Ultrasound viscous reduction effects on the proteolysis of soy protein isolate at a limited degree of hydrolysis: Changes in the functional characteristics and protein structure

Highlights

• •Power ultrasound effectively reduces the viscosity of high-concentration substrate.
• •Positive correlations exist between the viscosities of the substrate and hydrolysate.
• •Ultrasonication helps enhance enzymatic hydrolysis beyond the limited DH.
• •Ultrasonication helps retaining high hydrophobicity of the hydrolysate at limited DH.

Abstract

High-concentration soy protein isolate was subjected to ultrasonication for viscosity reduction to assist the process of limited enzymatic hydrolysis. Ultrasonication (20 kHz, 10 min, 160 W/L) effectively reduced the viscosity of soy protein isolate at a comparatively high concentration of 14 % (w/v) and promoted the limited enzymatic hydrolysis (controlled degree of hydrolysis of 12 %) with a higher peptide yield than that of the conventional method. The correlations between substrate viscosity and peptide yield, as well as the viscosities of the resulting hydrolysates, were studied. The findings revealed positive correlations between the viscosities of the substrate and hydrolysate, underscoring the potential impact of altering substrate viscosity on the final product. Furthermore, the utilization of ultrasonic viscosity reduction-assisted proteolysis has shown its capability to improve the functional and physicochemical properties, as well as the protein structure of the hydrolysate, while maintaining the same level of hydrolysis. It is worth noting that there were significant alterations in particle size (decrease), β-sheet content (increase), β-turn content (increase), and random coil content (increase). Interestingly, ultrasonication unexpectedly impeded the degradation of molecular mass in proteins during proteolysis, while increasing the hydrophobic properties of the hydrolysate. These findings aligned with the observed reduction in bitterness and improvement in emulsifying properties and water-holding capacity.

1. Introduction

Soy protein hydrolysate is highly valued in the food and supplement industries for its numerous health benefits. This hydrolyzed form of soy protein is easily digested and absorbed by the human body, making it a significant source of essential amino acids [1]. Soy protein hydrolysate has gained popularity as a natural dietary supplement for individuals seeking to boost their protein intake without consuming excessive amounts of fat and calories. However, the preparation of soy protein hydrolysate can be challenging due to the highly viscous nature of soy protein isolate (SPI), which hinders the movement of enzymes and the binding of protein molecules to the enzyme's active sites, resulting in a lower hydrolysis rate and peptide yield [2], [3]. To address this issue, researchers have been exploring various physical and chemical processing methods to enhance the enzymatic hydrolysis of SPI [4], including the use of green novel technologies to assist the process of enzymatic hydrolysis [5].

Ultrasound technology is considered a groundbreaking advancement in the food processing industries, offering a means to produce safe and high-quality products at an affordable cost. This technology boasts numerous benefits, including increased efficiency, reduced expenses, higher yields, improved product quality, environmental sustainability, and ease of operation [6]. During the protein treatment, ultrasonication involves the use of acoustic cavitation to induce physical shear forces that enhance the kinetic energy within protein molecules, resulting in changes to protein functionality, such as reducing viscosity [7]. However, there is disagreement concerning the findings. While Quaisie et al. [3] found that ultrasonic treatment reduced viscosity following a decrease in surface charge, intermolecular interactions, and protein particle size, yet an increase in the mean molecule kinetic energies, Karki [8] stated that ultrasonication induced the unfolding of protein structures, leading to the exposure of hydrophilic pores. This phenomenon ultimately results in an increase in viscosity due to the strong attraction between water molecules and the hydrophilic surface. Therefore, to achieve optimal results, it is crucial to investigate the applicable sonication condition against viscosity reduction as the indicator.

Numerous studies have indicated that ultrasonication alters the permeability of cells, allowing enzymes to be secreted, resulting in the enhancement of the enzymatic hydrolysis process and leading to an increase in the degree of hydrolysis (DH) [3], [9], [10], [11], [12]. However, this increase in DH can potentially lead to unfavourable changes in flavour and functionality [12], [13]. As such, the primary objective of this study is to mitigate this issue by controlling the DH to preserve the desired functionality of the products. Overall, the main objective of this study is to fill the existing research gap regarding ultrasound viscosity reduction-assisted proteolysis of soy protein isolate at relatively high concentrations and its effect on the peptide yield, functional properties and protein structure of the hydrolysate prepared at controlled degree of hydrolysis which has not been elucidated in the literature. The findings of this study are expected to contribute valuable scientific information and encourage the further utilization of ultrasound in reducing viscosity and facilitating the enzymatic hydrolysis of dense and viscous substrates.

2. Materials and methods

2.1. Materials

Soy protein isolate (SPI) was provided by Yuwang Ecological Food Industry Co., Ltd. (Shandong, China). Food-grade bromelain was purchased from Pangbo Biological Engineering Co., Ltd. (Nanning, China). Other reagents and solvents used in this study were of analytical grade.

2.2. Ultrasound-assisted viscosity reduction

Cylindrical power ultrasound described in our previous work [13] was utilized to reduce the viscosity of SPI. Treatment was conducted under a stable 55 °C temperature. A single factor experiment at the range of five-level experimental conditions was conducted on viscosity. Under 20 kHz frequency mode with 10 min treatment time, the substrate-liquid ratio (S/L ratio, 10 %, 12 %, 14 %, 16 % and 18 %, w/v) was optimized, followed by further optimization on the US frequency (20, 28, 35, 40 and 50 kHz), treatment time (5, 10, 15, 20 30 min) and US power density (61.5, 80.0,133.3, 143.3 and 160 W/L).

