Boosting NADPH with GABA: Converting aged seeds to functional soybean sprouts Authors: Song Yu, Jiaqi Yin, Fandi Meng, Kaiqiao Hua, Keyi Zhao, Fengming Li, Tong Zhao, Yingjie Wang, Yifei Zhang, Chunyu Zhang, Lihe Yu Published: 2026 ABSTRACT Seed aging impairs soybean germination and sprout nutrition. This study investigated how γ-aminobutyric acid (GABA) revitalizes aged seeds via pentose phosphate pathway (PPP) activation. Aged soybean seeds were treated with 0.1 mM GABA, 20 μM dehydroepiandrosterone (G6PDH inhibitor), or their combination, with pharmacological inhibition confirming PPP dependency. GABA restored germination from 35% to 72%, enhanced G6PDH activity to 225% of aged control, and upregulated GmG6PD2 expression (6.8-fold). Antioxidant capacity (FRAP) recovered to 83% of fresh seed levels, with AsA/DHA and GSH/GSSG ratios restored from 2.6 to 6.3 and 3.2 to 9.2, respectively. NADPH/NADP+ ratio increased from 1.4 to 3.4. Metabolomics (3876 metabolites) revealed GABA specifically activated PPP intermediates and isoflavone biosynthesis. Notably, endogenous GABA increased ∼3-fold, producing dual-functional sprouts with elevated antioxidants and high GABA content. This study demonstrates GABA revitalizes aged seeds by activating PPP to replenish NADPH, offering a promising strategy to mitigate storage losses. Introduction Soybean (Glycine max L.) seeds are a major source of protein and edible oil, while soybean sprouts, rich in protein, dietary fiber, vitamins, and bioactive isoflavones, are increasingly valued as functional foods (Swallah et al., 2023). However, seed aging during storage deteriorates vigor, reducing germination, weakening seedlings, and lowering sprout nutritional quality (Chhabra & Singh, 2019; da Silva et al., 2023; Lin et al., 2022), leading to substantial economic loss (Chhabra & Singh, 2019; Corbineau, 2024) and decreased crop yield with increased farmers' input costs (Gaur et al., 2020). Aging triggers metabolic collapse, characterized by severe depletion of NADPH and redox imbalance (Alipour et al., 2021; Fu et al., 2026; Kalemba et al., 2021; Kurek et al., 2019; Nietzel et al., 2020; Xin et al., 2014; Yazdanpanah et al., 2019), which limits germination and seedling growth and disrupts the ascorbate–glutathione cycle, impairing antioxidant biosynthesis and the accumulation of phenolics and flavonoids in sprouts (Foyer et al., 2020; Swallah et al., 2023). Thus, restoring NADPH pools is essential for reviving aged seeds and improving the nutritional quality of resultant sprouts (Aghdam et al., 2020). The pentose phosphate pathway (PPP), in which glucose-6-phosphate dehydrogenase (G6PDH) catalyzes the rate-limiting step, is the primary cellular source of NADPH (Yang et al., 2019). PPP-derived NADPH fuels antioxidant systems and provides precursors for biosynthesis, thereby linking redox balance to primary and secondary metabolism. Although PPP's role in stress responses of mature plants and fruits is well documented (Aghdam et al., 2020), its function in restoring seed vigor and enhancing sprout nutritional quality during germination remains poorly understood. Gamma-aminobutyric acid (GABA) improves germination in aged seeds (Afshari & Seyyedi, 2020; Shakhov et al., 2024) and alleviates aging-induced deterioration in several crops (Afshari & Seyyedi, 2020; Luo et al., 2018; Shakhov et al., 2024), but whether it also promotes antioxidant metabolite accumulation in resultant sprouts remains unknown. Endogenous GABA accumulates during germination, and GABA also serves as a bioactive compound with health benefits, including blood pressure regulation and cognitive enhancement (Hou et al., 2024; Lee et al., 2023). Recent evidence indicates that GABA can stimulate glycolysis and mitochondrial respiration (Che-Othman et al., 2020; Zhu et al., 2022), suggesting an energy-metabolism component to its priming action. We therefore hypothesize that GABA revitalizes aged soybean seeds by activating PPP to replenish NADPH, thereby restoring redox balance and the biosynthetic capacity for bioactive compounds in the resultant sprouts. To test this, we combine pharmacological inhibition with physiological, biochemical, and metabolomic analyses on germinated sprouts to elucidate the mechanism by which GABA revitalizes seeds and enhances sprout functional quality. To address these knowledge gaps, the present study aims to evaluate the effects of GABA priming on germination vigor and antioxidant compound accumulation in sprouts derived from aged soybean seeds, elucidate the role of PPP activation in GABA-mediated NADPH restoration and redox homeostasis, and unravel the metabolic cascade linking PPP activation to enhanced biosynthesis of functional metabolites, thereby assessing the potential of GABA-treated aged seeds for producing dual-functional sprouts with elevated antioxidant capacity and high GABA content. Materials and methods Chemicals and reagents All chemicals and reagents used in this study were of analytical or HPLC grade and were obtained from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise stated. The main reagents included: methanol, acetonitrile, and formic acid (LC-MS grade); 2,3,5-triphenyltetrazolium chloride (TTC), γ-aminobutyric acid (GABA), dehydroepiandrosterone (DHEA), nitroblue tetrazolium (NBT), L-glutamic acid, hydrogen peroxide (H₂O₂), thiobarbituric acid (TBA), Folin-Ciocalteu reagent, aluminum chloride hexahydrate, 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB), l-ascorbic acid, reduced glutathione (GSH), oxidized glutathione (GSSG), β-nicotinamide adenine dinucleotide phosphate (NADP+/NADPH), glucose-6-phosphate (G6P), GABA assay kit, ATP assay kit, and glutathione assay kit. TRIzol reagent was purchased from Invivogen (San Diego, CA, USA). PrimeScript™ RT reagent Kit was obtained from Takara Biomedical (Shiga, Japan). SYBR Green Master Mix was from Applied Biosystems (Foster City, CA, USA). NADPH/NADP+ assay kit was supplied by Abcam (Cambridge, UK). Plant material, artificial aging, and experimental treatments Fresh, plump, and disease-free soybean (Glycine max L. cv. Heinong 84) seeds (500 g) were used in this study. Of these, 400 g were subjected to artificial aging at 40 °C and 100% relative humidity for 48 h in a controlled environment chamber to induce significant decline in germination vigor (da Silva et al., 2023; Xin et al., 2014; Zhang et al., 1995), while the remaining 100 g served as non-aged control. Following aging, seeds were allocated into five experimental groups: non-aged control (CKN), aged control (CKO), and three treatment groups comprising aged seeds treated with 20 μM DHEA—a specific glucose-6-phosphate dehydrogenase (G6PDH) inhibitor (Li et al., 2020; Liu et al., 2007; Liu et al., 2013; Wang et al., 2008; Wei et al., 2019; Yu et al., 2005), 0.1 mM GABA, and combined 0.1 mM GABA +20 μM DHEA (GABA+DHEA). GABA concentration was selected based on Afshari and Seyyedi (2020), while DHEA concentration (20 μM) was optimized through preliminary experiments: 5 μM showed insufficient inhibition, 20 μM reduced germination to ∼10%, while 50 and 100 μM completely prevented germination, indicating that 20 μM specifically inhibits G6PDH without non-specific toxicity. Seeds were soaked in respective solutions (distilled water for CKN/CKO, 0.1 mM GABA, 20 μM DHEA, or combined GABA+DHEA) for 30 min, then rinsed, transferred to germination boxes, and maintained in darkness at 25 ± 1 °C. Seeds harvested 12 h after treatment were stained with TTC for seed vigor assessment and the stained seeds were quantified spectrophotometrically. Since early germination effects were more pronounced, seedling morphology was photographed at 24 h after treatment. Germinated seeds collected on days 1, 2, and 3 were used for physiological, biochemical, and qRT-PCR analyses, while only day 3 samples were subjected to widely-targeted metabolomics analysis (Table S1). Assessment of germination characteristics and seedling growth Cellular metabolic activity of soybean sprouts was first assessed visually using the TTC reduction assay, with relative vitality ranked based on red color intensity by eye (five replicates). Uniform tissue sections were incubated in 0.5% TTC solution at 30 °C in darkness for 4 h. Viable cells reduce colorless