Generation of zebrafish rtf1 null mutants with CRISPR/Cas9.

(a) Schematic of CRISPR/Cas9 mutagenesis of the rtf1 gene. Two mutant alleles, rtf1LA2678 and rtf1LA2679, were recovered after targeting of rtf1 exon 3. Both alleles are predicted to disrupt translation of the Rtf1 protein and eliminate the HMD, Plus3, Pol II interaction, and PAF1C interaction domains. (b) Agarose gel electrophoresis results of genotyping rtf1 mutants by PCR. Deletions in rtf1LA2678 and rtf1LA2679 alleles can be distinguished from wild type allele using primers rtf1-e3-F and rtf1-e3-R. (c) Western blot detecting Rtf1 and β-actin (loading control) proteins in lysates of wild type and rtf1 mutant embryos. Image is representative of two independent experiments.

Characterization of early cardiac development in Rtf1 deficient embryos.

(a) Representative images of RNA in situ hybridization detecting myl7 expression in 24 hpf zebrafish embryos. (b) Quantification of myl7 signal intensity in control and Rtf1 deficient embryos at 24 hpf. Numbers on bars indicate the number of embryos analyzed. (c) Representative images of RNA in situ hybridization detecting mef2ca expression in 8 somite stage zebrafish embryos. (d) Quantification of mef2ca signal intensity in control and Rtf1 deficient embryos at the 8 somite stage. Numbers on bars indicate the number of embryos analyzed. (e) Representative images of RNA in situ hybridization detecting nkx2.5 expression in 8 somite stage zebrafish embryos. (f) Quantification of nkx2.5 signal intensity in control and Rtf1 deficient embryos at the 8 somite stage. Numbers on bars indicate the number of embryos analyzed. (g) Representative images of RNA in situ hybridization detecting tbx5a expression in 8 somite stage zebrafish embryos. (h) Quantification of tbx5a signal intensity in control and Rtf1 deficient embryos at the 8 somite stage. Numbers on bars indicate the number of embryos analyzed. (i) Representative images of RNA in situ hybridization detecting tbx20 expression in 8 somite stage zebrafish embryos. (j) Quantification of tbx20 signal intensity in control and Rtf1 deficient embryos at the 8 somite stage. Numbers on bars indicate the number of embryos analyzed. Scale bars in a, c, e, g, and i represent 0.1 mm.

Failure of cardiac differentiation upon loss of Rtf1 in the mammalian cardiac mesoderm.

(a) Diagram of generation of cardiac mesoderm-specific Rtf1 knockout mouse embryos. A Mesp1: Cre insertion allele was bred into an Rtf1 exon 3 floxed background. Heterozygous Rtf1 floxed male mice with Mesp1:Cre were bred to homozygous Rtf1 floxed females, resulting in one quarter of embryos lacking Rtf1 activity in the cardiac mesoderm. (b) Representative images of RNA in situ hybridization detecting Nkx2.5 (wild type n = 7, Rtf1 CKO n = 5), Tbx20 (wild type n = 16, Rtf1 CKO n = 4), and MLC2v (wild type n = 1, Rtf1 CKO n = 1) gene expression in Cre-negative (wild type siblings) and Cre-positive homozygous Rtf1 floxed (Rtf1 CKO) embryos at E8.5 or E9.5.

Failure of cardiac differentiation from Rtf1 knockdown mouse embryonic stem cells.

(a) Diagram of shRNA knockdown of Rtf1 in mouse embryonic stem cells (mESCs) and differentiation of plated EBs. mESCs were transduced with non-target control or Rtf1 shRNA lentivirus and selected with puromycin. Transduced mESCs were grown into embryoid bodies (EBs) using the hanging drop method which were then plated to examine differentiation into cell types including beating cardiomyocyte clusters. (b) Western blot verifying reduction of Rtf1 protein in Rtf1 shRNA mESCs (∼70% reduced based on densitometry) compared to unchanged level of the loading control protein β-actin. Image is representative of three independent experiments. (c) qPCR analysis of Myh6, Nkx2-5, and Nppa expression in non-target control (NT) and Rtf1 shRNA plated EBs. Data are normalized to the mean expression level in NT shRNA samples and error bars indicate the standard error of the mean. (d) Mean percentage (± standard error) of plated EBs exhibiting beating cardiomyocyte differentiation (cEBs) in NT and Rtf1 shRNA plated EBs. (e) qPCR analysis of Brachyury expression in non-target control (NT) and Rtf1 shRNA plated EBs. (f) qPCR analysis of Afp expression in non-target control (NT) and Rtf1 shRNA plated EBs. (g) qPCR analysis of Ncam1 expression in non-target control (NT) and Rtf1 shRNA plated EBs. Data in e, f, and g are normalized to the mean expression level in NT shRNA samples at day 0 (D0) and error bars indicate the standard error of the mean. *: p < 0.05, ***: p < 0.001.

