Efficient Combination of Complex Chromatography, Molecular Docking and Enzyme Kinetics for Exploration of Acetylcholinesterase Inhibitors from Poria cocos

Poria cocos (P. cocos) is a traditional Chinese medicinal product with the same origin as medicine and food. It has diuretic, anti-inflammatory and liver protection properties, and has been widely used in a Chinese medicine in the treatment of Alzheimer’s disease (AD). This study was conducted to explore the activity screening, isolation of acetylcholinesterase inhibitors (AChEIs), and in vitro inhibiting effect of P. cocos. The aim was to develop a new extraction process optimization method based on the Matlab genetic algorithm combined with a traditional orthogonal experiment. Moreover, bio−affinity ultrafiltration combined with molecular docking was used to screen and evaluate the activity of the AChEIs, which were subsequently isolated and purified using high-speed counter−current chromatography (HSCCC) and semi−preparative high-performance liquid chromatography (semi−preparative HPLC). The change in acetylcholinesterase (AChE) activity was tested using an enzymatic reaction kinetics experiment to reflect the inhibitory effect of active compounds on AChE and explore its mechanism of action. Five potential AChEIs were screened via bio−affinity ultrafiltration. Molecular docking results showed that they had good binding affinity for the active site of AChE. Meanwhile, the five active compounds had reversible inhibitory effects on AChE: Polyporenic acid C and Tumulosic acid were non-competitive inhibitors; 3−Epidehydrotumulosic acid was a mixed inhibitor; and Pachymic acid and Dehydrotrametenolic acid were competitive inhibitors. This study provided a basis for the comprehensive utilization of P. cocos and drug development for the treatment of AD.


Introduction
Poria cocos (P. cocos) is a well-known traditional East Asian medicinal plant that grows around the roots of pine trees in China, Japan, Korea, and North America [1,2]. Through research, summary, and analysis of the current status of traditional Chinese medicine in the treatment of Alzheimer's disease (AD), Gong Jing found that the use of P. cocos ranks among the top 10 most commonly used unilateral drugs [3]. Hu Pan discussed the ancient medication rules for AD based on the traditional Chinese medicine inheritance computing platform. Of these, the core drug combination "Poria cocos-Licorice" has a good effect on the treatment of AD [4]. Zhou Siduo used an improved micro-screening model based on the Ellman method to screen 49 kinds of traditional Chinese medicine extracts for acetylcholinesterase (AChE) activity and calculate the inhibition rate. The results showed that P. cocos had a better ability to inhibit AChE [5].

Optimization of Extraction of P. cocos Reflux
The ethanol reflux extraction process of P. cocos was optimized using the Matlab genetic algorithm combined with an orthogonal experiment. The quadratic regression model of total triterpene content was used as the objective function, and the optimal extraction conditions were searched by the genetic algorithm. The initial population size was 50, the single point mutation probability was 0.01, the single point crossover probability was 0.85, the evolution was 100 generations, and the results were searched randomly 10 times. The optimal extraction conditions were 19.87 times the ethanol dosage, 1.06 h of extraction time, 1.09 h extraction time, and 80% ethanol concentration. Under these conditions, the theoretical value of Pachymic acid content reached 0.83%. For operability of the experimental conditions, the extraction conditions were further modified to 20 times the ethanol dosage, 1.0 h extraction time, and 80% ethanol concentration. Under the optimized conditions, the active components were extracted from P. cocos. The eluent was condensed at 45 °C by rotary vaporization to a dry state, and the concentrate was evaporated and transformed into a powder for subsequent use.

