Stimulation of Nuclear Export and Inhibition of Nuclear Import by a Ran Mutant Deficient in Binding to Ran-binding Protein 1*

Receptor-mediated nucleocytoplasmic transport is dependent on the GTPase Ran and Ran-binding protein 1 (RanBP1). The acidic C terminus of Ran is required for high affinity interaction between Ran and RanBP1. We found that a novel Ran mutant with four of its five acidic C-terminal amino acids modified to alanine (RanC4A) has an ∼20-fold reduced affinity for RanBP1. We investigated the effects of RanC4A on nuclear import and export in permeabilized HeLa cells. Although RanC4A promotes accumulation of the nuclear export receptor CRM1 at the cytoplasmic nucleoporin Nup214, it strongly stimulates nuclear export of GFP-NFAT. Since RanC4A exhibits an elevated affinity for CRM1 and other nuclear transport receptors, this suggests that formation of the export complex containing CRM1, Ran-GTP, and substrate is a rate-limiting step in export, not release from Nup214. Conversely, importin α/β-dependent nuclear import of bovine serum albumin, coupled to a classical nuclear localization sequence is strongly inhibited by RanC4A. Inhibition can be reversed by additional importin α, which promotes the formation of an importin α/β complex. These results provide physiological evidence that release of Ran-GTP from importin β by RanBP1 and importin α is critical for the recycling of importin β to a transport-competent state.

Molecular transport between the cytoplasm and the nucleus occurs through nuclear pore complexes (NPCs), 1 large supramolecular structures that span the nuclear envelope (for review, see Refs. 1 and 2). Small molecules such as ions and metabolites cross the NPC by passive diffusion. In contrast, most macromolecules are transported through the NPC by signal-and energy-dependent processes. Much of the signal-dependent nucleocytoplasmic transport is mediated by nucleocy-toplasmic shuttling receptor proteins of the importin ␤/karyopherin ␤ family. These receptors are thought to transfer their cargoes between the nucleus and the cytoplasm by sequentially interacting with a series of NPC proteins (nucleoporins; for review, see Refs. 3 and 4).
Transport mediated by importin/karyopherin ␤-type transport receptors is dependent on the small GTPase Ran, which shuttles between the cytoplasm and the nucleus (3). Due to the segregation of the RanGEF in the nucleus and the RanGAP in the cytoplasm, the GTP-bound form of Ran is thought to be concentrated in the nucleus and the GDP-bound form in the cytoplasm. Ran-GTP, which binds to all importin ␤-type import receptors, promotes the association of cargoes with export receptors and the dissociation of cargoes from import receptors in vitro and thus appears to regulate receptor loading and unloading in the nucleus (see below).
The best characterized nuclear import pathway involves cargo proteins carrying a basic amino acid-rich nuclear localization sequence (NLS). A trimeric import complex is formed by the binding of the cargo to the adapter protein importin ␣, which interacts with the import receptor importin ␤ via its importin ␤-binding (IBB) domain (3). After the import complex is translocated to the nucleoplasmic side of the NPC, Ran-GTP is thought to dissociate importin ␣ from importin ␤ and importin ␤ from nucleoporins. The resulting Ran-GTP-importin ␤ complex is exported back to the cytoplasm, whereas importin ␣ is recycled to the cytoplasm by the export receptor CAS (5).
In the most extensively studied nuclear protein export pathway, Ran-GTP and a cargo protein containing a leucine-rich nuclear export sequence (NES) bind cooperatively to the export receptor CRM1 (6,7). After translocation of this trimeric export complex to the cytoplasmic side of the NPC, RanBP1 and the related Ran-GTP binding domains of the nucleoporin Nup358/ RanBP2 are thought to disassemble the export complex (8,9) as well as release it from a terminal cytoplasmic nucleoporin in the pathway, Nup214/CAN (10).
RanBP1 is a cytosolic protein that further enhances the rate of GTP hydrolysis on free Ran-GTP that is stimulated by the GTPase-activating protein RanGAP (11,12). When Ran-GTP is complexed with import or export receptors, such as occurs when import receptors are recycled or export complexes are translocated to the cytoplasm, the Ran-GTP is GAP-insensitive. In this case, RanBP1 strongly promotes GTPase stimulation by RanGAP because it releases Ran-GTP from the receptors (8,9). Thus, RanBP1 may play a crucial role in the termination of nuclear export reactions in the cytoplasm by promoting the dissociation of Ran from the receptors and, as a consequence, the hydrolysis of Ran-GTP.
An acidic stretch at the C terminus of Ran is important for its interaction with RanBP1, since a mutant in which the last six amino acids of Ran are removed (Ran⌬DEDDDL; Ran⌬C) does not interact with Ran-GTP and RanBP1 in the yeast two-hybrid system (13) and on blot overlays (13,14). Moreover, Ran⌬C has a greatly reduced affinity for RanBP1 in solution (15,16). Whereas the acidic tail of Ran strongly promotes the interaction with RanBP1, it is inhibitory for binding to importin ␤ (14,16), perhaps due to competition with an acidic loop in importin ␤ for binding to a basic patch on Ran (17).
