Carboxyl-Proximal Regions of Reovirus Nonstructura
http://www.100md.com
病菌学杂志 2005年第10期
Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, Massachusetts 02115
Department of Biochemistry, University of Wisconsin-Madison, Madison, Wisconsin 53706
Ph.D. Programs in Virology
Biological and Biomedical Sciences, Division of Medical Sciences, Harvard University, Cambridge, Massachusetts 02138
ABSTRACT
Mammalian orthoreoviruses are believed to replicate in distinctive, cytoplasmic inclusion bodies, commonly called viral factories or viroplasms. The viral nonstructural protein μNS has been implicated in forming the matrix of these structures, as well as in recruiting other components to them for putative roles in genome replication and particle assembly. In this study, we sought to identify the regions of μNS that are involved in forming factory-like inclusions in transfected cells in the absence of infection or other viral proteins. Sequences in the carboxyl-terminal one-third of the 721-residue μNS protein were linked to this activity. Deletion of as few as eight residues from the carboxyl terminus of μNS resulted in loss of inclusion formation, suggesting that some portion of these residues is required for the phenotype. A region spanning residues 471 to 721 of μNS was the smallest one shown to be sufficient for forming factory-like inclusions. The region from positions 471 to 721 (471-721 region) includes both of two previously predicted coiled-coil segments in μNS, suggesting that one or both of these segments may also be required for inclusion formation. Deletion of the more amino-terminal one of the two predicted coiled-coil segments from the 471-721 region resulted in loss of the phenotype, although replacement of this segment with Aequorea victoria green fluorescent protein, which is known to weakly dimerize, largely restored inclusion formation. Sequences between the two predicted coiled-coil segments were also required for forming factory-like inclusions, and mutation of either one His residue (His570) or one Cys residue (Cys572) within these sequences disrupted the phenotype. The His and Cys residues are part of a small consensus motif that is conserved across μNS homologs from avian orthoreoviruses and aquareoviruses, suggesting this motif may have a common function in these related viruses. The inclusion-forming 471-721 region of μNS was shown to provide a useful platform for the presentation of peptides for studies of protein-protein association through colocalization to factory-like inclusions in transfected cells.
INTRODUCTION
Viruses with ten-segmented, double-stranded RNA genomes from the family Reoviridae, genus Orthoreovirus, are believed to replicate in distinctive, cytoplasmic inclusion bodies (7, 10, 11, 13, 23, 24, 30, 33, 37-39, 41, 44, 45). These inclusions are commonly called viral factories (13, 30) or viroplasms (45) and are similar to cytoplasmic inclusions formed by other viruses in the same family. In cells infected by rotaviruses or orbiviruses, for example, these structures are called viroplasms (12, 40) or viral inclusion bodies (8, 43), respectively.
Many viruses sequester their replication machinery within localized structures or surfaces in infected cells: for example, herpes simplex virus (double-stranded DNA genome, family Herpesviridae) in nuclear inclusions (reviewed in reference 35), vaccinia virus (double-stranded DNA genome, family Poxviridae) in cytoplasmic inclusions (also called viral factories [reviewed in reference 29]), brome mosaic virus (single-stranded RNA [ssRNA] genome, family Bromoviridae) on the cytoplasmic face of the endoplasmic reticulum (32, 36), and flock house virus (ssRNA genome, family Nodaviridae) on the cytoplasmic face of mitochondria (27). The basis and consequences of such specific localizations are the subjects of active investigations in many laboratories. We anticipate that studies of mammalian orthoreovirus (reovirus) factories in our own laboratory will provide new insights into the still poorly characterized mechanisms for RNA packaging and replication by these and other viruses in the family Reoviridae, as well as on the general significance and mechanism of concentrating viral replication at particular sites within cells.
In early studies, reovirus factories were determined to contain fully and partially assembled viral particles, viral proteins, double-stranded RNA, microtubules, and "kinky" filaments proposed to be intermediate filaments but not membrane-bound structures or ribosomes (10, 11, 23, 33, 37, 39, 41). The factories have a peculiarly dense consistency that distinguishes them from the adjacent cytoplasm and causes them to appear highly refractile by phase-contrast microscopy. At least part of this property appears to reflect a protein matrix that suffuses the factories. The determinants or features of one or more viral proteins that would make it capable of forming such a matrix are not well understood but might involve a variety of different types of intersubunit interactions, as well as interactions with cellular factors.
Results from our laboratory and others have recently shown that a single reovirus protein, μNS, is sufficient for forming phase-dense globular inclusions in the cytoplasm of transfected cells (4, 7). The inclusions formed by μNS in such experiments are notably similar in appearance to globular reovirus factories formed in infected cells, as visualized by either phase-contrast or immunofluorescence (IF) microscopy (4, 7, 30), suggesting that μNS forms the matrix of the factories (7). Moreover, μNS can associate with several other reovirus proteins and recruit them to these inclusions in transfected cells. To date, the viral proteins that have been reported to be recruited to inclusions formed by reovirus μNS are microtubule-binding core protein μ2 (7); nonstructural and ssRNA-binding protein NS (4, 26); and the core surface proteins 1, 2, and 2 (6). Whole core particles released into the cytoplasm during cell entry are also recruited to μNS inclusions (6). Indirect evidence suggests that μNS may possess ssRNA-binding activity as well, which may be important for recruiting the viral plus-strand RNAs to the factories and/or retaining them there for replication and packaging in infected cells (1). However, the ssRNA-binding activity of NS is more firmly established (15-17, 19, 34, 42) and may play the more important role in ssRNA recruitment to or retention within the factories (4, 6, 26). Recent studies on the μNS homolog from avian orthoreoviruses have reached conclusions similar to these regarding the roles of μNS in forming the factory matrix and recruiting other viral proteins to the factories (44, 45). We explicitly specify in the present study when we intend a statement to encompass results for avian orthoreovirus; otherwise, we use the abbreviated name reovirus to represent mammalian orthoreovirus only.
The 80-kDa μNS protein is not a component of mature virions but is expressed to high levels in infected cells and is concentrated in the reovirus factories (7, 49). Other previous findings about μNS include its association with transcriptionally active particles isolated from infected cells (28), its binding to the surfaces of purified core particles in vitro (5), and its capacity to recruit entering core particles to inclusions (6). The μNS protein has two predicted coiled-coil segments in the carboxyl-terminal one-third of its sequence (25), but the oligomeric status of μNS has not been demonstrated. Little else is known about the structure of μNS, although domains have been predicted from the pattern of conserved and variable regions of sequence among the μNS alleles from different reovirus isolates (25). An isoform of μNS lacking 5 kDa from its amino terminus and called μNSC is also present in virus-infected cells (22, 46). The origin of μNSC is not certain, but it is proposed to result from either secondary initiation or cleavage near Met41 in μNS. A recombinant protein engineered to lack the first 40 residues of μNS, μNS(41-721), which should be similar to μNSC, also forms phase-dense globular inclusions in transfected cells (7).
Involvement of microtubules in forming reovirus factories has also been demonstrated. The M1 segment, which encodes the structurally minor core protein μ2, has been shown to determine both the timing of factory formation (24) and the morphology of factories (30). The filamentous morphology of factories formed by most reovirus strains has been attributed to microtubule association of the μ2 protein (30). When μNS and μ2 are coexpressed in cells without other viral proteins, μNS is redistributed with μ2 to filamentous inclusions (7), suggesting that these two proteins together determine the formation and morphology of filamentous viral factories. Aside from any effects of μ2, however, μNS also requires microtubules for undergoing condensation of its smaller inclusions into larger ones, as shown by the effects of the microtubule-depolymerizing drug nocodazole. This probably reflects a role for microtubule-based transport in the formation of μNS inclusions and the determination of viral factory morphology (7, 30).
By forming the matrix of viral factories, as well as recruiting other components to these structures, the reovirus μNS protein could serve to concentrate the components for viral replication and assembly, to arrange the components in specific ways to promote replication and assembly, and/or to sequester the components and their mediated functions from antiviral factors in the cellular environment. To better understand the role of μNS in forming the matrix of viral factories, we engineered constructs to express a panel of μNS truncations and point mutants and thereby determined the regions of this protein that are involved in forming factory-like inclusions in transfected cells. The results identified several different sequences in the C-terminal one-third of μNS that are necessary and sufficient for this activity of μNS.
MATERIALS AND METHODS
Cells and antibodies. CV-1 cells were maintained in Dulbecco modified Eagle medium (Invitrogen Life Technologies) containing 10% fetal bovine serum (HyClone Laboratories) and 10 μg of gentamicin solution (Invitrogen Life Technologies) per ml. Goat anti-mouse immunoglobulin G (IgG) and goat anti-rabbit IgG conjugated to Alexa 488 or Alexa 594 were obtained from Molecular Probes. Rabbit polyclonal antisera to μNS or μ2 have been described previously (5, 7, 30). For some experiments, we used protein A-purified rabbit anti-μNS or anti-μ2 IgG conjugated to Texas Red, Oregon Green, Alexa 488, or Alexa 594 by using kits obtained from Molecular Probes. Mouse monoclonal antibody (MAb) FK2 against conjugated ubiquitin (14) was purchased from Medical & Biological Laboratories. Mouse MAb JL8 against Aequorea victoria green fluorescent protein (GFP) was purchased from BD Biosciences. Mouse MAb HA.11 against an immunodominant epitope of influenza A virus hemagglutinin (HA) was purchased from Covance. Mouse MAb 3E10 specific for reovirus NS protein (3) was a gift from T. S. Dermody and coworkers (Vanderbilt University). All antibodies were titrated to optimize signal-to-noise ratios.
Expression constructs. Reovirus proteins were expressed from genes cloned into the mammalian expression vector pCI-neo (Promega). pCI-M3(T1L) and pCI-M3(T3D) to express μNS from type 1 Lang (T1L) or type 3 Dearing (T3D) reovirus have been described previously (7), as have pCI-M3(41-721) to express μNS residues 41 to 721 (7), pCI-M1(T1L) to express μ2 (30), and pCI-S3(T1L) to express NS (26). Vent polymerase, which was used for all PCRs, and other enzymes were from New England Biolabs unless otherwise stated.
To express μNS residues 1 to 683, site-directed mutagenesis was used to introduce a stop codon at nucleotides 2068 to 2070, followed by a Bsu36I site. QuikChange site-directed mutagenesis (Stratagene) was used according to the manufacturer's protocol with pFastBac-M3(T1L) (5) as a template, forward primer 5'-GGATACGATGAACTAACCTCAGGCTAAATCATTGCG-3', and reverse primer 5'-CGCAATGATTTAGCCTGAGGTTAGTTCATCGTATCC-3' (double underline, nucleotide change to add the stop codon; single underline, nucleotide change to add the restriction site). The region containing the desired mutations was excised by digestion with HindIII and then ligated to pFastBac-M3(T1L) that had been cut with HindIII to remove the same region, generating pFastBac-M3(1-683). The subcloned region was sequenced to confirm its correctness. The M3 gene was removed from pFastBac-M3(1-683) by digestion with SalI and NheI and then ligated to pCI-neo that had been cut with the same enzymes, generating pCI-M3(1-683).
To generate other μNS truncations, start and stop codons and restriction sites were introduced at different positions in the M3 gene by PCR amplification of the desired M3 region. The truncations were made in either T1L or T3D M3. We have found no difference in inclusion formation with T1L and T3D μNS (7, 26), suggesting that the truncations from these allelic μNS proteins are directly comparable. In addition, the T1L and T3D μNS proteins are 96% identical (25). Each PCR was performed with pGEM-4Z-M3(T1L) or pGEM-4Z-M3 (T3D) (5) as a template, and the primers are listed in Table 1. Each PCR product was cut with the restriction enzymes listed in Table 1 and then ligated to a plasmid that had been cut with these same enzymes, the plasmid being either pGEM-4Z (for pGEM-4Z-M3(1-173) and pGEM-4Z-M3(1-221) or the template plasmid (for all other constructs). The correctness of each construct was confirmed by sequencing. The truncated M3 genes were subcloned into pCI-neo by using the restriction enzymes listed in Table 2. Each construct was named for the residues of μNS that the expressed protein should ultimately contain (Tables 1 and 2).
The pEGFP-N1 and pEGFP-C1 vectors (BD Biosciences) were respectively used to express fusions of enhanced A. victoria GFP to the C or N terminus of the desired μNS region. pEGFP-N1-M3(T1L) was previously constructed to express GFP fused to the C terminus of μNS (μNS/GFP) (7). To express GFP fused to the N terminus of μNS residues 471 to 721, pGEM-4Z-M3(471-721) was cut with SalI and KpnI, and the excised fragment was then ligated to pEGFP-C1 that had been cut with the same enzymes, generating pEGFP-C1-M3(471-721). To express GFP fused to the N termini of even smaller C-terminal regions of μNS, an EcoRI site was introduced at the desired position in the M3 gene during PCR amplification. Each PCR was performed with pGEM-4Z-M3(T1L) (5) as a template, the forward primers listed in Table 3, and a reverse primer complementary to the 3' end of the M3 coding strand with an added BamHI site (underlined) (5'-GCAGGGGATCCGATGAATGGGGGTCGGGAAGGCTTAAGGG-3'). Each PCR product was cut with EcoRI and BamHI and was then ligated to pEGFP-C1 that had been cut with the same enzymes. The correctness of each construct was confirmed by sequencing, and the construct was named for the residues of μNS that the expressed protein contains (Table 3).
