Structural characteristics and emulsifying properties of myofibrillar protein-dextran conjugates induced by ultrasound Maillard reaction.

In this study, we investigated the effect of the ultrasound-assisted Maillard reaction on the structural and emulsifying properties of myofibrillar protein (MP) and dextran (DX) conjugates with different molecular weights (40, 70 and 150 kDa). Compared with classical heating, mild and moderate ultrasound-assisted methods (100-200 W) could accelerate the later stage of the Maillard reaction, which increased the degree of graft (DG) and the content of advanced Maillard reaction products (MPRs). Structural analysis revealed conjugates obtained by Maillard reaction induced the loss of ordered secondary structures (α-helix, β-sheets) and red-shift of maximum emission wavelength of intrinsic fluorescence spectrum. The conjugate containing 40 kDa DX exhibited higher extent of Maillard reaction compared to those containing 70 kDa and 150 kDa DX under various treating methods. Moreover, the ultrasound-assisted Maillard reaction could effectively improve the emulsifying behaviors. 100 W ultrasound-induced conjugates grafted by 70 kDa DX produced the smallest emulsion size with optimum storage stability. Confocal laser scanning microscopy and analytical centrifugal analyzer further confirmed MP grafted by 70 kDa DX with the assistance of 100 W ultrasound field could produce the smallest and most homogeneous MP-base emulsion with no flocculation. Our study demonstrated that mild ultrasound treatment resulted in well-controlled Maillard reaction, and the related glycoconjugate grafted with 70 kDa DX showed the greatest improvements in emulsifying ability and stability. These findings provided a theoretical foundation for the development of emulsion-based foods with excellent characteristics.

Introduction

Myofibrillar protein (MP), the main component of meat proteins, is an important ingredient in the food industry due to its excellent gelation and water-binding properties [1]. In recent year, there has been growing interest in the development of MP-base oil-in-water emulsions because of the potential applications in newly-type functional foods [2], [3]. However, MP is insoluble in solutions with relatively low ionic strength (<0.3 M NaCl) and easily aggregates into a complex structure, which is hard to diffuse to interface during emulsification [4]. Moreover, although nearly half of the amino acids in MP are hydrophobic, most of the hydrophobic residues are buried in the interior molecular structure, and the resulting low surface hydrophobicity further limits its application in emulsion systems [5]. Hence, many attempts, including physical, chemical and enzymatic methods, have been made to modify the structural and physicochemical properties of MP for improving its interfacial properties [6], [7].

One promising approach to enhance the functional properties of proteins is by direct conjugation with a hydrophilic polymer via Maillard reaction (MR) [8]. This spontaneous chemical reaction involves the covalent attachment of protein with carbonyl groups from reducing sugars, as well as the formation of Schiff bases and Amadori compounds [9]. Glycation (Maillard reaction) is considered as a green modification method as its ingredients are naturally edible [10], [11]. Compared with non-covalent protein-polysaccharide complexes, covalently bonded Maillard reaction products (MRP) are more stable under various environments and tolerant to changes in pH, temperature and ionic strength [12], [13]. In addition, MRP may endow a fine oxidation resistance to emulsions as some of the bioactive compounds formed during the final stages of the MR [14]. So far, successes have been reported in the improvement of the functional properties of MP, such as increasing the ionic and thermal stability by glycation with alginate oligosaccharide [15] and enhancing solubility by conjugation with konjac oligo-glucomannan [16]. However, studies regarding the emulsifying properties of MP glycoconjugates have not been performed to date, and thus a suitable glycation method is worth exploring.

The traditional methods used to prepare MRP include dry- and wet-heating treatments. These approaches are time-consuming, requiring several weeks (dry-heating alone) and tens of hours (wet-heating alone), respectively [17]. Moreover, some complex protein structures may prevent reactive amino acids from participating in the Maillard reaction [18], [19]. Recently, new emerging technologies, including pulsed electric field, microwave radiation, high-pressure homogenization and high pressure, have been applied to promote the Maillard reaction with less processing time [20], [21], [22]. High-intensity ultrasound (20 kHz) also shows great potential in accelerating the Maillard reaction [23]. Ultrasonic cavitation can increase the excluded volume and the flexibility of protein, which speeds up the insertion of hydrophilic groups into the protein structure [24], [25]. In addition, the local translational motions induced by ultrasound field increases the proximity of protein molecules to reducing sugars.

Dextran (DX), a neutral polysaccharide, is composed of a linear chain of glucose residues linked through α bonds (1 → 6). The conformation of dextran is highly flexible in aqueous solution a poor aggregation ability [26], and the neutral charge property of dextran also inhibits electrostatic complexing between proteins and polysaccharides [27], [28], which makes DX as an ideal material for glycosylation with proteins. Nevertheless, there is little information regarding the influence of the molecular weight of DX on the Maillard reaction process, and an in-depth investigation of the changes that occurs in the MP-DX system at the molecular level in ultrasound-induced Maillard-type modification is required.

Therefore, in this work, DX with different molecular weights (40, 70 and 150 kDa) was grafted to MP with the assistance of different ultrasound treatments (0–300 W). The degree of grafting (DG), browning intensity, and Fourier transform infrared (FT-IR) spectra were applied to evaluate the progress of Maillard reaction. In the meantime, the structural changes and aggregate behaviors were determined by fluorescence spectroscopy and laser diffraction technique, respectively. Moreover, the emulsifying properties of various MP/DX conjugates were investigated according to microstructure of MP-base emulsion as well as their flocculation and creaming stability, meanwhile, the possible mechanisms were also discussed. Special attention was paid to the relationship between enhanced emulsifying properties and the degree of glycation.

