Quality attributes of the set-style yoghurt from whole bovine milk as affected by an enzymatic oxidative cross-linking

Abstract

The impact of an enzymatic oxidative cross-linking of milk proteins on quality attributes of set-style yoghurt was investigated in the present study. Whole bovine milk was pretreated by a ternary system comprising horseradish peroxidase, glucose oxidase and glucose at 4°C for 48 h and fermented by a commercial direct vat set starter after table sugar addition and pasteurization. The measured bacterial count in the treated milk was less than 2 × 10³ CFU/mL. Electrophoresis results revealed that both the treated milk and the treated yoghurt contained protein polymers about 11%. The treated yoghurt had the same main composition as the control yoghurt, but higher apparent viscosity, elastic and viscous moduli, larger thixotropic loops area (3.19 vs. 1.97 Unit) and lower syneresis (21.0% vs. 33.8%). Texture profile and microstructure analysis showed that the treated yoghurt had higher hardness (106.6 vs. 92.0 g) or adhesiveness (450 vs. 367 g s), and finer network structure.

En el presente estudio, se investigó el impacto que una reticulación enzimática y oxidativa de proteínas de leche tiene en los atributos cualitativos del yogurt firme (o compacto). Previamente, se trató durante 48 horas leche entera de vaca con un sistema ternario compuesto por peroxidasa de rábano, glucosa-oxidasa y glucosa a 4°C, el cual fue fermentado mediante un cultivo inicial comercial añadido al contenedor después de agregar azúcar común y de realizar la pasteurización. La medición de bacterias en la leche tratada fue de menos de 2 × 10³ CFU/mL. Los resultados de electroforesis revelaron que tanto la leche tratada como el yogurt tratado contenían polímeros proteínicos a un nivel de 11%. Si bien se constató que el yogurt tratado tenía la misma composición principal que el yogurt de control, su viscosidad aparente y sus módulos elásticos y viscosos fueron más elevados, además de mostrar un área de círculos tixotrópicos más grande (3,19 vs. 1,97 unidad) y sinéresis más baja (21,0% vs. 33,8%). El análisis del perfil de textura y de la microestructura mostraron que el yogurt tratado tenía una firmeza (106,6 vs. 92,0 g) y una adherencia (450 vs. 367 g) más altas, así como una estructura de red más fina.

Keywords:

• whole bovine milk
• horseradish peroxidase
• glucose oxidase
• cross-linking
• set-style yoghurt
• quality attributes

Palabras claves:

• leche entera de vaca
• peroxidasa de rábano
• glucosa-oxidasa
• reticulación
• yogurt firme
• atributos cualitativos

Introduction

Yoghurt is one of most important fermented dairy products in the world. Some novel technologies have been developed or applied to improve yoghurt quality at molecular level of milk proteins, for example, by forming covalent bonds between or within the molecules with enzymes (Fægemand, Sørenson, Jørgensen, Budolfsen, & Qvist, 1999). Transglutaminase (TGase, EC 2.3.2.13) is one of typical enzymes capable of forming isopeptide bonds between the γ-carboxyamide groups and the amino groups and thus can be used to induce cross-linking of food proteins (Beermann & Hartung, 2012; Lorenzen, Neve, Mautner, & Schlimme, 2002) or glycosylation plus cross-linking of casein or soybean proteins (Jiang & Zhao, 2010, 2012). The influences of TGase treatment on functional properties of milk proteins had been widely studied (Fægemand & Qvist, 1997; Hiller & Lorenzen, 2009), while the impacts of TGase treatment on the textural and other attributes of yoghurt products were also well characterized (Bönisch, Huss, Weitl, & Kulozik, 2007; Lorenzen et al., 2002). Cross-linking of milk proteins by TGase prior to fermentation could improve gel structure of the set-style yoghurt, resulting in finer protein network with thin strands between the particles (Fægemand & Qvist, 1997; Farnsworth, Li, Hendricks, & Guo, 2006). TGase treatment of milk proteins in different extents conferred an increase of 400–600% in gel hardness (Fægemand & Qvist, 1997), an enhancement of 25% in gel strength (Gauche, Tomazi, Barreto, Ogliari, & Bordingnon-Luiz, 2009) or a decrease of 37.8% in syneresis (Șanli, Sezgin, Deveci, Șenel, & Benli, 2011) on the yoghurt samples or acid gels, as the results of the cross-linking of milk proteins. Cross-linking of milk proteins by TGase also gave rise to the yoghurt product one- or two-folds increase in viscosity (Farnsworth et al., 2006; Ozer, Kirmaci, Oztekin, Hayaloglu, & Atamer, 2007) and enhanced elastic modulus (Anema, Lauber, Lee, Henle, & Klostermeyer, 2005).

