Effects of cultivation methods and growth stages on bioactive compound profiles and antioxidant capacity in Ganoderma lucidum: a comparative analysis of normal versus antler-shaped fruiting bodies

Abstract

This study investigates how culture condition and growth stage affect morphological development, phenolic acid, flavonoid accumulation, and in vitro antioxidant capacity in Ganoderma lucidum (G. lucidum). We found that 0.5% CO₂ stress induces antler-shaped G. lucidum (AGL) and suppresses biomass accumulation. AGL shows higher antioxidant capacity early on, while conventionally grown fruiting bodies accumulate more phenolics and flavonoids at maturity, yielding greater overall antioxidant activity. Integrated metabolomics, WGCNA, network pharmacology, and molecular docking identified phenolic acids and flavonoids as potential functional metabolites driving these differences. They are predicted to exert antioxidant effects by modulating PTGS2, PPARG, and EGFR, thereby coordinately regulating NF-κB, PI3K-Akt, MAPK, HIF-1, and p53 pathways. Docking was consistent with stable binding of characteristic monomers to target proteins. Our findings support standardized, phenotype-targeted cultivation of G. lucidum and the development of natural antioxidant food ingredients.

Highlights

• •0.5% CO₂ stress induced antler morphology and inhibited fruiting body biomass in Ganoderma lucidum.
• •AGL showed stronger antioxidant capacity on day 7, and CK was superior at maturity (day 40).
• •Phenolic acids and flavonoids may be the key metabolites for antioxidant divergence of the two phenotypes.
• •WGCNA screened brown and green modules significantly enriched in antioxidant-related metabolites.
• •Network pharmacology and molecular docking predicted their multi-target synergistic antioxidant mechanism.

1. Introduction

As a macrofungus with medicinal value, Ganoderma lucidum (G. lucidum) is rich in bioactive compounds such as polysaccharides and triterpenoids that underlie its well-documented pharmacological properties including antioxidant, immunomodulatory, antitumor and hepatoprotective effects (He et al., 2023; Wu et al., 2024). G. lucidum also possesses distinctive nutritional value and is widely utilized in the modern functional food industry (Gao et al., 2025; Xia et al., 2026). In addition, G. lucidum contains antioxidant compounds including polyphenols and flavonoids, rendering it a promising raw material for the development of functional dietary supplements and natural antioxidant foods (Wu et al., 2024). During artificial large-scale cultivation, elevated carbon dioxide concentrations can significantly impair fruiting body differentiation and development, inducing antler-like morphological abnormalities in G. lucidum (AGL) (Sudheer, Taha, Manickam, Ali, & Cheng, 2018). Compared with normally cultivated G. lucidum (CK), these CO₂-induced morphological variants exhibit markedly altered growth kinetics and biomass accumulation patterns. During the early growth stage of G. lucidum, elevated CO₂ concentrations suppress the expression of genes involved in cell cycle regulation, DNA replication and proteasome function, thereby inhibiting cell proliferation and compromising cell wall biosynthesis, which leads to the characteristic antler-like morphology (Fang et al., 2026). Meanwhile, the developmental progression of the mycelium from juvenile to mature stages dynamically remodels secondary metabolic pathways, directly influencing the synthesis and accumulation of bioactive compounds (Xia et al., 2024), thereby modulating the medicinal and nutritional properties of G. lucidum. As demand for high-quality G. lucidum intensifies, stringent requirements are being placed on cultivation efficiency, harvest timing, post-harvest grading and processing, and functional product development.

Excessive accumulation of reactive oxygen species (ROS) and other free radicals in biological systems induces oxidative stress, thereby promoting inflammatory responses, tumor cell proliferation, and the pathogenesis and progression of chronic diseases including diabetes and osteoarthritis (Ansari, Ahmad, & Haqqi, 2020; Nam et al., 2024). Therefore, scavenging excess free radicals in biological systems and maintaining cellular redox homeostasis are essential for sustaining normal physiological functions. Compared with synthetic antioxidants, natural antioxidants exhibit superior safety profiles and reduced potential for adverse effects, rendering them more suitable as dietary exogenous antioxidants (Chen, Dong, Chu, Zhang, & Qin, 2026). Phenolic acids and flavonoids are natural antioxidant secondary metabolites endowed with potent antioxidant activity. They exert their effects through direct free radical scavenging and inhibition of lipid peroxidation, and additionally modulate the expression of microRNAs (miRNAs), thereby indirectly regulating intracellular antioxidant enzyme activity (Wu et al., 2022; Zeng et al., 2025). Phenolic acids and flavonoids are secondary metabolites in G. lucidum that exhibit potent biological activities and constitute the primary biochemical basis for its antioxidant effects.

Widely targeted metabolomics enables comprehensive and unbiased qualitative and quantitative profiling of diverse primary and secondary metabolites in biological samples, establishing it as a pivotal analytical approach in food science. Weighted gene co-expression network analysis (WGCNA) enables the construction of co-expression networks from large-scale metabolomics datasets, thereby accurately revealing intrinsic associations between physiological traits and characteristic metabolites; it has been widely adopted in food science research. For example, WGCNA was employed to analyze metabolite–phenotype correlations in Huangqin tea subjected to different roasting treatments (Zhang et al., 2025), whereas Xiao et al. applied WGCNA to investigate metabolite e-nose flavor profile associations in flower tea processed under varying drying conditions (Xiao et al., 2026). Network pharmacology enables comprehensive mechanistic analysis of multi-component interventions by integrating target prediction, pathway enrichment, and molecular docking, with the latter predicting ligand–protein binding interactions at the molecular level. Although initially developed for pharmaceutical research and development, it has been increasingly applied to elucidate the mechanisms of functional food components (Chen et al., 2026). The integrated multi-technique analytical platform has emerged as the mainstream research paradigm in food science for elucidating the mechanisms of bioactive food compounds and discovering natural functional ingredients.

Currently, the accumulation dynamics of phenolic acids and flavonoids across developmental stages in distinct G. lucidum phenotypes remain poorly characterized, and the key metabolites underpinning stage-dependent variations in antioxidant activity have not been systematically identified. This knowledge gap not only constrains the standardization and optimization of G. lucidum cultivation practices but also impedes the expansion and high-value development of its deep-processing industry. To address the aforementioned research gap, this study selected two G. lucidum phenotypes, and collected samples at three critical developmental stages. Growth parameters, in vitro antioxidant capacity, total phenolic content, total flavonoid content, total polysaccharide content, and soluble sugar content were measured. By integrating widely targeted metabolomics, WGCNA, network pharmacology, and molecular docking, we systematically elucidated how culture condition and growth stage impact the metabolic profile of G. lucidum. Potential phenolic acid and flavonoid monomers governing antioxidant activity were predicted, and their synergistic antioxidant mechanisms were interpreted from three complementary perspectives: compound–target interactions, molecular target engagement, and pathway-level modulation. This work provides a theoretical foundation for targeted optimization of G. lucidum cultivation and the development of natural antioxidant ingredients for food applications.