Materials and methods

Materials

Fresh golden threadfin bream (1–1.5 g) were purchased from Haixin Foods Co., Ltd. (Fujian, China), and the fish were slaughtered following standard industrial procedures. Fat was trimmed away, and muscle was sliced into 20-mm-thick chops. The samples were stored at −80 °C prior to use (within two days). Commercial soybean oil was obtained from a local supermarket. Dextran (molecular weights of 40, 70 and 150 kDa) was obtained from Yuanye Biotechnology Co. Ltd. (Shanghai, China). Nile red was obtained from Acros Organics (Shanghai, China). All chemicals and reagents in this study were analytical grade. Distilled deionized water was used for the preparation of all solutions.

Preparation of MP

MP was prepared according to our previous method [29] with slight modifications. Thawed golden threadfin muscle was homogenized in five volumes of solution A (pH 7.0, containing 0.1 M NaCl, 2 mM MgCl2 and 1 mM EDTA) for 60 s with a PRO-12S blender (German Pool Ltd., Hong Kong, China) at 11,000 r/min. The mixture was centrifuged (2000 × g for 20 min) at 4 °C (Avanti J-E; Beckman Coulter, Brea, CA, USA). The sediment was collected and washed three times using solution A under the same centrifugation conditions. The pellet was then collected and washed three times with solution B (0.05 M NaCl). The resulting pellet was resuspended and homogenized in four volumes of solution B, and then the pH was adjusted to 6.25 with 0.05 M HCl. The final MP pellet was stored on ice and utilized within 24 h. The protein concentration was determined by the Biuret method using BSA as the standard.

Preparation of MP/DX conjugates

According to our previous method described by Chen et al. [4] with slight modifications, MP (6.0%, w/v) and DX (40, 70 and 150 kDa; 12.0%, w/v) were fully dispersed by magnetic stirring at 4 °C overnight. Different MP/DX solutions were designated as MP/LDX (40 kDa), MP/MDX (70 kDa), and MP/HDX (150 kDa). In the next day, one-hundred-milliliter aliquots of the MP-XG solutions were transferred to a jacketed beaker (250 mL), and an ultrasound processor (Scientz Biotechnology Co., Ltd., Ningbo, China) with a 1.5-cm diameter titanium probe was used to sonicate the mixtures in a temperature-controlled shaking water bath set at 70 °C and 10 rpm. Samples were treated at 20 kHz with different levels of power output (100 W, 200 W, 300 W) for 40 min (pulse duration of on-time 2 s and off-time 2 s). The MP/DX dispersions were cooled to room temperature in an ice bath and dialyzed at 4 °C for 24 h after ultrasound treatment. Finally, the resulting products were lyophilized, finely ground, and kept in a desiccator for further use. The same MP/DX mixture was treated by water bath (WB) alone at 70 °C for 12 h. Samples were designated according to the grafting methods and molecular weight of DG. For example, 100-W-UMP-LDX means the MP/LDX conjugate was prepared with 100 W ultrasound treatment; 300-W-UMP-MDX means the MP/MDX conjugate was prepared with 300 W ultrasound treatment.

DG measurement

Free amine groups were measured using the O-phthaldialdehyde (OPA) method. The OPA solution was prepared as described by Wang et al. [12]. Two hundred microliters of conjugate solution were mixed with 4 mL of OPA reagent and distilled water was used as a blank. The absorbances were measured at 340 nm after incubation at 35 °C for 2 min. The DG of different WB-induced and ultrasound-induced conjugates was calculated as follows:

DG = (Ac − At)/A0 × 100%,where A0 and Ac are the concentrations of free amino groups of untreated MP and conjugates, respectively.

Ftir

The freeze-dried conjugates were mixed with KBr powder at a certain proportion of 1:100 (w/w), and then the resulting powder was pressed in a mortar in order to be pressed into slice. The FT-IR spectra were recorded at a range of 4000 to 400 cm−1 by a Thermo Nicolet 5700 FTIR spectrometer (Thermo Electron, Madison, WI, USA). EZOMNIC analysis software was applied to analyze the change in infrared spectrum with wave number as the coordinate and transmittance as the ordinate.

Measurement of the Maillard reaction progress

The early, intermediary, and advanced MRP could be determined by the absorbance measurements at 284, 304 and 420 nm, respectively, and this these could reflect the progress of the Maillard reaction. According to the method described by María et al. [30] with slight modifications, all the conjugate solution were diluted to a protein concentration of 5 mg/mL, and then the absorbance values at 284, 304 and 420 nm were determined using a Lambda 20 spectrophotometer (Perkin Elmer, USA) at 25 °C.

Measurement of the secondary structure

The secondary structure of MP conjugates was investigated according to a previously reported method [16] with minor modification. Each sample was diluted to 50 μg/mL and transferred to a 0.1-cm path length quartz cell. Molecular ellipticity was measured from 200 to 260 nm at a scan rate of 100 nm/min using an Applied Photophysics Chirascan Spectropolarimeter (Applied Photophysics, Co. Ltd., Leatherhead, UK). The response time was 2 s. The percent concentration of different secondary structure conformation units was calculated by Dichroweb, which is found on the online Circular Dichroism website (http://dichroweb.cryst.bbk.ac.uk).

Surface hydrophobicity

The surface hydrophobicity of the samples was measured using 8-anilino-1-naphthalenesulphonic acid (ANS) on the base of the method reported by Chen et al. [4] with some modifications. Each sample was diluted to 1 mg/mL and mixed with a 15 mM ANS solution (phosphate buffer, 20 μL, pH 7.0). After incubation for 20 min at 25 °C, fluorescence was measured using a fluorescence spectrophotometer (F-7000; Hitachi Corp., Japan) at an excitation wavelength of 380 nm and emission wavelengths of 410–570 nm. The fluorescence intensity was used to measure surface hydrophobicity.

Reactive sulfhydryl groups (R-SH)

The determination of reactive sulfhydryl compounds was performed using a method previously described by Li et al. [31] with some modifications. A 10-mM DTNB solution (phosphate buffer, 50 μL, pH 8.0) was added to 4 mL of sample solution (1 mg/mL) and mixed well. Samples were incubated for 20 min at 25 °C, and the absorbance was measured at 412 nm using a Lambda 20 spectrophotometer (Perkin Elmer, USA). The R-SH content was calculated from the molar extinction coefficient (EM = 13,600).