Another type of enzymes, oxidoreductases, also can be used to modify milk proteins (Fægemand, Otte, & Qvist, 1998). For example, lactoperoxidase plus H₂O₂ (Hiller & Lorenzen, 2009; Ozer, Grandison, Robinson, & Atamer, 2003) laccase (Ercili Cura et al., 2009; Hiller & Lorenzen, 2009), Trichoderma reesei tyrosinase (Ercili Cura et al., 2010), peroxidases plus H₂O₂ (Fægemand et al., 1998; Li & Zhao, 2009) had been used to treat milk proteins for functionality modification. In one recent study, a ternary oxidative system consisting of horseradish peroxidase (HRP, EC 1.11.1.7), glucose oxidase (EC 1.1.3.4) and glucose showed a potential to cross-link caseinate by dityrosine formation (Chang & Zhao, 2012). For the ternary system, glucose oxidase catalyzed glucose oxidation to produce H₂O₂. H₂O₂ then was used by HRP to induce dityrosine formation (i.e. cross-linking) of the proteins. The ternary system was different from other oxidative systems employed in these above-mentioned studies, as it contained two oxidases and one reaction (glucose oxidation) was coupled by another reaction (caseinate cross-linking). The covalent bonds newly formed in the treated proteins by the ternary system were different from those bonds formed by TGase. It implies that the proteins treated by the ternary system would have different properties than those proteins treated by TGase. TGase has been well studied for its application in dairy processing, but potential application of the ternary system in dairy products is not studied so far, especially whether the quality attributes of set-style yoghurt might be affected by the ternary system is unknown yet.

In the present study, the mentioned ternary system was applied to pretreat whole bovine milk prior to yoghurt fermentation. The impacts of the carried out treatment on some quality attributes (mainly rheological and textural properties) of the set-style yoghurt prepared thereof were investigated, in order to reveal whether this enzymatic treatment is another suitable approach for the quality improvement of set-style yoghurt or other dairy products.

Material and methods

Materials and chemicals

Commercial whole bovine milk negative in antibiotics was purchased from a local supermarket in Harbin, China. Horseradish peroxidase with a declared activity of 20 kU/g was purchased from Shanghai Guoyuan Biotech, Inc. (Shanghai, China). Bis-acrylamide and glucose oxidase (type X-S) from Aspergillus niger with an activity of 130 kU/g were obtained from Sigma-Aldrich Co. (St. Louis, MO, USA). Acrylamide and glycine were obtained from Amresco, Inc. (Solon, OH, USA), whereas Tris was obtained from Solarbio Sci. & Technol. Co., Ltd. (Beijing, China). Standard protein markers (SM0431) were purchased from MBI Fermentas China Co. (Shanghai, China). A commercial direct vat set (DVS) starter (YO-MIX 499) consisting of Streptococcus thermophilus and Lactobacillus bulgaricus was obtained from Danisco GmbH. (Beijing, China). Other chemicals used were of analytical grade, while the water used was redistilled water.

