Nuclear Import of the TATA-binding Protein: Mediation by the Karyopherin Kap114p and a Possible Mechanism for Intranuclear Targeting

All yeast strains were derived from DF5 (Finley et al., 1987) and the procedures for yeast manipulation were as described (Aitchison et al., 1995). KAP114 was deleted by integrative transformation of HIS3. Heterozygous diploids were sporulated and tetrads dissected to generate kap114 haploid strains (Aitchison et al., 1995). kap114/kap123 deletion strains were constructed by mating, sporulating, and dissecting the relevant strains. Diploid strains expressing the Kap114–protein A (PrA) and TBP-PrA fusion proteins were constructed by integrative transformation of the coding sequence of four and a half IgG binding repeats of Staphylococcus aureus PrA immediately upstream of the relevant stop codon as described (Aitchison et al., 1995). Haploid strains were generated by sporulation and dissection. Other strains were constructed by PrA integration downstream of TBP in the relevant haploid deletion strains.

After fixation in 3.7% formaldehyde for 20 min, immunofluorescence microscopy on yeast spheroplasts was done as previously described (Pemberton et al., 1997). PrA tags were visualized using rabbit anti–mouse IgG (preadsorbed against formaldehyde-fixed wild-type yeast cells) followed by Cy3-conjugated donkey anti–rabbit IgG (Jackson Labs). GFP-expressing cells were applied to slides and briefly permeabilized in methanol, followed by acetone. All cells were mounted in a 4′,6-diamidino-2-phenylindole (DAPI)–containing medium. All images were viewed under the 63× oil objective on a Zeiss Axiophot microscope, images were collected with a video imaging system, and manipulated in the computer program Adobe Photoshop.

Postnuclear, postribosomal cytosol was prepared from 4 liters of the TBP-PrA and Kap114-PrA strains and 1 liter of TBP-PrA/kap114 and 160 ml of TBP-PrA/kap114/kap123 strains grown to an OD₆₀₀ of 1.6 as described (Aitchison et al., 1996). Kap114-PrA or TBP-PrA and associated proteins were immunoisolated by overnight incubation of cytosol with rabbit IgG Sepharose (Cappel Laboratories) as described. After washing in TB (20 mM Hepes, pH 7.5, 110 mM KOAc, 2 mM MgCl₂, 1 mM DTT, 0.1% Tween 20), proteins were eluted from the Sepharose with a step gradient of MgCl₂ and precipitated with methanol before analysis by SDS-PAGE. Coomassie staining bands were excised and prepared for analysis by MALDI-TOF mass spectrometry (MS) and by peptide sequencing (Fernandez et al., 1994; Gharahdaghi et al., 1996).

The entire coding region of KAP114 was amplified by PCR and ligated into the vector pGEX5X-1 (Pharmacia Biotech). The resulting GST-Kap114 fusion protein was purified from induced bacteria following the manufacturer's instructions. Purified GST was obtained from the manufacturer (Pharmacia) and recombinant TBP expressed in bacteria as a His₆ fusion protein was a gift of Dr. P.A. Weill (Vanderbilt University, Nashville, TN). Approximately 0.5 μg or 1 μg of GST-Kap114 or 2 μg of GST was bound to 20 μl of a 50% slurry of glutathione-Sepharose in TB/ 0.1% casamino acids. After washing, Sepharose was incubated with ∼1 μg of TBP for 30 min in the above buffer. The supernatant was collected and constituted the unbound fraction, the Sepharose was washed six times in TB, and the bound fraction was collected by boiling in the sample buffer.

