| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 43, 42699-42709, October 24, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Biological Sciences, Stanford University, Stanford, California 94305-5020
Received for publication, July 3, 2003 , and in revised form, August 7, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
36 instead of full-length Nup1p display specific defects in Kap95p localization and Kap95p-mediated nuclear import. We conclude that a high affinity binding site for Kap95p at the nuclear basket increases the translocation efficiency of Kap95p import complexes across the NPC. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
One prominent feature of the NPC is its peculiar composition, with almost half of its 30 Nups containing peptide repeats that feature phenylalanine-glycine (FG) motifs at their core. FG repeat regions in Nups provide binding sites for karyopherins (for example see Ref. 1). The structure of a karyopherin bound to peptides containing FG repeats shows that the phenyl side chains play a critical role in the binding reaction (2, 3). With 13 different FG Nups in the Saccharomyces cerevisiae NPC, and 8-32 copies of each per NPC (4), there may be as many as 150 or more FG Nups throughout the NPC. This multitude of docking sites for karyopherins may provide a continuum of steps that biases movement of karyopherins in one direction along the NPC. The particular direction of Kap movement (import or export) may be dictated by the differences in binding strength to proximal versus distal Nups that are asymmetrically distributed across the NPC (5, 6).
In yeast, there are 15 different karyopherins (7). The Kap95p-Kap60p heterodimer is responsible for import of cargoes containing classical nuclear localization signals (cNLSs) (8). In general, Kap60p binds to the cNLS of the cargo molecule and Kap95p interacts with Nups at the NPC to mediate translocation (9). Nuclear import in this pathway is terminated by binding of Gsp1p-GTP to Kap95p, which accelerates the dissociation rate of Kap95p from the Kap60p-cargo complexes (10) and reduces binding of Kap95p to Nups (1, 9).
The mammalian homologue of Kap95p (karyopherin 
, importin ) displays a higher affinity to nucleoporins that reside in the nuclear basket of the NPC than to those in the central scaffold region or in the cytoplasmic filaments (5). This is consistent with the notion that a gradient of affinities along the NPC contributes to movement of karyopherin . In this study, we demonstrate that a similar, but much steeper, gradient of binding affinities exists between Kap95p and FG Nups in S. cerevisiae. We investigate the role of this gradient in Kap95p-mediated nuclear import in vivo by disrupting a high affinity binding site for Kap95p import complexes in Nup1p. Nup1p is one of three Nups located in the nuclear basket of the yeast NPC.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
36), which was verified by Western blot analysis of total cell extracts with anti-Nup1p antibodies. Strains MRY122 through MRY133 were obtained by transforming the indicated yeast strains with the following NLS-GFP constructs: pCu416-cNLS-YFP, copper-inducible expression of the SV40 T-antigen NLS fused to YFP and inserted into the BamHI site of pCu416 (ATCC); Pho4p-NLS-GFP (pNLS), constitutive expression of Pho4p aa 140-166 fused to three copies of GFP (12); Nab2p-NLS-GFP (rgNLS), Nab2p aa 198-252 fused to the C terminus of GFP (13); pGAD-cNLS-GFP, constitutive expression of the SV40 T-antigen NLS fused to the Gal4p activation domain and GFP (14). Expression was confirmed by Western blot analysis of total cell extract with polyclonal anti-GFP antibodies. cNLS-YFP was induced with 100 µM CuSO4 for 2 h. MRY128 was obtained by mating YD05244 to MRY121, followed by selection of His+ kanR sporulants and verification by Western blot using anti-Nup1p and anti-Nup2p antibodies. Other double mutants (MRY129-131) were obtained in a similar manner.
|
Preparation of Recombinant Proteins—Recombinant proteins were expressed as GST fusions using the pGEX-2TK vector (Amersham Biosciences); construction and purification of each has been described previously (1, 10, 15) unless otherwise indicated. Nup1p fragments were constructed by PCR amplification of regions encoding the amino acids depicted in Fig. 4. Site-directed mutagenesis was performed according to the Stratagene QuickChangeTM kit. Wild-type HIS-Gsp1p was purified and charged with GTP as described previously (9).
|
Solution Binding Assay—All solution binding assays were performed using purified recombinant proteins in binding buffer (20 mM Hepes, pH 6.8, 150 mM KOAc, 2 mM Mg(OAc)2, 1 mM dithiothreitol, 0.1% Tween 20) as described previously (10). Briefly, GST-Nups were incubated in batch with glutathione-Sepharose beads, and aliquots were incubated with soluble proteins as indicated for 1 h at 4 °C. Beads were sedimented and unbound proteins collected. After brief washing, bound proteins were eluted with Laemmli sample buffer. Bound and unbound proteins were resolved by SDS-PAGE and visualized with Coomassie Blue.
Nucleoporin competition experiments were conducted as described above, except that Kap95p and Kap60p were pre-incubated with immobilized GST-Nup100pC prior to adding the soluble competitor. Competition experiments using Nup100p-Kap complexes formed in yeast extracts were performed as follows. GST-Nup100pC was immobilized on glutathione beads, and the beads were incubated with yeast extract. Unbound proteins were washed off as described in Ref. 1, and soluble Nups were incubated with pre-assembled Nup100pC-Kap complexes for 1 h at 4 °C. Unbound (and released) proteins were collected, the beads were washed, and bound proteins remaining were extracted from the beads with 1 M NaCl. All proteins were separated by SDS-PAGE and visualized with Coomassie Blue.