Fluorescence spectrometry

The fluorescence spectra of samples were measured using the procedure reported by Li et al. [32] with minor changes. Conjugate solutions were prepared by diluting to a final protein concentration of 0.2 mg/mL. The fluorescence spectra were scanned with an excitation wavelength of 280 nm, and the emission spectra are recorded in the range of 300–450 nm using a fluorescence spectrometer (F-7000; Hitachi Corp., Japan). The voltage was set to 700 mV.

Solubility

The solubility was calculated as soluble protein against the original protein presented in samples and is measured as described by Chen et al. [33].

Particle size

Particle size distribution and polydispersity index (PDI) of conjugates were analyzed using a Nano ZS Zetasizer instrument (Malvern Instruments, Worcestershire, UK). According to the method of Chen et al. [34], each sample was diluted to 1 mg/mL and loaded into the cuvette (PCS8501). The refractive indices of dispersion and continuous phase were taken as 1.467 and 1.330, respectively, and particle size data are presented as mean diameter (nm).

Preparation of emulsions

Dispersions of different samples were mixed with soy oil (9:1, v/v), and sodium azide (NaN3) was added to retard microbial growth. The mixture was subsequently emulsified using an UltraTurrax blender (KA T18, IKA-Werke GmbH & Co. KG, Germany) at 12,000 rpm for 2 min. The pH of the emulsion was immediately adjusted to 7.0 using 0.1 M NaOH. Finally, a 5-mL emulsion sample was transferred to a glass test tube and tightly sealed with a black plastic cap.

Droplet size distribution

The distribution of emulsions was measured by using a Mastersizer 2000 (Malvern Instruments Co. Ltd., Worcestershire, UK). The emulsion samples were diluted ten times with phosphate buffer (20 mM, pH 7). A refractive index of 1.520 was used for emulsion particles and 1.331 for the phosphate buffer. The laser obscuration level was set at 10%, and the samples were equilibrated for 90 s inside the instrument before the data were collected. The particle sizes refer to the volume-weighted mean diameter, which is calculated as followswhere ni is the number of droplets with diameter di.

d3, 2 = Σnidi 3/Σnidi 2

Confocal scanning laser microscopy (CLSM)

The microstructure of the emulsion samples was observed using a Leica TCS SP8® microscope (Leica Microsystems GmbH, Wetzlar, Germany) with a 40× (NA 1.25) oil immersion objective lens. A 40 μL staining solution containing 0.1% (w/v) Nile Red was mixed with 1 mL emulsion samples. Approximately 50 μL of stained emulsions were placed on a laboratory-made welled slide. The stained emulsions were examined using an argon krypton laser having an excitation line of 488 nm and a helium neon laser (He/Ne) with excitation at 633 nm.

Creaming stability measurements of emulsion

The creaming stability of emulsion samples was measured by a LUMiSizer analytical centrifugal analyzer (L.U.M. GmbH, Germany), employing centrifugal separation to accelerate the instability phenomenon. The instrumental parameters used for the measurement were as follows: rotational speed, 3000 rpm; temperature, 25 °C; time interval, 30 s; and total time of the experiment, 4 h.

Statistical analysis

Samples were independently prepared in triplicate for analysis. Variance analysis and mean separations were done through Duncan’s multiple-range test (p < 0.05), using IBMSPSS Statistics computer software Version 23 (International Business Machines Corp., Armonk, NY, USA)

Results and discussion

Degree of grafting

As shown in Fig. 1, the DG of the MP/DX conjugates prepared by WB method increased as the molecular weight of grafted DX decreased, indicating the DX with lower molecular weight was more reactive in the Maillard reaction. This could be explained by weaker steric effect made LDX more available for active groups in MP molecules. Moreover, the DG of all conjugates significantly increased (p < 0.05) when ultrasound intensity increased from 100 to 200 W, but it rapidly decreased (p < 0.05) as ultrasound intensity further increased to 300 W. Ultrasound treatment could assist the dissociation of the complex protein structure [35], [36], causing nucleophilic residues, including lysine, histidine and guanidino arginine, to be exposed to the accessible region, which resulted in a greater progress of the Maillard reaction. Moreover, sonocatalysis effect might be another important reason led to a more frequent occurrence of glycation [23]. However, strong ultrasound treatment might heavily denature the protein, resulting in large aggregate through the enhancement of hydrophobic interactions and/or protein–protein hydrogen bonds [37], where the reactant residues were likely to be embedded into the newly-formed clusters before reacting with carboxide. In addition, the maximum DG value of UMP-HDX was 12.28% under 200 W, which was significantly lower (p < 0.05) than that of UMP-LDX (19.32%) and UMP-MDX (17.89%). On one hand, great steric hindrance of HDX reduced the chances of contact between protein and polysaccharide. On the other hand, the long polysaccharide chains bound to the protein could pose a shielding effect that masked the reactant residues for the following Maillard reaction.

Fig. 1

DG value of various MP/DX conjugates. Different capital letters (A-D) indicate significant differences among different Maillard reaction methods (p < 0.05), and different small letters (a-c) indicate significant differences among various treated DX (p < 0.05).

