Development, Characterization, Stability and Bioaccessibility Improvement of 7,8-Dihydroxyflavone Loaded Zein/Sophorolipid/Polysaccharide Ternary Nanoparticles: Comparison of Sodium Alginate and Sodium Carboxymethyl Cellulose

In this study, two polysaccharides [sodium alginate (ALG) and sodium carboxymethyl cellulose (CMC)] were selected to establish zein/sophorolipid/ALG (ALG/S/Z) and zein/sophorolipid/ALG (CMC/S/Z) nanoparticles to encapsulate 7,8-dihydroxyflavone (7,8-DHF), respectively. The results showed that polysaccharide types significantly affected performance of ternary nanoparticles, including CMC/S/Z possessed lower polydispersity index, particle size and turbidity, but higher zeta potential, encapsulation efficiency and loading capacity compared to ALG/S/Z. Compared to zein/sophorolipid nanoparticles (S/Z), both ALG/S/Z and CMC/S/Z had better stability against low pH (pH 3~4) and high ionic strengths (150~200 mM NaCl). Hydrophobic effects, electrostatic interactions and hydrogen bonding were confirmed in ternary nanoparticles fabrication via Fourier-transform infrared spectroscopy. Circular dichroism revealed that CMC and ALG had no evident impact on secondary structure of zein in S/Z, but changed surface morphology of S/Z as observed by scanning electron microscope. Encapsulated 7,8-DHF exhibited an amorphous state in ternary nanoparticles as detected by X-ray diffraction and differential scanning calorimetry. Furthermore, compared to S/Z, ALG/S/Z, and CMC/S/Z remarkably improved the storage stability and bioaccessibility of 7,8-DHF. CMC/S/Z possessed a greater storage stability for 7,8-DHF, however, ALG/S/Z exhibited a better in vitro bioaccessibility of 7,8-DHF. This research provides a theoretical reference for zein-based delivery system application.


Encapsulation Efficiency (EE) and Loading Capacity (LC)
The EE and LC of encapsulated 7,8-DHF were assessed by UPLC based on our previously described method [25]. Specifically, the nanoparticles were centrifuged at 10,000× g for 10 min at 4 • C using a centrifugal machine (3K15, Sigma, Osterode, Germany). The supernatant (containing loaded 7,8-DHF) was removed and diluted 5-fold with methanol. And an equal volume of the initial suspension was diluted in 5-fold methanol to obtain initial 7,8-DHF. Then, EE and LC were calculated according to the following equation: The spectral scanning range was 500~4000 cm −1 at a resolution of 4 cm −1 . The analytical results were carried out by OMNIC software version 8.0.

Circular Dichroism (CD) Spectrum
Secondary structural characteristics of nanoparticle dispersions under analysis were measured using a CD spectrometer (J-1500, JASCO, Tokyo, Japan). The concentration of complex dispersions was 0.2 mg/mL corresponding to zein. The buffer in CD experiments was deionized water (pH = 7). The secondary structure scanning region was 190~260 nm with a 0.1 cm path length. The bandwidth was 1.0 nm and scanning speed was 50 nm/min. The data were evaluated by Spectra Manager™ II Software equipped with CD spectrometer.

Differential Scanning Calorimetry (DSC)
Thermal behavior of 7,8-DHF and freeze-dried samples were studied via DSC (Mettler Toledo, Zurich, Switzerland). 2~10 mg of samples were accurately weighed and hermetically sealed in aluminum pans. An empty crucible under the same condition was used as a reference. Scanning calorimetry was performed from a range of 25 to 200 • C in N2 atmosphere at a heating rate of 10 • C /min under 30 mL/min flow.

X-ray Diffraction (XRD)
The crystalline characteristic of lyophilized nanoparticles and selected 7,8-DHF were recorded on an X-ray diffractometer (Bruker D8, Karlsruhe, Germany). This diffractometer was carried at 40 kV accelerating voltage and 40 mA tube current to produce copper Kα radiation. Soller slit and divergence slit were set at 2.5 • and 0.5 • , respectively, the 2θ angle was ranged from 5 • to 90 • .

Transmission Electron Microscopy (TEM)
10-fold diluted fresh nanoparticle dispersions were deposited on a copper grid with formvar-carbon coating. Then, the samples were air-dried for 5 min and stained with 2% uranyl acetate. TEM (JEM-1200 EX, Tokyo, Japan) was performed for microscopic observation at 120 kV accelerating voltage.

Field Emission Scanning Electron Microscope (FE-SEM)
The surface morphology of polysaccharide samples and freeze-dried nanoparticles was captured by FE-SEM (GeminiSEM 300, ZEISS, Germany). Before analysis, 3-6 nm of a thick gold layer was covered on the sample surfaces. The electron microscope acceleration voltage was 15.0 kV.

In Vitro Simulated Gastrointestinal Digestion
Based on Yuan et al.'s study via some amendments [28], briefly, 10 mL of simulated gastric fluid (SGF, 3.2 mg/mL pepsin and 2 mg/mL NaCl, pH = 2.5) and 10 mL of 7,8-DHF, DHF-S/Z, DHF-ALG/S/Z, and DHF-CMC/S/Z were mixed for incubating in a 37 • C water bath shaker for 60 min at 100 rpm. After SGF digestion, 10 mL of abovementioned simulated gastric digestive fluids were rapidly adjusted to pH 7.4 using 2 M NaOH. Whereafter, 10 mL of simulated intestinal fluid (SIF, 4 mg/mL pancreatin, 5 mg/mL bile salts, 6.8 mg/mL K2HPO4 and 8.8 mg/mL NaCl, pH = 7.4) was added into above-mentioned simulated gastric digestive fluids and incubated for 120 min at same temperature and revolving speed. Finally, the final digestive solution was centrifuged by 20,000× g centrifugal force for 1 h, and the supernatant (the mixed micelle phase containing 7,8-DHF) was collected. The bioaccessibility (%) was calculated based on the following equation: Bioaccessibility (%) = 7,8-DHF concentration in the micelles phases/7,8-DHF concentration in the formulation × 100 (4)

Statistical Analysis
Mean ± SD was presented via at least three times for all data. One-way ANOVA followed by Tukey's honestly significant difference post hoc tests were utilized to assess potential significant differences among groups. A p < 0.05 indicated significant differences between groups. Data analysis was carried out by Origin 2021 (Origin Lab Co., Northampton, MA, USA) and GraphPad Prism version 8.0 (GraphPad Sofware, San Diego, CA, USA).

