Surface β-strands in proteins:: Identification using an hydropathy technique

From a representative set of monomeric globular proteins with known three-dimensional structures, beta -strands with lengths greater than or equal to5 amino acids have been identified and catalogued. By ascertaining the accessible surface areas of the constituent residues in these strands, and by checking whether the exposed/buried pattern is 80% or more similar to that in an idealised surface strand, a subset of structures can be delineated in which the beta -strands are all sited on the surface of the protein. The corresponding sequence data show that about 50% of the residues are apolar (Val, Be, Leu, Phe, Tyr, Ala) and that the common occurrence of valine (14.3%), isoleucine (9.6%), and threonine (8.1%) is a characteristic feature. The frequencies of occurrence of those amino acids in the strands that face the aqueous environment and the interior have also been determined separately and show that most surface strands have a substructure of the form (apolar-X)(n), where X is approximately equally divided between apolar, charged, and hydrophilic residues. Using the frequency data thus obtained, allied with an algorithm to delineate potential surface beta -strands from characteristic hydropathy profiles, it is now possible to search through the sequences of proteins with unknown tertiary structures and make realistic predictions of the presence of this element of structure on the protein surface. In addition, new data are presented on the distribution of the various types of residues on the surface of proteins and in their interior. Significant differences were observed, not all of which have been identified previously. Furthermore, the distribution of the types of residue in a surface beta -strand was compared to that corresponding to the surfaces of all of the proteins in our database. Again, very characteristic differences were observed. These are helpful in recognising the presence of surface beta -strands. (C) 2000 Academic Press.