Abstract
Rapeseed meal is rich in protein, but its high-value utilization remains insufficient. In this study, high-temperature steam pretreatment (HTSP) was applied to rapeseed meal to reduce safety- and flavor-limiting factors and improve its suitability as a raw material for soy sauce fermentation. Compared with the CON group, HTSP significantly reduced the contents of glucosinolates, isothiocyanates, erucic acid, and phenolic-related compounds in rapeseed meal, with glucosinolates and isothiocyanates decreasing by 59.02% and 92.65%, respectively. Meanwhile, the contents of protein, soluble protein, and reducing sugars increased to 39.28, 16.42, and 1.18 g/100 g, corresponding to 1.11-, 2.01-, and 1.97-fold those of the CON group, respectively. Subsequently, HTSP-treated rapeseed meal was used as the main raw material for soy sauce fermentation. During the 1–15 d fermentation period, the sample fermented for 15 d showed the most favorable physicochemical and flavor characteristics. At this stage, the contents of total nitrogen, amino acid nitrogen, reducing sugars, total acidity, and soluble solids reached 1.16, 0.70, 2.13, 1.22, and 34.31 g/100 mL, corresponding to 4.30-, 3.33-, 3.23-, 30.50-, and 4.73-fold those on day 1, respectively. During fermentation, the taste profile shifted from bitterness and sourness toward sweetness and umami, while the aroma profile evolved from fruity notes to nutty, roasted, and floral notes. Correlation analysis indicated that Weissella and Zygosaccharomyces rouxii were important microbial taxa closely associated with the quality and flavor development of rapeseed meal soy sauce, and may contribute to the accumulation of taste compounds in the early stage and aroma formation in the middle and late stages, respectively. Overall, HTSP effectively enhanced the utilization potential of rapeseed meal as a fermentation substrate, supporting its high-value application in soy sauce brewing.
Keywords: Rapeseed meal, High-temperature steam pretreatment, Soy sauce, Weissella, Zygosaccharomyces rouxii
Highlights
-
•
High-temperature steam pretreatment enhanced fermentation ability of rapeseed meal.
-
•
On day 15, the sweetness and umami of the rapeseed meal soy sauce (SY) were enhanced.
-
•
Salt-tolerant bacteria and yeasts were the main microbes responsible for SY's aroma.
1. Introduction
Rapeseed is one of the most important oil crops in China and is widely processed for rapeseed oil production. As the third-largest edible vegetable oil worldwide, after soybean oil and palm oil, rapeseed oil occupies an important position in the global edible oil supply system. With the continuous development of the rapeseed oil industry, large quantities of rapeseed meal, the major by-product generated during industrial oil extraction, are produced annually (Xu et al., 2025). It has been estimated that approximately 51 million tons of rapeseed meal are generated each year by the global rapeseed processing industry, with an average market price of approximately 308.71–370.74 USD per ton (Betchem, Dabbour, Akter Tuly, Flavorta Billong, & Ma, 2024). These data suggest that rapeseed meal is an abundant and cost-effective oilseed meal resource. Rapeseed meal contains 34.4%-47.3% crude protein and has a relatively balanced amino acid profile, making it a promising plant protein resource (Du et al., 2025). However, its current utilization remains largely confined to the feed industry, whereas its high-value application in food and other related fields is still limited. This limitation is mainly attributed to the presence of antinutritional factors, pronounced bitter and astringent flavors, and relatively high crude fiber content. Glucosinolates are among the major antinutritional factors in rapeseed meal and exhibit relatively low intrinsic toxicity (Yan et al., 2024). However, they can be hydrolyzed by myrosinase to form isothiocyanates, which have been associated with impaired animal growth performance and thyroid dysfunction. Meanwhile, phenolic compounds such as sinapine and tannins can interact with proteins to form complexes, thereby reducing protein digestibility and contributing to the characteristic bitterness and astringency of rapeseed meal. In addition, rapeseed meal contains approximately 10%-15% crude fiber, and its compact fiber matrix may hinder protein release, reduce enzyme accessibility to protein substrates, and consequently impair protein digestibility and processing suitability (Liu, Jacquet, Xie, Jiang, & Blecker, 2025). Therefore, rapeseed meal should not be regarded simply as a high-protein raw material, but rather as a protein-rich resource whose utilization is jointly constrained by antinutritional factors, phenolic compounds, and the fiber matrix. In this context, reducing antinutritional factors, mitigating undesirable flavors, alleviating the barrier effect of the fiber matrix on protein release and utilization, and improving protein availability have become key scientific issues that need to be addressed to promote the high-value utilization of rapeseed meal.
In recent years, increasing evidence has shown that fermentation is an effective strategy for reducing antinutritional factors in rapeseed meal and improving its undesirable flavor attributes (Huang, Lin, Wang, Zhao, & Feng, 2025). For example, the co-fermentation of rapeseed meal with lactic acid bacteria, Bacillus subtilis, and Saccharomyces cerevisiae has been reported to reduce glucosinolate content by more than 90% and partially alleviate bitterness. In addition, submerged fermentation of rapeseed meal by Bacillus subtilis has been shown to markedly degrade sinapine, with reductions of 97.8% and 93.3% in the supernatant and solid residue, respectively, after 72 h of fermentation (Jin et al., 2014). These findings suggest that fermentation may be a promising approach for improving the nutritional and processing suitability of rapeseed meal. Among plant protein-based fermented foods, soy sauce is one of the most representative products. Its mature brewing process and characteristic flavor profile provide an important reference for the high-value utilization of plant protein resources. Traditional soy sauce is produced mainly using soybean as the core protein raw material. Soybean protein not only supplies nitrogen sources for microbial growth but also provides key precursors for the formation of amino acids and flavor compounds (Hong et al., 2025). However, the long-term dependence on soybean as the major protein substrate may increase raw material pressure in the soy sauce industry. Therefore, exploring alternative plant protein substrates for soy sauce brewing has become an important direction for product innovation and raw material diversification. In this context, the feasibility of using rapeseed meal as a novel protein substrate for soy sauce brewing deserves further investigation. Notably, rapeseed meal differs substantially from soybean in protein composition and molecular structure, which may affect substrate degradation during soy sauce fermentation. Soybean proteins mainly consist of glycinin, an 11S globulin, and β-conglycinin, a 7S globulin, whereas rapeseed meal proteins are predominantly composed of cruciferin, an 11S globulin, and napin, a 2S albumin (Lu, Alenyorege, Ouyang, Zhou, & Ma, 2022). Compared with soybean proteins, rapeseed meal proteins generally exhibit greater structural heterogeneity, stronger hydrophobicity, and a more compact and stable molecular conformation, which may limit the accessibility of microorganisms and their secreted enzymes to protein substrates. Furthermore, protein secondary structure may also influence enzymatic hydrolysis and digestibility. Previous studies have shown that a higher proportion of β-sheet structures in proteins is generally associated with reduced protein digestibility. The β-sheet contents of cruciferin and napin in rapeseed meal protein are approximately 31% and 38%, respectively, whereas soybean protein contains a relatively lower proportion of β-sheet structures, approximately 19% (San et al., 2026). These differences in secondary structure may partly explain the lower enzymatic hydrolysis efficiency and amino acid release capacity of rapeseed meal protein compared with soybean protein. Therefore, it is necessary to develop efficient pretreatment technologies suitable for rapeseed meal to improve the enzymatic accessibility of its protein substrates. Further evaluating the potential of rapeseed meal as a raw material for soy sauce brewing is important for broadening the raw material base of soy sauce production and promoting the high-value utilization of this underused oilseed meal resource.
Physical pretreatments have been widely used as effective approaches for modifying oilseed cakes and meals. These treatments can improve the structural and functional properties of raw materials, alleviate the barrier effect of the fiber matrix on protein release and utilization, and thereby provide new strategies for enhancing the fermentative utilization of oilseed meal substrates. For example, steam explosion has been applied to defatted soybean meal and was found to disrupt the compact fiber-protein complex structure, promote polysaccharide degradation and protein depolymerization, and significantly increase the release of reducing sugars and available nitrogen sources, thereby making the substrate more accessible to microorganisms (Zhao et al., 2025; Zhao et al., 2025). Furthermore, soy sauce prepared from steam explosion-treated soybean meal exhibited an amino acid nitrogen content approximately twice that of soy sauce prepared from untreated soybean meal, while the fermentation period was markedly shortened from the conventional 6–12 months to approximately 45 days (Zhang et al., 2022). In addition to steam explosion, HTSP has also shown considerable potential for improving the fermentation suitability of oilseed by-products. For instance, when HTSP was applied to walnut meal, the amino acid nitrogen content of the resulting walnut meal soy sauce reached 1.07 g/100 mL, representing a 42.70% increase compared with the untreated group. Notably, umami amino acids increased by 71.91% after HTSP treatment (Tan et al., 2025). Mechanistically, steam explosion mainly depends on the penetration of saturated steam into the material, followed by rapid depressurization, which induces intensive structural disruption. However, this rapid pressure-release process may also cause the loss of some flavor-related compounds through flash evaporation or entrainment (Zheng et al., 2025). In contrast, HTSP transfers heat to the protein-rich matrix through steam and regulates material structure through the vaporization of internal moisture, thereby improving the physicochemical and nutritional properties of protein-rich raw materials. Moreover, because HTSP does not involve violent instantaneous depressurization, it may reduce the loss of flavor compounds to some extent and thus show potential advantages in flavor retention and enhancement. In addition, HTSP has the advantages of relatively low equipment cost and simple operation, indicating its potential for large-scale processing and utilization of agro-industrial by-products. However, studies on the application of HTSP to improve the structural properties of rapeseed meal and enhance its suitability as a fermentable substrate remain limited.
To address these challenges, HTSP was applied to rapeseed meal in this study to reduce the contents of antinutritional factors and undesirable flavor-related compounds, alleviate the barrier effect of the fiber matrix on protein release and utilization, and thereby improve its suitability as a fermentable substrate. Subsequently, HTSP-treated rapeseed meal was used as the main raw material and inoculated with a composite starter culture consisting of Aspergillus oryzae strain JY309 and Zygosaccharomyces rouxii for koji preparation, followed by moromi fermentation to produce rapeseed meal soy sauce. Ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) and headspace gas chromatography-ion mobility spectrometry (HS-GC-IMS) were used to characterize the dynamic changes in non-volatile and volatile compounds, respectively, during different fermentation stages on days 1, 5, 10, and 15. Meanwhile, microbial community analysis was performed to elucidate the succession of dominant microbial taxa during rapeseed meal soy sauce fermentation and to explore their relationships with physicochemical parameters and flavor quality. This study provides a theoretical and technical basis for the high-value utilization and green processing of oilseed cake and meal resources, such as rapeseed meal, in fermented condiments.
2. Materials and methods
2.1. Materials and reagents
Rapeseed meal was purchased from Luoping Fuxin Agricultural Development Co., Ltd. (Qujing, Yunnan Province, China). Aspergillus oryzae strain JY309 and Zygosaccharomyces rouxii were obtained from Angel Yeast Co., Ltd. (Shanghai, China). Methanol, lead sulfate, and phosphoric acid were purchased from Shanghai Yuanye Biotechnology Co., Ltd. (Shanghai, China). Phenylethylamine, ethyl acetate, and n-hexane were obtained from Aladdin Industrial Corporation (Shanghai, China).
2.2. Determination of antinutritional factors and safety-related compounds in rapeseed meal under HTSP conditions
The contents of glucosinolates, isothiocyanates, erucic acid, and sinapine in rapeseed meal were quantified using the external standard method with the corresponding authentic standards. Glucosinolates were extracted with preheated methanol and 70% methanol, followed by enzymatic hydrolysis with sulfatase. The resulting desulfoglucosinolates were analyzed using an Agilent 1260 HPLC system equipped with a diode array detector (DAD; Agilent Technologies, Santa Clara, CA, USA) at a detection wavelength of 235 nm. Isothiocyanates were hydrolyzed in an aqueous hydrochloric acid solution at pH 3.0, extracted with ethyl acetate, reconstituted in 50% aqueous methanol, and subsequently determined using the same HPLC-DAD system at 210 nm. Sinapine was extracted by ultrasonic-assisted extraction with methanol and analyzed using an Agilent 1200 HPLC system (Agilent Technologies, Santa Clara, CA, USA) at 326 nm. All HPLC analyses were performed using a C18 column at a column temperature of 30 °C, a flow rate of 1.0 mL/min, and an injection volume of 20 μL. Prior to analysis, all samples were filtered through a 0.22 μm membrane filter.
Erucic acid was determined after ultrasonic extraction of the lipid fraction with petroleum ether, followed by methyl esterification with methanolic KOH. The resulting fatty acid methyl esters were analyzed using an Agilent 6890 gas chromatograph (Agilent Technologies, Santa Clara, CA, USA). Separation was performed on an Agilent DB-23 capillary column (30 m × 0.25 mm × 0.25 μm; Agilent Technologies, Santa Clara, CA, USA), with N2 used as the carrier gas.
The chromatograms of the standard compounds, including glucosinolates, isothiocyanates, erucic acid, and sinapine, are shown in Fig. S1, The chromatograms of these compounds in the CON and HTSP-treated rapeseed meal samples are presented in Figs. S2-S5, respectively.
Tannin and total phenolic contents were determined using a tannin assay kit and a plant total phenol assay kit, respectively (Nanjing Jiancheng Bioengineering Institute, Nanjing, China), according to the manufacturer's instructions.
2.3. Determination of physicochemical properties of rapeseed meal under HTSP conditions
The protein, moisture, and ash contents of rapeseed meal were determined according to ISO 5983-1 (2005), ISO 8184 (2016), and ISO 665 (2000), respectively. The soluble protein content was determined using the Bradford method with a Bradford Protein Assay Kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The procedure was followed according to the manufacturer's instructions. The content of reducing sugars was determined using the 3,5-dinitrosalicylic acid (DNS) method. The procedure is as follows: 0.5 g of rapeseed meal sample were added to 10 mL of deionized water, boiled for 10 min, filtered, and the filtrate was diluted to a final volume of 100 mL. Three milliliters of the solution were transferred to a 25-mL cuvette, and 1.5 mL of DNS reagent were added, mixed thoroughly, and heated in hot water for 5 min. After rapid cooling, the volume was adjusted to 25 mL with distilled water. Finally, the absorbance was measured at a wavelength of 540 nm (Tan et al., 2025). The color of rapeseed meal was measured using a colorimeter (NR145, Shenzhen 3nh Technology Co., Ltd., Shenzhen, China).
2.4. Morphological characterization
The morphological characteristics of rapeseed meal were observed using a scanning electron microscope (SEM, SU8010, Hitachi, Tokyo, Japan). A small amount of rapeseed meal powder was fixed onto conductive adhesive tape and sputter-coated with gold using a sputter coater at 10 mA for 45 s to improve electrical conductivity. The samples were then observed under an accelerating voltage of 3 kV using the SE2 secondary electron detector. SEM micrographs were acquired at magnifications of 500×, 800×, 1000×, and 2000× to characterize the microstructural features of the samples.
