Impact of reaction sequence on the physicochemical stability and in vitro digestibility of β-carotene-loaded emulsions stabilized by a ternary covalent complex

Abstract

The construction of ternary covalent complexes using soy protein isolate (SPI), epigallocatechin gallate (EGCG), and polydextrose (PD) via an alkaline-mediated Maillard reaction was investigated to stabilize β-carotene-loaded emulsions. This study investigated how the reaction sequence impacts emulsion stability and in vitro digestion. Results indicated that ternary complexes offered superior physical stability and antioxidant protection compared to SPI or binary systems(SPI-EGCG and SPI-PD). Crucially, the SPI-PD-EGCG sequence (protein-polysaccharide-polyphenol) outperformed the SPI-EGCG-PD sequence. The SPI-PD-EGCG emulsion demonstrated the highest thermal stability (85 ± 0.01% retention) and best storage performance after 21 days, characterized by the lowest oxidation levels (POV: 26.54 ± 0.12 mmol/L; TBARS: 4.51 ± 0.02 μmol/L), smallest particle size (516.77 ± 2.29 nm), and highest β-carotene retention (56 ± 0.01%). The SPI-PD-EGCG system showed the lowest free fatty acid release rate and β-carotene bioaccessibility, demonstrating that this ternary covalent complex offers superior protection for β-carotene in the simulated gastrointestinal environment by reducing its oxidative degradation during digestion.These findings highlight that systematically controlling the conjugation order is a vital strategy for designing efficient delivery systems for lipophilic bioactives. These findings provide a rational basis for designing efficient delivery systems for β-carotene and other fat-soluble nutrients.

Highlights

• •Reaction sequence regulated the structure of ternary covalent complexes.
• •SPI-PD-EGCG complex exhibited superior emulsifying and protective effects.
• •Denser interfacial layers enhanced physical and oxidative stability but reduced lipolysis and bioaccessibility.
• •SPI-PD-EGCG complex showed the lowest FFA release and β-carotene bioaccessibility, indicating superior upper gastrointestinal protection.

1. Introduction

As a fat-soluble carotenoid, β-carotene possesses a structure rich in unsaturated linkages, which renders it susceptible to isomerization and oxidative deterioration throughout manufacturing, storage, and digestion. (X. Xu et al., 2023). The utilization of β-carotene within functional dietary matrices faces restrictions due to intrinsic hurdles, particularly regarding its tendency to degrade(C. Yang et al., 2026). As a result, entrapment inside emulsified carrier platforms constitutes a key method regarding the resolution of these issues. As a natural amphiphile, SPI functions as an effective emulsifier.Factors like pH, ionic strength, thermal processing, and oxidative stress can compromise the stability of emulsions it stabilizes, leading to variations in gastrointestinal digestion(Lin et al., 2024). This can cause the premature release and degradation of encapsulated compounds. Covalent modification is a common approach to functionally enhance proteins, thereby improving emulsion stability and delaying the release of encapsulated substances. Compared to non-covalent interactions (e.g., hydrogen bonding, hydrophobic interactions, and electrostatic attraction), covalent bonds confer superior stability and irreversibility to the resulting complexes(Liu et al., 2016).

Alkaline treatment is a frequently used method for covalent conjugation between proteins and polyphenols. This method involves dispersing the protein and polyphenol in an alkaline solution and introducing oxygen to drive the reaction. During this process, the phenolic hydroxyl groups of the polyphenol are readily oxidized to o-quinones. These o-quinones can form dimers or polymers via side reactions. Alternatively, they may react with protein nucleophiles, establishing C—N and C—S covalent linkages on the phenolic ring. This ultimately yields stable covalent complexes with enhanced antioxidant and interfacial activities. For example, Abd et al. covalently conjugated β-lactoglobulin with caffeic acid under alkaline conditions (pH 8.5) for 24 h. The covalent interaction increased the antioxidant activity of the resulting complex. This wet-heating Maillard pathway involves a condensation mechanism that facilitates the attachment of proteinaceous amine groups to the reducing-end carbonyls of polysaccharide chains, resulting in covalent conjugates(A. Huang et al., 2024; Zhou et al., 2024). These conjugates exhibit improved solubility, thermal stability, and steric hindrance. For instance, Yang et al. found that Maillard reaction products of SPI and polysaccharides could maintain the physical stability of citral emulsions for extended periods under both heat treatment and simulated gastrointestinal digestion(Y. Yang, Cui, et al., 2015). Therefore, an important direction in developing efficient delivery systems is to construct ternary covalent complexes, which combine the dual functionalities of protein-driven emulsification and polysaccharide-mediated steric stabilization, while integrating the antioxidant capacity of polyphenols.

EGCG is the most abundant catechin compound in green tea. Its molecular structure is rich in active phenolic hydroxyl groups, making it an excellent hydrogen donor and electron donor. It can efficiently scavenge free radicals and interrupt oxidative chain reactions, thereby exhibiting outstanding antioxidant activity. PD is a water-soluble dietary fiber formed by random cross-linking of glucose, primarily via α-1,6-glycosidic bonds, with a highly complex molecular structure. As a high-quality prebiotic, PD possesses properties such as low caloric content, anti-cariogenic activity, and resistance to digestive degradation. It can improve food texture and mouthfeel and has been widely used in functional foods and dairy products. In a study by Yang et al., non-covalent complexes of lactoferrin, pectin, and EGCG were systematically prepared following different sequential orders. It was found that the sequential introduction of EGCG and pectin played a crucial role in defining the physicochemical attributes and performance of the composite emulsifier(W. Yang, Xu, et al., 2015). In a study by Zhang et al., WPI, CA, and HMP were utilized to develop emulsions encapsulating β-carotene, which were stabilized via ternary covalent structures prepared using varying reaction pathways. (Y. Zhang et al., 2023). The findings suggested that the WPI-HMP-CA complex conferred slightly enhanced protection to β-carotene in the emulsion.

