Structural and Metabolic Remodeling of Mixed Lactic Acid Bacteria-Fermented Wheat Germ and Its In Vitro and In Vivo Digestive Stability

Abstract

Fermentation utilizing a combination of lactic acid bacteria (LAB) is known to enhance the nutritional value of wheat germ extract via the production of functional bioactive compounds. In this work, fermentation via compound LAB significantly enhanced the antioxidant activity of wheat germ extract. Compared to the unfermented group (CON), the in vitro antioxidant indices of Lactobacillus fermented wheat embryo extract were increased significantly: DPPH^·+/ABTS^·+ clearance (67.87 ± 3.48%/71.44 ± 5.90%), FRAP value (1.33 ± 0.02 μmol Trolox/10 mg), and active substance content including GSH (78.04 ± 1.43 μmol/g), total phenols (0.53 ± 0.01 mg GAE/10 mg), and total flavonoids (0.032 ± 0.01 mg/10 mg). Moreover, the antioxidant activity and substances of lactic acid bacteria-fermented wheat embryo extract were improved after gastrointestinal digestion compared with CON. In the erastin-induced Drosophila oxidative stress model, LFWGC intervention significantly improved behavioral performance (12.6 ± 3.78 s of tube climbing and 101.2 ± 7.98 jumps) and increased in vivo antioxidant levels: DPPH^·+·+ clearance by 31.75 ± 0.62%, ABTS^·+ clearance by 50.11 ± 0.82%, FRAP to 0.89 ± 0.03 μmol Trolox/10 mg, and GSH (116.30 ± 1.95 μmol/g), total phenols (0.117 ± 0.01 mg GAE/mg), and total flavonoids (0.027 ± 0.002 mg/g). Mechanistically, LFWGC enriched the intestinal flora with Levilactobacillus and Pseudomonas by restoring Tsf1 protein function, upregulating the expression of the TSF1 and GPX4 genes, and activating the pentose phosphate and a-lanine–aspartate-glutamate metabolic pathways, thereby synergistically enhancing the antioxidant defense system. LAB fermentation effectively enhanced the antioxidant capacity of wheat germ extract, providing a theoretical foundation for the development of functional foods.

1. Introduction

Wheat germ, an abundant high-value byproduct of wheat processing, boasts one of the highest nutritional densities among cereal byproducts and is rich in high-quality protein including 8 essential amino acids, unsaturated fatty acids, vitamins, phenolic compounds [1], as well as glutathione GSH precursors, exhibiting significant functional potential in antioxidation, metabolic regulation and intestinal health maintenance [2]. However, natural wheat germ has three key limitations: most antioxidant components such as phenolics and flavonoids exist in glycoside-bound forms or associated with protein–polysaccharide complexes, resulting in a bioaccessibility of less than 30% [3]; antinutritional factors like phytic acid and tannins chelate minerals and inhibit digestive enzyme activity, and its high unsaturated fatty acid content makes it prone to lipid peroxidation during storage causing flavor deterioration and nutrient loss [4], all of which severely restrict its industrial application in functional foods.

Biotransformation technology provides an efficient approach for the high-value modification of wheat germ, among which lactic acid bacteria fermentation has attracted considerable attention due to its safety, mildness and ability to simultaneously achieve nutritional enhancement, structural optimization and stability improvement. The core mechanism of lactic acid bacteria lies in secreting hydrolase systems such as β-glucosidase, lipase and protease to degrade the cellulose–hemicellulose network of plant cell walls and glycoside-bound compounds [5], promote the release and structural modification of bound phenolics and flavonoids, while synthesizing secondary metabolites including Gama-aminobutyric acid, short peptides and exopolysaccharides. Previous studies have confirmed that fermentation with single strains such as Lactiplantibacillus plantarum and Lactobacillus acidophilus can increase the total phenolic content of wheat germ by 40~60% and achieve a phytic acid hydrolysis rate of over 55%, significantly enhancing its antioxidant and digestive stability [6]. However, obvious gaps remain in previous studies. Most focus on single strains or individual functional indicators such as only evaluating antioxidant activity, lacking systematic analysis of the enzyme system -complementarity–metabolic synergy effect of mixed strains, and failed to conduct integrated analysis of the in vitro structural changes, digestive stability of fermented wheat germ and its in vivo bioactivities including intestinal microbiota regulation and disease intervention, making it difficult to fully reveal its functional mechanisms [7].

Ferroptosis is a novel form of programmed cell death driven by iron-dependent lipid peroxidation, and dysfunction of its core pathway (GSH/GPX4 pathway) is closely associated with various diseases including neurodegenerative diseases and metabolic syndrome. Recent studies revealed a bidirectional regulatory crosstalk between the gut microbiota and ferroptosis: the gut microbiota can regulate host iron metabolism and lipid peroxidation levels through metabolites such as short chain fatty acids and phenolic acid derivatives while oxidative stress induced by ferroptosis can reshape the gut microbiota structure [8]. However, research on this crosstalk remains in its infancy, as the roles of intestinal pathogenic bacteria such as Aeromonas and specific metabolites such as capsiate have only been partially clarified [9], and the regulatory role of dietary derived complex biotransformation products in this process has not been addressed with the molecular mechanism urgently needing elucidation.

This study used the gshF gene (a key gene for GSH synthesis) and antioxidant activity as dual targets to screen high-activity lactic acid bacteria and establish single and mixed fermentation systems for wheat germ, systematically analyzing the effects of fermentation on wheat germ’s structure, metabolome, and antioxidant stability through structural characterization, metabolomics, and simulated gastrointestinal digestion experiments to clarify the synergistic advantages of mixed strains. Further, we used the erastin-induced Drosophila ferroptosis model for in vivo experiments, integrating behavioral assays (sleep, motor ability), physiological and biochemical indicators (iron metabolism, GSH content), gut microbiota analysis, and metabolomics to reveal the molecular mechanism by which mixed fermented wheat germ extract intervenes in ferroptosis through regulating the gut microbiota GSH/GPX4 pathway. This study aimed to provide technical support for the high-value biotransformation and utilization of wheat germ, as well as a scientific basis for dietary intervention strategies targeting ferroptosis-related diseases.

2. Materials and Methods

2.1. Chemicals and Materials

The modeling reagent erastin and the positive drug (Fer-1) used in this study were provided by Aladdin (Shanghai, China). Bovine serum albumin, SDS protein loading buffer, membrane transfer buffer, SWE rapid high-resolution electrophoresis buffer, Marker (10–180 kDa), protein-free rapid blocking solution, and HRP-labeled goat anti-rabbit IgG were all provided by Sever Biotech (Wuhan, China). Tap enzyme, 10× Tap enzyme, DEPC water, enzyme-free water, the total RNA extraction kit, the reverse transcription kit, and the total antioxidant capacity (T-AOC) kit (DPPH method) were all provided by Shenggong Biotech (Shanghai, China). The otal antioxidant capacity (ABTS method) kit, total antioxidant capacity (FRAP method) kit, total flavonoid content detection kit, and total phenol content detection kit were all provided by Biyuntian (Shanghai, China). MRS, MRS agar culture medium, and agar were provided by Beijing Road & Bridge (Beijing, China). Yeast powder was provided by Angel Yeast Co., Ltd. (Yichang, China), and white sugar was provided by Shukeman (Guangzhou, China). All solvents were prepared with distilled water. All chemicals were of analytical reagent grade (AR) unless otherwise stated.

2.2. Isolation and Identification of Lactic Acid Bacteria

A 25 g sample of naturally fermented Chinese cabbage from Harbin, Heilongjiang Province, China was placed in a conical flask containing 225 mL of sterile physiological saline (concentration 0.85%) and shaken at 120 rpm for 30 min to prepare a homogeneous bacterial suspension. Subsequently, a gradient dilution was performed, and 0.1 mL of bacterial suspensions at dilutions of 10⁻⁴, 10⁻⁵, and 10⁻⁶ were evenly spread on MRS agar plates. The plates were inverted and placed in a constant-temperature incubator for anaerobic cultivation at 37 °C for 48 h. Single colonies with uniform morphology, neat edges, and milky white color were selected and repeatedly streaked until pure strains were obtained. The pure strains were then inoculated onto MRS liquid medium and cultured at 37 °C for 24 h before collecting the bacterial cells. Genomic DNA was extracted and used as a template for 16S rDNA gene PCR amplification. The amplified products were verified by agarose gel electrophoresis and sequenced using the Sanger method. Sequencing results are shown in Appendix A Table A1. The obtained sequences were compared with known sequences in the GenBank database using BLAST +2.14.0, and highly homologous (98%) sequences were selected to complete strain identification. The identified pure strains were protected with 30% glycerol, aliquoted into cryovials, and stored long-term in an ultra-low temperature freezer (SNOWSONG, Guangzhou, China) at −80 °C.

2.3. Screening of Lactic Acid Bacteria with Antioxidant Ability

Lactic acid bacteria strains with potential glutathione (GSH) synthesis ability were screened by PCR amplification of the gshF gene using the primers listed in Appendix A Table A2. PCR was performed in a 20 μL reaction mixture containing [e.g., 10 μL 2× PCR Master Mix], [0.4 μM each primer], [template DNA], and nuclease-free water. The thermal cycling program was as follows: initial denaturation at 95 °C for 5 min, followed by 40 cycles of denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, and extension at 72 °C for 60 s, with a final extension at 72 °C for 8 min.

Strains showing positive amplification were further rescreened for antioxidant capacity based on DPPH radical scavenging activity. For the DPPH assay, each sample was measured in five replicates (n = 5). Results are expressed as mean ± standard deviation.

2.4. Lactic Acid Bacteria-Fermented Wheat Germ Extract

Wheat germ (Biyuntian Reagent Co., Ltd., Shanghai, China) was mixed with sterile deionized water at a ratio of 1:10 (w/v) to prepare the fermentation substrate, which was then inoculated with 5% (v/v) lactic acid bacteria suspension with a viable cell concentration of 1.3 × 10⁷ CFU/mL. The unfermented sample was set as the control group (CON), single-strain fermented groups were denoted as LFWG11, LFWG14, and LFWG16, and the mixed-strain fermented group was designated as LFWGC. Fermentation was conducted at 37 °C for 48 h. After incubation, the mixture was centrifuged at 8500 rpm for 6 min to collect the supernatant and residual precipitate, which were subsequently freeze-dried and stored at 4 °C.

