Yrb1p Interaction with the Gsp1p C Terminus Blocks Mog1p Stimulation of GTP Release from Gsp1p*

Mog1p, a multicopy suppressor ofgsp1, the temperature-sensitive mutant of theSaccharomyces cerevisiae Ran homologue, binds to GTP-Gsp1p but not to GDP-Gsp1p. The function of Mog1p in the Ran cycle is as yet unknown. This study found that Mog1p releases a nucleotide from GTP-Gsp1p but not from GDP-Gsp1p. Yrb1p, the S. cerevisiaehomologue of RanBP1, which is a strong inhibitor of RCC1-stimulated nucleotide release, also inhibited the Mog1p-stimulated nucleotide release from GTP-Gsp1p. At a concentration corresponding to the molar concentration of GTP-Gsp1p, Yrb1p completely inhibited the Mog1p-stimulated nucleotide release. Consistently, the Yrb1p·GTP-Gsp1p complex was more stable than the Mog1p·GTP-Gsp1p complex. Yrb1p did not inhibit the Mog1p-stimulated nucleotide release from GTP-Gsp1ΔC. The Gsp1ΔC protein lacks the final eight amino acids of the C terminus, and for this reason, the interaction between GTP-Gsp1ΔC and Yrb1p was strongly reduced. On the other hand, Mog1p binds to GTP-Gsp1ΔC more efficiently than to GTP-Gsp1p.

sensitive (ts) mutant of the S. cerevisiae Ran homologue (11). The disruptant of MOG1,⌬mog1, is temperature-sensitive for growth. In this mutant, both nuclear localization signal, nuclear localization signal-dependent and -independent nuclear protein pathways are abolished, but mRNA export seems to be normal (11). Interestingly, the overproduction of Ntf2p confers the ts ϩ phenotype to ⌬mog1 cells. Ntf2p is also essential for nuclear protein import (15,16).
In order to clarify the function of Mog1p in the Ran GTPase cycle, recombinant Mog1p was produced in Escherichia coli and purified. Interestingly, recombinant Mog1p releases a nucleotide from GTP-Gsp1p but not from GDP-Gsp1p. It did not interfere with Prp20p, the S. cerevisiae RCC1 homologue, or with Rna1p, the S. cerevisiae RanGAP1 homologue. Yrb1p, the S. cerevisiae RanBP1 homologue which inhibits the RCC1stimulated nucleotide release, also inhibited the Mog1p-stimulated nucleotide release from GTP-Gsp1p but not from GTP-Gsp1⌬C. Gsp1⌬C lacks the final eight amino acids of the C terminus, and for this reason, the interaction between Yrb1p and Gsp1⌬C was strongly reduced. However, Gsp1⌬C still binds to Mog1p, with an efficiency higher than when wild-type (wt) 1 Gsp1p binds to Mog1p. The strong interaction between Mog1p and Gsp1⌬C should be a clue toward clarifying Mog1p function in the Ran GTPase cycle, which is as yet unknown.

EXPERIMENTAL PROCEDURES
Strains and Medium-S. cerevisiae wild-type strain, YPH499, and E. coli strains, BL21(DE3) and XL1-blue, were used for producing recombinant proteins and for plasmid engineering, as described previously (17). The media used for yeast and bacteria have also been described previously (18).
Construction and Purification of Recombinant Proteins-Plasmids carrying the glutathione S-transferase (GST)-fused DNAs (pGEX-MOG1 (11), pGEX-RNA1 (19), pGEX-GSP1 (19), pGEX-YRB1⌬1-9 (19), and pGEX-PRP20 (19)) have been described previously. The deletion of the C terminus of Gsp1p was accomplished using polymerase chain reaction mutagenesis to introduce a stop codon and remove the final 8 amino acid residues of the open reading frame. The coding region of the GSP1 gene was amplified using as primers 5Ј-CAT TTA TAT TTA TCC ATG GCT GCC CCA G-3Ј and 5Ј-CCC TTA TAA ATCA GCA TAA GCT TAA TCT TAC AAT GGC AAA GCA G-3Ј. The resulting GSP1 C DNA was digested with the restriction enzymes NcoI and HindIII and then inserted into the NcoI/HindIII sites of pGEX-KG.