We have characterized a novel Ran mutant, RanC4A, which has four of its five C-terminal acidic residues exchanged with alanine (18), to examine the role of Ran-RanBP1 interactions in nuclear import and export in vitro. As Ran⌬C, RanC4A has a reduced affinity for RanBP1 and an increased affinity for importin ␤ and also for CRM1. Using well established in vitro nuclear transport systems, we show that RanC4A strongly stimulates CRM1-mediated nuclear export of GFP-NFAT, whereas it inhibits importin ␣-dependent nuclear import by importin ␤. Our findings shed light on rate-limiting steps of CRM1-mediated export and provide physiological evidence for a role of RanBP1 in the recycling of import receptors.

EXPERIMENTAL PROCEDURES
Cell Culture-HeLa cells were grown on either plastic dishes or coverslips in Dulbecco's modified Eagle's medium or in suspension in Joklik's modified S-MEM. Both media contained 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM glutamine. All tissue culture reagents were from Life Technologies, Inc.
Recombinant Transport Factors-The expression plasmid (pET3d (19)) for C-terminally mutated RanC4A (DEDDDL mutated to AAADAL, amino acids 211-216) used for the experiments in Table I and Figs. 1-3 was constructed by polymerase chain reaction using appropriate oligonucleotides containing NcoI/BamHI sites for cloning. Wildtype and mutant Ran were purified as described (20). Purified Ran was Ͼ95% pure, as judged by gel electrophoresis. Its quality was assessed by determining the concentration of GDP bound to Ran using reversephase HPLC analysis on a C-18 HPLC column (21).
The wild-type Ran and RanQ69L used for all experiments except Table I were purified as described (22), with minor modifications. RanC4A used for the experiments in Figs. 4 and 5 was expressed in Escherichia coli BL21 (DE3) as a GST fusion protein. After purification using glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech), the GST tag was cleaved with biotin-coupled thrombin (Novagen), as described by the manufacturer, followed by the removal of thrombin with streptavidin-agarose (Sigma) and of cleaved GST with glutathione-Sepharose 4B beads. Ran⌬C was expressed as a GST fusion protein and thrombin-cleaved as described above. Proteins were dialyzed against transport buffer (20 mM Hepes, pH 7.3, 110 mM KOAc, 2 mM Mg(OAc) 2 , 1 mM EGTA). No functional difference was evident between RanC4A obtained by cleavage from a GST fusion protein as compared with expression as an unfused protein.
Importins ␣ and ␤ (23), CRM1 (24), and mouse RanGAP (25) were prepared as described and dialyzed against transport buffer containing 2 mM dithiothreitol and 1 g/ml each of leupeptin, pepstatin, and aprotinin. RanBP1 (kindly provided by Dr. Iris Schmitt) was expressed in E. coli from pET11d. After precipitation of proteins from a bacterial lysate with ammonium sulfate (50 -75% cut), RanBP1 was further purified by gel filtration and ion exchange chromatography using a MonoQ column and dialyzed against transport buffer. GST-RanBP1 (15), Rna1p (26), and RCC1 (RanGEF; Ref. 27) were purified as described. All proteins were frozen in liquid nitrogen and stored at Ϫ80°C.
Intrinsic GTPase Activity-The intrinsic GTPase activity was measured by reverse-phase HPLC analysis (21). 400 M Ran-GTP was incu-bated at 37°C in 50 mM Tris, 5 mM MgCl 2 , 5 mM dithioerythritol, pH 7.4, at 37°C. Aliquots were removed at regular intervals and applied to HPLC analysis to measure the protein-bound GTP and GDP concentration. The HPLC data were expressed as ln((GTP ϩ GDP)/GTP) and plotted as a function of time. The GTP hydrolysis rate, k cat (min Ϫ1 ), was calculated by linear regression of the data with Kaleidagraph 3.0.5 (Synergy Software).
Interaction of Ran and Ran Mutants with Regulatory Proteins-The RanGAP-stimulated GTP hydrolysis of Ran analyzed in Table I was measured by radioactive filter assays as described previously (30) using different concentrations of the RanGAP of Schizosaccharomyces pombe, Rnalp.
The GST-RanBP1-Ran interaction was analyzed by determining the dissociation constant K d ϭ k off /k on using fluorescence spectroscopy (15). The association kinetics were measured under pseudo-first-order conditions with 200 nM Ran-mant-Gpp(NH)p and increasing concentrations of GST-RanBP1 (1, 2, 3, 4, 5, and 6 M). The signal of the Ran-bound fluorescence-labeled nonhydrolyzable GTP analog mant-Gpp(NH)p (excitation, 350 nm; emission, 440 nm) was followed in a SX16MV stopped flow system (Applied Photophysics). The kinetic constant k on of the Ran-GST-RanBP1 association was calculated with Kaleidagraph 3.0.5 (Synergy Software) as described previously (15). For the measurement of the dissociation kinetics (k off ), a 1:1 complex between 200 nM Ran-mant-Gpp(NH)p and GST-RanBP1 was generated in 20 mM K 2 HPO 4 /KH 2 PO 4 , pH 7.4, 5 mM MgCl 2 , 5 mM dithioerythritol, in a titration experiment at 25°C. The titration was followed in a SPEX Fluoromax fluorescence spectrometer (excitation, 350 nm; emission, 440 nm; Ref. 15). After formation of the GST-RanBP1-Ran complex, it was dissociated by adding a 50-fold molar excess of unlabeled Ran-Gpp(NH)p. The data of the dissociation reaction were fitted as firstorder exponentials with Kaleidagraph 3.0.5 to calculate the kinetic constant k off . The two kinetic constants were used to calculate the equilibrium constant K d ϭ k off /k on , which characterizes the affinity between Ran and RanBP1.
