Carrier-independent Nuclear Import of the Transcription Factor PU.1 via RanGTP-stimulated Binding to Nup153*

PU.1 is a transcription factor of the Ets family with important functions in hematopoietic cell differentiation. Using green fluorescent protein-PU.1 fusions, we show that the Ets DNA binding domain of PU.1 is necessary and sufficient for its nuclear localization. Fluorescence and ultrastructural nuclear import assays showed that PU.1 nuclear import requires energy but not soluble carriers. PU.1 interacted directly with two nucleoporins, Nup62 and Nup153. The binding of PU.1 to Nup153, but not to Nup62, increased dramatically in the presence of RanGMPPNP, indicating the formation of a PU.1·RanGTP·Nup153 complex. The Ets domain accounted for the bulk of the interaction of PU.1 with Nup153 and RanGMPPNP. Because Nup62 is located close to the midplane of the nuclear pore complex whereas Nup153 is at its nuclear side, these findings suggest a model whereby RanGTP propels PU.1 toward the nuclear side of the nuclear pore complex by increasing its affinity for Nup153. This notion was confirmed by ultrastructural studies using gold-labeled PU.1 in permeabilized cells.

Nuclear import of proteins occurs through the nuclear pore complex (NPC), 1 and in most cases involves an interaction between soluble carriers and nuclear localization sequences on the cargo protein (1,2). Most soluble carriers belong to the karyopherin/importin family and mediate transport either as monomers or as heterodimers. For example, classical nuclear localization sequences are bound by a heterodimer that consists of karyopherin ␣ (importin ␣) and karyopherin ␤1 (importin ␤). Members of the karyopherin ␤ family interact with nucleoporins and thus mediate the interaction between their cargoes and the NPC. A subset of nucleoporins contains FG repeats and have been implicated in nuclear import through interacting with cargo and carriers (3). Carrier-mediated nuclear import usually requires energy in the form of GTP, although exceptions have been reported (4,5). The GTP requirement is mediated by the small GTPase Ran, which cycles between GDP-and GTP-bound forms (RanGDP and RanGTP, respectively). RanGTP is present in the nucleus where it dissociates import complexes by binding to members of the karyopherin ␤ family. Carrier proteins such as karyopherin ␤1 (importin ␤), karyopherin ␤2 (transportin), and karyopherin ␣ (importin ␣) can enter the nucleus unassisted by soluble proteins, presumably by interacting with nucleoporins (6 -8).
Some proteins can enter the nucleus in the absence of any carriers. These include the hnRNP K protein (9), ␤ catenin (10 -12), the HIV Vpr protein (13), SMAD transcription factors (14,15), ERK2 (16,17), and STAT transcription factors (18). In most cases this is thought to involve a direct interaction between these proteins and nucleoporins, although contradictory results have been reported in the case of ␤ catenin (10,12). The role of energy in carrier-independent nuclear import is not clear. In the case of Vpr, ERK2, and STAT transcription factors, nuclear import occurs in the absence of energy (13, 16 -18). On the other hand, although initial results indicated that the nuclear import of ␤ catenin required energy (10), later reports suggested that ␤ catenin nuclear import was energyindependent (11,12). The energy requirements for the carrierindependent import of hnRNP K and SMAD transcription factors have not been reported.
PU.1 (also called Spi-1) (19 -21) belongs to the Ets family of transcription factors and binds to purine-rich recognition sites in gene promoters through its Ets DNA binding domain (22). It is expressed in several hematopoietic cell lineages and plays a pivotal role in the differentiation of myeloid cells and lymphocytes (23,24). The concentration of PU.1 plays an important role in determining cell lineage during hematopoiesis (25). In addition, several lines of evidence have implicated PU.1 in the pathogenesis of leukemia. PU.1 was originally identified as an oncogene that is up-regulated in mouse erythroleukemias induced by the Friend virus, and PU.1 transgenic mice develop erythroleukemia (26,27). In addition, frequent PU.1 gene mutations have been described in a series of patients with acute myeloid leukemia (28), although other studies found a much lower frequency of such mutations (29 -31). Finally, PU.1 is down-regulated in certain subsets of acute myeloid leukemia (32,33). Thus, the level and activity of PU.1 are important both in normal hematopoiesis and in leukemogenesis. Transcription factors can be regulated at several levels, including transcription, post-translational modifications, and nuclear import. Although several studies have addressed the regulation of PU.1 at the transcriptional/mRNA level (33)(34)(35)(36)(37) and at the post-translational level (38 -46), the mechanisms by which PU.1 is imported into the nucleus have not been studied.