2.5. Preparation of rapeseed meal soy sauce (SY)
After HTSP treatment at 120 °C and 0.103 MPa for 30 min, rapeseed meal was dried in a constant-temperature oven at 80 °C for 10 h according to a previous study (Tan, Peng, Wang, et al., 2025). The dried rapeseed meal was mixed with water at a mass ratio of 5:4 and equilibrated for 30 min. Subsequently, the hydrated rapeseed meal was blended with wheat flour at a mass ratio of 7:3 and steamed at 121 °C for 30 min. After cooling to 40 °C, the mixture was inoculated with a compound starter culture composed of Aspergillus oryzae strain JY309 and Zygosaccharomyces rouxii at a ratio of 1:1 (w/w), with an inoculation level of 0.1% (w/w), followed by koji fermentation. The koji substrate was incubated at 32 °C for 48 h and turned every 6 h. Koji samples were collected at 24, 36, and 48 h during fermentation. Dense sporulation on the koji surface was used as an indicator of the completion of koji fermentation. After koji fermentation, the resulting koji was mixed with brine (18 g/100 mL NaCl) at a ratio of 1:2 (w/w) and fermented at 42 °C. During moromi fermentation, moromi salinity was maintained by controlling both the NaCl concentration of the brine and the koji-to-brine mixing ratio. The moromi was thoroughly mixed before fermentation to ensure uniform salt distribution, and no additional salt or water was added during fermentation. Moromi samples were collected on days 1, 5, 10, and 15. For soy sauce extraction, each moromi sample was mixed with water at a ratio of 1:2 (moromi:water), followed by pressing, filtration, and clarification. The obtained SY samples were designated as SY-1, SY-5, SY-10, and SY-15, respectively.
2.6. Determination of enzyme activities during koji fermentation
To investigate the dynamic changes in enzymatic activities during the koji fermentation stage, samples were collected at 24, 36, and 48 h of fermentation. At each time point, 5 g of koji material was mixed with 30 mL of distilled water, extracted at 40 °C for 30 min, and diluted to a final volume of 100 mL to obtain the crude enzyme extract. Prior to enzyme activity assays, the extract was diluted 10-fold using different buffer solutions: lactic acid‑sodium lactate buffer (pH 3.0) for acidic protease, phosphate buffer (pH 7.2) for neutral protease, and borax‑sodium hydroxide buffer (pH 10.0) for alkaline protease and glucoamylase assays. Briefly, protease activity was measured as follows: 1 mL of the diluted enzyme solution was mixed with 1 mL of casein solution (2%, w/v) and incubated at 40 °C for 10 min. After adding 2 mL of trichloroacetic acid (0.4 mol/L), the reaction was continued for another 20 min. The mixture was then centrifuged, and 1 mL of the supernatant was mixed with 5 mL of sodium carbonate solution (0.4 mol/L) and 1 mL of Folin-Ciocalteu reagent. The reaction was conducted at 40 °C for 20 min, and the absorbance was measured at 660 nm to calculate enzyme activity. Glucoamylase activity was determined by mixing 0.1 mL of the diluted enzyme solution with 0.1 mL of soluble starch solution (1%, w/v), followed by incubation at 40 °C for 10 min. After the reaction, 1 mL of DNS reagent was added, and the mixture was boiled for 5 min. Upon cooling, 4.8 mL of distilled water was added, and the absorbance was measured at 520 nm to calculate enzyme activity.
2.7. Determination of physicochemical properties of SY
To investigate the changes in physicochemical parameters during the fermentation of SY, the contents of amino acid nitrogen, total nitrogen, reducing sugars, total acids, and total soluble solids were measured at different fermentation stages. Soy sauce samples (10 g each) were collected on days 1, 5, 10, and 15 of fermentation, extracted in a 70 °C water bath with continuous stirring for 4 h, cooled to room temperature, and then filtered. The filtrates were stored at 4 °C for subsequent analysis. The determination of amino acid nitrogen and total nitrogen was performed following previously reported methods. Briefly, the sample was titrated with 0.05 mol/L NaOH to pH 8.2, followed by the addition of 10 mL of 40% formaldehyde solution. Titration was continued to pH 9.2, and the nitrogen content was calculated based on the volume of NaOH consumed. The levels of reducing sugars, total acids, and total soluble solids were determined according to methods described in the literature (Jiang et al., 2023).
2.8. Determination of non-volatile compounds (non-VOCs) in SY by UPLC-MS/MS
An accurately weighed sample (0.5 g) was transferred into a 2 mL centrifuge tube, followed by the addition of 400 μL of methanol. The mixture was vortexed for 1 min and centrifuged at 12,000 rpm for 10 min at 4 °C. The supernatant was collected, concentrated to dryness, and reconstituted in 150 μL of 80% methanol aqueous solution containing 2-chloro-L-phenylalanine (4 mg/L) as an internal standard. The resulting solution was filtered through a 0.22-μm membrane prior to UPLC-MS/MS analysis. UPLC-MS/MS analysis was performed by Suzhou Panomix Biomedical Technology Co., Ltd. (Jiangsu, China; hereafter referred to as Panomix) using a Thermo Vanquish UHPLC system (Thermo Fisher Scientific, Waltham, MA, USA) equipped with an ACQUITY UPLC® HSS T3 column (2.1 × 100 mm, 1.8 μm) and coupled to a Thermo Q Exactive Focus mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). The flow rate was set at 0.3 mL/min, and the mobile phases and gradient elution conditions were set according to our previously reported method (Zhang et al., 2025; Zhang, Zhang, Huang, Zhou, & Wu, 2025).
MS data were acquired in both positive and negative ion modes using data-dependent acquisition (DDA). Full MS scans were acquired over an m/z range of 100–1000, and the three most intense precursor ions were selected for MS/MS fragmentation using higher-energy collisional dissociation (HCD). Dynamic exclusion was applied to avoid repeated acquisition of redundant MS/MS information.
The raw UPLC-MS/MS data were processed using the self-developed metabolomics data-processing platform of Panomix for peak extraction, peak alignment, peak integration, and normalization. Relative quantification of metabolites was performed based on normalized peak areas. Metabolite annotation was performed by matching accurate mass, retention time, and MS/MS fragmentation patterns against public databases, including HMDB, MassBank, KEGG, LipidMaps, and mzCloud, as well as the in-house metabolite database of Panomix.
In this non-targeted UPLC-MS/MS analysis, annotations not verified by authentic standards were regarded as putative. For isomeric metabolites with identical or nearly identical accurate masses and similar MS/MS fragmentation patterns, database-assigned identities were also treated as putative when isomer-specific evidence was insufficient.
Metabolite annotation confidence was classified according to the Metabolomics Standards Initiative (MSI). MSI level 1 indicated metabolites confirmed with authentic standards, representing the highest confidence level, with matched retention time, accurate mass, and MS/MS spectra under identical analytical conditions. MSI level 2 was assigned to putatively annotated metabolites with relatively high confidence; although not verified by authentic standards, these metabolites showed accurate mass and MS/MS spectral matches against public or in-house databases. MSI level 3 referred to putatively characterized compound classes or preliminary annotations based mainly on accurate mass, molecular formula, retention behavior, and/or diagnostic fragment ions, but lacked sufficient MS/MS evidence for unambiguous identification. MSI level 4 denoted unknown metabolic features for which only m/z, retention time, and peak intensity information were available; therefore, reliable compound annotation could not be achieved.
The total ion chromatograms (TICs) acquired in positive and negative ion modes are shown in Fig. S6 and Fig. S7, respectively.
2.9. Determination of volatile organic compounds (VOCs) in SY using HS-GC-IMS
The VOCs in SY were determined using HS-GC-IMS. The analysis was performed on a FlavourSpec® GC-IMS instrument (G.A.S., Dortmund, Germany) equipped with a CTC-PAL 3 static headspace autosampler (CTC Analytics AG, Zwingen, Switzerland). Chromatographic separation was carried out using an MXT-WAX capillary column (15 m × 0.53 mm, 1.0 μm; Restek, USA). Briefly, 1.0 mL of each sample was placed in a 20 mL headspace vial and incubated at 60 °C for 15 min. All samples were analyzed in triplicate. After incubation, 500 μL of headspace gas was injected in splitless mode. The agitation speed during incubation was set at 500 rpm, and the injection needle temperature was maintained at 110 °C. The GC column temperature was maintained at 60 °C, and the inlet temperature was set at 80 °C. High-purity nitrogen (purity ≥99.999%) was used as the carrier gas. The carrier gas flow rate was initially set at 2.0 mL/min and held for 2 min, then increased linearly to 10.0 mL/min within 8 min, further increased to 100.0 mL/min within 10 min, and finally maintained at 100.0 mL/min for 10 min. The total run time was 30 min.
In the IMS system, analytes were ionized in positive ion mode using a tritium (3H) ionization source. The ionized compounds were introduced into a 98 mm drift tube operated at an electric field strength of 500 V/cm and a temperature of 45 °C. High-purity nitrogen (purity ≥99.999%) was used as the drift gas at a flow rate of 150.0 mL/min. The shutter grid opening time was 100 μs, and the shutter grid closing voltage was set to 90 dgt.
A mixed standard containing six analytical-grade ketones, including 2-butanone, 2-pentanone, 2-hexanone, 2-heptanone, 2-octanone, and 2-nonanone, was analyzed to establish the retention time-retention index (RT-RI) calibration curve. The retention indices of the target VOCs were calculated based on their retention times and matched against the GC retention index database (NIST 2020) and the IMS drift time database integrated into VOCal software (version 0.4.07) for VOCs annotation. The Reporter, Gallery Plot, and Dynamic PCA plug-ins in VOCal software were used to generate three-dimensional spectra, two-dimensional spectra, difference spectra, fingerprint plots, and PCA plots. Peak signal intensities were used for the relative quantification of VOCs. The GC-IMS chromatograms of SY-1, SY-5, SY-10, and SY-15 are shown in Figs. S8-S11, respectively.
The Relative Odor Activity Value (ROAV) was used to evaluate the contribution of individual VOCs to the overall flavor of the sample. The ROAV is calculated as follows: ROAVi = 100 × (OAVi / OAVmax), where OAVmax is the highest odor activity value (OAV) among the VOCs, and OAVi is the OAV of a specific volatile compound. The OAV is calculated as: OAV = Ci / OTi, where Ci is the concentration of the volatile compound, and OTi is the threshold concentration of flavor compounds in water.
2.10. Metagenomic analysis (DNA extraction, library construction, and sequencing)
Total genomic DNA was isolated from SY using the Mag-Bind® soil DNA Kit (Omega Bio-tek, Norcross, GA, U.S.). To generate paired-end libraries, the DNA extract was sheared to an average fragment size of approximately 400 base pairs using the Covaris M220 (Gene Company Limited, China). Subsequently, paired-end sequencing was conducted on the Illumina NovaSeq platform (Illumina Inc., San Diego, CA, USA). Data analysis was carried out utilizing the Majorbio Cloud Platform (www.majorbio.com). Alpha diversity indices, including Chao1, Shannon and Simpson, were calculated using the Mothur software. Beta diversity was analyzed using QIIME, and the unweighted pair-group method with arithmetic mean (UPGMA) was applied for hierarchical clustering. Principal coordinate analysis (PCoA) and non-metric multidimensional scaling (NMDS) were employed to visualize community similarity. The Kruskal-Wallis's test was used to assess differences among groups. Venn diagrams and other microbial diversity visualizations were generated using R software packages.
2.11. Statistical analysis
All experiments were conducted with three biological replicates (n = 3), and the results are presented as mean ± standard deviation (SD), unless otherwise stated. Statistical differences among groups were analyzed by one-way analysis of variance (ANOVA), followed by Tukey's multiple comparisons test using GraphPad Prism 9 software (GraphPad Software, Boston, MA, USA). Different lowercase letters in the tables indicate significant differences among groups (p < 0.05). Error bars in the figures represent the SD of three independent biological replicates. Figures were prepared using Origin 2021 software (OriginLab, Northampton, MA, USA).
3. Results
3.1. Analysis of antinutritional factors and safety-related compounds in rapeseed meal under HTSP conditions
As shown in Table 1, compared with the CON group, HTSP treatment reduced the total glucosinolate content from 284.47 mg/kg to 116.57 mg/kg, corresponding to a decrease of 59.02%. It should be noted that no unified legal limit has been established specifically for “food-grade rapeseed meal” as a broad raw material category. However, in the specifications for some authorized partially defatted rapeseed powder as a novel food in the European Union, the limit for total glucosinolates is set at <0.3 mmol/kg, approximately equivalent to 120 mg/kg (EFSA Panel on Nutrition, Novel Foods and Food Allergens et al., 2020). In the present study, the total glucosinolate content in rapeseed meal after HTSP treatment was 116.57 mg/kg, which was below this reference level, indicating that HTSP treatment contributed to reducing the residue of this major antinutritional factor in rapeseed meal. Meanwhile, the isothiocyanate content decreased from 12.12 mg/kg to 0.89 mg/kg, with a reduction of 92.65%, suggesting that HTSP treatment also markedly reduced the residual level of glucosinolate-related degradation products. Erucic acid, as a fatty acid of safety concern in rapeseed meal, decreased from 3.22 mg/g to 2.25 mg/g, corresponding to a reduction of approximately 30.31%. According to European Union regulations, the maximum level of erucic acid in rapeseed oil is 20 mg/g. Although this limit mainly applies to oils and fats rather than rapeseed meal, the erucic acid content detected in the present study was far below this level, suggesting a relatively low erucic acid-related risk in the treated rapeseed meal.
Table 1.
Antinutritional factors and safety-related compounds in rapeseed meal.
| Antinutritional and safety-related compounds | CON | HTSP |
|---|---|---|
| Total glucosinolates (mg/kg) | 284.47 ± 0.86 | 116.57 ± 1.32 |
| Isothiocyanates (mg/kg) | 12.12 ± 0.38 | 0.89 ± 0.01 |
| Erucic acid (C22:1) (mg/g) | 3.22 ± 0.07 | 2.25 ± 0.05 |
| Sinapine (mg/g) | 10.13 ± 0.04 | 1.95 ± 0.02 |
| Tannins (mg/g) | 13.43 ± 1.04 | 10.41 ± 0.28 |
| Total phenolic content (mg/g) | 7.46 ± 0.18 | 5.50 ± 0.32 |
In addition, HTSP treatment also reduced the levels of phenolic-related compounds in rapeseed meal. The contents of sinapine, tannins, and total phenolics decreased from 10.13, 13.43, and 7.46 mg/g to 1.95, 10.41, and 5.50 mg/g, respectively, corresponding to reductions of 80.71%, 22.53%, and 26.22%. Sinapine and tannins are important phenolic-related compounds in rapeseed meal and are closely associated with bitterness, astringency, protein binding, and limited nutritional utilization (Chadni et al., 2023). The decrease in total phenolic content further suggests that HTSP treatment may alleviate the adverse effects of phenolic compounds on the processing suitability and potential flavor quality of rapeseed meal.
Overall, HTSP treatment effectively reduced the levels of total glucosinolates, isothiocyanates, erucic acid, and phenolic-related compounds in rapeseed meal, indicating that this treatment may improve the safety-related chemical characteristics and processing suitability of rapeseed meal as a raw material for food fermentation.
