Identification and characterization of starter lactic acid bacteria and probiotics from Columbian dairy products


Sigrid C.J. De Keersmaecker, Centre of Microbial and Plant Genetics, Katholieke Universiteit Leuven, Kasteelpark Arenberg 20, B-3001 Leuven, Belgium. E-mail:


Aims:  Considering the significant rise in the probiotic market in Columbia, and given the lack of reports concerning the microbial population and strain performance in products from different producers, this study aims at determining the number of viable starter bacteria and probiotics in bio-yoghurts available at the Columbian market, identifying the species and analysing the performance of the isolated strains in bile acid resistance, antagonistic activity against pathogens, and adherence capacity to human intestinal epithelial cells.

Methods and Results:  Seven bio-yoghurts were analysed for the bacterial species present. Species identification was carried out using 16S rRNA gene targeted PCR. The cultured bacteria were tested for bile acid resistance, adherence to a human intestinal epithelial cell line, and antagonism against the pathogen Salmonella enterica serovar Typhimurium. A total of 17 different strains were identified. Based on plate counting, all bio-yoghurts have at least total viable cells of ∼107 CFU ml−1. Streptococcus thermophilus and Lactobacillus delbrueckii ssp. bulgaricus were the most frequently isolated bacteria. Viable Bifidobacterium was only recovered from one product. However, after PCR analysis, DNA of this genus was confirmed in five out of seven products. Major differences were found for S. typhimurium antagonism. The adherence capacity to Caco-2 cells was observed in 10 of the isolated strains. In general, low survival to simulated gastric juice was observed.

Conclusions:  Some of the isolated strains have probiotic potential, although not all of them were present in the advised amount to exert beneficial health effects. However, the full correct scientific name of the isolated bacteria and their viable counts were not included on the product label.

Significance and Impact of the Study:  This is the first report describing the identification and functionality of starter bacteria and probiotics present in dairy products on the Columbian market.


Probiotics are defined as ‘live microorganisms which when administered in adequate numbers confer a health benefit on the host’ (FAO/WHO 2001). To date, probiotics have been mainly selected from the genera Lactobacillus and Bifidobacterium, because of their long history of safe use in dairy industry, and their natural presence in the human intestinal tract (Saxelin 1997).

The term starter lactic acid bacteria (SLAB) is used to describe bacterial strains capable of growth under the selective conditions of dairy production. A starter culture of traditional yoghurt normally includes a mixture of Streptococcus thermophilus and Lactobacillus delbrueckii ssp. bulgaricus. However, although L. delbrueckii ssp. bulgaricus was reported to survive the gastric transit (Mater et al. 2005), it is not native to humans in contrast to most probiotic bacteria for human consumption. Consequently, in order to improve yoghurt’s health value, the traditional L. delbrueckii ssp. bulgaricus in many products has been replaced by or combined with well-known probiotics like Lactobacillus acidophilus, Lactobacillus casei and Lactobacillus rhamnosus, which are frequently present in the human gastro-urogenital tract (Heller 2001; Vaughan et al. 2002; Reid and Bruce 2003).

In general, the selection criteria for starter cultures are their acidification rate and flavour-producing characteristics. For probiotics, selection is based on a detectable health effect for the host (Saarela et al. 2000; Teusink and Smid 2006). The essential characteristics for lactic acid bacteria (LAB) to be used as probiotics during manufacturing include the following: (i) recognition as safe (GRAS; generally recognized as safe); (ii) viability during processing and storage; (iii) antagonistic effect against pathogens; (iv) tolerance to bile acid challenge; and (v) adherence to the intestinal epithelium of the host among others (MacFarlane and Cummings 2002; Begley et al. 2005; Vesterlund et al. 2005; Lin et al. 2006). It is generally considered that minimum numbers required for a probiotic to provide a health benefit are 107 CFU ml−1 (Ross et al. 2005; Jayamanne and Adams 2006). Therefore, the viability and survival of probiotic bacteria during dairy processing and storage are important parameters for providing a health benefit (Corcoran et al. 2005; Lin et al. 2006). Fermented milk products containing probiotics were introduced on the Columbian dairy market at the beginning of this century. However, until now, there are no reports in scientific literature concerning the content and viability of probiotic and starter bacteria of these products. For this reason, we identified the bacterial species in seven yoghurts available on the Columbian market representing the seven bigger dairy food companies, and determined their in vitro performance. In addition, the viable counts of the starter bacteria and probiotics identified in the dairy products analysed, were determined.

