* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Nuclear and nucleolar targeting of human ribosomal protein
Survey
Document related concepts
Endomembrane system wikipedia , lookup
Magnesium transporter wikipedia , lookup
Protein phosphorylation wikipedia , lookup
Protein structure prediction wikipedia , lookup
Protein moonlighting wikipedia , lookup
Signal transduction wikipedia , lookup
Intrinsically disordered proteins wikipedia , lookup
Cell nucleus wikipedia , lookup
Protein–protein interaction wikipedia , lookup
Nuclear magnetic resonance spectroscopy of proteins wikipedia , lookup
Transcript
Oncogene (1999) 18, 1503 ± 1514 ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $12.00 http://www.stockton-press.co.uk/onc Nuclear and nucleolar targeting of human ribosomal protein S25: Common features shared with HIV-1 regulatory proteins Satoshi Kubota1, Terry D Copeland2 and Roger J Pomerantz1 1 The Dorrance H. Hamilton Laboratories, Center for Human Virology, Division of Infectious Diseases, Department of Medicine, Jeerson Medical College, Thomas Jeerson University, Philadelphia, Pennsylvania 19107; and 2ABL-Basic Research Program NCI-Frederick Cancer Research and Development Center P.O. Box B, Frederick, Maryland 21702, USA The nuclear and nucleolar targeting properties of human ribosomal protein S25 (RPS25) were analysed by the expression of epitope-tagged RPS25 cDNAs in Cos-1 cells. The tagged RPS25 was localized to the cell nucleus, with a strong predominance in the nucleolus. At the amino terminus of RPS25, two stretches of highly basic residues juxtapose. This con®guration shares common features with the nucleolar targeting signals (NOS) of lentiviral RNA-binding transactivators, including human immunode®ciency viruses' (HIV) Rev proteins. Deletion and site-directed mutational analyses demonstrated that the ®rst NOS-like stretch is dispensable for both nuclear and nucleolar localization of RPS25, and that the nuclear targeting signal is located within the second NOS-like stretch. It has also been suggested that a set of continuous basic residues and the total number of basic residues should be required for nucleolar targeting. Signal-mediated nuclear/nucleolar targeting was further characterized by the construction and expression of a variety of chimeric constructs, utilizing three dierent backbones with RPS25 cDNA fragments. Immuno¯uorescence analyses demonstrated a 17 residue peptide of RPS25 as a potential nuclear/ nucleolar targeting signal. The identi®ed peptide signal may belong to a putative subclass of NOS, characterized by compact structure, together with lentiviral RNAbinding transactivators. Keywords: ribosome; protein; nuclear targeting; rev; HIV-1 Introduction The mechanism of nuclear translocation of proteins has been characterized in signi®cant detail. It has been shown that nuclear import through the nuclear pore complexes consists of two distinct steps, in which karyopherin/importin and a small GTP-binding protein, Ran/TC4, play central roles (Akey and Goldfarb, 1989; Melchior and Gerace, 1995; GoÈrlich and Mattaj, 1996). A number of nuclear localization signals (NLS) have been identi®ed in a variety of nuclear proteins. Recently, alternative pathways of nuclear import, one of which is employed speci®cally by certain ribosomal proteins in yeast, have been reported as well (Siomi and Dreyfuss, 1995; Goldfarb, 1997; Rout et al., 1997). Correspondence: RJ Pomerantz Received 27 May 1998; revised 16 September 1998; accepted 16 September 1998 In contrast, targeting of proteins into the cell nucleolus by single distinct signals has been described in a limited number of proteins, primarily in RNA-binding transactivator proteins of human retroviruses ± human immunode®ciency virus type 1 (HIV-1) and human T-cell leukemia virus type I (HTLV-I) (Siomi et al., 1988, 1990; Kubota et al., 1989; Dang and Lee, 1989). The functional signi®cance of the nucleolar localization of these proteins is still controversial, although a possible role of the nucleolus as a storage site for the proteins is suggested in the case of the HIV1 Rev protein (Kubota et al., 1996), which shuttles between the cytoplasm and the nucleus, to enable the nuclear export of unspliced viral mRNAs (Malim et al., 1989a,b; Fischer et al., 1994; Kalland et al., 1994; Meyer and Malim, 1994; Richard et al., 1994). Despite its evident functional signi®cance as a pre-event of ribosomal assembly, nucleolar targeting of human ribosomal proteins has been characterized in only a few members, S6, S14 and L7a (Schmidt et al., 1995; Martin-Nieto and Roufa, 1997; Russo et al., 1997). Therefore, further analyses of the nucleolar targeting of other ribosomal proteins seem to be required to gain insights into not only ribosomal biogenesis, but also the role of the nucleolar localization of such RNAbinding transactivators. Ribosomes can be regarded as huge RNA-protein complexes, which are precisely organized to allow ecient protein biosynthesis. In eukaryotic cells, the biogenesis of ribosomes is carried out in the cell nucleolus (Sommerville, 1986; Warner, 1990), under the well-regulated expression of at least 80 ribosomal proteins, which are coordinated with cellular metabolism (Mager, 1988). The expression of eukaryotic ribosomal proteins has been shown to be controlled at both transcriptional (Agrawal and Bowman, 1987; Tasheva and Roufa, 1995) and post-transcriptional sites (Gritz et al., 1985; Mager, 1988; Aloni et al., 1992), predominantly at the level of translation. After expression in the cytoplasm, ribosomal proteins migrate into the nucleus, and reach the nucleolus. The transcription of pre-ribosomal RNAs (rRNA) also takes place in the nucleolus (Sollner-Webb and Mougey, 1991). Afterwards, the ribosomal proteins and rRNAs assemble together in the nucleolus. Eventually, assembled ribosomal subunits are exported out of the nucleus to be engaged in protein synthesis in the cytoplasm. Therefore, ribosomal proteins are initially imported into the nucleus towards the nucleolus, and later, return to the cytoplasm as a part of ribosome. However, as stated above, it has not been well-investigated how a vast number of human ribosomal proteins can accumulate Nuclear targeting of S25 S Kubota et al 1504 in the nucleolus, prior to ribosomal assembly, although it is one of the fundamental steps to clarify in the mechanism(s) of ribosomal biogenesis. Human ribosomal protein S25 (RPS25) is one of the highly basic members of human ribosomal proteins. Of note, the RPS25 gene was reported to be highly overexpressed in human leukemia (HL60) cells isolated for adriamycin-resistance, thus its cDNA was cloned and identi®ed from those cells (Li et al., 1991). Cloning and sequencing of the cDNA revealed RPS25 as a protein of 125 amino acid residues, which shares 56% homology to yeast ribosomal protein S31 (Li et al., 1991). A recent report assigned the chromosomal location of the RPS25 gene to 11q23.3 (Imai et al., 1994). Nevertheless, no further investigation has been reported until now. Recently, we have determined that RPS25 possesses common features with the NOS of the lentiviral RNA-binding