Ultrasound-modified soy protein-hyaluronic acid conjugates for swallowing-friendly high internal phase pickering emulsion gels: structure, properties, and applications

TL;DR

This study develops a new type of soft, swallow-friendly food using soy protein and hyaluronic acid, suitable for people with swallowing difficulties.

Contribution

The use of ultrasound-modified soy protein-hyaluronic acid conjugates to create high-performance, swallow-friendly Pickering emulsion gels.

Findings

Ultrasound treatment improved the emulsifying and structural properties of the SPI/HA conjugate.

The resulting gels had soft texture, low adhesiveness, and good swallowability, meeting dysphagia diet requirements.

The gels showed potential for nutrient delivery and 3D printing applications.

Abstract

With the increasing global aging population, there is a growing demand for safe and nutritious specialized foods tailored to individuals with dysphagia. However, designing foods with suitable texture, stability, and swallowability remains a significant challenge. This study explores the use of high-intensity ultrasound (HIU)-modified soy protein isolate (SPI)/hyaluronic acid (HA) complexes and conjugates to fabricate high internal phase Pickering emulsion gels (HIPPEGs) suitable for dysphagia patients. The formation of the SPI/HA conjugate was demonstrated by SDS-PAGE and degree of grafting. Compared to SPI alone, the HIU-treated conjugate exhibited smaller particle size, higher surface charge, greater surface hydrophobicity, and enhanced structural flexibility, leading to superior emulsifying performance. Rheological analysis confirmed that HIPPEGs stabilized by U-Conjugate possessed higher gel strength. Texture profile and IDDSI tests indicated a soft texture, low adhesiveness, and favorable swallowability—all meeting the requirements for dysphagia diets. LF-NMR and CLSM analyses further revealed uniform water distribution and a dense, homogeneous microstructure. Moreover, U-Conjugate-stabilized HIPPEGs demonstrated potential for in vitro quercetin delivery and suitability for 3D printing applications. In conclusion, this study provides an effective strategy for developing high-performance HIPPEGs with strong potential for application in dysphagia-friendly food design, particularly highlighting their capabilities in nutrient delivery and 3D printing.

Figures

Average particle size (A), particle size distribution (B), zeta potential (C), and protein flexibility (D), DG (E), and SDS-PAGE (F) of SPI, Mixture, and Conjugate in aqueous solution.

TEM images of SPI, Mixture, and Conjugate aqueous solutions at different magnifications. Scale bars: 5 μm for the 2000× images; 2 μm for the 5000× images.

UV-Vis spectra (A), FTIR spectra (B), surface hydrophobicity (C), intrinsic fluorescence spectra (D), and fluorescence excitation-emission matrix (EEM) spectra (E) of SPI, Mixture, and Conjugate in aqueous solution.

Secondary structure content (A), emulsifying activity index (EAI, B), emulsifying stability index (ESI, C) of SPI, Mixture, and Conjugate aqueous solutions; MRI images (D), LF-NMR (E), and peak area (F) of the formed HIPPEGs.

Hardness (A), adhesiveness (B), chewiness (C) from texture profile analysis (TPA), swelling ratio (D), and the gelation process (E-I) of HIPPEGs based on SPI, Mixture, and Conjugate.

IDDSI tests (fork pressure and spoon tilt tests) (A), shear rate sweep (B), frequency sweep (C), 3ITT (D), and 3D printing (E) tests of HIPPEGs.

Optical microscopy images (A), general appearance (B), encapsulation efficiency (EE) (C), and bioaccessibility (D) of SPI and U-Conjugate HIPPEGs.

Full text

Introduction

1

The World Health Organization (WHO) projects that the global population aged 60 and above will reach 1.5 billion by 2050 [1]. With this accelerating demographic shift toward aging, many older adults face swallowing difficulties due to conditions such as tooth loss and weakened tongue muscle function. These impairments can lead to serious health risks, including aspiration pneumonia, choking, and even mortality. In response, the International Dysphagia Diet Standardisation Initiative (IDDSI, 2019) highlights that foods considered “easy to swallow” should exhibit a soft texture and form a cohesive bolus in the mouth, thereby reducing mechanical irritation to the oropharyngeal tissues [2]. When designing such foods, in addition to texture modification [3], nutritional adequacy must also be considered. Macronutrients such as proteins, carbohydrates, and lipids serve as fundamental components that can be tailored into specific formulations to achieve soft structures and specialized textures—for instance, in the form of gelled solids or semi-solid foods.

Plant protein, as a sustainable resource with wide availability, low cost, and health benefits, holds promise for partially replacing animal protein, thereby reducing energy consumption and pollutant emissions [4]. Among plant proteins, soy protein isolate (SPI) stands out as a key ingredient in composite food systems due to its excellent foaming, emulsifying, film-forming, and gelation properties [5]. In practical applications, SPI often interacts with molecules such as polysaccharides and polyphenols, producing synergistic effects that further enhance its functional performance [6]. Hyaluronic acid (HA), a glycosaminoglycan widely present in animal connective tissues, consists of repeating disaccharide units of D-glucuronic acid and N-acetyl-D-glucosamine linked via alternating β-1,4 and β-1,3 glycosidic bonds [6]. HA can bind to proteins through covalent or non-covalent interactions, thereby contributing to various biological functions such as anti-inflammatory, anti-aging, and antioxidant activities [6]. Studies have shown that protein-polysaccharide composite systems can serve not only as emulsifiers for high internal phase Pickering emulsions (HIPPEs) but also as gelling agents, playing critical roles in both the emulsification and subsequent gelation processes involved in forming high internal phase Pickering emulsion gels (HIPPEGs) [7]. Such gels exhibit an oil phase volume fraction exceeding 74%, forming an adjustable viscoelastic three-dimensional network that combines the advantages of both Pickering emulsions and gels. This robust and printable network allows them to be precisely shaped into customized food structures via 3D printing, offering visually appealing and personalized nutritional solutions for individuals with dysphagia [8]. Furthermore, the network efficiently encapsulates and protects bioactive substances (such as vitamins, polyphenols, or probiotics) and controls their release in the gastrointestinal tract, enabling targeted delivery and enhanced nutritional benefits. However, the network strength and environmental stability of HIPPEGs produced by current preparation methods are often insufficient, resulting in limited functionality and difficulty in fully meeting the stringent requirements for serving as a matrix for customized nutritional foods.