Effect of Polysaccharides on Ternary Nanoparticles Fabrication
The effects of varying concentrations and types of polysaccharide on particle size of S/Z were illustrated in Figure 1. S/Z was approximate 82.81 nm particle size. The particle size of the ternary nanoparticles was differentially altered from that of S/Z after introduction with CMC and ALG into the binary system. After supplementation with CMC, the particle size of CMC/S/Z stepwise grew as the concentration of CMC increased ( Figure 1A). Moreover, the incorporation of ALG resulted in a significant increase in particle size (p < 0.05). And in comparison to CMC/S/Z, ALG/S/Z showed a greater particle size at the identical polysaccharide level ( Figure 1B). The above phenomenon might be attributed to the distinctive molecular structure of two studied polysaccharides and interaction force among molecules. The changes in particle size of zein/sophorolipid/polysaccharide ternary nanoparticles are possibly a result of sufficient anionic CMC and ALG being ab-sorbed on the surface of S/Z through the electrostatic attractions. The changes in PDI value of CMC/S/Z and ALG/S/Z was analogous to that of particle size. As the concentration of CMC and ALG increased, the PDI value of CMC/S/Z and ALG/S/Z gradually increased as well. However, ALG/S/Z showed a higher PDI value at the same polysaccharide level.
Foods 2021, 10, x. https://doi.org/10.3390/xxxxx www.mdpi.com/journal/foods tion with CMC and ALG into the binary system. After supplementation with CMC, the particle size of CMC/S/Z stepwise grew as the concentration of CMC increased ( Figure  1A). Moreover, the incorporation of ALG resulted in a significant increase in particle size (p < 0.05). And in comparison to CMC/S/Z, ALG/S/Z showed a greater particle size at the identical polysaccharide level ( Figure 1B). The above phenomenon might be attributed to the distinctive molecular structure of two studied polysaccharides and interaction force among molecules. The changes in particle size of zein/sophorolipid/polysaccharide ternary nanoparticles are possibly a result of sufficient anionic CMC and ALG being absorbed on the surface of S/Z through the electrostatic attractions. The changes in PDI value of CMC/S/Z and ALG/S/Z was analogous to that of particle size. As the concentration of CMC and ALG increased, the PDI value of CMC/S/Z and ALG/S/Z gradually increased as well. However, ALG/S/Z showed a higher PDI value at the same polysaccharide level. The zeta-potential of S/Z was −40.63 mV. When CMC and S/Z were combined, their charge values became more negative as the fraction of present CMC increased ( Figure 1C). On the other hand, the negative charges of ALG/S/Z initially declined and grew whereafter when the level of ALG was increased ( Figure 1D). When the mass ratio of The zeta-potential of S/Z was −40.63 mV. When CMC and S/Z were combined, their charge values became more negative as the fraction of present CMC increased ( Figure 1C). On the other hand, the negative charges of ALG/S/Z initially declined and grew whereafter when the level of ALG was increased ( Figure 1D). When the mass ratio of polysaccharide to zein was above 1:10, the negative charges of ALG/S/Z were higher than that of CMC/S/Z. This difference is potentially caused by the different molecular structures of the two polysaccharides. The zeta-potential of zein/sophorolipid/polysaccharide ternary nanoparticles had been dominated by negatively charged CMC and ALG. It is reported that electrostatic attraction played a critical role in the formation of protein-surfactantpolysaccharide complexes, for instance, zein-rhamnolipid-propylene glycol alginate [29], and zein-lecithin-propylene glycol alginate [30]. Meanwhile, the uniform trend of turbidity in ternary nanoparticles was consistent with that of PDI value and particle size, as the increasing trend of turbidity was primarily caused by micro-aggregation of complexes [31].

CMC and ALG Prevents S/Z Ternary Nanoparticles Precipitation at pH = 4
Our previous work confirms that S/Z exhibits weak stability in low pH conditions. To overcome this deficiency, we explore the influence of polysaccharide type (CMC and ALG) and concentration on S/Z stability at pH = 4. As shown in (Figure 2A), S/Z ternary nanoparticles were extremely unstable and aggregated together as expected. This phenomenon could be elucidated by the lack of appropriate electrostatic repulsion, which is required to conquer attractive interactions (e.g., van der Waals) among nanoparticles. Furthermore, at low pH values, the pKa of hydrophilic sugar residues present on sophorolipid molecules were reported to promote this instability [32]. After CMC was added, the PDI value, particle size, and turbidity of CMC/S/Z initially declined and subsequently jumped until a mass of zein to CMC was 5:1, the CMC/S/Z were possessed the lowest values for particle size, PDF, and turbidity (particle size ≈ 341.50 nm, PDI ≈ 0.335 and turbidity ≈ 2.257) (Figure 2A,C). Moreover, the incorporation of ALG similarly decreased and then increased PDI value, particle size and turbidity of ALG/S/Z. The particle size for ALG/S/Z was smallest at the 10:1 (353.62 nm), while the PDI value (0.335) and turbidity (2.509) were lowest at 5:1 mass ratio of zein to ALG ( Figure 2B,D). As seen from photograph ( Figure 2E,F), no floccules were observed at the bottom of the container at varying zein to polysaccharide mass ratios (10:1, 5:1, 3:1, and 2:1). A reason for lacking floccules at these mass ratios might be a result of appropriate CMC and ALG participation, which potentially increased the steric or electrostatic repulsions for S/Z against aggregation. Nevertheless, at a zein to polysaccharide mass of 1:1, there were sediments in both CMC/S/Z and ALG/S/Z. The underlying mechanism for these observations is bridging flocculation, i.e., the ability of a single CMC or ALG molecule to adsorb to the surfaces of two or more S/Z. Therefore, CMC or ALG adsorption results in the formation of clusters [33]. Furthermore, with the CMC and ALG concentrations raised, the charge values of CMC/S/Z and ALG/S/Z became increasingly negative. The intensification of these charges revalidates the idea that CMC/S/Z and ALG/S/Z experience sufficient electrostatic repulsion to avoid precipitation at varying mass ratios of 10:1~2:1. Based on these results, we subsequently selected CMC/S/Z and ALG/S/Z at a mass ratio of zein to polysaccharide at 5:1 as the optimum ratio to study in the following research.

