INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Many viruses depend on components of the nucleus to complete replication, and thus, at some stage, most viral genomes must enter the nucleus. Although several viruses are known to access the nucleus through the nuclear pores, little is known about the mechanisms responsible for viral transport. The nuclear pore complex (NPC) is an ~125-MDa structure, and in Saccharomyces cerevisiae, it has been estimated to contain approximately 30 different polypeptides (41, 44, 52). Several possible mechanisms for nuclear entry have been proposed. Viral particles could enter the nucleus intact, as is thought to be true of simian virus 40 (SV40) virions (37, 51). However, the sizes of virion or capsid particles are estimated to range from 26 to 125 nm and thus seem too large for most to be transported through the NPC (25). Maximal pore size estimates range from 9 to 26 nm (for reviews, see references 18 and 46). Alternatively, viral particles might enter the nucleus by means of a conformational change allowing the passage through nuclear pores. Another formal possibility is that virions dock at the nuclear pores and inject their contents into the nucleus, similar to what has been proposed for adenovirus (19). A more likely possibility for long-terminal-repeat (LTR)-containing retroelements is that a subassembly of the particle, which is called the preintegration complex (PIC), is transported into the nucleus by host cell machinery.

We study the nuclear entry of Tf1, an LTR-containing retrotransposon that propagates within the fission yeast Schizosaccharomyces pombe. The propagation of Tf1 requires many of the same processes used by retroviruses. Tf1 has coding sequences for Gag, protease (PR), reverse transcriptase (RT), and integrase (IN) proteins (30, 32). Once a full-length transcript of an LTR retroelement is produced and translated, the Gag, PR, RT, and IN proteins assemble with copies of the RNA to form virus or virus-like particles. RT converts the RNA into full-length double-stranded cDNA that is subsequently transported into the nucleus with IN. The cDNA is then inserted by IN into target sites in the host genome (4). The principal difference between retroviruses and LTR retrotransposons is that the viruses can escape the host cell and infect new cells.

In vivo assays for transposition demonstrate that Tf1 is highly active and generates transposition frequencies varying from 2 to 20% (31). Results of sucrose gradient sedimentation revealed that Tf1 Gag, IN, Tf1 mRNA, and reverse transcription products all assemble into virus-like particles (VLPs) that contain a 26-fold molar excess of Gag relative to IN (2, 28, 31).

The question of how Tf1 enters the nucleus is especially interesting because in yeasts the nuclear envelope (NE) does not break down during mitosis (26). For the integration of Tf1 to occur, the VLP or a PIC containing IN and cDNA must be transported across the nuclear membrane and into the nucleus. In higher eukaryotic cells, the nuclear envelope can also act as an obstacle of virus import that must be overcome. Lentiviruses, such as human immunodeficiency virus type 1 (HIV-1) and visna virus, are able to infect cells that have an intact NE by transporting large PICs through nuclear pores (7, 20, 33, 38, 50). In the case of other retroviruses, such as murine leukemia virus, the breakdown of the NE during cell division allows the PIC to interact with the host genome (42). However, avian sarcoma virus (ASV) does require nuclear import for efficient propagation (27).

Our previous studies of Tf1 showed that the Gag protein is localized in the nucleus (3, 8). As the result of a genetic screen, we identified Nup124p as a protein that is required for transposition and that a mutation in Nup124p caused a substantial defect in the nuclear import of Gag (3). Our studies of immunofluorescence and immunoelectron microscopies demonstrated that the Nup124p protein is a component of the NPC (3). We also showed that Nup124p possesses a specialized activity that is specifically required for the nuclear localization of transposon material and not the other cargos that were tested (3). It is indeed surprising that a strain containing a deletion of nup124 is defective for Tf1 transposition but nevertheless grows with wild-type rates (3).

