|Home | About | Journals | Submit | Contact Us | Français|
Kap123p is a yeast β-karyopherin that imports ribosomal proteins into the nucleus prior to their assembly into preribosomal particles. Surprisingly, Kap123p is not essential for growth, under normal conditions. To further explore the role of Kap123p in nucleocytoplasmic transport and ribosome biogenesis, we performed a synthetic fitness screen designed to identify genes that interact with KAP123. Through this analysis we have identified three other karyopherins, Pse1p/Kap121p, Sxm1p/Kap108p, and Nmd5p/Kap119p. We propose that, in the absence of Kap123p, these karyopherins are able to supplant Kap123p's role in import. In addition to the karyopherins, we identified Rai1p, a protein previously implicated in rRNA processing. Rai1p is also not essential, but deletion of the RAI1 gene is deleterious to cell growth and causes defects in rRNA processing, which leads to an imbalance of the 60S/40S ratio and the accumulation of halfmers, 40S subunits assembled on polysomes that are unable to form functional ribosomes. Rai1p localizes predominantly to the nucleus, where it physically interacts with Rat1p and pre-60S ribosomal subunits. Analysis of the rai1/kap123 double mutant strain suggests that the observed genetic interaction results from an inability to efficiently export pre-60S subunits from the nucleus, which arises from a combination of compromised Kap123p-mediated nuclear import of the essential 60S ribosomal subunit export factor, Nmd3p, and a ΔRAI1-induced decrease in the overall biogenesis efficiency.
In eukaryotic cells, macromolecules are actively and selectively transported between the nucleus and cytoplasm by the concerted action of soluble transport factors and the nuclear pore complex (NPC). Proteins destined for the nucleus carry a nuclear localization signal (NLS) (13), while substrates to be exported from the nucleus harbor nuclear export signals (NESs) (18, 21). The signals are recognized by a structurally related family of proteins, termed karyopherins (abbreviated as Kaps, but also known as importins or exportins) (22, 41, 57), that interact with the NPC and the small GTPase Ran to mediate translocation. The NPC is a large, octagonally symmetric structure, and its overall architecture is highly conserved from yeast to metazoans (reviewed in references 5, 44, 53, and 56). The yeast NPC contains multiple copies of ~30 protein components, termed nucleoporins or Nups. Twelve of these Nups contain characteristic degenerate repeated peptide motifs (GLFG, FXFG, PSFG, or FG) and are thus collectively termed FG-Nups. These nucleoporins provide multiple docking sites for cargo-bearing transporters and are present throughout the NPC, extending from the cytoplasmic filaments to the nuclear basket (45). It remains a mystery how the interactions between karyopherins and the NPC mediate vectorial transport; however, karyopherins appear to derive directional cues from both specific nucleoporins and Ran (reviewed in references 10, 40, 44, and 56).
In a rapidly growing cell, the complex process of ribosome biogenesis accounts for a major proportion of nucleocytoplasmic transport. Like all messenger transcripts, mRNAs encoding ribosomal proteins are exported to the cytoplasm, where translation occurs. Once synthesized, ribosomal proteins are imported into the nucleus and coordinately assembled with nascent rRNA to yield immature ribosomal subunits. These subunits are then exported to the cytoplasm where they undergo the final stages of maturation and assemble onto mRNA to carry out translation. Yeast cells that divide every 1.5 h must accordingly double their ribosomal content in this time frame, necessitating the import of ~150,000 ribosomal proteins per min (~1,000/NPC), while simultaneously exporting ~4,000 assembled subunits per min (55).
In Saccharomyces cerevisiae, it appears that the import leg of this biogenesis program is accomplished largely by the karyopherin Kap123p. Kap123p binds to many different ribosomal proteins including rpL25, rpS1a, rpL8a/b, rpL18a/b, rpL12a/b, rpL32, rpL11a/b, and rpL42a/b, and at least rpL25 requires Kap123p to be efficiently imported into the nucleus (46, 48). Interestingly, deletion of KAP123 from the yeast genome does not dramatically affect cell growth; thus, it is apparent that there are other ribosomal protein importers in yeast. One such candidate is Kap121p, which, among the karyopherin family members, is most similar to Kap123p. Kap121p also binds to several ribosomal proteins in the absence of Kap123p and, when overexpressed, suppresses the rpL25 import defect observed in Δkap123 strains (46).
