Journal of Cell Science 113, 1471-1480 (2000) 1471 Printed in Great Britain © The Company of Biologists Limited 2000 JCS1243

Nup2p is located on the nuclear side of the complex and coordinates Srp1p/importin-α export

Jennifer K. Hood, Jason M. Casolari and Pamela A. Silver* Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School and The Dana-Farber Cancer Institute, 44 Binney Street, Boston, MA 02115, USA *Author for correspondence (e-mail: [email protected])

Accepted 10 February; published on WWW 21 March 2000

SUMMARY

Proteins bearing canonical nuclear localization sequences displaced from the nuclear rim to the nuclear interior are imported into the nucleus by the importin/karyopherin- in ∆nup2 cells. We do not observe any biochemical α/β heterodimer. Recycling of the importin-α subunit to the interaction between Cse1p and Nup2p. Instead, we find cytoplasm requires the action of Cas, a member of the that Nup2p binds directly to Srp1p. We have localized importin-β superfamily. In the yeast Saccharomyces Nup2p to the interior face of the nuclear pore complex, and ceresivisiae, the essential CSE1 encodes a Cas have shown that its N terminus is sufficient for targeting homologue that exports the yeast importin-α Nup2p to the pore, as well as for binding to Srp1p. Taken Srp1p/Kap60p from the nucleus. In this report, we describe together, these data suggest that Nup2p is an important a role for the FXFG Nup2p, and possibly the NPC docking site in the Srp1p export pathway. related Nup1p, in the Cse1p-mediated nuclear export pathway. Yeast cells lacking Nup2p or containing a particular temperature-sensitive mutation in NUP1 Key words: Nup2p, Srp1p, Cse1p, Nucleoporin, Yeast, Nuclear accumulate Srp1p in the nucleus. Similarly, Cse1p is export

