Preparation, physicochemical characterization and stability of iron-supplemented caseins were studied. When ferrous iron (final concentrations of 0·25, 0·5, 1·0 and 1·5 mM) was added to sodium caseinate (final casein concentration 25 g/l), iron binding to caseins was complete. In parallel, an increase in absorbance at 280 nm, a decrease in fluorescence intensity, modifications of reversed-phase HPLC profile, a decrease in pH and an increase in free sodium concentration were observed. Moreover, no release of iron from iron-casein complexes was found between pH 6·5 and 3·7 or after heat treatments or in the presence of inorganic phosphate. Only EDTA disodium salt and trisodium citrate had significant effects on the release of iron bound to caseins.

Mineral acid casein (MAC) curd was subjected to a range of ionic strength, pH and temperature conditions to explore their effects on curd water-holding capacity, particularly the reversibility of shrinkage. The shrinkage of MAC curd on whey removal during the washing stage of casein manufacture could be attributed almost completely to the reduction in ionic strength and not specifically to either calcium or lactose removal. This shrinkage on whey removal during washing could not be reversed by adding whey back to the curd to attain the original ionic environment. In this sense it was irreversible. The pH during formation of MAC curd was more important than later pH adjustments in controlling curd water-holding capacity. The effect of a high precipitation pH in creating a tough, compact curd could not be reversed by subsequently decreasing the pH. However, the effect of a low precipitation pH in creating a soft, open curd could be partly reversed by subsequently increasing the pH. MAC curd shrank by 45% when the temperature was increased from 20 to 80°C. This temperature-dependent shrinkage was virtually fully reversible on cooling. It is suggested that the water-holding capacity of casein curd is governed largely by hydrophobic interactions and that once such interactions have occurred they cannot easily be disrupted by small changes in either pH or ionic strength.

After extended tryptic hydrolysis of large bovine casein micelles, a mineral-rich peptide fraction was recovered by ultracentrifugation. Its mineral part contained 72% of the colloidal Ca and 49% of the colloidal Pi originally present in the native micelle. Colloidal nitrogenous components were also present, amounting to 27% of the original N content. They contained most of the phosphopeptides and 82% of the micellar phosphoseryl residues. These tryptic peptides were characterized by reversed-phase HPLC on-line electrospray ion source–mass spectrometry analysis. Among the peptides produced 14 phosphopeptides were identified: αs2-CN(l–24), αs2-CN(1–21), αs1-CN(43–79), αs1-CN(35–79)7P, αs1-CN(35–79)8P, αs1-CN(37–79), αs1-CN(104–119), αs1-CN(104–124), β-CN(1–25), β-CN(1–28), β-CN(1–29), β-CN(30–97), β-CN(33–97) and β-CN(29–97). The proportion of the phosphopeptides interacting with colloidal calcium phosphate was correlated with their relative content of phosphoserine residues, since phosphopeptides containing more than four phos-phoserine residues were consistently present within this fraction. It also appeared that other types of peptides, some of them hydrophobic in nature, were also partly or completely present within the colloidal fraction, including αs1-CN(91–100), αs1-CN(152–193), αs1-CN(23–34), αs1-CN(125–193), αs1-CN(125–199), β-CN(177–209), β-CN( 184–209), β-CN(114–169) and β-CN(108–169). Their possible involvement in the micellar backbone is discussed.

κ-Casein was purified from a single batch of whole acid casein (κ-A variant) using different methods in order to compare their merits in producing a purified material with a carbohydrate and phosphate heterogeneity representative of the whole κ-casein complement in milk. Ion-exchange methods of purification gave products of higher purity than precipitation techniques involving final purification by ethanol fractionation, but all methods resulted in κ-caseins of apparently similar heterogeneity and chemical composition. The purified κ-caseins were hydrolysed with chymosin and the derived macropeptides isolated. These were all virtually identical as determined by reversed-phase chromatography and gel electrophoresis. Some observations on chymosin hydrolysis of κ-casein were made. In addition to formation of the major para-κ-casein (Glu1–Phe105) and macropeptide (Met106–Val169), chymosin hydrolysis at pH 6·6 also resulted in two minor para-κ-caseins with N-termini corresponding to Phe18 and Ser33 of κ-casein. At pH 5·5 and 4·5 para-κ-casein was rapidly hydrolysed into at least six fragments, one of which had an N-terminus corresponding to Trp76 of κ-casein. At pH 6·6, 5·5 and 4·5 the κ-casein macropeptide was stable to chymosin, but at pH 2·3 it was hydrolysed by chymosin into fragments with N-termini corresponding to Met106, He125, Ala138, Val139, Thr145 and Glu147 of κ-casein.

