2.1 Plant materials and chemicals

HMLF was collected (Heilongjiang Province, China, northern latitude 45°43′9″, east longitude 126°38′5″), shade dried with final moisture of 3.54%, pulverized (360 μm) and stored at room temperature. All standard items (≥98%), ACN (≥99.9%) and H₃PO₄ (85–90%) of HPLC grade used for HPLC were purchased from Sigma-Aldrich.

2.2 Preparation of crude extract

In a conical flask, 2 g of HMLF was extracted with a certain concentration of ethanol solution using ultrasonic-assisted extraction (UAE) in an ultrasonic bath (XM-500UVF, Kunshan, China). In the UAE, the parameters included liquid/solid ratio, extraction time, extraction temperature, ultrasonic power and ethanol concentration at four levels, which were methodically optimized by Taguchi OA Design (TOAD) to determine the effect of individual factors on the response variable of the total flavonoid content (Table 1). Under the optimal conditions, 1000 g of HMLF samples were extracted, concentrated and suspended in water for macroporous adsorption resin experiments. All experiments were performed in triplicate.

2.3 Enrichment of chemical compositions by macroporous adsorption resin

Static adsorption and desorption experiments were implemented as follows: The pretreated resins (1.0 g of dry resin, Shanghai Yuanye Biotechnology Co. LTD, China, Table S1) were weighed and combined with the sample solution (60 mL) in a 150 mL Erlenmeyer flask, which was then shaken and adsorbed for 6 h in a constant temperature shaker (25 ℃) at 100 rpm. When the adsorption equilibrium was achieved, the resins were filtered out of the solution and the concentrations of the target compounds were determined by HPLC. Subsequently, 60 mL of 60% ethanol was added to the separated resins for desorption at 100 rpm for 6 h. The remaining step was the same as before, with HPLC analysis being used to identify the desorption solution.

The 24 mL sample solution passed down the glass column (1.5 × 50 cm) wet-packed with 5 g (dry weight) of resin at a fixed flow rate of 3 BV/h, and the effluent was gathered and monitored at 1 BV intervals. After attaining adsorption equilibrium, the column was in turn eluted with ultrapure water and 3 BV different ethanol solutions (10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100%), and the results were monitored and analyzed. The enriched sample was eventually obtained for further preparative separation of target compounds by HSCCC.

2.4 Separation and purification of target compounds from HMLF by HSCCC

2.5 Total flavonoid content

The total flavonoid content (TFC) of the samples was calculated and represented as grams of rutin equivalents per 1 g of the extract (RE mg/g extract, y = 0.6756x + 0.0111, R² = 0.9991) as described in our prior work (Liu et al., 2022a).

2.6 Structural identification

ESI-MS/MS, FT-IR, ¹H and ¹³C NMR were used to structurally identify the isolated compounds with the Agilent 6460 Accurate-Mass (Agilent Technologies, CA, USA), Thermo Fisher Nicolet iS50 Fourier transform infrared spectrometer (Thermo Fisher Technologies, USA) and Bruker 500 MHz spectrometer (Bruker BioSpin Corporation, Billerica, MA).

2.7 Molecular docking

The molecular docking procedures of the isolated compounds against AChE and BuChE were performed with AutoDock Vina to investigate the binding mode of ligands with the receptors. The crystal structures of AChE (PDB ID; 3LII) and BuChE (PDB ID; 6ESJ) were obtained from the protein data bank (Dvir et al., 2010; Rosenberry et al., 2017). Before docking, the ligands and receptors were dehydrated, hydrogenated, charged, and saved in PDBQT format. In docking calculation, the receptor was considered semi-flexible while the ligand was flexible, and the grid matrix of the docking region was scaled to guarantee full coverage of the protein molecule (Liu et al., 2022c). The grids were set to (74, 74, 72) and (108, 112, 104), and the docking centers were (82.170, 89.554, −9.450) and (0, −10, −20) for AChE and BuChE, respectively. The best score with the suitable energetically favored conformations was applied for further analysis. The interaction of the isolated compounds with amino acid residues of AChE and BuChE was analyzed and represented using ChimeraX and Maestro 11.5.

2.8 Statistical analysis

Analysis of variance (ANOVA) was performed to confirm whether the process parameters are statistically significant to TFC using MINITAB 17. p < 0.05 was considered as statistically significant.

