A TALEN-induced mutation abrogates rif1 expression in zebrafish.

(A) Schematic illustration of the TALEN mutation in exon 8 of the zebrafish rif1 gene. The DNA and protein sequences at the TALEN site are shown for the wild-type and a 7 bp deletion mutant. The deletion causes a frame shift and introduces a premature termination codon. The affected amino acids are in red. (B) Diagram of the Rif1 protein. The 7 bp frame shift mutation occurs in the conserved Rif1 N-terminal domain (Rif1-N), causing truncation of the Heat repeats and loss of the nuclear localization signal (NLS) and the PP1 phosphatase SILK and RVSF interaction motifs of the C-terminal domain (CTD). (C) Expression levels of rif1 measured by quantitative RT-PCR analysis at pre-MBT (2.5 hours), dome (4.3 hours), shield (6 hours) and bud (10 hours) in wild-type (+/+), heterozygous (+/-) and homozygous (-/-) rif1 zebrafish. Error bars represent standard deviation of the mean. (D) Rif1 and Smc3 protein expression levels were analyzed by immunoblotting at dome and shield in wild-type (+/+) and homozygous (-/-) rif1 zebrafish.

Loss of rif1 affects zebrafish sex determination, epiboly, and primitive erythropoiesis.

(A) Schematic of the heterozygous rif1 mutant crosses used to test the essentiality of rif1. 766 fish were raised to sexual maturity (3 months) and then genotyped. (B) Quantitation of the genotypes from the heterozygous crosses shows the expected Mendelian ratios. (C) Counts of males and females from the heterozygous crosses shows fewer than expected rif1 mutant female fish. (D) Time-lapse imaging of wild-type and maternal-zygotic rif1 -/-early development shows a delay in epiboly completion in the mutants. (E) Expression of gata1 and B-globin at 24 hpf marks primitive erythroid cells. In situ hybridization shows that gata1 and B-globin expression at 24 hpf was decreased in rif1 homozygous mutants, indicating loss of primitive erythroid cells.

Rif1 loss reduces the general sharpening of RT patterns that occur during normal zebrafish development.

(A) Heat plot and dendograms showing hierarchical clustering of samples based on Euclidean distances calculated with RT values. (B) Principle component analysis reveals that 50% and 27% of variance in RT between samples is due to developmental stage (PC1) and genotype (PC2), respectively. (C) Autocorrelation analysis of RT values shows that RT profiles gradually gain structure during zebrafish development, and Rif1 loss delays that process. (D) Whole chromosome 1 RT profiles are shown for preMBT (top left), shield (top right), bud (bottom left) and 24 hpf (bottom right) stages. The RT profiles gradually develop definition, but Rif1 loss reduces the amplitude of the peaks and valleys. (E) Density plots of genome-wide RT values show a reduced dispersion of values from 0 in the rif1 mutants compared to wild-types. Vertical black lines within density plots represent 25 th and 75 th percentiles.

Rif1 loss does not prevent major changes in RT that occur during early zebrafish development.

(A) Scatterplot of changes in RT of 2,498 putative enhancers that are H3K27 acetylated during gastrulation (DARs). A positive shield-to-24 hpf change indicates a switch to earlier replication during gastrulation or segmentation. The majority of acetylated regions switch to earlier replication in both wild-type and rif1 mutants. (B) RT plotted across three regions with representative DARs. The positions of individual DARs are marked with triangles. (C-F) RT calculated for all of chr4: 28,400,000-70,500,000 (C-D) or chr22: 5,700,000-10,100,000 (E-F) at each stage and genotype. The broad early-to-late switches in RT during gastrulation occur in the rif1 mutants.

Rif1 loss affects gene expression preferentially during the earliest stages of zebrafish development.

