Guide to Genome-Wide Bacterial Transcription Factor Binding Site Prediction Using OmpR as Model

Abstract

Gene expression regulation in a cell plays a crucial role in the cellular response to environmental cues and other important biological processes (Bauer et al., 2010). A major mechanism of gene expression regulation is the binding of transcription factor (TF) protein to a specific DNA sequence in the regulatory region of a gene, thereby activating or inhibiting its transcription (Zhou & Liu, 2004). A TF often regulates multiple genes whose binding sites have similar but not identical sequences (Zhang et al., 2009). There is, however, a short, recurring pattern among the promoter sequences called a motif, and it is this motif that a TF recognizes and interacts with (D’haeseleer, 2006b). It is important to identify the set of genes a TF modulates, called its regulon, as this will advance our understanding of the regulatory network of an organism (D’haeseleer, 2006b; Tan et al., 2005). One way to identify the regulon is to determine a TF’s motif and subsequently use the motif to search for other candidate genes regulated by the TF. Traditionally, TF binding sites (TFBSs) are determined by various experimental approaches. Mutagenesis, DNase footprinting, gel-shift, and reporter construct assays are common methods for identifying the binding sites upstream of individual genes, but the throughput of these techniques is low (D’haeseleer, 2006b; Ladunga, 2010). In recent years, chromatin immunoprecipitation (ChIP) and systematic evolution of ligands by exponential enrichment (SELEX) are available to study protein-DNA interactions in a high throughput manner. Chromatin-immunoprecipitation of DNA cross-linked to a TF can be hybridized to a microarray (ChIP-chip) or sequenced (ChIP-seq) to obtain the TF’s cognate binding sites on the whole genomic scale (Homann and Johnson, 2010; Ladunga, 2010; Stormo, 2010). SELEX is an in vitro technique that measures the binding affinities of TFs for synthetic, randomly generated oligonucleotides, usually 10-30 bp long (D’haeseleer, 2006b; Ladunga, 2010; Stormo, 2010). Sequences that strongly bind to a TF in question will be selectively amplified for later identification (Schug, 2008). The major drawback of experimental approaches to determine TF recognition motifs is the time required and the relative high cost (Zhou & Liu, 2004). Moreover, some methods have specific requirements. For example, ChIP requires antibodies and certain growth conditions under which the transcription regulator is active (Tan et al., 2005). Even if a biologist can satisfy the requirements, the resolution of the regions containing the binding sites can span