To generate HA-tagged fusions, epitope-encoding forward primer 5'-CGTAGCTAGCGTCATGGCTTACCCATACGACGTCCCAGACTACGCTCTCGAGATGC-3' and reverse primer 5'-GCATCTCGAGAGCGTAGTCTGGGACGTCGTATGGGTAAGCCATGACGCTAGCTACG-3' (underlining indicates NheI and XhoI sites added to each primer) were annealed by boiling in 1x SSC (150 mM NaCl plus 15 mM sodium citrate [pH 7.0]), followed by cooling at room temperature for 1 h. The duplex oligonucleotide and vector pCI-neo were then digested with NheI and XhoI, and the products were ligated to generate pCI-neo-HA. The region encoding μNS(561-721) was amplified by PCR from the template pCI-M3(T1L) by using forward primer 5'-GCTAGAATTCATGTAGTCTGGATATGTATTTGAGACACCAC-3' and reverse primer 5'-GATCGATCCCGGGTCGGGAAGGCTTAAGGGATTAGGGCAA-3' (underlining indicates an EcoRI or SmaI site added to each primer). The PCR product and vector pCI-neo-HA were then sequentially digested with SmaI and EcoRI, and the products were ligated to generate pCI-neo-HA-M3(561-721). The region encoding μNS(471-721) was amplified by PCR from the template pCI-M3(T1L) by using the forward primer 5'-AGCTCTCGAGGTCATGTCCAGTGACATGGTAGACGGGATTAAAC-3' (underlining indicates an added XhoI site) and reverse primer 5'-CGAAGCATTAACCCTCAC-3'. The PCR product and vector pCI-neo-HA were then digested with XhoI and NotI, and the products were ligated to generate pCI-neo-HA-M3(471-721). To generate untagged μNS(561-721), PCR was performed by using the template pCI-M3(T1L), forward primer5'-AGCTGAATTCGTCATGGCTTGTAGTCTGGTATGTATTTGAGACAC-3' (underline indicates an added EcoRI site), and reverse primer 5'-CGAAGCATTAACCCTCAC-3'. The PCR product and vector pCI-neo were then digested with EcoRI and NotI, and the products were ligated to generate pCI-neo-M3(561-721). The correctness of the final constructs was confirmed by sequencing.
To generate a μNS(471-721) or full-length μNS protein with residues His569 and His570 changed to glutamine and Cys572 changed to serine, QuikChange site-directed mutagenesis was used according to the manufacturer's protocol with pCI-M3(471-721) or pCI-M3(T3D) as a template, forward primer 5'-GGATATGTATCTGCGACAACAAACTTCCATTAATGGTCATGC-3', and reverse primer 5'-GCATGACCATTAATGGAAGTTTGTTGTCGCAGATACATATCC-3' (double underlines, nucleotide changes to provide amino acid changes; single underline, nucleotide change to remove a DdeI site for screening purposes). After the QuikChange protocol, the mutated pCI-M3(471-721) or pCI-M3(T3D) plasmid was prepared for subcloning by cutting with NheI and EcoRI or by cutting with BlpI and NotI, respectively. The excised fragments were then ligated to pCI-neo or pCI-M3(T3D), respectively, that had been cut with the same respective pair of enzymes, generating the final versions of pCI-M3(471-721)QQS and pCI-M3(T3D)QQS. The correctness of these constructs was confirmed by sequencing. To generate a full-length μNS protein with Cys561 or Cys572 changed to Ser or His569, His570, or His576 changed to Gln, QuikChange site-directed mutagenesis was used according to the manufacturer's protocol with pCI-M3(T3D) as a template and the primers listed in Table 4. According to the QuikChange protocol, each mutated plasmid was prepared for subcloning by cutting with BlpI and NotI. In each case, the excised fragment was then ligated to pCI-M3(T3D) that had been cut with the same enzymes, generating the constructs as named in Table 4. The correctness of each subcloned region was confirmed by sequencing.
To generate a fusion of μNS residues 1 to 41 to the N terminus of GFP and μNS residues 471 to 721 to the C terminus of GFP, DNA encoding residues 1 to 41 of μNS and a large portion of GFP was removed from pEGFP-N1-M3(1-41) (7) by digestion with NheI and BsrGI. pEGFP-C1-M3(471-721) was also cut with these enzymes, and the two DNA fragments were ligated to generate pEGFP-M3(1-41)/GFP/M3(471-721). To generate a fusion of μNS residues 1 to 12 to the N terminus of GFP and μNS residues 471 to 721 to the C terminus of GFP, forward primer 5'-AATTC ATGGCTTCATTCAAGGGATTCTCCGTCAACACTGTTGCG-3' and reverse primer 5'-GATCCGCAACAGTGTTGACGGAGAATCCCTTGAATGAAGCCATG-3' were annealed to generate a small duplex encoding μNS residues 1 to 12 and having BamHI and EcoRI overhangs at the respective ends (underlined). This small duplex and plasmid pEGFP-N1 were both digested with BamHI and EcoRI and then ligated to yield pEGFP-N1-M3(1-12). This plasmid was then digested with NheI and BsrGI, and the excised fragment was ligated to pEGFP-C1-M3(471-721) that had been digested with the same enzymes, generating plasmid pEGFP-M3(1-12)/GFP/M3(471-721). The correctness of the resulting construct was confirmed by sequencing.
Transfections and IF microscopy. Cells were seeded the day before transfection at a density of 1.5 x 104 per cm2 in six-well plates (9.6 cm2 per well) containing glass coverslips (19 mm). Cells were transfected with 2 μg of DNA and 7 μl of Lipofectamine 2000 (Invitrogen Life Technologies) or 1.5 μl of DNA and 10 μl of Polyfect (Qiagen) according to the manufacturer's directions. Cells were further incubated for 18 to 24 h at 37°C before fixation for 10 min at room temperature in 2% paraformaldehyde in phosphate-buffered saline (PBS; 137 mM NaCl, 3 mM KCl, 8 mM Na2HPO4, 1 mM KH2PO4 [pH 7.5]) or 3 min at –20°C in ice-cold methanol. Fixed cells were washed three times with PBS, permeabilized, and blocked in PBS containing 1% bovine serum albumin and 0.1% Triton X-100 (PBSAT). Primary antibodies were diluted in PBSAT and incubated with cells for 25 to 40 min at room temperature. After three washes in PBS, secondary antibodies diluted in PBSAT were added and incubated with cells for 25 min at room temperature. Coverslips were incubated with 300 nM 4,6-diamidino-2-phenylindole (DAPI; Molecular Probes) in PBS for 5 min to counterstain cell nuclei, briefly washed in PBS, and mounted on glass slides with Prolong (Molecular Probes). Samples were examined by using a TE-300 inverted microscope (Nikon) equipped with phase and fluorescence optics, and images were collected digitally as described elsewhere (30). All images were processed and prepared for presentation by using Photoshop (Adobe Systems).
Immunoblot analysis. CV-1 cells were transfected as described for IF, and whole-cell lysates were collected 18 to 24 h posttransfection (p.t.). CV-1 cells (1.2 x 106) were washed briefly in PBS and then scraped into 1 ml of PBS and pelleted. The pelleted cells were resuspended in 30 μl of PBS containing protease inhibitors (Roche Biomedicals), lysed into sample buffer, boiled for 10 min, and subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were electroblotted from the gels to nitrocellulose in 25 mM Tris and 192 mM glycine (pH 8.3). Binding of antibodies was detected with alkaline phosphatase-coupled goat anti-mouse IgG (Bio-Rad Laboratories) and the colorimetric reagents p-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolylphosphate p-toluidine salt (Bio-Rad Laboratories).
Sequence comparisons. The following μNS and μNS-homolog sequences, each designated as complete coding sequences in GenBank, were compared: μNS of mammalian orthoreoviruses T1L (accession no. AAF13169), type 2 Jones (accession no. AAF13170), and T3D (accession no. AAF13171); μNS of avian orthoreoviruses 1733 (accession no. AAQ81873), S1133 (accession no. AAS78987), 1017-1 (accession no. AAS78988), 2408 (accession no. AAS78990), 601G (accession no. AAS78991), 750505 (accession no. AAS78992), 916SI (accession no. AAS78993), 918 (accession no. AAS78994), 919 (accession no. AAS78995), OS161 (accession no. AAS78996), R2 (accession no. AAS78997), and T6 (accession no. AAS78998); and NS1 of aquareoviruses grass carp 873 (accession no. AAM92735) and golden shiner (accession no. AAM92747). Sequences were compared by using the Bestfit and Pretty programs from the Genetics Computer Group Wisconsin package. The Multicoil program (http://multicoil.lcs.mit.edu/cgi-bin/multicoil) (48) was used for predicting coiled-coil regions.
RESULTS
The C terminus of μNS is required for forming factory-like inclusions. We have previously shown that a truncated μNS protein lacking the N-terminal 40 residues, μNS(41-721), collects in cytoplasmic, factory-like inclusions in transfected cells, similarly to full-length μNS (7). The N terminus of μNS is thus not required for forming these structures. To determine whether the C terminus of μNS is required, we engineered M3 gene constructs to express a series of truncated μNS proteins that lacked increasing numbers of residues from the C terminus (Table 2). Each of these plasmids was transfected into CV-1 cells, and the cell lysates were subjected to SDS-PAGE, followed by immunoblotting with polyclonal anti-μNS serum (5). The results verified the expression of an appropriately sized, μNS-derived protein from each construct (data not shown).
To determine the intracellular distribution of the C-terminally truncated μNS proteins, CV-1 cells were separately transfected with each of the plasmids and later fixed and immunostained with polyclonal anti-μNS antibodies (Fig. 1A, left and right columns). In addition, the cells were coimmunostained with MAb FK2, which recognizes conjugated ubiquitin (14), to determine whether any of these proteins were substantially misfolded and targeted for degradation by the ubiquitin-proteasome system (Fig. 1A, center and right columns). Each of the C-terminally truncated proteins was diffusely distributed in the cytoplasm and nucleus, in clear contrast to full-length μNS, which collected in globular inclusions as expected (Fig. 1A and data not shown). We conclude that all of these truncated proteins are negative for forming factory-like inclusions (summarized in Fig. 2) and thus that some portion of the smallest region we deleted from the C terminus of μNS, residues 714 to 721, is required for inclusion formation. None of the truncated proteins appeared to be aggregated or strongly colocalized with conjugated ubiquitin (Fig. 1A and data not shown), suggesting that they were not substantially misfolded.
All of the C-terminally truncated μNS proteins appeared to be partially localized to the nucleus (Fig. 1A, left column), even though μNS(1-683), μNS(1-700), and μNS(1-713) were above the 60-kDa limit for passive diffusion through nuclear pores (reviewed in reference 31) (Table 2). In addition, at least some of the truncated proteins appeared to be excluded from nucleoli (Fig. 1A, left column). The relevance of this nuclear pattern is unclear because full-length μNS is not detected in the nuclei of infected or M3-transfected cells (7, 30). When coexpressed with μ2(T1L), each of the C-terminally truncated proteins strongly colocalized with μ2 on microtubules (Fig. 1B and data not shown; summarized in Fig. 2). This was expected because each protein included μNS residues 1 to 41, which are known to be sufficient for association with μ2 (7, 26).
The N-terminal 470 residues of μNS are not required for forming factory-like inclusions but affect inclusion shape. To identify the minimal region of μNS required for inclusion formation, we engineered M3 gene constructs to express a series of truncated μNS proteins that lacked increasing numbers of residues from the N terminus (Table 2). Each of these plasmids was transfected into CV-1 cells, and the cell lysates were subjected to SDS-PAGE, followed by immunoblotting with the anti-μNS serum (5). The results verified expression of an appropriately sized, μNS-derived protein from each construct (data not shown).
The intracellular distribution of each of the N-terminally truncated μNS proteins was determined as described above for the C-terminal truncations, including coimmunostaining with MAb FK2. With the exception of μNS(363-721), which showed a distinctly aggregated pattern and strongly colocalized with conjugated ubiquitin (Fig. 3), each of the N-terminally truncated proteins collected in globular inclusions (Fig. 3, left and right columns, and data not shown). However, the proteins missing more than the 40 N-terminal residues of μNS often formed inclusions with more elongated shapes and enclosed fenestrations than those formed by full-length μNS or μNS(41-721) (Fig. 3, left and right columns). Although these inclusions had altered morphologies, they did not colocalize with conjugated ubiquitin (Fig. 3, center and right columns), suggesting that these truncated proteins were not substantially misfolded. The behavior of μNS(363-721), on the other hand, suggested that it was in fact largely misfolded.
When coexpressed with μ2(T1L), none of the N-terminally truncated μNS proteins colocalized with μ2 on microtubules or in inclusions (data not shown; summarized in Fig. 2). This was expected because each of these proteins lacks μNS residues 14 to 40, which are known to be required for association with μ2 (7, 26). From the results summarized in Fig. 2, we conclude that μNS residues 471 to 721 are a sufficient part of μNS for forming factory-like inclusions. These residues encompass the two predicted coiled-coil segments of μNS, the intervening "linker" between these segment, and the C-terminal "tail" that follows the second predicted coiled-coil segment (25). Given the altered morphologies of the inclusions formed by μNS proteins lacking more than the 40 N-terminal residues, we also conclude that some portion of residues 41 to 172 plays a distinguishable role in modulating inclusion morphology.
Residues 561 to 721 of μNS are sufficient for allowing a GFP-tagged protein, but not an HA-tagged or untagged protein, to form factory-like inclusions. To determine whether a region of μNS smaller than residues 471 to 721 may be sufficient for inclusion formation, and also to allow more ready detection of these smaller proteins, we next generated constructs to express GFP fused to the N terminus of selected μNS truncations (Table 3). Each of these plasmids was transfected into CV-1 cells, and the cell lysates were subjected to SDS-PAGE, followed by immunoblotting with GFP-specific MAb JL-8. The results verified production of an appropriately sized, GFP- and μNS-derived fusion protein from each construct (Fig. 4A).
To determine the intracellular distribution of each fusion protein, CV-1 cells were transfected and later immunostained with the GFP-specific MAb. Nonfused GFP was diffusely distributed in the cytoplasm and nucleus, and GFP fused to the C terminus of full-length μNS (μNS/GFP) collected in globular inclusions as previously shown (7) (also see Fig. 4B). GFP fused to the N terminus of μNS(471-721) [GFP/μNS(471-721)] also collected in inclusions (Fig. 4B), which appeared similar to those formed by μNS(471-721) (Fig. 3) or μNS/GFP (Fig. 4B). GFP fused to the N terminus of μNS(561-721) collected in inclusions as well, although a small fraction of the transfected cells (13%) displayed a diffuse distribution of this protein (Fig. 4B and data not shown). In contrast, GFP fused to the N terminus of μNS(614-721), μNS(625-721), or μNS(695-721) was diffusely distributed in the cytoplasm and nucleus of most or all cells expressing them (Fig. 4B and data not shown). From these results, we conclude that residues 561 to 721 are a sufficient part of μNS in GFP fusions for forming factory-like inclusions (summarized in Fig. 2), albeit at a lower efficiency than full-length μNS or μNS(471-721). The first predicted coiled-coil segment (25) is thus dispensable for inclusion formation by a μNS-GFP fusion. On the other hand, some portion of residues 561 to 613, in the linker between the two predicted coiled-coil segments, is required.
