Identification and Characterization of a Novel RanGTP-binding Protein in the Yeast Saccharomyces cerevisiae *

The small Ras-like GTPase Ran plays an essential role in the transport of macromolecules in and out of the nucleus and has been implicated in spindle (1, 2) and nuclear envelope formation (3, 4) during mitosis in higher eukaryotes. We identified Saccharomyces cerevisiae open reading frame YGL164c encoding a novel RanGTP-binding protein, termed Yrb30p. The protein competes with yeast RanBP1 (Yrb1p) for binding to the GTP-bound form of yeast Ran (Gsp1p) and is, like Yrb1p, able to form trimeric complexes with RanGTP and some of the karyopherins. In contrast to Yrb1p, Yrb30p does not coactivate but inhibits RanGAP1(Rna1p)-mediated GTP hydrolysis on Ran, like the karyopherins. Stu I- Pst I GSP1 fragments into the Sma I- Pst I sites of the two-hybrid vector pBTM116

pET9d-GST-TEV-YRB30 was generated by replacing a NcoI-SacI fragment containing the MAD2 gene from pET9d-GST-TEV-MAD2 (G. Stier, European Molecular Biology Laboratory (EMBL) Heidelberg, Germany) with the corresponding fragment from pGALPATG1L-YRB30. For expression of a His 6 -tagged fusion protein in E. coli the entire YRB30 ORF was ligated as a NcoI-BamHI fragment (see above) into pET9d-His 6 -TB adjacent to the Thrombin cleavage site (G. Stier, EMBL). The various plasmids containing deleted or mutated YRB30 genes were constructed in an analogous way.
Fusions between the bipartite NLS of ribosomal protein L25 under control of its own promoter and GFP-YRB30 were constructed by inserting YRB30 as PstI fragment from pNOPGFP2L-YRB30 into pRS314-L25NLS-GFP-ARC1 2 between GFP and the coding sequence of ARC1. The various plasmids containing deleted or mutated YRB30 genes were constructed by PCR in an analogous way.
For assessing the functionality of the C-terminal NES in the YRB30 ORF various PCR-generated BamHI fragments of YRB30 were introduced into the BclI site of pRS315-NMD3(⌬NES1ϩ2)-eGFP, which will be described elsewhere. 3 To check the correct orientation of the insert a NcoI site was introduced immediately upstream of the 3Ј-BamHI site.
For overexpression studies, YRB30 was expressed under control of the galactose-inducible GAL1-promoter from high copy number vector pRS426GAL1 (18). For this purpose, full-length and truncated YRB30 ORFs were cloned as PCR-generated SpeI-XhoI fragments. Expression of all proteins was verified by Western blot analysis using the anti-Yrb30p antiserum or commercially available anti-ProtA, anti-GFP, anti-His 6 , or anti-GST antibodies.
Disruption of YRB30-The YRB30 gene was disrupted by replacing the entire open reading frame of YRB30 with a PCR-generated HIS3cassette in the diploid strain W303-D. The disruption was verified by PCR, and haploid yrb30::HIS3 strains were obtained by sporulation and tetrad dissection of the heterozygous diploid strain. Mating of haploid disruption strains of opposite mating type led to homozygous diploid disruption strains.
Preparation of Rabbit Polyclonal Anti-Yrb30p Antiserum-A rabbit polyclonal antiserum against Yrb30p was raised against His 6 -Yrb30p fusion protein expressed in BL21(DE3) cells (23) and affinity purified over Talon beads (Clontech, Palo Alto, CA). Immunization of two rabbits was performed by a commercial antibody service (J. Pineda, Berlin, Germany). For detection of Yrb30p in immunoblots, the resulting antisera were used at a dilution of 1:4000.
Two-hybrid Screen-The two-hybrid screen was performed as described previously (24) but using L40⌬gal4 cells for transformation of the bait-plasmids.
In Vitro Protein Interaction/Pull-down Assays-In vitro interaction between recombinant proteins was assayed as described previously (25).
