The H2Bub1-deposition complex is required for human and mouse cardiogenesis

ABSTRACT De novo variants affecting monoubiquitylation of histone H2B (H2Bub1) are enriched in human congenital heart disease. H2Bub1 is required in stem cell differentiation, cilia function, post-natal cardiomyocyte maturation and transcriptional elongation. However, how H2Bub1 affects cardiogenesis is unknown. We show that the H2Bub1-deposition complex (RNF20-RNF40-UBE2B) is required for mouse cardiogenesis and for differentiation of human iPSCs into cardiomyocytes. Mice with cardiac-specific Rnf20 deletion are embryonic lethal and have abnormal myocardium. We then analyzed H2Bub1 marks during differentiation of human iPSCs into cardiomyocytes. H2Bub1 is erased from most genes at the transition from cardiac mesoderm to cardiac progenitor cells but is preserved on a subset of long cardiac-specific genes. When H2Bub1 is reduced in iPSC-derived cardiomyocytes, long cardiac-specific genes have fewer full-length transcripts. This correlates with H2Bub1 accumulation near the center of these genes. H2Bub1 accumulation near the center of tissue-specific genes was also observed in embryonic fibroblasts and fetal osteoblasts. In summary, we show that normal H2Bub1 distribution is required for cardiogenesis and cardiomyocyte differentiation, and suggest that H2Bub1 regulates tissue-specific gene expression by increasing the amount of full-length transcripts.

H2Bub1 in wild-type mouse hearts (e9.5, e11.5, e14.5, e16.5, and P0).The loading control for the complex components is total protein (shown next to the western blots with relative intensities given underneath) and for H2Bub1 is H2B.Corrected Total Cell Fluorescence (CTCF) for RNF20 in the compact myocardium is given.Heatmaps depicting the genes containing regions with decreased H2Bub1 occupancy, genes with decreased gene expression, and genes that have both decreased H2Bub1 occupancy and decreased gene expression when comparing iPSC and mesoderm (M), M and cardiac mesoderm (CMes), CMes and cardiac progenitor (CP), and CP and cardiomyocytes (CM).If there are less than 25 genes in a category, they are all listed.If there are more than 25 genes in a category, example genes are given below each heatmap.See supplemental data 1 and 2 for the complete list.The lost ChIP peaks were determined by a differential binding analysis for H2Bub1-ChIP-seq.
There are 860 genes downregulated in M, 1364 genes downregulated in CMes, 1446 genes downregulated in CP, and 3141 genes downregulated in CM.Heatmaps depicting the genes containing regions with increased H2Bub1 occupancy, genes with increased gene expression, and genes that have both increased H2Bub1 occupancy and increased gene expression when comparing iPSC and mesoderm (M), M and cardiac mesoderm (CMes), CMes and cardiac progenitor (CP), and CP and cardiomyocytes (CM).If there are less than 25 genes in a category, they are all listed.If there are more than 25 genes in a category, example genes are given below each heatmap.See supplemental data 1 and 2 for the complete list.The gained ChIP peaks were determined by a differential binding analysis for H2Bub1-ChIPseq.There are 648 genes upregulated in M, 1844 genes upregulated in CMes, 1704 genes upregulated in CP, and 3490 genes upregulated in CM.Heatmaps depicting the genes containing regions with constant H2Bub1 occupancy, genes with constant gene expression, and genes that have both constant H2Bub1 occupancy and constant gene expression when comparing iPSC and mesoderm (M), M and cardiac mesoderm (CMes), CMes and cardiac progenitor (CP), and CP and cardiomyocytes (CM).If there are less than 25 genes in a category, they are all listed.If there are more than 25 genes in a category, example genes are given below each heatmap.See supplemental data 1 and 2 for the complete list.The constant ChIP peaks were determined by a differential binding analysis for H2Bub1-ChIP-seq.

