Anti-Inflammatory Potential of Beesioside O: Target Prediction, Docking Studies, and Molecular Dynamics

Abstract

Triterpenoids with diverse structural features have shown considerable potential as pharmaceutical precursors for anti-inflammatory therapies. Beesioside O (BO), a representative triterpenoid (cycloartane triterpene saponin), has previously been reported to exhibit notable anti-HIV and anticancer activities. However, its anti-inflammatory mechanisms have not been fully elucidated. In this study, we investigated the anti-inflammatory activity and underlying molecular mechanisms of BO in LPS-induced RAW264.7 macrophages. In addition, NP AI Engine predictions, molecular docking, density functional theory (DFT) calculations, and molecular dynamics simulations were conducted to characterize the anti-inflammatory properties of BO further. The experimental results indicated that BO inhibited the mRNA expression levels of iNOS and COX-2. Moreover, it can regulate the phosphorylation of ERK at 3 h. Potential signaling pathways and targets were subsequently analyzed. The structural and electronic properties of BO were calculated using the B3LYP/6-311+G (d,p) basis set. The BO–ERK2 kinase complex was also constructed for simulation. Furthermore, a BO derivative was prepared through hydrolysis followed by acylation, and its anti-inflammatory activity was evaluated. Overall, this study provides deeper insight into the anti-inflammatory effects of BO and supports its potential for further development as an anti-inflammatory agent.

1. Introduction

Inflammation is a crucial component of the body’s defense system, activated in response to harmful stimuli, including infection, tissue injury, or environmental stress. However, prolonged or dysregulated inflammatory signaling can contribute to the onset and progression of numerous chronic conditions, including rheumatoid arthritis, metabolic syndromes, atherosclerosis, and neurodegenerative disorders [1]. These diseases are closely associated with aberrant activation of key inflammatory pathways—particularly NF-κB and MAPKs—and the excessive production of pro-inflammatory mediators and cytokines [2]. Therefore, identifying bioactive natural compounds capable of modulating specific inflammatory pathways has become a promising strategy for developing new anti-inflammatory therapeutics [3,4,5].

Triterpenoids represent a significant class of structurally diverse and bioactive compounds known for their wide range of pharmacological properties, including anti-inflammatory, anti-HIV, and anticancer activities. These properties have garnered substantial research interest in drug discovery over the decades [6,7,8]. Actaea vaginata (Maxim.) J. Compton, a perennial herb native to China and traditionally used to treat stomatitis and pharyngitis, is particularly rich in cycloartane triterpenoids [9]. Our previous phytochemical studies on this species have led to the isolation of several new cycloartane triterpene glycosides [10,11,12,13]. In this study, we investigated the anti-inflammatory activity and mechanisms of beesioside O (BO) (Figure 1). The potential signaling pathway and target prediction were performed using NPAIEngine. Molecular docking was subsequently performed on the online platform. The structural data and electronic properties of BO were calculated using Density Functional Theory (DFT) with the B3LYP/6-311+G (d,p) basis set. The complex of BO and ERK2 kinase was used to perform a molecular dynamics simulation. A BO derivative was also prepared through hydrolysis followed by acylation [14], and its anti-inflammatory activity was assessed.

Figure 1.

The structure of beesioside O (BO) and its derivative.

In natural-product-based drug discovery, molecular docking is widely used as a structure-based screening and hypothesis-building tool to propose how a compound may fit into a protein binding pocket, rank plausible targets or binding modes, and guide subsequent experimental design (e.g., selecting targets for pathway validation, suggesting residues for mutagenesis, or prioritizing analog synthesis) [15]. However, docking and related in silico approaches (including MD simulations and DFT calculations) provide probabilistic models rather than proof of direct binding; therefore, they should also be interpreted together with biological readouts and, when feasible, followed by target-engagement experiments [16].

2. Materials and Methods

2.1. Materials and Reagents

BO (purity > 95%) was isolated from A. vaginata. Methanol, acetonitrile (ACN), and formic acid of HPLC grade were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Ultrapure water used for sample preparation and mobile phase was supplied by Wahaha Corporation (Hangzhou, China). Other organic reagents used in this study were of analytical grade and obtained from Beijing Chemical Works (Beijing, China). TRIzol reagent was purchased from Invitrogen (Carlsbad, CA, USA), and the cDNA synthesis kit was obtained from Takara Biotechnology (Dalian, China). The primary antibodies used in this study were as follows: p38/p-p38, ERK/p-ERK, p-JNK, and GAPDH (Cell Signaling Technology, Danvers, MA, USA).

