Feeding during milking has been shown to influence milk production, milk flow and milking time as well as the secretion of the pituitary hormones oxytocin and prolactin, and the gastrointestinal hormone somatostatin. However, it is not known whether feeding before or after milking has any effect. The aim of the present study was to investigate how the timing of feeding relative to milking influences milk production and flow, milking time and hormone secretion. The trial was carried out over 9 weeks with 24 cows at varying stages of lactation. Each treatment period lasted for 3 weeks, including one registration week. The cows were fed ad lib. and were exposed to three treatments: feeding 1·5 h before milking (FBM), feeding at exactly the same time as milking (FDM) and feeding 1·5 h after milking (FAM). The most marked treatment effect was observed during morning milking. FDM resulted in higher milk production and higher yields of protein and lactose. FAM produced a lower fat yield and a lower fat content compared with FDM, and a lower lactose content than either FBM and FDM. Milking time was longer when cows were fed during milking, but no significant effects on milk flow were found. The amount of milk collected during the first 2 min of milking was lower when cows were fed after milking. Milking-related oxytocin and somatostatin secretion was lower in FAM than in FDM. The level of prolactin was lower when cows were fed before or after than during milking. More studies are needed to elucidate whether there is a long-term effect on milk production related to the discussed milking routines.

The effects on milk composition and processing characteristics of varying grass supply by changing stocking density and of offering a concentrate supplement were investigated. The experiment was conducted over 28 weeks of the lactation (April–October) using 48 spring-calved Friesian–Holstein cows. Three herds each of 16 cows were offered a restricted grass supply, a standard grass supply and a standard grass supply with a supplement of 3 kg concentrate/d. Treatment groups were grazed separately with a residence time of 3 d/paddock. Milk production, composition and processing characteristics such as renneting properties, ethanol stability and plasmin activity were measured weekly. Increasing stocking density above the standard system resulted in significant reductions in milk fat and protein yields, the concentrations of total protein, casein and whey proteins, and a deterioration in most processing characteristics. Imposing concentrate supplementation on the standard system increased total protein, casein and whey protein concentrations but generally did not improve processing characteristics except for ethanol stability. These results suggest that the standard grass supply in a rotational grazing paddock system can support efficient production of quality milk, and concentrate supplementation will not improve processing characteristics when an adequate supply of good quality herbage is available.

Forty-eight Manchega dairy ewes were used during a complete lactation in a 2×2 factorial design to determine the effects of supplementing diets with fat (calcium soaps of palm oil fatty acids, CSFA) and rumen undegradable protein (RUP) on milk production and composition. Factors tested were amounts of CSFA (0 or 200 g/kg) and RUP (300 or 450 g/kg crude protein) in the concentrate. RUP was altered by adding a mixture of maize gluten meal and blood meal. Lactation was divided into one nursing period (period 1, weeks 1–4), and three milking periods (periods 2–4, weeks 5–8, 9–14 and 15–21). Concentrates were given at 0–8 kg/d during periods 1 and 2, and at 0–6 kg/d in periods 3 and 4. Ewes grazed rotationally in an Italian rye-grass pasture and received a daily supplement of 0·8 kg vetch–oat hay during period 1, and 0·3 kg lucerne hay during periods 2–4. For the whole lactation, supplemental fat markedly increased milk fat content (+23%) and yield (+16%), and decreased milk protein content (−9%). The positive effect of feeding CSFA on milk fat content was more evident at the beginning of lactation; however, its negative effect on milk protein was more pronounced in late lactation. Supplementary RUP had little effect, increasing milk protein content only in period 3, when the crude protein content of pasture was lower. Milk yield and lamb growth were not affected by dietary treatments. The results indicated that CSFA can be useful for increasing the milk fat content of dairy ewes at pasture, which may help farmers to produce milk reaching the minimum requirements of fat content for the cheese industry.

