Characterization of the Auto-inhibitory Sequence within the N-terminal Domain of Importin α

An evolutionarily conserved auto-inhibitory within the N-terminal importin β binding (IBB) domain of importin regulates NLS-cargo binding to the NLS binding pocket on importin α . In this study, we have used site-directed mutagenesis coupled with in vitro binding assays and in vivo analyses to investigate the intramolecular interaction of the N-terminal IBB domain and the NLS binding pocket of Saccharomyces cerevisiae importin α , Srp1p. We find that mutations within the IBB domain that decrease the binding affinity of the auto-inhibitory sequence for the NLS binding pocket impact importin α function in vivo . In addition, the severity of the in vivo phenotype is directly correlated to the reduction of auto-inhibition measured in vitro suggesting that the in vivo phenotypes are directly related to the loss of auto-inhibitory function. We exploit a conditional auto-inhibitory mutant, srp1-55 , to study the in vivo functional overlap between the N-terminal IBB domain of importin α and other factors implicated in NLS-cargo release, Cse1p and Nup2p. We propose that the N-terminal IBB domain of importin α and Cse1p function together in NLS-cargo release while Nup2p contributes to cargo release/importin α recycling through a distinct mechanism. Cse1p and Nup2p in vivo . We propose that the auto-inhibitory N-terminal domain of importin α and Cse1p function together in NLS-cargo release, whereas Nup2p functions through a different mechanism in this essential process.


INTRODUCTION
All macromolecules that move into and out of the nucleus are transported through nuclear pore complexes, large proteinaceous channels that are embedded in the nuclear envelope (1,2).
Soluble factors are required to recognize, target, and transport most macromolecules through the nuclear pores (3,4). The best characterized nuclear transport process occurs via receptor recognition of classical nuclear localization signals (NLSs) 1 on protein cargoes targeted for nuclear import (3,5). These classical NLSs are typified by a single cluster of basic amino acids (monopartite) or two clusters of basic amino acids separated by a 10-12 amino acid linker (bipartite) (6,7).
Protein cargoes that contain classical NLSs are recognized in the cytoplasm by a heterodimeric receptor composed of importin/karyopherin α and importin/karyopherin β (3,8-a high affinity for NLS-cargo due to binding to importin β, and the nuclear form, which has a low affinity for NLS cargo. Importin α structural studies suggest that monopartite NLSs and the auto-inhibitory sequence bind to the NLS-binding pocket of importin α in a similar manner (22,24,25). Indeed, a conserved cluster of basic amino acids, that resembles a classical basic NLS, serves as the autoinhibitory sequence (yeast importin α residues 54 KRR 56 ) (22,23). In vitro analysis of the energetic landscape of NLS binding to importin α revealed details of the specific interactions required for NLS binding to the pocket of importin α (26). In particular, Hodel et al. defined the requirements for specific amino acid residues within an NLS that are critical for high affinity interaction with importin α (26). We hypothesize that the auto-inhibitory sequence depends on similar energetic molecular interactions with the NLS binding pocket. This would be consistent with its role in NLS-cargo release through a direct competition mechanism.
The present study further characterizes the auto-inhibitory sequence within the Nterminal domain of S. cerevisiae importin α. We have utilized site-directed mutagenesis, in vitro binding assays and in vivo analyses to investigate the intramolecular interaction between the Nterminal IBB domain and the NLS binding pocket of importin α. Through these analyses we demonstrate that the auto-inhibitory sequence binds to the NLS binding pocket through energetic interactions that are analogous to those for a monopartite NLS. We present data in support of our hypothesis that the auto-inhibitory function of the IBB domain is responsible for essential in vivo functions. Our experiments demonstrate that the severity of the in vivo phenotypes are directly correlated to the reduction of auto-inhibition measured in vitro, suggesting that the in vivo phenotypes are directly related to the loss of auto-inhibitory function. Furthermore, we present data in support of functional overlap between the N-terminal domain of importin α, 7

EXPERIMENTAL PROCEDURES
Strains, Plasmids and Chemicals-All chemicals were obtained from Sigma or USBiological unless otherwise noted. All DNA manipulations were performed according to standard methods (27), and all media was prepared by standard procedures (28). All yeast strains and plasmids used in this study are described in Table I. Expression and Purification of Recombinant Proteins-Assays were performed with purified recombinant S. cerevisiae proteins Srp1p (importin α) and Rsl1p (importin β). Fulllength His 6 -importin α (residues 1-542), His 6 -SV40 (SPKKKRKVEAS)-NLS-GFP and His 6 -Myc (PAAKRVKLD)-NLS-GFP were expressed in the E. coli strain BL21 (DE3) and purified by nickel affinity chromatography essentially as described (26,29). Importin β was expressed and purified as described previously (29).
