Introduction

The Korean ginseng belonging to Panax, family Araliaceae has been known as medicinal herbs and used for functional foods and dietary supplements which can provide various benefits such as immunity-boosting, anti-inflammatory, antioxidative activity, fatigue recovery, blood flow improvement, and memory improvement [1, 2]. For these reason, many researches identified the active ingredient known as ginsenoside and isolated by various analytical techniques such as TLC, GC, HPLC, capillary electrophoresis, and near infra-red spectroscopy [3]. Although simultaneous quantitation of ginsenoside was attempted by HPLC-UVD [4] and UHPLC–MS/MS [5], the HPLC method not only had low sensitivity but also had limited number of target compounds due to the overlapping of compounds. Since multiple reaction monitoring (MRM) mode allows the overlapping of retention time for each compound, the MS/MS with triple quadrupole instrument has widely been used for rapid and simultaneous analysis of many types of compounds. For profiling purposes, UHPLC–MS/MS has been generally used to compare the relative content of ginsenoside [6, 7]. Because the medical activity of ginsenosides depends on its quantities in ginseng or its formulation, a precise quantitation is very critical for evaluating of the medical activity. In this study, a rapid and simultaneous analysis of 26 ginsenosides (Rb1, Rb2, Rc, Rd, Re, Rf, Rg1, Rg2(R), Rg2(S), Rg3(S), Rg3(R), Rg5, Rg6, Rh1(R), Rh1(S), Rh2(R), Rh2(S), F1, F2, F3, F4, K, Mc, PPT(S), XVII, and Y) was achieved quantitatively with 26 min of chromatographic running time using MRM mode on UHPLC-MS/MS, including chromatographic separation of isomeric ginsenosides. To compensate the matrix effect in the UHPLC-MS/MS detection, matrix matched calibration was used and standard addition method was applied for precise quantitation. Various types of columns and elution solvents were compared for efficient, rapid and simultaneous separation of 26 ginsenosides. The UHPLC-MS/MS method developed was successfully applied to quantitation of those in commercial white Panax ginseng (WG) and red Panax ginseng (RG).

Materials and methods

Chemicals and reagents

Twenty-six reference standards (purity ≥ 98.0%) of ginsenosides (Rb1, Rb2, Rc, Rd, Re, Rf, Rg1, Rg2(R), Rg2(S), Rg3(R), Rg3(S), Rg5, Rg6, Rh1(R), Rh1(S), Rh2(R), Rh2(S), F1, F2, F3, F4, K, Mc, XVII, Y, and PPT(S)) were purchased from AMBO (Ambo institute, Seoul, Korea). HPLC grade methanol was purchased from Merck (Darmstadt, Germany) while LC-MS grade of methanol and formic acid were obtained from Merck (Darmstadt, Germany) and Sigma-Aldrich (St. Louis, Mo, USA), respectively.

White and red Panax ginseng

Six-year old white and red Panax ginseng root, cultivated in Geumsan-gun, Chungcheongnam-do, Rerpublic of Korea, were purchased from the commercial market.

Analytical standard solution

Individual stock solutions of 26 reference standards were prepared in methanol at the concentration of 1,000 µg/mL and working standard solutions with the concentrations of 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, and 50 µg/mL were prepared by dilution with methanol.

Selection of LC mobile phase and MRM optimization in UHPLC-MS/MS

UHPLC-MS/MS analysis was carried out on a Shimadzu LC-MS 8040 triple quadrupole mass spectrometer (Kyoto, Japan). Electrospray ionization (ESI) mode was used with de-solvation temperature of 250 °C and heat block temperature of 400 °C. The flow rate of nitrogen for nebulizing and drying gas was 3 L/min and 15 L/min, respectively.

The intensity of precursor ion was obtained in positive and negative full scan mode (50 ~ 1,200 m/z) by injecting of 5 µL standard working solution (10 µg/mL). Two different mobile phase modifiers of 0.1% formic acid in water and 5 mM ammonium formate in water were compared as an UHPLC-MS/MS eluent under the ratio of 50:50 with methanol. The optimal MRM transitions were obtained, using 0.1% formic acid in water/methanol as UHPLC-MS/MS eluent, by dissociation of selected the precursor ions to produce quantifier and qualifier ions (Table 1).

