Screening and characterizing the quality markers of Selaginella tamariscina (P. Beauv.) Spring using metabonomics and molecular networking

2.1 Materials and chemicals

Information on the five Selaginella samples is listed in Table 1. The herbs Selaginella tamariscina, Selaginella doederleinii, Selaginella uncinata, Selaginella delicatula, and Selaginella moellendorffii were collected in 2019 from Hunan Province, China. The plant was identified by Prof. Gui-Shan Tan (Xiangya Hospital, Central South University). The voucher specimens were deposited at the Herbarium Library of Hunan Institute for Drug Control. All plant samples were dried naturally, ground into powders, and then sifted via a 40-mesh sieve to remove particles with a certain diameter. The reference standard selaginellins (purity > 98 %) were isolated from S. tamariscina by our research group. Water was purified by a Milli-Q plus purification system (Millipore, USA). LC-MS-grade acetonitrile and formic acid were purchased from Merck (Darmstadt, Germany). The other reagents were of HPLC grade. The total materials for cultureing cell were purchased from Beijing Solarbio Science & Technology Co., ltd.

2.2 Sample preparation

Five Selaginella samples were extracted according to the methods described in a previous study with minor modifications (K. P. Xu et al., 2011). The herb powder (0.3 g) was refluxed and extracted with 70 % methanol (30 mL) for 1 h. After cooling and compensating for weight loss, the mixtures were centrifuged at 15000 × g for 15 min at 4 °C. Approximately 200 μL of supernatant from each sample was stored for UHPLC-QTOF-MS/MS analysis. A quality control sample was also prepared and mixed with each sample.

2.3 Metabolic profiling of five Selaginella plants by UHPLC-QTOF-MS/MS

The metabolites extracted from five Selaginella samples were analyzed using an Agilent 1290 InfinityⅡ/6540 Q-TOF-MS system (Agilent Technologies, Santa Clara, CA, USA) in electrospray ionization positive (ESI⁺) mode. Chromatographic separation was achieved using an Agilent ZORBAX Eclipse Plus C18 column (2.1 × 100 mm, 1.8 μm) at a flow rate of 0.4 mL/min. The mobile phase system consisted of (A) 0.1 % formic acid in water and (B) acetonitrile. The binary gradient elution program was set as follows: 0 min-10 min, 5–15 % B; 10–20 min, 15 %-35 % B; 30–40 min, 55 %-85 % B; 40–45 min, 85–95 % B; 45–55 min, 95 % B. The injection volume was 2 μL.

The ESI source conditions were set as follows: the sheath and auxiliary gas velocities were set at 10 and 15 arb, respectively; the gas temperature was set at 325℃, with the drying gas flow rate set at 8 L/min, while the sheath gas temp was set at 350℃, with the aux gas flow rate set to 11 L/min. The scan range was from m/z 50 to 1200 at a scan rate of 1.00 spectrum/sec. The collision energy was 15/25/35/45 eV in ESI⁺ mode.

2.4 Screening specific markers using multivariate statistical analyses

For metabolomics, the UHPLC-QTOF-MS raw data were converted to the matrix format of metabolite peak area using MassHunter Profinder (Version 6; Agilent Technologies, Santa Clara, CA, USA). PCA and HCA were employed to visually discriminate between metabolic phenotypes, and orthogonal projections to latent structure discriminant analysis (OPLS-DA) was adopted to screen the key distinguishing markers/markers responsible for the variance in the data, which were then validated through permutation analysis (200 times). The variable importance parameter (VIP) represents the contribution of each metabolite to the OPLS-DA model. Metabolites satisfying both VIP > 1 in the OPLS-DA model and p < 0.05 in Student’s test were selected as candidate distinguishing markers/markers. MassBank (https://massbank.eu/MassBank/) and comparison of reference materials were used for annotation of distinguishing markers/markers.