Fluorescence Anisotropy Assay-Fluorescence anisotropy measurements were carried out using an ISS PC1 fluorimeter fitted with polarization filters. The dissociation constants for the binding of SV40 NLS-GFP and Myc NLS-GFP to importin α were measured essentially as described previously (26,29). Briefly, SV40-NLS-GFP or Myc-NLS-GFP was diluted in PBS to the desired concentration (~20 nM) in a total volume of 1 ml in a 1-cm quartz cuvette. Changes 8 in the anisotropy of the GFP fluorophore were monitored as increasing amounts of wild-type or mutant importin α proteins were added to the assay volume. Changes in anisotropy were used to calculate the fraction of the GFP fluorophore bound, yielding a binding isotherm for the reaction.
The binding isotherm was then fit through nonlinear regression to a simple binding equation to obtain dissociation constants. All K d values are calculated as detailed at http://www.biochem.emory.edu/Hodel/Research/BindingCurves/ fitting_curves.htm. Binding energies were calculated using ∆G = RT ln K d where RT = -0.59 kcal/mol.
In Vivo Functional Analysis-The in vivo function of each of the importin α variants was tested using a plasmid shuffle technique (30). Plasmids encoding each of the importin α mutant proteins were individually transformed into SRP1 deletion cells (ACY324) containing the URA3 SRP1 maintenance plasmid, pAC876 (23). Single transformants were grown in liquid culture to saturation, serially diluted (1:10) and spotted on minimal media plates lacking leucine as a control or on fluoroorotic acid (5-FOA) plates. The drug 5-FOA eliminates the URA3 plasmidencoded wild-type importin α (pAC876) (30). Plates were incubated at the indicated temperatures for 3-5 days.
Microscopy-Direct fluorescence microscopy was used to localize GFP fusion proteins in live cells. For all experiments, cells were stained with DAPI (1 µg/ml) to visualize the DNA and confirm the location of the nucleus. The localization of the fusion proteins was monitored by directly viewing the GFP signal in living cells through a GFP-optimized filter (Chroma Technology) using an Olympus BX60 epifluorescence microscope equipped with a Photometrics Quantix digital camera.
Integration into the Yeast Genome-The importin α-GFP mutants were integrated at the endogenous importin α locus of wild-type (ACY192) cells using a standard integration strategy (23). The integration created a duplication at the endogenous importin α locus such that both endogenous importin α and importin α-GFP were each expressed from SRP1 promoters. TUB1-GFP was integrated at the URA locus as described previously (33).
In contrast to the importin α-GFP integrations, the A55-importin α mutant replaced the wild-type copy of importin α. To integrate A55-importin α, the A55 mutation was subcloned into the SRP1 open reading frame cloned in the LEU2 integrating plasmid, pRS305 (34), to create srp1-55-pRS305 (pAC1128). A55-importin α was then integrated at the endogenous SRP1 locus by linearization of srp1-55-pRS305 and transformation into the wild-type diploid ACY247. Transformants that grew on plates lacking leucine were selected and the presence of the A55-importin α mutation was confirmed by PCR and sequencing. The heterozygous diploid was subsequently sporulated and tetrads were dissected to generate the haploid A55-importin α strain, srp1-55 (ACY642). This integration strategy is designed to make A55-importin α the only copy of importin α expressed in the haploid strain.
