U4 snRNP inhibits premature cleavage and polyadenylation of pre-mRNAs

Abstract

The essential role of U4 snRNP in pre-messenger RNA (mRNA) splicing has been well established. In this study, we utilized an antisense morpholino oligonucleotide (AMO) specifically targeting U4 snRNA to achieve functional knockdown of U4 snRNP in HeLa cells. Our results showed that this knockdown resulted in global intronic premature cleavage and polyadenylation (PCPA) events, comparable to the effects observed with U1 AMO treatment, as demonstrated by mRNA 3′-seq analysis. Furthermore, our study suggested that this may be a common phenomenon in both human and mouse cell lines. Additionally, we showed that U4 AMO treatment disrupted transcription elongation, as evidenced by chromatin immunoprecipitation sequencing (ChIP-seq) analysis for RNAPII. Collectively, our results identified a unique role for U4 snRNP in the inhibition of PCPA and indicated a model wherein splicing intrinsically inhibits intronic cleavage and polyadenylation in the context of cotranscriptional mRNA processing.

The phenomenon of U1 snRNP telescripting involves the inhibition of intronic premature cleavage and polyadenylation (PCPA) of numerous pre-messenger RNAs (mRNAs) by U1 snRNP, thereby promoting transcription elongation of thousands of eukaryotic genes (1–5). This finding emerged from the use of a specific antisense morpholino oligonucleotide (AMO) targeting the 5′-end of U1 snRNA (1). Functional depletion of U1 snRNP through the application of this U1 AMO has been shown to result in splicing defects and intronic PCPA of pre-mRNAs in vivo (1–3), and whether the role of U1 in splicing is involved in U1 snRNP telescripting mechanism remains unknown (5, 6).

Results and Discussion

We previously demonstrated that U1 AMO disrupts the structure of U1 snRNP, thereby promoting intronic PCPA of pre-mRNAs (7). To further explore the potential role of splicing in the U1 telescripting mechanism, we employed an AMO targeting U4 snRNA, a crucial core component of the spliceosome (8, 9) (Fig. 1A). Results from gel shift assays and in situ hybridization assays suggest that U4 AMO specifically targets U4 snRNA (Fig. 1 B and C). To further support the notion that U4 AMO can disrupt U4 snRNP function, an in vitro splicing assay was conducted using the well-established MINX-MS2 pre-mRNA substrate (10, 11). As expected, U4 AMO, like U1 AMO, markedly decreased splicing efficiency compared to the negative control AMO (Fig. 1D). These results suggest that the U4 AMO specifically inactivates U4 snRNP function in vitro and in vivo.

Fig. 1.

U4 snRNP inhibition by U4 AMO activates intronic PCPA events on a transcriptome-wide scale. (A) Description of the AMO sequences used in this study. (B) Gel mobility shift assay using [α–32P]UTP-labeled U1 or U4 snRNA and indicated AMOs. (C) In situ hybridization was performed on HeLa cells transfected with 10 µM AMO using a Cy3-labeled DNA probe to the U4 snRNA. Nuclei were visualized with DAPI and merged images are shown. (D) [α–32P]UTP-labeled MINX-MS2 pre-mRNA was spliced in vitro in the presence of control, U1, or U4 AMOs at various concentrations (1, 2, and 10 µM). Splicing product identities are depicted to the right of the gel. (E) Bar graph represents the numbers and types of PASs displaying significantly differential usage between control and U1/U4 AMO-treated HeLa cells. (F) Scatter plots depicting the numbers of genes showing mRNA APA shifts upon U1/U4 AMO treatment in comparison with control AMO. See details in SI Appendix. (G) Venn diagram showing the overlapping PASs/genes/APA events regulated by U1 and U4 snRNP in HeLa cells. (H) IGV track screen shots showing mRNA-seq and 3′-seq results for rexo4 gene in control and U1/U4 AMO-treated HeLa cells. (I) Metagene plots of RNAPII ChIP-Seq reads in control and U4 AMO-treated HeLa cells for all actively expressed genes. Y axis represents ChIP-seq enrichment/input in a log2 scale.

