- Email: [email protected]
Microbial Succession of Debaryomyces hansenii Strains During the Production of Danish Surfaced-Ripened Cheeses
J. Dairy Sci. 85:478–486 American Dairy Science Association, 2002.
Microbial Succession of Debaryomyces hansenii Strains During the Production of Danish Surfaced-Ripened Cheeses K. M. Petersen,* S. Westall,† and L. Jespersen* *The Royal Veterinary and Agricultural University, Department of Dairy and Food Science, Food Microbiology, Rolighedsvej 30, 1958, Frederiksberg, Copenhagen, Denmark †The Technical University of Denmark, Biocentrum, Building 221, 2800 Lyngby, Denmark.
ABSTRACT
INTRODUCTION
Surface-ripened cheeses of the Danbo type were analyzed for the presence of yeasts with special emphasis on Debaryomyces hansenii. Samples were taken from pasteurized milk, brine, and inoculation slurries and from cheese surfaces during ripening at a Danish dairy. D. hansenii was found to be the dominant yeast species throughout the ripening period, whereas other yeast species such as Trichosporon spp., Rhodotorula spp., and Candida spp. were found in minor concentrations during early stages of cheese ripening. Mitochondrial DNA RFLP was used to show that several strains of D. hansenii were present from the onset of ripening. Thereafter, a microbial succession among the strains took place during the ripening. After 3 d of ripening, only one strain was found. This particular strain was found to be dominant in 16 additional batches of surface-ripened cheeses. We investigated the cause of the observed microbial succession by determining the variation in strains with regard to their ability to grow on lactate and at different pH and NaCl concentrations. The strains were shown to vary in their ability to grow on lactate. In a full factorial design at three levels with factor levels close to the actual levels on the cheese surface, differences in pH and NaCl tolerances were observed. The dominant strain was found to be better adapted than other strains to the environmental conditions existing in surface-ripened cheeses during production [e.g., lactate as the main carbon source, pH 5.5 to 6.0 and NaCl concentrations of 7 to 10% (wt/vol)]. (Key words: surface-ripened cheeses, Debaryomyces hansenii, mtDNA RFLP, microbial succession)
Surface-ripened cheeses are characterized by a surface layer consisting of a complex microflora including both yeasts and bacteria (Reps, 1993). This microflora contributes to the ripening of cheeses by the production of proteolytic and lipolytic enzymes, as well as aroma components (Lenoir, 1984; Fleet, 1990; Jakobsen and Narvhus, 1996). During the first days of ripening, yeasts are dominant and include Debaryomyces hansenii and, depending on the variety of cheese, Yarrowia lipolytica, Kluyveromyces lactis, and Candida zeylanoides (Fleet, 1990; Eliskases-Lechner and Ginzinger, 1995a). After approximately 5 d of ripening, the yeast population decreases, while there is an increase in the number of bacteria (Reps, 1993). The bacterial flora is primarily composed of Brevibacterium linens, Arthrobacter spp., Corynebacterium spp., and Micrococcus spp. (Seiler, 1986; EliskasesLechner and Ginzinger, 1995b; Valde´s-Stauber, 1997). The development of the bacterial surface flora has been shown to be dependent on the metabolism of lactic acid by yeasts, mainly D. hansenii (Leclercq-Perlat et al., 1999). This breakdown of lactate increases pH, which enables the growth of less acid-tolerant coryneform bacteria, in particular B. linens (Seiler and Busse, 1990; Rattay and Fox, 1998). Furthermore, the production of growth factors by yeasts appears to promote the growth and development of the bacterial flora (Valde´sStauber, 1997). Production of surface-ripened cheeses is based, in part, on spontaneous fermentation by adventitious flora. However, commercial strains of D. hansenii are available for use as starter cultures for cheese ripening. Such starter cultures may be added to the brine, sprayed on the cheese surface, or used as an inoculum together with smear from previously produced cheeses. Besides the added starter culture, other strains of D. hansenii originating from milk, brine, or the dairy environment might be present on the cheese surface (Leclerq-Perlat et al., 1999). Since D. hansenii appears to be a highly heterogeneous yeast species as evidenced by the ability to assimilate/ferment different carbon
Abbreviation key: MYGP = malt yeast peptone glucose, RH = relative humidity; SPO = saline peptone solution, YNB = yeast nitrogen base.
Received November 27, 2000. Accepted August 26, 2001. Corresponding author: K. M. Petersen; e-mail: [email protected].
478
479
MICROBIAL SUCCESSION OF YEASTS
compounds (Seiler and Busse, 1990; Nakase et al., 1998, van den Tempel and Jakobsen, 2000) strains of D. hansenii can be also expected to vary in their physiological properties. Microbial successions of yeast species during fermentation of various food products have been reported for years (Wood, 1998). However, little is known about microbial successions at the subspecies level among yeasts during cheese ripening. The development of molecular techniques for strain typing has enabled new types of succession studies, and recent investigations indicate that microbial succession at strain level is a natural phenomenon during fermentation of many food products. Microbial successions of strains belonging to Saccharomyces cerevisiae and Candida krusei, respectively, have been observed during spontaneous fermentation of maize dough (Hayford and Jespersen, 1999; Hayford and Jakobsen, 1999). Similarly, succession of S. cerevisiae strains during wine fermentation has been observed (Sabate et al. 1998). Although of potentially great importance for the dairy industry, microbial successions of D. hansenii strains during cheese ripening has not been investigated by molecular techniques. Mitochondrial DNA RFLP with the restriction enzymes HaeIII and HpaII has previously been proven to have a high discriminative power for typing of D. hansenii strains. It also meets the requirements for a simple method of subspecies typing of D. hansenii (Petersen et al., 2001). The objective of the present study was to determine the yeast flora on Danish semihard surface-ripened cheeses of the Danbo type during the actual events of cheese ripening. For D. hansenii microbial succession, strain dominance and genetic diversity were examined by mtDNA RFLP. Analysis of strain variations in the ability to grow on lactate and at different pH and NaCl concentrations were performed. MATERIALS AND METHODS Sampling The present study was conducted at a Danish dairy producing surface-ripened cheeses of the Danbo type. Samples were taken during cheese production from pasteurized milk, brine, inoculation slurries, and cheese surfaces during the ripening period. The samples were collected twice at 3-mo intervals (sampling periods I and II). In addition, samples were taken from the surface of 16 batches of cheeses after 3 d of ripening. All cheese samples were taken from the side of the cheeses, as described below.
Production of Cheeses Cheeses were made from pasteurized raw milk (74.6°C for 18 s). After the last pressing, the cheeses were salted in saturated brine (22%, wt/vol) at 11 to 13°C for 48 h, then inoculated with a slurry containing smear from the surface of 7-d-old cheeses and starter cultures of D. hansenii and B. linens. The cheeses were reinoculated after 4 d of ripening. After the first inoculation, cheeses were stored at 22°C and 98% relative humidity (RH) for 48 h and then for approximately 20 d at 18°C and 95% RH. Finally, the cheeses were stored for approximately 3 wk at 14°C and 85% RH, after which the smear was washed off and the cheeses were packed in a polyolefin (19 µm) film and further ripened for approximately 29 wk at 7 to 8°C. Determination of pH and NaCl Concentration on the Cheese Surface Surface measurements of pH were carried out on three batches of cheese by use of a surface electrode (InLab 426, Mettler-Toledo, Glostrup, Denmark) and a transportable pH meter (1120, Mettler-Toledo). The NaCl concentration on the cheese surfaces was measured in triplicate on three batches: after brining, after inoculation, after 1 d of ripening, and after 7 d of ripening. The NaCl concentration was measured according to the supplier’s manual with a sodium chloride meter HI931101 (Hanna Instruments, Padova, Italy) fitted with a sodium electrode FC300B (Hanna instruments). A sample from the side of the cheese was sliced (approximately 1 mm thick). From this slice, a sample of 1.0 g was transferred to a plastic tube and mixed with 9.8 ml of deionized water containing 0.2 ml of Ion Strength Adjuster (ISA solution HI7090, Hanna Instruments). The tubes were placed on a water bath at 50°C for 5 h and cooled to room temperature, and the NaCl concentration was measured. Results were reported as grams of NaCl per gram of cheese. Enumeration and Isolation of Yeasts Ten-milliliter samples were taken from the pasteurized milk, brine, and inoculation slurry. The samples were diluted in 90 ml of pH 7.0 saline-peptone solution (SPO). The SPO solution was made by combining 1.0 g of peptone (Difco, Detroit, MI), 8.5 g of NaCl (Merck, Darmstadt, Germany), 0.3 g of Na2HPO4, 2 H2O (Merck) in 1 L of distilled water. Cheese samples were taken from the cheese surface with a sterile tube. The samples (18.5 cm2) were added to 360 ml of SPO and homogenized for 60 s at full speed in a Seward Stomacher (Lab-blender, model 4001, Seward Medical UAC House, London, UK). Serial dilutions of the homogenate of both Journal of Dairy Science Vol. 85, No. 3, 2002
480
PETERSEN ET AL.