2.3. Determination of viscosity

The dynamic viscosity of the fluid was measured using an LC-NDJ-1 T rotational viscometer (Lichen Instrument Technology Co., Ltd., Zhejiang, China). The rotating spindle was immersed in the supernatants of the slurry or hydrolysates at room temperature. The selection of the spindle number, rotation speed, time and shear rate are specified in Table 1.

Table 1.

The selection of rotational viscometer parameter.

Fluid type	Rotating spindle	Rotation speed (rpm)	Time	Shear rate (1/s)
SPI 10 % (w/v)	#2	30	1 min	6.24
SPI 12 % (w/v)	#3	12	2 min	2.49
SPI 14 % (w/v)	#4	12	2 min	2.49
SPI 16 % (w/v)	#4	6	3 min	1.24
SPI 18 % (w/v)	#4	6	3 min	1.24
SPIH	#0	60	40 sec	12.48

2.4. Enzymatic hydrolysis

Soybean protein isolate solution was subjected to a 55 °C preheated double-layered beaker after ultrasound pre-treatment. Proteolysis was carried out using bromelain under optimized hydrolysis conditions determined through preliminary experiments with peptide yield as the indicator. After adjusting the pH to 6.5, 15 % (w/w) bromelain was added to the mixture. During the proteolysis, the pH levels of the mixture were maintained constant by utilizing the ZD-2 automatic titration system (INESA Scientific Instrument Co., Ltd., Shanghai, China) and the 0.5 M NaOH solution. After reaching 12 % DH, the enzyme was inactivated by heating at 100 °C for 15 min. Subsequently, the terms of SPIH and US-SPIH were used to denote soybean protein hydrolysate prepared with conventional and US-pretreated methods, respectively.

2.5. Degree of hydrolysis (DH)

DH was estimated by the pH-stat technique outlined in our previous works [13], [14].

2.6. Determination of peptide yield

Peptide content was determined following the slightly modified protocol described by Yang et al. [15]. A 1:1 mixture of the hydrolysate’s supernatant and 10 % TCA was mixed and left to stand for 10 min at 25 °C. The mixture was centrifuged (10,000 × g, 10 min) and the supernatant was subjected to Lowry protein concentration analysis using bovine serum albumin as the standard protein. The peptide content was calculated according to the following equation:

$$\text{Peptide content}\left(\%\right)=\frac{\text{protein concentration}\times \text{total volume of the hydrolysate}}{\text{protein net mass}}\times 100\%$$ (1)

2.7. Flavour (bitter intensity)

Bitter intensity of the samples was collected from electronic tongue (e-tongue) data acquisition and sensory evaluation. E-tongue used bitterness analysis system with a total of 5 sensors, i.e., AAE for umami, AE1 for astringency, C00 for bitterness, CA0 for sourness, and CT0 for saltiness, coupled with three electrodes. The inner solution containing 30 mM KCl mixed with 0.3 mM C₄H₆O₆ was used for the reference electrode, 100 mM HCl with 30 % EtOH for negative electrode cleaning solution and 10 mM KOH, 100 mM KCl and 30 % EtOH for the positive electrode cleaning solution. Sensory evaluation of the bitter taste was performed following the method described in our previous work [14].

2.8. Emulsifying properties

The emulsifying activity index (EAI) and emulsion stability index (ESI) were measured following the method described in our previous work [13].

2.9. Water- and oil-holding capacities

Water holding capacities (WHC) and oil holding capacities (OHC) were analyzed following the method described in our previous work [13].

2.10. Molecular weight distribution

The molecular weight of the samples was determined using Size Exclusion Chromatography (SEC), following the methodology outlined in our previous works [13], [14].

2.11. Hydrophobic properties

Following the methodology outlined in our previous work [14], surface hydrophobicity (H₀) and relative hydrophobicity (RH) were determined by ANS method and Reverse-Phase High-Performance Liquid Chromatography (RP-HPLC), respectively.

2.12. Determination of particle size and zeta potential

Lyophilized samples were dispersed in distilled water. After a 10 min centrifugation at 12,000 × g, the protein concentration was determined using the biuret method, and the samples were diluted to a protein concentration of 1 mg/mL. A disposable cuvette with a refractive index ranging from 20 to 40 % was used to measure the particle size (Litesizer 500 Particle Size Analyzer, Anton Paar GmbH, Austria) of the samples recorded at a solvent refractive index of 1.33, viscosity of 8.903 × 10–3 Pa.s, material refractive index of 1.45, and absorption coefficient of 0.001 m⁻¹.

2.13. Fourier transform infrared (FTIR) spectroscopy

Samples powder was subjected to FTIR spectrum instrument (Nicolet iS50 FTIR Spectrometer, Thermo Fisher Scientific Inc., USA) at 4000–600 cm⁻¹ scanning wavelengths. 36 scans were made with air as the blank. Data analysis was performed using OMNIC 8 (Thermo Fisher Scientific Inc., Canada) and Peakfit 4.12 (Systat Software Inc., USA) softwares. Second-derivative analysis of the IR-SD was performed on the decomposition of amide I band to obtain quantitative analysis of the secondary structure. Peak fitting was performed with linear baseline correction, and Fourier self-deconvolution was performed using a Gaussian curve fit [16], [17].

2.14. Scanning electron microscopy (SEM) measurement

The morphological structures of the samples were obtained by utilizing a scanning electron microscope (Zeiss Sigma 300, Germany) with an acceleration voltage of 15 kV [7]. The samples were smeared on a double-sided carbon taped on a copper sample-holder and coated with a thin layer of gold (±10 nm).