In situ light scattering, where light scattered from a sample is measured directly while the sample is heated in the instrument, is presented as a simple and effective technique for studying the heat-induced aggregation of β-lactoglobulin. This technique was shown to be applicable not only to monitoring the initial aggregation steps, but to following the overall aggregation process with time. The experiments gave results similar to measurements carried out after a heat-quench treatment, but were more informative. From experiments on a standard NIZO β-lactoglobulin sample, a strongly desalted standard NIZO sample, different genetic variants of β-lactoglobulin and a mixture of these, we concluded that the standard NIZO sample was suitable for studying heat-induced aggregation. This sample has been investigated more extensively. Results with β-lactoglobulin (10–100 g/1) at 65 °C fitted a kinetic model for the denaturation and aggregation of β-lactoglobulin. This model, which held for β-lactoglobulin dissolved in water at near neutral pH and at 60–75 °C, recognizes an initiation, propagation and termination reaction, by analogy with polymer radical chemistry. It gave a quantitatively correct description of the dependence of the scattering intensity on the initial β-lactoglobulin concentration. Salt composition, pH and temperature strongly influenced the aggregation of β-lactoglobulin. Particle size increased with salt concentration in the range studied (up to 20 mM-NaCl and 1·0 mM-CaCl2). When the pH increased from 6°9 to 8·0 particle size was strongly reduced, whereas it strongly increased when pH was lowered to 6·2. Between 61·5 and 70 °C temperature did not affect particle size, whereas aggregation rate strongly increased. These effects could be incorporated in the kinetic model via the reaction constants of the reaction kinetic pathway.

A microparticle-enhanced nephelometric immunoassay has been developed for the determination of caseinomacropeptide (CMP) in bovine milk. It is based on the nephelometric quantification of the competitive immunoagglutination of a microparticle–CMP conjugate with an anti-κ-casein (κ-CN) antiserum. This one step immunoassay was sensitive (detection limit in reaction mixture, 16μg/l), accurate (linear recovery of CMP in dilution overloading) and reproducible (CV 7–14% for within and between run precision). Because of the specificity of the polyclonal antiserum used, it was necessary to separate CMP from κ-CN by ultrafiltration before the quantification of bovine milk CMP. Under the conditions of milk ultrafiltration used, κ-CN was entirely retained (> 99·5%) but the concentration of CMP measured in milk ultrafiltrates was underestimated (by ∼25%) compared with its concentration in whole milk. Microparticle-enhanced nephelometric immunoassay of CMP, with a calibration range from 0·32 to 20 mg/1 for 20- fold diluted milk ultrafiltrate, allowed contamination of bovine milk by rennet whey as low as 5 ml/1 to be detected. Applied to ultrafiltrates from milk stored at 4 °C, this immunoassay also detected proteolysis of κ-CN not revealed by measurement of κ-CN concentration in milk. A statistical lower limit of 3·21 mg/1 was determined as the increase in CMP concentration in milk ultrafiltrates that indicated probable κ-CN proteolysis in the milk sample. Previously demonstrated to be an easy to perform method for assaying the main proteins of bovine milk, microparticle-enhanced nephelometric immunoassay thus also appeared to be appropriate to quantify CMP so as to detect slight contamination of milk by whey and to indicate the proteolysis of κ-CN during milk storage at low temperature.

