Method development for the determination of seven ginsenosides in three Panax ginseng reference materials via liquid chromatography with tandem mass spectrometry

Abstract

A new liquid chromatography-tandem mass spectrometry (LC–MS/MS) method was developed for the analysis of ginsenosides in three Panax ginseng reference materials (RMs). Extraction procedures were optimized to recover neutral and malonyl-ginsenosides using a methanol–water extraction under basic conditions. Optimized mass fragmentation transitions were obtained for the development of a multiple reaction monitoring (MRM) detection method with electrospray ionization in negative and positive ion mode. Mass fraction values were determined for ginsenosides Rb1, Rb2, Rc, Rd, Re, Rf, and Rg1 in the three ginseng materials (rhizomes, extract, and an oral dosage form). Quantitation of these seven compounds was accomplished with 4-methylestradiol and SRM 3389 Ginsenoside Calibration Solution serving as an internal standard (IS) and calibration standards, respectively. Mass fraction values for the seven ginsenosides ranged from 1.27 mg/g to 21.42 mg/g, 3.25 mg/g to 35.81 mg/g, and 0.56 mg/g to 2.51 mg/g for SRM 3384, SRM 3385, and RM 8664, respectively.

Introduction

Ginsenosides are active compounds found in the Panax family of plants which are often used as medicinal herbs in traditional Chinese medicine as well as an ingredient in dietary supplement commercial products. One of the species of genus Panax, Panax ginseng C.A. Meyer (Asian ginseng), has been used for thousands of years in Eastern medicine as an energizing tonic and remedy for weakness [1-4]. It has been reported that wild Panax ginseng plants have greater pharmacological activity than other species including cultivated plants, which significantly affects the price of material [1]. It is known that American ginseng (Panax quinquefolius L.) contains significantly lower levels of ginsenoside Rf along with varying ratios between other ginsenosides when compared to Asian ginseng [5]. However, physical characteristics are similar between roots of varying species which can make material authentication challenging. Therefore, it is not uncommon for commercially available ginseng products to contain adulterated materials, either inadvertently or economically motivated [6]. Knowledge of the ginsenoside compositions in these herbs is important for phytochemical determination of specific Panax species; indeed, specifications for ginsenoside quantification can be found in pharmacopeial monographs. While the root is the primary source of medicinal preparations from Panax ginseng, other plant parts also contain ginsenosides and other compounds with differing pharmacological activities [7], making the identification and quantification of ginsenosides an important aspect of biomedical research on ginseng health effects. In the dietary supplement industry, accurate identification of specific species based on phytochemical determination is essential for consumer safety, quality assurance, and good manufacturing practices (GMP).

The National Institute of Standards and Technology (NIST) in collaboration with the National Institutes of Health-Office of Dietary Supplements (NIH-ODS) have developed suites of reference materials (RMs) for natural product dietary supplements. RMs, such as those for Ginkgo biloba (SRMs 3246–3248) [8-10] and green tea Camellia sinensis (SRMs 3254-3256) [11-13], include value assignments for targeted, organic, and/or inorganic compounds and are used for standardization of supplement products. RM use promotes experimental rigor and supports manufacturing quality control efforts [14, 15]. With greater than 50% of adults in the USA reporting supplement use [10, 16], reference materials are essential for ingredient authentication, accurate compositional analysis, and for ensuring consumer confidence and safety. NIST has developed SRM 3389 Ginsenoside Calibration Solution [17], which serves as an authentic standard mixture for the validation of new and current analytical methods focusing on ginseng materials.

Traditional analytical techniques for the determination of ginsenosides include liquid chromatography (LC) coupled with ultraviolet–visible (UV–vis) absorbance spectrometry and/or single quadrupole mass spectrometry detectors [18-20]. However, the isolation of ginsenosides in roots and rhizomes of the plant generally produce complex samples that require more selective detection methods to achieve unbiased analytical results, such as tandem mass spectrometry (MS/MS) [21-23]. The work presented here shows the optimized extraction and LC–MS/MS analysis for the certification of seven ginsenosides in SRM 3384 (Ground Asian Ginseng (Panax ginseng C.A. Meyer) Rhizome) [24], SRM 3385 (Asian Ginseng (Panax ginseng) Extract) [25], and RM 8664 (Ginseng-Containing Solid Oral Dosage Form) [26]. It is important to note that SRM 3385 (extract) was prepared in bulk from the same rhizomes material in SRM 3384 which links both materials together. Reversed-phase liquid chromatography was used to separate ginsenosides Rb1, Rb2, Rc, Rd, Re, Rf, and Rg1 in complex samples, produced by basic extraction of ginseng matrices by sonication. Quantitation of these seven compounds was performed with 4-methylestradiol and SRM 3389 Ginsenoside Calibration Solution serving as an internal standard (IS) and calibrants, respectively. Knowledge of the ginsenoside compositions and mass fraction levels in various parts of medicinal herbs are important for phytochemical determination of specific Panax species and for investigating health effects of dietary supplementation.

