Nuclear Survivin Has Reduced Stability and Is Not Cytoprotective*

Survivin is an essential mitotic protein that is overexpressed in many cancers, and its presence is correlated with increased resistance to radiation and chemotherapy. Here we demonstrate that sending survivin into the nucleus accelerates its degradation in a cdh1-dependent manner, abolishes the radio resistance normally conferred to cells by its overexpression, and prevents survivin from inhibiting apoptosis without affecting its mitotic localization. Our data suggest that targeting survivin to the nucleus provides an efficient means of eliminating it from the cell and may prove a novel strategy in cancer treatment, particularly in combination with radiotherapy.

Survivin is an essential mitotic protein that can also inhibit apoptosis. It is up-regulated in the vast majority of human cancers, and unlike in normal proliferating cells, in cancer cells it can be present throughout interphase, indicating a loss of cell cycle regulation. Deregulated survivin expression has been reported at both mRNA and protein levels and correlates with increased resistance to radio-and chemotherapies. In tumor biopsies, survivin has been localized to the nucleus and cytoplasm or both, and a number of studies have implied that differences in patient prognosis correlate with differences in nuclear or cytoplasmic compartmentalization. However, there is no clear consensus from these studies (1).
We and others have recently shown that survivin is a nuclear cytoplasmic shuttling protein that is predominantly cytoplasmic in part because of an active nuclear exportation signal (NES) 3 in its linker region (2)(3)(4)(5)(6)(7)(8). Importantly, unlike wild type survivin, NES mutants are unable to protect cells against Xirradiation or TNF-related apoptosis-inducing ligand (TRAIL)induced apoptosis (4), suggesting that relocating survivin to the nucleus during interphase may be key to inhibiting its cytoprotective activity. These data highlight the importance of regulating, not only the level, but also the localization of survivin in cancer cells.
To date, any link between subcellular compartmentalization and survivin stability has not been addressed; such regulation would have implications in terms of protein behavior, both in the etiology of tumorigenesis and in the design of chemotherapeutic targeting of the protein. To examine the consequences of expressing survivin in the nucleus rather than the cytoplasm, we have fused wild type survivin-GFP to nuclear localization signals (NLS). We herein report that nuclear survivin is subject to accelerated proteosomal degradation and an abrogation of the cytoprotection otherwise afforded by its overexpression. Together our results suggest a possible mechanism for eliminating survivin from interphase cells with a concomitant sensitization to apoptotic stimuli.

EXPERIMENTAL PROCEDURES
Unless otherwise stated tissue culture reagents were from Invitrogen, and all other chemicals were from Sigma.
Cell Culture and Generation of Stable Lines-HeLa cells were maintained at 37°C with 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, penicillin/streptomycin, 500 g/ml G418, and fungizome. Lines made specifically for the study were survivin NLS(LANA) -GFP, survivin NLS(SV40) -GFP, and GFP NLS -GFP. All other lines have been described previously (4,9). Proteins of interest were expressed by FuGENE 6 (Roche Applied Science)-mediated transfection of pcDNA3.1 constructs and selected with G418 (500 g/ml). The cells stably expressing proteins of interest were maintained similarly but were grown in the presence of G418. Prior to experimentation, the lines were sorted using an LSRII fluorescence-activated cell sorter (BD Biosciences) to ensure homogeneous populations and used within five passages of sorting.
Nuclear Cytoplasmic Fractionation-The cells were harvested, washed in phosphate-buffered saline then resuspended in ice-cold hypotonic buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 1 mM dithiothreitol, plus protease inhibitors). The cells were sheared by passage through a 25-gauge needle 15 times. The lysates were centrifuged at 11,000 ϫ g for 20 min at 4°C, and the supernatant was collected. The pellet was resuspended in 20 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 0.42 M NaCl, 0.2 mM EDTA, 25% (v/v) glycerol, 1 mM dithiothreitol, plus protease inhibitors. The lysates were centrifuged at 21,000 ϫ g for 5 min to obtain nuclear fractions. Protein concentration was determined using a Bradford Reagent protein assay (Bio-Rad), and equal concentration of cytoplasmic and nuclear extracts were used for immunoblotting analyses.
