Advances in Saponin Diversity of Panax ginseng

Ginsenosides are the major bioactive constituents of Panax ginseng, which have pharmacological effects. Although there are several reviews in regards to ginsenosides, new ginsenosides have been detected continually in recent years. This review updates the ginsenoside list from P. ginseng to 170 by the end of 2019, and aims to highlight the diversity of ginsenosides in multiple dimensions, including chemical structure, tissue spatial distribution, time, and isomeride. Protopanaxadiol, protopanaxatriol and C17 side-chain varied (C17SCV) manners are the major types of ginsenosides, and the constitute of ginsenosides varied significantly among different parts. Only 16 ginsenosides commonly exist in all parts of a ginseng plant. Protopanaxadiol-type ginsenoside is dominant in root, rhizome, leaf, stem, and fruit, whereas malonyl- and C17SCV-type ginsenosides occupy a greater proportion in the flower and flower bud compared with other parts. In respects of isomeride, there are 69 molecular formulas corresponding to 170 ginsenosides, and the median of isomers is 2. This is the first review on diversity of ginsenosides, providing information for reasonable utilization of whole ginseng plant, and the perspective on studying the physiological functions of ginsenoside for the ginseng plant itself is also proposed.


Introduction
Panax ginseng Meyer (P. ginseng), known as the king of all herbs, has been frequently used as traditional medicine and healthy food in China, Korea, and Japan. In 2012, P. ginseng was approved as a new food resource by Chinese government, and it has been widely used as the raw material of healthcare products [1]. Ginseng contains a large amount and number of ginsenosides. More than 289 saponins were reported from eleven different Panax species [2]. In addition, at least 123 ginsenosides have been identified in different P. ginseng species, and these include both naturally occurring compounds and those from steaming and biotransformation [3]. In addition, 112 saponins were reported from raw or processed ginseng, including hydrolysates, semisynthetic, and metabolites [4]. Ginsenosides are known to possess a lot of biological activities including regulatory effects on immunomodulation, protection functions in the central nervous and cardiovascular systems, anti-diabetic, anti-aging, anti-carcinogenic, anti-fatigue, anti-pyretic, anti-stress, boosting physical vitality, and promotion of DNA, RNA, and protein synthesis activities [5][6][7][8][9]. In addition, the biosynthesis of triterpenoid is an important factor of saponin diversity. Consequently, biosynthetic mechanisms The history of ginsenoside isolation can be divided into three periods (before 1980 for Period I, 1980-2000 for Period II, after 2000 for Period III) based on the development of analytical techniques. The study on ginsenoside started in 1854. A ginsenoside-containing constituent was firstly isolated from American ginseng by American scholar Garriques [19], and subsequently, Japanese chemists reported panaquilon, panacon, panaxasapogenol, and ginsenin preliminarily separated from P. ginseng. For almost 100 years since the middle of the nineteenth century, it was difficult to obtain a pure ginsenoside due to the under development of separation techniques. In the early 1950s, with the development of separation technology and the invention of modern analytical instruments, such as GC, TLC, etc., the studies on the chemical ingredient of ginseng made remarkable progress. In 1963, for the first time, Shibata et al. reported the chemical property and structure of the panaxadiol separated from ginseng root [20]. In the 1970s, 17 ginsenosides were detected in ginseng, named as ginsenoside Ro, Ra, Rb1, Rb2, Rc, Rd, Re, Rf, Rg1, Rg2, Rg3, F1, F2, F3, Rb3, Rh, and 20-glucoginsenoside-Rf [21][22][23][24][25][26]. The second period began when the 13 C NMR technique was introduced into the structure analysis of ginsenosides. By comparison of the measured 13 C NMR spectroscopic data with known compounds, the accurate structure of new ginsenosides (G-Rh1, Rh2, Rh3, Rg4, Ra1, Ra2, Ra3, La, Rf2, Rs3, Ia, Ib, etc.) could be resolved from different parts of ginseng (root, steamed root, flower bud, stem, and leaf). In this period, more and more scientists focused on ginsenoside isolation, and most of ginsenosides were found in the aerial parts of ginseng [27][28][29][30][31][32][33][34][35][36]. The third period was defined by high-efficiency separation methods, as methods such as high-speed counter current chromatography (HSCCC), high performance centrifugal partition chromatography (HPCPC), and 2D NMR spectroscopic techniques were used for separating and identifying ginsenosides. The application of these powerful new techniques helps to identify the complex chemical structure, for instance, C17 side-chain variation and malonyl group. More than 50 new ginsenosides were isolated from 2000 to 2019, among which most of those possessed variations in the C17 side-chain, besides a part of malonyl ginsenosides [37][38][39][40][41].

