INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Mammalian orthoreoviruses (reoviruses) encode three nonstructural proteins, µNS, NS, and 1s, whose roles during viral infection remain poorly understood. This report concerns NS, which is known to be essential for reovirus replication based on the phenotype of a conditionally lethal (temperature-sensitive) mutant with its lesion in the NS-encoding S3 gene segment (9, 27). The NS protein comprises 366 amino acids and has a molecular mass of 41 kDa. Interaction of NS with the viral plus-strand RNAs in infected cells is well documented (3, 14). Moreover, when NS from infected cells is used to bind those RNAs in vitro, it protects 20- to 40-nucleotide fragments of the RNAs from RNase T1 digestion (34). Since the protected fragments include the 3' ends of at least some of the plus-strand RNAs, it was proposed that NS binds specifically to those regions (34). In addition, NS and two other reovirus proteins, µNS and 3, are found to associate with the viral plus-strand RNAs shortly after they are transcribed in infected cells and before minus-strand synthesis converts them into the double-stranded RNA (dsRNA) genome segments (3). These findings have led to a hypothesis that NS plays a role in selecting or condensing the viral plus-strand RNAs for packaging during early stages of particle morphogenesis (3, 14, 28, 34). A role for NS in translation of proteins from the reovirus plus-strand RNAs has also been suggested (10, 14).

Evidence for a direct role of NS in minus-strand synthesis is limited. The temperature-sensitive mutant whose lesion maps to the S3 gene segment (27) produces little or no dsRNA at restrictive temperatures (9, 15), but this indicates only that NS provides a required function at or before minus-strand synthesis in the replication cycle. Other previous work demonstrated that NS-containing ribonucleoprotein complexes isolated from reovirus-infected cells display a poly(C)-dependent poly(G) polymerase activity (12, 13). However, recombinant NS expressed in Escherichia coli or insect cells does not display this activity (10, 28), whereas recombinant reovirus protein 3 obtained from mammalian or insect cells does (35; D. L. Farsetta and M. L. Nibert, unpublished data). These findings suggest that the NS-containing complexes characterized by Gomatos et al. (12, 13) also contained 3, which is a known component of the reovirus RNA polymerase (22, 35). Thus, despite not having polymerase activity itself, NS might play a role in minus-strand synthesis within these 3-containing complexes.

In the present study, we performed additional characterizations of the RNA-binding properties of NS to learn more about its roles in reovirus replication. In initial experiments, we analyzed purified, baculovirus-expressed NS (10) for its capacity to bind single-stranded RNA (ssRNA) molecules in gel mobility shift assays. The results indicated that multiple units of NS bind to single molecules of ssRNA with positive cooperativity and in numbers dependent on RNA length such that each unit of NS covered approximately 25 nucleotides at saturation. Competition experiments with various nucleic acids addressed the specificity of NS binding. These studies revealed a preference for NS to bind single-stranded over double-stranded nucleic acids, with a slight preference for RNA over DNA, but they showed no preference for NS to bind reovirus over nonreovirus RNA. We also characterized the capacity of NS to bind ssRNA to which short DNA oligonucleotides were hybridized. Displacement of these oligonucleotides from the partial duplexes was observed, and further characterizations demonstrated that this strand displacement activity of NS is distinct from that of a classical helicase. The observed activities of NS are similar to those of several well-characterized single-stranded DNA (ssDNA) binding proteins, which are known to be involved in genomic double-stranded DNA (dsDNA) replication, suggesting that NS may play a related role in replication of the reovirus dsRNA genome.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Overexpression and purification of NS. NS was overexpressed in insect cells infected with a recombinant baculovirus containing the NS-encoding reovirus type 1 Lang S3 gene and purified as described previously (10). Although RNase A was used in the purification procedure, no nuclease activity was detected in the final purified preparations of NS. As noted in the previous study (10), the NS protein obtained by this protocol was 95% pure according to results with Coomassie-stained sodium dodecyl sulfate-polyacrylamide gels and appeared to be relatively homogeneous in that it migrated between 7- and 9S in 5 to 20% sucrose gradients in the absence of RNA. Nonetheless, whether these complexes represent one discrete type of NS oligomer (e.g., tetramer or hexamer) or a mixture and whether or not they can self-associate under some conditions to form higher-order multimers remain to be determined.

