H3.3 deposition counteracts the replication-dependent enrichment of H3.1 at chromocenters in embryonic stem cells

Abstract

Chromocenters in mouse cells are membrane-less nuclear compartments representing typical heterochromatin stably maintained during cell cycle. We explore how histone H3 variants, replicative H3.1/2 or replacement H3.3, mark these domains during the cell cycle in mouse embryonic stem cells, neuronal precursor cells as well as immortalized 3T3 cells. We find a strong and distinct H3.1 enrichment at chromocenters, with variation in mouse embryonic stem cells. Mechanistically, this H3.1 selective enrichment depends on the DNA Synthesis Coupled deposition pathway operating in S phase challenged when we target H3.3 deposition through the DNA Synthesis Independent deposition pathway mediated by HIRA. Altering the H3.1/H3.3 dynamics at chromocenters in mouse embryonic stem cells affects nuclear morphology and cell division. Here, we reveal opposing mechanisms for H3.1 and H3.3 deposition with different enforcement according to cell cycle and potency which determine their ratio at chromocenters and are critical for genome stability and cell survival.

Fig. 1. Histone variant H3.1 is enriched at chromocenters in mouse cells while H3.3 is not.

a Representative wide-field epifluorescence images of Pulse-labeled (TMR) H3.1 and H3.3-SNAP (red) and DNA counterstaining with DAPI (cyan) in mouse ESCs. Single planes are shown, and squared boxes indicate nuclei on the right panels. Scale bar: 10 µm. b Insets of representative ESC nuclei showing “Enriched” (1), “Even” (2, 4), and Excluded (3) H3 patterns at chromocenters for H3.1/H3.3. Arrows point to clusters of pericentric chromatin (or chromocenters) identified as DAPI-dense regions. Scale bar: 10 µm. c Representative epifluorescent images of endogenous H3.1/2 (left) and H3.3 (right) detected with antibodies (red) in ESCs, neuronal progenitor cells (NPCs) derived from ESCs, and NIH-3T3 cells, followed by DNA counterstaining (DAPI, cyan). Scale bar: 10 µm. d Images from STORM acquisitions in HiLO illumination mode of endogenous H3.1 (top) and H3.3 (bottom) in ESC counterstained with DAPI. Arrows indicate chromocenters. HILO widefield mode, before STORM acquisition. (left) and super-resolved (right) images are shown. On the right panel, a zoom-in of the squared chromocenter is shown. The color scale indicates the density of localizations. Scale bar: 10 µm for nucleus, 1 µm for the zoom inset. e Quantitative analysis of the proportion (in %) of cells exhibiting recurrent patterns of endogenous and exogenous H3 variants at PHC in different cell backgrounds. Bar plots shows the mean and standard deviations (s.d.) of 100 nuclei for each cell line from 4 experiments for ESCs and 3 for NPCs and 3T3. ANOVA two-way test was used for statistical analysis: ns (p > 0.05), *(p < 0.05), ** (p < 0.01). P values for endogenous H3 variants: H3.1/2 ESCs vs NPCs = 0.0013; H3.1/2 ESCs vs 3T3 = 0.0041; H3.3 ESCs vs NPCs = 0.0011; H3.3 ESCs vs 3T3 = 0.091. P values for exogenous H3 variants: H3.1 ESCs vs NPCs = 0.0010; H3.1 ESCs vs 3T3 = 0.0135; H3.3 ESCs vs NPCs = 0.619; H3.3 ESCs vs 3T3 = 0.9786. Source data are provided as a Source Data file. f Boxplot showing quantification of H3 variant enrichment at major satellite repeat elements in ESCs and NPCs by ChIP-Seq (SNAP-capture) from 1 replicate. The center of the boxplot is the median, the bounds of the box are the 1st and 3rd quartiles, diamonds are outliers, and the whiskers extend to 1.5 × IQR. The H3 variant enrichment is displayed as a Z-score of log2 enrichment of IP over input indicating enrichment when above 0 as indicated by red dotted line. Source data are provided as a Source Data file.