INTRODUCTION rounds of protein import, the import receptor must be recycled to the cytoplasm. Importin-β can exit the nucleus on its own move into and out of the nucleus through nuclear pore (Izaurralde et al., 1997; Kose et al., 1999), but Srp1p/importin- complexes (NPCs) embedded in the nuclear membrane. Their α must be exported by the exportin Cse1p/CAS (Hood and transport requires soluble receptor proteins of the Silver, 1998; Künzler and Hurt, 1998; Kutay et al., 1997; importin/exportin superfamily. These receptors shuttle Solsbacher et al., 1998). Cse1p/CAS binds to Srp1p/importin- between the nucleus and the cytoplasm, making contacts with α with Ran-GTP, forming a cooperative export complex that is NPC components, or , as they traverse the pore dissociated by Ran-GTP hydrolysis once it reaches the (reviewed by Pemberton et al., 1998; Adam, 1999). The cytoplasm (Kutay et al., 1997; Solsbacher et al., 1998). association of transport receptors with their cargoes and with The actual mechanism of nucleocytoplasmic translocation is the NPC is regulated by the nucleotide-bound state of the Ran not well understood, although a fairly detailed picture of the GTPase in the nuclear and cytoplasmic compartments of the structure of the NPC and the position of many of its cell (Marelli et al., 1998; Nakielny et al., 1999; Seedorf et al., components within that structure is available, especially for the 1999; reviewed by Moore, 1998; Adam, 1999). As a result, the yeast Saccharomyces cerevisiae (reviewed by Stoffler et al., Ran cycle dictates the direction of transport. 1999). Many nucleoporins contain repeats of the short amino Proteins that contain classical nuclear localization signals acid sequences FG, FXFG or GLFG. Numerous interactions (NLSs) of the SV40 or bipartite type are imported into the between nucleoporin repeats and importins/exportins have nucleus by importin-β, also called Kap95p in yeast (Görlich et been reported (Albertini et al., 1998; Fornerod et al., 1997; al., 1995a; Imamoto et al., 1995; Moroianu et al., 1995). This Hellmuth et al., 1998; Iovine et al., 1995; Radu et al., 1995; transport pathway is unique in that the association of importin- Shah et al., 1998; Titov and Blobel, 1999), but it is unclear β with NLS substrates requires an adapter protein, importin-α, which of these are significant for nuclear transport in vivo. also called Srp1p in yeast (Enenkel et al., 1995; Görlich et al., The yeast proteins Nup1p, Nup2p and Nsp1p belong to the 1995b). The trimeric import complex assembles in the class of nucleoporins that contain amino acid repeats of the cytoplasm, crosses the NPC, and then dissociates in the nucleus FXFG variety. NUP1 and NSP1 are essential in most strain due to the binding of Ran-GTP to Kap95p/importin-β (Görlich backgrounds, while NUP2 is nonessential in an otherwise wild- et al., 1996b; Rexach and Blobel, 1995). To allow further type background (Loeb et al., 1993). Genetic interactions 1472 J. K. Hood, J. M. Casolari and P. A. Silver between these three suggest that they have some University, Waltham, MA, USA). These were raised against a peptide functional overlap: the 175 amino acids at the N terminus of consisting of repeats of the amino acid sequence FSFG; they Nup2p are essential for viability in strains that express mutant recognize the three FXFG-containing yeast nucleoporins Nup1p, versions of Nup1p or Nsp1p that lack the amino termini of Nup2p and Nsp1p. Mouse monoclonal anti-GST antibodies were from these proteins (Loeb et al., 1993). In addition to this N-terminal Santa Cruz Biotechnology. domain, the Nup2p protein possesses a central domain that Immunofluorescence contains the FXFG repeats and a C-terminal Ran-binding Cells were grown to mid-log phase in 10 ml liquid yeast extract- domain similar to that found in the yeast Ran-binding proteins peptone-dextrose (YEPD) cultures at 25°C, then fixed in 3.7% Yrb1p and Yrb2p (Dingwall et al., 1995). Nup2p is of special formaldehyde for 30 minutes and processed for immunofluorescence interest because it is the only yeast nucleoporin that contains as previously described (Hood and Silver, 1998). Rabbit polyclonal such a Ran-binding domain. anti-Srp1p antibodies were used at 1:1000 dilution overnight. Both Nup1p and Nup2p have previously been shown to FITC-conjugated anti-rabbit secondary antibodies (Jackson interact with Srp1p in coimmunoprecipitation experiments ImmunoResearch) were used at 1:1000 dilution with a 1 hour (Belanger et al., 1994). In this work we have investigated the incubation. role of Nup2p in the Srp1p export pathway and have found Visualization of GFP/YFP-tagged proteins in live cells evidence to support the model that Nup2p serves as an initial Cells were grown to mid-log phase in liquid synthetic complete docking site for the formation of the Cse1p/Ran-GTP/Srp1p medium or in ‘dropout’ medium to select for the retention of plasmids. export complex. Fluorescent signal was observed using a Nikon fluorescence microsope and filters specific for GFP (Chroma Technology Corp.) or EYFP (Omega). Images were captured by a Princeton Instruments MATERIALS AND METHODS Micromax digital camera using Metamorph imaging software (Universal Imaging Corp.). Yeast strains and plasmids The ∆nup2-4::URA3 (MAT α, ura3-5, leu2-3, 112, trp1-289), ∆nup2- Expression and purification of GST-tagged proteins from 5::HIS3 (MAT α, ura3-52, leu2-3, 112, his3∆200, ade2, trp1∆63), E. coli nup1-8 (pCEN LEU2 nup1-8 in MAT α, his3∆200, trp1-1, ura3-5, pGEX expression plasmids were transformed into BL-21 E. coli. 5- leu2-3, 112, nup1-2::LEU2) and nup1-21 (pCEN LEU2 nup1-21 in ml LB cultures containing 100 µg/ml ampicillin (LB/AMP) were MAT α, his3∆200, trp1-1, ura3-5, leu2-3, 112, nup1-2::LEU2) strains grown overnight at 37¡C, then diluted into 1 L LB/AMP and allowed + were previously described (Loeb et al., 1993). An ADE version of to grow at 37¡C to an OD600 of 0.4. Isopropyl-1-thio-β- ∆nup2-5::HIS3 (PSY1946; MAT a, ura3, leu2, his3, trp1) was D-galactopyranoside (IPTG) was added to 0.1 mM to induce generated by crossing ∆nup2-5 to PSY1052 (MAT a, ura3, leu2, trp1, expression for 3 hours at 37¡C. Cultures were centrifuged and his3). The CSE1-GFP:TRP1 and XPO1-GFP:TRP1 strains have been pellets frozen overnight at −20¡C or in liquid nitrogen. Cell previously described (Hood and Silver, 1998). ∆nup2-4::URA3 CSE1- pellets were resuspended in ice-cold PBS containing 0.5 mM GFP:TRP1 and ∆nup2-4::URA3 XPO1-GFP:TRP1 strains were phenylmethylsulfonyl fluoride (PMSF) and 3 µg/ml each pepstatin A, created by transforming BstEII-linearized pPS1610 (pRS304 CSE1- leupeptin, aprotinin and chymostatin. Cells were lysed by sonication GFP) or XcmI-linearized pRS304 XPO1-GFP, respectively, into (3× 30 seconds with 1 minute on ice between sonications), Triton X- ∆nup2-4::URA3. The NUP2-EYFP:URA3 (PSY1835) strain was 100 was added to 1%, and lysates were clarified by centrifugation. constructed by M. Damelin (Damelin and Silver, 2000). Lysates were incubated for 30 minutes at room temperature with 1 ml The GST-Nup2p E. coli expression plasmid (pPS1944) was created 50% glutathione-Sepharose slurry (Amersham Pharmacia Biotech) by cloning the NUP2 ORF into the BamHI/XhoI sites of pGEX 4T-1 prewashed with PBS. Beads were washed 3× with 10 ml PBS/1% (Amersham Pharmacia Biotech). Sequence encoding amino acids 1- Triton X-100, then bound GST fusion proteins were eluted with 5× 556 of Nup2p was cloned into the BamHI/XhoI sites of pGEX 4T-1 500 µl of 10 mM reduced glutathione/50 mM Tris, pH 8. to generate the GST-Nup2∆RBD E. coli expression plasmid (pPS1945). The GAL-Nup2p-GFP plasmid (pPS1946) and the cleavage of GST-Srp1p truncated versions (Nup2∆C-GFP, pPS1947; Nup2∆N-GFP, Purified GST-Srp1p was dialyzed overnight in thrombin cleavage pPS1948) were constructed by PCR-amplifying the corresponding buffer (150 mM NaCl, 25 mM Tris, pH 8, 2.5 mM CaCl2, 0.1% β- regions of the NUP2 ORF with a BamHI site at the N terminus and a mercaptoethanol, 10% glycerol), then again for 4 hours in fresh buffer. XhoI site at the C terminus (plus an ATG start codon for the ∆N Thrombin (Sigma) was mixed with dialyzed GST-Srp1p (0.014 units construct), then cloning them into the BamHI/XhoI sites in the thrombin/µg GST-Srp1p) and incubated for 2 hours at room polylinker of pPS1303 (2µ URA3 GAL1-GFP). Versions of these temperature. Cleaved products were bound 2× to 250 µl glutathione- plasmids that express the fusion proteins under the control of the Sepharose (30 minutes at 4¡C) to remove GST. NUP2 promoter (Nup2p-GFP, pPS1951; Nup2∆C-GFP, pPS1952; Nup2∆N-GFP, pPS1953) were created by PCR-amplifying 1 kb Binding from yeast lysates upstream of the NUP2 ORF with EcoRI/BamHI ends and cloning into GST or GST-Nup2p beads were prepared by mixing 20 µl 50% pRS316 (CEN URA3), then cloning the GFP fusion fragments from glutathione-Sepharose slurry with 250 µl PBS/1% Triton X-100 pPS1946, pPS1947 and pPS1948 into the BamHI site after the NUP2 containing 0.05 µM GST or GST-Nup2p and incubating for 30 promoter. The GST-Srp1p expression plasmid was obtained from G. minutes at 4¡C. Cse1-GFP yeast lysate was prepared from 50 ml Schlenstedt (Universität des Saarlandes, Homburg, Germany). mid-log phase YEPD culture by glass bead lysis in PBSMT buffer (2 mM MgCl2, 1 mM EDTA, 0.5% Triton X-100 in PBS) plus Antibodies protease inhibitors (0.5 mM PMSF and 3 µg/ml each of pepstatin Rabbit anti-Srp1p antibodies were obtained from D. Görlich (Zentrum A, leupeptin, aprotinin and chymostatin) using a FastPrep bead für Molekulare Biologie der Universität Heidelberg, Germany). beater (Savant) as previously described (Hood and Silver, 1998). Antibodies to Cse1p (Hood and Silver, 1998), Kap95p (Koepp et al., GST/GST-Nup2p beads were centrifuged and the supernatant was 1996), and GFP (Seedorf et al., 1999) have been previously described. removed; beads were then incubated on a rocker for 1 hour at 4¡C Rabbit anti-FXFG antibodies were obtained from L. Davis (Brandeis with 250 µl 1.5 mg/ml Cse1-GFP lysate without or with 1 mM Nup2p functions in Srp1p nuclear export 1473

