Patterns of Sequence and 3-D Structure Variation in Families of Homologous Proteins: Lessons for Tertiary Templates and Comparative Modelling

Abstract

More than 20 years ago X-ray analysis (Adams et al., 1969) showed that insulin was a globular protein which could form dimers and hexamers. Although there were very few sequences of proteins in general available, there were several sequences of insulin (Fig. 1). When these were aligned, it was immediately apparent that much of the pattern of amino acid substitutions could be understood in terms of the three- dimensional structure. For example, the conserved disulphide bridges could be seen to stabilize the association of the A and B-chains and two of them were buried in the core of the protein structure. It was also apparent that many of the amino acid residues in the core were conserved as hydrophobic; for example leucine, isoleucine, valine and phenylalanine. But there was a problem with our structure analysis: residue (A16), an invariant leucine, was fully exposed to the solvent and surrounded by other surface residues that were both polar and variable. As a result of the concern over this structural feature, we were forced to consider remodelling the structure. We discovered that a better interpretation of the electron density could be obtained with a conformation in which the leucine contributed to the conserved hydrophobic core (Blundell et al., 1971).