RanBPM Has Proapoptotic Activities That Regulate Cell Death Pathways in Response to DNA Damage

The integrity of the genome is under constant threat not only from environmental toxins and radiation but also from by-products of normal cellular metabolism. In response to DNA damage, eukaryotic cells trigger signaling pathways to induce cell cycle checkpoints and establish DNA repair complexes (1, 2). The activation of apoptotic pathways is also an essential component of the DNA damage response, and defects in the activation of apoptotic pathways or in the apoptotic machinery lead to genomic instability (3). Conversely, chromosomal instability favors the inactivation of apoptotic pathways to select for resistant cells, and tumors often harbor inactivating mutations of genes that encode proapoptotic factors (BAX for instance) or factors that are involved in apoptosis regulation, such as p53 (4, 5). Thus, identifying the factors and mechanisms that link DNA repair and apoptosis is fundamental to our understanding of tumorigenesis and to develop strategies for the prevention of cancer development.

The intrinsic mitochondrial pathway is the primary pathway activated in response to DNA damage. The central event in this pathway is mitochondrial membrane depolarization, which is controlled by Bcl-2 family factors (6, 7). Antiapoptotic family members (such as Bcl-2, Mcl-1, and Bcl-X_L) prevent apoptosis by sequestering and neutralizing the proapoptotic members (such as Bax, Bad, Noxa, and PUMA) through direct interaction. Proapoptotic factors activate membrane permeabilization, releasing mitochondrial intramembrane proteins, such as cytochrome c, which initiate caspase-dependent apoptosis. Thus, the balance between cell life and death depends on the relative level of expression of proapoptotic and antiapoptotic members.

Ran-binding protein M (RanBPM; also called RanBP9) is a nucleocytoplasmic protein whose function remains largely unknown. Several recent reports have suggested that RanBPM contributes to the regulation of various cell signaling functions, including cell adhesion and migration (8-11), microtubule regulation (12, 13), as well as the regulation of gene transcription (14, 15). There is also evidence for RanBPM involvement in signaling pathways elicited by environmental signals. RanBPM is a phosphoprotein whose phosphorylation is modulated by stress stimuli such as osmotic shock and UV radiation (16). RanBPM is also phosphorylated in response to ionizing radiation (IR) at a consensus site recognized by the DNA damage–activated kinases ataxia telangiectasia mutated (ATM), ataxia telangiectasia mutated related (ATR), and DNA-dependent protein kinase (DNA-PK; ref. 17). In addition, the participation of RanBPM in apoptotic signaling pathways was suggested based on the ability of RanBPM to interact with cyclin-dependent kinase 11 isoform p46, a protein implicated in apoptotic signaling cascades (18). In addition, RanBPM was found to interact with the death domain of p75 neurotrophin receptor, a factor member of the tumor necrosis factor receptor family mediating programmed cell death in neurons (19). Finally, the association of RanBPM with homeodomain-interacting protein kinase 2 and with p73 has also been suggested to modulate DNA damage–induced apoptotic pathways (20, 21). However, the functional consequences of these interactions on the regulation of apoptosis remain largely unexplored.

Here, we further the link between RanBPM and apoptosis by establishing RanBPM as an activator of apoptotic pathways. We determined that RanBPM expression in HeLa cells activates caspase-3 activity and induces cell death. Consistent with these proapoptotic capabilities, small interfering RNA (siRNA)–mediated downregulation of RanBPM compromised the induction of apoptosis and increased cell survival in response to IR. Cells expressing reduced levels of RanBPM also showed altered Bax mitochondrial localization and increased levels of Bcl-2. In addition, endogenous RanBPM was found to translocate from the nucleus to the cytoplasm following IR treatment. These results suggest that RanBPM is a DNA damage–activated factor with proapoptotic activities capable of regulating the intrinsic cell death pathway.

We initially identified RanBPM in a yeast two-hybrid screen as interacting with octamer factor 1 (Oct-1), a transcription factor previously characterized as a regulator of cell survival in response to DNA damage (22, 23). Intrigued by preliminary reports implicating RanBPM in signaling pathways elicited by DNA damage (16, 17, 20), we decided to investigate a potential role for RanBPM in the DNA damage response.