Treatment of whole bovine milk and preparation of set-style yoghurt samples

The whole bovine milk was adjusted to a pH value of 6.8 by 2 mol/L NaOH solution, heated at 90°C for 5 min and cooled to ambient temperature. The milk of 1 L was poured aseptically into the sterilized containers, mixed with 200 U HRP, 6 U glucose oxidase and 0.05 mmol glucose per gram protein as the levels used in a previous study (Chang & Zhao, 2012), gently stirred for 5 min, stored at 4°C for 48 h and then served as treated milk. The milk treated as above but only by adding glucose oxidase and glucose was served as reference milk, while the milk treated as above but without the addition of the ternary system was served as control milk. The control, treated and reference milk prepared thereof were all subjected to microbiological and electrophoresis analysis.

The set-style yoghurt samples were prepared according to a reported procedure (Lee & Lucey, 2010). The control and treated milk mixed well with table sugar (at a level of 60 g/kg milk) were heated at 90°C for 5 min, rapidly cooled to about 42°C, and mixed with the DVS starter at a level of 0.03 g/kg milk, suggested by the producer, to obtain yoghurt samples in two groups, i.e. treated and control yoghurts. Then, yoghurt milk about 100 mL was poured into sterilized containers, placed in an incubator at 42°C for a fermentation time of 4 h till the value of pH fell to about 4.5. All prepared yoghurt samples were stored at 4°C for 24 h and assayed for main chemical composition, rheological, textural properties and microstructure.

Microbiological, chemical and electrophoresis analyses

Total bacterial count of the milk samples was assayed according to an IDF method (IDF, 1996), and expressed as colony-forming units per milliliter (CFU/mL) of the milk. Protein, fat, total solids and total reducing sugar contents (g/kg) were analyzed as the recommended methods (AOAC, 1995; IDF, 2001). The pH values were monitored by using a pH meter (Mettler, Toledo, DELTA-320 pH, Shanghai, China) calibrated by standard buffer solutions.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis under reducing conditions was performed as per the method of Laemmli (1970) by using a SDS Tris-glycine buffer system with 40 g/kg stacking gels and 120 g/kg resolving gels. The stained protein bands were quantitatively analyzed for protein content in the PhotoDoc-It Imaging System (UVP Inc., San Gabriel, CA, USA).

Rheological and textural evaluation

The assayed yoghurt samples was pretreated as the reported methods (Guyot & Kulozik, 2011; Ramchandran & Shah, 2010) and then measured for their apparent viscosity and visco-elasticity with the reference methods (Bourne, 2002; Meza, Verdini, & Rubiolo, 2009) in a Bohlin Gemini Rheometer (Malvern Instruments Limited, Worcestershire, UK) at 25°C. A cone-plate geometry (40 mm diameter, 4° angle and 150 μm gap) and shear rates from 0.1 to 10 per second were used in the measurement of apparent viscosity. Another cone-plate geometry (60 mm diameter, 0.5° angle and 15 μm gap) was used in the measurement of elastic (G′) and viscous (G″) moduli, with frequency sweeps in 0.1–10 Hz and 0.5% strain (within linear viscoelastic region).

Thixotropic behavior of the yoghurt samples was evaluated as per the method (Bönisch, Huss, Lauber, & Kulozik, 2007) in the Rheometer. The samples loaded on the inset plate were sheared at 500 per second for 1 min to diminish structural differences among the samples, and then equilibrated for 5 min to allow structural rebuilding. The flow curves were generated by measuring shear stress as a function of increasing shear rates (0.1–100 per second) within 30 min, holding the shear rate 100 per second within 3 min and deceasing shear rates from 100-0.1 per second within 30 min. The hysteresis loops area between the upward and downward flow curves was calculated by the software in the Rheometer.

Syneresis of the yoghurt samples was assayed by a centrifugation method (Farnsworth et al., 2006). The prepared yoghurt milk of 20 mL was fermented in sterilized centrifuging tubes of 50 mL at 42°C about 4 h until the pH value fell to 4.5 and centrifuged at 640 × g for 10 min at 25°C after a storage time of 24 h at 4°C. The supernatant was collected and weighed. Syneresis was calculated as percentage of the supernatant to the evaluated yoghurt sample on weight basis.