For the Gsp1p-GTP competition assays, GST-Nups were immobilized on glutathione beads and incubated with radiolabeled Kap95p (with or without Kap60p and cargo) for 20 min at room temperature. Unbound proteins were washed away, and equal volumes of buffer with or without Gsp1p-GTP were added. After 1 h at room temperature, beads were collected, washed briefly, and bound Kap95p was quantified by liquid scintillation. The amount of Kap95p bound to Nups was normalized to the amount bound after the addition of buffer only.
Equilibrium Affinity Assay—Affinities were calculated using purified radiolabeled Kap95p in solution binding assays. Kap95p was phosphorylated at an engineered site in the N terminus using bovine heart kinase and [32P]ATP (PerkinElmer Life Sciences). Unincorporated nucleotide was removed in a G-25 desalting spin column (Bio-Rad). For each experiment, a GST-Nup was immobilized on beads and incubated with increasing concentrations of radiolabeled Kap95p (with or without a molar excess of unlabeled Kap60p and cargo). The cargo used was the NLS of CBP80 (aa 1-30) fused to the maltose-binding protein. Incubations were for 2 h at room temperature in binding buffer with protease inhibitors (2 µg/ml leupeptin, 1 µg/ml pepstatin, 2 µg/ml aprotinin, and 20 µg/ml PMSF) and 1 mg/ml bovine serum albumin. Beads were then washed quickly, and bound radiolabeled Kap95p was quantified by liquid scintillation. Binding curves were fit to the data using GraphPad PrismTM software (Biosoft). Each curve was fit using duplicate points of 6-8 different Kap95p concentrations. Experiments were repeated for each pair using different concentrations of Nup within the bead volume; these gave similar KD values. Nup42p affinities were measured in the presence of 5 µM Kap60p and 5 µM cargo; Nup100pC affinities were measured in the presence of 1 µM Kap60p and 1 µM cargo; mcNup1p affinities were measured in the presence of 25 nM Kap60p and 50 nM cargo; and mcNup1p36 affinities were measured in the presence of 2.5 µM Kap60p and 2.5 µM cargo.
Direct GFP Fluorescence Microscopy—Yeast expressing GFP fusions were grown to mid-log phase in SCD-URA culture media and visualized live under a Nikon Eclipse E600 microscope. Images were captured and assembled using QED and Photoshop 7.0 software. Three different counters examined over 100 cells for each strain. Only fluorescent cells were counted. Cells were scored as "nuclear" if there was more fluorescence in the nucleus than in the cytoplasm. Because of higher expression levels in the nup136 strain than in the wild-type strain, pNLS-GFP pictures were not taken at the same exposure times.
In Vivo cNLS-GFP Import Assay—In vivo nuclear import assays were performed as described by Shulga et al. (14). Briefly, strains carrying the cNLS-YFP-expressing plasmid (encoding the SV40 NLS fused to YFP) were grown at 30 °C to an A600 = 0.5. After sedimentation, cells were washed and resuspended in 1 ml of culture medium containing 20 mM sodium azide and 20 mM 2-deoxyglucose. After 1 h at 4 °C, yeast were sedimented, washed, resuspended in 100 µl of medium containing 2% dextrose, and incubated at 4 °C. Nuclear import was initiated by transferring 5 µl of the cell suspension onto a microscope slide at room temperature. Localization of the cNLS-YFP fusion protein was determined by fluorescence microscopy. Only cells showing fluorescence were scored. Approximately 30 cells were counted in 1-min intervals, and results were pooled from four separate experiments. The relative rates of recovery were calculated by drawing linear regression lines through the initial portion of the time course (0-5 min) using GraphPad PrismTM software (Biosoft).
Preparation of S. cerevisiae Cell Extracts—For total cell extracts, 4 x 107 cells were harvested, washed with 1 ml of TE buffer, and resuspended in 50 µl of TE buffer. 50 µl of 3x Laemmli sample buffer was added, and samples were heated for 10 min at 95 °C. After vortexing with glass beads for 3 min, the liquid was collected and centrifuged at 15,000 rpm for 5 min. 15 µl of the supernatant was used for Western blot analysis. Cytosol extracts were prepared using French press cell lysis as described previously (1).
