Evaluation of the Botanical Origin of Estonian Uni- and Polyfloral Honeys by Amino Acid Content RIIN REBANE AND KOIT HERODES* Institute of Chemistry, University of Tartu, Jakobi 2, 51014 Tartu, Estonia The free amino acid content of 61 honey samples from Estonia has been determined by HPLC-UV with precolumn derivatization with diethyl ethoxymethylenemalonate. Analyzed samples were seven types of unifloral honeys and polyfloral honeys. The main amino acids found in Estonian honeys were proline and phenylalanine. The resulting data have been analyzed by t test and principal component analysis (PCA). t Test revealed that some amino acids (R-alanine, -alanine, asparagine, γ-aminobutyric acid, glutamine, glycine, histidine, ornithine, phenylalanine, proline, serine, and tryptophan) are more potent for assigning honey botanical origin than others. PCA enabled differentiation of some honey types by their botanical origin. In the space of the two first principal components, heather honeys form a cluster that is clearly separable from, for example, polyfloral honeys. It is concluded that analysis of the free amino acid profile may serve as a useful tool to assess the botanical origin of Estonian honeys. KEYWORDS: Honey authenticity; free amino acid analysis; diethyl ethoxymethylenemalonate derivati- zation; HPLC INTRODUCTION The honey market is a large part of the economy for many countries, but honeys of various geographical and botanical origins are differently valued. Due to this difference, cheaper honeys are more commonly labeled as more expensive types of honeys. The traditional method that allows verification of honey botanical and geographical origin is melissopalynology. However, this kind of pollen analysis is not conclusive and is extremely tedious and time-consuming and requires trained analysts (1). Although European Union food laws establish composition and quality parameters for honey, such figures have no relationship with the botanical or geographical origin (2). Therefore, Directive 2001/L10 from The Council of European Communities lists a need to develop methods for honey verifi- cation (3). Many studies have sought analytical markers of the botanical origin of honey, based on various honey components, one being amino acids. Twenty-six amino acids in honey account for 1% (w/w) of honey, the most abundant being proline. The origin of amino acids in honey is attributable to both animal and vegetal sources. Because pollen is the main source of honey amino acids, the amino acid profile of honey could be characteristic of their botanical origin (2). It has been shown that there is a relationship between the amino acid composition of honey and its origin, most commonly botanical origin. Davies (4) analyzed honeys from 11 different countries and found that the ratios between certain honey amino acids were different depending on the geographical origin. More research has been done on amino acid composition and botanical origin. It has been found that Spanish lavendel and eucalyptus honeys can be distinguished by their amino acid composition and that the amino acid compositions of rosemary, thyme, and orange blossom honeys are more similar (5). Another study on Spanish honeys supports those conclusions (6). Amino acid analysis of Italian honeys by gas chromatography showed that the amino acid content of different unifloral honeys varies, but no statistical analysis was carried out (7). Less work has been done about the relationship between the amino acid composition of polyfloral honeys and geographical origin. Analysis of Argentinian polyfloral honeys concluded that the amino acid composition of polyfloral honeys from different Argentinian regions varied (8). To the best of our knowledge no research has been carried out to distinguish poly- and unifloral honeys from each other by amino acid composition. For statistical treatment of amino acid composition, data may be used as absolute content of amino acid in honey or may be pretreated in several ways. The absolute concentration of amino acids in honey depends, for example, on the water content of honey, which may vary in the range of 16-20 g/100 g (9, 10). As a result, when honeys of the same botanical and geographical origin, but with different water contents, are considered, the absolute content of amino acids is different. In such situations the relative content of amino acids should better express the similarities between amino acid profiles. Which compound(s) should be used as a reference for relative amino acid content calculation? Ideally, the concentration of such a compound in honey should be independent of botanical and geographical * Corresponding author (e-mail koit.herodes@ut.ee; telephone +372 737 5259; fax +372 737 5264). 10716 J. Agric. Food Chem. 2008, 56, 10716–10720 10.1021/jf8018968 CCC: $40.75  2008 American Chemical Society Published on Web 10/31/2008