3.2. The analysis of the physicochemical properties of rapeseed meal under HTSP conditions
Table 2 shows the effects of HTSP on the physicochemical properties of rapeseed meal. Specifically, the contents of protein, soluble protein, reducing sugar, moisture, and ash in the CON group were 35.51, 8.16, 0.60, 6.53, and 5.28 g/100 g, respectively. After HTSP treatment, the basic composition of rapeseed meal changed markedly. The contents of protein, soluble protein, reducing sugar, and moisture increased to 39.28, 16.42, 1.18, and 10.21 g/100 g, respectively, corresponding to 1.11, 2.01, 1.97, and 1.56 times those of the CON group. In contrast, the ash content decreased to 5.02 g/100 g, corresponding to 0.95 times that of the CON group. Among these indicators, the increase in soluble protein content was the most pronounced, suggesting that HTSP may promote the dissociation, denaturation, or structural loosening of insoluble protein fractions in rapeseed meal, thereby improving their water solubility. This result is consistent with our previous findings in walnut meal, where HTSP significantly increased the soluble protein content of the raw material (Tan, Peng, Wang, et al., 2025). Furthermore, proteins with higher solubility are generally more readily utilized by microorganisms and can provide more available nitrogen sources for microbial growth and metabolism, thereby facilitating the fermentation process (Shay, Jing, Bhandari, Dayananda, & Prakash, 2023). Meanwhile, the increase in reducing sugar content may be associated with the disruption of the cell wall structure of rapeseed meal during HTSP. HTSP may promote the partial degradation of cellulose, hemicellulose, or lignin-related structures, thereby releasing reducing sugars such as glucose, xylose, and arabinose (Dhakal & Acharya, 2025). These reducing sugars can serve as carbon sources for microbial growth and may also participate in Maillard reactions during subsequent fermentation, providing precursors for flavor formation. The increased moisture content may further improve substrate hydration and mass transfer, creating more favorable conditions for microbial proliferation and enzymatic reactions. Therefore, by increasing the levels of available nitrogen sources, carbon sources, and moisture in rapeseed meal, HTSP enhanced the suitability of the raw material for fermentation and provided a basis for microbial metabolism and flavor formation during subsequent soy sauce fermentation.
Table 2.
Physicochemical indices of rapeseed meal.
| Physicochemical indices | CON | HTSP |
|---|---|---|
| Protein (g/100 g) | 35.51 ± 1.05 | 39.28 ± 1.05 |
| Soluble protein (g/100 g) | 8.16 ± 0.11 | 16.42 ± 0.42 |
| Reducing sugars (g/100 g) | 0.60 ± 0.14 | 1.18 ± 0.10 |
| Moisture (g/100 g) | 6.53 ± 0.27 | 10.21 ± 0.18 |
| Ash (g/100 g) | 5.28 ± 0.17 | 5.02 ± 0.07 |
| L* | 49.38 ± 0.42 | 32.55 ± 0.25 |
| a* | 5.26 ± 0.12 | 8.17 ± 0.17 |
| b* | 27.73 ± 0.75 | 12.50 ± 0.33 |
In addition to changes in basic composition, HTSP also affected the color characteristics of rapeseed meal. The L*, a*, and b* values of the CON group were 49.38, 5.26, and 27.73, respectively. After HTSP treatment, the L* and b* values decreased to 32.55 and 12.50, respectively, whereas the a* value increased to 8.17. These results indicate that HTSP reduced the brightness and yellowness of rapeseed meal while enhancing its redness, resulting in an overall darker and redder appearance. This color change may be related to Maillard reactions between reducing sugars and amino compounds during HTSP. Considering the increases in reducing sugar and soluble protein contents, it can be inferred that HTSP not only improved the substrate properties of rapeseed meal for fermentation but may also contribute to color development and flavor formation in subsequent soy sauce fermentation by promoting Maillard reactions or generating related intermediates.
Overall, HTSP improved the processing suitability of rapeseed meal from both compositional and color perspectives. On the one hand, HTSP increased the contents of protein, soluble protein, reducing sugar, and moisture, providing more abundant nitrogen sources, carbon sources, and a favorable reaction environment for microbial growth and metabolism. On the other hand, the color darkening induced by HTSP may be associated with Maillard reactions and is expected to contribute to color formation and flavor development during subsequent soy sauce fermentation.
3.3. The changes in the appearance and morphology of rapeseed meal after HTSP treatment
As shown in the Fig. 1, the CON group samples exhibited an irregular, granular aggregate structure, with a generally smooth surface and no obvious cracks or wrinkles. In contrast, the HTSP group displayed a more compact aggregation, which may be related to the increase in reducing sugars and soluble protein content during the treatment process. These components form stable macromolecular complexes through non-covalent interactions, thereby promoting particle aggregation. Notably, HTSP also significantly altered the surface morphology of rapeseed meal, with a marked increase in surface roughness and the formation of numerous wrinkled structures. The formation of these structural features may result from thermal expansion and internal structural rearrangement during the treatment process, which leads to significant changes in the surface morphology. From a fermentation perspective, microbial growth and metabolic activity are often closely related to their ability to adhere to the substrate surface. The complex wrinkled surface formed by HTSP provides more potential attachment sites for microorganisms, facilitating microbial colonization and biofilm formation, which helps to enhance fermentation efficiency (Sang et al., 2026). Additionally, rough surfaces generally exhibit stronger moisture absorption capacity, and adequate moisture is crucial for the growth of microorganisms, especially yeast, providing a more favorable growth environment. This is consistent with the increase in moisture content observed in the physicochemical analysis. In summary, HTSP treatment significantly altered the microstructural characteristics of rapeseed meal, potentially enhancing water absorption and improving microbial attachment on the surface, providing a structural basis for improved fermentation performance.
Fig. 1.
SEM images of rapeseed meal before and after HTSP treatment. Panels (a-d) show the CON group at magnifications of 500×, 800×, 1000×, and 2000×, respectively; panels (e-h) show the HTSP-treated group at the same magnifications. Scale bars represent 20 μm, 10 μm, 10 μm, and 4 μm, respectively.
3.4. Changes in enzymatic activities during koji preparation
The koji-making time is a critical process parameter that must be strictly controlled in soy sauce production (Gao et al., 2021). It directly affects microbial metabolic activity and enzyme secretion characteristics, thereby determining the formation of soy sauce flavor and its quality stability. Therefore, precisely controlling the koji-making time is essential for ensuring the consistency of soy sauce flavor and the controllability of product quality. Aspergillus oryzae is one of the core microorganisms in the koji-making process, capable of secreting various proteases and amylases. The proteases hydrolyze macromolecular proteins to generate amino acids and small peptides, providing important precursors for the formation of umami compounds in soy sauce. Amylases degrade starch into fermentable sugars, which not only provide a carbon source for microbial growth but also contribute to the color and flavor formation of soy sauce (Liu et al., 2022). Zygosaccharomyces rouxii, a typical salt-tolerant aroma-enhancing yeast, metabolizes and generates alcohols, acids, and ester compounds during the subsequent fermentation stage, further enriching the aromatic profile of soy sauce (Zhang, Zhang, et al., 2025). Thus, proteases and amylases play a key role in the formation of flavor quality during soy sauce fermentation. Based on the above background, this study systematically evaluated the changes in enzyme activity of alkaline protease, acid protease, neutral protease, and glucose amylase under different koji-making times (12 h, 24 h, and 48 h) to determine the optimal koji-making time. The results showed that during the koji-making process, the enzyme activities of different proteases exhibited significant temporal variations (Table 3). Specifically, the enzyme activity of alkaline protease reached its highest value (496.93 U/g) at 24 h, with no significant difference observed at 48 h, indicating that it efficiently hydrolyzes substrate proteins in the early stage of koji-making and rapidly releases available nitrogen sources. In contrast, the enzyme activities of acid protease (434.12 U/g) and neutral protease (511.48 U/g) peaked at 48 h, further promoting the continuous conversion of proteins into free amino acids and small peptides, thus providing sufficient and stable nitrogen support for microbial metabolism during the subsequent brine fermentation stage. The enzyme activity of glucose amylase reached its highest value (283.55 U/g) at 36 h, indicating that the hydrolysis efficiency of starch to glucose was highest at this stage, significantly increasing the available carbon source in the system and providing an energy foundation for the metabolic activities of Aspergillus oryzae and Zygosaccharomyces rouxii. Overall, the enzyme activities of the three proteases remained at relatively high levels at 48 h, suggesting that microbial metabolic activity was vigorous at this stage, facilitating continued enzyme secretion and deep protein hydrolysis. Although the enzyme activity of glucose amylase slightly decreased at 48 h compared to 36 h, it was still significantly higher than at 24 h. Considering the enzyme activity levels, microbial metabolic state, and process operability, this study ultimately determined the optimal koji-making time to be 48 h. This condition not only helps maintain a high and stable enzyme activity level but also promotes the full accumulation of flavor compounds in soy sauce, thereby enhancing the overall quality of the final product.
Table 3.
Changes in enzyme activities during the Koji-making process.
| Koji-making time (h) | Alkaline protease (U/g) | Acid protease (U/g) | Neutral protease (U/g) | Glucoamylase (U/g) |
|---|---|---|---|---|
| 24 | 496.93 ± 10.79 a | 320.51 ± 18.65 b | 347.51 ± 5.57 c | 141.85 ± 4.66 c |
| 36 | 443.01 ± 7.71 b | 351.79 ± 17.99 b | 425.86 ± 16.06 b | 283.55 ± 11.78 a |
| 48 | 471.41 ± 18.27 a | 434.12 ± 6.42 a | 511.48 ± 4.32 a | 191.17 ± 3.55 b |
Different letters attached to the values within the same row indicate the significantly different at p < 0.05 level.
3.5. Analysis of physicochemical properties of SY
Total nitrogen, amino acid nitrogen, reducing sugars, total acidity, and soluble solids content are important physicochemical indicators for evaluating fermentation progress, substrate degradation, and the accumulation of flavor precursors in soy sauce. As shown in Table 4, these indicators generally increased with prolonged fermentation time, indicating that macromolecular substrates, such as proteins and carbohydrates, were continuously degraded through microbial metabolism and enzymatic hydrolysis, accompanied by the gradual release of amino acids, reducing sugars, organic acids, and other low-molecular-weight soluble compounds.
Table 4.
Physicochemical parameters analysis of SY.
| Physicochemical parameters (g/100 ml) | SY-1 | SY-5 | SY-10 | SY-15 |
|---|---|---|---|---|
| Total nitrogen | 0.27 ± 0.01d | 0.28 ± 0.02c | 0.35 ± 0.01b | 1.16 ± 0.03a |
| Amino acid nitrogen | 0.21 ± 0.06b | 0.16 ± 0.02b | 0.21 ± 0.02b | 0.70 ± 0.01a |
| Reducing sugars | 0.66 ± 0.04c | 0.71 ± 0.03c | 0.82 ± 0.05b | 2.13 ± 0.05a |
| Total acidity | 0.04 ± 0.01b | 0.42 ± 0.03b | 0.55 ± 0.03b | 1.22 ± 0.03a |
| Soluble solids content | 7.25 ± 0.11d | 13.65 ± 0.08c | 17.31 ± 0.1b | 34.31 ± 0.21a |
Different letters attached to the values within the same row indicate the significantly different at p < 0.05 level.
Specifically, all physicochemical indicators reached their highest levels on day 15 within the fermentation period investigated in this study. Compared with the initial fermentation stage (SY-1), the contents of total nitrogen, amino acid nitrogen, reducing sugars, total acidity, and soluble solids in SY-15 increased from 0.27, 0.21, 0.66, 0.04, and 7.25 g/100 mL to 1.16, 0.70, 2.13, 1.22, and 34.31 g/100 mL, respectively, corresponding to 4.30, 3.33, 3.23, 30.50, and 4.73 times those of SY-1. According to the statistical analysis in Table 4, all indicators in SY-15 were significantly higher than those in the earlier fermentation samples, suggesting that substrate degradation, soluble component release, and flavor precursor accumulation reached relatively high levels by day 15. Notably, the amino acid nitrogen content in SY-15 reached 0.70 g/100 mL, meeting the Chinese national standard for first-grade soy sauce. This indicates that the accumulation of nitrogenous flavor precursors and the fermentation maturity of rapeseed meal soy sauce had reached a relatively high level at this stage. The simultaneous increases in total nitrogen and amino acid nitrogen reflect the continuous hydrolysis of proteins during fermentation. The increase in total nitrogen indicates the gradual release of proteins and nitrogen-containing components from the raw material, whereas the accumulation of amino acid nitrogen further suggests that proteins were progressively degraded into available nitrogenous compounds, such as free amino acids and small peptides. These low-molecular-weight nitrogenous compounds are not only important indicators of soy sauce fermentation maturity and quality grade, but are also closely associated with umami development and flavor compound formation (Diez-Simon, Eichelsheim, Mumm, & Hall, 2020). Consistent with the present findings, our previous study on walnut soy sauce and milk-enriched walnut soy sauce also showed that total nitrogen and amino acid nitrogen contents reached their highest levels on day 15, further supporting that a 15-day fermentation period can promote protein degradation and the accumulation of nitrogenous flavor precursors in soy sauce fermentation systems.
The increase in reducing sugar content indicates the continuous degradation of carbohydrates and the release of available sugars during fermentation. These sugars can serve as carbon sources for microbial growth and metabolism and also act as important substrates for Maillard reactions, thereby contributing to color development and aroma compound formation in soy sauce. Meanwhile, the increase in total acidity reflects the gradual accumulation of acidic metabolites, such as organic acids, which may help regulate the acidic environment of the fermentation system and influence flavor balance, taste coordination, and fermentation stability (Wei et al., 2013). In addition, the marked increase in soluble solids content further confirms the continuous release and accumulation of proteins, sugars, and other soluble components during fermentation, indicating an enhanced degree of substrate conversion in rapeseed meal soy sauce. This trend is also consistent with our previous findings on walnut soy sauce.
Overall, within the fermentation period investigated in this study, SY-15 exhibited the most favorable physicochemical profile. At this stage, nitrogen release, carbon source conversion, organic acid accumulation, and soluble component release were relatively sufficient, and the amino acid nitrogen content met the Chinese national standard for first-grade soy sauce, providing a solid basis for subsequent flavor formation and quality improvement.
3.6. Analysis of non-VOCs in SY
UPLC-MS/MS-based metabolomics was used to characterize the dynamic changes in non-VOCs during soy sauce fermentation, providing a basis for evaluating the optimal fermentation stage. After peak extraction, alignment, integration, normalization, and database-based annotation, a total of 745 non-VOCs were putatively annotated across four fermentation stages (SY1, SY5, SY10, and SY15). The annotation confidence levels of these metabolites were classified according to the MSI, and the corresponding MSI levels are provided in Table S1. Orthogonal partial least squares discriminant analysis (OPLS-DA) revealed significant separation between soy sauce samples from different fermentation stages, with good intra-group clustering, indicating clear differences in the composition of non-VOCs across fermentation stages (Fig. 2a: positive ion mode; Fig. 2b: negative ion mode). The R2X, R2Y, and Q2 values for the positive ion mode were 0.59, 1.00, and 0.983, respectively (Table S2), and for the negative ion mode, they were 0.566, 1.00, and 0.978 (Table S3), indicating that the model had good fit and high predictive reliability.