Indirect Immunofluorescence Microscopy—Yeast were grown to early log phase at 30 °C in YPD media and were fixed with 3.7% formaldehyde for 10 min. For Kap95p and Kap60p localizations, cells were collected by centrifugation, resuspended in phosphate buffer (100 mM KH2PO4, pH 6.5, 500 µM MgCl2) with 3.7% formaldehyde, and fixed further for 1 h at 30 °C. For Nup1p localizations, this 1-h fix was omitted. All cells were washed twice in phosphate buffer and resuspended in the same buffer containing 1.2 M sorbitol. Cell walls were digested with Zymolase 100T (ICN) for 20 min at 30 °C, and yeast spheroplasts were allowed to adhere to poly-L-lysine-treated glass slides. Spheroplasts were post-fixed for 6 min in dry ice-cold methanol followed by 30 s in acetone. Antibody incubations were done as described previously (15), using anti-Kap95p and anti-Kap60p antibodies at a 1:2000 dilution, and anti-Nup1p antibodies at a 1:5000 dilution.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
Equilibrium Binding Affinities of Kap95p to FG Nups—To further characterize the observed difference in binding between Kap95p complexes and FG Nups, their relative binding affinities were quantified (Fig. 1B). Nup42p, Nup100p, and Nup1p were chosen to represent the three regions of the NPC (see above). For Nup100p and Nup1p, full-length proteins were not used because of poor solubility. Instead, the regions containing all of the FG repeats were used and are referred to as mcNup1p (aa 332-1076) and Nup100pC (aa 1-640). mcNup1p, Nup100C, and Nup42p are functional as they efficiently capture karyopherins from yeast extracts (1). As Kap95p is likely in a complex with Kap60p and its cargo during translocation across the NPC in vivo, we also measured the affinity of Kap95p to FG Nups in the presence of Kap60p and cargo. In all cases, the concentrations of unlabeled Kap60p and cargo added were well above the KD of Kap95p-Kap60p complexes (KD 
0.15 nM) and Kap60p-cargo complexes (KD = 2.8 nM) (10) and were in molar excess of the highest concentration of radiolabeled Kap95p used in each experiment. Therefore, all Kap95p molecules should be bound to Kap60p or Kap60p-cargo complexes at the concentrations used.
We find that Kap95p-Kap60p-cargo complexes bind to bead-immobilized Nup42p with a KD = 1.7 ± 0.3 µM, to Nup100p with a KD = 90 ± 9 nM, and to Nup1p with a KD = 0.4 ± 0.02 nM (Fig. 1B and Table II). Each value measured remains constant at multiple Nup concentrations (data not shown). Given the number of FG repeats in each Nup (22 for Nup42p, 44 for Nup100pC, and 28 for Nup1pN), and assuming that Kap95p can bind to any FG repeat, it seems possible that multiple Kap95p molecules bind simultaneously to a Nup. However, the maximal stoichiometry of Kap95p-Nup complexes in the solution binding assay always appears to be 1:1 as judged by Coomassie Blue staining (data not shown). It is also interesting that the strength of Kap95p-Nup interactions measured does not correlate with the number of FG repeats present in each Nup; instead, the strength correlates to the location of each Nup within the NPC, with the weakest binding to Nup42p of the cytoplasmic filaments, intermediate binding to Nup100p of the central scaffold region, and the strongest binding to Nup1p of the nuclear basket (Table II). The difference in affinity is pronounced: a 4,000-fold increase in affinity for Kap95p import complexes from the cytoplasmic side of the NPC to the nucleoplasmic side.
|
Effect of Kap60p and Kap60p-Cargo Complexes on the Affinity of Kap95p to FG Nups—We also measured the affinity of Kap95p monomers and Kap95p-Kap60p heterodimers to the same three Nups, to examine how Kap60p and cargo modulate binding of Kap95p to Nups. We find that Kap60p increases 1.5-fold the affinity of Kap95p to Nup42p, but does not change the affinity of Kap95p to Nup100p (Table II). In stark contrast, Kap60p increases 
160-fold the affinity of Kap95p to Nup1p (10). The presence of cargo decreases 2-fold the affinity of Kap95p-Kap60p to Nup42p, increases 3-fold the affinity to Nup100pC, and lowers 8-fold the affinity to Nup1p (Table II) (10). Thus, Nup1p is the most sensitive to the presence of Kap60p or Kap60p-cargo complexes when binding Kap95p. Despite the effects of Kap60p and cargo, the relative binding trend between Kap95p complexes and the FG Nups remains the same in all situations: weak binding to Nup42p, moderate binding to Nup100p, and strongest binding to Nup1p (Table II).
Effect of Gsp1p-GTP on the Binding of Kap95p Complexes to Nup1p—In general, the binding of Kap95p to FG Nups is disrupted by the presence of Gsp1p-GTP (1, 9), which is most abundant in the nucleoplasm. Among the three FG Nups tested here, Nup1p is the only one that resides exclusively in the nucleoplasmic side of the NPC. Thus, we assume that binding of Kap95p import complexes to Nup1p in vivo could only happen if it occurs prior to the import complex contacting a free Gsp1p-GTP molecule, or in spite of the presence of Gsp1p-GTP. Two observations support the latter statement. First, Kap95p-Kap60p can be recovered bound to Nup1p in yeast extracts that have been supplemented with 1.2 µM Gsp1pQ71L-GTP under conditions that diminish binding of Kap95p-Kap60p to most other FG Nups (1). Second, the C terminus of Nup1p (aa 963-1076) has the unusual property of enhancing the Kap60p-mediated disassembly of Kap95p-Gsp1p-GTP complexes (19). Here we explored this subject further by comparing the relative sensitivity of the various Kap95p-Nup1p complexes to Gsp1p-GTP.