GMPPNP (Sigma). Unbound fractions were saved for gel analysis; seconds in 0.2 M Reynolds lead citrate. Samples were examined on a beads were washed 5× with 500 µl PBSMT and once with 500 µl Jeol transmission electron microscope. PBSM (no Triton), then resuspended in 50 µl PBSM plus 12.5 µl of 5× sodium dodecyl sulfate (SDS)-sample buffer (10% SDS, 25% glycerol, 250 mM Tris, pH 6.8, 500 mM DTT, 0.5% Bromophenol RESULTS Blue). ∆ In vitro binding Srp1p is trapped in the nucleus in nup2 and nup1- 21 cells 20 µl 50% glutathione-Sepharose slurry were mixed with 0.05 µM GST, GST-Nup2p or GST-Nup2∆RBD plus 0.01 µM recombinant Srp1p was previously shown to interact with both Nup1p and Srp1p in PBSMT buffer to a total volume of 250 µl. Binding mixtures Nup2p in coimmunoprecipitation experiments (Belanger et al., were incubated for 1 hour at 4¡C on a rocker. Unbound fractions were 1994). We hypothesized that Nup1p and Nup2p might be saved for gel analysis; beads were washed 5× with 500 µl PBSMT important pore contacts in the Cse1p-mediated Srp1p export and once with 500 µl PBSM (no Triton), then resuspended in 50 µl pathway. Accordingly, the localization of Srp1p in cells deleted PBSM plus 12.5 µl of 5× SDS-sample buffer. for NUP2 and in two nup1 mutant strains was determined by Anti-GFP immunoprecipitations indirect immunofluorescence using an anti-Srp1p antibody (Fig. 1A). In a wild-type background, the Srp1p signal was GAL-Nup2-GFP and truncation mutant plasmids (pPS1946, 1947, distributed throughout the cell, with some cells showing a 1948) plus the GAL-GFP vector (pPS1303) were transformed into ∆ ∆nup2 (PSY1946) cells. Cells were grown at 30¡C in ura−, 2% glucose slight concentration in the nucleus. Virtually all nup2 cells medium, then diluted 1:200 into 50 ml ura−, 2% raffinose and allowed showed a strong accumulation of Srp1p in the nucleus, similar to grow to mid-log phase (OD600=0.4). GFP protein expression was to that observed in cold-sensitive cse1-1 cells (Hood and Silver, induced with 2% galactose for 1 hour. Lysates were prepared as 1998; Solsbacher et al., 1998), suggesting the presence of a described above for lysate binding to GST-Nup2p. 20 µl 50% protein defect in export of Srp1p from the nucleus. This defect has also G-Sepharose slurry (Amersham Pharmacia Biotech) plus 2.5 µg anti- been recently reported by Booth et al. (1999). Overexpression GFP antibodies were added to 500 µl (1 mg/ml) yeast lysate for of Cse1p from a 2µ plasmid did not restore wild-type Srp1p immunoprecipitations (IPs). IPs were incubated for 1 hour at 4¡C with ∆ × localization in nup2 cells, but expression of a C-terminal rocking. Beads were pelleted in a microcentrifuge, washed 5 with Nup2p truncation lacking the Ran-binding domain (see Fig. 500 µl PBSMT and once with 500 µl PBSM, then resuspended in 50 µl PBSM plus 12.5 µl 5× SDS sample buffer. 3B) from a centromeric plasmid did partially rescue the Srp1p export defect (data not shown). Immunoblotting Srp1p also accumulated in the nuclei of some nup1-21 cells Protein samples were resolved in 10% SDS-polyacrylamide gels (approximately 50%), but this phenotype was less penetrant (Laemmli, 1970), then transferred to nitrocellulose membranes using than in ∆nup2 cells. nup1-8 cells showed no nuclear standard techniques (Ausubel et al., 1998). All antibodies were diluted accumulation of Srp1p (Fig. 1), nor did cells containing a in 5% powdered milk/PBST (PBS, 0.25% Tween 20). Primary temperature-sensitive mutation in NSP1 (data not shown). The antibody incubation was for 1 hour at room temperature or overnight nup1-21 allele expresses a truncated Nup1p protein that lacks at 4¡C; secondary antibody incubation was for 30 minutes at room 34 amino acids at its C terminus, while the nup1-8 allele temperature. Antibody dilutions were as follows: anti-GFP, 1:5000; expresses an N-terminally truncated protein lacking amino anti-Srp1p, 1:5000; anti-Kap95p, 1:1000; anti-FXFG, 1:500; anti- Cse1p, 1:5000; anti-GST, 1:1000; HRP-conjugated anti-rabbit- and acids 4-141. anti-mouse-IgG secondary antibodies (Jackson Immunoresearch), ∆ 1:5000. Immunoreactive bands were visualized using enhanced Nup1p levels are elevated in nup2 cells chemiluminescence (ECL kit, Amersham Pharmacia Biotech). Where Deletion of NUP170 alters the overall levels of other applicable, band intensities were quantified by densitometry with a nucleoporins, specifically decreasing the amounts of Nup1p Personal Densitometer SI and analyzed using Mac IQ ImageQuant and Nup2p in cells (Kenna et al., 1996). Nup170p binds to software (Molecular Dynamics). Nup1p and has been proposed to anchor Nup1p and Nup2p within the NPC. Since Nup1p and Nup2p may share contacts Electron microscopy within the NPC and exhibit partial functional overlap, we All chemicals and other materials were from Electron Microscopy hypothesized that deletion of NUP2 might allow increased Sciences unless otherwise noted. Cells were grown at 30¡C in YEPD incorporation of Nup1p into the NPC, thus protecting the to OD600=1, then fixed and spheroplasted as previously described (Fahrenkrog et al., 1998). For anti-GFP labeling, cells were protein from degradation and increasing the overall level of resuspended in 100 µl anti-GFP 1:10 in 0.1 M potassium phosphate Nup1p in cells. We performed immunoblotting on whole cell buffer, pH 6.5 (KPi)/0.1% BSA and incubated overnight at 30¡C with lysates from wild-type and ∆nup2 cells using an antibody shaking. After primary labeling, cells were washed 2× in 0.1% specific for the FXFG repeats of Nup1p, Nsp1p and Nup2p. BSA/KPi, then incubated with shaking for 4 hours at 30¡C in 100 µl Densitometry analysis indicated an approximately threefold 5-nm gold-conjugated protein A (Dr G. Posthuma, Utrecht, increase in the amount of Nup1p in ∆nup2 cells compared to Netherlands; 1:70 in KPi/0.1% BSA). Cells were washed 2× with wild type (Fig. 1B). Equal sample loading was achieved by KPi/0.1% BSA, then prefixed for 1 hour at room temperature with harvesting equal numbers of cells, measured by OD600 KPi/2% glutaraldehyde. After glutaraldehyde treatment, cells were absorbance, and was verified by Ponceau S staining of the blots washed 2× with KPi, then cells were embedded in low-melting agarose (FMC BioProducts). Samples were postfixed, dehydrated and and immunoblotting with anti-Srp1p antibodies. embedded in Epon resin as previously described (Fahrenkrog et al., ∆ 1998). Ultrathin sections (60-80 nm) were cut on a Reichert Cse1-GFP is mislocalized in nup2 cells microtome and placed on copper grids. Grids were stained for 1 Cse1p localizes primarily to the , with some minute in 50% acetone/50% saturated uranyl acetate and for 15 background in the nucleus and cytoplasm (Hood and Silver, 1474 J. K. Hood, J. M. Casolari and P. A. Silver