3.1 Optimization of the conditions for extracting HMLF

The extraction parameters for TFC in HMLF were determined methodically by TOAD. To examine the influence of individual parameters on the TFC response variable (Table 1), five control factors with a total of four levels were chosen as the independent variables in TOAD: liquid/solid ratio (A), extraction time (B), extraction temperature (C), ultrasonic power (D) and ethanol concentration (E). An analysis of variance (ANOVA) was carried out to determine whether the technical parameters were statistically significant using MINITAB 17. Fig. 3A depicted the main effect of the five variables, in which the X- and Y-axes reflected the selection range of each variable and the mean of TFC. It can be discovered that the main effect of the factor increased as the slope of the line connecting the specified ranges increased. This finding was statistically represented in Table 1 as the rank of the main effects of each of the identified factors. The F-value for each parameter reveals which parameter significantly affects TFC, and the higher the F-value, the more significant the factor. In Table 1, the liquid/solid ratio exhibited the highest F-value, with extraction temperature ranking second. The p-values of the factors were used to decide the statistical significance of the terms (Razmi et al., 2022). There was a statistically significant relationship between the TFC and the variables since the p-values for liquid/solid ratio and extraction temperature were less than 0.05. While there was no statistically significant relationship between the other three variables and TFC. The ranking of a factor’s main effect increased with its delta value (Tan et al., 2022), which conformed to the graphical analysis of the interaction illustrated in Fig. 3B. The ranks were liquid/solid ratio, extraction temperature, extraction temperature, ethanol concentration and ultrasonic power, in descending order. From the Pareto chart of five extraction parameters in Fig. S1A-E, it can be found that the influence on TFC was as follows: A₃ > C₃ > E₂ > B₃ > D₃. Among them, liquid/solid ratio, extraction temperature, ethanol concentration and extraction time possessed a superior influence on TFC, which was basically consistent with the results in Table 1. Therefore, the maximal TFC of 70.653 RE mg/g DW was obtained by using 60% ethanol with a liquid/solid ratio of 30 mL/g at 40 ℃ and 400 W of ultrasonic power for 30 min (A₃B₃C₃D₃E₂) by verification experiment.

Close

Fig. 3

Main effects plot for the mean of TFC (A) and interaction plot for TFC (B) in HMLF by Taguchi OA Design (TOAD) experiments; and A-E in the main effects for the mean of TFC represents Liquid/solid ratio (mL/g), Extraction time (min), Extraction temperature (℃), Ultrasonic power (W), Ethanol concentration (%).

Export to PPT

3.2 Preparation of the chemical compositions enriched by macroporous adsorption resin

Nine macroporous resins were evaluated for the absorption and desorption performances of chemical compositions from HMLF. It was discovered that weak-polar AB-8 resin exhibited better performance than others (Fig. 4A and B). Also, the parameters of the resin including feed volume and elution conditions, were optimized by a dynamic desorption experiment. The corresponding feed volumes for the chemical compositions on AB-8 resin were 9/4 BV, 9/4 BV, 7/3 BV, 7/3 BV, 7/3 BV, 7/3 BV, 8/3 BV and 8/3 BV, respectively. Considering all the target chemicals, a 3 BV sample solution on AB-8 resin was chosen for dynamic adsorption (Fig. 5A). As shown in Fig. 5B, the target compounds were hardly desorbed before 40% ethanol. The desorption rates of target chemicals rose dramatically after the ethanol concentration exceeded 40% and peaked at 70% ethanol. After achieving optimal adsorption equilibrium, the adsorbate-laden column was eluted with 40% and 70% ethanol at a flow rate of 3 BV/h. As can be seen from Fig. 5C, the non-target components could be removed after elution with 40% ethanol of 3 BV, and the concentrations of target compounds in the effluent increased significantly. Subsequently, 70% ethanol of 8 BV can elute most of the target compounds adsorbed on the resin column. The optimal elution strategies were 3 BV of 40% ethanol and 8 BV of 70% ethanol, respectively. Therefore, the optimum enrichment and separation parameters for target compounds were as follows: feed volume of 3 BV, gradient elution of 3 BV of 40% ethanol and 8 BV of 70% ethanol. Under the above conditions, 10.31 g of enriched product was obtained in the scale-up experiment. After one run of AB-8 resin treatment, the contents of target compounds in the enriched product reached 0.15%, 0.11%, 0.51%, 0.31%, 5.52%, 4.09%, 6.82% and 0.56% with the recovery yields of 82.47%, 79.62%, 81.95%, 84.77%, 80.15%, 79.34%, 80.06% and 88.23%, which were 7.32–, 6.96-, 9.55-, 10.44-, 8.92-, 8.44-, 9.30- and 8.93-fold to those in crude extracts, respectively. It can be found that the target components can be easily and effectively enriched.