Our team previously explored the effects of different reaction sequences (SPI-PD-EGCG and SPI-EGCG-PD) on the structural conformation, interfacial properties, and emulsifying performance of ternary covalent complexes composed of SPI, EGCG, and PD(Zheng et al., 2026). It was found that the reaction sequence critically governed the grafting efficiency: the SPI-PD-EGCG sequence (polysaccharide grafted first, followed by polyphenol) exhibited a higher polysaccharide grafting degree and a higher EGCG binding equivalent, along with a more disordered secondary structure. Emulsions stabilized by this complex achieved the smallest droplet size, the highest absolute zeta-potential, the lowest interfacial tension, and the highest interfacial protein adsorption rate, as well as superior storage stability and a stronger three-dimensional network. However, that study focused primarily on the structure-function relationships of the complexes themselves, without systematically investigating how the reaction sequence affects the stability of encapsulated bioactives under environmental stresses and during in vitro digestion. Therefore, on the basis of those findings, the present study further examines the influence of the two reaction sequences on the physicochemical stability, oxidative stability, and in vitro digestibility of β-carotene-loaded emulsions stabilized by the ternary covalent complexes.

2. Materials and methods

2.1. Materials

SPI was extracted in-house with a purity of 93.4 ± 0.28%. EGCG (≥98%), PD(≥90%), β-carotene (96%), soybean oil, thiobarbituric acid, methanol, n-butanol, n-hexane, isopropane, absolute ethanol, Nile Red, pepsin(>3000 U/mg), pancreatin(>250 U/mg), porcine bile salts, isooctane, isopropanol, ammonium thiocyanate, BaCl₂, and FeSO₄ were all purchased from Shanghai Yuanye Biotechnology Co., Ltd.

2.2. Preparation of samples

2.2.1. Extraction of SPI

Defatted soy flour was accurately weighed and mixed thoroughly with deionized water at a solid-to-liquid ratio of 1:10 (w/v). The pH of the mixture was adjusted to 8.0 with 2 mol/L NaOH, and the suspension was stirred at room temperature for 2 h. Subsequently, it was centrifuged at 9000 rpm for 30 min at 4 °C, and the supernatant was collected. The pH of the supernatant was then adjusted to 4.5 with 2 mol/L HCL. After standing for 1 h to allow precipitation, the mixture was centrifuged at 6500 rpm for 30 min at 4 °C, and the precipitate was collected. The precipitate was washed three times with deionized water, after which the pH was adjusted to 7.0. SPI was obtained by vacuum freeze-drying the final product(S. Yan et al., 2021). The protein purity of the obtained SPI, as determined by the Kjeldahl method, was 92.8% ± 0.31%.

2.2.2. Preparation of the SPI-EGCG covalent complex

An SPI solution (10 mg/mL) was stirred at 4 °C for 24 h to achieve full hydration. Subsequently, its pH was leveled to 9.0 using 1 mol/L NaOH. Subsequently, EGCG (0.015 wt%) was introduced into the protein solution under conditions where pH was held at 9.0 with 1 mol/L NaOH. Following 24 h of stirring, the mixture pH was readjusted to 7.0 using 1 mol/L HCL To separate free polyphenols, the solution was subjected to dialysis against distilled water (4 °C, 48 h), followed by lyophilization. The lyophilized product was maintained at 4 °C until needed(Sui et al., 2018).

2.2.3. Preparation of the SPI-PD covalent complex

The SPI-PD covalent complex was prepared via the wet-heating Maillard reaction. An SPI solution (10 mg/mL) was stirred at 4 °C for 24 h to achieve full hydration. Following SPI hydration, PD was introduced at a 4:1 mass ratio (w/w). The solution underwent 4 h of magnetic stirring, and its pH was adjusted to 7.0 using 1 mol/L HCL or NaOH prior to a 1 h heat treatment at 95 °C with stirring. To quench the reaction, the mixture was rapidly cooled to room temperature. The resulting product was collected as a powder via lyophilization(Pirestani et al., 2018).

2.2.4. Preparation of the SPI-PD-EGCG covalent complex

As detailed in Section 2.2.1, the ternary SPI-PD-EGCG covalent complex emulsifier was prepared by replacing the SPI in the alkaline treatment step with the pre-formed SPI-PD covalent complex.

2.2.5. Preparation of the SPI-EGCG-PD covalent complex

As detailed in Section 2.2.2, the ternary SPI-EGCG-PD covalent complex emulsifier was prepared by replacing the SPI in the Maillard reaction step with the pre-formed SPI-EGCG covalent complex.

2.3. Preparation of β-carotene emulsions

Emulsions were formulated with SPI and its binary and ternary complexes. The aqueous phase was prepared by dissolving the respective complex in deionized water, with its protein concentration adjusted to 1% (w/v). Subsequently, it was subjected to continuous stirring at 4 °C for 12 h to ensure full hydration.Separately, the oil phase was constituted by dissolving β-carotene in soybean oil, with stirring at room temperature to facilitate complete dissolution.The oil phase and aqueous phase were mixed at a mass ratio of 9:1. A coarse emulsion was first obtained by homogenizing the mixture using an IKA T25 homogenizer (IKA-Werke GmbH & Co. KG, Germany) at 12,000 rpm for 3 min. The coarse emulsion was then homogenized at 60 MPa for three passes using a high-pressure homogenizer (AH-BASIC, ATS Industrial Systems Ltd., Canada) to produce the fine emulsion. The β-carotene mass fraction in this emulsion was 0.1%(S. Yan et al., 2021).

2.4. Multiple light scattering

A 20 mL aliquot of each emulsion was introduced into a cylindrical glass cell with a flat bottom for scanning at 880 nm using a Turbiscan Lab instrument (Formulaction Co., L'Union, France). Over a 2 h period, both backscattering (BS) and transmission (T) profiles were automatically collected at 2 min intervals(D. Xu et al., 2016). The overall stability was quantified by the Turbiscan Stability Index (TSI), which is derived from the changes in T and BS signals over time. TSI was computed according to Eq. (1).

$$\mathrm{TSI}=\sum _i\frac{\left(\sum _h|{\mathit{scan}}_i\left(h\right)-{\mathit{scan}}_{i-1}\left(h\right)|\right)}{H}$$ (1)

where scan_i(h) refers to the average backscattering at a given measurement time (i), scan_i-1(h) is the corresponding value at the previous time point (i-1), and H is the sample height.

