The transport receptor Crm1 mediates the export of diverse cargos containing leucine-rich nuclear export signals (NESs) through complex formation with RanGTP. To ensure efficient cargo release in the cytoplasm, NESs have evolved to display low affinity for Crm1. However, mechanisms that overcome low affinity to assemble Crm1-export complexes in the nucleus remain poorly understood. In this study, we reveal a new type of RanGTP-binding protein, Slx9, which facilitates Crm1 recruitment to the 40S pre-ribosome-associated NES-containing adaptor Rio2. In vitro, Slx9 binds Rio2 and RanGTP, forming a complex. This complex directly loads Crm1, unveiling a non-canonical stepwise mechanism to assemble a Crm1-export complex. A mutation in Slx9 that impairs Crm1-export complex assembly inhibits 40S pre-ribosome export. Thus, Slx9 functions as a scaffold to optimally present RanGTP and the NES to Crm1, therefore, triggering 40S pre-ribosome export. This mechanism could represent one solution to the paradox of weak binding events underlying rapid Crm1-mediated export.
The hypoxia–inducible transcription factor (HIF) is a key component of the cellular adaptation mechanisms to hypoxic conditions. HIFα subunits are degraded by prolyl-4-hydroxylase domain (PHD) enzyme-dependent prolyl-4-hydroxylation of LxxLAP motifs that confer oxygen-dependent proteolytic degradation. Interestingly, only three non-HIFα proteins contain two conserved LxxLAP motifs, including the putative RNA helicase with a zinc finger domain HELZ. However, HELZ proteolytic regulation was found to be oxygen-independent, supporting the notion that a LxxLAP sequence motif alone is not sufficient for oxygen-dependent protein destruction. Since biochemical pathways involving RNA often require RNA helicases to modulate RNA structure and activity, we used luciferase reporter gene constructs and metabolic labeling to demonstrate that HELZ overexpression activates global protein translation whereas RNA-interference mediated HELZ suppression had the opposite effect. Although HELZ interacted with the poly(A)-binding protein (PABP) via its PAM2 motif, PABP was dispensable for HELZ function in protein translation. Importantly, downregulation of HELZ reduced translational initiation, resulting in the disassembly of polysomes, in a reduction of cell proliferation and hypophosphorylation of ribosomal protein S6.
Nuclear export of mRNAs and pre-ribosomal subunits (pre40S and pre60S) is fundamental to all eukaryotes. While genetic approaches in budding yeast have identified bona fide export factors for mRNAs and pre60S subunits, little is known regarding nuclear export of pre40S subunits. The yeast heterodimeric transport receptor Mex67-Mtr2 (TAP-p15 in humans) binds mRNAs and pre60S subunits in the nucleus and facilitates their passage through the nuclear pore complex (NPC) into the cytoplasm by interacting with Phe-Gly (FG)-rich nucleoporins that line its transport channel. By exploiting a combination of genetic, cell-biological, and biochemical approaches, we uncovered an unanticipated role of Mex67-Mtr2 in the nuclear export of 40S pre-ribosomes. We show that recruitment of Mex67-Mtr2 to pre40S subunits requires loops emanating from its NTF2-like domains and that the C-terminal FG-rich nucleoporin interacting UBA-like domain within Mex67 contributes to the transport of pre40S subunits to the cytoplasm. Remarkably, the same loops also recruit Mex67-Mtr2 to pre60S subunits and to the Nup84 complex, the respective interactions crucial for nuclear export of pre60S subunits and mRNAs. Thus Mex67-Mtr2 is a unique transport receptor that employs a common interaction surface to participate in the nuclear export of both pre-ribosomal subunits and mRNAs. Mex67-Mtr2 could engage a regulatory crosstalk among the three major export pathways for optimal cellular growth and proliferation.
