Mechanistic insights into Grifola frondosa-driven fermentation of Rice-wheat bran: Metabolomic profiling, molecular dynamics, and enhanced antioxidant efficacy

TL;DR

This study shows that fermenting rice and wheat bran with Grifola frondosa boosts bioactive compounds and antioxidant effects, making it a promising functional food ingredient.

Contribution

The study introduces a novel fermentation method using Grifola frondosa to enhance the bioaccessibility and antioxidant properties of rice-wheat bran.

Findings

Fermentation increased polysaccharide, polyphenol, and protein contents significantly after 13 days.

Metabolomics identified 445 differentially expressed metabolites, with ferulic acid methyl ester highly upregulated.

FRBWB showed improved antioxidant activity and reduced oxidative stress in zebrafish embryos.

Abstract

Rice bran (RB) and wheat bran (WB) are rich in bioactives but poorly utilized due to limited bioaccessibility. Here, a fungal fermentation strategy was developed to valorize RB-WB blends. A novel fermented product (FRBWB) was obtained via Grifola frondosa (G. frondosa) fermentation, with conditions optimized for bioactive enrichment. After 13 days of fermentation, polysaccharide, polyphenol, and soluble protein contents increased from 86.56 to 126.93 mg/g, 1.38 to 1.67 mg/g, and 53.69 to 93.06 mg/g, respectively (P < 0.05), accompanied by marked improvements in water-holding capacity and solubility. The process altered substrate's microstructure, resulting in a looser, more porous surface. Untargeted metabolomics identified 445 differentially expressed metabolites (DEMs) between FRBWB and the unfermented control (RBWB), among which ferulic acid methyl ester showed the highest upregulation. Molecular docking and molecular dynamics simulations suggested a stable interaction between ferulic acid methyl ester and lignin peroxidase, dominated by van der Waals forces, electrostatic interactions, and hydrogen bonding (ΔEMMPBSA = −79.73 ± 1.47 kJ/mol), which may help rationalize the observed enhancement of phenolic release. Functionally, FRBWB exhibited significantly enhanced in vitro antioxidant activity and conferred superior protection against AAPH-induced oxidative stress in zebrafish embryos, improving survival (92% at 50 μg/mL), hatching rate (87.69%), and reducing reactive oxygen species (ROS), apoptosis, and lipid peroxidation (P < 0.05). Overall, this study suggests that G. frondosa fermentation represents a sustainable strategy for upgrading RB–WB by-products into functional food ingredients with potential antioxidant benefits.

•A novel rice-wheat bran fermented product was developed by Grifola frondosa.•A 13-day fermentation enhanced bioactive contents, functional properties, and formed a porous low-crystallinity structure.•Metabolomics uncovered 445 differential metabolites with ferulic acid methyl ester highly upregulated.•Molecular dynamics revealed a stable binding of ferulic acid methyl ester to lignin peroxidase.•The product alleviated oxidative stress in zebrafish through multiple antioxidant pathways.

A novel rice-wheat bran fermented product was developed by Grifola frondosa.

A 13-day fermentation enhanced bioactive contents, functional properties, and formed a porous low-crystallinity structure.

Metabolomics uncovered 445 differential metabolites with ferulic acid methyl ester highly upregulated.

Molecular dynamics revealed a stable binding of ferulic acid methyl ester to lignin peroxidase.

The product alleviated oxidative stress in zebrafish through multiple antioxidant pathways.

Figures

Graphical abstract.Fig. 1

Dynamic changes in the content, physicochemical properties and structure of rice bran (RB) and wheat bran (WB) complex during fermentation by *G. frondosa*. (A) Polysaccharide content. (B) Polysaccharide content. (C) Soluble protein content. (D) Water-holding capacity (WHC). (E) Oil-holding capacity (OHC). (F) Solubility (S). (G) Swelling power (SP). (H) Scanning electron microscopy (SEM) images: a1 and a2 show samples at 0 days of fermentation at magnifications of 2000× and 5000×, respectively; b1 and b2 show samples after 13 days of fermentation at magnifications of 2000× and 5000×, respecti

Metabolite profiling of fermented and unfermented polyphenolic extracts analyzed by ultra-performance liquid chromatography–tandem mass spectrometry (UPLC–MS/MS). (A) Principal component analysis (PCA). (B) Orthogonal partial least squares discriminant analysis (OPLS-DA) score plot. (C) OPLS-DA validation plot. (D) Volcano plot showing differentially expressed metabolites (DEMs) between fermented rice bran (RB) and wheat bran (WB) product (FRBWB, n = 3) and unfermented RB and WB product (RBWB, n = 3). Red points represent significantly upregulated DEMs (log_2_fold change (FC) > 1), green point

Docking simulations between seven key differentially expressed metabolites (DEMs) and three *G. frondosa* enzymes. Compounds 1–7 are (1) ferulic acid methyl ester, (2) p-coumaric acid ethyl ester, (3) caffeate, (4) apigenin, (5) quercetin-3-O-glucoside, (6) chlorogenate, and (7) luteolin 7-O-(6″-malonylglucoside)*, each docked with (A–G) laccase, (H—N) manganese peroxidase, and (O—U) lignin peroxidase.Fig. 4

Molecular dynamics simulations of the lignin peroxidase–ferulic acid methyl ester complex. (A) Root mean square deviation (RMSD). (B) Radius of gyration (Rg). (C) Root mean square fluctuation (RMSF). (D) Distance between the ligand and catalytic site. (E) Buried solvent-accessible surface area (Buried SASA). (F) Overlay of sampled conformations. (G) Total van der Waals (VDW)/electrostatic (ELE) binding energies. (H) Residue-based energy contributions. (I) Hydrogen bond evolution. (J) Docked complex showing key interaction types.Fig. 5

Effect of *G. frondosa* fermentation on glycosidic differentially expressed metabolites (DEMs) in rice bran (RB) and wheat bran (WB) complex. (A) Proportion of glycosidic DEMs within the overall DEM pool. (B) Primary and secondary classifications of glycosidic DEMs. (C) Tertiary classification of glycosidic DEMs. (D) Variation trends of glycosidic DEMs. (E) Boxplot visualization of key glycosidic DEMs. FRBWB: fermented RB and WB product. RBWB: unfermented RB and WB product.Fig. 6

Full text

Introduction

1

As agricultural by-products, rice bran (RB) and wheat bran (WB) are rich in bioactive compounds. RB contains abundant phenolics, γ-oryzanol, and vitamin E,with antioxidant, anticancer, and antidiabetic activities (Kodape et al., 2025; Tan et al., 2023), whereas WB is characterized by high levels of dietary fiber, protein, and bioactives such as arabinoxylan and ferulic acid (Cheng et al., 2022; Onipe et al., 2021). However, their high-value utilization remains limited because many bioactive constituents are tightly embedded within complex cell wall matrices composed of cellulose, hemicellulose, lignin, and protein networks, resulting in low bioaccessibility. In addition, oxidation susceptibility, physicochemical instability, and coarse texture further restrict their application in food systems (De Bondt et al., 2020; Ronie et al., 2025; Spaggiari et al., 2021; Wang et al., 2024). As a result, RB and WB are often underutilized as low-value feed materials or discarded. Although various physical, chemical, and enzymatic approaches have been explored, these methods frequently suffer from limited efficiency, high energy demand, or incomplete release of bound bioactives. Therefore, developing efficient and eco-friendly bioprocessing strategies that can simultaneously disrupt structural barriers and enhance functional properties is essential for upgrading RB and WB into value-added nutritional or functional ingredients.

Previous studies have explored stabilization and selective extraction strategies to improve the utilization of RB, including solvent and enzymatic approaches aimed at preserving nutritional quality and extending shelf life (Das et al., 2025; Liu et al., 2019). Fermentation has proven particularly effective in enhancing the release of phenolics and sugars from WB, improving protein solubility and digestibility, and refining flavor and texture (Arte et al., 2015; Nemes et al., 2025). Co-fermentation with yeast and lactic acid bacteria increases WB bioactivity and aroma, while fermentation-induced changes in volatile profiles enhance the sensory quality of RB (Tang et al., 2024; Wang et al., 2025). Recent fermentation studies further indicate that fermentation outcomes strongly depend on substrate composition, microbial selection, and process parameters, as demonstrated in blended fruit fermentations, microbial strain screening, and agricultural by-product valorization (Choi et al., 2025; Kiran et al., 2025; Luo et al., 2024), while untargeted metabolomics has been applied to resolve fermentation-driven changes in phenolic and antioxidant profiles (Lang et al., 2025). Moreover, fermentation can significantly increase total phenolic content and antioxidant capacity, which is essential for the development of high-value functional foods (Chen, Zhang, et al., 2025). Accordingly, starter selection is critical for achieving both efficient cell wall disruption and antioxidant enrichment. Grifola frondosa (G. frondosa) is of particular interest because it produces antioxidant-related metabolites, including phenolic acids, ergothioneine, flavonoids, and glutathione, and secretes ligninolytic enzymes (Motoda et al., 2025) such as laccase and peroxidases (Fernandes et al., 2011) during fermentation, facilitating structural modification of lignocellulosic matrices (Jang et al., 2025), consistent with genome-level evidence for lignin depolymerization capacity in white-rot fungi (Lodi et al., 2025). Besides, G. frondosa synthesizes bioactive polysaccharides with immunomodulatory activities (Jiang et al., 2022; Yang et al., 2007). However, systematic investigations linking G. frondosa–driven structural remodeling with metabolite transformation and antioxidant functionality in mixed bran substrates remain limited.

