Global H3.3 dynamic deposition defines its bimodal role in cell fate transition

Abstract

H3.3 is a histone variant, which is deposited on genebodies and regulatory elements, by Hira, marking active transcription. Moreover, H3.3 is deposited on heterochromatin by Atrx/Daxx complex. The exact role of H3.3 in cell fate transition remains elusive. Here, we investigate the dynamic changes in the deposition of the histone variant H3.3 during cellular reprogramming. H3.3 maintains the identities of the parental cells during reprogramming as its removal at early time-point enhances the efficiency of the process. We find that H3.3 plays a similar role in transdifferentiation to hematopoietic progenitors and neuronal differentiation from embryonic stem cells. Contrastingly, H3.3 deposition on genes associated with the newly reprogrammed lineage is essential as its depletion at the later phase abolishes the process. Mechanistically, H3.3 deposition by Hira, and its K4 and K36 modifications are central to the role of H3.3 in cell fate conversion. Finally, H3.3 safeguards fibroblast lineage by regulating Mapk cascade and collagen synthesis.

Fig. 1

Reprogramming induces transcriptomic and chromatin rewiring. a Schematics of the cellular reprogramming indicating the time-points at which chromatin and RNA were collected for libraries preparation. b PCA of ATAC-Seq libraries. c Differential GO analysis revealing enriched biological processes in D0, D9T− and D16S+-accessible genes. The colour ranges from white (no enrichment) to dark red (high enrichment). d PCA of RNA-Seq libraries. e Heatmap demonstrating the dynamic expression of differentially expressed genes between D0 and iPSCs. The boxes to the left indicate genes belonging to each cluster. The values are per-row normalized FPKM and colour ranges from dark blue (low expression) to dark red (high expression). f Average enrichment profile of mESC H3K27ac, H3K56ac, H3K27me3 and H3K9me3 around the TSS of genes belonging to Cluster I and II (top) and Cluster III and IV (bottom). The y-axis represents average normalized number of fragments at the corresponding genomic regions indicated in the x-axis

Fig. 2

Dynamic deposition of H3.3 during cellular reprogramming. a Average distribution of H3.3 peaks on the indicated genomic regions during cellular reprogramming. b PCA of H3.3 ChIP-Seq libraries at the indicated time-points. c–f Average enrichment profile of H3.3 reads, belonging to the libraries at the indicated time-points, around the genebodies of Cluster I (c), II (d), III (e) and IV genes (f). The y-axis represents average normalized number of fragments mapping to the corresponding regions indicated in the x-axis. g UCSC screenshot of the dynamic chromatin accessibility, gene expression and enrichment of H3.3 on Adcy7 (Cluster I gene) during reprogramming. h UCSC screenshot demonstrating the dynamic chromatin accessibility, gene expression and enrichment of H3.3 on Terf2 (Cluster III gene) during reprogramming

Fig. 3

H3.3 acts as barrier during the early stage of reprogramming. a Line plot demonstrating the dynamic H3.3 enrichment on the fibroblast genes-set during cellular reprogramming. Y-axis represents the normalized number of reads mapping to the genebodies of these genes at the indicated time-points (x-axis). b Line plot demonstrating the dynamic H3.3 enrichment on the mesenchymal genes-set during cellular reprogramming. Y-axis represents the normalized number of reads mapping to the genebodies of these genes at the indicated time-points (x-axis). c Heatmap revealing the dynamic expression of the top 250 genes, which are depleted of H3.3 enrichment at any time-point during cellular reprogramming. The values are per-row normalized FPKM values and colour ranges from dark blue (low expression) to dark red (high expression). d Protein–protein interaction network formed by the top 250 genes, which lose H3.3 enrichment at any time-point during cellular reprogramming. Each node represents a protein and the cellular localization of the protein is indicated by the colour of the node according to the key. Black lines (edges) represent protein–protein interaction. e Heatmap revealing the enrichment of transcription factor motifs in the intergenic regions bound by H3.3 in the reprogramming cells (both successful and unsuccessful). Enrichment ranges from not enriched (white) to highly enriched (dark red). f Schematics of the siH3.3 knockdown experiment (top). The bar chart below represents the number of AP-stained colonies (y-axis) observed in wells in which the knockdown of H3.3 had been performed. Non-targeting siNT constructs were used as controls. Values are mean ± s.e.m from independent replicate experiments (n = 3). Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation

Fig. 4

Gain of H3.3 at pluripotency genes is required. a Line plot demonstrating the dynamic H3.3 enrichment on the pluripotency genes-set during cellular reprogramming. Y-axis represents the normalized number of reads mapping to the genebodies of these genes at the indicated time-points (x-axis). b Line plot demonstrating the dynamic H3.3 enrichment on the epithelial genes-set during cellular reprogramming. Y-axis represents the normalized number of reads mapping to the genebodies of these genes at the indicated time-points (x-axis). c, d Average enrichment profile of H3.3 reads, of the indicated time-points, around Nanog-bound (c) and Eset-bound (d) genomic loci. The y-axis represents average normalized number of fragments mapping to the corresponding regions indicated in the x-axis. e Venn diagram demonstrating uniquely and commonly bound regions among D0 H3.3 (left), D16S+ H3.3 (middle), iPSC H3.3 (right) with mESCs Oct4, Sox2 or Nanog (OSN). f Differential GO analysis revealing the enriched biological processes by genes commonly bound by OSN and H3.3 at the indicated time-points. The colour ranges from white (no enrichment) to dark red (high enrichment). g Heatmap revealing the enrichment of the motifs of the indicated transcription factors in the intergenic regions bound by H3.3 in the reprogramming cells (both successful and unsuccessful). Enrichment ranges from not enriched (white) to highly enriched (dark red). h Schematics of the knockdown experiment (top). The bar chart below represents the number of AP-stained colonies (y-axis) observed in wells in which the knockdown of the H3.3 had been performed. Non-targeting siNT constructs were used as controls. Values are mean ± s.e.m from independent replicate experiments (n = 3). Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation

Fig. 5

Role of H3.3 in various induced cell fate changes. a Schematics of the transdifferentiation process indicating the time-points at which chromatin was collected for preparing H3.3 ChIP-Seq libraries. b Venn diagram revealing the number of uniquely and commonly bound genes among day 0 H3.3, day 4 H3.3 and iHP H3.3. c Bar chart revealing the significantly enriched biological processes for genes which are bound by H3.3 on day 4 and in iHP (dark grey) and those bound by H3.3 at day 0 and day 4 (orange). X-axis represents the enrichment score of the hypergeometric enrichment test. d Schematics of the knockdown experiment (top). The bar charts below represent the number of cobblestones formed (left) and CFU numbers (right), observed in wells in which the knockdown of the H3.3 had been performed. Non-targeting siRNA constructs (siNT) were used as controls. The images above the bar charts are representative images of the wells in which the counting took place. Scale bars equal 100 µm. Values are mean ± s.e.m from independent replicate experiments (n = 3). Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation. e Bar chart demonstrating the level of expression of the indicated genes at day 1 and day 2 of the induced neuronal differentiation after the knockdown of H3.3 (siH3.3). Non-targeting siRNA (siNT) were used as controls. Y-axis represents the fold change over control sample. *p-value <0.05. Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation (n = 3). f Bar chart demonstrating the level of expression of the indicated genes at day 4 and day 6 of the induced differentiation to neuronal lineage after the knockdown of H3.3 (siH3.3). Non-targeting siRNA (siNT) were used as controls. Y-axis represents the fold change over control sample. *p-value <0.05. Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation (n = 3)

Fig. 6

Hira is crucial for the role of H3.3 in reprogramming. a Schematics of the H3.3 knockdown ATAC-Seq experiment (top). Venn diagram revealing the number of uniquely and commonly accessible regions in cells in which H3.3 was knocked-down. The pie charts demonstrate the genomic distribution of the accessible regions. b Bar chart demonstrating the hypergeometric test performed on loci with significant accessibility changes. Y-axis denotes the percentage of regions in each category. *p-value <0.05. **p-value >0.1 using Benjamini–Hochberg statistics. c Schematics of the early time-point knockdown experiment (top). The bar chart below represents the number of colonies (y-axis) observed in wells in which the knockdown of the indicated genes had been performed. Non-targeting siRNA constructs were used as controls (siNT). Values are mean ± s.e.m from independent replicate experiments (n = 3). Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation. Representative images of the wells in which the counting took place (top). d Schematics of the late time-point knockdown experiment (top). The bar chart below represents the number of colonies (y-axis) observed in wells in which the knockdown of the indicated genes had been performed. Non-targeting siRNA constructs were used as controls (siNT). Values are mean ± s.e.m from independent replicate experiments (n = 3). Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation. Representative images of the wells in which the counting took place (top). e Enrichment of motifs in the indicated intergenic regions. Enrichment ranges from not enriched (white) to highly enriched (dark red). f Schematics of the overexpression experiment (top). The bar chart below represents the number of GFP+ colonies (y-axis) observed in wells in which the overexpression of the indicated constructs (x-axis). Cells overexpressed with empty vectors were used as controls. Values are mean ± s.e.m from three independent replicate experiments. Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation. Representative images of the wells in which the counting took place (top). g Dynamic enrichment of H3.3 (left), H3K4me3 (middle) and H3K36me3 (right) over fibroblast genes. Colour ranges from dark blue (low enrichment) to dark red (high enrichment)

Fig. 7

H3.3 regulates processes associated with fibroblast lineage. a Venn diagram correlating genes with reduced expression upon knockdown of H3.3 and genes with reduced H3.3 enrichment as cells progress via the successful route of cellular reprogramming. b GO analysis of the functional targets of H3.3. X-axis represents the GO term enrichment score. c, d Protein–protein interaction network of functional H3.3 targets associated with collagen synthesis (c) and Mapk cascade pathway (d). e Schematics of the knockdown experiment (top). The bar chart below represents the number of colonies (y-axis) observed in wells in which the knockdown of the indicated genes had been performed. Non-targeting siRNA constructs were used as controls (siNT). The images above the bar charts are representative images of the wells in which the counting took place. Values are mean ± s.e.m from independent replicate experiments (n = 3). Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation. f Bar chart demonstrating the level of H3.3 enrichment on the genes indicated in the x-axis upon the knockdown of H3.3 (siH3.3). Non-targeting siRNA (siNT) were used as controls. Y-axis represents the enrichment level in the ChIP sample over input. Error bars represent standard deviation (n = 3). g Schematics of the knockdown experiment (top). The bar chart below represents the number of colonies (y-axis) observed in wells in which the knockdown of the indicated genes had been performed. Non-targeting siRNA constructs were used as controls (siNT). The images above the bar charts are representative images of the wells in which the counting took place. Values are mean ± s.e.m from independent replicate experiments (n = 3). Two-tailed t-test was used for statistical analysis. Error bars represent standard deviation. h Dynamic deposition of H3.3 during cellular reprogramming. H3.3 is enriched at genes associated with Mapk cascade or collagen synthesis where it cooperates with the AP1–Fra1 complex to safeguard the fibroblast identity. During reprogramming, H3.3 are disassembled from the lineage genes. Reconfiguration of H3.3 to the nucleosomes of the reprogrammed pluripotent lineage ensure the establishment of the new fate. The effect of H3.3 is exerted via its deposition by Hira and the methylation of its K4 and K36 residues