Comprehensive Profiling and Quantification of Ginsenosides in the Root, Stem, Leaf, and Berry of Panax ginseng by UPLC-QTOF/MS

The effective production and usage of ginsenosides, given their distinct pharmacological effects, are receiving increasing amounts of attention. As the ginsenosides content differs in different parts of Panax ginseng, we wanted to assess and compare the ginsenosides content in the ginseng roots, leave, stems, and berries. To extract the ginsenosides, 70% (v/v) methanol was used. The optimal ultra-performance liquid chromatography-quadrupole time of flight mass spectrometry (UPLC-QTOF/MS) method was used to profile various ginsenosides from the different parts of P. ginseng. The datasets were then subjected to multivariate analysis including principal component analysis (PCA) and hierarchical clustering analysis (HCA). A UPLC-QTOF/MS method with an in-house library was constructed to profile 58 ginsenosides. With this method, a total of 39 ginsenosides were successfully identified and quantified in the ginseng roots, leave, stem, and berries. PCA and HCA characterized the different ginsenosides compositions from the different parts. The quantitative ginsenoside contents were also characterized from each plant part. The results of this study indicate that the UPLC-QTOF/MS method can be an effective tool to characterize various ginsenosides from the different parts of P. ginseng.


Introduction
In the herbal markets and food industries of Korea and other East Asian countries, Panax ginseng C.A. Meyer is one of the most important herbal plants. Its roots have been used as a main constituent in oriental medicine [1,2]. P. ginseng root contains many bioactive compounds including ginsenosides, peptides, polyacetylenic alcohols, fatty acids, and essential oils [3]. Among these compounds, ginsenosides are considered critical ingredients, having various pharmacological effects, such as enhancing the immune system, and anti-tumor and anti-diabetes activities [4][5][6][7]. Each ginsenoside has specific activities, and the efficacy of ginseng is determined by the ginsenoside contents. Due to its biological utility, the production and use of ginsenosides has been receiving increased attention. More than 150 ginsenosides have been isolated and identified from the P. ginseng plant [8,9]. In particular, its distinct parts including the roots, leaves, stems, and berries are considered as good sources of ginsenosides [10][11][12]. The qualitative and quantitative ginsenoside contents may differ depending on the plant part of P. ginseng [13,14]. Hence, to evaluate the efficacy of different ginseng parts and their effective production of ginsenosides, profiling the various ginsenosides from the distinct parts of P. ginseng is required.
To assess the levels of various ginsenosides from a complex P. ginseng extract, a sensitive and reliable analytical method is necessary. To analyze ginsenosides, high performance liquid chromatography (HPLC) coupled with a UV detector [15,16], or an evaporative light scattering detector (ELSD) [17,18] are typically used. However, due to the limited availability of standard samples, only a few ginsenosides have been quantified. Liquid chromatography (LC) coupled with electrospray ionization/mass spectrometry (ESI/MS) has emerged as a robust tool for the comprehensive profiling of various ginsenosides [19][20][21]. Within MS, a full mass scan based on quadrupole time of flight (QTOF)/MS is a sensitive detector used to measure the exact mass values of target and non-target compounds [22,23]. The ultra-performance LC (UPLC) system provides high resolution separation using a small particle size column, and is effective for analyzing various compounds in a complex mixture [24].
In this study, UPLC-QTOF/MS with an in-house library was used to perform a comprehensive profiling of ginsenosides from the roots, leave, stems, and berries of P. ginseng. To construct an in-house library and the quantification of ginsenosides, we used a total of 58 standard samples, including malonyl-ginsenosides (MGRs), isolated in our laboratory. The qualitative and quantitative contents of ginsenosides were characterized in the distinct parts of P. ginseng.

