Two-stage selection of sequences from a random phage display library delineates both core residues and permitted structural range within an epitope.

Libraries of random peptides can be screened to identify species which interact with antibodies or receptors. Similarly, maps of native molecular interactions can frequently be deduced by screening a limited set of peptide fragments derived from sequences within a native antigen or ligand. However, the existence of cross-reactive sequences that mimic original epitopes and the limited replaceability of amino acid residues suggest that the sequence space accessible by a receptor can be much broader. Definition of this space is of particular importance where structural information is required for peptidomimetic or drug design. We have used a two-stage selection scheme to expand the sequence space accessible by a phage display library and to define peptide epitopes of the anti-FLAG octapeptide monoclonal M2 antibody. Affinity selection of a primary library of 2 x 10(6) random decapeptides identified a non-contiguous core of three residues in the binding motif Tyr-Lys-Xaa-Xaa-Asp. A second stage library with 2 x 10(7) individual clones bearing the core motif but with the remaining flanking and internal residues re-randomized permitted access to a broader sequence space represented in a library equivalent to several orders of magnitude larger. Data here demonstrate that extended access to binding sequence space permitted by multi-stage screening of phage display libraries can reveal not only essential residues required for ligand binding, but also the ligand structural range permitted within the receptor binding pocket.