The aim of this study was to identify whey-derived peptides with angiotensin I-converting enzyme (ACE) inhibitory activity. The bovine whey proteins α-lactalbumin and β-lactoglobulin were hydrolysed with pepsin, trypsin, chymotrypsin, pancreatin, elastase or carboxypeptidase alone and in combination. The total hydrolysates were fractionated in a two step ultrafiltration process, first with a 30 kDa membrane and then with a 1 kDa membrane. Inhibition of ACE was analysed spectrophotometrically. The peptides were isolated by chromatography and identified by mass and sequencing analysis. The most potent inhibitory peptides were synthesized by the 9-fluorenylmethoxycarbonyl solid phase method. Inhibition of ACE was observed after hydrolysis with trypsin alone, and with an enzyme combination containing pepsin, trypsin and chymotrypsin. Whey protein digests gave a 50 % inhibition (IC50) of ACE activity at concentration ranges within 345–1733 μg/ml. The IC50 values for the 1–30 kDa fractions ranged from 485 to 1134 μg/ml and for the <1 kDa fraction from 109 to 837 mg/ml. Several ACE-inhibitory peptides were isolated from the hydrolysates by reversed-phase chromatography, and the potencies of the purified peptide fractions had IC50 values of 77–1062 μM. The ACE-inhibitory peptides identified were α-lactalbumin fractions (50–52), (99–108) and (104–108) and β-lactoglobulin fractions (22–25), (32–40), (81–83), (94–100), (106–111) and (142–146).

We have investigated the capacities of Geotrichum candidum strains to produce sulphides from methionine. This attribute is very important in cheese technology because of the flavouring potential of sulphur compounds. A spectrophotometric procedure using 5,5′-dithiobis(2-nitrobenzoic acid) to determine sulphides was tested on a collection of G. candidum strains, and confirmed by gas chromatography–mass spectrometry. The strains were distinguished on the basis of their ability to produce methanethiol. Gas chromatography–mass spectrometry also made it possible to identify other sulphides, such as dimethyl disulphide, dimethyl trisulphide and dimethyl sulphide. Using sonicated cells, the specific production of these four sulphides was studied in presence of L-[S-methyl-2H]methionine. Both methanethiol and dimethyl sulphide were produced from methionine, but two different pathways were used by G. candidum.

31P-NMR has been tested as a technique for determining the phosphorus content of milk samples in both the inorganic form (Pi) and bonded to caseins as phosphoserine (SerP). 31P-NMR made possible the simultaneous determination of Pi and SerP in ∼ 1·7 h. The determination is based on measuring the area under the resonance in question and interpolating in a calibration curve prepared using external standards of known phosphorus contents and prepared in the same manner. The concentration–area plots obtained were linear in the ranges 162–1993 mg P/l for Pi and 38–402 mg P/l for casein SerP. The repeatability of the method was good. When the same milk samples were analysed by both 31P-NMR and the classic colorimetric method the correlation was good. Polyphosphates added to commercial milk were also detected, as diphosphates. The NMR response for diphosphate was linear between 58 and 576 mg P/l and preliminary results indicated that quantitative analysis could be achieved in commercial milk.

Casein fractions and their breakdown products in Iberico-type cheeses made from the milk of cows, ewes or goats were analysed by capillary electrophoresis in order to characterize them. The actions of plasmin and chymosin on caseins were evaluated by comparing the electropherograms of caseins from milk and from cheese, both with and without treatment with plasmin. Characteristic capillary electrophoresis patterns were obtained for cheeses made from the milk of each of the three species, and the main components were identified. Caprine para-κ-casein and bovine β-caseins, eluting at the first and at the last part of the electropherogram respectively, were found to be indicative of the presence of the milks of these species.