Table 1 Optimized MRM transitions of ginsenosides in UHPLC-MS/MS analysis
Full size table

Separation of 26 ginsenosides

To separate 26 ginsenosides, four kinds of C18 UHPLC columns such as kinetex (2.1 x 100 mm, 2.6 μm, Phenomenex, Torrance, CA, USA), Luna (2.0 x 100 mm, 3 μm, Phenomenex, Torrance, CA, USA), BEH (2.1 x100 mm, 1.7 μm, Waters, Milford, MA, USA), and HSS T3 C18 (2.1 x 150 mm, 1.8 μm, Waters, Milford, MA, USA) were compared with 0.1% formic acid in water (mobile phase A) and methanol (mobile phase B) as mobile phase. The gradient program used for the comparing was as follows: 0–1 min (50% B); 1–45 min (50–95% B); 45–50 min (95% B); 50–50.1 min (95–50% B); 50.1–60 min (50% B). In addition, shorter HSS T3 C18 column (2.1 x 50 mm, 1.8 μm, Waters, Milford, MA, USA) was further tested to minimize analytical time, using the following gradient program: 0–0.1 min (30% B); 0.1–1 min (30–50% B); 1–22 min (50–88% B); 22-23 min (88% B); 23–23.1 min (30% B); 23.1–26 min (30% B) with the flow rate of 0.3 mL/min.

Sample preparation for ginsenosides analysis

Six year-old white and red Panax ginseng roots were freeze-dried for 3 days in a Freeze-Dryer (Il-Shin Lab Tech. Co., Korea), and pulverized by a ball-mill instrument (MM 400, Retsch, GmbH, Haan, Germany). Fifty milligrams of each ginseng powder were extracted with 1 mL of 70% methanol in an ultra-sonic bath for 30 min at 30 °C followed by centrifugation at 15,000 rpm at 4 °C. The supernatant was used for quantitation.

Quantitation by standard addition method

An aliquot (64 μL) of ginseng extract without dilution was used as quantitation group I (QG-I), and another aliquot was diluted by 100 (QG-II), 400(QG-III), and 8000 times (QG-IV) with 70% methanol, respectively. Each of QG-I, -II, -III, and -IV were spiked with working standard solution (16 μL) for standard addition quantitation. Five microliter of QG-I, -II, -III, and -IV was injected and quantitation was performed by calculating x-intercept value when y = 0 of calibration curve equation y = ax + b. In white Panax ginseng, QG-I was applied for ginsenoside F1, F2, F3, K, Mc, PPT(S), Rg1, Rh2(R, S), XVII, and Y, while QG-II for F4, Rg2(R, S) and Rg6. QG-III was for Rd, Rf, Rg3(R, S), Rg5 and Rh1(R, S) and QG-IV was for Rb1, Rb2, Rc and Re. In red Panax ginseng, QG-I was applied for ginsenoside F1, F2, F3, K, Mc, PPT(S), Rh2(R, S), XVII, and Y, while QG-II for F4, Rg2(R, S), and Rg6. QG-III was for Rd, Rf, Rg3(R, S), Rg5, and Rh1(R, S), and QG-IV was for Rb1, Rb2, Rc, Re, and Rg1.

Results and discussion

Selection of LC mobile phase and optimization of the MRM transitions

Since ion adduct types and intensity tend to rely on the mobile phase modifier such as formic acid or ammonium formate for ESI of UHPLC-MS/MS [8], 0.1% formic acid in water and 5 mM ammonium formate in water, were tested to check for the adduct types and intensity of the precursor ion on MS scan mode. In positive scan mode, most precursor ions of ginsenosides showed better intensity in the water with 0.1% formic acid than in the water with 5 mM ammonium formate. In addition to the low intensity in 5 mM ammonium formate, the precursor ions were not even generated for F4, Rb1, Rb2, Rc, Re, Rg5, and XVII. The most of ginsenosides produced sodium adduct precursor ion [M + Na]+ as a precursor ion while PPT(S), Rg3(R), Rg3(S), Rh2(R), and Rh2(S) showed [M + H-2H2O]+ precursor ion as also reported in literature [9]. In negative mode, both formic acid and ammonium formate resulted in similar intensity, for Rf, Rg2(R, S), Rg6, F4, Y, and Rg5 producing [M-H]− precursor ion. Their intensities were higher than sodium adducted precursors [M + Na]+ from the positive mode. Therefore, the mobile phase for the MRM optimization was chosen to water containing 0.1% formic acid (A) and methanol (B) as UHPLC-MS/MS eluent. After collision induced dissociation of precursor ions, the product ion providing the highest intensity was selected as the quantifier ion while the second highest product ion was selected as the qualifier ion (Table 1, Additional file 1: Fig. S1).

Separation of 26 ginsenosides including isomers on various analytical columns

It is very important to quantitate the constitutional and stereisomeric isomer of ginsenosides independently because each isomer was known to have different activity [10,11,12]. Analysis of ginsenoside isomers required the derivatization for GC analysis [13] or 2-dimensional UHPLC-MS/MS for the separation [14]. Separating isomers using MS detection is challenging due to the similarity of the MS spectra and MRM transition between isomers. Therefore, chromatographic separation is required to distinguish the isomers. In this study, we tried to resolve the isomers chromatographically using the C18 column which is commonly used in UHPLC-MS/MS analysis.