Classification of Saponins Identified from P. ginseng
Although most ginsenosides have a rigid four-trans-ring steroid skeleton, they produce multiple pharmacological and biological effects that are different from one another due to minor variations on: (1) Type of sapogenins; (2) number, type, and site of glycosyl units; and (3) modification of C17 side-chains [11,42,43]. Therefore, the study of ginsenoside structure will help to elucidate the mechanism of multiple functions of ginsenosides. The reported ginsenosides are classified into protopanaxadiol type (PPD), protopanaxatriol type (PPT), oleanolic acid type (OA), and C17 side-chain varied (C17SCV) subtypes according to their determined sapogenin structures ( Figure 1). The glycosyl components of saponin were mainly β-d-glucopyranosyl group, followed by α-l-rhamnopyranosyl group, a few binding α-l-arabinopyranosyl group and β-d-xylopyranosyl group, and the β-d-glucopyranosiduronyl group only appears in saponins with oleanolic acid-type (OA) sapogenin. In dammarane-type triterpenoid saponins, β-d-glucopyranosyl group (2→1)-β-d-glucopyranosyl oligosaccharide chains occur more frequently, and are mostly bound to C-3 of sapogenin to generate oxyglycoside; β-d-glucopyranosyl group (2→1)→α-l-rhamnopyranosyl group oligosaccharide chains are mostly bound to C-6 of sapogenin to form oxyglycoside. The tetracyclic parent nucleuses are relatively stable, whether they are PPT and/or PPD type. Moreover, the substituents that occur in the C17 side-chains often undergo oxidation, reduction, cyclization, and epimerization, contributing to diversity in chemical structure [12,16]. Table 1 displays the molecular formulas, molecular masses, and structural categories of 170 ginsenosides, isolated from different parts of P. ginseng. As a result, four ginsenosides are OA type, 59 ginsenosides are PPD type, 42 ginsenosides are PPT type, and 65 ginsenosides are C17CSV type. Among them, four PPD-type ginsenosides (Rb1, Rb2, Rc, Rd), three PPT-type ginsenosides (Re, Rf, Rg1), and one OA-type ginsenoside Ro (the structures are shown in Figure 2) are the most abundant in P. ginseng, and account for more than 70% of the total saponins [5].
