|
|
||||||||
Unité Mixte de Recherche Génie et de Microbiologie des Procédés Alimentaires, F-78 850 Thiverval-Grignon, France
Corresponding author: M.-N. LecClercq-Perlat; e-mail: perlat@grignon.inra.fr.
| ABSTRACT |
|---|
|
|
|---|
Key Words: cheese biofilm color L*C*h° spectro-colorimetry biochemical composition
Abbreviation key: ASN = acid-soluble nitrogen, DCM = decidified cheese medium (pH 5.8), YEG = yeast extract glucose broth
| INTRODUCTION |
|---|
|
|
|---|
The aim of this study was to monitor the development of color of a B. linens strain occurring on the surface of a cheesy medium after deacidification by a yeast strain. The color of the bacterial biofilms was analyzed with the L*a*b* system after 14 or 28 d incubation under controlled ambient conditions.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Preparation of Kluyveromyces marxianus and Debaryomyces hansenii Cheesy Cultures
All equipment was sterilized, and the operations involved in this preparation were carried out under aseptic conditions.
A preculture of each yeast strain was prepared in 20-mL aliquots of yeast extract glucose broth (YEG) (Biokar, Paris, France) in 150-mL-conical flasks. This broth was autoclaved (15 min, 121°C) before inoculation and incubation for 2 d at 25°C on a shaker (250 rpm). Cultures were grown in a broth containing: 10 g of trypcase (Biokar), 5 g of Casaminoacid (Biokar), 5 g of autolytic yeast extract (Biokar), 30 mL of 60% sodium lactate (Prolabo, Paris, France), 5 g of potassium dihydrogenophosphate (KH2PO4) (Prolabo), 5 g of sodium chloride (NaCl) (Prolabo) and deionized water (1 L). After adjusting the pH to 7.0 and pouring 200 mL into a 1-L conical flask, the medium was autoclaved (121°C, 20 min) and inoculated with the yeast preculture (1%). The medium was then incubated while being shaken for at least 48 h at 25°C and at 150 rpm. After incubation, cells were harvested by centrifugation (9000 x g, 20 min) and resuspended in 50 mL of 0.9% NaCl. This suspension was homogenized with an Ultra-Turrax T25 grinder (Janke and Kunkel, IKA-Labortechnik, Staufel, Germany) for 1 min at 20,500 rpm and mixed with 50 mL of sterile 10% glycerin-milk. This suspension was poured into 8-mL tubes, which were frozen at 20°C.
One week before the cheese-making test, a tube was thawed and its viable cell counts determined by plating on a YEG agar and incubating for 48 h at 25°C. The purity of the yeast starter was determined by inoculating on amphotericin lactate agar (Piton, 1988) for 72 h at 25°C.
Preparation of Brevibacterium linens
The preparation of B. linens was the same as for the yeast strains, except a lactate-based medium was used for the preculture (Leclercq-Perlat et al., 2000a). Cell counts were done on the corresponding agar medium.
Cheese Medium Preparation
One-day-old Muenster-type cheeses, obtained from Coopérative Agricole Laitière (BP2; 54450 Blâmont, France), were used to prepare the model medium. They were not salted and were not inoculated with the surface ripening flora. They were frozen at 80°C to prevent enzymatic reactions (Ajit et al., 1976; Guéguen and Schmidt, 1992) and used within 6 mo.
The model cheese medium was prepared in 2 steps: 1) biological deacidification with ripening yeast and 2) inoculation of B. linens on the medium surface. All stages of this preparation were done under a microbiological hood.
Biological deacidification.
After thawing, 240 g of cheese was ground in a mortar and transferred to a 1-L flask containing 160 mL of sterile water (120°C, 20 min). This preparation was heat to 100°C for 10 min. When its temperature was close to 60°C, it was homogenized with an Ultra-Turrax T25 grinder (Janke and Kunkel, IKA-Labortechnik) (1 min, 24,000 rpm). After cooling down to 30°C, one of the yeast starters was inoculated (106 cfu/mL). After a second homogenization (Ultra-Turrax, 40 s, 13,500 rpm), 300 mL of the medium was transferred to a 1-L conical flask with a cellulose cork that was previously sterilized (120°C, 20 min). Then, this cheesy medium (initial pH close to 4.8) was incubated in darkness at 25°C while being shaken (150 rpm) until the pH reached 5.8. Three independent trials were run to test the repeatability of this medium preparation.
