The degradation of bovine β-casein by plasmin was used as a model system to investigate the applicability of an enzyme recycle reactor for the continuous production and isolation of peptide fractions. A simple, ultrafiltration-type membrane reactor was used incorporating a commercially available hollow-fiber module normally employed for dialysis purposes (artificial kidney). The results were compared with those obtained in a batch reactor process, using HPLC to monitor the course of proteolysis.

Peptide fragments were isolated and were characterized using amino-acid analysis and N-terminal end-group determination. Sixteen peptide fragments were identified, which together accounted for virtually all the potential cleavage sites in β-casein. In the system studied the N-terminal half of β-casein appeared to be more sensitive to plasmic hydrolysis than the rest of the molecule.

Casein phosphopeptides (GPP) were produced by tryptic hydrolysis of sodium caseinate and further purified by precipitation and chromatography on QAE-Sephadex A-25. Their physico-chemical properties were compared with the properties of an enzymically dephosphorylated equivalent preparation (DPP). Binding of Ca2+ to the peptides was measured using a Ca selective electrode and was found to increase with pH and to show 1/1 stoicheiometry Ca/Porg in CPP at pH 6·5 a.nd 7·6. Klotz plots indicated equivalent binding sites at these two pH values, but some heterogeneity was seen at pH 3·5. In contrast, DPP did not bind significant amounts of Ca2+.

CPP effectively inhibited the formation of insoluble calcium phosphates at different Ca/P ratios. The effective CPP concentration was 10 mg/1 and complete stability of calcium phosphate solutions was obtained at about 100 mg/1. This stabilizing effect was dependent on the presence of organic P.

The in vivo formation of biologically active caseinopeptides was studied. It was proved that bioactive peptides were released in the small intestine of minipigs in the course of luminal digestion of diets containing bovine casein. An opioid peptide and a phosphopeptide were isolated from jejunal chyme and were chemically characterized. The opioid peptide has been identified as a fragment of β-casein (60–70). This peptide, named β-casomorphin-ll, displayed substantial opioid activity in an opiate receptor-binding assay. The caseinophosphopeptide has been shown to be a fragment of αs1-casein (66–74). Casein-derived phosphopeptides exhibit a potent ability to form soluble complexes with Ca and trace elements. Evidence exists that casomorphins and caseinophosphopeptides participate in the regulation of nutrient entry.

Casein and other milk proteins in maternal colostrum and milk, the earliest food of the newborn, should not only be considered as a nutritional supply but also as a source of biologically active peptides. Some of them isolated from casein and lactotransferrin were active on platelet function. They inhibited both aggregation of ADP-treated platelets and binding of [125I] fibrinogen to ADP-treated platelets. Their behaviour was compared to that of fibrinogen peptides possessing similar effects: once more similarities between the milk and blood-clotting phenomena could be observed.

Maternal milk should not only be considered as a nutrient, but also as a protecting agent against aggressions from the neonate's new environment. Breastfeeding facilitates transmission of a passive immunity by multifunctional factors which have a direct effect on the neonate's resistance to bacterial and viral infections. Among these factors are the main milk proteins, the caseins: during enzymic digestion of human and bovine caseins, immunomodulating peptides are released. Corresponding synthetic peptides stimulated in vitro phagocytic activity of murine and of human macrophages and exerted in vivo a protective effect against Klebsiella pneumoniae infection of mice. These data suggest that casein peptides may exert a stimulating function on the immune system of the newborn.

Opioid antagonists were sought in the fragments of κ-casein which were obtained by chemical synthesis and enzymic digestion. A synthetic bovine κ-casein peptide (35–41), Tyr-Pro-Ser-Tyr-Gly-Leu-Asn (casoxin A) showed opioid antagonist activity at 200 μM in the guinea pig ileum assay. A synthetic peptide Tyr-Pro-Tyr-Tyr (casoxin B) which is found in bovine and human κ-casein, also showed opioid antagonist activity at 100 μM. Another opioid antagonist peptide (casoxin C) was isolated from tryptic digests of bovine κ-casein by reverse-phase HPLC. The structure of the peptide was Tyr-Ile-Pro-Ile-Gln-Tyr-Val-Leu-Ser-Arg, which corresponded to κ-casein (25–34). Casoxin C was active at 5 μM in the guinea pig ileum assay. Thus, bovine κ-casein contains three potential opioid antagonist sequences.