Radiolabeling and Immunoprecipitation-In vivo labeling was carried out by incubating 10 6 cells with 50 mCi/ml [ 35 S]methionine. To determine the rate of protein turnover, the cells were pulsed as above and chased for up to 16 h in the presence of an excess of unlabeled amino acids. After radiolabeling cells were lysed for 30 min on ice in 500 l of radioimmune precipitation assay buffer (20 mM Tris, pH 8, 137 mM NaCl, 0.5 mM EDTA, 10% glycerol, 1% Nonidet P-40, 0.1% SDS, 1% deoxycholate with 1 mM ␤-glycerophosphate, and 1 g/ml each of the protease inhibitors 4-(2-aminoethyl) benzenesulfonyl fluoride (AEBSF), chymostatin, leupeptin, antipain, pepstatin A), containing 2 mM MgCl 2 and 25 units/ml benzonase (VWR). The lysates were then cleared, and the supernatants were incubated for 1.5 h at 4°C with 2 g of a polyclonal anti-survivin (Novus) antibody. Protein G-Sepharose beads were then added (40 l of a 50% slurry in lysis buffer), and the samples were incubated for a further 2 h at 4°C. The samples were then washed and separated by SDS-PAGE, and band intensities were quantified from the dried gel using a Storm 860 PhosphorImager (GE Healthcare). Cold immunoprecipitations were performed similarly and analyzed by immunoblotting.
Fluorescence Microscopy-The cells were grown on poly-Llysine-coated coverslips, then fixed with 4% formaldehyde (Science Services), and permeabilized with 0.15% Triton, both in phosphate-buffered saline for 5 and 2 min, respectively (37°C). Interphase cells were probed with anti-lamin B antibodies (C20: 1/500, Santa Cruz), and mitotic cells were probed with antitubulin antibodies (1/2000; B512), followed by anti-goat or anti-mouse Texas red secondary antibodies (1/200; Vector Labs). The cells were mounted in Vectashield with 4Ј,6Ј-diamino-2-phenylindole and viewed using an inverted Olympus microscope fitted with an ϫ63 oil immersion lens, (NA 1.35). The images were captured using a Hammamatsu CCD camera and Delta Vision Spectris software (Applied Precision). JPEG snap shots were prepared as three-dimensional projections of deconvolved z-stacks. Fields of cells were photographed using a Zeiss Axioplan microscope, fitted with an ϫ40 objective, and operated using Simple PCI software.
Reverse Transcription-PCR-RNA was extracted from 10 7 asynchronously growing HeLa cells using the RNAqueous kit (Ambion). RNA samples were incubated for 1 h at 37°C with RNase-free DNase (Promega) to eliminate any contaminating DNA. After inactivation of the DNase (70°C for 10 min), RNA was precipitated with 1 volume of isopropanol and then resuspended in RNase free water. 4 g of each sample was used for cDNA synthesis using a First-Strand cDNA synthesis kit (GE Healthcare). cDNA for exogenous survivin-GFP (and variants) was amplified using a forward primer, which annealed to the 5Ј end of survivin open reading frame and a reverse primer, which annealed to the 5Ј-end of a GFP open reading frame.
Drug Treatments, Cell Synchronization, and FACS Analysis-To inhibit protein translation, the cells were treated with 50 g/ml cycloheximide. To inhibit protein degradation mediated by the proteasome, the cells were treated with 50 M MG132 for 6 h or 20 M MG132 for 16 h. When working with the NLStagged versions of survivin, MG132 treatment prior to cycloheximide treatment was necessary to enable detection of these proteins at the outset of the experiment. To inhibit CRM1-dependent nuclear export, the cells were treated with 6 -10 ng/ml leptomycin B (LMB; VWR) for 4, 6, or 12 h as indicated. For G 1 synchrony, the cells were treated overnight with 400 M mimosine. Cell cycle distribution was determined by measuring the DNA content using flow cytometry. Briefly, 10 5 cells were harvested, washed, and fixed with 70% ice-cold ethanol. The cells were then washed with phosphate-buffered saline and resuspended in 200 l of propidium iodide solution containing 50 g/ml propidium iodide and 100 g/ml RNase A (MP Biomedicals, UK). Propidium iodide-stained cells were analyzed with a FACScan cytometer using CellQuest software (Becton Dickinson).
