Effect of milk pH at heating on protein complex formation and ultimate gel properties of free-fat yoghurt

Abstract

The effect of milk pH before heating on casein–whey protein interactions and ultimate gel properties of the free-fat yoghurt was investigated. Reconstituted skim milk at different pH values (6.4, 6.8 and 7.2) was heated at 80 °C for 30 min. The type of protein and size of casein micelle in milk were determined. The storage modulus (G′), loss tangent (tan δ), flow behaviour as well as microstructure, firmness and water holding capacity of the yoghurt samples were measured. Heating milk at pH 7.2 formed mostly soluble protein complexes whereas at pH 6.4 micelle bound complexes was dominant. However, heating milk at pH 6.8 resulted in a relatively compact protein network due to a balanced contribution from both soluble protein/κ-casein complexes and whey protein-casein micelle associated complexes. Yoghurt prepared with milk heated at pH 6.8 showed significantly higher G′ values, shorter gelation times, higher water holding capacity, firmness and more compact protein network compared to those at pH 6.4, 7.2 and unheated milk. The obtained results demonstrated that milk pH adjustment before heating could be an important factor governing uniform quality yoghurt production.

Introduction

Heat stability of milk protein is greatly affected by milk pH. Alteration of milk pH prior to heat treatment affects the degree of association between denatured whey proteins and casein micelles (Ozcan et al. 2015). It was found that 82.2% of denatured whey proteins were associated with the casein micelle at pH 6.3, and only 30% and 0.0–5.0%, respectively were associated at pH 6.7 and 7.1 in rennet coagulated gel (Kethireddipalli et al. 2010). Consequently, it has also been demonstrated that the size of casein micelles was increased upon the association of denatured whey proteins with casein micelle (Anema and Li 2003). When milk is heated at pH 6.3, most of the denatured whey proteins tend to interact with κ-casein at the surface of the micelles and form micelle bound complexes (Anema and Li 2003; Anema et al. 2004). Whereas, milk heating at high pH (> 7.0), denatured whey proteins have a preference to interact with dissociating κ-caseins and form soluble protein complexes (Guyomarc’h et al. 2003). The heat-induced denaturation of whey proteins resulted in the protein unfolding, which may expose the reactive sulfhydryl groups of β-lactoglobulin to initiate the thiol-disulfide (SH-SS) interchange reaction with casein micelles and other unfolded whey proteins (Famelart et al. 2004; Sadeghi et al. 2014; Sava et al. 2005). In heated milk, there is the potential for whey proteins to form irreversible protein–protein interactions, present as either micelle-bound complexes that have occurred by forming disulphide bond with κ-casein at the surface of the micelles or soluble protein complex at the serum of milk (Anema and Li 2003; Guyomarc’h et al. 2003; Patel et al. 2007). The degree of such interaction determine textural and rheological properties of yoghurt.

Yoghurt texture is a crucial attribute for the mouth feeling, which is an important parameter for the acceptance of the product by the consumer (Abou-Soliman et al. 2017; Mudgil et al. 2017). The physical attributes of set yoghurt, including the whey separation and perceived firmness, are crucial aspects of the quality and consumer acceptance. Many factors influence the textural properties of yoghurt including heating milk prior to manufacture. Heating milk leads to increase the water holding capacity as well as firmness (Lucey et al. 1997). The effect of heating milk on the textural properties of yoghurt is correlated with the denaturation rate of whey proteins and formation of the micelle bound and/or the soluble protein complexes (Anema et al. 2004; Lucey et al. 1998; Vasbinder et al. 2003). The formation of different proportions of micelle bound and soluble protein complexes exhibit distinct properties in the yoghurt gel (Anema et al. 2004; Lucey et al. 1998). Most studies have been focused on the improving the water holding capacity (WHC) and firmness of yoghurt by the application of heat treatment of milk (Chever et al. 2014; Damin et al. 2009; Delikanli and Ozcan 2014). It must be taken into account that the milk pH value can change slightly all over the year which in turn might produce different proportion of protein complexes after heat treatment. As a result, inconsistent yoghurt texture can be found regarding the reasoning behind the formation of various protein complexes.

Therefore, the main objective of this research was to investigate the effect of different pH values of milk prior to heating on the casein–whey protein interactions as well as their effects on resulting set yoghurt. Also, the firmness, WHC, rheological properties, and gel microstructure of free-fat yoghurt were investigated.