2.5. Scanning Electron Microscopy

Wheat germ samples were freeze-dried (lyophilized) to constant weight, gently fractured to expose the internal microstructure, mounted on aluminum stubs with conductive carbon tape, and sputter-coated with a thin conductive gold layer (approximately 10 nm) prior to imaging. Samples were imaged using a scanning electron microscope (Gemini300, ZEISS, Oberkochen, Germany) at an accelerating voltage of 2.0 kV and a magnification of 5.0k×. SEM observations were performed on five independently prepared samples (n = 5), and representative micrographs were selected from multiple randomly chosen fields per sample.

2.6. Fourier Transform Infrared Spectroscopy

Spectra were recorded using a Fourier transform infrared (FTIR) spectrometer (PerkinElmer, Waltham, MA, USA) at a resolution of 4 cm⁻¹ over the wavenumber range of 4000–400 cm⁻¹. A fresh background spectrum was collected before each measurement, and raw spectra were processed by background subtraction and baseline correction; when appropriate, spectra were vector-normalized for comparison. Each sample was measured in five replicates (n = 5), and the reported spectrum represents the mean of the replicate measurements.

2.7. Ultraviolet-Visible Spectroscopy

Samples were scanned over the wavelength range of 200–800 nm using a -UV–Vis spectrophotometer (PerkinElmer, Waltham, MA, USA), with ultrapure water used as the blank for calibration. -UV–Vis spectra were blank-corrected (water baseline subtracted) and, where applicable, normalized to facilitate comparison among samples. All -UV–Vis measurements were performed in five replicates (n = 5), and results are presented as mean ± standard deviation.

2.8. Determination of Antioxidant Capacity

Sample pretreatments were consistent across antioxidant assays: 1 mL of fermented wheat germ extract and 10 mg of homogenized Drosophila samples were used for each test. DPPH free radical scavenging capacity, ABTS free radical cation scavenging capacity, and ferric reducing antioxidant power (FRAP) were determined using commercial assay kits (Beyotime Biotechnology, Shanghai, China). Briefly, sample extracts were mixed with the corresponding working reagents according to the kit protocol, followed by incubation in the dark at room temperature for the specified duration. Absorbance was then measured at the characteristic wavelengths (515 nm for DPPH^·+, 734 nm for ABTS^·+, 593 nm for FRAP) using a microplate reader. Scavenging rates or relative antioxidant activity were calculated based on the standard curves and formulas provided in the kit instructions, with blank and standard controls included in each batch of tests. Antioxidant capacities were expressed as μmol Trolox equivalents per gram of sample (μmol TE/g), calculated from a Trolox calibration curve generated in each assay run. A Trolox positive control (mid-point standard) and a reagent blank were included on every microplate as internal and negative controls, respectively. All determinations were performed in five replicates (n = 5).

2.9. Determination of Bioactive Compounds

For bioactive compound quantification, 1 mL of fermented wheat germ extract and 10 mg of homogenized Drosophila samples were analyzed. Total phenolic and total flavonoid contents were determined using commercial kits (Suzhou Keming Biotechnology Co., Ltd., Suzhou, China). Samples were mixed with color-developing reagents, incubated at the recommended temperature and time, and measured at 760 nm (total phenolics) and 510 nm (total flavonoids) to obtain absorbance values, with contents calculated against corresponding standard curves. In addition, glutathione (GSH) levels were measured using the alloxan-based method: samples were deproteinized, mixed with alloxan reagent, and incubated at room temperature, and absorbance was detected at 305 nm. GSH concentration was calculated using a preestablished standard curve, with parallel blank controls to exclude background interference. Total phenolic content was expressed as mg gallic acid equivalents per gram of sample (mg GAE/g) using a gallic acid calibration curve, and total flavonoid content was expressed as mg quercetin equivalents per gram of sample (mg QE/g) using a quercetin calibration curve. For quantitative determinations, standard curves were prepared for each analytical batch, and a mid-level standard was assayed as an internal quality control together with a reagent blank. GSH content was expressed as μmol GSH per gram of sample (μmol/g) based on a reduced glutathione standard curve, with parallel blanks included to correct background signals. All determinations were performed in five replicates (n = 5). Tannin and sphytic acid were additionally quantified to evaluate changes in anti-nutritional factors. Tannin content was determined using a tannin assay kit (Suzhou Keming Biotechnology Co., Ltd., Suzhou, China) following the manufacturer’s instructions, and absorbance was measured at the specified wavelength. Phytic acid was determined using a commercial phytic acid assay kit (Suzhou Keming Biotechnology Co., Ltd., Suzhou, China) according to the manufacturer’s protocol. Reduction rate (%) was calculated as (C_CON − C_sample)/C_CON × 100, where C_CON is the content in unfermented wheat germ (CON) and C_sample is the content in each fermented group. All determinations were performed in five replicates (n = 5).

2.10. In Vitro Digestion Model

Fermented wheat germ extract solution was mixed with an equal volume of simulated gastric fluid (SGF, pH 2.0) and incubated with shaking (150 rpm) in a 37 °C water bath for 30 min. Subsequently, the gastric digestion mixture was combined with an equal volume of simulated intestinal fluid (SIF, pH 7.4) and incubated under the same conditions for an additional 30 min. For SGF preparation (pH 2.0), 2.0 g/L NaCl and 3.2 mg/mL pepsin (≥2500 U/mg, from porcine gastric mucosa) were dissolved in deionized water, and the pH was adjusted to 2.0 with 1 mol/L HCl. SIF (pH 7.4) was prepared by dissolving 6.8 g/L KH₂PO₄ and 10 mg/mL pancreatin (≥400 U/mg, from porcine pancreas) in deionized water, with pH adjusted to 7.4 using 1 mol/L NaOH. Both fluids were freshly prepared before the experiment and sterilized by 0.22 μm polyethersulfone membrane filtration to prevent microbial contamination. After gastric digestion, 2 mL aliquots were collected and mixed with an equal volume of 0.5 mol/L NaHCO₃ to inactivate pepsin. Following intestinal digestion, the reaction was terminated by heating at 80 °C for 10 min. The treated samples were centrifuged at 8000 rpm for 10 min at 4 °C; the supernatants were filtered through a 0.45 μm membrane and stored at −20 °C until analysis. Antioxidant indices and bioactive compounds in the digested samples were determined following the methods described in Section 2.7 and Section 2.8.

2.11. Ferroptosis Induction and Intervention Treatment in Drosophila melanogaster

3–5-day-old male Canton-S wild-type Drosophila melanogaster were obtained from the KIM Laboratory, School of Life Sciences, Harbin Institute of Technology, and maintained in an artificial climate incubator at 25 °C and 60% relative humidity under a 12 h light/dark cycle. Drosophila were divided into groups as described in Table A3: the healthy control group (CON), model group (MOD), unfermented wheat germ extract intervention group (WG), single-strain fermented wheat germ extract intervention groups (LFWG11, LFWG14, and LFWG16), mixed-fermentation wheat germ extract intervention group (LFWGC), and ferrostatin-1 (Fer-1, positive control; POS) group. Except for CON, flies in MOD, WG, all fermented intervention groups, and POS were administered 500 μL of 10 μM erastin (Aladdin, Shanghai, China) every three days to induce ferroptosis, while CON received an equal volume of sterile normal saline. To distinguish fermentation-specific effects from those of the wheat germ matrix, the WG group received unfermented wheat germ extract under erastin challenge (unfermented wheat germ extract + erastin), in parallel with the fermented extract groups. Additionally, flies in each intervention group were fed 500 μL of the corresponding wheat germ extract once daily, and flies in POS were administered 500 μL of 10 μM Fer-1 (Aladdin, Shanghai, China) once daily. Dose rationale: wheat germ extracts were administered at 0.2 mg/mL (w/v), a concentration selected from preliminary range-finding experiments (0.05–0.5 mg/mL) to ensure dietary tolerability and stable intake while providing measurable protection against erastin-induced oxidative injury; 0.2 mg/mL consistently produced protective effects without observable adverse impacts on fly viability or feeding behavior under our conditions. At the end of the experiment, all flies were anesthetized by freezing and stored at −80 °C for subsequent analyses [10]. Sample size and randomization: for survival and behavioral assays, 30 flies per group (n = 30) were used (Section 2.12, Section 2.13, Section 2.14 and Section 2.15), a group size routinely applied in Drosophila behavioral studies and yielding stable estimates in our pilot runs. After collection and acclimation, flies were randomly assigned to experimental groups using a computer-generated random number sequence; allocation was performed into coded vials with equal numbers per group. Vials were labeled with anonymized codes until data collection and primary analyses were completed.

2.12. Survival Rate Assay

Thirty Drosophila per group were selected, and the number of dead individuals in each group was recorded after seven days of normal feeding and calculate the survival rate. Flies were drawn from the randomized allocation described in Section 2.11 (n = 30 per group). Mortality was checked once daily at a comparable circadian time (same Zeitgeber time window each day; ZT2–ZT4), and the observer was blinded to treatment using vial codes.

2.13. Drosophila Sleep Monitoring

The sleep monitoring system developed by the KIMI team of the Harbin Institute of Technology was used to detect the sleep state of fruit flies; the sleep time of fruit flies was recorded every 3 h, and the sleep time of fruit flies was continuously recorded for 48 h. Sleep recording was initiated at a fixed circadian phase (lights-on, ZT0) after a short acclimation period, and locomotor/sleep data were collected continuously for 48 h. Data extraction and analysis were performed using group codes to maintain blinding to treatment.

2.14. Drosophila Climbing Assay

Thirty Drosophila per group were selected and transferred to glass tubes (height: 20 cm, inner diameter: 3.2 cm), and the tubes were tapped gently to allow all Drosophila to fall to the bottom before the time taken for all individuals to climb past the 6 cm mark inside the tubes was recorded. Climbing performance was assessed at a comparable circadian time for all groups (ZT2–ZT6) under the same lighting conditions. The experimenter recording outcomes was blinded to treatment using coded vials.