TTC to insoluble red formazan via mitochondrial dehydrogenases. Subsequently, quantitative TTC analysis was performed to confirm the visual observations (Wang et al., 2023). For spectrophotometric quantification, stained tissues were collected and homogenized with 5 mL of 95% ethanol, heated at 80 °C for 30 min, then centrifuged at 5000 ×g for 10 min. The absorbance of the supernatant was measured at 530 nm using a microplate reader (BioTek Instruments, VT, USA). Quantitative TTC analysis was performed with three biological replicates, with the relative TTC reduction capacity expressed as absorbance per gram fresh weight (A₅₃₀/g FW). Germination was monitored daily. Seeds were considered germinated when the radicle protruded ≥2 mm, calculated according to the International Rules for Seed Testing (Allen & Alvarez, 2020) with minor modifications. Germination Percentage (GP): GP (%) = (Number of germinated seeds at final count / Total number of seeds) × 100. Germination Energy (GE): GE (%) = (Number of seeds germinated at specific time point / Total number of seeds) × 100. In this study, GE was recorded on days 1, 2, and 3 to characterize the germination kinetics (Mrđa et al., 2011). Germination Index (GI): GI = Σ(Gₜ/Dₜ), where Gₜ is the number of seeds germinated on day t, and Dₜ is the corresponding day. This index reflects the cumulative germination performance over the measurement period. Vigor Index (VI): VI = GI × S, where S is the mean radicle length (cm) measured on the same day as the GI calculation. Thus, VI integrates both germination speed and seedling growth, providing a comprehensive measure of seed vigor (Alipour et al., 2021). After 3 days of germination, ten uniform sprouts per treatment were randomly selected for morphological measurements. Hypocotyl length and radicle length were measured using a digital vernier caliper, and fresh weight was determined using an analytical balance after blotting with absorbent paper. Three biological replicates were performed. Enzyme activity assays Samples (germinating seeds/radicles) were collected at 1, 2, and 3 days, frozen in liquid nitrogen, and stored at −80 °C. Enzyme extracts were prepared in appropriate cold extraction buffers. Activities of key antioxidant and metabolic enzymes were determined spectrophotometrically. Superoxide dismutase (SOD): Activity was assayed by measuring its ability to inhibit the photochemical reduction of nitroblue tetrazolium (NBT) at 560 nm (Yan & Mao, 2021). Catalase (CAT): Activity was determined by monitoring the decomposition of H₂O₂ at 240 nm (ε = 39.4 mM−1 cm−1) (Yan et al., 2020). Ascorbate peroxidase (APX): Activity was measured by tracking the oxidation of ascorbate at 290 nm (ε = 2.8 mM−1 cm−1) (Yan & Mao, 2021). Glutathione reductase (GR): Activity was assayed by measuring NADPH oxidation at 340 nm (ε = 6.22 mM−1 cm−1) during the reduction of oxidized glutathione (GSSG) (Aghdam et al., 2020). Glucose-6-phosphate dehydrogenase (G6PDH): Activity was determined by monitoring the rate of NADP+ reduction to NADPH at 340 nm (ε = 6.22 mM−1 cm−1) (Tao et al., 2022; Yang et al., 2019). Glutamate decarboxylase (GAD): Activity was measured by quantifying GABA production (Xu & Hu 2014). The assay mixture (enzyme extract, 50 mM L-glutamic acid, pH 5.8) was incubated at 40 °C for 1 h, then stopped with 0.2 mL of 0.5 M HCl. After centrifugation, GABA in the supernatant was determined with a commercial GABA assay kit, and activity was expressed as U/g fresh weight. Gene expression of related enzyme To investigate the molecular response under different treatments, total RNA was extracted from soybean sprouts subjected to five conditions: fresh seeds (CKN), aged seeds (CKO), aged seeds treated with GABA (GABA), aged seeds treated with DHEA (DHEA), and aged seeds treated with GABA+DHEA (GABA+DHEA). Samples were collected at 1, 2, and 3 days post-germination. Total RNA was isolated using TRIzol reagent following the manufacturer's instructions, with quality assessed using NanoDrop spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA) and agarose gel electrophoresis. First-strand cDNA was synthesized from 1 μg of total RNA using a PrimeScript™ RT reagent Kit (Takara) with oligo(dT) primers. Quantitative real-time PCR (qRT-PCR) was performed to analyze the expression levels of four target genes: GmMSD2, GsoCAT4.1, PAL, and GmG6PD2. The 20 μL reaction mixture contained 10 μL of 2× SYBR Green Master Mix, 0.8 μL each of forward and reverse primers (10 μM), 2 μL of diluted cDNA, and 6.4 μL of nuclease-free water. Gene-specific primer sequences are listed in Table S2, with ACT11 as the internal reference gene (Jian et al., 2008). The thermal cycling program was: initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s, with melt curve analysis at the end. Each sample was analyzed in three biological replicates, each with three technical replicates. Relative gene expression was calculated using the 2 − ΔΔCt method, with CKN day 1 as the calibrator. Biochemical analyses in 3-day-old sprouts Three-day-old soybean sprouts from all groups were harvested, immediately frozen in liquid nitrogen, and used for comprehensive biochemical profiling. Membrane integrity (Relative electrolyte leakage, REL): Electrolyte leakage was measured according to Zhang et al. (1995) with minor modifications. Twenty seed sprout tissues (0.2 g) were immersed in 20 mL deionized water, and electrical conductivity was measured using a conductivity meter (DDS-307 A, Shanghai INESA Scientific Instrument Co., Ltd.) at the initial reading (C₁) and after boiling for 10 min (C₂). REL was calculated as: REL (%) = (C₁/C₂) × 100. Hydrogen Peroxide (H₂O₂) Content: Measured using the titanium sulfate method (Yan et al., 2020). Lipid Peroxidation: Estimated as malondialdehyde (MDA) content using the thiobarbituric acid (TBA) reaction (Yan & Mao, 2021). Total Phenolic Content: Determined using the Folin-Ciocalteu reagent (Zhao et al., 2021). Total Flavonoid Content: Measured using the aluminum chloride colorimetric method (Zhao et al., 2021). GABA Content: Quantified using a commercial GABA assay kit based on an enzymatic method (Ding et al., 2016; Luo et al., 2018). Adenylate Nucleotides (ATP, ADP, AMP) and Energy Charge: Extracted with perchloric acid and analyzed by high-performance liquid chromatography (HPLC). The energy charge was calculated as (ATP + 0.5ADP) / (ATP + ADP + AMP) (Aghdam & Alikhani-Koupaei, 2021). Total Antioxidant Capacity: Assessed using the Ferric Reducing Antioxidant Power (FRAP) assay (Hu et al., 2023). Ascorbate Pool: Total ascorbate (AsA + DHA) and reduced ascorbate (AsA) were measured spectrophotometrically. Dehydroascorbate (DHA) was calculated by difference, and the AsA/DHA ratio was determined (Wang et al., 2025). Glutathione Pool: Total glutathione (GSH + GSSG) and reduced glutathione (GSH) were quantified enzymatically using GR and DTNB. The GSSG content and GSH/GSSG ratio were calculated (Aghdam et al., 2020). NADPH Pool: NADPH and NADP+ were extracted and quantified using a commercial enzymatic cycling assay kit. The NADPH/NADP+ ratio was calculated (Aghdam et al., 2020; Tao et al., 2022). Widely targeted metabolomic profiling: Extraction, analysis, and data processing Fresh soybean sprouts from five treatment groups (CKN, CKO, DHEA, GABA, and GABA+DHEA; Table S1) were collected, immediately frozen in liquid nitrogen, and lyophilized. Three biological replicates were performed for each treatment group. Approximately 100 mg of freeze-dried powder was extracted using pre-cooled methanol/water (70:30, v/v) following Kang et al. (2020) and Zhang et al. (2023). Samples were vortexed, subjected to low-temperature ultrasonication for 30 min, incubated at 4 °C overnight, then centrifuged. The extraction was repeated, and the combined supernatants were filtered through a 0.22 μm membrane, concentrated, and reconstituted in methanol/water for LC-MS/MS analysis. Quality control (QC) samples were prepared by pooling equal aliquots from all experimental samples and injected every 10 samples to monitor the stability of the LC-MS system. The analysis was conducted using a UPLC system (Shimadzu, Kyoto, Japan) coupled with a triple quadrupole-linear ion trap mass spectrometer (QTRAP® 6500+, SCIEX, Framingham, MA, USA) following