Structural and proteomic characteristics of the extracted protein from thermal sterilized soy seeds and their effects on IgE reactivity, antioxidant activity and interfacial properties

Abstract

The study assessed protein structural and proteomic characteristics from thermal-sterilized (boiling and autoclaving) soy seeds and their effects on allergenicity, antioxidant activity and interfacial properties. Results demonstrated that boiling and autoclaving caused protein degradation and a reduction of soluble protein content and particle size. Scanning electron microscopy, Fourier transform infrared, ultraviolet and fluorescence spectroscopy showed that autoclaving is more effective than boiling in causing protein aggregation and secondary and tertiary structural changes. Proteomic analysis using LC/MS-MS revealed that autoclaving caused more changes of protein composition than boiling, with lower contents of major allergenic proteins and less reduction in linear epitopes. Thus, the autoclaved samples exhibited more reduction (81.5 %) of IgE reactivity than boiled samples (43.5 %). Additionally, autoclaving increased antioxidant activity more than boiling but caused a greater reduction in interfacial foaming and emulsifying) properties. This study provides a foundation for producing hypo-allergenicity soy products through industry sterilization techniques.

Graphical abstract

Highlights

• •Autoclaving and boiling altered protein structural and proteomic characteristics.
• •Autoclaving caused a higher reduction of IgE reactivity than boiling.
• •Autoclaving caused a higher increase of antioxidant activity than boiling.
• •Autoclaving caused a higher decreased foaming and emulsifying properties than boiling.
• •Autoclaving was more effective than boiling in producing low allergenic soy products.

1. Introduction

Soy proteins exhibit excellent nutritional values and health effects, making it been applied to various food products (Hammer et al., 2025; Sui et al., 2021). However, soy is one of the “Big Nine” allergenic foods and soy allergy is triggered by soy protein (Pi et al., 2024). Soy proteins are divided into the 2S, 7S, 11S, and 15S fraction, in which 7S and 11S fraction are key allergenic constituents (Pi, Sun, et al., 2021; Sui et al., 2021). It was reported that soy allergy resulted in various allergic symptoms such as conjunctivitis, angioedema, vomiting and even anaphylactic shock and death, seriously endangering 0.5 % of the global population's health (Pi, Sun, et al., 2021). Hence, accelerating the reduction in soy allergenicity is crucial for developing soy-based foods suitable for allergic individuals.

Thermal processing methods like boiling and autoclaving are industry sterilization techniques widely adopted in food production. These cost-effective, user-friendly processes effectively eliminate microorganisms and degrade anti-nutritional compounds in foodstuffs. Their simplicity and reliability make them go-to solutions for large-scale food production. Boiling and autoclaving were reported to decrease the allergenicity in soy, peanut, shellfish, milk, and egg (Pi et al., 2024). As for soy, limited studies revealed that boiling of soymilk for 5 min caused the reduction of protein electrophoretic band intensity, the increase of α-helix, ultraviolet absorption spectrum, surface hydrophobicity, and free sulfhydryl and the decrease of random coil to make proteins unfold and loosen, thereby affecting the conformational epitopes of proteins to decrease soy allergenicity (Lu et al., 2022). However, heating of soy protein isolate (Pi, Liu, Sun, Ban, Cheng, et al., 2023)), β-conglycinin (Li et al., 2021), and β-conglycinin α’ subunit (Xi et al., 2021; Xi & He, 2020) to affect allergenicity were widely investigated, which cannot simulate the actual processing situation, as heat processing often deals with soy seeds rather than their proteins. Additionally, the mechanism for the explanation of thermal processing to decrease soy allergenicity based on proteomic characteristics is limited. Previous studies have demonstrated that proanthocyanidins modification caused the reduction of major soybean allergens and the masking or destruction of dominant epitopes according to proteomic analysis based on LC/MS-MS, thus reducing the IgE reactivity in soy protein (Pi, Liu, Sun, Ban, Liang, et al., 2023). Similar results were shown in whey protein hydrolysates, demonstrating that the combined enzymes treatment increased conformational disruption and altered linear epitope hydrolysis sites based on LC/MS-MS, thus decreasing the IgE reactivity in whey protein (Pang et al., 2025). Furthermore, while reducing allergenicity, the functional properties of soy must be taken into account for the development of high-quality hypoallergenic soy-based foods, which is also limited. Previous studies have reported the increase of DPPH• radical scavenging activity total phenolic content after bitter bean was boiled (Muhialdin et al., 2020). The contrary results were shown in interfacial properties, showing a reduction of emulsifying and foaming properties in the extracted protein from autoclaved peanut, which was attributed to the protein aggregation and the reduction of surface hydrophobicity (Pi et al., 2022). Therefore, thermal processing affects food various food quality such as allergenicity, antioxidant activity and interfacial properties.

In the study, we assessed the alteration of allergenicity (IgE reactivity) and antioxidant activity (DPPH• and ABTS •+ radical scavenging activity), interfacial properties (emulsifying and foaming properties) in the extracted protein from thermal sterilized (boiling and autoclaving) soy seeds. The structural characteristics were determined by scanning electron microscopy (SEM), Fourier transform infrared (FTIR) spectroscopy, ultraviolet (UV) absorption spectroscopy, 8-anilino-1-naphthalene sulfonic acid (ANS) fluorescence probe, and Ellman methods. Proteomic characteristics were measured by LC/MS-MS. The relationship between allergenicity or functionality and structural or proteomic characteristics will be established, providing the insight into the effect of thermal sterilized technologies on soy allergenicity and functionality.

2. Methods and materials

2.1. Experimental materials

Soy samples (Glycine max Linn.) were sourced from Northeast Agricultural University. Serum from five individuals with soy allergies (Table S1) were procured from Chongqing Manuik Technology Co., Ltd. (Chongqing, China). Other chemicals were obtained from Sigma-Aldrich (St. Louis, MO, USA).

2.2. Preparation of thermal sterilized soy seeds

Raw soy seeds were soaked in distilled water at a 1:20 (w/v) ratio for 12 h, followed by a 2-min draining period. The seeds were then subjected to two different thermal sterilized treatments: boiling in water at 100 °C for 20 min and autoclaving in a high-pressure steam sterilizer (ZDX-35BI, ShenAn Medical, China) at 121 °C for 20 min. These processes yielded boiled soy (BS) and autoclaved soy (AS), respectively.

2.3. Extraction of soy protein

The isolation of soy proteins was adapted from Lu et al. (2022) with minor adjustments, which was performed identically for raw and thermally treated samples. Soybeans were ground into powder and subjected to lipid removal using n-hexane in a 1:15 (w/v) ratio for 3 h. This defatting procedure was carried out twice, with the defatted soy powder subsequently separated via centrifugation at 10,000×g for 20 min. Following thorough drying, the defatted flour was homogenized in distilled water (pH 8.5), agitated for 3 h, and then centrifuged to collect the supernatant. This process was repeated twice. The combined supernatants from all three extractions served as the protein isolate mixture. The samples were freeze-dried until use.

2.4. SDS-PAGE

Each sample (0.5 mg/mL) was mixed with loading buffer (10 mM Tris–HCl buffer with pH 8.0 containing 1 % β-mercaptoethanol, 1 % SDS, 0.02 % bromophenol blue and 40 % (v/v) glycerol) and heated at 100 °C for 3 min. A 10 μL aliquot was loaded onto a 12 % separating gel with a 5 % stacking gel, and then performed electrophoresis. After staining with Coomassie Brilliant Blue R-250G and destaining in acetic acid solution, the gel was imaged using an Odyssey cLx scanner (LI-COR, USA) (He et al., 2020).