2.15. Statistical analysis

Three independent experiments (mean ± standard deviation) were conducted and one-way ANOVA at 0.05 level was performed to evaluate the statistical significance between groups using SPSS 24.0 software (SPSS Inc., Chicago, IL, USA). Graphs were generated using the advanced software tools, Origin 2023 software (OriginLab Corporation, Massachusetts, USA) and GraphPad Prism 8 software (Dotmatics Ltd., San Diego, California).

3. Results and discussion

3.1. Ultrasound-assisted viscosity reduction

3.1.1. Effect of substrate-liquid ratio on the viscosity of US pre-treated SPI

Viscosity is an important functional property of food which plays a major role in affecting food processing and is of practical interest in all types of beverages. Food materials with high protein content were reported to be high in viscosity, making it the main limiting factor in the utilization of these products in food processing, including enzymatic hydrolysis [18]. The soy protein isolate used in this study had an 86.202 ± 0.726 % high protein content on a wet basis, and its dynamic viscosity correspondingly showed a relatively high resistance to flow, limiting its utilization in food processing, especially at higher concentrations. Therefore, novel technologies have been proposed for improving the functional properties of proteins.

Considering that viscosity is dependent on concentration, the effect of US pre-treatment on SPI dynamic viscosity is presented at varying concentration, i.e., 10, 12, 14, 16, 18 % (w/v). Under 20 kHz US frequency with 10 min treatment time, the substrate-liquid ratio was screened to find an applicable solution concentration for US-assisted viscosity reduction. As shown in Fig. 1, SPI dynamic viscosity increased as its concentration increased for both untreated and treated samples. Notably, it can be observed that US treatment exhibited significant changes (p < 0.05) on SPI dynamic viscosity at the concentration up to 14 % (w/v), where increases were found at the concentrations of 10 % and12 %, yet decrease was observed at the concentration of 14 %. Besides, non-significant increase (p > 0.05) was found at the concentration of 16 % and no change was found when the concentration reached 18 %. These outcomes align with previous studies [3], [8] where ultrasonication could both increase and reduce viscosity. Remarkably, the ultrasonication effect on viscosity was also found to be dependent on concentration, where the concentration of 14 % was considered the optimal concentration for US-assisted viscosity reduction of SPI. The significant increase in viscosity found at lower concentration could be due to the formation of more aggregates induced during sonication at lower concentration, while a non-significant change was found at higher concentration slurry as it comprises higher viscosity that resist the ultrasound transmission through the aqueous medium [19], [20].

Fig. 1.

Effect of US treatment on SPI dynamic viscosity profile at varying concentration. Groups with * are significantly different within the group, i.e., control and US treated (p < 0.05).

3.1.2. Effect of ultrasonic frequency on SPI viscosity

The effect of different ultrasonic frequencies (20, 28, 35, 40, and 50 kHz) on the dynamic viscosity of SPI solution was investigated. Analysis of the data presented in Table 2A revealed no significant difference in viscosity profiles among the samples. However, a trend was observed where increasing the ultrasonic frequency tended to increase the dynamic viscosity of SPI. A statistically significant difference (p < 0.05) was noted between the viscosity of SPI treated with the lowest frequency of 20 kHz and the highest frequency of 50 kHz. Compared to the control group, a non-significant decrease in viscosity (p > 0.05) was observed when SPI was treated with 20 kHz ultrasonic frequency. Slight decreases to no change (p > 0.05) were observed in the 25–35 kHz treatment group, while slight increases (p > 0.05) were seen in the 40–50 kHz treatment group. Although the result was in contrast to the research undertaken by Arzeni et al. [21] where an increase in viscosity was found when SPI was treated with 20 kHz US, the study conducted by Zisu et al. [22] supported the potential of 20 kHz US frequency in reducing viscosity. According to previous studies [20], [23], [24], low-frequency ultrasonic waves could penetrate high viscous products more effectively than high-frequency ultrasonic waves. This is due to the dissipation of viscous energy within the liquid, causing damping of acoustic waves and reducing the level of ultrasonic energy received. According to these results, 20 kHz US frequency was adopted as the most applicable US frequency for reducing the viscosity of SPI.

Table 2.

Effect of US frequency, treatment time and power density on SPI dynamic viscosity profile and SPIH peptide yield and viscosity. Left Y axis is presented in a reverse direction. Means that do not share a letter or * in the same color are significantly different (p < 0.05).

SPI dynamic viscosity profile	SPIH peptide yield and viscosity

3.1.3. Effect of ultrasonic treatment time on SPI viscosity

The effect of US treatment time (5, 10, 15, 20 30 min) on SPI viscosity is presented in Table 2B. The data reveals a decrease in dynamic viscosity of SPI with increasing ultrasonication treatment time up to 10 min. However, these differences were not deemed statistically significant (p > 0.05) when compared to the control group. Significantly, increases in viscosity were found as the US treatment time further increased (p < 0.05). This phenomenon may be attributed to various mechanisms associated with sonication, including acoustic waves, mass transportation, and heat changes during the treatment, as implied by previous studies [23], [25], [26], [27]. As treatment time increased, more heat was generated in the sample, which resulted to the inevitable dissipating in vapor. According to the results, a treatment time of 10 min was acceded as the most applicable US treatment time for reducing SPI viscosity.

3.1.4. Effect of ultrasonic power density on viscosity

Table 2C illustrates the impact of different power densities on the change in SPI dynamic viscosity. The results indicate a decrease in SPI viscosity as the ultrasonic power density increases. Compared to the control, a significant reduction (p < 0.05) was found when SPI was exposed to the highest power density of 160 W/L. Since the decrease in viscosity was reported to be related to the cavitation effect, which causes molecular unfolding of proteins, resulting in a more ordered polymer along the flow field [28], Bernardi et al. [20] suggested that increasing the US amplitude is necessary for high-viscosity samples that show resistance to ultrasound cavitation, in order to promote sample cavitation and generate mechanical vibrations. Henceforth, it could be concluded that a low frequency (20 kHz), high intensity (100 % amplitude, 160 W/L power density) US treatment for 10 min was identified as the most effective parameters for reducing SPI viscosity at a concentration of 14 % (w/v).

