Portal de Periódicos da CAPES

Dbp5, a DEAD-box protein required for mRNA export, is recruited to the cytoplasmic fibrils of nuclear pore complex via a conserved interaction with CAN/Nup159p

1999; Springer Nature; Volume: 18; Issue: 15 Linguagem: Inglês

10.1093/emboj/18.15.4332

1460-2075

Caroline Schmitt,

RNA and protein synthesis mechanisms

Article2 August 1999free access Dbp5, a DEAD-box protein required for mRNA export, is recruited to the cytoplasmic fibrils of nuclear pore complex via a conserved interaction with CAN/Nup159p Christel Schmitt Christel Schmitt University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland Search for more papers by this author Cayetano von Kobbe Cayetano von Kobbe University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland Search for more papers by this author Angela Bachi Angela Bachi European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Nelly Panté Nelly Panté Institute of Biochemistry, Swiss Federal Institute of Technology (ETH), Universitätstrasse 16, CH-8092 Zürich, Switzerland Search for more papers by this author João P. Rodrigues João P. Rodrigues Institute of Histology and Embryology, Faculty of Medicine, University of Lisbon, 1699 Lisboa, Codex, Portugal Search for more papers by this author Cécile Boscheron Cécile Boscheron University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland Search for more papers by this author Guillaume Rigaut Guillaume Rigaut European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Matthias Wilm Matthias Wilm European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Bertrand Séraphin Bertrand Séraphin European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Maria Carmo-Fonseca Maria Carmo-Fonseca Institute of Histology and Embryology, Faculty of Medicine, University of Lisbon, 1699 Lisboa, Codex, Portugal Search for more papers by this author Elisa Izaurralde Corresponding Author Elisa Izaurralde University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland Search for more papers by this author Christel Schmitt Christel Schmitt University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland Search for more papers by this author Cayetano von Kobbe Cayetano von Kobbe University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland Search for more papers by this author Angela Bachi Angela Bachi European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Nelly Panté Nelly Panté Institute of Biochemistry, Swiss Federal Institute of Technology (ETH), Universitätstrasse 16, CH-8092 Zürich, Switzerland Search for more papers by this author João P. Rodrigues João P. Rodrigues Institute of Histology and Embryology, Faculty of Medicine, University of Lisbon, 1699 Lisboa, Codex, Portugal Search for more papers by this author Cécile Boscheron Cécile Boscheron University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland Search for more papers by this author Guillaume Rigaut Guillaume Rigaut European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Matthias Wilm Matthias Wilm European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Bertrand Séraphin Bertrand Séraphin European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Maria Carmo-Fonseca Maria Carmo-Fonseca Institute of Histology and Embryology, Faculty of Medicine, University of Lisbon, 1699 Lisboa, Codex, Portugal Search for more papers by this author Elisa Izaurralde Corresponding Author Elisa Izaurralde University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland Search for more papers by this author Author Information Christel Schmitt1, Cayetano von Kobbe1, Angela Bachi2, Nelly Panté3, João P. Rodrigues4, Cécile Boscheron1, Guillaume Rigaut2, Matthias Wilm2, Bertrand Séraphin2, Maria Carmo-Fonseca4 and Elisa Izaurralde 1 1University of Geneva, Department of Molecular Biology, 30 quai Ernest-Ansermet, CH-1205 Geneva, Switzerland 2European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany 3Institute of Biochemistry, Swiss Federal Institute of Technology (ETH), Universitätstrasse 16, CH-8092 Zürich, Switzerland 4Institute of Histology and Embryology, Faculty of Medicine, University