Yoghurt texture was evaluated by a texture analyzer, Texturometer (Model TA-XT2, Stable Micro Systems Ltd., Surry, UK) as the reference method of Bourne (1978) with a 5 kg load cell. The samples (60 mm diameter × 60 mm height) held in the containers were equilibrated to ambient temperature. A probe of 35 mm diameter (A/BE35) and two cycles were applied to a sample depth of 30 mm, with a constant crosshead velocity of 1 mm/s and a surface trigger of 10 g. From the generated force–time curves, four textural indices (hardness, adhesiveness springiness and cohesiveness) were calculated by using the TA-XT. Dimension Ver. 3.7 was software equipped in the Texturometer.

Microstructure observation

Sample fixation and examination followed a reference method (Sandoval-Castilla, Lobato-Calleros, Aguirre-Mandujano, & Vermon-Carter, 2004) with some modifications. Gels cubes (4 mm × 4 mm × 3 mm) prepared with a scalpel were fixed in 0.1 mol/L phosphate buffer (pH 6.8), containing 25 g/kg glutaraldehyde solution for 24 h at 4°C. The fixed specimens were dipped in the buffer (pH 6.8) for 10 min, washed three times with the buffer for 10 min interval and dehydrated in graded solutions of ethanol, ethanol/tert-butyl alcohol and tert-butyl alcohol for 15 min interval, respectively. The specimens were frozen by immersion in liquid nitrogen and freeze dried for 48 h and then kept in a desiccator to prevent moisture absorption. The dried specimens were mounted onto holders by an adhesive carbon membrane, coated with gold in a Hummer VI sputtering system (Matsushita Electric Industrial Co., Osaka, Japan), observed and photographed by Hitachi S-3400N Scanning Electron Microscope (Hitachi High-Technologies Co., Tokyo, Japan) at an accelerating voltage of 5 kV.

Statistical analysis

All preparations and analyses were carried at three times. All data were expressed as means or means ± standard deviations. The differences between the means of multiple groups were analyzed by one-way analysis of variance (ANOVA) with Duncan’s multiple range tests by using SPSS 16.0 software (SPSS Inc., Chicago, IL, USA). MS Excel 2003 (Microsoft Corporation, Redmond, WA, USA) was used to report the data.

Results and discussion

Treatment of whole bovine milk by the ternary oxidative system

When the whole bovine milk was treated by the ternary system at 4°C for 48 h, its pH value fell into 6.7 as the formation of gluconic acid by the glucose oxidase and glucose. Practical assaying results (Table 1) indicated that total bacterial count detected in the control milk after the treatment was about (36 ± 9) ×10³ CFU/mL, while that in the treated and reference milk was about (17 ± 5) and (5 ± 3) × 10² CFU/mL, respectively. That is, the ternary system and glucose oxidase plus glucose could retard bacterial growth in the milk during the treatment significantly (P < 0.05). Electrophoretic results in Figure 1 highlight a fact that the ternary system induced cross-linking of milk proteins in the treated milk, as some protein polymers with molecular weights (MW) greater than 116.0 kDa were only detected in the treated milk (Lane 2) but not found in the control (Lane 1) and reference (Lane 3) milk. These protein polymers amounted to about 11% of the total proteins, showing only part of the milk proteins was cross-linked. The control and treated milk were thus subjected for lactic fermentation to prepare control and treated yoghurt, respectively. The reference milk had no protein polymers and was not studied in later experiments.

Table 1. Total bacterial count detected in three milk samples after a treatment at 4°C for 48 h.

Tabla 1. Recuento bacteriano total detectado en tres muestras de leche tras el tratamiento a 4°C durante 48 horas.

	Control milk	Treated milk	Reference milk
Total bacterial count (CFU/mL)	(36 ± 9) ×10^3 C	(17 ± 5) ×10^2 B	(5 ± 3) ×10^2 A

Note: Different capital letters after the data in the same row as superscripts indicate that one-way ANOVA of the mean values is significantly different (P < 0.05).