In this report, we demonstrated that the Gag protein of Tf1 possessed sequences that were sufficient to cause the nuclear localization of fusion proteins. Mutations in a nuclear localization signal (NLS) in the N terminus of Gag caused a significant defect in transposition as well as the nuclear import of Gag and the Tf1 cDNA. Interestingly, the NLS activity that we identified in the N-terminal domain of Gag was sufficient when fused to GFP-LacZ to cause import into the nucleus that depended on the nuclear pore factor Nup124p. In addition, the NLS in this domain could be separated from another sequence element that specified the requirement for Nup124p. When this independent element was inserted next to the NLS of SV40, it was sufficient to cause nuclear import to become dependent on Nup124p.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Yeast strains, media, and genetic procedures. Yeast strains used in this study are listed in Table 1. Yeast transformation and the preparation of S. pombe minimal liquid and plate media were done as described previously (36). Selective plates contained Edinburgh minimal media (EMM) and 2 g of "dropout" mixture/ml, a powder with adenine, uracil, and all amino acids except for nutrients that are absent as required for selection (43). A concentration of 10 µM vitamin B1 (thiamine) was added to minimal medium to repress the nmt1 promoter. 5-Fluoroorotic acid (5-FOA) (U.S. Biologicals, Swampscott, Mass.) was used at 1 mg/ml in EMM. YES 5-FOA/G418 plates were made from YES medium containing 1 mg of 5-FOA/ml and 500 µg (corrected for purity) of geneticin (Gibco) per ml.

Plasmid construction. Many plasmids used for this study were constructed using PCR cloning techniques. To avoid problems related to inadvertent mutations created during PCR, the plasmids were made in duplicate from independent PCRs, and the properties of each plasmid were studied in parallel. The high-fidelity enzyme Pfu or Turbo Pfu (Stratagene) was used for all PCRs. Plasmids used in this study are also listed in Table 1.

Mutations in the transposon. Point mutations have been created within the sequences of the three NLSs of Gag by using the fusion PCR technique described previously (28). Starting from the Tf1-neo assay plasmid pHL891-19, the sequence for the N-terminal NLS of Gag, KRIR, was converted into either AAIR (double mutation) or AAIA (triple mutation). The two restriction sites XhoI and AvrII that were unique in the plasmid pHL891-19 were used for the cloning of the fusion-PCR products. Resulting plasmids that contain these new mutant versions of Tf1 were named pHL1891 and pHL1893, respectively. The fusion PCR technique was also exploited to generate point mutations within the two putative C-terminal NLSs of Gag, using the two unique restriction sites AvrII and BsrGI of the Tf1-neoAI assay plasmid pHL449-1. The two plasmids pHL1724 and pHL1726 were the Tf1-neoAI assay plasmid pHL449-1 with the sequence for the C-terminal NLS C1 (RKPKK) converted into either RAPAK or AAAAK, respectively. A sequence encoding a two-alanine replacement of the NLS C2 (KKRR converted into AARR) was also created within the Tf1-neoAI assay plasmid to generate the plasmid pHL1728. The plasmid pHL1837 contained the pHL449-1 backbone and the combination of sequences encoding a quadruple mutation of the NLS C1 (as in pHL1726) and a double mutation of the NLS C2 (as in pHL1728). The NLS-less version of Tf1 was generated as follows. The two unique restriction sites XhoI and AvrII in the plasmid pHL891-19 were used to insert a fusion PCR product in which the first 10 amino acids of Gag were replaced by an NgoMI restriction site. Two complementary oligonucleotides encoding a FLAG epitope were cloned into the NgoMI site. The resulting plasmid, pHL1757, encoded a FLAG epitope with the sequence ADYKDDDDKG.

Construction of chimeric proteins. All of the green fluorescent protein (GFP) fusion constructs that were needed for the microscopy study were derived from the vector pSGP502 (39). To examine the ability of different portions of Gag to localize in the nucleus, different sections of Gag were PCR amplified and then fused in frame to the green fluorescence reporter gene GFP-LacZ that was carried on the vector pSGP502. For this purpose, the wild-type sequence of Tf1 that was carried on the plasmid pHL449-1 was used as a PCR template. The oligonucleotides that were used for PCR amplifications all contained a KpnI restriction site, so that the PCR products could be then digested with KpnI and subsequently cloned into the unique KpnI site that is located upstream of the GFP-LacZ sequence of the vector pSGP502. The sequences of the junctions between Gag and GFP were verified by sequencing. The Gag-GFP fusions that contained different mutant versions of the amino- and carboxyl-termini of Gag were constructed in the same manner, except that the template plasmids that were used for PCR amplification were either pHL1891, pHL1893, pHL1724, pHL1726, pHL1728, or pHL1837. The detailed information about the Gag sequences contained in each individual construct is given in Table 1. The expression of the Gag-GFP-LacZ constructs was driven by the ATG codon of Gag and an inducible nmt1 promoter.

