Movement Protein Pns6 of Rice dwarf phytoreovirus Has Both ATPase and RNA Binding Activities

Cell-to-cell movement is essential for plant viruses to systemically infect host plants. Plant viruses encode movement proteins (MP) to facilitate such movement. Unlike the well-characterized MPs of DNA viruses and single-stranded RNA (ssRNA) viruses, knowledge of the functional mechanisms of MPs encoded by double-stranded RNA (dsRNA) viruses is very limited. In particular, many studied MPs of DNA and ssRNA viruses bind non-specifically ssRNAs, leading to models in which ribonucleoprotein complexes (RNPs) move from cell to cell. Thus, it will be of special interest to determine whether MPs of dsRNA viruses interact with genomic dsRNAs or their derivative sRNAs. To this end, we studied the biochemical functions of MP Pns6 of Rice dwarf phytoreovirus (RDV), a member of Phytoreovirus that contains a 12-segmented dsRNA genome. We report here that Pns6 binds both dsRNAs and ssRNAs. Intriguingly, Pns6 exhibits non-sequence specificity for dsRNA but shows preference for ssRNA sequences derived from the conserved genomic 5′- and 3′- terminal consensus sequences of RDV. Furthermore, Pns6 exhibits magnesium-dependent ATPase activities. Mutagenesis identified the RNA binding and ATPase activity sites of Pns6 at the N- and C-termini, respectively. Our results uncovered the novel property of a viral MP in differentially recognizing dsRNA and ssRNA and establish a biochemical basis to enable further studies on the mechanisms of dsRNA viral MP functions.


Introduction
Cell-to-cell movement is required for both local and systemic infection by plant viruses. Viruses encode movement proteins (MP) to facilitate such movement. The specific movement mechanisms vary among viruses. Some viruses, such as Tobacco mosaic virus, encode MPs that can alter the size exclusion limit (SEL) of plasmodesmata [1] and bind RNAs [2], and may thus move as ribonucleoprotein complexes [3]. Other viruses, including Grapevine fanleaf virus, move through tubular structures formed inside modified plasmodesmata that are induced by the viral MPs [4]. Many MPs are multifunctional. The 25 K MP (TGBp1, a triple gene block component) encoded by Potato virus X (PVX), has RNA binding, RNA helicase and Mg 2+ -dependent ATPase activities in vitro [5,6]. It can increase the SEL of plasmodesmata in trichome cells of Nicotiana clevelandii [7]. It also functions as an RNA silencing suppressor that is important for PVX spreading within a host plant [8,9]. A recent study indicated that the 25 K protein interacts with Argonaute proteins (AGO 1-4) and mediates their degradation through the proteasome pathway [10].
Rice dwarf phytoreovirus (RDV) is a member of the genus Phytoreovirus within the family Reoviridae that replicates in both host plants and insect vectors. The genome of RDV is composed of 12 double-stranded RNAs (dsRNAs) [11]. The sense strand RNAs from all genome segments of RDV contain a 59 terminal consensus sequence (59 GGCAAA---or 59 GGUAAA---) and a 39 terminal consensus sequence (---UGAU 39 or ---CGAU 39) [12]. Many other reoviruses have similar sequences at the ends of their genomic dsRNAs [13,14]. Although such sequence conservation suggests its functional significance, the biological role of these 59and 39-terminal conserved sequences for RDV and other reoviruses remain unclear. In rotavirus, another member of the family Reoviridae, the terminal sequences of mRNAs form the minimal cis-acting signal required for minus-strand synthesis [15,16].
The RDV virion has an outer shell composed of structural proteins P2, P8 and P9, and a core composed of structural proteins P1, P3, P5 and P7 as well as the genomic dsRNAs. The functions of nonstructural proteins Pns4, Pns6, Pns10, Pns11 and Pns12 have previously been characterized. Pns4 can be phosphorylated in vivo and is localized at the periphery of viroplasms [17]. Pns6 is involved in cell-to-cell movement [18]. Pns10 is an RNA silencing suppressor in plants [19,20,21] and is involved in the formation of tubular structures between neighboring insect cells [22]. Pns11 is a nucleic acid-binding protein [23]. Pns12 is essential for the formation of cytoplasmic inclusions and is involved in virion assembly [24]. Utilizing a complementation approach, Li et al. (2004) determined that Pns6, encoded by segment S6, is responsible for the movement of RDV between cells and can restore the cell-to-cell movement ability of a PVX 25 K deletion mutant [18]. In addition, Pns6 was shown to localize to plasmodesmata in epidermal cells of both Nicotiana tabacum bombarded and RDV-infected rice leaves. However, Pns6 did not suppress RNA silencing in cells [19].
To advance our understanding of the cell-to-cell movement mechanisms of dsRNA viruses, we studied the biochemical functions of RDV Pns6. We report here that Pns6 binds both dsRNAs and ssRNAs. Intriguingly, Pns6 exhibits non-sequence specificity for dsRNA but shows preference for ssRNA sequences derived from the conserved genomic 59-and 39-terminal consensus sequences of RDV. Furthermore, Pns6 exhibits magnesiumdependent ATPase activities. Mutagenesis identified the Pns6 RNA binding and ATPase activity sites at the N-and C-termini, respectively. Our results uncovered the novel property of a viral MP in differentially recognizing dsRNA and ssRNA and establish a biochemical basis to enable further studies on the mechanisms of dsRNA viral MP functions.

