Nab2p is required for poly(A) RNA export in Saccharomyces cerevisiae and is regulated by arginine methylation via Hmt1p.

From transcription to translation, mRNA is complexed with heterogeneous nuclear ribonucleoproteins (hnRNP proteins) that mediate mRNA processing, export from the nucleus, and delivery into the cytoplasm. Although the mechanism is unknown, export of mature mRNA from the nucleus is a critical regulatory step in gene expression. Analyses of hnRNP proteins have shown that many of these proteins are required for this essential cellular process. In this study, we characterize the Saccharomyces cerevisiae Nab2 protein, which was first identified as a poly(A) RNA-binding protein (Anderson, J. T., Wilson, S. M., Datar, K. V., and Swanson, M. S. (1993) Mol. Cell. Biol. 13, 2730-2741). Our work indicates that poly(A) RNA export from the nucleus is dependent upon a functional Nab2 protein; correspondingly, export of Nab2p from the nucleus is dependent upon ongoing RNA polymerase II transcription. Furthermore, we show that Nab2p is modified within its RGG domain by the type I protein-arginine methyltransferase, Hmt1p. Our experiments demonstrate that arginine methylation is required for the export of Nab2p from the nucleus and therefore establish an in vivo effect of this modification. Overall, these experiments provide evidence that Nab2p is an hnRNP protein that is required for poly(A) RNA export and whose export from the nucleus is regulated by Hmt1p.

The evolutionary divergence of prokaryotes and eukaryotes also marks the evolution of compartmentalization of the genetic material in the form of a membrane-bound nucleus. In eukaryotes, movement of macromolecules (i.e. proteins and RNA) between the nucleus and the cytoplasm occurs through the nuclear pore complex, which is embedded within the nuclear envelope (1,2). The presence of the nucleus also compartmentalizes the nuclear transcriptional machinery from the cytoplasmic translational machinery. mRNA serves as the linker molecule between these two cellular processes.
Active genes are first transcribed to pre-mRNA via RNA polymerase II and then processed within the nucleus to form mature mRNA transcripts (3). These processing events occur co-transcriptionally and include the addition of a 5Ј 7-methylguanosine cap (4), the splicing of introns (5), and cleavage of the 3Ј end followed by polyadenylation (4,6). Fully processed transcripts are then exported from the nucleus to the cytoplasm where they can be translated into functional proteins at ribosomes (7,8). Export of mRNA from the nucleus serves as an essential checkpoint in the regulation of gene expression. However, the detailed mechanism of the export of mature mRNA transcripts from the nucleus is poorly understood.
From transcription to translation, mRNA is bound by various heterogeneous nuclear ribonucleoproteins (hnRNPs) 1 that serve to regulate the mRNA life cycle (9). Approximately 30 human hnRNP proteins have been identified that have been implicated in various stages of mRNA processing and export (10). Some of these hnRNP proteins shuttle between the nucleus and the cytoplasm (11). They are first imported into the nucleus and later accompany their mRNA cargo from the nucleus to the cytoplasm. The mammalian hnRNPA1 protein is a well studied example of a shuttling hnRNP thought to function in the export of mRNA from the nucleus (11,12).
Due to the dynamic role that hnRNP proteins play in mRNA metabolism and transport, it is not surprising that these proteins must be highly regulated (10,13). Shuttling hnRNP proteins are imported into the nucleus where they bind to mRNA. They are then exported in complex with mRNA to the cytoplasm, where they release mRNA, and are finally re-imported into the nucleus to restart the cycle. Each of these steps requires regulation of hnRNP function, which may occur through protein-protein interactions and/or post-translational modifications. hnRNP proteins interact with the cellular transport machinery upon entry into and exit from the nucleus, and regulatory requirements must be satisfied in order for these interactions to occur. Furthermore, various hnRNP proteins have been reported to be methylated (14), phosphorylated (15,16), and glycosylated (17), all of which could affect proteinprotein interactions.
Recently, there has been a heightened interest in the regulation of protein function by arginine methylation (18,19). Many hnRNP proteins are methylated on arginine residues by protein-arginine methyltransferase 1, a mammalian type I arginine methyltransferase that catalyzes the asymmetric di-* This work was supported in part by National Institutes of Health grants (to A. H. C., X. Z., and X. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Both authors contributed equally to this work. § Emory University Minority Graduate Fellow supported by Emory University.
¶ Supported by a National Institutes of Health Biology Program training grant.
ʈ Sponsored by the Emory University Scientific Undergraduate Research Experience Program funded by the Howard Hughes Medical Institute.
** To whom correspondence should be addressed: Dept. of Biochemistry, Emory University, 1510 Clifton Rd., NE, Atlanta, GA 30322. Tel.: 404-727-4546; Fax: 404-727-3954; E-mail: acorbe2@emory.edu. methylation of arginine residues (14). The methyltransferase recognition sequence within hnRNP A1 has been characterized as (F/G)GGRGG(G/F), and studies of other substrates have suggested that the RGG (arginine-glycine-glycine) domain found in many RNA-binding proteins is the site of methylation (14,20). Additionally, protein-arginine methyltransferase 1 has recently been shown to modify both the STAT1 transcription factor (21) and histone H4 (22), with effects on proteinprotein interactions. The impact of this modification on the in vivo function of hnRNP proteins, however, is still unclear.
The type I arginine methyltransferases are conserved throughout all eukaryotic organisms. In Saccharomyces cerevisiae, over 85% of protein arginine methylation is catalyzed by the type I methyltransferase, Hmt1p (also known as Rmt1p) (23). Although HMT1 is not essential for cell viability, studies have shown that loss of HMT1 affects the dynamic localization of several yeast hnRNP proteins (24,25). Npl3p and Hrp1p, two shuttling yeast hnRNP proteins, are methylated by Hmt1p, and their export from the nucleus is regulated by arginine methylation (25). These examples suggest a role for arginine methylation in nuclear transport, specifically in the export of mRNA-binding proteins from the nucleus and possibly an indirect role in mRNA export.