FTiR

FTIR spectroscopy is an effective method to analyze the protein-polysaccharide interactions and structural changes at the molecular level. As can be seen from Fig. 2, for the native MP, the peaks around 1650 cm−1 were attributed to amide I and were related to C = O stretch; the elastic vibration at 1530 cm−1 (amide II) was from N-H bend and C-N stretch; the amide III (1150 cm−1) peaks were formed by the interaction of N–H angular vibration and CN stretching vibration. In addition, the peak near 2120 cm−1 was influenced by C-O bending, and the peak at 3350 cm−1 was related to the changes in O–H and N–H bending of the protein molecules. After glycation with DX, the absorption peaks of amide I, amide II and amide III were red-shifted to various degrees. Abdelhedi et al. [38] also reported an increase in the covalent cross-linking between protein and polysaccharide, which contributed to the certain shift of these bands. Moreover, the absorption intensity of those absorption bands became stronger and was accompanied by a wider absorption band at 3350 cm−1, indicating some new chemical bonds were formed after Maillard reaction. Maillard reaction involved an increased number of schiff base (C = N), pyrazine (C-N), and Amadori compounds (C = O) [39]. For MP conjugates grafted by LDX, the absorption intensity of all bands was stronger than those of MDX and HDX under both WB and ultrasound conditions. This suggested LDX could exert a greater effect on the MP structure and induced more MRPs formation. Furthermore, conjugates obtained by ultrasound-assisted method enlarged these effects, and the distinct absorption bands of 200 W UMP-LDX, UMP-MDX and UMP-HDX were larger than other corresponding samples, reflecting that moderate ultrasound treatment had greater potential to accelerate Maillard reaction and exerted more marked modification effect on structural properties of glycoconjugates.

Fig. 2

FT-IR spectra of MP and various MP/DX samples: (A) WB-induced conjugates, (B) 100 W ultrasound-induced conjugates, (C) 200 W ultrasound-induced conjugates, (D) and 300 W ultrasound-induced conjugates.

Absorbance of the MP/DX conjugate

The absorbance at 284 nm (A284) was measured as an indicator of the early products of the Maillard reaction; the schiff base of the MPR could be monitored by the absorbance at 304 nm (A304); the absorbance at 420 nm (A420) represented the advanced stage of the Maillard reaction [40]. These indexes are commonly used to evaluate the extent of the Maillard reaction [41]. As shown in Fig. 3, it could be observed that the DX with lower molecular weight led to greater absorbance of the conjugate. MP glycated with LDX under both classical heating and ultrasound treatment showed higher A284, A304 and A420 than those of MDX and HDX, again confirming polysaccharide with a lower molecular weight contributed more to the progress of Maillard reaction at all stages. These results were partly in accordance with the previous studies carried out in the WP-DX systems, where the Maillard reaction between whey protein and the DX with lower molecular weight took a shorter time with better reaction kinetics [41].

Fig. 3

Absorbance at 284 nm (A), 304 nm (B) and 420 nm (C) of various MP/DX conjugates. Different capital letters (A-D) indicate significant differences among different Maillard reaction methods (p < 0.05), and different small letters (a-c) indicate significant differences among various treated DX (p < 0.05).

Moreover, 100–200 W ultrasound-induced MPR exhibited significantly higher A420 (p < 0.05), compared with the corresponding WB samples, while A284 and A304 were significantly decreased (p < 0.05). It seemed that ultrasound-assisted glycation accelerated Maillard reaction mainly at its later stage, resulting in more advanced MPR at the expense of its early products. Wang et al. [42] also pointed out the browning intensity (A420) of BPI (bean Vigna radiate protein)-glucose conjugate prepared by ultrasound treatment was higher than that formed by WB method. Hence, it could be demonstrated that a high-intensity shear field induced by ultrasound promoted the polymerization of intermediate products during Maillard reaction. Nevertheless, Chen et al. [33] suggested side reactions, such as caramelization, could also be enhanced by ultrasound, which might cause a decrease in advanced MPR. In addition, the A284, A304 and A420 were significantly (p < 0.05) lower at 300 W ultrasound treatment than those of the wet method or 100–200 W treatments. This implied the suppression effect of graft process under strong ultrasound treatment was mainly due to the inhibition of Maillard reaction in the early stage.

Structural properties

Secondary structure

The secondary structures of MP and various MP/DX conjugates were presented in Table 1. The main structure of native MP was α-helix (48.21%) and β-sheet (20.7%). Generally, α-helical structures are stabilized by hydrogen bonds between the carbonyl oxygen (–CO) and amino hydrogen (NH–) of single polypeptide chain, while β-sheets are supported by inter-chain hydrogen bonds between peptide chains [31]. These structures were decreased along with increasing in β-turn and random coil contents after glycation. Covalent binding of polysaccharides to proteins during Maillard reaction involved a condensation between the carbonyl groups and ε-amino groups, which were mainly distributed in the α-helix and β-sheet regions [43]. The potential mechanism for these changes might be that the grafting of polysaccharide induced protein denaturation and unfolding, which promoted the disruption of both intramolecular and intermolecular hydrogen-bonded structures. Hence, more MRP formation meant more ordered structures (α-helix and β-sheet) were transformed into disordered structures (β-turns and random coils).

Table 1

Effect of different Maillard reaction methods on secondary and tertiary structures of MP.

Samples	Secondary structure				Tertiary structure
	α-helix	β-sheet	β-turn	random coil	S0-ANS	R-SH
MP	48.21 ± 3.31a	20.7 ± 0.57a	17.84 ± 1.8e	13.25 ± 1.23d	582.78 ± 16.09c	5.79 ± 0.19e
MP-LDX	42.63 ± 2.76b	16.94 ± 0.3b	20.79 ± 1.17d	19.64 ± 2.33c	653.3 ± 44.89a	7.83 ± 0.12c
MP-MDX	45.12 ± 1.53ab	18.27 ± 1.28ab	18.81 ± 0.35e	17.8 ± 1.87c	657.92 ± 50.32a	7.31 ± 0.87 cd
MP-HDX	47.33 ± 1.89a	19.08 ± 1.97a	17.97 ± 2.75e	15.62 ± 2.05 cd	660.34 ± 21.08a	6.75 ± 0.65d
100 W UMP-LDX	32.48 ± 4.07d	11.53 ± 1.06d	30.85 ± 1.66ab	25.14 ± 0.76ab	668.7 ± 39.64a	8.96 ± 0.39ab
100 W UMP-MDX	37.31 ± 0.44 cd	13.92 ± 0.49c	26.07 ± 3.08b	22.7 ± 1.84b	659.64 ± 20.57a	9.27 ± 0.68a
100 W UMP-HDX	43.57 ± 1.88b	17.16 ± 2.12b	21.91 ± 2.3 cd	17.36 ± 0.97c	671.72 ± 44.29a	7.89 ± 1.17c
200 W UMP-LDX	29.64 ± 0.15e	10.78 ± 0.73e	32.11 ± 1.46a	27.47 ± 0.21a	661.5 ± 62.10a	8.49 ± 0.75b
200 W UMP-MDX	34.21 ± 0.82d	13.54 ± 1.45c	27.8 ± 2,19b	24.45 ± 2.55ab	674.22 ± 48.92a	8.65 ± 0.69b
200 W UMP-HDX	40.26 ± 3.02c	15.88 ± 1.86b	22.47 ± 0.89c	21.39 ± 2.37b	656.69 ± 56.31a	7.13 ± 0.94 cd
300 W UMP-LDX	30.38 ± 2.9e	19.41 ± 1.57ab	23.08 ± 2.45c	27.13 ± 3.89a	606.9 ± 10.57b	7.06 ± 0.08 cd
300 W UMP-MDX	33.96 ± 0.34d	20.75 ± 2.44a	20.02 ± 1.94d	25.27 ± 4.12ab	618.72 ± 36.9b	6.67 ± 0.36d
300 W UMP-HDX	39.92 ± 2.75c	21.23 ± 0.95a	17.68 ± 3.67e	21.17 ± 3.48b	590.01 ± 18.79c	6.37 ± 0.67d

