BIG3 Inhibits the Estrogen-Dependent Nuclear Translocation of PHB2 via Multiple Karyopherin-Alpha Proteins in Breast Cancer Cells

We recently reported that brefeldin A-inhibited guanine nucleotide-exchange protein 3 (BIG3) binds Prohibitin 2 (PHB2) in cytoplasm, thereby causing a loss of function of the PHB2 tumor suppressor in the nuclei of breast cancer cells. However, little is known regarding the mechanism by which BIG3 inhibits the nuclear translocation of PHB2 into breast cancer cells. Here, we report that BIG3 blocks the estrogen (E2)-dependent nuclear import of PHB2 via the karyopherin alpha (KPNA) family in breast cancer cells. We found that overexpressed PHB2 interacted with KPNA1, KPNA5, and KPNA6, thereby leading to the E2-dependent translocation of PHB2 into the nuclei of breast cancer cells. More importantly, knockdown of each endogenous KPNA by siRNA caused a significant inhibition of E2-dependent translocation of PHB2 in BIG3-depleted breast cancer cells, thereby enhancing activation of estrogen receptor alpha (ERα). These data indicated that BIG3 may block the KPNAs (KPNA1, KPNA5, and KPNA6) binding region(s) of PHB2, thereby leading to inhibition of KPNAs-mediated PHB2 nuclear translocation in the presence of E2 in breast cancer cells. Understanding this regulation of PHB2 nuclear import may provide therapeutic strategies for controlling E2/ERα signals in breast cancer cells.


Introduction
Prohibitin 1 and 2 (PHB and PHB2) proteins are highly conserved in eukaryotic cells and exhibit diverse subcellular localization with different functions [1][2][3]. These molecules are primarily observed in inner mitochondrial membranes via their N-terminal transmembrane according to the respective depositor's recommendations. The cell line stocks that were used in this study had been properly stored in liquid nitrogen. We monitored the cell morphology of these cell lines by microscopy and confirmed that they maintained their morphology by comparing images with the original morphologic images as described above. No Mycoplasma contamination was detected in any of the cultures using a Mycoplasma Detection kit (Roche, Basel, Switzerland).
Nuclear/cytoplasmic fractionation MCF-7 cells and COS-7 cells were treated as described above, and the nuclear and cytoplasmic fractions were prepared using the NE-PER nuclear and cytoplasmic extraction reagent (Thermo Fisher Scientific) as described previously [8]. α/β-Tubulin and lamin B1 were used as loading controls for the cytoplasmic and nuclear fractions, respectively.
Immunocytochemical staining COS-7 cells were seeded at 1 × 10 5 cells per well (Laboratory-Tek II Chamber Slide System; Nalge Nunc International, Naperville, IL, USA) under E2-free or 10 nM E2 conditions as described below. The COS-7 cells were co-transfected with HA-PHB2, FLAG-KPNAs (KPNA1, 2, 5 or 6) and FLAG-ERα, and then treated with 10 nM E2 for 24 h. The cells were then fixed with phosphate buffered saline (PBS) containing 4% paraformaldehyde at 4°C for 30 min and rendered permeable with PBS containing 0.1% Triton X-100 at 4°C for 2 min. Subsequently, the cells were covered with 3% BSA in PBS for 1 h to block non-specific hybridization followed by incubation with anti-HA antibody diluted at 1:500 for another 1 h. After washing with PBS, the cells were stained with Alexa 594-conjugated anti-rat secondary antibody (Molecular Probe, Eugene, OR, USA) diluted at 1:1,000 for 1 h. The nuclei were counter-stained with 4 0 , 6 0 -diamidine-2 0 -phenylindole dihydrochloride (DAPI). Fluorescent images were obtained with an Olympus IX71 microscope (Olympus, Tokyo, Japan). The nuclear intensity of translocated PHB2 were calculated using MetaMorph software (Molecular Devices, Tokyo, Japan), and are expressed as the ratio of translocated PHB2 at four different points. For endogenous BIG3 staining, MCF-7 cells were incubated with anti-BIG3 diluted at 1:500 followed by with Alexa 488-conjugated anti-rat secondary antibody.