transactivators, Rev proteins. Most interestingly, another ribosomal protein, L32 of yeast, has been shown to bind the structured target of its own mRNA to inhibit splicing (Li et al., 1995), which is quite similar to the function of the lentiviral Rev proteins. As ®rst steps in the functional characterization, we set up a system for the expression and analysis of RPS25, focusing on the primary preassembly event, pursued its nuclear/nucleolar targeting property, and demonstrated critical features involved in nuclear/nucleolar targeting, in comparison with other NOSs of lentiviral RNA-binding transactivators. Results Expression and characterization of an epitope-tagged RPS25 In general, it is dicult to raise a highly speci®c antibody against a ubiquitous protein in experimental animals. Indeed, all the recent similar works on human ribosomal proteins have employed either epitopetagging systems (Moreland et al., 1985; Michael and Dreyfuss, 1996; Martin-Nieto and Roufa, 1997) or even b-galactosidase fusions (Schmidt et al., 1995; Russo et al., 1997). Following the more well-accepted former strategy, an RPS25 expression vector, pcFLRPS25, was designed (Figure 1a). In our system, the expressed RPS25 was tagged with an eight residue epitope of a murine monoclonal antibody (Anti-FLAG M2), to enable its detection. Cos-1 cells were transfected with pcFL-RPS25, and the transient expression of the FLAG-RPS25 protein was examined by immunoblotting using the corresponding monoclonal antibody. As displayed in Figure 1b a distinct signal was speci®cally detected in the lysate of Cos-1 cells transfected with pcFL-RPS25. As expected, the apparent molecular weight of the expressed RPS25 in SDS ± PAGE was reasonably larger than the calculated molecular weight of RPS25 (14 KD), which is due to the addition of the FLAG epitope and linker residues (2 KD), and its high positive charge density. Immuno¯uorescence analyses of the transfected cells clearly demonstrated the expression and nuclear localization of RPS25, with a dominant accumulation in the cell nucleoli (Figure 1c and d) and also weak signals from the cytoplasm (Figures 1c and 3b). The expression of exogenous FLAG-RPS25 did not seem to cause major damage in nucleolar structure, which had been observed by the over-expression of another nucleolar protein, HIV-1 Rev (Nosaka et al., 1993). As illustrated in Figure 1e, the morphology of the nucleoli appeared to be unaltered with a variety of levels of FLAG-RPS25 expression. Similar experiments with human 293T cells also yielded the same results (data not illustrated). Since the added FLAG epitope [DYKDDDDK] contains a net negative charge, it may be of concern Nuclear targeting of S25 S Kubota et al 1505 Figure 1 Establishment of a FLAG-tagged RPS25 expression and detection system. (a) Structure of the plasmid for the expression of a FLAG-tagged RPS25 (FL-RPS25). Abbreviations; CMV: cytomegalovirus promoter, FLAG: the FLAG epitope-encoding region (in the stippled box). The cDNA for RPS25 was a composite of synthetic oligonucleotides (solid black box) and a DNA fragment from a cDNA clone of HeLa cells (hatched box with the name of the clone). Numbers over a box indicate codon numbers counted from the initiation codon of the cDNA. (b) Expression of RPS25 in Cos-1 cells detected by Western blotting analysis. An arrowhead indicates the speci®c signal. Positions of protein molecular weight standards are shown on the left of the panel. (c) Expression of RPS25 in Cos-1 cells detected by indirect immuno¯uorescence analysis. Non-transfected cells, around the three cells with positive signal, serve as internal negative controls. Some cells in this ®eld also represent the morphological variation among their two dimensional views. Apparent sizes of the cells and the nuclei are, frequently, not uniform in Cos-1 cell cultures. (d) Immuno¯uorescence analysis of the Cos-1 cells expressing RPS25 at high power magni®cation (61000), in comparison with phasecontrast photomicrographs of the same cells. (e) No major morphological alteration of the nucleoli by the expression of FLAGRPS25 at a variety of levels. Individual cells expressing low to high (top to bottom) levels of FLAG-RPS25 are displayed. (f) Veri®cation of the subcellular distribution of RPS25 by a Myc-tagged expression and detection system. Structure of the plasmid to express a carboxy-terminal Myc-tagged RPS25 is shown at the top. Abbreviations; SV: SV40 promoter, Myc: an oligonucleotide encoding a c-Myc epitope (Siomi and Dreyfuss, 1995; Dang and Lee, 1989; Kalderon et al., 1984). Numbers indicate the same items as shown in a. Immunostaining of the transfected Cos-1 cells was performed with a murine monoclonal antibody (9E10) against the Myc epitope and is illustrated Nuclear targeting of S25 S Kubota et al 1506 that such an acidic stretch of amino acid residues might have some interaction with a basic stretch of amino acid residues of RPS25, to aect its native subcellular localization. To verify if such an undesirable eect is caused by the FLAG epitope, control experiments were carried out. The other RPS25 expression plasmid, entitled pCHRPS25-Myc, was constructed to express an RPS25, tagged with an epitope of c-Myc protein at the carboxy terminus (Figure 1f). This Myc-tag can be detected by a murine monoclonal antibody, which has been previously utilized (Siomi and Dreyfuss, 1995; Dang and Lee, 1989). Transfection of pCHRPS25-Myc into Cos-1 cells and subsequent immuno¯uorescence analysis revealed the same subcellular distribution of the Myctagged RPS25 as the FLAG-tagged RPS25, suggesting that the FLAG epitope at the amino terminus should not aect the intact subcellular localization of RPS25. In addition, the eect of the FLAG epitope on another nucleolar-localizing protein, HIV-1 Rev, was evaluated. Of importance, the subcellular distribution of the Rev, which was tagged by the FLAG epitope, was demonstrated to be the same as the wild-type Rev protein (data not illustrated). The results of these control experiments strongly suggest the utility of the FLAG tag in studying nuclear-nucleolar targeting of proteins. corresponding region, for nuclear/nucleolar localization, since a majority of classical NLSs do contain such an arginine residue. In the mutant, pcFL-RPS25m1, the arginine codon was changed to a glycine codon (Figure 4a). Despite the mutation, in comparison with the wild-type, no signi®cant change in subcellular localization was observed (Figure 4b), which indicates that the arginine residue itself is not critical for either nuclear, or nucleolar, localization of RPS25. In contrast to the m1-mutation, extensive mutation of ten residues, including the arginine, to unrelated eight residues (two residues deleted) has completely abolished the dominant nuclear targeting of RPS25. This second mutant, entitled pcFL-RPS25m2, expressed an RPS25 protein with a diuse intracellular distribution, both inside or outside of the nuclear membrane (Figure 4). Apparently, the immunofluorescent signal from and around the nucleus frequently a The region required for nuclear localization of RPS25 To initially estimate which motif is necessary for