Investigating the enhancing effects of traditional processing on the pharmacological and electrochemical functions of Polygonum multiflorum using THSG as a key marker

2.1 Chemical and reagents

HPLC-grade acetonitrile and methanol were purchased from Merck (Darmstadt, Germany). THSG was isolated from Radix PM in the Department of Microbiology, Immunology, and Biopharmaceuticals, National Chiayi University, by Dr. Lih-Geeng Chen. The structure of THSG was identified by ¹H and ¹³C NMR. The purity of THSG was found to be 98% by HPLC determination. Emodin and physcion were purchased from Extrasynthese, France. Ascorbic acid, dimethyl sulfoxide (DMSO), dopamine, DPPH, FeCl₃·6H₂O, TPTZ (2,4,6-tripyridyl-s-triazine), Trolox, trifluoroacetic acid (TFA), and dipotassium hydrogen phosphate (K₂HPO₄) were purchased from Sigma-Aldrich (St. Louis, MO, USA). All chemicals were analytical grade for this experiment.

2.2 High-performance liquid chromatographic (HPLC) analysis of THSG, emodin, and physcion in PM

Quantitative analysis of THSG was carried out using an HPLC system equipped with an LC-10ATvp pump, SPD-10A UV-VIS detector, SIL-9A auto-sampler, CTO-10A column oven, and C-R8A chromatopac data recorder (Shimadzu, Tokyo, Japan). Separation was achieved on a Purospher® RP-18e reversed-phase column (5 μm, 4 mm i.d. × 250 mm; Merck, Darmstadt, Germany), with the column temperature set at 40°C. The mobile phase consisted of acetonitrile and 0.05% trifluoroacetic acid (TFA) in water (20:80, v/v), delivered isocratically at a flow rate of 1 mL/min. Detection of THSG was performed at 310 nm with an injection volume of 10 μL. For the analysis of emodin and physcion, the detection wavelength was adjusted to 280 nm, while other chromatographic conditions remained consistent. A gradient elution method was employed for stability evaluation: acetonitrile-0.05% TFA in water (20:80, v/v) from 0-10 min, followed by a shift to 60:40 (v/v) from 10-20 min, and finally to 80:20 (v/v) from 20-30 min.

2.3 PM processing method

To investigate the effects of various processing conditions on the chemical composition of PM, 10 g of dried PM was used for each treatment. The study design included five treatment categories: (1) Morphological classification (PM-1 to PM-6), in which raw PM samples were categorized based on surface appearance; (2) Cross-sectional morphology, where different tissue regions (stem, elongation zone, epidermis, endodermis, and cortex) were separated and individually categorized; (3) Drying at different temperatures, where the raw PM was subjected to drying under three temperature conditions (–20°C, 50°C, and 80°C); (4) Storage period, where dried PM samples were stored at room temperature and analyzed at three time points (1 day, 3 months, and 6 months); and (5) Adjuvant material treatments, where samples were soaked for 30 min in 100 mL of either distilled water (control), standard black soybean juice (prepared by boiling 75 g of black soybeans in 1.5 L of water for 2 h), or concentrated black soybean juice (prepared using 150 g of black soybeans under the same conditions). After the soaking step, samples were directly steamed for 1 h using an electric steamer to simulate traditional processing methods. All treatments were conducted in triplicate to ensure experimental reproducibility.

2.4 PM collection

All samples of commercial products were collected from Guangdong Province, China, or purchased from pharmacies in China. A voucher specimen (Batch 2013-PM) was deposited in the Graduate Institute of Pharmacognosy, Taipei Medical University, Taipei, Taiwan.

2.5 Sample preparation

Prior to HPLC analysis, PM samples underwent extraction using 75% methanol. A sonication process lasting 2 h was applied to facilitate the efficient release of bioactive compounds. The extraction ratio was maintained at 1:5 w/v. After extraction, the supernatant was filtered through a 0.45 μm PVDF membrane filter (FP Vericel, Pall Corporation, Ann Arbor, MI, USA) to remove particulates and obtain a clear solution. This sample preparation protocol ensured high clarity and consistency, which are essential for reliable chromatographic quantification.

