Antigen‐binding site anatomy and somatic mutations in antibodies that recognize different types of antigens

The number of antibody structures co-crystallized with their respective antigens has increased rapidly in the last few years, thus offering a formidable source of information to gain insight into the structure-function relationships of this family of proteins. We have analyzed here 140 unique middle-resolution to high-resolution (<3 Å) antibody structures, including 55 in complex with proteins, 39 with peptides, and 46 with haptens. We determined (i) length variations of the hypervariable loops, (ii) number of contacts with antigen, (iii) solvent accessible area buried upon binding, (iv) location and frequency of antigen contacting residues, (v) type of residues interacting with antigens, and (vi) putative somatic mutations. Except for somatic mutations, distinctive profiles were identified for all the variables analyzed. Compared with contacts, somatic mutations occurred with less abundance at any given position and extended beyond the regions in contact, with no clear difference among antibodies that recognize different types of antigens. This observation is consistent with the fact that although antigen recognition accomplished by shape and physicochemical complementarity is selective in nature, the somatic mutation process is stochastic and selection for mutations leading to improved affinity is not directly related to contact residues. Thus, the knowledge emerging from this study enhances our understanding of the structure-function relationship in antibodies while providing valuable guidance to design libraries for antibody discovery and optimization.