Gene clustering and copy number variation in alkaloid metabolic pathways of opium poppy

Abstract

Genes in plant secondary metabolic pathways enable biosynthesis of a range of medically and industrially important compounds, and are often clustered on chromosomes. Here, we study genomic clustering in the benzylisoquinoline alkaloid (BIA) pathway in opium poppy (Papaver somniferum), exploring relationships between gene expression, copy number variation, and metabolite production. We use Hi-C to improve the existing draft genome assembly, yielding chromosome-scale scaffolds that include 35 previously unanchored BIA genes. We find that co-expression of BIA genes increases within clusters and identify candidates with unknown function based on clustering and covariation in expression and alkaloid production. Copy number variation in critical BIA genes correlates with stark differences in alkaloid production, linking noscapine production with an 11-gene deletion, and increased thebaine/decreased morphine production with deletion of a T6ODM cluster. Our results show that the opium poppy genome is still dynamically evolving in ways that contribute to medically and industrially important phenotypes.

Chromosome-level genome assembly of Ophiorrhiza pumila reveals the evolution of camptothecin biosynthesis

“…One possible explanation for the positive selection of gene clusters is the reduced rate of recombination between genes involved in local adaptation 55 , 59 . Positive selection of gene clusters does have a possible role in providing chemotypes that may facilitate ecological/local success for a species or cultivar for successful propagation, as was reported for opium poppy 56 . The conserved nature and collinearity of metabolic gene clusters of the Ophiorrhiza genome across MIA-producing plant species suggest a potential means to select genes for functional studies.…”

“…Benzylisoquinoline alkaloid biosynthesis is one of the best-known specialized metabolic pathways, with enzymes forming gene clusters within the opium poppy genome. Nevertheless, the nature of gene clustering was reported to be heterogeneous, with the thebaine and noscapine pathways being highly clustered, and the morphine and sanguinarine pathways being scattered 56 . These results suggest the possibility of the active evolution of genome architecture through a combination of natural and artificial selection for specialized metabolite biosynthesis centered at gene clusters.…”

PR10/Bet v1‐like Proteins as Novel Contributors to Plant Biochemical Diversity

“…In contrast to NCS and THS where genomic clustering with BIA biosynthetic genes was evident, a similar arrangement was not found for NISO. NISO genes are within several hundred kb of other PR10/Bet v1‐like genes and approximately 25 Mb from the reticuline 7‐ O ‐methyltransferse ( 7OMT ) gene [101] . Nevertheless, the abundance of NISO transcripts within the laticifer cells is consistent with the expression profile of COR and other morphine biosynthetic genes.…”

“…With focused effort, PR10/Bet v1‐like proteins are likely to join the cadre of protein families for which the molecular evolution of catalytic lineages from a noncatalytic ancestor is understood [157,158] . The recent expansion of transcriptomic and genomic resources has made available the full complement of PR10/Bet v1‐like proteins from a wide range of plant species [57,101,159–161] . By characterizing phylogenetic interrelationships, as well as the catalytic properties of extant and resurrected ancestral members, the evolutionary trajectories which opened the gateway to benzylisoquinoline and amaryllidaceae alkaloid biosynthesis can be revealed.…”

CRISPR‐based assessment of genomic structure in the conserved SQUAMOSA promoter‐binding‐like gene clusters in rice

“…Gene clustering is defined here as the relatively close spatial proximity of specific genes to each other on a chromosome, in contrast to singleton genes, which are spatially far apart on a chromosome or on separate chromosomes. Clustering in plant genes has been shown to exist at distances of up to 10 Mb and may predict coexpression or other types of interaction within a network (Li et al, 2020). SPL genes in particular have been noted previously to include four clusters containing 10 genes (Xie et al, 2006;Zhang et al, 2015).…”