Drosophila Ebi mediates Snail-dependent transcriptional repression through HDAC3-induced histone deacetylation

Abstract

The Drosophila Snail protein is a transcriptional repressor that is necessary for mesoderm formation. Here, we identify the Ebi protein as an essential Snail co-repressor. In ebi mutant embryos, Snail target genes are derepressed in the presumptive mesoderm. Ebi and Snail interact both genetically and physically. We identify a Snail domain that is sufficient for Ebi binding, and which functions independently of another Snail co-repressor, Drosophila CtBP. This Ebi interaction domain is conserved among all insect Snail-related proteins, is a potent repression domain and is required for Snail function in transgenic embryos. In mammalian cells, the Ebi homologue TBL1 is part of the NCoR/SMRT-HDAC3 (histone deacetylase 3) co-repressor complex. We found that Ebi interacts with Drosophila HDAC3, and that HDAC3 knockdown or addition of a HDAC inhibitor impairs Snail-mediated repression in cells. In the early embryo, Ebi is recruited to a Snail target gene in a Snail-dependent manner, which coincides with histone hypoacetylation. Our results demonstrate that Snail requires the combined activities of Ebi and CtBP, and indicate that histone deacetylation is a repression mechanism in early Drosophila development.

Figure 1

Similar developmental defect in embryos derived from ebi^k16213 germline clones and sna¹⁸ embryos. Cuticle preparations of newly hatched wild type (A), ebi^k16213 germline clone mutant (B) and sna¹⁸ zygotic mutant (C) embryos. The ‘U-shape' phenotype is shared between ebi and sna mutants, suggesting a common biological function. ebi mRNA is maternally distributed in the whole embryo in wt (D), but is undetectable in ebi^k16213 germline clone embryos (E) hybridized with a digoxigenin-labelled antisense ebi probe. All embryos shown in this study are oriented with anterior to the left.

Figure 2

Snail target genes are expanded into the presumptive mesoderm in ebi mutant embryos. Ventral or ventro-lateral views of the embryos are shown. Wild-type (wt) (A, C, E, G) and ebi germline clone embryos (B, D, F, H) were hybridized with digoxigenin-labelled RNA probes or a Snail antibody. Rhomboid (rho) is expressed in two lateral stripes in pre-cellular stage wt embryos (A), but is derepressed into the ventral region in ebi mutants (B). Short gastrulation (sog) expression is similarly altered in the ebi mutant (D) compared to wt (C). Embryos at stage 9 have invaginated the mesoderm, and single-minded (sim) is only expressed along the ventral midline in wt (E), but is expanded to the mesoderm in ebi mutant embryos (F). Snail protein expression in ebi mutant cellularized embryos (H) is comparable to wt embryos (G). (I–L) Snail but not Krüppel (Kr) is unable to repress a modified rho neuroectoderm enhancer (NEE)-lacZ reporter gene in ebi mutant embryos. Two different lacZ reporter genes were introduced into wt embryos (I, K) or embryos derived from ebi germline clones (J, L) and stained with a lacZ RNA antisense probe. Ventro-lateral views of cellularized embryos are shown. (I, J) A lacZ reporter driven by a rho NEE containing synthetic Snail-binding sites. lacZ expression is repressed by endogenous Snail in the mesoderm in wt (I) but is derepressed in ebi mutant embryos (J). A lacZ reporter driven by a similar NEE that contains synthetic Kr-binding sites can be repressed by Kr in the central domain of the embryo in both wt (K) and ebi mutants (L). Schematic drawings of the reporter genes are shown underneath the embryo images.