GST-Nup1p was immobilized on glutathione beads and incubated with 0.2 µM Kap95p-Kap60p-cargo complexes. The amount of radiolabeled Kap95p bound was then measured in the absence and presence of 0.2 or 1 µM Gsp1p-GTP. Binding of Kap95p import complexes to Nup1p was hardly affected by the presence of up to 1 µM Gsp1p-GTP (Fig. 2A), which approximates the estimated concentration of Gsp1p-GTP in the nucleoplasm. However, a 36-aa deletion of the C terminus of Nup1p (Nup1p36) caused Kap95p binding to Nup1p36 to be more responsive to Gsp1p-GTP, rendering nearly half of the Kap95p unable to bind Nup1p36 when the concentrations of Gsp1p-GTP and Kap95p were equivalent (Fig. 2A). To examine this issue further, we repeated the binding of Kap95p to mcNup1p using higher ratios of Gsp1p-GTP to Kap95p with or without Kap60p and cargo (Fig. 2B). We found that a 2-3-fold molar excess of Gsp1p-GTP was sufficient to prevent over 90% of the binding of Kap95p monomers to Nup1p, but a 60-fold molar excess of Gsp1p-GTP over Kap95p was required for the same effect in the presence of Kap60p or Kap60p-cargo complexes. Thus, the presence of Kap60p reduced the effects of Gsp1p-GTP on Kap95p-Nup1p interaction, consistent with the previous observation that Kap60p accelerates the dissociation of Kap95p-Gsp1p-GTP complexes (19). As expected, Gsp1p-GDP had no effect on the interaction between Kap95p or Kap95p complexes with Nup1p (data not shown). Given the very rapid rate of association and dissociation of Kap95p import complexes with Nup1p (kon 8.3 x 107; koff 3.3 x 10-2) (10), the effect of Gsp1p-GTP in accelerating the dissociation of Kap95p-Kap60p-cargo complexes (10), and the converse effect of Kap60p in accelerating the dissociation of Kap95p-Gsp1p-GTP complexes (19), we estimate that most complexes in the solution binding assays above (except the Kap60p-cargo complex) associated and dissociated more than 100 times during the 1-h equilibrium binding reaction.
|
Nup1p Can Specifically Capture Kap95p-Kap60p Complexes from Nup100p—As Kap95p import complexes move across the NPC, they likely encounter Nup100p before encountering Nup1p, given the location of these Nups at the NPC (Fig. 8) (4). We therefore tested whether the more distal Nup1p could "capture" or out-compete Kap95p from Nup100p (Fig. 3), because Kap95p has a much higher affinity to Nup1p than to Nup100p (Table II). Kap95p and Kap60p were first incubated with immobilized GST-Nup100pC, and then soluble mcNup1p or Nup100pC was added to compete for karyopherin binding. When no competitor was added, Kap95p-Kap60p complexes were recovered bound to GST-Nup100pC (Fig. 3A, lane 4). When soluble Nup100pC was added as competitor, Kap95p-Kap60p was again recovered bound to the bead-immobilized Nup100pC, and, surprisingly, a fraction of the soluble Nup100pC also bound to the immobilized complex (lane 5). This result suggests that Kap95p has two binding sites for Nup100p, because neither Nup100pC nor Kap60p bind GST-Nup100pC on their own (Fig. 3A, lane 2; Fig. 1A). Finally, when mcNup1p was added as a soluble competitor, most of Kap95p-Kap60p was recovered in the unbound fraction (lane 6) in a complex with mcNup1p (data not shown). Similar results were obtained when these competition reactions were conducted in the presence of cargo (data not shown). We conclude that Nup1p can out-compete Nup100p for binding Kap95p import complexes as predicted from the differences in measured affinities.
|
|
Nup42p and Nup100p bind equally well to six and eight karyopherins, respectively, in yeast extracts (1), yet Nup1p and Nup2p bind preferentially to one karyopherin, the Kap95p-Kap60p heterodimer. Consistently, we find that Kap95p also binds to Nup2p with high affinity in the presence of Kap60p (KD 70 pM) (Table II). Such high affinity and selectivity implies that Nup1p and Nup2p function as exclusive high affinity docking sites for Kap95p-Kap60p complexes at the nuclear basket; hence, these Nups should only capture Kap95p-Kap60p from other FG Nups. To test this notion, we conducted a competition assay using Nup100p-Kap complexes formed in yeast extracts (Fig. 3B). As expected, GST-Nup100pC coated beads captured many karyopherins from yeast extracts (lane 1) (1). These beads were then incubated with soluble Nup2p, mcNup1p, or Nup100pC as competitors. Strikingly, Kap95p and Kap60p were the only Kaps out-competed from Nup100pC by Nup2p and mcNup1p, as judged by their recovery in the unbound fraction (lane 3). In contrast, addition of soluble Nup100pC did not compete effectively for Kaps bound to the immobilized Nup100pC (lane 4), as expected because of a lower concentration of soluble Nup100C than immobilized GST-Nup100C. These results demonstrate that Nup1p and Nup2p effectively and specifically out-competed Nup100p for binding Kap95p-Kap60p.
The C Terminus of Nup1p Contains a High Affinity Binding Site for Kap95p Import Complexes—The in vivo relevance of a high affinity binding site for Kap95p import complexes at the nuclear basket was examined using a mutation in Nup1p that reduces its affinity for Kap95p. First, we mapped the domain of Nup1p responsible for this high affinity interaction by examining the binding of purified Kap95p, Kap60p, and cargo to three Nup1p fragments: nNup1p (aa 1-332), mNup1p (aa 332-963), and cNup1 (aa 963-1076) (Fig. 4). We found that Kap95p monomers bound to cNup1p better than to nNup1p or mNup1p (Fig. 4B, lane 3), and that Kap95p-Kap60p heterodimers bound to all three Nup1p fragments (lane 4). In stark contrast, Kap95p-Kap60p-cargo complexes bound only to cNup1p (Fig. 4B, lane 5).