Astringent off-flavours were observed in commercial samples of UHT-milk heated by direct-steam injection. The off-flavour components from these milk samples were isolated by the chloroform-methanol extraction procedure of Harwalkar & Elliott. The astringent-tasting components in these extracts were partly purified by isoelectric precipitation at pH 7·0 and by size exclusion chromatography on Sephadex G-50 columns. The extracted off-flavour components and the partly purified samples were further analysed by FPLC using a Mono-Q anion exchange column and PAGE. The FPLC and PAGE patterns of the extracts from astringent and non-astringent samples of milk showed marked differences and the astringent flavour correlated with one of the fractions resolved by FPLC. This fraction was found to be heterogeneous by PAGE. The astringent off-flavours in UHT-sterilized milk were γ-casein-like breakdown products of casein.

A series of hydrolysates from commercial sodium caseinate and rennet casein were prepared using different commercial proteolytic enzymes. Increasing degree of hydrolysis, DH, resulted in markedly changed solubility of the products. The level of bitterness varied with the enzyme used. Non-bitter hydrolysates of sodium caseinate were obtained with a DH of 10–12 and protein solubilities of 99%. The rennet casein hydrolysates exhibited low levels of bitterness and high solubility (90%), even at high DH values (DH = 55).

To improve acid clotting properties, enzymic dephosphorylation of caseins with calf intestinal alkaline phosphatase (CAP) or potato acid phosphatase (PAP) was investigated. Greater dephosphorylation was achieved using αs1- or β-casein as substrates, compared to whole casein or skim milk. Electrophoresis of PAP-modified caseins revealed bands with lower mobility and a multibanded pattern in the β-casein region which was similar to that of human β-casein. On the other hand, CAP modification produced electrophoretic bands having lower mobility of the β-casein component, but with higher mobility in the αs1-casein component as well as increased net negative charge in the CAP-casein. PAP-casein formed a fine dispersion upon acidification to pH 4. with a microstructure similar to that of acidified human casein. Greater initial rates of hydrolysis by pepsin at pH 4 were observed for both CAP- and PAP-modified caseins, compared to bovine and human caseins. The rate and extent of hydrolysis remained high for CAP-casein but tended to level off with PAP-casein during sequential digestion with pepsin and pancreatin. There may be advantages in the use of partial dephosphorylation to improve acid clotting and digestibility properties of bovine casein for infant feeding.

The rapid isolation of major bovine caseins in gram quantities was investigated. Whole casein was precipitated from individual cow's milk by adjusting the pH to 4·6 and the precipitated casein was suspended in 4·5 M urea (pH 8·0) containing 0·02 M imidazole and 0·03 M β-mercaptoethanol, and bound on a QAE Zeta Prep 250 cartridge. Stepwise elution with the urea/imidazole β-mercaptoethanol buffer and varying amounts of NaCl gave five well resolved peaks, which were identified by polyacrylamide gel electrophoresis and fast protein liquid chromatography to be pure γ-casein, κ-casein. β-casein, β-casein and αs-casein, respectively. The ion exchange cartridge was regenerated by flushing with buffer containing 0·50 Μ-NaCl followed by equilibration with starting buffer before separation of next sample. The time required to run each sample including cartridge regeneration and equilibration was 4 hours.

The phosphorylation of different caseins by a number of tyrosine-specific kinases has been studied. The different kinases showed a range of specificities, suggesting differences in recognition mechanisms. The site of phosphorylation of αs2-casein by one phosphorylase has been identified, and compared with analogous sites in other proteins.

The effect of Ca2+. binding and ionic strength on the physicochemical characteristics of phosphoserine residues was studied on O-phospho-DL-serine, bovine β-casein and its phosphopeptide (1–25) using 31P NMR. pK in various experimental conditions were determined.

The pK of the phosphoserine residues of β-casein and its phosphopeptide (1–25) respectively ranged from 6·46 to 7·21 and from 6·57 to 7·10. pK of O-phospho-DL-serine, β-casein and its phosphopeptide decreased when Ca2+ was bound to the phosphoserine residue or when ionic strength was increased. The binding of Ca2+ to the phosphopeptide (1–25) took place at first on phosphoserine residues 17, 18, 19 which had the highest pK, then, when these were saturated, on residue 15 whose pK was the lowest. When the four sites had bound Ca2+, peaks corresponding to a different complex form appeared in the spectrum.

Micellar calcium phosphate has the chemical composition and physicochemical properties that are consistent with it being a complex of the phosphate centres of casein with an acidic amorphous calcium phosphate. Similar acidic amorphous calcium phosphates have been prepared in the laboratory and for these, as well as for micellar calcium phosphate, the most appropriate crystalline model compound from which the short-range structure may be derived is brushite, CaHPO4.2H2O. The predicted secondary structures around sites of phosphorylation in the Ca2+-sensitive caseins often comprise an α-hclix-loop-α-helix motif with the sites of phosphorylation in the loop region. This motif may be important in linking the colloidal calcium phosphate with casein in native casein micelles.

