| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Unité Mixte de Recherche Génie et Microbiologie des Procédés Alimentaires, Institut National de la Recherche Agronomique (INRA), F-78 850 Thiverval-Grignon, France
1 Corresponding author: perlat{at}grignon.inra.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: chamber atmospheric composition • CO2 and O2 • microbial and biochemical aspects • mold soft cheese ripening
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Environmental factors (temperature, hygrometry, and gaseous atmospheric composition in ripening chamber) play a determining role in microbial development and enzymatic process (van den Tempel and Nielsen, 2000; Bonaïti et al., 2004). For French mold soft cheeses, ripening temperature is kept at 12°C. Although increasing the ripening temperature (3 to 4°C) is one of the easiest and most economically feasible strategies to accelerate ripening (Nunez et al., 1991), soft cheeses (in which the surface microflora has important enzymatic activities) cannot undergo substantial temperature increases without significantly modifying cheese quality (poor texture, unpleasant taste; Ramet, 2000). For mold cheeses, ripening chamber hygrometry is close to 95%. Consequently, a part of the cheese water evaporates to the atmosphere leading to a decrease in cheese mass and water content (Hardy et al., 2000). If relative humidity (RH) is higher than this value, P. camemberti development is poor and mycelia are brown. Moreover, if RH is lower than 95%, the water loss is excessive; therefore, the resulting cheese is too dry.
Although more data are available on packaging films and their effects on soft cheeses, little information is available on the effect of the gaseous environment in ripening chambers on the growth of microorganisms and its consequences on cheese qualities. Carbon dioxide extends the shelf life of cheeses by inhibiting mold growth (Doyon et al., 1997; Pintado and Malcata, 2000a,b; Colchin et al., 2001). However, the atmospheric composition of the ripening chamber has been shown to change the overall ripening process due to its effects on microorganism physiology (Kiermeier and Wolfseder, 1972a,b; Champagne et al., 2003). A low O2 concentration reduces aerobic microorganism metabolisms (Champagne et al., 2003), whereas addition of CO2 to the atmospheric composition may increase bactericidal or fungicidal effects (Taniwaki et al., 2001). Moreover, according to Champagne et al. (2003), CO2 and O2 concentrations used in experiments focused on postwrapping changes (10 to 30% CO2 and 10 to 40% O2) are very different from those used in industrial ripening chambers. Little work has involved the effects of CO2 and O2 on the growth of flora throughout the ripening process and the consequences on cheese quality (Kiermeier and Wolfseder, 1972a,b; Weissenfluh and Puhan, 1987; Roger et al., 1998). Kiermeier and Wolfseder (1972a,b) have shown that oxygen consumption is correlated to CO2 production during the ripening of soft mold-ripening cheeses. Weissenfluh and Puhan (1987) and Roger et al. (1998) have shown that respiration rate of P. camemberti decreased when O2 partial pressure was reduced. However, the growth dynamics of strains used to ripen Camembert cheeses and their contribution to overall ripening need more investigation, in relation to time and gaseous atmospheric composition of chamber.
The aim of this study was to monitor microflora development and biochemical changes of Camembert-type cheeses throughout all the ripening process (from d 0 to 40), in relation to gaseous atmospheric composition of the ripening chamber (from d 0 to 16), especially CO2 concentration. The approach involved manufacturing a Camembert-type cheese on a pilot scale under aseptic environmental conditions, seeding with a ripening culture (K. lactis, G. candidum, P. camemberti, or B. aurantiacum). After draining and molding, the cheeses were ripened under 5 different gaseous atmospheric compositions and under controlled conditions of temperature and RH. The effects of CO2 and O2 levels during chamber ripening were studied in relation to microbial growth, biochemical dynamics, and appearance of cheeses. The set of cheeses used here was the same as used in the companion study (Picque et al., 2006).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The preparations of K. lactis and B. aurantiacum were carried out as previously described in Leclercq-Perlat et al. (2004). Geotrichum candidum (strain D) and P. camemberti (strain R) were lyophilized. Their preparation was described in Leclercq-Perlat et al., 2004. One week before use, a single tube of each starter was thawed and the viable cell concentration of each was determined. For K. lactis and G. candidum, this was done by surface plate counting on yeast extract glucose chloramphenicol (YEGC, Biokar, Beauvais, France) after 2 to 3 d of incubation at 25°C. For B. aurantiacum it was determined by using a lactate medium after 5 to 6 d of incubation at 25°C, and for P. camemberti by using YEGC containing 6 g of sodium chloride/100 g (6 to 8 d; 20°C).
