Lin H et al. (2017), KDM3A-mediated demethylation of histone H3 lysi...

XB-ART-54148

Development 2017 Oct 15;14420:3674-3685. doi: 10.1242/dev.144113.

Show Gene links Show Anatomy links

KDM3A-mediated demethylation of histone H3 lysine 9 facilitates the chromatin binding of Neurog2 during neurogenesis.

Lin H , Zhu X , Chen G , Song L , Gao L , Khand AA , Chen Y , Lin G , Tao Q .

???displayArticle.abstract???
Neurog2 is a crucial regulator of neuronal fate specification and differentiation in vivo and in vitro However, it remains unclear how Neurog2 transactivates neuronal genes that are silenced by repressive chromatin. Here, we provide evidence that the histone H3 lysine 9 demethylase KDM3A facilitates the Xenopus Neurog2 (formerly known as Xngnr1) chromatin accessibility during neuronal transcription. Loss-of-function analyses reveal that KDM3A is not required for the transition of naive ectoderm to neural progenitor cells but is essential for primary neuron formation. ChIP series followed by qPCR analyses reveal that Neurog2 promotes the removal of the repressive H3K9me2 marks and addition of active histone marks, including H3K27ac and H3K4me3, at the NeuroD1 and Tubb2b promoters; this activity depends on the presence of KDM3A because Neurog2, via its C-terminal domain, interacts with KDM3A. Interestingly, KDM3A is dispensable for the neuronal transcription initiated by Ascl1, a proneural factor related to neurogenin in the bHLH family. In summary, our findings uncover a crucial role for histone H3K9 demethylation during Neurog2-mediated neuronal transcription and help in the understanding of the different activities of Neurog2 and Ascl1 in initiating neuronal development.

???displayArticle.pubmedLink??? 29042477
???displayArticle.link??? Development

Species referenced: Xenopus laevis
Genes referenced: ascl1 ebf2 eef1a1 egr2 en2 fgf8 h3-3a hoxb9 isl1 kdm3a kdm4a krt12.5 mapk1 mnx1 msx1 myc myt1 neurod1 neurod4 neurog1 neurog2 nog pax6 rax smo snai1 sox1 sox3 sox9 tbxt tlx1 tubb2b ventx2 zic1
GO keywords: histone demethylation [+]
???displayArticle.antibodies??? Acetylated H3f3 Ab37 H3f3 Ab30 H3f3 Ab35 H3f3 Ab36 Mapk1 Ab1 Mapk1 Ab10 Methyl-h3f3a Ab38 h3f3 Ab37 h3f3 Ab38 h3f3a Ab32 kdm3a Ab1 kdm4a Ab1 neurog2 Ab1
???displayArticle.morpholinos??? ascl1 MO7 kdm3a MO1 kdm3a MO2 neurog2 MO2

???attribute.lit??? ???displayArticles.show???

	Fig. 1. Expression of KDM3A in early Xenopus embryos and the effects of its depletion on neurogenic potential. (A-E) The spatial expression pattern of kdm3a mRNA revealed through whole-mount in situ hybridization. (F) Western blot analysis of KDM3A in embryos injected with or without KDM3A MO (3A MO). 3A MO (80 ng) was injected into both cells at the twocell stage. Embryos were then collected at the indicated stages (stage 11-18) and subjected to western blot analysis using an antibody directed against mouse KDM3A. Histone H3 served as a loading control. (G) Western blot analysis showing that injecting the MO-resistant kdm3a mRNA rescued the expression of KDM3A protein in 3A MO-injected embryos. Control MO (cMO) and 3AMO (60 ng) were injected into both cells at the two-cell stage. 3A MO-resistant kdm3a mRNA (500 pg) was injected at the one-cell stage. Embryos were then subjected to western blotting at stage 11. The numbers in G show the relative pixel intensities of bands when the intensity of cMO-injected lane was set to 1.00. (H-J) RTqPCR analyses for gene expression in animal cap explants (AC) with indicated treatments. cMO or KDM3A MO (80 ng) was injected at the one-cell stage. mRNA (100 pg) encoding FGF8a (H), Ascl1 (I) or Neurog2 (J) was injected into animal pole of control embryos or MO-injected embryos at the two-cell stage. Two biological repeats and three technical repeats for each biological repeat were performed. The results from these six repeats were combined together and are represented as mean±s.e.m. The expression level of each examined gene in the group of animal cap explants (AC) that received only mRNA injection was set to 1 for normalization. Two-tailed Student t-test was performed to assess significance. ns, not significant. P<0.05; **P<0.005.