To construct protein-based HIPPEGs with enhanced stability and functionality, physical, chemical, or biological modifications are commonly employed to improve their physicochemical and functional properties, thereby enhancing the overall stability of the composite systems [9]. Current studies have prepared protein-based HIPPEGs through heat-induced gelation of protein particles [10] or simple protein–polysaccharide complexation [11]. However, the structure and performance of proteins are susceptible to factors such as temperature, pH, and biopolymer concentration, often leading to diminished gel functionality. In contrast, protein–polysaccharide conjugates formed via covalent cross‑linking exhibit greater stability and are less affected by environmental conditions [12]. For instance, Li et al. demonstrated that lactoferrin–hyaluronic acid conjugates exhibit superior emulsifying properties and thermal stability compared to their non-covalent counterparts [13]. Furthermore, Yan et al. promoted the covalent binding between tilapia collagen and hyaluronic acid using 1‑ethyl‑3‑(3‑dimethylaminopropyl) carbodiimide (EDC)/N‑hydroxysuccinimide (NHS), confirming that EDC/NHS can activate amino and carboxyl groups to form stable amide bonds, thereby enhancing water‑binding capacity, thermal stability, and gel structural strength [14]. Nevertheless, the efficiency and extent of covalent cross‑linking directly influence the performance of the final product. To further improve the protein/polysaccharide cross‑linking efficiency mediated by EDC/NHS, the introduction of an efficient and economical physical modification technique—high‑intensity ultrasound (HIU) as a pretreatment—has emerged as a promising strategy. Studies have shown that HIU can alter protein structures through intense shear and cavitation effects, exposing more originally buried active groups (–NH_2_ and –COOH) and increasing the number of sites available for covalent binding [15]. Meanwhile, the acoustic streaming induced by HIU can enhance micromixing and mass transfer within the system, accelerating collisions and reactions between cross‑linking agents and functional groups, thereby promoting the construction of a more efficient and uniform cross‑linked network [16]. Combining HIU physical pretreatment with EDC/NHS chemical cross‑linking offers a novel pathway for constructing structurally stable and high‑performance protein–polysaccharide covalent complexes and their subsequent HIPPEGs.

To validate the effectiveness of this synergistic strategy, this study systematically investigates the structural and functional properties of SPI and HA complexes formed under HIU treatment through non-covalent or covalent interactions. Furthermore, it analyzes the formation mechanism, structural characteristics, and functional performance of HIPPEGs constructed from these complexes. The potential applicability of such HIPPEGs as dysphagia-friendly foods was evaluated using the International Dysphagia Diet Standardisation Initiative (IDDSI) framework. To advance this research, the study was expanded to include 3D printing technology, demonstrating the feasibility of HIPPEGs for precise structuring and customization as personalized nutrition carriers. This research is expected to expand the variety of foods suitable for individuals with swallowing difficulties, thereby addressing their specific nutritional requirements.

Materials and methods

2

Materials

2.1

Low temperature defatted soybean meal was obtained from Shandong Yuwang Group. Hyaluronic acid (40–80 kDa) was purchased from Macklin Biochemical Technology Co., Ltd. (Shanghai, China). Soybean oil was acquired from a local supermarket (Harbin, China). All reagents were of analytical grade.

Sample preparation

2.2

A stock solution of soy protein isolate (SPI, 80 mg/mL) and a hyaluronic acid (HA, 10 mg/mL) aqueous solution were prepared and allowed to hydrate overnight. The control sample was prepared by mixing SPI stock solution (80 mg/mL) with deionized water at a 3:1 vol ratio, followed by pH adjustment to approximately 7. For the experimental groups, four distinct procedures were employed:

(1) Mixture: SPI and HA solutions were mixed at a 3:1 vol ratio under stirring for 1 h, followed by dialysis for 6 h and subsequent pH adjustment to 7;

(2) U-Mixture: the same mixing ratio and stirring conditions were applied as in group 1, followed by ultrasonication for 20  min (3  s on, 3  s off; 150  W; equipped with a 10–mm diameter titanium probe immersed in 20  mL of protein dispersion placed in a 50–mL beaker, at a frequency of 20  kHz), dialysis, and pH adjustment;

(3) Conjugate: EDC (26.6 mg) and NHS (16 mg) were added to 5 mL of HA solution and stirred for 30 min to activate HA, which was then mixed with SPI at a 3:1 ratio and stirred for at least 6 h, followed by dialysis and pH adjustment;

(4) U-Conjugate: the same activation and mixing steps as in group 3 were performed, followed by dialysis, pH adjustment, and finally ultrasonication under the same conditions as in group 2 to obtain the final solution.

Determination of degree of grafting (DG)

2.3

DG was quantified via an o-phthalaldehyde (OPA) assay [17]. A fresh OPA reagent was prepared daily by mixing 1 mL of methanolic OPA (40 mg/mL), 50 mL of 0.1 mol/L sodium tetraborate, 5 mL of 20% (w/v) SDS, and 200 μL of β-mercaptoethanol, then bringing the final volume to 100 mL with distilled water. This reagent was stored in amber glass. Prior to measurement, all samples were adjusted to a concentration of 2 mg/mL. For the assay, 200 μL of sample solution was reacted with 4 mL of OPA reagent at 35 °C for 2 min, and the absorbance at 340 nm was recorded using a UV–visible spectrophotometer. A blank consisting of distilled water and OPA reagent was used for baseline correction. Absorbance values for non-glycosylated SPI and glycosylated samples were denoted as A_0_ and A_1_, respectively. DG was then derived according to the following equation:

$$[eqn]$$

SDS-PAGE

2.4

All solutions (diluted to a protein concentration of 2 mg/mL) were mixed with 5 × loading buffer containing SDS and dithiothreitol at a 4:1 (v/v) ratio and heated at 95 °C for 5 min to denature the proteins. Then, 10 μL of each sample was loaded onto an SDS-PAGE gel (12% separating gel with a stacking gel). Electrophoresis was initially performed at a constant voltage of 80 V for the stacking phase. After 15 min, the voltage was increased to 120 V for the separation phase. Finally, the gel was stained with Coomassie Brilliant Blue solution (G-250), destained with distilled water, and images were captured using a gel documentation system [18].

Particle size and zeta-potential

2.5

A Zetasizer 2000 device (Malvern Instrument Co. Ltd. Worcestershire, UK) was used to measure the particle size and zeta-potential of SPI-HA mixtures and conjugates. The samples were diluted with deionized water to a protein concentration of 1 mg/mL prior to measurement.

Transmission electron microscope (TEM) measurements

2.6

The microstructure of the SPI, Mixture, and Conjugate group was examined by TEM (H-7650, Hitachi, Japan). Samples were diluted to a suitable concentration (1 mg/mL) and deposited onto a copper mesh grid. Images were acquired at various magnifications at 100 kV.

Protein flexibility

2.7

A 1 mg/mL trypsin solution was prepared, and the protein sample was diluted to the same concentration with Tris-HCl buffer (pH 8.0). The two solutions were combined at a 16:1 (v/v) ratio and incubated for 10 min, followed by the addition of an equal volume of 10% (w/v) trichloroacetic acid. The supernatant, obtained after centrifugation (5,000 × g, 25 min), was collected for absorbance measurement at 280 nm [19].

