J. Dairy Sci. 9 4 :5278–5288 doi:1 0.3168/jds.2010-3932 © American Dairy Science Association®, 2 011 . Effect of increasing the colloidal calcium phosphate of milk on the texture and microstructure of yogurt T . O zcan, *† D. H orne ,* and J . A. L ucey *‡1 * Department of Food Science, University of Wisconsin-Madison, Madison 53706-1565 † Department of Food Engineering, Uludag University, 16059 Gorukle, Bursa, Turkey ‡ Wisconsin Center for Dairy Research, Madison 53706-1565 A BSTRACT decreased. Loss tangent values at pH 5.1 were similar in all samples except in gels with a CCP level of 128%. For The effect of increasing the colloidal calcium phos- dialyzed milk, the whey separation levels were similar phate (CCP) content on the physical, rheological, and in gels made from milk with up to 107% CCP but in- microstructural properties of yogurt was investigated. creased at higher CCP levels. Microstructure of yogurt The CCP content of heated (85°C for 30 min) milk was gels made from milk with 100 to 107% CCP was similar increased by increasing the pH by the addition of alkali but very large clusters were observed in gels made from (NaOH). Alkalized milk was dialyzed against pasteur- milk with higher CCP levels. By dialyzing heated milk ized skim milk at approximately 4°C for 72 h to attempt against pasteurized milk, we may have retained some to restore the original pH and soluble Ca content. By heat-induced Ca phosphate on micelles that normally adjustment of the milk to pH values 7.45, 8.84, 10.06, dissolves on cooling because, during dialysis, pasteur- and 10.73, the CCP content was increased to approxi- ized milk provided soluble Ca ions to the heated milk mately 107, 116, 123, and 128%, respectively, relative system. Yogurt texture was significantly affected by to the concentration in heated milk. During fermenta- increasing the casein-bound Ca (and total Ca) content tion of milk, the storage modulus (G ) and loss tangent of milk as well as by the alkalization procedure involved values of yogurts were measured using dynamic oscilla- in that approach. tory rheology. Large deformation rheological properties Key words: y ogurt, c olloidal calcium phosphate, rhe- were also measured. The microstructure of yogurt was ology , m icrostructure observed using fluorescence microscopy, and whey sepa- ration was determined. Acid-base titration was used to evaluate changes in the CCP content in milk. Total Ca I NTRODUCTION and casein-bound Ca increased with an increase in the Yogurt is a semisolid dairy product made by fer- pH value of alkalization. During acidification, elevated mentation of milk with Streptococcus thermophilus and buffering occurred in milk between pH values 6.7 to Lactobacillus delbrueckii ssp. bulgaricus cultures. Ag- 5.2 with an increase in the pH of alkalization. When gregation and gelation of CN occurs due to the reduc- acidified milk was titrated with alkali, elevated buffer- tion in charge repulsion with the decrease in milk pH ing occurred in milk between pH values 5.6 to 6.4 with (Lucey and Singh, 1997). Within the CN micelles, CN an increase in the pH of alkalization. The high residual molecules are held together primarily by hydrophobic pH of milk after dialysis could be responsible for the interactions and (insoluble or casein-bound) colloidal decreased contents of soluble Ca in these milks. The pH calcium phosphate (CCP) crosslinks (Horne, 1998; Fox of gelation was higher in all dialyzed samples compared and Brodkorb, 2008). These CCP crosslinks are dis- with the heated control milk, and the gelation pH was solved with a decrease in milk pH (Pyne and McGann, higher with an increase in CCP content. The sample 1960) and caseins are liberated into the serum phase with highest CCP content (128%) exhibited gelation (Dalgleish and Law, 1989). The extent of liberation of at very high pH (6.3), which could be due to alkali- caseins depends on the temperature at acidification; at induced CN micellar disruption. The G values at pH 30°C, a decrease in pH causes virtually no liberation 4.6 were similar in gels with CCP levels up to 116%; of protein (Dalgleish and Law, 1989). Thus, during yo- at higher CCP levels, the G values at pH 4.6 greatly gurt fermentation, which is performed at temperatures >30°C, no dissociation of CN likely occurs. M any factors influence the texture and physical Received October 14, 2010. Accepted August 7, 2011. properties of yogurt gels, including heat treatment 1 Corresponding author: j alucey@facstaff.wisc.edu (Dannenberg and Kessler, 1988; van Vliet and Keetels, 5278 COLLOIDAL CALCIUM PHOSPHATE AND YOGURT TEXTURE AND MICROSTRUCTURE 5279 1995), incubation temperature (Lee and Lucey, 2004), MATERIALS AND METHODS rate of acidification (Horne, 2003; Anema, 2008), and fortification with milk proteins (Sodini et al., 2004). Materials Ozcan-Yilsay et al. (2007) used a different approach Low-heat skim milk powder with a whey undena- to alter yogurt texture. They added low concentra- tured protein nitrogen index of 6.60 mg/g (Bradley et tions of trisodium citrate to milk to reduce the level al., 1992) was supplied by Dairy Farmers of America of CCP cross-linking between CN. Higher gel stiffness (Fresno, CA). Yogurt starter culture (Streptococcus and decreased whey separation were observed in yogurt thermophilus and Lactobacillus delbrueckii