Materials and methods

Materials

Skim milk powder (protein 35.19%, lactose 52.34%, fat 0.92%, ash 6.61% and moisture 4.94%) (Difco™ skim milk, Wako, Osaka, Japan) was used for yoghurt preparation. A commercial freeze-dried yoghurt culture consisting of Streptococcus thermophilus and Lactobacillus delbrueckii subsp. bulgaricus (FD-DVS YFL-904) was obtained from Chr. Hansen (Horsholm, Denmark).

Adjustment of pH and heat treatment of milk samples

Reconstituted skim milk (RSM) of 10.7% total solids (wt/wt) was prepared by adding 12.67 g of SM in 100 g of deionized water. The milk samples were stirred for at least 3 h at room temperature and were kept overnight at 4 °C to obtain complete hydration. The pH values of the milk samples were adjusted to 7.2, 6.8, and 6.4 with the addition of 1 M NaOH and 1 M HCl thereafter milk samples were heated at 80 °C for 30 min in a thermostatically controlled water bath. After heat treatment, the milk samples were rapidly cooled in an ice bath. The pH values of heat treated samples were readjusted back to pH 6.8 before yoghurt manufacture to ensure a similar acidification rate during the fermentation process.

Centrifugation of milk samples

To obtain serum protein (supernatant) from colloidal protein(pellet) from each of heated and unheated milk, 1 mL milk sample was transferred into 1.5 mL Eppendorf (Eppendorf, Japan) tubes in duplicate and centrifuged at 20,100 × g at 25 °C for 60 min using a bench centrifuge (Hitachi Koki Centrifuge, Tokyo, Japan). After centrifugation, the clear supernatants were carefully separated from the precipitants. Thereafter the supernatants and precipitants were dispersed in Laemmli sample buffer at 1:20 and 1:40 (v/v), respectively to prepare stock samples. The protein compositions of the supernatants and precipitants were determined by electrophoresis.

Polyacrylamide gel electrophoresis for protein profile

Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was performed using a Bio-Rad mini-gel slab electrophoresis system (Bio-Rad Laboratories, Richmond, CA, USA) as described previously (Nguyen et al. 2013). The supernatants and precipitants were run under both non-reducing and reducing conditions. Under the reducing condition, 5 µL β-mercaptoethanol (0.5%, v/v) was added to 1 mL of samples, which was then heated at 95 °C for 10 min while in case of non-reducing conditions, the samples were analyzed without further addition of β-mercaptoethanol. The gels were prepared, run, stained and de-staining as described previously (Chevalier et al. 2009). The intensities of the protein bands were integrated using image J software (Rasband, W.S., Image J., U. S. National Institutes of Health, Bethesda, Maryland, USA).

Particle size measurement

The casein micelle size of the heated and unheated milk samples was measured by dynamic light scattering (DLS) at 20 °C using a Malvern Zetasizer Nano-ZS (Malvern Instruments Ltd., Malvern, UK). Milk samples were diluted fivefold with calcium-imidazole buffer (5 mM CaCl₂, 20 mM imidazole, 30 mM NaCl, pH 7.0) and allowed to stand for 10 min prior to measurements as previously reported (Taterka and Castillo 2015).

Yoghurt preparation

RSM were divided into 4 equal portions, the first one: unheated milk of pH 6.7 (not adjusted) as control; heated milk at pH 6.4, 6.8 and 7.2 respectively. For heated milk pH was readjusted again at 6.8 before adding starter culture. The starter culture was prepared as described previously (Lucey et al. 1998). Milk samples were inoculated with a 2% (wt/wt) mother culture in a plasc cup and incubated at 41 °C until the pH of samples reached 4.6 for preparing the set yoghurt. After completing fermentation, the yoghurts were cooled in ice water and stored at 4 °C until further analysis.