Results
Sequence analysis of Pns6 reveals potential RNA acid binding region and ATPase/helicase motif In our previous study, Pns6 was shown to restore the cell-to-cell movement activity of a PVX 25 K deletion mutant [18]. The 25 K protein of PVX is a multifunctional protein that contains different functional domains [25]. We predicted the potential functional domains of Pns6 based on amino acid sequence. As Figure 1A shows, the N-terminal region of Pns6 is rich in basic amino acids, potentially containing an RNA-binding site. In the internal region of Pns6, two transmembrane helicase domains were predicted at amino acid positions 207-228 and 254-271. A putative GKS motif, potentially related to ATPase/helicase activity, was present at amino acid positions 125-127 ( Figure 1B).

Pns6 binds both single-stranded and double-stranded RNA
During RDV replication, both dsRNAs and ssRNAs are produced. Therefore, both forms of RNA (Table 1) were used to test the RNA binding activity of Pns6. We first tested RNA probes of non-RDV sequences, including a human a-actin sequence (ssA) and a random dsRNA sequence (dsR), in northwestern blotting assays. Figure 2 shows that Pns6 could bind both probes. We then performed electrophoresis mobility shift assay (EMSA) to further confirm the RNA binding activity of Pns6. As shown in Figure 3A, Pns6 bound dsR as well as an RDV-specific sequence (S3-5) with equal efficiency. Pns6 also bound ssRNAs ( Figure 3B). However, binding to a random ssRNA sequence (ssR) was weaker than binding to RDV-specific ssRNA sequences (S3-5s, S3-5a, S3-3s and S3-3a). These data establish that Pns6 has RNA binding activities and suggest that it appears to have sequence preference for ssRNAs. This was further tested as described below.
Many MP-RNA complexes were disrupted at high salt concentrations [26,27]. As shown in Figure 3C&D, Pns6 bound both dsRNA (S3-5) and ssRNA (S3-5a) in the range of 50-200 mM NaCl. The Pns6-RNA complexes underwent gradual dissociation with increasing salt concentrations, as expected of the influence of salt concentrations on protein-RNA interactions. In one control, BSA did not bind these RNAs. In another control, the RDV P9 protein, prepared in the same manner as Pns6 from recombinant E. coli, showed no interaction with the RNAs. These controls ruled out the possibility that the observed Pns6-RNA interaction was due to (i) non-specific electrostatic interactions or (ii) contaminating proteins from E. coli that bound the RNAs. Rather, they indicate that Pns6 interacted with the RNAs via specific molecular recognition. This specificity is further supported by the following experimental results showing selectivity of Pns6 for some RNA sequences and identifying the Pns6 domain for RNA binding.
Pns6 binds preferentially ssRNAs derived from the terminal consensus sequences of RDV genome Genomic RNAs of RDV consist of 12 dsRNA segments containing consensus sequences at their 59 and 39 ends. As described above, Pns6 showed stronger binding to ssRNA probes containing these conserved sequences than to a random ssRNA ( Figure 3B). We conducted additional experiments to determine whether Pns6 has higher affinity for the terminal consensus sequences in double-stranded as well as single-stranded forms. Four unlabeled dsRNAs were used to compete with the radiolabeled conserved dsRNA S3-5 derived from RDV, including a dsRNA of random sequence (dsR), a nonconserved RDV dsRNA (S3m) and conserved dsRNAs from the 59 and 39 termini of RDV S3 segments (S3-5 and S3-3), respectively. The relative amounts of free labeled dsRNAs were used to determine the binding abilities of the competitors. Figure 4A shows that the four competitors had equivalent competitive abilities. The data further establish that Pns6 binds dsRNA in a sequence-non-specific manner.
To test sequence preference of ssRNA for Pns6 binding, three ssRNA competitors with random sequences and different lengths (ssR, ssB and ssA) were used to compete with a conserved ssRNA sequence (S3-5a). Figures 4B and 4C show that short random ssRNAs (21-nt ssR and 50-nt nucleotide ssB) had extremely weak competing abilities. The binding ability of the longer random ssRNA (270-nt ssA) was between that of the short random ssRNAs and that of the conserved ssRNA. Three nonconserved RDV ssRNA sequences of different lengths (S3mss, M2s and M2a) were then used to compete with S3-5a. Figure 4D shows that the binding efficiency of these competitors was similar to that of the random ssRNAs. Four ssRNAs of the RDV conserved sequences (S3-5a, S3-5s, S3-3a and S3-3s) were finally used as competitors, and they had high binding abilities ( Figure 4E). S3-5s and S3-3s had higher competition efficiency than S3-5a and S3-3a ( Figure 4E), suggesting that Pns6 has stronger affinity for sensestranded RDV ssRNAs than for anti-sense ssRNAs. Furthermore, the higher binding affinity of Pns6 with S3-5a and S3-5s suggest that Pns6 had a preference for the 59-end sense-stranded RDV RNAs. All together, these data indicate that Pns6 has preference for ssRNA derived from the terminal consensus sequences of the RDV genome.

