Characteristics of imitation Mozzarella cheese manufactured without emulsifying salts using a combination of culture‐based acid curd and micellar casein concentrate

Abstract The objectives of this study were to develop a process to produce acid curd from micellar casein concentrate (MCC) using starter cultures and to manufacture imitation Mozzarella cheese (IMC) using a combination of acid curd and MCC that would confer emulsification ability to the caseins without the use of emulsifying salts (ES). The formulations were targeted to produce IMC with 49.0% moisture, 20.0% fat, 18.0% protein, and 1.5% salt. In the IMC formulation made without ES (FR‐2:1), the acid curd was blended with MCC so that the formula contained a 2:1 ratio of protein from acid curd relative to MCC. IMC with ES was also produced as a control. The melt and stretch characteristics of IMC made from FR‐2:1 were similar to those of control IMC. We conclude that IMC can be made without ES using a 2:1 ratio of protein from acid curd relative to MCC.


| INTRODUC TI ON
Microfiltration (MF) is a membrane process that is utilized to fractionate casein (CN) and serum protein (SP) from skim milk using a 0.1 μm semipermeable membrane. The skim milk is force driven through the membrane to separate CN (retentate side) and SP (permeate side) based on their sizes (0.1-0.4 μm CN vs. 0.003-0.01 μm SP). The retentate is called micellar casein concentrate (MCC), which is mostly native casein with approximately 9.0% total protein (TPr) and 13.0% total solids (TS). Micellar casein concentrate is a high protein ingredient that is typically manufactured in 3 MF stages using a 3× concentration factor (CF) with diafiltration (DF).
The IMC is a dairy, partial dairy, or nondairy food depending on the sources of protein and fat used in the formulation. The typical protein source for IMC is rennet casein and fat sources can range from milk fat to vegetable oils/fats depending on the final nutritional, functional, and cost targets. This type of cheese is one of the popular analogue cheeses in the United States due to its applications in pizza (Bachmann, 2001;O'Riordan et al., 2011). It also offers manufacturers the flexibility to produce a final product with desired characteristics (shreddability, meltability, flowability, stretchability, chewiness, oiling off, and browning on baking) that are more consistent over a longer shelf-life when compared to natural cheese (Bachmann, 2001;Guinee et al., 1999;O'Riordan et al., 2011). The meltability and stretchability characteristics are important in IMC due to the application of this type of cheese in pizza.
The IMC has the same basic principles that are used in the manufacture of PC in terms of process and equipment. It is prepared by blending dairy (rennet casein and milk permeate) and nondairy ingredients (edible oils/fat, protein, and emulsifying salts: ES, NaCl, acidulant, and water) with the aid of heat and shear to produce a homogeneous product (Hammam, Beckman, et al., 2022;Metzger & Hammam, 2020;O'Riordan et al., 2011 (Guinee et al., 2004), which prevents oil separation in IMC. In the presence of heating and mixing, a homogeneous IMC is produced.
Acid curd is a protein concentrate (>80% total protein as a percentage of total solids), which can be obtained by precipitating casein at pH 4.6 (isoelectric point) using starter cultures or acids without the use of rennet. The colloidal calcium phosphate in the micelles is dissolved in the whey at a pH of 4.6, and this results in acid curd with low mineral or calcium content. In contrast to the low mineral content in acid curd, MCC has a high level of casein-bound calcium with a pH of 6.5-6.7. If acid curd is mixed with MCC ( Figure 1), it may be possible to create a partially deaggregated casein network without the use of ES. The ratio of acid curd to MCC will have an impact on the level of deaggregation and the pH of the final product. In our previous patent, we hypothesized that a ratio of two parts of protein from acid curd to one part of protein from MCC created a partially deaggregated casein network similar to a typical PC that utilizes ES ( Figure 1) (Metzger & Hammam, 2020), and we hypothesized that this can also occur in IMC formulations.
Acid curd can be produced from skim milk in a process similar to Cottage cheese manufacture. There is a possibility of using MCC instead of skim milk to produce acid curd. Making acid curd from MCC has advantages as compared to skim milk, since manufacturing MCC using MF results in milk-derived whey protein as a by-product which can be utilized in many value-added applications, particularly making whey protein isolate. In contrast, acid curd produced from skim milk results in acid whey as a by-product, which is more challenging to utilize. The typical composition of MCC (three stages using 3× CF with DF) is >9% true protein (TP) and >13% TS (Zulewska et al., 2009). This MCC could be used immediately in making acid curd or diluted to lower protein levels before making acid curd if required. In our previous studies, we produced acid curd from MCC with different protein content (3%, 6%, and 9%). We found that MCC with 9% protein is the optimum ingredient to produce acid curd (pH = 4.6) using lactic acid Metzger & Hammam, 2020).
Although direct acids such as lactic acid take less time to produce the acid curd, those acids are costly and produce acid curd with a bland flavor compared to the varieties of starter cultures that have less cost and can be utilized to develop flavors in the curd.
Many consumers are perceiving ES as chemicals which are reducing the consumption of products like IMC. As a result, manufacture of IMC with no ES would meet the consumer desire. Therefore, the objectives of this study were to develop a process to produce acid curd from MCC (~9% TP and ~13% TS) using starter cultures and to determine if IMC could be produced in the absence of ES using a combination of acid curd and MCC in the formulations.