For the pull-down assay from yeast extract a culture from wt strain W303-1A was grown at 30°C to an A 600 of 1.5. Cells were harvested by centrifugation and lysed using a Fritsch Pulverisette 6 (Fritsch GmbH, Idar-Oberstein, Germany). Endogenous glutathione in the supernatant was removed by gel filtration using G25-fine-Sepharose (Amersham Biosciences). Loading of extract with GTP␥S was done according to loading of Gsp1p with GTP (25). Purified GST-Yrb30p (ϳ15 g) was rebound to 50 l of GSH-Sepharose in universal buffer (25) containing 5 mM ␤-mercaptoethanol and one tablet of Complete EDTA-free protease inhibitor mix (Roche Diagnostics) per 50 ml, washed, and incubated with ϳ800 g of protein of the corresponding supernatant. After washing the bound proteins were eluted in sample buffer and analyzed with SDS-PAGE and Coomassie Blue staining, as well as mass spectrometry.
Purification of ProtA-TEV-Yrb30p from Yeast-Affinity purification of ProtA fusion proteins was performed essentially as described (26).
Release Assay to Determine Inhibition of RanGAP-induced GTP Hydrolysis-GST-Yrb30p or GST-Yrb1p as a control were incubated in universal buffer (25) with Gsp1pGTP. After washing, the complexes were incubated for 12 min at 4°C with His 6 -Rna1p or buffer alone. The released proteins were collected. After washing the bound fraction was eluted by boiling with sample buffer. Human Ran loaded with GTP and Rna1p from Schizosaccharomyces pombe were kindly provided by R. Bischoff (Deutsches Krebsforschungszentrum, Heidelberg, Germany).
Fluorescence Microscopy-Classical and confocal fluorescence microscopy of living cells expressing GFP(S65T) fusion proteins was done as described previously (28).
Miscellaneous-SDS-PAGE, immunoblot, and preparation of yeast cell extracts were performed according to standard protocols. 2 K. Galani and E. Hurt, unpublished data. 3 T. Gerstberger and E. Hurt, manuscript in preparation.

Two-hybrid Screen for Novel Yrb (Yeast Ran Binders)-
To identify novel yeast Ran-binding proteins a two-hybrid screen was performed using both wt (40 million of interactions tested) and two dominant mutant forms (G21V and T26N; 29 million or 46 million of interactions tested, respectively) of yeast Ran, Gsp1p. These mutant forms are locked in the GTP-bound and the GDP-bound state, respectively (29). Among the preys identified, we pulled out a number of already known Ran-binding proteins, such as members of the karyopherin family, the yeast homolog of RanBP1 (Yrb1p), and Mog1p (Table II). In addition, we identified a number of candidate novel Ran-binding proteins (Yrbs), most of them in the screen with the wt form of Gsp1p. Comparably few preys were identified with the mutant forms of Gsp1p.
Yrb30p Interacts with Gsp1p in Vivo-As a first approach to confirm a physical interaction between these candidate Ranbinding proteins and Gsp1p, we fused selected ORFs (see Table  II) to a TEV-cleavable ProtA moiety and expressed the fusion proteins under control of the NOP1 or the GAL1 promoter. After affinity purification over IgG-Sepharose we probed the column loads and the TEV eluates with a specific antiserum against Gsp1p. In none of the cases we could detect an enrichment of Gsp1p in the eluate (data not shown).
The fact that one of the ORFs, YGL164C (YRB30), was found as the only prey besides Yrb1p in the two-hybrid screen with the G21V mutant, suggested that the protein bound specifically to the GTP-bound form of Gsp1p, which may be underrepresented in our cell lysate because of RanGAP1(Rna1p)-mediated GTP hydrolysis. We hypothesized that a putative Yrb30p-Gsp1pGTP complex would be stabilized in the rna1-1 mutant strain, in which GTP hydrolysis is inhibited. We therefore expressed and affinity-purified the ProtA-tagged Yrb30p driven from a galactose-inducible GAL1 promoter from both wild-type and rna1-1 mutant yeast cells. As predicted, we could enrich Gsp1p using ProtA-Yrb30p when purified from the rna1-1 background (Fig. 1). In contrast, no Gsp1p could be copurified with ProtA-Yrb30p from the wt strain or with ProtA alone. A similar result was obtained if Yrb30p was expressed from the constitutive NOP1 promoter (data not shown).