Given that A. victoria GFP is known to weakly dimerize (9) and also that GFP/μNS(561-721) was less efficient at forming factory-like inclusions than was GFP/μNS(471-721), we hypothesized that the GFP tag may partially complement an otherwise-required contribution of the first predicted coiled-coil segment of μNS for forming inclusions. We therefore tested another version of μNS(561-721), this one fused to an epitope of influenza virus HA (47) at its N terminus [HA/μNS(561-721)]. After expression in transfected CV-1 cells and immunostaining with tag-specific MAb HA.11, HA/μNS(561-721) was diffusely distributed in the cytoplasm and nucleus and thus not concentrated in inclusions (Fig. 5). Fusion of the HA tag to the N terminus of μNS(471-721) [HA/μNS(471-721)], in contrast, caused little or no reduction in its capacity to form factory-like inclusions (Fig. 5). We also generated and tested a third version of μNS(561-721), this one lacking any tag. After expression in transfected CV-1 cells and immunostaining with anti-μNS antibodies, untagged μNS(561-721) was diffusely distributed in the cytoplasm and nucleus and thus not concentrated in inclusions (Fig. 5). Untagged μNS(471-721) tested in parallel formed factory-like inclusions (Fig. 5) as seen in the preceding experiments (see Fig. 3). Based on these results, we conclude that the more N-terminal predicted coiled-coil segment of μNS is required for inclusion formation in the absence of a fusion tag such as GFP that can independently self-associate.
Putative metal-chelating residues His570 and Cys572 are required for factory-like inclusion formation. In an effort to identify specific residues in μNS that may be required for inclusion formation, we compared the deduced protein sequences of μNS homologs from mammalian and avian orthoreoviruses, as well as from aquareoviruses (see Materials and Methods for the GenBank accession numbers) (2, 25, 45). In all, the μNS homologs derived from 17 isolates in the three groups: 3 mammalian isolates from the Orthoreovirus genus, 12 avian isolates from the Orthoreovirus genus, and 2 piscine isolates from the Aquareovirus genus. The overall identity scores for any two μNS homologs from separate groups are small: <30% in each case (2, 45) (the present study and data not shown). Despite this degree of divergence, the C-terminal one-third of each of these μNS homologs contains two predicted coiled-coil segments, of similar lengths and spacing, separated by a linker and followed by a C-terminal tail (data not shown), as previously described for the mammalian and avian isolates (25, 45).
Interestingly, in the linker between the predicted coiled-coil segments, we identified a small consensus motif common to all of the examined μNS homologs (Fig. 6A). This sequence,Ile/Leu-x-x-Tyr-Leu-x-x-His-Thr/Val-Cys-Ile/Val-Asn (where"x" represents nonconserved positions), includes two residues (underlined) with strong potential to chelate transition metal ions such as Zn2+. The two residues correspond to His570 and Cys572 in the mammalian orthoreovirus μNS proteins. Each of the μNS homologs contains other His and/or Cys residues flanking the consensus motif, but the position and spacing of these residues is not conserved among the examined sequences (Fig. 6A). In mammalian orthoreovirus μNS, the other conserved His and Cys residues in this region are Cys561, His569, and His576. The consensus motif spans residues 563 to 574 in the mammalian orthoreovirus μNS proteins and is thus near the beginning of the minimal C-terminal region of μNS that we showed to be sufficient for inclusion formation in GFP fusions (Fig. 4).
To determine whether the conserved residues with metal-chelating potential in the consensus motif are important for inclusion formation, we first generated constructs encoding Gln (Q) substitutions for both His570 and the adjacent residue, His569, as well as a Ser (S) substitution for Cys572 (Fig. 6A). The constructs were generated with all three of these mutations in the setting of either full-length μNS or μNS(471-721). We then transfected CV-1 cells with these mutant plasmids and costained the cells with anti-μNS antibodies and MAb FK2 for conjugated ubiquitin. Both proteins, designated μNS(1-721)QQS and μNS(471-721)QQS, were diffusely distributed in the cytoplasm and nucleus (Fig. 6B and data not shown), and neither strongly colocalized with conjugated ubiquitin (data not shown). This distribution was in sharp contrast to the inclusions in which both full-length μNS and μNS(471-721) concentrated (data not shown for this experiment, but see previous figures), demonstrating that one or more of residues His569, His570, and Cys572 is important for inclusion formation.
We next generated constructs encoding single mutations at residues Cys561 (to Ser), His569 (to Gln), His570 (to Gln), Cys572 (to Ser), or His576 (to Gln) within full-length μNS. The mutant plasmids were transfected into CV-1 cells, and the cells were costained with anti-μNS antibodies and MAb FK2 for conjugated ubiquitin. Both μNS(1-721)H570Q and μNS(1-721)C572S were diffusely distributed in the cytoplasm and nucleus (Fig. 6B). In contrast, μNS(1-721)C561S, μNS(1-721)H569Q, and μNS(1-721)H576Q all collected in globular inclusions indistinguishable from those formed by wild-type μNS (Fig. 6B). None of the mutant proteins strongly colocalized with conjugated ubiquitin (data not shown). From these findings, we conclude that His570 and Cys572 are specifically required for μNS to form factory-like inclusions in transfected cells.
NS recruitment to factory-like inclusions containing μNS residues 1 to 12 connected by GFP to μNS residues 471 to 721. We have previously shown that NS colocalizes with full-length μNS but not with μNS(41-721) or μNS(13-721) in cotransfected cells, indicating that the N terminus of μNS is required for recruiting NS (7, 26). Because both NS and the N-terminal 41 residues of μNS fused to the N terminus of GFP [μNS(1-41)/GFP] are diffusely distributed in cells, we encountered difficulties in using IF microscopy to determine whether the N terminus of μNS is sufficient for association with NS (7, 26). Upon finding in the present study that GFP/μNS(471-721) collected in factory-like inclusions in transfected cells (Fig. 4B), we recognized that this protein provided a new platform on which to assay protein-protein associations through redistribution to its distinctive inclusions. When GFP/μNS(471-721) was coexpressed with NS, NS remained diffusely distributed in the cytoplasm and nucleus and did not colocalize with the GFP/μNS(471-721) inclusions (Fig. 7A). This was expected, because some portion of μNS residues 1 to 13 is required for recruiting NS to μNS inclusions (26).
To determine whether μNS residues 1 to 41 can direct recruitment of NS to the inclusions of GFP/μNS(471-721), we constructed a plasmid to express these residues fused to the N terminus of GFP/μNS(471-721). When expressed in transfected CV-1 cells, this protein, μNS(1-41)/GFP/μNS(471-721), collected in inclusions similar to those formed by μNS(471-721) or GFP/μNS(471-721) (data not shown). When NS was coexpressed with μNS(1-41)/GFP/μNS(471-721) and coimmunostained with anti-NS MAb 3E10 (3) and the anti-μNS serum, NS strongly colocalized with the μNS(1-41)/GFP/μNS(471-721) inclusions (Fig. 7A). This was in stark contrast to the diffuse distribution of NS expressed alone (26) or coexpressed with GFP/μNS(471-721) (Fig. 7A). When coexpressed with μ2, μNS(1-41)/GFP/μNS(471-721) was redistributed to filamentous inclusions (data not shown), as expected because of the presence of μNS residues 14 to 40 (26) (also see Fig. 2 and 7B). These results newly demonstrate that μNS residues 41 to 470 are dispensable for recruiting NS (summarized in Fig. 7B).
We additionally constructed a plasmid to express μNS residues 1 to 12 fused to the N terminus of GFP/μNS(471-721). When expressed in transfected CV-1 cells, this protein, μNS (1-12)/GFP/μNS(471-721), collected in factory-like inclusions (data not shown). Moreover, when coexpressed, NS was strongly recruited to the μNS(1-12)/GFP/μNS(471-721) inclusions (Fig. 7A). When coexpressed with μ2, μNS(1-12)/GFP/μNS(471-721) was not redistributed to filamentous inclusions (data not shown), as expected because of the absence of μNS residues 13 to 41 (26). These results demonstrate that μNS residues 13 to 470 are dispensable for recruiting NS (summarized in Fig. 7B) and suggest that μNS residues 1 to 12 may be sufficient for this association.
DISCUSSION
Numerous observations suggest that the inclusion bodies in reovirus-infected cells are sites of viral replication and assembly (4, 6, 33, 38, 39). The roles of nonstructural protein μNS in forming the matrix of these viral factories and recruiting other components to them have become a focus of recent studies in this field (4, 6, 7). Given the recent findings and the absence of strong evidence for other functions of this protein, we hypothesize that the primary role of μNS in reovirus infection is to build and organize the factories to promote replication and assembly. We find it intriguing that reovirus may have a protein dedicated to this role, and we therefore wish to learn more about how the different regions of μNS carry out specific activities toward this end.
Current summary of μNS functional regions. Truncation and deletion analyses have been a fruitful approach for identifying discrete regions of μNS involved in its different activities. For example, the current study demonstrated that the C-terminal one-third of μNS (residues 471 to 721) is a sufficient part of this protein for forming factory-like inclusions in transfected cells in the absence of infection or other viral proteins. The present study also identified several smaller regions of sequence, and even single residues, within this third of μNS that are required for inclusion formation.
An otherwise-uncharacterized region of μNS (residues 41 to 172) affected the shape of the inclusions in the present study, causing them to be more compact and less fenestrated when present in the inclusion-forming protein. This effect could reflect a direct interaction of some portion of residues 41 to 172 with the C-terminal, inclusion-forming region of μNS or with some cellular protein, which in an unknown manner produces more compact inclusions. This effect could also reflect an indirect mechanism, such as a reduction in the turnover rate of the inclusion-forming protein such that holes do not develop within the inclusions.
Previous reports have shown that the N-terminal 40 residues of μNS are involved in associations with at least two other reovirus proteins: microtubule-binding protein μ2 and ssRNA-binding protein NS (7, 26). Moreover, the specific residues necessary or sufficient for these activities have been shown to be separable. Residues 1 to 41 of μNS are sufficient for association with μ2 (7), but only some portion of residues 14 to 40 of μNS is necessary for this activity (26). In contrast, some portion of residues 1 to 13 of μNS is necessary for association with NS (26), and residues 1 to 12 of μNS may be sufficient for this activity (the present study). Other, complementary evidence suggests the N-terminal 40 residues of μNS represent a discrete domain. A form of μNS lacking 5 kDa of N-terminal sequence is produced in infected cells concomitantly with the full-length protein (7, 46). Although the origin of this smaller form, called μNSC, remains in question, the fact that it is similar to μNS(41-721), which lacks both μ2 and NS association activities, suggests that it and full-length μNS may play distinguishable roles in infection. Interestingly, a μNSC-like form of μNS has been recently identified in avian orthoreovirus and found not to associate with the avian orthoreovirus NS protein as well, suggesting that the different activities of μNS and μNSC in this regard may be a conserved point of regulation for reoviruses in general (45). The avian orthoreovirus μ2 protein has not yet been examined for μNS association.
Except for the modulation of inclusion shape described in the present study and a predicted coiled-coil segment from residues 518 to 561 (25), the large central portion of μNS spanning residues 41 to 560 remains without well-characterized activities. This region from positions 41 to 560 may contain the residues responsible for μNS associations with other reovirus components, including the individual core-surface proteins, 1, 2, and 2 and whole core particles (6, 7). In fact, these activities are shared by μNS and μNSC, indicating they do not require the N-terminal 40 residues of μNS. Further truncation/deletion analyses are needed to determine whether each of the individual core-surface proteins, as well as core particles, may associate with μNS/μNSC through a distinct set of residues, such as those for μ2 and NS in the unique N-terminal region of μNS.
Other components with which μNS/μNSC may yet be shown to associate include the reovirus RNA-dependent RNA polymerase 3, the reovirus outer-capsid proteins, the reovirus RNA molecules, and any possible number of cellular factors. For each of these potential associations, again, further truncation/deletion analyses could aid in ascertaining whether each component may associate with μNS or μNSC through a distinct set of residues in μNS. In any case, the emerging picture of μNS is one of a protein involved in multiple interactions to form the factory matrix and to recruit other components to the factories, with a modular organization of its primary sequence for performing these many distinct activities.
How does μNS form factory-like inclusions? There are many possible interactions that could be involved in forming the three-dimensional structure that each inclusion likely represents, even before the addition of other viral components as found in the factories in infected cells. For example, μNS-μNS interactions may be all that are required for forming factory-like inclusions. Alternatively, interactions of μNS with one or more cellular factor may be necessary, with the cellular factor(s) acting as bridges between μNS monomers or oligomers. The presence of predicted coiled-coil segments in μNS has led to the suggestion that μNS forms a basal small oligomer, possibly a dimer, through -helical coiled-coil interactions (25). If so, then inclusions are likely built by linking these small oligomers together, with or without the help of cellular factors. In any case, according to the new results, μNS(471-721) must be capable of mediating or instigating all of the necessary interactions for inclusion formation. The smaller regions within μNS(471-721) that are required for forming inclusions might be directly involved in these interactions.
The consensus motif that spans residues 563 to 574 of μNS and that is partially conserved in the homologous proteins of avian orthoreoviruses and aquareoviruses includes residues His570 and Cys572, which are required for inclusion formation. Given that His and Cys residues have strong potential to chelate transition metal ions such as Zn2+, we hypothesize that metal chelation by His570 and Cys572 is a part of their role in forming inclusions. Moreover, considering that His570 and Cys572 are so closely spaced and that nearby residues with the same potential for metal chelation—Cys561, His569, and His576—are dispensable for inclusion formation, we hypothesize that His570 and Cys572 form half of an intermolecular metal-chelating motif, similar to the zinc hook of Rad50 (18) or the zinc clasp of CD4/8 and Lck (20, 21). The latter two motifs provide four zinc-chelating residues, two from each participating subunit, and contribute to homodimerization in the case of Rad50 or heterodimerization in the case of CD4/8 and Lck. Of course, further work is needed to determine whether μNS residues His570 and Cys572 indeed chelate a metal ion and, if so, whether this chelation may contribute to μNS homodimerization or μNS heterodimerization with a cellular protein, either of which could be required for inclusion formation.
