Many psychrotrophic bacteria contaminating raw milk produce phospholipase that withstands pasteurization and UHT treatments. This enzyme acts on the milk fat globule membrane and exposes triacylglycerides to the action of lipase. Phospholipase production by various isolates of Pseudomonas was investigated. The isolates were cultured aerobically at 8 °C in nutrient broth, McKellar's minimal salts medium, Chrisope's medium, and skim milk. Each strain produced phospholipase during the 50 h incubation. Enzyme production varied significantly (P < 0·001) with strain and growth medium. Strains varied significantly (P < 0·001) in their enzyme production in each medium and during the incubation time as well. Strain, incubation time, and the growth medium significantly influenced (P < 0·001) heat stability of the enzyme activity. Pasteurization reduced the activity, but did not eliminate it in skim milk.

Cathepsin D, the principal indigenous acid proteinase in bovine milk, is a lysosomal proteinase, which exists in milk in four forms, including the inactive zymogen procathepsin D. The thermal inactivation kinetics of bovine cathepsin D, isolated from spleen and milk, were studied under isothermal conditions, using a specific HPLC assay to determine residual activity. Inactivation of the blood enzyme preparation followed first order kinetics, with z-values in phosphate buffer (pH 6·7) and skimmed milk of 6·5 and 7·6 °C, respectively, the enzyme being far more stable in the latter environment. Inactivation kinetics of the enzyme purified from milk were more complex, and could be best approximated by a double exponential model. Again, stability was higher in milk than in buffer. The double exponential model may indicate differing heat stabilities of isoforms of the enzyme, or stabilization of the enzyme by some milk constituent. It is clear that the enzyme can survive, at least partially, processes such as heating at 55 °C for 30 min during manufacture of high-cook cheese varieties (45% survival), and HTST pasteurization (8% survival), and thus may contribute to proteolysis in a range of dairy products.

At atmospheric pressure, inactivation of lactoperoxidase (LPO) in milk and whey was studied in a temperature range of 69–73 °C and followed first order kinetics. Temperature dependence of the first order inactivation rate constants could be accurately described by the Arrhenius equation, with an activation energy of 635·3±70·7 kJ/mol for raw bovine milk and 736·9±40·9 kJ/mol for diluted whey, indicating a very high temperature sensitivity. On the other hand, LPO is very pressure resistant and not or only slightly affected by treatment at pressure up to 700 MPa combined with temperatures between 20 and 65 °C. Both for thermal and pressure treatment, stability of LPO was higher in milk than in diluted whey. Besides, a very pronounced antagonistic effect between high temperature and pressure was observed, i.e. at 73 °C, a temperature where thermal inactivation at atmospheric pressure occurs rapidly, application of pressure up to 700 MPa exerted a protective effect. At atmospheric pressure, LPO in diluted whey was optimally active at a temperature of about 50 °C. At all temperatures studied (20–60 °C), LPO remained active during pressure treatment up to 300 MPa, although the activity was significantly reduced at pressures higher than 100 MPa. The optimal temperature was found to shift to lower values (30–40 °C) with increasing pressure.

We studied the effects of goat and cow milk fat on the digestive utilization of this nutrient and on some of the biochemical parameters that are related to the metabolism of lipids, using rats with a resection of 50% of the distal small intestine and control animals (transected). The fat content in all the diets was 10% but the lipid quality was varied: the standard diet was based on olive oil, while the other two diets included fat obtained from lyophilized goat milk and cow milk, respectively. The digestive utilization of the fat was lower in the resected animals than in the transected ones for all three diets studied. In both resected and transected animals, the apparent digestibility coefficient (ADC) of the fat was greater with the standard diet (olive oil) than with diets whose fat content was provided by goat or cow milk. The digestive utilization of the fat was greater in the transected and resected rats receiving a diet of goat's milk (rich in medium-chain triglycerides) than those given a cow-milk-based diet, and more closely approached the values obtained for olive oil. The consumption of goat milk reduced levels of cholesterol while levels of triglycerides, HDL, GOT and GPT remained within the normal ranges, for both transected and resected animals. The advantageous effect of goat milk on the metabolism of lipids with respect to cow milk suggests that the former should be included in the diet in cases of malabsorption syndrome.

