|Home | About | Journals | Submit | Contact Us | Français|
Nucleosome assembly protein 1 (Nap1) is widely conserved from yeasts to humans and facilitates nucleosome formation in vitro as a histone chaperone. Nap1 is generally localized in the cytoplasm, except that subcellular localization of Drosophila melanogaster Nap1 is dynamically regulated between the cytoplasm and nucleus during early development. The cytoplasmic localization of Nap1 is seemingly incompatible with the proposed role of Nap1 in nucleosome formation, which should occur in the nucleus. Here, we have examined the roles of a putative nuclear export signal (NES) sequence in yeast Nap1 (yNap1). yNap1 mutants lacking the NES-like sequence were localized predominantly in the nucleus. Deletion of NAP1 in cells harboring a single mitotic cyclin gene is known to cause mitotic delay and temperature-sensitive growth. A wild-type NAP1 complemented these phenotypes while nap1 mutant genes lacking the NES-like sequence or carboxy-terminal region did not. These and other results suggest that yNap1 is a nucleocytoplasmic shuttling protein and that its shuttling is important for yNap1 function during mitotic progression. This study also provides a possible explanation for Nap1's involvement in nucleosome assembly and/or remodeling in the nucleus.
Chromatin is one of the hallmarks of eukaryotes. The eukaryotic genome DNA is complexed with chromosomal proteins to form the chromatin structure in the nucleus. Nuclear reactions such as DNA replication, transcription, DNA repair, and recombination take place on the chromatin and are regulated by the disruption and assembly of chromatin, namely chromatin remodeling. Recently, a variety of chromatin remodeling factors has been identified, including histone modification enzymes and ATP-dependent factors (reviewed in references 1, 2, 22, and 42). The histone chaperone family is one of the chromatin remodeling factors that binds to core histones and facilitates assembly and remodeling of the chromatin in an ATP-independent manner. This family includes various kinds of proteins, such as nucleoplasmin (31) and N1/N2 (28) in Xenopus, nucleosome assembly protein 1 (designated Nap1 in this paper) and Nap2 (13, 48), template-activating factor Iα (TAF-Iα) and TAF-Iβ/SET (41), chromatin assembly factor 1 (CAF-1) (6), dNLP (17), Spt6 (5), DF31 (9), CIA (40), TAF-III/B23 (45), and so on.
Nap1 was first identified from HeLa cell extracts as a protein that stimulates nucleosome formation from DNA and purified core histones (14). Nap1 directly binds to core histones and transfers them onto naked DNA (15, 36, 54). Recent studies suggest that Nap1 is involved in transcriptional regulation through chromatin remodeling (32). For example, Nap1 remodels the chromatin-like structure of the adenovirus genome DNA and stimulates transcription as well as replication (24). Nap1 acts synergistically with ATP-dependent chromatin remodeling factors such as NURF and ACF to remodel the chromatin structure (18, 19, 23, 43). Furthermore, Nap1 and Nap2 are found to interact with p300, a histone acetyltransferase, and to stimulate transcription from p21 and E2F promoters (20, 49). Considering the roles of Nap1 in nucleosome formation and/or transcription, Nap1 should exist in the nucleus. However, it is widely accepted that Nap1 is localized in the cytoplasm but not in the nucleus (26, 35; unpublished observations). One exceptional observation is that Drosophila melanogaster Nap1 is present in nuclei limitedly during some stages of development (16). A possible explanation for this contradiction is that Nap1 is transiently present in nuclei and regulatedly exported from nuclei, although the static subcellular localization of Nap1 is cytoplasmic. Here, we have addressed this issue and further tried to understand the physiological meaning of nuclear localization and nucleocytoplasmic shuttling of Nap1.
In the yeast Saccharomyces cerevisiae, deletion of NAP1 causes no phenotypes or very weak phenotypes, if any (26, 50). In yeasts, there are four cyclin B homologs, Clb1, Clb2, Clb3, and Clb4. Nap1 physically and genetically interacts with Clb2 (25, 26). Deletion of NAP1 in a Clb2-dependent strain, in which CLB1, CLB3, and CLB4 are disrupted, shows a temperature-sensitive growth phenotype and an elongated bud phenotype (25, 26). It is suggested that the elongated bud phenotype is caused by a prolonged mitotic delay because disappearance of Clb2 at the end of mitosis is delayed in the ΔCLB1 ΔCLB3 ΔCLB4 ΔNAP1 strain (25). Different kinds of factors and genes are known to be involved in the elongated bud phenotype. A cdc28-1N mutation shows the elongated bud phenotype (52); a clb2ts mutant, in which CLB1, CLB3, and CLB4 are disrupted, shows the elongated bud phenotype at nonpermissive temperatures (10); and overexpression of Swe1 kinase, which inactivates Cdc28 by phosphorylation, also causes the elongated bud phenotype (4, 34). These observations suggest that the compromised activity of Cdc28-Clb kinase causes the elongated bud phenotype and that a target protein(s) of this kinase is responsible for the elongated bud phenotype. Gin4 is a protein kinase that is localized in the bud neck and activated during mitosis by phosphorylation. Mutations in this gene also cause the elongated bud phenotype (3). The mitosis-specific phosphorylation of Gin4 requires Nap1 and Clb2 (3). Gin4 phosphorylation is downstream of Clb2 because an ectopic expression of destruction box-deleted Clb2 causes phosphorylation of Gin4 in interphases (3). It is interesting that Gin4 physically and genetically interacts with Nap1 (3). Thus, Nap1 may play an important role during mitotic progression from Cdc28-Clb regulation to cytokinesis, although an exact in vivo function(s) of yeast Nap1 (yNap1) remains unclear. Identification of interactors of Nap1 further emphasizes this point. Nap1 interactors identified by biochemical methods and two-hybrid screening (8) includes Clb2 (26), Nbp1 (50), Gin4 (3), CK2 (33), Hta1, Kcc4, Yrb1, Jip1 (53), Yap6, and Kap95 (21). Note that Yrb1 and Kap95 are closely related to the nuclear import mechanism (30). Recently, Mosammaparast et al. reported that Kap114 is involved in the nuclear import of Nap1 (38).