| Preparation of MCC solution
The MCC solution (pH ~6.6) was prepared ( Figure 2) and standardized by mixing MCC powder (CasPro 8500, Lot # NF8109A1, Milk Specialties Global, Eden Prairie, MN 55344), milk permeate powder (product lot: 19113D40, Idaho Milk Products, ID), and water to produce a recombined MCC with an average of 13.0% TS, 9.0% TPr, and 2.0% lactose (the minimum lactose amount that is required to be fermented by the starter cultures to get the pH of 4.6). Techwizard software (Excel-based formulation software program provided by Owl Software) was used to prepare 1 L of the recombined MCC. Powder ingredients were mixed with water using a magnetic stirring plate for 1 h at room temperature and then batch pasteurized at 65°C for 30 min.

| Manufacture of culture-based acid curd
Thermophilic cultures (i455, Batch no 3489654, Chr Hansen) were added at a rate of 0.005% to the recombined MCC and incubated at 43°C (Major Science, Saratoga, CA 95070, USA) for about 15 h to decrease the pH to 4.6 ( Figure 2). After reaching the pH of 4.6, the curd was cut and stirred gently during heating to 50°C for 1 h. The whey was subsequently drained, and the curd was washed with water at a 1:1 ratio, pressed, and kept in the freezer for further analyses. This trial was repeated three times.

| Manufacture of imitation Mozzarella cheese
Techwizard was also used to develop the IMC formulations (Metzger & Roland, 2017) to produce IMC with 49.0% moisture, 20.0% fat, 18.0% protein, and 1.5% salt. The percentage of ingredients utilized in IMC formulations is shown in Table 1. In the formulation of 2:1 (FR-2:1), the amount of protein from acid curd and MCC in the mixture was adjusted to have a ratio of 2:1, respectively. The ingredients of FR  Water (0.5 g) was added to each canister to compensate for the water that evaporated during mixing and cooking in RVA. The sample was stirred for 2 min at 1000 rpm and subsequently at 160 rpm for 1 min.
Ten canisters from each batch were cooked in the RVA. The canisters were poured into plastic molds (diameter = 28.3 mm; height = 25 mm) to measure the melt temperature using dynamic stress rheometry (DSR) and melt diameter using the Schreiber melt test.

| The end apparent cooked viscosity
The end apparent cooked viscosity was measured following the same procedures we performed in our previous studies (Hammam, Beckman, et al., 2022). The end apparent cooked viscosity of the IMC was measured by the end of the cooking time using the RVA at 95°C by calculating the mean value of the last five values, which is referred to as the end apparent viscosity.
The end apparent cooked viscosity was measured in all canisters of each batch.

| Dynamic stress rheometry (DSR)
The DSR was performed following previous studies (Hammam, Beckman, et al., 2022). The IMC sample was prepared by cutting the cheese into slices (2 mm thick and 28.3 mm diameter) using a wire cutter. A stress sweep test of the IMC was performed at a frequency of 1.5 Hz, and stress ranged from 1 to 1000 Pa at 20°C using a rheometer with parallel plate geometry (MSR 92, Anton Paar, Graz, Austria). The stress sweep experiment revealed that the maximum stress limit for the linear viscoelastic region was 500 Pa. The dynamic rheological properties of the IMC were then analyzed with a dynamic temperature ramp test that ranged from 20 to 90°C with a ramp rate of 1°C/min using a frequency of 1.5 Hz and constant stress of 500 Pa. The temperature at which tan δ = 1 (G″/G′) was referred to as the cheese melting temperature. A duplicate test was performed on each batch.