2.5. Standard curve for β-carotene

To begin, 5 mg of β-carotene standard was dissolved in n-hexane and the solution was made up to 25 mL to prepare the initial stock. From this, a 500 μL aliquot was taken and its volume was increased to 10 mL with n-hexane, producing the intermediate solution. From this intermediate solution, aliquots of 1, 2, 3, 4, and 5 mL were separately pipetted and each was brought to a final volume of 10 mL with n-hexane, resulting in a series of standard solutions with final β-carotene concentrations of 1, 2, 3, 4, and 5 μg/mL, respectively. Using deionized water as a blank control, the absorbance of these standard solutions was measured at 450 nm. A standard curve of absorbance versus β-carotene concentration was established. The curve equation was determined to be $\mathrm{Y}=0.2852x-0.0306,R^2=0.9999$ with a coefficient of determination (R²) of 0.9999.

2.6. Retention rate of β-carotene

First, a 20 μL aliquot of the emulsion was diluted 50-fold. This diluted sample underwent triple extraction, each time with 1 mL of a 1:1 (v/v) ethanol/n-hexane mixture. The organic (upper) phases from all extractions were pooled, and the absorbance of this combined extract was read at 450 nm. The β-carotene concentration was subsequently quantified by interpolation from a pre-established standard curve.

$$\text{Retention of}\mathrm{\beta }-\text{carotene}\left(\%\right)=\frac{\mathrm{C}}{{\mathrm{C}}_0}\times 100$$ (2)

where C represents the actual concentration of β-carotene at any given stage, and C₀ represents its initial concentration.

2.7. Particle size and zeta potential

Subsequent to a 1000-fold dilution using deionized water to mitigate multiple scattering, the particle size and zeta potential were assessed with dynamic light scattering with three replicates per sample. The measurements utilized preset refractive indices of 1.460 for the particles and 1.330 for the aqueous dispersant.

2.8. Determination of environmental stability

2.8.1. pH stability

The pH of the fresh emulsions prepared with different complexes was adjusted within a range of 2.0 to 9.0 using appropriate concentrations of NaOH or HCl solutions. The emulsions were then kept at room temperature in the dark for 24 h. Subsequently, their particle size and zeta potential were measured.

2.8.2. Ionic strength stability

The ionic strength stability was examined by combining 8 mL of fresh emulsion with NaCl solutions to achieve final concentrations of 50, 100, and 200 mM. Following a 24 h storage period at 4 °C, the emulsions were then characterized by measuring changes in particle size, zeta potential, and visual appearance.

2.8.3. Thermal stability

Heat treatment was conducted by incubating emulsion samples at 90 °C in a water bath for 30 min, followed by cooling to ambient temperature. Subsequent to cooling, the droplet size, zeta potential, and extent of β-carotene retention were determined.

2.9. Storage stability

The β-carotene emulsions stabilized by different complexes were aliquoted (8 mL per tube) and maintained under refrigeration at 4 °C. The stability was monitored over a 21-day period. At designated time points (days 1, 7, 14, and 21), the emulsions were visually inspected for phase separation (creaming/sedimentation). Furthermore, their particle size, zeta potential, visual appearance, and β-carotene retention rate were measured.

2.10. Oxidative stability

2.10.1. Determination of peroxide value

Based on the method of Chen et al.(Y. Chen, Wang, et al., 2018) with modifications, emulsion peroxide content was determined. A 0.3 mL aliquot of the sample was mixed with 1.5 mL of an isooctane/ isopropanol blend (3:1 volume ratio). The mixture was vortexed for a duration of 1 min and centrifuged at 5000 rpm for 5 min to separate the organic phase. Subsequently, 200 μL of the supernatant was combined with 2.8 mL of a methanol/n-butanol solution (2:1, v/v). Next, 50 μL of 3.94 M ammonium thiocyanate followed by 50 μL of ferrous ion reagent were introduced. The optical density was recorded at 510 nm, and peroxide levels were determined via a standard curve.

2.10.2. Determination of thiobarbituric acid reactive substances (TBARS)

MDA concentration was determined using a TBA assay. The protocol involved combining 0.3 mL of the sample with 1.8 mL of distilled water, after which 4.0 mL of TBA solution was introduced. The reaction tube was incubated at 100 °C, immediately quenched on ice, and centrifuged (15 min, 10,000 rpm). The developed color in the supernatant was measured at 532 nm, and concentrations were derived from a standard curve prepared with 1,1,3,3-tetraethoxypropane(Hu & Zhang, 2022).

2.11. In vitro digestion experiment

A two-phase in vitro digestion protocol, adapted from Yan et al.(Y. Yan et al., 2020) with modifications, was used to assess the digestive characteristics of β-carotene emulsions stabilized by ternary covalent complexes.

Simulated gastric fluid (SGF) preparation: SGF was freshly prepared by dissolving 2.0 g of NaCl, 7.0 mL of HCl (37%), and 3.2 g of pepsin (>3000 U/mg) in deionized water, and the volume was adjusted to 1 L. The final pH was adjusted to 2.0 ± 0.1 using 1 M HCl.

Simulated intestinal fluid (SIF) preparation: SIF was prepared by dissolving 6.8 g of KH₂PO₄, 8.5 g of NaCl, and 10.0 g of pancreatin (>250 U/mg) in deionized water, with the volume adjusted to 1 L. Porcine bile salts (53.5 mg/mL) were prepared separately. The final pH of the SIF was adjusted to 7.0 ± 0.1 using 1 M NaOH.

Gastric digestion phase: Freshly prepared β-carotene emulsion (10 mL) was mixed with 20 mL of pre-warmed (37 °C) SGF. The pH of the mixture was adjusted to 2.0 ± 0.1 using 0.1 M HCl. The mixture was incubated in a shaking water bath at 37 °C with constant agitation at 100 rpm for 2 h.

Intestinal digestion phase: To the gastric digest (30 mL), the following pre-warmed (37 °C) components were added: 2 mL of SIF, 4.7 mL of bile salt solution (53.5 mg/mL), and 3.3 mL of pancreatin solution (24 mg/mL in SIF). The pH of the mixture was immediately adjusted to 7.0 ± 0.1 using 0.1 M NaOH. The mixture was incubated at 37 °C for 2 h under continuous shaking at 100 rpm. At the end of the intestinal phase, the digest was heated at 85 °C for 10 min to inactivate pancreatin, followed by cooling to room temperature for further analysis.