Two new variants have been detected and characterized for the goat β-lactoglobulin gene at the cDNA level and confirmed at the genomic level. The two polymorphisms are located on exon 7 of the gene. One of the polymorphic sites is produced by a single nucleotide substitution in position +4601, allowing a polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP) genotyping procedure to be developed using SacII restriction enzyme. The other polymorphic position contains a 10 bp long insertion at position +4641 that can be detected by capillary electrophoresis of the PCR product amplified with a fluorescent primer. The association of these two polymorphisms was also investigated, resulting in the description of two new alleles. Both of these contained the point mutation at the SacII site, with or without the 10 bp insertion at position +4641. The distribution of these new polymorphisms was studied in a population of males of four different goat breeds. The gene frequencies for these variants were similar in Spanish and French breeds.

Growth of propionibacteria in complex media was independent of the initial number of cells; in contrast, growth of propionibacteria in milk and whey did not occur if the initial level of cells was <106 cfu/ml. Addition of vitamins, minerals or complex nitrogen sources to the milk or whey, or incubation under anaerobic conditions had no effect on the lack of growth. Addition of freeze-dried whey, prepared from skim milk reconstituted from powder, to a complex medium prevented growth from low inocula in the complex medium, demonstrating the presence of an inhibitor or inhibitors in the whey. The inhibitor(s) was heat stable, had a low molecular mass and retained its activity for at least 4 weeks at 20 °C. Pregrowth of some lactic acid bacteria, used as starter cultures in Swiss-type cheese manufacture, in milk for 2 weeks at 20 °C removed the inhibition, which explains how propionibacteria develop in Swiss-type cheese from low numbers even though they are inhibited in milk.

DNA fingerprints of lactic acid bacteria were generated by polymerase chain reaction using a primer based on the repetitive elements found in the genome of Streptococcus pneumoniae (BOX-PCR). The method made it possible to identify 37 isolates from raw milk, industrial starters and yogurt. Differentiation at species, subspecies and strain level was possible for Lactobacillus delbrueckii subsp. lactis, Lb. delbrueckii subsp bulgaricus and Str. thermophilus. BOX-PCR was also applied to studying the strain composition of a starter culture and the direct detection of strains in commercial fermented milk.

Transgenic mice were produced carrying the coding region of the rabbit κ-casein gene linked to the upstream region of the rabbit whey acidic protein gene. Mice from the highest-expressing line produced 2·5 mg rabbit κ-casein/ml in their milk. The foreign protein was associated with the casein micelles and altered micelle size, though in the high-expressing line rabbit κ-casein also segregated into the whey fraction obtained after centrifuging the milk samples. Milk from transgenic mice had the same overall protein content as that from non-transgenic mice, except for the transgene product. However, litters fed with this transgenic mouse milk grew less well than litters given milk from non-transgenic mice. This reduction in growth was not related to changes in mammary gland structure or mammary cell morphology. Preliminary results indicated that milk from the transgenic mice had a higher viscosity.

The stability of the water-soluble B vitamins and the fat-soluble vitamins A and E of three liquid infant milks was followed throughout 12 months of storage at 20, 30 and 37 °C. Two samples were ‘follow-on’ milks from two different manufacturers and the other a ‘junior’ milk from one of these. The levels of water-soluble vitamins and vitamin E remained constant during storage at the three temperatures. In contrast, vitamin A decreased in samples from one manufacturer at all storage temperatures, and in samples from the other only at 37 °C. A kinetic equation was derived that made it possible to predict the influence of storage temperature on the shelf life of these kinds of products, together with the storage period during which levels of vitamin A recommended as adequate would remain.

Goat whey was hydrolysed by pepsin in an ultrafiltration membrane enzymic reactor coupled with a 30 kDa mineral membrane. Peptides collected in the permeate were resolved using reversed-phase HPLC. Their sequences were determined by amino acid analysis, second order derivative spectra analysis and mass spectrometry. Owing to the resistance of β-lactoglobulin (β-lg) towards pepsin, the majority of peptides identified were derived from α-lactalbumin (α-la). Pepsin showed a broad specificity of hydrolysis sites and generated a wide range of products from dipeptides to very large peptides containing disulphide bridges. The molecular masses of peptides resulting from α-la degradation were between 150 and 6900 Da: 36% were < 600 Da, 24% were 600–2000 Da and 40% were > 2000 Da.