Yeast Genetic Analyses-Synthetic interactions between A55-importin α (srp1-55) and other nuclear transport factors were tested by creating double and triple mutant strains. These strains were made by crossing the various haploid mutant strains [srp1-55, cse1-1 (35), ∆nup2 (Research Genetics)]. The srp1-55 strain was covered by a SRP1 URA3 plasmid (pAC876) and mated to each of the mutants to be tested. The resultant heterozygous diploids were sporulated and dissected to generate the appropriate double and triple mutant haploid strains. For suppression analysis, high copy plasmids (2µ) expressing various nuclear transport factors were transformed into srp1-55 (ACY642) covered by a SRP1 URA3 plasmid (pAC876). Genetic interactions (synthetic growth defects/lethality and suppression) were assessed by growing single colonies in liquid culture to saturation, serially diluting (1:10) and spotting on minimal media plates as a control or on fluoroorotic acid (5-FOA) plates. Plates were incubated at the indicated temperatures for 3-6 days.
FACS Analysis-Cells were prepared for FACS analysis by staining with propidium iodide (36). Briefly, cells were ethanol fixed at 4°C, washed and resuspended in 1ml of 50 mM sodium citrate, pH 7.0. Cells were then treated with 0.08 mg/ml Rnase A for 1 h at 50°C, followed by 0.25 mg/ml proteinase K for 1 h at 50°C, before incubation in 8 µg/ml propidium iodide. Each sample was analyzed with a FACS Caliber cytometer.

Generation of Mutant Importin α Proteins-Previous analyses of the N-terminal domain
of importin α protein revealed a conserved cluster of basic amino acids that serves as an autoinhibitory sequence (mouse importin α residues 49 KRR 51 corresponding to yeast importin α residues 54 KRR 56 ) (22,23). Structural studies suggested that the auto-inhibitory sequence resembles an NLS and binds to the NLS-binding pocket in a manner similar to the binding of a monopartite NLS (22,24,25). In order to test this prediction, we individually substituted each of the basic amino acids to alanine in the 54 KRR 56 cluster of yeast importin α (Srp1p) to generate importin α mutants referred to as A54 (K54A), A55 (R55A) and A56 (R56A) (Fig. 1).
Auto-inhibitory Function of Importin α Mutants-The A54 and A55 mutations are in residues that correspond to NLS residues that are critical for a high affinity interaction with the NLS binding pocket (26). Thus, we would predict that the auto-inhibitory domains of the A54 and A55 proteins should have weaker binding to the NLS binding pocket and should therefore have a decrease in auto-inhibitory function when compared to wild-type importin α. To examine both NLS binding and auto-inhibitory function for each importin α variant, we used a quantitative, fluorescence anisotropy, solution binding assay (23,26,29). Full-length wild-type importin α binds weakly to a classical SV40-NLS due to the N-terminal auto-inhibitory IBB domain. However, when the same experiment is carried out in the presence of a stoichiometric amount of importin β, the N-terminal auto-inhibition of full-length importin α is relieved and much tighter binding (~ 30-fold) to the SV40-NLS is observed. In contrast, our previous work demonstrates that the A3-importin α mutant ( 54 KRR 56 →AAA) has decreased auto-inhibitory function without impacting importin β binding (23). Due to this decreased auto-inhibition, the A3 protein binds to SV40-NLS-GFP ~7-fold more tightly than full-length wild-type importin α (23). This assay actually examines three aspects of importin α function: binding to NLS cargo, auto-inhibitory function (measured by the ability of full-length importin α to bind NLS cargo in the absence of importin β), and binding to importin β (based on the relief of auto-inhibition).
To examine NLS binding and the auto-inhibitory function for each importin α protein that contains a single amino acid change, we performed the fluorescence anisotropy assay using a monopartite SV40-NLS cargo. The assay was carried out with each importin α protein in the presence and absence of importin β (Fig. 2). Typical curves for binding of SV40-NLS-GFP to the mutant importin α proteins in the absence of importin β are shown in Fig. 2A. As described in Experimental Procedures, these curves are used to calculate K d values for the interaction between NLS-cargo and importin α (see Table II). The K d values can be used to determine the change in free energy (∆G) to compare the wild-type and mutant proteins.
