Analysis of antigenic surfaces of proteins.

The processes underlying molecular evolution have proved difficult to understand due to the complexity and obscurity of the selective pressures at work. The strong selective pressure to optimize antigen recognition means that antibody paratopes are more favorable systems than most in which to investigate these processes, as are viral epitopes, which evolve under an inverse pressure, rapidly changing to avoid recognition. Because recognition is a property of the surfaces of these molecules we expect that their evolutionary development may be read in the changes in the 3-dimensional array of chemical groups displayed on their surface. We have analyzed the bulk properties of these surfaces and find that there are significant differences in exposed amino acid preferences among 1) a control group of immunologically secluded proteins, 2) binding surfaces of immunoglobulins, and 3) the outer surfaces of picornaviruses. Compared to the control group, the immunoglobulin complementarity determining regions possess a relative excess of serine residues whereas picornaviruses suppress serine but overuse threonine residues, suggesting that the differing selective pressure has led to perturbations in the population of amino acid types on the surface of these proteins. Although these changes may be rationalized in terms of the structure and chemistry of the different side chains, we suggest that there may be a further, genetic component behind the observations: point mutations in the respective codons for serine and threonine lead to markedly different forms of structural variability. The high rate of reselection observed for these residues is in line with this second mechanism.