Lamb rennet pastes were prepared by the procedure most commonly used by Idiazabal cheese manufacturers. We studied the effects on their coagulating and lipolytic activities of the state of the stomach at the time of death (full of milk or empty), the amount of NaCl added, the origin of the lambs and paste storage time. Coagulating activities were generally between 155 and 363 units/g tissue. Pastes prepared from stomachs of lambs from slaughterhouse flocks had significantly higher coagulating activities than those of lambs from separate flocks. No significant decrease in coagulating activity was observed after 1 year storage at 4 °C. Chymosin represented 75–80% of the total coagulating activity with the remainder being pepsin. Rennet paste extracts with pH < 4·7 did not have increased coagulating activities when their pH was lowered to 2·0, while those with pH > 5·2 had activities 1·5-fold those before treatment. Lipase activity was higher in extracts of rennet pastes prepared using the stomachs of lambs that arrived at the slaughterhouse in the morning just prior to slaughter than in those prepared with the stomachs of lambs that had arrived on the previous evening. However, the reverse was the case for esterase activity. Activating the coagulating activity by pH cycling completely destroyed both lipolytic activities. Storage at 4 °C for > 1 year did not affect esterase activity but lipase activity decreased substantially after 4–5 months. Lipase, but not esterase, activity was responsible for the liberation of short-chain free fatty acids from ovine milk fat.

Camel calf rennet (CCR) and buffalo calf rennet (BCR) were prepared from dried abomasa to study their physicochemical properties and electrophoretic behaviour and to carry out an immunological characterization of the rennet proteins. CCR was more thermostable than BCR. The milk clotting activity of both rennets increased as pH decreased. The optimum temperatures for CCR and BCR were 50 and 45 °C respectively. CCR was more sensitive to increased CaCl2 in milk than BCR. Addition of NaCl to milk in the range 0–100 g/l resulted in a marked decrease in the clotting activity of both rennets. When the rennets were treated with acetone, the activity of BCR was completely destroyed, but that of CCR was unaffected. The proteolytic activity of CCR was higher than that of BCR and pepsin towards both camel and cows' milk caseins at pH 6·0. SDS-PAGE electrophoretic patterns of CCR and BCR proteins gave two major bands with molecular masses estimated as 52 and 39 kDa for CCR and 50 and 35 kDa for BCR. Immunodiffusion and immunoelectrophoresis using anti-CCR serum demonstrated immunological cross reactivity between CCR and BCR.

Whey protein isolate (WPI), either untreated or pretreated at 80 °C for 30 min, was incubated with a proteinase from Bacillus licheniformis until a gel was formed. Standardized reaction times, directly linked to the degree of hydrolysis, were obtained from plots of the relative amount of peptides released v. reaction time obtained under different conditions (enzyme concentration, temperature, pH, NaCl addition). This provided a connection between the gelation profile and the degree of hydrolysis. In the case of untreated WPI, gelation occurred at lower degrees of proteolysis when the enzyme concentration was decreased, demonstrating that a rate-limiting aggregation process occurred at the same time as the proteolysis in a manner similar to the renneting of milk. This was not the case for preheated WPI, when gelation was found to take place at a constant degree of proteolysis, independent of the enzyme concentration. In this case, the mechanism could be described by assuming the thermally induced aggregates present in this substrate had progressively more stabilizing peptide segments shaved off, resulting in increased attraction between individual aggregates that ultimately led to gelation. Results obtained at 40–60 °C supported this, as we found no effect of temperature on the degree of proteolysis at gelation for the untreated WPI, whereas the degree of proteolysis decreased with increasing temperature when heated WPI was hydrolysed. The effect of pH and NaCl addition on the process was to reduce repulsion between the aggregating species so that gelation was induced at a decreased degree of proteolysis.