Thus, to assess the impact of each amino acid change within the auto-inhibitory sequence, we compare the K d values and calculate the change in free energy, ∆G, for the binding of each importin α protein to NLS-cargo. As previously demonstrated, wild-type importin α binds to SV40-NLS-GFP weakly (K d ~ 500 nM), but the affinity increases ~30-fold (K d ~ 18 nM) in the presence of importin β (23). As a control, A3-importin α, which has decreased autoinhibitory function, binds to SV40-NLS-GFP more tightly (K d ~ 73 nM) than full-length wildtype importin α (K d ~ 500 nM). A54-importin α binds to SV40-NLS-GFP with a similar affinity (K d ~83 nM) to A3-importin α suggesting that K54 is the most critical residue in the 54 KRR 56 auto-inhibitory sequence. A55-importin α binds to SV40-NLS-GFP more tightly (K d ~240 nM) than wild-type importin α, an ~2-fold increase in affinity. A56-importin α binds to SV40-NLS-GFP with a similar affinity (K d ~ 1300 nM) to wild-type importin α. To compare each of the importin α proteins, the change in free energy for the binding of each importin α protein to NLS cargo in the absence ( ) and presence ( ) of importin β  (Fig. 2B). In the presence of importin β, the affinity of each of the mutant importin α proteins is similar to wild-type importin α suggesting that each mutant retains normal binding to importin β (Table II). These binding results also demonstrate that each of the mutant importin α/β complexes binds NLS-cargo with approximately the same affinity as the wild-type importin α/β complex.

Correlation of Importin α Function with NLS Binding-Previous structural studies
suggested that the auto-inhibitory sequence, KRR, resembles an NLS and binds to the NLSbinding site in a similar way to a monopartite NLS (sites P2-P4) (22,24). Alignment of NLS sequences with the auto-inhibitory sequence shows that residue K54 of importin α binds to importin α at the same position (P2) as the essential lysine of a monopartite NLS (Fig. 2C) (26).
Mutation of K54 to A54 significantly decreases the auto-inhibitory function of importin α ( Fig.   2A). The decrease in auto-inhibition can be expressed as a change in free energy (∆∆G) for each importin α mutant when compared to wild-type protein (Fig. 2C). As shown in Fig. 2C, residues critical for auto-inhibitory function correlate with residues that are critical for an NLS binding to the NLS binding pocket of importin α, where mutation of the P2 binding residue of the NLS to alanine causes the largest decrease in binding affinity (26). The NLS residue that binds importin α in position 3 (P3) has an intermediate contribution and the position 4 (P4) residue has a fairly weak contribution to the NLS binding energy (Fig. 2C). The in vitro auto-inhibitory behavior of A55-importin α and A56-importin α correlates with the energetic contributions of the corresponding residues within the SV40 and Myc monopartite NLSs (Fig. 2C). Thus, the auto-inhibition of each of the importin α mutants correlates with the energetic contribution of each residue to a functional NLS.
In Vivo Function of Importin α Mutants-Our in vitro experiments demonstrate that variants of importin α with point mutations within the 54 KRR 56 auto-inhibitory sequence exhibit a range of auto-inhibitory functions (Table II). To test the effect of different levels of autoinhibition on importin α function in vivo, each mutant was transformed into yeast cells deleted for the endogenous importin α gene (SRP1) and plasmid shuffle (see Experimental Procedures) was used to replace a plasmid borne wild-type copy of SRP1 (Fig. 3A). This results in ∆SRP1 cells that express each of the mutant importin α proteins from their own promoter on a low copy centromeric plasmid as the only copy of importin α. Controls demonstrate that a wild-type importin α plasmid can complement ∆SRP1 cells while neither a vector alone nor the autoinhibitory defective importin α (A3) can functionally replace SRP1 (Fig. 3A, compare the control and 5-FOA plates). Results shown in Fig. 3A indicate that cells expressing A54-importin α have a pronounced cold-sensitive phenotype (no growth at 16°C) and grow extremely slowly even at 30°C. Cells expressing A55-importin α grow more slowly than wild-type at 30°C and are coldsensitive at 16°C. Cells expressing A56-importin α grow similar to cells expressing wild-type importin α at all temperatures.
