In this study, we developed mathematical models to describe the growth kinetics of Staphylococcus aureus on natural cheeses. A five-strain mixture of Staph. aureus was inoculated onto 15 g of Brie and Camembert cheeses at 4 log CFU/g. The samples were then stored at 4, 10, 15, 25, and 30 °C for 2–60 d, with a different storage time being used for each temperature. Total bacterial and Staph. aureus cells were enumerated on tryptic soy agar and mannitol salt agar, respectively. The Baranyi model was fitted to the growth data of Staph. aureus to calculate kinetic parameters such as the maximum growth rate in log CFU units (rmax; log CFU/g/h) and the lag phase duration (λ; h). The effects of temperature on the square root of rmax and on the natural logarithm of λ were modelled in the second stage (secondary model). Independent experimental data (observed data) were compared with prediction and the respective root mean square error compared with the RMSE of the fit on the original data, as a measure of model performance. The total growth of bacteria was observed at 10, 15, 25, and 30 °C on both cheeses. The rmax values increased with storage temperature (P<0·05), but a significant effect of storage temperature on λ values was only observed between 4 and 15 °C (P<0·05). The square root model and linear equation were found to be appropriate for description of the effect of storage temperature on growth kinetics (R2=0·894–0·983). Our results indicate that the models developed in this study should be useful for describing the growth kinetics of Staph. aureus on Brie and Camembert cheeses.

Experimental Camembert cheeses were made to investigate the effects on browning of the following factors: inoculation with Yarrowia lipolytica, the use of Penicillium candidum strains with different proteolytic activity, the addition of tyrosine, and the addition of Mn2+, thus leading to 16 different variants of cheese. Two physical colour parameters were used to describe browning, depending on the location in the cheeses: a whiteness index for the outside browning (mould mycelium), and a brownness index for the inside browning (surface of the cheese body). Mn2+ promoted a significant increase of browning at both locations, whereas Yar. lipolytica had the opposite effect. Outside browning was significantly more intense when using the Pen. candidum strain with higher proteolytic activity. A significant interaction was found between Yar. lipolytica and Pen. candidum. The yeast had no effect in combination with a low proteolytic strain of Pen. candidum, but significantly reduced proteolysis and browning in combination with a high proteolytic strain of Pen. candidum. We further confirmed that both strains of Pen. candidum were able to produce brown pigments from tyrosine, and thus both are presumably responsible for the browning activity in this type of cheese.

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.