Materials and methods

Culture conditions for the isolation of viable bacteria

A total of seven dairy products (A–G) were purchased at supermarkets in Columbia. Three samples from each product were assayed, and for each assay, triplicate results were obtained. Description and microbial composition labels on these products are listed in Table 1. Bacterial isolation was performed as previously described (Lin et al. 2006) with minor modifications. Briefly, 1 ml of the yoghurt sample was mixed with 9-ml sterilized phosphate buffered saline (PBS) (130-mmol l−1 NaCl, 10-mmol l−1 sodium phosphate buffer, pH 7·2) and stirred thoroughly. After serial (10-fold) dilutions with PBS, 100 µl of each dilution was spread on de Man Rogosa Sharpe (MRS) agar for LAB isolation, M17 agar for Lactococcus isolation (CM0785; Oxoid, Drongen, Belgium), ST agar (Dave and Shah 1997) for S. thermophilus isolation, RB agar (Hartemink and Rombouts 1999) for Bifidobacterium isolation, and kanamycin aesculine azide agar base (KAAAB) for the isolation of streptococci and enterococci (CM591; Oxoid, Drongen, Belgium). The agar plates (1·5%) were incubated at 30 and 37°C under aerobic and anaerobic conditions for 48 or 72 h. The anaerobic condition was achieved by the use of anaerobic jars (BBL Gas-pack Anaerobic Systems, VWR International, Haasrode, Belgium). In order to confirm the recovery of Bifidobacterium in the RB medium, a sample of Yoghurt Activia (Danone, Belgium) containing ‘Bifidus Essensis’ according to the label, was included as positive control.

Table 1.   Commercial probiotic dairy products analysed during this study
ProductDescription and compositions labelled
ABaby yoghurt, 113 g, Lactobacillus casei
BNatural yoghurt, 180 g, probiotic and lactic cultures
CLight yoghurt, 180 g, L. casei, Lactobacillus acidophilus
DYoghurt light (flavoured), 150 g, Streptococcus thermophilus, Lactobacillus delbrueckii ssp. bulgaricus, Lactobacillus acidophilus, Bifidobacterium spp.
ELight yoghurt, 150 g, lactic ferment
FNatural yoghurt, 210 ml, lactic cultures and probiotics
GNatural yoghurt, 200 g, probiotic and specific lactic cultures

Total DNA isolation

The total DNA from all the dairy products was isolated as previously described (Theunissen et al. 2005). The total DNA from the purified strains was isolated as previously described (De Keersmaecker et al. 2006a).