In this study, we show that yNap1 is a nucleocytoplasmic shuttling protein. A region of yNap1 involved in histone binding was found to be important for nuclear import of yNap1, and nuclear export of yNap1 was dependent on a nuclear export signal (NES)-like sequence. To elucidate the role of the Nap1 nucleocytoplasmic shuttling, we used a yNAP1-disrupted strain in complementation assays and showed that nucleocytoplasmic shuttling, but not cytoplasmic localization of yNap1, is important for mitotic progression. Our findings solve the contradiction in Nap1 localization and show the possibility that Nap1 acts as a histone chaperone in vivo. In addition, these data provide a cue to its physiological function during mitotic progression.
Except where noted, all yeast strains were grown at 25°C in yeast extract-peptone-dextrose (YPD) medium. For protein expression induced by galactose, single colonies were picked up and first inoculated into synthetic defined (SD) medium containing 2% glucose. Cells were grown at 25°C until the optical density at 600 nm (OD600) reached 0.8. Then cells were collected by centrifugation at 800 × g and washed twice with phosphate-buffered saline (PBS). The collected cells were grown in synthetic galactose (SG) medium containing 2% galactose for 6 to 8 h, allowing the induced expression of proteins.
To construct expression vectors of yNap1 derivatives in yeast cells, promoter and terminator fragments of the yNAP1 gene were amplified by PCR with primers 5′-GTGGCGGCCGCCGGGCACTCCCTTCTTTCTTCC-3′ and 5′-GGAGGATCCACTAGTTCTAGAGATCTTGCGCTTTGCTCTTGGTCC-3′ for the promoter and 5′-ATCTCTAGAACTAGTGGATCCTCCCAACGCACTTCGCAAGAGTG-3′ and 5′-CCCCTCGAGTAACTCGTCCAACTTGAAGGCCCC-3′ for the terminator, with the yeast genomic DNA as a template. Overlap extension was performed by PCR with primers 5′-GTGGCGGCCGCCGGGCACTCCCTTCTTTCTTCC-3′ and 5′-CCCCTCGAGTAACTCGTCCAACTTGAAGGCCCC-3′, with promoter and terminator PCR products as templates. PCR products were digested with NotI and XhoI and cloned into NotI- and XhoI-digested pRS316 plasmid, generating pRS316-yNap1pt, which contains a unique BamHI site between the promoter and the terminator fragments. pBluescript-Flag was constructed from BamHI- and XhoI-digested pBluescript SK(+) (Stratagene) and a DNA fragment made by annealing two synthetic oligonucleotides, 5′-GATCCGCCGCCACCATGGACTACAAGGATGACGACGACAAGCATATGC-3′ and 5′-TGCAGCATATGCTTGTCGTCGTCATCCTTGTAGTCCATGGTGGCGGCG-3′. yNAP1 cDNA fragments were amplified by PCR with primers 5′-AGCGCAAGCATATGTCAGACCCTATCAGAACGAAAC-3′ and 5′-GCGGGATCCTTATGACTGCTTGCATTCAGGAG-3′. The amplified DNA fragment was phosphorylated with T4 polynucleotide kinase (Toyobo), blunted with Klenow fragment, and digested with NdeI. Then the fragment was cloned into NdeI- and EcoRV-digested pBluescript-Flag (pBS-Flag-yNap1). For yNap1ΔNES, DNA fragments corresponding to the N- and C-terminal segments were prepared by PCR with primers 5′-AGCGCAAGCATATGTCAGACCCTATCAGAACGAAAC-3′ and 5′-CAACTTCGAATAGCTCGCTTTGACCCCCCACATACCCGCTGT-3′ and primers 5′-CAAAGCGAGCTATTCGAAGTTG-3′ and 5′-GCGGGATCCTTATGACTGCTTGCATTCAGGAG-3′, respectively, with the yNap1 cDNA as a template. The two segments were ligated by overlap extension PCR with primers 5′-AGCGCAAGCATATGTCAGACCCTATCAGAACGAAAC-3′ and 5′-GCGGGATCCTTATGACTGCTTGCATTCAGGAG-3′. Amplified DNA fragments were phosphorylated with T4 polynucleotide kinase, blunted with Klenow fragment, and digested with NdeI. The fragment was cloned into NdeI- and EcoRV-digested pBS-Flag and designated pBS-Flag-yNap1ΔNES. For yNap1 (L99A L102A), DNA fragments corresponding to the N- and C-terminal segments were prepared by PCR with primers 5′-AGCGCAAGCATATGTCAGACCCTATCAGAACGAAAC-3′ and 5′-CGCTTTGTGCAGTCTTAGCGCTCAGCAGC-3′ and primers 5′-GCTGAGCGCTAAGACTGCACAAAGCGAGC-3′ and 5′-GCGGGATCCTTATGACTGCTTGCATTCAGGAG-3′, respectively, with the yNap1 cDNA as a template. These two segments were ligated by overlap extension PCR with the primers 5′-AGCGCAAGCATATGTCAGACCCTATCAGAACGAAAC-3′ and 5′-GCGGGATCCTTATGACTGCTTGCATTCAGGAG-3′. Amplified DNA fragments were phosphorylated with T4 polynucleotide kinase, blunted with Klenow fragment, and digested with NdeI. The fragment was cloned into NdeI- and EcoRV-digested pBS-Flag and designated pBS-Flag-yNap1 (L99A L102A). Flag-yNap1 wild-type (WT) and mutant fragments, except for ΔC1 mutants, were excised from pBS-Flag-yNap1 derivatives by digestion with BamHI and cloned into the BamHI site of pRS316-yNap1pt or pYES2. Flag-yNap1ΔC1 and yNap1ΔNESΔC1 were excised from pBS-Flag-yNap1 or pBS-Flag-yNap1ΔNES by digestion with BamHI and XhoI and ligated with the plasmid pYES2, which was digested with