*2.5. Analyses*

All analyses were performed as previously described [1,2,7]. In brief, gross composition, yoghurt bacteria counts, water holding capacity (WHC), firmness/cohesiveness and ACE-IA determinations were carried out by means of the Milkoscan, colony count technique, centrifugation, texture profile analysis and RP-HPLC, respectively.

#### *2.6. Statistical Analysis*

The analysis of variance (ANOVA) was used to analyze the effect of the mixture composition (milk base) and of the treatment on the characteristics of the yoghurt-type gels and for the differences among the means of the Least Significant Difference test was used (LSD, *p* < 0.05). Statistical analysis was carried out by means of the Statgraphics, Centurion V (Manugistics Inc., Rockville, MA 20852, USA).

### **3. Results and Discussion**

Sodini et al. [3] summarized that the rheological properties and microstructure of yoghurt are related to the heating conditions of the milk base from 75 ◦C for 1 min–5 min to 95 ◦C for 5 min–10 min; the later conditions result in >99% denaturation of the β-lactoglobulin (β-lg). The heat treatment induces the formation of various types of complexes, which are key factors for the configuration of acid milk gels and yoghurt: i. between casein micelles and denatured whey proteins, ii. between κ-casein and β-lg and iii. between denatured whey proteins [25]. Their extent and distribution are differentiated within the range from pH 6.5 to pH 6.7. At low pH, the denatured whey proteins bind onto the casein micelles thus increasing the particle size while at higher pH they participate in soluble complexes with solubilized κ-casein [26].

From Table 2, it is evident that only the control skim sheep milk base was stable under the usual heating conditions for the yoghurt manufacture –i.e., 95 ◦C/5 min. The addition of WPC or WPHs even at low level decreased the heat stability and trisodium citrate addition was necessary. The heat stability of sheep milk is lower than that of cow milk because ovine micelles are more mineralized, contain more β-casein, are less hydrated than their bovine counterparts and their size increase substantially during heating. In fact, heating at >80 ◦C causes an increase of the ovine micelle size by >50%, which along with the high casein content favour micelle-micelle interaction and aggregation [13,14]. The later phenomenon is expected to be more pronounced in the skim sheep milk of the present experiments due to the increase of protein concentration caused by the removal of fat. Moreover, the calcium content of WPC and WPHs may also adversely affect the heat stability of the mixtures of Table 2. Excessive calcium before heat treatment results in the formation of large aggregates that decrease the heat-stability while calcium combined with acidification induce the gelation of denatured whey protein polymers [27].

Therefore, both the low pH and the elevated calcium content of the milk bases before acidification could be responsible for the poor heat stability of the sheep milk bases in the experiments of Section 2.2. Milder heating conditions and the addition of trisodium citrate before heat treatment were used for the enriched skim milk bases. As shown in Table 2, the pH of milk bases supplemented with approximately 7 mM sodium citrate was higher compared to the control before and after heat treatment, on average by 0.12 pH and 0.16 pH units, respectively. The addition of sodium citrate in sheep or goats milk enhances their heat stability by linking with ionic calcium and solubilizing both the colloidal calcium phosphate and the calcium linked to the phosphoseryl residues. Moreover, sodium citrate increases the milk pH by approximately 0.1 units resulting in the increase of small-sized casein micelles and of their negative charge that do not favour heat aggregation [13,14]. In regard to cow milk yoghurt gels, using >25 mM trisodium citrate causes detrimental disruption of casein micelles, while the addition of 10 mM–20 mM trisodium citrate improves the texture by chelating calcium. Calcium chelation induces solublization of colloidal calcium phosphate, thus enhancing the formation of crosslinks [28]. The milk bases were stable under the conditions of HP treatment, which reduced the pH by 0.18 units, on average.

#### *3.1. Properties of Yoghurt-type Gels*

The effect of the experimental factors on the physical and compositional properties of yoghurt-type gels was analyzed by the multifactor ANOVA and it is shown in Table 3.


**Table 3.** Effect of experimental factors on the properties of yoghurt-type gels, expressed as *p*-values.

It is evident (Table 3) that the physical properties (water holding capacity, firmness and cohesiveness) were affected significantly (*p* < 0.05) by the type of enrichment (composition) of the milk base and by the processing (heat- or high pressure-treatment). Moreover, there was a significant (*p* < 0.05) combined effect of these factors (A × B) on the physical properties. The days of storage affected significantly (*p* < 0.05) the pH and the firmness of gels.