Some portion of the C-terminal eight residues of μNS (Phe-Ser-Val-Pro-Thr-Asp-Glu-Leu) is also required for inclusion formation. Since this region contains no His or Cys residues, it must contribute to forming inclusions by a mechanism distinct from that hypothesized for His570 and Cys572. Although this region is not highly conserved in the homologous proteins of avian orthoreoviruses and aquareoviruses, each of the μNS homologs terminates in Leu and has at least one acidic residue and no basic residues in its C-terminal eight positions. Further work is needed to determine whether these C-terminal residues may contribute to μNS-μNS interaction or to μNS interaction with a cellular protein, either of which could be additionally required for inclusion formation.
Evidence in the present study suggests that the more N-terminal of the predicted coiled-coil segments in μNS (25) is required for inclusion formation but that the loss of function resulting from its deletion can be largely rescued by fusion to GFP. We interpret this result to indicate that the role of the first predicted coiled-coil segment is in self-association, which GFP can also mediate (9), but we acknowledge that this interpretation requires further testing. Although the more C-terminal one of the two predicted coiled-coil segments in μNS has not been directly tested for its role in inclusion formation, we expect that at least part of it is also required. Coiled-coil interactions between μNS subunits could, for example, mediate the formation of basal oligomers, which then interact through other motifs to form the inclusions. Alternatively, one or both of the coiled-coil segments could mediate hetero-oligomerization with a cellular protein. As noted in Results, predicted coiled-coil segments are also found flanking the proposed metal chelation sequences in each of the μNS homologs from avian orthoreoviruses and aquareoviruses, suggesting a conserved function for these motifs.
Formation of μNS inclusions as a tool to study protein-protein associations inside cells. The characteristic subcellular localization of μNS and its inclusion-forming derivatives—in cytoplasmic, phase-dense, globular inclusions—has proven a potent tool for studying associations between μNS and other viral proteins in infected or transfected cells (4, 6, 7). In the present study, we acquired original evidence that GFP/μNS(471-721) may provide a useful platform for examining possible associations between any two proteins in transfected cells. In particular, full-length proteins or protein regions expressed as fusions to GFP/μNS(471-721) should localize to the distinctive, globular inclusions induced by the region from positions 471 to 721. The capacity of this fused protein or protein region to associate with another, "test" protein can then be assayed by examining cells in which the test protein has been coexpressed with the inclusion-forming fusion protein. If the test protein localizes to the globular inclusions, then it and the fused protein or protein region can be concluded to associate. Of course, nonfused GFP/μNS(471-721) should be tested in parallel as a negative control for the specificity of test protein localization to the inclusions. A range of relative expression levels of the two proteins could also be tested, in case relative overexpression of the test protein may retard inclusion formation by the other. We are currently extending our studies of the feasibility of this approach as a general one for studying protein-protein associations within the "native" cellular environment in which the proteins normally reside and function.
ACKNOWLEDGMENTS
We express our sincere gratitude to Elaine Freimont and Jason Dinoso for laboratory maintenance and technical assistance and to other members of our lab for helpful discussions. We also thank John Patton and John Parker for reviews of a draft of the manuscript.
This study was supported in part by NIH grants R01 AI47904 (M.L.N.) and F32 AI56939 (C.L.M.) and by a junior faculty grant from the Giovanni Armenise-Harvard Foundation (M.L.N.). T.J.B. and M.M.A. were also supported by NIH grant T32 AI07245 to the Viral Infectivity Training Program. J.A.H. was also supported by NIH grant T32 GM07226 to the Biological and Biomedical Sciences Training Program. At earlier stages of this work, C.L.M. was also supported by NIH grant T32 AI07061 to the Combined Infectious Diseases Training Program.
T.J.B. and M.M.A. contributed equally to this study.
REFERENCES
Antczak, J. B., and W. K. Joklik. 1992. Reovirus genome segment assortment into progeny genomes studied by the use of monoclonal antibodies directed against reovirus proteins. Virology 187:760-776.
Attoui, H., Q. Fang, F. M. Jaafar, J. F. Cantaloube, P. Biagini, P. De Micco, and X. De Lamballerie. 2002. Common evolutionary origin of aquareoviruses and orthoreoviruses revealed by genome characterization of golden shiner reovirus, grass carp reovirus, striped bass reovirus and golden ide reovirus (genus Aquareovirus, family Reoviridae). J. Gen. Virol. 83:1941-1951.
Becker, M. M., M. I. Goral, P. R. Hazelton, G. S. Baer, S. E. Rodgers, E. G. Brown, K. M. Coombs, and T. S. Dermody. 2001. Reovirus NS protein is required for nucleation of viral assembly complexes and formation of viral inclusions. J. Virol. 75:1459-1475.
Becker, M. M., T. R. Peters, and T. S. Dermody. 2003. Reovirus NS and μNS proteins form cytoplasmic inclusion structures in the absence of viral infection. J. Virol. 77:5948-5963.
Broering, T., A. McCutcheon, V. Centonze, and M. Nibert. 2000. Reovirus nonstructural protein μNS binds to reovirus cores but does not inhibit their transcription activity. J. Virol. 74:5516-5524.
Broering, T. J., J. Kim, C. L. Miller, C. D. Piggott, J. B. Dinoso, M. L. Nibert, and J. S. L. Parker. 2004. Reovirus nonstructural protein μNS recruits viral core surface proteins and entering core particles to factory-like inclusions. J. Virol. 78:1882-1892.
Broering, T. J., J. S. L. Parker, P. L. Joyce, J. Kim, and M. L. Nibert. 2002. Mammalian reovirus nonstructural protein μNS forms large inclusions and colocalizes with reovirus microtubule-associated protein μ2 in transfected cells. J. Virol. 76:8285-8297.
Brookes, S. M., A. D. Hyatt, and B. T. Eaton. 1993. Characterization of virus inclusion bodies in bluetongue virus-infected cells. J. Gen. Virol. 74:525-530.
Campbell, R. E., O. Tour, A. E. Palmer, P. A. Steinbach, G. S. Baird, D. A. Zacharias, and R. Y. Tsien. 2002. A monomeric red fluorescent protein. Proc. Natl. Acad. Sci. USA 99:7877-7882.
Dales, S. 1963. Association between the spindle apparatus and reovirus. Proc. Natl. Acad. Sci. USA 50:268-275.
Dales, S., P. Gomatos, and K. C. Hsu. 1965. The uptake and development of reovirus in strain L cells followed with labeled viral ribonucleic acid and ferritin-antibody conjugates. Virology 25:193-211.
Eichwald, C., J. F. Rodriguez, and O. R. Burrone. 2004. Characterization of rotavirus NSP2/NSP5 interactions and the dynamics of viroplasm formation. J. Gen. Virol. 85:625-634.
Fields, B. N., C. S. Raine, and S. G. Baum. 1971. Temperature-sensitive mutants of reovirus type 3: defects in viral maturation as studied by immunofluorescence and electron microscopy. Virology 43:569-578.
Fujimuro, M., H. Sawada, and H. Yokosawa. 1994. Production and characterisation of monoclonal antibodies specific to multi-ubiquitin chains of polyubiquitinated proteins. FEBS Lett. 349:173-180.
Gillian, A. L., and M. L. Nibert. 1998. Amino terminus of reovirus nonstructural protein sigma NS is important for ssRNA binding and nucleoprotein complex formation. Virology 240:1-11.
Gillian, A. L., S. C. Schmechel, J. Livny, L. A. Schiff, and M. L. Nibert. 2000. Reovirus nonstructural protein NS binds in multiple copies to single-stranded RNA and shares properties with single-stranded DNA binding proteins. J. Virol. 74:5939-5948.
Gomatos, P. J., O. Prakash, and N. M. Stamatos. 1981. Small reovirus particle composed solely of NS with specificity for binding different nucleic acids. J. Virol. 39:115-124.
Hopfner, K. P., L. Craig, G. Moncalian, R. A. Zinkel, T. Usui, B. A. Owen, A. Karcher, B. Henderson, J. L. Bodmer, C. T. McMurray, J. P. Carney, J. H. Petrini, and J. A. Tainer. 2002. The Rad50 zinc-hook is a structure joining Mre11 complexes in DNA recombination and repair. Nature 418:562-566.
Huismans, H., and W. K. Joklik. 1976. Reovirus-coded polypeptides in infected cells: isolation of two native monomeric polypeptides with affinity for single-stranded and double-stranded RNA, respectively. Virology 70:411-424.
Huse, M., M. J. Eck, and S. C. Harrison. 1998. A Zn2+ ion links the cytoplasmic tail of CD4 and the N-terminal region of Lck. J. Biol. Chem. 273:18729-18733.
Kim, P. W., Z. Y. Sun, S. C. Blacklow, G. Wagner, and M. J. Eck. 2003. A zinc clasp structure tethers Lck to T-cell coreceptors CD4 and CD8. Science 301:1725-1728.
Lee, P. W. K., E. C. Hayes, and W. K. Joklik. 1981. Characterization of anti-reovirus immunoglobulins secreted by cloned hybridoma cell lines. Virology 108:134-146.
Mayor, H. D. 1965. Studies on reovirus. 3. A labile, single-stranded ribonucleic acid associated with the late stages of infection. J. NCI. 35:919-925.
Mbisa, J. L., M. M. Becker, S. Zou, T. S. Dermody, and E. G. Brown. 2000. Reovirus μ2 protein determines strain-specific differences in the rate of viral inclusion formation in L929 cells. Virology 272:16-26.
McCutcheon, A. M., T. J. Broering, and M. L. Nibert. 1999. Mammalian reovirus M3 gene sequences and conservation of coiled-coil motifs near the carboxyl terminus of the μNS protein. Virology 264:16-24.
Miller, C. L., T. J. Broering, J. S. L. Parker, M. M. Arnold, and M. L. Nibert. 2003. Reovirus NS protein localizes to inclusions through an association requiring the μNS amino terminus. J. Virol. 77:4566-4576.
Miller, D. J., M. D. Schwartz, and P. Ahlquist. 2001. Flock house virus RNA replicates on outer mitochondrial membranes in Drosophila cells. J. Virol. 75:11664-11676.
Morgan, E. M., and H. J. Zweerink. 1975. Characterization of transcriptase and replicase particles isolated from reovirus-infected cells. Virology 68:455-466.
Moss, B. 2001. Poxviridae: the viruses and their replication, p. 2637-2671. In D. M. Knipe and P. M. Howley (ed.), Fields virology. Lippincott Williams & Wilkins Co., Philadelphia, Pa.
Parker, J. S., T. J. Broering, J. Kim, D. E. Higgins, and M. L. Nibert. 2002. Reovirus core protein μ2 determines the filamentous morphology of viral inclusion bodies by interacting with and stabilizing microtubules. J. Virol. 76:4483-4496.
Quimby, B. B., and A. H. Corbett. 2001. Nuclear transport mechanisms. Cell. Mol. Life Sci. 58:1766-1773.
Restrepo-Hartwig, M. A., and P. Ahlquist. 1996. Brome mosaic virus helicase- and polymerase-like proteins colocalize on the endoplasmic reticulum at sites of viral RNA synthesis. J. Virol. 70:8908-8916.
Rhim, J. S., L. E. Jordan, and H. D. Mayor. 1962. Cytochemical, fluorescent-antibody and electron microscopic studies on the growth of reovirus (ECHO 10) in tissue culture. Virology 17:342-355.
Richardson, M. A., and Y. Furuichi. 1985. Synthesis in Escherichia coli of the reovirus nonstructural protein NS. J. Virol. 56:527-533.
Roizman, B., and A. E. Sears. 2001. Herpes simplex viruses and their replication, p. 2231-2295. In D. M. Knipe and P. M. Howley (ed.), Fields virology. Lippincott Williams & Wilkins Co., Philadelphia, Pa.
Schwartz, M., J. Chen, M. Janda, M. Sullivan, J. den Boon, and P. Ahlquist. 2002. A positive-strand RNA virus replication complex parallels form and function of retrovirus capsids. Mol. Cell 9:505-514.
Sharpe, A. H., L. B. Chen, and B. N. Fields. 1982. The interaction of mammalian reoviruses with the cytoskeleton of monkey kidney CV-1 cells. Virology 120:399-411.
Silverstein, S. C., and S. Dales. 1968. The penetration of reovirus RNA and initiation of its genetic function in L-strain fibroblasts. J. Cell Biol. 36:197-230.
Silverstein, S. C., and P. H. Schur. 1970. Immunofluorescent localization of double-stranded RNA in reovirus-infected cells. Virology 41:564-566.
Silvestri, L. S., Z. F. Taraporewala, and J. T. Patton. 2004. Rotavirus replication: plus-sense templates for double-stranded RNA synthesis are made in viroplasms. J. Virol. 78:7763-7774.
Spendlove, R. S., E. H. Lennette, J. N. Chin, and C. O. Knight. 1964. Effect of antimitotic agents on intracellular reovirus antigen. Cancer Res. 24:1826-1833.
Stamatos, N. M., and P. J. Gomatos. 1982. Binding to selected regions of reovirus mRNAs by a nonstructural reovirus protein. Proc. Natl. Acad. Sci. USA 79:3457-3461.
Theron, J., H. Huismans, and L. H. Nel. 1996. Site-specific mutations in the NS2 protein of epizootic haemorrhagic disease virus markedly affect the formation of cytoplasmic inclusion bodies. Arch. Virol. 141:1143-1151.
Touris-Otero, F., M. Cortez-San Martin, J. Martinez-Costas, and J. Benavente. 2004. Avian reovirus morphogenesis occurs within viral factories and begins with the selective recruitment of NS and A to μNS inclusions. J. Mol. Biol. 341:361-374.
Touris-Otero, F., J. Martinez-Costas, V. N. Vakharia, and J. Benavente. 2004. Avian reovirus nonstructural protein μNS forms viroplasm-like inclusions and recruits protein NS to these structures. Virology 319:94-106.
Wiener, J. R., J. A. Bartlett, and W. K. Joklik. 1989. The sequences of reovirus serotype 3 genome segments M1 and M3 encoding the minor protein μ2 and the major nonstructural protein μNS, respectively. Virology 169:293-304.
Wilson, I. A., H. L. Niman, R. A. Houghten, A. R. Cherenson, M. L. Connolly, and R. A. Lerner. 1984. The structure of an antigenic determinant in a protein. Cell 37:767-778.
Wolf, E., P. S. Kim, and B. Berger. 1997. MultiCoil: a program for predicting two- and three-stranded coiled coils. Protein Sci. 6:1179-1189.
