The combined effect of probiotic cultures and incubation final pH on the quality of buffalo milk yogurt during cold storage

Abstract The combined effects of starter culture type (SCT) and incubation final pH (IFpH) on the physicochemical and organoleptic properties of buffalo milk yogurt containing 3 g·100 g−1 milk fat were investigated throughout 20 days of storage at 4°C. The postacidification kinetics fitted to zero‐order reaction for all buffalo milk yogurt samples. The reaction rate constants of the buffalo milk yogurt samples containing YC‐X11, ABY‐2, and ABT‐4 cultures were 0.010, 0.007, and 0.004 g·100 g−1·day−1, respectively. Regardless of the IFpH, the absence of Lactobacillus delbrueckii subsp. bulgaricus in the starter culture increased the syneresis. L*, a*, and b* values were not affected by the IFpH and the SCT. ABY‐2 culture increased the amount of organic acids during cold storage in comparison with the YC‐X11, while its effect on the proportions of saturated and unsaturated fatty acids was not significant. The results of sensory evaluation revealed that a more acceptable buffalo milk yogurt can be manufactured by using probiotic ABY‐2 culture.


| INTRODUCTION
Final composition and physicochemical and sensory aspects of a fermented milk product such as yogurt are influenced mainly by chemical composition of milk (Akgun, Yazici, & Gulec, 2016) and processing conditions (Nguyen, Ong, Kentish, & Gras, 2014). The effects of milk composition on the mentioned properties of buffalo milk yogurt are well described in our previous study (Akgun et al., 2016). Starter culture type (SCT) and incubation final pH (IFpH) are considered as other important factors which affect overall quality of buffalo milk yogurt during cold storage.
Streptococcus thermophilus and Lactobacillus delbrueckii subsp. bulgaricus are typical strains in yogurt manufacturing. Yogurt that contains probiotic bacteria such as Lactobacillus acidophilus and Bifidobacteria (Tamime & Robinson, 1999) is popular in dairy industry due to their health-promoting properties (Ravula & Shah, 1998). Also, a wide range of products with different flavors, textures, and consistencies were obtained by mixed strains of these microorganisms (Mckinley, 2005). Olson and Aryana (2008) stated that an excessively high inoculated level of L. acidophilus prolonged the incubation time and resulted in an inferior quality in cow milk yogurt. Up to now, other studies on production of bioyogurt have mainly focused on evaluating the viability of probiotic bacteria in fermented milk products and their potential health benefits (Maragkoudakis et al., 2006). However, to the best of our knowledge, there is no comprehensive study on the evaluation of the effects of SCT and IFpH on the physicochemical and organoleptic properties of buffalo milk yogurt during cold storage. The process variables affecting the gel formation and syneresis in buffalo milk yogurt are also not well understood.
The combined assessment of organic and fatty acid profiles in yogurt with the activity of different starter cultures is a key issue to identify the effect of SCT on the flavor and aroma characteristics of the final product. The organic acid profile in a fermented dairy food is an indicator of the metabolic activity of added bacterial cultures (Maragkoudakis et al., 2006). Vénica, Perotti, and Bergamini (2014) reported that there was difference between organic acid profiles of yogurt samples having similar population of starter cultures, which was attributed to the changes in matrix structure of yogurt.
Moreover, organic acids play an important role as natural preservatives in yogurt and they also affect the final characteristics of the product and cell viability of probiotics (Shah, 2017). In this context, Cruz et al. (2013) stated that the diversity of proteolytic activity of probiotic cultures caused production of additional organic acids which leads to more complex yogurt structure. The effect of SCT on the distribution of organic acids in buffalo milk yogurt was not explained enough in the literature. Fatty acids have also important role in the organoleptic and nutritional qualities of milk products.
Especially, short-chain fatty acids are involved in developing the sensory quality of products (Beshkova, Simova, Frengova, & Simov, 1998). It showed that yogurt is a good source of n − 3 fatty acids and the change in fatty acids during cold storage depends on milk source (Serafeimidou, Zlatanos, Kritikos, & Tourianis, 2013). Jia, Chen, Chen, and Ding (2016) stated that fatty acid profile of goat milk yogurt strictly depends on amount of carbohydrates and the ratio of starter cultures. Naydenova, Iliev, and Mihaylova (2014) showed that yogurt manufacturing increased the C18:3 fatty acid percentage of the total fatty acids in comparison with the ratio in raw buffalo milk, while it decreased the omega-6 to omega-3 fatty acid ratio. However, the effects of processing conditions and cold storage on the fatty acids content of buffalo milk yogurt are not well described.
SCT may also have an effect on the acidification kinetics during yogurt manufacturing. The final textural and organoleptic properties of the yogurt product are also affected by the IFpH. The interactions between whey protein and casein micelles due to the level of IFpH may result in the formation of different gel structure. Improper texture, acidity, and hardness which reduce the acceptability of the final product may occur due to the extent of IFpH.
In this study, the main objectives are (1) to determine the combined effects of SCT and IFpH on the physicochemical and organoleptic properties of buffalo milk yogurt and (2) to investigate the biochemical and sensory changes during cold storage of buffalo milk yogurt comparatively. The effects of SCT on the organic acid and fatty acid profiles of buffalo milk yogurt were also evaluated in detail.