Ran (Gsp1p) was expressed in bacteria and purified and loaded with GTP or GDP as described (Floer and Blobel, 1996). Oligonucleotides (TATA, 5′-CTGTATGTATATAAAACG-3′; M1, 5′-CTGTATGTAGAGAAAACG-3′) were annealed by standard procedures and diluted in TB. Purified TFIIA was a gift from Dr. S. Hahn (Fred Hutchison Cancer Research Center, Seattle, WA) (Ranish et al., 1992). Kap114-PrA and TBP were isolated in batch by incubation with IgG Sepharose overnight, ∼1–1.5 ml of cytosol (equivalent to 33–50 ml of yeast grown to an OD₆₀₀ of 1.6) per dissociation was used. After washing in TB, the Sepharose was divided into separate microfuge tubes for each experiment, resulting in ∼10–20 μl of Sepharose per tube. Ran, DNA, and TFIIA were added as indicated to give a final volume of 100 μl. 0.1 mM GTP was included, except when using RanGDP. Tubes were rotated for 50 min at room temperature. The entire contents of the tube were transferred to a 2-ml disposable column and the liquid was drained and collected. The Sepharose was washed with a further 100 μl of TB, together these fractions constituted the eluted fraction. The columns were washed with 10 ml of TB and drained. Proteins bound to the Sepharose were eluted either with 200 μl of 1 M MgCl₂ followed by 200 μl of 4.5 M MgCl₂, or directly with 4.5 M MgCl₂ to elute the Kap114-PrA and bound TBP from the Sepharose, this constituted the bound fraction. All fractions were precipitated and separated by SDS-PAGE and immunoblotted. Amido black–stained blots were observed to see the excess unbound Ran in the eluted fraction that acted as an internal standard for the protein loading in this fraction. Procedures for Western blotting and subsequent detection by enhanced chemiluminescence were as described (Amersham). An affinity-purified rabbit polyclonal antibody was used to detect TBP. This antibody also interacts with PrA, so the levels of Kap114-PrA were compared as an internal standard for the bound fractions. All dissociations were carried out several times and representative experiments are shown.

In yeast, TBP is encoded by the gene SPT15 (Eisenmann et al., 1989). Analysis of the TBP amino acid sequence revealed no known NLSs. To find the Kaps that are responsible for the import of TBP, we sought to isolate proteins that interacted with the low abundance, cytosolic pool of TBP. This pool most likely represents newly synthesized TBP en route to the nucleus. We genomically tagged TBP with an in frame carboxy-terminal fusion of the IgG-binding domain of PrA. Therefore, haploid cells expressed a single PrA-tagged copy of TBP (TBP-PrA) from its endogenous promoter. The tagged version of TBP was functional as it complemented growth and was localized to the nucleus (see below). TBP-PrA and associated proteins were isolated by incubation of a postribosomal cytosol with IgG-Sepharose and elution with a step gradient ranging from 0.05 to 4.5 M MgCl₂. The eluted fractions were analyzed by SDS-PAGE and Coomassie blue staining (Fig. 1). As expected, the TBP-PrA, which binds IgG with high affinity because of the PrA moiety, was eluted at high concentrations of MgCl₂. A few faint bands were visible in every fraction suggesting that they may be nonspecific contaminants or keratin. A band of ∼116 kD, visible mainly in the 0.25- and 1-M MgCl₂ eluates, was in the expected size range for an associated Kap (Pemberton et al., 1998; Wozniak et al., 1998). The band was excised and analyzed by MS. The band represented a protein encoded by the uncharacterized open reading frame YGL241W (also called Hrc1004). YGL241W is predicted to encode a protein of 1,004 amino acids with an expected molecular mass of 113.9 kD. Comparison of this protein sequence with the database showed strongest homology with Kap family members (15–17% identity), suggesting that this protein was, indeed, a member of the Kap superfamily (Görlich et al., 1997; Pemberton et al., 1998; Wozniak et al., 1998). Consistent with its role in nuclear transport (described below) and the standard nomenclature for other yeast Kaps, we will refer to this protein as Kap114p.

To verify that the interaction of Kap114p with TBP was specific, we tagged Kap114p as before with PrA (Kap114-PrA). Using cytosol prepared from the Kap114-PrA strain we purified by affinity chromatography those proteins that interacted with Kap114p. Kap114-PrA eluted at the high concentrations of MgCl₂, whereas several bands, including one of ∼27 kD, eluted at 0.25–1 M MgCl₂ (Fig. 2 a). The 27-kD band was excised and subjected to MS analysis and peptide sequencing which showed that this protein was indeed TBP. These data confirm that Kap114p and TBP interact in the cytosol, suggesting that TBP might represent an import substrate for Kap114p. Other bands detected in this experiment may also include Kap114p interacting proteins and, hence, potential import substrates for Kap114p.

To determine whether the interaction between Kap114p and TBP was direct, a complex was formed between bacterially expressed, purified GST-Kap114 and bacterially expressed, purified His₆-tagged TBP (Fig. 2 b). In control experiments very little TBP could be detected bound to glutathione-Sepharose alone (Fig. 2 b, lane 1) or to GST bound–Sepharose (Fig. 2 b, lane 2). However, a much greater amount of TBP bound to GST-Kap114 bound– Sepharose, suggesting that the binding was due to the presence of Kap114p (Fig. 2 b, lanes 3 and 4). TBP binding was also seen to increase with increasing amounts of GST-Kap114 immobilized on the Sepharose (Fig. 2 b, lanes 3 and 4). These results suggested that TBP and Kap114p interacted directly.

