The Karyopherin Kap122p/Pdr6p Imports Both Subunits of the Transcription Factor Iia into the Nucleus

All strains used were derived from Saccharomyces cerevisiae wild-type DF5α (Finley et al. 1987) and its derivative kap122Δ (MATα, lys2-801, leu2-3, 112, ura3-52, his3-200, trp1-1 (am), pdr6::URA3). Yeast strains were grown at 30°C in yeast extract/peptone/glucose (YPD); all yeast manipulation was performed according to described protocols (Sherman et al. 1986).

For deletion of PDR6 in wild-type strain DF5α, the HIS3 gene was used as a selective marker to replace the PDR6 open reading frame by integrative transformation in a haploid DF5α strain. HIS3 replacement cassette was generated by PCR amplification of markers from pRS 306 with primers that contained 60 nucleotides flanking the PDR6/KAP122 open reading frame from 5′ and 3′ ends (Aitchison et al. 1995). HIS3 marker was switched to URA3 marker by recombination (plasmids were gifts of Dr. F.R. Cross, Rockefeller University). Deletion of genes was confirmed by PCR on total yeast DNA with internal primers.

Carboxy-terminal genomic KAP122-PrA fusion constructs were created by integrative transformation of PCR-amplified constructs with four and a half IgG-binding domains of protein A immediately upstream of PDR6/KAP122 stop codon, followed by a not-in-frame HIS5 (Schizosaccharomyces pombe) selection marker (donated by Dr. R. Beckmann, Rockefeller University). TOA1 and TOA2 protein A tagging was similar to that of KAP122. Primers used for PCR amplification of protein A/HIS5 cassette contained 60 nucleotides directly upstream of the stop codon of the relevant gene for the 5′ primer and 60 nucleotides 150 bp downstream of the stop codon for the 3′ primer. Haploid yeast cells were transformed by electroporation. This resulted in expression of the chimeric protein A fusion construct under the control of the endogenous promoter.

Fractionation and immunoisolation of protein A fusion proteins were performed as described (Aitchison et al. 1996). For a typical isolation, 500 ml of postribosomal supernatant (cytosol) was prepared from a 6-liter YPD culture with a density of 1.7 at A₆₀₀. Cytosol was incubated with 200 μl of rabbit IgG-Sepharose beads at 4°C overnight. After washing with transport buffer (TB: 20 mM Hepes-KOH, pH 7.5, 110 mM KOAc, 2 mM MgCl₂, 1 mM DTT, 0.1% Tween 20), bound proteins were eluted with a step gradient of MgCl₂ from 50 to 4,500 mM. Proteins were precipitated, resolved by SDS-PAGE on a 4–20% acrylamide gel (Novex), and stained with Coomassie blue. Proteins of interest were excised and prepared for MALDI-TOF spectrometry and/or sequencing.

Yeast cells were fixed in 3.7% formaldehyde for 15 min and cell walls were digested. Indirect immunofluorescence was carried out according to published protocols (Wente et al. 1992). Protein A moieties of fusion proteins were detected with rabbit IgG that had been preadsorbed to wild-type yeast spheroplasts; Nab2p was detected by rabbit polyclonal antiserum to Nab2p (Aitchison et al. 1996), Npl3p was detected by mouse monoclonal antiserum to Npl3p (Wilson et al. 1994), the appropriate Cy3-conjugated anti–rabbit or anti–mouse IgG was used for visualization. Nuclei were visualized with the DNA binding stain 4′,6-diamidino-2-phenylindole (DAPI).

A Nup1p fragment containing the FXFG repeat region (amino acids 432–816) and a Nup2p fragment containing the FXFG repeat region (amino acid 186–561) were expressed as glutathione S-transferase (GST) fusion proteins as described (Enenkel et al. 1995; Rexach and Blobel 1995). Proteins of bacterial lysates were separated by SDS-PAGE and transferred to nitrocellulose. Overlay assays were performed as described (Aitchison et al. 1996). Yeast cytosol from the Kap122p-PrA–expressing strain was diluted 1:1 with TB-5% milk and incubated on the blot overnight at 4°C. Bound proteins were detected with rabbit antibodies to mouse IgG and HRP-conjugated donkey anti–rabbit antibodies and enhanced chemiluminescence.