2.5. Soluble protein content

The protein extraction mixture (section 2.3) was mixed with Coomassie reagent in a 1:5 (v/v) ratio, incubated at 25 °C for 10 min, and then measured the absorbance at 595 nm by a microplate reader (SpectraMax reg iD3, BeckmanCoulter, US) with a standard of bovine serum albumin (Pi et al., 2022).

2.6. Particle size

The sample (0.1 mg/mL) was assessed by a Mastersizer particle size analyzer (Malvern Instruments Co., Ltd., Worcestershire, UK) (Pi, Fu, et al., 2021).

2.7. Structural analysis

2.7.1. SEM

The sample (0.1 mg/mL) was applied to a mono crystal silicon wafer and dried under vacuum at low temperatures. Subsequent the photographs were captured using the SEM (SU8000, HITACHI, Japan) (He et al., 2020).

2.7.2. FTIR spectroscopy

Each freeze-dried sample was mixed with potassium bromide, tableted, and measured by an FTIR spectrometer (Nicolet iS10, Thermo Fisher, US) under 400–4000 cm⁻¹. The number of scans was 32, and the resolution was 2 cm⁻¹ (Li et al., 2021).

2.7.3. UV absorption spectroscopy

The sample (0.125 mg/mL) was measured by a T9 UV–visible spectrophotometer (Puxi Analytical Instrument Co. Ltd, Beijing, China) under 200–500 nm (Zhang et al., 2022).

2.7.4. ANS fluorescence probe

The sample (0.25 mg/mL) was mixed with ANS solution (8.0 mM, pH 7.4) at a volume ratio of 400:1, incubated for 2 h, and then measured under 420–600 nm emission wavelengths and 390 nm excitation wavelengths at a scanning speed of 240 nm/min by a cary eclipse fluorescence spectrometer (F-7100, HITACHI, Japan) (Lu et al., 2022).

2.7.5. Ellman methods

The free SH group content was assessed using Ellman methods (He et al., 2020). Firstly, 1.04 g of Tris, 0.68 g of glycine and 0.12 g of EDTA were dissolved in 10 mL deionized water and adjusted the pH 8.0, obtaining the Tris-Gly solution. Then, each sample (0.5 mg/mL) was incubated with Ellman's reagents (4 mg/mL 5,5’-Dithiobis-(2-nitrobenzoic acid) in Tris-Gly solution) at a volume ratio of 27:1 for 15 min, and measured at 412 nm. Finally, the free sulfhydryl (SH) content was calculated according to the formula (1).

$$\text{Free}\mathrm{SH}\text{group content}\left(\text{μmol}/\mathrm{g}\text{of protein}\right)=73.5\times \mathrm{O}{\mathrm{D}}_{412}\times \mathrm{D}/\mathrm{C}$$ (1)

where OD₄₁₂ was the absorbance of the sample at 412 nm, C was sample concentration (mg/mL), and D was the dilution ratio.

2.8. Proteomics analysis

2.8.1. LC/MS-MS

The LC/MS-MS was conducted according to Pi, Liu, Sun, Ban, Cheng, et al. (2023). The sample (150 μg) was first treated with 1 M dithiothreitol at 56 °C for reduction, followed by alkylation using 60 mM iodoacetamide at 25 °C. The processed samples were filtered through a 10-kDa molecular weight cutoff filter and then rinsed and washed to obtained sample extraction. The sample extraction was incubated with trypsin overnight at 37 °C, and the reaction was quenched by adding a 10 % C₂HF₃O₂ solution. The resulting peptides were purified using C18 solid-phase extraction tips, dried under vacuum, and stored at −20 °C until further analysis. Peptide separation and identification were carried out using an EASY-nLC 1200 system (Thermo Fisher Scientific, USA). Approximately 5 μg of the peptide mixture was reconstituted in 0.1 % aqueous formic acid, loaded onto a C18 pre-column, and then separated on a C18 analytical column with a gradient of 0.1 % formic acid in 80 % acetonitrile at a flow rate of 300 nL/min. Following separation, the peptides were analyzed using a Q Exactive Plus-Orbitrap mass spectrometer. Full MS scans were acquired over a mass range of 350–2000 m/z with a spray voltage of 1.8 kV and a capillary temperature of 300 °C. Tandem MS (MS/MS) was performed within a scan range of 200–2000 m/z.

2.8.2. Protein identification

The obtained mass data were analyzed using Proteome Discoverer 24 Software (Thermo Fisher Scientific) under the soy database (https://www.uniprot.org/). Protein was identified under the False Discovery Rate (FDR) Threshold (≤1 %) and peptide count (≥1). Protein and peptide abundance were measured using the label-free quantification method with raw samples as the standard (Pi, Liu, Sun, Ban, Cheng, et al., 2023).

2.9. Allergenicity

The allergenicity was evaluated by the IgE reactivity using direct ELISA (Pi, Sun, Liu, Wang, Hong, et al., 2023). Protein samples (5 μg/mL, 100 μL) were dispensed into microtiter plate wells and allowed to coat overnight at 4 °C. Following a wash with PBST buffer (0.01 M phosphate buffer containing 0.05 % Tween 20) and a 60-min blocking step with skim milk at 37 °C, the well was incubated with human sera for 1 h at 37 °C. After another wash cycle, 100 μL of goat anti-human IgE HRP conjugates (diluted 1:4000 in PBS) were introduced and incubated for 1 h. Subsequent to further washing, 100 μL of TMB substrate solution was added for a 13-min incubation period before the reaction was stopped with 100 μL of 0.2 M sulfuric acid. After absorbance at 450 nm was measured, the IgE reactivity was calculated according to the formula (2).

$$\text{The}\mathrm{IgE}\text{reactivity}\left(\%\right)=\frac{{\mathrm{OD}}_{450}\left(\text{Other samples}\right)}{{\mathrm{OD}}_{450}\left(\mathrm{Raw}\text{sample}\right)}\times 100\%$$ (2)

2.10. Antioxidant capacities

2.10.1. DPPH• radical scavenging activity

The sample solution (0.2 mg/mL) was mixed with DPPH• solution (0.2 mM, methanol), then incubated in darkness at room temperature for 30 min. The absorbance at 517 nm was detected. The DPPH• radical scavenging activity were calculated according to formula (3) (Zhang et al., 2023).

$$\text{The DPPH}\dot{}\text{radical scavenging rate}\left(\%\right)=\frac{{\mathrm{A}}_0-{\mathrm{A}}_{\text{sample}}}{{\mathrm{A}}_0}\times 100\%$$ (3)

where, A_sample and A₀ were the absorbance of sample and distilled water, respectively.

2.10.2. ABTS•+ radical scavenging activity

The sample solution (0.2 mg/mL) was mixed with ABTS •+ solution (3.5 mM, potassium persulfate solution), then incubated in darkness at room temperature for 10 min. The absorbance at 734 nm was detected. The ABTS •+ radical scavenging activity were calculated according to formula (4) (Zhang et al., 2023).

$$\text{The ABTS}\dot{}+\text{radical scavenging rate}\left(\%\right)=\frac{{\mathrm{A}}_0-{\mathrm{A}}_{\text{sample}}}{{\mathrm{A}}_0}\times 100\%$$ (4)

where, A_sample and A₀ were the absorbance of sample and distilled water, respectively.

