The soybean (Glycine max L.) seeds of “Zhexian NO.9,” produced in the experimental farm of Zhejiang Academy of Agricultural Science, were used in our study. The seeds were stored in the airtight dry container at room temperature for 1, 6, 12, 18, and 24 months for naturally aging treatment (NAT), respectively.

The 1-month stored soybean seeds were used for the accelerated aging test (AAT) according to the method of Missaoui and Hill (2015) with small modifications. Briefly, 100 soybean seeds samples were placed into a seed-aging box and incubated at 45°C for 48 h under 100% relative humidity. After that, processed seeds were subjected to 2 days of drying under room temperature prior to the germination test.

Preliminary tests were conducted to determine the condition for ultrasonic treatment (Supplementary Table 1). Then, soybean seeds were subject to ultrasonic treatment for 15 min with the instrument ultra sonicator (BOKE, BKE-1002DHT; dimension, L × W × D: 300 mm × 240 mm × 150 mm; power, 100 W; and frequency, 40 kHz). Distilled water was used as the control.

Soybean seeds were disinfected with sodium hypochlorite solution (0.1%) for 15 min before the seed germination test. Briefly, rolled towels placed with 50 soybean seeds for each treatment were soaked with purified water and incubated at 25°C. The number of germinated seeds was counted at 0–7 days of germination. At 0, 1, and 3 days of germination, the soybean seeds were collected and used for further biochemical analysis. Soybean seeds were planted in sand at 25°C under the 12/12 h light/dark cycle. On day 7, the seedling establishment ratio (SER) was measured based on the formula SER = Number of normal seedlings/Number of tested seeds (Zhang et al., 2007). The seed germination and seedling emergence tests were performed with four biological replications.

Determination of hydrogen peroxide (H₂O₂) and superoxide anion (O₂⋅^–) content was analyzed using spectrophotometry. Briefly, 0.3 g of seed samples were homogenized in 3 ml of acetone and centrifuged at 12,000 rpm for 15 min. Later, 1 ml of supernatant was collected and mixed with 0.1 ml of titanium sulfate solution (5%) and 0.2 ml of ammonia water (25%). After centrifugation again at 12,000 rpm for 15 min, the precipitate was collected and dissolved with 3 ml of sulfuric acid (2 mM). Seeds sample was incubating in 3 ml of a reaction mixture [10 mM sodium phosphate buffer (pH 7.8), 10 mM sodium azide (NaN₃), and 0.05% (m/v) nitroblue-tetrazolium (NBT)]. Finally, H₂O₂ level was measured based on supernatant absorbance value (OD) at 390 nm with the method of Hasanuzzaman et al. (2021), and O₂⋅^– level was calculated based on the absorbance of the supernatant at 580 nm according to the method of Soares et al. (2021). All the samples were assayed with four biological replicates.

Malondialdehyde (MDA) content was measured for the determination of lipid peroxidation with the method of Padilla et al. (2021). Briefly, the seeds sample was homogenized with 0.05 M phosphate buffer (pH 7.8) and centrifuged at 10,000 rpm for 15 min. The supernatant was mixed with 1 ml of 5% (w/v) thiobarbituric acid-trichloroacetic acid solution and used for spectrophotometric determination. MDA level (μmol g^–1 FW) were calculated by the formula MDA = [(OD₅₃₂−OD₆₀₀) × A × C]/(0.155 × W). In the formula, A and C corresponded to the volume of the reaction solution and extraction solution, respectively, and W represented the fresh weight of the sample. The analysis of MDA was performed with four biological replicates.

Antioxidant enzymes analysis was performed based on the method of Soares et al. (2021). Potassium phosphate buffer (pH 7.5) was utilized for seeds homogenization. The homogenates were centrifuged at 12,000 × g for 20 min to collect supernatant for the measurement of activities of superoxide dismutase (SOD), peroxidase (POD), catalase (CAT), ascorbate peroxidase (APX), and acetaldehyde dehydrogenase (ALDH). The analysis of antioxidant enzymes was performed with four biological replicates.

Anthrone-H₂SO₄ colorimetry was performed to determine soluble sugar content in soybean seeds according to the method of Yi and Dong (2008) with four biological replicates. Sucrose level was measured by the resorcinol method and estimated by absorbance value at 480 nm (Shi et al., 2016). In addition, fructose content was measured by the method of Cai et al. (2016). The glucose level in soybean seed was measured through high-performance liquid chromatography (HPLC) according to the method of Bailly et al. (2010).

