XB-ART-56064Proc Natl Acad Sci U S A January 1, 2019; 116 (28): 14049-14054.
Histone H2B monoubiquitination regulates heart development via epigenetic control of cilia motility.
Genomic analyses of patients with congenital heart disease (CHD) have identified significant contribution from mutations affecting cilia genes and chromatin remodeling genes; however, the mechanism(s) connecting chromatin remodeling to CHD is unknown. Histone H2B monoubiquitination (H2Bub1) is catalyzed by the RNF20 complex consisting of RNF20, RNF40, and UBE2B. Here, we show significant enrichment of loss-of-function mutations affecting H2Bub1 in CHD patients (enrichment 6.01, P = 1.67 × 10-03), some of whom had abnormal laterality associated with ciliary dysfunction. In Xenopus, knockdown of rnf20 and rnf40 results in abnormal heart looping, defective development of left-right (LR) asymmetry, and impaired cilia motility. Rnf20, Rnf40, and Ube2b affect LR patterning and cilia synergistically. Examination of global H2Bub1 level in Xenopus embryos shows that H2Bub1 is developmentally regulated and requires Rnf20. To examine gene-specific H2Bub1, we performed ChIP-seq of mouse ciliated and nonciliated tissues and showed tissue-specific H2Bub1 marks significantly enriched at cilia genes including the transcription factor Rfx3 Rnf20 knockdown results in decreased levels of rfx3 mRNA in Xenopus, and exogenous rfx3 can rescue the Rnf20 depletion phenotype. These data suggest that Rnf20 functions at the Rfx3 locus regulating cilia motility and cardiac situs and identify H2Bub1 as an upstream transcriptional regulator controlling tissue-specific expression of cilia genes. Our findings mechanistically link the two functional gene ontologies that have been implicated in human CHD: chromatin remodeling and cilia function.
PubMed ID: 31235600
PMC ID: PMC6628794
Article link: Proc Natl Acad Sci U S A
Genes referenced: arl13b ccdc88a chrd.1 cplane1 ctrl dand5 dnah11 dnah7 dnah9 foxj1 h2bc21 hk1 kidins220 kif3a lhx6 nodal npat pitx2 psmd6 rfx3 rnf20 rnf40 rpl8 sgce spata6l ube2b
Morpholinos: rnf20 MO1 rnf40 MO1 ube2b MO1
Article Images: [+] show captions
|Figure 1. Mutations affecting the core RNF20 complex are identified in patients with congenital heart disease. (A) Summary of patient variants, inheritance, and phenotypes affecting the core components of the RNF20 complex (RNF20, RNF40, and UBE2B). CAVC, complete atrioventricular canal; HLHS, hypoplastic left heart syndrome; L-TGA, levo-transposition of the great arteries; PA, pulmonary atresia; RAA, right aortic arch; RAI, right atrial isomerism; TAPVR, total anomalous pulmonary venous return; TOF, tetralogy of Fallot. (B) Diagrams of RNF20, RNF40, and UBE2B with patient variants (red arrows) indicated. (C) TEM of tracheal cilia from a patient with the RNF20 variant (Top) and control (Bottom). Blue arrows point to outer dynein arms; the red arrow points to IDAs. High-magnification electron micrographs were obtained from greater than 10 ciliated cells.|
|Figure 4. H2Bub1 marks are enriched at cilia genes in multiciliated tissue. (A) A heatmap depicting the top 20 significant regions with increased H2Bub1 occupancy in multiciliated tissue compared with nonciliated tissue. The bolded genes are cilia-related genes depicted in C. (B) The top 20 significant gene ontology terms from the regions that have increased H2Bub1 occupancy in multiciliated tissue compared with nonciliated tissues. The cilia-related gene ontology terms are shown in blue. (C) The H2Bub1 binding profile, depicted using fold enrichment against random distribution values ranging from 0 to 10, across the Rfx3, Kif3a, and Dnah7 genes in multiciliated tissue (oviduct; red) and nonciliated tissue (liver; blue). The arrows indicate the TSS and the direction of transcription.|
|Figure S1: (A) Schematic of H2B K120 ubiquitination (H2Bub1) complex. (B) Detailed description of patient mutations affecting RNF20-core complex. Predicted deleteriousness by CADD score and MetaSVM. Mutation tolerance is shown as pLI score and mis-z score. Rank Heart Expression is the percentile of expression by RNAseq of e14.5 mouse heart. Mutation frequency is shown in both genomAD exome database (125,748 exomes), and genomAD genome (15,708 genomes) database. Evolutionary conservation is shown across 46 species.|
|Figure S2: List of 45 genes identified in the RNF20 interactome generated via STRING. Genes with de-novo mutations in patients with CHD are highlighted in blue.|