| Materials
Buffalo milk was obtained from farmers in Kizilirmak Delta in Bafra, Turkey. Plain starter cultures were obtained from Peyma Chr.

| Yogurt manufacturing
The manufacturing procedures for the three different buffalo milk yogurt samples are shown in Figure 1. The milk was standardized to 3 g/100 g fat ratio by the addition of the removed cream to the raw milk using the Pearson square calculation. The mixtures were poured into 200 g polystyrene yogurt cups. Yogurt-making procedure was replicated three times to a total of 18 vats.

| Physicochemical analyses
The pH of the buffalo milk yogurt samples was measured during fermentation and 20 days of cold storage using a pH meter (Eutech CyberScan pH 100, Singapore). Total solids, fat, protein, ash, and titratable acidity were analyzed according to the methods described by Bradley et al. (1992) and AOAC (2000).
F I G U R E 1 Process for manufacturing buffalo milk yogurt samples

| Viscosity
The viscosity of buffalo milk yogurt was determined following the method described previously (Akgun et al., 2016)

| Syneresis
Syneresis was measured according to the method described by Wacher-Rodarte et al. (1993). Five milliliter of yogurt sample was centrifuged at 2,208g for 20 min at 4°C, and the separated whey was measured after 1 min. The syneresis rate (%) was expressed as the volume of separated whey per 100 g of yogurt. The analyses were carried out in triplicate.

| Color
L*, a*, and b* values of buffalo milk yogurt were determined using a sensing Chroma Meter (Konica Minolta CR-400 Series, Japan). Prior to measurement, the instrument was calibrated with its white reference tile. Then, three readings were taken from the surface of each buffalo milk yogurt, and finally the average value was used.

| Organic acids analysis
The HPLC method described by Fernandez-Garcia and McGregor (1994) was carried out for the analysis of organic acids with some following modifications. Yogurt samples (5 g) were weighed into 30 ml Teflon centrifuge tubes and diluted to 25 ml by 0.01 N H 2 SO 4 . The tubes were mixed for 1 min on a vortex mixer. Resulting mixtures were centrifuged at 7000 g for 7 min at 5°C in a refrigerated centrifuge (Sigma 3K30, Germany). Supernatant fractions were filtered through 0.45μm nylon filter (Supelco Iso-Disc ™ , N-25-4 Nylon, 25 mm) into an HPLC vial.
The separation of organic acid was achieved using Shimadzu UFLC system (Shimadzu Corporation, Kyoto, Japan) equipped with ODS-3 column (Shim-pack, 150 × 4.6 mm). Twenty microliter of filtered supernatant fractions was injected. Organic acid was separated isocratically at 0.7 ml/min, at 65°C, using 10 mmol/L H 2 SO 4 as the mobile phase. Lactic, acetic, formic, oxalic, citric, succinic, propionic, and butyric acids were detected at 210-280 nm with PDA detector (SPD-M20A, Shimadzu, Japan). Individual organic acids were identified and quantified using external standards (Sigma Aldrich, Supelco, St. Louis, USA). Quantification of organic acids was performed from the standard curves obtained using five-point solutions of predetermined concentrations. Triplicate extractions and injections were conducted for each sample.