Surface seeding of B. linens.
After incubation, the yeast cheesy medium was pasteurized (80°C, 10 min) to kill the yeast cells yet without altering the medium characteristics. After cooling to 70°C, a sterile (120°C, 20 min) agar (8%), containing 7 g of sodium chloride, previously heated to 70°C, was added. This mixture was homogenized (Ultra-Turrax, 24,000 rpm, 30 s) immediately, and 100 mL was transferred to sterile glass dishes (95 mm in diameter, 30 mm high) closed with a glass stopper. This preparation resulted in a cheese-type medium whose thickness was equal to half the thickness of an industrial Muenster cheese. After drying for 1 h, the bacterium was inoculated (3 x 106 cfu/cm2) onto the surface and transferred to a ripening chamber that had been previously thermostated (12°C) and humidified (relative humidity close to 95%). Aseptic conditions were employed at all steps.
Three independent trials were carried out. For each trial, 3 control dishes were not inoculated with the bacterium and were used as a reference sample for color and sterility. Three dishes inoculated with B. linens were sampled daily for analytical purposes between d 2 and 14 and twice a week between d 14 and 28.
Analysis of Bacterial Biofilms
The surface was divided into 4 equal parts. Two quarters, opposite each other, were used for enumeration, while a third quarter was used for pH determinations and the last quarter for color measurements.
Microbial counts.
Under a microbiological hood, the 2 quarters for enumeration were scratched with a sterile scraper and transferred to a flask containing 40 mL of sterile NaCl solution (0.9%). Then, this suspension was homogenized with a sterile Ultra-Turrax grinder (15,000 rpm, 30 s) and poured into Petri dishes. The suspension was plated on the amphotericin-based lactate agar medium surface using a Spiral Plater (Interscience, Saint Nom La Bretèche, France), and dishes were incubated at 25°C for 7 d in darkness. Then, B. linens colonies were counted on the medium and differentiated based on size, appearance, and pigmentation.
pH measurements.
The biofilm pH was determined with a 632 pH-meter (Metrohm, Herisau, Switzerland) fitted with a surface Ingold electrode (Prolabo). pH values were the arithmetic means of 3 measurements from a single quarter.
Color measurements.
Objective measurements of color can be carried out with a spectrocolorimeter (Dufossé et al., 2001). This device mimics the psycho-sensorial mechanism of human perception of color and integrates both the light source spectrum and the object color spectrum into the reflectance spectrum (Dufossé et al., 2001). The reflectance spectrum is further filtered so that the red, green, and blue spectra are extracted as 3 coordinates then processed to provide the 3-dimensional L*a*b* response (Dufossé et al., 2001). Such responses are given in the C.I.E. L*a*b* standards (1978). L*, or lightness, scales the luminous intensity from 0 (black) to 100 (white). Perpendicular to the L* axis, chromatic coordinates a* and b* can be considered using 4 orthogonal directions with +a* (red), +b* (yellow) at 90°, and a* (green) at 180°, and b* (blue) at 270°. Beside these attributes, the (a*, b*) combination also determines the parameters hue angle (h°) and chroma (C*). The h° gives the predominant wavelength composing the color. We considered C* (chroma or saturation equal to
) accounts for the vividness or the color purity from the distance between the (a*, b*) point and the origin. For B. linens, the C* color value appears to be the best estimate for correlation (Guyomarch et al., 2000; Dufossé et al., 2001). Therefore, the L*C*h° system was used.
The biofilm color was determined with a CM-2002 spectrocolorimeter (Minolta, Carrières sur Seine, France), driven by a SpectraMagic 1.01 software (Minolta). The reference illuminant was D65 (standard daylight), and the geometry was d/8: incident light was diffuse and the observation angle was 10°, according to the CIE 15.2 publication and ISO-7724/1 recommendations. Data were logged into the L*C*h° colorimetric system. A Petri dish cover was used as a support and the quarter reserved for color measurement was placed in it with the B. linens-covered face of cheese medium agar facing up, thus facing incident light during color measurement of the biofilms. The quarter reserved for color measurement was analyzed by taking 3 independent measurements in a single session.
Analysis of Cheese Suspensions (Reference and After Yeast Deacidification)
Viable cell counts of yeasts.
For the cheese suspensions before and after pasteurization, viable cell counts of D. hansenii or K. marxianus were measured by counting on YEGC (Biokar). The suspension was plated on the surface of the medium using a Spiral Plater. The plates were incubated at 25°C for 2 to 3 d.