ssp. bul- when low levels of TSC were used. Ozcan-Yilsay et al. garicus, YC-087) was obtained from Chr. Hansen Inc. (2007) suggested that low levels of CCP removal facili- (Milwaukee, WI). Fresh pasteurized cow’s milk was tated greater rearrangement and molecular mobility of obtained from the UW-Madison Food Science Dairy the micelle structure, which may have helped increase Plant. the formation of crosslinks between strands in yogurt gel networks. In contrast, when most of the CCP was Milk Preparation and Dialysis dissolved, complete micelle disruption occurred, which caused the formation of very weak yogurt gels. To increase the CCP content of milk, we followed McGann and Pyne (1960) developed a method to the method described by McGann and Pyne (1960). increase the CCP content of milk. The pH value of Reconstituted skim milk (10.7% wt/vol) was preheated cold (~0°C) milk is increased by the addition of base at 85°C for 30 min in a thermostatically controlled wa- (NaOH). The increased pH alters the Ca equilibrium terbath and then cooled rapidly with ice water to 0 to and a shift occurs from soluble to insoluble (casein- 2°C. Immediately after cooling, 2 N NaOH was added bound) Ca. Dialysis of the pH-adjusted milk against to milk to reach concentrations of 0, 0.5, 1, 1.5, and 2% bulk milk restores the soluble Ca content but the CCP (vol/vol) by the slow addition of alkali with continuous content remains high and the overall Ca content is in- stirring at approximately 0°C. Milks were stirred for 1 creased (McGann and Pyne, 1960). This method has h and the pH was then recorded at approximately 0°C. been used in a variety of studies to examine the effect The pH values of the samples containing 0, 0.5, 1, of the CCP concentration on milk properties such as 1.5, and 2% (vol/vol) NaOH were 6.68, 7.45, 8.84, heat stability (Fox and Hoynes, 1975; Singh and Fox, 10.06, and 10.73, respectively. The resulting mixtures 1987). Recently, Anema (2009) used the method of were dialyzed using a dialysis membrane with a mo- McGann and Pyne (1960) to alter the CCP content of lecular weight cut-off of 6 to 8 kDa. One liter of each milk and studied the effect of CCP on acid gels made treated milk sample was dialyzed against 10 L of fresh by the addition of glucono-δ-lactone (GDL). Anema pasteurized skim milk with regular changes of milk. (2009) reported that increasing the CCP content of Total dialysis time was 72 h, with around 10 changes milk resulted in a slight increase in the gelation pH of milk during this period. The bags were left to reach of GDL-induced gels but had no consistent effect on equilibrium at 0 to 5°C and stirred for 72 h. By the end storage modulus (G ) values at pH 4.6. The properties of this dialysis period, the pH values of all milks had of model acid gels made with GDL have been reported decreased. At the end of the dialysis procedure, the pH to differ greatly from yogurt gels made by bacterial values of the samples containing 0, 0.5, 1, 1.5, and 2% fermentation (Lucey et al., 1998b), probably because of NaOH (vol/vol) were 6.69, 6.75, 6.86, 6.92, and 6.93, the markedly different rate of acidifications with these respectively. 2 approaches (some proteolysis may also occur in fer- Some milk samples were used for chemical analysis; mented milks). Anema (2009) altered the CCP content 0.02% (wt/wt) sodium azide was added to prevent bac- of milk before high heat treatment (80°C for 30 min), terial growth in these samples. For yogurt fermentation, but it is known that the CCP content of milk influences starter culture was prepared according to the method heat stability (Singh and Fox, 1987) so we heated the described by Ozcan et al. (2008). Before the addition milk before altering the CCP content. At very high of culture, milk was rewarmed at 60°C for 30 min to temperatures (i.e., >110°C), CCP can be involved in try to restore CN or milk salt changes induced by the cross-linking of proteins, resulting in the formation of cold storage used for the dialysis procedure. Milk were protein aggregates (Singh, 1994). We are not aware of cooled to 42°C and inoculated with 2% (wt/wt) work- any published study on the properties of yogurts made ing culture. The pH was recorded every 5 min during from milk with an elevated CCP content. The present fermentation as described by Ozcan et al. (2008). The study was performed to investigate the effect of increas- initial acidification rate (i.e., for a 1.0 pH change from ing the CCP content of milk on yogurt gelation. the original starting pH) was determined from the pH Journal of Dairy Science Vol. 94 No. 11, 2011 5280 OZCAN ET AL. profiles and expressed as pH milliunits per minute (mU/min; Lee and Lucey, 2004). Acid-Base Buffering Properties Buffering curves of milks were determined by the acid-base titration method described by Lucey et al. (1993b). The area between the acid and base titration curves was used to estimate the CCP content of milk (Lucey et al., 1993b, 1996). The area under the base titration curve from pH 4.2 to the original milk pH was subtracted from the area under the acid titration curve from the pH of the original milk to pH 4.2. Buffering areas for the treated samples were compared with the heated control sample, which was set as 100% (Table 1). We assumed that changes in the buffering capacity as a result of