Insights into the improvement of bioactive phytochemicals, antioxidant activities and flavor profiles in soy whey fermented by various native lactic acid bacteria

Abstract

Soy whey, a tofu production byproduct rich in nutrients, is typically discarded. This study investigated five natural lactic acid bacteria (LAB) strains isolated from spontaneously fermented soy whey (FSW) for their fermentation performance in soy whey medium. The growth characteristics, total phenols and flavonoids compounds, antioxidant activities, and aroma profiles were evaluated. The results demonstrated that all five LAB strains exhibited excellent adaptability in soy whey fermentation. After fermentation, the total phenols and flavonoids compounds increased, total isoflavone glycosides (daidzin, glycitin and genistin) decreased by 38–47%, while their corresponding aglycones (daidzein, glycitein, and genistein) significantly increased to 36.38–40.52 mg/L, resulting in improved bioavailability. A total of 33 volatile compounds were identified, Off-flavors like hexanal (beany odor) decreased, while new esters (ethyl hexanoate and butyl acetate) enhanced the flavor profile. Sensory evaluation showed LAB fermentation improved soy whey's taste, texture and acceptability, with strain 3–1 performing best. Results demonstrate LAB fermentation effectively reduces anti-nutritional factors while enhancing functionality and palatability, supporting fermented soy whey's potential as a functional beverage.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13205-025-04500-5.

Introduction

Tofu is a widely consumed soybean product in Asian countries and is highly valued for its superior quality, palatable flavor, and exceptional nutritional profile (Chen et al. 2022a; Li et al. 2017). Notably, the tofu production process generates substantial by-products, with approximately 8.95 kg of soy whey being produced per kilogram of tofu manufactured (Zhang et al. 2022). Soy whey is not only rich in basic nutrients such as carbohydrates, proteins, fats, and minerals, but also contains bioactive compounds like oligosaccharides, isoflavones, and vitamins. However, the majority of tofu producers discard it as waste since there are no economical processing techniques, which wastes resources and may pose environmental issues (Chen et al. 2022a). Therefore, there is an urgent need to explore efficient ways to utilize soy whey.

Fermented soy whey (FSW), produced through the natural fermentation of soy whey, was commonly employed as a coagulant in traditional Chinese tofu production to produce soy whey tofu. Reusing residual soy whey from tofu processing in subsequent batches can reduce overall waste generation (Chua and Liu 2019). Furthermore, the dominant lactic acid bacteria (LAB) within the FSW microbial community play a significant role in fermentation, influencing both the flavor and texture of tofu products (Dai et al. 2023). Even the concentration of total isoflavones and aglycones in FSW tofu was improved due to the biotransformation of LAB in FSW, which indicates that FSW is not only an acid coagulant for tofu but also a good candidate for functional beverage resources. Therefore, a functional FSW beverage could be a useful tactic for recycling soy whey from the production. Notably, LAB-mediated biotransformation in FSW enhances the concentration of total isoflavones and aglycones in the resulting tofu, suggesting that FSW serves not merely as an acidic coagulant but also as a promising resource for functional beverages. Consequently, developing functional FSW beverages could represent an effective strategy for recycling soy whey from tofu production.

However, natural FSW is inherently variable due to its spontaneous fermentation process, with its quality heavily influenced by human factors and environmental variables such as temperature, season, and geographical location (Dai et al. 2021) On the other hand, the microbes in natural FSW include both beneficial bacteria like LAB and dangerous spoiling bacteria like Escherichia coli. Furthermore, the bioactive phytochemicals and flavor profiles of FSW beverages might depend on the fermented microorganisms, especially the LAB (Tindjau et al. 2023a). Until now, although the effect of some LAB on FSW beverages has been investigated, most of them are derived from dairy products which might not be the best fit for soy whey (Xiao et al. 2015). Only a few LAB strains isolated from natural FSW have been investigated. For example, Lacticaseibacillus casei YQ336, which was identified from FSW and capable of rapidly producing large amounts of lactic acid, has been used for the pure culture fermentation of soy whey (Qian et al. 2023). Actually, natural FSW is a valuable resource for isolating LAB with unique characteristics, such as the ability to transform bioactive compounds, produce pleasant flavors, or generate various organic acids. However, to date, the specific LAB strains from FSW remain limited, and their fermentation properties are still largely unexplored.