Full text Figures and data Side by side Abstract eLife digest Introduction Results Discussion Materials and methods References Decision letter Author response Article and author information Metrics Abstract The transport receptor Crm1 mediates the export of diverse cargos containing leucine-rich nuclear export signals (NESs) through complex formation with RanGTP. To ensure efficient cargo release in the cytoplasm, NESs have evolved to display low affinity for Crm1. However, mechanisms that overcome low affinity to assemble Crm1-export complexes in the nucleus remain poorly understood. In this study, we reveal a new type of RanGTP-binding protein, Slx9, which facilitates Crm1 recruitment to the 40S pre-ribosome-associated NES-containing adaptor Rio2. In vitro, Slx9 binds Rio2 and RanGTP, forming a complex. This complex directly loads Crm1, unveiling a non-canonical stepwise mechanism to assemble a Crm1-export complex. A mutation in Slx9 that impairs Crm1-export complex assembly inhibits 40S pre-ribosome export. Thus, Slx9 functions as a scaffold to optimally present RanGTP and the NES to Crm1, therefore, triggering 40S pre-ribosome export. This mechanism could represent one solution to the paradox of weak binding events underlying rapid Crm1-mediated export. https://doi.org/10.7554/eLife.05745.001 eLife digest Plants, fungi, and animals store their genetic material within the nucleus of each of their cells. This structure is surrounded by a double layer of membrane that prevents the contents of the nucleus from mixing with the contents of the rest of the cell (namely the cytoplasm). Exchange of material between the nucleus and cytoplasm occurs through pores embedded within the nuclear membrane. To travel through one of these pores, large molecules (also called cargos) require the assistance of so-called ‘transport receptors’ such as the Crm1 protein. This protein recognizes and binds to the part of a cargo molecule called a ‘nuclear export signal’, and the Crm1 protein also binds to another protein called RanGTP. Nuclear export signals bind weakly to Crm1, which in turn ensures that these cargos are easily released in the cytoplasm once transport is completed. However, this weak binding means that it has remained a mystery how Crm1 is able to efficiently transport cargos out of the nucleus to begin with. Now, Fischer et al. have analyzed how one cargo that contains a nuclear export signal, namely molecules called ribosome precursors, assembles with Crm1. The experiments identified another protein called Slx9 that shuttles rapidly between the inside and the outside of the nucleus. Fischer et al. observed that Slx9 binds directly to RanGTP and brings it together with the ribosome precursor cargo. When a complex of Slx9, RanGTP, and cargo is assembled, it further recruits Crm1 to the cargo. Thus, Slx9 acts as a scaffold to bring cargo into contact with Crm1 and RanGTP, and a tight complex is formed that enables the export of the cargo out of the nucleus. Yeast cells lacking Slx9 delay export of ribosome precursors out of the nucleus. These findings imply the existence of yet-unidentified proteins like Slx9 that help Crm1 to rapidly transport diverse cargos out from the nucleus and into the cytoplasm. https://doi.org/10.7554/eLife.05745.002 Introduction In all eukaryotes, transport between the nucleus and the cytoplasm is channeled through nuclear pore complexes (NPCs) embedded within the nuclear envelope (Tran and Wente, 2006; D'Angelo and Hetzer, 2008). Yeast NPCs are approximately 60 MDa (Fernandez-Martinez et al., 2012) and are composed of multiple copies of about 30 nucleoporins (Rout et al., 2000). Cargos of different sizes and charges pass through the central transport channel of the NPC, which is filled with a meshwork of natively unfolded Phe-Gly (FG)-repeats present in FG-nucleoporins (Frey and Görlich, 2007; Terry and Wente, 2009). These nucleoporins generate a permeability barrier that allows passive diffusion of small molecules, such as ions and metabolites (Frey and Görlich, 2007; Patel et al., 2007). Macromolecules (>40 kDa) require the assistance of nuclear transport receptors, including members of the importin-β-like family, to efficiently overcome this selectivity barrier (Macara, 2001; Ribbeck and Görlich, 2001; Rout and Aitchison, 2001). These transport receptors, also termed importins and exportins, mediate the majority of molecular exchange between the nucleus and cytoplasm (Cook et al., 2007). Transport receptors recognize their cargo via specific signal sequences (Cook and Conti, 2010; Xu et al., 2010; Güttler and Görlich, 2011) and translocate them through the NPC by transiently interacting with FG-repeats. The small GTPase Ran coordinates the