Molecules 2020, 25, x FOR PEER REVIEW  3 of 19 Although most ginsenosides have a rigid four-trans-ring steroid skeleton, they produce multiple pharmacological and biological effects that are different from one another due to minor variations on: (1) Type of sapogenins; (2) number, type, and site of glycosyl units; and (3) modification of C17 sidechains [11,42,43]. Therefore, the study of ginsenoside structure will help to elucidate the mechanism of multiple functions of ginsenosides. The reported ginsenosides are classified into protopanaxadiol type (PPD), protopanaxatriol type (PPT), oleanolic acid type (OA), and C17 side-chain varied (C17SCV) subtypes according to their determined sapogenin structures ( Figure 1). The glycosyl components of saponin were mainly β-D-glucopyranosyl group, followed by α-L-rhamnopyranosyl group, a few binding α-L-arabinopyranosyl group and β-D-xylopyranosyl group, and the β-Dglucopyranosiduronyl group only appears in saponins with oleanolic acid-type (OA) sapogenin. In dammarane-type triterpenoid saponins, β-D-glucopyranosyl group (2→1)-β-D-glucopyranosyl oligosaccharide chains occur more frequently, and are mostly bound to C-3 of sapogenin to generate oxyglycoside; β-D-glucopyranosyl group (2→1)→α-L-rhamnopyranosyl group oligosaccharide chains are mostly bound to C-6 of sapogenin to form oxyglycoside. The tetracyclic parent nucleuses are relatively stable, whether they are PPT and/or PPD type. Moreover, the substituents that occur in the C17 side-chains often undergo oxidation, reduction, cyclization, and epimerization, contributing to diversity in chemical structure [12,16]. Table 1 displays the molecular formulas, molecular masses, and structural categories of 170 ginsenosides, isolated from different parts of P. ginseng. As a result, four ginsenosides are OA type, 59 ginsenosides are PPD type, 42 ginsenosides are PPT type, and 65 ginsenosides are C17CSV type. Among them, four PPD-type ginsenosides (Rb1, Rb2, Rc, Rd), three PPT-type ginsenosides (Re, Rf, Rg1), and one OA-type ginsenoside Ro (the structures are shown in Figure 2) are the most abundant in P. ginseng, and account for more than 70% of the total saponins [5].

Spatial Distribution of Ginsenosides in Different Parts
The Venn diagram (Figure 3) shows the number of ginsenosides commonly and separately shared by the following four groups: R&S (roots, rhizomes, and steamed roots), L&S (leaves and stems), F&P (fruits and fruit pedicels), and F&B (flowers and flower buds). Among them, the number of unique ginsenosides in group R&S, F&P, L&S, and F&B are 52, 15, 14, and 36, respectively, accounting for 30.6%, 8.8%, 8.2%, and 21.2% of the number of total ginsenosides, respectively. The

Spatial Distribution of Ginsenosides in Different Parts
The Venn diagram (Figure 3) shows the number of ginsenosides commonly and separately shared by the following four groups: R&S (roots, rhizomes, and steamed roots), L&S (leaves and stems), F&P (fruits and fruit pedicels), and F&B (flowers and flower buds). Among them, the number of unique ginsenosides in group R&S, F&P, L&S, and F&B are 52, 15, 14, and 36, respectively, accounting for 30.6%, 8.8%, 8.2%, and 21.2% of the number of total ginsenosides, respectively. The result gives some explanation why ginseng root is designated as medicinal parts rather than the other parts. Sixteen ginsenosides are commonly existed in all tissues, and among them, there are nine PPD type (Rc, Rd, Rb2, Rb1, Rb3, m-ginsenoside Rb1, m-ginsenoside Rc, m-ginsenoside Rb2, m-ginsenoside Rd), six PPT type (Re, Rg1, Rf, 20(R)-ginsenoside Rg2, Notoginsenoside R1, m-ginsenoside Re), one OA type (Ro), and none of C17SCV type. Numbers of ginsenosides shared by R&S and F&P, F&P and L&S, L&S and F&B, R&S and F&B were 32 (18.8%), 37 (21.7%), 24 (14.1%), and 19(11.2%), respectively. In addition, 13 malonyl-ginsenosides were existing specifically in flowers and buds; however, none of them was observed in fruit. This implies that these malonyl-ginsenosides show not only spatial specificity, but also temporal specificity. Here in, we speculate that malonyl-ginsenosides may play a physiological role during tissue development.   As indicated by Figure 4, the numbers of PPD-type ginsenosides (blue bar) are highest in R&S, F&P, and L&S, while the C17SCV-type ginsenoside is highest in F&B. Interestingly, C17SCV-type ginsenosides exhibit significant variation among different groups. Only nine C17SCV-type ginsenosides are shared by more than two groups, whereas the other 58 C17SCV-type ginsenosides are unique to a particular group. For the OA-type ginsenoside, three are specific to group R&S (Polyacetyleneginsenoside-Ro, Ginsenoside Ro methyl ester, Calenduloside-B) and one (Ginsenoside Ro) is commonly shared by all parts.

