Determination of survival, identity and stress resistance of probiotic bifidobacteria in bio-yoghurts


Martin R. Adams, School of Biomedical and Molecular Sciences, University of Surrey, Guilford GU2 7XH, UK. E-mail:


Aims:  To determine the level of bifidobacteria in bio-yoghurts in the UK, identify the species, and compare the resistance of common Bifidobacterium spp. to acidity and oxidative stress.

Methods and Results:  A storage trial of bio-yoghurts was carried out to determine the level and survival of bifidobacteria. The 16S rRNA gene targeted PCR was used to identify the species. Acid tolerance was determined by introducing the organisms to pH-adjusted skimmed milk and enumerating during storage at 4°C. Oxidative stress resistance was determined using the H2O2 disc diffusion assay technique. Nine of 10 bio-yoghurts contained bifidobacteria at levels >106 CFU g−1 at the time of purchase. The viability of the organism decreased during storage and on expiry only five products retained viability >106 CFU g−1 while two others were very close to the target population. Bifidobacterium animalis ssp. lactis showed superior survival abilities and stress tolerance compared with Bifidobacterium longum, Bifidobacterium breve, Bifidobacterium bifidum, Bifidobacterium adolescentis and Bifidobacterium longum biotype infantis.

Conclusions: Bifidobacterium animalis ssp. lactis, the only Bifidobacterium spp. found in bio-yoghurts, had the greatest resistance to acidity and oxidative stress.

Significance and Impact of the Study:  The technological properties of B. animalis ssp. lactis make it suitable for inclusion in bio-yoghurts although its putative health benefits need further investigation.


Probiotic bacteria can be defined as living micro-organisms, which, when consumed in sufficient numbers, exert health benefits beyond inherent basic nutrition (Guarner and Schaafsma 1998). Bifidobacterium, which is normally part of the human and animal gut microflora, is one genus that has been associated with such properties. The potential beneficial roles of bifidobacteria in the human intestine such as prevention of diarrhoeal diseases and protection from some forms of cancer have been documented and reviewed (Shah 2001). As a consequence, much effort has been devoted to the incorporation of bifidobacteria into fermented dairy products, baby foods, livestock feed supplements and pharmaceutical preparations (Crittenden 1999).

To have any effect, bifidobacteria must remain viable until they reach the intestinal tract. This requires their survival in the food vehicle during its shelf-life and their resistance to the acidic conditions of the stomach and bile salts in the small intestine (Kailasapathy and Rybka 1997). In view of these barriers, it is regarded as essential that: (i) carrier foods contain at least 106 CFU g−1; (ii) the species are of human origin; and (iii) the total intake of the product is c. 300–400 g per week (Samona and Robinson 1994). A recent meta-analysis has suggested that levels of lactobacilli of 107 or 109 CFU g−1 may be necessary to elicit anti-diarrhoeal effects (Van Niel et al. 2002) and currently there are proposals within the EU that the minimum level of probiotic bacteria in foods be set at 107 CFU g−1 (EU/AGRI/38743/2003rev3).

Several authors have reported low viable counts of bifidobacteria (Iwana et al. 1993; Micanel et al. 1997) and the presence of bifidobacteria of animal origin in bio-yoghurts (Biavati et al. 1992; Iwana et al. 1993), but no comparable studies from the UK have been published.

The objectives of this study were to determine the level of bifidobacteria in bio-yoghurts in the UK, identify the strains and compare the resistance of commonly used organisms to the acidity encountered in such products and oxidative stress.

Materials and methods

Determination of the population level of Bifidobacterium spp. in bio-yoghurts in the UK

Ten bio-yoghurts (A–J) available in the UK market were purchased 3 weeks prior to expiration, stored at 4°C and tested for viable bifidobacteria every week until 3 weeks past the expiry date. At each sampling, 1 g of yoghurt was aseptically weighed into a stomacher bag and mixed with 9 ml of maximum recovery diluent (MRD; Oxoid Ltd, Hampshire, UK). Serial dilutions were made using MRD with 0·05% cysteine (Sigma Ltd, Poole, UK). A volume of 100 μl was spread onto triplicate plates of Bifidobacterium Iodoacetate Mediacum (BIM; Muńoa and Pares 1988). The plates were incubated at 37°C in an anaerobic gas pack system (Oxoid Ltd) for 5 days. Colonies were counted using a Quebec colony counter (Bibby Sterilin Ltd, Stratfordshire, UK) and counts expressed as CFU g−1. The pH values were determined using Hanna HI8424 portable pH meter (Hanna Instruments, Milan, Italy). At each sampling, two new yoghurt pots were opened. The experiment itself was also replicated.