alkalization were due to alterations in the CCP content (i.e., the alkalization procedure did not influence buffering from protein side chains). Ultrafiltration and Ca Determinations A Prep/Scale-TFF membrane (Millipore, Billerica, MA), which was made from regenerated cellulose and had a molecular weight cut-off of 10 kDa, was used to obtain UF permeates of milks. The total Ca concentra- tion of milk and soluble Ca contents of UF permeate were determined using inductively coupled plasma-op- tical emission spectrometer (Vista-MPX Simultaneous ICP-OES, Varian Inc., Palo Alto, CA). The wavelength of plasma emission used to measure the Ca content was 317.9 nm (Park, 2000). Casein-bound Ca was calculated using the following equation (White and Davies, 1958): Casein-bound Ca = total Ca – Ca in UF permeate. Yogurt Properties Yogurt gel formation was determined by dynamic low amplitude oscillatory rheometry (Paar Physica UDS 200 controlled stress rheometer, Physica Messtechnik GmbH, Stuttgart, Germany) as described by Ozcan et al. (2008). Whey separation values of yogurts were determined using the method described by Lucey et al. (1998a). The fluorescence microscopy method de- scribed by Choi et al. (2007) was followed to examine the microstructure of yogurt gels. Statistical Analysis Statistical analysis was conducted by ANOVA using the statistical software SAS (version 9.1, SAS Institute Inc., Cary, NC). Experiments were replicated at least 3 times. Fisher’s least significant difference test was Journal of Dairy Science Vol. 94 No. 11, 2011 Table 1. Effects of increasing the colloidal calcium phosphate (CCP) content on the rheological and physical properties of milk and yogurt1 Ca (mg/100 g) pH after Loss tangent G value2 Yield Whey alkali pH after Casein- Gelation pH at value at pH 4.6 stress3 Yield separation3 Sample addition dialysis Soluble bound time (min) gelation pH 5.1 (Pa) (Pa) strain3 (%) Heated milk — — 28.4ab 86.0b 138d 5.34d 0.55a 143ab 48a 0.53a 4.13c Milks with CCP contents (%)4 of: 100 6.68e 6.69d 31.3a 88.4b 145c 5.65c 0.54a 135ab 39ab 0.49ab 3.38d 107 7.45d 6.75c 30.1a 93.8b 155b 5.76b 0.51b 128bc 35b 0.44b 4.64b 116 8.84c 6.86b 25.8b 103.7a 156b 5.78b 0.50b 151a 34b 0.51a 5.60a 123 10.06b 6.92a 21.5c 113.3a 166a 5.79b 0.51b 114c 18c 0.38c 5.90a 128 10.73a 6.93a ND5 ND 145c 6.34a 0.30c 11d 3d 0.45b ND a–eValues with different letters within the same column are significantly different (P < 0.05). 1Values are means of triplicates. 2G = storage modulus. 3These properties were determined when the pH of yogurt gels reached 4.6. 4Heated milk was alkalized and dialyzed to increase the CCP content, the CCP content was determined from the area between the acid and base titration curves in the pH range from the initial milk pH to 4.2. 5ND = not determined. COLLOIDAL CALCIUM PHOSPHATE AND YOGURT TEXTURE AND MICROSTRUCTURE 5281 carried out to evaluate differences in treatment means fermentation compared with the heated control milk at a significance level of P < 0.05. (not dialyzed). For the remainder of the fermentation process, the slopes of the pH profiles appeared similar. RESULTS The initial rate of acidification was 7.2, 6.4, 5.8, 5.9, 6.0, and 5.1 mU/min for the heated control milk (not dia- Acid-Base Buffering Properties lyzed) and samples with CCP contents of 100, 107, 116, 123, and 128%, respectively. Because the heated milk The acid-base buffering properties of milks are shown not dialyzed and heated milk dialyzed (100% CCP) had in Figure 1. All milk samples exhibited a buffering peak similar buffering profiles (Figure 1), this difference in at around pH 5.0 during acid titration, which was due initial acid development was related to an alteration to the solubilization of CCP (Figure 1a; Lucey et al., in bacterial fermentation. An increase was observed in 1993b). Higher buffering was observed between pH val- the initial pH value of dialyzed milks compared with ues 6.7 to 5.2 in milk with an increase in the pH of alka- the heated control milk, which could have contributed lization. The buffering observed during the titration of to slower bacterial fermentation. During the prolonged acidified milks with base is shown in Figure 1b. During dialysis procedure, some change to minor milk con- the back-titration of milk with base, a buffering peak stituents may have occurred that contributed to the was observed at around pH 6.0 due to formation of slower bacterial fermentation of the dialyzed sample insoluble Ca phosphate (Lucey et al., 1993b). Greater without alkalization (100% CCP) compared with the buffering was observed between pH values 5.5 to 6.4 in control milk. The pH values of milks at a fermentation milk with an increase in the pH of alkalization. During time of 150 min increased in the following order: 128% the titration of acidified milk with base, the buffering CCP >123% CCP = 116% CCP >107% CCP >100% peak was shifted to slightly lower pH values with an CCP > heated control milk. This trend agreed with increase in the pH of alkalization. the increased buffering in samples with an increase in the CCP content (Figure 1). The time to reach pH 4.6 Soluble and Casein-Bound Ca increased as follows: 116% CCP = 123% CCP = 128% CCP >107% CCP >100% CCP > heated control milk. The effect of alkalization of milk followed by dialysis on the soluble and casein-bound