the nuclear localization of RPS25, six deletion cDNA mutants of the FLAG-RPS25 protein were constructed. Three constructs were generated to express the RPS25 mutants with amino-terminal deletions, and the others for carboxy-terminal deletion mutants (Figures 2a and 3a). These constructs were transfected and characterized in the same manner as the wild-type protein. Unexpectedly, two amino-terminal deletion mutants, dN1 and dN2, demonstrated approximately the same subcellular distribution as the wild-type by immuno¯uorescence analyses (Figure 2b). Namely, the ®rst lysine-rich stretch (residues 4 ± 20), which contains nine lysine residues out of 17, was entirely dispensable for nuclear/nucleolar targeting. Extension of the deletion up to residue 43 completely abolished the proper nuclear/nucleolar localization of RPS25, and a diuse distribution throughout the cell was observed (dN3: Figure 2b). Similar analyses with the other three carboxy-terminal deletion mutants have revealed that removal of the 53 carboxy-terminal residues did not signi®cantly alter the nuclear/nucleolar targeting of RPS25 (dC1 and dC2: Figure 3b), whereas the other mutant which lacks the 94 carboxy-terminal residues lost nuclear targeting ability (dC3: Figure 3b). Therefore, the nuclear localization signal (NLS) was thought to reside between residues 21 and 72 of the RPS25 protein. Extensive analyses were further carried out by sitedirected mutagenesis of RPS25 cDNA. Since there is no NLS-like basic residue stretch between residues 42 and 72, we suspected a potential NLS/NOS between residues 21 and 42 of RPS25. As such, the mutagenesis was targeted to this region. The ®rst point mutant was constructed to examine the requirements of the arginine residue, which is the only one found in the b Figure 2 Amino-terminal deletion mutants of RPS25, and their subcellular localization. (a) Schematic representation of the primary structures of the deletion mutants. At the top, the amino acid sequence of the trimmed area is shown in single letter codes, in which basic residues are shown in bold letters. Numbers above boxes denote residue numbers counted from the initiation methionine of RPS25. Deleted regions are represented as solid lines. (b) Immuno¯uorescence analysis of the amino-terminal deletion mutants, by an anti-FLAG antibody. The names of the expressed and detected proteins are shown above the photomicrographs Nuclear targeting of S25 S Kubota et al a exclusion (Figure 4b and c). The distinct nuclear localization and the intensity of the nuclear signal in each cell, was not at all aected by this minor deletion, which indicates the continuous cluster of lysine residues plays a role, not in the nuclear, but rather in the nucleolar localization of the RPS25 protein. Nuclear/nucleolar targeting signal of RPS25 b Figure 3 Carboxy-terminal deletion mutants of RPS25 and their subcellular localization. (a) Representation of the primary structures of the deletion mutants. Numbers above boxes denote residue numbers counted from the initiation methionine of RPS25. Deleted regions are represented as solid lines. At the end of each line, some extra residues, which were added through construction procedures, are also shown in single letter codes. (b) Detection of the carboxy-terminal deletion mutants by immuno¯uorescence analyses. The names of the deletion mutants are shown above the photomicrographs seems to be relatively stronger, in the case of mutants m2, as well as dN3 and dC3, which is probably due to the even distribution inside and outside of the nuclei and the three-dimensional shape of the cells (Figures 2, 3 and 4). Thus, at least some of the ten residues of RPS25 are required for its nuclear localization. Nucleolar targeting property of RPS25 As observed in Figure 2b, the removal of aminoterminal 43 residues caused nucleolar exclusion as well as the loss of nuclear targeting in RPS25. Furthermore, creation and characterization of another RPS25 mutant yielded yet another insight into a sequence that aects, not the nuclear translocation, but only the nucleolar accumulation of RPS25. The mutant, RPS25m3, harbors a deletion of three continuous lysine residues in this speci®c area (Figure 4a). Expression and detection of this mutant showed signi®cant attenuation of nucleolar accumulation caused by the mutation, despite no evident nucleolar In order to identify the NLS/NOS of RPS25, three dierent fusion constructs were generated. As an initial step, the 41 amino-terminal residues with the FLAG epitope, which includes two NOS-like motifs, were fused to a short fragment of the rat preproinsulin two gene at the amino terminus (Figure 5a). Expression and immuno¯uorescence analysis demonstrated a clear nuclear/nucleolar localization of this small fusion protein, FL-NOQDINS (Figure 5b). As a control, a similar fusion construct with the corresponding region from an RPS25 mutant, pcFL-RPS25m2 (Figure 4), the nuclear/nucleolar targeting property of which is inactivated by mutation, was prepared (Figure 5a). A diuse intracellular distribution pattern was observed with transient expression and immuno¯uorescence analysis of the mutant plasmid, entitled pcFLNOQm2DINS. These ®ndings indicate a NOS within residues 1 ± 41 of RPS25. In order to further examine the ability of the amino terminal 41 residues to function as a NOS for a larger protein, E. coli bgalactosidase was next chosen as a fusion counterpart for the RPS25 fragment (Figure 6a). The b-galactosidase is a bacterial 116 kD protein, which is thus found outside of the nucleus when it is expressed in eukaryotic cells. This was also reproducible in our laboratory (Figure 6b: CH110). In contrast, once the 41 amino-terminal residues of RPS25 were fused to its amino terminus, the fusion protein was transported into the nucleus, and also accumulated in the nucleolus, although at a reduced level compared to the insulin fusion protein (Figure 6b: CHNOQ). This ®nding indicates that the RPS25 fragment functioned as an active NLS/NOS at the amino terminus. The other RPS25-b-galactosidase fusion protein, with residues 1 ± 21 of RPS25, failed to localize in the nucleus, in concordance with the mutational analysis of RPS25 (Figure 6b: CH-NOH). Although the b-galactosidase fusion system has been widely accepted as a marker for nuclear targeting of proteins, it has been suspected that this prokaryotic protein partially migrates through the nuclear membrane, which may make results ambiguous (Siomi and Dreyfuss, 1995; Kalderon et al., 1984). Therefore, it may not be a perfect reporter for nuclear/nucleolar targeting. Thus, to gain more reliability using a dierent system, we constructed another fusion-protein expression plasmid, in the backbone of chicken muscle pyruvate kinase (PK). PK has been shown to be stably excluded from the nucleus, hence the utility of a myc-tagged PK cDNA, as a marker gene for nuclear targeting, has already been established (Siomi and Dreyfuss, 1995; Dang and Lee, 1989; Kalderon et al., 1984). Immunofluorescence analyses of PK, expressed in Cos-1 cells, has shown extremely clear nuclear-exclusion, which seems more striking and reliable than b-galactosidase (Figure 7). 