Flora Danica lyophilizate (CHN11, Chr. Hansen, Arpajon, France) was the chosen mesophilic lactic acid bacteria culture. It was prepared according to Leclercq-Perlat et al. (2004). Before each cheese making run, 1 aliquot of the starter was cultured at 22°C in 1.5 L of sterile skim milk for 16 h.
The cheeses were prepared on a pilot scale under aseptic conditions in a sterilized 2-m3 cheese-making chamber in which coagulation, cutting, draining, and shaping of the curds were carried out (Leclercq-Perlat et al., 2004). For each cheese-making trial, 100 L of milk were used to make 45 cheeses. The raw milk was obtained from the experimental farm of the Institut National Agronomique Paris-Grignon (Thiverval-Grignon, France). It was standardized at 2.9 g of fat/100 g by mixing skim and full-cream milk, and contained more than 3.2 g of protein/100 g. The cheese making and ripening of the Camembert-type cheeses was described in detail previously (Leclercq-Perlat et al., 2004). Ripening took place under the same conditions (13° C for all chamber ripening, 24 h at 85% RH, then under defined RH from d 2 to 14, and thereafter at 85% RH) except for RH (95 ± 2% instead of 93 ± 1%) and the wrapping time (on d 16 instead of d 14). However, the system used to maintain RH in the chamber in previous experiments was changed to account for salt evaporation. Indeed, in this study, RH was controlled without salt baths (potassium nitrate, KNO3 in saturation) because during the preliminary CO2 runs, a significant quantity of water evaporated into the atmosphere and was not absorbed by the salt baths. Thus, a system based on the principle of cold point, involving a cooling and heating system plunged into a water tank, was used.
From d 0 to 16, cheeses were ripened in chamber under 5 controlled atmospheres as described in detail I the companion article (Picque et al., 2006): 1) continuously renewed atmosphere (CRA) under which CO2 and O2 levels remained at 0 and 21%, respectively; 2) periodically renewed atmosphere (PRA) under which the ripening chamber was sealed but each day, if necessary, the CO2 concentration was decreased to 2% by air injection (O2 decreased to 17 to 18%); 3) no renewed atmosphere (NRA) under which the chamber was sealed during chamber ripening (accumulation of CO2 up to 20%, disappearance of O2); 4) CO2 concentration maintained at 2% from d 0 to 16, with O2 close to 17 to 18% (2CO2); and 5) CO2 concentration kept constant at 6% from d 0 to 16, with O2 ~12 to 13% (6CO2).
Each controlled atmosphere trial was carried out in duplicate, except NRA. Only one run was carried under NRA because the rind color was slightly brown, and cheeses exhibited another major flaw: the surface was greasy and irregular; the underrind was completely runny.
For all runs, the effects of CO2 and O2 on the microbiological and biological aspects of cheeses were recorded. Under PRA and NRA, CO2 concentration increased in relation to time and it was calculated by determining the average for each phase of the timeline described below.
For each run, all the cheeses were turned on d 5. Between d 0 and 16, 1 cheese was removed daily for analysis. On d 16, all remaining cheeses were wrapped, and left to finish ripening at 4°C under unknown RH until d 40. During this period (d 16 to 40), 1 cheese was removed weekly for analysis.
Microbial Counts and Biochemical Analysis
These analyses were carried out as previously described by Leclercq-Perlat et al. (2004).
Viable cell counts of K. lactis and of G. candidum in both the rind and the core were measured by counting in YEGC dishes (25°C; 2 to 3 d). Spores of P. camemberti in the rind were determined by counting on YEGC medium containing 60 g of NaCl/L to limit G. candidum growth (20°C, 6 to 8 d). Viable cells of B. aurianticum in the rind were determined on amphotericin-containing lactate agar (25°C, 7 d).
Dry matter and pH as well as lactose and lactate concentrations in both the rind and the core were measured. Nitrogen fractions [total N, acid-soluble nitrogen (ASN), NPN, and free ammonium concentrations] were measured for the rind only. The thickness of the cheese creamy underrind was measured using the first 15 mm at 6 points on each face. Then, the arithmetic average of these 12 measurements was calculated. The appearance of the cheese (microorganism cover, surface appearance, and texture of the underrind) was estimated as proposed by Bonaïti et al. (2004).