By comparison to the import leg, ribosomal subunit export is less straightforward, as assembly and export appear to be temporally and physically coupled. Initially, the 35S rRNA is transcribed by RNA polymerase I and the 5S rRNA is separately transcribed by RNA polymerase III. Over 60 trans-acting factors are responsible for nucleotide modification and maturation of the ribosomal subunits prior to their export to the cytoplasm (17, 34). In short, several early associating large and small subunit proteins, and both proteineaceous and small nucleolar RNA trans-acting factors, assemble onto the primary 35S transcript, yielding a 90S preribosomal particle. The 35S rRNA is spliced and trimmed by the consecutive action of exo- and endonucleases to produce three products: a 20S rRNA within a 43S precursor particle, and 5.8S and 25S rRNAs that, together with the RNA polymerase III-transcribed 5S rRNA, assemble into the pre-60S particle. These particles are then independently exported through the NPC to the cytoplasm by a receptor-mediated, energy-dependent process (7), where they undergo final maturation and assembly onto mRNA. By monitoring the assembly of green fluorescent protein (GFP)-tagged ribosomal proteins into ribosomal subunits in specific yeast mutants, it has been established that the Ran cycle and several nucleoporins and karyopherins are required for this process (27, 52). Unfortunately, the intimate links between ribosomal subunit import, assembly, and export have made it difficult to ascertain the functional role of specific factors in the export process.
To further understand the import and assembly processes, we employed a genetic interaction screen to identify components in the ribosome biogenesis pathway that interact with the major ribosomal protein nuclear import factor, Kap123p. In this screen, we identified several karyopherins that likely function redundantly to Kap123p in ribosomal protein import. In addition, we identified a mutant allele of the gene encoding Rai1p, a protein previously shown to be involved in 60S subunit biogenesis (58). Interestingly, cells lacking Rai1p and Kap123p, but not other karyopherins assayed, were defective in 60S ribosomal subunit export. Here, we report on the characterization of this interaction and provide a model for the functional link between Kap123p and Rai1p.
The S. cerevisiae strains used in the study were derivatives of the diploid DF5 strain, unless specified otherwise, and are listed in Table Table1.1. All yeast genetic manipulations were performed according to established procedures (23). Gene knockout marker modules were switched as required by using “marker swap” plasmids, as described previously (11).
Deletion of the RAI1 open reading frame (ORF) was accomplished by integrative transformation of DF5 cells with a PCR-synthesized HIS3 marker containing short flanking homology to the upstream and downstream regions of the RAI1 ORF, generated using the following oligonucleotides: RAI1-5′ (5′-CCG GAA TTC AAG CTT ATG GGT GTT AGT GCA AAT TTG-3′) and RAI1-3′ (5′-CGG GAA TTC TTT CAA AGA TTT TCT CCA CTC-3′). Transformants were selected on synthetic complete (SC) plates lacking histidine (SC-His plates) and sporulated, and tetrads were dissected to obtain a haploid Δrai1 strain. Δsxm1, Δnmd5, and Δkap123-w303 strains were obtained in a similar manner, using the following oligonucleotides: SXM1-5′ (5′-AGA GAT TCC TTG CAG GTA ATT CTG GAA TTT GTT TCT CAA CAC GGT GAA GCT CAA AAA CTT AAT-3′), SXM1-3′ (5′-AAA AAG AAA CAA CTT TTA TAT TTG TAT ATT AGA GTA TAA ACT GTC GAC GGT ATC GAT AAG CTT-3′), NMD5-5′ (5′-ACC CCG GCT GAT CAA GAA CTA TTC ATG GGA ATT ATG AAT GCC GGT GAA GCT CAA AAA CTT AAT-3′), NMD5-3′ (5′-AGG CAA CAA ACT TTG AGC ATA ATA TCC TCT CTC TTC TAT CTA GTC GAC GGT ATC GAT AAG CTT-3′), KAP123-5′ (5′-GCG CTT GAG TTG CTT CAA GTG CAT G-3′), and KAP123-3′ (5′-ACA AGA TGA CCT GAA CTT GCG CGT A-3′).
The KAP123 gene was isolated as a genomic 6.85-kb XhoI-XbaI fragment from an S. cerevisiae phagemid library (ATCC 87021) and cloned into the SalI and XhoI sites of pRS316 or pRS317 to yield pRS316-KAP123 or pRS317-KAP123, respectively. The ADE3 gene (nucleotides [nt] 8136 to 11,677 on chromosome VI) was amplified by PCR from yeast genomic DNA to generate a PCR product with SacI linkers which was subsequently cloned into the SacI site of pRS316-123 to yield the pJA8 plasmid.
The RPL2, RPL3, RPL25, and NOP1 ORFs were amplified by PCR from yeast genomic DNA and cloned into the EcoRI and HindIII sites of pYX242, which contains the in-frame coding sequence for Aequoria victoria GFP (43). The resulting recombinant plasmids were termed rpL2-GFP, rpL3-GFP, rpL25-GFP, and Nop1p-GFP, respectively.
The RAI1, YGL247W, PDE1, KAP95, KAP121, SXM1, and NMD5 ORFs were amplified by PCR from yeast genomic DNA and cloned into the PstI and SacI sites of pRS314, and the resulting recombinant plasmids were termed pYGL246C, pYGL247W, Pde1p-pRS314, Kap95p-pRS314, Kap121p-pRS314, Sxm1p-pRS314, and Nmd5p-pRS314, respectively.