To further investigate the impact of U4 AMO on the transcriptome, total RNAs were isolated from HeLa cells transfected with U4 AMO. mRNA-seq analysis revealed a significant increase in intronic reads, indicating a role of U1/U4 in splicing (Dataset S1). Utilizing IRFinder-S software with defined parameters (Fold-change ≥ 2, FDR < 0.05) (12), we have identified a subset of genes showing statistically significant changes in intron retention following U1/U4 AMO transfection (Dataset S2). The analysis identified that 4,378 and 4,573 genes displayed the enrichment of one or more retained introns in the U1 AMO and U4 AMO transfected samples, respectively, with 3,480 genes exhibiting accumulation in both samples (Dataset S2). Utilizing the IGV genome browser, we observed a significant decrease in signal downstream of intronic cryptic PAS in U1 AMO-treated cells for numerous affected genes (Fig. 1H), aligning with its established function in suppressing PCPA. Notably, a similar trend was observed in multiple genes in U4 AMO-treated cells (Fig. 1H), warranting further exploration of U4’s impact on PCPA at the transcriptome level. To investigate this, we performed 3′-seq analysis on control, U1, and U4 AMO-treated samples using the Lexogen mRNA 3′-seq kit, which enables accurate quantification of global PAS usage (13). Utilizing our published QuantifyPoly(A) pipeline (14), it was observed that intronic PASs were significantly up-regulated following U1 AMO treatment, with 15,309 intronic PASs (Fold-change ≥ 2, FDR < 0.05) identified across 5,836 genes (Fig. 1E). These results are in line with previous research conducted by our team and others (1–3, 7, 15, 16). In contrast, the analysis revealed that 12,166 up-regulated intronic PASs were identified in U4 AMO-transfected samples, distributed among 5,227 genes (Fig. 1E). It is noteworthy that a significant proportion of these intronic PASs and genes exhibited coregulation by both U1 and U4, as evidenced by the presence of overlapping PASs and genes (Fig. 1G). To validate the accuracy of the 3′-seq/mRNA-seq data, a subset of 11 PCPA target genes was randomly selected for RT-qPCR analysis, which ultimately confirmed the initial findings (Dataset S3). Given the propensity for intronic PAS utilization induced by U1 AMO to cause widespread alterations in mRNA alternative polyadenylation (APA) profiles, particularly with regard to distal to proximal (D-to-P) APA shift, we proceeded to analyze APA profiles using the QuantifyPoly(A) pipeline. As expected, treatment with U4 AMO resulted in a significant APA modification in 6,252 genes, with 4,902 genes showing a D-to-P APA shift and 1,350 genes exhibiting a P-to-D APA shift (Fig. 1F). Similarly, treatment with U1 AMO led to a D-to-P APA shift in 5,596 genes and a P-to-D APA shift in 1,085 genes (Fig. 1F). Significantly, both cohorts displayed 4,350 genes that demonstrated a D-to-P APA shift compared to the NC AMO (Fig. 1G). Our results collectively demonstrate that the suppression of U4 snRNP by AMO leads to intronic PCPA events similar to those observed with U1 snRNP depletion in HeLa cells.

To explore the extent of intronic PCPA inhibition by U4, we performed 3′-seq analyses in additional human (HEK-293) and mouse (NIH3T3 and Neuro-2a) cell lines, with two biological replicates for each cell line. Utilizing the same experimental parameters and bioinformatics methodologies as previously mentioned, consistent outcomes were observed across all cell lines examined (Dataset S4). These results collectively indicate that the suppression of intronic PCPA by U4 may be a prevalent occurrence in mammalian cells.

Based on the association of the U1 telescripting mechanism with transcription regulation (3, 7), we hypothesized that U4 AMO may also disrupt transcription. To investigate this hypothesis, ChIP-seq analysis was conducted in HeLa cells using an antibody specific to the RNAPII CTD (C-terminal domain). Three biological replicates were performed for each sample, revealing a notable reduction in the overall ChIP signal across the gene body, as evidenced by RNAPII metagene plots of all expressed genes, including intronic PCPAed, non-PCPAed genes, and intronless genes (Fig. 1I) (Dataset S5). To confirm the ChIP-seq results, eleven regions were randomly selected for ChIP-qPCR analysis. The results aligned with the ChIP-seq data, validating its accuracy (Dataset S6). Further research is necessary to elucidate the connection between U4 AMO-induced transcription defect and PCPA.

Materials and Methods

AMOs were ordered from Gene Tools. AMO transfections were performed using a Bio-Rad Gene pulser at 960 μF and 280 V. After electroporation, cells were cultured for 12 h. Total RNA was extracted from cells using the TRIzol (Thermo Scientific). 3′-seq was conducted using the QuantSeq Rev 3′ mRNA-seq library prep kit (Lexogen, Cat. 016.24). The NovaSeq platform was utilized for sequencing 3′-seq/mRNA-seq/ChIP-seq libraries. We used our published bioinformatics pipeline for 3′-seq data analysis (Fig. 1 E–H) (14). See details in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Dataset S02 (XLSX)

Dataset S03 (XLSX)

Dataset S04 (XLSX)

Dataset S05 (XLSX)

Dataset S06 (XLSX)

Data, Materials, and Software Availability

Deep sequencing data have been deposited in GEO (GSE262385) (17). All other data are included in the manuscript and/or supporting information.