Nuclear Import of Chromatin Remodeler Isw1 Is Mediated by Atypical Bipartite cNLS and Classical Import Pathway - Vasicova - 2012 - Traffic - Wiley Online Library
Advertisement
The protein Isw1 of Saccharomyces cerevisiae is an imitation-switch chromatin-remodeling factor. We studied the mechanisms of its nuclear import and found that the nuclear localization signal (NLS) mediating the transport of Isw1 into the nucleus is located at the end of the C-terminus of the protein (aa). We show that it is an atypical bipartite signal with an unconventional linker of 19 aa (KRIR X19 KKAK) and the only nuclear targeting signal within the Isw1 molecule. The efficiency of Isw1 nuclear import was found to be modulated by changes to the amino acid composition in the vicinity of the KRIR motif, but not by the linker length. Live-cell imaging of various karyopherin mutants and in vitro binding assays of Isw1NLS to importin-& revealed that the nuclear translocation of Isw1 is mediated by the classical import pathway. Analogous motifs to Isw1NLS are highly conserved in Isw1 homologues of other yeast species, and putative bipartite cNLS were identified in silico at the end of the C-termini of imitation switch (ISWI) proteins from higher eukaryotes. We suggest that the C-termini of the ISWI family proteins play an important role in their nuclear import.In eukaryotic organisms, DNA is organized into a compact and highly dynamic chromatin structure that affects processes dependent on DNA accessibility such as gene expression, DNA replication and chromosome segregation. Several mechanisms exist to alter the structure of chromatin, and ATP-dependent chromatin remodeling is one of them. The group of ATP-dependent chromatin remodelers is comprised of four main families: SWI/SNF, ISWI, CHD and INO80. The Isw1 protein of Saccharomyces cerevisiae belongs to the group of imitation switch (ISWI) proteins. ISWI proteins are evolutionarily conserved proteins that form multi-subunit complexes []. Isw1 was found to form two distinct complexes. The Isw1a complex, formed of the proteins Isw1 and Ioc3, is involved in the repression of transcription initiation []. Its absence reduces the stability of the silent state at the yeast silent mating HMR locus [] and it is also required for transcriptional silencing at the HML locus []. The Isw1b complex formed by the Isw1 and Ioc2 and Ioc4 subunits regulates transcription elongation, coordinates elongation with termination and pre-mRNA processing []. Isw1 also has a role in preventing cryptic initiation within genes [] and functions in parallel with the NuA4 and Swr1 complexes in the repression of stress-induced genes []. ISWI proteins possess an ATPase activity that enables the remodeling of chromatin by sliding, displacing or positioning nucleosomes. Next to the well-conserved ATPase domain at the N-terminus, ISWI proteins contain a unique combination of the HAND, SANT and SLIDE domains. Structural studies of the Drosophila melanogaster ISWI protein [] led to the identification of SANT and SLIDE as substrate recognition domains. Homologues of these domains are critical for in vivo functions of S. cerevisiae Isw1, in particular the SANT and SLIDE domains are necessary for the association of Isw1 with chromatin and for interactions with subunits within the Isw1 complexes. Both domains are involved in Ioc2 and Ioc4 binding, whereas the contact between Ioc3 and Isw1 is ensured by the SLIDE domain alone []. Recently, Yamada and coworkers [] crystallized a part of the yeast Isw1 protein composed of the HAND, SANT and SLIDE domains with Ioc3, thus confirming their interaction.The protein Isw1 is a large protein of 131&kDa that exerts its functions in the nucleus and thus needs to enter the nucleus via an active process through nuclear pores like other molecules larger than 40&kDa. Nuclear pores are formed by highly conserved nuclear pore complexes (NPCs) consisting of nucleoporins []. Cargo molecules are transported through nuclear pores by transport molecules & karyopherins. The yeast genome encodes at least 14 proteins of the karyopherin-& family and one adaptor protein, importin-&. Each &-karyopherin can mediate distinct import or export events by direct binding to a nuclear localization signal (NLS) or a nuclear export sequence (NES) on its substrate molecule (cargo). Some NLSs are recognized and translocated into the nucleus directly by karyopherin-&s (importin-&s) []. However, the best-characterized means of nuclear import is the classical import pathway, where a cargo carrying a classical NLS (cNLS) is bound by the adaptor protein importin-&/Srp1/Kap60 and is directed into the nucleus in complex with importin-&/Rsl1/Kap95. The cargo/importin-&/importin-& ternary complex is disassembled in the nucleus upon binding RanGTP to importin-& and interaction of the liberated IBB domain, nucleoporins (yeast Nup2, mammalian Nup50, Nup153) and exportin Cse1 with importin-& []. Both importins are subsequently recycled to the cytoplasm and become available for another round of transport.Importin-& is composed of two structurally and functionally distinct domains. The N-terminal domain, called the importin-& binding (IBB) domain, contains stretches of basic amino acid residues enabling the interaction of importin-& with importin-& that ensures translocation of the cargo/importin-&/importin-& complex into the nucleus. When the IBB domain is not associated with importin-&, it has an autoinhibitory role and competes with the cargo cNLS for the cNLS binding site in the C-terminal domain of importin-& [reviewed in []]. The C-terminal domain consists of 10 tandem armadillo (ARM) repeats forming major and minor grooves that are responsible for cNLS binding [reviewed in []]. The cNLSs recognized by importin-&, which are the most studied nuclear targeting sequences, exhibit monopartite or bipartite patterns. The monopartite cNLSs, [e.g. the SV40 large T antigen with the PKKKRK motif, []], typically contain a short stretch of basic amino acid residues and can bind to either of the two binding sites on importin-&. Although, the major site is considered to be the high-affinity binding site for typical monopartite cNLSs [], short monopartite NLSs have been recently identified that only interact with the minor groove []. The bipartite cNLS contain two clusters of basic residues separated by a short linker, for example the Xenopus nucleoplasmin NLS (KRPAAATKKAGQAKKKK) []. The basic N- and C-terminal elements of bipartite NLSs interact with the minor and major binding pockets, respectively []. It was long thought that the typical length of the amino acid spacer between the two basic parts of a bipartite NLS is 10&12 aa []. However, evidence is mounting that length of the linker region between the two parts of a bipartite signal can vary considerably (8 to &30 aa). The composition of the spacer influences the efficiency of the nuclear transport, most likely by affecting the cargo affinity and binding to importin-& []. These atypical bipartite NLSs were identified in several organisms including yeasts [], plants [], mice [] and humans [].In this work, we analyzed the mechanisms of the nuclear transport of the yeast chromatin remodeling factor Isw1. We provide evidence that the NLS is located at the end of the C-terminus of Isw1. It is an atypical bipartite cNLS with an extended linker of 19 aa that is recognized by importin-&, mediating its transfer to the nucleus via the classical import pathway. The efficiency of Isw1NLS nuclear import was found to be affected by the residues adjacent to the N-terminal part of the NLS but not by its atypical linker length. There are some indications that the nuclear localization of Isw1 might be regulated upon mild heat stress. In silico, we identified motifs analogous to Isw1NLS in Isw1 homologues of other yeast species and putative bipartite cNLSs at the ends of the C-termini of ISWI proteins from higher eukaryotes.ResultsNLS is located at the end of C-terminus of Isw1To understand the mechanisms of the nuclear translocation of Isw1 in S. cerevisiae, we first analyzed the Isw1 amino acid sequence using available predictive tools