Water-holding capacity

The water-holding capacity (WHC) was determined using the centrifuge method. A 25 g yoghurt (Y) was centrifuge at 1250 × g for 25 min as described previously (Mahomud et al. 2017). The expelled whey (WE) was carefully removed, and the pellet was weighed. The WHC was calculated (Akalın et al. 2012) as

$$\text{WHC}\left(\%\right)=\frac{\text{Y}-\text{WE}}{\text{Y}}\times 100$$

Firmness of the yoghurt

The firmness of yoghurt samples was measured with a texture analyzer (Sun Rheometer CR-200D, Sun Scientific Co., Tokyo, Japan) as reported earlier (Sah et al. 2016). A 30-mm cylindrical probe with a constant velocity of 100 mm/min was set in the cylindrical containers to sample depth of 20 mm holding the test sample. A single-compression cycle test with 2-N load cell was used. The firmness was determined as the maximum force (N) require to penetrate the sample to 20 mm depth.

Rheological properties of yoghurt

Rheological measurements were performed using a controlled stress rheometer (AR2000, G2KG, TA Instruments, Newcastle, DE, USA). The measuring geometry consisted of two coaxial cylinders (diameters 37.03 and 35.01 mm). On addition of starter culture to the milk, the mixture was mixed properly and then 13 mL of the inoculated milk was poured into the rheometer. After loading, the sample was oscillated at a frequency of 1.0 Hz with an applied strain of 0.1%. Rheological properties were studied using a time sweep test with the measurement of storage modulus (G′) and loss tangent (δ). Data were taken every 5 min for 420 min at 41 °C. The storage modulus (G′) describes the deformation energy stored in the yoghurt during the shearing and it shows the elastic behavior of the yoghurt samples. Lower G′ indicate weeker gel. Where as loss tangent (δ) is the ratio of viscous to elastic properties (G″/G′), indicates the relaxation behavior of bonds (Lucey et al. 1997). Frequency sweep and shear rate sweep tests were also studied using a cone and plate geometry (40 mm diameter, 2° angle, and 48 µm gap) that has been described previously (Sah et al. 2016).

Confocal laser scanning microscopy

The yoghurt microstructure was studied using a confocal laser scanning microscope (LSM710, Carl Zeiss microscopy GmbH, Jena, Germany) as previously described (Ong et al. 2011; Mahomud et al. 2017). Acridine orange (0.5% wt/wt, Sigma Chemical Co., St Louis, MO, USA) was used to stain the proteins. 300 μL of acridine orange solution was mixed with 100 g of selected milk samples and 2% (wt/wt) starter culture was added. The milk samples were then placed on to the concave region of a glass petri dish (35 mm diameter) and covered with a cover slip (Iwaki Asahi Glass Co., Chiba, Japan), ensuring on air was trapped. The Petri dish was then incubated at 41 °C until the pH reached 4.6. CLSM observation was performed in a dark room and the excitation wavelength was set at 488 nm. The images were captured with a 63 × oil immersion objective lens (numerical aperture = 1.4) and the emitted signal passed through a 515–530 nm filter. At least five CLSM images were recorded for each yoghurt treatment. A typical image is presented in each of the figures. This procedure was slightly modified from Ozcan-Yilsay et al. (2007).

Statistical analysis

Statistical analysis was performed using Kaleida Graph (Version 4.1.1, Synergy Software, Reading, PA, USA). All measurements were carried out in three replications. Mean values were compared using the Tukey’s honest significant difference test at p < 0.05 significance level.

Results and discussion

Distribution pattern of proteins in the serum and micellar phases of skim milk

The supernatants and precipitants obtained after centrifugation of each of the pH adjusted-heated and unheated SM dispersions were analyzed using SDS-PAGE. The PAGE patterns of reduced and non-reduced samples are presented in Fig. 1. Reducing and non-reducing SDS-PAGE was carried out to investigate the formation of covalent aggregation between denatured whey proteins and κ-casein. For non-reducing SDS-PAGE analysis (Fig. 1a, b), the band pattern for both β-lactoglobulin and κ-casein of heated and unheated samples showed similar and low intensity (Fig. 1a, b, lane 5, 6, 7, 8). In reducing SDS-PAGE analysis, as shown in Fig. 1a, b, the band pattern of β-lactoglobulin and κ-casein of heated pH-adjusted samples were higher than those of unheated samples indicating formation of heat-induced aggregation of β-lactoglobulin with κ-casein in either the serum or micellar phase. The band of β-lactoglobulin and κ-casein disappeared in the supernatant of heated milk at pH 6.4 (Fig. 1a, lane 2) whereas they appeared in the micellar phase of that sample (Fig. 1b, lane 2) which could be explained by the formation of protein complexes and it was treated as micelle bound complexes (Donato and Guyomarc’h 2009). Similarly, the band of β-lactoglobulin and κ-casein appeared in the supernatant of heated milk at pH 7.2 (Fig. 1a, lane 4) whereas they disappeared in the micellar phase of that sample (Fig. 1b, lane 4) which could be explained by the formation of protein complexes and it was treated as soluble protein complexes (Guyomarc’h et al. 2003). On the other hand, the band of β-lactoglobulin and κ-casein appeared in both supernatant and micellar phase of heated milk at pH 6.8 (Fig. 1a, b, lane 3) which could be explained by the formation of both micelle bound and soluble protein complexes (Anema and Li 2003; Donato and Guyomarc’h 2009).