CMC and ALG Prevents S/Z Ternary Nanoparticles Precipitation at pH = 4
Our previous work confirms that S/Z exhibits weak stability in low pH conditions. To overcome this deficiency, we explore the influence of polysaccharide type (CMC and ALG) and concentration on S/Z stability at pH = 4. As shown in (Figure 2A), S/Z ternary nanoparticles were extremely unstable and aggregated together as expected. This phenomenon could be elucidated by the lack of appropriate electrostatic repulsion, which is required to conquer attractive interactions (e.g., van der Waals) among nanoparticles. Furthermore, at low pH values, the pKa of hydrophilic sugar residues present on sophorolipid molecules were reported to promote this instability [32]. After CMC was added, the PDI value, particle size, and turbidity of CMC/S/Z initially declined and subsequently jumped until a mass of zein to CMC was 5:1, the CMC/S/Z were possessed the lowest values for particle size, PDF, and turbidity (particle size ≈ 341.50 nm, PDI ≈ 0.335 and turbidity ≈ 2.257) (Figure 2A,C). Moreover, the incorporation of ALG similarly decreased and then increased PDI value, particle size and turbidity of ALG/S/Z. The particle size for ALG/S/Z was smallest at the 10:1 (353.62 nm), while the PDI value (0.335) and turbidity (2.509) were lowest at 5:1 mass ratio of zein to ALG ( Figure 2B,D). As seen from photograph ( Figure 2E,F), no floccules were observed at the bottom of the container at varying zein to polysaccharide mass ratios (10:1, 5:1, 3:1, and 2:1). A reason for lacking floccules at these mass ratios might be a result of appropriate CMC and ALG participation, which potentially increased the steric or electrostatic repulsions for S/Z against aggregation. Nevertheless, at a zein to polysaccharide mass of 1:1, there were sediments in both CMC/S/Z and ALG/S/Z. The underlying mechanism for these observations is bridging flocculation, i.e., the ability of a single CMC or ALG molecule to adsorb to the surfaces of two or more S/Z. Therefore, CMC or ALG adsorption results in the formation of clusters [33]. Furthermore, with the CMC and ALG concentrations raised, the charge values of CMC/S/Z and ALG/S/Z became increasingly negative. The intensification of these charges revalidates the idea that CMC/S/Z and ALG/S/Z experience sufficient electrostatic repulsion to avoid precipitation at varying mass ratios of 10:1~2:1. Based on these results, we subsequently selected CMC/S/Z and ALG/S/Z at a mass ratio of zein to polysaccharide at 5:1 as the optimum ratio to study in the following research.

Physical-Chemical Stability Study on Ternary Nanoparticles
The optimum delivery system for food nutraceuticals must accommodate flexible pH and ionic environments during beverages processing, storage, and gut passage. Therefore, pH and salt stability testing of delivery systems is necessary to evaluate colloidal particles' functionality.

Effect of pH
As seen from Figure 3A, S/Z exhibited extensive aggregation at pH 3.0 and 4.0, exhibiting an increased particle size. This increase in size could be due to the fact that weakened electrostatic repulsion was weakened among binary nanoparticles. However, after adding CMC and ALG, compared to S/Z, the stability of both nanoparticles was enhanced noteworthily within pH 3.0 to 4.0. Most notably, from pH 3 to 9, both CMC/S/Z and ALG/S/Z had excellent stability. It is reported that certain zein-surfactant or zein-polysaccharide binary nanoparticles, such as zein-rhamnolipid, zein-alginate, and zein-chondroitin sulfate, experienced a significant size increase at low pH levels [34][35][36]. Unexpectedly, zein-surfactant-polysaccharide ternary nanoparticles exhibited superior pH stability. There are some yellow floccules in the bottom of the container of S/Z at pH 3.0 and 4.0 but no floccules in CMC/S/Z and ALG/S/Z ( Figure S4).

Physical-Chemical Stability Study on Ternary Nanoparticles
The optimum delivery system for food nutraceuticals must accommodate flexible pH and ionic environments during beverages processing, storage, and gut passage. Therefore, pH and salt stability testing of delivery systems is necessary to evaluate colloidal particles' functionality.

