SRSF6 Regulates the Alternative Splicing of the Apoptotic Fas Gene by Targeting a Novel RNA Sequence

Abstract

Alternative splicing (AS) is a procedure during gene expression that allows the production of multiple mRNAs from a single gene, leading to a larger number of proteins with various functions. The alternative splicing (AS) of Fas (Apo-1/CD95) pre-mRNA can generate membrane-bound or soluble isoforms with pro-apoptotic and anti-apoptotic functions. SRSF6, a member of the Serine/Arginine-rich protein family, plays essential roles in both constitutive and alternative splicing. Here, we identified SRSF6 as an important regulatory protein in Fas AS. The cassette exon inclusion of Fas was decreased by SRSF6-targeting shRNA treatment, but increased by SRSF6 overexpression. The deletion and substitution mutagenesis of the Fas minigene demonstrated that the UGCCAA sequence in the cassette exon of the Fas gene causes the functional disruption of SRSF6, indicating that these sequences are essential for SRSF6 function in Fas splicing. In addition, biotin-labeled RNA-pulldown and immunoblotting analysis showed that SRSF6 interacted with these RNA sequences. Mutagenesis in the splice-site strength alteration demonstrated that the 5' splice-site, but not the 3' splice-site, was required for the SRSF6 regulation of Fas pre-mRNA. In addition, a large-scale RNA-seq analysis using GTEX and TCGA indicated that while SRSF6 expression was correlated with Fas expression in normal tissues, the correlation was disrupted in tumors. Furthermore, high SRSF6 expression was linked to the high expression of pro-apoptotic and immune activation genes. Therefore, we identified a novel RNA target with 5' splice-site dependence of SRSF6 in Fas pre-mRNA splicing, and a correlation between SRSF6 and Fas expression.

Keywords: Fas; SRSF6; alternative splicing.

Figure 1

SRSF6 regulates the alternative splicing of Fas pre-mRNA. (A) (Upper) Schematic representation of the Fas minigene. Cassette exons are indicated by the gray boxes. Flanking exons are indicated by empty boxes. Introns are indicated by lines. Vector sequences of the minigene are indicated by arcs. Primers used in RT-PCR are indicated by arrows. (Middle) RT-PCR analysis of minigenes with RNAs extracted from untreated, control vector-treated (pcDNA3.1) (mock), and SRSF6-overexpressed HEK293T cells and HCT116 cells. GAPDH was used as a loading control. Immunoblotting with anti-SRSF6 antibody for these cells was also performed with α-tubulin as a loading control. (Lower) Statistical analysis graphs of RT-PCR with p values. Standard deviations (SD) calculated from three independent experiments are indicated by error bars: ** p < 0.01, * p < 0.05. (B) (Upper) Schematic representation of genomic Fas gene and its alternative splicing products. Exons and introns are indicated by boxes and lines, respectively. Primers used for RT-PCR are indicated by arrows. (Middle) RT-PCR analysis with RNAs extracted from untreated, non-silencing shRNA-treated, SRSF6-shRNA-treated HEK293T, and HCT116 cells with GAPDH as a loading control. RT-PCR and immunoblotting analysis of these cells were also performed with GAPDH and α-tubulin as the loading controls. (Lower) Statistical analysis graphs of RT-PCR with p values. Uncropped figures are shown in Supplementary Figure S2.

Figure 2

SRSF6 contacts a novel RNA sequence to promote cassette exon inclusion. (A) Sequences of wild type, deleted (ΔSRSF6), and substituted mutant (M1 and M2) minigenes are shown. Deleted RNA sequences are indicated in green. Substituted nucleotides are indicated by red. (B) (Upper) RT-PCR analysis of mutant minigenes with RNAs from untreated, empty vector-treated, and SRSF6-overexpressed HEK293T cells. (Lower) Statistical analysis graphs of RT-PCR with p values. (C) (Upper) Chemically synthesized biotin-labeled ΔSRSF6, M1 and + (positive control) RNA sequences. (Lower) RNA-pulldown and immunoblotting analysis of the biotin-labeled RNAs and anti-SRSF6 antibody. Uncropped figures are shown in Supplementary Figure S3.

Figure 3

5′ splice-site (5′SS), but not 3′ splice-site (3′SS), strength affects SRSF6 function on Fas pre-mRNA splicing. (A) (Upper) Schematic representation of the 5′W minigene. Mutated sequences are shown in red. (Middle) RT-PCR analysis of the 5′W minigene with RNAs from untreated, empty vector-treated, and SRSF6-overexpressed cells with GAPDH as a loading control. Immunoblotting of these cells with anti-SRSF6 antibody was also performed with α-tubulin as a loading control. (Lower) Statistical analysis graphs of RT-PCR with p values. (B) (Upper) Schematic representations of WT, 5-5, and 6-6 minigenes. Stronger and weaker 3′ splice-sites are indicated by upper and lower arrows, respectively. (Middle) RT-PCR analysis 5-5 and 6-6 minigenes with RNAs from untreated, empty vector-treated and SRSF6-overexpressed cells. (Lower) Statistical analysis graphs of RT-PCR with p values. Standard deviations (SD) calculated from three independent experiments are indicated by error bars: ** p < 0.01, ns p > 0.05. Uncropped figures are shown in Supplementary Figure S3.

Figure 4

The co-expression landscape of SRSF6 and Fas genes varies between normal and tumor transcriptomes. Based on SRSF6 and FAS gene expression (TPM units), we explored their co-expression landscape in (A) normal colon samples from the GTEx database, and (B) tumor samples of colon cancer from the TCGA database. Interestingly, SRSF6 and Fas genes were positively co-expressed in normal colon samples (R = 0.800), whereas they were negatively co-expressed in tumor samples (but weaker than the normal samples) (R = −0.202).

Figure 5

SRSF6 expression in normal samples promotes pro-apoptotic signaling and immune cell activation. (A) Based on differential gene expression analysis (R DESeq2), we identified significant expression changes between SRSF6-high and -low colon samples. Log2 fold changes and −log 10-adjusted p-values of differentially expressed genes (DEGs) are shown on the x-axis and y-axis, respectively. (up- and down-regulation were colored red and blue based on fold changes, respectively) (B) We performed pathway enrichment analysis of significantly up-regulated DEGs and identified immune cell activation among SRSF6-high normal samples (hypergeometric tests p-values < 0.01). The top 10 enriched pathways are shown with their respective statistical significances (i.e., −log 10 p-value).