TEXT

Top Abstract Text References

The genus Bifidobacterium is the third most numerous bacterial population in the human intestine after the genera Bacteroides and Eubacterium (4). It is also the predominant population during childhood (9). Some species have been observed in high numbers in human feces (1). Because of these characteristics, it has been proposed that Bifidobacterium be used as an indicator of fecal contamination (8, 14), although this proposal is controversial. Several studies pointed out species-specific relationships between certain Bifidobacterium species and their hosts which suggest the use of these species to determine the origin of fecal contamination ( 3, 12, 17, 19). Moreover, Bifidobacterium spp. are unlikely to grow in water because of their obligate anaerobic physiology and complex growth requirements. Such characteristics support the possible use of Bifidobacterium as a fecal indicator. Several Bifidobacterium species have been incorporated in yogurts, as they are considered to have health-promoting properties (9). The availability of easy and inexpensive methods for detection, identification, and enumeration of Bifidobacterium is consequently important in environmental and food microbiology. Several media for selective or differential isolation have been described (2, 11, 13, 15-17). Most of them have complex compositions that include antibiotics as growth inhibitors, require long incubation times, or show low bacterial recovery levels. These factors hinder their routine use for monitoring when Bifidobacterium is used as a fecal indicator or for the enumeration of bifidobacterial populations in dairy products. In this study, a new selective medium for Bifidobacterium spp. is described. It has been evaluated for the enumeration of bifidobacteria in yogurts.

Forty-two bacterial strains were used in this study. The media used to cultivate the different strains are listed in Table 1, and the strains are listed in Tables 2 and 3. Bifidobacterium strains were grown at 37°C for 48 h in anaerobic jars (GasPak; BBL). Strains other than Bifidobacterium spp. were grown at 37°C for 24 h. Bacteroides fragilis was also cultured in anaerobic jars. The enumeration of bacteria in yogurts was carried out with the differential LS-Cennan medium (ADSA, Barcelona, Spain) (7) and incubation of plates at 37°C for 48 h in anaerobic jars (GasPak; BBL). Bacterial strains in yogurts were identified by colonial morphology on the differential medium according to the method described by Eloy in 1976 (7), Gram staining, and cell morphology.

An initial list of the potential components of the new medium was based on the components of previously described media (16) and the physiological characteristics of the genus (10). The significance of the effects these components have on growth and selectivity for Bifidobacterium were studied by the statistical model 2^k (5, 6), which analyzes the importance of k factors in a determinate event. Fifteen factors (k = 15) were selected for evaluation at two levels or concentrations each, as follows: meat extract, 2 and 5 g/liter; yeast extract, 3 and 7 g/liter; starch, 0.5 and 2 g/liter; lactulose, 1 and 5 g/liter; L-cysteine, 0.5 and 2 g/liter; sodium chloride, 1 and 5 g/liter; casein, 1 and 5 g/liter; maltose, 1 and 5 g/liter; peptone, 2 and 5 g/liter; tryptone, 2 and 5 g/liter; riboflavin, 0.25 and 1 mg/liter; thiamine, 0.25 and 1 mg/liter; pantothenic acid, 0.5 and 2 mg/liter; propionic acid, 1 and 5 ml/liter; and pH levels of 5 and 6.5. According to these parameters, 2¹⁵ media could be defined. The number of potential media was reduced to 40 combinations (statistical design, 2^kp = 40) in order to minimize analyses. The bacterial strains used to evaluate the media was Bifidobacterium pseudolongum DSM 20099^T due to its high growth rate. A McFarland no. 3 cell suspension in 0.25× Ringer solution supplemented with 0.5 g of L-cysteine/liter was prepared by growing B. pseudolongum on Columbia blood agar (CBA) supplemented with 5 g of glucose per liter and 0.5 g of L-cysteine per liter for 48 h. Ten-fold dilutions were made from these suspensions in 0.25× Ringer solution supplemented with 0.5 g of L-cysteine/liter. Four replicates of plating for each experimental medium were performed. Plates were inoculated with aliquots of 10 µl of each dilution and incubated at 37°C for 48 h in anaerobic jars. The media were compared by counting the B. pseudolongum colonies after 48 h of culture. This permitted a comparison of the levels of recovery on the different media tested to establish the basic composition of the new selective medium by using the components with a positive effect on the growth of Bifidobacterium. The use of some inhibitors for bacterial species physiologically or ecologically related to Bifidobacterium (Enterobacteriaceae, Enterococcus spp., and Lactobacillus spp.) was also studied. The inhibitors and their concentrations were determined according to previous studies (11, 20) and were as follows: methylene blue, 0.130 g/liter; lithium chloride, 2 g/liter; propionic acid, 5 ml/liter; crystal violet, 0.4 mg/liter; dodecyl sulfate sodium salt, 0.1 g/liter; propionic acid sodium salt, 3 g/liter; sodium sulfite, 2 g/liter; fuchsin basic, 0.5 g/liter; and a pH of 5.5. Each inhibitor was tested by including it in the basic composition of the medium established in the previous study. The colony counts of the media with or without inhibitors were compared by using B. pseudolongum DSM 20099, Lactobacillus bulgaricus CECT 4005, and Escherichia coli CN13. A McFarland no. 3 cell suspension for each strain was made in 0.25× Ringer solution supplemented with 0.5 g of L-cysteine/liter after 48 h of growth on CBA. Ten-fold dilutions were performed in 0.25× Ringer solution supplemented with L-cysteine. Plates of each experimental medium with inhibitor were inoculated in the same way as plates in the 2^kp experiment described earlier. Plates were incubated at 37°C for 48 h in anaerobic jars. Those inhibitors that reduced the colony counts for L. bulgaricus and E. coli but not for B. pseudolongum were added to the final composition of the new Bifidobacterium medium (BFM).

