- Email: [email protected]
The BARD1 BRCT domain contributes to p53 binding, cytoplasmic and mitochondrial localization, and apoptotic function
Cellular Signalling 27 (2015) 1763–1771
Contents lists available at ScienceDirect
Cellular Signalling journal homepage: www.elsevier.com/locate/cellsig
The BARD1 BRCT domain contributes to p53 binding, cytoplasmic and mitochondrial localization, and apoptotic function Varsha Tembe a,1, Estefania Martino-Echarri a,1, Kamila A. Marzec a,1, Myth T.S. Mok a,2, Kirsty M. Brodie a, Kate Mills a, Ying Lei a, Anna DeFazio a, Helen Rizos a,3, Emma Kettle b, Ross Boadle b, Beric R. Henderson a,⁎ a b
Centre for Cancer Research, The Westmead Millennium Institute for Medical Research, The University of Sydney, Westmead, NSW 2145, Australia Electron Microscopy Unit, Institute for Clinical Pathology and Medical Research, Westmead Hospital, Westmead, NSW 2145, Australia
a r t i c l e
i n f o
Article history: Received 7 May 2015 Accepted 15 May 2015 Available online 3 June 2015 Keywords: BARD1 p53 Mitochondria Apoptosis BCL-2
a b s t r a c t BARD1 is a breast cancer tumor suppressor with multiple domains and functions. BARD1 comprises a tandem BRCT domain at the C-terminus, and this sequence has been reported to target BARD1 to distinct subcellular locations such as nuclear DNA breakage sites and the centrosome through binding to regulatory proteins such as HP1 and OLA1, respectively. We now identify the BRCT domain as a binding site for p53. We first confirmed previous reports that endogenous BARD1 binds to p53 by immunoprecipitation assay, and further show that BARD1/ p53 complexes locate at mitochondria suggesting a cellular location for p53 regulation of BARD1 apoptotic activity. We used a proximity ligation assay to map three distinct p53 binding sequences in human BARD1, ranging from weak (425–525) and modest (525–567) to strong (551–777 comprising BRCT domains). Deletion of the BRCT sequence caused major defects in the ability of BARD1 to (1) bind p53, (2) localize to the cytoplasm and mitochondria, and (3) induce Bax oligomerization and apoptosis. Our data suggest that BARD1 can move to mitochondria independent of p53, but subsequently associates with p53 to induce Bax clustering in part by decreasing mitochondrial Bcl-2 levels. We therefore identify a role for the BRCT domain in stimulating BARD1 nuclear export and mitochondrial localization, and in assembling mitochondrial BARD1/p53 complexes to regulate specific activities such as apoptotic function. © 2015 Elsevier Inc. All rights reserved.
1. Introduction The BRCA1-associated RING domain 1 (BARD1) protein is a putative tumor suppressor and a key binding partner of the breast and ovarian cancer susceptibility protein, BRCA1 [1,2]. The BARD1 gene is mutated in a subset of breast and ovarian cancer patients [3–5], and was first discovered in a genetic screen for BRCA1-interacting proteins [6]. BARD1 was later confirmed as the major in vivo partner for BRCA1 [1], and the two proteins interact through their N-terminal tail sequences [7]. The formation of a stable BRCA1/BARD1 dimer has a positive influence on BARD1 stability [8], nuclear localization [9] and DNA repair function Abbreviations: BRCA1, breast cancer regulatory protein-1; BARD1, BRCA1-associated RING domain 1; CMX-Ros, chloromethyl-X-rosamine; FCS, fetal calf serum; mtHSP70, mitochondrial heat shock protein 70; NES, nuclear export sequence(s); NLS, nuclear localization sequence; OMM, outer mitochondrial membrane; PBS, phosphate-buffered saline; PCNA, proliferating cell nuclear antigen; YFP, yellow fluorescence protein. ⁎ Corresponding author at: Westmead Millennium Institute, 176 Hawkesbury Road (PO Box 412), Westmead, NSW 2145, Australia. Tel.: +61 2 86273730. E-mail address: [email protected] (B.R. Henderson). 1 These authors contributed equally to this work. 2 Current address: School of Biomedical Sciences, Faculty of Medicine, The Chinese University of Hong Kong. 3 Current address: School of Advanced Medicine, Macquarie University, NSW 2109, Australia.
http://dx.doi.org/10.1016/j.cellsig.2015.05.011 0898-6568/© 2015 Elsevier Inc. All rights reserved.
[10,11]. Furthermore, the BRCA1/BARD1 dimer acts as an E3 ubiquitin (Ub) ligase whose enzymatic activity is impaired by cancer associated gene mutations in the BRCA1 RING domain [12,13]. BRCA1/BARD1 Ub ligase activity is thought to contribute to the regulation of DNA repair [11], mitotic cell division [14] and centrosome duplication [15,16]. The BRCA1/BARD1 heterodimer is also important for cell viability, and loss of either BRCA1 or BARD1 in mice leads to chromosomal abnormalities and early embryonic death due to severe cell proliferation defects [11, 17]. Human BARD1 comprises 777 amino acids and, like BRCA1, has an N-terminal RING finger domain and two C-terminal BRCT domains [18]. Unlike BRCA1, BARD1 contains a centrally located ankyrin repeat domain known to mediate protein–protein interactions [19]; the role of this domain in BARD1 is still under investigation although it was reported to contribute to association with p53 in mouse [20]. BARD1 can shuttle between the nucleus and cytoplasm [21,22], and it was previously proposed that while nuclear localization of BARD1 correlated with its role in DNA repair and cell survival [21,23], the cytoplasmic accumulation of BARD1 at mitochondria correlated with induction of apoptosis [24]. BARD1 is transcriptionally up-regulated in response to genotoxic stress in rodent cells [25]. Furthermore, the overexpression of exogenous BARD1 leads to apoptosis and this is partly dependent on
1764
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
functional p53, but is independent of and inhibited by BRCA1 as demonstrated in p53 or BRCA1-deficient cell lines [25,26]. We previously discovered a pool of BARD1 at mitochondria whose expression correlated with an increase in Bax oligomerization, mitochondrial membrane permeability and apoptosis [24]. Given that BARD1 apoptotic activity is p53-dependent, and that p53 itself is known to translocate to mitochondria and induce apoptosis [27], we hypothesized that p53 might regulate BARD1 apoptotic function by either (i) recruiting it to mitochondria, (ii) binding BARD1 at mitochondria and/or (iii) modulating its ability to stimulate Bax oligomerization. In this study we identify a key role for the BRCT domain in p53 binding, recruitment to the cytoplasm/mitochondria and apoptotic function. These data reveal that while p53 is not required for targeting of BARD1 to mitochondria, it does form complexes with BARD1 at this organelle and stimulates its role in Bax oligomerization.
2. Materials and methods 2.1. Cell culture and transfections Human MCF-7 breast cancer cells were grown under standard tissue culture conditions in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS). The cell line was mycoplasmanegative. Transient transfections were performed using Lipofectamine2000 (Life Technologies, Vic, Australia) according to manufacturer instructions. Briefly, at 24 h after seeding, cells were transfected at 50% confluence with 2 μg of DNA (per well in a 6-well plate) or 4 μg DNA (for a T25 flask for Western blots). At 6 h post-transfection, the transfection mix was replaced with medium containing 10% FBS. Cells were fixed and processed 24 h post-transfection for fluorescence microscopy.
2.4. Immunoprecipitation of mitochondrial proteins (a) BARD1 pull-down. 50 μl of Dynalbeads (Life Technologies) was used per immunoprecipitation. Beads were washed in 0.1 M phosphate buffer pH 7.4 for 2 min at room temperature. Washed beads were then resuspended in 0.1 M PBS pH 7.4 and incubated with 1 μg IgG or BARD1 antibody overnight at 37 °C, beads were washed with PBS with 0.1% (w/v) BSA, followed by a wash with 0.2 M Tris pH 8.5 for 4 h at 37 °C, and PBS/Triton X-100 for 10 min. Beads were resuspended in 50 mM Tris pH 7.5, 150 mM NaCl, 0.05% NP-40 (NET2) buffer and then incubated with 2 mg total or mitochondria-enriched lysate, washed and analyzed by SDS-PAGE and immunoblotting. (b) p53 pull-down. For IP of p53 we used Protein A-Sepharose beads (GE Healthcare Bio-Sciences) equilibrated in lysis buffer (100 mM NaCl, 50 mM Tris–HCl pH 7.5, 0.5% NP-40). Beads (30 μl/reaction) were pre-coated with 3 μg rabbit p53 antibody (sc-6243, Santa Cruz Biotech, CA) or rabbit IgG (Sigma, MO) for 1.5 h at 4 °C and washed with lysis buffer three times. Cells were lysed in lysis buffer for 30 min on ice, then centrifuged 15,000 ×g for 10 min at 4 °C. 100 μg extract was removed, denatured in 2× Laemmli buffer and run as an input sample. 1 mg of the supernatant (total protein lysate) was incubated with beads for 1.5 h at 4 °C and immunocomplexes pelleted, washed in RIPA buffer for 5 min before centrifugation, then denatured and analyzed by immunoblotting. The following antibodies were used for detection: GFP monoclonal (11814460001, Roche, 1:1000); BARD1 rabbit polyclonal (A300-263A, Bethyl Laboratories, 1:1000), 53BP1 monoclonal (PC712, Oncogene, 1:1000) and p53 monoclonal (sc126, Santa Cruz Biotec, 1:1000).
2.2. Plasmid DNA constructs
2.5. Immunoblotting
The human BARD1 coding sequence was originally provided by Prof. Richard Baer (Columbia University, NY). Most of the pYFP–BARD1 plasmids have been described previously [21,24]. In general, pYFP–BARD1 plasmids were generated by PCR-amplifying the YFP cDNA from the pEYFP-C1 vector (Life Technologies) and inserting it into the NotI site of the respective pFLAG–BARD1 plasmids. pYFP–BARD1 (551–777) was described previously [24]. New pYFP–BARD1 plasmids expressing different BARD1 sequences were constructed by amplifying BARD1 cDNA sequences and inserting them into the indicated restriction sites of pEYFP-Cl: (BARD1 425–525, Xho I/Sal I; 525–567, Sac I/Xma I). The forward and reverse primer sequences used are outlined in the Supplementary Methods section. The BARD1 internal deletion constructs BARD1 Δ425–525 and Δ525–567 were constructed by PCR-amplifying the flanking sequences of each deletion and re-ligating and inserting them into pEYFP-C1. Full details of the primer sequences and restriction sites are in the Supplementary Methods. Plasmid sequences were checked by DNA sequencing.