Physicochemical measurements.
For each cheese suspension before and after yeast deacidification and when the pH reached 5.8, measurements of pH, and lactose and lactate concentrations were carried out as described by Leclercq-Perlat et al. (1999). Total N, acid-soluble N, NPN, and ammonium fractions were carried out as described by Leclercq-Perlat et al. (2000b).
Statistical Analyses
The repeatability of the measurements was tested by one-way single analysis of variance, which was performed with the Statgraphics Plus 3.1 software (Stagraphics, Manugistics, Rockville, MD).
When comparing the viable cell counts, pH and color parameters, the significance of the considered factor was estimated by calculating the F-value and the probability (P) that differences between the means were due to chance, with a
risk of 5%.
| RESULTS |
|---|
|
|
|---|
|
|
|
Table 2
also shows the changes in acid-soluble N, NPN, and ammonium concentrations of the cheese media before and after deacidification by D. hansenii and K. marxianus.
The ASN index of the cheesy medium before deacidification 3.5 ± 0.5 g of ASN/100 g of total nitrogen (%) and for D. hansenii it remained more and less at this value during deacidification. But for K. marxianus, the ASN index reached 6.0 ± 0.2%, which was significantly higher (1.7 times) than the inoculation medium.
The same results were found for the NPN indexes and ammonium concentrations. For D. hansenii, the NPN index and ammonium concentration remained close to the inoculated cheese suspension one. However, for K. marxianus the NPN index and ammonium concentration increased, respectively, 2 and almost 6 times, respectively.
Analysis of Bacterial Biofilms
pH changes in bacterial biofilms.
The changes in pH during growth on the deacidified medium appeared to be very dependent on the yeast used for deacidification (Figure 2
). When D. hansenii was the deacidifying yeast, the pH of the surface growth of B. linens increased almost linearly with a mean rate close to 0.110 ± 0.004 units x d1 between d 0 to d 12, to reach 7.00 ± 0.05 at d 12. Then until d 28, this pH continued to increase linearly but at a smaller mean rate (0.030 ± 0.002 units x d1), reaching 7.51 ± 0.08 at d 28. For K. marxianus deacidification, from d 0 to d 8, the pH of the bacterial surface remained close to 5.93 ± 0.05. Then until d 11, the pH decreased at a mean rate of 0.129 ± 0.001 units x d1, reaching 5.45 ± 0.05 at d 11. From d 11 to 21, it increased against at a mean rate of 0.15 ± 0.01 units x d1, reaching 7.1 ± 0.1 at d 21. Then until d 28, it continued to increase very slowly, reaching 7.4 ± 0.1 at d 28.
|
|
Metric saturation evolution (C*).
The chroma values C* of the bacterial biofilms, versus time, are shown in Figure 4
. Regardless of the time, the saturations for DCMDh were slightly greater than those of DCMKm, but the overall change in the 2-color saturations was similar. For DCMDh the metric saturation of B. linens biofilms increased quickly between d 2 and 10, with a mean rate of 1.80 x 0.05 units x d1. Then, between d 10 to 14, the increase was slow, reaching 35 ± 1 at d 14. It remained at this value until d 28. For DCMKm the metric saturation of B. linens biofilms remained constant (15 ± 1) between d 0 and 4. Then until d 12, it increased at a mean rate of 1.59 ± 0.08 units x d1. After d 14 and until d 28, it remained constant (31 ± 1).