X-irradiation and Clonogenic Survival-The cells were seeded at low density (500 -1000 cells/dish) in 9-cm 2 Petri dishes and allowed 2 h to attach, before exposure to X-irradiation using an Hs-X-Ray System (A.G.O. Installations Ltd., Reading, UK). Seven days post-irradiation, the colonies were stained with methylene blue (1 h at room temperature), dried, and then rinsed with H 2 O, and colonies of 50 cells or more were counted.
Apoptosis Assays-To induce apoptosis by the extrinsic caspase-8/caspase-3 pathway, exponentially growing cells were treated with 250 g/ml recombinant human TRAIL (Pepro Tech EC Ltd) for 60 or 90 min. The cells were lysed (45 min at room temperature) in mammalian protein extraction buffer, MPER (Pierce), supplemented with 1 mM EDTA, 1 g/ml pepstatin A, and 1 mM 4-(2-aminoethyl) benzenesulfonyl fluoride (AEBSF), at a concentration of 10 6 cell equivalents/ml. The lysates were then cleared, snap frozen in liquid nitrogen, and stored at Ϫ80°C.
To determine apoptotic activity, tetrapeptide cleavage assays were performed in a 96-well plate. Briefly, 5 ng/ml of the caspase-3-specific tetrapeptide substrate (DEVD-AMC; Biomol) was incubated at 37°C for 1 h with 20 -50 l of whole cell lysate prepared in MPER (Pierce) in 20 mM HEPES, pH 7.5, with 10% glycerol and 1 mM dithiothreitol. Relative fluorescence release was measured using a Spectramax Gemini fluorimeter (Molecular Devices) with excitation set at 380 nm and emission at 440 nm.
Cell Viability Assay-The cells were seeded at a density of 10 4 /well in a 24 well dish and then irradiated at the doses indicated. Seven days later, the cells were incubated with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, and cell viability was assessed using a Spectramax Gemini fluorimeter (Molecular Devices).

Endogenous Survivin Is Preferentially Degraded in the
Nucleus-It has previously been shown that survivin is subject to proteasome-mediated degradation, because levels of endogenous survivin increase after treatment with MG132 (17). This increase is not due to accelerated synthesis because immunoprecipitation of endogenous survivin pulse-labeled with [ 35 S]methionine for 2 h actually showed decreased incorporation of 35 S (thus decreased synthesis) in the presence of MG132 (data not shown). To investigate any dependence of survivin stability on subcellular compartmentalization, we have fractionated asynchronous HeLa cells following MG132 treatment and analyzed the level of endogenous survivin in nuclear and cytoplasmic fractions (Fig. 1A). Using tubulin and XRCC1 as markers of cytoplasmic and nuclear fractions, respectively, we observed a selective increase in nuclear levels of endogenous survivin following MG132 treatment. Consistent with this finding, when endogenous survivin was sequestered in the nucleus by treatment with the exportin inhibitor LMB, its expression was reduced (Fig. 1B) by ϳ30%. Together, these data suggest that nuclear survivin may be less stable than survivin localized to the cytoplasm.
Generation of Stable Lines Expressing Survivin NLS -GFP-Survivin is a nuclear cytoplasmic shuttling protein that is primarily cytoplasmic when overexpressed. To further investigate post-translational regulation of survivin levels, we sought to send survivin to the nucleus. To this end we fused full-length human survivin to two separate NLS sequences, the bipartite LANA sequence RRHERPTTRRIRHRKLRS (10) and the monopartite SV40 T-antigen NLS sequence PKKKRKV (11), hereinafter referred to as survivin NLS(LANA) -GFP and survivin NLS(SV40) -GFP, respectively. Because these survivin con-structs are expressed from the cytomegalovirus promoter, they are not subject to transcriptional regulation; thus they enable us to investigate changes in protein level attributed solely to post-translational regulation. HeLa cell lines were generated that stably overexpressed these versions of survivin and were FACS-sorted to homogeneity prior to use. As shown in Fig. 2A, survivin-GFP was predominantly cytoplasmic, whereas both survivin NLS(LANA) -GFP and survivin NLS(SV40) -GFP were retained in the nucleus (see Fig. 2, B and C). Lines were also generated that expressed GFP or GFP NLS -GFP, for use as controls (data not shown). Importantly, the presence of an NLS on survivin did not hamper its localization in mitosis, where both constructs were found at the same locations as survivin-GFP: the centromeres, midzone, and midbody, during prometaphase, anaphase, and cytokinesis, respectively, (Fig. 2, D-F).