Chemical and microbiological composition of four Argentinean kefir grains from different sources as well as characteristics of the corresponding fermented milk were studied. Kefir grains CIDCA AGK1, AGK2 and AGK4 did not show significant differences in their chemical and microbiological composition. In contrast, protein and yeast content of AGK3 was higher than in the other grains. Although grain microflora comprised lactobacilli, lactococcus, acetic acid bacteria and yeast, we found an important difference regarding species. Lactococcus lactis subsp. lactis, Lactobacillus kefir, Lactobacillus plantarum, Acetobacter and Saccharomyces were present in all types of kefir grain. While Leuconostoc mesenteroides was only isolated from grains CIDCA AGK1 and Lactococcus lactis subsp. lactis biovar diacetylactis, Lactobacillus parakefir and Kluyveromyces marxianus were only isolated from CIDCA AGK2 grains. All grains produced acid products with pH between 3·5 and 4·0. The apparent viscosity of AGK1 fermented milk was greater than the product obtained with AGK4. All fermented milks had inhibitory power towards Escherichia coli but AGK1 and AGK2 supernatants were able to halt the bacterial growth for at least 25 h. Grain weight increment in AGK1, AGK2 and AGK3 during growth in milk did not show significant differences. Despite their fermenting activity, AGK4 grains did not increase their weight.

Studies on camels' milk, whether with respect to concentration or bioavailability of trace elements from this milk, are limited and warrant further investigation. The object of this study was to analyse the concentration and distribution of zinc, copper, selenium, manganese and iron in camel milk compared to those in human milk, cows' milk and infant formula under similar experimental conditions. Camels' milk and cows' milk were collected from local farms, human milk samples were obtained from healthy donors in Kuwait and infant formula was purchased locally. Milk fractionation was performed by ultra-centrifugation and gel-column chromatography. The concentration of trace elements was analysed by atomic absorption spectrometry and that of protein was determined spectrophotometrically. The concentration of manganese and iron in camels' milk was remarkably higher (7–20-fold and 4–10-fold, respectively) than in human milk, cows' milk and infant formula. The zinc content of camels' milk was higher than that of human milk but slightly lower than in cows' milk and infant formula. The concentration of copper in camels' milk was similar to that of cows' milk but lower than in human milk and infant formula. The selenium content of camels' milk was comparable to those of other types of milk. Approximately 50–80% of zinc, copper and manganese in camels' milk were associated with the casein fraction, similar to that of cows' milk. The majority of selenium and iron in camels' milk was in association with the low molecular weight fraction. It is recommended that camels' milk be considered as a potential source of manganese, selenium and iron, perhaps not only for infants, but also for other groups suspected of mild deficiency of these elements. Further investigations are required to confirm this proposal.

Hispánico cheese, a semi-hard Spanish variety, was manufactured from a mixture of pasteurized cows' and ewes' milks (4[ratio ]1) using a commercial mesophilic LD-type starter comprising Lactococcus lactis subsp. cremoris, Lc. lactis subsp. lactis, Lc. lactis subsp. lactis var diacetylactis and Leuconostoc mesenteroides subsp. cremoris. Varying amounts (0–1·0 g/kg) of an Enterococcus faecalis INIA 4 culture in milk were added as a bacteriocin-producing adjunct. Differences in pH between cheeses manufactured with and without the bacteriocin producer did not exceed 0·11 pH units. Starter lactococci lost viability more rapidly in cheeses made with the bacteriocin producer, which reached counts of up to 6×107 cfu/g during ripening. Aminopeptidase activity in 1-d-old cheese made from milk inoculated with 1·0 g bacteriocin-producing culture/kg was twice that in control cheese. Degrees of overall proteolysis and levels of total free amino acids in 45-d-old cheese made with 1·0 g bacteriocin-producing culture/kg were 1·80-fold and 2·17-fold those in control cheese of the same age. Inoculating milk with 1·0 g/kg bacteriocin-producing culture reduced the level of hydrophobic peptides in the resultant cheese, increased the concentrations of 3-methyl-1-butanal, diacetyl and acetoin, and resulted in the highest scores for flavour quality and flavour intensity throughout ripening.