In an inverse approach, we coexpressed GFP-tagged Yrb30p with ProtA-tagged Gsp1p in its wt form or in its G21V mutant form, respectively, under control of the NOP1 promoter in wt cells. Proteins were purified over ProtA-Sepharose and eluted by TEV cleavage. The eluates were analyzed by immunoblotting using a GFP antiserum. In agreement with the two-hybrid data we detected GFP-Yrb30p coeluting with ProtA-Gsp1(G21V) but not with the wt form of Gsp1p (Fig. 1). We conclude that Yrb30, initially found in a two-hybrid screen with Gsp1(G21V), interacts specifically with the GTP-bound form of Gsp1p in vivo.
Yrb30p Binds Directly to Gsp1p-To test whether Yrb30p could bind to Gsp1p directly, we performed pull-down assays using recombinant proteins produced in E. coli. We immobilized GST-Yrb30p, as well as GSTp alone, as a negative control and GST-Yrb1p (yRanBP1) and GST-Kap95p (yImp␤) as positive controls on GSH-Sepharose beads and analyzed binding of His 6 -tagged Gsp1p, loaded either with GTP or GDP, to these beads. Like the already known RanGTP-binding proteins, Yrb30p showed clear binding of Gsp1pGTP but not of Gsp1pGDP in this assay, which indicates that these proteins interact directly and specifically ( Fig. 2A). Moreover, we could demonstrate that Yrb30p binds human Ran with the same specificity (data not shown). Ϫ) and yrb30::HIS3 rna1-1 (rna1-1; ϩ) cells containing a plasmid expressing ProtA-Yrb30p or ProtA alone under control of the GAL1 promoter were grown in SRC-Leu at 23°C. After induction with YPGal the cells were shifted for 2 h to 37°C. The ProtA fusion proteins were affinity-purified on IgG-Sepharose and eluted using acetic acid. Loads and eluates were analyzed by SDS-PAGE, Coomassie Blue staining, and immunoblotting using a ␣Gsp1p antiserum. Prominent bands were identified by mass spectrometry. The relative mobility of Gsp1p is indicated. Asterisks indicate the purified ProtA fusion proteins, and degradation products of ProtA-Yrb30p are indicated by open circles.
Proteins that bind to yeast Ran probably via three-hybrid interactions involving karyopherins. c ORFs selected for further analysis.
As an alternative approach to test the specificity of the interaction between Yrb30p and Gsp1pGTP, we tested whether we could pull down endogenous Gsp1p from a yeast extract using an excess of immobilized GST-Yrb30p. To convert all GTPases in the yeast extract to their GTP-bound state, half of the extract was loaded with GTP␥S whereas the other half was mock-treated (see "Experimental Procedures"). Indeed, Yrb30p was able to pull down Gsp1p that was verified by mass-spectrometry but only from the GTP␥S-treated extract (Fig. 2B), suggesting a high specificity and affinity for Gsp1pGTP.
Identification of the RanGTP-binding Domain of Yrb30p-Having established the direct interaction between Yrb30p and Gsp1pGTP we were interested in delimiting the domain of Yrb30p required for binding to Gsp1pGTP. For this purpose we created various N-and C-terminal truncations of GST-Yrb30p and tested them for binding to His 6 -Gsp1pGTP as outlined above.
As shown in Fig. 3A only the N-terminal 127 amino acids of Yrb30p were dispensable for RanGTP binding. Hence, the minimal RBD of Yrb30p comprises residues 128 to 440. This domain is much larger than the RBD of RanBP1 (ϳ130 residues) and does not show any sequence similarity to the RBDs of RanBP1 or importin ␤, suggesting a novel motif for RanGTP binding (Fig. 3B, RBD).  For the latter experiment, the plasmids encoding proteins 1-3 were introduced in a NMD3 shuffle strain, and the transformants were analyzed for growth on 5-fluoro-orotic acidcontaining medium, which selects against the plasmid carrying the essential NMD3 wt gene.
