Ran binding domains promote the interaction of Ran with p97/beta-karyopherin, linking the docking and translocation steps of nuclear import.

Nuclear protein import is accomplished by two sequential events: docking at the nuclear pore complex followed by ATP-dependent translocation across the nuclear envelope. Docking of nuclear targeted proteins requires a 56-kDa nuclear localization signal receptor (α-karyopherin, importin-α, SRP1α) and a 97-kDa protein (β-karyopherin, importin-β). Components necessary for translocation include the Ran/TC4 GTPase and NTF2/B-2. The functions of these factors at a molecular level remain unclear. We have now found that a complex of Ran, in the GTP-bound state, with either the Ran binding protein, RanBP1, or an isolated Ran binding domain binds with high affinity and specificity to β-karyopherin to form a ternary complex. We find that a C-terminal truncation mutant of Ran, Δ-DE Ran, also binds to β-karyopherin and that Δ-DE Ran can associate with a cytosolic, multiprotein complex that contains β-karyopherin and another Δ-DE Ran binding protein of 115/120 kDa. These data suggest a physical link between docking and translocation mediated by a Ran GTPase-Ran binding protein complex.

The mediated nuclear transport of proteins that bear nuclear localization signals (NLS) 1 proceeds through two distinct steps. Docking at the nuclear envelope is followed by energy-dependent translocation of the proteins through the nuclear pore complex (1)(2)(3). Recognition and docking of NLS substrates require a 9 S cytosolic complex that includes p56/␣-karyopherin and p97/␤-karyopherin (4,5). The ␣-karyopherin is the rat homolog of human SRP1␣, Xenopus importin-␣, and yeast SRP1 and functions as a receptor for NLS-containing proteins (5)(6)(7)(8). The 97-kDa protein, ␤-karyopherin, also called importin-␤, is required for docking of the NLS substrate, but the exact function of this protein is not known (5). It has been suggested that ␤-karyopherin acts as an adapter that links the ␣-karyopherin-NLS substrate complex with the nuclear pore complex through nucleoporin peptide repeats (5,9,10). The docking complex disassembles as a late event in translocation leaving ␤-karyopherin associated with the nuclear pore complex while ␣-karyopherin enters the nucleus with the NLS substrate (10).
ATP-dependent translocation of NLS-containing proteins through the nuclear pore complex is mediated by the Ran GTPase and a 10-kDa protein, NTF2/B-2 (11)(12)(13). While these components are known to be necessary for the import of NLScontaining proteins, their mechanism of action has not been elucidated. Ran is a 25-kDa Ras-related GTPase that is predominantly localized to the nucleus (14,15). Expression of a GTPase-deficient mutant of Ran, or depletion of the Ran GTPase-activating protein, Rna1p, blocks nuclear import both in vitro and in intact cells (16,18,19), 2 demonstrating that GTP hydrolysis is necessary for nuclear transport. After GTP hydrolysis, Ran-GDP may enter the nucleus where guanine nucleotide exchange can be stimulated by the chromatin-associated factor, RCC1 (19). Ran-GTP would then exit the nucleus to complete the cycle (for review see Ref. 20).
Two potential downstream effectors that bind to Ran-GTP have been identified, RanBP1/HTF9A, a 29-kDa cytosolic protein, and RanBP2/NUP358, a 360-kDa nuclear pore protein (21)(22)(23). These proteins associate with Ran through one or more conserved Ran binding domains (RanBDs). These domains stabilize Ran in the GTP-bound state and co-activate the GTPase-activating protein, Rna1p (24 -26). Additionally, antibodies raised against RanBP2 inhibit nuclear translocation, which suggests that RanBP2 may be a necessary constituent in the nuclear import mechanism (22). No requirement for RanBP1 in nuclear protein import has yet been established.
Removal of the acidic C terminus of Ran reduces binding to RanBPs but enhances binding to two unidentified proteins of 90 and 115/120 kDa (24). In this report we identify the 90-kDa protein as ␤-karyopherin. Direct interaction of ␤-karyopherin to wild-type Ran is promoted by association of Ran with RanBP1 or with an isolated Ran binding domain. These results suggest that ␤-karyopherin links the docking and translocation steps of nuclear import by forming a ternary complex with the Ran GTPase and a RanBP.