Cell lysates were denatured in 100 mM Tris–HCl (pH 6.8), 20% glycerol, 0.01% bromophenol blue, 10% β-mercaptoethanol, 5% SDS then separated by SDS-PAGE and analyzed by Western blot. Approximately 80 μg of protein extract was loaded per lane, resolved by SDS-PAGE and transferred onto a nitrocellulose membrane (Millipore Corporation). The membranes were blocked with 5% skim milk powder in PBST (PBS containing 0.2% Tween 20) for 1 h followed by incubation with primary antibody for 2 h, using antibodies against BARD1 (A300263A, Bethyl Laboratories, 1:1000), PCNA (610665, BR Transduction Laboratories, 1:5000), β-tubulin (T6074, Sigma, 1:3000), mtHSP70 (MA3-028, Affinity Bioreagents, 1:300) and p53 (sc-6243, Santa Cruz Biotec, 1:500). Blots were incubated with secondary horseradish peroxidase-conjugated antibodies (1:10,000, Sigma) for 1 h followed by detection using enhanced chemiluminescence (ECL; Amersham Biosciences). For detection of ectopic YFP–BARD1 proteins, cells were transfected with YFP–BARD1 plasmid and processed for mitochondrial extraction 24 h later. 2.6. Immunofluorescence microscopy
2.3. Isolation of mitochondrial extracts Mitochondrial fractions were enriched from cells using the Qproteome mitochondrial isolation kit (Qiagen). Cells were lysed to isolate cytosolic proteins. Plasma membranes and organelles such as nuclei, mitochondria and endoplasmic reticulum were pelleted by centrifugation at 1000 ×g for 10 min. The pellet was then resuspended in Disruption Buffer, using a narrow-gauge needle and re-centrifuged to pellet nuclei and cell debris. The supernatant was again centrifuged at 6000 ×g for 10 min to pellet the mitochondria, which were snap frozen in liquid nitrogen, thawed on ice and analyzed by immunoprecipitation and/or western blot analysis.
Cells were grown on coverslips at 60% density and fixed in 3.7% formalin/PBS for 20 min, followed by permeabilization with 0.2% Triton X-100/PBS for 10 min at 24 h post-transfection, and incubated with various antibodies as described [24]. For analysis of mitochondrial membrane permeability, live cells were incubated with 100 nM MitoTracker CMX-Ros (Molecular probes) in medium for 30 min at 37 °C, then washed and fixed in ice-cold acetone:methanol (1:1) for 3 min at room temperature. For antibody staining of fixed cells, cells were blocked with 3% bovine serum albumin in PBS for 1 h, followed by addition of primary antibodies as follows: rabbit polyclonal Ab against BARD1 (A300-263A, Bethyl Laboratories, 1:500); rabbit polyclonal
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
1765
antibody against BARD1 Exon 4 (NB100-319, Novus Biologicals, 1:500), mouse anti-p53 monoclonal antibody (sc-126, Santa Cruz Biotec, 1:500). Bound antibodies were detected with either AlexaFluor 488conjugated (1:500; Molecular Probes, Inc.) or biotin-conjugated (1:500; DAKO) secondary anti-mouse or anti-rabbit antibodies. Biotinylated secondary antibodies were incubated with avidin D-Texas Red (Vector Laboratories, 1:1000 dilution) or Alexafluor 594 diluted 1:1000 tertiary stains. Coverslips were mounted with Vectashield aqueous mountant (Vector Laboratories) and cells photographed using an Olympus BL51 fluorescence microscope at × 400 magnification. A SPOT32 camera was used for general image capture. Cell images for deconvolution were taken using the Olympus IX71 Deltavision Core deconvolution microscope (GE Healthcare Life Sciences, Australia) equipped with a CoolSNAP HQ2 camera. Images collected were further resolved using Softworx deconvolution software. Other images were collected as z-stacks using an Olympus FV1000 confocal microscope. 2.7. Analysis of apoptosis by Bax oligomerization and Hoechst staining of nuclei Transfected cells were scored by single cell assay at 24 h posttransfection for appearance of abnormal nuclei by Hoechst 33258 (Sigma, MO) staining of chromatin, indicative of apoptosis. Scoring of Bax-positive cells was also performed as previously described [24], where MCF-7 cells were fixed in methanol–acetone and stained with Bax monoclonal antibody (Cell Signaling), and scored for intense clustering of Bax fluorescence. 2.8. Duolink proximity ligation assay MCF-7 cells were plated in 8 well chamber slides at 50% confluency 24 h before transfection of YFP-tagged BARD1 using K2 (Biontex Laboratories) transfection reagent. At 48 h post-transfection cells were fixed using 3.7% formalin solution in PBS for 20 min. Cells were then permeabilized with PBS containing 0.2% Triton-X-100 for 10 min. The PLA assay was performed using the commercial Duolink (Olink Bioscience) kit, following supplier instructions to detect interactions between YFP–BARD1 proteins (anti-GFP Ab, Roche, 11814460001) and endogenous p53 (anti-p53 Ab, Santa Cruz Biotechnology, sc-6243), BCL-2 (anti-Bcl-2 Ab #2872, Cell Signaling) or Bax (anti-Bax Ab #2772, Cell Signaling). The total number of PLA interactions per cell was scored in cells that were selected for moderate YFP expression only. Data from two different experiments were combined and analyzed by dot plot using GraphPad Prism software, with at least 100 cells scored per sample and per experiment. In each case, single antibody controls were tested to ensure low background levels and YFP empty vector was used as a control for off-target interaction signals. Single antibody controls are shown in Fig. 2C and are representative of every Duolink PLA experiment shown in this paper. 3. Results and discussion 3.1. Endogenous human BARD1 forms complexes with p53 at mitochondria Previous reports identified binding of p53 to mouse BARD1 [20,25]. To show that human p53 and BARD1 form complexes in MCF-7 breast cancer cells, we employed immunoprecipitation (IP) assays and confirmed detection of endogenous BARD1/p53 complexes in cell lysates (Fig. 1A). A reverse IP using p53 antibody detected capture of BARD1 and 53BP1 (positive control), but not transiently expressed yellow fluorescent protein (YFP; negative control). We next complemented this experiment by using a Duolink proximity ligation assay (Duolink PLA), an in situ protein–protein interaction method using specific primary antibodies and secondary fluorescent-amplification reactions, to show that ectopic YFP–BARD1 formed many complexes with endogenous p53 in MCF-7 cells compared to YFP-only control (see red dots in
Fig. 1. BARD1 forms p53 complexes at mitochondria. (A) Evidence that BARD1 binds p53 at mitochondria. Co-immunoprecipitation assays were performed using total extracts (2 mg) of MCF-7 cells (see Methods). Proteins were immunoprecipitated (IP) with BARD1 polyclonal antibody (left panel) or p53 polyclonal antibody (right panel) and with an IgG control, separated by SDS-PAGE and analyzed by Western blot. Filters were probed with anti-p53 and BARD1 antibodies as shown, revealing a specific association. This data is typical of two separate experiments. (B) Use of Duolink PLA to detect BARD1/p53 complexes in cells. MCF-7 cells were transfected with YFP and YFP–BARD1 plasmids, then 48 h posttransfection were processed with the Duolink PLA (refer to Methods for details) to detect BARD1/p53 interactions. Representative images are shown of YFP transfected cells (green) showing interactions (red dots) between ectopic YFP–BARD1 (detected with GFP primary antibody) and endogenous p53 (detected using p53 antibody) together with nuclear staining in blue. (C) Mitochondrial fractions of MCF-7 cells were immunoprecipitated with BARD1 and IgG control antibody and then probed with BARD1, p53 and mtHsp70 antibodies. The data shown are from two separate experiments. For input lanes, 200 μg of mitochondrial and total protein was loaded as a control.
Fig. 1B; for controls see Fig. 2). In particular, the Duolink PLA data indicated that the majority of YFP–BARD1/p53 complexes formed in the cytoplasm. Since both BARD1 [24] and p53 [27] are each known to locate at mitochondria, we tested for localized complex formation by IP and showed that BARD1 associated with p53 in mitochondria-enriched fractions (Fig. 1C). A known mitochondrial partner protein of p53, the mitochondrial form of HSP70, also appeared to associate with BARD1 at mitochondria (Fig. 1C). These results identify, for the first time, the presence of cytoplasmic BARD1/p53 complexes at mitochondria, and these
1766
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
localized interactions could contribute to the known role of p53 in BARD1 pro-apoptotic function [2]. 3.2. Mapping of BARD1 p53-binding sites: Role of BRCT domain p53 was previously reported to bind a C-terminal region in mouse BARD1 [20]. To further characterize the human BARD1/p53 interaction we expressed a series of C-terminal BARD1 fragments (see Fig. S1) in MCF-7 cells and tested their ability to associate with p53 in order to map the interaction domain. Initial experiments by IP were unsuccessful and therefore the more sensitive Duolink PLA approach was employed. As shown in Figs. 1B and 2B, when compared to background YFP expression alone, the ectopic form of wild-type YFP–BARD1 displayed consistent interactions with p53 as indicated by the numerous red dots (cell images in Fig. 2B) and these were quantified by scoring the red dots in transverse images of transfected cells captured and deconvolved by Deltavision microscopy (graphs in Fig. 2C). The experiments identified three sites of varying affinity in the C-terminus of BARD1, and of these the BRCT domains (551–777) bound p53 most strongly (summarized in Fig. 2A). It is noteworthy that internal deletion of the p53-binding sequence 525–567 caused a modest ~ 50% reduction in detectable BARD1/p53 complexes, whereas deletion of the BRCT domains caused almost complete loss of p53 binding. A similar loss of p53 interaction was observed with the BARD1 cancer mutation, Q564H, providing a strong validation of this assay system as the Q564H mutation was previously shown to disrupt p53 binding by IP [25]. It is not yet clear how a single mutation such as Q564H in the “linker” region can ablate p53 binding, however it might elicit some subtle alteration in BARD1 protein conformation. 3.3. The BRCT domain is required for cytoplasmic localization of BARD1 independent of p53 We previously characterized the nuclear import and export sequences that facilitate BARD1 shuttling between the nucleus and cytoplasm of cells [21,22]. While the p53 binding domains we identified here do not overlap the transport signals, we tested the impact of deleting the p53 binding sites on subcellular localization. A range of YFPtagged BARD1 fragments were transiently expressed in MCF-7 cells, and the cells then imaged by microscopy and scored for nuclear +/− cytoplasmic distribution. None of the internal deletions (i.e. Δ425–525 or Δ525–567) tested altered the nuclear-cytoplasmic equilibrium of BARD1. Unexpectedly, we discovered that deletion of the BRCT domains in the sequence 1–593 caused a striking nuclear accumulation of BARD1 (Fig. 3). This was observed under different fixation conditions in a range of experiments, and is surprising given that this fragment retains the potent nuclear export sequence (NES) at the N-terminus. The data suggest that the C-terminal BRCT domain is somehow required for optimal nuclear export of BARD1, perhaps by mediating a conformation that prevents the NES from being masked by other internal sequences, and thus allowing accessibility by the CRM1 export receptor. Given that p53 binds to the BRCT domain of BARD1 (Fig. 2), we tested whether p53 might contribute to BARD1 localization. However, the overexpression of GFP-p53 or knockdown of p53 had no influence on endogenous BARD1 or YFP–BARD1 in MCF-7 cells (KAM and BRH, unpublished data). Moreover, the Q564H mutation in full-length BARD1, which prevents binding to p53, did not alter the nuclear-cytoplasmic distribution of BARD1 (e.g. see Fig. 4C). Therefore the BRCT domain stimulates cytoplasmic localization of BARD1 independent of p53. 3.4. Loss of p53 binding does not prevent BARD1 mitochondrial localization in MCF-7 cells We next determined whether p53 contributed to the mitochondrial localization of BARD1. First we compared the YFP–BARD1 deletion fragments tested above. The internal deletion of sequences 425–525 or
