Extracellular signal-dependent nuclear import of Stat1 is mediated by nuclear pore-targeting complex formation with NPI-1, but not Rch1
1997; Springer Nature; Volume: 16; Issue: 23 Linguagem: Inglês
10.1093/emboj/16.23.7067
ISSN1460-2075
Autores Tópico(s)interferon and immune responses
ResumoArticle1 December 1997free access Extracellular signal-dependent nuclear import of Stat1 is mediated by nuclear pore-targeting complex formation with NPI-1, but not Rch1 Toshihiro Sekimoto Toshihiro Sekimoto Department of Anatomy and Cell Biology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Naoko Imamoto Naoko Imamoto Department of Anatomy and Cell Biology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Koichi Nakajima Koichi Nakajima Department of Molecular Oncology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Toshio Hirano Toshio Hirano Department of Molecular Oncology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Yoshihiro Yoneda Corresponding Author Yoshihiro Yoneda Department of Anatomy and Cell Biology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Toshihiro Sekimoto Toshihiro Sekimoto Department of Anatomy and Cell Biology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Naoko Imamoto Naoko Imamoto Department of Anatomy and Cell Biology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Koichi Nakajima Koichi Nakajima Department of Molecular Oncology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Toshio Hirano Toshio Hirano Department of Molecular Oncology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Yoshihiro Yoneda Corresponding Author Yoshihiro Yoneda Department of Anatomy and Cell Biology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan Search for more papers by this author Author Information Toshihiro Sekimoto1, Naoko Imamoto1, Koichi Nakajima2, Toshio Hirano2 and Yoshihiro Yoneda 1 1Department of Anatomy and Cell Biology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan 2Department of Molecular Oncology, Biomedical Research Center, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka, 565 Japan The EMBO Journal (1997)16:7067-7077https://doi.org/10.1093/emboj/16.23.7067 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info In response to interferon-γ (IFN-γ), Stat1 is tyrosine phosphorylated and translocates to the nucleus where it activates transcription. In this study, we identified factors which mediate the nuclear import of Stat1. Tyrosine-phosphorylated Stat1 associated with the β subunit (a 97 kDa component) of the nuclear pore-targeting complex via the NPI-1 family, but not the Rch1 family, of α subunit (a 58 kDa component) as a result of IFN-γ stimulation. Antibodies against NPI-1 or β subunit consistently inhibited the IFN-γ-dependent nuclear import of Stat1 in living cells, although antibodies reactive to Rch1 had no effect. Solution binding assays with deletion mutants of NPI-1 showed that the Stat1-binding domain of NPI-1 was located in the carboxy-terminal region, which is clearly distinct from the SV40 large T antigen nuclear localization signal (NLS)-binding region. These results indicate that the extracellular signal-dependent nuclear transport of Stat1 is mediated by NPI-1, but not Rch1, in conjunction with β subunit, and that these factors participate in, not only constitutive, but also the conditional nuclear import of proteins. Introduction Studies of cytokines have revealed essential components of the cytokine signaling pathway. A family of latent cytoplasmic transcription factors termed Stats (proteins acting as signal transducers and activators of transcription) are activated by a variety of cytokines and growth factors. In response to various cytokines, the Stat family proteins are transported to the nucleus where they activate their specific gene transcription, and thus Stat family proteins play critical roles in the signal transduction pathways for a variety of cytokine receptors (Darnell et al., 1994; Ihle and Kerr, 1995; Fukada et al., 1996; Nakajima et al., 1996). Each cytokine activates a different Stat protein, for example interferon-γ (IFN-γ) activates Stat1 (Shuai et al., 