Valorisation of the Italian biodiversity: specialised metabolism in Rosid clade

Plant biodiversity represents an exceptional reservoir of chemical diversity, yet only a limited fraction of plant species has been investigated for their specialised metabolism and potential applications. This thesis is embedded within the Mediterranean Phytocomplex Library (MPL), a bioprospecting initiative designed to represent the Italian flora and to explore its phytochemical space through a metabolomic approach. The focus of this work is the Rosid clade, one of the largest and most diverse clades in the Italian flora, represented in the MPL by 172 species. The sampling strategy was to collect a number of species roughly proportional to family size, allowing broad coverage of chemical space and providing a robust overview of metabolic diversity across the clade, particularly within the largest Rosid families. A comparative untargeted metabolomic analysis was performed using LC–MS, generating a feature quantification matrix based on 516 samples belonging to Rosid clade (172x3) analysed within a single experimental framework and comprising 1,598 manually validated metabolites. This represents one of the most extensive metabolomic datasets generated for the Rosid clade to date. The HeatMaps and PCA analyses revealed that metabolomic similarity does not strictly mirror phylogenetic relationships. Instead, chemical space appears partially decoupled from taxonomy, reflecting the combined effects of ancient conserved biosynthetic pathways and lineage-specific metabolic innovations. Widespread metabolite classes, such as flavonoids and terpenoids, were broadly conserved across taxa, whereas other pathways, including glucosinolate biosynthesis, were restricted to specific lineages, notably the Brassicales. Comparative analyses of major families, Fabaceae, Rosaceae and Brassicaceae, highlighted distinct chemical space. Within Rosaceae, metabolomic comparisons between the subfamilies Amygdaloideae and Rosoideae revealed contrasting dominance of flavonoids and hydrolysable tannins, respectively, while maintaining partial overlap in metabolite classes. These results confirm that the underlying biosynthetic pathways originated prior to subfamily divergence and were subsequently differentially expanded. Similar lineage-specific patterns were observed in Fabaceae, characterised by isoflavonoid enrichment, and in Brassicales, defined by glucosinolate diversity. To valorise the chemical diversity showed in the Rosid collection, high promising species, exhibiting a much high relative abundance of specific chemical classes compared to all the other species, were identified. In parallel, selected species were screened using three bioassays relevant to agricultural defence. Six species exhibited high bioactivity in antifeedant or antisettling assays against Spodoptera littoralis and Myzus persicae, respectively. In several cases, the observed bioactivity was consistent with previously reported compounds, whereas in other species the bioactivity had not been reported previously. Overall, this work demonstrates that large-scale metabolomic profiling combined with functional bioassays provides a robust framework for investigating plant chemical diversity and identifying bioactive species. The results indicate that metabolomic patterns are only partially aligned with phylogenetic relationships, reflecting a complex distribution of specialised metabolites across plant lineages. MPL, and consequently the Rosid collection, has proven itself as a valuable platform for the exploration of plant chemical diversity with potential applications.