Fig. S6 .
Fig. S6.H2Bub1 is located in the heterochromatic regions.(A) Graphs of the four clusters identified in Figure 3A across five stages of CM differentiation (iPSC, Mesoderm (M), Cardiac Mesoderm (CMes), Cardiac Progenitor (CP), and Cardiomyocyte (CM)).Cluster 1 indicates high H2Bub1 levels, cluster 2 indicates low H2Bub1 levels, cluster 3 indicates moderate H2Bub1 levels, and cluster 4 indicates no H2Bub1.(B) Western blot for H2Bub1 in five stages of CM differentiation (iPSC, M, CMes, CP, and CM).The loading control for H2Bub1 is H2B.Numbers indicate average quantification (imageJ) of H2Bub1 normalized to H2B over two replicates.(C) Percent of H2Bub1 peaks overlapping with the B compartment in each stage of CM differentiation (iPSC, M, CP, and CM) comparing observed (gray) and expected (black) values.Data are shown as mean ± SEM (n = 3).Pvalues comparing observed and expected values were manually calculated by comparing the actual values to the expected values, * p < 0.05, ** P < 0.01 (see Materials and Methods).P-values comparing observed values between stages were calculated by computing a z score, * p < 0.05, ** P < 0.01.(D) Percent of H2Bub1 peaks on the same gene as an activating chromatin mark (H3K4me3 shown in solid bars) and two repressive chromatin marks (H3K27me3 shown in bars with white dashes and H3K9me3 with black dashes) in iPSCs and CMs.Data are the ratio of observed to expected values.Data are shown as mean ± SEM (n = 3).P-values comparing observed values to expected values were calculated by Chi Square tests and adjusted for multiple comparisons,* p < 0.05, ** P < 0.01.P-values comparing observed values between stages were calculated using an unpaired 2-tailed, heteroscedastic t-test, * p < 0.05, ** P < 0.01, N.S. is non-significant.(E) H2Bub1 occupancy over H3K27me3 peaks in iPSCs (blue) and CMs (red).(F-G) Peak annotations (F) and Compartment information (G) for the H2Bub1 peaks on the same gene as H3K4me3.(H) Peak annotations for H2Bub1 peaks in the A compartment and in the B compartment.(I) Expression levels (from RNA-seq) for genes with H2Bub1 peaks in the A compartment and in the B compartment.2-tailed, Mann-Whitney test, ** p < 0.01.
Fig. S11.Characteristics of the RNF20 +/-mutants.(A) Diagram of the RNF20 protein with its domains (Ring-Finger domain is shown in blue) and where the created RNF20 mutations are located (with a star).Lower four panels are sequencing traces of the wild-type and two independent RNF20 +/-iPSC lines demonstrating CRISPR-generated mutations.(B) Western blot for RNF20 and H2Bub1 in two stages of CM differentiation (iPSC and Mesoderm (M)).The loading control for RNF20 is GAPDH and for H2Bub1 is H2B.Numbers indicate average quantification (imageJ) of H2Bub1 normalized to H2B over three replicates.Next to the western blot is a H2Bub1-deposition complex schematic illustrating the RNF20 +/- iPSC mutant, which leads to increased total H2Bub1 levels.

Fig. S15 .E
Fig. S15.Amount of full-length transcript controls.(A-B) Histograms showing the average length in Kb of each of 10,000 random quantity matched gene sets to the calcium and sarcomere gene sets.The vertical dashed red lines show the average length in Kb of the calcium genes (A) and the sarcomere genes (B).(C-D) Log2 of fold change in transcript abundance between wild-type and UBE2B -/-mutants isshown at each position.The genes shown in (C) are the sarcomere genes (n = 70 genes) that are differentially expressed between wild-type and UBE2B -/-mutants (n = 3, for each of 2 cell lines).The genes shown in (D) are "randomly" selected quantity and size-matched genes to the sarcomere gene set (n = 3, for each of 2 cell lines).30 "random" plots were created from the sarcomere gene set: 10 from genes that are upregulated between wild-type and both UBE2B -/-cell lines, 10 that are non-regulated between wild-type and both UBE2B -/-cell lines, and 10 that are down-regulated between wild-type and both UBE2B -/-cell lines.(E) A histogram showing the average expression in TPM of each of the "random" quantity and size-matched gene sets described in (D).The vertical dashed red line shows the average expression in TPM of the sarcomere genes.(F) A histogram showing the log average expression in TPM of all of the 3 rd Quartile genes (greater than 33.940Kb and less than 93.323Kb).The vertical dashed red lines show the average expression in TPM of the calcium genes (left) and sarcomere genes (right).(G)Expression of an example calcium gene, RYR2, shown in Figure5D.5' on the left of the diagram (WT (blue), UBE2B -/-1 (green), and UBE2B -/-2 (red)).Zoom in of exon 1 (red box region) is shown on the right.(H)Example log2 of fold change gene traces for one calcium gene (RYR2) and non-calcium short genes (TMEM117, CCDC178 and ERC1).The short genes do not have accumulation of H2Bub1 near the center of the gene and do not have decreased full-length transcripts.5' on the left of the diagram. Rnf20