2.15. Drosophila Jumping Assay

Thirty Drosophila per group were selected and transferred to glass tubes (height: 20 cm, inner diameter: 3.2 cm), and the tubes were rotated at a constant speed to induce jumping behavior before the number of jumps per individual within 1 min was recorded. Jumping performance was assessed at a comparable circadian time for all groups (ZT2–ZT6) under the same lighting conditions. The experimenter recording outcomes were blinded to treatment using coded vials.

2.16. Metabolomics Analysis

A total of 100 mg of each fermented/unfermented wheat germ extract and Drosophila sample was freeze-dried and ground into powder, then mixed with 1 mL of 80% methanol (v/v). Before extraction, an internal standard solution (L-2-chlorophenylalanine, 10 μg/mL; 20 μL per sample) was added to each tube to monitor extraction efficiency and instrument stability. The mixture was vortexed for 1 min, ultrasonically extracted at 4 °C for 30 min, and centrifuged at 12,000 rpm for 15 min. The supernatant was collected and filtered through a 0.22 μm organic-phase filter membrane for subsequent LC-MS/MS analysis. A pooled QC sample was prepared by combining equal aliquots of supernatants from all biological samples and processed identically; an extraction blank (80% methanol with internal standard) was included to assess background contamination. The injection order was randomized, and the LC-MS system was conditioned with multiple QC injections prior to acquisition; during the run, a QC sample was injected every 8 samples to evaluate analytical reproducibility. Raw peak areas were first normalized to the internal standard, and features with poor repeatability in QC samples (RSD > 30%) were removed. Signal drift and batch effects were corrected using QC-based robust LOESS signal correction (QC-RLSC) based on periodic QC injections. MS and MS/MS spectral data were matched against public metabolomic databases including HMDB (http://www.hmdb.ca/), (accessed on 5 December 2025), Metlin (https://metlin.scripps.edu/), (accessed on 10 December 2025), and the Meiji self-built database to obtain metabolite information. For downstream statistical analyses, the normalized peak area matrix was log10-transformed and Pareto-scaled. Principal component analysis (PCA) was used to assess overall clustering and QC stability, and orthogonal partial least squares-discriminant analysis (OPLS-DA) was performed for group discrimination with 7-fold cross-validation and 200-permutation tests to evaluate model overfitting. Differential metabolites were screened by combining multivariate and univariate statistics: features with VIP > 1.0 in OPLS-DA and |log2 fold change| ≥ 1 were further tested by two-tailed Student’s t-test (two-group comparisons) or one-way ANOVA (multiple groups) when data met normality (Shapiro–Wilk test) and homoscedasticity (Levene’s test); otherwise, the Mann–Whitney U test or Kruskal–Wallis test was used. p values were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR), and metabolites with FDR-adjusted p < 0.05 were considered statistically significant.

2.17. 16S rRNA Sequencing

Genomic DNA was extracted from Drosophila samples, and its purity and concentration were determined using a Nanodrop One spectrophotometer (Thermo Fisher Scientific, Bar Harbor, ME, USA) before the target fragment was amplified via PCR with a Bio-Rad S1000 thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA). After successful amplification, it was verified by agarose gel electrophoresis, the PCR product was purified, and the sequencing library was constructed using the ALFA-SEQ DNA Library Prep Kit, followed by analysis performed by Shanghai Majorbio Bio-pharm Technology Co., Ltd. (Shanghai, China; www.majorprotein.com) (accessed on 12 December 2025).

2.18. Quantitative Real-Time PCR

Total RNA was extracted from Drosophila tissues and reverse transcribed into cDNA as the template, with transferrin receptor 1 (Tfr1) and gpx4 genes primer sequences (Sangon Biotech, Shanghai, China) listed in Table A2. RT-qPCR was performed using TB Green^® Premix Ex Taq™ (RR420Q, Takara Bio Inc., Beijing, China) on a qTOWER3 real-time PCR system (Analytik Jena, Jena, Germany), and relative mRNA expression levels were analyzed via the 2^−ΔΔCt method with β-actin as the housekeeping gene.

2.19. Western Blot

Total protein was extracted from Drosophila tissues and treated with RIPA lysis buffer containing the protease inhibitor PMSF, followed by separation of protein samples via SDS-PAGE and transfer onto a PVDF membrane. After blocking with protein-free rapid blocking buffer, the membrane was incubated overnight at 4 °C with a primary antibody against Tsf1 (Wuhan Servicebio Technology Co., Ltd., Wuhan, China) and subsequently with a secondary goat anti-rabbit IgG-HRP antibody (Wuhan Servicebio Technology Co., Ltd., Wuhan, China) at room temperature for 2 h. Protein bands were visualized using ECL detection reagent, imaged with a UVP Gel Studio PLUS Imager (Analytik Jena AG, Jena, Germany), and quantified via Image-Pro (Plus 6.0) Analyzer software.

2.20. Data Analysis

GraphPad Prism 8.0.2 and R software(4.5.2) were used for graphing, while IBM SPSS Statistics 23 and GraphPad Prism 8.0.2 were employed to calculate and present means with 95% confidence intervals. Unfermented wheat germ extract and healthy Drosophila served as blank control groups: The Student’s t-test was used to analyze differences between the Drosophila model group and the healthy control group, while one-way analysis of variance (ANOVA) followed by Dunnett’s multiple comparison test was applied to determine significant differences among the remaining groups. All differences were considered statistically significant at p < 0.05, and experimental data are presented as mean ± standard deviation (SD). Data distribution was examined using the Shapiro–Wilk normality test and Levene’s test for homogeneity of variances. When assumptions were violated, nonparametric tests (Mann–Whitney U test for two groups, or Kruskal–Wallis test followed by Dunn’s multiple comparison test for multiple groups) were applied. For analyses involving multiple comparisons in metabolomics and microbiota, p values were corrected using the Benjamini–Hochberg FDR method unless stated otherwise. For microbiota, alpha-diversity indices were compared using Wilcoxon rank-sum (two groups) or Kruskal–Wallis (multiple groups) tests; beta-diversity was calculated based on Bray–Curtis distances and visualized by PCoA, with group differences assessed by PERMANOVA (adonis in the vegan R package; 999 permutations). Differential taxa were identified using LEfSe (LDA score > 2.0) and confirmed by Wilcoxon/Kruskal–Wallis tests with FDR-adjusted p < 0.05.

3. Results

3.1. Screening and Validation of Lactic Acid Bacteria with Glutathione Synthesis Capacity

Based on the two indicators, the gshF gene and DPPH radical scavenging activity of the strains (Figure 1A,B), three strains (Lactiplantibacillus plantarum HUCF11, 14, and 16) were identified from 10 lactic acid bacteria isolates with the target band, directly confirming that their genomes harbor functional elements related to GSH synthesis. Meanwhile, under the same fermentation conditions, these three target strains significantly increased the GSH level in the system (Figure 1C), and this elevation was highly synchronized with the enhancement of DPPH radical scavenging capacity (Figure 1D). This screening result is consistent with the earlier findings [11], directly validating the reliability of the gene screening–functional verification technical route. To verify if changes in the system before and after fermentation are accompanied by alterations in chemical bonds, structural characterization of the samples was conducted using SEM, FTIR spectroscopy, and -UV–Vis spectroscopy, among other techniques.

Figure 1.

Screening and validation of lactic acid bacteria. (A) Screening of Lactobacillus with GSH synthesis capacity. (B) DPPH free radical scavenging capacity of LAB. (C) GSH content of fermented wheat germ (D) DPPH free radical scavenging capacity of fermented wheat germ. CON indicates non-fermented wheat germ extract by lactic acid bacteria. LFWG11, 14, 16, and LFWGC refer to different samples of fermented wheat germ extracts prepared by mixed fermentation of HUCF11, 14, and 16. Data are expressed as mean ± SD (n = 5). Values are mean ± SD (n = X). Different lowercase letters indicate significant differences among groups (one-way ANOVA followed by Tukey’s post hoc test, p < 0.05).

3.2. Scanning Electron Microscopy

The natural water-soluble components of wheat germ in the unfermented group (CON) exhibited a smooth, dense, and fissured structure (Figure 2A), which is a direct reflection of the aggregation pattern formed by intrinsic intermolecular hydrogen bonds and hydrophobic interactions. After fermentation, all groups (LFWG11, 14, 16, and LFWGC) exhibited a wrinkled, layered, and complex morphology, which is consistent with the mechanism observed in lactic acid bacteria-fermented oats where exopolysaccharide (EPS)-mediated component reorganization induces microstructural wrinkling [12]. Notably, EPS was not directly quantified in this study; therefore, the contribution of EPS is supported by evidence from the literature and warrants targeted quantification in future work.

Figure 2.

Structural analysis of unfermented and fermented wheat germ. (A,B) Scanning electron microscope of the unfermented and fermented group (5000×). (C) FTIR spectra. (D) -UV–Vis absorption spectra. CON indicates non-fermented wheat germ extract by lactic acid bacteria. LFWG11, 14, 16, and LFWGC refer to different samples of fermented wheat germ extracts prepared by mixed fermentation of HUCF11, 14, and 16. Data are expressed as mean ± SD (n = 5).

The irregular and loose particles of the unfermented group (CON) (Figure 2B) reflected the natural embedded structure of starch granules, protein bodies, and fibers in wheat germ. After fermentation, the significant remodeling of particle morphology in all groups was closely associated with the cascade reactions of microbial enzymolysis and component reorganization [13,14]; the rounded particles observed in the single-strain groups (LFWG11, 14, and 16) align with the finding Yang et al. [15]. The degradation of cereal starch by exogenous enzymes (e.g., amylase and protease) can alter particle morphology and enhance component solubility and digestibility, implying improved nutritional availability of fermented wheat germ in this study. In contrast, the particles in LFWGC group were more dense, and the aggregated densification of the cereal microstructure might enhance its textural properties and nutrient retention capacity as a functional ingredient, which stems from the dual effects of enzyme system synergy and anti-nutritional factor degradation by multiple strains. In addition to reflecting EPS-mediated component reorganization, the wrinkled and multilayered morphology implies a fermentation-induced loosening of the native embedded matrix of wheat germ [16,17]. In the unfermented sample, starch granules, protein bodies, and dietary fibers are tightly packed and stabilized by intrinsic intermolecular hydrogen bonds and hydrophobic interactions, which constrains water penetration and limits enzyme accessibility [18,19]. After fermentation, microbial enzymolysis and concurrent macromolecular redistribution (potentially assisted by EPS) disrupt the compact architecture, yielding rougher surfaces and layered/porous domains [20,21]. Such structural remodeling can (i) increase the effective surface area and diffusion pathways for water/solvent, facilitating the release of soluble components, and (ii) expose starch/protein substrates that were previously embedded, thereby improving the accessibility of digestive enzymes and supporting the observed enhancement in solubility and digestibility [22,23]. To investigate changes in intermolecular interactions, FTIR, UV–Vis spectroscopy analyses were further performed.