Transposition and homologous recombination assays. Tf1 transposition frequencies were determined as described previously (29). Briefly, Tf1 transposition was monitored by placing a neo-marked Tf1 element under the control of the inducible nmt1 promoter. The neo gene allowed cells to grow in the presence of 500 µg of G418/ml; thus, Tf1 transposition activity could be correlated with the ability of cells to grow on G418-containing medium. S. pombe strains that contained a Tf1-neo plasmid were grown as patches on EMM-ura dropout agar plates in the absence of thiamine to induce transcription of the nmt1-Tf1-neo fusion. After 4 days of 32°C incubation, these plates were then replica printed to medium containing 5-FOA to eliminate the URA3-Tf1-neo plasmid (5). Finally, 5-FOA^r patches were printed to medium containing both 5-FOA and G418 and incubated at 32°C for 2 days to detect strains that became resistant to G418 as the result of insertions of the neo-marked Tf1 element into the genome (30, 31).

Isolation of Tf1 cDNA from whole-cell extracts. Total cDNA content was extracted from stationary-phase cells grown under inducing conditions (absence of thiamine). Approximately 10⁹ cells (100 optical density at 600 nm [OD₆₀₀] units) were used for each preparation. These preparations, as described previously (29), were used to generate materials for cDNA blot analysis of BstXI digests (2). To examine the accumulation of cDNA in the cells, the filters were hybridized with a 1.0-kb neo probe derived from a BamHI digest of pGH54 (24).

Protein extraction and immunoblot detection of Tf1 Gag and IN. Total proteins were extracted from cells grown under inducing conditions (absence of thiamine) by using a previously published protocol (2). Protein extracts were collected, and an equal volume of 2× sample buffer (2) was added. The mixture was boiled for 3 min, and 25 µg of total protein from each sample was loaded onto sodium dodecyl sulfate-10% polyacrylamide gel electrophoresis (SDS-10% PAGE) gels for immunoblot analysis. Standard electrotransfer techniques were used (47) with Immobilon-P membranes (Millipore). The detection method used was the ECL system as described by the manufacturer (Amersham), except that the secondary antibody, horseradish peroxidase-conjugated donkey anti-rabbit immunoglobulin, was used at a 1:10,000 dilution. The primary polyclonal antisera used for each filter were from production bleeds 660 (anti-Gag) and 657 (anti-IN) (31).

Indirect immunofluorescence. The anti-FLAG M2 monoclonal antibody (Eastman Kodak, New Haven, Conn.) and FITC-Oregon green 488 goat anti-mouse immunoglobulin G IgG antibody (Molecular Probes) were used for immunofluorescence experiments essentially as previously described (8). To visualize nuclear DNA, cells were stained with 1 µg of DAPI (4',6'-diamidino-2-phenylindole)/ml. Detection of FLAG-tagged Gag in intact cells was achieved by incubating mutant and wild-type cells containing the FLAG-tagged Gag under inducing conditions. Cells were mounted on glass slides with mounting solution (1 mg of p-phenylenediamine [catalog no. P-1519; Sigma], 1 µg of DAPI/ml in 50% glycerol). The cells were examined by using fluorescence microscopy with a Zeiss Axioscope equipped with UV and fluorescein isothiocyanate optics. Images were collected and imported into Adobe Photoshop 4.0.1 for figure presentation.

GFP microscopy. To visualize the cellular localization of GFP-LacZ fusion proteins, cells expressing GFP fusions were grown in selective media under inducing conditions (absence of thiamine) to an OD₆₀₀ of 0.3. Cells were then washed once with 1× PBS and resuspended in a 1× PBS solution that contained 2.5 µg of Hoechst dye/ml before being transferred to microscope slides. Cells were mounted onto the microscope slides using Vectashield (Vector Laboratories Inc., Burlingame, Calif.). GFP images were collected and imported into Adobe Photoshop 5.0.1 for figure presentation.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The Tf1 Gag protein contains two distinct karyophilic domains. Our previous studies showed that the Gag protein of Tf1 is localized in the nucleus of cells in stationary phase (3, 8). In this study, we first investigated the question of whether Gag contains any amino acid sequences with a karyophilic property. For this purpose, we examined the ability of domains of Gag to direct the GFP into the yeast nucleus. Gag was divided into five sections that were fused to the N terminus of a chimeric coding sequence that included GFP and LacZ (Fig. 1A). The molecular masses of these fusion proteins were approximately 140 kDa, and thus they were unlikely to enter the nucleus by passive diffusion. Their expression in yeast cells was under the control of the inducible nmt1 promoter. Immunoblot analysis using antiserum against LacZ showed that each fusion protein possessed its predicted molecular weight and was accumulated to similar levels in the cells (data not shown). Examination of the localization of GFP signals in live cells was performed with cells that were grown under inducing conditions and harvested at an OD₆₀₀ of 0.3. We found that two of the five domains of Gag contained the ability to cause nuclear localization (Fig. 1B). Fusions of residues 1 to 50 and residues 200 to 251 were found to possess nuclear localizing activity (Fig. 1B). As a control, cells expressing the GFP-LacZ reporter protein alone displayed GFP signals that distributed evenly throughout the cytoplasm. It is interesting to note that the LacZ fusion protein was not occluded from the nucleus. This suggests that small amounts of LacZ did enter the nucleus without the addition of an NLS. However, a significant concentration of LacZ fusion protein was localized in the nucleus when fused to the classical NLS from SV40 large antigen (Fig. 1B).