2.2. General Procedure for the Synthesis of Compound 1a

BO (0.05 mmol) was first converted to its aglycone by enzymatic hydrolysis. Then the aglycone (0.038 mmol) in anhydrous pyridine (1.0 mL) with DMAP (1 equiv) and 2,2-dimethylsuccinate anhydride (10 equiv) was stirred at 150 °C for 2 h in a microwave oven (Biotage, San Jose, CA, USA). The reaction mixture was diluted with EtOAc, neutralized with HCl (1 M), and extracted with EtOAc [14]. The combined organic layers were washed with brine, dried over MgSO₄, and concentrated in vacuum to afford the crude product, followed by semipreparative HPLC eluting (CH₃CN/H₂O, 83:17, 2.0 mL/min) to give 1a (t_R = 26 min, 2.7 mg).

2.3. Purification and Structural Characterization of Compound 1a

Semi-preparative HPLC was performed on a CXTH LC3050N equipped with two pumps of P3000 and a UV3000 detector (Beijing Chuang Xin Tong Heng Science and Technology Co., Ltd., Beijing, China), and a ZORBAX Eclipse XDB-C₁₈ (9.4 × 250 mm, 5 μm, Agilent, Santa Clara, CA, USA) column. The eluent consisted of 58% acetonitrile containing 0.1% formic acid and 42% water. Silica gel (100~200 and 300~400 mesh, Qingdao Marine Chemical Co., Ltd., Qingdao, China) and YMC C₁₈ gel (50 μm, YMC Co., Ltd., Kyoto, Japan) were used for column chromatography. TLC was carried out on GF254 plates (Zhi Fu Huang Wu Pilot Plant of Silica Gel Development, Yantai, China) and visualized by spraying with 10% H₂SO₄ in EtOH followed by heating. NMR experiments were recorded on a Bruker AV III-600 NMR spectrometer (Bruker, Billerica, Germany). HRESIMS spectra were acquired on an LTQ Orbitrap XL spectrometer (Thermo Fisher Scientific, Boston, MA, USA).

2.4. Cell Line and Culture

RAW264.7 macrophage cell line (ATCC TIB-71) was obtained from the American Type Cell Culture Collection (Manassas, VA, USA), and was cultured with RPMI 1640 medium supplemented with 10% FBS and 1% antibiotics in a 37 °C, 5% CO₂ humid cell incubator.

2.5. Cell Viability

Cells (1 × 10⁵/well) were seeded into 96-well culture plates and incubated overnight at 37 °C. Subsequently, the cells were treated with various concentrations of BO (2, 4, 8, 10, 15, 20, and 30 μM) for 24 h. After removing the medium, 100 μL/well of MTT solution (0.5 mg/mL) was added, and the cells were incubated for an additional 4 h at 37 °C. MTT stopping buffer was then added to each well, and the plates were further incubated for 3 h to ensure complete dissolution of the formazan crystals. The absorbance was measured at 570 nm using a microplate reader.

2.6. Nitric Oxide Assay

NO production was assessed in RAW264.7 cells. Cells were seeded in 96-well plates and cultured overnight. The cells were then pretreated with various concentrations of BO (2, 4, 8, 10, 15, 20, and 30 μM) for 24 h. Following BO treatment, cells were stimulated with LPS for 30 min. The culture supernatants were subsequently collected, and NO levels were measured using Griess reagent described previously [17,18].

2.7. RNA Extraction and RT-qPCR

After cell culture and treatment, total RNA from RAW264.7 cells was extracted using TRIzol reagent following the manufacturer’s instructions. The isolated RNA was then reverse-transcribed into cDNA using a cDNA synthesis kit. Both semi-quantitative PCR and real-time quantitative PCR (RT-qPCR) were performed to evaluate the mRNA expression levels of target genes. For semi-quantitative PCR, reaction mixtures and thermal cycling conditions were optimized based on the specifications of the primers and enzymes, and PCR products were analyzed by agarose gel electrophoresis. RT-qPCR was carried out using the TB Green Premix Ex Taq II FAST qPCR kit (Takara Biotechnology, Dalian, China) according to the manufacturer’s instructions. Relative mRNA expression levels were calculated using the 2^−ΔΔCq method, and the housekeeping gene GAPDH was used as an internal reference gene. The primers used in this study were synthesized by Sangon Biotech (Shanghai, China) and are listed in Table 1.