designed to identify NLS signals. These analyses found most of the putative NLSs to be within the first 200 N-terminal amino acids and did not identify any NLS in the Isw1 fragment that Pinskaya and coworkers [] detected in the nucleus (Table&). To determine the correct position of NLS and evaluate the significance of all predicted NLSs, we constructed a series of plasmids carrying Isw1 truncations fused to GFP (Figure&A). Expression of these ISW1-deletion variants (Table ) was controlled by the GAL1 promoter. Strains with the endogenous ISW1 deleted (isw1&D) and carrying constructed Isw1-GFP chimeras were analyzed by live-cell imaging. As shown in Figure&B, the fragments Isw1(1&200)-GFP and Isw1(1&1002)-GFP containing in silico-predicted putative NLSs, and Isw1(1&1060) carrying the predicted NLSs and the intact SLIDE domain (aa988&1052), were found in the cytoplasm, whereas the remaining C-terminal fragment Isw1()-GFP localized exclusively to the nucleus. This result suggested that the NLS is located within these last 69 aa residues. To exclude the possibility that because of its small size (35&kDa) Isw1()-GFP accumulated in the nucleus by passive diffusion through the nuclear pore followed by irreversible binding and retention in the nucleus, we fused the same Isw1 fragment (aa) with two GFP molecules so that the resulting size (62&kDa) of the Isw1NLS-GFP2 fusion was above the assumed limit for passive diffusion. We found that the GFP signal only accumulated in the nucleus (Figure&C). To further confirm the Isw1() fragment as the NLS-bearing region, we cloned this sequence upstream of the Escherichia coli lacZ gene into the plasmid pYC2-CTlacZ. Immunofluorescence microscopy proved that the signal found within aa of Isw1 was strong enough to mediate the nuclear translocation of &-galactosidase as an example of a heterologous cytoplasmic protein (Figure&C).Table&1.&In silico-predicted NLSs of Isw1PredictNLS []No NLS predictedPSORTII []Monopartite&20 & RKKH&83 & PESNKKR&157 & RRRK&766 & PTKRERKcNLS Mapper []Bipartite&84 & ESNKKRYLLKDANGKKFDLEGTTKRFEH&103 & EGTTKRFEHLLSLSGLFKHFIESKAAKDPKFFigure&1. NLS of Isw1 is located within C-terminal fragment aa. A) Schematic presentation of the full-length Isw1 and its deletion variants fused to GFP or lacZ. Their observed localization, either nuclear (N) or cytoplasmic (C) together with calculated N/C ratios are indicated (ND & not determined). The Isw1 domain's SNF2-N-ATP binding (aa199&480), Helicase-C (HC; aa541&616), HAND (H; aa776&883), SANT (S; aa884&933), SLIDE (SL; 940&1058) and in silico-predicted NLSs (black boxes) are indicated. B) Localization of Isw1 and its truncated variants fused with GFP in the BY isw1&D strain was examined by live-cell microscopy 2&3&h after induction with 2% galactose. C) The truncation Isw1() containing Isw1-NLS was fused to tandem GFP. Its expression in the BY isw1&D strain was driven by the ADH1 promoter and exponentially growing cells were observed by live-cell microscopy to map the GFP signal. The localization of Isw1NLS fused with Escherichia coli lacZ (pGalIsw1()-lacZ) was examined by indirect immunofluorescence using a monoclonal antibody against &-galactosidase. The nucleus was visualized by DAPI staining. Scale bar: 4&&m.Deletion analysis thus revealed that the nuclear targeting signal is located within the last 69 aa of the Isw1 molecule in the vicinity of the protein- and DNA-binding domain SLIDE.Isw1NLS is atypical bipartite signal with extended linkerThe canonical NLSs reported so far are usually stretches of basic amino acids. The C-terminal Isw1() fragment contains three regions rich in basic residues, KDRMKK (aa, basic region 1, BR1), KRIR (aa, BR2) and KKAK (aa, BR3). Further Isw1 deletion variants were constructed to unambiguously position Isw1NLS (Figure&A). The nearly full-length protein, Isw1(1&1105), containing all three basic regions, was found exclusively in the nucleus (Figure&B). The amino acid residues
were therefore not required for Isw1 nuclear targeting. However, the shorter fragment Isw1(1&1090), carrying the first two basic regions, BR1 and BR2, did not enter the nucleus and the GFP signal was only observed in the cytoplasm. The same held for the fragments Isw1(1&1080) containing BR1 and a part of BR2, and Isw1() carrying the second half of BR2 and the intact BR3. This suggested that the middle stretch of basic residue BR2 was indispensable but not sufficient for the nuclear localization of Isw1, and that the third motif was also essential. The last construct, pIsw1()-GFP2 carrying BR2 and BR3, was only detected in the nucleus, thus confirming that Isw1NLS is located between amino acid residues 1079 and 1105. The motifs KRIR and KKAK may represent a bipartite NLS, because both of them are indispensable for Isw1 nuclear targeting.Figure&2. Isw1NLS is bipartite signal with extended linker. A) Schematic representation of Isw1's basic regions (BR1, BR2 and BR3, underlined) and deletion derivatives fused to single or double GFP. The observed localizations of the GFP chimeras, N & nucleus, C & cytoplasm and its calculated ratios are indicated. B) Expression of Isw1 deletion derivatives fused to single GFP molecule was induced by 2% galactose (2&3&h) and that of Isw1()-GFP2 was constitutive from the ADH1 promoter. The localization of the constructs in the BY isw1&D cells was observed by live-cell microscopy. C) Live-cell imaging with DSU of BY isw1&D cells carrying either the intact or mutated variants in the KR (K1079A, R1080A) and the KKAK (K1102A, K1103A and K1015A) motifs, constructed in the context of the short Isw1()-GFP2 fragment or the full-length protein. Expression of the full-length Isw1-GFP, intact or mutant variants, was regulated by the GAL1 promoter in the BY isw1&D cells and localization of the GFP signal was followed after 4&h of 2% galactose induction at 30&C. The GFP signal is shown in the upper panel, the corresponding staining of nuclei with DAPI in the middle and the combination of both signals (GFP & green, DAPI & blue) in the bottom row. The calculated N/C ratios are indicated. Scale bar: 4&&m.To confirm that Isw1 really contains a bipartite NLS formed by the BR2 and BR3 regions, we performed site-directed mutagenesis. In BR2, the basic residues typical of a bipartite NLS KR were replaced with alanine residues, as were all lysines in BR3. As shown in Figure&C, the mutations K1079A and R1080A resulted in the accumulation of the GFP signal in the cytoplasm. The same held for the substitution of aa KKAK with AAAA. The N/C ratios calculated for both mutated NLS variants, 0.9&&&0.2 for KR substitution and 0.8&&&0.1 for KKAK substitution, clearly confirmed a defect in their nuclear import. Because both regions were found to be vital for NLS functionality, we conclude that Isw1NLS is a bipartite signal with an extended linker.To determine whether the nuclear import of the full-length Isw1 protein is also dependent on the identified bipartite NLS, we replaced either the KR or KKAK motifs of Isw1NLS with AA or AAAA, respectively, and reconstituted the full-length Isw1-GFP fusion protein expressed from the GAL1 promoter (pGalIsw1(1079AA)-GFP and pGalIsw1(1102AAAA)-GFP). As shown in Figure&D, the full-length Isw1-GFP carrying the substitution 1079KR&AA was strongly accumulated in the cytoplasm, although a slight signal from the location of the nucleus was also observed, as was indicated by the calculated N/C ratio 1.2&&&0.2. The second mutated variant of the full-length Isw1-GFP carrying the 1102KKAK&AAAA substitution was completely mislocalized to the cytoplasm, as was also shown by the N/C ratio 0.8&&&0.1. Taken together, these results demonstrated that the bipartite NLS discovered in the context of the truncated form of Isw1, Isw1(aa), was also indispensable for the efficient nuclear import of the full-length Isw1 protein.Putative NLS motifs were identified at C-termini of Isw1 yeast homologuesTo determine whether a pattern similar to Isw1NLS can be found in the C-terminal parts of Isw1 homologues from other yeast species, we performed BLAST analyses and multiple sequence alignments using ClustalW. These analyses revealed that the homology