movement of importins and exportins between the nucleus and the cytoplasm, and directs the compartment-specific binding and release of transported cargos (Fried and Kutay, 2003; Pemberton and Paschal, 2005; Cook et al., 2007). Ran exists in both GDP- and GTP-bound forms, and the two states are asymmetrically distributed, with RanGTP significantly enriched in the nucleus (Nakielny and Dreyfuss, 1999; reviewed in Görlich and Kutay, 1999). This gradient of RanGTP is established through the spatial separation of regulators of the Ran-cycle (Izaurralde et al., 1997). Whereas the Ran guanine nucleotide exchange factor, RCC1 (Prp20 in yeast) (Bischoff and Ponstingl, 1991; Fleischmann et al., 1991), localizes to the nucleus, the Ran GTPase-activating protein, RanGAP1 (Rna1 in yeast), is found in the cytoplasm (Hopper et al., 1990; Matunis et al., 1996). Interactions between RanGTP and transport receptors are crucial for the directionality of nucleocytoplasmic exchange (Nachury and Weis, 1999). In the nucleus, RanGTP induces the release of imported cargos from importins (Rexach and Blobel, 1995; Görlich et al., 1996). In addition, RanGTP promotes the interaction of cargos with exportins for their transport to the cytoplasm (Fornerod et al., 1997; Kutay et al., 1997; Stade et al., 1997; Solsbacher et al., 1998). In budding yeast, ribosome assembly accounts for a major proportion of the nucleocytoplasmic transport (Rout et al., 1997; Schlenstedt et al., 1997; Sydorskyy et al., 2003; Kressler et al., 2012; Schütz et al., 2014). mRNAs encoding ribosomal proteins (r-proteins) are exported into the cytoplasm. Newly synthesized r-proteins are imported into the nucleus and then targeted to the nucleolus for incorporation into nascent pre-ribosomes (Schütz et al., 2014). Additionally, >300 transiently interacting non-ribosomal assembly factors aid the construction and maturation of ribosomes (Bassler et al., 2001; Dragon et al., 2002; Grandi et al., 2002; Schäfer et al., 2003; Gerhardy et al., 2014). Correctly assembled pre-ribosomal particles are transported through NPCs into the cytoplasm (Tschochner and Hurt, 2003; Panse and Johnson, 2010). In addition to other cargos, it is estimated that each yeast NPC facilitates the export of ∼25 pre-ribosomal particles every minute (Warner, 1999). Transporting pre-ribosomal cargos from the nucleus through the NPC into the cytoplasm, therefore, represents a major task for the export machinery. The Ran-cycle-dependent exportin Crm1 plays an essential role in exporting pre-ribosomal particles to the cytoplasm (Hurt et al., 1999; Moy and Silver, 1999; Gadal et al., 2001; Johnson et al., 2002). Crm1 recognizes and directly binds leucine-rich nuclear export signals (NESs) on cargos in the presence of RanGTP to form a Crm1-export complex (Fornerod et al., 1997). Although an essential NES-containing adaptor, Nmd3, has been identified for 60S pre-ribosome export (Ho and Johnson, 1999; Ho et al., 2000; Gadal et al., 2001), a similar adaptor to recruit Crm1 to the 40S pre-ribosome remains elusive. It has been suggested that multiple NES-containing adaptors such as Ltv1 and Rio2 recruit Crm1 (Seiser et al., 2006; Zemp et al., 2009; Merwin et al., 2014), thereby guaranteeing efficient 40S pre-ribosome export. The essential mRNA and 60S pre-ribosome transport receptor Mex67-Mtr2, which does not directly utilize the RanGTP gradient, also facilitates nuclear export of the 40S pre-ribosomal cargo (Faza et al., 2012). Despite the identification of several components of the export machinery, assembly steps and mechanisms that prepare the pre-ribosomal cargo for transport through the NPC remain largely unexplored. To initiate export, Crm1 must cooperatively bind RanGTP and its NES-containing cargo in the nucleus, to form a trimeric export complex (Petosa et al., 2004; Dong et al., 2009; Monecke et al., 2013). NESs have evolved to maintain relatively low affinity to Crm1 to avoid defects in disassembly of the export complex in the cytoplasm (Engelsma et al., 2004; Kutay and Güttinger, 2005). Mechanisms that promote Crm1-export complex assembly and thereby ensure export of the NES-containing cargo at a reasonable rate remain poorly understood. Here, we identify yeast Slx9 as a new type of RanGTP-binding protein that promotes assembly