| Fatty acids analysis
Fatty acid composition was determined after methylation (IUPAC, 1990) by a Shimadzu gas chromatograph (ModelGC-2010, Shimadzu Corporation, Kyoto, Japan) using a Optima ® FAP column (60 m × 0.25 mm I.D., 0.25 μm) (Macherey-Nagel, Düren, Germany). The temperatures of the injector port and detector were held at 260°C and 280°C, respectively. The injection volume was 1 μl. The carrier gas was helium at a pressure of 150 kPa. The split used was 1:100. The temperature of the column was held at 140°C for 2 min, raised to 170°C at 10°C/min and 215°C at 10°C/min, held at 215°C for 5 min, raised again to 230°C at 5°C/min and held at 230°C for 5 min, and finally raised to 240°C at 10°C/min and held at 240°C for 35 min. The identification of fatty acids was carried out by comparing their retention times with that of a standard mixture of fatty acids (Supelco 37 Component FAME Mixture, Cat. No. 18919-1 Amp, St. Louis, MO, USA). The final concentration of fatty acids was reported as mg/g fat. All analyses were performed in triplicate.

| Sensory analysis
Buffalo milk yogurt samples were organoleptically examined by a group of 15 panelists at the University of Ondokuz Mayis, Food Engineering Department, according to the method modified from Bodyfelt, Tobias, and Trout (1988), with maximum scores of 10, 5, 5, and 10 for flavor, body and texture, appearance and color, and general acceptability, respectively. The highest and the lowest numbers indicate liking extremely and disliking extremely, respectively. Initially, the panelists were trained in 2-hr sessions prior to evaluation to be familiar with attributes and scaling procedures of yogurt samples. All buffalo milk yogurt samples were coded with three-digit random numbers and presented to the panelists on a tray in individual booths. Orders of serving were completely randomized. Panelists were structured to cleanse their palate with plain crackers and water before tasting each sample.

| Kinetic data analyses
The kinetic data analyses were done according to the kinetic model suggested by Fogler (2010). It was mentioned that the integral method uses a trial and error procedure to find reaction order. In the integral method, the reaction order was guessed and the differential equation used to model the batch system was integrated (Equation 1) where C is the concentration of the substance, k the reaction rate constant, and n the reaction order.
For example, integrating with C = C 0 at t = 0, Zero order; C = C 0 ± kt was obtained and a plot of the concentration of a substance as a function of time will be linear with slope (±k) for a zero-order reaction carried out in a constant volume batch reactor.

| Statistical analysis
Statistical analysis of the data for the effects of the factors on IFpH, viscosity, syneresis, color, organic and fatty acid profiles, and sensory (1) dC dt = ± kC n properties was performed by one-way and three-factor randomized complete block design using SPSS 13.0 statistical software. The mean differences were analyzed using Tukey's multiple-range test at 5% significance levels.

