Metabolomic and Pharmacologic Insights of Aerial and Underground Parts of Glycyrrhiza uralensis Fisch. ex DC. for Maximum Utilization of Medicinal Resources

The roots of Glycyrrhiza spp. have been utilized in Traditional Chinese medicine (TCM) for thousands of years. Non-traditional (aerial) parts constitute a large portion of the biomass of Glycyrrhiza plants and are mostly discarded after harvesting the roots and rhizomes. Through comparative phytochemical and anti-inflammatory activity analyses, this study explored the potential benefits of the aerial parts of Glycyrrhiza uralensis Fisch. ex DC. as medicinal materials. First, a combined approach based on GC/MS and UHPLC-ESI-QTof MS analysis was adopted for the identification and quantitative examination of medicinally important compounds from G. uralensis. Additionally, a bioassay-guided fractioning of ethanolic extracts of G. uralensis leaf material was performed and its anti-inflammatory activity was tested. The aerial portion of G. uralensis was rich in medicinally important compounds. Two compounds (henicosane-1 and decahydroisoquinoline-2) were found to exert a significant anti-inflammatory effect, inhibiting the release of pro-inflammatory mediators (NO and PGE2) and cytokines (IL-1β, IL6, and TNF-α), without exerting cytotoxic effects. Moreover, both compounds down-regulated iNOS and COX-2 mRNA expression. These results suggest that non-traditional parts of G. uralensis are suitable sources of bioactive metabolites that can be explored for medicinal purposes.


INTRODUCTION
Glycyrrhiza uralensis Fisch. ex DC. as well as Glycyrrhiza glabra L. (Fabaceae), commonly known as licorice, are traditional plants recognized through ages for their multiple health benefits and medicinal uses. Particularly, G. uralensis is mentioned in the pharmacopoeia of China, Russia, and other countries (Wang et al., 2020). "Licorice" is obtained from the underground parts of G. uralensis and related species. Different compounds including triterpenoid saponins, flavanones, chalcones, and coumarins have been isolated from the roots of Glycyrrhiza spp. (Wang et al., 2020). Glycyrrhiza spp. are widely cultivated, since these contain most of the bioactive compounds that are responsible for their medicinal and culinary attributes as a flavoring agent and spice (Bell et al., 2011;Dong et al., 2014). Currently licorice is used at different stages of processing grains and oil products, meat products, beverages, candies, jellies, dried fruits, seeds, and soy sauce etc. (Montoro et al., 2011). The roots of this plant are used to treat influenza, coughs, and liver damage in traditional medicinal formulations (Zarubaev et al., 2016).
Previous studies have shown that the extracts of the roots of G. uralensis contain antioxidant, anti-inflammatory, antiviral, cytotoxic, antidiabetic, inhibitors of angiotensinconverting enzyme 2 and transmembrane protease, serine 2, skin-whitening, hepatoprotective, and cholinergic properties (Ahn et al., 2010;Gou et al., 2020;Isbrucker and Burdock 2006;Wu et al., 2020). However, the aerial portion of this plant is of lesser importance to cultivators and usually constitutes an agro-industrial waste after the harvest of the roots or rhizomes that corresponds to merely one fourth of the whole biomass of the plant (Figure 1). It is worth mentioning that the aerial parts of G. uralensis also contains liquiritin and some other medicinally important compounds ( Table 1). While previous studies are mostly limited to some major compounds identified from the roots such as glycyrrhizin, liquiritin, liquiritigenin, and isoliquiritigenin (Ji et al., 2016), the information related to the medicinal importance of many other compounds present in the foliar portion of this plant is scarce. Therefore, in this study we examined the leaves of G. uralensis for the presence of anti-inflammatory compounds. To the best of our knowledge, we described for the first time the presence of two bioactive compounds in the leaves of this plant, and reported the effect of the isolated compounds on the production of important pro-inflammatory mediators.     (Weckwerth et al., 2004). The whole process was performed twice. The material was filtered through a Whatman No. 1 filter paper. The solvent was evaporated using rotary evaporator under low-pressure to obtain a semi-solid consistency.