525–567 did not prevent mitochondrial localization as determined by western blot analysis of mitochondria-enriched fractions of MCF-7 cells (Fig. 4B). As shown above, the C-terminal truncation of the BRCT domain greatly reduced the cytoplasmic pool of BARD1 and hence the amount available to locate at mitochondria, however analysis of the cytoplasmic pool showed that the BARD1 (1–593) mutant still had the capacity to associate with mitochondria (Fig. 4A,B). This mutant no longer binds p53, suggesting that p53 is not required for the targeting of BARD1 to mitochondria in MCF-7 cells. This is further supported by expression of the Q564H mutant, which does not bind p53 but located at mitochondria (Fig. 4B). 3.5. Deletion of the BARD1 BRCT domain/p53-binding site disrupts Bax oligomerization BARD1 apoptotic function is dependent on p53 [25,26]. Hence we investigated whether p53 is required for BARD1 to induce apoptosis through Bax clustering at mitochondria. First we mapped the BARD1 fragments required for Bax oligomerization in MCF-7 cells (Fig. 4A). Cells were transfected with YFP–BARD1 sequences and then stained for Bax and counterstained with Hoechst dye to assess pre-apoptotic changes in nuclear integrity by microscopy (Bax cell images in Fig. 4C left panel; quantification in Fig. 4C right panel graphs; for example of Hoechst staining see Supplementary Fig. S2). Relative to wild-type BARD1, shorter fragments of BARD1 displayed weak Bax-stimulating activity. Notably, removal of the ankyrin repeats (Δ425–525) did not eliminate Bax oligomerization. We tested whether loss of either of the two main p53-binding sites impacted on apoptosis, and observed that deletion of the moderate p53-binding site (see Δ525–567) reduced Bax oligomerization and apoptosis by ~ 30% (Fig. 4C). Of particular note, deletion of the strong p53-binding site within the C-terminal BRCT domain (see 1–593) eliminated Bax oligomerization and apoptosis. Similar results were seen with the non-p53-binding Q564H mutant. These results suggest that the p53-dependence of BARD1 apoptotic activity correlates with binding of p53 to BARD1 C-terminal sequences. Since p53 was not required to transport BARD1 to the cytoplasm or to mitochondria (Figs. 3 and 4), it is possible that p53 interacts with BARD1 at the mitochondria to promote its role in apoptosis. 3.6. BARD1 binds and down-regulates Bcl-2 at mitochondria: a potential mechanism for Bax oligomerization The expression levels of the Bcl-2 family proteins, which consist of anti-apoptotic and pro-apoptotic members, determine life or death of a cell. To investigate possible intracellular mechanisms for YFP– BARD1-induced apoptosis, we tested for binding of BARD1 to the antiapoptotic factor Bcl-2. No interaction was detectable between endogenous BARD1 and Bcl-2 by IP (Fig. 5A), whereas a modest association could be detected using the more sensitive Duolink PLA (Fig. 5B). BARD1 revealed no association with Bax under similar conditions (Fig. 5B, lower panel graph). Next, the expression of Bcl-2 protein was tested. MCF-7 cells were fractionated into total and mitochondrial/ cytosolic extracts and analyzed for endogenous Bcl-2 by Western blot. In total extracts (Fig. 5C), the expression of Bcl-2 remained unchanged after overexpression of YFP–BARD1. In contrast, mitochondrial Bcl-2 expression was moderately down-regulated in two separate experiments (see two panels in Fig. 5C). Transient expression of YFP–BARD1 was confirmed by probing with anti-GFP antibody. Thus, a small pool of BARD1 associates with Bcl-2, possibly through complex formation with p53, and triggers its down-regulation at mitochondria. Bcl-2 is a potent inhibitor of apoptosis and is overexpressed in a wide variety of malignancies [28,29]. In addition, the down-regulation of Bcl-2 has been shown to inhibit tumor growth and induce apoptosis. These results implicate BARD1 as a possible regulator of Bcl-2 and indicate that down-regulation of Bcl-2 may play a role in BARD1 induced
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
1767
Fig. 2. BARD1 binds p53 through C-terminal sequences 525–567 and 551–777. (A) Schematic of BARD1 showing the key domains (NES, NLS, Ankyrin repeats and BRCT domains) and fragments used to map the p53 binding region. A summary of the results from Duolink PLA interaction studies is shown at the right. (B) The PLA method was used to map the BARD1 p53 binding site. MCF-7 cells were transfected with YFP and different YFP–BARD1 plasmids. Representative images are shown of YFP transfected cells (green) showing interactions (red dots) between ectopic YFP–BARD1 (detected with GFP primary antibody) and endogenous p53 (using p53 antibody), with nuclear staining in blue. (C) The total number of Duolink PLA interactions per moderate YFP-expressing cell were scored. Data from at least two different experiments were combined and shown on the Graphpad Prism dot plot, with at least 100 cells scored per sample and per experiment. Levels of significance above background are indicated (**, p ≤ 0.01 and ***, p ≤ 0.001). Note that the two comparative graphs shown each represent at least 2 experiments, but performed 6 months apart. The degree of binding activity for each BARD1 fragment relative to wild-type (WT, set to 100%) is shown below the graphs for ease of comparison.
apoptosis. To test this, we performed a reconstitution experiment to test whether forced expression of Bcl-2 could prevent Bax regulation by BARD1. Indeed, as shown in Fig. 5D, transient expression of Flagtagged BARD1 (wild-type) induced measurable Bax clustering at
mitochondria and this was completely blocked by co-expression of GFP-Bcl-2. In comparison, co-expression of GFP-Bax caused hyperstimulation of oligomeric Bax in BARD1 transfected cells (Fig. 5D). Similar results were observed for apoptosis, and provide support for a
1768
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
Fig. 3. Loss of the BRCT domain induces nuclear accumulation of BARD1. pEYFP-C1 and a series of plasmids that encode YFP-tagged BARD1 sequences were transiently expressed at moderate levels in MCF-7 cells and analyzed for subcellular localization. The cells were fixed and YFP fluorescence detected by microscopy, and the cellular distribution patterns scored (N800 cells were counted per sequence across 3 individual experiments performed in duplicate) and plotted. Note that the truncation of the C-terminal BRCT domain (1–593) is the only deletion to shift BARD1 to the nucleus.
functional link between BARD1, Bcl-2 down-regulation and Bax oligomerization.
4. Conclusions Here we show for the first time that endogenous BARD1 forms p53 protein complexes at mitochondria and identify the tandem BRCT domain as a major binding site for p53. Deletion mapping experiments also revealed a new role for the BRCT domain in driving cytoplasmic localization of BARD1, although this did not appear to depend on binding to p53. We propose that p53 can associate with BARD1 at mitochondria to stimulate certain functions, such as apoptotic activity. BARD1 and p53 are nuclear-cytoplasmic shuttling proteins with the potential to translocate from nucleus to mitochondria to trigger an apoptotic response [21,27,30]. We found that p53 is not essential for targeting of BARD1 to mitochondria but it did contribute to BARD1dependent apoptosis (Fig. 4). Previously, Feki et al. [20] mapped p53binding in mouse BARD1 to the C-terminal sequence 394–604 comprising the ankyrin and linker domains. We show here that in human BARD1 p53 associates with three distinct sequences, the strongest being the tandem BRCT domain (Fig. 2). The glutamine amino acid Q564 lies within both the extended linker (525–567) and BRCT domain (551–777) fragments of BARD1 that displayed strong binding to p53 (Fig. 2). The cancer-linked mutation of this residue, Q564H, abolishes p53 binding, and since this mutation was reported not to alter BARD1
structure [19], it may represent a specific contact site for p53 or affect binding to p53 through a more subtle shift in conformation. We confirmed the interaction between BARD1 and p53 by immunoprecipitation assays, and found that endogenous BARD1 formed in vivo complexes with p53 and another p53-associated factor, mtHsp70, in mitochondrial extracts. Although we have focused on p53 regulation of BARD1, it is anticipated that BARD1 can elicit some reciprocal regulation of p53. This is suggested by earlier studies which found that BARD1 is transcriptionally up-regulated in response to genotoxic stress, leading to increased p53 levels through stabilization and phosphorylation of p53 at serine-15 [2]. Preliminary electron microscopy imaging (see Fig. S3) detected BARD1 at the outer surface and inner matrix of mitochondria in MCF-7 cells, similar to that shown previously for p53 [27]. We speculate that a pool of BARD1 and p53 could attach to the outer mitochondrial surface and get imported together by mtHsp70, although more detailed analyses are required to define precisely where p53/ BARD1 complexes locate within mitochondria. Given the role of BARD1 in nuclear DNA repair [10,11], and involvement of p53 in maintaining mitochondrial DNA base excision repair [31,32], it is conceivable that BARD1 performs a related function in mitochondria separate from (or perhaps linked to) its ability to regulate apoptosis. In response to DNA damage, we observed that BARD1 displays a modest increase at mitochondria [24], suggesting the possibility that BARD1–p53 complexes might increase at mitochondria after DNA damage stress. We also note that a DNA damage-induced phosphorylated pool of BARD1 increases at mitochondria as it does in the nucleus [33].
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
1769
Fig. 4. The BRCT domain stimulates BARD1 mitochondrial targeting and Bax oligomerization. (A) A series of BARD1 deletion fragments were analyzed for mitochondrial localization and ability to induce Bax oligomers and apoptosis. The results are shown summarized from Western blots and IF staining, from negative (−) to positive (+ to ++++, weak to strongest), in comparison to p53-binding ability and the nuclear-cytoplasmic distribution pattern localization. (B) YFP–BARD1 sequences were transiently expressed in MCF-7 cells, and then tested for detection in mitochondrial fractions. Mitochondria-enriched (M) and cytosolic (C) (100 μg/lane) extracts were analyzed by western blotting and YFP–BARD1 was detected using a monoclonal anti-GFP antibody (Roche, 1:1000). All of the BARD1 sequences were detectable in the mitochondrial fraction. Antibodies against mitochondria (mtHsp70), cytosol (tubulin), and nuclear/cytoplasmic (PCNA) markers were used to demonstrate the purity of the fractions. The data presented are typical of two or more experiments. (C) YFP–BARD1 plasmids were expressed in MCF-7 cells, and at 24 h post-transfection cells were stained for Bax (a monoclonal Ab targeting the N-terminus of Bax) and analyzed by fluorescence microscopy (representative cell images shown in left panel, Bax staining in red). Bax clustering is clearly evident above background in cells transfected with wild-type BARD1 and specific BARD1 sub-fragments. The transfected cells were then scored for % cells displaying Bax oligomerization and those that displayed abnormal nuclei after Hoechst staining (see Methods) as shown in right panel.