Construction of the plasmids pGEM-DS4 and pGEM-3'DS4. The reovirus type 3 Dearing S4 gene was amplified from a linearized plasmid (19) by PCR (29) with Vent DNA polymerase (New England Biolabs, Beverly, Mass.) and primers (Integrated DNA Technologies, Coralville, Iowa) representing the 5' and 3' ends of S4 plus EcoRI and PstI sites, respectively. The purified PCR product was digested with EcoRI and PstI, again purified, and ligated into those sites in the pGEM-4 plasmid (Promega, Madison, Wis.). A PCR with other primers was used to amplify nucleotides 738 to 1196 of the S4 gene from the new plasmid as well as the majority of pGEM-4 sequences between the PstI and EcoRI sites. Another PCR was used to amplify nucleotides 1 to 536 of the S4 gene. A final PCR was used to join and amplify the purified products from the first two reactions (8). The purified product from the final reaction was digested with SpeI and XhoI and ligated to the purified SpeI-XhoI fragment of the cloned S4 gene to generate new plasmid pGEM-DS4, with the entire S4 gene positioned immediately 3' to the SP6 RNA polymerase promoter such that the first nucleotide of the SP6 transcript was S4 nucleotide 1 (23).

Generation of RNA probes and competitors by in vitro runoff transcription. To synthesize linear DNA templates with the desired termini for in vitro runoff transcription, we performed PCR with different primer pairs (Integrated DNA Technologies) (Table 1) and Deep Vent DNA polymerase (New England Biolabs). For making either full-length or partial (5'S4-121, 5'S4-151, 5'S4-181, 5'S4-211, 5'S4-271, and 3'S4-121) S4 plus-strand transcripts, one primer represented the plus strand of pGEM-DS4 or pGEM-3'DS4 in vector sequences 5' to the SP6 promoter and the second primer represented the minus strand of pGEM-DS4 or pGEM-3'DS4 in S4 sequences beginning either at the 3' end (relative to the plus strand) or the indicated (Table 1) internal position of S4. For making the 121-nucleotide vector transcript (pGEM-121), one primer represented the plus strand of pGEM-4Z in sequences 5' to the SP6 promoter and the second primer represented the minus strand of pGEM-4Z in the appropriate downstream sequences. The fragments were isolated from 1% agarose gels and purified using the Qiaex II kit (Qiagen, Valencia, Calif.).

Gel mobility shift assays. Various amounts of purified NS in phosphate-buffered saline (137 mM NaCl, 3 mM KCl, 8.4 mM Na₂HPO₄, 1.6 mM KH₂PO₄) were incubated with various amounts of ³²P-labeled RNA in gel shift buffer (12 mM Tris-HCl [pH 7.5], 0.1 mg of bovine serum albumin/ml, 10 mM -mercaptoethanol, 0.2% Tween 20, 10 U of RNasin) (32) for 15 min at room temperature in a total volume of 10 µl. The samples were mixed with gel loading buffer (25 mM Tris, 192 mM glycine, 0.25% bromophenol blue, 0.25% xylene cyanole FF, 30% glycerol) and subjected to electrophoresis through either a 1% agarose-Tris-acetate-EDTA gel (40 mM Tris-acetate, 1 mM EDTA) at 90 V or a 5% polyacrylamide Tris-glycine native gel (described above) at 10 mA. The times of electrophoresis varied depending on the size of the RNA probe used in the assay. The gels were dried onto filter paper under vacuum, and the samples were visualized by phosphorimaging (Molecular Dynamics, Sunnyvale, Calif.). The amount of radioactivity in the samples was quantitated using ImageQuant software (Molecular Dynamics). Concentrations of NaCl from 0 to 300 mM had little effect on the binding of NS to RNA. The standard binding reaction mixture used throughout this study contained less than 30 mM NaCl.

Immunoblots. Proteins in the 5% Tris-glycine polyacrylamide gels were blotted onto a nylon membrane (Bio-Rad, Hercules, Calif.) at 30 V overnight at 4°C in transfer buffer (25 mM Tris, 192 mM glycine). The membrane was probed with the anti-NS antibody (10) at a dilution of 1/500 using the protocol supplied with the Bio-Rad color development reagent kit. The secondary antibody was visualized with color development reagents 5-bromo-4-chloro-3-indoylphosphate p-toluidine salt and p-nitroblue tetrazolium chloride (Bio-Rad).

Hill plots. A Hill plot was generated for the 121-, 151-, and 181-nucleotide RNAs in Fig. 3 as well as for three other assays done with a 121-nucleotide RNA and NS amounts from 0.24 to 6.0 pmol (1 µg  24 pmol) (26). The amount of bound NS was calculated as follows. First the amount of radiolabeled RNA in each shifted band was quantitated and expressed as a percentage of total RNA in the reaction mixture. Next the percentage was multiplied by the number of NS units in each band (e.g., one NS unit in the fastest-migrating band and four in the slowest). Finally, the amounts of NS in all four bands were summed and the sum was divided by the total amount of protein added to each sample. The amount of unbound NS = 1  the amount of bound NS. The log₁₀ (bound NS/unbound NS) versus the log₁₀ (NS concentration) was plotted, and the best-fit line to the linear portion of the graph was calculated using Excel (Microsoft, Redmond, Wash.). The slope of the line is the Hill coefficient (40) and was calculated for all six assays and averaged.

