Nuclear export signals: small domains with large impact
Engelsma, D.H.
Citation
Engelsma, D. H. (2008, October 16). Nuclear export signals: small domains with large impact. Retrieved from https://hdl.handle.net/1887/13258
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/13258
Note: To cite this publication please use the final published version (if
applicable).
Homodimerization Antagonizes Nuclear Export of Survivin
Traffic 8, No. 11, p 1495-1502, November 2007
Dieuwke Engelsma 1,† , Jose A. Rodriguez 2,3,† , Alexander Fish 4 , Giuseppe Giaccone 2 Maarten Fornerod 1,*
1Department of Tumor Biology, The Netherlands Cancer Instute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands
2Department of Medical Oncology VUMC, Boelelaan 1117, 1081 HV Amsterdam, The Netherlands
3Current address: Department of Genecs, Physical Anthropology and Animal Physiology, University of the Basque Country, P.O. Box 644, 48080, Bilbao, Spain
4 Department of Molecular Carcinogenesis, The Netherlands Can- cer Instute, Plesmanlaan 121, 1066 CX Amsterdam, The Neth- erlands
† These authors contributed equally to this work.
Abstract
Survivin plays separate roles during different phases of the cell cycle. In mitosis, Survivin is a key regulator of cell division, while in interphase, Survivin is able to protect cells from apoptosis. Survivin shules between nucleus and cytoplasm under the influence of one or more nuclear export signals (NESs). Paradoxically, our data show that Survivin poorly binds CRM1 in vitro because hydrophobic residues of the NES are occupied in homodimer contacts. We show that NES-preserving dimerizaon mutants behave as monomers in solu-
on, show dramacally increased CRM1 binding and are more efficiently exported in vivo than wild-type Survivin. These data indicate that Survivin contains a monomer-specific NES and that dimerizaon modu- lates cytoplasmic access of the protein. Our findings have implicaons for both the mitoc and interphase roles of survivin.
Introduction
Survivin was first described as a baculovirus inhibitor of apoptosis (IAP) repeat containing protein upregulated in cancer cells (1) and was later found to be a member of the chromosomal passenger complex. Survivin is usu- ally expressed at low levels during interphase, which rise during G2. Characteriscally of chromosomal pas- sengers, Survivin is localized at the centromeres at the beginning of mitosis and relocalizes to the overlapping microtubules of the spindle midzone in anaphase. In telophase, Survivin is localized at the midbody and is degraded aer cytokinesis (2,3). Survivin also has a clear role as an IAP when overexpressed in cancer cells (4,5). Nuclear or cytoplasmic accumulaon of Survivin in tumor cells have been differently correlated with dis- ease outcome [reviewed in Li et al. (6)], indicang that regulaon of nucleocytoplasmic localizaon may be an
important factor in disease. Survivin has been shown to exit the nucleus by means of the CRM1 pathway (7–10).
CRM1 is an export receptor (exporn) that binds nuclear export signal (NES)-containing proteins (11–13). Most NESs are short hydrophobic sequences conforming to the consensus Φ-X2-3-Φ-X2-3-Φ-X-Φ, with Φ being a hydrophobic residue (14). The CRM1–NES interacon is weak and is smulated in the nucleus by the small GT- Pase Ran in its GTP-bound state (15). In the cytoplasm, RanGTP is hydrolyzed by the cytoplasmic factors RanGAP and RanBP1, resulng in destabilizaon of the complex and release of the NES-containing protein [reviewed in Görlich & Kutay (16)]. CRM1 can be specifically inhibited by the fungal toxin leptomycin B (LMB) (11,17). Although the presence of a sequence conforming to the NES con- sensus was noted in the linker between the N-terminal BIR domain and the C-terminal coiled-coils, Survivin was inially reported to be exported through a nonclassical export mof located in its C-terminus (9). However, two different funconal export sequences have been recent- ly described in the linker (7,8,10), and one of them was reported necessary and sufficient for nuclear export of Survivin (8).
In this study, we resolve these conflicng issues and pro- vide evidence that two CRM1-dependent localizaon el- ements (one in the linker and another in the C-terminus) operate in Survivin. Moreover, we show that the central NES of Survivin is only acve in Survivin monomers.
Results and discussion
Survivin contains two cytoplasmic localizaon elements Following the original report that Survivin is exported from the nucleus by means of the CRM1 pathway (9), a number of studies have further specified the require- ments of this export event (7,8,10,18). However, there are several inconsistencies between these studies. Rodri- guez et al. (9) idenfied the C-terminus of Survivin(119–
142) as the determinant for CRM1-mediated nuclear ex- port; Knauer et al. idenfied a central region (aa 89–98) as a classical NES (8,10,18) and Colnaghi et al. (7) suggest an NES in the parally overlapping region of amino acids 96–104.
