REPSA: general combinatorial approach for identifying preferred ligand-DNA binding sequences.

Most DNA-binding ligands, ranging from protein transcription factors to small molecule antineoplastic agents, recognize duplex DNA with some degree of sequence specificity. Determining this binding specificity is important for biochemists, molecular biologists, and medicinal chemists. Some information can be obtained through the study of defined DNA sequences, but a full picture of a ligand's binding specificity can only be obtained through combinatorial means, whereby vast libraries of sequences are screened. Several combinatorial methods have been developed for the study of ligand-DNA interactions, all of which require the physical separation of ligand-bound DNA from uncomplexed DNA before amplification by PCR. Here, we describe the novel combinatorial method Restriction Endonuclease Protection Selection and Amplification (REPSA). REPSA selects for ligand-bound DNAs through their inhibition of an enzymatic process-cleavage by a type IIS restriction endonuclease-which inactivates templates for subsequent PCR amplification. We have used REPSA to identify the preferred binding sites of oligonucleotides, proteins, and small molecules on duplex DNA. Unlike conventional combinatorial methods, REPSA is amenable to the study of mixtures of native ligands with relatively unknown identities and properties. Thus, REPSA is a powerful, versatile, general method for the combinatorial determination of ligand-binding specificity and a functional means of ligand discovery.