of a Crm1-export complex on the 40S pre-ribosome-associated NES-containing adaptor Rio2. Our data raise the possibility of a yet-unidentified family of RanGTP-binding proteins that act as scaffolds to optimally present RanGTP and NES-containing cargos to Crm1, orchestrating a non-cooperative stepwise assembly that drives fast and efficient Crm1-mediated export. Results slx9-1 causes defects in 40S pre-ribosome nuclear export Slx9 is a 24-kDa basic protein that co-enriches with pre-ribosomal particles in the 40S maturation pathway (Gavin et al., 2002; Faza et al., 2012) and is required for efficient nuclear export of 40S pre-ribosomes (Li et al., 2009; Faza et al., 2012). However, the precise contribution of Slx9 to 40S pre-ribosome export has remained unclear. To investigate the function of yeast Slx9, we generated slx9 variants by random mutagenesis and analyzed the growth of the resulting strains at different temperatures. One allele, slx9L108P, hereafter termed slx9-1, caused slow growth at temperatures between 20°C and 30°C, indistinguishable from slx9∆ cells (Figure 1A, top panel). Like slx9∆, slx9-1 cells were not impaired in growth at 37°C (Figure 1A). Western analysis of whole cell lysates revealed that Slx9 and Slx9-1 were present at similar levels (Figure 1A, bottom panel), indicating that impaired growth of the slx9-1 strain is not due to reduced levels of the mutant protein. As previously observed, Slx9-GFP localized primarily to the nucleolus, where it co-localized with the nucleolar marker Gar1-mCherry, as well as to the nucleoplasm (Faza et al., 2012 and Figure 1B). Slx9-1-GFP displayed an identical localization (Figure 1B), indicating that the mutant protein is correctly targeted to the nucleolus and nucleoplasm. Slx9 maximally co-enriched with Enp1-TAP that purifies both the 90S and 40S pre-ribosomes (Faza et al., 2012). A similar purification from slx9-1 cells revealed that Enp1-TAP co-enriched at least as much Slx9-1 mutant protein as Slx9 (Figure 1C). Together, these data show that Slx9-1 is correctly expressed, localized, and recruited to 40S pre-ribosomes. Figure 1 Download asset Open asset slx9-1 phenocopies the slx9∆ mutation. (A) The slx9-1 allele does not complement the slow growth of slx9∆ cells. Top: SLX9, slx9∆, and slx9-1 cells were spotted in 10-fold dilutions on SD-plates and grown at the indicated temperatures for 3–6 days. Bottom: Slx9 protein levels from whole cell extracts derived from the indicated strains were determined by Western analysis using antibodies directed against Slx9. Levels of the protein Arc1 served as a loading control. (B) Slx9-1 localizes to the nucleolus/nucleoplasm. Cells expressing Gar1-mCherry and Slx9-GFP or Slx9-1-GFP were grown until mid-log phase. Localization of the indicated fusion proteins was analyzed by fluorescence microscopy. Gar1-mCherry served as a nucleolar marker. Scale bar = 5 µm. (C) Slx9-1 is recruited to the early 40S pre-ribosome. Enp1-TAP was isolated by tandem affinity purification (TAP) from the indicated strains. Calmodulin-eluates were separated on a 4–12% gradient gel and analyzed by either silver staining or Western using the indicated antibodies. The ribosomal protein uS7 served as a loading control. (D) slx9-1 cells are impaired in nuclear export of 40S pre-ribosomes. Top: localization of uS5-GFP was monitored by fluorescence microscopy. Bottom: localization of 20S pre-rRNA was analyzed by FISH using a Cy3-labeled oligonucleotide complementary to the 5′ portion of ITS1 (red). Nuclear and mitochondrial DNA was stained by DAPI (blue). Scale bar = 5 µm. https://doi.org/10.7554/eLife.05745.003 Previous studies showed that slx9∆ cells accumulate the small subunit reporter uS5-GFP (yeast Rps2, nomenclature according to Ban et al., 2014) and 20S pre-rRNA in the nucleoplasm (Li et al., 2009; Faza et al., 2012), indicating a defect in 40S pre-ribosome export. Using these reporters, we tested whether slx9-1 cells have defects in 40S pre-ribosome export. Whereas WT cells displayed cytoplasmic uS5-GFP localization, slx9-1 cells showed a strong nuclear accumulation of this reporter, similar to that observed in slx9∆ cells (Faza et al., 2012 and Figure 1D, top panel). As expected, fluorescence in situ hybridization (FISH) of 20S pre-rRNA in WT cells showed a strong