Identification of currently used Bifidobacterium spp. by PCR

The 16S rRNA gene targeted PCR technique was used to identify the Bifidobacterium spp. isolated. The Bifidobacterium spp. were isolated from products using the selective medium BIM and purified cultures were stored frozen on beads (Protect; Technical Service Consultants Ltd, Lancashire, UK) at −80°C. The pure organisms were streaked, in the form of a lawn, onto De Man Rogosa Sharpe agar (MRS; supplemented with 0·05% cysteine; De Man et al. 1960) plates and incubated anaerobically at 37°C for 24 h. The cells were harvested by adding 1 ml of MRS broth and scraping into a 1·5 ml microfuge tube (Axygen Scientific, Inc., Union City, CA, USA). The DNA was extracted using guanidium thiocyanate according to Pitcher et al. (1989). The 16S rRNA gene was amplified (PCR) using the primers pA (5′-AGAGTTTGATCCTGGCTCAG-3′) and pE* (5′-CCGTCAATTCCTTTGAGTTT-3′; Sigma Genosys Ltd, Haverhill, UK) and the cyclic amplification procedure included 5 min at 95°C, 35 cycles of 1 min at 94°C, 1 min at 55°C, 1 min at 72°C and another 5 min at 72°C. The PCR products were purified using the Qiagen PCR purification kit (Qiagen Ltd, West Sussex, UK) and the mixture was prepared with the primer pD (5′-GTATTACCGCGGCTGCTG-3′; Sigma Genosys Ltd) for sequencing. Sequencing was carried out using the DNA sequencer CEQ 2000XL DNA Analysis System (Beckman Coulter Inc., Fullerton, CA, USA). The base sequencing results were submitted on-line to the National Center for Biotechnology Information (NCBI, US National Library of Medicine, Madison, WI, USA; and were matched, using a BLAST search (Basic Logical Alignment Search Tool; Altschul et al. 1990), with the existing database of bifidobacteria.

Determination of the survival of Bifidobacterium spp. at 4°C in skimmed milk

The survival of Bifidobacterium longum NCTC11818, Bifidobacterium breve NCIMB702258, Bifidobacterium longum biotype infantis NCIMB702205 (BLI), Bifidobacterium adolescentis NCIMB702204, Bifidobacterium bifidum NCIMB702203 and Bifidobacterium animalis ssp. lactis USCC50051 (BAL; a strain isolated from a commercial brand and kept in the University of Surrey Culture Collection, Guildford, UK) was determined. The organisms were grown and subcultured on MRS agar three times to ensure the purity and viability. The organisms were then inoculated into MRS broth and incubated anaerobically at 37°C for 24 h. Cells were harvested by centrifugation (6000 g for 20 min at 25°C), washed twice with 0·1% sterile saline solution and resuspended in MRD, which served as the inoculum.

The ultra-high temperature (UHT) skimmed milk (<0·1% fat) was used as the survival medium. The concentrated bifidobacteria inoculum was introduced into 500 ml portion of sterile skimmed milk in Duran bottles (pH adjusted to 4·25 using 88% lactic acid; BDH Chemicals Ltd, Poole, UK) to have a final bifidobacteria population of c. 108 CFU ml−1. The bottles were stored in the cold room (4°C) and samples were taken every 12 h for B. longum, B. breve, B. bifidum, B. adolescentis and BLI, and every 3 days for BAL to determine the bifidobacterial population. A sample (1 ml) of milk was aseptically withdrawn and serially diluted using MRD. A volume of 100 μl was spread onto triplicate MRS agar, and plates were anaerobically incubated at 37°C for 48 h. Colonies were counted using a Quebec colony counter (Bibby Sterilin Ltd) and counts expressed as CFU ml−1.

Another experiment was carried out to compare the acid tolerance of B. longum NCTC11818 with that of BAL at pH 4·0, 4·25 and 4·5. Cultures were prepared as above in UHT skimmed milk (<0·1% fat) adjusted to pH 4·0, 4·25 and 4·5, and stored at 4°C. Samples were taken every day for B. longum and every week for BAL to determine the bifidobacteria population as described above.

Determination of the oxidative stress resistance of Bifidobacterium spp.

The oxidative stress resistance of B. longum NCTC11818, B. breve NCIMB702258, BLI, B. adolescentis NCIMB702204, B. bifidum NCIMB702203 and BAL was determined. The organisms were grown in MRS broth until the culture reached an optical density of 0·4 at 600 nm using a UV-visible spectrophotometer (Model: Unicam Helios; Unicam Limited, Cambridge, UK). A volume of 2·5 ml of the broth was introduced into 50 ml of 0·4% MRS agar [0·2 g of bacteriological agar (Oxoid Ltd) + 50 ml of MRS broth] and mixed well. A volume of 4 ml of inoculated 0·4% MRS agar was poured onto preprepared MRS agar plates and allowed to solidify. A sterile paper disc (6 mm; Whatman International Ltd, Maidstone, UK) was placed on each Petri dish and a volume of 10 μl of 3% H2O2 (Sigma-Aldrich Ltd, Poole, UK) was dispensed onto the paper disc. The plates were incubated anaerobically at 37°C for 24 h and the diameter of the inhibition zones around the paper discs were measured.