Effect of pH
As seen from Figure 3A, S/Z exhibited extensive aggregation at pH 3.0 and 4.0, exhibiting an increased particle size. This increase in size could be due to the fact that weakened electrostatic repulsion was weakened among binary nanoparticles. However, after adding CMC and ALG, compared to S/Z, the stability of both nanoparticles was enhanced noteworthily within pH 3.0 to 4.0. Most notably, from pH 3 to 9, both CMC/S/Z and ALG/S/Z had excellent stability. It is reported that certain zein-surfactant or zein-polysaccharide binary nanoparticles, such as zein-rhamnolipid, zein-alginate, and zein-chondroitin sulfate, experienced a significant size increase at low pH levels [34][35][36]. Unexpectedly, zein-surfactant-polysaccharide ternary nanoparticles exhibited superior pH stability. There are some yellow floccules in the bottom of the container of S/Z at pH 3.0 and 4.0 but no floccules in CMC/S/Z and ALG/S/Z ( Figure S4).

Effect of Ionic Strengths
Particle size of S/Z was relatively low by 0 to 100 mM NaCl concentrations (F 3B). However, when increased to higher ionic strengths at 150~200 mM NaCl conce tions, precipitates were produced along with increasing particle size. This forming a gation patterns for nanoparticles at high ionic strengths could be explained by incre counterions (such as Cl − and Na + ), depressed electrostatic repulsion via electro screening and nanoparticle charge neutralization [21]. However, in the existence of and ALG, the particle size of CMC/S/Z and ALG/S/Z was increased with the NaC centration increased. And adding CMC and ALG reduced the sensitivity of S/Z to a gation under high ionic strength condition. There are two possible reasons for th proved physical stability of CMC/S/Z and ALG/S/Z. Firstly, CMC and ALG prev counterion neutralization and strong electrostatic repulsions, which could potential strain complex particles from coalescence via enhancing repulsive forces among par Secondly, an increase of the steric repulsion among particles could have occurred a ysaccharide molecules were attached to S/Z. Some floccules were also observed at th tom of the tube for S/Z at 150 and 200 mM NaCl ( Figure S5). However, there we floccules at the bottom of the tube for CMC/S/Z and ALG/S/Z, in which stable coll systems were both developed.

Encapsulation of 7,8-DHF
The influence of polysaccharide types (CMC and ALG) and 7,8-DHF concentr on PDI value, particle size, EE and LC value in different delivery systems were sum rized in Table 1. When the mass ratio of 7,8-DHF to zein was 1:5, the EE and LC of S/Z were 82.42 and 7.49%, respectively. However, compared to DHF-S/Z, DHF-CM and DHF-ALG/S/Z exhibited better EE value, especially DHF-CMC/S/Z nanopar (88.63%) (p < 0.05). These findings suggested that polysaccharide addition improve embedding ability of S/Z via non-covalent interactions. There are two hydroxyl grou the 7,8-DHF molecule that can interact with the hydroxyl groups of polysaccharide the tyrosyl of zein. These interactions can form hydrogen bonds, possibly resulting increase of EE. These phenomena indicated that polysaccharides and S/Z showed a ergistic effect on the EE of 7,8-DHF. When mass ratio of 7,8-DHF to zein was 1:10, colloidal delivery systems possessed high EE for 7,8-DHF (above 98.21%). In contr unloaded colloidal delivery systems, the particle size of loaded colloidal delivery sys

Effect of Ionic Strengths
Particle size of S/Z was relatively low by 0 to 100 mM NaCl concentrations ( Figure 3B). However, when increased to higher ionic strengths at 150~200 mM NaCl concentrations, precipitates were produced along with increasing particle size. This forming aggregation patterns for nanoparticles at high ionic strengths could be explained by increasing counterions (such as Cl − and Na + ), depressed electrostatic repulsion via electrostatic screening and nanoparticle charge neutralization [21]. However, in the existence of CMC and ALG, the particle size of CMC/S/Z and ALG/S/Z was increased with the NaCl concentration increased. And adding CMC and ALG reduced the sensitivity of S/Z to aggregation under high ionic strength condition. There are two possible reasons for the improved physical stability of CMC/S/Z and ALG/S/Z. Firstly, CMC and ALG prevented counterion neutralization and strong electrostatic repulsions, which could potentially restrain complex particles from coalescence via enhancing repulsive forces among particles. Secondly, an increase of the steric repulsion among particles could have occurred as polysaccharide molecules were attached to S/Z. Some floccules were also observed at the bottom of the tube for S/Z at 150 and 200 mM NaCl ( Figure S5). However, there were no floccules at the bottom of the tube for CMC/S/Z and ALG/S/Z, in which stable colloidal systems were both developed.

Encapsulation of 7,8-DHF
The influence of polysaccharide types (CMC and ALG) and 7,8-DHF concentration on PDI value, particle size, EE and LC value in different delivery systems were summarized in Table 1. When the mass ratio of 7,8-DHF to zein was 1:5, the EE and LC of DHF-S/Z were 82.42 and 7.49%, respectively. However, compared to DHF-S/Z, DHF-CMC/S/Z and DHF-ALG/S/Z exhibited better EE value, especially DHF-CMC/S/Z nanoparticles (88.63%) (p < 0.05). These findings suggested that polysaccharide addition improved the embedding ability of S/Z via non-covalent interactions. There are two hydroxyl groups in the 7,8-DHF molecule that can interact with the hydroxyl groups of polysaccharides and the tyrosyl of zein. These interactions can form hydrogen bonds, possibly resulting in an increase of EE. These phenomena indicated that polysaccharides and S/Z showed a synergistic effect on the EE of 7,8-DHF. When mass ratio of 7,8-DHF to zein was 1:10, three colloidal delivery systems possessed high EE for 7,8-DHF (above 98.21%). In contrast to unloaded colloidal delivery systems, the particle size of loaded colloidal delivery systems was growing when increasing the mass ratio of 7,8-DHF to zein. In general, 7,8-DHF, as a hydrophobic compound, is expected to be located within hydrophobic region of zein, ultimately possessing an impact on the size and stability of particles. 7,8-DHF encapsulation might impair interactions among hydrophobic groups of zein and increase particle size.
Additionally, the encapsulation of 7,8-DHF caused an increase of PDI value in contrast to unloaded complexes. A uniform result was reported by Sun et al. [37]. The formation of free 7,8-DHF aggregates might help to explain the above observed phenomenon. Values are the means ± SD (n = 3). a-e : Different letters in the same column indicate significant differences (p < 0.05) based on one-way ANOVA analysis followed by Tukey's honest significant difference post hoc tests.