2. Materials and methods

2.1. Sample treatment

The G. lucidum strain was obtained from the Shihezi Agricultural Reclamation Academy (Xinjiang). The substrate comprised 70% sawdust, 15% bran, 12% corn flour, 2% gypsum, and 1% quicklime (w/w), adjusted to 60% moisture and pH 6.8. It was packed into sterile 1-kg polypropylene bags, autoclaved (121 °C, 2 h), and inoculated with mycelial agar plugs. Following inoculation, the spawn bags are transferred to a dedicated mycelium growth chamber and incubated in the dark at 26 °C for 30 days. Only bags exhibiting complete mycelial colonization and no signs of contamination are subsequently transferred to the fruiting chamber. Two treatments were implemented in adjacent rooms under identical temperature (28 ± 1 °C) and relative humidity (80–90% RH) conditions: CK, involving three one-hour ventilation events per day with CO₂ concentration maintained below 0.1%; and AGL, implementing intermittent ventilation with ventilation frequency dynamically regulated to maintain canopy-level CO₂ concentration at approximately 0.5%, as continuously monitored and feedback-controlled by the Sensirion SCD40 sensor (Note: CO₂ concentration and ventilation mode were varied concurrently; thus, their individual effects cannot be disentangled). Day 0 marked completion of full substrate colonization. Growth parameters and samples were collected on days 7, 20, and 40: CK_1–CK_3 and AGL_1–AGL_3 denote corresponding time points for each treatment (Fu, Yu, Zhan, & Yu, 2026). All samples were flash-frozen in liquid nitrogen and stored at −80 °C until analysis.

2.2. Phenotype and physiology measurements

All assays were conducted in triplicate following the respective manufacturer's protocols.

DPPH was quantified using the Norminkoda NMW0111 kit with Trolox standards (50–200 μg/mL); absorbance was measured at 515 nm (anhydrous ethanol blank) and expressed as μmol Trolox/g. ABTS was quantified using the Norminkoda NMW0110 kit with Trolox standards (0.1–0.6 μmol/mL); absorbance was measured at 405 nm (assay solvent blank) and expressed as μmol Trolox/g. FRAP was quantified using the Norminkoda NMK0001 kit with Trolox standards (0–1.0 μmol/mL); absorbance was measured at 590 nm (PBS blank) and expressed as μmol Trolox/g. Total phenol content (TPC) was determined by the Folin–Ciocalteu method with gallic acid standards (0–500 μg/mL); absorbance was read at 760 nm at room temperature (60% ethanol blank) and reported as mg /g. Total flavonoid content (TFC) was measured using the Norminkoda NMW0120 kit with rutin standards (0–500 μg/mL); absorbance was recorded at 510 nm (60% ethanol blank) and expressed as mg/g. Total polysaccharide content (TPS) was quantified using the Grace Biotechnology G0593W kit: polysaccharides were ethanol-precipitated and detected by the phenol–sulfuric acid method at 488 nm (glucose standards, 0.025–0.5 mg/mL; distilled water blank), reported as mg/g. Soluble sugar content (SSC) was measured using the Norminkoda NMW0209 kit: sugars reacted with anthrone reagent at 95 °C and absorbance was read at 620 nm (glucose standards; distilled water blank), expressed as mg/g. Malondialdehyde content (MDA) was determined using the Norminkoda NMW0161 kit (TBA method, dual-wavelength correction at 532/600 nm) with MDA standards (1.25–80 μmol/L) and distilled water blank, reported as μmol/g. Fresh weight (FW) was determined immediately after harvest: fruiting bodies were carefully excised from substrate bags, gently blotted to remove residual mycelium and debris, and weighed on a Liangping FA104 balance (precision: 0.01 g); data are given in grams (g).

2.3. Metabolomics analysis

Widely targeted metabolomics was performed by MetWare Biotechnology Co., Ltd. (Wuhan), following standardized protocols.

Sample preparation: G. lucidum fruiting bodies were freeze-dried (Scientz-100F), homogenized (Retsch MM 400, 30 Hz, 1.5 min), and stored as fine powder at −80 °C. For extraction, 50 mg (±0.01 mg) powder was weighed (MS105DM balance) and mixed with 1200 μL pre-chilled (−20 °C) 70% methanol (v/v) containing internal standards. The mixture was vortexed 30 s every 30 min for 3 h (6 cycles), centrifuged (12,000 ×g, 4 °C, 3 min), filtered through a 0.22 μm PTFE membrane, aliquoted into LC-MS vials, and stored at −80 °C.

UPLC-ESI-MS/MS analysis: ExionLC™ AD coupled to SCIEX QTRAP 6500+; Agilent Poroshell 120 SB-C18 column (1.8 μm, 2.1 × 100 mm); mobile phases: 0.1% formic acid in water (A) and in acetonitrile (B); gradient: 95–5% A over 10 min, hold 1 min, re-equilibrate 4 min; flow rate 0.35 mL/min; column temp 40 °C; injection volume 2 μL. Data acquired in both positive and negative ESI modes.

MS parameters: source temp 500 °C; spray voltage +5500 V (pos)/−4500 V (neg); GS1/GS2/CUR = 50/60/25 psi; CAD high; MRM mode with nitrogen collision gas; DP and CE optimized per transition; retention time–scheduled MRM. Metabolites identified against MetWare's proprietary database and grouped by ionization mode (positive/negative) for separate data processing.

2.4. Weighted co-expression network analysis

Joint association analysis of G. lucidum physicochemical traits and widely targeted metabolomics data was conducted using WGCNA, adapted from (Xia et al., 2026). WGCNA was performed on the normalized metabolite abundance matrix from all 18 samples (3 biological replicates per treatment group). A signed co-expression network was constructed using Pearson correlation coefficients as similarity measures. The soft-thresholding power was set to 18 to achieve a scale-free topology (R² > 0.85). Modules were identified using the dynamic tree cut algorithm (mergeCutHeight = 0.15; minimum module size = 30).

