• candidate genes
• genetic variability
• IRAP
• LTR-Retrotransposons
• phenolic compounds
• RNA-seq
• Stevia rebaudiana
• steviol glycosides

Stevia rebaudiana Bert. is a promising crop, belonging to the Asteraceae family, especially given the content of the steviol glycosides (SGs) and phenolic compounds (PCs) present in its leaves. SGs are calorie-free sweeteners, 100 times more potent than sucrose, and have antibacterial, antioxidant, and antidiabetic properties. To date, 64 SGs have been found in stevia leaf extracts at various concentrations; the most prevalent are stevioside and rebaudioside A, followed by rebaudioside B-E, dulcoside A, and steviolbioside. Similarly, PCs have attracted interest due to their positive effects on health and numerous biological functions. More than 30 PCs found in stevia leaves have been characterised. The main classes of phenols include flavonoids, chlorogenic acids, and a polyphenol family of esters that includes hydroxycinnamic acids and quinic acid. S. rebaudiana is a diploid species (2n = 22), and the genome comprising 1416 Mb was sequenced using a combination of PacBio and Hi-C, two cutting-edge sequencing techniques. Repetitive elements account for 80.11% of the stevia genome. This Thesis aimed to characterise the stevia genome's repetitive elements and the stevia germplasm's genetic and phytochemical diversity to identify candidate genes for the biosynthesis of SGs and PCs that are crucial for future breeding programs.
In the first section of the Thesis, the stevia genome sequence was used to identify and characterise full-length long terminal repeats retrotransposons (LTR-REs), which constitute the majority of the genome. Then the genetic diversity of 25 S. rebaudiana accessions from the Department of Agriculture, Food, and Environment (DAFE) at University of Pisa was investigated with the Inter-Retrotransposon Amplified Polymorphism (IRAP) protocol exploiting a collection of isolated full-length elements. Retrotransposons from Gypsy superfamily were twice reduntant than Copia and contained a high disproportion of Chromovirus/Tekay lineage elements. A significant portion of the genome was covered by the SIRE and Angela lineages from Copia elements. All isolated full-length elements were relatively recent, with a proliferation peak occurring between 1-2 million years ago. A comparison of the retrotranscriptase sequences revealed the existence of an earlier time frame during which LTR-RE proliferation was significant. The genomic structures of the various accessions were examined using the IRAP fingerprints, which revealed a sizable amount of genetic variation. The four ancestral subpopulations from which the examined accessions are part of, turned out to be mixed populations.
In the second section of the Thesis, 11 stevia genotypes from various origins belonging to the DAFE collection, that were part of a 3-year field trial arranged in a randomised block experimental design, were characterised at biochemical level. The 3-years old genotypes were examined after 3-year field trial at two harvesting times: the vegetative phase (July 2020) and the pre-flowering stage (September-October 2020). The amount of specific SGs and PCs was determined using HPLC, while the amounts of total phenols, total flavonoids, DPPH (2,2′-diphenyl-1-picrylhydrazyl), and FRAP (Ferric Reducing Antioxidant Power) were analysed using spectrophotometric assays. Although many genes play important roles in the biosynthesis of SGs and PCs, their contribution to the biosynthesis of these specific compounds has been poorly investigated in stevia. RNA-Seq has been successfully used to find candidate genes encoding enzymes and TFs that may be involved in SGs biosynthesis in stevia and PCs in other species. In these transcriptomic analyses a single pairwise comparison between three genotypes with high and three with low SGs and PCs content, selected among the 11 genotypes from DAFE collection, was performed. A total of 2997 differentially expressed genes (DEGs) were identified. Numerous known genes participating in SGs biosynthesis were found, including 2-C-methyl-D-erythritol 4-phosphate cytidylyltransferase (CMS), kaurene synthase (KS), UGT85C2, and UGT91D2. In addition, it was found that the genotypes with high PCs content had several overexpressed genes involved in monolignol biosynthesis. The weighted gene co-expression network analysis (WGCNA) also identified genes that are positively correlated with a number of biochemical traits. Particularly, UGT85A8 and UGT76H1 were putatively involved in PCs glycosylation and specific SGs biosynthesis, including rebaudioside E. The biosynthesis of specific SGs and PCs may be regulated by various TFs, including WRKY65, MYB12, and MYB30.
In conclusion, this work of Thesis identified and described the LTR-REs of S. rebaudiana and confirmed the viability of utilising molecular markers based on repeated sequences in this species. In addition, a wealth of putative candidate genes involved in the biosynthesis of specific SGs and PCs were proposed. Overall, these data could help guide future stevia breeding programs.