The RAI1 ORF was amplified by PCR from yeast genomic DNA and cloned into the EcoRI and HindIII sites of p12-GFP2-NLS (pKW431; a gift from K. Weis, University of California, San Francisco) (51), which resulted in the replacement of the classical NLS sequence with the RAI1 coding sequence, in frame with a mutant NES (p12) followed by two GFP coding sequences. The resulting recombinant plasmid was termed Rai1p-GFP.
Multicopy plasmids pNmd3p and pNmd3Δ100 were constructed by PCR amplification of the NMD3 gene, including 250 nt upstream of the initiation codon in the case of pNmd3p, or including 250 nt upstream but lacking the 3′ 300 coding nucleotides for pNmd3Δ100. These fragments were then ligated into the PstI site of pRS425.
A colony-sectoring assay was used to isolate KAP123 synthetic fitness mutants as described previously (3, 8, 32). The strains chosen for complementation were transformed with an S. cerevisiae genomic cDNA library cloned into the vector pSB32 (provided by J. Rine, University of California, Berkeley ). Leu+ transformants were screened to identify sectoring (s+) colonies. Sec+ colonies were transferred to 5′-fluoroorotic acid (5′-FOA)-containing media (9) to confirm their ability to lose the pJA8 covering plasmid. Plasmids were purified from each complemented strain and sequenced. Genes within the complementing genomic fragments were amplified by PCR and subcloned individually into the plasmid pRS314. The resulting plasmids were transformed into the corresponding mutant to identify the complementing gene.
Cells expressing GFP-tagged proteins were grown in selective media, and the GFP chimeras were visualized directly by fluorescence microscopy using either a Zeiss Axioskop 2 (Carl Zeiss, Inc.) and a Spot camera (Diagnostic Instruments Inc.), or a confocal microscope (LSM 510 NLO; Carl Zeiss, Inc.). Further processing of images was performed using Adobe Photoshop 6.0 (Adobe Systems Inc.).
For Western blot analysis, proteins were resolved on sodium dodecyl sulfate (SDS)-7 to 15% polyacrylamide gels and transferred to nitrocellulose membranes by standard methods (6, 39). The following antibodies were used as primary antibodies: affinity-purified rabbit anti-mouse immunoglobulin G (IgG; Cappel; Organon Teknika Corp.), mouse monoclonal anti-Tcm1p (rpL3) (kindly provided by J. L. Woolford, Carnegie Mellon University, Pittsburgh, Pa.), rabbit polyclonal anti-GFP (kindly provided by M. Rout, The Rockefeller University, New York, N.Y.), and anti-Nop1p (D77) (kindly provided by G. Blobel, The Rockefeller University). Horseradish peroxidase-conjugated anti-mouse and anti-rabbit Ig (Amersham Biotech) were used as secondary antibodies, and immunoreactive bands were visualized using the SuperSignal West Pico chemiluminescent substrate (Pierce).
Immunoprecipitations from yeast whole-cell lysates were performed as previously described (12, 38). After incubation in cell extracts, IgG-Sepharose beads were washed 10 times with 1 ml of wash buffer containing 20 mM Na2HPO4 (pH 7.5), 150 mM NaCl, 0.1 mM MgCl2, 0.1% Tween 20, 4 μg of pepstatin A/ml, and 180 μg of phenylmethylsulfonyl fluoride/ml. Bound proteins were eluted in wash buffer containing either 0.2 or 1 M magnesium chloride and trichloroacetic acid precipitated. Copurified proteins were resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and visualized by Coomassie blue or silver staining (50). Bands of interest were then excised from the gel, subjected to in-gel digestion, and identified by mass spectrometry (25).
RNA was isolated by the hot-phenol technique (6). The analysis of rRNA by Northern hybridization was performed as described previously (28). The following primers were used: A3-B1L (5′-CCA GTT ACG AAA ATT CTT G-3′), A2-A3 (5′-TGT TAC CTC TGG GCC C-3′), E-C2 (5′-GGC CAG CAA TTT CAA GTT A-3′), 25S (5′-CTC CGC TTA TTG ATA TGC-3′), D-A2 (5′-GCT CTC ATG CTC TTG CCA-3′), and 18S (5′-CAT GGC TTA ATC TTT GAG AC-3′).
For polysome preparations, yeast cultures were grown to an optical density at 600 nm (OD600) of 0.6, treated with cycloheximide (100 μg per ml of culture), harvested, and disrupted by glass bead lysis as described previously (19, 33). For each sample, 200 A260 units of cell lysate supernatants was loaded onto a 36-ml linear 7-to-42% sucrose gradient and centrifuged for 6 h at 141,000 × g in a Beckman SW28 rotor at 4°C. Sucrose gradient fractions were analyzed by continuous monitoring at 254 nm with a UV monitor (UV-M II; Pharmacia LKB).