Panax ginseng Samples
Panax ginseng was cultivated in the experimental field of the National Institute of Horticultural and Herbal Science, Rural Development Administration (RDA) at Chungbuk, Eumseong, Korea (36 • N, 127 • E), according to the "ginseng GAP standard cultivation guide" protocol developed by RDA, Republic of Korea [36]. A voucher specimen (NIHHS-14S01) was deposited at the Herbarium of the Department of Herbal Crop Research, National Institute of Horticultural and Herbal Science, Rural Development Administration, Eumseong, Republic of Korea. Six years into its growing season, in the month of July, the roots, stems and leaves, and berries were harvested, and stored in a deep freezer at −80 • C until further processing.

Extraction of Ginsenosides
The samples of ginseng roots, stems and leaves, and berries were dried at 40 • C in a forced-air, convection-drying oven for 48 h after washing, and then weighed. For the ginseng root, the main roots were used for experiments after removing the lateral and fine roots. Each sample was ground (<0.5 mm) using a mixer (Hanil, Seoul, Korea) and thoroughly mixed, and subsamples were homogenized further using a Retsch MM400 mixer mill (Retsch GmbH, Haan, Germany) for analyses. Fine powder was weighed (200 mg), suspended in 2 mL of 70% (v/v) methanol, and ultrasonically extracted for 30 min at 50 • C. Each sample was centrifuged at 13,500 rpm for 5 min. The solution was filtered through a syringe filter (0.22 µm) and injected directly into the UPLC system. Next, MS analysis was performed using a Waters Xevo G2-S QTOF MS (Waters Corp., Milford, MA, USA) operating in negative ion mode. The mass spectrometers performed alternating highand low-energy scans, known as the MS E acquisition mode. The operating parameters were set as follows: cone voltage 40 V; capillary 3.0 kV; source temperature 120 • C; desolvation temperature 550 • C; cone gas flow 30 L/h; and desolvation gas flow at 800 L/h. Accurate mass measurements were obtained by means of an automated calibration delivery system that contained an internal reference [Leucine-enkephalin, m/z 554.262 (ESI-)]. Data were collected between 100 and 2000 m/z.

Data Processing and Multivariate Analysis
All MS E data were collected and processed using UNIFI 1.8 (Waters Corp., Milford, MA, USA). Data within UNIFI 1.8 were passed through the apex peak detection and alignment processing algorithms. The intensity of each ion was normalized with respect to the total ion count to generate a data matrix having RT, m/z value, and the normalized peak area. Charged species, salt adducts, and fragments were all automatically aligned and grouped. The three-dimensional data, including peak number (RT-m/z pair), sample name, and normalized peak areas, were exported to the EZinfo software 3.0.3 (UMETRICS) for PCA. The processed data were also entered into the MetaboAnalyst website (http://www.metaboanalyst.ca) to perform HCA.

UPLC-QTOF/MS with an In-House Library to Profile Various Ginsenosides
The sensitivity of MS-based analytical methods has increased to enable the detection of compounds with limited abundance in biological samples. Thus, LC-MS methods have been widely used to analyze major compounds in various herbal medicines. For the detailed characterization of P. ginseng, assessing the levels of both the high-and low-abundance ginsenosides is important. In our previous study [25], UPLC-QTOF/MS was used to analyze 30 ginsenosides. The optimal UPLC method was effective for separating various ginsenosides. In the identification of ginsenosides, QTOF/MS was also sensitive and reliable for measuring the exact mass values of the compounds. In this study, the LC-MS method was simultaneously applied to profile 58 ginsenosides. Figure S1 represented the molecular structures of these ginsenosides. First, these ginsenoside standards were purchased or obtained from the experiments by isolating and purifying them from P. ginseng. In our laboratory, we collected four MGRs that contained a malonyl residue, M-Rb1, M-Rb2, M-Rc, and M-Rd, which is linked to the glucose in the compounds. Isolating and purifying the MGRs from the extracted mixture of P. ginseng is difficult in practice. Due to their instability when exposed to heat, care must be taken when handling and preserving MGR compounds.
Next, for the comprehensive profiling of various ginsenosides, 58 ginsenoside standards were used to optimize the UPLC-QTOF/MS conditions. By using the UPLC system with an ACQUITY BEH C18 column, these standards were well separated after 30 min at a flow rate of 450 µL/min. The base peak intensity (BPI) chromatograms of the 58 ginsenosides are shown in Figure 1. Among these ginsenosides, 16 isomeric well-separated pairs were found: C 36  . For the exact identification of ginsenosides, we constructed an in-house library using the UNIFI software. Information about the compound's name, molecular formula, retention time (RT), mass accuracy, and adduct was listed in the library for the 58 ginsenosides (Table 1). Next, a validation study was performed to prove the performance of the UPLC-QTOF/MS method for profiling the ginsenosides. For this, the standard curves of the 58 ginsenosides were constructed, and their linearity range and correlation (R 2 ) were calculated. The limits of detection (LODs) of each standard are provided in Table S1. The R 2 of each ginsenoside analysis was, at minimum, above 0.9908, proving the high reliability of the quantitative analysis of the ginsenosides. LODs also showed the high sensitivity of this method for detecting ginsenosides. For example, in the case of Ra3, Ro, M-Rb2, Rd, GyL, GyLI, Rg3, Mc, and Cy, the two picogram (pg) of these compounds was detected by UPLC-QTOF/MS.  Table 1.