Here we dissect the nuclear import of PU.1 using both classical fluorescence-based nuclear import assays and ultrastructural import assays and show that it is independent of soluble carriers but requires energy. The karyopherin independence is explained by direct interactions between PU.1 and nucleoporins Nup62 and Nup153. The affinity of PU.1 for Nup153 is increased dramatically in the presence of RanGMPPNP (GMP-PNP is a nonhydrolyzable analog of GTP), indicating the formation of a ternary PU.1⅐RanGTP⅐Nup153 complex. In contrast, RanGMPPNP does not enhance the binding of PU.1 to Nup62. Because Nup62 is located close to the midplane of the NPC and Nup153 is at the nuclear end of the NPC, these findings suggest that RanGTP creates a gradient of increasing affinity for PU.1 toward the nuclear side of the NPC. This notion is supported by ultrastructural studies using gold-labeled PU.1 in permeabilized cells. These findings provide an explanation for the energy requirement of PU.1 nuclear import. Using GFP-PU.1 fusions, we show that the Ets domain of PU.1 is necessary and sufficient for its nuclear localization in transfected cells. The Ets domain also accounts for the bulk of the interaction of PU.1 with Nup153 and RanGMPPNP, underscoring the importance of this interaction for nuclear import. Based on these findings we suggest a model for the energy-dependent translocation of PU.1 through the NPC.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-HeLa cells were grown at 37°C with 5% CO 2 in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin.
Plasmids that express GFP-PU.1 and its derivatives were transfected into HeLa cells grown on coverslips using FuGENE 6 (Roche Applied Science) according to the manufacturer's instructions.
In Vitro Nuclear Import Assays-HeLa cells grown on coverslips were washed twice with PBS and then washed twice with transport buffer (TB) consisting of 20 mM HEPES⅐KOH, pH 7.3, 110 mM potassium acetate, and 2 mM magnesium acetate. The cells were permeabilized with 35 g/ml digitonin (Calbiochem) in TB containing protease inhibitors (complete EDTA-free protease inhibitor mixture tablet, Roche Applied Science) for 5 min at room temperature. After two washes with 50 l of TB, the cells were incubated with 30 l of the import mix for 20 min at room temperature. All import mixes contained the import substrate and 2 mg/ml bovine serum albumin (Sigma) in TB. The energy-regenerating system consisted of 2 mM ATP (Sigma), 5 mM phosphocreatine (Sigma), 20 units/ml creatine phosphokinase (Sigma), and 2 mM GTP (Sigma). The complete nuclear import assay with carriers included 0.6 g of karyopherin ␣3, 0.5 g of karyopherin ␤1, and 0.6 g of RanGDP, in addition to the above. The cells were then washed twice with TB followed by fixation with 4% formaldehyde (Polysciences) in TB for 15 min at room temperature. The cells were washed twice with TB and mounted with ProLong antifade reagent (Molecular Probes). The cells were analyzed by confocal microscopy with a Leica TCS SP spectral confocal microscope.
