The Mammalian Ecdysoneless Protein Interacts with RNA Helicase DDX39A To Regulate Nuclear mRNA Export

Abstract

The mammalian orthologue of ecdysoneless (ECD) protein is required for embryogenesis, cell cycle progression, and mitigation of endoplasmic reticulum stress. Here, we identified key components of the mRNA export complexes as binding partners of ECD and characterized the functional interaction of ECD with key mRNA export-related DEAD BOX protein helicase DDX39A. We find that ECD is involved in RNA export through its interaction with DDX39A. ECD knockdown (KD) blocks mRNA export from the nucleus to the cytoplasm, which is rescued by expression of full-length ECD but not an ECD mutant that is defective in interaction with DDX39A. We have previously shown that ECD protein is overexpressed in ErbB2⁺ breast cancers (BC). In this study, we extended the analyses to two publicly available BC mRNA The Cancer Genome Atlas (TCGA) and Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) data sets. In both data sets, ECD mRNA overexpression correlated with short patient survival, specifically ErbB2⁺ BC. In the METABRIC data set, ECD overexpression also correlated with poor patient survival in triple-negative breast cancer (TNBC). Furthermore, ECD KD in ErbB2⁺ BC cells led to a decrease in ErbB2 mRNA level due to a block in its nuclear export and was associated with impairment of oncogenic traits. These findings provide novel mechanistic insight into the physiological and pathological functions of ECD.

Keywords: ECD; ErbB2; RNA export; RNA helicase; RNA processing; RNA splicing; ecdysoneless; hSGT1; oncogenesis.

FIG 1

ECD interacts with DDX39A and other components of the mRNA export machinery. (A to E) HEK-293T (A to D) or MCF10A (E) cell lysates were immunoprecipitated (IP) with the antibodies indicated at the top, followed by Western blotting (WB) with the antibodies shown on the left. ALY and CRM1 were used as positive controls. CBL, mouse IgG (mIgG), and rabbit IgG (rIgG) were used as negative controls; 100-μg aliquots of lysate protein were used in the input lane. (F) Twenty micrograms of GST (negative control) or GST fusion with full-length ECD (1 to 644 aa) was incubated with 1 mg of protein lysate from HEK-293T cells transfected with FLAG-tagged RUVBL1 or FLAG-tagged DDX39A, and the GST pulldown proteins were analyzed by Western blotting with the indicated antibodies. The membrane was stained with Ponceau S to visualize the GST fusion proteins to assess comparable fusion protein use for pulldowns (indicated by arrows). (G) GST or GST fusion with full-length ECD (1 to 644) or indicated ECD mutants were incubated with protein lysate of HEK-293T cells transiently transfected with FLAG-tagged-DDX39A and FLAG-tagged RUVBL1, and the bound proteins were analyzed by Western blotting with the indicated antibodies. The membrane was stained with Ponceau S to visualize the GST fusion proteins (indicated by arrows). The experiment shown is a representative of at least three repeats with comparable results. (H) To validate ECD commercial antibody, Western blotting was performed in various cell lysates from control and ECD siRNA-treated MCF10A, 76NTERT, MDA-MB-231 and HeLa cells and ECD-overexpressing MCF10A stable as well as DOX-inducible (as indicated); glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as loading control. (I) ECD colocalizes with DDX39A in the nucleus. MCF10A cells untreated or treated with 20 ng/μl of leptomycin B (LMB) for 4 h and fixed in 4% paraformaldehyde (PFA) were subjected to immunofluorescence staining using anti-ECD rabbit polyclonal and anti-DDX39 mouse monoclonal antibodies followed by imaging using 63× Zeiss confocal microscope (scale bars, 10 μm).