(A) Matrix with colors and dendrograms representing Euclidean distances between 3’ mRNA-Seq samples. The order of samples is based on hierarchical clustering and is the same for rows and columns. Each sample’s genotype and developmental stage is marked by a color to the left of its row. (B) Principal component analysis (PCA) plot showing clusters of samples based on similarities of 3’ mRNA-Seq data. The first two principal components are plotted, which represent 29.1% and 14.8% of the variance. Both the Euclidean distance and PCA analyses show that biological replicates are most similar, followed by samples from the same genotype and then stage. Rif1 mutant and wild-type gene expression differs most at dome and shield stages. (C) Fold-change in expression between rif1 -/-and wild-type embryos versus normalized mean 3’ mRNA-Seq read counts. Black data points represent differentially expressed genes (DEGs) with statistically significant (adjusted Wald test p-value < 0.05; Log2FC > 0.5) changes in expression in the rif1 mutants. (D) Counts of DEGs from C with lower (Down) or higher (Up) expression in rif1 mutants.

Figure 5—figure supplement 1. Most DEGs are at normal levels in maternally-contributed mRNA. Heat plots showing the Log2 fold change in mRNA expression of rif1 mutants vs wild types for all Down-DEGs (A) or Up-DEGs (B). Genes are displayed in rows and developmental stages are in columns. Rows were split by k-means clustering. Colors in the row annotations to the right of each heat plot denote the stage(s) at which each DEG was identified. The majority of 1-cell DEGs are expressed normally at other stages, and the majority of DEGs at other stages are expressed normally at 1-cell.

ZGA transcribed or degraded mRNAs are deregulated in rif1 mutants.

(A-B) Heat plots showing the expression (tags per million scaled to maximum) of rif1 mutant Dome Up (A) or Down (B) DEGs (rows) across 18 developmental time points (columns) in wild type embryos. Expression data is from White, et al. (White et al., 2017). The legend across the top of each heat plot denotes six developmental phases. (C-D) Log2 odds ratios from Fisher’s exact tests of associations between dome Up (C) or Down (D) DEGs and the developmental phases in A and B. The total number of each DEG expressed in each phase is indicated by the size of the point, and point fills indicate statistical significance. Dome Up or Down DEGs are over-represented in the Blastula/Gastrula or pre-ZGA phases, respectively. (E-F) Scaled QuantSeq counts for dome Up (E) or Down (F) DEGs across all developmental stages. Overall, dome Up or Down DEGs are normally up-or down-regulated by shield stage, so normal expression of dome DEGs is generally restored by shield. (G) Gene set enrichment analysis showing that genes known to be expressed during the first waves of zebrafish transcription are overexpressed upon Rif1 loss. The genes are ranked by Log2FC in expression between dome stage wild type and rif1 omf201/omf201 embryos, as measured by QuantSeq. (H) Gene set enrichment analysis showing that genes, which are strictly zygotic or maternal, are often up-or down-regulated at dome stage in rif1 mutants, respectively.

Labeling of nascent mRNAs reveals that the role for Rif1 in regulating ZGA transcription is pervasive.

(A, D) Volcano plots showing the Log2-fold change in T-to-C conversions from SLAM-seq for each transcript between wild type and rif1 mutant embryos at preMBT (A) or dome (D) stage. P values were calculated using the beta-binomial test and were corrected using the Benjamini & Hochberg method. The colors represent the density of overlapping data points. (B, E). Barplots of the number of differentially labeled genes (DLGs; >1.5-fold change in T-to-C conversions in rif1omf201/omf201 mutants relative to wild-types and BH-adjusted p value <= 0.1) in A or D. (C, F) Boxplots of T-to-C conversion percentages at preMBT (C) or dome (F) stage for the following: all genes; DEGs upregulated in rif1 mutants from this study (256 Cell Up, Dome Up, or Shield Up), genes transcribed during the first ZGA wave from Heyn et al. or Lee et al., or genes identified by Harvey et al. as maternal or zygotic (Harvey et al., 2013; Heyn et al., 2014; Lee et al., 2013).

Most DEGs are at normal levels in maternally-contributed mRNA.

Heat plots showing the Log2 fold change in mRNA expression of rif1 mutants vs wild types for all Down-DEGs (A) or Up-DEGs (B). Genes are displayed in rows and developmental stages are in columns. Rows were split by k-means clustering. Colors in the row annotations to the right of each heat plot denote the stage(s) at which each DEG was identified. The majority of 1-cell DEGs are expressed normally at other stages, and the majority of DEGs at other stages are expressed normally at 1-cell.