the platform of Baimaike Biotechnology Co., Ltd. (Cao et al., 2023; Kang et al., 2020). Chromatographic separation was achieved on an Agilent SB-C18 column (2.1 × 100 mm, 1.8 μm; Agilent Technologies, Santa Clara, CA, USA) at 40 °C using water/acetonitrile both with 0.1% formic acid as mobile phases with gradient elution. MS/MS analysis was performed in positive and negative ion modes using scheduled multiple reaction monitoring (sMRM). Ion source parameters were: ESI temperature, 550 °C; ion spray voltage, ±5500 V; curtain gas, 35 psi. Data acquisition used Analyst® 1.6.3 software (SCIEX). Metabolite identification was performed by comparing mass spectrum (MS2 fragments, Q3) and retention time (RT) with our self-built database containing curated standards (Chen et al., 2013), with a match score threshold >80%. Raw MS data were processed using SCIEX OS software (version 2.0). Metabolites detected in ≥80% of samples within a group were retained, and missing values were imputed with the minimum value. Multivariate statistical analysis was performed as described by Zhang et al. (2023) and Cao et al. (2023): PCA was conducted to assess overall metabolic variation, followed by OPLS-DA to maximize group separation. Differentially accumulated metabolites (DAMs) were defined with VIP >1.0, fold change ≥2 or ≤ 0.5, and p-value <0.05. DAMs were annotated and mapped onto KEGG pathways to elucidate metabolic shifts. Statistical analysis All experiments were performed with at least three independent biological replicates unless otherwise specified. Data are presented as mean ± standard deviation (SD). Statistical significance among treatments was analyzed using two-way analysis of variance (ANOVA) followed by Duncan's multiple range test at p < 0.05. Statistical analyses were performed using SPSS software (version 21.0). Results Integrated phenotypic assessment The five treatments generated a clear gradient in both cellular viability and sprout morphology: CKN > GABA > GABA+DHEA ≈ CKO > DHEA. Cellular metabolic vitality was assessed using the TTC reduction assay (Fig. 1A). CKN exhibited the deepest red coloration, followed by GABA, GABA+DHEA, CKO, and DHEA. Quantitative TTC analysis (Fig. 1B) showed CKN had the highest absorbance (0.81 ± 0.05 A₅₃₀/g FW), while aging (CKO) reduced it to 45.7% of fresh seed levels. GABA treatment restored TTC reduction to 79.0% of CKN (73.0% recovery relative to CKO), while DHEA suppressed it to only 22.2% of CKN, with the combined treatment showing an intermediate value of 50.6%.Fig. 1Phenotypic assessment of soybean sprouts under different seed treatments. (A) TTC staining at 12 h and seedling morphology at 24 h after treatment. (B) TTC reduction capacity (A₅₃₀/g FW) at 12 h. (C) Hypocotyl length at 3 d. (D) Radicle length at 3 d. (E) Fresh weight at 3 d. Data are means ± SD (n = 3 biological replicates). Different letters indicate significant differences among treatments (P < 0.05). CKN, fresh seeds; CKO, aged seeds; GABA, aged seeds +0.1 mM GABA; DHEA, aged seeds +20 μM DHEA; GABA+DHEA, aged seeds + combined GABA and DHEA.Fig. 1 Phenotypic assessment of soybean sprouts under different seed treatments. (A) TTC staining at 12 h and seedling morphology at 24 h after treatment. (B) TTC reduction capacity (A₅₃₀/g FW) at 12 h. (C) Hypocotyl length at 3 d. (D) Radicle length at 3 d. (E) Fresh weight at 3 d. Data are means ± SD (n = 3 biological replicates). Different letters indicate significant differences among treatments (P < 0.05). CKN, fresh seeds; CKO, aged seeds; GABA, aged seeds +0.1 mM GABA; DHEA, aged seeds +20 μM DHEA; GABA+DHEA, aged seeds + combined GABA and DHEA. Morphological observation revealed that CKN sprouts exhibited robust hypocotyls and radicles, while CKO displayed shortened and abnormal sprouts; GABA substantially improved sprout morphology compared to CKO, whereas DHEA caused severe deformities, with GABA+DHEA showing intermediate characteristics (Fig. 1A). Quantitative parameters showed similar trends (Fig. 1C-E). CKN displayed the highest hypocotyl length (4.9 ± 0.3 cm), radicle length (4.5 ± 0.2 cm), and fresh weight (1.68 ± 0.12 g). Aging (CKO) significantly reduced these parameters to 1.9 cm, 1.7 cm, and 0.72 g, respectively. GABA substantially restored growth to 3.8 cm, 3.2 cm, and 1.25 g. DHEA caused the most severe inhibition (1.1 cm, 0.9 cm, 0.43 g), while GABA+DHEA showed intermediate values close to CKO (2.1 cm, 1.3 cm, 0.77 g). Quantitative germination dynamics Germination parameters are presented in Fig. 2.Fig. 2Seed germination and vigor indices. Germination percentage (A), germination energy (B), germination index (C), and vigor index (D) in soybean seeds after treatment for 1, 2 and 3 days. VI was calculated as GI × radicle length (cm) for each corresponding day. Data were analyzed using SPSS 21.0 software. Each treatment consisted of three biological replicates. Different letters above the bars indicate significant differences (P < 0.05) among both treatment groups and days.Fig. 2 Seed germination and vigor indices. Germination percentage (A), germination energy (B), germination index (C), and vigor index (D) in soybean seeds after treatment for 1, 2 and 3 days. VI was calculated as GI × radicle length (cm) for each corresponding day. Data were analyzed using SPSS 21.0 software. Each treatment consisted of three biological replicates. Different letters above the bars indicate significant differences (P < 0.05) among both treatment groups and days. Germination percentage (GP): Fresh seeds (CKN) achieved 95% germination by day 3. Aged seeds (CKO) showed severe viability loss, dropping to 35%. GABA treatment restored germination to 72%, representing approximately 76% of fresh seed performance. DHEA treatment alone further suppressed germination to 11%. The combined GABA+DHEA treatment resulted in 49%, demonstrating that DHEA counteracted GABA's beneficial effect by approximately 32 percentage points. Germination energy (GE): CKN reached 85% GE by day 2 and stabilized. CKO exhibited drastically low GE (25%). GABA treatment increased GE to 60%, restoring it to about 71% of the CKN level. DHEA alone severely inhibited this process, resulting in only 8% GE. The combined group showed 35% GE, approximately 42% lower than the GABA-only group. Germination index (GI) and vigor index (VI): Both parameters increased progressively over the 3-day germination period (Fig. 2C-D), reflecting cumulative germination and radicle growth. For GI (Fig. 2C): CKN showed values of 55, 83, and 95 on days 1, 2, and 3, respectively. CKO drastically reduced GI to 8, 20, and 35. GABA elevated GI to 20, 54, and 72. DHEA alone kept GI very low (2, 5, and 11), while the combined treatment showed intermediate values (12, 31, and 49). Radicle length measurements were: CKN (1.1, 2.7, and 4.5 cm), CKO (0.4, 0.9, and 1.7 cm), GABA (0.7, 1.9, and 3.2 cm), DHEA (0.2, 0.5, and 0.9 cm), and GABA+DHEA (0.3, 0.7, and 1.3 cm) on days 1, 2, and 3, respectively. For VI (Fig. 2D), calculated as GI × radicle length, CKN showed values of 60.5, 224.1, and 427.5 cm. CKO had very low VI (3.2, 18.0, and 59.5 cm), and GABA improved VI to 14.0, 102.6, and 230.4 cm. DHEA alone nearly abolished vigor (0.4, 2.5, and 9.9 cm), whereas the combined treatment gave intermediate values (3.6, 21.7, and 63.7 cm). Complete raw data are presented in Table S3. Enzyme activities in soybean sprouts Key enzyme activities across the 3-day germination period are shown in Fig. 3.Fig. 3Antioxidant enzyme, G6PDH and GAD activities. Activities of SOD (A), CAT (B), APX (C), GR (D), G6PDH (E) and GAD (F) in soybean sprouts after treatment for 1, 2 and 3 days. Data were analyzed using SPSS 21.0 software. Each treatment consisted of three biological replicates. Different letters above the bars indicate significant differences (P < 0.05) among both treatment groups and days.Fig. 3 Antioxidant enzyme, G6PDH and GAD activities. Activities of SOD (A), CAT (B), APX (C), GR (D), G6PDH (E) and GAD (F) in soybean sprouts after treatment for 1, 2 and 3 days. Data were analyzed using SPSS 21.0 software. Each treatment consisted of three biological replicates. Different letters above the bars indicate significant differences (P < 0.05) among both treatment groups and days. SOD activity (Fig. 3A): CKN