Analysis of the code relating sequence to conformation in globular proteins. Development of a stereochemical alphabet on the basis of intra-residue information

Abstract

1. The relation of primary sequence to all residue backbone conformations was explored to test out starting conformations for protein folding. 2. Information theory was used to obtain measures of information which quantitate the role of each residue in determining its own conformation; i.e. intra-residue information. 3. The information measures are plotted as a function of ϕ, ψ peptide-backbone angles and ϕ, ψ contour maps obtained for each of the 20 amino acids. These show characteristic differences between residues. 4. To find practical ways of relating sequence to ϕ, ψ angles, several types of stereochemical alphabet were investigated. The value of these was tested by using them to predict the ϕ, ψ angles of nine different proteins. 5. A difference plot was constructed to show regions of the sequence that require little or no information extra to the intra-residue information in order to predict a correct conformation. These regions are suggested to be candidates for nucleating sites in the protein.