Soy protein isolated was purchased from Beijing OKA Biotechnology Co., Ltd. (Beijing, China). Camellia oil was bought from a local supermarket. Bamboo was provided by China National Bamboo Research Center. All the other chemicals were analytical grade.

Bamboo shoot protein concentrate was extracted from bamboo shoot following a previous method with some modifications (12). Briefly, the bamboo shoots were mixed with MilliQ water at 1:10 (w/v) and the pH was adjusted to 8.0 using 0.1 M NaOH solution. Subsequently, the mixture was constantly stirred in a water bath at 45°C for 2 h and then filtered with 80-mesh sieves. After centrifugation at 8,000 rpm for 25 min at 4°C, the supernatants were collected, and the precipitates were redissolved for centrifuge again. Then, BPC was obtained by spray drying and stored in a refrigerator at−20°C for further analysis.

The SPI was dissolved in MilliQ water at a concentration of 4 wt%. BPC was added to the SPI solution with mass ratios of 0:1, 1:4, 2:4, 3:4, and 4:4 (wt/wt), respectively. To prepare SPI/BPC complexes, a certain amount of camelia oil (10% v/v) was added into the solution of SPI/BPC with the aqueous phase by continuously stirring, followed by the pH adjustment to 8.0 using 0.1 M NaOH solution. The mixture was further homogenized using a high-speed homogenizer (IKA T25 digital ULTRATURRAX®, Germany) at 15,000 rpm for 3 min. Afterward, the emulsions were prepared by high-pressure homogenization (Nano DeBEE30, BEE International, USA) at 25,000 psi for 5 min at 25°C.

The three-phase contact angle was measured on an OCA 20 contact angle meter (Data Physics Co. Ltd., Germany). Prior to measurement, the SPI/BPC complexes were pressed to the films with a thickness of about 2 mm. A syringe was used to transfer 3.5 μl of the solutions onto the surface, where the solutions were kept for 15 min. The contact angle was calculated by the SCA software.

The infrared spectra of the particles were recorded on an FTIR instrument (Nicolet iS50, Thermo Fisher Scientific, China). The spectra were scanned with the absorbance mode of 4 cm⁻¹ in the wavenumber range of 4,000–400 cm⁻¹. All tests were conducted with air as the background.

The secondary structure of SPI and BPC was quantitatively determined by the analysis of the amide I region (1,700–1,600 cm⁻¹) in the FTIR spectra. The contents of α-helix, β-sheet, β-turn, and random coil components were calculated using Peakfit 4.12 software (13).

The droplet size or diameter of the emulsions stabilized by SPI/BPC complex was measured by dynamic light scattering according to the method of Chantrapornchai et al. (14) with slight modification. Samples were diluted 1,000 times by the MilliQ water, and then the droplet size of the emulsions was measured on a Zetasizer Nano-ZS90 (Malvern Instruments, Worcestershire, UK). The refractive index of water and camellia oil were 1.33 and 1.462, respectively. All measurements were carried out at 25°C.

The zeta potential of SPI/BPC emulsions was measured by the Zetasizer Nano-ZS90 (Malvern Instruments, Worcestershire, UK) at 25°C. Each sample was measured three times.

The morphology characterization of emulsion droplets was performed using the TEM (Tecnai G2 Spirit, Thermo FEI, Czech Republic) at an accelerating voltage of 120 kV with the negative staining method (15). The samples (0.4 mg/ml) were added to a copper grid stained with phosphotungstic acid (1% w/v). The excess stain was removed, and the copper grids were dried at room temperature.

The IFT of the emulsions was measured on a tensiometer (OCA 20, Data Physics Co. Ltd., Germany) with the pendant drop method. The droplet of distilled emulsion was suspended on the tip of a syringe (22 gauge, I-quip), and then the pictures of the droplet were captured by a video camera. The IFT values were calculated by Young-Laplace equation (16).

where R₁ and R₂ are the principal radii of curvature, γ is the liquid surface tension coefficient, ΔP is the Laplace pressure across the interface.

A strain sweep test was performed to determine the linear viscoelastic region from 0.01 to 100% at 1 Hz. The storage modulus (G′) and loss modulus (G″) of sample were measured by frequency sweeping from 0.1 to 100 Hz at a constant stain of 1% (19).

