Protein A Tagging and Gene Replacement

Protein A tagging and gene replacement of SXM1 with HIS3 were performed using published procedures (Aitchison et al., 1995). For protein A tagging, a PCR product was generated that contained a protein A, HIS3, URA3 cassette flanked by the final 60 nucleotides of the coding region and 60 nucleotides downstream of the stop codon. SXM1 was replaced with HIS3 by generating a PCR product corresponding to the HIS3 gene flanked by 60 nucleotides directly upstream and 60 nucleotides directly downstream of the coding region of SXM1. Yeast cells were transformed by electroporation. Correct integration was verified by PCR. Heterozygous diploids were sporulated and dissected to generate haploid strains.

Whole Yeast Protein Extracts

Protein extracts were prepared using a method similar to a published procedure (Yaffe and Schatz, 1984). Briefly, 1 ml of a saturated, overnight YPD culture was added to 140 μl of 1.85 M NaOH/7.4% 2-mercaptoethanol. After 2 min at room temperature, 140 μl of 50% (wt/vol) trichloroacetic acid was added. The mixture was spun for 5 min at 4°C. After aspiration of supernatant, the pellet was washed with 95% ethanol and dried. The pellet was then resuspended in 100 μl of sample buffer and heated at 95°C for 5 min. After centrifugation, 1–5 μl were subjected to SDS-PAGE.

Immunofluorescence

Indirect immunofluorescence was carried out as described (Rout et al., 1997). Briefly, yeast cells were fixed with 3.7% formaldehyde for 5 and 20 min. After digestion of the cell wall, spheroplasts were attached to glass slides. Protein A moieties were visualized by probing with rabbit IgG that had been preadsorbed to wild-type yeast spheroplasts (Cappel Laboratories, Malvern, PA) followed by Cy3-conjugated donkey anti–mouse IgG (Jackson ImmunoResearch, West Grove, PA). Nuclear DNA in fixed cells was visualized with 4′,6-diamidino-2-phenylindole (DAPI). Direct immunofluorescence was used to visualize live, unfixed cells transformed with a plasmid encoding an SV-40 NLS–green fluorescent protein (GFP) fusion (Shulga et al., 1996).

Cell Fractionation and Immunoisolation

Fractionation and immunoisolation of Sxm1–protein A (PrA) were as described (Aitchison et al., 1996). Briefly, postnuclear, postribosomal cytosol from 160 ml of a YPD culture with an OD₆₀₀ of 1.6 was prepared. Rabbit IgG–Sepharose (Cappel Laboratories) was added to the cytosol. After overnight incubation, the Sepharose was extensively washed, and bound proteins were eluted with a step gradient from 50 to 4,500 mM MgCl₂. For mass spectrometry and microsequencing, eluted fractions were precipitated and resolved on a 10–20% acrylamide gel (Novex, San Diego, CA).

Northern Blot

Northern blot to analyze pre-tRNA processing was as described (Yoo and Wolin, 1997) with minor modifications. Briefly, total RNA from ~3 × 10⁷ cells was isolated using Tri Reagent (Sigma Chemical Co., St. Louis, MO) according to the manufacturer's instructions. This RNA was separated on a 10% TBE-Urea gel (Novex) and transferred in a tank apparatus to Hybond-N (Amersham Corp., Arlington, Heights). To assess overall RNA integrity, the samples were loaded in duplicate, and the half that was not transferred was stained with ethidium bromide and visualized with UV light. Oligonucleotides (Yoo and Wolin, 1997) and hybridization conditions (Tarn et al., 1995) have been described.

Genomic Tagging of SXM1 and Localization of Tagged Sxm1p

The carboxy terminus of the genomic copy of SXM1 was directly fused, in frame, to the IgG-binding domains of Staphylococcus aureus PrA via integrative transformation into diploid cells. Sporulation and tetrad dissection yielded four viable colonies on YPD plates. The four strains produced grew at the same rate at 16, 30, and 37°C. Two of the four strains grew on plates lacking His and Ura. These strains also expressed a 140-kD protein reactive with rabbit IgG by immunoblotting, as expected for a fusion protein composed of Sxm1p and protein A.

