Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Hedgehog signaling pathway wikipedia , lookup
Cellular differentiation wikipedia , lookup
Protein phosphorylation wikipedia , lookup
Protein (nutrient) wikipedia , lookup
Magnesium transporter wikipedia , lookup
Cell nucleus wikipedia , lookup
Signal transduction wikipedia , lookup
Nuclear magnetic resonance spectroscopy of proteins wikipedia , lookup
Protein moonlighting wikipedia , lookup
Journal of General Virology (2004), 85, 185–196 DOI 10.1099/vir.0.19352-0 Precursor of human adenovirus core polypeptide Mu targets the nucleolus and modulates the expression of E2 proteins T. W. R. Lee,1 F. J. Lawrence,2 V. Dauksaite,3 G. Akusjärvi,3 G. E. Blair1 and D. A. Matthews2 1 School of Biochemistry and Molecular Biology, University of Leeds, Leeds LS2 9JT, UK Correspondence David Matthews 2 Division of Virology, Department of Pathology and Microbiology, University Walk, Bristol University, Bristol BS8 1TD, UK [email protected] 3 Department of Medical Biochemistry and Microbiology, Uppsala University, Box 582, S-75123 Uppsala, Sweden Received 13 May 2003 Accepted 25 September 2003 We have examined the subcellular localization properties of human adenovirus 2 (HAdV-2) preMu and mature Mu (pX) proteins as fusions with enhanced green fluorescence protein (EGFP). We determined that preMu is exclusively a nucleolar protein with a single nucleolar accumulation signal within the Mu sequence. In addition, we noted that both preMu–EGFP and Mu–EGFP are excluded from adenovirus DNA-binding protein (DBP)-rich replication centres in adenovirus-infected cells. Surprisingly, we observed that cells in which preMu–EGFP (but not Mu–EGFP) is transiently expressed prior to or shortly after infection with Ad2 did not express late adenovirus genes. Further investigation suggested this might be due to a failure to express pre-terminal protein (preTP) from the E2 region, despite expression of another E2 protein, DBP. Deletion mutagenesis identified a highly conserved region in the C terminus of preMu responsible for these observations. Thus our data suggest that preMu may play a role in modulating accumulation of proteins from the E2 region. INTRODUCTION Human adenovirus particles consist of approximately 12 proteins enclosing a genome of linear double-stranded DNA of approximately 36 kb. This viral DNA is covalently linked to viral terminal protein and non-covalently bound to the viral core proteins Mu, V and VII (reviewed by Shenk, 2001). Protein V, found only in mastadenoviruses (Davison et al., 2003), is believed to form a link between the viral DNA–core protein complex and the viral capsid (Matthews & Russell, 1998b), while proteins VII and Mu are tightly associated with the viral DNA (Chatterjee et al., 1985; Vayda et al., 1983). Previous studies showed that protein V is associated with the nucleolus during the late phase of infection and, in the nucleus, V is excluded from viral DNA binding proteinrich centres where single-stranded DNA accumulates (Matthews & Russell, 1998b). Additional nuclear and nucleolar targeting sequences in protein V have been analysed, in which two highly basic nucleolar targeting regions were identified (Matthews, 2001). In addition data were presented that protein V may disrupt nucleolar function by affecting the subcellular localization of the nucleolar antigens nucleolin and B23. Why adenovirus might disrupt nucleolar function is unknown, but recent 0001-9352 G 2004 SGM data suggest that B23, for example, may have importance in initiating adenoviral DNA replication (Okuwaki et al., 2001). Nucleoli are known to be the sites of ribosome formation, where rRNA is synthesized, processed and incorporated into ribosomes (Pederson, 1998; Scheer & Hock, 1999). In adenovirus infection the formation of 18S and 28S rRNA is reduced (Castiglia & Flint, 1983) and late in infection the nucleoli are disrupted (Puvion-Dutilleul & Christensen, 1993). Our interest in these phenomena led us to examine Mu protein because it is arginine-rich and contains sequences that are similar to a nucleolar targeting signal found in protein V (Matthews, 2001). Mu (also known as pX) is synthesized as a 79 aa precursor that undergoes both N- and C-terminal cleavage by a viruscoded protease to form mature Mu (19 aa). It has been shown that Mu protein precipitates DNA in vitro, suggesting that Mu helps condense DNA in the virus core by means of a charged-based interaction between the nine arginine residues in Mu and the phosphate DNA backbone (Anderson et al., 1989). Consistent with this, a recent study demonstrated that Mu protein enhanced transfection efficiency by up to 11-fold when used as an adjunct to liposome-mediated gene therapy (Murray et al., 2001). Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 Printed in Great Britain 185 T. W. R. Lee and others There is no published in situ data on the intracellular localization of Mu (or its precursor protein) during infection. Indeed, such a study is complicated by the observation that a monoclonal antibody against Mu crossreacts with a similar region within protein VII (Lunt et al., 1988). Consequently, we decided to use tagged versions of Mu and preMu to examine the subcellular localization of the proteins in an infected cell, relative to viral DNA replication centres. We found that preMu contains a nucleolar localization signal, the key element of which is located within Mu. We found no evidence that preMu or Mu affect the subcellular distribution of the nucleolar antigens nucleolin or B23. In addition, we could not demonstrate any gross defects in de novo rRNA synthesis. However, our data are consistent with preMu (but not Mu) being able to modulate accumulation of viral proteins derived from the E2 region. METHODS Cells and viruses. HeLa cells, cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10 % foetal calf serum, penicillin (100 IU ml21) and streptomycin (100 mg ml21), were used to propagate human wild-type 2 adenovirus (HAdV-2). Cells were infected and viruses purified as previously described (Matthews & Russell, 1994). For transfection of plasmid constructs, HeLa cells were grown on glass coverslips in six-well dishes. The cells were transfected with 2 mg of each plasmid using Lipofectamine 2000 (Invitrogen). Cloning of recombinant derivatives of preMu and ASF/SF2. Regions of the open reading frame for preMu protein were amplified from HAdV-2 DNA using oligonucleotide primers and a PCR kit (PFX; Invitrogen). Synthetic oligonucleotides corresponding to the open reading frame of Mu were made and annealed to generate a DNA fragment for cloning. In this manner we also generated mutants of Mu with selected arginine codons replaced by alanine codons. The DNA fragments were cloned into a CMV promoterbased mammalian expression plasmid (pcJMA2egfp; a kind donation from J. Askham; Askham et al., 2000) to express the amino acid sequences produced as N-terminal fusions to enhanced green fluorescence protein (EGFP; Clontech). The open reading frame for the splicing factor ASF/SF2 was PCR-amplified from EGFP-ASF/SF2 (Sleeman et al., 1998; kindly provided by A. I. Lamond) and cloned into pds-RedC1 (Clontech). In addition, synthetic oligonucleotides were designed to enable us to replace the open reading frame of EGFP with the amino acid sequences corresponding to the Myc or FLAG tags. The sequences of primers used in this manuscript are available on request. Fluorescent imaging. Eighteen to twenty hours after transfection the cells were fixed using 4 % formaldehyde (v/v in PBS). Cells were washed in PBS prior to the coverslips being mounted with Vectashield containing 49,6-diamidino-2-phenylindole (DAPI; Vector Laboratories). Detection of EGFP-tagged proteins was performed using a Zeiss Axiovert 135TV microscope with a Neofluor 406 oilimmersion lens. Immunofluorescence. Formaldehyde-fixed cells on coverslips were permeabilized using Triton X-100 (1 %, v/v, in PBS) prior to blocking with dried skimmed milk (1 %, w/v, in PBS) for 1 h at room temperature. The following primary antibodies were used: anti-B23 (kindly provided by B. Valdez; Perlaky et al., 1997), anti-protein V (Matthews & Russell, 1998b), anti-protein VI, (Matthews & Russell, 186 1994), anti-DBP (Russell et al., 1989), anti-hexon, anti-penton base (kind donations from W. C. Russell), and anti-terminal protein (kindly provided by R. T. Hay; Webster et al., 1997). Appropriate secondary antibodies were linked to Texas Red or FITC (Vector Laboratories). Cells were mounted and viewed as described above; alternatively images were collected using a laser confocal microscope (Leica TCS SP) and a PlanApo 1006 UV oil-immersion lens. Fluorouridine labelling of transfected cells. HeLa cells were transfected with preMu–EGFP or Mu–EGFP constructs as described and 18 h post-transfection the cells were incubated with 15 mM 59fluorouridine (59FU from Sigma) for 15 min. Cells were then fixed and permeabilized as outlined and FU incorporated into RNA was detected using monoclonal anti-BrdU antibody (Sigma) as described previously (Boisvert et al., 2000). Western blotting. HeLa cells were grown in six-well dishes without coverslips and transfected as described above. After 18 to 20 h, cells were harvested and prepared for SDS-PAGE. Western blotting was performed using rabbit antibodies against EGFP (Santa-Cruz), and secondary anti-rabbit immunoglobulin linked to HRP (Vector Laboratories). Detection was performed using ECL (Amersham-Pharmacia). Infection/transfection studies. In order to assess the location of preMu–EGFP and Mu–EGFP during infection, cells were infected with HAdV-2 at an m.o.i. of 5 p.f.u. per HeLa cell. This inoculation was performed either 7 h before transfection with plasmid, simultaneous with transfection, 7 h after transfection or 24 h after transfection. Immunofluorescence patterns were observed at 20, 26 or 30 hours following infection. RESULTS PreMu and Mu-EGFP fusion protein distribution in uninfected HeLa cells Fig. 1(A) shows the deletion mutants of preMu generated, and localization of the EGFP fusion products in uninfected cells. In each case the intracellular localization of these products was consistent across >90 % of transfected cells. The expression of full-length protein was confirmed by Western blotting (see Fig. 6D) and nucleolar targeting was confirmed by comparison with anti-B23 antibody and phase-contrast microscopy (data not shown). Full-length preMu–EGFP (1–79EGFP) demonstrated exclusive nucleolar targeting, as did 1–50EGFP. However, 32–79EGFP and Mu (32–50EGFP) demonstrated nucleolar targeting with a cytoplasmic background (representative images shown in Fig. 1C). Mutations of Mu that affect nucleolar localization In order to assess which regions of Mu are important for nucleolar targeting, a series of Mu mutants was generated with arginine domains replaced with alanines (Fig. 1B). This did not ablate nucleolar