The study of hnRNP proteins in S. cerevisiae has been useful in investigating the regulation and function of hnRNP proteins in mRNA metabolism and export. Many yeast hnRNP proteins, including Pab1p (poly(A)-binding protein 1), Pub1p (poly(U)binding protein 1), Npl3p (nuclear protein localization 3), Hrp1p (hnRNP-like protein 1), and Nab2p (nuclear polyadenylated RNA binding 2), were identified in a screen for proteins that cross-link to polyadenylated RNA in vivo (26,27). Identification in this screen suggested a role for these proteins in mRNA metabolism. Subsequent studies have confirmed the role of Pab1p in mRNA stability and in the control of poly(A) tail length (28), Pub1p in mRNA stability (29), and Hrp1p in 3Ј end cleavage and polyadenylation (30). Studies of Npl3p suggest its involvement in various aspects of mRNA processing and export from the nucleus (31)(32)(33).
The relationship between Nab2p and poly(A) RNA has not yet been investigated in detail. Experiments have shown that Nab2p binds directly to RNA homopolymers in vitro, suggesting a direct interaction between Nab2p and RNA (27). Consistent with such a hypothesis, the Nab2 protein contains three domains characteristic of RNA-binding proteins: a glutaminerich region, an RGG domain, and a novel zinc finger domain. The glutamine-rich region (residues 101-171), consisting of a QQQP tetrapeptide repeat, has been found in many eukaryotic transcription factors and RNA-binding proteins (34 -36). This domain is hypothesized to form a coiled-coil structure and to function in protein-protein interactions (37). The RGG domain (residues 205-229) of Nab2p is characteristic of many RNAbinding proteins and, as mentioned earlier, is thought to be a recognition sequence for arginine methylation (20). The RGG domain of Nab2p also overlaps with its experimentally defined nuclear targeting sequence (38). The zinc finger domain (residues 262-474) of Nab2p consists of a seven CX 5 CX 4 -6 CX 3 H repeat consensus sequence, which shows some similarity to the zinc finger domain of the largest subunits of RNA polymerases I-III (27,39). Both the identification of Nab2p as a poly(A)binding protein and its domain structure strongly suggest a role for Nab2p in the mRNA life cycle.
Like other hnRNPs involved in mRNA transport, Nab2p has a dynamic cellular localization. The steady state localization of Nab2p is nuclear (27); however, further analyses have indicated that Nab2p shuttles between the nucleus and the cytoplasm (38,40). Other studies (41) have shown that Nab2p is imported into the nucleus by the import receptor, Kap104p. Kap104p binds to residues 200 -249 of Nab2p, and this region has been shown to be sufficient for nuclear import (38). Little information is available, however, regarding the requirements for export of Nab2p from the nucleus.
Although it is established that Nab2p is a poly(A) RNAbinding protein that shuttles between the nucleus and the cytoplasm (27,38,40), the cellular function of Nab2p is not known. Our investigation seeks to analyze the export of Nab2p from the nucleus and its relationship with polyadenylated RNA. Our results indicate that Nab2p is necessary for efficient bulk mRNA export from the nucleus and that Nab2p requires active transcription of RNA polymerase II transcripts for its own nuclear export. We also show that Nab2p is methylated within the RGG domain by the arginine methyltransferase Hmt1p. In addition, we provide evidence that arginine methylation is required for efficient Nab2p export. These results indicate that Nab2p is a yeast hnRNP protein necessary for proper mRNA metabolism and/or export. Furthermore, they demonstrate that Nab2p export is regulated by active RNA polymerase II transcription and arginine methylation.

EXPERIMENTAL PROCEDURES
Strains, Plasmids, and Chemicals-All DNA manipulations were performed according to standard methods (42), and all media were prepared by standard procedures (43). All yeast strains and plasmids used are described in Table I. All chemicals were obtained from Sigma, United States Biological (Swampscott, MA), or Fisher unless otherwise noted.
Plasmid Construction-The pAC636 plasmid consists of the NAB2 gene with 400 bp upstream and downstream of the open reading frame (ORF) amplified from genomic DNA and cloned into pRS316 (URA, CEN) (44) at XbaI and XhoI sites. pAC717 contains the NAB2 sequence from pAC636 cloned into pRS315 (LEU, CEN) (44) via XbaI and XhoI sites. pAC719 contains the NAB2 ORF with 400 bp of upstream sequence cloned into pAC45 (URA, 2) via SmaI and XhoI sites to create a C-terminal GFP fusion. pAC753 contains the NAB2-GFP sequence from pAC719 cloned via SmaI into pRS315 (LEU, CEN). pAC757 consists of the NAB2 ORF with 300 bp of downstream sequence amplified from genomic DNA and cloned via BamHI and XhoI into pGEX4T-3 (Amersham Biosciences) to create an N-terminal GST-tagged protein. pAC785 contains the NAB2 sequence from pAC757 cloned via BamHI and XhoI sites into pET28a (Novagen) to create an N-terminal His 6tagged protein. pAC1100 contains the NAB2 sequence from pAC719 cloned via SmaI and XhoI into pAC496 to create a C-terminal c-myc(3ϫ) fusion protein. pAC1103 is a His 6 -HMT1 expression plasmid created by subcloning from pGEX-RMT1 (23) into a modified pET28b vector (Novagen) containing an N-terminal His 6 sequence.