The resulting fusion protein (Sxm1–PrA) was used to determine the subcellular distribution of Sxm1p by indirect immunofluorescence microscopy. Using a fixation time of 5 min, Sxm1–PrA appeared predominantly in the nucleus (data not shown). This nuclear concentration is consistent with that seen for an Sxm1p–GFP fusion as reported by Seedorf and Silver (1997). After more prolonged fixation, however, the cytoplasmic pool of Sxm1–PrA appeared at least as concentrated as that in the nucleus (Fig. ​(Fig.2).2). Localization of Sxm1p in both the nucleus and cytoplasm is consistent with a role for it in nucleocytoplasmic transport.

Open in a separate window

Figure 2

Sxm1–PrA is localized throughout the cell. The PrA moiety was visualized by indirect immunofluorescence. For reference, the right panel represents the nuclei of the same field of cells by staining for DNA with DAPI.

Immunoisolation and Identification of Sxm1p-associated Proteins

If Sxm1p is indeed a factor involved in transporting newly synthesized proteins from the cytoplasm to the nucleus, some portion of the cytoplasmic pool of Sxm1p can be expected to be complexed with its nucleus-bound cargo. As such, we should be able to immunoisolate this transport complex from the cytoplasm of our Sxm1–PrA strain by virtue of the high affinity of PrA for IgG (Aitchison et al., 1996; Rout et al., 1997). We fractionated the Sxm1–PrA strain into crude nuclear and cytoplasmic fractions and subjected the cytoplasmic fraction to IgG-Sepharose chromatography. Four major Sxm1–PrA–associated proteins were eluted using either low pH (data not shown) or a magnesium chloride gradient (Fig. ​(Fig.3).3). The associated proteins eluted in the 100 and 250 mM MgCl₂ fractions, whereas the Sxm1–PrA/IgG interaction was not disrupted until ~1,000 mM MgCl₂. By Coomassie blue staining, there appeared to be similar concentrations of the associated proteins as compared with Sxm1–PrA, indicating that much of the cytosolic Sxm1–PrA exists complexed to these proteins.

Open in a separate window

Figure 3

Immunoisolation of cytosolic Sxm1–PrA interacting proteins. Whole cytosol from an Sxm1–PrA strain was subjected to IgG-Sepharose chromatography. Fractions from the final wash and MgCl₂ gradient were separated by SDS-PAGE. Molecular mass standards are indicated to the left in kilodaltons. The gel was stained with Coomassie blue R.

After SDS-PAGE and staining with Coomassie blue, associated protein bands were individually excised from the gel and digested with endoprotease Lys-C. The resulting peptides were subjected to microsequencing (Fernandez et al., 1994) and/or mass spectrometry (Gharahdaghi et al., 1996). The 38-kD band was identified as Lhp1p, which is the yeast homologue of the human autoantigen La. This assignment was made by sequencing the peptide VIEALRSSEILEVSADGENV, which corresponds to amino acids 83–102 of Lhp1p. From the 22-kD band, the peptide EQLSGQTPVQ was sequenced, corresponding to amino acids 38–47 of the 60S ribosomal protein Rpl16p. Two other ribosomal proteins, Rpl25p and Rpl34p, were identified through mass spectrometry of the doublet centered at 15 kD.

Lhp1p, Unlike Other Nuclear Proteins, Is Mislocalized in a Strain Missing Sxm1p

An Sxm1p-deficient strain was generated by replacement of SXM1 by the HIS3 gene. After sporulation and tetrad dissection, all four spores were viable on YPD plates. No growth phenotype was observed for the sxm1::HIS3 strain at any temperature. The absence of a growth phenotype is consistent with a previous study (Seedorf and Silver, 1997). Lhp1p, like the La protein of vertebrates (Simons et al., 1996), is concentrated in the nucleus at steady state. Like SXM1, LHP1 is nonessential for growth (Yoo and Wolin, 1994; Lin-Marq and Clarkson, 1995). We characterized the impact of SXM1 deletion on nuclear import of Lhp1p. As we had identified Lhp1p as a potential import substrate for Sxm1p, a dependence on Sxm1p for nuclear localization of Lhp1p would secure the definition of Sxm1p as a karyopherin. To examine the localization of Lhp1p, we constructed genomic Lhp1–PrA fusions in both a wild-type strain and in our sxm1::HIS3 strain. As expected, Lhp1–PrA colocalized with DAPI-stained DNA in the nucleus of the wild-type strain (Fig. ​(Fig.44 a, top). In marked contrast, in the sxm1::HIS3 strain, Lhp1-PrA was mislocalized to the cytoplasm, and in fact appears to be excluded from the nucleus in many cells (Fig. ​(Fig.44 a, bottom). Importantly, our sxm1 strain retained its ability to import the endogenous nuclear protein Npl3p as well as a reporter bearing the SV-40 large T nuclear localization sequence (Fig. ​(Fig.44 b). From these results, it is clear that Sxm1p is necessary for the nuclear import of Lhp1p and that this import operates in parallel to other nuclear import pathways.