Gelation of α-lactalbumin (α-la) incubated with a protease from Bacillus licheniformis (BLP) at 50 °C for 4 h was monitored using small oscillatory shear and the large deformation properties of final gels were characterized by uniaxial compression. Transmission electron microscopy was used to visualize the microstructure. Gels made from α-la (10 g/l) using BLP were almost transparent, although somewhat whitish, and they were more than 20 times stiffer (measured as complex modulus) than equivalent gels made from β-lactoglobulin (β-lg) at the same concentration. The microstructure of the gels consisted of non-branching, apparently hollow strands with a uniform diameter close to 20 nm, similar in overall structure to microtubules. Adding Ca2+ in amounts of 50 or 100 mM changed the spatial distribution of the strands and resulted in a reduction in the failure stress recorded in uniaxial compression. Apart from affecting the microstructure, Ca2+ was shown to be essential for the formation of the gels. It is proposed, that the mechanism behind the self-assembly of the partially hydrolysed α-la into long tubes is a spatially restricted creation of ionic bonds between Ca2+ and carboxyl acid groups on peptide fragments resulting from the action of BLP on α-la. Proteolysis of α-la with BLP in the presence of Ca2+ thus results in formation of a strong gel with a microstructure not previously observed in food protein systems.

The ability of Listeria monocytogenes to grow at low temperatures and increase to hazardous levels during storage of dairy products is of great concern to the dairy industry. List. monocytogenes was detected in 3·6% of 774 cows' milk samples from 114 farms in Central Spain (Gaya et al. 1998). Ewes' and goats' milks produced in Spain had contamination levels of 2·2 and 2·6% respectively (Rodríguez et al. 1994; Gaya et al. 1996). List. monocytogenes may survive during the manufacture and ripening of most cheese varieties, with higher risk in soft surface-ripened cheeses owing to the rise in pH during ripening (Maisnier-Patin et al. 1992; Giraffa & Carminati, 1997).

The efficacy with which actively growing bacteriocin-producing cultures inhibit List. monocytogenes in some cheese varieties has been demonstrated. Nisin-producing starters inhibited List. monocytogenes V7 in Camembert cheese (Maisnier-Patin et al. 1992) and List. innocua BL86/26 in raw ewes' milk Manchego cheese (Rodríguez et al. 1998). A lacticin 3147-producing starter culture (McAuliffe et al. 1999) reduced List. monocytogenes Scott A levels in cottage cheese. Inoculation of milk with enterocin AS-48-producing Enterococcus faecalis INIA 4 inhibited List. monocytogenes strains Ohio and Scott A during the manufacture and ripening of Manchego cheese (Nuñez et al. 1997).

Semi-hard cheese varieties from raw or pasteurized cows' milk mixed with different proportions of ewes' and/or goats' milk represent ∼ 50% of Spanish cheese production. The objective of the present work was to evaluate the antilisterial activity of three bacteriocin-producing lactic acid bacteria strains isolated from milk (Rodríguez et al. 2000) during the manufacture and ripening of a semi-hard cheese manufactured from raw milk.

The aim of the present work was to investigate the role of whey protein denaturation on the acid induced gelation of casein. This was studied by determining the effect of whey protein denaturation both in the presence and absence of casein micelles. The study showed that milk gelation kinetics and gel properties are greatly influenced by the heat treatment sequence. When the whey proteins are denatured separately and subsequently added to casein micelles, acid-induced gelation occurs more rapidly and leads to gels with a more particulated microstructure than gels made from co-heated systems. The gels resulting from heat-treatment of a mixture of pre-denatured whey protein with casein micelles are heterogeneous in nature due to particulates formed from casein micelles which are complexed with denatured whey proteins and also from separate whey protein aggregates. Whey proteins thus offer an opportunity not only to control casein gelation but also to control the level of syneresis, which can occur.