Analysis by gas chromatography–mass spectrometry of pasteurized milk with a fruity (pineapple-like) off-odour and a sour, rancid and soapy taste indicated the presence of concentrations at μg/ml levels of ethyl butanoate, ethyl hexanoate, ethyl octanoate, ethyl decanoate, octanoic acid, decanoic acid and dodecanoic acid. The off-odour and taste were attributed to the presence of these compounds in the milk. Microbiological examination confirmed that the milk was also contaminated with a series of psychrotrophic bacteria including Yersinia intermedia, Pseudomonas putida and Rahnella aquatilis. Growth of isolates of these bacteria in UHT milk at 23 °C for 7 d showed that Yer. intermedia produced significant quantities of the C4–C12 alkanoic acids; Ps. putida produced only small quantities of these acids and Rah. aquatilis produced none. In addition, Yer. intermedia and Ps. putida also produced small but significant quantities of the corresponding ethyl esters. In milk inoculated with both Yer. intermedia and Ps. putida, the quantity of ethyl esters produced was greater than that found in cultures containing only one of the isolates. These studies indicated that Yer. intermedia was the principal source of the alkanoic acids in the tainted milk and that the major producer of the corresponding ethyl esters was Ps. putida. This is the first report that Yer. intermedia and Ps. putida can cause an off-odour or taste in dairy products.

To explain the limited nutritional value of milk cultured with proteinase-positive (Prt+) strains of Lactococcus lactis for the subsequent growth of dairy lactococci, we investigated further the time courses of modifications in the free amino acid and peptide contents of cultured milk. When growing in milk for up to 24 h, Prt+ strains of Lc. lactis progressively accumulated amino acids and casein-derived peptides. The growth of proteinase-negative (Prt−) wild-type strains and peptide transport mutants of Lc. lactis in cultured milk showed that casein-derived peptides could sustain growth up to 5×108 cfu/ml, depending on the extent of casein degradation during the preliminary growth of Prt+ strains and the Prt− strains. Of the casein-derived oligopeptides, <25% were transported into the cell and used for Lc. lactis growth. However, they played a prominent role, contributing 90% to growth. In contrast, di- and tripeptides did not contribute to growth, suggesting that either few were released from caseins or they did not supply essential amino acids.

We have developed a method to quantify the resistance to freezing and frozen storage of lactic acid starters, based on measuring the time necessary to reach the maximum acidification rate in milk (tm) using the Cinac system. Depending on the operating conditions, tm increased during the freezing step and storage. The loss of acidification activity during freezing was quantified by the difference (Δtm) between the tm values of the concentrated cell suspension before and after freezing. During storage at −20 °C, linear relationships between tm and the storage time were established. Their slope, k, allowed the quantitation of the decrease in acidification activity during 9–14 weeks of frozen storage. The method was applied to determine the resistance to freezing and frozen storage of four strains of lactic acid bacteria and to quantify the cryoprotective effect of glycerol.