Identification of the isolated bacteria

The 16S rRNA gene of the isolated bacteria was amplified by PCR using the Pro-26 (5′-AGAGTTTGATCCTGGCTCAG-3′) and Pro-27 (5′-AAGGAGGTGATCCAGCCGCA-3′) primers (Eurogentec, Liège, Belgium) corresponding to the universal primers for the amplification of bacterial 16S rRNA (referred to as BSF8/20 and BSR1541/20) described in the Ribosomal DNA Primer Database (Wilmotte et al. 1993). PCR was carried out in a Personal Mastercycler (Eppendorf, VWR International, Haasrode, Belgium) according to the manufacturer’s instructions and as previously described (De Keersmaecker et al. 2006a). The PCR fragments of 1·5 kb were sequenced by the chain-termination dideoxynucleoside triphosphate method (Sanger et al. 1977) with the BigDye® Terminator V3.1 CycleSequencing Kit, using the ABI 3100-Avant Genetic Analyzer (Applied Biosystems, Lennik, Belgium). Databases were screened for similarities by using BLAST (Altschul et al. 1990, 1997) and the alignment of overlapping fragments was performed with the VectorNTI Advance™ 10 ContigExpress software (Informax, Oxford, United Kingdom). Furthermore, in order to determine whether the analysed dairy products (including the positive control Yoghurt Activia), contained Bifidobacterium spp., PCR amplification was carried out using DNA isolated from the dairy products as template, and the genus-specific primers Pro-488 (5′-GTSCAYGARGGYCTSAAGAA-3′) and Pro-489 (5′-CCRTCCTGGCCRACCTTGT-3′) (Eurogentec), corresponding to the primers B11 and B12 for the amplification of the hsp60 gene (Delcenserie et al. 2005). Sequencing and sequence analysis was performed as described earlier.

Antibacterial activity assay

The antimicrobial activity was assayed using the antagonist method previously described (Parret et al. 2003) with minor modifications. Briefly, 1 µl of a stationary-phase culture (16 h), corresponding to ∼108–109 CFU ml−1 was spotted onto MRS agar plates and incubated for 24 h at 37°C (MRS supplemented with 0·05% cysteine was used for Bifidobacterium). To prevent further cell growth, the plates were subsequently exposed to chloroform vapour (30 min) and subsequently overlaid with 3 ml of Luria Bertani (LB) soft agar (10% NaCl, 10% BactoPeptone, 5% yeast extract, 0·5% agar) seeded with 50 µl of a cell culture of the indicator strain, i.e. Salmonella enterica serovar Typhimurium strain SL1344 (Hoiseth and Stocker 1981) or strain ATCC 14028 (Miller et al. 1989) at a concentration of 108 CFU ml−1. The antagonist activity was assessed after overnight incubation at 37°C by determining the formation of clear zones of growth inhibition of the indicator strain around the test colonies. The inhibition zones were classified as (-) no visible inhibition, (+) 1- to 6-mm inhibition, (++) 7- to 12-mm inhibition, and (+++) more than 12-mm inhibition. The results were further confirmed using 24-h spent-culture supernatant (SCS) from the selected strains. Briefly, SCS was obtained from 24-h MRS culture of LAB by centrifugation (4000 g, 20 min) and filtered through a 0·22-µm pore sterile filter (Millipore, Brussels, Belgium). The cultures were all in the stationary phase and had a counting of ∼108–109 CFU ml−1. Fifteen microlitres of SCS, with or without neutralization (to pH 7·0 with 1-mmol l−1 NaOH), were added into a 5-mm-diameter well of a LB agar plate (1·5% agar). Subsequently, the plates were overlaid with 3 ml of soft agar (0·5%) containing the indicator strain, and antagonist activity was assessed as described earlier. For the antimicrobial assay activity, the commercial strain L. rhamnosus GG (ATCC 53103) (Sherwood and Gorbach 1996) was used as a positive control.

Survival of isolated strains in simulated gastric juice environment

The preparation of simulated gastric juice and survival experiments were performed as previously described (Corcoran et al. 2005) with minor modifications. Cultures of the isolated bacteria were grown overnight (16 h) in 15-ml MRS medium. The cultures were subsequently centrifuged at 4000 g at 4°C for 20 min, and washed once in an equal volume of cold PBS. The pellets were resuspended in 5-ml PBS, and the volume equivalent to approximately 108–109 CFU ml−1 was further centrifuged and resuspended in the appropriate volume of either simulated gastric juice or PBS. The suspensions were incubated at 37°C for 90 min with constant stirring. The samples were taken at 0, 30, 60 and 90 min, serially diluted in PBS, plated on MRS medium, or MRS supplemented with 0·05% cysteine in the case of Bifidobacterium, and incubated at 37°C for 72 h. For Bifidobacterium, strict anaerobic incubation was used. For the survival assay, L. rhamnosus GG was used as a positive control.