Competition assays. A radiolabeled RNA probe (0.24 pmol each) was mixed with the unlabeled competitors in various relative amounts, after which 4.8 pmol of NS was added, and the samples were incubated for 15 min at room temperature. For one set of samples, NS was combined with radiolabeled RNA probe and incubated for 15 min as described above. The competitor was then added, and the samples were incubated for an additional 15 min or 2 h. Equivalent weights of RNA (concentration determined by A₂₆₀ [30]) were used. The radiolabeled RNA probe used for both competition experiments was the 121-nucleotide fragment from the 5' end of S4 in the pGEM-DS4 plasmid. The nucleic acid competitors used for competition between ss- and dsRNA and ss- and dsDNA were the following: ssRNA, full-length in vitro-transcribed S4 gene; dsRNA, reovirus type 3 Dearing genomic RNAs; ssDNA, circular M13 plasmid containing rhinovirus sequences; dsDNA, SmaI-linearized pGEM-4Z vector containing the reovirus type 1 Lang S3 gene (10). The unlabeled RNA competitors used for competitions were the following: the 5' 121 nucleotides of S4 (same RNA used as the labeled probe), the 3' 121 nucleotides of S4, and 121 nucleotides from the multiple-cloning region of the pGEM-4Z vector (Promega) that did not contain reovirus sequences. The amount of radiolabeled RNA bound was determined as the amount of RNA in the upper four shifted bands versus the total amount of RNA. All results are expressed relative to the sample containing no competitor.

RNA-DNA and RNA-RNA partial duplexes. Nineteen-nucleotide DNA oligonucleotides (Integrated DNA Technologies) were designed to be complementary to the 3' end, the middle, or the 5' end of the 5'S4-121 RNA (Table 2). The oligonucleotides were 5'-end-labeled with polynucleotide kinase (New England Biolabs) and [-³²P]ATP, extracted with phenol-chloroform, and precipitated with ethanol. The oligonucleotides were then hybridized to the 5'S4-121 RNA in hybridization buffer (100 mM NaCl, 50 mM Tris-HCl, 2 mM EDTA) at 90°C for 5 min and cooled to room temperature.

ATPase and displacement assays. In the NS samples for the ATPase assay, 75 pmol of purified NS was incubated with 3 µCi of [-³²P]ATP (NEN) in gel shift buffer with or without EDTA (12.5 mM). In the core samples for the ATPase assay, 6 × 10⁹ reovirus cores were incubated with 3 µCi of [-³²P]ATP (NEN) in nucleoside triphosphatase buffer (50 mM Tris-morpholinoethanesulfonic acid [MES] [pH 8.5], 5 mM MgCl₂) (25) with or without EDTA (12.5 mM). The RNA-DNA hybrid (0.11 pmol) was added to all samples, which were incubated at room temperature or 35°C (cores) for 30 min in a total volume of 5 µl each. One microliter of each sample was then spotted onto a polyester-backed polyethyleneimine-cellulose thin-layer chromatography plate (Sigma) and developed with ascending solvent (1 M formic acid, 0.5 M LiCl). The reaction products were visualized by phosphorimaging (Molecular Dynamics). To assay for DNA oligonucleotide displacement activity in parallel with ATPase assays, samples were prepared in the same manner except that no [-³²P]ATP was added. The entire volume of each sample was then analyzed on a 5% polyacrylamide Tris-glycine native gel as described above. For other experiments in which DNA or RNA oligonucleotide displacement activity was measured, various amounts of purified NS and 0.24 to 0.26 pmol of RNA-DNA or RNA-RNA were added to gel shift buffer in a total volume of 10 µl and then the mixtures were incubated at room temperature for 15 min before running the entire sample on a 5% polyacrylamide Tris-glycine native gel.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Gel mobility shift assays for NS binding to 1,196- or 151-nucleotide ssRNAs. The RNA-binding activity of purified NS was examined using an agarose gel mobility shift assay. In initial experiments, increasing amounts of NS were incubated with runoff transcripts representing full-length ssRNA plus strands of the reovirus type 3 Dearing S4 gene (data not shown). More than one band was seen in the mobility shift assay with this long RNA (1,196 nucleotides); however, most of the intermediate-size complexes migrated as a smear, making it difficult to distinguish the number of different complexes.