milk, brine, inoculation slurries, and cheese samples were made in 9 ml of SPO, and 0.1 ml of dilution was spread onto malt-yeast-peptone-glucose (MYGP) agar. MYGP was composed of 3.0 g of malt extract (Difco), 3.0 g of yeast extract (Difco), 10.0 g of glucose (Merck), 5.0 g of bactopeptone (Difco), and 20.0 g of agar (Difco) per liter of distilled water. After adjusting pH 5.6, 50 µg/ml of both chloramphenicol (Merck) and chlortetracycline (Sigma, St. Louis, MO) were added. Plates were incubated at 25°C for 5 d. Identification of Yeasts Twenty representative yeast isolates were selected from each sample for further characterization. The yeast isolates were streaked onto MYGP agar and incubated at 25°C for 5 d until pure cultures were obtained. Purified isolates were maintained at −80°C in yeast extract peptone glucose broth [5.0 g of yeast extract (Difco), 10.0 g of bactopeptone (Difco), 10.0 g of glucose (Merck)] containing 20% (vol/vol) glycerol. A total of 880 isolates were identified to the species level by microand macromorphological characteristics according to Kurtzman and Fell (1998). Representative isolates were further characterized by use of the API ID 32 C kit (Bio Merie`ux SA, Marcy-l’Etoile, France). Typing of D. hansenii Isolates The succession of D. hansenii strains during cheese ripening was followed by use of mtDNA RFLP. Mitochondrial DNA RFLP profiles were obtained according to Petersen et al. (2001). DNA purified from 5 ml of overnight culture (approximately 109 cells) was resuspended in 50 µl of TE buffer (10 mM Tris-Base (Sigma), 1 mM EDTA (Merck)). Ten microliters of the purified DNA was digested with 5 U of restriction enzymes HaeIII or HpaII (New England BioLabs Inc., Beverly, MA) overnight at 37°C. The restriction fragments were analyzed by electrophoresis through a 1% (wt/vol) NAagarose (Amersham Pharmacia Biotech, Uppsala, Sweden) gel in 1 × TBE buffer at 100 V for 3 h. TBE buffer was composed of 89 mM Tris-base, 89 mM boric acid (Merck), and 2 mM EDTA. A λDNA/HindIII (Promega, Madison, WI) marker was used. The restriction fragments were visualized by ethidium bromide staining and UV transillumination. Variations in the Ability to Grow on Lactate Four isolates were studied for their ability to grow on lactate in yeast nitrogen base (YNB, Difco). Twentyfive milliliters of 1 × YNB broth containing 1.0% (wt/ vol) glucose (Merck) and buffered at pH 5.8 by the addiJournal of Dairy Science Vol. 85, No. 3, 2002
tion of 40 mM Na2HPO4, 2 H2O (Merck) and 0.46 M NaH2PO4, H2O (Merck) was inoculated with a loopful of yeast cells taken from MYGP plates (5 d at 25°C) and incubated for 48 h at 25°C with shaking at 120 rpm. Of this culture 2 × 108 cells were used to inoculate 200 ml of YNB broth containing 1.0% (wt/vol) lactic acid (Merck) at pH 5.2. The cells were grown at 25°C with shaking at 120 rpm. Optical density at 600 nm (UV1201 Spectrophotometer, Shimadzu, Brøndby, Denmark) and pH (pHC2401-8, Radiometer, Copenhagen, Denmark) were recorded until stationary phase was reached. The growth experiments were performed in duplicate. Variations in pH and NaCl Tolerance The ability of two D. hansenii isolates to grow at various pH values and NaCl concentrations was tested in a full factorial design at three levels (32–structure) with factor levels close to the levels found on the cheese surface. The doubling times of D. hansenii isolates grown on the surface of YNB-agar (Difco) with 1% (wt/ vol) lactic acid (Merck) were determined by the use of a Bactometer M123-2 (Bio Me`rieux SA). In total, nine media with different combinations of pH and NaCl concentrations were produced as follows: 1.0 L of basic medium [6.7 g of YNB (Difco), 12 ml of 90% lactic acid (Merck), 2.0 g of KH2PO4 (J. T. Baker, Phillipsburg, NJ), 20 g of agar (Difco) per liter of deionized water] was autoclaved for 15 min at 121°C. The pH of three aliquots of the medium was adjusted to 5.0, 5.5, and 6.0, respectively, by the addition of sterilized 2 M KOH-solution (Merck). Each of the three aliquots was divided into three portions of 100 ml. Sterilized NaCl (J. T. Baker) was added to produce concentrations of 4.0, 7.0, and 10.0% (wt/vol) at each pH value. One Bactometer module (Bio Me`rieux SA) was used for each of the nine media. Each well in the module was filled with 1 ml of medium. To calculate the standard variation, the wells of two additional modules were filled with YNB medium at pH 5.5 and 7.0% (wt/vol) NaCl, i.e., triplicate examinations were performed for this medium. Suspensions of D. hansenii isolates with a concentration of 1.2 × 107 cfu/ml (estimated by microscopy) in 0.1% peptone water [0.1% (wt/wt) peptone (Difco) in deionized water] were made by transferring cells from cultures, which had grown on MYGP plates for 5 d at 25°C. The concentrations of the suspensions were checked by plating on MYGP agar. Serial dilutions were made in 0.1% peptone water and 0.1 ml of the 100, 10−1, 10−2, and 10−3 dilutions were transferred in duplicate to each of the media in the Bactometer wells. The modules were placed in the Bactometer processing
MICROBIAL SUCCESSION OF YEASTS
481
Table 1. Variations in pH at the cheese surface. pH ± SD1
Sample Cheese Cheese Cheese Cheese Cheese Cheese Cheese Cheese
after inoculation 1d 2d 3d 4 d (after reinoculation) 6d 7d 8d
5.4 5.4 5.4 6.2 5.6 6.1 6.7 7.3
± ± ± ± ± ± ± ±
0.04 0.05 0.06 0.34 0.03 0.30 0.06 0.14
1
Standard deviation for three batches of surface ripened cheeses.
units at 22°C and changes in capacitance measured until a rapid decrease in capacitance was recorded (detection time). For each medium, the measured DT was plotted as a function of log cfu in the wells at inoculation. The corresponding slope α was determined by use of linear regression. Doubling time was determined as g = −α log2 (Nielsen, 1991). The response surfaces of the doubling time as a function of pH and NaCl concentrations were analyzed by use of multiple linear regression. Statistical analyses were performed by the use of a modeling and design PC program (MODDE version 4.0, UMETRI AB, Umea˚, Sweden). RESULTS Determination of pH and NaCl Concentration on the Cheese Surface The average pH on the surface of three batches of surface-ripened cheese is shown in Table 1. Immediately after inoculation, pH was 5.4 and remained constant for the first 2 d of ripening. After 3 d of ripening, pH increased to 6.2. After reinoculation, pH decreased to 5.6 and then increased again. After 8 d of ripening, pH was increased to 7.3. The NaCl concentration on the cheese surface was measured for three batches of surface-ripened cheeses. The NaCl concentration on the cheese surface after brining was measured at 8.2 ± 0.3% (wt/wt). After inoculation, the NaCl concentration was measured at 7.5 ± 0.5% (wt/wt), after 1 d of ripening, a considerable decrease in the NaCl concentration was observed [(4.1 ± 0.1% (wt/wt)] and after 7 d of ripening the NaCl concentration was measured at 2.6 ± 0.1 % wt/wt. Enumeration and Identification of Yeasts on the Cheese Surface Samples of pasteurized milk, brine, inoculation slurries, and cheese surfaces were taken at 3-mo intervals and examined for yeast. The concentration of yeasts in
Figure 1. Enumeration of yeasts (cfu × 106/cm2) isolated from the cheese surface during production of surface-ripened cheeses (sampling periods I and II).
the brine was 6.7 × 102 ± 1.3 × 102 cfu/ml for the first sampling period (I) and 2.8 × 102 ± 1.7 × 102 cfu/ml for the second sampling period (II). D. hansenii was found to be the predominant yeast species in brine from both sampling periods. Inoculation slurries used for both inoculation and reinoculation varied in yeast concentration from 1.6 × 105 ± 3.5 × 103 − 4.7 × 105 ± 7.2 × 104 cfu/ml and consisted exclusively of D. hansenii. For both sampling periods no yeasts (<10 cfu/ml) were found in the milk after pasteurization. The yeast counts during cheese production are shown in Figure 1. Quantitative differences in the concentration of yeasts were observed between samples from the two sampling periods. After brining, the yeast counts in samples from sampling periods I and II reached 1.4 × 103 ± 2.8 ×102 and 7.2 Journal of Dairy Science Vol. 85, No. 3, 2002
482
PETERSEN ET AL.
