Conservative interactions in protein evolution

Patterns of sequence conservation are widely used in studies of proteins. Protein families, and more recently also protein superfamilies, are characterized by a shared three-dimensional structure and patterns of conservation of amino acid sequence. Obtaining a comprehensive understanding of the factors that drive the conservation of residues at specific positions in a protein is an important challenge in understanding protein structure, function, and evolution. We explore this question in the PAS domain superfamily. PAS domains form a diverse superfamily of signaling proteins defined by a weak but characteristic pattern of sequence conservation. The 9 most strongly PAS-conserved residues are located at structurally inconspicuous positions that remain largely unstudied. Thus, while evolution indicates that these residues are critical, the mechanistic reason for their conservation remains unresolved. Two candidate functions are in PAS domain folding/stability and in allosteric switching for signaling. Our experimental data on photoactive yellow protein argue against a role in allosteric switching or protein stability. However, we observed that mutations at these residues often substantially reduce the degree of protein production. Further analysis revealed conserved residues that affect protein stability and in vivo protein production, indicating that both effects drive evolutionary conservation in PYP, but through two largely distinct sets of residues. Bioinformatics analysis of contact maps in PAS domains revealed conserved networks of interactions, where side chain identity is variable, but a cluster of inter-residue interactions is retained. We propose that such conservative interactions is important for efficient protein production.