Effect of Fibril Entanglement on Pickering Emulsions Stabilized by Whey Protein Fibrils for Nobiletin Delivery

The aim of the study was to investigate the effects of whey protein isolate (WPI) fibrils entanglement on the stability and loading capacity of WPI fibrils-stabilized Pickering emulsion. The results of rheology and small-angle X-ray scattering (SAXS) showed the overlap concentration (C*) of WPI fibrils was around 0.5 wt.%. When the concentration was higher than C*, the fibrils became compact and entangled in solution due to a small cross-sectional radius of gyration value (1.18 nm). The interfacial behavior was evaluated by interfacial adsorption and confocal laser scanning microscopy (CLSM). As the fibril concentration increased from 0.1 wt.% to 1.25 wt.%, faster adsorption kinetics (from 0.13 to 0.21) and lower interfacial tension (from 11.85 mN/m to 10.34 mN/m) were achieved. CLSM results showed that WPI fibrils can effectively absorb on the surface of oil droplets. Finally, the microstructure and in vitro lipolysis were used to evaluate the effect of fibrils entanglement on the stability of emulsion and bioaccessibility of nobiletin. At C* concentration, WPI fibrils-stabilized Pickering emulsions exhibited excellent long-term stability and were also stable at various pHs (2.0–7.0) and ionic strengths (0–200 mM). WPI fibrils-stabilized Pickering emulsions after loading nobiletin remained stable, and in vitro digestion showed that these Pickering emulsions could significantly improve the extent of lipolysis (from 36% to 49%) and nobiletin bioaccessibility (21.9% to 62.5%). This study could provide new insight into the fabrication of food-grade Pickering emulsion with good nutraceutical protection.


Introduction
Pickering emulsions have been widely applied in foods, pharmaceuticals, and cosmetics due to their outstanding characteristics, such as extraordinary stability, controlled release of components, and targeted nutraceutical delivery [1][2][3]. The formation and stability mechanisms of Pickering emulsions are affected by some factors, such as particle concentration, oil fraction, pH, ionic strength, and temperature [4][5][6].
The particle concentration is one of the most important and easily adjusted factors during Pickering emulsion formation [7]. In general, droplet size decreases with increasing particle concentration (at a fixed oil-to-water ratio) until it reaches a plateau level. This effect is attributed to two factors: (1) more particles are available to cover the oil droplet surfaces with concentration increasing; (2) sufficient particles cover all the oil droplet surfaces against re-coalescence [8]. The droplet sizes of emulsion also play an important role in emulsion stability due to the lower interfacial layer caused by large droplet sizes [9]. Feng et al. reported that the apparent viscosity and modulus, and stability of food-grade gelatin nanoparticle-stabilized Pickering emulsions were improved by increasing the nanoparticle concentration [7]. Gao et al. found that Pickering emulsions stabilized

Rheology Analysis
The viscosity of WPI fibrils was measured by a rheometer (Malvern Panalytical Limited, Malvern, England, UK) using a 60 mm cone and plate, with a shear rate ranging from 0.01 to 1000 s −1 . All samples were measured at 25 • C. The obtained viscosity, as a function of shear rate data, was fitted to Cross models (Equation (1)) to obtain the relevant parameters [26]. The shape and curvature of a flow curve can be described through this model and the behavior at unmeasured shear rates can be predicted as well.
where η 0 is the zero-shear viscosity; η ∞ is the infinite shear viscosity; K is the cross constant, which is indicative of the onset of shear-thinning; . γ is the shear rate or strain rate; and m is the shear thinning index, which ranges from 0 (Newtonian) to 1 (infinitely shear thinning).

SAXS Analysis
The samples were prepared with a series of WPI fibril concentrations (0.1, 0.25, 0.5, 1.0, 1.25 wt.%). A detector located 2631 mm from the sample was set to collect the scattering intensity, and the wavelength of X-ray radiation was 1.033 Å. The final scattering data were obtained by averaging 20 curves, with the scattering vector Q ranging from 1.0 × 10 −2 to 4.1 × 10 −1 Å −1 .

Dilatational Interfacial Rheology
The change in the interface tension (mNm −1 ) at the oil-water interface with adsorption time (t) was measured by droplet shape analysis (DSA30S, KRUSS, Hamburg, Germany). The WPI fibril solution was collected in a syringe with a diameter of 1.8 mm. The needle tip was immersed in a rectangular glass tank filled with soybean oil. The interfacial tension value was recorded at 1 frame/s for the first 1 h and 0.1 frame/min thereafter, and the value was monitored continuously for 10,800 s. All measurements were performed at 25 • C.