of Lisbon, 1699 Lisboa, Codex, Portugal ‡C.Schmitt, C.von Kobbe and A.Bachi contributed equally to this work *Corresponding author. E-mail: [email protected] The EMBO Journal (1999)18:4332-4347https://doi.org/10.1093/emboj/18.15.4332 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Dbp5 is a DEAD-box protein essential for mRNA export from the nucleus in yeast. Here we report the isolation of a cDNA encoding human Dbp5 (hDbp5) which is 46% identical to yDbp5p. Like its yeast homologue, hDbp5 is localized within the cytoplasm and at the nuclear rim. By immunoelectron microscopy, the nuclear envelope-bound fraction of Dbp5 has been localized to the cytoplasmic fibrils of the nuclear pore complex (NPC). Consistent with this localization, we show that both the human and yeast proteins directly interact with an N-terminal region of the nucleoporins CAN/Nup159p. In a conditional yeast strain in which Nup159p is degraded when shifted to the nonpermissive temperature, yDbp5p dissociates from the NPC and localizes to the cytoplasm. Thus, Dbp5 is recruited to the NPC via a conserved interaction with CAN/Nup159p. To investigate its function, we generated defective hDbp5 mutants and analysed their effects in RNA export by microinjection in Xenopus oocytes. A mutant protein containing a Glu→Gln change in the conserved DEAD-box inhibited the nuclear exit of mRNAs. Together, our data indicate that Dbp5 is a conserved RNA-dependent ATPase which is recruited to the cytoplasmic fibrils of the NPC where it participates in the export of mRNAs out of the nucleus. Introduction RNA helicases are enzymes that use energy derived from nucleoside triphosphate hydrolysis to unwind double-stranded RNA (reviewed by Schmid and Linder, 1992; Gorbalenya and Koonin, 1993; Fuller-Pace, 1994; de la Cruz et al., 1999). In most cases, the NTPase activity is stimulated by or is dependent on RNA binding, and therefore RNA helicases are considered as RNA-dependent NTPases. Although the precise mechanism by which these enzymes unwind RNA remains to be established, RNA helicases are thought to unwind short duplex regions in RNA molecules in a non-processive manner (Schmid and Linder, 1992; Gorbalenya and Koonin, 1993; Fuller-Pace, 1994). Moreover, RNA helicases may also be implicated in disrupting RNA–protein interactions (Lorsch and Herschlag, 1998a,b; see also Staley and Guthrie, 1999). Members of the RNA helicase protein family are defined by the presence of seven evolutionarily conserved motifs. The conserved motifs define the helicase core of the protein. Based on the particular consensus sequences of these conserved motifs, RNA helicases have been grouped into two major superfamilies of proteins (SFI and SFII; Gorbalenya and Koonin, 1993). The function of some of the conserved motifs has been elucidated by studying the effects of mutations in ATP and RNA binding, ATP hydrolysis and unwinding activity. Motifs I and II, known as Walker motifs A and B, are defined as the NTPase motifs and are involved in NTP binding and hydrolysis (Walker et al., 1982; Rozen et al., 1989), while motif VI has been implicated in RNA binding (Pause et al., 1993). Depending on the particular consensus sequence in the conserved motif II, SFII RNA helicases are divided further into the DEAD-box and the DExH-box protein families. The prototype of an RNA helicase of the DEAD-box family is the translation initiation factor eIF4A (Linder et al., 1989). In the presence of ATP and Mg2+, eIF4A unwinds short RNA duplexes in a non-processive manner (Rogers et al., 1999). Processivity is conferred by a second protein, eIF4B, which stimulates the helicase activity of eIF4A allowing the enzyme to unwind longer and more stable RNA duplexes (Rozen et al., 1990; Rogers et al., 1999). Comparison of the crystal structure of the hepatitis C virus (HCV) NS3 RNA helicase domain (a SFII DECH-box helicase) with that of bacterial DNA helicases suggests that the folding of the conserved motifs is very similar in both types of enzymes (Subramanya et al., 1996; Yao et al., 1997; Cho et al., 1998; Kim et al., 1998; Velankar et al., 1999). Due to the high conservation