FIGURE 1. Endogenous survivin is preferentially degraded in the nucleus.
A, nuclear (lanes N) and cytoplasmic (lanes C) fractionation was carried out on HeLa cells that had been incubated in the absence (Ϫ) or presence (ϩ) of MG132 (50 M, 6 h). An increase in endogenous survivin was apparent upon proteasome inhibition in the nuclear (compare lanes 2 and 4) but not the cytoplasmic fraction (lanes 1 and 3). B, HeLa cells were incubated in the absence (Ϫ) or presence (ϩ) 6 ng/ml LMB for 12 h to inhibit exportation of survivin from the nucleus. This treatment alone caused a 30% reduction in survivin expression.

FIGURE 2. Expression of survivin-GFP and survivin NLS -GFP constructs in HeLa cells. A-C, interphase cells stably expressing the constructs indicated
were probed with anti-lamin B antibodies (red) to show the nuclear margins, and 4Ј,6Ј-diamino-2-phenylindole to visualize the DNA (blue). The right panels show a representative field of cells from each population. D-F, mitotic cells as above were probed with anti-tubulin antibodies (red) and 4Ј,6Ј-diamino-2phenylindole (blue). NLS fusion caused nuclear sequestration of survivin-GFP in interphase but did not alter localization during mitosis. FEBRUARY 8, 2008 • VOLUME 283 • NUMBER 6

Nuclear Survivin Is Rapidly Degraded
Survivin -NLS -GFP Is Degraded More Rapidly than Survivin-GFP-From our low magnification fluorescence data in Fig. 2 (A-C, right panels), we noted that the level of expression of survivin NLS -GFP in both lines appeared lower than for the line expressing survivin-GFP. By reverse transcription-PCR, we confirmed that the transcripts to these forms were present (Fig.   3A), and therefore mRNA was still being expressed. We also ascertained that these forms did not have different rates of protein synthesis (Fig. 3G). Thus we reasoned that the differential expression was due to differences in the rates of protein turnover. To examine this, we immunoblotted whole cell extracts from asynchronous populations of cells expressing survivin- GFP or survivin NLS -GFP. Strikingly, whereas survivin-GFP was abundantly present in untreated asynchronous cells, survivin NLS(LANA) -GFP was barely detectable (Fig. 3B). Furthermore, whereas 6 h treatment with MG132 caused only a modest (1.08-fold) increase in survivin-GFP expression (Fig. 3, B and  C), survivin NLS -GFP levels rose at steady rate to 3.9-fold (Fig. 3,  B and C), demonstrating that the stability of survivin NLS -GFP is proteasome dependent, as is the case for the endogenous protein (Fig. 1A). Similar results were obtained with survivin NLS(SV40) -GFP. As a control we also compared GFP expression in cells expressing GFP and GFP NLS -GFP. No differences in GFP expression were observed in these lines, indicating that the increased rate of turnover was not an artifact of the tag (data not shown).
To determine the relative stability of these versions of survivin, cells were treated with the translational inhibitor, cycloheximide (Fig. 3, D and E). Because of the rapid clearance of the nuclear forms of survivin, this experiment had to be carried out after pretreatment with MG132 (Fig. 3E, lanes 1 and 2). Note, 16 h of treatment with MG132 did not affect cell cycle stage as assessed by FACS analysis (data not shown). Over a 16-h time course, survivin NLS(LANA) -GFP was degraded much more rapidly than survivin-GFP (Fig. 3, D and E), as is evident by the quantitation in Fig.  3F. We also noted that the addition of an NLS to GFP itself did not decrease the stability of GFP (data not shown).