FIG 2

RNA-independent interaction of ECD with DDX39A. (A to D) MCF10A cell lysates treated with Benzonase/RNase were immunoprecipitated (IP) with the antibodies indicated at the top, followed by Western blotting (WB) with the antibodies shown on the left. TERT and dyskerin were used as positive controls. Mouse IgG (mIgG) and rabbit IgG (rIgG) were used as negative controls; 2% aliquots of lysate protein were used in the input lanes. (E) Phosphorylation-independent interaction of ECD with DDX39A. 293T cells were transfected with FLAG-tagged WT (wild-type) ECD or FLAG-tagged ECD phosphorylation site mutants 3S/A (503S/A, 505S/A, and 518S/A) and 6S/A (503S/A, 505S/A, 518S/A, 572S/A, 579S/A, and 584S/A) or untreated (nontransfected cells [NTC]). Cell lysates were immunoprecipitated (IP) with anti-FLAG antibody coupled to M2 FLAG-agarose beads (Sigma-Aldrich), followed by Western blotting (WB) with the antibodies shown on the left; 2% aliquots of lysate protein were used in the input lanes. Blue arrow indicates the DDX39A band at higher exposure.

FIG 3

ECD KD decreases poly(A) mRNA export from the nucleus to the cytoplasm. (A to F) RNA FISH analysis in MCF10A (A to C) and 76NTERT (D to F) cells with control siRNA (Ctl; as a negative control), two independent siECDs (number 1 and number 2), or siDDX39A (as a positive control). (A and D) WB shows ECD or DDX39A KD. (B and E) Cells were fixed and hybridized with 12 μM oligo(dT) 22-nt Quasar 570 probe and then imaged using a LSM 710 Zeiss confocal microscope at ×63 magnification. Last column shows enlarged view of cells shown in the insets. (C and F) Quantification of nucleus-to-cytoplasmic signal ratio (N/C) was conducted using ImageJ software in at least 100 cells, and graphs were plotted by normalizing to the control cells to calculate the fold change. (G to I) ECD^fl/fl cells were either treated with GFP (control) or Cre-GFP adenovirus to delete ECD. (G) WB shows Cre-mediated knockdown of ECD; β-actin was used as a loading control. (H) Cells were fixed and imaged as discussed above. (I) Quantifications of at least 25 cells using ImageJ to measure nucleus-to-cytoplasmic signal ratio (N/C) by normalizing to the control cells to calculate the fold change. Significance of the ratios was calculated using Student’s two-tailed t test; *, P ≤ 0.05, error bars represent standard errors from the mean (SEs). Means ± SEs were derived from three different experiments. Bars, 10 μm.

FIG 4

ECD interaction with DDX39A is required for mRNA export. RNA FISH analysis was performed on Ecd^fl/fl MEFs overexpressing either vector, ECD, DDX39A, or ECD-P617G after treatment with GFP (control) or Cre-GFP adenovirus to delete ECD. (A) WB shows Cre-mediated knockdown of ECD. Exogenous ECD indicated by red marks. β-Actin was used as a loading control. (B) Cells were fixed and hybridized with 12 μM oligo(dT) 22-nt Quasar 570 probe and then imaged at ×63 magnification. Bars, 10 μm. (C) Quantifications of nucleus-to-cytoplasmic signal ratios were conducted using ImageJ software in at least 75 cells, and graphs were plotted by normalizing to the control cells. Significance of the ratios was calculated using Student’s two-tailed t test between groups as indicated, error bars represent standard errors from the means. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001.