Preparation of Nobiletin-Loaded Pickering Emulsions Stabilized by WPI Fibrils
Nobiletin was dissolved in soybean oil by heating (100 • C for 1 h) in an oil bath to a final concentration of 0.5 wt.%. The nobiletin-loaded Pickering emulsion was prepared as described above.

Characterization of WPI Fibril-Stabilized Pickering Emulsion Confocal Laser Scanning Microscopy
Confocal laser scanning microscopy (FV1200, Olympus, Tokyo, Japan) was performed to confirm the formation of Pickering emulsions. Prior to confocal laser scanning microscopy analysis, samples were dyed by ThT. Images were obtained using a 20× magnification lens at an excitation wavelength of 488 nm. ImageJ software was used to estimate the mean droplet diameters.

Optical Microscopy
The microstructure images of the Pickering emulsion stabilized by WPI fibrils were obtained using an optical microscope (CX40, Sunny Optical Technology Co., Ltd., Yuyao, China) equipped with a camera. Thirty microliters of Pickering emulsion were deposited on the glass slide and covered with a coverslip. All samples were measured with a Freshly prepared samples of Pickering emulsions were diluted 10 times using deionized water, and the pH was adjusted to pH 2.0, 3.0, 4.0, 5.0, 6.0, and 7.0, respectively, with 0.1 M NaOH solutions.

Effect of Ionic Strength
Freshly prepared samples of Pickering emulsions were diluted 10 times using 0 mM to 200 mM NaCl solutions.

Effect of Storage Time
After the preparation of emulsions, the freshly prepared sample emulsions were stored at room temperature. The microstructure of the emulsions was determined at regular storage periods (1, 7, and 28 d). This study used an in vitro model consisting of oral, gastric, and intestine phases slightly modified from Brodkorb et al. [27]. Salivary amylase was not added in the oral digestion phase in this study. Fresh emulsions were prepared with an oil fraction of 50% (w/w). The samples, including soybean oil and WPI fibril-stabilized Pickering emulsion containing 0.25 g of oil, were mixed with 4 mL of simulated salivary fluid (SSF), 0.025 mL of 0.3 M CaCl 2 , and 0.225 mL of pure water for 2 min. Eight milliliters of simulated gastric fluid (SGF), 0.005 mL of 0.3 M CaCl 2 , 0.4 mL of 5 M HCl, and 0.448 mL of pure water were added to oral digestive juice successively, and freshly dissolved pepsin was added to achieve enzyme activity of 2000 U/mL and to initiate the gastric digestion process. Subsequently, the reaction mixture was incubated under continuous stirring in a temperature-controlled oil bath (37.0 ± 0.1 • C) for 2 h. Eight milliliters of simulated intestinal fluid (SIF), 0.04 mL of 0.3 M CaCl 2 , 10 mM bile salt, and 3.16 mL of pure water were added to the gastric digesta, and the pH was adjusted to 7.0 with the addition of NaOH. Pancreatin was added to gastric digesta to obtain an enzyme activity of 100 U/mL to start intestinal digestion. The mixture was incubated under continuous agitation in an oil bath (37.0 ± 0.1 • C) for 2 h, and 0.1 M NaOH was added manually to maintain the pH at 7.0 during lipolysis. The volume of added NaOH solution was recorded over time during the intestinal digestion. Lipolysis in samples was characterized by the release of free fatty acids (FFA). The fraction of FFA released was calculated as follows: where M lipid is the molecular mass of the triacylglycerol oil (in g/mol); V NaOH is the volume of NaOH solution used to neutralize the released FFA (in L); m NaOH is the molarity of NaOH solution (in mol/L); and W lipid is the total mass of the initial triacylglycerol oil. The molecular mass of soybean oil was 876.56 g/mol.