3.2. Ultrasonic viscosity reduction-assisted proteolysis

3.2.1. Effect of ultrasonic frequency on peptide yield and hydrolysate’s viscosity

A 14 % (w/v) SPI solution was pre-treated with US at varying frequencies of 20, 28, 35, 40 and 50 kHz, followed by enzymatic hydrolysis with bromelain to reach a desired DH of 12 %. The impact of these treatments on peptide yield and the dynamic viscosity of the hydrolysate is detailed in Table 2(a). The results indicate a notable reduction in viscosity (12.05 %, p < 0.05) in hydrolysates pre-treated with a 20 kHz ultrasound frequency. Conversely, no significant effects were observed at higher ultrasound frequencies (p > 0.05). On the other hand, peptide yield analysis revealed a significant increase in all US-pretreated hydrolysates (p < 0.05) compared to the control, with no significant differences among the treated samples (p > 0.05). Notably, it can be observed that treatment with low frequency resulted in higher peptide yields, with the 20 kHz US treatment yielding the highest peptide yield of 77.608 ± 1.010 %, which was 8.549 %, 2.873 %, 2.471 %, 3.187 % and 2.727 % higher than the control, 28, 35, 40 and 50 kHz, respectively. According to these results, the 20 kHz US frequency was deemed the most applicable US frequency for US-assisted proteolysis of SPI. This result aligns with the results presented in section 3.1.2.

To study the effect of the original viscosity of the substrate prior to hydrolysis on peptide yield and the dynamic viscosity of the hydrolysate, a two-tailed bivariate correlation analysis between the variables was performed. The Pearson correlation analysis between the viscosity of the substrate and the hydrolysates revealed a non-significant moderate positive correlation (R² of 0.504, p > 0.05). Unfortunately, short of evidence could be found to explain this finding. Meanwhile, the correlation between substrate viscosity and peptide yield, prepared at various US frequencies, exhibited a non-significant moderate negative correlation (R² of −0.533, p > 0.05). Previous studies by Kanosue, Kojima and Ohkata [29] and Quaisie et al. [3] reported similar results indicating a negative correlation between substrate viscosity and enzymatic reaction rate. It was hypothesized that the observed phenomenon could be attributed to the restricted movement of enzymes in highly viscous solutions.

3.2.2. Effect of ultrasonic pre-treatment time on peptide yield and hydrolysate’s viscosity

As shown in Table 2(b), US pre-treatment showed to have a significant increase in peptide yield (p < 0.05), with the increase generally improving as the length of US pre-treatment time increased. The lowest increase of 8.286 % was found in the 5-minute pre-treatment sample, where the peptide yield reached 77.420 ± 0.952 % (compared to the control). Conversely, the highest increase of 12.172 % was found in the 30-minute pre-treatment sample, with a peptide yield of 80.199 ± 0.491 %. However, the means between the pre-treated samples did not show to exhibit any significant change (p > 0.05). The result of the hydrolysates’ dynamic viscosity showed a decrease up to 10 min of pre-treatment time, followed by a sudden increase as the pre-treatment time further increased up to 15 min (p < 0.05). The change in hydrolysates’ dynamic viscosity among the untreated and pre-treated samples did not exhibit any significant difference between the means (p > 0.05), except for the 10-minute pre-treatment time (p < 0.05), where the highest decrease of 12.049 % compared to the control was found. Although these results align with the previous findings in section 3.1.3, the correlation analysis between substrate viscosity and proteolysis results did not reveal any significant correlation (p > 0.05). Moderate positive correlations with R² values of 0.522 and 0.518 were found for SPI viscosity – peptide yield and SPI viscosity – SPIH viscosity, respectively. Interestingly, there was a correlation change between substrate viscosity and peptide yield. Remarkably, following the change in pre-treatment time, the change in substrate viscosity exhibited a positive correlation with the change in peptide yield, which contrasts with the previous result reported in section 3.2.1. Regrettably, there has been a lack of research conducted on this particular issue. Therefore, it is recommended that further study be pursued in order to gain a better understanding of this issue.

3.2.3. Effect of ultrasonic power density on peptide yield and hydrolysate’s viscosity

As presented in Table 2(c), the viscosity of the hydrolysate exhibited a slight increase at low power density of 61.5 W/L with an increase rate of 0.52 % (p > 0.05) compared to the control. A sudden decrease in viscosity was observed at a low-medium power density of 80.0 W/L, with a decrease rate of 9.17 % (p > 0.05) compared to the control. Further increase in power density up to 133.3 W/L resulted in a 7.39 % increase in viscosity (p > 0.05), although the viscosity remained 2.46 % lower than the control sample (p > 0.05). Following this, viscosity showed a gradual decrease with further increases in power density, and a significant alteration (p < 0.05) was found when samples were subjected to the highest US power density (160 W/L), resulting in the lowest viscosity of 1.47 cp (decreased by 12.05 %, compared to the control). Additionally, peptide yield increased gradually with the rise in power density, with increase rate ranging from 4.95 % to 8.55 %. Notably, significant increases in peptide yield were observed when the power density reached 80 W/L and above (p < 0.05). These results align with the result in section 3.1.4, and according to the Pearson correlation analysis, substrate viscosity formed a strong positive correlation with hydrolysates viscosity (R² of 0.757, p > 0.05). Similar to the result in section 3.2.1, SPI viscosity formed a high negative correlation with peptide yield, with R² value of −0.773 (p > 0.05).