Inhibition of BIG3-PHB2 interactions by the dominant negative peptide ERAP
The 13 amino acid peptides derived from the PHB2-binding domain of BIG3 (codons 165-177) were covalently linked at the NH 2 terminus to a membrane-transducing 11 polyarginine sequence (11R) to construct the ERAP peptide [9]. MCF-7 cells were treated with 10 nM E2 ± 10 μM ERAP. BIG3-PHB2 interactions were assessed using co-immunoprecipitation followed by immunoblotting, as described above.

Identification of interacting regions between PHB2 and BIG3
To identify the BIG3-binding regions of PHB2, HEK293T cells were plated in 6 cm plates at 1 × 10 6 cells per well and individually co-transfected with two different constructs corresponding to partial PHB2 (PHB2 1-189 , PHB2 190-299 ) and BIG3 as described in construction of expression vectors. Forty-eight hours later, the interactions between HA-PHB2 and FLAG-BIG3 were assessed using co-immunoprecipitation followed by immunoblotting, as described above.

Statistical analysis
Statistical significance was calculated using Student's t test to evaluate gene expression. A difference of P < 0.05 was considered statistically significant.

Interactions between PHB2 and KPNAs
Because previous reports have shown that PHB2 has a putative nuclear localization signal (NLS) at 86-RPRK-89 [4,13], we hypothesized that PHB2 is translocated to the nucleus via its interaction with Karyopherin-alpha family (KPNAs, importin-alpha) proteins in breast cancer cells. There are seven known subtypes of KPNA proteins in human cells [11,17]. We first investigated the endogenous expression of KPNAs gene (KPNA1 to 7) in 11 breast cancer cell lines and normal mammary gland tissue via semi-quantitative RT-PCR. The results showed that KPNA1 to 6 was highly expressed in all 11 breast cancer cell lines as well as in normal mammary gland tissue, while focal KPNA7 expression was observed in the SK-BR-3 and HCC1143 cell lines ( Fig 1A). Next, to determine which KPNAs (KPNA1 to 6) interact with PHB2, we constructed plasmids designed to express HA-tagged PHB2 (HA-PHB2) and FLAGtagged KPNAs (FLAG-KPNA1 to 6) (see Materials and Methods). These plasmids were cotransfected into COS-7 cells, and the proteins were immunoprecipitated with anti-FLAG antibody. Immunoblot analysis of the precipitates using anti-HA antibody indicated that HA-PHB2 was strongly bound to FLAG-KPNA1,-KPNA2,-KPNA5 and-KPNA6 and slightly bound to FLAG-KPNA3 and-KPNA4, respectively (Fig 1B). To determine whether the predicted classic NLS (86-RPRK-89) of PHB2 is required for its nuclear translocation, we generated constructs containing NLS mutants of PHB2 in which three conserved amino acids (R86, R88 and K89) had been substituted with alanine, respectively. Then, HA-tagged PHB2 NLS mutants and FLAG-tagged KPNAs (FLAG-KPNA1,-KPNA2,-KPNA5, and-KPNA6) were cotransfected into COS-7 cells, respectively, and the proteins were immunoprecipitated with anti-FLAG antibody followed by immunoblot analysis with anti-HA antibody. Unexpectedly, these substitutions did not abolish the interactions between PHB2 and any of the KPNA proteins, including FLAG-KPNA1,-KPNA2,-KPNA5, and-KPNA6 (Fig 1C). These findings suggest that PHB2 may interact with KPNAs via a non-NLS or one or more novel NLS sequences.