TOA1, TOA2, and KAP122 open reading frames were amplified from yeast genomic DNA by PCR using synthetic oligonucleotide primers with incorporated restriction sites for subcloning into relevant Escherichia coli expression vectors.

TOA1 was subcloned into NdeI and BamHI sites of kanamycin resistance–conferring pET 28a (Novagen, Inc.) to allow expression of Toa1p with an amino-terminal His tag. TOA2 was subcloned into NcoI-BamHI sites of ampicillin resistance–conferring pET 19b (Novagen, Inc.) to allow coexpression of Toa2p with Toa1p.

E. coli strain BL21 gold (Stratagene) was cotransformed with both the His₆-Toa1p– and the Toa2p-expressing plasmids, and E. coli cells coexpressing both subunits of TFIIA were grown at 37°C in LB medium containing 60 μg/ml kanamycin and 100 μg/ml ampicillin to a density of 0.7 at A₆₀₀. Protein expression was induced by 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 10 h at 17°C. E. coli were lysed in TB with added protease inhibitor cocktails (Boehringer Mannheim) using the French pressure cell. The bacterial lysate was incubated with Talon™ resin (CLONTECH Laboratories) for 30 min at room temperature (RT). Talon™ beads were washed with lysis buffer (TB-0.1% Tween 20) and TB buffer containing 2 mM ATP, 10 mM Mg(OAc)₂, and 10 mM imidazole. For protein binding assays, the His₆-Toa1p/Toa2p complex was used either still bound to Talon™ beads or was eluted with TB containing 80 mM imidazole.

KAP122 was subcloned into BamHI-EcoRI sites of the derivative of pGeX4T3 (Pharmacia Biotech) with additional Tev protease cleavage site (Chook, Y.M., and G. Blobel, 1999), to allow expression of Kap122p with a potentially cleavable amino-terminal GST tag. Kap122p-expressing E. coli cells were grown in LB medium containing 100 μg/ml ampicillin to a density of 0.7 at A₆₀₀ and GST-Kap122p expression was induced by 1 mM IPTG for 3 h at 30°C. GST-Kap122p fusion protein was purified from bacterial lysates according to the manufacturer's protocol (Pharmacia Biotech).

Binding assays were performed in TB with the addition of 10% glycerol and 0.1% casaminoacids (Invitrogen Corp.) to prevent nonspecific interactions. 10 μl of glutathione-Sepharose beads containing either 2 μg immobilized GST-Kap122p or 2 μg immobilized GST were incubated with 5 μg of purified His₆-Toa1p/Toa2p complex in a total volume of 100 μl for 1 h at RT. Beads were collected by centrifugation at 1,000 g for 30 s, washed five times by mixing with 1 ml of TB followed by sedimentation, and were resuspended in 15 μl of sample buffer. Proteins in one half of each sample were resolved by SDS-PAGE on a 4–20% acrylamide gel (Novex, Inc.) and stained with Coomassie blue.

200 μl of GST-Kap122p E. coli lysate was incubated with 10 μl Talon™ resin at RT for 1 h to deplete the lysate of endogenous bacterial proteins that interact nonspecifically with Talon™ beads. Beads were sedimented at 1,000 g for 30 s, and the lysate was passed through a Micro Bio-Spin^® chromatography column (Bio-Rad Laboratories) to exclude any remaining Talon™ beads. 50 μl of TB containing 50% glycerol and 0.5% casaminoacids was added to 200 μl of precleared lysate and the mixture was incubated for 1 h at RT with 2 μg of His₆-Toa1p/Toa2p complex immobilized on 10 μl of Talon™ beads or with 10 μl of empty Talon™ beads as a control. At the end of incubation, beads were sedimented and washed five times with 1 ml of TB. Talon™ beads with bound proteins were resuspended in sample buffer and half of each sample was resolved by SDS-PAGE on a 4–20% acrylamide gel (Novex, Inc.) and stained with Coomassie blue.