The objective of this work was to examine whether variation in the amino acid supply to cows could be a reason for the reduced casein content and poorer renneting properties of milk that often occur in late summer, or whether these effects are related to proteolysis in the raw milk. In a 2×2×2 factorial design, we investigated the effects of sward (clover v. rye-grass) and supplementary feed with a high or low level of rumen-soluble N or of rumen undegradable protein on milk protein composition during the grazing season. A total of 32 Danish Holstein cows were included in the experiment. Milk protein and casein contents and the ratios casein N[ratio ]total N and casein[ratio ]true protein were at a minimum in late summer, whereas the contents of urea, non-protein N and whey protein were higher during this period. These seasonal effects were unrelated to either the type of supplementary feed or the type of sward; neither were they clearly related to proteolysis, although casein[ratio ]true protein was related to the proteose peptone content. The results indicated that whey proteins other than α-lactalbumin or β-lactoglobulin accounted for the higher proportion or concentration of whey protein in late summer. Based on a principal component analysis including variables such as citric acid, lactose and non-protein N, we suggest that the cows' energy supply during this period may be a critical factor in determining the milk protein composition, although our results were not conclusive. There was an interaction between the supplement of rumen undegradable protein and type of sward. When clover was grazed, a high supplement increased the concentrations of protein and casein in milk and the κ-casein[ratio ]total casein ratio. When rye-grass was grazed, the opposite response was found, and overall milk protein yield was not affected. The very low N content of clover in early summer reduced milk protein and casein protein during this period.

Exogenous growth hormone was administered subcutaneously to five lactating goats during the post-peak period of lactation. Milk yields increased significantly by ∼20% in response to growth hormone. Blood and milk samples were taken in the periods before, during and after growth hormone treatment. The concentrations of glucose in milk increased significantly by ∼50% in the period following growth hormone treatment at a time corresponding to the increase in milk yield. There was a transient increase in plasma glucose concentrations immediately following growth hormone treatment before either milk glucose concentrations or milk yields were increased. Both free and total IGF-1 concentrations in plasma increased slowly following growth hormone treatment. The increase in plasma IGF-1 corresponded to the increase in milk yields and milk glucose concentrations. Concentrations of IGF-1 in milk increased more rapidly than those in plasma, rising by ∼150% following growth hormone treatment, and were starting to decline by the time that milk yield and milk glucose concentrations were at their maximum. As milk glucose concentrations have been shown previously to reflect changes in the intracellular concentration of glucose, the results indicate that part of the mechanism by which growth hormone stimulates milk production is by increasing the intracellular availability of glucose for lactose synthesis. The results also suggest that changes occur in the concentrations of IGF-1 in the environment of the mammary gland before changes are observed in the general circulation, and that these are reflected in the changed concentrations in milk.

The concentrations of cations (calcium and magnesium) and anions (inorganic phosphate, citrate and chloride) have been determined during the acidification of casein micelle suspensions in the pH range 6·7–1·5. The effects of casein concentration (27, 55, 83 and 144 g/kg) and ionic strength (0, 10 and 20 g/kg NaCl added) were investigated. Acidification resulted in solubilization of calcium, magnesium, inorganic phosphate and citrate ions. However, the solubilization curves were different and depended on the casein concentration. Increasing ionic strength by adding NaCl had no effect on acid-induced mineral solubilization. These results were compared with those obtained during milk acidification and discussed in relation to the mineral solubilization that occurs during curd acidification in cheese manufacture.

Fractions enriched with α-lactalbumin (α-la) and β-lactoglobulin (β-lg) were produced by a process comprising the following successive steps: clarification–defatting of whey protein concentrate, precipitation of α-lactalbumin, separation of soluble β-lactoglobulin, washing the precipitate, solubilization of the precipitate, concentration and purification of α-la. The present study evaluated the performance of the process, firstly on a laboratory scale with acid whey and then on a pilot scale with Gouda cheese whey. In both cases soluble β-lg was separated from the precipitate using diafiltration or microfiltration and the purities of α-la and β-lg were in the range 52–83 and 85–94% respectively. The purity of the β-lg fraction was higher using acid whey, which does not contain caseinomacropeptide, than using sweet whey. With the pilot scale plant, the recoveries (6% for α-la; 51% for β-lg) were disappointing, but ways of improving each step in the process are discussed.

The development of the Maillard reaction in pasteurized, UHT and in-bottle sterilized dietetic milks was studied. In these products damage caused by heat treatments could increase as a result either of the addition of various ingredients or of manufacturing processes that alter their content of reducing carbohydrates. Protein damage was evaluated by measuring furosine by reversed-phase ion-paired HPLC. The levels of furosine detected made it possible to assess the amounts of biologically unavailable lysine. In all milks analysed blocked lysine values were <340–350 mg/g total lysine, the level at which lysine becomes the limiting amino acid in milk. Pasteurized dietetic milks had levels of blocked lysine similar to that in ordinary pasteurized cows' milk. In some UHT and in-bottle sterilized dietetic milks their different composition resulted in an increase in the blocked lysine content. In some in-bottle sterilized milks, protein damage greatly reduces the beneficial effects of milk as a dietary supplement. Lactose-free milks, which are more susceptible to protein deterioration because of their higher content of reducing carbohydrates, were also analysed after storage at 20°C and at [les ]4°C. At the end of their recommended storage times, they contained limited amounts of blocked lysine only if they had been stored at [les ]4°C.