The damage-dependent phosphorylation of BARD1 occurs in the Cterminal BRCT domain [34], the same sequence shown here to bind p53, and hence it is possible that post-translational modifications influence the action of BARD1–p53 complexes at mitochondria. Transient expression of YFP–BARD1 induces Bax oligomerization at mitochondria (Fig. 4) [24]. In viable cells, Bax is monomeric and located in the cytosol or loosely attached to the outer mitochondrial membrane (OMM). Changes in conformation that expose its N-terminus stimulate oligomerization and its insertion into the OMM, an integral step toward apoptosis. We found that BARD1 expression did not alter Bax protein
levels at mitochondria, suggesting more a post-translational change in Bax folding and self-association than changes in its trafficking. In this regard, the transient expression of BARD1 caused a specific reduction of endogenous Bcl-2 in mitochondria-enriched fractions (Fig. 5). Bcl-2 can inhibit apoptosis by binding to Bax and blocking Bax oligomerization [35]. Previously, p53 was reported to bind Bcl-2 at mitochondria [36], reduce its expression in MCF-7 cells [37], and induce mitochondrial membrane permeabilization [36,38]. More notably, p53 can disrupt the Bcl-2/Bax complex by directly binding to a flexible loop regulatory domain (FLD) in Bcl-2, resulting in enhanced Bax insertion in the OMM
1770
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
Fig. 5. The ability of YFP–BARD1 to induce Bax oligomerization correlates with loss of Bcl-2 at mitochondria. (A) MCF-7 total extract was subjected to immunoprecipitation using BARD1 antibody but did not detect endogenous BARD1–Bcl-2 complexes. (B) A Duolink PLA was used to test for binding between transiently expressed YFP–BARD1 and endogenous Bcl-2 or Bax. Duolink PLA-positive complexes are seen as red dots. Quantification of the number of PLA complexes per cell indicated that YFP–BARD1 showed weak binding to Bcl-2 and no binding to Bax, relative to YFP control. (C) MCF-7 cells, untransfected (unt) or transfected with YFP–BARD1, were fractionated into mitochondrial (M), cytosolic (C) and total extracts 24 h post-transfection and assessed by western blot. Left panel: 60 μg of total extracts was analyzed by western blot, and probed with antibodies to GFP, β-tubulin, Bcl-2, and Bax. β-Tubulin was used as a loading control. Middle and right panels: 100 μg of mitochondrial and cytosolic extract protein was subjected to SDS-PAGE and western blot analysis, and probed with monoclonal antibodies against Bcl-2 and Bax. The two panels shown for Bcl-2 represent two separate experiments. The integrity of the mitochondrial/cytosolic fractions was confirmed by staining with nuclear/cytosolic (PCNA), cytosolic (β-tubulin) and mitochondrial marker (mthsp 70). The transfection of YFP–BARD1 was confirmed by probing with anti-GFP antibody. (D) Bcl-2 overexpression blocks Bax oligomerization. Flag-tagged BARD1 and GFP-tagged Bcl-2 or Bax were expressed in MCF-7 cells alone or in combination, and cells scored after 24 h for Bax oligomerization or apoptosis (Hoechst staining of nuclei). Cell images show that overexpressed BARD1 and Bax, but not Bcl-2, elicit intense Bax staining. Scoring of transfected cells showed that coexpression of GFP-Bcl-2 blocked the induction of Bax and apoptosis by BARD1.
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
and apoptosis [39]. Thus, based on our findings we hypothesize that BARD1 can bind to p53 and Bcl-2 at mitochondria, and under certain circumstances such as cellular stress caused by DNA damage, BARD1 can reduce Bcl-2 expression at mitochondria leading indirectly to a stimulation of Bax oligomerization and apoptosis. Author contribution VT, EM and KAM performed the majority of experiments and contributed to writing of the paper. MM, KB, KM, and YL created specific reagents such as plasmids and performed some of the supporting experiments. EK and RB contributed to the electron microscopy. AD and HR contributed to the conception of the study and writing of the manuscript. BRH designed the study, supervised the project and wrote the paper. Acknowledgments We thank Dr. Richard Baer for the original BARD1 plasmid and Dr Irmgard Irminger-Finger for discussions and encouragement of this work. This work was supported by grants to BRH from the Cancer Council New South Wales (project grant RG06-09) and the National Breast Cancer Foundation of Australia (grant NC-13-13), and to BRH and HR from the Australian Research Council of Australia (ARC; grant DP0881263). BRH is a Senior Research Fellow supported by the Cancer Institute New South Wales (grant 12/CDF/2-12). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.cellsig.2015.05.011. References [1] R. Baer, T. Ludwig, Curr. Opin. Genet. Dev. 12 (2002) 86–91. [2] I. Irminger-Finger, C.E. Jefford, Nat. Rev. Cancer 6 (2006) 382–391. [3] T.H. Thai, F. Du, J.T. Tsan, Y. Jin, A. Phung, M.A. Spillman, H.F. Massa, C.Y. Muller, R. Ashfaq, J.M. Mathis, D.S. Miller, B.J. Trask, R. Baer, A.M. Bowcock, Hum. Mol. Genet. 7 (1998) 195–202. [4] C. Ghimenti, E. Sensi, S. Presciuttini, I.M. Brunetti, P. Conte, G. Bevilacqua, M.A. Caligo, Gene Chromosome Cancer 33 (2002) 235–242. [5] S.-M. Karppinen, K. Heikkinen, K. Rapakko, R. Winqvist, J. Med. Genet. 41 (2004) e114. [6] L.C. Wu, Z.W. Wang, J.T. Tsan, M.A. Spillman, A. Phung, X.L. Xu, M.C. Yang, L.Y. Hwang, A.M. Bowcock, R. Baer, Nat. Genet. 14 (1996) 430–440. [7] P.S. Brzovic, P. Rajagopal, D.W. Hoyt, M.C. King, R.E. Klevit, Nat. Struct. Biol. 8 (2001) 833–837.
1771
[8] V. Joukov, J. Chen, E.A. Fox, J.B. Green, D.M. Livingston, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 12078–12083. [9] M. Fabbro, J.A. Rodriguez, R. Baer, B.R. Henderson, J. Biol. Chem. 277 (2002) 21315–21324. [10] U.K. Westermark, M. Reyngold, A.B. Olshen, R. Baer, M. Jasin, M.E. Moynahan, Mol. Cell. Biol. 23 (2003) 7926–7936. [11] E.E. McCarthy, J.T. Celebi, R. Baer, T. Ludwig, Mol. Cell. Biol. 23 (2003) 5056–5063. [12] H. Ruffner, C.A. Joazeiro, D. Hemmati, T. Hunter, I.M. Verma, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 5134–5139. [13] R. Hashizume, M. Fukuda, I. Maeda, H. Nishikawa, D. Oyake, Y. Yabuki, H. Ogata, T. Ohta, J. Biol. Chem. 276 (2001) 14537–14540. [14] V. Joukov, A.C. Groen, T. Prokhorova, R. Gerson, E. White, A. Rodriguez, J.C. Walter, D.M. Livingston, Cell 127 (2006) 539–552. [15] L.M. Starita, Y. Machida, S. Sankaran, J.E. Elias, K. Griffin, B.P. Schlegel, S.P. Gygi, J.D. Parvin, Mol. Cell. Biol. 24 (2004) 8457–8466. [16] B.R. Henderson, BRCA1 intracellular dynamics, transport and function, Scientifica2012.http://dx.doi.org/10.6064/2012/796808. [17] R. Hakem, J.L. de la Pompa, C. Sirard, R. Mo, M. Woo, A. Hakem, A. Wakeham, J. Potter, A. Reitmair, F. Billia, E. Firpo, C.C. Hui, J. Roberts, J. Rossant, T.W. Mak, Cell 85 (1996) 1009–1023. [18] T.-C. Ayi, J.T. Tsan, L.-Y. Hwang, A.M. Bowcock, R. Baer, Oncogene 17 (1998) 2143–2148. [19] D. Fox III, I.L. Trong, P. Rajagopal, P.S. Brzovic, R.E. Stenkamp, R.E. Klevit, J. Biol. Chem. 283 (2008) 21179–21186. [20] A. Feki, C.E. Jefford, P. Berardi, J.Y. Wu, L. Cartier, K.H. Krause, I. Irminger-Finger, Oncogene 24 (2005) 3726–3736. [21] J.A. Rodriguez, S. Schuchner, W.W. Au, M. Fabbro, B.R. Henderson, Oncogene 23 (2004) 1809–1820. [22] S. Schüchner, V. Tembe, J.A. Rodriguez, B.R. Henderson, J. Biol. Chem. 280 (2005) 8855–8861. [23] C.E. Jefford, A. Feki, J. Harb, K.H. Krause, I. Irminger-Finger, Oncogene 23 (2004) 3509–3520. [24] V. Tembe, B.R. Henderson, J. Biol. Chem. 282 (2007) 20513–20522. [25] I. Irminger-Finger, W.C. Leung, J. Li, M. Dubois-Dauphin, J. Harb, A. Feki, C.E. Jefford, J.V. Soriano, M. Jaconi, R. Montesano, K.H. Krause, Mol. Cell 8 (2001) 1255–1266. [26] M. Fabbro, S. Schuechner, W.W.Y. Au, B.R. Henderson, Exp. Cell Res. 298 (2004) 661–673. [27] N.D. Marchenko, A. Zaika, U.M. Moll, J. Biol. Chem. 275 (2000) 16202–16212. [28] D.G. Breckenridge, D. Xue, Curr. Opin. Cell Biol. 16 (2004) 647–652. [29] A. Gross, J.M. McDonnell, S.J. Korsmeyer, Genes Dev. 13 (1999) 1899–1911. [30] Y. Arima, M. Nitta, S. Kuninaka, D. Zhang, T. Fujiwara, Y. Taya, M. Nakao, H. Saya, J. Biol. Chem. 280 (2005) 19166–19176. [31] D. Chen, Z. Yu, Z. Zhu, C.D. Lopez, Cancer Res. 66 (2006) 3485–3494. [32] N.C. de Souza-Pinto, C.C. Harris, V.A. Bohr, Oncogene 23 (2004) 6559–6568. [33] K. Brodie, B.R. Henderson, Cell. Signal. 22 (2010) 291–302. [34] H.S. Kim, H. Li, M. Cevher, A. Parmelee, D. Fonseca, F.E. Kleiman, S.B. Lee, Cancer Res. 66 (2006) 4561–4565. [35] V. Mikhailov, M. Mikhailova, D.J. Pulkrabek, Z. Dong, M.A. Venkatachalam, P. Saikumar, J. Biol. Chem. 276 (2001) 18361–18374. [36] Y. Tomita, N. Marchenko, S. Erster, A. Nemajerova, A. Dehner, C. Klein, H. Pan, H. Kessler, P. Pancoska, U.M. Moll, J. Biol. Chem. 281 (2006) 8600–8606. [37] S. Haldar, M. Negrini, M. Monne, S. Sabbioni, C.M. Croce, Cancer Res. 54 (1994) 2095–2097. [38] A. Comel, G. Sorrentino, V. Capaci, G. Del Sal, FEBS Lett. 588 (2014) 2600–2609. [39] X. Deng, F. Gao, T. Flagg, J. Anderson, W.S. May, Mol. Cell. Biol. 26 (2006) 4421–4434.