In our previous work, five different LAB, including Lactobacillus plantarum (1–1), Lactobacillus rhamnosus (1–16), Lactobacillus fermentum (7–16), Lactobacillus casei (2–1), and L. paracasei (3–1) were isolated from natural FSW and identified. However, whether these strains are suitable for the FSW beverage is still unclear. Current research determined the in vitro tolerance of the five strains to simulated gastrointestinal conditions first. Furthermore, the bioactive compounds and antioxidant activity of the pure culture FSW were investigated. In addition, the volatile profile and sensory evaluation of the FSW beverage were conducted.

This was the first time to compare the effects of different native LAB from natural FSW on the FSW beverage's quality. Therefore, the results can highlight the potential for the production of functional foods and thus promote the recycling of food by-products.

Materials and methods

Materials and reagents

The soy whey used in this study was provided by Beijing Chaoyang District North Beach restaurant. 2,2-diphenyl-1-(2,4,6-trinitrophenyl) hydrazyl (DPPH), folin-ciocalteu’s phenol reagent, bile salts, pepsin, and trypsin were purchased from Macklin (Shanghai, China). Ultrapure water, HPLC-grade methanol, acetonitrile, and formic acid were provided by Meridian (Beijing, China). Glycitin, genistin, glycitein, and genistein were purchased from Beijing Jinlis Technology Co., Ltd. (Beijing, China). All reagents and standards were prepared with HPLC-grade water or methanol before use.

Microbial strains

The LAB strains including Lactobacillus plantarum (1–1), Lactobacillus rhamnosus (1–16), Lactobacillus fermentum (7–16), Lactobacillus casei (2–1), and Lactobacillus paracasei (3–1) were laboratory stock cultures originally isolated from FSW. These LAB strains were grown on De Man, Rogosa, and Sharpe (MRS) media at 37 °C. Cells were harvested using centrifugation (2000×g, 5 min, 4 °C). The strain was resuspended in sterile normal saline to create an alternate bacterial solution.

In vitro tolerance of LAB strains to simulated gastrointestinal conditions

Resistance to low pH

The effect of low pH (2) on the growth of five LAB strains was investigated using the methodology described by Kouadio et al. (2024). The pH of the MRS medium containing the activated strains was adjusted to 1.5. The number of viable bacteria (CFU/mL) was counted at 0 h and 3 h. The following formula was used to determine the calculated survival rate (SR%):

$$\text{SR\%}=\frac{\text{LogCFU}/\text{mL}\left(1\right)}{\text{LogCFU}/\text{mL}\left(0\right)}\ast 100$$ 1

where (0) and (1) are the viable count (CFU/mL) at 0 h and after 3 h of incubation.

Resistance to pepsin in acid conditions

The capacity of strains to withstand pepsin in acid conditions was evaluated using the method described by Wang et al., (Wang et al. 2024). Activated strains were inoculated into test tubes (0.13% NaCl; 0.02% KCl; 0.64% NaHCO₃; 0.1% pepsin; pH 2.5). After being serially diluted and collected at 0 and 3 h, the samples were plated on MRS agar. The survival rate in simulated gastric juice was calculated using the same formula to assess acidic pH resistance (Eq. 1).

Tolerance to bile salts

Activated strains were inoculated into test tubes (0.3% bile salts) and the survival rate was determined after 4 h at 37 °C. The survival rate was calculated using Eq. 1.

Resistance to trypsin and bile salts

The activated strains were inoculated into simulated gastric juice and cultured at 37 °C and 50 rpm for 3 h. The pH of the solution was then adjusted to 7.5 with HCl (1 M), followed by the addition of 1% trypsin and 0.3% bile salts to simulate intestinal fluid. The mixture was cultured at 37 °C and 50 rpm for an additional 4 h. The viable counts in simulated gastric juice at 0 h and 3 h, as well as in simulated intestinal fluid at 0 h and 4 h, were calculated (Assis et al. 2024). The survival rate was determined using Eq. 1.

Fermented soy whey (FSW) preparation

Fresh soy whey was processed within 12 h. After filtering, the soy whey was pasteurized for five minutes at 80 °C in sterile containers and then kept at 4 °C for storage. Afterward, 50 mL of soy whey was mixed with 100 μL of the LAB culture, and the mixture was cultured for five days at 37 °C and 50 rpm. 5 mL of fermented soy whey (FSW) was collected every 24 h for the following tests (Qi et al. 2021). The original SW served as the control.