KPNA mediates the nuclear translocation of PHB2
We next investigated the possibility that PHB2 is translocated to the nucleus via the KPNAs in the presence of E2. FLAG-KPNAs (KPNA1, KPNA2, KPNA5, and KPNA6), FLAG-ERα and HA-PHB2 were co-transfected into COS-7 cells, followed (24 h later) by E2 stimulation and immunoblot analysis (see Materials and Methods). The results showed that the overexpression of KPNA1, KPNA5 or KPNA6 but not KPNA2 led to the nuclear translocation of PHB2 in the presence of E2 (Fig 2A). Immunocytochemical staining analysis confirmed that HA-PHB2 localized to the nucleus in the presence of E2 when FLAG-KPNA1,-KPNA5 or-KPNA6 was overexpressed (Fig 2B and 2C). These results suggest that the overexpression of KPNA1, KPNA5, and KPNA6 may induce the nuclear translocation of PHB2 in the presence of E2 and ERα in mammals. However, overexpression of KPNA2 did not. Although we cannot rule out the possibility of KPNA2 as a candidate carrier protein for PHB2 nuclear translocation, in this study we focused on KPNA1, KPNA5, and KPNA6 in further analyses to elucidate the mechanism of PHB2 nuclear translocation via these KPNAs.

Interactions between endogenous PHB2 and KPNAs in breast cancer cells
To further validate the KPNA-mediated E2-dependent nuclear translocation of endogenous PHB2 in breast cancer cells, we examined the knockdown effect of each KPNA via siRNA treatment on the subcellular distribution of endogenous PHB2 in MCF-7 cells after BIG3 knockdown followed by E2 treatment. In the presence of E2, the knockdown of KPNA1, KPNA5 and KPNA6 led to a remarkable decrease in the amount of PHB2 in the nuclei of the BIG3-depleted cells, respectively; however, the knockdown of KPNA2 did not (Fig 3A). Subsequently, we evaluated the possibility of the ERα-dependent nuclear translocation of PHB2 in the presence of E2 in breast cancer cells as previously reported [18]. We first examined the nuclear translocation of ERα and PHB2 after treatment with E2 and ERAP, a dominant-negative peptide inhibitor that inhibits BIG3-PHB2 interactions [8] in MCF-7 cells. We observed that ERAP led to a decrease in cytoplasmic PHB2, thereby substantially increasing the amount of nuclear PHB2 in the presence of E2 in EGFP siRNA-transfected cells (Fig 3B), but did not change the amount of ERα protein in either the cytoplasm or nuclei of EGFP siRNA-transfected cells regardless of treatment with either E2 or ERAP. Moreover, the knockdown of KPNA1, KPNA2, KPNA5, and KPNA6 had no effect on the nuclear translocation of ERα in comparison with that of PHB2 (Fig 3B). We examined the expression of each KPNA in cytoplasm and nuclear fractions of each KPNA-depleted MCF-7 cells. The results showed that knocking down of KPNA1, KPNA5 and KPNA6, but not KPNA2, led to abolish PHB2 nuclear-translocation ( Fig 3C). Furthermore, knocking down of KPNA1 and KPNA6 also led to abolish nuclear-translocation of all of KPNAs, and depletion of KPNA5 led to moderately reduction of nuclear-translocation of all of KPNAs. These findings suggest that the nuclear import of PHB2 released from BIG3 by siRNA or ERAP is mediated by KPNA1, KPNA5 and KPNA6 in an E2-dependent manner in breast cancer cells.
We next investigated whether the endogenous PHB2 forms a complex with KPNA1, KPNA5, or KPNA6 in the nuclear or cytoplasmic fractions of MCF-7 cells after ERAP treatment by co-immunoprecipitation experiments with PHB2 and ERα antibodies. In the presence of E2, ERAP treatment led to a remarkable increase the interactions of PHB2 with KPNA1, KPNA5, and KPNA6 in the cytoplasm and nucleus even after 1h (Fig 4A). In addition, in the presence of E2, the amount of nuclear KPNAs (KPNA1, KPNA5, and KPNA6) which bound to PHB2 was gradually decreased after ERAP treatment in a time-dependent fashion (Fig 4A), suggesting that PHB2 released from BIG3 by ERAP treatment rapidly interacts with KPNAs (KPNA1, KPNA5, and KPNA6) in the cytoplasm, thereby leading to its rapid nuclear translocation, followed by its binding to nuclear ERα. On the other hand, KPNAs (KPNA1, KPNA5, and KPNA6) did not co-immunoprecipitate with nuclear ERα even after ERAP treatment ( Fig  4B), suggesting that KPNAs could be detached from PHB2 immediately after nuclear import of PHB2. More importantly, we also demonstrated that each PHB2-KPNA complex had no effect when only one of the three KPNAs (KPNA1, KPNA5, or KPNA6) was knocked down (Fig 4C), suggesting that the E2-dependent nuclear translocation of endogenous PHB2 is required for its binding to multiple KPNAs (KPNA1, KPNA5, and KPNA6) in breast cancer cells. These findings indicated that PHB2 released from BIG3 by ERAP rapidly interacts with multiple KPNAs (KPNA1, KPNA5, and KPNA6) in the cytoplasm and that this is followed by KPNA-mediated nuclear translocation in the presence of the E2 stimulus.