2.11.1. Particle size and zeta potential during in vitro digestion

The particle size and Zeta potential of the emulsions during in vitro digestion were precisely measured using a dynamic light scattering nanoparticle size and Zeta potential analyzer. Prior to measurement, the digestion samples were diluted 1000-fold with buffer solutions corresponding to the digestion pH.

2.11.2. Free fatty acid release rate

Throughout the simulated intestinal phase, the system pH was automatically maintained at 7.0 via titration with 0.1 mol/L NaOH at 20 min intervals. The volume of titrant added during digestion was recorded and used to calculate the rate of free fatty acid (FFA) release rate. The FFA release rate was calculated according to the following formula:

$$\text{Free fatty acid release rate}\left(\%\right)=\frac{V_{\mathit{NaOH}}\times m_{\mathit{NaOH}}\times M_{\mathit{lipid}}}{W_{\mathit{lipid}}\times 2}\times 100$$ (3)

where $V_{\text{NaOH}}$ is the volume of NaOH consumed (mL), $M_{\text{NaOH}}$ is the molarity of the NaOH solution (mol/mL), $M_{\text{lipid}}$ is the average molecular weight of soybean oil (876 g/mol), and $W_{\text{lipid}}$ is the mass of oil in the initial emulsion (g).

2.11.3. Bioaccessibility

The small intestinal digesta from each group was separated by centrifugation (10,000 rpm, 40 min). The intermediate aqueous layer, representing the “micellar” phase formed during digestion, was collected. This fraction was filtered to eliminate any crystallized β-carotene, and approximately 1 mL was retained for analysis. Bioaccessibility, defined as the ratio of β-carotene concentration in this micellar fraction to that in the initial emulsion, was calculated. The β-carotene content in both matrices was determined via extraction and quantification (Qian et al., 2012). The calculation formula is as follows:

$$\text{Bioaccessibility}\left(\%\right)=100\times \frac{{\rho }_{\mathit{micelle}}}{{\rho }_{\mathit{initial}}}$$ (4)

where ${\rho }_{\text{micelle}}$ and ${\rho }_{\text{initial emulsion}}$ represent the mass concentration of β-carotene in the micellar fraction and the initial emulsion, respectively.

2.11.4. Confocal laser scanning microscopy (CLSM)

To visualize the emulsion microstructure during simulated digestion, confocal microscopy was utilized. The lipid phase was stained using a Nile Red stock solution (1 mg/mL in ethanol) with an excitation wavelength of 543 nm. A 2 mL aliquot of the digest was incubated with 100 μL of the dye solution. This mixture was then applied onto a microscope slide, covered, and examined under the confocal microscope for imaging(Z. Huang et al., 2022).

2.12. Data analysis

All experimental measurements were conducted with three independent replicates, and the results are presented as mean ± standard deviation. Differences among group means were assessed by one-way analysis of variance (ANOVA), with post-hoc comparisons performed using Tukey's test at a significance level of p < 0.05. Graphical representations of the data were generated with Origin 2019 software (OriginLab, Northampton, MA, USA).

3. Results and discussion

3.1. Multiple light Sscattering

Changes in BS intensity and the TSI were assessed to evaluate emulsion stability during storage. Fig. 1A and B reveal that the BS profiles of emulsions stabilized by ternary covalent complexes (SPI-PD-EGCG and SPI-EGCG-PD) displayed greater stability relative to those stabilized by single SPI or binary complexes (SPI-EGCG and SPI-PD). Concurrently, the TSI values for the ternary complex-stabilized emulsions increased more slowly and reached lower final values. This indicates a clear advantage of ternary complexes in inhibiting droplet flocculation, coalescence, and gravitational separation, with their stability being significantly superior to that of binary systems (SPI-EGCG and SPI-PD) (Su et al., 2020).

Fig. 1.

Delta backscattering prolles(A), Turbiscan stability index (B) of emulsion stabilized by SPI and its binary and ternary complex samples.

The physicochemical robustness of the ternary covalent complexes differed notably contingent upon the synthesis pathway. Within this group, the system emulsified utilizing the conjugate prepared in the SPI-PD-EGCG sequence exhibited the smallest change in BS profile throughout the testing period, along with the lowest and most slowly increasing TSI value. This suggests that the ternary covalent complex emulsion prepared in the SPI-PD-EGCG order was the most stable. This enhanced stability likely arises from the specific reaction sequence that fosters a denser, more cohesive and elastic interfacial film, thereby more effectively inhibiting droplet aggregation(Jian et al., 2023). In comparison, although the emulsion stabilized by SPI-EGCG-PD was also superior to binary systems(SPI-EGCG and SPI-PD), its BS and TSI profiles showed relatively greater fluctuation. This demonstrates that the reaction sequence modulates the physical stability of the emulsion by influencing the interfacial structure and spatial arrangement of the complex.

In summary, the application of ternary covalent complexes can considerably enhance the physical stability of β-carotene emulsions, a performance which is crucially influenced by the specific reaction sequence employed. Within the system studied, the ternary complex prepared following the SPI-PD-EGCG reaction sequence exhibited the optimal stabilizing effect, providing a basis for its application in efficient delivery systems.

3.2. pH stability

As indicated in Fig. 2A and B the emulsion stabilized solely by SPI displayed the greatest particle size enlargement and the minimal absolute zeta potential over the pH 2.0–9.0 interval, suggesting its comparatively inferior stability(K. Zhang et al., 2024). The performance of emulsions stabilized by binary covalent complexes (SPI-EGCG and SPI-PD) showed improvement, but still displayed significant particle size increase and zeta potential approaching zero near the isoelectric point (pH 4–5), leading to aggregation. Conversely, emulsions stabilized by ternary covalent complexes (SPI-EGCG-PD and SPI-PD-EGCG) demonstrated smaller particle sizes and levated absolute zeta potential throughout the tested pH spectrum.This is due to the interfacial film formed through rapid adsorption and synergistic cross-linking within the ternary complexes, coupled with their superior ability to reduce interfacial tension, which promotes more uniform oil droplet dispersion and reduces particle size (Y. Huang et al., 2025). Furthermore, the ternary complexes provide stronger electrostatic repulsion and steric hindrance (Y. Huang et al., 2025), effectively mitigating the tendency for aggregation around the isoelectric point, thereby conferring broader pH adaptability and better dispersion stability.