|
|
| DISCUSSION |
|---|
|
|
|---|
During deacidification after the lag phase, growth of D. hansenii and K. marxianus were exponential with a generation time (g) of 3.60 ± 0.02 h and 2.18 ± 0.01 h, respectively. These values were higher than those on the cheese surfaces reported earlier (Leclercq-Perlat et al., 1999) and corresponded to those observed by Guéguen and Schmidt (1992) and Soulignac (1995). Growth of K. marxianus was faster than that of D. hansenii. In addition, according to Guéguen and Schmidt (1992), K. marxianus is one of the yeasts that grows quickly and produces a higher biomass. Debaryomyces hansenii was less efficient at deacidifying the curd than K. marxianus. During the pigmentation tests of B. linens on the surface of DCM, the pH of the bacterial biofilm increased from 5.8 ± 0.1 at d 0 to 7.4 ± 0.2 at d 24, regardless of the yeast used in deacidification. However, the increase in pH appeared to be dependent on the deacidification yeast used because a pH of 7.4 was reached 7 d sooner on DCMDh than on DCMKm. This fact suggested that the metabolites produced during deacidification were different and their actions on the bacterial biofilm color were not negligible. The yeast cells were killed by the heating step so, the interactions involved are due to differences in the composition of DCM and not to the biological activities of the yeast during biofilm growth. The same effect was also seen on bacterial biofilm growth. The number of B. linens cells for DCMDh was maximal (close to 4 x 109 cfu x cm2) on d 7, but for DCMKm maximum cell number (close to 2 x 109 cfu x cm2) was reached after 14 d of incubation. Previously Lecocq and Guéguen (1994) and Lecocq et al. (1996)), using another yeast (Geotrichum candidum), showed that extrinsic factors played an important role in the interactions between the yeast and B. linens during cheese ripening. They attributed the variations in appearance of cheeses provided by different factories to the effect of strains and to the interactions between the 2 microorganisms used. However, they did not specify the type of interactions (indirect or intracellular) that could exist between these microorganisms.
Except for the lightness (L*) value, all the colorimetric parameters (a*, b*, C*, and h°) increased more quickly and their level was higher when the deacidification of the cheese medium was realized with D. hansenii. The pigmentation of B. linens was directly related to the yeast used to deacidify the cheese medium. Commercially available surface starters do not reflect the microbial composition of the cheese surface (Bockelmann and Hoppe-Seyler, 2001). Too much emphasis is put on B. linens because of the orange pigmentation, but the red-brown or orange pigments are most likely due not only to orange-pigmented B. linens but also to the yellow-pigmented Arthrobacter species in the surface flora. The mechanisms of how different shades of red are developed are not yet understood.
Due to the extreme difference in color found in the present study, it was decided to look at the interactions between yeast and B. linens. In particular, the carbon and nitrogen concentrations of the cheese media after yeast deacidification (DCM) were compared with the ones of the reference cheese medium (before deacidification) thinking that they played a role in color differences. According to Roostita and Fleet (1996b), the properties of yeasts, which influence their occurrence and growth in cheeses, were fermentation/assimilation of lactose, production of proteolytic and extracellular lipolytic enzymes, utilization of lactic and citric acids, and growth at low temperature.
The substrate levels confirmed that D. hansenii consumed lactose and lactate simultaneously as previously shown by Soulignac (1995) on a liquid medium or by Leclercq-Perlat et al. (1999) on a model cheese. Debaryomyces hansenii 304 used lactose and lactate simultaneously and its growth was quicker with a lactose-lactate mixture than with lactate alone. Kluyveromyces marxianus consumed lactose and lactate in succession (Soulignac, 1995) and their consumption was higher than those of D. hansenii. To simulate the deacidification phenomenon by D. hansenii in soft cheese, Riahi (2003) showed that these yeast cells used lactose for their growth and lactate for their maintenance. Guéguen and Schmidt (1992) found that K. marxianus was one of the yeasts that grew quickly and produced a higher biomass. From our study, we deduced that it may use lactose and lactate for its growth and proteins for its maintenance. This fact would explain that the DCM reached the pH 5.8 quicker (2 times) with K. marxianus than with D. hansenii. D. hansenii strains vary in their ability to use lactate (Petersen et al., 2002). During the growth of K. marxianus on lactose, glucose was detected in the culture medium (results not shown), but Bacci Junior et al. (1996) did not find ß-galactosidase activity, and Berruga et al. (1997) have shown that the rates of lactose disappearance were poorly correlated with intracellular ß-galactosidase levels. Kluyveromyces marxianus cells register maximal intracellular lactase activity, and large fluctuations in enzymatic activity were observed during incubation, presumably as a result of the reciprocal effects between lactase activity and lactose concentrations (Berruga et al., 1997). In addition, Carvalho-Silva and Spencer-Martins (1990) described the transport mechanisms for lactose, glucose, and galactose and the occurrence of extracellular lactose hydrolysis. Most of activities seemed to be cell associated (Brady et al., 1995) and strain dependent. These studies and the growth of K. marxianus in the present study could explain that the concentrations in carbon substrates of the DCMKm were negligible. The levels of nitrogen fractions in the deacidified cheese media were considerably different between the yeast strains. The increase in ASN, NPN and ammonium fractions in DCM by K. marxianus showed the existence of a proteolytic activity of the K. marxianus used.
