Integrator enforces the fidelity of transcriptional termination at protein-coding genes

Abstract

Integrator regulates the 3′-end processing and termination of multiple classes of noncoding RNAs. Depletion of INTS11, the catalytic subunit of Integrator, or ectopic expression of its catalytic dead enzyme impairs the 3′-end processing and termination of a set of protein-coding transcripts termed Integrator-regulated termination (IRT) genes. This defect is manifested by increased RNA polymerase II (RNAPII) readthrough and occupancy of serine-2 phosphorylated RNAPII, de novo trimethylation of lysine-36 on histone H3, and a compensatory elevation of the cleavage and polyadenylation (CPA) complex beyond the canonical polyadenylation sites. 3′ RNA sequencing reveals that proximal polyadenylation site usage relies on the endonuclease activity of INTS11. The DNA sequence encompassing the transcription end sites of IRT genes features downstream polyadenylation motifs and an enrichment of GC content that permits the formation of secondary structures within the 3′UTR. Together, this study identifies a subset of protein-coding transcripts whose 3′ end processing requires the Integrator complex.

Fig. 1.. Integrator regulates transcriptional termination at protein-coding genes.

(A) HMM algorithm used to analyze PRO-seq data at the 3′ end of genes and define IRT genes. (B) Example of the IRT gene DELE1 with the HMM predicted extension below. (C) PRO-seq heatmaps and mean density ratio spanning the poly(A) site to the HMM extension termination of 1315 IRT gene transcripts in control and INTS11-depleted cells. (D) Boxplot of the HMM predicted extension length of 1315 IRT gene transcripts in control and INTS11-depleted cells. (E) Cumulative distribution of eCLIP signal ratio log₂(INTS11/INPUT) at TES (−/+100) of control (blue) and extended genes (red). The significance between the distribution was calculated using the Kolmogorov-Smirnov (KS) test. The control genes (n = 1315) used for this analysis have similar expression levels of those observed in the extended genes. (F) PRO-seq Genome Browser example of the IRT gene OAZ1 at the 3′ end in control and INTS11-depleted cells. (G) PRO-seq Genome Browser example of the IRT gene DELE1 at the 3′ end in control and INTS11-depleted cells.

Fig. 2.. Integrator depletion induces transcriptional readthrough at protein-coding genes.

(A) Average profile of RNAPII spanning the poly(A) site to the HMM extension termination of 1315 IRT genes in INTS11-depleted cells. (B) Average profile of pSer²-RNAPII spanning the poly(A) site to the HMM extension termination of 1315 IRT genes in INTS11-depleted cells. (C) Average profile of CSTF64 spanning the poly(A) site plus additional 10-kb downstream from N- TES. Profile at IRT genes without and with INTS11 shRNA induction. (D) Average profile of CPSF73 spanning the poly(A) site plus additional 10-kb downstream from TES. Profile at IRT genes without and with INTS11 shRNA induction. (E) Boxplot of IRT gene groups separated by PRO-seq extension strength. (F) Boxplot of the extension length ratio in IRT gene groups separated by PRO-seq extension strength. (G) Boxplot of the RNAPII extension strength in IRT gene groups separated by PRO-seq extension strength. (H) Boxplot of the pSer²-RNAPII extension strength in IRT gene groups separated by PRO-seq extension strength.

Fig. 3.. Steady-state transcripts undergo 3′ end extension following the loss of Integrator.

(A and B) RNA-seq heatmaps and mean density ratio spanning the poly(A) site to the HMM extension termination of 1315 IRT gene transcripts in (A) control and (B) INTS11-depleted cells. (C) Volcano plot of differentially expressed IRT genes (n = 1315 genes, fold change = 1.5, q < 0.05) after INTS11 depletion. (D) Boxplot of down-regulated IRT genes (q < 0.05) separated by PRO-seq extension strength.

Fig. 4.. Defective termination is accompanied by de novo deposition of H3K36me3.

(A) Average profile of H3K36me3 spanning the poly(A) site to the HMM extension termination of 1315 IRT genes in INTS11-depleted cells. (B) Average profile of H3K36me3 in IRT gene groups separated by PRO-seq extension strength spanning the poly(A) site to the HMM extension termination in INTS11-depleted cells. (C) Average profile of H3K36me2 in IRT gene groups separated by PRO-seq extension strength spanning the poly(A) site to the HMM extension termination in INTS11-depleted cells. (D) Mean density ratio of H3K36me3 in the extended regions of the top 100 genes ranked by differential H3K36me3 signal. The top plot denotes the differential expression status of each gene. The bottom plot displays the PRO-seq extension strength of each gene. (E) H3K36me3 Genome Browser example of the IRT gene DELE1 in INTS11-depleted cells. (F) H3K36me3 Genome Browser example of the IRT gene OAZ1 in INTS11-depleted cells.

Fig. 5.. The catalytic activity of INTS11 is required for transcriptional termination at protein-coding genes.

(A and B) PRO-seq heatmaps and mean density ratio spanning the poly(A) site to the HMM extension termination of 1315 IRT gene transcripts in (A) WT INTS11 and (B) E203Q INTS11 rescue cells. (C) PRO-seq Genome Browser example of the IRT gene OAZ1 at the 3′ end in WT-INTS11 and E203Q-INTS11 rescue cells. (D) PRO-seq Genome Browser example of the IRT gene DELE1 at the 3′ end in WT INTS11 and E203Q INTS11 rescue cells. (E and F) RNA-seq heatmaps and mean density ratio spanning the poly(A) site to the HMM extension termination of 1315 IRT gene transcripts in (E) WT INTS11 and (F) E203Q INTS11 rescue cells.

Fig. 6.. INTS11 catalyzes the 3′ end cleavage of protein-coding transcripts at canonical termination sites.

(A) 3′ RNA-seq Genome Browser example of the IRT gene DELE1 in WT INTS11 and E203Q INTS11 rescue cells. (B) CM of canonical versus downstream termination site usage in WT INTS11 rescue cells. (C) CM of canonical versus downstream termination site usage in E203Q INTS11 rescue cells. (D) Motif analysis at canonical 3′ end and downstream extended transcript 3′ end. A window of −/+150 nt from the peak was used for the analysis; top 5 motifs were shown. (E) CPSF73 ChIP-seq profiles at 3′ end in the downstream region of IRT genes. Profiles without (blue line) and with INTS11 shRNA induction (red line). (F) CSTF64 ChIP-seq profiles at 3′ end in the downstream region of IRT genes. Profiles without (blue line) and with INTS11 shRNA induction (red line).

Fig. 7.. Model depicting the role of Integrator in the termination of pre-mRNA transcripts.

IRT transcripts are enriched in APA sequence AAUAUA motifs. In the absence of Integrator, RNAPII moves beyond the proximal PAS sites and the extended transcripts are cleaved at multiple downstream sites.