mRNA splicing is modulated by intronic microRNAs

Abstract

Splicing of transcripts is catalyzed by the spliceosome, a mega-complex consisting of hundreds of proteins and five snRNAs, which employs direct interactions. When U1 snRNA forms high-affinity binding, namely more than eight base pairs, with the 5'SS, the result is usually a suppressing effect on the splicing activity. This likely occurs due to the inefficient unwinding of U1/5'SS base-pairing or other regulatory obstructions. Here, we show in vitro and in patient-derived cell lines that pre-microRNAs can modulate the splicing reaction by interacting with U1 snRNA. This leads to reduced binding affinity to the 5'SS, and hence promotes the inclusion of exons containing 5'SS, despite sequence-based high affinity to U1. Application of the mechanism resulted in correction of the splicing defect in the disease-causing VCAN gene from an individual with Wagner syndrome. This pre-miRNA/U1 interaction can regulate the expression of alternatively spliced exons, thus extending the scope of mechanisms regulating splicing.

Keywords: Cell biology; Functional aspects of cell biology; Molecular biology.

Graphical abstract

Figure 1

Pre-miR-211-U1 snRNA binding promotes splicing (A) Binding of pre-miRNAs to the U1 snRNA. Score distribution (using an EMBL-EBL tool⁴¹) of the pre-miRNAs that occur once in an up to 10Kb intron. The higher the score, the stronger the predicted binding between a pre-miRNA and the U1 snRNA. The red dot represents pre-miR-211, the yellow dot represents pre-let-7g, and the orange dot represents pre-miR-7-1. (B) U1-5′SS score sequence alignment.^, Base-pairing miss-matches between the U1 snRNA and the splice sites are denoted in red bold letters. (C) RT-qPCR analysis of GFP expression following 24 h of overexpression in the HEK 293T cell line. Fold changes of spliced GFP products normalized to the 100 5′SS score. In the absence of primary miR-211, GFP mini-genes showed the most efficient splicing at the 95 5′SS score. The paired Student’s t test was used for the statistical analysis (n = 3, ∗∗p < 0.005). Data are represented as mean ± SEM. (D) As in C, with miR-211 or control plasmid overexpressed in trans. For miR-211 overexpressed, GFP spliced product expression was directly correlated with the 5′SS score and decreased as the 5′SS score decreased. Overexpression of the control plasmid was like in C. The paired Student’s t test was used for the statistical analysis (n = 3, ∗p < 0.05, ∗∗p < 0.005). Data are represented as mean ± SEM. (E) As in C, with miR-211 or a control plasmid overexpressed in cis (for both transient and stable transfections). GFP spliced product expression was directly correlated with the 5′SS score and decreased as the 5′SS score decreased. The paired Student’s t test was used for the statistical analysis (n = 3, ∗∗p < 0.005). Data are represented as mean ± SEM.

Figure 2

pre-miR-7-1/pre-let-7g or mut-pre-miR-7-1 overexpression binding affinities to U1 snRNA and the effect on the splicing mechanism (A) As in Figure 1E, transient transfection when pre-let-7g replaced pre-miR-211 in the intron. GFP spliced product expression does not depend on pre-let-7g overexpression and GFP expression is the highest with the 95 5′SS score. The paired Student’s t test was used for the statistical analysis (n = 3, ∗∗p < 0.005). Data are represented as mean ± SEM. (B) As in Figure 1E, transient transfection when pre-miR-7 replaced pre-miR-211 in the intron. GFP spliced product expression does not depend on pre-miR-7 overexpression and GFP expression is the highest with the 95 5′SS score. The paired Student’s t test was used for the statistical analysis (n = 3, ∗p < 0.05, ∗∗p < 0.005). Data are represented as mean ± SEM. (C) As in Figure 1E, transient transfection when mut-pre-miR-7 replaces pre-miR-211 in the intron. GFP spliced product expression is directly correlated with the 5′SS score and decreased as the 5′SS score decreased. The paired Student’s t test was used for the statistical analysis (n = 3, ∗p < 0.05, ∗∗p < 0.005). Data are represented as mean ± SEM. (D) Streptavidin pull-down of biotinylated-labeled RNA. The fold change of U1 recovery represents the amount of the U1 snRNA recovered by the streptavidin pull-down assay compared to the quantity of the U1 snRNA loaded on the magnetic beads. U1 recovery was significantly higher following hybridization of U1 snRNA with pre-miR-211 or mut-pre-miR-7-1 than with WT pre-miR-7-1. Therefore, U1 snRNA binding to pre-miR-211 and mut-pre-miR-7-1 was stronger than to WT pre-miR-7-1. The paired Student’s t test was used for the statistical analysis (n = 3, ∗∗p < 0.005). Data are represented as mean ± SEM.

Figure 3

miR-211-U1 interaction corrects the VCAN splicing deficiency in fibroblasts derived from patients suffering from Wagner syndrome (A) Enrichment plot of miR-211 target genes. The top portion of the plot shows where the target gene set appeared in the ranked list of genes. The bottom portion of the plot shows the value of the running enrichment score. The enrichment score is calculated by walking down the ranked list of genes, increasing a running-sum statistic when a gene is in the gene set and decreasing it when it is not. The magnitude of the increment depends on the correlation of the gene with the phenotype. The enrichment score is the maximum deviation from zero encountered in walking the list (designated by the dashed red line). A negative enrichment score indicates gene set enrichment at the bottom of the ranked list. (B) Histogram of a 5′SS score of 24,860 exons (excluding the first exons and UTRs). The mean 5′SS score is 81, with only 168 (<1%) exons harboring a score of 100 (median score = 81.87, ISQ = 11.04, SD = 8.2). (C) Boxplots of the mean differences in exon splicing events (SE) between overexpression and control experimental pairs by 5′SS score quantiles, and the 100 score. Means are designated by a red dot. The mean of the 100 5′SS group is significantly greater than the mean of each quantile group and all the other scores combined. The mean of 10,000 random permutations is designated by a blue cross (X). (D) The VCAN E7-E8:E6-E8 expression ratio was analyzed by RT-qPCR in fibroblasts derived from individuals with Wagner syndrome. Exon 7 expression was significantly higher in the mutant cells versus the control cells for miR-211 or miR-182 OE. The paired Student’s t test was used for the statistical analysis (n = 3, ∗p < 0.05, ∗∗p < 0.005). Data are represented as mean ± SEM.

Figure 4

A suggested miRNA-splicing regulation model based on our results (A–D) Pre-miRNA-U1 snRNA binding affinity. A high base match score reflects, as indicated previously, a strong miR-U1 interaction. (A) miR-211, (B) let-7g, (C) miR-7, (D) mut-miR-7.