(GenBank TM accession number AAA35271.1). These proteins show sequence identities of 30 to 40% and similarities of 50 to 60% with highly conserved sequence blocks that are spread over the entire sequence and not restricted to the RanGTPbinding domain.
Characterization of the Yrb30p-binding Site of Gsp1pGTP-It was shown previously (30) that importin ␤ and RanBP1 bind to different surfaces on RanGTP thus enabling the formation of ternary complexes. To get an idea about the epitope on RanGTP recognized by Yrb30p, we tested how His 6 -Kap95p (yeast importin ␤) and His 6 -Yrb1p (yRanBP1) would interfere with the in vitro interaction between GST-Yrb30p and His 6 -Gsp1pGTP. In a first experiment, GST-Yrb30p was bound to GSH-Sepharose beads and incubated with a mixture of His 6 -Gsp1pGTP and His 6 -Yrb1p. With increasing concentrations of His 6 -Yrb1p the amount of His 6 -Gsp1pGTP bound to GST-Yrb30p was decreasing suggesting that Yrb1p and Yrb30p compete for an overlapping binding site on Gsp1GTP (Fig. 4A). To test whether Yrb30p could build trimeric complexes, like Yrb1p, with karyopherins, we incubated GST-Yrb30p pre-bound GSH-Sepharose with His 6 -Kap95p and His 6 -Gsp1p either in its GTP-or GDP-bound form. As shown in Fig. 4B Yrb30p was able to build trimeric complexes with Kap95p and Gsp1pGTP. In contrast to Yrb1p, no trimeric complexes were observed with His 6 -Gsp1pGDP. In summary, the binding site for Yrb30p on RanGTP appears to overlap with the one for RanBP1 but is distinct from the one for importin ␤.
Modulation of the RanGTPase Cycle by Yrb30p-All RanGTP-binding proteins known so far, including the karyopherins and the RanBP1-related proteins, inhibit RanGEF (RCC1)-mediated GTP-to-GDP exchange on Ran, presumably by competition with RCC1 for binding to RanGTP. Accordingly, GST-Yrb30p also had an inhibitory effect on RCC1-mediated guanine nucleotide exchange (Fig. 5A). The dissociation constant of the Yrb30p-Gsp1pGTP complex derived from the concentration of GST-Yrb30p necessary for half-maximal inhibition was 10 nM, which is ϳfive times higher than the value for Yrb1p (21) but still significantly lower than for importin ␤, suggesting a rather high affinity.
Because Yrb30p showed similar binding characteristics like Yrb1p, the coactivator of RanGAP-mediated GTP hydrolysis on Ran, we checked whether Yrb30p would show a similar biochemical behavior. RanGAP-mediated GTP hydrolysis on RanGTP was measured as release of 32 P from Ran loaded with [␥-32 P]GTP as described earlier (24). Addition of GST-Yrb30p did not coactivate the reaction but acted as an inhibitor like karyopherins (Fig. 5B). We were able to confirm this result in pull-down assays where we incubated pre-formed complexes between His 6 -Gsp1pGTP and GST-Yrb1p or GST-Yrb30p, respectively, with recombinant His 6 -Rna1p (Fig. 5C). Under the applied conditions, we observed release of Gsp1p only from Yrb1p but not from Yrb30p.
Evidence for a Shuttling Mechanism of Yrb30p-One of the conserved sequence blocks (Fig. 3B) in the C terminus of Yrb30p (residues 377 to 413) resembles a canonical NES for the nuclear export receptor CRM1 (L-X 2-3 -(F,I,L,V,M)-X 2-3 -L-X-(I,L)) (31). Because this finding suggests that Yrb30p shuttles between nucleus and cytoplasm, like importin ␤ and RanBP1, we tested this sequence for functionality.