Biochemical analysis

Determination of OD₆₀₀ nm

The growth curves of five LAB strains during fermentation were determined using a wavelength of visible light at 600 nm (Xiang 2021).

Determination of pH and total acidity

The pH value of samples was measured using a pH meter (Shanghai Lei Magnetic Instrument Co., Ltd., China). The titratable acidity of the samples was measured following the method of Plangpin and Athapol (2023). Briefly, using phenolphthalein as an indicator, each sample was titrated until the solution turned pink and remained so for 30 s. The titratable acidity was calculated using the following formula:

$$\text{C}=\frac{0.1\text{N NaOH}\times \text{K}\times \text{F}\times (\text{V}1)}{\text{V}}\times 1000$$ 2

where F is the dilution ratio of the fermentation solution, K is the acid conversion factor (K = 0.090), V1 is the titration volume of NaOH (mL), and V is the fermentation liquid volume (mL).

Determination of total sugar and reducing sugar content

Total sugar content was measured following Yan and Zhang with modifications (Yan et al. 2023). Absorbance at 490 nm was used to calculate total sugar content from a glucose standard curve.

Reducing sugar content was measured following Hu et al. (2008). Briefly, 500 μL A (FSW) and 1000 μL B (DNS solution) were mixed in a tube and heated in a boiling water bath for 15 min. The absorbance at 520 nm was measured following the addition of 5 mL of distilled water.

Determination of total phenols and flavonoids

Determination of total phenol content

The total phenolic content was measured following Yan et al. (2023) with minor adjustments. In summary, distilled water was used to dilute 10 µL of the samples ten times. Subsequently, the diluted sample was mixed with 2 mL of Folin-Ciocalteu and 2 mL of Na₂CO₃, and it was left at room temperature in the dark for 40 min. At 760 nm, absorbance was measured. The samples’ results were presented as milligrams of gallic acid.

Determination of total flavone content

The method described by Xu et al. (Xu et al. 2023a) was used to calculate the total phenolic flavonoid. In short, 0.5 mL of NaNO₂ was added after 1 ml of sample had been combined with 2 mL of distilled water. One mL of AlCl₃ was added after 5 min, stirred, and allowed to incubate for an additional 5 min. Following that, 2 mL of a 2 mol/L NaOH solution was added, and the mixture was given 5 min to react. The absorbance of the reaction mixture was determined at 510 nm. The samples’ results were presented as milligrams of rutin equivalent (RE).

Determination of soy isoflavones content

Soy isoflavones were quantified using the methods of Chen (Chen et al. 2022b) and coworkers. The supernatant obtained after analysis was subjected to high-performance liquid chromatography (HPLC). Twenty microliters of the filtered supernatant were injected into the HPLC system with a C18 column (InertSustaain AQ-C18 5 um, Columns 4.6*250 mm) at 30 ℃ using the SPD-M20A detector. The mobile phase consisted of water (A) with 0.5% glacial acetic acid and acetonitrile (B), as detailed in Table S1.

The results are expressed based on the daidzin, glycitin, genistin, daidzein, glycitein, and genistein standard curves.

Antioxidant activity assays

Determination of ABTS (2, 2'-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid)) radical scavenging activity

ABTS radical scavenging activity was assessed by applying a modified method from Qi et al. (2021) method. Briefly, 7 mmol/L ABTS and 2.45 mmol/L K₂S₂O₈ were combined and left to incubate for 16 h in the dark. After that, ethanol was used to dilute the solution (absorbance at 734 nm was 0.70 ± 0.02). After 6 min of reaction between a 0.6 mL sample and 1.8 mL of the ABTS solution, absorbance at 734 nm was measured. The formula below was used to determine the ABTS radical scavenging percentage:

$$\text{ABTS radical scavenging activity}\left(\%\right)=\left(\frac{Acontrol-Asample}{\mathit{Acontrol}}\right)\times 100$$ 3

Determination of DPPH (1,1-diphenyl-2-picryl-hydrazyl radical) radical scavenging activity

DPPH radical scavenging activity was assessed using the method described by Tonolo et al. and absorbance at 517 nm (Tonolo et al. 2023). The DPPH radical scavenging percentage was calculated using Eq. 3.