Zweerink, H. J., M. J. McDowell, and W. K. Joklik. 1971. Essential and nonessential noncapsid reovirus proteins. Virology 45:716-723.(Teresa J. Broering , Mich)
Department of Biochemistry, University of Wisconsin-Madison, Madison, Wisconsin 53706
Ph.D. Programs in Virology
Biological and Biomedical Sciences, Division of Medical Sciences, Harvard University, Cambridge, Massachusetts 02138
ABSTRACT
Mammalian orthoreoviruses are believed to replicate in distinctive, cytoplasmic inclusion bodies, commonly called viral factories or viroplasms. The viral nonstructural protein μNS has been implicated in forming the matrix of these structures, as well as in recruiting other components to them for putative roles in genome replication and particle assembly. In this study, we sought to identify the regions of μNS that are involved in forming factory-like inclusions in transfected cells in the absence of infection or other viral proteins. Sequences in the carboxyl-terminal one-third of the 721-residue μNS protein were linked to this activity. Deletion of as few as eight residues from the carboxyl terminus of μNS resulted in loss of inclusion formation, suggesting that some portion of these residues is required for the phenotype. A region spanning residues 471 to 721 of μNS was the smallest one shown to be sufficient for forming factory-like inclusions. The region from positions 471 to 721 (471-721 region) includes both of two previously predicted coiled-coil segments in μNS, suggesting that one or both of these segments may also be required for inclusion formation. Deletion of the more amino-terminal one of the two predicted coiled-coil segments from the 471-721 region resulted in loss of the phenotype, although replacement of this segment with Aequorea victoria green fluorescent protein, which is known to weakly dimerize, largely restored inclusion formation. Sequences between the two predicted coiled-coil segments were also required for forming factory-like inclusions, and mutation of either one His residue (His570) or one Cys residue (Cys572) within these sequences disrupted the phenotype. The His and Cys residues are part of a small consensus motif that is conserved across μNS homologs from avian orthoreoviruses and aquareoviruses, suggesting this motif may have a common function in these related viruses. The inclusion-forming 471-721 region of μNS was shown to provide a useful platform for the presentation of peptides for studies of protein-protein association through colocalization to factory-like inclusions in transfected cells.
INTRODUCTION
Viruses with ten-segmented, double-stranded RNA genomes from the family Reoviridae, genus Orthoreovirus, are believed to replicate in distinctive, cytoplasmic inclusion bodies (7, 10, 11, 13, 23, 24, 30, 33, 37-39, 41, 44, 45). These inclusions are commonly called viral factories (13, 30) or viroplasms (45) and are similar to cytoplasmic inclusions formed by other viruses in the same family. In cells infected by rotaviruses or orbiviruses, for example, these structures are called viroplasms (12, 40) or viral inclusion bodies (8, 43), respectively.
Many viruses sequester their replication machinery within localized structures or surfaces in infected cells: for example, herpes simplex virus (double-stranded DNA genome, family Herpesviridae) in nuclear inclusions (reviewed in reference 35), vaccinia virus (double-stranded DNA genome, family Poxviridae) in cytoplasmic inclusions (also called viral factories [reviewed in reference 29]), brome mosaic virus (single-stranded RNA [ssRNA] genome, family Bromoviridae) on the cytoplasmic face of the endoplasmic reticulum (32, 36), and flock house virus (ssRNA genome, family Nodaviridae) on the cytoplasmic face of mitochondria (27). The basis and consequences of such specific localizations are the subjects of active investigations in many laboratories. We anticipate that studies of mammalian orthoreovirus (reovirus) factories in our own laboratory will provide new insights into the still poorly characterized mechanisms for RNA packaging and replication by these and other viruses in the family Reoviridae, as well as on the general significance and mechanism of concentrating viral replication at particular sites within cells.
In early studies, reovirus factories were determined to contain fully and partially assembled viral particles, viral proteins, double-stranded RNA, microtubules, and "kinky" filaments proposed to be intermediate filaments but not membrane-bound structures or ribosomes (10, 11, 23, 33, 37, 39, 41). The factories have a peculiarly dense consistency that distinguishes them from the adjacent cytoplasm and causes them to appear highly refractile by phase-contrast microscopy. At least part of this property appears to reflect a protein matrix that suffuses the factories. The determinants or features of one or more viral proteins that would make it capable of forming such a matrix are not well understood but might involve a variety of different types of intersubunit interactions, as well as interactions with cellular factors.
Results from our laboratory and others have recently shown that a single reovirus protein, μNS, is sufficient for forming phase-dense globular inclusions in the cytoplasm of transfected cells (4, 7). The inclusions formed by μNS in such experiments are notably similar in appearance to globular reovirus factories formed in infected cells, as visualized by either phase-contrast or immunofluorescence (IF) microscopy (4, 7, 30), suggesting that μNS forms the matrix of the factories (7). Moreover, μNS can associate with several other reovirus proteins and recruit them to these inclusions in transfected cells. To date, the viral proteins that have been reported to be recruited to inclusions formed by reovirus μNS are microtubule-binding core protein μ2 (7); nonstructural and ssRNA-binding protein NS (4, 26); and the core surface proteins 1, 2, and 2 (6). Whole core particles released into the cytoplasm during cell entry are also recruited to μNS inclusions (6). Indirect evidence suggests that μNS may possess ssRNA-binding activity as well, which may be important for recruiting the viral plus-strand RNAs to the factories and/or retaining them there for replication and packaging in infected cells (1). However, the ssRNA-binding activity of NS is more firmly established (15-17, 19, 34, 42) and may play the more important role in ssRNA recruitment to or retention within the factories (4, 6, 26). Recent studies on the μNS homolog from avian orthoreoviruses have reached conclusions similar to these regarding the roles of μNS in forming the factory matrix and recruiting other viral proteins to the factories (44, 45). We explicitly specify in the present study when we intend a statement to encompass results for avian orthoreovirus; otherwise, we use the abbreviated name reovirus to represent mammalian orthoreovirus only.
The 80-kDa μNS protein is not a component of mature virions but is expressed to high levels in infected cells and is concentrated in the reovirus factories (7, 49). Other previous findings about μNS include its association with transcriptionally active particles isolated from infected cells (28), its binding to the surfaces of purified core particles in vitro (5), and its capacity to recruit entering core particles to inclusions (6). The μNS protein has two predicted coiled-coil segments in the carboxyl-terminal one-third of its sequence (25), but the oligomeric status of μNS has not been demonstrated. Little else is known about the structure of μNS, although domains have been predicted from the pattern of conserved and variable regions of sequence among the μNS alleles from different reovirus isolates (25). An isoform of μNS lacking 5 kDa from its amino terminus and called μNSC is also present in virus-infected cells (22, 46). The origin of μNSC is not certain, but it is proposed to result from either secondary initiation or cleavage near Met41 in μNS. A recombinant protein engineered to lack the first 40 residues of μNS, μNS(41-721), which should be similar to μNSC, also forms phase-dense globular inclusions in transfected cells (7).
Involvement of microtubules in forming reovirus factories has also been demonstrated. The M1 segment, which encodes the structurally minor core protein μ2, has been shown to determine both the timing of factory formation (24) and the morphology of factories (30). The filamentous morphology of factories formed by most reovirus strains has been attributed to microtubule association of the μ2 protein (30). When μNS and μ2 are coexpressed in cells without other viral proteins, μNS is redistributed with μ2 to filamentous inclusions (7), suggesting that these two proteins together determine the formation and morphology of filamentous viral factories. Aside from any effects of μ2, however, μNS also requires microtubules for undergoing condensation of its smaller inclusions into larger ones, as shown by the effects of the microtubule-depolymerizing drug nocodazole. This probably reflects a role for microtubule-based transport in the formation of μNS inclusions and the determination of viral factory morphology (7, 30).
By forming the matrix of viral factories, as well as recruiting other components to these structures, the reovirus μNS protein could serve to concentrate the components for viral replication and assembly, to arrange the components in specific ways to promote replication and assembly, and/or to sequester the components and their mediated functions from antiviral factors in the cellular environment. To better understand the role of μNS in forming the matrix of viral factories, we engineered constructs to express a panel of μNS truncations and point mutants and thereby determined the regions of this protein that are involved in forming factory-like inclusions in transfected cells. The results identified several different sequences in the C-terminal one-third of μNS that are necessary and sufficient for this activity of μNS.
MATERIALS AND METHODS
Cells and antibodies. CV-1 cells were maintained in Dulbecco modified Eagle medium (Invitrogen Life Technologies) containing 10% fetal bovine serum (HyClone Laboratories) and 10 μg of gentamicin solution (Invitrogen Life Technologies) per ml. Goat anti-mouse immunoglobulin G (IgG) and goat anti-rabbit IgG conjugated to Alexa 488 or Alexa 594 were obtained from Molecular Probes. Rabbit polyclonal antisera to μNS or μ2 have been described previously (5, 7, 30). For some experiments, we used protein A-purified rabbit anti-μNS or anti-μ2 IgG conjugated to Texas Red, Oregon Green, Alexa 488, or Alexa 594 by using kits obtained from Molecular Probes. Mouse monoclonal antibody (MAb) FK2 against conjugated ubiquitin (14) was purchased from Medical & Biological Laboratories. Mouse MAb JL8 against Aequorea victoria green fluorescent protein (GFP) was purchased from BD Biosciences. Mouse MAb HA.11 against an immunodominant epitope of influenza A virus hemagglutinin (HA) was purchased from Covance. Mouse MAb 3E10 specific for reovirus NS protein (3) was a gift from T. S. Dermody and coworkers (Vanderbilt University). All antibodies were titrated to optimize signal-to-noise ratios.
Expression constructs. Reovirus proteins were expressed from genes cloned into the mammalian expression vector pCI-neo (Promega). pCI-M3(T1L) and pCI-M3(T3D) to express μNS from type 1 Lang (T1L) or type 3 Dearing (T3D) reovirus have been described previously (7), as have pCI-M3(41-721) to express μNS residues 41 to 721 (7), pCI-M1(T1L) to express μ2 (30), and pCI-S3(T1L) to express NS (26). Vent polymerase, which was used for all PCRs, and other enzymes were from New England Biolabs unless otherwise stated.
To express μNS residues 1 to 683, site-directed mutagenesis was used to introduce a stop codon at nucleotides 2068 to 2070, followed by a Bsu36I site. QuikChange site-directed mutagenesis (Stratagene) was used according to the manufacturer's protocol with pFastBac-M3(T1L) (5) as a template, forward primer 5'-GGATACGATGAACTAACCTCAGGCTAAATCATTGCG-3', and reverse primer 5'-CGCAATGATTTAGCCTGAGGTTAGTTCATCGTATCC-3' (double underline, nucleotide change to add the stop codon; single underline, nucleotide change to add the restriction site). The region containing the desired mutations was excised by digestion with HindIII and then ligated to pFastBac-M3(T1L) that had been cut with HindIII to remove the same region, generating pFastBac-M3(1-683). The subcloned region was sequenced to confirm its correctness. The M3 gene was removed from pFastBac-M3(1-683) by digestion with SalI and NheI and then ligated to pCI-neo that had been cut with the same enzymes, generating pCI-M3(1-683).
To generate other μNS truncations, start and stop codons and restriction sites were introduced at different positions in the M3 gene by PCR amplification of the desired M3 region. The truncations were made in either T1L or T3D M3. We have found no difference in inclusion formation with T1L and T3D μNS (7, 26), suggesting that the truncations from these allelic μNS proteins are directly comparable. In addition, the T1L and T3D μNS proteins are 96% identical (25). Each PCR was performed with pGEM-4Z-M3(T1L) or pGEM-4Z-M3 (T3D) (5) as a template, and the primers are listed in Table 1. Each PCR product was cut with the restriction enzymes listed in Table 1 and then ligated to a plasmid that had been cut with these same enzymes, the plasmid being either pGEM-4Z (for pGEM-4Z-M3(1-173) and pGEM-4Z-M3(1-221) or the template plasmid (for all other constructs). The correctness of each construct was confirmed by sequencing. The truncated M3 genes were subcloned into pCI-neo by using the restriction enzymes listed in Table 2. Each construct was named for the residues of μNS that the expressed protein should ultimately contain (Tables 1 and 2).
The pEGFP-N1 and pEGFP-C1 vectors (BD Biosciences) were respectively used to express fusions of enhanced A. victoria GFP to the C or N terminus of the desired μNS region. pEGFP-N1-M3(T1L) was previously constructed to express GFP fused to the C terminus of μNS (μNS/GFP) (7). To express GFP fused to the N terminus of μNS residues 471 to 721, pGEM-4Z-M3(471-721) was cut with SalI and KpnI, and the excised fragment was then ligated to pEGFP-C1 that had been cut with the same enzymes, generating pEGFP-C1-M3(471-721). To express GFP fused to the N termini of even smaller C-terminal regions of μNS, an EcoRI site was introduced at the desired position in the M3 gene during PCR amplification. Each PCR was performed with pGEM-4Z-M3(T1L) (5) as a template, the forward primers listed in Table 3, and a reverse primer complementary to the 3' end of the M3 coding strand with an added BamHI site (underlined) (5'-GCAGGGGATCCGATGAATGGGGGTCGGGAAGGCTTAAGGG-3'). Each PCR product was cut with EcoRI and BamHI and was then ligated to pEGFP-C1 that had been cut with the same enzymes. The correctness of each construct was confirmed by sequencing, and the construct was named for the residues of μNS that the expressed protein contains (Table 3).
