A Novel Nuclear Import Pathway for the Transcription Factor TFIIS

The yeast strains used in this study were Saccharomyces cerevisiae wild-type DF5α (Finley et al., 1987) and its derivative, kap119Δ (MATα, lys2-801, leu2-3, 2-112, ura3-52, his3-Δ200, trp1-1 [am], NMD5::HIS3). Yeast strains were grown at 30°C in YPD or in minimal medium (YNBD). Required supplements were added. Manipulation of yeast cells was performed according to standard methods (Rose et al., 1990). Bacterial transformation was performed essentially as described (Ausubel et al., 1997).

To delete the KAP119 gene in wild-type strain DF5α, the HIS3 gene was used as a selective marker for insertion into the genomic locus. Deletion cassettes containing the HIS3 gene and 60 nucleotides of each the 5′ and 3′ untranslated regions of the KAP119/NMD5 open reading frame (ORF) were constructed by PCR. For protein A tagging of Kap119p and TFIIS, a PCR product was generated that contained a protein A, HIS3, and URA3 cassette (Aitchison et al., 1995) flanked by 60 nucleotides directly upstream of the KAP119/NMD5 stop codon and 60 nucleotides downstream of the coding region of the corresponding gene. Yeast cells were transformed by electroporation and the ORF replaced by integrative transformation in a diploid strain. Heterozygous diploids were induced to sporulate, and the meiotic progeny were examined by standard tetrad analysis (Pemberton et al., 1997).

Overlay assays were performed essentially as described (Rout et al., 1997). Equal amounts of Escherichia coli lysates were separated by SDS-PAGE and transferred to nitrocellulose. Nitrocellulose strips were incubated with yeast cytosol from strains expressing either Kap119–PrA or Kap95–PrA at 4°C for 14 h. Cytosol from the Kap119–PrA strain was used undiluted, Kap95–PrA cytosol was diluted 1:5 with transport buffer (20 mM Hepes-KOH, pH 7.5, 110 mM KOAc, 2 mM MgCl₂, 0.1% Tween 20) containing 5% dried skim milk plus 0.001 vol of a protease inhibitor cocktail (Rout and Kilmartin, 1990). Bound PrA fusion proteins were detected with rabbit IgG (Cappel, Malvern, PA) and anti-rabbit IgG-coupled HRP (Amersham, Arlington, IL) was used as the secondary antibody, and blots were developed using the enhanced chemiluminescence (ECL) system (Amersham). All incubations were performed in transport buffer containing 5% dried skim milk.

Indirect immunofluorescence microscopy was performed as described by Rout et al. (1997). Yeast cells were fixed with 3.7% formaldehyde for 5 min and cell walls were digested. The PrA tags were detected with rabbit IgG, Nab2p with polyclonal mouse antiserum (Aitchison et al., 1996), and Npl3p using antibody 1E4 (Wilson et al., 1994). CY3-conjugated donkey anti–mouse and anti–rabbit IgG as well as FITC-conjugated donkey anti– mouse IgG were used as 6 μg/ml solutions for visualization. GFP reporter proteins were detected directly in fixed cells. DNA was visualized by 4′,6-diamidino-2-phenylindole (DAPI). Immunoblotting experiments were performed using anti-rabbit IgG-coupled HRP (Amersham) as secondary antibody, and blots were developed using the ECL system (Amersham). Nop1p was detected with monoclonal mouse antiserum D66 (Aris and Blobel, 1998) and 3-phosphoglycerate kinase with monoclonal mouse antiserum 22C5-D8 (Molecular Probes, Eugene, OR).

Spheroplasting of yeast cells and preparation of nuclei and postribosomal supernatant were described (Rout and Kilmartin, 1990, 1994; Wilson et al., 1994; Melchior et al., 1995; Aitchison et al., 1996). PrA-tagged Kap119p and TFIIS were immunoisolated according to Aitchison et al. (1996). Proteins were analyzed by SDS-PAGE and stained by Coomassie blue. Bands of interest were excised and prepared for matrix-assisted laser desorption/ ionization time-of-flight (MALDI-TOF) mass spectrometry.

Recombinant yeast Ran (Gsp1) was prepared and nucleotide exchange performed as described by Floer and Blobel (1996). After nucleotide exchange RanGDP was incubated with Rna1p to convert remaining RanGTP into the GDP-bound state: 50 μM RanGDP and 4 μM RanGAP were incubated in a final volume of 100 μl for 12 h at 21°C. All further RanGDP incubations were performed without previous removal of RanGAP.

