Human MI-ER1 alpha and beta function as transcriptional repressors by recruitment of histone deacetylase 1 to their conserved ELM2 domain

Abstract

mi-er1 (previously called er1) was first isolated from Xenopus laevis embryonic cells as a novel fibroblast growth factor-regulated immediate-early gene. Xmi-er1 was shown to encode a nuclear protein with an N-terminal acidic transcription activation domain. The human orthologue of mi-er1 (hmi-er1) displays 91% similarity to the Xenopus sequence at the amino acid level and was shown to be upregulated in breast carcinoma cell lines and tumors. Alternative splicing at the 3' end of hmi-er1 produces two major isoforms, hMI-ER1alpha and hMI-ER1beta, which contain distinct C-terminal domains. In this study, we investigated the role of hMI-ER1alpha and hMI-ER1beta in the regulation of transcription. Using fusion proteins of hMI-ER1alpha or hMI-ER1beta tethered to the GAL4 DNA binding domain, we show that both isoforms, when recruited to the G5tkCAT minimal promoter, function to repress transcription. We demonstrate that this repressor activity is due to interaction and recruitment of a trichostatin A-sensitive histone deacetylase 1 (HDAC1). Furthermore, deletion analysis revealed that recruitment of HDAC1 to hMI-ER1alpha and hMI-ER1beta occurs through their common ELM2 domain. The ELM2 domain was first described in the Caenorhabditis elegans Egl-27 protein and is present in a number of SANT domain-containing transcription factors. This is the first report of a function for the ELM2 domain, highlighting its role in the regulation of transcription.

FIG. 1.

Human MI-ER1α and MI-ER1β function as transcriptional repressors in vivo. Cells were transfected with the G5tkCAT reporter plasmid alone or with the pM plasmid containing an NLS and the GAL4 DBD (GAL4) alone or fused to hmi-er1α (GAL4-α) or hmi-er1β (GAL4-β); the amount of plasmid (in micrograms) used for transfection is indicated below each bar. Cells were harvested 48 h after transfection, and the amount of CAT protein (in nanograms per 100 μg of cellular protein) was determined as described in Materials and Methods. Expression values for all constructs were normalized to the CAT expression level obtained with G5tkCAT alone (relative CAT expression). Shown are the average values and standard deviations from at least three independent experiments. (A) Relative CAT expression in HeLa cells transfected with increasing amounts of GAL4-α or GAL4-β plasmid; in each case, the amount of GAL4 empty vector was adjusted so that the total amount of DNA used in each transfection was constant. The amount of GAL4-α or GAL4-β protein expressed in each sample was determined by Western blotting using an anti-GAL4 antibody. A representative blot is shown. (B) Relative CAT expression in C33A, HEK 293 (293K), and NIH 3T3 cells transfected with the G5tkCAT reporter plasmid alone or cotransfected with the indicated GAL4 construct. Shown are the average values and standard deviations from three independent experiments.

FIG. 2.

Repression by hMI-ER1α and hMI-ER1β occurs through a HDAC-dependent mechanism. (A) HeLa cells were cotransfected with 0.8 μg of the G5tkCAT reporter plasmid and 0.8 μg of the GAL4, GAL4-α, or GAL4-β plasmid and cultured in the presence or absence of TSA. Cells were harvested 48 h after transfection, and the amount of CAT protein (in nanograms per 100 μg of cellular protein) was determined as described in Materials and Methods. The values for GAL4-α- and GAL4-β-transfected cells are presented as a proportion of the value obtained with the GAL4 empty vector (relative CAT expression). Shown are the average values and standard deviations from three independent experiments. (B) ³⁵S-labeled TNT mixtures programmed with cDNA encoding luciferase (Luc), Myc-α (α), or Myc-β (β) were loaded directly on the gel (lanes 1 to 3) or incubated with unlabeled TNT mixtures programmed with hdac1 cDNA and subjected to immunoprecipitation (IP) with anti-HDAC1 (lanes 4 to 6) as described in Materials and Methods. Proteins were visualized by SDS-PAGE and autoradiography. (C) ³⁵S-labeled TNT mixtures programmed with cDNA encoding luciferase or HDAC1 were loaded directly on the gel (lanes 1 and 2) or incubated with unlabeled TNT mixtures programmed with cDNA encoding Myc-α (lanes 3 and 4) or Myc-β (lanes 5 and 6) and subjected to IP with anti-Myc; proteins were visualized as described for panel B. In panels B and C, the positions of luciferase, HDAC1, Myc-α, and Myc-β proteins are indicated.

FIG. 3.

