Bioassay-directed analysis-based identification of relevant pyrrolizidine alkaloids

Abstract

Pyrrolizidine alkaloids (PAs) are produced by various plant species and have been detected as contaminants in food and feed. Monitoring programmes should include PAs that are present in relevant matrices and that exhibit a high toxic potential. The aim of the present study was to use a bioassay-directed analysis approach to identify relevant PAs not yet included in monitoring programmes. To that end, extracts of Heliotropium europaeum and H. popovii were prepared and analysed with LC-MS/MS for the presence of 35 PAs included in monitoring programmes, as well as for genotoxic activity in the HepaRG/γH2AX assay. Europine, heliotrine and lasiocarpine were found to be the most abundant PAs. The extracts showed a higher γH2AX activity than related artificial mixtures of quantified known PAs, which might point to the presence of unknown toxic PAs. The H. europaeum extract was fractionated and γH2AX activities of individual fractions were determined. Fractions were further analysed applying LC-Orbitrap-MS analysis and Compound Discoverer software, identifying various candidate PAs responsible for the non-explained genotoxic activity. Altogether, the results obtained show that bioassay-directed analysis allows identification of candidate PAs that can be included in monitoring programmes.

Keywords: Bioassay; HepaRG cells; LC-MS/MS; LC-Orbitrap-MS; Pyrrolizidine alkaloids; γH2AX assay.

Fig. 1

Chemical structures of some frequently occurring PAs. Unsaturated PAs can be divided in different categories according to their type of necine base (retronecine, heliotridine, otonecine) and their type of esterification (cyclic diesters, open diesters or monoesters). The necine base can be oxidised at the nitrogen atom, giving rise to PA N-oxides (not shown)

Fig. 2

The effect of increasing concentrations of reduced (R) and non-reduced (NR) extracts of H. europaeum, H. popovii and chamomile (M. recutita) on viability of HepaRG cells (squares, right Y-axes) and γH2AX induction (bars, left Y-axes). For each condition, mean values (± SD) from two independent experiments are presented

Fig. 3

Comparison of the effects of reduced extracts of H. europaeum and H. popovii and artificial PA mixtures on cell viability and γH2AX induction. HepaRG cells were exposed to different dilutions (up to 9× diluted) of plant extracts and artificial mixtures in DMSO and analysed for their effects on cell viability (squares, right Y-axes) and γH2AX induction (bars, left Y-axes). For each condition, mean values (± SD) from two independent experiments are presented

Fig. 4

Concentration–response data and modelling for the PAs europine, heliotrine, and lasiocarpine and a ternary mixture thereof. A HepaRG cells were exposed to increasing concentrations of europine, heliotrine, and lasiocarpine and a ternary mixture as referred in Supplementary Table 5. After 24 h, cells were subjected to the WST-1 and γH2AX assays. B, C Concentration–response modelling of γH2AX induction data was performed as described in the materials and methods section using PROAST software. In the right-hand-side legend of the plots, a number of PROAST annotations and corresponding values are given that are related to the fitted model (including RPFs); for description of the annotations, see Supplementary Fig. 3. The values at the x-axis are concentration equivalents of the reference PA, europine. Green diamonds represent europine, red crosses heliotrine, black triangles lasiocarpine, and blue triangles the europine–heliotrine–lasiocarpine ternary mixture. The obtained curves represent the four-parameter Hill model. Concentration–response modelling and analysis for the single compounds (B) and the single compounds plus their mixture (C) show an overall fit of the concentration–response curves, pointing to dose addition as further described in the text

Fig. 5

Effects of different dilutions of the 10 fractions of reduced H. europaeum extract on viability of HepaRG cells (circles, right Y-axes) and γH2AX induction (bars, left Y-axes). For each condition, mean values (± SD) from two independent experiments are presented

Fig. 6

γH2AX activity of active fractions of reduced H. europaeum extract (Fig. 5) expressed in riddelliine equivalents of quantified PAs (Table 2) compared to the concentration–response curve for riddelliine-induced γH2AX activity. Riddelliine data were taken from Louisse et al. (2019). The highest concentration of fraction 9 was excluded because of substantial cytotoxicity (see Fig. 5)

Fig. 7

LC–Orbitrap-MS chromatograms and mass spectra of fractions 7–10 of the reduced extract of H. europaeum. A Fraction 7: 5′-acetyleuropine + 7-acetyleuropine. B Fraction 8: heleurine. C Fraction 9: 7-tigloyleuropine. D Fraction 10: 7-angeloylheliotrine + 5′-acetyllasiocarpine