Genome organisation and chromatin structure in Escherichia coli

We have analysed the complete sequence of the Escherichia coli K12 isolate MG1655 genome for chromatin-associated protein binding sites, and compared the predicted location of predicted sites with experimental expression data from 'DNA chip' experiments. Of the dozen proteins associated with chromat in in E. coli, only three have been shown to have significant binding prefe rences: integration host factor (IHF) has the strongest binding site prefer ence, and FIS sites show a weak consensus, and there is no clear consensus site for binding of the H-NS protein. Using hidden Markov models (HMMs), we predict the location of 608 IHF sites, scattered throughout the genome. A subset of the IHF sites associated with repeats tends to be clustered aroun d the origin of replication. We estimate there could be roughly 6000 FIS si tes in E. coli, and the sites tend to be localised in two regions flanking the replication termini. We also show that the regions upstream of genes re gulated by H-NS are more curved and have a higher AT content than regions u pstream of other genes. These regions in general would also be localised ne ar the replication terminus. (C) 2001 Societe francaise de biochimie et bio logie moleculaire / Editions scientifiques et medicales Elsevier SAS.