To analyze the RanGAP activity for all experiments other than in Table I, Ran was loaded with [␥-32 P]GTP as described (10). RanGAP assays (except for Table I; see above) were performed essentially as described (32), using 30 -60 nM total Ran, 4 nM RanGAP, 400 M GTP, and, in some reactions, a 20 M concentration of the NES peptide from the minute virus of mouse NS2 protein (CVDEMTKKFGTLTIHDTEK (32,33)) or a 20 M concentration of the NLS peptide from the SV40 large T antigen (CGGGPKKKRKVED). After 10 min at 20°C, reactions were stopped by adding 1 ml of stop buffer (7% charcoal, 10% ethanol, 0.1 M HCl, 10 mM NaH 2 PO 4 ). After centrifugation, the released [ 32 P]phosphate in 700 l of the supernatant was measured by scintillation counting. Background values of samples with radioactive Ran added immediately to the stop buffer were subtracted, and GTP hydrolysis was expressed as the percentage of the maximal value of recovered radioactivity, as obtained in the absence of inhibiting transport receptors.
Nuclear Transport Assays-For nuclear import assays on adherent cells, HeLa cells were grown on coverslips and permeabilized with 30 g/ml digitonin (Calbiochem) in transport buffer containing 2 mM dithiothreitol and 1 g/ml each of leupeptin, pepstatin, and aprotinin. The permeabilized cells were incubated in nuclear import reactions for 30 min at 30°C as described (34), except the import substrate was a fusion protein between the amino terminus of human SRP␣ and ␤-galactosidase (IBB-␤Gal; 112 g/ml; modified from Ref. 35; kindly provided by Dr. S. Lyman). Cells were fixed with 3.7% formaldehyde in phosphate-buffered saline, permeabilized with 0.2% Triton X-100 in phosphate-buffered saline, and then analyzed by indirect immunofluorescence microscopy using a monoclonal antibody against ␤-galactosidase (Promega) to detect the IBB-␤Gal. Nuclear import was visualized by confocal microscopy using a Bio-Rad MRC1024 confocal unit attached to a Zeiss Axiovert S100TV microscope. All digital data sets were processed identically using Adobe Photoshop (Adobe). When suspension HeLa cells were used for import assays, the reactions were performed as described in Ref. 36, containing 2.5 mg/ml cytosol, unless otherwise indicated. The nuclear import substrate for suspension assays was Cy5-BSA-NLS, which was prepared as described (37). Flow cytometry was performed using a dual laser cytometer (FACSCaliber, Becton Dickinson). Reactions were standardized by assigning a fluorescent value of 100 arbitrary units to a reaction resulting in optimal import. The nuclear export assay using GFP-NFAT-transfected HeLa cells was carried out as described (37).
Antibodies, Immunoprecipitations, and Western Blotting-The anti-CRM1 peptide antibody was prepared and affinity-purified as described (37). The monoclonal antibody RL2 is described elsewhere (38). The monoclonal anti-importin ␤-antibody (ascites fluid; Affinity BioReagents, Inc.) used for immunoprecipitations was diluted 1:1 with 5 mg/ml BSA in transport buffer and stored at Ϫ20°C. Monoclonal anti-Ran antibody was from Transduction Laboratories. The polyclonal antiimportin ␤ and anti-importin ␣ antibodies were raised in rabbits against recombinant proteins and used after affinity purification.
For coimmunoprecipitation of nucleoporins, 3 ϫ 10 6 digitonin-permeabilized cells were solubilized for 20 min on ice in 1 ml of Nonidet P-40 buffer (1% Nonidet P-40, 50 mM Tris-HCl, pH 8, 300 mM NaCl, 5 mM EDTA, 5 mM EGTA, 15 mM MgCl 2 , 60 mM ␤-glycerophosphate, 100 M KF, 100 M NaVO 4, 2 mM dithiothreitol, 1 g/ml each of leupeptin, pepstatin, aprotinin). After centrifugation at 14,000 ϫ g for 20 min at 4°C, the lysate was precleared with protein G-agarose beads (Life Technologies) for 30 min, and anti-CRM1 antibody (10 g/ml) was added to the supernatant. Samples were incubated at 20°C for 1.5 h, and immunoreactive proteins were collected with protein G-agarose. The beads were washed four times with Nonidet P-40 buffer, and precipitated proteins were analyzed by SDS-polyacrylamide gel electrophoresis followed by immunoblotting. For coimmunoprecipitation of importin ␤, 2 ϫ 10 6 permeabilized cells were incubated with cytosol, import ligand, and an ATP-regenerating system (i.e. under the conditions used in nuclear import reactions). After centrifugation (1,000 ϫ g for 1 min and 14,000 ϫ g for 5 min), the pelleted nuclei were discarded, and the supernatant was precleared with protein G-agarose beads for 30 min. Anti-importin ␣ antibody was added to 10 g/ml. Samples were then incubated for 1.5 h at 20°C, and immunoreactive proteins were collected with protein G-agarose. The beads were washed four times with transport buffer, and precipitated proteins were analyzed by SDSpolyacrylamide gel electrophoresis followed by immunoblotting.