Milk fat globule membranes (MFGM) were isolated from milk from cows injected intraperitoneally with dl-α-tocopherol acetate. The fatty acid composition and content, and the contents of tocopherols, retinol and β-carotene were determined and compared with the composition and content in the original cream sample. Intraperitoneal injection of 10 g dl-α-tocopherol acetate elevated the α-tocopherol content in the milk fat from 13–30 to 50–70 εg α-tocopherol/g milk fat 2–3 d after injection. The increase depended on the α-tocopherol status of the cow prior to injection. The concentrations of retinol and β-carotene in the milk fat were unchanged after the α-tocopherol injections. MFGM fatty acids made up 18–27 g/kg total fatty acids in the milk fat. However, the proportions of monounsaturated and polyunsaturated fatty acids were higher in MFGM than in total milk fat, while the proportion of saturated fatty acids was lower in MFGM (P < 0·001). Thus, the longchain polyunsaturated fatty acids in MFGM constituted ∼ 40–70 g/kg total milk fat. α-Tocopherol was the only fat-soluble vitamin detected in MFGM; γ-tocopherol, retinol and β-carotene were detected only in the cream. A significant relationship between α-tocopherol contents in cream and MFGM was found: (α-tocopherol in MFGM fatty acids) =–1652 × (1/ln(α-tocopherol in cream fatty acids)3)+ 97 (r = 0·857, P < 0·001) when the contents of α-tocopherol were expressed as εg/g fatty acid. According to this equation the upper limit for incorporation of α-tocopherol into MFGM is 97±5 εg/g MFGM fatty acids. Thus, under normal farming conditions the α-tocopherol content will be highest in MFGM. However, in cream with an α-tocopherol concentration < 15 εg/g cream fatty acids the concentration in MFGM will be lower and furthermore will decrease rapidly. The possible importance of this relationship in relation to the oxidation of milk fat is discussed.

Monoclonal antibodies against peptidases of Lactococcus lactis were isolated and characterized: PEPN1–4 against a lysyl aminopeptidase PepN, PEPT1–5 against a tripeptidase PepT and PEPD1–3 against a dipeptidase PepD. These monoclonal antibodies reacted specifically with their respective antigens in crude cell extracts of Lc. lactis subspp. cremoris and lactis. A number of monoclonal antibodies cross reacted with proteins of other (lactic acid) bacteria. PEPT1, 2, 4 and 5 cross reacted weakly with a 35 kDa protein in Lactobacillus delbrueckii, while PEPT1 and PEPT2 reacted with proteins in the cell-free extract of Streptococcus thermophilus and Clostridium fervidus. Of the four isolated monoclonal antibodies against PepN, only PEPN3 cross reacted weakly with a 90 kDa protein in Escherichia coli cell-free extract, and the other three antibody species against PepN cross reacted with 80 kDa proteins of Lb. casei, Lb. delbrueckii, and Str. bovis, but not of Esch. coli. Of the three monoclonal antibodies against PepD, only PEPD1 and PEPD2 cross reacted with 40 kDa proteins of Lb. casei, Lb. delbrueckii and Str. bovis. All PEPN, PEPD and PEPT antibodies reacted with components in cell-free extracts of eleven different Lc. lactis strains, indicating that the peptidases of these strains were very similar to those of Lc. lactis subsp. cremoris WG2. However, Lc. lactis subsp. hordniae appeared to differ from the other Lc. lactis subspecies since only PEPT1, 2 and 5 reacted with a protein in the cell-free extract. Immunogold labelling of Lc. lactis WG2 with the isolated monoclonal antibodies revealed that PepN, PepD and PepT were located intracellularly. The intracellular location of these peptidases is discussed in relation to the supply of essential amino acids and peptides.

The heat-induced binding of whey proteins to milk fat globule membranes in whole milk was investigated by quantitative electrophoresis and laser scanning densitometry. Both α-lactalbumin and β-lactoglobulin bound to the surfaces of fat globules when milk was heated in a water bath in the temperature range 65–85 °C. The interaction behaviour of α-lactalbumin did not seem to change with temperature, and the total amount of protein bound was ∼ 0·2 mg/g fat contained in the cream. The quantity of βlactoglobulin interacting with the milk fat globules increased with temperature from 02 to 0·7 mg/g fat between 65° and 85 °C. Even in whole milk heated at batch pasteurization temperatures (60–65 °C), α-lactalbumin and β-lactoglobulin were found attached to the fat globules. The interactions of the whey proteins with intact fat globule membranes were also investigated in milk heated in an industrial system (a pilot scale UHT and high temperature short time module), and the results were compared with those from the laboratory treatment (simple batch heating). The binding of the whey proteins to fat globules differed between milk heated by UHT using indirect steam heating or direct steam injection (DSI). However, the surface load in milk treated by DSI was not comparable to that of milk treated by batch heating or indirect steam heating, because of the changes in fat globule size and membrane composition caused by the DSI process.