targeting; however, substitution of the arginines in positions 33 and 34 (mutant 1), 38 and 39 (mutant 2a) or 45 and 46 (mutant 3) resulted in an apparent modest reduction in nucleolar targeting. Substituting the two arginines at positions 40 and 41 (mutant 2b) appeared to distinctly reduce the nucleolar targeting (Fig. 1C viii). The significance of these changes in Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 Journal of General Virology 85 Properties of adenovirus preMu protein Fig. 1. Identification of regions of Mu protein involved in nuclear and nucleolar targeting. (A) Schematic showing the regions of preMu protein fused to EGFP. On the left, each protein is identified by the amino acids from the full-length preMu protein sequence, expressed N-terminal to the EGFP sequences. On the right, the subcellular distribution of each fusion protein is indicated as No. (nucleolar), Np. (nucleoplasmic) and/or Cyt. (cytoplasmic). (B) Mutants of Mu, formed by substitution of selected arginine codons with alanine codons. Mu protein is shown, with arginines highlighted, followed by each mutant with substitute alanines underlined. All mutants were formed as N-terminal fusions to EGFP. On the right the subcellular localizations are described, as above: nucleolar (No.), nucleoplasmic (Np.) and cytoplasmic (Cyt). (C) Subcellular localization of representative deletion mutants of preMu, and five Mu mutants. (i) Expression pattern of GFP alone; (ii) DAPI staining of cell transfected with preMu–EGFP; (iii) preMu–EGFP expression pattern; (iv) DAPI staining of cell transfected with Mu–EGFP; (v) Mu–EGFP expression pattern; (vi) Mu mutant 1; (vii) Mu mutant 2a; (viii) Mu mutant 2b; (ix) Mu mutant 3; (x) Mu mutant 4. All images were scanned from slide film and labelled using Adobe Photoshop 4.0. targeting, which although consistent across multiple experiments are difficult to quantify, is unclear. Substituting the arginine in the sequence position 48 (mutant 4) made no detectable difference to the nucleolar targeting characteristics of Mu. In addition, we noted that evenly distributing the arginines within Mu (mutant 5) did not affect the ability to localize EGFP in the nucleolus. PreMu–EGFP and Mu–EGFP did not abrogate rRNA synthesis or affect the localization of nucleolar antigens An in situ rRNA synthesis assay (Boisvert et al., 2000) using fluorouridine labelling demonstrated that preMu–EGFP had no effect on RNA synthesis in the nucleolus. In addition, as with protein V (Matthews, 2001), blocking rRNA http://vir.sgmjournals.org synthesis using actinomycin D did not affect the localization of preMu–EGFP (data not shown). In contrast to protein V (Matthews, 2001), preMu–EGFP and Mu–EGFP did not affect the gross distribution of nucleolin or B-23. However, we could not find cells expressing high levels of preMu– EGFP after 48 h. We were also unable to isolate stable cell lines expressing preMu–EGFP using an inducible promoter system. Together these two observations indicate that expression of preMu–EGFP is toxic to the cells. PreMu–EGFP and Mu–EGFP were excluded from the DBP-rich viral DNA replication centres during adenovirus infection To determine if the localization of preMu–EGFP or Mu– EGFP was affected by adenovirus infection, we studied the Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 187 T. W. R. Lee and others effects of transfecting these EGFP fusion protein expression plasmids into adenovirus-infected cells. As protein V had previously been shown to associate with the nucleolus and was excluded from the DBP-rich areas (Matthews & Russell, 1998a) we first assessed whether this also occurred with preMu–EGFP and Mu–EGFP protein (Fig. 2A, B). Immunofluorescent labelling of DBP enabled us to demonstrate that, as with protein V, both preMu–EGFP and Mu–EGFP protein were excluded from DBP-rich areas. Other than this exclusion, the distribution of both preMu–EGFP and Mu–EGFP was similar to that seen in uninfected cells. Expression of preMu–EGFP in HeLa cells blocked late adenovirus gene expression In order to further clarify the location of preMu–EGFP and Mu–EGFP proteins in infected cells a series of fluorescent antibody staining experiments was undertaken to establish which adenoviral proteins co-localized with preMu– and Mu–EGFP-tagged variants. We transfected cells with a range of EGFP-tagged constructs and infected the cells at the same time with adenovirus as described. We then assayed the expression of protein VI at various times >20 h postinfection to ensure that late gene expression had begun. To our surprise we observed that there was no detectable adenovirus protein VI in any cells expressing preMu–EGFP fusion, whilst un-transfected cells did express protein VI (Fig. 3A). This effect was seen in cells with a wide range of expression levels of preMu–EGFP (see Fig. 3A), suggesting that preMu–EGFP inhibited the expression of protein VI, a product of late gene transcription. The m.o.i. selected ensured that >95 % of cells were infected and were producing protein VI. The expression of other proteins derived from adenovirus late genes, such as hexon, fibre, penton base and protein V, was also inhibited but expression of the intermediate gene product protein IX was reduced, but not ablated (data not shown). By comparison, Mu–EGFP fusion protein expression had no effect on the production of late proteins (Fig. 3B). We also determined that a protein V–EGFP fusion (Matthews, 2001) did not