To understand these discrepancies, we wished to rigor- ously test the nuclear export acvity of different regions of Survivin using a live nuclear export assay. MCF7 cells were transfected with yellow fluorescent protein (YFP)- tagged deleon mutants and imaged before and aer 20 min of addion of the CRM1 inhibitor LMB. As shown in Figure 1A, full-length (1–142) Survivin localized to the
Homodimerization of Survivin Antagonizes
Nuclear Export of Survivin
74
Survivin Dimerization Antagonizes Nuclear Export
cytoplasm, as did a mutant containing the previously suggested NESs but lacking the C-terminus (1–109) or a mutant consisng of the C-terminus only (119–142). The previously idenfied NES region by itself (89–109) also mediated nuclear exclusion. In all cases, the YFP-proteins lost their preferenal cytoplasmic localizaon within 20 min upon addion of the CRM1 inhibitor LMB (Figure 1A), arguing for a rather direct effect of CRM1 in the cytoplas- mic localizaon of these proteins. It has been reported that mutaon of the 89–98 NES (L96A/L98A) or 96–104 NES (L98A) caused Survivin to become (partly) nuclear (7,8). Surprisingly, we found that the L96A Survivin mu- tant was sll predominantly cytoplasmic in all cells (Fig- ure 1B), although its nuclear exclusion was significantly lower in comparison to wild-type Survivin (Figure 1C). To exclude mutaon bias or cell-line-specific effects, we also transfected an L96A/L98A mutant in both MCF7 cells and HeLa cells. In both cell lines, this mutant was cytoplasmic to the same extent (Figure S1A). We noted that placing the YFP moiety at the C-terminus resulted in a reduced
cytoplasmic localizaon of Survivin L96A (Figure S1B), suggesng that fusion at the C-terminus may interfere with the C-terminal localizaon signal. Strikingly, upon introducon of the L96A mutaon in Survivin(1–109) or Survivin(89–109) export was completely inhibited. These data indicate that both previously described CRM1-de- pendent cytoplasmic localizaon elements operate in Survivin and that one of them is crically dependent on L96.
We previously reported (9) that residues 89–109 are not sufficient to promote nuclear export using a Rev(1.4)–
green fluorescent protein (GFP) assay system (19), sug- gesng that the 89–98 fragment is a very weak NES. We noted the presence of addional hydrophobic residues upstream of amino acid 89 and tested the possibility that they might contribute to CRM1 binding, similar to what has been reported for the Rev NES (20). Indeed, addion of five residues (Figure 3A) to the N-terminus resulted
Figure 1. Survivin contains two CRM1-dependent cytoplas- mic localization elements.
A) Subcellular localizaon of YFP–
Survivin(1–142), YFP–Survivin(1–
109), YFP–Survivin (89–109) or YFP–
Survivin(119–142) before or aer 20 min of exposure to 100 nm of LMB.
B) Subcellular localizaon of YFP–
Survivin(1–142)L96A, YFP–Survivin (1–109)L96A and YFP–Survivin (89–
109)L96A, treated with LMB as indi- cated. C) Distribuon of nucleocyto- plasmic raos (C/N) of YFP–Survivin (1–142) versus YFP–Survivin(1–142) L96A. Box plots show lower and up- per quarle (boxed), with median value indicated as a horizontal line.
The smallest and largest observaons (whiskers) are connected to the box with dashed lines and outliers are indicated by circles. p-value accord- ing to a Mann–Whitley U-test. D–F) Subcellular localizaon of Survivin’s short (aa 89–109) and extended (84–
109) NES fused to either YFP (E) or Rev(1.4)–GFP (F). Nucleocytoplasmic raos (D) as in (C). Bars, 10 μm. WT, wild type.
in an increased cytoplasmic localizaon of YFP as well as efficient export of Rev(1.4)–GFP (Figure 1C–E), indicat- ing that the more complete NES lies within amino acids 84–109.
The central survivin NES is monomer specific Studies by Knauer et al. showed that Survivin interacts with CRM1 in a reculocyte lysate (8,10,18). This inter- acon was dependent on RanGTP and the central NES region (amino acids L96 and L98). Somewhat surpris- ingly, Colnaghi et al. (7) reported that the CRM1/Sur- vivin interacon was not dependent on amino acid L98.