The previous study showed that the ternary system was able to induce cross-linking of caseinate, leading to the formation of some protein polymers (Chang & Zhao, 2012). It was also found that treatment of milk proteins by lactoperoxidase led to the formation of protein polymers (Hiller & Lorenzen, 2009). When Trametes hirsute laccase (Ercili Cura et al., 2009) and Trichoderma reesei tyrosinase (Ercili Cura et al., 2010) were used to treat casein and milk proteins, the bands of protein polymers were detected by SDS-PAGE analysis. The aforementioned three studies provided support to the present study that the ternary system could induce protein cross-linking in the milk. One study reported that the skimmed milk powders treated by TGase at 3–10 U/g protein for 2 h had average degrees of protein polymerization of 23.6–33.9%, respectively (Guyot & Kulozik, 2011). Electrophoretic results from other two reported studies (Jaros, Jacob, Otto, & Rohm, 2010; Zhang, Liu, & Zhao, 2011) also showed that TGase treatment at 3–20 U/g protein for 2–4 h resulted in cross-linking of the most of caseinate. Different to the three reported results, the ternary system only induced lower cross-linking (about 11%) for the milk proteins than TGase. However, the ternary system showed an ability to control potential bacterial growth during the treatment besides cross-linking of the milk proteins, as some H₂O₂ was formed by the added glucose oxidase and glucose. H₂O₂ was more effective than refrigeration to retard bacterial growth in raw milk (Ozer et al., 2003). The bacterial growth in the treated or reference milk were thus retarded by the formed H₂O₂, leading to the detected bacterial count in lower level (2 × 10³ CFU/mL or less).

Impacts on quality attributes of set-style yoghurt

The chemical analysis results revealed that protein, fat, total solids and reducing sugar contents of the control and treated yoghurt samples were about 25, 27, 150 and 90 g/kg, respectively. For the measured four indices, no statistical difference (P > 0.05) was found between the two groups of yoghurt samples. It means that the carried out treatment totally had no impact on the main chemical composition of the yoghurt samples. However, the carried out treatment showed a slight impact on yoghurt fermentation (i.e. lactic acid production). A fermentation time of 4 h ensured milk coagulation in the control and treated milk, but the treated yoghurt had a slightly higher pH value than the control yoghurt (4.43 vs. 4.35) after the storage of 24 h at 4°C.

Electrophoretic results in Figure 1 indicate that the control yoghurt had no protein polymers (Lane 4) but the treated yoghurt contained some protein polymers (Lane 5). These protein polymers had MW greater than 116 kDa, amounted to about 11% of the total proteins. Unsurprisingly, this result was the same to the electrophoretic result for the treated milk.

Figure 1. SDS-PAGE analysis of the milk and set-style yoghurt samples under reducing conditions.

Figura 1. Análisis SDS-PAGE de muestras de leche y de yogurt firme bajo condiciones de reducción.

Notes: Lane M, standard protein markers; Lane 1, control milk; Lane 2, treated milk; Lane 3, reference milk; Lane 4, control yoghurt; Lane 5, treated yoghurt. Standard protein markers used and their molecular weights (MW) (in kDa) are as follows: egg albumin lysozyme (14.4), β-lactoglobulin (18.4), REase Esp 98 I (25.0), lactate dehydrogenase (35.0), ovalbumin (45.0), bovine serum albumin (66.2) and β-galactosidase (116.0).

Rheological properties of the control and treated yoghurt samples were evaluated, and the results are given in Figure 2. All samples showed typical shear-thinning behavior, and the treated yoghurt had higher apparent viscosity, elastic and viscous moduli than the control yoghurt (Figure 2a–c) during evaluation. Importantly, higher elastic modulus could result in the yoghurt better storage stability (Damin, Alcântara, Nunes, & Oliveira, 2009). The measured elastic modulus was much higher than the viscous modulus for all yoghurt samples (Figure 2b), declaring these samples had more solid-like (elastic) properties (Lee & Lucey, 2010; Ramchandran & Shah, 2010). At the same time, the treated yoghurt had progressively higher shear stress at the same shear rate (Figure 2c) and greater thixotropic loops area (Table 2, 3.19 vs. 1.97 Unit) than the control yoghurt. This result points out a fact that the evaluated sample (i.e. the treated yoghurt) was denser and needed greater shear stress to destroy the network (Bönisch, Huss, Lauber, et al., 2007; Guyot & Kulozik, 2011). This indicates that the treated yoghurt had a more stable structure than the control yoghurt, showing a helpful impact of the ternary system on yoghurt quality.