Fig. 2.
Analysis of non-VOCs in SY. (a) OPLS-DA analysis in positive ion mode; (b) OPLS-DA analysis in negative ion mode; (c) Cluster heatmap analysis of non-VOCs in SY.
Furthermore, based on the criteria of VIP > 1, p < 0.05, and |log₂FC| ≥ 1, 66 differential non-VOCs were screened from the 745 putatively annotated non-VOCs. These compounds included amino acids and their derivatives (26), heterocyclic compounds (12), sugars and sugar derivatives (6), vitamins (4), aromatic compounds (4), flavonoids (3), carboxylic acids and their derivatives (3), and other compounds (8). Their annotation confidence levels were classified according to the MSI criteria, and the corresponding MSI levels are provided in Table S4. Hierarchical clustering analysis (Fig. 2c) showed that SY1 and SY5 clustered together, while SY10 and SY15 formed another group, further confirming that fermentation time significantly affects the composition of non-VOCs in soy sauce.
To further characterize the compositional changes of differential non-VOCs during SY fermentation, their relative abundances were summarized in Table S4. Amino acids and their derivatives, sugars and sugar derivatives, aromatic compounds, and heterocyclic compounds were the dominant differential non-VOC categories during SY fermentation. Specifically, as fermentation progressed, the relative abundance of amino acids and their derivatives gradually decreased from 57.60% on day 1 to 34.53% on day 15. Among them, the proportions of bitter-related amino acid derivatives, such as Lysopine and Leucylleucine, significantly decreased from 11.85% and 14.85% to 3.56% and 3.40%, respectively. This change may be closely related to the Maillard reaction, which involves the reaction between amino acids and reducing sugars. This not only enhances the aroma of soy sauce but also consumes some amino acids, especially bitter amino acids, leading to a decrease in the overall content of amino acids and bitter amino acids (Gao et al., 2021). In contrast, the proportion of sugars and their derivatives gradually increased during fermentation, with the total content rising from 19.91% to 25.68%. Notably, the increase in sorbitol was particularly significant, reaching its highest value (19.45%) on day 15. The increase in sugar content was primarily attributed to enzymatic hydrolysis during fermentation, where microorganisms (such as Aspergillus oryzae and Zygosaccharomyces rouxii) secrete various enzymes to break down starch and cellulose in the raw materials, releasing more sugars and thus promoting the increase in sugar content (Li et al., 2023). Aromatic compounds continuously accumulated throughout the fermentation process, reaching their highest value (23.57%) on day 15. Among these compounds, the relative abundance of putatively annotated 4-hydroxy-1-(3-pyridyl)-1-butanone increased from 7.36% to 23.01%, indicating the accumulation of pyridine-related non-VOCs during fermentation. Additionally, heterocyclic compounds, which are closely associated with Maillard reaction-related processes, steadily increased and reached 8.76% on day 15, suggesting the accumulation of Maillard reaction-related non-VOCs during fermentation.
In summary, microbial metabolism during fermentation, particularly protein degradation, carbohydrate conversion, and Maillard reaction-related processes, markedly altered the composition of non-VOCs in soy sauce. As fermentation progressed, some bitterness- and sourness-related metabolites decreased, whereas sweetness-, umami-, and Maillard reaction-related compounds gradually accumulated. These changes suggest that the non-VOC profile on day 15 may contribute to improved taste coordination and enhanced flavor precursor accumulation in rapeseed meal soy sauce.
3.7. HS-GC-IMS analysis
In this study, HS-GC-IMS was employed to systematically characterize the changes in VOCs during the fermentation of SY. Based on retention index and IMS drift time matching with the GC-IMS library, a total of 61 VOC-related signal features were annotated, including esters (15), aldehydes (11), ketones (11), alcohols (12), carboxylic acids (6), heterocyclic compounds (4), and sulfur-containing compounds (2). To visualize the dynamic changes in VOCs across fermentation stages, a fingerprint map was constructed (Fig. 3). In this map, each row represents a sample, and each column corresponds to the signal intensity of a specific compound across different samples; brighter colors indicate higher relative signal intensities.
Fig. 3.
Fingerprint analysis of VOCs in SY.
In the early fermentation stage (SY1, region A), high levels of compounds such as 2-butanone, 3-hydroxy (buttery, creamy), acetic acid ethyl ester (fresh, fruity, sweet, grassy), ethyl propanoate (grape, pineapple, fruity, rum), 1-Butanol, 3-methyl- (whiskey, banana, fruity), 1-octen-3-ol (mushroom, lavender, rose, hay), 3-octanone (vegetable, mushroom, cheese, fruity), and 1-propanol (alcoholic, pungent) were detected, contributing fruity, alcoholic, and buttery notes. By day 5 (SY5, region B), major VOCs included 1-butanol (wine), ethanol (aromatic), butanoic acid ethyl ester (pineapple, fruity, ester-like, whiskey), 2-pentanol (fusel oil, green), and 1-hexanol (fresh, fruity, wine-like, sweet, green), resulting in dominant fruity and alcoholic aromas. On day 10 (SY10, region C), the accumulation of acetone (fresh, apple, pear), 3-methylbutanal (chocolate, fatty), 3-penten-2-one (fruity), and (E)-2-heptenal (spicy, green vegetables, fresh, fatty) indicated the development of fruity and fatty sensory attributes. During the late fermentation stage (day 15, SY15, region D), the VOCs profile was dominated by compounds such as propanal (pungent, green, grassy), n-valeraldehyde (green, grassy, faint banana, pungent), butyl formate (plum), 3-methyl-2-butenal (fruity), ethyl 2-hydroxypropanoate (fruity), 2,5-dimethylpyrazine (nutty, peanut, mouldy, earthy, potato, fatty, cocoa powder), 1-(2-furanyl)ethanone (fatty, sweet, caramel-like, nutty, tobacco), and 5-methyl furfural (spicy, caramel, woody). These compounds imparted fruity, nutty, and caramel-like characteristics to the soy sauce. Collectively, these key VOCs shaped the stage-specific aroma profiles of SY, providing essential insights into the mechanisms underlying its aroma development during fermentation.
Table 5 further presents a comprehensive comparison of the relative abundances of VOCs in SY across various fermentation stages, highlighting the dynamic evolution of odor-active compounds during the process. The major classes of VOCs included alcohols, esters, ketones, aldehydes, carboxylic acids, and sulfur-containing compounds. Overall, alcohol and esters were predominantly enriched in the early stages of fermentation. At the same time, ketones accumulated mainly during the mid-stage, and aldehydes and carboxylic acids were significantly elevated in the later stages. Alcohols, primarily derived from sugar and amino acid metabolism under aerobic conditions, are key contributors to floral, fruity, and alcoholic aromas in soy sauce (Feng et al., 2024). From day 1 to day 10, alcohols accounted for more than 40% of total VOCs, but their proportion sharply declined to 21.96% by day 15. This decrease may be attributed to their oxidation into aldehydes or transformation into esters (Zhang et al., 2022). For example, 1-Butanol, 3-methyl-, derived from isoleucine and leucine via the Ehrlich pathway, decreased from 14.99% on day 1 to 6.05% on day 15. This compound is associated with unpleasant notes, and its reduction is likely to improve overall aroma profile (Wang, Meng, & Song, 2022). Similarly,
Table 5.
VOCs identified in SY samples.
| No. | Compounds | Aroma description | Relative abundance (%) |
|||
|---|---|---|---|---|---|---|
| SY1 | SY5 | SY10 | SY15 | |||
| Esters (15) | ||||||
| 1 | 1-Butanol, 3-methyl-, acetate | sweet, banana, fruity | 2.66 ± 0.52 a | 0.39 ± 0.02 b | 0.59 ± 0.02 b | 0.35 ± 0.01 b |
| 2 | Ethyl 3-methyl-2-butenoate | null | 0.04 ± 0.01 b | 0.03 ± 0.001b | 0.04 ± 0.01 b | 0.15 ± 0.02 a |
| 3 | Ethyl 2-hydroxypropanoate | fruity | 0.19 ± 0.02 c | 0.3 ± 0.01 b | 0.2 ± 0.01 c | 0.92 ± 0.03 a |
| 4 | 2-Methylbutyl acetate | fruity | 0.05 ± 0.01 d | 0.1 ± 0.01 c | 0.18 ± 0.01 b | 1.14 ± 0.03 a |
| 5 | Acetic acid ethyl ester | fresh, fruity, sweet, grassy | 21.26 ± 0.12 a | 11.63 ± 0.14 b | 9.26 ± 0.1 c | 4.8 ± 0.16 d |
| 6 | Butanoic acid ethyl ester | pineapple, fruity, ester, whiskey | 0.1 ± 0.01 d | 3.42 ± 0.06 a | 2.62 ± 0.02 b | 1.1 ± 0.05 c |
| 7 | 2-methyl-1-propyl acetate | fruity, raw pear and raspberrie | 0.05 ± 0.01 c | 0.09 ± 0.002 a | 0.06 ± 0.001 b | 0.03 ± 0.001 d |
| 8 | Ethyl 2-methylpropanoate | sweet, fruity, alcoholic, rummy | 0.03 ± 0.01 c | 0.11 ± 0.01 a | 0.11 ± 0.01 a | 0.08 ± 0.01 b |
| 9 | Ethyl propanoate | grape, pineapple, fruity, rum | 0.62 ± 0.16 a | 0.36 ± 0.01 b | 0.41 ± 0.01 b | 0.27 ± 0.02 b |
| 10 | Acetic acid propyl ester | fruity, pear | 1.69 ± 0.23 a | 0.08 ± 0.01 b | 0.11 ± 0.01 b | 0.1 ± 0.001 b |
| 11 | Butyl formate | plum | 0.22 ± 0.01 c | 0.12 ± 0.01 d | 0.3 ± 0.01 b | 1.24 ± 0.02 a |
| 12 | 2-methyl-2(E)-butenyl acetate | null | 0.07 ± 0.01 c | 0.09 ± 0.01 c | 0.6 ± 0.02 b | 0.7 ± 0.03 a |
| 13 | 2-Methylbutanoic acid, methyl ester | apple | 0.11 ± 0.02 d | 0.14 ± 0.01 c | 0.2 ± 0.01 b | 0.3 ± 0.01 a |
| 14 | Butanoic acid 3-methyl, ethyl ester | sour and sweet | 0.03 ± 0.001 d | 0.55 ± 0.03 a | 0.17 ± 0.001 b | 0.13 ± 0.02 c |
| 15 | (Z)-3-Hexenyl formate | green | 0.23 ± 0.02 a | 0.16 ± 0.01 b | 0.13 ± 0.01 c | 0.11 ± 0.01 c |
| Total | 27.35 | 17.57 | 14.98 | 11.42 | ||
| Aldehydes (11) | ||||||
| 16 | Benzaldehyde | bitter almond, cherry, nutty | 1.3 ± 0.06 c | 0.95 ± 0.27 d | 4.47 ± 0.04 a | 3.08 ± 0.13 b |
| 17 | 2-furaldehyde | sweet, woody, almond, bready | 0.17 ± 0.02 c | 0.13 ± 0.01 c | 1.11 ± 0.05 b | 1.76 ± 0.16 a |
| 18 | (E)-2-octenal | fresh cucumber, fatty, green herbal, banana, green leaf | 0.14 ± 0.01 b | 0.27 ± 0.06 a | 0.22 ± 0.01 a | 0.07 ± 0.01 c |
| 19 | phenylacetaldehyde | hyacinth, sweet fruity, almond, cherry, clover honey, cocoa | 0.51 ± 0.02 c | 0.55 ± 0.01 c | 1.4 ± 0.04 b | 2.53 ± 0.1 a |
| 20 | (Z)-2-pentenal | null | 0.31 ± 0.02 b | 0.09 ± 0.01 c | 0.96 ± 0.02 a | 0.12 ± 0.01 c |
| 21 | Propanal | pungent, green grassy | 0.48 ± 0.03 b | 0.28 ± 0.03 d | 0.34 ± 0.01 c | 0.55 ± 0.02 a |
| 22 | 3-methylbutanal | chocolate, fat | 0.23 ± 0.03 b | 0.2 ± 0.02 b | 0.94 ± 0.01 a | 0.05 ± 0.001 c |
| 23 | (E)-2-heptenal | spicy, green vegetables, fresh, fatty | 0.23 ± 0.01 c | 0.77 ± 0.09 b | 1.02 ± 0.01 a | 0.18 ± 0.01 c |
| 24 | 5-methyl furfural | spices, caramel wood | 0.3 ± 0.02 c | 0.33 ± 0.02 bc | 0.35 ± 0.02 b | 0.64 ± 0.03 a |
| 25 | 3-Methyl-2-butenal | fruity | 0.17 ± 0.03 d | 0.26 ± 0.01 c | 0.37 ± 0.02 b | 1.24 ± 0.07 a |
| 26 | n-valeraldehyde | green grassy, faint banana, pungent | 0.13 ± 0.01 b | 0.03 ± 0.01 c | 0.05 ± 0.001 c | 0.26 ± 0.02 a |
| Total | 3.97 | 3.86 | 11.23 | 10.48 | ||
| Ketones (11) | ||||||
| 27 | Cyclopentanone | pleasant | 0.41 ± 0.01 d | 0.5 ± 0.02 c | 0.8 ± 0.03 b | 2.1 ± 0.04 a |
| 28 | Acetone | fresh, apple, pear | 7.12 ± 0.43 d | 10.81 ± 0.11 b | 13.32 ± 0.16 a | 8 ± 0.22 c |
| 29 | 2-Butanone | fruity, camphor | 0.72 ± 0.1 d | 3.34 ± 0.15 c | 5.1 ± 0.02 b | 5.54 ± 0.13 a |
| 30 | 3-Penten-2-one | Fruity, turns into spicy during storage | 0.11 ± 0.01 c | 0.21 ± 0.01 b | 0.29 ± 0.01 a | 0.07 ± 0.001 d |
| 31 | 1-Octen-3-one | strong earthy, mushroom, vegetable, fishy, chicken | 0.2 ± 0.02 ab | 0.21 ± 0.04 a | 0.2 ± 0.001 a | 0.15 ± 0.01 b |
| 32 | Cyclohexanone | strong pungent, earthy | 0.45 ± 0.06 b | 0.42 ± 0.02 b | 0.44 ± 0.001 b | 0.61 ± 0.02 a |
| 33 | 2-methyl-2-hepten-6-one | citrus, fruity, mouldy, ketone | 0.19 ± 0.01 c | 0.38 ± 0.01 a | 0.29 ± 0.02 b | 0.09 ± 0.01 d |
| 34 | 1-(2-furanyl)ethanone | fatty, sweet, caramel, nutty, tobacco | 0.11 ± 0.02 c | 0.12 ± 0.01 c | 0.16 ± 0.01 b | 0.24 ± 0.01 a |
| 35 | 2-Butanone, 3-hydroxy | butter, cream | 5.8 ± 0.96 a | 5.27 ± 0.36 a | 3.74 ± 0.12 b | 1.57 ± 0.1 c |
| 36 | 2-Pentanone | acetone, fresh, sweet fruity, wine | 0.73 ± 0.01 c | 1.3 ± 0.01 b | 1.25 ± 0.01 b | 2.48 ± 0.05 a |