Determination of superoxide dismutase (SOD) activity

The SOD inhibition rate was determined following the method outlined by Li et al. (2021). At approximately 25℃, 450 µL of Tris–HCl (50 mM with 1 mM EDTA·2Na (ethylene diamine tetra acetic acid·2Na), pH = 8.2) and 100µL of fermented SW were mixed, followed by the addition of 40 µL of B (25 mM pyrogallol-HCl solution) and incubated for 25 min at room temperature. The reaction was halted by adding 8 mM HCl. Absorbance was then measured at 325 nm. In the control group, an equal amount of water was used instead of fermented soy whey. The SOD activity was calculated using Eq. 3.

Determination of volatile compounds

The volatile compounds in soy whey and FSW extracts were quantified using headspace solid-phase microextraction coupled (Agilent, USA) with SPME–GC–MS (Supelco, USA), compared with the NIST database. Quantification used the internal standard method with 2-methyl-3-heptanone. Specifically, four microliters of the internal standard 1-heptanol (16.32 ng/µL) were added to 4 mL of fermented SW samples. For 45 min at 60 °C, the SPME fiber was placed inside the vial for extraction. The full-scan assay followed Wu et al. (2022), and compound concentrations were calculated using the formula:

$${\rho }_{\text{x}}=\frac{Ax\times \rho \text{i}}{\mathit{Ai}}$$ 4

where ρx (ng/mL) and Ax indicate the concentration and peak area of the substance to be measured, respectively, and ρi and Ai denote the concentration and peak area of the internal standard, respectively.

Sensory evaluation

A quantitative descriptive analysis was conducted by a panel of 10 experienced sensory evaluators (5 men and 5 women, aged 20–30 years, from Beijing Technology and Business University) and approved by the Ethics Review Committee of Beijing Technology and Business University (approval date 9/20/2024). The order in which the samples were displayed was random. Panelists had to rate the samples on a scale of 0 (not satisfied) to 5 (very satisfied). Five sensory characteristics were identified: flavor, taste, texture, color, and overall acceptability (Jiang et al. 2024). This study was reviewed and approved by the Beijing Technology and Business University and informed consent was obtained from each subject before they participated in the study.

Statistical analysis

All the samples were tested in triplicate. The data were analyzed using one-way ANOVA in IBM SPSS Statistics 26 (IBM, New York, USA). Tukey’s test was performed to determine if P < 0.05 indicated a significant difference. Origin 2021b (Origin Lab, Massachusetts, USA) was utilized for data analysis.

Results and discussion

Resistance to simulated gastrointestinal tract conditions

A high tolerance to acidic conditions (pH 2) in the stomach is crucial for the effective probiotic function of LAB, as it ensures their survival and activity during digestion (Kouadio et al. 2024). All tested LAB strains maintained survival rates above 73% under simulated gastrointestinal conditions (Table 1). Strain 7–16 showed the lowest acid resistance (79.01 ± 1.78% at pH 2), while strain 1–16 demonstrated the lowest bile salt tolerance (73.58 ± 2.05%). Notably, strain 1–1 consistently displayed the highest survival rates across all stress conditions (88.30–97.96%). The low pH of natural FSW, which enables LAB to flourish in acidic settings, is likely the reason for the strong acid tolerance of the LAB strains isolated from natural FSW. All of the LAB strains in this investigation showed strong growth at pH 2.5 and in the presence of 0.3% pepsin, in addition to their tolerance to low pH, underscoring their remarkable resistance to severe gastric conditions.

Table 1.

Resistance of five kinds of LAB strains to simulated gastrointestinal conditions

Survival Rate (%)
Strains	pH 2	Pepsin (0.1%)/pH 2.5	Bile Salts (0.3%)	Trypsin (1%)/bile salts (0.3%)
3–1	83.33 ± 0.91b	94.08 ± 1.29b	80.62 ± 3.66a	85.12 ± 2.80a
2–1	81.35 ± 0.51b	91.58 ± 2.88bc	75.31 ± 3.64b	80.91 ± 0.44a
1–16	81.18 ± 1.05bc	89.98 ± 2.41cd	73.58 ± 2.05b	79.63 ± 2.07a
7–16	79.01 ± 1.78c	86.77 ± 1.78d	74.18 ± 1.19b	82.33 ± 2.14a
1–1	88.30 ± 0.81a	97.96 ± 0.84a	76.63 ± 0.75ab	83.59 ± 3.01a