Unlike previous studies that employed Aspergillus (Shin et al., 2019), mushrooms (Xu et al., 2018), or Lentinula (Li et al., 2024) to enhance phenolic content and bulk antioxidant capacity in cereal bran, this study proposes a G. frondosa–based co-fermentation strategy to valorize a RB and WB complex by systematically linking fermentation time with structural remodeling and functional outcomes. To achieve this, comprehensive physicochemical characterization, polysaccharide structural analysis, untargeted metabolomics, enzyme activity profiling, and molecular docking were integrated to elucidate how fermentation-driven enzymatic processes regulate bioactive release and antioxidant performance. Furthermore, both in vitro assays and zebrafish-based in vivo models were employed to evaluate the biological relevance of these fermentation-induced changes. Collectively, this work suggests the potential of agricultural by-products such as RB and WB to be transformed into high-value antioxidant resources, offering a theoretical and practical foundation for their further development and utilization (Fig. 1).Fig. 1. Graphical abstract.Fig. 1

Materials and methods

2

Fungal strain reactivation and fermentation substrate preparation

2.1

G. frondosa (strain no. 14078) was obtained from the China Center of Industrial Culture Collection (CICC). RB and WB were purchased from a local market in Hohhot, China (40°48′ N, 111°39′ E), air-dried under shade to reduce moisture, then ground and passed through a 60-mesh sieve (approximately 250–380 μm), and subsequently stored in a dry environment until use. The G. frondosa was reactivated on potato dextrose agar (PDA) at 30 °C, after which actively growing mycelia were transferred to potato dextrose broth (PDB) and cultured at 30 °C for 5 days to obtain pelletized inoculum.

Fermentation substrates were prepared by blending WB (47.6%), RB (47.6%), glucose (4%), KH_2_PO_4_ (0.15%), MgSO_4_ (0.15%), and peptone (0.5%) with distilled water (80% moisture) and adjusting the pH to 5.5. The substrate was inoculated with 10% (w/w) G. frondosa inoculum mycelial pellets and incubated at 28 °C and 80% relative humidity to produce the fermented product (FRBWB). Multiple fermentation times (0, 3, 5, 7, 9, 11, 13, and 15 days) were evaluated in this study. In order to maintain experimental consistency, for each fermentation interval, an unfermented control was maintained under the same conditions but without inoculum, yielding the unfermented product (RBWB). After fermentation, all samples were reground, passed through a 60-mesh sieve, and stored for further analysis. All primary reagents, instruments, and equipment employed in this study were listed in Supplementary material 1, and all chemicals were of analytical grade unless otherwise specified.

Determination of principal bioactive components and physicochemicall–structural analyses

2.2

We first quantified the principal bioactive components—namely polysaccharides, polyphenols, and soluble proteins—in FRBWB at different time points (0, 3, 5, 7, 9, 11, 13, and 15 days). In brief, following our previous protocols, polysaccharide content was measured by the phenol–sulfuric acid method (An et al., 2024), total polyphenol content was determined leveraging Folin–Ciocalteu assay (Liu et al., 2022), and soluble protein content was assessed by Coomassie brilliant blue method (Bradford, 1976). To further characterize the physicochemical properties, we evaluated their water-holding capacity (WHC), oil-holding capacity (OHC), solubility (S), and swelling power (SP). WHC and OHC were determined according to previously reported methods with minor modifications(Chau & Cheung, 1998). Specifically, 0.5 g of sample was placed in a 10 mL tube and weighed (m1), mixed with 5 mL distilled water, equilibrated 30 min, and centrifuged at 4000 rpm for 15 min. The supernatant was discarded, and the tube was inverted on filter paper for 10 min before reweighing (m2). WHC (g/g) was calculated as (m2 − m1)/0.5. For OHC, 0.5 g of sample was mixed with 10 mL soybean oil in a 15 mL tube, incubated at 25 °C for 90 min, and centrifuged at 8000 rpm for 30 min; after discarding the supernatant, OHC (g oil/g) was determined as [(W1 − W0)/(0.5)] × 100%, where W1 and W0 represent the final and initial weights, respectively. Both S and SP were measured with reference to a previously published method (Li, Gu, et al., 2023), with minor modifications. Briefly, 0.2 g sample in 20 mL water was heated at 80 °C for 30 min (shaking every 5 min), cooled, and centrifuged at 3000 rpm for 15 min. The supernatant was transferred to a pre-weighed aluminum dish, dried at 105 °C for 2–3 h, and weighed (m1); the wet precipitate in the tube was weighed (m2). S (%) was calculated as (m1/0.2) × 100%, while SP (g/g) was computed as [m2/(0.2 × (1 − S/100))]. Additionally, structural analyses including scanning electron microscopy (SEM), X-ray diffraction (XRD), and thermogravimetric analysis (TGA) on 0-day and 13-day samples. SEM (Li, Wang, et al., 2023) was conducted using a Hitachi SU8010 scanning electron microscope (Hitachi High-Tech Corporation, Tokyo, Japan) operated at an accelerating voltage of 10 kV and a magnification of 500×. XRD (Cen et al., 2024) were recorded on a Rigaku Ultima IV X-ray diffractometer (Rigaku Corporation, Tokyo, Japan) operated at 45 kV and 40 mA, with a scanning range of 5° to 40° (2θ), a step size of 0.013°, and a dwell time of 0.3 s per step. TGA was carried out using a TA Instruments TGA Q500 analyzer (TA Instruments, New Castle, DE, USA) with a heating rate of 10 °C/min, from 30 to 800 °C under a nitrogen atmosphere (Lanjewar et al., 2025).

Particularly, based on the maximum levels of polysaccharides and polyphenols, as well as optimal WHC, OHC, S, and SP values at various fermentation times (0, 3, 5, 7, 9, 11, 13, and 15 days), we identified the most suitable fermentation duration for producing the final product.

Polysaccharide extraction, monosaccharide analysis, and structural characterization

2.3

According to the previously determined optimal fermentation time, fermented and unfermented products were prepared. For extraction, 1 g of each sample was mixed with 16 mL distilled water, heated at 80 °C for 30 min, cooled, and centrifuged at 4000 rpm for 10 min. The supernatants were collected, concentrated by rotary evaporation at 60 °C, oven-dried, and precipitated with four volumes of 95% ethanol at 4 °C overnight. The precipitates were recovered by centrifugation, freeze-dried, and stored at −20 °C as crude polysaccharide fractions for subsequent analyses. Monosaccharide composition was analyzed by 1-phenyl-3-methyl-5-pyrazolone (PMP) pre-column derivatization followed by high-performance liquid chromatography (HPLC) (Boonla et al., 2015; Chen et al., 2020). Briefly, crude polysaccharides (10 g) were hydrolyzed with 2 M trifluoroacetic acid at 110 °C for 5 h under nitrogen. The hydrolysates were derivatized with PMP under alkaline conditions at 70 °C for 60 min, acidified, and extracted with chloroform. The aqueous phase was filtered (0.45 μm) and analyzed using an Agilent 1200 HPLC system (Agilent Technologies, Santa Clara, CA, USA). For further structural characterization, dried polysaccharide samples were analyzed by Fourier-transform infrared (FT-IR) spectroscopy using KBr pellets on a Nicolet iS10 FT-IR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA), and their molecular weight distribution was determined by gel permeation chromatography (GPC) using a Waters Alliance e2695 system equipped with a refractive index detector (Waters Corporation, Milford, MA, USA), following established methods (Chen et al., 2020; Wu et al., 2019).

Extraction of polyphenolic compounds and amino acid profiling

2.4

In a manner similar to crude polysaccharide extraction protocol, both fermented and unfermented water extracts were concentrated by rotary evaporator at 60 °C and subsequently oven-dried. The dried concentrates were mixed with four volumes of 95% ethanol and incubated at 4 °C overnight. After centrifugation, the supernatants were collected, lyophilized, and designated as fermented and unfermented polyphenolic extracts, which were stored at −20 °C until analysis. Polyphenolic profiles were determined by ultra-performance liquid chromatography–tandem mass spectrometry (UPLC–MS/MS) system (Chen et al., 2020; Wu et al., 2019). Separation was achieved on an Agilent SB-C18 column (2.1 × 100 mm, 1.8 μm) using water (A) and acetonitrile (B), both containing 0.1% formic acid, with a gradient of 5–95–5% B over 34 min. The column temperature was 40 °C, flow rate 0.35 mL/min, and injection volume 4 μL. MS detection was performed on an AB Sciex QTRAP 6500+ mass spectrometer (AB Sciex, Framingham, MA, USA) equipped with an electrospray ionization (ESI) source operated in both positive and negative ion modes, using multiple reaction monitoring (MRM). Amino acid composition was determined using L-8900 automatic amino acid analyzer (Hitachi High-Tech, Tokyo, Japan). Samples were hydrolyzed with 8 M HCl under nitrogen at 110 °C for 24 h, evaporated to dryness, redissolved in 0.01 M HCl, and analyzed against 17 amino acid standards. Chromatographic separation was conducted on an AminoPac PA-10 column using double-distilled water, 0.25 M NaOH, and 1.0 M sodium acetate as the mobile phase.