Construction of the NAB2 Deletion-A complete deletion of the S. cerevisiae NAB2 ORF was created by using a PCR-based strategy (45). The HIS3 gene was amplified with 45 base pairs of flanking sequence homologous to the 5Ј and 3Ј ends of the NAB2 coding sequence. The resulting PCR product was transformed into the diploid yeast strain ACY247, and His ϩ transformants were selected. PCR was used to determine which transformants contained the desired replacement of NAB2 by HIS3. Heterozygous diploids were then transformed with a URA3 plasmid containing the wild type NAB2 gene (pAC636). Cells that were phenotypically His ϩ Ura ϩ were subsequently sporulated and subjected to tetrad dissection to generate the haploid deletion ACY429, where the viability of cells deleted for the essential NAB2 gene is maintained by the URA3 NAB2 plasmid (pAC636).
Expression and Protein Purification-The His 6 -NAB2 expression plasmid (pAC785) was transformed into the Escherichia coli strain BL21 (DE3). E. coli cells containing the His 6 -NAB2 plasmid were grown in LB media containing 50 g/ml kanamycin, and expression was induced (at A 600 ϭ 0.6) with 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside for 5 h at 30°C. Cells were harvested by centrifugation at 4000 rpm for 20 min at 4°C and resuspended in lysis buffer (20 mM Tris-HCl, pH 8.0, 0.1% igepal, 4 mM phenylmethylsulfonyl fluoride, 2 mM EDTA, 0.5 M NaCl). Cells were lysed via French press, and the soluble fraction was collected after centrifugation at 13,000 rpm for 30 min at 4°C. The soluble supernatant was loaded onto a HiTrap nickel chelator column (Amersham Biosciences), washed with Buffer A (50 mM Na 2 HPO 4 , pH 7.4, 0.25 M NaCl) containing 0.1% Triton X-100, and eluted with a stepwise gradient of 0.05, 0.15, and 0.5 M imidazole in Buffer A. The purity of eluted fractions was assessed by SDS-PAGE Coomassie staining, and purified His 6 -tagged Nab2 protein was detected by immunoblot analysis using an antibody specific to the His 6 tag. Fractions containing purified protein were dialyzed against 2 mM Tris-HCl, pH 8.0, and stored at Ϫ80°C in 10% glycerol. For Hmt1p purification, His 6 -Hmt1p (pAC1103) was expressed as described for the His 6 -Nab2 protein; however, cultures were induced with isopropyl-1-thio-␤-D-galactopyranoside overnight at 22°C. The protein was purified to homogeneity using HiTrap nickel chelating, HiTrap Q and Sephacryl S300 columns (Amersham Biosciences).
Production of Nab2p Antibody-Polyclonal antisera against purified His 6 -Nab2 protein were raised by immunizing New Zealand White rabbits (Covance). Rabbits were injected with the purified protein five times, using 250 g for the first injection and 125 g per subsequent injection (3 weeks between injections). Antisera were screened for their ability to detect specifically endogenous Nab2p in yeast extracts, His 6 -Nab2p expressed in E. coli lysate, and purified His 6 -Nab2 protein. The antibody was affinity-purified via 0.1 M glycine, pH 2.5, elution from a column consisting of epoxy-conjugated His 6 -Nab2p. The purified Nab2p antibody was diluted 1:50,000 for immunoblot analysis and indirect immunofluorescence.
Immunoblot Analysis-Cultures were grown to saturation (or as indicated) and then harvested by centrifugation at 3000 rpm for 3 min. Cell pellets were washed twice with sterile H 2 O and once in PBSMT (phosphate-buffered saline, 2.5 mM MgCl 2 , 0.5% Triton X-100). Cells were then resuspended in 500 l of PBSMT supplemented with protease inhibitors (0.5 mM phenylmethylsulfonyl fluoride and 3 g/ml each of aprotinin, leupeptin, chymostatin, and pepstatin). One volume of glass beads was added to each sample, and cells were lysed with 30-s pulses for 3 min in a bead beater (Biospec Products). The lysate was clarified by centrifugation at 13,000 rpm and assayed for total protein concentration using a Bio-Rad protein assay kit. Immunoblot analysis was performed as described by Towbin et al. (46). Nab2 protein was detected by a 1:50,000 dilution of the purified anti-Nab2p rabbit polyclonal antibody. Myc-tagged proteins were detected using a 1:200 dilution of the 9E10 mouse monoclonal c-Myc antibody (Santa Cruz Biotechnology). GFP-tagged proteins were detected using a 1:5000 dilution of an anti-GFP rabbit polyclonal antibody (47). GST-tagged proteins were detected by incubation with a 1:1000 dilution of an anti-GST (B-14) mouse monoclonal antibody (Santa Cruz Biotechnology). His 6tagged proteins were detected using a 1:5000 dilution of an anti-His 6 (H-15) rabbit polyclonal antibody (Santa Cruz Biotechnology).
Construction and Functional Analysis of the nab2-1 Mutant Allele-Standard mutagenesis techniques were used to generate conditional alleles of NAB2 (48). Several conditional alleles were identified from this procedure, one of which, nab2-1, was used for this analysis. To assess the function of the Nab2-1 mutant protein, a standard plasmid shuffle technique was performed (49). The nab2-1 allele on a centromeric LEU2 plasmid (pAC1038) was transformed into NAB2 deletion cells (ACY429) containing the URA3 NAB2 maintenance plasmid (pAC636). Cells were counted and serially diluted in sterile H 2 O to obtain 10,000, 1000, 100, 10, or 1 cell per 3-l volume. These dilutions were spotted onto complete media (YEPD) or media containing 5-fluoroorotic acid (5-FOA) which selects for cells that have lost the wild type URA3 maintenance plasmid (50). Growth of cells that express only the nab2-1 allele was then examined at 18 and 30°C. The plasmid shuffle technique was also used to assess the function of Nab2p-GFP (pAC753) and Nab2p-myc (pAC1100).