Whereas these studies indicate that CRM1 physically interacts with Survivin amino acids 89–104, it did not exclude reculocyte-lysate-derived bridging factors and therefore could not unambiguously idenfy the Survivin NES. To address this issue, we expressed recombinant Survivin and performed a pull-down assay with recom- binant CRM1, which was immobilized using an in-frame IgG binding domain (21) (Figure 2A). The specificity of the NES interacon with CRM1 was evaluated by addion of
RanGTP (11). Surprisingly, whereas a GFP-NES protein interacted with CRM1 in a RanGTP-smulated manner (lanes 5–6), Survivin was retained on the CRM1 column, both in the absence or presence of RanGTP (lanes 2–3).
A similar interacon was also seen on a z-tagged trans- porn1 column (lane 8). These results suggest that the observed CRM1/Survivin binding was aspecific, rather than an NES-type interacon. To further confirm this lack of direct NES-type interacon, we tested the ability of recombinant Survivin to bind CRM1 in the recombinant CRM1 GTPase acvang protein assay (21,22). This as- say probes the ability of a potenal NES substrate to pro- tect RanGTP from RanGAP-smulated hydrolysis in the presence of CRM1, which is a measure of NES substrate/
CRM1/RanGTP formaon. Consistent with the results of the pull-down assay, recombinant full-length Survivin showed a very weak NES-type interacon with CRM1 (Figure 2B), with an affinity at least two orders of mag- nitude larger than the well-characterized protein kinase inhibitor NES (PKI NES) (23). The isolated Survivin central NES pepde (84–109) interacted with CRM1 with an in- termediate affinity, approximately 10-fold lower than the PKI NES, and the interacon was completely disrupted by mutang leucine 98 to alanine (Figure 2B). Together with the pull-down data, these results suggest that Survivin contains an NES at amino acids 84–109, which is masked in the full-length protein. X-ray crystallography and bio- Figure 2. Interaction between recombinant Survivin
and CRM1. A) CRM1 pull-down assay. Immobilized CRM1 (lanes 1–6) or transporn1 (lanes 7–8) was incubated with Sur- vivin, GFP-NES and/or RanGTP as indicated above the lanes.
Flow-throughs (FT) and eluons with 0.2% (E1) and 2% SDS (E2) were analyzed on Coomassie-Blue-stained SDS polyacrylamide gels. B) Ran GTPase protecon assay. Affinies of Survivin, Sur- vivin NES(84–109), Survivin NES(84–109)L96A and the PKI NES for CRM1 were measured using their abilies to inhibit RanGAP- smulated, CRM1-dependent hydrolysis of RanGTP (21). 100%
GTP hydrolysis indicates lack of binding. Results are the average of two experiments, except for circled data points that are from
Figure 3. NES-preserving mutations in the Survivin di- merization region increase the affinity for CRM1. A) Schemac overview of the Survivin dimerizaon region with predicted NESs underlined (8) or boxed (7). Van der Waals and hydrogen bond dimer interacons are shown in dashed and solid lines, respecvely, with the arrowhead poinng to the acceptor group (24). The horizontal dashed line represents the addional residues required for opmal NES funcon. In the consensus NES sequence, Φ is any hydrophobic amino acid, X any other amino acid. B) CRM1 Ran GTPase protecon assay using wild-type Sur- vivin, SurvivinF101A/L102A or SurvivinL98E/F101A/L102A that was untreated or treated with 2 m urea was performed as in Fig- ure 2B. Results are average of two to four experiments.
76
Survivin Dimerization Antagonizes Nuclear Export
chemical studies indicate that Survivin is a dimer (24–26).
The dimerizaon interface overlaps with one of the sub- sequently idenfied central NES domains (8,10,18) (Fig- ure 3A). We therefore wished to test CRM1 binding of monomeric Survivin. As a first test, we exposed Survivin to 2 m urea, which is reported to disrupt protein dimers (27–29). Indeed, urea-exposed Survivin has a significant- ly higher affinity for CRM1 than the untreated protein (Figure 3B). As a control, urea treatment did not affect the affinity of PKI for CRM1 (data not shown). While these results are consistent with dedimerizaon-induced binding of Survivin to CRM1, the effects of urea cannot be exactly defined. Therefore, we wished to mutagenize the dimerizaon interface of Survivin. Many of the dimer contacts are part of the NES consensus sequence (Figure
3A) and therefore cannot be used to test the effect of dimerizaon on NES funcon. However, phenylalanine 101 and leucine 102 may fall outside the NES, although it should be noted that these mutaons would negavely affect the NES as proposed by Colnaghi et al. (7). We test- ed the interacon of recombinant SurvivinF101A/L102A with CRM1. Strikingly, the mutant showed a drascally improved affinity for CRM1, which was not further affect- ed by urea (Figure 3B). Addional mutaon of leucine 98 (SurvivinL98E/F101A/L102A) abolishes all detectable CRM1 binding. This suggests that NES-mediated nuclear export is negavely modulated by dimerizaon. Also, this result suggests that residues 96–104 is not the locaon of the NES, as CRM1 binding would be expected to be strongly reduced by mutaon of F101 and L102 (23).