1507 Nuclear targeting of S25 S Kubota et al 1508 Based on the data from the amino-terminal deletions of RPS25, RSP25-insulin and RPS25-bgalactosidase fusion proteins, we hypothesized that residues 25 ± 41 of RPS25 should involve a NLS/NOS. For veri®cation, we fused the RPS25 fragment to the carboxy terminus of the Myc-PK protein (Myc-PKNOR) (Figure 7a). During the construction procedure of pcDNA3Myc-PK-NOR, the carboxy terminal 86 residues of PK were removed. In such a case, although unlikely, there still was a formal possibility that removal of the 86 residues might aect the subcellular localization of PK itself. Therefore, as a further control experiment, another plasmid used to express PK, with a carboxy terminal truncation of 87 residues, was constructed. Immunostaining of these a expressed proteins demonstrated that, as anticipated, the fusion protein was capable of migrating into the nucleus and accumulating in the nucleolus, while the PK with carboxy terminal deletion revealed the same intracellular distribution pattern, with clear nuclear exclusion, as the full-length PK (Figure 7b). The subcellular distribution of Myc-PK-NOR was not precisely uniform throughout all of the expressing cells. Some cells demonstrated complete nuclear accumulation, while others still revealed a cytoplasmic signal, as well as a nuclear/nucleolar signal (Figure 7). However, it was evident that residues 25 ± 41 of RPS25 acted as a NLS/NOS at the carboxy terminus of Myc-PK, to actively guide this cytoplasmic protein into the nucleus and the nucleolus. c b Figure 4 Site-directed mutational analyses targeted to the region which is critical for the nuclear and nucleolar targeting of RPS25. (a) Structure of the mutants. Primary structures of the native RPS25 with the FLAG epitope, and mutants, are represented. Numbers over a box denote residue numbers, as seen in Figures 1 and 2. Amino acid residues around the aected area are shown in single letter codes. Basic and mutated residues are shown in bold letters. Deletion and amino acid residue substitutions are represented as a solid bar, and a hatched box, respectively. (b) Subcellular distribution of the mutants in comparison with the wild-type protein. Abbreviated names of intact and mutant proteins are shown above the photomicrographs. (c) Attentuated nucleolar accumulation of the mutant, m3. Two additional examples with phase-contrast photomicrographs are shown Nuclear targeting of S25 S Kubota et al 1509 a b Figure 5 Insulin fusions with the amino-terminal fragment of RPS25 and its mutant. (a) Primary structures of the fusion constructs. A crossed box with `DINS2' represents a carboxyterminal fragment of rat preproinsulin 2. Numbers over the boxes and other abbreviations represent the same as in Figure 1. Mutations introduced into FL-NOQm2DINS are described in the middle of this panel in an analogous manner to Figure 4a. The code `wt' stands for the wild-type RPS25. The insulin fusion constructs express proteins with predicted molecular weights of 7.1 ± 7.3 kD. (b) Subcellular location of the fusion proteins illustrated in a. The names of the fusion proteins are displayed above each photomicrograph Figure 7 Pyruvate kinase fusion with the NOS of RPS25. (a) Primary structure of Myc-tagged chicken muscle pyruvate kinase and its derivatives. PK and RPS are the abbreviations for pyruvate kinase, and RPS25, respectively. Numbers above and below boxes represent residue numbers corresponding to PK and RPS25, respectively. (b) Subcellular distribution of Myc-tagged wild-type or truncated PK, with or without the NOS of RPS25 (residue 25 ± 41). Two representative views are shown for the Myc-tagged RPS25-PK-fusion protein. The name of the expressed protein is given above each photomicrograph Discussion Figure 6 Fusion proteins of b-galactosidase with amino terminal fragments of RPS25. (a) Primary structures of the b-galactosidase fusion proteins. Numbers over the boxes represent residue numbers of RPS25, counted from the initiation methionine. Crossed boxes indicate assemblies of gene products of gpt, trpS, and lacZ of E. coli which are involved in the parental and control plasmid, pCH110 (Pharmacia). The lacZ fusion constructs express proteins with predicted molecular weights of 116 ± 120 kD. (b) Subcellular location of the fusion proteins illustrated in a. The names of the fusion proteins are displayed above each photomicrograph The FLAG-epitope tagging system has been initially developed and utilized to characterize the nuclear/ nucleolar targeting of human RPS25, and enabled us to perform further mutational analyses to dissect its structure-function relationships. Using this system, we have demonstrated the distinct nuclear/nucleolar localization of RPS25, analysed its sequence requirements, and identi®ed a new NOS in RPS25. This system is shown to be excellent in analysing the ®rst dynamic event of the long journey of RPS25 ± from ribosome, back to ribosome, since nuclear/nucleolar targeting itself is not to be aected by the FLAGRPS25 accumulation that occurs after this initial event. The utility of this system for the analysis of late events, after nucleolar-targeting of RPS25, is not yet established, although our preliminary data have suggested considerable cytoplasmic FLAG-RPS25 (Figure 3b and subcellular fractionation analysis-data not illustrated), and a signi®cant, but limited amount of FLAG-RPS25 incorporated in puri®ed ribosomes (data not illustrated). The strong nuclear-dominant distribution of FLAG-RPS25 may have been caused by the combination of Cos-1 (or 293T) cells and pcFL-RPS25 plasmids. This system yields high levels of transient protein expression, through the intracellular amplification of plasmids supported by endogenous SV40 large T antigen. However, along with cell growth, ribosomes Nuclear targeting of S25 S Kubota et al 1510 are not synthesized in an unlimited fashion. Overexpression of exogenous FLAG-RPS25 should not increase, or even may cause possible inhibitory eects on overall ribosomal biogenesis; for synthesis of other components are coordinately controlled. Consequently, only a limited portion of FLAG-RPS25 could be incorporated into ribosomes in the present system, which caused strong nuclear accumulation of the remaining FLAG-RPS25. Currently, we are improving this system to obtain lower levels of stable exogenous RPS25 expression for more natural analysis of postnucleolar events of RPS25, as previously performed for another ribosomal protein (Martin-Nieto and Roufa, 1997). In those studies, comparative examination of the distribution of nucleolar components, in the absence or presence of exogenous RPS25 and its mutants, will be involved as well, in order to estimate possible eects of those proteins on nucleolar functions. In this study, a short (17 residue) and distinct NOS has been identi®ed. Such a compact NOS is not commonly found in eukaryotic cellular proteins. More generally, nucleolar localizing properties seem to be acquired by cooperation of several functional domains (Russo et al., 1997; Peculis and Gall, 1992; SchmidtZachmann and Nigg, 1993; Yan and MeÂleÁse, 