Ca content is shown Rheological Properties in Table 1. Casein-bound Ca increased with an increase in the pH of alkalization. The soluble Ca content of the The effects of increasing the CCP content on the dialyzed sample without alkalization was slightly (but rheological and physical properties of milk and yogurt not significantly) higher than that of the heated control are summarized in Table 1. All dialyzed samples had milk. Milks with a CCP content of 116% exhibited a de- longer gelation times than the heated control milk even crease in soluble Ca content compared with the dialyzed though the dialyzed samples had a higher gelation pH control or milk with CCP content of 107%. Increasing than the heated control milk. For the dialyzed samples, the CCP content of milk to 123% resulted in a further gelation time increased with an increase in the pH of significant decrease in the soluble Ca content. Milks alkalization until samples containing 128% CCP when with a CCP content of 128% were difficult to ultrafilter the gelation time decreased again. Samples containing as these samples became translucent with some visual 107, 116, and 123% CCP had similar pH values at ge- precipitation. We did not determine the casein-bound lation and the pH value at gelation was higher than Ca for the milk with a CCP content of 128%. Milks the dialyzed control milk. The pH of gelation for the with CCP contents of 116 and 123% had significantly sample containing 128% CCP was very high (6.3). elevated casein-bound Ca contents compared with the The effects of increasing the CCP content of milk on dialyzed control or the milk with a CCP content of the G and loss tangent (LT) values of yogurt as a func- 107%. Total Ca levels also exhibited a significant in- tion of pH are shown in Figure 3. The heated control crease with an increase in the pH of alkalization. milk gelled at low pH values (~5.34) and the G values were lower than that of the other samples until around pH Profiles pH 4.7, when they increased sharply (Figure 3a). Af- ter gelation, the G values for the samples containing The pH profiles during acidification with 2% (wt/ 107 and 116% CCP were higher than those of samples wt) starter culture at 42°C are shown in Figure 2. The containing 100 and 123% CCP. The sample containing pH profile of the dialyzed sample without alkalization 128% CCP exhibited an unusual G profile; although exhibited a slower rate of pH change from approxi- apparent gelation occurred at a high pH value, no sub- mately 100 min after culture addition until the end of stantial increase in the G values was observed during Journal of Dairy Science Vol. 94 No. 11, 2011 5282 OZCAN ET AL. Figure 1. Acid-base buffering curves of milks: (a) titration of milk from initial pH to pH 3.0 with 0.5 N HCl, and (b) back-titration of acidi- fied milk from pH 3.0 to pH 9.0 with 0.5 N NaOH. Heated (85°C for 30 min) milk ( ) and dialyzed milk containing 100% (), 107% (), 116% (□), 123% (), and 128% () colloidal calcium phosphate compared with the heated milk. fermentation and the G values at pH 4.6 were very low LT values decreased slowly for the yogurt made from (~11 Pa). milk containing 128% CCP. The LT profiles for yogurt gels are shown in Fig- The yield stress values of yogurt gels decreased with ure 3b. Apart from the gel made from milk containing an increase in the pH of alkalization with a sharp re- 128% CCP, all samples exhibited a maximum in LT duction in yield stress values observed in samples con- during acidification. The pH value of the LT maximum taining more than 116% CCP (Table 1). No consistent in the heated control milk sample was slightly lower trends were observed for the yield strain values of gels compared with that of the dialyzed samples. The LT made from the heated control or dialyzed samples. values for yogurt made from milk containing 128% CCP decreased to <0.5 at pH values >6.2, in agreement with Whey Separation and Microstructure the very high gelation pH of this sample (Table 1). The whey separation levels of gels made from the During the remainder of the fermentation process, the dialyzed sample without alkalization were lower than Journal of Dairy Science Vol. 94 No. 11, 2011 COLLOIDAL CALCIUM PHOSPHATE AND YOGURT TEXTURE AND MICROSTRUCTURE 5283 Figure 2. pH profiles as function of time for yogurts made with 2% (wt/wt) starter culture at 42°C from heated (85°C for 30 min) milk ( ) and dialyzed milk containing 100% (), 107% (), 116% (□), 123% (), and 128% () colloidal calcium phosphate compared with the heated milk. that of the heated control milk. Whey separation levels bound Ca (and increased total Ca content). This pro- in yogurt gels increased with an increase in the pH cedure altered both milk pH and mineral equilibria, of alkalization, and thus with an increase in the CCP which greatly affected the properties of CN micelles level. Weak gels were formed from milk containing and the yogurt gels made from these treated milks. 