Profiling of Various Ginsenosides from the Roots, Stems, Leaves, and Berries of Panax ginseng
In this study, the constructed UPLC-QTOF/MS with an in-house library was applied to profile various ginsenosides from the roots, stems, leaves, and berries of P. ginseng. The numbers of individual samples were as follows: ginseng root (GR) (n = 10), stem and leaf (GS) (n = 10), and berry (GB) (n = 10). To extract various ginsenosides from individual samples, 70% (v/v) methanol was used. The extracts were analyzed using UPLC-QTOF/MS; each piece of data was then processed using the UNIFI software. Figure 2 shows the representative BPI chromatograms of ginsenoside extracts from GR, GS, and GB. UPLC was able to separate the ginsenosides and other metabolites in 30 min. The intensities of several peaks were different dependent on the sample. For example, overall, the GR sample had lower intensity peaks compared to the other samples. The GS sample had higher intensity peaks than the GB sample, eluted for 5-10 min and 18-26 min. However, the GB sample had higher intensity peaks than the GS sample, eluted for 5-10 min. These demonstrated that GR, GS, and GB have distinct metabolite compositions, including ginsenosides.  First, all the processed data were exported to the EZinfo software for principal component analysis (PCA). PCA helps visualize the general clustering trends among individual samples, and the scatter of the points indicates the similarities or differences between the metabolite data of the samples. In the score plot, the GR, GS, and GB were well-differentiated ( Figure 3A). Moreover, in the loading plot, we identified the main variables that differentiate the three groups in the score plot. Each point represented the m/z-RT pairs of molecules. Based on the in-house library, the nine variables of the loading plot were identified as several ginsenosides ( Figure 3B). Next, in the data processing, a total of 39 ginsenosides were identified from the three samples. Using the in-house library, the numbers of identified ginsenosides from three samples were as follows: 27 in GR, 32 in GS, and 29 in GB. The ginsenosides datasets were then subjected to hierarchical clustering analysis (HCA). As a result, the three groups were clearly differentiated. The different levels of the 39 ginsenosides in the ginseng roots, stems and leave, and berries are described in the heatmap in Figure 4. This demonstrated the major ginsenosides that were found in each ginseng part.