2.6 MALDI HDMS analysis condition

Samples for MALDI-MSI analysis were prepared using the Leica Cryostat CM3050S (United States). The PM was sliced and directly fixed onto the Leica Cryostat machine and sliced with a thickness of 10 µm, with optimal cutting temperature (OCT) medium at −20°C. The matrix, 25 mg/mL DHB (2,5-dihydroxybenzoic acid) in 90:10% MeOH:H₂O, was sprayed for further analysis. The Waters SYNAPT G2 8K MALDI-Q-TOF (Waters Group, USA) was utilized to perform the High-Definition MS (HDMS) experiment. The PM sample mass spectra were acquired in negative ion mode with a mass range of 100-1200 Da and a sampling resolution of 100 μM. All imaging analyses were conducted using High-Definition Imaging (HDI) software.

2.7 Antioxidant activities of THSG

The DPPH free radical scavenging assay was conducted based on established methods (Tsai et al., 2022), with all procedures performed under dark conditions to minimize the photodegradation of light-sensitive components. A 5 mg/mL ascorbic acid stock solution was prepared in deionized water and serially diluted in ethanol to construct the standard calibration curve. For the assay, 100 μL of the sample, ascorbic acid standard, ethanol control, or blank was combined with 150 μL of freshly prepared DPPH reagent in a 96-well microplate. After a 30-min incubation at room temperature, absorbance was measured at 517 nm using a microplate spectrophotometer. The ferric-reducing antioxidant power (FRAP) assay was carried out in accordance with previously published protocols (Tsai et al., 2022). A Trolox stock solution (1000 μg/mL) was obtained by diluting a 500 μg/mL solution with ethanol and double-distilled water at a 2:3 ratio, followed by serial dilution to generate the standard curve. In each reaction, 50 μL of sample, standard, or blank was mixed with 1450 μL of FRAP working reagent. All samples were analyzed in triplicate, and absorbance readings were recorded at 593 nm.

2.8 Microbial fuel MFC of THSG

To evaluate the bioenergy-stimulating capability of PM, a dual-chamber H-type microbial fuel cell (DC-MFC) was employed. The system featured graphite electrodes (Grade: IGS743, Central Carbon Co., Ltd., Taiwan) functioning as both anode and cathode, each with an active surface area of approximately 1.649 cm². These electrodes were immersed in the respective electrolyte and culture media during operation. The two 200 mL chambers were separated by a proton exchange membrane (Nafion® NR-212, DuPont™), with an effective membrane area of 0.000452 m² (inner diameter: 1.2 cm). The cathodic chamber contained an electrolyte composed of potassium ferricyanide (K₃Fe(CN)₆; JTBaker®, USA) and dipotassium hydrogen phosphate (K₂HPO₄; Sigma-Aldrich, USA), dissolved in deionized water. The anodic chamber was inoculated with Aeromonas hydrophila (NCBI Taxonomy ID: 644), a Gram-negative, rod-shaped, electroactive bacterium known for its ability to transfer electrons directly to electrodes, degrade organic substrates, and form conductive biofilms. The bacterial strain (NIU01) was cultured in LB broth at 30°C with shaking at 125 rpm for 12 h until reaching an optical density (OD₆₀₀) of approximately 2.1, and 200 mL of this culture was added to the anode chamber to initiate current generation. The PM extract was prepared by dissolving 1.2 g of the sample in 12 mL of deionized water. To assess dose-dependent effects, different concentrations (250-2000 ppm) were added to the anodic chamber while keeping the total volume at 200 mL. Following each treatment, the system, including both chamber and electrodes, was rinsed with sterile deionized water. A consistent bacterial density (OD₆₀₀ ≈ 2.1) was maintained throughout the experiment. Control measurements were taken before the 250 ppm test and after the 2000 ppm trial. For benchmarking purposes, dopamine (0.1 mL of 0.6 M solution) was added post-blank as a reference standard. To record electrical performance, a data acquisition system (DAS 5020; Jiehan Tech Corp., Taiwan) monitored current (I_MFC) and voltage (V_MFC) across an external 1 kΩ resistor. Power and current densities were calculated using the following equations:

with $A_{Anode}$ signifying the effective working area of the graphite anode.

For precise determination of V_MFC and I_MFC at maximum power, linear sweep voltammetry (LSV) was performed using an electrochemical workstation (Jiehan 5600, Jiehan Technology Corp., Taiwan). The LSV curve was used to identify the voltage at which the product of current and voltage (P = IV) peaked, thereby indicating the optimal operating point of the MFC.