Each value represents the mean of three replicates, accompanied by their respective standard deviations. Values bearing the same lowercase letter (a, b, c, d, e, etc.) are not statistically significantly different at the 5% level within the same column

Furthermore, all strains exhibited survival rates exceeding 70% when exposed to digestive enzymes simulating small intestinal conditions, including trypsin (1%) and bile salts (0.3%). The acid tolerance and enzymatic resistance of LAB are critical for delivering probiotic benefits, as these properties ensure both gastrointestinal survival and intestinal colonization capacity (Kouadio et al. 2024). While strain 1–1 consistently achieved the highest survival rates under acidic and enzymatic stresses, strain 3–1 also maintained high viability (> 80%) in bile salt challenges. To further investigate the adaptation of these strains in soy whey and the characterization of the FSW, all five strains were used as the starter microorganisms for the pure culture fermentation of soy whey comparatively.

Effect of fermentation on the growth of LAB

Compared to other strains, the dominant LAB isolated from natural FSW demonstrated better adaptation to soy whey. Figure 1a shows the growth curves (OD600) of the five LAB strains during fermentation. Notably, all strains exhibited immediate growth without a detectable lag phase, suggesting their inherent adaptation to the soy whey environment. After 24 h fermentation, the OD600 values reached approximately 1.89, indicating robust microbial proliferation.

Fig. 1.

Metabolic activity and substrate utilization during soy whey fermentation by five LAB strains. a Bacterial growth (OD600nm), b pH value, c Titratable acidity, d total sugar (phenol–sulfuric acid method) and e Reducing sugar content (DNS assay). Values are presented as mean ± standard deviation. Error bars represent standard deviations

This suggests that soy whey provides the perfect conditions for LAB growth, allowing them to multiply greatly. The results were consistent with the findings for Lactobacillus casei YQ336 (Qian et al. 2023). The remarkable adaptability to soy whey was demonstrated by the result that strains 1–1 and 2–1 grew even after 24 h of fermentation, with the OD600 values only starting to drop after 48 h. The LAB strains selected from natural FSW exhibited superior growth and adaptation in soy whey fermentation (Feng et al. 2024).

Effect of fermentation on pH and titration acidity

pH and titration acidity plays a crucial role in assessing the fermentation process (Xu et al. 2023b). Acid production decreases substrate pH and increases titration acidity levels.

Figures 1b and c show the differences in the pH and titration acidity between five different LAB. After 24 h of fermentation with LAB, the pH of soy whey significantly (p < 0.05) decreased from 6.05 ± 1.10 to between 3.89 and 4.19. This resulted in an increase in titratable acidity from 14.4 ± 1.8 g/L to about 41.40–63.00 g/L. pH and titratable acidity stabilized after 24 h of fermentation. The pH value and titration acidity of 1–1 exhibited the most significant changes, and its OD₆₀₀ nm was likewise the highest, which was consistent with our previous research on the remarkable adaptability of Lactobacillus plantarum at low pH values (Zhao et al., 2024). Acids produced during fermentation have the potential to function as antibacterial agents, protecting fermented beverages in the process (Qian et al. 2023).

Effect of fermentation on carbohydrate content

Figure 1d and e has shown the changes in soy whey's total sugar and decreasing sugar content following fermentation. While the lowering sugar levels varied based on the starter microbe, the total sugar content of FSW reduced considerably after 24 h of fermentation. The decline in total sugar was due to LAB entering the logarithmic growth phase and metabolizing sugars into acids, increasing acid production and reducing sugar content. While strains 1–1 and 2–1 exhibited a gradual decline in reducing sugars, strains 3–1, 1–16, and 7–16 showed a paradoxical increase in reducing sugar content (p < 0.05). We hypothesize that these strains may hydrolyze soy whey oligosaccharides into monosaccharides through extracellular enzymatic activity, thereby transiently elevating reducing sugar levels before complete catabolism (Tindjau et al. 2023a).

The results demonstrated a rapid metabolic activity of the LAB within the first 96 h, during which they effectively utilized the carbon source present in the soy whey, in conjunction with changes in the pH and sugar content of the FSW. As a result, the pH and total sugar content quickly decreased. Furthermore, the simultaneous depletion of carbon sources and the pH decline resulted in a considerable reduction and stabilization of the microbes' metabolic activity rates (Wu et al. 2024).

