Transcriptome driven characterization of curly- and smooth-leafed endives reveals molecular differences in the sesquiterpenoid pathway

Endives (Cichorium endivia L.) are popular vegetables, diversified into curly/frisée- and smooth/broad-leafed (escaroles) cultivar types (cultigroups), and consumed as fresh and bagged salads. They are rich in sesquiterpene lactones (STL) that exert proven function on bitter taste and human health. The assembly of a reference transcriptome of 77,034 unigenes and RNA-sequencing experiments were carried out to characterize the differences between endives and escaroles at the gene structural and expression levels. A set of 3,177 SNPs distinguished smooth from curly cultivars, and a SNP-supported phylogenetic tree separated the cultigroups into two distinct clades, consistently with the botanical varieties of origin (crispum and latifolium, respectively). A pool of 699 genes maintained differential expression pattern (core-DEGs) in pairwise comparisons between curly vs smooth cultivars grown in the same environment. Accurate annotation allowed the identification of 26 genes in sesquiterpenoid biosynthesis pathway, which included several germacrene A synthase, germacrene A oxidase and costunolide synthase members (GAS/GAO/COS module), required for the synthesis of costunolide, a key precursor of lactucopicrin- and lactucin-like sesquiterpene lactones. The core-DEGs contained a GAS gene (contig83192) that was positively correlated with STL levels and recurrently more expressed in curly than smooth endives, suggesting a cultigroup-specific behavior. The significant positive correlation of GAS/GAO/COS transcription and STL abundance (2.4 fold higher in frisée endives) suggested that sesquiterpenoid pathway control occurs at the transcriptional level. Based on correlation analyses, five transcription factors (MYB and WRKY) were inferred to act on contig83192/GAS and specific STL, suggesting the occurrence of two distinct routes in STL biosynthesis.