Twenty-four dairy Leuconostoc strains isolated from French commercial starters and four reference strains were screened for 58 morphological and biochemical characters. Hierarchical clustering analysis was performed on the results of dextran production and carbohydrate fermentation tests. Most strains fitted into three clusters at a similarity level of 65% and were identified as belonging to the species lactis or mesenteroides. Fourteen strains were then selected and further characterized for their technological properties. They were examined for growth, acidification kinetics and proteolytic capacity when growing in milk, and peptidase activities and diacetyl production after growth in MRS broth. Principal component analysis was carried out to group strains and compare their technological properties. This allowed selection of strains of dairy interest. Special attention was given to the flavouring capacity of Leuconostoc lactis isolates that could be useful in designing commercial starters.

A hydrolase activity that cleaves lysyl-p-nitroanilide (Lys-pNA) has been purified from the cytoplasm of Lactococcus lactis subsp. cremoris AM2 by chromatography on DE52, DEAE Affi-Gel Blue Gel, Hydroxyapatite Bio-Gel HTP and Phenyl Sepharose. The purified aminopeptidase was found to have a native Mr of 50 000–55 000 by gel filtration chromatography and by FPLC gel filtration on Superose 12 and to be composed of a single polypeptide chain following SDS-PAGE. Enzyme activity was almost completely inhibited by EDTA, amastatin, puromycin and bestatin, while the sulphydryl-reactive agents p-chloromercuribenzoate and iodoacetamide were inhibitory. The enzyme was found to be very unstable during the purification procedures at 4°C and its stability was greatly improved when 10 ml glycerol/l and 2 mm-dithiothreitol were included in the purification buffers. The purified enzyme was found to hydrolyse a wide range of dipeptides, tripeptides and longer peptides provided that proline was not present in the penultimate position from the N-terminus or that a pyroglutamyl residue was not present at the N-terminus. While neither Asp-pNA nor Pro-pNA was hydrolysed by the purified enzyme, the release of N-terminal acidic residues from peptides was observed in addition to the release of N-terminal proline from Pro–Leu–Gly–NH2, Pro–Leu–Gly–Gly and Pro–His–Pro–Phe–His–Leu–Phe–Val–Tyr. This ability of Lys-pNA hydrolase to release N-terminal proline residues was employed in concert with a purified aminopeptidase P preparation to release alternate N-terminal amino acids from Tyr–Pro–Phe–Pro–Gly. The complementary action of these enzymes represents an alternative mechanism to that of post-proline dipeptidyl aminopeptidase for metabolism of proline-containing peptides.

Experimental cheeses were prepared in triplicate from pasteurized milk inoculated with Debaryomyces hansenii under aseptic conditions. Three cheesemaking replicates, with efficient control of environmental parameters (temperature, relative humidity, atmospheric composition) showed similar ripening characteristics. Deb. hansenii grew only on the cheese surface, where its oxygen demand was satisfied, especially during the first 24 h (mean generation time, 5·8 h). Salting in a sterile saturated brine solution reduced its growth and decreased viability. Growth was slower after 48 h because of the decrease in ripening temperature (mean generation time, 94 h). The total count of Deb. hansenii was maximum (≈3×107 yeasts/mm2) after 6 d ripening and its viable cell concentration was ≈2×106cfu/mm2. This difference was due to the ‘non-viability’ of part of the population. The viable Deb. hansenii concentration was highly correlated (r2>0·95) with the lactate concentration in the inner part and with the surface and inner lactose concentrations, up to day 10 of ripening. This emphasized the importance of the diffusion of carbon substrate from the inner part to the surface of the cheese during ripening. The pH of the inner part depended significantly on the lactate and lactose concentrations. Surface pH was significantly related to inner lactate concentration, temperature and relative humidity. This also demonstrated the controlling role of carbon source diffusion.

Branched-chain amino acids (AA) are mostly metabolized in the splanchnic area, but some are metabolized within the mammary gland and thus could contribute to the synthesis of non-essential AA (Wohlt et al. 1977). The extraction rate of leucine from plasma by the mammary gland is particularly high (64% in the goat; Roets et al. 1983), in excess of that used for the synthesis of milk proteins (Davis & Mepham, 1976; Wohlt et al. 1977; Roets et al. 1983). Thus, although mammary leucine is mainly used for milk protein, it also provides amino nitrogen and carbon sources for the synthesis of non-essential AA.

