Comparison of transcriptional activity profiling by metabolic labeling or nuclear RNA sequencing

Abstract

The application of high-throughput sequencing to cellular transcriptome profiling (RNA-seq) has enabled significant advances in our understanding of gene expression in plants. However, conventional RNA-seq data reports mainly cytoplasmic transcript abundance rather than actual transcription rates. As a result, it is less sensitive to detect unstable and low-abundance nuclear RNA species, such as long non-coding RNAs, and is less directly connected to chromatin features and processes such as DNA replication. To bridge this gap, several protocols have been established to profile newly synthesized RNA in plants and other eukaryotes. These protocols can be technically challenging and present their own difficulties and limitations. Here we analyze newly synthesized nuclear RNA metabolically labeled in vivo with 5-ethynyl uridine (EU-nuclear RNA) in maize (Zea mays L.) root tips and compare it with the entire nuclear RNA population. We also compare both nuclear RNA preparations to conventional RNA-seq analysis of cellular RNA. The transcript abundance profiles of protein-coding genes in nuclear RNA and EU-nuclear RNA were tightly correlated with each other (R² = 0.767), but quite distinct from that of cellular RNA (R² = 0.170 or 0.293). Nuclear and EU-nuclear RNA reads are frequently mapped across entire genes, including introns, while cellular reads are predominantly mapped to mature transcripts. Both nuclear and EU-nuclear RNA exhibited a greater ability to detect both protein-coding and non-coding expressed genes.