Fig. 1.

SDS-PAGE electrophoretograms of supernatants (a) and precipitants (b) obtained from reconstituted skim milk. Unheated milk (lane 1 = reducing, lane 5 = non-reducing), and heated milk at pH 6.4 (lane 2 = reducing, lane 6 = non-reducing), pH 6.8 (lane 3 = reducing, lane 7 = non-reducing), and pH 7.2 (lane 4 = reducing, lane 8 = non-reducing), respectively. (i) bovine serum albumin, (ii) α-casein, (iii) β-casein, (iv) κ-casein, (v) β-lactoglobulin, (vi) α-lactalbumin

Table 1 showed the β-lactoglobulin and κ-casein fractions in serum and micellar phase of the milk samples. In serum phase, presence of those protein fractions were significantly higher (p < 0.05) in milk sample heated at pH ≥ natural pH of milk and the lowest was for the sample heated at pH 6.4, while in unheated milk both fractions showed intermediate results. On the other hand, in micellar phase, the fraction of κ-casein was significantly (p > 0.05) higher and the fraction of β-lactoglobulin was significantly lower (p > 0.05) in unheated milk than heated milk compare to heated milk samples. However, Among the heated milk samples, both protein fractions showed the following trend: sample heated at pH 6.4 > pH 6.8 > pH 7.2.

Table 1.

Protein fractions of serum and micellar phases measured from SDS-PAGE and Z-average particle diameter of casein micelles obtained from different reconstituted skim milk samples

Milk sample	Protein fraction in serum (%)		Protein fraction in micelle (%)		Particle size (nm)
	κ-casein	β-lactoglobulin	κ-casein	β-lactoglobulin
Unheated	8.63 ± 0.28a	8.17 ± 0.95a	17.08 ± 0.92a	6.67 ± 0.29a	186.75 ± 3.68a
pH 6.4	0.43 ± 0.13b	4.16 ± 0.30b	10.25 ± 0.42b	15.76 ± 0.23b	308.33 ± 2.20b
pH 6.8	24.43 ± 0.64c	33.71 ± 0.38c	12.18 ± 1.01c	12.41 ± 0.49c	194.81 ± 3.46a
pH 7.2	42.42 ± 0.77d	49.59 ± 0.81d	8.65 ± 0.38d	7.36 ± 0.30a	166.71 ± 2.49c

Results are expressed as means ± standard deviation from three trials

Different lowercase superscripts in the same column are significantly different (p < 0.05)

The PAGE patterns suggested that the distribution of β-lactoglobulin and κ-casein in heated milk at pH 7.2 and 6.8 showed more covalent aggregates in the supernatant compared to that of pH 6.4 (Table 1, protein fraction in serum). This was due to the dissociation of κ-casein from the casein micelles to the serum phase when milk was heated at higher pH than natural milk pH (Anema 2008). On the other hand, the protein band of β-lactoglobulin and κ-casein obtained from heated milk at pH 6.4, and pH 6.8 (Table 1, protein fraction in micelle) showed more covalent aggregates in the precipitant compared to those of pH 7.2. This was due to the preferential heat-induced aggregation of β-lactoglobulin and κ-casein on the surface of the casein micelles (Anema and Li 2003; Donato and Guyomarc’h 2009). The formation of soluble protein and micelle bound complexes was highly dependent on the milk pH before heating, which influences the formation of the covalent protein complexes between β-lactoglobulin and κ-casein in the serum and micellar phase (Anema and Li 2003; Vasbinder et al. 2003).

