- Email: [email protected]
Association of the Quadruply Phosphorylated β-Casein from Human Milk with the Nonphosphorylated Form
Association of the Quadruply Phosphorylated β-Casein from Human Milk with the Nonphosphorylated Form S. M. Sood* and C. W. Slattery*,† *Department of Biochemistry and †Department of Pediatrics, School of Medicine, Loma Linda University, Loma Linda, CA 92350
ABSTRACT
INTRODUCTION
Human β-casein (β-CN) is phosphorylated at levels from zero (β-CN-0P) to five (β-CN-5P). The major constituent is the 4P form (∼35%), whereas the 0P form (∼5%) has been implicated in the formation of a framework upon which the forms with higher levels of phosphorylation may aggregate. At 4°C in 0.01 M imidazole and 0.02 M NaCl, pH 7, with a 1:1 (wt:wt) 0P:4P ratio and a total protein concentration of 3 mg/ml, the s20,w was 1.4 S (monomer). Laser light scattering gave a radius of ∼4.5 nm. As the temperature, T, increased, s20,w increased to 2 S. At 25°C, peaks of 9.5 S and 2 S were observed. This transition T was different from that of either form. At 37°C, a single peak was again observed with s20,w of 17.5 S, compared with 42 S for the 0P and 14 S for the 4P form. Laser light scattering at 37°C revealed a polymer of ∼16 nm radius and D20,w of 1.55 cm2/s. A combination of D20,w and s20,w gave a relative molecular mass suggesting about 45 monomers per polymer. An incubation of 3 h or more at 37°C caused further aggregation, characteristic of the 0P form, and supported the concept of framework formation. At pH 6.6, s20,w was 38 S compared with 1.4 S at pH 10.4. Hydrostatic pressure did not have a large effect but supported a soap micelle-like structure for the polymer. The turbidity of the mixture increased with the amount of CaCl2 and T until the protein precipitated. The properties of the 1:1 mixture of these human βCN are intermediate but probably more biased toward those for the 4P form. (Key words: human milk, human β-casein, proteinprotein interactions, protein-ion interactions)
The formation of CN micelles in milk comes about through the stabilizing properties of κ-CN and prevents precipitation of the calcium-sensitive proteins, usually α- and β-CN (Waugh and von Hippel, 1956). Because the amount of α-CN in human milk is not significant (Cavaletto et al., 1994; Rasmussen et al., 1995), the calcium-sensitive proteins in this system are mainly βCN. Human β-CN is unique and interesting in the sense that it is phosphorylated at six different levels, from 0 (β-CN-0P) to 5 (β-CN-5P) per molecule (Groves and Gordon, 1970). Knowledge concerning the interactions that take place between these different forms prior to and during micelle formation is of paramount importance for understanding the structure of the human milk micelles and how they function in this system. The β-CN composition of the micelles is of interest because about 40% of the total has three or fewer phosphorylated groups, with the β-CN-2P form making up about 26% of the total β-CN (Sood et al., 1985). It has been reported that at least three phosphate groups are needed for cross-linking of the micellar caseins by colloidal calcium phosphate (Aoki et al., 1992). Furthermore, human β-CN has been shown to have anti-Haemophilus influenzae activity, which is detected only in molecules with three or more phosphate groups (Kroening et al., 1999). The major component overall is not the β-CN-5P or fully phosphorylated form, which is the predominant form in the milk of most other species, but the β-CN4P form at about 33% of the total (Sood and Slattery, 1997). All of the different forms have been purified, and their self-association properties, in the presence and absence of Ca+2 ions, have been reported (Sood et al., 1985, 1988, 1990, 1992; Sood and Slattery, 1993, 1997). Additionally, under conditions in which human milk micelles tend to dissociate, the forms with fewer phosphorylated groups, especially β-CN-0P and β-CN-1P, have been shown to stay in the micelle and perhaps form a framework with which other entities may interact upon the reversal of the stressful conditions (Sood et al., 1997, 1998). The nature of the interactions between the major constituent of the system and these minor entities thus needs to be elucidated. For the present report we chose to study the interactions between
Abbreviation key: β-CN-0P to β-CN-5P = phosphorylation level ranging from zero to five, as indicated by number preceding P, D20,w = diffusion coefficient corrected to 20°C and water solution, s20,w = sedimentation coefficient corrected to 20°C and water solution.