Screening of Potential AChEIs in P. cocos by UF-LC-MS
According to the chromatographic results of the ligand (potential AChEIs) in P. cocos, in the range of 25-55 min, 12 components were eluted from P. cocos extract, five of which exhibited an affinity to the AChE receptor. The results were confirmed using AChE at varying levels of activity. As illustrated in Figure 2, the retention time of compound 1 was 36.144 min, and the inhibition rates of 0.5, 1.0, and 2.0 U/mL AChE were 24.10, 24.34, and 24.22%, respectively; For compound 2, the retention time was 39.562 min, and the inhibition rates at 0.5, 1.0, and 2.0 U/mL AChE were 36.59, 37.46, and 37.32%, respectively; For compound 3, the retention time was 41.032 min, and the inhibition rates at 0.5, 1.0, and 2.0 U/mL AChE were 13.12, 13.88, and 13.78%, respectively; For compound 4, the retention time was 46.066 min, and the inhibition rates at 0.5, 1.0, and 2.0 U/mL AChE were 38.32, 38.60, and 38.52%, respectively; For compound 5, the retention time was 55.058 min, and the inhibition rates at 0.5, 1.0, and 2.0 U/mL and AChE were 11.87, 12.03, and 11.90%, respectively. The mixture of P. cocos extract with 1.0 U/mL AChE exhibited the highest binding degree. The binding strength of the compound and AChE was determined as expressed in Equation (1). The order of combined strength capabilities of the active ingredient and AChE was 4 (38.60%) > 2 (37.46%) > 1 (24.34%) > 3 (13.88%) > 5 (12.03%). The five active ingredients that were identified by mass spectrom-

Optimization of Extraction of P. cocos Reflux
The ethanol reflux extraction process of P. cocos was optimized using the Matlab genetic algorithm combined with an orthogonal experiment. The quadratic regression model of total triterpene content was used as the objective function, and the optimal extraction conditions were searched by the genetic algorithm. The initial population size was 50, the single point mutation probability was 0.01, the single point crossover probability was 0.85, the evolution was 100 generations, and the results were searched randomly 10 times. The optimal extraction conditions were 19.87 times the ethanol dosage, 1.06 h of extraction time, 1.09 h extraction time, and 80% ethanol concentration. Under these conditions, the theoretical value of Pachymic acid content reached 0.83%. For operability of the experimental conditions, the extraction conditions were further modified to 20 times the ethanol dosage, 1.0 h extraction time, and 80% ethanol concentration. Under the optimized conditions, the active components were extracted from P. cocos. The eluent was condensed at 45 • C by rotary vaporization to a dry state, and the concentrate was evaporated and transformed into a powder for subsequent use.

2.2.
Screening and Evaluation of Potential AChE Inhibitors in P. cocos 2.2.1. Screening of Potential AChEIs in P. cocos by UF-LC-MS According to the chromatographic results of the ligand (potential AChEIs) in P. cocos, in the range of 25-55 min, 12 components were eluted from P. cocos extract, five of which exhibited an affinity to the AChE receptor. The results were confirmed using AChE at varying levels of activity. As illustrated in Figure 2, the retention time of compound 1 was 36.144 min, and the inhibition rates of 0.5, 1.0, and 2.0 U/mL AChE were 24.10, 24.34, and 24.22%, respectively; For compound 2, the retention time was 39.562 min, and the inhibition rates at 0.5, 1.0, and 2.0 U/mL AChE were 36.59, 37.46, and 37.32%, respectively; For compound 3, the retention time was 41.032 min, and the inhibition rates at 0.5, 1.0, and 2.0 U/mL AChE were 13.12, 13.88, and 13.78%, respectively; For compound 4, the retention time was 46.066 min, and the inhibition rates at 0.5, 1.0, and 2.0 U/mL AChE were 38.32, 38.60, and 38.52%, respectively; For compound 5, the retention time was 55.058 min, and the inhibition rates at 0.5, 1.0, and 2.0 U/mL and AChE were 11.87, 12.03, and 11.90%, respectively. The mixture of P. cocos extract with 1.0 U/mL AChE exhibited the highest binding degree. The binding strength of the compound and AChE was determined as expressed in Equation (1). The order of combined strength capabilities of the active ingredient and AChE was 4 (38.60%) > 2 (37.46%) > 1 (24.34%) > 3 (13.88%) > 5 (12.03%). The five active ingredients that were identified by mass spectrometry were (1) Tumulosic acid, (2) Polyporenic acid C, (3) 3-Epidehydrotumulosic acid, (4), Pachymic acid, and (5) Dehydrotrametenolic acid.