Surface hydrophobicity (H0)

2.8

H_0_ was measured using ANS as a fluorescence probe [20]. A 4 mL sample solution (0.5 mg/mL) was mixed with 20 μL of 8 mmol/L ANS. After incubation in the dark for 10 min, fluorescence was measured with an excitation wavelength of 390 nm and an emission range of 400–600 nm, using a slit width of 5 nm.

Circular dichroism (CD) spectrum

2.9

The far-UV CD spectra of the protein samples (0.1 mg/mL) were acquired at 25 °C over 190–260 nm on a Chirascan V100 spectropolarimeter (Applied Photophysics Ltd., England). Following acquisition, the secondary structure composition was calculated using the software (https://dichroweb.cryst.bbk.ac.uk).

Fourier infrared spectroscopy (FTIR)

2.10

The freeze-dried sample (2.0 mg) was first homogenized with 150 mg of KBr and compressed into slice. The FTIR spectrum was subsequently acquired over the 600–4000 cm^−1^ range employing a spectrophotometer (8400S, Shimadzu, Japan).

Ultraviolet–visible spectroscopy

2.11

The ultraviolet–visible (UV–Vis) absorption spectra were acquired between 200 and 500 nm with a UV spectrophotometer. For measurement, the sample solutions were diluted to protein concentration of 0.5 mg/mL.

Intrinsic fluorescence spectrum

2.12

The intrinsic fluorescence and excitation-emission matrix (EEM) spectra of the samples were acquired using a fluorescence spectrophotometer (F-7000, Hicathi, Tokyo, Japan). The intrinsic fluorescence was measured with an excitation wavelength of 280 nm and an emission scan from 290 to 450 nm. The EEM spectra were collected by scanning excitation wavelengths from 200 to 400 nm in 5 nm increments and recording the corresponding emission spectra from 200 to 500 nm [21].

Emulsifying properties

2.13

Emulsion activity index (EAI) and emulsion stability index (ESI) were evaluated as follows. Protein sample solutions (4 mg/mL) were mixed with soybean oil at a 7:3 vol ratio. The mixture was then homogenized using a high-shear mixer (T25, IKA Instrument Equipment Co., Ltd, China) at 12,000 rpm for 1 min. Subsequently, 50 μL of each emulsion was diluted in 5 mL of 0.1% (w/v) SDS solution. Absorbance was measured at 500 nm immediately (A_0_) and after 10 min (A_10_). EAI and ESI were calculated using the equations below [19]:

$$[eqn]$$ $$[eqn]$$

where N is the dilution factor; φ is the oil volume fraction; A_0_ and A_10_ represent the absorbance of the dispersion at 0 and 10 min, respectively; and C is the protein concentration in (g/mL).

Preparation of HIPPEGs

2.14

Quercetin (Que, final concentration 0.1% w/v) was first dissolved in soybean oil to form the quercetin-loaded oil phase. The complex and conjugate solutions served as the aqueous phase. The aqueous phase was mixed with the quercetin-loaded oil phase at a ratio of 25:75 (v/v) and homogenized at 13,000 rpm for 1.5 min to form HIPPEs. The resulting HIPPEs were subsequently converted into HIPPEGs by heating in a water bath at 85 ℃ for 20 min, followed by cooling to room temperature (25℃). The gels were then allowed to stand overnight to stabilize, yielding the final HIPPEGs.

Textural profile analysis (TPA)

2.15

The textural characteristics of the HIPPEGs were evaluated with a texture analyzer (Stable Micro Systems Ltd., Godalming, UK). TPA was conducted by applying two sequential compression cycles to each specimen. The instrument settings were configured as follows: pre-test speed at 2 mm/s, test speed at 2 mm/s, post-test speed at 2 mm/s, trigger force at 5 g, compression distance of 15.0 mm, and a maintained temperature of 25 °C.

CLSM

2.16

First, Nile blue (1%, w/v) and Nile red (1%, w/v) dye solutions were prepared in isopropanol and stored in the dark. Subsequently, 2.5 mL of the sapmle solution was mixed with 40 μL of the Nile blue solution, and 7.5 mL of oil was mixed with 25 μL of the Nile red solution, each being homogenized thoroughly. An emulsion gel was then prepared according to the method described in Section 2.12. Using a 200 μL micropipette tip, approximately 1 mg of the gel was gently collected, placed on a glass slide, compacted, covered with a coverslip, and finally sealed with nail polish.

Rheological properties

2.17

The rheological properties of the gel samples were evaluated using a dynamic shear rheometer (MARS IQ AIR, HAAKE, Germany) with a 35 mm diameter parallel plate geometry. A constant gap of 1 mm and a strain of 1% were maintained throughout the tests. To minimize solvent evaporation, the sample edges were coated with silicone oil. The temperature of the samples was increased from 25 to 85 °C, held at 85 °C for 20 min, and then cooled back to 25 °C at the same rate. During the entire temperature sweep, the loss modulus (G″) and storage modulus (G′) of the emulsion gels were continuously recorded. Shear rate sweeps from 1 to 100 s^−1^ were conducted to determine emulsion gel viscosity. To mimic 3D printing conditions, a three-interval thixotropy test (3ITT) evaluated the material's shear recovery behavior. Each of the three stages lasted 200 s: (i) an initial low shear rate of 1 s^−1^ representing the pre-printing state; (ii) a high shear rate of 100 s^−1^ simulating extrusion during printing; and (iii) a return to 1 s^−1^ to assess recovery after printing. Frequency sweeps between 0.1 and 10 Hz were also performed at 1% strain to measure the G′ and G″ without damaging the network structure [20].

Swelling ratio (SR)

2.18

The gel samples were cut into cylinders of comparable size, immersed in deionized water for 12 h, and the swelling ratio was calculated using the formula [22]:

$$[eqn]$$

where M_0_ and M are the weights of the gel before and after swelling, respectively.

LF-NMR

2.19

LF-NMR (Bruker Optik GmbH, Ettlingen, Germany) was employed to measure the relaxation times (T_2_) and assess the distribution of water in the emulsion gels. The waiting time was 2500 ms and the data was obtained from 12,000 echoes. The Magnetic resonance imaging (MRI) of the samples was acquired using a low-field NMR analyzer with a spin-echo imaging sequence. The resulting grayscale images were processed using OsiriX software [23].

IDDSI tests

2.20

The texture of the prepared HIPPEGs was evaluated against the International Initiative for the Standardization of Dysphagia Diets (IDDSI) framework, which classifies dysphagia-friendly foods into eight levels (0–7). The categorical correspondence of the samples was determined using standard IDDSI tests: the spoon tilt test for cohesiveness and the fork pressure test for deformability [24].