Fig. 2.

Particle size(A),Zeta potential(B)of emulsion stabilized by SPI and its binary and ternary complex samples treated with different pH; Particle size(C),Zeta potential(D),C/Co(E)of emulsion stabilized by SPI and its binary and ternary complex samples treated with heat.

The two ternary complexes prepared via different reaction sequences exhibited differences in pH stability. Emulsions prepared with the SPI-PD-EGCG sequence exhibited the most consistent particle size across the pH range tested. This was reflected in relatively lower peak sizes under highly acidic conditions (pH 2.0) and within the isoelectric point region (pH 4.0). It also maintained higher absolute zeta potential values under neutral to alkaline conditions. This may be because this sequence is more conducive to constructing a synergistically efficient interfacial structure, where the polysaccharide layer provides excellent physical resistance to coalescence, and the polyphenol layer reinforces chemical stability, resulting in superior overall stability across a pH range. In the interfacial structure formed by the protein-polyphenol-polysaccharide reaction sequence, the integrity and thickness of the physical barrier might be slightly inferior to those of the protein-polysaccharide-polyphenol sequence. Consequently, when confronting charge neutralization and protein denaturation induced by extreme pH, its physical stability appears somewhat less robust.

In summary, ternary covalent complexes can notably improve the stability of β-carotene emulsions under varying pH conditions, thereby extending their potential applications. This enhancement stems primarily from the covalent integration of PD and EGCG. This combination works synergistically to reinforce both steric hindrance and electrostatic repulsion within the interfacial layer. The SPI-PD-EGCG reaction sequence demonstrates potential within this system for constructing a superior interfacial structure and achieving better pH stability.

3.3. Thermal stability

Evaluation of thermal stability via particle size, zeta potential, and β-carotene retention showed that heating induced a significant particle size increase in all emulsions (Fig. 2C-E), alongside a decrease in absolute zeta potential. This indicates droplet aggregation and enhanced shielding of surface charges (Salminen & Weiss, 2014). The poorest thermal stability was observed for the emulsion stabilized solely by SPI. It displayed the most substantial changes, characterized by the greatest particle size increase and the lowest β-carotene retention (69%). Emulsions stabilized by binary covalent complexes (SPI-EGCG, SPI-PD) exhibited improved metrics across all indicators, with β-carotene retention rates rising to 74–76%. This suggests that covalent modification alleviated thermal aggregation and protected the active compound to some extent.In contrast, emulsions stabilized by ternary covalent complexes demonstrated the smallest increase in particle size, relatively well-maintained zeta potential, and the highest β-carotene retention rate, indicating optimal thermal stability. This is likely because the covalent conjugation among protein, polyphenol, and polysaccharide can raise the thermal denaturation temperature and provide enhanced steric hindrance that prevents droplet aggregation (Bi et al., 2017). This shows that ternary complexes more effectively resist protein denaturation and droplet coalescence induced by heat treatment, offering better protection for β-carotene.

When comparing ternary complexes prepared using varied reaction sequences, the one stabilized by SPI-PD-EGCG showed relatively better thermal stability. It showed the smallest increase in particle size, the least reduction in absolute zeta potential, and achieved the highest β-carotene retention (85%). This is due to the reaction sequence of protein-polysaccharide-polyphenol constructing a structure where the polysaccharide layer acts as a physical barrier, while the polyphenol layer serves an anchoring function. This structure proved more effective in resisting droplet coalescence caused by protein denaturation during heating, resulting in minimal changes to particle size and zeta potential. In comparison, the structure built by the protein-polyphenol-polysaccharide sequence exhibited relatively lower synergistic efficiency between its physical barrier and conformational stability under thermal stress(Bi et al., 2017).

In summary, ternary covalent complexes are effective in improving both the thermal stability of emulsions and their capacity for offering enhanced protection to β-carotene. The reaction sequence is a key factor influencing their performance. The SPI-PD-EGCG sequence constructed an interfacial structure within this system where the physical barrier (PD) and the antioxidant (EGCG) demonstrate greater spatial and functional synergy(Jian et al., 2023), leading to superior stabilizing performance under thermal conditions.

3.4. Ionic strength stability

Data in Fig. 3 demonstrate that rising Na⁺ concentration led to a substantial increase in particle size and a sharp decline in absolute zeta potential for emulsions stabilized by SPI or the binary SPI-EGCG complex. These systems displayed severe phase separation even at 50–100 mM salt, indicating that electrostatic shielding critically undermined their stability (H. Chen, Ji, et al., 2018). The binary SPI-PD complex, where grafted polysaccharides provided strong steric hindrance, was less affected by salt ions and showed significantly improved stability. In contrast, emulsions stabilized by the two ternary covalent complexes (SPI-EGCG-PD and SPI-PD-EGCG) demonstrated the most outstanding salt tolerance. At all tested ionic strengths, they displayed minimal change in particle size and zeta potential and retained a homogeneous macroscopic state without observable phase separation. The superior performance is due to the synergistic electrostatic and steric effects offered through covalent bonding. This combination more effectively shielded the emulsion from salt-induced structural breakdown than was possible with binary complexes(SPI-EGCG and SPI-PD) (Yong Wang et al., 2011).

Fig. 3.

Particle size(A),Zeta potential(B), Appearance(C) of emulsion stabilized by SPI and its binary and ternary complex samples treated with different salt. Note: Samples in panel C from left to right are SPI, SPI-EGCG, SPI-PD, SPI-EGCG-PD, and SPI-PD-EGCG.