One important cauonary note is that while the Survivin crystals clearly indicate dimerizaon through the central 89–102 domain (24–26), predicted dimerizaon mutants have so far not been shown to be monomers (10,30). We therefore tested the molar mass of wild-type Survivin in the presence or absence of 2 m urea and the dimerizaon mutant SurvivinF101A/L102A using mulangle laser light scaering (MALLS) in line with gel filtraon analysis. For wild-type Survivin, we measured a molar mass of 37 kDa (Figure 4A), close to double the calculated molar mass of His-tagged Survivin (17.3 kDa). In contrast, the molar mass of wild-type Survivin in the presence of 2 m urea or SurvivinF101A/L102A was measured as 19 or 23 kDa, respecvely, close to the predicted monomeric state.
The conversion from dimer to monomer is also seen in the 280-nm absorpon peaks, which are clearly shied toward a lower molecular weight. The shi in the pres- ence of 2 m urea is limited possibly because the protein is partly denatured and therefore would have a larger ef- fecve radius. Collected fracons from the size-exclusion Figure 4. The NES-preserving Survivin mutation F101A/
L102A or 2 M urea disrupts the Survivin dimer in solution.
A) Mulangle laser light scaering measurements of molar mass together with absorbance at 280 nm of recombinant His-tagged Survivin or SurvivinF101A/L102A in the presence or absence of 2 m urea. Measurements were in sequence with gel filtraon anal- ysis, of which eluon volumes are noted on the x-axis. Dashed lines indicate observed molar mass at the peak fracons. B) Coomassie-Blue-stained SDS polyacrylamide gels of 0.5 mL (wild type) or 1 mL (SurvivinF101A/L102A) fracons from (A).
Figure 5. Survivin dimerization mutants show increased CRM1-dependent nuclear export in vivo. A) Subcellu- lar localizaon of YFP–Survivin(84–109), YFP–Survivin(84–109) F101A/L102A, YFP–Survivin(1–109) and YFP–Survivin(1–109) F101A/L102. B) Distribuon of nucleocytoplasmic raos as in Fig- ure 1C. Bar, 10 μm. WT, wild type.
column were analyzed on a Coomassie-Blue-stained gel to confirm the purity of the protein (Figure 4B).
These results confirm that Survivin forms a dimer in soluon and that 2 m urea or NES-preserving point mutaons in the dimerizaon interface are sufficient to disrupt the dimer. Furthermore, these data sub- stanate our proposal that dimerizaon of Survivin blocks CRM1 interacon. The lack of CRM1 interacon with wild-type Survivin in vitro seems inconsistent with previous results using glutathione S-transferase–
Survivin and in vitro–translated CRM1 (8,10,18). How- ever, we note that Survivin was in great molar excess over CRM1 in these experiments and that only a small fracon of monomeric Survivin would have been re- quired to bind a substanal fracon of CRM1.
Dimerizaon of survivin antagonizes nuclear export To examine the effect of constuve monomerizaon of Survivin on nuclear export in live cells, we intro- duced the dimerizaon mutaons in the extended YFP–NES (84–109) and in YFP–Survivin fragment 1–109 (the C-terminus was removed to avoid inter- ference from the C-terminal cytoplasmic localizaon element). MCF7 cells were transfected with these constructs and the nucleocytoplasmic rao of the flu- orescent proteins was measured. A clearly increased cytoplasmic localizaon of SurvivinF101A/L102A as compared with wild type was observed in the con- text of both protein fragments (Figure 5), indicang that nuclear export was increased by the F101A/
L102A mutaon. These data, therefore, further sub- stanate the idea that dimerizaon of Survivin blocks CRM1 interacon. Inclusion of the C-terminal region (aa 110–142) in this analysis was not possible as the nucleocytoplasmic raos of full-length Survivin but not full-length SurvivinF101A/L102A were strongly dependent on expression levels (Figure S2), making the nucleocytoplasmic raos of the two proteins dif- ficult to compare.