1992, 1993a), while interleukin-6 (IL-6) activates Stat3 and Stat1 (Akira et al., 1994; Fujitani et al., 1994; Zhong et al., 1994; Nakajima et al., 1995). After treatment of cells with IFN-γ, a specific tyrosine residue (Tyr701) of Stat1 is phosphorylated by receptor-associated tyrosine kinases, after which the activated Stat1 is translocated into the nucleus to activate target genes directly by binding to specific promoter sequences (Darnell et al., 1994; Ihle and Kerr, 1995). In contrast to the detailed understanding of the activation step and the binding of Stat1 to DNA, very little is known about the translocation step from the cytoplasm to the nucleus in this signaling pathway. Nuclear import of karyophilic proteins into the nucleus has been shown to be directed by short amino acid sequences, termed nuclear localization signals (NLSs). The selective protein transport process into the nucleus can essentially be divided into two steps; (i) NLS-dependent binding to the cytoplasmic face of the nuclear pore, followed by (ii) an energy-dependent translocation through the nuclear pore complex (NPC). In previous studies, we found that the first step in the nuclear import of SV40 T antigen NLS-conjugated bovine serum albumin (T-NLS–BSA) occurs via the formation of a functional complex (termed the nuclear pore-targeting complex; PTAC) in the cytoplasm, targeting the nuclear pores (Imamoto et al., 1995a). It was found that in subsequent steps, a small G protein, Ran (Melchior et al., 1993; Moore and Blobel, 1993) and the Ran-interacting factor p10/NTF2 (Moore and Blobel, 1994; Paschal and Gerace, 1995) are required for the translocation of the NLS-containing protein into the nucleus through NPCs. The PTAC is a stable complex, composed of a karyophilic protein and two essential components, which we have identified from mouse Ehrlich ascites tumor cells (Imamoto et al., 1995b,c). One of these, a 58 kDa component of PTAC (PTAC58), functions as an NLS receptor. A variety of homologous proteins were identified almost simultaneously from various organisms, and individually named as importin α, Rch1, hSRP1α, karyopherin α and NPI-1/hSRP1 in mammals and SRP1 in yeast (Yano et al., 1992; Cortes et al., 1994; Cuomo et al., 1994; Görlich et al., 1994; Moroianu et al., 1995; O'Neill and Palese, 1995; Weis et al., 1995). These proteins consist of eight repeats known as the armadillo structure, which are rich in hydrophobic amino acids. From the cDNA sequences and the deduced amino acid sequences of these proteins, human homologs can be classified into two major groups, Rch1/hSRP1α and NPI-1/hSRP1 (∼50% amino acid identity with each other). The 58 kDa component of PTAC which we identified previously appears to correspond to the mouse homolog of human Rch1/hSRP1α (Imamoto et al., 1995b). In this report, the term 'α subunit' will be used collectively for all these molecules. The other, 97 kDa, component of the PTAC (PTAC97) functions to target the PTAC to the NPC by binding to PTAC58 bound to a karyophile (Imamoto et al., 1995c). Its homologs from various species have been isolated and are referred to as importin β, karyopherin β and p97 in mammals and Kap95p in yeast (Chi et al., 1995; Enenkel et al., 1995; Görlich et al., 1995; Radu et al., 1995). Since these proteins have a very high amino acid identity with each other in mammals (>95%), we will refer to them as the 'β subunit' in this study. Recently, a novel receptor-mediated nuclear import pathway was demonstrated by use of a novel transport signal in the hnRNP A1, termed M9. The M9-directed nuclear import was found to be mediated by a specific interacting protein, named transportin, which is distantly related to the β subunit (Pollard et al., 1996). Apart from this finding, in yeast, three additional proteins, homologous to Kap95p, were identified from the complete yeast database search. One of these, Kap104p, has been shown to act in the nuclear import of a specific class of proteins (Aitchison et al., 1996). The α and β subunits and Ran were shown to be essential factors in the nuclear import of basic-type NLS-containing