Extract Derivatization and GC/MS Analysis
Gas chromatography coupled with mass spectrometry (GC/MS) analysis was performed as described by Weckwerth et al. (2004) with some modifications. Here, 13C-ribitol (0.02 μg/μL) was used as an internal standard. Dried samples were derivatized using standard MOX and MSTFA mixtures as described by Mari et al. (2013). The clear supernatant was obtained after centrifugation and poured into clean GC-vials for analysis. The sample (1 µL) was injected in a Shimadzu GC/MS-QP2010 SE (Shimadzu, Japan) instrument at the constant temperature of 230°C in splitless mode. Chromatographic separation was performed using HP-5MS capillary column (30 m × 0.25 mm × 0.25 µm) and helium as carrier gas at a constant flow rate of 1.0 ml/min. The GC/MS temperature gradient used during analysis was same as adopted by Mari et al. (2013). Mass analyzer was set at full scan mode (40-800 m/ z) and the ion source temperature was maintained at 250°C, with EI ionization at 70 eV.

UHPLC ESI-QTof MS analysis
Previously prepared plant extracts were dissolved in pure methanol following sonication for 5 min. The obtained solution was centrifuged, and the supernatant was passed through cellulose filters (0.2 µm pore size). Afterward, the sample (0.2 μL) was injected in an ultra-performance liquid chromatography coupled with electrospray ionizationquadrupole time-of-flight mass spectrometry (UHPLC-ESI-QTof MS/MS) instrument. The chromatographic separation ( Figure 2) was performed on a Waters ACQUITY UHPLC I-class system (Waters Corporation, Dublin, Ireland) fitted with a Waters ACQUITY UHPLC BEH C18 column. The mobile phases were deionized H 2 O containing 0.1% of formic acid (A) and MeOH containing 0.1% of formic acid (B) at constant flow rate of 0.7 ml/min. The gradient was set as follows: 5% B at 0 min, linearly increasing from 5 to 10% B within 5 min, from 10 to 100% B within 22.5 min, and held at 100% B for 2.5 min. The chromatographic system was coupled with Waters Xevo QTof-MS system via an electrospray ionization (ESI) interface operating in full scan mode. The ESI source and MS parameters were set as adopted by (Muema et al., 2017).

Compound Identification and Data Analysis
MzMine version 2.30 (mzmine.github.io) was used for both qualitative and quantitative analysis of both GC/MS and LC/MS data. The alignment was carried out as a function of retention time, using a tolerance window of 0.2 min and 10 ppm mass accuracy (Molina-Calle et al., 2017). Metabolites were identified by comparing mass spectra with spectral libraries (NIST and Wiley), online database MassBank (http://www.massbank.jp/) and previously published literature (Zhang and Ye 2009;Siracusa et al., 2011;Farag et al., 2012;Ammar et al., 2017;Yu et al., 2017;Hefny Gad et al., 2018). Metabolites were identified with a spectral match factor higher than 800. The resulting data matrix based on the relative abundance of metabolites of different parts of G. uralensis was exported into the online tool ClustVis (https:// biit.cs.ut.ee/clustvis/) to create heat maps and principal component analysis (PCA) plots. All samples were analyzed three times and mean data was used to perform statistical analysis. Furthermore, comparative quantifications of different medicinally important compounds were performed for their prevalence in aerial and below ground parts of G. uralensis.

Isolation and Identification of
Anti-inflammatory Compound/s from the Leaves of G. uralensis