View larger version (87K): [in this window] [in a new window]   FIG. 1.   Gel mobility shift assay with a 151-nucleotide fragment of reovirus S4 plus strand and immunoblot analysis of a similar assay with anti-NS serum. (A) Increasing amounts of purified NS were incubated with 0.21 pmol of radiolabeled ssRNA comprising the 5' 151 nucleotides of the reovirus S4 plus strand (5'S4-151). The samples were then subjected to electrophoresis in a 5% Tris-glycine polyacrylamide native gel until the free RNA was near the bottom of the gel. The radiolabeled RNA in the dried gel was visualized by phosphorimaging. NS-RNA complexes with six distinct mobilities are indicated at left. (B) Increasing amounts of NS were incubated with and without unlabeled 5'S4-151 RNA (0.8 pmol) and subjected to electrophoresis as described above. An immunoblot assay was performed to detect the protein. The amounts of RNA and protein in these samples were increased so that the ratios were similar to those in panel A but the protein was more easily detected. Arrows, comparable bands in panels A and B.

Gel mobility shift assay for NS binding to ssRNAs of other lengths. The 5'S4-151 RNA appeared to be shifted into six distinct bands over the range of amounts of NS (Fig. 1A). The presence of six shifted bands suggested that if the larger amounts of NS were fully saturating the RNA, then a maximum of six units of NS were binding to each RNA molecule. Moreover, since the RNA was 151 nucleotides long, each unit of NS must be covering about 25 nucleotides if the NS units are evenly distributed. To test this hypothesis, three additional plus-strand RNAs, differing in length by 30 nucleotides (5'S4-121, 5'S4-181, and 5'S4-211), were generated as runoff transcripts from the same S4 clone used to generate the 5'S4-151 RNA. The new RNAs were incubated with increasing amounts of NS and were analyzed in the polyacrylamide gel mobility shift assay. As predicted from initial findings with 5'S4-151 (Fig. 1A), 5'S4-121 was shifted into five distinguishable complexes, 5'S4-151 was shifted into six distinguishable complexes, 5'S4-181 was shifted into seven distinguishable complexes, and 5'S4-211 was shifted into eight distinguishable complexes with increasing amounts of NS (Fig. 2A). To compare the mobilities of the different complexes formed with each RNA, all four RNAs were incubated with the same amount of NS (18 pmol) and subjected to electrophoresis in adjacent lanes of the same gel (Fig. 2B). This analysis demonstrated that with each 30-nucleotide increase in RNA length over this range, one new lower-mobility complex was seen. In later experiments, 5'S4-271 RNA was also tested and found to be shifted into 11 distinguishable complexes with increasing amounts of NS (data not shown). Considering the size of each RNA and the number of distinguishable complexes formed with each, we concluded that the binding unit of NS covers a 24- to 27-nucleotide region of RNA and that the total number of NS units that bound per RNA molecule is determined by the overall length of the RNA.

View larger version (45K): [in this window] [in a new window]   FIG. 2.   Gel mobility shift assays with S4 plus-strand RNA fragments ranging in size from 121 to 211 nucleotides in 30-base increments. Radiolabeled ssRNA fragments comprising the following 5' portions of the S4 plus strand were synthesized: 121 nucleotides, 5'S4-121; 151 nucleotides, 5'S4-151; 181 nucleotides, 5'S4-181; and 211 nucleotides, 5'S4-211. (A) Increasing amounts of purified NS were separately incubated with 0.21 to 0.22 pmol of each RNA, and the samples were analyzed as described for Fig. 1A. NS-RNA complexes with five to eight distinct mobilities are indicated to the left of each gel. (B) Purified NS (18 pmol) was incubated with each of the four RNAs as described for panel A (RNA sizes are indicated above the lanes) and subjected to electrophoresis as described for Fig. 1A. To provide better separation of the shifted complexes, the gel was run longer in this experiment such that the free RNA was run off the bottom.

Binding cooperativity. Since multiple units of NS bound to a single RNA molecule, the binding could have exhibited positive, negative, or no cooperativity. To address these possibilities, the ratio of bound to total NS in each sample was calculated for the 121-, 151-, and 181-nucleotide RNAs from Fig. 2A. In addition, three other experiments using the 121-nucleotide RNA and smaller protein amounts were performed. Hill plots were generated for all six data sets (40). A representative plot for the 121-nucleotide RNA is shown (Fig. 3). The linear region of this curve has a slope >1 indicating that NS binds ssRNA with positive cooperativity. When the data from all six experiments were averaged and NS was assumed to bind ssRNA as a monomer, the Hill coefficient was calculated at 1.43 ± 0.10. Since the RNA-binding unit of NS is thought to be a small oligomer (10), the calculations were also done for NS binding as a dimer, trimer, and tetramer. The Hill coefficient increased by 0.02 to 0.03 unit with each step in oligomer order (data not shown), indicating that NS binding to ssRNA exhibited positive cooperativity regardless of the oligomeric nature of the binding unit.

