The His-tagged PRAS40 protein was purchased from Genscript Co., Ltd (Nanjing, China). The herbs were obtained from Tongrentang Co., Ltd (Beijing, China). For reference, Indole (No. I104724) and palmatine chloride (No. P274975) were acquired from Aladdin Co., Ltd (Shanghai, China). All chemicals were analytically pure unless stated otherwise. MHY1485 (MCE, Monmouth Junction, NJ, United States), an agonist of mTOR. KingFisher magnetic particle processor (Thermo Fisher Scientific) was utilized for affinity selections, and the bioactive metabolites were identified via the UPLC-ESI-MS/MS system (UPLC, ExionLC™ AD).

Weigh the herbs and add distilled water. After soaking for 30 min, boil for 30 min and then filter to obtain EXD. The ratio is shown in Table 1.

Briefly, the PRAS40 protein was incubated with Ni-NTA magnetic beads (10 mg) in PBS (pH 7.4) to coat the proteins for 30 min. Then, the beads were rinsed thrice with PBS + 1 mL Tween-20 (0.05% v/v, incubated for 1 h with EXD in PBS + Tween with mixing, and rinsed again with buffer + Tween 5 times to remove unbound metabolites. The resulting beads carrying bound metabolites were redissolved in elution buffer (10 mM imidazole, pH 8.5) to obtain the bioactive metabolites from EXD. These metabolites were analyzed using the Agilent column SB-C18 (2.1 mm × 100 mm, 1.8 µm) with the mobile phase comprising solvents A (0.1% formic acid + pure water) and B (0.1% formic acid + acetonitrile). For the identification of the bioactive metabolites, a linear gradient program from 95% A, 5% B to 5% A, 95% B within 9 min until a composition of 5% A, 95% B was kept for 1 min. The column was run at the temperature of 40°C, 2 μL injection volume, and the 0.35 mL/min flow rate. The effluent was alternatively connected to an ESI-triple quadrupole-linear ion trap (QTRAP)-MS.

Thermo Scientific UltiMate 3,000 for HPLC quantification. Chromatographic conditions: palmatine and terephthalic acid: mobile phase:0.01 mol/L Potassium dihydrogen phosphate and methanol(1:1), flow rate:1 mL/min, detection wavelength was 270 nm, Sample Volume 10 μL. 4-Nitrophenol: mobile phase: water and methanol (0.35:0.65), flow rate:1 mL/min, detection wavelength was 280 nm, Sample Volume 10 μL. Sinapinaldehyde, Indole and 3-Chloroaniline: water and methanol (0.3:0.7), flow rate: 0.8 mL/min, detection wavelength was 270 nm, Sample Volume 10 μL. The contents of each component were calculated according to the standard curve and the peak area of the sample.

Before the experiment, each sample was placed in a vacuum machine for 10 min to remove gas from the sample. Use the sampling needle that comes with the instrument to draw 300 uL of PRAS40 (5 μM) into the sample pool, and use the titration needle to draw 60 uL of the metabolites (50 μM) as the titrant. The conditions were set as 20 repeated injections at 25°C with an interval of 165 s between each injection. The equilibrium dissociation constant (Kd), stoichiometry (n), enthalpy (ΔH) and entropy (ΔS) were obtained by nanoAnalyzer software. For subsequent analysis, the metabolite with the highest thermodynamic parameter was selected.

The sdf files of ligand small molecules were found in the PubChem database and converted into a PDB file through OpenBabel3.1.1. For receptor structure preparation, the 3D structure of the PRAS40 was downloaded from the UniProt database (Uniprot: Q96B36) and subjected to hydrogenation, water removal, and charge addition through AutoDock Tools 1.5.6. And convert the format of active ingredient and target protein into pdbqt format. Finally, molecular docking was performed through AutoDock Vina 1.2.2.

PC12 cells were provided by Stem Cell Bank, Chinese Academy of Sciences (Shanghai, China). Cells were cultured in Rosewell Park Memorial Institute (RPMI)-1640 medium (Invitrogen, Carlsbad, CA, United States) containing 10% (v/v) fetal bovine serum (FBS, Gibco) and 1% (v/v) penicillin-streptomycin-glutamine in 100 cm² cell culture flask. The flask was placed in a humidity incubator at 37°C with 5% CO2.

Viabilities of PC12 cells were measured using cell counting kit-8 (CCK-8) assay (Beyotime Biotechnology, Shanghai, China) in this research. Briefly, PC12 cells were propagated in 96-well plates and incubated in a humidified incubator at 37°C for 24 h. Then different concentrations of H₂O₂ added into plates to stimulate PC12 cells for 24 h. After that, PC12 cells were treated with CCK-8 (10 μL) solution in fresh RPMI 1640 for 25 min. Each well’s absorbance was assessed via a microplate reader (Infinite M200 PRO; Tecan) at 450 nm.