RESULTS
Biochemical Characterization of RanC4A-The acidic C terminus of Ran promotes the interaction between Ran and RanBP1 (15,16,39), whereas it appears to impede the inter-action between Ran and importin ␤ (16,40). To investigate the importance of this Ran region for nuclear import and export, we have characterized a novel Ran mutant, RanC4A, which has its C-terminal sequence DEDDDL (amino acids 211-216) mutated to AAADAL. We expected that this mutant, where the acidic C terminus is largely neutralized, would be structurally more similar to wild-type Ran than the previously described C-terminal deletion mutant Ran⌬C, where different effects on nuclear import have been described in different assay systems (13,41). Since RanC4A exhibited a much stronger effect on nuclear export as compared with Ran⌬C (see below), we have focused our functional analysis on RanC4A, although Ran⌬C was also analyzed in most experiments to attempt to resolve the discrepancies previously reported with functional assays (13,41).
The biochemical characteristics of RanC4A, as compared with wild-type Ran (RanWT) and the previously analyzed RanQ69L mutant (the standard for our export inhibition studies; see below), are summarized in Table I (see also Ref. 18). As described before, RanQ69L had a strongly reduced intrinsic GTPase activity and was largely insensitive to RanGAP (31). RanC4A had an about 3-fold increased GTPase activity, when stimulated by RanGAP (0.261 min Ϫ1 as compared with 0.074 min Ϫ1 for RanWT). The k cat /K m of RanC4A for the exchange factor RanGEF (RCC1), on the other hand, was decreased 3-fold as compared with RanWT. A more dramatic change was observed in the affinity of RanC4A for RanBP1. As a result of an increased off-rate, the apparent affinity of RanC4A is ϳ20fold lower, as compared with wild-type Ran (74.4 versus 3.7 nM). Thus, the mutation in RanC4A, which neutralizes the acidic C terminus, most strongly affects the binding of the protein to RanBP1. Very likely, RanC4A also has a strongly diminished affinity for the Ran-binding domains of Nup358/ RanBP2. Our findings are consistent with the strong decrease in affinity of the C-terminal deletion mutant Ran⌬C for RanBP1 that was reported previously (15,16,39).
CRM1 Accumulates at Nup214 in the Presence of RanC4A-We used the RanC4A mutant to test the model that RanBP1 promotes the release of the export complex containing CRM1, Ran, and substrate from Nup214 at the cytoplasmic side of the NPC (10). This model was based on the finding that CRM1 accumulates at Nup214 in permeabilized cells preincubated with the RanQ69L mutant, reflecting an apparent transport arrest at a terminal nucleoporin binding site. Since CRM1 can be released from the NPC in these cells with exogenous RanBP1, the export inhibition is presumed to be due to a The rate constant k cat (min Ϫ1 ) of the intrinsic GTP hydrolysis was calculated by linear regression using a plot (ln(GTP/GTP ϩ GDP) versus time) with time points at up to 9 h (not shown).
b The RanGAP-stimulated Ran-GTP hydrolysis was determined in filter assays using 1 M Ran and 0.2 nM Rna1p and treated as first order kinetics to determine the rate constant k obs .
c The association rate constant k on was calculated by linear regression using a plot (k obs versus GST-RanBP1 concentration) with GST-RanBP1 concentrations of up to 6 M (not shown).
d The reaction rate constant of the Ran-mant-Gpp(NH)p-GST-RanBP1 complex dissociation, k off , was calculated as first-order exponential fits using a plot (relative fluorescence versus time) with readings for up to 100 min (not shown). e K D ϭ k off /k on . f The reaction rates of RCC1-catalyzed GDP dissociation on Ran-mant-GDP were measured as a decrease in fluorescence, and the kinetic data were fitted exponentially to calculate the first-order rate constant k obs . The values for k obs were plotted against the Ran-GDP concentration, and this plot was used to determine the kinetic constants K M and k cat (Ref. 31; data not shown). See "Experimental Procedures" for details.
RanQ69L-mediated occlusion of the Ran-binding domains of Nup358/RanBP2 that normally mediate the release step.