View larger version (20K): [in this window] [in a new window]   FIG. 3.   Hill plot of NS binding to ssRNA to determine cooperativity. A gel shift assay with 0.24 pmol of radiolabeled 5'S4-121 RNA was performed as described for Fig. 2A except that the concentrations of purified NS ranged from 0.24 to 6 pmol. The amount of NS bound at each concentration was calculated as described in Materials and Methods. Log10 (bound NS/1  bound NS) was plotted relative to log10 (NS concentration), and a best-fit line was calculated for the linear portion of the graph (the equation and coefficient of determination for the line are shown in the box).

Competition assays with single- and double-stranded nucleic acids. Previous work demonstrated that NS does not bind detectably to dsRNA or dsDNA in filter-binding or sedimentation assays (14, 28) but does bind to ssDNA in the filter-binding assay (28). To compare the affinities of NS for ssRNA and other nucleic acids in a more quantitative fashion, we performed competition assays. Radiolabeled 5'S4-121 RNA was combined with unlabeled competitor nucleic acids (reovirus ssRNA, reovirus dsRNA, ssDNA, or dsDNA) before the addition of NS. The RNA-protein complexes were then analyzed using the polyacrylamide gel mobility shift assay (Fig. 4A). Both ssRNA and ssDNA competed efficiently with the labeled ssRNA probe for NS binding, whereas dsRNA and dsDNA competed much less efficiently (Fig. 4A). For example, when 10-fold more ssRNA or ssDNA competitor was added only 23% ± 2% or 37% ± 11%, respectively, of the radiolabeled 5'S4-121 RNA was bound. In contrast, when 20-fold more dsRNA or dsDNA competitor was added, 82% ± 7% or 98% ± 4%, respectively, of the radiolabeled 5'S4-121 RNA was bound. In addition, with either single- or double-stranded competitors, RNA competed with only a slightly greater efficiency for binding to NS than did DNA (Fig. 4A). In summary, NS bound efficiently to single-stranded, but not double-stranded, nucleic acids and exhibited a slight preference for RNA over DNA.

View larger version (32K): [in this window] [in a new window]   FIG. 4.   Competition assays for NS binding to ssRNA using ssRNA, dsRNA, ssDNA, or dsDNA as the competitor. Each data point represents the mean from three experiments, and the error bars represent the standard deviation of the mean. (A) Increasing concentrations of unlabeled competitor ssRNA, dsRNA, dsDNA, or ssDNA (see Materials and Methods) were combined with 0.24 pmol of radiolabeled 5'S4-121 ssRNA. Purified NS (4.8 pmol) was added to each sample, and the sample was incubated. The samples were subjected to electrophoresis and visualized as described for Fig. 1A. The upper four shifted RNA bands were included in the quantitation of bound RNA. The ratios of competitor to probe RNA were calculated from weights rather than molar amounts of nucleic acid to reflect numbers of potential NS binding sites rather than numbers of nucleic acid molecules. (B) The assay was performed as described for panel A except that the unlabeled competitor ssRNAs were 121 nucleotides from the 5' end of the S4 plus strand (5'S4-121), the 3' end of the S4 plus strand (3'S4-121), or the vector sequence (pGEM-121). For one set of samples, NS was added to the labeled RNA prior to addition of the 5'S4-121 competitor.