Studies on the immunomodulatory activities of ruminant milk and colostral whey fractions were undertaken. By comparing with boiled colostral whey in a preliminary experiment, a putative heat-labile immunostimulatory factor for antibody responses was found to be present in ovine colostral whey. Studies were then undertaken in sheep in which the efferent prefemoral lymphatic ducts were cannulated bilaterally, and immune responses in the node were measured following subcutaneous injection in the flank fold of whey protein preparations of various purities. A significant sustained decline of efferent lymphocyte output was observed following injection with autologous crude milk whey or colostral whey preparations, but no changes were observed in interferon-gamma levels in lymph plasma. Two bovine milk whey fractions (lactoperoxidase and lactoferrin) of high purity were compared in bilaterally cannulated sheep. A transient decline over the first 6 h was seen in the efferent lymphocyte output and lymph flow rate after injection of both fractions. A significant difference was seen between the two fractions in interferongamma levels in lymph at 6 h after injection. However, no significant changes in the proportion of the various efferent lymphocyte phenotypes were seen following either treatment. Whereas both fractions showed a significant inhibitory effect in a dose-dependent manner on the proliferative response of T lymphocytes, but not B lymphocytes, to mitogenic stimulation in vitro, no similar changes were seen following in vivo stimulation with these two fractions.

The water-holding capacities of mineral acid casein (MAC) curd, rennet casein curd and whey protein concentrate (WPC) gel were determined as a function of pH and temperature by drainage of liquid on a stationary, horizontal screen. As precipitation pH was varied over the range 4·3–5·55, the water-holding capacity of MAC curd ranged from 2·4 to 5·2 g water/g dry casein with minimum water-holding capacity at precipitation pH 5·3. Samples of dried MAC curd over the same range of precipitation pH did not vary significantly in water-holding capacity, and all absorbed ∼2·0 g water/g dry casein. Both rennet casein curd and WPC (170 g total solids/kg) gels increased in water-holding capacity as the pH of the suspending liquid increased in the range 3·5–10, but the changes with pH were much smaller than for MAC curd. For both types of casein curd and for WPC gel, water-holding capacity decreased with increase in temperature in the range 5–90 °C. The results are interpreted largely in terms of changes in electrostatic interactions with pH and changes in hydrophobic interactions with temperature.

The milk whey protein β-lactoglobulin has been isolated from ovine milk, purified and crystallized in a form suitable for X-ray crystallography. The crystals, which diffract to 1·9 Å resolution, belong to the trigonal space group P3121 with unit cell lengths a = b = 99·8 Å, c = 67·7 Å and unique angle γ = 120·0°. Although there have been many crystal forms reported for bovine β-lactoglobulin, none has been reported with these unit cell parameters. A preliminary native data set has been collected and a molecular replacement solution obtained, using the structure of dimeric bovine β-lactoglobulin as the search model. The importance of the ovine structure in relation to that of the bovine is discussed.

