Copy number variations shape the structural diversity of Arabidopsis metabolic gene clusters and are associated with the climatic gradient

Metabolic gene clusters (MGCs) encode at least three different enzymes for a common biosynthetic pathway. Comparative genome analyses highlighted the role of duplications, deletions and rearrangements in MGC formation. We hypothesized that these mechanisms also contribute to MGC intraspecies diversity and play a role in adaptation. We assessed copy number variations (CNVs) of four Arabidopsis thaliana MGCs in a population of 1,152 accessions, with experimental and bioinformatic approaches. The MGC diversity was lowest in marneral gene cluster (one private deletion CNV) and highest in the arabidiol/baruol gene cluster where 811 accessions had gene gains or losses, however, there were no presence/absence variations of the entire clusters. We found that the compact version of thalianol gene cluster was predominant in A. thaliana and more conserved than the noncontiguogus version. In arabidiol/baruol cluster we found a large insertion in 35% of analyzed accessions, that contained duplications of the reference genes CYP705A2 and BARS1. The BARS1 paralog, which we named BARS2, encoded a novel oxidosqualene synthase. Unexpectedly, in accessions with the insertion, the arabidiol/baruol gene cluster was expressed not only in roots but also in leaves. Additionally, they presented different root growth dynamics and were associated with warmer climates compared to the reference-like accessions. We also found that paired genes encoding terpene synthases and cytochrome P450 oxidases had higher copy number variability compared to non-paired ones. Our study highlights the importance of intraspecies variation and nonreference genomes for dissecting secondary metabolite biosynthesis pathways and understanding their role in adaptation and evolution.