Seed samples (2 g) were homogenized into powder and extracted with 0.5 mol/l perchloric acid (10 ml) at 4°C for 5 min. The homogenization was centrifuged at 12,000 rpm for 10 min, and the supernatant (5 ml) was collected and filtered through a 0.22 μm membrane filter. The determinations of ATP and energy charge were performed by HPLC with the method of Emami and Kempken (2019). Data were expressed as the average values from four biological replicates.

Enzyme-linked immunosorbent assay (ELISA) was used to determine the activities of lipase (LIPG), GK, α-glycerol phosphate oxidoreductase (GPDH), citrate synthase (CS), aconitase (ACO), isocitrate lyase (ICL), malate synthase (MS), malate dehydrogenase (MDH), succinate dehydrogenase (SDH), fumarase (FUM), malate degydrogenase (MDH), phosphoenolpyruvate carboxykinase (PCK), enolase, phosphoglycerate kinase (PGK), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), aldolase, fructose-1,6-bisphosphatase (FBPase), UDPG1c pyrophosphorylase (UDPG PPase), sucrose synthase (SUSY), sucrose-6-P synthase (SPS), sucrose phosphatase (SPP), pyrophosphatase (PPase), and invertase (INV) with the ELISA kit (Mlbio, Shanghai, China). Briefly, soybean seeds were extracted using the extraction buffer, and then the chromogenic reaction was carried out with the Chromogen Solution. The color change was tested by absorbance value at 450 nm. The activities of TAG and glycometabolism-related enzymes in soybean seeds were then determined by comparing the OD of these samples to the standard curve.

The PrimeScript RT Reagent Kit (Vazyme, Nanjing, China) was employed to extract RNA from seeds and prepare cDNA through reverse transcription. The CFX96 Touch Real-Time PCR system (Biorad) was utilized for RT-qPCR procedures by using the ChamQ SYBR qPCR Master Mix (Vazyme, Nanjing, China). Supplementary Table 2 presents all the primers utilized in this study, with GmTubulin being the internal control. The fold change of expression (FC) was determined with the formula of FC = EΔCt, where E stands for the average gene amplification efficiency, ΔCt indicates the difference in average Ct values for all biological replicates between two compared samples. The normalized results were displayed in a form of mean SD. The RT-PCR assays were performed with four independent biological replicates and three technical replicates.

Data were statistically analyzed by one-way ANOVA using Statistical Analysis System (SAS) software. Besides, the least significant difference at p < 0.05 (LSD₀.₀₅) was adopted for multiple comparisons. Percentage data were converted to arcsintrans values by y = arcsin [sqrt (x/100)] before statistical analysis.

The harvested soybean seeds were stored for 1, 6, 12, 18, and 24 months, respectively, and then subjected to germination test and seedling establishment (Figure 1). The results showed that the soybean seeds maintained a high-germination rate of 90% following both 1 and 6 months of storage. Long-term storage (12, 18, and 24 months) significantly reduced the germination rate of soybean seeds. On day 7 of germination, 12 and 18 months stored sunflower seeds achieved 71 and 65.5% germination rates, respectively, while 66% of soybean seeds lost their germination capacity following 24 months of storage.

Similarly, the soybean seedling establishment rate showed a substantial decrease with the increase in storage time. No significant differences in seedling establishment rate were observed between soybean seeds stored for 1 and 6 months, reaching approximately 80%. By contrast, for the seeds stored for 12, 18, and 24 months, the seedling establishment rates decreased significantly, which were 61.50, 49.75, and 20.50%, respectively. Besides, the seedling establishment rates were all lower than the germination rates after storing for different times. In summary, the aforementioned results revealed that natural aging significantly decreased the germination and seedling establishment ability of soybean seeds.

The result proved that UWT exhibited a positive effect on germination and seedling establishment from NAT soybean seeds (Figure 2). Following 12 and 18 months of storage, the UWT seeds exhibited significantly higher rates of germination and seedling emergence in comparison with the control. However, no remarkable difference in seed germination and seedling emergence was detected between NAT and NAT + UWT treatments in soybean seeds stored for 1, 6, and 24 months.