BamHI and XhoI. To construct yNap1 expression vectors for HeLa cells, Flag-yNap1 fragments excised from pBS-Flag-yNap1 plasmids with BamHI were cloned into BamHI-digested pcDNA3 (Invitrogen). To construct Flag-yNap1ΔC1 and Flag-yNap1ΔNESΔC1 expression vectors, plasmids pBS-Flag-yNap1 and pBS-Flag-yNap1ΔNES were digested with BamHI and XhoI, blunted with Klenow, and cloned into the EcoRV site of pcDNA3. To construct enhance green fluorescent protein (EGFP)-Nap1 expression vectors, Nap1 coding sequences were cut out with NdeI and BamHI from the pBS-Flag-yNap1 plasmids and cloned into the BglII site of pEGFP-C1 (Clontech) by blunt-end ligation. To construct the glutathione S-transferase (GST)-fused recombinant yNap1 protein expression vector, yNap1 cDNA fragments digested with NdeI and BamHI and then blunted with Klenow fragment were cloned into pGEX-5X-3 (Amersham Biosciences), which had been digested with SalI and filled with Klenow fragment. To construct pYES2-Clb2-EGFP, two DNA fragments were made by PCR as follows. The Clb2 fragment was obtained by PCR with oligonucleotides 5′-CGCTAGCGAATTCCGCCACCATGTCCAACCCAATAGAAAACACA-3′ and 5′-TGGTGGCGACCGGTGGATCTTCATGCAAGGTCATTATATCATAGCC-3′ as primers and with CLB2 cDNA as a template. The EGFP fragment was obtained by PCR with oligonucleotides 5′-GATCCACCGGTCGCCACCATGG-3′ and 5′-CGGATCCTCGAGTTACTTGTACAGCTCGTCCATGCC-3′ as primers and with pEGFP-N1 (Clontech) as a template. Then overlap extension PCR was performed with oligonucleotides 5′-CGCTAGCGAATTCCGCCACCATGTCCAACCCAATAGAAAACACA-3′ and 5′-CGGATCCTCGAGTTACTTGTACAGCTCGTCCATGCC-3′ as primers and with the two DNA fragments as templates. The overlap extension products were gel purified, digested with EcoRI and BamHI, and cloned into EcoRI- and BamHI-digested pYES2 to generate pYES2-Clb2-EGFP. To construct yCP, a low-copy-number plasmid carrying the NAP1 gene, the coding region of NAP1 was cloned between SmaI and SphI sites in a TRP1/CEN4-based yeast plasmid, YCp19, so that the region between the SmaI and SphI sites in URA3 was replaced by NAP1. To construct yEP, a high-copy-number plasmid carrying the NAP1 gene, the coding region of NAP1 was cloned between the SmaI and SphI sites in a TRP1, URA3, 2-based plasmid, YEp24.
To prepare recombinant GST-yNap1, Escherichia coli BL21(DE3) cells were transformed with pGEX-5X-3 derivatives and grown in Lurai-Bertani medium containing 100 μg of ampicillin/ml. Isopropyl-β-d-thiogalactopyranoside (IPTG) was added at a concentration of 1 mM when the OD600 reached 0.3, and cells were further incubated for 3 h. The cells were harvested by centrifugation and washed with PBS. Recombinant proteins were purified by using glutathione Sepharose beads according to the manufacturer's instructions (Amersham Biosciences).
Single colonies of DK168 and its transformants were grown at 25°C in YPD for 12 h. Cells were plated onto slide glasses, covered with coverslips, and subjected to examination under a microscope (Axioscope; Zeiss). To examine the temperature sensitivity of each strain, single colonies were inoculated into 2 ml of YPD medium and incubated at 25°C for 24 h. Cell concentrations were adjusted to ODs of 0.1, 0.01, and 0.001 with YPD, and 2 μl of each dilution was spotted onto SD plates lacking His, Trp, Leu, and Ura and incubated at either 25 or 37°C for 1 or 2 days. To examine the growth rate of yNap1-overexpressing strains, a single colony of each strain was streaked onto SD and SG plates lacking His, Trp, Leu, and Ura and incubated at 25°C for 2 or 3 days.
The NAP1 gene was disrupted by insertion of LEU2. To this end, a LEU2 gene sandwiched with upstream and downstream flanking regions of a nap1 gene was amplified by PCR from the genomic DNA of DK168, whose NAP1 is disrupted by LEU2. The Δnap1::LEU2 DNA was transformed into PSY1785 and PSY1784 strains carrying WT KAP114 and disrupted KAP114 alleles, respectively (37). The disruption of the NAP1 gene was confirmed by PCR.
Exponentially growing yeast cells were harvested by centrifugation and washed twice with PBS. Cells were suspended in a cell lysis buffer containing 62.5 mM Tris-HCl (pH 6.8), 10% glycerol, 145 mM β-mercaptoethanol, and 0.25% sodium dodecyl sulfate (SDS). Glass beads were added, and cells were lysed by vigorously shaking and boiling at 98°C for 5 min. The lysates were cleared by centrifugation at 15,000 × g for 5 min, and their protein concentrations were determined by using the protein assay kit (Bio-Rad). Aliquots (equivalent to 10 μg of protein) were subjected to SDS-7.5% polyacrylamide gel electrophoresis (PAGE) followed by Western blotting with the antibodies indicated in each figure legend.