3D printing

2.21

The model was sliced using Slic3r software and processed by the FOODDOT-D1 3D food printer (Shiyin Technologies Co. Ltd., Hangzhou, Zhejiang, China). The printing parameters were as follows: layer height = 1.0  mm, nozzle retraction speed = 15  mm·s^−1^, nozzle travel speed = 15  mm·s^−1^, and infill density = 20%. The model is a rectangular block measuring 15  mm × 15  mm × 8 mm. The printing accuracy was comprehensively evaluated by measuring the top-view projected area of the cubic sample using ImageJ analysis software, combined with measuring its height using a ruler.

In vitro digestion

2.22

Simulated gastrointestinal tract (GIT) digestion

2.22.1

To simulate in vitro gastrointestinal digestion, a modified INFOGEST static method was employed [20]. Initially, 5.0 mL of sample was blended with an equal volume of simulated salivary fluid (SSF, pH 6.8) supplemented with NaCl (1.6 mg/mL), KCl (0.2 mg/mL), and mucin (0.2 mg/mL). This mixture was stirred for 10 min at 37 °C. The gastric phase was then initiated by combining the blend in a 1:1 (v/v) ratio with simulated gastric fluid (SGF, pH 1.2) containing pepsin (3.2 mg/mL) and NaCl (1.8 mg/mL). After pH adjustment to 2.0, the incubation continued for 120 min at 37 °C under shaking (100 rpm). For the intestinal phase, the pH was raised to 7.0, and simulated intestinal fluid (SIF) was incorporated at a 1:2 (v/v) ratio. The SIF consisted of bile salt (20 mg/mL), pancreatin (10 mg/mL), and CaCl_2_ (0.75 mol/L). Digestion was finalized by shaking the system at 37 °C for a further 120 min.

Encapsulation efficiency (EE)

2.22.2

Que was extracted from the emulsion gel using ethyl acetate. Specifically, 1 mL of the Que-loaded emulsion gel was mixed with 3 mL of ethyl acetate and vortexed for 5 min. After vortexing, the mixture was allowed to stand for 30 min to achieve phase separation, and this extraction procedure was repeated three times. The ethyl acetate phase was collected, and its absorbance was measured at 374 nm using an ultraviolet spectrophotometer (T9, PUXI, China). A standard curve for Que was generated by measuring the absorbance of ethyl acetate solutions containing known concentrations of Que in a gradient. The encapsulation efficiency (EE) was calculated as follows:

$$[eqn]$$

Bioaccessibility

2.22.3

To determine quercetin (Que) bioaccessibility, the compound released into the mixed micellar fraction after intestinal digestion was quantified. Following the digestion process, samples underwent centrifugation at 10,000 × g for 20 min at 25 °C, which generated a clear intermediate layer containing the Que-loaded phospholipid-bile salt micelles. Next, 1 mL of this micellar phase was thoroughly mixed with 5 mL of ethyl acetate, and its absorbance was recorded at 374 nm. The final quercetin concentration was calculated using a pre-established standard curve.

$$[eqn]$$

where C_s_ and C_Original_ are the solubilized Que content of the micelle and the amount of Que added to the digestion system, respectively.

Statistical analysis

2.23

Data are presented as mean ± standard deviation. Statistical significance (p < 0.05) between sample means was determined by one-way analysis of variance (ANOVA) using SPSS 26.0 software (SPSS Inc., Evanston, IL, USA).

Results and discussion

3

Particle Size, zeta Potential, and microstructure of mixtures and conjugates

3.1

As illustrated in Fig. 1A, the native SPI exhibited an average particle size of 758.9 ± 213.6 nm, which is consistent with the self-aggregation tendency commonly observed in SPI [25]. The particle size distribution (Fig. 1B) further confirmed that the majority of SPI particles fell within two primary ranges: 150–400 nm and 500–2000 nm. Upon the introduction of HA, the average particle size of the resulting SPI–HA mixture decreased, accompanied by a discernible shift in the distribution profile toward smaller sizes. This reduction can likely be attributed to the inherent negative charges of HA, which enhance electrostatic repulsion within the system [26]. In contrast to native SPI, the conjugates, formed under the catalysis of EDC and NHS, showed a marked decline in average particle size (431.7 ± 121.6 nm), with the majority of particles distributed in the 20–80 nm range. Moreover, when subjected to HIU treatment, both the U-Mixture and U-Conjugate exhibited a pronounced reduction in particle size. This phenomenon is primarily ascribed to the ultrasonic cavitation effect, which disrupts aggregate structures and promotes protein conformational rearrangements, thereby facilitating the formation of finer composite particles.Fig. 1. Average particle size (A), particle size distribution (B), zeta potential (C), and protein flexibility (D), DG (E), and SDS-PAGE (F) of SPI, Mixture, and Conjugate in aqueous solution.

The zeta-potential results provide further evidence for the alterations in the particle size of the complexes (Fig. 1C). Due to the inherent negative charge of HA, its introduction into the SPI system led to enhanced electrostatic repulsion in both the Mixture and Conjugate [13]. The HIU treatment further increased the electronegativity of U-Conjugate, which can be attributed to the disruption of protein structures by ultrasound. Specifically, HIU unfolds protein molecules and disrupts protein aggregates, thereby increasing the net surface charge of proteins [27].

Protein flexibility was evaluated for SPI, the Mixture, and the Conjugate (Fig. 1D). Trypsin susceptibility, attributed to the broad specificity of trypsin toward proteins, was used as an indicator of protein flexibility. The presence of HA, with its abundant hydroxyl and carboxyl groups, induced structural changes in the proteins, resulting in increased flexibility in both the Mixture and Conjugate [28]. Furthermore, the shear and cavitation effects generated by HIU significantly increased the number of potential trypsin-binding sites on the protein surface. It has been reported that proteins with greater flexibility exhibit enhanced interfacial adsorption capacity [19] The trend in particle size variation was further corroborated by TEM images (Fig. 2). In particular, U-Conjugate, characterized by its smaller particle size and stronger electrostatic repulsion, is likely to possess enhanced stability, superior interfacial properties, and improved functionality.Fig. 2TEM images of SPI, Mixture, and Conjugate aqueous solutions at different magnifications. Scale bars: 5 μm for the 2000× images; 2 μm for the 5000× images.

DG and SDS-PAGE

3.2

DG was quantified via an OPA assay (Fig. 1E). Compared to the conjugate prepared without ultrasonication (15.1 ± 1.3%), the cross-linking degree of the ultrasonication-assisted EDC/NHS-prepared conjugate (31.4 ± 4.6%) was significantly enhanced. Appropriate ultrasonication generates shear forces and induces free radical formation, promoting conformational changes in proteins and polysaccharides, which exposes more reactive groups and thereby facilitates cross-linking reactions [17]. In addition, the local thermal effect induced by ultrasonic cavitation helps reduce the reaction energy barrier, increases the collision frequency between protein and polysaccharide molecules, and synergistically improves the grafting efficiency [29].

