ZNF366 is an estrogen receptor corepressor that acts through CtBP and histone deacetylases

Abstract

The regulation of gene expression by estrogen receptor-alpha (ERalpha) requires the coordinated and temporal recruitment of diverse sets of transcriptional co-regulator complexes, which mediate nucleosome remodelling and histone modification. Using ERalpha as bait in a yeast two-hybrid screen, we have identified a novel ERalpha-interacting protein, ZNF366, which is a potent corepressor of ERalpha activity. The interaction between ZNF366 and ERalpha has been confirmed in vitro and in vivo, and is mediated by the zinc finger domains of the two proteins. Further, we show that ZNF366 acts as a corepressor by interacting with other known ERalpha corepressors, namely RIP140 and CtBP, to inhibit expression of estrogen-responsive genes in vivo. Together, our results indicate that ZNF366 may play an important role in regulating the expression of genes in response to estrogen.

Figure 1

The amino acid sequence and tissue distribution of human ZNF366. (A) The deduced amino acid sequence of human ZNF366 is shown in the single letter amino acid code, with the cysteine (C) and histidine (H) residues of the 11 zinc fingers in ZNF366 highlighted, as are the conserved phenylalanine (F) and leucine (L) residues. Potential CtBP binding motifs and an LXXLL motif are underlined. (B) A schematic representation of ZNF366 from different species is shown, together with the position of the zinc fingers and the putative CtBP and LXXLL motifs, the sequences for these motifs in human ZNF366 being shown. The numbers below the representation of ZNF366 from each species refer to percentage amino acid sequence homology relative to human ZNF366. The numbers to the right refer to the predicted number of amino acids encoded by the ZNF366 genes from different species. The region encoding 277–1053 amino acids was used for alignment in the case of the chicken ZNF366. ZNF366 sequences used here have GenBank accession nos NM_152625 (human), NM_001004149 (mouse), XM_226715 (rat), XP_544370 (dog) and XP_429153 (chicken). The Xenopus tropicalis ZNF366 sequence was derived from the ENSEMBL database, from sequences having the Gene ID ENSGALG00000015008. (C) Northern blot of 2 μg of poly(A)+ RNA from the tissues indicated, probed with ZNF366 cDNA. Size markers are shown on the right.

Figure 2

Interaction of ZNF366 with ERα. (A and B) Lysates prepared from COS-1 cells transiently transfected with ERα and ZNF366-FLAG were immunoprecipitated with mouse IgG, or with ERα or FLAG antibodies. Input represents 5% of the total volume of lysate used in the immunoprecipitations. (C) Cultures of PL1α yeast cells transformed with pBridge(Mod), pBridge(Mod)-ERα-ΔLBD or pBridge(Mod)-ERα, together with pACTII or ZNF366 were spotted at various dilutions, as indicated. Yeast growth was performed over 7 days at 30°C, in the presence of 17β-estradiol (E2; 1 nM), 4-hydroxytamoxifen (OHT; 100 nM) or in the absence of ligand (no ligand), on plates lacking tryptophan, leucine and uracil. (D) COS-1 cells transiently transfected with ERα−ΔNLS and/or FLAG-ZNF366 were visualized by immunofluoresecent staining, as described in Materials and Methods. Dapi was used to visualize nuclei. Estrogen (100 nM), OHT or ICI (100 nM) were added as indicated.

Figure 3

In vitro interaction between ER and ZNF366. (A and B) Schematic representations of ERα, ZNF366 and deletion mutants are shown. All N-terminal ZNF366 fusion proteins started at amino acid 9, thereby avoiding potential internal translation start sites. (C and D) GST binding assays were carried out by incubation of ³⁵S-labelled ERα or mutants with GST or GST–ZNF366 fusion proteins, in the absence of ligand (NL) or in the presence of 100 nM E2, OHT or ICI. E2 was present throughout in part (D). Input lanes represent 20% of the total volume of the in vitro translation reaction used in the binding assay.

