This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fenner, B. J.
Right arrow Articles by Kwang, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fenner, B. J.
Right arrow Articles by Kwang, J.

 Previous Article  |  Next Article 

Journal of Virology, July 2006, p. 6822-6833, Vol. 80, No. 14
0022-538X/06/$08.00+0     doi:10.1128/JVI.00079-06
Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Sequestration and Protection of Double-Stranded RNA by the Betanodavirus B2 Protein

Beau J. Fenner, Winnie Goh, and Jimmy Kwang*

Animal Health Biotechnology, Temasek Life Sciences Laboratory, National University of Singapore, Singapore 117604

Received 12 January 2006/ Accepted 24 April 2006


arrow
ABSTRACT
 
Betanodavirus B2 belongs to a group of functionally related proteins from the sense-strand RNA virus family Nodaviridae that suppress cellular RNA interference. The B2 proteins of insect alphanodaviruses block RNA interference by binding to double-stranded RNA (dsRNA), thus preventing Dicer-mediated cleavage and the subsequent generation of short interfering RNAs. We show here that the fish betanodavirus B2 protein also binds dsRNA. Binding is sequence independent, and maximal binding occurs with dsRNA substrates greater than 20 bp in length. The binding of B2 to long dsRNA is sufficient to completely block Dicer cleavage of dsRNA in vitro. Protein-protein interaction studies indicated that B2 interacts with itself and with other dsRNA binding proteins, the interaction occurring through binding to shared dsRNA substrates. Induction of the dsRNA-dependent interferon response was not antagonized by B2, as the interferon-responsive Mx gene of permissive fish cells was induced by wild-type viral RNA1 but not by a B2 mutant. The induction of Mx instead relied solely on viral RNA1 accumulation, which is impaired in the B2 mutant. Hyperediting of virus dsRNA and site-specific editing of 5-HT2C mRNA were both antagonized by B2. RNA editing was not, however, observed in transfected wild-type or B2 mutant RNA1, suggesting that this pathway does not contribute to the RNA1 accumulation defect of the B2 mutant. We thus conclude that betanodavirus B2 is a dsRNA binding protein that sequesters and protects both long and short dsRNAs to protect betanodavirus from cellular RNA interference.


arrow
INTRODUCTION
 
Viral suppression of innate host antiviral responses forms an essential part of the replication cycle of animal viruses, as its efficiency governs not only whether a virus will commence replication but also the level of intracellular accumulation of viral components and, thus, progeny virions. Much effort has recently been focused on one innate antiviral response, namely RNA interference (RNAi), and an emerging class of viral suppressors that block this response (21, 29, 32). Moreover, it is now becoming apparent that, in the largest class of viruses, the positive-strand RNA viruses, viral suppression of RNAi is a virtually ubiquitous adaptation to cope with this efficient and destructive host response.

Among the first animal virus suppressors of RNAi to be characterized on the basis of this ability were the B2 proteins of the alphanodaviruses (20, 22, 33), a group of small bipartite positive-strand animal viruses that infect mammals and insects. These viruses possess a bipartite genome consisting of the 3.1-kb RNA1, which encodes the viral replicase and subgenomic B2-encoding RNA3, and the 1.4-kb RNA2, which encodes the single nucleocapsid protein. Initial investigations indicated that infection of insect cells by flock house virus (FHV), a model alphanodavirus, induces host RNA interference, presumably in an attempt by the host to specifically destroy the accumulating viral RNA (20). To combat this, the virus was found to protect itself through the action of B2, a small nonstructural protein whose function had previously remained obscure. The B2 proteins of FHV, and also of Nodamura virus, have since been found to suppress RNAi at the level of double-stranded RNA (dsRNA), a viral replication intermediate and key inducer for many innate host antiviral pathways. Suppression appears to be through the nonspecific sequestration by B2 of short interfering RNA (siRNA) that would otherwise be loaded into RNA-induced silencing complex and facilitate sequence-specific RNA silencing, and also through B2 coating the viral genome and thus preventing the initial formation of siRNA via the action of the Dicer RNase (2, 23, 25, 33). The nonspecific dsRNA binding activity of the FHV B2 protein explains why it is able to act across kingdoms, being able to block RNAi in both animals and plants (20). It has also been suggested that such activity might render B2 capable of blocking additional dsRNA-dependent host antiviral responses, such as the interferon response (33). Indeed, other viral dsRNA binding proteins, such as the vaccinia E3L and influenza NS1 proteins, have broad activity spectra that extend to both RNAi and the interferon response (22).

Piscine betanodaviruses are another subgroup of nodaviruses that lethally infect a myriad of fish species worldwide and cause severe economic losses in the aquaculture industry (27). Like their alphanodavirus counterparts, the betanodaviruses also possess a B2 protein which, unlike the B2 proteins of the alphanodaviruses, is highly conserved at the amino acid level among the betanodaviruses. Betanodavirus B2 is a 75-amino-acid protein with a molecular size of 8.5 kDa and a pI of 5.16 and shares little identity with known proteins present in sequence databases. The protein localizes throughout the cytoplasm during the early stages of betanodavirus infection of fish cells and subsequently enters the nucleus (6). We recently demonstrated that betanodavirus B2 is important for the intracellular accumulation of viral RNA in a variety of permissive and nonpermissive cell types (6). This phenotype was explained, at least in part, by the ability of betanodavirus B2 to suppress RNAi in animal cells (6), a finding which has also been extended to plant cells (14).

In the current study, we further our understanding of the role of B2 in betanodavirus infection by showing that this protein, like its alphanodavirus counterparts, is a sequence-nonspecific dsRNA binding protein that, by coating the viral dsRNA, prevents the generation of siRNA from long viral dsRNA by Dicer and also binds to short dsRNAs of more than 10 bp in length. This activity enables B2 to interact, presumably in a competitive fashion, with the dsRNA binding domain (dsRBD) of ADAR1, a cellular dsRNA binding protein. We go on to show that, despite this activity, B2 is unable to block the induction of the dsRNA-dependent interferon response in fish cells. We also show that B2 is able to block RNA editing of virus dsRNA and host mRNA by exogenously expressed ADAR in fish and mammalian cells, but this activity does not seem to explain the inability of B2 mutants to accumulate in fish cells.


arrow
MATERIALS AND METHODS
 
Cells and viruses. SB (derived from the barramundi perch, Lates calcarifer) cell culture and infection by greasy grouper nervous necrosis virus (GGNNV) were performed as described previously (12), while BSRT7/5 (1) and HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. SB cells were grown at 28°C, and BSRT7/5 and HeLa cells were grown at 37°C in a 5% CO2-humidified chamber.