The E. coli strains, BL21 (DE3)/pGEX-MOG1, BL21(DE3)/pGEX-PRP20, BL21 (DE3)/pGEX-YRB1⌬1-9, and BL21(DE3)/pGEX-RNA1, were each cultured individually in 750 ml of SB medium at 30°C up to an OD 660 of 0.4, treated with isopropyl-1-thio-␤-D-galactopyranoside (final concentration 0.2 mM) for 4 h, and then collected by centrifugation at 3,200 rpm for 15 min. Cells were lysed, and GST-fused proteins were purified using glutathione-Sepharose 4B beads (Amersham Pharmacia * This work was supported by grants-in-aid for specially promoted research from the Ministry of Education, Science, Sports, and Culture of Japan. 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. ‡ To whom correspondence should be addressed. Tel.: 81-92-642-6175; Fax: 81-92-642-6183; E-mail: tnishi@molbiol.med.kyushu-u.ac.jp. Biotech) as described (9). GST-fused proteins were eluted by reduced glutathione as described (20) or else treated with thrombin as described (9). After digestion with thrombin, GST beads were spun down. The resulting supernatant was adjusted to the MonoQ column buffer (20 mM Hepes (pH 7.5), 10 mM KP i (pH 7.4), and 1 mM DTT) and then charged onto a MonoQ column. Both Mog1p and Prp20p were eluted between 300 and 400 mM NaCl, whereas Yrb1p and Rna1p were eluted between 100 and 200 mM and between 400 and 500 mM NaCl, respectively.
The E. coil strains, BL21 (DE3)/pGEX-wtGSP1 or GSP1⌬C, were cultured in 7.5 liters of SB medium at 30°C up to an OD 660 of 0.4, treated with isopropyl-1-thio-␤-D-galactopyranoside (final concentration 0.2 mM) for 4 h, and then collected by centrifugation at 3,200 rpm for 15 min. After cell lysis, GST-fused Gsp1p was purified in the presence of 1 mM GTP or GDP, using glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech) as described (20). After digestion with thrombin, GST beads were spun down. The resulting supernatant containing Gsp1 proteins was incubated with 1 mM GTP or GDP, adjusted to 25 mM Pipes (pH 6.5), 10 mM KP i (pH 6.5), 1 mM MgCl 2 , 1 mM DTT, and charged onto the SO 3 Ϫ column. GTP-Gsp1p and GDP-Gsp1p were eluted at between 500 and 600 mM and between 150 and 250 mM NaCl, respectively.
Biosensor Analysis-Real time interaction analysis was performed with a BIAcore 2000 biosensor instrument (Biacore, Uppsala, Sweden) (23). CM5 sensor chips and amine coupling kits were obtained from Amersham Pharmacia Biotech. The monoclonal antibody to the GST (Amersham Pharmacia Biotech) was immobilized on the CM5 sensor chip as recommended by the manufacturers, and then 0.2 M GST-fused Mog1p or Yrb1p were injected to be trapped on the sensor chip through the mAb to the GST. Binding experiments were performed through the injection of purified recombinant GTP-wtGsp1p or Gsp1⌬C in buffer containing 25 mM Tris-HCl (pH 7.5), 100 mM NaCl, 5 mM MgCl 2 , and 0.005% Tween 20. Evaluation and calculation of the binding parameters were carried out according to the manual Bia evaluation 2.1 software.

