Structural characteristics of food protein-originating di- and tripeptides using principal component analysis

Anna Iwaniak,Monika Hrynkiewicz,Justyna Bucholska,Małgorzata Darewicz,Piotr Minkiewicz

doi:10.1007/s00217-018-3087-3

Abstract

This paper presents the use of principal component analysis (PCA) to study the effect of specific physicochemical attributes on bitterness of di- and tripeptides originating from food proteins. Peptide sequences were derived from the BIOPEP-UWM database of sensory peptides and amino acids. Descriptors defining the physicochemical properties of amino acids forming the analyzed peptides were study variables. They were derived from ProtScale program and Biological Magnetic Resonance Data Bank. Finally, PCA was carried out for 51 dipeptides/12 variables, and 51 tripeptides/18 variables using STATISTICA®13.1 software. PCA allowed reducing the input datasets to 4 principal components (PCs) for dipeptides and to 5 PCs for tripeptides. The impact of the following properties on the bitterness of peptides was observed: relatively high molecular weight, bulkiness, increasing number of carbon and hydrogen atoms of amino acids forming the sequences. These properties characterized the N- (negative correlations) and C-terminal residue (positive correlations) of both di- and tripeptides. An additional property affecting peptide bitterness was amino acids’ hydrophobicity. Our results were consistent with scientific reports on structure–bitterness of peptides. Thus, we find PCA a chemometric approach helpful in broadening the knowledge about the function of peptides resulting from their chemical nature.

Highlights

Food provides the organism with energy as well as essential components necessary to maintain the life functions [1]
Descriptors such as molecular weight (MW), B, P and HKD were obtained from the ProtScale program available at: http://web.expasy.org/protscale [23], whereas nCat and nHat were taken from the Biological Magnetic Resonance Data Bank [24]
Bitter di- and tripeptide separate datasets were studied using principal component analysis (PCA) for few reasons. They were the most numerous group of bitter peptides listed in BIOPEP-UWM database comparing to longer chain sequences [19] and such sets of di- and tripeptides enabled to obtain well-conditioned matrices [25]

Summary

Introduction

Food provides the organism with energy as well as essential components necessary to maintain the life functions [1]. A continuously increasing load of information about biological phenomena (e.g., derived from experimental data) has contributed to the development of bioinformatic (in silico) tools which, e.g., support the research by helping plan the experiments and/or analyze data and interpret the results Examples of such tools include databases of biological and chemical information of compounds, programs for the prediction of physicochemical properties of molecules, for the simulation of processes taking place in the living systems, and for the analysis of the function of a molecule taking. The above-mentioned disciplines are useful in explaining processes and/or phenomena [9] by transformation of the multi-dimensional datasets to find general regularities between the variables and/or to define some attributes (variables) that may potentially affect properties of a molecule [10, 12] Such an approach may be used to analyze the bitter taste of peptides considering their structure defined by numerical physicochemical attributes. The accessibility to databases and in silico programs providing the information about the bitter peptide sequences as well as variables describing the physicochemical properties of single amino acids forming the peptides encouraged us to undertake the study aimed at applying the PCA to define the properties represented by PCs and deciding about the bitterness of food protein-originating di- and tripeptides

Materials and methods

Results and discussion

Compliance with ethical standards