Recombinant Saccharomyces cerevisiae Ran was prepared and loaded with GDP or GTP as described (Floer and Blobel 1996). Protein A–tagged Toa2p with bound Kap122p was immunoisolated from a postribosomal supernatant as described above. After washing, IgG-Sepharose beads were resuspended in TB and divided into several equal fractions. These fractions were incubated either with TB alone, or with 4 μM RanGDP or RanGTP for 60 min at room temperature. 1 mM GTP was included, except when using RanGDP. Beads were transferred to a column, and the drained liquid, together with a subsequent 100-μl TB wash, was collected; this constituted the released material. After washing columns with TB, remaining bound proteins were eluted with 1,000 and 4,500 mM MgCl₂ step gradient. Fractions eluted at 1,000 and 4,500 mM were combined and comprised the bound material. Proteins were precipitated, resolved by SDS-PAGE, and stained with Coomassie blue.

15 μl of Talon™ beads with 3 μg of immobilized His₆-Toa1p/Toa2p complex were washed five times with 1 ml of TB, incubated with precleared GST-Kap122p lysate, as described above, to obtain a GST-Kap122p/His₆-Toa1p/Toa2p complex. Three equal aliquots of beads were incubated with either 4 μM RanGTP, 4 μM RanGDP, or TB alone for 2 h at RT in a final volume of 40 μl. 1 mM GTP was included, except when using RanGDP. Beads were sedimented by centrifugation and 30 μl of supernatant were collected and passed through Micro Bio-Spin^® chromatography column (Bio-Rad Laboratories) to exclude any contaminating Talon™ beads. The drained liquid represented the released material. Beads were washed five times with 1 ml of TB and resuspended in 15 μl of sample buffer. Proteins in one half of each bound and released sample were resolved by SDS-PAGE on 4–20% acrylamide gel (Novex, Inc.) and stained with Coomassie blue.

In all cases described here, we found that the genomic replacement of a gene by its protein A–tagged version did not result in any apparent changes in growth rates under the conditions used in our experiments and when compared with those of an otherwise isogenic strain. This indicated that PrA tagging did not apparently interfere with the function of the essential proteins that were tested here; however, this cannot be confirmed for the Kap122p-PrA strain because of the absence of a discernible phenotype of the KAP122 deletion strain.

Immunofluorescence microscopy of a KAP122-PrA haploid strain showed that Kap122p-PrA is located both in the cytoplasm and the nucleus (Fig. 1), which is consistent with the localization of other protein A–tagged Kaps and their function in shuttling between the nucleus and the cytoplasm.

Karyopherins have been previously shown to interact with peptide repeat–containing nucleoporins (Rexach and Blobel 1995; Aitchison et al. 1996; Fornerod et al. 1997; Rout et al. 1997; Pemberton et al. 1997; Rosenblum et al. 1997; Albertini et al. 1998; Marelli et al. 1998). To test whether Kap122p interacts with nucleoporins, we expressed FXFG repeat–containing fragments of the nucleoporins Nup1p and Nup2p in E. coli, separated proteins of bacterial lysates by SDS-PAGE, transferred them to nitrocellulose, and performed overlay assays with Kap122p-PrA cytosol. Kap122p-PrA interacted strongly and specifically both with Nup1p and Nup2p fragments (Fig. 2). This strongly suggests that Nup1p and Nup2p are among the nucleoporins that bind to Kap122p.

Kaps that function in protein import form stable complexes with their import substrates in the cytoplasm (Aitchison et al. 1996; Pemberton et al. 1997; Rosenblum et al. 1997; Rout et al. 1997; Albertini et al. 1998). To isolate such complexes, we prepared a postribosomal cytosol fraction from the KAP122-PrA haploid strain and incubated it with IgG-Sepharose. Bound proteins were eluted with a step gradient of MgCl₂, separated by SDS-PAGE, and stained with Coomassie blue (Fig. 3). The Kap-bound substrates usually elute at MgCl₂ concentration between 100–1,000 mM MgCl₂, whereas the PrA-tagged Kap elutes at 4.5 M MgCl₂ (Aitchison et al. 1996). Indeed, a major band with an apparent M_r of 150 kD, which likely represented Kap122-PrA, eluted at 4.5 M MgCl₂. The numerous minor bands in this fraction, as well as in fractions eluting at lower MgCl₂ concentrations, are likely degradation products of Kap122-PrA retaining their PrA moiety, as confirmed by immunoblotting with rabbit IgG (data not shown). A number of proteins were present primarily in the 100- and 250-mM MgCl₂ eluates and these are putative substrates for Kap122p. One of these bands migrating at ∼14 kD was excised, analyzed by mass spectrometry and peptide microsequencing after digestion with trypsin (Gharahdaghi et al. 1996; Fernandez et al. 1998), and thereby identified as the small subunit (Toa2p) of the general transcription factor IIA (TFIIA). These data suggested that Toa2p might be an import substrate for Kap122p.