Molecular docking and molecular dynamics simulation

2.5

With the intention of elucidating the binding modes and putative mechanisms by which the selected compounds influence the key enzymes present in G. frondosa fermentation, both molecular docking and molecular dynamics simulation were performed. Adapting previously established protocols (Cao et al., 2025; Li, Zeng, et al., 2025), we proceeded as follows. At the outset, the amino acid sequence of target enzymes was retrieved from UniProt database (https://www.uniprot.org/) and its highest-resolution crystal structure obtained from PDB database (https://www.rcsb.org/) and preprocessed in PyMOL (v 3.1.1) by removing crystallographic water molecules and non-essential ligands and adding polar hydrogens (Seeliger & de Groot, 2010). The corresponding ligand structures for the selected differentially expressed metabolites (DEMs) were retrieved from PubChem database (https://pubchem.ncbi.nlm.nih.gov/) and converted to the pdbqt format after adding Gasteiger charges through MGLTools (v 1.5.7) (Arcon et al., 2019). Docking grids were defined to encompass the entire active-site region of each enzyme. For each target, the grid box was centered on the catalytic pocket and was large enough to cover the binding cavity and surrounding residues, ensuring that all plausible ligand poses could be sampled. AutoDock Vina (v 1.5.7) (Trott & Olson, 2010) was run with an exhaustiveness value of 8 and a maximum of 20 output poses per ligand. A fixed random seed was used to ensure reproducibility. For each protein–ligand pair, docking poses were ranked by Vina binding affinity, and the lowest-energy conformation without severe steric clashes was selected for further analysis. Generally, lower binding energies indicate more stable ligand–receptor conformations, and any compound with a binding energy of ≤ −5.0 kcal/mol was deemed likely to exhibit appreciable binding activity. In addition, a hydrogen bond count of ≥2 suggested higher binding specificity. For molecular dynamics simulations, the ligand–receptor pose with the most favorable binding energy was selected as the representative system. Molecular dynamics simulation was conducted through GROMACS (v 2022). The AMBER14SB force field was used for the protein, the GAFF force field for the ligand, and the TIP3P model for water. The protein–ligand complex was placed in a periodic simulation box, solvated, and neutralized with counter-ions. Hydrogen bonds were constrained using the LINCS algorithm, a 2 fs integration time step was applied, long-range electrostatics were treated with the Particle Mesh Ewald (PME) method with a 1.2 nm cutoff, and non-bonded interactions were truncated at 10 Å and updated every 10 steps. System temperature (298 K) and pressure (1 bar) were maintained by the V-rescale thermostat and Berendsen barostat, respectively. After 100 ps of NVT and 100 ps of NPT equilibration, a 200 ns production run was carried out, saving snapshots every 10 ps. Trajectories were analyzed using VMD and PyMOL, and MM-PBSA binding free energies (ΔE_MMPBSA_) between ligand–receptor were calculated with the g_mmpbsa module.

Determination of lignin peroxidase, manganese peroxidase, and feruloyl esterase activities

2.6

Activities of lignin peroxidase, manganese peroxidase, and feruloyl esterase were determined in RBWB and FRBWB using commercial colorimetric assay kits (Jiangsu Aidisheng Biological Technology Co., Ltd., China). Briefly, about 0.1 g of sample was homogenized on ice in 1 mL of extraction buffer and centrifuged at 4 °C (12,000 rpm for 10 min). The supernatant was used for enzymatic measurements following the manufacturer's instructions. Lignin peroxidase activity was monitored at 310 nm (readings at 10 s and 5 min, 30 °C), manganese peroxidase at 470 nm (10 s and 10 min, 30 °C), and feruloyl esterase at 340 nm (immediately and after 30 min incubation at 40 °C).

Assessment of the antioxidant activity of both FRBWBCP and FRBWBCP

2.7

In vitro antioxidant assays

2.7.1

Following our previously developed protocol s(An et al., 2024; Liu et al., 2024) we evaluated the in vitro antioxidant capacity of both fermented and unfermented products. Specifically, three key assays were performed: (i) 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical scavenging, (ii) hydroxyl radical scavenging, and (iii) total reducing power. Butylated hydroxyanisole (BHA) and ascorbic acid (VC) were included as positive controls in all assays. In addition to enable quantitative comparison of antioxidant capacity among samples, dose–response curves were constructed for each assay, and the half maximal inhibitory concentration (IC_50_) values were calculated accordingly.

Zebrafish maintenance, embryo collection, and in vivo antioxidant assays

2.7.2

This experiment was approved by the Animal Ethics Committee of Inner Mongolia Agricultural University (NND2024121). AB wild-type zebrafish (purchased from the China Zebrafish Resource Center) were maintained in a recirculating aquaculture system at (28.0 ± 1) °C under a 14/10-h light/dark cycle. Fish were fed daily with commercial feed and newly hatched Artemia nauplii. Uneaten feed and waste were removed after feeding, and one-third of the tank water was replaced each day. Aeration was continuously provided, and water quality parameters were monitored regularly. Prior to the experiment, males and females (1:2 ratio) were placed in spawning chambers separated by a divider. After 10–12 h dark adaptation, fish were exposed to light to induce spawning; embryos were collected 60 min later, rinsed several times with distilled water, and transferred to 12-well plates for assay. Fermented and unfermented products were tested at 12.5, 25, 50, 100, and 200 μg/mL (eight replicates per concentration, 15–20 embryos per replicate). After 1 h exposure, 0.5 mol/L 2,2′-azobis(2-amidinopropane) dihydrochloride (AAPH) was added and embryos were incubated for an additional 24 h, following the protocol previously reported (An et al., 2024; An et al., 2025). Untreated embryos served as controls. Embryo mortality and hatching rates were recorded every 12 h up to 72 h. Each trial was performed in triplicate to ensure reproducibility.

Embryo locomotor activity was video-recorded at 24, 30, 36, 42, and 48 h post-fertilization (hpf) through a Hikvision (DS-2CD3T46WD-I3) (Hangzhou Hikvision Digital Technology Co., Ltd., China) industrial camera under a fixed settings (f/2.8, 400 μs exposure, 3648 × 2736 pixels). Both the camera and LED light source were positioned 38 cm above the sample to ensure consistent illumination. Each recording lasted approximately 3–5 min. Still frames extracted from the videos were analyzed with a preference for lateral-view images to improve morphological measurements. For enhancement, Gaussian blur filtering was applied in ImageJ. Beginning at 60 hpf, heart rates were measured over 1 min intervals, and lateral images were collected to determine yolk sac area and body length. Eight randomly selected fish per replicate were evaluated. At 72 hpf, larvae were incubated in the dark (28.0 ± 1 °C) with 20 μg/mL 2′,7′-dichlorofluorescin diacetate (DCF-DA) (1 h), 7 μg/mL acridine orange (AO) (0.5 h), or 25 μg/mL 1,3-bis(diphenylphosphino)propane (DPPP) (1 h), rinsed, anesthetized with 3-aminobenzoic acid ethyl ester methanesulfonate, and imaged by fluorescence microscopy (Olympus IX73 inverted fluorescence microscope, Olympus Corporation, Tokyo, Japan). Relative fluorescence intensity was normalized to the control group and used to evaluate intracellular reactive oxygen species (ROS) production, cell death, and lipid peroxidation. The rates of ROS production, cell death, and lipid peroxidation were calculated as follows: ROS (or cell death or lipid peroxidation) (%) = (Fluorescence intensity of treated group / Fluorescence intensity of control group) × 100%.

Statistical analysis

2.8

Unless otherwise stated, all experiments included three independent biological replicates (n = 3), with each replicate being measured in triplicate (technical replicates). The results are presented as mean ± standard deviation (SD). For comparisons among more than two groups (e.g., different fermentation times, treatment concentrations, or sample types), one-way analysis of variance (ANOVA) was performed by SAS (v 9.2, SAS Institute Inc., Cary, NC, USA), followed by Duncan's multiple range test for post hoc pairwise comparisons. Differences were considered statistically significant at P < 0.05. All figures were generated using OriginPro 2021C (OriginLab Corporation, Northampton, MA, USA). For the compositional analysis of active components, raw mass spectrometry data were processed using Analyst (v 1.6.3, AB Sciex, Framingham, MA, USA) to identify metabolites. Principal component analysis (PCA) was carried out in R software (v 4.2.2) implementing standard base package, and orthogonal partial least squares discriminant analysis (OPLS-DA) was performed with the OPLSR. Anal function in MetaboAnalystR to characterize group differences. DEMs were identified using Student's t-test (P < 0.05), combined with variable importance in projection (VIP) > 1 and |log_2_fold change (FC)| > 1. For in vitro antioxidant assays, dose–response curves were constructed using GraphPad Prism by plotting log10-transformed concentrations against normalized responses, followed by nonlinear regression analysis (dose–response inhibition model with variable slope) to calculate IC_50_ values. Image processing was performed with OpenCV (v 3.8) and Python (v 3.7.5), focusing on viability indicators in zebrafish embryos. Morphological parameters such as body length and yolk sac area were quantified in the RGB space for structural feature analysis. This approach enabled precise quantification of embryonic development parameters.