Using the Kap114-PrA–tagged strain it was possible to determine the localization of the Kap114p by visualization of the PrA tag. The Kap114-PrA was localized to the cytoplasm and nucleus, which would be consistent with its function as a nucleocytoplasmic shuttling protein (Fig. 3). To determine whether the gene encoding Kap114p was essential for cell growth the gene was replaced by genomic integration with HIS3. Viable haploid HIS3 yeast strains deleted for KAP114 (Δkap114) were obtained (data not shown), suggesting that the KAP114 gene was not essential for cell growth. Deletion strains were also grown on rich medium at 37°C and 17°C, however, no growth defect was detected under these conditions (data not shown).

To ascertain whether TBP was imported into the nucleus by Kap114p in vivo, the localization of TBP-PrA in the KAP114 deletion strain (Δkap114) was determined. As expected TBP-PrA appeared nuclear in wild-type cells, whereas in Δkap114 cells a large proportion of the TBP-PrA was now localized to the cytoplasm (Fig. 4 a). These results were confirmed using an antibody to detect endogenous TBP in the Δkap114 strain demonstrating that the mislocalization was not dependent on the PrA moiety (data not shown). Thus the nuclear import of TBP appears to be partially abrogated in the absence of Kap114p, confirming that Kap114p imports TBP in vivo. The localization of Lhp1p-GFP and a human La–GFP reporter, which have been previously shown to be substrates for other Kap-mediated import pathways (Rosenblum et al., 1998), were also analyzed in the Δkap114 strain. These proteins were correctly localized to the nucleus in this strain (Fig. 4 b), suggesting that we were not observing a general defect in nuclear transport and the effect on TBP localization was specific to deletion of KAP114.

As strains deleted for KAP114 appeared to have no growth defect under the described conditions and some of the TBP was nuclear in the absence of Kap114p (Fig. 4 a, as evidenced by the nuclear foci visible in some cells in the Δkap114 strain), it seemed likely that a fraction of the TBP was imported into the nucleus via an alternative pathway. Cytosol was made from the TBP-PrA/Δkap114 strain to determine which Kaps interacted with TBP in the absence of Kap114p. Coomassie blue staining of TBP-PrA–copurifying proteins revealed one in the correct molecular weight range for a Kap (Fig. 5 a). MS analysis revealed that this Kap was Kap123p, which has been previously shown to be involved in the import of ribosomal proteins (Rout et al., 1997; Schlenstedt et al., 1997). Kap123p is the most abundant Kap and is not encoded by an essential gene, although deletion strains grow very slowly at 37°C (Rout et al., 1997; Pemberton, L.F., unpublished data). TBP was PrA-tagged in the Δkap123 strain and the localization of TBP-PrA was determined by immunofluorescence. TBP-PrA did not appear to be mislocalized in this strain at any temperature (Fig. 6). As Kap114p was still present in this strain we made a double deletion of KAP114 and KAP123 (to generate Δkap114/Δkap123), to determine whether they were synthetically lethal. This strain was also viable and phenotypically appeared similar to the single mutants, in that TBP was mislocalized to a similar degree as in Δkap114 cells and that the cells grew poorly at 37°C as observed in the Δkap123 strain (Fig. 6).

These results suggested that the essential protein TBP may in fact have several routes into the nucleus and to find further import pathways for TBP, TBP-PrA–copurifying proteins were isolated from the Δkap114/Δkap123 strain as before. A very faint band in the correct molecular weight range for a Kap was detected by Coomassie blue staining of the proteins after separation by SDS-PAGE (data not shown), however, there was insufficient material for MS analysis. Another Kap, Kap121p, has been previously shown to have overlapping functions with Kap123p and to compensate for a lack of Kap123p in the import of ribosomal proteins (Rout et al., 1997; Schlenstedt et al., 1997). We determined whether the TBP-PrA–interacting band of the Δkap114/Δkap123 strain represented Kap121p by probing the eluted fractions, after transfer to nitrocellulose, with a polyclonal antibody to Kap121p (Marelli et al., 1998). This analysis showed that this band indeed contained Kap121p (Fig. 5 b). Kap121p is encoded by an essential gene and a temperature sensitive strain, pse1-1, has been described (Seedorf and Silver, 1997). A pse1-1 strain expressing TBP-PrA was grown at the restrictive temperature for 3 h to determine the effect on TBP localization. After this treatment, TBP-PrA was still localized to the nucleus (Fig. 6). This suggests that although Kap121p and Kap123p may play a role in the import of TBP, this role is likely to be redundant with Kap114p. Interestingly, Kap121p could also be detected by immunoblot in TBP-PrA isolation experiments in the presence of Kap114p (data not shown). These data suggest that Kap114p mediates the bulk of TBP nuclear import, as evidenced by the fact that only a kap114 deletion leads to mislocalization of TBP, even in the presence of an otherwise wild-type complement of Kaps.