proteins (Görlich et al., 1994; Imamoto et al., 1995b,c), which are constitutively transported to the nucleus immediately after their synthesis in the cytoplasm. On the other hand, extracellular signal-dependent nuclear protein import pathways have also been studied recently. It was shown recently that tyrosine phosphorylation of Stat1 is essential for its active nuclear import, and Ran and its GTPase activity are required for the IFN-γ-dependent nuclear translocation of Stat1 (Sekimoto et al., 1996). Furthermore, Ran was found to be required for signal-dependent nuclear import of glucocorticoid receptor (Carey et al., 1996). However, the mechanism and factors that participate in the first step of the signal-dependent nuclear transport of proteins which pre-exist in the cytoplasm remain unknown. This study reports an attempt to identify the factors which mediate the targeting of Stat1 from cytoplasm to nuclear pores after cytokine stimulation. An in vitro solution binding assay showed that tyrosine-phosphorylated Stat1 forms a stable complex with NPI-1, but not with PTAC58 (Rch1), and the β subunit of the nuclear pore-targeting complex in response to IFN-γ. Furthermore, microinjection of anti-NPI-1, but not anti-PTAC58, antibodies into the cytoplasm inhibited the nuclear import of Stat1. These results indicate that although SV40 T antigen NLS-containing substrate is transported by both Rch1 and NPI-1, the IFN-γ-dependent nuclear import of Stat1 is mediated by NPI-1, but not by Rch1, in conjunction with β subunit. An NPI-1 deletion experiment shows that the Stat1-binding domain of NPI-1 is distinct from the conventional NLS-binding domain, which, in turn, indicates that NPI-1 binds to two different karyophiles at different domains. The data collected herein show, for the first time, that the extracellular signal-dependent nuclear import of proteins requires NPI-1 and the β subunit for targeting the nuclear pores. Results Tyrosine-phosphorylated Stat1 associates with β subunit In order to determine whether the factors required for nuclear import of SV40 T antigen NLS-containing substrate also participate in the IFN-γ-dependent nuclear transport of Stat1, we examined the interaction between Stat1 and PTAC58 or β subunit. PTAC58 and β subunit were prepared as GST fusion proteins. These GST fusions show transport activities in the in vitro semi-intact cell assay using SV40 T antigen NLS–BSA conjugate (T-NLS–BSA) as the substrate, and the activities are comparable with proteins which are devoid of the GST portion (Imamoto et al., 1995c). HeLa whole cell extracts treated with or without IFN-γ were incubated with either GST–PTAC58 or GST–β subunit fusion proteins which were absorbed to glutathione–Sepharose. Stat1 failed to associate with GST–PTAC58 (Figure 1A, lanes 3 and 4), whereas SV40 large T antigen did associate with the GST–PTAC58 after incubation with whole cell extract of SV40-transformed human fibroblasts (Wi38 VA13) (data not shown). In contrast, the interaction of GST–β subunit with Stat1 was dependent on IFN-γ treatment (Figure 1A, lane 6). Precipitated Stat1 was tyrosine phosphorylated and the interaction between Stat1 and GST–β subunit reached a maximum at 15 min after IFN-γ treatment, and decreased thereafter in a time-dependent manner (Figure 1B). This time course of interaction between GST–β subunit and Stat1 paralleled the tyrosine phosphorylation of Stat1 (data not shown; Shuai et al., 1992). These data indicate that the β subunit associates with the activated form of Stat1. Figure 1.The indirect interaction of β subunit with tyrosine-phosphorylated Stat1. (A) HeLa whole cell extract with or without IFN-γ treatment was incubated with GST–PTAC58 or GST–β subunit fusion protein absorbed to glutathione–Sepharose. Precipitates were separated on 7.5% SDS–PAGE, transferred to nitrocellulose and immunoblotted with anti-Stat1 antibody. (B) Time course of interaction between Stat1 and β subunit. HeLa whole cell extract treated with IFN-γ for the indicated times was incubated with GST–β subunit, and the precipitated Stat1 was detected with anti-Stat1 or