Repo-Man/PP1 regulates heterochromatin formation in interphase

Abstract

Repo-Man is a protein phosphatase 1 (PP1) targeting subunit that regulates mitotic progression and chromatin remodelling. After mitosis, Repo-Man/PP1 remains associated with chromatin but its function in interphase is not known. Here we show that Repo-Man, via Nup153, is enriched on condensed chromatin at the nuclear periphery and at the edge of the nucleopore basket. Repo-Man/PP1 regulates the formation of heterochromatin, dephosphorylates H3S28 and it is necessary and sufficient for heterochromatin protein 1 binding and H3K27me3 recruitment. Using a novel proteogenomic approach, we show that Repo-Man is enriched at subtelomeric regions together with H2AZ and H3.3 and that depletion of Repo-Man alters the peripheral localization of a subset of these regions and alleviates repression of some polycomb telomeric genes. This study shows a role for a mitotic phosphatase in the regulation of the epigenetic landscape and gene expression in interphase.

Figure 1. Repo-Man is enriched at the periphery of interphase nuclei.

(a) Summary of Repo-Man interactors identified in previous studies. In green interactions with nuclear envelope proteins, blue with histones; yellow phosphatases (PP2A—mitotic exit onset only) and (PP1). (b) HeLa cells were transfected with GFP:Repo-Man (red) then fixed and stained for Nup153 (green). (c) HeLa cells immunostained for endogenous Repo-Man (red) (a,b,d,e) transfected with GFP:Nup153 (a,c) or co-immunostained for endogenous Nup153 (green) (d,f). Example of PLA signals (red) using Repo-Man and Nup153 antibodies (g,h). Scale bar, 10 μm. (d) HeLa cells were transfected with GFP:Nup153 (b,d,f) or GFP alone (a,c,e) and PLA (red) was performed using Repo-Man and GFP antibodies (c,d). (e) Quantification and cellular distribution of PLA signals as described in d from two independent experiments (Fisher ****P-value<0.0001). (f) Percentage and cellular distribution of PLA signals in Repo-Man (green) or Control RNAi (grey) (Chi-Square, ****P-value<0.0001). (g) Electron Microscopy image of Repo-Man cell line expressing the peripheral N terminus domain fused to GFP. Immuno-electron microscopy was conducted using an anti-GFP antibody. Black arrow shows accumulation at the edge of the NPC (white arrow) and Red arrows show accumulation on heterochromatin adjacent to the nuclear envelope (see Supplementary Fig. 1), scale bar, 500 nm. (h) Quantification of the experiment in g. Numbers represent the density of labelling in each of the indicated sub-compartments as the number of gold particles μm⁻² (see materials and methods). Total number of gold particles counted was 1,057.

Figure 2. Nup153 is necessary for Repo-Man targeting to the nuclear periphery.

(a) Anaphase from HeLa stable cell line expressing GFP:Repo-Man^N (N terminus Repo-Man), stained for Nup153 (red). Late anaphase showing Repo-Man accumulation at the chromosome periphery and co-localization with Nup153. (b) HeLa cells transfected with Control (1) or Nup153² RNAi oligo (2) and stained for endogenous Nup153 (red) and endogenous Repo-Man (green). Zoom-in image of Nup153 and Repo-Man at the chromosome periphery in control cells (3). Scale bar, 5 μm. (c) Quantification (line profile analysis) of Repo-Man enrichment at the chromosome periphery. (d) Interphase nucleus of GFP: Repo-Man^N HeLa stable cell line showing the nuclear localization of the construct with enrichment at the nuclear periphery. The quantification of Repo-Man distribution was measured as line profile across the nucleus (yellow line) for the experiments in c and g. (e) Profiles of Repo-Man (green) and H3K9me2 (gray) signals across the nucleus. Repo-Man enrichment was measured as the ratio between the average of the two maximum intensity values (Max1 and Max2) by the median of the values in the plateau (c,g). (f) HeLa cells expressing GFP:Repo-Man^N were transfected with control or Nup153 siRNA oligos and the GFP profiles were analysed as in e. Lower panels are representations of GFP:Repo-Man^N localization in a section of a nucleus. (g) Quantification of Repo-Man enrichment at the nuclear periphery in HeLa cells stably expressing GFP:Repo-Man^N or transiently transfected with GFP: Repo-Man^FL (full-length Repo-Man) after RNAi with Control oligos (grey bars) or with a single (Nup153²) or combination (Nup153^1&2) Nup153 oligos (green bars). Data in c and g were analysed with Mann–Whitney test (****P<0.0001), n are depicted in the figures. In box plots in g, central line represents the median, box limits are the 25th and 75th percentiles and whiskers extend to 1.5 × interquartile range.