Yeast cell extracts were prepared from exponentially growing cells cultured in SG (without Ura) for 6 to 8 h. The cell cultures were harvested by centrifugation and frozen in liquid nitrogen. The cells were sonicated in lysis buffer (50 mM HEPES-KOH [pH 7.6], 50 mM KCl, 0.5 mM EGTA, 1 mM MgCl2, 10% glycerol, 2 μg of aprotinin/ml, 2 μg of leupeptin/ml, 2 μg of pepstatin A/ml, 1 mM phenylmethylsulfonyl fluoride) in the presence of glass beads. The cell lysates were obtained from sonicated lysates by centrifugation. The protein concentration was determined by the Bradford method. The soluble yeast cell extracts (200 μg of protein) and anti-Flag antibody-conjugated agarose beads (Sigma) were mixed and incubated at 4°C for 30 min. Then the beads were washed four times with the lysis buffer, and bound proteins were separated by SDS-7.5% PAGE followed by immunoblotting with anti-Flag or anti-Clb2 (Santa Cruz) antibodies.
To prepare resins bound by GST or GST fusion proteins, glutathione-Sepharose beads (Amersham Bioscience) were mixed with E. coli lysates containing GST or GST fusion proteins, washed three times with a binding buffer containing 500 mM NaCl, 50 mM Tris (pH 7.9), 1 mM EDTA, and 1% Triton X-100, and washed three times with a buffer containing 50 mM KCl, 50 mM HEPES-KOH (pH 7.6), 1 mM EGTA, 1 mM MgCl2, and 0.01% Tween 20. To prepare [35S]methionine-labeled yNap1 proteins, in vitro transcription-translation reactions were performed according to the manufacturer's instructions (Invitrogen). As a template for transcription with T7 RNA polymerase, pYES2-Clb2-EGFP was digested with NcoI, which removes the EGFP coding region downstream of Clb2. In vitro-translated Clb2 was incubated with the resins bound by GST or GST-Nap1 at 4°C for 1 h. Beads were washed 3 times with binding buffer and eluted with an SDS sample buffer. Bound materials were separated by SDS-10% PAGE and visualized by Coomassie brilliant blue (CBB) staining and autoradiography.
Supercoiling assays were performed essentially as described previously (11). Recombinant yNap1 and its derivatives were prepared as GST fusion proteins.
HeLa cells were grown in minimal essential medium supplemented with 10% fetal calf serum. For the subcellular localization study, HeLa cells were split and grown on coverslips placed in 6-cm-diameter tissue culture dishes for 16 h before transfection. Cells in each dish were transfected with 10 μg of mammalian expression vectors by the standard calcium phosphate method (27). At 6 h posttransfection, the medium was changed and cells were further incubated for 18 h. Cells were washed twice with PBS and fixed for 10 min at room temperature with 4% formaldehyde that was freshly prepared from paraformaldehyde. Coverslips were washed with 0.5% Triton X-100 in PBS for 10 min and washed three times with PBS. To visualize Flag-tagged yNap1 proteins, fixed cells were serially incubated with 1% nonfat dried milk in PBS for 30 min, with 10 μg of anti-Flag M2 antibody (Sigma)/ml in 1% nonfat dry milk in PBS for 1 h, with rhodamine-conjugated anti-mouse immunoglobulin G (Cappel) for 1 h, and with 1 μg of 4′,6′-diamino-2-phenylindole dihydrochloride (DAPI)/ml in PBS for 10 min. The cells were then observed under a fluorescence microscopy (Axioscope). To measure the fluorescent intensity in the nucleus and cytoplasm, EGFP-Nap1, EGFP-Nap1ΔNES, EGFP-Nap1ΔC1, and EGFP-Nap1ΔNESΔC1 were expressed in HeLa cells by conventional calcium phosphate transfection and fixed with 4% paraformaldehyde for 10 min and DNA was stained with DAPI. Transfected cells were analyzed by the laser scanning cell cytometer LSC2 (Olympus).
Indirect immunostaining of yeast cells with anti-Flag antibody was performed as described previously (51). DNA was visualized by DAPI staining.
Nap1 has been identified as a protein that facilitates formation of nucleosomes from DNA and histones in vitro. Therefore, Nap1 should be present in the nucleus where nucleosome formation occurs. However, it is reported that the subcellular localization of Nap1 in mammalian and yeast cells is restricted to the cytoplasm (26, 35). To get insight into a physiological function(s) of Nap1 as a histone chaperone, we initiated this study by carefully reexamining the subcellular localization of Nap1. In a previous study (16), a putative NES-like sequence was pointed out in Drosophila Nap1, leading to a hypothesis that Nap1 may be imported into and exported from the nucleus. To study this point further, we looked for NES-like sequences in Nap1 homologs of other species. We found NES-like sequences in Nap1's in humans, nematodes, and yeasts (Fig. (Fig.1A).1A). To examine whether the putative NES is functionally active, we constructed expression vectors for Flag-tagged yNap1ΔNES mutant protein, which lacks the region between amino acid positions 88 to 102 encompassing the NES-like sequence, and two kinds of amino acid substitution mutant proteins in the NES-like sequence, Flag-yNap1 (L88A V92A) and Flag-yNap1(L99A L102A) (Fig. (Fig.1).1). We tested the subcellular localization of these yNap1 mutant proteins in mammalian (Fig. (Fig.2)2) and yeast (see Fig. Fig.4)4) cells. Flag-tagged proteins were visualized by using anti-Flag antibody combined with a fluorescein-conjugated secondary antibody, whereas nuclei were visualized by staining DNA with DAPI. WT Flag-yNap1 is predominantly localized in the cytoplasm (Fig. (Fig.2a2a and e). In contrast, Flag-yNap1ΔNES and Flag-yNap1 (L99A L102A) are predominantly localized in the nucleus (Fig. (Fig.2b,2b, c, d, f, g, and h). We could not see expression of Flag-yNap1 (L88A V92A) for unknown reasons (data not shown). These results show that the NES-like sequence is functional and responsible for the export of yNap1 to the cytoplasm. Therefore, it is quite possible that the cytoplasmic localization of yNap1 is not static and yNap1 is dynamically shuttled between the nucleus and the cytoplasm.