3.3. FTIR and UV–Vis Spectroscopy

Characteristic absorption peaks of the unfermented group (CON) at 3293 cm⁻¹ (O–H stretching vibration), 1049 cm⁻¹ (C–O stretching vibration), and 992 cm⁻¹ (C=C–H bending vibration) reflect the intrinsic functional group profiles of components such as polysaccharides, polyphenols, and unsaturated fatty acids in wheat germ (Figure 2C). After fermentation, redshift of the C–O and C=C–H absorption peaks indicated that macromolecular components in wheat germ (e.g., polyphenols, carotenoids, unsaturated fatty acids) underwent structural modification via enzymolysis or oxidation by lactic acid bacteria, leading to extension of the conjugated π–π system. The absorption peak of UV–Vis spectroscopy in the range of 210–270 nm (Figure 2D) is a characteristic response of phenolic substances, indicating that more polyphenolic components with antioxidant potential were present in wheat germ after lactic acid bacteria fermentation.

3.4. Untargeted Metabolomic Comparative Analysis of Lactic Acid Bacteria Fermented Wheat Germ Extracts

Principal component analysis (PCA, Figure 3A) of untargeted metabolomics revealed distinct clustering separation among the unfermented group (CON), single-strain fermentation groups (LFWG11, 14, and 16), and mixed-strain fermentation group (LFWGC) in the score plot, with the LFWGC group showing the farthest distance from the other groups indicating a more thorough remodeling effect of mixed strain fermentation on the wheat germ metabolome. This aligns with the conclusion reported by [24] that multi-strain fermentation expands metabolite diversity through metabolic synergy.

Figure 3.

Analysis of the differences in metabolites of fermented wheat germ extract. (A) OPLS-DA analysis, (B,C) Volcano map of differential metabolites. (D) KEGG pathway database analysis. (E) Major chemical metabolites or lipidome analysis. (F) KEGG topology analysis. (G) RDA analysis. CON indicates non-fermented wheat germ extract by lactic acid bacteria. LFWG11, 14, 16, and LFWGC refer to different samples of fermented wheat germ extracts prepared by mixed fermentation of HUCF11, 14, and 16. Data are expressed as mean ± SD (n = 5).

Differentially abundant metabolites screened with VIP > 1 and |log₂ fold change| > 1 (Table A4), along with volcano plots and metabolic pathway analyses (Figure 3B–F), further confirmed the regulatory specificity of the mixed strains (LFWGC). Single strain fermentation exhibited obvious limitations: certain strains (e.g., LFWG11) increased 4-nitrophenol (a harmful substance that induces reactive oxygen species (ROS) production) and either failed to downregulate deoxycholic acid (a potential intestinal irritant) or even elevated its level. In contrast, the mixed strain not only significantly downregulated these two types of substances but also eliminated additional harmful components untouched by single strains, including 13(S)-HOT, 13-HODE (lipid peroxides that directly exacerbate oxidative stress), and 2,6-xylidine (a potential toxicant).

In addition, the mixed strains achieved more thorough degradation of redundant metabolites (e.g., citric acid, D-(+)-malic acid, α,α-trehalose), while only sporadic downregulation of these metabolites was observed in some single strains further reflecting the synergistic effect of mixed strains in eliminating undesirable components. This regulation optimizes the compositional structure of the fermentation system, creating a purer metabolic environment for the enrichment of functional components. The components significantly upregulated by the mixed strains were all high-value functional substances, including indole-3-lactic acid (anti-inflammatory and antioxidant), daidzein (antioxidant and cardioprotective), pseudouridine (immunomodulatory), and valylproline (intestinal mucosal repair), with no strain specific fluctuations observed in single strains (e.g., 2-hydroxy-4-(methylthio)butyric acid was only upregulated in some single strains).

This stable enrichment was attributed to the metabolic network synergy of the mixed strains, where different strains participate in pathways such as aromatic amino acid metabolism and isoflavone biosynthesis, collectively promoting the continuous synthesis of functional components [25]. In addition, the mixed strains simultaneously enriched nutrients including DL-arginine (an essential amino acid), 5-guanylic acid (involved in nucleic acid metabolism), and D-alanyl-D-alanine (mucosal barrier repair), forming a synergistic network covering nutrition, functionality, and intestinal health. While single strains were mostly limited to upregulating components of a single category (e.g., only small-molecule peptides or isoflavones) and lacked such multi-dimensional synergy, this aligned with metabolic pathway analysis results showing significant enrichment of multiple pathways (e.g., phenylalanine metabolism, glutathione metabolism) in the mixed strains reflecting their systematic quality improvement advantage in metabolic regulation. Single strains exhibited a contradiction where beneficial and harmful components were upregulated simultaneously (e.g., LFWG11 increased both 2,6-diaminohexanoic acid and 4-nitrophenol), whereas the mixed strains, through inter-strain metabolic interactions, selectively enriched only beneficial components and avoided harmful accumulation, achieving targeted quality improvement.

Redundancy analysis (RDA, Figure 3G) clearly revealed the correlation between differential metabolites and antioxidant indices (DPPH radicals, ABTS radicals, and FRAP), with the LFWGC group showing a stronger metabolite–antioxidant activity correlation than the single-strain groups (mean r² increased by 0.23), attributed to its diverse differential metabolites covering multiple antioxidant mechanisms to achieve multi-target synergistic enhancement. To clarify its digestive stability and in vivo bioavailability, in vitro and in vivo digestion experiments were further conducted on the fermented wheat germ extracts.

3.5. Stability of Bioactive Substances and Antioxidant Activity Under In Vitro Simulated Digestion

Total phenolics, total flavonoids, and GSH are core antioxidant bioactive substances in wheat germ. The contents of these bioactive substances in the unfermented group (CON) were significantly lower than those in all fermented groups (Figure 4A–C), indicating that lactic acid bacteria fermentation can promote the release and/or formation of bioactive substances via enzymolysis and metabolic conversion. However, because we did not directly quantify extracellular amylase, protease, or esterase activities in the present work, we describe these enzymes as literature-supported, strain-dependent capacities of Lactiplantibacillus plantarum and related LAB rather than asserting secretion for all strains. Specifically, amylolytic L. plantarum strains have been reported and L. plantarum can serve as a producer/host for α-amylase [26], and L. plantarum strains are also reported to possess proteolytic/peptidase systems and esterase/lipase activities that can contribute to macromolecule depolymerization and ester-bond cleavage in fermented matrices [27,28]. These activities are therefore plausibly involved in degrading macromolecules (e.g., polysaccharides, proteins, and lipids) in wheat germ, facilitating the release of bound phenolics and flavonoids. For GSH, strains harboring the bifunctional glutathione synthetase gene gshF (e.g., LFWG11, 14, and 16) are expected to support de novo GSH biosynthesis [29]. The LFWGC group, through inter-strain enzyme system complementarity (e.g., different strains participating in phenolic liberation/biotransformation, GSH synthesis, and polysaccharide degradation), achieved synergistic increases in total phenolics, total flavonoids, and GSH as well as enhanced antioxidant activity (DPPH, ABTS radicals, and FRAP) (Figure 4D and Figure A1A,B). Importantly, the higher post-digestion retention of phenolics/flavonoids and the sustained antioxidant indices support improved digestive stability/bioaccessibility of the bioactives rather than a purely pH-driven effect. This stability is plausibly attributable to (i) microbial biotransformation enabled by strain-specific beta-glucosidases/esterases that remodel phenolic conjugates and can improve persistence under gastrointestinal conditions [30,31], (ii) protective interactions/complexation between polyphenols and macromolecular matrices (proteins/polysaccharides), which can form physical barriers and controlled-release structures that mitigate enzymatic and oxidative degradation during digestion [32,33], and (iii) elevated GSH, which may further preserve antioxidant functionality by maintaining a more reducing microenvironment and limiting phenolic oxidation [34], which is consistent with the conclusion by [35] that multi-strain fermentation enhances functional component synthesis through metabolic network synergy. To address anti-nutritional factors, the reduction rates of tannin and sphytic acid were determined (Figure A2A,B). Compared with the single-strain fermentations (LFWG11, LFWG14, and LFWG16), the mixed fermentation (LFWGC) achieved the highest reduction rates, reaching approximately 33.78 ± 1.95% for tannins and 50.41 ± 3.54% for phytic acid, whereas the single-strain groups reduced tannins by ~14–21% and phytic acid by ~23–36%. These results indicate that mixed fermentation more effectively alleviates anti-nutritional constraints, which may contribute to the improved functional performance observed in the LFWGC group.

Figure 4.

Bioactive substances and antioxidant activity. (A–C) Total phenolic, flavonoid and GSH Content in undigested fermented group. (OS) (D) DPPH free radical scavenging capacity in OS. (E–G) Total phenolic, flavonoid and GSH Content in SGF. (H) DPPH free radical scavenging capacity in SGF (I–K). Total phenolic, flavonoid and GSH Content in SIF. (L) DPPH free radical scavenging capacity in SIF. n = 5 per group. CON indicates non-fermented wheat germ extract by lactic acid bacteria. LFWG11, 14, 16, and LFWGC refer to different samples of fermented wheat germ extracts prepared by mixed fermentation of HUCF11, 14, and 16. Data are expressed as mean ± SD (n = 5). Where each point represents an independent biological replicate, and the bars represent the mean ± standard deviation (n = 5). A significant difference between samples is indicated by different capital lowercase letters (p < 0.05).

