Chromatin architecture transitions from zebrafish sperm through early embryogenesis

Abstract

Chromatin architecture mapping in 3D formats has increased our understanding of how regulatory sequences and gene expression are connected and regulated in a genome. The 3D chromatin genome shows extensive remodeling during embryonic development, and although the cleavage-stage embryos of most species lack structure before zygotic genome activation (pre-ZGA), zebrafish has been reported to have structure. Here, we aimed to determine the chromosomal architecture in paternal/sperm zebrafish gamete cells to discern whether it either resembles or informs early pre-ZGA zebrafish embryo chromatin architecture. First, we assessed the higher-order architecture through advanced low-cell in situ Hi-C. The structure of zebrafish sperm, packaged by histones, lacks topological associated domains and instead displays "hinge-like" domains of ∼150 kb that repeat every 1-2 Mbs, suggesting a condensed repeating structure resembling mitotic chromosomes. The pre-ZGA embryos lacked chromosomal structure, in contrast to prior work, and only developed structure post-ZGA. During post-ZGA, we find chromatin architecture beginning to form at small contact domains of a median length of ∼90 kb. These small contact domains are established at enhancers, including super-enhancers, and chemical inhibition of Ep300a (p300) and Crebbpa (CBP) activity, lowering histone H3K27ac, but not transcription inhibition, diminishes these contacts. Together, this study reveals hinge-like domains in histone-packaged zebrafish sperm chromatin and determines that the initial formation of high-order chromatin architecture in zebrafish embryos occurs after ZGA primarily at enhancers bearing high H3K27ac.

Figure 1.

Chromatin architecture in the developing zebrafish sperm and embryos. (A) Schematic of the time points collected. Samples collected for low-cell in situ Hi-C experiments: sperm, 2.25 hpf (pre-ZGA, 128-cell), 4 hpf, 5.3 hpf, and 24 hpf. The onset of transcription during zygotic genome activation (ZGA) is ∼3 hpf in zebrafish. The transcription activity is portrayed by the green background. (B) Contact matrices of each time point from Chromosome 18 (top) inlet is marked in dashed gray box; Chr 18: 20–30 Mb (bottom), 25 kb resolution in log scale. Flares detected in sperm time point are marked by black arrows. (C) Correlation matrix of each time point from Chr 18. The first eigenvector (PCA1) for the normalized observed/expected ratio is shown below the panel to determine A/B compartment status. (D) Contact matrices of each time point from the Drosophila S2 spike-in Chr 2L: 5–15 Mb, 25 kb resolution in log scale.

Figure 2.

Impact of alternative chorion removal procedures on perceived chromatin architecture. (A) Brightfield images of embryos collected at pre-ZGA/128-cell: early dechorionated (at 128-cell stage; top left), late dechorionated (just before fixation; bottom left). pre-ZGA/128-cell embryo fixed and the DNA was stained with DAPI (cyan) dechorionated at one-cell stage (top right) and dechorionated before fixation (bottom right) with 40× obj scale bar = 10 µM. (B) Contact matrices from pre-ZGA/128-cell, whole Chromosome 11 (early dechorionated, top left; late dechorionated, bottom left) inlet is marked in dashed gray box; partial Chr 11: 19–26 Mb (early dechorionated, top right; late dechorionated, bottom right), 25-kb resolution in log scale. (C) Contact matrices from the Drosophila S2 spike-in, Chr 2L: 5–15 Mb (early dechorionated, top; late dechorionated, bottom), 25-kb resolution in log scale. (D) Contact matrices from Chr 11: 19–26 Mb pre-ZGA/128-cell, 24 hpf, and sperm down-sampled to 100 M valid pairs (left of dashed line). Simulated contact matrices (right of dashed line) are imaged of pre-ZGA/128-cell with increasing percentages of 24 hpf (top) or sperm (bottom) valid pairs. Metaplots for the boundaries called at 24 hpf (25-kb resolution) are plotted below each matrix.

Figure 3.

Characterization of chromatin architecture established at enhancers and super-enhancers at 4 hpf. (A) Super-enhancer (SE) plot using the ROSE algorithm, which ranks enhancers based on histone H3K27ac (Zhang et al. 2018) and H3K4me1 (Bogdanovic et al. 2012). ChIP-seq data at 4 hpf in zebrafish embryos. Data are separated into four groups: Group 1 (purple), Group 2 (red), Group 3 (blue), and SE group (green). N = total number in each partition. (B) Heatmaps of insulation score at enhancers. Insulation maps at 4, 5.3, and 24 hpf ranked by insulation strength at 4 hpf, centered on enhancers from each respective group in Figure 4A. Positive insulation (red) indicates increased contacts, and negative insulation (blue) indicates a lack of contacts. (C) Comparisons of chromatin factor and attribute occupancy at enhancers. Metaplot of log₂ fold enrichment of histone H3K27ac ChIP-seq (Zhang et al. 2018), RNA Pol II ChIP-seq, Rad21-cohesin ChIP-seq, ATAC-seq, and Click-iT-seq (Chan et al. 2019) signal over input are plotted, centered on enhancers from each respective group in Figure 4A: super-enhancers (Super Enh, green), Group 3 (blue), Group 2 (red), and Group 1 (purple). (D) Proportional distribution of different enhancer regions; no enhancer (No Enh, yellow), super-enhancers (SE, green) Group 3 (Grp3, blue), Group 2 (Grp2, red), and Group 1 (Grp1, purple) detected with positive insulation score 0.1–0.2, 0.2–0.3, >0.3. The proportional distribution of each positive insulation score detected over the entire genome is depicted on the right; 0–0.1 (63%), 0.1–0.2 (27%), 0.2–0.3 (8%), >0.3 (1%). The bracket highlights the positive insulation score used in the bar graph on the left. (E) Groups from A overlap with ATAC-seq peak signal across enhancers regions were analyzed using HOMER Motif Analysis to determine potential TF binding. Similarity to known binding motifs is indicated by Pearson R values in shaded red, and motif frequency is indicated by circle size. T-box transcription protein family of motifs (TBX Fam), Kruppel-like factor protein family of motifs (KLF FAM), SRY-box transcription factor protein family of motifs (SOX).