affect late gene expression Fig. 2. Location of preMu–EGFP and Mu–EGFP in adenovirus-infected cells. Cells were infected with HAdV-2 and simultaneously transfected with plasmid encoding preMu–EGFP (A) or Mu–EGFP (B). After 20 h the cells were fixed and expression of DBP was detected using monoclonal antiserum. In these confocal images the cytoplasm is not visible. (A) (i) Location of preMu–EGFP; (ii) immunofluorescent labelling of DBP; (iii) merge. (B) (i) Location of Mu–EGFP; (ii) immunofluorescent labelling of DBP; (iii) merge. Scale, 10 mm. 188 Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 Journal of General Virology 85 Properties of adenovirus preMu protein Fig. 3. PreMu–EGFP but not Mu–EGFP blocked late gene expression in adenovirus-infected cells. Cells were infected with HAdV-2 and simultaneously transfected with plasmids encoding preMu–EGFP (A) or Mu–EGFP (B). After 30 h the cells were fixed and expression of protein preVI was detected using polyclonal antiserum. (A) (i) DAPI staining of cells; (ii) location of three cells expressing notably different amounts of preMu-EGFP; (iii) immunofluorescent labelling of protein VI. (B) (i) DAPI staining of cells; (ii) location of Mu-EGFP expressing cells; (iii) immunofluorescent labelling of protein VI. All images were scanned from slide film and labelled using Adobe Photoshop 4.0. (data not shown). Finally, we determined that expression of preMu–EGFP 7 h before, after or during infection with adenovirus precluded late gene expression for at least 30 h post-infection. Expression of preMu–EGFP in HeLa cells inhibited preTP expression We also examined preTP expression since, like DBP, it is also a product of the E2 transcription unit. In adenovirusinfected cells, those cells transfected with preMu–EGFP and simultaneously infected with adenovirus failed to express preTP after 20 h (Fig. 4A), whereas we had previously shown that expression of DBP could be readily detected (Fig. 2A). Mu–EGFP, protein V–EGFP and EGFP expression did not affect the expression of preTP in these assays. Molin & Akusjärvi (2000) demonstrated that overexpression of essential splicing factor ASF/SF2 during adenovirus http://vir.sgmjournals.org infection resulted in a failure to express late genes (but not DBP) and postulated that this was due to reduced expression of preTP. We therefore used a RED–ASF/SF2 construct to determine if the results obtained using preMu–EGFP could be repeated. Our data (Fig. 4B, C) show that, like preMu–EGFP, RED–ASF/SF2 apparently blocks the expression of preTP but not that of DBP. We also noted that, like preMu–EGFP, RED–ASF/SF2 was excluded from the DBP-rich replication centres. Overexpression of RED–ASF/SF2 mislocalized a proportion of preMu–EGFP in uninfected cells This similarity of effect of preMu–EGFP and RED–ASF/ SF2 on late gene expression led us to examine whether overexpression of RED–ASF/SF2 would affect the intracellular localization of preMu-EGFP or vice versa. We therefore transfected uninfected cells with plasmids expressing both RED–ASF/SF2 and preMu–EGFP. In Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 189 T. W. R. Lee and others Fig. 4. PreMu–EGFP and RED–ASF/SF2 reduced preTP expression in adenovirus-infected cells whereas RED–ASF/SF2 expression did not affect DBP expression in infected cells. Cells were infected with HAdV-2 and simultaneously transfected with appropriate plasmids. After 20 h the cells were fixed and expression of protein was detected using appropriate antiserum. (A) PreMu–EGFP and preTP expression. (i) DAPI staining of cells; (ii) distribution of cells expressing preMu–EGFP; (iii) immunofluorescent labelling of preTP. (B) RED–ASF/SF2 and preTP expression. (i) DAPI staining of cells; (ii) immunofluorescent labelling of preTP; (iii) distribution of cells expressing RED–ASF/SF2. (C) RED–ASF/SF2 and DBP expression. (i) Immunofluorescent labelling of DBP; (ii) distribution of RED-ASF/SF2; (iii) merge. Scale, 10 mm. >90 % of cells examined we observed mislocalization of a proportion of preMu–EGFP into the nucleus. This nuclear preMu–EGFP corresponded with RED–ASF/SF2 in 190 extranucleolar sites (Fig. 5A). This effect on the distribution of preMu–EGFP was only evident in cells overexpressing RED–ASF/SF2. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 Journal of General Virology 85 Properties of adenovirus preMu protein Fig. 5. Uninfected cells expressing preMu–EGFP and RED–ASF/SF2 indicated a proportion of preMu–EGFP was present in the same regions of the nucleus as RED–ASF/SF2. In both cases, uninfected cells were transfected with appropriate plasmids and examined 20 h later. (A) Cells transfected with preMu–EGFP and RED–ASF/SF2 and examined 20 h later. (i) Location of preMu–EGFP in the nucleus and nucleolus. (ii) RED–ASF/SF2. (iii) Merge. (B) Cells transfected with preMu– EGFP only and counter-stained 20 h later with monoclonal antiserum to detect the splicing factor sc35. (i) PreMu–EGFP in the nucleus and nucleolus (ii) Anti-sc35. (iii) Merge. Note that in this set of images the preMu-EGFP image has been deliberately overexposed compared to (A) (i) in order that the location of very low levels on preMu–EGFP in the nucleus could be detected. In all cases the white bar represents 10 mm. In contrast, Fig. 5(B) shows that we consistently observed only a very low level diffuse staining of preMu–EGFP in the nucleus of cells not transfected with RED–ASF/SF2. For clarity, we deliberately raised