Three common C18 columns were tested first because they have widely been used for the analysis of ginsenosides [3, 15,16,17,18]. The stereoisomer of ginsenosides, Rh2(R, S) and Rg3(R, S) were not separated on those three columns, instead, the HSS T3 C18 column (2.1 x 150 mm, 1.8 μm, Waters, Milford, MA, USA) gave a very good resolution for Rg3(S)/Rg3(R), and Rh2(S)/Rh2(R) (Fig. 1), resulting in the successful separation of 26 ginsenosides within 60 min. However, the analytical running time of 60 min with the column of 150 mm length was considered to be quite long, so shorter HSS T3 C18 column of 50 mm length was tried again to analyze in 26 min, saving 34 min of analytical time (Fig. 2b).

Fig. 1
figure 1

Separation of stereoisomeric ginsenoside Rh2 isomers (a), and Rg3 isomers (b) with various C18 columns (kinetex, luna, BEH, and HSS T3)

Full size image
Fig. 2
figure 2

UHPLC-MS/MS TIC chromatogram of 26 ginsenosides on HSS T3 C18 column (2.1 x 150 mm, 1.8 μm) (a) (1; Re, 2; Rg1, 3; Rf, 4; Rg2(S), 5; F3, 6; Rh1(S), 7; Rg2(R), 8; Rh1(R), 9; F1, 10; Rb1, 11; Rc, 12; Rb2, 13; F4, 14; Rg6, 15; Rd, 16; PPT(S), 17; XVII, 18; Rg3(S), 19; Rg3(R), 20; F2, 21; Mc, 22; Y, 23; Rg5, 24; K, 25; Rh2(S), 26; Rh2(R)), HSS T3 C18 column (2.1 x 50 mm, 1.8 μm) (b)

Full size image

Quantitation of 26 ginsenosides in red and white Panax ginseng using matrix matching and standard addition method

In HPLC-UVD analysis of ginsenosides, solvent standard solutions have been used for quantitation as the external calibration method thanks to the lack of matrix effect [19]. However, in the UHPLC-MS/MS analysis, the matrix-matched calibration to compensate the matrix effects by a co-extracted matrix, which is a general and inevitable phenomenon, is an essential procedure for precise calibration and quantitation [8]. Although several studies used UHPLC-MS/MS quantitation, their methodology for the matrix matching or the quantitation was not described in detail [5, 18, 20].

The endogenous concentrations of ginsenosides in sample extracts that are already present in the extract makes the quantitation difficult since the external standard calibration using the solvent standard cannot be applicable for the quantitation. Therefore, we decided to use the standard addition method for precise calibration and quantitation. SANTE/11813/2017 guideline [21] suggested a standard addition method in which calibration curve range is in less than 5 times of the area of original level to prevent distortion of the calibration curve. Another problem is the large difference of each ginsenoside concentration in ginseng, ranging from a few mg/kg to several thousands of mg/kg, which means high fold dilution is needed to fit in the calibration range. Therefore, to fit the ginsenoside of higher concentration into the calibration range of MS/MS, original extract (QG-I) was diluted to make QG-II (100 times dilution), -III (400 times dilution) and -IV (8000 times dilution) and standard solution of various concentration was added for quantitation with standard addition method.

The matrix matched calibration was performed by mixing 64 µL of the ginseng extract and 16 µL of the working standard solution. Three or four point of calibration was applied as recommended by SANTE/11813/2017 [21]. Highly correlated equations were obtained for each ginsenoside (Table 2; Fig. 3) from white and red Panax ginseng samples to calculate ginsenoside contents in precise way.

Table 2 Concentration of ginsenosides in red and white Panax ginseng
Full size table
Fig. 3
figure 3

Representative standard addition calibration curve (WG; white Panax ginseng and RG; red Panax ginseng)

Full size image

In white Panax ginseng, major ginsenoside of Rb1, Rb2, Rc, Rd, Re, Rf, and Rg1 [22] were observed at relatively higher concentration as expected [5, 23, 24]. The S-isomer was eluted faster than the R-isomer with the higher concentration [6, 20]. Among other minor ginsenoside, Mc and Y were analyzed for the first time in this study, and Mc was quantitated. Rg2(S) was observed at higher concentration compared to other reports. Most of the ginsenoside in white Panax ginseng were observed in red Panax ginseng with higher concentration than white Panax ginseng [5, 23, 25]. Major ginsenoside in red Panax ginseng in this study contained higher amount than that of reported literature [5, 24, 25]. Compound PPT(S) and Y were not detected in both of white and red Panax ginseng.