	Fig. 2. KDM3A is required for the primary neurogenesis in Xenopus. (A) A schematic showing the experimental design for B and C. (B,B′) Embryos at stage 24/25 stained with red-Gal and in situ hybridized with tubb2b. cMO or 3A MO (40 ng) was injected into one cell at the two-cell stage. The 3A MO-resistant kdm3a mRNA (300 pg) together with 200 pg β-gal mRNAwas subsequently injected into the 3A MO-injected cell at the two-cell stage. The expression levels of tubb2b were classified into three categories: ‘normal’, as seen in cMO-injected embryos; ‘partial’, as seen in some of KDM3A depleted embryos; and ‘complete’ loss, as seen in the remainder of the KDM3A-depleted embryos. Injection of the MO-resistant kdm3a mRNA partially rescues the expression of tubb2b in KDM3A morphants. The numbers on the top of histograms in B′ are sums from two independent experiments. (C,C′) Embryos at stage 16/17 stained with red-Gal and in situ hybridized with tubb2b. cMO and 3A MO (40 ng) were injected into one cell at the two-cell stage. mRNA (100 pg) encoding neurog2 or ascl1 together with 200 pg β-gal mRNA was subsequently injected into the MO-receiving cell at the two-cell stage. Phenotypes were classified into four categories based on whether ectopic neurons were induced and/or whether tubb2b was ‘partially’ or ‘completely’ lost in the MO-injected side. The numbers on the top of histograms in C′ are the total number of samples from two independent experiments.
	Fig. 3. Neurog2 fails to transactivate its neuronal targets in KDM3A morphant embryos. (A) Embryos at stage 17/18 in situ hybridized with Xenopus neurog2. cMO or 3A MO (80 ng) was injected into both cells at the two-cell stage. (B) qPCR results showing that KDM3A depletion differentially affects the expression of neural progenitor genes (sox2 and neurog2) and neuronal genes (neurod1 and tubb2b). ns, no significance. *P<0.001. cMO or 3A MO (80 ng) was injected into both cells at the two-cell stage. Embryos were processed for RT-qPCR at stage 16. (C,C′) Embryos at stage 18 stained with red-Gal and in situ hybridized with neurod1. cMO or 3A MO (40 ng) was injected into one cell at the two-cell stage. mRNA encoding Neurog2 (100 pg) together with 200 pg β-gal mRNA was subsequently injected into the MO-receiving cell at the two-cell stage. (D,E) ChIP-qPCR detection of H3K27ac (D) and H3K4me3 (E) at the neurod1 and tubb2b promoters. cMO or 3A MO (80 ng) was injected into one-cell stage embryos. mRNA encoding Neurog2 (100 pg) was injected into both cells of half of the MO-injected embryos at the two-cell stage. ef1α expression was examined as a negative control. Locations of the qPCR primers are schematically shown in the x axes. The ChIP-qPCR experiments shown in D,E were combined from two independent repeats. Three technical replicates were made in each independent repeat. P<0.01; ***P<0.005; ns, not significant (according to a two-tailed Student’s t-test).
	Fig. 4. KDM3A is associated with Neurog2. (A) Schematics of the domain organization of Neurog2 and KDM3A proteins. (B-D) CoIP followed by western blot analyses for the interaction between Neurog2 and KDM3A. mRNAs encoding the full-length epitope-tagged Neurog2 and KDM3A (B), deletion mutants of Neurog2 and the full-length KDM3A (C), or the full-length Neurog2 and deletion mutants of KDM3A (D) were microinjected into two-cell stage embryos, which were then collected at stage 14 and subjected to CoIP followed by western blot analyses with indicated antibodies. Bands around the 55 kDa and 40 kDa marks in the top blot are non-specific. (E) qPCR analysis of gene expression in cMO-, 3A MO- or Neurog2 MO-injected embryos. cMO, 3A MO or Neurog2 MOs with were injected into one-cell stage embryos at indicated doses followed by RT-qPCR for tubb2b and myt1 expression at stage 16. P<0.05; P<0.01; **P<0.005 (two-tailed Student’s t-test). Doses of mRNAs injected for CoIP in B-D were 3 ng 6MT-KDM3A and 1.5 ng HA-Neurog2 (B); 3 ng HA-KDM3A, 0.5 ng WT Neurog2, 1.5 ng Neurog2-δC, 0.3 ng Neurog2- δN and 1.5 ng bHLH (C); and 50 pg HA-GFP, 1.5 ng HA-GFP-Neurog2, 3 ng 6MT-KDM3A- δC and 50 pg 6MT-KDM3A-δN (D). Numbers by the blots shown in B-D indicate the positions of molecular weight.