The selectivity of BFM was determined by comparing the growth of 26 strains belonging to 24 Bifidobacterium species (Table 2) and 16 strains from other bacterial genera (Table 3). In order to measure the recovery on BFM agar, two media were used for comparison: CBA supplemented with glucose and L-cysteine, an enriched medium, and a previously described Bifidobacterium medium, BL medium (21). Bifidobacterial strains were grown in Reinforced Clostridial Medium (RCM) for 48 h at 37°C in 25-ml glass tubes. An aliquot of 0.1 ml from this culture was again inoculated in RCM and incubated for 24 h at 37°C. Ten-fold dilutions of this bacterial culture were performed in 0.25× Ringer solution supplemented with 0.5 g of L-cysteine/liter. Aliquots (10 µl) of 10-fold dilutions were plated in duplicate on BFM, CBA supplemented with glucose and L-cysteine, and BL medium. Plates were incubated anaerobically for 48 h at 37°C, and colonies were counted. The recovery of strains other than Bifidobacterium spp. was also evaluated by colony counting on BFM and on the appropriate medium (Table 1) after incubation at 37°C for 24 h.

The enumeration of Bifidobacterium spp. in three different commercial yogurts was performed twice to evaluate the use of BFM. Ten-fold dilutions from each yogurt were prepared in 0.25× Ringer solution supplemented with 0.5 g of L-cysteine/liter and plated on BFM and LS-Cennan medium (ADSA) in duplicate. Colonies of different morphology were counted after 48 h of incubation at 37°C in anaerobic jars. Morphologically different colonies on BFM agar and LS-Cennan medium were characterized by Gram staining and the fructose-6-phosphate phosphoketolase test (7, 10).

Eleven of the 15 factors tested enhanced the growth of B. pseudolongum DSM 20099 and constitute the composition of BFM. Low concentrations of meat extract and tryptone had a greater effect on the colony numbers than did high concentrations. In contrast, high concentrations of yeast extract, starch, lactulose, L-cysteine, sodium chloride, peptone, riboflavin, and thiamine yielded colony counts higher than those yielded with low concentrations. No effect on the colony enumeration for B. pseudolongum DSM 20099 was observed with casein, maltose, or pantothenic acid. Methylene blue (16 mg/liter), propionic acid (99%; 5 ml/liter), lithium chloride (2 g/liter), and a pH of 5.5 did not inhibit colony growth of B. pseudolongum DSM 20099, although they did inhibit colony growth of strains of other genera, such as Lactobacillus spp. and E. coli. The new medium (BFM) was established with the following components: meat extract (2 g/liter; Merck, Darmstadt, Germany), yeast extract (7 g/liter; Scharlau, Barcelona, Spain), starch (2 g/liter; Panreac, Barcelona, Spain), L-cysteine hydrochloride (0.5 g/liter; Sigma, St. Louis, Mo.), sodium chloride (5 g/liter; Pancreac), peptone 5 g/liter; Difco, Detroit, Mich.), tryptone (2 g/liter; Scharlau), lactulose (5 g/liter; Merck), riboflavin (1 mg/liter; Merck), thiamine chloride hydrochloride (1 mg/liter; Merck), methylene blue (16 mg/liter; Merck), lithium chloride (2 g/liter; Merck), and propionic acid (99%; 5 ml/liter; Sigma). All components except propionic acid were mixed and resuspended in distilled water. The medium was sterilized by autoclaving at 121°C for 20 min and then allowed to cool to 55°C. Propionic acid was then added, and the pH was adjusted to 5.5 with sterile 10 N NaOH. When BFM agar plates were made, agar-agar (ADSA) was added at a concentration of 15 g/liter. The principal carbon source in this medium is lactulose. The growth stimulation of this oligofructose for Bifidobacterium observed in this study was consistent with that of other studies (9, 18). Two vitamins (riboflavin and thiamine) acted as growth factor promoters. Propionic acid, lithium chloride, and methylene blue inhibited the growth of some related bacterial species. The low pH (5.5) of BFM inhibited the growth of Enterobacteriaceae.