Culture of the intestinal epithelial cell line Caco-2

The Caco-2 human colon adenocarcinoma cells (ATCC HTB 37) (Pinto et al. 1983) were routinely grown in plastic flasks (75 cm2, Falcon) at 37°C in a 95% air–5% CO2 atmosphere in Dulbecco’s modified Eagle’s minimal essential medium (DMEM)/F-12 (1 : 1 v/v) medium [DMEM with 15-mmol l−1 HEPES buffer, l-glutamine, pyridoxine-HCl, without phenol red (GibcoBRL, Merelbeke, Belgium)], supplemented with 10% fetal bovine serum (FBS; HyClone, Erembodegem, Belgium). For maintenance of the cells, twice a week, the cultures were washed, trypsinized [trypsin-EDTA without Ca2+ and Mg2+ (GibcoBRL, Merelbeke, Belgium)] and seeded in new flasks. To form monolayers, Caco-2 cells were suspended in trypsin-EDTA, centrifuged at 1200 g for 7 min at 4°C, resuspended in the appropriate volume of DMEM/F-12 and seeded in 12-well culture plates. The culture medium was replaced every other day. Caco-2 cells were used between passages 54 and 56.

Adherence assay of the isolated strains

For the in vitro adherence inhibition assay, differentiated monolayers (15 days) were washed two times with sterile, prewarmed PBS. To each well, 1·5 ml of each bacterial suspension was added (109 CFU ml−1) and further incubated for 45 min at 37°C in a 95% air–5% CO2 atmosphere. Subsequently, the cells were washed twice with prewarmed PBS. Trypsin was added (100 µl) to the cells and further incubated for 5 min (at 37°C in a 95% air–5% CO2). Finally, 900 µl of PBS were added, the suspensions were serially diluted in PBS, plated out on MRS plates, and incubated at 37°C for 72 h. For Bifidobacterium animalis ssp. lactis (G-2), quantification of adhesion was performed by real-time PCR using the Bifidobacterium-specific primers (forward 5′-GCGTGCTTAACACATGCAAGTC-3′) and (reverse 5′-CACCCGTTTCCAGGAGCTATT-3′), and the Bifidobacterium spp. Taqman Probe (5′-TCACGCATTACTCACCCGTTCGCC-3′) as previously described (Penders et al. 2005). Finally, for the adherence assay, L. rhamnosus GG (Elo et al. 1991) and L. acidophilus JCM 1132 (ATCC 4356) (Kimoto et al. 1999) were used as positive and negative controls, respectively.


Identification of LAB

For all dairy products tested, anaerobic incubation conditions of the various plated media resulted in higher cell numbers as compared with aerobic conditions (data not shown). The summary of the identified strains is listed in Table 2. Streptococcus thermophilus and L. delbrueckii ssp. bulgaricus were among the most frequently isolated bacteria from the products, which was expected, as these two species are the main starter cultures of yoghurt. Apart from the control used (Yoghurt Activia) (data not shown), viable Bifidobacterium cells were recovered from product G, although the presence of Bifidobacterium was not mentioned on the label. In contrast, although product D claimed to contain Bifidobacterium spp., no viable bacteria belonging to this genus were recovered. However, additional PCR analysis using specific Bifidobacterium primers indicated the presence of DNA of this genus in products D and A, B, C and G as well (Fig. 1). All the dairy products analysed show an average of total viable cell counts ∼107 CFU ml−1. However, this is not the case for each strain separately (Table 2).

Table 2.   Viable anaerobic counts of commercial dairy products and strain identification
ProductCounting log10 (CFU ml−1) Species identified
  1. Each value in the table represents the mean value ± standard deviation of nine data points (triplicates of three samples per product).

  2. *In this particular case, the value corresponds to aerobic counts.