View larger version (64K): [in this window] [in a new window]   FIG. 1.   Gag contains two distinct karyophilic domains. (A) Construction of Gag-GFP-LacZ chimeric proteins. Shown is a schematic representation of the five domains of Tf1 Gag that were fused to the GFP-LacZ reporter. Roman numerals used to name the constructs correspond to the numbers used in panel B. The five domains of Gag with the location of their ATG codons are indicated as dark shaded boxes. The numbers above the shaded boxes indicated the residues of Gag that were contained in the corresponding fusion proteins. The GFP and LacZ sequences are represented by the hatched and white boxes, respectively. The two putative NLSs that were identified by a computer-based search are indicated by two vertical black boxes within the Gag sequence. The expression of the fusion proteins was under the control the inducible nmt1 promoter. (B) Ability of the five domains of Gag to direct the GFP-LacZ reporter proteins into the nucleus. The localization of GFP fusion proteins was then examined to determine the ability of the corresponding Gag residues to direct the protein into the nucleus (top panels). In the control strains, we included GFP-LacZ with or without the NLS of SV40. DNA was counterstained using Hoechst dye (bottom panels). Bar, 10 µm.

View larger version (45K): [in this window] [in a new window]   FIG. 2.   Mutational analysis of the NLS activity identified in the C-terminal and the N-terminal domains of Gag. The sequences of the last 50 residues (top panel) and the first 50 residues (bottom panel) of Gag were fused to GFP-LacZ. The hatched and white boxes represent the GFP and LacZ sequences, respectively. The two putative NLSs that have been identified by a computer-based search are named C1 and C2. Their corresponding amino acid sequences are also underlined (RKPKK and KKRR, respectively). The NLS in the N-terminal domain of Gag, KRIR, is underlined. The black circles denote the positions of alanine substitution mutations that were created within the NLS sequences. The cellular localization of the corresponding GFP fusion proteins are indicated by either N (nuclear) or C (cytoplasmic). The first 10 residues of Gag contained karyophilic activity, but this was somewhat less efficient than that of the full-length 50 amino acids of the Gag N terminus. Its cellular localization is therefore indicated as N/c. aa, amino acid.

The NLS activity of the C terminus of Gag is due to two distinct amino acid sequences that were dispensable for Tf1 transposition. To determine whether the two NLSs predicted to be in the C terminus of Gag played a role in nuclear localization, we generated amino acid substitutions in the protein that consisted of residues 200 to 251 of Gag fused to the N terminus of GFP-LacZ. The original amino acid sequence of NLS C1, RKPKK, was converted to either RAPAK or AAAAK (Fig. 2, top). In a similar manner, the original sequence of NLS C2, KKRR, was converted to AARR. Results of fluorescence microscopy showed that the karyophilic property of this C-terminal fragment of Gag was not altered when either one of the two putative NLSs was mutated. Interestingly, the nuclear localization of the GFP-LacZ fusion was totally disrupted when both NLS C1 and C2 were simultaneously mutated (Fig. 2, top). Images of cells with the intact C-terminal domain and the version with substitutions in both C1 and C2 are shown in Fig. 3. The expression level of the Gag-GFP-LacZ fusions was measured by immunoblot analysis using LacZ antibodies and was found to be unaffected by the presence of the point mutations (data not shown).

View larger version (41K): [in this window] [in a new window]   FIG. 3.   Analysis of the NLS activity identified in the N-terminal and C-terminal domains of Gag. Shown is localization in wild-type cells of fusion proteins that contained GFP-LacZ and residues from the N-terminal and C-terminal domains of Gag. These are representative images for the proteins described in Fig. 2. The localization of the reporter was detected by green GFP signals. The headings indicate the amino acid (AA) residues of Gag that were fused to GFP-LacZ or the type of mutations that were made in the NLSs of the 50-amino-acid domains. The positions of nuclei were indicated by staining of DNA with Hoechst dye. Bars, 10 µm.