of residues at consensus positions throughout the conserved core, it has been proposed that the structure of the helicase core in other members of the family will closely resemble that of HCV NS3 enzyme. The HCV NS3 helicase consists of three domains arranged in a Y shape (Yao et al., 1997; Cho et al., 1998; Kim et al., 1998). The NTPase motifs I and II are located in domain 1, and define one branch of the Y. Motif VI is located at the opposite branch, in domain 2. Both domains are connected by a flexible hinge that harbours motif III. At the stem of the Y, domain 3 does not contain any conserved motif and is specific to the NS3 enzyme. Domains 1 and 2 undergo large conformational changes upon binding RNA and the nucleotide cofactor (Yao et al., 1997; Cho et al., 1998; Kim et al., 1998; see also Velankar et al., 1999). Apart from the helicase core, most helicases have variable N- and/or C-terminal unique extensions which have been suggested to play a role in determining substrate specificity and subcellular localization of the enzyme. For instance, the non-conserved N-terminal domain of Prp16p is required for its specific binding to the spliceosome (Wang and Guthrie, 1998). However, specificity cannot be conferred solely by the unique extensions as eIF4A consists essentially of a helicase core without extensions and it is a highly specific enzyme. Therefore, residues exposed at the surface of the helicase core are certainly engaged in the establishment of specific interactions. The observation that, in spite of the functional and structural similarities of their helicase cores, most helicases found in Saccharomyces cerevisiae are encoded by essential genes and do not have redundant functions strongly indicates that these enzymes are highly specific (reviewed by de la Cruz et al., 1999). RNA helicases are present in all organisms including prokaryotes, and are likely to participate in all steps of cellular RNA metabolism. In a search for S.cerevisiae DEAD-box protein genes using PCR-based strategies, Chang et al. (1990) identified five putative RNA helicases. These were called DEAD-box proteins (Dbp) 1–5. Recently, Snay-Hodge et al. (1998) and Tseng et al. (1998) have shown independently that yeast Dbp5p (yDbp5p) plays an essential role in mRNA export from the nucleus in yeast cells. Furthermore, both groups have reported that Dbp5 is evolutionarily conserved. This conclusion was based on the existence of cDNAs encoding Dbp5 homologues in mouse, Xenopus laevis, Dictyostelium discoideum and Schizoaccharomyces pombe, and on the observation that an antibody raised against yDbp5p recognized a protein of similar size in several species (Snay-Hodge et al., 1998; Tseng et al., 1998). yDbp5p is mainly cytoplasmic, but a fraction of the protein associates with nuclear pore complexes (NPCs) (Snay-Hodge et al., 1998; Tseng et al., 1998). Here we report the isolation of a cDNA encoding human Dbp5 (hDbp5). Human Dbp5 is mainly cytoplasmic, but a fraction of the protein associates with the nuclear rim. By immunoelectron microscopy in HeLa and yeast cells and in Xenopus laevis oocytes, we have localized a fraction of Dbp5 to the cytoplasmic fibrils of the NPC. Consistent with this localization, we have identified the nucleoporin CAN/Nup159p (von Lindern et al., 1992; Kraemer et al., 1994, 1995; Gorsch et al., 1995) as a directly interacting Dbp5 partner. CAN/Nup159p has been localized to the cytoplasmic fibrils of the NPC (Kraemer et al., 1994, 1995; Boer et al., 1997) and therefore is a good candidate for mediating the association of Dbp5 with the NPC. Indeed, we show that in a conditional yeast strain in which Nup159p is degraded when shifted to the non-permissive temperature, yDbp5p dissociates from the NPC and localizes entirely to the cytoplasm. Together, these results indicate that Dbp5 is recruited to the NPC via an evolutionarily conserved interaction with CAN/Nup159p. To investigate the functional conservation of hDbp5, we generated a collection of ATPase-deficient mutants. These mutants were tested by microinjection in X.laevis oocytes for their ability to inhibit RNA nuclear export in a