Dietary influence of kefir on microbial activities in the mouse bowel


Domingo Marquina Department of Microbiology III, Biology Faculty, Complutense University of Madrid, 28040 Madrid, Spain (e-mail:


Aims: In this work the microflora present in kefir, a fermented milk product, was studied together with the effect of kefir administration on different groups of indigenous bacteria of mouse bowel.

Methods and Results: Kefir microflora was composed of lactic acid bacteria, acetic acid bacteria and yeasts. Yeast population was composed of Saccharomyces cerevisiae, S. unisporus, Candida kefir, Kluyveromyces marxianus and K. lactis. The streptococci levels in kefir treated mice increased by 10-fold and the levels of sulfite-reducing clostridia decreased by 100-fold. The number of lactic acid bacteria increased significantly.

Conclusions: The administration of kefir significantly increased the lactic acid bacteria counts in the mucosa of the bowel. Ingestion of kefir specifically lowered microbial populations of Enterobacteriaceae and clostridia.

Significance and Impact of the Study: This is the first long-term study about the effects of the kefir administration on the intestinal microflora of mice.


The term probiotic refers to live microorganisms that reside in the gastrointestinal tract and have beneficial effects on the host (De Simone et al. 1991; Lee and Salminen 1995). More than four hundred species of bacteria are estimated to survive in the gastrointestinal tract and these endogenous bacteria comprise the intestinal microflora. Even more species may be present but not culturable by the traditional plate count methods. After passage through the stomach and the small intestine, some probiotics survive and become established transiently in the large bowel. In order to survive in and colonize the gastrointestinal tract, probiotic bacteria need to express high tolerance to acid and bile and to have the ability to adhere to intestinal surfaces (Lee and Salminen 1995; Kirjavainen et al. 1998; Fujiwara et al. 2001).

Possible health effects of probiotics include immune system stimulation, alleviation of lactose intolerance, hypocholesterolaemic effect, and prevention of cancer recurrence (De Simone et al. 1991; Reddy and Rivenson 1993; Lee and Salminen 1995). Probiotics have been shown to possess inhibitory activity toward the growth of pathogenic bacteria such as Escherichia coli, Listeria monocytogenes, Salmonella spp., and others (Harris et al. 1989; Ashenafi 1991; Chateau et al. 1993; Letellier et al. 2001). This inhibition is supposedly due to the production of inhibitory compounds such as bacteriocins, hydrogen peroxide, organic acids and competitive adhesion to the epithelium.

Kefir is a fermented milk which originated in the Caucasus countries. This acidic fermented milk is slightly carbonated and has the presence of small amounts of alcohol. Kefir distinguishes itself from the traditional fermented milks (yogurt) in that it is made only from kefir grains, which are composed of bacteria and yeasts (Marshall and Cole 1985). Kefir beverage contains live microflora from the grains, which are removed by filtration after the fermentation process. Kefir grains are groups of lactic acid bacteria (lactobacilli, lactococci and Leuconostoc), acetic acid bacteria (Acetobacter aceti) and yeasts (Saccharomyces cerevisiae, Candida kefir, Kluyveromyces marxianus) held together by a matrix of different exopolysaccharides (kefiran).

The microbial populations that inhabit the intestine of mice have been extensively studied over the recent years (Tannock 1979; Tannock et al. 1988). The mouse gastrointestinal tract therefore provides an excellent model to study the effect of different types of diet on this ecosystem. Mice can be maintained under controlled conditions, and different parts of the gastrointestinal tract can be sampled after killing the animals thus avoiding the need to rely on data obtained from faecal examination. It is apparent that the nature of the diet and of the experimental animal must be considered in future studies of the intestinal microbiota of humans. The extrapolation of data from mice to humans must be avoided, although in humans development of the gastrointestinal epithelium follows a developmental programme similar to that observed in rodents (Montgomery et al. 1999).

The aims of the present work were to study with bowel samples the influence of the kefir administration on the microbial ecology of the gastrointestinal tract of mice and to evaluate the changes on these microbial populations.


Animals and preparation of diets

Six-month-old female Swiss mice were used after a 1-month quarantine period. Animals were housed four per cage, and cages were changed twice a week. Water was available continuously from bottles.