Fig. 1. (A) Srp1p localization in ∆nup2, nup1-8 and nup1- 21 cells. Polyclonal anti-Srp1p antibodies were used to localize endogenous Srp1p in wild type, ∆nup2-5, nup1-8 and nup1-21 cells grown at 25¡C. Nuclei are visualized by 4′,6-diamidino-2-phenyl-inadole (DAPI) staining, and Nomarski images show phase views of the cells. (B) Anti- FXFG immunoblot of whole cell extracts from wild-type and ∆nup2 cells. The FXFG antibody recognizes Nup1p, Nsp1p and Nup2p (marked by arrows).

envelope (Fig. 2B). An integrated GFP fusion of another exportin, Xpo1p, showed no difference in localization in wild-type versus ∆nup2 cells (Fig. 2B). GFP fusions of the importins Kap95p/importin β, Pse1p and Sxm1p were also not mislocalized in ∆nup2 cells (data not shown). Furthermore, Cse1p- GFP was not mislocalized in nup1-8 or nup1-21 cells. These results indicate that the absence of Nup2p specifically disrupts the normal distribution of Cse1p 1998). Disruption of the Cse1p-mediated export pathway by in the cell, suggesting that Nup2p plays an important role in the cse1-1 mutation or by a mutation in the guanine nucleotide the Cse1p-mediated export pathway. exchange factor for yeast Ran, prp20-1, causes Cse1p to accumulate in the (Hood and Silver, 1998). To The N terminus of Nup2p binds Srp1p directly determine the effect of the ∆nup2 deletion on Cse1p Since the lack of Nup2p altered the subcellular localization of localization, a C-terminal fusion of Cse1p to the green Cse1p, we looked for a biochemical interaction between the fluorescent protein (GFP) was integrated into the genome of two proteins. Recombinant Nup2p was purified from E. coli as wild type and ∆nup2 strains such that the fusion protein was an N-terminal glutathione S-transferase (GST) fusion. GST- expressed from the endogenous CSE1 promoter as the only Nup2p immobilized on glutathione-Sepharose beads was form of Cse1p in these cells. As shown by immunoblotting incubated with lysate from Cse1-GFP cells. The binding with anti-Cse1p and anti-GFP antibodies, Cse1-GFP was experiment was done in the absence or presence of the non- expressed at the same level in both wild-type and ∆nup2 cells hydrolyzable GTP analog GMPPNP, since some transport (Fig. 2A). In the ∆nup2 cells, Cse1-GFP was localized to the receptor/nucleoporin interactions have been shown to depend nucleoplasm with no visible concentration at the nuclear on the nucleotide-bound state of Ran (Marelli et al., 1998;

Fig. 2. Cse1p-GFP and Xpo1p-GFP localization in ∆nup2 cells. CSE1-GFP and XPO1-GFP were integrated into wild type and ∆nup2-4 cells, replacing the endogenous CSE1 and XPO1 genes. (A) Top, anti-GFP immunoblot of wild type (lane 1), CSE1-GFP (lane 2) and ∆nup2-4 CSE1-GFP (2) lysates. Bottom, anti-Cse1p immunoblot of the same samples. The anti-Cse1p antibody recognizes two cross-reacting bands above the endogenous Cse1p band. Equal amounts of total protein were loaded in each lane. (B) GFP fluorescence in living cells. Nup2p functions in Srp1p nuclear export 1475

Fig. 3. Nup2p binding interactions. (A) GST-Nup2p (lanes 1,2), GST (lanes 3,4), or no recombinant protein (lanes 5,6) was bound to glutathione-Sepharose beads and incubated with yeast lysate from Cse1p-GFP cells in the absence (lanes 1,3,5) or presence (lanes 2,4,6) of 1 mM GMPPNP. Bound and unbound fractions were analyzed by immunoblotting with anti-Srp1p, anti-GFP and anti-Kap95p antibodies. (B) In vitro binding of recombinant Srp1p to GST-Nup2p (lane 2) or GST-Nup2∆RBD (lane 3). Lane 1, control with GST alone. Only bound fractions are shown. Top, anti-GST immunoblot; bottom, anti-Srp1p immunoblot. (C) Nup2p domains. Numbers refer to amino acids. N, N-terminal domain; FXFG, repeat domain; RBD, Ran- binding domain. (D) Anti-GFP immunoprecipitations from ∆nup2-5 cells expressing GFP (lane 1), Nup2p-GFP (lane 2), Nup2∆C-GFP (lane 3) or Nup2∆N-GFP (lane 4). Top, anti- Srp1p immunoblot; bottom, anti-GFP immunoblot. Left, equal total protein amounts of whole lysates; right, equal volumes of immunoprecipitated proteins. IgGh, immunoglobulin heavy chain.

Seedorf et al., 1999). The bound and unbound fractions were remained a formal possibility that the interaction was mediated analyzed by SDS-PAGE and immunoblotting with antibodies by a third protein. Therefore, recombinant Srp1p was purified to GFP, Srp1p and Kap95p. Srp1p associated very strongly from E. coli as an N-terminal GST fusion protein. The GST with GST-Nup2p under both conditions, but no association of moiety was then cleaved by thrombin protease treatment. The Cse1-GFP was detected with or without GMPPNP (Fig. 3A, thrombin treatment yielded approximately equal amounts of lanes 1 and 2). Kap95p bound weakly to GST-Nup2p in the two cleavage products that ran as a doublet on SDS-PAGE gels absence of GMPPNP (lane 1), but this binding was abolished (Fig. 3B), suggesting that a second site within Srp1p was in the presence of the nucleotide analog (lane 2). None of the cleaved by thrombin in addition to the cleavage site in the three proteins bound to GST or to the beads alone. This result linker between GST and Srp1p. A good consensus thrombin suggests that Nup2p may function as a terminal site in cleavage site does exist in the N terminus of the protein after Kap95p/Srp1p-mediated NLS protein import. The trimeric amino acid 134. Cleavage at this position would remove the import complex may interact with Nup2p via Srp1p, then be region of Srp1p that binds to Kap95p, referred to as the disrupted by the binding of Ran-GTP to Kap95p, leaving Srp1p importin-β binding (IBB) domain (Görlich et al., 1996a; Weis bound to Nup2p. et al., 1996) and would be predicted to cause an increase in gel The interaction between Nup2p and Srp1p was previously mobility consistent with that observed for the lower band of shown by coimmunoprecipitation experiments (Belanger et al., the Srp1p doublet. Both forms of recombinant Srp1p were 1994), and the efficiency of Srp1p copurification with GST- recovered in the bound fraction when incubated with GST- Nup2p suggested that this interaction was direct, but it Nup2p and glutathione-Sepharose beads (Fig. 3B, lane 2). 1476 J. K. Hood, J. M. Casolari and P. A. Silver

