An attempt to locate the non-helical and permissively helical sequences of proteins: application to the variable regions of immunoglobulin light and heavy chains.

Abstract

From a consideration of (varphi, Psi) values of the amino acids of myoglobin, lysozyme, the alpha and beta chains of horse oxyhemoglobin, tosyl-alpha-chymotrypsin, and carboxypeptidase A, an empirical procedure of predicting whether amino-acid residues in proteins are in a non-helical or may be in a helical conformation has been developed. The conformation of an amino acid at any position n is considered to be influenced by its nearest neighbors (the amino acids at positions n + 1 and n - 1), and the (varphi, Psi) values of the middle amino acid n for the various tripeptide sequences in the known proteins are tabulated. If helical, the (varphi, Psi) values are plotted to define a helical (varphi, Psi) domain. A 20 x 20 table for all tripeptides (n - 1)-(n)-(n + 1) taken sequentially for the entire chain was constructed; it lists the number of instances in which helical and non-helical conformations for the amino acids at position n were found. Certain sequences are found to be associated exclusively with non-helical and others exclusively with helical conformations, whereas many sequences may be either helical or non-helical. The distribution of non-helical residues serves to limit stretches of permissively helical regions; these are then further examined by the helical wheel method. As applied to cytochrome c from 18 species, the only permissively helical segment found was the stretch 91-101 near the C-terminus. For the variable regions of three light and three heavy chains of immunoglobulins, upper limits of 12 and 17% alpha-helix, respectively, were obtained.