Elucidating the Inhibitory Mechanism of Raisin Polyphenols Against α-Amylase: Insights from Multispectral Analysis and Computational Simulations

1. Introduction

Against the backdrop of major public health challenges associated with global dietary shifts, diabetes and its complications have become the ninth leading cause of death worldwide, after only cancer and cardiovascular disease[1]. Extensive evidence shows that postprandial hyperglycemia is closely associated with type 2 diabetes. Accordingly, controlling postprandial blood glucose has become a central goal of diabetes management[2]. One effective approach is to intervene in carbohydrate digestion, because α-amylase is the key enzyme that hydrolyzes α-1,4-glycosidic bonds in starch and therefore plays an essential role in glucose release. Inhibition of α-amylase can delay starch digestion and subsequent glucose absorption[3,4]. However, although currently used synthetic inhibitors, such as acarbose, voglibose, and miglitol, are effective, they are often associated with gastrointestinal side effects and limited long-term tolerability[5,6]. Identifying safer and more effective alternatives has therefore become an important research priority. In this context, natural polyphenols have attracted considerable attention because of their broad bioactivity and favorable safety profile. Polyphenols have been extensively investigated as natural inhibitors owing to their remarkable biological activities and favorable safety profiles[7]. For example, Zhu et al.[7] explored the inhibitory mechanism of polyphenols from Flos Trollii against α-amylase, while Wu et al.[8] evaluated the α-amylase inhibitory effects of polyphenols derived from blueberry leaves.

Xinjiang is one of the major raisin-producing regions in China, and the “Green Pearl” raisins from Turpan have received national and international recognition[9]. However, studies on the molecular mechanisms by which raisin polyphenolic compounds inhibit α-amylase remain limited. An integrated strategy that combines multispectral techniques with computational modeling has been widely used to investigate α-amylase inhibition at the molecular level[10,11]. Experimentally, inhibitory activity assays and enzyme kinetics quantify inhibitory potency and identify inhibition modes[12], whereas multispectral methods, such as fluorescence spectroscopy and circular dichroism, provide direct evidence of ligand-enzyme interactions and conformational changes[13]. Computationally, molecular docking predicts preferred binding conformations and key intermolecular forces within the active site, whereas molecular dynamics simulations characterize complex stability, conformational flexibility, and binding free energy under near-physiological conditions[14]. By integrating experimental validation with computational prediction, the binding behavior, conformational changes, intermolecular interactions, and dynamic stability of α-amylase inhibitors can be systematically characterized, thereby establishing a more complete “structure-activity-mechanism” relationship[15,16]. For example, Liao et al.[17] used this strategy to clarify the inhibitory mechanism of Zn²⁺ on pancreatic and salivary α-amylases. Zn²⁺ binding induced distinct secondary-structure changes, including decreased α-helix content and increased β-sheet, β-turn, and random coil contents. Zn²⁺ coordinated with His217 and formed a cation-π interaction with Trp244, locking the dihedral angle C–CA–CB–CG in a planar conformation and rigidifying the active pocket. This restriction suppressed the opening-closing motion required for catalysis and thereby inhibited enzymatic activity. Similarly, Avwioroko et al.[18] characterized the interaction between citric acid and α-amylase and examined its inhibitory potential, antiglycation activity, and related mechanisms through spectroscopic, kinetic, and molecular docking analyses. Comparative studies that integrate multispectral techniques with molecular dynamics simulations remain limited, even though these approaches are important for systematically clarifying the inhibition mechanisms, binding behaviors, conformational changes, and dynamic stability of purified polyphenols and their key monomeric compounds. To date, the inhibitory mechanism of purified raisin polyphenols against α-amylase has not been reported, and the structure-activity-mechanism relationship of their characteristic phenolic monomers remains unclear at the molecular level.

Against this background, this study combined multispectral techniques with computational modeling to clarify how raisin polyphenols inhibit α-amylase. The binding interactions between macroporous resin-purified raisin polyphenols and α-amylase, as well as the accompanying conformational changes, were systematically examined. Through structure-activity analysis, the molecular basis for the inhibitory effects of raisin polyphenols as natural α-amylase inhibitors was elucidated, thereby revealing the microscopic basis of their potential role in blood glucose regulation. These findings provide a theoretical and methodological foundation for the development of functional foods designed to regulate blood glucose levels.

2. Materials and Methods

2.1. Chemicals and Materials

Raisins from Xinjiang, China, were purchased commercially. Analytical-grade reagents included anhydrous sodium carbonate, Folin-Ciocalteu reagent, ABTS, DPPH, ascorbic acid, and ethanol (Beijing InnoChem Science & Technology Co., Ltd., China); macroporous resins (XAD7HP, X-5, D101, AB-8, and NKA-II; Shanghai Yuanye Bio-Technology Co., Ltd., China); PPA (50 U/mg), acarbose, ferulic acid, quercetin, isoquercetin, and phosphate-buffered saline (PBS) (Shanghai Macklin Biochemical Co., Ltd., China); and 3,5-dinitrosalicylic acid (DNS; Shanghai Aladdin Biochemical Technology Co., Ltd., China). All reagents were of analytical grade and were used without further purification. Deionized water was used throughout.

