HP1 loses its chromatin clustering and phase separation function across evolution

Abstract

Heterochromatin protein 1 (HP1) is a multifunctional chromatin-associated protein conserved from fission yeast to mammals. HP1 has been suggested to drive heterochromatin formation via phase separation. However, there is seemingly conflicting evidence about HP1 phase-separating in different systems or not. Here, we assess the phase separation behavior of HP1 from fission yeast, fruit fly and mouse in vitro and in mammalian cells side-by-side. We find that HP1 from fission yeast and fly can undergo liquid-liquid phase separation and induce heterochromatin coalescence in mouse cells, in stark contrast to HP1 from mouse. Induced heterochromatin coalescence has only mild effects on gene expression. We link the decreasing phase separation propensity of HP1 homologs to their decreasing intrinsic disorder and their increasing sensitivity to HP1 paralogs antagonizing phase separation. Our work elucidates the relationship between phase separation, nuclear organization and gene expression, and highlights the evolutionary dimension of protein phase separation control.

Here, the authors compare HP1 from fission yeast, fly and mouse, and find that the propensity of HP1 to phase-separate and to cluster heterochromatin decrease in this order, suggesting an evolutionary adaptation of HP1 function.

Introduction

HP1 has originally been identified as a non-histone protein associated with heterochromatin in fruit flies¹. HP1 homologues were subsequently found in several species from fission yeast to mouse and human^2,3. HP1 family members have a conserved domain organization, with a chromodomain that can bind di/trimethylated lysine 9 of histone 3 (H3K9me2/3)^4,5 and a chromoshadow domain that can mediate dimerization⁶. Both of these structured domains are surrounded by flexible regions, including the linker or hinge region between the chromo- and chromoshadow domain and extensions at the N- and C-terminus. Several parts of HP1 are involved in binding RNA, DNA, chromatin as well as other protein partners^7,8.

Most organisms contain several HP1 paralogs that share a common domain organization and can form homo- and heterodimers^{6,9, 10, 11, 12–13}. Despite these similarities, different HP1 paralogs carry out overlapping but also non-overlapping functions^8,14,15. Fission yeast harbors two HP1 paralogs, Swi6 and Chp2, which have mostly been linked to the regulation of heterochromatin, albeit via distinct pathways^16,17. Fruit flies contain three ubiquitously expressed HP1 paralogs, HP1a, HP1b and HP1c, which regulate heterochromatin but also exert functions in euchromatin¹⁵. Mice and humans express three HP1 paralogs, HP1α, HP1β and HP1γ⁸, which have been linked to a broad range of functions, including the regulation of gene expression, DNA replication, DNA repair and nuclear mechanics^8,18,19. HP1a in flies and HP1β in mice are essential, in contrast to the other HP1 homologs^15,20.

One HP1 paralog in each of the above-mentioned model organisms, namely fission yeast Swi6, fly HP1a and mammalian HP1α, is prone to undergo liquid-liquid phase separation (LLPS)^{21, 22, 23–24}. Other HP1 paralogs are more refractory or antagonistic to LLPS^21,25. These biophysical differences are thought to be linked to the distinct functions of the respective proteins²⁶. Phase separation of HP1 might impact the global organization of the nucleus by regulating the coalescence of heterochromatin regions²⁷, which supposedly affects the gene-regulatory program of the cell. While it is clear that the HP1 paralogs listed above can undergo LLPS in the test tube, it is less clear if they do so in the cell, because it has proven technically challenging to detect LLPS in cells^28,29.

To comprehensively understand the role of HP1 proteins in heterochromatin organization across species, we lack a side-by-side comparison using the same set of assays. To address this issue, we studied HP1 from fission yeast, fruit fly and mouse in the test tube and in different types of mammalian cells, using a set of complementary biophysical techniques. We find that the intrinsic disorder of HP1 decreases from fission yeast to fly to mammals, which goes along with a loss of the capacity to phase-separate and drive heterochromatin coalescence in mammalian cells. We also find that the abundance of HP1 paralogs that antagonize LLPS increases from fission yeast to fly to mammals. Ectopic expression of HP1 from fly in mouse cells induces strong heterochromatin coalescence that goes along with only mild effects on gene regulation. This shows that LLPS of a heterochromatin protein can in principle reorganize chromatin, and that this reorganization does not necessarily have a strong impact on gene regulation. Expression of HP1 from fission yeast also induces heterochromatin coalescence and leads to the upregulation of ~100 genes in mouse cells. Our results suggest that the phase separation capacity of HP1 has been attenuated from fission yeast to fly to mammals, indicating that mammalian HP1 has abandoned its role in large-scale chromatin organization. More broadly, our work suggests that protein phase separation can be specifically tuned across evolution³⁰.

Results

The disorder of HP1 decreases from fission yeast to mammals

We first predicted the intrinsic disorder of HP1 from common model organisms that contain a homolog of mammalian HP1α³¹ using PONDR³². When plotting the cumulative predicted disorder against the divergence time of the different species versus human³³, we observed a monotonous decrease (Fig. 1a). We focused on the three HP1 homologs that have received most interest in the context of phase separation: Swi6 from the fission yeast Schizosaccharomyces pombe²³, HP1a from the fruit fly Drosophila melanogaster (dHP1a)²², and HP1α from the house mouse Mus musculus (mHP1α)²¹. All three HP1 homologs contain similar-sized chromo- and chromoshadow domains but differ with respect to the size of their disordered regions, which decreases from fission yeast to fly to mammals (Fig. 1b, and Supplementary Fig. 1a). We next sought to experimentally determine the size, shape, disorder and assembly state of the respective HP1 homologs in solution. We therefore expressed and purified recombinant HP1 versions (Supplementary Fig. 1b) and analyzed their structure by size-exclusion chromatography coupled to small-angle X-ray scattering (SEC-SAXS). Consistently with the prediction by PONDR, Swi6 was most disordered and flexible, followed by dHP1a and mHP1α (Fig. 1c, and Supplementary Fig. 1c). Based on the particle sizes extracted from the scattering data, all HP1 homologs were most likely present as dimers (Supplementary Data 1). Schematic models of the indicated HP1 dimers are shown in Fig. 1d. We conducted similar SEC-SAXS experiments with HP1 carrying N-terminal GFP tags, which have been shown to be functional in cells, i.e., to localize to heterochromatin and rescue phenotypes caused by the deletion of endogenous HP1^11,34,35. We observed the same trends as for untagged HP1 (Fig. 1e).

[See PDF for image]

Fig. 1

Evolutionary decrease of HP1 disorder.

a Predicted disorder of the HP1 homolog with highest disorder in the indicated species using PONDR. b Domain organization and predicted disorder of fission yeast Swi6, fly HP1a and mouse HP1α. The chromodomains (left) and chromoshadow domains (right) are highlighted. c Kratky plots and pair distance distribution functions P(r) for fission yeast Swi6 (magenta), fly HP1a (blue) and mouse HP1α (green) obtained by SAXS. d Schematic models of the indicated HP1 dimers. e Same as panel c but for GFP-tagged HP1 homologs.

The propensity of HP1 to undergo liquid-liquid phase separation follows its disorder

To compare the intrinsic capacity of HP1 homologs from different species to undergo LLPS in the test tube, we compared them side-by-side in the same buffers. We focused on the GFP-tagged proteins that allowed us to quantify their mobility. At a physiological ionic strength of 150 mM monovalent salt, mHP1α and dHP1a did not form visible condensates at concentrations of up to 400 µM, while fission yeast Swi6 formed small condensates above 50 µM (Fig. 2a, and Supplementary Fig. 2a). When lowering the ionic strength, all three HP1 homologs started to form condensates, which became larger and more numerous with decreasing ionic strength and increasing protein concentration (Fig. 2a, and Supplementary Fig. 2a). We confirmed this behavior by analyzing the turbidity of the samples, which can serve as a readout of condensate formation (Supplementary Fig. 2b). These results are consistent with previous studies of the untagged HP1 counterparts^22,23,25,36 and indicate that Swi6 has a higher capacity to undergo LLPS than the HP1 homologs from fruit fly and mammals. Next, we compared the behavior of HP1 homologs upon addition of 200 bp-sized DNA fragments (Fig. 2b). We fluorescently labeled 25% of the DNA fragments to visualize them by microscopy. At physiological ionic strength and protein concentrations up to 200 µM, Swi6 formed large condensates, whereas only few and small structures were observed for dHP1a and mHP1α (Fig. 2b, top). At the lower ionic strength of 70 mM, which has previously been used to study condensates of human HP1α^21,25, all the three HP1 homologs formed large condensates that contained DNA (Fig. 2b, bottom). We next used a variant of fluorescence recovery after photobleaching (FRAP), termed model-free calibrated half-FRAP (MOCHA-FRAP)³⁷, to study reconstituted HP1-DNA condensates. MOCHA-FRAP can retrieve information about the apparent viscosity and interfacial barrier of condensates via the half-time of recovery and the intensity decrease in the non-bleached half (“dip”), respectively³⁷. The strength of the apparent interfacial barrier reflects the tendency of molecules to preferentially diffuse inside of the condensate instead of exchanging with the surrounding medium. We found that all three HP1 homologs could form condensates with relatively similar apparent viscosities and interfacial barriers in vitro (Fig. 2c). Finally, we assessed the impact of ATP:magnesium, which has been shown to shape the material properties of protein-RNA condensates³⁸, on HP1 phase separation. Increasing ATP:magnesium concentrations enhanced HP1 condensate formation, giving rise to round-shaped condensates in a narrow concentration window and to more irregular assemblies at higher concentrations (Supplementary Fig. 3). These observations suggest that HP1 phase separation can be regulated by active processes that affect ATP:magnesium levels in the cell nucleus.

[See PDF for image]

Fig. 2

Evolutionary decrease in the LLPS propensity of recombinant HP1.

a, b Phase diagrams of GFP-tagged fission yeast Swi6, fly HP1a and mouse HP1α at different NaCl concentrations (a) and upon addition of DNA (b). All images are shown at the same magnification and were acquired with the same microscopy settings. Colored frames indicate condensate formation. Scale bars, 10 µm. c MOCHA-FRAP of condensates reconstituted with 50 µM of the indicated protein and DNA at 70 mM NaCl. The green and the violet line shows the intensity in the bleached and non-bleached half, respectively. The indicated dip depths, which correspond to the minimum intensities in the non-bleached halves, translate into apparent interfacial barriers of 0.02 kT (Swi6, mHP1α) and 0.03 kT (dHP1). The dashed gray line shows the dip depth that is expected in the absence of any interfacial barrier. Data are presented as means ± standard deviations. Source data are provided as a Source Data file.