2.11. Interfacial properties

2.11.1. Foaming properties

Foaming properties were determined by measuring foaming capacity (FC) and stability (FS). Sample solution (7.5 mL) was stirred, and then recorded the volume after 0 min and 10 min, respectively. FC and FS were calculated according to the (5), (6), respectively (Pi et al., 2022).

$$\mathrm{FC}\left(\%\right)=\frac{{\mathrm{V}}_0-\mathrm{V}}{\mathrm{V}}\times 100\%$$ (5)

$$\mathrm{FS}\left(\%\right)=\frac{{\mathrm{V}}_{10}-\mathrm{V}}{{\mathrm{V}}_0-\mathrm{V}}\times 100\%$$ (6)

where, V is the starting volume of the sample, clocking in at 7.5 mL. Meanwhile, V₀ and V₁₀ denote the volume of well-stirred sample post 0 min and after 10 min.

2.11.2. Emulsifying properties

Emulsifying properties were determined by measuring emulsion activity index (EAI) and stability index (ESI). In brief, a 2 mg/mL sample solution was blended with soybean oil in a 3:1 volume ratio and homogenized at 10,000 rpm for two minutes. Aliquots (50 μL) of the emulsion were collected immediately and after 10 min, combined with 200 μL of 0.1 % SDS solution, and analyzed spectrophotometrically at 500 nm. EAI and ESI were calculated according to the (7), (8), respectively (Hu & Li, 2022).

$$\mathrm{EAI}\left({\mathrm{m}}^2/\mathrm{g}\right)=\frac{2\times 2.303\times {\mathrm{A}}_0\times \mathrm{D}}{\mathrm{N}\times \mathrm{C}\times 10000}$$ (7)

$$\mathrm{ESI}\left(\min \right)=\frac{{\mathrm{A}}_0}{{\mathrm{A}}_0-{\mathrm{A}}_{10}}\times \mathrm{T}$$ (8)

where, A₀ and A₁₀ referred to the absorbance of the emulsion at the 0 min and 10 min, respectively. T, D, C, and N were time (10 min), dilution factor, protein content (g/mL), and volumetric oil fraction, respectively.

2.12. Statistical analysis

All experiments were repeated three times. Analysis of variance (ANOVA) and Duncan's mean comparison test were applied with a significance level of 0.05 by using SPSS software (Windows 2003, USA). The result data was expressed as “mean ±standard deviation”. Allergenic peptides were identified using the database (http://imed.med.ucm.es/Tools/antigenic.pl, https://webs.iiitd.edu.in/raghava/abcpred/ABC_submission.html) and the previously reported epitopes (Havenith et al., 2017, He and Xi, 2020; Helm et al., 2000; Liu et al., 2012; Sun et al., 2013; Vanga et al., 2019; Wang et al., 2014; Xi & He, 2020) (Table S2). If the peptide exhibited the average antigenic propensity above 0.5 and contained above 5 continuous amino acid fragments of the epitopes, the peptide is considered as allergenic peptides. Venn diagram, Peptide map, and heat map were drawn by some online tools (e.g., https://bioinfogp.cnb.csic.es/tools/venny/index.html; http://bioware.ucd.ie/peptigram/ and https://cloud.metware.cn/#/home). The correlation analysis was performed by Pearson's correlation analysis using by Origin 2021 (Origin Lab Corporation, Northampton, MA).

3. Result and discussion

3.1. SDS-PAGE

The SDS-PAGE profiles of protein extracts from RS, BS and AS samples were presented in Fig. 1A. Compared to RS, BS exhibited a reduced band intensity for higher molecular weight proteins (>43 kDa) alongside increased staining intensity for lower molecular weight fractions (<43 kDa) (land 2). These findings align with previous research by Cabanillas et al. (2012), who observed similar electrophoretic pattern shifts in heat-treated peanuts, with diminished staining in the 50–81 kDa range but enhanced intensity between 19 and 50 kDa. Notably, autoclaving produced more dramatic changes, causing the virtual disappearance of protein bands above 10 kDa (lane 3). Therefore, autoclaving caused more alteration of SDS-PAGE protein patterns than boiling due to its high temperature, suggesting more protein degradation (Pi et al., 2022). Comparative research has demonstrated that autoclaved silkworm pupa proteins exhibit greater band disappearance on SDS-PAGE gels compared to boiled samples when processed for equivalent durations (He et al., 2021). Thermal processing could also trigger protein denaturation and aggregation, along with chemical interactions between proteins and other components like carbohydrates and lipids, ultimately reducing protein solubility (Cabanillas et al., 2012; Kasera et al., 2012; Pi, Sun, et al., 2021), which might cause the loss of certain protein extraction to make SDS-PAGE protein patterns shallow and even disappear. At present, the World Health Organization and International Union of Immunological Societies (WHO/IUIS) Allergen Nomenclature Subcommittee has identified 8 soy allergens, namely Gly m 1 to Gly m 8 (Sub-Committee, 2025). Therefore, the allergen composition of was altered after soy was boiled and autoclaved (Fig. 1A), probably resulting in the reduction of allergenicity in soy. Similar results were shown by Wang et al. (2022), who found a reduction of allergenicity in the boiled and fried peanuts, accompanied by the reduced quantities of allergens Ara h 1 (64 kDa), Ara h 2 (20 kDa), and Ara h 3 (38 kDa).

Fig. 1.

The effect of boiling and autoclaving on the protein profiles (A), soluble protein content (B), particle sizes distribution (C), average particle sizes (D) after soy was boiled and autoclaved. Means with different letters (a–c) in the bars indicated significant difference (P < 0.05).

3.2. Soluble protein content

As shown in Fig. 1B, boiling and autoclaving led to a significant (P < 0.05) reduction of 31 % and 66.7 % in the soluble protein content, respectively. This result was occurred, probably because thermal sterilization caused the protein aggregation and chemical interactions between proteins and other components like carbohydrates and lipids (Pi, Sun, et al., 2021). Similar results were found by Kasera et al. (2012), who found that boiling of legume decreased the soluble protein content from 11.3 mg/mL to 3.70 mg/mL. Pi et al. (2022) also showed a substantial decrease of 72.7 % in the soluble protein found in autoclaved peanut, likely due to protein-aggregation or the formation of protein-carbohydrate or lipid complexes, which diminished the protein solubility.

3.3. Particle sizes

As shown in Fig. 1C, the majority percentage of particle sizes in RS samples was distributed about at 4000–7000 nm and 50–1000 nm. After boiling and autoclaving, the particle size was decreased. A similar shift in particle size distribution was shown by Pi, Fu, et al. (2021), who found that raw peanut extracts predominantly contained particles around 1000 nm and 200 nm, and autoclaving resulted in a significant redistribution, with the majority of particles clustering near 550 nm and 100 nm. As shown in Fig. 1D, boiling and autoclaving led to a significant 34.9 % and 16.5 % reduction (P < 0.05) in the average particle sizes (nm) of both two protein extractions namely boiled (212.8 ± 6.1 nm) and autoclaved (273.0 ± 12.0 nm) soy as compared to raw soy (327.0 ± 16.5 nm). Previous studies reported the increased peptide contents and hydrolysis degree in thermal processed (boiling and autoclaving) bitter beans, peanut seeds and sweet potato proteins, suggesting the occurrence of proteins-degradation (Muhialdin et al., 2020; Pi et al., 2020; Zhang et al., 2019). Therefore, the reduction of average particle sizes was observed in boiled and autoclaved samples, probably because proteins may undergo hydrolysis or degradation, resulting in smaller fragments and a decrease in average particle size. In addition, autoclaved samples exhibited a higher average particle size than boiled samples, probably resulting from a higher protein-aggregation induced by autoclaving.