High-Performance Liquid Chromatography (HPLC) Analysis of Nobiletin
The digests were collected after gastrointestinal digestion and centrifuged at 10,000× g for 40 min. The clear micelle phase was collected, and the nobiletin content in the micelle phase was determined by an UltiMate 3000 HPLC system (Dionex, Sunnyvale, CA, USA), with a SunFire C18 column (150 mm × 4.6 mm, 5 µm). The HPLC mobile phase consisted of (A) acetonitrile and (B) water. The elution conditions were as follows: 45% (A) and Foods 2022, 11, 1626 5 of 13 55% (B), and the running time was 15 min. The detection wavelength was 333 nm, and the concentration of nobiletin was determined using a standard curve of nobiletin.

Determination of Nobiletin Bioaccessibility
After HPLC quantification of nobiletin in the micelle phase, nobiletin bioaccessibility was determined by the following equation [28]: bioaccessibility = nobiletin content in the micelle phase total nobiletin content in the formulations × 100% (3)

Statistical Analysis
Each experiment was conducted in triplicate. All statistical analyses were performed using OriginPro 9.0. Duncan's test of SPSS Statistics 26 was used, and significance was set at p < 0.05.

Rheology
In polymer solution, there are three regimes: dilute, semi-dilute, and concentrated. The transition from the dilute to the semi-dilute regime depends on the concentration of the polymer solution, and entanglement occurs at concentrations above the overlap threshold (C*) [29]. The C* can be determined by measuring the concentration at which the viscosity of the polymer solution increases suddenly. Nanofibrils are, to some extent, similar to polymers, which have a few nanometers in diameter but up to several micrometers in length. As shown in Figure 1A, the viscosity of WPI fibril solutions at the same concentration decreased with increasing shear rates, indicating that all samples showed shear-thinning behavior, a type of non-Newtonian behavior. At the same shear rates, the viscosity of WPI fibrils increased with increasing WPI fibril concentration, which was attributed to molecule chain entanglement [26]. There was more entanglement among protein molecules as the WPI fibril concentration increased. In addition, at low enough shear rates, shear-thinning samples showed a constant viscosity, which was defined as the zero-shear viscosity (η 0 ). Therefore, to further understand the steady flow behavior of WPI fibrils, the Cross model was applied in Figure 1B. At low concentrations (0.1-0.25 wt.%), the η 0 of the WPI fibrils did not change significantly, indicating that the WPI fibrils were flexible and moved freely in the solution. As the concentration increased to 0.5 wt.%, the η 0 increased abruptly, which indicated that it was a critical overlap concentration (C*) [26]. At this concentration, WPI fibrils were closer to each other, and entanglement occurred between the WPI fibrils. As the concentration increased further, the density of entanglement between the WPI fibrils increased, which led to an increase in the viscosity of the WPI fibrils.

SAXS
Small-angle X-ray scattering (SAXS) is a powerful method that probes the structure of biomolecules at the nanoscale [30]. Figure 2A shows the scattering intensity profiles of WPI fibrils at various concentrations. In a previous study, β-Lg fibrils exhibited a rigid rod-like structure at low concentrations (0.2 and 0.3 wt.%) since the slope of the low Q region of the scattering profiles was −1 [6]. Guinier analysis is a straightforward approach to determining the cross-sectional radius of gyration (R c ), and the shape of the primary nanostructure is calculated via Guinier fitting in the Q range of 0.03-0.05 Å [31]. Guinier plots of WPI fibrils are presented in Figure 2B. The R c of WPI fibrils from 0.1 wt.% to 1.25 wt.% was 11.45 Å, 15.55 Å, 13.75 Å, 12.57 Å, and 11.84 Å, respectively. The value of R c increased when WPI fibril concentration increased, and it decreased after the concentration exceeded 0.5 wt.%, which suggested that there were two distinct, concentration-dependent scaling regions in the R c of WPI fibril solutions. These regions are analogous to the threshold between dilute and semi-dilute regions in WPI fibrils dissolved in isolated form. Moreover, there was hardly any intermolecular interaction when the concentration of WPI fibrils was lower than 0.5 wt.%. Above 0.5 wt.%, the fibril size decreased upon the increase in fibril concentration, which may be due to intermolecular interpenetration [32]. The threshold concentration between the two distinct scaling regions was very close to the C* of the WPI fibril solution.