Dissociated ribosomal subunits were analyzed as described above, except that extracts were prepared in buffer containing 10 mM Tris-HCl (pH 7.4), 100 mM NaCl, 30 mM EDTA, and heparin (200 μg/ml). In addition, gradients contained 30 mM EDTA instead of MgCl2.
Nuclei were prepared by the method of Rout and Kilmartin (31, 47) with minor modifications. Cell cultures (4 liters) were harvested at an OD600 of 0.7 to 0.8, washed twice with water, and pretreated in ice-cold buffer containing 100 mM Tris-Cl (pH 9.5) and 10 mM dithiothreitol, after which cells were washed with 1 M sorbitol once and incubated for 3 h at 30°C with gentle shaking in 1 M sorbitol containing 1% Zymolyase 20T (ICN Biomedicals), 1% Mutanase (Sigma), and 10% Glusulase (NEN). Cells were pelleted at 2,000 × g, washed in 1.1 M sorbitol, and resuspended in 1.1 M sorbitol containing solution P (4 μg of pepstatin A/ml and 180 μg of phenylmethylsulfonyl fluoride/ml). Spheroplasts were loaded on a 10-ml 7.5% Ficoll 400-1.1 M sorbitol cushion and pelleted at 4,000 × g for 20 min in a Beckman JS13.1 rotor at 4°C. The sorbitol and Ficoll layers were removed, cell pellets were resuspended in 8% polyvinylpyrrolidone containing 2 mM dithiothreitol, 0.025% Triton X-100, and 1:100 solution P. The samples were homogenized using a Polytron (Brinkman Instruments), and the lysates were loaded onto a cushion containing 0.6 M sucrose, 8% polyvinylpyrrolidone containing solution P. Cytoplasmic supernatants were collected after centrifugation at 14,200 × g for 20 min in a Beckman JS13.1 rotor at 4°C. The pellet was washed by resuspending in a volume of buffer (10 mM bis-Tris [pH 6.5], 1 mM MgCl2, 1:100 solution P) equal to that of the pellet, and the suspension was centrifuged at 27,200 × g for 15 min in a Beckman JA25.5 rotor, at 4°C, to yield a nuclear fraction.
Kap123p plays an integral role in ribosome biogenesis by importing ribosomal proteins into the nucleus prior to their incorporation into assembling preribosomal particles (46, 48). To investigate how cells compensate for the loss of Kap123p and its interaction with other components of the ribosome assembly pathways, we have isolated mutants that, in combination with kap123 null mutants, exhibit a synthetic fitness defect. The genes encoding the β-karyopherins Sxm1p/Kap108p, Nmd5p/Kap119p, and Pse1p/Kap121p complemented strains sfA7 and sfA9, sf5, and sf21 mutants, respectively, suggesting that the nuclear import pathways mediated by these karyopherins overlap with those mediated by Kap123p (Fig. (Fig.1).1). Indeed, genetic and functional interactions have previously been reported between KAP123 and KAP121 (46, 49). Furthermore, Sxm1p/Kap108p interacts with ribosomal proteins, suggesting a role in importing this class of proteins (42). Such functional redundancy between multiple members of this family is underscored by the fact that each of these synthetic fitness mutants was viable, displaying synergistic fitness defects but not synthetic lethality at 30°C. To confirm these results, we constructed strains harboring double knockouts of genes encoding Sxm1p/Kap108p, Nmd5p/Kap119p, and Kap123p in all three combinations, along with a Δkap123/kap121ts double mutant strain. All of these strains were viable, but their growth rates were dramatically reduced (data not shown).
Interestingly, sf17 was not complemented by a gene encoding a karyopherin, but rather by the gene YGL246C/RAI1 (Fig. (Fig.1).1). Rai1p stabilizes and augments the function of Rat1p (58), a 5′-exonuclease required for the trimming of the 27A3 precursor rRNA to the 27SBs and degradation of excised pre-rRNA spacer fragments during 60S ribosomal subunit biogenesis (36, 54). In agreement with the original report on Rai1p (58), deletion of YGL246C was not lethal but compromised cell growth. Moreover, these cells showed no specific temperature sensitivity, growing at all temperatures tested (16, 23, 30, or 37°C) (data not shown). This generalized growth defect was complemented by reintroduction of a plasmid encoding Rai1p or a gene fusion encoding a Rai1p-GFP chimera (Fig. (Fig.1).1). Furthermore, diploid cells resulting from a cross of Δrai1 cells with sf17 cells displayed a similar growth defect as the haploid strains, suggesting that the mutation in sf17 was indeed allelic to RAI1 (data not shown). This mutation was termed rai1-1. To evaluate the specificity of the rai1/kap123 genetic interaction, double mutant strains containing Δrai1/Δsxm1, Δrai1/Δnmd5, and Δrai1/kap121ts were compared to cells containing single deletions of each gene and the rai1-1/Δkap123 and Δrai1/Δkap123 strains. In contrast to the dramatic growth defect associated with the Δrai1/Δkap123 double deletion, Δrai1/Δsxm1, Δrai1/Δnmd5, and Δrai1/kap121-34 cells grew only slightly slower then Δrai1 cells (data not shown). Direct DNA sequencing of the PCR-amplified rai1 gene in sf17 cells revealed an A→T transversion at position 443, resulting in a nonsense mutation of lysine 147 that predicts an expressed product containing approximately the amino-terminal one-third of Rai1p (Fig. (Fig.1).1). Interestingly, this mutation seemed more detrimental to cells than did complete loss of Rai1p, as the Δrai1/Δkap123 strain was not as defective in growth as the rai1-1/Δkap123 strain (data not shown).