In the nucleus, the dissociation of Kaps from their cognate import substrates is thought to be achieved by RanGTP and not RanGDP (Rexach and Blobel, 1995). To determine whether this is also the case for the Kap114p–TBP complex, we incubated the IgG-Sepharose bound Kap114-PrA-TBP complex, isolated from yeast cytosol, with RanGTP or RanGDP. The material eluted after the incubation with Ran was collected (constituting the unbound fraction) and the IgG-Sepharose bound material was eluted with 1 M MgCl₂ to dissociate the remaining TBP, followed by 4.5 M MgCl₂ to dissociate the Kap114-PrA (both fractions constituting the bound fraction). After separation by SDS-PAGE and transfer to nitrocellulose, TBP and Kap114-PrA were detected by immunoblotting with an anti-TBP polyclonal antibody. This antibody not only recognizes TBP but also Kap114-PrA, by virtue of its PrA tag. TBP was not eluted at high (9.5 μM, Fig. 7 a) or low concentrations of RanGDP (data not shown). A high concentration of RanGTP (9.5 μM) led to dissociation of more than half the TBP from the complex (Fig. 7 a). However at lower concentrations of RanGTP (480 nM) only a small amount of TBP was dissociated (Fig. 7 a). These results suggested that RanGTP rather than RanGDP was required to effect dissociation of TBP from Kap114p and the extent of dissociation depended on the concentration of RanGTP.

Many previously published Kap import substrate dissociation experiments have been carried out at RanGTP concentrations in the 5–10-μM range (Schlenstedt et al., 1997; Albertini et al., 1998; Jäkel and Görlich, 1998; Kaffman et al., 1998b; Senger et al., 1998) and even at these concentrations many of the dissociation experiments appeared to be fairly inefficient. The concentration of unbound RanGTP in the nucleus is unknown, however, a number of RanGTP-binding proteins such as the 14 Kaps, RanBP1, and the RanGEF, present in the nucleus, will compete for binding to RanGTP. Additional factors might, therefore, be necessary to stimulate RanGTP-dependent dissociation of import substrates from their cognate Kaps. To investigate this possibility we reduced the RanGTP in our experiments to a level where we saw virtually no dissociation of TBP from Kap114p. As TBP functions by binding the TATA box at the core promoter, we first determined whether an 18-bp ds oligonucleotide containing the CYC1 TATA box might stimulate the dissociation of TBP from Kap114p. This particular ds oligonucleotide was also used in the cocrystallization of TBP bound to DNA (Y. Kim et al., 1993; Geiger et al., 1996). Isolated Kap114-PrA-TBP complex was incubated with 380 nM RanGTP, and very little dissociation of TBP from the Kap was observed (Fig. 7 b, lanes 5 and 6). In these experiments the dissociated TBP fraction was collected and the bound fraction containing Kap114-PrA and bound TBP was eluted as one fraction with 4.5 M MgCl₂. Surprisingly, the addition of increasing concentrations of TATA-containing DNA, in the presence of RanGTP, led to the dissociation of most of the TBP from Kap114p (Fig. 7 b, lanes 9–12). Even with a low concentration of DNA (5 ng/μl or 430 nM) at least half the TBP was released from Kap114p (Fig. 7 b, lanes 7 and 8). In contrast, when the complex was incubated with a high concentration of DNA but no RanGTP, no dissociation was observed, demonstrating that DNA alone could not release TBP from Kap114p (Fig. 7 b, lanes 3 and 4). The DNA-mediated dissociation was dependent on RanGTP rather than RanGDP, as high concentrations of RanGDP (4.8 μM) and DNA (10 ng/μl) together caused no dissociation (Fig. 7 b, lanes 1 and 2). Using biotinylated ds TATA oligonucleotides captured on streptavidin beads, it was also possible to show that the released TBP was bound to DNA (data not shown). These experiments show that at this concentration of RanGTP, TBP is released from Kap114p only in the presence of DNA and suggests a possible mechanism for the specific release of TBP at its DNA target.

