Effect of pH at drainage on the physicochemical , textural and microstructural characteristics of mozzarella cheese from goat milk

Cheese may be defined as a structure of bi-continuous gel consisting of a porous protein matrix interspersed with fat globules. Mozzarella, a variety of pasta filata cheese originally made from buffalo milk in Italy, is one of the most widely consumed cheeses in the world (Vogt et al., 2015). It is characterized by its particular structure and plastic consistency (Niro et al., 2014), resulting from the thermo-mechanical stretching to which the curd is subjected during processing. This process aligns the casein fibers parallel, interlacing them with fat and serum channels (Vogt et al., 2015).


Introduction
Cheese may be defined as a structure of bi-continuous gel consisting of a porous protein matrix interspersed with fat globules.Mozzarella, a variety of pasta filata cheese originally made from buffalo milk in Italy, is one of the most widely consumed cheeses in the world (Vogt et al., 2015).It is characterized by its particular structure and plastic consistency (Niro et al., 2014), resulting from the thermo-mechanical stretching to which the curd is subjected during processing.This process aligns the casein fibers parallel, interlacing them with fat and serum channels (Vogt et al., 2015).
The physicochemical properties, and the textural and microstructural characteristics of mozzarella cheese are affected by several variables involved in processing, such as milk composition, whey pH prior to stretching, straining time and temperature, and storage conditions (Cavalier-Salou & Cheftel, 1991;Ennis & Mulvihill, 1999;Hennelly et al., 2005;Lee et al., 2004).However, the whey pH at draining is the most influential variable in the ability of the curd to be laminated and stretched in hot water (Yazici et al., 2010), determining the point at which the yarn is possible.
In conventional production of mozzarella cheese from cow milk, the pH drops to a range between 5.4 to 5.1 (Yazici et al., 2010), thus favoring the solubilization of micellar calcium phosphate and the hydration of the para-casein, and increasing the proportion of soluble colloidal calcium (Guinee et al., 2002), which can affect final product quality (Yun et al., 1995).
Goat milk is a natural source of oligosaccharides derived from lactose, conjugated linoleic acid, A-and B-complex vitamins, and calcium (Haenlein & Anke, 2011;Paz et al., 2014).It also tolerates different technological treatments, thus yielding products with the ability to meet consumer demands in terms of health, nutritional value, safety and sensory attributes, indicating a promising future for the dairy goat sector.Thus any effort to encourage the production and scientific research in this field is positive (García et al., 2014).
Several studies have reported the effects of pH on the physicochemical properties of mozzarella cheese made from cow milk (Yun et al., 1993;Joshi et al., 2002Joshi et al., , 2004;;Zisu & Shah, 2005) and buffalo milk (Yazici et al., 2010), but no in goat milk.In this sense, the definition of appropriate pH for the obtain of goat mozzarella cheese with appropriate characteristics, is of great interest.
Therefore, the aim of this study was contribute to standardize the making of mozzarella cheese from goat milk, by trying different whey pH at draining, in order to obtain a product with suitable physicochemical, textural and microstructural characteristics.

Materials
Pasteurized goat's milk variety Saanen and Creole (80:20) (15 L), was obtained from producers in the town of Vaqueros, Salta, Argentina in November; and it was transported to the Food Laboratory, of Facultad de Ciencias de la Salud, in 4-liter jerry cans of high-density polyethylene, in storage containers at 4 °C, in about 7 min.It was stored at 4 °C in cans of high-density polyethylene (4 L), for 7 minutes Mozzarella cheese was prepared with CHY-MAX TM clotting enzyme (Chr.Hansen, Córdoba, Argentina), 100% pure Chymosin produced by fermentation, which requires lower dosage and facilitates better control of manufacturing processes; lyophilized lactic bacteria for direct inoculation as growing starters: Streptococcus thermophilus (ST-M7  ) and Lactobacillus helveticus (LB-12  ) (Chr.Hansen, Córdoba, Argentina); calcium chloride (Anedra laboratory, Buenos Aires, Argentina) and sodium chloride (Merck, Buenos Aires, Argentina).