The physical and compositional properties of the yoghurt-type gels treated by heat or by high-pressure are presented in Table 4. The factor "composition"- that is the composition of the milk base mixture—was related to statistically significant differences (*p* < 0.05) mostly in the HT group. These differences can be assigned to: i. different heat treatments, and ii. differences in composition. It is evident that at day three, the YWPHt1 base treated at 75 ◦C for 5 min, i.e., under the mildest conditions, was significantly differentiated among the HT group in regard to the water holding capacity (WHC), firmness and cohesiveness. The extend of whey protein denaturation in YWPHt1 was expected to be the lowest, which is consistent with its lowest WHC. Extended heat denaturation of β-lg enhances the capability of the casein network to immobilize the serum [3]. On the other hand, in high-protein yoghurts the reduction of heating temperature from 95 ◦C to 75 ◦C for 5 min reduces firmness but improves the sensory properties [29]. Despite the highest total solids and protein content of YWPHt1, its firmness was the lowest in accordance to the above-mentioned effect of heating on the formation and distribution of complexes. It has to be noticed that the enrichment of skim milk - i.e., higher total solids and protein contents - and heat treatment at 85 ◦C or 90 ◦C for 5 min resulted in significantly higher (*p* < 0.05) WHC and firmness for YWP80, YWOHp0.5 and WPHt0.5 compared to the control Y0. Therefore, the supplementation with WP80 at 1% and with WPHs at 0.5% counteracted the effect of milder heating conditions. The addition of whey proteins in the yoghurt milk base increases also the ratio whey protein to casein that in turn increases WHC and affect also the viscoelasticity and flow behavior; the latter is related to heating conditions [3]. In particular, the whey protein to the casein ratio has been demonstrated as a crucial factor for the structure of non-fat stirred yoghurts [30]. The enrichment with WPHs resulted in lower WHC and firmness compared to the intact WP80. Apparently, due to hydrolysis, the WPHt and WPHp contained less intact native whey proteins, which are key components for the crosslinking within the yoghurt gel matrix, as reported above. Moreover, the solubility of a WPH may be reduced due to the exposure of hydrophobic areas of the molecules [9].


**Table 4.** Properties of yoghurt-type gels made from skim sheep milk bases enriched with WPHs.

Means of two experiments ± standard deviation. Symbols as in Table 2; <sup>1</sup> heat treatment conditions indicated in Table 2; <sup>2</sup> high hydrostatic pressure treatment at 600 MPa/55 ◦C/10 min. a–d lowercase letters indicate statistically significant differences (LSD, *p* < 0.05) within each type of treatment, i.e., within rows; A–B, indicate significant differences (*p* < 0.05) between heat treatments (HT) and high pressure (HP) treatments; \* indicates significant differences (*p* < 0.05) between three and 10 days.

Similar findings have been reported by other researchers. The use of the WPHt of the present study in reduced-fat cow milk base did not influence WHC but dramatically affected firmness, while the use of commercial WPHs of bovine origin had the opposite effect [2]. The addition of commercial WPHs in reduced-fat cow milk base at levels lower than 0.4% (*w*/*v*) affected negatively the texture of yoghurts due to less cross-linked microstructure [11]. The use of a tryptic WPH in buffalo milk at a ratio of 3% decreased the firmness and increased the syneresis of the sweetened yoghurt made there from [12].

The observation that the heating conditions were more important than the enrichment of the sheep milk base coincides with the findings for the HP-group of yoghurt-type gels. Less differences were observed between the HP milk bases (Table 2). These differences can be attributed solely to the enrichment since all were treated under the same conditions. Similar HP treatments of full-fat sheep milk induced >90% denaturation of β-lg and substantial reduction of α-la [22]. Firmness of the HP-group was significantly lower compared to the HT-group (Table 4) and this holds true for the control non-supplemented milk base Y0. WHC was not affected but firmness of YWP80 supplemented with 1% WP80 was significantly the lowest, half that of the control Y0. The opposite was true for cohesiveness. From the literature reviews [3,16,17] comes out that according to several studies, the HP treatment improve firmness and WHC. However, there are also opposite reports that coincide with our findings [15,29]. Similarly to the present study, the HP treatment of the cow milk base supplemented with whey protein concentrates and isolates resulted in weaker and less firm acid gels compared to heat treatment due to differences in the complexation of denatured β-lg [31,32]. The study of the effect of HP on reconstituted skim milk powder base combined with heating and in comparison to heat-treatment suggested that it can be used for the production of high protein drinking yoghurts of low viscosity [33]. Interestingly, the incorporation of WPHs in the HP treated milk bases had less detrimental effects on physical properties than the addition of non-hydrolyzed WP80. Again, a possible explanation is a favourable change of solubilization and the interactions between peptides and proteins, induced by partial hydrolysis [9].

The HP treatment of homogenized pasteurized full-fat sheep milk at 500 MPa/55 ◦C increase the firmness and the WHC of yoghurt compared to the typical 95 ◦C/5 min heat treatment [19]. The effect of HP on skim sheep milk has not been reported, but it is reasonable to expect that its high protein content could affect its behavior under various HP conditions. Additionally, the particularities of sheep milk and casein should be taken into consideration. Our previous study [23] showed that HP conditions similar to the present study affected the rennet clotting behavior of full-fat sheep milk in a different manner compared to cow milk. HP at 600 MPa decreased the size of ovine micelle by 40% [21]. At 400 MPa the κ-casein was extensively solubilized by >80%, much higher than the 22% observed in ovine milk [18]. Therefore, it could be assumed that soluble κ-casein/β-lg complexes were favoured under the applied HP conditions impairing thus crosslinking between micelles and β-lg and consequently the gel microstructure.