Fig. 2. H3.1 enrichment at chromocenters follows chromocenter replication status.

a Experimental scheme for visualizing endogenous H3.1/2 or H3.3 during the S phase. Aurora B and EdU staining allowed to resolve the cell cycle stage of individual cells. Cells are scored as G1 (negative to Aurora B and EdU), S phase (EdU positive), or G2 (negative to EdU and positive to Aurora B). S-phase designations are based on EdU S phase patterns with the Early S phase (E) defined by a diffused staining (euchromatin); Mid S phase (M) by a focused EdU labeling around DAPI foci (chromocenters); and late S phase (L) by specific foci staining at the nuclear periphery. b Left: Representative immunofluorescence images of ESCs after in vivo labeling with EdU (green), immunofluorescence staining of H3.1/2 and H3.3 (red), Aurora B (magenta), and DNA (DAPI, cyan). Scale bars, 10 μm. c Quantitative analysis of the proportion of cells displaying H3 variant enrichment patterns at PHC in different cell lines (ESCs, NPCs, NIH-3T3) during the G1, S, and G2 phases. Stacked histogram show the mean (in %) and bar standard deviation from 4 experiments for H3.1 in ESCs and 3 experiments for the other. For H3.1, n = 523, 329 and 275 nuclei for ESCs, NPCs, and 3T3, respectively. For H3.3 n = 328, 298, and 266 nuclei for ESCs, NPCs and 3T3 respectively. ANOVA two way test was used for statistical analysis: ns (p > 0.05), *(p < 0.05), **(p < 0.01). P values are shown in the source data file. Source data are provided as a Source Data file. d Schematic of FUCCI ESCs allowing visualization by microscopy of hCdt1-mCherry and Geminin-Citrine during the cell cycle. Cells are scored as G1 (mCherry++/Citrine−), Late G1/Early S (mCherry+/Citrine+), Mid/Late S (mCherry−/Citrine+), or G2 (mCherry−/Citrine++). e Representative immunofluoresce images of FUCCI ESCs expressing hCdt1 (red) or Geminin (green) with immunofluorescent staining of H3.1/2 (magenta) in G1, Early S, Late S, and G2 phases. Scale bars, 10 μm. f Quantitative analysis of the percentage of FUCCI ESCs exhibiting recurrent H3.1/2 patterns at chromocenters throughout the cell cycle. Bar plots show the mean and standard deviation of 100 nuclei for each cell cycle phase from 4 experiments. Source data are provided as a Source Data file.

Fig. 3. H3.1 enrichment at chromocenters requires SVM motif and is compromised in Suv39h dn ESCs.

a Left: Specific aminoacid residues are depicted for H3.1/2 (purple) and H3.3 (green) and grouped based on their similarities with the histone chaperone recognition motif (SVM or AIG). Aminoacid substitutions from original sequences are boxed. Right: experimental scheme b Representative epifluorescence images of H3-SNAP-Tag histones (red) in ESCs along with DNA counterstaining (DAPI, cyan). Scale bar 10 μm. c Quantification of cells exhibiting H3 patterns at PHC during Early, Mid, and Late S stages. Stacked histograms show the mean (in %) and bars s.d. from 3 experiments for all constructs exept H3.2 with 2 experiments. n = 115 for H3.1; 297 for H3.3; 120 for H3.2; 351 for H3.1 A31S; 186 for H3.3 S31A; 182 for H3.3 S31D. Source data are provided as a Source Data file. d left: Representative epifluorescence images of endogenous H3.1/H3.1 (Red) and H3.3 (Green) co-stained in WT and Suv39h dn ESCs along with DNA counterstaining (DAPI, cyan). Scale bar, 10 µm. Right: boxplots showing the ratio of H3.1/2 (magenta) and H3.3 (green) immunofluorescence signal inside and outside of pericentric heterochromatin (PHC) in WT and Suv39h dn cells from 4 experiments. The center of the boxplot is the median, the bounds of the box are the 1st and 3rd quartiles, dots are outliers, and the whiskers extend to 1.5x IQR. For ESC WT n = 244 for H3.1 and 218 for H3.3; for Suv39h dn ESCs n = 294 for H3.1 and 155 for H3.3. Mann–Whitney two-tailed was used: **** = p value < 0.0001, *** = p value < 0.001, ns non-significant = p value > 0.05. P values: H3.1 ES WT vs. H3.3 ES WT < 1.e⁻¹⁵; H3.1 ES WT vs. H3.1 Suv39dn < e⁻¹⁵; H3.1 ES WT vs. H3.3 Suv39dn < e⁻¹⁵; H3.3 ES WT vs. H3.1 Suv39dn = 0.00022247; H3.3 ES WT vs. H3.3 Suv39dn = 0.68257230; H3.1 in Suv39dn vs. H3.3 in Suv39dn = 0.2479430. Source data are provided as a Source Data file.