The influence of pH, temperature and Ca depletion on bovine casein micelle suspensions in D2O containing simulated milk ultrafiltrate was studied by 1H-NMR spectroscopy.

In the pH range of 5·8–7·5 the spectrum of the micelles showed very little pH dependence, indicating that no changes occurred in the dynamic behaviour of the proteins constituting the micelle.

The NMR spectrum of casein micelles was strongly temperature dependent, particularly in the temperature range of 60–98 °C. Increase in temperature resulted in a strong increase in spectral intensity concomitant with changes in the spectral characteristics. In micelle suspensions these changes were reversible, and indicated that at clevated temperatures the rigid structure of the casein micelle started to melt, leading to an increased mobility of appreciable parts of the proteins in the micelle.

Ca depletion of the casein micelles by addition of EDTA resulted in an increase in spectral intensity, which arose from the presence of casein components in the serum phase, The spectrum of these serum phase particlcs resembled closely the spectrum of a solution of total casein in simulated milk ultrafiltrate and was quite different from the spectrum of casein micelles.

The implications of these results with respect to models of the structure of bovine casein micelles are discussed.

The association of non-phosphorylated (0-P) and fully phosphorylated (5-P) human β-caseins was studied by fluorescence spectroscopy and laser light scattering. The tryptophan fluorescence intensity (FI) level increased between 20 and 35 °C, indicating a change in the environment of that residue. A similar transition occurred when ANS was used as a probe. Transition temperatures were slightly lower in 10 mM-CaCl2 but were not affected by an equivalent increase in ionic strength caused by NaCl. The magnitude of the FI change was less for the 5-P than the 0-P protein but was increased for both by CaCl2 addition. These FI data were characteristic of a conformational change and this was supported by fluorescence polarization which indicated that with CaCl2, tryptophan and ANS mobility increased at the transition temperature even though the extent of protein association also increased. Light scattering suggested that protein association proeeeded with the primary formation of submicellar aggregates containing 20–30 monomers which then associated further to form particles of minimum micelle size (12–15 submicelles), and eventually larger. The temperature of precipitation of the 5-P form in the presence of CaCl2 was lower than the conformational transition and suggested that both hydrophobic interactions and Ca bridges between phosphate esters on adjacent molecules are important in micelle formation.

The effects of pressure on the association states of enzyme-treated casein molecules were studied by monitoring the turbidity of the solutions under pressures up to 3000 kg/cm2. β-Casein polymers, partly degraded with immobilized trypsin, dissociated with increasing pressure up to a critical pressure (e.g. 1200 kg/cm2) and then reassociated under the higher pressure up to 3000 kg/cm2, following a parabolic turbidity-pressure curve. In the case of chymosin-treated κ-casein, a similar pressure dependence of turbidity was found under the same pressure conditions. Based on these results, the pressure effects on dissociation and reassociation of casein molecules were discussed in terms of the volume change of water around their hydrophobic moieties.

Small-angle X-ray scattering was performed on whole casein under submicellar (Ca2+ removed) and micellar (Ca2+ re-added) conditions. Submicellar scattering curves showed two Gaussian components which were interpreted in terms of a spherical particle with two concentric regions of different electron density, a relatively compact core of higher electron density and a looser shell. Normalized scattering curves and calculated distance distribution functions were consistent with this picture. Micellar scattering data, which can yield only cross-sectional information related to a window of scattered intensities, could be analysed by a sum of three Gaussians with no residual function. The two Gaussians with the lower radii of gyration were again taken to indicate the two concentric regions of different electron density of inhomogeneous spherical particles; the third Gaussian was shown to reflect the packing number of these particles within a cross-sectional portion of the micelle, which was 3:1 for this system. These results are a strong indication that submicellar inhomogeneous particles containing hydrophobically stabilized inner cores exist within the colloidal micelle.

Casein micelles were reassembled from mixtures of F2 and F3 subunits, the major subunits of bovine casein micelles, in varying ratios. The average radius of the reassembled micelles was inversely proportional to the F2 content. Based on this relationship, a model for casein micelle assembly is proposed in which the F3 subunit forms the core of the micelle with F2 subunit occurring on the surface. Micelle size can subsequently be defined by the F2 content.

Hydrodynamic radii from inelastic light-scattering experiments and radii of gyration from Zimm plots give an indication of the change of average casein micelle size when the pH is changed. Combination of the results of both types of measurements gives information on changes in the micelle protein matrix, i.e. changes in the voluminosity.

The voluminosity was also determined by the pellet method and by electron microscopy which also provided comparative data on size parameters.