Figure 3. Repo-Man is necessary and sufficient to establish a heterochromatic environment.

(a) Quantification of HP1 alpha foci after immunostaining of HeLa cells depleted of Repo-Man (green) or SDS22 (yellow) (1). Rescue of the HP1 foci numbers is achieved by a Repo-Man:GFP oligo-resistant construct in a Repo-Man siRNA background (brown) (2). Chi-Square (****P<0.0001). Typical image of HP1 foci in a control (a,b) or Repo-Man (c,d) RNAi. Scale bar, 5 μm. (b) DT40 cells containing a LacO array inserted in a single locus were transfected with GFP:LacI, GFP:LacI:Repo-Man(RM), GFP:LacI:Repo-Man(RM)^RAXA and the PP1 binding domain GFP:LacI:Ki67^PP1BD. Cells were fixed and stained with HP1 antibody (representative image shown in the inset). The enrichment was calculated as a ratio between the intensity at LacI spot (green arrow in the inset), and a random nuclear spot. (c) Correlation between the accumulation of GFP:LacI:Repo-Man(RM) at the LacO array and HP1 from the experiments in b, linear regression. (d–f) Intensities of H3K27me2/3 (d), H3K9me3 (e) and H3K9ac (f) staining in fixed HeLa cells after control (gray) or Repo-Man (green) RNAi. Cell numbers are depicted in the figure. Data sets were analysed with Mann–Whitney test between three replicates (****P<0.0001). In box plots, central line represents the median, box limits are the 25th and 75th percentiles and whiskers extend to 1.5 × interquartile range. (g,h) DT40 cells containing a LacO array inserted at a single locus were transfected with GFP:LacI (grey) or GFP:LacI:Repo-Man (green). Cells were fixed and stained with antibodies against H3K9ac, H3K9me3 and H3K27me2/3. The signal intensity levels were measured as described in b. (h) Correlation between GFP:LacI:Repo-Man(RM) enrichment at the LacO array and the levels of the active H3K9ac, linear regression. Stars indicate t-test unless stated otherwise (*P<0.05, **P<0.01, ***P<0.001 using two-three replicates). Error bars in a, b and g represent s.d.

Figure 4. Repo-Man interacts with modified histone H3.

(a) Recombinant GST tagged Repo-Man (C terminus domain) was incubated with a histone peptide array (Active Motif). The signal intensity was detected with an anti-GST antibody and quantified by LICOR. Preferential interactions of GST:Repo-Man^CTerm (C-terminus domain) are shown between two biological replicas. (b) Repo-Man has less affinity for peptides containing modified K27 residue (either methylated or acetylated) if the adjacent S28 is phosphorylated (S28P). (c) The histone peptide array was incubated with recombinant GST:Repo-Man^1-135 (dark green bars) or GST alone (white bars) and signals detected with an anti-GST antibody as in a. Error bars represent s.e.m. between two arrays. (d) InstantBlue staining of GST alone, GST:Repo-Man^CTerm or GST:Repo-Man^1-135 proteins incubated with HeLa nucleosomes (bound and unbound fractions are shown). (e) Endogenous Repo-Man and H3K27me2/3 interactions in interphase detected by PLA. Panel on the right shows the overlay of the PLA signals with the nuclear erosion script. Scale bar, 5 μm. (f) Counts of PLA signals in control and Repo-Man RNAi in two replicates as described in c. P-value was calculated using Mann–Whitney test (**P<0.01). n is depicted in the figure. In box plots, central line represents the median, box limits are the 25th and 75th percentiles and whiskers extend to 1.5 × interquartile range.

Figure 5. Repo-Man associates with repressive histone modifications and subtelomeric regions.