In initial transfection experiments, we determined that RanBPM ectopic expression in HeLa cells triggered significant cell death 24 to 48 hours after transfection, which was not observed on transfection of similar constructs expressing different cDNAs using identical transfection conditions (data not shown). To confirm a potential effect of RanBPM ectopic expression on cell viability, we expressed RanBPM from a neomycin selection–containing expression vector and selected transfected cells with G418 for several days (Fig. 1A). Whereas transfection of the vector alone gave rise to numerous G418-resistant colonies, cells transfected with RanBPM produced far less colonies (20-30 times less), supporting the notion that increased expression of RanBPM reduced cell viability. To determine if RanBPM-induced cell death was due to increased apoptotic activity, we measured caspase activity in RanBPM-transfected HeLa cells using a caspase-3 substrate. A strong induction of caspase activity was observed in cells transfected with RanBPM, but not in cells transfected with vector alone, suggesting that RanBPM expression activates apoptotic pathways (Fig. 1B). Additional caspase analyses done using extracts from cells transfected with EGFP and Oct-1 expression constructs confirmed that apoptosis was not appreciably induced by overexpression of these proteins, but that the effect was restricted to RanBPM ectopic expression (Fig. 1C). Finally, to confirm that cell death was triggered specifically by hemagglutinin (HA)-RanBPM expression, we performed indirect immunofluorescence experiments. Cells showing condensed nuclei, typical features of apoptosis, also showed expression of HA-RanBPM detected with an HA antibody (Fig. 1D), further linking RanBPM overexpression with cell death. Control experiments verified that HA staining in apoptotic nuclei was only observed in transfected cells and that the expression of HA-RanBPM, and not of another HA-expressing protein (HA-Oct-1), was responsible for this effect (Supplementary Fig. S1). Altogether, these results suggest that increased expression of RanBPM activates apoptotic pathways.

RanBPM displays three conserved motifs: a SPRY domain, which is thought to mediate protein-protein interactions; a LisH [Lissencephaly type 1 (LIS1) homologous] motif known to mediate protein dimer and tetramer formation; and a CTLH (C-terminal to LisH) motif of yet unknown function (refs. 24, 25; Fig. 2A). Interestingly, the COOH-terminal 100 amino acid residues of RanBPM are highly conserved between known RanBPM homologues and are predicted to adopt six-helical elements reminiscent of the death domain superfamily (12, 26).

To start investigating the protein motifs involved in RanBPM proapoptotic activity, we produced a RanBPM deletion mutant lacking four of the six helices present in the COOH-terminal domain (Fig. 2A). This mutation greatly compromised the ability of RanBPM to induce apoptosis and cell death, although it did not affect the level of expression of the protein (Fig. 2B and C). Strikingly, opposite effects were observed with a RanBPM mutant lacking the NH₂-terminal domain (mutant ΔN 252-729, Fig. 2D), as this mutation stimulated the ability of RanBPM to promote caspase-3 activity although its expression levels were severely reduced compared with full-length RanBPM (Fig. 2E). Interestingly, deletion of amino acids 1-101 (ΔN2 mutant) also had a detrimental effect on protein expression, albeit less pronounced than that of the ΔN 252 mutation (Fig. 2E). However, this deletion did not seem to enhance the ability of RanBPM to induce apoptosis as was observed with ΔN 252-729 (Fig. 2D and E). This suggested that the SPRY domain, which is partially deleted in ΔN 252-729, might be restraining RanBPM proapoptotic activity. Analysis of the subcellular localization of these mutants revealed no change in their subcellular localization compared with wild-type RanBPM (data not shown). Altogether, these results suggested the potential for RanBPM to act as a proapoptotic factor.