128% CCP and no whey separation could be measured. The addition of alkali to milk (without dialysis against The microstructures of yogurt samples are shown untreated milk) results in a decrease in soluble Ca and in Figure 4. Similar types of microstructures were ob- inorganic phosphate (van Dijk, 1992; Vaia et al., 2006). served in gels made from heated milk (Figure 4a) and Vaia et al. (2006) reported that adjustment of milk to those made from dialyzed milk containing 100% CCP pH 9 and 10 decreased the soluble Ca levels to approxi- (Figure 4b). The gels containing 107% (Figure 4c) and mately 20 and 3%, respectively, compared with the 116% CCP (Figure 4d) exhibited larger clusters than soluble Ca levels in untreated milk. The alkalization- the heated milk or 100% CCP sample. The network in dialysis procedure has been used previously to increase yogurts made from heated milk exhibited small clusters the CCP content of milk (McGann and Pyne, 1960; Fox (<5 μm) with extensive branching. Yogurts made from and Hoynes, 1975; Singh and Fox, 1987; Anema, 2009). milk containing 123% CCP exhibited very large, dense In all previous studies, this procedure was applied to (many greater than 50 μm) protein clusters (Figure either raw or low-heat-treated milk, whereas we used 4e). Large pores (>20 μm) could also be observed in severely heated milk (85°C for 30 min). Heat treatment this network. Yogurts made from milk containing 128% of milk results in an increase in casein-bound Ca and CCP did not exhibit a characteristic cross-linked net- a concomitant decrease in soluble Ca (Walstra and work but instead very large protein agglomerates could Jenness, 1984; de la Fuente, 1998); these changes are be observed (Figure 4f). reversed upon cooling (Pierre and Brulé, 1981; Pouliot et al., 1989). Dialysis of heated milk that was not alka- DISCUSSION lized (100% dialyzed sample) against low-heat-treated In our alkalization-dialysis procedure, milks were milk helped to maintain a high level of casein-bound Ca subjected to increasing pH and an elevation in casein- (Table 1) because the low-heat-treated milk provided Journal of Dairy Science Vol. 94 No. 11, 2011 5284 OZCAN ET AL. Figure 3. (a) Storage modulus (G ) and (b) loss tangent (LT) as function of pH for yogurts made from heated (85°C for 30 min) milk ( ) and dialyzed milk containing 100% (), 107% (), 116% (□), 123% (), and 128% () colloidal calcium phosphate compared with the heated milk. Yogurts were made at 42°C using 2% (wt/wt) starter culture. Journal of Dairy Science Vol. 94 No. 11, 2011 COLLOIDAL CALCIUM PHOSPHATE AND YOGURT TEXTURE AND MICROSTRUCTURE 5285 Figure 4. Microstructure of yogurt gels made from heated (85°C for 30 min) milk (a) and dialyzed milk containing 100% (b), 107% (c), 116% (d), 123% (e), and 128% (f) colloidal calcium phosphate compared with the heated milk. Yogurts were made at 42°C using 2% (wt/wt) starter culture. The protein matrix is white and pores are dark; scale bar = 20 μm. soluble Ca during dialysis and decreased the tendency sters, which would have to grow in size or number, or of the heated milk to dissolve its heat-induced CCP. binding of ionic Ca to phosphoserine residues that were Heated milk was dialyzed immediately after cooling. A not involved in stabilizing nanoclusters. We believe similar type of principle is involved in the alkalization- that alkalization resulted in an increase in insoluble dialysis procedure of McGann and Pyne (1960), except Ca phosphate content because we observed an increase that alkalization (instead of heat treatment) is used to in the buffering area between the initial pH of treated increase the casein-bound Ca level and dialysis against milk and pH 4.2 (Figure 1), which is indicative of an a milk with a higher soluble Ca level helps restore the increased concentration of insoluble Ca phosphate (Lu- soluble Ca level while maintaining the elevated CCP cey et al., 1993b, 1996). The acid-base buffering profiles content caused by the alkalization step. of CCP-enriched milk (Figure 1) appeared generally When milk was alkalized to very high pH values, ex- similar to those observed previously for milk (Lucey et tensive (72 h) dialysis was not sufficient to completely al., 1993b). The CCP-enriched milks created by the al- restore the original milk pH value. It could be that the kalization procedure did exhibit some minor differences additional (Ca) phosphate in milk may be increasing in buffering behavior, including higher buffering during the “natural” pH of milk. Exhaustive dialysis for 72 h acidification at pH values 6.7 to 5.2 compared with with up to 10 changes in milk used for dialysis should that of heated milk. Because alkalization also resulted have been ample time to attain equilibrium. We did not in an increase in the total Ca content in milk, it is prolong the dialysis time any further due to concerns likely that the pH solubility of CCP would be modified about possible proteolysis or cold aging affects that as well (which could affect the buffering properties). could negatively influence yogurt gelation. From our buffering profiles, we believe that alkaliza- Heat treatment of milk creates heat-induced Ca tion produces CCP that is mostly similar to the native phosphate, which exhibits similar pH solubilization CCP in milk. It is possible that alkalization induced a behavior compared with the original CCP, unless milk change in the type of CCP in milk but different forms is subjected to severe heating conditions; for example, of CCP have unique buffering behaviors (Upreti et al., 120°C for 10 min (Lucey et al., 1993a). Vaia et al. 2006) and we observed only a minor modification to the (2006) noted that the nature of the increase in casein- buffering