Contents lists available at ScienceDirect
Cellular Signalling journal homepage: www.elsevier.com/locate/cellsig
The BARD1 BRCT domain contributes to p53 binding, cytoplasmic and mitochondrial localization, and apoptotic function Varsha Tembe a,1, Estefania Martino-Echarri a,1, Kamila A. Marzec a,1, Myth T.S. Mok a,2, Kirsty M. Brodie a, Kate Mills a, Ying Lei a, Anna DeFazio a, Helen Rizos a,3, Emma Kettle b, Ross Boadle b, Beric R. Henderson a,⁎ a b
Centre for Cancer Research, The Westmead Millennium Institute for Medical Research, The University of Sydney, Westmead, NSW 2145, Australia Electron Microscopy Unit, Institute for Clinical Pathology and Medical Research, Westmead Hospital, Westmead, NSW 2145, Australia
a r t i c l e
i n f o
Article history: Received 7 May 2015 Accepted 15 May 2015 Available online 3 June 2015 Keywords: BARD1 p53 Mitochondria Apoptosis BCL-2
a b s t r a c t BARD1 is a breast cancer tumor suppressor with multiple domains and functions. BARD1 comprises a tandem BRCT domain at the C-terminus, and this sequence has been reported to target BARD1 to distinct subcellular locations such as nuclear DNA breakage sites and the centrosome through binding to regulatory proteins such as HP1 and OLA1, respectively. We now identify the BRCT domain as a binding site for p53. We first confirmed previous reports that endogenous BARD1 binds to p53 by immunoprecipitation assay, and further show that BARD1/ p53 complexes locate at mitochondria suggesting a cellular location for p53 regulation of BARD1 apoptotic activity. We used a proximity ligation assay to map three distinct p53 binding sequences in human BARD1, ranging from weak (425–525) and modest (525–567) to strong (551–777 comprising BRCT domains). Deletion of the BRCT sequence caused major defects in the ability of BARD1 to (1) bind p53, (2) localize to the cytoplasm and mitochondria, and (3) induce Bax oligomerization and apoptosis. Our data suggest that BARD1 can move to mitochondria independent of p53, but subsequently associates with p53 to induce Bax clustering in part by decreasing mitochondrial Bcl-2 levels. We therefore identify a role for the BRCT domain in stimulating BARD1 nuclear export and mitochondrial localization, and in assembling mitochondrial BARD1/p53 complexes to regulate specific activities such as apoptotic function. © 2015 Elsevier Inc. All rights reserved.
1. Introduction The BRCA1-associated RING domain 1 (BARD1) protein is a putative tumor suppressor and a key binding partner of the breast and ovarian cancer susceptibility protein, BRCA1 [1,2]. The BARD1 gene is mutated in a subset of breast and ovarian cancer patients [3–5], and was first discovered in a genetic screen for BRCA1-interacting proteins [6]. BARD1 was later confirmed as the major in vivo partner for BRCA1 [1], and the two proteins interact through their N-terminal tail sequences [7]. The formation of a stable BRCA1/BARD1 dimer has a positive influence on BARD1 stability [8], nuclear localization [9] and DNA repair function Abbreviations: BRCA1, breast cancer regulatory protein-1; BARD1, BRCA1-associated RING domain 1; CMX-Ros, chloromethyl-X-rosamine; FCS, fetal calf serum; mtHSP70, mitochondrial heat shock protein 70; NES, nuclear export sequence(s); NLS, nuclear localization sequence; OMM, outer mitochondrial membrane; PBS, phosphate-buffered saline; PCNA, proliferating cell nuclear antigen; YFP, yellow fluorescence protein. ⁎ Corresponding author at: Westmead Millennium Institute, 176 Hawkesbury Road (PO Box 412), Westmead, NSW 2145, Australia. Tel.: +61 2 86273730. E-mail address: [email protected] (B.R. Henderson). 1 These authors contributed equally to this work. 2 Current address: School of Biomedical Sciences, Faculty of Medicine, The Chinese University of Hong Kong. 3 Current address: School of Advanced Medicine, Macquarie University, NSW 2109, Australia.
http://dx.doi.org/10.1016/j.cellsig.2015.05.011 0898-6568/© 2015 Elsevier Inc. All rights reserved.
[10,11]. Furthermore, the BRCA1/BARD1 dimer acts as an E3 ubiquitin (Ub) ligase whose enzymatic activity is impaired by cancer associated gene mutations in the BRCA1 RING domain [12,13]. BRCA1/BARD1 Ub ligase activity is thought to contribute to the regulation of DNA repair [11], mitotic cell division [14] and centrosome duplication [15,16]. The BRCA1/BARD1 heterodimer is also important for cell viability, and loss of either BRCA1 or BARD1 in mice leads to chromosomal abnormalities and early embryonic death due to severe cell proliferation defects [11, 17]. Human BARD1 comprises 777 amino acids and, like BRCA1, has an N-terminal RING finger domain and two C-terminal BRCT domains [18]. Unlike BRCA1, BARD1 contains a centrally located ankyrin repeat domain known to mediate protein–protein interactions [19]; the role of this domain in BARD1 is still under investigation although it was reported to contribute to association with p53 in mouse [20]. BARD1 can shuttle between the nucleus and cytoplasm [21,22], and it was previously proposed that while nuclear localization of BARD1 correlated with its role in DNA repair and cell survival [21,23], the cytoplasmic accumulation of BARD1 at mitochondria correlated with induction of apoptosis [24]. BARD1 is transcriptionally up-regulated in response to genotoxic stress in rodent cells [25]. Furthermore, the overexpression of exogenous BARD1 leads to apoptosis and this is partly dependent on
1764
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
functional p53, but is independent of and inhibited by BRCA1 as demonstrated in p53 or BRCA1-deficient cell lines [25,26]. We previously discovered a pool of BARD1 at mitochondria whose expression correlated with an increase in Bax oligomerization, mitochondrial membrane permeability and apoptosis [24]. Given that BARD1 apoptotic activity is p53-dependent, and that p53 itself is known to translocate to mitochondria and induce apoptosis [27], we hypothesized that p53 might regulate BARD1 apoptotic function by either (i) recruiting it to mitochondria, (ii) binding BARD1 at mitochondria and/or (iii) modulating its ability to stimulate Bax oligomerization. In this study we identify a key role for the BRCT domain in p53 binding, recruitment to the cytoplasm/mitochondria and apoptotic function. These data reveal that while p53 is not required for targeting of BARD1 to mitochondria, it does form complexes with BARD1 at this organelle and stimulates its role in Bax oligomerization.
2. Materials and methods 2.1. Cell culture and transfections Human MCF-7 breast cancer cells were grown under standard tissue culture conditions in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS). The cell line was mycoplasmanegative. Transient transfections were performed using Lipofectamine2000 (Life Technologies, Vic, Australia) according to manufacturer instructions. Briefly, at 24 h after seeding, cells were transfected at 50% confluence with 2 μg of DNA (per well in a 6-well plate) or 4 μg DNA (for a T25 flask for Western blots). At 6 h post-transfection, the transfection mix was replaced with medium containing 10% FBS. Cells were fixed and processed 24 h post-transfection for fluorescence microscopy.
2.4. Immunoprecipitation of mitochondrial proteins (a) BARD1 pull-down. 50 μl of Dynalbeads (Life Technologies) was used per immunoprecipitation. Beads were washed in 0.1 M phosphate buffer pH 7.4 for 2 min at room temperature. Washed beads were then resuspended in 0.1 M PBS pH 7.4 and incubated with 1 μg IgG or BARD1 antibody overnight at 37 °C, beads were washed with PBS with 0.1% (w/v) BSA, followed by a wash with 0.2 M Tris pH 8.5 for 4 h at 37 °C, and PBS/Triton X-100 for 10 min. Beads were resuspended in 50 mM Tris pH 7.5, 150 mM NaCl, 0.05% NP-40 (NET2) buffer and then incubated with 2 mg total or mitochondria-enriched lysate, washed and analyzed by SDS-PAGE and immunoblotting. (b) p53 pull-down. For IP of p53 we used Protein A-Sepharose beads (GE Healthcare Bio-Sciences) equilibrated in lysis buffer (100 mM NaCl, 50 mM Tris–HCl pH 7.5, 0.5% NP-40). Beads (30 μl/reaction) were pre-coated with 3 μg rabbit p53 antibody (sc-6243, Santa Cruz Biotech, CA) or rabbit IgG (Sigma, MO) for 1.5 h at 4 °C and washed with lysis buffer three times. Cells were lysed in lysis buffer for 30 min on ice, then centrifuged 15,000 ×g for 10 min at 4 °C. 100 μg extract was removed, denatured in 2× Laemmli buffer and run as an input sample. 1 mg of the supernatant (total protein lysate) was incubated with beads for 1.5 h at 4 °C and immunocomplexes pelleted, washed in RIPA buffer for 5 min before centrifugation, then denatured and analyzed by immunoblotting. The following antibodies were used for detection: GFP monoclonal (11814460001, Roche, 1:1000); BARD1 rabbit polyclonal (A300-263A, Bethyl Laboratories, 1:1000), 53BP1 monoclonal (PC712, Oncogene, 1:1000) and p53 monoclonal (sc126, Santa Cruz Biotec, 1:1000).
2.2. Plasmid DNA constructs
2.5. Immunoblotting
The human BARD1 coding sequence was originally provided by Prof. Richard Baer (Columbia University, NY). Most of the pYFP–BARD1 plasmids have been described previously [21,24]. In general, pYFP–BARD1 plasmids were generated by PCR-amplifying the YFP cDNA from the pEYFP-C1 vector (Life Technologies) and inserting it into the NotI site of the respective pFLAG–BARD1 plasmids. pYFP–BARD1 (551–777) was described previously [24]. New pYFP–BARD1 plasmids expressing different BARD1 sequences were constructed by amplifying BARD1 cDNA sequences and inserting them into the indicated restriction sites of pEYFP-Cl: (BARD1 425–525, Xho I/Sal I; 525–567, Sac I/Xma I). The forward and reverse primer sequences used are outlined in the Supplementary Methods section. The BARD1 internal deletion constructs BARD1 Δ425–525 and Δ525–567 were constructed by PCR-amplifying the flanking sequences of each deletion and re-ligating and inserting them into pEYFP-C1. Full details of the primer sequences and restriction sites are in the Supplementary Methods. Plasmid sequences were checked by DNA sequencing.