FTIR
FTIR is a versatile tool for monitoring changes within the functional groups of biopolymers and evaluating intermolecular interactions between components and particles. As shown in Figure 4A, the spectrum of CMC and ALG included diversiform representative carbohydrate peaks. The broad 3446 and 3441 cm −1 peaks represented hydroxyl groups (O-H) stretching, and the sharp 1604 and 1615 cm −1 peaks are associated with carboxyl (-COO-) symmetrical stretching vibration [38]. The featured peaks of CMC and ALG spectra appear at 1418 and 1417 cm −1 , corresponding to rhamnogalacturonan moiety [39]. Multiple simultaneous vibration peaks in CMC and ALG at 900~1350 cm −1 were accredited to characteristic peaks of polysaccharides [24]. In our previous study, we confirmed that there are electrostatic attractions, hydrophobic interactions, and hydrogen bonding in S/Z. The FTIR spectrum of S/Z nanoparticles showed typical characteristic peaks at 3368, 1657, and 1545 cm −1 , respectively. When CMC and ALG were added, the peaks of O-H stretching vibration (3100~3500 cm −1 peaks) shifted from 3368 (S/Z) to 3382 (CMC/S/Z) and 3402 cm −1 (ALG/S/Z) [40]. This characteristic peak migration implied that there was a strong hydrogen bond between -OH groups in polysaccharides and amide group of glutamine in zein [41]. According to Liu et al.'s report [42], the 1657 cm −1 peak of zein at was the C=O stretching (amide I). 1545 cm −1 peak was primarily associated with bending of N-H coupled with the stretching of C-N (amide II). With CMC and ALG incorporation, the amide I and amide II characteristic peaks of ternary nanoparticles were switched to (1650 and 1651 cm −1 ) and (1544 and 1544 cm −1 ), respectively. These results revealed electrostatic attractions were related to the establishment process of CMC/S/Z and ALG/S/Z. Based on these results, we confirmed that ALG/S/Z possessed stronger hydrogen bonding and electrostatic attraction than CMC/S/Z. As seen from Figure 4B, the peaks at 3114, 1626, 1575, 1405, 1195, and 1071 cm −1 were the typical peaks of 7,8-DHF, which have been confirmed in our previous study [25,43]. Expectedly, these characteristic peaks of 7,8-DHF were vanished in both binary and ternary nanoparticles, indicating that DHF-S/Z, DHF-CMC/S/Z, and DHF-ALG/S/Z samples successfully encapsulated for 7,8-DHF.

CD Spectrum
In this study, CD spectrum analysis was applied to measure conformational changes (secondary structure) of zein in complexation (260 nm~190 nm). As seen from ( Figure 4C), two peaks at 209 and 223 nm with a zero-crossing were around 203 nm (typical secondary structure) in the zein spectrum [44]. As shown in (Table 2), the α-helix, β-sheet and β-turn

CD Spectrum
In this study, CD spectrum analysis was applied to measure conformational changes (secondary structure) of zein in complexation (260 nm~190 nm). As seen from ( Figure 4C), two peaks at 209 and 223 nm with a zero-crossing were around 203 nm (typical secondary structure) in the zein spectrum [44]. As shown in (Table 2), the α-helix, β-sheet and β-turn content of DHF-S/Z changed from 25.3 to 24.7 and 24.5%, 25.8 to 26.1 and 26.6% and 28.3 to 29.1 and 27.9%, respectively after adding CMC and ALG. These results suggested that the addition of CMC and ALG had no apparent impact on zein's secondary structure in DHF-S/Z. The behavior of CMC and ALG could possibly be the result of sophorolipid adsorption on the zein surface. This adsorption could have possibly resulted in insufficient contact between the polysaccharides and zein, which was difficult to further change the secondary structure of zein. In general, the amide I band shifts in FTIR spectra could also reflect secondary structure changes of the protein, including the band of α-helix, β-sheet, and β-turn [45]. In the FTIR spectra of our study ( Figure 4B), the band of amide I did not show significant change among DHF-S/Z, DHF-CMC/S/Z and DHF-ALG/S/Z. This lack of significant change further confirmed that CMC and ALG did not impact the secondary structure of zein.

DSC
Thermal properties of individual components and composite nanoparticles were studied via DSC. As shown in (Figure 4D), the embedded thermograms of 7,8-DHF displayed a narrow and sharp peak at 246.24 • C. This temperature peak was probably caused by the melting of 7,8-DHF crystals [46]. Furthermore, the representative endothermal peak of CMC and ALG was at approximately 96.33 and 113.33 • C, respectively. These findings confirmed that ALG had a higher thermostability than CMC because of its specific carbohydrate structure. However, the endothermic peak of S/Z was at approximately 60.33 • C, showing a low thermostability. After 7,8-DHF was encapsulated, no endothermic peaks of 7,8-DHF were found in DHF-S/Z, DHF-CMC/S/Z, and DHF-ALG/S/Z. The lack of endothermic peaks verified that 7,8-DHF was defined as an amorphous form rather than a crystalline form. Similar literature has reported in recent study on curcumin [47] and hyperoside [48]. In addition, the endothermic peak of DHF-S/Z nanoparticles was increased from 60.33 • C to 64.66 • C compared to S/Z. This increase in melting temperatures could be attributed to intermolecular interactions among 7,8-DHF, zein and sophorolipids [29]. Most importantly, after adding CMC and ALG, the endothermic peak of DHF-CMC/S/Z and DHF-ALG/S/Z rose to 70.01 and 76.33 • C in comparison to DHF-S/Z, respectively. The higher endothermic peak of DHF-CMC/S/Z and DHF-ALG/S/Z manifested that they possessed better thermal stability than DHF-S/Z. This new peak might be the result of CMC and ALG interaction enhancements for hydrophobic, electrostatic, or hydrogen bond interactions among different components in nanoparticles, which ultimately leads to a higher endothermic peak temperature.