The NLS in the N terminus of Gag is in the first 10 residues, and it is absolutely required for Tf1 transposition. Since a computer-based search for NLSs present within the N-terminal domain of Gag revealed no obvious candidates, we investigated whether this fragment truly contained an NLS element. For this purpose, we conducted an extensive search for a minimal sequence required for nuclear localization. Various fragments of the N terminus of Gag were fused to the N terminus of GFP-LacZ (Fig. 2, bottom). As mentioned earlier, the first 50 amino acids of Gag were able to direct the Gag-GFP-LacZ protein into the nucleus. Progressive deletions of Gag sequence that originated at residue 50 and extended to residue 21 did not inhibit the nuclear localization of the fusion proteins. A GFP-LacZ protein that contained only the first 10 amino acid residues of Gag was concentrated in the nucleus but with somewhat less efficiency than proteins with the first 20 amino acids of Gag. This result indicated that important information for the nuclear localization was contained within the first 10 amino acids of Gag. The importance of these 10 amino acid residues for nuclear localization was further confirmed by two additional deletions in the protein with amino acids 1 through 50 of Gag. We found that the two Gag-GFP-LacZ fusions that lacked the first 10 or 20 residues of Gag were distributed evenly throughout the cytoplasm (Fig. 2, bottom). Figure 3 contains images of a representative set of the strains listed in Fig. 2.

The NLS in the N terminus of Gag was required for transposition but not particle assembly. We next investigated whether the NLS activity of the first 10 amino acid residues of Gag was required for Tf1 transposition. Here too Tf1 activity was monitored using the assay that detected the resistance to G418 caused by the insertion of a neo-marked Tf1 element into the genome of S. pombe cells (31). A brief description of the assay follows. The assay plasmid (pHL449-1) carried a copy of Tf1 (Tf1-neoAI) that was placed under the control of the inducible nmt1 promoter. Patches of cells harboring the plasmid with Tf1-neoAI were induced for transposition by activating the expression of the nmt1 promoter. Before the induced patches were tested for resistance to G418, the plasmid with Tf1-neoAI was subjected to counterselection by replica-printing cells to medium that contained 5-FOA. Thus, only cells that received a transposition event (conferring a G418^r phenotype) and subsequently lost the assay plasmid (conferring a 5-FOA^r phenotype) grew on medium that contained G418 and 5-FOA (3, 8). Figure 4A presents the results of a transposition assay and shows that a patch of cells that contained wild-type Tf1-neoAI produced confluent growth on a G418-5-FOA plate. A frameshift mutation in the N terminus of IN caused a significant drop in transposition (1). In addition, a frameshift mutation in PR blocked the expression of RT and IN and produced no resistance to G418. In the context of Tf1-neoAI, we engineered the N terminus of Gag such that the first 10 amino acids were replaced by the 10-amino-acid FLAG epitope. This newly created version of Tf1-neoAI was named NLS-less Tf1-neoAI. The transposition assay was performed to determine whether the absence of the NLS affected Tf1 activity. Indeed, as presented in Fig. 4A, the NLS-less version of Tf1 exhibited a significant defect in transposition compared to that of the wild-type Tf1.