To confirm that each mutant protein is expressed at a similar level to wild-type importin α, we analyzed their expression using a C-terminal triple myc tag. Immunoblotting of the myctagged importin α proteins demonstrates that each of the mutant proteins is expressed at approximately the same level as wild-type importin α (Fig. 3B, compare lanes 3-6 with lane 2).
This suggests that none of the mutations in the N-terminal IBB domain of importin α significantly affect the level of the protein within the cell, but rather that the A54 and A55 mutations perturb the function of the importin α proteins.
Localization of the Mutant Importin α Proteins-We have previously found that mutations that decrease the auto-inhibitory function of importin α result in its accumulation within the nucleus (23). This accumulation is presumably due to the persistence of an NLScargo/importin α complex within the nucleus. In order to further examine the importin α variants, we analyzed the localization of each protein using C-terminal GFP tagged fusion proteins. These importin α-GFP fusion proteins were expressed from the endogenous importin α promoter. We integrated each importin α-GFP fusion protein at the endogenous importin α locus and visualized them as described previously (23). Wild-type importin α localizes to the nuclear rim and the cytoplasm in wild-type cells when visualized in this manner (Fig. 3C, panel E). However, A3-importin α-GFP accumulates within the nucleus (Fig. 3C, panel A). The localization of A54-importin α-GFP (Fig. 3C, panel B) is similar to that of A3-importin α-GFP.
A55-importin α-GFP also localizes to the nucleus but shows some cytoplasmic signal, suggesting that its localization is intermediate between A54-importin α and wild-type importin α Analysis of R54-importin α-Even conservative mutation of the P2 lysine to arginine in SV40 NLS results in an ~ 94-fold decrease in binding to the NLS binding pocket (26). To further test the hypothesis that the auto-inhibitory sequence interacts with importin α in the same manner as an NLS, we mutated K54 to an arginine residue to create R54-importin α (Fig. 1) and assessed the impact of this conservative mutation. R54-importin α binds to SV40-NLS-GFP with ~5.5-fold higher affinity than wild-type importin α (Table II) as assessed by fluorescence anisotropy. This demonstrates that R54-importin α is defective in auto-inhibition and provides additional evidence that a monopartite NLS and the auto-inhibitory sequence bind to the NLS binding pocket of importin α with similar energetics. In the presence of importin β the binding of SV40-NLS-GFP to R54-importin α is similar to wild-type importin α (Table II).
To test the function of the R54-importin α protein in vivo, we used the plasmid shuffle technique and the importin α-GFP localization assay (see Experimental Procedures). R54importin α was transformed into yeast cells deleted for the endogenous importin α gene (SRP1) and plasmid shuffle was used to replace the plasmid-borne wild-type copy of SRP1 (Fig. 4A).
Results show that cells expressing R54-importin α as the only copy of importin α grow slowly at 30°C and are not viable at 16°C. The R54-importin α protein was expressed at similar levels to wild-type importin α (data not shown). In vivo localization analysis shows that the R54-importin α-GFP protein accumulates within the nucleus similar to A54-importin α-GFP (Fig. 4B).
In Trans Localization of the Auto-inhibitory Sequence-The IBB domain of importin α, which contains the auto-inhibitory sequence, binds weakly to the NLS binding pocket in trans (K d~ 14 µm) (23). This suggests that the auto-inhibitory sequence should not bind to the NLS binding pocket with sufficient affinity to act as an NLS and direct a heterologous protein into the nucleus when expressed in vivo. To test whether the auto-inhibitory sequence can act like an NLS to direct a heterologous protein to the nucleus, we fused importin α residues 49-60 in frame with GFP-GFP to create IBB 49-60 -GFP-GFP. The localization of this fusion protein was compared both to GFP-GFP, which lacks any NLS, and to SV40-NLS-GFP-GFP, a positive control which contains a canonical monopartite NLS. As shown in Fig. 5, the control GFP-GFP protein is diffusely localized throughout the cell. In contrast, SV40-GFP-GFP accumulates in the nucleus. The IBB 49-60 -GFP-GFP is localized throughout the cell similar to GFP-GFP. This demonstrates that the auto-inhibitory sequence cannot efficiently act as an NLS and direct a protein to the nucleus when expressed in trans.