Total proteins in PC12 cells were isolated using RIPA Lysis Buffer (Beyotime Biotechnology). The protein concentration was determined using a BCA Protein Assay Kit (Beyotime Biotechnology, Shanghai, China) with a full-wavelength functional microplate reader (Infinite M200Pro, Tecan, Switzerland). Then equal concentration Proteins were separated using 12.5% SDS–PAGE and transferred to nitrocellulose membranes. After blocking in 10% nonfat dry milk for 1 h, phosphorylated proteins were blocked using bovine serum albumin (5% BSA, room temperature) for 2 h. The membranes were incubated overnight at 4°C with the following primary antibodies: anti-phospho-mTOR-(Ser2448) (ZRB1553), anti-m-TOR(SAB5700687), anti-LC3 (SAB5701328) (Sigma Aldrich, United States), anti-phospho-PRAS40-(Thr246) (Cell Signaling) and anti-PRAS40(Cell Signaling)antibodies. The membrane was then incubated with the corresponding secondary antibody for 1 h. After three washes with PBST, the membrane was visualized using an ECL Western blot detection kit (Merck, United States). The average optical density of the images was analyzed using ImageJ.

Adult Wistar rats (weight = 250 g, female) were acquired from the SPF Biotechnology Co., Ltd. [license number: SCXK (Jing) 2019-0010] and housed in ventilated cages (2-4/cage) at an environmentally controlled (22°C–24°C) conditions, a 12-h dark cycle, and ad libitum chow and water. The Care and Use of Laboratory Animals and ARRIVE guidelines were employed in a licensed laboratory for all animal investigations (license number: SYXK (Shaan) 2020-007). This investigation was authorized (No. IACUC-20211003) by the Welfare and Ethics Committee of the Laboratory Animal Center of Air Force Military Medical University, China. Rats were randomly divided into 5 groups(n = 8): SCI group; SCI+5 μM palmatine group; SCI+10 μM palmatine group; SCI+20 μM palmatine group; sham group.

Surgical treatment was performed as described previously (Li et al., 2022). Briefly, After anesthesia with 25 mg/kg sodium pentobarbital solution, laminectomy from T8 to T10 was performed through a midline incision in the thoracic spine. A 10 g impactor fell vertically on the surface of the exposed T9 spinal cord from a height of 50 mm, causing a contusion injury. Rats with a tail swing action were included in the experiment, otherwise excluded. In the sham operation group, only the T9 lamina was removed without vertical impact. The palmatine group was injected with palmatine 10 μL (5, 10, and 20 μM concentrations) in the intrathecal space after impact. The SCI group was also injected with an equal volume of saline. After surgery, assisted urination (twice a day) will be given until spontaneous urination is restored. The BBB scores were used to assess the locomotor ability of rats at the time of pre-operation and awakening(0), 1, 3, and 7 days.

The animals were perfused 7 days after SCI. 100 mg/kg of pentobarbital sodium solution was utilized for anesthetizing the rats; then, their spinal cord tissues were sampled, preserved in paraformaldehyde (4%), embedded in wax, sliced, deparaffinized, hydrated, and treated with sodium citrate antigen retrieval solution (1:50, Proteintech, United States). The samples were then blocked for 1 h using 5% sheep serum albumin (Boster Biological, China) in PBS at room temperature before overnight immunostaining with primary anti-LC3 antibodies and secondary antibodies. After mounting (DAPI Fluoromount-G, Southern Biotech, United States), the specimens were imaged via a confocal laser scanning microscope (Olympus BX51) and DP Controller software (Olympus, Japan).

The normality and homogeneity of variances were tested using SPSS (version 25.0). One-way analysis of variance (ANOVA) was utilized to analyze data from multiple groups. To determine differences between groups, multiple comparisons were conducted using Bonferroni post hoc tests. The error bars in all figures indicate the mean ± standard error of the mean (SEM). A significance level of p < 0.05 was considered statistically significant. Data analysis was performed using SPSS and GraphPad Prism (version 6.0c).

The EXD has been widely used for SCI treatment. Considering the pathological mechanism of mTOR, it was hypothesized that EXD has certain metabolites that interact with PRAS40. Therefore, PRAS40 was immobilized onto the surface of beads for screening the bioactive metabolites from EXD. As illustrated in Figure 1, six metabolites were identified according to the increased relative concentration in the samples (Table 2). Among these, 4-nitrophenol and 3-chloroaniline indicated poor drug-like properties due to a lack of compatibility with Lipinski’s rules (Chen et al., 2020). Whereas sinapinaldehyde and terephthalic acid were Pan-Assay Interference Compounds (PAINS), which could indicate positive readouts in biochemical assays via different mechanisms; however, such readouts are not linked with optimizable and processible compounds (Grigalunas et al., 2020). Therefore, palmatine and indole were selected to determine their association constants binding to PRAS40 using the ITC, a robust method for analyzing protein-ligand interactions.