Results

3

G. frondosa fermentation for 13 days augments bioactive constituents and physicochemical properties in RB and WB complex

3.1

To determine the optimal fermentation duration, we monitored polysaccharides, polyphenols, and soluble proteins in FRBWB. The polysaccharide content exhibited dynamic changes throughout the fermentation process. Starting from an initial value of 86.56 mg/g at day 0, it dropped to 59.43 mg/g at 3 days. Subsequently, the content surged significantly, peaking at 126.93 mg/g at 13 days (P < 0.05), before falling once more to 59.07 mg/g at 15 days (P < 0.05) (Fig. 2A). Polyphenol content began at 1.38 mg/g (day 0), significantly decreased to a plateau of 1.00–1.10 mg/g between days 3 and 9, then rose to a peak of 1.67 mg/g at day 13—significantly exceeding all other time points (P < 0.05)—before falling to a final value of 0.99 mg/g at day 15 (P < 0.05) (Fig. 2B). The soluble protein content exhibited an overall upward trend with increased fermentation time. At day 13, soluble protein content increased to 93.06 mg/g from an initial value of 53.69 mg/g at day 0, which was significantly higher compared to all other time points (P < 0.05) (Fig. 2C). Collectively, 13 days of G. frondosa fermentation resulted in the highest measured levels of polysaccharides, polyphenols, and soluble proteins among all tested time points. We next evaluated how fermentation time affected key physicochemical properties. WHC increased from 1.32 g/g at day 0 to 2.12 g/g at day 13, and OHC rose from 2.06 to 2.93 g oil/g over the same period (P < 0.05) (Fig. 2D, E). Similarly, S and SP, indices of processability, exhibited consistent trends. Compared with 0 d, S rose from 11.61% to 22.03% at 13 d (P < 0.05) (Fig. 2F). Analogously, SP also reached its peak on day 13 (P < 0.05) (Fig. 2G). Thus, a 13-day fermentation simultaneously maximized polysaccharide, polyphenol, and soluble protein contents and produced the most favorable WHC, OHC, S, and SP, supporting 13 days as the optimal fermentation duration for RB and WB complex.Fig. 2. Dynamic changes in the content, physicochemical properties and structure of rice bran (RB) and wheat bran (WB) complex during fermentation by G. frondosa. (A) Polysaccharide content. (B) Polysaccharide content. (C) Soluble protein content. (D) Water-holding capacity (WHC). (E) Oil-holding capacity (OHC). (F) Solubility (S). (G) Swelling power (SP). (H) Scanning electron microscopy (SEM) images: a1 and a2 show samples at 0 days of fermentation at magnifications of 2000× and 5000×, respectively; b1 and b2 show samples after 13 days of fermentation at magnifications of 2000× and 5000×, respectively. (I) X-ray diffraction (XRD) patterns. (J) Thermogravimetric analysis (TGA) curves. (K) Monosaccharide composition: Glucose (Glc), Galactose (Gal), Xylose (Xyl), Arabinose (Ara), Mannose (Man), Rhamnose (Rha), Galacturonic acid (GalA), Glucuronic acid (GlcA), Glucosamine (GlcN), and Fucose (Fuc). (L) FT-IR spectra recorded within 4000–500 cm^−1^. (M) Changes in molecular weight distribution. FRBWB: fermented RB and WB product. RBWB: unfermented RB and WB product. Data are presented as mean ± standard deviation (SD) (n = 3). Different lowercase letters indicate significant differences among groups (P < 0.05), while the same letters indicate no significant difference (P > 0.05), as determined by one-way ANOVA with Duncan's multiple range test.Fig. 2

G. frondosa fermentation changes SEM, XRD, TGA, monosaccharide composition and structural characteristics in RB and WB complex

3.2

In order to clarify structural changes induced by G. frondosa fermentation, we examined RB and WB complex by SEM, XRD, and TGA. SEM images at 0 days revealed compact, block-like surfaces composed of small, irregular fragments without visible pores (Fig. 2H**, a1/a2**). Conversely, at 13 days, the substrate displayed a looser, sponge-like structure with smaller pores and rough, uneven surfaces (Fig. 2H**, b1/b2**). XRD profile revealed a crystalline region (13°–23° 2θ) and a noncrystalline region (30°–40° 2θ). After 13 days of fermentation, the characteristic cellulose I peak near 2θ = 21.6°, became markedly weaker, accompanied by a decrease in crystallinity from 22.77% to 11.24% (Fig. 2I). TGA demonstrated three major weight-loss stages between 30 and 800 °C for both 0-day and 13-day FRBWB (Fig. 2J). The 13-day product exhibited a slightly greater mass loss in the main degradation stage (200–400 °C), with values of 58.29% (0-day) and 56.19% (13-day). Overall mass losses were comparable [74.52% (0-day) vs. 74.86% (13-day)], confirming similar thermal stability despite compositional changes.

After removing proteins from FRBWB and RBWB via ethanol precipitation, we obtained crude polysaccharide fractions from fermented and unfermented RB and WB complex. Making use of HPLC, FT-IR, and GPC we investigated their monosaccharide composition, functional groups, and molecular weights. Notably, G. frondosa fermentation increased the proportion of mannose (Man) in RB and WB complex, resulting in 10 monosaccharides in FRBWB versus 9 in RBWB, with fucose (Fuc) detected exclusively in FRBWB (Fig. 2K). FT-IR spectra suggested typical polysaccharide features in both samples, evidenced by: (1) a broad band at 3500–3000 cm^−1^ corresponding to O—H stretching and hydrogen bonding (Tanthana & Chuang, 2010); (2) peaks at 2900–3000 cm^−1^ for –CH2/–CH3 stretching and bending (Plichta et al., 2020); (3) a strong band near 1652.65 cm^−1^ from C

<svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="20.666667pt" height="16.000000pt" viewBox="0 0 20.666667 16.000000" preserveAspectRatio="xMidYMid meet"><metadata> Created by potrace 1.16, written by Peter Selinger 2001-2019 </metadata><g transform="translate(1.000000,15.000000) scale(0.019444,-0.019444)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z M0 280 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z"/></g></svg>

O stretching in glycosidic bonds (Yang et al., 2021); (4) a sharp peak around 1400 cm^−1^ indicating the presence of uronic acids (Wang et al., 2019); and (5) intense absorption at 1150–1000 cm^−1^ for asymmetric vibrations of C–O–C/C–OH pyran rings (An et al., 2024). A distinct peak at 526 cm^−1^ suggested that the fermented samples contained α-glycosidic linkages (An et al., 2024) (Fig. 2L). Analysis of molecular weight distributions revealed a shift from low to mid-high molecular weight polysaccharides after fermentation. FRBWB showed a higher proportion of polysaccharides in the 10–30 kDa range (51%) compared with RBWB (22.55%), whereas RBWB was dominated by low-molecular-weight fragments (1–10 kDa, 76.59%) (Fig. 2M).

G. frondosa fermentation modulates the metabolic profile of RB and WB complex

3.3

Polyphenolic extracts from fermented (n = 3) and unfermented (n = 3) RB and WB complex were obtained via ethanol extraction and analyzed by UPLC–MS/MS. PCA and OPLS-DA showcased clear separation between FRBWB and RBWB (Fig. 3A-C), with high explained variance and predictive power (R^2^X = 0.939, R^2^Y = 1, Q^2^ = 0.999, P < 0.05), indicating a clear separation between fermented and unfermented samples. Overall, 445 DEMs were identified, with 253 upregulated and 202 downregulated between FRBWB and RBWB, spanning nine major classes (VIP > 1, P < 0.05, |log_2_FC| > 1) (Fig. 3D and Supplementary material 2). Flavonoids (141) and phenolic acids (69) were most abundant, followed by terpenoids, alkaloids, lignans/coumarins, and others (Fig. 3E). Afterwards, Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis showcased that the DEMs were significantly enriched in pathways related to the biosynthesis and metabolism of various secondary metabolites, vitamins, and amino acids (P < 0.05) (Fig. 3F).Fig. 3. Metabolite profiling of fermented and unfermented polyphenolic extracts analyzed by ultra-performance liquid chromatography–tandem mass spectrometry (UPLC–MS/MS). (A) Principal component analysis (PCA). (B) Orthogonal partial least squares discriminant analysis (OPLS-DA) score plot. (C) OPLS-DA validation plot. (D) Volcano plot showing differentially expressed metabolites (DEMs) between fermented rice bran (RB) and wheat bran (WB) product (FRBWB, n = 3) and unfermented RB and WB product (RBWB, n = 3). Red points represent significantly upregulated DEMs (log_2_fold change (FC) > 1), green points indicate significantly downregulated DEMs (log_2_FC < 1), and gray points denote non-significant changes (Student's t-test P > 0.05). (E) Scatter plot of DEMs. (F) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the identified DEMs. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)Fig. 3

G. frondosa fermentation drives the stable binding of ferulic acid methyl ester to lignin-degrading enzymes as revealed by molecular docking and dynamics

3.4

Recent studies indicate that G. frondosa produces lignin-degrading enzymes—such as laccase (Xing et al., 2006) and peroxidases (including manganese peroxidase and lignin peroxidase) (Švagelj et al., 2007)—during fermentation (Fernandes et al., 2011). To confirm their involvement in the release of bioactive components from the RB and WB complex, we measured the activities of lignin peroxidase, manganese peroxidase, and feruloyl esterase in both RBWB and FRBWB. All three activities showed significant increases after fermentation: lignin peroxidase from 26.7267 to 51.6833 nmol/min/g (approximately 1.93-fold) (P < 0.05), manganese peroxidase from 24.3167 to 74.9800 nmol/min/g (approximately 3.08-fold) (P < 0.05), and feruloyl esterase from 17.8033 to 49.1133 nmol/min/g (approximately 2.76-fold) (P < 0.05) (Table 1).Table 1. Key enzyme activities in unfermented (RBWB) and fermented (FRBWB) rice bran and wheat bran complex.Table 1. Enzymes (nmol/min/g)Unfermented rice bran and wheat bran complex (RBWB)Fermented rice bran and wheat bran complex (FRBWB)P-valueLignin peroxidase26.7267 ± 4.16^b^51.6833 ± 9.12^a^0.0125Manganese peroxidase24.3167 ± 1.48^b^74.9800 ± 2.68^a^<0.0001Feruloyl esterase17.8033 ± 1.89^b^49.1133 ± 2.89^a^<0.0001Note: Different superscript letters in the same row indicate significant differences between groups (P < 0.05), and enzyme activities are expressed as mean ± standard deviation (SD).