To determine whether this effect was specific for ds TATA-containing DNA, several controls were carried out using mutant TATA-containing ds DNA, single stranded DNA, and total yeast RNA in the presence of 290 nM RanGTP (Fig. 8 a). Total yeast RNA did not stimulate dissociation, suggesting that we were not observing a nonspecific effect of large negatively charged molecules (Fig. 8 a, lanes 11 and 12). We saw a spectrum of activity using other DNA controls, some double and single stranded templates led to low levels of dissociation (data not shown), suggesting that they would not stimulate the release of TBP from the Kap. Other DNA controls including the ds oligonucleotide, M1, that was identical to the TATA oligonucleotide but included two mutations in the TATA box (Fig. 8 a, lanes 9 and 10) stimulated dissociation quite efficiently. Single stranded TATA DNA (Fig. 8 a, lanes 7 and 8) and single stranded M1 DNA (Fig. 8 a, lanes 9 and 10) stimulated some dissociation of the TBP from the Kap. However, most of the controls did not stimulate dissociation as well as the original TATA-containing ds oligonucleotides (compare all with Fig. 8 a, lanes 1 and 2).

As DNA stimulated the dissociation of TBP in the presence of RanGTP, it was possible that a TBP-interacting protein could also function in this way. We examined TFIIA as it has been shown to bind TBP alone and complexed with DNA (Ranish et al., 1992; Geiger et al., 1996; Tan et al., 1996). At the RanGTP concentrations (290 nM) used in this experiment, almost no dissociation of TBP from Kap114p was observed (Fig. 8 b, lanes 1 and 2) and the addition of ds TATA-containing DNA (2 ng/μl) with RanGTP resulted in the dissociation of more than half of the TBP (Fig. 8 b, lanes 5 and 6). The addition of TFIIA (65 nM) together with RanGTP, however, did not result in the dissociation of any TBP (Fig. 8 b, lanes 3 and 4). At this concentration, analysis of amido black–stained blots suggested that TFIIA was in excess of TBP (data not shown). RanGTP-dependent dissociation was also not observed when a higher concentration of TFIIA (200 nM) was used (data not shown). This suggested that TFIIA could not stimulate the RanGTP-dependent dissociation of TBP from the Kap in the absence of ds DNA, raising the possibility that Kap114p may prevent TBP from associating with TFIIA before it has bound DNA. When TFIIA was added together with RanGTP and DNA, as expected most of the TBP was dissociated from Kap114p in a similar manner to that seen with RanGTP and DNA alone (Fig. 8 b, lanes 7 and 8). However, the addition of TFIIA appeared to stimulate the dissociation seen with RanGTP and DNA. This may be due to the fact that TFIIA stabilizes the TBP–DNA interaction (Imbalzano et al., 1994).

Here we show that the major import pathway for TBP is mediated by the newly characterized Kap, Kap114p. TBP is also imported into the nucleus by Kap114p-independent pathways and we suggest that these pathways may be mediated by Kap123p and Kap121p. As expected for a Kap family member, the dissociation of TBP and Kap114p is RanGTP-dependent. However, we show here that this dissociation is greatly stimulated by the addition of TATA-containing ds DNA. This suggests a mechanism for the targeted dissociation of TBP at the promoter of genes to be transcribed.

