Uropathogenic Escherichia coli forms biofilm aggregates under iron restriction that disperse upon the supply of iron.

The transition between biofilm and planktonic cells has important consequences during infection. As a model system, we have investigated uropathogenic Escherichia coli (UPEC) strain 536, which forms large biofilm aggregates when grown in iron-restricted tissue culture media. The provision of both inorganic and physiological iron to the media induces dispersal. Aggregates do not disperse upon the addition of exogenous iron when cells are pretreated with either rifampicin or chloramphenicol as inhibitors of transcription or translation, respectively. Aggregates stain with the cellulose stain Calcofluor White, can be prevented by the addition of cellulase to the growth media, and aggregates are broken down in the absence of exogenous iron when cellulase is added. An extension of this study to 12 UPEC clinical isolates identified seven that form cellulose aggregates under iron restriction, and that disperse upon the provision of iron. Consequently, we hypothesize that iron restriction stimulates the formation of cellulose aggregates, which disperse as a result of new gene expression in response to the provision of iron.

Analysis of the CoIE1 stability determinant Rcd.

Multimer formation is an important cause of instability for many multicopy plasmids. Plasmid CoIE1 is maintained stably because multimers are converted to monomers by Xer-mediated site-specific recombination at the cer site. However, multimer resolution is not the whole story; inactivation of a promoter (Pcer) within cer causes plasmid instability even though recombination is unaffected. The promoter directs the synthesis of a short transcript (Rcd) which is proposed to delay the division of multimer-containing cells. Mapping of the 5' terminus of Rcd confirms that transcription initiates from Pcer. The 3' terminus shows considerable heterogeneity, consistent with a primary transcript of 95 nt being degraded via intermediates of 79 and 70 nt. Secondary structure predictions for Rcd are presented. Of four mutations which abolish Rcd-mediated growth inhibition, one reduces the activity of Pcer while the other three map to the rcd coding sequence and reduce the steady-state level of the transcript. RNA folding analysis suggests that these three mutant transcripts adopt a common secondary structure in which the major stem-loop differs from that of wild-type Rcd. A survey of 24 cer-like multimer resolution sites revealed six which contain Pcer-like sequences. The putative transcripts from these sites have similar predicted secondary structures to Rcd and contain a highly conserved 15 base sequence. To test the hypothesis that Rcd acts as an anti-sense RNA, interacting with its target gene(s) through the 15 nt sequence, we used DNA hybridization and sequence analysis to find matches to this sequence in the Escherichia coli chromosome. Our failure to find plausible anti-sense targets has led to the suggestion that Rcd may interact directly with a protein target.