After treatment with simulated gastric fluid (low pH, pepsin), the contents of total phenolics, total flavonoids, and GSH (Figure 4E–G) as well as antioxidant activity in the fermented groups remained significantly higher than those in the CON group (Figure 4H and Figure A1C,D), demonstrating the structural stability of bioactive substances in fermented wheat germ under gastric fluid conditions. The LFWGC group maintained the highest GSH content and ABTS scavenging activity after gastric fluid treatment, further confirming that diverse metabolites produced through multi-strain synergy (e.g., exopolysaccharides of varying molecular weights, multiple short peptides) form a more complex protective network, enhancing the stability of bioactive substances in gastric fluid.

After treatment with simulated intestinal fluid (neutral pH, trypsin), the bioactive substance contents (Figure 4I–K) and antioxidant activity (Figure 4L and Figure A1E,F) in the fermented groups decreased but remained significantly higher than those in the CON group (Figure A1G), suggesting that the bioactive substances in fermented wheat germ possessed bioavailability potential in the intestinal segment. This phenomenon was closely associated with the selective enzymolysis in intestinal fluid and the structural tolerance of components [36]. The LFWGC group exhibited the highest total phenolics content and FRAP value after intestinal fluid treatment, demonstrating that diverse metabolites produced through multi-strain synergy possess stronger structural stability in a trypsin containing environment.

Changes in the contents of total phenolics, total flavonoids, and GSH exhibited a significant positive correlation with antioxidant indices (DPPH radicals, ABTS radicals, and FRAP), confirming that these bioactive substances are the core contributors to the antioxidant capacity of fermented wheat germ. Differences in bioactive substance contents and antioxidant activity among the single strain groups (LFWG11, 14, and 16) stemmed from variations in strain specific enzyme systems (e.g., esterase activity, gshF gene expression efficiency), while the LFWGC group achieved diversification of the bioactive substance profile and complementation of antioxidant mechanisms (e.g., simultaneous possession of hydrogen transfer, electron transfer, and metal chelation capabilities) through multi-strain metabolic synergy. This is the key reason for its optimal performance under all treatment conditions. Based on this, Drosophila intervention experiments were further conducted to evaluate the protective effect of fermented wheat germ against in vivo oxidative damage.

3.6. In Vivo Protective Effect of Fermented Wheat Germ Extracts Against Oxidative Damage

The erastin-treated model group (MOD) exhibited significant increases in iron ion content and decreases in GSH and MDA levels (Figure 5A–D), which is consistent with classic ferroptosis phenotypes, in which iron overload promotes ROS generation via the Fenton reaction, while GSH depletion compromises the antioxidant function of GPX4 and accelerates lipid peroxidation. The positive control group (POS) showed improved iron metabolism and GSH homeostasis, further verifying the reliability of the model. The WG group (unfermented wheat germ extract + erastin) was included as a matrix control to separate the baseline effect of wheat germ from fermentation-derived benefits. Based on this model, the regulatory effects of fermented wheat germ extracts on antioxidant defense were further investigated.

Figure 5.

In vivo protective effect of fermented wheat germ extracts against oxidative damage. (A) Drosophila experimental design. (B) Iron content. (C) GSH content. (D) MDA level. (E–G) total phenols, flavonoids and GSH content of fermented wheat germ. (H–J) DPPH, ABTS free radical scavenging capacity and FRAP value. (K) Correlation analysis heatmap. n = 5 per group. CON indicates the healthy control group receiving normal saline (no erastin). MOD refers to the erastin-induced model group. WG denotes the intervention group treated with unfermented wheat germ extract under erastin challenge (unfermented wheat germ extract + erastin). LFWG11, LFWG14, and LFWG16 denote single-strain fermented wheat germ extract intervention groups. LFWGC denotes fermented wheat germ extracts prepared by mixed fermentation of HUCF11, HUCF14, and HUCF16. POS indicates the Fer-1 positive control group. Data are expressed as mean ± SD (n = 5). A significant difference between samples is indicated by different capital lowercase letters (p < 0.05). ** p < 0.01, *** p < 0.001.

As core antioxidant bioactive substances (Figure 5E–G), total phenolics, total flavonoids, and GSH were significantly higher in the LFWGC group than in the MOD group and showed a significant positive correlation with improvements in antioxidant indices (DPPH radicals, ABTS radicals, and FRAP) (Figure 5H–K). Compared with MOD, WG showed a modest increase in phenolics/flavonoids and antioxidant capacity, but the fermented wheat germ extracts—-especially LFWGC—-produced consistently greater increases, indicating that fermentation enhanced the antioxidant bioactive profile beyond the unfermented wheat germ matrix. Accordingly, the improvements in DPPH and ABTS radical-scavenging activity and FRAP values were stronger in the fermented groups than in WG, supporting a fermentation-specific contribution to oxidative stress resistance. Differences between the single-strain fermentation groups (LFWG11, 14, and 16) and the LFWGC group further highlighted the advantage of multi-strain metabolic synergy, where complementary enzyme systems of different strains (e.g., one strain efficiently releasing phenolics and another highly expressing gshF to synthesize GSH) achieved diversification of the bioactive substance profile and complementation of antioxidant mechanisms, which aligns with the finding by Luthfia Hastiani [37] that multi-strain fermentation enhances functional component synthesis through metabolic networks.

3.7. Fermented Wheat Germ Extracts Improve Sleep Disturbances and Behavioral Abnormalities in Ferroptotic Drosophila

The ferroptosis model group (MOD) exhibited significantly fragmented sleep patterns (Figure 6A), resulting from ROS production via the Fenton reaction induced by excessive iron ions these ROS damaged clock neurons (e.g., LNv neurons) and disrupted the expression rhythms of core circadian clock genes (per, tim) [38]. These findings indicate that ferroptosis-associated oxidative stress not only affects cellular viability but also impairs neurophysiological functions governing circadian regulation, suggesting that sleep fragmentation is a functional manifestation of ferroptotic injury in the nervous system.

Figure 6.

Sleep monitoring and locomotor ability. (A) sleep monitoring. (B,C) climbing and jumping ability. n = 5 per group. CON indicates non-fermented wheat germ extract by lactic acid bacteria. MOD refers to the erastin modeling module, while WG denotes the intervention group treated with unfermented wheat germ extract. LFWGC refer to different samples of fermented wheat germ extracts prepared by mixed fermentation of HUCF11, 14, and 16. Data are expressed as mean ± SD (n = 5). A significant difference between samples is indicated by different capital lowercase letters (p < 0.05).

After intervention with fermented wheat germ extracts, particularly in the LFWGC group, sleep rhythmicity was markedly restored and closely resembled that of the CON group. The LFWGC group exhibited superior improvement compared with the single-strain groups, which may be attributed to the synergistic effects of diverse bioactive metabolites generated during multi-strain fermentation. These results demonstrate that fermented wheat germ extracts effectively mitigate ferroptosis-induced circadian disruption, supporting their potential role in protecting neural clock systems from oxidative and iron-dependent damage.

Locomotor activity (climbing and jumping abilities; Figure 6B,C) reflects muscle integrity and systemic energy metabolism in Drosophila. The MOD group showed a significant decline in locomotor performance, likely due to lipid peroxidation—induced damage to muscle cell membranes and mitochondrial dysfunction caused by excessive ROS [39]. This decline further confirms that ferroptosis-related oxidative injury extends beyond neural tissue and compromises peripheral functional systems, including skeletal muscle.

Following supplementation with fermented wheat germ extracts, locomotor activity was significantly restored, with the LFWGC group demonstrating the most pronounced recovery. The restoration of motor function suggests that the extracts exert systemic protective effects, potentially through dual mechanisms involving direct antioxidant activity and modulation of iron homeostasis. The superior efficacy observed in the LFWGC group further supports the advantage of multi-strain fermentation, which may enhance metabolite diversity and functional complementarity.

Collectively, these findings indicate that fermented wheat germ extracts alleviate ferroptosis-induced physiological dysfunction at both behavioral and systemic levels. The improvements in sleep rhythm and locomotor activity provide functional evidence that dietary intervention can attenuate oxidative stress via mediating neural and muscular impairment. Moreover, the enhanced efficacy of the multi-strain fermented product underscores the importance of metabolic synergy in maximizing bioactivity. These results support the feasibility of fermented wheat germ extracts as a functional dietary supplement for mitigating ferroptosis-associated neurobehavioral and metabolic disturbances.

To further elucidate how fermented wheat germ extracts modulate the gut microecological environment during in vivo digestion, preliminary analyses were conducted to assess changes in gut microbiota structure and metabolomic profiles. These analyses aim to clarify whether the observed systemic protective effects are partially mediated through gut–microbiota–metabolite interactions, thereby providing mechanistic support for the functional outcomes described above.

3.8. Fermented Wheat Germ Extract Regulates the Microbiota Structure of Ferroptotic Drosophila

Gut microbiota diversity is a core indicator for maintaining intestinal microecological balance. The Chao and Ace (species richness) as well as Shannon (species diversity) indices of the MOD group were lower than those of the CON group (Figure 7A–C), indicating that oxidative stress induced by iron overload disrupted the living environment of the gut microbiota, leading to a sharp decrease in microbiota diversity. This reduction suggests that the ferroptosis-related high-ROS intestinal environment may suppress susceptible commensals and narrow community ecological niches, thereby weakening microbial resilience and stability. The diversity indices of the LFWGC group significantly recovered and were close to those of the POS and CON groups, which was closely associated with antioxidant components such as phenolics and GSH enriched in fermented wheat germ. The recovery of α-diversity implies that LFWGC supplementation helps reestablish a more favorable intestinal microenvironment, which is consistent with its antioxidant capacity and supports the restoration of microbial homeostasis under oxidative stress.

Figure 7.

Gut microbiota analysis. (A–C) α diversity index. (D) β diversity index. (E,F) Phylum and genus levels. (G) LEfSe analysis. (H,I) Differential genus. CON indicates non-fermented wheat germ extract by lactic acid bacteria. MOD refers to the erastin modeling module, while LFWGC refers to different samples of fermented wheat germ extracts prepared by mixed fermentation of HUCF11, 14, and 16. Data are expressed as mean ± SD (n = 3). * p < 0.05, ** p < 0.01, *** p < 0.001.

