Sylvie Pochet

Ripening and quality of Swiss-type cheese made from raw, pasteurized or microfiltered milk

Abstract Experimental mini-cheeses were made from raw (Ra), microfiltered (MF), pasteurized (Pa) (72 °C, 30s) or pasteurized mixed with microfiltration retentate (PR) milk to study the influence of the indigenous microflora and pasteurization on the quality of Swiss-type cheese. To estimate biochemical transformations during cheese ripening, several methods were used: nitrogen fractionation (water-soluble fraction and phosphotungstic acid (PTA)-soluble fraction), urea-polyacrylamide gel electrophoresis of caseins, reverse phase liquid chromatography of the water-soluble fraction, lactate and volatile fatty acids. Microbial populations were also enumerated. At the end of ripening, in comparison with MF and Pa milk cheeses, Ra and PR milk cheeses exhibited higher overall aroma intensity and pungency, characteristics which correlated with higher populations of facultatively heterofermentative lactobacilli (108 cfug−1), propionibacteria (108 cfug−1), and enterococci (106cfug−1). These cheeses had high levels of PTA-soluble N and acetic, propionic and isovaleric acids. MF and Pa milk cheeses, although somewhat different from one another, were very different from the two other types of cheese. Pa milk cheese had a lower pH than MF milk cheese, and contained a higher proportion of γ-caseins due to the activation of plasmin. Moreover, Pa milk cheese was more acidic, but demonstrated a higher overall aroma intensity. The addition of raw milk flora (retentate) to Pa milk restored almost all the biochemical and sensory characteristics of Ra milk cheese measured in this study.

Low occurrence of safety hazards in coagulase negative staphylococci isolated from fermented foodstuffs

Some coagulase negative staphylococci (CNS) species play an important role in the fermentation of meat and milk products and are considered as food-grade. However, the increasing clinical significance of CNS and the presence of undesirable and unsafe properties in CNS question their presence or use in food. Our goal was to assess the safety of CNS by developing a diagnostic microarray targeting 268 genes corresponding to safety hazards in a food context i.e. toxins (especially enterotoxins) and determinants of antibiotic resistance and biogenic amine production. Target genes were selected among staphylococci and Gram-positive species that may be in contact with CNS in foodstuffs. The diagnostic microarray was used to screen 129 strains belonging to the 2 dominant species isolated from foodstuffs (S. equorum and S. xylosus) and the 2 main species isolated both in foodstuffs and clinical samples (S. epidermidis and S. saprophyticus). Microarray data were further completed by antibiograms and measurement of biogenic amine production. Safety hazards associated with CNS were mostly limited to the presence of antibiotic resistance. Seventy-one percent of the strains possessed at least one gene encoding antibiotic resistance, while only one strain carried an enterotoxin gene. Most strains did not carry any genes encoding staphylococcal toxins (68%), non-staphylococcal toxins (95%) or decarboxylases involved in biogenic amine production (78%). Food safety hazards were more pronounced in S. epidermidis than in the three other species regardless the food or clinical origin of the strains. Seventy-six percent of the strains carrying genes encoding staphylococcal toxin and 69% of strains carrying 5 or more antibiotic determinants belonged to S. epidermidis species. The dominant antibiotic resistance targeted erythromycin, tetracycline and penicillin and were generally traced back to the presence of tetK and blaZ in the two latest cases. Six percent of the food-related strains produced significant amounts of biogenic amines in vitro without any of the corresponding genes detected, reflecting a lack of knowledge on genetic determinants of such production in staphylococci. This work gives a first picture of safety hazards within four species of CNS frequently isolated from food or clinical environment.

Effect of linseed oil supplementation on feed degradation and microbial synthesis in the rumen of ciliate-free and refaunated sheep

The consequences of a 6% linseed oil (LSO) supplementation on fibre digestion and microbial protein synthesis in the rumen were investigated in ciliate-free and refaunated sheep, in order to assess the role protozoa depletion plays in the various effects of oil feeding. The incorporation of LSO in the diet decreased protozoal numbers from 565 × 103 ml−1 to 191 × 103 ml−1 (P < 0.01). It did not modify significantly the rumen fluid kinetics. Defaunation and the supply of LSO shifted rumen fermentation to a higher propionateacetate molar ratio. Digesta flows at the duodenum were measured using ytterbium as a flow marker. The addition of LSO significantly reduced the digestion of hemicelluloses in the rumen, mainly by a direct inhibition of bacterial activity. Flow of microbial matter at the duodenum was estimated using 2,4-diaminopimelic acid or purine bases as microbial markers. Both defaunation and LSO supplementation decreased the amount of OM apparently digested in the rumen and increased the flow of microbial nitrogen at the duodenum, leading to an increase in the efficiency of microbial protein synthesis.