3.3. Effect of ultrasonic viscosity reduction-assisted proteolysis on the functional properties

3.3.1. Flavour (bitter intensity)

E-tongue analysis results (Fig. 2A) showed that SPI originally exhibited a high bitter intensity of 9.287 ± 0.160, yet limited hydrolysis by bromelain at 12 % DH successfully decreased (p < 0.05) the bitterness of SPI by 26.43 % (SPIH) and 25.65 % (US-SPIH). Similar observations were reported by Saha and Hayashi [30], who noted that the Q value of soybean protein was higher than the average Q value of bitter peptides. In addition to reducing bitterness, enzymatic hydrolysis also improved the umami flavor, richness as well as the saltiness of SPI, with improvements reached 7.38 %, 66.14 % and 138.01 %, respectively (p < 0.05). However, hydrolysis of SPI resulted in a higher aftertaste of bitterness and astringency, reaching the highest value of 3.673 and 4.510, respectively (p < 0.05). Notably, all samples did not exhibit sourness nor astringency, as the values were lower than the thresholds for these taste attributes, which are −13 and 0, respectively. Based on the PCA (Principal component analysis) results presented in Fig. 2B, the enzymatic hydrolysis of SPI is associated with the increase in the umami flavor, richness, saltiness, aftertaste-A, aftertaste-B, while simultaneously resulting in a decrease in bitterness. Consistent with the PCA analysis results, the Pearson correlation analysis (presented in Fig. 2C) revealed significant relationships among the taste variables (p < 0.05). Notably, bitterness was found to have a positive correlation with sourness and astringency, while showing a negative correlation with umami flavor, richness, saltiness, aftertaste-A, and aftertaste-B.

Fig. 2.

Radar graph of e-tongue (A), principal component analysis (B), Pearson correlation between the taste variables (C) and the compilation of bitter intensity of SPI, SPIH and US-SPIH (D). Means that do not share a letter or * in the same color are significantly different at p < 0.05 and ** at p < 0.01.

Due to the reported limitations of e-tongue in accurately replicating the intricate human taste perception, a sensory analysis was conducted by trained panellists to further explore the bitter intensity of the samples (Fig. 2D). The results of the sensory analysis revealed that SPI exhibited the lowest level of bitterness, with no significant difference compared to US-SPIH (p > 0.05). Conversely, SPIH was found to be the most bitter among the samples (p > 0.05). In contrast to the e-tongue analysis results, these results suggest that utilizing ultrasound-assisted viscosity reduction in conjunction with limited proteolysis has the potential to decrease the bitterness of SPIH. This aligns with the research conducted by Huang et al. [31]. The disparities between the e-tongue and sensory analysis results could be attributed to the differing mechanisms employed by each analysis technique. The e-tongue mainly relies on the physicochemical indicators [32] which may not effectively capture the holistic sensory experience.

3.3.2. Emulsifying properties

The emulsifying properties of SPI, SPIH and US-SPIH were observed and presented as emulsifying activity (EAI) and stability indexes (ESI) at varying pH levels in Table 3. Consistent with the literatures [33] EAI and ESI of all samples were significantly influenced by pH. SPI exhibited the lowest EAI of 1.45 ± 0.51 near its isoelectric point at pH 5, while the highest EAI of 14.59 ± 0.29 was found at pH 11, consistent with O’Flynn et al. [34]. On the other hand, the EAI of the hydrolysates increased with rising pH levels. Under extreme acidic and basic conditions, SPIH exhibited EAI lower than that of SPI, with reductions up to 57.56 % and 12.05 % at pH 3 and pH 11, respectively. However, the use of US-assisted hydrolysis led to a significant improvement in EAI compared to the conventional method (SPIH), with increases of 58.72 % and 48.25 % at pH 3 and pH 11, respectively (both p < 0.05). Nevertheless, US-SPIH still displayed lower EAI values than SPI. The increase in EAI by US-assisted hydrolysis may be attributed to the alterations in hydrophobicity. Recent studies [35], [36] suggested a strong positive correlation between emulsifying activities and hydrophobicity, although the results were not specifically described for different pH levels.

Table 3.

Emulsifying properties and secondary structure of SPI, SPIH and US-SPIH. Means with different superscripts are significantly different (p < 0.05).

pH	SPI	SPIH	US-SPIH
Emulsifying activity index (EAI, m2/g)
3	9.19 ± 0.57 fg	3.90 ± 0.05i	6.19 ± 0.47 h
5	1.45 ± 0.51j	9.44 ± 0.68 fg	8.23 ± 0.02 g
7	9.16 ± 0.02 fg	10.35 ± 0.97ef	12.20 ± 0.05 cd
9	10.73 ± 0.48def	11.84 ± 0.36cde	13.55 ± 0.50bc
11	14.59 ± 0.29b	12.83 ± 0.41bc	19.02 ± 0.25a
Emulsifying stability index (ESI, min)
3	10.24 ± 0.03f	10.32 ± 0.09f	10.16 ± 0.06f
5	12.53 ± 0.17de	12.85 ± 0.40 cd	10.77 ± 0.22ef
7	11.69 ± 0.21def	17.71 ± 0.98b	25.78 ± 0.97a
9	12.18 ± 0.08de	18.59 ± 0.58b	24.53 ± 0.85a
11	14.61 ± 0.37c	18.84 ± 1.87b	19.07 ± 0.84b
Secondary structure
α-helix	18.52 ± 0.78*	16.13 ± 0.35	16.26 ± 0.22
β-sheet	44.52 ± 1.23a	29.36 ± 0.90b	37.18 ± 1.00c
β-turn	17.16 ± 0.42a	26.28 ± 0.20b	28.94 ± 0.88c
Random coil	16.06 ± 0.67	16.81 ± 0.44	22.18 ± 0.78*