Isomers of Ginsenosides
The total 170 ginsenosides are divided into 69 molecular formula groups. Therefore, it is common that one molecular formula corresponds to several ginsenosides. ( Table 2). The molecular formula with the largest number of isomers is C 48 H 82 O 19 (molecular weight 962.5450), with a total of nine isomers; followed by C 51 H 84 O 21 (molecular weight 1032.5505) with a total of eight isomers, and C 41 H 70 O 13 (molecular weight 770.4816) with a total of seven isomers. The isomers median of 69 molecular formulas is 2, which means that one molecular formula corresponds to two isomers equally. Optical and position isomerism are the dominant types of ginsenoside isomers, whilst cis-trans isomerism and tautomerism are detected occasionally.

Mass Spectrometry-Based Metabolomics Analysis on P. ginseng
Recently, MS and its hyphenations with chromatographic separation techniques have emerged as an instrumental trend in ginsenoside analysis [93,94]. HPLC/MS can overcome the problems related to ginsenoside pre-analysis derivatization and the low abundance of molecular ions [95,96]. The use of on-line MS detection shows superior sensitivity and specificity compared with conventional UV and ELSD detection [97,98]. The sensitivity of MS detection can surpass 1000 times that of UV absorbance [99]. In addition, the possible matrix effects encountered with many Panax ginseng formulations may be compromised by MS [100]. Despite these advantages, MS remains costly for use in routine analysis. With the development of soft ionization techniques, HPLC/MS has been successfully applied for the qualitative and quantitative analyses of Panax ginseng [101]. Among the various mass spectrometry ionization techniques, electrospray mass spectrometry (ESI-MS) is the approach that is most commonly coupled with HPLC [15,102,103]. While ESI-MS suffers from matrix-induced ionization suppression difficulties [104], atmospheric pressure chemical ionization (APCI) can offer itself as one possible alternative [105]. Quadrupole time-of-flight mass spectrometry (QTOF-MS), a powerful tool for the identification of analytes, provides several advantages in structural analysis, such as a higher resolution and accuracy in mass measurements. Coupled with QTOF-MS, UPLC has been introduced for metabolite profiling and metabolomics purposes [99]. In recent years, orbitrap technology has achieved great breakthrough in resolution and scanning speed and realized the high-resolution detection of multi-stage mass spectrometry by combining the linear ion trap and quadrupole mass spectrometry, which can be widely applied in the development of new drugs [106].
According to the available literature, Wang et al. in 1999 [97] firstly identified ginsenosides by LC/MS/MS and differentiated P. ginseng and P. quinquefolius based on the ginsenoside Rg1/Rf and Rc/Rb2 ratios. A liquid chromatography-tandem mass spectrometry (LC/MS/MS) method was developed to distinguish Asian ginseng and North American ginseng. The method is based on the baseline chromatographic separation of two potential chemical markers: Rf and 24(R)-pseudo ginsenoside F11 [107]. Z X. et al. 2000 developed a similar LC/MS/MS method to determine ginsenoside in ginseng. Nine ginsenosides were determined, among which five of them were identified according to molecular weight [108]. In the late 1990s and early 2000s, the resolution of mass spectrometry was low and the number of identified ginsenosides was limited, which could be used for distinguishing Asian ginseng and American Ginseng, and identifying ginsenosides.