A functional HDAC1 coimmunoprecipitates with hMI-ER1α and hMI-ER1β in vivo. (A) Cell lysates from HeLa cells transiently transfected with myc tag empty vector or myc-α or myc-β plasmid were prepared, and equivalent amounts of protein from each sample were either added directly to sample buffer (lanes 1 to 3) or subjected to immunoprecipitation (IP) with anti-HDAC1 (lanes 4 to 6); Western blot (WB) analysis was performed using anti-Myc. The positions of the Myc-α and Myc-β proteins are indicated. (B) HeLa cells (1.5 × 10⁵ cells per sample) were transfected with myc tag empty vector or myc-α or myc-β plasmid and lysed, and the supernatants were subjected to IP with anti-Myc. Additional controls consisted of mock-transfected HeLa cell extracts immunoprecipitated with anti-HDAC1 or anti-Myc. Immunoprecipitates were assayed for HDAC activity in the presence or absence of 300 nM TSA as described in Materials and Methods. Shown are the average values and standard deviations from three independent experiments. (C) HeLa cell lysates were subjected to IP with nonimmune serum (lane 1) or anti-pan hMI-ER1 (lane 2), anti-hMI-ER1β-specific (lane 3), or anti-hMI-ER1α-specific (lane 4) antiserum. HDAC1 protein from an in vitro TNT mixture was loaded in lane 5. Western blot (WB) analysis was performed using anti-HDAC1. The position of the HDAC1 protein is indicated.

FIG. 4.

hMI-ER1 associates with HDAC activity through a region containing the ELM2 domain. (A) Deletion mutants of hMI-ER1α or β fused to GAL4 were transfected into HeLa cells (1.5 × 10⁵ cells per sample). The schematic on the left illustrates the constructs used and shows a scaled representation of the hMI-ER1 sequence and each of its domains. The individual domains are identified in the legend below the schematic, and the hMI-ER1 amino acid residues encoded by each construct are listed on the left. Cell extracts were prepared 48 h after transfection and subjected to immunoprecipitation with anti-GAL4. Immunoprecipitates were assayed for HDAC activity in the presence or absence of 300 nM TSA as described in Materials and Methods. The histogram shows the average values and standard deviations from three independent experiments. (B) The expression of the GAL4-hMI-ER1 fusion protein in each sample used in panel A was examined by Western blotting using an anti-GAL4 antibody. Indicated above each lane are the hMI-ER1 residues encoded by the construct.

FIG. 5.

The ELM2 domain of hMI-ER1 can recruit HDAC1 activity and repress transcription in vivo. (A) Deletion mutants of hMI-ER1(163-283) fused to GAL4 were transfected into HeLa cells for HDAC activity measurements or were cotransfected with the G5tkCAT reporter plasmid for transcriptional repression assays. The schematic illustrates the constructs used, and the hMI-ER1 amino acid residues encoded by each construct are listed on the left. ²¹³W→A and ²²⁶FL→AA constructs contain residues 163 to 283 with alanine substitutions at ²¹³W and ²²⁶FL, respectively. HDAC activity measurements were performed as described in the legend to Fig. 3B, and the average values and standard deviations from three independent experiments are shown. Repression was determined by measuring CAT expression levels as described in the legend to Fig. 2. (B) The expression of the GAL4-hMI-ER1 fusion protein in each sample used in panel A was examined by Western blotting using an anti-GAL4 antibody. Indicated above each lane are the hMI-ER1 residues encoded by each construct. (C) ³⁵S-labeled TNT mixtures programmed with cDNA encoding HDAC1 were loaded directly on the gel (Input; lane 1) or incubated with unlabeled TNT mixtures programmed with Myc-tagged constructs of the regions listed in panel A and then subjected to immunoprecipitation using anti-Myc (lanes 2 to 10). Proteins were visualized by SDS-PAGE and autoradiography. The position of the HDAC1 protein is indicated.

FIG. 6.

Alignment of ELM2 domains reveals additional conserved sequence. The ELM2 regions of proteins from the Pfam and GenBank databases were aligned using ClustalW. Shown is the amino acid sequence from the C-terminal end of the ELM2 domain to the beginning of the SANT domain in each protein. Residues belonging to these two domains are shaded. Highly conserved residues in the region C terminal to the ELM2 domain are shown with white lettering and highlighted in black. The consensus sequence is listed below the alignment; X represents any amino acid; φ represents Y, F, or H; and + represents a charged residue. The numbers listed above the alignment correspond to amino acid positions in the hMI-ER1 protein sequence. The accession numbers for the sequences used in this alignment (from top to bottom) are as follows: AF515447, O42194, AB033019, XM_125783, Q9UKL0, XM_127140, XM_127140, AJ311849, Q9JMK4, Q9R190, Q13330, O94776, Q9VNF6, Q9VNF6, Q9NHX6, Q9P2R6, and Q09228.