Open in a separate window

Open in a separate window

Figure 4

Deletion of SXM1 leads to mislocalization of Lhp1-PrA but not other nuclear proteins. (a) Wild-type haploid cells (top row) and sxm1 cells (bottom row), both with a genomic Lhp1-PrA fusion, are probed for Lhp1-PrA (left) and stained for DNA with DAPI (right). (b) Wild-type cells (top row) and sxm1 cells (bottom row) were probed for either an NLS-GFP reporter (left) or the endogenous protein Npl3p (right).

Deletion of SXM1 Does Not Markedly Affect tRNA Processing

By virtue of their in vivo interactions with small RNAs, La and Lhp1p have been suggested to have a number of functions in RNA biogenesis (Gottlieb and Steitz, 1989; Bachmann et al., 1990; Meerovitch et al., 1993; Yoo and Wolin, 1997). The first direct effect of Lhp1p on small RNA biogenesis has recently been described. Lhp1p was shown to be essential in vivo for 3' endonucleolytic maturation of pre-tRNA (Yoo and Wolin, 1997). In the absence of Lhp1p, pre-tRNA was matured by an exonucleolytic pathway that does not appear to contribute to pre-tRNA maturation in wild-type cells (Yoo and Wolin, 1997). Maturation is important not only for the biogenesis of tRNA, but it is also a prerequisite for tRNA export, as unprocessed RNAs are retained in the nucleus (Haselbeck and Greer, 1993). To determine if our Lhp1–PrA fusion is functional and to characterize further phenotypes of the SXM1 deletion, we analyzed maturation of tRNA ^Ser _CGA in our Lhp1–PrA strains that are wild type or deleted for SXM1. By Northern analysis of total RNA, it is possible to discern three distinct, intron-containing pre-tRNA ^Ser _CGA species in wild-type cells (Fig. ​(Fig.5,5, A–C; D is properly processed tRNA ^Ser _CGA). The intermediates have 5′ and 3′ extensions (Fig. ​(Fig.55 A), only a 3′ extension (B), and no extension (C).

As previously described (Yoo and Wolin, 1997), deletion of LHP1 leads to an alteration in precursors A and B; in our assay, both are dramatically reduced in abundance (Fig ​(Fig5,5, compare lanes 1 and 2). When the data are quantitated and normalized for intermediate C, whose abundance is independent of LHP1 genotype (Yoo and Wolin, 1997) (Fig. ​(Fig.5,5, bottom left) the peaks representing intermediates A and B are both clearly smaller. Deletion of SXM1 did not drastically affect the appearance of these intermediates for this pre-tRNA, although a minor effect was seen in precursor B (Fig. ​(Fig.5,5, compare lanes 3 and 4; see also Fig. ​Fig.5,5, bottom right). It is possible that even in the absence of Sxm1p, some Lhp1p is brought into the nucleus with low efficiency by another pathway. As Lhp1p is likely to act catalytically, inefficient transport or diffusion of Lhp1p into the nucleus could account for the appearance of intermediates A and B in the SXM1-deficient strain.

A Nuclear Import Pathway for Lhp1p

The La autoantigen directly binds all nascent RNA polymerase III transcripts (Stefano, 1984). In vivo, La is found in ribonucleoprotein particles with many classes of small RNAs, including pre-tRNA and U6 RNA, as well as the 5S rRNP (Hendrick et al., 1981; Rinke and Steitz, 1982, 1985). Specific nuclear retention and export of several classes of RNAs have been suggested to be mediated by La (Guddat et al., 1990; Boelens et al., 1995; Grimm et al., 1997). Lhp1p, the yeast homologue of La (Lin-Marq and Clarkson, 1995, Yoo and Wolin, 1994), was recently shown to be essential for accurate pre-tRNA processing (Yoo and Wolin, 1997). In wild-type cells, the 3′ end of tRNA is determined by an endonucleolytic cleavage facilitated by Lhp1p (Yoo and Wolin, 1997). As LHP1 is nonessential, it is evident that an alternative, exonucleolytic pathway can also produce mature, exportable tRNA. The relative kinetics and downstream events from each of these pathways is unknown. As Lhp1p binds to all RNA polymerase III transcripts and not just pre-tRNAs, it is likely to have additional roles in RNA biogenesis (Yoo and Wolin, 1997). Many of these other functions of Lhp1p are likely to be carried out in the nucleus as well. As a result, factors influencing the import of Lhp1p may be used to modulate its activities.