Whereas our results suggested the possibility of a proapoptotic role for RanBPM, we proposed to determine if endogenous RanBPM could fulfill such a function in response to proapoptotic stimuli. To investigate a physiologic role for RanBPM in the activation of apoptotic pathways, we assessed the effect of RanBPM downregulation on the apoptotic response induced by DNA damage (Fig. 3). The substantial reduction in RanBPM expression obtained through transient siRNA transfection resulted in a marked decrease of caspase-3 activation in response to 10 Gy of IR (>2.5-fold decrease at 96 and 120 hours; Fig. 3B). The activation of caspase-2, a stress- and DNA damage–induced caspase, was also severely blunted by RanBPM downregulation, with >2-fold reduction of activation at 96 and 120 hours (Fig. 3C). This effect seemed to be specific to RanBPM because we obtained a similar result with a second RanBPM siRNA targeting a different region of RanBPM mRNA (Supplementary Fig. S2). To determine if this decrease in caspase activation correlated with an effect on cell viability, we generated a stable cell line expressing RanBPM siRNA from the pSuper.Neo expression vector in which RanBPM expression was found to be efficiently downregulated (Fig. 4A). We measured the relative number of cells undergoing apoptosis (apoptotic index) by examination of apoptotic cell morphology following staining with Hoechst 33342 in control and RanBPM short hairpin RNA (shRNA) cells following IR treatment (Fig. 4A). Consistent with the reduction of caspase activity observed in our previous assay, we observed a significant decrease in apoptosis induced by DNA damage in RanBPM-deficient cells compared with control shRNA cells. This correlated with an increased overall survival of the RanBPM shRNA–expressing cells in colony-forming assay (Fig. 4B), further showing the protective effect of RanBPM downregulation on cell survival following DNA damage. The protective effect conferred by RanBPM downregulation was particularly apparent at high doses of IR, with >5-fold increase in survival observed at 8 Gy (5.3 ± 2.9 for control versus 29.7 ± 8.1 for RanBPM siRNA) and >10-fold at 12 Gy (1.3 ± 0.6 versus 24.6 ± 6.4). Finally, we verified that the reintroduction of RanBPM in the shRNA-downregulated cell line, through ectopic expression of a RanBPM cDNA bearing a point mutation in the sequence targeted by the shRNA, reinstated caspase activation in response to IR, confirming the specificity of RanBPM function in the activation of apoptotic pathways (Fig. 4C). Thus, these results suggest that RanBPM deficiency prevents the activation of apoptosis in response to DNA damage and promotes cell survival.

Western blot analysis revealed that RanBPM protein levels are not appreciably affected by IR treatment (see Fig. 6A), ruling out the modulation of RanBPM protein levels as a primary mechanism of its function in the apoptotic process. Thus, we turned our attention to a potential effect of DNA damage on RanBPM subcellular localization. Endogenous RanBPM is present in both nucleus and cytoplasm, but shows predominantly nuclear staining in proliferating HeLa cells (ref. 27; Fig. 5A). To verify the specificity of the antibody, we stained HeLa cells expressing RanBPM siRNA, which confirmed that RanBPM downregulation correlated with reduced or lack of staining in these cells (Fig. 5A). Following IR treatment of HeLa cells, we observed a change in the subcellular distribution of RanBPM, suggesting a relocalization of RanBPM from the nucleus to the cytoplasm (Fig. 5A). This relocalization was apparent at 24 hours and persisted at 48 and 72 hours after IR treatment and seemed to affect most cells, albeit to various degree. Quantification of these results confirmed the occurrence of a major shift of the nuclear pool of RanBPM to the cytoplasmic compartment within the first 48 hours following IR treatment (Fig. 5B). This relocalization of RanBPM to the cytoplasm was corroborated by Western blot analysis of cytoplasmic and nuclear fractions (Fig. 5C and D). We also observed a similar IR-induced relocalization of RanBPM in two other cell lines, HCT116 and HEK-293, indicating that this response was not cell type specific or a particularity of HeLa cells (Fig. 7D and E; Supplementary Fig. S3). Thus, the activation of apoptotic pathways by RanBPM in response to DNA damage may be regulated by nucleocytoplasmic trafficking.

IR-induced DNA damage activates apoptosis primarily through the mitochondrial pathway, which is controlled by the Bcl-2 protein family (7, 28). We sought to examine the mechanism of RanBPM proapoptotic function by evaluating the effect of RanBPM expression, or lack thereof, on members of the Bcl-2 family. We initiated this analysis by testing the expression and localization of Bax, which is central to the activation of the mitochondrial pathway activated by DNA damage (28). In most cell types, in response to apoptotic stimuli, Bax translocates to the mitochondrial membrane, oligomerizes, and triggers the release of cytochrome c and other proteins present in the intermembrane space (6, 29). However, in some cell types, including HeLa cells, Bax mitochondrial levels do not increase during apoptosis, and Bax activation occurs through its oligomerization and insertion in the mitochondrial membrane (30, 31).