FIG 5

ECD mRNA analyses in breast cancer The Cancer Genome Atlas (TCGA) and the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) data sets reveal ECD mRNA overexpression correlates with poor patient survival. (A) TCGA breast cancer data set comprising a total of 1,089 breast cancer cases, including 676 cases of ER⁺/PR⁺, 164 cases of HER2⁺, and 112 adjacent normal tissue samples, was analyzed for ECD mRNA expression. RNA-seq expression level read counts were normalized using the upper quartile fragments per kilobase to transcript per million mapped reads (FPKM-UQ) calculation, and ECD expression in box plots is represented as log base 2 compared to that in adjacent normal tissue. In TCGA cohorts, ECD mRNA expression is significantly higher in all breast cancer tissues than in adjacent normal breast tissues (P = 2.7e−08) (A), in ER⁺/PR⁺ breast cancer samples (P = 2.7e−12) (B), and in HER2⁺ breast cancer cases (P = 3e−04) (C). (D) The 10-year overall survival of 854 cases was significantly worse in ECD-high (238) than in ECD-low (616) mRNA-expressing patients. High ECD mRNA expression in PR⁻ subgroup (E) and HER2⁺ (F) subgroup of TCGA data set correlates with poor survival. (G) METABRIC cohort: Kaplan-Meier plot for survival curve of all the breast cancer patients shows poor prognosis with high ECD mRNA, including in the ER⁻ subgroup (H), PR⁻ subgroup (I), triple-negative breast cancer (TNBC) subgroup (J), and HER2⁺ subgroup (K). Cox regression analysis results are shown with the Kaplan-Meier plots.

FIG 6

ECD KD decreases ErbB2 protein and mRNA expression in breast cancer cells. ECD knockdown was performed using specific siRNA against ECD in BT-474, SKBR3 (ErbB2-overexpressing), and MDA-MB-468 (EGFR-overexpressing) breast cancer cell lines. Protein and total mRNA were isolated after 48 h of transfection. (A) Lysates were harvested and immunoblotted with the indicated antibodies. β-Actin was used as a loading control. (B) Control mRNA levels of ECD, ErbB2, or EGFR were detected by qRT-PCR; DHFR was used as a positive control, and GAPDH served as an internal control. Fold change over GAPDH was calculated and plotted and normalized with control. (C) MCF-7-vector and MCF-7-ErbB2 expressing cells were treated with ECD siRNA or control siRNA, and lysates were collected and blotted with indicated antibodies. (D) RNA was isolated, cDNA was prepared, and qRT-PCR was performed using ErbB2-specific primers. Fold change with respect to control after normalizing with β-actin was calculated and plotted. ECD downregulation decreases ErbB2 mRNA stability. (E and F) The stability of mRNA was analyzed in SKBR3 and BT-474 breast cancer cell lines. After ECD downregulation by siRNA, cells were treated with actinomycin D (5 μg/ml) at zero time point, and then total RNA was isolated at various time points (0, 2, 4, 6, 8 and 12 h) after actinomycin D treatment. cDNA was made, and levels of ErbB2 were measured by qRT-PCR. GAPDH was used as normalization control. (G) qRT-PCR using specific primers of indicated genes from RNA samples of SKBR3 cells treated with control and ECD siRNA. ECD knockdown was confirmed. HER2 (full form) and its isoforms, herstatin and Δ16HER2, were analyzed using specific sets of primers. β-Actin was used as an internal control. Fold change with respect to control after normalizing with β-actin was calculated and plotted. Significance of the ratios was calculated using Student’s two-tailed t test. ns, not significant; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001. Error bars represent means ± standard deviations (SDs) from six replicates in a representative experiment. The experiment was repeated three times independently.

