Favored and suppressed patterns of hydrophobic and nonhydrophobic amino acids in protein sequences

Hydrophobic amino acids of the group Leu, Ile, Val, Phe, and Met (LIVFM) are distributed in favored or suppressed patterns within protein sequences. The frequencies of all five-position combinations of [symbol: see text] = LIVFM and [symbol: see text] = non-LIVFM residues were analyzed in 48 proteins of known crystallographic structure. Some motifs were strongly preferred or suppressed; e.g., [symbol: see text] was favored (z = 3.5), while [symbol: see text] was suppressed (z = -3.4). In longer patterns, [symbol: see text] followed by [symbol: see text] and one [symbol: see text] was favored ([symbol: see text], z = 5.1), while conversion of the single hydrophobic residue to a pair was not ([symbol: see text], z = 0.8). Distributions of certain non-LIVFM amino acids around [symbol: see text] positions in strongly favored patterns were also favored or disfavored (Asp, Glu, Lys, Arg, Asn, Cys, Tyr, and Pro; for each magnitude of z > 2.0). While the strongly favored pattern [symbol: see text] was found in both alpha-helical and beta-strand sequences, it associated significantly with alpha-helices (z = 3.6 for the second-position alpha-helical phi and psi angles) but not with beta-strands (z = -1.1). Certain motifs of LIVFM and non-LIVFM residues might be selected if they lead efficiently to the local nucleations hypothesized to characterize molten globule intermediates in the folding of proteins.