SAXS
Small-angle X-ray scattering (SAXS) is a powerful method that probes the structure of biomolecules at the nanoscale [30]. Figure 2A shows the scattering intensity profiles of WPI fibrils at various concentrations. In a previous study, β-Lg fibrils exhibited a rigid rod-like structure at low concentrations (0.2 and 0.3 wt.%) since the slope of the low Q region of the scattering profiles was −1 [6]. Guinier analysis is a straightforward approach to determining the cross-sectional radius of gyration (Rc), and the shape of the primary nanostructure is calculated via Guinier fitting in the Q range of 0.03-0.05 Å [31]. Guinier plots of WPI fibrils are presented in Figure 2B. The Rc of WPI fibrils from 0.1 wt.% to 1.25 wt.% was 11.45 Å, 15.55 Å, 13.75 Å, 12.57 Å, and 11.84 Å, respectively. The value of Rc increased when WPI fibril concentration increased, and it decreased after the concentration exceeded 0.5 wt.%, which suggested that there were two distinct, concentration-dependent scaling regions in the Rc of WPI fibril solutions. These regions are analogous to the threshold between dilute and semi-dilute regions in WPI fibrils dissolved in isolated form. Moreover, there was hardly any intermolecular interaction when the concentration of WPI fibrils was lower than 0.5 wt.%. Above 0.5 wt.%, the fibril size decreased upon the increase in fibril concentration, which may be due to intermolecular interpenetration [32]. The threshold concentration between the two distinct scaling regions was very close to the C* of the WPI fibril solution.

SAXS
Small-angle X-ray scattering (SAXS) is a powerful method that probes the structure of biomolecules at the nanoscale [30]. Figure 2A shows the scattering intensity profiles of WPI fibrils at various concentrations. In a previous study, β-Lg fibrils exhibited a rigid rod-like structure at low concentrations (0.2 and 0.3 wt.%) since the slope of the low Q region of the scattering profiles was −1 [6]. Guinier analysis is a straightforward approach to determining the cross-sectional radius of gyration (Rc), and the shape of the primary nanostructure is calculated via Guinier fitting in the Q range of 0.03-0.05 Å [31]. Guinier plots of WPI fibrils are presented in Figure 2B. The Rc of WPI fibrils from 0.1 wt.% to 1.25 wt.% was 11.45 Å, 15.55 Å, 13.75 Å, 12.57 Å, and 11.84 Å, respectively. The value of Rc increased when WPI fibril concentration increased, and it decreased after the concentration exceeded 0.5 wt.%, which suggested that there were two distinct, concentration-dependent scaling regions in the Rc of WPI fibril solutions. These regions are analogous to the threshold between dilute and semi-dilute regions in WPI fibrils dissolved in isolated form. Moreover, there was hardly any intermolecular interaction when the concentration of WPI fibrils was lower than 0.5 wt.%. Above 0.5 wt.%, the fibril size decreased upon the increase in fibril concentration, which may be due to intermolecular interpenetration [32]. The threshold concentration between the two distinct scaling regions was very close to the C* of the WPI fibril solution.

Interfacial Adsorption Behavior
To evaluate the role of the WPI fibril concentration in the formation dynamics of the films at the oil-water interface, the interfacial adsorption behavior of the WPI fibrils in relation to concentration was investigated. Figure 3A shows the time evolution of interfacial tension for the WPI fibrils (0.1-1.25 wt.%) at the oil-water interface. Initially, the interfacial tension values of WPI fibrils with concentrations of 0.1-1.25 wt.% were 15.96 mN/m, 15.36 mN/m, 15.08 mN/m, 13.98 mN/m, and 13.47 mN/m, respectively. This suggested that a significant reduction in interfacial tension occurred when the concentration was higher than C*. The interfacial tension values of WPI fibrils rapidly decreased with adsorption time and exhibited a notable dependence on fibril concentration, which suggests a high concentration could positively promote the adsorption process [33]. The diffusion rate was used to explain the migration of proteins from the bulk phase to the oil-water interface [34]. K diff is an estimation of the rate of initial diffusion-controlled migration, and it was dependent on the concentration in the bulk phase [35]. The diffusion of fibrils from the bulk phase to the interface occurred at relatively short adsorption times (up to about 47.8 s) [35], and gradually reached equilibrium [36]. Figure 3B shows that the K diff values increased with increasing WPI fibril concentration. Thus, the rate of WPI adsorption was faster at higher concentrations. This suggests the concentration gradient is the driving force for the diffusion of WPI fibrils, which was supported by the previous results for whey protein isolate at the oil-water interfaces [35]. Liu et al. studied the diffusion of soy glycinin when the particle concentration was low (0.01 wt.%). While diffusion-controlled adsorption occurred at a high concentration (>0.5 wt.%) during the initial periods of adsorption, it can be inferred that diffusion is related to the C* [14].
suggests a high concentration could positively promote the adsorption process [33]. The diffusion rate was used to explain the migration of proteins from the bulk phase to the oil-water interface [34]. Kdiff is an estimation of the rate of initial diffusion-controlled migration, and it was dependent on the concentration in the bulk phase [35]. The diffusion of fibrils from the bulk phase to the interface occurred at relatively short adsorption times (up to about 47.8 s) [35], and gradually reached equilibrium [36]. Figure 3B shows that the Kdiff values increased with increasing WPI fibril concentration. Thus, the rate of WPI adsorption was faster at higher concentrations. This suggests the concentration gradient is the driving force for the diffusion of WPI fibrils, which was supported by the previous results for whey protein isolate at the oil-water interfaces [35]. Liu et al. studied the diffusion of soy glycinin when the particle concentration was low (0.01 wt.%). While diffusion-controlled adsorption occurred at a high concentration (>0.5 wt.%) during the initial periods of adsorption, it can be inferred that diffusion is related to the C* [14].