An interesting question is raised as to how Nap1 is imported into the nucleus. Several lines of the evidence suggest that histones could be involved in the nuclear import of Nap1. First, Nap1 interacts with H2A/H2B dimers in the cytoplasm (7, 16). Second, the N-terminal regions of histones act as nuclear localization signals (NLS) (39). Third, ectopically expressed NAP-2, one of the Nap1 family proteins, is translocated from the cytoplasm to the nucleus at the S phase, when histones are most actively synthesized in the cytoplasm and imported into the nucleus (47). Fourth, histone H2A/H2B forms a ternary complex with Kap114 and Nap1, and recently, it was shown that Kap114 is involved in the nuclear import of Nap1 (38). We therefore examined a possible relationship between histone binding and nuclear import of yNap1. We constructed vectors for the expression of the yNap1 mutant proteins yNap1ΔC1 and yNap1ΔNESΔC1, both of which lack the carboxy-terminal region of WT yNap1 and yNap1ΔNES (Fig. (Fig.1B),1B), respectively. The deleted region contains a highly acidic stretch and a small region homologous to TAF-I (24). As previously reported, yNap1ΔC1 has no stimulatory activity in nucleosome formation in vitro (11). Since histone binding is one of the most basic activities for histone chaperones and the binding ability is likely to depend on the acidity of histone chaperones, it is important to examine whether yNap1ΔC1 binds to histones. To this end, we performed GST pull-down assays with GST-Nap1 proteins and core histones purified from HeLa cells. As shown in Fig. Fig.3,3, WT yNap1 and yNap1ΔNES bind to core histones (Fig. (Fig.3,3, lanes 7 and 8). The histone-binding activity of yNap1ΔNES is approximately 80% of that of WT yNap1. It is clear that the amount of core histones binding to yNap1ΔC1 is less than that of WT yNap1 (Fig. (Fig.3,3, lanes 7 and 9). The ability of yNap1ΔC1 to bind to H2A/H2B is almost completely lost. Therefore, it is concluded that the C-terminal region is involved in H2A/H2B binding.
We next examined the involvement of the C-terminal region of yNap1 in its nuclear import. Flag-yNap1ΔNES is clearly localized in the nucleus (Fig. (Fig.2b,2b, c, f, and g). It is assumed that the subcellular localization could be changed by truncation of the C-terminal region from Flag-yNAP1ΔNES. The ectopically expressed Flag-yNap1ΔC1 and Flag-yNap1ΔNESΔC1 are localized in the cytoplasm in HeLa cells (Fig. (Fig.2i,2i, j, k, and l). Therefore, it is likely that the C-terminal region that is important for H2A/H2B binding is involved in the nuclear import of Nap1. To more precisely examine the subcellular localization of yNap1, we employed a novel microscope system, termed laser scanning cytometer (LSC2; Olympus) and measured the fluorescent intensity of the nucleus and of the peripheral area of the nucleus. To minimize the background staining, we used EGFP-tagged yNap1's for transient expression in HeLa cells. As shown in Table Table1,1, a majority of cells (89.1%) expressing EGFP-yNap1 showed the dominant EGFP signals in the cytoplasm. In 10.1% of the cells, EGFP signals were dominant in the nucleus. In sharp contrast, among cells expressing EGFP-yNap1ΔC1, we could not observe a cell in which the nuclear fluorescence is brighter than the cytoplasmic fluorescence. Thus, it is concluded that WT yNap1 is dominantly localized in the nucleus and that the C-terminal H2A/H2B binding region of yNap1 is important for the nuclear localization of yNap1. EGFP-yNap1ΔNES is a nuclear dominant in 89.2% of total cells expressing EGFP-yNap1ΔNES. This confirms that the NES-like sequence is involved in exclusion of yNap1 from the nucleus. Among cells expressing EGFP-yNap1ΔNESΔC1, we could not see the cell that is dominantly stained in the nucleus (data not shown). Again, the C-terminal region is likely to be essential for the nuclear localization of yNap1.
We tried to examine the physiological role of nucleocytoplasmic transport of yNap1 by genetic complementation assays with the DK168 strain (ΔNAP1 ΔCLB1 ΔCLB3 ΔCLB4). DK168 shows a temperature-sensitive growth and an elongated bud phenotype below the permissive temperature (25, 26). Before carrying out complementation assays, we first examined the subcellular localization of Flag-yNap1 mutant proteins in yeast DK168 cells (Fig. (Fig.4A).4A). Flag-yNap1 is dominantly localized in the cytoplasm, whereas Flag-yNap1ΔNES and Flag-yNap1 (L99A L102A) are localized dominantly in the nucleus (Fig. (Fig.4A).4A). Therefore, it is confirmed that the mechanism of nuclear import and export of yNap1 in yeast cells are similar to that in mammalian cells.