| Schreiber melt test
The IMC samples were cut into cylinders (diameter = 28.3 mm and height = 7 mm) and placed in Petri dishes. The dishes were transferred to a forced draft oven at 90°C for 7 min Muthukumarappan et al., 1999). After cooling the dishes to room temperature, the diameter of the melted IMC samples was measured in four different places using a Vernier caliper and reported F I G U R E 2 Schematic diagram for manufacturing of recombined micellar casein concentrate (MCC), acid curd, and imitation Mozzarella cheese (IMC). in millimeters. The melting area (A) was calculated using the radius I of the cheese (A = π r 2 ). This test was repeated four times for each batch.

| Stretchability test
The stretchability of IMC samples was measured using the method described by (Gunasekaran & Ak, 2003) with some modifications. The test was done by placing a cylinder of cheese (diameter = 28.3 mm and height = 7 mm) in a glass Petri dish (95.0 mm diameter) and left in a forced draft oven at 232°C for 3 min and 30 s (Muthukumarappan et al., 1999). It was cooled for 30 s and then a four-pronged fork was inserted into the cheese.
Subsequently, the fork was lifted vertically and the distance before breaking the cheese was measured in centimeters. This test was replicated four times.

| Statistical analysis
Statistical analysis of the data generated was performed to study the effect of formulations on the functional properties of IMC. An ANOVA was done using R software (R × 64-3.3.3, R Foundation for Statistical Computing, Vienna, Austria). Differences were tested using the least significant difference test at p < .05.

| Composition of acid curd and acid whey
The composition of acid curd produced from MCC is shown in Table 2. The mean composition of acid curd from three replicates was 25.8% TS, 23.7% protein, 0.9% ash, 0.2% lactose, 0.6% lactic acid, 0.2% Ca, and 0.1% P. The average pH of acid curd was targeted to have 4.6. The composition of acid curd depends on the composition of initial materials, final pH, and process conditions (e.g., cooking temperature, washing curds, and pressing) (Wong et al., 1976).
The step of washing the curd could decrease the ash, Ca, P, lactose, and lactic acid content in the final acid curd. Since the lactose is converted into lactic acid using starter cultures as a result of fermentation, lactose decreased while lactic acid increased as the pH reached 4.6. It is also expected that the ash content would decrease when the reconstituted MCC solution is converted into acid curd (Hill et al., 1985;Lucey & Fox, 1993;Wong et al., 1976). The Ca is converted from insoluble (colloidal form) to soluble form and released in the whey as the pH decreases (Dalgleish & Law, 1989;Guinee et al., 1993;Wong et al., 1976). As a result, the Ca and P were reduced as the pH decreased yielding a low ratio of Ca to P (Kindstedt & Kosikowski, 1988;Lucey & Fox, 1993), which results in low ash content.
The composition of acid whey (a by-product of making the curd from MCC) is illustrated in Table 2. The acid whey produced as a by-product during making the curd showed an average of 5.0% TS, 1.4% protein, 0.9% ash, 0.6% lactose, and 1.4% lactic acid. The total protein as a percentage of total solids in acid whey produced on the lab scale was 28.0%. Results in Table 2 showed that approximately 1.4% of lactose was required to reach a pH of 4.6 in the acid curd/ whey. Using 2.0% lactose MCC left around 0.6% of lactose in the lab scale acid whey. The loss of components in whey, especially protein, depends on the composition of the initial material as well as handling the curd in the cheese vat. The composition of acid whey produced as a by-product of making acid curd using MCC was similar to the acid whey produced from milk in other studies (Chandrapala et al., 2015;Chen et al., 2016;Saffari & Langrish, 2014). Those studies found that the TS of acid whey ranged from 5.0% to 7.0%, while protein and ash ranged from 0.5 to 1.0 and 0.5% to 1.0%, respectively. The protein TA B L E 1 Imitation Mozzarella cheese (IMC) formulations made with acid curd and micellar casein concentrate (MCC). in acid whey could increase to 1.4% as in our study with elevating the protein content in MCC (9.2% protein), which was expected. As a result, solids, ash, and protein content in acid whey can be changed based on the composition of starting material. The lactose content in acid whey was different compared to other studies since we started with 2.0% lactose MCC, not 4.5% lactose as in milk.