Table 2. Six measured properties of the control and treated yoghurt.

Tabla 2. Seis propiedades medidas del yogurt de control y del yogurt tratado.

Indices	Control yoghurt	Treated yoghurt
Hysteresis loop area (Unit)	1.97 ± 0.18^A	3.19 ± 0.07^B
Syneresis (%)	33.8 ± 1.3^B	21.0 ± 1.8^A
Hardness (g)	92.0 ± 3.5^A	106.6 ± 4.6^B
Adhesiveness (g s)	367 ± 17^A	450 ± 16^B
Springiness	0.940 ± 0.009^B	0.917 ± 0.007^A
Cohesiveness	0.296 ± 0.009^A	0.303 ± 0.004^A

Note: Different capital letters after the data in the same row as superscripts indicate that one-way ANOVA of the mean values is significantly different (P < 0.05).

Figure 2. Evaluation results for the apparent viscosity (a), elastic (G′) and viscous (G″) moduli (b) and thixotropic behaviors (c) of the control and treated yoghurt.

Figura 2. Resultados de la evaluación de la viscosidad aparente (a), módulos elásticos (G′) y viscosos (G″) (b) y comportamiento tixotrópicos (c) del yogurt de control y del yogurt tratado.

Formation of covalent bonds in milk proteins by cross-linking might confer the yoghurt gels enhanced rheological properties (Guyot & Kulozik, 2011; Ozer et al., 2003). TGase (Bönisch, Huss, Lauber, et al., 2007; Fægemand & Qvist, 1997; Ozer et al., 2007; Șanli et al., 2011) or TGase and glutathione (Bönisch, Huss, Weitl, et al., 2007) treatment of the milk resulted in the yoghurt significantly higher viscosity. At the same time, TGase and glutathione treatment of the milk also gave rise to the yoghurt greater hysteresis loops area (Bönisch, Huss, Weitl, et al., 2007). Treatment of milk by HRP in the presence of H₂O₂ led to the yoghurt sample higher elastic and viscous moduli (Wen, Kong, & Zhao, 2012). Due to larger molecular sizes of the cross-linked milk proteins by dityrosine formation, the treated yoghurt showed higher rheological properties such as viscosity, elastic and viscous moduli and thixotropic loops area than the control yoghurt, sharing accordance with these reported results.

Other measured properties of the control and treated yoghurt samples are given in Table 2, which also shows different quality attributes of these samples. Compared to the control yoghurt, the treated yoghurt had nearly 16% and 23% increase in hardness and adhesiveness (106.6 vs. 92 g, and 450 vs. 367 g s), respectively, but lower springiness (0.917 vs. 0.940). The cohesiveness of the treated yoghurt was not significantly (P > 0.05) impacted by the applied treatment. The treated yoghurt also had higher water holding capacity than the control yoghurt, reflected by lower syneresis (21.0 vs. 33.8%). Lower syneresis of the treated yoghurt was also supported by its stable structure as reflected in Figure 2c. These measured quality attributes also confirmed that the ternary system had a favorable impact the textural properties of the set-style yoghurt.