Protein expression and purification. The primers used in the current study are described in Table 1. A translational fusion between glutathione S-transferase (GST) and GGNNV B2 was created by inserting the B2 open reading frame, amplified from GGNNV1(2,0) using the primers B2-SalI-FWD and B2-NotI-REV, between the SalI and NotI sites of pGEX-4T3 (Amersham Biosciences). The resulting recombinant protein, GST-B2, and the control protein, GST, were expressed in Escherichia coli BL21(DE3) at 30°C, and the cells were harvested after a 4-h induction period in the presence of 1 mM IPTG (isopropyl-ß-D-thiogalactopyranoside). Cell extracts were prepared essentially as described previously (16), except that cells were broken by sonication using five 30-s pulses on setting 4 (Misonix ultrasonic processor XL), with 30-s cooling intervals on ice. Proteins were purified using glutathione-Sepharose 4B (Amersham Biosciences) according to the manufacturer's recommendation using either low (phosphate-buffered saline [PBS])- or high (PBS containing 1 M NaCl and 2% [wt/vol] Triton X-100)-salt washing buffer. Purified proteins were desalted using Microcon 5000 filters (Millipore) and stored in 50-mM Tris-HCl (pH 7.4) at –20°C. Proteins and nucleic acids were quantitated using a NanoDrop ND-100 spectrophotometer.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Primers used in the current study

His6-tagged dsRBD from zebra fish ADAR1 was expressed from recombinant pQE30 (QIAGEN) in E. coli M15 (pREP4) cells at 37°C as suggested by the vector manufacturer. Lysates were prepared from cell pellets in a denaturing buffer (100 mM NaH2PO4, 10 mM Tris-HCl, 8 M urea, pH 8) to prevent the dsRBD binding to E. coli nucleic acids. His6-tagged dsRBD was purified using Ni2+-nitrilotriacetic acid resin (QIAGEN), concentrated with Centricon 5000 (Millipore) filters, and refolded by stepwise reductions of urea concentration (8 M, 6 M, 4 M, 2 M, and finally none). Finally, buffer was exchanged for 50 mM Tris-HCl, pH 7.4, and the protein was stored at –20°C.

In vitro nucleic acid binding assays. Purified GST-B2 or GST proteins were incubated with synthetic DNA or RNA oligonucleotides of 10, 20, or 40 nucleotides (or base pairs) at concentrations indicated in Results. Binding reactions were typically performed in 20-µl volumes containing GST-B2 or GST, 0.1 µM DNA or RNA, 100 mM NaCl, and 50 mM Tris-HCl, pH 7.4. Reaction mixtures were incubated at 25°C for 30 min to allow protein binding to the nucleic acid, followed by nondenaturing polyacrylamide gel electrophoresis (PAGE) at 5 V/cm using 4% gels buffered with TBE (89 mM Tris-HCl, 89 mM borate, 2 mM EDTA). Nucleic acids were visualized by staining the gels for 5 min in 1 µg/ml ethidium bromide, followed by destaining for 5 min in distilled water and subsequent viewing with a Bio-Rad gel/ChemiDoc system.

In competition assays, 20- or 40-bp RNAs were end labeled with [{gamma}-33P]ATP in 10-µl reaction mixtures containing 10 pmol of oligonucleotide, 10 pmol (20 to 30 µCi) of [{gamma}-33P]ATP, 1 U of T4 polynucleotide kinase (New England Biolabs), and kinase buffer. Labeling reaction mixtures were incubated at 37°C for 45 min and terminated by 2 min of incubation at 95°C. Competition binding assays were performed in 10-µl volumes containing 5 nM-labeled dsRNA probe, different concentrations of unlabeled competitor, 500 nM GST-B2, 100 mM NaCl, and 50 mM Tris-HCl, pH 7.4. Reaction mixtures were incubated and electrophoresed as described above, and the resulting gels vacuum dried and visualized by autoradiography.

Preparation of poly(I:C)-agarose. Approximately 2 mg of poly(I:C) (Amersham Biosciences) was dissolved in 1 ml of cyanoborohydride coupling buffer (Sigma). A 1-ml bed volume of periodate-activated agarose beads (Sigma) was washed twice with 10 ml of PBS and then combined with the coupling buffer-poly(I:C) mixture. The slurry was then mixed at room temperature for 3 h, followed by mixing at 4°C for 20 h, and then centrifuged at 500 x g for 2 min. The beads were washed with 4 ml of coupling buffer, and residual reactive groups were blocked by incubation with coupling buffer containing 0.5 M Tris-HCl (pH 7.4) for 60 min at room temperature with gentle agitation. Finally, the beads were washed three times with 10 ml of PBS and then stored at 4°C in PBS, pH 7.4, containing 0.02% sodium azide. Poly(dI:dC) was used in place of poly(I:C) for the preparation of DNA-agarose beads.

In vivo nucleic acid binding assays. SB cells in 75-cm2 flasks were infected with GGNNV at a multiplicity of infection of 1, and cells were harvested by scraping into the medium at 24 h postinfection, followed by centrifugation at 1,000 x g for 5 min at 4°C. Cell pellets, containing approximately 107 cells, were stored at –80°C until required. Cell lysates were prepared by suspending the cell pellets in a hypotonic buffer (20 mM HEPES, pH 7.4, 10 mM KCl, 1.5 mM MgCl2, 1 mM dithiothreitol [DTT]; Roche Complete protease inhibitor cocktail) and swiftly passing the suspension 10 times through a 26-G needle. The lysate was then mixed with 200 µg dsRNA equivalent of poly(I:C)- or poly(dI:dC)-agarose (prepared as described above) suspended in binding buffer (100 mM KCl, 5 mM MgCl2, 1 mM DTT; Roche Complete protease inhibitor cocktail). The cell extract/agarose mixture was incubated at 4°C for 60 min, followed by two washes with the binding buffer containing 200 mM KCl, with each wash being followed by centrifugation at 200 x g at 4°C for 5 min. Proteins bound to the agarose beads were eluted by boiling the mixture in Laemmli sample buffer for 5 min and resolved by sodium dodecyl sulfate (SDS)-PAGE on a 15% gel. GGNNV B2 was detected by immunoblotting with anti-B2 as described previously (6).

Dicer inhibition assays. A 600-bp dsRNA derived from the GGNNV RNA1 was used as a Dicer substrate. The dsRNA was generated by first amplifying the GGNNV RNA1 region from nucleotides 1701 to 2300 using the primers RNA1-dsRNA-FWD1 and RNA1-dsRNA-REV1 and pGGNNV1(2,0) (6) as the RNA1 template. These primers incorporate 5' and 3' T7 RNA polymerase promoters into the 600-bp RNA1 PCR product, which, following PCR amplification, was gel purified with a QIAquick kit (QIAGEN) and then used as a template for RNA synthesis with T7 RNA polymerase (Stratagene). Template DNA was removed by digestion with RQ1 RNase-free DNase (Promega) and subsequent RNA extraction with Trizol reagent (Invitrogen). The purified dsRNA was quantitated by spectrometry and used as the Dicer enzyme substrate. Dicer cleavage reactions were performed in 10-µl volumes containing 1x Dicer buffer (2.5 mM MgCl2, 150 mM NaCl, 20 mM Tris-HCl, pH 8), 10 µM GST or GST-B2, and 0.01 µM of the 600-bp dsRNA. Reaction mixtures were incubated at 25°C for 30 min to facilitate GST-B2 dsRNA binding, and 1 U of recombinant human Dicer (Stratagene) was then added. Reaction mixtures were incubated for 20 h at 37°C, and the reaction products were resolved by nondenaturing PAGE as described above.

GST pull-down assays. A 5-µg quantity of purified GST or GST-B2 was bound to glutathione-Sepharose (10-µl bed volume) in PBS containing 100 mM NaCl, all in a final volume of 50 µl. To this mixture, approximately 50 ng of purified His6-tagged dsRBD and 10 µg of either poly(dI:dC) or poly(I:C) was added, and the reaction mixture was incubated at 25°C for 30 min. After six 5-min washes in 1-ml volumes of PBS containing 100 mM NaCl, the slurry was combined with 100 µl of Laemmli sample buffer and boiled. Proteins were resolved by SDS-PAGE and subsequently detected by immunoblotting with horseradish peroxidase-conjugated anti-His6 monoclonal antibody (BD Biosciences) and SuperSignal West Pico chemiluminescent substrate (Pierce).