Statistical Analyses
The evolutions of microbiological concentrations, biochemical kinetics, and the cheese appearance during the entire ripening period (d 0 to 40) were described by 2 or 3 descriptors. Except for K. lactis, the descriptors were determined by a Weibull model (using Statistica for Windows, Statsoft, Maisons-Alfort, France). The descriptors for K. lactis were calculated by linear regression. All the statistical tests were carried out using the 95% confidence level (P < 0.05). The reproducibility of the microbiological (viable cell counts) and the biochemical measurements (carbon substrate concentrations, pH, DM, nitrogen fractions, and underrind thickness) were studied according to Leclercq-Perlat et al. (2004). This analysis showed that 1) the hypotheses of equality of the means were satisfactory (1 – 
= 0.95), and 2) the risks of a false interpretation () were less than 0.05. For all statistical analyses, explained variance (Weibull modeling; R) or correlation coefficient (linear regression; R) was between 0.90 and 0.99, at a 95% confidence level, except for NPN maximal rate of PRA and 6CO2. For this descriptor, explained variance (R) was between 0.81 and 0.83 at a 95% confidence level. If R was higher than 0.9, the descriptor was highly significant, and if it was between 0.81 and 0.90, the descriptor was simply significant.
For each single measurement, and each sampling day, a 1-way ANOVA was used to determine the influence of CO2; ANOVA was carried out with Newman-Keuls test at a P-value of 0.01.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Microbiological Changes in Ripening Cheeses
For the average of 2 runs, the log10 of ripening microorganism counts of the rind vs. time were similar to those described in a previous study (Leclercq-Perlat et al., 2004). In the core, the overall changes in K. lactis and G. candidum concentrations were similar to those in the rind. However, their concentrations were around 250 times lower than in the rind (data not shown). As expected, P. camemberti and B. aurantiacum did not grow in the core.
Three phases were apparent in the timeline of K. lactis growth and several kinetic descriptors were calculated (Table 1
). Firstly, between d 0 and 5, the specific growth rate (µmax) during the exponential phase was 0.46 ± 0.04 d–1 and the associated generation time was close to 1.5 d. Secondly, from d 5 to 19, the viable count (Xmax) remained constant, close to 4 x 107 cfu/g. Thirdly, after d 19 and until d 40, the viable K. lactis count diminished slowly and exponentially with a specific death rate (Dmax) equal to 0.100 ± 0.007 d–1.
|
, Table 1
). Firstly, growth of G. candidum was measurable only after d 3 (lag phase). The second phase, between d 4 and 8, had a maximum specific growth rate (µmax) on d 6 equal to 0.7 ± 0.2 d–1 and associated generation time was about 1.0 d. In the third phase (stationary phase), between d 9 and 40, the maximal viable count remained constant at ~1 x 107 cfu/g of DM.
|
, Table 1). Then, from d 6 to 19, growth was slow and close to exponential with a specific growth rate (µmax) of 0.23 ± 0.05 d–1 occurring on 9 ± 3 d (Table 1). After d 20 and until d 40, these spore concentrations remained constant, close to 7.5 x 105 cfu/g of DM. Because the sampling technique destroyed the mycelia, only visual observation of the surface was possible; P. camemberti mycelia were seen from d 4. Between d 5 and 12, the mycelia grew and uniformly covered the entire cheese surface by d 12.
Three phases were apparent in the timeline of B. aurantiacum growth as previously shown (Leclercq-Perlat et al., 2004); B. aurantiacum counts in the rind remained nearly constant (at 6.3 x 104 cfu/g of DM) during the first 10 d of ripening. Between d 12 and 19, this species grew exponentially with a specific growth rate (µmax) equal to 0.46 ± 0.07 d–1 and the time taken to reach µmax reached 15 ± 1 d (Table 1). From d 20 to 40, B. aurantiacum counts remained constant, close to 3.4 x 108 cfu/g.
Biochemical Changes in Ripening Cheeses
pH dynamics under CRA conditions were similar to those observed in previous studies (Leclercq-Perlat et al., 2004) and showed 3 phases. Firstly, between d 0 and 3, pH values remained constant at 4.60 ± 0.06. Secondly, between d 5 and 7, pH values increased strongly with a deacidification rate (Vmax) close to 1.3 pH unit/d, reaching 7.5 on d 7. During this phase, rind pH presented an important standard deviation due to important differences between the 2 sides of each cheese. Thirdly, from d 7 to 40, rind pH did not significantly change, remaining close to 7.6 ± 0.2. The pH of the core changed slowly during ripening (Table 2). From d 0 to 8, the pH remained constant at 4.65 ± 0.05, and it increased slightly until the end of ripening, reaching 7.0 ± 0.2 on d 40.
|
). Lactose concentration was negligible on d 6 in the rind and on d 10 in the core.
|
). Lactate concentration became negligible on d 16 (wrapping) for the rind and on d 40 for the core.