Permeabilized cells were incubated with RanWT, RanQ69L, or RanC4A and subsequently analyzed for CRM1 and Ran distribution by indirect immunofluorescence microscopy. As shown in Fig. 1A, after incubation of cells with RanWT, CRM1 was largely depleted from the nuclei. In contrast, CRM1 accumulated at the nuclear envelope in the presence of RanQ69L, as described before (10). Incubation of cells with RanC4A also leads to the accumulation of CRM1 at the nuclear envelope, although to a lesser extent than with RanQ69L. Also, more CRM1 remained in the nucleoplasm in the presence of RanC4A, as compared with RanWT or RanQ69L. RanWT was largely lost from nuclei after the reaction (and after washing the cells), whereas RanQ69L became concentrated at the nuclear envelope, as described before (42). RanC4A, on the other hand, was detectable at a high level throughout the nucleus after the reaction. This retention of RanC4A in the nucleus probably resulted from an increased affinity for intranuclear importin ␤-related transport receptors (see below).
To identify the nucleoporins associated with CRM1 in cells treated with the Ran mutants and RanWT, we solubilized cells and carried out coimmunoprecipitation analysis with anti-CRM1 antibodies. As shown previously (10), Nup214 coprecipitated with CRM1 after incubation of permeabilized cells with RanQ69L, but not after incubation with RanWT (Fig. 1B, top). Incubation with RanC4A also resulted in coprecipitation of CRM1 with Nup214, although to a somewhat lesser extent than with RanQ69L. Immunoprecipitation of CRM1 was very similar under the different conditions (Fig. 1B, bottom). No other nucleoporins were detected in immunoprecipitates with CRM1 after pretreatment of cells with RanC4A, when samples were analyzed by Western blotting with the RL2 monoclonal antibody that reacts with a group of FG repeat nucleoporins (data not shown). Incubation of permeabilized cells with the C-terminal deletion mutant Ran⌬C also resulted in increased coprecipitation of Nup214 (data not shown).
Thus, these data indicate that CRM1 accumulates at Nup214 in the presence of RanC4A, which is strongly deficient in its interaction with RanBP1 and (by extension) the Ran-binding domains of Nup358/RanBP2. These findings provide independent support for our previous conclusion that the interaction of RanBP1 and/or the Ran-binding domains of Nup358/RanBP2 with the export complex is required for efficient release of CRM1 from Nup214 in a terminal step of export (10).
Stimulation of Nuclear Export by RanC4A-We next tested the effect of RanC4A and Ran⌬C on in vitro nuclear export mediated by CRM1. Surprisingly, RanC4A did not inhibit nuclear export of GFP-NFAT when transport was assayed in the presence of cytosol, which contains RanWT (data not shown). This is in contrast to RanQ69L, which strongly inhibited nuclear export under similar conditions (37). Furthermore, RanC4A strongly stimulated nuclear export in permeabilized cells when the analysis was done without added cytosol and under conditions where Ran is the only rate-limiting factor (Ref. 37; Fig. 2A). Stimulation of nuclear export by RanC4A was even more efficient than by RanWT, since high levels of export were obtained with much lower levels of RanC4A as compared with RanWT ( Fig. 2A). The C-terminal deletion mutant Ran⌬C also promoted nuclear export to some extent. Ran⌬C, however, did not yield the degree of stimulation of RanWT or RanC4A, even at saturating concentrations ( Fig.  2A). Release of GFP-NFAT from the nucleus in the presence of RanC4A still followed a physiological pathway, since export could be inhibited by wheat germ agglutinin (43), which occludes the NPC; by leptomycin B (44), which inhibits cargo binding to CRM1; and by a short peptide corresponding to the nuclear export sequence of the minute virus of mice (32), which competes for binding to CRM1 (data not shown). The stimulation of nuclear export obtained with RanC4A suggests that the release of the nuclear export complex from the cytoplasmic side of the NPC, which appears to be impaired with RanC4A, is not a rate-limiting step in export in our assay (see "Discussion").
It has been observed previously by blot overlay analysis that Ran⌬C binds more strongly to importin ␤ (and potentially to other importin ␤ family members) than does RanWT (16,40). We therefore tested whether the greater stimulation of nuclear export by RanC4A, as compared with RanWT, resulted from an increased affinity of the Ran mutant for CRM1, which could enhance the formation of a nuclear export complex. Since the binding of nuclear transport receptors to Ran-GTP inhibits GTP hydrolysis stimulated by RanGAP, RanGAP resistance can be used to determine the apparent affinities of receptors for Ran-GTP (8). In the case of CRM1, a trimeric GAP-resistant complex is formed by export substrate, Ran-GTP and CRM1 (32). We used a peptide comprising the strong nuclear export sequence of the NS2 protein of minute virus of mice (32) to generate a CRM1 export complex. As shown in Fig. 2B, halfmaximal inhibition of GTP hydrolysis could be obtained at much lower concentrations of CRM1 in reactions containing RanC4A, as compared with reactions containing RanWT (ϳ2 nM versus 40 nM). The apparent affinity of CRM1 for RanWT that we measured is consistent with the one obtained previously under similar conditions (32). Almost no inhibition of GTP hydrolysis was observed when the NES peptide was omit- ted in the reaction (data not shown). In contrast to CRM1, we did not detect a significant difference in the affinity of RanWT and RanC4A for the export receptor CAS (data not shown).