Mog1p Stimulates GTP Release from GTP-Gsp1p-Since
Mog1p specifically binds to GTP-Gsp1p (11), we first examined whether Mog1p stimulates the activity of RanGAP. In order to address this issue, recombinant GST fusion proteins of Mog1p and Rna1p were produced in E. coli, and highly purified by glutathione-Sepharose columns, before being cleaved to remove the GST through incubation with thrombin. The resulting Mog1p and Rna1p were further purified over MonoQ columns to become a single band stained with Coomassie Brilliant Blue (data not shown).
Ten pmol of [␥-32 P]GTP-wtGsp1p were mixed with both Rna1p and Mog1p and, as a control, Rna1p, Mog1p, or buffer alone. After incubation at 30°C for the indicated time, the radioactivity remaining on Gsp1p was quantified using a liquid scintillation counter. As shown in Fig. 1A, the RanGAP activity of Rna1p was enhanced by the addition of Mog1p. After incubation for 10 min, about 80% of the radioactivity of [␥-32 P]GTP-Gsp1p was released, compared with 40% in the case of Rna1p alone. Surprisingly, the radioactivity of [␥-32 P]GTP-wtGsp1p was significantly reduced by the addition of only Mog1p, although [␥-32 P]GTP-wtGsp1p was stable following incubation in buffer alone. In order to confirm further the effect of Mog1p on GTP-Gsp1p, we used another substrate, Gsp1⌬C, that lacks the final eight amino acid residues of the C terminus, instead of wtGsp1p. When 10 pmol of [␥-32 P]GTP-Gsp1⌬C were incubated with the same dose of Rna1p and Mog1p, the radioactivity was strongly reduced. For this reason, the doses of Rna1p and Mog1p, which reduced the radioactivity of [␥-32 P]GTP-Gsp1⌬C by around 40 -60% after incubation for 10 min, were used for a comparison. As shown in Fig. 1B, the RanGAP activity of Rna1p was enhanced by the addition of Mog1p. Furthermore, the radioactivity of [␥-32 P]GTP-Gsp1⌬C was reduced by the addition of Mog1p alone at a concentration 3-fold less than in the case of wtGsp1p.
In order to confirm the effect of Mog1p on Gsp1p, 10 pmol of [␥-32 P]GTP-wtGsp1p or Gsp1⌬C were mixed with the increasing doses of Mog1p and then incubated at 30°C for 10 min. At the doses of Mog1p higher than 10 pmol corresponding to the concentration of Gsp1p, the radioactivity of both wtGsp1p and Gsp1⌬C was significantly reduced (Fig. 1C). Similar to the case of Rna1p, Gsp1⌬C is more sensitive to the action of Mog1p, compared with wtGsp1p. The radioactivity of [␥-32 P]GTP-wtGsp1p and Gsp1⌬C was not reduced by the degradation of wtGsp1p and Gsp1⌬C, since the amount of wtGsp1p and Gsp1⌬C, which was estimated by immunoblotting analysis, was not changed by the addition of Mog1p during the incubation (Fig. 1D).