	Fig. 5. Neurog2 recruits KDM3A to neuronal targets. (A) Western blot detection of MT-tagged proteins from embryos microinjected with mRNAs encoding 6MT-Neurog2 and 6MT-Ascl1. The doses of 6MT-neurog2 (500 pg) and 6MT-ascl1 (200 pg) mRNA were optimized to ensure both proteins ectopically expressed at comparable levels for the ChIP-qPCR experiments shown in BD. (B) Anti-Myc ChIP followed by qPCR analyses at stage 15, indicating that overexpressed 6MTNeurog2 and 6MT-Ascl1 were able to bind the promoter region of neurod1 (left panel). ef1α and dll1 (delta-like 1) served, respectively, as negative and positive controls. (C) Anti-KDM3A ChIP-qPCR analyses at stage 15, indicating that overexpressed Neurog2 but not Ascl1 could increase KDM3A bound on the promoter region of neurod1. (D) Anti- H3K9me2 ChIP-qPCR analyses at stage 18, indicating that overexpressed Neurog2 but not Ascl1 decreased the H3K9me2 marks on the neurod1 promoter. (E) Anti-KDM3A ChIP-qPCR analyses at stage 15, indicating that Neurog2 MO (80 ng) but not standard MO reduced KDM3A bound on the neurod1 promoter. (F) Anti-H3K9me2 ChIP-qPCR at stage 18 showing that Neurog2 MO (80 ng) but not standard MO increased the H3K9me2 marks on the neurod1 promoter. The ChIP-qPCR experiments shown in B-F were combined from two biological repeats. Three technical replicates were made in each biological repeat. P<0.05; P<0.01; **P<0.005; ns, not significant (two-tailed Student’s t-test).
	Fig. 6. KDM3A-mediated demethylation of H3K9me2 enhances Neurog2 recruitment at the neurod1 promoter. (A) Western blot detection of soluble histone H3, H3K9me1, H3K9me2 and H3K9me3 in control and 2 ng kdm3a mRNA-injected embryos. (B) Embryos stained with red-Gal and in situ hybridized with neurod1 (left panel, stage 18) or tubb2b (right panel, stage 15). β-gal (200 pg) together with 1 ng kdm3a was injected into one cell at the two-cell stage. (C) Western blot detection of chromatin histone H3, H3K9me1, H3K9me2 and H3K9me3 in embryos without injection or injected with 80 ng cMO or 3A MO. See Materials and Methods for isolation of chromatin histones. (D) Anti-H3K9me2 ChIPqPCR analyses showing that 80 ng KDM3A MO increased the H3K9me2 marks on the neurod1 promoter. (E) Anti-H3K9me2 ChIPqPCR analyses showing that Neurog2 was unable to reduce the H3K9me2 marks on the neurod1 promoter when KDM3A was depleted by injecting 80 ng 3A MO. (F) Anti-Myc ChIPqPCR analyses showing that ectopic Neurog2 was unable to bind the neurod1 promoter when KDM3A was depleted (cMO or 3A MO: 80 ng). ChIP-qPCR results shown in D-F were combined from two biological repeats. Three technical replicates were made in each biological repeat. P<0.05; P<0.01; **P<0.005, ns, not significant (two-tailed Student’s t-test). (G) A summary of our current findings that posit the roles for KDM3A in regulating chromatin states and its collaboration with Neurog2 to initiate neuronal precursor differentiation. See details in the Discussion. 3681
	Fig. S1 (related to Fig. 1). MO-mediated depletion of KDM3A and assessment of the function of KDM3A in the developmental potential of naïve ectoderm. (A) Schematic depiction of the start codon spanning sequence of KDM3A CDS and the KDM3A MO target sequence. The sequence of a 5mis-match mutant version of KDM3A MO, designed cMO, shown in comparison with the KDM3A MO. (B) qPCR analyses of gene expression in animal cap explants. cMO: 80ng; 3A MO: 80ng. Noggin mRNA: 200 pg. WE: whole embryo; AC: animal cap explants. ns: no significance according to two-tailed Student t-test. (C) Animal cap explants treated with or without Activin protein. 80ng cMO and 3A MO were injected into the animal pole at the 2-cellstage and animal caps were dissected at the stage 8.5 and treated with or without Activin (5 ng/ml) for four hours. All explants were then cultured in simple saline to the sibling stage 18. (D, E) Western blot data showing that KDM3A MO/MO2 injection did not affect the expression of microinjected 6MT-Neurog2 (D) or 6MT-Ascl1 (E). cMO: 80 ng; 3A MO: 80 ng; 3A MO2: 80 ng. 6MT-neurog2 mRNA: 500 pg; MT-ascl1: 200 pg. MO and mRNA were sequentially injected into both cells at the 2-cell stage. Embryos were lysed at stage 11.
	Fig. S2 (related to Fig. 2). Assessment of the effects of KDM3A depletion on neural development. (A, B, C) Control (cMO) and KDM3A morphant (3A MO) embryos at stage 18in situ hybridized with indicated marker genes. (A) Neuronal genes. (B) Markers for neural progenitors (sox3, pax6) and the anteroposterior pattern (rx, krox20, en2, and hoxb9). (C) Markers for neural plate border, neural crest specification and epidermis. (D) Dorsal view of embryos at stage 18 showing that KDM3A depletion by injecting 60 ng 3A MO delays neural tube closure, which was partially rescued by injecting back of 500 pg the MO-resistant kdm3a mRNA.