Glutaraldehyde cross-linking. Prior to cross-linking reactions, protein buffers were exchanged for PBS using Centricon 5000 filters (Millipore) to prevent interference by Tris. Purified GST, GST-B2, or cleaved GST-B2 (1 to 2 µg each) was incubated in the presence or absence of 0.1% glutaraldehyde (grade 1; Sigma) at 0.1 or 1 M NaCl, in 25 mM sodium phosphate, pH 7.4, for 30 min at 25°C. GST-B2 was cleaved in a 1-ml reaction mixture containing 300 µg of GST-B2 and 2 U of thrombin (Amersham Biosciences) in PBS at 22°C for 20 h. Cross-linking reaction products were resolved by SDS-PAGE on 15% gels, and the proteins were visualized by staining with Coomassie blue R-250.

Interferon gene expression analysis. Following extraction, RNA samples were treated with RQ1 RNase-free DNase (Promega) to remove contaminating DNA. Primers that targeted the SB Mx gene were designed by first aligning the Mx cDNA sequences for bastard halibut, Chinese perch, gilthead sea bream, and orange-spotted grouper (GenBank accession numbers AB110446, AY392097, AF491302, and AY574372, respectively), all of which belong to the Percomorpha series of fishes, of which barramundi perch (Asian sea bass, Lates calcarifer) is also a member. The primer pair PERC-Mx-FWD1/PERC-Mx-REV1, which targets a 198-bp amplicon in the 3' region of the Mx gene, was designed based on the alignment for quantitative real-time reverse transcription-PCR (qRT-PCR) measurement of SB Mx mRNA levels. Reverse transcription reactions were performed using avian myeloblastosis reverse transcriptase (Promega) and quantitative PCRs were performed using a LightCycler FastStart DNA Master PLUS SYBR green I kit and LightCycler instrument (Roche) according to the manufacturers' instructions. Copy numbers of mRNAs were determined using a standard curve prepared using serial dilutions of T7 RNA polymerase in vitro-synthesized RNAs derived from linearized recombinant pGEM-T vectors containing the relevant PCR amplicons.

RNA editing assays. SB cells were first transfected with different combinations of pcDNA3.1(+) (Invitrogen), pcDNA-B2 (6), and pME18S-ADAR1Dr (GenBank accession no. BC044344) (ADAR1Dr is ADAR1 from Danio rerio [zebra fish]), using 20 µg of each plasmid and a previously described electroporation protocol (6). Following a 24-h incubation to allow B2 and ADAR1Dr expression, cytoplasmic extracts were prepared using NE-PER extraction reagent (Pierce Biotechnology) and aliquots were added to an RNA editing reaction mixture containing 1 pg (2.5 amol) of the above-mentioned 600-bp RNA1-derived dsRNA and reaction buffer (50 mM Tris-HCl, pH 7.9, 150 mM KCl, 5 mM EDTA, 20% glycerol, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride, and 0.2 mg/ml bovine serum albumin) to make a final volume of 20 µl. Following 2 h of incubation at 28°C, reactions were heat inactivated at 95°C for 2 min and the dsRNA was amplified by RT-PCR using the same primers used for amplification of the dsRNA. Products were resolved by agarose gel electrophoresis, the 600-bp DNA band was excised and gel purified (QIAquick kit; QIAGEN), and the DNA was sequenced to assess RNA editing, as described below.

Site-specific editing of the polymerase (Pol) II/5-HT2C transcript (38) in HeLa cells in the presence of ADAR1 expressed from pRK-ADAR1Rn (M. Higuchi and P. H. Seeburg) (ADAR1Rn is ADAR1 from Rattus norvegicus [Norway rat]) was essentially as reported elsewhere (38). Cells grown to 70 to 90% confluence were first transfected with 0.8 µg of pcDNA3.1(+) (Invitrogen) or pcDNA-B2 along with pRK-ADAR1Rn using Lipofectamine 2000 (Invitrogen) and plated in 24-well plates as recommended by the manufacturer. Following a 12-h incubation to allow B2 and ADAR1 expression, cells were secondarily transfected with 0.2 µg of the Pol II/5-HT2C editing substrate reporter. After a further 36 h of incubation, total RNA was prepared from the transfected cells using the Trizol reagent (Invitrogen) and treated with RQ1 RNase-free DNase (Promega) and the editing substrates were amplified by StrataScript reverse transcriptase and Pfu Turbo PCR (Stratagene) using the T7 promoter/BGH-REV1 primers. For all samples, an additional reaction was performed without reverse transcriptase. Products were resolved by agarose gel electrophoresis and, assuming that no DNA contamination was present in the corresponding reverse transcriptase-minus control reaction, were purified using a QIAquick gel purification kit (QIAGEN). Purified products were sequenced on both strands with the RT-PCR primers using an ABI Prism sequencing unit and BigDye reaction chemistry. Chromatogram analysis of sequencing results was performed using the ContigExpress module of Vector NTI Advance, version 9.1 (Invitrogen). Editing efficiency was determined by measuring the trace values of peaks A (unedited site) and G (edited site) in order to determine the G/(A+G) peak ratio, which was reported to 100 to determine the percentage of editing at each site, as described previously (38).

Yeast (Saccharomyces cerevisiae) two-hybrid screening. Saccharomyces cerevisiae strain AH109 (ade– his– leu– trp–; Clontech) was used as a host for two-hybrid screening experiments. The B2 bait vector construct was created by ligating the GGNNV B2 open reading frame to pGBKT7 (trp+; Clontech) as a 0.2-kb NcoI/BamHI PCR-generated fragment. Prey vectors were created by ligating GGNNV B2, GGNNV protein A, and zebra fish ADAR1 genes to pGADT7 (leu+; Clontech) as Pfu Turbo (Stratagene) PCR-generated fragments. ADAR1 and its dsRBD and Z{alpha} domains were amplified from pME18S-ADAR1Dr using primers DR-ADAR1-FWD1/DR-ADAR1-REV1 for full-length ADAR1, DR-ADAR1-dsRBD-FWD1/DR-ADAR1-dsRBD-REV1 for the dsRBD, and DS-ADAR1-ZA-FWD1/DS-ADAR1-ZA-REV1 for the Z{alpha} domain. Proteins B2 and A were amplified from GGNNV1(2,0) (6) using primers PB2-BAIT-FWD1/PB2-BAIT-REV1 and PA-BAIT-FWD1/PA-BAIT-REV1. Recombinant bait and prey vectors were confirmed by DNA sequencing and subsequently used to cotransform AH109 cells using the lithium acetate/polyethylene glycol method (8). Transformants were initially selected on selective dropout medium without Leu or Trp (–Leu/–Trp SD) medium (Clontech), and then colonies were suspended in PBS and serial dilutions were spotted onto –His/–Leu/–Trp and –Ade/–His/–Leu/–Trp SD agar plates to screen for protein-protein interactions.