2.9 Molecular docking

Structural data for five proteins associated with Parkinson’s disease and three involved in immunomodulatory activity were retrieved from the RCSB Protein Data Bank (https://www.rcsb.org/) in PDB file format. The Parkinson’s-related proteins (PDB IDs: 1JS3, 1XQ8, 2C65, and 3BWM) and immunomodulatory targets (PDB IDs: 2AZ5, 1ALU, and 1OY3) were imported into BIOVIA Discovery Studio for preprocessing. This included the removal of crystallographic water molecules and heteroatoms, followed by the addition of polar hydrogen atoms. Protein structures were further optimized using the “Prepare Protein” protocol, which ensured appropriate protonation states and corrected any missing side chains or atoms, making the targets suitable for molecular docking studies. In parallel, ligand structures of the key bioactive compounds from PM THSG, emodin, and physcion, along with reference drugs levodopa (for Parkinson’s disease) and thiopurine (for immunomodulatory evaluation), were obtained from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) in SDF format. These ligand files were processed using the “Prepare Ligands” function in Discovery Studio, which included charge assignment, tautomer enumeration, and ionization state prediction. The finalized ligand structures were then converted and saved in PDB format for docking experiments. The docking sites for each target protein were defined based on the literature and structural analysis. The X, Y, and Z coordinates, along with the docking box sizes used for the binding site definition, were as follows: 1JS3 (32.90, 29.37, 69.48; box size: 14.00), 1XQ8 (238.13, 83.73, -9.69; box size: 14.00), 3BWM (-13.74, -8.02, -13.28; box size: 14.00), 2C65 (56.11, 141.16, 16.55; box size: 14.00), 2AZ5 (-26.11, 67.32, 39.72; box size: 14.00), 1ALU (-9.81, -16.27, 4.45; box size: 14.00), and 1OY3 (-6.79, -18.27, -24.93; box size: 14.00). These coordinates were used to construct docking spheres that ensured accurate ligand placement within the active or predicted binding pockets of the respective proteins.

2.10 Principal component analysis

Correlation analysis between the three marker molecules and the extraction methods was performed using PCA. In brief, the data summarizing the content of marker substances from the cross-sectional morphology of PM, PM dried at various temperatures, and PM processed by steaming with different adjuvant materials were consolidated into a single dataset and subjected to principal component analysis (PCA) using R. The PCA was performed using the prcomp() function in R, centering and scaling the data to ensure equal weights for all variables. The number of principal components retained was determined based on the Scree plot and the proportion of variance explained by each component. The results were interpreted by examining the scores and loadings of the principal components, providing a visual representation of the data and indicating the importance of each variable in explaining the variance. All analyses were performed using R version 4.4.1, with the prcomp() function from the base R package, ggplot2, and factoextra used for data manipulation and visualization.

2.11 Statistical analysis

All experimental results are expressed as mean ± standard deviation (SD) based on three independent biological replicates (n = 3). To assess the variation in THSG, emodin, and physcion content across five anatomical regions of PM roots, namely the stem, elongation zone, epidermis, endodermis, and cortex, a one-way analysis of variance (ANOVA) was conducted using IBM SPSS Statistics software (Version 26.0; IBM Corp., Armonk, NY, USA). Tukey’s Honestly Significant Difference (HSD) test was employed for post hoc comparisons to identify statistically significant differences among the groups. A p-value<0.05 was considered statistically significant. Within each compound group, distinct lowercase letters were used to denote significantly different means across anatomical regions.

3.1 Linearity and calibration curve preparation

The HPLC spectrum of PM was optimized to establish the retention times for THSG, emodin, and physcion (Fig. 2) at 7.5, 19.2, and 26.2 min, respectively. The linear regression equations indicate the precise quantification of these compounds (Meng et al. 2021). Serial concentrations were prepared, with THSG ranging from 0.5 to 0.0125 mg/mL, emodin from 0.5 to 0.03125 mg/mL, and physcion from 0.25 to 0.03125 mg/mL.

Close

Fig. 2.

HPLC Profile of THSG (a), emodin (b), and physcion (c), and their calibration curves

Export to PPT

3.2 Precision and accuracy

As shown in Table 1, intra-day and inter-day precision for THSG, emodin, and physcion showed relative standard deviations (RSD) below 1% and 4%, respectively. Recovery error was under 5%, indicating the method’s reliability and reproducibility.