The isolation of a cytosolic Kap114p–TBP complex suggested that Kap114p mediated the import of TBP into the nucleus and analysis of the localization of TBP in the absence of Kap114p confirmed this. Our ability to reconstitute the Kap114p–TBP complex using recombinant proteins suggested that they interact directly. In addition, the Kap114p–TBP complex, isolated from yeast cytosol, was resistant to incubation with S7 nuclease and DNase, suggesting that the complex was not mediated by nucleic acid (data not shown). Strains lacking Kap114p are viable, in contrast to strains lacking TBP, which is encoded by an essential gene. In common with the Pdr6p/Kap122p–mediated import pathway for TFIIA (Titov, A., and G. Blobel, manuscript submitted for publication) and the import of ribosomal proteins by Kap123p proteins (Rout et al., 1997; Schlenstedt et al., 1997), this represents another example of a nonessential Kap importing an essential substrate. This led us to search for additional pathways into the nucleus for TBP. The isolation of TBP-interacting Kaps, Kap123p and Kap121p, suggested that they too might mediate import of TBP. In both a Kap123p deletion strain and in a Kap121p temperature sensitive strain, however, we saw no mislocalization of TBP. The fact that we could detect no other Kaps in a complex with TBP-PrA, and that these two Kaps were not isolated in analogous coprecipitation experiments with Lhp1p-PrA (Rosenblum, S., and G. Blobel, unpublished data) were indicative of a potential role for Kap123p and Kap121p in TBP import. The demonstration of direct binding and of an effect on nuclear import will be needed, however, to conclusively demonstrate the role of Kap123p and Kap121p in TBP import. It is likely that Kap114p mediates the bulk of TBP nuclear import, as evidenced by the fact that among the TBP-interacting Kaps, only a kap114 deletion leads to mislocalization of TBP, even in the presence of an otherwise wild-type complement of Kaps.

We do not know if these are the only Kaps responsible for the import of TBP into the nucleus. Kap121p is encoded by an essential gene (Chow et al., 1992), and we were, therefore, not able to test for synthetic lethality by deletion of all three genes in one strain. It is possible that there are additional Kap-mediated pathways for the import of TBP as Kap121p was only detected by immunoblotting, indicating that other Kaps could be interacting below the limits of detection by Coomassie blue staining. In strains lacking both Kap114p and Kap123p only very small amounts of TBP–copurifying Kap were detected by our assay. As some TBP was being imported into the nucleus to support normal transcription and cell growth (discussed below), it is possible that we are not able to detect very nonabundant Kaps in this way. The fact that we could not detect them may suggest that these Kaps interact with their substrates with lower affinity or the interaction is less stable. It is also possible that some TBP is entering the nucleus by a Kap-independent pathway or by diffusion. Diffusion is unlikely, however, as TBP is believed to dimerize in the cytoplasm, increasing its molecular weight to beyond the exclusion limit of the NPC (Coleman et al., 1995).

Like mammalian ribosomal proteins, this is an example of a single substrate that appears to have multiple secondary pathways into the nucleus (Jakel and Görlich, 1998). It may prove to be a general phenomenon, whereas one Kap plays a major role in the import of a given substrate, this function is redundant with several other Kaps. Such a multiply redundant, hierarchical import pathway also suggests that the coordinate regulation of the Kap family is likely to be very complex. This may help explain why there are so many nonessential or redundant Kaps. Of the fourteen β Kaps present in the yeast genome, only four are encoded by essential genes (Kap95p, Kap121p, Crm1p, and Cse1p), whereas deletion of the others lead to varying effects on cell growth (Chow et al., 1992; Toda et al., 1992; Shen et al., 1993; Xiao et al., 1993; Enenkel et al., 1995; Aitchison et al., 1996; Akada et al., 1996; Pemberton et al., 1997; Rosenblum et al., 1997; Rout et al., 1997; Schlenstedt et al., 1997; Seedorf and Silver, 1997; Albertini et al., 1998; Ferrigno et al., 1998; Chaves, S., and G. Blobel, unpublished data; Titov, A., and G. Blobel, manuscript submitted for publication). It is possible that the essential Kaps transport essential proteins, or they may play a more general role and can backup many pathways. Alternatively, they may carry out the bulk of nuclear transport, although their relative abundance would suggest that this is not the case (Rout et al., 1997). Kap121p has been shown to functionally overlap with Kap123p and can participate in the import of ribosomal proteins (Rout et al., 1997; Schlenstedt et al., 1997). If Kap123p and Kap121p also play a role in the import of TBP, it raises the possibility that the Kap family may be divided into functional subgroups. Future identification of additional substrates for these three Kaps will determine whether this subgroup coordinates the import of other proteins.