The effects of heat treatment of milk, and a range of rennet and glucono-δ-lactone (GDL) concentrations on the rheological properties, at small and large deformation, of milk gels were investigated. Gels were made from reconstituted skim milk at 30 °C, with two levels each of rennet and GDL. Together with controls this gave a total of sixteen gelation conditions, eight for unheated and eight for heated milk. Acid gels made from unheated milks had low storage moduli (G′) of < 20 Pa. Heating milks at 80 °C for 30 min resulted in a large increase in the G′ value of acid gels. Rennet-induced gels made from unheated milk had G′ values in the range ∼ 80–190 Pa. However, heat treatment severely impaired rennet coagulation: no gel was formed at low rennet levels and only a very weak gel was formed at high levels. In gels made with a combination of rennet and GDL unusual rheological behaviour was observed. After gelation, G′ initially increased rapidly but then remained steady or even decreased, and at long ageing times G′ values increased moderately or remained low. The loss tangent (tan δ) of acid gels made from heated milk increased after gelation to attain a maximum at pH ∼ 5·1 but no maximum was observed in gels made from unheated milk. Gels made by a combination of rennet and GDL also exhibited a maximum in tan δ, indicating increased relaxation behaviour of the protein–protein bonds. We suggest that this maximum in tan δ was caused by a loosening of the intermolecular forces in casein particles caused by solubilization of colloidal calcium phosphate. We also suggest that in combination gels made from unheated milk a low value for the fracture stress and a high tan δ during gelation indicated an increased susceptibility of the network to excessive large scale rearrangements. In contrast, combination gels made from heated milk formed firmer gels crosslinked by denatured whey proteins and underwent fewer large scale rearrangements.

Mutants of Lactococcus lactis producing excess carbon dioxide could be isolated on LDHA-20 agar (described by El Attar et al. Journal of Dairy Research67 641–646 2000). The use of these mutants in the manufacture of Roquefort cheese has the potential to improve the formation of openings in this cheese. The aim of this work was to examine the stability of these mutants, their enzymic activities and their metabolism of lactose and citrate during growth in milk. They produced less l-lactate than the parent strain and their lactate dehydrogenase activity was lower. Nevertheless none of the mutants produced no L-lactate at all and the most active gas generators among them generally produced 30–50 mM-L-lactate. Unexpectedly, all the strains produced some D-lactate, some > 10 mM. We found that carbon dioxide production by the mutants could be determined indirectly by assaying acetoin, citrate and 2,3-butanediol by high-performance liquid chromatography. Generally, spontaneous mutants were more stable than those obtained after treating with nitrosoguanidine or u.v. irradiation.

The fatty acid composition and freeze–thaw resistance of eight strains of thermophilic lactobacilli were studied. Seven of these contained the same polar and neutral lipids, the five major components making up 90% of the cellular fatty acid pool being 14[ratio ]0, 16[ratio ]0, 16[ratio ]1, 18[ratio ]1 and C19 cyclopropane (cyc19[ratio ]0). Strain comparison by means of cluster analysis based on the fatty acid ratios using the overlap coefficient revealed two well defined clusters. One was formed by three strains of species Lactobacillus delbrueckii subsp. lactis and Lb. delbrueckii subsp. delbrueckii, the other included five strains of the species Lb. delbrueckii subsp. bulgaricus, Lb. acidophilus and Lb. helveticus. Resistance of strains with a high content of unsaturated fatty acids (66–70%) decreased with increasing cyc19[ratio ]0 concentrations. In contrast, in strains with a low concentration of unsaturated fatty acids (42–49%), increasing cyc19[ratio ]0 levels were associated with increased freeze–thaw resistance.

Capillary electrophoresis using a hydrophilically coated capillary and a low pH buffer containing urea has been used to characterize processed cheeses. Different electrophoretic patterns were obtained depending on the ingredients used in the blend such as acid casein, rennet casein, sodium and calcium caseinates and skim milk powder. Isoelectric casein, and sodium and calcium caseinates were shown to contain intact non-glycosylated κ-casein (κ-CN), while rennet casein contained only trace amounts of κ-CN and mainly para-κ-CN. Therefore, the addition of casein or caseinate to processed cheeses has been detected by analysing the intact non- glycosylated κ-CN. Quantitation of intact non-glycosylated κ-CN in processed cheeses of known and unknown composition was carried out using a regression curve from standard mixtures of 150–550 g isoelectric casein/kg total rennet casein. This capillary electrophoresis method successfully confirmed the addition of isoelectric casein or caseinate to processed cheeses of known composition. The quantitative determination range was 0·605–3·688 mg κ-CN/ml. This method cannot be used for measuring additions of rennet casein or any caseinates that have been exposed to chymosin.