Transcription Factor GATA4 Regulates Cell Type-Specific Splicing Through Direct Interaction With RNA in Human Induced Pluripotent Stem Cell-Derived Cardiac Progenitors

Abstract

Background: GATA4 (GATA-binding protein 4), a zinc finger-containing, DNA-binding transcription factor, is essential for normal cardiac development and homeostasis in mice and humans, and mutations in this gene have been reported in human heart defects. Defects in alternative splicing are associated with many heart diseases, yet relatively little is known about how cell type- or cell state-specific alternative splicing is achieved in the heart. Here, we show that GATA4 regulates cell type-specific splicing through direct interaction with RNA and the spliceosome in human induced pluripotent stem cell-derived cardiac progenitors.

Methods: We leveraged a combination of unbiased approaches including affinity purification of GATA4 and mass spectrometry, enhanced cross-linking with immunoprecipitation, electrophoretic mobility shift assays, in vitro splicing assays, and unbiased transcriptomic analysis to uncover GATA4's novel function as a splicing regulator in human induced pluripotent stem cell-derived cardiac progenitors.

Results: We found that GATA4 interacts with many members of the spliceosome complex in human induced pluripotent stem cell-derived cardiac progenitors. Enhanced cross-linking with immunoprecipitation demonstrated that GATA4 also directly binds to a large number of mRNAs through defined RNA motifs in a sequence-specific manner. In vitro splicing assays indicated that GATA4 regulates alternative splicing through direct RNA binding, resulting in functionally distinct protein products. Correspondingly, knockdown of GATA4 in human induced pluripotent stem cell-derived cardiac progenitors resulted in differential alternative splicing of genes involved in cytoskeleton organization and calcium ion import, with functional consequences associated with the protein isoforms.

Conclusions: This study shows that in addition to its well described transcriptional function, GATA4 interacts with members of the spliceosome complex and regulates cell type-specific alternative splicing via sequence-specific interactions with RNA. Several genes that have splicing regulated by GATA4 have functional consequences and many are associated with dilated cardiomyopathy, suggesting a novel role for GATA4 in achieving the necessary cardiac proteome in normal and stress-responsive conditions.

Keywords: GATA4 transcription factor; RNA splicing; RNA-binding motifs; induced pluripotent stem cells; myocytes, cardiac.

Fig. 1:. GATA4 protein interacts with RNA splicing proteins in cardiac progenitors.

(A) Schematic of the experimental approach to identify human GATA4-interacting proteins in iPSC-derived cardiac progenitor cells (CPCs) by AP-MS using antibody to endogenous GATA4. Proteins isolated in an iPSC line lacking GATA4 were subtracted. (B) RNA binding proteins that were isolated with GATA4 from human iPS-CPs. The interaction network was realized using String. (C) Western blot (WB) of splicing factors indicated after Flag immunoprecipitation (IP) using anti-Flag antibody in HEK293T cells expressing Flag-GATA4. Protein expression in input lysate indicated. Cell lysate were treated with RNase and DNase before IP.

Fig. 2:. eCLIP demonstrates sequence-specific binding of GATA4 to RNA in human iPS-CPs.

(A) Numbers of GATA4-eCLIP peaks in human iPS-CPs. (B) Integrated genome viewer tracks centered on the GATA4 gene. OmniCLIP identified peak regions indicated as black boxes above. (C) Distribution of GATA4-eCLIP (hCPCs) binding sites or RbFOX2-eCLIP (HepG2) binding sites across different regions of gene bodies. (D) KEGG pathway analysis of GATA4-bound RNAs. (E) Weblogos depicting the most significantly enriched GATA4-eCLIP-binding motifs by eCLIP in human iPS-CPs. (F) The motifs 1 or 2 were the reverse complementary sequences of motifs 3 or 4, respectively. (G) EMSAs were performed by adding GATA4-knockout (KO) or GATA4-overexpressing (OE) HEK293T cell lysate to an IRDye 800 labeled RNA probe containing the GATA4-binding motif 4 sequence. EMSA was performed by incubating HEK293T GATA4-OE lysate with labeled probe in the presence of 200(+)-, 1000(++)- and 2000(+++)-fold molar excess of the unlabeled competitor RNA or unlabeled mutant-RNA competitor. (H) RNA-pull down assay of GATA4 with the GATA4 binding site on CESR2 mRNA for GATA4. Immunoprecipitated mRNA were reverse-transcribed and amplified by qRT-PCR. One-way ANOVA coupled with a Tukey test was used to assess significance. **P < 0.001.