To more carefully map the high affinity binding site for Kap95p import complexes in the C terminus of Nup1p, we mutated cNup1p as shown in Fig. 4A. First, all four FG repeats were eliminated by mutating phenylalanine residues to aspartic acid or isoleucine (cNup1p-noFG: F971I, F1008D, F1027D, and F1038D). There were two surprising results. First, binding of Kap95p monomers was diminished but not completely abolished (Fig. 4B, lane 3). Second, Kap95p-Kap60p-cargo complexes bound to cNup1p-noFG as well as to wild-type cNup1p (lane 5). We made additional truncations from either end of cNup1p and examined binding to Kap95p and Kap60p as before. Most notably, deletion of the C-terminal 36 aa of cNup1p (cNup1p36) abolished binding of Kap95p-Kap60p-cargo complexes to cNup1p (Fig. 4B). The same mutation in a larger Nup1p construct (mcNup1p36; aa 332-1040) also diminished binding of Kap95p-Kap60p-cargo complexes (compare Figs. 1 and 4B, bottom panels, lane 5). Quantitation of binding affinities revealed that the Kap95p-Kap60p-cargo complex bound mcNup1p36 with an affinity of 180 nM, which is 450-fold lower than the affinity of the same complex to "wild-type" mcNup1p (Table II). In addition, Kap95p monomers and Kap95p-Kap60p heterodimers bound to mcNup1p36 with affinities that were 438- and 224-fold lower than to mcNup1p, respectively (Table II). In essence, the 36-aa C-terminal truncation converted Nup1p from a high affinity binding site for Kap95p import complexes to a medium-affinity site.
In Vivo Disruption of the High Affinity Binding Site for Kap95p Import Complexes in Nup1p—We engineered a 108-bp deletion (corresponding to a 36-aa C-terminal truncation of Nup1p) in the chromosomal copy of NUP1 in an otherwise wild-type yeast strain. Although Nup1p is an essential nucleoporin in our strains (and in most other yeast strains) (20), the expression of Nup1p36 is tolerated. nup136 yeast grow at rates similar to wild-type at 25, 30, and 37 °C (Fig. 7, and data not shown), indicating that, overall, nuclear import was not severely impaired. A small percentage of nup136 yeast did, however, have a large cell size and an abnormal cell shape (data not shown). Importantly, the stability and subcellular localization of Nup1p36 (aa 1-1040) mutant were similar to those for wild-type Nup1p (Fig. 5, A and B).
|
|
To test for subtle defects in Kap95p localization and Kap95-mediated nuclear import, we first examined the localization of Kap95p and Kap60p in the nup136 strain via immunofluorescence microscopy (Fig. 5C). In a wild-type strain, 85% of cells display bright Kap95p fluorescence in the nucleus or at the nuclear rim, and 33% of cells have a little more Kap60p in the nucleus than in the cytoplasm (15, 17, 21, 22) (Fig. 5, C and D). In nup136 cells, only 33% showed bright nuclear staining of Kap95p, and only 9% of cells showed the slight nuclear accumulation of Kap60p (Fig. 5, C and D). The quantities of Kap95p and Kap60p were equivalent in both strains (data not shown). The reduced accumulation of Kap95p and Kap60p in the nuclei of nup136 yeast may reflect a reduction in Kap95p-Kap60p import efficiency, or an increase in the export efficiency of these Kaps. To distinguish between these possibilities, we examined the intracellular localization of a cNLS-bearing cargo.
Nuclear Accumulation of a cNLS Cargo Is Less Efficient in the nup136 Strain—We expressed a cNLS-YFP fusion (the NLS of SV40 T-antigen fused to YFP) in wild-type and nup136 strains and examined its steady-state distribution (Fig. 6, A and B). We also examined cells expressing rgNLS-GFP and pNLS-GFP reporter constructs, which are cargoes for Kap104p and Kap121p, respectively. The localization of cNLS-YFP is mostly nuclear in 76% of wild-type cells, whereas only 33% of nup136 cells show a similar nuclear accumulation (Fig. 6B). In contrast, the cellular localization of rgNLS-GFP and pNLS-GFP was not affected in the nup136 strain (Fig. 6, A and B).
|
Interestingly, we noted that the brightness of the NLS reporters was variable among nup136 cells. An examination of NLS reporter levels using Western blot analysis with anti-GFP antibodies shows that the expression of cNLS-YFP, rgNLS-GFP, and pNLS-GFP was elevated by 5-10-fold in the nup136 strain compared with wild-type (data not shown). As these reporter cargoes are expressed from a centromeric plasmid and selective medium was used for cell growth, the variability is probably not related to changes in plasmid copy number. It is therefore unclear why plasmid expression was affected by the nup136 truncation. However, the results clearly show that nup136 cells were specifically deficient in nuclear import of a Kap95p-dependent cargo (cNLS-YFP), but not Kap104p- or Kap121p-dependent cargoes (Fig. 6).
