Cofactor of BRCA1 modulates androgen-dependent transcription and alternative splicing

Abstract

Transcriptional activity of nuclear receptors (NRs) is influenced by a large number of coregulators that exert their actions predominantly at the transcription initiation step. Unlike most well-characterized NR coregulators, cofactor of BRCA1 (COBRA1), a subunit of the negative elongation factor (NELF), binds to estrogen receptor alpha (ERalpha) and modulates estrogen-dependent transcription by impeding the movement of RNA polymerase II (RNAPII) during the transcription elongation stage. Here we show that, in addition to ERalpha, COBRA1 also displays various degrees of affinity for several other NRs. In particular, COBRA1 binds strongly to androgen receptor (AR) via its ligand-binding domain (LBD). Small hairpin RNA (shRNA)-mediated reduction of endogenous COBRA1 enhances androgen-mediated transcription. The effect of COBRA1 knockdown can be rescued by a silent mutant COBRA1 that is refractory to the shRNA action. Using a reporter assay for alternative splicing, we also provide evidence for a role of COBRA1 in influencing the exon skipping/inclusion of nascent transcripts produced from an androgen-dependent promoter. These findings suggest that COBRA1 may coordinate multiple steps in ligand-dependent gene expression, which in turn ensures both the quantity and quality of hormone-stimulated gene products.

FIGURE 1. Interaction between COBRA1 and multiple steroid hormone receptors

A. Empty or Flag-COBRA1-expressing vector was co-transfected with AR, PRB, or GR expression vectors into HEK293T cells. 24 hours after transfection, whole cell lysate was prepared and immunoprecipitated with an α-Flag antibody. Immunoprecipitates were resolved by SDS-PAGE and blotted with α-AR, α-PR, or α-GR antibodies. 2% of the whole cell lysate was loaded as input control. B. The protein levels of Flag-COBRA1 in the whole cell lysate were determined by immunoblotting with the α-Flag antibody. α-Tubulin was used as a loading control.

FIGURE 2. COBRA1 binds to the ligand-binding domain of AR

A. GST-pulldown assay was performed to detect the interaction between in vitro translated ³⁵S-labeled AR and bacterially expressed GST-COBRA1. 1/15 of the in vitro translated AR used for GST-pull down was loaded as an input control. B. Full-length COBRA1 protein and various AR fragments were co-expressed in HEK293T cells, and the binding affinity was compared by co-immunoprecipitation. The diagram summarizes the interaction of COBRA1 with various AR fragments. AF-1: activation function 1; D: DNA binding domain; H: Hinge region; LBD: Ligand binding domain. C. Full-length AR was co-expressed with various Flag-COBRA1 fragments in HEK293T cells, and their interactions were examined with co-immunoprecipitation. The diagram is a summary of the binding affinity of COBRA1 fragments for the full-length AR protein. The shaded and solid boxes represent the NR-binding signature motifs LXXLL and ΦXXΦΦ, respectively (L for leucine; X for any amino acid; and Φ for hydrophobic residue).

FIGURE 3. Reduction of COBRA1 in breast cancer cells results in elevated transcriptional activation by AR

A. The inlet shows an α-COBRA1 immunoblot in T47D-derived cell lines that stably expressed short hairpin RNAs against EGFP control (shEGFP) or COBRA1 (shCOBRA1-2A and shCOBRA1-GA). α-Tubulin was used as a loading control. MMTV-luciferase reporter plasmid was transiently transfected in the control or two COBRA1 knockdown cell lines. The cells were treated with DHT (10 nM; open bars), R5020 (10 nM; shaded bars), or dexamethasone (Dex, 100 nM; solid bars) for 24 hours before harvesting for luciferase assay. Ethanol was used as vehicle. Steroid hormone-dependent transcription was expressed as fold increase over vehicle. B. A COBRA1 silent mutant (COBRA1) that is refractory to shCOBRA1-2A was co-expressed with the MMTV-luciferase reporter plasmid in stable knockdown cells expressing shCOBRA-2A. Transfected cells were treated with DHT (10 nM) for 24 hours before harvesting for luciferase assay. C. DHT (10 nM)–dependent transcription from the prostate-specific antigen (PSA) promoter in the control and two COBRA1 knockdown cells as measured by luciferase assay.

FIGURE 4. COBRA1 does not affect the ligand sensitivity of AR-mediated transcriptional activation

The control and COBRA1 knockdown cells were transfected with the MMTV-luciferase reporter plasmid, and subsequently treated with a serial 10-fold dilution of DHT, ranging from 10^-11 M to 10^-8 M. Transfected cells were also treated with AR antagonist flutamide (10^-6 M) in the presence of 10^-9 M DHT to ascertain the androgen-dependency of the observed luciferase expression.

FIGURE 5. Over-expression of COBRA1 reduces the skipping/inclusion ratio of androgen-dependent alternative splicing

COBRA1 and AR were co-expressed with the MMTV-CD44 mini-gene cassette in HEK293T cells. The transfected cells were treated with 1 nM R1881 for 24 hours before harvesting for alternative splicing assay. Two major products of CD44 mini-gene resulted from alternative splicing were designated as inclusion and skipping, respectively. The minor splicing product with one exon inclusion is also indicated in the figure, but not considered in the calculation of the skipping/inclusion ratio. The skipping/inclusion ratio was calculated by averaging two independent transfections in each assay. The diagram indicates various splicing events. The open and solid boxes in the diagram refer to the variable and invariable exons, respecitively. The arrows indicate the positions of the primers used in the PCR reactions.

FIGURE 6. COBRA1 knockdown increases the skipping/inclusion ratio of the alternative splicing

The MMTV-CD44 mini-gene were transfected into T47D derived stable cell lines expressing EGFP (control) or COBRA1-specific shRNAs (2A and GA). Cells were treated with DHT (10 nM) for 24 hours before harvesting for alternative splicing assay. Quantitation of the results as shown in the graph was performed in the same way as shown in Fig. 5.