Mediator Condensates Localize Signaling Factors to Key Cell Identity Genes

Abstract

The gene expression programs that define the identity of each cell are controlled by master transcription factors (TFs) that bind cell-type-specific enhancers, as well as signaling factors, which bring extracellular stimuli to these enhancers. Recent studies have revealed that master TFs form phase-separated condensates with the Mediator coactivator at super-enhancers. Here, we present evidence that signaling factors for the WNT, TGF-β, and JAK/STAT pathways use their intrinsically disordered regions (IDRs) to enter and concentrate in Mediator condensates at super-enhancers. We show that the WNT coactivator β-catenin interacts both with components of condensates and DNA-binding factors to selectively occupy super-enhancer-associated genes. We propose that the cell-type specificity of the response to signaling is mediated in part by the IDRs of the signaling factors, which cause these factors to partition into condensates established by the master TFs and Mediator at genes with prominent roles in cell identity.

Keywords: JAK/STAT; TGF-β; WNT; gene regulation; signaling pathway; transcriptional condensates.

Figure 1.. Signaling factors form signaling dependent condensates at super-enhancers in vivo

1A) Immunofluorescence for β-catenin, STAT3, SMAD3 and MED1 with concurrent RNA-FISH for Nanog nascent RNA demonstrating the presence of condensed nuclear foci of the signaling factors at the Nanog super-enhancer in mES cells. Cells were grown for 24 hours in the presence of CHIR99021, LIF and Activin A to activate the WNT, JAK/STAT and TGF-β signaling pathways respectively 24 hours prior to fixation. Hoechst staining was used to determine the nuclear periphery, highlighted with a dotted line.100x objective was used for imaging on a spinning disk confocal microscope. Average RNA-FISH signal and average IF signal centered on the RNA-FISH focus for each signaling factor from at least 10 images is shown. Average signaling factor IF signal around randomly selected nuclear positions is displayed in the right most panel. Scale bars indicate 5 µm. 1B) ChIP-seq tracks displaying occupancy of β-catenin, STAT3, SMAD3 and MED1 in mES at the super-enhancer associated with the Nanog gene. Read densities are displayed in reads per million per bin (rpm/bin) and the super-enhancer is indicated with a red bar. 1C) Immunofluorescence of mES cells for the signaling factors β-catenin, STAT3 and SMAD3 in unstimulated or stimulated conditions. Cells were stimulated for 24 hours with either CHIR99021, LIF, or Activin A to activate the WNT, JAK/STAT and TGF-β signaling pathways respectively 24 hours prior to fixation. Hoechst staining was used to determine the nuclear periphery, highlighted with a dotted line.100x objective was used for imaging on a spinning disk confocal microscope. Scale bars indicate 5 µm. 1D) Top left: Representative images of FRAP experiment of mEGFP-β-catenin engineered HCT116 cells. Yellow box highlights the punctum undergoing targeted bleaching. Top right: Quantification of FRAP data for mEGFP-β-catenin puncta. Bottom left: Representative images of FRAP experiment of mEGFP-HP1α engineered HCT116 cells. Yellow box highlights region undergoing targeted bleaching. Bottom right: Quantification of FRAP data for mEGFP-HP1α puncta. Bleaching event occurs at t = 0s. For both bleached area and unbleached control, background-subtracted fluorescence intensities are plotted relative to a pre-bleach time point (t = −4s). Data are plotted as mean +/− SEM (N=9). Images were taken using the Zeiss LSM 880 confocal microscope with Airyscan detector with a 63x objective. Scale bar indicates 2 µm. 1E) Live cell imaging of endogenously-tagged mEGFP-β-catenin in HEK293T cells stimulated with CHIR99021 and imaged over time. Representative images of cells imaged over a four hour time course in the top panels. Identified foci used for quantification in the bottom panels. Foci in the nucleus were called and quantified at different time intervals for three biological replicates (right panel). Images were acquired using a Zeiss LSM 880 confocal microscope with Airyscan detector and a 63x objective. Scale bar indicates Scale bar indicates 2 µm.