First, we tested whether the Yrb30p NES would function in a heterologous context. For this purpose we replaced both NESs in Nmd3p by the NES-containing C terminus of Yrb30 (residues 341-440). Nmd3p is an adaptor in the Crm1p-mediated nuclear export of the large ribosomal subunit and shuttles between nucleus and cytoplasm (32,33). Deletion of the NESs in Nmd3p is lethal and leads to nuclear accumulation of a GFP-Nmd3p fusion protein, which is cytoplasmic at steady state (Fig. 6). Insertion of the Yrb30p C terminus rescued both cell growth and cytoplasmic localization of GFP-Nmd3p devoid of its own NESs (Fig. 6). This result demonstrates the functionality of the Yrb30p NES in a heterologous context.
Next, we tested whether the sequence was recognized in vitro by yeast Crm1p/Xpo1p. For this purpose, we performed pull-down assays as described above using GST-Yrb30p, His 6 -Gsp1pGTP, and His 6 -yCrm1p. We were able to form trimeric complexes suggesting that the NES was indeed recognized by yCrm1p (Fig. 7). Complex formation was cooperative, because no binding of His 6 -yCrm1p to GST-Yrb30p was observed in the absence of His 6 -Gsp1pGTP (data not shown), and yCrm1p does not bind Gsp1pGTP alone (21). Similar complexes are formed by GST-Yrb1p where the formation is dependent on the integrity of the acidic C terminus of Gsp1p and the integrity of the FIG. 7. Complex formation among Yrb1p, Xpo1p, and Gsp1pGTP. A, the NES but not the RBD of Yrb30p is necessary for complex formation. Recombinant GST fusions of full-length Yrb30p and Yrb1p, as well as N-and C-terminally truncated forms of Yrb30p, were tested for complex formation with His 6 -Xpo1p in the presence of His 6 -Gsp1pGTP and His 6 -Gsp1pGDP. The bound fractions were analyzed by SDS-PAGE and Coomassie Blue staining, as well as immunoblotting using ␣Gsp1p antiserum. B, complex formation of Yrb30p is, in contrast to Yrb1p, not dependent on the integrity of the acidic C terminus of Gsp1p. Recombinant GST-Yrb30p, and GST-Yrb1p as a control, were tested for complex formation with His 6 -Xpo1p in the presence of full-length His 6 -Gsp1pGTP or a version lacking the last seven residues (⌬DE). Shown are the bound fractions of the pull-down assays using GSH-Sepharose analyzed by SDS-PAGE and Coomassie Blue staining.
RBD of Yrb1p. To show the NES dependence of the interaction we tested a C-terminally truncated Gsp1pGTP lacking the last seven residues (1-212; ⌬DE), an N-terminal truncation of Yrb30p defective in Gsp1pGTP binding (252-440) but retaining the NES, and a C-terminal truncation of Yrb30p (1-377), devoid of both NES and Gsp1pGTP binding, for complex formation. Complex formation was possible in all cases except the C-terminally truncated Yrb30p, suggesting that the NES is necessary and sufficient for complex formation (Fig. 7).
Third, we examined the localization of various GFP-Yrb30p fusion proteins in a wt background and in various export mutants. In a wt background and in the four export mutants, xpo1-1, cse1-2, ⌬msn5, and ⌬los1, the protein localizes exclusively to the cytoplasm (Fig. 8) (data not shown). However, deletion of residues 378 to 440 including the NES and part of the RBD leads to a mislocalization of the protein to the nucleoplasm of wt cells suggesting that this region also acts as a NES in the homologous context (Fig. 8). Interestingly, impairment of yCrm1p function either by mutation (xpo1-1) or inhibition by leptomycin B does not result in nuclear accumulation of GFP-Yrb30p unless a strong heterologous NLS is fused to the protein, or, alternatively, RanGTP binding is compromised (GFP-Yrb30 (252-440)p) (Fig. 8). In these cases, however, nuclear accumulation occurs very rapidly i.e. within 5 min upon switching to the restrictive conditions (data not shown). These results suggest that Yrb30p shuttles between nucleus and cytoplasm and is exported from the nucleus by two alternative pathways, one dependent on its NES and Crm1p and another one on its ability to bind RanGTP.
