HoxBlinc RNA Recruits Set1/MLL Complexes to Activate Hox Gene Expression Patterns and Mesoderm Lineage Development

Abstract

Trithorax proteins and long-intergenic noncoding RNAs are critical regulators of embryonic stem cell pluripotency; however, how they cooperatively regulate germ layer mesoderm specification remains elusive. We report here that HoxBlinc RNA first specifies Flk1(+) mesoderm and then promotes hematopoietic differentiation through regulation of hoxb pathways. HoxBlinc binds to the hoxb genes, recruits Setd1a/MLL1 complexes, and mediates long-range chromatin interactions to activate transcription of the hoxb genes. Depletion of HoxBlinc by shRNA-mediated knockdown or CRISPR-Cas9-mediated genetic deletion inhibits expression of hoxb genes and other factors regulating cardiac/hematopoietic differentiation. Reduced hoxb expression is accompanied by decreased recruitment of Set1/MLL1 and H3K4me3 modification, as well as by reduced chromatin loop formation. Re-expression of hoxb2-b4 genes in HoxBlinc-depleted embryoid bodies rescues Flk1(+) precursors that undergo hematopoietic differentiation. Thus, HoxBlinc plays an important role in controlling hoxb transcription networks that mediate specification of mesoderm-derived Flk1(+) precursors and differentiation of Flk1(+) cells into hematopoietic lineages.

Keywords: HoxBlinc lincRNA; SETD1A and MLL1 HMTs; chromatin looping; hoxb gene activation; mesoderm specification.

Figure 1. HoxBlinc specifies Flk1⁺ mesodermal cells

(A) Expression of HoxBlinc negatively correlates with Oct4 expression during EB hematopoietic differentiation. (B) Expression of hoxb genes was gradually induced upon EB differentiation. (C) Northern blot analysis of HoxBlinc RNA was performed in ESCs and day 6 EBs with or without Dox induced HoxBlinc KD. (D) FACS analysis of Flk1⁺ cells in induced HoxBlinc KD and control EBs. KD was induced at the day 2 epiblast stage and FACS analysis was carried out at day 4 of EB differentiation. (E) BL-CFC potential of EBs with or without Dox induced HoxBlinc KD at day 2 epiblast stage. (F) FACS analysis of CD31⁺/CD41⁺ endothelial/hematopoietic cells differentiated from the induced HoxBlinc KD at day 2 epiblast stage and control blast colonies. (G) RNA-seq analysis of induced HoxBlinc KD (+Dox) and control EBs (−Dox) at epiblast stage (day 2). RNA was isolated from day 6 differentiated EBs. HoxBlinc RNA regulated genes were annotated by Gene Ontology (GO) analysis. (H) Overlap between HoxBlinc RNA activated genes and total ESC bivalent genes.

Figure 2. HoxBlinc controls hematopoietic differentiation by regulating hoxb genes and hematopoietic transcription programs

(A) Scatter plots showing that expression levels of TFs and those encoding genes required for hematopoiesis (Left) or the NOTCH pathway (Right). Activated genes are shown in red, repressed genes in blue (only genes are shown that alter expression by more than 3-fold upon HoxBlinc KD). (B) qRT-PCR analysis of hoxb gene expression comparing induced HoxBlinc KD (+Dox) and control EBs (−Dox) at day 6. (C) qRT-PCR analysis of expression of genes encoding key hematopoietic TFs and markers in induced HoxBlinc KD (day 3 Flk1⁺ stage) and control EBs collected at day 6. (D) Schematic representation of hematopoietic differentiation from the purified FLK1⁺ cells following hemangioblast development. (E) FACS analysis of CD41 expression upon hematopoietic differentiation of sorted FLK1⁺ cells which were then induced for HoxBlinc KD and cultured in blast culture media. (F) CFC analysis of definitive hematopoietic colonies (Ery-D, GEMM, GM, and Mac) in induced HoxBlinc KD and control EB derived cells. (G) qRT-PCR analysis of hoxb gene expression upon hematopoietic differentiation of sorted FLK1⁺ cells that were then induced for HoxBlinc KD and cultured in blast culture media. (H) Analysis of the expression of hematopoietic specific TFs and markers during hematopoietic differentiation of sorted FLK1⁺ cells with and without HoxBlinc KD cultured in blast culture media. Data is presented as Mean ± SD from 3–4 independent experiments; *P<0.05; **P<0.01 by Student’s t-test.

Figure 3. KD of HoxBlinc inhibits hoxb gene expression and cardiogenic mesoderm formation

(A) Scatter plot showing expression levels of genes encoding TFs and markers important for heart development increased (red) or decreased (blue) by more than 3-fold upon HoxBlinc KD. (B) qRT-PCR analysis of expression of genes encoding TFs and markers important for heart development in induced HoxBlinc KD (at the epiblast stage) and control EBs (−Dox) collected at day 4 and day 6. Data is presented as Mean ± SD from 3–4 independent experiments; *P<0.05; **P<0.01. (C) FACS analysis of Flk1⁺/PDGFRα⁺ cells in induced HoxBlinc KD (at the epiblast stage) and control EBs collected at day 4. (D) qRT-PCR analysis of the expression of cTnT and actc1 in HoxBlinc KD and control EBs collected at days 0, 4, 7, and 10. (E) Green fluorescence visualization of the expression of the cardiomyocyte marker, MHC-GFP, in induced HoxBlinc KD (at the epiblast stage) and control EBs (−Dox) at day 12.