XRD
X-ray diffraction ranging from 5 • to 90 • at 2θ values was applied to examine 7,8-DHF's physical state in different nanoparticles. As presented in (Figure 4E), 7,8-DHF was highly crystalline with multiple sharp diffraction peaks at the range of 5~40 • . On the one hand, when the diffraction angles of ALG in the 2θ range were 13.2 • and 21.8 • , two flat peaks without sharp diffraction maximum appeared in the XRD spectrum. On the other hand, CMC only had a flat peak at a diffraction angle of 20.2 • , while no sharp diffraction peaks were emerged. This XRD hump indicated that both pure polysaccharides were totally in an amorphous form [21,32]. However, no obvious sharp diffraction peaks about crystalline form for 7,8-DHF were found in both binary and ternary nanoparticles. Such behavior indicated 7,8-DHF was completely loaded into nanoparticles with an amorphous state. Interestingly, in comparison to DHF-S/Z, the hump for DHF-ALG/S/Z was distinctly increased, while the hump of DHF-CMC/S/Z was only slightly increased as the diffraction angle was at 19.6 • . Thus, the polysaccharide types applied to construct nanocomposites can significantly affect XRD patterns. Moreover, no characteristic peaks for these two polysaccharides were observed. Collectively, due to changes in the interactions (hydrophobic, hydrogen bonding, and electrostatic) among zein, polysaccharide, and surfactant molecule, distinguishable behaviors of 7,8-DHF in DHF-S/Z, DHF-CMC/S/Z, and DHF-ALG/S/Z were observed.

Micromorphology
The microstructural features of loaded composite nanoparticles were analyzed via TEM. As observed in (Figure 5A), the diameter of DHF-S/Z was roughly 100 nm, this finding was in agreement with the results of dynamic light scattering. After adding CMC and ALG, DHF-CMC/S/Z and DHF-ALG/S/Z showed a similar spherical structure that was comparable to the DHF-S/Z. However, the diameter of DHF-CMC/S/Z and DHF-ALG/S/Z was larger than that of DHF-S/Z, exhibiting a particle size of 100~200 nm ( Figure 5B,C). These findings demonstrate that CMC and ALG were absorbed on the surface of DHF-S/Z. Additionally, the diameter of DHF-CMC/S/Z was larger than that of DHF-ALG/S/Z, which was consistent with the results of DLS. crystalline form for 7,8-DHF were found in both binary and ternary nanoparticles. Such behavior indicated 7,8-DHF was completely loaded into nanoparticles with an amorphous state. Interestingly, in comparison to DHF-S/Z, the hump for DHF-ALG/S/Z was distinctly increased, while the hump of DHF-CMC/S/Z was only slightly increased as the diffraction angle was at 19.6°. Thus, the polysaccharide types applied to construct nanocomposites can significantly affect XRD patterns. Moreover, no characteristic peaks for these two polysaccharides were observed. Collectively, due to changes in the interactions (hydrophobic, hydrogen bonding, and electrostatic) among zein, polysaccharide, and surfactant molecule, distinguishable behaviors of 7,8-DHF in DHF-S/Z, DHF-CMC/S/Z, and DHF-ALG/S/Z were observed.

Micromorphology
The microstructural features of loaded composite nanoparticles were analyzed via TEM. As observed in (Figure 5A), the diameter of DHF-S/Z was roughly 100 nm, this finding was in agreement with the results of dynamic light scattering. After adding CMC and ALG, DHF-CMC/S/Z and DHF-ALG/S/Z showed a similar spherical structure that was comparable to the DHF-S/Z. However, the diameter of DHF-CMC/S/Z and DHF-ALG/S/Z was larger than that of DHF-S/Z, exhibiting a particle size of 100~200 nm ( Figure 5B,C). These findings demonstrate that CMC and ALG were absorbed on the surface of DHF-S/Z. Additionally, the diameter of DHF-CMC/S/Z was larger than that of DHF-ALG/S/Z, which was consistent with the results of DLS. Furthermore, FE-SEM was applied to further observe the differences in surface microscopic morphology among individual composite nanoparticles ( Figure 6). As seen from ( Figure 6A), the surface morphology of DHF-S/Z expressed an irregular and rough shape consisting of many interlinked independent complexes. This morphological change was possibly attributed to sophorolipid adsorption on the surface of zein particles. The microphotograph of CMC and ALG depicted a shape similar to that of a silk ribbon (Figure 6D,E). When CMC and ALG were incorporated into DHF-S/Z, the micromorphology of DHF-CMC/S/Z and DHF-ALG/S/Z were spherical with uniform size and smooth surface ( Figure 6B,C). The particle sizes of DHF-CMC/S/Z and DHF-ALG/S/Z were larger than that of DHF-S/Z. The polysaccharide coating on the outer layer of DHF-S/Z nanoparticles via electrostatic attraction possibly caused this size difference [49]. The above results were consistent with DLS measurements. Furthermore, FE-SEM was applied to further observe the differences in surface microscopic morphology among individual composite nanoparticles ( Figure 6). As seen from ( Figure 6A), the surface morphology of DHF-S/Z expressed an irregular and rough shape consisting of many interlinked independent complexes. This morphological change was possibly attributed to sophorolipid adsorption on the surface of zein particles. The microphotograph of CMC and ALG depicted a shape similar to that of a silk ribbon ( Figure 6D,E). When CMC and ALG were incorporated into DHF-S/Z, the micromorphology of DHF-CMC/S/Z and DHF-ALG/S/Z were spherical with uniform size and smooth surface ( Figure 6B,C). The particle sizes of DHF-CMC/S/Z and DHF-ALG/S/Z were larger than that of DHF-S/Z. The polysaccharide coating on the outer layer of DHF-S/Z nanoparticles via electrostatic attraction possibly caused this size difference [49]. The above results were consistent with DLS measurements.