	Fig. S3 (related to Fig. 2). KDM3A is required for primary neurogenesis in Xenopus and depletion of KDM3A using a second MO (MO2). (A, B) Embryos at stage 18 in situ hybridized with tubb2b (A, dorsal view) or neurod1 (B, dorsal anterior view). 80 ng 3A MO or cMO was injected into both cells at the 2-cell stage. mRNA encoding Xneurog1 (100 pg), Xneurog3 (100 pg), or mouse Neurog2 (50 pg) was injected into two dorsal cells at the 4-cell stage. (C) A schematic depiction of KDM3A MO and MO2 targeting different locations that are critical for the translation control of kdm3a mRNA. (D) Western blot detection of KDM3A protein at stage 11 after KDM3A MO2 injection at the 2-cell stage (40 and 80 ng). (E) Semi-quantitative PCR analyses of gene expression in animal caps with indicated treatment. cMO: 80 ng. 3A MO2: 80 ng. neurog2 mRNA: 100 pg. (F) Effects of KDM3A MO2 injection on tubb2b expression analyzed through WISH (left column, dorsal view), neural tube closure (middle column, dorsal anterior view, yellow arrows demarcate the closing neural tubes), and later development (right column, lateral view, solid triangles indicate eyes). cMO: 60 ng. 3A MO2: 60 ng. 3A MO2-resistant kdm3a mRNA: 500 pg.
	Fig. S4 (related to Figs. 3 and 4). Assessment of functional and physical interactions between KDM3A and Neurog2. (A) WISH detection of myt1 in embryos at stage 18 (dorsal view). cMO and 3A MO: 80 ng. neurog2 mRNA:100 pg. (B) Embryos at stage 15/16 in situ hybridized with tubb2b (dorsal view). 3A MO and cMO: 80 ng. RA (1.0 M) was added to culture medium from stage 12 to stage 15/16. (C) Two different types of cells (RA-induced NE-4C cells and human induced-Neu neural stem cells) were lysed and subjected to CoIP for detecting interaction between the endogenous KDM3A and Neurog2. A KDM3A antibody was used to pull down endogenous KDM3A protein, and a Neurog2 antibody was used to detect the signal of endogenous Neurog2 protein.
	Fig. S5 (related to Fig. 5). Assessment of the activities of Ascl1 on neuronal gene expression and H3K9me2. (A) Semi-quantitative PCR analyses of gene expression in animal cap explants treated with Neurog2 or Ascl1. (B) Western blot detection of overexpressed 6MT-Neurog2 and 6MT-Ascl1 in the animal cap explants prepared through the same procedure as done in (A). (C, C’) WISH (C) and RT-qPCR (C’) detection of gene expression in control and Ascl1 splice blocking MO (Ascl1 sMO, 80 ng)-injected embryos at the stage 18. * P<0.05; P<0.01; * P<0.005, according to two-tailed Student t-test.. (D, E) Anti-H3K9me2 ChIP-qPCR analyses showing that 80 ng Ascl1 sMO did not alter the H3K9me2 marks on the promoter regions of neurod1 (D) or tubb2b (E). (F) ChIP-qPCR detection of KDM3A on the -36 bp position of tubb2b promoter. 6MT-neurog2 mRNA (500 pg) and 6MT-ascl1 mRNA (200 pg) were individually injected at the 2-cell stage and embryos were then harvested at the stage 15 followed by ChIP-qPCR procedures. (G-I) ChIP-qPCR data showing the effects of ectopic Neurog2 and Ascl1 on the promoter region of myt1. Both ectopic Neurog2 and Ascl1 were able to bind myt1 promoter (G). Only overexpressed Neurog2 increased the level of KDM3A (H), and decreased the H3K9me2 marks (I) on the promoter of myt1. * P<0.05; P<0.01; * P<0.005; ns: no significance, according to two-tailed Student’s t-test.
	Fig. S6 (related to Fig. 5). Neurog2 transactivates tubb2b in a KDM3A-dependent manner. (A) Schematics of the tubb2b promoter and upstream regulatory sequences that contain several degenerate E-box motifs. Locations of ChIP-qPCR primers are also shown. (B) Dual-luciferase reporter assays showing that the tubb2b promoter-driven luciferase reporter is responsive to overexpressed 100 pg neurog2 depending the presence of intact E-box motifs (E1 and E2). (C) ChIP-qPCR analysis of the binding of overexpressed 6MT-Neurog2 on the promoter region of tubb2b. (D) Anti-H3K9me2 ChIP-qPCR analyses showing that 80 ng 3A MO abolished the ability of ectopic Neurog2 (100 pg) to decrease theH3K9me2 marks on the tubb2b promoter. (E) Anti-KDM3A ChIP-qPCR analyses at stage 15 indicating that Neurog2 MO (80 ng) but not standard MO reduced KDM3A bound on the tubb2b promoter. (F) Anti-H3K9me2 ChIP-qPCR at stage 18 showing that Neurog2 (80 ng) but not control MO increased the H3K9me2 marks on the tubb2b promoter. * P<0.05; P<0.01; *P<0.005, ns: no significance, according to two-tailed Student’s t-test.
