Determining the DNA sequence elements required for binding integration host factor to two different target sites

Binding sites for the Escherichia coli protein integration host factor (IHF) include a set of conserved bases that can be summarized by the consensus sequence WATCAANNNNTTR (W is dA or dT, R is dA or dG, and N is any nucleotide). However, additional 5'-proximal bases, whose common feature is a high dA+dT content, are also thought to be required for binding at some sites. We examine the relative contribution of these two sequence elements to IHF binding to the H' and H1 sites in attP of bacteriophage lambda by using the bacteriophage P22-based challenge-phage system. IHF was unable to act as a repressor in the challenge-phage assay at H' sites containing the core consensus element but lacking the dA+dT-rich element. This indicates that both elements are required for IHF to bind to the H' site. In contrast, the core consensus determinant alone is sufficient for IHF binding to the H1 site, which lacks an upstream dA+dT-rich region. Fifty mutants that decreased or eliminated IHF binding to the H1 site were isolated. Sequence analysis showed changes in the bases in the core consensus element only, further indicating that this determinant is sufficient for IHF binding to the H1 site. We found that placement of a dA+dT-rich element upstream of the H1 core consensus element significantly increased the affinity, suggesting that the presence of a dA+dT-rich element enhances IHF binding.