Compound Extraction and Isolation
An overview of the purification process of compounds is shown in Supplementary Figure S1. Briefly, dried leaves of G. uralensis were ground into a fine powder prior to use. Leaf powder (∼1 kg) was first extracted with ∼20 L of EtOH at ∼ 77°C for 5 h. Afterward, the solvent was removed by rotary evaporator and lyophilized to obtain a dry material. This dried material was further extracted using an EtOH-H 2 O (v/v) based solvent system and a stepwise elution method with increasing EtOH concentration (from 20 to 100%) to yield five fractions (F1-F5). Subsequently, the selected bioactive sub-fractions, eluted at 60% EtOH (F 4.6) and 70% EtOH (F 4.7), were processed with column chromatography on silica gel to yield  five sub-fractions (A-E). Parent fractions were separated into subfractions based on thin layer chromatography (TLC) to obtain pure compounds. Sub-fraction 4.6°C was passed through a silica gel column to obtain four sub-fractions, 4.6Ca-4.6Cd. Then, subfraction 4.6 Cb was purified on silica gel eluted with EtOH-H 2 O (3:1 v/v) to obtain compound 1 (6.8 mg). Similarly, compound 2 (826.5 mg) was purified using EtOH-H 2 O (4:1 v/v) from subfraction 4.7Bc. Purified compounds were identified by comparing spectral data with previously published data and authentic pure reference compounds.

Cell Line and Cell Culture
The cell line RAW 264.7 was maintained in DMEM medium supplemented with 10% FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin. The cells were grown at 37°C and 5% CO2 in humidified incubator (ABI 371, Thermo Fisher Scientific Inc, United States).

Determination of NO Production
Nitric oxide (NO) production was measured with the Griess reagent (Sigma-Aldrich, Darmstadt, Germany). Briefly, RAW264.7 cells (1 × 10 5 cells/well) were incubated in a 48well plate with different concentrations of test materials for 1 h. Cells were stimulated with lipopolysaccharide (LPS, 1 μg/ml) for 24 h. Supernatants (100 μl) were collected, mixed with an equal volume of the Griess reagent, and incubated at 37°C for 10 min. Afterward, optical density (OD) was measured at 540 nm using a microplate reader (xMark, BIO-RAD, CA, United States). Acetylsalicylic acid (aspirin) (100 µM) was used as positive control in all subsequent assays (Gao et al., 2015). Each experiment was repeated twice, and measurements were taken in triplicate.

Measurement of Pro-Inflammatory Cytokine (PGE2, TNF-α, IL-1β, and IL-6) Production
The production of pro-inflammatory cytokine was determined by commercially available ELISA kits (Bio-Rad, CA, United States; Sigma-Aldrich, Darmstadt, Germany). RAW 264.7 cells (1 × 105 cells/well) were plated in 48-well plates and incubated with test material for 1 h prior to LPS (1 μg/ml) stimulation. Cell-free culture supernatants were collected for the determination of PGE2, IL-1β, IL-6, and TNF-α concentration according to the manufacturer's instructions.

RT-qPCR Analysis
Total RNA from RAW 264.7 cells was extracted using TRIzol reagent (Invitrogen, United States) according to the manufacturer's instructions. First strand complementary DNA (cDNA) was synthesized by using MMLV based reverse transcriptase kit (Invitrogen, United States).  Afterward, cDNA was amplified with gene-specific primers using 2Xn-Taq polymerase mixture (Enzynomics, Korea). The primer sequences are listed in Supplementary Table S1.

Statistical Analysis
Data were analyzed statistically by performing one-way analysis of variance (ANOVA) followed by Duncan's New Multiple Range Test using SPSS version 21 (Chicago, IL).