Figure 3

Dissection of an Ebi interaction domain in the Snail protein. (A) GST pull-down assay. A series of GST–Snail fusion proteins showed different binding abilities to in vitro-translated Ebi or CtBP. GST alone served as a negative control. Notably, Sna 1–40 and Sna 4–21 interact only with Ebi. (B) A schematic drawing summarizing the binding activities of all tested Sna fusions to Ebi or CtBP. In the right panel, ‘+' or ‘−' symbol indicates strong binding or weak binding/no binding, respectively. (C) Alignment of the N termini of insect Snail proteins with Clustal W (DNASTAR Lasergene). Amino acids 5 and 8–14 in Drosophila melanogaster Snail are identical among all insect Snail-related sequences. Flies (D.m., Drosophila melanogaster, D.p., Drosophila pseudoobscura, D.v., Drosophila virilis, D.g., Drosophila grimshawi), mosquito (A.g., Anopheles gambiae), silkworm (B.m., Bombyx mori), honeybee (A.m., Apis mellifera) and flour beetle (T.c., Tribolium casteneum) were compared. (D) Snail is able to interact with Ebi and CtBP simultaneously. Ebi and Snail were in vitro-translated and incubated with GST fusion proteins separately or together. GST–CtBP interacts strongly with Snail (lane 6), but not with Ebi alone (lane 7). However, in the presence of Snail, GST–CtBP can also bind to Ebi (lane 5, arrow), suggesting that Snail can bridge the interaction between CtBP and Ebi by binding both proteins at the same time. (E) GST–Ebi fusion proteins binding to in vitro-translated Snail. Full-length Ebi and the Ebi C terminus bind to Snail, but the Ebi N terminus does not interact.

Figure 4

The Ebi interaction domain in Snail is a potent repression domain. (A) Snail amino acids 4–21, 1–40, 1-40Δ5-25, 1-245 or 1-245Δ5-25 were fused to the tetracycline repressor (TetR) DNA-binding domain. These plasmids were co-transfected into S2 cells with a luciferase reporter gene driven by the actin 5C enhancer containing Tet operators, as well as an actin 5C-driven lacZ gene to control for transfection efficiency. Luciferase activity (normalized for β-galactosidase activity) of unfused TetR is set to 1, and activity of TetR–Snail fusion proteins was plotted relative to the TetR. The value in the plot is the average from at least three independent experiments and error bars indicate the standard deviation. A schematic drawing of the reporter plasmid is depicted below the histogram. (B–D) Lateral views of transgenic embryos expressing lacZ under the control of a modified rho NEE into which Gal4 upstream activating sequences (UASs) have been inserted. LacZ is expressed in ventral cells in embryos containing only the reporter gene (B). Embryos that additionally express a Gal4 DNA-binding domain–Sna 1–40 (C) or Gal4 DNA-binding domain–Sna 1–245 (D) transgene under control of the Kr CD enhancer are shown.

Figure 5

The Ebi interaction motif is essential for Snail-mediated repression. Cellularizing embryos were hybridized with digoxigenin-labelled snail (sna) (A, B) or rho (C, D) antisense RNA probes and oriented with anterior to the left and dorsal up (A, B) or shown from a ventro-lateral view (C, D). Snail is mis-expressed with the Kr CD enhancer in transgenic embryos that contain either full-length Snail coding region (A, C) or Snail with a deletion of amino acids 5–25 (B, D). snail expression is observed both in the ventral region (the endogenous pattern) and in the Kr domain. The expression level of Snail is similar in these transgenic strains. (C) rho expression pattern in the transgenic strain shown in (A). The ectopic snail stripe creates a gap (see arrow) in the rho lateral stripes, but not in the pattern present in the dorsal-most region of the embryo that is regulated independently of Snail. (D) rho expression pattern in the transgenic strain shown in (B). Mutant Snail does not alter the rho pattern, suggesting that amino acids 5–25 are essential for Snail-mediated repression. (E) GST pull-down assay showing that deletion of amino acids 5–25 abolishes the binding of Ebi to full-length Snail, whereas CtBP binding is unaffected.