Received April 11, 2000. Accepted August 8, 2000. Corresponding author: S. M. Sood; e-mail: [email protected]. 2000 J Dairy Sci 83:2766–2770
2766
2767
ASSOCIATION OF HUMAN MILK β-CASEINS
β-CN-4P, the major constituent, and the form present in the smallest amount (at approximately 5%), the βCN-0P form. MATERIALS AND METHODS The β-CN-0P and β-CN-4P proteins were prepared as described earlier (Sood et al., 1988; Sood and Slattery, 1997) by using a slightly modified procedure of Groves and Gordon (1970). Further purification by HPLC on Mono Q (Pharmacia LKB, Piscataway, NJ) (Slattery et al., 1989), yielded virtually pure samples. Mixtures of the proteins after freeze-drying were dissolved in 0.01 M imidazole buffer at pH 7 containing 0.02 M NaCl (low salt buffer). Sedimentation coefficients were determined with a Beckman Model E analytical ultracentrifuge (Beckman Instruments, Fullerton, CA) (Sood et al., 1985). The diffusion coefficients and size of the monomeric and polymeric proteins were determined by laser light scattering using a Nicomp Model 370 submicron particle sizing system and Nicomp analysis software (Pacific Scientific, Silver Spring, MD) (Sood et al., 1992). Sedimentation coefficients (s20,w) were corrected to standard conditions of 20°C and water solution, as were diffusion coefficients (D20,w). Pressure effects in ultracentrifugation experiments were assessed by overlayering the sample solution with oil or by changing the rotor speed (Harrington and Kegeles, 1973). Light absorbence at 400 nm in the absence and presence of different amounts of Ca+2 was measured at various temperatures as an indication of turbidity from protein aggregation. RESULTS AND DISCUSSION There are several processes involved in the formation of casein micelles from newly synthesized protein molecules, and the order in which order they occur is not known. For example, it is not known if phosphorylation by casein kinase precedes or follows protein aggregation. Nor is it known what, other than chance, may determine the unique phosphorylation pattern. Once a pattern is established, association between the entities with different levels of phosphorylation are assumed to establish an equilibrium distribution of products and reactants dependent upon the concentrations of the different forms and the constants governing each interaction. The presence of ionic Ca+2 further complicates the process.
Protein Ratio The initial sedimentation velocity experiments tested various ratios of β-CN-0P:β-CN-4P (wt:wt) from 1:15 up to 1:1 at 37°C with a total protein concentration of 3 mg/ml. The results are shown in Table 1. Although under these conditions s20,w was 42 S for the 0P form alone (Sood et al., 1988) and 14 S for the 4P form alone (Sood and Slattery, 1997), only single peaks with s20,w values close to that for the 4P form alone were observed, which was indicative of extensive interaction between the two forms leading to polymers of similar structure. Table 1 shows that as the amount of the 4P form increased in the mixture, the s20,w value decreased. This change was probably due to the electrostatic forces of repulsion from the negatively charged phosphate groups on the more highly phosphorylated form. Another interesting observation made during these studies was that the amount of material sedimenting to the bottom of the centerpiece during a run decreased as the amount of the 4P form in the mixture increased. At the 1:15 ratio there was very little deposit. Because intermolecular interactions probably begin with the interaction of a single molecule of each type, we chose to continue the studies using the 1:1 protein ratio, although the amount of the 0P form present in human milk is much less than that of the 4P form (1:7). Protein Stability There is a large difference in the aggregation behavior of the β-CN-0P and β-CN-4P forms of human β-CN with temperature, T. The 0P form is solubilized with some difficulty even at a low T (4°C), and it aggregates to become opaque but without precipitation at 37°C (Sood et al., 1988). In contrast, the 4P form is readily soluble and remains as a clear solution up to 37°C (Sood and Slattery, 1997). Although the results with different protein ratios indicate an extensive interaction between the two forms, the stability of these interaction products is of some interest. By using the 1:1 (wt:wt) β-CN-0P:β-CN-4P ratio, a stability test was performed by allowing the solution to stand at 37°C for different periods of time. After 3 h, the solution developed some turbidity (which cleared on cooling in ice water), and an additional faster peak (>800 S) was observed upon sedimentation. Also, an additional gel-like deposit was observed on the bottom of the centerpiece at the end of the run. With time, the
Table 1. Sedimentation velocity of mixtures of the 0P and 4P forms of human β-CN at different ratios in low salt buffer, pH 7.1, at 37°C and a protein concentration of 3 mg/ml. OP:4P Ratio s20,w (S)
1:1 17.5 ± 0.5
1:1.67 15.6 ± 0.1
1:3 16.0 ± 0.1
1:7 14.2 ± 0.1
1:15 13.1 ± 0.1
Journal of Dairy Science Vol. 83, No. 12, 2000
2768
SOOD AND SLATTERY
Table 2. Effect of temperature on the particle size of a 1:1 (wt:wt) mixture of the 0P and 4P forms of human β-CN in low salt buffer, pH 7.0, at a protein concentrations of 3 mg/ml. Temperature (°C) 4 10 20 25 30 33 37 37 (in 3.3 M urea + imidazole
s20,w (S) 1.4 ... 1.65 9.5 13.2 17.1 17.5 buffer, pH 7.0)
Diameter (nm) 9.0 9.1 12.4 23.0 25.8 27.8 31.3 7.5
turbidity increased. At 24 h, the solution was much more turbid than at 3 h, and a clear solution was not obtainable by cooling for several hours in ice water. The size of the faster peak and the relative amount of the gel-like material increased with time but the s20,w for the faster peak did not change up to 24 h. Addition of a protease inhibitor (phenylmethylsulfonylfluoride) did not eliminate the development of this turbidity. Furthermore, electrophoresis of the solutions on precast, 10% Tricine gels (Novex, San Diego, CA) in the presence of SDS after 3 and 24 h standing at 37°C, showed no detectable proteolysis. These results suggest that a dynamic exchange is established and rearranges the molecules in the aggregates. However, the equilibrium for the 0P form lies further toward aggregation such that the continual rearrangement results in some aggregates predominately containing the 0P form. These aggregates tend to grow to a much larger size than when limited by the charges on the phosphates of the 4P form. This finding is further evidence that, in the micelles, the few forms with fewer phosphates may develop a stable framework for redeposition of the more highly phosphorylated forms when conditions are normal. As a consequence of this behavior, measurements were made within the first few hours while only a single species was present after the initial interaction.