Fermentation of deproteinised whey with kefir grains CIDCA AGK1 was studied focusing on polysaccharide production from lactose. Kefir grains were able to acidify whey at different rates depending on the grain/whey ratio. During fermentation, kefir grains increased their weight and a water-soluble polysaccharide was released to the media. Exopolysaccharide concentration increased with fermentation time, reaching values of 57·2 and 103·4 mg/l after 5 days of fermentation in cultures with 10 and 100 g kefir grains/l, respectively. The polysaccharide fraction quantified after fermentation corresponded to the soluble fraction, because part of the polysaccharide became a component of the grain. Weight of kefir grains varied depending on the time of fermentation. Polysaccharide production was affected by temperature. Although the highest concentration of polysaccharide in the media was observed at 43 °C at both grain/whey ratios, the weight of the grains decreased in these conditions. In conclusion, kefir grains were able to acidify deproteinised whey, reducing lactose concentration, increasing their weight and producing a soluble polysaccharide.

A detailed kinetic study of hydroxymethylfurfural, lactulose and furosine formation was performed upon heating milk at temperatures between 90 °C and 140 °C. In case of prolonged heating, formation kinetics could be described by a fractional conversion model. Considering only the first phase of the model, kinetics could be simplified to a pseudo-zero order model. A first assessment of kinetic parameters was made by isothermal experiments. Data were analysed using both a 2-step linear and a 1-step non-linear regression method. Only for furosine, did the global 1-step regression approach seem to give better results than the individual 2-step regression approach. Next, the estimated parameters kref and Ea were re-evaluated under non-isothermal conditions by subjecting milk to a time variable temperature profile. Given the complexity of Maillard reaction, it seemed better to estimate kinetic parameters under non-isothermal conditions when using a simplified model. Formation of hydroxymethylfurfural, lactulose and furosine was characterized by an Ea value of 90·2 kJ/mol (k110 °C = 1·2 μmol/l, min), 99·1 kJ/mol (k110 °C = 51·5 mg/l, min) and 88·7 kJ/mol (k110 °C = 16·3 mg/100 g protein, min) respectively. Additionally, 90% joint confidence regions were constructed in order to obtain an accurate representation of the statistical confidence associated with the simultaneously estimated parameters.

Lactic acid bacteria (LAB) play an important role in food fermentation, as the products obtained with their aid are characterized by hygienic safety, storage stability and attractive sensory properties. A major aim of the research in this field is the selection of LAB strains that could be used. Hence it is very important to be able to apply a reliable method to distinguish a particular strain specifically and unambiguously, which allows studies of population dynamics of mixed cultures and monitoring starter strains during fermentation (Ramos & Harlander, 1990).

Molecular methods are a powerful alternative to the traditional differentiation of bacteria. A highly reproducible method for characterizing and distinguishing closely related strains, is represented by REA-PFGE (restriction endonuclease analysis by pulsed-field gel electrophoresis) performed by infrequently cutting endonucleases. Genetic differentiation of strains in several species of LAB has been successfully performed by this technique (Moschetti et al. 1997; Villani et al. 1997) obtaining very clear and reproducible restriction patterns (Moschetti et al. 1998).

In this study selected inoculated strains (lactococci or lactobacilli) were monitored by REA-PFGE during the whole process of water-buffalo ‘cacioricotta’ cheese-making. This product is a typical and traditional cheese in southern Italy produced from cow, goat, ewe or water-buffalo milk. The original technology of this preparation permits the recovery of whey proteins due to the high heat treatment of whole milk employed, allowing interesting yields to be achieved in terms of cheese. The use of starters is not common in traditional technology but low acid protection of the final product suggested the use of LAB as starter (Emaldi et al. 1987).