Production of Recombinant Proteins-Recombinant
Ran was expressed in Escherichia coli DH5␣ bacteria from a pET11a expression vector (Novagen) and purified by DEAE column chromatography as described previously (24). The ⌬-DE Ran mutant was created by the removal of the C-terminal sequence of Ran, PDEDDDL (24). GST fusion proteins of ⌬-DE Ran and RanBP1 were expressed from pGEX2T (Pharmacia Biotech Inc.) and purified by glutathione-Sepharose chromatography. Truncations of RanBP1 were made using polymerase chain reaction mutagenesis, and the products were subcloned into pGEX2T. The residues expressed by the RanBP1 truncations are RanBP1 ⌬-C (residues 1-160), RanBD (residues 27-160), and RanBP1 ⌬-N (residues 27-203). The integrity of the constructs was confirmed by DNA se-* This work was supported by National Institutes of Health Grant GM 50526 from the National Cancer Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. quencing. ␤-Karyopherin (a gift from Dr. A. Radu, Rockefeller Institute) was expressed in E. coli BL21(DE3) from the His 6 -tag vector pET-21b(ϩ) (Novagen). The protein was purified using a nickel-nitrilotriacetic acid column (Qiagen) as described in Ref. 5. Note that the recombinant protein lacks 12 amino acids at the N terminus but has an attached His 6 and a T7 tag that add 18 amino acids. This construct appears slightly larger than the 90-kDa protein in our gels, with an apparent size of 93 kDa.
Ran Overlay Assay-Ran overlay assays were performed as described in Ref. 24 with the following modifications. To form complexes of [␣-32 P]GTP-Ran with RanBPs, 2 g of Ran was added together with 4 g of GST-RanBP1 or -RanBP1 fragment, or with 8 g of ␤-karyopherin and incubated on ice with 10 Ci of [␣-32 P]GTP (specific activity, 3000 Ci/mmol) in 25 mM MOPS, pH 7.1, 1 mM EDTA. In some experiments unfused GST was included as a negative control. After 20 min, MgCl 2 was added (to 20 mM), and the incubation was continued for 15 min (4°C). Free [␣-32 P]GTP was removed by gel filtration through a G-25 Centrisep column (Princeton Separations, Adelphia, NJ) equilibrated with 25 mM MOPS, pH 7.1, 20 mM magnesium acetate. The amount of [␣-32 P]GTP complexed to Ran was quantitated by binding to nitrocellulose filters. Rat brain cytosol proteins were prepared as described previously (3). BHK21 cell proteins were prepared by addition of hot SDS sample buffer to 100-mm plates of confluent BHK21 cells that had been rinsed twice with phosphate-buffered saline. Cell lysate was passed through a 26-gauge syringe needle, and insoluble material was removed by centrifugation at 12,000 ϫ g for 5 min. Protein samples (150 g) were separated by 8% SDS-PAGE and transferred to nitrocellulose for analysis by the Ran overlay assay, as described previously (24). Equal amounts of Ran-bound [␣-32 P]GTP (approximately 300,000 cpm) were diluted into 10 ml of binding buffer for the overlay. After 30 min of incubation, overlays were washed and exposed to film (24).
Purification of the 90-kDa Protein by ⌬-DE Ran-Sepharose Chromatography-All steps were performed at 4°C. Cytosol was prepared from the brains of 65 rats as described in Ref. 3 in buffer B (20 mM HEPES-KOH, pH 7.3, 2 mM magnesium acetate, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 mg/ml aprotinin, 1 mg/ml leupeptin). The cytosol was adjusted to 150 mM sodium acetate and added to 100 ml of DE52 cellulose slurry (Whatman) equilibrated with buffer B plus 150 mM sodium acetate. The mixture was stirred gently for 30 min and then washed with 1 liter of buffer B plus 150 mM sodium acetate. Proteins were eluted with 2 ϫ 50 ml of buffer B plus 500 mM sodium acetate. The eluate was dialyzed overnight against buffer B plus 100 mM sodium acetate. Proteins that bind nonspecifically to GST and glutathione-Sepharose were removed by incubation with 100 g of GST bound to 100 l of glutathione-Sepharose beads, and the cleared supernatant was purified by ⌬-DE Ran affinity chromatography.
The ⌬-DE Ran affinity matrix was prepared by applying lysate from 1 liter of E. coli expressing GST-⌬-DE Ran (approximately 3 mg) onto 1 ml of glutathione-Sepharose beads. After 30 min, the beads were washed twice with phosphate-buffered saline and once with 25 mM MOPS, pH 7.1, plus 1 mM EDTA. The beads were resuspended in the same buffer plus 2 mM GTP and incubated for 30 min. Magnesium acetate (20 mM) was then added, and the ⌬-DE Ran-Sepharose was centrifuged for 2 min at 2000 ϫ g and washed with binding buffer (24).
Cleared supernatant was incubated with the ⌬-DE Ran-Sepharose for 1-2 h. The mixture was then washed three times with binding buffer plus 0.05% Tween 20 and once with binding buffer and then loaded onto a small column. Purified proteins were eluted with 4 ml of 10 mM glutathione in binding buffer.