It is suggested that the elongated bud phenotype of DK168 is caused by the prolonged mitotic delay followed by the inability of Gin4 kinase phosphorylation (3). Next, we performed complementation experiments with various yNap1 derivatives by using the DK168 strain. WT Flag-yNap1 complements the elongated bud phenotype (Fig. (Fig.5A,5A, b) and the temperature-sensitive (ts) phenotype (Fig. (Fig.5B).5B). In sharp contrast, Flag-yNap1ΔNES and Flag-yNap1 (L99A L102A) cannot complement the elongated bud phenotype (Fig. (Fig.5A,5A, c and d) and the ts phenotype (Fig. (Fig.5B).5B). The artificial mutations sometimes affect protein stability so that it is possible that yNap1ΔNES and yNap1 (L99A L102A) could not complement the phenotypes due to its instability and low abundance. Western blot analyses of lysates with anti-Flag antibody have revealed that Flag-yNap1, Flag-yNap1ΔNES, and Flag-yNap1 (L99A L102A) driven by the native promoter are equally expressed (Fig. (Fig.4B,4B, lanes 2 to 4). Therefore, cytoplasmic localization and/or nucleocytoplasmic shuttling of yNap1 is required for the putative mitotic function of yNap1. To distinguish these possibilities, we performed further complementation experiments with Flag-yNap1ΔC1 and Flag-yNap1ΔNESΔC1, which are predominantly localized in the cytoplasm (Fig. (Fig.2k2k and l). Flag-yNap1ΔC1 and Flag-yNap1ΔNESΔC1 do not complement the elongated bud phenotype (Fig. (Fig.5A,5A, e and f). As these mutant proteins are localized in the cytoplasm and the C-terminal region is important for the nuclear import of Nap1, it is likely that nucleocytoplasmic shuttling, but not cytoplasmic localization, of yNap1 is important for the mitotic progression.
Nucleocytoplasmic shuttling could be divided into two steps, that is, nuclear import and export. To elucidate the molecular mechanisms of the role of Nap1, it is necessary to clarify whether either one of the two steps or both steps are required for the mitotic function of Nap1. As NES-deficient Nap1's do not complement the mitotic function (Fig. (Fig.5),5), it is likely that the nuclear export of Nap1 is essential for its function. However, it is possible that the loss of the NES function reduces the level of nuclear import of Nap1, possibly due to the decrease of recycling of Nap1 for repeated nuclear import processes by the loss of its nuclear export. To test this, we performed overexpression experiments of NES-deficient yNap1's. Since a part of yNap1ΔNES is localized in the cytoplasm, it is possible that overexpression of yNap1ΔNES may result in the increase of the cytoplasmic portion of yNap1ΔNES. Thus, the increased amount of NES-deficient yNap1 may overcome the reduced nuclear import. Overexpression of yNap1 and its derivatives was performed by using a galactose-inducible multicopy plasmid, pYES2 (Fig. (Fig.6A).6A). The expression level of each galactose-inducible derivative was examined in comparison with that derived from a single-copy plasmid containing the native promoter by Western blot analysis with anti-Flag antibody (Fig. (Fig.4B,4B, lanes 5 to 10, and data not shown). It is shown that the expression level in the galactose-inducible derivatives tested here was at least five times or more than that derived from the native promoter. Under these conditions, the growth rate of each strain was observed. Overexpression of yNap1ΔNES and yNap1 (L99A L102A) could not complement the ts and elongated bud phenotypes (data not shown). To be more precise, the slight growth inhibition by overexpression of yNap1ΔNES or yNap1 (L99A L102A) was observed even at the permissive temperature (Fig. (Fig.6A).6A). yNap1ΔNES appears to inhibit cell growth more potently than yNap1 (L99A L102A). Possibly, yNap1 (L99A L102A), but not yNap1ΔNES, may retain a weak NES function, although they both show similar subcellular localizations (Fig. (Fig.22 and and4)4) and both are unable to complement the WT function (Fig. (Fig.5).5). These results show that the increased nuclear import of NES-deficient Nap1 is still insufficient for mitotic function of Nap1. Furthermore, the nuclear accumulations of NES-deficient Nap1's have a trans dominant-negative effect for mitotic progression. Therefore, it is strongly suggested that the nuclear export of Nap1 is essential for the mitotic function of Nap1.
We further examined the effect of the balance between nuclear import and export on the Nap1 function. As Flag-yNap1ΔC1 did not complement the mitotic function of Nap1 and the C-terminal region was shown essential for nuclear import of Nap1, it is possible that not only nuclear export but also nuclear import is important for the Nap1 function. For instance, it is possible that the loss of nuclear import causes reduced nuclear export, since nuclear import is the primary step for nuclear export. To know the effect of nuclear import of Nap1 on the Nap1 function, we examined a genetic interaction between NAP1 and KAP114, which is involved in the nuclear import of Nap1 (38) and other proteins, including TATA-binding protein and histone H2A/2B (37, 46). It has been reported that disruption of either NAP1 or KAP114 does not show any cell growth defect (25, 26, 37, 46). The NAP1 and KAP114 double-gene-disrupted strain PSY1784Δnap1 shows slower growth than the KAP114-positive strain PSY1785Δnap1 (Fig. (Fig.6B).6B). The decreased growth rate caused by NAP1 gene disruption was partially complemented by the NAP1 gene with both low- and high-copy-number plasmids. These results indicate the genetic interaction between these two genes. Disruption of KAP114 shows the reduced level of nuclear import of some proteins, such as histone H2A/2B and TATA-binding protein (37, 46), so that the genetic interaction between NAP1 and KAP114 suggested that Nap1 has some functions in nuclear import which is important for the cell growth.
Clbs play key roles in mitotic progression (10). It has been noted that the physical and genetic interactions between Clb2 and yNap1 have been indicated (25, 26). As clb2ts mutant shows the elongated bud phenotype at nonpermissive temperatures, it is likely that Nap1 regulates some Clb2 function through a direct interaction. Thus, we examined the Clb2-binding ability of yNap1 mutant proteins by GST pull-down assays. Clb2 was prepared by in vitro transcription-translation reactions in the presence of [35S]methionine. GST-fused WT yNap1, but not GST alone, interacts with Clb2 in a dose-dependent manner (Fig. (Fig.7A,7A, lanes 2 to 7), which is in good agreement with the previous result (3). yNap1ΔNES and yNap1 (L99A L102A) also interact with Clb2 as efficiently as WT yNap1 (Fig. (Fig.7,7, lanes 8 to 13). Therefore, the mutation in the NES sequence does not affect the interaction ability of Nap1 to Clb2.