To exclude the possibility that the level of survivin expression was due to changes in the rate of protein synthesis, we next pulse labeled survivin-GFP and survivin NLS -GFP cells with [ 35 S]methionine. First, the cells were treated with MG132 for 4 h, then exposed to [ 35 S]methionine, and incubated for a further 2 h (Fig. 3G). The lysates were then prepared from each population and survivin-GFP or survivin NLS -GFP immunoprecipitated from the extracts using anti-survivin antibodies. As shown in Fig. 3G survivin-GFP and survivin NLS -GFP incorporated [ 35 S]methionine to similar levels as quantified using a PhosphorImager (pixel intensities of bands 245246 and 246678, respectively). Next, we followed the pulse labeling with a cold chase after the removal of MG132 and the addition of cycloheximide. In accordance with our immunoblotting experiments in Fig. 3 (D-F), the rate of survivin NLS -GFP turnover was more rapid than survivin-GFP (Fig. 3H). Thus these data further indicate that survivin is less stable in the nucleus than in the cytoplasm.
Survivin Is Preferentially Degraded in the Nucleus-Next we made nuclear and cytoplasmic extracts from asynchronous cultures of the stable cell lines of interest and loaded equivalent numbers of cells/lane (Fig. 4). Using tubulin and XRCC1 as markers of cytoplasmic and nuclear fractions, respectively, we found that survivin-GFP, like endogenous survivin, was predominantly cytoplasmic (Fig. 4A), but, consistent with our fluorescent imaging, expression of the NLS fused versions, survivin NLS(LANA) -GFP (Fig. 4B), and survivin NLS(SV40) -GFP (Fig.  4C), was extremely low. Moreover, there appeared to be little difference in expression between the two compartments, which was surprising given that survivin NLS(LANA) -GFP and survivin NLS(SV40) -GFP were specifically sent to the nucleus. However, upon 6 h of MG132 treatment, the levels of all versions of survivin, wild type and NLS-fused, rose dramatically in the nucleus, further suggesting that survivin is less stable in the nuclear versus cytoplasmic compartment. We noted that the levels of the NLS-fused forms of survivin also increased in the cytoplasmic fraction upon treatment with MG132 (Fig. 4, B and C), illustrating the nucleo-cytoplasmic shuttling activity of the protein.
Survivin Is Degraded in the Nucleus in G 1 -Survivin expression is normally limited to the G 2 /M stages of the cycle, with a rapid decline in its levels as cells exit mitosis and enter G 1 . A combination of factors contributes to the reduction in survivin during G 1 , including transcriptional repression, externalization of the midbody at the end of mitosis, and proteolysis. This prompted us to further investigate survivin degradation during G 1 , using cells expressing the survivin-GFP, which is not subject to transcriptional repression. The cells were synchronized in G 1 using mimosine (Fig. 4D), then treated with 20 M MG132 (Fig. 4E), and fractionated (Fig. 4F) to assess exogenous levels of survivin-GFP in the cytoplasm versus the nucleus. As with the asynchronous population following MG132 treatment (Fig.  4A), survivin-GFP was selectively up-regulated in the nucleus after 6 h of MG132 treatment in G 1 -arrested cells (Fig. 4F).