View larger version (33K): [in this window] [in a new window]   FIG. 4.   Characterization of an NLS-less version of Tf1 in which the first 10 amino acids of Gag were replaced by a FLAG epitope. (A) Transposition assay based on G418r due to the insertion of Tf1-neoAI. Top row, three control strains were the wild-type strain (WT) containing either the wild-type Tf1-neoAI assay plasmid (YHL1282) or Tf1-neoAI with a frameshift created either in IN (YHL1554) or in PR (YHL1836) of Tf1. Bottom row, two independent yeast transformants of each of the two strains (YHL6773 and YHL6774) were tested. These strains contained two independent clones of the NLS-less version of the Tf1-neoAI assay plasmid. (B) The levels of Tf1 Gag and IN in strains that were wild type (lane 1, YHL1282) or contained the NLS-less version of Tf1 (lanes 4 and 5, YHL6773). The levels of Tf1 Gag and IN were also tested for a Tf1 version in which the N-terminal NLS in Gag was deleted (lanes 2 and 3). These immunoblots of S. pombe extracts were made from cells harvested either at log phase (left) or at stationary phase (right). The filters were probed simultaneously with anti-Gag and anti-IN antisera. The positions of Tf1 precursors, Gag, and IN are indicated by arrows. (C) Sucrose gradient analysis. The top panel is an immunoblot of an SDS-10% PAGE with fractions from a sucrose gradient that contained an extract from a wild-type Tf1-expressing strain (YHL1282). The antibody probe was anti-Gag serum. The bottom panel was also probed with anti-Gag serum and is an immunoblot of pooled fractions from a strain that expressed the NLS-less version of Tf1 (YHL6773). (D) Effects of the lack of the N-terminal NLS in Gag on the accumulation of cDNA in the cell. This DNA blot was used to measure the levels of Tf1 cDNA produced in the strain with the NLS-less version of Tf1(YHL6773). Three control strains were the wild-type strain containing versions of Tf1 that either were wild type (YHL1282) or contained a frameshift (fs) mutation in IN (YHL1554) or a frameshift mutation in PR (YHL1836). The DNA was digested with BstXI and probed with a neo-specific sequence. The 2.1-kb band was generated by the linear cDNA while the 9.5-kb band was produced by a vector sequence.

The NLS-less mutation in Tf1 inhibited the nuclear import of Gag. An indirect immunofluorescence study was conducted to determine whether the Gag protein produced by Tf1-neoAI with the NLS-less mutation was imported into the nucleus. As a control, we used a functional FLAG-tagged allele of wild-type Tf1 that contained a FLAG epitope inserted near the C terminus of Gag. We previously showed that this control version of Tf1, named Tf1(FLAG)-neoAI, expressed normal levels of Tf1 proteins and cDNA and possessed wild-type levels of transposition activity (3, 8). The nuclear import of Tf1 Gag was originally characterized with this strain (3, 8). Wild-type cells that were induced for expression of either NLS-less Tf1 or Tf1(FLAG)-neoAI were subjected to immunofluorescence staining with the use of the M2 anti-FLAG monoclonal antibody. Consistent with the data reported previously (3, 8), we found that the majority of cells with wild-type Tf1 produced a single primary focus of Gag signal (green) within the nucleus, as demonstrated by the colocalization of the FLAG-Gag signals and nuclear staining (black) (Fig. 5A). No FLAG signal was observed from cells that were not induced for Tf1 expression (data not shown). A totally different scenario was observed in the strain with the NLS-less version of Tf1-neoAI. FLAG-Gag signals were distributed evenly throughout the cytoplasm. In addition, positions corresponding to cell nuclei lacked FLAG-Gag signals, suggesting that the import of Gag protein was totally disrupted in these cells.

View larger version (126K): [in this window] [in a new window]   FIG. 5.   (A) Immunofluorescence analysis of Gag expressed by intact Tf1. Top, immunofluorescence of cells expressing wild-type Tf1. The wild-type strain YHL5896 contained a FLAG-Gag version of the Tf1-neoAI plasmid (pHL1277). In the top panels, the green fluorescein isothiocyanate (FITC) signals are specific for the FLAG-Gag protein and the blue signals indicate the locations of nuclei counterstained with DAPI. The panel in the upper right position is a merge of the FLAG-Gag signals produced by YHL5896 with an inverted black and white image of its DAPI stain. The merge was generated with Adobe Photoshop 4.0 with the screen function set at 65% opacity. Bottom, the same experiment was conducted except that the yeast strain (YHL6773) contained the NLS-less version of Tf1 (pHL1757). Bar, 10 µm. (B) Effects of the lack of the N-terminal NLS in Gag on cDNA homologous recombination. This cDNA recombination assay was used to detect the presence of cDNA in the nucleus. The same yeast strains that were used for the transposition assay presented in Fig. 4A were subjected to cDNA homologous recombination. These strains contained two independent clones of the NLS-less version of the Tf1-neoAI assay plasmid. Three control strains were the wild-type strain (WT) containing either the wild-type Tf1-neoAI assay plasmid (YHL1282) or Tf1-neoAI with a frameshift mutation created either in IN (YHL1554) or in PR (YHL1836) of Tf1.