SDS-PAGE is a reliable method for detecting the formation of covalent bonds between proteins and polysaccharides. As shown in the SDS-PAGE profiles of complexes and conjugates in Fig. 1F, the bands of SPI, Mixture, and U-Mixture at the positions of the 7S (50–71 kDa) and 11S (20–35 kDa) subunits are essentially consistent, indicating that the interactions between proteins and polysaccharides in Mixture and U-Mixture are primarily mediated by non-covalent interactions [18]. In contrast, Conjugate and U-conjugate exhibit distinct high-molecular-weight bands above 71 kDa, indicating the formation of macromolecular polymers. These results further confirm that, under EDC/NHS mediation, proteins and polysaccharides form conjugates and aggregates through covalent cross-linking [29].

Structural Characterization of mixtures and conjugates using Multispectral Techniques

3.3

UV–Vis spectroscopy was employed to investigate structural modifications of proteins and to analyze interactions between proteins and polysaccharides. As illustrated in Fig. 3A, the conjugated double bonds of aromatic amino acids in SPI exhibited an absorption peak near 280 nm [30]. The formation of both complexes and conjugates, along with the application of HIU, enhanced the energy required for π → π* transitions, which was reflected by an increase in UV absorption intensity [31]. The presence of HA induced an extension of the SPI peptide chains, leading to conformational changes in SPI. This structural rearrangement altered the microenvironments surrounding aromatic amino acid residues, thereby increasing the absorbance.Fig. 3UV-Vis spectra (A), FTIR spectra (B), surface hydrophobicity (C), intrinsic fluorescence spectra (D), and fluorescence excitation-emission matrix (EEM) spectra (E) of SPI, Mixture, and Conjugate in aqueous solution.

FT-IR serves as a powerful tool for investigating protein/polysaccharide interactions and associated structural changes. As shown in Fig. 3B, the FT-IR spectrum of SPI exhibits characteristic absorption bands in the following regions: the Amide A band (3000–3700 cm^−1^) is attributed to hydrogen bonding and N–H stretching; the Amide III band (1200–1450 cm^−1^) is associated with C–N stretching and N–H bending; the Amide II band (1500–1550 cm^−1^) corresponds to N–H deformation; while the Amide I band (1600–1700 cm^−1^) primarily arises from C

<svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="20.666667pt" height="16.000000pt" viewBox="0 0 20.666667 16.000000" preserveAspectRatio="xMidYMid meet"><metadata> Created by potrace 1.16, written by Peter Selinger 2001-2019 </metadata><g transform="translate(1.000000,15.000000) scale(0.019444,-0.019444)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z M0 280 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z"/></g></svg>

O stretching vibrations. Compared to SPI, the conjugate exhibited a significantly intensified and lower-wavenumber-shifted amide I band (from 1651.2 to 1627.6 cm^−1^) in its FTIR spectrum, indicating a substantial contribution from newly formed amide bonds to the CO stretching vibration [26]. A concurrent enhancement in the absorption of the amide II band further corroborates the formation of these amide linkages. Additionally, a blue shift along with altered intensity is observed in the amide A band of the Conjugate. These changes can be attributed to the EDC/NHS-mediated enhancement of protein–polysaccharide molecular interactions, which significantly increases both the hydroxyl content and hydrogen bonding within the conjugate [14].

Surface hydrophobicity is a key parameter for evaluating the number of hydrophobic groups exposed on a protein's surface, commonly employed to monitor structural changes that influence functional properties such as surface tension and emulsifying activity. As illustrated in Fig. 3C, HIU treatment, as well as the formation of complexes and conjugates, progressively increased the surface hydrophobicity of SPI, with U-Conjugate exhibiting the highest value. Such alterations may stem from interactions between HA and amino acid residues on SPI polypeptide chains, which modify the protein’s structural conformation and expose previously buried hydrophobic regions [32]. In a related context, Wang et al. also reported that while high concentrations of HA may introduce excessive hydrophilic groups—leading to a reduction in surface hydrophobicity of whey protein isolate—low concentrations of HA (<0.2%) can partially hinder ANS binding while sufficiently expanding the protein structure, thereby promoting hydrophobic group exposure and enhancing surface hydrophobicity [26].

The intrinsic fluorescence of proteins primarily arises from three aromatic amino acids—tyrosine (Tyr), tryptophan (Trp), and phenylalanine (Phe)—which contain benzene rings or conjugated double-bond systems. As a result, fluorescence spectroscopy is widely employed to monitor conformational changes and structural dynamics within proteins. As shown in Fig. 3D, compared with SPI, the Mixture exhibited a decrease in fluorescence intensity without a shift in the emission wavelength maximum. This suggests that HA quenches protein fluorescence through steric hindrance and electrostatic interactions, partially masking the intrinsic chromophores [26]. Furthermore, HIU treatment altered the tertiary structure of the protein, leading to a further reduction in fluorescence intensity in the Mixture, a result consistent with previous findings by Huang et al. [6]. When SPI and HA were cross-linked via amide bonds under EDC/NHS catalysis, the Conjugate showed a significant decrease in fluorescence intensity. A redshift in the maximum emission wavelength from 330 nm to 335 nm was also observed. This shift likely reflects structural reorganization within the Conjugate, where hydrophobic amino acid residues become exposed to a more hydrophilic microenvironment [33]. These results imply that covalent grafting of HA combined with HIU treatment promotes the unfolding and extension of polypeptide chains. The quenching mechanism may involve several processes, including ground-state complex formation, collisional quenching, structural rearrangement, and energy transfer [28]. Further evidence of structural changes in SPI was provided by EEM spectra (Fig. 3E). Peak I, associated with Tyr and Trp residues, originates from π → π* transitions and reflects the polarity of the local microenvironment [21]. A notable quenching of Peak I was observed in the Conjugate, indicating that HA may interact with SPI and induce conformational changes—consistent with the fluorescence quenching results discussed above.

CD spectroscopy serves as a sensitive probe for conformational changes in proteins and is widely utilized to characterize alterations in secondary structure. As illustrated in Fig. 4A, native SPI exhibits the following secondary structure composition: 34.5 ± 1.4% random coil, 13.6 ± 2.3% α-helix, 20.0 ± 0.4% β-turn, and 31.8 ± 3.2% β-sheet. In both the Mixture and Conjugate, an increase in α-helix content was observed, accompanied by a decrease in β-sheet and random coil structures. These changes are likely attributable to modifications in spatial conformation driven by the formation of hydrogen and covalent bonds within the system [34]. Furthermore, upon HIU treatment, a noticeable shift from α-helix to β-sheet structures was detected. This transition may result from the disruption of stabilizing interactions within the Mixture and Conjugate, leading to structural loosening and a reduction in the overall content of ordered protein motifs [32].Fig. 4. Secondary structure content (A), emulsifying activity index (EAI, B), emulsifying stability index (ESI, C) of SPI, Mixture, and Conjugate aqueous solutions; MRI images (D), LF-NMR (E), and peak area (F) of the formed HIPPEGs.