| Composition of IMC
The moisture content and pH of IMC are shown in  (Hammam, Beckman, et al., 2022;Kommineni et al., 2012;Salunke, 2013). However, we added 0.5 g of water into each canister (20 g) to compensate for the evaporated water from the cooked cheese (Purna et al., 2006).
The pH of the control IMC was 5.7, while the pH of the experimental IMC made from FR-2:1 was 5.4. Significant differences (p < .05) were detected in the pH of control and FR-2:1 IMC. We previously manufactured PCP with no ES using acid curd and MCC at the same ratio. The pH of the PCP was 5.4 Metzger & Hammam, 2020). The main roles of using ES in PC formulations are calcium sequestration and pH adjustment (Kapoor & Metzger, 2008;Shirashoji et al., 2010). ES sequesters the Ca from the casein network to produce a deaggregated casein network. ES exerted buffering action which could affect the amount of caseinbound calcium and thereby the pH of final PC (Brickley et al., 2008;Shirashoji et al., 2010). This might affect the characteristics of the final PC, as a result, the type of ES and typical amounts should be considered to have the desired PC (Kapoor & Metzger, 2008). The IMC made from FR-2:1 had low pH compared to control. It was found that the pH of IMC can range from 5.4 to 5.8 when ES was utilized (Bi et al., 2016;Jana et al., 2010;Noronha et al., 2008), which is similar to the typical pH of PC or PCP (Bulut-Solak & Akin, 2019; Kapoor & Metzger, 2008;Marchesseau et al., 1997). PC or PCP and IMC have the same basic principles and the same interactions during cooking of the cheese. Palmer and Sly stated that the emulsion stability of PC is poor when the pH is lower than 5.4 or higher than 5.8 (Palmer & Sly, 1943). The differences in pH of control IMC made with ES relative to IMC made from FR-2:1 could affect the structure and quality of final IMC and thereby its functional properties due to its effects on the protein interactions in the final IMC emulsion (Marchesseau et al., 1997;Meyer, 1973;Palmer & Sly, 1943).
As the pH of PC was reduced to 5.2, the protein-protein interaction increased (Marchesseau et al., 1997) because this pH is close to the isoelectric point of caseins (4.6). This induced the aggregation of protein, which, in turn, resulted in a poor emulsion of fat in IMC. On the other hand, the PC had an open structure when the pH was elevated to 6.1, which eventually led to a weaker emulsification (Marchesseau et al., 1997). Marchesseau also found that the pH of 5.7 resulted in PC with more uniform fat emulsion with a closely knit protein network (Marchesseau et al., 1997).

| Functional characteristics of IMC
The end apparent cooked viscosity of IMC is presented in Table 4.
The end apparent cooked viscosity is referred to as the cheese's flowability when completely melted, which is measured at the end of cooking time (Prow & Metzger, 2005 can be explained by the differences in pH of IMC. As the pH drops to the isoelectric point, the net negative charges on caseins reduce which increases the protein-protein interactions and this led to aggregation of protein and thereby poor emulsification (Kapoor, 2007).
The higher pH in control IMC resulted in a uniform fat emulsion with a closely knit protein network. This led to a higher melting temperature of IMC made with ES relative to that of IMC made from FR-2:1. As the pH elevated, the net negative charges on casein micelles increased, which promoted the calcium-mediated cross-links casein molecules, and this, in turn, strengthened the IMC gel network. Increasing the strength of IMC gel network led to restricting the movement of casein chains in IMC during reheating, which decreased the flowability, melt diameter, and melt area of IMC with higher pH made with ES (Kapoor, 2007). This phenomenon was not pronounced in the melting temperature and stretchability of IMC, but it was more differentiating in the melt diameter and melt area.

| CON CLUS IONS
Culture-based acid curd was produced from liquid MCC (>9% TPr and >13% TS). Acid curd and MCC can be mixed in a specific ratio and marketed to be ready for making different types of cheeses

ACK N OWLED G M ENTS
The authors acknowledge and thank the Midwest Dairy Foods Research Center (St. Paul, MN) for their financial support.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.