the sensitivity of the confocal microscope for Fig. 5(B) so that this low-level background nuclear staining of preMu–EGFP could be clearly seen. Note there was no correspondence with structures revealed by anti-sc35 serum which detects another essential splicing factor, sc35. Examination of preMu–EGFP transfected cells (Fig. 5A) revealed that its subnucleolar distribution within the cell nucleolus was similar to that of protein V–EGFP in uninfected cells (Matthews, 2001). Thus, it appeared to be associated with those regions of the nucleolus where rRNA processing predominates and was excluded from the centres where inactive rDNA is stored. http://vir.sgmjournals.org RED–ASF/SF2 did not accumulate in the nucleolus in any of our experiments, and did not co-localize with protein V–EGFP (data not shown). The deletion mutants 1–50EGFP, 32–50EGFP, 32–79EGFP, 1–69EGFP and 1–59EGFP also co-localized with RED–ASF/SF2, whereas 1–31EGFP and 51–79EGFP did not co-localize with RED– ASF/SF2 (Fig. 6A). Mutagenesis identified a region in the C-terminal one-third of preMu involved in blocking late gene expression In order to explore which domains of preMu are responsible for the effect on late gene expression, the preMu deletion series described above was assessed (Fig. 6A). Only those proteins containing the sequence 32–79EGFP affected late gene expression. Additional deletion mutants demonstrated Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 191 T. W. R. Lee and others Fig. 6. Identification of regions of preMu protein involved in inhibiting late gene expression in adenovirus-infected cells, and regions required for co-localization with ASF/SF2. (A) Schematic representation of deletion mutants of preMu–EGFP fusion protein as in Fig. 1, with indication of the effects of these mutants, and further deletion mutants 1–59EGFP and 1–69EGFP, on both the expression of adenovirus late gene products and co-localization of mutants with Red–SF2 in adenovirus-infected cells. (B) Amino acid sequence of a fragment of preMu (51–79), showing the cleavage sites for the further deletion mutants 1–59EGFP and 1–69EGFP, as well as the point mutants formed by replacing the isoleucine at position 63, proline at position 67 and isoleucine at position 69 with glycine in turn. (C) Comparison of amino acid sequence of fragment of preMu protein after the Mu region, showing sequences from representatives of the four official genera of adenoviruses: Mastadenovirus [human adenovirus serotypes 2 (WMADH2), 12 (A45393) and 40 (Q64858) and mouse adenovirus 1 (O10443)]; Atadenovirus [duck adenovirus 1 (NP044708)]; Aviadenovirus [fowl adenovirus 1 (CELO) (CAB69562)]; and Siadenovirus [frog adenovirus 1 (NP 062441)]. Conserved amino acids are in bold type, amino acids selected for mutation are denoted with an asterisk and amino acids required for blocking late gene expression are underlined. (D) Western blot of EGFP fusion constructs of key functional importance, to demonstrate that these mutants were expressed at the appropriate size. PreMu– EGFP and 1–69EGFP ablated the expression of late genes in HAdV-2-infected cells, whilst Mu–EGFP, EGFP alone and 1–59EGFP did not. that the sequence 1–69EGFP but not 1–59EGFP inhibited the expression of preTP and protein VI. In order to explore which regions of 32–69EGFP were responsible for this effect, site-directed mutagenesis was performed in turn at sites 63, 67 and 69 within a 1–69EGFP construct (Fig. 6B). These amino acids are highly conserved between adenovirus species (Fig. 6C), and have large hydrophobic sidegroups. In each case glycine (small hydrophilic) was used as a non-conservative substitution. Each of these mutants of 1–69EGFP continued to block the expression of late genes. The series of deletion and site-directed mutants described in this section all had the same subcellular localization as 192 that of 32–79EGFP, and were expressed at the expected size on Western blot (for mutants of key functional importance see Fig. 6D). Differentially tagged preMu fusion proteins also accumulate in the nucleolus in uninfected cells and blocked late gene expression in infected cells The recombinant preMu–EGFP protein is three times larger than preMu so we also C-terminally tagged our clone of preMu using Myc (12 aa), FLAG (8 aa) or dsRED (225 aa) Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 Journal of General Virology 85 Properties of adenovirus preMu protein Fig. 7. Differential tagging of preMu does not affect nucleolar localization or inhibition of late gene expression. (A) Cells were transfected with plasmid encoding preMu–Myc and allowed to express for 20 h prior to fixation and staining. (i) Location of preMu–Myc in uninfected cells as detected using an anti-Myc antiserum; (ii) Phase-contrast photograph of the same cells. (B) Cells were infected with HAdV-2 and simultaneously transfected with plasmids encoding preMu–Myc. After 30 h the cells were fixed and expression of protein preVI was detected using polyclonal antiserum. (i) Location of preMu–Myc-expressing cells; (ii) Immunofluorescent labelling of fibre; (iii) DAPI staining of cells. instead of EGFP. We used these fusion proteins to confirm the nucleolar localization of preMu (Fig. 7A) and the effect on expression of late genes; in this example we have shown that cells in which preMu–Myc is expressed exhibit greatly reduced levels of fibre protein (Fig. 7B). DISCUSSION Our data reveal that, as with protein V (Matthews & Russell, 