dominant-negative manner. A mutant protein containing a Glu→Gln change in the conserved DEAD-box exhibits a dominant-negative effect on the nuclear exit of mRNAs. These results indicate that the vertebrate homologue of Dbp5 is also required for the export of mRNAs from the nucleus. Results Cloning of the human homologue of Dbp5 To investigate whether the function of Dbp5 in mRNA export has been conserved in higher eukaryotes, we have cloned and characterized its human homologue. Human expressed sequence tags (ESTs) that were likely to represent Dbp5 cDNA clones were identified by database searches. Based on these human ESTs, and on the homology with the murine cDNA, we designed primers to clone the full-length hDbp5 cDNA by RT–PCR. Consideration of the deduced amino acid sequence of the clones obtained by RT–PCR indicates that hDbp5 is 94 and 46% identical to the murine and yeast proteins, respectively (Figure 1A). Figure 1.Cloning of human Dbp5. (A) Complete amino acid alignment of predicted Dbp5 protein sequences from human, mouse and yeast using ClustalW1.7. Residue identity between species is highlighted by an asterisk below the sequence. Conserved sequence motifs (denoted by roman numerals) found in the helicase superfamilies are boxed. The positions of the substitution mutants are shown below the primary amino acid sequence of human Dbp5. The sequence data for human Dbp5 have been submitted to DDBJ/EMBL/GenBank under accession No. Aj237946. (B) Protein samples from HeLa nuclear (N) and cytoplasmic S-100 (C) extracts were analysed by Western blot using polyclonal anti-hDbp5 antibodies. In parallel, 35S-labelled hDbp5 synthesized in vitro in rabbit reticulocyte lysates was analysed. Download figure Download PowerPoint To confirm that our cDNA encodes the full-length protein, we compared the migration on SDS–PAGE of the recombinant protein synthesized in vitro in rabbit reticulocyte lysates with that of the protein present in HeLa cell extracts. In HeLa cell extracts, hDbp5 was detected by Western blot using rabbit antibodies raised against the bacterially expressed recombinant protein. The in vitro translated hDbp5 migrates with the apparent molecular weight of the protein detected in HeLa nuclear or cytoplasmic extracts (Figure 1B). These results, together with the presence of a Kozak consensus sequence surrounding the putative initiation codon and the homology with the murine protein, indicate that the entire open reading frame is likely to be present in our cDNA. RNA-dependent ATPase and ATP-dependent RNA unwinding activities of human Dbp5 Members of the DEAD-box protein family are predicted to exhibit RNA-dependent NTPase and NTP-dependent RNA-unwindase activities (see Introduction). We have analysed the RNA-binding, ATPase and unwindase activities of recombinant hDbp5 in vitro and compared them with those of its yeast homologue. Human and yeast Dbp5 were expressed in Escherichia coli as GST fusions. RNA binding was tested by an electrophoretic gel mobility retardation assay. To this end, a 32P-labelled RNA probe was incubated with the recombinant purified proteins, and the resulting complexes were resolved in a native polyacrylamide gel and visualized by autoradiography (Figure 2A). Binding reactions were performed in either the absence or presence of various nucleotides that previously were reported to influence RNA-binding affinity (Lorsch and Herschlag, 1998a,b). These were ATP, AMP-PNP, ATPγS and ADP. In the absence of nucleotides, hDbp5 did not interact with the RNA or the complexes formed were unstable (Figure 2A, lane 2). Inclusion of nucleotides in the binding reactions increased the formation and/or stabilized hDbp5–RNA complexes (Figure 2A, lanes 3–6). It is noteworthy that the complexes obtained in the presence of the nucleotides tested exhibited different mobilities. This is consistent with previous studies showing nucleotide-induced conformational changes in eIF4A (Lorsch and Herschlag, 1998a,b) and in PcrA DNA helicase (Velankar et al., 1999). AMP-PNP stabilized the formation of a fast migrating complex (LC: low complex, Figure 2A, lane 4) while ATPγS and ADP