Fluorescence in Situ Hybridization-The intracellular localization of poly(A) RNA was assayed by fluorescence in situ hybridization (FISH) as described by Wong et al. (51). Briefly, cells expressing wild type Nab2p or the Nab2-1 mutant protein were grown to saturation at 30°C and subsequently diluted and incubated at 18°C overnight or at 30°C for 3 h. Cells were fixed with 5% formaldehyde (J. T. Baker Inc.), permeabilized with 300 g/ml zymolyase, and adhered to Teflon-coated slides (Cell Point) pretreated with 0.1% polylysine. Cells were subsequently permeabilized with 0.5% igepal, equilibrated with 0.1 M triethanolamine, pH 8.0, and incubated with 0.25% acetic anhydride to block polar groups. Cells were then incubated in prehybridization buffer and hybridized overnight to digoxigenin-labeled oligo(dT) probe (51). Wells were washed several times and then blocked in 0.1 M Tris, pH 9.0, 0.15 M NaCl, 5% heat-inactivated fetal calf serum, and 0.3% Triton X-100. Cells were incubated overnight in rhodamine-conjugated anti-digoxigenin antibody (1:200 dilution, Roche Molecular Biochemicals). Wells were then washed several times and cells were subsequently stained with 4Ј,6-diamidino-2-phenylindole-dihydrochloride (DAPI) (1 g/l) to detect chromatin.
Indirect Immunofluorescence and Microscopy-Indirect immunofluorescence was performed according to Wong et al. (51). Cultures were grown to log phase and then prepared and adhered to slides as described for FISH. Cells were then fixed to slides by methanol/acetone treatment, blocked with PBS/BSA (phosphate-buffered saline, 0.5% BSA), and then incubated with anti-GFP antibody (1:1000 dilution) or anti-Nab2p antibody (1:50,000 dilution) overnight. Samples were subsequently washed with PBS/BSA and incubated with fluoroisothiocyanate (FITC)-conjugated anti-rabbit secondary antibody (1:1000 dilution) for 2 h at room temperature. The wells were then washed and samples stained with DAPI (1 g/l). Finally, sample wells were washed several times with PBS and allowed to dry before addition of antifade (p-phenylenediamine in PBS, 90% glycerol). Samples were visualized using filters from Chroma Technology (Brattleboro, VT) and an Olympus BX60 epifluorescence microscope equipped with a photometric Quantix digital camera. For direct fluorescence, cells expressing GFP-tagged proteins were grown to 1 ϫ 10 7 cells per ml, and live cells were then examined for the GFP signal through a GFP-optimized filter. All images were captured using IP Lab Spectrum software. Nuclear Protein Export Assay-Nuclear export of Nab2p was examined as described previously (32) with some modifications. The Nab2p-GFP fusion protein (pAC719) was expressed in nup49-313 (ACY480), nup49-313 rpb1-1 (ACY436), or nup49-313 ⌬hmt1 (ACY437) yeast cells. A control NLS-LacZ-GFP plasmid (pAC613) was also analyzed in all strains to demonstrate nuclear integrity at the non-permissive temperature. Cells were grown to mid-log phase at 25°C in minimal media lacking uracil supplemented with 2% glucose. To inhibit protein synthesis, cycloheximide (100 g/ml) was added to each sample for 1 h at 25°C. Cultures were then either maintained at 25°C or shifted to 37°C for 5 h. Live cells were then examined for the GFP signal by direct fluorescence microscopy as described above.
Immunoprecipitation of Nab2p-The NAB2-myc plasmid (pAC1100) was transformed into either wild type (ACY402) or ⌬hmt1 (ACY577) cells. Cultures were grown to ϳ1 ϫ 10 7 cells per ml in minimal media lacking uracil supplemented with 2% glucose. Extracts were made as described for immunoblot analysis. Protein extracts were incubated overnight at 4°C with 10 l of agarose-conjugated anti-Myc (9E-10, Santa Cruz Biotechnology). The unbound fraction was collected, and the bound fraction was washed three times with PBSMT. Bound protein was eluted from the anti-Myc beads by adding 40 l of loading buffer (125 mM Tris-HCl, pH 6.8, 250 mM dithiothreitol, 5% SDS, 0.25% bromphenol blue, 25% glycerol). Unbound and bound fractions were resolved by SDS-PAGE and analyzed by Coomassie staining and immunoblot analysis.
In Vivo Methylation-The in vivo methylation assay conducted is a modification of a protocol described by Henry and Silver (24). Wild type (ACY402) and ⌬hmt1 (ACY577) cells expressing Nab2p-myc (pAC1100) or Npl3p-myc (pAC942) were grown to 1 ϫ 10 7 cells per ml at 30°C in minimal media lacking uracil supplemented with 2% glucose. Cycloheximide (150 g/ml) was then added to each culture followed by a 10-min incubation at 30°C. After this incubation, cells were pelleted and washed twice in minimal media lacking uracil and methionine supplemented with 2% glucose and cycloheximide (150 g/ml). Pellets were then resuspended in 1 ml of the same media supplemented with 150 Ci of either [ 35 S]methionine or [ 3 H]methionine. Cells were labeled for 90 min at 30°C and washed in PBSMT, and then protein extracts were collected, and Nab2p-myc or Npl3p-myc was immunoprecipitated as described above. Samples were then resolved by SDS-PAGE and analyzed via Coomassie staining, fluorography, and immunoblot analysis. Following Coomassie staining, the protein gel was dried and exposed to Kodak film at Ϫ80°C for 1-14 days. The signal from the 3 H-labeled samples was intensified using Amplify (Amersham Biosciences).

NAB2 Is Required for Poly(A) RNA Export-To test whether
Nab2p is required for the export of poly(A) RNA from the nucleus, it was necessary to generate a conditional allele of the essential NAB2 gene. Standard mutagenesis of the NAB2 ORF was used to generate a thermosensitive allele of NAB2, nab2-1. nab2-1 mutant cells grew more slowly than wild type cells at 30°C, and growth was severely impaired at 18°C indicating that this mutant displays a cold-sensitive phenotype (Fig. 1A).