The NLS activity of Gag is required for the uptake of Tf1 cDNA into the nucleus. Our previous studies demonstrated that the Gag protein, together with cDNA and other Tf1 proteins, forms large macromolecular structures named VLPs (28, 31). We hypothesized that if the nuclear import of Tf1 Gag was defective in the absence of its N-terminal NLS, then the import of Tf1 cDNA may also be altered in this mutant. To investigate this possibility, we used a cDNA recombination assay that was previously developed to measure the presence of cDNA in the nucleus (1). This assay measures homologous recombination between copies of Tf1 cDNA and Tf1 plasmid sequences. The Tf1-neoAI in the transposition assay plasmid contained an artificial intron in the reading frame of neo. The intron was in the opposite orientation of neo and because it could not be spliced from the neo mRNA, it inactivated the neo gene. However, the intron can be spliced from the Tf1 mRNA. Once the intron is spliced from the Tf1 mRNA and reverse transcription is complete, the cDNA is carried into the nucleus, where it recombines with its homologous sequence in the assay plasmid. This generates cells that are resistant to G418. Wild-type levels of this recombination indicate that normal levels of Tf1 cDNA are produced and that this cDNA is transported into the nucleus. The results presented in Fig. 5B show that the NLS-less version of Tf1-neoAI displayed a reduction in the homologous recombination of Tf1 cDNA compared to that of the wild-type strain or the strain with a frameshift in IN that shows the normal level of homologous recombination in the absence of any transposition. Since we found that the NLS-less mutation did not reduce the amounts of cDNA produced, the homologous recombination assay indicated that the absence of the N-terminal NLS in Gag reduced the nuclear uptake of Tf1 cDNA.

The nuclear pore factor Nup124p was specifically required for the NLS activity associated with the N-terminal domain of Gag. We previously showed that Nup124p, an FXFG nuclear pore factor, possesses a specialized activity that is specifically required for the nuclear localization of transposon material and not other proteins tested (3). In this experiment, we investigated whether the nuclear import of the Gag-GFP-LacZ fusion proteins required Nup124p. For this purpose, we determined whether the nuclear localization of GFP-LacZ fusion proteins that contained the NLS activities in either the N-terminal or the C-terminal domains of Gag was affected by the truncation of Nup124p protein encoded by the nup124-1 allele. This allele, isolated by random mutagenesis, caused a severe defect in Tf1 transposition and the nuclear import of Gag (3). As a control in this experiment, we also used a GFP-LacZ fusion that contained the classical NLS from SV40 large T antigen (Fig. 1B). This nuclear localization activity was not affected by the truncation of nup124-1 (Fig. 6A). Similarly, the nuclear localization of the C terminus of Gag (amino acids 200 to 251) was unaffected in the strain with the mutation in nup124-1 (Fig. 6A). Interestingly, the NLS activity of the N terminus of Gag (residues 1 to 50) was disrupted in the strain with the nup124-1 mutation. The GFP signals spread randomly throughout the cytoplasm. This finding indicated that the nuclear pore factor Nup124p played an important and specific role in the activity of the NLS that we mapped in the N-terminal domain of Gag.

View larger version (69K): [in this window] [in a new window]   FIG. 6.   The requirement of Nup124p for the function of the NLSs in Gag. (A) Nuclear localization of the N- and C-terminal domains of Gag fused to GFP-LacZ expressed in cells with the nup124-1 mutation. These images are a representative sample of the proteins tested in panel B. The localization of the reporter was detected as green GFP signals. The positions of nuclei were indicated by staining of DNA with Hoechst dye. As a control, we used a GFP-LacZ fusion protein that contained the classical NLS from SV40 large T antigen. Bar, 10 µm. (B) Summary of the localizations of the proteins with residues from the N-terminal domain of Gag as expressed in cells with the nup124-1 mutation. The positions are indicated for the residues with NLS activity and the residues that impose Nup124p dependence on import. AA, amino acid; WT, wild type.

View larger version (54K): [in this window] [in a new window]   FIG. 7.   Residues 10 to 30 of Gag confer the requirement of Nup124p on the nuclear import of a heterologous protein with the NLS of SV40. (A) Summary of localization data. Fusion proteins consisted of GFP-LacZ fused to the C termini of the SV40 NLS, the Tf1 NLS and Gag10-30, the SV40 NLS and Gag10-30, or the FLAG tag and Gag10-30. The localization of the GFP fluorescence is indicated for strains with wild-type nup124 (WT) and with nup124-1. (B) Localization of fusion proteins that contained Gag10-30. Strains that had wild-type nup124 or nup124-1 alleles expressed the fusion proteins FLAG-(Gag10-30)-GFP-LacZ or SV40 NLS-(Gag10-30)-GFP-LacZ. The localization of the reporter was detected as green GFP signals. The positions of nuclei were indicated by staining of DNA with Hoechst dye. Bar, 10 µm.