Next, we examined the in vivo Clb2 interacting abilities of WT yNap1, yNap1ΔNES, and yNap1 (L99A L102A) by coimmunoprecipitation assays (Fig. (Fig.7C).7C). Flag-tagged WT yNap1, yNap1ΔNES, and yNap1 (L99A L102A) proteins were expressed in DK168 strains by using the Gal1 promoter. Flag-tagged WT yNap1, yNap1ΔNES, and yNap1 (L99A L102A) proteins were precipitated by anti-Flag antibody (Fig. (Fig.7,7, upper panel). Clb2 was coprecipitated with WT Flag-yNap1, Flag-yNap1ΔNES and Flag-yNap1 (L99A L102A) (Fig. (Fig.7,7, lower panel). Therefore, the interaction between Nap1 and Clb2 is not dependent on either the NES-like sequence or the subcellular localization.
A question is raised whether the mitotic function of yNap1 is correlated with its nucleosome assembly activity. We next examined the nucleosome assembly activity of NES-mutated yNap1 proteins (Fig. (Fig.8B)8B) by using purified recombinant GST-yNap1 proteins (Fig. (Fig.8A).8A). In the supercoiling assay, a protein containing the nucleosome assembly activity facilitates nucleosome formation with core histones and a relaxed circular plasmid DNA in the presence of topoisomerase I. Removal of proteins after the reaction gives a supercoiled DNA. A relaxed DNA (Fig. (Fig.8B,8B, lane 9) generated by treatment of a supercoiled DNA with E. coli topoisomerase I (Fig. (Fig.8B,8B, lanes 1 and 2) was mixed with core histones that had been preincubated with yNap1 proteins. The supercoiled DNA was formed by GST-yNap1 in a dose-dependent manner (Fig. (Fig.8B,8B, lanes 3 and 4). GST-yNap1 (L99A L102A) has the same level of nucleosome assembly activity as GST-yNap1 (Fig. (Fig.8B,8B, lanes 3, 4, 7, and 8). Therefore, the substitution mutations in the NES region of yNap1 do not affect the nucleosome assembly activity in vitro. GST-yNap1ΔNES has less, but significant, activity than WT GST-yNap1 (Fig. (Fig.8B,8B, lanes 3, 4, 5, and 6). Densitometric scanning revealed that the activity of GST-yNap1ΔNES is approximately 50% of that of WT GST-yNap1 (Fig. (Fig.8B,8B, lanes 4 and 5). The low nucleosome assembly activity of yNap1ΔNES could be due to its low histone-binding ability, as shown in Fig. Fig.2.2. We could not find any strict correlation between the mitotic complementation ability and the nucleosome assembly activity.
Our analyses described in this paper started with a question of whether Nap1 is stably localized in the cytoplasm because the nucleosome assembly activity should be a nuclear function. By examining the subcellular localization and biochemical activities of mutant Nap1 proteins, we could propose a novel regulatory role of Nap1 in mitotic progression through its nucleocytoplasmic shuttling. We have shown that yNap1 is localized in the nucleus when the NES-like sequence is deleted or mutated (Fig. (Fig.22 and and4A).4A). The histone binding was found to be important for nuclear import of yNap1 (Fig. (Fig.22 and and3).3). Since NES-deficient Nap1's had no complementing activity for mitotic function of Nap1 (Fig. (Fig.5),5), it is suggested that nucleocytoplasmic shuttling of Nap1 is required for its function. Overexpression of NES-deficient Nap1's did not complement the mitotic function of Nap1 (Fig. (Fig.6).6). Therefore, it is concluded that nuclear import of Nap1 is insufficient and the nuclear export is essential for Nap1 function.
As yNap1 is predominantly localized in the cytoplasm (26) (Fig. (Fig.4A)4A) and Gin4, which is one of the Nap1 interactors which plays a critical role in mitosis, is localized in the bud neck, it is reasonably assumed that Nap1 acts as part of the mitotic function in the cytoplasm. However, our results suggest that nucleocytoplasmic shuttling, but not stable cytoplasmic localization, of Nap1 is required for the mitotic function of Nap1. First, C-terminal deletion mutants (Flag-yNap1ΔC1 and Flag-yNap1ΔNESΔC1), which are not effectively imported into the nucleus, thereby staying in the cytoplasm, have no complementing activity (Fig. (Fig.5A).5A). Second, overexpression of NES mutant proteins is still insufficient to complement the DK168 phenotypes (Fig. (Fig.6A).6A). Ectopically expressed NES-deficient yNap1's are predominantly localized in the nucleus, but a significant level of cytoplasmic staining is also observed (Fig. (Fig.4A).4A). Therefore, the cytoplasmic amount of the mutant yNap1 could be increased, as the total amount of mutant yNap1 proteins is increased by its overexpression. These observations strongly suggest that the constitutive cytoplasmic existence of Nap1 is not enough and that nucleocytoplasmic shuttling is required for the mitotic function of Nap1.