Fourier transform infrared spectroscopy and differential scanning calorimetry were used as complementary techniques to study changes in the secondary structure of β-lactoglobulin under various physicochemical conditions. The effects of pH (3–9), NaCl (0–2 M), and lactose, glucose and sucrose (100–500 g/l) in the temperature range 25–100 °C on the conformation sensitive amide I band in the i.r. spectrum of β-lactoglobulin in D2O solution were examined. The 1692 cm−1 band in the amide I band profile had not been definitively assigned in previous studies of the i.r. spectrum of β-lactoglobulin. The decrease in this band at ambient temperature with time or upon mild heating was attributed to slow H-D exchange, indicating that it was due to a structure buried deep within the protein. The disappearance of the 1692 cm−1 band on heating was accompanied by the appearance of two bands at 1684 and 1629 cm–1, assigned to β-sheets. The 1692 cm−1 band was therefore attributed to a β-type structure. β-Lactoglobulin showed maximum thermal stability at pH 3 and was easily denatured at pH 9. On denaturation, the protein unfolded into more extensive random coil structures at pH 9 than at pH 3. After 10 h at pH 9 (25 °C), β-lactoglobulin was partly denatured. Heating to 60–80 °C generally resulted in the loss of secondary structure. At all pH values studied, two new bands at 1618 and 1684 cm−1, characteristic of intermolecular β-sheet structure and associated with aggregation, were observed after the initial denaturation. Differential scanning calorimetry studies indicated that the thermal stability of β-lactoglobulin was enhanced in the presence of sugars. The Fourier transform i.r. results obtained provide evidence that sugars promoted the unfolding of β-lactoglobulin via multiple transition pathways leading to a transition state resisting aggregation.

Conversion of plasminogen to plasmin provides an important source of proteolytic activity in the bovine mammary gland. Plasminogen activator inhibitor 1 (PAI-1) plays a key role in limiting plasminogen activation. The PAI-1 biosynthetic capabilities of various bovine mammary cells were determined. The immortalized epithelial cell lines MAC-T and BME-UV and the myoepithelial cell line BMM-UV were used as model systems. Northern blot analysis indicated that both epithelial and myoepithelial cells contained PAI-1 mRNA. Bovine PAI-1 was encoded by a single mRNA species ∼ 3·0 kb long. BME-UV cells contained 2·0-fold (P < 0·01) the PAI-1 mRNA of MAC-T or BMM-UV cells. Reverse zymography indicated that both epithelial and myoepithelial cells synthesized PAI-1 protein with a molecular mass of ∼ 50 kDa.

Nine organisms were isolated from separate pasteurized milk samples after they had been incubated at 6 °C for 5 d (European Union preincubated count, PIC), four from high count samples (> 5 x 106 cfu/ml) and five from low count samples (< 103 cfu/ml). When the organisms were harvested without overt stress being applied and subjected to a simulated PIC using UHT whole milk, all except one isolate gave comparatively high (> 106 cfu/ml) counts. The imposition of a heat stress at 50 °C prior to a simulated PIC resulted in a segregation of the isolates into those giving high and those giving low counts, which reflected the PIC values of the milk samples from which they were originally isolated. When the isolates were subjected to a cold stress (25 to 4 °C) and inoculated into nutrient broth at 4 °C, the high count isolates were found to have significantly (P < 0·05) shorter lag phases than the low count isolates.

The effects of somatic cell count and stage of lactation on the yield and quality of Cheddar cheese were investigated. Cheese was manufactured in a pilotscale factory using milk of low bulk milk cell count (BMCC) from herds in early (LE) and late (LL) lactation, and milk of high BMCC from herds in early (HE) and late (HL) lactation. The deleterious effect of an elevated BMCC on product yield and quality in late lactation was clear. Cheese made from LL milk was significantly superior to that made from HL milk for most yield and quality characteristics measured. Stage of lactation also affected cheese yield and quality, as evidenced by the lower recovery of fat and poorer flavour score for cheese from LL milk compared with that manufactured from LE milk. The observed differences could be explained largely by differences in raw milk composition. We conclude that the effect of stage of lactation was magnified by an elevated BMCC, and that many of the problems encountered when processing late season milk could be overcome by containing mastitis at this time.