3.3 Influence of processing conditions on compound stability

Comprehensive analysis of PM samples revealed statistically significant differences in THSG content among different samples (p<0.05), with PM-4 exhibiting the highest concentration (62.7 ± 0.91 mg/g) (Table 2). Fig. 3 and Table 3 showed that THSG was most abundant in the epidermis (56.91 ± 0.24 mg/g), followed by the cortex and endodermis, and significantly lower in the stem and elongation zone (p<0.05). Emodin levels also varied significantly across tissue regions, while physcion content remained statistically unchanged. Drying temperature had a significant impact on compound stability (Table 4). THSG content was highest at −20°C and declined at 50°C and 80°C (p<0.05), indicating its sensitivity to heat. Emodin showed peak levels at 50°C and decreased significantly at −20°C and 80°C (p<0.05), suggesting moderate temperature facilitates extraction, while higher heat promotes degradation. Physcion content was highest at 80°C (p < 0.05), indicating greater thermal stability or release at elevated temperatures. As shown in Table 5, THSG content significantly decreased over time during storage, from 47.80±0.144 mg/g on Day 1 to 23.10±0.234 mg/g after 6 months (p<0.05), confirming its instability under long-term storage. Processing methods also influenced compound levels (Table 6). THSG content was significantly higher in samples processed with concentrated black soybean juice (33.67 ± 3.28 mg/g) than in raw or water-treated groups (p < 0.05), suggesting enhanced extraction or stability conferred by the adjuvant. Emodin content decreased significantly after processing, while physcion content was highest in the concentrated black soybean group (p<0.05). Overall, THSG was the most responsive to environmental and processing factors, supporting its use as a key marker for evaluating PM quality. Emodin and physcion, though less sensitive, offer complementary insights into thermal and matrix-related compound behavior. These trends emphasize THSG’s sensitivity and utility as a primary processing marker, while emodin and physcion provide complementary thermal insights.

Table 2. The content of marker substances of PM-1 to 6 (n=3 per group).

No.	Samples	THSG (mg/g)	Emodin (mg/g)	Physcion (mg/g)
1		29.0±0.920	0.11±0.003	0.01±0.000
2		37.9±1.020	0.16±0.000	0.01±0.000
3		37.6±0.520	0.16±0.000	0.01±0.000
4		62.7±0.910	0.03±0.000	0.02±0.000
5		51.4±0.280	0.17±0.032	0.02±0.004
6		1.4±0.040	-	-

Close

Fig. 3.

Cross-sectional morphology and anatomical structure of the PM root. The image highlights five major anatomical regions analyzed in this study: epidermis, cortex, endodermis, area of elongation, and stem. These distinct sections were separately sampled for quantification of marker compounds, including THSG, emodin, and physcion, to assess compound distribution across tissues

Export to PPT

3.4 Spatial mapping via MALDI imaging

MALDI-TOF imaging visualized the spatial distribution of THSG (Fig. 4), predominantly localized in the epidermal region of PM roots, in agreement with HPLC data. The molecular ion was identified at m/z 405.1186, confirming its presence and specificity (Liang et al., 2010a; Liang et al., 2010b).

Close

Fig. 4.

Spatial Distribution of THSG in PM root sections

Export to PPT

3.5 Bioenergy output and functional implication in MFCs

Using DC-MFCs, THSG’s effect on bioenergy output was quantified. The highest power density (PD) and amplification factor (1.3134 ± 0.0726-fold) were observed at 40 μM (Table 7, Fig. 5). Despite lacking a typical redox-active structure, THSG’s antioxidant properties may indirectly enhance electron transfer. Emodin and physcion, in contrast, acted as direct electron shuttles, with physcion showing the highest PD, aligning with previous studies on mitochondrial modulation and neurodegenerative relevance (Samudro et al., 2021; Xu et al., 2023). These findings imply that processed PM with elevated THSG may contribute to enhanced MFC performance through antioxidant synergy.

Close

Fig. 5.

Power density curves of MFCs supplemented with different concentrations of THSG

Export to PPT

3.6 Antioxidant activity evaluation of THSG

Antioxidant capacity was evaluated using DPPH and FRAP assays. THSG exhibited a DPPH IC₅₀ of 0.279 ± 0.001 mg/mL compared to ascorbic acid (0.103 ± 0.001 mg/mL) and strong FRAP activity (1361.609 ± 6.237 mg/g) (Table 8), indicating robust electron-donating capacity.