	Fig.S7 (related to Figs. 5 and 6). KDM3A facilitates the chromatin binding of Neurog2. (A, B) Anti-H3K9me2 ChIP-qPCR analyses indicating that overexpression of wild type 6MT-KDM3A (1 ng) but not a catalytic mutant 6MT-KDM3A (H1130Y, 1 ng) reduced the expression levels of H3K9me2 on the promoter regions of neurod1 (A) and tubb2b (B). (C, D) Anti-H3K27ac ChIP-qPCR analyses indicating that ectopic 6MT-KDM3A or a catalytic mutant 6MT-KDM3A (H1130Y) did not alter the H3K27ac marks on the promoter regions of neurod1 (C) or tubb2b (D). (E) Anti-H3K9me2ChIP-qPCR analyses showing 3A MO (80 ng) increased the H3K9me2 marks on the promoter region of tubb2b. (F) Anti-Myc ChIP-qPCR analyses indicating that 80 ng 3A MO but not cMO blocked the overexpressed 6MT-Neurog2 from accessing the promoter of tubb2b.

	Fig. 1. Expression of KDM3A in early Xenopus embryos and the effects of its depletion on neurogenic potential. (A-E) The spatial expression pattern of kdm3a mRNA revealed through whole-mount in situ hybridization. (F) Western blot analysis of KDM3A in embryos injected with or without KDM3A MO (3A MO). 3A MO (80 ng) was injected into both cells at the twocell stage. Embryos were then collected at the indicated stages (stage 11-18) and subjected to western blot analysis using an antibody directed against mouse KDM3A. Histone H3 served as a loading control. (G) Western blot analysis showing that injecting the MO-resistant kdm3a mRNA rescued the expression of KDM3A protein in 3A MO-injected embryos. Control MO (cMO) and 3AMO (60 ng) were injected into both cells at the two-cell stage. 3A MO-resistant kdm3a mRNA (500 pg) was injected at the one-cell stage. Embryos were then subjected to western blotting at stage 11. The numbers in G show the relative pixel intensities of bands when the intensity of cMO-injected lane was set to 1.00. (H-J) RTqPCR analyses for gene expression in animal cap explants (AC) with indicated treatments. cMO or KDM3A MO (80 ng) was injected at the one-cell stage. mRNA (100 pg) encoding FGF8a (H), Ascl1 (I) or Neurog2 (J) was injected into animal pole of control embryos or MO-injected embryos at the two-cell stage. Two biological repeats and three technical repeats for each biological repeat were performed. The results from these six repeats were combined together and are represented as mean±s.e.m. The expression level of each examined gene in the group of animal cap explants (AC) that received only mRNA injection was set to 1 for normalization. Two-tailed Student t-test was performed to assess significance. ns, not significant. P<0.05; **P<0.005.
	Fig. 2. KDM3A is required for the primary neurogenesis in Xenopus. (A) A schematic showing the experimental design for B and C. (B,B′) Embryos at stage 24/25 stained with red-Gal and in situ hybridized with tubb2b. cMO or 3A MO (40 ng) was injected into one cell at the two-cell stage. The 3A MO-resistant kdm3a mRNA (300 pg) together with 200 pg β-gal mRNAwas subsequently injected into the 3A MO-injected cell at the two-cell stage. The expression levels of tubb2b were classified into three categories: ‘normal’, as seen in cMO-injected embryos; ‘partial’, as seen in some of KDM3A depleted embryos; and ‘complete’ loss, as seen in the remainder of the KDM3A-depleted embryos. Injection of the MO-resistant kdm3a mRNA partially rescues the expression of tubb2b in KDM3A morphants. The numbers on the top of histograms in B′ are sums from two independent experiments. (C,C′) Embryos at stage 16/17 stained with red-Gal and in situ hybridized with tubb2b. cMO and 3A MO (40 ng) were injected into one cell at the two-cell stage. mRNA (100 pg) encoding neurog2 or ascl1 together with 200 pg β-gal mRNA was subsequently injected into the MO-receiving cell at the two-cell stage. Phenotypes were classified into four categories based on whether ectopic neurons were induced and/or whether tubb2b was ‘partially’ or ‘completely’ lost in the MO-injected side. The numbers on the top of histograms in C′ are the total number of samples from two independent experiments.
