Natural contaminants in bee pollen: DNA metabarcoding as a tool to identify floral sources of pyrrolizidine alkaloids and fungal diversity

The use of bee pollen as a food supplement has increased in recent years as it contains several nutrients and phytochemicals. However, depending on floral composition, bee pollen can be contaminated by pyrrolizidine alkaloids (PAs), PA N-oxides (PANOs) and toxigenic fungi found in plants, which may pose a potential health risk for consumers. Thus, a DNA metabarcoding approach based on internal transcribed spacer 2 (ITS2) region was used to identify the plant sources of 17 PAs/PANOs detected by a validated method in liquid chromatography coupled to mass spectrometry (LC-MS/MS), as well as floral and fungal diversity in 61 bee pollen samples. According to LC-MS/MS analysis, 67% of the samples contained PAs/PANOs with mean concentration of 339 & micro;g/ kg. The contamination pattern was characterised by lycopsamine- and senecionine-type PAs/PANOs. PA/PANOproducing plants were identified in 54% of the PA/PANO-contaminated samples analysed by DNA metabarcoding, which also allowed identifying the overall floral and fungal composition of 56 samples. To evaluate the performance of the molecular approach, a subset of 25 samples was analysed by classical palynology. Palynological analysis partially confirmed the results of DNA metabarcoding, which had a better performance in distinguishing pollens of different genera from Asteraceae (76%) and Brassicaceae (88%). However, the molecular analysis did not identify pollens from Castanea, Eucalyptus, Hedera and Salix, which were abundant in 11 samples according to palynology. On the other hand, the molecular analysis allowed identifying several fungal genera in 33 samples, including the toxigenic fungi Alternaria and Aspergillus, which were positively correlated to the plant genus Hypericum. Despite limitations in identifying some pollen types, these preliminary results suggest that the DNA metabarcoding could be applied in a multidisciplinary approach to give a picture of floral and fungal diversity, which can be sources of natural contaminants in bee pollen and would help to control its safety.

Automated summary

A study compared DNA metabarcoding and traditional palynology methods to identify PAs/PANOs in bee pollen, showing that the molecular method performed better in distinguishing pollen from different genera but had limitations in identifying certain pollen types. The study also found a correlation between certain fungal genera and the plant genus Hypericum, highlighting the potential application of DNA metabarcoding in assessing floral and fungal diversity in bee pollen for safety control.