peak with s20,w equal to 1.4 S and a radius of ∼4.5 nm by laser light scattering (Table 2). Above 20°C, two peaks were observed, one of ∼2 S that decreased in amount, and a faster peak that increased in amount as T increased until only the faster peak was present at 37°C. The S values for the faster peak are also shown in Table 2. The limiting value for s20,w of about 17.5 S at 37°C is intermediate between 14 S for the pure 4P form and 42 S for the pure 0P form. Extensive interaction in the absence of calcium is evident by the fact that the individual sedimentation peaks are not present and are replaced by a homogeneous single peak for the complex. The sedimentation coefficient (17.5 S) for a 1:1 weight ratio mixture is much closer to the value for the 4P form alone, reflecting the extent of the interaction, which then limits the size of the polymers because of the charge on the phosphate groups. The data in Table 2 show that the physical size of the polymers, as determined by laser light scattering, did not increase as drastically above 20°C as did the particle mass, indicated by the s20,w. Eventually, the mass and the size apparently began to level off. This finding is again consistent with a spherical soap micelle-like structure into which more monomers may be packed as T increases, without increasing the polymer diameter much. At 37°C, a polymer was obtained with ∼16 nm radius with D20,w of 1.55 × 10−7 cm2/s. Under these conditions, the combination of the s20,w and D20,w data may be used to calculate that the polymer has a molecular mass equivalent to approximately 45 monomers per polymer. These interactions at 37°C may be completely disrupted by 3.3 M urea buffer where the molecule has a diameter of ∼4 nm, a value reported for the monomers of the individual proteins. Pressure Effects on Association
At 4°C, the protein mixture showed all the characteristics of the individual monomeric proteins: a single
If there was a change in the volume of the polymer after association relative to the dissociated monomers, the pressure developed in the analytical ultracentrifuge would cause a change in the extent of association. Different pressures were applied by changing rotor speed or by overlayering with oil. Results of such experiments are shown in Table 3. Increasing the rotor speed or
Table 3. Pressure effects on the sedimentation of a 1:1 (wt:wt) mixture of the unphosphorylated and four phosphorylations forms a human β-CN in low salt buffer, pH 7.1, at 37°C.
Table 4. Effects of pH on the sedimentation velocity of a 1:1 (wt:wt) mixture of the 0P and 4P forms of human β-CN in low salt buffer, pH 7.1, at 37°C and a protein concentration of 3 mg/ml.
Sample volume1
rpm
s20,w (S)
pH
s20,w (S)
0.70 ml 0.70 ml 0.40 ml + 0.30 ml of oil
32,000 48,000 32,000
17.5 ± 0.1 15.3 ± 0.1 13.8 ± 0.1
6.6 7.1 8 10
38.2 17.5 8.2 1.4
T-Dependent Association
1
Protein concentration of 3 mg/ml.
Journal of Dairy Science Vol. 83, No. 12, 2000
± ± ± ±
0.1 0.1 0.1 0.1
2769
ASSOCIATION OF HUMAN MILK β-CASEINS Table 5. Appearance of solutions of a 1:1 (wt:wt) mixture of 0P and 4P forms of human β-CN in the presence and absence of calcium ions.1 CaCl2 concentration (mM) Temperature (°C)
0
2.5
5.0
7.5
10
19.0 20.1 20.7 21.0 21.6 22.8 25.4 26.8 27.6 30.3 33
Clear Clear Clear Clear Clear Clear Clear Clear Clear Clear Clear
Clear Clear Clear Clear Clear Clear Slightly cloudy Turbid Precipitate Precipitate
Clear Clear Clear Slightly cloudy Turbid Precipitate
Clear Slightly cloudy Slightly cloudy Cloudy Precipitate
Clear Slightly cloudy Turbid Turbid
1
In low salt buffer, pH 7.1, and a protein concentration of 3 mg/ml.
overlayering with oil caused a small decrease in the s20,w values, consistent with a decrease in association. This finding indicated that the structure of a polymer may be such that its volume was greater than the aggregate volume of the individual monomers. One could visualize this in a polymer with monomers associated in a soap micelle-like structure, resulting in an internal cavity filled with solvent. An increase in pressure would tend to cause the cavity to be smaller and leave less room for monomers to fit into the particle, resulting in an overall smaller particle with a lower sedimentation coefficient. This structure would also be consistent with formation of a polymer of limited size as T is increased, as suggested in Table 2.
ter process would lead to larger aggregates, which, in the absence of a stabilizing protein such as κ-CN, would eventually precipitate. The 0P protein also precipitates in the presence of Ca+2 ions by a mechanism that is not completely understood. The visual appearance of the 1:1 mixture at different levels of Ca+2 and different T is shown in Table 5. With 2.5 mM CaCl2 and at a T in which the solution starts getting cloudy (25.5°C), sedimentation coefficients were measured, and two peaks were observed. The faster peak was quite skewed and had a value of ∼31 S for s20,w, whereas the slow peak was that of a monomer. Turbidimetry at 400 nm for different Ca+2 concentrations gave a more precise measurement of the increase in size of the protein particles to the level where light was scattered (Figure 1).
