The samples were homogenized and sub-samples were diluted in phosphate buffered saline for plating on selective media (MacConkey agar)

supplemented with 100 μg ml-1 streptomycin sulfate. The lower limit of detection in fecal plate counts was 102 CFU (g feces)-1 for 100 μl of the diluted solution per plate. The remaining samples were stored at -80°C. Colony forming units (CFUs) were monitored per gram feces. Phenotypic determination Crude colicin lysates were prepared according to the procedure of Suit et al [42] and stored at 4°C see more until use. Twenty colonies of streptomycin-resistant E. coli from fecal pellets obtained from each mouse at four-week intervals were assayed for the production of growth inhibition zones on plates pre-inoculated with a sensitive lawn (E. coli strain BZB1011). Confirmation of the identity of the colicin produced was provided

by a strain’s ability to grow in the presence of its own colicins (100 μl of crude colicin lysate spread onto LB plates), due to the immunity protein it produces. The zones of inhibition of each strain were documented using an imaging and documentation system (Bio-Rad, Hercules, CA). Statistical analysis Each cage was treated as an independent sample and an average of the two co-caged mice was determined. The average number of CFUs per cage was compared at two times, 0 and 112 days, using a CH5424802 nmr one-way ANOVA. In addition, for each of these times we employed two orthogonal contrasts to test for differences in CFUs among groups of strains that were chosen a priori. One contrast served to compare the average number of CFUs of the colicin-free strain with that of the other (colicinogenic) strains. The second served to compare the average

number of CFUs of the colicinogenic strains. A repeated-measure ANOVA was conducted to test for differences in the persistence of the various strains over time treating strain as a between-subject factor and time as a within-subject factor. The effects of strain type and time (i.e. beginning vs. end of the experiment) on strain doubling time were tested with a two-way ANOVA with both strain and time treated as fixed factors. All statistical analyses were done with the STATISTICA 2007 (StatSoft, Tulsa, OK). Acknowledgements This work was supported by National Institutes of Health grants R01GM068657-01A2 and R01A1064588-01A2 PtdIns(3,4)P2 to M.A. Riley. References 1. Gorbach S, Bartlett JG, Blacklow NR: Infectious Diseases. Philadelphia: Lippincott, Williams, and Wilkins 2003. 2. Guarner F: Enteric flora in health and disease. Digestion 2006,73(Suppl 1):5–12.PubMedCrossRef 3. Altenhoefer A, Oswald S, Sonnenborn U, Enders C, Schulze J, Hacker J, Oelschlaeger TA: The probiotic Escherichia coli strain Nissle 1917 interferes with invasion of human intestinal epithelial cells by different enteroinvasive bacterial pathogens. FEMS Immunol Med Microbiol 2004, 40:223–229.PubMedCrossRef 4.