Competition assays with reovirus and nonreovirus ssRNAs. Previous work demonstrated NS binding to reovirus RNA in infected cells (3, 14) and to nonreovirus RNA in vitro (14, 28). However, no experiments addressed whether NS exhibits a preference for reovirus RNA (14, 28). Sequences at the 5' and 3' termini of all 10 reovirus RNAs are conserved (2) and therefore may be involved in distinguishing reovirus from nonreovirus RNA for binding by proteins such as NS during steps in reovirus replication. To test this hypothesis, we used a competition assay to compare the affinities of NS for RNAs representing the 5' end of the reovirus S4 plus strand, the 3' end of the reovirus S4 plus strand, and nonreovirus sequences derived from pGEM-4Z. Radiolabeled 5'S4-121 RNA was incubated with unlabeled competitor RNAs (5'S4-121, 3'S4-121, or pGEM-121) before addition of NS. The RNA-protein complexes were then analyzed using the polyacrylamide gel mobility shift assay (Fig. 4B). In this assay, all three competitor RNAs competed to approximately the same level with radiolabeled 5'S4-121 RNA for binding to NS. For example, when fivefold more 5'S4-121 competitor RNA was added, only 13% ± 3% of the radiolabeled 5'S4-121 RNA was bound by NS. When fivefold more 3'S4-121 competitor RNA was added, only 22% ± 12% of the radiolabeled 5'S4-121 RNA was bound. Last, when fivefold more pGEM-121 competitor RNA was added, only 13% ± 9% of the radiolabeled 5'S4-121 RNA was bound. Although these data suggested that NS does not preferentially bind reovirus sequences, it was possible that both ends of the RNA together might be required to confer specificity. However, the amount of 5'S4-121 RNA bound to NS in the presence of the full-length ssRNA competitor in Fig. 4A was similar to the amount bound in the presence of each of the competitor RNAs in Fig. 4B. Thus, all ssRNAs appeared to compete to about the same level with radiolabeled 5'S4-121 RNA for binding to NS. This suggests that NS does not distinguish between RNA sequences corresponding to the 5' and 3' ends of the S4 gene and does not exhibit preference for binding to reovirus sequences.

Gel mobility shift assays for NS binding to partially duplex RNA-DNA hybrids and evidence for a strand displacement activity. We hypothesized that NS may require a single-stranded 5' or 3' end of RNA to initiate binding. To test this hypothesis, an end-radiolabeled DNA oligonucleotide complementary to 19 nucleotides at the 3' end of the 5'S4-121 RNA (Table 2) was hybridized to that RNA. The RNA-DNA hybrid was then incubated with increasing amounts of NS and analyzed using the polyacrylamide gel mobility shift assay (Fig. 5). The mobility of this RNA-DNA hybrid with a duplex region at the RNA 3' end was retarded by the smaller amounts of NS (Fig. 5A). Moreover, the amount of NS that shifted the majority of this RNA-DNA hybrid (30 pmol of NS to shift 0.26 pmol of hybrid in Fig. 5A) was comparable to the amount that shifted the majority of the 121-nucleotide RNA (24 pmol of NS to shift 0.22 pmol of RNA in Fig. 2A). Similar results were observed when other end-radiolabeled DNA oligonucleotides (Table 2) were hybridized to either the middle or the 5' end of the 5'S4-121 RNA (data not shown). We therefore concluded that NS does not require a large single-stranded region at either the 5' or the 3' end of RNA to initiate binding.

View larger version (75K): [in this window] [in a new window]   FIG. 5.   Gel shift assay with ssRNA having an RNA-DNA duplex region at its 3' end. A 19-nucleotide DNA oligonucleotide that was exactly complementary to the 3' end of the RNA was 32P-labeled at its 5' end. (A) The oligonucleotide was hybridized to unlabeled ssRNA (5'S4-121) and purified. Differing amounts of purified NS were then mixed with 0.26 pmol of hybrid and assayed as described for Fig. 1A. (B) An unhybridized 32P-labeled oligonucleotide was incubated in the presence of increasing amounts of NS and also assayed as described for Fig. 1A. For both panels, the amounts of NS added to samples are indicated beneath the lanes, and the positions of the RNA-DNA hybrid and the DNA oligonucleotide are also indicated.

Evidence that NS is not an ATP-dependent helicase. Since NS displaced DNA oligonucleotides from RNA-DNA hybrids, we recognized that the protein might possess helicase activity. If so, then NS like other RNA or DNA helicases, should require hydrolysis of ATP as a source of energy for the displacement activity (reviewed in reference 16). Although ATP was not added to the displacement assay mixtures described in the previous section, it is possible that there was contaminating ATP from the reaction mixture used to radiolabel the DNA oligonucleotides (see Materials and Methods). To investigate whether ATP hydrolysis was required for NS to displace the DNA oligonucleotides, the displacement assay was repeated in the absence or presence of EDTA, a known inhibitor of ATPase and RNA helicase functions through its chelation of divalent cations (16, 25). In addition, [-³²P]ATP was added to parallel samples, which were then analyzed by thin-layer chromatography to determine if ATP had been hydrolyzed. The NS protein displaced DNA from the RNA-DNA hybrid in the absence or presence of EDTA (Fig. 6A) but did not detectably hydrolyze ATP in either case (Fig. 6B). Reovirus cores, which were analyzed as a positive control for EDTA-sensitive ATPase activity (25), hydrolyzed ATP in the absence, but not in the presence, of EDTA (Fig. 6B). These data suggest that NS does not require hydrolysis of ATP during displacement of the DNA oligonucleotide from the RNA-DNA hybrid.