When subjected to high salt concentrations, the SPI-PD-EGCG-stabilized emulsion showed superior stability compared to other ternary complexes. This was corroborated by its negligible change in particle size, marginal decrease in zeta potential even at 200 mM Na⁺, and effective maintenance of a homogeneous appearance. This is likely due to a more optimized interfacial structure formed by this specific reaction sequence, where the polysaccharide chains create a protective layer that effectively shields the droplets from variations in ionic strength(Yong Wang et al., 2011). Comparatively, the emulsion stabilized by the SPI-EGCG-PD sequence performed slightly less effectively, particularly showing slightly weaker macroscopic homogeneity at very high salt concentrations. It is hypothesized that in this sequence, the initial binding of EGCG to SPI might have restricted the full extension and coverage of the subsequent PD polysaccharide chains at the interface, resulting in a spatial barrier that is relatively less efficient in resisting high-salt stress.

In summary, ternary covalent complexes can significantly enhance the ionic strength stability of β-carotene-loaded emulsions. The incorporation of the polysaccharide PD plays a key role in resisting salt-induced flocculation. The reaction sequence further modulates the salt tolerance of the emulsions by influencing the conformation and spatial arrangement of polysaccharide chains within the interfacial structure. The SPI-PD-EGCG sequence afforded superior physical stability at high ionic strength in this system.

3.5. Storage stability

As shown in Fig. 4, throughout the 21-day duration, the droplet diameter for all groups grew over time, and a decline in the zeta potential magnitude was observed, while β-carotene levels fell. Of these, dispersions emulsified using SPI in isolation displayed the most inferior robustness, characterized by a drastic rise in droplet diameter and distinct phase separation detected during the final storage phase, coupled with the minimal β-carotene retention ratio. The stability of emulsions stabilized by the binary covalent complexes (SPI-EGCG and SPI-PD) showed improvement; however, the SPI-EGCG group still exhibited phase separation relatively early.In contrast, emulsions stabilized by the two ternary covalent complexes demonstrated significantly optimized storage performance. They showed a slower increase in particle size, better maintenance of zeta potential, the highest β-carotene retention rates, and were able to maintain a homogeneous macroscopic appearance throughout the entire storage period. This indicates that these emulsions possess a stronger and more stable interfacial layer, making them less prone to aggregation and oxidative degradation of the active compound during storage (Jian et al., 2023).

Fig. 4.

Particle size(A),Zeta potential(B), C/C₀(C), Appearance(D) during storage of emulsions stabilized by SPI and its binary and ternary complex samples. Note: Samples in panel D from left to right are SPI, SPI-EGCG, SPI-PD, SPI-EGCG-PD, and SPI-PD-EGCG.

Between the two ternary complexes, the emulsion stabilized by the SPI-PD-EGCG sequence demonstrated relatively superior long-term storage stability. It exhibited the smallest increase in particle size, the slowest decline in zeta potential, and the highest β-carotene retention rate after 21 days. This is likely because the SPI-PD-EGCG reaction sequence constructs a physical protective layer with the polysaccharide as the backbone, which synergizes with the antioxidant. In addition to substantially suppressing physical coalescence and creaming of the dispersed phase throughout the holding period, this framework also provides a more direct shield for the core active substance against oxidative depletion.In summary, emulsions stabilized by the SPI-PD-EGCG ternary covalent complex possess better storage stability compared to those stabilized by single protein or binary covalent complexes(SPI-EGCG and SPI-PD), owing to their stronger and more stable interfacial layer(Jian et al., 2023). This identifies them as an effective delivery system for β-carotene.

3.6. Oxidative stability

The oxidative stability of β-carotene-loaded emulsions during storage was significantly influenced by the properties of the stabilizers. As shown in Fig. 5A and B, the emulsion stabilized by native SPI exhibited a rapid and significant increase in both lipid peroxides and TBARS values, indicating poor protection against lipid oxidation. Covalent conjugation with EGCG (SPI-EGCG) significantly suppressed the formation of peroxides and the increase in TBARS values, highlighting the direct antioxidant effect of the grafted polyphenol. The SPI-PD conjugate also showed reduced peroxide and TBARS values compared to SPI alone, with its effect primarily observed in delaying the later stages of oxidation (TBARS), possibly due to its enhanced physical barrier properties(Perreault et al., 2017).Notably, the emulsions stabilized by the ternary covalent complexes, SPI-EGCG-PD and SPI-PD-EGCG, demonstrated the most effective inhibition of lipid oxidation during the 21-day storage period. This superior performance can be attributed to a synergistic mechanism: the physical barrier provided by covalently linked polysaccharide chains hinders the penetration of pro-oxidants, while the grafted EGCG actively quenches free radicals at the interface.

Fig. 5.

Lipid hydroperoxides(A),TBARS(B), Particle size(C),Zeta potential(D) of emulsions stabilized by SPI and its binary and ternary complex samples in vitro digestion.

Among the different reaction sequences, the slightly better performance observed in the system emulsified via SPI-PD-EGCG aligns with data from ionic strength stability tests, suggesting the formation of a more robust and more compact surface coating that acts as an advanced mechanical shield(Perreault et al., 2017). This barrier prevents droplets from interacting with co-oxidants within the continuous medium. Research indicates that rigid particulates positioned along the oil-water boundary hinder the interaction connecting the emulsified oil and O₂; therefore, dispersions displaying improved physical stability often exhibit superior oxidative resistance(Yitong Wang et al., 2026). This further supports that the specific SPI-PD-EGCG reaction sequence optimizes the interfacial structure for maximal antioxidant protection. In summary, the application of the SPI-PD-EGCG ternary covalent complex can effectively delay lipid oxidation in the emulsion system.