stabilized a slow migrating complex (lanes 5 and 6, respectively). The upper complex (UC) corresponds to the major complex obtained with a mutant protein having a Glu→Gln change in the conserved DEAD-box (see Figure 7C). Surprisingly, under the same conditions, GST–yDbp5p did not form stable complexes with the RNA probe (data not shown). This may reflect a difference in affinity and/or in sequence preference between these proteins. Figure 2.RNA binding, RNA-dependent ATPase and ATP-dependent unwinding activities of recombinant hDbp5. (A) Recombinant hDbp5 binds RNA in a nucleotide-dependent manner. A gel mobility retardation assay was performed with purified recombinant GST–hDbp5. In lanes 3–6, the nucleotides indicated above the lanes were added. The concentration of the nucleotides in the binding reactions was 2 mM, and that of the recombinant protein 0.2 mg/ml. The positions of the free RNA probe (lanes 1 and 7) and of the hDbp5–RNA complexes (lanes 2–6) are indicated on the left. The upper and lower complexes may represent two distinct conformations of hDbp5 bound to the RNA. (B) RNA-dependent ATPase activity of purified human and yeast Dbp5 proteins. In lanes 2–7 and 9–14, ATP hydrolysis was stimulated in the presence of the polynucleotides indicated above the lanes. In lanes 8 and 15, no polynucleotide was added. Lane 1 shows the ATP input. The position of ATP and ADP is indicated on the left of the panel. (C) hDbp5 immunoprecipitated from HeLa cell S–100 extracts unwinds RNA duplexes in vitro. Lane 1, RNA duplex; lane 2, the sample was boiled before loading onto the gel; lane 3, unwinding activity co-immunoprecipitated with anti-hDbp5 polyclonal antibodies; lanes 4 and 5, no unwinding activity was observed when ATP or Mg2+ were omitted; lane 6, background activity selected on beads pre-coated with the pre-immune serum. The RNA duplex and the monomer are indicated on the left of the gel. (D) Recombinant GST–hDbp5 unwinds RNA duplexes upon incubation with S-100 extracts. HeLa cell extracts were incubated with glutathione–agarose beads pre-coated with GST–hDbp5 (lane 4) or with GST fused to Dbp5 mutants E243Q+V386N (lane 5), R429Q (lane 6) and E243Q (lane 7). Unwinding activity was only observed on beads coated with the wild-type protein (lane 4) and with mutant R429Q. No unwinding activity was observed when the extracts were omitted (lane 3). The bands below the duplex in lane 3 are likely to represent partial degradation of the RNA probe. Symbols are as in (C). Download figure Download PowerPoint Figure 3.Human Dbp5 is localized in the cytoplasm and nuclear rim. (A–C) HeLa cells were transfected with pEGFP-C1 hDbp5 using the calcium phosphate method. Approximately 20 h after transfection, cells were fixed in formaldehyde and observed directly with the fluorescence microscope. The fusion protein is detected predominantly in the cytoplasm. Arrows point to a rim staining at the nuclear periphery. (D–F) HeLa cells transfected with pEFGP-N3 hDbp5 were permeabilized with digitonin for 2 min before fixation in formaldehyde. The cells were then incubated sequentially with an anti-nucleoporin monoclonal antibody (mAb414) and an appropriate secondary antibody conjugated to Texas red. The superimposition of (D) and (E) confirms that GFP–Dbp5 co-localizes with nucleoporins. Bar, 10 μm. Download figure Download PowerPoint Previously it has been reported that the ATPase activity of recombinant yDbp5p is stimulated by poly(U), poly(A), poly(I), poly(C), poly(G) and tRNA (Tseng et al., 1998). We have compared the ATPase activity of hDbp5 with that of its yeast counterpart in vitro in the presence of different polynucleotides (Figure 2B). In the presence of poly(A), poly(U) or poly(C), both proteins hydrolyse >80% of the ATP present in the reaction, in 10 min at 37°C (Figure 2B, lanes 2–4 and 9–11, respectively). Addition of tRNA resulted in ∼50% hydrolysis (lanes 6 and 13) while, in the presence of poly(G) or of single-stranded DNA, <40% of the input ATP was hydrolysed. Omission of polynucleotides resulted in ∼10% hydrolysis for both proteins (Figure 2B, lanes 8 and 15). This residual activity is