Principles of Protein Structure - H. Schulz 1982

Prediction of secondary structure from amino acid sequence
Physicochemical methods
Methods based on stereochemical data

Reverse turns consist predominantly of polar residues. A simple stereochemical approach to predicting chain turns was proposed by Kuntz [203], who noted that turns are located almost exclusively on the protein surface. Based on this, he suggested that all triplets of polar residues (designated as T in Table 6.2) adopt a reverse-turn conformation.

Helices situated between the Hydrophobic core and the solvent can be identified by the arrangement of their nonpolar residues. Using graphs spanning five turns of the helix, Schiffer and Edmundson [376, 377] demonstrated that nonpolar residues cluster along one edge of the helix to form nonpolar arcs. An example of such an "α-helical wheel" is shown in Fig. 6.1. To predict helices, the authors examined the presence of nonpolar triplets at positions i, i + 3, i + 4 (relative positions 1-4-5), as well as at positions i, i − 3, i − 4 (1-2-5) within a given Amino Acid Sequence. These positions were then designated as helical nucleation sites, around which the "helical wheels" were constructed. It was assumed that the helix would propagate in both directions until the nonpolar arc was interrupted by polar residues or a Pro residue. This method was designed to locate helices shielding the hydrophobic core from the solvent, as in the case of Myoglobin; detecting helices in other locations using this approach is more challenging.

The arrangement of nonpolar residues in α-helices was also examined by Palau and Puigdomènech [378], who used a considerably larger benchmark dataset. These authors showed that the frequency of nonpolar triplets at relative positions 1-2-5 and 1-4-5 in α-helices is significantly higher than would be expected from the Amino Acid Composition of these helices. Frequently, these triplets were arranged in continuous chains such as 1-2-5-"1"-"4"-"5" → 1-2-5-8-9, thereby forming arcs on the "α-helical wheel" (Fig. 6.1). Thus, the observations of Schiffer and Edmundson, originally based on data from only three Proteins, were confirmed for a larger dataset.

Class="center">

Fig. 6.1. Helical wheel of the C-terminal α-Helix of adenylate kinase. (a) The wheel represents a PROJECTION OF THE positions of all side chains along the helical axis onto a plane. Residue 1 is Val-179, and the final residue at position 16 is Lys-194. Nonpolar residues (circled) are located on one side of the helix, which faces the hydrophobic core in the three-dimensional Structure [186]; the other side of the helix forms part of the outer surface of the molecule. (b) Positions of the side chains on a cylindrical diagram of the α-helix (Fig. 5.2). Residues forming nonpolar arcs are circled.

A method for identifying the arrangement of residues in helices and β-pleated sheets. The consideration of nonpolar residue nucleation sites was also utilized by Lim in his Secondary structure predictions [379, 380]. Similar to Palau and Puigdomènech [378], he drew attention to the arrangement of triplets at relative positions 1-2-5 and 1-4-5, as well as doublets at positions 1-4 and 1-5. If the side chains were excessively bulky, cluster formation was not considered. METABOLISM/18.html">The Influence of polar residues was also taken into account, given that large polar residues can stabilize adjacent nonpolar clusters by shielding them from the solvent. In predicting β-structures, a distinction was made between internal, semi-surface, and surface β-sheets. Surface sheets possess both polar and nonpolar faces and therefore exhibit an alternation of polar and nonpolar residues. Conversely, in internal sheets, nonpolar residues are located on both sides, allowing these structures to be identified by a continuous alternation of nonpolar residues. Such segments are typically capped by large polar residues at the chain termini. In semi-surface sheets, all side chains are oriented obliquely relative to the surface, as seen, for example, in the C-terminal β-sheet of the central β-structure of adenylate kinase shown in Fig. 7.6. Such β-sheets may lack bulky nonpolar residues, as there is no need to shield them from the solvent.

Lim formulated his concepts as 22 rules for α-predictions and 14 rules for β-predictions. He compiled a table of 17 anti-helical doublets and 26 anti-helical triplets, and grouped amino acid residue types in 14 different ways. The β-conformation was postulated solely for regions where an α-helix was not predicted. It should be noted that Lim's rules contain a large Amount of Information—that is, a high number of variable parameters—and consequently, they can always be reconciled with a benchmark dataset, even if that dataset is large. Therefore, the validity of this method can only be tested by applying it to unknown structures.