The above results indicated that Mog1p has either the ability to activate the GTPase of Gsp1p or to stimulate the nucleotide release from GTP-Gsp1p. In order to address this issue, we used [␣-32 P]GTP instead of [␥-32 P]GTP. Ten pmol of [␣-32 P]GTP-wtGsp1p or Gsp1⌬C were mixed with increasing doses of Rna1p and, as a control, Mog1p. After incubation at 30°C for 10 min, the radioactivity remaining on wtGsp1p or Gsp1⌬C was quantified using a liquid scintillation counter. As expected from the function of RanGAP that activates the hy-drolysis of GTP-Ran (24), the radioactivity of [␣-32 P]GTP-wtGsp1p or Gsp1⌬C was not reduced by the addition of Rna1p ( Fig. 2A). To the contrary, it was significantly reduced by the addition of Mog1p at doses higher than 10 pmol (Fig. 2B), as observed in Fig. 1C. These results indicate that Mog1p has the activity to release a nucleotide from GTP-Gsp1p, rather than to activate Gsp1pGTPase. We further confirmed this finding with TLC analysis. Ten pmol of [␣-32 P]GTP-wtGsp1p and Gsp1⌬C were mixed with 100 fmol of Rna1p, at which dose the radioactivity of ␣-32 P was not reduced ( Fig. 2A), and with 300 pmol of Mog1p, at which dose the radioactivity remaining on [␣-32 P]GTP-wtGsp1p and Gsp1⌬C was reduced to less than 20% of the original value (Fig. 2B). After incubation for 10 min at 30°C, the reaction mixtures were boiled, and the remaining nucleotides were separated by thin layer chromatography. As a control, the original mixtures without incubation were also boiled in order to analyze the nucleotides bound to wtGsp1p and Gsp1⌬C. The radioactivity of [␣-32 P]GTP and [␣-32 P]GDP was analyzed by autoradiography. As shown in Fig. 2C, all of the [␣-32 P]GTP bound to wtGsp1p or Gsp1⌬C was changed to [␣-32 P]GDP upon incubation with Rna1p (Fig. 2C, wtGsp1p,  lanes 3 and 4; Gsp1⌬C, lanes 9 and 10). In contrast, the amount of GDP was not increased by incubation with Mog1p, indicating that Mog1p has no ability to activate Gsp1pGTPase.
Mog1p Does Not Interfere with the Nucleotide Exchange Activity of Prp20p-Thus far, only RCC1 has been known to have the ability to exchange the nucleotide on Ran (6,11). RCC1 stimulates a nucleotide release from GTP-or GDP-Ran (25). In this regard, we addressed the question of whether or not Mog1p also stimulates a nucleotide release from GDP-Gsp1p. Ten pmol of [ 3 H]GDP-wtGsp1p or Gsp1⌬C and, as a control, [ 3 H]GTP-wtGsp1p or Gsp1⌬C were mixed with increasing doses of Mog1p. After incubation at 30°C for 10 min, the radioactivity remaining on Gsp1p was quantified using a liquid scintillation counter. The radioactivity of [ 3 H]GDP-wtGsp1p did not decrease, whereas the radioactivity of [ 3 H]GTP-wtGsp1p was clearly reduced by the same preparation of Mog1p (Fig. 3). The radioactivity of After incubation at 30°C for 10 min, the reactions were stopped, and the radioactivity remaining on Gsp1p was counted (C). The vertical axis shows the ratio of radioactivity after incubation, compared with the value in the absence of Mog1p (0 mol). The reaction mixtures were boiled after incubation, separated by 11.25% SDS-PAGE, transferred to polyvinylidene difluoride membrane, and assayed for the presence of Gsp1p by immunoblotting analysis using anti-Gsp1p antibodies as a probe (D). Lanes 1-6, wtGsp1p, and lanes 7-12, Gsp1⌬C. The doses of Mog1p were as follows: lanes 1 and 7, 0 pmol; lanes 2 and 8, 1 pmol; lanes 3 and 9, 10 pmol; lanes 4 and 10, 30 pmol; lanes 5 and 11, 100 pmol; lanes 6 and 12, 300 pmol.
Prp20p and Mog1p by kinetic analysis. Prp20p is the S. cerevisiae RCC1 homologue. Ten pmol of [ 3 H]GDP-wtGsp1p were mixed with both Prp20p (0.1 pmol) and Mog1p (60 pmol) and, as a control, with Prp20, Mog1p, or buffer alone. After incubation at 30°C for the indicated time, the radioactivity remaining on Gsp1p was quantified using a liquid scintillation counter (Fig. 4). As reported (19), the radioactivity of [ 3 H]GDP-wtGsp1p was reduced by the addition of Prp20p (Fig. 4A). The remaining radioactivity of [ 3 H]GDP-wtGsp1p was almost identical between the mixture containing both Prp20p and Mog1p and the mixture containing Prp20p alone. This was not due to the inactivation of Mog1p, since the same preparation of Mog1p released nucleotide from [ 3 H]GTP-wtGsp1p (Fig. 4B). Upon the addition of Prp20p, the Mog1p-stimulated nucleotide release from [ 3 H]GTP-wtGsp1p was further enhanced. The identical results were obtained by using [ 3 H]GDP-and GTP-Gsp1⌬C as substrates (Fig. 4, C and D). In the case of Gsp1⌬C, we used 20 pmol as the dose of Mog1p for comparison, as described in Fig.  1. We thus concluded that Mog1p does not interfere with the Prp20-stimulated nucleotide release from GDP-Gsp1p or GTP-Gsp1p. Both Mog1p and Prp20p additively enhanced a nucleotide release from GTP-Gsp1p.