HP1 paralogs antagonizing phase separation have become more prominent in mammals

As the HP1 family comprises several paralogs that can homo- and heterodimerize^{6,9, 10, 11, 12–13}, their relative stoichiometry and mutual interactions seem critical to understanding HP1 phase separation in cells. When assessing the abundance of HP1 paralogs in different species based on publicly available quantitative mass spectrometry data³⁹, we observed that Swi6 and dHP1a are more abundant than the other HP1 paralogs in fission yeast and fruit fly, respectively. In contrast, HP1α is less abundant than the other HP1 paralogs in mouse and human (Fig. 3a, and Supplementary Fig. 4). This suggests that mammalian HP1β and HP1γ, which are more refractory to phase separation²¹, might interact or heterodimerize with HP1α and thereby reduce its capacity to undergo LLPS. Mechanistically, this could occur because attractive interactions between HP1α/HP1β or HP1α/HP1γ complexes are weakened due to charge repulsion between acidic residues that are present in HP1β and HP1γ^40,41. It has indeed been shown that recombinant human HP1β and HP1γ can dissolve reconstituted condensates containing human HP1α and DNA²⁵, which could reflect the ability of HP1β/HP1γ to sequester DNA and/or to interact with HP1α and thereby reduce its propensity to form condensates.

[See PDF for image]

Fig. 3

Evolutionary appearance of HP1 paralogs antagonizing phase separation.

a Average abundance of different HP1 paralogs in different species. Data were taken from PaxDB³⁹. See Supplementary Fig. 4 for data from individual tissues. Data are presented as weighted means ± weighted standard deviations. b Recombinant RFP-mHP1β impaired condensate formation of GFP-mHP1α but not of GFP-dHP1a and GFP-Swi6. All protein concentrations were 125 µM, NaCl concentrations were 75 mM. Scale bar, 5 µm. c The RFP-mHP1β^I161E mutant did not impair GFP-mHP1α condensate formation. Conditions were the same as in panel b, images are shown at the same magnification. d Fluorescence anisotropy assays indicating that recombinant RFP-mHP1β can disrupt GFP-mHP1α homodimers but not GFP-dHP1a or GFP-Swi6 homodimers. lw, low concentration (200 nM); hi, high concentration (2 µM). For RFP-mHP1β titrations, GFP-mHP1α, GFP-dHP1a and GFP-Swi6 concentrations were kept at 2 µM, NaCl concentrations were kept at 150 mM. No condensates were observed under these conditions. Data are presented as means ± standard deviations. e Fluorescence anisotropy assays indicating that recombinant RFP-mHP1β^I161E cannot disrupt GFP-mHP1α homodimers. Conditions were the same as in panel d. Data are presented as means ± standard deviations. f Dissolution kinetics of GFP-mHP1α condensates upon addition of RFP-mHP1β. Conditions were the same as in panel b. Scale bar, 2 µm. g Fluorescence anisotropy images during dissolution of GFP-mHP1α condensates by RFP-mHP1β. Conditions were the same as in panel b. Scale bar, 2 µm. Source data are provided as a Source Data file.

We decided to use our set of HP1 homologs to elucidate how HP1β affects phase separation. We first induced condensate formation of GFP-Swi6, GFP-dHP1a and GFP-mHP1α by lowering the ionic strength to 75 mM without adding DNA, ruling out potential effects due to DNA sequestration. We subsequently added RFP-mHP1β at equimolar ratios and observed that GFP-Swi6 and GFP-dHP1a condensates were unperturbed while GFP-mHP1α condensates vanished (Fig. 3b). To test if mHP1β requires an intact dimerization domain to perturb GFP-mHP1α condensates, we purified RFP-mHP1β^I161E, which contains a point mutation in its chromoshadow domain that perturbs dimerization⁴². At an equimolar ratio of RFP-mHP1β^I161E, GFP-mHP1α condensates remained intact (Fig. 3c).

We next wondered if RFP-mHP1β can disrupt homodimers formed by GFP-Swi6, GFP-dHP1a and GFP-mHP1α, which is expected if mHP1β heterodimerizes with the respective HP1 homologs. To address this question, we decided to use steady-state fluorescence anisotropy assays. GFP-HP1 homodimers should show a smaller fluorescence anisotropy than GFP-HP1 monomers or GFP-HP1/RFP-HP1 heterodimers, as homotypic Förster Resonance Energy Transfer (homo-FRET) between the two GFP moieties in the GFP-HP1 homodimer induces a partial depolarization of the signal⁴³. We measured the fluorescence anisotropy of GFP-Swi6, GFP-dHP1a and GFP-mHP1α at concentrations of 200 nM and 2 µM, which is below and similar/above the reported dimerization constants of human HP1α²¹, dHP1a⁴⁴ and CFP-Swi6⁴⁵, respectively. We found anisotropy values of ~0.30 for the dilute and ~0.25 for the concentrated samples (Fig. 3d, first two points in each plot), which is similar to previously published values of GFP monomers and GFP dimers, respectively⁴³. We next added increasing amounts of RFP-mHP1β to the different GFP-HP1 samples. We found that the fluorescence anisotropy of GFP-mHP1α increased to the value of the monomer, while only a smaller increase was seen for GFP-Swi6 and GFP-dHP1a (Fig. 3d, black points and colored dashed lines). These results suggest that RFP-mHP1β can disrupt GFP-mHP1α homodimers but not GFP-Swi6 and GFP-dHP1a homodimers. The smaller increase in anisotropy seen for GFP-Swi6 and GFP-dHP1a likely reflects the ability of RFP-mHP1β to form higher-order complexes with intact GFP-Swi6 and GFP-dHP1a homodimers (Supplementary Fig. 5). As a control, we crosslinked the GFP-mHP1α sample before adding RFP-mHP1β, thereby stabilizing homotypic complexes including GFP-mHP1α homodimers. As expected, we observed only a small increase in anisotropy when adding RFP-mHP1β to this crosslinked sample (Fig. 3d, bottom, gray points and gray dashed line). Next, we added RFP-mHP1β^I161E to GFP-mHP1α to assess if RFP-mHP1β requires an intact chromoshadow domain to disrupt GFP-mHP1α homodimers. We observed only a small increase in anisotropy (Fig. 3e), which is consistent with the model that RFP-mHP1β disrupts GFP-mHP1α homodimers via heterodimerization.

To further elucidate the antagonistic effect of RFP-mHP1β on GFP-mHP1α condensates, we conducted time-lapse experiments of GFP-mHP1α condensates after addition of RFP-mHP1β (Fig. 3f). We found that RFP-mHP1β was initially enriched at the surface of GFP-mHP1α condensates, subsequently partitioned into the condensates, and finally triggered condensate dissolution, which proceeded through the formation of a donut-shaped intermediate that was depleted of GFP-mHP1α in its interior. Under the conditions used here, condensate dissolution was completed after ~10 min. We also imaged the fluorescence anisotropy of GFP-mHP1α during condensate dissolution and found that the anisotropy increased while condensates dissolved (Fig. 3g), especially where RFP-mHP1β was enriched, indicating that condensate dissolution induced by RFP-mHP1β proceeds via disruption of GFP-mHP1α homodimers.

Taken together, these results point to an evolutionary adaptation of HP1 phase separation through the emergence of abundant HP1 paralogs that antagonize LLPS by limiting the abundance of HP1 homodimers in the cell.

HP1 from fission yeast and fly but not from mouse can drive heterochromatin coalescence in mammalian cells via phase separation

We next sought to compare the behavior of HP1 homologs from different species side-by-side in the complex environment of a living cell. We chose NIH3T3 mouse fibroblasts as a model system, which contain dense clusters of heterochromatin⁴⁶, also referred to as chromocenters, which are compatible with fluorescence microscopy and MOCHA-FRAP experiments³⁷. Mouse heterochromatin can readily be visualized in living cells using the DNA stain Hoechst, and it is enriched in di/trimethylated lysine 9 of histone H3 (H3K9me2/3), which is recognized by the chromodomain of HP1^4,5. As expected, all three HP1 homologs localized to heterochromatin foci (Fig. 4a). Upon expression of GFP-tagged Swi6 and dHP1a, the size of heterochromatin foci increased compared to wildtype cells, while their number decreased (Fig. 4a, top/center). These effects were dose-dependent, as reflected by the gradual size increase of heterochromatin foci that scaled with the expression level of GFP-tagged Swi6 and dHP1a (Fig. 4a, b, top/center). To put the indicated expression levels into context, the level of endogenous mHP1α in mouse fibroblasts is ~1 µM, that of endogenous dHP1a in fly embryos is ≳10 µM, and that of Swi6 in fission yeast is ~2 µM^{11,12,47, 48–49} (see “Methods” for details). Co-expression of a fluorescent TALE that binds to major satellite sequences⁵⁰ showed that enlarged heterochromatin foci contained pericentric satellite repeats (Fig. 4c). Cells expressing GFP-Swi6 also contained additional dense chromatin foci that were enriched in Swi6 and Hoechst but were devoid of TALEs (Fig. 4c, top). Neither expression of GFP-Swi6 nor expression of GFP-dHP1a abolished mHP1α localization at pericentric heterochromatin (Supplementary Fig. 6a). Only mild effects on heterochromatin organization were observed when expressing GFP-tagged mHP1α (Fig. 4a, b, bottom), which is in line with previous observations made upon overexpression of untagged mHP1α²⁴. We obtained similar results in human U2OS cells, where increasing levels of GFP-Swi6 and GFP-dHP1a induced the formation of heterochromatin clusters with increasing sizes (Supplementary Fig. 7). These observations indicate that GFP-Swi6 and GFP-dHP1a can drive heterochromatin coalescence in different cell types, while GFP-mHP1α produces only mild effects. We next used MOCHA-FRAP to probe the properties of the different HP1 homologs at mouse heterochromatin foci (Fig. 4d). Mouse HP1α did not show an apparent interfacial barrier that would be indicative of LLPS (Fig. 4d, bottom; dip: 0.08), in agreement with previous work^24,37. In stark contrast, Swi6 and dHP1a exhibited interfacial barriers (Fig. 4d, top/center; dips: 0.21 and 0.37, corresponding to apparent barriers of 0.01 and 0.04 kT, respectively), indicating that heterochromatin coalescence induced by these proteins goes along with the formation of phase-separated condensates. The dynamics of chromatin within coalesced heterochromatin foci was not altered as judged by half-FRAP of coexpressed H2B-mCherry (Supplementary Fig. 8a).

