Green fluorescent protein for detection of the probiotic microorganism Escherichia coli strain Nissle 1917 (EcN) in vivo.

Probiotic microorganisms are defined as viable nutritional agents conferring benefit to the health of the human host. Especially, Escherichia coli strain Nissle 1917 (EcN) was shown to be equally effective as mesalazine in the maintenance of remission in ulcerative colitis (UC). Presumably, the therapeutic effect of EcN is linked to the presence of the strain in the region of interest; however, it remains difficult to follow the orally administered strain on its passage through the complex microbial environment of the intestine in vivo, inhabited dominantly by various E. coli strains, using traditional culturing methods. In this study we transformed EcN and a wild-type E. coli from a laboratory rat (EcR) with a plasmid carrying a gfp gene (pUC-gfp) to obtain EcN- and EcR-GFP to allow in vivo detection without alteration of strain-specific characteristics. Analysis of different strain-specific characteristics included the measurement of stimulation of IL-8 secretion and adhesion in vitro using the epithelial cell line HT-29. The kinetics of intestinal distribution in mice and colonization properties in rats following oral administration was studied in vivo. Detectability of the strain in histologic specimens was analysed using fluorescence microscopy and immunohistochemistry. The identity of fluorescent E. coli strains isolated from stool samples, Peyer's patches (PP) and mesenteric lymph nodes (MLN) was determined by REP-PCR. We were able to demonstrate that EcN and EcN-GFP do not differ in stimulation of IL-8 secretion or adhesion to HT-29 cells. In vivo, EcN-GFP colonies were readily detectable by fluorescence microscopy in luminal samples and also by immunohistochemistry in histological sections allowing analysis of the kinetics of the intestinal passage following oral administration. Translocation of fluorescent and non-fluorescent bacteria into PP and MLN was noted at 6 h post oral administration. EcN-GFP was detectable initially for 14 days in faecal samples of rats, while EcR-GFP was detectable throughout the whole experiment (45 days). Challenge with ampicillin at day 45 demonstrated continuing presence of EcN-GFP in small numbers by reappearing fluorescent colonies. The plasmid was not stable in vivo since non-fluorescent EcN colonies were detected also in faecal samples by REP-PCR. In summary, transformation of EcN to obtain EcN-GFP in our study had no detectable influence on the probiotic microorganism regarding adhesion on and induction of IL-8 secretion of HT-29 cells and allows the detection in mixed microbial environments in vivo but the stability of EcN-GFP in vivo is limited.