Analysis of whole-cell extracts from control shRNA and RanBPM shRNA cells showed that Bax protein levels were unaffected by RanBPM downregulation (Fig. 6A). However, analysis of heavy membrane fractions revealed that the levels of Bax associated with the mitochondria were strikingly decreased in extracts from RanBPM shRNA cells compared with control shRNA cells (Fig. 6B). Because Bax localization and activation are largely dependent on Bcl-2 regulation (7, 32), we analyzed Bcl-2 levels in cells exhibiting reduced expression of RanBPM. Bcl-2 protein levels were found markedly increased in RanBPM shRNA cells, in both cytoplasmic fractions and whole-cell extracts (Fig. 6C and D). This increase was observed in two independent RanBPM siRNA clonal derivatives (Supplementary Fig. S4). Further, we confirmed that the increase of Bcl-2 protein levels was triggered by the lack of RanBPM expression because reintroduction of RanBPM in the RanBPM shRNA cells (via RanBPM si-mt transfection) reduced Bcl-2 protein levels close to those observed in control HeLa cells (Fig. 6D). Thus, RanBPM proapoptotic activity is mediated, at least in part, by its regulation of factors of the mitochondrial apoptotic pathway.

Finally, to verify that the effects of RanBPM depletion were not due to a HeLa cell–specific mechanism, we produced HCT116 clonal derivatives expressing RanBPM shRNA. As for HeLa cells, downregulation of RanBPM in HCT116 cells correlated with a strong increase of Bcl-2 protein levels (Fig. 7A). Further, RanBPM depleted HCT116 cells also displayed decreased caspase activation in response to IR (Fig. 7B) and increased survival as measured by colony assay (Fig. 7C), confirming that RanBPM also regulates apoptotic pathways in these cells.

We present evidence that RanBPM functions as an activator of the mitochondrial apoptotic pathway. By the use of two complementary approaches, overexpression and downregulation, we showed that RanBPM regulates the activation of caspases, modulates the levels and localization of Bcl-2 family members, and regulates cell survival in response to DNA damage. Further, we determined that RanBPM intracellular localization is actively regulated in response to IR, further suggesting a role for this factor in the cellular response to DNA damage.

We have identified two regions of RanBPM that regulate its proapoptotic function in opposite ways. First, we found that the extreme COOH terminus of RanBPM is required for RanBPM proapoptotic activity, as deletion of this domain prevented caspase activation mediated by the full-length RanBPM protein. This domain was previously determined to interact with fragile X mental retardation protein and named CRA (for CT11-RanBPM; ref. 12). Sequence alignment revealed this domain to be extremely conserved in RanBPM homologues and predicted to adopt a six-α-helical structure reminiscent of death fold domains (12, 26). Interestingly, in addition to fragile X mental retardation protein, RanBPM COOH-terminal domain was reported to interact with the death domain of p75 neurotrophin receptor, a factor member of the tumor necrosis factor receptor family mediating programmed cell death in neurons (19). Thus, this domain may be involved in the activation of proapoptotic factors through protein-protein interactions.

Conversely, the SPRY domain may be functioning to restrain RanBPM proapoptotic activity, as partial deletion of this domain stimulated the ability of RanBPM to activate caspase activity, even though the expression of the resulting mutant was severely reduced. We noted that deletion of the first 100 amino acids of RanBPM (RanBPM ΔN 101) also resulted in reduced expression of the mutant protein compared with full-length RanBPM, suggesting that this region contains sequences important for protein folding and/or stability. Thus, the strong decrease in ΔN 252 protein expression may be partially due to the deletion of the extreme NH₂-terminal domain, but may also result from the loss of the repressive function of the SPRY domain. Because this deletion confers stronger proapoptotic activities, most cells expressing this mutant would be eliminated, resulting in the apparent low level of expression. In support to this, immunofluorescence analysis indicated a dramatic reduction in the number of cells expressing this mutant, rather than a reduction in the overall level of expression (data not shown). SPRY domains are involved in protein interaction in a wide variety of cellular functions (25). Interestingly, an inhibitory role for a SPRY domain in caspase activation has previously been documented. The SPRY domain (also called B30.2 domain) of Pyrin, a member of the pyrin domain (PYD) subfamily of the death domain superfamily, has been shown to interact with caspase-1 and other components of the inflammasome (33, 34). Recently, the Pyrin SPRY domain was reported to exert an inhibitory effect on caspase-1 activation and subsequent pro-interleukin-1β processing (34). This suggests the possibility of a similar negative regulation of yet undefined caspase(s) by the SPRY domain of RanBPM.