Volatile sulphur compounds are major flavouring compounds in many traditional fermented foods including cheeses. These compounds are products of the catabolism of L-methionine by cheese-ripening microorganisms. The diversity of L-methionine degradation by such microorganisms, however, remains to be characterized. The objective of this work was to compare the capacities to produce volatile sulphur compounds by five yeasts, Geotrichum candidum, Yarrowia lipolytica, Kluyveromyces lactis, Debaryomyces hansenii, Saccharomyces cerevisiae and five bacteria, Brevibacterium linens, Corynebacterium glutamicum, Arthrobacter sp., Micrococcus luteus and Staphylococcus equorum of technological interest for cheese-ripening. The ability of whole cells of these microorganisms to generate volatile sulphur compounds from L-methionine was compared. The microorganisms produced a wide spectrum of sulphur compounds including methanethiol, dimethylsulfide, dimethyldisulfide, dimethyltrisulfide and also S-methylthioesters, which varied in amount and type according to strain. Most of the yeasts produced methanethiol, dimethylsulfide, dimethyldisulfide and dimethyltrisulfide but did not produce S-methylthioesters, apart from G. candidum that produced S-methyl thioacetate. Bacteria, especially Arth. sp. and Brevi. linens, produced the highest amounts and the greatest variety of volatile sulphur compounds including methanethiol, sulfides and S-methylthioesters, e.g. S-methyl thioacetate, S-methyl thiobutyrate, S-methyl thiopropionate and S-methyl thioisovalerate. Cell-free extracts of all the yeasts and bacteria were examined for the activity of enzymes possibly involved in L-methionine catabolism, i.e. L-methionine demethiolase, L-methionine aminotransferase and L-methionine deaminase. They all possessed L-methionine demethiolase activity, while some (K. lactis, Deb. hansenii, Arth. sp., Staph. equorum) were deficient in L-methionine aminotransferase, and none produced L-methionine deaminase. The catabolism of L-methionine in these microorganisms is discussed.

Bovine β-lactoglobulin (β-lg), α-lactalbumin (α-la) and bovine serum albumin (BSA), dispersed in ultrafiltration permeate, that had been prepared from whey protein concentrate solution (100 g/kg, pH 6·8), were heated at 75 °C. The consequent protein aggregation was studied by one-dimensional (1D) and two-dimensional (2D) polyacrylamide gel electrophoresis (PAGE). When 100 g β-lg/kg permeate solution was heated at 75 °C, cooled and examined, large aggregates were observed. These aggregates were partially dissociated in SDS solution to give monomers, disulphide-bonded dimers, trimers and larger aggregates. When mixtures of β-lg and α-la or BSA were heated, homopolymers of each protein as well as heteropolymers of these proteins were observed. These polymer species were also observed in a heated mixture of the three proteins. Two-dimensional PAGE of mixtures demonstrated that these polymers species contained disulphide-bonded dimers of β-lg, α-la and BSA, and 1:1 disulphide-bonded adducts of α-la and β-lg, or BSA. These results are consistent with a mechanism in which the free thiols of heat-treated β-lg or BSA catalyse the formation of a range of monomers, dimers and higher polymers of α-la. It is likely that when whey protein concentrate is heated under the present conditions, BSA forms disulphide-bonded strands ahead of β-lg and that α-la aggregation with β-lg and with itself is catalysed by the heat-induced unfolded BSA and β-lg.

In the present study, 124 enterococcal strains, isolated from traditional Italian cow, goat and buffalo cheeses, were characterized using phenotypic features and randomly amplified polymorphic DNA polymerase chain reaction (RAPD–PCR). The RAPD–PCR profiles obtained with four primers and five different amplification conditions were compared by numerical analysis and allowed an inter- and intraspecific differentiation of the isolates. Whole-cell protein analysis by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) was used as a reference method for species identification. The strains were identified as Enterococcus faecalis (82 strains), E. faecium (27 strains), E. durans (nine strains), E. gallinarum (four strains) and E. hirae (two strains). Species recognition by means of RAPD–PCR was in agreement with the SDS–PAGE results except for eight strains of E. faecium that clustered in separated groups. On the other hand, phenotypic identification based on carbohydrate fermentation profiles, using the rapid ID 32 STREP galleries, gave different results from SDS–PAGE in 12·1% of the cases. The majority of the strains had weak acidifying and proteolytic activities in milk. One E. faecium strain showed vanA (vancomycin resistance) genotype while four strains showed a β-haemolytic reaction on human blood. Several strains showed antagonistic activity towards indicator strains of Listeria innocua, Clostridium tyrobutyricum and Propionibacterium freudenreichii subsp. shermanii.