×102 ± 3.6 × 102 cfu/cm2, respectively, and D. hansenii accounted for 44% of the yeast population and 55%, respectively. Other identified yeast species were Trichosporon spp., Rhodotorula spp., and Candida spp. An increase in the concentration of yeasts was observed in both sampling periods as a result of inoculation (1.4 × 106 ± 7.3 × 105 and 6.7 × 104 ± 1.2 × 104 cfu/cm2, respectively) and D. hansenii accounted for 80 (sampling period I) and 90% (sampling period II). After 1 d of ripening, the concentration of yeasts was slightly decreased or constant for both sampling periods (2.3 × 105 ± 1.2 × 105, and 2.7 × 104 ± 2.2 × 104 cfu/cm2, respectively) and D. hansenii accounted for 85 and 100%, respectively. After 3 d of ripening, the yeast counts started to increase. A less pronounced increase was observed in sampling period I, compared with sampling period II (4.8 × 106 ± 1.7 × 106 and 6.9 × 106 ± 1.3 × 106 cfu/ cm2). At that time and throughout the ripening period the yeast flora consisted exclusively of D. hansenii. Just before reinoculation (after 4 d of ripening), the yeast count was increased to 8.1 × 106 ± 1.9 × 106 cfu/cm2 in sampling period I and 1.2 × 107 ± 1.4 × 106 cfu/cm2 in sampling period II. Just after reinoculation (4 d), the concentrations of yeasts were decreased for both sampling periods (5.9 × 106 ± 9.0 × 104 and 7.0 × 106 ± 1.1 × 106 cfu/cm2, respectively). After 5 d, it was shown that reinoculation did not result in a significant increase in sampling period I (6.6 × 106 ± 1.2 × 106 cfu/cm2), whereas in sampling period II the reinoculation did result in a major increase in yeast counts (1.6 × 107 ± 4.4 × 106 cfu/cm2). For both sampling periods, a decrease in yeast counts were observed after 7 d (1.4 × 106 ± 5.7 × 105 cfu/ cm2 and 1.0 × 107 ± 5.5 × 105 cfu/cm2), which continued throughout the remaining ripening period (Figure 1). Mitochondrial DNA Restriction Profiles of D. hansenii Isolates The mtDNA RFLP profiles obtained by digestion with HaeIII for the D. hansenii isolates from sampling period I are shown in Figure 2. In the brine (Figure 2A), four different mtDNA RFLP profiles of D. hansenii were found (denoted A, B, C, and D); two of these appeared to be dominant (A and C). None of these four mtDNA RFLP profiles were found for the isolates on the cheese surface after brining (Figure 2B). Two other contrary mtDNA RFLP profiles were found (denoted E and F); one of these was dominant (E). In the inoculation slurry (Figure 2C), only two isolates with two new mtDNA RFLP profiles were found, one of these (denoted G) was identical to the mtDNA RFLP profile of the added starter culture (result not shown), the other (denoted H) was dominant. Both the mtDNA RFLP profile corresponding to the starter culture (G) and the dominant Journal of Dairy Science Vol. 85, No. 3, 2002
mtDNA RFLP profile (H) from the inoculation slurry were found for the isolates on the cheese surface after inoculation (Figure 2D), together with mtDNA RFLP profile E. After 3 d of ripening (Figure 2E), only isolates with mtDNA RFLP profile H were found on the cheese surface. After 27 d of ripening (Figure 2F) mtDNA RFLP profile H was still the only profile found for the isolates on the cheese surface. Digestion with HpaII instead of HaeIII (results not shown) did not result in mtDNA RFLP profiles, which led to a further division of the D. hansenii isolates. All mtDNA RFLP profiles of D. hansenii isolates obtained for both sampling periods are shown Table 2. In both sampling periods, a microbial succession of D. hansenii strains was observed. Isolates with mtDNA RFLP profile H were found in highest concentration just after inoculation (55 and 90%, respectively). After 1 to 4 d of ripening isolates with mtDNA RFLP profile H were almost exclusively present on the cheese surface. Besides, only minor variations were observed between samples from sampling periods I and II. Isolates with mtDNA RFLP profile C found in the brine used in sampling period I was not found on the surface of cheeses from sampling period I after brining; however, isolates with this mtDNA RFLP profile were found on the cheese surface after brining in sampling period II. In addition, isolates with mtDNA RFLP profile G corresponding to the starter culture was found on the cheese surface after brining in sampling period II; isolates with this mtDNA RFLP profile were not found in either the brine or on the cheese surface after brining in sampling period I. Isolates with mtDNA RFLP profile E were found on cheeses after brining in both sampling periods; however, after inoculation isolates with mtDNA RFLP profile E was only found on cheeses in sampling period I. Enumeration and Identification of Yeasts from 16 Batches of Surface-ripened Cheeses after 3 d of Ripening To see whether the dominance of D. hansenii isolates with mtDNA RFLP profile (H) were a general phenomenon, we examined the yeast flora on the cheese surface of 16 different batches of cheeses after 3 d of ripening. The average yeast counts of the 16 batches were 4.1 × 106 cfu/cm2, with a standard deviation of 1.9 × 106 cfu/ cm2. In total, 320 isolates were tested. D. hansenii was the only yeast species found on cheeses from these 16 batches. Of the 16 batches, 14 batches consisted only of D. hansenii isolates with mtDNA RFLP profile H, identical to the dominant mtDNA RFLP profile from sampling periods I and II. The mtDNA RFLP profiles found on the remaining two batches consisted of 80% isolates with mtDNA RFLP profile H and 20% isolates
483
MICROBIAL SUCCESSION OF YEASTS
Figure 2. Mitochondrial DNA RFLP profiles of Debaryomyces hansenii isolates obtained by digestion with HaeIII. A: Brine. B: Cheese after brining. C: Inoculation slurry. D: Cheese after inoculation. E: Cheese after 3 d of ripening. F: Cheese after 27 d of ripening. M: Marker λDNA/HindIII.
with profile G, identical to the mtDNA RFLP profile of the starter culture.
H and G. The isolate with mtDNA RFLP profile F grew very slowly on lactate. The results were confirmed for two independent experiments.
Variations in the Ability to Grow on Lactate To explain the dominance of isolates with mtDNA RFLP profile H, we investigated the ability of isolates with mtDNA RFLP profile E and F (isolated from the cheese after brining), G (starter culture) and H (dominant isolate) to grow on lactate. The variations in the ability to growth on YNB added 1.0% (wt/vol) lactic acid are shown in Figure 3. A close correlation between change in extracellular pH and yeast cell growth as determined by OD600 was observed for all the investigated isolates (results not shown). The isolate with mtDNA RFLP profile H had the shortest lag phase, while the growth rate appeared to be similar to the growth rates of the isolates with mtDNA RFLP profiles G and E. The isolate with mtDNA RFLP profile E had a longer lag phase than both the isolates with profile
Variations in pH and NaCl Tolerances The response surfaces of the doubling time as a function of pH and NaCl concentration for D. hansenii isolates with mtDNA RFLP profile G (starter culture) and with mtDNA RFLP profile H (dominant profile), analyzed by multiple linear regression, are shown in Figure 4A and B, respectively. The R2 values for the two isolates were 0.92 and 0.81, respectively. Both isolates were able to grow at all investigated combinations of pH (5.0 to 6.0) and NaCl concentrations [4.0 to 10.0% (wt/vol)]. For both isolates, only a minor influence of pH on the doubling times was observed at all examined NaCl concentrations, whereas NaCl concentration had a major influence on the doubling time. However, the influence of NaCl concentration on the doubling time Journal of Dairy Science Vol. 85, No. 3, 2002
484
PETERSEN ET AL. Table 2. Mitochondrial DNA RFLP profiles of Debaryomyces hansenii isolated during production of surface ripened cheeses. Sample Brine
Cheese after brining
Inoculation slurry Cheese after inoculation
Cheese 1 d Cheese 4 d (before re-inoculation) Cheese 4 d (after re-inoculation) Cheese 7 d Cheese 27 d
Sampling period I mtDNA RFLP profiles1 40% 40% 10% 10% 86% 14%
A C B D E F
nd2 50% C 25% E 25% G
70% H 30% G 55% 27% 18% 86% 14%
Sampling Period II mtDNA RFLP profiles1
nd2
H E G H G
90% H 10% G 100% H
100% H
100% H
100% H
90% H 10% G 100% H 100% H
100% H 100% H
1
The letters A to H refer to the mtDNA RFLP profiles in Figure 2. Not determined.
2
was more pronounced for the isolate with profile G than for the isolate with profile H. The lowest observed doubling time (corresponding to the fastest growth rate) for the isolate with profile G was 3.7 ± 1.9 h at pH 5.5 and 4.0% (wt/vol) NaCl. The doubling time of the isolate with profile G increased at pH 5.0 from 4.8 ± 1.9 h at 4.0% NaCl (wt/vol) to 11.3 ± 1.9 h at 10.0% (wt/vol) NaCl. The lowest doubling time 2.9 ± 1.2 h for the isolate with profile H was measured at pH 5.0 and 4.5% (wt/ vol) NaCl. The doubling time of the isolate with profile H at pH 5.0 increased from 3.0 ± 1.2 h at 4.0% (wt/vol) NaCl to 5.9 ± 1.2 h at 10.0% (wt/vol) NaCl. In general, the isolate with profile G grew more slowly than the isolate with profile H at combinations of pH and NaCl concentrations corresponding to the conditions occurring on the cheese surface. At pH 5.5 and 10.0% (wt/vol) NaCl, which is the examined combination of pH and NaCl concentration, that resembles best the conditions on the cheese surface at the time of inoculation, the doubling times of both isolates were 9.6 ± 1.9 and 5.9 ± 1.2 h, respectively. When pH increased to 6.0 and NaCl concentration decreased to 7.0% (wt/ vol), corresponding to the conditions on the cheese surface after a few days of ripening, the doubling times of both isolates were 6.2 ± 1.9 and 3.5 ± 1.2 h, respectively. The sensitivity of the isolate with profile G to high NaCl concentrations was further confirmed by the fact that, at 12.0% (wt/vol) NaCl growth was extremely slow at Journal of Dairy Science Vol. 85, No. 3, 2002
all examined pH values with a lag phase longer than 200 h, while growth of the isolate with profile H was significantly faster (results not shown). DISCUSSION In the present study D. hansenii was found to be the dominant yeast species on the cheese surface throughout the ripening period. This finding is in agreement with results obtained for surface-ripened cheeses of the Tilsiter type (Eliskases-Lechner and Ginzinger, 1995). Also, studies on the surface of blue-veined cheeses demonstrated D. hansenii as the dominant yeast species (Roostita and Fleet, 1996, Van den Tempel and Jakobsen, 1998). However, in another study including different types of surface-ripened cheeses (Limburger, Romadour, Tilsiter, Mu¨nster, Weinka¨se, and Harzer), the dominant yeast flora was found to consist of D. hansenii as well as one or more of the following species; Galactomyces geotrichum, Pichia membranifaciens, and Kluyveromyces marxianus (Valde´s-Stauber et al., 1997). By mtDNA RFLP, a microbial succession of D. hansenii strains was observed during cheese ripening. Isolates with several mtDNA RFLP profiles were present at the beginning of the ripening process. However, after 3 d of ripening and throughout the ripening period only isolates with mtDNA RFLP profile H were present. The dominance of isolates with profile H was further con-
MICROBIAL SUCCESSION OF YEASTS
Figure 3. Variations in pH during growth of Debaryomyces hansenii isolates in YNB added 1% (wt/vol) lactate. Isolate with mtDNA RFLP profile E (isolated from the cheese after brining): (♦). Isolate with mtDNA RFLP profile F (isolated from the cheese after brining): (䊏). Isolate with mtDNA RFLP profile G (starter culture): (▲). Isolate with mtDNA RFLP profile H (dominant isolate): (●).