These data indicate that RanC4A can increase the rate of formation of a CRM1 export complex, as compared with RanWT. The ability of RanC4A to stimulate nuclear export at lower concentrations than RanWT suggests that the initial formation of the export complex containing export substrate, CRM1, and Ran-GTP is the rate-limiting step in export. Although the release of the export complex from the cytoplasmic side of the NPC is impaired with RanC4A, this step is not rate-limiting in the export cycle in our conditions. Apparently, the 20-fold reduced affinity of RanC4A for RanBP1 (or the Ran-binding domains of Nup358/RanBP2) still suffices for release of the export complex from the cytoplasmic side of the NPC.
Inhibition of Nuclear Import by RanC4A-RanBP1 has been proposed to mediate the release of Ran-GTP from the import receptor importin ␤ and from CAS, the export receptor for importin ␣, to promote the recycling of nuclear import receptors (8,9). Consistent with the prediction that RanBP1 has an important role in the nuclear import cycle, expression of a C-terminal deletion mutant of Ran has been found to inhibit nuclear import in vivo (41). By contrast, the same mutant was found to stimulate nuclear import similar to RanWT under certain in vitro conditions (13). To address this discrepancy, we tested the effects of RanC4A and Ran⌬C on nuclear import in permeabilized cells. In the presence of cytosol, import of BSA-NLS was strongly inhibited by RanC4A (Fig. 3A), but not by RanWT (data not shown). The C-terminal deletion mutant Ran⌬C also strongly inhibited nuclear import in the presence of cytosol (Fig. 3B).
The 20-fold reduced affinity of RanC4A for RanBP1 suggests that nuclear import in the presence of RanC4A might be restored to normal levels by the addition of high levels of exogenous RanBP1. Although we found that RanBP1 did not stimulate nuclear import when added together with RanC4A, a meaningful conclusion could not be drawn from this experiment, since RanBP1 itself inhibits nuclear import under similar conditions (data not shown).
To further investigate the basis for inhibition of nuclear import by RanC4A, we used RanGAP assays (as in Fig. 2B) to determine the apparent affinity of RanWT and RanC4A for importin ␤. As shown in Fig. 3C, importin ␤ has a higher apparent affinity for RanC4A than for RanWT, although the difference is not as pronounced as seen with CRM1. Halfmaximal inhibition of GTP hydrolysis was obtained with ϳ0.4 nM importin ␤ and RanC4A, as compared with ϳ2 nM importin ␤ and RanWT. Our results with RanC4A indicate that the C terminus of Ran negatively affects the interaction of the pro- tein with importin ␤, confirming previous findings with Ran⌬C (16,40).
A priori, the increased affinity of RanC4A for importin ␤ is unlikely to inhibit the nuclear import phase of the import cycle. Consistent with this premise, we found by coprecipitation experiments that the putative Ran-GTP-mediated release of importin ␤ from nucleoporins including Nup153 (45,46) does not appear to be inhibited by RanC4A (data not shown). However, the increased affinity of these proteins may impair efficient recycling of import receptors and their release from Ran-GTP, thereby indirectly inhibiting nuclear import. Impaired receptor recycling might be manifest by an increase in the intranuclear level of the receptors. To examine whether transport receptors accumulate in the nucleus in the presence of RanC4A, cells were incubated in the presence of cytosol with RanWT or RanC4A and subsequently fractionated into a nuclear and a cytosolic pool. RanC4A became somewhat concentrated in the nuclear fraction as compared with RanWT, consistent with the immunofluorescence data in Fig. 1A. However, importin ␣, importin ␤, and CAS were largely or exclusively recovered in the cytosolic fraction, irrespective of the type of Ran used during the incubation (data not shown). Hence, RanC4A does not lead to nuclear accumulation (or as a consequence, cytosolic depletion) of either one of those transport factors.
RanC4A also could inhibit the recycling of import receptors by inhibiting the release of Ran-GTP from the Ran-GTP-importin ␤ complex that is exported from the nucleus, which apparently is needed to form an import-competent importin ␣/␤ complex. In vitro studies have suggested that the release of Ran-GTP from importin ␤ is mediated by RanBP1 and importin ␣ (8,9). To investigate whether RanC4A caused any changes in the cytosolic levels of the importin ␣/␤ complex generated by importin ␤ recycling, we incubated permeabilized cells with RanWT or RanC4A in the presence of cytosol, sedimented the nuclei by centrifugation, and immunoprecipitated proteins from the supernatant (i.e. the cytosolic fraction) using an antibody against importin ␣. As shown in Fig. 4A, RanC4A strongly decreased the level of coprecipitated importin ␤, as compared with an incubation with RanWT, whereas levels of precipitated importin ␣ were similar in both conditions. Therefore, it appears that RanC4A impairs the release of Ran-GTP from cytosolic importin ␤, thereby causing a decrease in the amount of the transport-competent importin ␣/␤ complex in the transport assays.