Srp1p also bound efficiently to recombinant Nup2p lacking the to the FXFG repeats of Nup2p (data not shown). Nup2p-EYFP C-terminal Ran-binding domain (lane 3), but did not bind to appears to be fully functional, in that it does not cause any GST alone (lane 1). Thus, Srp1p does bind directly to Nup2p Srp1p mislocalization, any growth defect in wild-type cells, or and this binding does not require the Nup2p Ran-binding any synthetic growth phenotype in nup1-8 or nup1-21 strain domain, nor is it likely to require the N terminus of Srp1p. backgrounds (data not shown). Nup2p-EYFP localizes to the Booth et al. have also recently shown that the Srp1p IBB nuclear envelope in a punctate pattern characteristic of NPCs domain is dispensible for binding to Nup2p (Booth et al., (Fig. 4A). There was no difference in Nup2p-EYFP 1999). localization when the fusion construct was integrated into The Nup2p domain structure is depicted schematically in nup1-8 or nup1-21 strains (data not shown), suggesting that Fig. 3C. To further narrow the Srp1p binding domain of Nup2p, Nup1p is not required for the pore localization of Nup2p. In full-length Nup2p, Nup2p lacking the N-terminal 175 amino addition, several experiments that addressed a possible acids (Nup2∆N), and Nup2p lacking both the FXFG and Ran biochemical interaction between Nup1p and Nup2p gave binding-domains (Nup2∆C) were expressed in yeast as C- negative results. Therefore, the genetic interaction between terminal GFP fusions under the control of the galactose these two nucleoporins is not due to their anchoring each other inducible (GAL) promoter. These fusion proteins were in the NPC. immunoprecipitated with antibodies to GFP and the bound To determine what part of Nup2p targets the protein to fractions were analyzed by immunoblotting with anti-Srp1p NPCs, Nup2p-GFP, Nup2∆C-GFP and Nup2∆N-GFP fusion antibodies (Fig. 3D). Srp1p co-immunoprecipitated equally proteins were expressed from centromeric plasmids under the well with the full-length and Nup2∆C fusion proteins (lanes 2 control of the NUP2 promoter in ∆nup2 cells. Both Nup2p- and 3), but not with Nup2∆N-GFP or with GFP alone (lanes 1 GFP and Nup2∆C-GFP showed nuclear rim localization that and 4). Therefore, the N terminus of Nup2p is both necessary was indistinguishable from the localization of Nup2p-EYFP and sufficient for Srp1p binding in cell lysates. expressed from the integrated construct (Fig. 4B, left panels). In stark contrast, however, Nup2∆N-GFP was diffusely The Nup2p N terminus is required for NPC distributed throughout cells and showed no visible localization concentration at the nuclear envelope. The NPC association of A C-terminal Nup2p fusion to the enhanced yellow fluorescent Nup2p-GFP and Nup2∆C-GFP was confirmed by expressing protein (EYFP) was integrated into the yeast genome such that the same constructs in a rat3-1 mutant strain. rat3-1 cells have it was expressed under the control of the Nup2p promoter, a temperature-sensitive defect in mRNA export and exhibit replacing the endogenous protein. Nup2p-EYFP expression clustering of NPCs to one or several regions of the nuclear was confirmed by immunoblotting with antibodies to GFP and envelope at permissive and non-permissive temperatures (Li et al., 1995). Nup2p-GFP localized to discrete, bright dots in rat3-1 cells, consistent with its association with clustered NPCs (Fig. 4B, right panels). Similar clustering was observed for Nup2∆C-GFP in rat3-1 cells, but Nup2∆N-GFP remained diffusely localized in this strain. These results indicate that the N terminus of Nup2p is necessary and sufficient for targeting the protein to NPCs. Nup2p-GFP and Nup2∆C-GFP were also expressed at high levels from the GAL promoter in ∆nup2 cells. The GFP signal filled the nuclei and was concentrated in very bright spots at the nuclear periphery, but remained absent from the cytoplasm (data not shown). This accumulation of Nup2p-GFP and Nup2∆C-GFP inside the nucleus at high levels of expression suggested that Nup2p is on the nuclear side of the NPC rather than on the cytoplasmic face.

Fig. 4. (A) Nup2p-EYFP fluorescence in wild-type cells. The fusion protein was expressed from an integrated construct under the control of the endogenous NUP2 promoter. (B) Localization of truncated Nup2p proteins. Nup2p-GFP, Nup2∆C-GFP and Nup2∆N-GFP were expressed from centromeric plasmids in ∆nup2-5 (left panels) or rat3-1 (right panels) cells at 25¡C and visualized by GFP fluorescence. Nup2p functions in Srp1p nuclear export 1477

Fig. 5. Localization of Nup2p- EYFP by immuno-electron microscopy. (A) Ultrathin cross section of a yeast nucleus from Nup2p-EYFP cells labeled with anti-GFP/protein A-5 nm gold prior to embedding in Epon resin. Arrows point to two nuclear pores. Lower panels show enlarged views of NPCs from the same cell in the upper panel. C, cytoplasm; N, nucleus. Bar, 200 nm. (B) Histogram of the perpendicular distance from the central plane of the NPC to the site of labeling for 68 gold particles in 12 cells.