Comparative Metabolomics of Roots and Aerial Parts of G. uralensis
Roots of G. uralensis are mostly used in Chinese traditional medicine. Since the aerial parts are normally discarded, we wonder if these parts could also be utilized. In an analysis of the biomass distribution, it was found that roots accounted about 30% of the whole plant biomass (Figure 1). Therefore, medicinal evaluation of the aerial parts must be performed. Considering the lack of research dealing with the use of the non-traditional (aerial) parts of G. uralensis for the exploration of medicinally valuable compounds, a preliminary study was performed focusing on the comparative metabolomics of the roots and aerial parts of this plant. The extracts of aerial and below ground parts of G. uralensis were analyzed by GC/MS and UHPLC-ESI-QTof MS/MS analysis (Figure 2). Analysis of the mass spectrum data led to the identification of 79 compounds from the leaves of G. uralensis. The identified compounds can be seen in Tables 1 and 2, along with the main identification parameters obtained from the existing databases. Based on our comparative study, both traditional (roots) and non-traditional (aerial) parts showed a varying profile of different compounds like phenolic, saponins, flavonoids, flavonoid glycosides, coumarins, chalcones, and tannins. To provide a global overview, the relative abundance of the compounds in different parts of the plant is shown in a heatmap (Figure 3). The most abundant group of compounds included phenolic acids and their derivatives, which were identified in both positive and negative ionization mode that generated [M + H] and [M-H] precursor ions ( Table 1). Another group identified in the leaf samples was the medicinally valued flavanones and glycoside compounds, which are characteristic of Glycyrrhiza spp. Other important groups of compounds putatively identified in the leaf samples were saccharides, tannins, and sulfoxides. In addition, lipids and their derivatives were identified in the samples in [M + H] and [M-H] modes. Overall, the results show that the leaves of G. uralensis contain a high diversity of all the examined classes of compounds when compared with the roots and shoots (Figure 2).
The root extract contained higher concentrations of most of the medicinally important compounds ( Table 2). Moreover, these compounds showed different abundance among roots shoots and leaves. For instance, roots contained approximately ten times more flavanones (13.64% of the total mass extract) compared with the leaves (4.07% of the total mass extract). Similarly, compounds like liquiritin, glycyrrhetic acid, feraulic acid, and isoquercitrin were found abundantly in the roots when compared with the leaves and shoot. Alternatively, some compounds like caffeic acid, glyasperin C, isolicoflavonol, and isolicoflavonol were found in the leaves but not in the shoots. Quantitative data showed that the leaves of G. uralensis contained some phenolic acids (sinapic acid and p-coumaric acid), isoviolanthin, and glycyuralin B in higher concentrations in comparison with the roots and shoots. Indeed, phenolic content in 1 g of leaf extract were equivalent to 9.63% of the total mass extract, which is approximately three times lower as compared to the root extracts (2.87% of the total mass extract).
Furthermore, a quantitative data set was created to compare metabolic profile of different plant parts by performing PCA analysis. It showed great extent of variability in the chemical composition of extracts obtained from the roots, shoots, and leaves of G. uralensis. PCA plot showed three distinct groups corresponding to different plant parts (Figure 4).

Effect of G. uralensis Leaf Extracts on LPS-induced NO Production
To determine the anti-inflammatory effects of G. uralensis leaf extracts, we initially investigated the inhibitory effects of crude extracts (at concentration of 25,50,75 and 100 μg/ml) and fractionated leaf extracts (at concentration of 2.5,5.7.5 and 10 μg/ml) against NO production using LPS-induced RAW 264.7 cells. Among the tested fractions, F4 showed the highest inhibitory effect against NO release ( Figure 5).

Identification of the Active Compounds and Their Inhibitory Effect in LPS-Induced NO and PGE2 Production
Firstly, the compounds were putatively identified with MS analysis (Figure 6). Afterward, the identification was confirmed by comparing retention and molecular indices with pure authentic internal standards. Compounds belonging to fraction four, were identified as henicosane (1) and decahydroisoquinoline (2) also known as perhydroisoquinoline.
To assess the inhibitory effect of purified compound 1 and 2 in LPS stimulated RAW 264.7 cells, the production of NO was measured by the Griess reaction and PGE2 by ELISA, respectively. As shown in Figures 7A,B the production of NO and PGE2 was markedly increased by stimulation with LPS. In contrast, treatment with compound 1 and 2 inhibited the production of both mediators in a concentration-dependent manner ( Figures 7A,B). This remarkable effect was not related to the nonspecific cytotoxicity, since both compounds showed non-significant effects on RAW 264.7 cell viability, as determined by the methyltetrazolium (MTT) assay ( Figure 7C). Hence, the inhibition of NO and PGE2 was due to a direct inhibitory effect of the test compounds.