anti-phosphotyrosine antibodies. (C) Cytosol (C) and nuclear extract (N) were prepared from HeLa cells treated with (+) or without (−) IFN-γ. Extracts were incubated with GST–β subunit, and associated Stat1 was detected as for (A). C(+) and N(+), cytosol and nuclear extract prepared from IFN-γ-treated cells, respectively; C(−), cytosol prepared from untreated cells. Download figure Download PowerPoint We attempted to determine whether β subunit associates with Stat1 directly or indirectly. We precipitated Stat1 with the GST–β subunit from cytosol or nuclear extracts of HeLa cells with or without IFN-γ treatment. Although both cytosol and nuclear extract treated with IFN-γ contain tyrosine-phosphorylated Stat1 (data not shown), Stat1 was precipitated with β subunit from cytosol but not from the nuclear extract (Figure 1C, lanes 1 and 2). Stat1 was precipitated more efficiently on incubation with a mixture of both cytosol and nuclear extracts with IFN-γ treatment (Figure 1C, lane 3). Furthermore, Stat1 was also precipitated from nuclear extracts with IFN-γ treatment when cytosol without IFN-γ stimulation was added (Figure 1C, lane 4). These results indicate that the β subunit interacts with activated Stat1 indirectly and that additional factor(s) exist in the unstimulated cytosol which are required for this association. The NPI-1 family, but not the Rch1 family, associates with tyrosine-phosphorylated Stat1 It is well known that β subunit associates with a karyophile via the α subunit (Görlich et al., 1995; Imamoto et al., 1995c). In humans, two types of α subunit, NPI-1 and Rch1, have already been isolated. PTAC58 is a mouse homolog of Rch1 (93% amino acid identity). Therefore, we prepared GST–NPI-1, and examined its interaction with Stat1. The GST–NPI-1 fusion protein showed in vitro transport activity for T-NLS–BSA in conjunction with β subunit and Ran, and this activity was comparable with that of proteins devoid of the GST portion (data not shown). Using a whole cell extract of HeLa cells treated with IFN-γ, it was found that only GST–NPI-1, but neither GST nor GST–PTAC58, associated with Stat1 (Figure 2A upper). The precipitated Stat1 was tyrosine phosphorylated (Figure 2A lower). Figure 2.The interaction of NPI-1 with tyrosine-phosphorylated Stat1. (A) HeLa whole cell extract with or without IFN-γ treatment was incubated with GST, GST–NPI-1 or GST–PTAC58 adsorbed to glutathione–Sepharose. Precipitates were separated on 7.5% SDS–PAGE, transferred to nitrocellulose and immunoblotted with anti-Stat1 and anti-phosphotyrosine antibodies. (B) NIH-3T3 whole cell extract with or without murine IFN-γ treatment was used as for (A) instead of HeLa cells. Download figure Download PowerPoint Since PTAC58 is a mouse protein, it is possible that human Stat1 does not interact with mouse PTAC58, but does so with its human homolog, Rch1. To clarify this problem, we attempted the same experiment using mouse NIH-3T3 cells with or without murine IFN-γ treatment. The result showed that intrinsic mouse Stat1 in NIH-3T3 cells associated with human NPI-1, but not with mouse PTAC58 (Figure 2B), which is consistent with the results obtained with HeLa cell extracts. Since these results were obtained using an in vitro solution binding assay employing recombinant proteins, in order to confirm whether native NPI-1 and β subunit interact with Stat1 in vivo, we transiently expressed His6-tagged Stat1 in human 293T cells. After IFN-γ stimulation, the expressed His6-Stat1 was precipitated with Ni-NTA–agarose and the bound proteins were analyzed by immunoblotting with anti-NPI-1 antibodies, or anti-PTAC58 antibodies which specifically recognize the human counterpart, Rch1. The results showed that the expressed His6-Stat1 interacted with NPI-1, but not with Rch1, in an IFN-γ-dependent manner (Figure 3). In this series of experiments, β subunit was not detected by immunoblotting, probably as a result of a large amount of co-migrating His6-Stat1 in SDS–PAGE. These results indicate that activated Stat1 associates with NPI-1, but not with Rch1 in vivo. Figure 3.The interaction of endogenous NPI-1 with