KPNAs-mediated PHB2 inhibits nuclear ERα transactivation in breast cancer cells
Our previous reports showed that intrinsic PHB2 released from BIG3 by ERAP directly binds to both nuclear-and membrane-associated ERα [8]. We validated the knockdown effect of each KPNA (KPNA1, KPNA5, and KPNA6) on the interactions between endogenous PHB2 and nuclear ERα in BIG3-depleted cancer cells. The results showed that the depletion of only BIG3 led to interactions between endogenous PHB2 released from BIG3 with nuclear ERα but that the depletion of BIG3 and KPNA1, KPNA5, and KPNA6 did not (Fig 5A), indicating that PHB2 binding to nuclear ERα in cancer cells is KPNA-mediated. Similar results were observed with ERAP treatment in the presence of E2 in KPNA-depleted cells, respectively (Fig 5B).  Next, to elucidate the effect of each KPNA on nuclear ERα transcriptional activity, we knocked down BIG3 and all KPNAs and examined the expression of TFF1 (an ERα-target gene) via qRT-PCR. We confirmed the upregulation of both the BIG3 and TFF1 genes (which have been identified as ERα target genes) in E2-stimulated MCF-7 cells (Fig 5C; siEGFP). By contrast, the knockdown of BIG3 expression led to the significant suppression of E2-induced TFF1 expression, whereas knocking down KPNA1, KPNA5, and KPNA6 caused a remarkable up-regulation of TFF1 expression in BIG3-depleted MCF-7 cells, respectively (Fig 5C, S5 Fig). Taken together, our data clearly demonstrate that KPNA1, KPNA5, and KPNA6 primarily regulate the E2-dependent nuclear translocation of endogenous PHB2 released from BIG3 in the presence of E2. Immunoblotting analysis was performed to evaluate the interactions between ERα and PHB2 in BIG3-and KPNA (KPNA1, KPNA5, and KPNA6)-depleted MCF-7 cells. MCF-7 cells were treated with siBIG3 and each siKPNA, followed by E2 ± ERAP for 24 h. Then, the nuclear fractions were immunoprecipitated with anti-ERα antibody and were immunoblotted with antibodies against the indicated proteins. The data are expressed the fold increase over E2-treated siBIG3-transfected cells of right and left panels, respectively (set at 1.0). ND: not detected. This experiment was performed using the nuclear fractions used in Fig 3A; (B) The interaction between ERα and PHB2 released by E2 and ERAP in the nuclear fractions was evaluated. MCF-7 cells depleted of each KPNA were treated with E2 ± ERAP for 24 h, and the nuclear fractions were immunoprecipitated with anti-ERα antibody. The data are expressed the fold increase over E2-treated siEGFP-transfected cells of right and left panels, respectively (set at 1.0). This experiment was performed using the nuclear fractions used in Fig 3B; (C) The TFF1 expression levels following treatment with siBIG3 and siKPNA were evaluated using real-time PCR. The data are expressed as the fold increase over the untreated cells (set at 1.0) and represent the means ± SD of two independent experiments (**P<0.01, ***P<0.001 in a two-sided Student's t-test); (D) Immunoblotting analysis was performed to identify the KPNA-binding regions in PHB2. The lysates from COS-7 cells transfected with the indicated HA-PHB2 constructs and FLAG-KPNAs were immunoprecipitated with an anti-FLAG antibody; (E) Immunoblotting analysis was performed to identify the BIG3-binding region in PHB2. The lysates from HEK293T cells transfected with the indicated HA-PHB2 constructs and FLAG-BIG3 were immunoprecipitated with an anti-