1993; Michael and Dreyfuss, 1996). Among these proteins, no common or prototypic NOS motif has been presented, suggesting the complexity of the nucleolar localizing activity. Also, investigations of the other human ribosomal proteins revealed such features in their nucleolar targeting. In S6, portions of approximately 60 residues were shown as minimum continuous fragments to confer nuclear/nucleolar targeting, in which two independent basic amino acid stretches seemed to cooperate (Schmidt et al., 1995). In the case of L7a, the amino-terminal 100 residue fragment was necessary and sucient for the nuclear/nucleolar targeting, which was also enabled by the combination of 17 residue amino terminal peptide and another portion (residue number 52 ± 100) in the fragment (Russo et al., 1997). However, most interestingly, several small and distinct NOSs, which are similar to the NLS/NOS of RPS25, have been identi®ed in retroviral RNA-binding trans-regulator proteins. In particular, the Rev protein of HIV-1 possesses a NOS/RNA-binding domain, which contains a similar motif to the NLS/NOS of RPS25 (Kubota et al., 1989) (Figure 8a). HIV-1 Rev is a 116 amino acid nucleocytoplasmic shuttling protein with a predominant localization in the nucleolus (Malim et al., 1989a; Kalland et al., 1994; Meyer and Malim, 1994; Richard et al., 1994), binds to a structured target on viral unspliced and singly-spliced mRNAs, and enables these incompletely-spliced RNAs to accumulate in the cytoplasm for the expression of structural and enzymatic viral protein precursors (Malim et al., 1989b; Fischer et al., 1994; Zapp and Green, 1989; Heaphy et al., 1990; Malim and Cullen, 1991). The putative NLS/NOSs from the Rev proteins of HIV-2 and simian immunode®ciency virus (SIV)mac also retain the basic amino acid residue motif (Figure 8a). Although belonging to another class of retroviruses, human T-cell leukemia/lymphoma virus type I (HTLVI) possesses a functional equivalent of Rev, which is termed Rex (Seiki et al., 1988). A small and distinct NOS was ®rst identi®ed in this protein (Siomi et al., Figure 8 Comparative alignment of the NOS of RPS25 with other ribosomal and retroviral proteins. Numbers over each amino acid sequence represent residue numbers of each protein. (a) A basic residue motif shared among the NOSs of RPS25 and the Rev proteins of primate lentiviruses. NOSs of simian immunode®ciency virus (SIV)mac and HIV-2 are putative. Identical residues are indicated by bars, whereas non-identical basic residues are linked by dots. Shared motif is indicated at the bottom. `B' stands for a basic residue. Lower case letters represent the residues shared by only three proteins. (b) Sequence comparison of identi®ed NOSs. Continuous clusters of three or more basic amino acid residues are boxed. Other basic residues are underlined. Total number of basic residues involved in each NOS is also shown at the right. (c) Conservation of the NOS sequence of human RPS25 in yeast RPS31, and a common basic residue motif between a basic domain of Visna virus Rev and the non-conserved, amino-terminal basic residue cluster of RPS25. Identical residues and non-identical basic residues are indicated by bars and dots, respectively. The region of a basic residue motifs shared by Visna virus Rev and RPS25 and by the NOS of human RPS25 and its homologue in yeast RPS31, are framed by the large boxes. Sources of amino acid sequences are as follows; human RPS25 and yeast RPS31: (Li et al., 1991) HIV-1 (strain BRU:GenBank accession #K02013), HIV-2 (strain ROD:GenBank accession #M15390), and SIVmac (strain MM142:GenBank accession #M16403): Visna virus: (Schoborg and Clements, 1994) 1988). Such a compact NOS can be found in the other RNA-binding trans-regulator of HIV-1, Tat (Dang and Lee, 1989; Siomi et al., 1990). The major function of Tat is to increase the expression of all viral genes (Sodroski et al., 1985), through the binding to its structured RNA target on viral gene transcripts. Taken together, these compact NOSs are commonly characterized by the involvement of a functional NLS herein, hence they do not require cooperation of the other functional domain for nuclear targeting. In their primary structures, although there are signi®cant variations in sequence, a few common features can be pointed out. First, all the identi®ed compact NOSs retain continuous stretches of basic residues; one continuous stretch of at least four basic residues, or two continuous stretches of three basic residues. Second, the total number of basic residues is more Nuclear targeting of S25 S Kubota et al than nine (Figure 8b). These features are necessary but not sucient for nucleolar targeting property. For example, the other basic residue stretch at the amino terminus of RPS25 and a highly basic region in RPS6 which are not fully demonstrated as NOSs themselves (Schmidt et al., 1995), also possess these characteristics. However, although high positive charge density is not sucient to provide nucleolar targeting ability, accumulated mutational analysis data on the de®ned compact NOSs have already pointed out the critical requirement of these features for their function (Siomi et al., 1988; Kubota et al., 1989; Dang and Lee, 1989; Malim et al., 1989a). Also, in our study, removal of the continuous lysine residues in the NOS caused signi®cant loss of nucleolar accumulation. Therefore, such NOSs, as shown in Figure 8, may be regarded as members of the compact NOS family, a particular subclass of NOSs whose majority are composed of more than one domain which do not juxtapose. Unlike nuclear localization, the mechanisms for nucleolar localization appear to be complex. Since no membrane exists around the nucleolus, it has been believed that direct or indirect binding to nucleolar components should play central roles in nucleolar localization. Thus, nucleolar localization can be accomplished by three dierent types of interactions: nucleolar or nucleolar-associated protein-mediated capture, rDNAmediated capture, and nucleolar RNA-mediated capture of nucleolar localization signals. In the case of RPS25, we hypothesized that the nucleolar localization may be mediated by RNA-protein interactions, as the NOS of RPS25 shares common features with the NOS of RNA-binding trans-regulators, which are thought to function also as RNA-binding domains (Tan et al., 1993). RNA binding analyses are being performed to investigate interactions between RPS25 and speci®c RNAs. At present, preliminary results indicate direct binding of the NOS of RPS25 to ribosomal RNAs (data not illustrated). Our characterizations of RNA binding properties may clarify which type of molecular interaction(s) plays a central role in the nucleolar localization of RPS25. Yeast ribosomal protein S31 (RPS31) is a homologue of human RPS25, with 56% homology in amino acid sequence (Li et al., 1991). Comparison of the primary structures of these two ribosomal proteins reveals interesting characteristics. The NOS region, particularly the basic residue motif, is quite wellconserved between these proteins (Figure 8c), which suggests an indispensable functional role of this region ± probably