Effect of Purified Compounds in LPS-induced iNOS and COX-2 Expression
Furthermore, the anti-inflammatory effect of the purified compounds was correlated with the expression levels of inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2), as revealed by RT-qPCR analysis. LPS simulation significantly increased the expression of these inflammation related genes (Figure 8), whereas, the presence of compound 1 and 2 significantly attenuated their induction in a concentrationdependent manner (Figure 8).

Effect of Purified Compounds on LPS-Induced Pro-inflammatory Cytokines Production
The inhibitory effect of purified compounds on LPS-simulated RAW 264.7 cell was further analyzed by measuring the changes in the release and the transcription levels of pro-inflammatory cytokines (IL-1β, IL-6, and TNF-α) using ELISA and RT-qPCR analysis, respectively. As can be seen in Figure 9, treatment with compound 1 and 2 lowered the expression levels of all tested LPS-induced pro-inflammatory cytokines at both mRNA and protein levels in a concentration-dependent manner.

DISCUSSION
Glycyrrhiza spp. plays an important role in many prescriptions used in complementary and alternative medicines (Ayeka et al., 2017). These plants are used in traditional Chinese medicine to treat many diseases and act as ingredients in the confectionary industry in Japan. To obtain a chemical profile of the roots and aerial parts of G. uralensis, an analytical method based on GC/MS and UHPLC/MS was developed. Simultaneously acquired UHPLC/MS total ion chromatograms for the roots, shoots, and leaves extracts of G. uralensis are show in Figure 2; while the identities, retention times, and observed molecular and fragment ions for metabolites are presented in Tables 1 and 2.
Metabolites were identified by matching their m/z attributes with those reported in the literature, standard compound libraries (NIST and Wiley), and the "Japan Mass Bank" (Horai et al., 2010). When possible, the identification was confirmed with standard compounds available in-house.
In this study, the overall chemical profile of G. uralensis in terms of the types and contents is in agreement to previous studies (Bai et al., 2020;Kitagawa et al., 1993;Song et al., 2017;Yu et al., 2021). The roots and aerial parts showed the presence of varying abundance of different classes of phytochemicals such as total phenolics, flavonoids, tannins, and saccharides. As shown in Table 2, some of these medicinally valuable compounds were quantified by the normalization of peak areas with authentic internal standards. Remarkably, more than 40 phenolic compounds were identified in the leaves of G. uralensis, suggesting that their abundance is related to the medicinal usefulness of the aerial parts when compared with the roots FIGURE 9 | Inhibitory effect of purified compounds on pro-inflammatory cytokine production in RAW 264.7 cells. Cells (1.0 × 105 cells/ml) were stimulated by LPS (1 μg/ml) for 24 h in the presence of compounds (2.5, 5, and 10 µ/ml). Supernatants were collected, the TNF-α, IL-6 and IL-1β production in the supernatants was determined by ELISA (A-C). Gene expression was analyzed by quantitative RT-PCR analysis using gene specific primers (D-F). Values are the mean ± standard error of triplicate experiments. Values are the mean ± standard error of triplicate experiments. *p < 0.05, **p < 0.01 and ***p < 0.001 for the comparison with the LPSstimulated group.  (Table 1). Some previous studies have also reported presence of same types of medicinally important flavonoids (Fukai et al., 1991;Yuldashev 1998) and phenolics (Nomura et al., 2002) in aerial and belowground parts of G. uralensis. Moreover, O-and C-glycosylated forms were also identified from the aerial parts of G. uralensis. The C-glucosides entities were proposed based on their fragmentation pattern, which involved the sugar moiety by losses of 2, 3, or 4 (HCHO) (De et al., 2012 100%). The concentration of quercetin 3-O-glycoside was higher in the leaves than in the roots, whereas the opposite was seen for kaempferol 7-O-glucoside that was present in higher quantities in the root extracts ( Table 1). The same types of glycosides have been reported in aerial parts of G. uralensis (Jia et al., 1992).