Characterization of WPI Fibril-Stabilized Emulsions
The visual observation of freshly prepared Pickering emulsions at different concentrations is shown in Figure 4A. The WPI fibrils with various concentrations stabilized the Pickering emulsions and exhibited obvious creaming. The droplet sizes of the emulsions stabilized by WPI fibrils, with a fixed oil-phase fraction of 0.5, are presented in Figure 4B. The droplet size decreased with increasing WPI fibril concentration (from 0.1 wt.% to 0.5 wt.%), and the average droplet size remained relatively constant when the WPI fibril concentration increased from 0.5 wt.% to 1.25 wt.%. This suggested that there was a minimal concentration (0.5 wt.%) for the formation of homogenous WPI fibril-stabilized Pickering emulsions. CLSM was used to observe the morphology and interface properties of Pickering emulsions. The CLSM images of WPI-stabilized Pickering emulsions, with different concentrations and a fixed oil-phase fraction of 0.5, are shown in Figure 4C. WPI fibrils

Characterization of WPI Fibril-Stabilized Emulsions
The visual observation of freshly prepared Pickering emulsions at different concentrations is shown in Figure 4A. The WPI fibrils with various concentrations stabilized the Pickering emulsions and exhibited obvious creaming. The droplet sizes of the emulsions stabilized by WPI fibrils, with a fixed oil-phase fraction of 0.5, are presented in Figure 4B. The droplet size decreased with increasing WPI fibril concentration (from 0.1 wt.% to 0.5 wt.%), and the average droplet size remained relatively constant when the WPI fibril concentration increased from 0.5 wt.% to 1.25 wt.%. This suggested that there was a minimal concentration (0.5 wt.%) for the formation of homogenous WPI fibril-stabilized Pickering emulsions. CLSM was used to observe the morphology and interface properties of Pickering emulsions. The CLSM images of WPI-stabilized Pickering emulsions, with different concentrations and a fixed oil-phase fraction of 0.5, are shown in Figure 4C. WPI fibrils were labeled with ThT, which can only bind with fibrillar structures, and bright green fluorescence was observed at various concentrations, indicating all WPI fibrils were able to absorb at the oil-water interface [35]. In addition, when the concentration was low (<0.5 wt.%), the image showed relatively large emulsion droplets, which may be due to insufficient coverage of the surface of the emulsion droplets by WPI fibrils [10]. When the WPI fibril concentration increased to 0.5 wt.%, the droplet size became smaller and homogenous. This suggests the surface of droplets was sufficiently covered by fibrils at high concentrations (from 0.5 wt.% to 1.25 wt.%), and the emulsion became stable due to the electrostatic or/and steric repulsions [37]. Thus, when the WPI concentration was lower than C*, single-molecule chains existed in isolation with a low amount of WPI fibrils. When the WPI fibril concentration was greater than C*, the viscosity of the system increased due to a high degree of polymer entanglement [26], which could hinder the movement and collision of droplets.
nous. This suggests the surface of droplets was sufficiently covered by fibrils at high concentrations (from 0.5 wt.% to 1.25 wt.%), and the emulsion became stable due to the electrostatic or/and steric repulsions [37]. Thus, when the WPI concentration was lower than C*, single-molecule chains existed in isolation with a low amount of WPI fibrils. When the WPI fibril concentration was greater than C*, the viscosity of the system increased due to a high degree of polymer entanglement [26], which could hinder the movement and collision of droplets.  Figure 5 illustrates the appearance and microstructure of the Pickering emulsion stabilized by WPI fibrils (0.1 wt.%-1.25 wt.%, v/v) at pH 2 during storage for up to 28 d. The appearance of the samples did not change significantly when the emulsions were stored for 28 d. Additionally, the droplet was still stable without aggregation when the WPI fibril concentration was low (0.1 wt.%), suggesting that WPI fibrils are outstanding emulsifiers. Figure 6 shows the emulsion microstructures at various pH values and ionic strengths. No strong aggregation or droplet collapse was observed, which indicated that highly stable emulsion droplets were formed [38]. A slight increase in emulsion droplet size was observed when the pH was closer to the isoelectric point (pI), which was possibly due to a decrease in the electrostatic repulsion force at pH closer to the pI [39,40].  Figure 5 illustrates the appearance and microstructure of the Pickering emulsion stabilized by WPI fibrils (0.1-1.25 wt.%, v/v) at pH 2 during storage for up to 28 d. The appearance of the samples did not change significantly when the emulsions were stored for 28 d. Additionally, the droplet was still stable without aggregation when the WPI fibril concentration was low (0.1 wt.%), suggesting that WPI fibrils are outstanding emulsifiers. Figure 6 shows the emulsion microstructures at various pH values and ionic strengths. No strong aggregation or droplet collapse was observed, which indicated that highly stable emulsion droplets were formed [38]. A slight increase in emulsion droplet size was observed when the pH was closer to the isoelectric point (pI), which was possibly due to a decrease in the electrostatic repulsion force at pH closer to the pI [39,40].