[See PDF for image]

Fig. 4

Evolutionary decrease in the capacity of HP1 to drive heterochromatin coalescence.

a Expression of GFP-tagged Swi6 and dHP1a induced heterochromatin coalescence in mouse NIH3T3 cells, as opposed to expression of mHP1α. Expression levels determined via calibration with a GFP standard (Supplementary Fig. 9a) are indicated in the top right corners. Low, 0.0–2.5 μM; med, 2.5–5.0 μM; high, > 5.0 μM. All cells are shown at the same magnification. Scale bar, 5 μm. b Area of heterochromatin foci versus GFP-HP1 expression levels. Groups are defined as in panel a. *, p < 0.05; n.s., not significant (two-sided Welch’s t-test). See Statistics and reproducibility for further details. c Localization of TALEs that recognize major satellite sequences in cells expressing GFP-Swi6 (top) or GFP-dHP1a (bottom). Non-transfected cells were included as a reference (bottom right in each image). Scale bar, 5 μm. d MOCHA-FRAP of the different GFP-tagged HP1 homologs in cells with expression levels of ~4 μM. The green and the violet line shows the intensity in the bleached and non-bleached half, respectively. The dip depths, which correspond to the minimum intensities in the non-bleached halves, indicate that there is an interfacial barrier for dHP1a and Swi6 but not for mHP1α (the violet curve crosses the dashed gray line for dHP1a and Swi6 but not for mHP1α). Data are presented as means ± standard deviations. e Expression of chimeric versions of GFP-Swi6 and GFP-dHP1a, which contained the chromo- and chromoshadow domain of mHP1α, also induced heterochromatin coalescence. Scale bar, 5 μm. Source data are provided as a Source Data file.

Next, we asked if the unstructured regions of HP1 are responsible for driving heterochromatin coalescence. To this end, we constructed two chimeric proteins, GFP-Swi6^swap and GFP-dHP1a^swap, in which we replaced the structured chromo- and chromoshadow domains with the respective domains from mHP1α but kept the unstructured regions unchanged (Fig. 4e). Expression of GFP-Swi6^swap and GFP-dHP1a^swap induced heterochromatin coalescence in a dose-dependent manner (Fig. 4e, and Supplementary Fig. 6b), indicating that the chromatin clustering activity of HP1 is mediated by its disordered regions whose properties differ across species.

We next assessed HP1 phase separation in bipotential mouse embryonic liver (BMEL) cells lacking all three endogenous HP1 paralogs⁵¹. These triple-knockout (TKO) cells exhibit compact chromocenters that contain pericentric satellite repeats (Supplementary Fig. 6c) but do not exhibit an enrichment of H3K9me2/3⁵¹. This phenotype is in line with the role of HP1 in stabilizing H3K9 methyltransferases⁵² and with the dispensability of HP1 for chromocenter formation^51,53. First, we overexpressed GFP-tagged mHP1α in these cells and found that it formed small droplet-like structures that neither colocalized with chromocenters visualized by Hoechst nor with nucleoli stained by RFP-tagged nucleolin (Fig. 5a, top). Some of these mHP1α structures were dynamic (Supplementary Fig. 8b, and Supplementary Movie 1), suggesting that they were not persistently attached to chromatin. When GFP-mHP1α was expressed for longer times (> 48 h), HP1 structures accumulated at the periphery of chromocenters and subsequently covered the entire chromocenters (Fig. 5a, center). In BMEL cells that express endogenous HP1β/HP1γ and also contain chromocenters that are enriched in H3K9me2/3⁵¹, GFP-mHP1α did not form similar droplet-like structures but rather bound to chromocenters (Fig. 5a, bottom). We next expressed GFP-dHP1a and GFP-Swi6 in BMEL TKO cells. Both proteins formed droplet-like structures (Fig. 5b, first and third row) and colocalized with compact chromocenters when expressed at high levels for longer times (Fig. 5b, second and fourth row). Prolonged expression of GFP-dHP1a or GFP-Swi6 in BMEL TKO cells also induced heterochromatin coalescence (Fig. 5b, second and fourth row). These results show that heterochromatin coalescence can occur in the absence of endogenous mHP1α and corroborate the model that mHP1β antagonizes LLPS of mHP1α in mouse cells.

[See PDF for image]

Fig. 5

HP1β antagonizes phase separation in mouse cells.

a GFP-mHP1α formed droplet-like structures (arrows) when expressed for 24 h in BMEL TKO cells lacking all three endogenous HP1 paralogs (top). After prolonged expression in TKO cells, GFP-mHP1α bound chromocenters (center). GFP-mHP1α also bound chromocenters in BMEL cells lacking only endogenous HP1α (bottom). b GFP-Swi6 and GFP-dHP1a also formed droplet-like structures in BMEL TKO cells. After prolonged expression at high levels, both proteins bound chromocenters and induced heterochromatin coalescence. Cells in both panels are shown at the same magnification. Red signals in both panels represent the nucleolar marker RFP-nucleolin. Scale bars, 5 μm.

HP1-driven heterochromatin coalescence has only mild effects on gene regulation

We next wondered how HP1 phase separation and heterochromatin coalescence affect gene expression. To this end, we generated stable NIH3T3 cell lines expressing the different GFP-tagged HP1 homologs under the control of a doxycycline-inducible promoter (Supplementary Fig. 9b–d), and profiled their transcriptome using RNA-seq. We found almost no differentially expressed genes in cells expressing GFP-mHP1α and GFP-dHP1a at ~2–5 µM for 48 h (Fig. 6a), although a fraction of cells expressing GFP-dHP1a contained coalesced heterochromatin and although MOCHA-FRAP indicated that GFP-dHP1a phase-separates at such expression levels (Fig. 4d). In contrast, we found 132 differentially expressed genes (including predicted genes) in cells expressing GFP-Swi6 at similar levels for 48 h (Fig. 6a and Supplementary Data 2–4). These genes were not enriched in the proximity of (peri)centromeres (Fig. 6b), and only 16% of them were located in the constitutive pericentromere-associated domains that were previously mapped in four different mouse cell types⁵⁴. More than 80% of them changed their expression levels in the same direction in TKO livers that lack all three HP1 paralogs⁵¹, with 12% of them showing significant changes in TKO livers. Accordingly, at least some of the genes that are sensitive to GFP-Swi6 expression seem to be regulated by endogenous HP1, which might get displaced upon GFP-Swi6 expression. Gene ontology analysis showed that differentially regulated genes were preferentially associated with processes that are pertinent to the function of multicellular organisms rather than the function of unicellular organisms like fission yeast, including the regulation of the extracellular matrix (Fig. 6c). The levels of major satellite repeat transcripts, which originate from heterochromatin foci, did not significantly change upon expression of any of the GFP-HP1 homologs (Fig. 6d). When estimating their absolute levels based on spike-in controls, we found that each NIH3T3 cell contained ~2 attograms of them (Fig. 6d). Considering a single repeat unit per transcript and a heterochromatin volume of ~75 µm³ per cell⁵⁵, this translates into a concentration of ~300 pM within heterochromatin domains, which is much lower than the micromolar RNA concentrations that have been shown to drive phase separation of mHP1α in vitro⁵⁶. In summary, our results indicate that heterochromatin coalescence induced by condensate formation of GFP-HP1 has relatively mild effects on gene expression, suggesting that these processes might play more prominent roles in regulating other cellular functions^18,57.

[See PDF for image]

Fig. 6

Decoupling between HP1-driven heterochromatin reorganization and gene regulation.

a Expression of GFP-tagged dHP1a and subsequent heterochromatin coalescence had only mild effects on gene expression, while expression of GFP-tagged Swi6 significantly altered the expression of 132 genes (117 genes upregulated, red; 15 genes downregulated, blue; p_adj < 0.05). p-values were obtained with a Wald test, p_adj-values were corrected for multiple testing. b Genomic positions of differentially expressed genes (DEGs) upon induction of GFP-Swi6. DEGs are not enriched near pericentromeres, which are located close to the left end of telocentric mouse chromosomes. c Manhattan plot showing gene ontology enrichments of DEGs. BP, biological processes; CC, cellular components. p_adj-values were obtained with a hypergeometric test and were corrected for multiple testing. d Estimated levels of major satellite repeat transcripts did not significantly change upon expression of GFP-tagged HP1 homologs. e Two mechanisms modulate (liquid) phase separation of HP1 across evolution: (i) The intrinsic disorder of HP1 decreases from fission yeast to mammals, and (ii) the abundance of HP1 paralogs that antagonize LLPS increases. These changes go along with alterations of HP1 function. Source data are provided as a Source Data file.

Discussion

In the present work, we study the capacity of HP1 homologs from different species to phase-separate and thereby impact chromatin organization and gene expression. We provide evidence that HP1 loses its capacity to undergo LLPS and drive heterochromatin coalescence from fission yeast to fruit fly to mouse, which have diverged from a common ancestor with humans ~1275, 686 and 87 million years ago³³, respectively. On the one hand, the intrinsic disorder of HP1 homologs decreases from fission yeast to mammals, limiting the propensity of HP1 to undergo LLPS. On the other hand, HP1 paralogs that antagonize LLPS increase in abundance from fission yeast to mammals (Fig. 6e). We find that HP1 from fly and fission yeast can undergo LLPS in mouse cells and can drive heterochromatin coalescence. Notably, only HP1 from fission yeast had a strong effect on gene expression in mouse cells, while HP1 from fly, which was as potent as Swi6 in driving heterochromatin coalescence, had only mild effects for the expression levels and time points we tested. These observations indicate that enforced HP1-driven heterochromatin coalescence can in principle induce global alterations of nuclear architecture, in line with recent work in flies^22,58,59, while these alterations per se do not necessarily have a strong impact on gene expression. Notably, coalescence of heterochromatin domains might play other roles in the cell, for example in regulating genome stability and nuclear mechanics^18,57,60. Mouse HP1 seems to have only a weak role in regulating large-scale nuclear organization, which is consistent with previous studies that have shown that mouse heterochromatin can form dense clusters in the absence of HP1^51,61 and which suggests that the role of HP1 in gene regulation is decoupled from a global architectural function. Other proteins might have taken on roles in regulating heterochromatin architecture, for example proteins that bind to methylated cytosines, including MeCP2⁶², which have appeared across evolution in organisms with increasing levels of cytosine methylation⁶³. It should be noted that other protein families, such as high mobility group proteins⁶⁴, zinc fingers and transcription factors⁶⁵, also regulate heterochromatin clustering in mammals, establishing together a redundant system to spatially organize heterochromatin.