Skim milks were concentrated by ultrafiltration. Cream was added to the retentates to give cheesemilks standardized to a casein: fat ratio of ∼ 0·74 with protein levels ranging from 30 to 46 g/1. Pasteurized cheesemilks were renneted on a volume basis (22 ml single strength calf rennet/100 1) and converted to Cheddar cheese in 500 1 cheese vats. Set temperatures were reduced from 31 to 28 °C with increasing milk protein level to normalize curd firming rates and prevent curd shattering on cutting the curds. Cheesemaking was otherwise as normal for Cheddar cheese. The proportions of milk fat and protein lost in the cheese whey were not significantly influenced by milk protein level. Moisture-adjusted cheese yields increased with milk protein at a rate similar to that predicted by the Van Slyke cheese yield equation. However, owing to the negative correlation between cheese moisture and milk protein concentration, actual yields increased at a lower rate with respect to milk protein than moisture-adjusted yields. Increasing milk protein levels resulted in significant (P < 0·01–0·001) decreases in the concentrations of moisture, moisture-in-non-fat cheese solids and fat-in-dry matter in the cheese and increases (P < 0·05–0·001) in the levels of protein, salt-in-moisture, Ca and P. While increasing milk protein concentration resulted in significant (P < 0·05–0·01) reductions in the levels of water-soluble N at all stages of the 270d ripening period, it had little influence on the sensory scores awarded for aroma/flavour at 180 or 270d.

Listeria monocytogenes and List. innocua were isolated from commercial soft ripened and blue-veined cheeses manufactured in France, mainly from Brie cheese made from unpasteurized milk. Five isolates were List. monocytogenes serotype 1/2 and two were List. innocua. Examination of Bleu d'Auvergne cheese with the cryoscanning electron microscope showed that many conidia spores were present in the blue veins in close contact with the cheese surface. There were few conidia spores in the Brie, mostly on the outside of the cheese but not in contact with the surface. High concentrations of free dodecanoic (lauric) acid (1·77–2·50 g/kg cheese) and tetradecanoic (myristic) acid (2·54–6·38 g/kg cheese) were found in the veins of the blue cheese, but concentrations in the white regions were much lower. Free lauric and myristic acids were not detected in the Brie cheeses. There was no difference in the overall fatty acid composition of the fat in the surface ripened and blue-veined cheeses, although higher concentrations of free medium-chain fatty acids were found in a blue cheese compared with a surface ripened cheese. The pH and fat content were higher in regions with obvious fungal growth, the blue veins of Fourme d'Ambert and the rind of Brie. Free lauric acid dissolved in butteroil inhibited multiplication in broth at pH 7·0 of a test strain of List. monocytogenes isolated from Bleu d'Auvergne. Some inhibition was seen with hexanoic, octanoic, decanoic and tetradecanoic acids. We suggest that the presence of localized concentrations of free medium-chain fatty acids (dissolved in the fat) in the blue veins of blue mould ripened cheese could act as natural preservatives and inhibit the growth of listerias in conditions where (if present), one would otherwise expect them to grow.

Hydrolysis of β-lactoglobulin B (β-lg B) by pepsin, a process slow at ambient conditions, is facilitated at a moderately high hydrostatic pressure such as 300 MPa, corresponding to an apparent volume of activation ΔV# = −63 ml mol−1 at pH 2·5, 30 °C and Γ/2=0·16. Digestion of β-lg by trypsin and thermolysin is likewise enhanced by pressure, and the pressure effect has been traced to pressure denaturation of β-lg B, which by high-pressure fluorescence spectroscopy has been shown to have a large negative volume of reaction, ΔV° = −98 ml mol−1, at pH 6·7, 30 °C and Γ/2 = 0·16. Pressure denaturation is only slowly reversed following release of pressure and the enhanced digestibility is maintained at ambient pressure for several hours.

The aim of this work was to develop new oriental style dairy products coagulated with culture filtrates from lao-chao, a traditional fermented rice product. The fermenting fungal cultures screened were Rhizopus javanicus, Rhiz. oryzae, Rhiz. chinensis, Rhiz. oligosporus, Aspergillus oryzae, Mucor racemosus and chiu-yao, a commercial starter. All the inocula liquefied a steamed glutinous rice base and produced culture filtrates with both milk-clotting and proteolytic activities. Of the fungal products tested, Rhiz. javanicus was the most successful, producing a good yield of culture filtrate with the desired milk-clotting activity, and the resultant yogurt-like product was more acceptable to consumers than any except that from chiu-yao. The firmness was less acceptable, but this could be improved by using mixed pure cultures of Rhiz. javanicus and M. racemosus.