Kap123p copurified with TBP-PrA in the absence of Kap114p, whereas Kap121p was detected in small quantities in both the presence and absence of Kap114p. As Kap114p would be present in a wild-type yeast cell, it is possible that some TBP import is constitutively mediated by other Kaps or that other Kaps participate in TBP import only under specific conditions. These conditions could be dictated by the available carbon source, stress, or phase of the cell cycle amongst others. Several Kaps have been shown to have more than one substrate, and the fact that there are many more nuclear proteins than Kaps suggests that this will be a general phenomenon (Aitchison et al., 1996; Rosenblum et al., 1997; Rout et al., 1997; Schlenstedt et al., 1997; Albertini et al., 1998; Ferrigno et al., 1998; Kaffman et al., 1998b). We observed several Kap114- PrA copurifying bands in addition to TBP, suggesting that it also may have other import substrates and future characterization should determine the identity of these proteins.

The Kaps may be coordinated in a complex network of overlapping substrate specificities, raising the problem of how several Kaps can recognize the same substrates while distinguishing between others. The recognition of a substrate by a Kap is determined by the NLS contained in the substrate. The crystal structure of TBP has been elucidated both alone and complexed with DNA, TFIIA, and TFIIB (Nikolov et al., 1992, 1995, 1996; J.L. Kim et al., 1993; Y. Kim et al., 1993; Geiger et al., 1996; Tan et al., 1996). This has allowed the precise binding sites of these molecules within TBP to be determined. It is not yet known whether the TBP NLS recognized by Kap114p overlaps with one of these sites, such as the DNA-binding domain, or whether the Kap recognizes another domain of TBP. It is possible that other Kaps, for example Kap121p and Kap123p, may recognize the same NLS in TBP as Kap114p. However, comparison of TBP with known Kap123p and Kap121p substrates has not yet revealed any similarities at the amino acid level that may represent a consensus NLS.

To function in transcription, TBP in the nucleus must be first dissociated from Kap114p and then interact with DNA and other components of the transcriptional machinery. All the Kaps are believed to bind RanGTP with their homologous amino-terminal domain as shown in the structure of mammalian Kap β2 complexed with RanGTP (Chook and Blobel, 1999). As expected RanGTP, which is believed to be the predominant form of Ran in the nucleus, is necessary for the dissociation of TBP from Kap114p. RanGDP does not appear to be able to function in this process. However, we observed in vitro that not all the TBP was dissociated from Kap114p, even at high (10 μM) RanGTP concentrations. Previously published Kap import substrate dissociations have been carried out at RanGTP concentrations in the 5–10-μM range, however, many of these dissociation experiments appeared similarly inefficient (Schlenstedt et al., 1997; Albertini et al., 1998; Jakel and Gorlich, 1998; Kaffman et al., 1998b; Senger et al., 1998). Although the total Ran concentration in the whole cell has been suggested to be as high as 5–10 μM, it is difficult to predict the concentration of unbound RanGTP in the nucleus (Bischoff and Ponstingl, 1991; Gygi et al., 1999). As there are at least 16 RanGTP binding partners present in the nucleus, which bind RanGTP with dissociation constants in the micromolar to nanomolar range, the concentration of unbound RanGTP may be much lower than this (Görlich et al., 1997; Kutay et al., 1997; Schlenstedt et al., 1997; Senger et al., 1998; Floer and Blobel, 1999). At lower (190–380 nM) concentrations of RanGTP, we observed almost no dissociation of TBP from Kap114p that led us to search for stimulatory factors. We identified ds TATA–containing DNA as one such factor. As nonnucleosomal ds TATA–containing DNA would be the available form of DNA at genes about to undergo transcription, this suggests a mechanism whereby TBP may only be dissociated from the Kap at its point of function.

The dissociation caused by TATA-containing DNA was most efficient, however, other DNAs could also stimulate dissociation of TBP from Kap114p to varying extents. This lack of specificity in vitro may be because TBP can bind all DNA with some affinity (Chen and Struhl, 1988; Coleman and Pugh, 1995). In vivo, other GTFs such as TFIIA and TFIIB would be present and there is evidence to suggest that they increase the specificity of TBP for TATA sequences (Imbalzano et al., 1994). Alternatively, in vivo the initial dissociation at the promoter may be TATA-independent; once dissociated, TBP could then move to the TATA box itself to form a high affinity interaction (Coleman and Pugh, 1995). In this way, the dissociation would still most likely occur at the promoter as only a small proportion of DNA in the nucleus is actually available for TBP binding. Most DNA is incorporated into nucleosomes, and naked, nonnucleosomal DNA is only likely to be found at promoters of genes that are about to be transcribed or after the replication fork (Sogo et al., 1986).