Table 1.

Primers used in this study.

Genes		Sequence (5′-3′)
iNOS	Forward	CATTGATCTCCGTGACAGCC
	Reverse	CATGCTACTGGAGGTGGGTG
COX2	Forward	GGGAGTCTGGAACATTGTGAA
	Reverse	GCACATTGTAAGTAGGTGGACTGT
GAPDH	Forward	CACTCACGGCAAATTCAACGGCACA
	Reverse	GACTCCACGACATACTCAGCAC

2.8. Western Blot

RAW264.7 cells were seeded and pretreated with BO (10 μM). After stimulation with LPS for different time points, cells were collected and lysed in RIPA buffer to extract total protein. Protein concentrations were determined using a BCA Protein Assay Kit (Cwbio, Taizhou, China). Equal amounts of protein (30 μg) were separated by SDS-PAGE and subsequently transferred onto PVDF membranes. The membranes were washed with TBS buffer and blocked with 5% BSA for 2 h at room temperature. After three washes with TBST, the membranes were incubated with primary antibodies overnight at 4 °C. The following day, the membranes were washed three times with TBST and incubated with HRP-conjugated secondary antibodies for 2 h at room temperature. Finally, after three additional washes with TBST, the target protein bands were visualized using an ECL detection system and were relatively quantified as relative intensity (RI) with ImageJ (v1.54k) software.

2.9. Target Prediction and Docking

The online platform NPAIEngine (https://npaiengine.yaozh.com/) accessed on 5 December 2024 was used to predict potential targets and signaling pathways. Then, molecular docking was carried out with a possible target identified and BO. Discovery Studio 2016 and PyMOL 1.8 were used for visualization and 2D interaction analysis.

2.10. DFT Calculations

A three-dimensional model of BO was built in the ChemBio3D Ultra 14.0 program and pre-minimized with the MM2 forcefield (minimum RMS gradient = 0.01). Geometrical optimization and frequency calculations were performed with Gaussian 16 (Rev. C. 01) using DFT at the B3LYP/6-311+G (d,p) level with Grimme’s D3 (BJ) dispersion correction (GD3BJ) [19]. Default convergence thresholds were used. The absence of imaginary frequencies confirmed that the optimized structure corresponds to a true minimum. Molecular surfaces and MEP were rendered with GaussView 6.0.

2.11. Molecular Dynamics Simulation

All MD simulations were run with Amber22/AmberTools22. ERK2 was described with the ff19SB force field, and BO parameters were generated with GAFF via antechamber. System assembly and solvation were carried out in tleap. The solvated complex was minimized for 10,000 steps (steepest descent followed by conjugate gradient). The system was heated gradually to 300 K, equilibrated under NVT conditions, and then simulated for 100 ns, saving coordinates every 1 ps. Trajectories were analyzed with cpptraj, and plots were generated with Xmgrace [20].

3. Results and Discussion

3.1. Isolation and Molecular Characterization of BO

BO was isolated using Semi-preparative HPLC, and the molecular character was elucidated in detail based on NMR and HRESIMS [21]. To clarify its structure, we collected NMR data (Figure 2A–D).

Figure 2.

NMR analysis of BO. (A) ¹³C NMR. (B) HMBC. (C) ¹H NMR. (D) HSQC.