To our knowledge, no information is available on the transfer and distribution of plasma leucine amino nitrogen to milk protein AA. Using the technique for chromatographic fractionation of AA recently developed by Casseron et al. (1997), we studied the specific 15N enrichment of casein (CN) AA in the goat given a single intravenous injection of [15N]leucine.

The use of transgenic animals to manipulate milk composition has considerable potential, both for the production of biomedical proteins and for the direct manipulation of milk composition for the improvement of dairy animals and their products (for reviews, see Wall et al. 1992; Yom & Bremel, 1993). Promoters from a number of milk protein genes from a variety of species have been tested for their ability to direct the expression of foreign proteins to the mammary gland (for review, see Maga & Murray, 1995).

β-Lactoglobulin (β-lg) is the major whey protein produced in ruminant milk and is part of the normal milk composition of most mammals except humans and rodents (Pervaiz & Brew, 1985). It is expressed at high levels in the mammary gland and is developmentally regulated. Transgenic mice have been produced using the complete ovine (Simons et al. 1987; Shani et al. 1992) and caprine (Ibañez et al. 1997) β-lg genes. In general, high levels of expression were obtained with the ovine β-lg gene, and expression was also seen in a position-independent manner (Whitelaw et al. 1992). Lower levels of expression were reported using the caprine β-lg gene. Here we report the production of transgenic mice using the bovine β-lg gene. We describe high expression, position-dependent, and copy number-related expression of bovine β-lg protein in the milk of six lines of transgenic mice.

In the lactating mammary gland, epithelial cells secrete triacylglycerols in the form of droplets enveloped by an apical surface membrane. This membrane is known as the milk fat globule membrane (MFGM; Mather & Keenan, 1983). MFGM-associated proteins have been studied and employed in developing antibodies against surface antigens of breast epithelial cells, which are used in breast cancer immunodiagnosis and histopathology (Salinas et al. 1987; Larocca et al. 1991; Peterson et al. 1995). So far, only a small number of proteins have been analysed and their sequence identified in bovine, murine and human MFGM; among them are butyrophilin (Jack & Mather, 1990; Ishii et al. 1995; Taylor et al. 1996), MFG-E8 (Stubbs et al. 1990), PAS 6/7 (Hvarregaard et al. 1996) and lactadherin or breast antigen BA46 (Couto et al. 1996; Taylor et al. 1997). Several minor proteins have yet to be characterized, since it is not easy to isolate them in large quantities from the membrane. SDS gel patterns give useful information about MFGM proteins, such as concentration, relative molecular mass and presence of carbohydrate. Over forty membrane components have been separated by electrophoretic techniques from bovine MFGM (Mather et al. 1980).

The research reported here combined SDS-PAGE with sequencing analysis and describes the composition of human MFGM, with the exception of high molecular mass mucin, which only penetrates an acrylamide gel of 40 g/l. Mucins have been extensively studied and the sequence predicted from cDNA (Gendler et al. 1990). Surprisingly, identification of the protein bands in the present study revealed that three proteins alone constituted the major components of human MFGM: xanthine oxidase (EC 1.1.3.22), butyrophilin and lactadherin. Lactadherin belongs to a family of proteins possessing epidermal growth factor-like and factor V/VIII C1/C2-like domains, including bovine PAS 6/7, guinea pig GP55 (Hvarregaard et al. 1996) and murine MFG-E8 (Stubbs et al. 1990). In a previous investigation, we characterized lactadherin (formerly breast antigen BA46) and its truncated 30 kDa form as components of healthy human MFGM (Giuffrida et al. 1998). Human butyrophilin has recently been cloned and sequenced (Taylor et al. 1996); the presence of two extracellular immunoglobulin superfamily domains suggested a potential cell surface receptor function. This study was aimed at identifying and characterizing the multiple forms of the major proteins of MFGM.

Lactobacilli are of great commercial value because they are widely used as starters in food fermentation. The species Lactobacillus delbrueckii, which comprises the three subspecies bulgaricus, lactis and delbrueckii, is important in dairy products and vegetables. The subspecies bulgaricus is present in yogurt, and subspecies lactis is recovered from whey, starter cultures and cheeses (Stiles & Holzapfel, 1997). It is unusual to recover Lb. delbrueckii subsp. delbrueckii (Lb. delbrueckii) from dairy products, from which Lb. delbrueckii subsp. lactis (Lb. lactis) and subsp. bulgaricus (Lb. bulgaricus) are typically isolated. However, given the similarity of these two latter subspecies, a clear identification of isolates on the basis of phenotype criteria alone is often problematic (Dellaglio, 1989; Millière et al. 1996).