2.2. Adsorption Kinetic Curves and Fitting Models

Adsorption kinetic curves and corresponding fitting models were obtained with slight modifications to a previously reported method[19]. Exactly 5 g of D101 macroporous resin was weighed and added to 25 mL of raisin polyphenol solution. Adsorption was carried out with shaking at 150 rpm and 25℃ for different contact times. At each time point, the concentration of the raisin polyphenol solution was determined, and the adsorption capacity of the macroporous resin for raisin polyphenols was calculated using Equation (1).

$$ln\left(Q_e-Q_t\right)=-k_{1}t+ln{Q}_e$$

(1)

$${\displaystyle \frac{t}{Q_t}}={\displaystyle \frac{1}{k_2{Q}_e^2}}+{\displaystyle \frac{t}{Q_e}}$$

(2)

The adsorption kinetics of polyphenols on the resin were evaluated using the pseudo-first-order model (2) and the pseudo-second-order model (3).

2.3. Adsorption Isotherms and Fitting Models

Adsorption isotherms and corresponding fitting models were determined with slight modifications to a previously reported method[20]. Five grams of the selected D101 macroporous resin was mixed with 25 mL of aqueous raisin polyphenol solution at different concentrations in a conical flask and shaken for 12 h at 25, 35, and 45℃. The polyphenol content and equilibrium adsorption capacity of D101 resin were then determined, and the equilibrium data were fitted with the Langmuir equation (4) and the Freundlich equation (5).

$${\displaystyle \frac{Q_e}{C_e}}={\displaystyle \frac{1}{Q_m}}{C}_e+{\displaystyle \frac{1}{Q_m{K}_L}}$$

(3)

$$ln{Q}_e={\displaystyle \frac{1}{n}}ln{C}_e+ln{K}_F$$

(4)

where Q_m is the theoretical maximum adsorption capacity (mg/g); K_L is the equilibrium constant of the Langmuir model (L/mg); K_F is the equilibrium constant of the Freundlich model [(mg/g)(L/mg)1/n]; and 1/n is an empirical constant that reflects the adsorption driving force of the adsorbent.

2.4. Dynamic Adsorption and Desorption Curves

Dynamic adsorption and desorption experiments were performed in a 20 × 300 mm chromatographic column packed with 20 g of pretreated macroporous resin by the wet method. A 250 mL aliquot of raisin polyphenol extract (3.5 mg/mL) was loaded onto the column at flow rates of 2.4, 3.6, and 4.8 BV/h, with 5 mL fractions collected for absorbance determination. The breakthrough point was defined as the eluate absorbance reaching one-tenth that of the feed solution, and the corresponding breakthrough curve was constructed.

At room temperature, the chromatographic column was rinsed with distilled water at 2.4 BV/h to remove water-soluble impurities from the macroporous resin. After the water eluate became colorless, the resin column was eluted with 90% ethanol at 2.4 BV/h. The ethanol eluate was collected every 2 mL, and its polyphenol concentration was determined by the Folin-phenol method. The elution profile was then monitored and an elution curve was constructed.

2.5. UHPLC-MS/MS Identification of Raisin Polyphenols

Phenolic compounds were identified using an UltiMate 3000 high-performance liquid chromatograph coupled to an Orbitrap Exploris 480 high-resolution mass spectrometer. Chromatographic conditions were as follows: ACQUITY UPLC HSS T3 column (2.1 × 100 mm, 1.8 μm); mobile phase A, deionized water containing 0.1% formic acid; and mobile phase B, acetonitrile containing 0.1% formic acid. The gradient elution program was as follows: mobile phase B 0-30% (0-10 min), mobile phase B 30-40% (10-25 min), mobile phase B 40-50% (25-30 min), mobile phase B 50-70% (30-40 min), mobile phase B 70-100% (40-45 min), mobile phase B 100% (45-60 min), mobile phase B 100-0% (60-60.5 min), and mobile phase B 0% (60.5-70 min). The injection volume was 1 μL, the flow rate was 0.4 mL/min, and the column temperature was 35℃.

Mass spectrometry was performed with an electrospray ionization (ESI) source in negative ion mode. The MS₁ resolution was set to 70,000 and the MS₂ resolution to 17,500. The capillary voltage was 3.2 kV, the nebulizer pressure was 1.5 bar, the dry gas temperature was 320℃, the carrier gas flow rate was 15 L/min, and the mass scan range was 100-1,200 m/z.