Beta diversity analysis (NMDS, Figure 7D) showed that the MOD group exhibited significant separation in microbiota structure from the CON and LFWGC groups, while the LFWGC group was closer to the CON group indicating that mixed strain intervention can restore microbiota structure from ferroptosis-induced disruption to a healthy state. This clustering pattern indicates that iron-overload stress reshaped the overall community structure, whereas LFWGC intervention shifted the microbial community toward the normal configuration, suggesting a corrective effect rather than a nonspecific change in composition.

Microbiota composition analysis (Figure 7E,F) combined with comparison of differential genera (Figure 7G–I) revealed that for harmful bacteria inhibition, Pseudomonas, a potential pathogen that secretes pro-inflammatory substances to exacerbate oxidative stress, was significantly enriched in the MOD group. The enrichment of this genus is consistent with the inflammatory and oxidative milieu expected in the ferroptosis model and may further amplify intestinal stress through pro-inflammatory metabolites and impaired barrier-associated processes. Meanwhile, opportunistic pathogens such as Enterobacter and Stenotrophomonas that were elevated in the MOD group were significantly downregulated in the LFWGC group, reducing the risk of intestinal inflammation and endotoxin release. These changes suggest that LFWGC may reduce the expansion of stress-tolerant, potentially pro-inflammatory taxa that typically proliferate under oxidative conditions.

In terms of beneficial bacteria enrichment, Lactiplantibacillus was significantly enriched in the LFWGC group. As a representative lactic acid bacterium-associated genus, its enrichment is generally linked to improved intestinal ecological balance and may contribute to barrier protection and redox regulation through organic acid production and related metabolic activities. Additionally, the increase in genera such as Brevundimonas and Sphingomonas may participate in phenolic metabolism and antioxidant synthesis, echoing the previous metabolomic finding of antioxidant metabolite enrichment. Taken together, the microbial shifts observed in the LFWGC group support the interpretation that multi-strain fermented wheat germ extract not only restores community diversity but also suppresses potentially harmful taxa while favoring genera associated with metabolic functions relevant to antioxidant utilization. This provides a microecological basis for the functional improvements observed in the ferroptosis model.

3.9. Compound Fermented Wheat Germ Extract Causes Ferroptosis-Induced Disruption of the Metabolic Structure of Drosophila

Partial least squares discrimination analysis (PLS-DA) of metabolomics (Figure 8A) showed that the MOD group exhibited significant clustering separation in metabolic profile from the CON and LFWGC groups, indicating that ferroptosis triggers systematic disruption of the metabolic network in Drosophila. Meanwhile, the metabolic profile of the LFWGC group was significantly closer to that of the CON group, demonstrating that mixed-strain fermented wheat germ extract can achieve the reversal of metabolic abnormalities and restoration of metabolic homeostasis through synergistic regulation via multiple metabolic pathways. Volcano plots (Figure 8B,C) further confirmed that the LFWGC group exhibited a large number of differential metabolites compared to the MOD group (VIP > 1 and |log₂ fold change| > 1), with most of these metabolites reversing the abnormal trends observed in the MOD group reflecting the targeted metabolic regulatory characteristics of mixed strain intervention.

Figure 8.

Metabolomic and gene, protein expression analysis. (A) OPLS-DA analysis. (B,C) volcano plot of differential metabolites, (D) KEGG pathway analysis. (E) RDA analysis. (F) Tfr1 gene expression analysis. (G) Gpx4 gene expression analysis. (H,I) GPX4 protein expression analysis. (J) Correlation analysis. n = 5 per group. CON indicates non-fermented wheat germ extract by lactic acid bacteria. MOD refers to the erastin modeling module, while LFWGC refers to different samples of fermented wheat germ extracts prepared by mixed fermentation of HUCF11, 14, and 16. Data are expressed as mean ± SD (n = 5). A significant difference between samples is indicated by different capital lowercase letters (p < 0.05). * p < 0.05, ** p < 0.01, *** p < 0.001.

Differential metabolite analysis (Table A5) revealed the precise restoration of ferroptosis-related metabolic abnormalities by mixed strain intervention, particularly in sulfur-containing amino acid and glutathione metabolism: 2-Hydroxy-L-Methionine, an intermediate in methionine metabolism, was significantly downregulated in the MOD group but restored to CON group levels in the LFWGC group, while N-Acetyl-D-Cysteine, a GSH derivative abnormally elevated in the MOD group, was reversed by the LFWGC group. Phenolic and organic acid metabolism: Quinic acid, a precursor for phenolic synthesis, was elevated in the MOD group but restored to normal levels in the LFWGC group, while syringic acid (a potent antioxidant phenolic acid) showed no difference in the MOD group but was activated via microbiota metabolism in the LFWGC group. These substances quench free radicals by donating hydrogen atoms, with their conjugated structures enhancing antioxidant activity. For lipid and nucleotide metabolism: 9(Z),11(E)-conjugated linoleic acid, a lipid with anti-inflammatory and antioxidant properties, was downregulated in the MOD group but reversed in the LFWGC group, while 2-hydroxyadenine (a purine metabolite) was elevated in the MOD group but reduced in the LFWGC group. The former inhibits membrane lipid peroxidation, while the latter reduces nucleic acid oxidative damage, collectively maintaining the stability of cellular structures and genetic material [40].

Metabolic pathway enrichment analysis (Figure 8D) showed that the LFWGC group significantly enriched pathways including purine metabolism, glutathione metabolism, and sulfur metabolism with all pathway impact values > 0.5 indicating that mixed-strain intervention achieves systematic restoration of metabolic function through multi-pathway synergy. RDA analysis (Figure 8E) further confirmed that these differential metabolites showed a strong positive correlation with antioxidant indices such as DPPH and ABTS radicals (r² > 0.8), demonstrating that the functional restoration of metabolites is the core material basis for the enhanced antioxidant activity of fermented wheat germ extracts, ultimately forming a cascade effect of metabolic regulation, enhanced antioxidant activity, and phenotypic improvement.

3.10. qPCR and Western Blot Analysis

Tfr1 mRNA expression in the MOD group was significantly altered compared to the CON group (Figure 8F), reflecting transcriptional stress induced by iron overload, while the LFWGC group significantly restored Tfr1 gene expression. This indicates that the extract inhibits abnormal intracellular iron accumulation by regulating the synthesis of iron transporters, suppressing the substrate supply for the Fenton reaction at the source, which constitutes the critical first step in blocking the ferroptotic cascade [41].

Gpx4 gene expression was significantly downregulated in the MOD group (Figure 8G), reflecting transcriptional repression of Gpx4 synthesis during ferroptosis, while the LFWGC group significantly upregulated Gpx4 transcription to levels close to those of the POS group, indicating that the mixed strain fermented wheat germ extract activates Gpx4 gene transcription. Western blot results (Figure 8H,I) showed that GPX4 protein expression was almost undetectable in the MOD group, but the LFWGC group significantly restored its expression. This synergistic activation of transcription and translation directly enhances GPX4 enzymatic activity and improves the scavenging efficiency of lipid peroxides [42].

Correlation analysis (Figure 8J) revealed that beneficial bacterial genera (e.g., Lactiplantibacillus) exhibited significant positive correlations with antioxidant indices (e.g., GSH, total phenolic content) and behavioral parameters (climbing time), which aligns with previous findings that Lactiplantibacillus enhances host GSH synthesis and antioxidant enzyme activity, confirming that enrichment of such probiotic genera strengthens the antioxidant system to improve locomotor function. Core antioxidant metabolites (e.g., syringic acid) showed strong positive correlations with GSH and FRAP, which aligns with reports that syringic acid possesses potent free-radical-scavenging activity. Ferroptosis-related targets (Gpx4 mRNA, GPX4 protein) exhibited positive correlations with beneficial bacterial genera and antioxidant metabolites providing supportive evidence that gut microbiota derived metabolites inhibit ferroptosis by upregulating Gpx4, a mechanism that links microbial/metabolic changes to the suppression of iron-dependent lipid peroxidation. In contrast, pathogenic bacterial genera (e.g., Pseudomonas) exhibited negative correlations with antioxidant and behavioral indices, which aligns with their role in exacerbating oxidative stress [43], underscoring that its downregulation restores host homeostasis. Collectively, mixed strain fermentation enriches beneficial bacterial genera and antioxidant metabolites, synergistically enhancing the antioxidant system, upregulating ferroptosis inhibiting targets, and improving behavioral phenotypes establishing a multi-level mechanistic link between the gut microbiota, metabolic profiles, and host ferroptotic resilience.

4. Discussion

This study systematically investigated the structural, metabolic, and functional modifications of wheat germ induced by fermentation with Lactiplantibacillus plantarum strains, individually and in combination. The discussion is framed around key themes of fermentation-induced structural remodeling, system stabilization, digestive stability, and the underlying mechanisms for alleviating oxidative stress. In this discussion, we integrate the above findings and emphasize their physiological relevance and mechanistic implications to provide a clear separation from Section 3.

Fermentation led to profound microstructural reorganization of wheat germ, transforming its native smooth and dense morphology into a wrinkled, layered architecture. This restructuring is attributed to the synergistic action of microbial enzymolysis and metabolite-mediated reassembly, consistent with the mechanism observed in LAB-fermented oats where EPS-mediated component reorganization induces microstructural wrinkling [44]. Strain-secreted enzymes such as amylases and proteases degrade macromolecular networks, disrupting intrinsic intermolecular forces. The specificity of enzyme systems among different strains [45] accounts for variations in wrinkling degree observed in single-strain fermentations. Concurrently, microbial metabolites including exopolysaccharides (EPSs) and short peptides reassociate with residual components via hydrogen bonding and electrostatic forces, promoting the formation of complex aggregated structures. The more uniform and pronounced wrinkling in the mixed-strain fermented group (LFWGC) reflects metabolic complementarity, where diverse enzyme systems and metabolite profiles facilitate more extensive component interaction and structural integration.