As discussed above, the nuclear export of Nap1 is important for the mitotic function of Nap1. However, we could not rule out the possibility that nuclear import of Nap1 is also critical because nuclear import is the primary step for nuclear export. We addressed this point by using KAP114 mutants. Kap114 physically interacts with Nap1 and is involved in the nuclear import of TATA-binding protein (37, 46) and histone H2A/2B (38). Kap114 and Nap1 make a complex with H2A/2B (38). The nuclear import of Nap1 is shown to be dependent on its histone-binding domain (Fig. (Fig.2).2). Therefore, it is likely that Nap1 assists the Kap114 function in nuclear import of some proteins as a nuclear import partner. Indeed, a double gene disruption (kap114 and nap1) resulted in a decreased growth rate phenotype (Fig. (Fig.6B),6B), although a single gene disruption of either one of these two genes did not give any growth defect. This synthetic phenotype between two genes suggests that Nap1 may be involved in nuclear import of some nuclear proteins.
Experimental results strongly suggest that the nuclear export of Nap1 is essential for its mitotic function. Overexpression of NES-deficient Nap1s caused the cell growth inhibition (Fig. (Fig.6A).6A). Such growth inhibition was not observed when they were expressed from the authentic yNAP1 promoter on a single-copy plasmid (Fig. (Fig.5B).5B). A yNap1 itself may be inhibitory when stably localized in the nucleus. On this line, it is noted that NAP1 has the chromatin disruption activity (24) as well as the nucleosome assembly activity (36, 54). The cell cycle is likely to be inhibited at the mitotic phase because nuclei of these cells were localized at the bud neck (data not shown). Such inhibition was not observed in NAP1-positive strains, a WT YPH499 strain, and a DK166 strain (ΔCLB1 ΔCLB3 ΔCLB4) (data not shown). Under the same conditions, yNap1-NLS, which is a fusion yNap1 protein with simian virus 40 NLS at the C terminus of yNap1 and predominantly localized in nucleus, could complement the DK168 phenotypes and did not have any inhibitory effect on the cell growth (data not shown). Thus, the inhibitory effect is suppressed by endogenous Nap1. These observations suggest that nuclear yNap1 does not have a mitotic cell arrest ability, but an export-deficient Nap1 that accumulates in the nucleus provides the inhibitory effect on cell growth.
An alternative explanation for the growth inhibition would be that the yNap1 mutants in the nucleus may anchor a yNap1 interactor(s) that exerts its normal function transiently in the nucleus and is needed to be exported to the cytoplasm. The target of Nap1 during its nuclear export is an open question. Several lines of evidence suggest that Clb2 is one of the candidates. First, Clb2 is also a nucleocytoplasmic shuttling protein and is dominantly localized in the nucleus, in particular during the G2/M phase (12). Second, an NES mutation of Clb2 causes an elongated bud phenotype (12). Therefore, we examined the difference of Clb2 localization in the absence and presence of NAP1. We confirmed that Nap1 and NES-mutated Nap1s bind to Clb2 (Fig. (Fig.7).7). We tried to detect the difference of Clb2-EGFP localization in NAP1-positive and -negative strains. However, the difference is not so significant between two strains.
Although systematic analyses are needed, we would like to propose the term shuttling chaperone for Nap1. The shuttling chaperone may modulate nuclear import and/or export of proteins rather than itself. A variety of Nap1-interacting proteins are being identified, and their intracellular localization patterns are closely related to the regulation of mitotic progression. It is possible that these interacting proteins are also targets of the Nap1 shuttling chaperone activity. Thus, it is interesting and necessary to examine the involvement of Nap1 in the subcellular localization of these interacting proteins and to ask whether their aberrant localization eliminates the proper mitotic progression.
We have discussed the function of Nap1 associated with its nucleocytoplasmic shuttling. It is interesting how the shuttling is regulated. It has long been thought that Nap1 is localized in the cytoplasm in a variety of species. Since the NES-like sequence in yNap1 is well conserved among species, the conclusion from our analyses with yeast Nap1 is probably applicable for other species. It is reported that the fly Nap1 is temporally localized in the nucleus at some stages during development (17). We have shown that EGFP-Nap1 is dominantly detected in cytoplasm, whereas in 10.1% of cells expressing EGFP-yNap1, the signal is dominantly detected in nucleus (Table (Table1).1). These observations point out that Nap1 could be localized in the nucleus under a certain condition(s). It is possible that the rate of nuclear import and export is varied depending on cell conditions. Based on the histone-binding activity and nucleosome assembly activity of Nap1, it is assumed that Nap1 functions as a chaperone for histones during nuclear import and in the cytoplasm as well as at the deposition to DNA. Thus, the rate of the nuclear import of Nap1 might be increased at the S phase, since histones are most actively synthesized at the S phase. The modification may also be involved in the regulation of nuclear transport and nuclear function, if any, of Nap1. It is reported that a cytoplasmic fraction of Nap1, but not the nuclear fraction, is possibly phosphorylated by CK2 and that the histone-binding activity is stimulated by phosphorylation of Nap1 by CK2 (48).
The balance of nuclear import and export of Nap1 may be important for the nuclear function of Nap1. Nap1 could be involved in not only de novo assembly of nucleosomes but also structural change of chromatin, although we do not know an exact function of Nap1 in the nucleus. Nap1 family proteins are shown to be involved in transcription regulation (20, 24, 49). Our preliminary experiments with one-hybrid assays indicated that Nap1 stimulates the transcription from chromatin under a certain condition (data not shown). To consider the nuclear function of Nap1, it is worthwhile to note that the nuclear localizations of NES-mutated yNap1s show speckled patterns in the nuclei (Fig. (Fig.2g,2g, h, and j). Determination of the nuclear domain may support understanding of the nuclear function of Nap1.
We thank Doug Kellogg for providing yeast strains and helpful discussion. kap114Δ strains were kindly provided by Steve Buratowski and Pamela A. Silver. We thank Yoshimasa Kiyomatsu (Olympus ProMarketing, Inc.) for permitting us to use the laser scanning cytometer and for technical support. We thank Yuki Yamaguchi for critical comments on the manuscript.
This work was supported by a grant-in-aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan and by a grant from the Bioarchitect Research Project of RIKEN.