Protein interaction affinities were measured by first inoculating 250-ml flasks containing 50 ml of –His/–Leu/–Trp SD medium with single colonies of the relevant yeast transformant and incubating the flasks with shaking at 250 rpm at 30°C. The growth rate of the transformants was measured by monitoring the optical density at 600 nm of the culture over time, and the calculated growth rate constants were taken as an indirect measure of protein interaction affinity.


arrow
RESULTS
 
Recombinant B2 binds nucleic acids. Our previous work (6) has indicated that the betanodavirus B2 protein antagonizes RNA interference in animal cells, but the mechanism of its activity remained unknown. Fortunately, however, the purification of a recombinant B2 protein using an E. coli expression host allowed us to shed light on the nature of this small nonstructural protein. B2 was expressed in E. coli as a soluble GST-B2 fusion protein of approximately 35 kDa (Fig. 1A, lane 2) and was validated by immunoblotting with polyclonal anti-B2 and monoclonal anti-GST antibodies (Fig. 1A, lanes 3 and 4). During this analysis, we noted that the yield of GST-B2 was more than 20-fold lower than the yield of nonrecombinant GST (Fig. 1A, compare lanes 1 and 2). Spectrophotometric analysis of the purified proteins to measure protein concentrations revealed that GST-B2, when purified under standard low-salt conditions with PBS washing off the glutathione-Sepharose resin, absorbed strongly at 260 nm, while nonrecombinant GST purified under the same conditions had a more typical protein absorption spectrum with a maximum absorbance at 280 nm (Fig. 1C). The spectral characteristics of the seemingly pure GST-B2 protein preparation suggested the presence of high concentrations of nucleic acid. This was confirmed by resolving the protein on a nondenaturing polyacrylamide gel and staining with either Coomassie blue or ethidium bromide. Inspection of the banding pattern of GST-B2 purified under low-salt conditions revealed the presence of high-molecular-weight banding that comigrated with nucleic acids (Fig. 1B, lanes 2 and 5), while nonrecombinant GST was free from nucleic acids (Fig. 1B, lanes 1 and 4). This finding clearly indicated that B2 was binding nucleic acids from E. coli cells, though the type of nucleic acid being bound was unknown. Overnight incubation of the GST-B2 protein preparation with DNase I and RNase A failed to eliminate or even reduce the observed nucleic acid banding pattern (data not shown), suggesting that either GST-B2 was protecting the bound nucleic acid fraction from digestion or the nucleic acids were neither dsDNA nor single-strand RNA or both. Digestion of GST-B2 with proteinase K and separation of the free nucleic acids on an agarose gel showed that most of the bound nucleic acids were of a relatively small size, corresponding to 50 to 100 bp of dsDNA (Fig. 1D).


Figure 1
View larger version (34K):
[in this window]
[in a new window]
  		FIG. 1. Recombinant GGNNV B2 copurifies with nucleic acids in an E. coli expression system. (A) Denaturing SDS-PAGE analysis of purified GST and GST-B2 and the corresponding immunoblots of GST-B2 with anti-GST and anti-B2. M, marker. (B) Nondenaturing PAGE of GST and GST-B2 purified under low- and high-salt conditions. Proteins were visualized by Coomassie staining, and nucleic acids were visualized by staining with ethidium bromide. The position of the low-mobility banding that corresponds to high-molecular-weight GST-B2/nucleic acid complexes is indicated beside the gel. (C) Scanning spectrophotometric analysis of GST and GST-B2 purified under low- and high-salt conditions, showing the strong absorption of low-salt-purified GST-B2 at 260 nm compared with the major peaks at 280 nm for low-salt-purified GST and high-salt-purified GST and GST-B2. (D) Agarose gel electrophoretogram showing the increase in ethidium staining of nucleic acids bound by GST-B2 before (–) and after (+) digestion of GST-B2 with proteinase K. Discrete nucleic acid bands are indicated by pointers. M, marker.

In order to remove the bound nucleic acids from GST-B2, we repurified the protein using a high-salt washing buffer (PBS with 1 M NaCl and 1% Triton X-100), which was sufficient to yield a more typical protein absorption spectrum (Fig. 1C), and eliminated more than 95% of the high-molecular-weight protein and nucleic acid banding seen after nondenaturing PAGE (Fig. 1B, compare lanes 2 and 3 and lanes 5 and 6).

B2 binds specifically to double-stranded RNA. To determine the nature of the nucleic acids being bound by the recombinant B2 protein, we incubated high-salt-purified GST-B2 or nonrecombinant GST (negative control) with an equimolar concentration of purified single- or double-stranded DNA or RNA oligonucleotides of 10, 20, or 40 residues in length. The sequences of the DNA and RNA oligonucleotides were chosen arbitrarily from the GGNNV RNA1 genome sequence and corresponded to the regions from positions 1429 to 1438, 1429 to 1448, and 1429 to 1468, for the 10-, 20-, and 40-residue DNA/RNA oligonucleotides, respectively. Separation of the binding reaction products by nondenaturing PAGE revealed that, of all the different oligonucleotides, only the 20- and 40-residue double-stranded RNAs were bound by B2, as evidenced by the presence of a band shift in the nondenaturing gels, while no shift was observed in any of the nonrecombinant GST controls (Fig. 2).


Figure 2
View larger version (37K):
[in this window]
[in a new window]
  		FIG. 2. GGNNV B2 binds to double-stranded RNA oligonucleotides of more than 10 base pairs in length. Protein-nucleic acid mixtures were prepared at equimolar (0.5 µM) concentrations in PBS, incubated at 25°C for 30 min, and then combined with a one-sixth volume of loading buffer (50% [wt/vol] sucrose, 0.2% [wt/vol] bromophenol blue). Following incubation, the mixtures were resolved on 7% TBE polyacrylamide gels buffered at 10 V/cm and the nucleic acids were stained with ethidium bromide at 1 µg/ml.

We next looked at the affinity of the recombinant B2 protein for dsRNA by incubating each of the three dsRNAs (10, 20, and 40 bp) at 0.1 µM with B2 concentrations ranging from 0.1 to 10 µM. In the case of the 20- and 40-bp dsRNAs, there was a gradual increase in the amount of bound dsRNA, with increasing concentrations of B2 (Fig. 3A and B), while no significant binding of the 10-bp RNA was observed even at 10 µM B2 (data not shown), which represents a 100-fold molar excess of B2. This finding suggests that B2 binds only dsRNAs of more than 10 bp in length. Moreover, B2 bound the 40-bp RNA more efficiently than it bound the 20-bp RNA, with almost all of the 40-bp RNA bound at a B2 concentration of 1 µM, at which concentration all of the 20-bp dsRNA remained unbound (Fig. 3A and B). We did not observe any significant binding of single- or double-stranded DNA, single-stranded RNA, or DNA-RNA hybrids of 10, 20, or 40 residues even with 10 µM B2 (data not shown), indicating that B2 has little or no affinity for these nucleic acids. Additionally, no change to the binding pattern of B2 was observed when we used scrambled dsRNAs with different nucleotide sequences (data not shown), revealing that B2 binds dsRNA in a sequence-nonspecific fashion.


Figure 3
View larger version (41K):
[in this window]
[in a new window]
  		FIG. 3. GGNNV B2 binds to double-stranded RNA molecules of more than 20 bp with high affinity in vitro. Binding reaction mixtures contained the indicated concentrations of recombinant B2 protein and either 20-bp (A) or 40-bp (B) dsRNA. Double-stranded RNA was visualized by staining with ethidium bromide. (C) Quantitation of recombinant B2 binding to dsRNA based on densitometric analysis of the gel images in panels A (20-bp) and B (40-bp) using the histogram function of Adobe Photoshop CS2. (D) Hill plot of GST-B2 binding to 40-bp dsRNA. The Hill coefficient of 1.39 was obtained from the linear regression line and corresponding equation, shown beside the line. Linear regression using only the central three data points was performed to obtain the maximum gradient, and the resulting line (dashed) and equation are shown. In competition binding assays, labeled 20-bp probe (E) or 40-bp probe (F) was incubated in the presence of GST-B2 and increasing concentrations of unlabeled 40-bp (E) or 20-bp (F) competitor. Competition curves are displayed in panel G, with the dashed lines indicating the molar excess of unlabeled competitor required to compete out 50% of the bound labeled probe.