3.7 Molecular docking and target affinity

LibDock computational docking (Table 9) indicated strong binding affinities of THSG and physcion to Parkinson’s disease and immunomodulatory protein targets. For Parkinson’s disease targets, THSG exhibited a notable affinity for protein 1JS3 (LibDock score: 103.70) (Chen et al., 2022). Emodin and physcion exhibited strong interactions, particularly physcion with protein 2C65 (LibDock score: 119.51). Regarding immunomodulatory targets, THSG showed the highest binding affinities among tested compounds for proteins 2AZ5 (LibDock score: 99.34), 1ALU (LibDock score: 111.21), and 1OY3 (LibDock score: 99.23). These interactions, visualized in Fig. 6, indicate stable binding modes and extensive hydrogen bonding and hydrophobic interactions. Conventional drugs levodopa and thiopurine displayed comparatively lower binding affinities than these natural compounds. These results suggest that processed PM, which increases THSG content, may enhance neuroprotective and immunomodulatory efficacy at the molecular level.

Close

Fig. 6.

Binding interaction and 2D receptor-ligand interaction model of different Parkinson’s disease proteins (a-e) and immunomodulation key targets (f-g) with the ligand that obtained the highest docking score (a) IJS3-THSG; (b) 1XQ8-T/HSG; (c) 3BWM-THSG; (d) 2C56-physcion; (e) 2AZ5-THSG; (f) 1ALU-THSG; (g) 1OY3-THSG

Export to PPT

Close

Fig. 6.

Binding interaction and 2D receptor-ligand interaction model of different Parkinson’s disease proteins (a-e) and immunomodulation key targets (f-g) with the ligand that obtained the highest docking score (a) IJS3-THSG; (b) 1XQ8-T/HSG; (c) 3BWM-THSG; (d) 2C56-physcion; (e) 2AZ5-THSG; (f) 1ALU-THSG; (g) 1OY3-THSG

Export to PPT

Close

Fig. 6.

Binding interaction and 2D receptor-ligand interaction model of different Parkinson’s disease proteins (a-e) and immunomodulation key targets (f-g) with the ligand that obtained the highest docking score (a) IJS3-THSG; (b) 1XQ8-T/HSG; (c) 3BWM-THSG; (d) 2C56-physcion; (e) 2AZ5-THSG; (f) 1ALU-THSG; (g) 1OY3-THSG

Export to PPT

3.8 Principal component analysis

PCA analysis identified three principal components (Fig. 7). THSG had the greatest contribution, followed by emodin and physcion. The PCA biplot (Fig. 8) showed strong correlations between THSG and drying temperature, as well as epidermal extraction, confirming their influence on THSG yield. Based on the PCA results in Figs. 7 and 8, THSG was identified as the primary differentiating factor among the marker compounds under various extraction and processing conditions. This is evidenced by the longest eigenvector (red arrow) in Fig. 8, reflecting THSG’s dominant contribution to the overall variance, and by Fig. 7, where principal component 1 (Dim1), which explains 59.2% of the total variance, is primarily influenced by THSG. These results indicate that THSG is the most sensitive compound to processing variations, particularly drying temperature and tissue-specific extraction (e.g., epidermal layer), and thus represents a robust marker for optimizing extraction or processing protocols in PM. Although THSG exhibited the strongest response, emodin and physcion also contributed to the variation, albeit to a lesser extent. For multi-parameter optimization approaches, emodin may serve as a secondary marker, while physcion, being the least responsive, may have limited utility in differentiation. In summary, THSG is the most reliable indicator for assessing processing impacts, and future studies aiming to enhance extraction efficiency should prioritize THSG as a key optimization parameter.

Close

Fig. 7.

Scree plot of the PCA analysis on the marker compounds THSG, emodin, and physcion, along with the various factors involved in the extraction process.

Export to PPT

Close

Fig. 8.

Eigenvector contributions of THSG, emodin, and physcion to the principal components of the data.

Export to PPT

Together, the results form a cohesive narrative in which processing with black soybean juice significantly increases THSG content, particularly in the epidermal tissues of PM roots. Elevated THSG levels enhance antioxidant capacity, correlate with stronger molecular docking affinities to neuroprotective and immunomodulatory targets, and moderately improve power output in MFCs. Although THSG does not function as a classical electron shuttle, its antioxidant properties synergize with the redox-active behaviors of emodin and physcion, which contribute directly to electron transfer. These findings highlight the dual role of THSG as both a chemical quality marker and a functional bioactive compound. Ultimately, traditional processing not only augments PM’s pharmacological efficacy but also expands its potential into modern bioelectrochemical applications, bridging traditional herbal medicine with emerging technologies.