Figure 4

ZNF366 is a repressor of ERα activity. (A) COS-1 cells were transfected with ERE-3-TATA-luc (100 ng), ERα (100 ng), ZNF366 and the RLTK renilla luciferase reporter (100 ng). Results represent the mean of three independent experiments individually corrected for transfection efficiency against renilla luciferase activity. Error bars represent the standard error of the mean. The activity for ERα in the presence of E2 and in the absence of ZNF366 was taken as 100%. All other activities are shown relative to this. The amounts of ZNF366 transfected were 0 ng (lanes 1, 7, 13 and 19), 0.1 ng (lanes 2, 8, 14 and 20), 1 ng (lanes 3, 9, 15 and 21), 10 ng (lanes 4, 10, 16 and 22), 30 ng (lanes 5, 11, 17 and 23) or 100 ng (lanes 6, 12, 18 and 24). (B and C) Reporter gene assays were performed following transfection of 100 ng ERα−ΔLBD (B) or ERα−ΔAF1 (C), as for (A). (D) COS-1 cells were co-transfected with 100 ng Gal4 DBD, the Gal4 DBD fused to full-length RIP140 or ZNF366 together with LexA-VP16 and the Lex-Gal-luc reporter gene. Relative reporter gene activities from three independent experiments are shown. (E) Shown are RT–PCR carried out using PCR primers for GREB1, TERT, ZNF366, Lamin A/C and GAPDH, using total RNA prepared from PE04 cells transfected with non-targeting control, Lamin A/C, or ZNF366 siRNA.

Figure 5

ZNF366 interacts with CtBP in vitro and in vivo. RIP140 fragments (A), ZNF366 fragments (B), CtBP1 (A and E) or ERα−ΔLBD (A) fused to GST, were immobilized on glutathione beads and incubated with ³⁵S-labelled ZNF366 (A and C), CtBP1 (A, B) or RIP140 (B). Bound proteins were eluted, resolved by 10% SDS–PAGE and exposed to autoradiography. (D) Mammalian 2-hybrid assay was performed by co-transfecting 100 ng Gal4 DBD or Gal4 fusions with ZNF366 558–744 amino acids, and 100 ng of VP16 or VP16 fused to CtBP. (E) COS-1 cells were co-transfected with LexA-VP16 and the Lex-Gal-luc reporter plasmid, together with the Gal4 DBD or Gal4 fused to 558–744 amino acids of ZNF366 in which the N-terminal (M1), the C-terminal (M2) or both (M1/M2) CtBP binding motifs have been mutated. Relative reporter gene activities from three independent experiments are shown. (F and G) Whole cell lysates prepared from COS-1 cells transiently transfected with CtBP1 and FLAG-tagged ZNF366 or FLAG-tagged ZNF366 mutated in the CtBP binding sites were immunoprecipitated using antibodies to CtBP1, followed by western blotting for CtBP1 and FLAG (F) or using the FLAG antibody followed by western blotting for FLAG and CtBP1 (G). In each case control immunoprecipitations were performed using mouse IgG.

Figure 6

Repression of ERα activity by ZNF366 involves HDACs. (A) COS-1 cells were transfected with ERE-3-TATA-luc (100 ng), ERα (100 ng) and ZNF366 (10 ng) or ZNF366 in which the CtBP sites have been mutated (ZNF366-Mut). E2 (10 nM) was present throughout. Results represent the mean of three independent experiments. (B) SAHA and E2 were added 24 h following transfection with ERα and ZNF366 and cells processed as above. GST pulldowns were carried out by incubation of ³⁵S-labelled HDAC 1, 3, 4 or 6, with GST-ZNF366 fusion proteins. The input lanes represent 20% of the total volume of the in vitro translation reaction used in the binding assay.

Figure 7

ZNF366 acts as a corepressor for endogenous estrogen-responsive genes. (A) MCF7 cells cultured in estrogen-free medium for 3 days were transfected with ZNF366 or control vector. E2 (10 nM) was added 24 h following transfection and lysates prepared after a further 24 h, were immunoblotted for cathepsin D (CTD), pS2, FLAG-ZNF366 and β-actin. (B and C) MCF7 and MDA-MB-231 cells were transfected as above and cell counts obtained 72 h after the addition of E2. The means of three experiments are shown, error bars representing the standard error of the mean.