Quantitative analysis of the gel images obtained for B2 binding to 20- and 40-bp RNAs suggested that B2 had higher affinity for the 40-bp dsRNA (Fig. 3C). We confirmed this observation by performing competitive binding assays (Fig. 3E and F) with labeled 20- or 40-bp dsRNA and unlabeled competitor. Strikingly, quantitation of these data revealed that B2 bound 40-bp dsRNA with at least a 180-fold-higher affinity than it bound 20-bp dsRNA (Fig. 3G), suggesting that siRNAs in the 20-bp range are not the favored target of B2.

The minor laddering effect observed in Fig. 3E and F hinted at cooperative binding by B2. Analysis of the cooperativity of B2 binding using a Hill plot (36) of the B2/40-bp dsRNA gel shift data revealed a clear positive cooperativity, with a Hill coefficient of 1.39 (Fig. 3D). Moreover, using only the central linear section of the Hill plot, as is common in such analyses (5, 9), a coefficient of 2.06 was obtained (Fig. 3D). Thus, we can conclude that B2 binds to dsRNA in a positively cooperative manner.

Given that dsRNA binding by B2 would need to occur under physiological conditions (~100 mM salt) to be biologically relevant, we sought to determine how efficiently B2 bound dsRNA under different salt concentrations. Thus, fixed concentrations of recombinant B2 and the 40-bp dsRNA probe were incubated in the presence of 0 to 500 mM NaCl (Fig. 4A). The results indicated that the interaction between B2 and dsRNA was largely unaffected by salt concentrations of up to 100 mM, decreasing by only 12% from 0 to 100 mM NaCl (Fig. 4B). Beyond 100 mM NaCl, however, the stability of the protein-RNA complex decreased substantially. From 100 to 500 mM NaCl, the amount of B2 bound to dsRNA dropped by almost 30% (Fig. 4B). We therefore concluded that B2 can stably bind dsRNA under physiological salt concentrations, but it is sensitive to concentrations above these levels.


Figure 4
View larger version (32K):
[in this window]
[in a new window]
  		FIG. 4. The interaction between B2 and double-stranded RNA is stable at physiological salt concentrations. (A) Gel shift assay showing interaction of recombinant B2 to 40-bp dsRNA under different salt concentrations. Binding reaction mixtures contained recombinant B2 at 1 µM and 0.1 µM 40-bp dsRNA in 50 mM Tris-HCl buffer, pH 7.4, with the indicated concentrations of NaCl. Binding reaction mixtures were incubated at 25°C for 30 min, the products were resolved on a 4% polyacrylamide gel, and dsRNA was visualized by staining with ethidium bromide. Note the warping of the unbound dsRNA migration front at high-salt concentrations. (B) Quantitation of recombinant B2 binding to dsRNA based on densitometric analysis of the gel using the histogram function of Adobe Photoshop CS2.

B2 protects long dsRNA from Dicer cleavage. A plethora of small nonstructural virus-encoded dsRNA binding proteins have been characterized in recent years (4, 11, 13, 33) and most of these proteins have been implicated in the suppression of one or more of the host cell's innate immune responses, including the dsRNA-induced interferon response and RNA interference. Our previous work (6) revealed that B2 antagonized RNA interference in vivo, and we sought to understand how this phenomenon occurred. Given that B2 was clearly capable of binding dsRNA, we reasoned that B2 might protect long dsRNAs and, in particular, viral dsRNA replication intermediates from Dicer-mediated cleavage. Such protection would prevent the formation of siRNAs that would otherwise be incorporated into the RNA-induced silencing complex and subsequently guide the destruction of viral RNA by RNA interference. We thus synthesized a long dsRNA using GGNNV RNA1 (nucleotides 1701 to 2300) as a template to represent a transient viral dsRNA that would be present during virus replication and added a 1,000-fold molar excess of B2 protein or nonrecombinant GST as a negative control. Following a 30-min incubation to facilitate dsRNA binding by B2, we added recombinant human Dicer enzyme and allowed the cleavage reaction to proceed until the untreated dsRNA control was completely digested into siRNAs (Fig. 5, compare lanes 1 and 2). Dicer cleavage was unaffected by nonrecombinant GST, but was completely blocked by B2 (Fig. 5, compare lanes 3 and 4). In the latter, B2 formed a high-molecular-weight complex with the dsRNA. To rule out the possibility that the observed complex was not simply a complex between Dicer-generated siRNAs and B2, we incubated the reaction mixture at 95°C for 5 min to denature the B2 protein and then slowly cooled the reaction mixture to allow the RNA to reanneal. All of the observed dsRNA recovered from this reaction comigrated with the original untreated 600-bp RNA (Fig. 5, compare lanes 1 and 5), indicating that B2 effectively protected the long dsRNA from Dicer cleavage and thus prevented the formation of siRNAs.


Figure 5
View larger version (18K):
[in this window]
[in a new window]
  		FIG. 5. GGNNV B2 protects long double-stranded RNA from Dicer-mediated cleavage, blocking the formation of short interfering RNAs in vitro. A 600-bp dsRNA derived from the GGNNV RNA1 sequence (lane 1) was digested with recombinant human Dicer (Stratagene) in the absence (–) of other proteins (lane 2) or in the presence (+) of GST (lane 3) or GST-B2 (lane 4). Lane 5 was the same reaction as that of lane 4 except that the mixture was incubated at 95°C for 5 min to denature the GST-B2 protein and release the 600-bp dsRNA. Following incubation, the reaction products were resolved by nondenaturing PAGE and the dsRNA was visualized by staining with ethidium bromide.

Native B2 expressed in fish and mammalian cells binds dsRNA. We next sought to confirm the dsRNA binding properties of B2 in vivo. For this experiment, we prepared dsRNA-agarose beads using the synthetic dsRNA poly(I:C) for pull-down immunoblots with GGNNV-infected SB cells or BSRT7/5 cells transfected with pGGNNV1(2,0) or pGGNNV1{Delta}B2(2,0) (6). The pGGNNV1(2,0) plasmid expresses full-length GGNNV RNA1, while pGGNNV1{Delta}B2(2,0) contains a point mutation in the B2 start codon (T2773C) that ablates B2 translation (6). Poly(I:C) was confirmed to be a suitable substrate for B2 binding by incubating purified recombinant B2 with poly(I:C) or poly(dI:dC) (dsDNA control). The incubation with poly(I:C) but not poly(dI:dC) resulted in the formation of a high-molecular-weight complex with B2 (Fig. 6A), indicating that B2 can bind to this synthetic dsRNA. The incubation of SB and BSRT7/5 cell extracts with dsRNA or dsDNA beads and subsequent immunoblotting with a polyclonal anti-B2 antibody revealed the presence of B2 in only GGNNV-infected SB and pGGNNV1(2,0) transfected BSRT7/5 cell extracts that had been incubated with dsRNA beads (Fig. 6B). Interestingly, far more B2 was detected in the input protein samples for both cell types than was seen after dsRNA-agarose pull down, suggesting that much of the B2 protein present in the extracts was already occupied with cellular dsRNA and was thus unable to bind to the dsRNA-agarose. Nonetheless, this experiment clearly demonstrated that the native B2 protein expressed during GGNNV infection or from a transfected RNA1 replicon is able to bind dsRNA.