2.5. Network pharmacology analysis

Metabolite structures (SMILES) were retrieved from PubChem and submitted to SwissTargetPrediction; targets with prediction probability >0.2 were retained. Antioxidant-related disease targets were extracted from GeneCards (keyword: “antioxidant”; GeneCards score ≥ 5). Overlapping targets were identified using Metware Cloud, visualized as a Venn diagram, and imported into Cytoscape 3.9.1. Core metabolites were prioritized by degree centrality; the top five were selected as key active components (Fei et al., 2025). Their targets were used to build a PPI network in STRING (Homo sapiens; confidence ≥0.9), then re-imported into Cytoscape for topology-based filtering: targets with betweenness, closeness, and degree centrality ≥ respective medians were retained (Peng et al., 2024). Enrichment analysis (GO and KEGG) was performed in Metascape (Homo sapiens); results were visualized via the Microbiome Bioinformatics Visualization Cloud Platform.

2.6. Molecular docking

Core antioxidant target protein structures were retrieved from the PDB; corresponding ligands were obtained from PubChem and converted to PDB format using Open Babel 2.3.1. Proteins were prepared in PyMOL 2.5.4: co-crystallized ligands and water molecules removed; polar hydrogens added. Docking was performed with AutoDock Tools 1.5.7 (parameterization) and AutoDock Vina 1.2.2 (sampling, scoring); the lowest-energy pose per ligand–target pair was selected as optimal. Results are reported as binding energy (kcal/mol) and visualized via MicroBioCloud. 3D interactions were rendered in PyMOL 2.5.4; 2D interaction diagrams were generated using LigPlot 4.5.3.

2.7. Statistical analysis

All experiments included three independent biological replicates per condition to ensure reliability and reproducibility. Statistical analysis was performed in SPSS 26.0 (IBM). Two-way ANOVA (culture condition × growth stage) with Tukey's post hoc test was used; significance was set at p < 0.05 (Nie et al., 2026). Values with different lowercase letters differ significantly (p < 0.05). Principal component analysis (PCA), Orthogonal partial least squares discriminant analysis (OPLS-DA), correlation analysis, venn diagrams, volcano plots, and hierarchical clustering heatmaps were generated using Metware Cloud with default parameters.

3. Results and discussion

3.1. Phenotypic and physiological analysis

Phenotypic traits and fresh weight were measured for AGL and CK at 7, 20, and 40 days. At 7 days (Fig. 1A and J), both groups remained at the mycelial emergence stage with no morphological differences. At 20 days, CK developed small kidney-shaped fruiting bodies with white, tender marginal tissue; AGL showed arrested cap differentiation and produced branched antler-like structures. At 40 days, CK caps were fully expanded and marginal tissue fully regressed; AGL branches elongated and thinned, retaining white, tender tissue only at tips. Fresh weight increased progressively in both groups: no significant difference at 7 days; CK was significantly greater than AGL at 20 days (p < 0.05); CK was significantly greater than AGL at 40 days (2.31-fold higher, p < 0.01). These results indicate that elevated CO₂ induces antler-like morphology and suppresses biomass accumulation in G. lucidum in a time-dependent manner (Sudheer et al., 2018).

Fig. 1.

Phenotypic and physiological characteristics of CK and AGL at three time points: day 7, day 20, and day 40: (A) Representative phenotypes, (B) DPPH radical scavenging activity, (C) ABTS radical scavenging activity, (D) Ferric reducing antioxidant power (FRAP), (E) Total phenols content, (F) Total flavonoid content, (G) Total polysaccharide content, (H) Soluble sugar content, (I) Malondialdehyde (MDA) content, (J) Fresh weight.

Antioxidant capacity of the fruiting bodies was assessed at 7, 20, and 40 days using DPPH, ABTS, and FRAP assays (Fig. 1B, C and D). DPPH scavenging activity was significantly higher in AGL than in CK at both 7 and 20 days (p < 0.05), with no significant difference observed at 40 days. In the ABTS assay, AGL exhibited significantly higher activity than CK at 7 and 20 days; notably, CK reached only 65% of AGL's activity at day 20. By day 40, CK surpassed AGL by approximately 28% (p < 0.05). FRAP results followed a consistent temporal pattern: AGL > CK at day 7, no significant difference at day 20, and CK > AGL at day 40 (p < 0.05). These findings indicate that AGL exhibits stronger antioxidant activity during early development, whereas CK demonstrates superior antioxidant capacity at harvest (day 40). Dynamic changes in TPC and TFC confirmed this trend (Fig. 1E and F): at day 7, both were higher in AGL; at day 20, TPC remained slightly higher in AGL while TFC shifted to favor CK; by day 40, both TPC and TFC were significantly higher in CK, indicating that polyphenols and flavonoids are the among the potential endogenous contributors to in vitro antioxidant activity. Regarding carbohydrate metabolism and oxidative stress markers (Fig. 1G, H and I), TPS levels were consistently higher in CK across all time points, while MDA levels were also consistently elevated in CK (up to 1.80-fold higher at day 20), suggesting comparatively lower oxidative damage in AGL. This is because high CO₂ concentrations profoundly inhibit cellular metabolism and reduce respiratory rate, thereby decreasing ROS production. Consequently, MDA content declines and morphological distortion increases. SSC increased initially and then declined. Overall, although AGL showed stronger antioxidant activity early on, CK exhibited superior comprehensive antioxidant capacity at harvest (40 days), with significantly higher TPC and TFC and lower oxidative stress markers, indicating greater suitability for food applications.

The R² values for the two-factor ANOVA ranged from 0.767 to 0.996, indicating a strong model fit. Significant interactions between growth stage and culture condition were observed for most response variables (Table 1). The main effect of culture condition exerted the strongest influence on FRAP and TPC (η²p = 0.930 and 0.924, respectively; both P < 0.001), substantially exceeding that of growth stage. This suggests that antioxidant capacity and phenolic accumulation were predominantly governed by CO₂ concentration. In contrast, TFC exhibited no significant main effects but a highly significant interaction effect (η²p = 0.740, P < 0.001). For TPS, the interaction was non-significant (F = 0.19, P = 0.831, η²p = 0.030), indicating that polysaccharide accumulation was driven primarily by the additive effects of the individual factors. SSC was significantly affected by both main effects and their interaction. Overall, culture conditions had the greatest influence on most indicators, followed by their interaction with growth stage and then growth stage alone.

Table 1.

Two-way ANOVA of the effects of growth stage, culture condition, and their interaction on G. lucidum.