Emulsifying Property analysis

3.4

The EAI reflects the initial ability of proteins and their complexes to form emulsions, whereas the ESI indicates their capacity to maintain stability over time. It has been documented that native SPI exhibits emulsifying activity due to its amphiphilic nature and ability to adsorb at the oil–water (O/W) interface [35]. As illustrated in Fig. 4B and C, both the EAI and ESI of the conjugate were significantly enhanced compared to those of SPI. This suggests that the covalent conjugation between SPI and HA enhances interfacial adsorption behavior and improves the stability of the resulting emulsion. The improved emulsion stability observed in the conjugate can be attributed to a reduction in droplet size and an increase in electrostatic repulsion [35]. On one hand, smaller droplet sizes increase the total interfacial area available for stabilization, contributing to the formation of a denser and more cohesive emulsion system. On the other hand, the elevated absolute zeta potential strengthens electrostatic repulsion among droplets, thereby suppressing aggregation and promoting overall system stability. Furthermore, previous studies have elaborated on the mechanisms underlying protein–polysaccharide interactions, including both complexation and conjugation, and emphasized their critical roles in emulsion formation [12]. These interactions contribute to stability by forming a physical barrier that prevents molecular proximity, thereby inhibiting aggregation through steric hindrance and restricted mobility, ultimately helping to maintain system integrity.

LF-NMR and MRI analysis

3.5

Patients with dysphagia often experience reduced saliva secretion, making it necessary to consume foods with high water-holding capacity [23]. To further investigate the moisture distribution within the HIPPEGs, MRI imaging was performed. As illustrated in Fig. 4D, the red regions in the pseudo-color images correspond to the distribution of hydrogen protons (water) in the gels, indicating favorable water retention properties across all samples [36]. The incorporation of HA, which is rich in hydroxyl groups, enhances the interaction between the gel matrix and water molecules [37]. Moreover, the HIPPEGs based on Conjugate and U-Conjugate exhibited a more homogeneous moisture distribution. This phenomenon may be attributed to the role of Conjugate in promoting the formation of a denser gel network, thereby facilitating more efficient entrapment of water molecules [38].

The formation of a gel network structure typically inhibits the mobility of water molecules. Therefore, this study employed LF-NMR to analyze the migration behavior of water in different HIPPEGs. As shown in Fig. 4E, the transverse relaxation time (T_2_) of each HIPPEG group is presented. The T_2_ spectrum generally consists of three relaxation components: T_21_ (1–10 ms) is attributed to bound water tightly associated with the gel macromolecules; T_22_ (30–100 ms) corresponds to immobilized water with restricted mobility within the gel network; and T_23_ (>100 ms) represents free water that can move freely outside the network [39]. Compared to SPI-based HIPPEGs, the free water peak (T_23_) in gels based on Conjugate and U-Conjugate showed a significant shift toward the T_22_ direction, indicating a transformation of free water into immobilized water. Quantitative analysis results (Fig. 4F) further demonstrated that compared to SPI-HIPPEGs, the proportions of immobilized water (P_22_) and bound water (P_21_) increased in the Conjugate group. This suggests that covalent grafting of HA enhances the interaction between SPI and water molecules, thereby reducing water molecule mobility. Additionally, compared to the Mixture and Conjugate groups, the P_21_ and P_22_ values further increased in the ultrasonically treated U-Mixture and U-Conjugate samples, indicating that HIU treatment promotes the formation of a denser gel network, thereby enhancing the ability to immobilize water.

TPA

3.6

TPA was employed to characterize the textural properties of the HIPPEGs. All samples exhibited a soft gel-like texture (Fig. 5A–C). In comparison to SPI-based HIPPEGs, those formulated with Mixture and Conjugate demonstrated increased hardness. This enhancement can be attributed to the more robust three-dimensional gel network formed through non-covalent or covalent interactions between SPI and HA, which contributes to greater structural firmness. A higher hardness in emulsion gels generally indicates a denser gel microstructure [31]. Adhesiveness is associated with the effort required by the tongue to propel the food bolus through the pharynx and is linked to the risk of obstruction. HIPPEGs prepared from Mixture and Conjugate showed lower adhesion compared to SPI-based gels, suggesting easier detachment from oral surfaces such as the palate or teeth, a characteristic beneficial for individuals with dysphagia [40]. The incorporation of HA also influenced the chewiness of SPI-based HIPPEGs. Emulsion gels with higher chewiness may be particularly suitable for edentulous elderly individuals or postoperative patients undergoing swallowing rehabilitation, as such textures can facilitate safe and controlled oral processing [41].Fig. 5. Hardness (A), adhesiveness (B), chewiness (C) from texture profile analysis (TPA), swelling ratio (D), and the gelation process (E-I) of HIPPEGs based on SPI, Mixture, and Conjugate.

Swelling properties

3.7

Swelling refers to the volumetric expansion of gels upon solvent absorption, with the equilibrium state determined by the balance between the elastic retractive force of the polymer network and the swelling force driven by osmotic pressure. As illustrated in Fig. 5D, SPI-based HIPPEGs exhibited the highest swelling ratio (SR) at 36.0 ± 0.7%, which can be attributed to their relatively loose and macroporous three-dimensional network. This structural configuration facilitates rapid and extensive water penetration, leading to a high degree of swelling at equilibrium. In contrast, HIPPEGs based on either Mixture or Conjugate exhibited a reduction in their SR. This suggests that HA promotes the formation of a more compact network structure. While such a structure may enhance water retention capacity (i.e., reduced water loss after swelling), it more significantly restricts the initial penetration and ingress of water molecules, thereby lowering the overall extent of swelling [20]. Previous studies have reported that the presence of smaller oil droplets can contribute to the formation of a tight network, further impeding water permeation [22].