Mog1p-stimulated GTP Release from GTP-Gsp1p was Blocked by Yrb1p-RanBP1 has been known to inhibit RCC1stimulated nucleotide release (26). In addition, the S. cerevisiae RanBP1 homologue, Yrb1p, also inhibits the Prp20p-stimulated nucleotide release (19). Based on these previous reports, we examined the functional relationship between Yrb1p and Mog1p. Ten pmol of [␣-32 P]GTP-wtGsp1p were mixed with 100 pmol of Mog1p and increasing doses of Yrb1p. After incubation at 30°C for 10 min, the radioactivity remaining on Gsp1p was quantified using a liquid scintillation counter. At the doses of Yrb1p, higher than 10 pmol corresponding to the concentration of [␣-32 P]GTP-wtGsp1p, Yrb1p completely inhibited the Mog1p-stimulated nucleotide release from GTP-Gsp1p (Fig. 5,  open circles). This finding indicates that the interaction between Yrb1p and GTP-wtGsp1p is important for Yrb1p to inhibit Mog1p-stimulated GTP release.
Ran lacking the C-terminal amino acids conserved in the Ran family is unable to bind tightly to RanBP1 (17,27,28). If this is the case regarding the interaction between Yrb1p and Gsp1p, then the C-terminal amino acids of Gsp1p may be essential for Yrb1p to inhibit the Mog1p-stimulated nucleotide release. Indeed, when GTP-Gsp1⌬C was used as a substrate instead of GTP-wtGsp1p, the activity of Yrb1p for inhibiting the Mog1p-stimulated nucleotide release was almost com-  1-6) or Gsp1⌬C (lanes 7-12) were mixed with GAP buffer alone (lanes 1, 2, 7, and 8), 100 fmol of Rna1p (lanes 3,  4, 9, and 10), and 300 pmol of Mog1p (lanes 5, 6, 11, and 12). The mixtures were incubated at 30°C for 10 min. The reaction mixtures prior to incubation (odd-numbered lanes) and after incubation (evennumbered lanes) were boiled, separated by thin layer chromatography, and analyzed by autoradiography. pletely abolished (Fig. 5, closed circles). For a comparison, we used the dose of Mog1p that released 60% of the nucleotides from [␣-32 P]GTP-wtGsp1p or [␣-32 P]GTP-Gsp1⌬C. Even at 10 pmol of Yrb1p, which corresponds to the concentration of GTP-Gsp1⌬C, there was no activity of Yrb1p for inhibiting Mog1pstimulated nucleotide release, suggesting that Yrb1p inhibits Mog1p-stimulated nucleotide release from Gsp1p by directly binding to Gsp1p.