The structures of differential metabolites were determined by following steps: 1) Use the exact mass to determine elemental compositions of metabolites. 2) Use MS/MS fragments to deduce the possible structural motif. 3) Use molecular networking to search the potential candidates as differential metabolites. 4) Confirm the structures of metabolites by comparing the retention time and MS/MS fragmentation of the sample with authentic standards.

2.5 Molecular networking analysis of distinguishing markers in five Selaginella plants

The raw UHPLC-QTOF-MS data were converted to.mgf format using MS Convert software ( https://www.proteowizard.sourceforge.net) and uploaded to the GNPS online platform ( https://gnps.ucsd.edu/ProteoSAFe/static/gnps-splash.jsp). The GNPS parameters were as follows: precursor ion mass and MS/MS fragment ion tolerance < 0.02 Da, matched peaks > 6, and cosine score > 0.60. The MS cluster size was set to 2 for all sample extracts. Other parameters were set as defaults. The cluster network diagram was downloaded and visualized using Cytoscape 3.8.0.

2.6 Isolation, purification and identification of the Q-markers in S. tamariscina

The top-2 distinguishing markers were isolated, purified and characterized from S. tamariscina extract. The herbs of S. tamariscina (1.0 kg) were extracted five times with 70 % ethanol (5 L) under reflux for 1 h. The 70 % ethanol extract was dried under reduced pressure, redissolved in 80 % methanol, and then centrifuged (2000 g, 5 min) to obtain the supernatant. This supernatant was subjected to polyamide chromatography using a gradient solvent system of ethanol (10%, 30%, 50%, 70%, and 100%), yielding five combined fractions. The 50 % ethanol fraction, Subfraction ST.2–2, was further purified by using a preparative Gilson HPLC system equipped with an RS Tech Optima Pak C18 column (10 × 250 mm, 10 μm particle size) with ACN-H₂O mobile phase containing 0.1% formic acid (0–30 min: 38% ACN, 30–33 min: 38–100% ACN, 33–41 min: 100% ACN, 41–43 min: 100–38% ACN, 43–45 min: 38% ACN). UV detection was performed at 225 nm, and the Q-markers were identified and characterized by molecular networking and NMR.

2.7 Network pharmacology prediction and cytotoxicity assay of the Q-markers of S. tamariscina

The network pharmacology settings were as follows: PharmMapper Server (https://www.lilab-ecust.cn/pharmmapper/) was adopted to forecast the target of the Q-markers. The structural formulas of the Q-markers were saved in mol2 format and then uploaded into PharmMapper Server, and the top-50 targets were collected. All targets were subjected to pathway enrichment analysis (KEGG analysis) by using the String database ( https://cn.string-db.org/) and KEGG online database (https://www.genome.jp/kegg/).

To verify the target, the cytotoxicity of the Q-markers of S. tamariscina was measured. The cell counting kit-8 (CCK-8) approach was applied for in vitro assessment of the cytotoxicity of each tested marker toward two human liver cancer cell lines (SMMC-7721 and HepG₂). Selaginellin/Selaginellin A (1 mM) was dissolved with 1 % dimethylsulfoxide (DMSO) as a stock solution at 20 °C for in vitro experiments. Briefly, cells were cultured at 37 °C under 5 % CO₂ conditions in Dulbecco’s modified Eagle’s medium (DMEM) containing 1 % FBS and 0.1 % penicillin–streptomycin solution. The tested markers dissolved in dimethyl sulfoxide were diluted with DMEM to prepare gradient concentrations (100 µM, 50 µM, 25 µM, 12.5 µM, 6.25 µM, and 3.125 µM). The medium was sterilized using a 0.22µm pore-size filter. Cells were counted, and 4000 cells per well were seeded in a 96-well cell culture plate. Three replicates were made for each measurement. After continuous exposure to the tested compounds at serial concentrations for 48 h, 10 µL of CCK-8 solution was added, and the absorbance was determined at 450 nm by a microtiter plate reader. Cisplatin (purity of 98 % MCE, USA) was used as a positive control. IC50 values were calculated using GraphPad Prism version 8 for Windows (GraphPad Software, San Diego, California USA, https://www.graphpad.com).