A-17·37 ± 0·03Lactobacillus paracasei
A-29·06 ± 0·07Streptococcus thermophilus
B-17·62 ± 0·01Leuconostoc mesenteroides
B-27·78 ± 0·13S. thermophilus
C-16·23 ± 0·22Lactobacillus casei
C-28·43 ± 0·07S. thermophilus
D-18·84 ± 0·08S. thermophilus
E-16·16 ± 1·01Gluconobacter oxidans*
E-26·10 ± 0·09Lactobacillus delbrueckii ssp. bulgaricus
E-37·59 ± 0·16S. thermophilus
F-17·23 ± 0·23L. delbrueckii ssp. bulgaricus
F-28·11 ± 0·12L. delbrueckii ssp. lactis
F-33·63 ± 0·30L. paracasei
F-48·55 ± 0·12S. thermophilus
G-17·36 ± 0·11L. delbrueckii ssp. bulgaricus
G-25·00 ± 0·01Bifidobacterium animalis ssp. lactis
G-36·31 ± 0·05S. thermophilus
Figure 1.

 PCR amplification of Bifidobacterium DNA from probiotic products. Products are named as A to G. Positives bands of approximately 217 bp were observed in products A, B, C, D and G. Positive controls correspond to Yoghurt Activia (Danone-Belgium), and to total DNA from Bifidobacterium lactis Bb12. Negative controls correspond to total DNA from Lactobacillus casei and MilliQ water.

Antagonistic activity of the LAB against Salmonella typhimurium strains SL1344 and ATCC 14028

The inhibition capacity of the isolated strains against S. typhimurium is summarized in Table 3. Strain F-3 is the only one exerting strong antimicrobial activity on both Salmonella strains tested. This is also the case for the control strain L. rhamnosus GG which has been previously reported to antagonize Salmonella spp. (Hudault et al. 1997; De Keersmaecker et al. 2006b). Strains A-1 and B-1 shared the same rather strong inhibitory pattern against the indicator strains. On the contrary, strains B-2, F-1, F-4 and G-2 did not inhibit the growth of the indicator strains. Interestingly, strain C-1 showed full inhibition against strain SL1344, but no inhibition against strain ATCC 14028.

Table 3.   Antimicrobial activity against the pathogen Salmonella typhimurium
Isolated strainsTest organism*pH†
ATCC 14028SL1344
  1. *Interpretation of zone diameter of inhibition: −, no visible inhibition; +, 1–6 mm, ++, 7–12 mm; +++, more than 12 mm.

  2. †pH of the spent-culture supernatant after 24 h of growth of the isolated strain.

A-1 Lactobacillus paracasei+++++4·08
A-2 Streptococcus thermophilus++4·40
B-1 Leuconostoc mesenteroides+++++4·18
B-2 S. thermophilus4·70
C-1 Lactobacillus casei+++3·90
C-2 S. thermophilus+4·67
D-1 S. thermophilus+++++3·86
E-1 Gluconobacter oxidans++4·70
E-2 Lactobacillus delbrueckii ssp. bulgaricus+3·99
E-3 S. thermophilus+4·50
F-1 L. delbrueckii ssp. bulgaricus4·00
F-2 L. delbrueckii ssp. lactis+3·80
F-3 L. paracasei++++++3·82
F-4 S. thermophilus4·75
G-1 L. delbrueckii ssp. bulgaricus+3·95
G-2 Bifidobacterium animalis ssp. lactis4·45
G-3 S. thermophilus+4·67
Lactobacillus rhamnosus GG (LGG). Control strain++++++3·81

Similar antagonistic patterns for each of the selected LAB strains were obtained when using 24-h SCS with pH indicated in Table 3. However, upon neutralization of the SCS, the inhibitory capacity of all tested SCS was lost (data not shown).

