Gazdag E et al. (2016), Activation of a T-box-Otx2-Gsc gene network ind...

XB-ART-51942

Development 2016 Apr 15;1438:1340-50. doi: 10.1242/dev.127936.

Show Gene links Show Anatomy links

Activation of a T-box-Otx2-Gsc gene network independent of TBP and TBP-related factors.

Gazdag E , Jacobi UG , van Kruijsbergen I , Weeks DL , Veenstra GJ .

???displayArticle.abstract???
Embryonic development relies on activating and repressing regulatory influences that are faithfully integrated at the core promoter of individual genes. In vertebrates, the basal machinery recognizing the core promoter includes TATA-binding protein (TBP) and two TBP-related factors. In Xenopus embryos, the three TBP family factors are all essential for development and are required for expression of distinct subsets of genes. Here, we report on a non-canonical TBP family-insensitive (TFI) mechanism of transcription initiation that involves mesoderm and organizer gene expression. Using TBP family single- and triple-knockdown experiments, α-amanitin treatment, transcriptome profiling and chromatin immunoprecipitation, we found that TFI gene expression cannot be explained by functional redundancy, is supported by active transcription and shows normal recruitment of the initiating form of RNA polymerase II to the promoter. Strikingly, recruitment of Gcn5 (also known as Kat2a), a co-activator that has been implicated in transcription initiation, to TFI gene promoters is increased upon depletion of TBP family factors. TFI genes are part of a densely connected TBP family-insensitive T-box-Otx2-Gsc interaction network. The results indicate that this network of genes bound by Vegt, Eomes, Otx2 and Gsc utilizes a novel, flexible and non-canonical mechanism of transcription that does not require TBP or TBP-related factors.

???displayArticle.pubmedLink??? 26952988
???displayArticle.pmcLink??? PMC4852510
???displayArticle.link??? Development
???displayArticle.grants??? [+]

Species referenced: Xenopus
Genes referenced: bmp4 crx eomes fgf8 foxd3 gsc h3-3a h4c1 kat2a lhx1 mark3 nodal nodal1 not otx2 rhob taf6 tbp tbpl1 tbpl2 vegt zic1
???displayArticle.antibodies??? Acetylated H3f3a Ab15 Acetylated h4c1 Ab6 Kat2a Ab1 Kat2a Ab2 Polr2a Ab2 Polr2a Ab6 Tbp Ab1