All the data are expressed as mean ± SD. Means with the different superscript letters within the same column are significantly different (p < 0.05).

With the same DX molecular weight, more structural features were lost for the conjugates assisted by 100–200 W ultrasound treatment than corresponding WB samples, and the main reasons were summarized as followed: (1) Local high temperature, shear forces, and shock waves induced by cavitation bubbles destroyed the protein–protein interactions within the complex MP structure, where more hydrogen-bonded structures were dissociated. (2) Maillard reaction might introduce complex groups to hydrogen-bonded sites in the protein molecule, and external ultrasound field further enhanced this phenomenon, in this case, molecular repulsion induced by grafted DX suppressed the refolding of glycoconjugates. (3) During Maillard reaction, the remaining long-chain polysaccharides could interact with protein through non-covalent interactions, which rearranged the hydrogen bond scheme of MRP and further contributed to shielding of the inter-molecular hydrogen bonds. Additionally, further increase in ultrasound intensity to 300 W caused a significant increase (p < 0.05) in β-sheet composition at the expense of β-turns. Strong ultrasound-induced thermal effects, accompanied with excessive formation of reactive free radicals, promoted the heavily aggregation of denatured protein rather than remaining in an unfolded state [44], in which weakly hydrogen-bonded β-turn structure was transformed to a strongly hydrogen-bonded β-sheet.

Tertiary structure

In general, hydrophobic and sulfhydryl groups (SH) were encased in the compact MP interior [20]; thus, changes in surface hydrophobicity (S0-ANS) and reactive sulfhydryl (R-SH) group contents could reflect the degree of tertiary structure changes. With the introduction of DX, both S0-ANS and R-SH were increased in varying degrees (Table 1), which might be explained by the fact that glycation induced the uncoiling of ordered structures, leading to the exposure of active residues originally embedded in MP molecules. As the molecular weight of DX decreased, the R-SH of glycoconjugates increased significantly (p < 0.05), while the S0-ANS decreased slightly (p > 0.05). The degree of denaturation and unfolding increased with an increase in grafted polysaccharide, which accounted to the higher R-SH content. However, the introduction of more polyhydroxy molecules into MP meant more hydrophilic groups were coupled on the surface of the conjugates, resulting in a decrease in surface hydrophobicity.

The R-SH content of UMP-LDX, UMP-MDX and UMP-HDX further increased under 100 W ultrasound treatment, but it was decreased when 200–300 W ultrasound field was applied. The increase in ultrasound intensity induced higher injection speed into the swirling cavitation chamber, causing a stronger drop in pressure in the vortex center [37], which resulted in a stronger cavitation phenomenon for molecules stretching. This effect increased the degree of Maillard reaction, and increased Maillard reaction further promoted the dissociation and unfolding of MP. Nevertheless, a rapid increase in temperature and a great amount of free radical formation made the protein more reactive, in which the ultrasound-induced aggregate re-appeared with the consequent masking and inactivation of SH groups. Moreover, the R-SH content of 100-W-UMP-MDX sample was significantly higher (p < 0.05) than that of other samples, which was partly opposite to the change in DG value. This was due to that more grafted polysaccharide might increase the chances to mask the neighboring R-SH. Thus, the mild glycation process was more suitable for generating exposed functional groups. In addition, there was no significant change (p > 0.05) in S0-ANS for glycoconjugates with the same DX molecular weight under 0–200 W ultrasound treatments, which suggested the ultrasound-induced and glycation-induced exposure of hydrophobic regions could well compensate for the embedded hydrophilic groups, therefore stabilized the hydrophilicity/hydrophobicity balance.

Fluorescence analysis

The fluorescence spectra at 300–450 nm was measured to characterize the microenvironmental changes around the tryptophan (Trp), tyrosine (Tyr) and phenylalanine (Phe) residues [45]. As shown in Fig. 4, when excited at 280 nm, the maximum emission intensity (λmax) of MP was at 328 nm, confirming most of the chromophore groups were active in hydrophobic environment. The λmax of the glycoconjugates grafted with LDX, MDX and HDX were red-shifted to 333, 332 and 330 nm, respectively. This further proved that glycation was accompanied by the unfolding of peptide chains, and those amino acids tended to move toward a hydrophilic environment under a more flexible conformation. For the ultrasound-assisted MRPs, the fluorescence intensity was further increased, and a more obvious red-shit of the λmax was observed under 100 W treatment. The reason for this phenomenon might be that protein molecules stretched and became loose. The increase in ultrasound intensity at 200 W led to a decrease in fluorescence intensity for various glycoconjugates with no marked change in λmax. With a high level of DG value, DX tended to form a molecular layer around the MP, which reduced the detectable autofluorescence value. Liu et al. [46] deemed that the density of the outward polysaccharide layer was depended on the nature of the proteins (denaturation, unfolding and aggregation) and DG value. Wang et al. [42] also reported ultrasound pre-treatment contributed to reducing fluorescence intensity of protein-polysaccharide conjugates. Furthermore, further increase in ultrasound intensity caused an obvious blue-shit of the λmax, indicating those amino acids were converted to more inaccessible microenvironment [47], [40]. Hydrophobic interactions coupled with protein–protein hydrogen bonds played the dominated role under strong ultrasound field rather than protein-polysaccharide interaction.