To determine whether Toa2p is indeed a transport substrate for Kap122p, we examined the localization of Toa2p in wild-type (WT) and KAP122 deletion (kap122Δ) strains. There were no apparent differences in growth rates between kap122Δ and an isogenic wt strain in YPD medium at 30°C. For these experiments, TOA2 was genomically tagged with PrA in isogenic wt and kap122Δ strains. Immunofluorescence microscopy showed that in wild-type cells Toa2p-PrA was located primarily in the nucleus, whereas in the kap122Δ cells Toa2p-PrA was mislocalized largely to the cytoplasm (Fig. 4). The diminished nuclear localization of Toa2p-PrA in the kap122Δ strain is consistent with Kap122p representing the principal Kap for import of Toa2p.

Localization of several other substrates, whose transport is mediated by distinct Kaps, was not affected by KAP122 deletion. The nuclear import of Npl3p, mediated by Kap111p (Pemberton et al. 1997; Senger et al. 1998) and Nab2p, mediated by Kap104p (Aitchison et al. 1996), appeared similar in both wild-type and kap122Δ strains (Fig. 5), showing that deletion of KAP122 does not generally affect nucleocytoplasmic transport.

Yeast TFIIA has been shown to consist of two subunits: a small subunit (Toa2p, calculated M_r 13.5 kD) and a large subunit (Toa1p, calculated M_r 32 kD) that migrates as a 43-kD protein in SDS-PAGE (Hahn et al. 1989; Ranish and Hahn 1991; Ranish et al. 1992). Separately expressed Toa1p and Toa2p are insoluble and unable to complement transcription systems unless both subunits are renatured together (Ranish et al. 1992); this suggests that both subunits of TFIIA are unlikely to exist as separate entities in the cell (Ranish et al. 1992; Geiger et al. 1996). Crystallographic data of TFIIA bound to a DNA–TATA-binding protein (TBP) complex revealed that the two subunits are intertwined with each other to form a heterodimer, and that the heterodimer interacts with TBP and also with the phosphate backbone of DNA (Geiger et al. 1996; Tan et al. 1996). Hence, it is clear that in the nucleus Toa1p and Toa2p interact with the DNA–TBP complex as a heterodimer.

Our data above (Fig. 3) suggested that, in the cytoplasm, Toa2p existed in a complex with Kap122p-PrA and it was possible that Toa1p was part of this complex. As Toa2p and Toa1p bind as a dimer to the DNA–TBP in the nucleus, we investigated whether a Toa1p–Toa2p complex existed also in the cytoplasm. Therefore, we analyzed a postribosomal cytosol from a haploid TOA2-PrA strain by IgG-Sepharose affinity chromatography. A band migrating at ∼110 kD was eluted between 100 and 1,000 mM MgCl₂ (Fig. 6), and this band was confirmed to be Kap122p by mass spectrometric analysis. As expected, Toa2p-PrA eluted at 4.5 M MgCl₂. However, there was another major band, migrating at ∼45 kD, which coeluted with Toa2p-PrA (Fig. 6). By mass spectrometric analysis this band was identified as Toa1p. The elution of Toa1p at 4.5 M MgCl₂ suggested that Toa1p and Toa2p form a tight cytoplasmic complex that interacts with Kap122p.