of most yeast Isw1 proteins decreases after the C-terminus of the SLIDE domain. Despite their heterogeneity, the C-termini of analyzed yeast Isw1-like proteins carry some conserved motifs (Figure&A). The first motif, KxRM/LxE/KED, is located immediately adjacent to the SLIDE domain and overlaps a putative SUMO site in S. cerevisiae (MKKE) and Ashbya gossypii (MKEE). However, our deletion analysis showed that this region (BR1) is not required for the nuclear import of Isw1. The second motif, KRxREE, conserved in all analyzed proteins, includes the N-terminal part of the bipartite Isw1NLS. The C-terminal part of Isw1NLS, the KKAK motif, is not conserved in however a stretch of basic amino acid residues was found in all examined Isw1-like sequences at various distances (18&64 aa) from the KRxREE motif. Furthermore, these non-conserved basic motifs also overlap a putative SUMO site in the Isw1 of S. cerevisiae (AKIE), Candida glabrata (PKIE), Lachancea thermotolerans (PKTE) and Zygosaccharomyces rouxii (VKLE).Figure&3. Yeast Isw1 homologues contain conserved KRxREE motif covering the N-terminal part of S. cerevisiae Isw1NLS. A) The C-termini of Iws1 yeast homologues were aligned using BLAST, ClustalW and manually. The yeast Isw1 sequences were as follows: S. cerevisiae & Saccharomyces cerevisiae (P38144.2) C. glabrata & Candida glabrata (CAG58154.1); V. polyspora & Vanderwaltozyma polyspora (EDO19437.1); L. thermotol. & Lachancea thermotolerans (KLTHOHO5566p); Z. rouxii & Zygosaccharomyces rouxii (ZYRO0G21780p); Ashbya gossypii & Ashbya gossypii (ALF040Wp). Conserved amino acids are labeled in red and C-terminal motifs of putative bipartite NLSs are underlined. B) Live-cell microscopy using DSU was used to assess the location of the intact KRIREE motif within Isw1()-GFP2 and its variants carrying the mutations KRIAAA (R1082A, E1083A, E1084A) and KRIRAA (E1083A, E1084A) constitutively expressed in the BY isw1&D cells. Scale bar: 4&&m. C) The mean fluorescence intensity was measured in the nucleus (N) and the cytoplasm (C) and the N/C ratio was calculated. The mean N/C ratios were calculated from more than 50 cells of each strain and their distribution curves were plotted. Quantification was done with three independent experiments.The conserved motif KRxREE, found in all the analyzed Isw1 homologues, has not yet been described. As demonstrated above, the first two residues from this motif, KR, resembling the N-terminal motif of the classical bipartite NLS, are important for the functioning of S. cerevisiae Isw1NLS. To determine whether the conserved amino acids REE in close proximity to the essential 1079KR motif also contribute to NLS functionality, we substituted the amino acids R1082A, E1083A and E1084A in pIsw1()-GFP2. Figure&B documents that the mutated Isw1NLS partially entered the nucleus, but a significant portion of the GFP signal remained in the cytoplasm. To analyze this microscopic observation quantitatively, we calculated the mean nuclear/cytoplasmic (N/C) ratio of the GFP-fluorescence intensity, reflecting the affinity of the NLS for the import receptor [] (Figure&B). The N/C ratio of the triple mutant dropped to 1.8&&&0.4 compared to the wild-type Isw1NLS N/C&=&7.0&&&3.2, pointing to a significant impact of the REE motif on Isw1NLS functionality. Further mutagenesis revealed that arginine R1082 is of major importance, because substitutions of the glutamic residues E1083A and E1084A alone only partially reduced the nuclear import of the Isw1NLS mutated fragment, as the mean N/C ratio decreased to 4.1&&&1.3 (Figure&B). We also plotted N/C ratio distribution curves for the cell population, which better reflect the statistical significance of changes in the N/C ratio []. The wild-type Isw1 protein exhibited a significantly wider distribution of N/C ratio in the cell population compared to both mutated forms, clearly indicating the importance of amino acids adjacent to the core N-terminal motif KR (Figure&C).On the basis of these results, we propose Isw1NLS to be an atypical bipartite NLS with an extended linker of 19 aa, KRIR X19 KKAK. Because putative bipartite NLSs with unconventionally long linkers were identified at the C-termini of all the analyzed Isw1 yeast homologues and the well-conserved motif KRxREE was proved to be very important for the nuclear import of the S. cerevisiae Isw1 protein, we suggest that C-termini play an important role in the nuclear import of yeast Isw1-like proteins.Integrity of extended linker within its bipartite NLS is not essential for Isw1 nuclear importAs the conventional length of a classical-bipartite NLS linker is proposed to span 10&12 aa according to the prototypical sequence of nucleoplasmin, we were further interested in whether the extended linker of 19 aa is essential for the nuclear import of Isw1. We therefore constructed deletions of 9 aa within the linker sequence from both its N- (aa) and C- (aa) terminal parts in the context of the short Isw1NLS-carrying fragment, pIsw1()-GFP2 and of the full-length Isw1 protein, pGalIsw1-GFP. In both deletions, the two important motifs, the N-terminal KRIREE and C-terminal KKAK together with the preceding serine (SKKAK) were kept, but the length of the linker was adjusted to contain 12 aa (KR-x12-KKAK) in total, similar to the prototypical bipartite NLS sequence of nucleoplasmin (Figure&A). The localization patterns of all deletion variants visible in Figure&B,C did not change markedly compared to that of the intact Isw1NLS. This is also evident from the calculated mean N/C ratios that in the short Isw1NLS fragment yielded highly variable values of 12.0&&&4.1 and 10&&&3.9 and a significantly wider N/C ratio distribution in the cell population than the intact Isw1NLS. As judged from the mean N/C ratios and from plotted N/C ratio distributions in the cell population, both versions of Isw1NLS&linker shortening resulted in the formation of NLSs with an enhanced nuclear distribution, pointing to a higher affinity toward the import receptor (importin-&) than was observed for the intact fragment pIsw1()-GFP2 (Figure&B). However, both deletions of the linker in the context of the full-length Isw1 protein did not have any apparent impact on the distribution of N/C ratios and thus on its affinity towards importin-& (Figure&C).Figure&4. Shortening Isw1NLS-linker length does not negatively affect NLS function. A) Schematic presentation of two types of linker deletions. B) Linker deletions constructed in the short Isw1NLS fragment were analyzed in the BYisw1&D strain to localize the GFP signal. The calculated N/C ratios and their distribution curves are shown. C) Localization of the GFP signal of linker deletions constructed in the full-length Isw1 (FL-Isw1) with indicated N/C ratios and their distribution curves. The GFP signal was followed after 4&h of 2% galactose induction at 30&C in the BY isw1&D cells. Scale bar: 4&&m.In summary, the shortening of Isw1NLS did not have any negative impact on the nuclear import of Isw1, either in the context of the short NLS fragment or in the full-length version, thus the reasons for such an atypical linker length within the bipartite Isw1NLS are not evident and can only be speculated upon.Isw1 enters nucleus via classical import pathwayDeletion analysis and site-directed mutagenesis revealed that Isw1 carries the bipartite NLS with a linker of atypical length. All bipartite sequences that have been experimentally determined so far were recognized and transported by the classical importin-&/& pathway, therefore we questioned whether Isw1 is also transported to the nucleus via this pathway.We compared the subcellular distribution of Isw1NLS carried by the plasmid pGalIsw1()-GFP in the temperature-sensitive mutants of importin-& (srp1-31, srp1-54), importin-& (rsl1-4) and the exportin of importin-& (cse1-1) to that of the wild-type W303 and another essential karyopherin mutant pse1-1. The expression of Isw1()-GFP was induced by galactose at both permissive and restrictive temperatures. At restrictive temperatures the appropriate karyopherins were inactivated prior to induction (see section Materials and Methods for conditions). In