Nup2p lies on the nuclear face of NPCs The direct binding of Nup2p to Srp1p and the defects in the Srp1p export pathway observed in ∆nup2 cells are consistent with Nup2p serving as a docking site for formation of the Cse1p/Ran-GTP/Srp1p export complex (see Fig. 6). Furthermore, Booth et al. (1999) have recently demonstrated that Cse1p and yeast Ran-GTP (Gsp1p- GTP) compete with Nup2p for Srp1p binding, which is also consistent with this model. However, these data could also support a model in which Nup2p is involved in the terminal step in Srp1p export, binding to Srp1p and releasing from the trimeric export complex. To enable a distinction to be made between these two models, we localized Nup2p-EYFP in the integrated strain by immuno-electron microscopy. Paraformaldehyde-fixed cells were subjected to pre-embedding labeling with anti-GFP antibodies and gold-conjugated secondary antibodies according to the protocol of Fahrenkrog et al. (1998) with some modifications (see Materials and Methods). The cells were then embedded in Epon resin and ultrathin-sectioned for electron microscopy. As shown in Fig. 5A, the gold particles were clustered at the nuclear face of NPCs. For 12 cells, 88 out consistent with Nup2p forming part of the nuclear basket, as of 99 (89%) gold particles were tightly NPC-associated. depicted schematically in Fig. 6. Therefore, we propose that Background labeling in cells lacking a GFP epitope was Nup2p does indeed serve as a docking site for the initial negligible (data not shown). Cells expressing an integrated formation of the Cse1p/Ran-GTP/Srp1p export complex. version of Nsp1p-EYFP were subjected to immunogold labeling as a control. Nsp1p has previously been localized to three sites in the NPC: at the cytoplasmic and nuclear DISCUSSION peripheries of the central gated channel, and at a distal site on the nuclear basket (Fahrenkrog et al., 1998). In our preparation, We have shown that Nup2p plays an important role in the Nsp1p-EYFP labeling was concentrated at the cytoplasmic Cse1p-mediated Srp1p nuclear export pathway, since ∆nup2 periphery of NPCs (data not shown), indicating that the fixation cells accumulate Srp1p inside the nucleus and also mislocalize protocol did not destroy epitopes on the cytoplasmic face of Cse1p to the nuclear interior. Many interactions between the NPC and validating the localization of Nup2-EYFP to the nucleoporins and nuclear transport receptors have already been nuclear side of the NPC. identified; however, we were unable to detect a biochemical The perpendicular distance from the central plane of the interaction between Cse1p and Nup2p. Instead, the Cse1p NPC to the site of labeling was measured for 68 gold particles export cargo, Srp1p, interacts quite strongly with Nup2p in in 12 cells. These distances are plotted on the histogram in Fig. coprecipitation experiments using yeast lysates and in direct 5B. The greatest number of gold particles was located binding experiments using recombinant proteins. The N approximately 35 nm from the NPC central plane. This is terminus of Nup2p is necessary and sufficient for this binding 1478 J. K. Hood, J. M. Casolari and P. A. Silver