Binding regions of PHB2 and each KPNA
To determine the binding regions of PHB2 and each KPNA, we constructed plasmids designed to express HA-PHB2 deletion mutants-HA-PHB2 (1-189) and , respectively-and FLAG-KPNAs (FLAG-KPNA1, KPNA5, and KPNA6) (see Materials and Methods). These plasmids were co-transfected into COS-7 cells, and the proteins were then immunoprecipitated with anti-FLAG antibody. Immunoblot analysis of the precipitates using anti-HA antibody indicated that HA-PHB2 (1-189) but not HA-PHB2 (190-299) was bound strongly to all FLAG-KPNA1,-KPNA5, and-KPNA6 proteins (Fig 5D). Notably, we demonstrated that HA-PHB2 (1-189) but not HA-PHB2 (190-299) was also bound to the FLAG-BIG3 protein (Fig 5E). Taken together, these data raised the possibility that BIG3 structurally overlay the KPNAs (KPNA1, KPNA5, and KPNA6) binding region(s) of PHB2 (excluding the ERAP binding region), leading to the inhibition of KPNA-mediated PHB2 nuclear translocation in the presence of E2 in breast cancer cells and resulting in constitutive E2-dependent ERα transcriptional activity.

Discussion
Previous studies have shown that PHB2 directly binds to nuclear ERα, resulting in suppression of ERα transcriptional activity by competing with the co-activator SRC-1 to bind ERα [5] and by recruiting histone deacetylase 1 [HDAC1; ref. 6] and a co-repressor, NcoR [5,6], in breast cancer cells, suggesting that PHB2 acts as a co-repressor of ERα. However, this is controversial because endogenous PHB2 is abundantly expressed in ERα-positive breast cancer cells. Therefore, it was unclear how PHB2 is inactivated in ERα-positive cancer cells. Our previous studies demonstrated that endogenous BIG3 mainly localized in cytoplasm (S7 Fig) and sequesters PHB2, thereby causing the loss of function of PHB2 and resulting in constitutive ERα transcriptional activation in breast cancer cells. This information suggests that the BIG3-PHB2 complex plays a critical role in promoting ERα-positive breast cancer cell growth [8][9][10]. However, the mechanism by which BIG3 blocks E2-dependent PHB2 nuclear translocation in breast cancer cells remains unclear. In this study, we demonstrated that KPNAs (KPNA1, KPNA5, and KPNA6) primarily interact with PHB2 released from BIG3 via siRNA or ERAP treatment in the cytoplasm and that this is followed by KPNA-mediated nuclear translocation in the presence of E2 stimulation. Notably, endogenous PHB2 preferentially bound to BIG3 in MCF-7 cells regardless of the abundant presence of KPNAs, suggesting the possibility that endogenous PHB2 has a high affinity to BIG3 protein compared with that to KPNAs in breast cancer cells. In addition, when KPNA1, KPNA5, or KPNA6 was knocked down by siRNA, the inhibition of PHB2 nuclear translocation occurred in the presence of E2 in BIG3-depleted cells, indicating that all three (KPNA1, KPNA5, and KPNA6) are required for the nuclear translocation of PHB2. Interestingly, human KPNA1, KPNA5, and KPNA6 are classified into the same group by their amino acid sequences and share a minimum 80.7% identity [17,19,20]. On the other hand, overexpression of only KPNA1, KPNA5, or KPNA6 also led to PHB2 nuclear translocation in COS-7 cells (Fig 2A and 2B). A possible reason for this discrepancy is due to that each KPNA may interact with PHB2 through a mutual, complementary, or reciprocal relationship in overexpression experiments. Moreover, although nuclear proteins are generally known to be transported into the nucleus by forming complexes with KPNA and KPNB heterodimers [11,12], we found only nuclear-translocation of PHB2 via KPNAs. Therefore, further analyses are needed to elucidate the involvement of endogenous KPNB in PHB2 nuclear-translocation in breast cancer cells (Fig 6). Furthermore, we demonstrated that overexpression of KPNA2 interacts