Immunoblot analysis of protein extracts collected from nab2-1 cells grown at all temperatures indicated that the mutant protein is expressed and stable at all temperatures tested (data not shown).
To determine whether Nab2p is required for poly(A) RNA export from the nucleus, we examined poly(A) RNA localization by FISH analysis of nab2-1 cells. As shown in Fig. 1B, poly(A) RNA is distributed throughout the cell in wild type cells grown at either 30 or 18°C (Fig. 1B, panels A and D). In contrast, the nab2-1 cells show accumulation of poly(A) RNA within the nucleus following a shift to 18°C, as evidenced by co-localization with DAPI staining of chromatin (Fig. 1B, panels J and K). This mislocalization is observed even at 30°C where nab2-1 cells grow more slowly than wild type cells (Fig. 1B, panel G). It should also be noted that the mutant nab2-1 cells were significantly larger than wild type cells at both 30 and 18°C (Fig. 1B, panels C, F, I, and L). A similar requirement for Nab2p in poly(A) RNA export was observed when Nab2p was depleted from cells through glucose repression of a galactoseinducible NAB2 (data not shown). These results indicate that Nab2p is required for proper export of poly(A) RNA from the nucleus.
Localization of Nab2p-To analyze the dynamic intracellular localization of Nab2p, we first generated two reagents: a Nab2-GFP fusion protein and a polyclonal Nab2p antibody. For the GFP fusion protein, a plasmid shuffle technique was used to determine whether Nab2p-GFP could functionally replace the essential endogenous Nab2 protein ( Fig. 2A). Nab2p-GFP was FIG. 1. Characterization of the nab2-1 mutant. NAB2 deletion cells expressing either wild type NAB2 (pAC717) or the nab2-1 allele (pAC1038) were grown to saturation at 30°C before being subjected to either growth analysis or FISH. A, cells were counted and spotted onto 5-FOA, and growth was assessed at 30 and 18°C. B, cultures were shifted to 18°C overnight or maintained at 30°C. FISH was performed on fixed cells as described under "Experimental Procedures" (panels A, D, G, and J). Cells were stained with DAPI to localize chromatin within nuclei (panels B, E, H, and K). DIC images are also shown (panels C, F, I, and L).

Nab2p Is Required for mRNA Export and Regulated by Hmt1p
able to complement the lethal phenotype of a ⌬nab2 strain, suggesting that the Nab2-GFP fusion protein is functional in vivo. Immunoblot analysis using an anti-GFP antibody (data not shown) also indicated that the full-length Nab2-GFP fusion protein was expressed in these cells.
To facilitate the analysis of the endogenous Nab2 protein, a polyclonal Nab2p antibody was generated and affinity-purified (see "Experimental Procedures"). The specificity of the antibody was assessed by immunoblot analysis (Fig. 2B). The Nab2p antibody recognized a protein of ϳ72 kDa in wild type yeast cells, which corresponds to what has previously been reported for Nab2p (27) (Fig. 2B, lane 1). Furthermore, when cells expressed a functional Nab2-GFP fusion protein as the only copy of Nab2p, the antibody recognized a band of 100 kDa (Fig. 2B, lane 2) that corresponded both to the predicted size of Nab2p-GFP and to the signal observed when the extracts were analyzed with an anti-GFP antibody (data not shown).
We next examined the steady state localization of Nab2p. Consistent with previous reports (27,38,40), we observed nuclear localization for Nab2p-GFP in live cells (Fig. 2C, panels A and B), Nab2p-GFP in fixed cells probed with an anti-GFP antibody (panels C-E), and endogenous Nab2p in fixed cells probed with the anti-Nab2p antibody (panels F-H). Co-localization with DAPI-stained nuclei in fixed cells further confirmed that the localization of Nab2p at steady state is nuclear (Fig. 2C, panels C, D, F, and G). The data from Fig. 2 confirm that the Nab2-GFP fusion protein created is functional for proper cell growth and localization of Nab2p and also demonstrate that the antibody generated specifically recognizes Nab2p.
Active RNA Polymerase II Transcription Is Required for Nab2p Export-To analyze the export of Nab2p from the nucleus, a NUP49-based nuclear export assay was used (32). This assay utilizes a nup49-313 nucleoporin temperature-sensitive allele that exhibits a defect in protein import at the nonpermissive temperature of 37°C but has no effect on protein and RNA export from the nucleus (53,54). Specifically, import of Nab2p into the nucleus is significantly decreased in a nup49-313 strain at 37°C, and Nab2p-GFP is observed in the cytoplasm (data not shown). This genetic background, therefore, allows us to dissect effects on Nab2p import from those on Nab2p export.
Consistent with a previous report (40), we observed that Nab2p-GFP shuttles between the nucleus and the cytoplasm in the NUP49-based nuclear export assay (Fig. 3, panels A-D). This is indicated by the detection of Nab2p-GFP in the cytoplasm in the nup49-313 strain at the non-permissive temperature (37°C) (Fig. 3, panels C and D). A control NLS-GFP protein remains within the nucleus because it is not exported to the cytoplasm (Fig. 3, panels G and H). GFP signal in the cytoplasm is attributed solely to exported protein because translation of new protein is inhibited by cycloheximide in these experiments. Similar results, indicating that Nab2p-GFP is exported from the nucleus, were obtained in a heterokaryon shuttling assay (data not shown). These data confirm that Nab2p-GFP shuttles between the nucleus and the cytoplasm.