Effect of fermentation on total phenols and flavonoids compounds

Phenolic compounds are known for their antioxidant and anti-inflammatory health benefits, primarily due to their ability to donate hydrogen atoms or electrons to free radicals (Indira et al. 2024). As shown in Fig. 2a, the total phenolic content in FSW increased significantly after LAB fermentation (p < 0.05). As fermentation progressed, the total phenol content reached its peak at 24 h, increasing from 54 ± 1.38 mg GAE/L to a range of 63.91–67.72 mg GAE/L. Subsequently, a transient decline occurred at 48 h, possibly due to microbial metabolic shifts or substrate depletion during the mid-fermentation phase. The total phenolic content then gradually recovered and stabilized between 48 and 96 h, likely through secondary metabolite production or microbial adaptation. This phenomenon is similar to the study by Qi et al. (2021). During fermentation, LAB produces β-glucosidase that hydrolyzes sugars, which plays a key role in increasing polyphenol content (). Meanwhile, Qian's study has also mentioned that β-glucosidase is the primary enzyme responsible for carbohydrate cleavage during microbial fermentation. It enhances the bioactivity of phenolic compounds and increases the phenolics in soy whey. Thus, the high total phenolic content observed in the FSW samples in this study can be attributed to the production and increased bioactivity of free phenols and flavonoids during fermentation. Furthermore, phenolic compounds in beverages are generally stable, and processing or storage does not significantly accelerate their degradation (Shahidi et al. 2007).

Fig. 2.

Changes in bioactive compounds and antioxidant capacities during soy whey fermentation by five LAB strains. a Total phenolics content, b total flavonoid content, c antioxidant activities based on ABTS radical clearance rate, d DPPH radical clearance rate, and e SOD activation of soy whey during fermentation by five lactic acid bacteria strains. Values are presented as mean ± standard deviation. Error bars represent standard deviations

Soy whey contains a variety of flavonoids, which have strong antioxidant and anti-tumor properties and can scavenge free radicals (Indira et al. 2024). The total flavonoid concentration increased throughout fermentation and stabilized after around 72 h. The sample with the highest capacity to convert flavonoids was 3–1, which during 48 h of fermentation reached a maximum value of 55.66 ± 3.29 mg RE/L. The increase in total phenol compounds (TFC) can be attributed to the production of enzymes and organic acids by LAB during the fermentation of soy whey, which facilitates the release of flavonoids from their bound forms. The metabolic activity of LAB could also contribute to the generation of new flavonoid molecules (Yang and Zhou 2023; Tu et al. 2019).

Effect of fermentation on isoflavones compounds

Figure 3 illustrates the variations in isoflavone content. In the unfermented soy whey, the concentrations of the six isoflavones were as follows: daidzin (8.73 mg/L), glycitin (5.39 mg/L), genistin (1.48 mg/L), daidzein (1.46 mg/L), glycitein (2.51 mg/L), and genistein (2.2 mg/L). All five strains reduced the total isoflavone glycosides, where the 7–16 led to a 47% reduction, while other strains achieved at least 38% reduction. Concurrently, the aglycone content increased dramatically, reaching peak total levels of 40.52 mg/L (1–16), 40.38 mg/L (2–1), 38.45 mg/L (7–16), 37.72 mg/L (3–1), and 36.38 mg/L (1–1) after 96-h fermentation. The decrease in isoflavone glycosides and the rise in isoflavone aglycones indicate that β-glycosidase catalyzes the cleavage of the glycosidic link, converting glycosides to aglycones (Yao et al. 2024). The results of this experiment proved that 1–1, 2–1, and 3–1 had better abilities to convert isoflavones. Meanwhile, the total aglycones content produced by the five LAB strains in this experiment not only exceeded the 31.83 mg/L reported by Tindjau et al. but also reached this level after approximately 96 h of fermentation, whereas Tindjau et al. achieved it after 8 days (Tindjau et al. 2024). This result suggests that LAB strains isolated from natural FSW are not only highly adaptable but also possess a robust ability to convert isoflavones. These capabilities allow them to convert soy whey into beverages with higher nutritional value.

Fig. 3.