Quantification of Ginsenosides from the Roots, Stems and Leaves, and Berries of Panax ginseng
Next, we quantified the 39 ginsenosides in the individual samples. The standard curve of each ginsenoside was applied to determine its actual amount in the ginseng samples. The quantification was based on 10 biological replicate analysis of each ginseng sample (n = 10), and the amounts are described as mean ± SD in Table 2. This displays the qualitative and quantitative ginsenoside contents in the different P. ginseng parts. Previously, Wang et al. has reported the HPLC-UV-based quantification of 12 ginsenosides from ginseng root, rootlet, leaf, and berry [14]. From the five parts such as root, root-hair, rhizome, stem, and leaf, Shi et al. quantified seven ginsenosides by HPLC-UV [13], and Qu et al. quantified 12 ginsenosides by HPLC-ELSD [36]. Compared to these reports, this study detail characterized a total of 39 ginsenosides with high-or low-abundance from ginseng root, stem and leaf, and berry. As our results, ginsenosides, such as nR4, Ra2, Ra3, and Ra1, were uniquely found in GR. Furthermore, GS uniquely contained Rb3, Mc, CY, Rh2, and Rh1-R. F1 and Rg4 were only detected in GB. Although most of unique ginsenosides are low-abundant, their profiles are valuable to characterize the ginsenoside contents of each ginseng part. In the quantification, GR had highest amounts of Rf, Rb1, and Ro, compared to the other samples. GS had highest amounts of Rg1, vR4, F5, Rh1, F3, Rc, Rb2, Rd, CO, F2, Rg3, CK, and Rk1. GB had highest amounts of gRf, nR1, Re, nR2, Rg2, Rg2-R, M-Rb1, M-Rc, M-Rb2, M-Rd, F4, and Rg5. Previously, two papers have reported that GR had the highest amounts of Rb1, GB had the highest amounts of Rg2, and GS had the highest amounts of Rg1, Rh1, Rd, and Rg3 [14,36]. Interestingly, these results have been matched to ours. Rb3 and Rh1-R are also equally analyzed from GS [13,14]. On the other hand, Wang et al. has reported that Re and Rc are mainly consisted in GS and GR, respectively [14]. As our results, GB had highest amounts of Re, and this is matched to another previous paper [37]. The quantification of several ginsenosides can differ depending on ginseng's cultivars and growing conditions. Each ginsenoside has specific activities, and the efficacies of GR, GS, and GB are determined by the ginsenoside contents. Here, we focused on the pharmacological activities of ginsenosides which are mainly or uniquely consisted in each ginseng part. By searching previous literatures [5,6,15,[38][39][40][41][42][43][44][45][46], we found six kinds of activities in 17 ginsenosides (Table S2). In summary, three ginsenosides (Rf, Rb1, Ro) consisted in GR shows anti-inflammation effect and lymphocyte proliferative capacity. Moreover, 10 ginsenosides (Rg1, Rc, Rb2, Rd, F2, Rg3, CK, Rb3, Rh2, Rh1-R) consisted in GS shows six activities (anti-cancer, anti-diabetes, anti-inflammation, lymphocyte proliferative capacity, immunomodulating activity, chemopreventive effect). Lastly, four ginsenosides (Re, Rg2, Rg2-R, F1) consisted in GB shows four activities (anti-cancer, anti-diabetes, anti-inflammation, lymphocyte proliferative capacity). These results demonstrated that not only GR, but also GS and GB are good sources to obtain ginsenosides having pharmacological activities. By detail profiling of various ginsenosides, it was possible to find the efficacies of ginseng root, stems, leaves, and berries.

Conclusions
In conclusion, this study optimized an analytical method based on UPLC-QTOF/MS with an in-house library to profile 58 ginsenosides. Furthermore, using this method, we determined the quantitative contents of 39 ginsenosides from GR, GS, and GB. These results demonstrated that not only the ginseng root, but also the stems, leaves, and berries are good sources of ginsenosides. Each ginsenoside has specific activities. Thus, in the basis of ginsenosides profiling and quantification, we can assess the pharmacological utility of different ginseng parts. Moreover, ginsenoside contents may differ depending on ginseng's cultivars, growing years, growing regions, etc. Finally, the proposed UPLC-QTOF/MS method for ginsenoside profiling is a good tool for evaluating various ginseng samples as sources for ginsenosides.
Supplementary Materials: Supplementary materials are available online. Figure S1. Molecular structures of 58 ginsenosides listed in Table 1. Table S1. Validation study for the analysis of the 58 ginsenosides using UPLC-QTOF/MS. Table S1. Pharmacological activities of 17 ginsenosides which are mainly or uniquely consisted in ginseng root (GR), stem and leaf (GS), and berry (GB).