	Fig. 3. Neurog2 fails to transactivate its neuronal targets in KDM3A morphant embryos. (A) Embryos at stage 17/18 in situ hybridized with Xenopus neurog2. cMO or 3A MO (80 ng) was injected into both cells at the two-cell stage. (B) qPCR results showing that KDM3A depletion differentially affects the expression of neural progenitor genes (sox2 and neurog2) and neuronal genes (neurod1 and tubb2b). ns, no significance. *P<0.001. cMO or 3A MO (80 ng) was injected into both cells at the two-cell stage. Embryos were processed for RT-qPCR at stage 16. (C,C′) Embryos at stage 18 stained with red-Gal and in situ hybridized with neurod1. cMO or 3A MO (40 ng) was injected into one cell at the two-cell stage. mRNA encoding Neurog2 (100 pg) together with 200 pg β-gal mRNA was subsequently injected into the MO-receiving cell at the two-cell stage. (D,E) ChIP-qPCR detection of H3K27ac (D) and H3K4me3 (E) at the neurod1 and tubb2b promoters. cMO or 3A MO (80 ng) was injected into one-cell stage embryos. mRNA encoding Neurog2 (100 pg) was injected into both cells of half of the MO-injected embryos at the two-cell stage. ef1α expression was examined as a negative control. Locations of the qPCR primers are schematically shown in the x axes. The ChIP-qPCR experiments shown in D,E were combined from two independent repeats. Three technical replicates were made in each independent repeat. P<0.01; ***P<0.005; ns, not significant (according to a two-tailed Student’s t-test).
	Fig. 4. KDM3A is associated with Neurog2. (A) Schematics of the domain organization of Neurog2 and KDM3A proteins. (B-D) CoIP followed by western blot analyses for the interaction between Neurog2 and KDM3A. mRNAs encoding the full-length epitope-tagged Neurog2 and KDM3A (B), deletion mutants of Neurog2 and the full-length KDM3A (C), or the full-length Neurog2 and deletion mutants of KDM3A (D) were microinjected into two-cell stage embryos, which were then collected at stage 14 and subjected to CoIP followed by western blot analyses with indicated antibodies. Bands around the 55 kDa and 40 kDa marks in the top blot are non-specific. (E) qPCR analysis of gene expression in cMO-, 3A MO- or Neurog2 MO-injected embryos. cMO, 3A MO or Neurog2 MOs with were injected into one-cell stage embryos at indicated doses followed by RT-qPCR for tubb2b and myt1 expression at stage 16. P<0.05; P<0.01; **P<0.005 (two-tailed Student’s t-test). Doses of mRNAs injected for CoIP in B-D were 3 ng 6MT-KDM3A and 1.5 ng HA-Neurog2 (B); 3 ng HA-KDM3A, 0.5 ng WT Neurog2, 1.5 ng Neurog2-δC, 0.3 ng Neurog2- δN and 1.5 ng bHLH (C); and 50 pg HA-GFP, 1.5 ng HA-GFP-Neurog2, 3 ng 6MT-KDM3A- δC and 50 pg 6MT-KDM3A-δN (D). Numbers by the blots shown in B-D indicate the positions of molecular weight.
	Fig. 5. Neurog2 recruits KDM3A to neuronal targets. (A) Western blot detection of MT-tagged proteins from embryos microinjected with mRNAs encoding 6MT-Neurog2 and 6MT-Ascl1. The doses of 6MT-neurog2 (500 pg) and 6MT-ascl1 (200 pg) mRNA were optimized to ensure both proteins ectopically expressed at comparable levels for the ChIP-qPCR experiments shown in BD. (B) Anti-Myc ChIP followed by qPCR analyses at stage 15, indicating that overexpressed 6MTNeurog2 and 6MT-Ascl1 were able to bind the promoter region of neurod1 (left panel). ef1α and dll1 (delta-like 1) served, respectively, as negative and positive controls. (C) Anti-KDM3A ChIP-qPCR analyses at stage 15, indicating that overexpressed Neurog2 but not Ascl1 could increase KDM3A bound on the promoter region of neurod1. (D) Anti- H3K9me2 ChIP-qPCR analyses at stage 18, indicating that overexpressed Neurog2 but not Ascl1 decreased the H3K9me2 marks on the neurod1 promoter. (E) Anti-KDM3A ChIP-qPCR analyses at stage 15, indicating that Neurog2 MO (80 ng) but not standard MO reduced KDM3A bound on the neurod1 promoter. (F) Anti-H3K9me2 ChIP-qPCR at stage 18 showing that Neurog2 MO (80 ng) but not standard MO increased the H3K9me2 marks on the neurod1 promoter. The ChIP-qPCR experiments shown in B-F were combined from two biological repeats. Three technical replicates were made in each biological repeat. P<0.05; P<0.01; **P<0.005; ns, not significant (two-tailed Student’s t-test).