Together, the data presented here show the pres- ence of an NES within the Survivin dimerizaon re- gion, which is only accessible to CRM1 in the Survivin monomer. While this dimer-induced masking is very efficient in vitro, in living cells the dimerizaon-pro- ficient NES is sll recognized and exported by CRM1, albeit less efficiently than the dimerizaon-deficient NES. This indicates that in vivo, Survivin rapidly switches between monomeric and dimeric states. In- teresngly, cytoplasmic access of Survivin is reported to smulate its anapoptoc funcon (7,10,31) and nuclear or cytoplasmic Survivin localizaon is a sig- nificant risk factor in mulple types of cancer [re- viewed in Li et al. (6)]. Unl present, the underlying regulatory mechanisms of Survivin localizaon were unknown. Our current data suggest that alteraons in factors that regulate dimer to monomer transion might contribute to the preferred subcellular localiza-
on of Survivin and are therefore important to further invesgate. Furthermore, Survivin dimerizaon may con- stute a potenal therapeuc target.
Whether CRM1 interacon has a role in the mitoc func-
on of Survivin is unclear at present. An L98A mutaon, which disrupts CRM1 interacon, is able to rescue the mitoc funcon of Survivin (7). However, an L96A/L98A double mutant is not (8,10,18). Perhaps this double mu- tant interferes with more than just CRM1 binding. Even if the NES is not absolutely required for Survivin’s mitoc funcon, it may sll act in mitoc regulaon. NES/CRM1 interacons are greatly destabilized aer nuclear enve- lope breakdown, when the RanGTP concentraon drops under the influence of RanGAP1 and RanBP1 (32). The CRM1/NES interacon therefore may be a ‘double lock’
for the mitoc funcon of Survivin in interphase, reduc- ing its nuclear access and prevenng the dimerizaon interface from unwanted nuclear interacons.
Conclusions
The mitoc regulator and anapoptoc protein Survivin contains two CRM1-dependent cytoplasmic localizaon elements. One of those, located in the linker between the BIR and coiled-coil domains, interacts directly with CRM1 but is masked in Survivin dimers. Constuve Survivin monomers show an increased nuclear export as compared with wild-type Survivin. The dimerizaon interface of Survivin, therefore, is a regulatory region of the protein in interphase.
Methods
Plasmids and plasmid construcon
Cloning of human Survivin complementary DNA (cDNA) and the generaon of YFP–Survivin and YFP–
Survivin(119–142) was reported previously as in Rodri- guez et al. (9). The L96A, L96A/L98A, F101A/L102A and L98E/F101A/L102A point mutaons were introduced into YFP–Survivin using polymerase chain reacon (PCR)- based site-directed mutagenesis. These constructs were subsequently used as templates in PCR-generated de- leon mutants YFP–Survivin(1–109), YFP–Survivin(89–
109), YFP–Survivin(84–109), YFP–Survivin(1–109)L96A, YFP–Survivin(89–109)L96A, YFP–Survivin(1–109)F101A/
L102A and YFP–Survivin(84–109)F101A/L102A. In all cas- es, the PCR products were cloned as HindIII/BamHI frag- ments into pEYFP-C1 (Clontech, Palo Alto, CA, USA). Poly- merase chain reacons were also used to amplify short DNA fragments encoding Survivin residues 84–109 and 89–109, which were cloned as BamHI/PinAI fragments into the pRev(1.4)–GFP nuclear export assay vector [Henderson & Eleheriou (19)] to generate pRev(1.4)–
Survivin(84-109)–GFP and pRev(1.4)–Survivin(89-109)–
GFP. To make the pCDNA–SurvivinL96A–YFP construct, the cDNA encoding SurvivinL96A was first cloned into pCDNA3.1 (Invitrogen). Subsequently, a PCR-amplified DNA fragment encoding YFP was cloned in frame at the
78
Survivin Dimerization Antagonizes Nuclear Export
3’ end of Survivin cDNA. For recombinant expression, cDNAs encoding wild-type, F101A/L102A and L98E/
F101A/L102A mutant Survivin were cloned into the bacterial expression vector pQE80 (Qiagen) to gen- erate pQE-Survivin, pQE-SurvivinF101A/L102A and pQE-SurvivinL98E/F101A/L102A. In all cases, the ab- sence of unwanted mutaons in the inserts was veri- fied by sequencing. Plasmid sequences are available on request.