Figure 6
View larger version (48K):
[in this window]
[in a new window]
  		FIG. 6. Native B2 expressed during GGNNV infection and from plasmid-based RNA1 replicons binds dsRNA. (A) Poly(I:C) is a suitable substrate for B2 binding. Purified GST-B2 was incubated in the presence of poly(dI:dC) (dsDNA) or poly(I:C) (dsRNA) at 5 µg, and the products were separated by PAGE on a 4% nondenaturing Tris-glycine gel. Protein was visualized by staining with Coomassie blue, and nucleic acid was visualized with ethidium bromide. (B) B2 is recoverable from infected cell lysates by pull down with poly(I:C). GGNNV-infected SB cell extracts were incubated with poly(dI:dC)- or poly(I:C)-agarose, and the bound proteins were resolved by SDS-PAGE, followed by immunoblotting with anti-B2 polyclonal antibodies. Untreated cell extracts were used as an input control for the presence of B2. BSRT7/5 cells were transfected with pGGNNV1(2,0) (RNA1) or pGGNNV1{Delta}B2(2,0) (RNA1{Delta}B2) and cell extracts used for agarose pull-down assays. See Materials and Methods for experimental details.

B2 interacts with itself and other dsRNA binding domains. The sequence-nonspecific dsRNA binding nature of the B2 protein suggested to us that this protein, when expressed in any cell system with dsRNA, should compete with cellular dsRNA binding proteins for available dsRNA substrates. Significantly, many cellular dsRNA binding proteins are involved in innate antiviral immunity, such as the Dicer RNase, the PKR dsRNA-inducible protein kinase, and the adenosine deaminases that act on dsRNA (ADARs). To probe the potential interactions that B2 might have with such proteins in the cell, we generated a series of yeast two-hybrid expression constructs using GGNNV B2 for the bait vector and B2, protein A (the viral replicase), or zebra fish ADAR1 for the prey vector. Of these proteins, ADAR1 possesses both dsRNA and Z-DNA binding domains, while protein A presumably generates dsRNA during viral replication but is not known to bind dsRNA per se. In addition, we also prepared vectors containing only the dsRBD or Z{alpha} domains of ADAR1. Transformants were then screened for protein-protein interactions on medium- and high-stringency selective media. As one might have expected given its apparent ability to cooperatively accumulate on available dsRNA substrates, B2 was able to interact with itself and with B2-B2 cotransformants growing on even the high-stringency medium (Table 2). Interaction also occurred between B2 and ADAR1, though with less affinity than that of the B2-B2 interaction (Table 2). This interaction apparently occurred due to the dsRBD of ADAR1, as B2-dsRBD cotransformants, but not B2-Z{alpha} cotransformants, retained the ability to grow under stringent selection conditions. The same growth patterns were observed when the bait and prey vectors were switched (data not shown), confirming the specificity of these interactions. Interestingly, transformants containing B2 and the dsRBD grew under high-stringency conditions, while those containing B2 and full-length ADAR1 grew under only medium-stringency conditions (Table 2). The B2-dsRBD interaction was not, however, as strong as the B2-B2 interaction, as evidenced by the high growth rate of B2-B2 cotransformants compared to that of the B2-dsRBD cotransformants in medium-stringency medium (Table 2). No interactions were observed between B2 and protein A, though it should be noted that these experiments were performed in the absence of an RNA1 replicon that could serve as a substrate for protein A, and thus, an interaction between these two proteins cannot be ruled out entirely.


View this table:
[in this window]
[in a new window]
  		TABLE 2. GGNNV B2 can interact with itself and with a dsRNA binding domain in a yeast two-hybrid systema

To confirm the putative interaction between B2 and the dsRBD of ADAR1, we first purified the same dsRBD region used in the two-hybrid analysis as a recombinant His6-tagged protein. The purified protein had an apparent molecular size of 26 to 27 kDa, which corresponded well with the predicted protein size of 25.7 kDa (Fig. 7A, left lane). We then incubated purified GST or GST-B2 immobilized on glutathione-Sepharose with His6-tagged dsRBD in the presence or absence of poly(I:C). Pull down of the GSH-Sepharose and subsequent immunoblotting with anti-His6 revealed that His6-tagged dsRBD interacted with GST-B2 in the presence of only poly(I:C) (Fig. 7A). This finding clearly demonstrated that B2 can interact with itself and with another dsRNA binding protein through shared dsRNA substrates.


Figure 7
View larger version (49K):
[in this window]
[in a new window]
  		FIG. 7. B2 is a monomeric protein that interacts with the double-stranded RNA binding domains of ADAR1 through shared dsRNA substrates. (A) Interaction of B2 with ADAR1 dsRBD through dsRNA binding, as shown by GST pull down. Purified His6-tagged dsRBD was incubated with GST or GST-B2 bound to glutathione-Sepharose in the presence of either poly(dI:dC) (dsDNA) or poly(I:C) (dsRNA). Pulled down proteins were resolved by SDS-PAGE, and His6-tagged dsRBD was detected by anti-His6. (B) B2 exists as a monomeric protein in solution. Purified GST, GST-B2, or thrombin-digested GST-B2 was incubated in the presence (+) or absence (–) of 0.1% glutaraldehyde (Glut) at 0.1 or 1 M NaCl for 30 min at 25°C. Proteins were resolved by SDS-PAGE on 15% polyacrylamide gels and stained with Coomassie blue.

The possibility that B2 self-interacted by oligomerization in addition to dsRNA binding was also investigated. The functionally similar FHV B2 protein has recently been shown to form a homodimer in solution (23), while the canonical dsRBDs of RNase III and ADAR bind dsRNA substrates as monomers, though both of these enzymes are catalytically active as dimers (7, 19). To elucidate the oligomeric status of GGNNV B2, we incubated GST (a homodimer), GST-B2, and cleaved GST-B2 with the cross-linking reagent glutaraldehyde at 0.1 and 1 M NaCl and separated the products by SDS-PAGE. Cross-linking markedly reduced protein staining by Coomassie blue, but cross-linked oligomers were still clearly visible (Fig. 7B). As expected, the bulk of the GST formed a dimer at both salt concentrations, though small quantities of tetrameric and other oligomeric complexes were also present (Fig. 7B, lanes 1, 2, 7, and 8). GST-B2 oligomerized as both a dimer and other higher-order oligomers (Fig. 7B, lanes 3, 4, 9, and 10). Importantly, however, the cross-linking of cleaved GST-B2 revealed that while GST formed dimers and other higher-order oligomers, B2 alone remained monomeric under both physiological (100 mM) and high-salt (1 M) conditions (Fig. 7B, lanes 5, 6, 11, and 12). Thus, unlike FHV B2, GGNNV B2 appears to exist as a monomer in solution.