Origin of the Putrescine-Producing Ability of the Coagulase-Negative Bacterium Staphylococcus epidermidis 2015B

ABSTRACT A multiplex PCR method, aimed at the detection of genes associated with biogenic amine production, identified the odc gene encoding ornithine decarboxylase in 1 of 15 strains of Staphylococcus epidermidis. The ability of the positive strain, S. epidermidis 2015B, to produce putrescine in vitro was demonstrated by high-performance liquid chromatography (HPLC). In this strain, the odc gene was detected on plasmid DNA, suggesting that the ability to form putrescine is carried by a mobile element, which explains the fact that the trait is strain dependent within the S. epidermidis species. A 6,292-bp nucleotide sequence harboring the putative odc gene was determined. S. epidermidis ornithine decarboxylase (ODC) showed 60 to 65% sequence identity with known ODCs of Gram-positive as well as Gram-negative bacteria. Downstream of the odc gene, a gene encoding a putative amino acid transporter was found that shared 59% sequence identity with the ornithine/putrescine exchanger (PotE) of Escherichia coli. Cloning and expression of the potE gene of S. epidermis 2015B in Lactococcus lactis demonstrated that the gene product transported ornithine and putrescine into the cells and efficiently exchanged putrescine for ornithine. Analysis of the flanking regions showed high identity levels with different S. epidermidis plasmid sequences, which would confirm the plasmidic location of the odc operon. It follows that the odc and potE gene pair encodes a putrescine-producing pathway in S. epidermis 2015B that was acquired through horizontal gene transfer.

Impact of Gram-negative bacteria in interaction with a complex microbial consortium on biogenic amine content and sensory characteristics of an uncooked pressed cheese.

The impact of Gram-negative bacteria on sensory characteristics and production of volatile compounds as well as biogenic amines (BA) in the core of an uncooked pressed type model cheese was investigated in the presence of a defined complex microbial consortium. Eleven strains of Gram-negative bacteria, selected on the basis of their biodiversity and in vitro BA-production ability, were individually tested in a model cheese. Four out of 6 strains of Enterobacteriaceae (Citrobacter freundii UCMA 4217, Klebsiella oxytoca 927, Hafnia alvei B16 and Proteus vulgaris UCMA 3780) reached counts close to 6 log CFU g⁻¹ in the model cheese. In core of cheeses inoculated with Gram-negative bacteria, only slight differences were observed for microbial counts (Enterococcus faecalis or Lactobacillus plantarum count differences below 1 log CFU g⁻¹), acetate concentration (differences below 200 mg kg⁻¹) and texture (greater firmness) in comparison to control cheeses. Cheese core colour, odour and volatile compound composition were not modified. Although ornithine, the precursor of putrescine, was present in all cheeses, putrescine was only detected in cheeses inoculated with H. alvei B16 and never exceeded 2.18 mmol kg⁻¹ cheese dry matter. Cadaverine was only detected in cheeses inoculated with H. alvei B16, K. oxytoca 927, Halomonas venusta 4C1A or Morganella morganii 3A2A but at lower concentrations (<1.05 mmol kg⁻¹ cheese dry matter), although lysine was available. Only insignificant amounts of the detrimental BA histamine and tyramine, as well as isopentylamine, tryptamine or phenylethylamine, were produced in the cheese model by any of the Gram-negative strains, including those which produced these BA at high levels in vitro.

Effect of genetic potential and level of feeding on milk protein composition

Two groups of 15 multiparous cows in mid-lactation were used in a Latin square design experiment with 4-week experimental periods. The genetic milk protein concentration level was high in the first group and low in the second. Each group of cows was given in a random order three feeding levels that covered 85, 100 and 115% of energy requirements and 90, 110 and 125% of nitrogen requirements, respeetively. In both groups, increasing level of feeding induced a significant increase in milk yield (+ 2.4 kg/d between lowest and highest levels) and in protein concentration (+ 1.7 g/kg). The proportion of paracasein in total proteins was not altered by either genetics or nutrition. The proportion of casein in total proteins was slightly increased by 0.5 percentage points (P < 0.05) with the intermediate level of feeding. Plasmin and plasminogen activities were not significantly modified by the genetic milk concentration level. Plasmin activity significantly increased with nutrient supplementation, but only in animals of low genetic potential (+ 21% between low and high levels, P < 0.01). Casein composition was not significantly altered by the genetics or level of nutrition. Over the whole range of individual measurements taken (n = 90), the relationships between casein or paracasein and total protein concentrations were linear and very narrow (R2 = 0.92 and 0.95, respectively). The proportion of casein or paracasein in total proteins significantly decreased as plasmin activity increased.