3.2. Characterization of BO Derivative

Colorless needle, ¹H NMR (600 MHz, pyridine-d₅) δ_H 0.25 (1H, d, J = 4.2 Hz, H-19a), 0.43 (1H, d, J = 4.2 Hz, H-19b), 0.60 (1H, q, J = 12.5 Hz, H-6a), 0.95 (1H, m, H-7a), 0.96 (3H, s, H-29), 0.98 (3H, s, H-30), 1.05 (3H, s, H-18), 1.05 (H, m, H-12b), 1.07 (1H, m, H-11a), 1.22 (1H, m, H-1a), 1.25 (1H, m, H-5), 1.27 (1H, m, H-7b), 1.38 (3H, s, H-21), 1.42 (1H, m, H-6b), 1.48 (6H, s, H₃-3′), 1.53 (3H, s, H-28), 1.57 (3H, s, H-27), 1.58 (1H, m, H-1b), 1.63 (H, m, H-12a), 1.63 (1H, m, H-22b), 1.65 (3H, s, H-26), 1.72 (1H, J = 12.5 Hz, 4.2 Hz, H-8), 1.90 (1H, d, J = 6.6 Hz, H-17), 1.95 (1H, m, H-2a), 1.99 (1H, m, H-23b), 2.01 (1H, m, H-22a), 2.06 (1H, m, H-11b), 2.07 (3H, s, 15-COCH₃), 2.36 (1H, m, H-2b), 2.59 (1H, m, H-23a), 2.60 (1H, d, J = 15.6 Hz, H-2′a), 2.95 (1H, d, J = 15.6 Hz, H-2′b), 4.47 (1H, dd, J = 11.4 Hz, 4.2 Hz, H-3), 4.83 (1H, d, J = 7.2 Hz, H-16), 5.67 (1H, br, s, H-15); HRESIMS m/z 681.4028 [M + Na]⁺ (Cacld. for C₃₅H₅₈O₉Na m/z 681.3979).

Rationale for derivatization: we introduced a bulky, lipophilicity-modulating acyl substituent at C-3 after hydrolysis to probe whether altering the polarity and steric environment of the scaffold could improve cellular activity, while maintaining the cycloartane core. The lack of improved activity suggests that this single-point modification is not sufficient to enhance the anti-inflammatory phenotype under our assay conditions, and it provides an initial (negative) SAR data point for future analogue design.

3.3. BO Anti-Inflammatory Activity

To evaluate if BO in concentrations used in the experiments is safe to RAW264.7 cells. Cell viability was measured using the MTT assay. The BO/BO-DMS treatment group (2–30 μmol/L) compared to the control group showed no significant difference in cell viability (Figure 3A,B). The production of NO was a vital inflammatory event that produced a series of pro-inflammatory factors that can lead to cell death [22]. To check the effect of BO on NO release, RAW264.7 cells were stimulated with LPS (1 μg/mL) and treated [23], showing a concentration-dependent decrease in the presence of BO. To examine whether its activity could be further improved, we carried out a primary structural modification with a privileged substituent introduced into the side chain at C-3 of BO. However, the resulting derivative showed no evidence of enhanced anti-inflammatory activity (Figure 3C,D).

Figure 3.

Cell viability and NO production. (A,B) RAW264.7 cells were treated with various concentrations of BO or BO-DMS for 24 h. The cell viability was determined by MTT assay as described in the section on materials and methods. (C,D) Cells were pretreated with various concentrations of BO and BO-DMS (2, 4, 8, 10, 15, 20, and 30 μM) for 24 h. Following BO treatment, cells were stimulated with LPS for 30 min. The NO content was determined by the Griess reagent. The data are presented as means ± SD (n = 3). Letters marked on the histogram indicate the order of values of each data, different letters indicate a significant difference (p < 0.05).

3.4. BO Inhibits LPS-Induced iNOS and COX-2

Macrophages play a pivotal role in cytokine production and immune regulation [24]. They are generally classified into two phenotypes: pro-inflammatory M1 macrophages and anti-inflammatory M2 macrophages [25]. M1 macrophages are characterized by the expression of iNOS and the production of pro-inflammatory mediators. Previous studies have demonstrated that modulation of macrophage polarization can effectively improve the inflammatory microenvironment [26]. The secretion of inflammatory mediators involves complex signaling cascades that trigger the transcription of pro-inflammatory genes [27]. To evaluate the effect of BO on inflammatory gene expression, RAW264.7 cells were treated with LPS to induce upregulation of iNOS and COX-2 mRNA. Compared with the LPS-stimulated group, BO treatment significantly reduced the mRNA expression levels of COX-2 and iNOS (Figure 4), indicating that BO suppresses the transcription of key pro-inflammatory genes in activated macrophages [28].

Figure 4.