Cellular Function of Yrb30p-Interestingly, there are two proteins in S. pombe that are homologous to Yrb30p. These two proteins bind specifically to the GTP-bound form of Gsp1p in vitro suggesting that they are conserved not only on a structural but also on a functional level (data not shown). Thus far, we did not find any homologue in higher eukaryotes suggesting that the function of Yrb30p is restricted to fungi.
To explore the function of this novel RanGTP-binding protein, we created complete deletions of the corresponding ORFs in S. cerevisiae (W303-1A), S. pombe (single and double knock-outs), and A. gossypii. Neither of these mutations had any obvious phenotype under normal growth conditions. For the S. cerevisiae null mutant we tested in addition growth at different temperatures and on different media formulations, as well as mating and sporulation, haploid invasive growth, and localization of various nucleocytoplasmic transport cargoes. None of these processes was affected by the deletion of YRB30. Neither could we find any synthetic growth phenotype of the deletion with other mutations in the Ran GTPase cycle or the nucleocytoplasmic transport machinery. Mutations tested included gsp1-1, rna1-1, prp20 -1,  yrb1-51, ⌬mog1, ⌬yrb4, srp1-31, xpo1-1, crm1-1, -2, -3, ⌬los1, ⌬msn5, cse1-2, ⌬nup2, ⌬yrb2, and ⌬nup42. On the other hand, galactose-induced overproduction of the fulllength protein from a high copy number plasmid under control of the GAL1 promoter did not suppress the temperature sensitivity of gsp1-1, rna1-1, prp20 -1, or yrb1-51 strains. However, the GAL1-YRB30 construct was able to partially suppress the toxicity of overproduced Gsp1(G21V)p (data not shown). Because the same effect is also seen with a GAL1-YRB1 construct (data not shown), this phenotype confirms again the in vivo interaction between Yrb30p and RanGTP. During these overproduction studies we noticed a slight inhibition of vegetative growth by overproduction of full-length Yrb30p independent of the genetic background. This phenotype was much more pronounced in case of the overproduction of C-terminally truncated forms of Yrb30p, such as Yrb30p (1-377), which strongly inhibits the vegetative growth of a wt strain (Fig. 9). It remains to be seen whether this phenotype is dependent on the nuclear accumulation of these constructs (see Fig. 8). Transformants pre-grown in selective medium with raffinose as sole carbon source were spotted in serial dilutions on selective plates containing either glucose or galactose as sole carbon source and grown for 2 or 3 days, respectively.

DISCUSSION
Effector or target proteins of small Ras-like GTPases are the molecular links between these molecular switches and the cellular processes regulated by them. Given that the variety of cellular processes Ran had been implicated (6,34), it seemed not unlikely that there were several effectors of Ran, each one specific for a given cellular process. However, today we know that probably all three processes that are controlled by Ran, nucleocytoplasmic transport, microtubule aster formation, and nuclear envelope formation, are mediated by one type of effector proteins, the karyopherins (2,9). This family of proteins thus represents, in a strict sense, the only effector or target protein of Ran. Besides, however, there is a family of RanGTPbinding proteins, RanBP1 and its relatives (Yrb1p, Yrb2p, and Nup2p in S. cerevisiae), that plays an important role as modulators of the Ran GTPase system (6).
In this study, we set out to identify further RanGDP-or RanGTP-binding proteins in the yeast S. cerevisiae using a two-hybrid approach. Among the proteins identified were already known Ran-binding proteins such as karyopherins, RanBP1 and Mog1p. Interestingly, latter Ran-binding protein was identified using as a bait the Gsp1p(T26N) mutant form of yeast Ran, which is supposed to be locked in the empty or RanGDP-bound state (35). This finding is in agreement with recent publications (36,37) where Mog1p is implicated in the release of RanGDP from its import receptor NTF2 and in the subsequent guanine-nucleotide exchange on RanGDP to RanGTP by RanGEF (RCC1 in metazoans, Prp20p in S. cerevisiae). The only novel protein among the identified preys whose physical interaction with Ran could be confirmed by various means was YGL164c, which was termed Yrb30p. Consistent with its identification in the screen with Gsp1p(G21V), where yeast RanBP1 (Yrb1p) was also identified, Yrb30p binds specifically and with high affinity (10 Ϫ7 M) to the GTP-bound form of yRan (Gsp1p).