the strains W303 and pse1-1, Isw1()-GFP localized to the nucleus at both permissive and restrictive temperatures and was barely detectable in the cytoplasm (Figure&A). In rsl1-4, srp1-31, srp1-54 and cse1-1 mutant cells, a significant GFP signal was found in the cytoplasm even under permissive conditions, although a nuclear signal was also detected. No important difference in the GFP signal distribution was observed between permissive or restrictive temperatures, as demonstrated by the calculated N/C ratios shown in Figure&D. To prove that the Isw1NLS nuclear import in srp1-31, srp1-54 and cse1-1 cells was reduced because of mutations in importin-& and in its exportin Cse1, and not due to a general effect on nuclear import, we assessed the localization of the N-terminal (IBB) domain of importin-& (aa1&60) in these strains. The IBB domain of importin-& is imported to the nucleus by importin-&, thus its nuclear import is independent of importin-& and Cse1. We fused the IBB domain with GFP and placed it under the control of the GAL1 promoter (pGalIBB-GFP). The size of the resulting fusion protein was 34.5&kDa, nearly the same as that of Isw1()-GFP (35&kDa). As shown in Figure&B, IBB-GFP localized strictly to the nucleus in srp1-31, srp1-54 and cse1-1 cells, as represented by the high N/C ratios in Figure&D. From these results, we concluded that the cytosolic accumulation of Isw1()-GFP in the strains srp1-31, srp1-54 and cse1-1 was a consequence of mutations in importin-& and Cse1. To further support our hypothesis that the classical import pathway mediates the nuclear import of Isw1NLS, we assessed the localization of the Isw1()-GFP fusion protein in a strain lacking the nucleoporin Nup2 (nup2&D). Nup2 affects the release of a cargo from importin-& on the nuclear side of the nuclear pore and recycling of importin-& back into the cytoplasm []. As is obvious from Figure&C, a strong cytoplasmic GFP signal appeared in nup2&D cells and the calculated N/C ratio (2.9&&&0.8) markedly decreased compared to the control BY isw1&D strain (8.6&&&0.8), clearly pointing to an import defect.Figure&5. Isw1()-GFP fragment carrying Isw1NLS enters nucleus via classical import pathway. A) Expression of Isw1()-GFP was induced by 2% galactose in all indicated strains. The permissive temperature was 25&C, with the exception of the cse1-1 strain, which was grown at 30&C. At restrictive temperature (37&C, 15&C for cse1-1), the mutants were first inactivated for 2&h and then the expression of Isw1NLS was induced for 2&5&h. The cse1-1 mutant was inactivated at 15&C for 12&h followed by galactose induction (5&8&h). GFP fluorescence was observed by live-cell imaging. B) As a control, expression of the importin-&-binding domain (aa1&60, IBB) of importin-& was induced by 2% galactose in the mutants of the classical import pathway. Cultivation conditions and fluorescence microscopy were the same as in part A. C) The expression of Isw1()-GFP was induced by 2% galactose (2&h) in the BY isw1&D and nup2&D cells grown at 30&C to the exponential phase. The GFP signal was assessed by live-cell microscopy. Scale bar: 4&&m. D) The N/C ratios were calculated from more than 50 cells of each strain and plotted along with the standard deviation values.To assess the nuclear import of the full-length Isw1 protein in the strains mutated in karyopherins of the classical import pathway, we used the plasmid pGalIsw1-GFP, which provided a galactose-regulated expression of the full-length Isw1 in a C-terminal fusion with GFP. The same importin mutant strains were used as in the above experiments with the Isw1 truncation, and similarly a cytoplasmic accumulation of the Isw1-GFP signal, indicating an import defect was observed in the srp1-31, srp1-54, rsl1-4 and cse1-1 mutants even at permissive temperature (Figure&A). As demonstrated in Figure&B, the mean N/C ratios calculated for the observations at permissive temperatures were significantly lower compared to the values obtained for the wild-type strain W303, reflecting the apparent cytoplasmic accumulation of the Isw1NLS-GFP signal in these strains. At restrictive temperatures, the mean N/C ratios further declined compared to permissive conditions. In cse1-1 cells, FL-Isw1-GFP was not expressed under the galactose induction under restrictive conditions (15&C), because no GFP signal was detectable. However, cse1-1 cells displayed an elevated cytoplasmic GFP signal even at permissive temperature, suggesting that the FL-Isw1 nuclear import was impaired. Also, distribution curves of the N/C ratios measured for individual cells in the tested mutants (Figure&C) clearly demonstrate a changed pattern of full-length Isw1 nuclear import even under permissive conditions, particularly in the importin-& mutants. Interestingly, under restrictive conditions the N/C ratios and their distribution not only changed in the analyzed mutants in the classical import pathway, but also in the wild-type strains (Figure&D).Figure&6. Nuclear import of full-length Isw1 is mediated by importin-&. A) Expression of the full-length Isw1-GFP was induced by 2% galactose in the indicated strains. The permissive temperature was 25&C, with the exception of the cse1-1 strain, which was grown at 30&C. At restrictive temperature (37&C), the mutants were first inactivated for 2&h and then Isw1-GFP expression was induced for 2&5&h. The strain cse1-1 carrying pGalIsw1-GFP did not grow at 15&C. The GFP signal was observed by live-cell imaging. Scale bar: 4&&m. B) N/C ratios were calculated from more than 50 cells of each strain and plotted along with the standard deviation values. C) Distribution curves of calculated N/C ratios for the mutants in the classical import pathway were plotted against the reference strains under permissive and restrictive conditions to indicate subtle differences that were difficult to see in the microscopic images presented in (A).In summary, changes in the nuclear translocation of the full-length Isw1 as well as of the short Isw1NLS fragment in the mutants of the classical import pathway suggest that Isw1 nuclear transport is predominantly mediated by the classical import pathway.Isw1NLS interacts with importin-&To confirm that the classical import pathway is implicated in the nuclear translocation of Isw1 and to determine whether Isw1 interacts with importin-&, we performed GST pull-down assays with recombinant importin-& and the lysate of yeast cells harboring pGalIsw1()-GFP. GST was fused to the N-termini of both the full-length importin-& (Imp&FL) and without its IBB domain (Imp&&DIBB). GST-fused recombinant importin-& variants and GST alone embedded in glutathione-agarose beads were combined with the yeast lysate containing the Isw1NLS-GFP fusion protein. Isw1NLS was pulled down more efficiently by importin-& with a deleted IBB domain than by the full-length protein (Figure&). This is in agreement with the observation that the IBB domain competes with a cNLS for the cNLS-binding sites on importin-& []. GST alone did not pull down any Isw1NLS-GFP. Note that the Isw1()-GFP fragment carrying Isw1NLS-GFP migrated in SDS-PAGE as two distinct bands and the upper form bound to importin-& with a weaker affinity than the one in the lower band. This suggested that Isw1NLS could be post-translationally modified and this modification might negatively affect the Isw1NLS binding to importin-&. However, our mass-spectrometric (MS) analyses did not detect any post-translational modification of the Isw1()-GFP fragment. On the other hand, MS experiments revealed several N-terminal peptides missing from the mass spectra of the lower band compared to that of the upper band, indicating the possibility of a second translation initiation site at methionine 1071. This second initiation site, 10-aa residues away from the first one, is probably as favorable as the first one and the yeast cells transformed with pGalIsw1()-GFP produce a similar amount of both versions of the Isw1NLS carrying fragment. The Isw1()-GFP fragment lacking M1071, just 20-aa residues shorter, migrated as a single band in SDS-PAGE (data not shown) thus supporting our hypothesis of a second translation initiation