Rheological properties of reconstituted skim milk coagulated with plant coagulants Cynara cardunculus L., Cynara humilis L. and chymosin was monitored by dynamic low amplitude oscillation. There are no published reports on the modelling of the gelation behaviour of milk by plant coagulants. Three mathematical models, Scott Blair, Douillard and Carlson, were fitted to the storage modulus (G′) as function of time curves. For all coagulants, Scott Blair model was the most efficient in modelling the gelation process, and gave both the smallest residuals and standard error of residuals, Se (P < 0·0001). Douillard model gave the poorest fit and in particular it was not able to predict the initial part of the gelation curves. Carlson model had an intermediate behaviour and, in the case of chymosin, it gave results that were quite comparable to Scott Blair model. The parameters of the Scott Blair model were different for plant coagulants and chymosin. Chymosin had the longest rate constant (τ) and the time shift coefficient (ts) was also different between vegetable coagulants and chymosin (P < 0·0005). These results are in agreement with the overall trends for gelation profiles obtained for vegetable coagulants and chymosin. In the beginning of gelation both plant coagulants produced gels with slightly higher G′ values than chymosin, but after longer incubation times chymosin gels had higher G′ values. It was concluded that the Scott Blair model was the best equation to follow the gelation of milk induced by both plant coagulants as well as chymosin. Modelling is an important and useful method for comparing the gelation process in gels formed by different types of coagulants.

The taste-active compounds of a Camembert cheese selected for its intense bitterness defect were investigated. The water-soluble fraction (WSE) was extracted with pure water and fractionated by successive tangential ultrafiltrations and nanofiltration. The physicochemical assessment of these fractions led to the construction of a model WSE which was compared by sensory evaluation to the crude water-soluble extract, using a panel of 16 trained tasters. As no significant difference was perceived, this model WSE was then used directly or mixed with other cheese components for omission tests. Among the main taste characteristics of the WSE (salty, sour, umami and bitter), bitterness was found to be due to small peptides whose mass distribution was obtained by RPHPLC-MS (400–3000 Da) and whose taste properties are discussed.

The behaviour of Streptococcus thermophilus in combination with Lactococcus lactis subsp. cremoris or subsp. lactis mesophilic starters in experimental Cheddar cheese is reported. In a standard manufacturing procedure employing a 38 °C cook temperature, even very low levels (0·007%) of Str. thermophilus combined with normal levels of the mesophilic starter (1·7%) resulted in increased rates of acid production, the formation of significant amounts of galactose (∼ 13 mmol/kg cheese), and populations nearly equivalent to those of the mesophilic lactic starter in the curd before salting. At a 41 °C cook temperature, the Str. thermophilus attained a higher maximum population (∼ log 8·2 colony forming units (cfu)/g) than the Lc. lactis subsp. cremoris (∼ log 6·8 cfu/g) and formed more galactose (∼ 28 mmol/kg). Lactobacillus rhamnosus, deliberately added to a cheese made using Str. thermophilus starter and which contained 24 mmol galactose/kg at day one, utilized all the galactose during the first 3 months of cheese ripening. Adventitious non-starter lactic acid bacteria had the potential to utilize this substrate too, and a close relationship was demonstrated between the increase in this flora and the disappearance of the galactose. Some possible consequences for cheese quality of using Str. thermophilus as a starter component are discussed.

Two main genetic variants of κ-casein, A and B, are found in Holstein cows. A positive association between the κ-casein B genetic variant and milk characteristics of economic importance have been reported (reviews: Creamer & Harris, 1997; Ng-Kwai-Hang, 1998). This can be explained, at least in part, by a higher κ-casein content of the casein in milk of cows having the κ-casein B genetic variant (Donnelly et al. 1984; Law, 1993). We have recently reported the presence of a differential allele-specific expression of the κ-casein gene in Holstein cows by comparing the content of allele-A and allele-B specific mRNA in the mammocytes of lactating cows genotyped κ-casein AB (Robitaille & Petitclerc, 2000). For some cows, the A- and B-specific κ-casein gene alleles were expressed in a similar manner. These cows are named HH here to indicate that they were homozygous for κ-casein gene expression. The other group of cows, cows HL, presented a differential allele specific expression of the κ-casein gene allele B which was overexpressed compared to allele A. The genetic polymorphism governing this differential allele-specific expression of the κ-casein gene is likely to affect the physico-chemical properties of casein micelles and, hence, the technological properties of milk, as κ-casein is considered to be the stabilizing component of the milk micelles. The objective of the study was to evaluate the effect of this expression polymorphism of the κ-casein gene on milk ethanol stability.