Evolution of the gelation process

3.8

The gelation behavior of HIPPEGs stabilized by SPI, Mixture, and Conjugate was systematically investigated using rheological analysis. As shown in Fig. 5E–I, no distinct gel point (i.e., crossover of storage modulus (G′) and loss modulus (G″)) was observed during the initial stage, indicating that gelation had already been initiated in all samples under the combined effects of shear and thermal treatment during homogenization. This phenomenon can be attributed to heat-induced protein denaturation, which exposes hydrophobic amino acid residues and promotes the rearrangement of intermolecular interactions—such as disulfide bonds, hydrogen bonding, and hydrophobic associations—thereby driving the formation of a three-dimensional gel network [42]. Throughout the temperature ramp (25–85 °C), holding, and cooling (85–25 °C) stages, both the G′ and G″ of all samples increased markedly, reflecting progressive strengthening of the gel structure and development of the network. Specifically, after approximately 30 min of thermal gelation, SPI-stabilized HIPPEGs reached final G′ and G″ values of 4187.9 and 602.9 Pa, respectively. In contrast, HIPPEGs stabilized by U-Conjugate exhibited significantly higher values, with G′ and G″ reaching approximately 8389.2 and 1117.1 Pa under the same conditions, indicating superior gel strength. These results suggest that the formation of U-Conjugate facilitates the development of a more robust gel architecture, likely due to structural modifications in the protein that enhance the contribution of hydrophobic interactions and hydrogen bonding to the formation and stabilization of the gel matrix [43].

Swallowability of HIPPEGs

3.9

By providing a common language for food and liquid characteristics, the IDDSI framework enables the systematic classification of diets based on swallowing needs [28]. Typically, a safe particle size for swallowing is considered to be between 2–4 mm, which corresponds to the gap between the prongs of a standard fork (4 mm), thereby providing a practical reference for evaluating food particle dimensions [44]. During the fork pressure test, the force exerted by the tongue during swallowing is approximately 17 kPa, comparable to the pressure applied by a fingertip [45]. All HIPPEGs demonstrated ease of crushing and deformation under minimal pressure—insufficient to cause blanching of the thumb. They exhibited no surface clumping, formed distinct patterns, and did not regain their original shape after fork removal (Fig. 6A). This indicates that all samples can be effortlessly divided into small portions under gentle pressure, making them suitable for mastication by individuals with dysphagia and easily manageable with utensils such as forks, spoons, or chopsticks [39]. However, SPI-stabilized HIPPEGs showed a tendency to flow between the fork prongs and formed short tails at the ends, suggesting compromised swallowability. This behavior may be attributed to the lower hardness and storage modulus (G′) of the emulsion gel, resulting in a looser network structure. Such texture properties could elevate the risk of choking during ingestion, and individuals with reduced tongue strength may find it difficult to propel the bolus effectively toward the pharynx. In contrast, HIPPEGs based on Mixture and Conjugate exhibited sufficient cohesiveness to accumulate on and maintain their shape over forks and spoons. With a slight flick, they detached easily from the utensil. Overall, except for SPI-HIPPEGs which showed some residue in the spoon tilt test and did not meet the Level 4 standard applicable to patients with dysphagia, all other samples satisfied the Level 5 criteria (i.e., a moist, easily breakable food texture) [45]. This characteristic aligns closely with the requirements for dysphagia-friendly diets, significantly reducing the risk of food adhering to the oral or pharyngeal mucosa and thereby enhancing the safety of the swallowing process [39].Fig. 6IDDSI tests (fork pressure and spoon tilt tests) (A), shear rate sweep (B), frequency sweep (C), 3ITT (D), and 3D printing (E) tests of HIPPEGs.

CLSM images of HIPPEGs stabilized by SPI and U-Conjugate are presented in Fig. S1. Protein-rich regions are shown in green, while oil-rich regions appear red. The control sample (SPI-HIPPEGs) displayed larger oil droplets, notable aggregation, and wider interstitial spaces within the protein network. This may be due to the insufficient robustness of the interfacial film formed by SPI, coupled with weak electrostatic repulsion among protein molecules, promoting droplet coalescence and aggregation, ultimately leading to larger droplet sizes [46]. In contrast, U-Conjugate-stabilized HIPPEGs exhibited a more homogeneous microstructure and uniform oil droplet distribution. This improvement can be attributed to structural modifications in the protein induced by HA via hydrogen bonding, along with increased exposure of hydrophobic groups. These changes enhance the diffusion of SPI at the oil–water interface and promote the formation of a stable interfacial film [38]. Additionally, strengthened electrostatic repulsion likely suppressed droplet aggregation, thereby contributing to the development of a denser network structure.

Rheological properties and 3D printing of HIPPEGs

3.10

By measuring the rheological properties of the prepared HIPPEGs, including apparent viscosity, viscoelasticity, and shear recovery behavior, this study systematically simulates and evaluates their performance during the three stages of 3D printing: extrusion, recovery, and self-support [47].

Shear rate sweep tests were employed to characterize the rheological behavior of the HIPPEGs during the extrusion stage. An ideal 3D printing food material should exhibit shear-thinning properties to ensure smooth passage through the printing nozzle under shear. As shown in Fig. 6B, when the shear rate increased from 0.1 to 100 s^−1^, the apparent viscosity of all samples continuously decreased, indicating favorable shear-thinning behavior in these emulsion gels. This phenomenon may be attributed to the disruption of the gel structure under shear forces and the subsequent realignment of molecules along the flow direction. Furthermore, high viscosity contributes to supporting the layer-by-layer deposition structure during the printing process, thereby enhancing the printability of the material [39]. Compared to the HIPPEGs stabilized by SPI, those stabilized by U-Conjugates exhibited higher apparent viscosity. This increase can likely be attributed to the HIU treatment and the incorporation of HA, which improved the microstructure of the emulsion gels, enhanced intermolecular interactions during gelation, and ultimately led to the formation of a denser and more homogeneous network [48].

During the self-supporting stage, the viscoelastic properties and sufficient mechanical strength (primarily reflected by the G′) of the material are crucial for maintaining the shape fidelity of the printed structure [49]. The mechanical properties of 3D printing materials can be assessed using the G′ and G″ obtained from frequency sweep tests. As illustrated in Fig. 6C, the G′ values of all HIPPEG samples were significantly higher than their corresponding G″ values across the entire measured frequency range. This behavior indicates a typical solid-like elastic response of the system, providing further evidence for the formation of a highly cross-linked network structure within the gels [44]. It is noteworthy that compared to the HIPPEGs stabilized by SPI, those stabilized by U-Conjugates exhibited higher values for both the G′ and G″. This result suggests that the combined effect of HIU treatment and HA incorporation facilitated the formation of a stronger and more densely packed intermolecular network, thereby enhancing the rigidity of the printed structure. Adequate mechanical strength assists the material in better maintaining shape stability during the self-supporting stage, which is crucial for ensuring the structural integrity of the final printed product.