We next analyzed the ability of importin ␣ to release the inhibition of GTP hydrolysis imposed by importin ␤ on either RanWT or RanC4A in vitro. After formation of a complex between importin ␤ and RanWT or RanC4A, GAP assays were performed with increasing concentrations of importin ␣ together with a high concentration of RanBP1, so that the latter would not be rate-limiting for stimulation of GTP hydrolysis. As shown in Fig. 4B, much higher concentrations of importin ␣ are required to release inhibition of GTP hydrolysis in the RanC4A-importin ␤ complex, as compared with the complex containing RanWT. These results suggest that under conditions where nuclear import is inhibited by RanC4A, importin ␣ may be a rate-limiting nuclear import factor; the cytosolic concentration of importin ␣ may be adequate only to generate reduced levels of the importin ␣/␤ complex.
Importin ␣ Restores RanC4A-inhibited Nuclear Import-We directly tested whether generation of an importin ␣/␤ complex was limiting for nuclear import in vitro by adding increasing amounts of importin ␣ to reactions in the presence of cytosol. In the absence of RanC4A, increasing concentrations of exogenous importin ␣ inhibited nuclear import of BSA-NLS, probably because free importin ␣ competed with importin ␣-␤ complexes for the cargo (Fig. 5A). In the presence of RanC4A, which strongly inhibited import in the absence of added importin ␣, low concentrations of exogenous importin ␣ restored import nearly to the uninhibited level. These results show that although the level of cytosolic importin ␣, as detected by Western blotting, does not change upon incubation of cells with RanC4A (see above), importin ␣ is rate-limiting for nuclear import when RanC4A is present.
A number of proteins have been shown to be imported into the nucleus by direct interaction with importin ␤ without the participation of importin ␣ (47,48). We used IBB-␤Gal (35) as an artificial, importin ␣-independent import substrate. As shown in Fig. 5B, nuclear import of IBB-␤Gal in the presence of cytosol was not inhibited by RanC4A. Furthermore, in the presence of importin ␤, strong import of IBB-␤Gal could be obtained by adding either RanWT or RanC4A (Fig. 5B).
These results suggested that IBB-␤Gal was able to function like importin ␣ to release Ran-GTP from the Ran-GTP-importin ␤ complex and allow importin ␤ recycling to an importcompetent state. To directly test this, we examined the inhibition of GTP hydrolysis imposed on RanWT by importin ␤. As shown in Fig. 5C, in the presence of RanBP1, IBB-␤Gal stimulated GTP-hydrolysis at similar concentrations as importin ␣. Therefore, the importin ␤-interacting domain of importin ␣ alone appears to promote GTP-hydrolysis on Ran-importin ␤ complexes. Interestingly, IBB-␤Gal was unable to stimulate nuclear import of BSA-NLS in the presence of RanC4A (data not shown), suggesting that the interaction of the IBB domain with importin ␤ directly leads to the formation of a competing import complex.
These results clearly demonstrate that importin ␣ becomes rate-limiting when nuclear import is inhibited by RanC4A. Inhibition appears to result from inefficient release of Ran-GTP from importin ␤ by importin ␣ (together with RanBP1) and, as a result, a decreased regeneration of import-competent importin ␣/␤ complexes. Since the IBB domain of importin ␣ efficiently stimulates GTP-hydrolysis on Ran-GTP-importin ␤ complexes, nuclear import of IBB-␤Gal (and probably of other, importin ␣-independent substrates) is not sensitive to inhibition by RanC4A. DISCUSSION We have characterized a novel Ran mutant, RanC4A, which has four out of five of the acidic residues at the very C terminus mutated to alanine. The major biochemical phenotype of this mutant is an ϳ20-fold reduced affinity of the GTP-bound form for RanBP1, as compared with wild-type Ran. This difference is small compared with the 2,000 -10,000-fold reduction in affinity of the C-terminal deletion mutant Ran⌬C for RanBP1 (15,16). In the latter case, the reduced affinity results from a 100-fold decrease of the association rate constant k on , combined with a 100-fold increase in the dissociation rate constant k off (15). For RanC4A, the mutation mainly affects k off , suggesting that the initial recognition of RanBP1 still is functional for RanC4A, although it is impaired in the case of Ran⌬C. The one acidic residue that remains unchanged in RanC4A (D214) and/or the C-terminal leucine may contribute to the remaining affinity of this mutant for RanBP1. Another prominent biochemical property of RanC4A is its increased affinity for the transport receptors CRM1 and importin ␤. For the latter protein, this may be explained by competition between an acidic loop of importin ␤ and the acidic tail of Ran for binding to a basic patch in Ran (17). Since CAS did not show differences in its apparent affinity for RanWT or RanC4A, this binding mechanism apparently does not apply to all importin ␤-type receptors.