Molecular Docking Simulation of Active Compounds and AChE
The active pocket of AChE mainly contains two ligand active binding sites; one of which is the acylation site and the other is the peripheral anion binding site. The acylation site is near the bottom of the pocket and contains a very important residue, SER (serine). The peripheral anion binding site is located at the edge of the pocket and contains two important residues, the aromatic residue TRP (tryptophan) and the negatively charged residue ASP (Aspartic). The active sites of AChE were analyzed by ligand expansion and SiteMap search. The ligand expansion method was used to define the amino acid within 0.5 nm of the eutectic compound in the crystal structure of acetylcholinesterase as the active site.
The complex conformation of AChE with Tumulosic acid, Polyporenic acid C, 3−Epidehydrotumulosic acid, Pachymic acid, and Dehydrotrametenolic acid was simulated at 298 K using an Amber99sb−ILDN force field. The crystal structure of human AChE (PDB code: 2ACK) was downloaded from the Protein Database. Before the molecular docking experiment, Autodock Tool 1.5.4 software was used to hydrogenate the protein and calculate the Gasteiger charge. Non−polar hydrogen atoms were combined to determine the atomic type, and the non-integer charge on the amino acid residues in AChE was corrected. Key potential target proteins were downloaded from the PDB database. PyMOL software was used for the pretreatment of protein receptor molecules, and Chemdraw 3D software was used for the pretreatment of active compound ligands. The pre-processed potential targets and active compounds were imported into the docking software AutoDock Vina 1.2.0 for molecular docking, and the results were saved in pdbqt format. Huperzine A is the AChEIs extracted from natural products with the best clinical effect in the treatment of AD. Therefore, it was simulated with the target protein, and the binding ability of five active components with AChE in P.cocos was compared by the binding energy of the dock binding energy. In this process, a dodecahedron was used as the simulation system, TIP3P was used as the filled water molecule model, and Na + was used as the counterbalance ion to balance the system charge. The results are shown in Figure 3.

Molecular Docking Simulation of Active Compounds and AChE
The active pocket of AChE mainly contains two ligand active binding sites; one of which is the acylation site and the other is the peripheral anion binding site. The acylation site is near the bottom of the pocket and contains a very important residue, SER (serine). The peripheral anion binding site is located at the edge of the pocket and contains two important residues, the aromatic residue TRP (tryptophan) and the negatively charged residue ASP (Aspartic). The active sites of AChE were analyzed by ligand expansion and SiteMap search. The ligand expansion method was used to define the amino acid within 0.5 nm of the eutectic compound in the crystal structure of acetylcholinesterase as the active site.
The complex conformation of AChE with Tumulosic acid, Polyporenic acid C, 3− Epidehydrotumulosic acid, Pachymic acid, and Dehydrotrametenolic acid was simulated at 298 K using an Amber99sb−ILDN force field. The crystal structure of human AChE (PDB code: 2ACK) was downloaded from the Protein Database. Before the molecular docking experiment, Autodock Tool 1.5.4 software was used to hydrogenate the protein and calculate the Gasteiger charge. Non−polar hydrogen atoms were combined to determine the atomic type, and the non-integer charge on the amino acid residues in AChE was corrected. Key potential target proteins were downloaded from the PDB database. PyMOL software was used for the pretreatment of protein receptor molecules, and Chemdraw 3D software was used for the pretreatment of active compound ligands. The pre-processed potential targets and active compounds were imported into the docking software AutoDock Vina 1.2.0 for molecular docking, and the results were saved in pdbqt format. Huperzine A is the AChEIs extracted from natural products with the best clinical effect in the treatment of AD. Therefore, it was simulated with the target protein, and the binding ability of five active components with AChE in P.cocos was compared by the binding energy of the dock binding energy. In this process, a dodecahedron was used as the simulation system, TIP3P was used as the filled water molecule model, and Na + was used as the counterbalance ion to balance the system charge. The results are shown in Figure 3.