2.6. Assessment of Antioxidant Capacity of Raisin Polyphenols

The hydroxyl radical-scavenging activity was determined according to a previously reported method[21]. Briefly, 1.00 mL each of 9.0 mM FeSO₄, 9.0 mM salicylic acid in ethanol, raisin polyphenol solution at varying concentrations, and 8.8 mM H₂O₂ was mixed to form the reaction system. The mixture was incubated at 37℃ for 30 min, cooled to 25℃, and the absorbance was measured immediately at 530 nm with a UV spectrophotometer.

The DPPH radical-scavenging activity was determined according to an established method[22]. Two milliliters of 0.1 mM DPPH ethanolic solution was mixed with 2.00 mL of raisin polyphenol solution at various concentrations. The mixture was incubated in the dark at 25℃ for 30 min, and the absorbance was measured at 519 nm.

The ABTS radical-scavenging activity was determined as previously described[23]. A working solution was freshly prepared by diluting the ABTS stock solution with distilled water to obtain an absorbance of approximately 0.70 at 750 nm. Then, 0.30 mL of raisin polyphenol solution at different concentrations was mixed with 3.00 mL of the ABTS working solution. After incubation in the dark at 25℃ for 6 min, the absorbance was recorded immediately at 734 nm.

2.7. Inhibitory Capacity of Raisin Polyphenols toward PPA

The α-amylase inhibition assay was performed with slight modifications based on the method of Xie et al.[24]. PPA was prepared as a 0.05 mg/mL solution in PBS buffer (25 mM, pH 6.8). Equal volumes of this PPA solution and each of the five polyphenol solutions (0.5-15 mg/mL) were mixed and incubated at 37℃ for 10 min. Subsequently, 100 μL of 10 mg/mL starch solution was added, and the mixture was incubated again at 37℃ for 10 min. Then, 150 μL of DNS reagent was added, and the reaction mixture was heated in a boiling water bath for 10 min. After cooling in an ice bath under light protection, the absorbance was measured at 540 nm. Acarbose served as the positive control. Four parallel groups were established: blank (A₁), blank control (A₂), sample (A₃), and sample control (A₄). The α-amylase inhibition rate was calculated using Equation (5):

$$\mathrm{Inhibitory} \mathrm{activity}\left(\%\right)=\left(1-{\displaystyle \frac{A_1-A_2}{A_3-A_4}}\right)\text{×}100$$

(5)

2.8. Kinetic Analysis of α-Amylase Inhibition

Kinetic analysis of α-amylase inhibition by purified raisin polyphenols (PRP) and their major monomeric components was performed with minor modifications based on a published protocol[25]. The reaction system contained soluble starch at varying concentrations (2.5-15 mg/mL) as the substrate and polyphenols at 0-4 mg/mL as the inhibitor, with PPA (0.05 mg/mL) as the enzyme. The initial reaction rate was determined spectrophotometrically, and the inhibition kinetics were analyzed using Lineweaver-Burk plots, in which the reciprocal of the reaction rate (1/V) was plotted against the reciprocal of the substrate concentration (1/[S]). The kinetic parameters (K_m and V_max) were obtained from the following models:

$$\mathrm{Competitive} \mathrm{inhibition}\text{:} {\displaystyle \frac{1}{V}}={\displaystyle \frac{K_m}{V_{\mathit{max}}}}\text{×}1+{\displaystyle \frac{\left[I\right]}{K_i}}\text{×}{\displaystyle \frac{1}{\left[S\right]}}+{\displaystyle \frac{1}{V_{\mathit{max}}}}$$

(6)

$$\mathrm{Noncompetitive} \mathrm{inhibition}: {\displaystyle \frac{1}{V}}={\displaystyle \frac{K_m}{V_{\mathit{max}}}}\text{×}1+{\displaystyle \frac{\left[I\right]}{K_i}}\text{×}{\displaystyle \frac{1}{\left[S\right]}}+{\displaystyle \frac{1}{V_{\mathit{max}}}}\text{×}1+{\displaystyle \frac{\left[I\right]}{K_i}}$$

(7)

where [I] represents the inhibitor concentration, and K_i denotes the inhibition constant.

2.9. Fluorescence Quenching Assay of α-Amylase

Fluorescence quenching assays were performed according to Huang et al.[26] and Liu et al.[27] by incubating PPA (1.0 mg/mL) with polyphenol solutions at varying concentrations (0.00625-0.4 mg/mL) at a 1:1 volume ratio for 15 min at 310 K. The fluorescence spectra were recorded with a Hitachi F-7100 spectrofluorometer (Japan) using an excitation wavelength of 280 nm and an emission range of 300-400 nm. Measurements were carried out in a 1 cm quartz cuvette, and the spectra were corrected by subtracting the corresponding buffer or ligand blanks. Quenching constants and binding-site numbers were determined using the Stern-Volmer and double-logarithmic equations:

$${\displaystyle \frac{F_0}{F}}=1+K_q{\mathsf{\tau }}_0[Q]=1+K_{\mathit{SV}}\left[Q\right]$$

(8)

$$\mathit{log}{\displaystyle \frac{F_0-F}{F}}=\mathit{log}{K}_a+n\mathit{log}\left[Q\right]$$

(9)

where F₀ and F represent the fluorescence intensities of PPA before and after polyphenol binding, respectively; K_q is the bimolecular quenching rate constant; and τ₀ is the intrinsic fluorescence lifetime of the enzyme in the absence of quencher. For α-amylase, τ₀=2.97×10^-8 s (Prendergast et al., 1983); [Q] is the molar concentration of the polyphenolic ligand; Ka is the binding constant; and n represents the number of binding sites.

