Untargeted metabolomics suggests a trade-off between phenolamides and flavonoids in the pollen of Petunia hybrida with contrasting flower colours

BACKGROUND: Pollen is vital for reproduction of flowering plants. Several nutritional and biological benefits for humans and pollinators are described and related to its richness and diversity of metabolites. However, the chemical composition of pollen from several horticulturally significant plant species, such as those in the genus Petunia, has not been thoroughly characterised. Here, we present the first comprehensive description of the chemical profile of two distinct P. hybrida lines: V26 (violet flowers) and W115 (white flowers) using untargeted metabolomics. Our workflow started from pollen sampling and isolation, followed by a 3-in-1 liquid phase extraction for wide-range metabolite recovery, and data acquisition by ultra-high-performance liquid chromatography coupled to high-resolution tandem mass spectrometry (UHPLC-HRMS/MS) using three chromatographic columns (C8, C18, and HILIC) with positive and negative ionisation. For data analysis, we implemented a user-friendly and reproducible data processing pipeline based on open-source computational tools. RESULTS: The P. hybrida pollen is rich in glycosylated flavonoids, phenolamides, and lipids, which were detected mostly in non-polar phase extracts analysed by reversed-phase chromatography. Simple phenylpropanoids, fatty acids, amino acids, and terpenoids were annotated to a lesser extent. Statistical analyses integrated with molecular networking demonstrated a distinct metabolic dichotomy: phenolamide derivatives are predominantly present in the pollen of the V26 line, while flavonoids are accumulated in the pollen of W115. This finding suggests different regulation of the phenylpropanoid metabolism in pollen of P. hybrida lines with differing flower colours and sheds light on hypotheses of ecological roles of pollen secondary metabolites, for example, in plant-pollinator interactions. CONCLUSIONS: Our findings suggest a metabolic trade-off in the phenylpropanoid pathway in the two studied P. hybrida lines. These cultivar-specific chemical signatures may have significant implications for pollen viability and interactions with pollinators. Furthermore, our analytical and computational workflow serves as a robust template for the deep metabolic profiling of other under-characterised and complex natural matrices.