in nuclear/nucleolar targeting for these moieties. Nevertheless, only limited homology can be found in the amino-termini of these proteins. Most of the lysine residues in this area (residues 4 ± 20) of human RPS25 are absent in the corresponding area of yeast RPS31. Together with the fact that this region is not de®nitely required for the nuclear/nucleolar localization in human RPS25, it is hypothesized that the region may have evolved towards an accessory function, that is required only in higher eukaryotes. Interestingly, this section of RPS25 shares a distinct motif of basic residues, with the putative RNA-binding domain of the Visna virus Rev protein (Figure 8c), which has been shown to be required for multimerization and binding to the target RNA of the Visna virus Rev protein (Tiley et al., 1991). RNA-binding by a ribosomal protein is one of the key interactions involved in the control of ribosomal protein expression. Bacterial ribosomal proteins S4 and S15 bind to their own mRNAs and repress translation (Tang and Drapier, 1990). In yeast, ribosomal protein L32 (RPL 32) binds to a purine-rich stem loop of its unspliced mRNA, and blocks its splicing to prevent over-production of RPL32 (Li et al., 1995; Eng and Warner, 1991), which has similarities to the interaction of the Rev proteins and the structured targets on their unspliced viral mRNAs. Since introns are rare in yeast, it is plausible that this kind of RNA-binding regulatory function may have developed more extensively in higher eukaryotes. As such, it can be suggested that the `Visna Rev-like' region may have been furnished as an alternative RNA-binding domain for a speci®c regulatory function, acquired during evolution. It can also be suggested that the evolutional pathway of RNA-binding ribosomal proteins, such as yeast L32, may be similar to the Rev proteins of HIVs and other lenti-retroviruses. Comparative investigation of the ribosomal proteins of higher eukaryotes and RNA-binding lentiviral proteins may yield comprehensive insights into the understanding of both critical families of proteins. Materials and methods Construction of an epitope-tagged expression system of RPS25 Basic RPS25 expression plasmid The major portion of RPS25 cDNA, which encodes the carboxy-terminal 84 amino acid residues, was obtained from an isolated clone of lZAP II cDNA library of HeLa cells (Stratagene). The corresponding cDNA fragment was ampli®ed through a polymerase chain reaction (PCR), using an RPS25-speci®c sense primer with a ¯anking HincII site (RPC: 5'GACGTCGACAAGCTCAATAACTTAGTCTTG-3') and an RPS25-speci®c anti-sense primer with a ¯anking XhoI site (RPX: 5'-GAGCTCGAGTACAGCTGGTTGGACCTATTC-3'). The obtained amplicon was digested with HincII and XhoI, subcloned into pBluescriptSK(7) (Stratagene) and the resultant plasmid was designated, pBSRPS-C. Between the upstream HindIII site and the HincII site of pBSRPS-C, four synthetic oligonucleotides were assembled, providing the amino-terminal 41 codons of RPS25 (Li et al., 1991) to modify pBSRPS-C into pBSRPS-F, that contains a full-length RPS25 cDNA. Eukaryotic expression machinery was provided by pcREVM1 (Malim et al., 1989a). At the unique BglII site, which is immediately downstream of the initiation codon, double-stranded synthetic oligonucleotides encoding the octapeptide [DYKDDDDK] entitled FLAG epitope (IBI/Kodak) was inserted into pcREVM1. Utilizing the internal HindIII site inscribed in the FLAG oligonucleotides, the HIV-1 Rev cDNA was replaced by the RPS25 cDNA through HindIII ± XhoI digestion and fragmentexchange with pBSRPS-F. The resultant plasmid, pcFLRPS25, contains RPS25 cDNA with the FLAG peptide at the amino terminus, which is driven by the cytomegalovirus (CMV) promoter. The structure of pcFLRPS25 is illustrated in Figure 1a. Myc-tagged RPS25 expression plasmid The full-length cDNA of RPS25 was re-ampli®ed from pcFL-RPS25 by PCR, in order to construct another plasmid for the expression and characterization of RPS25 with a dierent tag-epitope. A sense primer (NRP: 5'-TAGGTACCAAGCTTAATGCCGCCTAAG-3'), which anneals from the HindIII site in the FLAG region towards the RPS25 cDNA in pcFL-RPS25, and an anti-sense primer (RPE: 5'- 1511 Nuclear targeting of S25 S Kubota et al 1512 GGGGTACCGCATCTTCACCGGC-3') with an internal KpnI site which is located immediately upstream of the termination codon of RPS25 cDNA, were synthesized and used for PCR ampli®cation. The amplicon was puri®ed, digested by HindIII and KpnI and ligated between the unique HindIII and KpnI sites of pCH110 (Pharmacia). After digestion of this plasmid, entitled pCHRPS25, with KpnI and MluI, a 4.7 kbp fragment was puri®ed and ligated with a double-stranded synthetic oligonucleotide, encoding a c-Myc epitope (Siomi and Dreyfuss, 1995) to be pCHRPS25-Myc. As shown in Figure 1f, this construct contains an RPS25 cDNA with a c-Myc epitope at the carboxy terminus, which is driven by an SV40 promoter for eukaryotic expression. Deletion mutants Amino-terminal deletion mutants were obtained, utilizing PCR. Three deletion mutants of RPS25 cDNA, each of which lacks 10, 20, or 43 amino-terminal codons, respectively, were ampli®ed by corresponding speci®c sense primers and RPX (see the ®rst subsection of Materials and methods) from pcFL-RPS25. Plasmids pcFL-RPS25 dN1, dN2, and dN3 were established by substituting the RPS25 cDNA with these PCR-ampli®ed cDNA mutants, utilizing the ¯anking restriction enzymatic sites (Figure 2a). Two carboxy-terminal deletion mutant expression vectors were constructed by enzymatic digestion and re-ligation of the basic plasmids, pBSRPS-F or pcFL-RPS25. RPS25 cDNA contains two internal SacI sites. Removal of a 40 bp fragment between these SacI sites, by digestion and religation of pBSRPS-F, resulted in a frame shift at codon 91, and an extra lysine and termination codons downstream. The XhoI ± HindIII fragment of the pBSRPS-F variant, which contained the deleted cDNA, was exchanged with the corresponding fragment of pcFL-RPS25, and the resultant plasmid was entitled pcFL-RPS25dC1. Similarly, using internal three PvuII sites in RPS25 cDNA, pcFL-RPS25dC2 was constructed. Digestion and self-ligation of the major fragment of pcFL-RPS25 resulted in the removal of 150 bp and 30 bp internal fragments, which caused a frame-shift at codon 72, which was followed by two extra and one nonsense codons. Plasmid pcFL-RPS25dC3 was obtained via PCR. An RPS25 cDNA fragment, which encodes the amino terminal 34 amino acid residues, was ampli®ed from pcFL-RPS25 by corresponding speci®c primers with a HindIII site upstream and an XhoI site downstream. The cDNA fragment was exchanged with the full-length RPS25 cDNA in pcFL-RPS25 to obtain pcFL-RPS25dC3. In this construct, four extra codons and a nonsense codon follows codon 34 of RPS25. Structures of these carboxy-terminal deletion mutant proteins are displayed in Figure 3a. Internal site-directed mutants For the construction of pcFL-RPS25m1, a point mutation was introduced into pcFL-RPS25 by a PCR-mediated technique. NRP was utilized for the sense primer here. As an anti-sense primer, NRG: (5'-ACGCGTCAGGTACCGGCCCAACTTTGCC3') with a