site within the Isw1()-GFP fragment. The reason why the upper band was bound less efficiently than the lower one could simply be due to the shorter fragment having better access to the importin-& NLS-binding pockets.Figure&7. Isw1NLS interacts with importin-&. GST-fused importin-& variants Imp&FL, Imp&&DIBB and GST alone purified from Escherichia coli and immobilized on glutathione-agarose beads were combined with equal amounts of the lysate prepared from the yeast cells carrying Isw1NLS-GFP (pGalIsw1()-GFP). After eluting the bound proteins by boiling in 2& SDS & Laemmli loading buffer, the samples were resolved in 10% SDS-PAGE. The GST-bound importin-& variants were detected with a polyclonal anti-GST antibody and the interacting Isw1NLS fused to GFP with anti-GFP-HRP conjugate.The in vitro-binding assay thus confirmed that the Isw1NLS present in the Isw1()-GFP fragment is bound by importin-&, the adaptor protein of the classical import pathway, and mass spectrometry showed that there is no dominant post-translational modification that could affect this interaction. We thus conclude that the classical import pathway is implicated in Isw1 nuclear targeting.Isw1 is transferred to nucleus without participation of its complex partnersWe noticed that in strains with mutated karyopherins of the classical import pathway grown under restrictive conditions, the full-length Isw1-GFP protein still entered the nucleus and the shorter, NLS-carrying fragment Isw1()-GFP exhibited a higher mean N/C ratio than GFP alone, which demonstrates the presence of a residual active nuclear import. Both findings might indicate that either mutated karyopherins were not fully inactivated under restrictive conditions, or that other proteins were involved in the translocation of Isw1 into the nucleus. We excluded the participation of other non-essential karyopherins in Isw1 nuclear import using a set of deletions in individual karyopherins (Figure S1). The nuclear localization of the human Isw1 homologue (hSNFH2) was described to be affected by its interaction partners in complexes []. Because Isw1 of S. cerevisiae forms two complexes [], we analyzed the localization of the full-length Isw1-GFP in strains lacking proteins Ioc2 and Ioc3, which are necessary for the formation of the complexes Isw1a and Isw1b, respectively. In the ioc2&D, ioc3&D double-deletion strain, the Isw1-GFP signal always remained in the nucleus (Figure S2). To exclude the possibility that the only situation in which Isw1 is imported into the nucleus as the complex Isw1a and/or Isw1b is when the classical import pathway is inactive, we further constructed a strain carrying the temperature-sensitive allele of importin-& srp1-31, and both deletions ioc2&D and ioc3&D. The result obtained with this strain was the same as that of the strain carrying the srp1-31 mutation alone. Isw1-GFP was found to be localized in the cytoplasm as well as in the nucleus (Figure S2). The additional removal of Ioc2 and Ioc3 thus did not lead to a reduction in the Isw1FL-GFP nuclear signal present in srp1-31 cells. We conclude that neither Isw1's interaction partners in the Isw1a and Isw1b complexes of S. cerevisiae nor non-essential karyopherins have a detectable influence on the nuclear import of Isw1. However, we cannot exclude the participation of another importin that competes with the major Isw1 importin, importin-&, for the identified atypical bipartite NLS but is only able to bind Isw1NLS when importin-& is mutated.Taken together, the data obtained either with the truncated or with the full-length protein showed that the import of the S. cerevisiae chromatin remodeling factor Isw1 into the nucleus depends solely on the bipartite nuclear targeting signal KRIR X19 KKAK located at the end of the C-terminus of the protein molecule and that the nuclear translocation of Isw1 is predominantly mediated by the classical import pathway.DiscussionMany proteins are functionally nuclear proteins and need to be ultimately imported into the nucleus. Chromatin remodeling factor Isw1 is one of them and in this work we studied its nuclear translocation, identified the nuclear targeting signal and the corresponding import pathway.Our in silico analyses using NLS predictive tools localized most putative S. cerevisiae Isw1 NLSs into the first 200 aa, however, the group of J. Mellor [] observed the nuclear localization of an Isw1 peptide corresponding to the 248 C-terminal amino acids. We confirmed this observation and specified that the Isw1 NLS KRIR X19 KKAK is an atypical bipartite NLS with an unconventionally long linker located at the end of the C-terminus of the protein (aa). Kosugi and coworkers [] postulated that a bipartite cNLS is comprised of two imperfect monopartite cNLSs, one corresponding to the class 3/class 4 in its N-terminal region, and the second to the class 1/class 2 of cNLSs in its C-terminal region. However the KRIR motif (aa), which constitutes the N-terminal part of Isw1NLS, does not resemble any of the defined classes of monopartite sequences. The C-terminal part of Isw1NLS, KKAK (aa) is similar to the class 2 monopartite sequence. The first two N-terminal residues, KR (aa) from the KRIR motif, are essential for an effective NLS function in the context of the short NLS-carrying fragment. However in the context of the full-length Isw1 protein, a very low nuclear signal (N/C &1.2) was observed in variants with the mutated KR motif, which could be attributed to the remaining intact C-terminal motif KKAK of the bipartite Isw1NLS, which might function autonomously as a weak monopartite NLS. Also, we cannot exclude that other parts of the Isw1 molecule (not present in the short NLS-carrying fragment) might contribute to this low-affinity nuclear translocation. The C-terminal motif KKAK is indisputably indispensable for Isw1 nuclear import in both the short and full-length versions. Our results clearly show that Isw1NLS is the bipartite signal in which both parts are individually incapable of ensuring an efficient translocation of Isw1 to the nucleus, but together with the linker with the unconventional length of 19 aa, serve as an efficient nuclear targeting signal.Structural studies of mouse, human and yeast importin-&'s using X-ray crystallography [] set up the optimal basic/bipartite consensus sequence for binding to importin-&. The KR X10&12 K(K/R)X(K/R) motif was established for the yeast model and the effective linker length was assigned to be 10&12 aa residues, which is equal to the distance between the major and the minor binding sites on importin-& []. However, bipartite NLSs with an extended linker that can span even more than 30 aa residues were experimentally identified in various proteins originating from humans, mice, plants and yeasts and their number has been continually increasing over the last few years []. Although atypical, these bipartite NLSs are recognized, bound and transported by importin-&. An efficient binding to importin-& not only depends on the basic stretches of a NLS but also on the linker composition, determining its flexibility, which affects NLS conformation and thus the binding of both basic regions to the minor and major pockets of importin-& []. Replacing the amino acids in the linker of a bipartite NLS for acidic residues can increase NLS functionality in an additive manner []. There are seven acidic residues in total within the Isw1NLS 19-aa linker, and the mutagenesis of two glutamic residues at positions 1083 and 1084 in the vicinity of the N-terminal part of Isw1NLS substantially reduced the nuclear import of Isw1. The composition of the adjacent part of the Isw1NLS linker thus has a very significant impact on the NLS functionality. However, shortening of the linker from either the N- or C-terminus by 9 aa did not negatively affect the functionality of Isw1NLS. It instead enhanced nuclear import in the context of the short Isw1NLS-carrying fragment, probably due to an even better accommodation of the Isw1NLS fragment with a shorter linker to the importin-& molecule. Thus, it seems that the overall length of the linker sequence is not crucial for the function of Isw1NLS under used