Genetic Analysis of NLS Release Factors-Through analysis of the auto-inhibitory defective mutant of importin α, A3, we suggested a model where the auto-inhibitory activity of importin α is required for NLS-cargo release and the subsequent Cse1p-dependent recycling of importin α to the cytoplasm (23). Cse1p and Nup2p have also been reported to affect NLS-cargo release from importin α (19,21). It has been suggested that these factors may cooperate or act sequentially to facilitate NLS-cargo release and may therefore functionally overlap. As shown in Fig. 3A, the A55 mutation in importin α causes a conditional growth phenotype which correlates with its defect in auto-inhibitory function. This cold-sensitive phenotype can be exploited for genetic analysis. As a genetic test for functional overlap between the auto-inhibitory function of the IBB domain of importin α, Cse1p, and Nup2p, we tested for any exacerbated growth defects (synthetic lethality) in double or triple mutant cells.
The A55-importin α allele (srp1-55) was integrated into the genome to avoid copy number effects (Experimental Procedures). Since CSE1 is essential, we used a well characterized cold-sensitive allele, cse1-1 (35). NUP2 is not essential (37) and therefore we utilized a complete deletion of the open reading frame, ∆nup2. If Cse1p and Nup2p are involved in NLS-cargo release in the nucleus, cells with combined mutations in the importin α autoinhibitory function, CSE1 and/or NUP2 might be more growth compromised than any of the single mutants. We therefore generated each of the double mutants (srp1-55 cse1-1, srp1-55 ∆nup2, and cse1-1 ∆nup2) and the triple mutant (srp1-55 cse1-1 ∆nup2) maintained by a plasmid borne wild-type importin α (SRP1) as described in Experimental Procedures. As previously reported (21,35,38) and shown in Fig. 6A, cse1-1 cells are cold-sensitive at 16°C and ∆nup2 cells do not have any detectable growth defect. Results shown in Fig. 6A demonstrate that the srp1-55 cse1-1 double mutant is inviable (compare the control and 5-FOA plates). This synthetic lethal phenotype was observed at all temperatures tested. In addition, the srp1-55 ∆nup2 and cse1-1 ∆nup2 double mutants grow more slowly than either single mutant. As shown in Fig. 6A, these phenotypes are more pronounced at cold temperatures.
As a second genetic test for functional overlap, we assessed high copy suppression of the srp1-55 cold-sensitive phenotype. Importin β, Cse1p and Nup2p were expressed from a high copy plasmid and their ability to suppress the cold-sensitive phenotype of srp1-55 was examined as described in Experimental Procedures. Controls demonstrate that a wild-type importin α plasmid can complement the cold-sensitive phenotype of srp1-55 cells, while a vector alone cannot (Fig. 6B, compare the control and 5-FOA plates). Results shown in Fig. 6B demonstrate that importin β cannot complement the cold-sensitive phenotype of the srp1-55 cells. This is consistent with our in vitro analyses, Fig. 2B, where we show that the mutant importin α proteins bind to importin β with the same affinity as wild-type importin α. Nup2p also does not suppress the srp1-55 cold-sensitive phenotype. In contrast, overexpression of Cse1p partially suppresses the cold-sensitive phenotype of srp1-55 cells.

srp1-55 Cells Accumulate in G 2 /M at 18 o C-Importin α is required for the execution of mitosis as cells with conditional mutations of importin α arrest with a G 2 /M phenotype (39).