PrA-tagged Kap119p and bound TFIIS were coimmunoisolated as described above. After the final wash, IgG-Sepharose–bound Kap119-PrA/ TFIIS was divided into three 50-μl portions and placed into siliconized reaction tubes. The beads were resuspended in 240 μl transport-buffer (20 mM Hepes-KOH, pH 7.5, 150 mM KOAc, 2 mM MgCl₂, 1 mM DTT, 0.1% Tween 20, 0.1 mM PMSF, 1 μg/ml leupeptin and 1 μg/ml pepstatin A) and the slurry was incubated in batch with buffer alone, RanGDP, or RanGTP (final concentration 5 μM) in a final volume of 250 μl for 60 min at 21°C. The beads were then collected by centrifugation at 2,000 g for 30 s and unbound proteins were removed in the supernatant and collected for further analysis. Beads were washed twice with 1 ml of transport buffer. Analysis of the bound proteins was performed as described for immunoisolation except that the elution of bound proteins was performed only in two steps with 250 mM and 4,500 mM MgCl₂.

Nmd5p (for nonsense-mediated mRNA decay) (GenBank/ EMBL/DDBJ accession number U31375) is a yeast protein that was originally identified in a two-hybrid screen with the cytoplasmic protein Upf1p (up frame shift mutation) as a bait (Atkin et al., 1995; He and Jacobsen, 1995). Upf1p functions in the rapid degradation of mRNAs that contain a premature stop codon (He and Jacobsen, 1995). However, a direct interaction between Upf1p and Nmd5p has not been demonstrated and it is not clear what function, if any, Nmd5p has in Upf1p-mediated mRNA degradation. As we show in this paper that Nmd5p functions as a Kap we suggest the alternative name Kap119p in keeping with previously proposed yeast nomenclature (Enenkel et al., 1995; Aitchison et al., 1996). The amino acid sequence of Kap119p is 19% identical with that of Kap95p and 26% identical with that of Kap108p (Rosenblum et al., 1997, 1998) (data not shown).

The KAP119 gene was replaced by KAP119-PrA coding for a chimeric Kap119p that is COOH terminally fused to IgG-binding domains of Staphylococcus aureus protein A. The resulting Kap119–PrA strain showed no difference in its growth rate compared with wild-type cells. By immunofluorescence (Fig. 1,A) and by cell fractionation (Fig. 1,B) the Kap119–PrA was localized to the cytoplasm and the nucleus indicating that the protein might shuttle between these two compartments. Cell fraction analyses (Fig. 1 B) indicated that about one-third of the total Kap119–PrA was localized to the nucleus. Considering that the nucleus occupies less than one-third of the cellular volume, Kap119p appears to be slightly more concentrated in the nucleus than in the cytoplasm.

To coimmunoisolate potential transport substrate(s), a cytosol of the Kap119–PrA strain was incubated with IgG Sepharose. After extensive washing, the bound material was eluted with a MgCl₂ step gradient ranging from 50 mM to 4.5 M MgCl₂. The proteins in the eluted fractions were analyzed by SDS-PAGE and stained with Coomassie blue (Fig. 2). One major protein of ∼35 kD eluted at 50 and 100 mM MgCl₂. In agreement with the previously observed elution behavior of transport substrates from PrA-tagged Kaps, the 35-kD band was a strong candidate for an Kap119p-bound transport substrate. As expected (Aitchison et al., 1996; Pemberton et al., 1997; Rosenblum et al., 1997; Rout et al., 1997), the Kap119–PrA eluted between 1.0 and 4.5 M MgCl₂. By mass spectrometry, the 35-kD band was identified as TFIIS (also called PPR2, YSII, DST1, STALPHA, or YGL043) (Hubert et al., 1983; Davies et al., 1990; Clark et al., 1991; Nakanishi et al., 1992) by virtue of eight tryptic peptides between 1,000 and 1,915 D. The matching peptides, with 0.5 D tolerance, covered 29% of the TFIIS sequence. The TFIIS gene is not essential (Nakanishi et al., 1992) and codes for a protein of 309-amino acid residues with a calculated M_r of 34.8 kD. Homologous proteins have been identified in Drosophila, various mammalian genomes and in the vaccinia virus genome (for review see Kassavatis and Geiduschek, 1993). TFIIS binds to RNA polymerase II (Rappaport et al., 1988), and stimulates cleavage and elongation of nascent RNA in the transcription elongation complex (Izban and Luse, 1992; Nakanishi et al., 1992; Borukhov et al., 1993). As TFIIS contains a zinc-binding domain, it may directly interact with nucleic acid (Agarwal et al., 1991).