experimental conditions, which has not always been the case for other extended linkers []. The reasons for the presence of such an atypical-length linker within bipartite Isw1NLS are not clear and can only be speculated upon, because the presumption frequently thought for longer linkers, that they might be needed to adopt a favorable conformation for an efficient binding to the importin to assure their nuclear import, seems to not be valid for Isw1NLS.The Isw1 molecule consists of several domains important for its function and interactions within various complexes and these are spread throughout the molecule. The end of the C-terminus is the only &unstructured& part, according to its crystal structure&[], and the location of the bipartite NLS behind the domains important for Isw1 protein functionality at the unstructured C-terminus might enable Isw1 to form a favorable conformation for its efficient binding to importin-&. In silico analysis of the sequences of Isw1 homologues from other yeast species found a new conserved motif spanning the N-terminal of Isw1NLS, and supports our hypothesis of a favorable context for the presence of an NLS at the C-termini of ISWI proteins. Unconventional bipartite NLSs with long linkers are usually not recognized by NLS predictive tools. Only cNLS Mapper [] was designed to identify S. cerevisiae bipartite cNLS with linkers up to 20-aa residues long. This software, however, detected neither the intact Isw1NLS nor the putative bipartite NLSs in the proteins homologous to Isw1. Interestingly, when the sequences of the full-length Isw1 protein with either of two linker deletions were submitted to cNLS Mapper, a bipartite NLS was found in both versions spanning the region identified by ourselves as the Isw1 bipartite NLS. cNLS Mapper and/or PSORTII also identified bipartite cNLS motifs with a linker of 11&14 aa at the end of the C-termini of human (hSNF2H), plant (A. thaliana CHR11, CHR17), frog (X. leavis Smarca5) and fly (D. melanogaster ISWI) Isw1 homologues (Figure S3). These data and the fact that we were able to experimentally prove the presence of a monopartite NLS in the last 58 aa of the second S. cerevisiae ISWI protein, Isw2 (our unpublished results), allows us to suggest that the C-termini of ISWI family proteins play an important role in their nuclear localization.With the help of live-cell imaging and GST-pulldown assays, we were able to prove that the classical import pathway is involved in the nuclear trafficking of Isw1. The affected Isw1 import in the mutants of importin-& (srp1-31 and srp1-54) was even detectable at the permissive temperature, and analysis of the mutant cells under restrictive conditions gave similar results. Although a little surprising, this observation is in agreement with the first characterization of these importin-& mutants by Yano and coworkers [], who noticed their mutant phenotypes at as little as 25&C, and a further temperature shift to 38&C had only a minor effect. The fact that transfer to restrictive conditions did not have any obvious additional effect on Isw1NLS localization might suggest that the mutated importins (especially importin-& alleles) have an altered molecular structure that prevents their efficient specific binding to Isw1NLS, and so affects the Isw1 import rate even under permissive conditions. To support this hypothesis, we compared the steady-state distribution of SV40NLS-GFP and Isw1NLS-GFP. Our results (Figure S4) clearly show that the nuclear import of both reporters is affected in strains with mutated karyopherins of the classical import pathway, even at permissive temperature, and that the effect on import is significantly stronger with Isw1NLS. It thus seems that under permissive conditions, Isw1NLS is much more sensitive toward changes in the structural conformation of classical importins-& and -& than the prototypical classical monopartite SV40NLS. Further support that alterations in the nuclear import of the short Isw1NLS-GFP fusion protein even at permissive temperature are specific for the binding of this NLS to the classical import pathway karyopherins comes from the localization pattern of the IBB domain (imported independently of importin-&) in the same mutant strains. The localization of the IBB domain was unaffected, as can be seen from the N/C ratios (Figure&). Finally, a kinetic import assay using azide and 2-deoxy-glucose performed on cells with mutated classical karyopherins carrying the Isw1()-GFP fusion protein also demonstrated that Isw1NLS needs the classical import pathway to be functional for its nuclear translocation. At restrictive conditions (37&C) we did not observe any re-import of Isw1NLS-GFP into the nucleus over a period of 15&min in importin-& and -& mutant cells, and the GFP signal remained diffusely localized in the analyzed cell, as shown in Figure S5.Analyses of images from live-cell microscopy revealed a large scattering of calculated N/C ratio values in the cell population of the analyzed strains. A similar observation was described by Hodel et al. [], who showed that the distribution of N/C ratio values measured for a single cell varied in the population, and this distribution reflected the affinity of a cargo NLS to an import receptor. To put it simply, strong NLSs have a higher N/C ratio with a wider distribution when measured in a single cell than weak NLSs, and these populations can overlap each other. We used this analysis to describe the steady-state localization of the full-length Isw1-GFP in strains carrying mutated importin-& and -&. The distribution curves of N/C ratios calculated for each classical karyopherin mutant strain cultivated at permissive temperature were within a narrow range of lower N/C ratios in srp1-31 and srp1-54, indicating an import defect compared to the wide distribution in both reference strains W303 and BY4741. At the restrictive temperature, the most remarkable difference in the N/C ratio distribution from the wild-type strains was observed in the srp1-31 importin-& mutant. The N/C ratio distribution in the strains srp1-54, rsl1-4 was essentially unchanged. Distribution curves obtained for the restrictive temperature imply remarkable changes in the N/C ratio distribution between the restrictive and permissive temperature in the reference strains W303 and BY4741, indicating an effect on import of Isw1-GFP at the restrictive temperature. This points to a very interesting possibility that has not been previously considered, the regulation of Isw1 at the level of its nuclear translocation under mild-heat shock conditions. This would be a very interesting subject for further studies, because Isw1 was identified as a repressor of stress-response genes including those responsive to mild-heat shock [].As was mentioned above, we detected a nuclear Isw1NLS-GFP signal in the mutants of importin-&, even under restrictive conditions when the protein was considered to be nonfunctional. Similarly, a nuclear signal of a cargo protein was observed by other authors when searching for the cNLSs of various S. cerevisiae proteins []. The portion of the cargo signal detected in the nucleus was always specific to the cargo and the importin-& mutant, indicating that different importin-& mutations affected the nuclear import of distinct cargoes through distinct mechanisms. The importin-& mutants used in our study are not mutated in the cNLS binding sites. The mutant allele srp1-31 carries a point mutation in the first armadillo (ARM1) repeat and the allele srp1-54 in ARM9 []. As described by Conti et al. [], ARM1 and ARM9 are not directly involved in cNLS recognition but make an essential contribution to the maintenance of the NLS-binding site's structure. It is thus conceivable that Isw1NLS binding to the mutated importin-& was not completely abolished, but only substantially reduced, which may explain why a small portion of the Isw1-GFP signal could be observed in the nucleus, even after the assumed inactivation of importin-&. This nuclear signal was more prominent in cells of the importin-& mutants carrying the full-length Isw1 protein than in those with the small Isw1NLS-bearing fragment Isw1(), indicating