2.10. CD Analysis

CD spectroscopy was used to evaluate secondary-structure changes in PPA upon interaction with polyphenols. The measurements followed the procedure of Avinash Kumar et al. [28] with minor modifications. Far-UV CD spectra of PPA (2.5 μM), either alone or in the presence of 20 μM polyphenols (PRP, ferulic acid, isoquercetin, or quercetin), were recorded at 25℃ using an Applied Photophysics qCD spectrometer equipped with a 0.1 cm pathlength quartz cuvette. All spectra were baseline-corrected by subtracting the corresponding PBS and polyphenol solution spectra. Secondary-structure deconvolution was performed with the DichroWeb server (SELCON3 algorithm, reference set 7, 190-260 nm) to quantify changes in α-helix and β-sheet content.

2.11. Molecular Docking

Molecular docking simulations were conducted to predict the binding modes and interactions between polyphenols and PPA. The crystal structure of PPA (PDB ID: 1HX0) was prepared by removing water molecules, adding hydrogen atoms, and automatically defining the active site. Docking was performed using AutoDock 4.2.6[29] with grid boxes of 100 × 100 × 100 ų covering the entire protein. Ligand structures (ferulic acid, quercetin, and isoquercetin) were energy-minimized using the Amber14 force field before docking. The resulting binding conformations were visualized and analyzed with PyMOL 2.3.0 to evaluate optimal poses and key intermolecular interactions.

2.12. MD Simulations

MD simulations were performed to examine the stability and interaction dynamics of PPA-polyphenol complexes. All simulations were carried out using GROMACS 2022.1[30] with the Amber14sb force field. The protein-ligand complexes were solvated in an 8.0 nm³ cubic box containing TIP3P water molecules under periodic boundary conditions. After energy minimization, each system underwent a 100 ns production run in the NPT ensemble. The Leapfrog integrator was used with a 2 fs time step, LINCS constraints were applied to all bonds, and temperature and pressure were maintained with the V-rescale thermostat (298.15 K) and Parrinello-Rahman barostat, respectively. Trajectory analyses were performed using standard GROMACS utilities, and visualization was carried out with VMD [31].

3. Results and Discussion

3.1. Screening of Macroporous Resins and Process Optimization

To enrich polyphenols from raisin extract, five macroporous resins (XAD7HP, X-5, D101, NKA-II, and AB-8) were evaluated, and their physical properties are summarized in Supplementary Table S1. The adsorption and desorption performance of these resins is shown in Supplementary Figure S1(a). Among them, D101 resin exhibited the best overall performance, combining high adsorption capacity with efficient desorption, and was therefore selected for subsequent purification experiments. Figure 1(b) shows that the adsorption capacity increased as the polyphenol concentration in the solution increased. This trend indicates that a higher initial concentration enhanced the driving force for mass transfer and promoted adsorption. However, when the concentration exceeded 3.5 mg/mL, the adsorption ratio began to decline. As shown in Figure 1(c), the adsorption capacity of D101 resin first increased and then gradually decreased as the pH of the sample solution increased, reaching a maximum at pH 5. This trend may be attributed to stronger hydrogen-bonding interactions between polyphenols and the resin under mildly acidic conditions. At higher pH, deprotonation of phenolic hydroxyl groups can weaken these interactions and reduce adsorption. Figure S1(d) indicates that desorption efficiency increased markedly as the ethanol concentration in the eluent rose to 90%, but decreased beyond this point. At lower ethanol concentrations, the hydrogen-bonding interactions between polyphenols and the resin may not be disrupted sufficiently, whereas excessively high ethanol concentrations may co-elute other adsorbed impurities, such as proteins and pigments.

Based on these observations, a sample-solution pH of 5, a polyphenol concentration of 3.5 mg/mL, and an eluent containing 90% ethanol (v/v) were selected for subsequent experiments.

3.2. Adsorption Kinetics and Fitting Models

Adsorption kinetics describe the relationship between the adsorption rate and the time required to reach equilibrium and are therefore important for equipment selection and process design[32]. As shown in Figure 1(a), the adsorption capacity of D101 resin for polyphenols increased rapidly during the first 100 min, after which the rate of increase slowed between 100 and 260 min. The rapid initial uptake can be attributed to the large number of available adsorption sites on the resin surface. As these surface sites became progressively occupied, mass-transfer resistance increased and the adsorption rate gradually decreased until equilibrium was reached.