Physicochemical Stability of WPI Fibril-Stabilized Pickering Emulsions
nous. This suggests the surface of droplets was sufficiently covered by fibrils at high concentrations (from 0.5 wt.% to 1.25 wt.%), and the emulsion became stable due to the electrostatic or/and steric repulsions [37]. Thus, when the WPI concentration was lower than C*, single-molecule chains existed in isolation with a low amount of WPI fibrils. When the WPI fibril concentration was greater than C*, the viscosity of the system increased due to a high degree of polymer entanglement [26], which could hinder the movement and collision of droplets.  Figure 5 illustrates the appearance and microstructure of the Pickering emulsion stabilized by WPI fibrils (0.1 wt.%-1.25 wt.%, v/v) at pH 2 during storage for up to 28 d. The appearance of the samples did not change significantly when the emulsions were stored for 28 d. Additionally, the droplet was still stable without aggregation when the WPI fibril concentration was low (0.1 wt.%), suggesting that WPI fibrils are outstanding emulsifiers. Figure 6 shows the emulsion microstructures at various pH values and ionic strengths. No strong aggregation or droplet collapse was observed, which indicated that highly stable emulsion droplets were formed [38]. A slight increase in emulsion droplet size was observed when the pH was closer to the isoelectric point (pI), which was possibly due to a decrease in the electrostatic repulsion force at pH closer to the pI [39,40].    Figure 7 illustrates the visual appearance, microstructure, and average droplet size of the Pickering emulsion stabilized by WPI fibrils (0.1 wt.%-1.25 wt.%) at pH 2 after load-  Figure 7 illustrates the visual appearance, microstructure, and average droplet size of the Pickering emulsion stabilized by WPI fibrils (0.1-1.25 wt.%) at pH 2 after loading 0.5 wt.% nobiletin. The emulsion droplets presented structural integrity without droplet collapse, and there was no obvious droplet coalescence. The results suggest WPI fibrils functioned well as Pickering emulsifiers to deliver nobiletin. The average droplet size of the Pickering emulsion decreased from 44.4 µm to 38.5 µm when the concentration of WPI fibril increased from 0.1 wt.% to 0.5 wt.%, and the droplet size became constant at around 39 µm, with increasing WPI fibril concentration. When the concentration of WPI fibrils was low (0.1 wt.% and 0.25 wt.%), the amount of WPI fibrils was insufficient to ensure particle surface coverage, and a smaller interfacial area was needed to prevent the coalescence of droplets. To decrease the interfacial area, the droplet size must increase [41]. When increasing the concentration of WPI fibrils to 0.5 wt.%, the sufficient coverage of WPI fibrils can facilitate the occurrence of smaller emulsion droplets with a larger surface area.  Figure 7 illustrates the visual appearance, microstructure, and average droplet size of the Pickering emulsion stabilized by WPI fibrils (0.1 wt.%-1.25 wt.%) at pH 2 after loading 0.5 wt.% nobiletin. The emulsion droplets presented structural integrity without droplet collapse, and there was no obvious droplet coalescence. The results suggest WPI fibrils functioned well as Pickering emulsifiers to deliver nobiletin. The average droplet size of the Pickering emulsion decreased from 44.4 μm to 38.5 μm when the concentration of WPI fibril increased from 0.1 wt.% to 0.5 wt.%, and the droplet size became constant at around 39 μm, with increasing WPI fibril concentration. When the concentration of WPI fibrils was low (0.1 wt.% and 0.25 wt.%), the amount of WPI fibrils was insufficient to ensure particle surface coverage, and a smaller interfacial area was needed to prevent the coalescence of droplets. To decrease the interfacial area, the droplet size must increase [41]. When increasing the concentration of WPI fibrils to 0.5 wt.%, the sufficient coverage of WPI fibrils can facilitate the occurrence of smaller emulsion droplets with a larger surface area.