The TBP-interacting protein TFIIA did not stimulate the dissociation of TBP in the absence of TATA-containing DNA. It is possible that association of TBP with Kap114p prevents TBP from prematurely interacting with other PIC components before reaching the promoter. TFIIA did stimulate dissociation in the presence of DNA, which may be due to its ability to stabilize TBP–TATA interactions (Imbalzano et al., 1994). It is not yet known whether other TBP-interacting proteins such as TBP-associated factors, TFIIB, negative regulators of TBP, and components of the SAGA complex can stimulate the dissociation of TBP from Kap114p, either with or without DNA. Future experiments will be needed to determine the sequence of events of PIC complex formation in vivo. There is believed to be enough TBP present in the cell to participate in transcription from all 6,000 plus promoters at one time (Lee and Young, 1998). There are at least three mechanisms by which the availability of TBP and its ability to function at specific promoters is negatively regulated. These include the proteins or protein complexes Mot1p, NC2, and Nots (Lee and Young, 1998). It is likely that much of the cellular TBP is negatively regulated by these factors, which may act by sequestering TBP (Lee and Young, 1998). Our in vitro dissociation experiments would argue that the Kap participates in this regulation, by only releasing the TBP it has imported, at its point of function.

By reducing the amount of RanGTP we observed a dissociation requirement for further factors that may prove to be a general phenomenon in nuclear transport. Previous studies using Kap111p/Mtr10p have shown that the addition of total cellular RNA also stimulated the RanGTP-dependent dissociation of the mRNA-binding protein Npl3p, from Kap111p (Senger et al., 1998). The DNA-stimulated, RanGTP-mediated targeted dissociation of TBP from Kap114p we observe in vitro suggests that in vivo dissociation might be effected at much lower concentrations of RanGTP than previously thought. For targeted dissociation to occur, Kap114p would be predicted to travel with TBP from the nucleoplasmic face of the NPC to the promoter, and this raises the possibility that this Kap may also be functioning in intranuclear delivery of TBP. Very little is known about the movement of molecules within the nuclear interior and it is possible that Kaps function not only in nucleocytoplasmic transport but may also mediate the efficient transport of proteins within the nucleus.

It was surprising that although much of the TBP was mislocalized in the Δkap114 deletion strain, we could detect no growth defect. It is possible that the small amount of TBP that was imported into the nucleus was sufficient to support transcription. Whether this small amount of TBP is now no longer subject to the negative regulation discussed above, and is all functioning in transcription, remains to be determined. In the absence of Kap114p, TBP imported by other Kaps such as Kap121p and Kap123p, may also be dissociated in a DNA-stimulated, RanGTP-dependent fashion at its point of function and it will be interesting to determine whether this will be a generalized theme in nuclear transport. The experiments presented here for TBP and previously published data for TFIIA suggest that at least two components of the PIC are imported by separate pathways (Titov, A., and G. Blobel, manuscript submitted for publication). It remains unclear how other components of the PIC are imported into the nucleus and whether their import by separate pathways plays a role in their regulation.

In summary, we have identified a new pathway for protein import into the nucleus, mediated by the novel transport factor Kap114p. This pathway constitutes the major import pathway for TBP, however, other Kaps also appear to participate in TBP import. The import of TBP is one of the first examples in yeast of a substrate having multiply redundant, hierarchical import pathways and suggests that the Kaps may be coordinated in a complex network of overlapping substrate specificities. We also present data suggesting that TBP is only released by the Kap at its point of function, raising the possibility that the Kaps may play a role in the intranuclear targeting of their substrates as well as nucleocytoplasmic transport.

We thank Monique Floer for purified Ran, P. Anthony Weil for TBP antibody and purified His₆-TBP, Rick Wozniak (University of Alberta, Edmonton, Canada) for Kap121p antibody, Pam Silver (Dana-Farber, Boston, MA) for the pse1-1 strain, Steven Hahn for purified TFIIA, Mike Rout (The Rockefeller University) for the Δkap123 strain, Farzin Gharahdaghi and Joe Fernandez of The Rockefeller University Protein DNA Technology Center for MS and peptide sequencing and Evette Ellison for technical support.

DAPI

4′,6-diamidino-2-phenylindole

ds

double stranded

GTF

general transcription factor

Kap

karyopherin

MS

mass spectrometry

NLS

nuclear localization sequence

NPC

nuclear pore complex

PIC

preinitiation complex

PrA

protein A

RNAP

RNA polymerase

TBP

TATA-binding protein

TBP-PrA

PrA-tagged copy of TBP