The emulsifying stability of SPI showed to exhibit a slight increase following the increase of pH levels, while remaining generally stable near neutral pH. In contrast, the ESI of the hydrolysates (both SPIH and US-SPIH) had gradual increases under acidic conditions and sudden increases at neutral pH, which then stabilized with further increases of pH. Generally, all samples exhibited comparable ESI (p > 0.05) at acidic conditions (pH 3–5). At pH 7–9, conventional enzymatic hydrolysis greatly improved ESI by 51.50 % and 52.63 % respectively, compared to that of SPI (both p < 0.05). Meanwhile, at the same pH range, US viscosity reduction-assisted proteolysis further improved ESI by 45.57 % and 31.95 % respectively, compared to that of SPIH (both p < 0.05). At extreme basic condition (pH 11), conventional enzymatic hydrolysis greatly improved ESI by 28.95 % compared to that of SPI (p < 0.05), meanwhile further slight increase was found by US viscosity reduction-assisted proteolysis, although it was not statistically significant compared to SPIH (1.22 %, p > 0.05). The results indicated that enzymatic hydrolysis plays a crucial role in enhancing EAI and ESI of SPI, with US viscosity reduction-assisted proteolysis offering further improvements in both parameters. However, it is important to notice that the improvement was dependent to pH levels. These results align with existing literature [37], [38], [39].

3.3.3. Water holding capacity (WHC) and oil holding capacity (OHC)

Previous studies [19], [40] have highlighted the importance of WHC and OHC in determining the quality and appeal of various food product, especially for enhancing the textural properties and preserving the flavor during food processing. As presented in Fig. 3A, SPI had a high WHC of 4.44 g_water/g_protein but a low OHC of 1.59 g_oil/g_protein, which was lower than values reported in the literatures [41], [42]. However, after hydrolysis with bromelain at 12 % DH, SPIH showed a decrease in WHC to 1.03 g_water/g_protein, but an increase in OHC to 3.55 g_oil/g_protein. These results also contradicted the previous studies [42], [43] that suggested an improvement in both WHC and OHC after enzymatic hydrolysis, however both studies also enlightened that the increase was affected with DH which might explain the difference found in this study. Interestingly, our study found that US viscosity reduction-assisted proteolysis showed led to a significant increase in WHC by 16.60 % (p < 0.05) with a non-significant decrease in OHC by 4.81 % (p > 0.05). This result also contradicted previous research [19], [44], [45], [46] which reported a decrease in WHC due to the exposed hydrophobic groups by sonication. The enhancement of WHC and OHC despite any changes in hydrophobicity is indeed a promising development by the utilization of ultrasonication. However, further research is necessary to fully comprehend this phenomenon.

Fig. 3.

Water- and oil-holding capacities (A), particle size (B) and zeta potential (C) as well as particle size distribution (D) of SPI, SPIH and US-SPIH.

3.4. Effect of ultrasonic viscosity reduction-assisted proteolysis on protein structure

3.4.1. Molecular weight distribution

As presented in Fig. 4A, the large molecules present in SPI have undergone degradation, resulting in the formation of smaller molecules through hydrolysis. In the SPIH samples, the majority of molecules ranged from 1 to 3 kDa. It is worth noting that the total of small molecules below 3 kDa accounted for 83.24 %, which aligns with the reported range for most bioactive peptides [47]. Interestingly, under the same hydrolysis conditions and DH, the decrease in viscosity caused by ultrasound unexpectedly impeded the degradation of protein molecular mass, particularly for molecules smaller than 3 kDa. This was evident from the observed increase in larger molecular weight (>3 kDa) and decrease in smaller molecular weight (<3 kDa), which could be attributed to the reduced hydrolysis time required to achieve the same DH after ultrasonication. However, this finding contradicts the results reported by Liang et al. [48], where an increase in smaller MW (<3 kDa) and a decrease in larger MW (>3 kDa) were observed, corresponding to an increase in DH. Notably, both studies demonstrated that ultrasonication had the most significant impact on protein breakdown at the 3 kDa crossover point.

Fig. 4.

SEC spectra (A), surface hydrophobicity (B) and relative hydrophobicity distribution (C) of SPI, SPIH and US-SPIH.

In the US pre-treated sample group, a noteworthy increase of 35.14 % (p < 0.05) was observed in the larger molecule fractions ranging from 6.4 to 10 kDa. Conversely, a significant decrease of 3.91 % (p < 0.05) was noted in the smaller molecule fractions ranging from 1 to 3 kDa, compared to the control group. The unexpected hindrance of protein molecular mass degradation through ultrasound-assisted limited proteolysis suggests that ultrasound treatment may develop a protective effect on the protein structure, preventing excessive breakdown during the proteolysis. Furthermore, these findings provide an explanation for the reduced bitterness observed in the US-SPIH samples.

3.4.2. Hydrophobic properties

The surface hydrophobicity (H₀) of SPI was measured at 953.995 ± 27.937. The H₀ of its hydrolysates proteolyzed by bromelain at 12 % DH using conventional and US pre-treatment methods is presented in Fig. 4B. The use of ultrasonic viscosity reduction-assisted proteolysis resulted in a significant increase in the H₀ of SPIH prepared at the same DH level (p < 0.05). Specifically, the H₀ of US-SPIH was 2.60 times higher compared to the hydrolysate prepared using the conventional method (H₀ of 9.399 ± 0.856). Meanwhile, as shown in Fig. 4C, the RH of US-SPIH exhibited a slightly higher content of high and medium hydrophobic properties compared to the hydrolysate prepared by conventional method, with increases of 0.230 % and 20.146 %, respectively. These results align with the results reported by previous studies [15], [49], [50]. The results suggest that the improved functional properties observed in the hydrolysate may be attributed to the ultrasonication treatment prior to hydrolysis. This technique has been reported to enhance the H₀ by leveraging its cavitation effect, which can lead to the alteration or loss of certain hydrophilic groups [51]. This process stretches the protein molecule and exposes hydrophobic interaction sites that were previously concealed within the molecule [9], [15]. However, further research is needed to fully understand the impact of this process on RH.