Casein micelles size distribution in skim milk

Dynamic light scattering technique was used to investigate the size of the casein micelles in milk heated at different pH values. Table 1 shows the Z-average particle diameter of casein micelles of each pH adjusted-heated and unheated skim milk samples. The casein micelle diameter of milk sample heated at pH 6.4 was significantly (p > 0.05) higher than all other samples and the lowest was for the unheated milk sample. According to particle diameter, milk samples follow the following trend: pH 6.4 > pH 6.8 > unheated > pH 7.2. The highest particle size (308.33 nm) was observed for the milk sample heated at pH 6.4 while the lowest (166.71 nm) was for the sample at pH 7.2 (Table 1) which was due to the fact that denatured whey proteins interacted with casein micelles at pH 6.4. Whereas, at higher pH (7.2), κ-caseins dissociated from the casein micelles (Anema 2008). Subsequently, their size (166.71 nm) was reduced to below than that of the unheated sample (186.75 nm), which was supported by our SDS-PAGE results (Fig. 1). When milk is heated at pH < the natural pH (6.7–6.8), denatured whey proteins interacts with κ-caseins on the surface of the casein micelles resulting in larger micelle size (Vasbinder et al. 2003).

Water holding capacity of free-fat yoghurt

The water holding capacity of free-fat yoghurt prepared from heated SM at different pH values and unheated SM is shown in Table 2. The highest WHC (49.9%) was observed for yoghurt sample prepared by heating milk at pH 6.8 whereas the lowest (32.12%) was for the control sample (unheated milk). Yoghurt samples prepared by heating milk at pH 6.4 and 7.2, the WHC values were about 40.18% and 45.34% respectively. The WHC of yoghurt gel prepared from heating milk at different pH and unheated milk followed the following order: pH 6.8 > pH 7.2 > pH 6.4 > unheated milk.

Table 2.

Water holding capacity and firmness of yogurt made from different reconstituted skim milk samples

Sample	Water holding capacity (%)	Firmness (N)
Unheated	32.12 ± 0.98a	0.67 ± 0.05a
pH 6.4	40.18 ± 1.51b	0.92 ± 0.04b
pH 6.8	49.92 ± 1.37d	1.14 ± 0.05c
pH 7.2	45.34 ± 0.81c	0.98 ± 0.06c

Results are expressed as means ± standard deviation from three trials

Different lowercase superscripts in the same column are significantly different (p < 0.05)

The WHC of yoghurts prepared by heating SM at pH 7.2 was significantly (p < 0.05) higher than that of heated at pH 6.4 (Table 2). The result pointed out the importance of soluble protein complex over the micelle bound complexes to increase the water holding capacity of yoghurt. However, the highest WHC of yoghurt sample obtained by heating milk at pH 6.8 indicated the inevitability of the combined effect of both micelle-bound and soluble protein complexes to increase the WHC of yoghurt (Xu et al. 2015). Heating milk influences the denaturation of whey proteins and their interaction with micellar κ-casein, which results in forming compact protein network, immobilization of a large amount of free water and ultimate positive impact on the WHC of yoghurt (Akalın et al. 2012; Jørgensen et al. 2015).

Rheological properties of free-fat yoghurt

Firmness is considered as one of the most important attributes of yoghurt quality. The firmness of free-fat yoghurt made from unheated and heated SM at different pH values are shown in Table 2. In general, the firmness of yoghurt sample prepared form heated milk was significantly (p < 0.05) higher than the sample prepared from unheated milk and the highest value (1.14 N) was observed for the yoghurt prepared from heated milk at pH 6.8. On the other hand, firmness of yoghurt sample made from heated milk at pH 7.2 was significantly (p < 0.05) higher than that of at pH 6.4 (Table 2). The increased firmness of yoghurt samples prepared from heated SM was due to the interaction of denatured whey proteins with κ-caseins in the casein micelle, forming micelle bound complexes as well as the interactions of denatured whey protein with κ-caseins in the serum phase, forming soluble protein complexes (Guyomarc’h et al. 2003; Vukić et al. 2018; Xu et al. 2015). Heating milk at different pH influences the type such interaction and ultimate yoghurt gel firmness. It can be assumed that milk heated at pH 6.8 produced substantial proportion of whey proteins associated-casein micelle and soluble whey protein-κ casein complexes (details in “Distribution pattern of proteins in the serum and micellar phases of skim milk” section), which enhance the firmness of the yoghurt gel (Akalın et al. 2012; Lakemond and van Vliet 2008). Therefore, a balance of both bound and soluble complexes is necessary to produce firmer and stiffer yoghurt gel firmness (Ozcan et al. 2015).