Electron Microscopy-Recombinant PU.1 was conjugated to 5-nm colloidal gold beads (BBI International) as described previously (47). HeLa cells were seeded at about 30% confluence the day before the experiments in a 24-well plate. Cells were washed twice with PBS and twice with transport buffer and were permeabilized with digitonin (as described above). Permeabilized cells were incubated with 200 l of import mix for 30 or 60 min at room temperature with gentle rocking every 5 min. The import mix contained gold-conjugated PU.1 and 2 mg/ml bovine serum albumin in TB, in the presence or absence of the energy-regenerating system (see above). After the import reaction, the cells were washed twice with TB and fixed with 2.5% glutaldehyde in 0.1 M sodium cacodylate, pH 7.4. The samples were prepared as described (48) and examined with a JEOL 100CX electron microscope at an accelerating voltage of 80 kV.
Immunofluorescence Microscopy-HeLa cells grown on coverslips were washed twice with PBS and then fixed with 4% formaldehyde in PBS for 15 min at room temperature. After two washes with PBS, the cells were permeabilized with 0.2% Triton X-100 in PBS for 5 min at room temperature. The cells were washed twice with PBS and then incubated with the mouse monoclonal antibody mAb414 (Covance), that recognizes several nucleoporins and therefore highlights the nuclear membrane, at 1:2000 dilution in PBS containing 2% bovine serum albumin for 1 h at room temperature. After three washes in PBS, the cells were incubated with Cy5-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc.) (1:200 dilution in PBS containing 2% bovine serum albumin) for 30 min at room temperature. The cells were washed three times with PBS and then mounted with ProLong antifade reagent. The cells were analyzed by confocal microscopy with a Leica TCS SP spectral confocal microscope.
Plasmids-The human PU.1 open reading frame was obtained by PCR from a pGEX-3X vector kindly provided by Dr. Elizabeth Eklund and subcloned into the BamHI and XhoI sites of pGEX-6P1 (Amersham Biosciences). The resulting 6P1-PU.1 vector was used to produce recombinant PU.1 protein for ultrastructural and binding studies. The EGFP open reading frame was amplified with EcoRI and BamHI flanking sites by PCR from the pIRES2-EGFP vector (BD Biosciences) and subcloned along with PU.1 (flanked by BamHI and XhoI sites) between the EcoRI and XhoI sites of pGEX-6P1 by triple ligation. The resulting 6P1-GFP-PU.1 vector was used to produce GST-GFP-PU.1 protein for binding assays and GFP-PU.1 protein for in vitro import assays. The PU.1 open reading frame was divided into PU(1-149) and PU(150 -264) using the SmaI site and into PU(1-94) and PU(94 -264) using the NcoI site. These fragments were subcloned into pGEX-4T1 (Amersham Biosciences) and into pcDNA3 (Invitrogen) along with an N-terminal GFP tag in a manner similar to that described for full-length PU.1. The vector pEGFP-C1 (BD Biosciences) was used to express GFP as a control in transfection studies.
The open reading frame of Nup62 was amplified from Human Testis Marathon Ready cDNA (BD Biosciences) by PCR and was cloned into the EcoRI/XhoI sites of pcDNA3.1/HisB (Invitrogen). Nup62N encoding amino acids 1-265 of Nup62 and Nup62C encoding amino acids 178 -522 of Nup62 were cloned into EcoRI/XhoI sites of pcDNA3.1/HisB. Nup62N was also subcloned into the EcoRI/XhoI sites of pGEX4T-3 for recombinant protein expression.
Protein Binding Assays-Nup153 and Nup62, full-length and deleted, were produced using the TNT T7 quick coupled transcription/ translation system (Promega) in the presence of Tran 35 S-label (ICN, Irvine, CA). The amounts of translation products from different deletion constructs were compared by SDS-PAGE and autoradiography, and the amounts used for binding assays were adjusted accordingly to produce bands of equal intensity. Binding assays were carried out essentially as described previously (51,53). For a binding reaction, 10 l of beads of immobilized recombinant protein were incubated for 1 h at 4°C with either in vitro translated or recombinant protein in 36 l of TB with 0.1% Tween 20 (TB-T). The supernatants and the material that remained bound to the beads after three washes in cold TB-T were separated in 4 -20% polyacrylamide gradients by SDS-PAGE. For in vitro translated proteins, the gel was dried and exposed to film. Recombinant PU.1 was detected by immunoblotting using anti-PU.1 antibody (Santa Cruz Biotechnology, sc-352), horseradish peroxidase-conjugated anti-rabbit antibody (Santa Cruz Biotechnology, sc-2054), and the SuperSignal West Pico substrate system (Pierce). The ratio of the unbound to bound materials analyzed was ϳ1:4.

