Segmentation of a protein into structural elements: analysis and classification of segments.

A radial location measure (designated N14) derived from the X-ray crystallographic data represents an extent of exposure of an amino acid residue to solvent or location of the residue in a globular protein. A protein structure can be segmented along the chain by cutting at minima of a smoothed profile of N14. The segment so defined is a structural element traversing from a turn to the next turn and usually contains a regular secondary structure of alpha-helix or beta-strand in the middle. The characters of segments obtained for 87 proteins were examined in terms of four parameters: (1) the average angular moment (a quantity similar to the hydrophobic moment) at 40 degrees, M(40 degrees); (2) the average angular moment at 100 degrees, M(100 degrees); (3) the average value of location measures; and (4) the length of a segment. The segments were classified into 10 groups according to relative distances among the segments in the space spanned by the four parameters as components. Segments located in connecting regions (c segment) were expressed by three parameters: M(40 degrees), M(100 degrees), and the average value of location measures, and were classified into eight groups. Features of the segments were analyzed in terms of secondary structures, folding types of proteins, and occurrence of segment pairs. Protein structures are described by the arrangement of characterized segments and c segments. Cutting positions for the segmentation could be assigned from a smoothed profile calculated from its amino acid sequence, and the positions were in coincidence with the experimental ones with an accuracy of 77%. The significance of this approach to the description of protein structures in terms of segments is discussed.