Statistical analysis

The experiments were carried out in triplicate, unless otherwise stated, and analysis of variance (anova) was determined using the software Statistical Analysis Systems, SAS version 8 for Windows (SAS Institute Inc., Cary, NC, USA). Mean comparisons were performed using the Duncan's multiple range test in SAS. The probability level of 5% (α = 0·05) was used to indicate the significance.


Nine of ten bio-yoghurts contained >106 CFU g−1 of bifidobacteria when purchased (Table 1). All the products showed a steady decline in viable numbers during storage and five of ten contained >106 CFU g−1 of bifidobacteria on expiry, whereas two other products contained levels very close to the target level (Table 1). Brand A contained an initial bifidobacterial population of 105 CFU g−1, which decreased to 103 CFU g−1 on expiry (Table 1). The initial pH values of yoghurts ranged from 4·18 to 4·46 and on expiry they ranged from 3·98 to 4·12. There were no significant differences (P > 0·05) in pH values among the products tested.

Table 1.  Viable bifidobacterial populations in bio-yoghurts in the UK
BrandMean population (CFU g−1) 3 weeks before expiryMean population (CFU g−1) on expiryLog reduction
  1. Each value in the table is a mean ± SD of four determinations in two separate trials.

A1·0 × 105 (±0·31)1·1 × 103 (±0·19)1·96
B6·2 × 106 (±0·13)1·4 × 105 (±0·26)1·65
C1·0 × 108 (±0·09)2·6 × 106 (±0·12)1·59
D2·4 × 108 (±0·25)5·5 × 106 (±0·24)1·64
E1·3 × 108 (±0·19)9·1 × 106 (±0·36)1·15
F2·0 × 106 (±0·12)1·3 × 105 (±0·51)1·19
G1·8 × 107 (±0·31)4·3 × 106 (±0·42)0·62
H1·6 × 107 (±0·13)9·1 × 105 (±0·11)1·24
I1·6 × 107 (±0·23)1·3 × 106 (±0·42)1·09
J3·1 × 107 (±0·23)9·5 × 105 (±0·33)1·51

The 16S rRNA partial sequencing results showed that the best-matched bifidobacterial species isolated from these products was BAL. According to BLAST results, isolates from products A, B, C, D, E, F, G, H, I and J showed 98% (485/491), 99% (413/417), 97% (330/340), 97% (452/462), 99% (466/470), 99% (468/472), 99% (376/379), 97% (332/342), 98% (483/491) and 99% (466/470) of nucleotide sequence alignment, respectively, with the reference strain BAL.

Probiotic effects in humans are more commonly associated with Bifidobacterium spp. of human origin (Reddy and Rivenson 1993). To investigate the technological feasibility of incorporating human isolates into yoghurts, the acid and oxygen tolerance of several species of human origin was compared with that of BAL isolated from bio-yoghurts. All Bifidobacterium spp. except BAL declined in numbers and decreased below 106 CFU ml−1 in 7 days of storage at 4°C in pH-adjusted skimmed milk (Fig. 1). Bifidobacterium breve showed slightly better survival than the other human strains (Fig. 1), but BAL survived for more than 5 weeks above 106 CFU ml−1.

Figure 1.

Survival of Bifidobacterium longum (bsl00066), Bifidobacterium breve (bsl00001), Bifidobacterium adolescentis (bsl00084), Bifidobacterium longum biotype infantis (○), Bifidobacterium bifidum (•) and Bifidobacterium animalis ssp. lactis (□) in skimmed milk at 4°C with pH adjusted with lactic acid to 4·25. Error bars show the SD of the mean of nine determinations in three experiments.

The survival of B. longum (typical of the human strains tested and often claimed to be present in bio-yoghurts) was markedly affected by the different pH values (Fig. 2); the lower the pH the faster the decline in viability. BAL showed no significant loss of viability over 2 weeks at 4°C at different pH values (P < 0·05) and declined by <1 log cycle.

Figure 2.

Survival of Bifidobacterium longum NCTC11818 [pH 4·0 (□), 4·25 (bsl00066) and 4·5 (bsl00001)] and Bifidobacterium animalis ssp. lactis USCC50051 [pH 4·0 (○), 4·25 (•) and 4·5(bsl00084)] in skimmed milk at 4°C. Error bars show the SD of the mean of nine determinations in three experiments.