To test whether the decrease in nuclear accumulation of cNLS-YFP in nup136 yeast is the result of a reduced rate of nuclear import, we utilized an in vivo nuclear import assay (14) (Fig. 6C). During metabolic poisoning of yeast with sodium azide, nucleocytoplasmic transport stops and the cNLS-YFP reporter protein equilibrates between the cytoplasm and nucleoplasm. Upon removal of the poison, nuclear import kinetics can be quantified as transport resumes and cNLS-YFP accumulates in the nucleoplasm. As there are no tethering sites for this artificial cargo in the yeast nucleoplasm, the rate of accumulation reflects the nuclear import rate minus the rate of diffusion out of the nucleus (14). We find that wild-type yeast accumulated nuclear fluorescence at an initial rate of 14.2 ± 2.0%/min, whereas nup136 yeast accumulated nuclear fluorescence at a slower initial rate of 3.1 ± 1.1%/min (Fig. 6C). These results are consistent with the notion that Nup1p36 slows down the nuclear accumulation of cNLS-bearing cargoes in vivo because of its reduced binding affinity for Kap95p import complexes. It is also possible (although unlikely) that the 36-aa deletion in Nup1p causes the NPC to become more permeable to small proteins (such as cNLS-YFP), thereby increasing their diffusion rate out of the nucleus.
Overexpression of a Kap95p-Kap60p-dependent Cargo Slows Growth of Yeast with Mutations in Nuclear Basket FG Nups—During the course of our experiments, we observed that expression of a different cNLS reporter (cNLS-GFP) slows growth of nup136 yeast but not wild-type yeast (Fig. 7A), even though both strains grew at the same rate without overexpression of this cargo (data not shown). This cargo contains the same SV40 T-antigen NLS as the cNLS-YFP cargo, but is expressed at a higher level and is larger (45 versus 26 kDa) because of a Gal4p activation domain present in the construct. The growth defect is specific for a Kap95p-Kap60p-dependent cargo, as expression of a Kap104p-dependent cargo (rgNLS-GFP, 35 kDa) or a Kap121p-dependent cargo (pNLS-GFP, 60 kDa) did not slow growth of nup136 yeast (Fig. 7A). These results indicate that nuclear import defects in the nup136 strain are likely restricted to the Kap95p-Kap60p transport pathway.
We also examined the effect of cNLS-GFP overexpression in strains lacking Nup2p (nup2) or Nup60p (nup60) (Fig. 7B). Like Nup1p, these FG Nups are major components of the nuclear basket (4, 18) and bind mostly to Kap95p-Kap60p in yeast extracts (1). In fact, Nup60p tethers Nup2p to the nuclear basket, and in a nup60 strain, Nup2p is mostly localized to the nucleoplasm rather than at the NPC (15, 23). We find that nup2 and nup60 yeast grew as well as wild-type yeast, but overexpression of the cNLS-GFP cargo slowed their growth (Fig. 7B). In comparison, overexpression of the cNLS-GFP in nup100 and nup42 mutant yeast had no effect on cell growth (Fig. 7B). Thus, mutations in nuclear basket Nups make yeast particularly sensitive to overexpression of a Kap95p-Kap60p-dependent cargo, whereas mutations of other Nups do not.
Double Mutants in Nuclear Basket FG Nups Slow Down Cell Growth—To examine the genetic relationship between Nup1p and other FG Nups, we introduced the nup136 truncation in several nucleoporin knockout strains (Fig. 7C). We find that nup136 nup2 double mutant yeast grow slowly at room temperature (Fig. 7C) and cannot grow at 37 °C (data not shown). Likewise, nup136 nup60 double mutant yeast display a similar, but less severe slow growth phenotype (Fig. 7C). In contrast, nup136 nup42 or nup136 nup100 double mutant yeast are not impaired for growth (Fig. 7C). The synthetic lethality between the nup136 and nup2 or nup60 mutations emphasizes the importance of an intact nuclear basket for normal cell growth, and is consistent with the notion that nuclear basket Nups function to increase the efficiency of the essential Kap95p-mediated nuclear import pathway.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
36 yeast (Figs. 5 and 6). It has been suggested that all karyopherins move through the NPC via low affinity interactions with Nups, and that high affinity interactions would slow down or prevent transport because of long interaction times between Kaps and Nups (24). However, our results clearly show that Kap95p import complexes bind to a nuclear basket FG Nup with very high affinity (Fig. 1; Table II), and that this binding is important for efficient nuclear import of Kap95p complexes in yeast (Figs. 5 and 6). The notion that only low affinity sites are involved in transport is based on the extrapolation of binding affinities from rates of in vitro nuclear import reactions fitted to a Michaelis-Menten equation (24). We argue that such an extrapolation may be inadequate, as it assumes that translocation across the NPC resembles a classical enzymatic reaction, which is likely not the case.
An Affinity Gradient for Kap95p Import Complexes along the Yeast NPC—Our results support the notion that FG Nups provide a gradient of affinity for Kap95p along the yeast NPC (Fig. 8). This affinity gradient is best described as having three steps: weak binding of Kap95p import complexes to cytoplasmic filament FG Nups (e.g. Nup42p), moderate binding to centrally located FG Nups (e.g. Nup100p), and strong binding to FG Nups in the nuclear basket (e.g. Nup1p). Although only a small subset of FG Nups was examined here, we believe that these results reflect the overall binding preferences of Kap95p for FG Nups in general; this view is based on the cumulative results from a number of reconstitution experiments that tested most of the FG Nups under similar conditions (1, 25, 26) (data not shown). Consistent with our results, it was recently shown that vertebrate importin also displays binding preferences for FG Nups along the NPC, with the highest affinity toward a nuclear basket FG Nup (Nup153), medium affinity to centrally located FG Nups (Nup62, Nup58, Nup54), and lowest affinity to a cytoplasmic filament FG Nup (Nup358) (5). Together, these results reinforce the notion that an affinity gradient exists for Kap95p/importin along the eukaryotic NPC.