To generate HA-tagged fusions, epitope-encoding forward primer 5'-CGTAGCTAGCGTCATGGCTTACCCATACGACGTCCCAGACTACGCTCTCGAGATGC-3' and reverse primer 5'-GCATCTCGAGAGCGTAGTCTGGGACGTCGTATGGGTAAGCCATGACGCTAGCTACG-3' (underlining indicates NheI and XhoI sites added to each primer) were annealed by boiling in 1x SSC (150 mM NaCl plus 15 mM sodium citrate [pH 7.0]), followed by cooling at room temperature for 1 h. The duplex oligonucleotide and vector pCI-neo were then digested with NheI and XhoI, and the products were ligated to generate pCI-neo-HA. The region encoding μNS(561-721) was amplified by PCR from the template pCI-M3(T1L) by using forward primer 5'-GCTAGAATTCATGTAGTCTGGATATGTATTTGAGACACCAC-3' and reverse primer 5'-GATCGATCCCGGGTCGGGAAGGCTTAAGGGATTAGGGCAA-3' (underlining indicates an EcoRI or SmaI site added to each primer). The PCR product and vector pCI-neo-HA were then sequentially digested with SmaI and EcoRI, and the products were ligated to generate pCI-neo-HA-M3(561-721). The region encoding μNS(471-721) was amplified by PCR from the template pCI-M3(T1L) by using the forward primer 5'-AGCTCTCGAGGTCATGTCCAGTGACATGGTAGACGGGATTAAAC-3' (underlining indicates an added XhoI site) and reverse primer 5'-CGAAGCATTAACCCTCAC-3'. The PCR product and vector pCI-neo-HA were then digested with XhoI and NotI, and the products were ligated to generate pCI-neo-HA-M3(471-721). To generate untagged μNS(561-721), PCR was performed by using the template pCI-M3(T1L), forward primer5'-AGCTGAATTCGTCATGGCTTGTAGTCTGGTATGTATTTGAGACAC-3' (underline indicates an added EcoRI site), and reverse primer 5'-CGAAGCATTAACCCTCAC-3'. The PCR product and vector pCI-neo were then digested with EcoRI and NotI, and the products were ligated to generate pCI-neo-M3(561-721). The correctness of the final constructs was confirmed by sequencing.
To generate a μNS(471-721) or full-length μNS protein with residues His569 and His570 changed to glutamine and Cys572 changed to serine, QuikChange site-directed mutagenesis was used according to the manufacturer's protocol with pCI-M3(471-721) or pCI-M3(T3D) as a template, forward primer 5'-GGATATGTATCTGCGACAACAAACTTCCATTAATGGTCATGC-3', and reverse primer 5'-GCATGACCATTAATGGAAGTTTGTTGTCGCAGATACATATCC-3' (double underlines, nucleotide changes to provide amino acid changes; single underline, nucleotide change to remove a DdeI site for screening purposes). After the QuikChange protocol, the mutated pCI-M3(471-721) or pCI-M3(T3D) plasmid was prepared for subcloning by cutting with NheI and EcoRI or by cutting with BlpI and NotI, respectively. The excised fragments were then ligated to pCI-neo or pCI-M3(T3D), respectively, that had been cut with the same respective pair of enzymes, generating the final versions of pCI-M3(471-721)QQS and pCI-M3(T3D)QQS. The correctness of these constructs was confirmed by sequencing. To generate a full-length μNS protein with Cys561 or Cys572 changed to Ser or His569, His570, or His576 changed to Gln, QuikChange site-directed mutagenesis was used according to the manufacturer's protocol with pCI-M3(T3D) as a template and the primers listed in Table 4. According to the QuikChange protocol, each mutated plasmid was prepared for subcloning by cutting with BlpI and NotI. In each case, the excised fragment was then ligated to pCI-M3(T3D) that had been cut with the same enzymes, generating the constructs as named in Table 4. The correctness of each subcloned region was confirmed by sequencing.
To generate a fusion of μNS residues 1 to 41 to the N terminus of GFP and μNS residues 471 to 721 to the C terminus of GFP, DNA encoding residues 1 to 41 of μNS and a large portion of GFP was removed from pEGFP-N1-M3(1-41) (7) by digestion with NheI and BsrGI. pEGFP-C1-M3(471-721) was also cut with these enzymes, and the two DNA fragments were ligated to generate pEGFP-M3(1-41)/GFP/M3(471-721). To generate a fusion of μNS residues 1 to 12 to the N terminus of GFP and μNS residues 471 to 721 to the C terminus of GFP, forward primer 5'-AATTC ATGGCTTCATTCAAGGGATTCTCCGTCAACACTGTTGCG-3' and reverse primer 5'-GATCCGCAACAGTGTTGACGGAGAATCCCTTGAATGAAGCCATG-3' were annealed to generate a small duplex encoding μNS residues 1 to 12 and having BamHI and EcoRI overhangs at the respective ends (underlined). This small duplex and plasmid pEGFP-N1 were both digested with BamHI and EcoRI and then ligated to yield pEGFP-N1-M3(1-12). This plasmid was then digested with NheI and BsrGI, and the excised fragment was ligated to pEGFP-C1-M3(471-721) that had been digested with the same enzymes, generating plasmid pEGFP-M3(1-12)/GFP/M3(471-721). The correctness of the resulting construct was confirmed by sequencing.
Transfections and IF microscopy. Cells were seeded the day before transfection at a density of 1.5 x 104 per cm2 in six-well plates (9.6 cm2 per well) containing glass coverslips (19 mm). Cells were transfected with 2 μg of DNA and 7 μl of Lipofectamine 2000 (Invitrogen Life Technologies) or 1.5 μl of DNA and 10 μl of Polyfect (Qiagen) according to the manufacturer's directions. Cells were further incubated for 18 to 24 h at 37°C before fixation for 10 min at room temperature in 2% paraformaldehyde in phosphate-buffered saline (PBS; 137 mM NaCl, 3 mM KCl, 8 mM Na2HPO4, 1 mM KH2PO4 [pH 7.5]) or 3 min at –20°C in ice-cold methanol. Fixed cells were washed three times with PBS, permeabilized, and blocked in PBS containing 1% bovine serum albumin and 0.1% Triton X-100 (PBSAT). Primary antibodies were diluted in PBSAT and incubated with cells for 25 to 40 min at room temperature. After three washes in PBS, secondary antibodies diluted in PBSAT were added and incubated with cells for 25 min at room temperature. Coverslips were incubated with 300 nM 4,6-diamidino-2-phenylindole (DAPI; Molecular Probes) in PBS for 5 min to counterstain cell nuclei, briefly washed in PBS, and mounted on glass slides with Prolong (Molecular Probes). Samples were examined by using a TE-300 inverted microscope (Nikon) equipped with phase and fluorescence optics, and images were collected digitally as described elsewhere (30). All images were processed and prepared for presentation by using Photoshop (Adobe Systems).
Immunoblot analysis. CV-1 cells were transfected as described for IF, and whole-cell lysates were collected 18 to 24 h posttransfection (p.t.). CV-1 cells (1.2 x 106) were washed briefly in PBS and then scraped into 1 ml of PBS and pelleted. The pelleted cells were resuspended in 30 μl of PBS containing protease inhibitors (Roche Biomedicals), lysed into sample buffer, boiled for 10 min, and subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were electroblotted from the gels to nitrocellulose in 25 mM Tris and 192 mM glycine (pH 8.3). Binding of antibodies was detected with alkaline phosphatase-coupled goat anti-mouse IgG (Bio-Rad Laboratories) and the colorimetric reagents p-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolylphosphate p-toluidine salt (Bio-Rad Laboratories).
Sequence comparisons. The following μNS and μNS-homolog sequences, each designated as complete coding sequences in GenBank, were compared: μNS of mammalian orthoreoviruses T1L (accession no. AAF13169), type 2 Jones (accession no. AAF13170), and T3D (accession no. AAF13171); μNS of avian orthoreoviruses 1733 (accession no. AAQ81873), S1133 (accession no. AAS78987), 1017-1 (accession no. AAS78988), 2408 (accession no. AAS78990), 601G (accession no. AAS78991), 750505 (accession no. AAS78992), 916SI (accession no. AAS78993), 918 (accession no. AAS78994), 919 (accession no. AAS78995), OS161 (accession no. AAS78996), R2 (accession no. AAS78997), and T6 (accession no. AAS78998); and NS1 of aquareoviruses grass carp 873 (accession no. AAM92735) and golden shiner (accession no. AAM92747). Sequences were compared by using the Bestfit and Pretty programs from the Genetics Computer Group Wisconsin package. The Multicoil program (http://multicoil.lcs.mit.edu/cgi-bin/multicoil) (48) was used for predicting coiled-coil regions.
RESULTS
The C terminus of μNS is required for forming factory-like inclusions. We have previously shown that a truncated μNS protein lacking the N-terminal 40 residues, μNS(41-721), collects in cytoplasmic, factory-like inclusions in transfected cells, similarly to full-length μNS (7). The N terminus of μNS is thus not required for forming these structures. To determine whether the C terminus of μNS is required, we engineered M3 gene constructs to express a series of truncated μNS proteins that lacked increasing numbers of residues from the C terminus (Table 2). Each of these plasmids was transfected into CV-1 cells, and the cell lysates were subjected to SDS-PAGE, followed by immunoblotting with polyclonal anti-μNS serum (5). The results verified the expression of an appropriately sized, μNS-derived protein from each construct (data not shown).
To determine the intracellular distribution of the C-terminally truncated μNS proteins, CV-1 cells were separately transfected with each of the plasmids and later fixed and immunostained with polyclonal anti-μNS antibodies (Fig. 1A, left and right columns). In addition, the cells were coimmunostained with MAb FK2, which recognizes conjugated ubiquitin (14), to determine whether any of these proteins were substantially misfolded and targeted for degradation by the ubiquitin-proteasome system (Fig. 1A, center and right columns). Each of the C-terminally truncated proteins was diffusely distributed in the cytoplasm and nucleus, in clear contrast to full-length μNS, which collected in globular inclusions as expected (Fig. 1A and data not shown). We conclude that all of these truncated proteins are negative for forming factory-like inclusions (summarized in Fig. 2) and thus that some portion of the smallest region we deleted from the C terminus of μNS, residues 714 to 721, is required for inclusion formation. None of the truncated proteins appeared to be aggregated or strongly colocalized with conjugated ubiquitin (Fig. 1A and data not shown), suggesting that they were not substantially misfolded.
All of the C-terminally truncated μNS proteins appeared to be partially localized to the nucleus (Fig. 1A, left column), even though μNS(1-683), μNS(1-700), and μNS(1-713) were above the 60-kDa limit for passive diffusion through nuclear pores (reviewed in reference 31) (Table 2). In addition, at least some of the truncated proteins appeared to be excluded from nucleoli (Fig. 1A, left column). The relevance of this nuclear pattern is unclear because full-length μNS is not detected in the nuclei of infected or M3-transfected cells (7, 30). When coexpressed with μ2(T1L), each of the C-terminally truncated proteins strongly colocalized with μ2 on microtubules (Fig. 1B and data not shown; summarized in Fig. 2). This was expected because each protein included μNS residues 1 to 41, which are known to be sufficient for association with μ2 (7, 26).
The N-terminal 470 residues of μNS are not required for forming factory-like inclusions but affect inclusion shape. To identify the minimal region of μNS required for inclusion formation, we engineered M3 gene constructs to express a series of truncated μNS proteins that lacked increasing numbers of residues from the N terminus (Table 2). Each of these plasmids was transfected into CV-1 cells, and the cell lysates were subjected to SDS-PAGE, followed by immunoblotting with the anti-μNS serum (5). The results verified expression of an appropriately sized, μNS-derived protein from each construct (data not shown).
The intracellular distribution of each of the N-terminally truncated μNS proteins was determined as described above for the C-terminal truncations, including coimmunostaining with MAb FK2. With the exception of μNS(363-721), which showed a distinctly aggregated pattern and strongly colocalized with conjugated ubiquitin (Fig. 3), each of the N-terminally truncated proteins collected in globular inclusions (Fig. 3, left and right columns, and data not shown). However, the proteins missing more than the 40 N-terminal residues of μNS often formed inclusions with more elongated shapes and enclosed fenestrations than those formed by full-length μNS or μNS(41-721) (Fig. 3, left and right columns). Although these inclusions had altered morphologies, they did not colocalize with conjugated ubiquitin (Fig. 3, center and right columns), suggesting that these truncated proteins were not substantially misfolded. The behavior of μNS(363-721), on the other hand, suggested that it was in fact largely misfolded.
When coexpressed with μ2(T1L), none of the N-terminally truncated μNS proteins colocalized with μ2 on microtubules or in inclusions (data not shown; summarized in Fig. 2). This was expected because each of these proteins lacks μNS residues 14 to 40, which are known to be required for association with μ2 (7, 26). From the results summarized in Fig. 2, we conclude that μNS residues 471 to 721 are a sufficient part of μNS for forming factory-like inclusions. These residues encompass the two predicted coiled-coil segments of μNS, the intervening "linker" between these segment, and the C-terminal "tail" that follows the second predicted coiled-coil segment (25). Given the altered morphologies of the inclusions formed by μNS proteins lacking more than the 40 N-terminal residues, we also conclude that some portion of residues 41 to 172 plays a distinguishable role in modulating inclusion morphology.
Residues 561 to 721 of μNS are sufficient for allowing a GFP-tagged protein, but not an HA-tagged or untagged protein, to form factory-like inclusions. To determine whether a region of μNS smaller than residues 471 to 721 may be sufficient for inclusion formation, and also to allow more ready detection of these smaller proteins, we next generated constructs to express GFP fused to the N terminus of selected μNS truncations (Table 3). Each of these plasmids was transfected into CV-1 cells, and the cell lysates were subjected to SDS-PAGE, followed by immunoblotting with GFP-specific MAb JL-8. The results verified production of an appropriately sized, GFP- and μNS-derived fusion protein from each construct (Fig. 4A).
To determine the intracellular distribution of each fusion protein, CV-1 cells were transfected and later immunostained with the GFP-specific MAb. Nonfused GFP was diffusely distributed in the cytoplasm and nucleus, and GFP fused to the C terminus of full-length μNS (μNS/GFP) collected in globular inclusions as previously shown (7) (also see Fig. 4B). GFP fused to the N terminus of μNS(471-721) [GFP/μNS(471-721)] also collected in inclusions (Fig. 4B), which appeared similar to those formed by μNS(471-721) (Fig. 3) or μNS/GFP (Fig. 4B). GFP fused to the N terminus of μNS(561-721) collected in inclusions as well, although a small fraction of the transfected cells (13%) displayed a diffuse distribution of this protein (Fig. 4B and data not shown). In contrast, GFP fused to the N terminus of μNS(614-721), μNS(625-721), or μNS(695-721) was diffusely distributed in the cytoplasm and nucleus of most or all cells expressing them (Fig. 4B and data not shown). From these results, we conclude that residues 561 to 721 are a sufficient part of μNS in GFP fusions for forming factory-like inclusions (summarized in Fig. 2), albeit at a lower efficiency than full-length μNS or μNS(471-721). The first predicted coiled-coil segment (25) is thus dispensable for inclusion formation by a μNS-GFP fusion. On the other hand, some portion of residues 561 to 613, in the linker between the two predicted coiled-coil segments, is required.