2.8 Content determination of Q-markers of S. tamariscina samples using UHPLC-QQQ-MS in multiple reaction monitoring mode

Preparation of mixed standard solutions. Standard substances of marker 1 and marker 2 were weighed precisely and dissolved in methanol to prepare mixed stock solutions of reference standards with concentrations of 0.00123 ∼ 1.234 μg/mL and 0.01960 ∼ 0.9804 μg/mL, respectively.

Chromatographic analyses of Q-markers 1–2 were performed with an Agilent 1290 InfinityⅡ/6540 Q-TOF-MS system (Agilent Technologies, Santa Clara, CA, USA). Chromatographic separation was carried out at 30 °C on an Agilent ZORBAX Eclipse Plus C18 column (2.1 × 100 mm, 1.8 μm). The mobile phase consisted of 0.1% formic acid in water (A) and acetonitrile (B), and gradient elution was performed, introducing 20–60% B at 0–8 min, 60–95% B at 8–11 min, and 95% B at 11–18 min. The sample injection volume was 4 μL, and the flow rate was 0.4 mL/min.

The absolute contents of Q-markers 1–2 were measured on an Agilent 1290 InfinityⅡ/6470 Triple Quadrupole MS system (Agilent Technologies, Santa Clara, CA, USA) equipped with an ESI⁺ source. Ionization of the two markers was performed in ESI⁺ mode. The MS and MS/MS spectra of the two markers were acquired by direct infusion of each standard solution. The multiple reaction monitoring (MRM) transitions were m/z 513.2 → 297.1 for Q-marker 1 and m/z 483.2 → 389.1 for Q-marker 2 (Table 2.). The other parameters were as follows: drying gas: N₂; flow rate: 8 L/min; drying gas temperature: 250 °C; sheath gas temperature: 325 °C; sheath gas flow: 11 L/min; capillary voltage: 3500 V; fragmentor voltage: 195/175 V for Q-markers 1–2; and collision energy: 22/28 V for Q-markers 1–2.

The quantitative determination methodology was used to determine metrics including linearity, accuracy, precision and repeatability. The calibration curves were established by plotting the peak area (Y) against the concentrations (X) of the two marker standards. Based on signal-to-noise ratios of approximately 3 and 10, the limit of detection (LOD) and limit of quantification (LOQ) for each sample were determined, respectively. The accuracy was assessed by analysis of the recovery. The recovery study was performed by adding known concentrations of standards into certain amounts of S. tamariscina samples to achieve different concentrations of 80 %, 100 %, and 120 % (n = 9). The precision was determined by intraday and interday variations. The mixed standards were detected (n = 6) within one day for evaluation of the intraday precision. The method repeatability was determined by analyzing S. tamariscina samples for six replicates.

3.1 Using PCA and HCA to discriminate between the metabolic phenotypes of five Selaginella plants and disguise S. tamariscina from its four adulterants

Chemometrics is an efficient method for exploring potential Q-markers of TCM (Yang et al., 2017). Principal component analysis (PCA) is an unsupervised statistical method used in chemometrics, according to dimension reduction without prior knowledge of group membership, which is useful to separate systematic variation from noise and is applied to perform pattern recognition and phenotypic differentiation of TCM (Sun et al., 2021). As displayed in Figure 1A, a PCA scatter plot was used to determine whether the chemical compound differences of the five selaginella plants were sufficiently differentiated and to screen metabolic markers responsible for the various species. The results showed that all the samples were clearly separated by species, suggesting abundant chemical diversity among species. In addition, hierarchical clustering analysis of metabolic profiling also showed that the five Selaginella species can be divided into five branched groups corresponding to their genetic or species diversity. Chemical similarity for clustering was assessed using Euclidean distance and the complete linkage method. Shorter horizontal lines in the dendrogram indicate higher similarity between samples. The S. tamariscina (JB) samples were similarly clustered with S. doederleinii (SL) samples in one minor branch, which suggested that S. tamariscina and S. doederleinii were chemically similar. The S. delicatula samples and S. uncinata (CYC) samples were aggregated into another branch. The S. nmoellendorffii (JN) samples were from another relatively independent branch, and their metabolites were significantly different from those of the other four Selaginella plants (Fig. 1B). The above results suggested that the PCA and HCA methods could differentiate between these five Selaginella plants.

