The competition between splicing and 3' processing shapes the human transcriptome

Eukaryotic pre-mRNA processing steps, including splicing and 3' processing, are tightly coordinated, but the underlying mechanisms remain poorly understood. Previous studies proposed that the splicing factor U1 snRNP inhibits 3' processing at intronic polyadenylation (IPA) sites through a splicing-independent mechanism, called telescripting. However, we found that global or gene-specific perturbation of splicing by targeting multiple splicing factors, including U1 snRNP, U2 snRNP, U2AF, and SF3b led to activation of 3' processing at IPA sites. Inhibiting different splicing factors activated overlapping and distinct IPA sites and such specificity was determined, at least in part, by alterations in RNA polymerase II elongation and termination. Conversely, we showed that blocking pre-mRNA 3' processing promoted splicing globally. These results strongly suggest that splicing and 3' processing are competing processes that shape the transcriptome. Finally, as splicing inhibition-induced shifts to IPA site usage can lead to gene inactivation, including tumor suppressor genes, the use of general splicing inhibitors to treat human diseases may pose a significant risk. Overall design: Examination of 30 samples by PAS-seq, a type of 3' end mRNA-sequencing. There are 3 biological replicates of 50uM control AMO treatment, 3 biological replicates of 50uM U2 AMO treatment, 3 biological replicates of control shRNA treatment, 3 biological replicates of U2AF1 shRNA treatment, 3 biological replicates of U2AF2 shRNA treatment, 3 biological replicates of DMSO treatment, 3 biological replicates of OTS964 (OTS) treatment, 3 biological replicates of Pladienolide B (PB) treatment, 3 biological replicates of 25uM control AMO treatment, and 3 biological replicates of 25uM U1 AMO treatment. All samples were prepared from HEK293T cells. Examination of 42 samples by 4sU-seq. There are 3 biological replicates of 50uM control AMO treatment, 3 biological replicates of 50uM U2 AMO treatment, 3 biological replicates of control shRNA treatment, 3 biological replicates of U2AF1 shRNA treatment, 3 biological replicates of U2AF2 shRNA treatment, 3 biological replicates of DMSO treatment, 3 biological replicates of OTS964 (OTS) treatment, 3 biological replicates of Pladienolide B (PB) treatment, 3 biological replicates of 25uM control AMO treatment, and 3 biological replicates of 25uM U1 AMO treatment. All samples were prepared from wild-type HEK293T cells. In addition, we generated a HEK293T cell line in which a FKBP12 F36V-degron was fused to the N-terminus of INTS11 via CRISPR/Cas9. We performed 4sU-seq in this cell line following treatment with 25 µM control AMO + DMSO (3 replicates), 25 µM control AMO + dTAG to induce INTS11 degradation (3 replicates), 25 µM U1 AMO + DMSO (3 replicates), and 25 µM U1 AMO + dTAG to induce INTS11 degradation (3 replicates). Examination of 41 samples by mRNA-seq. There are 3 biological replicates of 50uM control AMO treatment, 3 biological replicates of 50uM U2 AMO treatment, 3 biological replicates of control shRNA treatment, 3 biological replicates of U2AF1 shRNA treatment, 3 biological replicates of U2AF2 shRNA treatment, 3 biological replicates of DMSO treatment, 3 biological replicates of OTS964 (OTS) treatment, 3 biological replicates of Pladienolide B (PB) treatment, 3 biological replicates of 25uM control AMO treatment, and 2 biological replicates of 25uM U1 AMO treatment. All samples were prepared from wild-type HEK293T cells. In addition, we generated a HEK293T cell line in which a FKBP12 F36V-degron was fused to the N-terminus of INTS11 via CRISPR/Cas9. We performed mRNA-seq in this cell line following treatment with 25 µM control AMO + DMSO (3 replicates), 25 µM control AMO + dTAG to induce INTS11 degradation (3 replicates), 25 µM U1 AMO + DMSO (3 replicates), and 25 µM U1 AMO + dTAG to induce INTS11 degradation (3 replicates).