|Figure S3: Schematic of one cell and one cell in two-cell embryo Xenopus injections and predicted effect on asymmetrically expressed mRNA markers and heart looping. Early embryos are visualized from the dorsal aspect. The LRO is visualized from the ventral aspect. Pitx2cstage embryos are shown as a composite of the left and right sides. Tadpoles are shown from the ventral aspect. Normal cilia are shown in blue, cilia affected by MO injection are shown in red. Domains of cerl2 (coco, dand5) expression are shown in red, and nodal expression is shown in blue. pitx2c expression is shown in green.|
|Figure S5: (A) In situ hybridization images with rnf20, rnf40 and ube2b specific probes in X.tropicalis during important stages of development. 8-cell, gastrula (stage 10), LRO formation (GRP, stage 16), neurula (stage 18), and early tadpole (stage 24). LRO is outlined at stage 16, and arrows indicate the developing kidney in the tadpole stage images. Representative examples of 3 independent experiments. (B) RNF20 is expressed in the nuclei of LRO of Xenopus embryos at stages 16-17 as indicated by staining with anti-RNF20 antibody. LRO structure was identified by the presence of monociliated cells labeled with anti-acetylated tubulin, nuclei were visualized by Hoechst staining. Anterior (A)- Posterior (P) and Left (L)- Right (R) axes are indicated. The image is a maximum intensity projection of Z-stack. (C) Fluorescence images of Xenopus epidermis labeled with anti-Rnf20 Ab (green), Hoechst (blue) and anti-acetylated tubulin (magenta). Top panel shows a deeper section of the embryos whereas the bottom panel shows the epidermal surface of the tadpole.|
|Figure S6: (A) Rnf20 expression in the mouse LRO. The LRO structure was identified by presence of monociliated cells labeled with anti-Arl13b, nuclei were visualized by Hoechst staining. Anterior (A)-Posterior (P) and Left (L)-Right (R) axes are indicated. The image is the maximum intensity projection of a Z stack encompassing the entire LRO. (B) RNF20 expression in mouse multiciliated tissues visualized by immunostaining with anti-RNF20 antibody (Green). Multiciliated cells were identified by ciliary labeling with anti-acetylated tubulin (Red), and nuclei by Hoechst staining. (C) Western blot of 20µg/lane adult mouse tissue probed with anti-Rnf20 antibody showing expected 110kD band. Loading was normalized to total protein determined by BCA assay and confirmed by SimplyBlue SafeStain (Coomassie G-250).|
|Figure S7: Cilia appear structurally normal after rnf20 depletion (A) Top left panel: LROs from embryos treated with either control MO or rnf20 MO stained with anti-acetylated-tubulin (AcTub, green) to detect primary cilia, Rhodamine phalloidin (red) marks the cell borders. White outline highlights the LRO. Bottom left panel: High magnification of LROs from the top panel. Top right: Quantification of motile and immotile cilia in LROs of embryos injected with either control MO or rnf20 MO. n=total number of embryos. Bottom right: Average length of LRO cilia. (B) Cilia morphology and distribution at the epidermal surface in control MO (left) and rnf20 MO (right) after injection at the one-cell stage. Top panels show distribution of MCCs on the surface of the embryo, bottom panels at higher magnification show morphology of individual MCCs. Cilia were stained using anti-acetylated tubulin antibody|
|Figure S9: A heatmap of H2bub1 Chip-seq H2B ubiquitination surrounding the transcription start site (TSS) (-1Kb, + 10Kb) in ciliated tissue (oviduct) and non-ciliated tissue (liver). Genes are grouped by H2Bub1 levels into four clusters: high, moderate, low, and none. The average profile, depicted using fold enrichment against random distribution values, across this region for each cluster in each tissue is shown above the heatmap. A zoomed in heatmap of the "low" cluster highlights the differences between the ciliated and non-ciliated tissues.|
|Figure S11: (A) Experimental design of RNA-seq experiment in Xenopus. (B) PCA plot of RNAseq experiment at hours 12, 13, 14, 15, 16 for two control replicates (blue circles) and three rnf20MO replicates (red circles).|
|Figure S12: (A) Upper panel: Expression of rfx3, dnah7, and dnah9 by in situ hybridization in control, rnf20 MO injected, and rnf20 MO and rfx3 mRNA co-injected Xenopus embryos. LROs are shown. Central panel: quantification of the in situ experiment. Blue-% of embryos with normal expression, green-% of embryos with decreased expression. Lower panel(A), (B), and (C): RNA seq data comparing expression profiles in scrambled MO injected control embryos (blue) and rnf20 MO injected embryos (red) at 0-18 hourspost fertilization. The area highlighted in blue corresponds to the time of LRO formation and function. (A) Lower panel: Expression profiles of genes analyzed by in situ hybridization. (B) Expression profiles of inner dynein arm heavy chains (IDAs) in rnf20 MO and controls. (C) Expression profiles of outer dynein arm heavy chains (ODAs) in rnf20 MO and controls.|