| 37 | 3-Octanone | mouldy, ketone, green, waxy, vegetable, mushroom, cheese, fruity | 1.45 ± 0.01 a | 1.37 ± 0.04 b | 0.4 ± 0.01 d | 0.48 ± 0.01 c |
| Total | 17.29 | 23.93 | 25.99 | 21.33 | ||
| Alcohols (12) | ||||||
| 38 | 1-Butanol, 3-methyl- | whiskey, banana, fruity | 14.99 ± 0.32 a | 14.25 ± 0.12 b | 13.17 ± 0.09 c | 6.05 ± 0.29 d |
| 39 | 1-Hydroxy-2-propanone | pungent, caramel, fresh | 0.32 ± 0.03 c | 0.42 ± 0.02 b | 0.7 ± 0.06 a | 0.78 ± 0.04 a |
| 40 | (E)-3-hexen-1-ol | moss, fresh | 0.18 ± 0.01 c | 0.23 ± 0.02 c | 0.44 ± 0.02b | 1.57 ± 0.07 a |
| 41 | 1-Propanol, 2-methyl- | fresh, alcoholic, leather | 7.3 ± 0.02 b | 8.09 ± 0.09 a | 6.33 ± 0.03 c | 1.78 ± 0.06 d |
| 42 | 1- butanol | wine | 0.75 ± 0.02 d | 5.11 ± 0.04 a | 2.94 ± 0.01 b | 2.64 ± 0.08 c |
| 43 | ethanol | aromaticity | 17.24 ± 0.96 a | 17.76 ± 0.48 a | 15.31 ± 0.14 b | 6.59 ± 0.1 c |
| 44 | 1-Octen-3-ol | mushroom, lavender, rose, hay | 0.75 ± 0.01 a | 0.63 ± 0.01 b | 0.28 ± 0.02 c | 0.14 ± 0.01 d |
| 45 | 1-Penten-3-ol | ethereal, green, tropical fruity | 0.2 ± 0.01 c | 0.31 ± 0.02 b | 0.37 ± 0.01 a | 0.36 ± 0.01 a |
| 46 | 2-Pentanol | Fusel Oil, Green | 0.14 ± 0.01 c | 0.29 ± 0.01 a | 0.15 ± 0.02 b | 0.1 ± 0.01 d |
| 47 | 1-Propanol | alcohol, pungent | 3.23 ± 0.02 a | 1.37 ± 0.04 b | 0.85 ± 0.01 d | 0.93 ± 0.02 c |
| 48 | 2-Hexanol | null | 0.07 ± 0.01 c | 0.08 ± 0.002 c | 0.12 ± 0.01 b | 0.81 ± 0.03 a |
| 49 | 1-Hexanol | fresh, fruity, wine, sweet, green | 0.33 ± 0.05 b | 0.49 ± 0.02 a | 0.16 ± 0.002 c | 0.21 ± 0.01 c |
| Total | 45.5 | 49.03 | 40.82 | 21.96 | ||
| Carboxylic acids(6) | ||||||
| 50 | Acetic acid | spicy | 0.85 ± 0.04 b | 0.83 ± 0.12 b | 0.86 ± 0.11 b | 5.69 ± 0.16 a |
| 51 | Propanoic acid | yogurt, vinegar | 0.45 ± 0.1 b | 0.5 ± 0.09 b | 0.52 ± 0.02 b | 2.03 ± 0.04 a |
| 52 | 1-butanoic acid | strong acetic acid, cheese, butter, fruity | 0.94 ± 0.08 b | 0.94 ± 0.16 b | 0.91 ± 0.06 b | 6.31 ± 0.62 a |
| 53 | 3-Methylbutanoic acid | sour, foot sweat, cheese | 1.02 ± 0.05 b | 0.81 ± 0.06 b | 0.84 ± 0.1 b | 6.84 ± 0.61 a |
| 54 | 2-Methylbutanoic acid | pungent and spicy cheese, fruity | 0.3 ± 0.03 b | 0.27 ± 0.03 b | 0.28 ± 0.02 b | 2.57 ± 0.09 a |
| 55 | 2-Methyl propanoic acid | yogurt, rancid cream | 0.78 ± 0.03 b | 0.67 ± 0.04 b | 0.72 ± 0.06 b | 5.63 ± 0.15 a |
| Total | 4.34 | 4.02 | 4.13 | 29.07 | ||
| Heterocyclic compounds(4) | ||||||
| 56 | 2-Methylpyrazine | nutty, mouldy, roast, earthy | 0.25 ± 0.01 c | 0.24 ± 0.02 c | 0.36 ± 0.01 b | 1.11 ± 0.02 a |
| 57 | 2,5-Dimethylpyrazine | nutty, peanut, mouldy, earthy, potato, fatty, cocoa powder | 0.07 ± 0.001 c | 0.08 ± 0.002 c | 0.14 ± 0.01 b | 0.83 ± 0.02 a |
| 58 | 4,5-Dimethylthiazole | Green, Nut, Roast | 0.04 ± 0.01 c | 0.04 ± 0.01 c | 0.19 ± 0.01 b | 0.38 ± 0.01 a |
| 59 | Thiazole | putrid | 0.61 ± 0.01 d | 0.8 ± 0.02 b | 0.69 ± 0.01 c | 1.05 ± 0.04 a |
| Total | 0.97 | 1.16 | 1.38 | 3.37 | ||
| Sulfur-containing compounds(2) | ||||||
| 60 | allyl disulfide | stinky, garlic | 0.15 ± 0.01 b | 0.11 ± 0.01 c | 0.15 ± 0.02 b | 0.34 ± 0.02 a |
| 61 | 3-(methylsulfanyl)propanal | onion, meat, fruity | 0.43 ± 0.08 c | 0.3 ± 0.02 d | 1.28 ± 0.03 b | 2.05 ± 0.08 a |
| Total | 0.58 | 0.41 | 1.43 | 2.39 | ||
Different letters attached to the values within the same row indicate the significantly different at p < 0.05 level.
1-Propanol, 2-methyl- and ethanol peaked on day 5 at 8.09% and 17.76%, respectively, followed by a rapid decline, suggesting a marked reduction in alcoholic aroma in the later stages. Esters, which contribute fresh, fruity, and floral notes, are synthesized via microbial esterification, alcoholysis, acidolysis, and transesterification (Shangpliang & Tamang, 2023). Their total content showed a declining trend throughout fermentation, with ethyl acetate dropping from 21.26% on day 1 to 4.8% on day 15. Previous studies suggest that ester levels are generally higher in Japanese-style soy sauce than in Chinese-style, potentially due to differences in microbial community composition and fermentation techniques (Ju, Sun, Zhang, Li, & Hou, 2023). Ketones, often generated via microbial carbohydrate metabolism or fatty acid oxidation, impart pleasant aromas (Wang et al., 2025). Representative compounds include acetone and 3-hydroxy-2-butanone. Acetone peaked at 13.32% on day 10, imparting apple and pear notes. In contrast, 3-hydroxy-2-butanone decreased from 5.8% on day 1 to 1.57% on day 15, indicating a gradual decline in buttery aroma over time. Aldehydes, primarily produced through microbial metabolism of amino acids, are potent odorants with low detection thresholds, typically contributing malty, fruity, or cheesy aromas (Liu et al., 2024). Benzaldehyde peaked at 4.47% on day 10, generated via phenylalanine degradation, imparting cherry and nutty notes. Phenylacetaldehyde reached 2.53% on day 15, characterized by floral and honey-like aromas, possibly derived from Strecker degradation of phenylalanine. Carboxylic acids, mainly formed through fatty acid oxidation or further metabolism of aldehydes, were significantly enriched in the late fermentation stage (Liu et al., 2021). Acetic acid (5.69%) and propionic acid (2.03%) were prominent, along with branched-chain fatty acids such as 3-methylbutanoic acid (6.84%) and 2-methylpropanoic acid (5.63%). While acetic acid contributes a distinct sourness, branched-chain acids may introduce cheesy or sweaty notes. Among sulfur-containing compounds, 3-(methylthio) propanal reached 2.05% on day 15. This compound, characterized by its distinctive potato and cheese aromas, is a key signature compound in both rapeseed meal and Japanese-style soy sauces, primarily derived from the thermal degradation of methionine (Al-Dalali et al., 2024). Additionally, pyrazines were gradually enriched in the later stages as products of the Maillard reaction, contributing roasted and nutty aromas to the final soy sauce (Liu et al., 2023).
Generally, VOCs with a ROAV greater than one are considered to have a potential contribution to the overall aroma of a product (Fu et al., 2024). In this study, ROAV analysis was performed to screen potential key odor-active compounds at different fermentation stages of SY. As shown in Table 6, Table 1-octen-3-one showed the highest ROAV (100) on days 1, 5, and 10, and still maintained a relatively high level on day 15 (60.56), suggesting its consistent contribution to the mushroom-like aroma throughout fermentation. By day 15, phenylacetaldehyde became the dominant odor-active compound, with its ROAV increasing to 100, contributing pronounced floral and honey-like notes. During the early fermentation stage (SY1), compounds such as isoamyl acetate (ROAV 40.6), 1-Butanol, 3-methyl- (ROAV 45.51), and phenylacetaldehyde (ROAV 25.86) jointly contributed to the fresh, fruity and floral aroma profile of the soy sauce. As fermentation progressed, the ROAV of 1-Butanol, 3-methyl-, acetate gradually decreased, indicating a weakening of fruity notes, which was consistent with the fingerprint results obtained from HS-GC-IMS analysis. On day 5 (SY5), Butanoic acid 3-methyl, ethyl ester showed a relatively high ROAV (39.1), potentially enhancing the sweet-sour aroma characteristics of the product. By days 10 and 15, the ROAV of 2-furaldehyde increased to 54.89 and 69.83, respectively, suggesting its important contribution to woody, roasted, and nutty notes in the middle and late stages of aroma development in rapeseed meal soy sauce.
Table 6.
The VOCs with ROAVs above one in SY using HS-GC-IMS.
| Compounds | Odor threshold | Relative odor activity value |
|||
|---|---|---|---|---|---|
| SY-1 | SY-5 | SY-10 | SY-15 | ||
| Esters (15) | |||||
| 1-Butanol, 3-methyl-, acetate | 0.0001 | 40.6 | 5.52 | 8.68 | 4.14 |
| Ethyl 3-methyl-2-butenoate | 0.025 | 0 | 0 | 0 | 0.01 |
| Ethyl 2-hydroxypropanoate | 250 | 0 | 0 | 0 | 0 |
| 2-Methylbutyl acetate | 0.011 | 0.01 | 0.01 | 0.02 | 0.12 |
| Acetic acid ethyl ester | 91 | 0 | 0 | 0 | 0 |
| Butanoic acid ethyl ester | 0.002 | 0.08 | 2.41 | 1.94 | 0.65 |
| 2-methyl-1-propyl acetate | 0.001 | 0.07 | 0.12 | 0.09 | 0.04 |
| ethyl 2-methylpropanoate | 0.00002 | 2.36 | 7.61 | 8.31 | 4.56 |
| Ethyl propanoate | 0.01 | 0.09 | 0.05 | 0.06 | 0.03 |
| Acetic acid propyl ester | 0.01 | 0.26 | 0.01 | 0.02 | 0.01 |
| Butyl formate | 0.8 | 0 | 0 | 0 | 0 |
| 2-methyl-2(E)-butenyl acetate | n.f. | – | – | – | – |
| 2-Methylbutanoic acid, methyl ester | 0.0003 | 0.55 | 0.66 | 1.01 | 1.19 |
| Butanoic acid 3-methyl, ethyl ester | 0.00002 | 2.6 | 39.1 | 12.45 | 7.74 |
| (Z)-3-Hexenyl formate | 0.0001 | 3.47 | 2.28 | 1.89 | 1.33 |
| Aldehydes (11) | |||||
| Benzaldehyde | 0.75089 | 0 | 0 | 0.01 | 0 |
| 2-furaldehyde | 0.00003 | 8.67 | 6.33 | 54.89 | 69.83 |
| (E)-2-octenal | 0.003 | 0.07 | 0.13 | 0.11 | 0.03 |
| Phenylacetaldehyde | 0.00003 | 25.86 | 25.89 | 69.12 | 100 |
| (Z)-2-pentenal | 0.98 | 0 | 0 | 0 | 0 |
| Propanal | 0.0151 | 0.05 | 0.03 | 0.03 | 0.04 |
| 3-methylbutanal | 0.0011 | 0.31 | 0.26 | 1.27 | 0.05 |
| (E)-2-heptenal | 0.04 | 0.01 | 0.03 | 0.04 | 0.01 |
| 5-methyl furfural | 1.11 | 0 | 0 | 0 | 0 |
| 3-Methyl-2-butenal | 0.5 | 0 | 0 | 0 | 0 |
| n-valeraldehyde | 0.0002 | 0.97 | 0.23 | 0.34 | 1.56 |
| Ketones (11) | |||||
| Cyclopentanone | 47 | 0 | 0 | 0 | 0 |
| Acetone | 1.5 | 0.01 | 0.01 | 0.01 | 0.01 |
| 2-Butanone | 0.002 | 0.55 | 2.36 | 3.78 | 3.29 |
| 3-Penten-2-one | 1.2 | 0 | 0 | 0 | 0 |
| 1-Octen-3-one | 0.000003 | 100 | 100 | 100 | 60.56 |
| Cyclohexanone | 0.067 | 0.01 | 0.01 | 0.01 | 0.01 |
| 2-methyl-2-hepten-6-one | n.f. | – | – | – | – |
| 1-(2-furanyl)ethanone | n.f. | – | – | – | – |
| 2-Butanone, 3-hydroxy | n.f. | – | – | – | – |
| 2-Pentanone | 1.38 | 0 | 0 | 0 | 0 |
| 3-Octanone | 0.0214 | 0.1 | 0.09 | 0.03 | 0.03 |
| Alcohols (12) | |||||
| 1-Butanol, 3-methyl- | 0.0005 | 45.51 | 40.29 | 38.99 | 14.37 |
| 1-Hydroxy-2-propanone | n.f. | – | – | – | – |
| (E)-3-hexen-1-ol | 0.11 | 0 | 0 | 0.01 | 0.02 |
| 1-Propanol, 2-methyl- | 0.001 | 11.09 | 11.44 | 9.38 | 2.11 |
| 1- butanol | 0.4592 | 0 | 0.02 | 0.01 | 0.01 |
| ethanol | 950 | 0 | 0 | 0 | 0 |
| 1-Octen-3-ol | 0.0015 | 0.76 | 0.6 | 0.28 | 0.11 |
| 1-Penten-3-ol | 0.3581 | 0 | 0 | 0 | 0 |
| 2-Pentanol | 16 | 0 | 0 | 0 | 0 |
| 1-Propanol | 8.5056 | 0 | 0 | 0 | 0 |
| 2-Hexanol | 1.5082 | 0 | 0 | 0 | 0 |
| 1-Hexanol | 0.0056 | 0.09 | 0.12 | 0.04 | 0.04 |
| Carboxylic acids (6) | |||||
| Acetic acid | 99 | 0 | 0 | 0 | 0 |
| Propanoic acid | 2.19 | 0 | 0 | 0 | 0 |
| 1-butanoic acid | 0.0009 | 1.59 | 1.48 | 1.49 | 8.35 |
| 3-Methylbutanoic acid | 0.49 | 0 | 0 | 0 | 0.02 |
| 2-Methylbutanoic acid | 0.54 | 0 | 0 | 0 | 0.01 |
| 2-Methyl propanoic acid | 6.5505 | 0 | 0 | 0 | 0 |
| Heterocyclic compounds (4) | |||||
| 2-Methylpyrazine | 60 | 0 | 0 | 0 | 0 |
| 2,5-Dimethylpyrazine | – | – | – | – | |
| 4,5-Dimethylthiazole | 0.47 | 0 | 0 | 0 | 0 |
| Thiazole | 0.038 | 0.02 | 0.03 | 0.03 | 0.03 |
| Sulfur-containing compounds (2) | |||||
| allyl disulfide | 0.00002 | 11.49 | 8.01 | 10.8 | 20.08 |
| 3-(methylsulfanyl)propanal | 0.001 | 0.65 | 0.43 | 1.9 | 2.44 |
In summary, the composition and relative abundance of VOCs in SY exhibited distinct stage-specific variations throughout the fermentation process. In the early fermentation phase, alcohols and esters were predominant, imparting alcoholic and fruity notes to the product. As fermentation progressed, ketones, aldehydes, carboxylic acids, sulfur-containing compounds, and pyrazines gradually accumulated, significantly enhancing nutty, cheesy, and roasted aromas. ROAV analysis revealed that 1-octen-3-one consistently served as a primary aroma-active compound throughout the entire fermentation period, contributing a characteristic mushroom-like note. Ester compounds played a dominant role in the early stage. At the same time, aldehydes-particularly phenylacetaldehyde-emerged as key contributors in the later stage, further reflecting a shift in aroma profile from fruity to nutty, roasted, and floral characteristics. The consistency between ROAV-based evaluations and relative VOC abundance further supported the stage-dependent accumulation of key aroma compounds and their critical roles in shaping the overall aroma of SY during fermentation.