profiles. bound Ca caused by alkalization of milk was not known Anema (2009) adjusted the CCP content of unheated but several options existed: forming insoluble (casein- milk by acidification/alkalization followed by dialysis bound) Ca phosphate and becoming part of nanoclu- against the original unheated milk. After alteration of Journal of Dairy Science Vol. 94 No. 11, 2011 5286 OZCAN ET AL. the CCP content, milks were heated and then acidified 1981; Gastaldi et al., 1994; Philippe et al., 2003). It with GDL. In our study, we first heated the milk before is also not clear where the additional CCP is in the adjusting the CCP content, because adjustment of the micelles. Does it form new nanoclusters or larger nano- CCP content is known to greatly alter the heat stability clusters? of milk (Singh and Fox, 1987). Anema (2009) also var- We cannot have a fixed size and functionality for ied the GDL levels to obtain similar overall acidification the nanoclusters in native CN micelles, as suggested times in samples, whereas we added a constant level by McMahon and Oommen (2008), because this would of starter culture to all milks. Anema (2009) observed imply the creation of more (new) nanoclusters and re- only a slight increase (<0.1 pH unit) in the gelation quire, for the same reason, more (new) phosphoserine pH when the CCP content of milk was increased to clusters and more (new) caseins in the alkalized milk. 115%. It is possible that the larger change in gelation However, we cannot have additional caseins because pH obtained in our CCP-enriched milk, compared with additional casein does not pass the dialysis barrier. We that of Anema (2009), could be due to the use of heated propose that when the pH is increased, the serine phos- milk for alkalization-dialysis and the different method phates become more negatively charged and less favor- of acidification in our study, which was by fermentation ably inclined to associate with the Ca phosphate, which of lactose by starter culture. We observed that the gela- has 2 outcomes. First, the micelle tends to dissociate tion pH increased in alkalized-dialyzed milks, but these (becomes less turbid) and this tendency increases the milks also had longer gelation times probably because higher the pH value of the milk (especially at pH >9; of their high buffering (Figure 1). Addition of Ca to Odagiri and Nickerson, 1965; Thompson and Farrell, milk has been reported to increase the hydrophobicity 1973; Vaia et al., 2006; Huppertz et al., 2008).The sec- of micelles (Philippe et al., 2003) and decrease the zeta ond outcome of alkalization is that Ca phosphate tends potential (Dalgleish, 1984; Philippe et al., 2003), which to precipitate at high pH values. If this treated milk is could help to increase the gelation pH. dialyzed against a milk of normal pH, the pH is brought We did not observe any significant effect on G values back down and the Ca binding activity of the serine at pH 4.6 when the CCP content was increased, at least phosphates are restored. However, if the pH decrease up to 116% (Table 1). Anema (2009) also reported that is slow, the Ca phosphate nanoclusters may grow and elevating the CCP content of milk by 115% had little thus more CCP is incorporated into the CN micelle. effect on the G values at pH 4.6 for GDL-induced gels. What does this do for the properties of the micelle? It Ozcan-Yilsay et al. (2007) suggested that low levels of is likely that the incorporation will take time, as will CCP removal facilitated greater rearrangement of the reassembly of the micelle following the (partial) dis- micelle structure, which helped to increase the forma- sociation on increasing the pH value. The re-formation tion of crosslinks between strands in yogurt gel net- of CN micelles after alkali-induced dissociation has works and produced gels with higher G values at pH been studied and it appears that these micelles largely 4.6. In our study, the additional CCP did not result in resemble native micelles (Huppertz et al., 2008). a decrease in the G values at pH 4.6, possibly because At the highest CCP level (128%), unusual gelation of the high gelation pH observed for these samples com- behavior was observed, including very high gelation pH pared with that of the heated control milk. values, weak gels, and an agglomerated, poorly cross- Removal of CCP from CN micelles results in acid linked yogurt gel network. To attain such a high CCP gels with higher LT at pH 5.1 (Ozcan-Yilsay et al., level, milk had to be adjusted to pH 10.7. Beeby and 2007; Anema, 2009; Famelart et al., 2009). The higher coworkers (Beeby and Kumetat, 1959; Beeby and Lee, LT values reflect the greater mobility of CN in the net- 1959) reported that viscosity greatly increased when work due to the loss of CCP crosslinks. We observed milk was adjusted to pH values around 11. McGann significantly lower LT values at a pH of approximately and Pyne (1960) also noted a large increase in viscosity 5.1 with an increase in the CCP content compared when the CCP of milk was increased by >120%. Hemar with gels made from the heated control milk (Table et al. (2000) observed that milk could undergo alkaline- 1). Anema (2009) reported that for GDL-induced gels, induced gelation at pH 12. Vaia et al. (2006) proposed the LT values from the gelation point until around pH