Close

Fig. 4

Adsorption capacity (A) and desorption ratio (B) of RU, HY, ISQ, QOR, CHA, CAA, HI and QOG using nine different macroporous resins.

Export to PPT

Close

Fig. 5

Dynamic breakthrough curves (A), dynamic desorption curves (B) and profile of desorption of RU, HY, ISQ, QOR, CHA, CAA, HI and QOG (C). (A-C) ^a Concentration of HY, ISQ and HI, ^b Concentration of RU, QOR, CHA, CAA and QOG.

Export to PPT

3.3 Selection of TPSS

A suitable TPSS is imperative to effectively separate target compounds by HSCCC. The K value is the most essential factor influencing the screening of the solvent system. The ideal K value should be in the range of 0.2–2.0 to get a satisfactory separation. A larger K value would result in prolonged retention time, inadequate resolution, and overly broad peaks, whereas a lower K value may be unable to separate adjacent compounds with poor peak resolution (Pan et al., 2020). This study investigated and determined the K values of different target chemicals in diverse hydrophilic and lipophilic TPSSs, as shown in Table 2. The polarity of target compounds in HMLF was moderate and the K values of the solvent system containing n-Butanol and CHCl₃ were excessively high. Thus, EtOAc-MeOH-H₂O (100:3:100, v/v/v) and Hex-EtOAc-MeOH-H₂O (1:5:1:5, v/v/v/v) with feasible K values and α were applied to separate target compounds from HMLF.

3.4 Purification and identification of target compounds

To accomplish the ideal separation of target compounds, the rotation speed should be high enough at a low mobile phase flow rate to enable for sufficient liquid–liquid extraction. The flow rate restricts the separation duration and quantity of stationary phase retained in the column, while low rotational speeds reduce the retention rate and extremely high rotational speeds widen the peaks owing to system vibration induced by the intense pulsing of the high pressure in the column (Li et al., 2022; Liu et al., 2022b). Five target compounds were separated and purified from 200 mg of the enriched product by HSCCC with the solvent system EtOAc-MeOH-H₂O (100:3:100, v/v/v) was performed at room temperature with a flow rate of 2 mL/min and a rotational speed of 900 rpm. The HSCCC chromatogram and the HPLC chromatogram of the five target compounds by HSCCC were demonstrated in Fig. 6. Subsequently, four target compounds were separated and purified from the enriched product by HSCCC with the solvent system Hex-EtOAc-MeOH-H₂O (1:5:1:5, v/v/v/v) at room temperature and a flow rate of 2 mL/min with a rotational speed of 850 rpm. The HSCCC chromatogram and the HPLC chromatogram of the four target compounds by HSCCC were demonstrated in Fig. 7. A total of 50.4 mg of target compounds were obtained from the eluent in a single run when combined with further crystallization. The final product purities were determined by HPLC to be>95% by the peak area normalization method with recovery yields of 93.91%-98.02%.

Close

Fig. 6

HSCCC chromatogram of five target compounds isolated from HMLF (A) and HPLC chromatogram of five HSCCC fractions (B). 1, CHA; 2, CAA; 3, QOR; 4, RU; 5, HI. For HSCCC, solvent system: EtOAc-MeOH-H₂O (100:3:100, v/v/v); rotation speed: 900 rpm; flow rate: 2 mL/min; detection wavelength: 254 nm; stationary phase: upper phase; retention percentage of the stationary phase: 71.7%.

Export to PPT

Close

Fig. 7

HSCCC chromatogram of four target compounds isolated from HMLF (A) and HPLC chromatogram of four HSCCC fractions (B). 1, HY; 2, ISQ; 3, HI; 4, QOG. For HSCCC, solvent system: Hex-EtOAc-MeOH-H₂O (1:5:1:5, v/v/v/v); rotation speed: 850 rpm; flow rate: 2 mL/min; detection wavelength: 254 nm; stationary phase: upper phase; retention percentage of the stationary phase: 66.7%.