A Graphic Illustration for the Formation and Stability Mechanism of Nanoparticles
Diverse technologies and measurements including EE, PDI, particle size, zeta potential, turbidity, CD, TEM, FE-SEM, DSC, XRD, and FTIR were used to make clear the formation and stability mechanism of zein/sophorolipid/polysaccharide ternary delivery system (Figure 7). After zein and sophorolipid were rapidly added into polysaccharide (polysaccharide: zein mass ratio ≤ 1:2), sufficient CMC or ALG acted as a shielding effect (electrostatic repulsion) and steric hindrance stabilizer was coated onto the surface of S/Z particles to prevent their sedimentation at low pH 3~4 range condition, showing a homogeneous PDI, turbidity and size based on DLS and UV. Certain internal drives (electrostatic interaction, hydrogen bonding and hydrophobic effect) participated in the formation of DHF-CMC/S/Z and DHF-ALG/S/Z, and CMC and ALG had no significant impact on secondary structure of zein in S/Z according to CD, FTIR, and FE-SEM. Upon employing EE, XRD, DSC, TEM analysis, 7,8-DHF was shown to be successfully encapsulated in CMC/S/Z and ALG/S/Z with relatively uniform sphericity, displaying a good entrapment efficiency.

A Graphic Illustration for the Formation and Stability Mechanism of Nanoparticles
Diverse technologies and measurements including EE, PDI, particle size, zeta potential, turbidity, CD, TEM, FE-SEM, DSC, XRD, and FTIR were used to make clear the formation and stability mechanism of zein/sophorolipid/polysaccharide ternary delivery system (Figure 7). After zein and sophorolipid were rapidly added into polysaccharide (polysaccharide: zein mass ratio ≤ 1:2), sufficient CMC or ALG acted as a shielding effect (electrostatic repulsion) and steric hindrance stabilizer was coated onto the surface of S/Z particles to prevent their sedimentation at low pH 3~4 range condition, showing a homogeneous PDI, turbidity and size based on DLS and UV. Certain internal drives (electrostatic interaction, hydrogen bonding and hydrophobic effect) participated in the formation of DHF-CMC/S/Z and DHF-ALG/S/Z, and CMC and ALG had no significant impact on secondary structure of zein in S/Z according to CD, FTIR, and FE-SEM. Upon employing EE, XRD, DSC, TEM analysis, 7,8-DHF was shown to be successfully encapsulated in CMC/S/Z and ALG/S/Z with relatively uniform sphericity, displaying a good entrapment efficiency.

Storage Stability of 7,8-DHF
During storage, preventing food neutraceuticals from heat or light exposure lenging, but critically necessary for mitigating degradation. To meet this applicatio short and long term storage were investigated under varying environments for 7,8 loaded binary and ternary nanoparticles. As shown in (Figure 8A [50], besides, due to the different chemical structures of CMC and ALG CMC/S/Z have higher EE, which caused more 7,8-DHF to be protected in the hydro part. These results are in agreement with previous studies that introduced curcumin embedded in zein and quaternized chitosan complexes [51], along with work showi quercetagetin was loaded using the zein-hyaluronic acid binary complexes [52].   [50], besides, due to the different chemical structures of CMC and ALG, DHF-CMC/S/Z have higher EE, which caused more 7,8-DHF to be protected in the hydrophobic part. These results are in agreement with previous studies that introduced curcumin being embedded in zein and quaternized chitosan complexes [51], along with work showing that quercetagetin was loaded using the zein-hyaluronic acid binary complexes [52].  [50], besides, due to the different chemical structures of CMC and ALG, D CMC/S/Z have higher EE, which caused more 7,8-DHF to be protected in the hydroph part. These results are in agreement with previous studies that introduced curcumin b embedded in zein and quaternized chitosan complexes [51], along with work showing quercetagetin was loaded using the zein-hyaluronic acid binary complexes [52].

In Vitro Simulated Gastrointestinal Digestion
A gastrointestinal tract (GIT) model was applied to study the digestive fate and accessibility of 7,8-DHF in different formulations. Particle size changes were monitor a specific digestion time (30, 60, 120, and 180 min), and the results were presented in ure 9A). The mean particle size of DHF-S/Z was significantly increased after 60 min