Given that A. victoria GFP is known to weakly dimerize (9) and also that GFP/μNS(561-721) was less efficient at forming factory-like inclusions than was GFP/μNS(471-721), we hypothesized that the GFP tag may partially complement an otherwise-required contribution of the first predicted coiled-coil segment of μNS for forming inclusions. We therefore tested another version of μNS(561-721), this one fused to an epitope of influenza virus HA (47) at its N terminus [HA/μNS(561-721)]. After expression in transfected CV-1 cells and immunostaining with tag-specific MAb HA.11, HA/μNS(561-721) was diffusely distributed in the cytoplasm and nucleus and thus not concentrated in inclusions (Fig. 5). Fusion of the HA tag to the N terminus of μNS(471-721) [HA/μNS(471-721)], in contrast, caused little or no reduction in its capacity to form factory-like inclusions (Fig. 5). We also generated and tested a third version of μNS(561-721), this one lacking any tag. After expression in transfected CV-1 cells and immunostaining with anti-μNS antibodies, untagged μNS(561-721) was diffusely distributed in the cytoplasm and nucleus and thus not concentrated in inclusions (Fig. 5). Untagged μNS(471-721) tested in parallel formed factory-like inclusions (Fig. 5) as seen in the preceding experiments (see Fig. 3). Based on these results, we conclude that the more N-terminal predicted coiled-coil segment of μNS is required for inclusion formation in the absence of a fusion tag such as GFP that can independently self-associate.
Putative metal-chelating residues His570 and Cys572 are required for factory-like inclusion formation. In an effort to identify specific residues in μNS that may be required for inclusion formation, we compared the deduced protein sequences of μNS homologs from mammalian and avian orthoreoviruses, as well as from aquareoviruses (see Materials and Methods for the GenBank accession numbers) (2, 25, 45). In all, the μNS homologs derived from 17 isolates in the three groups: 3 mammalian isolates from the Orthoreovirus genus, 12 avian isolates from the Orthoreovirus genus, and 2 piscine isolates from the Aquareovirus genus. The overall identity scores for any two μNS homologs from separate groups are small: <30% in each case (2, 45) (the present study and data not shown). Despite this degree of divergence, the C-terminal one-third of each of these μNS homologs contains two predicted coiled-coil segments, of similar lengths and spacing, separated by a linker and followed by a C-terminal tail (data not shown), as previously described for the mammalian and avian isolates (25, 45).
Interestingly, in the linker between the predicted coiled-coil segments, we identified a small consensus motif common to all of the examined μNS homologs (Fig. 6A). This sequence,Ile/Leu-x-x-Tyr-Leu-x-x-His-Thr/Val-Cys-Ile/Val-Asn (where"x" represents nonconserved positions), includes two residues (underlined) with strong potential to chelate transition metal ions such as Zn2+. The two residues correspond to His570 and Cys572 in the mammalian orthoreovirus μNS proteins. Each of the μNS homologs contains other His and/or Cys residues flanking the consensus motif, but the position and spacing of these residues is not conserved among the examined sequences (Fig. 6A). In mammalian orthoreovirus μNS, the other conserved His and Cys residues in this region are Cys561, His569, and His576. The consensus motif spans residues 563 to 574 in the mammalian orthoreovirus μNS proteins and is thus near the beginning of the minimal C-terminal region of μNS that we showed to be sufficient for inclusion formation in GFP fusions (Fig. 4).
To determine whether the conserved residues with metal-chelating potential in the consensus motif are important for inclusion formation, we first generated constructs encoding Gln (Q) substitutions for both His570 and the adjacent residue, His569, as well as a Ser (S) substitution for Cys572 (Fig. 6A). The constructs were generated with all three of these mutations in the setting of either full-length μNS or μNS(471-721). We then transfected CV-1 cells with these mutant plasmids and costained the cells with anti-μNS antibodies and MAb FK2 for conjugated ubiquitin. Both proteins, designated μNS(1-721)QQS and μNS(471-721)QQS, were diffusely distributed in the cytoplasm and nucleus (Fig. 6B and data not shown), and neither strongly colocalized with conjugated ubiquitin (data not shown). This distribution was in sharp contrast to the inclusions in which both full-length μNS and μNS(471-721) concentrated (data not shown for this experiment, but see previous figures), demonstrating that one or more of residues His569, His570, and Cys572 is important for inclusion formation.
We next generated constructs encoding single mutations at residues Cys561 (to Ser), His569 (to Gln), His570 (to Gln), Cys572 (to Ser), or His576 (to Gln) within full-length μNS. The mutant plasmids were transfected into CV-1 cells, and the cells were costained with anti-μNS antibodies and MAb FK2 for conjugated ubiquitin. Both μNS(1-721)H570Q and μNS(1-721)C572S were diffusely distributed in the cytoplasm and nucleus (Fig. 6B). In contrast, μNS(1-721)C561S, μNS(1-721)H569Q, and μNS(1-721)H576Q all collected in globular inclusions indistinguishable from those formed by wild-type μNS (Fig. 6B). None of the mutant proteins strongly colocalized with conjugated ubiquitin (data not shown). From these findings, we conclude that His570 and Cys572 are specifically required for μNS to form factory-like inclusions in transfected cells.
NS recruitment to factory-like inclusions containing μNS residues 1 to 12 connected by GFP to μNS residues 471 to 721. We have previously shown that NS colocalizes with full-length μNS but not with μNS(41-721) or μNS(13-721) in cotransfected cells, indicating that the N terminus of μNS is required for recruiting NS (7, 26). Because both NS and the N-terminal 41 residues of μNS fused to the N terminus of GFP [μNS(1-41)/GFP] are diffusely distributed in cells, we encountered difficulties in using IF microscopy to determine whether the N terminus of μNS is sufficient for association with NS (7, 26). Upon finding in the present study that GFP/μNS(471-721) collected in factory-like inclusions in transfected cells (Fig. 4B), we recognized that this protein provided a new platform on which to assay protein-protein associations through redistribution to its distinctive inclusions. When GFP/μNS(471-721) was coexpressed with NS, NS remained diffusely distributed in the cytoplasm and nucleus and did not colocalize with the GFP/μNS(471-721) inclusions (Fig. 7A). This was expected, because some portion of μNS residues 1 to 13 is required for recruiting NS to μNS inclusions (26).
To determine whether μNS residues 1 to 41 can direct recruitment of NS to the inclusions of GFP/μNS(471-721), we constructed a plasmid to express these residues fused to the N terminus of GFP/μNS(471-721). When expressed in transfected CV-1 cells, this protein, μNS(1-41)/GFP/μNS(471-721), collected in inclusions similar to those formed by μNS(471-721) or GFP/μNS(471-721) (data not shown). When NS was coexpressed with μNS(1-41)/GFP/μNS(471-721) and coimmunostained with anti-NS MAb 3E10 (3) and the anti-μNS serum, NS strongly colocalized with the μNS(1-41)/GFP/μNS(471-721) inclusions (Fig. 7A). This was in stark contrast to the diffuse distribution of NS expressed alone (26) or coexpressed with GFP/μNS(471-721) (Fig. 7A). When coexpressed with μ2, μNS(1-41)/GFP/μNS(471-721) was redistributed to filamentous inclusions (data not shown), as expected because of the presence of μNS residues 14 to 40 (26) (also see Fig. 2 and 7B). These results newly demonstrate that μNS residues 41 to 470 are dispensable for recruiting NS (summarized in Fig. 7B).
We additionally constructed a plasmid to express μNS residues 1 to 12 fused to the N terminus of GFP/μNS(471-721). When expressed in transfected CV-1 cells, this protein, μNS (1-12)/GFP/μNS(471-721), collected in factory-like inclusions (data not shown). Moreover, when coexpressed, NS was strongly recruited to the μNS(1-12)/GFP/μNS(471-721) inclusions (Fig. 7A). When coexpressed with μ2, μNS(1-12)/GFP/μNS(471-721) was not redistributed to filamentous inclusions (data not shown), as expected because of the absence of μNS residues 13 to 41 (26). These results demonstrate that μNS residues 13 to 470 are dispensable for recruiting NS (summarized in Fig. 7B) and suggest that μNS residues 1 to 12 may be sufficient for this association.
DISCUSSION
Numerous observations suggest that the inclusion bodies in reovirus-infected cells are sites of viral replication and assembly (4, 6, 33, 38, 39). The roles of nonstructural protein μNS in forming the matrix of these viral factories and recruiting other components to them have become a focus of recent studies in this field (4, 6, 7). Given the recent findings and the absence of strong evidence for other functions of this protein, we hypothesize that the primary role of μNS in reovirus infection is to build and organize the factories to promote replication and assembly. We find it intriguing that reovirus may have a protein dedicated to this role, and we therefore wish to learn more about how the different regions of μNS carry out specific activities toward this end.
Current summary of μNS functional regions. Truncation and deletion analyses have been a fruitful approach for identifying discrete regions of μNS involved in its different activities. For example, the current study demonstrated that the C-terminal one-third of μNS (residues 471 to 721) is a sufficient part of this protein for forming factory-like inclusions in transfected cells in the absence of infection or other viral proteins. The present study also identified several smaller regions of sequence, and even single residues, within this third of μNS that are required for inclusion formation.
An otherwise-uncharacterized region of μNS (residues 41 to 172) affected the shape of the inclusions in the present study, causing them to be more compact and less fenestrated when present in the inclusion-forming protein. This effect could reflect a direct interaction of some portion of residues 41 to 172 with the C-terminal, inclusion-forming region of μNS or with some cellular protein, which in an unknown manner produces more compact inclusions. This effect could also reflect an indirect mechanism, such as a reduction in the turnover rate of the inclusion-forming protein such that holes do not develop within the inclusions.
Previous reports have shown that the N-terminal 40 residues of μNS are involved in associations with at least two other reovirus proteins: microtubule-binding protein μ2 and ssRNA-binding protein NS (7, 26). Moreover, the specific residues necessary or sufficient for these activities have been shown to be separable. Residues 1 to 41 of μNS are sufficient for association with μ2 (7), but only some portion of residues 14 to 40 of μNS is necessary for this activity (26). In contrast, some portion of residues 1 to 13 of μNS is necessary for association with NS (26), and residues 1 to 12 of μNS may be sufficient for this activity (the present study). Other, complementary evidence suggests the N-terminal 40 residues of μNS represent a discrete domain. A form of μNS lacking 5 kDa of N-terminal sequence is produced in infected cells concomitantly with the full-length protein (7, 46). Although the origin of this smaller form, called μNSC, remains in question, the fact that it is similar to μNS(41-721), which lacks both μ2 and NS association activities, suggests that it and full-length μNS may play distinguishable roles in infection. Interestingly, a μNSC-like form of μNS has been recently identified in avian orthoreovirus and found not to associate with the avian orthoreovirus NS protein as well, suggesting that the different activities of μNS and μNSC in this regard may be a conserved point of regulation for reoviruses in general (45). The avian orthoreovirus μ2 protein has not yet been examined for μNS association.
Except for the modulation of inclusion shape described in the present study and a predicted coiled-coil segment from residues 518 to 561 (25), the large central portion of μNS spanning residues 41 to 560 remains without well-characterized activities. This region from positions 41 to 560 may contain the residues responsible for μNS associations with other reovirus components, including the individual core-surface proteins, 1, 2, and 2 and whole core particles (6, 7). In fact, these activities are shared by μNS and μNSC, indicating they do not require the N-terminal 40 residues of μNS. Further truncation/deletion analyses are needed to determine whether each of the individual core-surface proteins, as well as core particles, may associate with μNS/μNSC through a distinct set of residues, such as those for μ2 and NS in the unique N-terminal region of μNS.
Other components with which μNS/μNSC may yet be shown to associate include the reovirus RNA-dependent RNA polymerase 3, the reovirus outer-capsid proteins, the reovirus RNA molecules, and any possible number of cellular factors. For each of these potential associations, again, further truncation/deletion analyses could aid in ascertaining whether each component may associate with μNS or μNSC through a distinct set of residues in μNS. In any case, the emerging picture of μNS is one of a protein involved in multiple interactions to form the factory matrix and to recruit other components to the factories, with a modular organization of its primary sequence for performing these many distinct activities.
How does μNS form factory-like inclusions? There are many possible interactions that could be involved in forming the three-dimensional structure that each inclusion likely represents, even before the addition of other viral components as found in the factories in infected cells. For example, μNS-μNS interactions may be all that are required for forming factory-like inclusions. Alternatively, interactions of μNS with one or more cellular factor may be necessary, with the cellular factor(s) acting as bridges between μNS monomers or oligomers. The presence of predicted coiled-coil segments in μNS has led to the suggestion that μNS forms a basal small oligomer, possibly a dimer, through -helical coiled-coil interactions (25). If so, then inclusions are likely built by linking these small oligomers together, with or without the help of cellular factors. In any case, according to the new results, μNS(471-721) must be capable of mediating or instigating all of the necessary interactions for inclusion formation. The smaller regions within μNS(471-721) that are required for forming inclusions might be directly involved in these interactions.
The consensus motif that spans residues 563 to 574 of μNS and that is partially conserved in the homologous proteins of avian orthoreoviruses and aquareoviruses includes residues His570 and Cys572, which are required for inclusion formation. Given that His and Cys residues have strong potential to chelate transition metal ions such as Zn2+, we hypothesize that metal chelation by His570 and Cys572 is a part of their role in forming inclusions. Moreover, considering that His570 and Cys572 are so closely spaced and that nearby residues with the same potential for metal chelation—Cys561, His569, and His576—are dispensable for inclusion formation, we hypothesize that His570 and Cys572 form half of an intermolecular metal-chelating motif, similar to the zinc hook of Rad50 (18) or the zinc clasp of CD4/8 and Lck (20, 21). The latter two motifs provide four zinc-chelating residues, two from each participating subunit, and contribute to homodimerization in the case of Rad50 or heterodimerization in the case of CD4/8 and Lck. Of course, further work is needed to determine whether μNS residues His570 and Cys572 indeed chelate a metal ion and, if so, whether this chelation may contribute to μNS homodimerization or μNS heterodimerization with a cellular protein, either of which could be required for inclusion formation.
Some portion of the C-terminal eight residues of μNS (Phe-Ser-Val-Pro-Thr-Asp-Glu-Leu) is also required for inclusion formation. Since this region contains no His or Cys residues, it must contribute to forming inclusions by a mechanism distinct from that hypothesized for His570 and Cys572. Although this region is not highly conserved in the homologous proteins of avian orthoreoviruses and aquareoviruses, each of the μNS homologs terminates in Leu and has at least one acidic residue and no basic residues in its C-terminal eight positions. Further work is needed to determine whether these C-terminal residues may contribute to μNS-μNS interaction or to μNS interaction with a cellular protein, either of which could be additionally required for inclusion formation.