Whereas overexpression of RanBPM triggered caspase activation, downregulation of RanBPM prevented the activation of apoptotic pathways in response to DNA damage, thus establishing a function for RanBPM in the activation of apoptosis in response to genotoxic insults. RanBPM depletion had a marked effect on the overall cell survival in response to IR. This raises the possibility that RanBPM may also function to regulate other forms of cell death such as necrosis and autophagy that can also be triggered, in addition to apoptosis, in response to DNA damage (35). Endogenous levels of RanBPM were not increased in response to IR, but RanBPM subcellular localization was found to shift progressively from a mainly nuclear to a mostly cytoplasmic localization. Several factors that control cell proliferation and survival as well as those involved in DNA damage responses have been shown to be regulated through nucleocytoplasmic shuttling (36-38). Nucleocytoplasmic transport regulates not only protein localization but also protein function. Further investigations will be needed to determine if RanBPM export plays an active role in the IR-induced cellular response or is a mere consequence of the DNA damage response. However, because RanBPM functions to regulate the activation of apoptotic pathways, it is tempting to speculate that RanBPM is exported to the cytoplasm to activate factors of the apoptotic machinery. Alternatively, its presence in the nucleus may be preventing the activation of apoptosis in the absence of stimulus, and this repressive effect would be relieved through its cytoplasmic export. As RanBPM is phosphorylated in response to stress and genotoxic stimuli (16, 17), these modifications could be the signal to promote its transport out of the nucleus and/or could also activate its apoptotic function.

Thus far, and apart from caspase-3, we have found three proteins whose change of expression, activity, or localization can account for the apoptotic defects observed on RanBPM downregulation: Bcl-2, Bax, and caspase-2. Bcl-2 is a primary regulator of the mitochondrial apoptotic pathway, and its overexpression impairs the activation of apoptosis and is linked to cancer development (4, 32). Bcl-2 functions by interacting with and neutralizing members of the proapoptotic family. In particular, Bcl-2 has been shown to prevent Bax translocation and oligomerization, either through direct binding or by preventing its activation through BH3-family factors (6, 30, 39, 40). Thus, the increase in Bcl-2 expression could explain the decrease in mitochondrial Bax observed in RanBPM-downregulated cells. Consistent with previous observations, Bcl-2 upregulation was not associated with an increase in Bax overall levels (41).

Bcl-2 levels have been shown to be regulated by several mechanisms and pathways. First, Bcl-2 expression is regulated at the transcriptional level by several transcription factors including cyclic AMP–responsive element binding protein, p53, nuclear factor κB, Sp1, and Oct-1 (42-44). Bcl-2 can also be regulated by posttranslational modifications such as phosphorylation and ubiquitination (45). In turn, several studies have implicated RanBPM in transcriptional regulation (14, 15, 46), whereas others have linked RanBPM to the regulation of posttranslational modifications such as phosphorylation, ubiquitination, and sumoylation (10, 20, 46). Thus, RanBPM could control Bcl-2 transcriptionally and/or posttranslationally. We noted a steady decrease of Bcl-2 protein levels in response to IR in RanBPM shRNA cells. Such decrease was previously suggested to be causally involved in apoptosis activation (47-49). Yet, this IR-dependent (RanBPM-independent) Bcl-2 downregulation is not sufficient to fully activate apoptotic pathways in RanBPM-deficient cells, suggesting that RanBPM also affects the activation of other proapoptotic pathways.

Interestingly, we found that caspase-2 activation in response to IR was defective in the absence of RanBPM. Several studies indicate that caspase-2 activation is a requirement for mitochondrial membrane permeabilization and the apoptotic response induced by various agents, including DNA damage and H₂O₂ (50-52). Caspase-2 has been reported to contribute to Bax translocation and oligomerization at the mitochondrial membrane (51, 53). However, the signals controlling caspase-2 activation are still largely undefined, and caspase-2 activation has been found in turn dependent on and independent of Bcl-2 factors regulation (54-57). Further investigation will be needed to understand whether RanBPM controls the activation of both factors or if they are regulated by each other sequentially.