In Vitro Simulated Gastrointestinal Digestion
A gastrointestinal tract (GIT) model was applied to study the digestive fate and bioaccessibility of 7,8-DHF in different formulations. Particle size changes were monitored at a specific digestion time (30, 60, 120, and 180 min), and the results were presented in ( Figure 9A). The mean particle size of DHF-S/Z was significantly increased after 60 min SGF digestion (p < 0.05). This finding was possibly due to the fact that the S/Z were exposed to ionic strength along with low pH and partially digestion via pepsase. The low pH and ionic strength exposure likely weakened electrostatic repulsion forces among the nanoparticles [52]. Particle size reduction in DHF-S/Z post-SIF-exposure was attributed to the fact that SIF contains bile salt with strong emulsifying ability. Bile salt can bind many biopolymer molecules and induce bridging flocculation [21]. The particle size of DHF-CMC/S/Z was increased after exposure to the stomach phase but remained relatively constant during incubation in SIF. This behavior suggested that the existence of CMC strengthened intestinal stability of DHF-CMC/S/Z. However, the particle size of DHF-ALG/S/Z remained fairly steady throughout simulated GIT, only showing a large increase at 30 min during the SGF incubation. The different influence of polysaccharide type (CMC or ALG) on gastrointestinal fate of colloidal carriers can be attributed to different molecular characteristics. FE-SEM microscopic observation further confirmed that exposure to simulated gastrointestinal conditions had a significant effect on the morphology of the 7,8-DHF-loaded nano-complexes ( Figure 9C). For DHF-S/Z, irregular shapes have developed after they were added into the SGF and SIF, similar to an anomalous sheet structure. This shape acquisition was due to irregular aggregates appearing after SGF digestion. Furthermore, DHF-CMC/S/Z and DHF-ALG/S/Z exhibited a spherical shape. After gastrointestinal digestion, the nanoparticles possessed a relatively spherical morphology, resembling the cross-linked structure of large nanoparticles. Overall, CMC and ALG effectively protected the stability of DHF-S/Z nanoparticles through the GIT. Particularly, ALG performed extremely well.

Conclusions
In the present study, we compared the performance of ternary nanop existence of two selected polysaccharides (CMC and ALG) for encapsulatio 7,8-DHF but with low bioavailability. CMC/S/Z exhibited lower PDI, particle bidity, but higher zeta potential and loading capacity compared to ALG/S/Z. both polysaccharides supplementations promoted the EE value of 7,8-DHF cially CMC. Invitation of polysaccharides displayed positive effects on the f physical stability (pH and ionic strength stability) of ternary complexes. The ternary complexes mainly occurred via hydrophobic effects, hydrogen bond trostatic interactions. More significantly, compared to S/Z, ALG/S/Z, and CM After being exposed to simulated gastrointestinal conditions, 7,8-DHF bioaccessibility was measured after centrifugation and collection of micelle phases. As shown in Figure 9C, the S/Z was broken down when exposed to SGF digestion, and the core of 7,8-DHF was released, resulting in a low bioaccessibility. The bioaccessibility of DHF-S/Z gradually increased in SIF digestion, primarily due to the emulsibility of bile salts and sophorolipids. Another reason for this increase was the binding of hydrolyzed peptides of zein protein [33]. In the presence of CMC, low levels of bioaccessibility were exhibited in SGF digestion. This behavior was introduced by the vulnerability of CMC to acid, pepsin enzymes, and the alkali ions of SGF juices [53]. However, the bioaccessibility of DHF-CMC/S/Z was higher than that of DHF-S/Z, indicating a synergistic effect existed among CMC, sophorolipid and zein on controlling the release of 7,8-DHF. Furthermore, the presence of ALG further increased bioaccessibility to 75.46% ( Figure 9B), showing the highest bioaccessibility of 7,8-DHF after GIT digestion compared to DHF-CMC/S/Z and DHF-S/Z. These results indicated that ALG increased the solubilization capacity of mixed micelles in small intestine fluids. Collectively, our results demonstrated that encapsulating 7,8-DHF in polysaccharide coated S/Z nanoparticles can promote an appreciable increase in its in vitro bioaccessibility.

Conclusions
In the present study, we compared the performance of ternary nanoparticles in the existence of two selected polysaccharides (CMC and ALG) for encapsulation of bioactive 7,8-DHF but with low bioavailability. CMC/S/Z exhibited lower PDI, particle size and turbidity, but higher zeta potential and loading capacity compared to ALG/S/Z. Furthermore, both polysaccharides supplementations promoted the EE value of 7,8-DHF in S/Z, especially CMC. Invitation of polysaccharides displayed positive effects on the formation and physical stability (pH and ionic strength stability) of ternary complexes. The formation of ternary complexes mainly occurred via hydrophobic effects, hydrogen bonding and electrostatic interactions. More significantly, compared to S/Z, ALG/S/Z, and CMC/S/Z obviously enhanced the storage stability and in vitro bioaccessibility of 7,8-DHF. CMC/S/Z possessed a higher storage stability for 7,8-DHF. In contrast, ALG/S/Z had a better in vitro bioaccessibility of 7,8-DHF. Collectively, the results of this study indicate that selected polysaccharides containing composite nanoparticles are efficient at encapsulating, retaining, and delivering 7,8-DHF, and might therefore be utilized in dietary supplements and functional foods. Future work will focus on the applicability of DHF-ALG/S/Z and DHF-CMC/S/Z in complex water-phase beverage systems. Besides, the transepithelial transport mechanism of DHF-ALG/S/Z in an Caco-2 cell model, and in vivo pharmacokinetic studies in rat will be also studies.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/foods10112629/s1, Figure S1: The chemical structure of 7,8-DHF. Figure S2: Size distributions of nanoparticles with different zein to polysaccharide mass ratios, CMC/S/Z (A), ALG/S/Z (B). Figure S3: Size distributions of nanoparticles with different zein to polysaccharide mass ratios at pH = 4, CMC/S/Z (A), ALG/S/Z (B). Figure S4: The photograph of each colloidal particle at different pH conditions. Figure S5: The photograph of each colloidal particle at different NaCl concentrations.
Author Contributions: Conceptualization, methodology, resources, funding acquisition and writingoriginal draft preparation, Y.C.; software and formal analysis, J.P. and X.G.; investigation and data curation, L.W. and Y.S.; writing-review and editing, D.W.; project administration, Y.W.; validation and supervision, G.X. All authors have read and agreed to the published version of the manuscript.