Fig. 4

Intrinsic fluorescence emission spectra of MP and various MP/DX samples: (A) WB-induced conjugates, (B) 100 W ultrasound-induced conjugates, (C) 200 W ultrasound-induced conjugates, (D) and 300 W ultrasound-induced conjugates.

Solubility and particle size

As shown in Table 2, the solubility of native MP was 13.72%, which was almost insoluble with poor protein-water interaction. After glycation with various DX, the protein solubility increased to 60.89% (LDX), 55.03% (MDX) and 46.76% (HDX), respectively. The hydrophilic groups in the polysaccharides could enhance the affinity of proteins to water, meanwhile, steric effects provided by the glycated DX also suppressed the re-aggregation of protein. Therefore, although MP-LDX had a higher DG value, weaker steric effects of LDX could account for lower solubility than that of MP-MDX. It was worth noting that 100 W ultrasound treatment further improved glycoconjugates solubility, and 100-W-UMP-MDX showed the highest solubility at 84.61%. The ultrasound field triggered intense mechanical forces, which contributed to the breaking of insoluble aggregates into soluble ones [35]. On the other hand, the unfolded structure with more grafted polysaccharides increased the amount of binding water. The further increase in ultrasound intensity of 200 W caused a slight reduction (p > 0.05) of protein solubility. This finding was similar to some previous studies that dense polysaccharide layer would partly shield the charge groups on the protein surface, thus suppressing the electrostatic repulsion among conjugates [23]. However, heavy denaturation and aggregation of MP would occur under 300 W ultrasound treatment, which resulted in a marked decrease (p < 0.05) in the solubility of corresponding glycoconjugates.

Table 2

Protein Solubility, Particle size, and polydispersity index (PDI) of different conjugate samples.

Samples	Solubility (%)	Particle size (nm)	PDI
MP	13.72 ± 2.9 g	255.25 ± 19.87a	0.386 ± 0.09a
MP-LDX	60.89 ± 4.97c	114.19 ± 22.45e	0.207 ± 0.08f
MP-MDX	55.03 ± 6.33 cd	125.04 ± 12.86d	0.242 ± 0.03e
MP-HDX	46.76 ± 3.76e	150.74 ± 18.77 cd	0.311 ± 0.06c
100 W UMP-LDX	76.97 ± 3.35b	86.21 ± 11.61f	0.148 ± 0.09gh
100 W UMP-MDX	84.61 ± 0.94a	69.96 ± 5.32 g	0.116 ± 0.01 h
100 W UMP-HDX	58.36 ± 2.41c	125.65 ± 17.19d	0.219 ± 0.08f
200 W UMP-LDX	73.38 ± 4.36b	90.17 ± 9.86f	0.162 ± 0.02 g
200 W UMP-MDX	82.24 ± 1.52a	74.11 ± 4.97 g	0.123 ± 0.02 h
200 W UMP-HDX	56.77 ± 5.69c	133.49 ± 16.61d	0.245 ± 0.06e
300 W UMP-LDX	51.38 ± 4.18d	161.83 ± 10.92c	0.276 ± 0.07d
300 W UMP-MDX	46.53 ± 5.52e	159.4 ± 20.47c	0.294 ± 0.08 cd
300 W UMP-HDX	38.36 ± 2.67f	187.87 ± 17.46b	0.35 ± 0.06b

All the data are expressed as mean ± SD. Means with the different superscript letters within the same column are significantly different (p < 0.05).

In order to further investigate the aggregate behaviors of MP/DX conjugates, the mean particle size and polydispersity index (PDI) were measured. The mean particle size of native MP was 255.25 nm with a high PDI value (0.386), which reflected MP molecules were unstable and tended to aggregate into various particle sizes. In case of WB-induced conjugates, the mean particle sizes dropped to 114.19 nm (LDX), 125.04 nm (MDX), 150.74 nm (HDX), respectively, and PDI was also decreased significantly (p < 0.05). This proved Maillard reaction could somewhat dissociate existing protein aggregates and discourage further aggregates formation. Moreover, external ultrasound field processed complex and obvious effects on the aggregate behaviors of conjugates. The mean particle size and PDI decreased sharply after treatment at 100 W for different DX. In particular, the mean particle size of 100-W-UMP-MDX showed the greatest decrease with the lowest PDI. When the treatment intensity was enhanced to 200 W, there was no significant change (p > 0.05) in mean particle size and PDI compared with the corresponding 100-W-treated samples. Hence, it was assumed that the highly-ordered MP aggregates were dissociated into fragments, oligomers and their glycation products in mild or moderate ultrasound-assisted samples. Generally, MP began to associate through interactions of the HMM region at 40–60 °C, while strong heat-induced clustered aggregates were formed due to interactions of the LMM region above 60 °C [20]. Considering the temperature of Maillard reaction process was>60 °C, it was likely that the highly ordered MP structures were fragmented into smaller sub-filament structures by synergistic effects of suitable ultrasound intensity and Maillard reaction, and subsequently DX were grafted in the HMM and LMM regions, which hardly permitted heat-induced aggregation at these regions. However, the mean particle size and PDI were significantly increased (p < 0.05) as the ultrasound intensity further increased to 300 W. The strong ultrasound treatment could induce a greater extent of denaturation with exposure of LMM [48], [49], where they could rapidly couple with each other before interactions with DX. Overall, ultrasound-assisted Maillard reaction led to increase in solubility and protein-water interactions, reduction in particle size, as well as the formation of better amphiphilic molecules, which might contribute to emulsifying properties.