Figure 2.. Purified signaling factors can form condensates in vitro

2A) Domain structures of the signaling factors used in this manuscript. DBD: DNA binding domain, PID: protein interaction domain, CC: coiled coil domain, DD: dimerization domain, SH2: Src homology domain 2. The predicted intrinsically disordered regions (IDR) are indicated with red brackets. 2B) Representative confocal images of concentration series of droplet formation assay testing homotypic droplet formation of mEGFP-β-catenin, mEGFP-STAT3 and mEGFP-SMAD3. mEGFP alone is included as a control (left panels). Quantification of the partition ratio for the signaling factors (right panels). Partition ratio was calculated by dividing the average fluorescence signal inside the droplets by the average fluorescence signal outside the droplets for at least 10 acquired images at all concentrations tested. All assays were performed in the presence of 125mM NaCl and 10% PEG-8000 was used as a crowding agent. Scale bars indicate 2 µm. 2C) Dilution droplet assay for the signaling factors. Initial droplets were formed at 1.25µM and imaged. The remaining reaction mixture was then diluted 2-fold with reaction buffer containing 4M NaCl to obtain a final salt concentration of 2M NaCl. Representative images of droplets before and after dilution are displayed. 2D) Representative images of FRAP of in vitro droplets of mEGFP-fused β-catenin, STAT3 and SMAD3 showing recovery after photobleaching in the order of seconds. Droplet formation assays were performed in the presence of 125 mM NaCl and 10% PEG-8000. Scale bars indicate 2 µm. FRAP was performed with a spinning disk confocal miscroscope using a 150x objective. 2E) Signaling factors form droplets in the presence of nuclear extracts. HEK293T cells were transfected with β-catenin, STAT3 or SMAD3 and nuclear extracts imaged using a spinning disk confocal microscope with a 150x objective. Scale bar indicates 2 µm. 2F) Phase diagrams for β-catenin, STAT3 or SMAD3 showing concentrations of salt and protein in which factors separate into a light and a dense phase (black dots) and conditions in which only a light phase is present (white dots). Droplet formation assays were performed in the presence of 5% PEG-8000 at the concentrations depicted in the diagram. Droplets were imaged with a spinning disk confocal miscroscope with a 150x objective. Partition ratio was calculated for 10 images and proteins assessed to be in a one or two phase regime by comparing the partition ratio to that of a mEGFP control.

Figure 3.. Purified signaling factors are incorporated into Mediator condensates in vitro

3A) Schematic representation of addition of signaling factor to pre-existing MED1-IDR droplets. mCherry-MED1-IDR droplets were formed and placed in a glass dish and imaged before and after addition of mEGFP-tagged signaling factors. 3B) Representative images of signaling factor incorporation into MED-IDR droplets. Preformed mCherry-MED1-IDR droplets were imaged pre and post addition of mEGFP-tagged signaling factor solution for a total of 10 mins. Signaling factor was added 30 sec after imaging acquisition started. Last image displayed corresponds to the imaging end point. 10µM of MED1-IDR-mCherry in the presence of PEG-8000 was used for droplet formation and 10uM of either mEGFP-β-catenin, mEGFP-SMAD3 or mEGFP-STAT3 in the absence of PEG-8000 was added. Scale bars indicate 2 µm. 3C) Partition ratio was calculated for pre-formed MED1-IDR-mCherry droplets that were mixed with dilute GFP-tagged signaling factor using the same conditions as in B. At least 10 images were used for quantification. Droplets were called on merged channels and signal intensity for the GFP-tagged factor in the area within the droplet compared to the intensity of the area outside of the droplet. Star indicates p-value obtained by a t-test < 0.05. 3D) Representative images of in vitro droplet assays of signaling factors with purified Mediator showing the ability of β-catenin, STAT3 and SMAD3 to interact and partition into intact Mediator droplets. Reactions were performed in the presence of 10% PEG-8000 and 300 nM signaling factor and imaged using a spinning disk confocal microscope with a 150x objective. Scale bars indicate 2 µm. 3E) Limited dilution droplet assay with near physiological concentrations of β-catenin, STAT3 and SMAD3. Indicated concentrations of the signaling factors were either added to droplet formation buffer alone (125mM NaCL and 10% PEG-8000) or in combination with 10 µM MED1-IDR. Scale bars indicate 2 µm.