3.4. Micro-structure

As shown in Fig. 2, boiling and autoclaving caused noticeable micro-structure alterations in the proteins, transforming their appearance from uniform ellipses to fragmented forms. The once-smooth surfaces became visibly rough, with the formation of protein-aggregates becoming evident. This phenomenon indicated the degradation and aggregation of protein, which was consistent with Fig. 1. In addition, more protein-aggregates were observed in autoclaved samples (Fig. 2), suggesting a higher protein-aggregation induced by autoclaving, which was consistent with Fig. 1D. It was reported that protein-aggregations were related to the alteration of conformational structures (Luo et al., 2013). Therefore, conformational alterations were further investigated.

Fig. 2.

The effect of boiling and autoclaving on micro-structure after soy was boiled and autoclaved. The magnification is 10,000 times.

3.5. FTIR

As shown in Fig. 3A, variations in peak intensity and wavelength were evident across all samples. The RS spectrum displayed three prominent bands at 3273.75, 1636.33, and 1548.52 cm⁻¹, which are characteristic of protein amide A, amide I, and amide II vibrations, respectively (Pi, Liu, Sun, Ban, Liang, et al., 2023). However, when compared to RS samples, BS and AS samples exhibited noticeable changes in both the strength and location of these peaks. Notably, the autoclaved samples showed a complete disappearance of the amide II. The amide I peak arises from the stretching vibrations of peptide-bonded carbonyl groups (C O), while the amide II band results primarily from N—H bending and C—N stretching motions (Pi et al., 2022). Therefore, boiling and autoclaving caused the alteration of C O, N—H and C—N groups in soy protein.

Fig. 3.

The effect of boiling and autoclaving on secondary structure after soy was boiled and autoclaved. A. Fourier transform infrared spectroscopy, B. quantify the proportions of various secondary structural elements, C. the content of β-turn, α-helix, random coil, and β-sheet. Means with different letters (a–c) in the bars indicated significant difference (P < 0.05).

The 1700–1600 cm⁻¹ spectral range is particularly rich in data concerning protein secondary structures (Li et al., 2021). By analyzing this region with PeakFit v4.12, researchers were able to quantify the proportions of various secondary structural elements, as illustrated in Fig. 3B. The Amide I band (1600–1700 cm⁻¹) provided insights into a protein secondary structure, including β-turn (1660–1700 cm⁻¹), α-helix (1650–1660 cm⁻¹), random coil (1640–1650 cm⁻¹), and β-sheet (1600–1640 cm⁻¹) (Zhang et al., 2023). As shown in Fig. 3C, RS sample contained 13.1 % α-helix, 35.6 % β-sheet, 30.3 % β-turn and 21.0 % random coil content. The contents for BS sample were 14.6 %, 29.9 %, 35.0 % and 20.5 %, respectively, and for AS sample were 4.5 %, 90.2 %, 3.4 % and 1.9 %, respectively. Thus, autoclaving caused more alteration of the secondary structure than boiling in soy protein. In addition, the reduction of β-sheet suggested the exposure of hydrophobic groups (Zhao et al., 2020). Therefore, boiling caused the exposure of hydrophobic group, whereas autoclaving caused the masking of hydrophobic group.

3.6. UV absorption spectra

UV absorption spectra provide useful information for protein conformational structure, which is related with aromatic amino acid residues such as tryptophan and tyrosine residues (Li et al., 2021). As shown in Fig. 4A, boiling and autoclaving caused an increase of the UV absorbance at 280 nm, suggesting that tryptophan and tyrosine residues were exposed (Pi et al., 2022). Compared to BS samples, AS samples exhibited a lower UV absorbance, probably because autoclaving caused more protein-aggregation than boiling to mask tryptophan and tyrosine residues. This was consistent with the results of Fig. 2.

Fig. 4.

The effect of boiling and autoclaving on tertiary structure after soy was boiled and autoclaved. A. ultraviolet absorption spectra, B. surface hydrophobicity, C. the content of free sulfhydryl groups. Means with different letters (a–b) in the bars indicated significant difference (P < 0.05).

3.7. Surface hydrophobicity

Changes in the hydrophobic nature of proteins typically mirror shifts in their tertiary and quaternary structures, which were measured by ANS anion fluorescence probe. Compared with RS samples, the ANS-fluorescence intensity of BS samples was increased by 21.9 % (Fig. 4B), suggesting the exposure of more hydrophobic regions caused by structural changes (Li et al., 2021). In addition, the ANS-fluorescence intensities of AS samples were decreased by 35.6 % (Fig. 4B), suggesting the masking of hydrophobic regions or the destruction of exposed hydrophobic structures induced by the aggregation of protein (Pi et al., 2022). In addition, there was a blue shift of the ANS-fluorescence intensity in AS samples compared to BS samples, indicating that autoclaving caused more protein aggregation than boiling, which was consistent with the results of Fig. 2.

3.8. Free sulfhydryl (SH) groups

The SH can either exist as a free radical or participate in the formation of disulfide bonds, both of which are crucial for protein tertiary structure. As demonstrated in Fig. 4C, Free SH group contents were increased by 37.8 % and 9.4 % in BS and AS samples, which was attributed to the breakage of disulfide bonds and exposure of free SH (Pi, Fu, et al., 2021; Zhou et al., 2016). In addition, there was a lower content of free SH groups in AS samples than BS samples, indicating that autoclaving caused the masking of the exposed free SH groups, likely due to more protein aggregation in autoclaving (Pi et al., 2022). This result was consistent with the results of Fig. 2.

3.9. Protein proteomic analysis

3.9.1. Protein composition

The effect of boiling and autoclaving on protein composition in soy was shown in Fig. 5 using protein proteomic analysis based on LC/MS-MS. As shown in Fig. 5A, there were 27, 26, and 28 proteins in RS, BS and AS samples. The number of proteins was altered after soybean was boiled and autoclaved, probably because protein structural changes altered the cleavage site of trypsin (Xu et al., 2022). The alteration of protein composition was further assessed by heat map (Fig. 5B). RS exhibited a higher content of some proteins (A0A2K4ZJ99∼P2597), while BS exhibited a higher concentration of some proteins (A0A063XB29∼A0A375BY25) and AS exhibited a higher abundance of some proteins (P0DO16∼A0A375BT21). Therefore, besides for protein number, the content of proteins was altered after soybean was boiled and autoclaved. These results occurred probably because protein structural alteration affected the ease of access for trypsin to the peptide bonds, consequently impacting the protein's detection via LC/MS-MS to alter the quantity and composition of proteins. In addition, heat map analysis categorizes RS and BS as one type, indicating a similar protein composition in RS and BS samples. Therefore, autoclaving caused a higher alteration of protein composition than boiling, which was consistent with Fig. 1A.

Fig. 5.

The effect of boiling and autoclaving on protein (A, B) and allergenic protein composition (C) after soy was boiled and autoclaved.