Cell lysates were denatured in 100 mM Tris–HCl (pH 6.8), 20% glycerol, 0.01% bromophenol blue, 10% β-mercaptoethanol, 5% SDS then separated by SDS-PAGE and analyzed by Western blot. Approximately 80 μg of protein extract was loaded per lane, resolved by SDS-PAGE and transferred onto a nitrocellulose membrane (Millipore Corporation). The membranes were blocked with 5% skim milk powder in PBST (PBS containing 0.2% Tween 20) for 1 h followed by incubation with primary antibody for 2 h, using antibodies against BARD1 (A300263A, Bethyl Laboratories, 1:1000), PCNA (610665, BR Transduction Laboratories, 1:5000), β-tubulin (T6074, Sigma, 1:3000), mtHSP70 (MA3-028, Affinity Bioreagents, 1:300) and p53 (sc-6243, Santa Cruz Biotec, 1:500). Blots were incubated with secondary horseradish peroxidase-conjugated antibodies (1:10,000, Sigma) for 1 h followed by detection using enhanced chemiluminescence (ECL; Amersham Biosciences). For detection of ectopic YFP–BARD1 proteins, cells were transfected with YFP–BARD1 plasmid and processed for mitochondrial extraction 24 h later. 2.6. Immunofluorescence microscopy
2.3. Isolation of mitochondrial extracts Mitochondrial fractions were enriched from cells using the Qproteome mitochondrial isolation kit (Qiagen). Cells were lysed to isolate cytosolic proteins. Plasma membranes and organelles such as nuclei, mitochondria and endoplasmic reticulum were pelleted by centrifugation at 1000 ×g for 10 min. The pellet was then resuspended in Disruption Buffer, using a narrow-gauge needle and re-centrifuged to pellet nuclei and cell debris. The supernatant was again centrifuged at 6000 ×g for 10 min to pellet the mitochondria, which were snap frozen in liquid nitrogen, thawed on ice and analyzed by immunoprecipitation and/or western blot analysis.
Cells were grown on coverslips at 60% density and fixed in 3.7% formalin/PBS for 20 min, followed by permeabilization with 0.2% Triton X-100/PBS for 10 min at 24 h post-transfection, and incubated with various antibodies as described [24]. For analysis of mitochondrial membrane permeability, live cells were incubated with 100 nM MitoTracker CMX-Ros (Molecular probes) in medium for 30 min at 37 °C, then washed and fixed in ice-cold acetone:methanol (1:1) for 3 min at room temperature. For antibody staining of fixed cells, cells were blocked with 3% bovine serum albumin in PBS for 1 h, followed by addition of primary antibodies as follows: rabbit polyclonal Ab against BARD1 (A300-263A, Bethyl Laboratories, 1:500); rabbit polyclonal
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
1765
antibody against BARD1 Exon 4 (NB100-319, Novus Biologicals, 1:500), mouse anti-p53 monoclonal antibody (sc-126, Santa Cruz Biotec, 1:500). Bound antibodies were detected with either AlexaFluor 488conjugated (1:500; Molecular Probes, Inc.) or biotin-conjugated (1:500; DAKO) secondary anti-mouse or anti-rabbit antibodies. Biotinylated secondary antibodies were incubated with avidin D-Texas Red (Vector Laboratories, 1:1000 dilution) or Alexafluor 594 diluted 1:1000 tertiary stains. Coverslips were mounted with Vectashield aqueous mountant (Vector Laboratories) and cells photographed using an Olympus BL51 fluorescence microscope at × 400 magnification. A SPOT32 camera was used for general image capture. Cell images for deconvolution were taken using the Olympus IX71 Deltavision Core deconvolution microscope (GE Healthcare Life Sciences, Australia) equipped with a CoolSNAP HQ2 camera. Images collected were further resolved using Softworx deconvolution software. Other images were collected as z-stacks using an Olympus FV1000 confocal microscope. 2.7. Analysis of apoptosis by Bax oligomerization and Hoechst staining of nuclei Transfected cells were scored by single cell assay at 24 h posttransfection for appearance of abnormal nuclei by Hoechst 33258 (Sigma, MO) staining of chromatin, indicative of apoptosis. Scoring of Bax-positive cells was also performed as previously described [24], where MCF-7 cells were fixed in methanol–acetone and stained with Bax monoclonal antibody (Cell Signaling), and scored for intense clustering of Bax fluorescence. 2.8. Duolink proximity ligation assay MCF-7 cells were plated in 8 well chamber slides at 50% confluency 24 h before transfection of YFP-tagged BARD1 using K2 (Biontex Laboratories) transfection reagent. At 48 h post-transfection cells were fixed using 3.7% formalin solution in PBS for 20 min. Cells were then permeabilized with PBS containing 0.2% Triton-X-100 for 10 min. The PLA assay was performed using the commercial Duolink (Olink Bioscience) kit, following supplier instructions to detect interactions between YFP–BARD1 proteins (anti-GFP Ab, Roche, 11814460001) and endogenous p53 (anti-p53 Ab, Santa Cruz Biotechnology, sc-6243), BCL-2 (anti-Bcl-2 Ab #2872, Cell Signaling) or Bax (anti-Bax Ab #2772, Cell Signaling). The total number of PLA interactions per cell was scored in cells that were selected for moderate YFP expression only. Data from two different experiments were combined and analyzed by dot plot using GraphPad Prism software, with at least 100 cells scored per sample and per experiment. In each case, single antibody controls were tested to ensure low background levels and YFP empty vector was used as a control for off-target interaction signals. Single antibody controls are shown in Fig. 2C and are representative of every Duolink PLA experiment shown in this paper. 3. Results and discussion 3.1. Endogenous human BARD1 forms complexes with p53 at mitochondria Previous reports identified binding of p53 to mouse BARD1 [20,25]. To show that human p53 and BARD1 form complexes in MCF-7 breast cancer cells, we employed immunoprecipitation (IP) assays and confirmed detection of endogenous BARD1/p53 complexes in cell lysates (Fig. 1A). A reverse IP using p53 antibody detected capture of BARD1 and 53BP1 (positive control), but not transiently expressed yellow fluorescent protein (YFP; negative control). We next complemented this experiment by using a Duolink proximity ligation assay (Duolink PLA), an in situ protein–protein interaction method using specific primary antibodies and secondary fluorescent-amplification reactions, to show that ectopic YFP–BARD1 formed many complexes with endogenous p53 in MCF-7 cells compared to YFP-only control (see red dots in
Fig. 1. BARD1 forms p53 complexes at mitochondria. (A) Evidence that BARD1 binds p53 at mitochondria. Co-immunoprecipitation assays were performed using total extracts (2 mg) of MCF-7 cells (see Methods). Proteins were immunoprecipitated (IP) with BARD1 polyclonal antibody (left panel) or p53 polyclonal antibody (right panel) and with an IgG control, separated by SDS-PAGE and analyzed by Western blot. Filters were probed with anti-p53 and BARD1 antibodies as shown, revealing a specific association. This data is typical of two separate experiments. (B) Use of Duolink PLA to detect BARD1/p53 complexes in cells. MCF-7 cells were transfected with YFP and YFP–BARD1 plasmids, then 48 h posttransfection were processed with the Duolink PLA (refer to Methods for details) to detect BARD1/p53 interactions. Representative images are shown of YFP transfected cells (green) showing interactions (red dots) between ectopic YFP–BARD1 (detected with GFP primary antibody) and endogenous p53 (detected using p53 antibody) together with nuclear staining in blue. (C) Mitochondrial fractions of MCF-7 cells were immunoprecipitated with BARD1 and IgG control antibody and then probed with BARD1, p53 and mtHsp70 antibodies. The data shown are from two separate experiments. For input lanes, 200 μg of mitochondrial and total protein was loaded as a control.
Fig. 1B; for controls see Fig. 2). In particular, the Duolink PLA data indicated that the majority of YFP–BARD1/p53 complexes formed in the cytoplasm. Since both BARD1 [24] and p53 [27] are each known to locate at mitochondria, we tested for localized complex formation by IP and showed that BARD1 associated with p53 in mitochondria-enriched fractions (Fig. 1C). A known mitochondrial partner protein of p53, the mitochondrial form of HSP70, also appeared to associate with BARD1 at mitochondria (Fig. 1C). These results identify, for the first time, the presence of cytoplasmic BARD1/p53 complexes at mitochondria, and these
1766
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
localized interactions could contribute to the known role of p53 in BARD1 pro-apoptotic function [2]. 3.2. Mapping of BARD1 p53-binding sites: Role of BRCT domain p53 was previously reported to bind a C-terminal region in mouse BARD1 [20]. To further characterize the human BARD1/p53 interaction we expressed a series of C-terminal BARD1 fragments (see Fig. S1) in MCF-7 cells and tested their ability to associate with p53 in order to map the interaction domain. Initial experiments by IP were unsuccessful and therefore the more sensitive Duolink PLA approach was employed. As shown in Figs. 1B and 2B, when compared to background YFP expression alone, the ectopic form of wild-type YFP–BARD1 displayed consistent interactions with p53 as indicated by the numerous red dots (cell images in Fig. 2B) and these were quantified by scoring the red dots in transverse images of transfected cells captured and deconvolved by Deltavision microscopy (graphs in Fig. 2C). The experiments identified three sites of varying affinity in the C-terminus of BARD1, and of these the BRCT domains (551–777) bound p53 most strongly (summarized in Fig. 2A). It is noteworthy that internal deletion of the p53-binding sequence 525–567 caused a modest ~ 50% reduction in detectable BARD1/p53 complexes, whereas deletion of the BRCT domains caused almost complete loss of p53 binding. A similar loss of p53 interaction was observed with the BARD1 cancer mutation, Q564H, providing a strong validation of this assay system as the Q564H mutation was previously shown to disrupt p53 binding by IP [25]. It is not yet clear how a single mutation such as Q564H in the “linker” region can ablate p53 binding, however it might elicit some subtle alteration in BARD1 protein conformation. 3.3. The BRCT domain is required for cytoplasmic localization of BARD1 independent of p53 We previously characterized the nuclear import and export sequences that facilitate BARD1 shuttling between the nucleus and cytoplasm of cells [21,22]. While the p53 binding domains we identified here do not overlap the transport signals, we tested the impact of deleting the p53 binding sites on subcellular localization. A range of YFPtagged BARD1 fragments were transiently expressed in MCF-7 cells, and the cells then imaged by microscopy and scored for nuclear +/− cytoplasmic distribution. None of the internal deletions (i.e. Δ425–525 or Δ525–567) tested altered the nuclear-cytoplasmic equilibrium of BARD1. Unexpectedly, we discovered that deletion of the BRCT domains in the sequence 1–593 caused a striking nuclear accumulation of BARD1 (Fig. 3). This was observed under different fixation conditions in a range of experiments, and is surprising given that this fragment retains the potent nuclear export sequence (NES) at the N-terminus. The data suggest that the C-terminal BRCT domain is somehow required for optimal nuclear export of BARD1, perhaps by mediating a conformation that prevents the NES from being masked by other internal sequences, and thus allowing accessibility by the CRM1 export receptor. Given that p53 binds to the BRCT domain of BARD1 (Fig. 2), we tested whether p53 might contribute to BARD1 localization. However, the overexpression of GFP-p53 or knockdown of p53 had no influence on endogenous BARD1 or YFP–BARD1 in MCF-7 cells (KAM and BRH, unpublished data). Moreover, the Q564H mutation in full-length BARD1, which prevents binding to p53, did not alter the nuclear-cytoplasmic distribution of BARD1 (e.g. see Fig. 4C). Therefore the BRCT domain stimulates cytoplasmic localization of BARD1 independent of p53. 3.4. Loss of p53 binding does not prevent BARD1 mitochondrial localization in MCF-7 cells We next determined whether p53 contributed to the mitochondrial localization of BARD1. First we compared the YFP–BARD1 deletion fragments tested above. The internal deletion of sequences 425–525 or