In general, adsorption proceeds through three stages: film diffusion (boundary-layer diffusion), intraparticle diffusion, and final adsorption at active sites. Kinetic models describing intraparticle diffusion are not suitable for the entire dynamic adsorption process, but they can help identify the rate-limiting step. To further characterize the adsorption behavior and clarify the mechanism, the pseudo-first-order and pseudo-second-order models were fitted, as shown in Figure 1(b-c). The pseudo-first-order model is commonly used to describe the initial stage of adsorption, whereas the pseudo-second-order model is often applied to the overall adsorption process. As shown in Table 1, the adsorption capacity predicted by the pseudo-first-order model (Q_e = 2.57 mg/g) was closer to the experimental value (Q_e = 2.60 mg/g) than that predicted by the pseudo-second-order model (Q_e = 8.48 mg/g). In addition, the R² value of the pseudo-first-order model (0.993) was higher than that of the pseudo-second-order model (0.984). These results indicate that the pseudo-first-order model better describes the adsorption of raisin polyphenols on D101 resin.

3.3. Adsorption Isotherms

An adsorption isotherm describes the relationship between the equilibrium concentration of the adsorbate and the equilibrium adsorption capacity of the adsorbent once equilibrium has been reached. Analysis of adsorption isotherms helps clarify the interactions between adsorbents and adsorbates, optimize process parameters, and reduce energy consumption[33].

As shown in Figure 2(a), at a given concentration, the equilibrium adsorption capacity of the resin gradually decreased as temperature increased, indicating that elevated temperatures are unfavorable for adsorption. In addition, as the equilibrium concentration (C_e) increased, the amount of polyphenols adsorbed on D101 resin increased until saturation was approached.

As shown in Figure 2(b-c), the adsorption behavior of D101 resin at 25, 35, and 45℃ was analyzed using the Langmuir and Freundlich models. The Langmuir model assumes monolayer adsorption on a homogeneous solid surface, with equivalent adsorption sites and no interactions among adsorbed molecules. As shown in Table 2, the predicted maximum adsorption capacity (Q_m) decreased as temperature increased, indicating that adsorption of polyphenols on D101 resin was exothermic. In addition, the K_L value decreased with increasing temperature, suggesting that lower temperatures favored polyphenol adsorption on D101 resin. According to the Freundlich model, adsorption occurs on a heterogeneous surface, and each adsorption site has a different binding energy. In general, adsorption is favorable when 1/n lies between 0 and 1[20]. As shown in Table 2, the 1/n values obtained at the different temperatures were all within this range, indicating that polyphenols could be readily adsorbed on D101 resin. Overall, because the Langmuir model gave the highest R² values, it was considered the best model for describing polyphenol adsorption on D101 resin at the tested temperatures, further indicating predominantly monolayer adsorption.

3.4. Dynamic Adsorption and Desorption Curves

As shown in Figure 3(a), no absorption peak was observed in the initial eluate, indicating that no raisin polyphenols were detected. As the loading volume increased, the absorbance gradually rose and the resin column progressively approached saturation. With further loading, raisin polyphenols began to break through and appeared in the eluate. This result indicates that an excessively high loading flow rate shortens the contact time between the polyphenol solution and the resin particles, decreases adsorption efficiency, and causes premature breakthrough. Conversely, an excessively low loading flow rate prolongs the adsorption process and reduces production efficiency. Considering both adsorption performance and practical throughput, a loading flow rate of 2 BV/h was selected as the optimal condition, at which the breakthrough volume of the extract reached 4.7 BV.

As shown in Figure 3(b), under dynamic conditions, the raisin polyphenols adsorbed on D101 macroporous resin were readily eluted. Approximately 40 mL of 90% ethanol was sufficient to elute most of the polyphenols from the D101 resin, yielding a relatively concentrated elution peak.

3.5. Identification of Polyphenols in Dried Apricots by UHPLC-QTOF-MS/MS

The aim of this study was to identify the major polyphenolic components in raisin polyphenols using the UHPLC-QTOF-MS/MS method. The composition of the raisin polyphenols was analysed by UHPLC-QTOF-MS/MS in negative ion mode, and the total ion chromatogram (TIC) is shown in Figure S4 As shown in Table S3, a total of 13 phenolic compounds were identified, comprising 3 flavonoids, 8 phenolic acids, 1 hydroxycinnamic acid and 1 phenolic aldehyde derivative. Among these, quercetin had the highest content, accounting for 26.16% of the total polyphenol content; followed by isoquercitrin and ferulic acid, accounting for 20.48% and 13.15% of the total polyphenol content, respectively. Based on the above content analysis, this study selected these three polyphenolic compounds with the highest proportions for subsequent experimental research.