Our results reconcile previous experimental results obtained with HP1 in different systems, and highlight the difference between HP1 homodimers in the test tube and the more diverse HP1 pool in cells. The evolutionary decrease of the phase separation capacity of HP1 goes along with the acquisition of novel functions of HP1 outside of heterochromatin^7,14. We hypothesize that a strong propensity to undergo LLPS is not favorable for this functional diversification, as chromatin regions where HP1 exerts different activities, e.g., damaged euchromatic regions undergoing repair and intact silenced heterochromatic repeats, are not supposed to coalesce into the same condensates. Interestingly, it has been shown that the length and disorder of proteins generally increases from bacteria to eukaryotes and then remains fairly similar across eukaryotes^30,66. For the species considered here, HP1 represents a protein that goes against this trend, becoming shorter and less disordered from fission yeast to fly to mammals, indicating that the phase separation capacity of proteins can be specifically regulated to accompany their functional alteration across species. To further elucidate the evolutionary adaptation of HP1 phase separation, it would be interesting to systematically compare the phase separation capacity and functional spectrum of HP1 homologs in additional organisms, including plants where evolutionary changes in heterochromatin condensation and HP1 function have recently been reported⁶⁷.

In summary, our work sheds light on the link between heterochromatin organization, (liquid) phase separation of heterochromatin proteins and gene regulation, and highlights the evolutionary dimension of phase separation control.

Methods

Plasmids

The coding sequence for mouse HP1α (UniProt Q61686) was obtained from a plasmid we have previously described²⁴. The coding sequence for fruit fly HP1a (UniProt P05205) was obtained from pDam-Myc-HP1⁶⁸ (Addgene plasmid #59221). The coding sequence for fission yeast Swi6 (UniProt P40381) was obtained from the genomic DNA of a yeast strain carrying a GFP-Swi6 transgene, which was kindly provided by Sylvie Tournier and Yannick Gachet (CBI Toulouse). Coding sequences were amplified by PCR and inserted into pET28 backbones (EMD Biosciences) for expression in bacteria and into pSBtet-BP backbones⁶⁹ (Addgene plasmid #60496) for expression of the proteins with an N-terminal GFP-tag in mammalian cells. pET28 versions for the expression of the proteins with an N-terminal GFP-tag in bacteria were also constructed. To express the chimeric Swi6^swap variant in mammalian cells, the chromodomain of Swi6 (aa 81-143) and the chromoshadow domain of Swi6 (aa 267-328) were replaced by the chromodomain of mHP1α (aa 20-78) and the chromoshadow domain of mHP1α (aa 121-179), respectively. To express the chimeric dHP1a^swap variant in mammalian cells, the chromodomain of dHP1a (aa 24-82) and the chromoshadow domain of dHP1a (aa 147-205) were replaced by the chromo- and chromoshadow domain of mHP1α, respectively. In addition, the coding sequence for TagRFP (Evrogen) followed by the coding sequence for mouse HP1β¹¹ (UniProt P83917) was inserted into a pET28 backbone for expression in bacteria. In the same manner, a pET28 plasmid for the expression of the I161E mutant of HP1β in bacteria was constructed. The plasmids for the expression of TagRFP-mHP1α and TagRFP-nucleolin in mammalian cells were previously described^11,70. To visualize major satellite sequences, we constructed a plasmid encoding the respective TALE (Addgene plasmid #47878)⁵⁰ fused to mScarlet-I.

Protein expression and purification

Recombinant HP1 and GFP-HP1 versions carrying N-terminal His-tags were expressed in E. coli Rosetta cells. Cells were grown in ZYM autoinduction medium for 5 h at 37 °C and subsequently overnight at 18 °C. Cells were afterwards pelleted, resuspended in lysis buffer (50 mM Tris-Cl pH 7.5, 300 mM NaCl, 1 mM DTT) and sonicated for 1 min (5 cycles) using a microtip sonicator. Cleared cell lysates were loaded on centrifuge columns packed with 10 mL HisPur cobalt resin (Thermo Fisher) and were allowed to bind for 20 min at 4 °C on a lab roller. The resin was washed with lysis buffer supplemented with 10 mM (1^st wash) or 20 mM (2^nd wash) of imidazole, and eluted with elution buffer (50 mM Tris-Cl pH 7.5, 300 mM NaCl, 1 mM DTT, 150 mM imidazole). Eluates were subsequently subjected to size exclusion chromatography using a HiLoad 16/600 Superdex 200 pg column (Cytiva), concentrated with spin columns, and dialyzed into storage buffer (25 mM Tris-Cl pH 7.5, 300 mM NaCl, 1 mM DTT, 20% glycerol). The DNA contamination of the resulting eluates was below 0.3% (w/w) as judged by Qubit (after correction of GFP signals, if applicable). Recombinant proteins analyzed by SDS Polyacrylamide Gel Electrophoresis followed by StainFree imaging are shown in Supplementary Fig. 1b.

Small-angle X-ray scattering

SAXS experiments were performed at the European Synchrotron Radiation Facility (ESRF) on the BioSAXS beamline BM29^71,72. An online HPLC system (Shimadzu) was connected to the sample inlet valve of the BM29 sample changer. A BioSEC 300A (Agilent) column was equilibrated with at least five column volumes of online degassed buffer, and the baseline was monitored before each run. At a flow rate of 0.25 mL/min at room temperature, 100 µL of the protein samples mHP1α (3 mg/mL), dHP1a (4.6 mg/mL) and Swi6 (2.5 mg/mL), GFP-mHP1α (2.6 mg/mL), GFP-dHP1a (5.5 mg/mL) and GFP-Swi6 (4.0 mg/mL) were automatically injected into the column via an integrated syringe system. Data were collected using a Pilatus 2M detector at a distance of 2.81 m, covering a q-range of 0.008 to 0.45 Å⁻¹. A total of ~660 frames (2 s/frame) per run were collected. Initial data processing was automated using the Dahu pipeline, which generated radially integrated, calibrated, and normalized 1D profiles⁷³. Further processing was done using Scatter IV⁷⁴; 50 to 100 frames for background buffer and 30 frames with the highest protein concentration were selected and merged for data analysis and model fitting.

Schematic HP1 models

To construct the schematic HP1 models shown in Fig. 1d, an ensemble of 10,000 HP1 monomer conformations was generated using RANCH (Random Chains), which is part of the EOM package⁷⁵. The conformations of the chromodomains (CD) and chromoshadow domains (CSD) were fixed based on the structures available in the protein data bank (PDB, RCSB.org; 2RVL, CD mouse; 1Q3L, CD fruit fly; 2RSO, CD fission yeast; 3P7J, CSD fruit fly, also used for mouse; 1E0B, CSD fission yeast)⁷⁶, and the conformations of the other parts of the proteins were sampled as random chains. Symmetric dimer structures were then constructed by aligning the generated monomers to the structure of chromoshadow domain dimers available in the PDB (3P7J for mouse and fruit fly, 1E0B for fission yeast). The dimer structure yielding the pair distribution function with the highest similarity to the experimentally measured one was rendered in VMD⁷⁷, using the SURF (solvent accessible surface) and Tachyon packages. Of note, the experimental pair distribution functions cannot be accurately reproduced by a single HP1 dimer structure, which means that the schematic models shown in Fig. 1d represent only one of the possible conformations present in the ensemble.

Turbidity measurements

The turbidity of GFP-HP1 solutions was measured as the apparent absorbance at a wavelength of 600 nm using a Varioskan Flash plate reader (Thermo Fisher). Sample volumes of 20 μL in flat-bottom 384-well plates were used. Measurements were taken ~5–15 min after mixing.

Microscopy and MOCHA-FRAP of reconstituted condensates

Reconstituted HP1 condensates were prepared in buffers containing 25 mM Tris-Cl pH 7.5, 1 mM DTT, and the indicated amount of NaCl. If indicated, a 200 bp-sized DNA fragment corresponding to the Widom 601 sequence⁷⁸ with flanking linker DNA was added. The DNA fragment was obtained by AvaI digestion of a plasmid containing an array of 601 sequences⁷⁹ and subsequent amplification by PCR, using primers with or without a terminal Alexa Fluor 647 dye. Condensates were imaged using a Zeiss LSM 710 confocal microscope with a 63x oil immersion objective. Samples were pipetted into a 96-well plate with a coated glass bottom (89606, ibidi). Fluorescence recovery after photobleaching (FRAP) experiments were conducted with a line frequency of 400 Hz, an image size of 512 × 512 pixels, a pixel size of ~60 nm, no averaging, and one bleach iteration. Typically, at least 10 condensates were bleached for each condition. Data were analyzed using a previously published pipeline³⁷. The same gain and laser settings were used for each sample.

Fluorescence anisotropy

Fluorescence anisotropy measurements were conducted using a custom-built epifluorescence microscope that was equipped with polarization optics. Excitation light from a mercury light source (Olympus) was linearly polarized with a film polarizer (LPVISA050, Thorlabs). The light was reflected by a dichroic mirror (Olympus) towards a 63× NA 1.2 water immersion objective (Olympus). Emitted fluorescence light passed the dichroic mirror and an emission filter (Chroma 525/50 for GFP and Brightline 641/75 for RFP) and was subsequently split into parallel and perpendicular polarization components with a broadband polarizing beam splitter (PBS201, Thorlabs). Each component was focused on a distinct part of an EMCCD detector (Andor iXon Ultra 897 EM-CCD) that was read out in conventional mode. The optimal rotational position of the polarizer in the excitation light path was determined by maximizing the difference between the parallel and perpendicular components, $I_{\parallel }$ and $I_{\perp }$. Detected signals were background-corrected and quantified with ImageJ, and the corrected steady-state anisotropy r was calculated^80,81 as

1

$r=\frac{I_z-I_y}{I_z+2\cdot I_y}$The quantities $I_y$ and $I_z$ are defined as

2

$I_y=\frac{I_{\parallel }{K}_{\mathrm{b}}-g{I}_{\perp }{K}_{\mathrm{c}}}{K_{\mathrm{b}}^2+K_{\mathrm{a}}{K}_{\mathrm{b}}-K_{\mathrm{c}}^2-K_{\mathrm{a}}{K}_{\mathrm{c}}}$