firmed by analysis of 16 additional batches of surfaceripened cheeses after 3 d of ripening. These isolates originated from the smear of previously produced cheeses. Surprisingly, the mtDNA RFLP profiles of D. hansenii isolates in the brine were not found for isolates on the cheeses either after inoculation or during the ripening. This is not in accordance with the generally accepted
485
belief that yeasts present in the brine are essential for cheese ripening, as suggested by Seiler and Busse (1990). The background for the microbial succession between D. hansenii strains might be due to differences in physiological properties as well as an adaptation to the dairy environment. Investigation of the ability of the isolates to grow on lactate, as the only carbon source, showed that the isolate with mtDNA RFLP profile F (cheese after ripening) grew extremely slowly on lactate, which explains why isolates with this profile were not found to establish on the cheese surface. The longer lag phase observed on lactate for the isolate with mtDNA RFLP profile E (also found on the cheese after brining) compared with the isolates with mtDNA RFLP profile G and H explains why isolates with mtDNA RFLP profile E were outgrown on the cheese surface during ripening. The establishment of isolates with mtDNA RFLP profile H on behalf of isolates with mtDNA RFLP profile G were indicated by minor differences in their ability to utilize lactate. When pH and NaCl concentrations were taken into account, it was further shown that the isolate with mtDNA profile H had a much higher tolerance towards high NaCl concentrations, i.e., 7.0 to 10.0% (wt/vol), which was close to the NaCl concentrations found on the cheese surface. Although D. hansenii is a naturally occurring yeast and a widely used starter culture in both the meat and dairy industry, only few studies on the physiological properties of D. hansenii strains have been carried out. These studies mainly deal with the effect of pH and NaCl
Figure 4. Response surfaces of the doubling time as a function of pH and NaCl concentration analyzed by the use of multiple linear regression. A: Isolate with mtDNA RFLP profile G (starter culture). B: Isolate with mtDNA RFLP profile H (dominant isolate). Journal of Dairy Science Vol. 85, No. 3, 2002
486
PETERSEN ET AL.
on growth (Sørensen and Jakobsen, 1997; Almagro et al., 2000; Van den Tempel and Nielsen, 2000). Accordingly, the effect of pH and NaCl concentration appears to be strain dependent, which is in accordance with the results obtained in the present study. The variation among D. hansenii strains to utilize lactate was studied previously by Eliskases-Lechner and Ginzinger (1995). They observed strain differences in the ability to grow on lactate, which is similar to the observations made in the present study. Assimilation of lactate is an important physiological property of D. hansenii controlling its establishment on cheese surfaces. It is also a significant factor in the development of an appropriate bacterial microflora as well. Further investigations on the ability of D. hansenii strains to grow on lactate are required. Also, uptake and utilization of lactate by D. hansenii should be investigated by molecular techniques as has been done for Saccharomyces cerevisiae (Cassio et al., 1987; Rojo et al., 1998; Casal et al., 1999). The present investigations have shown the importance of studying growth and physiological properties of potential starter cultures at appropriate environmental conditions. Furthermore, such starter cultures should be optimized by propagation under conditions similar to those existing on the cheese surface in an effort to promote adaptation to environmental stress factors such as high NaCl concentration and low pH. The value of mtDNA RFLP for typing of D. hansenii strains has been proved. The technique was found useful for examining microbial successions on cheese surfaces and might be used in the future to control the presence and growth of potential starter cultures. ACKNOWLEDGMENTS This work is part of the FØTEK program supported by the Danish Dairy Research Foundation (Danish Dairy Board) and the Danish Government. The authors are grateful to Tina Schmidt Christensen for excellent technical assistance. The assistance of Søren Lillevang, Arla ˚ rhus, Denmark, is highly apFoods R&D Department, A preciated. REFERENCES Almagro, A., C. Prista, S. Castro, C. Quintas, A. Madeira-Lopes, J. Ramos, and M. C. Loureiro-Dias. 2000. Effects of salts on Debaryomyces hansenii and Saccharomyces cerevisiae under stress conditions. Int. J. Food Microbiol. 56:191–197. Casal, M., S. Paiva, R. P. Andrade, C. Gancedo, and C. Lea˜o. 1999. The lactate-proton symport of Saccharomyces cerevisiae is encoded by JEN1. J. Bacteriol. 181:2620–2623. Cassio, F., C. Lea˜o, and N. van Uden. 1987. Transport of lactate and other short-chain monocarboxylases in the yeast Saccharomyces cerevisiae. Appl. Environ. Microbiol. 53:509–513. Eliskases-Lechner, F., and W. Ginizinger. 1995a. The yeast flora of surface ripened cheeses. Milchwissenschaft 50:458–462. Journal of Dairy Science Vol. 85, No. 3, 2002
Eliskases-Lechner, F., and W. Ginizinger. 1995b. The bacterial flora of surface ripened cheeses with special regard to coryneforms. Lait 75:571–584. Fleet, G. H. 1990. Yeast in dairy products—A Review. J. Appl. Bacteriol. 68:199–211. Hayford, A. E., and M. Jakobsen. 1999. Characterization of Candida krusei strains from spontaneously fermented maize dough by profiles of assimilation, chromosome profile, polymerase chain reaction and restriction endonuclease analysis. J. Appl. Microbiol. 1:29–40. Hayford, A. E., and L. Jespersen. 1999. Characterization of Saccharomyces cerevisiae strains from spontaneously fermented maize dough by profiles of assimilation, chromosome polymorphism, PCR and MAL genotyping. J. Appl. Microbiol. 86:284–294. Jakobsen, M., and J. Narvhus. 1996. Yeast and their possible beneficial and negative effects on the quality of dairy products. Int. Dairy J. 6:755–768. Lenoir, J. 1984. The surface flora and ITS1-5.8S rDNA-ITS2 role in the ripening of cheese. Int. Dairy Fed. Annul. Bull. 171:3–20. Lerlercq-Perlat, M. N., A. Oumer, J. L. Bergere, H. E. Spinnler, and G. Corrieu. 1999. Growth of Debaryomyces hansenii on a bacterial surface-ripened soft cheese. J. Dairy Res. 66:271–281. Kurtzman, C. P., and J. W. Fell, eds. 1998. The Yeast, A Taxonomic Study, 26. 4th ed. Elsevier, Amsterdam, The Netherlands. 1055 pp. Nakase, T., M. Suzuki, H. J. Phaff, and C. P. Kurtzman. 1998. Debaryomyces Lodder & Kreger-van Rij Nom. Cons. Pages 157–163 in The Yeasts; A Taxonomic Study, 26. 4th ed. C. P. Kurtzman, and J. W. Fell, eds. Elsevier, Amsterdam, The Netherlands. Nielsen, P. V. 1991. Preservative and temperature effect on growth of three varieties of the heat-resistant mold, Neosartorya fischeri, as measured by an impedimetric method. J. Food Sci. 56:1735–1740. Petersen, K. M., P. L. Møller, and L. Jespersen. 2001. DNA typing methods for differentiation of Debaryomyces hansenii strains and other yeasts related to surface ripened cheeses. Int. J. Food Microbiol. 69:11–24. Rattay, F. P., and P. F. Fox. 1998. Aspects of Enzymology and biochemical properties of Brevibacterium linens relevant to cheese ripening: a review. J. Dairy Sci. 82:891–909. Reps, A. 1993. Bacterial surface-ripened cheeses. Pages 137–172 in Cheese, Chemistry, Physics and Microbiology. Vol 2. 2nd ed. P. F. Fox, ed. Chapman and Hall, London, UK. Rojo, E. E., B. Guiard, W. Neupert, and R. A. Stuart. 1998. Sorting of D-lactate dehydrogenase to the inner membrane of mitochondria. J. Biol. Chem. 273:8040–8047. Roostita, R., and G. H. Fleet. 1996. The occurrence and growth of yeasts in Camembert and Blue-veined cheeses. Int. J. Food. Microbiol. 28:393–404. Sabate, J., J. Cano, A. Querol, and J. M. Guillamo´n. 1998. Diversity of Saccharomyces strains in wine fermentations: Analysis for two consecutive years. Lett. Appl. Microbiol. 26:452–455. Seiler, H., 1986. Identification of cheese-smear coryneform bacteria. J. Dairy Res. 53:439–449. Seiler, H., and M. Busse. 1990. The yeasts of cheese brines. Int. J. Food Microbiol. 11:289–304. Sørensen, B. B., and M. Jakobsen. 1997. The combined effect of temperature, pH and NaCl on growth of Debaryomyces hansenii analysed by flow cytometry and predictive microbiology. Int. J. Food Microbiol. 34:209–220. Van den Tempel, T., and M. Jakobsen. 1998. Yeasts associated with Danablu. Int. Dairy J. 8:25–31. Van den Tempel, T., and M. Jakobsen. 2000. The technological characteristics of Debaryomyces hansenii and Yarrowia lipolytica and their potential as starter cultures for production of Danablu. Int. Dairy J. 10:263–270. Van den Tempel, T., and M. S. Nielsen. 2000. Effects of atmospheric conditions, NaCl, pH on growth and interactions between moulds and yeasts related to blue cheese production. Int. J. Food Microbiol. 57:193–199. Valde´s-Stauber, N., S. Scherer, and H. Seiler. 1997. Identification of yeasts and coryneform bacteria from the surface microflora of brick cheeses. Int. J. Food Microbiol. 34:115–119. Wood, B. J. B., ed. 1998. Microbiology of Fermented Foods. Vol 1. 2nd ed. Thomason Science, New York, NY.