The Ets Domain of PU.1 Is Necessary and Sufficient for Its
Nuclear Import in Intact Cells-Although PU.1 is normally expressed in hematopoietic cells, the in vitro assays used for dissecting the mechanisms of nuclear import are usually carried out in permeabilized epithelial cells such as HeLa cells (54). To determine whether PU.1 would undergo nuclear import in HeLa cells, a construct expressing a GFP-PU.1 fusion protein was transfected into HeLa cells. As shown in Fig. 1, GFP-PU.1 localized exclusively to the nucleus, whereas the GFP tag alone was distributed diffusely throughout the cell.
To identify the region of PU.1 responsible for its nuclear localization, deletion constructs of PU.1 fused to GFP were prepared and transfected into HeLa cells. As shown in Fig. 1, PU.1 fragments that contain the Ets DNA binding domain (amino acids 150 -264) are localized exclusively to the nucleus, whereas fragments that do not include the Ets domain are distributed diffusely throughout the cell in a pattern similar to that of the GFP control. Thus, the Ets domain is necessary and sufficient for the nuclear localization of PU.1.
Nuclear Import of PU.1 Requires Energy but Not Import Carriers-To elucidate the mechanisms by which PU.1 is im-ported into the nucleus, recombinant GFP-PU.1 fusion protein was used in a classical in vitro nuclear import assay (54). Cells were permeabilized with digitonin to solubilize the plasma membrane and washed to remove cytosolic components. This treatment leaves the nuclear membrane intact. When GFP-PU.1 was added to the permeabilized cells alone, it was localized mainly in the cytoplasm ( Fig. 2A, left panel). The addition of GTP and an energy-regenerating system resulted in import of GFP-PU.1 into the nucleus ( Fig. 2A, middle panel). For the sake of simplicity, we will refer to GTP and the energy regenerating system as "energy." A standard nuclear import mixture that includes karyopherin ␤1, karyopherin ␣, and RanGDP, in addition to energy, did not further increase import ( Fig. 2A,  right panel). Thus, GFP-PU.1 import required energy but not import carriers.
To confirm this result further and to exclude any artifact related to the GFP tag, an ultrastructural in vitro nuclear import assay was carried out using recombinant PU.1 conjugated to 5-nm gold beads. Cells were permeabilized with digitonin and incubated with gold-conjugated PU.1 in the presence or absence of energy. As shown in Fig. 2B 2. A, GFP-PU.1 nuclear import is energy-dependent and carrierindependent. Digitonin-permeabilized HeLa cells were incubated for 30 min with an import assay mixture containing GFP-PU.1 in the absence of energy (left panel), in the presence of energy (middle panel), and in the presence of energy and import carriers (right panel) (see "Experimental Procedures"). At least 200 cells for each condition were examined, and more than 95% showed a GFP distribution similar to that shown. Bar, 10 m. B, digitonin-permeabilized HeLa cells were incubated for 30 min at room temperature with the import assay mixture containing PU.1/5-nm gold conjugates in the presence or absence of energy. Samples were processed for electron microscopy as described under "Experimental Procedures," and beads in the nucleus and cytoplasm were counted in electron micrographs from five randomly selected areas for each condition. The average number of beads/m 2 of nucleus (N) and cytoplasm (C) is shown. Bar, 100 nm. nucleus in the presence of energy, confirming the energy-dependent and carrier-independent nuclear import of PU.1. Carrier-independent import is usually attributable to interactions between the protein and nucleoporins (10, 11, 14 -17). Therefore, these findings suggest that PU.1 might interact with nucleoporins.
