Date of Award
Doctor of Philosophy (PhD)
Currently, the most common birth defect is congenital heart disease (CHD). Whole-exome sequencing of CHD patients identified variants affecting chromatin modifier genes, which contribute to 2.3% of CHD. De-novo mutations affecting the core complex (RNF20, RNF40 and UBE2B) required for the monoubiquitination of H2B on K120 (H2Bub1) are enriched in CHD patients. This complex has many cellular roles, such as stem cell differentiation, but no described role in cardiac morphogenesis. This body of work describes the role of H2Bub1 in heart development in vivo in mice and in vitro in human induced pluripotent stem cell (iPSC)-derived cardiomyocytes. The patient variant in RNF20 was an early stop codon in a single allele, suggesting haploinsufficiency for RNF20, so we first analyzed Rnf20+/- mice. Unlike the patients, these mice have no discernible cardiac abnormalities. Additionally, the Rnf20-/- mice are pre-implantation lethal. Therefore, we created cardiac specific deletions of Rnf20, using both Nkx2.5-Cre and Isl1-Cre to analyze its role in the first and second heart fields respectively. The Rnf20fl/-::Nkx2.5-Cre+ mice are e12.5 lethal, and have incomplete septation and decreased compact myocardium, while the Rnf20fl/-::Isl1-Cre+ mice survive until adulthood and appear morphologically normal. We further investigated the Rnf20fl/-::Nkx2.5-Cre+ mice and determined that they have disorganized cardiac sarcomeres lacking an H zone and decreased expression of calcium signaling genes. We conclude that Rnf20 is necessary for embryonic viability and normal cardiac structure formation. Given the myocardial phenotype, we investigated H2Bub1 during cardiomyocyte development by in vitro differentiation of human iPSCs into cardiomyocytes. ChIP-seq for H2Bub1 in iPSC-derived cardiomyocytes demonstrates that H2Bub1 is dynamically regulated during differentiation: there are abundant genes with H2Bub1 marks in iPSCs and mesoderm, but only a few genes maintain H2Bub1 during the transition from cardiac mesoderm to cardiac progenitors. The set of selectively maintained genes is strongly enriched for sarcomeric calcium genes, and 10/11 of these calcium genes are associated with cardiomyopathy. There are conflicting published data about whether H2Bub1 is an activating or repressive mark. Our data suggest that H2Bub1 is selectively maintained on tissue-specific genes to promote their expression. These data support the mouse data by also showing a link between H2Bub1, calcium signaling, and cardiomyopathy. We then tested the effect of downregulation of RNF20-complex components on cardiomyocyte differentiation. Haploinsufficiency for RNF20 arrests differentiation into cardiac mesoderm. Notably, decreased RNF20 leads to a paradoxical increase in H2Bub1 at chromatin modifier genes. The lack of concordance of the decrease in RNF20 and the resulting increase in total H2Bub1 levels suggests that H2Bub1 deposition efficiency may reflect the ratios of the RNF20-complex components. Depletion of another RNF20-complex component, UBE2B, results in either beating cardiomyocytes (which seem indistinguishable from wild-type) or failure to differentiate into cardiac mesoderm. RNA-seq shows decreased expression of sarcomere and calcium genes. These data demonstrate that H2Bub1 levels need to be tightly regulated for normal heart development. We then determined that H2Bub1 preferentially marks tissue-specific long genes. We first compared H2Bub1-ChIP-seq data between the multiciliated oviduct and the nonmulticiliated liver, and identified enrichment in cilia genes among the genes that have increased occupancy in oviducts. Interestingly, cilia genes have long transcripts. Published data indicate a link between H2Bub1 and transcriptional elongation, but there is disagreement as to whether H2Bub1 has a positive or negative effect. Thus, we hypothesized that that preferential marking of tissue-specific long genes by H2Bub1 is related to transcriptional elongation. We next compared the shape of the H2Bub1 signal between wild-type and UBE2B-/- cardiomyocytes in long genes and short genes. While short genes give the predicted profile, long genes have a vastly different profile. Interestingly, we identified tissue-specific UBE2B-dependent H2Bub1 accumulation near the center of long genes, which correlates with efficient transcriptional elongation. Finally, we compared the shape of the H2Bub1 signal between mouse embryonic fibroblasts and mouse embryonic stem cells. These data also show H2Bub1 accumulation near the center of tissue-specific genes. Therefore, our data suggest that H2Bub1 positively regulates transcriptional elongation on long tissue-specific genes. In summary, our data show that the RNF20-complex is required for cardiac development. Mice with cardiac-specific deletion of Rnf20 have structural heart defects; human iPSCs with RNF20 and UBE2B mutations cannot efficiently differentiate into cardiomyocytes. In both mice and iPSC-derived cardiomyocytes, sarcomere and calcium signaling gene expression is dependent on normal H2Bub1 levels. We also identify tissue-specific UBE2B-dependent H2Bub1 accumulation in cardiomyocytes and MEFs that corresponds to transcriptional efficiency. Thus, normal H2Bub1 distribution is required for cardiac development, and H2Bub1 accumulation may act as a general mechanism for tissue-specific regulation of transcriptional elongation efficiency.
Barish, Syndi Marissa, "Tissue-Specific H2Bub1 is Required for Cardiomyocyte Development" (2021). Yale Graduate School of Arts and Sciences Dissertations. 299.