The RNA binding site of Pns6 is located at the N-terminal region
The richness of basic amino acids in the N-terminal region of Pns6 suggests a role of this region in RNA binding. To test this, we generated a series of Pns6 mutants to test for RNA binding. As shown in Figure 1B, constructs M1, M2, M3, M12, M23 and M13 are deletion mutants, and the GKS construct contains a mutation at amino acids 125-127 (GKS to AAA). As shown in Figure 5A, northwestern blotting assays showed that all mutants containing the M1 region (i.e., M1, M12 and M13) as well as M3 bound ssRNA, M2 and M23 failed to bind ssRNA, and substitution of GKS with AAA had minimal effect on ssRNA binding. Furthermore, while M1 bound dsRNA of conserved RDV  sequence, M3 did not exhibit clear binding activity (Fig. 5B).
These results indicate that the N-terminal region of Pns6 is responsible for binding ssRNAs as well as dsRNAs.

Pns6 is a magnesium-dependent ATPase
The putative GKS motif at amino acid positions 125 to 127 suggests that Pns6 may be an ATPase. We performed a thin layer chromatography (TLC) assay with purified His-tagged Pns6. Figure 6A shows that His-Pns6 can hydrolyze ATP into ADP and that the released ADP increased with increasing amounts of His-Pns6. This indicated that Pns6 has ATPase activity. BSA and P9 exhibited no ATPase activity, ruling out the possibility that the observed Pns6 ATPase activity was due to contaminating proteins from E. coli. It was previously reported that divalent cations play an important role in ATP hydrolysis reactions [28,29]. To examine whether these ions play a role in ATP hydrolysis by Pns6, magnesium or calcium was added to the TLC assays. As shown in Figure 6B, the ATPase activity of Pns6 increased with increasing concentrations of magnesium, reaching the highest level at 5 mM of magnesium. Increasing concentrations of calcium had little effects on the ATPase activity of Pns6 ( Figure 6C), indicating calcium-independence of this activity. Furthermore, when EDTA was added to the reaction system, the ATPase activity of Pns6 was efficiently inhibited, particularly when the concentration of EDTA was equal to or higher than that of magnesium ( Figure 6D). These data indicate that Pns6 has ATPase activities that depend on magnesium but not calcium.

Conserved motifs and active site of Pns6 ATPase
We performed TLC and colorimetric malachite green assays with purified mutant Pns6 proteins to locate the potential ATPase activity site. Mutants M1, M3, M13 and M23 showed ATPase activity equal to or higher than that of wild-type Pns6 in TLC assays ( Figure 7A). Mutants M2, M12 and GKS/AAA showed a significant reduction in ATPase activity. Consistent results with obtained from the colorimetric malachite green assays ( Figure 7B). These results suggest that the ATPase activity of Pns6 resides in the C-terminal region and that GKS is a conserved ATPase motif.