Oscillatory rheology is extensively used to study the viscoelastic behavior of food systems and monitor the formation and strengthening process of gels. Rheological parameters describing casein gels depend generally on the number and strength of bonds among casein particles, structure of particles and the spatial distribution of the strands forming these particles (Lucey et al. 1997). In this study storage modulus and loss tangent were used to describe youghurt gel properties. When the G′ is lower the ration is < 1 the youghurt gel is more elastic and vicevers when G′ lower then the ration > 1 means the gel is more viscous. The rheological properties of yoghurt were monitored every 5 min during acidification (up to 420 min). The storage modulus (G′) and tan delta (δ) of different yoghurt samples are shown in Fig. 2. Storage modulus (G′) was used to measure the deformation energy stored in the yoghurt during the shearing and it showed the elastic behaviour of the yoghurt samples. The G′ of yoghurt made from heated SM at pH 6.8 was higher than other samples (Fig. 2a). The G′ values at pH 4.6 for gels made (Vasbinder et al. 2003) from unheated and heated milk at different pH values increased in the following order: pH 6.8 > 7.2 > 6.4 > unheated milk. On the other hand, the storage modulus of yoghurt made from heated milk at pH 7.2 was higher than that made at pH 6.4 (Fig. 2a), indicated that soluble protein complexes were predominant attribute in increasing the storage modulus of yoghurt than micelle bound complexes.

Fig. 2.

Storage modulus (G′; a) and tan delta (δ; b) of yoghurts made from unheated milk (●), and heated milk at pH 6.4 (▲), pH 6.8 (■), and pH 7.2(◆), respectively

The reasons of increased G′ of yoghurt made from heated milk is due to the active participation of denatured whey proteins in gel structure by forming protein aggregate between the denatured β-lactoglobulin and κ-casein (Nguyen et al. 2013). These protein aggregates accelerate the formation of the number and strength of bonds between protein complexes resulting in higher G′ of yoghurt (Velez-Ruiz et al. 2013). The formation of a significant proportion of micelle bound and soluble protein complexes in the heated SM at pH 6.8 are responsible for increasing the yoghurt G′ (Lakemond and van Vliet 2008). The lower G′ (weaker gel) was observed in the yoghurt made from heated SM at either higher or lower pH values than the natural pH (~ 6.7–6.8) of milk, which would have resulted from the predominant contribution of soluble protein complexes and bound aggregates, respectively (Ozcan et al. 2015). Milk heated at pH 6.8 results in a large number of protein complexes between soluble denatured whey proteins, between the denatured whey proteins and the casein micelles, and between denatured whey proteins and soluble κ-caseins. As consequence, these complexes results in enhance number and strength of contact points and ultimate gel firmness (Anema 2008; Donato and Guyomarc’h 2009; Velez-Ruiz et al. 2013).

Yoghurts produced from heated milk at pH 6.8 have lower tan δ value (Fig. 2b), which means that they are less susceptible to syneresis, which was in line with the result of water holding capacity. The yoghurt made from unheated milk shows a later maximum in loss tangent (Fig. 2b). This maximum value of tan δ was caused by a loosening of the intramolecular forces in casein particles caused by solubilization of colloidal calcium phosphate (Lucey et al. 1998).

In the frequency sweep and shear rate sweep tests, yoghurt prepared from heated milk at pH 6.8 showed higher the storage modulus and higher the shear stress compared to those prepared from unheated milk or those heated at pH 6.4 and 7.2 (Fig. 3a, b). All samples showed thixotropic behaviour as the breakdown and rebuilding of protein structure by shearing did not follow the same path. Large hysteresis loop observed for the yoghurt made from heated milk than unheated milk. Among the heated milk sample, yoghurt prepared by heating milk at pH 6.8 demonstrated lager hysteresis loop than those at pH 6.4 and 7.2. Larger the area of hysteresis loop, the more energy consumed in the structural breakdown means more solid like nature for the set yoghurts (Donato and Guyomarc’h 2009). Moreover, yoghurt made from heated milk at pH 6.8 exhibited higher apparent viscosity value compared to others samples (Fig. 3c), indicating a high viscous nature and a more solid-like behaviour during overnight storage at 4 °C. Whereas, yoghurt made from unheated milk exhibit significantly lower apparent viscosities than those made from heated milk (Fig. 3c), and hence a more liquid-like behaviour. The apparent viscosities of yoghurt decreased with increasing shear rate indicating non-Newtonian such as pseudo-plastic fluids (Santos et al. 2019).