Microbial Succession of Debaryomyces hansenii Strains During the Production of Danish Surfaced-Ripened Cheeses K. M. Petersen,* S. Westall,† and L. Jespersen* *The Royal Veterinary and Agricultural University, Department of Dairy and Food Science, Food Microbiology, Rolighedsvej 30, 1958, Frederiksberg, Copenhagen, Denmark †The Technical University of Denmark, Biocentrum, Building 221, 2800 Lyngby, Denmark.
ABSTRACT
INTRODUCTION
Surface-ripened cheeses of the Danbo type were analyzed for the presence of yeasts with special emphasis on Debaryomyces hansenii. Samples were taken from pasteurized milk, brine, and inoculation slurries and from cheese surfaces during ripening at a Danish dairy. D. hansenii was found to be the dominant yeast species throughout the ripening period, whereas other yeast species such as Trichosporon spp., Rhodotorula spp., and Candida spp. were found in minor concentrations during early stages of cheese ripening. Mitochondrial DNA RFLP was used to show that several strains of D. hansenii were present from the onset of ripening. Thereafter, a microbial succession among the strains took place during the ripening. After 3 d of ripening, only one strain was found. This particular strain was found to be dominant in 16 additional batches of surface-ripened cheeses. We investigated the cause of the observed microbial succession by determining the variation in strains with regard to their ability to grow on lactate and at different pH and NaCl concentrations. The strains were shown to vary in their ability to grow on lactate. In a full factorial design at three levels with factor levels close to the actual levels on the cheese surface, differences in pH and NaCl tolerances were observed. The dominant strain was found to be better adapted than other strains to the environmental conditions existing in surface-ripened cheeses during production [e.g., lactate as the main carbon source, pH 5.5 to 6.0 and NaCl concentrations of 7 to 10% (wt/vol)]. (Key words: surface-ripened cheeses, Debaryomyces hansenii, mtDNA RFLP, microbial succession)
Surface-ripened cheeses are characterized by a surface layer consisting of a complex microflora including both yeasts and bacteria (Reps, 1993). This microflora contributes to the ripening of cheeses by the production of proteolytic and lipolytic enzymes, as well as aroma components (Lenoir, 1984; Fleet, 1990; Jakobsen and Narvhus, 1996). During the first days of ripening, yeasts are dominant and include Debaryomyces hansenii and, depending on the variety of cheese, Yarrowia lipolytica, Kluyveromyces lactis, and Candida zeylanoides (Fleet, 1990; Eliskases-Lechner and Ginzinger, 1995a). After approximately 5 d of ripening, the yeast population decreases, while there is an increase in the number of bacteria (Reps, 1993). The bacterial flora is primarily composed of Brevibacterium linens, Arthrobacter spp., Corynebacterium spp., and Micrococcus spp. (Seiler, 1986; EliskasesLechner and Ginzinger, 1995b; Valde´s-Stauber, 1997). The development of the bacterial surface flora has been shown to be dependent on the metabolism of lactic acid by yeasts, mainly D. hansenii (Leclercq-Perlat et al., 1999). This breakdown of lactate increases pH, which enables the growth of less acid-tolerant coryneform bacteria, in particular B. linens (Seiler and Busse, 1990; Rattay and Fox, 1998). Furthermore, the production of growth factors by yeasts appears to promote the growth and development of the bacterial flora (Valde´sStauber, 1997). Production of surface-ripened cheeses is based, in part, on spontaneous fermentation by adventitious flora. However, commercial strains of D. hansenii are available for use as starter cultures for cheese ripening. Such starter cultures may be added to the brine, sprayed on the cheese surface, or used as an inoculum together with smear from previously produced cheeses. Besides the added starter culture, other strains of D. hansenii originating from milk, brine, or the dairy environment might be present on the cheese surface (Leclerq-Perlat et al., 1999). Since D. hansenii appears to be a highly heterogeneous yeast species as evidenced by the ability to assimilate/ferment different carbon
Abbreviation key: MYGP = malt yeast peptone glucose, RH = relative humidity; SPO = saline peptone solution, YNB = yeast nitrogen base.
Received November 27, 2000. Accepted August 26, 2001. Corresponding author: K. M. Petersen; e-mail: [email protected].
478
479
MICROBIAL SUCCESSION OF YEASTS
compounds (Seiler and Busse, 1990; Nakase et al., 1998, van den Tempel and Jakobsen, 2000) strains of D. hansenii can be also expected to vary in their physiological properties. Microbial successions of yeast species during fermentation of various food products have been reported for years (Wood, 1998). However, little is known about microbial successions at the subspecies level among yeasts during cheese ripening. The development of molecular techniques for strain typing has enabled new types of succession studies, and recent investigations indicate that microbial succession at strain level is a natural phenomenon during fermentation of many food products. Microbial successions of strains belonging to Saccharomyces cerevisiae and Candida krusei, respectively, have been observed during spontaneous fermentation of maize dough (Hayford and Jespersen, 1999; Hayford and Jakobsen, 1999). Similarly, succession of S. cerevisiae strains during wine fermentation has been observed (Sabate et al. 1998). Although of potentially great importance for the dairy industry, microbial successions of D. hansenii strains during cheese ripening has not been investigated by molecular techniques. Mitochondrial DNA RFLP with the restriction enzymes HaeIII and HpaII has previously been proven to have a high discriminative power for typing of D. hansenii strains. It also meets the requirements for a simple method of subspecies typing of D. hansenii (Petersen et al., 2001). The objective of the present study was to determine the yeast flora on Danish semihard surface-ripened cheeses of the Danbo type during the actual events of cheese ripening. For D. hansenii microbial succession, strain dominance and genetic diversity were examined by mtDNA RFLP. Analysis of strain variations in the ability to grow on lactate and at different pH and NaCl concentrations were performed. MATERIALS AND METHODS Sampling The present study was conducted at a Danish dairy producing surface-ripened cheeses of the Danbo type. Samples were taken during cheese production from pasteurized milk, brine, inoculation slurries, and cheese surfaces during the ripening period. The samples were collected twice at 3-mo intervals (sampling periods I and II). In addition, samples were taken from the surface of 16 batches of cheeses after 3 d of ripening. All cheese samples were taken from the side of the cheeses, as described below.
Production of Cheeses Cheeses were made from pasteurized raw milk (74.6°C for 18 s). After the last pressing, the cheeses were salted in saturated brine (22%, wt/vol) at 11 to 13°C for 48 h, then inoculated with a slurry containing smear from the surface of 7-d-old cheeses and starter cultures of D. hansenii and B. linens. The cheeses were reinoculated after 4 d of ripening. After the first inoculation, cheeses were stored at 22°C and 98% relative humidity (RH) for 48 h and then for approximately 20 d at 18°C and 95% RH. Finally, the cheeses were stored for approximately 3 wk at 14°C and 85% RH, after which the smear was washed off and the cheeses were packed in a polyolefin (19 µm) film and further ripened for approximately 29 wk at 7 to 8°C. Determination of pH and NaCl Concentration on the Cheese Surface Surface measurements of pH were carried out on three batches of cheese by use of a surface electrode (InLab 426, Mettler-Toledo, Glostrup, Denmark) and a transportable pH meter (1120, Mettler-Toledo). The NaCl concentration on the cheese surfaces was measured in triplicate on three batches: after brining, after inoculation, after 1 d of ripening, and after 7 d of ripening. The NaCl concentration was measured according to the supplier’s manual with a sodium chloride meter HI931101 (Hanna Instruments, Padova, Italy) fitted with a sodium electrode FC300B (Hanna instruments). A sample from the side of the cheese was sliced (approximately 1 mm thick). From this slice, a sample of 1.0 g was transferred to a plastic tube and mixed with 9.8 ml of deionized water containing 0.2 ml of Ion Strength Adjuster (ISA solution HI7090, Hanna Instruments). The tubes were placed on a water bath at 50°C for 5 h and cooled to room temperature, and the NaCl concentration was measured. Results were reported as grams of NaCl per gram of cheese. Enumeration and Isolation of Yeasts Ten-milliliter samples were taken from the pasteurized milk, brine, and inoculation slurry. The samples were diluted in 90 ml of pH 7.0 saline-peptone solution (SPO). The SPO solution was made by combining 1.0 g of peptone (Difco, Detroit, MI), 8.5 g of NaCl (Merck, Darmstadt, Germany), 0.3 g of Na2HPO4, 2 H2O (Merck) in 1 L of distilled water. Cheese samples were taken from the cheese surface with a sterile tube. The samples (18.5 cm2) were added to 360 ml of SPO and homogenized for 60 s at full speed in a Seward Stomacher (Lab-blender, model 4001, Seward Medical UAC House, London, UK). Serial dilutions of the homogenate of both Journal of Dairy Science Vol. 85, No. 3, 2002
480
PETERSEN ET AL.