???displayArticle.gses??? GSE76991: Xenbase, NCBI

GSE76994: Xenbase, NCBI
GSE76995: NCBI

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

	Fig. 1. Analyses of tbp, tlf and tbp2 mRNA ablation, embryonic gene activation and α-amanitin treatment reveal TBP family-insensitive gene transcripts. (A) Box plots of the fold change of groups of transcripts upon TBP, TLF or TBP2 knockdown as measured by microarray. Expression ratio of stages 10.5 (early gastrula) and 7 (blastula) is also plotted (10/7). TBP-, TLF- and TBP2-dependent transcripts (first three panels) were identified using a consistent statistical change call between replicates for the initiation factor that was ablated, in addition to a fold change (log2 ratio<−1.5) in expression. The unaffected group of transcripts (fourth panel) was identified by selecting for developmentally upregulated transcripts (between stage 7 and 10.5) with no significant change upon injection of TBP family antisense oligos. Boxplots are standard box plots generated in R, showing the interquartile range (IQR) around the median; the whiskers extend from the minimum value to the maximum value unless the distance to the first and third quartiles is more than 1.5 times the IQR. The fold change cut off range is shaded gray in each panel. (B) Phenotype of control and α-amanitin-treated embryos at stage 10.5. (C) Effect of α-amanitin treatment on transcript levels of ‘unaffected’ TFI genes, demonstrating that these transcripts are newly made in the embryo, with the exception of gapdh (maternally provided control) and vegt which is also partly maternally provided. (D) K-means clustering of transcript expression ratios [antisense oligo-treated relative to control; stage 10.5 relative to stage 7; α-amanitin (α-ama) relative to control] of developmentally active transcripts (defined as transcripts that are either developmentally upregulated or affected by α-amanitin at stage 10.5). Each column represents one of two replicate experiments. A cluster of genes insensitive for TBP, TLF, or TBP2 depletion is marked on the left.
	Fig. 2. Analysis of TBP family triple knockdown (ablation of tbp, tlf and tbp2 mRNA). (A) Morphology of TKD embryos at stage 10.5, relative to water-injected control (left panel). (B) Control experiments verified efficient TBP protein depletion (western blot, left panel) and knockdown of tlf and tbp2 transcripts by RT-qPCR in TKD embryos relative to control (Ctrl). Tubulin (western) and gapdh serve as controls. (C) Expression levels of TFI gene transcripts (left, in blue) and transcripts that require TBP, TBP2 or TLF (light gray) in TKD embryos relative to control embryos, as determined by RT-qPCR. gapdh transcript levels (dark gray) served as control. (D) Box plots showing fold change (log2) of transcript levels of all expressed genes, decreased genes (DEseq FDR 0.1, see Materials and Methods) and TFI genes, as determined by RNA sequencing of ribosomal RNA-depleted samples of TKD embryos versus control embryos. (E) Recruitment of initiating RNAPII-Ser5 determined by ChIP-qPCR. 5′ and 3′ ends of TFI genes were analyzed. Comparison of signals obtained from control (light blue) and TKD embryos (dark blue) indicates significant recruitment of RNAPII-Ser5 to TFI genes under TKD conditions.
	Fig. 3. Characterization of TFI gene transcript localization and function. (A) Over-representation (fold enrichment) of TFI gene transcripts among localized expression in the animal, vegetal, dorsal and ventral parts of the embryo. TFI transcripts are not over-represented among ventrally expressed transcripts. Hypergeometric P-values are indicated. (B) Over-representation of TFI gene transcripts within four clusters of genes identified after ectopic induction of the organizer by Noggin (Nog) and Dickkopf-1 (Dkk1). Clusters 3 and 3-extended represent organizer genes (Hufton et al., 2006). TFI gene transcripts are moderately but significantly over-represented in these clusters. (C) Gene ontology (GO) term analysis of TFI genes relative to GO terms in all genes (blue) or in developmentally active genes (orange, defined as in Fig. 1D). The GO fold enrichments are reflected in the sizes of the circles (see scale in gray). Significance (FDR; top) and relevant GO terms (right) are indicated.
	Fig. 4. Analysis of Gcn5 function in early embryos. A number of TFI genes recruit Gcn5 to their promoter and require Gcn5 for normal expression. (A) Gcn5 binding at different regions around the transcription start site (TSS) of bmp4, gsc and fgf8 genes in X. tropicalis. ChIP was performed using two different polyclonal anti-GCN5 antibodies. Negative controls (3′ ends of genes and an intergenic region) are indicated. One of the antibodies (C26A10) is highly specific for Gcn5 in western blotting (cf. Fig. S2B). In addition, ChIP signals are reduced in Gcn5-AS-injected embryos (cf. panel C). (B) Depletion of gcn5 transcripts was verified by RT-qPCR. Expression levels normalized by maternal gapdh levels were determined for both 5′ and 3′ regions of the gcn5 mRNA and compared with a control, taf6. (C) Binding of Gcn5 protein to X. laevis promoters is reduced upon gcn5 knockdown, as assessed by ChIP-qPCR using anti-GCN5 C26A10 antibody in control (light blue) and Gcn5-knockdown (dark blue) embryos (stage 10.5). (D) Morphology of control (water-injected) and Gcn5-knockdown (KD, Gcn5-AS injected) embryos showing gastrulation defects at stage 10.5-11. (E) Statistics of rescue experiments performed by co-injecting in vitro-transcribed full-length human GCN5 mRNA (FL Gcn5) together with Gcn5-AS oligos to restore normal development (cf. Fig. S2). Statistics of three independent experiments are summarized. (F) Box plots showing fold change (log2) of transcript levels in duplicate samples of Gcn5-KD and control embryos (left panel), shown for all expressed genes, genes with decreased (Down) and increased (Up) transcripts (DEseq FDR 0.1), genes with decreased transcripts in TKD embryos and TFI genes. Right panel shows reciprocal analysis for TKD conditions. The fold changes are depicted in subsets of genes, decreased transcripts in TKD embryos (Down), TFI genes, transcripts that are decreased (Gcn5 Down) or increased (Gcn5 Up) in Gcn5-AS embryos.