Comparative survival of isolated strains in simulated gastric juice

Testing the survival of the isolated strains in simulated gastric juice showed that of 11 of the 17 analysed strains, the viable bacteria declined to undetectable levels after only 30 min of exposure (Table 4). Only strains A-1 and B-1 were able to survive the gastric juice exposure over 90 min. Lower survival rates were observed for strains C-1, E-2, F-2 and F-3, with the counts of the last three strains undetectable after 60 min of incubation. Strikingly, while strain F-1 (L. delbrueckii ssp. bulgaricus) exhibited poor survival in this system, a second L. delbrueckii ssp. bulgaricus (E-2) was a better survivor.

Table 4.   Effects of simulated gastric juice on the survival of lactic acid bacteria (LAB) from commercial products
StrainsCounting log10 (CFU ml−1) at indicated time points (min)
  1. *Values representing counting in simulated gastric juice at time point 0 min. Controls in phosphate buffered saline (PBS) at 0 min showed in all cases a counting between 108 and 109 CFU ml−1 (not shown).

  2. Each value in the table represents the mean value ± standard deviation (SD) from three independent experiments.

A-18·93 ± 0·624·37 ± 0·312·21 ± 0·180·43 ± 0·06
A-27·27 ± 0·70000
B-19·13 ± 0·614·14 ± 0·152·90 ± 0·631·03 ± 0·31
B-27·04 ± 0·45000
C-18·50 ± 0·463·09 ± 0·361·12 ± 0·000
C-26·85 ± 0·45000
D-18·85 ± 0·03000
E-17·39 ± 0·02000
E-28·83 ± 0·893·32 ± 0·74 00
E-37·49 ± 0·45000
F-18·37 ± 0·84000
F-28·75 ± 0·002·56 ± 0·0000
F-38·69 ± 0·003·04 ± 0·0000
F-47·53 ± 0·70000
G-17·17 ± 0·50000
G-29·29 ± 0·75000
G-38·65 ± 0·15000
LGG9·69 ± 0·00 8·48 ± 0·008·31 ± 0·128·06 ± 0·07

Adherence capacity of the LAB to Caco-2 cells

The adhesion results showed that four of the 17 analysed strains (A-1, B-1, F-2 and F-3) were able to adhere to the Caco-2 cells with no significant differences in comparison with the control strain L. rhamnosus GG (Fig. 2), which is a well-adhering strain (Elo et al. 1991; Tuomola et al. 2001). On the other hand, six strains showed intermediate adhesion (B-2, C-2, D-1, E-3, F-1 and G-2), while seven strains showed similar level of adhesion as the negative control strain (Kimoto et al. 1999).

Figure 2.

 Adherence capacity of isolated strains to Caco-2 cells. Codes refer to the strains specified in Table 2. Values represent the mean count (log10 CFU ml−1) of three independent experiments. Adherence of G-2 was determined by real-time PCR. The error bars are standard deviations.


For commercial-, industrial- and health-related reasons, probiotic and starter cultures used during the manufacturing of bio-yoghurts must be functional in order to promote human health at the site of action (Sanders and Huis in’t Veld 1999; Saarela et al. 2000). Although recent progress in probiotic research has been accomplished, not all of the probiotic bacteria available in the market are well characterized, or have adequate technological performance (Ross et al. 2005). Additionally, recent reports have clearly highlighted the poor quality of many probiotic products in terms of their contents and label information (Hamilton-Miller et al. 1996, 1999; Yeung et al. 2002; Temmerman et al. 2003a,b). This study showed that this is also the case for some of the probiotic products available in the Columbian market. During recent years, commercialization of new probiotic products in the Columbian market has witnessed a strong increase. However, the label information of the purchased products, analysed in this study, was rather vague, and did not indicate the full scientific name of the probiotic micro-organism present in the product. Indeed, although in this study, the isolation procedure was carried out before the expiry date, and a broad range of selective media was used, the diversity of the different identified bacteria was lower than that claimed on the product labels. Only one of the seven evaluated products actually contained one bacterium stated on the label (C-1). Moreover, one product (E) mentioned on its label to contain ‘lactic ferment’. Surprisingly, besides the starter bacteria L. delbrueckii ssp. bulgaricus (E-2) and S. thermophilus (E-3), we unexpectedly isolated the gram-negative acetic acid bacterium Gluconobacter oxydans (E-1), which is mostly associated with vinegar production, sorbose fermentation, and wine and beer spoilage (Gupta et al. 2001). Moreover, the number of viable cells present in the product was not stated on the label. To confer any health effect, it is generally believed that probiotic bacteria should be present in a product to a minimum of ∼107 CFU ml−1 or g−1, and should remain viable until they reach the intestinal tract (Ross et al. 2005). This requires their survival in the food vehicle and their resistance to the acidic conditions of the stomach and bile salts in the small intestine. The low recovery in this study could be related to the presence of nonviable cells in the product as a consequence of either low bacterial tolerance to processing conditions, or antagonistic microbial interactions between the microbial populations in the products (Saarela et al. 2000). Nevertheless, although some probiotics can exert their beneficial health effects even if they are dead (Jijon et al. 2004), these findings are raising questions about the initial selection of the probiotics by the companies. There is a need for adequate microbial quality control before claims by the manufacturers of functional food are made. Although the overall findings in this study do not deny the beneficial effect of the product itself, the increasing consumer’s knowledge of the relation between diet and health creates the necessity of having the correct information of the strains present in a product at the time of consumption, and their corresponding benefits.