3.7. In vitro simulated digestion experiment

3.7.1. Particle size and zeta potential

The gastrointestinal robustness of the colloidal systems was evaluated via measuring variations in droplet dimension and surface charge throughout the simulated digestive process (Fig. 5C and D). During the gastric phase, a notable enlargement in particle diameter was detected for every emulsion, coupled with a decline in the magnitude of the zeta potential. This is primarily ascribed to the intensely low pH environment, substantial salt concentration within the stomach fluid, and the degradation of the interfacial protein structure induced by pepsin (Singh & Sarkar, 2011). Within this group, dispersions emulsified using SPI and the binary SPI-EGCG complex displayed the most significant growth in droplet diameter along with the lowest stability. The binary SPI-PD complex showed a reduced increase in particle size, which is likely due to the steric hindrance provided by the covalently linked polysaccharide. In contrast, the two emulsions stabilized by ternary covalent complexes (SPI-EGCG-PD and SPI-PD-EGCG) maintained the smallest particle size and relatively higher absolute zeta potential values during the gastric phase. This may be attributed to their increased interfacial layer thickness, which effectively protects the emulsion droplets from the harsh gastric environment (Davidov-Pardo, Pérez-Ciordia, Marı´n-Arroyo & McClements, 2015). Upon entering the intestinal phase, their final particle size remained significantly smaller than that of other groups. This indicates that the ternary complexes can more effectively maintain the integrity of the interfacial layer in the digestive environment, mitigating droplet aggregation and thereby exhibiting superior digestive stability.

Between the ternary complexes prepared with different reaction sequences, the emulsion stabilized by SPI-PD-EGCG demonstrated the best stability throughout the entire digestion process. It consistently showed the smallest particle size and the relatively highest absolute zeta potential. This is likely because the polysaccharide (PD) can more effectively provide steric hindrance and charge stabilization in the gastric environment, resisting erosion by acid, enzymes, and ions. The subsequently incorporated EGCG further enhances the compactness and antioxidant capacity of the interfacial film. Together, these factors help preserve the integrity of the droplets during the intestinal phase. Comparatively, the performance of the emulsion stabilized by the SPI-EGCG-PD sequence was slightly inferior. It is hypothesized that in this sequence, the initial binding of EGCG to SPI might have restricted the full extension and coverage efficiency of the PD polysaccharide chains at the interface. This could result in a spatial barrier with relatively weaker resistance to the digestive environment, leading to slightly less effective control over particle size and zeta potential maintenance.

In summary, ternary covalent complexes can significantly enhance the stability of β-carotene-loaded emulsions in the simulated gastrointestinal environment, effectively inhibiting droplet aggregation and interfacial disruption during digestion. The SPI-PD-EGCG sequence achieves optimized synergy between steric hindrance and chemical stability, thereby providing better protection for the emulsion structure throughout the digestion process.

3.7.2. Bioaccessibility

Bioaccessibility was determined by measuring the β-carotene content in the mixed micellar phase after completion of emulsion digestion. As shown in Fig. 6A, the β-carotene-loaded emulsion stabilized by native SPI exhibited the highest bioaccessibility (41.5 ± 0.48%). This is because the interfacial protein film formed by SPI is relatively thin and loosely structured, making it susceptible to hydrolysis by pepsin and trypsin during digestion. Upon entering the small intestine, bile salts rapidly displace the residual proteins at the oil–water interface, allowing pancreatic lipase to anchor extensively and efficiently catalyze triglyceride hydrolysis. Sufficient lipid digestion generates a large number of mixed micelles, which in turn facilitate the substantial release of β-carotene from the oil phase into the micelles(Salvia-Trujillo et al., 2013). The bioaccessibility of emulsions stabilized by binary covalent complexes (SPI-EGCG and SPI-PD) decreased to approximately 38.2 ± 0.60% and 20.7 ± 0.41%, respectively. For ternary covalent complexes, bioaccessibility further decreased. Among the ternary complexes prepared with different reaction sequences, the SPI-PD-EGCG system exhibited the lowest bioaccessibility (12.5 ± 0.22%). This is likely attributed to its dense interfacial film and the inhibitory effect of EGCG on enzymes, which reduces micelle formation and consequently lowers bioaccessibility (Salvia-Trujillo et al., 2013). This result is positively correlated with the free fatty acid (FFA) release rate, because micelle formation is closely related to the extent of lipid hydrolysis; a lower degree of lipid hydrolysis makes micelle formation more difficult (Nan, Aiqian, M. & Harjinder, 2021), leading to reduced β-carotene bioaccessibility. At the same time, the conversion stability of β-carotene in this system was the highest, indicating that it can more effectively reduce the degradation loss of β-carotene during simulated gastrointestinal digestion, thereby exhibiting a stronger protective effect.

Fig. 6.

Bioaccessibility(A),FFA(B) of emulsions stabilized by SPI and its binary and ternary complex samples in vitro digestion.

3.7.3. Free fatty acid (FFA) release rate

The free fatty acid (FFA) release rate is influenced by the accessibility of lipase to the oil–water interface and is also related to the emulsion interfacial composition/interfacial film structure and the competitive adsorption of surface-active substances such as bile salts (Golding et al., 2011). As shown in Fig. 6B, all samples exhibited similar kinetic trends. During the first 20 min, all emulsions showed a relatively rapid FFA release rate due to the rapid adsorption of bile salts and lipase onto the oil–water interface. As digestion progressed (20–40 min), the FFA release rate decreased and stabilized after 60 min, which can be attributed to the saturation of bile salts at the interface and the accumulation of lipolysis products on the droplet surface, hindering further contact between lipase and lipids(Jiang et al., 2024). The emulsion stabilized by native SPI exhibited the fastest initial release rate and the highest final FFA release, indicating that the interfacial film formed by the single protein was relatively weak and was easily displaced by the more surface-active bile salts in the simulated small intestinal environment. Consequently, lipase rapidly adsorbed and anchored onto the oil droplet surface, leading to efficient lipid hydrolysis. In contrast, emulsions stabilized by covalently modified complexes showed varying degrees of reduction in FFA release, suggesting that the altered interfacial structure may restrict lipase activity and thereby reduce the extent of lipolysis (Ren et al., 2025). Notably, the most pronounced inhibitory effect was observed in the ternary covalent complexes. Among the ternary complexes prepared with different reaction sequences, the SPI-PD-EGCG system exhibited the slowest FFA release kinetics and the lowest final FFA release, indicating its stronger resistance to lipolysis. The interfacial film formed by SPI-PD-EGCG was more continuous and had higher surface coverage, minimizing the exposure of hydrophobic regions and thus substantially reducing the binding sites for lipase. This demonstrates that the interfacial structure constructed by SPI-PD-EGCG can significantly delay lipid digestion and, to some extent, limit the entry of β-carotene into mixed micelles during the intestinal digestion stage, suggesting its strong protective effect in the upper gastrointestinal tract.