Factor	R2	Growth stage			Culture condition			Interactive effects
		F	P	η2p	F	P	η2p	F	P	η2p
DPPH	0.7671	12.15	0.004**	0.5031	7.83	0.007**	0.5660	5.86	0.017*	0.4941
ABTS	0.9293	13.75	0.003**	0.5340	32.00	<0.001***	0.8421	39.98	<0.001***	0.8695
FRAP	0.9487	4.15	0.064	0.2568	79.85	<0.001***	0.9301	28.93	<0.001***	0.8282
TPC	0.9435	4.30	0.060	0.2639	73.18	<0.001***	0.9242	24.95	<0.001***	0.8062
TFC	0.7969	5.80	0.033*	0.3257	3.60	0.059	0.3752	17.04	<0.001***	0.7396
TPS	0.9285	97.25	<0.001***	0.8902	29.10	<0.001***	0.8291	0.19	0.831	0.0305
SSC	0.9956	333.46	<0.001***	0.9653	521.13	<0.001***	0.9886	668.68	<0.001***	0.9911
MDA	0.8937	57.04	<0.001***	0.8262	7.62	0.007**	0.5596	14.29	<0.001***	0.7043
FW	0.9770	83.54	<0.001***	0.8744	178.81	<0.001***	0.9675	34.57	<0.001***	0.8521

Notes: Values are expressed as mean ± SD (n = 3). R², coefficient of determination; η²p, partial eta squared. Significance levels: * P < 0.05, ** P < 0.01, *** P < 0.001. P values <0.001 are shown as <0.001.

Correlation analysis revealed synergistic relationships among the measured indicators (Fig. S1). Antioxidant capacity was highly consistent with its endogenous contributors: FRAP showed a strong positive correlation with TPC (r = 0.88), significantly exceeding those of DPPH (r = 0.71) and ABTS (r = 0.57) with TPC, and exhibited weaker correlations with TFC (r = 0.21, 0.11 and 0.58), indicating that reducing power was primarily driven by phenolics, with flavonoids playing a secondary role. In the metabolic dimension, ABTS correlated strongly and negatively with SSC (r = −0.88); DPPH also correlated negatively with SSC (r = −0.61); and MDA correlated positively with SSC (r = 0.73) and TPS (r = 0.45). These patterns suggest that soluble sugar accumulation is inversely associated with antioxidant capacity and positively linked to oxidative damage. At harvest (40 days), CK outperformed AGL in ABTS, FRAP, TPC, TFC, TPS, and fresh weight. These results indicate that conventional ventilation rather than high-CO₂ treatment more effectively promotes functional compound accumulation and commercial trait development, supporting superior edible and processing quality and informing CO₂-responsive cultivation optimization for quality-oriented production.

3.2. Changes of metabolites of G. lucidum

3.2.1. Metabolomic analysis

To elucidate the biochemical basis underlying the differences in antioxidant activity among various G. lucidum samples, this study performed a widely targeted metabolomics analysis using UPLC-MS/MS. As shown in Fig. 2A, a total of 867 metabolites were identified and classified into 12 major chemical classes: amino acids and derivatives (175, 20.18%), lipids (157, 13.38%), alkaloids (106, 12.23%), flavonoids (79, 9.11%), nucleotides and derivatives (77, 8.88%), organic acids (68, 7.84%), phenolic acids (58, 6.69%), lignans and coumarins (12, 1.38%), terpenoids (12, 1.38%), tannins (4, 0.46%), quinones (3, 0.35%), and others (116, 13.38%). Unsupervised PCA based on the full metabolite dataset (Fig. 2B) revealed distinct sample clustering in the score plot, indicating significant metabolic differentiation among groups. The first two principal components, PC1 and PC2, accounted for 56.41% and 16.22% of the total variance, respectively, with a cumulative contribution of 72.63%, indicating substantial differences in non-volatile metabolite profiles across experimental groups. The metabolite clustering heatmap (Fig. 2C) revealed clear separation between the CK and AGL, with consistent within-group metabolite accumulation patterns and significant intergroup differences in metabolite abundance, thereby corroborating the PCA-derived sample separation at the metabolic expression level. Following log₁₀ normalization of metabolite abundances, density distribution curves were generated (Fig. 2D). Metabolite abundances across the three developmental stages differed significantly between CK and AGL samples in peak position, distribution range, and density profile, indicating that cultivation condition markedly influences both the overall abundance levels and the distribution characteristics of metabolites during development. This observation is consistent with the developmental metabolite dynamics reported in (Fang et al., 2026).

Fig. 2.

Metabolite analysis of all G. lucidum samples measured by UPLC-MS: (A) Circular plot, (B) PCA, (C) Heatmap, (D) DAMs distribution map.

OPLS-DA was applied to discriminate between AGL and CK at each of the three developmental stages (Fig. S2A-C). Permutation testing yielded R²Y values of 1.000 and Q² values of 0.985, 0.989, and 0.990 for the three models, all exceeding threshold of 0.9. These results confirm that the OPLS-DA models are robust, free from overfitting, and capable of reliably distinguishing group-specific metabolic profiles. Differential metabolites were identified using the criteria FC ≥ 2 or FC ≤ 0.5, VIP > 1, and P < 0.05. Volcano plot analysis identified 213 differentially abundant metabolites (DAMs) in AGL_1 vs. CK_1 (73 upregulated, 140 downregulated), 275 in AGL_2 vs. CK_2 (102 upregulated, 173 downregulated), and 340 in AGL_3 vs. CK_3 (165 upregulated, 175 downregulated). The overall pattern shows that the total number of DAMs between AGL and CK increases progressively across the three developmental stages, and in each stage, downregulated metabolites outnumber upregulated ones (Table S1). Upregulated metabolites were predominantly amino acid derivatives, alkaloids, and free fatty acids, whereas lignans, coumarins, and most organic acids were predominantly downregulated. Among secondary metabolites, flavonoids were consistently more abundant in CK samples, while phenolic acids accumulated dynamically in both AGL and CK across developmental stages. Integrating differential metabolomics and antioxidant phenotypic data, this study prioritized phenolic acids and flavonoids as key investigative targets.