The data presented in Fig. 1B indicate that proper Nab2p function is required for efficient export of poly(A) RNA from the nucleus. We next wanted to investigate whether export of Nab2p from the nucleus is dependent upon the presence of poly(A) RNA and active RNA polymerase II transcription. The nuclear export assay described above was modified by combining the nup49-313 allele with the rpb1-1 temperature-sensitive allele (ACY436) in which RNA polymerase II transcription is inhibited at the non-permissive temperature, and consequently poly(A) RNA cannot be detected in these cells (32,55). The double mutant was then used to perform the nuclear export assay (Fig. 3, panels I-L). Although Nab2p-GFP was observed within the cytoplasm in the nup49-313 cells at 37°C indicating ongoing Nab2p export (Fig. 3, panels C and D), the nup49-313 rpb1-1 cells at 37°C showed only nuclear signal for Nab2p-GFP (Fig. 3, panels K and L). This suggests that active RNA polymerase II transcription is required for efficient Nab2p export from the nucleus and further indicates that Nab2p export is dependent upon ongoing poly(A) RNA synthesis.
Nab2p Is Methylated by Hmt1p in Vitro-To determine whether Nab2p is methylated by the S. cerevisiae arginine methyltransferase Hmt1p, an in vitro methylation assay was performed in which recombinant His 6 -tagged Nab2 protein was incubated with purified His 6 -tagged Hmt1 protein and the methyl donor [ 3 H]S-adenosylmethionine (AdoMet) (52). Reactions were analyzed by Coomassie staining (Fig. 4A) and fluorography (Fig. 4B). Recombinant Nab2p was methylated by Hmt1p in this assay as indicated by the 3 H signal observed (Fig. 4B, lane 6). Tritium incorporation is due to the transfer of the 3 H-labeled methyl group from AdoMet to Nab2p. A similar signal was also observed for His 6 -Nab2 protein expressed in E. coli lysate (Fig. 4B, lane 8); this signal is only observed in the presence of Hmt1p, as E. coli does not contain any proteinarginine methyltransferases (Fig. 4B, lane 7). The signal ob-FIG. 2. Localization of Nab2p. A, NAB2 deletion cells (ACY429) maintained by a wild type copy of NAB2 (pAC636) and expressing either wild type NAB2 (pAC717), NAB2-GFP (pAC753), or vector alone were spotted onto complete media (YEPD) or 5-FOA at 30°C as described under "Experimental Procedures." B, protein extracts (10 g of total protein) were collected from ⌬nab2 cells expressing either Nab2p (pAC717) (lane 1) or Nab2p-GFP (pAC753) (lane 2), and Nab2p was detected by immunoblot analysis with an anti-Nab2p antibody. C, cells expressing Nab2p-GFP were viewed by direct fluorescence microscopy (panel A) or fixed for indirect immunofluorescence using the anti-GFP antibody (panel C). The localization of endogenous Nab2p was also analyzed by indirect immunofluorescence using the anti-Nab2p antibody (panel F). Cells were stained with DAPI to localize chromatin within nuclei (panels D and G). DIC images are also shown (panels B, E, and H).
served at ϳ50 kDa in lanes 6 and 8 of the fluorograph in Fig. 4 is likely a degradation product of Nab2p, because a corresponding signal was not observed in E. coli lysate alone. The negative controls recombinant importin ␣ (Srp1p), which lacks the characteristic RGG domain, and Hmt1p (Fig. 4B, lanes 1 and 2) were not methylated. In contrast, methylation of recombinant hnRNPA1, which does contain an RGG domain and has been shown previously to be methylated in this assay (24), was detected in the presence of Hmt1p (Fig. 4B, lane 4). These data demonstrate that Nab2p is methylated by Hmt1p in vitro.
Nab2p Is Methylated by Hmt1p in Vivo-Although the previous experiment shows that Nab2p can be methylated by Hmt1p in vitro, it does not demonstrate that Nab2p is modified by Hmt1p in vivo. To test whether Nab2p is indeed an in vivo substrate of Hmt1p, an in vivo methylation assay was performed. To analyze Nab2p in vivo, however, it was first necessary to develop a method to purify specifically Nab2p from yeast lysates. A functional Nab2-myc fusion protein was therefore generated that could be specifically immunoprecipitated from wild type yeast cells using an ␣-myc antibody (data not shown). The myc-tagged protein immunoprecipitation was chosen for these experiments in order to create conditions under which we could isolate and directly compare results for Nab2p-myc with a control protein, Npl3p-myc (24). Isolation of these proteins by immunoprecipitation from whole cell lysates allows us to analyze the methylation status of these individual proteins.
The in vivo methylation assay utilizes a 3 H-labeled AdoMet methyl donor to detect arginine methylation, similar to the in vitro methylation assay. In the in vivo experiment, [ 3 H]AdoMet is generated by labeling cells with [ 3 H]methionine; methionine is then converted to AdoMet by the addition of adenosine. However, in vivo, methionine can also be incorporated into proteins through translation. It was therefore necessary to differentiate between 3 H labeling due to arginine methylation and 3 H labeling due to methionine incorporation through protein translation. In order to eliminate incorporation of the 3 H signal due to methionine incorporation via translation, we used cycloheximide to inhibit cytoplasmic protein synthesis. Cells expressing Nab2p-myc were grown in [ 35 S]methionine, where radioactive signal would be detected due to incorporation of methionine into proteins by translation (Fig. 5A). Nab2p-myc was immunoprecipitated from 35 S-labeled protein extracts that had been preincubated in the presence or absence of cycloheximide, and bound fractions were analyzed by autoradiography. As seen in Fig. 5A, radioactive signal was detected in the wild type and ⌬hmt1 cells that were not treated with cycloheximide but was not detected in cells treated with cycloheximide. This confirms that no incorporation of methionine through protein translation was observed under these conditions. Immunoblot analysis using the Nab2p antibody confirmed the presence of Nab2p in each of the samples (Fig. 5B). This preliminary experiment indicated that in subsequent experiments carried out under the same conditions any 3 H signal incorporated into Nab2p is due to protein methylation and not to methionine incorporation through protein translation.