Materials and methods

Reagents

Ginsenoside calibration standards Rb1, Rb2, Rb3, Rc, Rd, Re, Rf, Rg1, Rg2, Rg3, Rh1, and Rh2 were obtained from Cerilliant (Round Rock, TX, USA) and previously evaluated for impurities by Phytolab. Structures of the 12 compounds are shown in Fig. S1. SRM 3384 (Ground Asian Ginseng (Panax ginseng C.A. Meyer) Rhizome) and SRM 3385 (Ground Asian Ginseng (Panax ginseng C.A. Meyer) Extract) were obtained through Modern Nutrition and Biotech (Ridgefield, CT). RM 8664 (Ginseng-Containing Solid Oral Dosage Form) was prepared and packaged by High-Purity Standards (Charleston, SC). Internal standard 4-methylestradiol was obtained through Steraloids, Inc. (Newport, RI). LC–MS grade methanol (MeOH), 0.1% formic acid (FA) in acetonitrile (ACN) (v/v), and 0.1% FA in water (H₂O) (v/v) were purchased from Fisher Scientific (Pittsburgh, PA, USA). Glacial acetic acid and potassium hydroxide (KOH) were purchased from Sigma-Aldrich (St. Louis, MO).

Sample preparation

Small portions of each ginsenoside reference standard were massed and dissolved in approximately 1.5 mL accurately massed aliquots of MeOH. A standard mixture was prepared by gravimetrically combining 200 μL of each ginsenoside stock solution for use in LC–MS/MS method development.

Liquid chromatography-tandem mass spectrometry

All analyses were performed on an Agilent 1290 Series LC system equipped with a binary pump, degasser, autosampler, column compartment, and variable wavelength absorbance detector set at 200 nm, with an Agilent 6410B electrospray ionization (ESI) triple quadrupole mass selective detector. All separations were carried out on an octadecyl ACE 3 C₁₈ column (Advanced Chromatography Technologies, Aberdeen, Scotland) with the following parameters: 150 mm length, 4.6 mm inner diameter, and 3 μm average particle size. The separation conditions were as follows: 0.7 mL/min flow rate, 25 °C column temperature, and a two solvent mobile phase system. Mobile phase A consisted of ACN with 0.1% formic acid and mobile phase B consisted of water with 0.1% formic acid. Gradient elution conditions began after 10 min, from 24% A to 58% A over 28 min, followed by a second gradient from 58% A to 100% A over 7 min. The separation concluded with a 15 min wash step at 100% A and a 10 min re-equilibration step at the starting conditions. The sample injection volume was 2 μL. Analyte detection was accomplished using multiple reaction monitoring (MRM) in negative and positive ion mode for ginsenosides and IS, respectively. Retention times and optimized MRM precursor-product ion transitions used for detection are listed in Table 1.

Table 1.

Optimized MRM transition values for the 12 ginsenosides

Compound	Detection window (min)	Retention time (min)a	Predominant mass (m/z)	Transition product mass (m/z)	Fragmentation voltage (V)
Rg1	0	8.77 ± 0.04	845.5	799.5	125
Re		8.97 ± 0.04	991.5	945.6	150
Rf	20	22.26 ± 0.01	845.5	799.5	125
Rb1		23.11 ± 0.01	1153.6	1107.7	170
Rc		23.86 ± 0.01	1123.6	1077.6	170
Rg2		24.36 ± 0.01	829.5	783.5	140
Rh1		24.55 ± 0.01	683.4	637.4	130
Rb2		24.71 ± 0.01	1123.6	1077.6	150
Rb3		24.94 ± 0.01	1123.6	1077.6	160
Rd	27	27.89 ± 0.01	991.5	945.6	150
ISb	34	34.91 ± 0.01	269.2	173.1	87
Rg3		35.92 ± 0.01	829.5	783.5	145
Rh2	40	43.11 ± 0.01	667.4	621.4	130

^aThe uncertainty listed for each retention time is an expanded uncertainty about the mean of the values based on one standard deviation obtained from three measurements