3.4.3. Particle size and zeta potential

The alteration of protein function is directly linked to its molecular changes, primarily influenced by hydrophobicity and particle size [52]. To assess the particle size and distribution profiles of SPI, SPIH and US-SPIH, a dynamic light scattering technique was employed, with the results illustrated in Fig. 3B and D, respectively. SPI and its hydrolysates were shown to exhibit double peaks (Fig. 3D). After hydrolysis, the particle size distribution shifted forward. Conventional hydrolysis separated the peaks that were originally attached together. However, US viscosity reduction-assisted proteolysis reattached the peaks and shifted the particle size distribution backward towards the substrate particle size distribution, which indicates that US-SPIH had larger aggregates than the hydrolysate prepared by the conventional method. As presented in Fig. 3B, it can be observed that conventional enzymatic hydrolysis of SPI increased the particle diameter by 3.10 %, compared to the substrate which particle diameter was 226.485 nm (p > 0.05). Meanwhile, US viscosity reduction-assisted proteolysis decreased the particle diameter by 7.04 % compared to the substrate and 9.84 % compared to the hydrolysate prepared using the conventional method at the same DH (both p < 0.05). According to Vanhatalo and Dahl [53], the intensity of the hydrolysis treatment significantly impacts particle size, with mild enzymatic hydrolysis resulting in a wider range of particle sizes compared to the intensive treatment. This explains the increase in the particle diameter in SPIH observed in this study compared to the substrate molecule. Despite all hydrolysates in this study being controlled at 12 % DH, the particle diameter of US-SPIH was smaller compared to that of the conventional method and untreated substrate, aligning with the findings of Jambrak et al. [54] and Quaisie [7].

On the contrary, the zeta potential is often associated with the stability of protein solutions. Higher absolute zeta potential levels lead to increased electrostatic repulsion, improved particle separation, and enhanced solution stability. Understanding the relationship between zeta potential and a particle's surface charge is essential as it influences the aggregation and dispersion behaviour of the particle [55]. As shown in Fig. 3C, zeta potential exhibited a lower negative value after hydrolysis (p > 0.05) and the use of US viscosity reduction-assisted proteolysis technique resulted in a further decrease in the zeta potential of the hydrolysate (p > 0.05). These findings align with previous research conducted by Wang et al. [56] and Quaisie [7]. The results show that hydrolysis by bromelain led to a reduction in the electrostatic repulsion between droplets in SPI. The observed phenomenon could potentially be attributed to a decrease in electrostatic or spatial repulsion resulting from the aggregation and/or recombination of proteins [56].

3.4.4. Secondary structure

FTIR spectra were utilized to investigate the structural modifications that occurred in the functional groups of the protein molecules. Protein peptide bond consists of a group of amides [57], namely amide I (1600–1800 1/cm), amide II (1470–1570 1/cm), amide III (1250–1350 1/cm) and amide A (3300–3500 1/cm). As presented in Fig. 5A, shifts in the position of the spectral peaks were observed after hydrolysis, yet US viscosity reduction-assisted proteolysis method did not show a shift in the position of the spectral peaks compared to the conventional method. Enzymatic hydrolysis caused the amide I peak to shift from 1632.93 cm⁻¹ (SPI) to 1640.64 cm⁻¹ (both SPIH and US-SPIH), the amide II peak to shift from 1516.74 cm⁻¹ (SPI) to 1530.72 cm⁻¹ (both SPIH and US-SPIH), and the amide III peak to shift from 1393.32 cm⁻¹ (SPI) to 1394.76 cm⁻¹ (both SPIH and US-SPIH). Meanwhile, all samples did not exhibit a peak of amide A. The alteration in the in the location of the amide bands explains the changes in the secondary structure of SPI to SPIH. This result contrasts with the findings reported in the study by Musa et al. [58] and Quaisie [7] in which the US-pretreated hydrolysis method resulted in shifts of the spectral peaks compared to the conventional method. This difference may be attributed to the controlled DH during the hydrolysis in this study, as well as the differences in the sample source.

Fig. 5.

FTIR spectra (A) and the second-derivative spectra (B) of SPI, SPIH and US-SPIH. Continuous curve reflects the deconvolution of amide I and point lines represents the Gaussian curve fit bands.

The secondary structure of the samples was quantitatively analyzed by decomposing the amide I band according to Yang et al. [16], which involved identifying the presence of α-helix (1650–1660 cm⁻¹), β-turn (1660–1680 cm⁻¹), β-sheet (1610–1638 cm⁻¹ and 1680–1690 cm⁻¹) and random coil (1640–1650 cm⁻¹). As illustrated in Table 3 and Fig. 5B, α-helix and β-sheet contents decreased after enzymatic hydrolysis (p < 0.05), while the content of β-turn and random coil significantly increased (p < 0.05). Compared to the conventional method, the US viscosity reduction-assisted proteolysis method resulted in a higher β-sheet, β-turn and random coil (p < 0.05). However, there was a non-significant slight increase observed in the α-helix content (p > 0.05). The changes in the secondary structure after enzymatic hydrolysis aligned with the findings of Wang et al. [56]. However, the alterations observed in the US-SPIH, compared to the SPIH prepared using the conventional method, did not align with the results reported by Wen et al. [59] and Quaisie [7]. This discrepancy may be attributed to the controlled DH during hydrolysis, as well as the differences in the sample source and US equipment.