To further elucidate the molecular basis underlying the marked changes in specific DEMs, we performed molecular docking and molecular dynamics simulations to investigate potential interactions between key DEMs and these lignin-degrading enzymes. Seven DEMs were selected: (1) ferulic acid methyl ester (VIP = 1.28, log_2_FC = 10.60), (2) p-coumaric acid ethyl ester (VIP = 1.28, log_2_FC = −3.59), (3) caffeate (VIP = 1.24, log_2_FC = −1.26), (4) apigenin (VIP = 1.21, log_2_FC = −1.12), (5) quercetin-3-O-glucoside (VIP = 1.15, log_2_FC = 1.31), (6) chlorogenate (VIP = 1.18, log_2_FC = −1.15), and (7) luteolin 7-O-(6″-malonylglucoside)* (VIP = 1.28, log_2_FC = −3.91). Their 3D structures were retrieved from PubChem and docked against high-resolution crystal structures of laccase, manganese peroxidase, and lignin peroxidase obtained from UniProt and PDB. All seven compounds exhibited strong affinities (binding energies < −5 kcal/mol), suggesting stable ligand–receptor conformations (Fig. 4). Notably, luteolin 7-O-(6″-malonylglucoside)* and manganese peroxidase displayed the best binding energy of −11.30 kcal/mol, with 11 hydrogen bonds formed by residues including PRO142, VAL175, and SER172, underscoring high specificity (Supplementary material 3). Among the analyzed DEMs, ferulic acid methyl ester—which had the highest log_2_FC (VIP = 1.28, log_2_FC = 10.60)—bound to laccase, manganese peroxidase, and lignin peroxidase with energies of −5.90, −6.80, and − 7.00 kcal/mol, respectively. Since G. frondosa is known to secrete lignin-degrading enzymes (Motoda et al., 2025), such interactions have particular relevance for releasing active substances during RB and WB fermentation. We therefore performed further molecular dynamics simulation for lignin peroxidase–ferulic acid methyl ester complexes.Fig. 4. Docking simulations between seven key differentially expressed metabolites (DEMs) and three G. frondosa enzymes. Compounds 1–7 are (1) ferulic acid methyl ester, (2) p-coumaric acid ethyl ester, (3) caffeate, (4) apigenin, (5) quercetin-3-O-glucoside, (6) chlorogenate, and (7) luteolin 7-O-(6″-malonylglucoside)*, each docked with (A–G) laccase, (H—N) manganese peroxidase, and (O—U) lignin peroxidase.Fig. 4

To evaluate complex stability and binding, we monitored root mean square deviation (RMSD), radius of gyration (Rg), root mean square fluctuation (RMSF), center-of-mass distance, buried solvent-accessible surface area (Buried SASA), and MM-PBSA binding free energy. As indicated by stability measures, the complex's RMSD and Rg gradually converged during the simulation (Fig. 5A, B). RMSF revealed that certain N-terminal and local residues exhibited higher flexibility than others (Fig. 5C), while the evolution of center-of-mass distance (i.e., dock site–ligand separation) also stabilized (Fig. 5D). Buried SASA analysis demonstrated that ligand–protein contact area gradually became stable (Fig. 5E). Superimposed trajectories further suggested that the ligand remained associated with lignin peroxidase (Fig. 5F). In simulation that temporarily excluded solvation effects, both van der Waals (VDW) and electrostatic (ELE) interactions stabilized over time, suggesting increasing ligand–protein affinity (Fig. 5G). Under solvated conditions, the MM-PBSA method was used to compute interaction energies, including electrostatics (ΔE_ele_), van der Waals (ΔE_vdw_), polar solvation (ΔE_pol_), and nonpolar solvation (ΔE_nonpol_). Interestingly, ΔE_MMPBSA_ was driven predominantly by van der Waals interactions, followed by hydrophobic forces, and then electrostatics. The final complex's ΔE_MMPBSA_ = −79.73 ± 1.47 kJ/mol indicated a strong binding affinity. Decomposition analysis revealed that key protein residues—such as PHE-68 and PHE-179—made substantial contributions to binding (Fig. 5H and Supplementary material 4). Hydrogen bonding, mostly 0–3 bonds during simulation, underscored moderate electrostatic interactions (Fig. 5I). Eventually, structural inspection of the docked complex revealed that GLY-167 contributed a hydrogen bond, while PHE-179, HIE-196, PHE-68, and PRO-166 provided pi–pi stacking, pi–pi T-shaped, or pi–alkyl hydrophobic contacts. SER-193, ILE-64, and PRO-168 offered van der Waals interactions (Fig. 5J). Overall, these analyses support the presence of relatively stable protein–ligand interactions and highlight critical interactions throughout the simulation.Fig. 5. Molecular dynamics simulations of the lignin peroxidase–ferulic acid methyl ester complex. (A) Root mean square deviation (RMSD). (B) Radius of gyration (Rg). (C) Root mean square fluctuation (RMSF). (D) Distance between the ligand and catalytic site. (E) Buried solvent-accessible surface area (Buried SASA). (F) Overlay of sampled conformations. (G) Total van der Waals (VDW)/electrostatic (ELE) binding energies. (H) Residue-based energy contributions. (I) Hydrogen bond evolution. (J) Docked complex showing key interaction types.Fig. 5

G. frondosa fermentation modifies the profile of small-molecular glycosides in RB and WB complex

3.5

We observed that glycosidic compounds accounted for 18.10% of the total metabolome (Fig. 6A). Consequently, we further examined G. frondosa fermentation's impact on RB and WB glycosides and identified 314 glycosidic DEMs between FRBWB and RBWB (VIP > 1, P < 0.05, |log_2_FC| > 1), spanning mainly phenolic acids (47) and flavonoids (182), with smaller contributions from alkaloids, lignans/coumarins, terpenoids, tannins, quinones, and others (Fig. 6B). Subclass analysis indicated that flavones, flavonols, dihydroflavones, and phenolic acids were the most abundant groups. Overall, fermentation tended to increase smaller, free phenolic acids and flavonoid aglycones while decreasing glycosylated derivatives, suggesting enhanced release of low-molecular-weight phenolics. A third-level classification showed that glucosides dominated the glycosidic pool in both samples, but their proportion was slightly lower in FRBWB (67.04%) than in RBWB (68.37%) (Fig. 6C). This observation may imply that G. frondosa secretes β-glucosidases, leading to the liberation of free phenolics or polysaccharide fragments during fermentation (Londoño-Hernández et al., 2017). Focusing on 215 glucoside DEMs, we observed complex patterns of up- and downregulation among derivatives of medicagol, ferulic acid, gallic acid, protocatechuic acid, vanillic acid, kaempferol, and other phenolics (Fig. 6D, E).Fig. 6. Effect of G. frondosa fermentation on glycosidic differentially expressed metabolites (DEMs) in rice bran (RB) and wheat bran (WB) complex. (A) Proportion of glycosidic DEMs within the overall DEM pool. (B) Primary and secondary classifications of glycosidic DEMs. (C) Tertiary classification of glycosidic DEMs. (D) Variation trends of glycosidic DEMs. (E) Boxplot visualization of key glycosidic DEMs. FRBWB: fermented RB and WB product. RBWB: unfermented RB and WB product.Fig. 6

G. frondosa fermentation changes amino acid composition of RB and WB complex

3.6

We further explored the amino acid composition of RBWB and FRBWB (Table 2). Overall, fermentation led to noticeable increases in most amino acids. In particular, aspartic acid (Asp), threonine (Thr), serine (Ser), glutamic acid (Glu), glycine (Gly), valine (Val), and lysine (Lys) showed marked elevations in FRBWB. Cystine (Cys) was the only amino acid that decreased upon fermentation, potentially reflecting cleavage of disulfide bonds by microbial enzymes. The increased levels of essential amino acids (e.g., Val, Ile, Leu, Lys) suggest that G. frondosa fermentation may enhance the nutritional value of RB and WB complex by releasing or synthesizing more biologically accessible amino acids.Table 2. Amino acid composition in unfermented (RBWB) and fermented (FRBWB) rice bran and wheat brancomplex.Table 2. Amino acid (g / 100 g)Unfermented rice bran and wheat brancomplex (RBWB)Fermentedrice bran and wheat brancomplex (FRBWB)Aspartic Acid (Asp)0.840.93Threonine (Thr)0.170.40Serine (Ser)0.180.50Glutamic Acid (Glu)0.911.42Glycine (Gly)0.280.60Alanine (Ala)0.320.50Cystine (Cys)0.080.04Valine (Val)0.250.42Methionine (Met)0.060.08Isoleucine (Ile)0.150.30Leucine (Leu)0.310.49Tyrosine (Tyr)0.130.36Phenylalanine (Phe)0.160.34Lysine (Lys)0.170.35Histidine (His)0.100.16Arginine (Arg)0.320.66Proline (Pro)0.310.59