The descriptors of changes in ASN and NPN indexes as well as NH3 concentrations of the rind during all the ripening (d 0 to 40) are shown in Table 4. The ASN index remained constant at 18 ± 4 g of ASN/100 g of total N (%) during the first 3 d of ripening. Then, it increased from d 3 to d 6 at a maximal rate of 26%/d, reaching 100% on d 8. Afterwards, and until d 40, ASN remained constant at 100%.
|
; Table 4).
|
, Table 4).
Changes in Cheese Appearance
The rind color of the cheeses provided information about the ripening and microorganism dynamics. From d 0 to 1, color was uniformly white and bright. From d 1 to 5, cheeses became slightly cream colored and extremely dull due to growth of K. lactis. On d 4, growth of P. camemberti aerial mycelia was observed. These mycelia began to grow in spots on the aired face of cheeses. From d 10 to 16 (wrapping), rind color was white with short and uniform mycelium growth typical of P. camemberti. After d 10, the rind was close to 3 mm thick and the surface color remained white. However, on d 40, the outer circumference of each side of the cheeses was light cream colored, but no relationship with B. aurantiacum growth could be established.
The thickening of the underrind in relation to ripening time provided information about the ripening level because the underrind was linked to proteolysis and lipolysis dynamics (Table 5, Figure 3). Between d 0 and 5, the underrind thickness values remained negligible. Between d 5 and 14, thickness increased slightly, reaching about 2.0 mm on d 16. After d 16 and until the end of ripening (d 40), thickness greatly increased with a maximal rate (Vmax) equal to 0.63 ± 0.09 mm/d (Table 5), which occurred on d 27. On d 40, the underrind thickness was maximal and corresponded to about half the thickness of the cheese (15 mm).
|
|
lists the 3 descriptors (µmax, Xmax, Dmax) for growth of K. lactis as well as the 2 descriptors (µmax, Xmax) for growth of G. candidum, P. camemberti spores, and B. aurantiacum, in relation to ripening chamber CO2 concentration. Under PRA and NRA conditions, CO2 concentration increased in relation to time, and it was calculated by determining the average for each phase of the time line described below.
Whatever the CO2 concentration, K. lactis growth began on d 0 and finished between d 4 (CRA, NRA, and 6CO2) and d 6 (2CO2, PRA, Table 1). The mean specific growth rate (µmax) was statistically 0.5 ± 0.1 d–1. The concentration of CO2 in the chamber atmosphere did not have any statistically significant effects on µmax and Xmax. Except under NRA conditions, the death phase began around the end of chamber ripening (d 14 to 19) and it finished on d 40. It was characterized by mean specific death rates between 0.062 and 0.10 d–1. Under NRA, the K. lactis death period was completely shifted in relation to ripening time. It took place between d 8 and 16, when CO2 completely replaced O2, and out of all the atmospheric conditions, the death rate of K. lactis was the highest (0.122 ± 0.007 per d) in this case (NRA). The absence of O2 probably explained this highest rate and earliest mortality. The overall dynamics of K. lactis depended on CO2 concentration, and in particular, the timing of the death phase.
The ANOVA (data not shown) showed no effect of CO2 concentration on dynamics of K. lactis.
Regardless of CO2 concentration, growth of G. candidum vs. time exhibited the same overall changes (Figure 1A). Concentration of CO2 present in the ripening chamber showed very significant effects on G. candidum growth descriptors (µmax, Xmax; Table 1). Maximal growth rates of G. candidum were about twice as high as those obtained under CRA conditions (0% of CO2), but the associated time (time for µmax) was about the same value due to standard deviation. The lowest Xmax value was obtained under CRA. Under 6CO2 conditions the maximal concentration of G. candidum rose to 145 x 107 cfu/g. However, when CO2 concentration was higher than 6% under NRA, G. candidum Xmax (32 x 107 cfu/g) was reduced.