To investigate whether TFIIS was a nuclear import substrate of Kap119p we prepared a strain lacking KAP119. The kap119Δ strain was viable and grew on YPD plates but was characterized by a slightly extended doubling time compared with the wild-type haploid strain (Fig. 3 A). As TFIIS is primarily localized to the nucleus (Nakanishi et al., 1995) it should be mislocalized to the cytoplasm in the kap119Δ strain if Kap119p functions as its principal cognate Kap. To facilitate localization of TFIIS we protein A–tagged the endogenous TFIIS gene in a wild-type and a kap119Δ strain and investigated its cellular distribution by immunofluorescence and cell fractionation.

By immunofluorescence TFIIS–PrA was indeed largely mislocalized to the cytoplasm in the kap119Δ strain (Fig. 3,B, middle), whereas in the wild-type strain it was primarily localized to the nucleus (Fig. 3,B, top). As Kap108p is the closest relative of Kap119p in yeast (26% identical) (Rosenblum et al., 1997), we also determined the localization of TFIIS–PrA in a kap108Δ strain. We found that TFIIS– PrA was properly localized to the nucleus and was not mislocalized to the cytoplasm (Fig. 3 B, bottom), indicating that Kap108p is not involved in the nuclear import of TFIIS.

The immunofluorescence results were further substantiated by cell fractionation of wild-type and kap119Δ cells and comparison of the distribution of TFIIS relative to marker proteins (Fig. 3 C). Compared with the wild-type strain a clearly reduced amount of TFIIS–PrA was detected in the nuclear fraction of kap119Δ. We conclude that Kap119p functions as cognate Kap for the import of TFIIS into the nucleus.

We tested whether cytosolic TFIIS binds only to Kap119p or whether it can also bind to other Kaps in a strain where Kap119p was deleted. Cytosol from a wild-type or a kap119Δ strain was incubated with IgG–Sepharose, bound protein eluted by MgCl₂ step gradients, analyzed by SDS-PAGE, and then stained with Coomassie blue. In the wild-type strain a major band of ∼120 kD eluted a 100 mM MgCl₂ (Fig. 4, left). Mass spectrometric analysis showed that this band was Kap119p. The TFIIS–PrA eluted between 1.0 and 4.5 M MgCl₂. Judging from the staining intensity of the Kap119p and TFIIS–PrA bands, it appears that most of the cytosolic TFIIS–PrA was associated with Kap119p, suggesting that Kap119p is a major binding partner of TFIIS in the cytosol. Interestingly, in the cytosol prepared from the kap119Δ strain no clear candidates for additional Kaps were observed (Fig. 4, right) suggesting that TFIIS was not associated with another Kap in a strain deleted for Kap119p. This result is in agreement with the observed mislocalization of TFIIS reported above. We therefore conclude that Kap119p is the principal Kap for import of TFIIS.

To determine whether deletion of KAP119 generally affected transport, we localized several substrates, whose transport is known to be mediated by a distinct cognate Kap, to see whether they were also mislocalized in a kap119Δ strain. We localized NLS–green fluorescent protein (GFP) (Shulga et al., 1996), NES–GFP–NLS (Stade et al., 1997), Lhp1p–GFP (Rosenblum et al., 1998), Npl3p (Wilson et al., 1994), and Nab2p (Aitchison et al., 1996) in both wild-type and kap119Δ strains. These substrates are transported by Kap60p/Kap95p, Kap124p (Crm1p), Kap108p, Kap111p, and Kap104p, respectively. As shown in Fig. 5, the localization of these substrates (data not shown for Nab2p) appeared similar in both strains. Hence deletion of Kap119p does not generally affect transport nor does Kap119p-mediated import specifically impact any of these other pathways.