3.2.2. Differential metabolite analysis

Phenolic acids constitute a major class of natural secondary metabolites in G. lucidum and are widely recognized for their potent antioxidant activity (Cao et al., 2023). Integrating the number of differentially abundant phenolic acids across the three developmental stages with quantitative mass spectrometry peak area data (Fig. 3A; Table S2), we observed significant dynamic shifts in phenolic acid accumulation between AGL and CK during growth progression. During the AGL_1 vs. CK_1 comparison, five phenolic acids were upregulated and 20 were downregulated. Although the number of downregulated metabolites exceeded that of upregulated ones, the summed peak area of upregulated phenolic acids was significantly greater. Gentisic Acid was the most abundant phenolic acid in AGL_1, with a content 2.16-fold higher than in CK_1. This compound exhibits low cellular toxicity and, owing to its para-dihydroxy substitution, displays superior antioxidant activity relative to monohydroxy phenolic acids—approaching that of gallic acid, which bears three hydroxyl groups (Cunha et al., 2024). Thus, Gentisic Acid is a key contributor to the elevated phenolic acid–associated antioxidant activity observed during early AGL development. At the AGL_2 vs. CK_2 comparison, nine phenolic acids were upregulated and 17 were downregulated; although the number of downregulated metabolites remained greater than that of upregulated ones, quantitative peak area analysis indicated that the total phenolic acid content in AGL_2 was still higher than in CK_2, yet the difference had significantly narrowed and was no longer statistically significant. In the AGL_3 vs. CK_3 comparison, 12 phenolic acids were upregulated and 17 were downregulated; although the number of downregulated metabolites remained greater, the overall phenolic acid enrichment shifted entirely to CK_3. Tropic acid, 3-(4-Hydroxyphenyl)-propionic acid, DL-3-Phenyllactic acid, and 1-O-Caffeoyl-β-D-xylose are among the monomeric compounds highly enriched in CK_3. 1-O-Caffeoyl-β-D-xylose contains a polyhydroxy structure, which is vital for free radical scavenging and antioxidant activity of phenolic acids (Q. Zhang et al., 2026). However, increasing the number of hydroxyl groups does not linearly enhance antioxidant activity, as steric hindrance and intramolecular hydrogen bonding may restrict hydroxyl group accessibility or modulate their electron-donating capacity (R. Xie et al., 2024). Overall, AGL accumulated phenolic acids predominantly during the early growth stage, whereas CK exhibited substantial accumulation of diverse, highly bioactive phenolic acids at maturity.

Fig. 3.

Metabolite analysis of all G. lucidum samples measured by UPLC-MS: (A) Circular heatmap of phenolic acids (B) Circular heatmap of flavonoids (C) KEGG analysis of AGL_1 vs CK_1 (D) KEGG analysis of AGL_2 vs CK_2 (E) KEGG analysis of AGL_3 vs CK_3.

Flavonoids are a major class of plant-derived phenolic compounds, and extensive evidence supports their potent free radical scavenging and antioxidant activities (Yang, Trevethan, Wang, & Zhao, 2022), indicating their likely contribution to G. lucidum’s antioxidant capacity. Compared with terrestrial plants, the overall abundance of flavonoids in G. lucidum is relatively low; however, their accumulation levels are significantly correlated with G. lucidum’ s in vitro antioxidant activity. The number of differentially accumulated flavonoids across the three growth stages, combined with mass spectrometry-based quantification data (Fig. 3B; Table S2), indicates that flavonoid content was consistently higher in CK than in AGL at all three stages. During the AGL_1 vs. CK_1 comparison, five flavonoid metabolites were upregulated and 28 were downregulated, with downregulated metabolites clearly predominating in number. However, the fold changes of the upregulated compounds were notably high: wistin and quercetin-3-O-(6”-O-acetyl)glucoside exhibited fold changes of 20.99 and 19.62, respectively. These aglycones and their glycosylated derivatives all possess multiple hydroxyl substitutions on aromatic rings, a structural feature central to flavonoid antioxidant activity (Huynh, Wongmaneepratip, & Vangnai, 2024), thereby supporting the early antioxidant potential of AGL_1. Nevertheless, glycosylation generally attenuates direct antioxidant activity, as sugar moieties occupy key phenolic hydroxyl groups, thereby reducing molecular efficiency in free radical scavenging (Xie et al., 2022). At the AGL_2 vs. CK_2 comparison, nine flavonoids were upregulated and 16 downregulated, with downregulation still predominant. Based on peak area quantification, accumulation levels of several highly bioactive flavonoids, such as quercetagetin, robinetin, and myricetin, were significantly higher in CK than in AGL, indicating CK's clear advantage in flavonoid biosynthesis. In the AGL_3 vs. CK_3 comparison, total flavonoid content showed no significant difference based on peak area quantification. However, several bioactive monomers, including isoschaftoside, isorhamnetin-6-C-glucoside, and dihydrokaempferol, were markedly enriched in CK_3. Although glycosylation generally attenuates direct antioxidant activity, C-glycosides exhibit superior chemical stability and resistance to enzymatic hydrolysis compared with O-glycosides owing to their robust C—C glycosidic bonds (Xie et al., 2022). Notably, these compounds displayed very low fold changes (0.02, 0.04, and 0.06, respectively), indicating minimal accumulation in AGL_3 relative to CK_3; their multi-hydroxylated and C-glycosylated structures further confer stable antioxidant activity.

3.2.3. KEGG enrichment analysis

To further elucidate the regulatory mechanisms underlying metabolic pathway alterations in AGL and CK across developmental stages, KEGG pathway enrichment analysis was performed on differentially accumulated metabolites at days 7, 20, and 40 (Fig. 3C–E). The results showed that at day 7, nucleotide metabolism pathways were more enriched in CK, and flavonoid biosynthesis pathways also exhibited higher activity in CK, with corresponding flavonoids primarily accumulating in the control group. Phenolic acid–related pathways showed comparable enrichment levels between the two groups, whereas lipid and alkaloid biosynthesis pathways were significantly enriched in AGL. By day 20, amino acid metabolism pathway activity decreased in AGL, resulting in reduced accumulation of free amino acids; concurrently, alkaloid biosynthesis pathway enrichment shifted toward CK. In parallel, the phenylpropanoid metabolism pathway was progressively activated in AGL, driving sustained synthesis of phenolic acids and flavonoid glycosides (Dong & Lin, 2021). In contrast, pathways associated with soluble sugar and vitamin metabolism remained highly enriched throughout AGL. By day 40, long-term high CO₂ stress strongly suppressed the primary metabolic pathways in AGL, significantly reducing the enrichment of amino acid and nucleotide metabolism pathways and substantially decreasing the capacity for primary metabolite synthesis, consistent with previous studies (Fang et al., 2026). At this stage, the phenylpropanoid pathway activity in AGL reached its peak, with continuous upregulation throughout the pathway leading to a dominant accumulation of flavonoids at AGL_3. However, based on TFC results, the flavonoid content in CK remained higher than that in AGL, and CK also exhibited stronger antioxidant capacity. Overall, long-term high CO₂ stress primarily suppresses the operation of primary metabolic pathways such as amino acid and nucleotide synthesis in G. lucidum. In contrast, the phenylpropanoid pathway associated with flavonoids exhibits strong stress resistance and continuously increasing activity, enabling flavonoids to gradually become the dominant active compound in AGL.