3.6. Evaluation of the Antioxidant Capacity of Raisin Polyphenols

The antioxidant capacity of raisin polyphenols before and after purification was evaluated by measuring their scavenging activities against hydroxyl radicals (·OH), DPPH·, and ABTS·⁺ radicals[34]. As shown in Figure 4, the scavenging activities against all three radicals increased markedly in a concentration-dependent manner. After purification, the IC₅₀ values decreased significantly, from 247.27 ± 11.06 to 212.80 ± 12.30 μg/mL for ·OH, from 6.07 ± 0.09 to 3.70 ± 0.08 μg/mL for DPPH·, and from 47.05 ± 0.38 to 22.68 ± 0.16 μg/mL for ABTS·⁺. These results demonstrate that purification substantially enhanced the antioxidant activity of raisin polyphenols.

3.7. Inhibitory Activity of Raisin Polyphenols against α-Amylase

α-Amylase is a key digestive enzyme that hydrolyzes starch into oligosaccharides, which are subsequently converted to glucose and rapidly absorbed[35]. By slowing starch digestion and glucose release, α-amylase inhibitors can help attenuate postprandial hyperglycemia[36]. Therefore, inhibition of α-amylase is widely regarded as an effective strategy for diabetes management. As shown in Figure 5 and Table 3, concentration-dependent inhibition of α-amylase was observed for crude and purified raisin polyphenols as well as the three representative monomeric compounds. Among them, PRP exhibited the strongest inhibitory activity, with an IC₅₀ of 3.15 ± 0.16 mg/mL, followed by quercetin, ferulic acid, isoquercetin, and crude polyphenols, with IC₅₀ values of 4.44 ± 0.21, 6.02 ± 0.21, 11.20 ± 0.56, and 24.55 ± 0.23 mg/mL, respectively. The markedly stronger inhibition by PRP than by crude polyphenols suggests that purification enriched the active components and reduced interference from impurities such as sugars and proteins. The stronger activity of PRP relative to any single monomer may also reflect synergistic interactions among multiple polyphenols. In addition, quercetin inhibited PPA more strongly than ferulic acid and isoquercetin, which may be related to its greater number of phenolic hydroxyl groups and stronger interactions with the enzyme.

3.8. Kinetic Analysis of Enzyme Inhibition

The inhibition types and kinetic parameters of polyphenols toward α-amylase were determined using Lineweaver-Burk double-reciprocal plots to further clarify the inhibitory mechanism. As shown in Figure 6(a), the PRP lines intersected in the third quadrant, and both V_max and K_m decreased as PRP concentration increased, indicating mixed-type inhibition. Similarly, the Lineweaver-Burk plot for ferulic acid (Figure 6b) showed intersecting lines in the third quadrant, with reduced V_max and increased K_m, also suggesting mixed-type inhibition. These results indicate that both PRP and ferulic acid can bind to the free enzyme (E) and the enzyme-substrate complex (ES), forming inactive EI or ESI complexes. For quercetin (Figure 6c), Vmax remained essentially unchanged whereas Km increased, indicating a predominantly competitive inhibition pattern. Isoquercetin (Figure 6d) showed lines intersecting on the y-axis, with increased K_m and constant V_max, which is also characteristic of competitive inhibition, indicating that isoquercetin binds mainly to the free enzyme (E) to form an EI complex rather than interacting appreciably with ES. Overall, PRP exhibited mixed inhibition, which may be related to the mixed-type contribution of ferulic acid and the competitive contribution of quercetin and isoquercetin[7,37,38,39].

3.9. Mechanisms of Fluorescence Quenching

Binding of small molecules can alter the microenvironment of protein chromophores, and the resulting fluorescence quenching phenomenon is widely used to evaluate polyphenol-protein interactions[40,41]. Aromatic amino acid residues, especially tyrosine and tryptophan, are the major contributors to the intrinsic fluorescence of α-amylase[42]. As shown in Figure 7(a-d), PPA exhibited a strong fluorescence emission peak at 335 nm, and the addition of polyphenols caused a concentration-dependent decrease in fluorescence intensity, consistent with previous reports. After addition of PRP, quercetin, or isoquercetin[40], the emission maximum shifted to longer wavelengths, from 335 nm to 340, 344, and 354 nm, respectively. This red shift indicates that these ligands altered the microenvironment around aromatic residues, increased local polarity, and decreased hydrophobicity. In contrast, after addition of ferulic acid, the fluorescence peak shifted slightly to 333 nm, suggesting a more hydrophobic local environment around Trp residues, possibly because hydrogen bonding or hydrophobic interactions promoted burial of hydrophobic groups within the protein. Overall, the fluorescence changes indicate that polyphenol binding induced conformational rearrangement in PPA. Among the tested monomers, quercetin caused the largest decrease in PPA fluorescence intensity, which is consistent with its strong in vitro inhibitory activity.