Therefore, to further characterize the srp1-55 cells, we assessed the morphology of cells grown at either 30°C or 18°C and compared them to wild-type cells grown under the same conditions. Microscopic analysis revealed that the srp1-55 cells have similar morphology to wild-type cells when grown at 30°C. Interestingly, at 18°C the srp1-55 cells were larger and there was a greater proportion of large budded cells within the population as compared to wild-type cells (for example see Fig.7A, compare panels G and H). To determine if these large budded cells were arrested uniformly within the G 2 /M phase of the cell cycle, we examined the microtubules by integrating TUB1-GFP into wild-type and srp1-55 cells (Experimental Procedures). The TUB1-GFP integrated wild-type and srp1-55 cells were grown to log phase at both 30°C and 18°C and compared by microscopic analysis (Fig. 7A). As shown in Fig. 7A, panel D, there is a variation in spindle length in the large budded srp1-55 cells, therefore these cells are not uniformly arrested at the same point within G 2 /M of the cell cycle. This was also confirmed by staining the DNA with DAPI (data not shown). These data suggest that the auto-inhibitory defective cells are not arrested, but rather pass through G 2 /M phase more slowly than wild-type cells.
The microscopic analysis of the srp1-55 cells suggests that they spend more time in the G 2 /M phase of the cell cycle, thus we predict that a higher percentage of cells should have replicated (2N) DNA than wild-type cells. To examine the DNA content of the srp1-55 cells, we performed FACS analysis of log phase cultures grown at both 30°C and 18°C and compared it to wild-type cells grown under the same conditions. Cells were stained with propidium iodide, and the DNA content was analyzed by flow cytometry. The FACS profile of wild-type cells grown at either 30°C or 18°C are indistinguishable in the distribution of cells with 1N and 2N DNA content (data not shown). Cells expressing the A55-importin α protein have a similar profile to wild-type cells at 30°C (data not shown). As shown in Fig. 7B, the 2N DNA peak of srp1-55 cells grown at 18°C is markedly broader than that of srp1-55 cells grown at 30°C, which is consistent with an increased proportion of cells in G 2 /M, and hence slower progression through this phase of the cell cycle.
This G 2 /M phenotype of srp1-55 cells provides another assay to characterize the genetic interaction between the auto-inhibitory function of importin α and Cse1p. As shown in Fig. 6B, expression of Cse1p from a high copy plasmid suppresses the cold-sensitive phenotype of srp1-55 cells. Therefore, we tested whether Cse1p expressed from a high copy plasmid could also suppress the G 2 /M phenotype of srp1-55 cells. Controls demonstrate that a wild-type importin α plasmid can suppress the srp1-55 G 2 /M phenotype while a vector alone cannot (Fig. 7B).
Overexpression of Cse1p also suppresses the G 2 /M cell cycle phenotype of the srp1-55 cells. The ability of the IBB domain to regulate NLS-cargo binding to the NLS binding pocket is presumably dependent on the energy the IBB gains from the intramolecular interaction versus an intermolecular interaction. Indeed, our data support a model where the energy for the binding of the auto-inhibitory sequence to the NLS binding pocket of importin α is obtained from the in cis intramolecular interaction. We have previously shown that in trans the IBB binds with low affinity (µM) to the NLS binding pocket of importin α (23). Consistent with this previous result, we demonstrate that the auto-inhibitory sequence is not able to target a heterologous protein into the nucleus of yeast cells when expressed in trans, suggesting that the binding affinity for importin α is not sufficient for nuclear localization. Although the auto-inhibitory sequence can bind to the NLS binding pocket like an NLS, it would not be energetically favorable for it to interact with importin α with an affinity comparable to a functional NLS. However, the autoinhibitory sequence can compete with an NLS for binding to the NLS binding pocket due to the physical tethering of the IBB to the NLS binding pocket domain of importin α, which significantly increases its local concentration. This requirement for an in cis interaction between the IBB domain and the NLS binding pocket of importin α also presumably prevents the Nterminal domains of adjacent importin α proteins from interacting with NLS-cargo pockets intermolecularly and forming dimers.