	Fig. 6. KDM3A-mediated demethylation of H3K9me2 enhances Neurog2 recruitment at the neurod1 promoter. (A) Western blot detection of soluble histone H3, H3K9me1, H3K9me2 and H3K9me3 in control and 2 ng kdm3a mRNA-injected embryos. (B) Embryos stained with red-Gal and in situ hybridized with neurod1 (left panel, stage 18) or tubb2b (right panel, stage 15). β-gal (200 pg) together with 1 ng kdm3a was injected into one cell at the two-cell stage. (C) Western blot detection of chromatin histone H3, H3K9me1, H3K9me2 and H3K9me3 in embryos without injection or injected with 80 ng cMO or 3A MO. See Materials and Methods for isolation of chromatin histones. (D) Anti-H3K9me2 ChIPqPCR analyses showing that 80 ng KDM3A MO increased the H3K9me2 marks on the neurod1 promoter. (E) Anti-H3K9me2 ChIPqPCR analyses showing that Neurog2 was unable to reduce the H3K9me2 marks on the neurod1 promoter when KDM3A was depleted by injecting 80 ng 3A MO. (F) Anti-Myc ChIPqPCR analyses showing that ectopic Neurog2 was unable to bind the neurod1 promoter when KDM3A was depleted (cMO or 3A MO: 80 ng). ChIP-qPCR results shown in D-F were combined from two biological repeats. Three technical replicates were made in each biological repeat. P<0.05; P<0.01; **P<0.005, ns, not significant (two-tailed Student’s t-test). (G) A summary of our current findings that posit the roles for KDM3A in regulating chromatin states and its collaboration with Neurog2 to initiate neuronal precursor differentiation. See details in the Discussion. 3681
	Fig. S1 (related to Fig. 1). MO-mediated depletion of KDM3A and assessment of the function of KDM3A in the developmental potential of naïve ectoderm. (A) Schematic depiction of the start codon spanning sequence of KDM3A CDS and the KDM3A MO target sequence. The sequence of a 5mis-match mutant version of KDM3A MO, designed cMO, shown in comparison with the KDM3A MO. (B) qPCR analyses of gene expression in animal cap explants. cMO: 80ng; 3A MO: 80ng. Noggin mRNA: 200 pg. WE: whole embryo; AC: animal cap explants. ns: no significance according to two-tailed Student t-test. (C) Animal cap explants treated with or without Activin protein. 80ng cMO and 3A MO were injected into the animal pole at the 2-cellstage and animal caps were dissected at the stage 8.5 and treated with or without Activin (5 ng/ml) for four hours. All explants were then cultured in simple saline to the sibling stage 18. (D, E) Western blot data showing that KDM3A MO/MO2 injection did not affect the expression of microinjected 6MT-Neurog2 (D) or 6MT-Ascl1 (E). cMO: 80 ng; 3A MO: 80 ng; 3A MO2: 80 ng. 6MT-neurog2 mRNA: 500 pg; MT-ascl1: 200 pg. MO and mRNA were sequentially injected into both cells at the 2-cell stage. Embryos were lysed at stage 11.
	Fig. S2 (related to Fig. 2). Assessment of the effects of KDM3A depletion on neural development. (A, B, C) Control (cMO) and KDM3A morphant (3A MO) embryos at stage 18in situ hybridized with indicated marker genes. (A) Neuronal genes. (B) Markers for neural progenitors (sox3, pax6) and the anteroposterior pattern (rx, krox20, en2, and hoxb9). (C) Markers for neural plate border, neural crest specification and epidermis. (D) Dorsal view of embryos at stage 18 showing that KDM3A depletion by injecting 60 ng 3A MO delays neural tube closure, which was partially rescued by injecting back of 500 pg the MO-resistant kdm3a mRNA.
	Fig. S3 (related to Fig. 2). KDM3A is required for primary neurogenesis in Xenopus and depletion of KDM3A using a second MO (MO2). (A, B) Embryos at stage 18 in situ hybridized with tubb2b (A, dorsal view) or neurod1 (B, dorsal anterior view). 80 ng 3A MO or cMO was injected into both cells at the 2-cell stage. mRNA encoding Xneurog1 (100 pg), Xneurog3 (100 pg), or mouse Neurog2 (50 pg) was injected into two dorsal cells at the 4-cell stage. (C) A schematic depiction of KDM3A MO and MO2 targeting different locations that are critical for the translation control of kdm3a mRNA. (D) Western blot detection of KDM3A protein at stage 11 after KDM3A MO2 injection at the 2-cell stage (40 and 80 ng). (E) Semi-quantitative PCR analyses of gene expression in animal caps with indicated treatment. cMO: 80 ng. 3A MO2: 80 ng. neurog2 mRNA: 100 pg. (F) Effects of KDM3A MO2 injection on tubb2b expression analyzed through WISH (left column, dorsal view), neural tube closure (middle column, dorsal anterior view, yellow arrows demarcate the closing neural tubes), and later development (right column, lateral view, solid triangles indicate eyes). cMO: 60 ng. 3A MO2: 60 ng. 3A MO2-resistant kdm3a mRNA: 500 pg.