The normal diet chow contains corn and wheat flakes, soybean meal, ground corn, fish meal, brewers' dried yeast, soybean oil and animal liver meal.

The kefir diet was prepared as follows. A 140-g amount of kefir was mixed with 200 g of the normal diet chow. The mixture was divided into portions and then lyophilized and stored under sterile conditions.

The kefir and normal diet feedings were started when the animals were seven months old. A total of 20 females were employed in each dietary group.


All animals were housed in a building that was specially equipped for animal experiments. Animal caretakers, technicians, or anyone entering the animal holding area had to dress in a special suit, including cap, face mask, and shoe covers, to minimize the possibility of contamination. For examination of organs and sampling, mice were killed by carbon dioxide anaesthesia followed by cervical dislocation. Samples were taken three months and seven months after the start of the diet treatment.

Homogenates were prepared from lengths (5 cm) of small intestine and large intestine. These tissues were homogenized in brain heart infusion (BHI) broth by using a Teflon homogenizer. The resulting homogenates were maintained in a reduced condition in GasPak jars (BBL Microbiology Systems, Cockeysville, MD) and then, they were diluted (10-fold to 1010) in prerreduced BHI.

Enumeration of gastrointestinal microbes

Microbial counts were obtained by culturing homogenates of gastrointestinal organs on selective media as reported previously (Roach and Tannock 1979; Tannock 1979). Viable counts were obtained from brain heart infusion (BHI) agar that was incubated aerobically and anaerobically at 37 °C for 3 d (aerobic and anaerobic counts, respectively). Lactic acid bacteria counts were obtained from MRS agar at pH 5·5 (Oxoid) that was incubated aerobically at 37 °C for 5 d, from phenylethanol agar (Difco) that was incubated aerobically and anaerobically at 37 °C for 3 d (Gram-positive bacterial counts), from Azide Blood Agar (Oxoid) that was incubated aerobically and anaerobically at 37 °C for 3 d (mainly streptococcal counts), from Kanamycin Esculin Azide Agar agar (Oxoid) that was incubated aerobically at 37 °C for 2 d (Enterococcus counts), from SPS agar that was incubated anaerobically at 37 °C for 3 d (sulfite-reducing clostridial counts), from Eosin Methylene blue agar (Oxoid) that was incubated aerobically at 37 °C for 1 d (Enterobacteriaceae counts), and from BHI agar containing a Gram-negative anaerobic supplement (Oxoid) that was incubated anaerobically at 37 °C for 3 d (Gram-negative anaerobic bacterial counts). Anaerobic incubations were performed in Gaspak jars.

Yeasts were isolated in Yeast Morphology Agar (YMA) at 30 °C for 3 d (Santos et al. 2000).

Enumeration of kefir microbes

Microbial counts were enumerated by culturing homogenates of kefir samples on selective media as described before. Acetic acid bacteria were determined in Yeast Glucose Agar for Acetobacter containing: glucose 100 g l–1, yeast extract 10 g l–1, CaCO3 20 g l–1, agar 20 g l–1 and cycloheximide 1000 p.p.m. Appropriate dilutions were made in sterile distilled water.

Bacterial and yeast identification

The strains of Lactobacillus, isolated from MRS, were identified by using the API-50CH system. The yeasts, isolated from YMA–Rose Bengal–chloramphenicol (Marquina et al. 1992), were identified according to the classical methods described previously (Van der Walt and Yarrow 1984).


Previous reports of studies of the normal microbiota of mice and rats show that its composition, in general terms, is universal (Tannock et al. 1988). This probably reflects the similarities in anatomy and physiology of the gastrointestinal tracts of these rodents. A normal microbiota is detectable along the entire length of the gastrointestinal tract, but specific types of microorganisms are present in each region. The numerically predominant species are obligately anaerobic bacteria but a high number of microaerophilic and facultatively anaerobic bacteria are also present. The number of species and the population of bacteria are relatively stable in healthy adults. In addition to the normal microflora present in the gastrointestinal tract, different microorganisms are introduced with food. This transient microflora has an important effect on the upper gastrointestinal tract. This effect seems to be less important in the lower tract. In spite of its stability, the intestinal flora can be affected by a great variety of internal (gastric acid, bile salts) and external factors (diet, infections, ageing, antibiotics).

When the microbial content of the kefir was studied the most important group of bacteria was the lactic acid bacteria (Table 1). The lactobacilli present in kefir, isolated in MRS medium, were identified using API-50CH system. Lactococcus brevis and L. kefir were present in the analysed kefir samples but isolates of L. paracasei, L. plantarum, L. acidophilus and L. kefiranofaciens were found. Between the yeast isolates the nonlactose fermenting yeasts were predominant (Saccharomyces cerevisiae, S. unisporus), although some different lactose fermenting yeasts were identified (Kluyveromyces marxianus, K. lactis and Candida kefir). As in yogurt, the lactose content is reduced in kefir and the β-galactosidase level is increased as a result of fermentation. These yeasts contribute to the low lactose content of the final kefir product.

Table 1.  Microbial counts in kefir samples Thumbnail image of

From our results, kefir diet had a significant influence on the composition of the microbial flora. The most impressive differences were with clostridia and lactic acid bacteria. In this study, we demonstrated that administration of kefir could also increase the levels of lactic acid bacteria in the small bowel and large bowel (Fig. 1). Other workers have also showed that Lactobacillus administration can decrease the numbers of faecal Escherichia coli and anaerobic cocci (Lidbeck et al. 1987). In our study, the number of members of the Enterobacteriaceae on the large bowel mucosa slightly decreased. However, the anaerobic bacterial counts and the counts of Gram-negative anaerobic bacteria in the large bowel were significantly decreased and this was more pronounced in the 7-months-treated mice. Sulfite-reducing clostridia also decreased significantly. This suggests that after a period of establishment, the kefir microflora exercised an antagonistic effect against the anaerobic microflora. From a medical perspective, the inhibition of anaerobic microflora is of interest. Several studies have shown that Gram-negative anaerobic bacteria are frequently isolated from infected sites in patients with postoperative intra-abdominal septic complications (Nichols 1980).

Figure 1.

Microbial counts in small and large bowel at two sampling times. (i) Small bowel, 3 months; (ii) small bowel, 7 months; (iii) large bowel, 3 months; (iv) large bowel, 7 months. A, Aerobic bacteria; B, anaerobic bacteria; C, Gram-positive aerobic bacteria; D, Gram-positive anaerobic bacteria; E, Gram-negative anaerobic bacteria; F, aerobic bacteria on Azide Blood Agar; G, anaerobic bacteria on Azide Blood Agar; H, lactic acid bacteria; I, Enterobacteriaceae; J, Enterococcus spp.; K, sulphite-reducing bacteria; L, yeasts

The augmented proportion of enterococci observed in the small bowel during the treatment with the kefir diet was very important too. The number of enterococci increased 10-fold with kefir diet. Enterococci occur in similar environmental locations to Listeria spp. According to the results described by McKay (1990), Enterococcus faecium has an antimicrobial activity against a wide range of Listeria spp., although the significance of this interaction may be limited by the absence of nutrients, such as peptides, which appear essential for antimicrobial production.

Dietary effects on the composition of intestinal microflora may lend support to the observed phenomenon that the colonic contents of rats on a meat diet have greater infectivity in inducing the intra-abdominal abscesses as compared with the colonic contents of rats on a conventional chow diet (Weinstein et al. 1974). Previous reports have shown that in human a low-liquid diet or a meat diet did not strongly influence the composition of the human faecal microflora (Maier et al. 1974). So, dietary variations may have little influence on the microbial population in the human large intestine. On the other hand, the fact that the animals were placed on dietary controls at a young age may be critical in microflora colonization in the gastrointestinal tract. The present work reveals that kefir diet has a pronounced effect on the microbial composition of the mouse bowel. Kefir possesses antimicrobial activity in vitro against a wide variety of Gram-positive and Gram-negative bacteria (Cevikbas et al. 1994; Zacconi et al. 1995), and against fungi (Cevikbas et al. 1994). The effect of the diet consumed by the animal host, of course, may not be the result of the direct influence of dietary components on gastrointestinal microbes. The diet obviously affects the overall condition of the animal, and any changes produced in the physiology of the animal host will indirectly influence the microbes in the gastrointestinal ecosystem. It is apparent that the nature of the diet must be considered carefully in future ecologcal studies of the gastrointestinal tract.


This research was supported by the UCM-DANONE project PR 238/00-9455.