G. frondosa fermentation enhances in vitro antioxidant capacity of RB and WB complex

3.7

For the purpose of evaluating the in vitro antioxidant ability of FRBWB and RBWB, three assays were conducted: DPPH radical scavenging, hydroxyl radical scavenging, and reducing power. In all assays, both positive controls (BHA and VC) and the samples showed dose-dependent increases in antioxidant capacity. Notably, at equal mass-based extract concentrations (0.5–4 mg/mL), FRBWB consistently showed significantly higher antioxidant activities than RBWB in all three assays (P < 0.05). At 4 mg/mL, FRBWB reached a DPPH scavenging rate of 94.81%, comparable to VC and significantly surpassing BHA (P < 0.05) (Fig. 7A). Similarly, FRBWB showed superior hydroxyl radical scavenging and reducing power at all tested doses (0.5–4 mg/mL) compared with RBWB (P < 0.05). At 4 mg/mL, hydroxyl radical scavenging activity of FRBWB and RBWB reached 56.14% and 29.44%, respectively, while their reducing power values were 84.88 and 30.35 OD, nearly 2.8-fold (Fig. 7B, C). To quantify antioxidant potency, the IC_50_ values were calculated from dose-response curves using nonlinear regression. In agreement with the above trends, the IC_50_ values of FRBWB for DPPH and hydroxyl radical scavenging, and reducing power assays were 0.4539, 3.4010, and 0.9363 mg/mL, respectively—lower than those of RBWB (1.7730, 5.3000, and 1.6640 mg/mL). VC exhibited the lowest IC_50_ values (0.0835, 0.1114, and 0.1000 mg/mL), consistent with its role as the strongest positive control.Fig. 7In vitro and in vivo antioxidant evaluation of fermented rice bran (RB) and wheat bran (WB) product (FRBWB) and unfermented control (RBWB). (A–C) In vitro assays: (A) 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging activity as a function of sample concentration. (B) Hydroxyl radical scavenging activity as a function of sample concentration. (C) Reducing power measured as absorbance at 700 nm. Butylated hydroxyanisole (BHA) and ascorbic acid (VC) served as positive controls in these assays. (D–J) In vivo assays in zebrafish embryos under 2,2′-azobis(2-amidinopropane) dihydrochloride (AAPH)-induced oxidative stress: (D) survival rate at 12.5, 25, 50, 100 and 200 μg/mL to determine optimal dosing, thereafter, effects of 25 and 50 μg/mL on (E) mortality, (F) locomotor activity, (G) hatching rate, (H) heart rate, (I) yolk sac area, and (J) body length. (K–M) Oxidative markers in zebrafish under AAPH stress at 25 and 50 μg/mL: (K) Reactive oxygen species (ROS) production rate. (L) Cell death level. (M) Lipid peroxidation extent.Fig. 7

G. frondosa fermentation enhances in vivo antioxidant activity and protects zebrafish embryos against AAPH-induced oxidative damage

3.8

We also determined a safe concentration range of FRBWB and RBWB in zebrafish embryos. Initially, we determined safe concentration ranges by comparing embryonic survival rates at various doses (Fig. 7D). Viability decreased dose-dependently, with >95% survival at 12.5 μg/mL. At 25 μg/mL, survival rates were 95.42% (RBWB) and 93.75% (FRBWB), dropping below 80% at 200 μg/mL. Notably, FRBWB showed significantly lower toxicity than RBWB at ≥50 μg/mL (P < 0.05). Based on these observations, 25 and 50 μg/mL were selected for subsequent in vivo antioxidant assays. Based on these results, 25 and 50 μg/mL were selected for subsequent antioxidant assays as they induced moderate stress while maintaining acceptable viability, particularly for FRBWB. We next evaluated the protective roles of FRBWB and RBWB (25 and 50 μg/mL) against AAPH-induced oxidative stress in zebrafish embryos. AAPH exposure significantly increased mortality, impaired hatching, reduced heart rate, and enlarged yolk sac size (P < 0.05), indicating successful model establishment. Concretely, AAPH raised embryo mortality to 25%, while co-treatment with FRBWB or RBWB dose-dependently improved survival. At 50 μg/mL, FRBWB achieved a survival rate of 92%, significantly outperforming RBWB (P < 0.05) (Fig. 7E). Both extracts also mitigated AAPH-induced alterations in locomotor activity. FRBWB more strongly suppressed early hyperactivity at 24 hpf and better maintained activity stability by 48 hpf compared to RBWB (Fig. 7F). Hatching rates, reduced to 73.08% by AAPH, were gradually restored by both extracts, with FRBWB at 50 μg/mL showing superior efficacy (87.69%, P < 0.05) (Fig. 7G). Regarding heart rate, markedly lowered by AAPH (130.50% of control), was significantly alleviated by FRBWB, nearly returning to baseline at 50 μg/mL (Fig. 7H). Yolk sac enlargement (approximately 1.4-fold vs. control) was also effectively reversed by FRBWB, bringing it close to normal dimensions (Fig. 7I). Although body length changes were not statistically significant (P > 0.05), both extracts showed a trend toward counteracting AAPH-induced growth retardation (Fig. 7J). In summary, at the tested mass-based doses, FRBWB provided stronger in vivo protective than RBWB across multiple developmental and physiological endpoints, underscoring its potential as a functional ingredient with enhanced antioxidant capacity.

G. frondosa fermentation enhances antioxidant protection in zebrafish by reducing ROS, apoptosis, and lipid peroxidation

3.9

To assess oxidative damage, we measured ROS levels, apoptosis, and lipid peroxidation in AAPH-treated zebrafish. AAPH significantly increased ROS (141.06%), apoptosis (122.12%), and lipid peroxidation (122.86%) compared to the control (P < 0.05). Both RBWB and FRBWB reduced these markers dose-dependently, but FRBWB consistently exhibited stronger protective effects. Specifically, at 25 and 50 μg/mL, RBWB reduced ROS levels by 42.94% and 48.19%, respectively. FRBWB at all tested concentrations likewise suppressed ROS generation, lowering it to 95.43% and 87.23%. Moreover, FRBWB consistently outperformed RBWB at each dose (Fig. 7K). When zebrafish embryos were exposed to 25–50 μg/mL of RBWB or FRBWB, apoptotic cell death decreased significantly in a dose-dependent manner. At 50 μg/mL, FRBWB had a markedly greater inhibitory effect than RBWB (P < 0.05), suggesting a relatively strong protective effect against AAPH-induced cell death (Fig. 7L). Moreover, both RBWB and FRBWB reduced AAPH-induced peroxidation in a concentration-dependent manner. However, FRBWB afforded significantly stronger inhibition than RBWB (P < 0.05), as its lipid peroxidation level ranged between 78.44% and 66.87%, while RBWB remained at 86.06% and 79.95% (Fig. 7M). These findings suggest that FRBWB may more effectively counter AAPH-induced oxidative stress in zebrafish embryos than RBWB, as evidenced by reduced ROS accumulation, lower apoptosis rates, and diminished lipid peroxidation.

Discussion

4

G. frondosa fermentation as a bio-processing strategy to unlock the functional potential of RB and WB complex

4.1

Fermentation has proven to be an effective approach for enhancing the nutritional and functional properties of agricultural by-products such as RB and WB complex (Das et al., 2025; Nemes et al., 2025; Tang et al., 2024). In RB, fermentation has been shown to improve the bioavailability of bioactive compounds and antioxidant capacity (Ardiansyah., 2021; Hou et al., 2024), while in WB it increases soluble dietary fiber and protein and reduces antinutritional factors such as phytic acid, thereby promoting nutrient absorption (Liu et al., 2023; Zhao et al., 2017). Fermentation also facilitates the release of phenolic acids and flavonoids that underpin many health-promoting effects (Li et al., 2022; Webber et al., 2014). Although these advantages, RB and WB differ markedly in composition and cell-wall architecture, which may limit the efficiency of single-substrate fermentation. By combining them, a more diverse and complementary substrate matrix is created, allowing the broad enzymatic repertoire of G. frondosa to act on a wider range of polysaccharides and phenolic conjugates. This co-fermentation strategy is expected to enhance cell-wall disruption and the release of bound bioactives more effectively than fermenting either material alone (Chen, Zhong, et al., 2025; Li, Wang, et al., 2023). Building on this evidence, we employed G. frondosa for co-fermentation of RB and WB complex—denoted as FRBWB and RBWB, respectively—to exploit potential synergistic effects between the two substrates. We then systematically evaluated the physicochemical properties, structural modifications, and differential metabolite profiles resulting from fermentation. Furthermore, the antioxidant capacity of both fermented products was assessed through in vitro and in vivo models. Our findings support that G. frondosa fermentation effectively enhances the functional quality of RB and WB complex, highlighting their potential as valuable ingredients in the functional food industry.