Viscosity was investigated as a function of shear rate on a rotational rheometer (MCR 302, AntonPaar, Graz, Austria), and the measurements were performed using a plate/plate measuring setup (PP25). The shear rate was increased from 0.1 to 100 s⁻¹.

The model of Herschel-Bulkley was described by Eqs (2–5):

where σ (Pa) and σ₀ (Pa) are the shear stress and the yield stress, respectively. K (Pa Sⁿ) is the consistency index, γ (s⁻¹) represents the shear rate, and n represents the flow behavior index.

The POV was used to measure the primary oxidation products, which was detected using the method of Zhu et al. (20) with slight modification.

As an indicator of the secondary oxidation product, TBARS were evaluated using a previously reported method of Pan et al. (21) with slight modification.

All experiments were performed in triplicate with parallel samples. All the data were carried out by one-way ANOVA to evaluate significant differences (p < 0.05). Origin 9.0 software and SPSS18.0 were used for data analysis.

Pure SPI and BPC initially had the contact angle values of 76.6 and 51.3°, respectively, but the three-phase contact angle of BPC was reduced to 12.7° when standing for 7.5 min, while there was no obvious change of SPI in the three-phase contact angle. With the SPI/BPC ratio ranging from 4:1 to 2:1, the initial contact angle value of the complexes increased from 61.8 to 71.2° (Figure 1). This result might be attributed to the presence of the hydroxyl group in BPC, improving the surface hydrophobicity of the complexes by exposing the hidden hydrophobic area in the SPI (22). When the SPI/BPC ratio was 4:3 and 1:1, the contact angle of the complexes decreased to 49.0 and 39.6°, respectively, probably due to the presence of hydrophilic groups (such as OH group) carried by excessive BPCs (23). Previous results showed that the use of the particles with intermediate hydrophobicity as a stabilizer could improve the emulsion stability (24). Therefore, the SPI/BPC complex particles at 2:1 might have the best ability to stabilize the emulsions since the contact angle value was the closest to 90°.

Figure 2A shows the FTIR spectra of SPI and SPI/BPC particles at various mass ratios. There was a broad band over a wavenumber of 3,700–3,200 cm⁻¹, associating with the free stretching vibration of hydroxyl groups in proteins (25). Compared with SPI, the band intensity of SPI/BPC complexes at about 3,300 cm⁻¹ was enhanced, indicating hydrogen bonds formed between SPI and BPC. The peak positioned in the wavenumber range of 2,990–2,850 cm⁻¹ represented the CH antisymmetric and symmetric stretching of methyl (CH₃) and methylene (CH₂) groups that appeared in aliphatic side chain of proteins, and then the peaks (2960 and 2926 cm⁻¹) in the spectra were corresponding to the stretching (26). Moreover, the left shift of the C-N stretching and N-H bending vibration at 1,542 cm⁻¹ was observed, which might be associated with the electrostatic attraction between the COO group in SPI and NH3+ group in BPC, thereby promoting the formation of SPI/BPC complexes (27).

As shown in Figure 2B, compared with BPC (45.7%), the β turns and random coil content of SPI/BPC particles from 4:1 to 1:1 increased from 63.5 to 71.2%, respectively, reflecting the formation of a more flexible structure that favored the adsorption and unfolding of protein molecules at the oil–water interface and then the enhanced emulsifying properties (13).

The mean size of the SPI-stabilized emulsions was about 200 nm and basically unchanged during 7 days of storage at 25°C (Figure 3A). However, the BPC-stabilized emulsions remained stable for only a few hours (data not shown). When the SPI/BPC complex at the mass ratio of 2:1 was employed as a stabilizer, the mean size of emulsion droplets increased from 200 to 595 nm during 7 days storage. As the mass ratio of BPC in the complex further increased, an increasing trend of the emulsions in droplet size was observed. Especially at the ratio of 1:1, the emulsions were inclined to separate after 7 days. These results indicated that a proper BPC concentration favored the faster adsorption of the complex on the water-in-oil interface, resulting in decreased IFT and then smaller droplet size (28). Moreover, the use of SPI/BPC complex at different ratios as stabilizers did not significantly affect the droplet size distribution of emulsions (Figure 3B), which agreed well with the previous reports of Wang et al. (29).