Degradation of Nuclear Survivin Is Mediated by cdh1-Expression of survivin, and its partner protein aurora-B kinase, is known to be regulated by proteolysis as cells exit mitosis (15,17). Degradation of aurora-B has been demonstrated to be mediated by the APC activated by cdc20 and cdh1 (15); however, how survivin degradation is regulated has not been . Immunoblots were probed with anti-tubulin and anti-XRCC1 antibodies to indicate cytoplasmic and nuclear fractions, respectively. MG132 stabilized survivin-GFP in the nucleus but did not alter its expression in the cytoplasmic fraction, suggesting that it is preferentially degraded in the nucleus. MG132 caused increased expression of survivin NLS -GFP lines in both compartments, which may reflect the nuclear cytoplasmic shuttling activity of these proteins. D, to arrest cells in G 1 , the cells were treated for 16 h with mimosine, and the DNA content was analyzed by FACS. E, no change in G 1 distribution was observed upon the addition of MG132. F, nuclear cytoplasmic fractionation was carried out as in A on a G 1 enriched population of survivin-GFP cells. As in A, survivin-GFP was preferentially stabilized in the nuclear fraction upon MG132 treatment. FEBRUARY 8, 2008 • VOLUME 283 • NUMBER 6

Nuclear Survivin Is Rapidly Degraded
addressed. Thus to test whether survivin degradation was cdh1 or cdc20 dependent, we transiently overexpressed Myc-cdh1 or Myc-cdc20 (gifts from Dr. K. Ravid) in cells expressing the survivin constructs of interest. Immunoblotting analysis 24 h posttransfection revealed that cells expressing either Myc-cdh1 or Myc-cdc20 decreased the abundance of survivin-GFP, survivin NLS(LANA) -GFP and endogenous survivin in asynchronous cells (Fig. 5, A-C). This decrease in survivin levels was prevented by the addition of MG132 for 1.5 h post-transfection (Fig. 5D). (Note also, however, that the transfection efficiency with cdc20 was always lower than for cdh1). Quantitation of survivin expression from Fig. 5 (A-D) is shown in Fig. 5E and plotted as a fraction of the expression in control cells.
cdh1 is a nuclear protein (12), whereas cdc20, whose level is low in G 1 , is more membranous/cytoplasmic (13). Thus, having established that survivin is degraded preferentially in the nucleus, we next asked whether depletion of cdh1 could increase survivin levels. cdh1 was depleted by small interfering RNA from asynchronous HeLa cells, and protein lysates were prepared 24 h post-transfection and analyzed for survivin expression by immunoblotting. Despite an incomplete knock down of cdh1 (54%), survivin expression doubled under these conditions (Fig. 5F).

Nuclear Survivin Does Not Protect Cells against Apoptosis-
To assess whether nuclear survivin is able to inhibit apoptosis, the cell lines indicated were subjected to a clonogenic survival assay after exposure to increasing doses of ionizing X-irradiation. Consistent with our previous data (4), expression of survivin-GFP conferred resistance to X-irradiation, compared with cells expressing GFP (Fig. 6A) or GFP NLS -GFP (data not shown). By contrast no resistance to radiation was conferred by the lines expressing survivin NLS(LANA) -GFP and survivin NLS(SV40) -GFP; indeed, these lines exhibited some increased sensitivity to this treatment, although less so than cells expressing survivin L98A -GFP.
Next we induced apoptosis by treatment with recombinant TRAIL and measured caspase activity in a fluorogenic tetrapeptide cleavage assay using the caspase-3-specific substrate, DEVD-AMC (Fig. 6B). The lysates were prepared from cells expressing GFP, GFP-NLS-GFP, survivin-GFP, survivin NLS -GFP, or survivin L98A -GFP (as indicated) 0 or 60 min post-treatment with TRAIL and incubated for 1 h with DEVD-AMC. In these assays survivin-GFP conferred protection against TRAIL mediated apoptosis, but cells expressing survivin NLS -GFP exhibited similar and sometimes elevated levels of caspase-3 activity to GFP and GFP NLS -GFP controls. The kinetics of the induction of caspase activity was most rapid in cells expressing the mutant version, survivin L98A -GFP, which we previously showed was nuclear and pro-apoptotic (4).
Because survivin is rapidly degraded in the nucleus, we asked whether inhibiting proteolysis could restore the ability of survivin to inhibit apoptosis. The cell lines indicated were exposed to 5-Gy x-rays in the absence and presence of MG132. MG132 was removed after 6 h, and cell viability was analyzed 7 days later using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (Fig. 6C). In this assay inhibiting proteolysis rescued the anti-apoptotic activity of survivin possibly because of the increased cytoplasmic pool that accumulates under these conditions (Fig. 4C). To ascertain specifically whether the nuclear pool of survivin can be cytoprotective, we repeated the TRAIL assay (Fig. 6B) in the absence and presence of MG132, or MG132 and LMB. In this assay, MG132 treatment caused a decrease in caspase activity in controls and experimental samples, making it difficult to assess whether increased stability of the exogenously expressed protein specifically contributes to the reduced caspase activity. However, the additional treatment of LMB caused an increase in apoptosis in survivin-GFP cells, whereas the level of caspase 3 activity in survivin NLS -GFP cells appeared to be unaffected by either treatment. These data indicate that when stabilized and completely nuclear, survivin cannot inhibit apoptosis. Taken together our present data suggest that forced expression of survivin in the nucleus is sufficient to prevent it from inhibiting apoptosis in cultured human cells.