Discussion
Although our knowledge of the detailed RDV replication cycle is still very limited, it has been proposed that the this virus may have a similar lifecycle to animal reoviruses [14]. Early after infection, reoviruses become partially uncoated to form subviral particles. During this process, the viral mRNA, identical to the sense strand of the genomic dsRNA, is released from the subviral particles while the genomic dsRNA remains inside the particles. Once the viral genomic ssRNAs have accumulated to a high level and are packaged into virions, dsRNAs can be synthesized by viral RdRP (RNA-dependent RNA polymerase) inside the core. The virions can be observed in the cytoplasm at the sites where viral replication or translation have occurred [13,14].
Virions or viral ribonucleoprotein complexes (vRNPs) can be transported through plasmodesmata into adjacent cells. In the case of RDV, it is unclear whether the virus is transported in the form of virions or vRNPs. RDV virions spread among cells of insect vectors through tubular structures composed of RDV Pns10 [22]. These tubular structures are not found in the host plant cells. In our previous report, Pns6 but not RDV virions was localized to plasmodesmata in cell walls [18]. It is possible that the diameter of the RDV virion (about 70 nm) prevents it from moving through plasmodesmata. Our previous and current data support the hypothesis that RDV may move between cells in the form of vRNPs.
Pns6 binds dsRNA in a non-sequence-specific manner. Many known MPs, such as the 30 K protein of TMV, bind ssRNA cooperatively and sequence-non-specifically [30]. Interestingly, Pns6 shows sequence preference when it binds ssRNA. Specifically, it binds preferentially to the conserved 59-and 39-terminal consensus sequences of the RDV genome and exhibits a stronger binding affinity for the RDV 59-end sense strand sequence than with the corresponding antisense strand. This is a novel property unreported for other known plant viral MPs. However, a specific property of reoviruses is that their genomic RNAs contain terminal consensus sequences. In rotavirus, the 39-terminal consensus sequence is recognized by NSP3 and is related to the translation of viral mRNA [31] while the 59 and 39 sequences are also recognized by viral RdRP for efficient dsRNA synthesis [32,33]. Thus, our data also suggest the intriguing possibility that Pns6 has additional roles in viral replication and translation, an important issue to be addressed in future studies.
The ATPase activity of Pns6 may be important for RDV movement. Numerous plant virus MPs have ATPase activities [34]. In potexviruses, the ATPase activity of PVX 25 K may be required to provide the driving force to traffic viral RNA through plasmodesmata or to suppress silencing [9,35]. The conserved GKS sequence is required for the NTP binding function of many proteins [36]. In the viral RNA helicases, which exhibit NTPase activity and have been divided into three superfamilies, the GKS motif is shared by the conserved Walker A site of SF3 helicases and the conserved segment I of SF2 and SF3 helicases [37]. This motif is present in many proteins encoded by reoviruses (e.g., VP6 of the bluetongue virus) and possesses NTPase activity [38,39,40]. The importance of GKS motif for Pns6 ATPase activity is consistent with findings from other viral proteins. When the GKS motif is changed to GAA in Bamboo mosaic virus ORF1, to GAS in the Hepatitis E virus helicase motif I, or to GES in the Hepatitis C virus NS3, the ATPase activities of these proteins decrease by 70% [41,42,43].
In summary, our analyses of RDV Pns6 uncovered the novel property of a viral MP in differentially recognizing dsRNA and ssRNA. Such property, together with the identification of the RNA binding sites and ATPase activity site in the protein, establishes a biochemical basis to enable further studies on the mechanisms of dsRNA viral MP functions in movement and perhaps also other aspects of the viral life cycle.

Plasmid construction
Plasmids pGEM-S6, pGEM-S7 and pGEM-S9 were digested with BamH I and Sal I, and the resulting S6/S7/S9 segments were inserted into the prokaryotic expression vector pET28a to yield pET28a-S6, pET28a-S7, pET28a-S9. Primers for constructing the S6 mutant segments are listed in Table 2. Segments M13 and GKS were generated using overlap extension PCR [44]. All resulting PCR fragments were inserted into pBluescript II KS at the EcoR V site to generate pBS-mutant plasmids. After digesting the pBS-mutant plasmids with BamH I, the resulting mutant fragments were ligated into the prokaryotic expression vector pGEX-4T-1 at the BamH I site to generate pGEX-mutant plasmids. The BamH I/Xho I fragment released from the pGEXmutant was inserted into pET28a at the BamH I/Xho I site to produce pET28a-mutant plasmids.