Close

Fig. 1

Metabolic phenotype differentiation between S. tamariscina and four non-S. tamariscina species marketed as counterfeit TCM. (A) PCA model that distinguished S. tamariscina (JB) from four counterfeit herbs and (B) HCA model that distinguished the metabolic profile of S. tamariscina (JB) and four counterfeit herbs. SL refers to S. doederleinii; BY refers to S. delicatula; CYC refers to S. uncinate; and JN refers to S. moellendorffii.

Export to PPT

3.2 Using OPLS-DA to distinguish between markers of S. tamariscina and four species marketed as counterfeit TCM

The orthogonal partial least squares discriminant analysis (OPLS-DA) model is a powerful statistical modeling tool (Huang et al., 2018). Improved discrimination of five Selaginella species by chemical markers can be achieved by fully utilizing class information in the OPLS-DA model, which maximizes covariance between measured data and response variables. As shown by the OPLS-DA scatter plot in Fig. 2, the Selaginella samples were clearly separated according to their species, and key differential metabolites contributing to the group separation included multiple selaginellins. The metabolite samples of S. tamariscina and S. doederleinii (SL) were the most similar; therefore, the OPLS-DA model between the S. tamariscina sample and SL sample was selected as a representative example for analysis. These differential metabolites belonged to selaginellins, which were only detected in S. tamariscina samples, not in other SL samples. According to the VIP rank, the top distinguishing markers, marker 1 (m/z = 513.16888) and marker 2 (m/z = 483.15842), had the greatest contribution to group discrimination and were selected as Q-markers (Fig. 2). Similarly, targeted chromatographic analysis revealed that selaginellins were not detected in JN/SL/BY/CYC samples (Fig. 3). These two metabolites were targeted, isolated and purified from S. tamariscina extracts and were further identified as selaginellin and selaginellin A by NMR (Section 3.3). Based on these analyses, OPLS-DA revealed that selaginellin and selaginellin A may be specific markers in S. tamariscina, and both could be used to distinguish S. tamariscina from the other four adulterants.

Close

Fig. 2

Screening the distinguishing markers between S. tamariscina (JB) and four non-S. tamariscina species marketed as counterfeit TCM. (A) OPLS-DA model scatter diagram. (B) Permutation test performed 200 times for the validation of OPLS-DA. (C) Scatter plot from the OPLS-DA model used to screen differential metabolites with p(corr) > 0.5. (D) Scatter plot from the OPLS-DA model used to screen distinguishing markers (DM) with VIP > 0.5. (E) The top-1 distinguishing marker (DM1). (F) The top-2 distinguishing marker (DM2).

Export to PPT

Close

Fig. 3

The top-2 distinguishing markers were specific components, only distributed in the S. tamariscina sample.

Export to PPT

3.3 Structural characterization of the top-2 distinguishing markers using UHPLC-QTOF-MS/MS-based molecular networking and NMR

Molecular networking is a tandem mass spectrometry (MS/MS) data organizational approach that can efficiently cluster the analysis of natural products or metabolites based on MS/MS fragment patterns and greatly assist in metabolite annotation in metabolomics or TCM analysis (Perez De Souza, Alseekh, Brotman, & Fernie, 2020). In this study, molecular network visual analysis detected numerous nodes corresponding to distinguishing markers. Each node in the molecular network represents a distinguishing marker. The cosine score was a key index for evaluating the similarity between MS/MS fragments and multiple differential metabolites (Fig. 4).

Close

Fig. 4

Molecular network of unique markers responsible for distinguishing S. tamariscina from four Selaginella species used as counterfeit TCM. Green nodes denote the reference standard, and red nodes denote the extract.