Analysis of variance (Table 6) confirmed that these dynamics were the lowest under CRA. The presence of CO2 during chamber ripening promoted G. candidum growth, and after wrapping (d 16) the effect of CO2 on cell concentrations (Xmax) was obvious: Xmax presented a maximum in relation to CO2 concentration (Figure 1A).
|
, Table 1). Indeed, out of all the atmospheric conditions, the highest maximal sporulation rate value was under 2CO2. The highest CO2 concentration (NRA) had a significant effect on the maximal Xmax value, which reached about 2.5 x 106 spores/g (around 3 times higher than for the other conditions).
Analysis of variance (Table 6) confirmed that for P. camemberti sporulation, 1) CRA and 6CO2 showed the same progressions; 2) 2CO2 runs had the highest progressions between d 3 and 16; and 3) PRA (CO2 between 2 and 8%) was intermediate between 2CO2, 6CO2, and NRA.
Regardless of CO2 concentration, B. aurantiacum counts presented several progressions vs. time (data not shown). Under all conditions except NRA, bacterial concentration showed a lag phase from d 0 to 9. After this lag phase, and until d 16 (6CO2), d 20 (CRA, PRA) or d 33 (2CO2), B. aurantiacum grew with a maximal growth rate (µmax) depending on the CO2 concentration during chamber ripening (Table 1). In the absence of CO2 (CRA), µmax was the highest (0.46 ± 0.07 d–1)—35% higher than that obtained when the CO2 concentration was under 2 or 6%. After this growth phase, B. aurantiacum concentration remained constant, but its value depended on the CO2 level. When CO2 levels were less than 6%, Xmax was between 3 x 108 and 10 x 108 cfu/g. When CO2 levels were 6.0% (6CO2), Xmax was divided by about 300 in relation to 2CO2 and CRA. Under NRA (the highest CO2 level), B. aurantiacum did not grow, and its concentration remained constant, about 4 x 104 cfu/g during chamber ripening. After d 16 and until 40, no bacteria were found in the rind. The ANOVA confirmed that the highest CO2 concentrations present in the ripening chamber significantly inhibited B. aurantiacum growth (data not shown).
Biochemical Changes in Ripening Cheeses in Relation to Atmospheric Conditions
Table 2 shows the descriptors of cheese rind pH (Vmax and pHmax) in relation to CO2 concentration. Regardless of the CO2 concentration used during chamber ripening, no significant differences of pH descriptors were observed. The ANOVA confirmed that no significant differences existed between these dynamics (data not shown).
The atmospheric conditions of ripening (CO2 level) did not have a significant effect on lactose consumption rate in the rind (CLO, mean value 18 ± 3 mmol/kg of DM per d) and lactose disappearance in the core (DLO, mean value = 12 ± 2 mmol/kg of DM per d) at 95% confidence level (Table 3). The ANOVA confirmed that there were no significant differences (data not shown).
Initial presence (d 0) of CO2 seemed to have slightly negative effect on lactate consumption by the microorganisms present on the surface (CLT), but it did not have any effect on lactate disappearance rate (DLT), consecutive to lactate consumption in the core by G. candidum and K. lactis (Leclercq-Perlat et al., 2004), as well as to its diffusion from the core to the rind. Indeed, CLT of CRA and PRA was higher than for the other conditions (Table 3). Under NRA (results not shown) lactate consumption in the rind and in the core was stopped on d 10 and 23, reaching about 97 and 10 mmol/kg of DM, respectively. In the chamber, between d 10 and 16, CO2 concentration was the highest: equal to 19.2% on d 10 and remaining constant at 20% from d 11 to 16. At these CO2 levels, all microorganism activities slowed down. Lactate diffusion from the core to the rind continued until d 23 and stopped when the 2 lactate concentrations were equal. This could explain why these lactate rates (CLT and DLT) obtained under NRA were lower than under the other atmospheric conditions. The ANOVA confirmed that the lactate dynamics in the rind were similar under 6CO2, CRA, 2CO2, and PRA. The dynamics in the rind and in the core were the lowest under NRA conditions (data not shown).
Regardless of the CO2 concentration used during chamber ripening, the kinetic increases occurred during the same ripening time (from d 3 or 4 to d 6 or 7). No significant difference of initial ASN concentration was observed (Table 4). The maximal rate (Vmax) was the highest for 2CO2 and PRA conditions (Table 4). The maximal rate was lowest when CO2 level was negligible (CRA) or the highest when CO2 level was maximal (NRA). Moreover, when Vmax was the highest, the time for which ASN level was 100% was the lowest (2CO2); indeed, it seemed that this proteolytic index became quickly maximal. However, ANOVA showed that there were no significant differences (data not shown).