Standardization of the concentration of coagulant enzyme and starter cultures
Seven servings of goat milk, 100 mL each, were inoculated with starters, alone and combined, in different ratios and quantities, as recommended by the supplier (Table 1).The sample selected was the one reaching a pH of 6.0 in the shortest time, at which point the enzyme has been shown to present the highest proteolytic activity on casein (Van Hooydonk et al., 1986).
From the above test result, the milk temperature was raised up to 45 °C, and Chymosin was added at the minimum, medium and maximum concentration recommended by the supplier: 0.001, 0.002, and 0.003 g 100 mL -1 , respectively.The sample was allowed to stand for 30 min at the same temperature until a firm, homogeneous and glossy clot was obtained.The gel hardness was measured with a QTS texture analyzer (Brookfield), and the enzyme concentration providing the greatest consistency to the curd was chosen.

Cheese manufacture
CaCl 2 at 0.02 g 100 mL -1 was added to the pasteurized whole milk and then the selected concentrations of the starter cultures obtained in the previous assays were added.Chymosin was then added and allowed to clot for 30 min.The curd was cut into 2 x 2 x 2 cm cubes, i.e., large pieces to promote the moisture retention required for this type of soft to very soft cheese.The whey was partially removed and maturation took place under whey, up to a pH value that would allow subsequent kneading and stretching of the curd to obtain yarn-like threads.
According to the literature review, the general ranges of pH drainage for making pasta filata cheese (Cheddar, Daxaca, Mozzarella) from different milks (cow, buffalo, goat, sheep, and mixtures thereof), could range from 4.9 to 5.8 (Zisu & Shah, 2005;Yazici et al., 2010;Bermúdez-Aguirre& Barbosa-Cánovas, 2012;Morales-Celaya et al., 2012;Ganesan et al., 2014;Niro et al., 2014;Dng et al., 2014).At pH 5.0 (results not shown) yarn not occur, possibly by excessive demineralization mass, accompanied by protein denaturation and proximity to its isoelectric point that would more unstable.Therefore, this study tested three-point pH (Figure 1): Then, curd was immersed in hot water and spun until it was elastic and fibrous.After shaping (moulding), and to stop the cooking, it was then immersed in cold brine (5 ± 1 ºC) at 20 g 100 mL -1 NaCl, for 30 min.Finally, it was allowed to stand for 24 h before the analysis was performed.
The CY was expressed as the nutrient recovered from milk (Solorza & Bell, 1998), which is especially related to the content of protein, fat and calcium in the clot associated with moisture.The actual performance of each batch of cheese was expressed as kg of cheese produced per 100 kg of goat milk used in its production.

Texture Profile Analysis (TPA)
The textural characteristics (hardness, cohesiveness, adhesiveness, elasticity and chewiness) of the cheeses were analyzed with QTS (Brookfield) equipment.Details of the selected parameters have been previously described (Pons & Fiszman, 1996).Refrigerated specimens (4 °C) were cut with a cylindrical mold of 20 mm in diameter x 15 mm in length, and allowed to equilibrate at room temperature (~ 22°C) for 10 min.A 500N load cell with a flat plunger was used at a speed of 30 mm min -1 .A 50% compression was selected to permit deformation without breaking the sample.Data were collected through Pro v2.0 software.In addition, a Hewlett Packard C3180 scanner was used to photograph the macrostructure of the samples during the process.

Scanning Electron Microscopy (SEM)
A Jeol scanning electron microscope (JSM 6480 LV, Tokyo, Japan) was used.The equipment has a capacity acceleration voltage of between 0.2 and 30 kV, including sensors for secondary and backscattered electrons, working with high and low vacuum.The cheeses were analyzed using cylindrical cuts of 5 mm in diameter and 2 mm in height and fixed in 36% formaldehyde at refrigeration temperature for 2 months.Subsequently dehydrated with a graded series of ethanol (30, 50, 70, 80, 90 and 100 g per 100 g, 15 min each) and dried by critical point, they were metallized with gold and finally placed in the SEM tubes to determine their internal microstructure, according to that described by Morales-Celaya et al. (2012), with some modifications.The most representative SEM micrographs were selected for presentation.