with PHB2 but is not responsible for PHB2 nuclear translocation ( Fig  1B and 1C, Fig 2A and 2B). Accumulating evidence indicates that KPNA2 upregulation is significantly associated with poor prognosis in various human cancers [21][22][23][24]. Notably, cytoplasmic KPNA2 may have an oncogenic role by binding to the cytoplasmic tumor suppressor NBS1, which is involved in the PI3K/Akt signaling pathway. Accordingly, these evidences indicate the possibility that the role of KPNA2 binding to PHB2 may be independent of nuclear transport in breast cancer cells. Although further analyses are also needed to elucidate the detailed mechanisms of this process, when taken together, these findings and ours suggest that KPNA1, KPNA5, and KPNA6 are likely involved in the sufficient nuclear import of functional PHB2 proteins.
Another interesting finding of this study is that PHB2 possibly interacts with KPNAs via its non-NLS sequence(s) despite bearing a putative nuclear localization signal (NLS; 86-RPRK-89), as described previously [4,13]. The binding capability of PHB2 with each KPNA was comparable with that of the Ala-replaced NLS mutant PHB2 (R86A, R88A and K89A) (Fig 1C). Notably, the scores of NLS sequences within PHB2 predicted using the SeqNLS algorithm [25] and cNLS Mapper [26] were substantially less than the cut-off value (data not shown). In PHB2 Nuclear Translocation Mediated by Karyopherin-Alpha Proteins recent years, however, it has become increasingly apparent that a number of proteins do not follow these canonical pathways and instead utilize non-conventional mechanisms [27]. We also noted the potential of non-conventional nuclear import pathways due to the complicated binding structure of PHB2 and the KPNAs.
However, we determined the KPNAs binding region(s) within the 1 to 189th amino acid region of PHB2 through biochemical analyses (Fig 5D). Furthermore, we identified several candidate protein binding sites within the 1 to 189th amino acid region of PHB2 using the PSIVER (Protein-protein interaction SItes prediction serVER) software [28]. More importantly, our biochemical analysis data demonstrated that PHB2 interacts with BIG3 via its N-terminal portion as well as KPNAs (KPNA1, KPNA5, and KPNA6) (Fig 5E). Furthermore, we previously demonstrated that a dominant-negative peptide, ERAP, based on the residues Q165, D168 and Q173 in BIG3, which are essential for its hererodimerization with PHB2, competitively inhibits BIG3-PHB2 interactions [8]. According to these observations, we can propose that the nuclear translocation of PHB2 is mediated by KPNA (KPNA1, KPNA5, and KPNA6) binding via non-NLS or novel NLS sequence(s) within its N-terminal portion (with the exception of ERAPbinding amino acids) (Fig 5E). Taken together, these findings suggest that BIG3 structurally overlays each KPNA binding region of PHB2, thereby blocking PHB2-KPNA interactions and resulting in the inhibition of E2-dependent PHB2 nuclear translocation in breast cancer cells (Fig 6). Further studies to clarify the BIG3 and KPNA (KPNA1, KPNA5, and KPNA6)-binding regions of PHB2 are now underway.
In conclusion, our findings are the first to show that multiple KPNAs (KPNA1, KPNA5, and KPNA6) play key roles in PHB2 nuclear translocation in the presence of E2 stimulation. In breast cancers, however, BIG3 captures PHB2 through its KPNA (KPNA1, KPNA5, and KPNA6)-binding region(s), thereby inhibiting the E2-dependent suppressive ability of PHB2. These findings may shed light on currently unrecognized ERα signaling networks in breast carcinogenesis. Clarifications of the relationship among BIG3, PHB2, and the KPNAs (KPNA1, KPNA5, and KPNA6) will inform the development of novel therapeutic agents targeting protein-protein interactions, thereby enhancing the nuclear import of the tumor suppressor.