showed the highest activity, reaching 32 U/g FW by day 3. Aging (CKO) reduced SOD to 22 U/g FW. GABA treatment restored and exceeded the fresh seed level to 38 U/g FW. DHEA suppressed activity to 16 U/g FW. The combined treatment resulted in 30 U/g FW. CAT activity (Fig. 3B): CKN achieved 22.3 U/g FW by day 3. Aging caused a severe decline to 12.3 U/g FW. GABA boosted CAT to 25.4 U/g FW. DHEA alone resulted in extremely low activity (7.1 U/g FW). The combination yielded 17.9 U/g FW. APX activity (Fig. 3C): CKN showed the highest activity at 14.2 U/g FW. Aged seeds exhibited markedly reduced activity (6.2 U/g FW). GABA increased APX to 12.9 U/g FW. DHEA inhibited activity to 3.5 U/g FW. The co-treatment yielded 9.5 U/g FW. GR activity (Fig. 3D): CKN showed the highest activity (9.5 U/g FW). Aging reduced it to 3.8 U/g FW. GABA effectively restored GR to 9 U/g FW. DHEA alone resulted in minimal activity (2.5 U/g FW). The combination led to 6.5 U/g FW. G6PDH activity (Fig. 3E): This enzyme showed the most dramatic response. In CKN, G6PDH activity increased progressively: 15.3, 18.1, and 20.5 U/g FW on days 1, 2, and 3, respectively. CKO markedly reduced activity to only 32.0%, 37.6%, and 38.5% of fresh seed levels (4.9, 6.8, and 7.9 U/g FW). GABA treatment elevated activity to 210.2%, 222.1%, and 225.3% of the aged control (10.3, 15.1, and 17.8 U/g FW)—the most pronounced restoration among all parameters. DHEA treatment resulted in the lowest activity, with only 24.5%, 45.6%, and 59.5% of the aged control (1.2, 3.1, and 4.7 U/g FW). The combined treatment showed intermediate values of 132.7%, 144.1%, and 153.2% of the aged control (6.5, 9.8, and 12.1 U/g FW). GAD activity (Fig. 3F): Consistent with endogenous GABA accumulation, CKN showed progressive increase, reaching 8.5 U/g FW by day 3. Aging suppressed GAD to 3.2 U/g FW (37.6% of CKN). GABA substantially restored it to 7.1 U/g FW (83.5% of CKN). DHEA inhibited to 2.1 U/g FW. The combined treatment showed intermediate values (4.8 U/g FW). Peroxide and antioxidants in soybean sprouts All biochemical parameters were measured on day 3, with quantitative data summarized in Table 1.Table 1Effects of GABA and DHEA on oxidative stress, antioxidant capacity, redox status, and energy metabolism in 3-day-old sprouts from aged soybean seeds. CKN: Fresh seeds (control); CKO: Aged seeds; GABA: Aged seeds +0.1 mM GABA; DHEA: Aged seeds +20 μM DHEA (G6PDH inhibitor); GABA+DHEA: Combined treatment. Data represent mean ± SD (n = 3). Different letters within a row indicate significant differences (P < 0.05). Abbreviations: FW, fresh weight; MDA, malondialdehyde; REL, relative electrolyte leakage; TPC, total phenolic content; TFC, total flavonoid content; GAE, gallic acid equivalent; RE, rutin equivalent; FRAP, ferric reducing antioxidant power; AsA/DHA, ascorbate/dehydroascorbate ratio; GSH/GSSG, glutathione redox ratio; NADPH/NADP+, nicotinamide adenine dinucleotide phosphate ratio; ATP, adenosine triphosphate; EC, energy charge.ParameterCKN (fresh)CKO (old)GABADHEAGABA+DHEAH₂O₂ (μmol/g FW)1.5 ± 0.3e3.8 ± 0.2b2.1 ± 0.4d4.3 ± 0.2a3.1 ± 0.2cTBARS (μmol MDA/g FW)8.7 ± 1.2d26.3 ± 1.8b15.4 ± 2.1c31.4 ± 1.3a22.5 ± 2.6bREL (%)12.6 ± 1.4e47.5 ± 3.4b27.7 ± 2.5d61.7 ± 4.6a39.8 ± 3.1cTPC (mg GAE/g FW)4.7 ± 0.2a2.1 ± 0.2d3.9 ± 0.3b1.4 ± 0.2e2.9 ± 0.2cTFC (mg RE/g FW)2.3 ± 0.2a1.1 ± 0.1d1.8 ± 0.2b0.7 ± 0.1e1.4 ± 0.1cGABA (μg/g FW)27.4 ± 3.2c16.8 ± 2.1d83.1 ± 5.3a11.5 ± 1.4e48.9 ± 4.4bFRAP (μmol Fe2+/g FW)125.6 ± 9.553.2 ± 5.4103.7 ± 7.532.9 ± 4.271.6 ± 5.8AsA + DHA (μmol/g FW)3.4 ± 0.2a1.4 ± 0.1d2.7 ± 0.2b1.1 ± 0.1e2.1 ± 0.2cAsA/DHA Ratio8.3 ± 0.5a2.6 ± 0.3d6.3 ± 0.4b1.4 ± 0.1e4.2 ± 0.2cGSH + GSSG (nmol/g FW)356 ± 22a154 ± 14d291 ± 25b104 ± 11e226 ± 17cGSH/GSSG Ratio12.5 ± 0.9a3.2 ± 0.4d9.2 ± 0.7b1.5 ± 0.3e5.4 ± 0.5cNADPH+NADP (nmol/g FW)86.4 ± 5.2a41.8 ± 3.8d73.6 ± 6.5b26.8 ± 3.1e57.3 ± 4.6cNADPH/NADP Ratio4.4 ± 0.3a1.4 ± 0.1d3.4 ± 0.3b0.8 ± 0.1e2.3 ± 0.2cATP (nmol/g FW)194.5 ± 13.6a71.7 ± 8.2d148.5 ± 10.5b38.6 ± 5.6e103.7 ± 9.6cEnergy Charge (EC)0.88 ± 0.02a0.59 ± 0.03d0.79 ± 0.02b0.42 ± 0.04e0.71 ± 0.02c Table 1Effects of GABA and DHEA on oxidative stress, antioxidant capacity, redox status, and energy metabolism in 3-day-old sprouts from aged soybean seeds. CKN: Fresh seeds (control); CKO: Aged seeds; GABA: Aged seeds +0.1 mM GABA; DHEA: Aged seeds +20 μM DHEA (G6PDH inhibitor); GABA+DHEA: Combined treatment. Data represent mean ± SD (n = 3). Different letters within a row indicate significant differences (P < 0.05). Abbreviations: FW, fresh weight; MDA, malondialdehyde; REL, relative electrolyte leakage; TPC, total phenolic content; TFC, total flavonoid content; GAE, gallic acid equivalent; RE, rutin equivalent; FRAP, ferric reducing antioxidant power; AsA/DHA, ascorbate/dehydroascorbate ratio; GSH/GSSG, glutathione redox ratio; NADPH/NADP+, nicotinamide adenine dinucleotide phosphate ratio; ATP, adenosine triphosphate; EC, energy charge.ParameterCKN (fresh)CKO (old)GABADHEAGABA+DHEAH₂O₂ (μmol/g FW)1.5 ± 0.3e3.8 ± 0.2b2.1 ± 0.4d4.3 ± 0.2a3.1 ± 0.2cTBARS (μmol MDA/g FW)8.7 ± 1.2d26.3 ± 1.8b15.4 ± 2.1c31.4 ± 1.3a22.5 ± 2.6bREL (%)12.6 ± 1.4e47.5 ± 3.4b27.7 ± 2.5d61.7 ± 4.6a39.8 ± 3.1cTPC (mg GAE/g FW)4.7 ± 0.2a2.1 ± 0.2d3.9 ± 0.3b1.4 ± 0.2e2.9 ± 0.2cTFC (mg RE/g FW)2.3 ± 0.2a1.1 ± 0.1d1.8 ± 0.2b0.7 ± 0.1e1.4 ± 0.1cGABA (μg/g FW)27.4 ± 3.2c16.8 ± 2.1d83.1 ± 5.3a11.5 ± 1.4e48.9 ± 4.4bFRAP (μmol Fe2+/g FW)125.6 ± 9.553.2 ± 5.4103.7 ± 7.532.9 ± 4.271.6 ± 5.8AsA + DHA (μmol/g FW)3.4 ± 0.2a1.4 ± 0.1d2.7 ± 0.2b1.1 ± 0.1e2.1 ± 0.2cAsA/DHA Ratio8.3 ± 0.5a2.6 ± 0.3d6.3 ± 0.4b1.4 ± 0.1e4.2 ± 0.2cGSH + GSSG (nmol/g FW)356 ± 22a154 ± 14d291 ± 25b104 ± 11e226 ± 17cGSH/GSSG Ratio12.5 ± 0.9a3.2 ± 0.4d9.2 ± 0.7b1.5 ± 0.3e5.4 ± 0.5cNADPH+NADP (nmol/g FW)86.4 ± 5.2a41.8 ± 3.8d73.6 ± 6.5b26.8 ± 3.1e57.3 ± 4.6cNADPH/NADP Ratio4.4 ± 0.3a1.4 ± 0.1d3.4 ± 0.3b0.8 ± 0.1e2.3 ± 0.2cATP (nmol/g FW)194.5 ± 13.6a71.7 ± 8.2d148.5 ± 10.5b38.6 ± 5.6e103.7 ± 9.6cEnergy Charge (EC)0.88 ± 0.02a0.59 ± 0.03d0.79 ± 0.02b0.42 ± 0.04e0.71 ± 0.02c Effects of GABA and DHEA on oxidative stress, antioxidant capacity, redox status, and energy metabolism in 3-day-old sprouts from aged soybean seeds. CKN: Fresh seeds (control); CKO: Aged seeds; GABA: Aged seeds +0.1 mM GABA; DHEA: Aged seeds +20 μM DHEA (G6PDH inhibitor); GABA+DHEA: Combined treatment. Data represent mean ± SD (n = 3). Different letters within a row indicate significant differences (P < 0.05). Abbreviations: FW, fresh weight; MDA, malondialdehyde; REL, relative electrolyte leakage; TPC, total phenolic content; TFC, total flavonoid content; GAE, gallic acid equivalent; RE, rutin equivalent; FRAP, ferric reducing antioxidant power; AsA/DHA, ascorbate/dehydroascorbate ratio; GSH/GSSG, glutathione redox ratio; NADPH/NADP+, nicotinamide adenine dinucleotide phosphate ratio; ATP, adenosine triphosphate; EC, energy charge. Oxidative stress markers: Aging dramatically increased H₂O₂, TBARS, and REL. GABA substantially alleviated these markers, while DHEA further exacerbated oxidative damage. Antioxidant metabolites: Aging reduced total phenolic content (TPC) and flavonoids (TFC). GABA substantially recovered both parameters. Notably, endogenous GABA accumulated to approximately 303% of CKN levels—a 3-fold increase—in GABA-treated sprouts. DHEA suppressed both TPC and TFC further. Redox status: The AsA/DHA ratio dropped from 8.3 to 2.6, GSH/GSSG from 12.5 to 3.2, and NADPH/NADP+ from 4.4 to 1.4 due to aging. GABA substantially restored these ratios to 6.3, 9.2, and 3.4, respectively, while DHEA induced extreme oxidation. Energy status: ATP content declined to 37% of CKN, with energy charge dropping from 0.88 to 0.59. GABA restored ATP to 76% and EC to 0.79, while DHEA caused further collapse. Gene expression in soybean sprouts qRT-PCR analysis revealed dynamic expression patterns of four key genes over the 3-day germination period (Fig. 4).Fig. 4Transcriptional profiles of antioxidant and G6PDH-related genes. Gene expression of GmMSD2 (A), GsoCAT4.1 (B), GmG6PD2 (C), and PAL (D) in soybean sprouts after treatment for 1, 2 and 3 days. Data were analyzed using SPSS 21.0 software. Each treatment consisted of three biological replicates. Different letters above the bars indicate significant differences (P < 0.05) among both treatment