(A) Host genotype influences microbial community structure. Principal coordinates analysis of unweighted UniFrac distances from 2-10 week TRUC (n= 3) and Rag2^−/− (n =3) mice and their mothers. Abbreviations: A, B, C, individual pups colored by genotype, followed over time (A.2, A.4, A.6, A.8, and A.10 refers to animal A sampled at 2,4,6, and 10 weeks of age). (B) The distribution of order-level phylotypes in TRUC and Rag2^−/− fecal microbial communities. Percent relative abundance is plotted for each age group.

(A) The antibiotic sensitivities of colitis transmitted via TRUC cross-fostering are the same as spontaneous TRUC colitis. Antibiotic-treated pregnant TRUC females were used as foster mothers and treated with antibiotics in their water until weaning. Histologic colitis scores are shown for the fostered mice at 8 weeks of age. (B) Klebsiella pneumoniae (squares) and Proteus mirabilis (circles) are detected in the fecal microbiota of TRUC cross-fostered Rag2^−/− and WT mice at 8 weeks of age but not in 8 week old TRUC mice fostered by Rag2^−/− or WT mice. TRUC-fostered TRUC, Rag2^−/−-fostered Rag2^−/−, and WT-fostered WT are shown as controls. Limits of detection; 10^4.4 cfu/g dry weight of feces. Each filled square or circle represents a fecal sample from a different animal. (C) Left panel: Fecal bacterial counts for co-colonized gnotobiotic TRUC mice. Mean values ±1 S.D. are shown for Klebsiella pneumoniae (squares) and Proteus mirabilis (circles) (n=5 mice). Right panel: Histologic colitis scores of germ-free TRUC and germ-free TRUC mice co-colonized with Klebsiella pneumoniae and Proteus mirabilis from the TRUC mother in Fig. 1. (D) Left panel: Klebsiellapneumoniae and Proteus mirabilis fecal cfu in Rag2^−/− and WT mice treated every other day from 2-10 weeks of age with 10⁷ cfu of E. coli, Proteus mirabilis, Klebsiella pneumoniae, or a combination of both added to their drinking water (all strains isolated from the TRUC mother in Fig. 1). Right panel: Histologic scores for colitis as assayed at sacrifice at 10 weeks of age. Each filled circle represents a separate animal in the treatment group. p-values shown were calculated using the Mann-Whitney test.

(A) Culture-based identification of bacterial species present in fecal samples from the same animals and time points as analyzed in Fig. 1. Species observed in more than one animal or in one animal at more than one time point are shown. Inset is a summary of bacterial species-level differences in the fecal microbiota of TRUC versus Rag2−/− mice observed in 2A. (B) Upper panel: Histologic colitis scores demonstrate the in vivo antibiotic sensitivities of TRUC colitis. Each dot represents an individual mouse treated for four weeks with the indicated antibiotics dissolved in drinking water. VMNA refers to treatment with a combination of vancomycin, metronidazole, neomycin, and ampicillin. Horizontal bars represent the mean. p-value ≤0.0001 defined by Mann-Whitney test. Lower panel: Summary of in vitro antibiotic sensitivities for several species selectively detected in TRUC fecal microbiota. (C) Culture-based survey of Gram-negative aerobes present in fecal samples from TRUC (shaded circles) and Rag2^−/− (open circles) at 2-20 weeks of age. Fecal samples were collected and cultured twice at each time point from Rag2^−/− mice. (D) In vivo sensitivity of Klebsiella pneumoniae (squares) and Proteus mirabilis (circles), as defined by culture-base surveys of TRUC fecal samples collected 1d before (shaded symbol) and 1d after (open symbol) treatment with antibiotic. Each dot represents data from a fecal sample obtained from one mouse. Horizontal bars represent the mean value. (E) Fluorescence In Situ Hybridization using an Oregon-Green® 488 conjugated “Universal bacterial” 16S rRNA-directed oligonucleotide probe (EUB 338) demonstrates the presence of bacteria in the mucus layer and directly adjacent to the epithelium in TRUC mice. Upper panels: TRUC, lower panels: Rag2^−/−. A 10 micron scale bar for the panel is shown in the lower left of the first image. (F) Enterobacteriaceae (red), Klebsiella (red), and Proteus (red) were visualized adjacent to the epithelium in TRUC mice using fluor-conjugated 16S rRNA or 23S rRNA oligonucleotide probes [(pB-00914 (Enterobacteriaceae), pB-00352 (Klebsiella pneumoniae), pB-02110 (Proteus mirabilis)]. Sections were also hybridized with the EUB338 universal bacterial probe (green). Scale bars (10 micron) are shown for each image. (G) Enterobacteriaceae (red), Klebsiella (red), and Proteus (red) probe signals are seen adjacent to or along the epithelium in TRUC mice. Epithelial cell nuclei were stained with DAPI. White star symbols mark bacteria in (F) and (G). Scale bars (10 micron) are shown for each image.

(A) Successful immunotherapy by TNF-α blockade alters levels of culturable Enterobacteriaceae in the feces. TRUC mice were treated with curative anti-TNF-α (15mg/kg every week) (open circles) or isotype control (shaded circles) for four weeks, after which time therapy was stopped. Enterobacteriaceae levels were defined by culture of fecal samples obtained 1 day before, during, and after treatment (up to 14 weeks of age). Circles represent the mean value of anti-TNF-α mice (n=10) and isotype controls (n=10). Error bars represent the standard deviation. (B) Successful immunotherapy by T-reg infusion does not produce statistically significant differences in the levels of culturable Enterobacteriaceae species compared to vehicle-treated controls. TRUC mice were injected once with 75,000 CD4⁺CD62L^hiCD25⁺ T-regulatory cells (n=10) or PBS (n=9). (C) TNF-α production from T-bet^−/− Rag2^−/− MyD88^−/− bone marrow derived DCs co-cultured with heat-killed and live K. pneumoniae and P. mirabilis strains. Bars represent the mean value of triplicate determinations/sample. Error bars are 1 S.D. Data are representative of three independent experiments.