Protein preparation
Escherichia coli BL21 cells (TaKaRa) containing pET28a-S6, pET28a-S7, pET28a-S9 or pET28a-S6 mutant were used to produce His-tagged Pns6, P7, P9 and mutant proteins, and BL21 cells carrying pGEX-mutant constructions were used to produce GST-fused S6 mutant proteins. The cells were cultured in Luria-Bertani medium until they reached an OD 600 of 0.6, and then the recombinant proteins were induced by adding isopropyl b-Dthiogalactoside (0.5 mM) for 2 hours at 37uC or overnight at 18uC. The soluble recombinant proteins were purified using affinity chromatography with a Ni 2+ -chelating column or a Glutathione Sepharose 4B column. The His-tagged proteins were purified in a buffer containing 20 mM TrisCl and 500 mM NaCl, pH 7.0, and the GST fused proteins were purified in a buffer containing 40 mM Tris and 50 mM NaCl, pH 8.0. The resulting proteins were used for ATPase analysis and electrophoresis mobility shift assays (EMSA). Proteins in inclusion bodies were isolated and recovered for northwestern blotting analysis. Cells were harvested by centrifugation at 5,000 g at 4uC for 15 min and resuspended in lysis buffer (50 mM TrisHCl pH 7.0, 100 mM NaCl and 1 mM EDTA, 1% Triton X-100) at 4uC. After sonication and centrifugation at 20,000 g at 4uC for 15 min, the resulting pellet was washed once with the same lysis buffer. The washed inclusion bodies were resuspended in 6 M guanidine hydrochloride, 100 mM TrisHCl pH 7.0, 100 mM dithiothreitol, 1 mM EDTA. The solubilized proteins were separated by SDS-PAGE. The gel was stained with cold 0.5 M KCl. The respective band was then cut out, crushed, and mixed with the SDS-PAGE loading buffer. After 10 min incubation at 80uC, the sample was centrifuged for 15 min at 13,000 g and the supernatant was collected for northwestern blotting.

Probe preparation
The sequences of the synthesized RNA probes and the templates used for the in vitro transcription are presented in Table 1. Short ssRNAs and dsRNAs were synthesized and purified via HPLC by TAKARA Biotechnology (DALIAN) Co. The MA-XIscript Kit (Ambion) was used to transcribe long RNAs. The ssA-Dig probe was digoxin-labeled with the DIG RNA Labeling Kit (SP6/T7) (Roche). The radioactive probes were end-labeled with c-32 P-ATP using T4 polynucleotide kinase (NEB).

Northwestern blotting and electrophoresis mobility shift assay (EMSA)
Northwestern blotting was used to detect the nucleic acid binding activity of Pns6 and its mutants as described previously [45]. The recovered proteins were separated by SDS-PAGE and then transferred to a nitrocellulose membrane. The membrane was soaked in buffer A (20 mM Tris?HCl, pH 7.5, 50 mM NaCl, 0.1% Triton X-100, 1 mM DTT, 1 mM EDTA, pH 8.0 and 0.02%  Ficoll 400) overnight at 4uC to renature the proteins. The digoxinlabeled RNA probes or the radioactive RNA probes were then added to the buffer. After 30 min of incubation, the membranes were washed three times in buffer A for 30 min at 4uC followed by UV crosslinking. The RNA probes were detected using an antidigoxin antibody (AP conjugated, Roche) or via autoradiography with a PE Cyclone phosphor screen scanner system.
For EMSA, RNA probes were incubated with various concentrations of proteins for 20 min on ice in a 10 ml reaction system containing 50 mM NaCl, 50 mM Tris?HCl, pH 7.5, and 10% glycerol. Samples were separated by native PAGE in a 9% gel in 0.56TBE buffer for 80 minutes at 180 V, and then the gel was dried for autoradiography. In the competition assays, unlabeled RNAs were added to compete with the labeled RNA probes. The competition ability of the RNAs is related to the signal from the free RNA probes, which was quantified with ImageJ (version 1.4) software.

Thin layer chromatography (TLC) and colorimetric malachite green assay
In the TLC ATPase analysis, we used a 10 ml reaction system containing 50 mM Tris, 5 mM MgCl 2 , 1 mM DTT, 0.5 mM ATP and the respective proteins. a-32 P-ATP was added to the reactions at 0.05 mCi/ml. After a 30 min incubation at 37uC, 0.5 ml of the reaction product was dotted onto a PEI thin layer chromatography plate (Merck), and the plate was developed in 0.15 M LiCl and 0.15 M formic acid as described previously [6]. Images were obtained using a PE Cyclone phosphor screen scanner system.
The colorimetric malachite green assay was performed as described previously [46]. We used a 100 ml reaction system containing 50 mM Tris, 5 mM MgCl 2 , 1 mM DTT, 1 mM ATP and the respective protein. A 96-well microplate was used to analyze multiple samples. After 60 min of incubation at 37uC, 45 ml of the colorimetric mixture (0.07% malachite green, 3.7% ammonium sulfate, 2.27% ammonium molybdate tetrahydrate and 0.134% Tween-20, prepared 2 hours prior to the assay) was added to the reactions. To develop the reaction, 45 ml of 15% sodium citrate was added. The OD 570 of the samples was read by a TECAN SUNRISE basic scanner.