Export to PPT

The molecular node with a precursor ion at m/z = 513.16888 corresponded to distinguishing marker 1. Distinguishing marker 1 showed its protonated molecular ion [M + H]⁺ at m/z 513.16888 in ESI⁺ mode, matching the molecular formula of C₃₄H₂₅O₅ ⁺. The MS/MS spectra showed that the precursor ion at m/z 513.16888 generated a product ion at m/z 495.15869 due to the loss of H₂O. The fragment ion at m/z 495.15869 then generated fragment ions [M + H-C₆H₆O]⁺ at m/z 419.12714 and [M + H-H₂O-C₆H₆O]⁺ at m/z 401.11691, owing to benzyl cleavage. The fragment ion at m/z 419.12714 formed the fragment ion [M + H-C₆H₆O-CH₂O]⁺ at m/z 389.11636 because of cleavage of the primary alcohol. The fragment ion at m/z 389.11636 produced fragment ions [M + H-C₆H₆O-CH₂O-C₆H₄O]⁺ at m/z 297.09064 and [M + H-C₆H₆O-CH₂O-C₈H₄O-H₂O]⁺ at m/z 255.07990. Distinguishing marker 1 was isolated, purified from S. tamariscina and identified by NMR (Fig. 5). ¹H-NMR (500 MHz, DMSO‑d₆) δH: 7.69 (1H, d, J = 6.5 Hz, H-16), 7.47 (1H, t, J = 10.0, 2.0 Hz, H-3), 7.36 (1H, d, J = 6.5 Hz, H-17), 7.18 (1H, t, J = 10.0, 2.0 Hz, H-5), 6.98 (2H,d, J = 8.4 Hz, H-28, 32), 6.80 (2H, d, J = 8.4 Hz, H-20, 24), 6.76 (2H, d, J = 7.5 Hz, H-8, 12), 6.64 (2H, d, J = 8.4 Hz, H-29, 32), 6.62 (2H, d, J = 7.5 Hz, H-9, 11), 6.55 (2H, d, J = 8.4, H-21, 24), 6.37 (1H, d, J = 10.0, H-2), 6.30 (1H, d, J = 10.0 Hz, H-6) and 4.80 (2H, t, H-34); ¹³C-NMR (125 MHz, DMSO‑d₆) δC: 185.6 (C-1), 159.1 (C-7), 158.1 (C-30), 156.4 (C-22), 142.4 (C-13, 15), 140.2 (C-19), 140.4 (C-18), 140.2 (C-19), 130.6 (C-25), 133.0 (C-3), 132.7 (C-28, 32), 129.7 (C-8, 12), 129.5 (C-4, 17, 20, 24), 127.0 (C-16), 121.0 (C-14), 115.6 (C-2, 29, 31), 115.0 (C- 9, 11), 114.7 (C-20, 23), 112.2 (C-33), 98.6 (C-27) and 83.7 (C-26). The NMR signals of distinguishing marker 2 were consistent with selaginellin A in the literature (Zhang et al., 2012). The chromatographic behaviors and mass spectrometric fragment pattern of marker 1 were in accordance with selaginellin. Therefore, distinguishing marker 1 was identified as selaginellin.

Close

Fig. 5

Putative fragmentation pattern of the top-1 distinguishing marker (DM1) (selaginellin). (A) MS/MS spectrum of differential metabolite 1. (B) Possible fragmentation pattern of selaginellin.