Fig. 3.

Storage modulus (G′) as a function of frequency sweep (a), flow curves (b; shear stress vs. shear rate) and apparent viscosity as a function of shear rate (c) of yoghurts measured after one day of storage at 4 °C. Yoghurt milk bases were made from unheated milk (●), and heated milk at pH 6.4 (▲), pH 6.8 (■), and pH 7.2(◆), respectively

In general, yoghurt made from heated milk demonstrated lower gelation time than that made from unheated milk. The shortest gelation time (time point when G′ reached > 1 Pa) was observed for the sample prepared from heated milk at pH 6.8. During acidification, the heated milk starts to form a gel at a higher pH, typically pH 5.3–4.9, whereas in unheated milk it occurs at a pH < 4.6 (Lucey et al. 1998). The presence of denatured whey proteins, particularly β-lactoglobulin are thought to be responsible for lowering the gelation time (Vasbinder et al. 2003).

Microstructure of yoghurt

The microstructure of the yoghurt samples was observed to visualize the three-dimensional protein networks form inside gel. Before incubation, milk with 2% starter culture was stained with acridine orange solution to observe the protein network and pore sizes. Figure 4 shows the microstructure of the free-fat yoghurt samples. Remarkable differences were observed between the microstructure of yoghurt prepared from heated and unheated milk. In general, the microstructure of the yoghurt made from unheated milk showed a coarse network with little interconnectivity with larger pore spaces (Fig. 4a) than those prepared from heated milk at different pH (Fig. 4b–d). Among the heated milk, yoghurt prepared at pH 6.8 (Fig. 4c) was less porous, strongly branched cross-linked network structure with lot of connections than those prepared at pH 6.4 and 7.2 (Fig. 4b–d). The more compact network of yoghurt was the consequence of the interaction of denatured whey proteins with κ-caseins in the casein micelles, and from the interactions of denatured whey protein with κ-caseins in the serum phase (Xu et al. 2015).

Fig. 4.

Confocal laser scanning microscopy images of yoghurt gels made from unheated milk (a), heated milk at pH 6.4 (b), pH 6.8 (c), and pH 7.2 (d). The protein matrix is white and the pores are dark. Scale bar represents 10 µm

Commonly, finer and more closed networks in yoghurt show less syneresis than coarse, more open structures (Puvanenthiran et al. 2002). The results of the analysis of the WHC showed that the yoghurt made from heated milk at pH 6.8 had significantly less syneresis than those made from unheated milk or heated milk at pH 6.4 and 7.2 respectively, which can be explained using CLSM images. Other studies (Guyomarc’h et al. 2003; Ozcan et al. 2015) reported that there was no major microstructural difference observed in the yoghurt gels containing a high level of either micelle bound or soluble protein complexes, or those had both. In contrast, our result demonstrated that major microstructural differences were observed as a result of the different level of either micelle bound or soluble protein complexes. The possible reason for some of the conflicting results found in various studies related to the method of the acidification and different starting materials (Lucey et al. 1998; Mahomud et al. 2017).

Conclusion

The pH value of milk before heating affect the physical, rheological and microstructural properties of the free-fat yoghurt. Heating milk at pH 6.4 promoted the formation of micelle bound complexes as demonstrated by significantly lower WHC, firmness and storage modulus. Whereas milk heated at pH 7.2 enhanced to the formation of soluble protein complexes as illustrated by significantly higher WHC, firmness and storage modulus. However, milk heated at pH 6.8 (~ natural pH) produced an optimal proportion of both micelle bound and soluble protein complexes which resulted in yoghurt with high gel stiffness, high WHC, and finer microstructure. Under this condition, a complex gel network was formed containing numerous aggregating particles as well as number of contact points. The results of this study suggested that balance of both bound and soluble complexes contribute to rigidity and stiffness of free-fat set yoghurt. These finding demand further research to observe the sensory evaluation of the free-fat yoghurt for industrial application.