Its binding site on Gsp1pGTP overlaps with the one for yRanBP1 (Yrb1p) and is distinct from the one recognized by importin ␤. Despite these similarities, Yrb30p does not coactivate RanGAP-mediated GTP hydrolysis on Ran but rather inhibits this reaction, as well as RanGEF-mediated guaninenucleotide exchange, like karyopherins do. The existence of a functional NES for the major nuclear export receptor, Exportin 1 (CRM1) in metazoans or Xpo1p in S. cerevisiae, suggests that the protein, despite its exclusive cytoplasmic localization at steady state, shuttles between nucleus and cytoplasm. This suggestion is confirmed by the nuclear accumulation of a Cterminally truncated form of the protein, which is defective for RanGTP binding and devoid of the NES, in a wt background, and by the nuclear accumulation of a N-terminally truncated Yrb30p, harboring the NES but unable to bind to RanGTP directly, upon inhibition of Crm1p-dependent nuclear export. The lack of nuclear accumulation of full-length Yrb30p upon impairment of yCrm1p function may be explained by the existence of two alternative nuclear export pathways for Yrb30p, one dependent on the NES and Crm1p and another one on the integrity of the RanGTP-binding domain. In latter case, Yrb30p may leave the nucleus as piggy-bag of any RanGTP-karyopherin complex. Alternatively to such a shuttling mechanism, the cytoplasmic localization of Yrb30p could be explained by a strong retention mechanism and diffusion of the protein into the nucleus upon loss of this retention. In this model, the nuclear export pathways for Yrb30p would be needed as a kind of safeguard mechanism to avoid the occurrence of Yrb30p in the nucleus, which may be detrimental for the cell. At the moment we cannot distinguish between these two possibilities, but the very rapid kinetics of the nuclear accumulation of GFP-Yrb30 (252-440)p upon impairment of yCrm1p function rather argues for a shuttling mechanism.
Regardless of its interesting binding behavior and localization, the cellular function of Yrb30p is unclear. Despite the fact that Yrb30p and RanBP1 (Yrb1p) see the same epitope on RanGTP, they differ in their effect on the cytoplasmic Ran-GAP1 (Rna1p in S. cerevisiae)-mediated GTP hydrolysis on RanGTP in that Yrb30p inhibits the reaction whereas Yrb1p coactivates it. Based on this behavior and the formation of ternary complexes among Yrb30p, RanGTP, and karyopherins, we hypothesize that Yrb30p may stabilize specific export complexes toward dissociation by Yrb1p and Rna1p in the cytoplasm to allow their transport to further destinations in the cell. Such a hypothesis of a pathway-specific rather than a general function of Yrb30p in the Ran GTPase cycle is in agreement with the cellular abundance of Yrb30p, which is estimated to be only [1/20] of Yrb1p. This estimation is based on immunoblotting of whole cell extracts derived from strains containing ProtA-tagged versions of the chromosomal YRB30 and YRB1 genes (data not shown). However, the hypothetic modulation of the RanGTPase cycle or nuclear export appears to be dispensable for vegetative growth, because a deletion of YRB30 does not result in any obvious phenotype, neither in S. cerevisiae nor in S. pombe or A. gossypii. The dominant-negative effect of overproduced, C-terminally truncated forms of the protein on yeast vegetative growth may result from the sequestration of an essential protein whose interaction with Yrb30p is stabilized by the truncation of Yrb30p. Ran is an unlikely candidate for this interaction partner, because these C-terminal truncations abolish the interaction of Yrb1p with Gsp1pGTP (see Fig. 3A). The non-essential function of Yrb30p might be related to the exclusive conservation of Yrb30p in the fungal kingdom as there is no evidence for homologues of Yrb30p in higher eukaryotes. A prominent feature of fungi is their closed mitosis i.e. the integrity of the nuclear envelope during nuclear and cellular division. At the moment, however, we have no evidence for a role of Yrb30p in this context.