Extended Data Fig. 4: Condensation of the endogenous Utx is important for ESC differentiation.

a. Schematic of FLAG-mEGFP Knock-in (KI) at Utx in ESCs. b. Genomic PCR products by indicated primers for different clones. The predicted sizes are 961 bp (by F1+R1) and 954 bp (by F2+R2) for the KI clones. c. Immunoblotting for KI clones by green circles (neighbouring lanes are clones of mixed population). d. Schematic for generating a Utx KO ESC clone. e. Sequencing results showing the 4 bp loss in Utx in the KO clone. f. Immunoblotting for the Utx KO clone. g. Schematic for generating Utx ΔcIDR ESC clones. These clones were generated by cIDR deletion followed with selection for in-frame rejoined Utx. h. Genomic PCR products for different clones. The bottom right gel used the primer pair in yellow shown in g, with predicted size of 205 bp for ΔcIDR. The other gels used the primer pair in black in g, with predicted size of 1040 bp for ΔcIDR. i. Sequencing of the indicated ΔcIDR clones. The in-frame deleted genomic regions are shown in the square brackets. The deletion of amino acids additional to the cIDR was due to the design of the gRNAs to ensure efficient recognition of CRISPR/Cas9. These residues are all in the disordered region. A phenylalanine was also added (in red) by the repair process in clone #72. j. Immunoblotting for the Utx ΔcIDR clones. k. Schematic of FLAG-mEGFP KI at Utx in the Utx ΔcIDR ESC clones. l. Genomic PCR products for different clones using the F2 and R2 primer pair shown in a. m. Immunoblotting for mEGFP-tagged WT and UtxΔ cIDR clones. n. Representative confocal microscopic images of ESCs that did not have mEGFP KI (Negative) or had mEGFP KI in the Utx WT or ΔcIDR clones. Two different samples of ΔcIDR clone #25 that had relatively high and low mEGFP signal levels are here to show that the signal levels did not affect the fuzziness or how dispersed UTX was. o. Granularity of the mEGFP signal was plotted as mean ± SD. n = 28 Utx WT and 38 ΔcIDR cells, each containing two different clones shown in panel n. Granularity, or mEGFP signal fluctuation, was calculated as SD of the relative signal intensity of a line plot (normalized to the mean signal intensity of the entire line) at each pixel on the line through an individual cell. P value by two-sided t-test. The line plot was generated by ImageJ, and an example is shown as the yellow line on a mEGFP-ΔcIDR clone in n. p. Venn diagrams for genes that, compared to the corresponding WT embryoid bodies, were downregulated (left) or upregulated (right) by KO (cyan) or ΔcIDR (salmon, two clones) in late embryoid bodies (day 8 or 9 in differentiation). Fold changes are indicated. From RNA-seq of two 2 independent differentiation assays. P values by two-sided Fisher’s exact test. q. Significantly enriched Biological Processes in embryonic development for genes downregulated (↓) or upregulated (↑) over 2 fold by Utx KO or ΔcIDR in embryoid bodies late in differentiation, compared to the corresponding WT embryoid bodies. By Metascape. P values by two-sided hypergeometric test.