Because of Kap123p's well-established role in nucleocytoplasmic transport, we first investigated whether the localization of Rai1p was dependent on Kap123p. Rai1p was therefore C-terminally tagged with GFP and localized by fluorescence microscopy. In wild-type cells, Rai1p-GFP localized primarily to the nucleus but also revealed a distinct cytoplasmic pool. Surprisingly, we detected no change in the Rai1p-GFP nuclear signal in cells lacking Kap123p, Sxm1p/Kap108p, or Nmd5p/Kap119p, or in cells carrying temperature-sensitive mutants of Kap121p (kap121-34) or Kap95p (kap95-14) (data not shown). Furthermore, despite numerous attempts, including immunoaffinity purification, overlay, and solution binding assays, we failed to detect a physical interaction between Rai1p and Kap123p. Together, these data suggest that Rai1p is not imported into the nucleus by Kap123p, and they do not support the hypothesis that the genetic interaction reflects a physical interaction between Rai1p and Kap123p.
To determine the physical interactions made by Rai1p, an Rai1p-pA chimera was generated and affinity purified from whole-cell lysates using IgG-Sepharose. These experiments yielded a single major protein of approximately 120 kDa, which was identified as Rat1p by tandem mass spectrometry of excised gel slices (Fig. (Fig.2).2). This is in agreement with the data from Johnson's group, which also established a physical interaction between Rai1p and Rat1p (58). In those studies, Δrai1 strains were shown to accumulate 27A3 precursors and Rat1p activity was more robust in the presence of interacting Rai1p, suggesting that Rai1p and Rat1p function together in the trimming of the rRNA species to the 27SBs precursor (58). In agreement with this function for Rai1p, Δrai1 and rat1-1 are synthetically lethal (Fig. (Fig.2).2). Interestingly, we also noted that Rai1p-GFP was mislocalized to the cytoplasm in rat1-1 cells (Fig. (Fig.3).3). One scenario to explain this interesting result would be that in the absence of a fully functional Rat1p, Rai1p is no longer effectively tethered to Rat1p in the nucleus and has become free to move within the nucleoplasm and, perhaps, exit the nucleus.
One potential mechanism for this movement and exit may be through an interaction with assembling ribosomes. We therefore tested for such an interaction. Ribosomes, polysomes, and ribosomal subunits were resolved by linear sucrose gradient centrifugation from a Rai1p-pA-expressing strain (RAI1-A) and fractions from the gradient were probed for Rai1p-pA. Under these conditions, Rai1p-pA was found only in the load fraction (data not shown). This suggested that if Rai1p interacts with ribosomal particles, it is either not stable under these conditions or occurs at levels below the level of detection. We therefore sought to increase the concentration of 60S subunits in the nucleus by inhibiting 60S subunit export. To accomplish this, a mutant version of Nmd3p was expressed in RAI1-A cells. Nmd3p was recently identified as an adaptor protein that, through its binding to both the 60S subunit protein rpL10 and the karyopherin Crm1p, mediates the Ran-dependent export of the pre-60S particle from the nucleus (20, 26). It has also been shown that deletion of the C-terminal 100 amino acids of Nmd3p inhibits ribosomal subunit export in a dominant negative fashion (26). Thus, polysomes were fractionated from RAI1-A cells expressing nmd3Δ100 from a plasmid and the presence of Rai1p-pA in each fraction was determined by Western blotting. A monoclonal antibody to wild-type rpL3 (TCM1) was used to identify the 60S, 80S, and polysome fractions. Under these conditions, Rai1p-pA was detected in association with 60S (precursor) subunits (Fig. (Fig.4),4), but not with 40S subunits or polysome fractions. Moreover, affinity purification of Rai1p-pA from lysates derived from cells expressing nmd3Δ100 also showed association of the 60S subunit marker protein rpL3 with the Rai1p chimera (Fig. (Fig.4).4). Further examination of this Rai1p-pA-immunopurified fraction by mass spectrometry demonstrated that several ribosomal proteins associated with Rai1p-pA under these conditions. This included 22 60S ribosomal subunit proteins and 8 (early associating) 40S subunit proteins from a total of 33 identifications (Fig. (Fig.4B).4B). In addition to Rat1p, also associated with the fraction were Las1p and Grc3p (14, 16); however, the potential roles of these proteins in ribosome assembly remain to be investigated.