Evidence in the present study suggests that the more N-terminal of the predicted coiled-coil segments in μNS (25) is required for inclusion formation but that the loss of function resulting from its deletion can be largely rescued by fusion to GFP. We interpret this result to indicate that the role of the first predicted coiled-coil segment is in self-association, which GFP can also mediate (9), but we acknowledge that this interpretation requires further testing. Although the more C-terminal one of the two predicted coiled-coil segments in μNS has not been directly tested for its role in inclusion formation, we expect that at least part of it is also required. Coiled-coil interactions between μNS subunits could, for example, mediate the formation of basal oligomers, which then interact through other motifs to form the inclusions. Alternatively, one or both of the coiled-coil segments could mediate hetero-oligomerization with a cellular protein. As noted in Results, predicted coiled-coil segments are also found flanking the proposed metal chelation sequences in each of the μNS homologs from avian orthoreoviruses and aquareoviruses, suggesting a conserved function for these motifs.
Formation of μNS inclusions as a tool to study protein-protein associations inside cells. The characteristic subcellular localization of μNS and its inclusion-forming derivatives—in cytoplasmic, phase-dense, globular inclusions—has proven a potent tool for studying associations between μNS and other viral proteins in infected or transfected cells (4, 6, 7). In the present study, we acquired original evidence that GFP/μNS(471-721) may provide a useful platform for examining possible associations between any two proteins in transfected cells. In particular, full-length proteins or protein regions expressed as fusions to GFP/μNS(471-721) should localize to the distinctive, globular inclusions induced by the region from positions 471 to 721. The capacity of this fused protein or protein region to associate with another, "test" protein can then be assayed by examining cells in which the test protein has been coexpressed with the inclusion-forming fusion protein. If the test protein localizes to the globular inclusions, then it and the fused protein or protein region can be concluded to associate. Of course, nonfused GFP/μNS(471-721) should be tested in parallel as a negative control for the specificity of test protein localization to the inclusions. A range of relative expression levels of the two proteins could also be tested, in case relative overexpression of the test protein may retard inclusion formation by the other. We are currently extending our studies of the feasibility of this approach as a general one for studying protein-protein associations within the "native" cellular environment in which the proteins normally reside and function.
ACKNOWLEDGMENTS
We express our sincere gratitude to Elaine Freimont and Jason Dinoso for laboratory maintenance and technical assistance and to other members of our lab for helpful discussions. We also thank John Patton and John Parker for reviews of a draft of the manuscript.
This study was supported in part by NIH grants R01 AI47904 (M.L.N.) and F32 AI56939 (C.L.M.) and by a junior faculty grant from the Giovanni Armenise-Harvard Foundation (M.L.N.). T.J.B. and M.M.A. were also supported by NIH grant T32 AI07245 to the Viral Infectivity Training Program. J.A.H. was also supported by NIH grant T32 GM07226 to the Biological and Biomedical Sciences Training Program. At earlier stages of this work, C.L.M. was also supported by NIH grant T32 AI07061 to the Combined Infectious Diseases Training Program.
T.J.B. and M.M.A. contributed equally to this study.
REFERENCES
Antczak, J. B., and W. K. Joklik. 1992. Reovirus genome segment assortment into progeny genomes studied by the use of monoclonal antibodies directed against reovirus proteins. Virology 187:760-776.
Attoui, H., Q. Fang, F. M. Jaafar, J. F. Cantaloube, P. Biagini, P. De Micco, and X. De Lamballerie. 2002. Common evolutionary origin of aquareoviruses and orthoreoviruses revealed by genome characterization of golden shiner reovirus, grass carp reovirus, striped bass reovirus and golden ide reovirus (genus Aquareovirus, family Reoviridae). J. Gen. Virol. 83:1941-1951.
Becker, M. M., M. I. Goral, P. R. Hazelton, G. S. Baer, S. E. Rodgers, E. G. Brown, K. M. Coombs, and T. S. Dermody. 2001. Reovirus NS protein is required for nucleation of viral assembly complexes and formation of viral inclusions. J. Virol. 75:1459-1475.
Becker, M. M., T. R. Peters, and T. S. Dermody. 2003. Reovirus NS and μNS proteins form cytoplasmic inclusion structures in the absence of viral infection. J. Virol. 77:5948-5963.
Broering, T., A. McCutcheon, V. Centonze, and M. Nibert. 2000. Reovirus nonstructural protein μNS binds to reovirus cores but does not inhibit their transcription activity. J. Virol. 74:5516-5524.
Broering, T. J., J. Kim, C. L. Miller, C. D. Piggott, J. B. Dinoso, M. L. Nibert, and J. S. L. Parker. 2004. Reovirus nonstructural protein μNS recruits viral core surface proteins and entering core particles to factory-like inclusions. J. Virol. 78:1882-1892.
Broering, T. J., J. S. L. Parker, P. L. Joyce, J. Kim, and M. L. Nibert. 2002. Mammalian reovirus nonstructural protein μNS forms large inclusions and colocalizes with reovirus microtubule-associated protein μ2 in transfected cells. J. Virol. 76:8285-8297.
Brookes, S. M., A. D. Hyatt, and B. T. Eaton. 1993. Characterization of virus inclusion bodies in bluetongue virus-infected cells. J. Gen. Virol. 74:525-530.
Campbell, R. E., O. Tour, A. E. Palmer, P. A. Steinbach, G. S. Baird, D. A. Zacharias, and R. Y. Tsien. 2002. A monomeric red fluorescent protein. Proc. Natl. Acad. Sci. USA 99:7877-7882.
Dales, S. 1963. Association between the spindle apparatus and reovirus. Proc. Natl. Acad. Sci. USA 50:268-275.
Dales, S., P. Gomatos, and K. C. Hsu. 1965. The uptake and development of reovirus in strain L cells followed with labeled viral ribonucleic acid and ferritin-antibody conjugates. Virology 25:193-211.
Eichwald, C., J. F. Rodriguez, and O. R. Burrone. 2004. Characterization of rotavirus NSP2/NSP5 interactions and the dynamics of viroplasm formation. J. Gen. Virol. 85:625-634.
Fields, B. N., C. S. Raine, and S. G. Baum. 1971. Temperature-sensitive mutants of reovirus type 3: defects in viral maturation as studied by immunofluorescence and electron microscopy. Virology 43:569-578.
Fujimuro, M., H. Sawada, and H. Yokosawa. 1994. Production and characterisation of monoclonal antibodies specific to multi-ubiquitin chains of polyubiquitinated proteins. FEBS Lett. 349:173-180.
Gillian, A. L., and M. L. Nibert. 1998. Amino terminus of reovirus nonstructural protein sigma NS is important for ssRNA binding and nucleoprotein complex formation. Virology 240:1-11.
Gillian, A. L., S. C. Schmechel, J. Livny, L. A. Schiff, and M. L. Nibert. 2000. Reovirus nonstructural protein NS binds in multiple copies to single-stranded RNA and shares properties with single-stranded DNA binding proteins. J. Virol. 74:5939-5948.
Gomatos, P. J., O. Prakash, and N. M. Stamatos. 1981. Small reovirus particle composed solely of NS with specificity for binding different nucleic acids. J. Virol. 39:115-124.
Hopfner, K. P., L. Craig, G. Moncalian, R. A. Zinkel, T. Usui, B. A. Owen, A. Karcher, B. Henderson, J. L. Bodmer, C. T. McMurray, J. P. Carney, J. H. Petrini, and J. A. Tainer. 2002. The Rad50 zinc-hook is a structure joining Mre11 complexes in DNA recombination and repair. Nature 418:562-566.
Huismans, H., and W. K. Joklik. 1976. Reovirus-coded polypeptides in infected cells: isolation of two native monomeric polypeptides with affinity for single-stranded and double-stranded RNA, respectively. Virology 70:411-424.
Huse, M., M. J. Eck, and S. C. Harrison. 1998. A Zn2+ ion links the cytoplasmic tail of CD4 and the N-terminal region of Lck. J. Biol. Chem. 273:18729-18733.
Kim, P. W., Z. Y. Sun, S. C. Blacklow, G. Wagner, and M. J. Eck. 2003. A zinc clasp structure tethers Lck to T-cell coreceptors CD4 and CD8. Science 301:1725-1728.
Lee, P. W. K., E. C. Hayes, and W. K. Joklik. 1981. Characterization of anti-reovirus immunoglobulins secreted by cloned hybridoma cell lines. Virology 108:134-146.
Mayor, H. D. 1965. Studies on reovirus. 3. A labile, single-stranded ribonucleic acid associated with the late stages of infection. J. NCI. 35:919-925.
Mbisa, J. L., M. M. Becker, S. Zou, T. S. Dermody, and E. G. Brown. 2000. Reovirus μ2 protein determines strain-specific differences in the rate of viral inclusion formation in L929 cells. Virology 272:16-26.
McCutcheon, A. M., T. J. Broering, and M. L. Nibert. 1999. Mammalian reovirus M3 gene sequences and conservation of coiled-coil motifs near the carboxyl terminus of the μNS protein. Virology 264:16-24.
Miller, C. L., T. J. Broering, J. S. L. Parker, M. M. Arnold, and M. L. Nibert. 2003. Reovirus NS protein localizes to inclusions through an association requiring the μNS amino terminus. J. Virol. 77:4566-4576.
Miller, D. J., M. D. Schwartz, and P. Ahlquist. 2001. Flock house virus RNA replicates on outer mitochondrial membranes in Drosophila cells. J. Virol. 75:11664-11676.
Morgan, E. M., and H. J. Zweerink. 1975. Characterization of transcriptase and replicase particles isolated from reovirus-infected cells. Virology 68:455-466.
Moss, B. 2001. Poxviridae: the viruses and their replication, p. 2637-2671. In D. M. Knipe and P. M. Howley (ed.), Fields virology. Lippincott Williams & Wilkins Co., Philadelphia, Pa.
Parker, J. S., T. J. Broering, J. Kim, D. E. Higgins, and M. L. Nibert. 2002. Reovirus core protein μ2 determines the filamentous morphology of viral inclusion bodies by interacting with and stabilizing microtubules. J. Virol. 76:4483-4496.
Quimby, B. B., and A. H. Corbett. 2001. Nuclear transport mechanisms. Cell. Mol. Life Sci. 58:1766-1773.
Restrepo-Hartwig, M. A., and P. Ahlquist. 1996. Brome mosaic virus helicase- and polymerase-like proteins colocalize on the endoplasmic reticulum at sites of viral RNA synthesis. J. Virol. 70:8908-8916.
Rhim, J. S., L. E. Jordan, and H. D. Mayor. 1962. Cytochemical, fluorescent-antibody and electron microscopic studies on the growth of reovirus (ECHO 10) in tissue culture. Virology 17:342-355.
Richardson, M. A., and Y. Furuichi. 1985. Synthesis in Escherichia coli of the reovirus nonstructural protein NS. J. Virol. 56:527-533.
Roizman, B., and A. E. Sears. 2001. Herpes simplex viruses and their replication, p. 2231-2295. In D. M. Knipe and P. M. Howley (ed.), Fields virology. Lippincott Williams & Wilkins Co., Philadelphia, Pa.
Schwartz, M., J. Chen, M. Janda, M. Sullivan, J. den Boon, and P. Ahlquist. 2002. A positive-strand RNA virus replication complex parallels form and function of retrovirus capsids. Mol. Cell 9:505-514.
Sharpe, A. H., L. B. Chen, and B. N. Fields. 1982. The interaction of mammalian reoviruses with the cytoskeleton of monkey kidney CV-1 cells. Virology 120:399-411.
Silverstein, S. C., and S. Dales. 1968. The penetration of reovirus RNA and initiation of its genetic function in L-strain fibroblasts. J. Cell Biol. 36:197-230.
Silverstein, S. C., and P. H. Schur. 1970. Immunofluorescent localization of double-stranded RNA in reovirus-infected cells. Virology 41:564-566.
Silvestri, L. S., Z. F. Taraporewala, and J. T. Patton. 2004. Rotavirus replication: plus-sense templates for double-stranded RNA synthesis are made in viroplasms. J. Virol. 78:7763-7774.
Spendlove, R. S., E. H. Lennette, J. N. Chin, and C. O. Knight. 1964. Effect of antimitotic agents on intracellular reovirus antigen. Cancer Res. 24:1826-1833.
Stamatos, N. M., and P. J. Gomatos. 1982. Binding to selected regions of reovirus mRNAs by a nonstructural reovirus protein. Proc. Natl. Acad. Sci. USA 79:3457-3461.
Theron, J., H. Huismans, and L. H. Nel. 1996. Site-specific mutations in the NS2 protein of epizootic haemorrhagic disease virus markedly affect the formation of cytoplasmic inclusion bodies. Arch. Virol. 141:1143-1151.
Touris-Otero, F., M. Cortez-San Martin, J. Martinez-Costas, and J. Benavente. 2004. Avian reovirus morphogenesis occurs within viral factories and begins with the selective recruitment of NS and A to μNS inclusions. J. Mol. Biol. 341:361-374.
Touris-Otero, F., J. Martinez-Costas, V. N. Vakharia, and J. Benavente. 2004. Avian reovirus nonstructural protein μNS forms viroplasm-like inclusions and recruits protein NS to these structures. Virology 319:94-106.
Wiener, J. R., J. A. Bartlett, and W. K. Joklik. 1989. The sequences of reovirus serotype 3 genome segments M1 and M3 encoding the minor protein μ2 and the major nonstructural protein μNS, respectively. Virology 169:293-304.
Wilson, I. A., H. L. Niman, R. A. Houghten, A. R. Cherenson, M. L. Connolly, and R. A. Lerner. 1984. The structure of an antigenic determinant in a protein. Cell 37:767-778.
Wolf, E., P. S. Kim, and B. Berger. 1997. MultiCoil: a program for predicting two- and three-stranded coiled coils. Protein Sci. 6:1179-1189.
Zweerink, H. J., M. J. McDowell, and W. K. Joklik. 1971. Essential and nonessential noncapsid reovirus proteins. Virology 45:716-723.(Teresa J. Broering , Mich)