The human La protein (hLa) has recently been shown to contain a nuclear localization sequence capable of leading to the import of a normally cytosolic reporter (Simons et al., 1996). The hLa NLS bears several similarities to the consensus bipartite NLS (Dingwall and Laskey, 1991; Makkerh et al., 1996). One major difference, however, is the presence of several acidic residues in the hLa NLS. This NLS may have been assumed to mediate import via the karyopherin α/β1pathway. However, extrapolating from our data in yeast, it may be assumed that mammalian cells also have a parallel import pathway for this highly conserved protein mediated by a Sxm1p-type karyopherin. In this light, it should be noted that all of the proteins that we isolated in complex with Sxm1p have homologues in higher eukaryotes.

A Backup Pathway for the Import of Ribosomal Proteins

Like KAP123, both SXM1 and LHP1 are nonessential in yeast. Even so, the benefits of the Sxm1p pathway have outweighed its metabolic cost. SXM1 was originally isolated in a screen for genes able to complement the conditional loss of Pse1p (Seedorf and Silver, 1997). Kap123p has recently been shown to be the primary carrier of ribosomal proteins into the nucleus, possibly backed up by Pse1p (Rout et al., 1997). Interestingly, Seedorf and Silver (1997) have generated a strain lacking Kap123p with a conditional allele of Pse1p that rapidly accumulates nuclear mRNA at 36°C. Overexpression of SXM1 in this strain complements its growth defect and partially restores normal mRNA export (Seedorf and Silver, 1997). On the basis of having a phenotype relating to a defect in the export of macromolecules from the nucleus, Sxm1p has been proposed to be an export factor (Ullman et al., 1997).

Our results do not address the possibility that Sxm1p may have a role in the export of macromolecules from the nucleus. However, from our data we can conclude that Sxm1p is necessary for the import of Lhp1p. In addition, it is possible that the Sxm1p pathway also is able to import ribosomal proteins Rpl16p, Rpl25p, and Rpl34p. It has recently been shown that various karyopherins overlap in their ability to import ribosomal proteins (Rout et al., 1997). In fact, one of the ribosomal proteins that we isolated in complex with Sxm1p, Rpl16p, has also been isolated in complex with Kap123p (Rout et al., 1997). It is therefore likely that in cells lacking a functional Sxm1p import pathway, Rpl16p, Rpl25p, and Rpl34p would be imported by Kap123p or Pse1p. Likewise, it is possible that the genetic interaction seen by Seedorf and Silver (1997) between KAP123/PSE1 and SXM1 results not from an overlap in their ability to export mRNA but from their ability to import ribosomal proteins. Indeed, the accumulation of mRNA in the nucleus is a common phenotype, shown in one study to affect 2% of all temperature-sensitive strains examined (Kadowaki et al., 1994). 16 complementation groups were found, and the screen does not appear to be saturated (Kadowaki et al., 1994). Various modulators of nucleocytoplasmic transport have been shown to be associated with mRNA export phenotypes, including components of the NPC and regulators of the small GTPase Ran (Kadowaki et al., 1994; Doye and Hurt, 1997). In many of these cases, the RNA export phenotype is not a primary phenotype.

We thank Joe Fernandez and Farzin Gharahdaghi of the Rockefeller University Protein/DNA Technology Center for protein sequencing and mass spectral analysis. We are grateful to Christopher Yoo and Sandra Wolin for providing the lhp1 and LHP1 strains, John Aitchison and Mike Rout for Kap95-PrA cytosol, and Ulf Nehrbass, Monique Floer, and Michael Rexach for bacterially expressed proteins. The advice and suggestions of members of the Blobel lab are gratefully appreciated.