Charge Effects on Association Association into a soap micelle-like structure, as proposed above, would also be affected by the charge on the monomer molecules and could cause them to repel each other. Changing the charge, particularly on the 4P form, by changing the pH did cause increases or decreases in association, reflected in the s20,w as shown in Table 4. At pH 6.6, a value of ∼38 S was obtained for the s20,w. Apparently, there must be groups on the 0P protein as well on as the 4P form that can change their charge to repel each other and eliminate association altogether or disrupt the binding sites of these molecules. At pH 10, the molecules apparently exist as monomers, even at 37°C, indicating that charge effects override any hydrophobic interactions. Addition of CaCl2 Calcium ions can bind to the phosphate groups on the 4P protein and reduce its charge. They may also form bridges between phosphate groups on adjacent monomers either within or between polymers. The lat-
Figure 1. Absorbance at 400 nm (A400) as a function of temperature for a 1:1 (wt:wt) mixture of the 0P and 4P forms of human β-CN in low salt buffer, pH 7.1, with a protein concentration of 3 mg/ml. The curves represent 0.0 (䊉), 2.5 (䊏), 5.0 (▲), 7.5 (▼), and 10.0 mM (◆) CaCl2. Journal of Dairy Science Vol. 83, No. 12, 2000
2770
SOOD AND SLATTERY
These data are intermediate between similar numbers for 4P and 0P but are biased toward the 4P data (Sood, et al., 1988; Sood and Slattery, 1997). SUMMARY AND CONCLUSIONS The data for the association of the 4P and 0P forms of human β-CN suggest an intimate interaction between the forms so that the 0P form appears to be phosphorylated. A conformational change is likely in each protein with T that exposes hydrophobic sites and allows for increased association at higher T at which hydrophobic interactions are stronger. The association product may reach a limited size at about 37°C, which depends upon the charge on the proteins as regulated by pH and on the pressure on the solution. This finding is consistent with the formation of spherical, soap micelle-like mixed polymers. Added Ca+2 binds to the negative groups, particularly phosphates, on the proteins and decreases the net charge, allowing for increased association within the mixed polymers. Above the T at which a conformational change takes place, interpolymer association through hydrophobic interactions or calcium bridges causes formation of larger milk micellelike aggregates that precipitate in the absence of κ-CN. The results of experiments examining the stability of these protein aggregates with time are consistent with the suggestion that the forms of β-CN with fewer than three phosphorylated groups may associate together to provide a framework for the micelle that is stable to changes in environmental conditions such as pH, cooling, or average mineral (Ca+2) content. ACKNOWLEDGMENTS The authors express their deep appreciation to Ann Pearson for preparing this manuscript for submission.
Journal of Dairy Science Vol. 83, No. 12, 2000
REFERENCES Aoki, T., T. Umeda, and Y. Kako. 1992. The least number of phosphate groups for crosslinking of casein by colloidal calcium phosphate. J. Dairy Sci. 75:971–975. Cavaletto, M., A. Contisani, G. Giuffride, L. Napolitano, and A. Conti. 1994. Human αS1-Casein like protein: purification and N-terminal sequence determination. Biol. Chem. Hoppe-Seyler 375:149–157. Groves, M. L., and W. G. Gordon. 1970. The major component of human casein: a protein phosphorylated at different levels. Arch. Biochem. Biophys. 140:47–51. Harrington, W. F., and G. Kegeles. 1973. Pressure effects in ultracentrifugation of interacting systems. Methods Enzymol. 27D:306– 345. Kroening, T. A., J. H. Baxter, S. A. Anderson, R. G. Hards, L. Harvey, and P. Mukerji. 1999. Concentrations and anti-Haemophilus influenzae activities of beta-casein phosphoforms in human milk. J. Pediatr. Gastroenterol. Nutr. 28:486–491. Rasmussen, L. K., H. A. Due, and T. E. Petersen. 1995. Human αS1casein: purification and characterization. Comp. Biochem. Physiol. 111B:75–81. Slattery, C. W., S. M. Sood, and P. Chang. 1989. Hydrophobic interactions in human casein micelle formation: β-casein aggregation. J. Dairy Res. 56:427–433. Sood, S. M., P. Chang, and C. W. Slattery. 1985. Interactions in human casein systems: self-association of fully phosphorylated human β-casein. Arch. Biochem. Biophys. 242:355–364. Sood, S. M., P. Chang, and C. W. Slattery. 1988. Interactions in human casein system: self-association of nonphosphorylated human βcasein. Arch. Biochem. Biophys. 264:574–583. Sood, S. M., P. Chang, and C. W. Slattery. 1990. Interactions of triply phosphorylated human β-casein: monomer characterization and hydrodynamic studies of self-association. Arch. Biochem. Biophys. 277:415–421. Sood, S. M., P. Chang, and C. W. Slattery. 1992. Interaction properties of doubly phosphorylated β-casein: a major component of the human milk caseins. J. Dairy Sci. 75:2937–2945. Sood, S. M., P. J. Herbert, and C. W. Slattery. 1997. Structural studies on casein micelles of human milk; dissociation of β-casein of different phosphorylation levels induced by cooling and ethylenediaminetetaacetate. J. Dairy Sci. 80:628–633. Sood, S. M., P. J. Herbert, and C. W. Slattery. 1998. The pH-dependent dissociation of β-casein from human milk micelles: role of electrostatic interactions. J. Dairy Sci. 81:3149–3153. Sood, S. M., and C. W. Slattery. 1993. Interaction properties of singly phosphorylated human β-casein: similarities with other phosphorylation levels. J. Dairy Sci. 77:405–412. Sood, S. M. and C. W. Slattery. 1997. Monomer characterization and studies of self-association of the major β-casein of human milk. J. Dairy Sci. 80:1554–1560. Waugh, D. F, and P. H. von Hippel. 1956. κ-Casein and the stabilization of casein micelles. J. Am. Chem. Soc. 78:4576–4582.