To determine the cytoplasmic binding partners of Toa1p-PrA, we analyzed the postribosomal cytosol of a TOA1-PrA strain by IgG-Sepharose affinity chromatography. A band migrating at ∼110 kD and eluting between 100 and 1,000 mM MgCl₂ was identified by mass spectrometric analysis as Kap122p. Toa2p, identified by mass spectrometric analysis, coeluted with the Toa1p-PrA at 4.5 M MgCl₂ (Fig. 7), confirming the data in Fig. 6 that these two proteins form a tight cytoplasmic complex that interacts with Kap122p. Numerous degradation products of Toa1p-PrA, retaining their PrA moiety and identified by immunoblotting with rabbit IgG, were also visible in this fraction (data not shown). The apparent sensitivity of Toa1p to proteolysis could explain less than equimolar quantities of Toa1p that were isolated with either Kap122p-PrA (Fig. 3) or with Toa2p-PrA (Fig. 6, see also Fig. 10). Toa1p was similarly sensitive to proteolysis when coexpressed with Toa2p in E. coli (see Fig. 11 and Fig. 12). These smaller fragments of Toa1p that we observed when Toa1p was isolated from yeast or expressed in E. coli, may correspond to p55-related fragments observed when a human homologue of Toa1p was expressed in bacteria (DeJong and Roeder 1993).

As expected, immunofluorescence microscopy of Toa1p-PrA in a wt and a kap122Δ strain gave similar results to those obtained for Toa2p-PrA (Fig. 4): the primarily nuclear localization of Toa1p-PrA in the wt strain is diminished in the kap122Δ strain in favor of a diffuse cytoplasmic localization (Fig. 8).

Toa1p and Toa2p are essential for viability, whereas Kap122p is not. One solution to this apparent paradox is that one or several Kap(s) other than Kap122p can import these proteins into the nucleus in a kap122Δ strain. However, the immunofluorescence data of Fig. 4 and Fig. 8 suggested that if Toa1p–Toa2p were imported by alternative Kaps, this import would be less efficient than that mediated by Kap122p. Nevertheless, to search for alternative Kaps, a cytosol from a kap122Δ/TOA2-PrA strain was analyzed by IgG-Sepharose affinity chromatography (Fig. 9). As expected, Toa2p-PrA was eluted together with Toa1p (and its numerous degradation products) in the 4.5-M MgCl₂ fraction (Fig. 9). However, fractions eluted between 100 and 1,000 mM MgCl₂ did not show any visible bands in the 100-kD region and above, where Kaps migrate. Hence, if Kaps other than Kap122p can import the Toa1p–Toa2p, they might do so with lower efficiency than Kap122p and, therefore, are likely to be below the limits of detection of this assay.

Import complexes of substrate–Kap are dissociated by RanGTP (Rexach and Blobel 1995; Schlenstedt et al. 1997; Albertini et al. 1998; Jakel and Gorlich, 1998; Kaffman et al. 1998; Senger et al. 1998). To investigate whether this is also the case for the Toa1p/Toa2p/Kap122p complex, we used postribosomal cytosol from a TOA2-PrA strain to prepare an IgG-Sepharose–bound complex of Toa2p-PrA/Toa1p/Kap122p. This complex was incubated with either transport buffer alone, RanGDP, or RanGTP, and the material that was released from the IgG-Sepharose after completion of incubation was collected. Thereafter, the remaining IgG-Sepharose–bound proteins were eluted at 1.0 and 4.5 M MgCl ₂. Proteins were analyzed by SDS-PAGE and Coomassie blue staining (Fig. 10). Incubation with RanGTP clearly led to the release of most of the Kap122p (compare lanes 1–3). In contrast, incubation with RanGDP did not result in dissociation of Kap122p (lanes 4–6). These data indicated that the cytoplasmic Toa2p-PrA/Toa1p/Kap122p complex, like other import substrate–Kap complexes is sensitive to dissociation by RanGTP but not by RanGDP.

To determine whether Kap122p is able to interact directly with a Toa1p–Toa2p complex, we coexpressed both His₆-Toa1p and Toa2p in E. coli. The TFIIA subunits formed a soluble complex and were purified from bacterial lysate via the His₆ affinity tag at the amino terminus of Toa1p. Kap122p with an amino-terminal GST tag was also expressed in E. coli. We tested for binding between GST-Kap122p and the His₆-Toa1p/Toa2p complex when either was immobilized on GSH-Sepharose or Talon™ beads, respectively.

We found that purified and immobilized His₆-Toa1p/Toa2p complex bound GST-Kap122p present in a preincubated bacterial lysate depleted of its Talon™-binding proteins (Fig. 11, lane 1). Incubation of empty Talon™ resin with E. coli lysate containing GST-Kap122p did not result in any Kap122p binding (Fig. 11, lane 2). Likewise, GST-Kap122p, immobilized on GSH-Sepharose, bound the purified soluble His₆-Toa1p/Toa2p complex (Fig. 11, lane 3). A control using immobilized GST alone showed no binding of His₆-Toa1p/Toa2p (Fig. 11, lane 4).