that alkaline-induced disruption of CN micelles was not 5.1 were lower for the samples containing 115% CCP only related to charge but that increasing the milk pH compared with gels made with 100% CCP. improves the solvent quality for the caseins, thereby It is not clear how milk can be enriched with ad- leading to the disruption of CN micelles into their ditional CCP compared with the native CN micelles. constituent nanoclusters. It appeared that extensive Several studies have indicated that when Ca is added dissociation of micelles occurred at very high pH values to milk some of the added Ca becomes associated with because the samples were visually translucent. The CN micelles, presumably as CCP (Brule and Fauquant, micelle system may remain destabilized and partially Journal of Dairy Science Vol. 94 No. 11, 2011 COLLOIDAL CALCIUM PHOSPHATE AND YOGURT TEXTURE AND MICROSTRUCTURE 5287 aggregated even after dialysis, and when the pH is sub- Beeby, R., and K. Kumetat. 1959. Viscosity changes in concentrated sequently lowered during fermentation the samples with skim-milk treated with alkali, urea and calcium complexing agents. I. The importance of casein micelle. J. Dairy Res. 26:248–257. high CCP levels gelled at a very high pH value (Table Beeby, R., and J. W. Lee. 1959. Viscosity changes in concentrated milk 1). Ozcan-Yilsay et al. (2007) reported that complete treated with alkali, urea and calcium-complexing agents. II. The disruption of CN micelles resulted in the formation of influence of concentration, temperature and rate of shear. J. Dairy Res. 26:258–264. very weak yogurt gels from this treated milk. Bradley, R. L., E. Arnold, D. M. Barbano, R. G. Semerad, D. E. Smith, and B. K. Vines. 1992. Chemical and physical methods. Pages 433–531 in Standard Methods for the Examination of Dairy CONCLUSIONS Products. R. T. Marshall, ed. 16th ed. American Public Health Association, Washington, DC. Alkalization of heated (85°C for 30 min) milk and Brule, G., and J. Fauquant. 1981. Mineral balance in skim-milk and dialysis against pasteurized skim milk resulted in an in- milk retentate: Effect of physicochemical characteristics of the aqueous phase. J. Dairy Res. 48:91–97. crease in the concentrations of casein-bound and total Choi, J., D. S. Horne, and J. A. Lucey. 2007. Effect of insoluble cal- Ca. One possible explanation for this result could be cium concentration on rennet coagulation properties of milk. J. the growth of CCP nanoclusters. With the increase in Dairy Sci. 90:2612–2623.Dalgleish, D. G. 1984. Measurement of electrophoretic mobilities and milk pH, serine phosphate groups become more nega- zeta-potentials of particles from milk using laser Doppler electro- tively charged, which could weaken their interaction phoresis. J. Dairy Res. 51:425–438. with the CCP nanoclusters. The high pH favors the Dalgleish, D. G., and A. J. R. Law. 1989. pH induced dissociation of bovine casein micelles. II. Mineral solubilization and its relation to precipitation of additional CCP. Dialysis of alkalized casein release. J. Dairy Res. 56:727–735. milk should restore the binding activity of the serine Dannenberg, F., and H. G. Kessler. 1988. Effect of denaturation of phosphates and this could allow for growth of the ex- β-lactoglobulin on texture properties of set-style nonfat yoghurt. 1. Syneresis. Milchwissenschaft 43:700–704. isting nanoclusters. The pH of gelation was higher in de la Fuente, M. A. 1998. Changes in the mineral balance of milk all dialyzed samples compared with that of the heated submitted to technological treatments. Trends Food Sci. Technol. control milk. By dialyzing heated milk against pasteur- 9:281–288.Famelart, M.-H., G. Gauvin, D. Paquet, and G. Brule. 2009. Acid ized milk, we could have retained some heat-induced gelation of colloidal calcium phosphate-depleted preheated milk. Ca phosphate on micelles that normally dissolves on J. Dairy Sci. Technol. 89:335–348. cooling, because the pasteurized milk provided soluble Fox, P. F., and A. Brodkorb. 2008. The casein micelle: Historical as-pects, current concepts and significance. Int. Dairy J. 18:677–684. Ca ions to the heated milk system, thereby reducing Fox, P. F., and M. C. T. Hoynes. 1975. Heat stability of milk: Influ- the driving force to reverse this heat-induced shift in ence of colloidal calcium phosphate and β-lactoglobulin. J. Dairy the Ca equilibrium. Increasing the CCP (and total Ca) Res. 42:427–435.Gastaldi, E., O. Pellegrini, A. Lagaude, and B. Tarodo de al Fuente. content of milk did not greatly affect the G values at 1994. Functions of added calcium in acid milk coagulation. J. pH 4.6, LT values, or gel microstructure until the CCP Food Sci. 59:310–312., 320. content exceeded 107%. Alkalization of milk to high Hemar, Y., A. J. R. Law, D. S. Horne, and J. Leaver. 2000. Rheo-logical investigations of alkaline-induced gelation of skimmed milk pH values is known to cause dissociation of micelles. and reconstituted skimmed milk concentrates. Food Hydrocoll. When milks were alkalized to very high pH values 14:197–201. (10.7) before dialysis, the gels formed during yogurt Horne, D. S. 1998. Casein interactions: Casting light on the black boxes, the structure in dairy products. Int. Dairy J. 8:171–177. fermentation were very weak and the gelation pH was Horne, D. S. 2003. Casein micelles as hard spheres: Limitations of the very high. To generate high CCP levels in heated milk model in acidified gel formation. Colloids Surf. A Physicochem. required the use of high alkalization pH values, which Eng. Asp. 213:255–263.Huppertz, T., B. Vaia, and M. A. Smiddy. 