nucleolar Cy3-ITS1 signal (red) with virtually no nucleoplasmic staining. In contrast, slx9-1 cells displayed a nucleoplasmic signal of Cy3-ITS1 localization, which co-localized with the DAPI signal (Figure 1D, bottom panel). These data indicate that slx9-1 cells, like slx9∆ cells (Faza et al., 2012), are impaired in 40S pre-ribosome export. Therefore, we conclude that Slx9-1 is recruited to the 40S pre-ribosome but fails to fulfill its function in nuclear export of the pre-ribosomal cargo. Slx9 is a shuttling RanGTP-binding protein Mutations in MEX67 and MTR2 (mex67Δloop and mtr2Δloop116-137), which encode the essential transport receptor Mex67-Mtr2, are synthetically lethal when combined with the slx9Δ mutant (Faza et al., 2012). In addition, we found that slx9∆ displayed a synthetic growth defect with a strain expressing Rrp12-GFP (Figure 2A). Rrp12 is a 40S pre-ribosome export factor that directly interacts with FG-rich nucleoporins (Oeffinger et al., 2004). Based on these genetic interactions, we asked whether Slx9 functions as a novel export factor for the 40S pre-ribosome. A salient feature of an export factor is that it rapidly shuttles between the nucleus and the cytoplasm. To test this, we employed the established heterokaryon assay (Altvater et al., 2014). WT cells expressing Slx9-GFP were mated to kar1-1 cells, which are deficient in nuclear fusion after cell conjugation, leading to heterokaryon formation. In order to distinguish between the two nuclei, kar1-1 cells also contained Nup82-mCherry as a marker for nuclear pores. As controls, we used the shuttling 40S assembly factor Enp1 and the non-shuttling nucleolar protein Gar1 fused to GFP. Whereas Gar1-GFP was never seen in the nucleus of kar1-1 cells (red signal), both Enp1-GFP and Slx9-GFP localized to both nuclei (Figure 2B). These data are consistent with the shuttling of Slx9 between the nuclear and the cytoplasmic compartments. Figure 2 Download asset Open asset Slx9 is a RanGTP binding protein. (A) slx9-1 genetically interacts with factors involved in 40S pre-ribosome export. slx9-1 is synthetically lethal with mex67∆loop, mtr2∆loop116-137, or yrb2∆ and strongly synthetically enhanced with rrp12-GFP. Strains containing the indicated WT and mutant alleles were spotted in 10-fold serial dilutions on 5-FOA-SD or SD and grown at 20–30°C for 3–6 days. (B) Slx9 shuttles between the nucleus and the cytoplasm. Cells expressing Enp1-GFP, Gar1-GFP, or Slx9-GFP were mated with kar1-1 cells expressing Nup82-mCherry. The resulting heterokaryons were analyzed by fluorescence microscopy. Scale bar = 5 µm. (C) Slx9 directly binds to RanGTP. GST-Slx9 or GST-Ssb1C was immobilized on GSH-Sepharose before incubating with either buffer alone or buffer containing 2 µM His6-RanQLGTP, 50 nM Crm1-His6 or 2 µM His6-RanQLGTP, and 50 nM Crm1-His6. After washing, bound proteins were eluted in LDS sample buffer, separated by SDS-PAGE and visualized by Coomassie staining or Western blotting using the indicated antibodies. L = input. (D) Slx9 specifically interacts with the GTP-bound form of Ran. GST-Slx9, GST-Yrb1, or GST-Ntf2 was immobilized on GSH-Sepharose and incubated with buffer alone or 2 µM His6-Ran loaded with GDP or GTP. Analysis of the eluted proteins was carried out as described in (C). L = input. (E) Slx9-1 binding to RanGTP is impaired. Top: GST-Slx9 or GST-Slx9-1 immobilized on GSH-Sepharose was incubated with buffer alone or 2 µM His6-RanQLGTP. Analysis of the eluted proteins was carried out as described in (C). L = input. Bottom: bar graph depicts the bound His6-RanQLGTP Western blot signal normalized to GST-Slx9 and GST-Slx9-1 levels, respectively. Four independent experiments were performed and Western blots were quantified by software ImageJ (Version 1.44o). Error bars (S.D.) are indicated. https://doi.org/10.7554/eLife.05745.004 We wondered whether Slx9 functions directly in 40S pre-ribosome export as a NES-containing adaptor for the exportin Xpo1 (hereafter termed Crm1). We, therefore, investigated whether Slx9 binds Crm1 in the presence of Gsp1Q71L (Maurer et al., 2001) in vitro (equivalent to the human RanQ69L GTP-stabilized mutant, hereafter termed RanQLGTP; Bischoff et al., 1994). The C-terminal domain of Ssb1 (Ssb1C) that contains a functional NES (Shulga et al., 1999) served