In conclusion, we have characterized a novel proapoptotic function for RanBPM and revealed a critical role for this factor in the activation of cell death pathways triggered by DNA damage. Our results predict that RanBPM inactivation would enable cells with genomic alterations to escape cell death, therefore allowing the propagation of potential oncogenic mutations. Interestingly, decreased and even sometimes undetectable levels or altered pattern of RanBPM expression was previously observed in cancer cells from several tumor samples, suggesting that downregulation of this protein accompanies cancer development (16). Although further investigations will be needed to explore the molecular details of RanBPM action, our current finding suggest that RanBPM could be an important regulator of pathways that prevent tumorigenesis by promoting the elimination of cells with genomic alterations.

pCMV-HA-RanBPM (a gift of Dr. Mark Nelson, Department of Surgery, University of Arizona, Tucson, AZ) was described in ref. 18. pcDNA3-FLAG-RanBPM was generated by subcloning of the full-length RanBPM cDNA from pCMV-HA-RanBPM into pcDNA3-FLAG. RanBPM ΔN and ΔC mutant constructs were generated in pCMV-HA-RanBPM by restriction enzyme digests. pCMV-HA-RanBPM shRNA mutant construct (HA-RanBPM si-mt) was generated by introducing two silent point mutations at nucleotides 2,152/2,153 (TC to AG, numbering with respect to the first ATG of the RanBPM cDNA) in pCMV-HA-RanBPM by site-directed mutagenesis using PfuTurbo (Stratagene). pEGFP-C1 is from Clontech; pCGN-Oct-1 has been described elsewhere (22).

RanBPM siRNAs were purchased from Ambion: control siRNA, RanBPM siRNA#2 (upper strand: 5′-GGAAUUGGAUCCUGCGCAU-3), and RanBPM siRNA#1 (upper strand: 5′-GGCCACACAAUGUCUAGGA-3). pSuper-shRanBPM was generated by subcloning an oligonucleotide corresponding to RanBPM siRNA #2 (Ambion) into pSuper (Oligoengine) following the company's protocol. Similarly, pSuper-shControl was obtained by subcloning control shRNA (Ambion) in pSuper.

HeLa, HCT116, and HEK-293 cells were cultured in high-glucose DMEM supplemented with 10% fetal bovine serum at 37°C in 5% CO₂. All cell lines were obtained from the American Type Culture Collection. For irradiation experiments, cells were plated the night before irradiation at 50% to 60% confluency. Irradiations were done with a Faxitron RX-650 at a dose rate of 1.42 Gy/min.

Plasmid transfections in HeLa cells were carried out with ExGen 500 (MBI Fermentas). siRNA duplexes were transfected with siPORT NeoFX (Ambion) following the manufacturer's instructions, with a final siRNA concentration of 20 nmol/L. For clonal selection, G418 (geneticin, Bioshop Canada) was added to the media 24 h after transfection and carried out for 10 to 14 d before colony isolation.

Whole-cell extracts were prepared as described (22). For subcellular fractionations, cells were scraped and washed in ice-cold PBS and lysed in mitochondrial lysis buffer (20 mmol/L HEPES, 1 mmol/L EGTA mmol/L 1 mmol/L EDTA, 10 mmol/L KCl, and 1.5 mmol/L MgCl₂) with 50 strokes of a dounce homogenizer. After centrifugation, the pellet (nuclei) was incubated in nuclear buffer (20 mmol/L HEPES, 25% glycerol, 450 mmol/L NaCl, 1.5 mmol/L MgC₂, and 0.2 mmol/L EDTA) and centrifuged to collect the nuclear fraction. The supernatant was centrifuged at 10,000 × g for 20 min to collect the supernatant (cytoplasmic fraction). The pellet (heavy membrane) was washed in mitochondrial lysis buffer and resuspended in 1% CHAPS buffer (50 mmol/L Tris-HCl, 110 mmol/L NaCl, 50 mmol/L HEPES, 10% glycerol, and 0.5% NP40), incubated on ice for 15 min, and centrifuged to collect the mitochondrial fraction.

For Western blot analysis, extracts were resolved by SDS-PAGE (between 8% and 12%). Gels were transferred onto a polyvinylidene difluoride membrane and hybridized with the following antibodies: RanBPM 5M (ref. 27; Bioacademia, Japan), β-actin (I-19, Santa Cruz, CA), Bax (N-20, Santa Cruz), Bcl-2 (Cell Signaling), HA (HA-7, Sigma), γ-tubulin (a kind gift from Dr. Litchfield, University of Western Ontario, London, Ontario, Canada), Ku70 (AB-4, NeoMarkers), proliferating cell nuclear antigen (clone PC-10, Millipore), or cytochrome c oxydase IV (Cell Signaling). The blots were developed using an enhanced luminol reagent (Renaissance, NEN Life Sciences).