To examine protein methylation in vivo, cells were incubated in media containing [ 3 H]methionine and cycloheximide, and Nab2p-myc was immunoprecipitated from 3 H-labeled protein extracts of either wild type or ⌬hmt1 yeast cells. Npl3p-myc served as a positive control for an S. cerevisiae protein that had been shown previously to be methylated in vivo (24). Bound fractions were analyzed by Coomassie staining (Fig. 5C) and fluorography (Fig. 5D). As seen in Fig. 5D, the 3 H signal was observed for both Npl3p-myc and Nab2p-myc from the wild type extracts (lanes 1 and 3, respectively) but not from the ⌬hmt1 extracts (lanes 2 and 4, respectively). The stronger signal observed for Npl3p is most likely due to a difference in the number of methylated arginine residues. Npl3p contains 15 RGG repeats, compared with only 4 RGG repeats in Nab2p. Coomassie staining (Fig. 5C) and immunoblot analysis with the anti-Nab2p and anti-Myc antibody (data not shown) confirmed the relative amount and presence of Npl3p-myc and Nab2pmyc in each of the samples. This result indicates that Nab2p is methylated in vivo and shows that this methylation is dependent on Hmt1p. Arginine Methylation Is Required for Efficient Nab2p Export from the Nucleus-To begin to assess the functional significance of arginine methylation of Nab2p, we analyzed whether loss of the arginine methyltransferase, Hmt1p, affected Nab2p export from the nucleus. To perform this experiment, the nup49-313 allele was combined with a disruption of the HMT1 gene (ACY437) (25). As mentioned earlier, the HMT1 gene is not essential for viability (24); however, as shown in Fig. 5D, Nab2p is not methylated in these cells. This double mutant was used in the nuclear export assay to test whether arginine methylation affects the export of Nab2p from the nucleus. As shown in Figs. 3C and 6C, Nab2p-GFP was observed within the cytoplasm in the nup49-313 cells at 37°C. In contrast, the nup49-313 ⌬hmt1 double mutant cells at 37°C showed only nuclear localization of Nab2p-GFP (Fig. 6, panel G). This demonstrates that arginine methylation is required for efficient Nab2p export from the nucleus.
Arginine Methylation of Nab2p Occurs within the RGG Domain-Since Nab2p contains the characteristic RGG domain (Fig. 7A) found in other arginine-methylated proteins, we next tested whether the RGG domain of Nab2p is the site of methylation by Hmt1p. To address this question, three constructs were tested for methylation by Hmt1p: full-lengthGST-Nab2p, GST-⌬RGG-Nab2p in which residues 201-264 had been deleted, and GST-RGG-Nab2p consisting of residues 180 -256 of Nab2p (Fig. 7A) (38). An in vitro methylation assay was performed on E. coli lysates containing full-length GST-Nab2p, GST-⌬RGG-Nab2p, or GST-RGG-Nab2p (Fig. 7B). E. coli lysate (lane 1) and GST alone (lane 2) served as negative controls for methylation. Expression of GST-tagged proteins was detected by immunoblot analysis using an anti-GST antibody (data not shown), and radioactive 3 H signal was detected by fluorography (Fig. 7B). As shown in Fig. 4 and lane 3 of Fig. 7B, the full-length Nab2 protein was methylated by Hmt1p in vitro. Deletion of the RGG region resulted in loss of 3 H signal (Fig.  7B, lane 5); the Nab2p-RGG region alone, however, was methylated (Fig. 7B, lane 6). These data indicate that the Nab2p RGG domain is both necessary and sufficient for arginine methylation. We therefore conclude that methylation of Nab2p occurs within the RGG domain. DISCUSSION Previous work (27) has shown that the Nab2 protein, which contains classical RNA-binding motifs, interacts with poly(A) RNA. The experiments presented in this study support the hypothesis that Nab2p is a classic shuttling hnRNP protein.
Our results indicate a functional relationship between the proper export of Nab2p and polyadenylated RNA from the nucleus. Specifically, we have shown that functional Nab2p is required for efficient poly(A) RNA export from the nucleus and, furthermore, that ongoing transcription of poly(A) RNA is necessary for export of Nab2p from the nucleus. Nab2p and poly(A) RNA are therefore dependent upon one another for their efficient export from the nucleus.
Although we have shown that Nab2p is required for proper localization of poly(A) RNA, specific conclusions regarding the role of Nab2p in mRNA export are premature. It is well documented that mRNA is not exported from the nucleus unless the transcript has been properly processed (8). Therefore, we cannot conclude that Nab2p has a direct effect on mRNA export. Nab2p could play an essential role in one of the maturation events of mRNA and have an indirect effect on mRNA export. Further investigation will be necessary to address this issue and pinpoint the specific function of Nab2p in vivo.
The data presented in this study functionally establish that Nab2p falls into a growing class of shuttling hnRNP proteins. Other members of this family in S. cerevisiae include Npl3p and Hrp1p. Npl3p may facilitate the export of mRNA from the nucleus (32), whereas Hrp1p has a biochemically characterized role in 3Ј end formation and polyadenylation of mRNA (30). Nab2p has characteristics that are both similar to and distinct from these proteins. Comparison of the primary amino acid sequences of Npl3p, Hrp1p, and Nab2p reveals little sequence homology outside the characteristic RGG domain present in many RNA-binding proteins. Npl3p and Hrp1p contain two central RNP-type RNA-recognition motifs (56 -58), whereas Nab2p contains a seven repeat zinc finger domain that has been shown to be crucial for RNA binding activity in vitro (27,39). This difference could suggest that Nab2p has a distinct specificity for its poly(A) RNA substrate or recognizes a specific sequence within a class of poly(A) RNA. Further analysis of the different RNA-binding motifs will be necessary to test this hypothesis of mRNA specificity.