The number of 60S subunit proteins identified by mass spectrometry suggests that pre-60S particles associate with Rai1p. The source of the relatively few 40S proteins identified is unclear. In proteomics approaches to identify components of purified subunit precursors, small numbers of ribosomal proteins from other subunits are commonly detected. In the study here, this may be because Rai1p can interact with the 60S particles as well as 80S particles, through the 60S subunit. Alternatively, as suggested by Harnipicharnchai et al. (24), some 40S subunit proteins may associate with 60S precursor particles. Interestingly, six of the eight small subunit proteins identified here were also identified in association with 66S precursor particles by Harnipicharnchai et al. (24).
Together, the above data support and extend previous studies demonstrating a role for Rai1p in 60S ribosomal subunit biogenesis; however, they fail to explain the specific genetic interaction observed between rai1 and kap123. Northern blotting of rRNA was used to identify precursors that accumulate in rai1 and kap123 strains (Fig. (Fig.5).5). As expected, rai1 strains accumulated 35S and 27S rRNA species (27SA2 and 27SA3). Interestingly, these precursors accumulated slightly more in sf17 and Δkap123/Δrai1 cells. In addition, Δrai1, sf17, and Δkap123/Δrai1 cells accumulated 7S precursors and a fragment that corresponds to A3-E. The A3-E fragment is proposed to accumulate if the precise order of processing is disrupted (35) and may require Rat1p and/or Rai1p for trimming to 5.8S. On the other hand, the presence of 7S precursors suggests that exosome function is compromised in the absence of Rai1p, but remains intact in the absence of Kap123p. Interestingly, there was also a detectable increase in the 20S signal in cells lacking Kap123p, suggesting a weak 40S biogenesis defect. Together, these data suggest that the loss of Kap123p marginally alters both the 60S and 40S biogenesis pathways. However, because the rRNA processing defect in Δrai1 cells was not dramatically exacerbated or altered by the additional loss of Kap123p, it is not likely that the genetic interaction between kap123 and rai1 is due to a catastrophic block in either the 40S or 60S assembly pathway.
The ribosome and polysome profiles from each relevant strain were also examined (Fig. (Fig.6).6). In agreement with the results of Northern blotting analysis, Δrai1 cells also had a paucity of assembled 60S subunits. This was revealed by an overall decrease in the mature 60S/40S subunit ratio and by the accumulation of halfmers, small shoulders on 80S and polysome peaks which result from kinetically stalled 40S subunits that have threaded onto mRNA but are unable to initiate translation due to a lack of cytoplasmic 60S subunits. In contrast, cells lacking Kap123p showed relatively normal polysome profiles. Interestingly, free 40S subunits and halfmers did not accumulate to the same extent in Δkap123/Δrai1 cells as they did in Δkap123 cells, suggesting that the 40S/60S subunit ratio was somehow normalized in Δkap123/Δrai1 cells. This “normalizing” effect of KAP123 deletion was confirmed by analysis of the 40S/60S subunit ratios (data not shown). Nevertheless, Δkap123/Δrai1 cells, like Δrai1 cells, showed a dramatic decrease in the total ribosomal content (Fig. (Fig.6).6). These data are consistent with the loss of Kap123p leading to an overall decrease in ribosome biogenesis efficiency, which is not specific to either subunit but upstream of Rai1p activity. Because Kap123p interacts with many early and late assembling ribosomal proteins, the observed epistasis likely reflects Kap123p's role in the import of many proteins required for optimal ribosome assembly, perhaps including early assembly factors.
To further investigate this specific Δrai1/Δkap123 genetic interaction, we examined the localization of GFP fusions of three large subunit ribosomal proteins. It has been shown previously that carboxy-terminal GFP tags of some 60S ribosomal subunit proteins are faithfully integrated into functional ribosomes (27, 52); therefore, plasmids encoding C-terminal GFP chimeras of rpL2, rpL3, and rpL25 were constructed. When expressed in wild-type cells, each chimera cofractionated with 60S subunits in 7-to-42% sucrose gradients and yielded diffuse cytoplasmic GFP signals, thereby indicating that each of these fusions was integrated into 60S subunits (Fig. (Fig.7).7). While GFP-chimera reporters showed normal cytoplasmic distributions in Δrai1 or Δkap123 cells, double null Δkap123/Δrai1 strains exhibited a striking nuclear accumulation of the GFP signals from each 60S subunit reporter (Fig. (Fig.7).7). The ratio of nuclear to cytoplasmic 60S particles was quantified by isolating nuclei and comparing the abundance of 60S particles in the nuclear and cytoplasmic fractions by sucrose density gradient centrifugation under low Mg2+ conditions (Fig. (Fig.8).8). In addition, the presence of rpL3 in the nuclear and cytoplasmic fractions was assayed by Western blotting (Fig. (Fig.8).8). Taken together, these data suggest that the reporter proteins accumulated in the nucleus are assembled into pre-60S particles. Thus, Δkap123/Δrai1 cells exhibit a 60S, postassembly, nuclear export block that was not observed in cells harboring either mutation alone.