1998b), preMu protein demonstrates nuclear and nucleolar targeting. Furthermore, the single nucleolar accumulation motif within the preMu protein corresponds to Mu. In this regard we can apparently separate nuclear targeting from nucleolar accumulation as has been described for protein V (Matthews, 2001). For example, clone 1–50EGFP is nuclear and nucleolar whereas 32–50 is also cytoplasmic, suggesting that this clone is not actively targeted to/retained within the nucleus. Since there is no nuclear localization motif within 1–31EGFP we suggest that the motifs required to target or retain 1–50EGFP within the nucleus span the http://vir.sgmjournals.org N-terminal cleavage site for Mu. This could imply that a bipartite nuclear localization signal spans this cleavage site. We suggest that, with the possible exception of the arginine at position 48, all the arginine motifs of Mu, particularly at positions 40 and 41, contribute to nucleolar accumulation. Nucleolar localization sequences (NoLS) similar to Mu have been described previously (Jarrous et al., 1999; Liu et al., 1997). The nucleolar targeting of a 23 aa fragment of ribonuclease P subunit 38 is dependent on all nine lysine residues spaced through the sequence, though two adjacent lysines are most critical for entry to the nucleolus (Jarrous et al., 1999). A single arginine present in the sequence had no role in nucleolar localization. Similarly, a comparison of homology between NoLS of 13 different proteins revealed that all are between 11 and 20 aa in length, contain between two and six arginine or lysine domains, with a gap between domains of between one and three non-basic amino acids (Liu et al., 1997). Whilst these sequences are all broadly similar, there is no absolute consensus sequence. Van Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 193 T. W. R. Lee and others Eenennaam et al. (2001) suggested that highly basic domains non-specifically target the nucleolus, and having entered the organelle by passive diffusion they are retained by binding to negatively charged molecules within the nucleolus. In common with these previously described NoLS, substitution of certain basic amino acids within the Mu NoLS reduces the intensity of the nucleolar targeting, but none of these amino acids are absolutely critical in this regard. Indeed, when the arginines are spaced evenly there is no apparent reduction in nucleolar accumulation. This may be due to either the existence of multiple redundant NoLS within the sequence or, more likely, targeting Mu to the nucleolus is primarily charge-based rather than dependent on a specific pattern of amino acids. PreMu-EGFP and Mu-EGFP do not affect nucleolar antigens (B23 and nucleolin) nor do they affect rRNA synthesis and consequently their role in the nucleolus will require further investigation. Interestingly, neither protein V–EGFP nor preMu–EGFP are affected by inhibition of rRNA synthesis, something that markedly effects the distribution of rRNA processing proteins. Potentially they interact with proteins/structures within the nucleolus unaffected by treatment with actinomycin D. Examples of such nucleolar proteins are, RNA pol I, fibrillarin and p120 (Perlaky et al., 1997). Significantly, we have been able to lengthen the list of adenovirus proteins that may affect nucleolar structure and function. We have shown that expression of preMu as a fusion protein in adenovirus-infected HeLa cells inhibits late gene expression. We have used specific antisera to a range of late viral proteins in order to establish that the defect is widespread. Normally, the progression to the late phase requires replication of viral DNA. Thus, the simplest conclusion is that there is a failure to enter the late phase, probably resulting from lack of DNA replication. Consistent with this, Western blotting revealed that preMu–EGFP was not cleaved by the viral protease in our infection/ transfection assays (data not shown). This is to be expected because the viral protease is a late gene and there is a failure to enter the late phase in preMu–EGFP-transfected cells. We had already established that viral DBP was being expressed in cells transfected with recombinant preMu proteins and we noted that there was precedence for this type of observation. A recombinant adenovirus that expressed ASF/SF2 under the control of an inducible promoter has been described (Molin & Akusjarvi, 2000). This report noted that when induced, ASF/SF2 blocked late gene expression but expression of DBP was evident. We have previously noted unpublished evidence that overexpression of ASF/SF2 inhibited preTP expression (Molin & Akusjarvi, 2000). The lack of preTP expression would account for the lack of DNA replication and the loss of late protein expression. How ASF/SF2 might do this is not understood at present but it seems reasonable to postulate that there is a failure of the splicing pathway required to generate preTP mRNA. 194 Our data demonstrate that preMu–EGFP has similar effects to ASF/SF2 in blocking adenovirus late protein expression. Consistent with a possible effect on DNA replication we noted that protein IX expression was reduced but not abolished. We therefore wanted to examine if our transfection/infection data could be repeated using the dsRED fusion of the splicing factor ASF/SF2 and we observed that both preMu–EGFP and RED–ASF/SF2 block preTP expression but allowed DBP expression. Thus our data provided further evidence