Color analysis
A ColorTec PCM colorimeter  (Accuracy Microsensor Inc., Pittsford, USA) was used for instrumental color evaluation.The CIE Lab color scale (L*a*b*) was used with a D 65 illuminant (standard daylight) and measuring angle of 10º.The visual sensations that are sent to the brain create the three dimensions of colour judgment response that is often referred to as three-dimensional colour space.In the CIE Lab system, these dimensions are expressed as: L* related to lightness varying from black (zero) to white (100), and other two related to chromaticity, a* from green (-a*) to red (+a*) and b* from blue (-b*) to yellow (+b*).The L*, a* and b* parameters were determined according to the Commission Internationale de l'Eclairage.Using reference plates, the apparatus was calibrated in the reflectance mode with specular reflection excluded.A 10-mm quartz cuvette was used for the readings.Measurements were performed in triplicate using the inner section of the cheeses immediately after unpacking.

Statistical analysis
All analyses were carried out in triplicate.The means of the results were evaluated using analysis of variance (ANDVA), and Duncan test was used to compare significant differences (P<0.05) by the statistical software StatProTM (Microsoft  Dffice Excel   2007, version 12.0.4518.1014).The data are presented as mean ± standard deviation of replicate measurements.

Physicochemical analysis of milk
The results of the physicochemical analysis of the pasteurized goat milk are shown in Table 2.
The pH was 6.66; this value depends mainly on the stability of casein and is affected by the weaning period, the diet, the breed and the presence of carbon dioxide bubbles released during milking, cooling and transport of milk (Albenzio & Santillo, 2011).The acidity was lower than the range set (14-22 ºD) by Administración Nacional de Medicamentos, Alimentos, y Tecnología Médica (2014).The presence of casein, minerals and ions increase the acidity titratable, during the spring-summer period.The protein content was similar to those reported in the literature (Albenzio & Santillo, 2011;Paz et al., 2014), but lower than the minimum established by law (2.8 g 100 g -1 ) (Administración Nacional de Medicamentos, Alimentos, y Tecnología Médica, 2013).The fat was higher than the minimum set (3.0 g 100 g -1 ) by Administración Nacional de Medicamentos, Alimentos, y Tecnología Médica (2014), mainly influenced by the diet of the animals (Park et al., 2007).The non-fat dry matter was also below the established minimum (9.0 g 100 g -1 ) (Administración Nacional de Medicamentos, Alimentos, y Tecnología Médica, 2013).The minimum content of protein, fat dry matter not less than the minimum established by law and reduced ability to produce mozzarella cheese goat milk compared with cow, would turn into a difficult procedure to perform (Niro et al., 2014).

Standardization of the concentration of coagulant enzyme and starter cultures
Sample 5 showed the shortest pre-maturation time (30 minutes), with a concentration of 0.010: 0.015 g 100 mL -1 of milk with Streptococcus thermophilus and Lactobacillus helveticus, respectively.While the initial pH of goat milk is around 6.7, this may change if one resorts to the use of freeze-dried lactic cultures or organic acids, with the latter products of low sensory quality (Morales-Celaya et al., 2012).Furthermore, it has been shown that the lower the pH, the faster the coagulation time (Hannon et al., 2006).The effect of low pH in milk accelerates the activity of the clotting enzyme, reduces the electrostatic repulsion between casein micelles, and alters the distribution of calcium between micelles and serum phase (Dng et al., 2012), resulting in a shorter product development time.
Regarding the clotting enzyme, it was observed that the higher the concentration (0.003 g Chymosin per 100 mL -1 milk), the greater the hardness of the gel (9 g).