3.3. WGCNA of physiology and metabolites

A differential metabolite co-expression network was constructed using WGCNA, and association analysis was performed between all metabolites and six sample groups (CK_1–CK_3, AGL_1–AGL_3) as well as physiological indicators: DPPH, ABTS, FRAP, TPC, TFC, TPS, SSC, MDA and FW. Based on metabolite expression patterns, components were divided into nine modules (Fig. 4A). As shown in Fig. 4B, the turquoise module showed a highly significant positive correlation with CK_3 (r = 0.73) and negative correlations with AGL_1, AGL_2, and AGL_3; the brown and green modules exhibited significant positive correlations with AGL_1 (r = 0.98 and r = 0.71, respectively). The turquoise, brown, and green modules displayed distinct correlation patterns across CK and AGL developmental groups, indicating that high CO₂ and growth stage specifically impact G. lucidum metabolite synthesis and accumulation. The association results between modules and physiological indicators (Fig. 4C) showed that the brown and green modules were generally positively correlated with antioxidant-related parameters including DPPH, ABTS, FRAP, TPC, and TFC. In particular, the brown module exhibited significant positive correlations with TPC (r = 0.79), FRAP (r = 0.82), and DPPH (r = 0.66), while the green module displayed similar characteristics. Both brown and green modules also showed moderate positive correlations with ABTS and TFC, suggesting that metabolites in these modules primarily impact G. lucidum antioxidant capacity. Meanwhile, the turquoise module correlated significantly and positively with TPS (r = 0.80), MDA (r = 0.69), and FW (r = 0.73), whereas the blue module showed significant positive correlations with TPS (r = 0.61), SSC (r = 0.69), and MDA (r = 0.92). These associations suggest that both modules are linked to primary metabolite accumulation and enhanced G. lucidum growth.

Fig. 4.

Nine physiological indicators (DPPH, ABTS, FRAP, TPC, TFC, TPS, SSC, MDA and FW) of all G. lucidum samples and WGCNA analysis of all metabolites: (A) Cluster dendrogram, (B) Module–group trait relationships, (C) Module–trait relationships, (D) Venn diagram of WGCNA and DAMs.

Metabolite composition analysis was performed on the brown and green modules, which were significantly positively correlated with antioxidant indicators (Table S1). The brown module was enriched in phenolic acids, flavonoids, amino acid derivatives, unsaturated fatty acids, lysophospholipids, and nucleosides. Top hub metabolites included 2,4,6-trihydroxybenzoic acid, gentisic acid, and quercetin glycosides, compounds with polyhydroxylated structures conferring direct radical-scavenging activity (Feng et al., 2023). Also prominent were putrescine, arginine derivatives, dipeptides, and lipids related to linoleic and linolenic acids, which indirectly enhance antioxidant capacity by inhibiting lipid peroxidation and maintaining redox homeostasis (Zhang et al., 2020). The green module was primarily composed of amino acids and unsaturated fatty acids, with α-linolenic acid and crepenynic acid as characteristic components. It contained minor amounts of kaempferol glycosides and tangeretin, along with indole-3-acetic acid. Compared to the brown module, the green module had a higher proportion of amino acid derivatives, and the two modules exhibited functional complementarity in metabolic pathways. Venn diagram analysis of the intersection between brown and green module metabolites and DAMs (Fig. 4D, Table S1) identified 95 metabolites unique to the brown and green modules, 334 metabolites exclusive to differential screening, and 90 overlapping metabolites. The overlapping metabolites accounted for 48.65% of the brown and green module metabolites, indicating robust module selection. Four overlapping phenolic acids were identified, including 2,4,6-trihydroxybenzoic acid and 4-hydroxybenzoic acid, most featuring ortho-dihydroxy or trihydroxy structures associated with strong antioxidant activity. This suggests that phenolic acids are potential drivers of antioxidant capacity differences in G. lucidum. Flavonoids showed no overlapping metabolites, likely due to their low overall abundance and module-specific expression, preventing detection in the intersection of the two analyses. Consequently, flavonoids contribute secondarily to antioxidant activity.

3.4. Network pharmacology results and analysis

Using AGL_3 and CK_3 samples as study subjects, all metabolites were sorted in ascending order by log₂FC, and the top 10 phenolic acid and flavonoid metabolites were separately selected for network pharmacology analysis.

As shown in Fig. 5A, the Venn diagram intersection analysis revealed a total of 687 human antioxidant-related targets identified through comprehensive screening, and 151 potential target proteins predicted based on phenolic acid chemical structures. The intersection of these two sets yielded 56 common targets. By constructing a protein interaction network based on shared targets, topological analysis identified key hub targets such as PTGS2, PPARG, EGFR, ESR1, and HSP90AA1 (Fig. 5B, Table S3). Existing studies indicate that PTGS2 promotes ferroptosis, a form of cell death dependent on lipid peroxidation, by disrupting cellular redox homeostasis (Dai et al., 2024); PPARG may regulate the transcription of downstream antioxidant genes, thereby enhancing the synthesis of endogenous antioxidant enzymes in the body (Muzio, Barrera, & Pizzimenti, 2021); EGFR not only directly inhibits pro-apoptotic factors by activating the PI3K-Akt signaling cascade, but also enhances cellular antioxidant capacity through transcriptional reprogramming (Ma et al., 2024); ESR1 belongs to the NR3 subfamily of nuclear hormone receptors and plays a crucial role in reproduction, bone, and cardiovascular systems (Olczak, Orzechowska, Bednarek, & Lipiński, 2023); while HSP90AA1 does not directly eliminate reactive oxygen species (ROS), it can stabilize antioxidant pathway proteins and modulate transcription factors via its chaperone function, thus participating in cellular adaptation and repair following oxidative damage (Shih et al., 2022). Therefore, these multiple core targets collectively form the molecular framework underlying the biological effects of phenolic acids. As shown in Fig. 5C, based on node connectivity values (Table S4), the top five phenolic acid metabolites are 3-(4-hydroxyphenyl)propionic acid, vanillic acid, DL-3-phenyllactic acid, tropic acid, and methyl 2,4,6-trihydroxybenzoate. The phenolic hydroxyl group is the core functional moiety responsible for antioxidant activity. Based on this structure-activity relationship analysis, methyl 2,4,6-trihydroxybenzoate and vanillic acid exhibit higher antioxidant activity than the other compounds.

Fig. 5.

Network pharmacology analysis of phenolic acids in AGL_3 vs CK_3: (A) Venn diagram of antioxidant and phenolic acid intersection targets, (B) Network diagram of intersection targets–core targets, (C) Phenolic acid–target network diagram, (D) GO enrichment bubble plot, (E) KEGG enrichment bubble plot.