expressed Stat1. His6-tagged Stat1 was transiently expressed in human 293T cells and the cells were treated with or without IFN-γ. Ni-NTA–agarose-bound proteins from the cell extract were separated on SDS–PAGE and immunoblotted with anti-NPI-1, anti-PTAC58, anti-β subunit and anti-Stat1 antibodies. As a control experiment, the cell extract was prepared from mock-transfected 293T cells. Download figure Download PowerPoint To confirm further that NPI-1 and β subunit participate in the transport of Stat1 from cytoplasm to the nucleus in vivo, affinity-purified polyclonal anti-NPI-1 and anti-β subunit antibodies were used. Both antibodies recognized one band when a whole cell extract of HeLa cells was analyzed by immunoblotting (Figure 4A). When recombinant NPI-1 and PTAC58 were resolved, anti-NPI-1 antibodies had no cross-reactivity with PTAC58, although anti-PTAC58 antibodies cross-reacted slightly with NPI-1 (data not shown). First, by using these antibodies, we checked ability of these antibodies to inhibit the nuclear import of T-NLS–BSA. Microinjected anti-PTAC58 and anti-β subunit antibodies inhibited the nuclear import of T-NLS–BSA effectively, but anti-NPI-1 antibodies did so only slightly (Figure 4B). To assess the effects of these antibodies on signal-dependent nuclear import of Stat1, we prepared hemagglutinin (HA)-tagged Stat1 to detect intracellular localization of injected Stat1 by indirect immunofluorescence with anti-HA antibodies, as described previously (Sekimoto et al., 1996). Stat1, when microinjected into the cytoplasm of HeLa cells, was translocated into the nucleus dependent on IFN-γ stimulation (Figure 4C). Microinjected anti-PTAC58 antibodies, which have already been shown to strongly inhibit the nuclear import of T-NLS–BSA (Imamoto et al., 1995b; Figure 4B), had no effect on the IFN-γ-dependent nuclear import of Stat1 (Figure 4D). In contrast, affinity-purified anti-NPI-1 antibodies inhibited the IFN-γ-dependent nuclear import of Stat1 (Figure 4D). Furthermore, affinity-purified anti-β subunit antibodies strongly inhibited the nuclear import of Stat1 when injected 30 min prior to IFN-γ stimulation (Figure 4D). It is also noteworthy that, unlike anti-Ran (Sekimoto et al., 1996) or anti-β subunit antibodies, cytoplasmic injection of anti-NPI-1 antibodies, but not anti-PTAC58 antibodies caused the distinct aggregation of Stat1 in the cytoplasm. A similar aggregation pattern was observed when T-NLS–BSA was co-injected with anti-PTAC58 antibodies (Figure 4B), and this aggregation was supposed to be the result of precipitation of T-NLS–BSA with Rch1 by the antibodies (Imamoto et al., 1995b). These results confirm that NPI-1, but not Rch1, and β subunit are involved in the IFN-γ-dependent nuclear translocation of Stat1. Figure 4.Effect of affinity-purified antibodies against nuclear import factors on the nuclear import of Stat1. (A) HeLa whole cell extracts were resolved by SDS–PAGE and immunoblotted with anti-NPI-1, anti-PTAC58, anti-β subunit or their pre-immune IgG. (B) Pre-immune IgG or affinity-purified anti-NPI-1, anti-PTAC58 or anti-β subunit antibodies (25 mg/ml) were co-injected with FITC-labeled T-NLS–BSA and, after 30 min incubation, the cells were fixed. (C) Purified HA-tagged Stat1 protein was microinjected into the cytoplasm of HeLa cells and the cells were treated with IFN-γ (50 ng/ml) for 30 min. After fixation of the cells, the intracellular localization of Stat1 was examined by indirect immunofluorescence with anti-HA antibody. (D) Pre-immune IgG or affinity-purified anti-NPI-1, anti-PTAC58 or anti-β subunit antibodies (25 mg/ml) were microinjected with Stat1 into the cytoplasm of HeLa cells. Thirty minutes later, the cells were stimulated with IFN-γ (50 ng/ml) for 30 min. Download figure Download PowerPoint It has been proposed that tyrosine-phosphorylated Stat1 forms a homodimer between phosphotyrosine and the SH2 domain of each Stat1, and the dimerization is required for its DNA binding (Shuai et al., 1994). It is known that the DNA binding of Stat1 is inhibited in the presence of phosphotyrosine, but not phosphoserine and