View larger version (77K): [in this window] [in a new window]   FIG. 6.   Analysis of ATP hydrolysis during strand displacement by NS. Either 12.5 mM EDTA (+) or H2O () was added to 75 pmol of purified NS (NS) or 6 × 109 reovirus core particles (cores) in each sample. Cores were included as a positive control for ATP hydrolysis. All samples were analyzed in duplicate. (A) The RNA-DNA hybrid described for Fig. 5 (0.11 pmol) was added to each sample in one set. These samples were then analyzed as described for Fig. 1A. The positions of the RNA-DNA hybrid and the DNA oligonucleotide are indicated. (B) Three microcuries of [-32P]ATP and 0.11 pmol of the RNA-DNA hybrid were added to each sample in the other set. Products of the reaction with ATP were resolved by thin-layer chromatography and analyzed by phosphorimaging. The positions of unlabeled ATP, ADP, and AMP markers, as determined by UV absorption, are indicated.

Test for strand displacement activity of NS using a partially duplex RNA molecule. Regions of RNA-RNA duplex are generated during the normal course of reovirus replication, certainly in the genomic dsRNA segments and probably also in regions of secondary structure within the free plus-strand transcripts. The strand displacement activity that we demonstrated for NS using RNA-DNA hybrids may therefore reflect an essential activity of this protein at disrupting regions of RNA-RNA duplex during one or more steps in reovirus replication. The poor capacity of NS to bind to fully duplex RNA molecules (Fig. 4A) suggested that partial duplexes were more appropriate for further testing. In experiments with one such partially duplex RNA, we found that NS could not displace a fully complementary 17-nucleotide RNA oligonucleotide (Table 2) from the 3' end of a longer RNA to which it was hybridized (data not shown). Notably, however, this RNA-RNA duplex had a higher calculated melting temperature than the RNA-DNA duplexes used in the preceding experiments (Table 2), which may have limited the capacity of NS to act on it. Further experiments are warranted before concluding that NS can or cannot disrupt dsRNA regions in partially duplex molecules.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

NS binds in multiple copies to single ssRNA molecules, but there are unresolved questions about the nature of these complexes. Results from the gel mobility shift assays indicate that multiple units of NS can bind to a single ssRNA molecule. Such molecules, bound by multiple units of NS, may represent the large nucleoprotein complexes obtained from both reovirus-infected mammalian cells (14) and insect cells infected with a recombinant baculovirus expressing NS (10). By binding in multiple units to the viral plus-strand RNAs (minus-strand RNAs are not released from template plus strands during reovirus replication [1, 31]), NS may provide multiple sites for interactions with other viral or cellular proteins that may play roles in translation, packaging, or minus-strand synthesis. Questions remaining to be answered include whether the multiple units of NS that bind to an RNA molecule stretch out like beads on a necklace or self-associate to form a more compact structure.

What are the unit of NS that binds to ssRNA and the number of nucleotides of RNA that each unit contacts? The purified NS protein migrates as 7- to 9S complexes in 5 to 20% sucrose gradients (10), consistent with oligomers containing three to six 41-kDa monomers of NS. As stated earlier, whether these complexes represent one discrete type of NS oligomer or a mixture and whether they can self-associate to form higher-order multimers in the absence of RNA remain unknown. We nonetheless hypothesize that one of these oligomers or multimers represents the unit of NS that binds to ssRNA in a number of copies dependent on RNA length. The molar ratio of NS to RNA at saturation in this study may indicate the number of NS subunits in the binding unit, that is, if the concentrations of NS are not limiting and if most or all of the added NS is competent for RNA binding. In this case, we can calculate that saturative binding was achieved with NS in 60- to 120-fold molar excess to each RNA in Fig. 2A (12 to 24 pmol of NS for about 0.2 pmol of RNA). Considering that only five to eight units of NS were bound to each RNA in the 121- to 211-nucleotide range at saturation, we can further calculate that each binding unit contained 8 to 24 subunits of NS. If the RNA-free, 7- to 9S NS oligomer is a tetramer, for example, this suggests that a multimer comprising two to six tetramers constitutes each binding unit. Given that some of the purified NS in our preparations might not have been competent for RNA binding, we recognize that these numbers indicate an upper limit for the size of the binding unit. It remains possible that each binding unit comprises one copy of a discrete type of NS oligomer that sediments in the 7- to 9S range in the absence of RNA (10).