Bifidobacterium animalis ssp. lactis USCC50051 was also more tolerant of oxidative stress than human bifidobacterial species showing significantly (P < 0·05) smaller inhibition zones in the H2O2 disc diffusion assay (Fig. 3). In marked contrast to acid tolerance results, among human strains, BLI was the most resistant of the oxidative stress while B. adolescentis was the most sensitive.

Figure 3.

Comparison of oxidative stress resistance of Bifidobacterium spp. Error bars show the SD of the mean of nine determinations in three experiments.


Bifidobacteria should be present at >106 CFU g−1 in bio-yoghurts on expiry to have health benefits (Kurmann and Râsíc 1991). All the bio-yoghurts tested (except one brand) achieved this level when purchased. However, the sooner the products are consumed, the greater the chances of consuming an adequate dose as only five of 10 commercial bio-yoghurts contained this level at their ‘best before’ date. Brand A, which showed the lowest counts of bifidobacteria, was a stirred yoghurt containing fruits and poor survival of bifidobacteria in this product may be related to different processing conditions and composition. Low viable bifidobacterial counts in bio-yoghurts have been previously reported (Iwana et al. 1993; Micanel et al. 1997) although it was not clear whether the enumeration was carried out on expiry. A similar study showed adequate bifidobacteria counts in bio-yoghurts in the USA on expiry (Shin et al. 2000). Microbial interactions could contribute to the viability of bifidobacteria in bio-yoghurts. Different producers use different combinations of yoghurt starter and probiotic bacteria, and these combinations could have mutualistic or antagonistic effects on bifidobacteria.

One product, tested in this study, claimed to contain B. longum whereas another claimed to contain B. lactis. Eight brands did not specify the Bifidobacterium spp. present. The only bifidobacterial species found in this study was BAL. Meile et al. (1997) isolated, from yoghurts, a hitherto new bifidobacterial species and classified it as B. lactis. However, it was later demonstrated that both B. lactis and B. animalis strains belong to the same species and it was further suggested that they could be unified as the B. animalis spp. divided into two subspecies, B. animalis ssp. lactis and B. animalis ssp. animalis (Ventura and Zink 2003). Bifidobacterium animalis is a species of animal origin (Ishibashi and Shimamura 1993). The presence of B. animalis in bio-yoghurts in Europe (Biavati et al. 1992; Iwana et al. 1993) and Canada (Roy et al. 1996) has been reported.

The acid tolerance of bifidobacteria is important as the organism has to withstand the acidity in yoghurts and the gut, it has been reported to have health benefits. For the digestion of food, the pH of gastric juice has to be around 3–4 (Takiguchi and Suzuki 2000) and the pH of yoghurt varies from 3·8 to 4·36 (Shah et al. 1995). The results in the present study showed that BAL survived significantly longer (P < 0·05) in high acid environments over other species. Bifidobacterium breve showed slightly better survival ability over other human-associated species. The superior survival ability of BAL in acidic conditions has been observed elsewhere (Matsumoto et al. 2004). They reported that of 17 strains nine Bifidobacterium spp., BAL had the highest acid tolerance. In their study, HCl was used as the acidulent. In our study, lactic acid, a more potent antimicrobial than HCl, was used to simulate the acidic environment in the product. When bacteria are exposed to acidic conditions, a pH homeostasis is maintained by discharging H+ from the cell (Booth 1985). This process is dependent on the activity of H+-ATPase, the enzyme responsible for maintaining the H+ concentration between the cell and environment. A higher H+-ATPase activity in BAL at low pH compared with other Bifidobacterium spp. has been observed (Matsumoto et al. 2004).

Bifidobacteria are strictly anaerobic and under aerobic conditions highly toxic reactive oxygen intermediates such as superoxides, hydroxy radicals and peroxides are formed within the cells (Talwalkar and Kailasapathy 2004). Exposure to oxygen may occur during processing and after opening pots before consumption. Oxidative stress resistance is a measure of the ability of bacteria to survive such conditions. In this study, BAL showed significantly higher (P < 0·05) oxidative stress resistance than the human-associated species, thus, demonstrating its technological superiority.

BAL appears to be superior to human-associated Bifidobacterium spp. in terms of its technological properties such as acid tolerance and resistance to oxidative stress. Although there is some evidence of beneficial effects of BAL in the case of atopic eczema (Isolauri et al. 2000) and infant diarrhoea (Chouraqui et al. 2004; Weizman et al. 2005), the health benefits of animal bifidobacterial species have not been researched to the same extent as human bifidobacterial species.