groups and days.Fig. 4 Transcriptional profiles of antioxidant and G6PDH-related genes. Gene expression of GmMSD2 (A), GsoCAT4.1 (B), GmG6PD2 (C), and PAL (D) in soybean sprouts after treatment for 1, 2 and 3 days. Data were analyzed using SPSS 21.0 software. Each treatment consisted of three biological replicates. Different letters above the bars indicate significant differences (P < 0.05) among both treatment groups and days. Compared to CKN, aging (CKO) consistently downregulated all four genes throughout germination. GABA treatment markedly upregulated: GmMSD2 (7-fold), GsoCAT4.1 (7.7-fold), GmG6PD2 (12.4-fold), and PAL (5.4-fold) relative to CKO, with peak expression on day 2. DHEA further suppressed expression below CKO levels. The combined treatment showed expression levels higher than the DHEA group but significantly lower than the GABA group across all time points. Comprehensive metabolomic profiling of soybean sprouts Metabolic characterization Instrument performance was validated by consistent signal detection and overlapping TIC chromatograms (Fig. S1). A total of 3876 compounds were identified across 18 biochemical classes, with lipids (661), ketones/aldehydes/esters (586), terpenoids (462), and sugars (308) being most abundant. Quantitative analysis revealed lipids as the dominant class (17.06%), followed by ketones/aldehydes/esters (15.12%) and terpenoids (11.92%). Multivariate statistical analysis PCA and OPLS-DA clearly separated the five treatment groups (Figs. S2-S3). PC1 and PC2 explained 40.93% and 17.29% of total variance, respectively. Biological replicates clustered tightly within groups, confirming substantial metabolic differences among treatments. Analysis of differentially abundant metabolites Volcano plot and heatmap analyses identified: 886 DAMs for CKO vs. CKN, 204 for GABA vs. CKO, 261 for DHEA vs. CKO, and 399 for GABA+DHEA vs. CKO (Fig. S4; Table 2). Heatmap analysis revealed distinct metabolic patterns across treatment groups (Fig. 5). Key metabolic shifts are summarized below:Table 2Top 30 differentially accumulated metabolites in soybean sprouts from GABA-treated aged seeds (GABA) compared to aged control (CKO) after 3 days of germination, identified by widely targeted metabolomics (LC-MS/MS). Metabolites are ranked by Fold Change (FC). VIP > 1.0 was used as the threshold for variable importance in the projection. Class I: Primary chemical class.NAMECLASSFCVIPIsoliquiritinFlavonoids/Chalcones17,609.571.1899Methyl DeacetylasperulosidateTerpenoids/Iridoid2888.761.12472-(alpha-L-rhamnopyranosyloxy)-benzylamineSugars/Other727.421.1893cis-Zeatin ribosideNucleotides/Purine Nucleosides333.931.1835sedumoside A3Lipid/Fatty Acyls309.321.18932-Azaspiro[4.5]dec-8-ene-1,7-dione,4,6-dihydroxy-2-methoxy-3-methylene-10-propylidene-Others/Other201.601.1905chlorantene BLipid/Isoprene lipid158.321.1903(5S,6S)-(+)-5-hydroxy-4,6-dimethoxy-6-trans-styryl-5,6-dihydro-2H-pyran-2-onKetones, Aldehydes, esters/Polyketide119.311.1902MarginatoxinLignans/Other74.931.17562,6,10,14,34-Hexatriacontapentaene-1,19,23,27,31-pentolAlcohols/Other73.151.1339Didemnin A (9CI)Organic Acid/Other49.091.1899Dihydrocyclomicrophyllidin-ATerpenoids/Triterpenoids29.961.1904Lansioside BTerpenoids/Diterpene15.901.1902(3-carboxymethylindol-1yl)-1-deoxy-beta-D-glucopyranosideNucleotides/Other14.081.1904abiesadine VLipid/Isoprene lipid13.671.1900ptesculentosideSugars/Glycosides11.521.1902N-Fructoryl cysteinylalanine + C2H5SAmino acids/Peptide12.611.1905Oxacyclooctacosa-17,19,21,23,25-pentaen-2-oneKetones, Aldehydes, esters/Other11.221.1901(6S)-6-(b-D-glucopyranosyloxy)docosanoic acidLipid/Fatty Acids2.941.18412-hydroxy-7,8-dihydro-beta-ionolTerpenoids/Sesquiterpenoid0.22371.1892Hygric AcidOrganic Acid/Carboxylic Acids0.21381.1857LysoPE 16–0(2n isomer)Lipid/LPE0.07731.1883Curculigoside BSugars/Glycosides0.07111.17441,15-Dioxacyclooctacosane-2,16-dioneKetones, Aldehydes, esters/Other0.06451.1889tert-Butyl 3,3-difluoro-5-hydroxypiperidine-1-carboxylateKetones, Aldehydes, esters/Other0.08411.1878Picrasan-21-oic acid,15-[[(2E)-4-(acetyloxy)-3,4-dimethyl-1-oxo-2-pentenyl]oxy]-13,20-epoxy-1,11,12-trihydroxy-2,16-dioxo-,methyl esterKetones, Aldehydes, esters/Other0.0031031.1861(4S,7R,8S,9S,13Z,16S(E))-4,8-dihydroxy-5,5,7,9,13-pentamethyl-16-[1-methyl-2-(2-methyl-oxazol-4-yl)-vinyl]-oxacyclohexadec-13-ene-2,6-dioneKetones, Aldehydes, esters/Polyketide0.0062911.1640Naphthalene,1-butyl-1,2,3,4-tetrahydro-4-methyl-Others/Naphthalenes0.0008461.1897HispidulinFlavonoids/Other2.34E-051.1850IsoformononetinFlavonoids/Isoflavonoids1.12E-051.1877Fig. 5Heatmap of the differentially abundant metabolites. This heatmap includes (A) CKO vs. CKN group, (B) DHEA vs. CKO group, (C) GABA vs. CKO group, (D) GABA+DHEA vs. CKO group, (E) GABA+DHEA vs. DHEA group, and (F) GABA+DHEA vs. GABA group. Note: Abscissa: sample name; Ordinate: differentially abundant metabolites. Different colors in the heatmap represent the normalized values for the relative content of the differentially abundant metabolites; red dots represent upregulated metabolites; and green dots represent downregulated metabolites. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)Fig. 5 Table 2Top 30 differentially accumulated metabolites in soybean sprouts from GABA-treated aged seeds (GABA) compared to aged control (CKO) after 3 days of germination, identified by widely targeted metabolomics (LC-MS/MS). Metabolites are ranked by Fold Change (FC). VIP > 1.0 was used as the threshold for variable importance in the projection. Class I: Primary chemical class.NAMECLASSFCVIPIsoliquiritinFlavonoids/Chalcones17,609.571.1899Methyl DeacetylasperulosidateTerpenoids/Iridoid2888.761.12472-(alpha-L-rhamnopyranosyloxy)-benzylamineSugars/Other727.421.1893cis-Zeatin ribosideNucleotides/Purine Nucleosides333.931.1835sedumoside A3Lipid/Fatty Acyls309.321.18932-Azaspiro[4.5]dec-8-ene-1,7-dione,4,6-dihydroxy-2-methoxy-3-methylene-10-propylidene-Others/Other201.601.1905chlorantene BLipid/Isoprene lipid158.321.1903(5S,6S)-(+)-5-hydroxy-4,6-dimethoxy-6-trans-styryl-5,6-dihydro-2H-pyran-2-onKetones, Aldehydes, esters/Polyketide119.311.1902MarginatoxinLignans/Other74.931.17562,6,10,14,34-Hexatriacontapentaene-1,19,23,27,31-pentolAlcohols/Other73.151.1339Didemnin A (9CI)Organic Acid/Other49.091.1899Dihydrocyclomicrophyllidin-ATerpenoids/Triterpenoids29.961.1904Lansioside BTerpenoids/Diterpene15.901.1902(3-carboxymethylindol-1yl)-1-deoxy-beta-D-glucopyranosideNucleotides/Other14.081.1904abiesadine VLipid/Isoprene lipid13.671.1900ptesculentosideSugars/Glycosides11.521.1902N-Fructoryl cysteinylalanine + C2H5SAmino acids/Peptide12.611.1905Oxacyclooctacosa-17,19,21,23,25-pentaen-2-oneKetones, Aldehydes, esters/Other11.221.1901(6S)-6-(b-D-glucopyranosyloxy)docosanoic acidLipid/Fatty Acids2.941.18412-hydroxy-7,8-dihydro-beta-ionolTerpenoids/Sesquiterpenoid0.22371.1892Hygric AcidOrganic Acid/Carboxylic Acids0.21381.1857LysoPE 16–0(2n isomer)Lipid/LPE0.07731.1883Curculigoside BSugars/Glycosides0.07111.17441,15-Dioxacyclooctacosane-2,16-dioneKetones, Aldehydes, esters/Other0.06451.1889tert-Butyl 3,3-difluoro-5-hydroxypiperidine-1-carboxylateKetones, Aldehydes, esters/Other0.08411.1878Picrasan-21-oic acid,15-[[(2E)-4-(acetyloxy)-3,4-dimethyl-1-oxo-2-pentenyl]oxy]-13,20-epoxy-1,11,12-trihydroxy-2,16-dioxo-,methyl esterKetones, Aldehydes, esters/Other0.0031031.1861(4S,7R,8S,9S,13Z,16S(E))-4,8-dihydroxy-5,5,7,9,13-pentamethyl-16-[1-methyl-2-(2-methyl-oxazol-4-yl)-vinyl]-oxacyclohexadec-13-ene-2,6-dioneKetones, Aldehydes, esters/Polyketide0.0062911.1640Naphthalene,1-butyl-1,2,3,4-tetrahydro-4-methyl-Others/Naphthalenes0.0008461.1897HispidulinFlavonoids/Other2.34E-051.1850IsoformononetinFlavonoids/Isoflavonoids1.12E-051.1877 Top 30 differentially accumulated metabolites in soybean sprouts from GABA-treated aged seeds (GABA) compared to aged control (CKO) after 3 days of germination, identified by widely targeted metabolomics (LC-MS/MS). Metabolites are ranked by Fold Change (FC). VIP > 1.0 was used as the threshold for variable importance in the projection. Class I: Primary chemical class. Heatmap of the differentially abundant metabolites. This heatmap includes (A) CKO vs. CKN group, (B) DHEA vs. CKO group, (C) GABA vs. CKO group, (D) GABA+DHEA vs. CKO group, (E) GABA+DHEA vs. DHEA group, and (F) GABA+DHEA vs. GABA group. Note: Abscissa: sample name; Ordinate: differentially abundant metabolites. Different colors in the heatmap represent the normalized values for the relative content of the differentially abundant metabolites; red dots represent upregulated metabolites; and green dots represent downregulated metabolites. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Aging effects (CKO vs. CKN): Severe ATP depletion (FC ≈ 8.5 × 10−5), reduced sugar alcohols (sorbitol FC = 0.36, mannitol FC = 0.36) and phospholipids (PE 38:2 FC = 0.40, PI 18:2 FC = 0.20); increased salicylic acid (FC = 19.23) and isoflavones (genistein FC = 2.53, formononetin FC = 2.50). GABA restoration (GABA vs. CKO): Notable ATP recovery (FC = 2.11) and marked upregulation of antioxidant flavonoids—particularly isoliquiritin (FC = 17,609.57) and bavachalcone (FC = 2.96). Key PPP intermediates also accumulated, such as (S)-2-Hydroxybutyric Acid (FC = 68,860.45). DHEA inhibition (DHEA vs. CKO): Accumulation of sugar alcohols (sorbitol, dulcitol, d-mannitol, FC > 2) but severe depletion of flavonoid glycosides (Kaempferol-3-O-α-L-rhamnoside FC = 0.34, (−)-Epigallocatechin FC ≈ 6.8 × 10−5). Increased fatty acids (2-Hydroxyhexadecanoic acid FC = 8.25) suggested membrane peroxidation. Combined treatment (GABA+DHEA vs. CKO): ATP accumulation (FC = 3.14) but PPP-linked flavonoids remained suppressed. Interaction analyses: GABA rescued PPP-related flavonoids when combined with DHEA (vs. DHEA alone): Isoliquiritin (FC = 16,121), bavachalcone (FC = 2.56), 1-Oleoylglycerophosphocholine (FC = 3.64). However, compared to GABA alone, membrane damage markers increased (2-Hydroxyhexadecanoic acid FC = 2.73, luteolin FC = 3.07), confirming DHEA's antagonistic effect. To further validate the reliability of the metabolomic analysis, characteristic fragment ions (Q3) and retention times (RT) of key differential metabolites (e.g., 2-acetyl-4-hydroxy-3-methoxybenzoic acid, cis-Zeatin riboside, and ptesculentoside) were extracted and compared with Q3 values from our self-built database, with detailed identification parameters presented in Fig. S6 and Table S3. Venn diagram identified 8 metabolites differentially abundant across all comparisons (Fig. S5), primarily lipids, ketones/aldehydes/esters, and terpenoids. KEGG pathway enrichment analysis DAMs were primarily enriched in amino acid metabolism, carbohydrate metabolism (especially PPP-related pathways), secondary metabolite biosynthesis, and transport/genetic information processing (Fig. 6; Table S5).Fig. 6KEGG pathway of the differential metabolites. (A) CKO vs. CKN group, (B) DHEA vs. CKO group, (C) GABA vs. CKO group, (D) GABA+DHEA vs. CKO group, (E) GABA+DHEA vs. DHEA group, and (F) GABA+DHEA vs. GABA group. Note: Abscissa: rich factor; Ordinate: path name. The dot size represents the number of enriched differentially abundant metabolites, and the dot color represents the p value; a redder dot represents a more significant enrichment.Fig. 6 KEGG pathway of the differential metabolites. (A) CKO vs. CKN group, (B) DHEA vs. CKO group, (C) GABA vs. CKO group, (D) GABA+DHEA vs. CKO group, (E) GABA+DHEA vs. DHEA group, and (F) GABA+DHEA vs. GABA group. Note: Abscissa: rich factor; Ordinate: path name. The dot size represents the number of enriched differentially abundant metabolites, and the dot color represents the p value; a redder dot represents a more significant enrichment. CKO vs. CKN: 55 metabolites mapped to 48 pathways, with major changes in isoflavone biosynthesis (6 metabolites), ABC transporters (10 metabolites), PPP-related fructose/mannose and galactose metabolism, flavonoid/flavonol biosynthesis, purine metabolism, and aminoacyl-tRNA biosynthesis. DHEA vs. CKO: 18 metabolites across 12 pathways. PPP-related pathways and ABC transporters showed 3 differentially abundant metabolites each; flavonol/flavonoid and isoflavone biosynthesis were downregulated. GABA vs. CKO: 10 metabolites across 10 pathways. PPP-related pathways (fructose/mannose and galactose metabolism) were modulated; isoflavone biosynthesis, plant hormone signal transduction, and zeatin biosynthesis were also affected. GABA+DHEA vs. CKO: 28 metabolites mapped to 23 pathways, with ABC transporters (6 metabolites), fructose/mannose metabolism (4 metabolites), and galactose metabolism significantly enriched. Interaction comparisons: Combined vs. DHEA showed changes in plant hormone signal transduction, isoflavone biosynthesis, lysine biosynthesis, and aminoacyl-tRNA biosynthesis. Combined vs. GABA showed significant changes in flavonoid/flavonol biosynthesis and plant hormone signal transduction. Discussion GABA as a metabolic repair agent for seed vigor and sprout quality This study demonstrates that exogenous GABA effectively revitalizes aged soybean seeds (Fig. 1), restoring germination vigor and significantly enhancing antioxidant bioactive metabolites in the resultant sprouts. Consistent with previous findings in aged wheat and medicinal pumpkin seeds (Afshari & Seyyedi, 2020; Shakhov et al., 2024), GABA markedly improved germination percentage, energy index, and vigor index in aged soybean seeds (Fig. 2A-D). Concomitantly, GABA treatment substantially elevated the levels of major antioxidant bioactive compounds, including total phenolics, total flavonoids, and total antioxidant capacity, in the resultant sprouts (Table 1). Metabolomic analysis further revealed that this enhancement was driven by the dramatic upregulation of specific potent antioxidant metabolites, particularly Isoliquiritin (a potent flavonoid, FC = 17,609.57), Bavachalcone (an antioxidant flavonoid, FC = 2.96), and 3,3′-di-O-methylellagic acid 4,4′-di-O-β-glucoside (a polyphenol with strong ROS scavenging capacity, FC = 7.90) (Table S6). Notably, endogenous GABA content was also dramatically increased to ∼3-fold of fresh seed levels (Table 1), endowing the sprouts with enhanced functional value. This seed deterioration under aging conditions stems from cumulative damage to cellular structures and metabolic functions, including membrane lipid peroxidation, electrolyte leakage, and degradation of stored mRNAs and ribosomes (Sano et al., 2020). GABA priming not only rescues seed viability but also fundamentally restores the biosynthetic capacity of aged seeds, enabling production of sprouts with nutritional profiles (rich in phenolics, flavonoids, ascorbate, and glutathione) comparable to those from fresh seeds. As shown in Table 1, aged control (CKO) sprouts exhibited 50–60% reductions in key antioxidant pools compared to CKN sprouts, with collapsed redox ratios (AsA/DHA, GSH/GSSG). GABA restored these parameters to levels approaching fresh seed benchmarks. GABA treatment markedly induced expression of GmMSD2 (4.2-fold), GsoCAT4.1 (5-fold), PAL (3.8-fold), and GmG6PD2 (6.8-fold), with peak expression on day 2 (Fig. 4), providing molecular basis for restored enzyme activities (Fig. 3A-D) and non-enzymatic antioxidant pools. Notably, membrane integrity was significantly restored by GABA treatment, as evidenced by reduced relative electrolyte leakage (REL) from 47.5% in aged controls to 27.7% in GABA-treated sprouts, consistent with the decreased TBARS levels. Another key finding is that GABA treatment dramatically elevated GABA content in sprouts to ∼303% of fresh seed levels (∼3-fold increase; Table 1), which may have implications for functional food development. The significant accumulation of GABA in sprouts following treatment may represent a novel strategy for producing natural GABA-enriched sprouts. Notably, this simultaneous enhancement of both antioxidant capacity and GABA content establishes the metabolic foundation for developing a “dual-functional” sprout product, concept that is further discussed in section 4.4.2. The significant upregulation of glutamate decarboxylase (GAD) activity in GABA-treated aged seeds (Fig. 3F) provides direct enzymatic evidence for the mechanism underlying endogenous GABA accumulation. GAD is the key rate-limiting enzyme in the GABA shunt, catalyzing the decarboxylation of glutamate to GABA. The restored GAD activity in GABA-treated sprouts (reaching 83.5% of fresh seed levels) correlates well with the ∼3-fold increase in endogenous GABA content (Table 1), suggesting that