DISCUSSION
Survivin is a nucleo-cytoplasmic shutting protein that is predominantly cytoplasmic when overexpressed in cultured cells (2,9). We and others have recently shown that this subcellular localization is dependent upon CRM1 (␤-exportin) and a revlike NES in the linker region of survivin, between its BIR domain and the C-terminal ␣-helix (2-4, 6 -8). Here we demonstrate that survivin is preferentially degraded in the nucleus in a cdh1/APC-dependent manner. These findings are consistent with cdh1-mediated degradation of survivin in vitro (14), and the nuclear localization of cdh1 in G 1 (12). Somewhat paradoxically, survivin appears to be devoid of destruction motifs recognized by the APC/C. However, the mitotic partner protein of survivin, aurora-B, has three putative D-boxes, a KEN box, and an A-box, and mutation of the cdh1-specific KEN and A-boxes stabilize aurora-B, suggesting that its destruction is mediated preferentially by cdh1 (15,16). In addition, aurora-B also exhibits accelerated clearance upon overexpression of cdh1, exhibits increased stability upon cdh1 depletion, and co-immunoprecipitates with cdh1 in mitotic extracts (15,16). Thus, because survivin and aurora-B are both destroyed at the end of mitosis (14 -17), it is formally possible that survivin relies on the consensus sequences of aurora-B for destruction. We are currently testing this hypothesis. . Survivin NLS -GFP is not cytoprotective. A, HeLa cells stably expressing the constructs indicated were seeded at low density, exposed to X-irradiation, and colonies of 50 or more cells counted 7 days post-irradiation. The surviving fraction was plotted in logarithmic scale. Overexpression of survivin-GFP, but not survivin NLS -GFP, protected cells against irradiation. Note that neither survivin NLS -GFP line was as sensitive to irradiation as survivin L98A -GFP. The data are representative of three independent experiments. Each experiment was performed in triplicate, and error bars show standard deviation from the mean. Paired t test analysis revealed that at 2.5-Gy irradiation there was a significant difference between survivin-GFP expressing populations and those expressing survivin NLS(LANA)-GFP (p ϭ 0.031), survivin NLS(SV40)-GFP (p ϭ 0.047), or survivin L98A -GFP (p ϭ 0.037). The differences were not significant at higher doses of irradiation. B, caspase-3 activity assay. Apoptosis was induced by the addition of recombinant TRAIL. The cell lysates were analyzed for their ability to cleave the caspase-3-specific substrate (DEVD-AMC), and relative fluorescence release was measured spectroscopically. Overexpression of survivin-GFP inhibited caspase-3 activity, but activity remained high in survivin NLS -GFP lines. A paired t test comparing TRAIL-treated survivin-GFP cells with those expressing the NLS-fused survivin-GFP constructs or survivin L98A -GFP revealed a significant difference in each case. p values in italics (above control and survivin-GFP samples) were obtained by comparison with HeLa cells. We noted also that cells expressing GFP-NLS(LANA) -GFP also showed a significant difference in caspase-3 activity compared with HeLa cells alone; thus we compared GFP-NLS(LANA)-GFP and survivin NLS(LANA)-GFP samples. In this case the difference was also significant (p ϭ 0.046). C, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide Assay indicating mean survival with error bars indicating standard deviation. The cell lines indicated were exposed to 5-Gy irradiation in the absence or presence of 50 M MG132. MG132 was maintained in the medium for 6 h post-irradiation. Cell viability, assessed 7 days post-irradiation using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay, revealed that the expression of survivin-GFP prevented cell death induced by 5-Gy irradiation and was unaffected by MG132 treatment (compare gray and white bars). A paired t test (5 Gy versus 5 Gy ϩ MG132 samples) revealed no significant difference in GFP and survivin-GFP expressing populations but a significant difference (p ϭ 0.03) in the survivin NLS -GFP cells. By contrast, the decreased viability observed with survivin NLS -GFP with exposure to 5 Gy (survivin NLS -GFP, gray bars) was restored upon stabilization of the protein with MG132 (survivin NLS -GFP, white bars). D, a caspase-3 activity assay, as described for B, was performed on cells that had been pretreated with no MG132 50 M MG132 (6 h) or with 50 M MG132 and 10 ng/ml LMB (6 h). The time indicated is the duration of exposure to TRAIL; the mean and standard deviation from one experiment performed in triplicate is shown and is representative of two independent experiments. MG132 treatment alone reduced the extent of apoptosis in each sample, but interestingly, co-treatment with LMB increased the caspase activity in survivin-GFP expressing cells and had no significant effect (p Ͼ 0.05) on cells expressing survivin NLS -GFP. These data indicate that nuclear survivin cannot inhibit apoptosis.