GO enrichment analysis (Fig. 5D) revealed that the targets were significantly enriched in biological processes such as inflammatory response, regulation of oxidative stress, apoptosis, and metabolism of small molecules, while cellular components were primarily associated with nuclear envelope and membrane-related structures. This suggests that phenolic acids not only reduce free radical generation by suppressing excessive inflammation and oxidation-induced abnormal apoptosis, but also alleviate the accumulation of peroxidation products by remodeling the intracellular metabolic environment (Bouyahya et al., 2024). KEGG pathway enrichment analysis (Fig. 5E) revealed that the targets were primarily concentrated in two key pathways: NF-κB and PI3K-Akt. Phenolic acids can block inflammatory cascades by inhibiting the abnormal activation of the NF-κB pathway, while positively activating the PI3K-Akt pathway to upregulate antioxidant enzymes such as SOD and GSH-Px (Afnan et al., 2022). Network pharmacology results predict at the molecular level that phenolic acids may be the core functional components responsible for CK_3's antioxidant capacity.

As shown in Fig. 6A, the Venn diagram intersection analysis revealed a total of 687 human antioxidant-related targets identified through comprehensive screening, while 183 potential targets were predicted for flavonoids based on structural analysis. The overlap between these two sets yielded 79 common targets. Based on these targets, a protein-protein interaction (PPI) network was constructed, and key hub targets such as TNF, PTGS2, PPARG, EGFR, and BCL2 were selected according to their connectivity (Fig. 6B; Table S3). TNF is a key pro-oxidative factor; its overexpression leads to ROS accumulation and enhanced lipid peroxidation (Zhang et al., 2023). PPARG and EGFR jointly participate in the regulation of the PI3K/Akt signaling pathway, playing an important role in scavenging free radicals and alleviating oxidative damage (Muzio et al., 2021). BCL2, a classic anti-apoptotic protein, inhibits cell death by blocking the mitochondrial apoptotic pathway and also reduces intracellular ROS levels, thereby mitigating oxidative stress-induced cellular injury (Amstad et al., 2001). The compound-target network (Fig. 6C, Table S4) reveals that various flavonoids, including dihydrokaempferide, cirsilineol, gallocatechin 3-O-gallate, aromadendrin, and kaempferol-4′-O-glucoside, exhibit characteristics of multiple targets per compound and shared targets among multiple flavonoids. This suggests that the observed antioxidant effects may result from synergistic interactions among these compounds, consistent with the concept of multi-component and multi-target synergy observed in network pharmacology analysis (Fei et al., 2025).

Fig. 6.

Network pharmacology analysis of flavonoids in AGL_3 vs CK_3: (A) Venn diagram of antioxidant and flavonoids intersection targets, (B) Network diagram of intersection targets–core targets, (C) Phenolic acid–target network diagram, (D) GO enrichment bubble plot, (E) KEGG enrichment bubble plot.

GO enrichment analysis (Fig. 6D) showed that the targets were significantly enriched in biological processes such as response to oxidative stress, apoptosis, lipid metabolic process, and response to hypoxia, indicating that flavonoids can eliminate excess free radicals, block oxidative stress-induced apoptosis, and reduce the production of lipid peroxides by optimizing lipid metabolism. KEGG enrichment analysis (Fig. 6E) revealed significant enrichment of targets in pathways including PI3K-Akt, MAPK, HIF-1, and p53. Flavonoids activate the PI3K-Akt pathway to enhance the activity of endogenous antioxidant enzymes, while simultaneously suppressing the abnormal activation of MAPK and HIF-1, thereby reducing tissue damage caused by oxidative inflammation (Jomova et al., 2025). Based on integrated network pharmacology data, flavonoids are predicted to possess a complete molecular regulatory network for antioxidant effects; the substantial accumulation of flavonoids in CK3 may be one of the reasons for the enhanced antioxidant capacity.

Integrated analysis of the two network pharmacology datasets reveals that phenolic acids and flavonoids share multiple key hub targets, including PTGS2, PPARG, and EGFR. These shared targets constitute an important molecular basis for the synergistic antioxidant effects of these two classes of active compounds. In GO functional enrichment analysis, both compound classes showed highly similar enrichment patterns, primarily concentrating on biological processes such as regulation of oxidative stress, inflammatory response, apoptosis, and lipid metabolic process. The difference lies in the fact that phenolic acids were additionally significantly enriched in various small-molecule metabolic pathways, while flavonoids specifically enriched response to hypoxia-related pathways. In KEGG enrichment analysis, PI3K-Akt and MAPK were key signaling pathways commonly regulated by both compounds, conferring protective effects through upregulation of endogenous antioxidant enzymes and suppression of oxidative damage. The distinction is that phenolic acids mainly target the NF-κB pathway to block inflammatory cascades, whereas flavonoids specifically modulate the HIF-1 and p53 signaling pathways. This study found that both phenolic acids and flavonoids are predicted to be synergistic bioactive components underlying the antioxidant activity of G. lucidum, with CK_3 promoting their simultaneous enrichment, thereby enhancing the in vitro antioxidant capacity of G. lucidum through synergistic effects.

3.5. Molecular docking results and analysis

To further investigate the potential mechanisms of key antioxidant compounds in G. lucidum, this study conducted molecular docking analyses between active components and core targets (Fig. S3, Table S5). Results from Fig. S3A showed that the binding energies of five phenolic acids with EGFR, ESR1, HSP90AA1, PPARG, and PTGS2 were all below −5.5 kcal/mol, indicating spontaneous and stable binding to target proteins (Chen et al., 2026). Among them, 3-(4-hydroxyphenyl)propionic acid and DL-3-phenyllactic acid exhibited the strongest binding affinity to HSP90AA1, with binding energies of −7.1 kcal/mol and −7.6 kcal/mol, respectively. As shown in Fig. S3B, the binding energies of five flavonoids with BCL2, EGFR, PPARG, PTGS2, and TNF were all less than −6.8 kcal/mol, demonstrating overall superior binding ability compared to phenolic acids. In particular, Kaempferol-4′-O-glucoside and dihydrokaempferol exhibited binding energies of −9.9 and −10.0 kcal/mol with PTGS2, representing the most favorable interactions within the flavonoid group. EGFR, PPARG, and PTGS2 emerged as high-affinity common targets shared by both compound classes.