Lipolysis and Bioaccessibility of Nobiletin in WPI Fibril-Stabilized Pickering Emulsions
Hydrophobic compounds have higher solubility in lipids and can be incorporated into the micelle core and then absorbed through the intestinal lining when lipids are hydrolyzed by lipase and micellized with bile salts, which can make the hydrophobic component become bioaccessible [28]. The in vitro lipolysis model is a useful tool to evaluate the lipid digestion kinetics and the bioaccessibility of target compounds in the delivery system [28]. During lipid digestion, pH will decrease due to the continuous release of fatty acids. To maintain the optimum pH for enzymatic digestion, sodium hydroxide was

Lipolysis and Bioaccessibility of Nobiletin in WPI Fibril-Stabilized Pickering Emulsions
Hydrophobic compounds have higher solubility in lipids and can be incorporated into the micelle core and then absorbed through the intestinal lining when lipids are hydrolyzed by lipase and micellized with bile salts, which can make the hydrophobic component become bioaccessible [28]. The in vitro lipolysis model is a useful tool to evaluate the lipid digestion kinetics and the bioaccessibility of target compounds in the delivery system [28]. During lipid digestion, pH will decrease due to the continuous release of fatty acids. To maintain the optimum pH for enzymatic digestion, sodium hydroxide was added to the digestion buffer [42]. In this study, the rate of lipolysis was determined by monitoring the volume of 0.1 M sodium hydroxide solution. Figure 8A shows the total amount of FFAs released from WPI fibril Pickering emulsion and soybean oil. During lipid digestion, all samples were digested rapidly in the initial stage (10-30 min), which indicated that lipase could access the emulsified lipids and catalyze the conversion of triacylglycerols into FFA and monoacylglycerols [43]. However, there were some differences between the extent of digestion and the digestion rates of Pickering emulsion and soybean oil. The FFA released from Pickering emulsions stabilized by all concentrations of WPI fibrils was significantly higher than that in soybean oil, except for a concentration of 0.1 wt.% WPI fibrils. The more FFA released, the higher the degree of lipolysis. Since lipolysis is an interfacial process and WPI fibril-stabilized Pickering emulsion droplets showed a larger interfacial area with a small droplet size, the Pickering emulsion had more opportunity to come in contact with the digestive enzymes [44]. In addition, the interface composition of the Pickering emulsion plays an important role in lipid digestion [45]. As shown in Figure 8B, the bioaccessibility of nobiletin in the Pickering emulsion with 0.5 wt.% WPI fibril was 62.53% ± 0.19%, which was much higher than that in soybean oil (21.91% ± 0.10%). It suggested that the higher the degree of lipid digestion, the more micelles are formed, and nobiletin will enter the micelles and be absorbed effectively.
wt.% WPI fibrils. The more FFA released, the higher the degree of lipolysis. Since lipolysis is an interfacial process and WPI fibril-stabilized Pickering emulsion droplets showed a larger interfacial area with a small droplet size, the Pickering emulsion had more opportunity to come in contact with the digestive enzymes [44]. In addition, the interface composition of the Pickering emulsion plays an important role in lipid digestion [45]. As shown in Figure 8B, the bioaccessibility of nobiletin in the Pickering emulsion with 0.5 wt.