Chen et al. [109] established a chemical finger-print metabolomics approach using ultra-high-performance liquid chromatography combined with quadrupole time-of-flight mass spectrometry (UPLC-QTOF/MS). The method was successfully used to authenticate and evaluate Panax Ginseng of various commercial grades. Using UPLC-QTOF-MS/MS, Zhang et al. evaluated the overall quality of commercially available white ginseng and red ginseng, and investigated their characteristic chemical composition indicators. Fifty-one major chromatographic peaks of white ginseng and red ginseng samples were separated within 24 min [110]. By means of UPLC-DAD-QTOF-MS/MS, Wang et al. conducted qualitative and quantitative analysis of ginsenosides of cultivated ginseng and mountain ginseng. A total of 131 ginsenosides were detected in cultivated ginseng and mountain ginseng, and all the components were completely separated within 10 min, among which contents of 19 typical ginsenoside were accurately quantified. This method has been validated for quality evaluation of ginseng and identification of cultivated ginseng and mountain ginseng [13]. Zhang et al. Quickly and comprehensively identified the ginsenosides using high-resolution time-of-flight mass spectrometry, electrospray dual-spray ion source, and negative ion mode. A total of 95 saponins in suncured ginseng were identified within 11 min, providing a feasible basis for the quality control of suncured ginseng [111]. With the emergence of high-resolution mass spectrometry and the development of high-throughput screening technologies, several time-saving methods were established for commercial ginseng product evaluation.
Since 2015, Orbitrap mass spectrometer had been applied in ginsenoside detection. In 2017, a total of 101 malonyl-ginsenosides were firstly systematic analyzed by hybrid LTQ-Orbitrap mass spectrometer after UHPLC separation, and ten potential malonyl-ginsenoside markers were discovered for the discrimination of P. ginseng, P. quinquefolius, and P. notoginseng [112]. Shi et al. established an untargeted profiling strategy on a linear ion-trap/Orbitrap mass spectrometer coupled to ultra-high performance liquid chromatography to analyze malonyl-ginsenosides in several Panax species. Finally, 178 malonyl-ginsenosides were characterized from roots, leaves, and flower buds of P. ginseng, P. quinquefolius, and P. notoginseng [113]. To investigate the variation of ginsenosides among different processed red ginseng, Zhong et al. tested steamed, vinegared and dried red ginseng samples by UPLC-Q-Orbitrap MS. In total, 32 ginsenosides were identified and ginsenosides m-Rb1, Rh1, F1, 20(R)-Rh1, Rg5, and Rs5 were only found in red ginseng processed by vinegar [114]. With the development of Orbitrap and multi-mass spectrometry techniques, ginsenosides with complex structures, such as malonyl and C17 side-chain variation, have been increasingly detected, and the types of ginsenosides have been greatly extended.

Conclusions
In this review, we summarized the existing studies related to saponin analysis of P. ginseng, and sorted out the information of structural characteristic, spatial distribution, and isomer of 170 ginsenosides. There are 16 common ginsenosides present in all parts of P. ginseng. In contrast, each part has unique ginsenosides, and ginsenosides in different parts show obvious structural diversity. It should be emphasized that ginseng aerial parts can regenerate every year, and there is a large amount of rare ginsenosides in stems, leaves, and flower buds. In light of previous research results of the rare ginsenoside bioactivity in red ginseng, it seems that the aerial parts of P. ginseng are highly worth developing and utilizing. A conclusion can also be drawn that C17SCV-type ginsenosides and malonyl-ginsenoside are rich in flowers and buds. Therefore, a hypothesis that ginsenosides have physiological roles in ginseng plant development is proposed. The rapid development of high-performance liquid chromatography and mass spectrometry techniques significantly raise the throughput and accuracy of ginsenoside determination.
In the future, (1) with the continuous advancement of detection and identification technology, the analysis method of ginsenosides will develop in the direction of being more sensitive, convenient, and environmentally-friendly, with high-throughput and high-precision. By leveraging these technologies, more monomer compounds will be separated and identified from ginseng, which will develop the knowledge of the diversity of chemical structure of ginsenosides. (2) It is necessary to conduct further research on spatial distribution of ginsenosides in different parts of ginseng, and multidisciplinary collaborations among genomics, proteomics, metabonomics, and transcriptomics could be used to study the physiological functions of ginsenosides. (3) With increasing separation of ginsenosides possessing a complex structure, such as malonyl and C17 side-chain variation, the pharmacological action and pharmacokinetics of these ginsenosides would be further studied to clarify the efficacy of ginseng.