ABSTRACT
INTRODUCTION
Human β-casein (β-CN) is phosphorylated at levels from zero (β-CN-0P) to five (β-CN-5P). The major constituent is the 4P form (∼35%), whereas the 0P form (∼5%) has been implicated in the formation of a framework upon which the forms with higher levels of phosphorylation may aggregate. At 4°C in 0.01 M imidazole and 0.02 M NaCl, pH 7, with a 1:1 (wt:wt) 0P:4P ratio and a total protein concentration of 3 mg/ml, the s20,w was 1.4 S (monomer). Laser light scattering gave a radius of ∼4.5 nm. As the temperature, T, increased, s20,w increased to 2 S. At 25°C, peaks of 9.5 S and 2 S were observed. This transition T was different from that of either form. At 37°C, a single peak was again observed with s20,w of 17.5 S, compared with 42 S for the 0P and 14 S for the 4P form. Laser light scattering at 37°C revealed a polymer of ∼16 nm radius and D20,w of 1.55 cm2/s. A combination of D20,w and s20,w gave a relative molecular mass suggesting about 45 monomers per polymer. An incubation of 3 h or more at 37°C caused further aggregation, characteristic of the 0P form, and supported the concept of framework formation. At pH 6.6, s20,w was 38 S compared with 1.4 S at pH 10.4. Hydrostatic pressure did not have a large effect but supported a soap micelle-like structure for the polymer. The turbidity of the mixture increased with the amount of CaCl2 and T until the protein precipitated. The properties of the 1:1 mixture of these human βCN are intermediate but probably more biased toward those for the 4P form. (Key words: human milk, human β-casein, proteinprotein interactions, protein-ion interactions)
The formation of CN micelles in milk comes about through the stabilizing properties of κ-CN and prevents precipitation of the calcium-sensitive proteins, usually α- and β-CN (Waugh and von Hippel, 1956). Because the amount of α-CN in human milk is not significant (Cavaletto et al., 1994; Rasmussen et al., 1995), the calcium-sensitive proteins in this system are mainly βCN. Human β-CN is unique and interesting in the sense that it is phosphorylated at six different levels, from 0 (β-CN-0P) to 5 (β-CN-5P) per molecule (Groves and Gordon, 1970). Knowledge concerning the interactions that take place between these different forms prior to and during micelle formation is of paramount importance for understanding the structure of the human milk micelles and how they function in this system. The β-CN composition of the micelles is of interest because about 40% of the total has three or fewer phosphorylated groups, with the β-CN-2P form making up about 26% of the total β-CN (Sood et al., 1985). It has been reported that at least three phosphate groups are needed for cross-linking of the micellar caseins by colloidal calcium phosphate (Aoki et al., 1992). Furthermore, human β-CN has been shown to have anti-Haemophilus influenzae activity, which is detected only in molecules with three or more phosphate groups (Kroening et al., 1999). The major component overall is not the β-CN-5P or fully phosphorylated form, which is the predominant form in the milk of most other species, but the β-CN4P form at about 33% of the total (Sood and Slattery, 1997). All of the different forms have been purified, and their self-association properties, in the presence and absence of Ca+2 ions, have been reported (Sood et al., 1985, 1988, 1990, 1992; Sood and Slattery, 1993, 1997). Additionally, under conditions in which human milk micelles tend to dissociate, the forms with fewer phosphorylated groups, especially β-CN-0P and β-CN-1P, have been shown to stay in the micelle and perhaps form a framework with which other entities may interact upon the reversal of the stressful conditions (Sood et al., 1997, 1998). The nature of the interactions between the major constituent of the system and these minor entities thus needs to be elucidated. For the present report we chose to study the interactions between
Abbreviation key: β-CN-0P to β-CN-5P = phosphorylation level ranging from zero to five, as indicated by number preceding P, D20,w = diffusion coefficient corrected to 20°C and water solution, s20,w = sedimentation coefficient corrected to 20°C and water solution.
Received April 11, 2000. Accepted August 8, 2000. Corresponding author: S. M. Sood; e-mail: [email protected]. 2000 J Dairy Sci 83:2766–2770
2766
2767
ASSOCIATION OF HUMAN MILK β-CASEINS
β-CN-4P, the major constituent, and the form present in the smallest amount (at approximately 5%), the βCN-0P form. MATERIALS AND METHODS The β-CN-0P and β-CN-4P proteins were prepared as described earlier (Sood et al., 1988; Sood and Slattery, 1997) by using a slightly modified procedure of Groves and Gordon (1970). Further purification by HPLC on Mono Q (Pharmacia LKB, Piscataway, NJ) (Slattery et al., 1989), yielded virtually pure samples. Mixtures of the proteins after freeze-drying were dissolved in 0.01 M imidazole buffer at pH 7 containing 0.02 M NaCl (low salt buffer). Sedimentation coefficients were determined with a Beckman Model E analytical ultracentrifuge (Beckman Instruments, Fullerton, CA) (Sood et al., 1985). The diffusion coefficients and size of the monomeric and polymeric proteins were determined by laser light scattering using a Nicomp Model 370 submicron particle sizing system and Nicomp analysis software (Pacific Scientific, Silver Spring, MD) (Sood et al., 1992). Sedimentation coefficients (s20,w) were corrected to standard conditions of 20°C and water solution, as were diffusion coefficients (D20,w). Pressure effects in ultracentrifugation experiments were assessed by overlayering the sample solution with oil or by changing the rotor speed (Harrington and Kegeles, 1973). Light absorbence at 400 nm in the absence and presence of different amounts of Ca+2 was measured at various temperatures as an indication of turbidity from protein aggregation. RESULTS AND DISCUSSION There are several processes involved in the formation of casein micelles from newly synthesized protein molecules, and the order in which order they occur is not known. For example, it is not known if phosphorylation by casein kinase precedes or follows protein aggregation. Nor is it known what, other than chance, may determine the unique phosphorylation pattern. Once a pattern is established, association between the entities with different levels of phosphorylation are assumed to establish an equilibrium distribution of products and reactants dependent upon the concentrations of the different forms and the constants governing each interaction. The presence of ionic Ca+2 further complicates the process.