Identification of the 90-kDa Protein That Binds to ⌬-DE Ran-Proteins purified from the ⌬-DE Ran Sepharose column were precipitated with 0.1% sodium deoxycholate plus 10% trichloroacetic acid for 16 h and collected by centrifugation. Pellets were washed with acetone, dried, and resuspended in 50 l of a solution containing 0.1 M NaOH and 0.1% deoxycholate. Proteins were separated by SDS-PAGE and transferred to Immobilon p sq (Millipore). Proteins were stained with Ponceau S, and the 90-kDa protein was excised. Approximately 70 pmol of the excised protein was generated for sequence identification. Digestion with trypsin and sequence analysis were performed by the Harvard Microchemistry Facility, Cambridge, MA.
Superose 12 FPLC-Approximately 200 g of brain cytosol proteins, batch-eluted from DE52 cellulose, were incubated with 2 mM GTP for 10 min on ice with or without 3 g of GST-⌬-DE Ran that had been preloaded with GTP. This mixture was loaded onto a Superose 12 column (Pharmacia Biotech Inc.), equilibrated in 20 mM Tris-HCl, pH 7.8 at 4°C, and 50 mM NaCl. The flow rate was 0.25 ml/min, and 0.5-ml fractions were collected. The 4.5-ml void volume was collected in frac-tions 1-8. Fractions 9 -26 were precipitated with 10% trichloroacetic acid as described above, and the proteins were separated by 8% SDS-PAGE and then transferred to nitrocellulose. Transfers were overlaid with [␣-32 P]GTP-GST-⌬-DE Ran as described above. After exposure to x-ray film, the nitrocellulose was immunoblotted with a 1:500 dilution of either anti-Ran12 antiserum (17) or monoclonal anti-GST antibody (Santa Cruz Biotechnology, CA). Horseradish peroxidase-coupled antirabbit or anti-mouse antibodies were then applied at a 1:20,000 dilution and detected by chemiluminescence (Kirkegaard and Perry Laboratories, Gaithersburg, MD).

RESULTS AND DISCUSSION
We showed previously that deletion of the C terminus of Ran reduces the affinity of Ran-GTP for the Ran-binding proteins RanBP1 and RanBP2 (17). In addition, the C-terminal deletion mutant of Ran, ⌬-DE Ran, binds avidly to unidentified proteins of 90 and 115/120 kDa (24). These proteins (⌬-DE RanBPs) associate specifically with ⌬-DE Ran that is in the GTP-bound state; they inhibit release of GTP and are located primarily in the soluble fraction of cell extracts (not shown). The affinity of the proteins for ⌬-DE Ran appeared to be higher than that to wild-type Ran, possibly as a consequence of the unmasking of a region on Ran normally blocked by the C terminus. We reasoned that the formation of a RanBP-Ran heterodimer may present the Ran in a conformation able to bind more efficiently to the 90-and 115/120-kDa proteins.
To test this hypothesis, BHK21 cell extract was separated by SDS-PAGE, transferred to nitrocellulose, and incubated with a complex of [␣ 32 P]GTP-Ran and RanBP1. As predicted, the formation of a Ran-RanBP1 complex dramatically increased the apparent affinity of Ran for the 90-and 115/120-kDa ⌬-DE RanBPs (Fig. 1). The RanBP1 also effectively competed out binding on the nitrocellulose to endogenous RanBP1 and RanBP2 from BHK21 cells. The association of Ran with the 90and 115/120-kDa proteins was further intensified when Ran was complexed with an isolated RanBD of RanBP1 (Fig. 1) or of RanBP2 (not shown). Removal of the C terminus from RanBP1 also enhanced the binding of Ran to the ⌬-DE RanBPs (Fig. 1). These results show that the behavior of the ⌬-DE Ran mutant can be mimicked by the heteromeric association of full-length Ran with a Ran binding domain. The data suggest that this association causes a change in the conformation of the C terminus of the Ran GTPase that exposes a binding site for the 90and p115/120-kDa proteins. These proteins are potentially important effectors of Ran function. We therefore pursued the identification of the 90-kDa protein by affinity purification using a GST fusion of ⌬-DE Ran. Fusion to GST did not affect the binding of ⌬-DE Ran to the 90and 115/120-kDa proteins or increase binding to nonspecific proteins in overlay assays (Fig. 1, lane 5). Using the rat brain cytosol fractions defined by Moore and Blobel (3), we observed that Ran is found exclusively in fraction B, and the RanBPs and ⌬-DE RanBPs were both present in fraction A, the component required for docking of NLS substrates (3,11). We therefore prepared a modified fraction A from rat brain cytosol as an initial step in the purification of the 90-kDa protein. This partially purified material was then affinity-purified using a glutathione-Sepharose matrix coupled to GST-⌬-DE Ran. The affinity step yielded the 90-and 115/120-kDa proteins, purified nearly to homogeneity (Fig. 2, lane 2). Peptide sequences of two separate trypsin fragments of the 90-kDa protein were identical to sequences within p97/␤-karyopherin, an essential component of the 9 S nuclear docking complex (Table I) (4).