Content of soy isoflavones content during 0–96 h fermentation with five LAB strains of soy whey. a daidzin; b glycitin, c genistin, d daidzein, e glycitein, and f genistein. Values: mean ± SD (n = 3). Different letters indicate significant differences (P < 0.05)

Changes in antioxidant activities during fermentation

Three indicators—ABTS clearance rate, DPPH clearance rate, and SOD activation—were used to assess the antioxidant activity of soy whey both before and after fermentation. The results, shown in Fig. 2c, d, and e, revealed that after fermentation with LAB, the ABTS and DPPH clearance rate of soy whey increased significantly (p < 0.05). However, the SOD activity decreased after fermentation for 24 h. This may be due to the progress of fermentation, during which LAB produces a significant number of acids (such as acetic acid and lactic acid). This acid production leads to a decrease in pH, altering the optimal pH for SOD enzymes and thereby inhibiting SOD activity.

According to a previous study by Qian et al. (2023), two main factors were identified that contributed to the enhanced antioxidant activity resulting from the LAB fermentation of soy whey. First, LABs themselves possess certain antioxidant activity. Second, during fermentation, LAB releases flavonoids, polyphenols, and organic acids, all of which further enhance antioxidant activity. The five strains of LAB used in this study were isolated from naturally FSW thus they were well adapted to soy whey, which contributed to their strong ABTS and DPPH radical scavenging abilities.

To further explore the relationship between antioxidant activity and LAB metabolites in soy whey, we conducted a Pearson correlation analysis. As shown in Fig. 4a, titratable acidity, daidzein, and genistein were positively correlated with antioxidant activity. It indicates that the improvement of antioxidant activities by LAB should be mainly attributed to the enhanced titratable acidity, daidzein, and genistein levels in fermented soy whey. Tindjau et al. reported that the antioxidant activity of the soy whey beverage was significantly enhanced by fermentation with Bifidobacterium animalis (Tindjau et al. 2023b). The findings showed that the soy whey beverage's improved nutritional profile and increased antioxidant capacity were both influenced by the rise in isoflavone aglycone during LAB fermentation. These results are consistent with the current study's findings.

Fig. 4.

Changes in physicochemical properties and sensory characteristics of soy whey during fermentation by five LAB strains. a Thermograph of correlations among physicochemical indices and active substances. “*” indicates a significant correlation (p < 0 .05), “**” represents a very significantly positive correlation (p < 0.01). b Comparison of the number of volatile compounds during fermentation by five lactic acid bacteria strains. c Comparison of volatile compounds content of soy whey during fermentation by five lactic acid bacteria strains. d Quantitative descriptive analysis on the sensory profile of the soy whey at 72 h

Changes in volatile flavor components in FSW based on GC–MS

The volatile flavor components in unfermented and fermented samples can be found in Table S2 and Table S3. A total of 33 volatile flavor substances were detected, including 8 alcohols, 8 aldehydes, 7 ketones, 5 esters, 4 acids, and 1 phenolics. Figure 4b and c clearly illustrate the changes in both the quantity and composition of flavor substances following the fermentation of soy whey by different starter microorganisms. Each of the five LAB strains differed in their ability to improve flavor, with 3–1 performing well, as it produces more flavor substances. Overall, fermentation resulted in the production of aromatic aldehydes (such as nonanal and decanal) and esters (such as ethyl hexanoate and butyl acetate), while simultaneously reducing the concentration of "beany" aldehydes (such as hexanal). These findings align with the results reported by Tindjau et al. (2024).

Content of Hexanal in FSW were significantly reduced or even metabolized completely after fermentation (Table S3). Alongside this decrease in aldehydes, the equivalent acids or alcohols increased. For example, hexanal may undergo enzymatic oxidation to produce hexanoic acid or reduction to hexanol by an enzymatic reduction reaction (Tindjau et al. 2024). Nonanal, Octanal, 1-nonanal, and decanal with fruity and fresh flavors, respectively, were produced following LAB fermentation (Yuan et al. 2024). Additionally, strains 2–1, 3–1, and 1–1 contained more nonanal at the end of fermentation.