Protein Ratio The initial sedimentation velocity experiments tested various ratios of β-CN-0P:β-CN-4P (wt:wt) from 1:15 up to 1:1 at 37°C with a total protein concentration of 3 mg/ml. The results are shown in Table 1. Although under these conditions s20,w was 42 S for the 0P form alone (Sood et al., 1988) and 14 S for the 4P form alone (Sood and Slattery, 1997), only single peaks with s20,w values close to that for the 4P form alone were observed, which was indicative of extensive interaction between the two forms leading to polymers of similar structure. Table 1 shows that as the amount of the 4P form increased in the mixture, the s20,w value decreased. This change was probably due to the electrostatic forces of repulsion from the negatively charged phosphate groups on the more highly phosphorylated form. Another interesting observation made during these studies was that the amount of material sedimenting to the bottom of the centerpiece during a run decreased as the amount of the 4P form in the mixture increased. At the 1:15 ratio there was very little deposit. Because intermolecular interactions probably begin with the interaction of a single molecule of each type, we chose to continue the studies using the 1:1 protein ratio, although the amount of the 0P form present in human milk is much less than that of the 4P form (1:7). Protein Stability There is a large difference in the aggregation behavior of the β-CN-0P and β-CN-4P forms of human β-CN with temperature, T. The 0P form is solubilized with some difficulty even at a low T (4°C), and it aggregates to become opaque but without precipitation at 37°C (Sood et al., 1988). In contrast, the 4P form is readily soluble and remains as a clear solution up to 37°C (Sood and Slattery, 1997). Although the results with different protein ratios indicate an extensive interaction between the two forms, the stability of these interaction products is of some interest. By using the 1:1 (wt:wt) β-CN-0P:β-CN-4P ratio, a stability test was performed by allowing the solution to stand at 37°C for different periods of time. After 3 h, the solution developed some turbidity (which cleared on cooling in ice water), and an additional faster peak (>800 S) was observed upon sedimentation. Also, an additional gel-like deposit was observed on the bottom of the centerpiece at the end of the run. With time, the
Table 1. Sedimentation velocity of mixtures of the 0P and 4P forms of human β-CN at different ratios in low salt buffer, pH 7.1, at 37°C and a protein concentration of 3 mg/ml. OP:4P Ratio s20,w (S)
1:1 17.5 ± 0.5
1:1.67 15.6 ± 0.1
1:3 16.0 ± 0.1
1:7 14.2 ± 0.1
1:15 13.1 ± 0.1
Journal of Dairy Science Vol. 83, No. 12, 2000
2768
SOOD AND SLATTERY
Table 2. Effect of temperature on the particle size of a 1:1 (wt:wt) mixture of the 0P and 4P forms of human β-CN in low salt buffer, pH 7.0, at a protein concentrations of 3 mg/ml. Temperature (°C) 4 10 20 25 30 33 37 37 (in 3.3 M urea + imidazole
s20,w (S) 1.4 ... 1.65 9.5 13.2 17.1 17.5 buffer, pH 7.0)
Diameter (nm) 9.0 9.1 12.4 23.0 25.8 27.8 31.3 7.5
turbidity increased. At 24 h, the solution was much more turbid than at 3 h, and a clear solution was not obtainable by cooling for several hours in ice water. The size of the faster peak and the relative amount of the gel-like material increased with time but the s20,w for the faster peak did not change up to 24 h. Addition of a protease inhibitor (phenylmethylsulfonylfluoride) did not eliminate the development of this turbidity. Furthermore, electrophoresis of the solutions on precast, 10% Tricine gels (Novex, San Diego, CA) in the presence of SDS after 3 and 24 h standing at 37°C, showed no detectable proteolysis. These results suggest that a dynamic exchange is established and rearranges the molecules in the aggregates. However, the equilibrium for the 0P form lies further toward aggregation such that the continual rearrangement results in some aggregates predominately containing the 0P form. These aggregates tend to grow to a much larger size than when limited by the charges on the phosphates of the 4P form. This finding is further evidence that, in the micelles, the few forms with fewer phosphates may develop a stable framework for redeposition of the more highly phosphorylated forms when conditions are normal. As a consequence of this behavior, measurements were made within the first few hours while only a single species was present after the initial interaction.