Figure 4.. Phase separation of β-catenin is dependent on its IDRs

4A) Left: Diagram of the different forms of mEGFP-β-catenin proteins tested. Right: Representative confocal images of a concentration series of droplet formation assays testing homotypic droplet formation for mEGFP, mEGFP-β-catenin, mEGFP-N-terminal-IDR, mEGFP-Armadillo, mEGFP-C-terminal-IDR, mEGFP-chimera, mEGFP-2xIDR, mEGFP-full-length aromatic mutant, and mEGFP-chimera-aromatic mutant. Droplet assays were performed in 125mM NaCL and 10% PEG-8000. Scale bar indicates 1um. 4B) Representative confocal images of heterotypic droplet formation assays mixing 10 µM MED1-IDR-mCherry with 10µM of wild type full length mEGFP-β-catenin or full length aromatic mutant mEGFP-β-catenin. Scale bar indicates 1 µm. 4C) Partition ratio of factors was quantified for at least 10 images each. Droplets were called on merged channels and signal intensity for the factor in the area within the droplet compared to the intensity of the area outside the droplet.

Figure 5.. Addressing of β-catenin and activation of target genes is dependent on its IDRs

5A) Schematic of the ChIP experiment. TdTomato-tagged wild type or aromatic mutant β-catenin were stably integrated in mES cells under a doxycycline-inducible promoter. Doxycycline and an inhibitor of the WNT pathway was added to the media 24 hours prior to crosslinking. ChIP was performed using antibodies against TdTomato. TRE = Tetracycline responsive element. 5B) ChIP-qPCR of ectopically-expressed wild type and aromatic mutant β-catenin at Myc, Sp5, and Klf4 enhancers. Error bars indicate standard deviation of three replicates. Stars indicate p-values obtained by a t-test < 0.05. 5C) RT-qPCR of mRNA levels after ectopic expression of wild type or aromatic mutant β-catenin of Myc, Sp5, and Klf4. Error bars indicate standard deviation of three replicates. Stars indicate p-values obtained by a t-test < 0.05. 5D) Luciferase assay using a synthetic WNT-reporter containing 10 copies of the consensus TCF/LEF motif where wild type or aromatic mutant was overexpressed in HEK293T cells. Average of 3 biological replicates is shown. Error bars show the standard deviation. Star indicates p-value obtained by a t-test < 0.05.

Figure 6.. β-catenin-condensate interaction can occur independent of TCF factors

6A) Immunofluorescence of β-catenin in Lac-U2OS cells transfected with a Lac binding domain-CFP or a Lac binding domain-CFP-MED1-IDR construct, imaged with a 100x objective on a spinning disk confocal microscope. Hoechst staining was used to determine the nuclear periphery, highlighted with a dotted line. Quantification shows the relative intensity of β-catenin in CFP foci. Scale bar indicates 5µm. 6B) Fluorescence imaging of overexpressed TdTomato-tagged wild type or aromatic mutant β-catenin in U2OS 2–6-3 cells co-transfected with a Lac binding domain-CFP or a Lac binding domain-CFP-MED1-IDR construct, imaged with a 100x objective on a spinning disk confocal microscope. Hoechst staining was used to determine the nuclear periphery, highlighted with a dotted line. Quantification shows the relative intensity of over-expressed β-catenin forms in called CFP foci. Scale bar indicates 5µm. 6C) ChIP-qPCR for β-catenin-GFP-chimera and chimera mutant at the enhancers of SOX9, SMAD7 and KLF9 in HEK293T cells. Error bars show the standard deviation of the mean. Stars indicate p-values obtained by a t-test < 0.05 6D) Luciferase assay of cells over-expressing β-catenin-mEGFP-chimera or mutant chimera in combination with a synthetic WNT-reporter containing 10 copies of the consensus TCF/LEF motif. Average of 3 biological replicates is shown. Untransfected control and WT FL-β-catenin came from the same experiment and are the same as in Figure 5, but displayed in two different graphs. Error bars show the standard deviation. Stars indicate p-values obtained by a t-test < 0.05.

Figure 7.. Both IDRs and Armadillo domains enable selective occupancy of super-enhancer genes

7A) Cartoon depicting the different forms of β-catenin used in ChIP-seq experiments. 7B) ChIP-sequencing tracks of Nanog and mir290 showing binding of β-catenin-armadillo repeats and IDRs to super-enhancer associated genes. Read densities are displayed in reads per million per bin (rpm/bin) and the super-enhancer is indicated with a red bar. C) Quantification of ChIP-seq read densities of super-enhancers (SE) and typical enhancers (TE) of the different forms of β-catenin.