RanGTP has been reported to dissociate several import substrate/Kap complexes (Rexach and Blobel, 1995; Izaurralde et al., 1997; Siomi et al., 1997). To test whether RanGTP would dissociate TFIIS from Kap119p, we incubated the Sepharose-bound Kap119–PrA/TFIIS complex with recombinant RanGTP. After extensive washing, the Sepharose was split into three equal pools. One pool was further incubated with buffer alone while the two others were incubated with either RanGDP or RanGTP. Nearly all of the bound TFIIS (Fig. 6, lanes 3–5) was released from Kap119-PrA (Fig. 6, lanes 9–11) after incubation with RanGTP (Fig. 6, compare lane 5 with 11), but not after incubation with buffer (Fig. 6, compare lane 3 with 9) or with RanGDP (Fig. 6, compare lane 4 with 10). A portion of the added RanGTP remained bound to Kap119– PrA (Fig. 6, lane 5) whereas RanGDP did not bind (Fig. 6, lane 4). Hence, incubation of the complex with RanGTP resulted in dissociation of TFIIS and binding of RanGTP to Kap119–PrA. Importantly, identical amounts of Kap119–PrA were eluted from each pool.

During the course of nuclear import, the import complex needs to dock at the NPC. This docking is thought to be mediated by a subset of nucleoporins that bear peptide-repeat motifs, including Nup1p, Nup2p, and Nup159p (Radu et al., 1995b). Peptide repeat-containing domains of Nup1p, Nup2p, and Nup159p (Rexach and Blobel, 1995; Kraemer et al., 1995) were expressed in E. coli. Proteins of bacterial lysates were separated by SDS-PAGE, transferred to nitrocellulose, and then incubated with cytosol of the Kap119–PrA or Kap95–PrA strain and detected by luminol-based chemiluminescence (Fig. 7). As previously shown for Kap95–PrA (Iovine et al., 1995; Kraemer et al., 1995; Rexach and Blobel, 1995; Pemberton et al., 1997; Rosenblum et al., 1997), Kap119–PrA also bound to the tested nucleoporins: binding of Kap119–PrA to Nup1p and Nup2p was weaker than that of Kap95–PrA, whereas the extent of binding to Nup159p was similar for both PrA-tagged Kaps. These results indicate that Kap119p shares overlapping binding sites for peptide repeat containing Nups with other Kaps.

As mentioned in the introduction, substrates for nine Kaps had been previously identified. These Kaps function in export and import, and transport several classes of proteins and tRNA. With the experiments presented here, we have identified an additional class of import substrates, represented by TFIIS, which is an RNA polymerase II transcription elongation factor (Davies et al., 1990; Nakanishi et al., 1992). We have shown that Kap119p is the principal nuclear import factor for TFIIS and that TFIIS is largely mislocalized in strains devoid of KAP119, indicating that other Kaps are incapable of efficient import of TFIIS. Although, TFIIS appears to be the major substrate of Kap119p in the growth conditions used here, we cannot exclude the possibility that other proteins are also imported (or exported) by Kap119p. In fact, after submission of this article it was reported that Kap119p was required for the nuclear import of the MAP kinase Hog1p that occurs as an initial response to high osmolarity in the growth medium (Ferrigno et al., 1998). In addition, because of the high cytoplasmic ratio of RanGDP/RanGTP (for review see Corbett and Silver, 1997) export substrates would not be expected to be Kap associated in the cytoplasm, and we would be unable to identify them using the approach we have taken here.

Although a slight amount of TFIIS still localized to nuclei in kap119Δ cells (the reasons for this Kap119p independent nuclear localization remain to be clarified but could be due to passive diffusion and/or piggyback import together with subunits of RNA polymerase II) TFIIS was not efficiently imported by other Kaps and deletion of Kap119p did not affect the import of substrates by other pathways. Further, TFIIS, being involved in elongation of RNA polymerase II transcripts (for review see Kerppola and Kane, 1991; Kassavatis and Geiduschek, 1993), represents a novel import substrate. TFIIS is the first RNA polymerase II accessory protein that has been shown to require a dedicated Kap for specific import. It has been shown that proteins involved in mRNA processing/export require their dedicated Kaps for import (Aitchison et al., 1995; Pemberton et al., 1997; Senger at al., 1998), so it is not surprising that factors involved in mRNA production would need distinct Kaps as well. Interestingly, Kap119p and Kap108p (Rosenblum et al., 1997) are the most similar pair of yeast Kaps (26% identical). We have not detected any direct overlap in their functions with regard to the import of TFIIS and Lhp1p. However, these two main substrates for the Kap119p and Kap108p pathways may have evolutionarily-related functions, TFIIS being involved in RNA polymerase II transcription (Nakanishi et al., 1992; Kassavatis and Geiduschek, 1993) and Lhp1p interacting with RNA polymerase III transcripts (Yoo and Wolin, 1997).