Cell extracts were prepared in lysis buffer [1 mmol/L KCl, 10 mmol/L HEPES (pH 7.4), 1.5 mmol/L MgCl₂, 1 mmol/L DTT, 1 mmol/L phenylmethylsulfonyl fluoride, 5 μg/mL leupeptin, 2 μg/mL aprotinin, and 10% glycerol]. Caspase activity was measured as previously described (58) in caspase assay buffer [25 mmol/L HEPES, (pH 7.4), 10 mmol/L DTT, 10% sucrose, 0.1% CHAPS containing either 10 μmol/L caspase-3 substrate, N-acetyl-Asp-Glu-Val-Asp-(7-amino-4trifluoromethyl-coumarin (DEVD-AFC), or 10 μmol/L caspase-2 substrate, N-acetyl-Val-Asp-Val-Ala-Asp-AFC (VDVAD-AFC; BIOMOL International, L.P.)]. The fluorescence produced by DEVD-AFC cleavage was measured on a SpectraMax M5 fluorimeter (excitation 400 nm, emission 505 nm) over a 2-h interval. Caspase activity was calculated as the ratio of the fluorescence output in treated samples relative to corresponding untreated controls. For caspase assays of transfected samples, mock-transfected cells were used as controls.

Apoptosis was assessed by analyzing nuclear morphology in Hoechst 33342 stained cells. Cells were stained live with Hoechst 33342 (1 μg/mL; Sigma-Aldrich) and were visualized with a fluorescence microscope (IX70, Olympus). Images were captured with a charge-coupled device camera (Q-imaging) using Northern Eclipse software (Empix Imaging). A minimum of 500 cells were counted for each sample analyzed, and the fraction of cells displaying an apoptotic nuclear morphology (chromatin condensation, nuclear blebbing, and/or fragmentation) was determined.

For clonogenic assays, cells were plated at single-cell density (200-1,000 per 6-cm dish), irradiated 6 to 8 h after plating, and incubated for 10 to 14 d to allow for colony growth. Colonies were fixed and stained with crystal violet. Colonies of at least 50 cells were scored as survivors. The number of colonies of irradiated samples was normalized to that of unirradiated controls. For crystal violet staining, the cells were washed with PBS, stained with 0.5% crystal violet in 20% methanol, and rinsed thrice with PBS.

Cells were plated on coverslips and incubated overnight or treated as described in the figure legends. Cells transfected with pCMV HA-RanBPM were grown on poly-l-ornithine–treated coverslips, incubated with 20 mmol/L BAF (Boc-D-FMK), and fixed 30 h after transfection. Cells were fixed with 3% paraformaldehyde, permeabilized in 0.5% Triton X-100 for 10 min, and preblocked in 5% fetal bovine serum diluted in PBS. Coverslips were incubated overnight with primary antibodies, RanBPM (Ab 5295, Abcam) or HA (HA-7, Sigma). After several washes in PBS, samples were incubated with secondary antibodies rabbit anti-goat Alexa Fluor 594 (Invitrogen, A-11080) or donkey anti-mouse Alexa Fluor 555 (Invitrogen, A31570). Cells were mounted with ProLong Gold antifade with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen). Visualization was done with an Olympus BX51 microscope with a 40× objective, and images were captured with the Image-Pro Plus software (Media Cybernetics, Inc.). For quantification analysis, coverslips were blinded by a third party and coded images were scored independently by two individuals. Immunofluorescent cells were classified into three categories (N > C, N = C, and C > N). For each time point, at least 100 cells per sample were scored by each individual at each time point and results were averaged from the two independent counts.

Differences between two groups were compared using unpaired two-tailed t test, and ANOVA was used when comparing multiple groups. Results were considered significant when P < 0.05.

No potential conflicts of interest were disclosed.

We thank Mark Nelson for providing the pCMV-HA-RanBPM plasmid, Sean Cregan and Michael Poulter for providing protocols, antibodies, and advice, and Stephen Ferguson and Sean Cregan for critical comments on the manuscript.