Fig. 3:. GATA4 knockdown in human iPS-CPs induces alternative splicing changes.

(A) Quantification of the alternative splicing events (n=1599) upon GATA4-knockdown (KD) by event type (FDR <0.05 from rMATS, junction counts ≥5) obtained from RNA-seq (n = 4). (B and C) Representative Sashimi plots depicting alternative splicing pattern of EIF4A2 in hiPS-CPs with siRNA targeting GATA4 (siGATA4) or negative control siRNA (siNC) and exon inclusion levels in each condition across four replicates (Rep). (D) Representative relative exon usage plots depicting alternative GATA4 splicing patterns in human iPS-CPs with green lines indicating change in exon inclusion upon GATA4 knockdown with siRNA compared to negative control siRNA. (E) Intersection of alternatively spliced genes upon GATA4-KD defined by rMATs (FDR <0.05, junction counts ≥5) and differentially expressed gene upon GATA4-KD (FDR<0.05, FC≥1.5).

Fig. 4:. GATA4 regulates mRNA splicing through direct interaction with mRNAs.

(A) Intersection of alternatively spliced genes upon GATA4-knockdown (KD) defined by rMATs (FDR <0.05, junction counts ≥5), genes with GATA4-eCLIP tag, and genes with GATA4-CHIP tag. (B) Quantification of GATA4 function in regulating splicing of 89 genes that were differentially spliced upon GATA4 KD and were bound to GATA4 based on an eCLIP binding peak. (C) Schematic representation of the in vitro splicing assay. (D) Exon inclusion level of significantly changed exon 3 of SLX4IP upon use of GATA4 siRNA compared to negative control (NC) siRNA. Data are shown as means ± SEM (n=4). Statistical significance was assessed using Student’s t test (**P < 0.001). (E) Integrated genome viewer tracks centered on gene SLX4IP. OmniCLIP identified peak regions indicated as black box. GATA4 binding motifs indicated in red. (F) Sequence information of the WT or mutant SLX4IP pre-mRNA template used in in vitro splicing assay. (G) In vitro splicing of SLX4IP minigene reporter transcripts in Hela nuclear-extracts (NE), with or without the addition of GATA4 or TBX5 protein. Bands indicate sizes of splicing events schematized with pre-mRNA length versus two alternatively spliced events. (H and I) Quantification of relative exon inclusion level in (G). Data are shown as means ± SEM (n=3). One-way ANOVA coupled with a Tukey test was used to assess significance. ns., non-significant; **P < 0.001.

Fig. 5:. GATA4 regulated physiologically-relevant alternative splicing, resulting in functionally distinct protein products.

(A) Intersection of alternatively spliced exons upon GATA4-knockdown (KD) and alternatively spliced exons between human dilated cardiomyopathy (DCM) patient hearts and healthy donor hearts. (B) Schematic of short (WT) and long (CACNA1C e21+e22) isoforms of CACNA1C with differential inclusion of exon 21. (C) Predicted protein topology of the CACNA1C, showing the exon21/22 (red) localizing to the IIIS2 transmembrane segment and part of the linker region between IIIS1 and IIIS2. (D) Whole cell calcium current recordings from neonatal rat ventricular myocytes infected with adenovirus containing green fluorescent protein (GFP) or GFP + CACNA1C^e21+e22. Voltage protocol used to elicit L-type calcium current is shown at the bottom, right. (E) Current-voltage relationships for GFP infected (black, n=8) or CACNA1C^e21+e22 infected (red, n=13) neonatal rat ventricular myocytes (NRVMs), normalized to cell capacitance. Data are shown as means ± SEM. (F) Peak L-type calcium current density at 0 mV for GFP infected (black, n=8) or CACNA1C^e21+e22 infected (red, n=13) NRVMs. Error bars denote standard deviation. P=0.008 by Mann-Whitney test.

Fig. 6:. Model of RNA splicing mediated by GATA4 in human cardiomyocytes through direct binding with RNA.