| Compositional analysis and physicochemical properties
After the first day of cold storage, the fat, total solids, ash, and protein contents of buffalo milk yogurt samples were 3.1 ± 0.0 g/100 g, 13.4 ± 0.2 g/100 g, 0.8 ± 0.0 g/100 g, and 4.2 ± 0.1 g/100 g, respectively. No significant (p > .05) effects of the SCT and IFpH) on the final compositions of the buffalo milk yogurt samples were observed since they were directly related with the composition of buffalo milk.
The initial pH of the buffalo milk was standardized to reduce variability in the length of fermentation as stated by Nguyen et al. (2014).
It was observed that the SCT primarily affected the acidification ki- Considering the whole period of shelf-life, the titratable acidity increased with the postacidification during cold storage. As shown in   It was also indicated that faster exopolysaccharide production can be occurred when mixed cultures were used (Shihata & Shah, 2002).
The syneresis values of the buffalo milk yogurt samples containing YC-X11 and ABY-2 cultures ranged from 39.00% to 41.67%. The IFpH did not significantly affect the syneresis (p > .05). As shown in Table 2, the absence of L. delbrueckii subsp. bulgaricus in the culture increased the syneresis significantly (p < .05). In this case, it was also easily mentioned that there was no significant difference in T1 and T2 again (p > .05). Şanlı, Şenel, Sezgin, and Benli (2014) mentioned that the body of low-fat yogurt could be improved  Table 2). Therefore, the primary reason for higher syneresis was considered to be the SCT instead of the fermentation period. Means with same letters (A-C) in a row within the category data are not significant at p > .05. Means with same letters (a-d) in a column within the category data are not significant at p > .05. Table 2, a significant decrease in syneresis values of X1, X2, T1, and T2 was recorded after 20 days of cold storage (p < .05). Mahmood, Abbas, and Gilani (2008) reported similar trends in yogurt samples made with different commercial probiotic cultures. As it was indicated by Mende et al. (2016), the interaction between polysaccharide molecules and protein network is related to the acidity of the medium.

As shown in
It is generally assumed that polysaccharides that are directly attached to protein network are charged and acidity of a medium changes the charges of protein molecules. Repulsive or associative interactions or segregative phase separation can be occurred depending on pH and, thus, protein charge. Unlike the other buffalo milk yogurt samples, a significant increase was observed in the syneresis values of Y1 and Y2 after 10 days of cold storage (p < .05). The increase in syneresis might be related to the count of L. acidophilus as stated before (Olson & Aryana, 2008).
Whiteness (L*), red/greenness (a*), and yellow/blueness (b*) values of the buffalo milk yogurt samples are shown in Table 3. The mean L* value of buffalo milk yogurt samples was higher than that found in dahi (Raju & Pal, 2009), and was lower than that of found in probiotic yogurt (Olson & Aryana, 2008). L*, a*, and b* values of all buffalo milk yogurt samples were not significantly affected by the IFpH and the SCT (p > .05). It was also observed that 20 days cold storage did not significantly affect the L* values of X2, Y1, Y2, T1, and T2 (p > .05), and only the L* values of X1 (p < .05) were significant. Although the decrease in L* values of all the buffalo milk yogurt samples was observed in this study, these reductions were statistically insignificant (p > .05). The a* values of probiotic buffalo milk yogurt samples were closer to that found in a probiotic yogurt by Olson and Aryana (2008). As shown in Table 2, an irregular trend was observed in syneresis values of probiotic buffalo milk yogurt samples, which was attributed to the amount of syneresis during cold storage (Shirai et al., 1992),

| Sensory evaluation
The changes in sensory properties of the buffalo milk yogurt samples during cold storage are shown in Table 4. It is clear that the IFpH did not significantly affect the flavor, texture, appearance, color, and general acceptability of all the buffalo milk yogurt samples (p > .05). The highest flavor and general acceptability scores were obtained for the yogurts made with ABY-2 culture (p < .05).
However, the effects of the SCT on the body, texture, appearance, and color of the buffalo milk yogurt samples were insignificant The most noted defects by the panelists were insufficient acidity and plain taste for X1 and X2 at the 1st day of the cold storage.
Thirteen percent of the panelists also indicated that X1 has rougher and grainier texture than the other buffalo milk yogurt samples.