Although RanGTP is the major determinant of import complex dissociation (16, 17,19), the N-terminal auto-inhibitory domain of importin α, Cse1p and Nup2p also play a role in the release of NLS-cargo into the nucleus (19,(21)(22)(23)25,29). The trimeric import complex is disassembled by the RanGTP-triggered dissociation of importin β to release a dimeric NLScargo/importin α complex into the nucleus. The importin α auto-inhibitory function is essential to efficiently dissociate this dimeric intermediate (23). In this study we use genetic analyses of the auto-inhibitory domain of importin α, CSE1 and NUP2 to provide data in support of a model where these factors have overlapping functions in vivo that facilitate release of NLS-cargo from importin α. This in vivo data is consistent with a recent study by Gilchrist et al. where they demonstrated that each of these interactions is able to increase the dissociation rate of the dimeric NLS-cargo/importin α intermediate in vitro (19).
Our genetic data suggest that Cse1p and Nup2p have distinct functions in NLS-cargo release. This conclusion is supported by two lines of genetic evidence. First, cse1-1 and srp1-55 are synthetically lethal, but ∆nup2 and srp1-55 show only a modest synthetic interaction.
Second, the cold-sensitive phenotype of srp1-55 cells is suppressed by Cse1p but not by Nup2p.
Independent biochemical experiments support the idea that Cse1p and Nup2p act through distinct mechanisms in NLS-cargo release (Gilchrist and Rexach, personal communication). As shown here and suggested previously (40), Cse1p and Nup2p demonstrate only minor genetic interactions. Thus, although the evidence suggests that both Cse1p and Nup2p participate in NLS-cargo release and/or importin α recycling, these proteins appear to function through distinct mechanisms.
A previous study also demonstrated a genetic interaction between importin α and Nup2p (41). This study utilized the srp1-31 allele of importin α (42). It should be noted that the srp1-31 allele has not been functionally characterized and therefore it is not known what step within the importin α transport cycle is affected. In contrast, importin β binding, NLS binding and auto-inhibitory function have all been analyzed for the srp1-55 mutant. Analysis of the genetic interactions of srp1-55 can therefore be interpreted in terms of the auto-inhibitory function of importin α. In srp1-31, serine residue 116 is substituted with a phenylalanine residue (42). This residue is outside of the NLS binding pocket and is not within the IBB domain of importin α (24,42). In the future it may be interesting to determine whether the srp1-31 mutant is specifically defective in a particular step in the importin α transport cycle so that it is possible to more definitively interpret studies with srp1-31 cells.
The ability of Cse1p to suppress the cold-sensitive phenotypes of srp1-55 cells suggests an intimate relationship between NLS-cargo dissociation and recycling of importin α to the cytoplasm. There is also evidence that NLS-cargo and Cse1p cannot bind to importin α simultaneously (13,14). This presumably prevents the recycling of importin α that is still bound to NLS-cargo in the nucleus and thus futile cycles of nuclear transport. It is not known how importin α and Cse1p interact, although we have shown that an auto-inhibitory defective mutant of importin α still interacts with Cse1p (23). This suggests that high copy suppression of the srp1-55 cold-sensitive phenotype by CSE1 is not due to a change in the binding affinity of the two proteins, but rather an overlap of function between the two proteins. Future studies of the importin α/Cse1p interaction should allow us to dissect the mechanism of NLS-cargo release and importin α recycling.
Mutant srp1-55 cells accumulate in the G 2 /M phase of the cell cycle. This observation is consistent with a previous report where the conditional srp1-31 mutation causes mitotic cell cycle defects (39). Loeb et al. suggested that the importin α dependent transport of cell cycle regulators into the nucleus is critical for cell cycle progression. Indeed, the slow growth phenotype of srp1-55 cells could be due to the inefficient release of NLS-cargoes within the nucleus, in particular, specific cargoes required for mitosis. For example, a recent study identified a critical cargo, TPX2, of the importin α/β complex whose release from importin α is essential for mitotic progression in Xenopus (15). This highlights the importance of understanding how cargoes are efficiently dissociated from importin α within the nucleus to mediate their cellular function. Future studies of the A55-importin α protein may allow us to identify these mitotic cargoes.