Effects of BO on mRNA expression of iNOS and COX2 in RAW264.7 cells. Both semi-quantitative PCR (A) and real-time quantitative PCR (RT-qPCR) were performed to evaluate the mRNA expression levels of target genes (B,C). The data are presented as means ± SD (n = 3). **** indicates a significant difference between the LPS group and BO + LPS groups (p < 0.0001).

3.5. BO Modulates MAPK Pathways in LPS-Induced RAW264.7 Cells

The numerous activators of the extracellular space that stimulated intracellular inflammatory response are regulated by the MAPK signaling pathway, which is highly conserved in eukaryotic cells [29,30]. There are three subfamilies, divided into p38 MAKs, ERK, and JNKs that compose the MAPK family [22]. The recent understandings have shown that the MAPK family is closely related to apoptosis, proliferation, and inflammation [31,32]. When macrophages were stimulated by LPS, the MAPK signaling was proven to mediate the pro-inflammatory cytokine release and nuclear translocation. Researchers have found that activation of the MAPK signaling cascade originates from hierarchical three-tiered phosphorylation and occurs from the cell membrane to cellular organelles, cytoplasm, and nucleus [33,34]. Therefore, inhibiting the phosphorylation of the MAPK pathway is crucial to attenuate the expression of pro-inflammatory and inflammatory cytokines [35,36]. To examine whether BO modulates inflammation through the MAPK pathway in RAW264.7 cells, we measured changes in the phosphorylation levels of ERK, JNK, and p38 following LPS stimulation with or without BO treatment. Our results showed that pretreatment with BO significantly reduced LPS-induced phosphorylation of ERK, suggesting that BO suppresses the ERK/MAPK signaling axis. In contrast, BO (10 μM) did not noticeably affect the phosphorylation levels of JNK or p38 under the same LPS challenge (Figure 5). These findings indicate that BO alleviates the inflammatory response primarily by inhibiting ERK phosphorylation within the MAPK pathway.

Figure 5.

Effects of BO on LPS-induced activation of ERK signaling pathways. Cells were plated at a density of 5 × 10⁶ cells/dish in 60 mm culture dishes and treated with LPS and BO for the indicated time points. After preparation of the total protein, the phosphorylated and total forms of ERK, p38, and JNK were measured by Western blot. RI indicates relative intensity of bands; it was calculated as the relative intensity ratio of each band to GAPDH, and was showed as normalized value with that of control.

3.6. Density Functional Theory (DFT) Calculations

To investigate the electronic properties of BO, DFT calculations were performed. The optimized molecular geometry at the B3LYP/6-311+G (d,p) level of theory was used to generate the Molecular Electrostatic Potential (MEP), as shown in Figure 6. BO exhibited negative electron density near the xylosyl ring, indicating regions of nucleophilic reactivity.

Figure 6.

The optimized structure (A) and corresponding MEP (B) of BO. In panel (A), the red spheres represent oxygen atoms, while gray and white spheres represent carbon and hydrogen atoms, respectively. In panel (B), the MEP is mapped onto the electron density surface, where the color scale indicates the electrostatic potential distribution: red regions correspond to more negative potential (electron-rich sites, prone to nucleophilic interactions and hydrogen-bond acceptors), whereas blue regions indicate more positive potential (electron-deficient sites, prone to electrophilic interactions and hydrogen-bond donors). Green/yellow areas represent intermediate potentials.

3.7. Target Prediction, Molecular Docking, and Dynamics Simulations

The IL-17 signaling pathway was identified to be a potential signaling pathway on the platform NPAIEngine with a p value of 1.049 × 10⁻⁸. Then, ERK2 kinase (PDB: 3SA0) was selected as the target for molecular docking. The grid box was centered at the active region with dimensions adjusted to the length of the tested compound (center: −12.257, 4.564, −6.13; dimensions: 17.0, 22.0, 23.0 in x, y, and z directions, respectively). BO exhibited strong binding energy towards ERK2 kinase with a value of −8.9 kcal/mol, forming one hydrogen bond with LYS114 (Figure 7A–C). Consistent with the western blot data showing reduced ERK phosphorylation, ERK2 was prioritized here as a computational target; nevertheless, ERK phosphorylation is a downstream signaling event and can be modulated indirectly by upstream mechanisms.