These experiments using E. coli-expressed Toa1p–Toa2p and Kap122p show that Kap122p interacts directly with Toa1p–Toa2p complex.

We investigated whether the complex of recombinant GST-Kap122p/His₆-Toa1p/Toa2p is sensitive to dissociation by RanGTP but not RanGDP. To this end, we prepared a Talon™-bound complex of GST-Kap122p/His₆-Toa1p/Toa2p recombinant proteins (Fig. 11, lane 1). Beads were divided into three equal fractions and incubated with RanGTP, RanGDP, or TB alone. At the end of incubation, released material was collected, beads were extensively washed with TB, and bound and released proteins were analyzed by SDS-PAGE and Coomassie blue staining (Fig. 12). Incubation with RanGTP resulted in almost complete release of GST-Kap122p from Talon™-bound His₆-Toa1p/Toa2p complex (Fig. 12, lanes 1 and 2). Incubation with RanGDP did not release GST-Kap122p from the Toa1p–Toa2p complex (Fig. 12, lanes 3 and 4), neither did control incubation with TB in the absence of Ran (Fig. 12, lanes 5 and 6). These results confirm the specificity of the interaction between recombinant Kap122p and the recombinant Toa1p–Toa2p complex as well as the sensitivity of this interaction to dissociation by RanGTP.

Based on sequence similarity with karyopherin βs, the product of the PDR6 gene of S. cerevisiae was previously suggested to be a member of the Kap β family (for review see Pemberton et al. 1998). In this paper, we report that the hitherto uncharacterized product of the PDR6 gene does indeed function as a Kap and, therefore, named it Kap122p. We show that Kap122p functions in the nuclear import of the complex of large and small subunits, Toa1p, and Toa2p, of TFIIA. The relationship between the observed drug resistant phenotype of PDR6 and the function of Kap122p/Pdr6p in nuclear import of TFIIA (or of other proteins) remains to be elucidated.

We found that Kap122p is localized both in the cytoplasm and the nucleus, which is consistent with its function of shuttling between these two compartments. Cytosolic Kap122p exists as a complex with the small and large subunit of TFIIA. The precise stoichiometry of this complex remains to be determined. Based on biochemical and crystallographic data, it is unlikely that the two subunits exist as separate entities (Ranish et al. 1992; Geiger et al. 1996; Tan et al. 1996). Therefore, one possibility is that Kap122p also functions as a chaperone, and that immediately after synthesis in the cytoplasm, each subunit associates with Kap122p. In this scenario, each subunit would contain a Kap122p cognate NLS. Each of the subunit/Kap122p heterodimers would associate, via interaction between the two subunits, to form a tetramer, which is imported into the nucleus. Alternatively, only one of the subunits may contain a Kap122p-cognate NLS. After synthesis, this subunit could associate with Kap122p and with the other subunit to form a heterotrimer that would be imported into the nucleus. After import, in each of these two scenarios, RanGTP (Fig. 10 and Fig. 12) would dissociate the TFIIA heterodimer from Kap122p. Our data here argue against a third possibility, namely that a subunit/Kap122p heterodimer is imported separately because we found a stable interaction between the two subunits in the cytoplasm. In fact, while the Toa1p–Toa2p complex was dissociated from Kap122p by MgCl₂ concentrations between 100 and 1,000 mM, the interaction between the two subunits resisted dissociation at these MgCl₂ concentrations (Fig. 6, Fig. 7, Fig. 9, and Fig. 10).

PrA-tagged Kap122p was found to be associated in the cytoplasm with other proteins (Fig. 3). We do not yet know the identity of these proteins and whether they represent contaminants or alternative import substrates for Kap122p. However, it is clear that these other proteins are not part of the Toa1p/Toa2p/Kap122p complex as they were not copurified in a reverse pullout with PrA-tagged Toa1p or Toa2p (Fig. 6 and Fig. 7). As in the case of other PrA-tagged Kaps, several degradation products of Kap122p-PrA retaining their protein A moiety and eluting predominantly in the 4,500 mM MgCl₂ fraction were observed and confirmed by immunoblotting (data not shown).