The data presented above suggest that the genetic interaction between Δkap123 and Δrai1 is manifested at the late stages of 60S assembly. Because the only known role for Kap123p is in nuclear import, we hypothesized that an essential factor required for driving the export process was not imported efficiently in Δkap123 cells and that, in the absence of Rai1p, this factor becomes limiting. To investigate this possibility, sf17 cells were transformed with multicopy plasmids encoding two essential proteins required late in the 60S subunit assembly and that export Nmd3p (15, 26) and rpL10p (20) and the karyopherins Kap121p and Nmd5p. Interestingly, overexpression of Nmd3p, but not rpL10, Kap121p, or Nmd5p, was sufficient to rescue both the sectoring phenotype (data not shown) and the FOA sensitivity of sf17 cells (Fig. (Fig.99).
To determine if the nuclear localization of Nmd3p was altered as a result of loss of Kap123p, we assayed Nmd3p nuclear import in Δkap123 cells. Nmd3p is known to have one NLS and two NES sequences (20, 26). Galactose-inducible Nmd3p-GFP chimeras, either containing or lacking an NES (Nmd3-NLS-GFP and Nmd3-NLS/NES-GFP), were expressed in kap123 and kap95 mutant strains (26). As shown in Fig. Fig.9,9, Nmd3-NLS-GFP was nuclear in wild-type cells, but mislocalized to the cytoplasm in Δkap123 cells. In contrast, the reporter remained nuclear in kap95 cells, even after temperature shifts that have previously been shown to mislocalize classical NLS reporters (37).
These results suggest that, in addition to the previously reported role in the import of several ribosomal proteins, Kap123p also imports Nmd3p, an essential late-acting assembly and export factor. The presence of a fraction of cells with some nuclear staining, along with cytoplasmic staining in the Δkap123 background, suggests that Nmd3p, like other essential Kap123p cargoes, can enter the nucleus by alternate means, albeit with reduced efficiency. Under normal conditions, the import defect associated with the loss of Kap123p alone does not appear to significantly compromise ribosome biogenesis. However, in the absence of Rai1p, the limited nuclear quantities of late-acting factors such as Nmd3p are not sufficient to support efficient export, leading to the accumulation of partly assembled 60S subunits in the nucleus and a growth defect that greatly exceeds that observed for either deletion alone.
Here, we used a synthetic fitness genetic screen to identify components that support Kap123p function and, when mutated, render cells dependent on the presence of Kap123p. Not surprisingly, this screen yielded several karyopherins: Kap121p, Sxm1p/Kap108p, and Nmd5p/Kap119p. Indeed, Kap121p has previously been shown to compensate for the loss of Kap123p, likely by importing several ribosomal proteins in its absence (46, 48). In addition, Sxm1p/Kap108p has previously been shown to interact physically with the ribosomal proteins rpL11A/B, rpL25, and rpL31A/B (42), and overproduction of this karyopherin suppresses a kap121ts mutant phenotype (41, 49), suggesting further complementarity between Kap123p, Kap121p, and Sxm1p/Kap108p. Moreover, the high degree of similarity between Sxm1p/Kap108p and Nmd5p/Kap119p (4) and the identification of NMD5 as genetically interacting with KAP123 suggest a similar and likely overlapping function for this karyopherin. Interestingly, the use of several karyopherins to import different ribosomal proteins by yeast cells seems to be conserved in metazoans. Numerous vertebrate karyopherins, including Kap β/importin β, Kap β2/transportin, RanBP5, RanBP7, and importin 11, have been implicated in ribosomal protein import (22). Furthermore, a metazoan Kap123p orthologue, importin β4, has recently been characterized (29). Thus, data from different systems suggest that ribosomal protein import and assembly makes use of a variety of import (and export) karyopherins and that their functional overlap may permit the loss of individual factors. However, it remains unclear to what extent these karyopherins functionally overlap under normal conditions and whether this multiple redundancy is exploited by cells to globally control the transport of classes of different molecules.
Beyond redundant import pathways, the genetic screen revealed an interaction between the ribosome assembly factor RAI1 and KAP123. In an attempt to understand the molecular bases for this interaction, we investigated how the loss of both proteins specifically affected the process of ribosome biogenesis. As shown previously, cells lacking Rai1p revealed a 60S assembly defect, accumulating 27S rRNA and halfmers in polysome profiles. Remarkably, the loss of Kap123p had little effect on ribosome assembly or export, as we detected no obvious rRNA accumulation or ribosomal subunit export defects in cells lacking Kap123p, but cells lacking both Rai1p and Kap123p displayed a more complex phenotype. These cells showed a moderate augmentation of the 27S rRNA defect, normalization of the 40S/60S ratio, an overall decrease in the number of ribosomes, and an accumulation of assembled pre-60S subunits in the nucleus. Furthermore, although other karyopherins import ribosomal proteins, the genetic interaction between Rai1p and Kap123p was specific: out of the five yeast Kaps tested, including all those known to import ribosomal proteins, only KAP123 was able to rescue sf17 cells and only kap123/rai1 mutants displayed pre-60S ribosomal subunit export defects. Together these data suggest that, in these cells, the 60S biogenesis program was attenuated at a late, postassembly, preexport step.