that not only might the binding affinity of FL-Isw1 to the importin-& mutants differ from that of the short Isw1NLS fragment, but also that other proteins (e.g. complex partners or other karyopherins) might be involved in the nuclear translocation of Isw1. Although there is increasing evidence that some nuclear protein complexes are assembled in the cytoplasm and subsequently imported into the nucleus as a whole complex [], and nuclear import of the human Isw1 homologue hSNF2H was described to be affected by its complex partners [], we clearly showed that the formation of S. cerevisiae Isw1 complexes was not required for Isw1 nuclear trafficking. This was further supported by an assessment of the localization of the Isw1(1&1090)-GFP fusion protein missing only the last 39 aa and of the Isw1(1102AAAA)-GFP mutant. Both proteins, containing all the domains essential for Isw1's functions and interactions, with the exception of the motif KKAK of the identified bipartite NLS, were never observed in the nucleus. Consistent with this, the results of the mutagenesis of both NLS motifs in the context of full-length Isw1 indicate that the C-terminal part of Isw1NLS (motif KKAK) might function as a very low-affinity monopartite signal that could be recognized either by the importin-& or by an as-yet unidentified importin. Significant participation of non-essential importins in Isw1 nuclear import was excluded (Figure S1). However we cannot exclude that an importin, even one of those tested, competes with the major Isw1 importin, importin-& in binding to the identified atypical bipartite NLS (or its C-terminal motif), most probably only under conditions when importin-& is mutated, thus contributing to the observed nuclear signal in the importin-& mutants.In conclusion, we have identified the nuclear targeting signal at the end of the C-terminus of the S. cerevisiae chromatin remodeling factor Isw1 and the corresponding import pathway. The signal sequence is an atypical bipartite NLS with an extended linker that meets all the criteria for a functional cNLS. Similar motifs or putative bipartite cNLSs were identified in silico at the end of the C-termini of Isw1 homologues from other yeast species and higher eukaryotes, respectively, and it would be of interest to find any experimental evidence of their role in nuclear import. Analysis of Isw1 localization under mild-heat shock raised the very interesting question of whether Isw1 might be regulated at the level of its nuclear translocation in response to some stress conditions.Materials and MethodsStrains and mediaThe S. cerevisiae strains used in this work are listed in Tables& and S1. E. coli strain DH5& [F& rec A1 supE44 endA1 hsdR17 (rk&, mk+) gyrA96 relA1 thi-1 &D(lacIZYA-argF)U169deoR (&P80d&D(lacZ)) M15] was used as a host in cloning procedures and E. coli BL21 Star&(DE3) [F& ompT hsdSB(rB&, mB&) gal dcm rne131 (DE3)] (Invitrogen) for the production of recombinant proteins. Standard bacterial and yeast cultivation media and temperatures were used, if not stated otherwise, and standard methods were used for all DNA manipulations [].Table&2.&S. cerevisiae strains used in this studyW303MAT&a ura3-1 leu2-3,112 trp1-1 his3-11,15 ade2-1 can1-100R. Rothstein (Columbia University, NY)BY&4741MAT&a his3&D1 leu2&D0 met15&D0 ura3&D0EUROSCARFBY&4741 isw1&DMAT&a his3&D1 leu2&D0 met15&D0 ura3&D0 isw1::kanMX4EUROSCARFBY&4741 nup2&DMAT&a his3&D1 leu2&D0 met15&D0 ura3&D0 nup2::kanMX4EUROSCARFS288CMAT&& SUC2 mal mel gal2 CUP1 flo1 flo8-1InvitrogenPSY&1040MAT&& cse1-1 ura3-52 trp1-&D901 his3-11,15 ade2-101P. Silver (Harvard Medical School, Boston)PSY&1103MAT&a rsl1-4 ura3-52 leu2&D1 trp1&D63P. Silver (Harvard Medical School, Boston)PSY&1201MAT&a pse1-1 ura3-52 leu2&D1 trp1&D63P. Silver (Harvard Medical School, Boston)NOY 612MAT&& srp1-31 ura3-1 leu2-3,112 trp1-1 his3-11 ade2-1can1-100M. Nomura, (University California, Irvine, (46))NOY672MAT&a srp1-54 ura3-1 leu2-3,112 trp1-1 his3-11 ade2-1 can1-100M. Nomura, (University California, Irvine, (46))YTT829MATa ura3-1 leu2-3,112 trp1-1 his3-11,15 ade2-1 can1-100 RAD5 ioc2::KanMX4 ioc3::nat1T. Tsukiyama, (Fred Hutchinson Cancer Research Center, Seattle)CRY1567MAT&& srp1-31 ura3-1 leu2-3,112 trp1-1 his3-11,15 ade2-1 can1-100 ioc2::KanMX4 ioc3::nat1This studyPlasmids constructionsAll plasmids constructed and used in this study are listed in Table&. Oligonucleotides used in the construction of plasmids are listed in Table S2. The ISW1 coding region together with the 5&UTR and 3&UTR sequences was amplified by PCR from the chromosomal DNA of S. cerevisiae S288C using the primers PT03 and PT04. The obtained DNA fragment was digested with BamHI and ligated into pBSKS, resulting in the plasmid pIsw1. For localization studies, two C-terminal fusions of Isw1 with GFP were produced, the first regulated by the natural promoter (pIsw1-GFP) to be used in non-essential karyopherin deletion strains, and the second by the inducible GAL1 promoter (pGalIsw1-GFP) to localize Isw1 in essential karyopherin mutants. pIsw1-GFP was generated by amplification of the ISW1 gene with its promoter from the plasmid pISW1 using the oligonucleotides Isw1cgfp5 and Isw1cgfp3, and by insertion of the resulting SacI-BamHI fragment into the vector pUG35. Regulated Isw1-GFP fusion was achieved by PCR amplification of the ISW1 gene without the promoter using oligonucleotides 5Galisw1 and 3revGFPSac, pISW1-GFP was used as the template. The PCR product was cloned as a SacI fragment into the plasmid pYC2/CT downstream from the GAL1 promoter, producing pGalIsw1-GFP.Table&3.&Plasmids used in this studypBluescript KS(pBSKS)AmpRStratagenepUG35AmpR, URA3, ARS/CENJ. Hegemann (H. Heine University, Duesseldorf, Germany)pIsw1AmpRThis studypIsw1-GFPAmpR, URA3, ARS/CENThis studypYC2/CTAmpR, URA3, ARS/CEN, GAL1InvitrogenpGalIsw1-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1(1&200)-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1(1&1000)-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1(1&1060)-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1()-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1()-GFPAmpR, URA3, ARS/CEN, GAL1This studypYC2/CT/lacZAmpR, lacZ, URA3, ARS/CEN, GAL1InvitrogenpGalIsw1()-lacZAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1(1&1080)-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1()-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1(1&1090)-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1(1&1105)-GFPAmpR, URA3, ARS/CEN, GAL1This studypKW431AmpR, URA3, 2&&[]pIsw1()-GFP2AmpR, URA3, 2&&This studypIsw1()-GFP2AmpR, URA3, 2&&This studypIsw1()-GFP2-1102AAAAAmpR, URA3, 2&&This studypIsw1()-GFP2-1079AAAmpR, URA3, 2&&This studypIsw1()-GFP2-1082AAAAmpR, URA3, 2&&This studypIsw1()-GFP2-1083AAAmpR, URA3, 2&&This studypGalIsw1(aa)-GFP-1102AAAAAmpR, URA3, ARS/CEN, GAL1This studyBG1805Srp1AmpR, URA3, ARS/CEN, GAL1Open BiosystemspGalIBB-GFPAmpR, URA3, ARS/CEN, GAL1This studypGal-GFPAmpR, URA3, ARS/CEN, GAL1This studypGEX-5x-3AmpRGE HealthcarepGSTSrp1FLAmpRThis studypGSTSrp1&DIBBAmpRThis studypGalIsw1(1102AAAA)-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1(1079AA)-GFPAmpR, URA3, ARS/CEN, GAL1This studypIsw1()-GFP2-(&D)AmpR, URA3, 2&&, ADH1This studypIsw1()-GFP2-(&D)AmpR, URA3, 2&&, ADH1This studypGalIsw1(&D)-GFPAmpR, URA3, ARS/CEN, GAL1This studypGalIsw1(&D)-GFPAmpR, URA3, ARS/CEN, GAL1This studyA series of ISW1-deletion variants fused to GFP was generated & Isw1(1&200), Isw1(1&1002), Isw1(1&1060), Isw1(), Isw1(1&1105), Isw1(1&1090), Isw1(1&1080) and Isw1(). PCR primers were designed in such a way that the 5&-end primers always contained the ATG codon preceding the first amino acid of the Isw1 fragment and 3&-end primers enabled in-frame Isw1-C-terminal fusion with GFP. The plasmid pGalIsw1-GFP was used as the template and the PCR products obtained were cloned as SacI-SalI fragments into pGalIsw2-GFP (I. Malcova, unpublished) to replace the full-length ISW2 gene. The resulting Isw1 variants were designated pGalIsw1(aaX-aaY)-GFP, where aaX represents the first amino acid and aaY the last amino acid in the fragment.Isw1NLS fusions of a greater mass than the passive