that both preMu–EGFP and ASF/SF2 apparently modulate accumulation of E2 region proteins. Previous studies have established that transcription and splicing are separate from DNA replication and DBP-rich regions of the infected nucleus (Bridge et al., 1993, 1995; Pombo et al., 1994). As expected therefore, our RED–ASF/ SF2 clone was also excluded from the DBP-rich centres. The exclusion of proteins preMu–EGFP and Mu–EGFP from the adenoviral DBP-rich areas is compatible with a model in which preMu influences mRNA biogenesis. However, we also note that preMu–EGFP in infected cells does have a significant nuclear component not normally seen in uninfected cells. This might reflect preMu being involved in disruption of nucleolar function as well as in viral RNA biogenesis. In support of this idea we observed that co-expression of RED–ASF/SF2 together with preMu–EGFP in uninfected cells leads to a proportion of preMu–EGFP accumulating in the nucleus with RED–ASF/SF2. Clearly, this phenomenon requires overexpression of Red–ASF/SF2 but it indicates that preMu–EGFP and Red–ASF/SF2 might colocalize under certain conditions in uninfected cells. Intriguingly, only clones with the arginine-rich Mu sequence present acted in this manner. However, this property of preMu–EGFP is not sufficient to block late gene expression in infected cells. For example, Mu–EGFP is affected by RED–ASF/SF2 in uninfected cells but is not associated with ablation of late gene expression in infected cells. Moreover, as they are both excluded from DBP-rich centres in infected cells we postulate that preMu and ASF/SF2 might be present in the same complex during infection. Thus, the nature of the relationship between preMu– EGFP and RED–ASF/SF2 is unclear but it is notable that ASF/SF2 can be isolated from purified nucleoli (Andersen et al., 2002). Deletion mutagenesis enabled us to determine that the C-terminal portion of preMu was required to affect the accumulation of preTP. Examination of this region in a wide range of adenoviruses revealed that this region is highly conserved amongst all the major lineages of adenovirus. We attempted to define individual amino acids that might be key to preMu’s ability to block preTP but were unsuccessful. In the current absence of structural data on preMu, this might imply that an extended region of preMu is involved in blocking preTP expression; such ideas are difficult to test. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 Journal of General Virology 85 Properties of adenovirus preMu protein The data we had accumulated pointed to a role for preMu in mRNA biogenesis, and the similarities with ASF/SF2 point to the simple conclusion that preMu may also be capable of modulating splice-site selection. However, splicing of the E2 region is not well understood at present and so the potential for particular effects on the accumulation of preTP mRNA cannot be excluded at this stage (e.g. specific effects on preTP mRNA transport). Indeed, it has not yet been shown that ASF/SF2 modulates the splicing of E2 mRNA, so claims that both these proteins modulate the splicing of E2 must be treated with caution at this time. That late viral proteins might affect splice selection has some precedence in the adenovirus system. Recent evidence suggests that unscheduled expression of pre-IIIa protein (derived from L1) caused alterations to the accumulation of late mRNA species (Molin et al., 2002). Clearly, it is possible that several members of the late family of proteins influence post-transcriptional events during the late phase of infection. We were conscious of the possibility that the EGFP tag may have a significant influence on preMu function. However, preMu–Myc, preMu–FLAG and preMu–dsRED proteins all showed identical characteristics to preMu– EGFP. Therefore, we consider that the use of a number of alternate tags provides some measure of assurance that the effects we observe result from the properties of preMu and not from the fusion protein as a whole. PreMu is the first adenovirus late protein that has been shown to have an effect on accumulation of early proteins in infected cells. Prior to this study, the role of preMu, other than as a precursor to Mu, was unknown. We note that the effects are likely to be mediated through the highly conserved C-terminal one-third of preMu. This suggests that modulation of E2 protein expression by preMu may be conserved amongst all the adenovirus genera. In addition, our data regarding the nucleolar targeting ability of Mu protein further enhances our understanding of the relationship of adenovirus with the nucleolus and suggests that Mu may also play a role in the disruption of the nucleolus. Further investigation will focus on the mechanism of preMu’s effect on E2 mRNA, and the activity of preMu in the nucleolus. ACKNOWLEDGEMENTS We would like to thank R. T. Hay, W. C. Russell and B. Valdez for antisera. We also thank the Molecular Medicine Unit Confocal User Group for invaluable assistance in using the microscope. The research is supported by the Medical Research Council (D. A. M. and T. W. R. L.), the Wellcome Trust (F. J. L.) and the Swedish Cancer Society (G. A. and V. D.). D. A. M. and T. W. R. L. are Fellows of the Medical Research Council. Anderson, C. W., Young, M. E. & Flint, S. J. (1989). Characterization of the adenovirus 2 virion protein, mu. Virology 172, 506–512. Askham, J. M., Moncur, P., Markham, A. F. & Morrison, E. E. (2000). Regulation and