milk, brine, inoculation slurries, and cheese samples were made in 9 ml of SPO, and 0.1 ml of dilution was spread onto malt-yeast-peptone-glucose (MYGP) agar. MYGP was composed of 3.0 g of malt extract (Difco), 3.0 g of yeast extract (Difco), 10.0 g of glucose (Merck), 5.0 g of bactopeptone (Difco), and 20.0 g of agar (Difco) per liter of distilled water. After adjusting pH 5.6, 50 µg/ml of both chloramphenicol (Merck) and chlortetracycline (Sigma, St. Louis, MO) were added. Plates were incubated at 25°C for 5 d. Identification of Yeasts Twenty representative yeast isolates were selected from each sample for further characterization. The yeast isolates were streaked onto MYGP agar and incubated at 25°C for 5 d until pure cultures were obtained. Purified isolates were maintained at −80°C in yeast extract peptone glucose broth [5.0 g of yeast extract (Difco), 10.0 g of bactopeptone (Difco), 10.0 g of glucose (Merck)] containing 20% (vol/vol) glycerol. A total of 880 isolates were identified to the species level by microand macromorphological characteristics according to Kurtzman and Fell (1998). Representative isolates were further characterized by use of the API ID 32 C kit (Bio Merie`ux SA, Marcy-l’Etoile, France). Typing of D. hansenii Isolates The succession of D. hansenii strains during cheese ripening was followed by use of mtDNA RFLP. Mitochondrial DNA RFLP profiles were obtained according to Petersen et al. (2001). DNA purified from 5 ml of overnight culture (approximately 109 cells) was resuspended in 50 µl of TE buffer (10 mM Tris-Base (Sigma), 1 mM EDTA (Merck)). Ten microliters of the purified DNA was digested with 5 U of restriction enzymes HaeIII or HpaII (New England BioLabs Inc., Beverly, MA) overnight at 37°C. The restriction fragments were analyzed by electrophoresis through a 1% (wt/vol) NAagarose (Amersham Pharmacia Biotech, Uppsala, Sweden) gel in 1 × TBE buffer at 100 V for 3 h. TBE buffer was composed of 89 mM Tris-base, 89 mM boric acid (Merck), and 2 mM EDTA. A λDNA/HindIII (Promega, Madison, WI) marker was used. The restriction fragments were visualized by ethidium bromide staining and UV transillumination. Variations in the Ability to Grow on Lactate Four isolates were studied for their ability to grow on lactate in yeast nitrogen base (YNB, Difco). Twentyfive milliliters of 1 × YNB broth containing 1.0% (wt/ vol) glucose (Merck) and buffered at pH 5.8 by the addiJournal of Dairy Science Vol. 85, No. 3, 2002
tion of 40 mM Na2HPO4, 2 H2O (Merck) and 0.46 M NaH2PO4, H2O (Merck) was inoculated with a loopful of yeast cells taken from MYGP plates (5 d at 25°C) and incubated for 48 h at 25°C with shaking at 120 rpm. Of this culture 2 × 108 cells were used to inoculate 200 ml of YNB broth containing 1.0% (wt/vol) lactic acid (Merck) at pH 5.2. The cells were grown at 25°C with shaking at 120 rpm. Optical density at 600 nm (UV1201 Spectrophotometer, Shimadzu, Brøndby, Denmark) and pH (pHC2401-8, Radiometer, Copenhagen, Denmark) were recorded until stationary phase was reached. The growth experiments were performed in duplicate. Variations in pH and NaCl Tolerance The ability of two D. hansenii isolates to grow at various pH values and NaCl concentrations was tested in a full factorial design at three levels (32–structure) with factor levels close to the levels found on the cheese surface. The doubling times of D. hansenii isolates grown on the surface of YNB-agar (Difco) with 1% (wt/ vol) lactic acid (Merck) were determined by the use of a Bactometer M123-2 (Bio Me`rieux SA). In total, nine media with different combinations of pH and NaCl concentrations were produced as follows: 1.0 L of basic medium [6.7 g of YNB (Difco), 12 ml of 90% lactic acid (Merck), 2.0 g of KH2PO4 (J. T. Baker, Phillipsburg, NJ), 20 g of agar (Difco) per liter of deionized water] was autoclaved for 15 min at 121°C. The pH of three aliquots of the medium was adjusted to 5.0, 5.5, and 6.0, respectively, by the addition of sterilized 2 M KOH-solution (Merck). Each of the three aliquots was divided into three portions of 100 ml. Sterilized NaCl (J. T. Baker) was added to produce concentrations of 4.0, 7.0, and 10.0% (wt/vol) at each pH value. One Bactometer module (Bio Me`rieux SA) was used for each of the nine media. Each well in the module was filled with 1 ml of medium. To calculate the standard variation, the wells of two additional modules were filled with YNB medium at pH 5.5 and 7.0% (wt/vol) NaCl, i.e., triplicate examinations were performed for this medium. Suspensions of D. hansenii isolates with a concentration of 1.2 × 107 cfu/ml (estimated by microscopy) in 0.1% peptone water [0.1% (wt/wt) peptone (Difco) in deionized water] were made by transferring cells from cultures, which had grown on MYGP plates for 5 d at 25°C. The concentrations of the suspensions were checked by plating on MYGP agar. Serial dilutions were made in 0.1% peptone water and 0.1 ml of the 100, 10−1, 10−2, and 10−3 dilutions were transferred in duplicate to each of the media in the Bactometer wells. The modules were placed in the Bactometer processing
MICROBIAL SUCCESSION OF YEASTS
481
Table 1. Variations in pH at the cheese surface. pH ± SD1
Sample Cheese Cheese Cheese Cheese Cheese Cheese Cheese Cheese
after inoculation 1d 2d 3d 4 d (after reinoculation) 6d 7d 8d
5.4 5.4 5.4 6.2 5.6 6.1 6.7 7.3
± ± ± ± ± ± ± ±
0.04 0.05 0.06 0.34 0.03 0.30 0.06 0.14
1
Standard deviation for three batches of surface ripened cheeses.
units at 22°C and changes in capacitance measured until a rapid decrease in capacitance was recorded (detection time). For each medium, the measured DT was plotted as a function of log cfu in the wells at inoculation. The corresponding slope α was determined by use of linear regression. Doubling time was determined as g = −α log2 (Nielsen, 1991). The response surfaces of the doubling time as a function of pH and NaCl concentrations were analyzed by use of multiple linear regression. Statistical analyses were performed by the use of a modeling and design PC program (MODDE version 4.0, UMETRI AB, Umea˚, Sweden). RESULTS Determination of pH and NaCl Concentration on the Cheese Surface The average pH on the surface of three batches of surface-ripened cheese is shown in Table 1. Immediately after inoculation, pH was 5.4 and remained constant for the first 2 d of ripening. After 3 d of ripening, pH increased to 6.2. After reinoculation, pH decreased to 5.6 and then increased again. After 8 d of ripening, pH was increased to 7.3. The NaCl concentration on the cheese surface was measured for three batches of surface-ripened cheeses. The NaCl concentration on the cheese surface after brining was measured at 8.2 ± 0.3% (wt/wt). After inoculation, the NaCl concentration was measured at 7.5 ± 0.5% (wt/wt), after 1 d of ripening, a considerable decrease in the NaCl concentration was observed [(4.1 ± 0.1% (wt/wt)] and after 7 d of ripening the NaCl concentration was measured at 2.6 ± 0.1 % wt/wt. Enumeration and Identification of Yeasts on the Cheese Surface Samples of pasteurized milk, brine, inoculation slurries, and cheese surfaces were taken at 3-mo intervals and examined for yeast. The concentration of yeasts in
Figure 1. Enumeration of yeasts (cfu × 106/cm2) isolated from the cheese surface during production of surface-ripened cheeses (sampling periods I and II).
the brine was 6.7 × 102 ± 1.3 × 102 cfu/ml for the first sampling period (I) and 2.8 × 102 ± 1.7 × 102 cfu/ml for the second sampling period (II). D. hansenii was found to be the predominant yeast species in brine from both sampling periods. Inoculation slurries used for both inoculation and reinoculation varied in yeast concentration from 1.6 × 105 ± 3.5 × 103 − 4.7 × 105 ± 7.2 × 104 cfu/ml and consisted exclusively of D. hansenii. For both sampling periods no yeasts (<10 cfu/ml) were found in the milk after pasteurization. The yeast counts during cheese production are shown in Figure 1. Quantitative differences in the concentration of yeasts were observed between samples from the two sampling periods. After brining, the yeast counts in samples from sampling periods I and II reached 1.4 × 103 ± 2.8 ×102 and 7.2 Journal of Dairy Science Vol. 85, No. 3, 2002
482
PETERSEN ET AL.
