Engineering the binding site of an antibody against N-glycolyl GM3: from functional mapping to novel anti-ganglioside specificities.

The structurally related gangliosides N-glycolyl GM3 and N-acetyl GM3 are potential targets for tumor immunotherapy. 14F7 is a monoclonal antibody able to discriminate the tumor-specific antigen N-glycolyl GM3 from the closely related N-acetyl GM3 on the basis of the presence of a single additional hydroxyl group in the former. A combinatorial phage display strategy, based on the screening of a large library followed by refined mutagenesis, allowed a thorough exploration of the binding chemistry of this unique antibody. Three essential features of the heavy chain variable region were identified: two aromatic rings (in positions 33 and 100D) contributing to the binding site architecture and an arginine residue (position 98) critical for recognition. Directed evolution of 14F7 resulted in novel variants that cross-react with the tumor-associated antigen N-acetyl GM3 and display recurrent replacements: the substitution W33Q and the appearance of additional arginine residues at several positions of CDR H1. Successful conversion of such engineered variable regions into whole cross-reactive anti-GM3 immunoglobulins validated our phage-based approach to study and modify the lead antibody 14F7. The resulting family of closely related antibodies offers new tools to study the mechanisms of cell death induced by antibodies targeting gangliosides. In vitro directed evolution was useful to overcome the technical limitations to obtain anti-ganglioside antibodies. The case of 14F7 illustrates the power of combining library screening with focused site-directed randomization for a comprehensive scanning of protein interactions.