HSCCC Separation of Potential AChE Inhibitors from P. cocos
According to the conditions described in Section 4.7.1, the active components were separated from P. cocos ethanol extract, as shown in  The results showed that Huperzine A and Pachymic acid only acted on the acylating active site of AChE, Tumulosic acid and 3−Epidehydrotumulosic acid only acted on the peripheral anion acting site of AChE, and Polyporenic acid C and Dehydrotrametenolic acid could simultaneously act on both the active site and the peripheral anion acting site of AChE. The average binding energies of the active Tumulosic acid (1), Polyporenic acid C (2), 3−Epidehydrotumulosic acid (3), Pachymic acid (4), Dehydrotrametenolic acid (5), and Huperzine A (6) were −7.77 kcal/mol, −7.88 kcal/mol, −7.37 kcal/mol, −8.01 kcal/mol, −6.92 kcal/mol and −7.51 kcal/mol, as shown in Table S1. The larger the absolute value of the binding energy, the better the binding effect and the stronger the inhibition. These results demonstrated that the model effect is good, and it is consistent with the results of the ultrafiltration experiments, which further verifies the reliability of the ultrafiltration experiments.
2.3. Isolation of Active Compounds from P. cocos 2.3.1. HSCCC Separation of Potential AChE Inhibitors from P. cocos According to the conditions described in Section 4.7.1, the active components were separated from P. cocos ethanol extract, as shown in Figure 4. When we used PET : EtOAc : MeOH : H 2 O (4.0:1.0:3.0:2.0, v/v/v/v) as the solvent system, the overall elution time was approximately 350 min; after 270 min, the baseline tended to level off. Three clear chromatographic peaks observed in the spectrogram corresponded to the active compounds 2, 3, and 4. HPLC was used to determine the purity of the isolated active ingredients, and the purities of the four components were 2 (96.29%), 3 (95.40%) and 4 (97.31%). The purity of the three active ingredients is higher than 95%, which should be considered during follow−up research. 2023, 28, x FOR PEER REVIEW 7 of 16 2, 3, and 4. HPLC was used to determine the purity of the isolated active ingredients, and the purities of the four components were 2 (96.29%), 3 (95.40%) and 4 (97.31%). The purity of the three active ingredients is higher than 95%, which should be considered during follow−up research.

Semi−preparative HPLC Separation of Potential AChEIs from P. cocos
Both HSCCC and semi−preparative HPLC can be used to isolate an active compound from purified substances; however, the separation abilities of the two methods vary. Therefore, on the basis of separation by HSCCC, semi-preparative HPLC was used to further separate the active components of P. cocos. The separation conditions are discussed in Section 4.7.2. Two high−purity monomer compounds were eluted using semi−preparative HPLC, as shown in Figure 5. HPLC detected active compounds 1 and 5, and the purities of the active compounds were 98.11% and 97.82%, respectively.

Semi−Preparative HPLC Separation of Potential AChEIs from P. cocos
Both HSCCC and semi−preparative HPLC can be used to isolate an active compound from purified substances; however, the separation abilities of the two methods vary. Therefore, on the basis of separation by HSCCC, semi-preparative HPLC was used to further separate the active components of P. cocos. The separation conditions are discussed in Section 4.7.2. Two high−purity monomer compounds were eluted using semi−preparative HPLC, as shown in Figure 5. HPLC detected active compounds 1 and 5, and the purities of the active compounds were 98.11% and 97.82%, respectively.
Both HSCCC and semi−preparative HPLC can be used to isolate an active compound from purified substances; however, the separation abilities of the two methods vary. Therefore, on the basis of separation by HSCCC, semi-preparative HPLC was used to further separate the active components of P. cocos. The separation conditions are discussed in Section 4.7.2. Two high−purity monomer compounds were eluted using semi−preparative HPLC, as shown in Figure 5. HPLC detected active compounds 1 and 5, and the purities of the active compounds were 98.11% and 97.82%, respectively.

Structural Identification of Active Compounds
The potential AChE inhibitors in P. cocos extract were analyzed using UPLC−Q−Exactive−MS. The five bioactive components in P. cocos were isolated using HSCCC and semi−preparative HPLC. The chromatograms are shown in Figure 6, and the purity of the four target compounds exceeded 95%. The HPLC determined purities of compounds 1, 2, 3, 4, and 5 were 98.11%, 96.29%, 95.40%, 97.31%, and 97.82%, respectively. The high-purity active ingredients were than processed for the next enzymatic reaction kinetic study.