Recombinant protein expression and purificaon His6-tagged Survivin, SurvivinF101A/L102A and Sur- vivinL98E/F101A/L102A were expressed from pQE80 in BL21(DE3)pRep4. Cells were lysed by sonicaon in buffer A [500 mm NaCl, 20 mm HEPES–KOH (pH 7.9), 8.7% glycerol] supplemented with 2.5 mm 2-mercap- toethanol and protease inhibitors (complete without EDTA, Roche) (buffer A+). The lysate was cleared by ultracentrifugaon, then passed over Ni-NTA agarose (Qiagen) and eluted with 500 mm imidazole in buffer B [200 mm NaCl, 20 mm HEPES–KOH (pH 7.9), 8.7%
glycerol, 2.5 mm 2-mercaptoethanol]. Proteins were dialyzed against PBS containing 8.7% glycerol, 2.5 mm 2-mercaptoethanol. For light-scaering experiments, proteins were dialyzed against 200 mm NaCl, 20 mm HEPES–KOH (pH 7.9) and concentrated using Vivaspin concentrators with 10-kDa cutoff (Vivascience).
2-Mercaptoethanol or DTT was present at all mes to prevent oxidaon and subsequent aggregaon of Survivin. His-tagged CRM1 was expressed in Escheri- chia coli TG1 from plasmid pQE60-hCRM1. Cultures were induced with 0.5 mm isopropyl-ß-d-thiogalacto- pyranoside for 6 h at 37°C. Cells were resuspended in buffer A+ and disrupted by passage through an EmulsiFlex C5 (Avesn). The lysate was cleared by ul- tracentrifugaon and subsequently precipitated with 35% ammonium sulfate. His-tagged CRM1 was affinity purified by HiTrap column (1 mL) (Amersham) loaded with Ni2+, and protein was eluted with a linear gradi- ent of imidazol (0–500 mm) in buffer B. Subsequent purificaon was performed by buffer exchange to RQ buffer [20 mm HEPES–KOH (pH 7.9), 10 mm 2-mer- captoethanol, 0.1 mm phenylmethylsulphonyl fluo- ride] containing 100 mm NaCl and Resource Q (Am- ersham) column chromatography. CRM1 was eluted at 180 mm NaCl. Z-tagged CRM1 and transporn1 were expressed as previously described (21,22). Ran and Rna1p were expressed and purified according to Izaurralde et al. (33) and Ran was loaded with GTP ac- cording to the method described previously (34).
CRM1 pull-down assay
Pull-down assays were performed as described (35).
About 1.5 μm zz-CRM1 and zz-transporn1 were in- cubated with 3 μm Survivin or 3 μm GFP–PKI NES and 4.5 μm RanGTP when indicated. Beads were eluted in 0.2% SDS on ice or in 2% SDS at room temperature.
CRM1 RanGAP assays
CRM1 RanGAP assays were performed as described (21).
When indicated, Survivin was supplemented with 2 m urea and incubated on ice for 30 min before complexes with CRM1/Ran[γ-32P]GTP were formed. The PKI NES used had the sequence CSNCELALKLAGLDINKTGG (Invit- rogen) and the Survivin NESs encompassed amino acids 84–109.
Mulangle laser light scaering
To determine the molecular weight of the proteins, a size- exclusion column (Superdex 75 30/10; GE Healthcare) was used in sequence with a MALLS detector, miniDAWN Tristar (Wya Technology Corp.) and a refracve index detector. The scaering data were analyzed using astra 4.9 soware (Wya Technology Corp.), and the molecu- lar mass of eluted peaks was calculated. Columns were equilibrated in 20 mm HEPES, pH 7.9; 200 mm NaCl and 1 mm DTT in presence or absence of 2 m urea. Proteins were loaded in 0.1 mL at 1.8 mg/mL (wild-type Survivin), 1.5 mg/mL (wild-type Survivin with urea) or 1.3 mg/mL (SurvivinF101A/L102A).
Cell culture and transfecons
MCF7 and HeLa cells were transfected with 300 ng DNA per well of a six-well dish using Fugene (Roche) accord- ing to manufacturer’s instrucons. Aer 24 h, living cells in carbonate-buffered saline were imaged using a Leica AOBS confocal microscope. Cells were treated for 20 min with 100 nm of LMB when indicated.
Image analysis
For quanficaons of nucleocytoplasmic raos approxi- mately 50 cells were imaged. ImageJ was used to mea- sure average pixel intensies of a nuclear and a cytoplas- mic region. Stascal analysis was done in R.
Acknowledgements
We thank Susanne Lens for discussions and sharing un- published data, Simone Span for general technical as- sistance, Lenny Brocks and Lauran Oomen for assistance with confocal microscopy, Minoru Yoshida for LMB, Tia Sixma and Anastassis Perrakis for discussions and Antho- ny Uren, Rob Wolthuis and Sjn Heessen for comments on the manuscript.