3.9.2. Allergenic protein composition

Gly m 4, Gly m 5, and Gly m 6 are major allergens in soy (Pi, Sun, et al., 2021). As demonstrated in Fig. 5C, Gly m 4 (SAM22; Uniport P26987), Gly m 5 (Beta-conglycinin; P0DO16, P11827, F7J077, P25947), Gly m 6 (Glycinin; P04776, P04405, P11828, P02858, P04347) were identified in the protein after soy was boiled and autoclaved. Although these major soy allergens were exhibited in all samples (consistent with Fig. 5A), BS and AS showed a 12.2 % and 44.2 % reduction of major soy allergens. There was 99.4 % and 98.2 % reduction of Gly m 4 and 70.8 % and 26.0 % of Gly m 5 after soy was boiled and autoclaved. However, boiling caused 160.5 % increase in Gly m 6, while autoclaving induced 29.5 % reduction in Gly m 6. Therefore, besides for protein composition, boiling and autoclaving also caused the alternation in major soy allergen compositions, which varied from the types of major soy allergens, resulting from structural changes.

3.9.3. The peptide composition in major allergenic protein

The alteration of peptide compositions in major allergenic protein was evaluated by peptide mapping (Fig. 6A–C) and corresponding heat map analysis (Fig. 6a–c). As demonstrated in Fig. 6A (P26987, Gly m 4), five peptides (AA22–33, AA40–54, AA40–55, AA55–65, AA56–65) were detected in the protein after soy was boiled and autoclaved, in which AA22–33 (ALVTDADNVIPK), AA40–54 (SVENVEGNGGPGTIK) and AA40–55 (SVENVEGNGGPGTIKK) were regarded as allergic peptides because they included the previously reported epitopes such as TDADN and EGNGGPGTIK (Table S2) (Vanga et al., 2019). Additionally, compared to the RS samples, BS and AS samples exhibited a lower abundance in AA22–33, AA40–54 and AA40–55, which suggested that these allergic peptides were masked due to protein structural changes (Pi, Sun, Liu, Wang, Hong, et al., 2023). As for Gly m 5, it contained α, α’ and β subunits. As demonstrated in Fig. 6B (α subunit of Gly m 5, P0DO16), BS and AS samples also exhibited a decrease of abundance in some allergic peptides [e.g., FFEITPEK (AA86-112), NPFLFGSNR (AA152-162), LITLAIPVNKPGR (AA306-318), TISSEDKPFNLR (AA422-434), SRDPIYSNK (AA575-584)]. Similar results were shown in the α’ (Fig. 6B, P11827) and β subunit (Fig. 6B, F7J077, P25974) of Gly m 5. As for Gly m 6, it contains G1-G5 subunit. As shown in Fig. 6C (G5 subunit of Gly m 6, P04347), there was also the reduction of abundance in some allergic peptides such as GAIGFAFPGCPETFEKPQQQSSR (AA100-122) and VFYLAGNPDIEHPETMQQQQQQK (AA186-208). However, boiled and autoclaved samples exhibited an increased abundance in the allergic peptide such as QIVTVEGGLSVISPK (AA260-274), whereas the allergic peptide KQIVTVEGGLSVISPK (AA259-274) showed a decreased and increased abundance in boiled and autoclaved samples, respectively. Therefore, the alteration of allergic peptides varied from the treatment. Similar results were shown in the G1 (Fig. 6C, P04776), G2 (Fig. 6C, P04405), G3 (Fig. 6C, P11828), and G4 (Fig. 6C, P02858) subunit of Gly m 6. As shown Fig. 6a–c, heat map analysis categorizes AS and BS as one type in Gly m 4 (Fig. 6a) and Gly m 5 (Fig. 6b), while categorizing AS and RS as one type in G1 (Fig. 6c, P04776), G2 (Fig. 6c, P04405), G3 (Fig. 6c, P11828), and categorizing BS and RS as one type in G4 (Fig. 6c, P02858) and G5 (Fig. 6c, P04347) subunit of Gly m 6. Therefore, the alteration of allergic peptides varied from the allergen types. Besides for conformational epitopes, linear epitopes can also trigger allergic reaction (Pang et al., 2026; Pi et al., 2025). Table 1 summarized changes in linear IgE epitopes in BS and AS samples. It was shown that the total content of linear IgE epitopes was decreased by 62.8 % and 47.7 % after soy was boiled and autoclaved, involving Gly m 4, the α and β subunit of Gly m 5, and the G4 subunit of Gly m 6, which suggested that many linear IgE epitopes were masked or destroyed.

Fig. 6.

The peptide map (A–C) and corresponding heat map (a–c) of peptides in major soy allergens after soy was boiled and autoclaved. For the peptide map, the color depth, the width and length of each band represent the relative abundance, amino acid coverage and overlap ratio. Decreased abundance at a position of peptide maps means that the corresponding peptide is masked, and vice versa.

Table 1.

The alteration of linear allergic peptides in major soy allergens.