3

$I_z=\frac{g{I}_{\perp }\left(K_{\mathrm{a}}+K_{\mathrm{b}}\right)-I_{\parallel }\left(K_{\mathrm{a}}+K_{\mathrm{c}}\right)}{K_{\mathrm{b}}^2+K_{\mathrm{a}}{K}_{\mathrm{b}}-K_{\mathrm{c}}^2-K_{\mathrm{a}}{K}_{\mathrm{c}}}$

4

$K_{\mathrm{a}}=\frac{2-3\cos \theta +{\cos }^3\theta }{3}$

5

$K_{\mathrm{b}}=\frac{1-3\cos \theta +3{\cos }^2\theta -{\cos }^3\theta }{12}$

6

$K_{\mathrm{c}}=\frac{5-3\cos \theta -{\cos }^2\theta -{\cos }^3\theta }{4}$Here, $\theta$ is the collection angle of the lens and the g-factor corrects the difference in detection sensitivity for the two different polarizations. The g-factor was measured as the ratio of parallel to perpendicular signal when a 1 mM solution of fluorescein was excited with perpendicularly polarized light. For each experiment in Fig. 3d, e, samples were prepared in 50 mM Tris-Cl pH 7.5, 150 mM NaCl and 10 % glycerol, incubated for 1 h at room temperature and pipetted on a 96 well-plate (89606, Ibidi). Anisotropy curves were fitted to a generic sigmoidal function of the form

7

$r\left(x\right)=r_0+\frac{r_{\infty }-r_0}{1+\exp \left(-k\cdot \left(x-x_0\right)\right)}$Here, $x$ is the (logarithmic) ratio between RFP- and GFP-tagged HP1 versions, $r_0$ and $r_{\infty }$ are the anisotropy values in the plateaus, and $x_0$ determines the position of the inflection point. For the experiments in Fig. 3g, the NaCl concentration was lowered to 75 mM NaCl to trigger condensate formation (using the same conditions as in Fig. 3b). A Gaussian blur with a standard deviation of one pixel was applied to the images.

Protein crosslinking and SDS-PAGE

Crosslinked GFP-mHP1α samples for fluorescence anisotropy experiments (Fig. 3d, bottom) were prepared by adding EDC (22980, Thermo Fisher) and NHS (24500, Thermo Fisher) to final concentrations of 1 mM and 5 mM, respectively, to 100 µL of a 2 µM GFP-mHP1α sample in PBS and by then incubating at room temperature for 1 h. The crosslinking reaction was quenched by addition of Tris-Cl to a final concentration of 100 mM. Crosslinked GFP-HP1/RFP-mHP1β complexes analyzed by gel electrophoresis (Supplementary Fig. 5) were prepared by incubating 2 µM GFP-mHP1α, GFP-dHP1a or GFP-Swi6 with increasing amounts of RFP-mHP1β at RFP-mHP1β/GFP-HP1 ratios of 0.1, 0.5, 1.5, 4, 6 in PBS for 1 h. Subsequently, complexes were crosslinked by addition of EDC (22980, Thermo Fisher) and NHS (24500, Thermo Fisher) to final concentrations of 1 mM and 5 mM, respectively, and incubated at room temperature for 1 h. Crosslinked reactions were quenched by addition of Tris-Cl to a final concentration of 100 mM and loaded onto a polyacrylamide gel (Mini-PROTEAN TGX Stain-Free Precast Gels, Bio-Rad). Electrophoresis was carried out at 180 V and gels were imaged at a Typhoon Trio Imager (GE), using a 488 nm laser to excite GFP and a 532 nm laser to excite RFP.

Cell culture and transfection

NIH3T3 and U2OS cells were obtained from ATCC (CRL-1658, HTB-96), bipotential mouse embryonic liver (BMEL) cells lacking HP1α or lacking all three HP1 paralogs were established by Florence M. Cammas (IRC Montpellier). Cells were cultured at 37 °C in a humidified atmosphere with 5% CO₂. NIH3T3 cells were grown in Dulbecco’s Modified Eagle Medium (DMEM; 31053028, Gibco) supplemented with 10% fetal bovine serum (FBS; 11550356, Life Technologies), 1× Penicillin/Streptomycin (15140122, Gibco), 1× Sodium Pyruvate (11360070, Gibco) and 1× GlutaMAX Supplement (35050061, Gibco). BMEL cells were grown in Roswell Park Memorial Institute medium (RPMI 1640; 61870010, Gibco) supplemented with 10% fetal bovine serum, 10 μg/mL insulin (I9278, Sigma), 30 ng/mL IGFII (I8904, Sigma), and 50 ng/mL EGF (E5160, Sigma). U2OS cells were grown in DMEM supplemented with 10% fetal bovine serum, 1x Penicillin/Streptomycin, 1x Sodium Pyruvate, 1× GlutaMAX Supplement and 1x MEM Non-essential Amino Acid Solution (M7145, Sigma). For microscopy experiments, cells were grown on LabTek chambered cover slides (155411, Nunc). Cells were transfected with Lipofectamine 2000 (11668019, Invitrogen) according to the manufacturer’s protocol. Expression of GFP-HP1 fusions was induced by adding 500 ng/mL doxycycline (D3447, Sigma) to the culture medium.

Generation of stable cell lines

Stable NIH3T3 cell lines were generated by co-transfecting NIH3T3 cells with the respective pSBtet-BP plasmid and another plasmid coding for a hyperactive Sleeping Beauty transposase⁸² (Addgene #34879). Puromycin (ant-pr-1, InvivoGen) was added one day after transfection to a final concentration of 2 ng/mL.

Live-cell microscopy and MOCHA-FRAP

Live-cell experiments were conducted using a Zeiss LSM 880 Airyscan microscope with a 63x oil immersion objective. Fluorescence recovery after photobleaching (FRAP) experiments were conducted with a line frequency of 400 Hz, an image size of 512×512 pixels, a pixel size of ~60 nm, no averaging, and one bleach iteration. Data were analyzed using a previously published analysis pipeline³⁷. The same gain and laser settings were used for each sample, and count rates were converted into concentrations using a calibration curve made by fluorescence correlation spectroscopy measurements of serial dilutions of a GFP solution using the same microscope.

Tracking

HP1 foci were tracked using custom-written code in Python⁸³. Mean-squared displacements were fitted with a (simplified) confined diffusion model that has previously been used to describe the mobility of nuclear bodies⁸⁴:

8

$MSD\left(t\right)=r_{\mathrm{c}}^2\cdot \left(1-\exp \left(-\frac{4Dt}{r_{\mathrm{c}}^2}\right)\right)$Here, r_c is the confinement radius and D is the diffusion coefficient. Fit results are indicated in Supplementary Fig. 8b.

Image analysis

Images were analyzed with custom scripts written in R⁸⁵ and Python⁸³. Nuclei and heterochromatin foci were segmented in z-projected images based on their increased intensity, and their sizes were subsequently determined from the segmentation results. The average intensity of each cell nucleus was also measured and converted to a concentration using a recombinant GFP standard that was imaged side-by-side at the same microscope using the same laser and detector settings. A representative calibration curve is shown in Supplementary Fig. 9a.

Total RNA extraction, RNA sequencing and RNA-seq analysis

For RNA-sequencing experiments, cells were incubated for 48 h with doxycycline to induce the expression of GFP-tagged HP1 homologs. Cells were inspected at the microscope to confirm heterochromatin coalescence in cells expressing GFP-tagged Swi6 and dHP1a. As not all cells expressed the GFP-tagged HP1 homolog, cells were subjected to fluorescence-activated cell sorting (FACS) to obtain only GFP-positive cells (Supplementary Fig. 10). The same sorting gate was used for all three cell lines. Part of the sorted cells was kept for quantification (Supplementary Fig. 9b–d), the rest was resuspended in TRI reagent (T9424, Sigma) supplemented with ERCC spike-in controls (4456740, Thermo Fisher; per million of sorted cells, 0.05 µL spike-in controls were added). Total RNA was extracted from the latter samples using the MasterPure Complete DNA and RNA Purification Kit (MC85200, Biosearch Technologies) after digestion of DNA with Baseline-ZERO DNase (DB0715K, Biosearch Technologies). Strand-specific library preparation and deep sequencing was carried out by BGI Genomics, yielding ~50 million clean reads per sample. Reads were mapped with STAR⁸⁶ (using default parameters), and differential expression analysis was conducted in R⁸⁵ using TEtranscripts⁸⁷ and DESeq2⁸⁸. For the latter, cells expressing one of the GFP-tagged HP1 homologs were compared to the three non-induced samples. To assess the location of differentially expressed genes with respect to pericentromere-associated domains (PADs)⁵⁴, the genomic coordinates of the latter were downloaded (GSE65618_sat4C_HSMM_PAD_calls.txt) and converted to GRCm39/mm39. The overlap between differentially expressed genes and PADs called in all four cell types (mouse embryonic stem cells, neural precursor cells, astrocytes and thymocytes) was determined using R⁸⁵. Gene ontology analysis was conducted using the gprofiler2 package⁸⁹. To estimate the amount of major satellite repeat transcripts per cell, we first plotted the counts on ERCC spike-ins versus their known absolute amounts per cell, and fitted a linear regression model (Supplementary Fig. 9e). Based on the slopes of these curves, we concluded that a concentration of 1 attogram/cell corresponded to 100-280 counts. Using these conversion factors, we converted counts on “GSAT_MM:Satellite:Satellite” to attograms/cell. These values are plotted in Fig. 6d.

Quantification of expression levels by cytometry and in-gel fluorescence imaging

To quantify expression levels of sorted cells, we calibrated the fluorescence signal detected by the cytometer using Rainbow calibration particles (Thermo Fisher). To this end, we sorted Rainbow calibration particles using the same settings we used for cell sorting before RNA-seq, and we subsequently imaged the sorted calibration particles by confocal microscopy, using the same excitation/detection settings that we used to image cells. Based on the calibration curve that relates intensities to concentrations (Supplementary Fig. 9a), we converted the fluorescence signals detected by the cytometer to concentrations. Prior to conversion, signals detected by the cytometer, which correspond to the integrated intensities of the sorted objects, were multiplied with the ratio between the volume of calibration particles, which we determined by microscopy to be 56 µm³, and the nuclear volume of NIH3T3 cells, which amounts to ~900 µm³⁹⁰. The resulting concentrations of sorted cells are shown in Supplementary Fig. 9b.