mutation at codon 41 and a HaeIII site, was synthesized. PCR-ampli®cation of pcFL-RPS25 with these primers yielded a mutated cDNA fragment encoding residues 1 ± 41 of RPS25, with the 41st codon of a glycine, instead of the arginine codon. The PCR product was digested by HindIII and HaeIII, and ligated with HindIII ± HincII digested pBSRPS-C to obtain a full-length cDNA with the [41R-G] mutation. The mutant cDNA was built in the same backbone as pcFL-RPS25, to become pcFL-RPS25 m1. In the process of verifying the nucleotide sequence of the expected pcFL-RPS25dC1, one of the molecular clones was shown to harbor unanticipated spontaneous minor deletions in RPS25 cDNA. Five nucleotides [AAGAA] at codons 31 ± 34, and a single nucleotide [G] at codon 41, were missing. These deletions caused an introduction of eight unrelated amino acid residues between residues 32 ± 41 in the expressed protein, instead of ten intact residues. This molecular clone was designated as pcFL-RPS25dC1/m2. Since the plasmid harbors a carboxy-terminal deletion, as well as the internal mutation in RPS25 cDNA, a 0.6 kbp NcoI ± MscI fragment of pcFL-RPS25dC1/m2, which included the mutated area, was exchanged with the corresponding fragment of pcFLRPS25 to obtain pcFL-RPS25m2 lacking the carboxyterminal deletion. Deletion of three internal codons for [31KKK] was introduced by a ligation-mediated, multiple-step PCR technique. A 57 base plus-strand oligonucleotide with a 3' ¯anking HaeIII site, which encodes residues 21 ± 41 except 31 ± 33 of RPS25, was synthesized. The oligonucleotide was used to generate a double-stranded form, using a hemi-PCR with a short primer (NRM: 5'-TTACGCGTCAGGTACCGGCCGAACTTTGCC-3') complementary to the 3' end of the oligonucleotide, using Pfu polymerase (Stratagene) to avoid extra additions of nucleotides at the 3' end. The double-stranded oligonucleotide was ligated with the other double-stranded oligonucleotides encoding residues 1 ± 20 of RPS25, which had been used for the construction of intact RPS25 cDNA. Fragments, that were ligated in order, were then selectively ampli®ed by PCR with NRP and NRM primers. The amplicon was digested with HindIII and HaeIII and subcloned between the HindIII ± HincII sites of pBSRPSC. Transfer of the mutant cDNA to the eukaryotic expression backbone was carried out, as described for the wild-type. The obtained plasmid was designated as pcFLRPS25m3. Structures of the mutants described herein are illustrated in Figure 4a. All the RPS25 cDNA mutants were sequenced from both directions to verify the proper introduction of mutations. DNA sequencing was performed via an automated DNA sequencing system (Applied Biosystems). Structures of RPS25 fusion-protein expression plasmids RPS25-insulin fusion construct Since the poly(A) addition signal of pcFL-RPS25 originates in the rat preproinsulin two gene, the plasmid contains the fragment of the 3' end of the protein-encoding region of rat preproinsulin two gene, as well as the poly(A) signal itself. By taking advantage of this feature, a fusion cDNA was created. Ligation of the assembled oligonucleotides with an upstream HindIII site and a downstream HaeIII site, which encodes residues 1 ± 41 of RPS25, with a HindIII ± PvuII digested 3.5 kbp fragment of pcFL-RPS25, yielded a fused open reading frame that consists of the FLAG epitope, 41 amino-terminal residues of RPS25, and 25 carboxy-terminal residues of rat preproinsulin 2. Schematic representation of the fusion protein is displayed in Figure 5a, as the translational product of pcFL-NOQDINS. As a control, a corresponding region of the cDNA of an RPS25 mutant with abolished nuclear/nucleolar localization was isolated and inserted in the same backbone as pcFLNOQDINS. An anti-sense primer for amplifying the mutated cDNA fragment, corresponding to residues 1 ± 41 of RPS25, from pcFL-RPS25m2, was designed and synthesized (NRMM: 5'-GGGGCCAACTTTGCCTTTGGAC-3'). PCR ampli®cation of pcFL-RPS25m2 with NRP and NRMM yielded an RPS25m2 cDNA fragment corresponding to residues 1 ± 41 of the wild-type (Figure 5a). The insulin fusion construct was obtained through the same construction procedure, as described above, in which the PCR product was used, instead of the assembled oligonucleotides. In the resultant fusion construct, pcFL-NOQm2DINS, the mutation was copied from pcFL-RPS25m2, with an alternative codon of tryptophan, instead of serine, at the forty-®rst codon of the wild-type RPS25. RPS25-b-galactosidase constructs The parental plasmid for the expression of the bacterial lacZ gene was Nuclear targeting of S25 S Kubota et al commercially available (pCH110; Pharmacia). Two complementary oligonucleotides encoding residues 1 ± 21 of RPS25 were synthesized and hybridized to each other to form a double-stranded DNA fragment, with HindIII and KpnI over-hangs. This fragment was ligated with a 6.9 kbp HindIII ± KpnI fragment of pCH110 to establish pCHNOH. Similar procedures were applied to obtain pCHNOQ with a HindIII ± KpnI digested PCR amplicon, which had been ampli®ed by NRP and NRM primers from pcFLRPS25. The expressed protein from pCH-NOQ contains residues 1 ± 41 of RPS25 and b-galactosidase (Figure 6a). Myc-tagged pyruvate kinase-RPS25 fusion construct The construction of pcDNA3Myc-PK has been described previously (Siomi and Dreyfuss, 1995). A sense primer with a ¯anking KpnI site and an anti-sense primer with a ¯anking NotI site were designed to create a DNA fragment, which encodes residues 25 ± 41 of RPS25. The fragment was ampli®ed by PCR from pcFL-RPS25, digested by KpnI and NotI and substituted with a 270 bp KpnI ± NotI fragment of pcDNA3Myc-PK. The resultant plasmid was designated, pcDNA3Myc-PK-NOR. An alternative negative control, pcDNA3Myc-PK-C, was constructed as follows. Original pcDNA3Myc-PK was double-digested by KpnI and XbaI and the major fragment was puri®ed. The isolated DNA fragment was treated with T4 DNA polymerase for 15 min at 168C to trim and ®ll the KpnI and XbaI overhangs, respectively, then self-ligated by T4 ligase. This blunting ligation procedure removed the tyrosine codon for residue 443 of PK, and a termination codon that originated in the XbaI overhang was placed in frame immediately downstream, which was con®rmed by DNA sequencing. Primary structures of the expressed proteins are illustrated in Figure 7a. Cell culture and DNA transfection Cos-1 and 293T cells were maintained in Dulbecco's modi®ed minimum essential medium (D-MEM) supplemented with 10% fetal bovine serum (FBS) at 378C. A liposome-mediated DNA transfection system (LipofectAMINE: GIBCO/BRL) was utilized for the immunofluorescence analyses. Twenty-four hours prior to transfection, 2.56104 Cos-1 or 293T cells were seeded into each well of Lab-Tek eight-well chamber slides (Nunc, Inc.). DNA transfection was carried out following a manufacturer's optimized protocol with 200 ng of each plasmid DNA. For immunoblotting, DEAE-dextran-mediated DNA transfection was performed, according to an established procedure (Cullen, 1987). In this case, Cos-1 cells (26105 cells) were prepared in 35 mm tissue culture plates, 24 h before transfection. In each experiment, 500 ng