possible roles for Nup2p in the Srp1p nuclear export pathway. This result implicates Nup2p in the formation of the Cse1p/Ran-GTP/Srp1p export complex inside the nucleus (as depicted in Fig. 6) rather than in the dissociation of this Cse1 complex in the cytoplasm. Srp1 Ran- The direct binding between Nup2p and Srp1p and the GTP importance of this interaction for export of Srp1p from the nucleus represents a unique role for a nucleoporin in a nuclear Cytoplasm transport pathway. That is, in the case of Nup2p and Srp1p, the 3 nucleoporin interacts with the transport cargo itself, and not only with the transport receptor. This interaction is facilitative rather than essential, similar to the role of Yrb2p in Xpo1p/Crm1p-mediated export of leucine rich NES-containing Nucleus 1 proteins (Taura et al., 1998; Noguchi et al., 1999). Yrb2p and Nup2p both contain a Ran-binding domain as well as FXFG Nup2 Ran- Kap95 Nup2 repeats. In the case of Yrb2p, both domains are required to GTP Cse1 Srp1 Srp1 rescue the NES protein export defect of ∆yrb2 cells (Taura et Ran- NLS GTP al., 1998). For Nup2p, the N-terminal domain that binds to 2 Srp1p can at least partially rescue the Srp1p mislocalization defect of ∆nup2 cells without the FXFG or Ran-binding Ran- GTP Kap95 domains (Booth et al., 1999 and our unpublished results). Perhaps Nup2p immobilizes Srp1p at the NPC in a NLS conformation that is competent for the formation of an export complex. Then, more transient interactions between Fig. 6. Model for the role of Nup2p in Srp1p export. (1) Nup2p is a Cse1p/Ran-GTP and the FXFG and Ran-binding domains of potential terminal site for the Kap95p/Srp1p/NLS import pathway. Nup2p may bring all three components of the export complex Kap95p is dissociated from Srp1p by the binding of Ran-GTP, while into close proximity, enabling their cooperative binding and Srp1p remains bound to Nup2p. (2) Nup2p serves as a scaffold for simultaneous release from Nup2p. the formation of the Cse1p/Ran-GTP/Srp1p export complex, holding In addition to binding to Srp1p, the Nup2p N terminus also Srp1p at the pore until it is freed by binding to Cse1p/Ran-GTP. (3) serves another important role in targeting Nup2p to the NPC. The formation of the trimeric export complex releases Srp1p from It is possible that the N terminus interacts with other Nup2p, and the complex travels through the pore into the cytoplasm. nucleoporins and/or with a nuclear import receptor. Almost nothing is known about the in vivo process of NPC assembly interaction. This region of the protein also targets Nup2p to the and what types of interactions may be important for bringing NPC through an unknown mechanism. NPC components to the right place in the pore. The Kap95p subunit of the NLS import receptor copurifies The potential role of Nup1p in Srp1p export is unclear. The with GST-Nup2p in the absence of the non-hydrolyzable GTP nup1-8 mutation, which produces an N-terminally truncated analog GMPPNP. In this situation, the Ran contained in the Nup1p protein that is delocalized from the NPC (Bogerd et al., yeast lysate would be expected to be in the GDP-bound form, 1994), did not cause mislocalization of Srp1p to the nucleus. due to the action of the GTPase activating protein Rna1p. In In fact, nup1-8 cells showed even less nuclear Srp1p signal contrast, when GMPPNP is added to the yeast lysate, Kap95p than wild-type cells by immunofluorescence. However, the no longer copurifies with GST-Nup2p, although Srp1p is still nup1-21 mutant strain, which expresses a truncated Nup1p that present in the bound fraction. Since Ran-GTP is known to lacks the C terminus, does show some accumulation of Srp1p dissociate Kap95p from Srp1p (Görlich et al., 1996b; Rexach in the nucleus. This differs from the result reported by Booth and Blobel, 1995), it is likely that the association of Kap95p et al. (1999), in which a strain deleted for NUP1 (in a with GST-Nup2p in the absence of GMPPNP occurs via Srp1p. background where this gene is non-essential) showed no such Thus, Nup2p may represent a site where the terminal step in defect. This discrepancy could suggest a dominant effect of the NLS protein import takes place, that is, where Gsp1p-GTP nup1-21 allele. The amino acids deleted in the proteins releases Kap95p from Srp1p, freeing the NLS substrate and expressed from both the nup1-8 and nup1-21 alleles have been leaving Srp1p bound to Nup2p (see Fig. 6, step 1). shown to bind Srp1p independently (Booth et al., 1999; Floer With Srp1p tethered to the NPC by Nup2p after a round of et al., 1997), whereas Nup1p and Nup2p cannot bind import, it may be poised to be picked up by Cse1p/Ran-GTP simultaneously to Srp1p (Belanger et al., 1994). This suggests for rapid return to the cytoplasm (Fig. 6, steps 2 and 3). We that the two nucleoporins compete for the same binding site on therefore hypothesize that the immobilization of Srp1p by Srp1p and raises the possibility that, in the nup1-21 mutant Nup2p facilitates the formation of the Cse1p/Ran-GTP/Srp1p strain, the Nup1p N-terminal Srp1p binding site has a dominant export complex. A further role for Nup2p in positioning the negative effect on Srp1p export. export complex in the vicinity of other nucleoporins important The nup1-21 allele causes temperature sensitivity as well as for the export pathway (such as Nup1p or Nsp1p, for example) a defect in import, while the nup1-8 allele does remains a possibility but has not yet been demonstrated. The not exhibit either of these phenotypes (Bogerd et al., 1994). localization of Nup2p to the nuclear side of the NPC is the key Both alleles, however, are synthetically lethal with the ∆nup2 piece of data that allowed us to discriminate between two deletion, suggesting a functional connection between these two Nup2p functions in Srp1p nuclear export 1479 nucleoporins (Loeb et al., 1993). The protein import defect of karyopherin heterodimer that targets import substrate to mammalian nuclear nup1-21 may be explained by slowed recycling of Srp1p to the pore complexes. J. Biol. Chem. 270, 16499-16502. cytoplasm in these cells. Like Nup2p, Nup1p has been Fahrenkrog, B., Hurt, E. C., Aebi, U. and Panté, N. (1998). Molecular architecture of the yeast nuclear pore complex: localization of Nsp1p localized to the nuclear side of the NPC (Rout et al., 2000). We subcomplexes. J. Cell Biol. 143, 577-588. have also shown that the amount of Nup1p is higher in ∆nup2 Floer, M., Blobel, G. and Rexach, M. (1997). Disassembly of RanGTP- cells than in cells that express Nup2p. This result may suggest karyopherin beta complex, an intermediate in nuclear protein import. J. Biol. that the two nucleoporins compete for similar contacts within Chem. 272, 19538-19546. Fornerod, M., Deursen, J. v., Baal, S. v., Reynolds, A., Davis, D., Murti, the NPC. In the absence of Nup2p, more Nup1p may be K. G., Fransen, J. and Grosveld, G. (1997). The human homologue of incorporated into the pore at sites normally occupied by Nup2p yeast CRM1 is in a dynamic subcomplex with CAN/Nup214 and a novel and may compensate for the normal function of Nup2p in the nuclear pore component Nup88. EMBO J. 16, 807-816. Srp1p export pathway, at least at a level sufficient for viability. Görlich, D., Henklein, P., Laskey, R. A. and Hartmann, E. (1996a). A 41 α β Srp1p/importin-α has the interesting property of being both amino acid motif in importin- confers binding to importin- and hence transit into the nucleus. EMBO J. 15, 1810-1817. a nuclear transport cargo and a transport receptor. Because of Görlich, D., Kostka, S., Kraft, R., Dingwall, C., Laskey, R. A., Hartmann, its importance in the classical NLS import pathway, rapid E. and Prehn, S. (1995a). Two different subunits of importin cooperate to recycling of Srp1p to the cytoplasm is essential to the cell. recognize nuclear localization signals and bind them to the nuclear envelope. Nup2p facilitates this recycling by tethering Srp1p to the NPC Curr. Biol. 5, 383-392. Görlich, D., Panté, N., Kutay, U., Aebi, U. and Bischoff, F. R. (1996b). on the nucleoplasmic side. It has not yet been determined Identification of different roles for RanGDP and RanGTP in nuclear protein whether a similar Srp1p anchor may exist on the cytoplasmic import. EMBO J. 15, 5584-5594. side of the NPC. If so, in an analogous manner to the events Görlich, D., Vogel, F., Mills, A. D., Hartmann, E. and Laskey, R. A. that are proposed to occur on the inner side of the NPC, Srp1p (1995b). Distinct functions for the two importin subunits in nuclear protein may remain tethered to the NPC after the dissociation of the import. Nature 377, 246-248. Hellmuth, K., Lau, D. M., Bischoff, F. R., Künzler, M., Hurt, E. and Simos, export complex. This would allow Srp1p to shuttle back and G. (1998). Yeast Los1p Has Properties of an Exportin-Like forth through the NPC channel, but never be released into the Nucleocytoplasmic Transport Factor for tRNA. Mol. Cell Biol. 18, 6374- cytoplasm or nucleoplasm unless the normal dynamics of 6386. import or export pathways are disrupted by mutations in any Hood, J. K. and Silver, P. A. (1998). Cse1p Is Required for Export of Srp1p/Importin-α from the Nucleus in Saccharomyces cerevisiae. J. Biol. of the components of these pathways. Such an arrangement Chem. 273, 35142-35146. would maximize the efficiency of nuclear import. Imamoto, N., Shimamoto, T., Takao, T., Tachibana, T., Kose, S., Matsubae, M., Sekimoto, T., Shimonshi, Y. and Yoneda, Y. (1995). In vivo evidence We are grateful to Ueli Aebi and Birthe Fahrenkrog at the for involvement of 58 kDa component of nuclear pore-targeting complex in University of Basel for the opportunity to learn the technique of nuclear protein import. EMBO J. 14, 3617-3626. immuno-electron microscopy. We also thank Maria Ericsson in the Iovine, M. K., Watkins, J. L. and Wente, S. R. (1995). The GLFG repetitive Harvard Medical School Department of Cell Biology for EM region of the nucleoporin Nup116p interacts with Kap95p, an essential yeast technical assistance. Thanks to Marc Damelin, Amy Hitchcock, and nuclear nmport factor. J. Cell Biol. 131, 1699-1713. Izaurralde, E., Kutay, U., Kobbe, C. v., Mattaj, I. W. and Görlich, D. Anne McBride for helpful comments on the manuscript. This work (1997). The asymmetric distribution of the constituents of the Ran system was supported by grants from the National Institutes of Health and is essential for transport into and out of the nucleus. EMBO J. 16, 6535- the Human Frontiers Science Program. J.K.H. is a recipient of a 6547. training grant from the National Cancer Institute to the Dana-Farber Kenna, M. A., Petranka, J. G., Reilly, J. L. and Davis, L. I. (1996). Yeast Cancer Institute. N1e3p/Nup170p is required for normal stoichiometry of FG nucleoporins within the nuclear pore complex. Mol. Cell. Biol. 16, 2025-2036. Koepp, D. M., Wong, D. H., Corbett, A. H. and Silver, P. A. (1996). Dynamic localization of the nuclear import receptor and its interactions with REFERENCES transport factors. J. Cell Biol. 133, 1163-1176. Kose, S., Imamoto, N., Tachibana, T., Yoshida, M. and Yoneda, Y. (1999). Adam, S. A. (1999). Transport pathways of macromolecules between the Beta-Subunit of Nuclear Pore-targeting Complex (Importin-beta) Can Be nucleus and the cytoplasm. Curr. Opin. Cell Biol. 11, 402-406. Exported from the Nucleus in a Ran-independent Manner. J. Biol. Chem. Albertini, M., Pemberton, L. F., Rosenblum, J. S. and Blobel, G. (1998). 274, 3946-3952. A Novel Nuclear Import Pathway for the Transcription Factor TFIIS. J. Cell Künzler, M. and Hurt, E. C. (1998). Cse1p functions as the nuclear export Biol. 143, 1447-1455. receptor for importin alpha in yeast. FEBS Lett. 433, 185-190. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Kutay, U., Bischoff, F., Kostka, S., Kraft, R. and Görlich, D. (1997). Export Smith, J. A. and Struhl, K. (1998). Current Protocols in Molecular Biology of importin-alpha from the nucleus is mediated by a specific nuclear (ed. V. B. Chanda). New York: John Wiley & Sons, Inc. transport factor. Cell 90, 1061-1071. Belanger, K. D., Kenna, M. A., Wei, S. and Davis, L. I. (1994). Genetic and Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly physical interactions between Srp1p and nuclear pore complex proteins of the head of bacteriophage T4. Nature 227, 680-685. Nup1p and Nup2p. J. Cell Biol. 126, 619-630. Li, O., Heath, C. V., Amberg, D. C., Dockendorff, T. C., Copeland, C. S., Bogerd, A. M., Hoffman, J. A., Amberg, D. C., Fink, G. R. and Davis, L. Snyder, M. and Cole, C. N. (1995). Mutation or deletion of the I. (1994). Nup1 mutants exhibit pleiotropic defects in nuclear pore complex Saccharomyces cerevisiae RAT3/NUP133 gene causes temperature- function. J. Cell Biol. 127, 319-332. dependent nuclear accumulation of poly(A)+ RNA and constitutive Booth, J. W., Belanger, K. D., Sannella, M. I. and Davis, L. I. (1999). The clustering of nuclear pore complexes. Mol. Biol. Cell. 6, 401-417. Yeast Nucleoporin Nup2p Is Involved in Nuclear Export of Importin Loeb, J. D. J., Davis, L. I. and Fink, G. R. (1993). NUP2, a novel yeast alpha/Srp1p. J. Biol. Chem. 274, 32360-32367. nucleoporin, has functional overlap with other proteins of the nuclear pore Damelin, M. and Silver, P. A. (2000). Mapping interactions between nuclear complex. Mol. Biol. Cell. 4, 209-222. transport factors in living cells reveals distinct pathways through the nuclear Marelli, M., Aitchison, J. D. and Wozniak, R. W. (1998). Specific pore complex. Mol. Cell. 5, 133-140. Binding of the Karyopherin Kap121p to a Subunit of the Nuclear Pore Dingwall, C., Kandels-Lewis, S. and Seraphin, B. (1995). A family of Ran Complex Containing Nup53p, Nup59p and Nup170p. J. Cell Biol. 143, binding proteins that includes nucleoporins. Proc. Natl. Acad. Sci. USA 92, 1813-1830. 7525-7529. Moore, M. S. (1998). Ran and Nuclear Transport. J. Biol. Chem. 273, 22857- Enenkel, C., Blobel, G. and Rexach, M. (1995). Identification of a yeast 22860. 1480 J. K. Hood, J. M. Casolari and P. A. Silver