Fluorescence quenching mechanisms are generally classified as static quenching, which involves formation of a quencher-fluorophore complex; dynamic quenching, which arises from collisions between the quencher and the fluorophore; or a combination of both[43]. Static quenching indicates that polyphenols form ground-state, nonfluorescent complexes with PPA, reflecting stable binding[44]. To further clarify the mechanism of polyphenol-induced PPA fluorescence quenching, the fluorescence data were analyzed using the Stern-Volmer equation. As shown in Figure 7(e), the F₀/F curves versus [Q] for PRP and the three monomeric phenols showed good linearity, indicating a single dominant quenching mechanism. The quenching constants (k_q) for PRP, ferulic acid, isoquercetin, and quercetin were 5.89, 4.62, 3.27, and 2.08 (×10¹² L mol^-1·s^-1), respectively, all far above the maximum diffusion-controlled collisional quenching constant (2×10¹⁰ L mol^-1·s^-1). Therefore, the quenching of PPA by all four polyphenols can be assigned to a static mechanism. The binding constants (K_a) were 21.99, 18.04, 17.50, and 16.95 (mL mg^-1), respectively. These data indicate that PRP showed the highest apparent binding affinity overall, while ferulic acid exhibited the highest apparent K_a among the three monomers. In addition, the n values for all four polyphenols were close to 1, indicating one major binding site on PPA[45].

3.10. CD Analysis

Conformational changes in the secondary structure of PPA were investigated by circular dichroism spectroscopy, as shown in Figure 8 and Table 4. PPA exhibited two characteristic negative peaks at 208 and 226 nm, which are typical of an α-helix structure and correspond to the π-π* and n-π* transitions, respectively[46]. The addition of polyphenols caused clear changes in these signals, indicating alterations in the secondary structure of PPA. As shown in Table 4, after binding with quercetin, isoquercetin, ferulic acid, and PRP, the α-helix content changed from 30.7% to 32.5%, 18.0%, 35.7%, and 25.8%, respectively. The β-sheet content changed from 20.5% to 14.8%, 31.2%, 15.9%, and 28.4%, respectively, whereas the β-turn content changed from 17.3% to 25.3%, 19.0%, 16.8%, and 18.5%, respectively. The random coil content also changed from 31.5% to 27.4%, 31.8%, 31.6%, and 27.3%, respectively. These results indicate that polyphenol binding induced a clear rearrangement of PPA secondary structure, which could disrupt the structural environment required for efficient substrate binding and catalysis.

3.11. Molecular Docking Simulation

To investigate the binding sites, affinities, and interaction types of different polyphenols with PPA, binding-mode analyses were performed for the simulated conformations of the PPA-ferulic acid, PPA-quercetin, and PPA-isoquercetin systems, as shown in Figure 9. As shown in Figure 9(a-b), ferulic acid was stably accommodated in a PPA cavity formed by His299, Trp58, Leu165, Tyr62, and Trp59, where it established both hydrogen-bonding and hydrophobic interactions. Specifically, ferulic acid formed one hydrogen bond with His299 and hydrophobic interactions with several surrounding residues, which enhanced ligand-protein binding. As shown in Figure 9(c-d), quercetin was stably bound within a cavity composed of His305, Asp356, Trp59, Gln63, Leu165, Tyr62, Asp300, His299, Asp197, and Arg195, engaging in both hydrogen bonds and hydrophobic interactions. As shown in Figure 9(e-f), isoquercetin occupied a cavity comprising Gly306, Tyr151, Leu165, Ile148, His201, Leu162, Val163, Val157, Asp197, Ala198, His305, and Asp300, also through hydrogen-bonding and hydrophobic interactions. Isoquercetin formed four hydrogen bonds with Asp197, Ala198, His305, and Asp300, along with hydrophobic contacts with eight surrounding residues, which enhanced its binding to PPA. Notably, among the three monomers, quercetin showed the richest network of hydrogen-bonding and hydrophobic interactions, indicating a more compact binding mode with PPA. This result is consistent with the in vitro inhibition assays.

3.12. MD Simulations

Figure 10(a) illustrates the time-dependent evolution of RMSD values for the protein systems during the simulation[47]. The RMSD values of the PPA-ferulic acid, PPA-quercetin, and PPA-isoquercetin complexes remained low and gradually stabilized over the 100 ns trajectory, with average values of 0.119 ± 0.008, 0.133 ± 0.010, and 0.159 ± 0.021 nm, respectively. These small fluctuations indicate that all three systems maintained good structural stability during the simulation[43]. As shown in Figure 10(b), the Rg values displayed only slight fluctuations, with average values of 2.309 ± 0.006, 2.321 ± 0.006, and 2.332 ± 0.007 nm for the PPA-ferulic acid, PPA-quercetin, and PPA-isoquercetin complexes, respectively. The minimal variation in Rg indicates that the overall compactness of the PPA-polyphenol complexes was maintained throughout the simulation.