The PCA was performed to highlight the varying metabolic profiles of the aerial and underground parts of G. uralensis. Preprocessed metabolomics data sets from different plant parts were analyzed to generate a PCA plot in which three different groups could be discriminated, thus indicating the varying distribution of components in the leaves, roots, and shoots of G. uralensis (Figure 4). In our study, the integration of data obtained from GC/MS and UHPLC/QTof/MS into a single matrix for PCA allowed the clear separation of extracts from the different plant parts, thereby highlighting the importance of both techniques for sample classification.
In our effort to screen bioactive compounds from the aerial parts of G. uralensis, a bio-guided fractioning allowed the isolation of two phytochemicals (henicosane-1 and decahydroisoquinoline-2) that exhibited significant antiinflammatory effects. Furthermore, the mechanisms mediating this effect were investigated using RAW 264.7 cells. The results revealed that both the compounds significantly reduced the production of NO and PGE2, as well as the mRNA levels of iNOS and COX-2 in LPS-stimulated RAW 264.7 macrophages (Figures 5, 7).
Henicosane belongs to alkanes that are found in various eukaryotic organisms (Coates et al., 2014). These compounds are considered among the main constituents in the most of the plants (Mathis and Ourisson, 1964). Long chain alkanes have been widely isolated from plant fractions possessing medicinal properties (Aiello et al., 2000;Bush and McInerney 2013;Ghasemi Pirbalouti et al., 2014). Methane, a simplest alkane has shown the protective effect to inhibit some inflammatory signals caused by LPS in macrophages and suppress immune response in mice by intensifying IL-10 expression through PI3K/ AKT/GSK-3β pathway (Zhang et al., 2016). Second bioactive compound (decahydroisoquinoline) purified in study is an isoquinoline alkaloid. The phytochemical and biological investigation of different plants have led to the isolation of several isoquinoline alkaloids with medicinal properties (Iranshahy et al., 2014;Khan and Kumar 2015;Haider et al., 2018;Bala et al., 2019). An increasing number of recent studies have reported that alkaloids are effective for treating inflammatory disorders and bring good ground for hope of drug development (Peng et al., 2019).
Macrophages play a key role in the immunopathological phenomena during inflammation and infection owing to their phagocytic and cytotoxic capacities (Laskin and Pendino 1995;Mosser and Edwards 2008;Decano et al., 2016). Proinflammatory mediators (NO and PGE2) and cytokines (IL-1β, IL-6, and TNF-α) are overproduced by macrophages under inflammation (Fujiwara and Kobayashi 2005;Jin et al., 2008). Lipopolysaccharides (LPS)s are the main components of the cell wall of Gram-negative bacteria, which upon recognition by murine macrophages, elicit their activation with a distinctive up-regulation of iNOS expression (Nathan and Xie 1994). In fact, high levels of NO production are of crucial importance in the process of macrophage response (MacMicking et al., 1997). Therefore, the suppression of NO is considered an important therapeutic target to treat inflammation (Batkhuu et al., 2002;Lee et al., 2008;Yoshitake et al., 2008). In our study, we successfully established that the compounds C1 and C2 isolated from the leaves of G. uralensis are strong suppressors of NO production by LPS-stimulated RAW 264.7 murine macrophages. This was accompanied by the inhibition of PGE2 and inflammatory cytokines (IL-1β, IL-6, and TNF-α), as assessed by ELISA and qRT-PCR analysis (Figures 8, 9). As both test compounds showed an effect in all the evaluated downstream targets, including iNOS and COX-2 enzymes, our findings suggest that the anti-inflammatory effect of compound 1 and 2 from G. uralensis, may be due to the targeting of upstream signaling such as that related to the mitogen-activated protein (MAP) kinase or the nuclear factor (NF)-κB signal pathways.

CONCLUSION
Our findings suggest that henicosane 1) and decahydroisoquinoline 2) isolated from the leaves of G. uralensis are valuable anti-inflammatory metabolites. This study supports the pharmacological importance of the nontraditional aerial parts of G. uralensis as potential sources of new natural compounds for the treatment of inflammation. Secondly, these aerial parts, which are currently considered an agro-industrial waste, can be used to recover liquiritin and some other medicinally valuable components.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material.