Chemical fingerprinting and quantitative analysis of a Panax notoginseng preparation using HPLC-UV and HPLC-MS

Background Xuesaitong (XST) injection, consisting of total saponins from Panax notoginseng, was widely used for the treatment of cardio- and cerebro-vascular diseases in China. This study develops a simple and global quality evaluation method for the quality control of XST. Methods High performance liquid chromatography-ultraviolet detection (HPLC-UV) was used to identify and quantify the chromatographic fingerprints of the XST injection. Characteristic common peaks were identified using HPLC with photo diode array detection/electrospray ionization tandem mass spectrometry (HPLC-PDA/ESI-MSn). Results Representative fingerprints from ten batches of samples showed 27 'common saponins' all of which were identified and quantified using ten reference saponins. Conclusion Chemical fingerprinting and quantitative analysis identified most of the common saponins for the quality control of P. notoginseng products such as the XST injection.


Background
Xuesaitong (XST) injection, consisting of total saponins from Panax notoginseng (Sanqi), was widely used for the treatment of cardiovascular and cerebrovascular diseases in China. As total saponins (including ginsenosides and notoginsenosides) in the XST injection are its active ingredients, quality control of total saponins in the XST injection is critical for its safety, efficacy and stability. Single or simultaneous determination of main components of the total saponin extracts from P. notoginseng using high performance liquid chromatography-ultraviolet detection (HPLC-UV) [1][2][3][4][5], high performance liquid chromatography-evaporative light scattering detection (HPLC-ELSD) [6], high performance liquid chromatography-mass spectroscopy (HPLC-MS) [7][8][9][10][11][12][13] have been reported but over half of the total saponins were not quantified in these studies due to the lack of saponin references or poor chromatographic resolution. A comprehensive and systematic quality control of saponin extracts is much needed. Fingerprint analysis is currently developed for quality control in Chinese medicine [14][15][16][17][18][19][20][21][22][23][24][25][26] and has been accepted by the WHO for the assessment of herbal medicines [27]. The State Food and Drug Administration (SFDA) of China requires all herbal medicine-derived injections and related materials to use chromatographic fingerprints [28] in standardization.

Materials and reagents
Acetonitrile and methanol (HPLC grade) were purchased from Merck (Darmstadt, Germany). Acetic acid glacial (HPLC grade) was from Tedia (Fairfield, OH, USA). The water used was purified by Milli-Q system (Millipore, USA). Reference compounds, namely notoginsenoside R 1 , ginsenoside Rg 1 , Rg 2 , Rh 1 , Rb 1 , Rb 2 , Rd, Re, 20(S)-Rg 3 and 20(R)-Rg 3 were purchased from Jilin University (Shenyang, China). The structures of these compounds are shown in Figure 1. Mixed standard stock solution containing accurately weighed reference compounds was directly prepared in 80% aqueous methanol (v/v). Working standard solutions were prepared by diluting the stock solution with 80% aqueous methanol (v/v) to obtain a series of concentrations for the calibration curves.

HPLC-MS n instrumentation and chromatographic conditions
Analysis was performed on an Agilent 1100 series LC system equipped with a binary solvent delivery system, an auto-sampler, a column temperature controller, a photo diode array detector and a Finnigan LCQ Deca XP plus ion trap mass spectrometer (Thermo Finnigan, USA) via an ESI interface. The chromatographic conditions were the same for HPLC-UV as described in the previous section. The operating parameters for MS in the negative mode were as follows: collision gas, ultrahigh-purity helium (He); nebulizing gas, high purity nitrogen (N 2 ); ion spray voltage, -4.5 kV; sheath gas (N 2 ) at a flow rate of 60 arbitrary units; auxiliary gas (N 2 ) at a flow rate of 20 arbitrary units; capillary temperature, 350°C; capillary voltage, -15 V; tube lens offset voltage, -30 V. Full scan data acquisition was performed from m/z 80 to 1800 in MS scan mode. The MS n spectra were obtained with the collision energy for collision-induced dissociation adjusted to 30%-40% of maximum and the isolation width of precursor ions was 2.0Th.