Acetic acid was the primary volatile compound in soy whey fermentation, accounting for approximately 80% of the total acid content. It emerged as the most predominant acid in FSW produced by LAB, likely due to the acetate kinase pathway of the phosphogluconate pathway or the metabolic conversion of citric acid. Acetic acid is characterized by a slightly sweet flavor, but it also possesses a strong, irritating odor (Wu et al. 2022; Bai et al. 2024). Its high concentration and distinct sensory properties play a crucial role in shaping the overall flavor profile of the FSW.

Esters are typically created during fermentation through the esterification and dehydration reactions between organic acids and alcohols. One of the main features of the LAB fermentation of soy whey is the rise in ester concentrations. LAB facilitates the formation of esters by catalyzing the conversion of acids, like acetic acid and hexanoic acid into their corresponding esters, like ethyl acetate and ethyl hexanoate. The final fermented product's fruity, sweet, and aromatic flavors are influenced by these esters. These ester molecules contribute significantly to the development of a more complex and appealing flavor profile in addition to improving the sensory quality of soy whey (Dai et al. 2023).

According to Chen et al. (2019), alcohols are essential for flavor formation since they provide sophisticated scents and act as solvents for other aromatic molecules. During the LAB fermentation of soy whey, the alcohol content significantly increased due to the fermentation of carbohydrates (Dai et al. 2023). Significant increases in 1-hexanol and ethanol were produced during the fermentation process, which adds to FSW's distinctive sweetness. Among the LAB strains, 3–1 exhibited a substantial increase in ethanol (from 34.42 ± 8.90 to 102.94 ± 42.25 ng/g), while 1–16 produced the highest 2-hexadecanol concentration (6.52 ± 2.49 ng/g at 96 h). The stimulation of the phosphogluconate pathway's acetate kinase pathway by Lactobacillus during fermentation is primarily responsible for the rise in alcohol levels. This metabolic shift alters the availability of external electrons, facilitating the reduction of aldehydes to their corresponding alcohols (Chen et al., 2024). The production of alcohols during fermentation highlights the important role of LAB in shaping the sensory characteristics of the FSW.

Six novel ketone molecules, including acetoin (3-hydroxybutanone), 6-methyl-5-hepten-2-one, 2-pentanone, 4-s-butoxy-2-butanone, 2-nonanone, and nerylacetone, were found in the FSW during fermentation. The contents of acetone and acetoin were significantly increased in the FSW, which contributed to the development of fruity, sweet, and creamy flavor notes in soy whey. The elevated ketone content in soy whey fermented beverages is likely a result of decarboxylation or oxidation processes mediated by LAB (Qi et al. 2021). Ketones are known for their buttery and creamy flavors, while acetone adds subtle fruity and floral characteristics. Consequently, the ketones produced during LAB fermentation enhance the overall sensory profile of soy whey, making its flavor more complex and enticing.

Sensory evaluation

Combining the above indicators, FSW for 72 h was chosen for sensory evaluation. Panelists assessed sensory attributes-flavor, taste, texture, color, and overall acceptability using a 5-point hedonic scale. The taste of the FSW varied with different LAB strains. As shown in Fig. 4d, the organoleptic properties of the fermented soy whey are significantly improved compared to the unfermented soy whey. 1–1 has the lowest acceptability score (below 4), most likely because it creates a high amount of acid, resulting in an extremely strong sour flavor. In contrast, sample 3–1 achieved the highest composite score for sensory attributes, likely due to the elevated production of esters during fermentation. This enhanced the flavor profile, resulting in a more desirable and balanced taste with fruity, floral, and complex notes.

Conclusions

Five different LAB strains isolated from natural FSW exhibited good intestinal tolerance and fermentation performance, indicating that natural FSW could be a valuable source of probiotics. Each of the five LAB strains shows good adaptation for the soy whey and various abilities to produce acid and transform the bioactive compounds and volatile substances. Comparative analysis of the volatile profile of pure-culture FSW and unfermented soy whey revealed that fermentation by the native LAB significantly enhances the flavor of FSW by reducing hexanal levels and promoting ester formation. Moreover, fermentation promotes the conversion of glycosidic isoflavones to aglycone isoflavones, thereby enhancing the antioxidant activity of the FSW. This study fully demonstrates that the preparation of FSW beverages through native LAB fermentation is an effective method for the reuse of soy whey.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This study was financially supported by the Program for the National Natural Science Foundation of China (No.32272253, 31771960).

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Declarations

Conflict of interest

The authors declare no competing financial interest.