Moroianu, J., Blobel, G. and Radu, A. (1995). Previously identified protein Seedorf, M., Damelin, M., Kahana, J., Taura, T. and Silver, P. A. (1999). of uncertain function is karyopherin α and together with karyopherin β Interactions between a Nuclear Transporter and a Subset of Nuclear Pore docks import substrate at nuclear pore complexes. Proc. Natl. Acad. Sci. Complex Proteins Depend on Ran GTPase. Mol. Cell Biol. 19, 1547- USA 92, 2008-2011. 1557. Nakielny, S., Shaikh, S., Burke, B. and Dreyfuss, G. (1999). Nup153 if an Shah, S., Tugendreich, S. and Forbes, D. (1998). Major binding sites for the M9-containing mobile nucleoporin with a novel Ran-binding domain. nuclear import receptor are the integral nucleoporin Nup153 and the EMBO J. 18, 1982-1995. adjacent nuclear filament protein Tpr. J. Cell Biol. 141, 31-49. Noguchi, E., Saitoh, Y., Sazer, S. and Nishimoto, T. (1999). Disruption of Solsbacher, J., Maurer, P., Bischoff, F. R. and Schlendstedt, G. (1998). the YRB2 gene retards nuclear protein export, causing a profound mitotic Cse1p Is Involved in Export of Yeast Importin alpha from the Nucleus. Mol. delay, and can be rescued by overexpression of XPO1/CRM1. J. Biochem. Cell Biol. 18, 6805-6815. (Tokyo) 125, 574-585. Stoffler, D., Fahrenkrog, B. and Aebi, U. (1999). The nuclear pore complex: Pemberton, L. F., Blobel, G. and Rosenblum, J. S. (1998). Transport routes from molecular architecture to functional dynamics. Curr. Opin. Cell Biol. through the nuclear pore complex. Curr. Opin. Cell Biol. 10, 392-299. 11, 391-401. Radu, A., Moore, M. S. and Blobel, G. (1995). The peptide repeat domain Taura, T., Krebber, H. and Silver, P. A. (1998). A member of the Ran-binding of nucleoporin Nup98 functions as a docking site in transport across the protein family, Yrb2p, is involved in nuclear protein export. Proc. Natl. nuclear pore complex. Cell 81, 215-222. Acad. Sci. USA 95, 7427-7432. Rexach, M. and Blobel, G. (1995). Protein import into nuclei: Association Titov, A. A. and Blobel, G. (1999). The Karyopherin Kap122p/Pdr6p Imports and dissociation reactions involving transport substrate, transport factors, Both Subunits of the Transcription Factor IIA into the Nucleus. J. Cell Biol. and nucleoporins. Cell 83, 683-692. 147, 235-245. Rout, M. P., Aitchison, J. D., Suprapto, A., Hjertaas, K., Zhao, Y. and Weis, K., Ryder, U. and Lamond, A. I. (1996). The conserved amino-terminal Chait, B. T. (2000). The yeast nuclear pore complex: composition, domain of hSRP1 alpha is essential for nuclear protein import. EMBO J. 15, architecture and transport mechanism. J. Cell Biol. 148, 635-651. 1818-1825.