To quantify expression levels of sorted cells via in-gel fluorescence imaging (Supplementary Fig. 9c, d), the latter were lysed in RIPA buffer (150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris pH 8) supplemented with complete protease inhibitor cocktail (05056489001, Roche) for 30 min at 4 °C. Subsequently, samples were centrifuged at 450 x g in a table top centrifuge. For the recombinant protein standard, the respective protein amounts were diluted in RIPA buffer in the same final volumes as the cell lysate. Samples were denatured for 10 min at 60 °C in Laemmli Sample Buffer (1610747, Bio-Rad) supplemented with 50 mM DTT. Under these conditions, GFP remained autofluorescent. Proteins were separated on a Criterion TGX Stain-Free Any kD SDS polyacrylamide gel (5678123, Bio-Rad) and GFP fluorescence signals were imaged at a Typhoon Trio Imager (GE). Subsequently, proteins were visualized via Stain-Free imaging at a ChemiDoc system (Bio-Rad). Note that the Stain-Free image likely contains a contribution from the GFP fluorescence signal and was not used for quantification. In order to estimate the nuclear concentration of GFP fusions, the obtained mass was converted to moles and was then divided by the cell number and the nuclear volume of NIH3T3 cells, which amounts to ~900 µm³⁹⁰. For the conversion from mass to moles, the molecular weight of the recombinant protein standard was used (MW of dHP1a-GFP: 50 kDa).

Protein disorder prediction

To predict the disorder of individual HP1 homologs, we obtained their amino acid sequences using UniProt, using the following IDs: P45973 (human), Q61686 (mouse), A0A8J0U8N9 (clawed frog), A1L223 (zebrafish), P05205 (fruit fly), P40381 (fission yeast). We then submitted the sequences to PONDR³² and summed the disorder score for each amino acid to obtain the cumulative predicted disorder that is plotted in Fig. 1a. Divergence times for the respective organisms were obtained from Timetree³³.

Expression levels from published mass spectrometry data

The expression levels of individual HP1 homologs in ppm were obtained from the PaxDB database³⁹ that contains published quantitative mass spectrometry data from different organisms that are normalized in a comparable manner. The values plotted in Fig. 3a correspond to the weighted averages obtained for the whole organisms based on several publications that are listed in PaxDB along with the weights. Error bars correspond to the weighted standard deviations of the values that enter the weighted averages. The Swi6 concentration reported in the text was derived from the copy number of 19,400¹² and the nuclear volume of fission yeast of ~13 µm³⁹¹. In fruit fly, previous studies have estimated the stoichiometry between dHP1a and nucleosomes to 1:15 in third instar larva⁴⁹, and to 1:4–1:8 in 2-4 h-old fly embryos⁴⁸ (197,600 dHP1a copies, 1.5–3.0 million copies of each core histone, 2 molecules of each core histone per nucleosome). Assuming a nucleosome concentration of ~100 µM like in other species^92,93, these values translate into dHP1a concentrations of 7–25 µM.

Statistics and reproducibility

Sample sizes were chosen according to standard practices in the field. No data were excluded from the analyses. Experiments were not randomized. The Investigators were not blinded to allocation during experiments and outcome assessment. The microscopy images presented in Fig. 2a, b and Supplementary Fig. 2a and Supplementary Fig. 3 are representative images of a pool of at least three independent experiments. The FRAP curves presented in Fig. 2c are the means of 7 (GFP-mHP1α), 6 (GFP-dHP1a) and 8 (GFP-Swi6) independent experiments (errors represent standard deviations). The anisotropy curves presented in Fig. 3d are the means of three independent experiments, errors represent standard deviations. The microscopy images presented in Fig. 3b, c, f, g are representative images of a pool of 5 experiments. All quantifications in Fig. 4b are presented as dot plots / box plots (with the upper and lower bounds of the box corresponding to the first and third quartiles, the line in the center to the median of the data, and the whiskers showing the rest of the distribution without outliers), using data from 12 (GFP-mHP1α, low levels), 19 (GFP-mHP1α, medium levels), 8 (GFP-mHP1α, high levels), 5 (GFP-dHP1a, low levels), 10 (GFP-dHP1a, medium levels), 26 (GFP-dHP1a, high levels), 8 (GFP-Swi6, low levels), 10 (GFP-Swi6, medium levels) and 33 (GFP-Swi6, high levels) cells. Outliers were not plotted. Two-sided Welch’s t-tests were performed to compare the distributions, yielding a p-value of 0.0009 when comparing the groups with low and high expression of GFP-Swi6, a p-value of 0.0047 when comparing the groups with low and high expression of GFP-dHP1a, and a p-value of 0.1671 when comparing the groups with low and high expression of GFP-mHP1α. The microscopy images presented in Fig. 4a are representative images of these groups. The microscopy images presented in Fig. 4c, e are representative images of a pool of 4 (Fig. 4c, top), 4 (Fig. 4c, bottom) and 17 (Fig. 4e) experiments. The FRAP curves presented in Fig. 4d are the means of 14 (GFP-mHP1α), 8 (GFP-dHP1a) and 28 (GFP-Swi6) independent experiments (errors represent standard deviations). All quantifications in Supplementary Fig. 7b are presented as dot plots / box plots (with the upper and lower bounds of the box corresponding to the first and third quartiles, the line in the center to the median of the data, and the whiskers showing the rest of the distribution without outliers), using data from 17 (GFP-mHP1α, low levels), 18 (GFP-mHP1α, high levels), 11 (GFP-dHP1a, low levels), 22 (GFP-dHP1a, high levels), 15 (GFP-Swi6, low levels) and 22 (GFP-Swi6, high levels) cells. The microscopy images presented in Supplementary Fig. 7a are representative images of these groups. The microscopy images presented in Fig. 5 are representative images of a pool of 15 (BMEL TKO cells expressing GFP-mHP1α), 21 (BMEL TKO cells expressing GFP-dHP1a) and 7 (BMEL TKO cells expressing GFP-Swi6) cells.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Acknowledgements

We thank Sylvie Tournier and Yannick Gachet for providing the coding sequence for Swi6, Bas van Steensel for providing a plasmid encoding dHP1a (Addgene plasmid #59221), Maria-Elena Torres-Padilla for providing a plasmid encoding a TALE against mouse major satellite sequences (Addgene plasmid #47878), Kerstin Bystricky for providing a plasmid encoding H2B-mCherry, and Odile Mondesert for providing a plasmid encoding CENP-A-mCherry. We thank Zoé Ferrand and Katalin Tóth for support and equipment for fluorescence anisotropy experiments, respectively, and Elodie Riant for support with flow cytometry. Confocal microscopy and cytometry experiments were carried out at the TRI-Genotoul facilities. S.B. received a fellowship from La Region Occitanie, and D.L. received a fellowship from FRM (Fondation pour la Recherche Médicale). M.D.T. and S.H. were supported by the French National Research Agency (ANR-21-CE11-0037). Part of this work was funded by grants from the European Research Council (ERC-2018-StG 804023 and ERC-2024-CoG 101170239 to F.E.). Views and opinions expressed are, however, those of the authors only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them.

Author contributions

Acquisition of data: S.B., D.L., F.M., S.H., M.D.T., M.A. Material preparation: S.B., D.L., F.M., C.N., F.M.C. Analysis of data: S.B, F.M., D.L., S.H., M.D.T., F.E. Drafting of manuscript: F.E. Reviewing of manuscript: S.B, F.M., D.L., M.A., F.M.C., S.H., M.D.T., F.E. Supervision: F.E. Study design and coordination: F.E.

Peer review

Peer review information

Nature Communications thanks Michael Hendzel and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

RNA-sequencing data were deposited at GEO under accession number GSE271404.  are provided with this paper.

Code availability

No custom algorithms that are central to the research presented here have been utilized. Scripts for image analysis and MOCHA-FRAP analysis are available in the GitHub repositories https://github.com/dymochro/ImageQuant and https://github.com/dymochro/MOCHA, respectively.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-025-61749-3.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. James, TC; Elgin, SC. Identification of a nonhistone chromosomal protein associated with heterochromatin in Drosophila melanogaster and its gene. Mol. Cell. Biol.; 1986; 6, pp. 3862-3872.1:CAS:528:DyaL2sXhtFCnsb8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/3099166][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC367149]

2. Lorentz, A; Ostermann, K; Fleck, O; Schmidt, H. Switching gene swi6, involved in repression of silent mating-type loci in fission yeast, encodes a homologue of chromatin-associated proteins from Drosophila and mammals. Gene; 1994; 143, pp. 139-143.1:CAS:528:DyaK2cXlt1yiurk%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/8200530]

3. Singh, PB et al. A sequence motif found in aDrosophila heterochromatin protein is conserved in animals and plants. Nucleic Acids Res; 1991; 19, pp. 789-794.1:CAS:528:DyaK3MXlsFaitLY%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/1708124][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC333712]

4. Jacobs, SA; Khorasanizadeh, S. Structure of HP1 chromodomain bound to a lysine 9-methylated histone H3 tail. Science; 2002; 295, pp. 2080-2083.2002Sci..295.2080J1:CAS:528:DC%2BD38Xit1Ohsb8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/11859155]

5. Nielsen, PR et al. Structure of the HP1 chromodomain bound to histone H3 methylated at lysine 9. Nature; 2002; 416, pp. 103-107.2002Natur.416.103N1:CAS:528:DC%2BD38Xitlaisbg%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/11882902]

6. Thiru, A et al. Structural basis of HP1/PXVXL motif peptide interactions and HP1 localisation to heterochromatin. EMBO J.; 2004; 23, pp. 489-499.1:CAS:528:DC%2BD2cXhtV2qsbY%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/14765118][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1271814]

7. Hiragami, K; Festenstein, R. Heterochromatin protein 1: a pervasive controlling influence. Cell Mol. Life Sci.; 2005; 62, pp. 2711-2726.1:CAS:528:DC%2BD28XhsVWqsw%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/16261261][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11139183]

8. Maison, C; Almouzni, G. HP1 and the dynamics of heterochromatin maintenance. Nat. Rev. Mol. Cell Biol.; 2004; 5, pp. 296-304.1:CAS:528:DC%2BD2cXivV2itr8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15071554]

9. Nielsen, AL et al. Heterochromatin formation in mammalian cells. Mol. Cell; 2001; 7, pp. 729-739.1:CAS:528:DC%2BD3MXjsVent7Y%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/11336697]

10. Rosnoblet, C; Vandamme, J; Völkel, P; Angrand, P-O. Analysis of the human HP1 interactome reveals novel binding partners. Biochem. Biophys. Res. Commun.; 2011; 413, pp. 206-211.1:CAS:528:DC%2BC3MXht1ehsLbJ [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21888893]

11. Müller-Ott, K et al. Specificity, propagation, and memory of pericentric heterochromatin. Mol. Syst. Biol.; 2014; 10, 746. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25134515][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4299515]

12. Sadaie, M et al. Balance between distinct HP1 family proteins controls heterochromatin assembly in fission yeast. Mol. Cell. Biol.; 2008; 28, pp. 6973-6988.1:CAS:528:DC%2BD1cXhsVCjsb%2FF [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18809570][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2593388]

13. Lee, D. H., Ryu, H. W., Kim, G. W. & Kwon, S. H. Comparison of three heterochromatin protein 1 homologs in Drosophila. J. Cell Sci. 132, jcs222729 (2019).