likely to be due to the presence of a contaminant activity from E.coli. Thus, the RNA-dependent ATPase activity of both hDbp5 and yDbp5p is stimulated most efficiently by poly(A), poly(U) and poly(C), while tRNA, poly(G) and ssDNA are less efficient. Few members of the DEAD-box protein family have been shown to unwind RNA helices in vitro (de la Cruz et al., 1999). eIF4A requires a cofactor, eIF4B, in order to unwind RNA duplexes in a processive manner (Rozen et al., 1990; Rogers et al., 1999). The ATP-dependent helicase activity of yDbp5p was observed when a protein A-tagged version of Dbp5 was expressed in yeast and purified using IgG–Sepharose, suggesting that yDbp5p may require a cofactor to unwind RNA duplexes (Tseng et al., 1998). Similarly, when hDbp5 was immunoprecipitated from HeLa cell S-100 extracts using anti-hDbp5 antibodies coupled to protein A–Sepharose, an unwinding activity was recovered on the beads (Figure 2C, lane 3). The observed RNA-unwinding activity required both ATP (Figure 2C, lane 4) and Mg2+ (Figure 2C, lane 5) and was not recovered on beads coated with the pre-immune serum (Figure 2C, lane 6). This activity is probably due to the endogenous hDbp5 protein present in the extracts, as in Western blots the antibodies do not exhibit major cross-reactivities with other proteins (Figure 1B, and data not shown). However, to rule out completely that the observed activity was derived from another unwindase cross-reacting with the antibody, we tested whether recombinant hDbp5 could unwind RNA duplexes upon incubation with HeLa cell S-100 extracts. HeLa cell extracts were incubated with glutathione–agarose beads pre-coated with GST–hDbp5. Upon binding and after extensive washes, beads were tested for the presence of an RNA-unwinding activity. We expected that if hDbp5 requires a cofactor, we should be able to recover this cofactor on the beads. Figure 2D shows that an unwinding activity was recovered on beads pre-coated with GST–hDbp5 (lane 4). No activity was observed when HeLa cell extracts were omitted (lane 3). As a control, three Dbp5 mutants fused to GST were used. Mutants E243Q and E243Q+V386N are ATPase deficient (see below, Figure 7) and therefore cannot unwind the RNA duplex even in the presence of extracts (Figure 2D lanes 5 and 7). In contrast, mutant R429Q, which exhibits a reduced ATPase activity compared with the wild-type protein (Figure 7), unwound the RNA duplex upon incubation with the extracts, albeit with a reduced efficiency (Figure 2D, lane 6 versus lane 4). Together, these results suggest that hDbp5 is an ATP-dependent RNA unwindase that requires a cofactor to exhibit unwinding activity. Note that the substrate employed in the assays shown in Figure 2C and Figure 2D consisted of an RNA duplex of 25 nucleotides (ΔG° = approximately −44 kcal/mol at 37°C; Turner et al., 1988), flanked by 3′-terminal extensions of ∼78 nucleotides of single-stranded RNA. Thus, Dbp5 does not require a 5′ single-stranded region on the substrate. Human Dbp5 is localized in the cytoplasm and at the nuclear rim Yeast Dbp5p has been localized both to the cytoplasm and to the nuclear envelope (Snay-Hodge et al., 1998; Tseng et al., 1998; see Figure 5). To investigate the localization of hDbp5, we generated fusion proteins by tagging hDbp5 with the green fluorescent protein (GFP) at its N- or C-terminus. hDbp5–GFP fusions could be vizualized directly upon transfection. Figure 3A–C shows the results obtained when GFP was fused to the N-terminus of hDbp5. In transfected HeLa cells, the majority of the GFP-tagged protein was detected in the cytoplasm, while the nucleoplasm was largely free from staining. Moreover, a fraction of the protein appeared to be concentrated around the nucleus. A similar subcellular localization of hDbp5 was found when the GFP tag was fused to its C-terminus (Figure 3D). The nuclear rim staining became more apparent when cells were treated with digitonin prior to fixation (Figure 3D). The nuclear rim-associated fraction of Dbp5 co-localizes with the labelling produced by monoclonal antibody 414 directed against nucleoporins (Figure 3E and F). Thus, hDbp5, like