Sample preparation
Ten samples of the XST injection (Batch No. 20090307, 20090510, 20090310, 20081018, 9042213, 20090312, 20090421, 20090512, 20090504, 20090203), manufactured by three Chinese pharmaceutical companies, were obtained either from pharmacies or factories. For HPLC-PDA-MS n analysis, a certain volume of the injection, according to its nominal content of total saponins, was transferred to a 50 ml volumetric flask and was diluted with 80% aqueous methanol (v/v) to obtain total saponins at a concentration of about 1 mg/ml. For HPLC-UV analysis, the injection was diluted with 80% aqueous methanol (v/v) to obtain total saponins at a concentration of about 0.5 mg/ml. Prior to analysis, the sample solutions were filtered through a 0.45 μm nylon membrane (Whatman, Britain). Spiked injection was produced by mixing sample solutions with the reference solutions at the ratio of 1:1.

Data analysis
Data analysis was carried out with Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine (version 2004A, National Committee of Pharmacopoeia, China) recommended by the SFDA.

Optimization of HPLC separation
We optimized the separation conditions including the column, mobile phase, detection wavelength, elution gradient and column temperature in this study. Four reversed-phase columns, Agilent Zorbax Eclipse SB-C 18 columns (250 mm × 4.6 mm, 5 μm; 150 mm × 4.6 mm, 3.5 μm; 100 mm × 2.1 mm, 1.8 μm) and Ultimate™ XB-C 18 column (250 mm × 4.6 mm, 5 μm) were tested. The results showed that all four columns obtained good peak resolutions in 75 min, 75 min, 45 min and 75 min respectively; however, only two columns with the length of 250 mm (Zorbax Eclipse SB-C 18 and Ultimate™ XB-C 18 ) produced more peaks in chromatograms. Ulti-mate™ XB-C 18 column (250 mm × 4.6 mm, 5 μm) was selected in the fingerprint analysis due to its lower cost than Zorbax Eclipse SB-C 18 column.
The effects of mobile phase composition on chromatographic separation were also studied. The cetonitrile/ water system produced more sharp peaks than the methanol/water system; the addition of 0.01% acetic acid in the acetonitrile/water system further improved the peak shape. Moreover, as the retention time of some components such as ginsenoside 20(S)-Rg 3 and 20(R)-Rg 3 was long, gradient elution was used in HPLC analysis. Satisfactory separation was achieved in 75 min.
There was no strong absorption for most of saponins in the region of ultraviolet and visible spectra due to their structural characteristics, eg lack of conjugation groups in the molecular structures. As the end adsorption wavelength 203 nm is suitable for the assay of ginsenosides and notoginsenosides [1][2][3][4][5], it was selected as the detection wavelength in the experiment. Furthermore, the effects of column temperature on chromatographic separation were also examined. Four column temperatures, namely 20, 25, 30 and 35°C were tested.   We found that at 30°C most peaks in chromatography had good resolutions; therefore, 30°C was chosen as the column temperature for the fingerprint analysis.

HPLC-UV fingerprinting of the XST injection
To standardize the fingerprints, we analyzed ten samples using the optimized HPLC-UV method. Peaks found in all ten samples with good resolution were assigned as 'characteristic peaks' and there were 27 characteristic peaks in the fingerprint chromatograms (Figure 2A). The software of Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine was used to evaluate these chromatograms. To exclude the effects of the solvent and baseline fluctuation, we selected the chromatographic data of these ten samples and treated them within the time frame of 28 min to 75 min. The similarities of chromatograms for the ten samples to the reference fingerprints were established using the means of all chromatograms (Additional file 1). The results showed that the ten samples possessed similarities to the reference fingerprints (Additional file 2). While the HPLC-UV fingerprints from different batches and companies varied, the 27 characteristic peaks were common in all samples. Therefore, the detection of these common peaks in HPLC fingerprints is useful in assessing the quality of the XST injection.