Structural Identification of Active Compounds
The potential AChE inhibitors in P. cocos extract were analyzed using UPLC−Q− Exactive−MS. The five bioactive components in P. cocos were isolated using HSCCC and semi−preparative HPLC. The chromatograms are shown in Figure 6, and the purity of the four target compounds exceeded 95%. The HPLC determined purities of compounds 1, 2, 3, 4, and 5 were 98.11%, 96.29%, 95.40%, 97.31%, and 97.82%, respectively. The high-purity active ingredients were than processed for the next enzymatic reaction kinetic study.

Identification of Potential AChEIs Using 1 H-NMR and 13 C−NMR Spectroscopy
The 1 H−NMR and 13 C−NMR data were in good agreement with those of the corresponding compounds. The main features of compounds 1, 2, 3, 4 and 5 are summarized in Table S2.

Study on Enzymatic Reaction Kinetics of the Inhibition of AChE by Active Compounds
According to the electroswimming strip introduced in 4.7.2, an electroswimming diagram of the standard mixture solution is obtained (Figure 7). It can be observed that acetylcholine (ACh) and choline (Ch) achieved adequate separation within 20 min, and the separation reoccurrence was excellent. The 1 H−NMR and 13 C−NMR data were in good agreement with those of the corresponding compounds. The main features of compounds 1, 2, 3, 4 and 5 are summarized in Table S2.

Study on Enzymatic Reaction Kinetics of the Inhibition of AChE by Active Compounds
According to the electroswimming strip introduced in 4.7.2, an electroswimming diagram of the standard mixture solution is obtained (Figure 7). It can be observed that acetylcholine (ACh) and choline (Ch) achieved adequate separation within 20 min, and the separation reoccurrence was excellent.

Inhibitory Types of Active Compounds in P. cocos on AChE
The effects of 10 mmol/L changes in the concentration of AChE (500, 1000, 1500, 2000, and 2500 U/L) determined the enzyme activity by changing the concentration of AChE (100, 300, 500, 800, and 1000 μmol/L). The results are illustrated in Figure 8(1(a),  2(a), 3(a), 4(a), and 5(a)). Under the conditions of monomer compounds with varying concentrations, all lines pass through the origin, and with the increase in inhibitor concentration, the slope of the lines gradually decreases, indicating that the inhibition types  5(b)). The results showed that the intersection of the double-reciprocal curves of inhibitors Pachymic acid and Dehydrotrametenolic acid with different concentrations was located on the Y−axis, which suggested that they were competitive inhibitors of AChE. The intersection of the double−-reciprocal curves of Polyporenic acid C and Tumulosic acid inhibitors with different concentrations was located on the X−axis, suggesting that they were non-competitive inhibitors of AChE. The intersection point of the double reciprocal curve of inhibitor 3−Epidehydrotumulosic acid with different concentrations was found in the quadrant, so it was speculated that it was a mixed inhibitor of AChE. Since the intersection point is located in the second quadrant, it is further speculated to be non-competitive-competitive inhibition. The Michaelis-Menten equation and the kinetic parameters of AChE inhibition of active ingredients are shown in Table S3.

Discussion
Current clinical drugs used to treat AD inhibit AChE activity in the body, preventing the degradation of ACh in the synapses and increasing the number of nerve cells in ACh. Therefore, the screening of AChEIs is key to the drug development process. In addition, traditional Chinese medicine is often the source of new active compounds; therefore, screening AChEIs from traditional Chinese medicine is an important trend in AD drug development.  5(b)). The results showed that the intersection of the double-reciprocal curves of inhibitors Pachymic acid and Dehydrotrametenolic acid with different concentrations was located on the Y−axis, which suggested that they were competitive inhibitors of AChE. The intersection of the double−-reciprocal curves of Polyporenic acid C and Tumulosic acid inhibitors with different concentrations was located on the X−axis, suggesting that they were non-competitive inhibitors of AChE. The intersection point of the double reciprocal curve of inhibitor 3−Epidehydrotumulosic acid with different concentrations was found in the quadrant, so it was speculated that it was a mixed inhibitor of AChE. Since the intersection point is located in the second quadrant, it is further speculated to be non-competitive-competitive inhibition. The Michaelis-Menten equation and the kinetic parameters of AChE inhibition of active ingredients are shown in Table S3.