The high affinity of RanC4A for CRM1, importin ␤ and possibly other transport receptors probably accounts for the enhanced retention of RanC4A in the nucleus as seen by immunofluorescence and Western blotting. The increased affinity of RanC4A for CRM1 would favor the formation of a nuclear export complex containing CRM1, Ran-GTP, and substrate, explaining the strong stimulation of export of GFP-NFAT by RanC4A, as compared with RanWT. Release of this complex by RanBP1 or the Ran-binding domains of Nup358/RanBP2 from its terminal binding site on the cytoplasmic site of the NPC appears to be impaired, since CRM1 accumulates at Nup214 as seen by coimmunoprecipitation with the anti-CRM1 antibody. Nevertheless, the submicromolar affinity of RanC4A or even the micromolar affinity of Ran⌬C for Ran-binding domains may still allow functional interactions because of the high local concentration of the latter at the NPC. As a result, termination of export can occur, even in the presence of Ran⌬C. However, it should be noted that Ran⌬C, with its drastically reduced affinity for RanBP1 or Ran binding domains of Nup358/RanBP2, is less efficient in promoting export as compared with RanC4A, with its rather modest decrease in affinity. This discrepancy may result from the difference in affinities of the two proteins for RanBP1 or from other, as yet unidentified, properties of Ran⌬C. Taken together, our results suggest that the initial formation of the export complex in the nucleus, which is promoted by RanC4A, is the most rate-limiting step in nuclear export.
Preincubation of permeabilized cells with RanQ69L leads to a similar accumulation of CRM1 at Nup214 as RanC4A, as described previously (10), although it occurs for a different reason. In this case, the Ran-binding sites of Nup358/RanBP2 probably are inactivated by the stable binding of RanQ69L, so that they are no longer available for releasing export complexes from the NPC. Efficient export under this condition depends on adding soluble RanBP1 (10). The fact that we obtain accumulation of CRM1 at Nup214 using two Ran mutants, which negatively affect the RanBP1-Ran interaction by two different mechanisms, strongly supports a role for RanBP1-like domains in release of the CRM1 export complex from Nup214.
Nuclear import of BSA-NLS in permeabilized cells supplemented with cytosol is strongly inhibited by RanC4A or Ran⌬C, in agreement with the inhibition of import seen in cultured cells transfected with Ran⌬C (41). Our results clearly show that the adapter protein importin ␣ becomes rate-limiting un- der conditions of import inhibition by RanC4A. In the study of Ren et al. (13), where Ran⌬C promoted nuclear import as efficiently as RanWT in vitro, transport in permeabilized cells was reconstituted with a high concentration of partially purified importin ␣, which can explain the discrepancy with our study, involving unfractionated cytosol.
Importin ␣ engages in at least two interactions important for nuclear import. First, it binds to proteins carrying a nuclear localization signal. Second, together with RanBP1 it dissociates Ran-GTP from importin ␤, thereby releasing the inhibition of GTP hydrolysis on Ran-GTP-importin ␤ complexes, after their export from the nucleus. This reaction is favored in the presence of NLS-substrate, as suggested before (49). The first function should not be affected by RanC4A, and we assume that importin ␣-BSA-NLS complexes are readily formed in the absence or presence of RanC4A. These complexes (or importin ␣ alone), however, appear to inefficiently release RanC4A from importin ␤, since higher concentrations of importin ␣ are required for the release reaction, even at saturating concentrations of RanBP1. This probably is due to the tight binding between importin ␤ and RanC4A, which impedes the dissociation of this complex and subsequent formation of importin ␣/␤ complexes. As a result, their level is reduced in the presence of RanC4A, as demonstrated by coimmunoprecipitation.
We showed that the dissociation of Ran-GTP-importin ␤ requires only RanBP1 plus the IBB domain of importin ␣, which interacts with importin ␤. A substrate like IBB-␤Gal, therefore, is efficiently imported into the nucleus in permeabilized cells supplemented with cytosol in the presence of RanC4A. Since IBB-␤Gal did not stimulate import of BSA-NLS in the presence of RanC4A, we conclude that the release reaction is directly coupled to the formation of import complexes containing IBB-␤Gal rather than BSA-NLS. Our results on nuclear import in vitro provide physiological evidence for a role of RanBP1 and importin ␣ in the recycling of importin ␤ after its export from the nucleus, supporting previous biochemical data (8,9), where it was shown that importin ␣ together with RanBP1 is required for the release of GTPase-inhibition on Ran-GTP-importin ␤ complexes.
Two types of importin ␤-dependent nuclear import occur in cells. The first requires an adapter protein such as importin ␣, whereas the second does not, since the import substrate binds directly to importin ␤. In the context of intact cells, the two types of import cargoes may compete for importin ␤. Cargoes like the human immunodeficiency virus Rev protein, which are adaptor-independent (50), could be kinetically favored in import because of their ability to release Ran-GTP from importin ␤ in a fashion that is directly linked to the formation of an importin ␤-cargo complex that is poised for nuclear import.
In conclusion, RanC4A is the first Ran mutant described to promote nuclear export of proteins carrying a leucine-rich NES and importin ␤-dependent/importin ␣-independent nuclear import but to strongly inhibit importin ␣-dependent transport. This mutant should prove useful for future functional studies of nuclear transport.