Optimized Extraction Procedure for Quantifying Norisoprenoids in Honey and Honey Food Products Christine Guyot-Declerck, Fabienne Chevance, Guillaume Lermusieau, and Sonia Collin* Universite ´ Catholique de Louvain, Unite ´ de Brasserie et des Industries Alimentaires, Place Croix du Sud 2 / Bte 7, B-1348 Louvain-la-Neuve, Belgium Norisoprenoids appear as promising compounds for authenticating unifloral honeys. So far, however, no method has been optimized for their isolation from a matrix so rich in sugars. In this framework, an original extraction procedure based on the use of Amberlite XAD-16 was developed. Recovery factors were determined and compared with those obtained with another resin (XAD-2). This was done for different model media and various norisoprenoids. In aqueous or alcoholic solutions, the efficiency of both resins proved very high. As expected, addition of honey decreased the adsorption of nonpolar compounds. This effect was much more pronounced with the lower-porosity XAD-2 support. Sugar addition markedly improved the recovery factors obtained with the XAD-16 resin in the case of more polar norisoprenoids. Keywords: Honey; norisoprenoid; extraction; XAD-16; XAD-2 INTRODUCTION Norisoprenoids constitute one of the most important classes of flavor compounds identified in nature. Derived from carotenoids, they occur in many higher plants, especially tobacco (Enzell, 1985). Many have also been isolated from honey, where they constitute markers of floral origin. Australian blue gum honeys were authen- ticated by D’Arcy et al. (1997) on the basis of their 8,9- dehydrotheaspirone and 3-oxo-R-ionone contents, whereas 3,5,6-trihydroxymegastigm-7-en-9-one enabled Tan et al. (1989a; 1990) to identify New Zealand thyme honeys. High norisoprenoid contents have also been found in heather (Ha ¨ usler and Montag, 1989, 1991; Tan et al., 1989b; Guyot et al., 1999) and strawberry tree honeys (Dalla Serra et al., 1999). Although these both come from plants belonging to the Ericaceae family, they appeared distinguishable on the basis of a few specific compounds, i.e. megastigma-4,6,7-trien-3,9-dione (III, Figure 1) in the former and C 10 norisoprenoids and 5,6- epoxy-9-hydroxymegastigma-3,7-dione in the latter. As regards heather honeys, Guyot et al. (1999) recently demonstrated that Calluna vulgaris samples can be distinguished from Erica arborea honeys on the basis of their megastigm-4-en-7-yne-3,9-dione (I) and dehy- drovomifoliol (VI) contents. Although they appear so promising for authenticating unifloral honeys, norisoprenoids remain insufficiently studied because of difficulties in extracting them from such a sugar-rich matrix. So far, nonselective liquid/ liquid extractions with solvents such as dichloromethane, diethyl ether, or ethyl acetate have been used for honeys. In the case of foods with lower sugar contents, such as fruits or fruit juices, other isolation procedures have been developed to isolate free norisoprenoids and glycosides. Based on the use of nonionic resins, such techniques offer the advantage of eliminating, by simple washing with water, numerous interfering substances such as sugars and acids, without significant loss of glycosides or aroma compounds (Gunata et al., 1985). In all cases, Amberlite XAD-2 has been chosen as the solid phase (Versini et al. (1988), grape; Winterhalter and Schreier (1988), quince; Sakho et al. (1997), mango; Knapp et al. (1997), white-fleshed nectarine; Chassagne et al. (1998), passion fruit). In the case of sugar-rich matrixes, data are scant. Regarding the best resin for extracting honey flavonoids, the chemical nature and physical characteristics of Amberlite XAD resins seem to have low effect on the extraction yield (in most cases higher than 80%, Tomas-Barberan et al., 1992). The use of XAD-2, however, has been recommended to minimize the amount of methanol needed for desorption. All optimizations have been performed, unfortunately, in sugarless aqueous solutions. XAD-2 was recently chosen by Dalla Serra et al. (1999) to isolate aroma compounds from strawberry tree honeys, but without preliminary tests. To our knowledge, no research concerning Am- berlite resins has focused on optimizing norisoprenoid recoveries in sugar matrixes. The aim of the present work was, therefore, to develop an original extraction method allowing accurate quantification of noriso- prenoids in unifloral honeys. Several parameters such as the nature of the Amberlite resin, the type of solvent, the elution volume, the extraction time, and the honey/ resin ratio were optimized. As heather honey is commonly used in the fabrication of some Belgian special beers, the chosen parameters were then applied to different model media including beer, to determine the influence of the matrix, and mainly its sugar content. MATERIALS AND METHODS Materials. A unifloral heather honey (Calluna vulgaris, France), stored at 4 °C until used, was chosen because of its well-known high norisoprenoids content (Ha ¨ usler and Montag, 1989, 1991; Tan et al., 1989b; Guyot et al., 1999). Screening for floral purity was based on pollen analyses (Louveaux et al., 1978), sensory tests, electrical conductivity, pH, titratable * To whom correspondence should be addressed. Phone: (32) 10 47 29 13. Fax: (32) 10 47 21 78. E-mail: collin@inbr.ucl.ac.be. 5850 J. Agric. Food Chem. 2000, 48, 5850-5855 10.1021/jf000504g CCC: $19.00 © 2000 American Chemical Society Published on Web 11/03/2000