Fig. 4. Targeted H3.3 deposition via the HIRA complex alters local H3.1 enrichment at chromocenters.

a Top: scheme depicting HIRA Clover chimeric protein fused to TALE protein recognizing Major Satellite repeat DNA sequences. Bottom: representative immunofluorescent images of ESCs non transfected with TALE construct (NT) and expressing TALE constructs Ctr, HIRA wt and HIRA mut. Clover (green), endogenous H3.1/2 (magenta), H3.3 (red), and DNA (DAPI, cyan) are shown. Arrows point to chromocenter. Scale bars, 10 μm. b Quantitative analysis of the percentage of cells expressing Ctr, HIRA wt, and HIRA mut displaying “enriched”, “even”, and “excluded”. Bar plot shows the mean and bars s.d. from 3 experiments. 100 nuclei were counted per condition. Dots indicate mean of each experiment. Source data are provided as a Source Data file. ANOVA two way test was used for statistical analysis: ns (p > 0.05), *(p < 0.05), **(p < 0.01). P values for H3.3 patterns at PHC: even Ctr vs HIRA wt = 0.0455; even HIRA wt vs HIRA mut = 0.0152; excluded Ctr vs HIRA wt = 0.045; excluded HIRA wt vs HIRA mut = 0.0152. P values for H3.1/2 patterns at PHC: enriched Ctr vs HIRA wt = 0.0006; even HIRA wt vs HIRA mut = 0.0002; even Ctr vs HIRA wt = 0.0021; even HIRA wt vs HIRA mut = 0.0005. Source data are provided as a Source Data file. c Top: scheme depicting approach to select cells non-transfected (NT) non expressing TALE construct from cells expressing HIRA wt or HIRA mut for quantification of H3 variant enrichment at chromocenter. Bottom: boxplots of enrichment values of H3.3 and H3.1/2 at chromocenters in ESCs expressing HIRA wt or HIRA mut and nontransfected (NT) ESCs. The center of the boxplot is the median, the bounds of the box are the 1st and 3rd quartiles and the whiskers extend to max and min. For each HIRA wt and HIRA mut transfection, non-transfected cells (NT non-expressing TALE construct) and cells expressing HIRA wt or HIRA mut are indicated. Data are from three experiments. For H3.3 HIRA wt transfection n = 206 and 109 for NT and HIRA wt; for H3.3 HIRA mut transfection, n = 191 for NT and 41 for HIRA mut; for H3.1/2 HIRA wt transfection, n = 227 for NT and 96 for HIRA wt, for H3.1/2 HIRA mut transfection, n = 161 for NT and 50 for HIRA mut. Kruskal-Wallis test was used for statistical analysis: ns (p > 0.05), *(p < 0.05), ***(p < 0.001), **** (p < 0.0001). P values for H3.3 enrichment NT(wt) vs HIRA wt = 0.0000121; NT(wt) vs NT(mut) = 0.2957; HIRA wt vs HIRA mut = 0.95; NT(mut) vs HIRA mut = 0.0002. P values for H3.1/2 enrichment NT(wt) vs HIRA wt = 0.0207; NT(wt) vs NT(mut) = 0.0920; HIRA wt vs HIRA mut = 0.0004; NT(mut) vs HIRA mut = 0.5713. Source data are provided as a Source Data file.