Spectroscopic analyses indicated that fermentation altered intermolecular interactions and conjugate systems within wheat germ components. A redshift in C–O and C=C–H absorption peaks suggested extension of conjugated π-systems in polyphenols and carotenoids, consistent with fermentation-induced enhancement of antioxidant conjugate structures [46]. Conversely, a blueshift of the O–H peak indicated strengthened hydrogen bonding between microbial metabolites (e.g., organic acids, EPS) and wheat germ constituents. These metabolites, rich in polar groups, form an intensive hydrogen bond network with polysaccharides, proteins, and polyphenols [47], enhancing the structural stability of the matrix, which aligns with findings that hydrogen bond formation between LAB metabolites and substrates can improve food matrix stability [48].

In vitro digestion simulations revealed that fermentation significantly increased the content and stability of bioactive compounds. The release of bound phenolics and flavonoids was facilitated by enzymatic degradation, while GSH was synthesized de novo by strains possessing the gshF gene, a screening criterion validated in prior work [49]. The mixed-strain group exhibited the highest levels of these bioactivities, consistent with the conclusion that multi-strain fermentation enhances functional component synthesis through metabolic network synergy [50]. Following simulated gastric digestion, bioactive retention remained high, attributed to fermentation-induced complexation between proteins, polysaccharides, and bioactive compounds, which confers resistance to nonspecific pepsin degradation [51], as aggregated structures stabilized by hydrogen bonds and hydrophobic interactions are less likely to dissociate under acidic conditions [52]. After intestinal digestion, levels remained significantly elevated compared to unfermented controls. The viscous microenvironment formed by EPSs may delay enzymatic degradation and diffusion [53], enhancing intestinal bioaccessibility. Notably, mixed fermentation markedly decreased key anti-nutritional factors (tannins and phytic acid) relative to the single-strain groups (Figure A2A,B). This enhancement may be attributed to the complementary enzyme systems and acidification dynamics among strains: organic acid production lowers pH and promotes phytate hydrolysis, while strain-specific hydrolytic activities (e.g., phytase- and tannase-like activities) can act synergistically to degrade phytate complexes and tannin-associated polymers [54]. Consequently, the mixed consortium not only increased antioxidant-related bioactive compounds but also reduced inhibitors of mineral/protein utilization [55], providing a more comprehensive explanation for the superior overall performance of LFWGC.

The in vivo efficacy was validated using a Drosophila model of erastin-induced ferroptosis. The LFWGC extract most effectively counteracted oxidative damage, restoring GSH levels, reducing MDA, and chelating excess iron ions. Improvements in sleep fragmentation and locomotor activity were closely associated with the antioxidant capacity of the extract. Sleep rhythm disruption in the model group is linked to ROS-induced damage to clock neurons [56], while decreased locomotor activity results from muscle cell damage induced by lipid peroxidation [57]. The mechanism involves direct ROS scavenging, iron chelation to suppress Fenton reactions, and potential neuroprotective effects of enriched metabolites such as indole-3-lactic acid and daidzein. The superior behavioral recovery in the mixed-strain group aligns with its more diverse metabolite profile, enabling multi-target regulation. Importantly, these behavioral phenotypes provide functional evidence that ferroptosis-associated oxidative stress compromises circadian regulation and systemic performance. The recovery of sleep rhythmicity and locomotor capacity, most prominently in the LFWGC group, supports a multi-target protective effect involving antioxidant defense together with attenuation of iron-driven ROS generation. The stronger efficacy of LFWGC relative to single-strain products is consistent with a broader metabolite spectrum generated by mixed fermentation, which may enhance functional complementarity under ferroptotic stress.

In addition to the above interpretations, alternative explanations for the enhanced antioxidant performance and microbiota modulation should be explicitly considered. Acidification is an inherent feature of lactic acid fermentation and can increase phenolic solubility/bioaccessibility while also reshaping microbial communities [58,59]. Nevertheless, pH reduction alone is insufficient to account for the present findings. All fermented treatments exhibited broadly comparable acidification, yet the mixed-strain product consistently showed higher antioxidant capacity, improved bioactive retention during simulated digestion, and stronger in vivo mitigation of ferroptosis. Moreover, the persistence of antioxidant advantages after gastrointestinal simulation—where buffering partially attenuates acidity—suggests that durable molecular stabilization and metabolite-mediated interactions, rather than transient low pH, are major contributors.

A second possibility is that fermentation promotes liberation of bound phenolics in a strain-independent manner. Indeed, enzymatic hydrolysis during fermentation commonly releases matrix-bound phenolics and can elevate antioxidant indices [60]. However, fermentation can also remodel phenolic structures via microbial biotransformation, yielding derivatives with altered stability and activity [61]. In this context, the higher post-digestion retention of total phenolics/flavonoids in the mixed-strain group is more consistent with coordinated enzymatic repertoires and cross-feeding that expand metabolite diversity and favor formation of more stable derivatives and/or macromolecule-associated complexes [62,63].

Finally, microbiota shifts may reflect a general response to fermented substrates, which deliver organic acids, peptides, and other fermentable components that can broadly influence community composition. Population-scale multi-omics analyses have reported systematic associations between fermented-food intake and gut microbiome/metabolome features [64]. However, the greater recovery of diversity and stronger suppression of stress-associated opportunistic genera observed for the mixed-strain product, relative to single-strain counterparts, is compatible with a broader and more functionally diverse metabolite spectrum produced by designed or naturally interacting co-cultures [65]. Collectively, while acidification and general fermentation effects likely contribute, the integrated biochemical, digestive-stability, in vivo, and ecological patterns support a mixed-strain–specific advantage attributable to metabolic complementarity and increased metabolite complexity rather than nonspecific fermentation alone [66,67].

Although the Drosophila ferroptosis model provides a tractable whole-organism platform for rapid in vivo screening and mechanistic interrogation, its translational scope is inherently constrained by interspecies differences in gastrointestinal architecture and physiology, xenobiotic biotransformation, neuroendocrine/immune regulation, and tissue complexity [68,69]. Moreover, ferroptosis-associated phenotypes and their modulation in flies should be interpreted as hypothesis-generating evidence rather than direct support for clinical efficacy, given that key antioxidant-defense architectures and ferroptotic programs can diverge across taxa [70,71]. Wheat germ is widely consumed as a food ingredient; however, fermentation may substantially reshape the composition, bioaccessibility, and relative abundance of bioactive constituents, with potential implications for tolerability and interaction profiles [72]. Therefore, prior to human intervention, fermented wheat germ extracts should be manufactured under food-grade specifications and subjected to standardized safety evaluation in mammalian systems, including acute oral toxicity testing and repeated-dose studies with clinical chemistry and organ-function endpoints [73,74].

Fermentation also positively modulated the gut microbiota in ferroptotic Drosophila. Oxidative stress-induced dysbiosis, characterized by reduced diversity [75] and enrichment of potential pathogens like Pseudomonas, was significantly reversed by the LFWGC intervention. Recovery of microbial diversity correlated with increased levels of beneficial genera. This restoration is linked to antioxidant components in fermented wheat germ, which improve the intestinal redox microenvironment and selectively promote probiotic growth, highlighting the advantage of multi-strain metabolic synergy in remodeling intestinal microecology [76]. Specifically, reduced Chao, ACE, and Shannon indices in the model group indicate that an ROS-enriched intestinal environment constrains ecological niches and destabilizes community assembly, thereby selecting for stress-tolerant facultative taxa [77]. NMDS ordination further demonstrated a clear separation between the model and control microbiomes, whereas the LFWGC group shifted toward the control cluster, supporting a corrective restructuring of global community composition. At the genus level, enrichment of Pseudomonas and increases in opportunistic Enterobacteriaceae (e.g., Enterobacter) and Stenotrophomonas may aggravate inflammatory and endotoxin-related risk, whereas their suppression following LFWGC supplementation is consistent with alleviation of dysbiosis-associated stress [78,79]. Meanwhile, enrichment of Lactiplantibacillus and increases in Brevundimonas and Sphingomonas—genera reported to harbor enzymatic capacities for phenolic conversion and aromatic compound transformation—align with the enhanced antioxidant metabolite profile and provide a plausible microecological basis for systemic protection [80].

Metabolomic and molecular analyses elucidated the systemic mechanisms. Mixed-strain fermentation enriched a spectrum of functional metabolites while reducing harmful compounds, achieved through complementary enzyme systems across strains, a phenomenon aligned with findings that multi-strain fermentation expands metabolite diversity through metabolic synergy [81] and promotes continuous synthesis of functional components via collective pathway participation [82]. At the molecular level, the LFWGC extract upregulated the expression of Gpx4 mRNA and GPX4 protein, a key inhibitor of ferroptosis, while helping restore the expression of the iron transporter gene Tfr1. This regulation inhibits abnormal iron accumulation, suppressing the Fenton reaction substrate at its source [83], and directly enhances the scavenging efficiency of lipid peroxides [84]. Correlation analysis further supported a mechanistic network where beneficial bacterial genera and core antioxidant metabolites, such as syringic acid which can be activated via microbiota-mediated processes [85], showed strong positive correlations with antioxidant indices and ferroptosis-inhibiting targets. Conversely, pathogenic genera like Pseudomonas, known to exacerbate oxidative stress [86], were negatively correlated with these beneficial parameters. We further propose that detoxification of 4-nitrophenol and 2,6-xylidine in the mixed culture proceeds through a sequential, redox-coupled co-metabolic network enabled by strain-level metabolic complementarity and cross-feeding [87]. For 4-nitrophenol, initial NAD(P)H-dependent reduction to aminophenolic intermediates followed by aromatic hydroxylation and ring-cleavage is consistent with established microbial 4-nitrophenol catabolic logic and pathway diversity across taxa [88,89]. For 2,6-xylidine, we hypothesize an initial monooxygenase-mediated activation (N-hydroxylation and/or aromatic hydroxylation) that increases ring reactivity, followed by dioxygenation and cleavage into short-chain organic acids; such aerobic aromatic attack is commonly initiated by aromatic ring-hydroxylating dioxygenases and related oxygenases [90]. Because these xenobiotics are unlikely to serve as sole carbon sources for Lactiplantibacillus plantarum, their turnover is most parsimoniously interpreted as co-metabolism supported by exchange of growth substrates (e.g., organic acids, peptides) and enhanced intracellular redox cycling in the mixed-strain consortium, thereby sustaining the reducing power required for sequential transformations [91].