An affinity gradient may be unique to the Kap95p-mediated import pathway in S. cerevisiae. We observed previously that nuclear basket Nups do not bind tightly to other importins, except for Kap123p, which binds well to Nup60p (1). We therefore speculate that most yeast importins do not "sense" an affinity gradient that involves nuclear basket Nups as high affinity docking sites. This raises an interesting possibility that the yeast nuclear basket is tailored specifically for the Kap95p-Kap60p transport pathway.
Directional Movement of Kap95p down an Affinity Gradient?—Our results support the hypotheses that high affinity binding sites for Kaps in asymmetrically located Nups play key roles in transport (4, 27). High affinity binding of Kap95p to nuclear basket FG Nups may bias the movement of Kap95p import complexes from the central channel of the NPC toward the nuclear basket. If Kap95p-Kap60p heterodimers could bind two FG Nups simultaneously, it would then be possible for Kap95p to "probe" affinity differences between adjacent FG Nups while bound to both Nups simultaneously. Given that binding affinities are dictated by molecular association and dissociation rates (KD = koff/kon), and assuming similar association rates for Kap95p-FG Nup interactions, then differences in Kap-Nup affinities would strictly reflect differences in the mean interaction time of a Kap95p-Nup complex (i.e. a higher affinity would reflect a longer interaction time). If Kap95p-Kap60p were bound simultaneously to two Nups of equal affinity, then both Nups would remain bound for roughly equal amounts of time and would make a stable Nup-Kap-Nup complex (as seen in Fig. 3A, lane 5). However, if a Kap95p-Kap60p heterodimer binds simultaneously to Nups of different affinity, it will likely end up bound to the higher affinity nucleoporin by dissociating more rapidly from the low affinity Nup. This "hand-off" mechanism could produce a net directional movement of Kap95p complexes from low affinity sites to high affinity sites within the NPC. Given the highly flexible nature of FG repeat regions of Nups (e.g. they lack secondary structure and are composed mostly of random coil) (28), collisions between FG Nups anchored at different positions in the NPC may occur with high enough frequency to support the rapid translocation of Kap95p complexes from one part of the NPC to another via consecutive hand-offs between FG Nups. Ultimately, the highest affinity binding site would need to be emptied to allow uninterrupted flux of Kap95p-Kap60p-cargo complexes into the nucleus. This is accomplished by the rapid dissociation of Kap95p import complexes from Nup1p (koff 3.3 x 10-2; t
21 s), which occurs in a Gsp1p-GTP independent manner (10). Once a Kap95p import complex dissociates from Nup1p, Gsp1p-GTP would bind Kap95p (Fig. 8), causing the accelerated release of the Kap60p-cargo complex (10). Then, Cse1p-Gsp1p-GTP or Nup2p would bind to Kap60p, causing the accelerated release of the NLS cargo as the final step of nuclear import (10).
Effect of Kap60p and Cargo on Kap95p Binding to FG Nups—We found that Kap60p monomers and Kap60p-cargo complexes modulate Kap95p binding to FG Nups (Figs. 1, 2, and 4), with the most drastic effects seen with Nup1p (Table II). As Nup1p (but not Nup42p or Nup100p) binds Kap60p directly (Fig. 1), the large increase in Kap95p affinity to Nup1p in the presence of Kap60p may be the result of both Kaps binding simultaneously to Nup1p. In that regard, it is interesting that the isolated region of Kap60p that binds Kap95p (the Kap60p IBB domain) does not stimulate binding of Kap95p to the middle region of Nup1p (data not shown), whereas full-length Kap60p does (Fig. 4B). Similarly, the IBB domain of mammalian importin 
does not increase the affinity of importin to FG Nups (5). These results support the notion that full-length Kap60p contacts Nup1p directly as part of a Kap95p-Kap60p heterodimer, rather than binding exclusively through Kap95p. At present, it is unclear how cargo modulates the binding of Kap95p-Kap60p to the FG Nups, but we suggest it does so by inducing a conformational change in Kap60p that modulates its direct binding to Kap95p and/or the FG Nups.