Allergic peptides	Soy allergen	Intensity
		RS	BS	AS
ALVTDADNVIPK*	P26987	290.7	2.7	6.6
SVENVEGNGGPGTIK*		299.2	0.8	0
SVENVEGNGGPGTIKK*		300	0	0
Gly m 4 (P26987)		889.9	3.5	6.6
FFEITPEK*	P0DO16	275.6	19.2	5.2
FFEITPEKNPQLR*		283.5	0	16.5
EEDEDEQPRPIPFPRPQPR		118.2	29.5	152.2
GSEEEDEDEDEEQDER*		186.7	38.4	75
GSEEEDEDEDEEQDERQFPFPRPPHQK		0	0	300
NKNPFLFGSNR*		279.2	3.7	17.1
NPFLFGSNR*		294.7	2.9	2.3
QEEDEDEEQQR*		223.3	76.7	0
VEKEECEEGEIPRPRPRPQHPER		0	0	300
EEQEWPR		73.8	51.9	174.4
EEQEWPRK		29.6	14.5	255.9
EPQQPGEKEEDEDEQPRPIPFPRPQPR		5.4	11.4	283.2
ESEESEDSELR*		300	0	0
ESEESEDSELRR*		276	0	24
ESYFVDAQPK*		268	0.1	31.8
KQEEDEDEEQQR		114.3	54	131.7
LITLAIPVNKPGR*		285.2	10.7	4.1
QEEEHEQREEQEWPR		7.1	12.2	280.7
QFPFPRPPHQK		98.2	13.3	188.5
QFPFPRPPHQKEER		0	0	300
VPSGTTYYVVNPDNNENLR*		300	0	0
TISSEDKPFNLR*		300	0	0
KTISSEDKPFNLR*		295.3	4.7	0
SRDPIYSNK*		297.4	2	0.7
The α subunit (P0DO16) of Gly m 5		4311.5	345.2	2543.3
RFYLAGNQEQEFLQYQPQK	P11828	96	69	135
FYLAGNQEQEFLQYQPQK		39	214	47.1
KLQGENEEEEKGAIVTVK		16.4	9.3	274.3
LQGENEEEEKGAIVTVK		15	45	240
NAMFVPHYNLNANSIIYALNGR		78.3	221.7	0
VFDGELQEGQVLIVPQNFAVAAR		84.7	206.4	8.9
The α'subunit (P11828) of Gly m 5		329.4	765.4	705.3
AILTLVNNDDR*	F7J077	295.6	0	4.4
IPAGTTYYLVNPHDHQNLK*		300	0	0
LAIPVNKPGR*		203.5	80.6	15.8
TISSEDEPFNLR*		300	0	0
NPIYSNNFGK*		266.4	0	33.6
FFEITPEK*		275.6	19.2	5.2
FFEITPEKNPQLR*		283.5	0	16.5
ESYFVDAQPQQK*		216.9	5.3	77.8
The β subunit (F7J077) of Gly m 5		2141.5	105.1	153.3
AILTLVNNDDR*	P25974	295.6	0	4.4
IPAGTTYYLVNPHDHQNLK*		300	0	0
LAIPVNKPGR*		203.5	80.6	15.8
TISSEDEPFNLR*		300	0	0
NPIYSNNFGK*		266.4	0	33.6
FFEITPEK*		275.6	19.2	5.2
FFEITPEKNPQLR*		283.5	0	16.5
ESYFVDAQPQQK*		216.9	5.3	77.8
The β subunit (P25974) of Gly m 5		2141.5	105.1	153.3
The β subunit (F7J077 + P25974) of Gly m 5		4283	210.2	306.6
Gly m 5		8923.9	1320.8	3555.2
RPSYTNGPQEIYIQQGK*	P04776	252.9	47.1	0
GIFGMIYPGCPSTFEEPQQPQQR*		150	92.8	57.1
RFYLAGNQEQEFLK		52.5	36.3	211.2
FYLAGNQEQEFLK		53.1	195.5	51.4
NAMFVPHYNLNANSIIYALNGR		78.3	221.7	0
The G1 subunit (P04776) of Gly m 6		586.8	593.4	319.7
RFYLAGNQEQEFLK	P04405	52.5	36.3	211.2
FYLAGNQEQEFLK		53.1	195.5	51.4
QQEEENEGSNILSGFAPEFLK		63.8	125.5	110.7
EAFGVNMQIVR		131.9	205.9	262.2
NLQGENEEEDSGAIVTVK*		178.1	39.9	82
KPQQEEDDDDEEEQPQCVETDKGCQR		34.5	31	234.5
VFDGELQEGGVLIVPQNFAVAAK		83	187.2	29.8
The G2 subunit (P04405) of Gly m 6		596.9	821.3	981.8
RFYLAGNQEQEFLQYQPQK	P11828	96	69	135
FYLAGNQEQEFLQYQPQK		39	214	47.1
KLQGENEEEEKGAIVTVK		16.4	9.3	274.3
LQGENEEEEKGAIVTVK		15	45	240
NAMFVPHYNLNANSIIYALNGR		78.3	221.7	0
VFDGELQEGQVLIVPQNFAVAAR		84.7	206.4	8.9
The G3 subunit (P11828) of Gly m 6		329.4	765.4	705.3
VFYLAGNPDIEYPETMQQQQQQK*	P02858	377.3	100.1	122.5
KQIVTVEGGLSVISPK		45.7	35.3	219
QIVTVEGGLSVISPK		40.6	82.8	176.6
RGQLLVVPQNFVVAEQAGEQGFEYIVFK		97.3	202.7	0
GQLLVVPQNFVVAEQAGEQGFEYIVFK		80.3	219.7	0
YEGNWGPLVNPESQQGSPR*		161.3	138.7	0
The G4 subunit (P02858) of Gly m 6		802.5	779.3	518.1
GAIGFAFPGCPETFEKPQQQSSR*	P04347	292.3	7.7	0
VFYLAGNPDIEHPETMQQQQQQK*		302.1	123.8	174.1
KQIVTVEGGLSVISPK		45.7	35.3	219
QIVTVEGGLSVISPK		40.6	82.8	176.6
HEDDEDEDEEEDQPRPDHPPQRPSRPEQQEPR		89.4	22.5	188.1
The G5 subunit (P04347) of Gly m 6		770.1	272.1	757.8
Gly m 6		3328.5	3560.4	3310.9
Total (Gly m 4∼Gly m 6)		13,142.3	4884.7	6872.7

Note: The underline indicates the fragment that overlaps with the reported epitope. “*” represents the epitope changes potentially responsible for the reduced allergenicity.

3.10. Allergenicity

Soy allergy is an IgE-mediated allergic reaction (Pi, Sun, et al., 2021), and soy allergenicity is reflected by the alteration of IgE reactivity based on ELISA. As shown in Fig. 7A, the IgE reactivity was decreased from 103.5 ± 3.9 % to 58.5 ± 2.1 % and 19.1 ± 1.0 % after soy was boiled and autoclaved, respectively. Therefore, boiling and autoclaving could cause the reduction of allergenicity in soy. In addition, the autoclaved soy exhibited a higher reduction (81.5 %) of allergenicity than boiled samples (43.5 %). Similar results were shown in boiled (Wang et al., 2022) and autoclaved peanut (Pi, Fu, et al., 2021). Cai et al. (2016) further found that autoclaving caused more reduction of allergenicity in rAra h 2.02 than boiling, which was attributed to the decreased binding capacities of the three core IgE-binding epitopes induced by structural changes. Therefore, in this study, the boiled and autoclaved soy exhibited the reduction of allergenicity, which was related to the structural changes and epitope alteration.

Fig. 7.

The effect of boiling and autoclaving on allergenicity, antioxidant activity and interfacial properties after soy was boiled and autoclaved. A. IgE reactivity, B. antioxidant activity, C. foaming properties, D. emulsifying properties. Means with different letters (a–c) in the bars indicated significant difference (P < 0.05).

Food allergenicity is depended on its allergens and epitopes (linear and conformational epitopes) (Pang et al., 2026; Pi, Sun, Liu, Wang, Hong, et al., 2023). As shown in Fig. 1A, protein degradation was observed in boiled and autoclaved soy, which caused the generation of peptides to make proteins unrecognize by serum (Kim et al., 2011; Tao et al., 2016). As a result, some linear epitopes were destroyed, involving the epitopes ALVTDADNVIPK and SVENVEGNGGPGTIKK from Gly m 4, FFEITPEK and GSEEEDEDEDEEQDER from α subunit of Gly m 5, AILTLVNNDDR and IPAGTTYYLVNPHDHQNLK from β subunit of Gly m 5, RPSYTNGPQEIYIQQGK and GIFGMIYPGCPSTFEEPQQPQQR from G1 subunit of Gly m 6, NLQGENEEEDSGAIVTVK from G2 subunit of Gly m 6, VFYLAGNPDIEYPETMQQQQQQK and YEGNWGPLVNPESQQGSPR from G4 subunit of Gly m 6, GAIGFAFPGCPETFEKPQQQSSR and VFYLAGNPDIEHPETMQQQQQQK from G5 subunit of Gly m 6 (the epitopes marked with an asterisk “*” in Table 1), which resulted in the decrease of IgE reactivity in BS and AS samples (Fig. 7A). Of note, although AS samples exhibited lower destruction of linear epitopes than BS samples (Table 1), AS samples showed a higher reduction of allergenicity than BS samples, probably due to the alteration of allergen content and conformational epitopes. What's more, boiling and autoclaving caused the decrease of protein solubility (Cabanillas et al., 2012; Kasera et al., 2012), resulting in the reduction of allergen contents in purified soluble protein (Fig. 5C). Therefore, AS samples showed a higher reduction of allergenicity than BS samples, resulting from a lower content of major soy allergens. Similar results were shown by Turner et al. (2014), who found that boiling caused the loss of allergenic proteins to decrease the allergenicity in peanut. Cabanillas et al. (2015) also demonstrated that the allergenicity of raw, fried, and roasted peanuts was decreased after autoclaving due to the degradation of protein, resulting in the reduction of content in peanut allergens such as Ara h 1, Ara h 2 and Ara h 6.