Physicochemical analysis of goat mozzarella cheese
The data of the physicochemical analysis of the mozzarella cheeses made from goat milk are summarized in Table 3.
The pH of the samples were slightly higher than the drainage pH prior to kneading and stretching, for each cheese.Moisture of the MC50 and MC53 samples was higher than that of MC56; however, they were all "high humidity" according to the general classification of cheese (46 to 54.9 g 100 g -1 ) (Administración Nacional de Medicamentos, Alimentos, y Tecnología Médica, 2013).The protein content was similar among the products, but greater at higher pH of the whey.
The lower retention of fat was obtained in the MC56 sample, which had to be subjected to heating for a longer time for the strains to form; this could have caused a greater loss of fat in the hot water used for stretching the curd, caused by the smaller fat globules present in the goat milk (Paz et al., 2014).However, all the samples exceeded the minimum required by law for this type of cheese (35.0 g 100 g -1 ) (Administración Nacional de Medicamentos, Alimentos, y Tecnología Médica, 2013).Total ashes were similar between samples.Furthermore, calcium retention was higher at higher pH (Dng et al., 2014); this was due to the shorter time required for solubilization of micellar calcium phosphate and to the increase in the proportion of colloidal soluble Ca (Guinee et al., 2002).With respect to sodium, although our country does not record benchmarks in terms of content for this type of cheese, the values were lower than those of commercial mozzarella reported in other studies (Agarwal et al., 2011).As to phosphorus, all samples had a Ca/P ratio greater than 1, so this would not be the factor triggering bone loss (Teegarden et al., 1998).
The cheese yield was lower at pH 5.6, because of the prolonged exposure to stretching and kneading in hot water, with resulting loss of moisture and fat.

Texture Profile Analysis (TPA)
The macro-photographs, taken during the manufacture of cheese (Figure 2) show that at first the curd lacked structure, because 15 minutes would not be enough for the clotting enzyme Table 2. Physicochemical parameters of milk used in the production of mozzarella cheese.
to have an effect; at 90 minutes protein aggregation gradually formed, and at 150 minutes the curd looked more even and continuous.In addition, there was a big difference between the curd and the finished product, where clearly a fibrous structure was observed.The characteristic texture of the pasta filata cheese can be explained by the structural rearrangement that casein molecules (αS, β and κ, which are part of the decalcified micelles) suffer when the curd is subjected to heating and mechanical work, altering its β-plate and α-helix conformation (Ren et al., 2013).Stretching spatially in one direction positions and "aligns" the proteins as if they were "threads".In addition, butterfat would function as a lubricant aligning the casein fiber (Yazici et al., 2010).
Regarding the texture, the MC56 sample showed the highest hardness value (Table 3), which could be related to the lower moisture content caused by long kneading and stretching in warm water, and to the higher content of protein and, therefore, of micellar calcium retained in the curd, all of this resulting in a less adhesive product.
For the remaining parameters (cohesiveness, elasticity and chewiness), no significant differences were observed among the samples.