Figure 7.

Molecular docking and molecular dynamics (MD) simulation of the ligand bound to ERK2 kinase (PDB ID: 3SA0). (A) 2D interaction map between the ligand and surrounding amino-acid residues. (B) Predicted binding pose of the ligand in the ERK2 active site, highlighting the key contact with Lys114. (C) Overall view of the ERK2–ligand complex (protein shown as cartoon and ligand as sticks). (D) RMSD profiles of the protein and ligand over the 100 ns MD trajectory. (E) Time evolution of the number of hydrogen bonds formed between the ligand and ERK2 during the MD simulation.

Molecular dynamics (MD) studies were conducted to complement the docking simulation results by analyzing the molecular movements of ERK2 upon ligand binding. The docking poses of compound 1 with ERK2 were subjected to a 100 ns MD simulation. Figure 7D,E illustrates that the RMSD of the protein fluctuated between 1.5 and 2.0 Å, while the RMSD of BO fluctuated around 2.0 Å. Overall, the RMSD values showed less fluctuation within the 100 ns timeframe, indicating the reliability of the docking poses.

The predicted pose places BO within the canonical kinase active-site region and suggests a stabilizing hydrogen bond with Lys114, together with multiple noncovalent contacts that can collectively support a feasible binding mode. Because BO is a large, highly functionalized saponin, we interpret the docking score primarily as a qualitative indicator of compatibility between the ligand shape/polar surface and the pocket environment rather than as a quantitative binding-affinity estimate. The relatively stable RMSD profiles over the 100 ns trajectory and the persistence of protein–ligand hydrogen bonding are consistent with a kinetically stable complex on the simulated timescale. At the same time, MD stability does not prove biological target engagement; it only indicates that the docked configuration is not rapidly destabilized in an explicit-solvent model under the chosen force-field parameters.

To substantiate ERK2 as a direct molecular target, future work should include target-engagement and functional assays (e.g., in vitro ERK2 kinase activity inhibition, thermal shift/CETSA, pull-down/affinity methods, or pharmacological competition with known ERK inhibitors). In addition, because RAW264.7 is an immortalized murine macrophage line with well-known LPS hypersensitivity, confirmation in primary macrophages and/or alternative inflammatory models would strengthen the mechanistic interpretation.

4. Conclusions

In summary, BO exhibited anti-inflammatory activity in LPS-stimulated RAW264.7 macrophages, as evidenced by reduced NO production and decreased iNOS and COX-2 mRNA expression. Mechanistically, BO selectively attenuated ERK phosphorylation under the tested conditions, whereas JNK and p38 were not markedly affected. Computational analyses (DFT, target prediction, docking, and MD) provided a coherent, hypothesis-generating model in which BO could adopt a feasible binding mode in ERK2 and form stabilizing noncovalent contacts (including a predicted interaction with Lys114), with the simulated complex remaining stable over 100 ns. The first-pass C-3 derivatization did not improve activity, offering an initial SAR datapoint for future analogue exploration. Overall, the combined experimental and computational results support BO as a promising natural-product scaffold for further optimization; definitive assignment of ERK2 as a direct target will require dedicated target-engagement validation and confirmation in additional inflammatory models (revised to moderate mechanistic claims and add prospects/limitations).

Author Contributions

Q.Q. and Q.-Y.Z.: Methodology, Formal analysis, Investigation, Data curation, Writing—original draft. L.J., Z.H., and Z.-X.Z.: Software, Formal analysis, Investigation. H.-F.W.: Visualization, Conceptualization, Resources, Supervision, Writing—review and editing, Data curation. J.Z.: Project administration, Supervision, Funding acquisition, Writing—review and editing. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors on request.

Conflicts of Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Funding Statement

This work was financially supported by the Beijing Natural Science Foundation [L252129], the CAMS Innovation Fund for Medical Sciences [CIFS 2021-I2M-1-071], Natural Science Foundation of Hunan Province of China [No. 2020JJ4493], the Foundation of Hunan Double First-rate Discipline Construction Projects of Bioengineering [No. YYZW2019-36], Huaihua Key Laboratory of Natural Products Research and Utilization in Wuling Mountain Area [2023R2206], and the State Scholarship Fund from China Scholarship Council.