2008. Reformation of casein negatively affected yogurt gelation properties. particles from alkaline-disrupted casein micelles. J. Dairy Res. 75:44–47. Lee, W. J., and J. A. Lucey. 2004. Structure and physical properties of ACKNOWLEDGMENTS yogurt gels: Effect of inoculation rate and incubation temperature. J. Dairy Sci. 87:3153–3164. This work was supported by The Commission of Sci- Lucey, J. A., C. Gorry, and P. F. Fox. 1993a. Acid-base buffering prop- entific Research Projects of Uludag University (Bursa, erties of heated milk. Milchwissenschaft 48:438–441.Lucey, J. A., C. Gorry, B. O’Kennedy, M. Kalab, R. Tan-Kinita, and Turkey; project number: YDP (Z)-2010/6) and Univer- P. F. Fox. 1996. Effect of acidification and neutralization of milk sity of Wisconsin-Madison. on some properties of casein micelles. Int. Dairy J. 6:257–272. Lucey, J. A., B. Hauth, C. Gorry, and P. F. Fox. 1993b. The acid-base buffering properties of milk. Milchwissenschaft 48:268–272. REFERENCES Lucey, J. A., P. A. Munro, and H. Singh. 1998a. Whey separation in acid skim milk gels made with glucono-δ-lactone: Effects of heat Anema, S. G. 2008. Effect of temperature and rate of acidification on treatment and gelation temperature. J. Texture Stud. 29:413–426. the rheological properties of acid skim milk gels. J. Food Proc. Lucey, J. A., and H. Singh. 1997. Formation and physical properties of Preserv. 32:1016–1033. acid milk gels: A review. Food Res. Int. 30:529–542. Anema, S. G. 2009. Role of colloidal calcium phosphate in the acid ge- Lucey, J. A., M. Tamehana, H. Singh, and P. A. Munro. 1998b. A com- lation properties of heated skim milk. Food Chem. 114:161–167. parison of the formation, rheological properties and microstructure Journal of Dairy Science Vol. 94 No. 11, 2011 5288 OZCAN ET AL. of acid skim milk gels made with a bacterial culture or glucono-δ- Singh, H., and P. F. Fox. 1987. Heat stability of milk: Influence of col- lactone. Food Res. Int. 31:147–155. loidal and soluble salts and protein modification on the pH-depen- McGann, T. C. A., and G. T. Pyne. 1960. The colloidal phosphate dent dissociation of micellar κ-casein. J. Dairy Res. 54:523–534. of milk. III. Nature of its association with casein. J. Dairy Res. Sodini, I., F. Remeuf, S. Haddad, and G. Corrieu. 2004. The relative 27:403–417. effect of milk base, starter, and process on yogurt texture: A re- McMahon, D. J., and B. S. Oommen. 2008. Supramolecular structure view. Crit. Rev. Food Sci. Nutr. 44:113–137. of the casein micelle. J. Dairy Sci. 91:1709–1721. Thompson, M. P., and H. M. Farrell. 1973. The casein micelle—The Odagiri, S., and T. A. Nickerson. 1965. Micellar changes in skimmilk forces contributing to its integrity. Neth. Milk Dairy J. 27:220– treated with alkali or acid. J. Dairy Sci. 48:1157–1160. 239. Ozcan, T., J. A. Lucey, and D. Horne. 2008. Effect of tetrasodium Upreti, P., P. Buhlmann, and L. E. Metzger. 2006. Influence of calcium pyrophosphate on the physicochemical properties of yogurt gels. and phosphorus, lactose, and salt-to-moisture ratio on Cheddar J. Dairy Sci. 91:4492–4500. cheese quality: pH buffering properties of cheese. J. Dairy Sci. Ozcan-Yilsay, T., W. J. Lee, D. Horne, and J. A. Lucey. 2007. Effect 89:938–950. of trisodium citrate on rheological and physical properties and Vaia, B., M. A. Smiddy, A. L. Kelly, and T. Huppertz. 2006. Solvent- microstructure of yogurt. J. Dairy Sci. 90:1644–1652. mediated disruption of bovine casein micelles at alkaline pH. J. Park, Y. W. 2000. Comparison of mineral and cholesterol composition Agric. Food Chem. 54:8288–8293. of different commercial goat milk products manufactured in USA. van Dijk, H. J. M. 1992. The properties of casein micelles. 6. Behav- Small Rumin. Res. 37:115–124. iour above pH 9, and implications for the micelle model. Neth. Philippe, M., F. Gaucheron, Y. Le Great, F. Michel, and A. Garem. Milk Dairy J. 46:101–113. 2003. Physicochemical characterization of calcium-supplemented van Vliet, T., and C. J. A. M. Keetels. 1995. Effect of preheating skim milk. Lait 83:45–59. milk on the structure of acidified milk gels. Neth. Milk Dairy J. Pierre, A., and G. Brulé. 1981. Mineral and protein equilibria between 49:27–35. the colloidal and soluble phase of milk at low temperature. J. Walstra, P., and R. Jenness. 1984. Dairy Chemistry and Physics. John Dairy Res. 48:417–428. Wiley & Sons Inc., New York, NY. Pouliot, Y., M. Boulet, and P. Paquin. 1989. Observations on the heat- White, J. C. D., and D. T. Davies. 1958. The relation between the induced salt balance changes in milk. II. Reversibility on cooling. chemical composition of milk and the stability of the caseinate J. Dairy Res. 56:193–199. complex. I. General introduction, description of samples, methods Pyne, G. T., and T. C. A. McGann. 1960. The colloidal phosphate of and chemical composition of samples. J. Dairy Res. 25:236–255. milk. 2. Influence of citrate. J. Dairy Res. 27:9–17. Singh, H. 1994. Crosslinking of milk proteins on heating concentrated milk at 120°C. Int. Dairy J. 4:477–489. Journal of Dairy Science Vol. 94 No. 11, 2011