The zeta potential of emulsions stabilized by SPI and SPI/BPC are shown in Table 1, and the absolute value of the zeta potential was used for comparison. When the mass ratio of BPC in the complex increased from 4:1 to 1:1, the zeta potential decreased significantly from −26.00 to −20.57 mV compared with the pure SPI stabilized emulsion, which was −38.13 mV, and this may be attributed to the hydrophobic interactions between SPI and BPC (30). Meanwhile, when the zeta potential decreased, the increase in the size of emulsions may be caused by the decreasing electrostatic repulsion and interfacial strength formed by SPI and BPC as the ratio of BPC increased (31).

The droplets of the pure SPI-stabilized emulsions possessed spherical shape and uniform size, with a diameter of approximately 200 nm (Figure 4), in agreement with the result of droplet size measurements. In the previous study, protein attached to the O/W interface could lead to a protein network structure (32). With the addition of BPC, the oil droplets were surrounded by SPI/BPC complexes, but the excess BPC may not absorb the oil-water interface, thus affecting the characteristic of the emulsions.

The interfacial behavior of a stabilizer is considered an important factor influencing emulsion formation and stability. The reduced IFT and the mechanical energy barrier formed at the interface can prevent the emulsion system against coalescence (33). As shown in Table 1, SPI had the highest IFT value of 50.72 mN/m. With the increasing ratio of BPC in the complex, the IFT value of the complex decreased to 50.25 (4:1), 49.91 (2:1), 49.16 (4:3), and 47.46 (1:1) mN/m, respectively. According to Karaca et al. (34), the more soluble proteins favored to reduce IFT because the proteins could shift to the interface. This phenomenon has also been proven by Zhang et al. (35), who found that the IFT of all the four samples decreased as the SPI content increased, which helped more SPI molecules get adsorbed on the air-water interface.

For all the emulsions, the G′ values were higher than G″ values at any given frequency (Figure 6A), suggesting the dominant elastic properties of the emulsions (42). Both G′ and G″ values of the emulsions at different SPI/BPC ratios increased with the increasing frequency ranging from 0.01 to 10 Hz, indicating that the modulus was frequency-dependent (43). In addition, with the ratio of BPC in the complex increased from 4:1 to 1:1, the G′ value was gradually increased to 3.66 Pa, suggesting a progressive enhancement in the gel-like structures (44).

The apparent viscosity (η) of the emulsions stabilized by SPI/BPC complex at various mass ratios gradually decreased with the shear rates increasing from 0.1 to 100 s⁻¹ (Figure 6B), indicating a typical non-Newtonian pseudoplastic behavior. As the ratio of BPC increased from 1:0 to 1:1, the η value of the complex stabilized emulsions varied from 0.005 to 0.154 Pa·s at a shear rate of 10 s⁻¹, which could lead to enhanced interaction between the droplets (44).

The Herschel–Bulkley model was used to evaluate the flow properties of the emulsions. As shown in Figure 6C, the shear stress vs. shear rate data at all the mass ratios showed a good correlation with Herschel–Bulkley models (R² > 0.99). The K value increasing with the increased ratio of BPC in the complex reflecting the change in the viscosity of emulsion and the shear-thinning behavior of the emulsions as a non-Newtonian fluid. (45). Furthermore, the n values of all emulsion were <1 (Table 2), which confirmed the pseudoplastic fluid characteristics. The increased ratio of BPC in the emulsions decreased the n value, suggesting that the deviation from Newtonian behavior gradually increased (46).

Figure 7 shows the changes in POV and TBARS of the emulsions as a function of storage time at 25°C. POV indicates the levels of hydroperoxides formed from fatty acid radicals during oxidation processes, while TBARS shows the content of malondialdehyde that is a secondary lipid oxidation product from polyunsaturated fatty acids (47). After 7 days of storage, the POV value of the emulsions stabilized by the SPI/BPC complex at the mass ratio of 1:1 increased remarkably from almost zero to 171 μmol/kg oil, revealing the occurrence of lipid oxidation in the emulsions during storage. In addition, the POVs of emulsions increased from 89.3 to 171 μmol/kg oil with the ratio of SPI/BPC increasing from 4:1 to 1:1. The changes in TBARS contents had a similar trend to those of POVs, indicating that the primary oxidation products might be transformed into secondary oxidation products (48). Generally, emulsion system trended to be oxidized due to the existence of an oil–water interface, and some unabsorbed molecules could affect the lipid oxidation process like the excess BPC content in the ratio of 1:1 (49).