as a positive control. Unlike Ssb1C (Maurer et al., 2001 and Figure 2C, lane 8), Slx9 was unable to form a trimeric export complex with Crm1 and RanQLGTP (Figure 2C, lane 4). Surprisingly, these studies revealed that Slx9 directly bound RanQLGTP (Figure 2C, lane 2 and 4). Therefore, although Slx9 does not contain a functional NES, it is a Ran-binding protein. Since Slx9 is a shuttling protein, we tested whether Slx9 interacts with both RanGTP and RanGDP in vitro. As controls, we used Ntf2, the import factor for RanGDP (Ribbeck et al., 1998; Smith et al., 1998 and Figure 2D, lane 8) and the yeast RanBP1 homolog Yrb1 that binds to RanGTP (Schlenstedt et al., 1995 and Figure 2D, lane 6). We found that, like Yrb1, Slx9 interacted exclusively with RanGTP (Figure 2D, lanes 2 and 3). Based on these data, we conclude that Slx9 is a shuttling RanGTP-binding protein. Slx9-1 is impaired in binding RanQLGTP The slx9-1 mutant did not rescue the slow growth and impaired 40S pre-ribosome export of slx9∆ cells (Figure 1A). Furthermore, like slx9∆ cells, slx9-1 cells genetically interacted with mex67 and mtr2 mutants (mex67∆loop and mtr2∆116-137) and rrp12-GFP (Figure 2A). These findings prompted us to test whether Slx9-1 binds to RanQLGTP in vitro, using the assay described above. We found a decrease of approximately 50% in the levels of RanQLGTP bound to Slx9-1 as compared to Slx9 (Figure 2E). Based on these data, we conclude that Slx9-1 is modestly impaired in binding RanQLGTP. The basic patch of RanGTP contributes to Slx9 binding A conserved basic patch on Ran is involved in the interaction with known Ran-binding proteins (Nilsson et al., 2001). Based on homology to human Ran, arginine 142, and lysine 143 in yeast Ran were mutated to alanine (RanQLGTPRKAA) or glutamate (RanQLGTPRKEE) and the contribution of this basic patch to Slx9:RanGTP complex formation was analyzed in vitro. In agreement with previous studies (Nilsson et al., 2001), these RanQLGTP mutants bound weakly to the importin β-like transport receptor, Kap123 (Figure 3A, compare lane 10 with lanes 11 and 12), and interacted more strongly with the RanBP1 homolog Yrb1 (Figure 3A, compare lane 6 with lanes 7 and 8). Pull down studies of Slx9 and these Ran mutants showed that the interactions between Slx9 and these two Ran mutants were impaired, with the charge reversal mutant having a more severe effect than the alanine mutant (Figure 3A, compare lane 2 with lanes 3 and 4). Altogether, these results suggest that, similar to Kap123, Slx9 binding to RanQLGTP involves the basic patch. Figure 3 Download asset Open asset The basic patch and acidic tail of Ran modulates interactions with Slx9. (A) The basic patch of RanQLGTP contributes to Slx9 binding. GST-Slx9, GST-Yrb1, or GST-Kap123 immobilized on GSH-Sepharose was incubated with buffer alone or 2 µM Ran (His6-RanQLGTP, His6-RanQLGTPRKAA, or His6-RanQLGTPRKEE). After washing, bound proteins were eluted in LDS sample buffer, separated by SDS-PAGE and visualized by Coomassie staining or Western blotting using the indicated antibody. L = input. (B) The acidic tail of RanQLGTP negatively regulates interactions with Slx9. GST-Slx9, GST-Yrb1, or GST-Kap123 immobilized on GSH-Sepharose was incubated with buffer alone or 2 µM Ran (His6-RanQLGTP or His6-Ran∆CQLGTP). Analysis of the eluted proteins was carried out as described in (A). L = input. https://doi.org/10.7554/eLife.05745.005 The acidic C-terminal tail of Ran reduces Slx9-binding The C-terminal acidic tail of Ran (-DEDDAL) plays a crucial role in the interaction with small RanGTP-binding proteins such as Yrb1. As expected, RanQLGTP lacking the C-terminal acidic tail (Ran∆CQL GTP) failed to interact with Yrb1 in vitro (Maurer et al., 2001 and Figure 3B, lane 6). We tested whether the acidic tail contributes to the interaction between RanGTP and Slx9. In contrast to Yrb1, Ran∆CQLGTP bound stronger to Slx9 compared to RanQLGTP (Figure 3B, compare lane 2 and 3). This enhanced interaction was specific, since Kap123 bound Ran∆CQLGTP and RanQLGTP to a similar extent (Figure 3B, compare lanes 8 and 9). These data suggest that the C-terminal acidic tail of Ran negatively regulates RanGTP:Slx9 interactions. SLX9 genetically interacts with the RanGTP- and Crm1-binding protein Yrb2 Cells lacking the