When CO2 concentration was lower or equal to 2% (CRA, 2CO2), the NPN rates (Vmax) (Figure 2A, Table 4) were between 4 and 5 times higher than those obtained for the other runs (CO2 level 
3%), and the time for which NPN level was maximal was the lowest (d 6 to 7). No significant differences of the maximal NPN were found regardless of CO2 concentration in the chamber. Under 6CO2 and NRA (the highest CO2 level) conditions, the maximal concentration of NPN was obtained after wrapping, on d 19 and 18, respectively.
Analysis of variance (Table 7) confirmed that under PRA, the NPN dynamic was the lowest. The NPN dynamics of CRA (starting on d 9) were the same as for 2CO2 (starting on d 7) but with a delay of 2 d. Under 6CO2 and NRA conditions, there was a weak acceleration of NPN production, from d 19 and d 16 until d 40, respectively.
|
; Table 4). Firstly, no production of NH3 was observed before d 7. Secondly, except under NRA, NH3 production occurred during chamber ripening and finished with wrapping (d 16). Under NRA, NH3 increase occurred from d 12 until 8 d after wrapping. Thirdly, an initial presence of CO2 or an important concentration of CO2 induced lower production of NH3.
Whatever the atmospheric conditions, the ANOVA study detected no significant effect from d 0 to 6 (Table 7B). Under NRA, regardless of day, the NH3 concentration was the lowest. From d 6 until 40, CRA and PRA runs presented the fastest dynamics and the highest values of NH3 concentrations. Under 2CO2 or 6CO2, ammonium concentration dynamics formed an intermediate group.
Influence of Atmospheric Composition on Cheese Appearance
Figure 3 shows the progression of underrind thickness in relation to ripening time (from d 0 to 40) and to atmospheric conditions used during chamber ripening. Table 5 presents the duration of lag period (TUR = 0 mm) and fast increase phase as well as its kinetic descriptors (Vmax, time for Vmax) and the type of ripening.
Two different ripening progressions were observed (Figure 3). In the presence of CO2 on d 0 (2CO2 and 6CO2), the underrind thickness increased greatly between d 6 and 26. After this phase, the underrind thickness remained constant, reaching about half the thickness of the cheeses on d 34. The ripening of these 2 tests was termed fast-slow. For the runs free of CO2 on d 0 (CRA, PRA, and NRA), the underrind thickness increased at a low rate from d 6 to 16. After that, and until d 40, it increased quickly, reaching half the thickness of the cheeses on d 40; this ripening was termed slow-fast. The CO2 concentration of atmospheric composition in the chamber did not have a significant effect on Vmax (Table 5) but it did have a significant effect on time associated with Vmax, as can be seen by the shape of the curves (Figure 3).
When CO2 concentration was higher than 6% during the ripening period (PRA, 6CO2, NRA) the underrind was too runny. The higher the CO2 concentration and the longer the time in ripening, the poorer the underrind quality.
Regardless of CO2 concentration, P. camemberti mycelium began to grow on the same day (d 5). However, when growth started, appearance of cheese surface, color, and density of mycelium as well as microorganism growth depended on CO2 level. When CO2 concentration was 0 or 2%, the appearance of the cheese surface corresponded to the 3 targeted visual criteria (white or light cream color, density of mycelium, and the uniformity of these 2 criteria). Increases in CO2 concentration above 2% resulted in worse cheese surface appearance.
On d 12, under NRA conditions, the cheese color was gray with some brown spots not significantly distributed; the cheese edges were also brown. The aerial mycelia of P. camemberti were no longer observed in the middle of the surface. Concerning color and appearance, the cheese surface was similar to G. candidum biofilms found on cheese ripened by this yeast. Moreover, the cheese completely lost its shape: the rind collapsed all around the core such that the cheese acquired a bowler-hat shape; wrapping did not change the shape. Under 6CO2 and PRA conditions, the cheese color was white and gray and the central core collapsed completely. Moreover, the CO2 level present in the chamber de graded the shape. The rind center was more or less collapsed, and the experts characterized this underrind as runny.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
20%) and low levels of O2 (
1%) in the chamber, which induced lysis of K. lactis cells. In fact, K. lactis growth was not sensitive to CO2 and O2 concentrations. However, its death (period and rate) was related to atmospheric conditions (CO2 and O2).