References
1. Ambrosini G, Adida C, Aleri DC. A novel an-apoptosis gene, survivin, expressed in cancer and lymphoma. Nat Med 1997;3:917–921.
2. Uren AG, Wong L, Pakusch M, Fowler KJ, Burrows FJ, Vaux DL, Choo KH. Survivin and the inner centromere protein INCENP show similar cell-cycle localizaon and gene knockout pheno- type. Curr Biol 2000;10:1319–1328.
3. Skoufias DA, Mollinari C, Lacroix FB, Margolis RL. Human sur- vivin is a kinetochore-associated passenger protein. J Cell Biol 2000;151:1575–1582.
4. Chakravar A, Zhai GG, Zhang M, Malhotra R, Latham DE, Delaney MA, Robe P, Nestler U, Song Q, Loeffler J. Survivin enhances radiaon resistance in primary human glioblas- toma cells via caspase-independent mechanisms. Oncogene 2004;23:7494–7506.
5. Lu B, Mu Y, Cao C, Zeng F, Schneider S, Tan J, Price J, Chen J, Freeman M, Hallahan DE. Survivin as a therapeuc target for radiaon sensizaon in lung cancer. Cancer Res 2004;64:2840–2845.
6. Li F, Yang J, Ramnath N, Javle MM, Tan D. Nuclear or cyto- plasmic expression of survivin: what is the significance? Int J Cancer 2005;114:509–512.
7. Colnaghi R, Connell CM, Barre RM, Wheatley SP. Sepa- rang the an-apoptoc and mitoc roles of survivin. J Biol Chem 2006;281:33450–33456.
8. Knauer SK, Bier C, Habtemichael N, Stauber RH. The Survivin-Crm1 interacon is essenal for chromosomal passenger complex localizaon and funcon. EMBO Rep 2006;7:1259–1265.
9. Rodriguez JA, Span SW, Ferreira CG, Kruyt FA, Giaccone G.
CRM1-mediated nuclear export determines the cytoplasmic localizaon of the anapoptoc protein Survivin. Exp Cell Res 2002;275:44–53.
10. Stauber RH, Rabenhorst U, Rekik A, Engels K, Bier C, Knau- er SK. Nucleocytoplasmic shuling and the biological acvity of mouse survivin are regulated by an acve nuclear export signal. Traffic 2006;7:1461–1472.
11. Fornerod M, Ohno M, Yoshida M, Maaj IW. CRM1 is an export receptor for leucine-rich nuclear export signals. Cell 1997;90:1051–1060.
12. Stade K, Ford CS, Guthrie C, Weis K. Exporn 1 (Crm1p) is an essenal nuclear export factor. Cell 1997;90:1041–1050.
13. Fukuda M, Asano S, Nakamura T, Adachi M, Yoshida M, Yanagida M, Nishida E. CRM1 is responsible for intracellu- lar transport mediated by the nuclear export signal. Nature 1997;390:308–311.
14. la Cour T, Gupta R, Rapacki K,
Skriver K, Poulsen FM, Brunak S. NESbase version 1.0:
a database of nuclear export signals. Nucleic Acids Res 2003;31:393–396.
15. Kutay U, Gu nger S. Leucine-rich nuclear-export signals:
born to be weak. Trends Cell Biol 2005;15:121–124.
16. Görlich D, Kutay U. Transport between the cell nucleus and the cytoplasm. Annu Rev Cell Dev Biol 1999;15:607–660.
17. Wolff B, Sanglier JJ, Wang Y. Leptomycin B is an inhibitor of nuclear export: inhibion of nucleo-cytoplasmic transloca-
on of the human immunodeficiency virus type 1 (HIV-1) Rev protein and Rev-dependent mRNA. Chem Biol 1997;4:139–
147.
18. Knauer SK, Kramer OH, Knosel T, Engels K, Rodel F, Kovacs AF, Dietmaier W, Klein-Hitpass L, Habtemichael N, Schweitzer A, Brieger J, Rodel C, Mann W, Petersen I, Heinzel T et al. Nu- clear export is essenal for the tumor-promong acvity of survivin. Faseb J 2007;21:207–216.
19. Henderson BR, Eleheriou A. A comparison of the acvity, sequence specificity, and CRM1-dependence of different nuclear export signals. Exp Cell Res 2000;256:213–224.
20. Paraskeva E, Izaurralde E, Bischoff FR, Huber J, Kutay U, Hart- mann E, Lührmann R, Görlich D. CRM1-mediated recycling of snurporn 1 to the cytoplasm. J Cell Biol 1999;145:255–264.