The selectivity of BFM did not affect the growth of 23 of the 26 Bifidobacterium strains tested. Most of them showed round, blue colonies approximately 2 mm in diameter a few minutes after removal from the anaerobic jar. There was no growth of colonies of three strains (Bifidobacterium bifidum DSM 20456, Bifidobacterium coryneforme DSM 20216, and Bifidobacterium pullorum DSM 20433) on BFM, but two others strains of B. bifidum (DSM 20082 and DSM 20239) grew on BFM. No colony growth was observed on BFM for the following strains: Lactobacillus acidophilus CECT 903, Vibrio furnissii LMG 7910, Vibrio cholerae CECT 658, Aeromonas hydrophila CECT 398, Salmonella typhimurium WG49, E. coli WG5, Proteus penneri CECT 864, Edwardsiella tarda CECT 849, Yersinia enterocolitica CECT 754, and Streptococcus thermophilus CECT 986. Four tested isolates that do not belong to the Bifidobacterium genus grew on both media (Table 3). However, these isolates were easily differentiated from Bifidobacterium spp. because of their small diameters (<0.5 mm) and differential colonial morphologies.

For most of the 26 Bifidobacterium strains, similar colony counts were obtained on BFM, CBA, and BL medium (Table 2). However, lower colony counts on BFM agar with respect to CBA and BL medium were observed for Bifidobacterium adolescentis DSM 20083, B. bifidum DSM 20456, Bifidobacterium catenulatum DSM 20103, B. coryneforme DSM 20216, Bifidobacterium longum DSM 20219, and B. pullorum DSM 20433 (Table 2). No colony growth on BFM agar was observed for most of the studied strains other than Bifidobacterium spp. The colony counts on BL agar for these strains were similar to these obtained on rich media. BL medium has been described as an optimal medium for Bifidobacterium (16). However, BFM is easier to prepare and proved to be more selective than BL medium (Table 3). A lower level of recovery from BFM agar was observed for L. bulgaricus CECT 4005. Only three strains other than Bifidobacterium (Lactobacillus plantarum CECT 748, Enterococcus faecalis CECT 481, and Enterococcus faecium CECT 410) showed colony counts on BFM agar similar to those obtained on rich media, although their colony sizes were smaller on BFM (Table 3) as described earlier.

Three colony morphologies were observed on LS-Cennan medium inoculated with the yogurt tested. Gram staining and cell morphology observed via optical microscopy confirmed the presumptive relationship between genus and colony morphology for the usual bacterial components of this type of yogurt (Streptococcus, Lactobacillus, and Bifidobacterium) (7). Large, irregular, red colonies were gram-positive rods (Lactobacillus spp.); large, round, red colonies were gram-positive cocci (Streptococcus spp.); and small, round colonies were gram-positive polymorphic cells (Bifidobacterium spp.). The fructose-6-phosphate phosphoketolase test confirmed that the latter isolates belonged to the genus Bifidobacterium (10). Two colonial morphologies, small and large, were observed on BFM agar. Both belonged to Bifidobacterium spp. according to the results with Gram stain and the fructose-6-phosphate phosphoketolase test (10). Large colonies presented both morphological colony types when they were again streaked on BFM. No differences in colony counts of bifidobacteria on BFM agar and LS-Cennan medium were observed for the three yogurts. The colonial morphologies associated with Lactobacillus spp. and Streptococcus spp. on LS-Cennan medium (7) were not found on BFM agar (Table 4). BFM is suitable for the bifidobacteria present in yogurt. It inhibits the growth of Lactobacillus and Streptococcus and allows a differential growth of bifidobacteria.

In conclusion, the simple composition of BFM makes it a feasible medium for routine monitoring purposes. Its selectivity could provide a differential tool for the recovery of bifidobacteria in environmental and food microbiology. The recovery on BFM agar suggests the practical use of BFM for enumeration of Bifidobacterium in routine monitoring of fermented dairy products such as yogurts.