phosphothreonine, which means that free phosphotyrosine disrupts the homodimerization of Stat1 (Sadowski et al., 1993). We confirmed that increasing amounts of free phosphotyrosine caused inhibition of DNA binding of Stat1 (Figure 5A). To determine whether tyrosine-phosphorylated Stat1 binds to NPI-1 as a homodimer, we assessed the effect of free phosphotyrosine, phosphoserine or phosphothreonine on the interaction between Stat1 and NPI-1. We found that phosphotyrosine, but not phosphoserine or phosphothreonine, inhibited the interaction of Stat1 with NPI-1 in a dose-dependent manner (Figure 5B, upper panel), whereas these phosphorylated amino acids had no effects on the interaction between NPI-1 and SV40 T antigen (Figure 5B, lower panel). These results clearly indicate that homodimerization of Stat1 is required for its interaction with NPI-1. Figure 5.Effects of phosphotyrosine on the binding of Stat1 to DNA or NPI-1. (A) HeLa whole cell extract with or without IFN-γ treatment was incubated with high affinity SIE of c-fos gene-conjugated Sepharose in the presence of increasing amounts of phosphoserine, phosphotyrosine or phosphothreonine (2.5, 5 and 20 mM). Gels were washed extensively, and bound proteins were detected by immunoblotting using anti-Stat1. (B) HeLa whole cell extract with or without IFN-γ treatment, or whole cell extract of SV40-transformed human fibroblasts (Wi38 VA13 cells) was incubated with GST–NPI-1 in the presence of increasing amounts of phosphoserine, phosphotyrosine or phosphothreonine (2.5, 5 and 20 mM). The bound proteins were detected by immunoblotting with anti-Stat1 or anti-SV40 T antigen antibodies. Download figure Download PowerPoint Stat1 does not have a conventional NLS Since the NLSs previously identified consist of a cluster of basic amino acid residues in the primary structure (Dingwall and Laskey, 1991) and NPI-1 is known to recognize the NLSs, we examined whether the primary structure of Stat1 contains a functional NLS belonging to the NLSs identified previously. To determine the region of Stat1 which is critical for IFN-γ-dependent nuclear import, we constructed several deletion mutants of Stat1 and expressed them transiently in human FL cells. Unlike wild-type Stat1, however, none of these mutants was translocated to the nucleus, even in the presence of IFN-γ (data not shown). Therefore, since human Stat1 contains three regions of basic amino acid clusters in its deduced amino acid sequence (Figure 6A), we introduced mutations into these regions by site-directed mutagenesis, in order to assess their contribution to the nuclear import of Stat1. Because the substitution of three or four amino acids caused low expression and instability of the mutant Stat1, we constructed nine mutants containing single or double mutations, expressed them in Escherichia coli and purified them to homogeneity. When microinjected, all the recombinant mutant proteins were transported into the nucleus in a manner similar to the wild-type, although the R88L mutant showed a slightly reduced nuclear import activity (Figure 6B). However, when the region in the vicinity of Arg88, 78LLQHNIRKSKRNLQ91, was inserted into the GST–green fluorescence protein (GFP) fusion, this protein was not transported into the nucleus in the presence or absence of IFN-γ, although the insertion of the NLS of SV40 T antigen showed efficient import activity of the fusion protein (data not shown). These data indicate that Stat1 does not contain a classical NLS, as defined by the basic amino acid cluster. Figure 6.Nuclear import activity of amino acid substitution mutants of basic amino acid clusters of Stat1. (A) Three basic regions exist in the deduced amino acid sequence of human Stat1. (B) Mutants in these basic regions were purified to homogeneity and microinjected into the cytoplasm of human FL cells. The intracellular localization of injected mutants after stimulation with IFN-γ is shown. N, nucleus; C, cytoplasm. Download figure Download PowerPoint The Stat1-binding domain of NPI-1 is located at its C-terminus Since it was found that NPI-1 is capable of interacting with Stat1 