The rheological behavior of HIPPEGs during the recovery stage was evaluated via 3ITT. As shown in Fig. 6D, both HIU treatment and HA incorporation significantly increased the apparent viscosity of HIPPEGs during the first interval. When the shear rate was raised to 100 s^−1^, the aligned structure of emulsion droplets within the gels was disrupted, leading to a pronounced decrease in apparent viscosity. This stage simulates the transient response of the material as it is extruded through the printing nozzle. Subsequently, upon the restoration of the shear rate to 1 s^−1^, all HIPPEGs exhibited partial viscosity recovery, indicating that the gel structure was capable of rebuilding after extrusion. This structural recovery thereby provides the necessary mechanical strength to support subsequent layer deposition [39]. It is noteworthy that the viscosity in the third stage was significantly lower than that in the initial stage. This decrease may be attributed to the partially irreversible damage inflicted on the emulsion gel network by the cyclic shear process. Compared to the HIPPEGs stabilized by SPI, those stabilized by U-Conjugates demonstrated superior network recovery capability and greater resistance to external shear perturbations. Consequently, they were able to return to a relatively stable state more rapidly after structural disturbance. Insufficient post-printing viscosity, conversely, can readily lead to the collapse of the 3D-printed structure.

An ideal 3D-printed product should exhibit clear contours, uniform lines, a smooth surface, and structural stability. A comparative analysis of the actual printing outcomes for HIPPEGs stabilized by SPI and U-Conjugates (Fig. 6E) reveals distinct advantages for the latter. In single-layer printing, HIPPEGs stabilized by U-Conjugates produced finer lines and more defined square boundaries, benefits attributable to their favorable shear-thinning behavior and superior viscosity recovery capability. During the layer-by-layer deposition of 3D structures, models printed from this system showed uniform interlayer consistency and well-defined geometry, with no significant collapse or deformation. The printing accuracy of HIPPEGs was evaluated by measuring the height in the front view and the area in the top view of the printed objects (Fig. S2). The original model had a height of 8 mm and a top-view area of 2.25 cm^2^. The experimental results showed that the SPI-based HIPPEGs prints had a front-view height of 6.83 ± 0.47 mm and a top-view area of 2.66 ± 0.18 cm^2^, while the U-Conjugate-based HIPPEGs prints exhibited a corresponding height of 7.94 ± 0.05 mm and an area of 2.28 ± 0.03 cm^2^. The observed differences can be attributed to the insufficient mechanical properties of SPI-HIPPEGs: their lower mechanical strength leads to an increased extrusion line width, while the accumulation of gravitational effects during layer-by-layer deposition results in reduced structural support, thereby causing a decrease in printed height. These results collectively indicate that HIPPEGs stabilized by U-Conjugates possess excellent suitability for 3D printing and can meet the structural stability requirements for high-precision food printing applications.

Bioactive substance delivery

3.11

The in vitro digestion characteristics of HIPPEGs containing quercetin were investigated. Previous studies have indicated that HIPPEGs stabilized by U-Conjugates possess desirable stability and application potential, thus this system was selected for subsequent simulated gastrointestinal digestion. During the gastric digestion phase, the microstructure of U-Conjugate-stabilized HIPPEGs remained largely intact, demonstrating good stability (Fig. 7A). This phenomenon may be attributed to their dense network structure, which shields the reactive sites of proteins from hydrolysis by pepsin, thereby enhancing the structural retention capacity in the gastric environment [50]. In addition, the covalent binding of HA increased the electrostatic repulsion in the HIPPEGs system, which may have further suppressed pepsin activity through an electrostatic shielding effect, thereby hindering its adsorption and hydrolytic activity at the oil–water interface [51]. In contrast, SPI-stabilized HIPPEGs exhibited the formation of large oil droplets after gastric digestion. This may be due to the combined effects of the high ionic strength of the simulated gastric fluid and the hydrolytic action of pepsin, which disrupted the gel network and consequently compromised the stability of the emulsion gel [20]. Upon entering the intestinal digestion phase, the droplet structures of both emulsion gels were largely disrupted in the weakly alkaline intestinal fluid, with no discernible oil droplets observed in the micrographs, indicating that the intestinal environment significantly compromised the gel network (Fig. 7A). The overall morphological changes of the emulsion gels in simulated digestive fluids are depicted in Fig. 7B; they maintained a gel state in the gastric fluid but underwent structural disintegration in the intestinal fluid. Encapsulation efficiency measurements revealed that both HIPPEG systems exhibited high encapsulation efficiency for quercetin (Fig. 7C).Fig. 7. Optical microscopy images (A), general appearance (B), encapsulation efficiency (EE) (C), and bioaccessibility (D) of SPI and U-Conjugate HIPPEGs.

The bioaccessibility of encapsulated lipophilic bioactive compounds is a critical metric for evaluating the efficacy of emulsion gels as delivery systems. As shown in Fig. 7D, the bioaccessibility of quercetin was significantly higher in HIPPEGs stabilized by U-Conjugates compared to those stabilized by SPI. This enhancement may be attributed to the HIU treatment and HA incorporation, which reduced the extent of oil droplet aggregation during digestion, resulting in the formation of smaller droplets. The increased specific surface area facilitated greater contact between pancreatic lipases and the lipid interface, thereby enhancing lipolysis efficiency and promoting the release of quercetin [3].

Conclusion

4

This study confirms that the SPI–HA conjugate, prepared via HIU treatment, is a highly effective stabilizer for HIPPEGs and is particularly suitable for dietary design in dysphagia patients. The successful formation of the SPI/HA conjugate was confirmed through SDS-PAGE analysis. Furthermore, the results indicated that ultrasonic treatment significantly enhanced the degree of grafting. Structural analysis indicated that the synergistic effect of covalent bonding and ultrasonication significantly reduced the average particle size of SPI, endowing the conjugate with a higher surface negative charge, increased surface hydrophobicity, and enhanced protein flexibility due to structural unfolding. These structural improvements directly contributed to its superior functional performance: HIPPEGs stabilized by the U-Conjugate exhibited markedly enhanced gel strength after a set gelation time, with a storage modulus (G′) of approximately 8389.2 Pa—nearly twice that of conventional SPI-based HIPPE-Gs (4187.9 Pa). Regarding swallowing safety, the gels deformed and fractured readily under mild pressure with low adhesiveness, allowing them to pass the IDDSI tests successfully and effectively mitigate swallowing risks. Furthermore, analyses using LF-NMR, CLSM, and swelling tests revealed a homogeneous internal moisture distribution and a dense network structure, which ensured excellent water retention and physical stability, thereby facilitating safe swallowing. Additionally, U-Conjugate-stabilized HIPPEGs demonstrated potential for in vitro quercetin delivery and suitability for 3D printing applications. In summary, this study provides a solid theoretical foundation and an efficient material solution for developing functional foods with desirable texture and enhanced swallowing safety.

CRediT authorship contribution statement

Yang Wang: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Zhanqiang Ma: Methodology, Investigation. Yueru Liu: Methodology, Investigation. Yuetong Gong: Software, 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.