Collectively, mixed-strain fermentation by L. plantarum strains confers significant advantages over single-strain fermentation in structurally modifying wheat germ, enriching bioactive and detoxifying components, enhancing digestive stability, and systemically mitigating oxidative stress in vivo. The core mechanism lies in metabolic synergy, enabling complementary enzymatic activities, diversified metabolite synthesis, and multi-pathway regulation. These findings support the development of mixed-strain fermented wheat germ as a functional food ingredient with potential for managing oxidative stress-related conditions. Further studies are required to establish human-relevant dosing, safety, and real-world feasibility, including mammalian validation and standardized toxicology assessment.

5. Conclusions

Using lactic acid bacteria harboring the gshF gene as functional strains, this study systematically investigated the effects of single and mixed-strain fermentation on the structural modification, component transformation, and in vitro/in vivo bioactivities of wheat germ, clarifying the synergistic advantages and underlying mechanisms of mixed strain fermentation. Mixed lactic acid bacteria fermentation, relying on enzyme system complementarity and metabolic synergy, achieves effective structural remodeling of wheat germ and targeted regulation of its bioactive components, while enhancing the stability of bioactive substances during gastrointestinal digestion. This addresses the core bottlenecks limiting the practical application of natural wheat germ, including poor bioaccessibility and high oxidative vulnerability. Validation in a ferroptotic Drosophila model demonstrated that mixed strain fermented wheat germ extract exerts significant intervention effects on ferroptosis-related pathological phenotypes, which is attributed to multi-dimensional synergistic mechanisms involving gut microbiota modulation, metabolic profile optimization, and GSH/GPX4 pathway activation. Notably, the intervention efficacy of mixed strain fermentation is superior to that of single strain fermentation. In summary, mixed lactic acid bacteria fermentation provides an efficient technical approach for the high-value utilization of wheat germ. The prepared fermented wheat germ extract possesses prominent antioxidant capacity, favorable digestive stability, and high in vivo bioavailability, thereby holding substantial application potential in the dietary intervention of ferroptosis-associated diseases.

Appendix A

Table A1.

Source and identification results of the experimental strains.

Bacterial Strain	Source	Identification Result
HUCF14	Natural fermented sauerkraut	Lactiplantibacillus plantarum
GC5	Natural fermented oats	Limosilactobacillus fermentum
GC6	Natural fermented oats	Limosilactobacillus fermentum
GC8	Natural fermented oats	Lacticaseibacillus paracasei
HUCF16	Natural fermented sauerkraut	Lactiplantibacillus plantarum
HUCF11	Natural fermented sauerkraut	Lactiplantibacillus plantarum
H8	Natural fermented sauerkraut	Lacticaseibacillus paracasei
DQ2	Naturally fermented kelp	Lactiplantibacillus plantarum
DQ6	Naturally fermented kelp	Ligilactobacillus salivarius
AD7	Natural fermentation of celery	Ligilactobacillus salivarius

Table A2.

Primer Information.

	Primer Name		Sequence
PCR	gshF	F	GAATTCCATATGATGGAATTAGATGCCGTTGGTAAGGAATTG
	gshF	R	CCGCTGGATCTTCATTTTTAAACAATGCATCCAACAA
qPCR①	gpx4	F	TACCGATCAAAGTACGTGCCC
	gpx4	R	CCGTCTGGCCACAGATAGTC
qPCR②	Tfr1	F	GAGATAGAAACGGCCTGGCT
	Tfr1	R	GGCACAAATAACCAAGTGCCA

Table A3.

Drosophila experiment design.

Constituencies	Quantity	Intervene	Time
blank control(CON)	300 × 5	500 μL saline solution	15 d
Model control(Model)	300 × 5	500 μL (10 µm/mL) erastin	15 d
Intervention control(LFWG)	300 × 5	500 μL LFWG(0.2 g freeze-dried powder + 1 mL saline solution)	15 d
positive control(POS)	300 × 5	500 μ (10 µm/mL) Fer-1	15 d

Table A4.

Differing metabolites of fermented wheat germ.

Common Differential Metabolites	LFWG11 vs. CON	LFWG14 vs. CON	LFWG16 vs. CON	LFWGC vs. CON
(2E)-3-(3,4-dimethoxyphenyl)prop-2-enoicacid	—	—	—	↓
1-Glyceryllinoleate	—	—	↑	—
13(S)-HOT	—	—	—	↓
13-HODE	—	—	—	↓
2,6-Diaminohexanoicacid	↑	↑	—	—
2,6-Xylidine	↑	—	—	↓
2-Hydroxy-4-methylthiobutanoicacid	↑	↑	↑	—
2-Hydroxycaproicacid	↑	↑	↑	—
2-Hydroxyphenylalanine	↑	—	—	—
3,6,9,12,15,18-Hexaoxaicosane-1,20-diol	↑	—	—	↓
4-Nitrophenol	↑	—	—	↓
4-Pyridoxicacid	↑	↑	↓	—
5′-Guanylicacid	↑	↑	↑	↑
5-Oxo-D-proline	↑	—	—	↓
6-Methyladenosine	↑	↑	—	—
6-β-D-Glucopyranosyl-8-β-D-ribopyranosylapigenin	↑	—	—	↓
Alanyltyrosine	↑	↑	—	—
Aspartame	↑	—	—	—
Citricacid	↓	—	—	↓
D-(+)-Malicacid	↓	—	—	↓
D-Alanyl-D-alanine	↑	↑	↑	↑
D-myo-Inositol1,4-bisphosphate	↑	↑	↑	—
DL-Arginine	↑	↑	—	↑
Daidzein	↑	—	↑	↑
DeoxycholicAcid	↑	—	—	↓
Ferulicacid	—	—	—	↓
Gluconicacid	—	—	—	↓
Glucose1-phosphate	↑	↑	↑	—
Glycyl-L-leucine	↑	↑	↑	—
Glyoxylicacid	—	↑	↑	—
Guanine	↑	—	—	↑
Hexaoxyethyleneglycol	—	—	—	↓
Indole-3-lacticacid	↑	↑	↑	↑
Isethion	—	—	—	↓
L-(−)-3-Phenyllacticacid	↑	↑	↑	—
N-Formylmethionine	↑	—	↑	↑
Pimelicacid	—	—	—	↓
Pseudouridine	↑	↑	↑	↑
S-Adenosylhomocysteine	—	—	—	↓
Syringic acid	—	—	—	↓
Uricacid	↑	↑	↑	—
Valylproline	↑	↑	↑	↑
P-Acetamidophenol	↑	—	—	↑
Trans-3-Indoleacrylicacid	—	↑	↑	—
A,α-Trehalose	—	—	—	↓
A-Aspartylphenylalanine	↑	↓	↑	↑
A-Linolenicacid	—	—	—	↓
N1-Acetylspermine	—	—	—	↓
Phenylacetaldehyde	—	↑	↑	↑
Phthalicaciddipentylester	↓	↓	—	↓

Note: ↑, ascend; ↓, descend; —, zero difference.

Table A5.

Differential metabolome of Drosophila.

Metabolite	MOD vs. CON	LFWGC vs. CON	LFWGC vs. MOD
2-Hydroxy-L-Methionine	↓	↑	↓
N-Acetyl-D-Cysteine	↑	—	↑
Methionine Sulfoxide	↓	↑	↓
2-Hydroxyadenine	↑	↓	—
Phloretin 2′-O-Glucuronide	↑	—	—
Quinic Acid	↑	—	↑
Deoxycholic Acid 3-Glucuronide	↓	—	—
L-Ascorbic Acid 2-Glucoside	↑	↓	—
Coriolic Acid	↑	↓	—
Dihydrobiopterin	↑	↓	—
Syringic Acid	↑	—	—
7,8-Dihydrobiopterin	↑	↓	—
9 (Z),11 (E)-Conjugated Linoleic Acid	↓	—	↓

Note: ↑, ascend; ↓, descend; —, zero difference.

Appendix B

Figure A1.

Digestion evaluation in vitro and volcano map of differential metabolites. (A,B) ABTS free radical scavenging capacity and FRAP value in OS. (C,D) ABTS free radical scavenging capacity and FRAP value in SGF. (E,F) ABTS free radical scavenging capacity and FRAP value in SIF. (G) Correlation heatmap. (H,I) Volcano map of differential metabolites. CON indicates non-fermented wheat germ extract by lactic acid bacteria. MOD refers to the erastin modeling module, while, LFWGC refer to different samples of fermented wheat germ extracts prepared by mixed fermentation of HUCF11, 14, and 16. Data are expressed as mean ± SD. Values are mean ± SD (n = 5). Different lowercase letters indicate significant differences among groups (one-way ANOVA followed by Tukey’s post hoc test, p < 0.05). * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure A2.

Reduction rate of tannins (A) and phytic acid (B) in CON, LFWG11, LFWG14, LFWG16, and LFWGC groups. Values are mean ± SD (n = 5). Different lowercase letters indicate significant differences among groups (one-way ANOVA followed by Tukey’s post hoc test, p < 0.05).

Author Contributions

Writing—original draft, Z.L. and D.L.; Writing—review and editing, Q.K., D.F., C.H., Y.H. and Y.M. (Yiying Ma); Validation, Z.L., D.L.; Supervision, Z.L., Y.M. (Yongqiang Ma), C.H. and Q.K.; Project administration, Z.L., C.H. and Q.K.; Visualization, Z.L., D.L. and Y.M. (Yiying Ma); Funding acquisition, Z.L.; Methodology, D.L., D.F. and Q.K.; Data curation, D.L. and S.Y.; Investigation, X.W. and S.Y.; Formal analysis, X.W., Y.M. (Yongqiang Ma) and Y.H. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2025) in the National Genomics Data Center (Nucleic Acids Res 2025), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA030468) and are publicly accessible at https://ngdc.cncb.ac.cn/gsa (accessed on 24 September 2025).

Conflicts of 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.

Funding Statement

National Natural Science Foundation of China (32302051), Natural Science Foundation of Heilongjiang Province (PL2024C009), WuxiYoung Elite Scientists Sponsorship Program (No. TJXD-2024-101), Doctoral Research Start-up Support Program of Harbin University of Commerce (No. 22BQ83).