Given the role of UWT in promoting seed vigor of soybean seeds following the storage process, we further explored the effect of UWT on the germination of artificially aged soybean seeds (Supplementary Figure 1). As a result, artificial accelerated aging remarkably decreased soybean seed germination and seedling emergence rates. Nevertheless, UWT made no significant effect on artificially aged soybean seed germination or seedling emergence. Generally, we speculated that the germination of NAT and AAT soybean seeds might respond differently to UWT. After proving the role of UWT in improving the seed vigor of NAT soybean seeds, only 12 months of stored seeds were adopted for further tests.

Overaccumulation of ROS and MDA in soybean seeds during storage was an important reason for the deterioration of seed vigor. To study the mechanism whereby UWT improved the vigor of aged soybean seeds, we determined the H₂O₂, O₂⋅^–, and MDA contents in soybean seeds during early germination. As shown in Figure 3, UWT effectively alleviated the overaccumulation of ROS and MDA in aged soybean seeds. On 1 and 3 days of germination, the H₂O₂, O₂⋅^–, and MDA contents of NAT + H₂O were significantly higher than those without NAT seeds. While NAT + UWT seeds exhibited significantly lower contents of H₂O₂, O₂⋅^–, and MDA than the NAT group on 1 and 3 days of germination. Besides, UWT also significantly reduced the MDA content of aged soybean seeds on 0 day of germinated (Figure 3).

Considering the effect of UWT on the ROS and MDA levels in aged soybean seeds during early germination, we further evaluated the activities of antioxidant enzymes and ALDH. The results revealed that NAT dramatically lowered the activities of SOD, APX and ALDH as compared with fresh seeds during early germination (Figures 4A,D,E). UWT significantly enhanced the SOD, CAT and ALDH activities in NAT soybean seeds on 1 and 3 d of germination (Figures 4A,B,E). However, no significant difference in POD activity was observed between NAT + H₂O and NAT + UWT during early germination (Figure 4C). Consistently, the qPCR analysis also revealed significantly lower expressions of corresponding genes in NAT soybean seeds in comparison with fresh seeds, including GmSOD1, GmSOD3, GmCAT1, GmAPX2, Gm APX3, and GmALDH1 (Figure 5). Besides, significantly higher transcriptional levels of GmSOD1, GmAPX2, and GmALDH1 have been observed in NAT + UWT seeds compared with NAT + H₂O treatment at 0 day of germination. UWT also increased the transcriptions of GmSOD3, GmAPX3, and GmCAT1 in NAT soybean seeds at 1 and 3 days of germination (Figure 5).

During seed imbibitions, the soluble sugars are the primary energy source for early germination and seedlings establishment, including sucrose, glucose, and fructose (Eastmond et al., 2015). To better investigate the effect of UWT in improving the germination and seedling establishment of aged soybean seeds, we further determined the levels of soluble sugar, sucrose, glucose, and fructose (Figure 6). Sugars quantification revealed that NAT markedly lowers the contents of soluble sugar, sucrose, and fructose during early germination time. Interestingly, UWT significantly elevated the levels of soluble sugar, sucrose, and fructose in NAT soybean seeds. Moreover, significantly higher levels of ATP and energy charge were detected in NAT + UWT seeds as compared with NAT + H₂O seeds (Figures 6E,F). Nevertheless, the glucose content was not significantly affected by NAT or UWT during soybean seed germination (Figure 6D).

During early oil-seed germination, TAGs were decomposed into soluble sugars through complex metabolism, thus providing energy for seed germination. The catabolism of TAGs involved several key enzymes, mainly including LIPG, GK, GPDH, CS, ACO, LCL, MS, MDH, PCK, Enolase, PGK, GAPDH, Aldolase, FBPase, UDPGPPase, SUSY, SPS, SPP, PPase, and INV. This study revealed that the activities of TAG catabolism-related enzymes continued to elevate during early germination time (Figure 7). In addition, NAT significantly inhibited the activities of a couple of key GNG-involved enzymes. Obviously, NAT failed to affect the activities of LIPG, GK, GPDH, and enzymes involved in GC and TAC. However, what is noteworthy is that UWT drastically enhanced the activities of several key GNG-related enzymes in the aged soybean seeds germinated for 1 and 3 days, including PCK, FBPase, SUSY, SPP, PPase, and INV. Similarly, no significant differences in the activities of LIPG, GK, GPDH, and enzymes involved in GC and TAC were detected between the NAT + UWT and the NAT + H₂O treatment (Figure 7).