Two Kaps have been identified in higher eukaryotes that are similar to Kap119p, RanBP7 from Xenopus and RanBP8 from humans (Görlich et al., 1997). In addition, TFIIS is highly conserved between yeast and humans (for review see Kassavatis and Geiduschek, 1993). Although yeast TFIIS appears to contain a consensus bipartite NLS (K₆₆KMISSWKDAINKNKRSR₈₃) our data here suggest that it is not imported by Kap60p/Kap95p. A consensus NLS is not discernible in the human TFIIS sequence. It remains to be seen if RanBP7 or RanBP8 import higher eukaryotic homologues of TFIIS. In addition, it is not yet clear whether evolution has maintained all the identified yeast Kapβs, or has diversified them even more (as seems to be the case with Kapβ2 [Siomi et al., 1997] [Bonifaci, N., G. Blobel, and A. Radu, unpublished results]), or has delegated some of their functions to an expanded Kapα reservoir (Kohler et al., 1997; Seki et al., 1997; Takeda et al., 1997; Nachury et al., 1998; Pemberton et al., 1998; Rosenblum et al., 1998). The advantages of maintaining a dedicated Kap for import of TFIIS are likely to be the regulation of TFIIS transport.

Judging from their relative staining intensities, it appears that the transport substrates and their cognate Kaps are present in import complexes in nearly stoichiometric amounts (Aitchison et al., 1995; Pemberton et al., 1997; Rosenblum et al., 1997) (Fig. 4). Apart from these complexes being required for import, their stability is likely to be physiologically relevant. The physiological benefits of complex formation between a cytoplasmic transport substrate and its cognate Kap may be to prevent cytoplasmic degradation of the transport substrate or to protect reactive sites on transport substrates that are designed to interact with nucleic acids and/or proteins following import into the nucleus. It is even possible that the cytoplasmic complexes of Kaps and substrates prevent unwanted diffusion of small substrates into the nucleus. Alternatively, the formation of complexes between Kaps and transport substrates may facilitate reversible posttranslational modification(s), such as phosporylation/dephosphorylation, for each round of shuttling between the cytoplasm and nucleoplasm. Although there is presently no evidence for such reversible modifications, they might be useful in regulating the frequency of shuttling.

As we have shown, RanGTP was able to dissociate most of the TFIIS from Kap119p. Due to the nuclear localization of the RanGEF (Ohtsubo et al., 1989; Aebi et al., 1990) and cytoplasmic localization of RanGAP (Hopper et al., 1990; Becker et al., 1995; Corbett et al., 1995), the concentration of RanGTP can be expected to be higher in the nucleus than in the cytoplasm (for review see Corbett and Silver, 1997). As such, RanGTP may be a sensor for the nucleoplasm, consistent with the ability of RanGTP to dissociate the Kap119p import complex. In summary, our data here identify Kap119p as a new member of the Kapβ family that is involved in the import of the transcription factor TFIIS.

We thank F. Gharahdaghi of the Rockefeller University Protein/DNA Technology Center (New York, NY) for the mass spectral analysis, M. Rout (Rockefeller University) for Kap95–PrA cytosol, J. Aitchison (University of Alberta, Edmonton, Canada) for the Nab2p antiserum, M. Swanson (University of Florida, Gainesville, FL) for the 1E4 antibody, K. Weis (University of California, San Francisco, CA) for the NES–GFP– NLS reporter, D. Kraemer (Universität Würzburg, Würzburg, Germany) and M. Rexach (Stanford University, Standford, CA) for E. coli expression constructs and M. Floer for Ran and RanGAP and much practical help. We are grateful to K. Yoshida and the other members of the Blobel Lab for stimulating discussions and practical help.

DAPI

4′,6-diamidino-2-phenylindole

GFP

green fluorescent protein

Kap

karyopherin

NES

nuclear export sequence

NLS

nuclear localization sequence

NPC

nuclear pore complex

Nups

nucleoporins

PrA

protein A from S. aureus