as well as the SV40 T antigen NLS substrate, we tried to determine if Stat1 and SV40 T antigen NLS interact with the same domain of NPI-1. For this purpose, we prepared several deletion mutants of NPI-1 and assessed their binding ability with Stat1 and SV40 large T antigen. These deletion mutants of NPI-1 were expressed and purified as GST fusion proteins (Figure 7A). We first checked the binding activity of these mutants with SV40 large T antigen by using Wi38 VA13 cell extracts. Four mutants (294–538, 338–538, 404–538 and 425–538) failed to bind to the SV40 large T antigen. In contrast, deletion mutants of the C-terminal region (1–492, 1–475 and 78–475) bound to the SV40 large T antigen to the same extent as the full-length NPI-1 (Figure 7B, upper). The same results were obtained when the bipartite-type NLS of nucleoplasmin-conjugated BSA was used as substrate (data not shown), suggesting that the binding domain of NPI-1 for the conventional basic-type NLSs is located in the armadillo repeats. Using HeLa cell extracts treated with IFN-γ, all the N-terminal deletion mutants of NPI-1 were found to associate with Stat1, whereas the C-terminal deletions abolished the Stat1-binding activity of NPI-1 (Figure 7B, lower). Although attempts were made to prepare a shorter form of NPI-1, to define the essential Stat1-binding domain, these short mutant constructs were not expressed as complete entities in E.coli. Figure 7.Mapping of the NPI-1 region which interacts with Stat1. (A) Deletion mutants of NPI-1 used in this study. All mutants were expressed as GST fusion protein in E.coli and purified using glutathione–Sepharose. (B) The ability of these mutants to bind to SV40 T antigen or Stat1 is shown. Wi38 VA13 or HeLa whole cell extracts treated with IFN-γ were incubated with various deletion mutants of NPI-1, and bound proteins were detected by immunoblotting with anti-SV40 T antigen or anti-Stat1 antibodies. (C) Effects of T-NLS–BSA on the binding of NPI-1 to SV40 T antigen or Stat1. Wi38 VA13 or HeLa whole cell extracts treated with IFN-γ were incubated with GST–NPI-1 in the presence of wild-type or reverse type T-NLS–BSA (0.1 and 1 μM), and bound proteins were detected by immunoblotting with anti-SV40 T antigen or anti-Stat1 antibodies. Download figure Download PowerPoint In order to determine whether NPI-1 binds to two cargo proteins (Stat1 and T-NLS–BSA) simultaneously, competition experiments were done. The interaction between SV40 large T antigen and NPI-1 was competed with wild-type T-NLS–BSA, but not reverse type T-NLS–BSA (Figure 7C, upper). In contrast, the presence of T-NLS–BSA did not interfere with the interaction between Stat1 and NPI-1 (Figure 7C, lower). Addition of wild-type SV40 T antigen NLS peptides (0.1 mM) caused complete inhibition of the interaction between SV40 T antigen and NPI-1, but NLS peptide had no effect on the interaction between Stat1 and NPI-1 even at 1 mM (data not shown). These results indicate that the binding site of Stat1 is distinct from that of the conventional basic-type NLS and, thus, that NPI-1 contains at least two binding sites for karyophiles, and Stat1 and conventional NLS-containing substrate can bind to NPI-1 simultaneously. In vivo inhibition of nuclear import of Stat1 and SV40 T-NLS substrate by distinct dominant-negative forms of NPI-1 It is well known that the amino-terminal region of NPI-1 is essential for β subunit binding, and deletion of this region causes loss of import activity. As expected, NPI-1 mutants 294–538 and 78–475 bind to Stat1 and SV40 T antigen NLS, respectively (Figure 7B), but not to β subunit (data not shown). Therefore, we examined whether they can inhibit the nuclear protein import in a dominant-negative manner. When 294–538 was microinjected with Stat1 or T-NLS–BSA into the cytoplasm of HeLa cells, the IFN-γ-dependent nuclear import of Stat1 was strongly inhibited, but not the nuclear import of T-NLS–BSA (Figure 8). In contrast, co-injection of 78–475 only inhibited the nuclear import of T-NLS–BSA. Injection of GST or GST–NPI-1 (full-length) had no effect on the nuclear import of
Referência(s)