Droplet size of emulsion

The average particle size (d32) of the emulsions formed by various glycoconjugates were shown in Table 3. The emulsion stabilized by native MP had an average droplet size of approximately 73.72 μm, and it completely delaminated after 7 days of storage. Large protein aggregates made MP hard to diffuse to the droplets, at the same time, these protein particles were less effective in covering oil droplets, which would account for the undesirable emulsifying ability and stability. Conversely, the presence of DX exhibited a more stable droplets size of 30.12 μm (MP-LDX), 34.19 μm (MP-MDX), and 40.34 μm (MP-HDX), respectively, which did not delaminate within 15 days. In this study, glycation with DX decreased the emulsion droplet size via three possible routes: (1) Glycation process improved the amphiphilic properties of MP, therefore the resulting glycoconjugates were prone to be tightly adsorbed at the surface of oil droplets. (2) The increase in solubility and molecular flexibility promoted either structural rearrangement or proteins cross-linking at the interface. (3) DX of the outer MP-DX layer could provide steric repulsion and increase bulk viscosity, thereby locking the droplets in place and retarding their upward movement. This was in agreement with the result from Zhang [28], who found covalently linked OPI–dextran conjugates facilitated a viscoelastic protein film against flocculation.

Table 3

Effects of different Maillard reaction methods on volume-weighted average particle size (D2,3) of various MP/DX-base emulsions.

Samples	Average emulsion size (D2,3)
	0 day	5 days	15 days
MP	73.72 ± 6.39a	ND	ND
MP-LDX	30.12 ± 2.98f	74.53 ± 3.1d	202.35 ± 9.65bc
MP-MDX	34.19 ± 3.32e	81.25 ± 2.57 cd	224.09 ± 4.88bc
MP-HDX	40.34 ± 1.77c	92.46 ± 6.95b	296.3 ± 7.96ab
100 W UMP-LDX	21.33 ± 1.2 h	40.86 ± 4.47 g	98.73 ± 2.53e
100 W UMP-MDX	18.59 ± 0.38i	35.82 ± 3.19 h	84.1 ± 5.23f
100 W UMP-HDX	28.88 ± 0.69f	65.08 ± 4.5e	181.57 ± 8.68c
200 W UMP-LDX	26.78 ± 2.55 fg	60.39 ± 1.73ef	146.65 ± 5.5d
200 W UMP-MDX	25.53 ± 1.46 g	53.8 ± 3.68f	129.77 ± 7.34d
200 W UMP-HDX	38.91 ± 3.39c	86.52 ± 0.97c	257.84 ± 8.2b
300 W UMP-LDX	37.45 ± 0.74 cd	95.03 ± 5.48b	313.89 ± 8.7a
300 W UMP-MDX	42.07 ± 2.99c	130.89 ± 7.60a	ND
300 W UMP-HDX	66.7 ± 4.82b	ND	ND

All the data are expressed as mean ± SD. Means with the different superscript letters within the same column are significantly different (p < 0.05).

Under the same conditions, it was showed that 100 W ultrasound-assisted conjugates were more effective in improving the emulsification properties than those using the WB method. Mild ultrasound treatment further increased protein solubility and molecular flexibility with higher surface-activity, which accelerated diffusion, absorption and structural rearrangement to an appropriate conformation during emulsification. On the other hand, the increase in grafted polysaccharide reinforced steric hindrance and increased the density of the viscoelastic layer around the emulsion droplets [50]. Emulsion stabilized by 200-W-treated conjugates showed larger droplet size than those of corresponding 100 W samples for 0–15 days, indicating emulsions were more prone to coalescence and creaming during storage. Generally, protein is unlikely to lose its amphiphilic properties after being grafted with nearly 12.28–19.32% DX. Hence, it was proposed that glycoconjugates with excess polysaccharide became large and less flexible, which might spend more time to anchor in oil droplet. The stability of the 300-W-conjugate-coated emulsions were further reduced, especially for the 300-W-UMP-HDX sample, which was completely aggregated and flocculated after 7 days of storage. The large ultrasound-induced protein and conjugate aggregates tended to be randomly arranged at the interface, which would not effectively reduce interfacial tension, meanwhile, some high-molecular-weight conjugates might act as flocculation bridges among the droplets, accelerating flocculation, coalescence and creaming. In addition, the emulsion stabilized by 100-W-UMP-MDX had the smallest droplet size with the minimum change during storage. This suggested the 100 W ultrasound treatment with modest molecular weight DX contributed to a well-controlled Maillard reaction for emulsifying modification of MP, in this case, this pre-treated conjugate could cross-link into a thick, continuous and viscoelastic film on the emulsion surface. This physical barrier could stabilize the oil droplets against phase separation during storage.

CLSM

The microstructure of emulsions was directly observed by confocal laser scanning microscopy. As expected, the MP-stabilized emulsion showed a large and heterogeneous microstructure involved in flocculation (Fig. 5). In case of MP/DX conjugates, the extent of flocculation decreased in some degree, and relatively smaller oil droplets were observed for the MP-LDX sample although those droplets still appeared to be loosely connected with each other. These results might be mainly associated with the more absorbed protein caused by an increase in molecular hydrophobicity, and it helped to reduce the free energy required to create a new interface. Moreover, grafted DX enhanced the interaction between proteins and the aqueous phase, improving the stability of conjugates at the interface. However, the relatively lower DG of WB-treated conjugates meant extra steric hindrance provided by DX was relatively weak, and could not completely suppress the flocculation.

Fig. 5

Confocal laser scanning microscopy images of fresh emulsions emulsified by various MP-DX and UMP-DX with selected Maillard reaction methods.

