236 M. KLVANA ET AL. Copyright © 2006 John Wiley & Sons, Ltd. Phytochem. Anal. 17: 236–242 (2006) DOI: 10.1002.pca Phytochemical Analysis Phytochem. Anal. 17: 236–242 (2006) Published online 14 June 2006 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/pca.913 Analysis of Secondary Metabolites from Eschscholtzia californica by High-performance Liquid Chromatography MAYA KLVANA, 1 JINGKUI CHEN, 1 FRANÇOIS LÉPINE, 2 ROBERT LEGROS 1 and MARIO JOLICOEUR 1 * 1 Bio-P2 Research Unit, Department of Chemical Engineering, École Polytechnique de Montréal, PO Box 6079 Centre-ville Station, Montréal, Québec, Canada H3C 3A7 2 Centre de Microbiologie et Biotechnologie, Institut Armand-Frappier, Université du Québec, 513 Boul. des Prairies, Laval, Québec, Canada H7V 1B7 Received 28 March 2005; Revised 31 August 2005; Accepted 16 September 2005 Abstract: A rapid and precise analytical HPLC method has been developed for screening the major benzophenanthridine alkaloids produced by cell cultures of Eschscholtzia californica, namely, sanguinarine, chelirubine, macarpine, chelerythrine and chelilutine. Separation was achieved on a C 18 reversed-phase column with gradient elution using acetonitrile and 50 mM phosphoric acid. Detection was performed by both fluorescence (λ ex 330 nm, λ em 570 nm) and photodiode array, leading to good selectivity and precision in determining peak purity. A simple and quick sample preparation protocol was elaborated involving a methanolic extraction for the measurement of intracellular concentrations of the alkaloids and a solid phase extraction for their quantification in culture medium. Owing to the non-availability of commercially standards, a method for the purification of chelirubine, macar- pine and chelilutine by semi-preparative HPLC was developed. Coupled together, the isolation method and the analytical method were highly reliable for screening the alkaloids of interest produced by E. californica. Copyright © 2006 John Wiley & Sons, Ltd. Keywords: HPLC; HPLC-MS, NMR; benzophenanthridine alkaloids; sanguinarine; Eschscholtzia californica. Phytochemical Analysis INTRODUCTION Benzophenanthridine alkaloids, especially sanguinar- ine and chelerythrine, have attracted a great deal of interest with respect to their medical properties. In addition to being used in dental products on account of its antibacterial activity (Grenby, 1995; Eley, 1999), sanguinarine has recently been shown to be pro- mising as a cancer treatment (Ahmad et al., 2000; Weerasinghe et al., 2001a, b; Slaninova et al., 2001). Similarly, arising from its ability to inhibit protein kinase C, chelerythrine offers potential as an anti- tumour drug (Chmura et al., 2000). Both of these compounds are produced by cells of Eschscholtzia californica within the alkaloidal metabolic pathway (Klvava et al., 2004). The major alkaloids most com- monly reported in this plant include sanguinarine, chelirubine, macarpine, chelerythrine, chelilutine and sanguirubine, as well as the dihydro-intermediates of all of the above. Most of the alkaloids found in E. californica can be produced aseptically and under controlled environmental conditions using cell cultures. Research related to the in vitro production of these secondary metabolites has succeeded in improving production through elicitation (Schumacher et al., 1987; Brodelius et al., 1989; Byun et al., 1992), in-situ product recovery (Byun and Pedersen, 1991; Williams et al., 1992; Dutta et al., 1994) and, more recently, through metabolic engineering of enzymes involved in the alkaloid metabolic pathway (Facchini et al., 1996; Haider et al., 1997; Park et al., 2002). However, be- cause only some of the constituents of the metabolic pathway are of commercial value, it is necessary fully to understand the dynamics of the metabolic flux. Thus, an accurate and precise analytical method is needed in order to measure levels of the main benzo- phenanthridine alkaloids. Many different HPLC methods have been used to evaluate benzophenanthridine alkaloid production. Most of these methods are lengthy, involving run times of more than 45 min (Schumacher et al., 1987; Tanahashi and Zenk, 1990; Tome et al., 1999; Fabre et al., 2000; Park et al., 2002), and many require tedious sample preparation owing to poor selectivity of detection. Indeed, the vast majority of reported methods employ UV detection at wavelengths around 280 nm, leading to poorly defined and, sometimes, overlapping peaks that render subsequent quantifica- tion unreliable (Schumacher et al., 1987; Collinge and Brodelius, 1989; Park et al., 2002). In order to increase selectivity, Chauret et al. (1990) and Chauret and Archambault (1991) took advantage of the fluorescence properties of the alkaloids and proposed a simple sample preparation involving extraction with acidified methanol. Unfortunately, the results could not be replicated owing to the low concentrations of meta- bolites in the cells and, especially, in medium, which * Correspondence to: M. Jolicoeur, Bio-P2 Research Unit, Department of Chemical Engineering, École Polytechnique de Montréal, PO Box 6079 Centre-ville Station, Montréal, Québec, Canada H3C 3A7. E-mail: Mario.Jolicoeur@polymtl.ca Contract/grant sponsor: The Canadian Foundation for Innovation and the Fonds Nature et Technologies du Québec (NATEQ).