% WPI fibril was 62.53% ± 0.19%, which was much higher than that in soybean oil (21.91% ± 0.10%). It suggested that the higher the degree of lipid digestion, the more micelles are formed, and nobiletin will enter the micelles and be absorbed effectively. The effect of WPI fibrils concentration on the interfacial structure of emulsions and the bioaccessibility of nobiletin is depicted in a schematic diagram (Figure 9). When the concentration is less than C*, WPI fibrils move freely in the aqueous solution and have a high steric probability of adsorption to the oil-water interface. However, the amount of WPI fibrils is not enough to cover the full interface, resulting in large droplet size and uneven distribution of Pickering emulsions. When the concentration is close to C*, the oil droplets can be effectively stabilized, and the emulsion exhibited homogenous at an appropriate concentration necessary for full surface coverage. If continued to increase WPI fibrils, the fibrils penetrated each other and became aggregated and entangled on the surface of oil droplets. The flocculation or aggregates in emulsions may slow their rate of lipid digestion because the floc or aggregates would prevent lipase molecules from reaching the lipid droplets [46]. The in vitro bioaccessibility of nobiletin was mainly affected by the degree of lipolysis. After the lipid is digested, nobiletin enters the micelle's structure along with it and is further absorbed by the small intestinal epithelial cells. When the interfacial structure is dense (c~C*) or there are still numerous fibrils existing in the aqueous phase (c > C*), un-adsorbed fibrils might exert a depletion effect and make Pickering emulsions have less opportunity to come in contact with the digestive enzymes. The effect of WPI fibrils concentration on the interfacial structure of emulsions and the bioaccessibility of nobiletin is depicted in a schematic diagram (Figure 9). When the concentration is less than C*, WPI fibrils move freely in the aqueous solution and have a high steric probability of adsorption to the oil-water interface. However, the amount of WPI fibrils is not enough to cover the full interface, resulting in large droplet size and uneven distribution of Pickering emulsions. When the concentration is close to C*, the oil droplets can be effectively stabilized, and the emulsion exhibited homogenous at an appropriate concentration necessary for full surface coverage. If continued to increase WPI fibrils, the fibrils penetrated each other and became aggregated and entangled on the surface of oil droplets. The flocculation or aggregates in emulsions may slow their rate of lipid digestion because the floc or aggregates would prevent lipase molecules from reaching the lipid droplets [46]. The in vitro bioaccessibility of nobiletin was mainly affected by the degree of lipolysis. After the lipid is digested, nobiletin enters the micelle's structure along with it and is further absorbed by the small intestinal epithelial cells. When the interfacial structure is dense (c~C*) or there are still numerous fibrils existing in the aqueous phase (c > C*), un-adsorbed fibrils might exert a depletion effect and make Pickering emulsions have less opportunity to come in contact with the digestive enzymes.

Conclusions
In conclusion, the overlap concentration (C*) of WPI fibrils was around 0.5 wt.%. The interfacial absorption ability was improved with the concentration increasing. WPI fibrils at various concentrations can effectively stabilize Pickering emulsion and showed a longterm stability at room temperature. At C* of WPI fibrils, the droplet size of the Pickering Figure 9. Schematic diagram of influence of WPI fibrils concentration on the emulsion structure and the bioaccessibility of nobiletin.

Conclusions
In conclusion, the overlap concentration (C*) of WPI fibrils was around 0.5 wt.%. The interfacial absorption ability was improved with the concentration increasing. WPI fibrils at various concentrations can effectively stabilize Pickering emulsion and showed a long-term stability at room temperature. At C* of WPI fibrils, the droplet size of the Pickering emulsion became homogenous, and the emulsion was stable at various pHs and ionic strengths. WPI fibrils stabilized Pickering emulsion could significantly improve the degree of lipolysis and the bioaccessibility of nobiletin, especially at C* of WPI fibrils, indicating that the entanglement of WPI fibrils was an important parameter for designing food-grade Pickering emulsions. It also provides an important insight for broadening the applications of hydrophobic nutraceuticals.