Transcriptome assemblies for two drug-type cannabis chemotypes by long-read RNA sequencing.

Abstract

Cannabis sativa has undergone over 10,000 years of domestication, resulting in extensive genetic and phenotypic diversity among cultivated chemotypes. Increased medical and recreational use of specialized metabolites accumulating in cannabis glandular trichomes-primarily the cannabinoids ∆9-tetrahydrocannabinol (THC) and cannabidiol (CBD)-has amplified research interest and a need to develop supporting genomic tools. Here, we present PacBio Iso-Seq-based transcriptome assemblies for two contrasting cannabis drug-type chemotypes (THC- and CBD-dominant) and their characterization. These assemblies encompass approximately 60% of the annotated loci in the cs10 reference genome, consistent with the commonly expressed fraction of the genome, and identify 1145 novel transcribed loci not present in the cs10 reference. Each assembly defines >50,000 transcripts and 15,000 alternative splicing events. Their accuracy is exemplified by confirming the conservation of alternative splicing events for Rubisco activase and a serine/argine-rich protein (SR45). We further highlight their utility by characterizing a novel cis-regulatory long non-coding RNA associated with the transcription factor NITRATE REGULATORY GENE 2. In addition, alternative splicing events for SPX DOMAIN 4, a key regulator of phosphate homeostasis, identified expression of transcripts encoding proteins with altered domain structure. Quantification of transcript abundances of these genes across different organs and varying phosphate supplies revealed isoform-specific expression patterns that differ between chemotypes, suggesting novel regulatory mechanisms for nitrogen and phosphate acquisition not previously described in either model or crop plant species. Our transcriptome assemblies provide a rich resource for the functional characterization of transcript and protein diversity in cannabis.