	Fig. S4 (related to Figs. 3 and 4). Assessment of functional and physical interactions between KDM3A and Neurog2. (A) WISH detection of myt1 in embryos at stage 18 (dorsal view). cMO and 3A MO: 80 ng. neurog2 mRNA:100 pg. (B) Embryos at stage 15/16 in situ hybridized with tubb2b (dorsal view). 3A MO and cMO: 80 ng. RA (1.0 M) was added to culture medium from stage 12 to stage 15/16. (C) Two different types of cells (RA-induced NE-4C cells and human induced-Neu neural stem cells) were lysed and subjected to CoIP for detecting interaction between the endogenous KDM3A and Neurog2. A KDM3A antibody was used to pull down endogenous KDM3A protein, and a Neurog2 antibody was used to detect the signal of endogenous Neurog2 protein.
	Fig. S5 (related to Fig. 5). Assessment of the activities of Ascl1 on neuronal gene expression and H3K9me2. (A) Semi-quantitative PCR analyses of gene expression in animal cap explants treated with Neurog2 or Ascl1. (B) Western blot detection of overexpressed 6MT-Neurog2 and 6MT-Ascl1 in the animal cap explants prepared through the same procedure as done in (A). (C, C’) WISH (C) and RT-qPCR (C’) detection of gene expression in control and Ascl1 splice blocking MO (Ascl1 sMO, 80 ng)-injected embryos at the stage 18. * P<0.05; P<0.01; * P<0.005, according to two-tailed Student t-test.. (D, E) Anti-H3K9me2 ChIP-qPCR analyses showing that 80 ng Ascl1 sMO did not alter the H3K9me2 marks on the promoter regions of neurod1 (D) or tubb2b (E). (F) ChIP-qPCR detection of KDM3A on the -36 bp position of tubb2b promoter. 6MT-neurog2 mRNA (500 pg) and 6MT-ascl1 mRNA (200 pg) were individually injected at the 2-cell stage and embryos were then harvested at the stage 15 followed by ChIP-qPCR procedures. (G-I) ChIP-qPCR data showing the effects of ectopic Neurog2 and Ascl1 on the promoter region of myt1. Both ectopic Neurog2 and Ascl1 were able to bind myt1 promoter (G). Only overexpressed Neurog2 increased the level of KDM3A (H), and decreased the H3K9me2 marks (I) on the promoter of myt1. * P<0.05; P<0.01; * P<0.005; ns: no significance, according to two-tailed Student’s t-test.
	Fig. S6 (related to Fig. 5). Neurog2 transactivates tubb2b in a KDM3A-dependent manner. (A) Schematics of the tubb2b promoter and upstream regulatory sequences that contain several degenerate E-box motifs. Locations of ChIP-qPCR primers are also shown. (B) Dual-luciferase reporter assays showing that the tubb2b promoter-driven luciferase reporter is responsive to overexpressed 100 pg neurog2 depending the presence of intact E-box motifs (E1 and E2). (C) ChIP-qPCR analysis of the binding of overexpressed 6MT-Neurog2 on the promoter region of tubb2b. (D) Anti-H3K9me2 ChIP-qPCR analyses showing that 80 ng 3A MO abolished the ability of ectopic Neurog2 (100 pg) to decrease theH3K9me2 marks on the tubb2b promoter. (E) Anti-KDM3A ChIP-qPCR analyses at stage 15 indicating that Neurog2 MO (80 ng) but not standard MO reduced KDM3A bound on the tubb2b promoter. (F) Anti-H3K9me2 ChIP-qPCR at stage 18 showing that Neurog2 (80 ng) but not control MO increased the H3K9me2 marks on the tubb2b promoter. * P<0.05; P<0.01; *P<0.005, ns: no significance, according to two-tailed Student’s t-test.
	Fig.S7 (related to Figs. 5 and 6). KDM3A facilitates the chromatin binding of Neurog2. (A, B) Anti-H3K9me2 ChIP-qPCR analyses indicating that overexpression of wild type 6MT-KDM3A (1 ng) but not a catalytic mutant 6MT-KDM3A (H1130Y, 1 ng) reduced the expression levels of H3K9me2 on the promoter regions of neurod1 (A) and tubb2b (B). (C, D) Anti-H3K27ac ChIP-qPCR analyses indicating that ectopic 6MT-KDM3A or a catalytic mutant 6MT-KDM3A (H1130Y) did not alter the H3K27ac marks on the promoter regions of neurod1 (C) or tubb2b (D). (E) Anti-H3K9me2ChIP-qPCR analyses showing 3A MO (80 ng) increased the H3K9me2 marks on the promoter region of tubb2b. (F) Anti-Myc ChIP-qPCR analyses indicating that 80 ng 3A MO but not cMO blocked the overexpressed 6MT-Neurog2 from accessing the promoter of tubb2b.