peak with s20,w equal to 1.4 S and a radius of ∼4.5 nm by laser light scattering (Table 2). Above 20°C, two peaks were observed, one of ∼2 S that decreased in amount, and a faster peak that increased in amount as T increased until only the faster peak was present at 37°C. The S values for the faster peak are also shown in Table 2. The limiting value for s20,w of about 17.5 S at 37°C is intermediate between 14 S for the pure 4P form and 42 S for the pure 0P form. Extensive interaction in the absence of calcium is evident by the fact that the individual sedimentation peaks are not present and are replaced by a homogeneous single peak for the complex. The sedimentation coefficient (17.5 S) for a 1:1 weight ratio mixture is much closer to the value for the 4P form alone, reflecting the extent of the interaction, which then limits the size of the polymers because of the charge on the phosphate groups. The data in Table 2 show that the physical size of the polymers, as determined by laser light scattering, did not increase as drastically above 20°C as did the particle mass, indicated by the s20,w. Eventually, the mass and the size apparently began to level off. This finding is again consistent with a spherical soap micelle-like structure into which more monomers may be packed as T increases, without increasing the polymer diameter much. At 37°C, a polymer was obtained with ∼16 nm radius with D20,w of 1.55 × 10−7 cm2/s. Under these conditions, the combination of the s20,w and D20,w data may be used to calculate that the polymer has a molecular mass equivalent to approximately 45 monomers per polymer. These interactions at 37°C may be completely disrupted by 3.3 M urea buffer where the molecule has a diameter of ∼4 nm, a value reported for the monomers of the individual proteins. Pressure Effects on Association
At 4°C, the protein mixture showed all the characteristics of the individual monomeric proteins: a single
If there was a change in the volume of the polymer after association relative to the dissociated monomers, the pressure developed in the analytical ultracentrifuge would cause a change in the extent of association. Different pressures were applied by changing rotor speed or by overlayering with oil. Results of such experiments are shown in Table 3. Increasing the rotor speed or
Table 3. Pressure effects on the sedimentation of a 1:1 (wt:wt) mixture of the unphosphorylated and four phosphorylations forms a human β-CN in low salt buffer, pH 7.1, at 37°C.
Table 4. Effects of pH on the sedimentation velocity of a 1:1 (wt:wt) mixture of the 0P and 4P forms of human β-CN in low salt buffer, pH 7.1, at 37°C and a protein concentration of 3 mg/ml.
Sample volume1
rpm
s20,w (S)
pH
s20,w (S)
0.70 ml 0.70 ml 0.40 ml + 0.30 ml of oil
32,000 48,000 32,000
17.5 ± 0.1 15.3 ± 0.1 13.8 ± 0.1
6.6 7.1 8 10
38.2 17.5 8.2 1.4
T-Dependent Association
1
Protein concentration of 3 mg/ml.
Journal of Dairy Science Vol. 83, No. 12, 2000
± ± ± ±
0.1 0.1 0.1 0.1
2769
ASSOCIATION OF HUMAN MILK β-CASEINS Table 5. Appearance of solutions of a 1:1 (wt:wt) mixture of 0P and 4P forms of human β-CN in the presence and absence of calcium ions.1 CaCl2 concentration (mM) Temperature (°C)
0
2.5
5.0
7.5
10
19.0 20.1 20.7 21.0 21.6 22.8 25.4 26.8 27.6 30.3 33
Clear Clear Clear Clear Clear Clear Clear Clear Clear Clear Clear
Clear Clear Clear Clear Clear Clear Slightly cloudy Turbid Precipitate Precipitate
Clear Clear Clear Slightly cloudy Turbid Precipitate
Clear Slightly cloudy Slightly cloudy Cloudy Precipitate
Clear Slightly cloudy Turbid Turbid
1
In low salt buffer, pH 7.1, and a protein concentration of 3 mg/ml.
overlayering with oil caused a small decrease in the s20,w values, consistent with a decrease in association. This finding indicated that the structure of a polymer may be such that its volume was greater than the aggregate volume of the individual monomers. One could visualize this in a polymer with monomers associated in a soap micelle-like structure, resulting in an internal cavity filled with solvent. An increase in pressure would tend to cause the cavity to be smaller and leave less room for monomers to fit into the particle, resulting in an overall smaller particle with a lower sedimentation coefficient. This structure would also be consistent with formation of a polymer of limited size as T is increased, as suggested in Table 2.
ter process would lead to larger aggregates, which, in the absence of a stabilizing protein such as κ-CN, would eventually precipitate. The 0P protein also precipitates in the presence of Ca+2 ions by a mechanism that is not completely understood. The visual appearance of the 1:1 mixture at different levels of Ca+2 and different T is shown in Table 5. With 2.5 mM CaCl2 and at a T in which the solution starts getting cloudy (25.5°C), sedimentation coefficients were measured, and two peaks were observed. The faster peak was quite skewed and had a value of ∼31 S for s20,w, whereas the slow peak was that of a monomer. Turbidimetry at 400 nm for different Ca+2 concentrations gave a more precise measurement of the increase in size of the protein particles to the level where light was scattered (Figure 1).