In a previous report we found that a mutant form of survivin that accumulates in the nucleus could no longer protect cells against ionizing radiation or TRAIL-induced apoptosis (4). Corroborating data were recently presented by Stauber and coworkers (6 -8). However, these experiments raised the question as to whether subcellular relocalization alone was responsible for abrogating the anti-apoptotic activity of survivin or whether the effect was mutant specific. Here, we have artificially forced wild type human survivin expression in the nucleus and observed that this relocation prevented survivin from acting as an inhibitor of apoptosis. Furthermore, in some cases we actually noted an increase in sensitivity to apoptotic stimuli, the reason for which is unclear. One possibility may be that the subcellular localization of the exogenous protein influences the localization of the endogenous protein. In a recent study Temme et al. (5) also found that cells were more sensitive to apoptosis when they forced survivin expression in the nucleus, and interestingly they linked this observation to enhanced transcription of p53, and the pro-apoptotic genes, Bad and Bax.
Our present data appear to contradict the recent work by Stauber et al. (6), who reported that nuclear sequestration of murine survivin via deletion of the NES increased the stability of the protein, thus suggesting that it is preferentially degraded in the cytoplasm. However, it is possible that deletion of these residues could have affected the folding or stability of survivin specifically rather than increased its stability as a result of its subcellular relocalization. (Note also that our experiments used stable cell lines rather than transiently transfected cells, which could have contributed to the different results). Differential stability caused by subcellular compartmentalization has been noted for a number of proteins including p53, whose localization and stability is altered upon DNA damage (18). Furthermore, the survivin isoform Delta-Ex3, which is nuclear when overexpressed (19,20), is also cleared from the cell more rapidly than wild type survivin (21) and may explain why endogenous survivin DeX3 is difficult to detect at the protein level (20,22). Interestingly, it has recently been reported that survivin degradation can also be facilitated by the XIAP association factor, XAF-1, in a proteasome-dependent manner, which suggests that multiple pathways for ensuring the removal of survivin from interphase cells exist (24).
Finally, survivin has a functional NES but no NLS. Thus one outstanding question is how is survivin gaining access to the nucleus? Although the endogenous protein is small enough to enter the nucleus by diffusion even if dimerized, this is unlikely given the behavior of the GFP tagged form. Of the known binding partners of survivin, aurora-B has sequences that correspond to NLSs but appear nonfunctional in a nuclear targeting assay, and INCENP has three functional NLSs (3). However, when overexpressed in MCF cells, neither aurora-B nor INCENP was able to influence survivin localization, which remained predominantly cytoplasmic (3). Another candidate for nuclear targeting is TD60, an RCC1-like protein that has a putative NLS and that co-localizes with the chromosomal passenger proteins (23). However, it should be noted that chromosomal passenger proteins have cell cycle-dependent expression, and whether they are present in interphase cells when survivin is overexpressed is unknown.
In conclusion, we have demonstrated that relocating survivin to the nucleus accelerates its degradation and prevents it from protecting cells against irradiation and inhibiting apoptosis. We have also shown that the presence of an NLS on survivin does not affect its mitotic function. Thus sequestering survivin in the nucleus could be very helpful in cancer therapy because it would resensitize cells to radiation without affecting the proliferation of noncancerous cells.