PU.1 Interacts Directly with Nucleoporins-A group of nucleoporins containing FG repeats have been implicated in mediating nuclear import (3), and proteins that undergo carrier-independent nuclear import tend to interact with this group of nucleoporins. For example, Nup153 interacts with several proteins that undergo carrier-independent nuclear import including karyopherin ␤1, karyopherin ␤2, ERK2, and SMAD2 (17,(55)(56)(57). Nup153 is for the most part located asymmetrically on the nuclear side of the NPC, but its C terminus is thought to be flexible and has been localized to both sides of the NPC (58 -61). Another FG repeat nucleoporin, Nup62, is implicated in nuclear import and interacts with several ␤ karyopherins (3). Down-regulation of Nup153 and Nup62 by poliovirus or rhinovirus infection leads to inhibition of nuclear import (62,63).
To determine whether PU.1 interacts with nucleoporins, recombinant GST-GFP-PU.1 was immobilized on glutathione-Sepharose 4B beads and used in binding assays with in vitro translated Nup62 and Nup153.
Recombinant GST-GFP immobilized on glutathione-Sepharose 4B beads was used as a negative control. PU.1 and control beads were incubated with in vitro translated full-length Nup62, an N-terminal fragment that includes its FXFG repeats, or a C-terminal fragment that does not contain FG repeats (Fig. 3A). As shown in Fig. 3B, PU.1 interacted specifically with full-length Nup62 and with the N-terminal FG repeat portion of Nup62 (Nup62N), but not with its C-terminal portion (Nup62C). Because these reactions contain reticulocyte lysate, it is possible that the interaction of Nup62 with PU.1 is mediated by molecules present in the lysate. To exclude this possibility, recombinant Nup62N, fused to GST, was purified from Escherichia coli, immobilized on glutathione-Sepharose 4B beads, and incubated with purified recombinant PU.1. GST similarly immobilized on beads was used as a control. As shown in Fig. 3C, PU.1 bound to Nup62N specifically in the absence of any other factors, demonstrating direct binding of PU.1 to the FG repeat region of Nup62.
To determine whether PU.1 binds to Nup153, similar binding assays using in vitro translated full-length Nup153 and a series of deletions were carried out (Fig. 4). Full-length Nup153 bound to PU.1 specifically (Fig. 4A). Several N-and C-terminal Nup153 deletion mutants were tested for binding to PU.1 (Fig.  4B). All deletion constructs bound to PU.1, indicating that the PU.1 binds to more than one site on Nup153. Interestingly, the C-terminal region of Nup153 which includes the FG repeats and zinc finger domains (64) appears to have an inhibitory effect on the binding of PU.1. To confirm that the interaction with Nup153 is direct, a recombinant fragment consisting of amino acids 346 -1175 of Nup153, fused to GST, was purified from E. coli and immobilized on glutathione-Sepharose 4B beads. As shown in Fig. 4C, this fragment interacted specifically with purified recombinant PU.1.
These findings show that PU.1 interacts directly with two nucleoporins implicated in nuclear import, Nup153 and Nup62, and thus provide an explanation for the carrier-independent nuclear import of PU.1. However, they do not explain the requirement of this import for energy.