525–567 did not prevent mitochondrial localization as determined by western blot analysis of mitochondria-enriched fractions of MCF-7 cells (Fig. 4B). As shown above, the C-terminal truncation of the BRCT domain greatly reduced the cytoplasmic pool of BARD1 and hence the amount available to locate at mitochondria, however analysis of the cytoplasmic pool showed that the BARD1 (1–593) mutant still had the capacity to associate with mitochondria (Fig. 4A,B). This mutant no longer binds p53, suggesting that p53 is not required for the targeting of BARD1 to mitochondria in MCF-7 cells. This is further supported by expression of the Q564H mutant, which does not bind p53 but located at mitochondria (Fig. 4B). 3.5. Deletion of the BARD1 BRCT domain/p53-binding site disrupts Bax oligomerization BARD1 apoptotic function is dependent on p53 [25,26]. Hence we investigated whether p53 is required for BARD1 to induce apoptosis through Bax clustering at mitochondria. First we mapped the BARD1 fragments required for Bax oligomerization in MCF-7 cells (Fig. 4A). Cells were transfected with YFP–BARD1 sequences and then stained for Bax and counterstained with Hoechst dye to assess pre-apoptotic changes in nuclear integrity by microscopy (Bax cell images in Fig. 4C left panel; quantification in Fig. 4C right panel graphs; for example of Hoechst staining see Supplementary Fig. S2). Relative to wild-type BARD1, shorter fragments of BARD1 displayed weak Bax-stimulating activity. Notably, removal of the ankyrin repeats (Δ425–525) did not eliminate Bax oligomerization. We tested whether loss of either of the two main p53-binding sites impacted on apoptosis, and observed that deletion of the moderate p53-binding site (see Δ525–567) reduced Bax oligomerization and apoptosis by ~ 30% (Fig. 4C). Of particular note, deletion of the strong p53-binding site within the C-terminal BRCT domain (see 1–593) eliminated Bax oligomerization and apoptosis. Similar results were seen with the non-p53-binding Q564H mutant. These results suggest that the p53-dependence of BARD1 apoptotic activity correlates with binding of p53 to BARD1 C-terminal sequences. Since p53 was not required to transport BARD1 to the cytoplasm or to mitochondria (Figs. 3 and 4), it is possible that p53 interacts with BARD1 at the mitochondria to promote its role in apoptosis. 3.6. BARD1 binds and down-regulates Bcl-2 at mitochondria: a potential mechanism for Bax oligomerization The expression levels of the Bcl-2 family proteins, which consist of anti-apoptotic and pro-apoptotic members, determine life or death of a cell. To investigate possible intracellular mechanisms for YFP– BARD1-induced apoptosis, we tested for binding of BARD1 to the antiapoptotic factor Bcl-2. No interaction was detectable between endogenous BARD1 and Bcl-2 by IP (Fig. 5A), whereas a modest association could be detected using the more sensitive Duolink PLA (Fig. 5B). BARD1 revealed no association with Bax under similar conditions (Fig. 5B, lower panel graph). Next, the expression of Bcl-2 protein was tested. MCF-7 cells were fractionated into total and mitochondrial/ cytosolic extracts and analyzed for endogenous Bcl-2 by Western blot. In total extracts (Fig. 5C), the expression of Bcl-2 remained unchanged after overexpression of YFP–BARD1. In contrast, mitochondrial Bcl-2 expression was moderately down-regulated in two separate experiments (see two panels in Fig. 5C). Transient expression of YFP–BARD1 was confirmed by probing with anti-GFP antibody. Thus, a small pool of BARD1 associates with Bcl-2, possibly through complex formation with p53, and triggers its down-regulation at mitochondria. Bcl-2 is a potent inhibitor of apoptosis and is overexpressed in a wide variety of malignancies [28,29]. In addition, the down-regulation of Bcl-2 has been shown to inhibit tumor growth and induce apoptosis. These results implicate BARD1 as a possible regulator of Bcl-2 and indicate that down-regulation of Bcl-2 may play a role in BARD1 induced
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
1767
Fig. 2. BARD1 binds p53 through C-terminal sequences 525–567 and 551–777. (A) Schematic of BARD1 showing the key domains (NES, NLS, Ankyrin repeats and BRCT domains) and fragments used to map the p53 binding region. A summary of the results from Duolink PLA interaction studies is shown at the right. (B) The PLA method was used to map the BARD1 p53 binding site. MCF-7 cells were transfected with YFP and different YFP–BARD1 plasmids. Representative images are shown of YFP transfected cells (green) showing interactions (red dots) between ectopic YFP–BARD1 (detected with GFP primary antibody) and endogenous p53 (using p53 antibody), with nuclear staining in blue. (C) The total number of Duolink PLA interactions per moderate YFP-expressing cell were scored. Data from at least two different experiments were combined and shown on the Graphpad Prism dot plot, with at least 100 cells scored per sample and per experiment. Levels of significance above background are indicated (**, p ≤ 0.01 and ***, p ≤ 0.001). Note that the two comparative graphs shown each represent at least 2 experiments, but performed 6 months apart. The degree of binding activity for each BARD1 fragment relative to wild-type (WT, set to 100%) is shown below the graphs for ease of comparison.
apoptosis. To test this, we performed a reconstitution experiment to test whether forced expression of Bcl-2 could prevent Bax regulation by BARD1. Indeed, as shown in Fig. 5D, transient expression of Flagtagged BARD1 (wild-type) induced measurable Bax clustering at
mitochondria and this was completely blocked by co-expression of GFP-Bcl-2. In comparison, co-expression of GFP-Bax caused hyperstimulation of oligomeric Bax in BARD1 transfected cells (Fig. 5D). Similar results were observed for apoptosis, and provide support for a
1768
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
Fig. 3. Loss of the BRCT domain induces nuclear accumulation of BARD1. pEYFP-C1 and a series of plasmids that encode YFP-tagged BARD1 sequences were transiently expressed at moderate levels in MCF-7 cells and analyzed for subcellular localization. The cells were fixed and YFP fluorescence detected by microscopy, and the cellular distribution patterns scored (N800 cells were counted per sequence across 3 individual experiments performed in duplicate) and plotted. Note that the truncation of the C-terminal BRCT domain (1–593) is the only deletion to shift BARD1 to the nucleus.
functional link between BARD1, Bcl-2 down-regulation and Bax oligomerization.
4. Conclusions Here we show for the first time that endogenous BARD1 forms p53 protein complexes at mitochondria and identify the tandem BRCT domain as a major binding site for p53. Deletion mapping experiments also revealed a new role for the BRCT domain in driving cytoplasmic localization of BARD1, although this did not appear to depend on binding to p53. We propose that p53 can associate with BARD1 at mitochondria to stimulate certain functions, such as apoptotic activity. BARD1 and p53 are nuclear-cytoplasmic shuttling proteins with the potential to translocate from nucleus to mitochondria to trigger an apoptotic response [21,27,30]. We found that p53 is not essential for targeting of BARD1 to mitochondria but it did contribute to BARD1dependent apoptosis (Fig. 4). Previously, Feki et al. [20] mapped p53binding in mouse BARD1 to the C-terminal sequence 394–604 comprising the ankyrin and linker domains. We show here that in human BARD1 p53 associates with three distinct sequences, the strongest being the tandem BRCT domain (Fig. 2). The glutamine amino acid Q564 lies within both the extended linker (525–567) and BRCT domain (551–777) fragments of BARD1 that displayed strong binding to p53 (Fig. 2). The cancer-linked mutation of this residue, Q564H, abolishes p53 binding, and since this mutation was reported not to alter BARD1
structure [19], it may represent a specific contact site for p53 or affect binding to p53 through a more subtle shift in conformation. We confirmed the interaction between BARD1 and p53 by immunoprecipitation assays, and found that endogenous BARD1 formed in vivo complexes with p53 and another p53-associated factor, mtHsp70, in mitochondrial extracts. Although we have focused on p53 regulation of BARD1, it is anticipated that BARD1 can elicit some reciprocal regulation of p53. This is suggested by earlier studies which found that BARD1 is transcriptionally up-regulated in response to genotoxic stress, leading to increased p53 levels through stabilization and phosphorylation of p53 at serine-15 [2]. Preliminary electron microscopy imaging (see Fig. S3) detected BARD1 at the outer surface and inner matrix of mitochondria in MCF-7 cells, similar to that shown previously for p53 [27]. We speculate that a pool of BARD1 and p53 could attach to the outer mitochondrial surface and get imported together by mtHsp70, although more detailed analyses are required to define precisely where p53/ BARD1 complexes locate within mitochondria. Given the role of BARD1 in nuclear DNA repair [10,11], and involvement of p53 in maintaining mitochondrial DNA base excision repair [31,32], it is conceivable that BARD1 performs a related function in mitochondria separate from (or perhaps linked to) its ability to regulate apoptosis. In response to DNA damage, we observed that BARD1 displays a modest increase at mitochondria [24], suggesting the possibility that BARD1–p53 complexes might increase at mitochondria after DNA damage stress. We also note that a DNA damage-induced phosphorylated pool of BARD1 increases at mitochondria as it does in the nucleus [33].
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
1769
Fig. 4. The BRCT domain stimulates BARD1 mitochondrial targeting and Bax oligomerization. (A) A series of BARD1 deletion fragments were analyzed for mitochondrial localization and ability to induce Bax oligomers and apoptosis. The results are shown summarized from Western blots and IF staining, from negative (−) to positive (+ to ++++, weak to strongest), in comparison to p53-binding ability and the nuclear-cytoplasmic distribution pattern localization. (B) YFP–BARD1 sequences were transiently expressed in MCF-7 cells, and then tested for detection in mitochondrial fractions. Mitochondria-enriched (M) and cytosolic (C) (100 μg/lane) extracts were analyzed by western blotting and YFP–BARD1 was detected using a monoclonal anti-GFP antibody (Roche, 1:1000). All of the BARD1 sequences were detectable in the mitochondrial fraction. Antibodies against mitochondria (mtHsp70), cytosol (tubulin), and nuclear/cytoplasmic (PCNA) markers were used to demonstrate the purity of the fractions. The data presented are typical of two or more experiments. (C) YFP–BARD1 plasmids were expressed in MCF-7 cells, and at 24 h post-transfection cells were stained for Bax (a monoclonal Ab targeting the N-terminus of Bax) and analyzed by fluorescence microscopy (representative cell images shown in left panel, Bax staining in red). Bax clustering is clearly evident above background in cells transfected with wild-type BARD1 and specific BARD1 sub-fragments. The transfected cells were then scored for % cells displaying Bax oligomerization and those that displayed abnormal nuclei after Hoechst staining (see Methods) as shown in right panel.