×102 ± 3.6 × 102 cfu/cm2, respectively, and D. hansenii accounted for 44% of the yeast population and 55%, respectively. Other identified yeast species were Trichosporon spp., Rhodotorula spp., and Candida spp. An increase in the concentration of yeasts was observed in both sampling periods as a result of inoculation (1.4 × 106 ± 7.3 × 105 and 6.7 × 104 ± 1.2 × 104 cfu/cm2, respectively) and D. hansenii accounted for 80 (sampling period I) and 90% (sampling period II). After 1 d of ripening, the concentration of yeasts was slightly decreased or constant for both sampling periods (2.3 × 105 ± 1.2 × 105, and 2.7 × 104 ± 2.2 × 104 cfu/cm2, respectively) and D. hansenii accounted for 85 and 100%, respectively. After 3 d of ripening, the yeast counts started to increase. A less pronounced increase was observed in sampling period I, compared with sampling period II (4.8 × 106 ± 1.7 × 106 and 6.9 × 106 ± 1.3 × 106 cfu/ cm2). At that time and throughout the ripening period the yeast flora consisted exclusively of D. hansenii. Just before reinoculation (after 4 d of ripening), the yeast count was increased to 8.1 × 106 ± 1.9 × 106 cfu/cm2 in sampling period I and 1.2 × 107 ± 1.4 × 106 cfu/cm2 in sampling period II. Just after reinoculation (4 d), the concentrations of yeasts were decreased for both sampling periods (5.9 × 106 ± 9.0 × 104 and 7.0 × 106 ± 1.1 × 106 cfu/cm2, respectively). After 5 d, it was shown that reinoculation did not result in a significant increase in sampling period I (6.6 × 106 ± 1.2 × 106 cfu/cm2), whereas in sampling period II the reinoculation did result in a major increase in yeast counts (1.6 × 107 ± 4.4 × 106 cfu/cm2). For both sampling periods, a decrease in yeast counts were observed after 7 d (1.4 × 106 ± 5.7 × 105 cfu/ cm2 and 1.0 × 107 ± 5.5 × 105 cfu/cm2), which continued throughout the remaining ripening period (Figure 1). Mitochondrial DNA Restriction Profiles of D. hansenii Isolates The mtDNA RFLP profiles obtained by digestion with HaeIII for the D. hansenii isolates from sampling period I are shown in Figure 2. In the brine (Figure 2A), four different mtDNA RFLP profiles of D. hansenii were found (denoted A, B, C, and D); two of these appeared to be dominant (A and C). None of these four mtDNA RFLP profiles were found for the isolates on the cheese surface after brining (Figure 2B). Two other contrary mtDNA RFLP profiles were found (denoted E and F); one of these was dominant (E). In the inoculation slurry (Figure 2C), only two isolates with two new mtDNA RFLP profiles were found, one of these (denoted G) was identical to the mtDNA RFLP profile of the added starter culture (result not shown), the other (denoted H) was dominant. Both the mtDNA RFLP profile corresponding to the starter culture (G) and the dominant Journal of Dairy Science Vol. 85, No. 3, 2002
mtDNA RFLP profile (H) from the inoculation slurry were found for the isolates on the cheese surface after inoculation (Figure 2D), together with mtDNA RFLP profile E. After 3 d of ripening (Figure 2E), only isolates with mtDNA RFLP profile H were found on the cheese surface. After 27 d of ripening (Figure 2F) mtDNA RFLP profile H was still the only profile found for the isolates on the cheese surface. Digestion with HpaII instead of HaeIII (results not shown) did not result in mtDNA RFLP profiles, which led to a further division of the D. hansenii isolates. All mtDNA RFLP profiles of D. hansenii isolates obtained for both sampling periods are shown Table 2. In both sampling periods, a microbial succession of D. hansenii strains was observed. Isolates with mtDNA RFLP profile H were found in highest concentration just after inoculation (55 and 90%, respectively). After 1 to 4 d of ripening isolates with mtDNA RFLP profile H were almost exclusively present on the cheese surface. Besides, only minor variations were observed between samples from sampling periods I and II. Isolates with mtDNA RFLP profile C found in the brine used in sampling period I was not found on the surface of cheeses from sampling period I after brining; however, isolates with this mtDNA RFLP profile were found on the cheese surface after brining in sampling period II. In addition, isolates with mtDNA RFLP profile G corresponding to the starter culture was found on the cheese surface after brining in sampling period II; isolates with this mtDNA RFLP profile were not found in either the brine or on the cheese surface after brining in sampling period I. Isolates with mtDNA RFLP profile E were found on cheeses after brining in both sampling periods; however, after inoculation isolates with mtDNA RFLP profile E was only found on cheeses in sampling period I. Enumeration and Identification of Yeasts from 16 Batches of Surface-ripened Cheeses after 3 d of Ripening To see whether the dominance of D. hansenii isolates with mtDNA RFLP profile (H) were a general phenomenon, we examined the yeast flora on the cheese surface of 16 different batches of cheeses after 3 d of ripening. The average yeast counts of the 16 batches were 4.1 × 106 cfu/cm2, with a standard deviation of 1.9 × 106 cfu/ cm2. In total, 320 isolates were tested. D. hansenii was the only yeast species found on cheeses from these 16 batches. Of the 16 batches, 14 batches consisted only of D. hansenii isolates with mtDNA RFLP profile H, identical to the dominant mtDNA RFLP profile from sampling periods I and II. The mtDNA RFLP profiles found on the remaining two batches consisted of 80% isolates with mtDNA RFLP profile H and 20% isolates
483
MICROBIAL SUCCESSION OF YEASTS
Figure 2. Mitochondrial DNA RFLP profiles of Debaryomyces hansenii isolates obtained by digestion with HaeIII. A: Brine. B: Cheese after brining. C: Inoculation slurry. D: Cheese after inoculation. E: Cheese after 3 d of ripening. F: Cheese after 27 d of ripening. M: Marker λDNA/HindIII.
with profile G, identical to the mtDNA RFLP profile of the starter culture.
H and G. The isolate with mtDNA RFLP profile F grew very slowly on lactate. The results were confirmed for two independent experiments.
Variations in the Ability to Grow on Lactate To explain the dominance of isolates with mtDNA RFLP profile H, we investigated the ability of isolates with mtDNA RFLP profile E and F (isolated from the cheese after brining), G (starter culture) and H (dominant isolate) to grow on lactate. The variations in the ability to growth on YNB added 1.0% (wt/vol) lactic acid are shown in Figure 3. A close correlation between change in extracellular pH and yeast cell growth as determined by OD600 was observed for all the investigated isolates (results not shown). The isolate with mtDNA RFLP profile H had the shortest lag phase, while the growth rate appeared to be similar to the growth rates of the isolates with mtDNA RFLP profiles G and E. The isolate with mtDNA RFLP profile E had a longer lag phase than both the isolates with profile
Variations in pH and NaCl Tolerances The response surfaces of the doubling time as a function of pH and NaCl concentration for D. hansenii isolates with mtDNA RFLP profile G (starter culture) and with mtDNA RFLP profile H (dominant profile), analyzed by multiple linear regression, are shown in Figure 4A and B, respectively. The R2 values for the two isolates were 0.92 and 0.81, respectively. Both isolates were able to grow at all investigated combinations of pH (5.0 to 6.0) and NaCl concentrations [4.0 to 10.0% (wt/vol)]. For both isolates, only a minor influence of pH on the doubling times was observed at all examined NaCl concentrations, whereas NaCl concentration had a major influence on the doubling time. However, the influence of NaCl concentration on the doubling time Journal of Dairy Science Vol. 85, No. 3, 2002
484
PETERSEN ET AL. Table 2. Mitochondrial DNA RFLP profiles of Debaryomyces hansenii isolated during production of surface ripened cheeses. Sample Brine
Cheese after brining
Inoculation slurry Cheese after inoculation
Cheese 1 d Cheese 4 d (before re-inoculation) Cheese 4 d (after re-inoculation) Cheese 7 d Cheese 27 d
Sampling period I mtDNA RFLP profiles1 40% 40% 10% 10% 86% 14%
A C B D E F
nd2 50% C 25% E 25% G
70% H 30% G 55% 27% 18% 86% 14%
Sampling Period II mtDNA RFLP profiles1
nd2
H E G H G
90% H 10% G 100% H
100% H
100% H
100% H
90% H 10% G 100% H 100% H
100% H 100% H
1
The letters A to H refer to the mtDNA RFLP profiles in Figure 2. Not determined.
2
was more pronounced for the isolate with profile G than for the isolate with profile H. The lowest observed doubling time (corresponding to the fastest growth rate) for the isolate with profile G was 3.7 ± 1.9 h at pH 5.5 and 4.0% (wt/vol) NaCl. The doubling time of the isolate with profile G increased at pH 5.0 from 4.8 ± 1.9 h at 4.0% NaCl (wt/vol) to 11.3 ± 1.9 h at 10.0% (wt/vol) NaCl. The lowest doubling time 2.9 ± 1.2 h for the isolate with profile H was measured at pH 5.0 and 4.5% (wt/ vol) NaCl. The doubling time of the isolate with profile H at pH 5.0 increased from 3.0 ± 1.2 h at 4.0% (wt/vol) NaCl to 5.9 ± 1.2 h at 10.0% (wt/vol) NaCl. In general, the isolate with profile G grew more slowly than the isolate with profile H at combinations of pH and NaCl concentrations corresponding to the conditions occurring on the cheese surface. At pH 5.5 and 10.0% (wt/vol) NaCl, which is the examined combination of pH and NaCl concentration, that resembles best the conditions on the cheese surface at the time of inoculation, the doubling times of both isolates were 9.6 ± 1.9 and 5.9 ± 1.2 h, respectively. When pH increased to 6.0 and NaCl concentration decreased to 7.0% (wt/ vol), corresponding to the conditions on the cheese surface after a few days of ripening, the doubling times of both isolates were 6.2 ± 1.9 and 3.5 ± 1.2 h, respectively. The sensitivity of the isolate with profile G to high NaCl concentrations was further confirmed by the fact that, at 12.0% (wt/vol) NaCl growth was extremely slow at Journal of Dairy Science Vol. 85, No. 3, 2002
all examined pH values with a lag phase longer than 200 h, while growth of the isolate with profile H was significantly faster (results not shown). DISCUSSION In the present study D. hansenii was found to be the dominant yeast species on the cheese surface throughout the ripening period. This finding is in agreement with results obtained for surface-ripened cheeses of the Tilsiter type (Eliskases-Lechner and Ginzinger, 1995). Also, studies on the surface of blue-veined cheeses demonstrated D. hansenii as the dominant yeast species (Roostita and Fleet, 1996, Van den Tempel and Jakobsen, 1998). However, in another study including different types of surface-ripened cheeses (Limburger, Romadour, Tilsiter, Mu¨nster, Weinka¨se, and Harzer), the dominant yeast flora was found to consist of D. hansenii as well as one or more of the following species; Galactomyces geotrichum, Pichia membranifaciens, and Kluyveromyces marxianus (Valde´s-Stauber et al., 1997). By mtDNA RFLP, a microbial succession of D. hansenii strains was observed during cheese ripening. Isolates with several mtDNA RFLP profiles were present at the beginning of the ripening process. However, after 3 d of ripening and throughout the ripening period only isolates with mtDNA RFLP profile H were present. The dominance of isolates with profile H was further con-
MICROBIAL SUCCESSION OF YEASTS
Figure 3. Variations in pH during growth of Debaryomyces hansenii isolates in YNB added 1% (wt/vol) lactate. Isolate with mtDNA RFLP profile E (isolated from the cheese after brining): (♦). Isolate with mtDNA RFLP profile F (isolated from the cheese after brining): (䊏). Isolate with mtDNA RFLP profile G (starter culture): (▲). Isolate with mtDNA RFLP profile H (dominant isolate): (●).