FIG 7

ECD KD increases nuclear ErbB2 mRNA accumulation. SKBR3 and BT-474 cells were transfected with control siRNA, ECD siRNA, and DDX39A siRNA. (A) Western blotting was performed with indicated antibodies; β-actin was used as a loading control, confirming knockdown of ECD and DDX39A in SKBR3. (B and D) qRT-PCR shows decrease in ErbB2 mRNA in ECD or DDX39A KD cells in comparison to control siRNA-treated cells. The graph plotted shows ECD, ErbB2 and DDX39A mRNA levels normalized to β-actin, and fold change was calculated by normalizing to control in SKBR3 (B) and in BT-474 (D) cells. SKBR3 cells were subjected to subcellular RNA fractionation after treating the cells with control siRNA, ECD siRNA, and DDX39A siRNA. (C and E) Subcellular RNA fractionation after control siRNA, ECD siRNA, or DDX39A siRNA treatment followed by qRT-PCR of ErbB2 mRNA showing nuclear/cytoplasmic ratio (N/C) in SKBR3 (C) and BT-474 (E) cells. (F) qRT-PCR of MALAT1 (a long noncoding RNA) used as control for purification of nuclear fraction. (G) qRT-PCR showing nucleus-to-cytoplasmic ratio of 18S rRNA, U1 snRNA, and ErbB2 and ECD mRNA in control siRNA- and ECD siRNA-treated cells; each cytoplasmic and nuclear fraction was normalized to corresponding total fraction. Significance was calculated using Student’s two-tailed t test. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.00; ns, nonsignificant, error bars represent standard errors from the means. Means ± SEs were derived from three different experiments. ECD knockdown was performed using specific siRNA against ECD in SKBR3 cells. (H) RNA FISH analysis was carried out with 12 μM oligonucleotide anti-ErbB2 570 probe and then imaged at ×63 magnification; last column shows enlarged view of cells. White arrows show the accumulation of ErbB2 mRNA probe. (I) Quantifications of at least 25 cells using ImageJ showing nucleus-to-cytoplasmic ratio (N/C) by normalizing to the control cells to calculate the fold change. Significance of the ratios was calculated using Student’s two-tailed t test. *, P ≤ 0.05, error bars represent standard errors from the means. Means ± SEs were derived from three different experiments.

FIG 8

ECD knockdown decreases cell proliferation and anchorage independent of ErbB2-overexpressing breast cancer cells. (A) ECD was stably knockdown by two different shRNAs against ECD, and then KD was assessed by Western blotting. (B and C) Cell proliferation assays were performed using CellTiter-Glo luminescent cell viability assay; 2,000 cells per well were plated (as discussed in Materials and Methods). Readings were taken at the indicated days and log₁₀ transformed to meet ANOVA assumptions. Graphs were plotted based on the mean of log₁₀-transformed readings for each group at each time point. The corresponding standard errors were very small and not plotted. ANOVA indicated that mean log₁₀-transformed readings of the control group were higher than those of both ECD shRNA no. 1 and ECD shRNA no. 2 at days 6, 8, and 10 (all Tukey’s adjusted P values less than 0.0001). (D) For colony formation assay, 10,000 cells were plated in 6-well plate, and colonies were fixed and then stained with crystal violet after 10 days of plating. (E and F) Colonies were counted and presented as histograms for BT-474 (E) and SKBR3 (F) cells. ANOVA indicated that the mean number of colonies of the control group was higher than those of both ECD shRNA no. 1 and ECD shRNA no. 2 (all Tukey’s adjusted P values less than 0.001). (G and H) Soft agar colony formation assay was used to measure anchorage dependence; 20,000 cells were plated in 0.3% agarose in 6-well plates for 21 days (described in detail in Materials and Methods), and then colonies were stained with 0.05% crystal violet. Colonies were counted and plotted as histograms for BT-474 (G) and SKBR3 (H) cells. ANOVA indicated that the mean number of colonies of control in soft agar was higher than those of both ECD shRNA no. 1 and ECD shRNA no. 2 (all Tukey-adjusted P values less than 0.001). *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ns, nonsignificant. Means ± SEs were derived from three different experiments, each performed in triplicates.

FIG 9

Knockdown of ECD decreases invasion and migration ability of breast cancer cells. BT-474, an ErbB2-positive breast cancer cell line, was treated with control or two independent siRNAs against ECD. (A) Western blotting shows knockdown of ECD with two independent siRNAs. (B and C) Ten thousand cells were counted and plated on Boyden chambers for assessing their ability to migrate (B) and invade (C). After 24 h, cells that had invaded through Matrigel (for invasion assay) and migrated to the bottom surface were fixed and stained with propidium iodide. Pictures were taken, cells were counted, and histograms were plotted. The bar diagrams represent numbers of cells migrated or invaded. ANOVA indicated that the mean number of cells migrated for the control group was higher than those for both ECD siRNA no. 1 and ECD siRNA no. 2 (all Tukey’s adjusted P values less than 0.001). *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ns, nonsignificant. Means ± SEs were derived from three different experiments, each performed in triplicates.