Defining intrinsic hydrophobicity of amino acids' side chains in random coil conformation. Reversed-phase liquid chromatography of designed synthetic peptides vs. random peptide data sets

The two leading RP-HPLC approaches for deriving hydrophobicity values of amino acids utilize either sets of designed synthetic peptides or extended random datasets often extracted from proteomics experiments. We find that the best examples of these two methods provide virtually identical results--with exception of Lys, Arg, and His. The intrinsic hydrophobicity values of the remaining residues as determined by Kovacs et al. (Biopolymers 84 (2006) 283) correlates with an R(2)-value of 0.995+ against amino acid retention coefficients from our Sequence Specific Retention Calculator model (Anal. Chem. 78 (2006) 7785). This novel finding lays the foundation for establishing consensus amino acids hydrophobicity scales as determined by RP-HPLC. Simultaneously, we find the assignment of hydrophobicity values for charged residues (Lys, Arg and His at pH 2) is ambiguous; their retention contribution is strongly affected by the overall peptide hydrophobicity. The unique behavior of the basic residues is related to the dualistic character of the RP peptide retention mechanism, where both hydrophobic and ion-pairing interactions are involved. We envision the introduction of "sliding" hydrophobicity scales for charged residues as a new element in peptide retention prediction models. We also show that when using a simple additive retention prediction model, the "correct" coefficient value optimization (0.98+ correlation against values determined by synthetic peptide approach) requires a training set of at least 100 randomly selected peptides.