Alignment of deduced amino-acid sequences and three-dimensional models of FDS and CDS. (A) Multiple sequence alignment of FDS and CDS. The amino-acid positions are numbered relative to CDS in P. cinerariifolium (I13995). The consensus sequence highlights the five conserved domains identified in FDS. Sites under positive selection identified by PAML are indicated by asterisks (*sites with p>80%, **sites with p>95%, ***sites with p>99%). Amino-acids involved in the binding and catalysis of substrates [29–31, 60] are marked by blue arrows. Red triangles show four sites involved in the binding and catalysis of substrates in FDS and mutated in CDS. Species name abbreviations: Aas, Achillea asiatica; Atr, Artemisia tridentata; Aag, Aster ageratoides; Can, Capsicum annuum; Cas, Centella asiatica; Cin, Cichorium intybus; Cla, Chrysanthemum lavandulifolium; Csc, Cynara scolymus; Glu,Gentiana lutea; Gan, Gerbera anandria; Han, Helianthus annuus; Lvu, Leucanthemum vulgare; Hsa, Homo sapiens (NP_001129294) (The human sequence is included for alignment since it was used as a template in the homology modeling); Mpi, Mentha x piperita; Pci, Pyrethrum cinerariaefolium; Pco Pyrethrum coccineum; Sly, Solanum lycopersicum; Tmo, Taraxacum mongolicum; Tof,Taraxacum officinale. (B) Three-dimensional model of FDS (PDB: 2F8Z). R351 and F239 are involved in IPP-binding in FDS. The red dashed lines represent hydrogen bonds. The yellow sphere indicates water. (C) Distribution of positively-selected sites in the homology model of CDS. CDS has a structural model very similar to FDS: the arrangement of 10 core helices around a large central cavity. The sites indentified as positively-selected (p>80%) were clustered in the large central cavity (shown in pink).