firmed by analysis of 16 additional batches of surfaceripened cheeses after 3 d of ripening. These isolates originated from the smear of previously produced cheeses. Surprisingly, the mtDNA RFLP profiles of D. hansenii isolates in the brine were not found for isolates on the cheeses either after inoculation or during the ripening. This is not in accordance with the generally accepted
485
belief that yeasts present in the brine are essential for cheese ripening, as suggested by Seiler and Busse (1990). The background for the microbial succession between D. hansenii strains might be due to differences in physiological properties as well as an adaptation to the dairy environment. Investigation of the ability of the isolates to grow on lactate, as the only carbon source, showed that the isolate with mtDNA RFLP profile F (cheese after ripening) grew extremely slowly on lactate, which explains why isolates with this profile were not found to establish on the cheese surface. The longer lag phase observed on lactate for the isolate with mtDNA RFLP profile E (also found on the cheese after brining) compared with the isolates with mtDNA RFLP profile G and H explains why isolates with mtDNA RFLP profile E were outgrown on the cheese surface during ripening. The establishment of isolates with mtDNA RFLP profile H on behalf of isolates with mtDNA RFLP profile G were indicated by minor differences in their ability to utilize lactate. When pH and NaCl concentrations were taken into account, it was further shown that the isolate with mtDNA profile H had a much higher tolerance towards high NaCl concentrations, i.e., 7.0 to 10.0% (wt/vol), which was close to the NaCl concentrations found on the cheese surface. Although D. hansenii is a naturally occurring yeast and a widely used starter culture in both the meat and dairy industry, only few studies on the physiological properties of D. hansenii strains have been carried out. These studies mainly deal with the effect of pH and NaCl
Figure 4. Response surfaces of the doubling time as a function of pH and NaCl concentration analyzed by the use of multiple linear regression. A: Isolate with mtDNA RFLP profile G (starter culture). B: Isolate with mtDNA RFLP profile H (dominant isolate). Journal of Dairy Science Vol. 85, No. 3, 2002
486
PETERSEN ET AL.
on growth (Sørensen and Jakobsen, 1997; Almagro et al., 2000; Van den Tempel and Nielsen, 2000). Accordingly, the effect of pH and NaCl concentration appears to be strain dependent, which is in accordance with the results obtained in the present study. The variation among D. hansenii strains to utilize lactate was studied previously by Eliskases-Lechner and Ginzinger (1995). They observed strain differences in the ability to grow on lactate, which is similar to the observations made in the present study. Assimilation of lactate is an important physiological property of D. hansenii controlling its establishment on cheese surfaces. It is also a significant factor in the development of an appropriate bacterial microflora as well. Further investigations on the ability of D. hansenii strains to grow on lactate are required. Also, uptake and utilization of lactate by D. hansenii should be investigated by molecular techniques as has been done for Saccharomyces cerevisiae (Cassio et al., 1987; Rojo et al., 1998; Casal et al., 1999). The present investigations have shown the importance of studying growth and physiological properties of potential starter cultures at appropriate environmental conditions. Furthermore, such starter cultures should be optimized by propagation under conditions similar to those existing on the cheese surface in an effort to promote adaptation to environmental stress factors such as high NaCl concentration and low pH. The value of mtDNA RFLP for typing of D. hansenii strains has been proved. The technique was found useful for examining microbial successions on cheese surfaces and might be used in the future to control the presence and growth of potential starter cultures. ACKNOWLEDGMENTS This work is part of the FØTEK program supported by the Danish Dairy Research Foundation (Danish Dairy Board) and the Danish Government. The authors are grateful to Tina Schmidt Christensen for excellent technical assistance. The assistance of Søren Lillevang, Arla ˚ rhus, Denmark, is highly apFoods R&D Department, A preciated. REFERENCES Almagro, A., C. Prista, S. Castro, C. Quintas, A. Madeira-Lopes, J. Ramos, and M. C. Loureiro-Dias. 2000. Effects of salts on Debaryomyces hansenii and Saccharomyces cerevisiae under stress conditions. Int. J. Food Microbiol. 56:191–197. Casal, M., S. Paiva, R. P. Andrade, C. Gancedo, and C. Lea˜o. 1999. The lactate-proton symport of Saccharomyces cerevisiae is encoded by JEN1. J. Bacteriol. 181:2620–2623. Cassio, F., C. Lea˜o, and N. van Uden. 1987. Transport of lactate and other short-chain monocarboxylases in the yeast Saccharomyces cerevisiae. Appl. Environ. Microbiol. 53:509–513. Eliskases-Lechner, F., and W. Ginizinger. 1995a. The yeast flora of surface ripened cheeses. Milchwissenschaft 50:458–462. Journal of Dairy Science Vol. 85, No. 3, 2002
Eliskases-Lechner, F., and W. Ginizinger. 1995b. The bacterial flora of surface ripened cheeses with special regard to coryneforms. Lait 75:571–584. Fleet, G. H. 1990. Yeast in dairy products—A Review. J. Appl. Bacteriol. 68:199–211. Hayford, A. E., and M. Jakobsen. 1999. Characterization of Candida krusei strains from spontaneously fermented maize dough by profiles of assimilation, chromosome profile, polymerase chain reaction and restriction endonuclease analysis. J. Appl. Microbiol. 1:29–40. Hayford, A. E., and L. Jespersen. 1999. Characterization of Saccharomyces cerevisiae strains from spontaneously fermented maize dough by profiles of assimilation, chromosome polymorphism, PCR and MAL genotyping. J. Appl. Microbiol. 86:284–294. Jakobsen, M., and J. Narvhus. 1996. Yeast and their possible beneficial and negative effects on the quality of dairy products. Int. Dairy J. 6:755–768. Lenoir, J. 1984. The surface flora and ITS1-5.8S rDNA-ITS2 role in the ripening of cheese. Int. Dairy Fed. Annul. Bull. 171:3–20. Lerlercq-Perlat, M. N., A. Oumer, J. L. Bergere, H. E. Spinnler, and G. Corrieu. 1999. Growth of Debaryomyces hansenii on a bacterial surface-ripened soft cheese. J. Dairy Res. 66:271–281. Kurtzman, C. P., and J. W. Fell, eds. 1998. The Yeast, A Taxonomic Study, 26. 4th ed. Elsevier, Amsterdam, The Netherlands. 1055 pp. Nakase, T., M. Suzuki, H. J. Phaff, and C. P. Kurtzman. 1998. Debaryomyces Lodder & Kreger-van Rij Nom. Cons. Pages 157–163 in The Yeasts; A Taxonomic Study, 26. 4th ed. C. P. Kurtzman, and J. W. Fell, eds. Elsevier, Amsterdam, The Netherlands. Nielsen, P. V. 1991. Preservative and temperature effect on growth of three varieties of the heat-resistant mold, Neosartorya fischeri, as measured by an impedimetric method. J. Food Sci. 56:1735–1740. Petersen, K. M., P. L. Møller, and L. Jespersen. 2001. DNA typing methods for differentiation of Debaryomyces hansenii strains and other yeasts related to surface ripened cheeses. Int. J. Food Microbiol. 69:11–24. Rattay, F. P., and P. F. Fox. 1998. Aspects of Enzymology and biochemical properties of Brevibacterium linens relevant to cheese ripening: a review. J. Dairy Sci. 82:891–909. Reps, A. 1993. Bacterial surface-ripened cheeses. Pages 137–172 in Cheese, Chemistry, Physics and Microbiology. Vol 2. 2nd ed. P. F. Fox, ed. Chapman and Hall, London, UK. Rojo, E. E., B. Guiard, W. Neupert, and R. A. Stuart. 1998. Sorting of D-lactate dehydrogenase to the inner membrane of mitochondria. J. Biol. Chem. 273:8040–8047. Roostita, R., and G. H. Fleet. 1996. The occurrence and growth of yeasts in Camembert and Blue-veined cheeses. Int. J. Food. Microbiol. 28:393–404. Sabate, J., J. Cano, A. Querol, and J. M. Guillamo´n. 1998. Diversity of Saccharomyces strains in wine fermentations: Analysis for two consecutive years. Lett. Appl. Microbiol. 26:452–455. Seiler, H., 1986. Identification of cheese-smear coryneform bacteria. J. Dairy Res. 53:439–449. Seiler, H., and M. Busse. 1990. The yeasts of cheese brines. Int. J. Food Microbiol. 11:289–304. Sørensen, B. B., and M. Jakobsen. 1997. The combined effect of temperature, pH and NaCl on growth of Debaryomyces hansenii analysed by flow cytometry and predictive microbiology. Int. J. Food Microbiol. 34:209–220. Van den Tempel, T., and M. Jakobsen. 1998. Yeasts associated with Danablu. Int. Dairy J. 8:25–31. Van den Tempel, T., and M. Jakobsen. 2000. The technological characteristics of Debaryomyces hansenii and Yarrowia lipolytica and their potential as starter cultures for production of Danablu. Int. Dairy J. 10:263–270. Van den Tempel, T., and M. S. Nielsen. 2000. Effects of atmospheric conditions, NaCl, pH on growth and interactions between moulds and yeasts related to blue cheese production. Int. J. Food Microbiol. 57:193–199. Valde´s-Stauber, N., S. Scherer, and H. Seiler. 1997. Identification of yeasts and coryneform bacteria from the surface microflora of brick cheeses. Int. J. Food Microbiol. 34:115–119. Wood, B. J. B., ed. 1998. Microbiology of Fermented Foods. Vol 1. 2nd ed. Thomason Science, New York, NY.