A diversity optimized combinatorial library for the identification of Fc-fragment binding ligands

To generate a library covering a relatively wide area of the chemical space, molecular descriptors, and multivariate data analysis were combined to select the building blocks required for generating a diversity optimized library of putative Fc-fragment binding ligands. In such a method of library design, structural information about the target protein is not needed. Synthesis of the resulting 770 member virtual library was carried out using encoded beads, which facilitated rapid identification of the subsequent hits. The library was screened in an on-bead fluorescence assay with immunoglobulin G Fc-fragment of the subtype 4 to identify Fc-fragment binding ligands that would be useful for purifying monoclonal antibodies. An analysis of the positions of the hits in the chemical space revealed that the ones with the highest fluorescence were primarily concentrated in a particular part of the chemical space. One of the identified hits, when immobilized on amino sepharose, was able to purify a monoclonal antibody from crude cell supernatant with purity of 84% and a 70% recovery. The chemometric tools employed for the library design allowed the identification of the fraction of the available chemical space that would be preferred for a second generation library.