Identification of characteristic peaks
HPLC-PDA/ESI-MS n was used for the components analysis and all 27 characteristic peaks were identified. In the ESI-MS experiment, the molecular weight of each peak was also obtained. By comparing with the ESI-MS n data and HPLC retention time of standard sanponins ( Figure 2B and Additional file 3), we identified 10 peaks as notogisenoside Yao et al. Chinese Medicine 2011, 6:9 http://www.cmjournal.org/content/6/1/9 R 1 , ginsenoside Rg 1 , Re, Rb 1 , Rg 2 , Rh 1 , Rb 2 , Rd and 20(S)-Rg 3 , 20(R)-Rg 3 . A total of 17 peaks were identified tentatively with the aid of the ESI-MS n data and HPLC retention time of some saponins from previous reports [1][2][3][4][5][6][7][8][9][10][11][12][13]. All the identification results are shown in Table 1. In addition, The UV spectra of all peaks in the XST injection were obtained from the PDA chromatogram (Additional file 3). The results showed that among all the peaks in the chromatogram of the XST injection no strong UV absorption within the wavelength range from 210 nm to 400 nm was obtained, suggesting that the XST injection consisted of saponins with few other natural components possessing strong UV absorption, such as flavonoids, lignins, anthraquinones and alkaloids.

Determination of the main saponins in the XST injection
As shown in Figure 2A, 27 saponins were well separated, of which 25 were potentially identified ( Table 1). The ratio of total saponin peak area to all peaks (except for solvent peaks and baseline fluctuation in 0-28 min) in the chromatogram of each sample was beyond 95%. Thus, a method for quantification of the 27 saponins should provide a global and systematical evaluation for the quality control of the XST injection. However, it was difficult to obtain the reference compounds for all 27 saponins; we were only able to obtain ten, including notoginsenoside R 1 , ginsenoside Rg 1 , Re, Rb 1 , Rg 2 , Rh 1 , Rb 2 , Rd, 20(S)-Rg 3 and 20(R)-Rg 3 . Some reports [1][2][3] found that the slopes of regression equations for most of the determined saponins, such as notoginsenoside R 2 , R 4 , Fa, ginsenoside Rg 1 , Re, Rf, Rb 1 , Rg 2 , Rh 1 and Rd were approximately negatively correlated to their molecular weights by HPLC-UV at 203 nm (Additional file 4) and that the regression equations of some saponins with similar molecular weights were also close to each other under the same chromatographic condition (Additional file 5, 6, 7, 8 and 9). Ten saponins, namely R 1 , ginsenoside Rg 1 , Re, Rb 1 , Rg 2 , Rh 1 , Rb 2 , Rd, 20(S)-Rg 3 and 20(R)-Rg 3 were quantitatively determined and the rest 17 saponins without standard references were semi-quantified using substitutive  Table 2. The developed analytical method was successfully applied to analysis of ten batches of the XST injection. All of the 27 characteristic peaks were determined simultaneously and the results are in Table 3. In the XST injection, the content of ginsenoside Rb 1 was the most (26.17%-29.60%), followed by ginsenoside Rg 1 (20.50%-25.43%), Rd (6.82%-8.10%), notoginsenoside R 1 (5.29%-6.89%) and ginsenoside Re (2.91%-4.92%). The total content of the five saponins made up 61.69%-71.39% of the total saponins in the XST injection (total saponins nominal: 50 mg/ml). The ten saponins with available standard substances were quantitatively determined and made up 68.46%-75.85% of the total saponins nominal. Thus, combined with the semi-quantification data, 81.81%-95.71% components in the XST injection could be examined.

Conclusion
The fingerprint profiles of ten batches of samples showed 27 characteristic peaks. Ten of these 27 saponins in the XST injections were quantitatively determined with their standard references; the rest 17 saponins were semiquantified with the substitutive standard references.

Additional material
Additional file 1: The chromatogram of similarity analysis of the fingerprints of 10 samples.
Additional file 3: PDA Chromatograms. standard compounds (A) and a XST injection (C), and total ion current chromatograms of standard compounds (B) and a XST injection (D). 1-27 were the characteristic peaks, listed in Table 2 Additional