3.8. Microbial community dynamics during SY fermentation
3.8.1. alpha and beta-diversity analysis
Soy sauce fermentation is a complex microbiological process in which the composition and metabolic activity of microbial communities play a critical role in flavor development and quality control (Feng, Zeng, Lei, & Zhao, 2024). A comprehensive analysis of microbial diversity during fermentation is crucial for a deeper understanding of the microbial contribution to flavor compound formation. In this study, microbial succession was evaluated based on alpha and beta-diversity metrics throughout the fermentation process. For α-diversity analysis, the Chao1, Shannon, and Simpson (1-D) indices were employed. A higher Chao1 index indicates a greater number of species present within the community; a higher Shannon index reflects increased species richness and evenness; and a higher Simpson index suggests greater microbial diversity. As shown in Table 7, all three indices increased progressively during fermentation and peaked on day 10, with values of 22,051.19 (Chao1), 8.9 (Shannon), and 0.99 (Simpson), indicating a highly diverse, evenly distributed, and stable microbial community at this stage, with active microbial growth. By day 15, all indices declined, suggesting substrate depletion, reduced microbial proliferation, and a subsequent decrease in community diversity.
Table 7.
Microbial community α diversity analysis.
| Groups | Chao1 | Shannon | Simpson |
|---|---|---|---|
| SY-1 | 9435.97 ± 299.07 d | 4.33 ± 0.01 d | 0.85 ± 0.02 b |
| SY-5 | 13,466.95 ± 209.6 c | 4.47 ± 0.02 c | 0.86 ± 0.04 b |
| SY-10 | 22,051.19 ± 278.25 a | 8.9 ± 0.12 a | 0.99 ± 0.02 a |
| SY-15 | 18,633.59 ± 391.52 b | 8.26 ± 0.01 b | 0.98 ± 0.01 a |
To further investigate the dynamic changes in microbial community composition at different fermentation stages, β-diversity was assessed using Principal Coordinates Analysis (PCoA) and Non-Metric Multidimensional Scaling (NMDS) based on the Bray-Curtis distance matrix. As shown in Fig. 4a, the PCoA results revealed that samples from day 1 and day 5 clustered closely together, indicating relatively minor differences in microbial community composition during the early stages. In contrast, samples from days 10 and 15 were separated, suggesting that substantial community succession began after day 5. The NMDS results (Fig. 4b) were highly consistent with those of the PCoA, further confirming the reliability of the observed community shifts and the robustness of the analytical approach. These findings indicate that the microbial community structure during rapeseed meal soy sauce fermentation exhibits distinct stage-specific patterns.
Fig. 4.
Microbial community analysis during the fermentation process of SY. (a) PCoA analysis; (b) NMDS analysis; (c) Dominant microorganisms at the phylum level; (d) Dominant microorganisms at the genus level; (e) Dominant microorganisms at the species level.
3.8.2. Changes in bacterial and fungal community structures during SY fermentation
The dynamic succession of microbial communities during SY fermentation was systematically analyzed at the phylum, genus, and species levels. At the phylum level, the dominant microbial groups included Ascomycota, Proteobacteria, Chloroflexi, and Actinobacteria (Fig. 4c). In the early fermentation stages (days 1–5), Ascomycota was the overwhelmingly dominant phylum, with a relative abundance exceeding 75%. Members of Ascomycota are known to secrete a variety of hydrolytic enzymes that synergistically degrade substrate proteins and starches, generating small peptides, free amino acids, and reducing sugars (Li et al., 2023). These compounds serve as essential precursors for the synthesis of flavor compounds, thereby contributing to the rich and well-balanced aroma and taste of soy sauce (Zhang et al., 2024; Zhang et al., 2024). As fermentation progressed into the later stages (days 10–15), the relative abundance of Ascomycota decreased sharply to below 10%, likely due to substrate depletion and reduced oxygen availability. Concurrently, Proteobacteria, Chloroflexi, and Actinobacteria proliferated rapidly and became the dominant microbial groups. Proteobacteria exhibit strong adaptability and proteolytic capacity, promoting the accumulation of flavor compounds. Chloroflexi can maintain metabolic activity under low-oxygen conditions, ensuring the continuation of fermentation (Feng, Xie, et al., 2024). In addition to contributing to flavor development, Actinobacteria also possess antimicrobial properties, which help stabilize the fermentation system.
At the genus level, the dominant microbial genera during the fermentation of rapeseed meal soy sauce included Aspergillus, Weissella, Bellilinea, Aquabacterium, Zygosaccharomyces, and Corynebacterium (Fig. 4d). Among them, Aspergillus is a key microorganism in soy sauce fermentation, capable of efficiently secreting proteases that promote protein hydrolysis and increase the amino nitrogen content, thus serving as a primary driver of foundational flavor development. However, its relative abundance declined progressively during fermentation, dropping to approximately 6% by day 10, likely due to substrate depletion and a decrease in pH, which may have limited its growth. Weissella, an important lactic acid bacterium, can metabolize branched-chain amino acids into alcohols, acids, and esters, thereby significantly enhancing the aroma profile of soy sauce. Its metabolic pathway involves the conversion of branched-chain amino acids into α-keto acids, followed by aldehyde intermediates that are further transformed into branched-chain alcohols and organic acids, which subsequently form esters (Zhao, Xu, et al., 2025; Zhao, Zhang, et al., 2025). Bellilinea, a filamentous anaerobe, is capable of producing acetic acid via carbohydrate fermentation, contributing to the acidic dimension of flavor. Its relative abundance increased significantly by day 10. Aquabacterium, primarily involved in amino acid metabolism, also reached high abundance at day 10, indicating its essential role in flavor formation during mid-stage fermentation. Under high-salt, semi-solid-state fermentation conditions, Zygosaccharomyces showed significant positive correlations (P < 0.05) with key aroma compounds such as phenylethanol, 3-methyl-1-butanol, 2-methyl-1-butanol, benzaldehyde, and ethyl acetate, suggesting its critical role in aroma development. In addition, Corynebacterium was associated with protein hydrolysis and amino acid conversion, and its relative abundance peaked on day 15, indicating sustained metabolic activity in the later stage of fermentation.
At the species level, the relative abundance of most microbial species declined progressively throughout fermentation (Fig. 4e). Weissella cibaria was dominant on day 1, accounting for approximately 30% of the total community, but its abundance gradually decreased thereafter. In contrast, the relative abundance of Zygosaccharomyces rouxii increased significantly from day 1 to day 5, followed by a rapid decline from day 10 to day 15, which may be attributed to the continuous decrease in system pH, inhibiting the growth and metabolic activity of Zygosaccharomyces rouxii. Previous studies have also identified Zygosaccharomyces rouxii as a key functional microorganism in fermented chili paste, responsible for producing ethanol, linalool, α-terpineol, and ethyl palmitate, among other flavor-active compounds (Wang et al., 2024; Wang et al., 2024).
In summary, the microbial community structure during rapeseed meal soy sauce fermentation exhibited distinct stage-specific succession. As fermentation progressed, the microbial community shifted from being dominated by proteolytic fungi to one increasingly enriched with salt-tolerant and low-oxygen-adapted bacteria and yeasts, which play crucial roles in flavor development. At the phylum and genus levels, functional microbial groups such as Proteobacteria, Chloroflexi, and Actinobacteria became predominant in the middle and late fermentation stages, primarily contributing to the further degradation of proteins and amino acids and the transformation of flavor precursors. At the species level, microbial diversity gradually declined over time. Core functional species, including Weissella cibaria and Zygosaccharomyces rouxii, were highly abundant during the early to mid-fermentation stages but declined significantly in the later stage due to substrate depletion and environmental shifts. Overall, the temporal succession of the microbial community was closely linked to changes in fermentation conditions and the accumulation of flavor compounds, underscoring its critical regulatory role in determining the final quality of soy sauce.
3.8.3. LEfSe analysis
LEfSe analysis (LDA score > 4.0) was employed to identify significantly different dominant microbial taxa across fermentation stages, revealing the temporal succession patterns of the microbial community during soy sauce fermentation (Fig. 5a). On day 1 (SY-1), dominant taxa included Aspergillaceae (family), Eurotiales (order), Eurotiomycetes (class), Aspergillus (genus), Aspergillus oryzae (species), and lactic acid bacteria such as Lactobacillaceae (family), Lactobacillales (order), and Weissella (genus) (Fig. 5b). Among these, Aspergillus oryzae serves as a core functional microorganism during the koji-making stage. It efficiently secretes proteases and amylases, promoting the hydrolysis of proteins and starches into peptides, amino acids, and reducing sugars, which serve as crucial precursors for subsequent flavor development (Hu, Chen, Du, & Fang, 2023). Lactic acid bacteria contribute to early-stage aroma by producing organic acids through lactic fermentation. By day 5 (SY-5), the dominant taxa shifted to Ascomycota (phylum), Eukaryota (domain), Saccharomycetales (order), Saccharomycetes (class), Saccharomycetaceae (family), and Zygosaccharomyces rouxii (genus and species). Zygosaccharomyces rouxii, a halotolerant yeast, is capable of metabolizing amino acids via the Ehrlich pathway to generate key aroma compounds including esters, alcohols, and aldehydes, thereby imparting fruity, floral, and nutty notes to soy sauce (Zhang, Xiao, et al., 2024; Zhang, Xiong, et al., 2024). By day 10 (SY-10), the community structure transitioned from being fungal-dominated to bacterial-dominated. The major taxa included Proteobacteria, Chloroflexi, and Ignavibacteriae (phyla), Alphaproteobacteria, Betaproteobacteria, and Anaerolineae (classes), as well as Burkholderiales (order) and Anaerolineaceae (family). These bacteria play essential roles in degrading complex substrates and contributing to the production of organic acids and esters. Alphaproteobacteria may regulate key metabolic pathways to further enhance the flavor complexity of soy sauce. On day 15 (SY-15), the microbial community structure underwent further succession, with Verrucomicrobia and Acidobacteria emerging as dominant phyla. Verrucomicrobia possess polysaccharide-degrading capabilities, which may help reduce system viscosity and improve the textural properties of the soy sauce. Additionally, they facilitate the release of flavor-active compounds, thereby enhancing the final product's flavor layering and sensory harmony.
Fig. 5.
LEfSe analysis of dominant microbiota in SY. (a) Taxonomic cladogram from LEfSe analysis; (b) Species with significant differences with LDA score (LDA score > 4.0).
In summary, the microbial community structure during soy sauce fermentation followed a well-defined stage-specific succession pattern. In the early stage, fungi and lactic acid bacteria dominated, initiating substrate hydrolysis and establishing foundational flavors. In the mid-fermentation phase, Zygosaccharomyces rouxii became predominant, markedly contributing to the synthesis of aromatic compounds. In the late stage, a diverse array of bacterial taxa took over, promoting the accumulation of acids and esters through complex substrate degradation and metabolic regulation. This sequential microbial succession and synergistic metabolism supported the progressive evolution of soy sauce flavor from simple to complex and from basic to layered.
3.8.4. Correlation analysis between dominant microbial taxa, physicochemical parameters, and VOCs
To systematically elucidate the associations among dominant microbial taxa (LDA score > 4.5), physicochemical parameters, and key VOCs (ROAV >1), a Mantel test based on Pearson correlation coefficients was performed (Fig. 6). The results showed that the eight dominant microbial taxa were significantly correlated with most physicochemical parameters and VOCs, indicating that microbial community succession may be closely associated with the formation of physicochemical and flavor quality during rapeseed meal soy sauce fermentation.
Fig. 6.
Correlation analysis between dominant microbial taxa, physicochemical parameters, and VOCs.
In terms of physicochemical parameters, Weissella, especially Weissella cibaria and Weissella confusa, showed strong positive correlations with soluble solids and total acid (r = 0.9–1, p < 0.01). Previous studies have shown that Weissella is an important group of lactic acid bacteria in soy sauce and related high-salt fermentation systems. Members of this genus can produce organic acids through carbohydrate metabolism and are closely involved in the modulation of the fermentation environment and the accumulation of flavor precursors (Kuang et al., 2022). Therefore, the significant correlations between Weissella and both total acid and soluble solids in this study suggest that this genus may play an important role in shaping the acidic taste foundation and promoting the accumulation of soluble metabolites in rapeseed meal soy sauce. In addition, amino acid nitrogen was significantly and positively correlated with Zygosaccharomyces rouxii, Zygosaccharomyces, Weissella, W. cibaria, and W. confusa (p < 0.01). Amino acid nitrogen is an important indicator of protein degradation and umami formation, suggesting that these microorganisms may be closely associated with the accumulation of nitrogen-containing flavor precursors (Devanthi & Gkatzionis, 2019). It should be noted that the initial hydrolysis of proteins mainly depends on proteases and peptidases secreted by microorganisms such as Aspergillus, whereas Zygosaccharomyces and Weissella may contribute more to amino acid transformation, organic acid production, and modulation of the fermentation environment, thereby promoting the formation of the taste profile of rapeseed meal soy sauce. Except for Bellilinea, total nitrogen was significantly and positively correlated with the other dominant microbial taxa, further indicating that multiple microorganisms may jointly participate in nitrogen metabolism and the accumulation of nitrogen-containing flavor precursors during rapeseed meal soy sauce fermentation.