The damage-dependent phosphorylation of BARD1 occurs in the Cterminal BRCT domain [34], the same sequence shown here to bind p53, and hence it is possible that post-translational modifications influence the action of BARD1–p53 complexes at mitochondria. Transient expression of YFP–BARD1 induces Bax oligomerization at mitochondria (Fig. 4) [24]. In viable cells, Bax is monomeric and located in the cytosol or loosely attached to the outer mitochondrial membrane (OMM). Changes in conformation that expose its N-terminus stimulate oligomerization and its insertion into the OMM, an integral step toward apoptosis. We found that BARD1 expression did not alter Bax protein
levels at mitochondria, suggesting more a post-translational change in Bax folding and self-association than changes in its trafficking. In this regard, the transient expression of BARD1 caused a specific reduction of endogenous Bcl-2 in mitochondria-enriched fractions (Fig. 5). Bcl-2 can inhibit apoptosis by binding to Bax and blocking Bax oligomerization [35]. Previously, p53 was reported to bind Bcl-2 at mitochondria [36], reduce its expression in MCF-7 cells [37], and induce mitochondrial membrane permeabilization [36,38]. More notably, p53 can disrupt the Bcl-2/Bax complex by directly binding to a flexible loop regulatory domain (FLD) in Bcl-2, resulting in enhanced Bax insertion in the OMM
1770
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
Fig. 5. The ability of YFP–BARD1 to induce Bax oligomerization correlates with loss of Bcl-2 at mitochondria. (A) MCF-7 total extract was subjected to immunoprecipitation using BARD1 antibody but did not detect endogenous BARD1–Bcl-2 complexes. (B) A Duolink PLA was used to test for binding between transiently expressed YFP–BARD1 and endogenous Bcl-2 or Bax. Duolink PLA-positive complexes are seen as red dots. Quantification of the number of PLA complexes per cell indicated that YFP–BARD1 showed weak binding to Bcl-2 and no binding to Bax, relative to YFP control. (C) MCF-7 cells, untransfected (unt) or transfected with YFP–BARD1, were fractionated into mitochondrial (M), cytosolic (C) and total extracts 24 h post-transfection and assessed by western blot. Left panel: 60 μg of total extracts was analyzed by western blot, and probed with antibodies to GFP, β-tubulin, Bcl-2, and Bax. β-Tubulin was used as a loading control. Middle and right panels: 100 μg of mitochondrial and cytosolic extract protein was subjected to SDS-PAGE and western blot analysis, and probed with monoclonal antibodies against Bcl-2 and Bax. The two panels shown for Bcl-2 represent two separate experiments. The integrity of the mitochondrial/cytosolic fractions was confirmed by staining with nuclear/cytosolic (PCNA), cytosolic (β-tubulin) and mitochondrial marker (mthsp 70). The transfection of YFP–BARD1 was confirmed by probing with anti-GFP antibody. (D) Bcl-2 overexpression blocks Bax oligomerization. Flag-tagged BARD1 and GFP-tagged Bcl-2 or Bax were expressed in MCF-7 cells alone or in combination, and cells scored after 24 h for Bax oligomerization or apoptosis (Hoechst staining of nuclei). Cell images show that overexpressed BARD1 and Bax, but not Bcl-2, elicit intense Bax staining. Scoring of transfected cells showed that coexpression of GFP-Bcl-2 blocked the induction of Bax and apoptosis by BARD1.
V. Tembe et al. / Cellular Signalling 27 (2015) 1763–1771
and apoptosis [39]. Thus, based on our findings we hypothesize that BARD1 can bind to p53 and Bcl-2 at mitochondria, and under certain circumstances such as cellular stress caused by DNA damage, BARD1 can reduce Bcl-2 expression at mitochondria leading indirectly to a stimulation of Bax oligomerization and apoptosis. Author contribution VT, EM and KAM performed the majority of experiments and contributed to writing of the paper. MM, KB, KM, and YL created specific reagents such as plasmids and performed some of the supporting experiments. EK and RB contributed to the electron microscopy. AD and HR contributed to the conception of the study and writing of the manuscript. BRH designed the study, supervised the project and wrote the paper. Acknowledgments We thank Dr. Richard Baer for the original BARD1 plasmid and Dr Irmgard Irminger-Finger for discussions and encouragement of this work. This work was supported by grants to BRH from the Cancer Council New South Wales (project grant RG06-09) and the National Breast Cancer Foundation of Australia (grant NC-13-13), and to BRH and HR from the Australian Research Council of Australia (ARC; grant DP0881263). BRH is a Senior Research Fellow supported by the Cancer Institute New South Wales (grant 12/CDF/2-12). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.cellsig.2015.05.011. References [1] R. Baer, T. Ludwig, Curr. Opin. Genet. Dev. 12 (2002) 86–91. [2] I. Irminger-Finger, C.E. Jefford, Nat. Rev. Cancer 6 (2006) 382–391. [3] T.H. Thai, F. Du, J.T. Tsan, Y. Jin, A. Phung, M.A. Spillman, H.F. Massa, C.Y. Muller, R. Ashfaq, J.M. Mathis, D.S. Miller, B.J. Trask, R. Baer, A.M. Bowcock, Hum. Mol. Genet. 7 (1998) 195–202. [4] C. Ghimenti, E. Sensi, S. Presciuttini, I.M. Brunetti, P. Conte, G. Bevilacqua, M.A. Caligo, Gene Chromosome Cancer 33 (2002) 235–242. [5] S.-M. Karppinen, K. Heikkinen, K. Rapakko, R. Winqvist, J. Med. Genet. 41 (2004) e114. [6] L.C. Wu, Z.W. Wang, J.T. Tsan, M.A. Spillman, A. Phung, X.L. Xu, M.C. Yang, L.Y. Hwang, A.M. Bowcock, R. Baer, Nat. Genet. 14 (1996) 430–440. [7] P.S. Brzovic, P. Rajagopal, D.W. Hoyt, M.C. King, R.E. Klevit, Nat. Struct. Biol. 8 (2001) 833–837.
1771
[8] V. Joukov, J. Chen, E.A. Fox, J.B. Green, D.M. Livingston, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 12078–12083. [9] M. Fabbro, J.A. Rodriguez, R. Baer, B.R. Henderson, J. Biol. Chem. 277 (2002) 21315–21324. [10] U.K. Westermark, M. Reyngold, A.B. Olshen, R. Baer, M. Jasin, M.E. Moynahan, Mol. Cell. Biol. 23 (2003) 7926–7936. [11] E.E. McCarthy, J.T. Celebi, R. Baer, T. Ludwig, Mol. Cell. Biol. 23 (2003) 5056–5063. [12] H. Ruffner, C.A. Joazeiro, D. Hemmati, T. Hunter, I.M. Verma, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 5134–5139. [13] R. Hashizume, M. Fukuda, I. Maeda, H. Nishikawa, D. Oyake, Y. Yabuki, H. Ogata, T. Ohta, J. Biol. Chem. 276 (2001) 14537–14540. [14] V. Joukov, A.C. Groen, T. Prokhorova, R. Gerson, E. White, A. Rodriguez, J.C. Walter, D.M. Livingston, Cell 127 (2006) 539–552. [15] L.M. Starita, Y. Machida, S. Sankaran, J.E. Elias, K. Griffin, B.P. Schlegel, S.P. Gygi, J.D. Parvin, Mol. Cell. Biol. 24 (2004) 8457–8466. [16] B.R. Henderson, BRCA1 intracellular dynamics, transport and function, Scientifica2012.http://dx.doi.org/10.6064/2012/796808. [17] R. Hakem, J.L. de la Pompa, C. Sirard, R. Mo, M. Woo, A. Hakem, A. Wakeham, J. Potter, A. Reitmair, F. Billia, E. Firpo, C.C. Hui, J. Roberts, J. Rossant, T.W. Mak, Cell 85 (1996) 1009–1023. [18] T.-C. Ayi, J.T. Tsan, L.-Y. Hwang, A.M. Bowcock, R. Baer, Oncogene 17 (1998) 2143–2148. [19] D. Fox III, I.L. Trong, P. Rajagopal, P.S. Brzovic, R.E. Stenkamp, R.E. Klevit, J. Biol. Chem. 283 (2008) 21179–21186. [20] A. Feki, C.E. Jefford, P. Berardi, J.Y. Wu, L. Cartier, K.H. Krause, I. Irminger-Finger, Oncogene 24 (2005) 3726–3736. [21] J.A. Rodriguez, S. Schuchner, W.W. Au, M. Fabbro, B.R. Henderson, Oncogene 23 (2004) 1809–1820. [22] S. Schüchner, V. Tembe, J.A. Rodriguez, B.R. Henderson, J. Biol. Chem. 280 (2005) 8855–8861. [23] C.E. Jefford, A. Feki, J. Harb, K.H. Krause, I. Irminger-Finger, Oncogene 23 (2004) 3509–3520. [24] V. Tembe, B.R. Henderson, J. Biol. Chem. 282 (2007) 20513–20522. [25] I. Irminger-Finger, W.C. Leung, J. Li, M. Dubois-Dauphin, J. Harb, A. Feki, C.E. Jefford, J.V. Soriano, M. Jaconi, R. Montesano, K.H. Krause, Mol. Cell 8 (2001) 1255–1266. [26] M. Fabbro, S. Schuechner, W.W.Y. Au, B.R. Henderson, Exp. Cell Res. 298 (2004) 661–673. [27] N.D. Marchenko, A. Zaika, U.M. Moll, J. Biol. Chem. 275 (2000) 16202–16212. [28] D.G. Breckenridge, D. Xue, Curr. Opin. Cell Biol. 16 (2004) 647–652. [29] A. Gross, J.M. McDonnell, S.J. Korsmeyer, Genes Dev. 13 (1999) 1899–1911. [30] Y. Arima, M. Nitta, S. Kuninaka, D. Zhang, T. Fujiwara, Y. Taya, M. Nakao, H. Saya, J. Biol. Chem. 280 (2005) 19166–19176. [31] D. Chen, Z. Yu, Z. Zhu, C.D. Lopez, Cancer Res. 66 (2006) 3485–3494. [32] N.C. de Souza-Pinto, C.C. Harris, V.A. Bohr, Oncogene 23 (2004) 6559–6568. [33] K. Brodie, B.R. Henderson, Cell. Signal. 22 (2010) 291–302. [34] H.S. Kim, H. Li, M. Cevher, A. Parmelee, D. Fonseca, F.E. Kleiman, S.B. Lee, Cancer Res. 66 (2006) 4561–4565. [35] V. Mikhailov, M. Mikhailova, D.J. Pulkrabek, Z. Dong, M.A. Venkatachalam, P. Saikumar, J. Biol. Chem. 276 (2001) 18361–18374. [36] Y. Tomita, N. Marchenko, S. Erster, A. Nemajerova, A. Dehner, C. Klein, H. Pan, H. Kessler, P. Pancoska, U.M. Moll, J. Biol. Chem. 281 (2006) 8600–8606. [37] S. Haldar, M. Negrini, M. Monne, S. Sabbioni, C.M. Croce, Cancer Res. 54 (1994) 2095–2097. [38] A. Comel, G. Sorrentino, V. Capaci, G. Del Sal, FEBS Lett. 588 (2014) 2600–2609. [39] X. Deng, F. Gao, T. Flagg, J. Anderson, W.S. May, Mol. Cell. Biol. 26 (2006) 4421–4434.