Figure 6

Ebi associates with HDAC3 and both proteins are required for Snail-mediated repression in S2 cells. (A) Ebi can be co-immunoprecipitated with endogenous HDAC3 from Drosophila embryos and S2 cells. Panels 1 and 2: normal rabbit serum (NRS), guinea-pig HDAC3, rabbit HDAC1, mouse Ebi or mouse GFP antibody was used in immunoprecipitation, and a mouse Ebi antibody used in the western blot to detect the presence of Ebi. Lane 1 is 1% of the input embryo extract. The same membrane was reprobed with a rat HDAC3 antibody. Panels 3 and 4: lanes 1–3 are the input extract (2%) from normal S2 cells and cells treated with HDAC3 or ebi dsRNA, respectively. The guinea-pig HDAC3 antibody was used to immunoprecipitate HDAC3 and its associated proteins, and Ebi was detected by western blot. The blot was reprobed with a rat HDAC3 antiserum. Panels 5 and 6: the Ebi monoclonal was used in immunoprecipitation, and rat HDAC3 antiserum was used in western blot. The membrane was reprobed with the Ebi antibody. The input panels are shown at a longer exposure than the IP panels. (B) Efficiency of RNAi in the Sna 1–40/luc cell line was determined by western blot. Panel 1: Ebi protein levels are reduced by ebi RNAi treatment. Panel 2: HDAC3 protein levels are reduced by HDAC3 RNAi treatment, and to some extent by ebi or HDAC1 RNAi treatment. Panel 3: HDAC1 protein levels are reduced by HDAC1 RNAi, and to a lesser extent by HDAC3 RNAi. Panel 4: CtBP RNAi knocked down both CtBP protein isoforms as detected by a CtBP antibody. Panel 5: tubulin was used as a loading control. (C) RNAi assay performed in stable cell lines containing luciferase reporter alone (luc), luciferase+TetR–Sna 1–40 (Sna40/luc) or luciferase+TetR–Sna 1–245 (Sna245/luc). The repression activity of Snail 1–40 is attenuated by knock down of either Ebi or HDAC3, but not affected by knock down of CtBP or HDAC1. The activity of Sna 1–245 is affected by knockdown of Ebi, HDAC3 or CtBP. Ebi and CtBP RNAi treatment together gives a stronger effect. The luciferase activity in the control cell line is largely unaffected by the RNAi treatments. Error bars indicate the standard deviation. (D) The HDAC inhibitor TSA relieves Sna 1–40-mediated repression, suggesting that histone deacetylation is involved in Ebi-dependent repression. The cell lines expressing Sna 1–245 or Sna 1–245Δ5–25 are insensitive to TSA treatment, indicating that repression occurs by a CtBP-dependent mechanism.

Figure 7

Binding of Ebi to the rho NEE enhancer in the embryo is Snail-dependent and coincides with histone hypoacetylation. (A) In situ hybridization showing expression of snail and rho in wt embryos, in gd⁷ embryos where no Dorsal activator protein translocates to nuclei, in Toll^rm9/rm10 embryos that contain intermediate levels of nuclear Dorsal in all cells and in Toll^10b embryos where high levels of nuclear Dorsal activator are present in all cells. (B) Chromatin immunoprecipitation of the rho NEE enhancer using embryo extracts derived from different genetic backgrounds. Whole cell extract (WCE) represents the input DNA in each extract. In the upper panel, an Ebi-specific antibody was used to show that Ebi binding to the rho NEE occurs in wt embryos and in Toll^10b embryos that contain Snail, but not in embryos that lack Snail. The bottom panel uses an antibody specific to acetylated lysine 14 in histone 3 (H3K14ac). H3 acetylation of the rho NEE enhancer is observed in wt embryos and in Toll^rm9/rm10 embryos that contain the Dorsal activator but lack the Snail repressor. In embryos with both Dorsal and Snail present throughout the dorsal–ventral axis (Toll^10b), H3K14 is hypoacetylated, indicating that Snail recruits a histone deacetylase. The total amount of histone H3 on the rho NEE does not change in the different genotypes (middle panel). A quantification of the DNA present in the immunoprecipitates is shown at the bottom.