Although this and previous work (25,71) have begun to delineate the requirements for the export of individual hnRNP proteins from the nucleus, a complete study of mRNA export must consider all proteins implicated in mRNA and RNP export from the nucleus. For example, other hnRNP-like proteins in S. cerevisiae that have been shown to affect poly(A) RNA localization include Yra1p and Yra2p (59,60). Yra1p and Yra2p are members of the RNA and export factor-binding (REF) family of hnRNP-like proteins. These proteins have RNA binding activity in vitro and were identified based on their association with the Mex67p export factor (59,60). Mex67p is the essential S. cerevisiae homologue of the human TAP protein and is thought to mediate interactions between mRNA complexes and the nuclear pore complex, as it is known to directly bind FG-containing nucleoporins (61)(62)(63). Mex67p cross-links to poly(A) RNA (62), but this interaction may be mediated by associations with other proteins, such as Yra1p and Yra2p, and not direct interactions with mRNA (59). The relationship between the Mex67-mediated mRNA export pathway and Nab2p has yet to be investigated. Preliminary results suggest that MEX67 affects the export of Nab2p from the nucleus. In order to understand better the complete picture of how mRNA is exported from the nucleus, further experiments will be necessary to link the export of Nab2p with this and other mechanisms of RNP export. Our results indicate that Nab2p is methylated at arginine residues within its RGG domain via the type I arginine methyltransferase, Hmt1p. We have analyzed the effect of arginine methylation on the regulation of export of Nab2p from the nucleus. Our data indicate that Nab2p export from the nucleus is dependent upon arginine methylation. There is evidence that at least two other S. cerevisiae hnRNP proteins, Npl3p and Hrp1p, are methylated by Hmt1p, and their export is blocked in the absence of HMT1 (25). It is therefore possible that the effect on Nab2p export in ⌬hmt1 cells is indirectly due to the inhibition of export of another protein rather than a direct effect of the loss of methylation of Nab2p. In order to address this possibility, it will be necessary to create specific mutations within the RGG domain of Nab2p that affect its ability to be methylated, and to subsequently test whether these mutations specifically affect proper function and export of Nab2p.
The functional significance of arginine methylation of hnRNP proteins is still unclear. It seems contradictory that the loss of type I arginine methylation via deletion of HMT1 inhibits the export of at least three essential shuttling hnRNP proteins, and yet the HMT1 deletion exhibits no poly(A) RNA export defect and has no apparent effect on cell growth or viability. One possible conclusion from these results is that the essential function of each of the hnRNP proteins is within the nucleus. This is consistent with the essential role that Hrp1p plays in 3Ј end formation and polyadenylation (30). One hypothesis that is consistent with the existing data in the literature is that Nab2p and other essential hnRNP proteins may mediate critical RNA processing events within the nucleus and become physically associated with the maturing hnRNP particle during this process. Association of the appropriate complement of hnRNP proteins would then serve as a signal that the mRNA had undergone the necessary processing steps and was now competent for export to the cytoplasm. In this model, it would not be essential for the hnRNP proteins to escort mRNA to the cytoplasm but merely to signal that the mRNA was ready to be exported. The shuttling of these hnRNP proteins could therefore be a consequence of their essential function within the nucleus. Further analysis of each of the hnRNP proteins will be required to determine whether shuttling between the nucleus and the cytoplasm is required for their essential cellular functions. The dependence of the export of Nab2p, Npl3p, and Hrp1p from the nucleus on arginine methylation suggests a common mechanism of export shared between each of these shuttling hnRNP proteins. However, evidence suggesting distinct forms of regulation of these proteins has also been reported. Tom1p, a putative ubiquitin ligase, has been shown to have a specific effect on the efficient export of Nab2p from the nucleus, with no apparent effect on the export of Npl3p (40). The function of Tom1p and its relationship to Nab2p is still under investigation. Additionally, the relationship between the three hnRNP proteins (Nab2p, Npl3p, and Hrp1p) also deserves further investigation. The similarities and differences shared between them will not only help in understanding their individual functions but also may give insight into the regulation of mRNA processing and export and the export of hnRNP proteins.
Export of mRNA from the nucleus is a critically important process that is poorly understood at the mechanistic level. Due to the fact that mRNA is known to be packaged as RNPs in vivo, advances in the study of hnRNP proteins can aid in the understanding of mRNA processing and nuclear export. In order to advance our knowledge in this field, attempts have been made to delineate the mechanism of mRNA export through the study of the export of RNA-binding proteins. S. cerevisiae has proven to be a useful model system for dissecting the steps of mRNA processing and export along with identifying factors essential for these processes (64). Our investigation into the requirements for the export of Nab2p, a S. cerevisiae hnRNP protein, has helped to advance our understanding of the requirements for mRNA export. Experiments to establish the in vivo function of Nab2p will FIG. 6. Export of Nab2p from the nucleus is blocked in the absence of Hmt1p. nup49-313 cells (panels A-D) or nup49-313 ⌬hmt1 cells (panels E-H) expressing Nab2p-GFP (pAC719) were analyzed in the nuclear export assay as described under "Experimental Procedures." Cells were treated with cycloheximide and maintained at 25°C (panels A, B, E, and F) or shifted to 37°C (panels C, D, G, and H). DIC images are also shown.
FIG. 7. The Nab2 protein is methylated within the RGG domain. A, schematic of the RGG domain and deletion mutants of Nab2p. Wild type fl-Nab2p is the full-length 525-amino acid Nab2 protein fused at the N terminus to GST. The black box designates the RGG domain of Nab2p, the sequence of which is outlined above the schematic presentation. The GST-tagged mutants include the ⌬RGG-Nab2p mutant, which lacks residues 201-264, and the RGG-Nab2p mutant, which consists only of residues 180 -256 (38). B, the in vitro methylation assay was performed on E. coli lysates containing GST alone (lane 2), fulllength GST-Nab2 protein (lane 3), GST-⌬RGG-Nab2p (lane 5), or GST-RGG-Nab2p (lane 6) and then analyzed via fluorography. also provide further insight into the mechanisms of mRNA metabolism and nuclear export.