exogenous GABA treatment not only directly contributes to the GABA pool but also activates the endogenous GABA biosynthetic pathway (Xu & Hu, 2014). The PPP as a metabolic hub connecting germination and antioxidant biosynthesis The metabolic foundation underlying these phenotypic improvements lies in the systemic activation of the PPP-NADPH axis by GABA. Seed aging is closely associated with ROS accumulation, which causes oxidative damage and compromises seed vigor (Kurek et al., 2019; Li et al., 2024; Lin et al., 2022). Conversely, seed revitalization requires enhanced antioxidant systems to scavenge ROS and restore redox balance (Kibinza et al., 2011; Yan et al., 2020; Yan & Mao, 2021). Thus, NADPH restoration via PPP activation provides reducing power for both the AsA-GSH cycle to scavenge ROS and for antioxidant metabolite biosynthesis, representing a key mechanism by which GABA rescues aged seeds. Metabolomic analysis revealed that GABA treatment dramatically elevated metabolites with established antioxidant properties in aged seed-derived sprouts. Among the top 30 upregulated metabolites (Table S6), the majority belong to bioactive compound categories with known ROS scavenging capacity, including flavonoids (e.g., isoliquiritin, bavachalcone), polyphenols (e.g., 3,3′-di-O-methylellagic acid 4,4′-di-O-beta-glucoside, guisinol), organic acids (e.g., (S)-2-Hydroxybutanoic Acid, didemnin A), and nucleotides (e.g., cis-zeatin riboside, urate-3-ribonucleoside). These findings provide a metabolic basis for the enhanced antioxidant capacity (FRAP, AsA/DHA, GSH/GSSG ratios) observed in GABA-treated sprouts (Table 1). GABA activates PPP to restore NADPH pools in aged seeds, thereby repairing metabolic dysfunction and enhancing both germination vigor and nutritional quality of resultant sprouts. GABA treatment significantly increased G6PDH activity in aged seeds (Fig. 3E), thereby restoring the NADPH/NADP+ ratio and consequently recovering ATP content and energy charge (Table 1). This metabolic reprogramming was underpinned by a coordinated transcriptional response: GABA treatment strongly induced the expression of GmG6PD2, alongside key antioxidant genes such as GmMSD2 and GsoCAT4.1, and the phenylpropanoid pathway gateway gene PAL (Fig. 4). The substantial upregulation of these genes provides a direct molecular basis for the restored enzyme activities and enhanced phenolic and flavonoid accumulation observed in GABA-treated sprouts (Fig. 3, Table 1). Notably, Li et al. (2024) reported that aging stress activates the phenylpropanoid metabolism pathway in soybean seeds, leading to increased accumulation of phenylpropanoid-derived metabolites. The consistent activation of PAL by GABA treatment in our study aligns with these findings, suggesting that GABA may enhance phenolic and flavonoid biosynthesis by upregulating PAL expression, thereby amplifying the phenylpropanoid pathway that is naturally triggered by aging stress. The parallel restoration of the NADPH/NADP+ ratio and energy charge (Table 1) strongly supports a sequential relationship, whereby PPP-derived NADPH drives the recovery of mitochondrial ATP synthesis and energy metabolism. Using DHEA, a specific G6PDH inhibitor (Li et al., 2020; Wei et al., 2019; Yu et al., 2005), we demonstrated that PPP inhibition completely blocked all GABA-induced effects, confirming that PPP activation is essential for GABA-mediated seed revitalization. The restored NADPH pool then fuels the regeneration of the major antioxidant buffers, as evidenced by the significantly improved GSH/GSSG and AsA/DHA ratios in GABA-treated sprouts compared to the aged control (Table 1). Consistent with the transcriptional upregulation, GABA significantly enhanced the activities of enzymatic antioxidants (SOD, CAT, APX, and GR), effectively scavenging excess H₂O₂ and mitigating lipid peroxidation (MDA). This robust redox homeostasis rescued mitochondrial viability, as evidenced by the intensified TTC reduction, thereby restoring ATP synthesis and energy charge to levels approaching those of fresh seeds (Table 1). Consequently, the synergistic provision of reducing power (PPP-derived NADPH) and energy (mitochondrial ATP) drove the metabolic reprogramming required for the vigorous biosynthesis of antioxidant metabolites, ultimately revitalizing seed vigor. Thus, PPP activation by GABA initiates a cascade: NADPH restoration → antioxidant system reactivation → mitochondrial protection → ATP and metabolite biosynthesis recovery. The combined treatment group (GABA+DHEA) exhibited a typical “intermediate phenotype”: G6PDH activity and the NADPH/NADP+ ratio fell between those of the DHEA-only and GABA-only treatments (Fig. 3E; Table 1). This partial rescue likely results from: (1) residual G6PDH activity that was substantially higher than that in the DHEA-only group, due to incomplete enzymatic inhibition by DHEA at 20 μM and GABA-induced transcriptional upregulation of GmG6PD2 (Fig. 4); and (2) GABA-activated compensatory NADPH-generating pathways (e.g., malic enzyme), as indicated by the dramatic accumulation of (S)-2-Hydroxybutyric Acid (FC = 68,860.45) in GABA-treated sprouts. This partial rescue effectively excludes the possibility that DHEA's effects are solely due to non-specific toxicity—if DHEA inhibited seed vigor through non-PPP pathways, GABA supplementation would be expected to show either complete rescue or complete failure, rather than the intermediate phenotype observed. PPP-derived NADPH serves dual purposes: (1) directly fueling antioxidant systems to re-establish redox homeostasis; and (2) supplying precursors (e.g., erythrose-4-phosphate) for the shikimate and phenylpropanoid pathways leading to phenolic and flavonoid biosynthesis. Therefore, the sequential activation of PPP-mediated NADPH production, followed by AsA-GSH cycle restoration and subsequent mitochondrial function recovery, constitutes the core metabolic logic through which GABA revitalizes aged seeds. Limitations and future perspectives Our findings reveal that GABA-induced PPP activation repairs the metabolic foundation—restoring ATP synthesis, NADPH generation, and biosynthetic capacity in aged seeds. The dramatic upregulation of GmG6PD2 expression in aged seeds, exceeding fresh seed levels, indicates that the transcriptional machinery for PPP activation remains intact and responsive, even when endogenous activation signals are attenuated. However, certain limitations warrant consideration. The artificial aging protocol (40 °C, 100% relative humidity for 48 h) was carefully optimized to induce moderate aging, reducing germination from nearly 100% to approximately 35–40%, which simulates natural seed deterioration under improper storage conditions. This protocol aligns with previous methodologies reported in soybean aging studies (Xin et al., 2014; Zhang et al., 1995). We deliberately chose a moderate aging level rather than more severe aging (e.g., >72 h) because our preliminary experiments indicated that excessively aged seeds showed severely diminished metabolic capacity and minimal responsiveness to GABA treatment. The aged seeds used in this study retained 35% germination potential (vs. >95% in fresh seeds), representing moderate aging severity; whether GABA can achieve comparable effects on more severely aged seeds (e.g., <10% germination) remains to be investigated. Although we supplemented membrane damage indicators with both TBARS and Relative Electrolyte Leakage (REL) to provide tissue-level evidence, more advanced approaches such as ROS fluorescence staining at the subcellular level or mitochondrial ultrastructural observation via transmission electron microscopy would further strengthen the conclusions. The current evidence supporting PPP as the key pathway is primarily based on pharmacological inhibition using DHEA and correlative physiological data. Without genetic evidence from G6PDH knockout or overexpression lines, we cannot definitively conclude that PPP activation is the sole mechanism for GABA's effect. Although pharmacological inhibition alone cannot provide definitive causal evidence, DHEA has been widely employed as a specific G6PDH inhibitor in numerous plant physiological studies, including those on NADPH production and ROS metabolism in apple fruit (Li