Dysregulation of cardiac gene expression in the Rtf1 deficient lateral plate mesoderm.

(a,b) Projections of confocal z-stacks of hand2:GFP signal in 11 somite stage uninjected control (a) and rtf1 morphant (b) embryos. Scale bars represent 100 µm. (c) Diagram of lateral plate mesoderm (LPM) FACS RNA-seq experiment. Transgenic hand2:GFP embryos were injected at the 1-cell stage with rtf1 morpholino and grown to the 10-12 somite stage prior to dissociation and sorting based on GFP expression. RNA isolated from GFP-positive LPM cells was subjected to RNA-seq and compared to RNA from uninjected embryo LPM cells. (d) Heatmap of scaled expression levels of the top 40 genes most significantly differentially expressed (based on p-value) between uninjected control and rtf1 morphant hand2:GFP-positive LPM. High and low z-scores represent high and low gene expression, respectively. (e) Hierarchical clustering of Biological Process gene ontology (GO) terms based on semantic similarity for the genes most significantly downregulated in the rtf1 morphant hand2:GFP-positive LPM.

Rtf1 is required for transcriptional progression of cardiac precursors to a more mature state.

(a-c) UMAP plots of anterior lateral plate mesoderm (ALPM) and derivatives from merged and integrated single cell RNA-seq datasets of 11-12 somite stage uninjected control and rtf1 morphant embryos. Cells in plots are colored by cell type (a), sample (b), and pseudotime (c). (d,f) Expression dynamics plots displaying expression levels (y-axis) for selected genes in uninjected control (d) and rtf1 morphant (f) cells over the pseudotime trajectory (x-axis) from ALPM (root) to cardiac precursor state shown in c. (e,g) UMAP plots of uninjected control (e) and rtf1 morphant (g) cells colored by gene expression levels.

Rtf1’s Plus3 domain is necessary for cardiac progenitor formation.

(a) Schematics of Rtf1 wildtype (wt) and Rtf1 mutant constructs (ΔHMD and ΔPlus3). Domains are indicated by colored boxes (HMD: red, Plus3: blue, PolII and PAF1C interaction: grey). Deleted regions are represented by lines. Numbers refer to the amino acid positions in wildtype Rtf1. (b) Projections of confocal z-stacks of whole mount immunostaining detecting N-terminal FLAG-tagged Rtf1 constructs (red) expressed in 75% epiboly zebrafish embryos. Nuclei are labeled by DAPI staining (blue). Scale bars represent 20 µm. (c) Representative images of RNA in situ hybridization detecting nkx2.5 expression in 10 somite stage (10S) zebrafish embryos coinjected with rtf1 morpholino and mRNA encoding Rtf1 wildtype or mutant proteins. Scale bars represent 0.1 mm.

Rtf1-dependent transcriptional pausing regulates cardiogenesis.

(a) Cumulative frequency plot of pause release ratios (PRRs) of 6,078 genes with substantial RNA Pol II signal. PRRs in control (red) and flavopiridol treated rtf1 morphant (light blue) samples differed significantly from untreated rtf1 morphant PRRs (blue). ***: p < 2.2 x 10-16; Welch’s paired two-tailed t-test. The median PRR of control samples is indicated by a vertical grey dashed line. (b) Dot plot comparing PRRs in controls and rtf1 morphants. Each dot represents the PRR values for a single gene that differ significantly (colored point) or are not significantly different (grey point) between controls and rtf1 morphants. (c) Dot plot comparing PRRs in rtf1 morphants and flavopiridol-treated rtf1 morphants. Each dot represents the PRR values for a single gene that differ significantly (colored point) or are not significantly different (grey point) between rtf1 morphants and flavopiridol-treated rtf1 morphants. Dot colors in b and c are based on the density of points, with lighter colors indicating more dense points. (d-i) Representative ChIP-seq tracks displaying RNA Pol II read densities (y-axis) for control (red), rtf1 morphant (blue), and flavopiridol-treated rtf1 morphant (light blue) embryos at cardiac mesoderm-related genes including hand2 (d), gata5 (e), aplnrb (f), bmp4 (g), nkx2.7 (h), and rbfox1l (i). (j) Representative images of RNA in situ hybridization detecting myl7 expression in 24 hpf control and Rtf1 deficient (± flavopiridol) zebrafish embryos. (k) Quantification of myl7 signal intensity in control and Rtf1 deficient (± flavopiridol) embryos at 24 hpf. Numbers on bars indicate the number of embryos analyzed. (l) Representative images of RNA in situ hybridization detecting myl7 expression in 24 hpf control and Rtf1 deficient (± cdk9 morpholino) zebrafish embryos. (m) Quantification of myl7 signal intensity in control and Rtf1 deficient (± cdk9 morpholino) embryos at 24 hpf. Numbers on bars indicate the number of embryos analyzed.