Optimal fermentation duration balances bioactive accumulation, functionality, and structural remodeling

4.2

The temporal patterns observed in polysaccharide, polyphenol, and protein contents during fermentation indicate a dynamic balance between substrate utilization and metabolite accumulation. For G. frondosa, relatively long fermentation cycles and processing challenges such as phase separation have been reported (Wang et al., 2010), underscoring the need to define an appropriate endpoint. In our study, polysaccharides, polyphenols, and soluble proteins all reached their maximal levels at day 13, indicating that this duration provides a favorable balance between biosynthesis and degradation in RB and WB substrate. Previous studies have shown that RB and WB complex serve as efficient substrates for G. frondosa, enabling stable yields of high-quality polysaccharides within relatively short fermentation periods (Cui et al., 2024a, Cui et al., 2024b). Based on typical fungal fermentation behavior and previous reports, one possible explanation is that, during the early stage, G. frondosa utilizes part of the available carbohydrate pool or degrades pre-existing polysaccharides for growth, whereas at later stages, the activation of hydrolytic and biosynthetic pathways may promote the synthesis and accumulation of extracellular polysaccharides, phenolics, and protein precursors (Jiang et al., 2022; Zhang et al., 2018; Zhao et al., 2016).

Fermentation time also had a pronounced impact on physicochemical properties, WHC, OHC, S, and SP all peaking at day 13. The parallel increase in these functional properties and the structural loosening observed by SEM suggest that fermentation-induced disruption of the cell-wall matrix exposes additional hydrophilic and hydrophobic sites, thereby improving WHC and OHC (Fernandes et al., 2011). In line with previous reports on bran fermentation, such changes may involve partial modification of cellulose, hemicellulose, and lignin networks s (Guo et al., 2024). Consistent with these morphological and crystallinity changes, activities of lignin peroxidase and manganese peroxidase, together with feruloyl esterase, were significantly higher in FRBWB compared with RBWB. Furthermore, microbial metabolism likely generates small-molecule carbohydrates that further increase S and SP. TGA further indicated greater mass loss in the 200–400 °C range for day-13 samples—suggesting enhanced cellulose and hemicellulose decomposition (Wen et al., 2017)—and additional weight loss at higher temperatures (400–800 °C). (Ilyas et al., 2019). Collectively, these results demonstrate that G. frondosa fermentation for 13 days optimizes both bioactive compound accumulation and physicochemical functionality while inducing significant structural modifications to RB and WB complex.

Structural remodeling of polysaccharides by fermentation as a basis for enhanced bioactivity

4.3

Molecular weight distribution, monosaccharide composition, and glycosidic linkage patterns are key determinants of polysaccharide bioactivity (Hou et al., 2020; Ji et al., 2020). Consistent with this, our HPLC, FT-IR, and GPC analyses showed that G. frondosa fermentation does not merely increase polysaccharide yield, but also substantially remodels polysaccharide structure. FRBWB contained a broader monosaccharide repertoire, with elevated mannose (Man) and the appearance of fucose (Fuc)—a monosaccharide absent in RBWB—suggesting enzymatic remodeling of native backbones or selective incorporation of specific sugar residues during fermentatio (Ruffing & Chen, 2006). FT-IR spectra further supported these observations by revealing characteristic absorption peaks corresponding to hydroxyl groups (3500–3000 cm^−1^) (Tanthana & Chuang, 2010), C—H stretching/bending vibrations (2900–3000 cm^−1^) (Plichta et al., 2020), carbonyl groups in glycosidic linkages (∼1652 cm^−1^) (Yang et al., 2021), uronic acids (∼1400 cm^−1^) (Wang et al., 2019), pyranose ring vibrations (1150–1000 cm^−1^) (An et al., 2024), and a distinct α–glycosidic linkage signal at ∼526 cm^−1^ (An et al., 2024)—indicating that high-molecular-weight chains were formed or enriched through directional glycosidic bonding during fermentation. GPC analysis suggseted a pronounced shift from predominantly low-molecular-weight fragments in RBWB to mid-to-high molecular-weight fractions in FRBWB, implying both depolymerization of native polymers and subsequent chain rebuilding. In summary, optimization at day 13 not only maximizes bioactive compound yields but also enhances structural properties critical for functional performance—providing a strong foundation for linking these compositional changes to downstream improvements in antioxidant activity observed both in vitro and in vivo.

G. frondosa fermentation reshapes the polyphenolic profile and enhances bioactive metabolite transformation

4.4

Metabolomic profiling showed that G. frondosa fermentation profoundly remodels the polyphenolic landscape of RB and WB complex, with flavonoids and phenolic acids emerging as the predominant DEMs. Among these changes, ferulic acid methyl ester exhibited the highest log_2_FC, indicating the most pronounced upregulation under G. frondosa fermentation. In line with the metabolomic upregulation of ferulic acid derivatives, feruloyl esterase activity increased markedly in FBWB compared with RBWB, supporting enhanced cleavage of ester linkages within lignin–carbohydrate complexes. Concurrent increases in lignin peroxidase and manganese peroxidase activities suggest augmented oxidative depolymerization capacity toward aromatic domains, which is consistent with the observed rise in extractable phenolics. Ferulic acid itself is a phenolic acid widely found in plant cell walls (Zhang et al., 2020), known for its free-radical scavenging capacity, antioxidant effects, and potential to reduce chronic disease risks by mitigating oxidative stress (Salazar-López et al., 2017). Previous reports have shown that enzymatic treatments can efficiently release ferulic acid from RB (Uraji et al., 2013), while lactic acid bacteria can further transform it into vanillin and other phenolic derivatives (Kaur et al., 2013). WB is likewise recognized as a key source of ferulic acid; its bound form can be released by gut microbiota or targeted fermentation, potentially improving bioavailability and conferring beneficial effects on gut health (Duncan et al., 2016). During the fermentation of bran, coordinated enzyme and microbial activities help unlock ferulic acid's release and utilization (Xu et al., 2024). Besides direct enzymatic hydrolysis, fermentation-induced pH shifts promote the release of cell-wall-bound phenolics from RB and WB. In vitro digestion and colonic fermentation models of RB report a markedly higher release of bound phenolics under the relatively neutral pH conditions of colonic fermentation compared with gastric/intestinal digestion (Zhang et al., 2019), and fungal fermentations of RB have also demonstrated increased recovery of ferulic acid in parallel with pH-dependent enzyme production (Denardi de Souza et al., 2019). Additionally, fiber-rich materials such as bran may enrich butyrate-producing microbial populations in the intestine (Akagawa et al., 2021), which transform dietary fibers into short-chain fatty acids, thereby offering cost-effective and eco-friendly strategies for enhancing biomass productivity. The biphasic regulation observed here—upregulation of certain phenolic acids and flavonoids (e.g., ferulic acid methyl ester, quercetin-3-O-glucoside) and downregulation of others (e.g., p-coumaric acid ethyl ester, caffeate, chlorogenate)—likely reflects both substrate specificity and the metabolic versatility of G. frondosa. However, our data do not allow us to attribute the improved antioxidant activity to any single newly formed metabolite; the observed effects may instead result from the combined action and higher extractability of multiple phenolics, including both newly formed and pre-existing compounds. Recent work has likewise demonstrated that ferulic acid can form stable complexes with antioxidant-related proteins, such as Nrf2, with binding strength and conformational stability strongly modulated by environmental factors such as pH (Amshumala et al., 2025). These findings further support the notion that ferulic acid-based ligands can engage in stable and functionally relevant interactions with protein targets that participate in oxidative-stress regulation. For instance, ferulic acid and its derivatives have been reported to attenuate oxidative stress not only by directly scavenging ROS but also by activating the Nrf2/Keap1 pathway, thereby upregulating endogenous antioxidant enzymes such as SOD, CAT, and HO-1 (Chen et al., 2022). Within this context, molecular docking and molecular dynamics simulation further support the possibility of relatively stable binding of ferulic acid methyl ester to lignin peroxidase, primarily mediated by VDW interactions, moderate ELE forces, and hydrogen bonding. The convergence RMSD and Rg, along with a steady SASA, indicating that the ligand–enzyme complex maintained a relatively stable conformation over time. Energy decomposition further underscored the importance of specific residues (e.g., PHE-68, PHE-179) in stabilizing the complex. Our conclusions regarding ligand-induced modulation remain inferential, as we did not perform essential-dynamics/PCA or generate dynamic cross-correlation maps. Therefore, the in silico analysis should be interpreted as providing mechanistic plausibility rather than a direct explanation of the in vivo antioxidant outcomes. Nevertheless, the observed stability and binding energy patterns support the involvement of lignin-degrading enzymes in phenolic release, consistent with functional food studies where newly generated antioxidant constituents are linked to stable metabolite–protein interactions following food processing or biotransformation (Li et al., 2025; Quan et al., 2025).