^bInternal standard (IS): 4-methylestradiol detection is in positive ion mode

Extraction

Approximately 30 mg portions of SRM 3384, SRM 3385, and RM 8664 were weighed in separate centrifuge tubes. 500 μL aliquots of 4-methylestradiol in a mixture containing a volume fraction of 60% MeOH and 40% H₂O and 500 μL aliquots of 60% MeOH/40% water containing 0.4 mol/L KOH were mixed with each SRM by end-over-end rotation for 20 min. The extracts were then sonicated for 90 min at 45 °C. After sonication, 100 μL of 1 mol/L acetic acid was added, and the slurries were centrifuged at 3000 rpm for 5 min. SRM extracts were diluted × 0.25 and analyzed via the developed LC–MS/MS MRM method.

Results and discussion

LC–MS/MS optimization

In a previous study [18], Wilson and Sander reported the separation of 12 ginsenosides with the following LC–MS mobile phase gradient: isocratic elution of 22% ACN for 10 min, gradient elution to 58% ACN over 28 min, and followed by gradient elution to 100% ACN over 7 min with a hold for 15 min. As reported by Wilson and Sander, the separation of ginsenosides is highly sensitive to changes in the initial solvent conditions. Separations were studied for initial isocratic conditions that ranged from 22 to 26% ACN (see Fig. 1). The LC–MS/MS system was operated in the selective ion monitoring (SIM) mode with negative ionization. Increasing the initial isocratic conditions for ACN resulted in a decrease in the retention of all ginsenosides, with significant changes in the separation of Rc, Rg2, Rh1, Rb2, and Rb3. At an initial mobile phase composition of 22% ACN, these five ginsenosides are resolved as three partially overlapping peaks. With increased ACN, the separation of these components is significantly improved. However, the opposite trend is observed for the separation of Rg1 and Re and improved separations of these ginsenosides result for decreased percentages of ACN. For this reason, an initial isocratic condition of 24% ACN was chosen as a compromise for optimum separation of the 12 ginsenoside compounds. Average retention times for each analyte are reported in Table 1.

Fig. 1.

Separation of a ginsenoside standard mixture with different elution conditions. Initial isocratic conditions were held constant for 10 min and then followed by a linear gradient to 58% ACN in 19 min. Peak identification: (1) Rg1, (2) Re, (3) Rf, (4) Rb1, (5) Rc, (6) Rg2, (7) Rh1, (8) Rb2, (9) Rb3, (10) Rd, (11) Rg3, (12) Rh2

An MRM method was developed for the detection of the 12 ginsenosides using MS/MS in the negative ion mode. Each standard was individually analyzed by MS/MS to obtain mass values for the fragmentation transitions. Product ion peaks were observed through the formation of an FA adduct ion which is represented by the predominant mass [M − 2H + HCO₂H]⁻. Due to the significantly higher abundance levels in the negative ion mode for the 12 ginsenosides, the fragment transitions resulting from the formic acid adducts were chosen as the optimum values for the MRM method (Table 1).

A key advantage of MS/MS is the ability to isolate coeluting compounds based on secondary fragmentation patterns. For example, Rg1 and Re partially coelute using the optimum LC separation parameters; however, they are fully resolved in the MS domain due to differences in MRM transitions (845.5 → 799.5 m/z and 991.5 → 945.6 m/z, respectively). Therefore, full chromatographic separation of these two ginsenosides is not essential. In the case of Rb2, Rb3, and Rc, baseline resolution is necessary since MRM transitions are identical.

Wilson and Sander [18] chose Rh1 as an IS for SRM 3389 Ginsenoside Calibration Solution certification measurements since this compound was not present in the SRM calibration solution. However, the 12 targeted ginsenosides in this work are present in SRM 3384, SRM 3385, and RM 8664. MacCrehan [19] evaluated several estrogenic compounds for suitability as an IS given the hormone-like backbone of the ginsenosides. In addition to other requirements, a potential IS must be stable throughout the entire extraction process and storage protocols. In the current study, 4-methylestradiol was selected for use as an IS based on similarities in structural and chromatographic properties compared with the targeted ginsenosides. The MRM transitions for the IS were studied in both positive and negative ion mode. In the positive ion mode, an MRM transition was observed with a precursor ion of 269.19 m/z, [M − H₂O+H]⁺, and a product ion of 173.1 m/z. MRM transition optimization in negative mode did not produce detectable product ions. For this reason, MRM detection of 4-methylestradiol in the positive ion mode was selected for quantitation.