The gene expression analysis showed that the transcriptions of GmPCK1, GmFBPase1, GmFBPase2, GmSUSY3, GmSPP1, GmINV1, and GmINV3 were downregulated by NAT. While NAT made no effect on GmPPase2 expression during soybean seed germination. It should be noted that the regulation of UWT on GNG-related gene expression was opposite to that of NAT. UWT remarkably increased the transcriptions of GmPCK1, GmFBPase2, GmSPP1, and GmINV1 at 1 and 3 days of germination. Besides, the transcriptional levels of GmSUSY3 and GmINV3 in NAT + UWT seeds were significantly high than those in NAT seeds (Figure 8).

In agricultural production, improper storage methods or prolonged storage time often led to seed deterioration, which, in turn, inhibited the field seedling emergence (Yin et al., 2014, 2015; Fleming et al., 2017). Accordingly, a study on seed pretreatment methods to enhance the vigor of aging seeds from the perspectives of molecular mechanism and production application is a worthy objective of concern in agricultural production. Maffei (2014) reported that the magnetic field treatment could activate the absorption of water and nutrients, influence the cell membrane composition and integrity, thereby promoting seed germination. Similarly, Hu and Zhou (2014) found that novel-hydrated graphene ribbon accelerated the germination speed and improved the seedling characteristic from stored wheat seeds by enhancing the integrity of cell membranes and increasing the contents of soluble sugars, amino acids, and FAs. Cold plasma application could facilitate rapeseed germination with drought stress (Li et al., 2015). Besides, a nanopriming technology was developed by Mahakham et al. (2017) for promoting seed vigor from stored rice seeds with photosynthesized silver nanoparticles. This study found that ultrasonic treatment could boost the germination and seedling emergence of NAT soybean seeds, but failed to enhance the vigor of artificially aged soybean seeds.

Ultrasound, as a sound wave with a frequency above 20 kHz, is characterized by good directivity and penetration (Suslick, 1990). UWT of seeds is a novel physical method featuring simple operation, low cost, and eco-friendliness. Several reports have revealed the function of ultrasonic treatment on seed vigor improvement (Chen et al., 2012; Ran et al., 2015). UWT could activate the transportation of water and oxygen, the release of hydrolytic enzymes, accelerate the decomposition of protein and starch in the endosperm, and lead to improved germination of cereal crop seeds (Patero and Augusto, 2015). Nevertheless, UWT may also lead to nutrient leakage to adversely influence seed germination (Wang et al., 2012). Besides, Liu J. et al. (2016) proved the role of UWT on germination improvement in tall fescue and Russian wildrye aged seeds, but the underlying mechanism remained unclear. However, most of these studies focused on germination rather than seedling emergence of the aged seeds, especially in aging oil seeds. Further study is required to explore the intrinsic mechanism of UWT-induced promotion on seed germination ability. This study extended the practical uses of UWT, suggesting that UWT could reverse the inhibition of seed vigor of soybean induced by natural aging.

Reactive oxygen species exert a vital role in cellular metabolism, especially affecting the dormancy and aging of seeds (Oracz and Karpinski, 2016; Sachdev et al., 2021). During the seeds storage, the decrease of activities of antioxidant enzymes, lipid peroxidation, and substantial accumulation of ROS, MDA resulted in the deterioration of seed vigor (Corbineau and Bailly, 2007). Our previous study reported that naturally aging significantly increased the accumulations of H₂O₂, O₂⋅^–, and MDA and resulting in the acceleration of seed deterioration (Huang et al., 2021b). ROS has a dual function, which plays a positive role in seed development, germination, and resistance to environmental stress at appropriate concentrations, while generating toxic effects on the cells and plant tissues at excessive concentrations (Barba-Espin et al., 2011; Oracz et al., 2015). Herein, compared with the higher contents of ROS (H₂O₂, O₂⋅^–) and MDA in NAT soybean seeds than the fresh seeds, NAT + UWT treatment effectively alleviated the overproduction of ROS and MDA during 1 and 3 days germination. It is crucial to maintain ROS homeostasis during seed germination, so as to guarantee seed germination as well as subsequent seedling emergence (Yin et al., 2015; Deng et al., 2017). This study proposed that UWT could enable better ROS scavenging ability in seeds during early germination than the control. This is similar to the previous study showing that maintaining ROS homeostasis is one of the mechanisms underlying UWT improved seed germination under abiotic stress (Chen et al., 2013).

