The Exon Junction Complex Core Represses Cancer-Specific Mature mRNA Re-splicing: A Potential Key Role in Terminating Splicing

Abstract

Using TSG101 pre-mRNA, we previously discovered cancer-specific re-splicing of mature mRNA that generates aberrant transcripts/proteins. The fact that mRNA is aberrantly re-spliced in various cancer cells implies there must be an important mechanism to prevent deleterious re-splicing on the spliced mRNA in normal cells. We thus postulated that mRNA re-splicing is controlled by specific repressors, and we searched for repressor candidates by siRNA-based screening for mRNA re-splicing activity. We found that knock-down of EIF4A3, which is a core component of the exon junction complex (EJC), significantly promoted mRNA re-splicing. Remarkably, we could recapitulate cancer-specific mRNA re-splicing in normal cells by knock-down of any of the core EJC proteins, EIF4A3, MAGOH, or RBM8A (Y14), implicating the EJC core as the repressor of mRNA re-splicing often observed in cancer cells. We propose that the EJC core is a critical mRNA quality control factor to prevent over-splicing of mature mRNA.

Keywords: EIF4A3; EJC; MAGOH; RBM8A (Y14); TSG101; mRNA re-splicing; pre-mRNA splicing.

Figure 1

Identification of EJC core factors as repressor candidates of cancer-specific mRNA re-splicing: (a) The schematic structure of normally spliced TSG101 mRNA and cancer-specific re-spliced TSG101∆154–1054 mRNA. The primer sets for the detection are indicated with arrows. (b) To measure the generated normal TSG101 and re-spliced TSG101Δ154-1054 mRNAs, the extracted total RNAs were analyzed by RT–qPCR using the indicated specific primer sets in (a). To knock-down the expression of the endogenous EIF4A3 gene, MCF-7 cells were transfected with EIF4A3 siRNAs together with control siRNA. The cell extracts were then analyzed by RT–qPCR, and the relative amounts of EIF4A3 mRNA were plotted. The histograms represent the means ± standard deviations of three replicates (* p < 0.05). (c) To knock-down the expression of endogenous EJC core factor genes, MCF-7 cells were transfected with siRNA targeting EIF4A3, MAGOH, RBM8A, CASC3, or control siRNA (cont.). To observe the knocked-down mRNAs and generated normal TSG101 and re-spliced TSG101∆154–1054 mRNAs, the extracted total RNAs were analyzed by RT–PCR. The depletion of the targeted proteins was verified by immunoblotting (Figure S2).

Figure 2

Depletion of EJC core factors can induce cancer-specific mRNA re-splicing in non-cancerous MCF-10A cells: (a) Total RNAs extracted from MCF-10A or MCF-7 cells were analyzed by RT–PCR using specific primer sets (see Figure 1a). (b) To knock-down the expression of endogenous EJC core factor genes, MCF-10A cells were transfected with the indicated siRNAs. To observe knocked-down efficiency and TSG101 re-splicing activity, the extracted total RNAs were analyzed by RT–PCR. The depletion of the targeted proteins was verified by immunoblotting (Figure S2).

Figure 3

Activation of mRNA re-splicing by the knock-down of EJC core factors is not due to the repression of EJC-mediated NMD: (a) To knock-down the expression of endogenous NMD factor UPF1, MCF-10A cells were transfected with the indicated siRNAs. To observe the knocked-down mRNA and TSG101 re-splicing activity, the extracted total RNAs were analyzed by RT–PCR. MCF-7 cells (control) show the re-spliced mRNA that lacks in other lanes using MCF-10A cells. (b) HeLa cells were transfected with EIF4A3 or UPF1 siRNAs (or control siRNA). To evaluate the re-splicing activity with TSG101 mRNA and the NMD efficiency with GAS5 mRNA, the extracted total RNAs were analyzed by RT–qPCR using specific primer sets. The histograms represent the means ± standard deviations of three replicates (* p < 0.05, n.s. = not statistically significant p > 0.05).

Figure 4

Depletion of EJC peripheral factors, including an mRNA export factor, have no significant effects on mRNA re-splicing in MCF-7 cells: (a,b) To knock-down the expression of endogenous EJC peripheral factors (NXF1, ACIN1, PNN, RNPS1, SAP18), MCF-7 cells were transfected with the indicated siRNAs. To observe the knocked-down efficiency and TSG101 re-splicing activity, the extracted total RNAs were analyzed by RT–PCR.

Figure 5

RBM8A protein expression is lower in cancer cells compared with that in non-cancerous cells, and RBM8A overexpression represses mRNA re-splicing in cancer cells: (a) Whole-cell extracts from MCF-10A and MCF-7 cells were analyzed by immunoblotting with the indicated specific antibody. GAPDH was used as an endogenous control. (b) RBM8A protein expression was normalized to that of GAPDH and quantified by densitometry. The histograms represent the means ± standard deviations of three replicates (* p < 0.05). (c) The stable MCF-7 cell lines expressing control and RBM8A proteins (pSB-DsRed and pSB-RBM8A) were cultured with or without doxycycline (DOX) for 48 h. The total RNAs were analyzed by RT–PCR using the indicated primer sets. (d) To evaluate the re-splicing activity with TSG101 mRNA, the total RNAs were analyzed by RT-qPCR using specific primer sets. The histograms represent the means ± standard deviations of three replicates (* p < 0.05, n.s. p > 0.05).

Figure 6

Model of the EJC-triggered pathway leading to the termination of splicing. The spliced mRNA was shown to form a packed and compacted mRNP structure by binding between EJC cores and among EJC and SR proteins (EJC-SR nexus) together with other RNA-binding proteins (RBPs) [20,21]. We assume that splicing-dependent binding of EJC core to mature mRNA is essential to initiate the formation of this highly compacted mRNP structure.