RanGTP- and Crm1-binding protein Yrb2 (yrb2Δ) exhibit strong nucleoplasmic accumulation of uS5-GFP and 20S pre-rRNA as well as reduced abundance of 40S subunits (Moy and Silver, 2002; Altvater et al., 2012). In addition, mex67Δloop and mtr2Δloop116-137 are synthetically lethal with yrb2Δ (Faza et al., 2012). These findings led us to test whether SLX9 genetically interacts with YRB2. Both slx9∆ and slx9-1 were synthetically lethal with yrb2∆ (Figure 2A), suggesting that Slx9 and Yrb2 functionally overlap to ensure proper nuclear export of 40S pre-ribosomes. Slx9 binds the NES-containing 40S pre-ribosomal adaptor Rio2 Yrb2 and its human homolog, RanBP3, stimulate the assembly of Crm1-export complexes on certain NES-containing cargos by cooperatively binding Crm1 and RanGTP (Englmeier et al., 2001; Lindsay et al., 2001; Koyama et al., 2014). The strong genetic interaction between SLX9 and YRB2 raised the possibility that Slx9 also functions in Crm1-complex assembly. However, our interaction studies showed that Slx9 binds RanGTP, but not Crm1 (Figure 2C, lane 4). We, therefore, wondered whether Slx9 instead facilitates the assembly of a Crm1-export complex by bringing together RanGTP and the NES-containing adaptor. Two 40S pre-ribosome-associated factors, hLtv1 and hRio2, have been shown to bind Crm1 in the presence of RanGTP (Zemp et al., 2009). In agreement with these studies, we found that yeast Ltv1 and yeast Rio2 formed trimeric complexes with Crm1 and RanQLGTP via a cooperative mechanism in vitro (Figure 4A, lanes 4 and 12). Moreover, the C-terminal regions of these proteins are predicted to contain a NES (Zemp et al., 2009; Merwin et al., 2014 and Figure 4A, top panel), and indeed, Rio2 and Ltv1 mutant proteins lacking NESs were unable to form trimeric export complexes (Figure 4A, lanes 8 and 16). Western analyses revealed that Rio2, but not Ltv1, interacted weakly with RanQLGTP, independent of Crm1 (Figure 4A, lane 2). Figure 4 Download asset Open asset Slx9 directly binds the 40S pre-ribosome nuclear export signal (NES)-containing adaptor Rio2 and RanQLGTP. (A) Rio2 and Ltv1 export complex formation requires their C-terminal NESs. Top: the positions of the Rio2 and Ltv1 NESs are shown. Hydrophobic residues in these NESs are highlighted in red. Bottom: GST-Rio2, GST-Rio2∆NES, GST-Ltv1, or GST-Ltv1∆NES was immobilized on GSH-Sepharose, and complex formation was analyzed as in Figure 2C. L = input. (B) Slx9 directly interacts with Rio2. Immobilized GST-Rio2 or GST-Ltv1 was incubated with buffer alone or 0.5 µM Slx9. Conversely, immobilized GST-Slx9 was incubated with buffer alone or with lysate containing His6-Nmd3 or His6-Rio2. Analysis of the eluted proteins was carried out as described in Figure 2C. L = input. https://doi.org/10.7554/eLife.05745.006 Ltv1 and Rio2 co-enrich with the Enp1-TAP particle that also maximally co-purifies Slx9 (Figure 1C; Schütz et al., 2014) suggesting that these NES-containing proteins might interact with Slx9. To test this, we incubated immobilized GST-Rio2 or GST-Ltv1 with Slx9. Slx9 directly bound to Rio2 but not to Ltv1 (Figure 4B, left panel). Conversely, GST-Slx9 interacted with Rio2 but not with the essential 60S pre-ribosome-associated NES-containing adaptor Nmd3 (Figure 4B, right panel). Moreover, the GST-Rio2:Slx9 complex efficiently recruited RanQLGTP (Figure 5A). The level of recruitment was not affected by the prior presence or absence of Slx9, since GST-Rio2 saturated with Slx9 recruited similar amounts of RanQLGTP (Figure 5B, compare top and bottom panels). Also, the levels of RanQLGTP recruitment to GST-Slx9 were not affected by the presence or absence of Rio2 (Figure 5C, compare top and bottom panels). Altogether, these data suggest that Rio2 binds Slx9 and RanQLGTP using distinct surfaces. Figure 5 Download asset Open asset Slx9 binds to Rio2 and RanGTP using distinct binding surfaces. (A) GST-Rio2 was immobilized on GSH-Sepharose and incubated with buffer, 2 µM His6-RanQLGTP, or 0.5 µM Slx9 (+1). After washing, the GST-Rio2:Slx9 complex was incubated with 2 µM His6-RanQLGTP (+2). Analysis of the eluted proteins was carried out as described in Figure 2C. L = input. (B) RanGTP does not displace Slx9 from a preformed GST-Rio2:Slx9 complex. Top: immobilized GST-Rio2 was incubated with buffer