its yeast homologue, is predominantly cytoplasmic, but a fraction localizes to the nuclear rim. Figure 4.Identification of CAN/Nup159p as a direct interacting Dbp5 partner. (A) Tandem affinity purification of yDbp5p partners. Fractions eluted from the calmodulin affinity resin were analysed on SDS–PAGE followed by silver staining (lanes 2–5). Proteins indicated on the right of the gel were identified by mass spectrometry. (B) HeLa cell S-100 extracts were incubated with glutathione–agarose beads on which GST–hDbp5 or GST alone were immobilized (lanes 2 and 3, respectively). After incubation and extensive washes, bound proteins were eluted and analysed by SDS–PAGE followed by silver staining. Lane 4 shows proteins eluted from the GST–hDBP5-coated beads when the extracts were omitted. (C) [35S]Methionine-labelled fragments of CAN (residues 1–586) and of Nup159p (residues 1–417) were synthesized in vitro in rabbit reticulocyte lysates. Samples of 2 μl from the lysates were incubated with glutathione–agarose beads pre-coated with the recombinant proteins indicated above the lanes. One-tenth of the inputs (lane 1), one-third of the bound fractions (lanes 2–4) and one-twentieth of the unbound fractions (lanes 5–7) were analysed on SDS–PAGE followed by fluorography. (D) Lysates from E.coli expressing His6-tagged Nup159p (residues 1–417) were incubated with Ni-NTA beads (lane 9) or glutathione–agarose beads pre-coated with the recombinant proteins indicated above the lanes (lanes 3, 5 and 7). Bound proteins were eluted with SDS sample buffer and analysed on SDS–PAGE followed by Coomassie Blue staining. For each selection, the background obtained in the absence of lysates is shown (lanes 2, 4 and 6). The position of the His-tagged Nup159p N-terminal domain is indicated on the right of the gel. Download figure Download PowerPoint Figure 5.Localization of GFP-tagged Dbp5 in a yeast strain carrying a conditional nup159 allele. Yeast cells were grown overnight in liquid YDP at 23°C until mid-exponential phase (OD600 of 1). Cells were then either kept at 23°C or shifted for 1 h to 37°C. A 1 ml aliquot of cell suspension was centrifuged briefly and resuspended in 10 μl of water. A 3 μl aliquot of cell suspension was mounted on a microscope slide, sealed with a coverslip and immediately inspected. Nuclear rim staining could be observed readily in cells growing at the permissive temperature. Following incubation at 37°C, the rim staining was no longer observed and Dbp5–GFP localized to the cytoplasm. Scale bar, 10 μm. The lower panels show different fields magnified ∼1.5-fold. Download figure Download PowerPoint Identification of an evolutionarily conserved interaction between Dbp5 and CAN/Nup159p The nuclear rim association of Dbp5 in both HeLa and yeast cells suggests that a fraction of the protein interacts with components of the NPC. Furthermore, experiments shown in Figure 2D suggest that Dbp5 requires a cofactor in order to unwind RNA duplexes. To identify Dbp5 partners, we have used a novel approach, the tandem affinity purification (TAP) strategy, that allows efficient purification of protein complexes from yeast cell extracts (see Materials and methods). Because of the high degree of similarity between human and yeast Dbp5, their similar localization and function (see below), we expected that identification of its interacting proteins in yeast would lead to the identification of Dbp5 partners in vertebrate cells. The TAP tag was fused to the C-terminus of yDbp5 protein by integrating a DNA cassette into the genome of a haploid cell (Puig et al., 1998). The TAP tag consists of a calmodulin-binding peptide followed by a TEV protease cleavage site and two IgG-binding units of Staphylococcus aureus protein A (ProtA). Because the tagged protein is the only source of the essential yDbp5 function, and the tagged strain did not display a strong growth phenotype, we conclude that the TAP tag did not abolish Dbp5 protein function. Extracts were prepared from the tagged strain, and Dbp5 with its associated proteins was purified following the two-step affinity purification of the TAP method. Figu