Discussion
Current clinical drugs used to treat AD inhibit AChE activity in the body, preventing the degradation of ACh in the synapses and increasing the number of nerve cells in ACh. Therefore, the screening of AChEIs is key to the drug development process. In addition, traditional Chinese medicine is often the source of new active compounds; therefore, screening AChEIs from traditional Chinese medicine is an important trend in AD drug development.
Five potential AChEIs were screened from the alcoholic extract of P. cocos by UF−LC−MS, and their binding strengths were Tumulosic acid (24.34%), Polyporenic acid C (37.46%), 3−Epidehydrotumulosic acid (13.88%), Pachymic acid (38.60%), and Dehydrotrametenolic acid (2.03%). Using molecular docking technology, the active compound and the target protein AChE were docked and simulated, and the average binding energies were −7.77 kcal/mol, −7.88 kcal/mol, −7.37 kcal/mol, −8.01 kcal/mol, and −6.92 kcal/mol, respectively. The larger the absolute value of the binding energy, the better the binding effect and the stronger the AChE inhibitory ability. Therefore, the results of molecular docking show that the inhibitory ability of the five active components of AChE is as follows: Pachymic acid > Polyporenic acid C > Tumulosic acid > 3−Epidehydrotumulosic acid > Dehydrotrametenolic acid, which is consistent with the ultrafiltration results, which further verifies the accuracy of the ultrafiltration experimental results. UF−LC−MS coupled with molecular docking is a powerful method for screening biologically active compounds in botanical extracts.
The five active ingredients were successfully separated by HSCCC combined with semi−preparative HPLC, and the purities were 98.11%, 96.24%, 97.31%, 99.17%, and 97.82%, respectively. The results show that this method has a good effect on the separation of chemical components in P. cocos. The inhibition types and mechanisms of the active ingredients in AChE were investigated by combining HPCE with enzymatic reaction kinetics. The results showed that the five active ingredients were all reversible inhibitors, among which Polyporenic acid C and Tumulosic acid were non-competitive inhibitors of AChE, 3−Epidehydrotumulosic acid was a mixed inhibitor of AChE, and Pachymic acid and Dehydrotrametenolic acid were competitive inhibitors of AChE. Therefore, when different active ingredients are used in anti−Alzheimer's disease drugs, the influence of the underlying environment on the inhibitory effect should be considered to achieve the best effect.

Apparatus
The analysis of HPLC products was carried out on a Waters 2695 coupled with a Waters 2998 Diode array detector, analysis SunFireTMC18 Column (250 mm × 4.

Screening for Potential AChEIs
UF was used to screen the potential AChE inhibitors from the crude P. cocos extract. A PBS solution was used as the buffer solution for the AChE inhibitors. The reaction mixtures (200 µL) contained 90 µL of 0.5, 1.0, and 2.0 U/mL enzyme, 10 µL (20 mg/mL) sample, and 100 µL buffer solution, with a molecular mass of 30 kDa in 10 mM PBS solution buffer at 37 • C for 30 min. The control sample contained the same reaction mixture, but the sample was replaced with an equal amount of buffer solution. After incubation, each mixture was filtered through a YM−30 UF membrane with the molecular weight cutoff of 30 kDa using a centrifugal filter for 10 min, and the released active ingredients were identified using HPLC. The binding strength was used to characterize the binding strength of the compound and AChE. The calculation formula of the binding strength is illustrated in Equation (1). The A a and A b are the peak area of the blank and the experimental group, respectively.

HPLC Conditions
We used a C 18 (150 mm × 4.6 mm ID) analytical column, where acetonitrile and 0.1% phosphoric acid aqueous solution were used as mobile phases D and B, respectively, for gradient elution. The flow phase gradient program was as follows: 0~10 min, 55% D; 10~40 min, 55~100% D; 40~85 min, 100% D. The flow velocity, detection wavelength, sample volume, and column temperature were 0.4 mL/min, 242 nm, 15 µL, and 30 • C, respectively.  The following 16 sets of solvent systems were used to separate the active compounds in P. cocos. In order to investigate the K value of the iconic compound, we conducted pre-experiments on each solvent system. After the solvent systems were prepared, they were allowed to reach equilibrium. Subsequently, 2.0 mL of both the upper and lower phase was added to a test tube and 2.0 mg of P. cocos extract was added to the test tube and allowed to fully dissolve. The upper and lower phases of the obtained mixtures were filtered through a 0.45 µm filtration membrane and allowed to rest prior to HPLC detection.