3.4.5. Scanning electron microscopy (SEM)

The changes in morphology resulting from enzymatic hydrolysis and US-assisted proteolysis of SPI were examined using SEM, with the findings presented in Fig. 6. Upon enzymatic hydrolysis, SPIH molecules formed larger aggregates, thus larger particles size was formed compared to the original substrate. Conversely, US viscosity reduction-assisted proteolysis led to smaller aggregates and a greater reduction in particles sizes compared to the conventional method. Moreover, it is important to notice that SPIH relatively formed a tight structure and a plate-like surface. Similar to the findings of Tian et al. [60], enzymatic hydrolysis dissociated the chains of protein molecules to yield small peptides, yet failed to deteriorate its protein microstructure. US-SPIH, on the other hand, displayed more cracked surfaces, forming porous, spherical aggregates with smaller loose fragments. These results may be attributed to the pre-alteration of substrate molecules by US-pretreatment, which unfolded proteins and increased protein surface area. Research by Yang et al. [15] further supports these findings, indicating that sonication led to a more loosened protein structure, different degrees of deformation, and small cavities.

Fig. 6.

The scanning electron micrographs of SPI (a), SPIH (b) and US-SPIH (c) at 500 μm (left) and 20 μm (right).

3.4.6. Stepwise multiple regression analysis between viscosity and protein structure properties

To analyze the correlation between protein structural changes and viscosity, a stepwise multiple regression analysis method was employed to identify and eliminate independent variables. Subsequently, collinear independent variables were eliminated, and the model was tested using an F test to obtain the regression equation. Viscosity was assigned as the dependent variable, while molecular weight, surface hydrophobicity, relative hydrophobicity, particle size, zeta potential, α-helix, β-sheet, β-turn, and random coil were assigned as the independent variables. As presented in Table 4.1, the variables entering the regression model through stepwise regression for viscosity were surface hydrophobicity and β-sheet values, which fitted two significantly correlated regression models (both p < 0.01, presented in Table 4.2). As presented in Table 4.3, the resulting multiple regression model equation is y = 1454.389 + 30.663x₁ − 59.265x₂ (R² value of 1.000), with y, x₁, x₂ signify viscosity, surface hydrophobicity, and β-sheet content, respectively.

Table 4.1.

Entered Variables.

Target	Model	Variables Entered	Method
Viscosity	1	H0	Stepwise (Criteria: Probability-of-F-to-enter <= 0.050, Probability-of-F-to-remove >= 0.100).
	2	β-sheet	Stepwise (Criteria: Probability-of-F-to-enter <= 0.050, Probability-of-F-to-remove >= 0.100).

Table 4.2.

The results of analysis of variance.

Target	Model		Sum of Squares	df	Mean Square	F	Sig.
Viscosity	1	Regression	1.58E + 09	1	1.58E + 09	2.73E + 04	.000a
		Residual	4.03E + 05	7	5.76E + 04
		Total	1.58E + 09	8
	2	Regression	1.58E + 09	2	7.88E + 08	4.94E + 04	.000b
		Residual	9.57E + 04	6	1.59E + 04
		Total	1.58E + 09	8

^aPredictors: (Constant), H0

^bPredictors: (Constant), H0, β-sheet

Table 4.3.

Coefficient of regression equation.

Target	Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	95.0 % Confidence Interval for B
			B	Std. Error	Beta			Lower Bound	Upper Bound
Viscosity	1	(Constant)	−503.659	99.792		−5.047	0.001	−739.630	−267.688
		H0	29.946	0.181	1.000	165.371	0.000	29.518	30.375
	2	(Constant)	1454.389	448.737		3.241	0.018	356.370	2552.408
		H0	30.663	0.189	1.024	162.267	0.000	30.201	31.126
		β-sheet	−59.265	13.489	−0.028	−4.394	0.005	−92.271	−26.259

4. Conclusions

Ultrasonication effectively reduced the viscosity of soy protein isolate at a relatively high concentration of 14 % (w/v) with the advantage of low frequency ultrasound treatment (20 kHz) for a duration of 10 min, with a power density of 160 W/L. The results of ultrasonic viscosity reduction-assisted proteolysis demonstrate the efficacy of ultrasound in modifying the viscosity of hydrolysates prepared at controlled DH. The ultrasound treatment conditions for reducing substrate viscosity showed to be the most applicable conditions for US viscosity reduction-assisted proteolysis to yield high peptide content. Correlations have been observed between substrate viscosity and hydrolysate viscosity, as well as between substrate viscosity and peptide yield. Additionally, US viscosity reduction-assisted proteolysis has proved its role in improving other functional properties and altering the protein structure of the hydrolysate prepared under the same DH. In conclusion, this study highlights the potential of ultrasonication as a valuable technique for viscosity reduction and facilitating enzymatic hydrolysis of soy protein isolate. The findings suggest that this approach can positively impact the functional properties and protein structure of the resulting hydrolysate. Further research in this area could offer valuable insights for the food industry and contribute to the development of improved protein-based products.

CRediT authorship contribution statement

Yolandani: Conceptualization, Methodology, Formal analysis, Writing – original draft. Haile Ma: Conceptualization, Resources, Supervision. Dandan Liu: Methodology. Yu Cheng: Conceptualization, Supervision. Fredy Agil Raynaldo: Validation, Investigation. Mokhtar Dabbour: Validation, Investigation. Jiapin Chao: Methodology. Asad Ali: Writing – review & editing. Susu Yang: Writing – review & editing.

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.