Export to PPT

The structural identification of the isolated compounds were performed and confirmed by ESI-MS/MS, FT-IR, ¹H and ¹³C NMR spectra data in the analysis and by comparison with literature (Xu et al., 2022; Dall'Acqua et al., 2020; Santos et al., 2021) as shown in the Supporting information (Fig. S2-9). Ultimately, eight compounds obtained as chlorogenic acid (peak 1 in Fig. 6B, 3.2 mg, 96.91%), caffeic acid (peak 2 in Fig. 6B, 2.1 mg, 98.34%), quercetin-3-O-robibioside (peak 3 in Fig. 6B, 1.7 mg, 95.28%), rutin (peak 4 in Fig. 6B, 0.6 mg, 98.97%), hibifolin (peak 5 in Fig. 6B and peak 3 in Fig. 7B, 11.3 mg and 12.1 mg, 97.72%), hyperin (peak 1 in Fig. 7B, 10.4 mg, 98.04%), isoquercetin (peak 2 in Fig. 7B, 7.9 mg, 97.13%) and quercetin-3′-O-glucoside (peak 4 in Fig. 7B, 1.1 mg, 96.89%).

3.5 Molecular docking analysis of isolated compounds against acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE)

The molecular docking analysis of the AChE and BuChE inhibitory potentials of eight isolated compounds was performed using AutoDock Vina. The isolated compounds showed binding affinities ranging from −6.9 to −8.7 kcal/mol against AChE, as displayed in Fig. 8. Among the eight isolated compounds, the highest binding affinity against AChE was recorded as −8.7 kcal/mol for RU and HI. Hydrogen bonds with Thr238, Arg296, Gln369, Asn533 and Ser541 (Fig. 9A, Table S2), hydrophobic interactions with Pro235, Pro312, Val367, Pro368, Val370, Trp532, Leu536, Pro537 and Leu540 (Fig. S9A) in molecular docking results of RU. Hydrogen bonds with Gly234, Pro235, Thr238, Gln369 and His405 (Fig. 9C, Table S2), hydrophobic interactions with Pro235, Pro312, Pro368, Val370, Pro410, Trp532, Leu536 and Leu540 were detected for HI (Fig. S10C). Notably, the oxygen atoms O5 and O13 of RU formed two polar bonds with two amino acid residues Gln369 and Thr238 of AChE, similar to HI. The interactions of the isolated compounds with BuChE were also analyzed, and it was calculated that the highest binding affinity with −10.0 kcal/mol for ISQ (Fig. 8). ISQ generated hydrogen bonds with Gln67, Asp70, Trp82, Asn83, Gly115, Tyr128 and Ser198 on BuChE (Fig. 9E), and hydrophobic interactions with Pro84, Tyr114, Leu125, Ala199, Trp231, Pro285, Leu286, Val288, Phe329, Phe398 and Ile442 (Fig. S10E). The location of ligands bound depends on Phe 329 and Tyr 332 of the E-helix, whose side chains project into the active site gorge (Darvesh et al., 2008). Phe329 appears to play an important role together with Trp231 in stabilizing an aromatic system in the groove of the acyl-binding pocket. It can be found that all the molecules interacted with AChE or BuChE in the same region. The main residues at the active site from AChE and BuChE for all the compounds and standards were noted to be Tyr, Trp, Ser, Phe and Trp, Tyr, His, Pro (Amjad et al., 2023). Inhibitors frequently used in the treatment of AD, such as Donepezil, Rivastigmine, and Galantamine were approved by the FDA. Combined with previous binding affinity data of these inhibitors against AChE and BuChE by many researchers (Table S4), it was found that the isolated RU, HI and ISQ had higher binding affinity than the drug molecules (Islam et al., 2013; Ghalloo et al., 2022). It was discovered through the examination of the isolated compounds’ mechanism of binding inside the active sites of AChE and BuChE that the compounds fixed themselves inside the pocket by entwining numerous contacts (Larik et al., 2020). All of the data anticipated that the isolated compounds could have an inhibitory effect against AChE and BuChE. further predicted the possible inhibitory activity of the isolated compounds against AChE and BuChE.

Close

Fig. 8

The heatmap of the binding energy of isolated compounds (RU, HY, ISQ, QOR, CHA, CAA, HI and QOG) to AChE and BuChE and the 2D chemical structures of the ligands for molecular docking.

Export to PPT

Close

Fig. 9

Molecular interactions of isolated compounds against AChE (A-C) and BuChE (D-F). A, RU-AChE; B, HY-AChE; C, HI-AChE; D, QOR-BuChE; E, ISQ-BuChE; F, QOG-BuChE.

Export to PPT