No evidence for sequence specificity in NS binding to ssRNA. Since reovirus replication is cytoplasmic (reviewed in reference 24), there is no selective advantage for NS to have a strong preference for binding ssRNA over ssDNA. Accordingly, we found that NS binds to ssRNA with only slightly greater efficiency than to ssDNA. Our evidence that neither dsRNA nor dsDNA is an efficient competitor for NS binding to ssRNA is consistent with previous findings (28), but our further evidence for strand displacement from short duplex regions is novel (see below).

NS displaces DNA oligonucleotides from RNA-DNA hybrids but is not a classical helicase. Since the strand displacement activity of NS bound to RNA-DNA hybrids occurred independently of ATP hydrolysis and was inhibited by MgCl₂ but not by EDTA, the data strongly suggest that NS is not a classical, ATP-dependent helicase. Consistent with this conclusion is the fact that NS lacks the nucleoside triphosphate-binding motifs characteristic of those enzymes (16). Instead, it appears most likely that NS melts duplex regions as it completes its cooperative and saturative binding to a primarily ssRNA molecule, with the energy for melting provided by the binding event(s).

Similarities between NS and ssDNA binding proteins involved in dsDNA replication. The adenovirus DNA binding protein Ad-DBP (21, 42), the herpes simplex virus type 1 protein ICP8 (5), and the Epstein-Barr virus ssDNA-binding protein BALF-2 (38) are all similar to NS in that they can destabilize regions of a nucleic acid duplex independently of ATP hydrolysis and in a manner that is inhibited by MgCl₂ (5, 21, 38, 42). In addition to ATP-independent strand displacement activity, Ad-DBP and ICP8 share the following properties with NS: (i) binding in multiple units to a nucleic acid molecule in a sequence-independent manner; (ii) protection of the single-stranded nucleic acid from nuclease degradation (data for NS are not shown in this paper), (iii) strong preference for single-stranded over double-stranded nucleic acids; and (iv) positive cooperativity for nucleic acid binding (reviewed in references 7 and 20). These and other ssDNA binding proteins have been shown to be involved in dsDNA replication. For example, both in vitro and in vivo studies have shown that Ad-DBP is required for elongation of ssDNA by the adenovirus polymerase (7, 20), possibly involving the ATP-independent helix-destabilizing properties of Ad-DBP (21, 42). Two prokaryotic ssDNA binding proteins, E. coli SSB and bacteriophage T4 gene 32 protein (gp32), are also required for dsDNA replication and share nucleic acid-binding properties with the noted DNA animal virus proteins and NS but cannot destabilize DNA duplexes (7). Of course, a clear difference between NS and ssDNA binding proteins is the nucleic acid substrate for genome replication (RNA versus DNA). Nonetheless, considering the large number of similarities, we propose that NS plays a role in genome replication similar to that played by the ssDNA binding proteins. Specifically, NS may bind to the reovirus plus-strand RNAs and stabilize them in a conformation that allows the reovirus RNA polymerase and other potential cofactors to mediate minus-strand synthesis to produce the dsRNA gene segments. Some secondary structures in the plus-strand RNAs may be unwound directly by NS (similarly to Ad-DBP [21, 42]), whereas others may need to be unwound by another viral or cellular protein and then stabilized in the single-stranded state by NS binding (similarly to gp32 [7]). Studies to address interactions among NS, the polymerase, and other proteins should provide further insight into the potential role of NS in reovirus genome replication.

Similarities to the nonstructural proteins of other Reoviridae members. Other members of the virus family Reoviridae, all of which have dsRNA genomes, also encode nonstructural proteins with ssRNA-binding activity. For instance, the bluetongue virus (orbivirus) nonstructural protein NS2 (41 kDa) binds ssRNA and forms 7S complexes (39). Similarly, the rotavirus nonstructural protein NSP2 (35 kDa) has sequence-independent ssRNA-binding activity (18), forms 10S multimers (17), and interacts with the rotavirus RNA polymerase VP1 (17). Recent data indicate that the binding activities of NSP2 are in fact very similar to those of NS shown in this study, including binding to single ssRNA molecules in multiple copies, with positive cooperativity, with little or no specificity for rotavirus sequences, and with substantially greater affinity for ssRNA than dsRNA (37). NSP2 has not yet been reported to destabilize nucleic acid duplexes, as was NS in this study, but was reported to have nucleoside triphosphatase and autophosphorylation activities (37). The latter have not been demonstrated for NS (Fig. 6). It remains to be determined whether reovirus NS, rotavirus NSP2, bluetongue virus NS2, and similar nonstructural proteins from other Reoviridae members play strictly analogous roles in the replication cycles of their respective viruses.