To confirm that ␤-karyopherin is a Ran binding protein, recombinant ␤-karyopherin carrying a His 6 -tag (5) was tested for its ability to interact with [␣-32 P]GTP-Ran in an overlay assay. The conditions that enhance binding of Ran to ␤-karyopherin (Fig. 3, lower panel) paralleled those that promote binding of Ran to the 90-kDa protein from brain cytosol (upper panel). Wild-type Ran in the presence of GST alone did not detectably bind recombinant ␤-karyopherin (lane 1). As predicted, however, both ⌬-DE Ran (lower panel, lane 5) and a complex of wild-type Ran with the isolated Ran binding domain of RanBP1 (lane 2) did associate with recombinant ␤-karyopherin. The addition of ␤-karyopherin to the Ran-RanBP1 com-plex also abolished binding of the complex to immobilized ␤-karyopherin and to the 90-and 115/120-kDa brain proteins (lane 4). It did not reduce binding to RanBP2, however (lane 3), confirming that ␤-karyopherin interacts with a region on Ran distinct from the RanBD binding site. Note that the recombinant ␤-karyopherin was less efficient in competition for binding to the ⌬-DE Ran (lane 6). We do not currently understand the basis for this difference.
Endogenous ␤-karyopherin is present in the cytosol as part of a multisubunit complex (4), and we explored the possibility that Ran interacts with this complex. Proteins in the modified fraction A from rat brain cytosol were mixed with GST (Fig. 4A) or GST-⌬-DE Ran (Fig. 4B) loaded with GTP. Protein complexes were then separated by size exclusion chromatography. Fractions were analyzed by overlay with GST-⌬-DE Ran to detect the 90-and 115/120-kDa proteins. After exposure to x-ray film, the same nitrocellulose blot was washed to remove the probe and immunoblotted with either anti-Ran antiserum (Fig. 4B, bottom panel) or anti-GST antibody (not shown) to identify fractions containing GST-⌬-DE Ran. Two distinct peaks of ⌬-DE RanBPs were detected in fractions that correspond to sizes of approximately 700 and 200 kDa (Fig. 4A). The 700-kDa peak may represent the 9 S complex (4). The addition of GST-⌬-DE Ran-GTP to the cytosol did not affect the elution profile of the ⌬-DE RanBPs (Fig. 4B), but the GST-⌬-DE Ran preferentially associated with the 700-kDa peak. One interpretation of this result is that other proteins contained in the large complex increase the affinity of Ran for the ⌬-DE RanBPs.
Together these data fit well with the current model for me-

FIG. 3. Interaction of [␣-32 P]GTP-⌬-DE Ran with
␤-karyopherin. 150 g of rat brain cytosol proteins (upper panel) or 0.5 g of recombinant His 6 -␤-karyopherin (lower panel) were separated by 8% SDS-PAGE and transferred to nitrocellulose. [␣-32 P]GTP-Ran or [␣-32 P]GTP-GST-⌬-DE Ran (2 g of protein each; 300,000 cpm; specific activity, 3000 Ci/mmol) were mixed with the proteins indicated above the top panel, then were added separately to the nitrocellulose strips in 10 ml of binding buffer, and incubated for 30 min, as described in Fig.  1. GST was included in lane 1 as a negative control. ␤-kar, recombinant ␤-karyopherin. The nitrocellulose strips were then washed and exposed to x-ray film overnight. The positions of the 90-kDa protein, 115/120-kDa protein, and ␤-karyopherin are indicated by arrows. diated transport of karyophiles through nuclear pore complexes and, importantly, they link the two major steps in nuclear import: docking and translocation. We propose a model in which ␤-karyopherin serves as an adapter between the ␣-karyopherin-NLS substrate complex and a Ran-RanBP heterodimer. RanBP1 may stabilize this complex in the cytosol until it is displaced by RanBP2 at the nuclear pore. Hydrolysis of Ran-GTP, stimulated by the Ran GTPase-activating protein, Rna1p, then triggers dissociation of the complex and permits the entry of the NLS substrate-␣-karyopherin heteromer into the nucleus, perhaps with NTF2/B-2. The GDP-Ran released from the complex also enters the nucleus where RCC1 catalyzes the exchange of GTP for GDP on Ran such that the cycle can continue. The tools currently available will now allow for a more detailed biochemical analysis of the interactions between the growing number of components that regulate nuclear import.