Since RanBP1 makes a complex consisting of RCC1, Ran, and RanBP1 (26), another possibility is that Yrb1p may inhibit the Mog1p-stimulated nucleotide release by making the Yrb1p⅐Gsp1p⅐Mog1p complex. In order to address this question, Mog1p and Yrb1p were mixed with GTP-Gsp1p. As controls, Mog1p was mixed with Yrb1p or GTP-Gsp1p, and Yrb1p was mixed with GTP-Gsp1p as shown in Fig. 6. Those mixtures and the solution containing a single component were subjected to gel filtration analysis. Proteins contained in the resulting fraction were detected by immunoblotting analysis. Gsp1p, molecular mass of 25 kDa, eluted with a peak in fraction 57 (molecular mass of 27 kDa), and Mog1p, molecular mass of 24 kDa, eluted with a peak in fraction 55 (molecular mass of 32 kDa) (Fig. 6, A and B). In contrast, Yrb1p, molecular mass of 23 kDa, eluted with a peak in fraction 50 (molecular mass of 48 kDa) (Fig. 6C), indicating that Yrb1p may behave as a dimer in solution, as reported for RanBP1 (26). When mixed, the complex comprising Mog1p, Yrb1p, and Gsp1p was not detected. Instead, the Mog1⅐Gsp1p complex eluted with a peak in fraction 43 (molecular mass of 84 kDa (Fig. 6, A and B), the stoichiometry of which is either 2:1 or 1:2). The Yrb1p⅐Gsp1p complex eluted with a peak in fraction 46 (molecular mass of 66 kDa) (Fig. 6C). The stoichiometry of Yrb1p⅐Gsp1p complexes appears to be 2:1. It is noticeable that most of the Yrb1p bound to Gsp1p, when Yrb1p was mixed with Gsp1p alone or with Gsp1p plus Mog1p (Fig. 6, A and C). This is consistent with the fact that the formation of the Mog1p⅐Gsp1p complex was strongly reduced by the addition of Yrb1p (Fig. 6B). These results suggested that the Yrb1p⅐GTP-Gsp1p complex is more stable than the Mog1p⅐GTP-Gsp1p complex, and thereby Yrb1p inhibited the interaction between Mog1p and GTP-Gsp1p at doses higher than that corresponding to the concentration of GTP-Gsp1p.
Mog1p Binds to GTP-Gsp1p⌬C-The above results indicated that Mog1p was able to bind to GTP-Gsp1⌬C, whereas the same deletion of mammalian Ran is reported to reduce the affinity between GTP-Ran and RanBP1 by approximately 8,000-fold (17). In order to confirm the interaction between Mog1p and GTP-Gsp1⌬C, we performed real time protein-protein interaction analysis using BIAcore (Fig. 7). The monoclonal antibody (mAb) to glutathione S-transferase was immobilized on the sensor chip of the BIAcore to trap GST-fused Mog1p and, as a control, GST-fused Yrb1p. The interaction of Mog1p or Yrb1p with GTP-Gsp1⌬C and, as a control, GTP-wtGsp1p, was then determined by injecting increasing doses (M) of GTP-Gsp1⌬C or GTP-wtGsp1p. The relative response units obtained with the GST alone were subtracted as a background from those of GST-fused Mog1p and Yrb1p. The interaction between Yrb1p and GTP-Gsp1p was abolished by the disruption of the final eight amino acids of the C terminus, as reported for RanBP1 (Fig. 7, C and D) (17,27,28). In contrast, Mog1p bound to both GTP-Gsp1⌬C and GTP-wtGsp1p. Gsp1⌬C rather efficiently bound to Mog1p, compared with wtGsp1p (Fig. 7, A and B). The calculated association constants (k a , M Ϫ1 s Ϫ1 ) of the interaction of Mog1p and Yrb1p with wtGsp1p or Gsp1⌬C are as follows: Mog1p/GTP-wtGsp1p, 0.97 ϩ 0.47 ϫ 10 4 ; Mog1p/GTP-Gsp1⌬C, 1.40 ϩ 0.90 ϫ 10 4 ; Yrb1p/ GTP-wtGsp1p, 1.33 ϩ 0.70 ϫ 10 4 .