Extraction optimization

A previous study explored the extraction of ginsenosides from SRM 3384 [19]. Ginsenoside recovery was optimized for different solvents, extraction times, sonication and microwave-assisted extraction, and number of extraction cycles. Internal standard selection was also investigated. The extraction protocol recommended by White and MacCrehan was used for preliminary extraction studies and is similar to the protocol described in the “Extraction” section of the “Materials and methods” described above; however, they used 0.1 M KOH instead of 0.4 M [19]. Images of the raw materials and corresponding extracts of SRM 3384, SRM 3385, and RM 8664 are shown in Fig. 2.

Fig. 2.

SRM 3384 (Ground Asian Ginseng (Panax ginseng C.A. Meyer) Rhizome) (top), SRM 3385 (Ground Asian Ginseng (Panax ginseng C.A. Meyer) Extract) (middle), and RM 8664 (Ginseng-Containing Solid Oral Dosage Form) (bottom). Corresponding extracts are shown on the left inset image

Separations of extracts of the three SRMs are shown in Fig. 3 with total ion detection, and an LC–MS/MS MRM analysis of the SRM 3384 extract is shown in Fig. 4. Rg1 and Re elute at approximately 9 min and are partially separated. The majority of the constituents elute between 20 and 30 min, and 12 peaks of varying intensities are separated within this interval. Eight constituents were identified with the LC/MS/MS method based on MRM transitions and retention times of respective reference standards: Rf, Rb1, Rc, Rg2, Rb2, Rh1, Rb3, and Rd.

Fig. 3.

LC–MS total ion chromatograms for extracts of candidate SRMs 3384, SRM 3385, and RM 8664

Fig. 4.

LC–MS/MS analysis of seven ginsenosides (Rb1, RB2, Rc, Rd, Rf, Re, and Rg1) quantified in candidate SRM 3384. The internal standard used for the quantitation measurements was 4-methylestradiol. Similar chromatograms were obtained for candidates SRM 3385 and RM 8664 shown in Figs. S2 and S3, respectively

Ginsenosides are effectively extracted in methanol/water mixtures due to their water solubility [27, 28]. Two groups of ginsenosides are distinguished as the “Rb1 family” (Rb1, Rb2, Rc, and Rd) and the “Rg1 family” (Rg1, Rg2, Re, and Rf); these compounds may be esterified to form malonyl-ginsenosides. Extraction protocols are typically developed to include conversion of the malonyl-ginsenosides to their parent compounds for quantitation of total ginsenosides [19]. The addition of a base to the extraction solvent promotes the recovery of the parent compounds via hydrolysis. Previous studies have indicated that KOH is suitable base for this process; however, the optimum concentration of KOH for the extraction of SRM 3384 has not been investigated.

To evaluate the effectiveness of this conversion, signal intensities of the 12 ginsenosides were observed for KOH concentrations that ranged from 0.08 M to 0.8 M. In general, an increase in concentration of KOH increased the extraction levels for the majority of the 12 ginsenosides (data not shown). For ginsenosides Rb1, Rd, and Re, the highest recovery was achieved with 0.4 M KOH and for ginsenosides Rb2, Rc, and Rg1, the highest recovery was observed using 0.8 M KOH. The recovery of ginsenoside Rf was similar at all KOH concentrations. Because only small differences were observed over the range 0.4 M to 0.8 M KOH, 0.4 M KOH is specified to minimize potential hydrolysis of nontargeted compounds.

