Deciphering the structural code for proteins: Helical propensities in domain classes and statistical multiresidue information in α‐helices

Abstract

We made several statistical analyses in a large sample of nearly 4,000 helices (from 546 redundancy-controlled PDB protein subunits), which give new insights into the helical properties of globular proteins. In a first experiment, the amino acid composition of the whole sample was compared with the composition of two helical sample subgroups (the "mainly-a" and the " ( a / p )~ barrel" domain classes); we reached the conclusion that composition-based helical propensities for secondary structure prediction do not depend on the structural class.Running a five-residue window through the whole sample, the positional composition revealed that positive and negative residues are located throughout the helices and tend to neutralize the macrodipole effect. On this basis, we analyzed charged triplets using a running five-residue window. The conclusion was that only mixed charged residues [positive (+) and negative (-)I located at positions 1-2-5 and 1-4-5 are clearly favored. In these locations the most abundant are (--. . +) and (-. . + +), and this shows the existence of side chain microdipoles, which neutralize the large macrodipole of the helix.We made a systematic statistical analysis of charged, dipolar, and hydrophobic + aromatic residues, which enabled us to work out rules that should be useful for modeling and design purposes.Finally, we analyzed the relative abundance of all the different amphipathic double-arcs that are present in helices formed by octapeptides (8) and nonapeptides (18). All of the double-arcs that make up Schiffer and Edmundson's classical helical wheel are found in abundance in the sample.Keywords: a-helix; helical patterns; hydrophilic residues; hydrophobic residues; medium-range interactions Since Chou and Fasman's pioneering work (Chou & Fasman, 1974), it has been accepted that the bulk of consecutive amino acid residues within a polypeptide sequence, with a high average intrinsic propensity, defines the nucleus of secondary structure segments in a protein (a-helices, @strands, and reverse turns). Under this assumption, a number of secondary structure predictive procedures, based on sets of propensities, have been worked out, all of which are merely refinements of the Chou and Fasman (1974) initial method. The secondary structure still cannot be accurately predicted because a number of factors derived from short-and mediumrange interactions among neighboring residues and between residues and the solvent are not well understood, and therefore, they are not considered within the algorithms that are presently used for structural predictions.

Local Destabilization of the Tropomyosin Coiled Coil Gives the Molecular Flexibility Required for Actin Binding

“…If optimal flexibility and stability are more important than the identity of the interface amino acids per se, how are the Ala clusters present in the "ancestral" 42-residue tropomyosin period (25) conserved through gene duplications and evolution? Alanine has high helical propensity (72) and is the most abundant amino acid in R-helices, along with leucine (73). Ala also has a high frequency in coiled coils, second to leucine in the interface positions, although the frequency in tropomyosin is unusually high (24,74).…”