Charge Effects on Association Association into a soap micelle-like structure, as proposed above, would also be affected by the charge on the monomer molecules and could cause them to repel each other. Changing the charge, particularly on the 4P form, by changing the pH did cause increases or decreases in association, reflected in the s20,w as shown in Table 4. At pH 6.6, a value of ∼38 S was obtained for the s20,w. Apparently, there must be groups on the 0P protein as well on as the 4P form that can change their charge to repel each other and eliminate association altogether or disrupt the binding sites of these molecules. At pH 10, the molecules apparently exist as monomers, even at 37°C, indicating that charge effects override any hydrophobic interactions. Addition of CaCl2 Calcium ions can bind to the phosphate groups on the 4P protein and reduce its charge. They may also form bridges between phosphate groups on adjacent monomers either within or between polymers. The lat-
Figure 1. Absorbance at 400 nm (A400) as a function of temperature for a 1:1 (wt:wt) mixture of the 0P and 4P forms of human β-CN in low salt buffer, pH 7.1, with a protein concentration of 3 mg/ml. The curves represent 0.0 (䊉), 2.5 (䊏), 5.0 (▲), 7.5 (▼), and 10.0 mM (◆) CaCl2. Journal of Dairy Science Vol. 83, No. 12, 2000
2770
SOOD AND SLATTERY
These data are intermediate between similar numbers for 4P and 0P but are biased toward the 4P data (Sood, et al., 1988; Sood and Slattery, 1997). SUMMARY AND CONCLUSIONS The data for the association of the 4P and 0P forms of human β-CN suggest an intimate interaction between the forms so that the 0P form appears to be phosphorylated. A conformational change is likely in each protein with T that exposes hydrophobic sites and allows for increased association at higher T at which hydrophobic interactions are stronger. The association product may reach a limited size at about 37°C, which depends upon the charge on the proteins as regulated by pH and on the pressure on the solution. This finding is consistent with the formation of spherical, soap micelle-like mixed polymers. Added Ca+2 binds to the negative groups, particularly phosphates, on the proteins and decreases the net charge, allowing for increased association within the mixed polymers. Above the T at which a conformational change takes place, interpolymer association through hydrophobic interactions or calcium bridges causes formation of larger milk micellelike aggregates that precipitate in the absence of κ-CN. The results of experiments examining the stability of these protein aggregates with time are consistent with the suggestion that the forms of β-CN with fewer than three phosphorylated groups may associate together to provide a framework for the micelle that is stable to changes in environmental conditions such as pH, cooling, or average mineral (Ca+2) content. ACKNOWLEDGMENTS The authors express their deep appreciation to Ann Pearson for preparing this manuscript for submission.
Journal of Dairy Science Vol. 83, No. 12, 2000
REFERENCES Aoki, T., T. Umeda, and Y. Kako. 1992. The least number of phosphate groups for crosslinking of casein by colloidal calcium phosphate. J. Dairy Sci. 75:971–975. Cavaletto, M., A. Contisani, G. Giuffride, L. Napolitano, and A. Conti. 1994. Human αS1-Casein like protein: purification and N-terminal sequence determination. Biol. Chem. Hoppe-Seyler 375:149–157. Groves, M. L., and W. G. Gordon. 1970. The major component of human casein: a protein phosphorylated at different levels. Arch. Biochem. Biophys. 140:47–51. Harrington, W. F., and G. Kegeles. 1973. Pressure effects in ultracentrifugation of interacting systems. Methods Enzymol. 27D:306– 345. Kroening, T. A., J. H. Baxter, S. A. Anderson, R. G. Hards, L. Harvey, and P. Mukerji. 1999. Concentrations and anti-Haemophilus influenzae activities of beta-casein phosphoforms in human milk. J. Pediatr. Gastroenterol. Nutr. 28:486–491. Rasmussen, L. K., H. A. Due, and T. E. Petersen. 1995. Human αS1casein: purification and characterization. Comp. Biochem. Physiol. 111B:75–81. Slattery, C. W., S. M. Sood, and P. Chang. 1989. Hydrophobic interactions in human casein micelle formation: β-casein aggregation. J. Dairy Res. 56:427–433. Sood, S. M., P. Chang, and C. W. Slattery. 1985. Interactions in human casein systems: self-association of fully phosphorylated human β-casein. Arch. Biochem. Biophys. 242:355–364. Sood, S. M., P. Chang, and C. W. Slattery. 1988. Interactions in human casein system: self-association of nonphosphorylated human βcasein. Arch. Biochem. Biophys. 264:574–583. Sood, S. M., P. Chang, and C. W. Slattery. 1990. Interactions of triply phosphorylated human β-casein: monomer characterization and hydrodynamic studies of self-association. Arch. Biochem. Biophys. 277:415–421. Sood, S. M., P. Chang, and C. W. Slattery. 1992. Interaction properties of doubly phosphorylated β-casein: a major component of the human milk caseins. J. Dairy Sci. 75:2937–2945. Sood, S. M., P. J. Herbert, and C. W. Slattery. 1997. Structural studies on casein micelles of human milk; dissociation of β-casein of different phosphorylation levels induced by cooling and ethylenediaminetetaacetate. J. Dairy Sci. 80:628–633. Sood, S. M., P. J. Herbert, and C. W. Slattery. 1998. The pH-dependent dissociation of β-casein from human milk micelles: role of electrostatic interactions. J. Dairy Sci. 81:3149–3153. Sood, S. M., and C. W. Slattery. 1993. Interaction properties of singly phosphorylated human β-casein: similarities with other phosphorylation levels. J. Dairy Sci. 77:405–412. Sood, S. M. and C. W. Slattery. 1997. Monomer characterization and studies of self-association of the major β-casein of human milk. J. Dairy Sci. 80:1554–1560. Waugh, D. F, and P. H. von Hippel. 1956. κ-Casein and the stabilization of casein micelles. J. Am. Chem. Soc. 78:4576–4582.