Different probiotic characteristics were determined in this study, such as antimicrobial capacity of the isolated strains (Table 3). Antagonistic properties of probiotic strains are essential in order to prevent the infection and/or invasion of pathogenic bacteria. In this preliminary screening of the antimicrobial capacity of the strains present in the products, two different strains of the gastrointestinal pathogen S. typhimurium were selected as indicator strains. Interestingly, the two Lactobacillus paracasei isolated (A-1 and F-3), Leuconostoc mesenteroides (B-1) and S. thermophilus (D-1) have strong antagonistic activity against both Salmonella strains. Unexpectedly, the only L. casei isolated (C-1) showed full inhibition against strain SL1344, but no inhibition against ATCC 14028. These differences in susceptibility could be explained by reported differences in gene expression of some membrane-related Salmonella proteins (Lee et al. 2000; Feng et al. 2003). Interestingly, the only Bifidobacterium spp. isolated (G-2) did not show any antagonist effect against the indicator strains, although some Bifidobacterium strains have been previously reported to antagonize different intestinal pathogens (Servin 2004). The results observed when using neutralized SCS suggested that the antagonistic activity exerted by the tested strains is related to the acidic nature of their SCS. We have previously reported that the strong antimicrobial activity exerted by L. rhamnosus GG against Salmonella is mediated by lactic acid (De Keersmaecker et al. 2006b). However, it is important to consider that the production of antimicrobial compounds, such as organic acids, can be dependent on medium composition. Moreover, the in vitro results cannot be merely extrapolated to in vivo situations without further testing.

The determination of the acid tolerance of starter and probiotic bacteria is important in order to predict strain performance during gastric transit. The results in the present study showed that the survival rates of the isolated bacteria were low in comparison with the high survival rate of L. rhamnosus GG. Bifidobacteria have been reported to be highly sensitive to acid exposure (pH 3–5 for 3 h) with the exception of B. lactis and B. animalis (Matsumoto et al. 2004). However, in our hands, B. animalis ssp. lactis did not survive to the acid challenge, and no viable cells were observed after 30-min incubation in simulated gastric juice. Interestingly, in this study, all S. thermophilus strains were highly sensitive towards the simulated gastric juice, although it has been reported that S. thermophilus survives the gastrointestinal transit (Mater et al. 2005). The reason for different abilities to tolerate acidity is uncertain. Overall, our results show that the survival of the cultures varies among species (e.g. A-1 and F-3), and those differences are also apparent at the strain level (e.g. F-1 and F-2), thereby confirming previous reports (Corcoran et al. 2005). However, it must be noted that the tolerance of strains in bile acid broth systems may not truly reflect their ability to tolerate bile in vivo. Like other physiological stresses, it is difficult to simulate exact in vivo conditions in a laboratory setting, and not always all parameters that affect survival can be taken into account.

