Hydrophilic properties as a new contribution for computer-aided identification of short peptides in complex mixtures.

A new method to predict elementary amino acid (AA) composition of peptides (molar mass <1,000 g/mol) is described. This procedure is based on a computer-aided method using three combined analyses-reversed phase liquid chromatography (RPLC), hydrophilic interaction chromatography (HILIC) and capillary electrophoresis coupled with mass spectrometry-and using a software calculating all possible amino acid combinations from the mass of any given peptide. The complementarity between HILIC and RPLC was demonstrated. Peptide retention prediction in HILIC was successfully modelled, and the achieved prediction accuracy was as high as r²=0.97. This mathematical model, based on amino acid retention contributions and peptide length, provided the information about peptide hydrophilicity that was not redundant with its hydrophobicity. Correlations between respectively the hydrophobicity coefficients and RPLC retention time, hydrophilicity and HILIC retention time, and electrophoretic mobility and migration time were used for ranking all potential AA combinations corresponding to the given mass. The essential contribution of HILIC in this identification strategy and the need to combine the three models to significantly increase identification capabilities were both shown. Applied to an 18-standard peptide mixture, the identification procedure enabled the actual AA combination determination of the 14 di- to pentapeptides, in addition to an over 98 % reduction of possible combination numbers for the four hexapeptides. This procedure was then applied to the identification of 24 unknown peptides in a rapeseed protein hydrolysate. The effective AA composition was found for ten peptides, whereas for the 14 other peptides, the number of possible combinations was reduced by over 95 % thanks to the association of the three analyses. Finally, as a result of the information provided by the analytical techniques about peptides present in the mixture, the proposed method could become a highly valuable tool to recover bioactive peptides from undefined protein hydrolysates.

Modeling the separation of small peptides by cation-exchange chromatography.

Cation-exchange chromatography was applied for the separation of short synthetic peptide standards, with various charges and hydrophobicities. The methodology to develop simple, reproducible and accurate models, based on physicochemical peptide properties, was described. A multilinear regression method was used for the calculation of the models, and descriptors were chosen according to the observed phenomena. The predictive and interpretative ability of the chromatographic models was evaluated considering cross-validated data (root mean-squared error of calibration, root mean-squared error, root mean-squared error of cross-validation and the Fisher ratio). Hydrophobic coefficients for amino acids were calculated with or without consideration of peptide sequences. A simple model, with only two parameters (charge and hydrophobic coefficient) was built. It enabled an accurate prediction of short peptide elution (up to nine residues). As the model was intended for further characterization of complex mixtures of unknown peptides, some mixtures were analyzed to investigate possible interactions between molecules. Peptides eluted in exactly the same pattern as when injected alone, supporting the use of these models for complex mixtures of small peptides.