Export to PPT

The node with a precursor ion at m/z 483.15842 corresponded to distinguishing marker 2. Distinguishing marker 2 displayed a precursor ion of [M + H]⁺ at m/z 483.15842 in positive mode and precisely matched the molecular formula of C₃₄H₂₄O₃⁺. It generated a fragment ion [M + H-C₆H₆O]⁺ at m/z 389.11664, owing to benzyl cleavage. The fragment ion at m/z 271.07536 resulted from α-cleavage of the peak ion at m/z 389.11664. Fragment ions at m/z 371.10614, m/z 361.12167, and m/z 343.11191 were also generated from the ion at m/z 389.11664, owing to the neutral loss of H₂O, CO and CO + H₂O, respectively (Fig. 6). Moreover, ¹H-NMR (DMSO‑d₆, 500 M Hz) δ:9.91 (2H, brs, –OH), 9.43 (1H, s, –OH), 7.63 (1H, d, J = 7.3 Hz, 15-H), 7.58 (1H, t, J = 7.6 Hz, 16-H), 7.49 (1H, d, J = 7.5 Hz, 6-H), 7.35 (1H, d, J = 7.2 Hz, 17-H), 7.20 (1H, d, J = 7.4 Hz, 2-H), 6.96 (2H, d, J = 8.5 Hz, 28-H, 32-H), 6.81 (2H, d, J = 8.4 Hz, 20-H, 24- H), 6.75 (2H, d, J = 8.4 Hz, 8-H, 12-H), 6.66 (2H, d, J = 8.4 Hz, 29-H, 31-H), 6.64 (2H, d, J = 8.4 Hz, 9-H, 11-H), 6.56 (2H, d, J = 8.4 Hz, 21-H, 23-H), 6.38 (1H, d, J = 10.2 Hz, 5-H), 6.32 (1H, d, J = 10.2, 3-H) ¹³C-NMR (DMSO‑d₆, 125 M Hz) δ:183.5 (C-1), 159.3 (C-10), 158.7 (C-7), 158.1 (C-30), 156.6 (C-22), 142.4 (C-18), 140.4 (C-19), 139.6 (C-2), 138.5 (C-6), 130.5 (C-4), 132.8 (C-28), 132.8 (C-12), 132.8 (C-32), 130.5 (C-8), 130.5 (C-25), 130.0 (C-15), 129.9 (C-20), 129.5 (C-17), 129.5 (C-16), 129.4 (C-13, C-24), 127.9 (C-3), 127.9 (C-5), 124.0 (C-14), 115.5 (C-31), 115.1 (C-9), 115.1 (C-11), 115.1 (C-29), 114.7 (C-21), 114.7 (C-23), 112.1 (C-33), 93.6 (C-27) and 86.5 (C-26). The NMR signals of distinguishing marker 2 were consistent with selaginellin A in the literature (Zhang et al., 2012). The chromatographic behavior and mass spectrometric fragment mode of distinguishing marker 2 were in line with selaginellin A. Therefore, distinguishing marker 2 was identified as selaginellin A.

Close

Fig. 6

Putative fragmentation pattern of the top-2 distinguishing marker (DM1) (selaginellin A). (A) MS/MS spectrum of differential metabolite 2. (B) Possible fragmentation pattern of selaginellin A.

Export to PPT

Amentoflavone was distributed in both authentic S. tamariscina herbs and counterfeit herbs. Therefore, amentoflavone could not be applied as a Q-marker of S. tamariscina due to its lack of specificity. In contrast to amentoflavone, selaginellin A and selaginellin were also exclusively distributed in authentic S. tamariscina, clearly differentiating S. tamariscina herbs from adulterants. Therefore, selaginellin and selaginellin A were more suitable as Q-markers of S. tamariscina.

Another remaining distinguishing marker, which was clustered due to the similar structures and MS/MS fragments, also belonged to selaginellins and selaginellin analogs, including selaginellin, selaginellin A, selaginellin B and selaginpulvilin B. Most distinguishing markers in molecular networking showed similar fragment patterns in the tandem mass spectrum and molecular networking results, for example, neutral losses such as H₂O, CO and phenol (C₆H₆O), and their characteristic fragment ions were [M- C₆H₆O]⁺ (Fig. 4).