Antioxidases, mainly including POD, SOD, APX, and CAT, are crucial to the ROS scavenging system (Mittler, 2006; Hasanuzzaman et al., 2020; Sachdev et al., 2021). Lin et al. (2017) found that the activities of POD, CAT, APX, and SOD decreased gradually during paddy seed storage, showing positive correlations with the seed vigor. Similarly, our results revealed significantly lower activities of SOD, ALDH, and corresponding genes expressions in NAT soybean seeds compared with fresh seeds. By contrast, the antioxidant enzymes activities are closely related to ROS level in soybean seeds; to be specific, NAT + UWT seeds showed significantly increased SOD, ALDH, and CAT activities relative to NAT + H₂O seeds. Consistently, the gene expression analysis revealed that UWT significantly increased the transcription levels of GmSOD1, GmSOD3, GmCAT1, GmAPX2, and GmALDH1 in NAT soybean seeds at 1 and 3 days of germination. The above results were consistent with the previous study reporting that ultrasonic treatment could improve the germination of aged Russian wild rye (Psathyrostaehys juncea Nevski) and tall fescue (Festuca arundinacea) seeds by regulating antioxidant defense (Liu J. et al., 2016), showing that UWT had a certain effect on the efficacy of antioxidant enzyme system in aged seeds during imbibitions. The successful activating of antioxidant defense by UWT was also reported by several other studies (Chen et al., 2012; Yang et al., 2015).

Admittedly, soluble sugars, mainly sucrose and fructose, are the primary energy source during early germination (Quettier et al., 2008; Eastmond et al., 2015). Soybean seeds are abundant in TAG, and TAG catabolism provides substantial energy for the germination and seedling establishment processes (Theodoulou and Eastmond, 2012). The mobilization and conversion of stored TAG into soluble sugars involve multiple pathways and subcellular fractions (Li-Beisson et al., 2010). Triacylglycerol lipase (LIPG) hydrolyzes TAGs stored in oil to produce FAs and glycerol (Basnet et al., 2016). GK catalyzes the conversion of glycerol into glycerol 3-phosphate and enters the gluconeogenic pathway. FAs produce CoA-SH via the β-oxidation process, generating sucrose via the glyoxylate cycle, and GNG (Eastmond, 2006). In this study, NAT significantly inhibited soybean seed germination and seedling emergence, while UWT effectively alleviated the NAT-induced inhibition of soybean seed vigor. Further analyses demonstrated that NAT significantly lowered the activities of several key gluconeogenic enzymes such as PCK, FBPase, SUSY, and SPP, and also the corresponding gene expression levels. Consequently, the soluble sugar content during early germination declined drastically. PCK catalyzes the production of phosphoenolpyruvate from oxaloacetate, which represents the rate-limiting step in the metabolic pathway (Penfield et al., 2012). Fructose-1,6-bisphosphatase (FBPase) functions to catalyze the conversion of fructose-1,6-bisphosphate to fructose-6-phosphate (Graham, 2008; Quettier et al., 2008). SUSY and SPP are in charge of catalyzing UDPGlc conversion into sucrose, which then subsequently produces fructose and glucose by invertase (Kaur and Hu, 2009; Ma et al., 2011). Pyrophosphatase (PPase) is another key enzyme involved in GNG, which catalyzed the generation of phosphate from cytosolic pyrophosphate. A PPase-deficient Arabidopsis mutant (fugu5) shows a poor seedling emergence rate due to the block of sucrose production, even though the total glyoxysome-associated processes except for b-oxidation can be completed (Ferjani et al., 2011). In this study, UWT significantly increased the activity of PPase and GmPPase2 expression in NAT soybean seeds, suggesting that NAT treatment suppressed TAG conversion to sucrose, as supported by Zhou et al. (2019) indicating that the blocked FA catabolism accounted for the primary cause of the low-aged soybean seed vigor (Zhou et al., 2019). In contrast, UWT significantly augments the enzymes activities and corresponding-gene expressions associated with GNG, and also the accumulation of soluble sugars, sucrose, and fructose. Interestingly, NAT and UWT did not significantly affect the activities of LIGP, as well as of enzymes related to glyoxylate or tricarboxylate cycle. It was proposed that the inactivation of GNG might be a major reason for the deteriorated vigor of soybean seeds during storage. In addition, UWT could improve soybean seeds germination by weakening the GNG inhibition induced by NAT.