MicroRNA biogenesis is broadly disrupted by inhibition of the splicing factor SF3B1

In animals, microRNA (miRNA) biogenesis begins with cotranscriptional cleavage of the primary (pri-)miRNA by Microprocessor. Cotranscriptional splicing has been shown to influence Microprocessor cleavage for miRNAs hosted in introns of protein coding pri-miRNAs, but the impact of splicing on production of miRNAs hosted in long noncoding (lnc)RNAs is largely unknown. Here, we investigated the role of splicing in the biogenesis of miR-122, a lncRNA-hosted, highly expressed, medically important, liver-specific miRNA. We found that splicing inhibition by the SF3B1 inhibitor Pladienolide B (PlaB) led to strong and rapid reduction in transcription of endogenous, but not plasmid-encoded, pri-miR-122, resulting in reduced production of mature miR-122. To allow detection of rapid changes in miRNA biogenesis despite the high stability of mature miRNAs, we used SLAMseq to globally quantify the effects of short-term splicing inhibition on miRNA synthesis. We observed an overall decrease in biogenesis of mature miRNAs following PlaB treatment. Surprisingly, miRNAs derived from different genomic locations were similarly affected. Together, this study provides new insight into the emerging role for splicing in transcription, demonstrating novel biological importance in promotion of miR-122 biogenesis from a lncRNA, and shows that splicing is important for global miRNA biogenesis. To understand the effects of splicing inhibition on miRNA biogenesis we treated Huh7 cells with SF3B1 inhibitor Pladienolide B (Pla B) or DMSO for 4h. To detect changes at this short time-point cells were metabolically labelled with 500 μM 4‑thiouridine (4SU). RNA isolated from the cells in each of the three treatment groups (Ctrl; no 4SU + DMSO, DMSO; 4SU + DMSO, Pla B; 4SU + Pla B) was alkylated using iodoacetamide. Libraries were generated from the alkylated RNA using the NEXTflex kit and were size selected after pooling. Following alkylation and library prep 4SU-modified positions were read as T>C conversions during sequencing. We analysed four biological repeats per treatment. Two samples (DMSO A and DMSO B) produced a low number of reads on first sequencing. To address this, new libraries were prepared for these two samples (re-prep), and existing libraries were also re-amplified (re-amp). Both re-prepped and re-amplified libraries were sequenced, and reads from the three runs (original, re-prepped and re-amplified) were combined for analysis.