Scanning Electron Microscopy (SEM)
The microstructure changes influence the physicochemical properties of the cheese, so it is essential to study how they affect the processing conditions.Figure 3 shows the SEM micrographs of cheeses at different pH of draining.
The MC50, MC53 and MC56 samples have different structural arrangements according to the pH used for processing, mainly 25.70 ± 0.43 a 26.50 ± 0.51 a 28.87 ± 0.51 a Fat (g 100 g -1 ) 42.44 ± 0.05 b 45.07 ± 0.28 a 38.33 ± 0.28 c Total Ash (g 100 g -1 ) 2.64 ± 0.10 a 2.98 ± 0.12 a 3.27 ± 0.17 a Calcium (mg 100 g -1 ) 601.30 ± 1.00 c 725.95 ± 14.31 b 853.58 ± 24.45 a Sodium (mg 100 g -1 ) 284.97 ± 3.75 a 277.20 ± 9.50 a 298.70 ± 6.42 a Phosphorus (mg 100 g -1 ) 573.64 ± 0.29 c 620.25 ± 0.68 b 644.28 ± 0.84 a CY (g 100 g The following abbreviations are used: MC50, curd kneaded and stretched at pH 5.0; MC53, curd kneaded and stretched at pH 5.3; MC56, curd kneaded and stretched at pH 5.6.Mean ± standard deviation (n=9).Different letters on the same line indicate significant differences.regarding the presence and distribution of the casein threads (Figure 3).Components such as fat and moisture were removed from the cheese during preparation of the samples; therefore, the solid background of each micrograph represents the protein matrix, while the empty black areas represent air sacs, originally occupied by fat cells or serum that may have sublimated.The fat globules of milk have been considered responsible for forming the spherical empty spaces inside the protein matrix of mozzarella cheese (Dberg et al., 1993).
The presence of large conglomerates of protein that can be seen in the figures, possibly merged during the formation of the curd, resulting in a compact structure within the matrix.MC50 has few threads (thickness of about 21.5μm) and pores (between 24.4-55.7 μm diameter).In MC56 no threads were observed, but possible to obtain a product of adequate moisture, fat, protein and calcium, as well as higher cheese yield; lower hardness and microstructure revealed the presence of long, thin threads, which gave it a distinctive texture.Additionally, given the almost total absence of scientific articles in this type of spun paste cheese, the results provide a clearer characterization of this product.several small pores were present (between 3.42-39.0μm).Because of the high pH of this sample, the curd had to be subjected to longer time in hot water, thus causing loss of moisture in the product.When the volume of the aqueous phase diminished, casein micelles were able to get closer together, leaving little space for the formation of large pores.Finally, MC53 showed the presence of long (up to 60μm) and thin (5.60-9.20μm)threads, keeping the same network within the matrix and even in fissured surface areas (Figure 4).

Color analysis
The average L* values found for all the samples in this study (Table 3) were lower than those found by Niro et al. (2014) for pasta filata cheese (Caciocavallo) made from a mixture of cow milk and goat milk.In color evaluation, the parameter L* indicates the lightness and the capability of an object to reflect or transmit light on a scale of 0 to 100.Therefore, higher brightness values result in lighter objects (Ramos do Egypto Queiroga et al., 2013).
The parameter a* is higher than that observed by Niro et al. (2014).High values of a* (green component) in goat milk products have been primarily attributed to their fatty acid profiles.Cheeses made from goat milk are generally whiter because goats are capable of converting β-carotene into vitamin A and also produce milk fat globules that are smaller in diameter than those in cow milk (Lucas et al., 2008;Park, 2006).Thus, when the cheeses are made from milk mixtures, increasing values of a* are directly related to the higher proportion of goat milk (Sheehan et al., 2009).
Parameter b* (yellow component) showed no significant differences between samples.However, their values were lower than those reported by Niro et al. (2014).The increase in b* values has been associated with the Maillard reaction, which decreases brightness due to the production of browning compounds (Lucas et al., 2008).
In general, the tested samples showed high levels of luminosity (L*), with a predominance of the yellow component (b*), instead of the green component (a*), suggesting that white-yellow is the predominant color feature in this type of cheese.

Conclusions
To conclude, the goat milk had adequate aptitude for processing this type of cheese, but only at whey drainage pH 5.3 was it

Figure 1 .
Figure 1.Flow chart of processing of mozzarella cheeses made from goat milk.

Figure 2 .
Figure 2. Photographs of the different stages of the curd: (A) 15 min post addition of Chymosin, pH 6.57; (B) at 90 min post addition of Chymosin, pH 6.00; (C) at 150 min post addition of Chymosin, pH 5.77; and, (D) mozzarella cheese from goat's milk after kneading and stretching.

Figure 4 .
Figure 4. Dimensions of the yarns present in mozzarella cheese from goat, kneaded at different drainage pH: 5.0 (a) 5.3 (b) and 5.6 (c).

Table 1 .
Combinations of starter culture at various ratios and concentrations used to achieve a pH of 6.0.

Table 3 .
Physicochemical parameters, texture and color of the mozzarella cheeses made from goat milk.