Six ligand-receptor complexes with the optimal binding energies were selected for two- and three-dimensional conformational analysis (Fig. 7), where A–C represent phenolic acid docking systems and D–F represent flavonoid docking systems. In Fig. 7, 3-(4-hydroxyphenyl)-propionic acid and DL-3-phenyllactic acid are embedded within the hydrophobic cavity of HSP90AA1, stabilizing the complex through hydrogen bonding, hydrophobic interactions, and π-π stacking with amino acid residues such as LEU107, TYR139, and TRP162. Tropic acid interacts with PTGS2 via molecular forces formed with residues VAL228, PHE205, and TYR385 (Fig. 7C). Among the flavonoids, kaempferol-4′-O-glucoside effectively fits into the active pocket of PPARG, establishing a stable binding mode through π-π stacking and multiple hydrophobic interactions (Fig. 7D). Both dihydrokaempferol and kaempferol-4′-O-glucoside form hydrogen bonds and extensive hydrophobic interactions with key amino acid residues of PTGS2, relying on their flavonoid core and glycosylated structures. The network of non-covalent interactions—including hydrogen bonds, hydrophobic contacts, and π-π stacking that provides the essential structural basis for the high-affinity binding of flavonoids to their target proteins (Fig. 7E-F).

Fig. 7.

Molecular docking analysis of core metabolites with core targets, 3D and 2D representation: (A) HSP90AA1 and 3-(4-Hydroxyphenyl)propionic acid, (B) HSP90AA1 and DL-3-Phenyllactic acid, (C) PTGS2 and Tropic acid, (D) PPARG and Kaempferol-4’-O-glucoside, (E) PTGS2 and Dihydrokaempferide, (F) PTGS2 and Kaempferol-4’-O-glucoside.

Molecular docking results indicate that flavonoids have stronger binding affinity to the target proteins than phenolic acids. However, considering the metabolite content data from this study, phenolic acids are present in much higher total concentrations in G. lucidum and exhibit superior cell permeability and bioavailability in vivo, thus contributing more significantly to the overall antioxidant activity. 3-(4-hydroxyphenyl)propionic acid, and DL-3-phenyllactic acid are key phenolic acids, while kaempferol-4′-O-glucoside and dihydrokaempferol are core flavonoids. 3-(4-Hydroxyphenyl)propionic acid can directly scavenge reactive oxygen species via its phenolic hydroxyl group and activate the Nrf2 antioxidant pathway while inhibiting the NF-κB pro-inflammatory pathway (Zhang et al., 2023); DL-3-Phenyllactic acid exhibits broad-spectrum antibacterial activity and effectively suppresses the growth of pathogens such as Salmonella (Rodríguez, Salgado, Cortés, & Domínguez, 2012); as a phenolic acid derivative, it may possess both antioxidant and anti-inflammatory properties. Kaempferol-4′-O-glucoside and dihydrokaempferol can directly eliminate reactive oxygen species through intramolecular phenolic hydroxyl groups and synergistically activate the Nrf2 antioxidant pathway while suppressing the NF-κB inflammatory pathway, thereby collectively exerting antioxidant, anti-inflammatory, and cytoprotective effects (Hussain et al., 2022). These results show, at the molecular level, the potential mechanism underlying the higher in vitro antioxidant activity of CK_3 compared with AGL_3.

4. Conclusions

This study systematically investigated the effects of conventional cultivation and 0.5% CO₂ stress on the phenotype, biomass, accumulation of phenolic acids and flavonoids, and in vitro antioxidant activity of G. lucidum by integrating culture conditions with growth stage. Results showed that elevated CO₂ concentration induced the formation of antler-shaped deformed fruiting bodies and enhanced antioxidant capacity during early growth; however, prolonged exposure disrupted primary and secondary metabolic pathways, thereby inhibiting compound synthesis and biomass accumulation. In contrast, conventionally cultivated G. lucidum exhibited regular morphological development and stable metabolic homeostasis, enabling substantial accumulation of highly bioactive phenolic acids and flavonoids at maturity and resulting in superior overall antioxidant quality. The study predicted that phenolic acids may constitute the principal contributors to antioxidant activity in G. lucidum, whereas flavonoids play a supportive and synergistic role. WGCNA identified the brown and green modules as key metabolic clusters regulating the antioxidant properties of G. lucidum. Network pharmacology predicted that phenolic acids and flavonoids bind to core targets such as PTGS2, PPARG, and EGFR, modulating the NF-κB, HIF-1, and p53 pathways, respectively, and collectively activate the PI3K/Akt and MAPK pathways to exert antioxidant effects. Molecular docking further predicted the stability of interactions between active compounds and target proteins, thereby establishing the compound-target-pathway regulatory network of G. lucidum’s bioactive components. Considering practical production, positive ventilation methods should be adopted to produce G. lucidum with high antioxidant quality, and prolonged exposure to elevated CO₂ concentrations should be avoided. The integrated analytical system developed in this study can serve as a reference for related research on other edible and medicinal fungi. However, the experimental design failed to adequately isolate the effects of ventilation method from those of CO₂ concentration, and the study was limited to in vitro antioxidant assessment without in vivo validation at either the cellular or animal level. Future work incorporating in vivo efficacy assessment, elucidation of key gene regulatory mechanisms, and optimization of targeted cultivation and stage-specific harvesting techniques for G. lucidum will facilitate the industrial application of highly bioactive G. lucidum raw materials in natural antioxidant food development.

CRediT authorship contribution statement

Yongtao Xia: Writing – original draft, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation. Rui Ni: Methodology, Investigation, Data curation. Fuming He: Visualization. Fuyuan Liu: Investigation, Data curation. Tingting Wang: Methodology, Investigation, Data curation. Ziyu Yang: Visualization, Investigation. Chenjing Li: Data curation. Hanjun Weng: Visualization, Investigation. Dongmei Wu: Visualization, Investigation. Neng Gao: Visualization, Investigation. Zhihua Sun: Supervision, Funding acquisition, Conceptualization. Dingguo Zhang: Supervision, Funding acquisition. Li Zhang: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Xinyong Guo: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Formal analysis, Data curation, Conceptualization.

Funding

This study was supported by the “2024 Corps Plan for Scientific and Technological Research in Key Areas” (Project No. 2024AB026), “2024 Science and Technology Development Project of the Fourth Division, Kekedala City” (Project No. 2024GG014), “Science and Technology Finance Plan Project of the Seventh Division, Huyanghe City” (Project No. 2023C19), “Science and Technology Finance Plan Project of the Seventh Division, Huyanghe City” (Project No. 2026C01), “Corps Science and Technology Program” (Project No. 2025YD020), “Third Division, Tumushuke City: Public Bidding and Appointment Project” (Project No. KY2025JBGS06), and “Kashgar Prefecture Science and Technology Program” (Project No. KS2023054 and Project No. KS2022056).

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.

Data availability

Data will be made available on request.