Aromatic Side-Chain Interactions In Proteins. I. Main Structural Features

In a data set of 593 nonhomologous proteins from the PDB, we have analyzed the
pairing of phenylalanine, tyrosine, tryptophan, and histidine residues with their
closest aromatic partner. The frequency distribution of the shortest interatomic
distance of partners is bimodal with a sharp peak at approximately 3.8 A and a
wider one at a longer distance. Only the 3.8 A peak corresponds to direct
ring-ring interactions thus aromatic pairs. The aromatic pairs were separated
into two classes, near-sequence pairs and far-sequence pairs. Near sequence pairs
stabilize local structure, and far-sequence pairs stabilize tertiary structure.
Far-sequence pairs (74% of all pairs) mainly bridge two beta-strands, followed by
pairs that bridge a beta-strand and a helix, and pairs that bridge a beta-strand
and a random coil structure. Pairs that bridge helices are rare. The secondary
structure of the near-sequence pairs depends on the partner distance in the
sequence. When the partners are 1, 3, or 4 residues apart in the sequence, pairs
are mostly found in helical structures. When the partners are two apart, pairs
are mostly found in the same beta-strand. Analysis of the frequency of near
sequence pairs supports the hypothesis that aromatic pairing occurs after, rather
than before, the formation of secondary structures.