As shown in Figure 10(c), the SASA values of the systems remained stable throughout the simulation, with average values of 191.323 ± 2.816, 191.379 ± 2.388, and 195.837 ± 2.230 nm² for the PPA-ferulic acid, PPA-quercetin, and PPA-isoquercetin complexes, respectively. The stability of SASA indicates that PPA reached a relatively steady conformational state after ligand binding and that polyphenol binding did not induce large overall structural fluctuations.

Root mean square fluctuation (RMSF) reflects the flexibility of amino acid residues in a protein. The RMSF profiles of all residues in the three systems are shown in Figure 10(d). The average RMSF values for the PPA-ferulic acid and PPA-isoquercetin systems were generally lower than that for the PPA-quercetin system, indicating that these two ligands more strongly restricted local residue fluctuations and contributed to local structural stabilization of PPA.

As shown in Figure 10(e), the average numbers of hydrogen bonds formed between polyphenols and PPA throughout the simulation were 0.488, 1.895, and 4.416, respectively. These persistent hydrogen-bonding interactions indicate that the ligands maintained stable contacts with PPA throughout the trajectory.

Furthermore, the time-dependent variation in the distance between the centers of mass (DCOM) of the protein and the ligand was calculated to evaluate relative ligand localization and microscopic motion. As shown in Figure 10(f), the average DCOM values ± standard deviations for the PPA-ferulic acid, PPA-quercetin, and PPA-isoquercetin systems were 2.687 ± 0.023, 2.629 ± 0.090, and 1.654 ± 0.039 nm, respectively. Overall, the isoquercetin system showed the smallest average distance, indicating that its center of mass remained closest to that of the protein. The ferulic acid system showed the smallest fluctuation, indicating the most stable positional localization during the simulation. The quercetin system had an intermediate average distance but the largest standard deviation, implying greater translational or conformational mobility.

The free energy landscape (FEL) characterizes the free-energy changes of PPA during the simulation. The FEL surfaces of the three complex systems are shown in Figure 10(g-i), where blue indicates the energy minima. Comparison of the three FEL surfaces showed nearly single and smooth energy clusters, without multiple rough minima or large regions of scattered unstable states, indicating that the conformational and energetic states of the complexes were relatively stable.

The binding free energies between PPA and polyphenols were calculated using the MM/PBSA method over the 80-100 ns interval, during which the RMSD values were relatively stable. The results are summarized in Table S3. The total binding free energies (ΔG_bind) for the three complexes were -49.927, -61.586, and -47.729 kJ/mol, respectively. Among them, the PPA-quercetin system exhibited the most favorable overall binding free energy, consistent with the molecular docking results and the in vitro inhibition assays. Further energy-decomposition analysis showed that the non-polar interaction energies (ΔG_vdw + ΔG_np) were -123.021 kJ/mol for ferulic acid, -60.288 kJ/mol for quercetin, and -124.064 kJ/mol for isoquercetin, indicating that non-polar interactions favored ligand binding in all three systems. In contrast, the electrostatic interaction energies (ΔG_ele + ΔG_PB) were 68.998, 32.854, and 76.335 kJ/mol, respectively, all positive, indicating unfavorable net electrostatic contributions to binding in these systems.

Overall, non-polar interactions, mainly van der Waals forces and non-polar solvation free energy, were the primary driving forces for the binding of ferulic acid, quercetin, and isoquercetin to PPA. However, total binding stability depended on the balance among all energy terms. Although ferulic acid and isoquercetin showed larger favorable non-polar contributions, quercetin exhibited the most favorable overall ΔG_bind, indicating the strongest net binding affinity. This conclusion is consistent with the docking results and the in vitro inhibition data.

4. Conclusions

This study demonstrates that PRP are promising PPA inhibitors that can effectively suppress enzymatic activity. Using multispectral techniques combined with computational simulations, we systematically investigated the inhibitory mechanism of PRP toward PPA. The results indicate that PRP acts as a mixed-type inhibitor, whereas its major monomeric components display distinct inhibition behaviors: ferulic acid shows mixed-type inhibition, while quercetin and isoquercetin show predominantly competitive inhibition. Spectral analyses showed that these polyphenols bound to PPA, altered the microenvironment of aromatic residues, and induced secondary-structure rearrangement, thereby reducing catalytic activity. Molecular docking and molecular dynamics simulations further confirmed that hydrogen bonding and hydrophobic interactions dominate the stable binding of polyphenols to PPA, and MM/PBSA analysis indicated that quercetin had the most favorable overall binding free energy among the tested monomers. In addition, D101 macroporous resin enabled efficient enrichment of raisin polyphenols and markedly improved their bioactivity. These findings clarify the structure-activity relationship of raisin polyphenols as natural α-amylase inhibitors, enrich the theoretical basis for plant polyphenols in regulating carbohydrate metabolism, and support the development of raisin-derived hypoglycemic functional ingredients.