of plasmid DNA was used. Antibodies A murine monoclonal antibody for the detection of the FLAG tag was purchased from IBI/Kodak (Anti FLAG M2), and was used in immuno¯uorescence analyses at a dilution of 1 : 100, or in immunoblotting at a dilution of 1 : 1000, respectively. A murine monoclonal anti-b-galactosidase IgG was obtained from Boehringer-Mannheim Biochemicals, and used at a dilution of 1 : 10 000 for immuno¯uorescence studies. The anti-Myc 9E10 murine monoclonal antibody was kindly provided by Dr H Siomi (University of Pennsylvania). For immuno¯uorescence, 9E10 was applied to the ®xed cells at a dilution of 1 : 1000. Fluorescein-isothiocyanate (FITC)-conjugated anti-murine IgG serum, FITC-conjugated anti-rabbit IgG serum, and horseradish peroxidase (HRP)-conjugated antimurine IgG serum were purchased (Sigma), and used as detector antibodies for immuno¯uorescence (1 : 100) and immunoblotting (1 : 1000), respectively. Immunoblotting Forty-eight hours after transfection, cells were washed once and harvested in 500 ml of a RIPA buer (Adachi et al., 1992). Proteins in 360 ml of the clari®ed lysate were precipitated in 80% acetone, dissolved in 40 ml of 16sodium dodecyl sulfate (SDS) sample buer (NOVEX) with 2.5% 2-mercapto-ethanol and separated using SDSpolyacrylamide gel electrophoresis (PAGE) through a 14% pre-cast gel (NOVEX). Separated proteins were transferred onto a PVDF membrane (Polyscreen : Dupont NEN), blocked in 5% nonfat dry milk, and probed with the antiFLAG antibody for 1 h in 1% BSA/0.05% Tween 20/PBS at room temperature. After extensive washes, protein visualization was carried out following a manufacturer's protocol, using a commercial kit (Renaissance: Dupont NEN) with the HRP-conjugated antiserum. Indirect immuno¯uorescence Forty-eight hours after transfection, cells were washed once with PBS, ®xed in 3.5% formaldehyde/PBS for 20 min and permeabilized in 0.1% NP-40/PBS for 10 min at room temperature. Incubation in primary or secondary antibodies in 3% BSA/PBS was performed for 1 h at 378C, followed by extensive washes with PBS, as described previously (Kubota et al., 1991). After the last wash, cells were placed in a mounting medium (Omni¯uor: Virostat), and viewed with an epi¯uorescence microscope (Olympus). Generally, transfection eciency was approximately 15%, as analysed by immuno¯uorescence microscopy. All the photomicrographs were taken under a magni®cation of6400, with the exception of Figure 1, in which a few of these were taken at6200 or61000. Each photomicrograph represents a vast majority of the cells with similar subcellular localization, unless otherwise stated on the text. Acknowledgements The authors wish to thank Dr Haruhiko Siomi for the murine monoclonal antibodies and pcDNA3Myc-PK, Dr Bryan R Cullen for the anti-HIV-1-Rev serum, Dr Michael H Malim for pcRev and pcREVM1, Ms Eva Majerova for peptide synthesis, Ms Lisa Bobroski for digital photography and Ms Rita M Victor and Ms Brenda O Gordon for excellent secretarial assistance. This work was supported in part by US PHS grant AI36552 to RJP, and the National Cancer Institute, Department of Health and Human Services under contract with ABL. References Adachi Y, Copeland TD, Takahashi C, Nosaka T, Ahmed A, Oroszlan S and Hatanaka M. (1992). J. Biol. Chem., 263, 21977 ± 21981. Agrawal MG and Bowman LH. (1987). J. Biol. Chem., 202, 4868 ± 4875. Akey CW and Goldfarb DS. (1989). J. Cell Biol., 109, 971 ± 982. Aloni R, Peleg D and Meyuhas O. (1992). Mol. Cell. Biol., 12, 2203 ± 2212. Cullen BR. (1987). Methods Enzymol., 152, 684 ± 704. Dang C and Lee WMF. (1989). J. Biol. Chem., 264, 18019 ± 18023. Eng FJ and Warner J. (1991). Cell, 65, 797 ± 804. 1513 Nuclear targeting of S25 S Kubota et al 1514 Fischer U, Meyer S, Teufel M, Heckel C, LuÈhrman R and Rautmann G. (1994). EMBO J., 13, 4105 ± 4112. Goldfarb DS. (1997). Curr. Biol., 7, 1213 ± 1216. GoÈrlich D and Mattaj IW. (1996). Science, 271, 1513 ± 1518. Gritz L, Abovich N, Teem JL and Rosbash M. (1985). Mol. Cell. Biol., 5, 3436 ± 3442. Heaphy S, Dingwall C, Ernberg I, Gait MJ, Green SM, Karn J, Lowe AD, Singh M and Skinner MA. (1990). Cell, 60, 685 ± 693. Imai T, Sudo K and Miwa T. (1994). Genomics, 20, 142 ± 143. Kalderon D, Roberts BL, Richardson WD and Smith AE. (1984). Cell, 39, 499 ± 509. Kalland K-H, Szilvay AM, Brokstad KA, Sñtrevik W and Haukenes G. (1994). Mol. Cell. Biol., 14, 7436 ± 7444. Kubota S, Duan L-X, Furuta RA, Hatanaka M and Pomerantz RJ. (1996). J. Virol., 70, 1282 ± 1287. Kubota S, Nosaka T, Cullen BR, Maki M and Hatanaka M. (1991). J. Virol., 65, 2452 ± 2456. Kubota S, Siomi H, Satoh T, Endo S, Maki M and Hatanaka M. (1989). Biochem. Biophys. Res. Comm., 162, 963 ± 970. Li H, Dalal S, Kohler J, Vilardell J and White SA. (1995). J. Mol. Biol., 250, 447 ± 459. Li M, Latoud C and Center MS. (1991). Gene, 107, 329 ± 333. Mager WH. (1988). Biochim. Biophys. Acta, 949, 1 ± 15. Malim MH, BoÈhnlein S, Hauber J and Cullen BR. (1989a). Cell, 58, 205 ± 214. Malim MH and Cullen BR. (1991). Cell, 65, 241 ± 248. Malim MH, Hauber J, Le S-H, Maizel JV and Cullen BR. (1989b). Nature, 338, 254 ± 257. Martin-Nieto J and Roufa DJ. (1997). J. Cell Sci., 110, 955 ± 965. Melchior F and Gerace L. (1995). Curr. Opin. Cell Biol., 7, 310 ± 318. Meyer BE and Malim MH. (1994). Genes & Develop., 8, 1538 ± 1547. Michael WM and Dreyfuss G. (1996). J. Biol. Chem., 271, 11571 ± 11574. Moreland RB, Nam HG, Hereford LM and Fried HM. (1985). Proc. Natl. Acad. Sci. USA, 82, 6561 ± 6565. Nosaka T, Takamatsu T, Miyazaki Y, Sano K, Sato A, Kubota S, Sakurai M, Ariumi Y, Nakai M, Fujita S and Hatanaka M. (1993). Exp. Cell Res., 209, 89 ± 102. Peculis BA and Gall JG. (1992). J. Cell. Biol., 116, 1 ± 14. Richard N, Iacampo S and Cochrane A. (1994). Virology, 204, 123 ± 131. Rout MP, Blobel G and Aitchison JD. (1997). Cell, 89, 715 ± 725. Russo G, Ricciardelli G and Concetta P. (1997). J. Biol. Chem., 272, 5229 ± 5235. Schmidt C, Lipsius E and Kruppa J. (1995). Mol. Biol. Cell., 6, 1875 ± 1885. Schmidt-Zachmann and Nigg EA. (1993). J. Cell Sci., 105, 799 ± 806. Schoborg RV and Clements JE. (1994). Virology, 202, 485 ± 490. Seiki M, Inoue J-I, Hidaka M and Yoshida M. (1988). Proc. Natl. Acad. Sci. USA, 85, 7124 ± 7128. Siomi H and Dreyfuss G. (1995). J. Cell Biol., 129, 551 ± 560. Siomi H, Shida H, Nam SH, Nosaka T, Maki M and Hatanaka M. (1988). Cell, 55, 197 ± 209. Siomi S, Shida H, Maki M and Hatanaka M. (1990). J. Virol., 64, 1803 ± 1807. Sodroski J, Rosen C, Wong-Staal F, Salahuddin SZ, Popovic M, Arya S, Gallo RC and Haseltine WA. (1985). Science, 227, 171 ± 173. Sollner-Webb B and Mougey BB. (1991). Trends Biochem. Sci., 16, 58 ± 62. Sommerville J. (1986). Trends Biochem. Sci., 11, 438 ± 442. Tan R, Chen L, Buettner JA, Hudson D and Frankel AD. (1993). Cell, 73, 1031 ± 1040. Tang CK and Drapier DE. (1990). Biochem., 29, 4434 ± 4439. Tasheva ES and Roufa DJ. (1995). Genes & Develop., 9, 304 ± 316. Tiley L, Malim M and Cullen BR. (1991). J. Virol., 65, 3877 ± 3881. Warner JR. (1990). Curr. Opin. Cell Biol., 2, 521 ± 527. Yan C and MeÂleÁse T. (1993). J. Cell Biol., 123, 1081 ± 1091. Zapp ML and Green MR. (1989). Nature, 342, 714 ± 716.