Certification measurements

Quantitation was based on an averaged response factor model with the use of an internal standard. Seven ginsenosides (Rb1, Rb2, Rc, Rd, Re, Rf, and Rg1) were quantified in the three SRMs since these seven compounds were characterized and assessed for purity in SRM 3389 Ginsenoside Calibration Solution. Isotopically labeled internal standards are ideal; however, there is a lack of availability for various phytochemical labeled internal standards. 4-Methylestradiol was selected for use as an internal standard based on chemical similarity to the measurands, chromatographic properties, and availability. Since ion-suppression is generally accounted for by using an isotopically labeled internal standard, it must be examined when using a single non-co-eluting internal standard. Concentrations and peak areas for each measurand and internal standard in the samples were scrutinized in a series of dilutions to investigate potential anomalies. No ion suppression/enhancement effects were observed. Four calibrants were gravimetrically prepared by combining various proportions of SRM 3389 Ginsenoside Calibration Solution with the IS. Ten extracts of each SRM were analyzed by LC–MS/MS in conjunction with the four calibrant solutions. The LC–MS/MS MRM chromatograms for candidate SRM 3384 extractions are shown in Fig. 4. Separations of SRM 3385 and RM 8664 were nearly identical to SRM 3384 (see Figs. S2 and S3). A summary of the constituent levels determined for each SRM are reported in Table 2 and compared in Fig. 5. Ginsenoside levels determined for the ginseng rhizomes ranged from 1.27 mg/g (Rf) to 21.42 mg/g (Rb1). On average, the ginsenoside levels in the concentrated extract material (SRM 3385) were over 200% greater than those determined for the raw plant material (SRM 3384), and levels in the solid oral dosage material (RM 8664) were 85% lower than the raw plant material. The LC–MS/MS method provided excellent precision for the determination of ginsenosides in plants, extracts, and finished product SRMs with uncertainty estimates (RSD) that ranged from 2.3 to 8.3%. It is worth noting that the National Research Council Canada offers a Certified Reference Material for North American Ginseng Root Extract (GINX-1) [29]. As mentioned previously, American ginseng differs from Asian ginseng and contains a different chemical makeup in terms of ratios among ginsenosides. However, ginsenoside mass fraction values obtained here for SRM 3385 Asian Ginseng (Panax ginseng) Extract are comparable to the respective values detailed in the Certificate of Analysis for GINX-1 [29]. SRM 3385 (extract) is also part of a series of RMs that are interrelated and provides multiple matrices for analytical method evaluations.

Table 2.

Ginsenoside mass fraction values determined by LC–MS/MS in SRM 3384, SRM 3385, and RM 8664

	Mass fraction valuesa,b
Ginsenosides	SRM 3384	SRM 3385	RM 8664
Rb1	22.80 ± 0.47	34.8 ± 2.0	2.64 ± 0.15
Rb2	9.36 ± 0.18	17.9 ± 2.5	0.312 ± 0.016
Rc	7.90 ± 0.15	13.64 ± 0.54	0.334 ± 0.015
Rd	4.700 ± 0.093	13.1 ± 1.9	0.674 ± 0.026
Re	7.18 ± 0.13	17.8 ± 1.6	0.930 ± 0.038
Rf	1.354 ± 0.022	2.94 ± 0.64	0.639 ± 0.023
Rg1	4.372 ± 0.090	6.9 ± 1.2	0.583 ± 0.035

^aLC-MS/MS values quantified using an IS response factor approach

^bValues are expressed as x ± U_95%(x), where x is the mass fraction value and U_95%(x) is the expanded uncertainty of the mass fraction value. The true value of the analyte lies within the interval x ± U_95%(x) with 95% confidence. To propagate this uncertainty, treat the certified value as a normally distributed random variable with mean x and standard deviation U_95%(x)/2

Fig. 5.

Ginsenoside mass fraction values determined in candidate SRMs 3384, SRM 3385, and RM 8664 determined by LC–MS/MS. Error bars represent one standard deviation above and below the mean mass fractions

Conclusion

An LC–MS/MS MRM method was developed for the separation and quantitation of seven ginsenosides—active components in Panax ginseng C.A. Meyer (Asian ginseng). This method provides high selectivity and sensitivity for the analysis of ginsenosides in ginseng rhizomes, extracts, and oral dosage forms. Previously implemented ginseng extraction procedures were further optimized to maximize recovery of ginsenosides in raw plant materials. Levels of seven ginsenosides were determined in SRM 3384, SRM 3385, and RM 8664, for use in quality control for related dietary supplements.

Biographies

Hugh V. Hayes is a Research Chemist at the National Institute of Standards and Technology where he focuses on developing analytical methods and reference materials for botanical and vitamin-based dietary supplementation. He is also involved in coordinating Quality Assurance Programs (QAPs) which support the measurement needs within the dietary supplement communities.

Walter B. Wilson currently coordinates the Cannabis Research Program at the National Institute of Standards and Technology with a focus on developing analytical methods, reference materials, and administration of a Quality Assurance Program (CannaQAP). Currently he is an active member of the D37 Cannabis Committee in ASTM and the Cannabis Analytical Science Program (CASP) of AOAC international to help in the standardization of Cannabis analytical measurements.

Catherine A. Rimmer has a Ph.D. in Analytical Chemistry and is the Group Leader of the Organic Chemical Metrology Group in the Chemical Sciences Division at NIST. Her specific areas of research have focused on understanding liquid phase separations and detection, identification, and quantitation of plant metabolites in dietary supplements. She supports work in the areas of nutrition, food safety, environmental, clinical, and forensic measurements.