or increasing concentrations of His6-RanQLGTP (62.5 nM–32 µM). Bottom: immobilized GST-Rio2 was incubated with either buffer or 1 µM Slx9. The unbound Slx9 was washed away, and the resulting GST-Rio2:Slx9 complex was incubated with increasing concentrations of His6-RanQLGTP (62.5 nM–32 µM). Analysis of the eluted proteins was carried out as described in Figure 2C. L = input. (C) RanQLGTP does not displace Rio2 from a preformed GST-Slx9:Rio2 complex. Top: immobilized GST-Slx9 was incubated with buffer or increasing concentrations of His6-RanQLGTP (62.5 nM–32 µM). Bottom: immobilized GST-Slx9 was incubated with excess of Rio2. The unbound Rio2 was washed away, and the resulting complex GST-Slx9:Rio2 complex was incubated with increasing concentrations of His6-RanQLGTP (62.5 nM–32 µM). Analysis of the eluted proteins was carried out as described in Figure 2C. L = input. https://doi.org/10.7554/eLife.05745.007 Recruitment of Crm1 to Rio2:RanGTP is stimulated by Slx9 We next investigated whether the GST-Rio2:Slx9:RanQLGTP complex could directly recruit Crm1. To this end, pre-formed GST-Rio2:RanQLGTP or GST-Rio2:Slx9:RanQLGTP complexes (summarized in Figure 6A) were incubated with buffer alone or Crm1 (Figure 6B, lanes 3, 4 and 7, 8). Only the GST-Rio2:Slx9:RanQLGTP complex efficiently recruited Crm1 (Figure 6B, compare lanes 4 and 8). The Crm1-recruitment to a GST-Rio2:Slx9:RanQLGTP complex was dependent on the NES of Rio2, since a GST-Rio2∆NES:Slx9:RanQLGTP complex was unable to bind Crm1 (Figure 6B, compare lanes 8 and 10). Moreover, Crm1 recruitment was also dependent on the RanQLGTP bound to Rio2, since a GST-Rio2:Slx9 complex was unable to bind Crm1 (Figure 6C, lane 4). These studies indicate that, in order to recruit Crm1 in a non-cooperative manner, Rio2 must bind to both Slx9 and RanGTP. Figure 6 Download asset Open asset Slx9 promotes stepwise assembly of a Crm1-export complex on the NES of Rio2. (A) Flow chart depicting the experimental setup to assemble a Rio2:Slx9:RanQLGTP:Crm1 complex. Immobilized GST-Rio2 was sequentially incubated with Slx9 (red), RanQLGTP (purple), and Crm1 (green). Unbound protein was washed away after each incubation step. (B) Crm1 is recruited to the GST-Rio2:Slx9:RanGTP complex in a NES-dependent manner. Immobilized GST-Rio2 or GST-Rio2∆NES was incubated with buffer alone or 0.5 µM Slx9, followed by the stepwise addition of 0.2 µM His6-RanQLGTP and 50 nM Crm1-His6, as depicted in (A). After a final washing step, bound proteins were analyzed as in Figure 2C. L = input. (C) Crm1 is not recruited to the GST-Rio2:Slx9 complex. Immobilized GST-Rio2 was incubated with buffer alone or 0.5 µM Slx9, followed by addition of buffer, 50 nM Crm1-His6, or the stepwise addition of 0.2 µM His6-RanQLGTP and 50 nM Crm1-His6 as depicted in (A). Analysis of the bound proteins was carried out as described in Figure 2C. L = input. (D) Recruitment of Crm1 to a Rio2:Slx9-1:RanQLGTP complex is impaired. Immobilized GST-Rio2 was incubated with buffer alone, 0.5 µM Slx9 or 0.5 µM Slx9-1, followed by the stepwise addition of 0.2 µM His6-RanQLGTP and 50 nM Crm1-His6 as depicted in (A). Analysis of the bound p
The evolutionarily conserved protein Sem1/Dss1 is a subunit of the regulatory particle (RP) of the proteasome, and, in mammalian cells, binds the tumor suppressor protein BRCA2. Here, we describe a new function for yeast Sem1. We show that sem1 mutants are impaired in messenger RNA (mRNA) export and transcription elongation, and induce strong transcription-associated hyper-recombination phenotypes. Importantly, Sem1, independent of the RP, is functionally linked to the mRNA export pathway. Biochemical analyses revealed that, in addition to the RP, Sem1 coenriches with components of two other multisubunit complexes: the nuclear pore complex (NPC)-associated TREX-2 complex that is required for transcription-coupled mRNA export, and the COP9 signalosome, which is involved in deneddylation. Notably, targeting of Thp1, a TREX-2 component, to the NPC is perturbed in a sem1 mutant. These findings reveal an unexpected nonproteasomal function of Sem1 in mRNA export and in prevention of transcription-associated genome instability. Thus, Sem1 is a versatile protein that might stabilize multiple protein complexes involved in diverse pathways.