14. Hediger, F; Gasser, SM. Heterochromatin protein 1: don’t judge the book by its cover!. Curr. Opin. Genet. Dev.; 2006; 16, pp. 143-150.1:CAS:528:DC%2BD28Xis1ehtrc%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/16503133]

15. Vermaak, D; Malik, HS. Multiple roles for heterochromatin protein 1 genes in Drosophila. Annu. Rev. Genet.; 2009; 43, pp. 467-492.1:CAS:528:DC%2BD1MXhsF2qtrbO [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19919324]

16. Fischer, T et al. Diverse roles of HP1 proteins in heterochromatin assembly and functions in fission yeast. Proc. Natl. Acad. Sci. USA; 2009; 106, pp. 8998-9003.2009PNAS.106.8998F1:CAS:528:DC%2BD1MXntlyjt7o%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19443688][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2690032]

17. Motamedi, MR et al. HP1 proteins form distinct complexes and mediate heterochromatic gene silencing by nonoverlapping mechanisms. Mol. Cell; 2008; 32, pp. 778-790.1:CAS:528:DC%2BD1MXjtFGmug%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19111658][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2735125]

18. Strom, A. R. et al. HP1α is a chromatin crosslinker that controls nuclear and mitotic chromosome mechanics. Elife10, e63972 (2021).

19. Bártová, E et al. Function of heterochromatin protein 1 during DNA repair. Protoplasma; 2017; 254, pp. 1233-1240. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28236007]

20. Singh, PB. HP1 proteins-what is the essential interaction?. Genetika; 2010; 46, pp. 1424-1429.1:STN:280:DC%2BC3M7jtlyrtA%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21250545]

21. Larson, AG et al. Liquid droplet formation by HP1α suggests a role for phase separation in heterochromatin. Nature; 2017; 547, pp. 236-240.2017Natur.547.236L1:CAS:528:DC%2BC2sXhtVequrbI [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28636604][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5606208]

22. Strom, AR et al. Phase separation drives heterochromatin domain formation. Nature; 2017; 547, pp. 241-245.2017Natur.547.241S1:CAS:528:DC%2BC2sXhtVequrfN [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28636597][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6022742]

23. Sanulli, S et al. HP1 reshapes nucleosome core to promote phase separation of heterochromatin. Nature; 2019; 575, pp. 390-394.2019Natur.575.390S1:CAS:528:DC%2BC1MXitFWnur%2FP [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31618757][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7039410]

24. Erdel, F et al. Mouse Heterochromatin Adopts Digital Compaction States without Showing Hallmarks of HP1-Driven Liquid-Liquid Phase Separation. Mol. Cell; 2020; 78, pp. 236-249.e7.1:CAS:528:DC%2BB3cXjvVKnsLY%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32101700][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7163299]

25. Keenen, M. M. et al. HP1 proteins compact DNA into mechanically and positionally stable phase separated domains. Elife10, e64563 (2021).

26. Canzio, D; Larson, A; Narlikar, GJ. Mechanisms of functional promiscuity by HP1 proteins. Trends Cell Biol.; 2014; 24, pp. 377-386.1:CAS:528:DC%2BC2cXjsF2iurY%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24618358][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4077871]

27. Falk, M et al. Heterochromatin drives compartmentalization of inverted and conventional nuclei. Nature; 2019; 570, pp. 395-399.2019Natur.570.395F1:CAS:528:DC%2BC1MXhtFWqtLfE [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31168090][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7206897]

28. McSwiggen, DT; Mir, M; Darzacq, X; Tjian, R. Evaluating phase separation in live cells: diagnosis, caveats, and functional consequences. Genes Dev.; 2019; 33, pp. 1619-1634.1:CAS:528:DC%2BB3cXhtVSlt73J [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31594803][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6942051]

29. Musacchio, A. On the role of phase separation in the biogenesis of membraneless compartments. EMBO J.; 2022; 41, e109952.1:CAS:528:DC%2BB38Xislegurc%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35107832][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8886532]

30. Peng, Z et al. Exceptionally abundant exceptions: comprehensive characterization of intrinsic disorder in all domains of life. Cell Mol. Life Sci.; 2015; 72, pp. 137-151.1:CAS:528:DC%2BC2cXhtVSgsr7E [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24939692]

31. Zeng, W; Ball, AR; Yokomori, K. HP1: heterochromatin binding proteins working the genome. Epigenetics; 2010; 5, pp. 287-292.1:CAS:528:DC%2BC3cXpslKlsbs%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20421743]

32. Xue, B; Dunbrack, RL; Williams, RW; Dunker, AK; Uversky, VN. PONDR-FIT: a meta-predictor of intrinsically disordered amino acids. Biochim. Biophys. Acta; 2010; 1804, pp. 996-1010.1:CAS:528:DC%2BC3cXis1Sjtr0%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20100603][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2882806]

33. Kumar, S. et al. Timetree 5: an expanded resource for species divergence times. Mol. Biol. Evol. 39, msac174 (2022).

34. Wang, G et al. Conservation of heterochromatin protein 1 function. Mol. Cell. Biol.; 2000; 20, pp. 6970-6983.1:CAS:528:DC%2BD3cXmsVamsb8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10958692][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC88772]

35. Giauque, CC; Bickel, SE. Heterochromatin-associated proteins HP1a and Piwi collaborate to maintain the association of achiasmate homologs in drosophila oocytes. Genetics; 2016; 203, pp. 173-189.1:CAS:528:DC%2BC28XitFSit73P [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26984058][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4858772]

36. Wang, L et al. Histone modifications regulate chromatin compartmentalization by contributing to a phase separation mechanism. Mol. Cell; 2019; 76, pp. 646-659.e6.1:CAS:528:DC%2BC1MXhvVWktbnM [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31543422]

37. Muzzopappa, F et al. Detecting and quantifying liquid-liquid phase separation in living cells by model-free calibrated half-bleaching. Nat. Commun.; 2022; 13, 2022NatCo.13.7787M1:CAS:528:DC%2BB38XjtFOmtr%2FK [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36526633][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9758202]7787.

38. Yewdall, NA et al. ATP:Mg2+ shapes material properties of protein-RNA condensates and their partitioning of clients. Biophys. J.; 2022; 121, pp. 3962-3974.2022BpJ..121.3962Y1:CAS:528:DC%2BB38Xit12qur%2FK [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36004782][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9674983]

39. Huang, Q; Szklarczyk, D; Wang, M; Simonovic, M; von Mering, C. Paxdb 5.0: curated protein quantification data suggests adaptive proteome changes in yeasts. Mol. Cell Proteom.; 2023; 22, 100640.1:CAS:528:DC%2BB3sXitVeht7zN

40. Qin, W et al. HP1β carries an acidic linker domain and requires H3K9me3 for phase separation. Nucleus; 2021; 12, pp. 44-57.43236591:CAS:528:DC%2BB3MXlslajtbg%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33660589][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7939559]

41. Phan, T. M., Kim, Y. C., Debelouchina, G. T. & Mittal, J. Interplay between charge distribution and DNA in shaping HP1 paralog phase separation and localization. Elife12, RP90820 (2024).

42. Brasher, SV et al. The structure of mouse HP1 suggests a unique mode of single peptide recognition by the shadow chromo domain dimer. EMBO J.; 2000; 19, pp. 1587-1597.1:CAS:528:DC%2BD3cXislymt7g%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10747027][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC310228]

43. Heckmeier, PJ et al. Determining the stoichiometry of small protein oligomers using steady-state fluorescence anisotropy. Biophys. J.; 2020; 119, pp. 99-114.2020BpJ..119..99H1:CAS:528:DC%2BB3cXhtFens7rO [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32553128][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7335908]

44. Mendez, DL et al. The HP1a disordered C terminus and chromo shadow domain cooperate to select target peptide partners. Chembiochem; 2011; 12, pp. 1084-1096.1:CAS:528:DC%2BC3MXkvVKlsLc%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21472955][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3154745]

45. Canzio, D et al. A conformational switch in HP1 releases auto-inhibition to drive heterochromatin assembly. Nature; 2013; 496, pp. 377-381.2013Natur.496.377C1:CAS:528:DC%2BC3sXkvFOktr0%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23485968][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3907283]

46. Guenatri, M; Bailly, D; Maison, C; Almouzni, G. Mouse centric and pericentric satellite repeats form distinct functional heterochromatin. J. Cell Biol.; 2004; 166, pp. 493-505.1:CAS:528:DC%2BD2cXmvFehtb4%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15302854][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2172221]

47. Müller, KP et al. Multiscale analysis of dynamics and interactions of heterochromatin protein 1 by fluorescence fluctuation microscopy. Biophys. J.; 2009; 97, pp. 2876-2885.2009BpJ..97.2876M [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19948116][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2784559]

48. Bonnet, J et al. Quantification of proteins and histone marks in drosophila embryos reveals stoichiometric relationships impacting chromatin regulation. Dev. Cell; 2019; 51, pp. 632-644.e6.1:CAS:528:DC%2BC1MXitVSlsr%2FP [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31630981]

49. Lu, BY; Emtage, PC; Duyf, BJ; Hilliker, AJ; Eissenberg, JC. Heterochromatin protein 1 is required for the normal expression of two heterochromatin genes in Drosophila. Genetics; 2000; 155, pp. 699-708.1:CAS:528:DC%2BD3cXkslyqtbs%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10835392][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1461102]

50. Miyanari, Y; Ziegler-Birling, C; Torres-Padilla, M-E. Live visualization of chromatin dynamics with fluorescent TALEs. Nat. Struct. Mol. Biol.; 2013; 20, pp. 1321-1324.1:CAS:528:DC%2BC3sXhsFOiu7fL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24096363]

51. Saksouk, N et al. The mouse HP1 proteins are essential for preventing liver tumorigenesis. Oncogene; 2020; 39, pp. 2676-2691.1:CAS:528:DC%2BB3cXislSgsrc%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32020053]

52. Maeda, R; Tachibana, M. HP1 maintains protein stability of H3K9 methyltransferases and demethylases. EMBO Rep.; 2022; 23, e53581.1:CAS:528:DC%2BB38XjslSjsL4%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35166421][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8982598]

53. Mateos-Langerak, J et al. Pericentromeric heterochromatin domains are maintained without accumulation of HP1. Mol. Biol. Cell; 2007; 18, pp. 1464-1471.1:CAS:528:DC%2BD2sXktVSmtbk%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17314413][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1838966]

54. Wijchers, PJ et al. Characterization and dynamics of pericentromere-associated domains in mice. Genome Res; 2015; 25, pp. 958-969.1:CAS:528:DC%2BC2MXht1SltbrL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25883320][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4484393]

55. Cantaloube, S; Romeo, K; Le Baccon, P; Almouzni, G; Quivy, J-P. Characterization of chromatin domains by 3D fluorescence microscopy: An automated methodology for quantitative analysis and nuclei screening. Bioessays; 2012; 34, pp. 509-517.1:CAS:528:DC%2BC38Xnt1Gltrs%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22451098]

56. Novo, CL et al. Satellite repeat transcripts modulate heterochromatin condensates and safeguard chromosome stability in mouse embryonic stem cells. Nat. Commun.; 2022; 13, 2022NatCo.13.3525N1:CAS:528:DC%2BB38XhsF2mt73P [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35725842][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9209518]3525.