In absence of CO2, G. candidum dynamics were the lowest and, in presence of 6% of CO2, they appeared maximal (Table 6). These results were in accordance with those of Wells and Spalding (1975). Indeed, those authors have shown that G. candidum is able to grow under microaerophilic conditions and under elevated CO2 levels, and that it is stimulated by low O2 and high CO2 atmospheres. Moreover, Couriol et al. (2001) have shown that CO2 production was associated with G. candidum growth whereas Adour et al. (2002) have considered that G. candidum produced lower CO2 levels than P. camemberti mycelia in liquid media and pure cultivation.
Sporulation of P. camemberti was also increased by significant CO2 concentration in the chamber (>6%). The specific rate (µmax) was maximal for a CO2 level close to 2% (2CO2), whereas the maximal spore concentration (Xmax) was slightly higher under 6CO2 or NRA (Figure 1B). Using ANOVA, P. camemberti spore dynamics appeared maximal under 2CO2 and PRA (Table 6). These results are in accordance with those of Roger et al. (1998). Roger et al. (1998) have shown that a CO2 concentration of about 4% allowed for the best development of P. camemberti mycelia on Brie cheeses, but only if the O2 level was >16%. Similarly, Magan and Lacey (1984) have shown that CO2 concentrations close to 5% with low O2 concentrations stimulate growth of some Penicillium species on wheat extract medium if water activity and temperature are correctly chosen. Adour et al. (2002) have also shown that under their study conditions (liquid media, pure culture), CO2 promoted G. candidum growth (Table 1) but disadvantaged P. camemberti mycelium development. The poor development of P. camemberti visually observed for CO2 higher than 3% could also be explained as proposed by van den Tempel and Nielsen (2000). These authors have shown that 1) G. candidum was adapted to growth at low levels of oxygen and high levels of CO2; and 2) growth and sporulation of Penicillium roqueforti was negatively affected in the presence of G. candidum at high CO2 levels, irrespective of O2 levels. Although van den Tempel and Nielsen (2000) used P. roqueforti, the effects of CO2 and O2 could be similar or more important due to the higher sensitivity of P. camemberti to O2 (Doyon et al., 1997; Roger et al., 1998). The development of P. camemberti (mycelia as well as sporulation) in the presence of G. candidum was related to the CO2 level in chamber ripening. In fact, CO2 changed the equilibrium existing between these 2 strains. Indeed, the higher the CO2 concentration, the more significant the G. candidum concentration was in the stationary phase (Xmax) and the worse P. camemberti mycelium development (visual appearance) was.
In the absence of CO2 (CRA), exponential B. aurantiacum growth occurred on d 9 and continued after wrapping. The pH of the rind was favorable for B. aurantiacum but, surprisingly, its growth occurred 3 d after the rind pH reached its maximum (d 6), whereas in our previous study (Leclercq-Perlat et al., 2004), we have shown that growth of this species started as soon as pH of the rind reached neutral pH. Its maximal growth rate was obtained in absence of CO2 (Table 1). However, maximal concentration (Xmax) was the same if the O2 level in the chamber was higher than 16% (CRA, 2CO2, PRA). In the other runs (6CO2, NRA), its growth (Xmax) was negligible. Lower levels of O2 (<12 to 13%) with high CO2 concentration had significant inhibitor effects on B. aurantiacum evolution; B. aurantiacum is considered a strict aerobic strain, and no growth occurs in absence of O2 (Rattray and Fox, 1999).
No significant effect on carbon substrate dynamics of CO2 levels during chamber ripening was observed on the cheese rind or in the cheese core, except under NRA. Indeed, O2 concentration was high (11 to 18%) and did not limit glycolysis. However, under NRA, the lactate consumption was stopped on d 11 when CO2 was 20% and O2 was negligible. The absence of O2 explains the limitation of these oxidative metabolisms as observed here.
The CO2 level present in the chamber seemed to have no significant effects on ASN dynamics, but a slight effect of CO2 on ASN maximal rates was observed. However, it had a significant effect on NPN maximal rates (Table 7). Indeed, the Vmax values of NPN kinetics were the highest when CO2 concentration was the lowest (0 to 2% for CRA and 2CO2), and the time NPN concentration took to become maximal was the shortest (6 to 7 d). However, the final NPN concentration (measured to NPNmax) was not significantly different. According to Sousa and et al. (2001), P. camemberti is the main agent of amino acid release in Camembert cheeses. This suggests that high CO2 (>2%) could have an inhibitor action on peptidasic activities during ripening.