et al., 2020; Wei et al., 2019), drought tolerance in soybean roots (Liu et al., 2013), salt stress response in reed callus (Wang et al., 2008), nitric oxide-involved defense in red kidney bean roots (Liu et al., 2007), and G6PDH-dependent secondary metabolite biosynthesis in Taxus cell cultures (Yu et al., 2005), thereby supporting the reliability of our experimental approach. Genetic approaches using G6PDH-overexpressing or knockdown lines would provide more stringent causal evidence (Yang et al., 2019). A fundamental unresolved question is the initial signaling event: how is exogenous GABA perceived to trigger PPP activation? GABA might function as a signaling molecule through putative receptors, triggering Ca2+ signaling or transcriptional reprogramming leading to G6PDH upregulation. Metabolomic analysis revealed no significant changes in tricarboxylic acid cycle (TCA) intermediates in GABA-treated groups (Table S3), while PPP-related metabolites were systematically upregulated, supporting PPP rather than TCA as the primary target. It is worth noting that our preliminary experiments demonstrated that GABA treatment (0.1 mM) had negligible effects on germination parameters in fresh seeds (data not shown), indicating that GABA primarily functions as a restorative agent for aged seeds rather than a general germination promoter. This is consistent with the concept that exogenous treatments typically show more pronounced effects on seeds with compromised vigor. Additionally, the osmotic potential of 0.1 mM GABA was calculated to be approximately −0.00025 MPa, which is negligible compared to the osmotic requirement for seed germination (−0.5 to −1.0 MPa), ruling out any significant osmotic effect of the treatment. Future research should prioritize: (1) genetic validation using materials with altered G6PDH activity; (2) testing efficacy across diverse seed types; (3) integrating transcriptomic and proteomic analyses of early sprouts; (4) investigating postharvest shelf-life of sprouts from GABA-primed aged seeds; and (5) employing metabolic flux analysis with stable isotopes to directly test the metabolic flux hypothesis. Practical implications and industrial perspectives Valorization of aged seeds Globally, substantial proportions of stored soybean seeds lose viability annually due to aging, causing considerable economic losses (Chhabra & Singh, 2019). The present study demonstrates that GABA priming effectively transforms low-vigor aged seeds into viable planting materials, enhancing the nutritional profile of resultant sprouts and aligning with the “value-added utilization of suboptimal seeds” concept (Nevara et al., 2023). From an economic perspective, GABA priming offers clear advantages. The low concentration (0.1 mM) and brief soaking (30 min) minimize operational complexity and costs. Furthermore, GABA-treated seeds yield sprouts with significantly elevated antioxidant compounds and improved quality, potentially enhancing their market value as functional food ingredients. Notably, this valorization approach differs from traditional seed by-product utilization by directly converting aged seeds into high-value sprout products rather than processing them as animal feed or industrial raw materials (Farag et al., 2022). Production of dual-functional sprouts In light of the restored metabolic capacity, the GABA-treated sprouts from aged seeds can be clearly defined as ‘dual-functional’: they concurrently provide markedly enhanced antioxidant capacity, driven by PPP activation and NADPH restoration, and naturally enriched GABA, reaching approximately three-fold the level of fresh sprouts (Hou et al., 2024; Lee et al., 2023; Sakthivel et al., 2025). The elevated antioxidant profile in these sprouts is supported by recent evidence showing that specific processing interventions can significantly elevate health-promoting bioactives in plant-based foods (Du et al., 2025). These antioxidant nutrients are essential for human health, helping scavenge free radicals and potentially preventing chronic diseases including cardiovascular disorders, diabetes, and neurodegenerative diseases (Halliwell, 2024; Zujko & Witkowska, 2023). Importantly, soybean sprouts are naturally rich in bioactive isoflavones with established antioxidant and health-promoting properties (Kim & Kim, 2020), and our GABA treatment further enhanced these functional components, making the resultant sprouts a promising “dual-functional” food product. Previous studies have demonstrated that GABA treatment can regulate phenolics accumulation in soybean sprouts under stress conditions (Zhao et al., 2021), and soybean sprouts contain diverse bioactive compounds including isoflavones, phenolic acids, and flavonoids that contribute to human health benefits (Swallah et al., 2023). The “dual-functional” nature of these sprouts—simultaneously enriched in both GABA and multiple antioxidant compounds—distinguishes them from single-bioactive-component functional foods and may appeal to health-conscious consumers seeking comprehensive nutritional benefits, aligning with the growing global market trend toward multifunctional food products. Industrial feasibility This strategy can be integrated into existing seed processing and sprout production workflows without specialized equipment, making it suitable for small-to-medium agricultural enterprises. The treatment protocol is simple, rapid, and uses food-grade concentrations of GABA, which supports its practical implementation. Furthermore, the use of aged seeds as raw materials addresses the challenge of seed waste in agricultural production, contributing to sustainable food systems (Agil & Hosseinian, 2012; Gaur et al., 2020). Overall, GABA priming represents a cost-effective, sustainable approach to valorize aged soybean seeds for high-quality sprout production, with potential for industrial implementation. This study not only advances our understanding of seed revitalization mechanisms but also provides a practical solution to reduce agricultural waste while meeting consumer demand for functional foods. Conclusion This study demonstrates that exogenous GABA effectively revitalizes aged soybean seeds and enhances antioxidant bioactive compounds in resultant sprouts. Physiological, biochemical, and metabolomic analyses collectively identify NADPH replenishment via PPP activation as the central mechanism underlying GABA-mediated seed revitalization. This is the first report showing that GABA revitalizes aged seeds by activating PPP to replenish NADPH and produces “dual-functional” sprouts with elevated antioxidant capacity and high GABA content. The simple GABA protocol is cost-effective and suitable for large-scale implementation, transforming low-vigor aged seeds into high-quality functional sprouts while reducing agricultural waste. However, current conclusions are primarily based on pharmacological inhibition; genetic validation using G6PDH-modified lines is needed to establish definitive causal evidence. Future studies should explore the upstream signaling cascade by which GABA triggers PPP activation and investigate the stability of GABA-enriched sprouts during storage and processing. Overall, this work provides a sustainable strategy to valorize aged soybean seeds and addresses the growing consumer demand for functional foods with dual health benefits. CRediT authorship contribution statement Song Yu: Writing – original draft, Project administration, Methodology, Formal analysis. Jiaqi Yin: Writing – original draft, Project administration, Methodology, Formal analysis. Fandi Meng: Writing – original draft, Project administration, Methodology, Formal analysis. Kaiqiao Hua: Visualization, Validation, Software, Investigation, Data curation. Keyi Zhao: Visualization, Validation, Software, Investigation, Data curation. Fengming Li: Visualization, Validation, Software, Investigation, Data curation. Tong Zhao: Visualization, Validation, Software, Investigation, Data curation. Yingjie Wang: Visualization, Validation, Software, Investigation, Data curation. Yifei Zhang: Writing – review & editing, Conceptualization. Chunyu Zhang: Visualization, Validation, Software, Investigation, Data curation. Lihe Yu: Supervision, Resources. 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.