Adhesion to intestinal surfaces and subsequent colonization of the human gastrointestinal tract is considered to be one of the main criteria for selecting probiotics for human use. The adherent strains of probiotic bacteria are expected to persist longer in the intestinal tract and thus have better possibilities of showing metabolic and immunomodulatory effects than nonadhering strains (Saarela et al. 2000). Adhesion may also provide competitive exclusion of pathogenic bacteria from the intestinal epithelium (Coconnier et al. 1993). Our results showed that almost 60% of the isolated strains were able to adhere to Caco-2 cells in comparison with the well-recognized adherent strain L. rhamnosus GG (Elo et al. 1991; Tuomola et al. 2001). Remarkably, although L. acidophilus JMC1132 has been reported as a nonadherent strain (Kimoto et al. 1999), in our test, the strain showed adhesion to Caco-2 cells, although less than L. rhamnosus GG. These differences might be explained by differences in the adhesion assays applied.

Until now, there are no reports in the literature describing either the strain content of Columbian dairy products, or the specific probiotic characteristics of the present strains. Table 5 shows a summary of the bacterial performance from the isolated strains. In this study, four of the tested strains (A-1, B-1, F-2 and F-3) showed, in general, satisfactory in vitro probiotic characteristics. Nonetheless, F-3 was present in the product at a number lower than the one suggested being necessary to confer health benefit effects. Additionally, it became clear that probiotic behaviour varies among species and strain level, as exemplified by the seven different S. thermophilus isolated.

Table 5.   Summary of bacterial performance from the isolated strains
Product and species identifiedAntimicrobial activityGastric juice toleranceAdhesion to Caco-2 cells
  1. Interpretation of symbols used:

  2. Antimicrobial test: +, inhibition of either one or both indicator strains; −, no inhibition observed.

  3. Survival in simulated gastric juice: +, survival at least during the first 30 min; −, absence of viable cells after 30 min.

  4. Adherence to Caco-2 cells: +, more adherence observed in comparison with the negative control; −, less or similar adherence observed in comparison with the negative control.

A-1 Lactobacillus paracasei+++
A-2 Streptococcus thermophilus+
B-1 Leuconostoc mesenteroides+++
B-2 S. thermophilus+
C-1 Lactobacillus casei++
C-2 S. thermophilus++
D-1 S. thermophilus++
E-1 Gluconobacter oxidans+
E-2 L. delbrueckii ssp. bulgaricus++
E-3 S. thermophilus++
F-1 L. delbrueckii ssp. bulgaricus+
F-2 L. delbrueckii ssp. lactis+++
F-3 L. paracasei+++
F-4 Streptococcus thermophilus
G-1 L. delbrueckii ssp. bulgaricus+
G-2 Bifidobacterium animalis ssp. lactis+
G-3 S. thermophilus+

This study aimed to determine the microbial composition and strain characteristics present in probiotic dairy products available in the Columbian market. Although molecular techniques are highly sensitive and specific for the detection of bacterial DNA, classical culture techniques remain essential for the assessment of cell viability. This application could, therefore, serve as a valuable tool for the microbial quality control of probiotic preparations in order to provide reliable and safe products to the Columbian consumers. According to Euromonitor International (Chicago-USA), who speculated on growing competition in the probiotics market, the Latin America dairy industry is set to increase nearly 11% between 2002 and 2007 (


During this work, M. Perea Vélez was financially supported by a PhD grant (STWW-Project: IWT 000162) from the Flemish Institute for the Promotion of Innovation by Science and Technology (IWT-Vlaanderen, Brussels, Belgium). We thank Ir. Yasmine Delaedt for providing us with primers/probe and necessary equipment for the real-time PCR analysis, and Dr Ellen Somers, Ir. Maarten Fauvart and Mrs Adiela Vélez for their valuable graphical and technical assistance.