The ITC aims to elucidate the heats of binding that evolve from protein-ligand interactions; however, in principle, titration yields other heats, too (Greytak et al., 2022). Therefore, first, the background heat between the solvent used to dissolve these two metabolites (5% DMSO in H₂O) and the PBS buffer during titration was assessed. After deducting these background heats, the ITC traces and binding isotherms between the ligands and PRAS40 were obtained (Figure 2). The continuous line in the lower plots depicts a good data fit at a single-binding-site model used to acquire fitted parameters ( ∆ G , ∆ H , and ∆ S ) values (Table 3). The resulting K_d values of palmatine and indole binding to PRAS40 could be calculated from the ∆ G values. The K_d values of palmatine and indole were 3.53 × 10⁻⁷ and 2.62 × 10⁻⁶, respectively, indicating palmatine was a promising PRAS40 binding drug candidate for SCI treatment (Table 2). The higher K_a values of palmatine also confirmed the above conclusion.

Molecular docking is an effective approach for studying molecular interactions (Pinzi and Rastelli, 2019). Based on the above screening results, the 3D structure of PRAS40 was imported into AutoDock Tools and docked with palmatine and indole (Figures 3A–C). Molecular docking results showed that the binding energy of the PRAS40 to palmatine and indole was −5.06, −4.24 kcal/mol, indicating that the combination of palmatine and the protein is more stable. Among them, ARG76 of PRAS40 interacts with palmatine to form a conventional hydrogen bond interaction; PRO43, ARG70, ALA66, and HIS69 interact with palmatine to form a carbon-hydrogen bond interaction; CYS44 forms a Pi-Alkyl interaction with palmatine while GLY42, TYR46, THR43 and PRO35 generate van der Waals interactions with palmatine (Figures 3D–E). VAL190 of PRAS40 has a hydrogen bond interaction with indole, and CYS44 and HIS69 have a conventional hydrogen bond interaction. In addition, ALA66, ARG70, PRO43, GLY42, and THR73 of the protein have van der Waals interactions with indole (Figures 3F–G). The results of molecular docking are consistent with ITC. Therefore, the pharmacological activity of palmatine in vitro and in vivo was assessed.

PC12 cells were stimulated with different H₂O₂ concentrations for 24 h to mimic SCI cells, which indicated a dose-dependent (0–300 μM) reduction in cell viability (Figure 4A). Based on this, 200 μM of H₂O₂ was selected for SCI construction (viability = 50%). Subsequently, the cells were treated with palmatine (5, 10, and 20 μM concentrations) for 2 h, which revealed a dose-dependent alleviation of the expression level of p-PRAS40, while the level of LC3 was increased (Figures 4B–E). At the same time, the phosphorylation of mTOR, an important target of PRAS40, is also downregulated as the drug concentration increases. Altogether, these data demonstrated that palmatine inhibited the expression of p-mTOR by binding to PRAS40, activating the autophagic flux of PC12 cells.

Furthermore, the in vivo assessment was carried out by categorizing rats into five groups (n = 8/group). Following SCI, the treatment groups were injected with palmatine (5, 10, and 20 μM concentrations) in the intrathecal space. The remaining two groups were the sham-operated and negative control groups injected with saline. For each group, BBB scoring was performed during the 7 days treatment period. This score is negatively correlated with the severity of SCI. As illustrated in Figure 5A, palmatine at the two higher doses substantially increased the BBB scores than the control group, indicating a significant improvement in SCI rats. Lastly, the tumor tissue was harvested, and LC3 was quantitated by immunohistochemical staining. The treatment group indicated significantly positive expression of LC3, which was dose-dependently increased after palmatine treatment (Figure 5B). These results indicated the palmatine screened from EXD acts by activating the autophagic flux and could be a potential alternative mechanism explaining EXD used for SCI treatment.

To further verify whether palmatine regulates autophagy through the PRAS40/mTOR pathway, PC12 cells were co-incubated with MHY1485 (an agonist of mTOR) and palmatine after treating with H₂O₂. The sham group was PC12 cells treated with saline. The results showed that palmatine treatment significantly promoted autophagy in SCI cells (Figure 6B). However, activation of mTOR by MHY1485 attenuated the effect of palmatine. Phosphorylation of PRAS40 can be inhibited by palmatine, as well as decreased p-mTOR levels and increased LC3 levels in the palmatine group compared to the H2O2 group (Figures 6C, D). However, when mTOR is activated, the expression of LC3, which represents the level of autophagy, decreases, indicating that the promotion effect of palmatine on autophagy was reversed.