| Organic acid profile
In the following part of this study, organic acid and fatty acid contents of Y2 were analyzed and compared with X2, as the highest flavor and general acceptability scores were obtained for Y2 (p < .05) according to sensorial evaluation.
Lactic, acetic, citric, and butyric acids were identified as the main organic acids in X2 and Y2. The changes in the amount of these organic acids during cold storage are presented in Figure 5. Lactic acid and citric acid were the most abundant acids in Y2 and X2 samples during cold storage. Unlike X2, a little amount of formic acid, which might be due to the growth of Bifidobacteria as stated by Seckin and Ozkilinc (2011), was also detected in Y2 during cold storage (data not shown). Formic acid was not significantly affected by the storage time (p > .05). It was probably due to the limited growth of Bifidobacteria during cold storage. Adhikari, Grün, Mustapha, and Fernando (2002) also indicated that the metabolic inertness of the bifid bacterial cells can occur when they were added to the yogurt after fermentation. It was found that the concentrations of lactic acid, acetic acid, and butyric acid in Y2 were higher than those of X2. The effect of the SCT on lactic acid and acetic acid production was not significant (p > .05). However, a significant difference was observed in butyric acid concentration (p < .05). Similarly, Donkor, Henriksson, Vasiljevic, and Shah (2006) reported that lactic acid concentration in yogurts containing probiotic culture was higher than the control containing standard yogurt culture. They also stated that the differences in lactic acid concentration were not statistically significant (p > .05). In contrast to the concentrations of other acids, it can be stated that X2 has higher amount of citric acid than Y2 (Figure 5c). Adhikari et al. (2002) noted that the citric acid content of the yogurts containing probiotic strains was lower than that of the control containing standard starter  Adhikari et al. (2002) reported that there was no significant effect of the storage period on the concentration of butyric acid in set and stirred types of cow milk yogurt (p > .05). The SCT significantly affected the production of lactic acid, acetic acid, and citric acid during cold storage ( Figure 5). The increases in the amounts of all organic acids presented in Figure 5 were found to be linear for X2 during 20 days of cold storage. However, for Y2, the amounts of lactic acid, acetic acid, and citric acid increased more rapidly in the first 10 days of cold storage. Then, the rate of increase slowed down gradually and further slight increases were observed throughout the storage period.

| Fatty acid profile
The changes in concentrations of fatty acids in X2 and Y2 during cold storage are presented in Table 5. In both Y2 and X2 samples, was also easily noted that 20 days of cold storage did not significantly (p > .05) change the ratio of saturated fatty acids to unsaturated fatty acids. In general, the amounts of fatty acids in X2 and Y2 did not follow a regular trend during cold storage. Similar results were reported by Güler (2007), while Yadav et al. (2007) observed an increase in the amounts of fatty acids throughout the storage. Yadav et al. (2007) indicated that butyric acid was formed principally by lipolytic activity of lactic acid bacteria, and the lactobacilli having probiotic activity can produce conjugated linoleic acid (CLA) in milk during fermentation by lipolysis of natural milk fat. Alternatively, distinct changes in the butyric acid and linoleic acid contents of X2 and Y2 were not observed in this study, which can be attributed to the pH levels of the both Y2 and X2 samples. Kim and Liu (2002) also indicated that pH below 4.6 has more effect on CLA synthesis than SCT. Only, the decrease in linoleic acid content of X2 and Y2 samples was observed after 20 days of cold storage and it might be explained that the starter cultures used in this study utilized the linoleic acid as a substrate for CLA synthesis as stated by Yadav et al. (2007). Beshkova et al. (1998) stated that the hydrolytic activity of S. thermophilus toward milk fat is low, and the most important precursors of volatile fatty acids were amino acids.
They also indicated that formation of volatile fatty acids (C2-C10) was more active in mixed cultures than that of in pure ones. However, SCT did not significantly affect the amounts of volatile fatty acids in this study (p > .05).

| CONCLUSIONS
The results of the presented study highlighted that the processing conditions such as the SCT and the IFpH altered the physicochemical and organoleptic properties of buffalo milk yogurt samples con-