Reps (1993) noticed that all yeast strains synthesize proteolytic enzymes. Debaryomyces hansenii gave only very weak proteolytic reactions (Roostita and Fleet, 1996a). No exocellular peptidase activity was detected in these yeast strains isolated from Camembert cheese (Baroiller and Schmidt, 1990; Reps, 1993). Endocellular peptidase activity did not differ greatly among strains; activities at pH 5.5 were equal to or greater than those at pH 4. However, to understand the possible proteolytic contribution of D. hansenii during cheese ripening, Kumura et al. (2002) studied its growth and casein degradation at the cheese-ripening temperature. They showed that: 1) proteolytic activity was found in the intracellular fraction, and the enzyme, which was attached to the cell wall, primarily acted on beta-casein; and 2) more than 90% of the total proteolytic activity, which is responsible for the degradation of both
(s)- and ß-caseins, is within the cell. According to that study, when all easy carbon substrates were consumed, the contribution of yeast to cheese ripening depends on its susceptibility to cell lyses in addition to its proteolytic activity. In our study, the carbon substrates were not totally consumed by D. hansenii. The endopeptidases of D. hansenii and K. marxianus, isolated from cheese smear, hydrolyze casein at pH 5 to 7, optimum 5.8 (Szumski and Cone, 1962; Auberger and Schmidt, 1983) and also degrade polypeptides released from casein by B. linens enzymes (Reps, 1993). But no B. linens strain was added to the cheese medium used. According to Auberger and Schmidt (1983) and Guéguen and Schmidt (1992), the activity of endopeptidases was greater in Kluyveromyces species than in D. hansenii. The proteolytic activity of K. marxianus is important for cheese ripening (Nunez et al., 1981). Strains of Kluyveromyces spp. are generally used to produce some mold-ripened cheeses for which the proteolysis is important. To our knowledge, there is almost no information on the proteolytic activities of Kluyveromyces spp. (Flores et al., 1999), except for some reports on a preliminary identification of some endoproteinases and carboxypeptidases of K. lactis (Grieve et al., 1983).
The influences on the growth and on the biofilm color formation by B. linens can be explained by two facts: 1) when deacidification of the cheese medium was performed by the K. marxianus, B. linens did not have any more lactate to get energy for its growth, so peptides were used to this end. Its growth (2 x 109 cfu x cm2 between d 14 to 28) and its color biofilm (C* = 31 ± 1) were lower than with the other yeast; and (2) when deacidification of the cheese medium was carried out by the D. hansenii, B. linens had lactate as an easily metabolizable substrate, which may explain the important biomass obtained (4 x 109 cfu x cm2 between d 14 to 28). The color of the biofilm depends directly on this biomass. Indeed, according to Dufossé et al. (2001), these differences of biomass could explain the difference in color saturation. These authors have shown that specific production of carotenoid by B. linens may be predicted by the equation: Y = 0.0297 C* value 0.1357. From d 14 to d 28, using this equation, it was 0.79 ± 0.04 for DCMKm and 0.90 ± 0.04 mg of carotenoid/g DM for DCMDh. This result can also explain the observed difference in B. linens biofilm color.
In conclusion, pigmentation by B. linens was directly related to the yeast used to deacidify the cheese medium. Debaryomyces hansenii deacidified more slowly than K. marxianus and K. marxianus seems to be less favorable to color developed by B. linens than D. hansenii. In fact, the carbon and nitrogen substrates in the cheese medium after deacidification highlighted the fact that the media, when they reached pH 5.8, were completely different. The carbon and nitrogen sources available were directly related to the deacidification yeasts used.
| ACKNOWLEDGEMENTS |
|---|
|
|
|---|
Received for publication May 21, 2003. Accepted for publication September 17, 2003.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
J. Mounier, C. Monnet, T. Vallaeys, R. Arditi, A.-S. Sarthou, A. Helias, and F. Irlinger Microbial Interactions within a Cheese Microbial Community Appl. Envir. Microbiol., January 1, 2008; 74(1): 172 - 181. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Bonaiti, F. Irlinger, H. E. Spinnler, and E. Engel An Iterative Sensory Procedure to Select Odor-Active Associations in Complex Consortia of Microorganisms: Application to the Construction of a Cheese Model J Dairy Sci, May 1, 2005; 88(5): 1671 - 1684. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |