28Bifidobacterium animalis subsp. lactis IPLA R1 and Bifidobacterium longum IPLA 29 E44 strains were tested in vivo for their safety and ability to modulate the intestinal 30 microbiota. Chemically simulated gastrointestinal digestion showed considerably lower 31 survival of E44 than R1 strain, the first microorganism also being more sensitive to 32 refrigerated storage in 10%-skimmed milk at 4ºC. Harmful glycosidic activities were absent, 33 or at low levels, in the strains R1 and E44. Both strains were sensitive to most antibiotics and 34 resistant to aminoglycosides, a common feature in bifidobacteria. Similarly to several other 35 bifidobacteria strains, B. animalis subsp. lactis IPLA R1 displayed a moderate resistance 36 against tetracycline which correlated with the presence of tet(W) gene in its genome. The 37 general parameters indicating well-being status, as well as translocation to different organs 38 and histological examination of the gut tissues, revealed no changes induced by the 39 administration of bifidobacteria. 12 week-old male Wistar rats were distributed into three 40 groups, eight rats in each. Two groups were administered daily over 10 8 cfu of the 41 corresponding strain suspended in 10%-skimmed milk for 24 days, whereas rats in the 42 placebo group received skimmed milk without microorganisms added. The microbiota and 43 short chain fatty acids (SCFA) were monitored in faeces at different time points during 44 treatment and in caecum-content at the end of the assay. Quantitative PCR (qPCR) showed 45 that faecal and caecal Bifidobacterium levels were higher in bifidobacteria-fed rats than in the 46 placebo rats at the end of the intervention, whereas total anaerobic-plate counts did not show 47 significant differences. Quantification of B. animalis and B. longum by qPCR showed that, 48 independent of the microorganism administered, treatment with bifidobacteria resulted in 49 higher levels of B. animalis in the caecum. PCR-DGGE analysis of microbial populations 50 revealed a higher diversity of bands in caecum-content of rats fed B. animalis IPLA R1 than 51 in the placebo group and rats fed B. longum IPLA E44. Remarkably, although no variations in 52 3 the proportion of acetate, propionate and butyrate were found, at the end of the assay the total 53 SCFA concentration in the faeces of rats fed bifidobacteria was significantly higher and those 54 in caecum-content significantly lower, than that of the placebo group. This suggests a 55 displacement of the SCFA production to parts of the colon beyond the caecum in rats 56 receiving bifidobacteria. Therefore, the oral administration of B. animalis IPLA R1 and B. 57 longum E44 can be considered safe, these microorganisms having the ability to modulate the 58 intestinal microbiota of rats by influencing SCFA and the bifidobacterial population levels.