Under CRA, the quantities of ammonium obtained in the rind were much higher than those found by Leclercq-Perlat et al. (2004). However, they were in accordance with those obtained by von Mrowetz (1979), who studied ammonium contents during Camembert cheese ripening without reporting cheese-making and ripening conditions. The following sequence of events could explain these significant differences. When a saturated salt (KNO3) system was used to regulate RH, a part of KNO3 was vaporized into the atmosphere, and by consequence, was deposited in the rind. This salt absorption induced a significant decrease of water activity in the rind (Gobbetti et al., 1999). These authors have also shown that a decrease of water activity involved a negative effect on the peptidase activity of some lactic acid bacterium strains. This negative effect on proteolysis could be verified for the other strains. Under CRA and PRA, ammonium production was higher and their dynamics were the fastest compared with those of the other atmospheric conditions. Similar to NPN dynamics, this could be due to an inhibition of aminopeptidase and deaminase enzymes by CO2. Indeed, according to Hemme et al. (1982), deaminases (aminooxidases or oxidoreductases) convert amino acids (or their amines) into NH3 and cetonic acids or aldehydes, which requires O2. Many ripening microorganisms, in particular, surface bacteria, are considered to be the most active in proteolysis. Rattray and Fox (1999) have reviewed the proteolytic system of Brevibacterium linens and shown the presence of several aminopeptidases. In our study, when CO2 contents higher than 2% were associated with lower O2 concentration, B. aurantiacum growth was decreased; therefore, inducing lower oxido-oxidase activities. The same explanation could be made for P. camemberti mycelium growth.
The thickness of the creamy underrind provides information about overall proteolysis and lipolysis of the cheeses. The presence or absence of CO2 on d 0 induced different ripening patterns. Under 2CO2 and 6CO2 (CO2 level was not 0% on d 0), the overall ripening measured by the underrind thickness was termed fast-slow, whereas under other conditions (CO2 = 0% on d 0), it was termed slow-fast. However, for a CO2 level above 4%, the cheese underrind was extremely runny on d 40. G. candidum and P. camemberti are well known for their significant lipolytic and proteic systems (Choisy et al., 1997; Boutrou and Guéguen, 2005). The effects of CO2 on their growth (notably 6% on G. candidum and 2% on P. camemberti mycelia) and on their equilibrium might explain the underrind dynamics. B. aurantiacum is also well known for its important exo- and endocellular proteolytic activities (Lecocq and Guéguen, 1994; Rattray and Fox, 1999), and K. lactis has few endocellular proteolytic activities (Klein et al., 2002). These 2 strains have no significant lipolytic activities in comparison to G. candidum and P. camemberti. However, the important death rate of K. lactis in presence of the highest CO2 level could also favor the cheese underrind liquefaction by release of endocellular enzymes in the cheese (Klein et al., 2002). The role B. aurantiacum played was difficult to determine because it grows in the presence of significant CO2 levels when the O2 level was above 12% in relation to CO2 and O2 levels.
The CO2 atmospheric conditions had an effect on overall cheese ripening. Indeed, high concentrations of CO2 (>2%) degraded cheese color and P. camemberti mycelium growth as well as cheese shape and underrind quality (texture).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Received for publication July 19, 2005. Accepted for publication January 31, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Picque, M.-N. Leclercq-Perlat, and G. Corrieu Effects of atmospheric composition on respiratory behavior, weight loss, and appearance of Camembert-type cheeses during chamber ripening. J Dairy Sci, August 1, 2006; 89(8): 3250 - 3259. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||
), PRA (
), 6CO2 (
), and NRA (
) in the ripening chamber (from d 0 to 16, 12°C, relative humidity = 95%). — = modeling results. Atmospheric conditions: CRA = continuously renewed atmosphere under which CO2 and O2 levels remained at 0 and 21%, respectively; PRA = periodically renewed atmosphere under which the ripening chamber was sealed but each day, if necessary, the CO2 concentration was decreased to 2% by air injection (O2 decreased to 17 to 18%); NRA = no renewed atmosphere under which the chamber was sealed during chamber ripening (accumulation of CO2 up to 20%, disappearance of O2); 2CO2 = CO2 concentration maintained at 2% from d 0 to 16, with O2 close to 17 to 18%; and 6CO2 = CO2 concentration kept constant at 6% from d 0 to 16, with O2 ~12 to 13%.
) = CO2 concentration maintained at 2% from d 0 to 16, with O2 close to 17 to 18%; and 6CO2 (