Liberated phenolics and enriched amino acids as contributors to improved nutritional quality

4.5

Our investigation further revealed the regulatory influence of G. frondosa fermentation on small-molecular glycosides in RB and WB complex with most glycosidic DEMs being flavonoid and phenolic-acid derivatives. In particular, quercetin-3-O-glucoside and related flavonol glycosides, which were significantly altered upon fermentation(log_2_FC = 1.31), are well known to exert strong antioxidant activity by scavenging ROS and modulating redox-sensitive signaling pathways, including Nrf2-dependent transcription of phase II detoxifying enzymes (Lee et al., 2018; Lee et al., 2019; Shokoohinia et al., 2015). Fermentation shifted this pool toward free, low-molecular-weight phenolic acids and flavonoid aglycones at the expense of glycosylated forms, a pattern consistent with β-glucosidase-mediated deglycosylation by G. frondosa (Londoño-Hernández et al., 2017). In parallel, G. frondosa fermentation markedly altered the amino-acid profile, in agreement with previous observations (F.-J. Cui et al., 2024). Most amino acids—particularly aspartic acid (Asp), threonine (Thr), serine (Ser), glutamic acid (Glu), glycine (Gly), valine (Val), and lysine (Lys)—were increased, whereas cystine (Cys) decreased. These shifts likely reflect fungal proteolytic activity releasing amino acids from proteins and peptides (Escudero et al., 2010; Zhao et al., 2020), together with microbial biosynthesis and selective utilization of sulfur-containing residues (Okamoto et al., 2022). Crucially, essential amino acids (e.g., Val, Ile, Leu, and Lys) also increased after fermentation, indicating improved nutritional value for potential human dietary or animal feed formulations—offering a richer source of high-quality protein.

Structural transformation-driven antioxidant enhancement in fermented RB and WB complex

4.6

Polysaccharides are widely recognized as safe, non-toxic natural antioxidants (Ji et al., 2021; Tian et al., 2024). Consistent with this body of evidence, our in vitro assays demonstrated that FRBWB displayed superior DPPH scavenging, hydroxyl radical scavenging, and reducing power compared with the unfermented RBWB with activity at the highest tested concentration approaching that of VC and exceeding BHA. This trend epitomizes the typical concentration-dependent antioxidant behavior of polysaccharides (Li, Chen, Wang, Tian and Zhang, 2010). The synergy between RB and WB may further amplify this effect; in a related study, Lactobacillus plantarum 423 fermentation enhanced both hydroxyl and oxygen radical-scavenging activities of bran substrates, with RB exhibiting stronger antioxidant responses than WB (Wang et al., 2020). Several, likely complementary, mechanisms may underlie the enhanced antioxidant capacity of FRBWB. Enzymatic degradation and structural remodeling of the cell-wall matrix probably facilitate the release of bound phenolic acids and their conversion into more reactive free forms. At the same time, deglycosylation of glycosides and enrichment of free phenolic acids and flavonoid aglycones, together with improved amino-acid profiles, may contribute additional radical-scavenging potential (Mettupalayam Kaliyannan Sundaramoorthy and Kilavan Packiam, 2020). Notably, the significant increases in lignin peroxidase, manganese peroxidase, and feruloyl esterase activities observed in FRBWB are consistent with an enzyme-assisted cleavage of lignin–carbohydrate complexes and ester-linked phenolic conjugates, thereby facilitating the release and transformation of bound phenolics. Enhanced extractability of these antioxidant constituents, together with glycosidic remodeling, likely contributes to the stronger in vitro and in vivo antioxidant performance observed. Similar observations have been reported in other fermented bran systems, where increases in acids, ketones, and related antioxidant compounds broaden the spectrum of free-radical-quenching mechanisms in the final products (Liu et al., 2017; Wang et al., 2020). Taken together, our findings support that G. frondosa fermentation can substantially elevate the antioxidant efficacy of RB and WB complex, expanding its potential applications in functional foods or nutraceuticals. However, because all antioxidant assays were conducted at equal mass-based extract concentrations (mg/mL), the observed enhancement likely reflects the combined effects of higher levels and improved extractability of antioxidant constituents in FRBWB, rather than demonstrating intrinsically stronger antioxidant activity of individual molecules.

Effective mitigation of oxidative damage positions fermented RB and WB complex as a promising functional ingredient

4.7

Oxidative stress, driven by excessive ROS, is closely linked to membrane damage, loss of cellular integrity, and multiple chronic diseases (Yuan et al., 2023). Zebrafish provide a powerful in vivo model for studying such processes because of their transparent embryos, rapid development, and high genetic homology to humans, and have been widely used to evaluate antioxidant and toxicological effects of chemicals and natural products (Chowdhury & Saikia, 2022). Polysaccharides in particular have attracted attention in zebrafish-based studies, where they have been shown to reduce intracellular ROS, inhibit apoptosis, and activate protective pathways such as Nrf2/Keap1 (Jayawardena et al., 2020; Kang et al., 2015). Building on these insights, our zebrafish oxidative damage model demonstrated that FRBWB affords notably stronger protection than RBWB, improving survival, locomotor behavior, hatching rate, and yolk sac morphology, and more effectively reducing ROS accumulation, apoptosis, and lipid peroxidation. This superior performance is likely linked to the fermentation-induced enrichment and structural optimization of phenolics, essential amino acids, and polysaccharides, which together enhance free-radical scavenging and cellular resilience. These findings are in line with previous reports that G. frondosa-derived polysaccharides and their zinc complexes exhibit potent antioxidant, antibacterial, anti-ageing, and metabolic regulatory activities in vitro and in vivo (F. J. Cui et al., 2024; Ma et al., 2014; Zhang et al., 2017). While FRBWB exhibited enhanced antioxidant and protective effects in zebrafish, the decline in embryo survival rate observed when concentrations exceeded 50 μg/mL underscores the importance of safety considerations when formulating functional foods or feed products. Nevertheless, the robust in vivo antioxidant properties of FRBWB, as demonstrated here, position it as a promising functional ingredient or dietary supplement, thereby providing both theoretical and practical foundations for its broader application.

Limitations outline a clear path for future optimization and translation

4.8

Despite these promising findings, several limitations must be acknowledged. First, our fermentation system was optimized for a specific G. frondosa strain; thus, it may not be directly applicable to other fungal species or strains with differing enzymatic profiles. Second, although we identified 13 days as the optimal fermentation period for polysaccharide, polyphenol, and soluble protein accumulation, minor alterations in substrate composition, environmental conditions, or inoculum viability could shift this target duration. Our interpretation of fermentation-induced changes, such as the polysaccharide dynamics, remains speculative due to the lack of direct biochemical data on metabolic flux, despite supportive morphological and functional evidence. Furthermore, while our molecular dynamics simulation and standard analyses revealed binding and stability, PCA was not conducted. As a result, we cannot yet delineate how ferulic acid methyl ester reshapes the dominant collective motions of the enzyme. In our antioxidant assays, FRBWB and RBWB were compared at equal mass-based extract concentrations (mg/mL), rather than normalized to equivalent polyphenol, polysaccharide, or total antioxidant levels. Thus, the superior performance of FRBWB may partly reflect a higher content and extractability of antioxidant compounds per unit dry mass. We also did not include a Trolox reference group, so the absolute strength of protection relative to standard antioxidants could not be quantified. Future work will normalize dosing to comparable polyphenol or polysaccharide equivalents, expand evaluation to additional animal models or human studies, and incorporate transcriptomic or proteomic analyses to elucidate regulatory networks underlying fermentation-driven metabolite transformation, consistent with recent multi-omics investigations of microbial functional metabolism (Huang et al., 2025). Altogether, our work—utilizing RB and WB as substrates with G. frondosa fermentation—offers a sustainable approach to agricultural by-product valorization and supports FRBWB as both a nutritional supplement and a feed additive.

Conclusion

5

G. frondosa fermentation of RB and WB complex for 13 days effectively enhanced the accumulation and extractability of key bioactive components, including polysaccharides, polyphenols, and soluble proteins, while improving physicochemical functionality and inducing pronounced structural remodeling. These changes were accompanied by substantial shifts in the polyphenolic and glycosidic profiles and translated into enhanced antioxidant performance in both in vitro assays and a zebrafish oxidative stress model. Although the observed bioactivities likely arise from the combined effects of multiple fermentation-induced modifications rather than any single compound, the results highlight the potential of G. frondosa fermentation as a sustainable strategy for valorizing agricultural by-products into functional ingredients. Future studies focusing on dose normalization, targeted bioactive isolation, and validation in higher-order animal models will further support the translational application of fermented RB and WB products.

Note: (1) # indicates a significant difference (P < 0.05) between control group and AAPH-treated group. (2) Different lowercase letters denote a significant difference (P < 0.05) between RBWB treatments of various concentrations and AAPH/control group. (3) Different uppercase letters denote a significant difference (P < 0.05) between FRBWB treatments of various concentrations and AAPH/control group. (4) Identical labels indicate no significant difference (P > 0.05). (5) * indicates a significant difference (P < 0.05) between RBWB and FRBWB, with all statistical differences determined by one-way ANOVA followed by Duncan's multiple range test.

CRediT authorship contribution statement

Yuan Sun: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Conceptualization. Yue Zheng: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis. Na Liu: Methodology, Formal analysis. Mu Qier: Supervision, Project administration. Jingwei Qi: Supervision, Project administration. Xiaoping An: Supervision, Project administration, Methodology, Conceptualization.

Funding

This study was supported by the project of Hohhot Municipal Talent Program in Science and Technology Innovation Project (2022RC-IUR-4), the project of national Dairy Technology Innovation Center Project (2024-JSGG-030), the project of basic Scientific Research Operating Funds Program of Higher Education Institutions Directly Under the Inner Mongolia Autonomous Region (BR251001), and the project of Science and Technology Plan Project of Inner Mongolia Autonomous Region (2025KJHZ0049).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.