The use of DNA probes and methods based on polymerase chain reaction (PCR) have greatly facilitated diagnostic identification of bacteria. For the lactobacilli, DNA probes have been described for Lb. helveticus, Lb. acidophilus, Lb. fermentum, Lb. plantarum and most of the other Lactobacillus species (Pot et al. 1994; Tailliez et al. 1994; Quere et al. 1997). For Lb. delbrueckii, an EcoRI DNA fragment of the plasmid pY85 was used as a probe for this species, although it was not able to discriminate its three subspecies (Delley et al. 1990).

In our laboratory, a specific amplification of a DNA fragment of ∼1·7 kbp using universal primers for the amplification of ribosomal RNA (rRNA) genes was found only for dairy isolates of Lb. lactis and not for Lb. bulgaricus, Lb. delbrueckii, Lb. helveticus and Lb. acidophilus (G. Giraffa, P. de Vecchi and L. Rossetti, unpublished results). This prompted us to test the possible use of this fragment as a specific DNA probe for Lb. lactis. To this end, Southern and dot blot hybridization experiments were carried out with total DNA of several strains belonging to different lactic acid bacteria (LAB) species.

Brucellosis can be transmitted to man by direct contact with infected animals or through contaminated meat, milk and dairy products (Nicoletti, 1989). The analysis of Brucella spp. is carried out in the laboratory by microbiological or serological assays (Alton et al. 1988). The first are more specific but are also time-consuming and expose the analyst to the risk of infection (López-Merino, 1991). However, the latter can result in false positives owing to cross reactivity with other Gram-negative bacteria (Diaz-Aparicio et al. 1994). Because of these limitations, the amplification in vitro of specific DNA regions by the polymerase chain reaction (PCR) could represent a powerful tool for rapid and specific diagnostic analysis. In recent years, several PCR methods have been developed to amplify specific DNA sequences of Brucella strains (Herman & de Ridder, 1992; Romero et al. 1995; Valentino et al. 1997). In addition, direct analysis of Brucella in contaminated abortive tissues (Fekete et al. 1992), milk and blood (Leal-Klevezas et al. 1995; Rijpens et al. 1996) has been reported.

In this paper we describe a method for gene-specific PCR amplification of a 443 base pair (bp) fragment of Brucella DNA that belongs to a gene encoding for a 31 kDa outer membrane protein. This protein (BCSP-31) is a membrane antigen characteristic of the Brucella genus (Mayfield et al. 1988). The PCR method was developed for the analysis of soft cheeses. We focused our attention on Mozzarella, Pecorino and ricotta samples, because such products are not subjected to the natural microbial autopurification process of maturing. They are widely consumed in Italy and a relationship between infected foods and the areas where brucellosis is a human zoonosis is a possibility.

The analysis was performed without purification of DNA from bacteria. Indeed, after homogenization, the sample was subjected to thermal shock by freeze–thaw cycles that lysed bacteria and solubilized nucleic acids for subsequent PCR amplification. Amplified DNA fragments were separated by agarose gel electrophoresis and visualized by ethidium bromide staining. Several brands of soft cheeses and ricotta contaminated at different levels with Brucella cells were analysed by our procedure to evaluate the detection sensitivity and the repeatability of the method.

The composition and sensory characteristics of matured cheeses are controlled by a number of factors, among which the type of feed is important. The influence of feeding can be reflected by the presence in cheeses of terpenes and sesquiterpenes, compounds typically indicating their vegetable origin (Mariaca et al. 1997). Indeed, several investigators have already established that these compounds could characterize the forage even to a specific geographical location. Dumont & Adda (1978), Dumont et al. (1981), Guichard et al. (1987), Bosset et al. (1994) and Moio et al. (1996) were thus able to distinguish cheeses from lowland and upland regions. In most of these studies, the conditions under which milk was produced and processed were not completely controlled. Furthermore, all these studies analysed only the volatile fraction of cheeses and did not examine the volatile compounds in the forages used. Recently Bosset et al. (1994), in one part of the project described by Jeangros et al. (1997), showed that highland grass with a highly diversified botanical composition produces milk and cheeses with significantly different chemical compositions from those from lowland grass. To improve our understanding of the relationship between animal feed and cheese composition, we have investigated under controlled experimental conditions both the composition of the terpene and sesquiterpene fractions of four forages with different botanical diversities and the influence of those forages on the terpene fraction of cheeses.