Characterizing ripening of Gruyère de Comté: Influence of time × temperature and salting conditions on eye and slit formation

Abstract Gruyere de Comte is a Swiss-type, hard cheese, made of raw milk and produced exclusively in a limited area, in the East of France. Comte cheese is regularly salted on the surface throughout the ripening period, with simultaneous rubbing. This process leads to an important bacterial surface growth (10 11 CFU/g), a production of NH 3 (50 ppm), and a regular increase of the salt level. The kinetics of salt intake and its final level strongly influence the propionic fermentation and eye formation, as well as the primary proteolysis. Hydrolysis of α s 1 - CN decreases regularly with the salt level, whereas plasmin activity is maximum at 2·3% salt/moisture. Variations in the ripening time × temperature cycle influence the rate of hydrolysis of α s 1 - CN and to a lesser extent of β-CN. The improntance of the propionic fermentation increases with the temperature of ripening. The modifications of the rheological properties of Comte during ripening appear to be time × temperature dependent. Slit formation is closely related to cheese composition and also to the nature and extent of proteolysis. It was shown that when several of the following physicochemical characteristics: low moisture, high fat, high NaCl, low calcium and high level of protein breakdown are associated, the rheological properties are modified in such a way (increase of the shortness), that the cheeses are more likely to have slits. The relative importance of the anisotropy of the rheological properties explains the direction of the slits which are always parallel to the surfaces. A high correlation (r = 0·96) was found between the log of the fracture strain difference and the slit intensity. From all the results obtained in various studies, it appears that monitoring both the nature and extent of proteolysis through milk quality, cheese making, and ripening conditions, is a major challenge to produce regularly high quality Comte cheese.

Behavior of Escherichia coli O26:H11 in the presence of Hafnia alvei in a model cheese ecosystem

This study was designed to evaluate the capacity of three Hafnia strains to inhibit the growth of an E. coli strain O26:H11 in an uncooked pressed model cheese, in the presence or absence of a microbial consortium added to mimic a cheese microbial community. Inoculated at 2 log CFU/ml into pasteurized milk without Hafnia, the E. coli O26:H11 strain reached 5 log CFU/g during cheese-making and survived at levels of 4 to 5 log CFU/g beyond 40 days. Inoculated into milk at 6 log CFU/ml, all three tested Hafnia strains (H. alvei B16 and HA, H. paralvei 920) reached values close to 8 log CFU/g and reduced E. coli O26:H11 counts in cheese on day 1 by 0.8 to 1.4 log CFU/g compared to cheeses inoculated with E. coli O26:H11 and the microbial consortium only. The Hafnia strains slightly reduced counts of Enterococcus faecalis (~-0.5 log from day 1) and promoted Lactobacillus plantarum growth (+0.2 to 0.5 log from day 8) in cheese. They produced small amounts of putrescine (~1.3 mmol/kg) and cadaverine (~0.9 mmol/kg) in cheese after 28 days, and did not affect levels of volatile aroma compounds. Further work on H. alvei strain B16 showed that E. coli O26:H11, inoculated at 2 log CFU/ml, was inhibited by H. alvei B16 inoculated at 6 log CFU/ml and not at 4.5 log CFU/ml. The inhibition was associated neither with lower pH values in cheese after 6 or 24h, nor with higher concentrations of lactic acid. Enhanced concentrations of acetic acid on day 1 in cheese inoculated with H. alvei B16 (4 to 11 mmol/kg) could not fully explain the reduction in E. coli O26:H11 growth. A synergistic interaction between H. alvei B16 and the microbial consortium, resulting in an additional 0.7-log reduction in E. coli O26:H11 counts, was observed from day 8 in model cheeses made from pasteurized milk. However, E. coli O26:H11 survived better during ripening in model cheeses made from raw milk than in those made from pasteurized milk, but this was not associated with an increase in pH values. In vitro approaches are required to investigate the mechanisms and causative agents of this interaction. H. alvei B16 appears to be a promising strain for reducing E. coli O26:H11 growth in cheese, as part of a multi-hurdle approach.

A simple micro-batch ion-exchange resin extraction method coupled with reverse-phase HPLC (MBRE-HPLC) to quantify lactoferrin in raw and heat-treated bovine milk

Lactoferrin is an iron-binding cationic glycoprotein (pI = 8.7) beneficial for mammal health, especially udder and milk preservation. A new simple two-step method of quantification was developed. Lactoferrin in 1 mL of bovine skim milk was first adsorbed onto 100 mg of macroporous sulfonated-resin at pH 6.8 by rotary stirring for 90 min at 20-25 °C. After washing the resin, lactoferrin was desorbed using 1 mL of 2 M NaCl containing phenylalanine as a dilution marker, then fully resolved and quantified by RP-HPLC at 220 nm using a wide-bore C4 silica column. This robust, inexpensive and flexible method improves selectivity (no protein interference) and sensitivity compared to previous HPLC methods. In-laboratory validation demonstrated its linearity (25 to 514 µg Lf mL-1), accuracy (110 to 98% recovery), and precision (<4%), which were comparable to immuno-based methods. The results for individual raw cows milk were strongly correlated with results using an ELISA test.