function of the interaction between the APC tumour suppressor protein and EB1. Oncogene 19, 1950–1958. Boisvert, F. M., Hendzel, M. J. & Bazett-Jones, D. P. (2000). Promyelocytic leukemia (PML) nuclear bodies are protein structures that do not accumulate RNA. J Cell Biol 148, 283–292. Bridge, E., Carmo-Fonseca, M., Lamond, A. & Pettersson, U. (1993). Nuclear organization of splicing small nuclear ribonucleo- proteins in adenovirus-infected cells. J Virol 67, 5792–5802. Bridge, E., Xia, D. X., Carmo-Fonseca, M., Cardinali, B., Lamond, A. I. & Pettersson, U. (1995). Dynamic organization of splicing factors in adenovirus-infected cells. J Virol 69, 281–290. Castiglia, C. L. & Flint, S. J. (1983). Effects of adenovirus infection on rRNA synthesis and maturation in HeLa cells. Mol Cell Biol 3, 662–671. Chatterjee, P. K., Vayda, M. E. & Flint, S. J. (1985). Interactions among the three adenovirus core proteins. J Virol 55, 379–386. Davison, A. J., Benkő, M. & Harrach, B. (2003). Genetic content and evolution of adenoviruses. J Gen Virol 84, 2895–2908. Jarrous, N., Wolenski, J. S., Wesolowski, D., Lee, C. & Altman, S. (1999). Localization in the nucleolus and coiled bodies of protein subunits of the ribonucleoprotein ribonuclease P. J Cell Biol 146, 559–572. Liu, J. L., Lee, L. F., Ye, Y., Qian, Z. & Kung, H. J. (1997). Nucleolar and nuclear localization properties of a herpesvirus bZIP oncoprotein, MEQ. J Virol 71, 3188–3196. Lunt, R., Vayda, M. E., Young, M. & Flint, S. J. (1988). Isolation and characterization of monoclonal antibodies against the adenovirus core proteins. Virology 164, 275–279. Matthews, D. A. (2001). Adenovirus protein V induces redistribution of nucleolin and B23 from nucleolus to cytoplasm. J Virol 75, 1031–1038. Matthews, D. A. & Russell, W. C. (1994). Adenovirus protein–protein interactions: hexon and protein VI. J Gen Virol 75, 3365–3374. Matthews, D. A. & Russell, W. C. (1998a). Adenovirus core protein V interacts with p32 – a protein which is associated with both the mitochondria and the nucleus. J Gen Virol 79, 1677–1685. Matthews, D. A. & Russell, W. C. (1998b). Adenovirus core protein V is delivered by the invading virus to the nucleus of the infected cell and later in infection is associated with nucleoli. J Gen Virol 79, 1671–1675. Molin, M. & Akusjarvi, G. (2000). Overexpression of essential splicing factor ASF/SF2 blocks the temporal shift in adenovirus pre-mRNA splicing and reduces virus progeny formation. J Virol 74, 9002–9009. Molin, M., Bouakaz, L., Berenjian, S. & Akusjarvi, G. (2002). Unscheduled expression of capsid protein IIIa results in defects in adenovirus major late mRNA and protein expression. Virus Res 83, 197–206. Murray, K. D., Etheridge, C. J., Shah, S. I., Matthews, D. A., Russell, W., Gurling, H. M. & Miller, A. D. (2001). Enhanced cationic liposome-mediated transfection using the DNA-binding peptide mu (mu) from the adenovirus core. Gene Ther 8, 453–460. Okuwaki, M., Iwamatsu, A., Tsujimoto, M. & Nagata, K. (2001). Identification of nucleophosmin/B23, an acidic nucleolar protein, as a stimulatory factor for in vitro replication of adenovirus DNA complexed with viral basic core proteins. J Mol Biol 311, 41–55. REFERENCES Pederson, T. (1998). The plurifunctional nucleolus. Nucleic Acids Res Andersen, J. S., Lyon, C. E., Fox, A. H., Leung, A. K., Lam, Y. W., Steen, H., Mann, M. & Lamond, A. I. (2002). Directed proteomic Perlaky, L., Valdez, B. C. & Busch, H. (1997). Effects of cytotoxic 26, 3871–3876. analysis of the human nucleolus. Curr Biol 12, 1–11. http://vir.sgmjournals.org drugs on translocation of nucleolar RNA helicase RH-II/Gu. Exp Cell Res 235, 413–420. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 195 T. W. R. Lee and others Pombo, A., Ferreira, J., Bridge, E. & Carmo-Fonseca, M. (1994). Adenovirus replication and transcription sites are spatially separated in the nucleus of infected cells. EMBO J 13, 5075–5085. Sleeman, J., Lyon, C. E., Platani, M., Kreivi, J. P. & Lamond, A. I. (1998). Dynamic interactions between splicing snRNPs, coiled bodies Puvion-Dutilleul, F. & Christensen, M. E. (1993). Alterations of and nucleoli revealed using snRNP protein fusions to the green fluorescent protein. Exp Cell Res 243, 290–304. fibrillarin distribution and nucleolar ultrastructure induced by adenovirus infection. Eur J Cell Biol 61, 168–176. van Eenennaam, H., van der Heijden, A., Janssen, R. J., van Venrooij, W. J. & Pruijn, G. J. (2001). Basic domains target Russell, W. C., Webster, A., Leith, I. R. & Kemp, G. D. (1989). protein subunits of the RNase MRP complex to the nucleolus independently of complex association. Mol Biol Cell 12, 3680–3689. Phosphorylation of adenovirus DNA-binding protein. J Gen Virol 70, 3249–3259. Scheer, U. & Hock, R. (1999). Structure and function of the nucleolus. Curr Opin Cell Biol 11, 385–390. Shenk, T. (2001). Adenoviridae: the viruses and their replication. In Fields Virology, 4th edn, pp. 2265–2299. Edited by D. M. Knipe & P. M. Howley. Philadelphia: Lippincott Williams & Wilkins. 196 Vayda, M. E., Rogers, A. E. & Flint, S. J. (1983). The structure of nucleoprotein cores released from adenovirions. Nucleic Acids Res 11, 441–460. Webster, A., Leith, I. R., Nicholson, J., Hounsell, J. & Hay, R. T. (1997). Role of preterminal protein processing in adenovirus replication. J Virol 71, 6381–6389. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Sun, 18 Jun 2017 10:18:04 Journal of General Virology 85