Fig. 5. H3.3 deposition at mouse chromocenters impacts cell viability and cell division time.

a Representative immunofluorescent images H3K9me3 (left, red) and HP1α (right, red) of ESCs expressing Ctr, HIRA wt or HIRA mut constructs (Clover, green) along with DNA staining (DAPI, cyan). Scale bars, 10 μm. b Boxplots of enrichment values of H3K9me3 (left) and HP1a (right) in ESCs expressing HIRA wt, HIRA mut, or non-transfected cells (NT) from 4 experiments. The center of the boxplot is the median, the bounds of the box are the 1st and 3rd quartiles and the whiskers extend to max and min. For H3K9me3 HIRA wt transfection n = 335 and 220 for NT and HIRA wt; for H3K9me3 HIRA mut transfection, n = 222 for NT and 164 for Hira mut; for HP1α HIRA wt transfection, n = 276 for NT and 178 for HIRA wt; for HP1α HIRA mut transfection, n = 385 for NT and 160 for HIRA mut. ANOVA two-way was used for statistical analysis: ns (p > 0.05), *(p < 0.05), ***(p < 0.001), **** (p < 0.0001). P values for H3K9me3 enrichment: NT(wt) vs HIRA wt = 0.1035; NT(wt) vs NT(mut) = 0.5365; HIRA wt vs HIRA mut = 0.0226; NT(mut) vs HIRA mut = 0.1521. P values for HP1α enrichment NT(wt) vs HIRA wt = 0.0004; NT(wt) vs NT(mut) = 0.1033; HIRA wt vs HIRA mut = 0.0000002676; NT(mut) vs HIRA mut = 0.2767. Source data are provided as a Source Data file. c Series of 15 min time lapse images from ESCs expressing Ctr, HIRA mut and HIRA wt TALE fusions. Clover (green) and Differential Interference Contrast (DIC) images are shown. Time (top arrow) is set to 0 when cell rounds up at beginning of mitosis. Yellow boxed images correspond to cell division. Scale bars, 20 μm. d Quantitative analysis. Left: Boxplot of cell division time. The centre of the boxplot is the median, the bounds of the box are the 1st and 3rd quartiles and the whiskers extend to max and min. Dots indicate mean of each experiment. Data are from 31 nuclei for Ctr, 48 for HIRA wt, and 43 for HIRA mut cells per condition from 2 experiments. Mann–Whiney two-tailed test: *(p < 0.05), **(p < 0.01), ***(p < 0.001). P values Ctr vs HIRA wt = 0.0178; HIRA wt vs HIRA mut = 0.0039. Right: Proportion (in %) of “lost” after cell division. The bars represent the mean and error bars the s.d. Dots indicate mean of each experiment. Data are from 8 nuclei for Ctr, 24 for HIRA wt, and 10 for HIRA mut cells from 2 experiments. Mann–Whiney two-tailed test: *(p < 0.05), **(p < 0.01), ***(p < 0.001). P values Ctr vs HIRA wt = 0.0051; HIRA wt vs HIRA mut = 0.0224. Source data are provided as a Source Data file.

Fig. 6. Schematic model for H3 variant dynamics at chromocenters.

a H3.1/2 and H3.3 variants deposition at chromocenters depends on cell cycle and distinct deposition mechanisms, DNA Synthesis Coupled (DSC) and DNA Synthesis Independent (DSI), respectively. b In a normal ES cell cycle, during G1 phase as the increased transcriptional activity in chromocenter promotes H3.3 deposition by DSI allowing to counteract H3.1/2 enrichment. In S phase, during chromocenter replication, H3.1/2 enrichment is re-established and maintained in G2 prior to cell division. c Forced deposition of H3.3 by DSI throughout cell cycle by targeting HIRA to chromocenter (TALE-HIRA) results in continuous H3.3 deposition and compromises H3.1/2 enrichment at chromocenter. This interference in turn leads to defects in nuclear morphology and cell cycle division.