Kefir is a refreshing fermented milk with a slightly acidic taste obtained by incubating milk with kefir grains (Saloff-Coste, 1996). Kefir grains are a complex mixture of lactic acid bacteria and yeasts in a strong and specific association. They are characterized by an irregular form, a folded and uneven surface and a white or slightly yellow colour. They are tough and resilient and have a characteristic acid taste (Bottazzi et al. 1994). The basic microflora contains lactococci, homofermentative and heterofermentative lactobacilli, yeasts and acetic acid bacteria (Bottazzi et al. 1994; Rea et al. 1996). Among the yeasts isolated from grains and identified are Candida kefir, Saccharomyces cerevisiae, Sac. delbrueckii, Cand. holmii, Sac. unisporus and Sac. lipolytica (Angulo et al. 1993; Marshall, 1993; Garrote et al. 1997). Lactobacillus brevis, Lb. viridescens, Lb. casei, Lb. kefir, Lb. kefiranofaciens, Lb. kefirgranum, Lb. parakefir, Leuconostoc spp. and Lactococcus lactis are among the lactic acid bacteria present in the grains (Marshall et al. 1984; Toba et al. 1991; Takizawa et al. 1994; Garrote et al. 1997). The mixed microflora of yeasts and bacteria is held together by a matrix containing (g/kg dry weight) protein 340, polysaccharide 470 (Ottogalli et al. 1973).

The study of kefir grains in milk has been centred on the characteristics of the polysaccharide produced by lactobacilli within the grain (Yokoi et al. 1991). This polysaccharide, named kefiran, is composed of glucose and galactose (Yokoi et al. 1991). It has been suggested that proteins are incorporated from the growth media (Bassette & Acosta 1988), but no details about structure and composition are available.

Soyabeans are an important component of the diet in many countries and have been used to obtain fermented products such as sogurt (Mann, 1991). Special attention has been given to the growth of, and sugar utilization by, Bifidobacterium spp., Lb. delbrueckii subsp. bulgaricus and Streptococcus thermophilus in soya milk (Buono et al. 1990; Murti et al. 1993; Ankenman Granata & Morr, 1996). To our knowledge, there have been no attempts to ferment soya milk with kefir grains.

The aim of this study was to investigate the growth of kefir grains in soya milk and the composition of these grains, focusing on the matrix proteins.

Dulce de leche is a dairy-based confectionery product, widely consumed in Argentina and other Latin American countries as filling or topping for desserts, cakes and biscuits (Hansen, 1978; Bouzas, 1997). It is prepared by heat concentration of whole milk with added sucrose until it reaches 700 g total solids/kg dulce de leche. Sucrose is usually partly replaced by other sugars (not more than 400 g/kg of total sugars added according to Normas Mercosur, 1996) to avoid crystallization. Sodium bicarbonate is added to prevent coagulation of proteins as the pH of the mixture decreases and the isoelectric pH of proteins increases during heating.

Maillard browning is the main reaction of milk proteins during the manufacture of dulce de leche. This results in an attractive flavour and colour, but also leads to less desirable changes, such as loss of nutritional value and formation of toxic compounds (Rizzi, 1994; Friedman, 1996a). The natural reactants are free amino groups of milk proteins and lactose. However, as the addition of other reducing sugars is an accepted practice, the consequences may be quite different depending on the formulation used. In addition, other reactions leading to crosslinkages between protein chains contribute to the reduction in nutritional value (Mauron, 1990; Friedman, 1996b; Finot, 1997). The reduction of nutritional quality of milk proteins is attributed to the lower digestibility and the biological inactivation or destruction of essential amino acids. Lysine is the essential amino acid most affected as its free ε-amino group is very reactive (Hurrell, 1990; Mauron, 1990; Finot, 1997).

Several studies have investigated the physical and rheological characteristics (Pauletti et al. 1984, 1996), sensory properties (Sabioni et al. 1984; Buera et al. 1990; Hough et al. 1990), sugar composition and microbiological stability (Ferramondo et al. 1984) of dulce de leche. However, there is little information about the loss in available lysine during manufacture (Pavlovic et al. 1994).

The aim of this study was to examine the effects of the nature and quantity of various sugars commonly used in the manufacture of dulce de leche on the loss of available lysine.