The changes of protein structure were reflected by the alternation in SEM (Fig. 2), FTIR (Fig. 3), UV absorption spectroscopy, surface hydrophobicity and free SH groups (Fig. 4). As demonstrated in Fig. 2, protein aggregation was observed in the BS and AS samples, contributing to the masking of conformational and linear epitopes. Based on the results of Table 1, it was compiled that AS samples exhibited a greater protein aggregation than BS samples, casing more masking of conformational epitopes and less masking of linear epitopes, which was responsible for a higher decrease of IgE reactivity in AS than BS samples (Fig. 7A). As for secondary structure, it was reported that secondary structural changes such as the decrease of α-helix caused the masking/destruction of epitopes, thus decreasing the allergenicity in the boiled and autoclaved rAra h 2.02 and peanut (Cai et al., 2016; Pi, Fu, et al., 2021). As shown in Fig. 3C, autoclaving caused more alteration of secondary structure than boiling, probably resulting in more masking/destruction of conformational epitopes to cause a higher reduction of IgE reactivity (Fig. 7A). As for tertiary structure, an increase of UV absorption spectroscopy (Fig. 4A) and SH groups (Fig. 4C) and the alteration of surface hydrophobicity (Fig. 4B) were observed, contributing to the masking/destruction of conformational epitopes responsible for the reduction of IgE reactivity in BS and AS samples. Similar results were found by Pi, Fu, et al. (2021), who found that the allergenicity of autoclaved peanuts was reduced by 66.6 %, which was accompanied by the increase of ultraviolet absorption spectroscopy and the reduction of surface hydrophobicity and free SH groups. The structural changes such as the decrease of α-helix and the increase of UV absorption spectroscopy and free SH groups were also responsible for the reduction of allergenicity in boiled ovotransferrin for 30 min (Tong et al., 2012). Of note, there was still IgE reactivity in BS and AS samples, probably resulting from the existence of linear epitopes recognized by IgE (Table 1).

3.11. Antioxidant activity

The antioxidant activity of protein is one of the most vital biological activities, which is usually reflected by DPPH• and ABTS •+ radical scavenging ability (Ren et al., 2025). As demonstrated in Fig. 7B, the DPPH• radical scavenging ability was increased from 0.92 % to 13.35 % and 15.22 %, after soy was boiled and autoclaved, respectively, whereas the ABTS •+ radical scavenging ability was improved from 60.24 % to 80.20 % and 82.43 %. Therefore, boiling and autoclaving caused an increase of antioxidant activity, probably resulting from the alteration of structure, particle size, surface hydrophobicity, etc. (Zou et al., 2019). Boiling and autoclaving caused protein degradation (Fig. 1A) to generate low molecular weight peptides likely interacted with free radicals (Dai et al., 2024; Hu & Li, 2022). For example, there was generation of antioxidant peptide such as GSEEEDEDEDEEQDERQFPFPRPPHQK and VEKEECEEGEIPRPRPRPQHPER, QFPFPRPPHQKEER from α subunit of Gly m 5 (P0DO16) in AS samples (Table 1), which was predicated using the PeptideRanker platform (http://distilldeep.ucd.ie/PeptideRanker) with a score of above 0.5 and The Biopep database (http://www.uwm.edu.pl/biochemia) (Pang et al., 2025). Boiling and autoclaving also caused the decrease of proteins particle size to increase their surface area, which might have contributed to raising in antioxidant proficiency (Hu & Li, 2022; Yeasmin et al., 2024). In addition, the exposure of a greater number of hydrophobic sites also could cause an augmentation of the electron donor capacity to increase antioxidant activity (Dai et al., 2024; Hu & Li, 2022; Yeasmin et al., 2024). The enhanced antioxidant capacity of BS and AS samples might also be due to protein structural changes, exposing hidden amino acid residues and side chains with antioxidant capacity (Hu & Li, 2022). Boiling and autoclaving further caused changes in protein composition (Fig. 5A and B), favoring the release of peptides that can interact with and donate electrons to DPPH• and ABTS •+ (Hu & Li, 2022). Furthermore, the conversion from the α-helix structure to β-sheet structure also contributed to increasing the antioxidant activity of protein (Ren et al., 2025), probably resulting from the exposed amino acid residues and side chains with antioxidant capacity. Therefore, boiling and autoclaving caused an increase of antioxidant activity due to the generation of low molecular weight peptides (Fig. 1A), the reduction of particle size (Fig. 1D), the alteration of protein structure (Fig. 3 and Fig. 4), and changes in protein composition (Fig. 5A and B). In addition, the AS samples exhibited a higher antioxidant activity than BS samples, probably resulting from a higher conversion from the α-helix to β-sheet (Fig. 3C) and a higher change in protein composition (Fig. 5B), which caused the exposure of amino acid residues and side chains with antioxidant capacity and the release of peptides that interacted with and donate electrons to DPPH• and ABTS•+.

3.12. Interfacial properties

3.12.1. Foaming properties

The foaming properties of proteins are defined as their ability to form flexible, elastic and cohesive interface films (Saran et al., 2024), which is usually reflected by FC and FS (Hu & Li, 2022). As demonstrated in Fig. 6C, FC was decreased from 32.00 % to 23.11 % and 15.56 %, after soy was boiled and autoclaved, respectively, whereas the FS was decreased from 40.09 % to 33.27 % and 13.33 %. Therefore, boiling and autoclaving caused the decrease of foaming properties, probably because protein-aggregation (Fig. 2) caused the difficult formations of an elastic foam network at the air-water interface (Pi et al., 2022; Yeasmin et al., 2024). Additionally, boiling and autoclaving caused the decrease of soluble protein content (Fig. 1B), suggesting the decrease of solubility, which made it difficult to quickly move to the gas-liquid interface and then decreased the foaming properties. Similar results were shown by Pi et al. (2022), who found that autoclaving of peanut caused the reduction of FC and FS in the extracted proteins. It was reported that the exposure of hydrophobic groups and the reduction of particle size might contribute to the increase of foaming properties (Hu & Li, 2022). Hydrophobic groups were exposed to increase the protein adsorption rate at air-water interfaces, thereby increasing FC. Besides, the exposure of hydrophobic groups can enhance the protein's ability to adsorb gas-liquid interfaces, reducing surface tension, and thereby improving FS (Pi et al., 2025). In addition, the decrease of protein particle size make it faster move to the gas-liquid interface, which shorts the formation time of foam and improve the FC (Pi et al., 2025). The molecules with lower particle size are closely arranged at the interface, forming a more uniform interface facial mask, delaying bubble coalescence, and improving FS. As a result, besides for a higher protein-aggregation and a lower solubility, the AS soy exhibited a higher reduction (51.38 % for FC, 71.08 % for FS) of foaming properties than BS samples (27.38 % for FC, 27.82 %for FS), resulting from a lower surface hydrophobicity (Fig. 4B) and a higher particle size (Fig. 1D), which decreased protein adsorption rate at air-water interfaces and formed uneven interface facial mask to accelerate bubble coalescence.