Selaginellins are a novel class of active natural products primarily found in S. tamariscina. The unique polyphenolic skeletons in selaginellins, commonly characterized as tautomeric phenol-quinone methide, alkynylphenol, or uorene moieties, endow them with multiple bioactivities, including cytotoxicity, antimicrobial (Cao et al., 2010), hypoglycemic (Cao et al., 2010), and anti-inflammatory activities (Tuan, Thuy, Quan, & Hung, 2019). Selaginellin and selaginellin A were previously exclusively extracted from S. tamariscina(Zhang et al., 2012) and have been proven to exhibit antiapoptotic, antineoplastic, antioxidant, and antisenescence activities (Zhou et al., 2022).

3.4 Network pharmacology and cytotoxicity assessment revealing that selaginellin a and selaginellin were potential quality markers of S. tamariscina

A bioavailability-enhanced network pharmacological approach has been used to select quality markers of TCM (Duan et al., 2020). Network pharmacology predicted that selaginellin and selaginellin A possessed 65 unique targets. By using Cytoscape software, a pharmacology network was established, and the relationships between the quality markers and corresponding targets were shown (Fig. 7A). These targets were closely related to pathways in cancer, endocrine resistance, the IL-17 signaling pathway, progesterone-mediated oocyte maturation, epithelial cell signaling in Helicobacter pylori infection, the relaxin signaling pathway, pathogenic Escherichia coli infection, the estrogen signaling pathway, prostate cancer and the AGE-RAGE signaling pathway in diabetic complications (Fig. 7B). Sixteen targets were correlated with pathways in cancer, including MAPK1, CDK2, IGF1R, EGFR, PPARG, F2, CASP3, NQO1, HSP90AA1, MAPK10, CASP7, PIM1, AR, MAPK8, ESR1 and RXRA.

Close

Fig. 7

Network pharmacology used to predict the potential bioactivity of selaginellin and selaginellin A. (A) Network diagram of selaginellin and selaginellin A as well as their targets. (B) KEGG pathway annotation of the antitumor activity of selaginellin and selaginellin A.

Export to PPT

To verify the potential antitumor activity of the quality markers, the cytotoxicity of selaginellin and selaginellin A was further assessed. Notably, our present study first reported that selaginellin and selaginellin A exhibited moderate cytotoxicity toward the human carcinoma cell line HepG₂, with IC₅₀ values of 20 and 25 μM, respectively. Additionally, selaginellin and selaginellin A showed potent cytotoxicity toward SMMC-7721 human carcinoma cells, with IC₅₀ values of 23 and 27 μM, respectively.

Both specificity and bioactivity are essential for quality markers (Sun et al., 2021). Finally, multiple strategies have been utilized to screen compounds fulfilling the requirements of a Q-marker. Therefore, selaginellin and selaginellin A were proven to be unique and promising quality markers of S. tamariscina.

3.5 Content determination of the quality markers of S. tamariscina

To verify the measurability of the quality markers of S. tamariscina, the contents of selaginellin and selaginellin A in the S. tamariscina sample were determined with UHPLC-QQQ-MS and multiple reaction monitoring. The calibration curve, linear range, LOD and LOQ of the two markers are shown in Table 3. The two calibration curves exhibited good linearity (r > 0.999). The recoveries of selaginellin and selaginellin A were 105.2 %, 104.6 %, 101.5 %, 96.5 %, 99.0 %, 99.5 %, 98.4 %, 98.4 %, and 97.0 % and 92.9 %, 101.2 %, 97.7 %, 94.4 %, 92.1 %, 95.6 %, 92.5 %, and 93.8 %, with RSDs of 3.3 % and 3.3 %, respectively. The intraday variations in selaginellin and selaginellin A were 2.3 % and 1.4 %, respectively (Table 4). The repeatabilities of selaginellin and selaginellin A were 2.0 % and 2.3 %, respectively. These values were within the acceptable range, demonstrating that the method was accurate and precise. The content ranges of selaginellin and selaginellin A were 41.57–44.89 μg/g and 15.09–16.75 μg/g in the S. tamariscina samples, according to the corresponding calibration curves (Table 5), proving the practical measurability of the quality markers. Collectively, based on the aforementioned bioactivity and measurability, selaginellin and selaginellin A were characterized as quality markers of S. tamariscina.