Differential splicing alters subcellular localization of the alpha but not beta isoform of the MIER1 transcriptional regulator in breast cancer cells

Abstract

MIER1 was originally identified in a screen for novel fibroblast growth factor activated early response genes. The mier1 gene gives rise to multiple transcripts encoding protein isoforms that differ in their amino (N-) and carboxy (C-) termini. Much of the work to date has focused on the two C-terminal variants, MIER1α and β, both of which have been shown to function as transcriptional repressors. Our previous work revealed a dramatic shift in MIER1α subcellular localization from nuclear in normal breast tissue to cytoplasmic in invasive breast carcinoma, suggesting that loss of nuclear MIER1α may play a role in breast cancer development. In the present study, we investigated whether alternative splicing to include a cassette exon and produce an N-terminal variant of MIER1α affects its subcellular localization in MCF7 breast carcinoma cells. We demonstrate that this cassette exon, exon 3A, encodes a consensus leucine-rich nuclear export signal (NES). Inclusion of this exon in MIER1α to produce the MIER1-3Aα isoform altered its subcellular distribution in MCF7 cells from 81% nuclear to 2% nuclear and this change in localization was abrogated by mutation of critical leucines within the NES. Treatment with leptomycin B (LMB), an inhibitor of the nuclear export receptor CRM1, resulted in a significant increase in the percentage of cells with nuclear MIER1-3Aα, from 4% to 53%, demonstrating that cytoplasmic localization of this isoform was due to CRM1-dependent nuclear export. Inclusion of exon 3A in MIER1β to produce the N-terminal variant MIER1-3Aβ however had little effect on the nuclear targeting of this isoform. Our results demonstrate that alternative splicing to include exon 3A specifically affects the localization pattern of the α isoform.

Figure 1. Sequence encoded by exon 3A.

(A) Schematic illustrating the 5′end of the mier1 gene and the nucleotide & predicted amino acid sequences of exon 3A. The upper diagram shows the location of exon 3A relative to the 2 promoters, P1& P2. The lower diagram shows the sequence of exon 3A and illustrates alternative splicing to generate either mier1 (exons 2A+4; protein sequence (i)) or mier1-3A (exons 2A+3A+4; protein sequence (ii)). The MIER1 start codon, located at the end of exon 2A, is highlighted yellow, as is a downstream in-frame ‘atg’. For the analysis in (B), both highlighted ATGs were mutated to avoid the possibility of spurious translation initiation at the second ‘atg’; this double mutation changes the putative protein sequence from MLKM to QLKL. The predicted start methionine for MIER1-3A is highlighted in blue; the consensus NES is underlined and the hydrophobic residues are highlighted in green. For the analysis in (B), both the initiation and third codons were changed to produce an MFMF->IFIF mutant. (B) Effect of mutating the putative initiation codons on in vitro translation. ³⁵S-labelled rabbit reticulocyte lysates programmed with wild-type (lane 1) or either of the two mutant mier1-3A cDNAs (lanes 2–3) were immunoprecipitated with anti-MIER1 and analyzed by SDS-PAGE and fluorography; the position of full–length MIER1-3A is indicated by an arrowhead and the molecular weight standards are shown on the left. Mutation of the predicted initiation codon (MFMF->IFIF) located within exon 3A abrogates production of full-length MIER1-3A (lane 2), while mutation of upstream ATGs (MLKM->QLKL) has no effect (lane 3), thus confirming the N-terminal sequence of this isoform. Note that full-length MIER1 proteins migrate aberrantly on SDS-polyacrylamide gels, as has been reported for other proteins containing stretches of acidic residues , .

Figure 2. Subcellular localization of the MIER1α and MIER1-3Aα isoforms in MCF7 cells.

MCF7 cells were transfected with myc-tagged MIER1α, MIER1-3Aα or empty vector and analyzed by immunocytochemistry or immunoblotting with the 9E10 monoclonal antibody. (A) Western blot of extracts from MCF7 cells transfected with empty vector (lane 1), myc-tagged mier1α (lane 2) or myc-tagged mier1-3Aα (lane 3); staining was performed with the 9E10 antibody and confirms that a single protein in each cell extract is recognized by the antibody. The positions of the molecular weight standards are indicated on the left. (B) Histogram showing the results of 3 experiments; random fields were selected and the staining pattern of each cell within the field was scored visually according to the categories described in the Results & Discussion. 650–1,350 cells were scored for each construct. Plotted is the percentage of cells in each category ± S.D. (C) Illustrative examples of the observed staining pattern. Panels (i) & (ii) show brightfield (BF) and the corresponding phase contrast (PhC) views of a staining control, prepared without primary antibody. Panel (iii) shows cells expressing the myc-tag alone; examples of whole cell staining are indicated by arrowheads. Panels (iv) & (v) show cells expressing MIER1-3Aα & MIER1α, respectively; note the absence of nuclear staining in panel (iv) while nuclei in panel (v) are intensely stained (arrows). Scale bar = 50 µm for (i)–(iii) and 25 µm for (iv)–(v).

Figure 3. Function of exon 3A sequence in nuclear export.

(A) Effect of leptomycin B on localization of MIER1α and MIER1-3Aα. Cells were transfected, treated with 5 ng/ml LMB for 24 h and analyzed by confocal microscopy, using DAPI, 9E10 and a DyLight-488 secondary antibody . Histogram showing the results of 2 experiments; the staining pattern from random fields was scored visually according to the categories described in the Results & Discussion. Plotted is the percentage of cells in each category ± S.D; 75–90 cells were scored for each construct with each treatment. Note the increase in nuclear localization in treated cells expressing MIER1-3Aα. (B) and (C) Mutation of the NES consensus increases nuclear localization. MCF7 cells were transfected with plasmids encoding myc-tagged MIER1α, MIER1-3Aα or MIER1-3Aα containing a double mutation in the NES consensus (NES mutant) and analyzed by confocal microscopy using the antibodies described in (A). (B) Illustrative examples of cells expressing MIER1α (a–c), MIER1-3Aα (d–f) or the NES mutant (g–i); arrowheads indicate nuclei. (C) Histogram showing the results of 2 experiments; the staining pattern was scored as in (A). Plotted is the percentage of cells in each category ± S.D; 85–130 cells were scored for each construct.

Figure 4. Subcellular localization of the MIER1-3Aβ isoform in MCF7.

MCF7 cells were transfected with myc-tagged MIER1α, MIER1-3Aα, MIER1β, MIER1-3Aβ or empty vector and analyzed by confocal microscopy, performed as in the legend to Fig. 3. (A) Illustrative examples of cells expressing myc-tag alone (a–c), MIER1α (d–f), MIER1-3Aα (g–i), MIER1β (j–l) or MIER1-3Aβ (m–o). Arrowheads indicate nuclei and the arrow indicates staining in the cytoplasm. (B) Histogram showing the results of 4 independent experiments; the staining pattern was scored as in the legend to Fig. 3. Plotted is the percentage of cells in each category ± S.D; 55–160 cells were scored for each construct. Note that, unlike MIER1-3Aα, MIER1-3Aβ remains predominantly nuclear. (C) Bar graph showing the intracellular distribution of each construct. Pixel values for the nuclear and in the cytoplasmic areas were measured using Image J v1.46 and plotted as a proportion of the total signal. Shown is the proportion in each compartment, using measurements from 30–40 cells for each construct.