3.12.2. Emulsifying properties

Protein emulsifying properties refer to the ability of proteins to stabilize emulsions, which are reflected by EAI and ESI (Ren et al., 2025). As demonstrated in Fig. 7D, EAI was decreased from 2.15 m²/g to 1.67 m²/g and 1.08 m²/g, after soy was boiled and autoclaved, respectively, whereas the ESI was decreased from 42.30 min to 30.16 min and 20.78 min. Therefore, boiling and autoclaving caused the decrease of emulsifying properties, probably due to protein-aggregation (Fig. 2) (Hu & Li, 2022), which caused the low adsorption of soy protein at the oil-water interface (Dai et al., 2024). Additionally, boiling and autoclaving caused the decrease of soluble protein content (Fig. 1B), suggesting the decrease of solubility, which decrease the dispersion and adsorption at the interface to decrease emulsifying properties (Hu & Li, 2022). Similar results were shown by Pi et al. (2022), who found that autoclaving of peanut caused protein aggregation to decrease EAI and ESI in the extracted proteins. The AS sample exhibited a higher reduction (49.86 % for EAI, 50.87 % for ESI) of allergenicity than BS samples (22.33 % for EAI, 28.70 % for ESI), probably resulting from the more protein aggregation and less solubility in AS samples than BS samples. In addition, AS samples exhibited a higher particle size than BS samples (Fig. 1D), which was also not conducive to maintaining the stability of the oil–water interface for reducing ESI (Hu & Li, 2022). Furthermore, AS samples exhibited a lower surface hydrophobicity than BS samples (Fig. 4B), exposing less hydrophobic groups, which caused a lower surface interfacial tension and an increased proteins adsorption on oil droplet surfaces to decrease ESI (Yeasmin et al., 2024). Therefore, besides for a higher protein-aggregation and a lower solubility, the AS samples exhibited a higher reduction of emulsifying properties than BS samples, which was also attributed to a lower surface hydrophobicity (Fig. 4B) and a higher particle size (Fig. 1D).

3.13. Correlation analysis

Based on the above results, the IgE reactivity, antioxidant activity and interfacial properties varied from thermal sterilized treatment, which was influenced by several factors, including the physicochemical, structural and protein compositions. The correlations between the IgE reactivity, antioxidant activity or interfacial properties and these properties were further analyzed and shown in Fig. 8. The IgE reactivity showed a strong positive correlation with soluble protein content (1.00), particle size (0.51), α-helix (0.77), β-turn (0.76), random coil (0.86), surface hydrophobicity (0.58), total major soy allergen content (0.96), Gly m 4 content (0.88), Gly m 4 linear epitopes (0.88), Gly m 5 linear epitopes (0.71) and total linear epitopes (0.75), while exhibiting a negative correlation with β-sheet (−0.80) and UV (−0.67). Therefore, there was a reduction of the IgE reactivity in thermal sterilized soy seeds, resulting from the destruction/masking of linear epitopes (especially from Gly m 4, Gly m 5), the destruction/masking of conformational epitopes (mainly induced by the alteration of secondary structure, UV and surface hydrophobicity and the reduction of total major soy allergen content (especially Gly m 4). In addition, a lower soluble protein content contributed to a lower total major soy allergen content, thus decreasing the allergenicity. Similar results were shown by Pi, Sun, Liu, Wang, Hong, et al. (2023), who found a reduction of IgE reactivity in soy protein-proteinproanthocyanidins conjugates prepared by the alkali treatment, which was attributed to the reduction of major soy allergens and the masking or destruction of epitopes induced by structural changes. Antioxidant activity such as DPPH• and ABTS •+ radical scavenging ability also showed a strong positive correlation with β-sheet (0.53, 0.51), UV (0.89, 0.91), free SH (0.60, 0.62) while exhibiting a negative correlation with soluble protein content (−0.91, −0.90) and particle size (−0.78, −0.80), random coil (0.62, 0.60). Therefore, there was an increase of antioxidant activity in thermal sterilized soy seeds, probably because protein degradation caused the generation of low molecular weight peptides likely interacted with free radicals (Dai et al., 2024; Hu & Li, 2022), the reduced particle size increased in antioxidant proficiency (Hu & Li, 2022; Yeasmin et al., 2024), and structural changes (mainly β-sheet, UV, free SH) exposed hidden amino acid residues and side chains with antioxidant capacity (Hu & Li, 2022). Foaming properties such as FC and FS showed a strong positive correlation with soluble protein content (1.00, 1.00), α-helix (0.76, 0.86), β-turn (0.76, 0.86), random coil (0.85, 0.93), surface hydrophobicity (0.58, 0.71), while exhibiting a negative correlation with β-sheet (−0.79, −0.89) and UV (−0.68, −0.54). Therefore, there was a reduction of foaming properties in thermal sterilized soy seeds, probably because the decreased soluble protein content reduced the absorbed protein content in gas-liquid interfaces and structural changes caused the masking of hydrophobic groups to decrease the protein adsorption rate at air-water interfaces. Similar results were shown by Zhang et al. (2024), who found that the amount of adsorbed protein and protein structure was responsible for the alteration of foaming properties in pH-treated milk proteins based on Pearson's correlation analysis. Emulsifying properties such as EAI and ESI showed a strong positive correlation with soluble protein content (1.00, 0.99), α-helix (0.83, 0.84), β-turn (0.82, 0.74), random coil (0.91, 0.84), surface hydrophobicity (0.66, 0.55), while exhibiting a negative correlation with β-sheet (−0.85, −0.78), UV (−0.60, −0.70). Therefore, there was a reduction of foaming properties in thermal sterilized soy seeds, probably because the decreased soluble protein content caused the low adsorption of soy protein at the oil-water interface (Dai et al., 2024), and structural changes caused the masking of hydrophobic groups to cause a lower surface interfacial tension (Yeasmin et al., 2024).

Fig. 8.

Results of correlation analysis.

4. Conclusion

Soy contains excellent protein content, but is one of “Big Nine” allergenic foods seriously endangering human health. Industry sterilization techniques like boiling and autoclaving could cause protein degradation, aggregation, structural changes. Protein degradation caused the destruction of linear epitopes, whereas protein aggregation caused the masking of linear and conformational epitopes. The decrease of α-helix, increase of UV absorption spectroscopy and SH groups, and the alteration of surface hydrophobicity caused the masking of linear epitopes and the masking/destruction of conformational epitopes. As a result, boiling and autoclaving caused the reduction of IgE reactivity in soy. The autoclaved soy exhibited a higher reduction of allergenicity (IgE reactivity) than boiled soy due to a higher reduction of major allergenic proteins and a greater masking/destruction of conformational epitopes induced by structural changes. The autoclaved samples exhibited a higher DPPH• and ABTS •+ radical scavenging activity than boiled samples, probably resulting from a higher conversion from the α-helix to β-sheet and a higher change in protein composition. A higher reduction of interfacial properties such as foaming and emulsifying properties was observed in autoclaved soy than in boiled soy, resulting from a higher protein-aggregation and particle size, and a lower solubility and surface hydrophobicity. Therefore, autoclaving might be more promising and effective to produce low allergenic soy products with the desired functionality than boiling. Of note, in this study, the use of only five human serum samples for IgE reactivity is a limitation, and there are potential differences in IgE binding ability across individuals. Future research should expand the serum sample number and/or evaluate the IgE binding ability of each serum separately. Additionally, the allergenicity is only evaluated by the alteration of IgE reactivity based on ELISA, which is insufficient. Cell, animal allergy and clinical trials should be conducted to confirm the allergenicity alteration and hypo-allergenic properties in the future studies. The alteration of antioxidant activity and interfacial properties should also further be comprehensively measured by cell and rheological experiments, respectively. Moreover, future studies should also consider the alteration of nutritional composition and value.

CRediT authorship contribution statement

Jinshen Chu: Writing – original draft. Qingqing Cao: Methodology. Siyu Ren: Investigation. Zhenling Chen: Methodology. Xiaowen Pi: Writing – review & editing, Writing – original draft, Supervision, Project administration. Guohui Xue: Supervision.

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.

Appendix A. Supplementary data

Supplementary material.

Data availability

The data that has been used is confidential.