It is evident that cells lacking Rai1p are defective in 60S assembly, but why should the additional loss of Kap123p, whose only known function is in nuclear import, specifically cause an export defect? Consider ribosome assembly as a simple series of chemical reactions; the removal of products at each step contributes to the progression of the entire pathway. Alternatively, the failure to remove products at any step causes the accumulation of intermediates. Thus, considering that Rai1p interacts physically with Rat1p, a late-acting exonuclease, and that Rai1p was detected associated with assembled pre-60S subunits, we hypothesized that the protein(s) required at the late stages of biogenesis is not imported efficiently in kap123 strains and that, in combination with a mutation in RAI1, products downstream of Rai1p function were not efficiently processed to the next step, leading to the accumulation of (partially) assembled subunits. Thus, we speculate that the specific genetic interaction observed between KAP123 and RAI1 is due to a reduced efficiency of 60S subunit assembly, contributed by a lack Rai1p function, as well as an inability to import critical ribosomal assembly and export factors. Here we show that one such factor is Nmd3p. Overexpression of NMD3 rescued the slow-growth phenotype observed in sf17 (rai1-1/Δkap123), Δrai1/Δkap123, and Δrai1 cells (data not shown), and direct visualization of an Nmd3-GFP chimera demonstrated that efficient Nmd3p import into the nucleus requires Kap123p. While the mislocalization of Nmd3p is evident in kap123 cells, deletion of NMD3 is lethal; thus, it is likely that other factors can also import Nmd3p in the absence of Kap123p. Furthermore, it is also likely that inefficient import of other factors contributes to the genetic interaction observed here. One such candidate is rpL10, also a late-acting assembly and export factor for 60S subunits imported by Kap123p (20). Nevertheless, because NMD3, but not RPL10, expression is sufficient to suppress the growth defects associated with rai1/kap123 cells, it is apparent that Nmd3p mislocalization is a primary cause of the rai1/kap123 genetic interaction.
Surprisingly, among the mutants assayed, only rat1-1 cells mislocalized Rai1p-GFP from the nucleus to the cytoplasm. Considering the tight in vitro binding between Rat1p and Rai1p, we propose that the steady-state localization of Rai1p to the nucleus is a result of its interaction with Rat1p. Furthermore, it is interesting to speculate that Rai1p may be used by Rat1p to tether the assembling subunit, but that, in the absence of a functional Rat1p, Rai1p may exit the nucleus with the ribosomal subunit. It has previously been shown that loss of active Rat1p can be complemented by directing Xrn1p, a normally cytoplasmic exonuclease, to the nucleus (30). It is not yet known if quality control mechanisms exist to prevent incompletely assembled ribosomal subunits from exiting the nucleus, but it seems possible that under conditions where Rai1p becomes cytoplasmic the subsequent maturation of unprocessed rRNA could occur in the cytoplasm, under the direction of Xrn1p. It will be interesting to determine if the function of Xrn1p is also augmented by Rai1p and if the export defect observed here also results from an active quality control mechanism.
The synthetic fitness screen employed here revealed a complex genetic interaction between KAP123, a nuclear import factor, and RAI1, a ribosome biogenesis factor, which manifests itself in a ribosome subunit export defect. The data support a model where a cause of the defect is an inability to import sufficient quantities of the essential export factor Nmd3p to overcome the loss of Rai1p. It is particularly intriguing to speculate that the coordination of the late steps of 60S biogenesis and nuclear export involve a direct link between Rai1p and Nmd3p, perhaps during the loading of ribosomes with Nmd3p. However, this remains to be investigated. The findings presented here underscore the integration of ribosome assembly and nucleo-cytoplasmic exchange; however, a good understanding of the entanglement between these two pathways demands further identification and characterization of ribosome assembly factors and an understanding of their relationships with the nuclear import-export apparatus.
We thank Rosanna Baker, Adriana Antunz-de-Mayolo, Tatiana Iouk, Vladmir Titorenko, and Dwayne Weber for experimental assistance, Taras Makhnevych and Marcello Marelli for helpful discussions, and Mike Rout for sharing unpublished data, critical reading of the manuscript, and helpful discussions.
This work was supported by operating and salary support from Canadian Institutes for Health Research, Alberta Heritage Foundation for Medical Research, the Institute for Systems Biology, and Merck.