Many study results had indicated that cross-linking of milk proteins was helpful to yoghurt quality. Treatment of skim milk by TGase might obtain stiffer acidified gels (Fægemand & Qvist, 1997; Jacob, Nöbel, Jaros, & Rohm, 2011) or firmer yoghurt gels (Jacob et al., 2011; Lauber, Henle, & Klostermeyer, 2000). Cross-linking of the milk prior to yoghurt fermentation by TGase gave rise to less syneresis (Gauche et al., 2009; Guyot & Kulozik, 2011; Lauber et al., 2000; Lorenzen et al., 2002; Șanli et al., 2011), as the cross-linked protein molecules were able to stabilize the three-dimensional network of the yoghurt gels (Lorenzen et al., 2002). Oxidative cross-linking of caseinate or raw milk by laccase (Ercili Cura et al., 2009), Trichoderma reesei tyrosinase (Ercili Cura et al., 2010) or HRP in the presence of H₂O₂ (Wen, Liu, & Zhao, 2012) resulted in much stronger gels due to formation of covalent bonds in the cross-linked proteins. The yoghurts produced from the milk treated by three oxidases (lactoperoxidase, laccase and glucose oxidase) also showed lower whey drainage than the control yoghurt (Hiller & Lorenzen, 2011). These mentioned studies gave similar conclusion as the present study, indicating that treatment of the milk by the ternary system prior to yoghurt fermentation could modify texture and water holding capacity of the set-style yoghurt.

Microstructural characteristics of yoghurt samples

The observed microstructural characteristics of the treated and control yoghurt samples are depicted in Figure 3. The control yogurt had larger clusters of protein aggregates and a few contact points between the clusters (Figure 3a). However, the treated yoghurt had smaller aggregates linked together and increased connectivity between the aggregates (Figure 3b). This fact suggests the presence of more efficient inter-particle interactions in the protein network of the treated yoghurt and gives a direct support to those mentioned properties such as higher apparent viscosity, elastic and viscous moduli, gel hardness and lower syneresis.

Figure 3. Microstructural characteristics of the control (a) and treated (b) yoghurt observed under scanning electron microscope at a magnification of 4000. The labeled scale bar was 10.0 μm.

Figura 3. Características microestructurales del yogurt de control (a) y del yogurt tratado (b), observadas bajo un microscopio electrónico de barrido a una ampliación de 4000. La barra de escala etiquetada fue de 10,0 μm.

Cross-linking causes surface changes on the casein micelles and alters the interactions between the micelles (Ercili Cura et al., 2009; Schorsch, Carrie, Clark, & Norton, 2000). Thus, treatment of the whole milk by the ternary system in the present study would result in the treated yoghurt different microstructural characteristics than the control yoghurt. Two study results showed that Trichoderma reesei tyrosinase (Ercili Cura et al., 2009) and HRP in presence of H₂O₂ (Wen et al., 2012) treatment of the milk had impact on gel or yoghurt microstructure, as a finer network with thinner particle strands and smaller pores were formed. The present study shares consistence to the two mentioned studies.

The present results revealed that the ternary system was capable of inducing limited protein cross-linking in the milk and quality improvement in the set-style yoghurt prepared thereof and thus might be served as a potential approach as the widely applied TGase to favor quality attributes of some dairy products including set-style yoghurt. A detailed investigation is needed to reveal that which protein fraction is easily involved in the cross-linking. The modification in molecular structure, and more importantly, the impact on the behavior of casein micelles also need to be characterized in future study.

Conclusion

A ternary system comprising horseradish peroxidase, glucose oxidase and glucose could induce limited protein cross-linking in whole bovine milk and retard bacterial growth during the treatment. The carried out treatment had no impact on main composition but conferred enhanced apparent viscosity, elastic and viscous moduli, hardness and adhesiveness, greater thixotropic loops area and lower syneresis on the set-style yoghurt prepared thereof. This ternary system is thus potential as another bio-approach as the widely used transglutaminase to modify the rheological and textural properties of set-style yoghurt and other dairy products.

Acknowledgements

This study was funded by the Innovative Research Team of Higher Education of Heilongjiang Province (Project No. 2010td11) and the National High Technology Research & Development Programme (“863” Programme) of China (Project No. 2013AA102205). The authors thank the anonymous reviewers and editors for their valuable advice.