Regarding VOCs, esters, alcohols, and aldehydes showed the closest associations with the dominant microorganisms. Among them, ester compounds, such as 1-Butanol, 3-methyl-, acetate, ethyl 2-methylpropanoate, and Butanoic acid 3-methyl, ethyl ester, were significantly correlated with Z. rouxii (p < 0.01), suggesting that Z. rouxii may be an important microorganism involved in the formation of fruity esters. As a typical salt-tolerant yeast in soy sauce fermentation, Z. rouxii can utilize carbohydrates and amino acid-derived metabolites to produce VOCs, including alcohols, aldehydes, and esters. In particular, via the Ehrlich pathway, branched-chain and aromatic amino acids, such as leucine, valine, isoleucine, and phenylalanine, can be converted into the corresponding higher alcohols through transamination, decarboxylation, and reduction reactions (Zhang, Xiao, et al., 2024; Zhang, Xiong, et al., 2024). These higher alcohols can further undergo esterification with organic acids to form esters with fruity aroma characteristics. Therefore, the significant correlations between Z. rouxii and ester compounds such as 1-Butanol, 3-methyl-, acetate, ethyl 2-methylpropanoate, and Butanoic acid 3-methyl, ethyl ester may reflect its potential role in fruity aroma formation through amino acid degradation, higher alcohol production, and esterification reactions. Alcohol compounds, including 1-Butanol, 3-methyl- and 1-Propanol, 2-methyl-, were significantly and positively correlated with Z. rouxii, W. cibaria, and W. confusa (p < 0.01). Among them, 1-Butanol, 3-methyl- is generally produced from leucine via the Ehrlich pathway and is associated with alcoholic, malty, and fruity notes, whereas 1-Propanol, 2-methyl- is closely related to valine metabolism and can contribute alcoholic aroma characteristics. These results further indicate that amino acid degradation may provide important precursors for the formation of VOCs in rapeseed meal soy sauce. It should be noted that the Ehrlich pathway is mainly applicable to yeast-mediated formation of higher alcohols and related esters, whereas Weissella may indirectly affect the accumulation of alcohols and esters mainly through carbohydrate metabolism, organic acid production, and modulation of the fermentation environment. In addition, aldehyde compounds such as phenylacetaldehyde and 2-furaldehyde were significantly correlated with all eight dominant microbial taxa (p < 0.01), suggesting that their formation may be jointly influenced by microbial synergistic metabolism and non-enzymatic reactions. Phenylacetaldehyde is mainly associated with phenylalanine degradation and can contribute sweet and floral notes, whereas 2-furaldehyde is generally derived from sugar degradation and Maillard reactions and can impart caramel-like, almond-like, and roasted aromas to soy sauce. These compounds may jointly contribute to the formation of complex aroma attributes, including sweet, floral, and roasted notes, thereby enhancing the overall flavor complexity of rapeseed meal soy sauce.
In summary, Weissella and Z. rouxii appeared to be key microbial taxa closely associated with physicochemical quality and flavor formation during rapeseed meal soy sauce fermentation. Weissella may be mainly involved in carbohydrate metabolism, organic acid production, soluble metabolite accumulation, and the establishment of the taste profile, whereas Z. rouxii may promote the formation of fruity esters through amino acid transformation, the Ehrlich pathway, higher alcohol production, and esterification reactions. Together with the LEfSe analysis, these results suggest that Weissella and Z. rouxii may play important roles in the formation of the taste foundation during the early fermentation stage and the accumulation of aroma compounds during the middle-to-late fermentation stages, respectively. Their stage-specific succession and functional complementarity may jointly drive the flavor profile of rapeseed meal soy sauce from an initial fermentation-derived flavor toward a more complex and layered flavor profile. It should be noted that correlation analysis cannot directly prove causal relationships, but it provides important clues for elucidating the potential mechanisms by which key microorganisms participate in flavor formation in rapeseed meal soy sauce.
4. Discussion
Soy sauce is one of the most representative fermented condiments in East and Southeast Asian diets. Its characteristic flavor is mainly derived from the coordinated integration of umami and salty tastes with caramel-like aroma, which plays an important role in enhancing the flavor complexity and sensory acceptance of foods (Wang, Shi, et al., 2024; Wang, Wang, et al., 2024). Traditional soy sauce production generally involves two stages: koji fermentation and moromi fermentation. Koji fermentation is a solid-state fermentation process, during which raw materials are typically incubated for approximately 3 d after inoculation with Aspergillus (Zhang et al., 2023). During this stage, proteins and carbohydrates in the raw materials are gradually hydrolyzed by microbial enzymes into low-molecular-weight compounds, including peptides, free amino acids, and sugars. Subsequently, the koji is mixed with concentrated brine and transferred to the moromi fermentation stage. During this stage, halotolerant lactic acid bacteria, yeasts, and other microorganisms further utilize the peptides, amino acids, and sugars generated during koji fermentation for growth and metabolism, thereby contributing to the formation of organic acids, alcohols, esters, and other VOCs. Because this process involves multiple biochemical transformations, including continuous enzymatic hydrolysis, microbial community succession, Maillard reactions, Strecker degradation, lipid oxidation, and esterification reactions, traditional fermented soy sauce usually requires a long fermentation and maturation period, ranging from several weeks to several years, to develop a complex and stable flavor profile.
The 15-day fermentation period used in this study was not intended to fully simulate the long-term fermentation and maturation process of traditional soy sauce. Instead, it was designed as a short-cycle fermentation evaluation model to investigate the substrate degradation capacity, degree of protein hydrolysis, and preliminary accumulation of flavor-related precursors in HTSP-treated rapeseed meal within a relatively short fermentation period. After HTSP treatment, the chemical composition of rapeseed meal changed markedly, with increases in protein, soluble protein, reducing sugar, and moisture contents, particularly in soluble protein and reducing sugar. The increases in soluble protein and reducing sugar could provide more readily available nitrogen and carbon sources for microbial growth and metabolism, respectively, while the higher moisture content may improve substrate hydration and mass transfer efficiency, thereby creating more favorable conditions for microbial proliferation, enzymatic hydrolysis, and the generation of flavor precursors. Therefore, HTSP treatment may improve the substrate availability of rapeseed meal, thereby promoting protein hydrolysis, the release of nitrogen-containing flavor precursors, and the accumulation of soluble metabolites during short-cycle fermentation. These effects may support the development of a shortened fermentation process for rapeseed meal-based fermented condiments.
In recent years, similar short-cycle fermentation strategies have also been applied to other protein-based soy sauce or fermented condiment systems to rapidly evaluate raw material conversion efficiency and early-stage flavor formation potential. For example, in studies on walnut meal soy sauce and milk-enriched walnut meal soy sauce, a 15-day fermentation period was also adopted, and total nitrogen, amino acid nitrogen, total acid, and total soluble solids reached relatively high levels on day 15 (Battur et al., 2025). These results indicated that substantial substrate hydrolysis and soluble metabolite accumulation had occurred at this stage. Further analysis of VOCs showed that both walnut meal soy sauce and milk-enriched walnut meal soy sauce exhibited a transition from fruity and sweet notes in the early stage to more complex aroma characteristics, including creamy, mushroom-like, and vegetable-like notes, in the later stage of short-cycle fermentation. This suggests that a clear flavor evolution pattern could be observed within a 15-day fermentation period. Notably, milk-enriched walnut meal soy sauce showed more abundant accumulation of key aroma compounds in the later stage and more pronounced complex aroma characteristics, indicating that optimization of raw material composition may contribute to improving flavor complexity.
In this study, SY also exhibited favorable physicochemical quality on day 15 of fermentation, with total nitrogen, amino acid nitrogen, reducing sugar, total acid, and soluble solids reaching their highest levels. In particular, the amino acid nitrogen content was 0.70 g/100 mL. Using the Chinese national standard for first-grade brewed soy sauce as a reference, this value satisfies the corresponding requirement for amino acid nitrogen, suggesting that the 15-day fermentation period could effectively promote protein degradation, the release of soluble nitrogen-containing compounds, and the accumulation of key physicochemical components in this short-cycle fermentation system. Meanwhile, the simultaneous increases in total nitrogen and amino acid nitrogen further suggest that HTSP pretreatment may have improved the availability of rapeseed meal protein during fermentation, thereby promoting protein hydrolysis and the formation of amino acid-derived flavor precursors.
The dynamic changes in non-VOCs further supported this observation. During fermentation, some bitterness-related amino acids and their derivatives gradually decreased, whereas flavor-related metabolites, including sugars, aromatic compounds, and heterocyclic compounds, continuously accumulated and reached relatively high levels on day 15. These results suggest that short-cycle fermentation could promote the degradation and transformation of proteins and carbohydrates, facilitate the accumulation of flavor precursors, and potentially improve the taste quality of SY by reducing bitterness and enhancing flavor coordination. The analysis of VOCs also showed that the aroma composition of SY changed markedly during the 15-day fermentation period. In the early stage of fermentation, the samples mainly exhibited initial fermentation-related aroma characteristics, such as fruity and alcoholic notes. As fermentation progressed, volatile compounds associated with floral, nutty, roasted, and caramel-like aromas gradually increased, indicating that the aroma profile shifted from a relatively simple initial fermented flavor toward a more complex and layered flavor profile. ROAV analysis further showed that several key aroma-active compounds made relatively high contributions on day 15, suggesting that this short-cycle fermentation process was able to induce the formation of key aroma compounds and preliminarily enhance the aroma complexity of rapeseed meal soy sauce. This change may be related to the combined effects of protein hydrolysis, carbohydrate transformation, and microbial metabolism in promoting the accumulation of flavor precursors. Therefore, day 15 was selected as the key evaluation point for this short-cycle fermentation system. It should be emphasized that the selection of this time point does not imply that the flavor of rapeseed meal soy sauce had fully reached the maturity level of traditional long-term fermented soy sauce. Rather, it was based on the integrated changes in physicochemical parameters, non-VOCs, and VOCs observed under the short-cycle fermentation evaluation model used in this study. Extending the fermentation period may further affect metabolite composition and flavor balance, but its specific effects still require validation through subsequent long-term fermentation experiments. In the context of this study, the 15-day fermentation period effectively reflected the substrate conversion efficiency and preliminary flavor formation potential of HTSP-treated rapeseed meal in a short-cycle fermentation system.
From an industrial application perspective, the 15-day short-cycle fermentation system has potential advantages, including a shorter production cycle, higher processing efficiency, and potentially lower production cost, which may contribute to improving the utilization efficiency of oilseed cake and meal by-products in fermented condiments. However, before practical industrial application, the scale-up stability, batch-to-batch consistency, microbial safety, sensory quality, and storage stability of this process should be further evaluated. Therefore, the HTSP-assisted short-cycle fermentation process established in this study may serve as a promising rapid fermentation strategy, although further pilot-scale validation and process optimization are still required.
Finally, it should be noted that the interpretation of flavor complexity in this study was mainly based on the analysis of non-VOCs and VOCs. Owing to the lack of systematic sensory evaluation, consumer acceptance testing, and quantitative validation of key flavor compounds, these results should be interpreted as preliminary flavor formation characteristics under the short-cycle fermentation system and should not be regarded as equivalent to the flavor maturation process of traditional long-term fermented soy sauce. Therefore, future studies should combine extended fermentation experiments, sensory evaluation, consumer acceptance analysis, and quantitative determination of key flavor compounds to further verify the actual flavor quality and market application potential of SY.
5. Conclusion
This study demonstrated that HTSP effectively improved the suitability of rapeseed meal as a fermentation substrate. HTSP reduced the contents of safety- and flavor-limiting factors in rapeseed meal, including glucosinolates, isothiocyanates, sinapine, tannins, and total phenolics, while increasing the levels of protein, soluble protein, and reducing sugars. These changes provided more favorable substrate conditions for subsequent microbial growth, protein hydrolysis, and flavor precursor accumulation. Furthermore, SY prepared from HTSP-treated rapeseed meal showed promising fermentation performance and flavor development potential. The koji showed good growth after 48 h of koji fermentation. When moromi fermentation proceeded to day 15, total nitrogen, amino acid nitrogen, reducing sugars, total acid, and soluble solids all reached relatively high levels, indicating substantial protein degradation, release of soluble nitrogen-containing compounds, and metabolite accumulation in the fermentation system at this stage. Flavor analysis further showed that, as fermentation progressed, the taste profile of rapeseed meal soy sauce shifted from bitterness and sourness toward sweetness and umami, while the aroma profile gradually evolved from fruity notes to more complex aroma characteristics, including nutty, roasted, and floral notes. Correlation analysis indicated that Weissella and Zygosaccharomyces rouxii were important microbial taxa closely associated with quality formation and flavor evolution in rapeseed meal soy sauce, potentially contributing to taste compound accumulation during the early fermentation stage and aroma compound formation during the middle and late fermentation stages, respectively. Overall, HTSP enhanced the fermentative utilization potential of rapeseed meal by reducing limiting factors and improving substrate availability, thereby supporting its high-value application in fermented condiments.
CRediT authorship contribution statement
Feng Zhang: Writing – original draft. Xiaolin Huang: Software, Investigation. Muhammad Aaqil: Methodology. Renwang Huang: Data curation. Jiawen Yao: Software. Cunchao Zhao: Writing – review & editing, Funding acquisition.
Funding
Canola Oil Product Development Project (2024533517002757).
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.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.fochx.2026.104046.
Appendix A. Supplementary data
Supplementary Fig 8: GC–IMS chromatographic profile of SY-1.
Supplementary Fig 9: GC–IMS chromatographic profile of SY-5.
Supplementary Fig 10: GC–IMS chromatographic profile of SY-10.
Supplementary Fig 11: GC–IMS chromatographic profile of SY-15.
Data availability
The data that has been used is confidential.