Figure 4.

Inhibition of Crebbp/Ep300a causes loss of chromatin architecture around the established super-enhancers at 4 hpf. (A) Heatmaps of insulation score for drug-treated embryos. Treatments involve DMSO (vehicle), flavopiridol (FLAV), SGC-CP30 (SGC) for 4 h (which causes a developmental arrest for FLAV and SGC). Respective enhancer groups are ranked as in Figure 3B. Positive insulation (red) indicates increased contacts, and negative insulation (blue) indicates a lack of contacts. (B) Click-iT-seq of 4 hpf untreated (UNT) and SGC-CP30 (SGC, purple) (Chan et al. 2019) heatmaps centered on enhancers of each respective enhancer group ranked as in Figure 4A. (C) Model depicting the features present at regions displaying structure/positive interaction. Features displayed include enhancers (red), elevated histone H3K27ac (purple), active transcription (green circle), defined boundaries (cyan), and increased chromatin interactions as detected by positive interaction scores in Hi-C contact maps at both 4 and 5.3 hpf. Regions with increased interactions are typically coated with histone H3K27ac. These interactions and boundaries persist upon inhibition of RNA Pol II initiation at 4 hpf. In contrast, these contacts between boundaries are lost upon inhibition of Crebbp/Ep300a (lowering histone H3K27ac [dashed]) leading to decreased transcription and loss of higher-order chromatin structure; however, boundaries remain stable.

Figure 5.

Hi-C contact maps in zebrafish sperm display “hinge-like” domains. (A) Correlation matrix from Chr 13 (top) and Chr 24 (bottom) for zebrafish sperm (left) and 24 hpf (right). First principal component values to determine A/B compartment status (below). (B) Contact matrices from two regions Chr 13: 33.4–39.4 Mb (top) and Chr 24: 15.6–21.6 Mb (bottom) for sperm (left) and 24 hpf (right), 25-kb resolution in log scale. Flares detected in sperm are marked by black arrows and are not evident in 24 hpf embryos. (C) Contact maps for 24 hpf and sperm samples each presented for a 6-Mb region on Chr 24 at 25-kb resolution in log scale (top). First principal component values to determine A/B compartment status (middle). Heatmap of insulation scores for different window sizes (bottom). The hinge region is marked by a purple square, and distance is marked by an orange square.

Figure 6.

Global characterization of “hinge-like” domains in sperm chromatin architecture. (A) Histogram to depict the width distribution of the 333 hinges. (B) Histogram to depict the average distances between adjacent hinges. (C) Insulation scores at hinge-like domains in sperm, and insulation scores at the corresponding regions in embryos at 4, 5.3, and 24 hpf. Positive insulation (red) indicates increased contacts, and negative insulation (blue) indicates a lack of contacts. (D) Metaplot of log₂ fold enrichment of histone H3K27ac (Zhang et al. 2018), H3K27me3 (Irimia et al. 2012), and H3K4me3 (Zhang et al. 2016) ChIP-seq signal over input for sperm (green) and 24 hpf (blue) centered on the hinge-like domains. (E) Two schematics depict two speculative models of zebrafish sperm chromatin and hinge architecture. In both models, the sperm DNA (which is nucleosome-packaged, not protamine-packaged) is arranged into arrays of consecutive loops/petals, possibly similar to the loops described for condensed mitotic chromosomes in somatic cells (Gibcus et al. 2018). Within each petal DNA positions A–G represent the repeated “hinge” unit within each petal, with “D” representing the hinge center, and the segments A–C and E–G representing the edges of the hinge “petals.” A key feature of the contact map data is that locations equidistant from position D show increased interaction. Two models are presented to achieve this: (1) interactions are caused by contacts within each petal (left), or (2) interactions are caused by contacts between two petals (right). In both models, we propose a constraint on topology, which might involve the loading of ring-like proteins (e.g., condensin or cohesin complexes) at the hinge position D (left) which, together with the fixed hinge position D, create the hinge-like domain through loop extrusion. Ring-like proteins are represented by orange rings. Arrows on the left indicate potential locations where ring-like proteins might load, at hinge position D, to help form stable hinges.