PU.1 Interacts Differentially with Nucleoporins in the Presence of RanGMPPNP-As shown in Fig. 2, nuclear import of PU.1 requires energy. The requirement for energy in carrier-mediated nuclear import is attributed to the conversion of RanGDP to RanGTP; the latter binds to ␤ karyopherins resulting in dissociation of the carrier-cargo complex (2). The role of energy in carrier-independent nuclear import is not well understood. To determine whether the energy requirement in PU.1 nuclear import is also mediated by RanGTP, we sought to compare the effects of RanGDP and RanGTP on the interaction between PU.1 and nucleoporins. To avoid the possibility of GTP hydrolysis by Ran during incubation with reticulocyte lysate, we used Ran loaded with the nonhydrolyzable GTP analog GMPPNP (RanGMPPNP). PU.1, immobilized on beads, was incubated with in vitro translated Nup153 in the presence or absence of RanGDP or RanGMPPNP. As shown in Fig. 5A, RanGMPPNP dramatically enhanced the binding of Nup153 to PU.1, whereas RanGDP had no significant effect. This indicates that PU.1, Nup153, and RanGTP form a ternary complex. Binding assays with deletion mutants of Nup153 revealed that this binding was diminished greatly by deletion of amino acids 300 -609 (Fig. 5B). Thus, the Nup153 region between amino acids 300 and 609 is required for the formation of this complex. It is of interest to note that although deletion of the C terminus of Nup153 increases the binding of Nup153 to PU.1 in the absence of RanGMPPNP (Fig. 4B), this effect is abrogated in the presence of RanGMPPNP (Fig. 5B).
In contrast to the findings with Nup153, RanGMPPNP had a slightly inhibitory effect on the interaction between PU.1 and Nup62 (Fig. 5C). Thus, in the absence of RanGTP, PU.1 binds to Nup62 and Nup153 at comparable levels, whereas in the presence of RanGMPPNP it binds Nup153 much more strongly. Nup153 is present at the nuclear side of the NPC, whereas Nup62 is close to the midplane of the NPC. Therefore, these findings suggest that energy stimulates the nuclear import of PU.1 by favoring its movement toward the nuclear side of the NPC. Consistent with this notion, the Ets domain of PU.1, which is required for its nuclear import (Fig. 1), interacted preferentially with Nup153 both in the presence and absence of RanGMPPNP (Fig. 5D), underscoring the importance of this interaction for PU.1 nuclear import.
PU.1 Moves to the Nuclear Side of the NPC in the Presence of Energy-To determine whether GTP shifts NPC-associated PU.1 molecules toward the nuclear side of the NPC, PU.1 conjugated to 5-nm gold beads was applied onto permeabilized cells in the presence or absence of energy. Gold beads associated with the NPC were visualized by transmission electron microscopy, and their distances from the midplane of the NPC were measured. As shown in Fig. 6, the addition of energy resulted in the movement of gold beads to the nuclear side of the NPC. These data support the notion that RanGTP facilitates the nuclear import of PU.1 by increasing its affinity for the nuclear side of the NPC. DISCUSSION The data presented here suggest a model for the energydependent, carrier-independent translocation of a protein through the NPC (Fig. 7). PU.1 interacts directly with nucleoporins Nup62 and Nup153, and in the presence of RanGTP its interaction with Nup153 becomes much stronger. Because Nup153 is located at the nuclear end of the NPC, the addition of GTP to the import reaction favors movement of PU.1 toward the nucleus.
Several models have been suggested for translocation of carrier-cargo complexes through the NPC. Although there is general agreement on the importance of interactions between karyopherins and nucleoporins for transport, the main point of disagreement among different models concerns the mechanisms by which directionality of transport is assured. Some models hold that the FG repeats of various nucleoporins are essentially equivalent to one another and would facilitate the passage of an interacting karyopherin in either direction. Directionality of transport is then explained solely by the presence of RanGTP on the nuclear side and RanGDP on the cytoplasmic side (65)(66)(67). Other models, on the other hand, postulate a role for nucleoporins that are located asymmetrically on either the nuclear or cytoplasmic side of the NPC in the directionality of transport (68 -70).