Eluted fractions of GST:Repo-Man and GST alone incubated with nucleosomes (as in Fig. 4d) were analysed by Mass Spectrometry (a) or the DNA was extracted and sequenced by Illumina HiSeq (b–f). (a) Histone H3 PTMs identified in the GST:Repo-Man fraction only. In bold are modifications identified in all the replicate experiments and in red are coexistent PRC-dependent histone modifications found in Jung et al. (b) Overview of GST:Repo-Man binding sites genome-wide in two replicates. (c) Repo-Man hits at subtelomeric regions are higher than expected by chance. (d) Annotation of Repo-Man hits according to gene features or lamina association (Fisher P-values). TES: transcription end site; TSS: transcription start site; LADs: lamina associated domains. (e) Overlaps between Repo-Man hits and double histone modifications extracted from HeLa ENCODE data sets for H3K27ac, H3K4me3, H3K79me2, H3K27me3, H2AZ, H3K9me3 and H4K20me1 (Fisher P-values). (f) Single gene profiles of Repo-Man target genes PPP2R2C (1) and PDE9A (2) classified as polycomb repressed and heterochromatin associated (H3K9me3) respectively by the software ChromHMM. The chromosomes and the position of the gene (red line) are shown along with the representation of the genomic sequence (lines/squares are exons). Repo-Man binding sites distribution is shown for two independent data sets (light and dark green). Positioning of histone marks along the genomic window were extracted from the UCSC in HeLa cells (H2AZ, H3K9ac, H3K9me3, H3K27ac, H3K27me3, H4K20me1 and S2-PolII), reads in y axis=50. (g) H3K27me2/3 ChIP on chromatin from control and Repo-Man RNAi cells. Repo-Man RNAi enrichment is expressed over Control RNAi enrichment, calculated relatively to input DNA using same amount of DNA in PCR. Error bars represent s.e.m. t-test was applied. (*P<0.05, **P< 0.01, ***P<0.001).

Figure 6. Repo-Man depletion affects chromosome positioning.

(a) HT1080 cells containing a LacO array inserted at 13q22 and expressing GFP:LacI were fixed and stained for Nup153. The arrow indicates the integration site. (b) Position of the chr13q22 was measured using an erosion script software across five concentric shells (1—most outer shell to 5—most inner shell) after RNAi with control or Repo-Man oligos or transiently transfected with the dominant-negative Repo-Man^RAXA mutant. (c) 3D FISH with probe CTC-820M16 (red signal) mapping to the subtelomeric region of chromosome 14 performed on HeLa cells. (d) Quantification of spots location described in c, using the erosion script software. (Fisher test, *P<0.05, **P<0.01, ***P<0.001 using two-three replicates). (e) Enrichment of H3K27me2/3 at the nuclear periphery after Nup153 RNAi. (f) Enrichment of H3K9me3 at the nuclear periphery after Nup153 RNAi. Enrichment was calculated as in Fig. 2e. Mann–Whitney test (*P<0.05), n is depicted in the figure. In box plots in e and f, central line represents the median, box limits are the 25th and 75th percentiles and whiskers extend to 1.5 x interquartile range. (g) Differential expression of telomeric genes bound by Repo-Man between control and Repo-Man (green) or Nup153 (blue) RNAi. Delta–delta-CT method was used and normalized for GAPDH. Error bars=s.e.m. between three replicates. t-test was used for statistical analysis (*P<0.05, **P<0.01).

Figure 7. Repo-Man dephosphorylates S28 and regulates a phospho-switch necessary for heterochromatin maintenance.

(a–c) Repo-Man dephosphorylates H3S28P. (a) HeLa cells overexpressing GFP:Repo-Man^TA3, which prematurely associates with chromatin, were fixed and stained for H3S28P. Scale bar, 5 μm. (b,c) HeLa cells after control or Repo-Man RNAi were fixed and stained for H3S28P and tubulin. The intensity of H3S28P was measured in anaphase and cytokinesis (cyto) in two independent replicates (c). Scale bar, 10 μm. Mann–Whitney test (****P<0.0001), n is depicted in the figure. In box plots in c, central line represents the median, box limits are the 25th and 75th percentiles and whiskers extend to 1.5 × interquartile range. (d) Model: Repo-Man associates with modified Lysine 27 when the adjacent S28 is not modified. Through dephosphorylations of the nearby serine 10 and 28 it regulates HP1 enrichment and potentially the maintenance of H3K27 methylation respectively. These processes may contribute to the establishment and/or maintenance of a repressive environment. At the periphery, the position of this chromatin environment is also locally maintained via the interaction between Repo-Man and Nup153.