B2 does not antagonize the host interferon response. The ability of B2 to bind dsRNA in vivo suggested that it would antagonize host responses to viral infection that depend on dsRNA for their induction. Other viral dsRNA binding proteins, such as E3L from vaccinia virus, have well-characterized abilities to block dsRNA-dependent interferon induction (40). We decided to further probe the ability of B2 to block host responses other than RNAi by analyzing the effect of B2 on interferon-related gene expression. The Mx gene was chosen as a marker for interferon because it is known to be activated in fish cells in response to dsRNA and virus infection, and it is a commonly used marker for interferon-mediated gene regulation (3, 15, 18, 26, 34). As the gene sequence of Mx from barramundi perch was yet to be determined at the time of this study, we retrieved the previously determined Mx gene sequences of closely related fish species (series Percomorpha) from databases and, from the resulting alignment, developed a primer pair (see Materials and Methods) that targeted two conserved regions. An RT-PCR with the primers using SB RNA yielded a product of the expected size, and subsequent DNA sequencing and amino acid sequence alignment confirmed that the product was indeed from an Mx-like gene (Fig. 8A). Later comparisons of the DNA sequence with a recently determined L. calcarifer Mx gene sequence (39) revealed complete nucleotide identity, further confirming the identity of this product (data not shown). We thus used these primers for qRT-PCR quantitation of Mx gene expression. Mx was induced 25-fold at the late stages of GGNNV infection in SB cells (Fig. 8B). Interestingly, Mx was rapidly induced during the first 9 h of infection but this induction was then slowed and continued at a much lower rate until late in infection, while at the same time, RNA1 accumulation continued at a relatively constant rate. GGNNV infection thus induced Mx expression in SB cells, though the bulk of Mx mRNA appeared late in infection, when the majority of cells already displayed cytopathic effects indicative of approaching cell death. In an attempt to establish a link between Mx induction and B2, we next introduced capped synthetic RNA1 or RNA1{Delta}B2 transcripts into SB cells and monitored Mx expression. These replicons are capable of autonomous intracellular replication, though RNA1{Delta}B2 exhibits drastically reduced accumulation compared to that exhibited by the wild-type RNA1 replicon (6). We found that Mx was strongly induced in response to transfection and subsequent accumulation of RNA1, with Mx mRNA being induced more than 100-fold from 3 to 24 h posttransfection, compared to approximately 25-fold from 3 to 48 h during a native infection (compare Fig. 8B and C). As expected, the transfected RNA1 accumulated over the transfection period, increasing by 16-fold from 3 to 24 h posttransfection (Fig. 8C). This suggested that Mx was being induced in response to the accumulation of wild-type RNA1 and, thus, also in the presence of B2. In the case of RNA1{Delta}B2, Mx was not significantly induced over time, while the accumulation of RNA1{Delta}B2 reached only 14% of wild-type RNA1 levels (Fig. 8C). This result indicates that, during betanodavirus replication in SB cells, B2 does not suppress the induction of Mx, an interferon-stimulated gene.


Figure 8
View larger version (37K):
[in this window]
[in a new window]
  		FIG. 8. GGNNV induces expression of the SB Mx gene, and B2 does not antagonize this response. (A) Alignment of the SB Mx partial amino acid sequence deduced from the amplified cDNA using perch Mx-specific primers with previously characterized Mx proteins. Species are abbreviated as follows: Lc, Lates calcarifer (barramundi perch); Sc, Siniperca chuatsi (Chinese perch); Ec, Epinephelus coioides (grouper); Tr, Takifugu rubripes (pufferfish); Dr, Danio rerio (zebra fish); Hs, Homo sapiens (human). Mx amino acid sequences were obtained from GenBank. (B) RNA1 replication and concomitant Mx induction in SB cells infected with GGNNV. Cells were infected at a multiplicity of infection of 5, and RNA1 and Mx expression was monitored by qRT-PCR as described in Materials and Methods. (C) Correlation between Mx induction and RNA1 accumulation in SB cells transfected with RNA1 and RNA1{Delta}B2. Cells were transfected with 1 µg of capped GGNNV RNA1 or RNA1{Delta}B2 as described previously (6) and total RNA sampled over time. RNA1, Mx mRNA, and 18S rRNA levels were measured by qRT-PCR. Values shown represent the means of at least three independent determinations performed in duplicate, with the standard deviation not exceeding 10% of the mean for any point.

RNA editing by ADAR is blocked by B2. In addition to RNA interference and the interferon response, another dsRNA-dependent cellular antiviral response is RNA editing of long dsRNA substrates by ADARs. RNA editing is thought to target long dsRNAs for destruction by a cellular RNase (31, 35), but it can be antagonized by viral dsRNA binding proteins such as vaccinia E3L (24). We therefore reasoned that B2 might also interfere with dsRNA-dependent RNA editing. To explore this possibility, we first sought to determine whether GGNNV RNA1 was capable of being edited by ADAR1, a predominantly cytoplasmic adenosine deaminase that forms part of the cellular antiviral response (35). To accomplish this, we introduced a zebra fish ADAR1 (ADAR1Dr) expression plasmid or the empty parent vector into SB cells. After a 24-h incubation to allow ADAR1 expression, we prepared cytoplasmic extracts of the transfected cells and added them to the 600-bp RNA1-derived dsRNA mentioned above (Fig. 5). In the presence of ADAR1Dr, the sense-strand 3' 500 bases of this long dsRNA were edited at 16 distinct sites (Fig. 9A). An analysis of nucleotides adjacent to the edited adenine residues revealed a preference of ADAR1Dr for 5' and 3' A/U residues (Fig. 9B). We next sought to determine whether B2 could block the observed RNA editing by cotransfecting SB cells with ADAR1Dr and B2 expression plasmids. Strikingly, extracts prepared from these cotransfected cells displayed only 14% editing activity compared to that displayed by the control (Fig. 9C), clearly indicating that B2 is capable of blocking the activity of ADAR1Dr in an in vitro assay. In the absence of ADAR1Dr, total RNA editing was less than 5% of that of the control, probably due to low-level endogenous SB ADAR1 activity (Fig. 9C). The ability of B2 to block RNA editing was further examined by monitoring ADAR1-dependent, site-specific editing of the human 5-HT2C editing substrate (Fig. 9D) in HeLa cells using mammalian ADAR1 and plasmid-expressed B2. In the absence of B2, the A site of 5-HT2C was edited with approximately 70% efficiency, which is similar to the results of previous studies (38), while in the presence of B2, the editing efficiency was reduced to only 30% (Fig. 9E). Thus, B2 is also capable of antagonizing RNA editing of cellular mRNAs in vivo. To determine whether RNA editing played a role in the observed RNA1{Delta}B2 accumulation defect in SB cells, we sequenced the same 500-bp RNA1 region analyzed for in vitro RNA editing assays using RNA prepared from RNA1- and RNA1{Delta}B2-transfected SB cells as a template for RT-PCR. We did not, however, observe significant RNA editing of any adenosine residues for either of the two RNAs sampled at 3 or 24 h posttransfection (data not shown). Though this does not rule out the possibility that RNA1{Delta}B2 is edited and subsequently destroyed in SB cells, the lack of any detectable RNA editing suggests that this pathway is not a major factor contributing to the accumulation defect of RNA1{Delta}B2.