DISCUSSION
Mog1p is a novel protein binding to GTP. It is required for nuclear protein imports but not for mRNA export (11). This study found that Mog1p can release a nucleotide from GTP-Gsp1p, but not from GDP-Gsp1p, in a dose-dependent manner. In this context, Mog1p differs from a guanine nucleotide exchange factor of Gsp1p, RCC1 which releases a nucleotide from both GTP-Gsp1p, and GDP-Gsp1p (29). During the preparation of this manuscript, Steggerda and Paschal (30) reported that murine Mog1p has no nucleotide exchange activity on Ran, whereas it releases GTP from GTP-Ran similar to S. cerevisiae Mog1p. Mog1p does not interfere with the activities of Rna1p and Prp20p. Rather, it functions additively for both Rna1p and Prp20p. It is unknown whether or not Mog1p functionally interacts with Rna1p and Prp20p in vivo.
The exact localization of Mog1p has not yet been defined, except that Mog1p is localized within the nucleus when overproduced (11). Murine Mog1 is also localized in the nucleus (30). In the cytoplasm, Mog1p could help to reduce the concentration of GTP-Gsp1p, maintaining the gradient of GTP-Gsp1p concentration from the nucleus to the cytoplasm. Within the nucleus, on the other hand, Mog1p may be harmful for cell survival since it may reduce the nuclear concentration of GTP-Gsp1p. However, high level expression of Mog1p induced by the GAL promoter does not inhibit the growth of S. cerevisiae (data not shown).
Yrb1p is localized within the cytoplasm, possessing a nuclear export signal. The mutant of Yrb1p lacking the nuclear export signal accumulates within the nucleus (28), suggesting that Yrb1p may have a role to play in the nucleus. One of the possibilities is that the nuclear Yrb1p inhibits the Mog1pstimulated nucleotide release from GTP-Gsp1p in order to maintain the high nuclear concentration of GTP-Gsp1p. RanBP1, the mammalian homologue of Yrb1p, is reported to inhibit completely the RCC1-stimulated nucleotide release from GTP-Ran, at doses greater than that corresponding to the concentration of GTP-Ran (31). In a similar manner, Yrb1p inhibited Mog1p-stimulated nucleotide release from Gsp1p. However, there does not seem to be any direct interaction between Mog1p and Yrb1p (Fig. 6). The formation of the Mog1p⅐GTP-Gsp1p complex seems to be inhibited by the addition of Yrb1p, probably due to the fact that GTP-Gsp1p binds to Yrb1p more stably than to Mog1p (Fig. 7). Thus, Yrb1p may inhibit the interaction between Mog1p and GTP-Gsp1p by depleting the GTP-Gsp1p. This is consistent with the finding that the Mog1p-stimulated nucleotide release from the Gsp1⌬C was not inhibited by the addition of Yrb1p. It is possible that Yrb1p inhibits Mog1p-stimulated nucleotide release by making a complex of Mog1p⅐Gsp1p⅐Yrb1p, since Mog1p was coprecipitated with overexpressed GST-Yrb1p in vivo (data not shown), and murine Mog1 is reported to make a trimeric complex of Mog1⅐Ran⅐RanBP1 (30).
Both Yrb1p and Mog1p bind to GTP-Gsp1p. However, upon the C-terminal deletion of Gsp1p, the ability of Yrb1p to bind to GTP-Gsp1⌬C was profoundly reduced. In contrast, the interaction between Mog1p and Gsp1⌬C was somewhat strengthened, compared with that between Mog1p and wtGsp1p. These results indicate that Mog1p binds to GTP-Gsp1p in a manner different from that of Yrb1p. The method by which Mog1p binds to GTP-Gsp1p may reflect the biological function of Mog1p that still remains to be investigated. Mog1p is conserved in S. pombe (DDBJ/EMBL/GenBank TM accession number AL031179) and probably in humans (DDBJ/EMBL/Gen-Bank TM accession number N48015), indicating that Mog1p plays some important role in the Ran GTPase cycle. The present results provide us with the first clue toward clarifying the function of Mog1p.