The High Affinity Binding Site for Kap95-Import Complexes in the C Terminus of Nup1p—We mapped a high affinity binding site for Kap95p-Kap60p-cargo complexes on Nup1p to the C-terminal 113 aa, in a region that contains only four FG repeats (Fig. 4A). It was surprising that Kap95p binding to this region is not limited to FG repeats; their deletion diminished, but did not abolish, Kap95p binding (Fig. 4, A and B). This finding disagrees with the common notion that karyopherins interact exclusively with the phenylalanine residue of FG repeats in Nups (2, 3). Instead, a 36-aa truncation in the C terminus of Nup1p (which does not include FG repeats) lowers its affinity for Kap95p and the Kap95 complexes by >200-fold each (Table II) and abolishes binding of Kap60p monomers (Figs. 1 and 4, compare bottom panels). There is a cluster of positive charges in the C terminus of Nup1p that resembles a cNLS and the IBB domain of karyopherin . However, that domain of Nup1p does not compete with cNLSs for binding Kap60p and binds Kap60p-cargo complexes efficiently (data not shown). We suggest that monomeric Kap95p and Kap60p bind to this sequence in a manner that mimics their binding to the IBB domain of Kap60p (29). In the case of Kap95p-Kap60p heterodimers, we suggest that Kap60p occupies the cluster of positive charges in cNup1p, whereas Kap95p binds simultaneously to the FG repeats of cNup1p and to the IBB domain of Kap60p. This divalent Kap-Nup interaction can explain the high affinity of Kap95p-Kap60p binding to Nup1p (KD 0.05 nM) and to Nup2p (KD 0.07 nM) (Table II), which also binds Kap60p directly (10, 17).
We suggest that Nup1p and Nup2p share the function of serving as high affinity binding sites for Kap95p-Kap60p heterodimers in the nuclear basket of the NPC. Importantly, loss of the high affinity sites provided by these Nups cannot be tolerated by yeast, as demonstrated by the synthetic lethality of the nup136 nup2 mutant strain (Fig. 7B). This is consistent with previous genetic data that demonstrated links between NUP1 and NUP2 (30), and with the fact that Nup1p and Nup2p capture only Kap95p-Kap60p heterodimers from yeast extracts (1).
The Nup1p36 Truncation—The Nup1p36 truncation mimics the previously described nup1-21 mutant, which lacks the C-terminal 35 aa and displays defects in cell growth and nuclear morphology (31). However, the phenotype of our nup136 strain is different, as nup136 cells do not show growth or nuclear morphology defects. This difference could be the result of (i) the 1-aa difference in the C-terminal deletion, (ii) to distinct genetic backgrounds, and/or (iii) altered amounts of Nup1p in the nup1-21 strain (where Nup1p was expressed from a plasmid). Consistently, we noticed a more pronounced morphology defect in a nup136 strain when Nup1p36 was expressed at low levels from a plasmid (data not shown). In our strain, the C-terminal truncation does not alter Nup1p stability (Fig. 5A) and should not alter Nup1p structure or NPC structure, as the middle and C-terminal regions of Nup1p are natively unfolded and are largely devoid of secondary structure (28).
The 36-aa truncation in the C terminus of Nup1p could affect other Nup1p functions. For example, the C terminus of Nup1p stimulates dissociation of Kap95p-Gsp1p-GTP heterodimers in vitro in a Kap60p-dependent manner (19). This explains why the equilibrium binding of Kap95p import complexes to Nup1p appears unaffected by Gsp1p-GTP at high concentrations (Fig. 2), when in fact Gsp1p-GTP is dissociating Kap95p-Kap60p complexes in a futile cycle against the reverse reaction driven by Kap60p and cNup1p (19). Regardless, any reduction of Kap95p-Gsp1p-GTP dissociation in the nucleoplasm as a consequence of the nup136 mutation would favor Kap95p nuclear export, which should not affect import of cNLS-containing cargoes. Instead, we found that import of a Kap95p-Kap60p-dependent cargo is hindered in nup136 mutant yeast (Fig. 6).
Nup1p also functions as a mild KaRF (karyopherin release factor) in that it slightly destabilizes the interaction of NLS cargo with Kap95p-Kap60p heterodimers by increasing 16-fold their dissociation rate (10). However, this mild KaRF activity of Nup1p resides in its middle region (aa 333-962) and not in the C terminus (data not shown), suggesting that Nup1p KaRF activity is not compromised in the nup136 strain. This is consistent with the inability of Kap95p-Kap60p to dock a cargo to the middle region of Nup1p (Fig. 4, second panel, lane 5), even though it does so efficiently at the C terminus of Nup1p (third panel, lane 5).
Finally, Prp20p binds to the C terminus of Nup1p in vitro (19), and an inability of Prp20p to dock and function properly at the NPC could disrupt many karyopherin pathways. However, in the nup136 strain, we only found defects in the Kap95p-mediated pathway, and not in the Kap104p or Kap121p pathways (Fig. 6).
We conclude that in vivo disruption of the high affinity binding site for Kap95p in the C terminus of Nup1p reduces the efficiency of the Kap95p-dependent nuclear import pathway directly, rather than indirectly through a secondary effect. This conclusion emphasizes the importance of high affinity binding sites for Kap95p import complexes at the NPC in promoting efficient nuclear import.
|
FOOTNOTES |
|---|
To whom correspondence should be addressed. Tel.: 650-725-4814; Fax: 650-723-0155; E-mail: rexach{at}stanford.edu.
1 The abbreviations used are: cargo, a protein containing a nuclear transport signal; FG, phenylalanine-glycine; FG Nup, nucleoporin containing phenylalanine-glycine repeats; IBB, importin binding; Kap, karyopherin; NLS, nuclear localization signal; NPC, nuclear pore complex; Nup, nucleoporin; aa, amino acid(s); GST, glutathione S-transferase; GFP, green fluorescent protein; YFP, yellow fluorescent protein; cNLS, classic nuclear localization signal.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