We were able to reconstitute the Kap122p/Toa1p/Toa2p complex from recombinant proteins (Fig. 11). Moreover, this complex was sensitive to dissociation by RanGTP but not RanGDP (Fig. 12). These findings support the conclusion that Kap122p interacts directly with the Toa1p–Toa2p complex and that this interaction, like Kap122p/Toa1p/Toa2p interaction in the yeast cytosol, is sensitive to dissociation by RanGTP but resists dissociation by RanGDP.

It is known that RanGTP dissociates import substrates from Kaps by binding to the Kap (Rexach and Blobel 1995). The X-ray crystal structure of RanGTP complexed to Kap β2 (transportin) and Kap β1 (importin) has been determined (Chook and Blobel 1999; Vetter et al. 1999). Ran binding to Kap122p/Pdr6p has been previously investigated and no Ran binding was detected (Görlich et al. 1997). Consistent with this report, we have so far not been able to demonstrate binding of RanGTP to Kap122p in overlay or solution binding assays (data not shown). This might indicate that RanGTP binds Kap122p with very low affinity. Stable binding of RanGTP to Kap122p may require additional proteins. However, the Kap122p/Toa1p/Toa2p complex isolated from yeast (Fig. 10) or reconstituted from recombinant proteins (Fig. 11 and Fig. 12) could be dissociated by RanGTP but not RanGDP. These data confirm that RanGTP is directly involved in dissociating Kap122p from the Toa1p–Toa2p complex and provide support for the existence of functionally relevant interaction between RanGTP and Kap122p.

It appears that Kap122p is the principal Kap dedicated to the import of TFIIA because in a strain where KAP122 had been deleted there was a significant mislocalization of the two TFIIA subunits from the nucleus to the cytoplasm (Fig. 4 and Fig. 8). Surprisingly, KAP122 deletion is not lethal, whereas deletion of either of the two TFIIA subunits is. Therefore, Kaps other than Kap122p are likely to be involved in nuclear import of the TFIIA subunits. However, so far we have failed to identify alternative Kaps by cytosolic pullout experiments with Toa2p-PrA in a kap122Δ strain (Fig. 9). It is likely that import of the TFIIA subunits by these putative alternative Kaps proceeds with much lower efficiency than import by Kap122p, based on the significant reduction in the nuclear localization of the TFIIA subunits observed in a kap122Δ strain. Nevertheless, import of these two essential proteins in the absence of Kap122p appears to be sufficient, as there is no apparent difference in the growth rate between the kap122Δ and an isogenic wt strain. Alternative Kap(s) may be difficult to detect in an immunoisolation assay as they may bind to the two TFIIA subunits with lower affinity. There are precedents for essential proteins being imported by several Kaps of which the principal one is not essential. For example, ribosomal proteins have been shown to be imported by the abundant Kap123p (Rout et al. 1997; Schlenstedt et al. 1997). However, Kap123p is not essential, but the essential Kap121p can back up ribosomal protein import in a kap123Δ strain (Rout et al. 1997; Seedorf and Silver 1997). Alternatively, the Toa1p–Toa2p heterodimer is small enough (<60 kD) that it might diffuse through the NPC.

Kap122p is the third Kap that so far has been shown to be dedicated to the import of a general transcription factor, the others being Kap119p, which is involved in the nuclear import of the nonessential transcription factor TFIIS (Albertini et al. 1998), and Kap114p, which is involved in the nuclear import of the TATA binding protein (Pemberton et al. 1999). The advantages of maintaining dedicated Kaps for the import of specific transcription factors are likely to be in the regulatory realm and remain to be elucidated.

A. Titov would like to dedicate this paper to Philip B. Royle, friend and mentor, in gratefulness for his encouragement, faith, and inspiration.

We thank Roland Beckmann, Lucy Pemberton, and Jonathan Rosenblum for advice and reading the manuscript, J. Fernandez (Rockefeller University Protein/DNA Technology Center, New York, NY) for protein sequence analysis, F. Gharahdaghi (PDTC) for mass spectral analysis, M. Floer for Ran and RanGAP, and E. Ellison and M. Misiak for technical assistance.