References
- Al-Dalali S., He Z., Du M., Sun H., Zhao D., Li C.…Xu B. Influence of frozen storage and flavoring substances on the nonvolatile metabolite profile of raw beef: Correlation of lipids and lipid-like molecules with flavor profiles. Food Chemistry: X. 2024;24 doi: 10.1016/j.fochx.2024.101898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Battur M., Aaqil M., Zheng J., Lin H.X., Chuluunotgon B., Zorigtbaatar T.…Tian Y. Exploring the effects of milk-enriched walnut soy sauce: Insights from GC-IMS and metagenomics approach to flavor and microbial shifts. Food Chemistry: X. 2025;27 doi: 10.1016/j.fochx.2025.102364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Betchem G., Dabbour M., Akter Tuly J., Flavorta Billong L., Ma H. Experimental investigation into the implications of low-intensity magnetic field treatment on the structural and functional properties of rapeseed meal during biofermentation. Food Chemistry. 2024;446 doi: 10.1016/j.foodchem.2024.138858. [DOI] [PubMed] [Google Scholar]
- Chadni M., Boussetta N., Guerin C., Lagalle F., Zoghlami A., Perré P.…Ioannou I. Improvement of Sinapine extraction from mustard seed meal by application of emerging technologies. Foods. 2023;12(3):520. doi: 10.3390/foods12030520. https://www.mdpi.com/2304-8158/12/3/520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Devanthi P.V.P., Gkatzionis K. Soy sauce fermentation: Microorganisms, aroma formation, and process modification. Food Research International. 2019;120:364–374. doi: 10.1016/j.foodres.2019.03.010. [DOI] [PubMed] [Google Scholar]
- Dhakal N., Acharya B. Extraction techniques for the development of protein-enriched extracts from canola meal. Industrial Crops and Products. 2025;235 doi: 10.1016/j.indcrop.2025.121772. [DOI] [Google Scholar]
- Diez-Simon C., Eichelsheim C., Mumm R., Hall R.D. Chemical and sensory characteristics of soy sauce: A review. Journal of Agricultural and Food Chemistry. 2020;68(42):11612–11630. doi: 10.1021/acs.jafc.0c04274. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Du Z., Sang N., Yu B., Zheng P., Wang H., Wang Q., Chen D. Obtaining higher yield and quality rapeseed protein: Complex enzymatic hydrolysis assisted alkaline water extraction of cold pressed-extracted rapeseed meal. Food Chemistry: X. 2025;29 doi: 10.1016/j.fochx.2025.102837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- EFSA Panel on Nutrition, Novel Foods and Food Allergens, Turck D., Castenmiller J., De Henauw S., Hirsch-Ernst K.I.…Knutsen H.K. Safety of rapeseed powder from Brassica rapa L. and Brassica napus L. as a novel food pursuant to Regulation (EU) 2015/2283. EFSA Journal. 2020;18(7) doi: 10.2903/j.efsa.2020.6197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Feng Y., Xie Z., Huang M., Tong X., Hou S., Tin H., Zhao M. Decoding temperature-driven microbial community changes and flavor regulation mechanism during winter fermentation of soy sauce. Food Research International. 2024;177 doi: 10.1016/j.foodres.2023.113756. [DOI] [PubMed] [Google Scholar]
- Feng Y., Zeng J., Lei H., Zhao M. Effect of fermentation containers on the taste characteristics and microbiota succession of soy sauce. Food Chemistry. 2024;448 doi: 10.1016/j.foodchem.2024.139066. [DOI] [PubMed] [Google Scholar]
- Fu B., Zheng M., Yang H., Zhang J., Li Y., Wang G.…Yu E. The effect of broad bean diet on structure, flavor and taste of fresh grass carp: A comprehensive study using E-nose, E-tongue, TPA, HS-SPME-GC-MS and LC-MS. Food Chemistry. 2024;436 doi: 10.1016/j.foodchem.2023.137690. [DOI] [PubMed] [Google Scholar]
- Gao X., Feng T., Sheng M., Wang B., Wang Z., Shan P.…Ma H. Characterization of the aroma-active compounds in black soybean sauce, a distinctive soy sauce. Food Chemistry. 2021;364 doi: 10.1016/j.foodchem.2021.130334. [DOI] [PubMed] [Google Scholar]
- Hong K., Zhang H., Han M., Nie X., Fu X., Lei F., He D. A novel four-species microbial consortium for nutritional value improvement of rapeseed meal. Food Chemistry. 2025;478 doi: 10.1016/j.foodchem.2025.143712. [DOI] [PubMed] [Google Scholar]
- Hu G., Chen J., Du G., Fang F. Moromi mash dysbiosis trigged by salt reduction is relevant to quality and aroma changes of soy sauce. Food Chemistry. 2023;406 doi: 10.1016/j.foodchem.2022.135064. [DOI] [PubMed] [Google Scholar]
- Huang X., Lin H., Wang Z., Zhao M., Feng Y. Optimization of derivatization-gas chromatography/mass spectrometry (Der-GC/MS) for analyzing non-volatile metabolites in soy sauce koji-making process and their evolution patterns. Food Chemistry. 2025 doi: 10.1016/j.foodchem.2025.145152. [DOI] [PubMed] [Google Scholar]
- Jiang X., Zhang W., Li L., Xiao Z., Tang J., Wu C.…Zhou S. Characteristics of microbial community, taste, aroma of high-salt liquid-state secondary fortified fermented soy sauce. LWT. 2023;182 doi: 10.1016/j.lwt.2023.114792. [DOI] [Google Scholar]
- Jin H., Zhang X., Li K., Niu Y., Guo M., Hu C.…Huang F. Direct bio-utilization of untreated rapeseed meal for effective Iturin a production by Bacillus subtilis in submerged fermentation. PLoS ONE. 2014;9(10) doi: 10.1371/journal.pone.0111171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ju Y., Sun L., Zhang X., Li W., Hou L. Fractionation, identification and umami characteristics of flavor peptides in natural brewed soy sauce. Food Chemistry. 2023;425 doi: 10.1016/j.foodchem.2023.136501. [DOI] [PubMed] [Google Scholar]
- Kuang X., Su H., Li W., Lin L., Lin W., Luo L. Effects of microbial community structure and its co-occurrence on the dynamic changes of physicochemical properties and free amino acids in the Cantonese soy sauce fermentation process. Food Research International. 2022;156 doi: 10.1016/j.foodres.2022.111347. [DOI] [PubMed] [Google Scholar]
- Li J., Sun C., Shen Z., Tian Y., Mo F., Wang B.…Wang C. Untargeted metabolomic profiling of aspergillus sojae 3.495 and aspergillus oryzae 3.042 fermented soy sauce koji and effect on moromi fermentation flavor. LWT. 2023;184 doi: 10.1016/j.lwt.2023.115027. [DOI] [Google Scholar]
- Liu M., Deng N., Hou X., Zhang B., Li H., Wang J. Characterisation of flavour profiles and microbial communities of fermented peppers with different fermentation years by combining flavouromics and metagenomics. Food Chemistry. 2024;443 doi: 10.1016/j.foodchem.2024.138550. [DOI] [PubMed] [Google Scholar]
- Liu T., Ding K., Zhou X., Pan Z.-H., Zhao G., Yao Y. Steam explosion pretreatment of soy sauce residue for improving the soybean paste flavor. LWT. 2021;149 doi: 10.1016/j.lwt.2021.111914. [DOI] [Google Scholar]
- Liu X., Jacquet N., Xie J., Jiang X., Blecker C. Response surface methodology optimization of alkaline extraction of polysaccharides from rapeseed meal: Structural characterization and antioxidant activities. LWT. 2025;232 doi: 10.1016/j.lwt.2025.118431. [DOI] [Google Scholar]
- Liu Z., Xiao T., Wang J., Fu B., Li W., Hu Y.…Xu N. Analysis of the contribution of koji-making with Z. Rouxii on volatile compounds of soy sauce. LWT. 2023;183 doi: 10.1016/j.lwt.2023.114903. [DOI] [Google Scholar]
- Liu Z., Zhang X., Duan X., Kang B., Liu J., Fu C.…Xu N. Effect of fermentation conditions on the formation of ammonium salt in soy sauce. LWT. 2022;153 doi: 10.1016/j.lwt.2021.112492. [DOI] [Google Scholar]
- Lu F., Alenyorege E.A., Ouyang N., Zhou A., Ma H. Simulated natural and high temperature solid-state fermentation of soybean meal: A comparative study regarding microorganisms, functional properties and structural characteristics. LWT. 2022;159 doi: 10.1016/j.lwt.2022.113125. [DOI] [Google Scholar]
- San Y., Zheng L., Xing Y., Teng H., Ren A., Tang S.…Li B. A study on the structure and gel properties: Construction of co-assembed structures of rice globulin and soybean protein based on pH cycling technology. Food Hydrocolloids. 2026;172 doi: 10.1016/j.foodhyd.2025.111881. [DOI] [Google Scholar]
- Sang X., Zhen F., Zhang Q., Wang M., Qu B., Wang Y. Study on the optimization of enzyme-assisted cold isostatic pressure treatment of soybean meal for efficient fermentation. Food Chemistry. 2026;504 doi: 10.1016/j.foodchem.2026.148021. [DOI] [PubMed] [Google Scholar]
- Shangpliang H.N.J., Tamang J.P. Metagenomics and metagenome-assembled genomes mining of health benefits in jalebi batter, a naturally fermented cereal-based food of India. Food Research International. 2023;172 doi: 10.1016/j.foodres.2023.113130. [DOI] [PubMed] [Google Scholar]
- Shay N., Jing X., Bhandari B., Dayananda B., Prakash S. Effect of enzymatic hydrolysis on solubility and surface properties of pea, rice, hemp, and oat proteins: Implication on high protein concentrations. Food Bioscience. 2023;53 doi: 10.1016/j.fbio.2023.102515. [DOI] [Google Scholar]
- Tan C., Peng L., Huang X., Wang M., Zhang F., Mu H.…Zhao C. Microbiomics and flavoromics insights into the effect of high temperature steam pretreatment on the flavor of walnut meal-based soy sauce. Food Chemistry: X. 2025;31 doi: 10.1016/j.fochx.2025.103092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tan C., Peng L., Wang M., Mu H., Zhang F., Huang S.…Zhao C. High temperature steaming pretreatment improves quality and flavor of the walnut meal soy sauce. LWT. 2025;215 doi: 10.1016/j.lwt.2024.117249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang B., Shi Y., Zhang H., Hu Y., Chen H., Liu Y.…Chen L. Influence of microorganisms on flavor substances and functional components of sojae semen praeparatum during fermentation: A study integrating comparative metabolomics and high-throughput sequencing. Food Research International. 2024;187 doi: 10.1016/j.foodres.2024.114405. [DOI] [PubMed] [Google Scholar]
- Wang L., Wang Z., Chen Y., Chen J., Pan M., Cheong K.-L.…Zhong S. The effect of adding Gracilaria on flavor and quality of low-salt fermented soy sauce. LWT. 2024;210 doi: 10.1016/j.lwt.2024.116890. [DOI] [Google Scholar]
- Wang T., An J., Sha G., Bo N., Guan Y., Yang D.…Zhao M. A metagenomics analysis of microbial functional genes linked to vitamins metabolism in the pile fermentation of pu-erh tea. LWT. 2025;215 doi: 10.1016/j.lwt.2024.117215. [DOI] [Google Scholar]
- Wang X., Meng Q., Song H. Characterization of odor-active compounds in high-salt liquid-state soy sauce after cooking. Food Chemistry. 2022;373 doi: 10.1016/j.foodchem.2021.131460. [DOI] [PubMed] [Google Scholar]
- Wei Q., Wang H., Lv Z., Hu G., Li Y., Liu Y.…Lu F. Search for potential molecular indices for the fermentation progress of soy sauce through dynamic changes of volatile compounds. Food Research International. 2013;53(1):189–194. doi: 10.1016/j.foodres.2013.04.001. [DOI] [Google Scholar]
- Xu S., Huang D., Liu C., Gao Y., Li Q., Yu X. Development of nutrition-flavor dual process of rapeseed oil based on resource utilization of rapeseed cake. Food Chemistry. 2025;492 doi: 10.1016/j.foodchem.2025.145576. [DOI] [PubMed] [Google Scholar]
- Yan L., Xie C., Zhou J., Wang P., Tao Y., Yang R. Electrochemical advanced oxidation of glucosinolates derived from rapeseed meal: Parameter optimization, degradation mechanism and toxicity assessment. Chemical Engineering Journal. 2024;493 doi: 10.1016/j.cej.2024.152621. [DOI] [Google Scholar]
- Zhang F., Wang Y., Wang M., Tan C., Huang S., Mou H.…Zhao C. Structural characteristics and nonvolatile metabolites of theabrownins and their impact on intestinal microbiota in high-fat-diet-fed mice. Food Chemistry. 2025;463 doi: 10.1016/j.foodchem.2024.141317. [DOI] [PubMed] [Google Scholar]
- Zhang L., Xiong S., Du T., Xu Y., Madjirebaye P., Huang G.…Xiong T. Effect of microbiota succession on the dynamics of characteristic flavors and physicochemical properties during the soy sauce fermentation. Food Bioscience. 2023;54 doi: 10.1016/j.fbio.2023.102883. [DOI] [Google Scholar]
- Zhang L., Xiong S., Du T., Xu Y., Zhao X., Huang G.…Xiong T. Unraveling the core functional microbiota involved in metabolic network of characteristic flavor development during soy sauce fermentation. Food Bioscience. 2024;58 doi: 10.1016/j.fbio.2024.103697. [DOI] [Google Scholar]
- Zhang L., Zhang Y., Huang J., Zhou R., Wu C. Temperature-driven functional microbial interactions in soy sauce fermentation: Effects of Zygosaccharomyces rouxii and Wickerhamiella versatilis on flavor enrichment and biogenic amine reduction. International Journal of Food Microbiology. 2025;442 doi: 10.1016/j.ijfoodmicro.2025.111399. [DOI] [PubMed] [Google Scholar]
- Zhang W., Xiao Z., Gu Z., Deng X., Liu J., Luo X.…Jiang X. Fermentation-promoting effect of three salt-tolerant Staphylococcus and their co-fermentation flavor characteristics with Zygosaccharomyces rouxii in soy sauce brewing. Food Chemistry. 2024;432 doi: 10.1016/j.foodchem.2023.137245. [DOI] [PubMed] [Google Scholar]
- Zhang Y., Feng Y., Shi H., Ding K., Zhou X., Zhao G., Hadiatullah H. Impact of steam explosion pretreatment of defatted soybean meal on the flavor of soy sauce. LWT. 2022;156 doi: 10.1016/j.lwt.2021.113034. [DOI] [Google Scholar]
- Zhao C., Zhang Y., Lan F., Wu H., Li W., Lin W., Luo L. Impact of Staphylococcus carnosus DG06 inoculation on microbial and metabolic profiles during Cantonese soy sauce fermentation. Food Research International. 2025;211 doi: 10.1016/j.foodres.2025.116363. [DOI] [PubMed] [Google Scholar]
- Zhao S., Xu X., Li Z., Wang Q., Guo T., Dong B.…Zhao G. Long-chain fatty acid esters produced by Sporidiobolus pararoseus or Rhodotorula mucilaginosa enhance the fat flavor of soy sauce fermented by defatted soybeans. Food Chemistry. 2025;490 doi: 10.1016/j.foodchem.2025.145097. [DOI] [PubMed] [Google Scholar]
- Zheng J., Luo S., Aaqil M., Guo Y., Zhang F., Huang X.…Tian Y. From waste to flavor enhancer: Differential pressure explosion puffing modifies spent coffee grounds for microbial consortia-driven flavor synthesis in craft beer systems. Food Research International. 2025;218 doi: 10.1016/j.foodres.2025.116933. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that has been used is confidential.