21. Askjaer P, Bachi A, Wilm M, Bischoff FR, Weeks DL, Ogniewski V, Ohno M, Niehrs C, Kjems J, Maaj IW, Fornerod M. RanGTP- regulated interacons of CRM1 with nucleoporins and a shuling DEAD-box helicase. Mol Cell Biol 1999;19:6276–6285.
22. Bischoff FR, Görlich D. RanBP1 is crucial for the release of RanGTP from imporn beta-related nuclear transport factors.
FEBS Le 1997;419:249–254.
23. Wen W, Meinkoth JL, Tsien RY, Taylor SS. Idenficaon of a signal for rapid export of proteins from the nucleus. Cell 1995;82:463–473.
24. Chantalat L, Skoufias DA, Kleman JP, Jung B, Dideberg O, Margolis RL. Crystal structure of human survivin reveals a bow
e-shaped dimer with two unusual alpha-helical extensions. Mol Cell 2000;6:183–189.
25. Muchmore SW, Chen J, Jakob C, Zakula D, Matayoshi ED, Wu W, Zhang H, Li F, Ng SC, Aleri DC. Crystal structure and muta- genic analysis of the inhibitor-of-apoptosis protein survivin. Mol Cell 2000;6:173–182.
26. Verdecia MA, Huang H, Dul E, Kaiser DA, Hunter T, Noel JP.
Structure of the human an-apoptoc protein survivin reveals a dimeric arrangement. Nat Struct Biol 2000;7:602–608.
27. Hess JF, Budamagunta MS, Voss JC, FitzGerald PG. Structural characterizaon of human vimenn rod 1 and the sequencing of assembly steps in intermediate filament formaon in vitro using site-directed spin labeling and electron paramagnec resonance.
J Biol Chem 2004;279:44841–44846.
28. Yagi K, Sugiura N, Oama H, Oishi N. Structure and funcon of D-amino acid oxidase. VI. Relaon between the quaternary struc- ture and the catalyc acvity. J Biochem (Tokyo) 1973;73:909–
914.
29. Gasset M, Alfonso C, Neira JL, Rivas G, Pajares MA. Equilib- rium unfolding studies of the rat liver methionine adenosyltrans- ferase III, a dimeric enzyme with intersubunit acve sites. Bio- chem J 2002;361(Pt 2):307–315.
30. Shi Y. Survivin structure: crystal unclear. Nat Struct Biol 2000;7:620–623.
31. Temme A, Rodriguez JA, Hendruschk S, Gunes S, Weigle B, Schakel K, Schmitz M, Bachmann M, Schackert G, Rieber EP.
Nuclear localizaon of Survivin renders HeLa tumor cells more sensive to apoptosis by inducon of p53 and Bax. Cancer Le
2007;250:177–193.
32. Weis K. Regulang access to the genome: nucleocytoplasmic transport throughout the cell cycle. Cell 2003;112:441–451.
33. Izaurralde E, Kutay U, von Kobbe C, Maaj IW, Görlich D. The asymmetric distribuon of the constuents of the Ran system is essenal for transport into and out of the nucleus. EMBO J 1997;16:6535–6547.
80
Survivin Dimerization Antagonizes Nuclear Export
34. Bischoff FR, Klebe C, Kretschmer J, Wi nghofer A, Ponst- ingl H. RanGAP1 induces GTPase acvity of nuclear Ras-relat- ed Ran. Proc Natl Acad Sci U S A 1994;91:2587–2591.
35. Engelsma D, Bernad R, Calafat J, Fornerod M. Supraphysi- ological nuclear export signals bind CRM1 independently of RanGTP and arrest at Nup358. EMBO J 2004;23:3643–3652.
Supplementary Figure 1
B A
Supplementary Figure 2
Supplementary Figure 1. Influence of NES mutation, cell line and N or C-terminal YFP-tagging on the subcellular localiza- tion of Survivin. (A) Subcellular localizaon of YFP-SurvivinL96A/L98A in MCF7 and HeLa cells.
(B) Subcellular localizaon in MCF7 cells of YFP- SurvivinL96A and SurvivinL96A-YFP. Bar, 10 μm.
Supplementary Figure 2. Influence of expression level on nucleocyto- plasmic ratios of Survivin(1-109) and Survivin(1-142), carrying or not carry- ing the F101A/L102A mutation. Expres- sion levels are approximated by cytoplasmic pixel intensies in arbritrary units (AU). Coef- ficients (coef) are from least sum of squares linear fits. For Survivin(1-109) the data is from one of two experiments also shown in Fig 5B.