diffusion limit were constructed either with tandem GFP molecules (GFP2) or the E. coli lacZ gene. pIsw1()-GFP2 and pIsw1()-GFP2 were produced by replacing the SV40 NLS fragment in the pKW431 vector [kindly provided by M. Miller []]. Isw1NLS-carrying fragments were amplified on the pIsw1 template using oligonucleotides I11061HindIII, I11079HindIII, I11129EcoRI and cloned as HindIII-EcoRI fragments downstream of the ADH1 promoter. pGalIsw1()-lacZ was the pYC2/CT/lacZ vector with the Isw1NLS (aa) amplified using the primers I11061HindIII and I11129BamHI, and cloned as a HindIII-BamHI fragment.The control pGal-GFP plasmid was created by SacI, SalI digestion of pGalIsw2-GFP followed by isolation, filling-up the ends and religation of the GAL1 promoter and GFP. The pGalIBB-GFP plasmid expressing the IBB domain of the importin-&/Srp1, was generated by amplifying the SRP1 coding sequence spanning amino acids 1&60 from BG1805Srp1 (Open Biosystems) as the template using the primers Srp1fSac and Srp1r60Sal. The resulting PCR product was ligated into the pGalIsw2-GFP plasmid as a SacI-SalI fragment to replace the ISW2 gene. The identity of ISW1 inserts in all the plasmids was verified by sequencing and immunoblotting.Isw1 mutagenesisTo determine the functional significance of individual amino acids within Isw1NLS, a QuikChange II XL Site-directed mutagenesis kit (Stratagene) was employed to replace the amino acids in question with alanines using pIsw1(aa)-GFP2 as the template. Three sequential PCR reactions were performed to reconstruct the full-length Isw1 carrying the alanine substitutions in the KR and the KKAK motifs. In the first reaction, the C-terminal fragment of Isw1 (nt ) containing the specific mutation was amplified using the primers Isw1nt3180f and I11129r and the plasmid pGalIsw1()-GFP-1102AAAA or pISW1(aa)-GFP2-1079AA as the template. In the second reaction, the Isw1 fragment without any mutation (nt ) was amplified using the primers Isw1nt1291f and Isw1nt3233r and the plasmid pGalIsw1-GFP as the template. Both new PCR fragments were pooled and amplified as one fragment (nt) using the primers Isw1nt1291f and I11129r. The resulting fragment containing the mutation was cloned into the pGalIsw1-GFP plasmid as a SphI-SalI fragment to replace the wild-type fragment without the mutation, yielding pGalIsw1(1079AA)-GFP and pGalIsw1(1102AAAA)-GFP, respectively. The identity of ISW1 inserts in all the constructs was verified by sequencing.Isw1NLS linker shorteningTo shorten the extended linker of 19 aa within Isw1-NLS to the conventional length of 12 aa, the method of gene splicing by PCR-driven overlap extension [] was used. To delete aa, the fragments N1 (oligonucleotides Isw1nt1291f and I1LS-N1rev) and C1 (oligonucleotides I1LS-C1fwd and GFP40) were first amplified using pGalIsw1-GFP as the template. The internal oligonucleotides I1LS-N1rev and I1LS-C1fwd contain overlapping sequences, enabling two different gene regions to be joined together. The N1 and C1 fragments were then pooled and hybridized and the resulting chimeric product LS1 was used as the template for the subsequent PCR. To introduce the deletion into the context of the full-length Isw1, pGalIsw1(&D)-GFP, the fragment bearing deletion &D, was amplified using the oligonucleotides Isw1nt1291f and GFP40 and chimeric LS1 as the template. The PCR product was cloned into the pGalIsw1-GFP plasmid as a SphI-SalI fragment to replace the wild-type fragment. To make deletions within the short Isw1NLS fragment, pIsw1()-GFP2-(&D), the fragment bearing the deletion &D, was amplified using the oligonucleotides I11061HindIII and I11129EcoRI, and LS1 as the template. The obtained PCR product was cloned into the pIsw1()-GFP2 plasmid as a HindIII-EcoRI fragment to replace the intact IswI() fragment. An analogous approach was used to construct pGalIsw1(&D)-GFP and Isw1()-GFP2-(&D), but using a different set of oligonucleotides (Isw1nt1291f, I1LS-N2rev, I1LS-C2fwd and GFP40) for the generation of fragments N2, C2 and the final chimeric template LS2.Cultivation conditions for live-cell imagingYeast cell cultures were grown in SC (0.17% yeast nitrogen base medium without amino acids and ammonium sulfate, 0.5% ammonium sulfate, 2% glucose, supplemented with a complete or a strain-specific drop-out), in SCR1 (the same as SC but the only carbon source was 1% raffinose) or in SCR2 (the same as the SC medium, but the only carbon source was 2% raffinose). Solid media contained 2% agar. All strains except for temperature-sensitive mutants were cultivated at 30&C. Cells expressing GFP-fused proteins from their own promoter or from the ADH1 promoter were grown overnight in the SC medium, diluted into fresh medium (1:3) in the morning and grown for an additional 3&h prior to localization studies by live-cell microscopy. Cells expressing GFP-fused proteins under the control of the GAL1 promoter were grown overnight in SCR2. In the morning, the culture was diluted with SCR1 medium and grown for 3&h prior to adding galactose to a final concentration of 2% to induce the expression of GFP-chimeras. The GFP signal was followed after 2&5&h of induction (dependent on the strain). Thermosensitive strains (srp1-31, srp1-54, rsl1-4, pse1-1) were cultivated at 25&C. The expression of Isw1-GFP-chimeras was induced by adding galactose after the inactivation of the appropriate karyopherin (cultivation for 2&h at 37&C) and the GFP signal was monitored after 3&5&h of induction in 2% galactose at 37&C. The cold-sensitive strain cse1-1 was cultivated at 30&C as the permissive temperature. To inactivate the cse1-1 exportin, cells were cultured for 12&h at 15&C in SCR1, and the expression of Isw1-GFP-fusions was induced by adding galactose to 2%. The localization of the GFP signal was followed after 5&8&h of galactose induction at 15&C. To indicate the position of the nucleus, intact cells were incubated with DAPI (10&&g/mL) for 60&min.Live-cell microscopyThe localization of GFP-fused proteins was examined by direct fluorescence microscopy in living yeast cells (see above for cultivation conditions). The cells were centrifuged and washed to remove the growth medium, mounted on a coverslip and coated with a slice of 1% agarose prepared in the appropriate growth medium. The distribution of GFP-fused proteins was analyzed with a 100& PlanApochromat objective (NA 1.4) using an Olympus IX-71 inverted microscope equipped with a Hammamatsu Orca/ER digital camera and an Olympus Cell RTM detection and analyzing system (GFP filter block U-MGFPHQ, exc. max. 488, em. max. 507, DAPI filter block U-MNUA2, exc. max. 440, em. max. 500&520). The Olympus Disc Scanning Confocal Unit (DSU) was used for some observations. Images were processed using Olympus Cell-RTM and Adobe CS2 software. The images were captured in 12-bit format in all experiments. The quantification of the nucleus/cytoplasm GFP-signal ratio (N/C ratio) was performed as described previously [], except the average gray values (mean fluorescence intensities) were obtained in the Olympus Cell-RTM program using a round region of interest (ROI) with a size of 0.3&&m2. The level of GFP signal reflecting gene expression varied between strains and even in the cells of one strain population. Realizing that the expression level may affect the resulting N/C ratio, we calculated the mean N/C ratio from the cells that had the average gray values of the GFP-signal in their nucleus between 250 and 1000 pixels only. Thus, we always compared cell populations with a similar level of expression of analyzed GFP fusions.Indirect immunofluorescence microscopyBY4741 isw1&D cells harboring the plasmid pGalIsw1()-lacZ bearing the C-terminal fusion of Isw1NLS with the E. coli lacZ gene were fixed and processed according to the procedure described previously []. The primary antibody was a mouse monoclonal antibody against &-galactosidase (BG-02, Exbio, CR) that was used at a final dilution of 1:50 in 2% (w/v) BSA/PEM. Cy3-labeled goat anti-mouse antibody (GAM-Cy3; Jackson Laboratories) diluted 1:200 was used as the secondary antibody. Before observation, the labeled cells were mounted in the TBS buffer containing }