The mechanisms of carrier-independent nuclear import are even less clear. Several proteins have been shown to accumulate in the nucleus in the absence of exogenously added proteins in an in vitro nuclear import assay. These include carrier proteins such as karyopherin ␤1 (importin ␤), karyopherin ␤2 (transportin), and karyopherin ␣ (importin ␣) (6 -8), as well as noncarrier proteins including hnRNP K (9), ␤ catenin (10 -12), the HIV Vpr protein (13), SMAD transcription factors (14,15), ERK2 (16,17), and STAT transcription factors (18). In the case of karyopherin ␤1, karyopherin ␣, Vpr, ERK2, and STAT proteins, carrier-independent nuclear import occurs in the absence of energy (6, 8, 13, 16 -18, 71). In vitro nuclear import of karyopherin ␤2 (transportin) has been carried out in the presence of an energy-regenerating system and is cold-sensitive; thus it may require energy, but this has not been shown conclusively (7). Similarly, the energy requirements for the carrier-independent import of hnRNP K and SMAD transcription factors have not been determined conclusively. Initial results indicated that the nuclear import of ␤ catenin required energy (10), but later reports suggested that ␤ catenin nuclear import was energy-independent (11,12). To summarize, in some cases carrier-independent nuclear import has been shown not to require energy, whereas in others the energy requirements have not been determined clearly.
In this paper, we show that PU.1 enters the nucleus in the absence of exogenously added proteins and that this import clearly requires energy. The carrier-independent nuclear import of PU.1 can be explained in part by its direct interactions with nucleoporins (Figs. 3 and 4). However, these interactions do not account for the energy requirement of PU.1 nuclear import. Ultrastructural analysis of NPC-associated PU.1 in permeabilized cells showed that in the absence of energy, PU.1 was associated predominantly with the cytoplasmic side of the NPC (Fig. 6A). Association of PU.1 with the nuclear side of the NPC was observed only in the presence of energy (Fig. 6B). Furthermore, the binding of PU.1 to Nup153, which is present at the nuclear side of the NPC, increases dramatically in the presence of a RanGTP analog (Fig. 5). These data suggest a model whereby RanGTP resulting from the addition of energy propels PU.1 toward the nucleus by increasing its affinity to the nuclear side of the NPC (Fig. 7). Low affinity interactions of PU.1 with nucleoporins appear to target PU.1 to the NPC but are not sufficient to translocate it to the nuclear side of the NPC and into the nucleus. Once energy is added, nuclear Ran is converted to RanGTP, and PU.1 shifts to the nuclear side of the NPC where it forms a ternary complex with RanGTP and Nup153. Thus, it appears that both the RanGTP gradient and the asymmetrical distribution of nucleoporins at the NPC play a role in the nuclear translocation of PU.1.
The mechanism by which PU.1 dissociates from Nup153 to reach its destinations inside the nucleus remains to be determined. Although the majority of NPC-associated RanGAP is located on the cytoplasmic side of the NPC, there is some RanGAP on the nuclear side of the NPC as well (72). It is therefore possible that GTP hydrolysis mediated by RanGAP may occur at the nuclear side of the NPC resulting in dissociation of PU.1 from Nup153.
The formation of a ternary complex between PU.1, RanGTP, and Nup153 is somewhat reminiscent of export complexes that include an export carrier, cargo, and RanGTP; and RanGTPdependent binding of Nup153 to export receptors such as Crm1 and Xpo-t has been reported (56,73). These considerations raise the possibility that the PU.1⅐Nup153⅐RanGTP complex may be an export intermediate. However, the addition of GTP and energy to PU.1 in the permeabilized cell assay results in import rather than export of PU.1. Further, transfection studies using deletion constructs (Fig. 1) do not show evidence of an export signal: even when the entire nuclear targeting Ets domain is removed, PU.1 is distributed diffusely throughout the cell (similar to the GFP control) rather than being actively exported to the cytoplasm. Finally, Nup153 and a RanGTP analog interact preferentially with the Ets domain that is responsible for nuclear import of PU.1, further emphasizing a role for this interaction in nuclear import.
In summary, we have demonstrated energy-dependent, carrier-independent nuclear import of a transcription factor, and we show evidence that this import is mediated by direct interactions with nucleoporins that are skewed toward the nuclear side of the NPC in the presence of RanGTP. In view of the importance of PU.1 concentration in determining cell fate, it is of interest to determine whether these interactions can be regulated to control the access of PU.1 to its target genes during differentiation and leukemogenesis.