Figure 4. Re-expression of anterior hoxb2-b4 genes rescues Flk1⁺ mesoderm and HS/PCs in the HoxBlinc KD EBs

(A) qRT-PCR analysis showing expression levels of hoxb2-b4 genes in control or induced HoxBlinc KD EBs rescued with the vector or constructs expressing hoxb2-b4 genes. (B) FACS analysis of Flk1⁺ cells (Top) and CD41⁺/c-Kit⁺ HS/PCs in control and induced HoxBlinc KD EBs rescued with vector or constructs expressing the HoxB2-B4 genes. (C) Percentage of Flk1⁺ cells (Top) and CD41⁺/c-Kit⁺ HS/PCs (Bottom) in control and rescued EBs. (D) BL-CFC potential of EBs in control or induced HoxBlinc KD EBs rescued with the vector or constructs expressing hoxb2-b4. (E) Secondary hematopoietic CFC assays of definitive hematopoietic colonies (Ery-D, GEMM, and GM) in induced HoxBlinc KD and rescued EBs derived blast colony cells. (F–G) qRT-PCR analysis showing expression levels of TFs and markers required for hematopoiesis (F) or genes encoding NOTCH pathway in control or rescued EBs (G). Data is presented as Mean ± SD from 3–4 independent experiments; *P<0.05; **P<0.01 by Student’s t-test.

Figure 5. HoxBlinc interacts with the Set1/MLL1 complexes to activate target genes

(A) ChIP analysis of H3K4me3 enrichment at promoters of hoxb genes in induced HoxBlinc KD and control EBs. (B) Biotinylated HoxBlinc RNA and a 3’ domain of HoxBlinc RNA specifically interact with Setd1a in 6d EB NEs. (C) RIP showing that both Setd1a and MLL1 are associated with HoxBlinc before FLK1 specification (d3 EBs) and after FLK1 specification (d6 EBs). (D)HoxBlinc directly interacts with the SET domain of Setd1a and MLL1 proteins shown by incubating GST-Setd1a^SET or GST-MLL1^SET with in vitro transcribed RNAs. (E)HoxBlinc activates luciferase reporter gene transcription by recruiting TrxG HMT complexes. Top, diagram of the BoxB tethering reporter system in which GAL4 fused λN recruits BoxB-HoxBlinc fusion RNA associated TrxG complexes to a UAS driven luciferase reporter construct. Bottom, luciferase activity was analyzed 48 hrs post transfection with the indicated constructs. Data is presented as Mean ± SD from 3–4 independent experiments; *P<0.05; **P<0.01 by Student’s t-test.

Figure 6. HoxBlinc binds to anterior hoxb genes and regulates chromatin structure alterations at the anterior hoxb locus

(A) qRT-PCR analysis of RNA retrieved by the complementary HoxBlinc tiling probes and LacZ probes in day 3 and day 6 EBs. (B) ChIRP analysis of the HoxBlinc RNA enrichment at the HoxB locus in day 3 and day 6 EBs. (C–D) ChIP analysis showing binding of Setd1a (C) and MLL1 (D) at the promoters of hoxb genes in HoxBlinc KD and control d3 EBs prior to the Flk1 specification. (E) Schematic diagram representing the cluster organization of the hoxb genes. (F) 3C analysis showing interactions between the HoxBlinc locus and the anterior hoxb genes in ESCs and differentiated EBs at day 1 and 6. (G) 3C analysis showing interactions between the HoxBlinc locus and the anterior hoxb genes in HoxBlinc KD and control day 6 EBs. (H) Model depicting how HoxBlinc organizes the anterior hoxb genes into a CTCF mediated inducible active chromatin domain that facilitates long-range interactions between the HoxBlinc locus and the anterior hoxb1-b3 genes. Data is presented as Mean ± SD from 3 independent experiments; *P<0.05; **P<0.01 by Student’s t-test.

Figure 7. Deletion of the HoxBlinc locus by CRISPR-Cas9 genome editing inhibits Flk1⁺ mesoderm differentiation

(A) PCR based genotyping analysis of the HoxBlinc genomic KO. (B) qRT-PCR analysis of HoxBlinc and hoxb gene expression upon CRISPR-Cas9 mediated deletion at day 4 EBs. (C) Immunohistochemistry staining of ecodermal marker nestin and mesodermal marker α-SMA comparing WT and HoxBlinc KO teratomas. (D) qRT-PCR gene expression analysis of three germ layers in teratomas derived from the control, HoxBlinc KD, and CRISPR-Cas9 mediated homozygous HoxBlinc KO. (E) FACS analysis of Flk1⁺ cells (Left) and CD41⁺ HS/PCs (Right) in control and HoxBlinc KO EBs. (F–H) qRT-PCR analysis of the expression levels of TFs (F), markers (G), and Notch signaling pathway (H) required for hematopoietic/cardiac differentiation upon CRISPR-Cas9 mediated deletion. Data is presented as Mean ± SD from 3 independent experiments; *P<0.05; **P<0.01 by Student’s t-test.