Furthermore, more uniform and dispersed emulsion droplets were observed in the 100-W-ultrasound-assisted glycoconjugates than those of the corresponding WB samples. In particular, 100-W-UMP-MDX developed the smallest droplet size with no flocculation. The suitable change in MP conformation modified by synergism between ultrasound treatment and Maillard reaction improved the emulsifying behaviors. Moreover, increase in the amount of grafted DX within a proper range not only reinforced the layer covered by MP, but also contributed to interfacial viscoelastic response [51]. Nevertheless, when the 200 W ultrasound-assisted glycoconjugates with higher DG were applied, the flocculation phenomenon appeared again with a larger emulsion size. Like WB and 100-W ultrasound-treated samples, HDX exerted the negative influence. Long-chain-polysaccharides might suppress the structural rearrangement and cross-linking with conjugates at neighboring sites. As the ultrasound further increased to 300 W, all the corresponding emulsions became highly-flocculated with severe aggregation. A strong ultrasound field would induce highly-aggregated structure that could not be precisely and preferentially anchored at the interfacial phase. This was in agreement with the phenomenon observed from other strong ultrasound-assisted systems [52], where protein products were hard to interact with phase interface due to heavily ultrasound-induced molecular denaturation and aggregation.

Accelerated physical stability of emulsions

Analytical centrifugal analyzers are generally used to evaluate creaming stability, which is expressed as a space- and time-related transmission profile over the entire sample length. During centrifugation, the phase separation occurred; the light and opaque oil phase moved to the top while the heavy and transparent aqueous phase moved to the bottom of each sample cell. Therefore, fewer changes of the transmission and a gentle slope of the resultant curve indicated smaller degree of phase separation. As shown in Fig. 6, it could be found the creaming rate decreased to various degrees when MP/DX conjugates were used. In case of glycoconjugates, hydrophobic protein moieties provided rapid adsorption to the surface of oil droplets, and the hydrophilic polysaccharide moieties bound to the aqueous phase and prevented the aggregation of oil droplets via strong steric and/or electrostatic repulsion [36]. Maillard conjugates improved both hydrophobic and hydrophilic moieties at the interface, contributing to the creaming stability under thermal processing, prolonging storage period and in-vitro digestion conditions [53]. In addition, the reduction in the creaming rate was more obvious as the molecular weight of grafted DX decreased (higher DG). According to the structural changes of Maillard conjugates, tendency in solubility and amphipathicity accounted for this phenomenon.

Fig. 6

Evolution of transmission profiles of MP-base emulsions emulsified by various conjugates with selected Maillard reaction methods.

In comparison with WB-induced conjugates, emulsions stabilized by 100-W-treated conjugates exhibited stronger creaming stability. More flexible structures with higher grafted DX not only increased entanglement and cross-linking sites to assemble into multilayered structure, but also enhanced the steric hindrance against creaming. This effect was partly mitigated when ultrasound intensity increased to 200 W. As expected, emulsions stabilized by all 300-W-treated conjugates were rapidly phase separated as the slope of the resultant curve obviously increased. The changes in the curve of 300-W-UMP-MDX and 300-W-UMP-LDX samples were similar to those of the untreated MP sample, indicating negative ultrasound-induced influences could completely compensate for the positive effect of glycation. Furthermore, 100-W-UMP-MDX exhibited optimal creaming stability as the oil droplet was not clearly phase separated after 4 h centrifugation, further proved Maillard conjugates produced by 100 W ultrasound field with modest-molecular-weight DX was the most suitable products to develop MP-base emulsion. Since the relationship between glycation reaction conditions, DG value, and emulsifying properties was complicated, in our study, a partly unfolded MP molecules with modest amounts of grafted DX was the most stable conformation anchored in the interface. Therefore, mild ultrasound condition promoted and accelerated Maillard reaction, and a well-controlled Maillard reaction with low- and modest-molecular-weight DX endowed MP with beneficial structural properties and desirable emulsifying behaviors.

Conclusions

This study demonstrated that with ultrasound-assisted method, the grafting ability between myofibrillar protein (MP) and dextran (DX) through the Maillard reaction was markedly enhanced in terms of degree of grafting (DG) and FTIR. The extent of Maillard reaction was dependent on the ultrasound intensity and the molecular weight of DX. Compared with the traditional wetting method, mild ultrasound treatment accelerated the later stage of the Maillard reaction. Low-molecular-weight DX was more available for grafting with MP under both WB- and ultrasound-assisted methods. Highly ordered MP aggregates were disrupted into smaller oligomers and monomers under mild or moderate ultrasound field, meanwhile, MP molecules were prone to stretching and unfolding as more hydrophobic and reactive sulfhydryl residues were exposed, where some ordered secondary structures were gradually converted to disordered structures. Glycation also contributed to protein-water interactions and improved the solubility of grafted MP molecules with more hydrophilic groups. In contrast, strong ultrasound treatment was more likely to induce heavy denaturation and aggregation of MP than that of Maillard reaction, in which active residues were transformed into newly-formed aggregates with lower protein solubility and surface activity. Moreover, glycoconjugates formed from DX with different molecular weight induced by 100–200 W ultrasound treatments were capable of developing stable emulsion, which exhibited smaller average particle size than those of non-treated and 300-W-treated conjugates. Furthermore, glycoconjugates grafted with modest-molecular-weight DX exhibited better emulsifying ability and stability compared with those with corresponding low-molecular-weight and high-molecular-weight DX. The 100-W-UMP-MDX was the optimal emulsifying agent, and it could produce the most desirable and stable emulsion in different conditions. Therefore, it could be proved that a well-controlled Maillard reaction with suitable DG value contributed more to developing MP-based emulsion. Overall, ultrasound-assisted glycation method could effectively compensate for the insufficiencies of MP and broaden its application in the food industry.

CRediT authorship contribution statement

Zhiyu Li: Conceptualization, Methodology, Writing - original draft. Yimei Zheng: Methodology, Validation. Qian Sun: Investigation, Validation, Formal analysis. Jianyi Wang: Software, Methodology. Baodong Zheng: Resources. Zebin Guo: Conceptualization, Supervision, Project administration, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.