	Fig. 5. Increased promoter binding of Gcn5 in TBP family triple-knockdown embryos. (A) ChIP analysis of Gcn5, TBP and histone H3 and H4 acetylation under VP16 activating conditions in TKD and control embryos. These experiments were performed in a VP16 transcription activation assay combined with TBP family loss-of-function experiments (cf. Fig. S3). Loss of TBP on the promoter in TKD embryos is compensated by an increase in Gcn5 (first two panels). Histones H3 and H4 are acetylated in both TKD and control embryos (second two panels). (B) Gcn5 is recruited to TFI gene promoters in TBP family triple-knockdown embryos. ChIP reveals enhanced Gcn5 binding when TBP and TBP-related factors are depleted in TKD embryos (blue) compared with water-injected controls (light blue). The intronic region of nadh gene shows background levels. The mark3 gene requires normal TBP levels for expression.
	Fig. 6. TFI network analysis. (A) TFI genomic interaction network (blue box) based on ChIP data of Otx2, Vegt, Eomes, Gsc, T and Lhx1 in X. tropicalis. The most common binding combinations with representative TFI genes are depicted on the bottom row. The total number of TFI genes with the same binding combinations is indicated. For a complete overview, see Fig. S4. (B) Over-representation of TFI genes among genes bound by Otx2, Vegt, Eomes, Gsc, T and Lhx1. Hypergeometric P-values are indicated. (C) Indegree (number of inputs) of TFI gene transcripts (blue) and all genes (gray). TFI gene transcripts are often bound by four or more of the transcription factors Otx2, Vegt, Eomes, Gsc, T and Lhx1.
	Supplemental Figure S1. (A) Recruitment of RNA polymerase II (RNAPII) to TFI genes determined by ChIP‐qPCR using an antibody recognizing the (unphosphorylated) carboxy‐terminal domain (CTD). 5´ and 3´ ends of TBP family‐insensitive (TFI) genes were analyzed. Signals obtained from control (light blue) and TKD (dark blue) embryos indicate RNAPII enrichment relative to a negative control region. (B) RNAPII pausing index of TFI orthologs in X.tropicalis compared to other genes using RNAPII ChIP‐seq data (van Heeringen et al., 2014).
	Supplemental Figure S2. (A) Transcript levels of gcn5 and gapdh transcript levels in TBP, TLF, TBP2 triple knockdown (TKD) embryos and control embryos, as determined by RT‐qPCR. (B) Reactivity of C26A10 antibody with Xenopus Gcn5 in embryos injected with 1 ng of gcn5 mRNA, Gcn5‐AS oligonucleotide and in control embryos (weak signal). No cross‐reactive bands are observed. TBP is shown as control. For validation of TBP, TBP2 and RNAPII antibodies see (Akhtar and Veenstra, 2009; Jallow et al., 2004; Veenstra et al., 1999). (C) TUNEL assay for the detection of apoptosis in control and Gcn5‐AS‐injected embryos. (D) Morphology of non‐injected and water‐injected controls, Gcn5‐AS‐injected and rescue (Gcn5‐AS rescued with co‐injected full length human GCN5 mRNA or GCN5‐ΔHAT mRNA encoding a truncated GCN5 protein lacking the histone acetyltransferase domain) embryos are shown. Photos were taken at stages 11.5 and 23 (controls). (E) Summary of survival (normal development) statistics after embryo collection at stage 10.5 and 23. Gcn5-AS embryos were rescued by full length human GCN5 mRNA at a rate of 80-86% at stage 10.5 and 38-72% at stage 23. Embryos rescued with GCN5-ΔHAT exhibited 64-91% rescue efficiency at stage 10.5 and 21-30% at stage 23. (F) Gene ontology analysis of decreased transcripts in Gcn5-AS embryos.
	Supplemental Figure S3. VP16 transcription activation assay combined with TBP family loss‐of‐function experiments. See Fig. 5A for ChIP analysis in the assay system. (A) Schematic overview of VP16 assay. (B) Transcription in non‐activating and VP16‐activated conditions (left panel) and basal levels of transcription (right panel) are compared in TBP family triple knockdown (TKD) and control embryos. Embryos were staged according to the morphology of control embryos. Stages of 9, 10.5 and 12 were collected for analysis.
	Supplemental Figure S4. (A) The subset of the mesoderm specification network (Koide et al., 2005) as it relates to TFI genes (blue box) and BMP, FGF and Activin/nodal signaling. (B) Hierarchical clustering of TFI genomic interaction network as determined by assigning ChIP‐seq peaks of Lhx1, Gsc, Otx2, Eomes, Vegt and T (Xbra) to genic regions (Methods). In addition, expression changes in gsc and lhx1/otx2/otx5 morphant (Mo) embryos are shown (Yasuoka et al., 2014). Color intensity reflects number of peaks in the locus (ChIP, 0, 1 or >=2 peaks) or Log2 fold expression change (RNA‐seq, yellow decreased, blue increased, grey not determined). The genes shown to the right represent all named X. tropicalis TFI genes.

References [+] :

Adelman, Promoter-proximal pausing of RNA polymerase II: emerging roles in metazoans. 2012, Pubmed