View larger version (41K): [in this window] [in a new window]   FIG. 8.   A version of Tf1 with the NLS of SV40 possessed high levels of transposition activity. The NLS of Tf1 Gag was replaced with the NLS of SV40, and the resulting transposon was tested for transposition activity. The level of growth of each strain on medium containing G418 corresponded to the transposition activity of each element. Top row, three control strains were the wild-type strain (WT) containing either the wild-type Tf1-neoAI assay plasmid (YHL1282) or Tf1-neoAI with a frameshift (fs) created either in IN (YHL1554) or in PR (YHL1836) of Tf1. Second row, four independent transformants of Tf1(SV40 NLS)-neo in a strain with nup124+. Third row, four independent transformants of Tf1(SV40 NLS)-neo in a strain with nup124-1. Bottom row, four independent transformants of wild-type Tf1-neo in a strain with nup124-1.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of multiple NLSs within the Gag protein of Tf1. The transport of retroelement material across the nuclear membrane is a critical step that precedes the integration of these elements into the host genome. The question of how Tf1 enters the nucleus is especially interesting because in yeasts the NE remains intact during mitosis (26). We reported previously that the Tf1 element imports its material into the nucleus by a unique mechanism that has a requirement for the nuclear pore factor Nup124p (3). In this report, we investigated the nuclear import of the Gag protein particularly because of our interest in understanding this unusual dependence on the nuclear pore factor Nup124p (3). The examination of different portions of Gag for their ability to localize GFP-LacZ in the nucleus identified two distinct domains with NLS activity. These two karyophilic regions consisted of 50 amino acid residues and were located at the amino and carboxyl termini of Gag. Furthermore, by conducting a computer-based search, we identified within the C-terminal karyophilic domain two putative NLSs with high similarity to the pattern 4 type of the classical NLS found in the T antigen of SV40. This type of NLS is composed of four basic amino acids (K or R) or composed of three basic amino acids (K or R) and either H or P. By making point mutations within these NLSs, we showed that the basic residues were critical for their NLS activity. However, we were surprised that the NLS activity in the C-terminal domain was dispensable for Tf1 transposition. One formal possibility is that the transposition assay we developed in the laboratory could in some way bypass the need for the NLSs that would ordinarily be required in nature. It is also possible that these C-terminal NLSs are folded into particle structures that are not accessible to transport factors.

The requirement of Nup124p for nuclear localization of Tf1. The FXFG repeats of the type present in Nup124p have been proposed to play a role in mediating the docking of the transport receptors known as importins or karyopherins to the NPCs. Nup124p is unique in that it has been shown to possess a specific activity that is required for the nuclear localization of Tf1 material and not other proteins tested (3). Interestingly, we found that an N-terminal domain of Gag was sufficient to reconstitute the requirement of Nup124p for nuclear import. Further analysis of peptides fused to GFP-LacZ revealed the surprising result that the residues of Gag with the NLS activity were independent from the residues that conveyed the requirement for Nup124p. The ability of residues 10 to 30 of Gag to impose Nup124p dependence on a heterologous NLS was demonstrated with a protein that had the SV40 NLS followed by Gag(10-30)-GFP-LacZ. The ability of residues in Gag to impose the requirement of Nup124p on a heterologous NLS was also demonstrated in the context of the intact transposon. A version of Tf1 with the SV40 NLS in place of the NLS in residues 1 to 10 had wild-type transposition activity that retained its requirement for Nup124p. Within the context of the current understanding of nuclear import, NLSs are known to serve as interaction sites for importins. Our results include the important and novel finding that an NLS and its associated residues can direct nuclear import through a specific pathway that is dependent on a particular factor of the nuclear pore.

View larger version (19K): [in this window] [in a new window]   FIG. 9.   Two possible mechanisms for the nuclear import of Gag. (A) In general, cargo to be imported interacts with the transport receptors, importins. Since Nup124p appears to be specifically required for the import of Gag, we propose that Gag interacts with specific importins to be carried to the nuclear pore. The interaction between this complex and a specific FXFG nuclear pore factor, Nup124p, could be required for translocation. (B) Alternatively, we propose that Gag may interact directly with Nup124p and that this interaction is required for the subsequent translocation. We have reported such an interaction and proposed that it is responsible for the specific function of Nup124p that is required for Tf1 import (3).