57. Brändle, F; Frühbauer, B; Jagannathan, M. Principles and functions of pericentromeric satellite DNA clustering into chromocenters. Semin. Cell Dev. Biol.; 2022; 128, pp. 26-39. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35144860]

58. Zenk, F et al. HP1 drives de novo 3D genome reorganization in early Drosophila embryos. Nature; 2021; 593, pp. 289-293.2021Natur.593.289Z1:CAS:528:DC%2BB3MXoslKks7w%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33854237][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8116211]

59. Tortora, M. M. C., Brennan, L., Karpen, G. & Jost, D. HP1-driven phase separation recapitulates the thermodynamics and kinetics of heterochromatin condensate formation. Proc. Natl. Acad. Sci. USA120, e2211855120 (2023).

60. Mitrentsi, I et al. Heterochromatic repeat clustering imposes a physical barrier on homologous recombination to prevent chromosomal translocations. Mol. Cell; 2022; 82, pp. 2132-2147.e6.1:CAS:528:DC%2BB38XhtVGhtL7J [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35447083][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9616805]

61. Bosch-Presegué, L et al. Mammalian HP1 isoforms have specific roles in heterochromatin structure and organization. Cell Rep.; 2017; 21, pp. 2048-2057. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29166597]

62. Brero, A et al. Methyl CpG-binding proteins induce large-scale chromatin reorganization during terminal differentiation. J. Cell Biol.; 2005; 169, pp. 733-743.1:CAS:528:DC%2BD2MXltVanur4%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15939760][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2171616]

63. Hendrich, B; Tweedie, S. The methyl-CpG binding domain and the evolving role of DNA methylation in animals. Trends Genet; 2003; 19, pp. 269-277.1:CAS:528:DC%2BD3sXjtFeqt7w%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/12711219]

64. Jagannathan, M., Cummings, R. & Yamashita, Y. M. A conserved function for pericentromeric satellite DNA. Elife7 (2018).

65. Novo, CL et al. The pluripotency factor Nanog regulates pericentromeric heterochromatin organization in mouse embryonic stem cells. Genes Dev.; 2016; 30, pp. 1101-1115.1:CAS:528:DC%2BC28XhsFWrtb%2FF [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27125671][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4863740]

66. Wang, D; Hsieh, M; Li, W-H. A general tendency for conservation of protein length across eukaryotic kingdoms. Mol. Biol. Evol.; 2005; 22, pp. 142-147.1:CAS:528:DC%2BD2cXhtFansrbF [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15371528]

67. Zhang, W. et al. Evolutional heterochromatin condensation delineates chromocenter formation and retrotransposon silencing in plants. Nat. Plants10, 1215–1230 (2024).

68. van Steensel, B; Henikoff, S. Identification of in vivo DNA targets of chromatin proteins using tethered dam methyltransferase. Nat. Biotechnol.; 2000; 18, pp. 424-428. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10748524]

69. Kowarz, E; Löscher, D; Marschalek, R. Optimized Sleeping Beauty transposons rapidly generate stable transgenic cell lines. Biotechnol. J.; 2015; 10, pp. 647-653.1:CAS:528:DC%2BC2MXkvVWltr8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25650551]

70. Caudron-Herger, M et al. Alu element-containing RNAs maintain nucleolar structure and function. EMBO J.; 2015; 34, pp. 2758-2774.1:CAS:528:DC%2BC2MXhs1Cntr3J [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26464461][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4682651]

71. Tully, MD et al. BioSAXS at European Synchrotron radiation facility - extremely brilliant source: BM29 with an upgraded source, detector, robot, sample environment, data collection and analysis software. J. Synchrotron Radiat.; 2023; 30, pp. 258-266. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36601945][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9814054]

72. Pernot, P et al. Upgraded ESRF BM29 beamline for SAXS on macromolecules in solution. J. Synchrotron Radiat.; 2013; 20, pp. 660-664.1:CAS:528:DC%2BC3sXpsVemsr8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23765312][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3943554]

73. Kieffer, J et al. New data analysis for BioSAXS at the ESRF. J. Synchrotron Radiat.; 2022; 29, pp. 1318-1328. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36073892][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9455220]

74. Tully, M. D., Tarbouriech, N., Rambo, R. P. & Hutin, S. Analysis of SEC-SAXS data via EFA deconvolution and Scatter. J. Vis. Exp. https://doi.org/10.3791/61578 (2021).

75. Tria, G; Mertens, HDT; Kachala, M; Svergun, DI. Advanced ensemble modelling of flexible macromolecules using X-ray solution scattering. IUCrJ; 2015; 2, pp. 207-217.1:CAS:528:DC%2BC2MXjs1Gnu7c%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25866658][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4392415]

76. Berman, HM et al. The protein data bank. Nucleic Acids Res; 2000; 28, pp. 235-242.2000spam.book...B1:CAS:528:DC%2BD3cXhvVKjt7w%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10592235][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC102472]

77. Humphrey, W., Dalke, A. & Schulten, K. VMD: visual molecular dynamics. J. Mol. Graph.14, 33–8, 27 (1996).

78. Lowary, PT; Widom, J. New DNA sequence rules for high affinity binding to histone octamer and sequence-directed nucleosome positioning. J. Mol. Biol.; 1998; 276, pp. 19-42.1:CAS:528:DyaK1cXhs1yju7Y%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/9514715]

79. Huynh, VAT; Robinson, PJJ; Rhodes, D. A method for the in vitro reconstitution of a defined “30 nm” chromatin fibre containing stoichiometric amounts of the linker histone. J. Mol. Biol.; 2005; 345, pp. 957-968.1:CAS:528:DC%2BD2MXis1agsw%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15644197]

80. Ojha, N; Rainey, KH; Patterson, GH. Imaging of fluorescence anisotropy during photoswitching provides a simple readout for protein self-association. Nat. Commun.; 2020; 11, 2020NatCo.11..21O1:CAS:528:DC%2BB3cXmt1Wqug%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31911590][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6946710]21.

81. Axelrod, D. Carbocyanine dye orientation in red cell membrane studied by microscopic fluorescence polarization. Biophys. J.; 1979; 26, pp. 557-573.1979BpJ..26.557A1:CAS:528:DyaE1MXkslSns7o%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/263688][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1328570]

82. Mátés, L et al. Molecular evolution of a novel hyperactive sleeping beauty transposase enables robust stable gene transfer in vertebrates. Nat. Genet.; 2009; 41, pp. 753-761. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19412179]

83. Van Rossum, G. & Drake, F. L. Python 3 Reference Manual. (CreateSpace, Scotts Valley, CA, 2009).

84. Görisch, SM et al. Nuclear body movement is determined by chromatin accessibility and dynamics. Proc. Natl. Acad. Sci. USA; 2004; 101, pp. 13221-13226.2004PNAS.10113221G [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15331777][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC516551]

85. Team, R. C. R: A LANGUAGE and ENVIRONMENT FOR STATISTICAL COMPUTING. R Foundation for Statistical Computing, Vienna, Austria (2015).

86. Dobin, A et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics; 2013; 29, pp. 15-21.1:CAS:528:DC%2BC38XhvV2gsbnF [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23104886]

87. Jin, Y; Tam, OH; Paniagua, E; Hammell, M. TEtranscripts: a package for including transposable elements in differential expression analysis of RNA-seq datasets. Bioinformatics; 2015; 31, pp. 3593-3599.1:CAS:528:DC%2BC28Xht1CitL%2FN [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26206304][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4757950]

88. Love, MI; Huber, W; Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol.; 2014; 15, [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25516281][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4302049]550.

89. Kolberg, L., Raudvere, U., Kuzmin, I., Vilo, J. & Peterson, H. gprofiler2 - an R package for gene list functional enrichment analysis and namespace conversion toolset g:Profiler. F1000Res. 9, ELIXIR-709 (2020).

90. Katiyar, A et al. Nuclear size changes caused by local motion of cell boundaries unfold the nuclear lamina and dilate chromatin and intranuclear bodies. Soft Matter; 2019; 15, pp. 9310-9317.2019SMat..15.9310K1:CAS:528:DC%2BC1MXitVejsr3E [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31674621][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6870954]

91. Gonzalez, Y et al. Nuclear shape, growth and integrity in the closed mitosis of fission yeast depend on the Ran-GTPase system, the spindle pole body and the endoplasmic reticulum. J. Cell Sci.; 2009; 122, pp. 2464-2472.1:CAS:528:DC%2BD1MXpvFGmu7g%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19571115][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2704881]

92. Hihara, S et al. Local nucleosome dynamics facilitate chromatin accessibility in living mammalian cells. Cell Rep.; 2012; 2, pp. 1645-1656.1:CAS:528:DC%2BC3sXlt1ygsg%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23246002]

93. Weidemann, T et al. Counting nucleosomes in living cells with a combination of fluorescence correlation spectroscopy and confocal imaging. J. Mol. Biol.; 2003; 334, pp. 229-240.1:CAS:528:DC%2BD3sXos1Chtrk%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/14607115]