UPLC-Q-Exactive Conditions
After complete stratification, 1 mL of each upper and lower phase was taken and dried, and then 1 mL of methanol was dissolved for HPLC determination. The formula for calculating the distribution coefficient is shown below: In the above formula, A upper is the peak area of the target compound in the upper phase, and the A lower is the peak area of the target compound in the lower phase.
A nonlinear correlation was used to analyze the K values of the target compounds, and the optimal volume ratios were calculated. The K values of all compounds were in the range of 0.5−2.0, and thus the solvent systems were considered ideal, as illustrated in Table 2. Based on the obtained results, we selected PET : EtOAc : MeOH : H 2 O (4.0 : 1.0 : 3.0 : 2.0, v/v/v/v) as the solvent system. The other conditions of the HSCCC experiment were as follows: a detection wavelength of 254 nm, a mobile phase velocity of 1.5 mL/min, a screw speed of 800 rpm, a temperature of 25 • C, a sample volume of 5 mL, and a sample concentration of 50 mg/mL. P. cocos extract (50 mg) was accurately weighed and dissolved in 5 mL of 100% MeOH solution followed by ultrasound-assisted filtration through a 0.45 µm filtration membrane.
The active components in the P. cocos extract were separated via semi-preparative HPLC, and the HPLC conditions (including the composition of mobile phase, flow, and injection volume) were optimized. The mobile phase typically consists of water, acetonitrile, and MeOH. Considering the polarity, viscosity, and separation effect, the acetonitrile and H 2 O solution were selected as mobile phases A and B, respectively. The detection wavelength and injection volume were 242 nm and 5 mL, respectively, and the gradient elution procedure is as follows: 10% A, 0~5 min; 10~100% A, 5-60 min; 100% A, 60~150 min.

Solution Preparation
Borax buffer solution: We precisely weighed 0.0381 g borax, dissolved it in deionized water, and kept it in a 100 mL volumetric flask to obtain a buffer solution with a concentration of 1.0 mmol/L (pH 6.80).
NaOH solution: We precisely weighed 0.04 g of the drug, dissolved it in 100 mL ultrapure water, and kept it in a 100 mL volumetric flask to obtain 0.1 mol/L NaOH solution for later use.
AChE solution: We accurately weighed 50 mg 200 U/g AChE and dissolved it in 5 mL 1 mmol/L borax buffer solution; thus, 2.0 U/L AChE solution was obtained. We stored the solution in the dark at −20 • C and diluted it with 1 mmol/L borax buffer solution before use.
Substrate ACh solution: We accurately weighed and dissolved 19.77 mg ACh in 5 mL 1.0 mmol/L borax buffer solution, and a substrate solution with a concentration of 20 mmol/L was obtained. We stored it in the dark at −20 • C and diluted it with 1 mmol/L borax buffer solution before use.

Electrophoresis Method
The selected models were measured by the variation in the substrate ACh and product choline (Ch) in the enzyme reaction. The variations in ACh and Ch were measured using the capillary electroswimming method. The electroswimming strip was as follows: a capillary column temperature of 25 • C and a PDA detection wavelength of 230 nm. The washing sequence and timing of the capillary columns are listed in Table 3.

Conclusions
In this study, a rapid screening method was established for the active compounds of anti-Alzheimer's disease in P. cocos. We used core basic theories such as mathematical model, mass spectrometry, chromatography and computer molecular docking to comprehensively screen the active ingredients against AD in P. cocos. In doing so, we established the identification method of a compound structure based on LC−MS analysis and the separation method of active compounds with high efficiency, and rapid and successful recovery. We also explored the action mechanism of active compounds against AChE. This provides a theoretical basis for the targeted screening and isolation of active compounds and clinical drug development. We believe that this method can be used to identify bioactive compounds for the development of novel anti-Alzheimer's disease therapeutics.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules28031228/s1, Table S1: Molecular docking analysis of potential acetylcholinesterase ligand in Huperzine A and P.cocos.; Table S2: Main features of AChE inhibition by active compounds.; Table S3: Kinetic parameters of AChE inhibition by active compounds.