3.7.4. Confocal laser scanning microscopy (CLSM)

CLSM was employed to visualize the microstructure of emulsions at different digestion stages, and the changes in droplet morphology and aggregation degree were analyzed in conjunction with particle size data (Fig. 7). In the initial emulsions, the SPI-stabilized emulsion exhibited large oil droplets with pronounced aggregation, indicating limited emulsifying capacity. After covalent modification with EGCG or PD, the binary complexes (SPI-EGCG and SPI-PD) produced emulsions with more regular droplet shapes and a more uniform size distribution, confirming that covalent modification improves the interfacial adsorption and droplet-stabilizing ability of SPI(Ju et al., 2020). Emulsions stabilized by ternary covalent complexes (SPI-EGCG-PD and SPI-PD-EGCG) showed even smaller droplet sizes and much better spatial uniformity. Notably, the SPI-PD-EGCG-stabilized emulsion possessed the smallest and most uniformly distributed droplets, which is consistent with the particle size data (SPI-PD-EGCG: 516.77 nm vs. SPI-EGCG-PD: 589.34 nm). The formation of finer droplets provides a larger surface area for emulsifier adsorption, indicating that the SPI-PD-EGCG complex possesses high interfacial activity and can form an effective stabilizing film (Hebishy et al., 2015). After 2 h of gastric digestion, all emulsions exhibited varying degrees of flocculation and droplet size increase. This is attributed to the strongly acidic pH (2.0) and high salt concentration in simulated gastric fluid, which screen electrostatic repulsion, as well as pepsin-induced degradation of interfacial proteins. The SPI- and SPI-EGCG-stabilized emulsions showed the most severe aggregation, with droplet sizes increasing sharply from initial values of 894 nm and 645 nm to approximately 1850 nm and 1520 nm, respectively, accompanied by irregular large aggregates. The SPI-PD-stabilized emulsion showed a smaller increase in particle size due to the steric hindrance provided by the covalently grafted PD. In contrast, emulsions stabilized by ternary covalent complexes (SPI-EGCG-PD and SPI-PD-EGCG) maintained relatively smaller droplet sizes after the gastric phase, with significantly reduced aggregation. Particularly for the SPI-PD-EGCG-stabilized emulsion, droplet boundaries remained relatively clear, and only a few small aggregates were observed. This is attributed to the polysaccharide (PD) layer effectively resisting gastric acid and enzyme erosion, while the covalent cross-linking of EGCG enhances the compactness of the interfacial film. Following 2 h of intestinal digestion, the number of emulsion droplets decreased and their sizes shrank due to the hydrolysis of triglycerides by pancreatin. The droplet sizes of SPI- and SPI-EGCG-stabilized emulsions decreased to approximately 650 nm and 580 nm, respectively, but CLSM images showed numerous irregular fragments and some incompletely hydrolyzed large oil droplets, indicating incomplete lipolysis. The SPI-PD-stabilized emulsion exhibited droplet sizes reduced to about 520 nm with relatively uniform distribution. Remarkably, emulsions stabilized by ternary covalent complexes showed even smaller and more uniform droplet sizes after the intestinal phase, which corresponds well with the FFA release data showing that SPI-PD-EGCG had the lowest final FFA release.

Fig. 7.

CLSM of emulsions stabilized by SPI and its binary and ternary complex samples in vitro digestion.

The CLSM observations are fully consistent with the macroscopic stability data. The smallest initial droplet size and most compact interfacial film of the SPI-PD-EGCG-stabilized emulsion directly contributed to its excellent physical stability under pH, heat, salt, and storage conditions. The minimal flocculation degree and most intact droplet structure of the SPI-PD-EGCG emulsion during the gastric phase reduced the oxidative loss of β-carotene in the acidic environment. The uniform and fine micellar morphology of the SPI-PD-EGCG emulsion during the intestinal phase, together with its sustained FFA release profile, indicates that its interfacial network structure can modulate lipase adsorption to achieve controlled lipolysis, thereby maximizing β-carotene micellarization. In conclusion, CLSM images provide direct visual evidence that the SPI-PD-EGCG ternary covalent complex maintains superior interfacial structural integrity throughout digestion.

4. Conclusion

This study constructed β-carotene-loaded emulsions stabilized by a single protein, binary covalent complexes, and ternary covalent complexes with different reaction sequences by controlling the covalent conjugation order among SPI, EGCG, and PD. The influence of these covalent complexes on emulsion stability, the protective effect on β-carotene, and the digestion characteristics were compared.The experimental results demonstrated that, compared to emulsions stabilized by the protein alone or binary covalent complexes(SPI-EGCG and SPI-PD), the emulsion stabilized by the ternary covalent complex exhibited significantly improved stability against various environmental stresses, including pH changes, heat treatment, ionic strength, and storage. Furthermore, the interfacial structure constructed using the SPI-PD-EGCG reaction sequence demonstrated superior overall performance compared to the SPI-EGCG-PD sequence.The β-carotene emulsion stabilized by the SPI-PD-EGCG complex showed the most outstanding physical stability and the highest β-carotene retention rate under all tested stress conditions (pH, ionic strength, heat, and storage). During in vitro simulated digestion, the SPI-PD-EGCG-stabilized emulsion exhibited a lower free fatty acid release rate and the lowest bioaccessibility (12.5 ± 0.22%). The SPI-PD-EGCG ternary covalent complex demonstrated excellent upper gastrointestinal tract protection in the simulated gastrointestinal environment, providing a theoretical basis for its further development as a functional factor delivery carrier. This study is the first to elucidate the reaction sequence-dependent regulatory mechanism of the SPI-PD-EGCG ternary covalent complex, offering new insights for the design of food macromolecule ternary complexes.

CRediT authorship contribution statement

Baoning Zheng: Writing – original draft, Software, Methodology, Formal analysis, Data curation. Ying Zhu: Writing – review & editing. Bingyu Sun: Investigation. Linlin Liu: Validation. Mingshou Lv: Validation. Xiuqing Zhu: Writing – review & editing. Yuyang Huang: Visualization, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.

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.

Data availability

The data that has been used is confidential.