitats and commercial herbs. Furthermore, according to the contents of these 43 constituents, PCA was performed to classify and distinguish between AVF and its adulterant PHF, and GRA implied that the relationship of the two species was applied to evaluate the quality of the samples. All of the results demonstrated that PHF was not as good as AVF from the quality point of view, and perhaps it was not the ideal substitute for AVF. In general, the proposed method was quite useful for the overall assessment of the quality of AVF, and this study may provide the foundation and support for normalization and standardization of AVF.
Supplementary Materials: The following are available online. Figure S1, Total ion chromatogram (TIC) of AVF(A) and PHF(B). Figure S2, Representative extract ion chromatograms of multiple reaction monitoring (MRM) chromatograms of the 43 investigated compounds. Figure S3, Chemical structures of 43 compounds analyzed in this study. Table S1, Effects of extraction method, solvent, solvent to sample ratios, and extraction time on the extraction efficiency of investigated. Table S2, Contents of 43 constituents in samples of AVF and PHF.