Figure 9
View larger version (27K):
[in this window]
[in a new window]
  		FIG. 9. B2 antagonizes RNA editing by ADAR in vitro. (A) Sequence of the GGNNV RNA1-derived dsRNA used for in vitro editing assays incorporating ADAR1Dr. Edited residues from the top are shown in bold and underlined. (B) Site specificity of ADAR1Dr using the RNA1-derived dsRNA. A total of 16 adenosine residues were edited in the in vitro reaction, and the frequency of editing for residues with different 5' and 3' neighbors is displayed. (C) Antagonism of RNA editing by B2 in vitro. Cytoplasmic extracts prepared from SB cells transfected with different combinations of pcDNA3.1(+) (vector), pME18S-FL3-ADAR1Dr, and pcDNA-B2 were added to the 600-bp RNA1-derived dsRNA and RNA editing determined as described in Materials and Methods. Values represent the means of three independent experiments, with error bars indicating the standard deviations. (D) The 5-HT2C RNA editing substrate was used to monitor ADAR1Rn-dependent RNA editing. Adenosine residues at the 5-HT2C A to E editing sites are bold. Representative DNA sequencing chromatograms showing the RNA editing of each substrate that occurred in HeLa cells are shown below. The peak heights of the overlapping A (unedited) and G (edited) residues at the 5-HT2C A site were compared to determine the efficiency of RNA editing (see Materials and Methods). (E) Suppression of ADAR1Rn-dependent RNA editing in HeLa cells by GGNNV B2. Cells were transfected with the 5-HT2C editing substrate reporter and different combinations of pcDNA3.1(+) (vector), pRK-ADAR1Rn, and pcDNA-B2, as described in Materials and Methods. Values represent the means of three independent experiments, with error bars indicating the standard deviations.


arrow
DISCUSSION
 
Our finding that B2 is a sequence-nonspecific dsRNA binding protein is perhaps not surprising given the similar phenotypes exhibited by alpha- and betanodavirus B2 mutants and the abilities of both proteins to block RNAi in a variety of cell types. Moreover, despite the lack of sequence similarity between the proteins (6), the dsRNA binding characteristics of GGNNV and FHV B2 are remarkably similar. Both can bind to dsRNA of approximately 20 bp in length but bind to longer dsRNA with higher affinities (25). These findings suggest that, as for B2's ability to block RNAi, the major mode of action is preventing Dicer-mediated cleavage of long dsRNAs originating from viral replication intermediates. This would then beg the question of how, given the affinity of B2 for dsRNA, the protein is separated from the bound dsRNA during replication by the viral replicase. It may be that B2 dissociates as the replicase melts the dsRNA helix and then associates as the duplex reforms. Given that B2 does not visibly associate with the viral replicase in the mitochondrial membrane vesicles formed during GGNNV infection (6, 10), any association would probably occur in the initial stages of infection where replication intermediates form in the cytoplasm.

It is inevitable that the nonspecific dsRNA binding activity of B2 will lead to competition with cellular proteins that normally bind dsRNA. We have shown here that B2, in addition to cumulatively binding to dsRNA, interacts with ADAR1 and, in particular, its dsRNA binding domain. Notably, B2 interacted with itself with greater affinity than with the dsRBD, suggesting that B2 has a higher affinity for dsRNA than does the dsRBD. One would have to assume that in order to prevent siRNA formation the affinity of B2 for dsRNA must be greater than that of host dsRNA binding antiviral proteins such as Dicer. Recent work with the B2 RNAi antagonist of FHV has indicated that this protein has an apparent Kd for siRNAs of approximately 1 nM (2), which compares favorably with the 50-to-60-nM Kd value of Dicer (28, 37). A comparison of data presented here and elsewhere (25) for GST-tagged GGNNV B2 and FHV B2 indicates that both proteins have similar affinities for dsRNA. GGNNV B2 bound >90% of a 40-bp dsRNA when present at a molar ratio of approximately 30:1, while FHV bound >90% of a 44-bp dsRNA at about 20:1 (25), suggesting that GGNNV B2 should have a Kd for dsRNA similar to that of FHV B2.

Interferon induction in response to virus-derived dsRNA is a potent antiviral strategy that encompasses a variety of different systems that function to destroy viral RNA, down regulate translation and induce apoptosis. An emerging trend is that viral suppressors of RNAi also inhibit the induction of the interferon response or components of the interferon system, like the dsRNA-dependent protein kinase PKR and RNase L (32, 40). It was therefore surprising to discover that B2 does not suppress the induction of Mx, an interferon-responsive gene, following transfection of SB cells with RNA1, compared to results with the RNA1{Delta}B2 mutant. Instead, we observed that Mx induction was strictly linked to the ability of RNA1 to accumulate and, therefore, its ability to generate dsRNA replication intermediates. Given that B2 would associate tightly with available dsRNA, including the replication intermediates, it may be that RNA1 replication eventually overwhelms the available B2 and dsRNA then becomes available to cytoplasmic interferon signaling components. In any case, our new findings are corroborated by Chen et al. (3), who have recently shown that betanodavirus strongly induces Mx expression in many tissue types during a native infection of grouper, lending further support to our suggestion that, in the context of fish infection by betanodavirus, B2 does not appear to function as an interferon antagonist.

The ability of B2 to block RNA editing by ADAR1 is consistent with its function as a dsRNA binding protein and also with our finding that B2 appears to interact with itself, presumably by binding to cellular dsRNAs, with higher affinity than with the dsRBD of ADAR1. Our finding that B2 is highly effective at blocking ADAR1 hyperediting of long dsRNAs but far less effective at blocking site-specific editing of cellular mRNAs is remarkably similar to previous work with the vaccinia E3L protein (24). Clearly it would be beneficial for the virus to protect its replication intermediates from being hyperedited by ADAR1 if that process resulted in the dsRNA being targeted for destruction in the cell. Insofar as its additional effect of blocking the editing of cellular mRNAs is concerned, this would be expected to be a residual activity and one which, in the case of the 5-HT2C substrate here, may be minimal due to the short length of dsRNA; the edited 5-HT2C dsRNA region is only 11 bp (see Fig. 9D) and would not likely be a good binding substrate for B2 given the lack of significant binding of a 10-bp dsRNA in vitro. The importance of RNA editing in the fish antiviral response during GGNNV infection remains unclear, as neither wild-type RNA1 nor RNA1{Delta}B2 was detectably edited following its introduction into fish cells. It is worth noting, however, that this work represents the first demonstration that a fish ADAR enzyme is capable of hyperediting dsRNA. Moreover, the zebra fish ADAR1 used here displayed editing site preferences somewhat similar to those seen in ADARs from other species. Like the Xenopus enzyme (17), zebra fish ADAR1 exhibited a clear preference for 5' A/U residues. Further experimentation is still needed to determine whether dsRNA hyperediting, such as that detected here, leads to dsRNA destruction in fish cells as it does in mammalian cells (30).

Despite their marked dissimilarity at the protein level, it now seems that B2s from both the alpha- and betanodaviruses possess the singular function of binding to and sequestering dsRNA to protect viral replication intermediates from the cellular RNAi antiviral machinery. Though we have shown that its dsRNA binding ability enables B2 to interfere with other dsRNA-dependent processes, at this point, its only biologically important role from the viewpoint of facilitating viral RNA accumulation is the suppression of RNAi. An important question that remains unanswered, however, is whether the suppression of RNAi but not the interferon response by B2 is simply due to the temporal separation of these two antiviral mechanisms in fish, with RNAi being more important at the early stages of infection and the interferon response being important relatively later on. This is the subject of our current research.


arrow
ACKNOWLEDGMENTS
 
We thank M. Higuchi and P. H. Seeburg (Max Planck Institute for Medical Research, Germany) for the pRK-ADAR1Rn plasmid and P. Vitali and J. Cavaillé (Paul Sabatie University, France) for the Pol II/5-HT2C plasmid.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Animal Health Biotechnology, Temasek Life Sciences Laboratory, 1 Research Link, National University of Singapore, Singapore 117604. Phone: (65) 6872 7473. Fax: (65) 6872 7007. E-mail: kwang{at}tll.org.sg. Back


arrow
REFERENCES
 
    1
  1. Buchholz, U. J., S. Finke, and K. K. Conzelmann. 1999. Generation of bovine respiratory syncytial virus (BRSV) from cDNA: BRSV NS2 is not essential for virus replication in tissue culture, and the human RSV leader region acts as a functional BRSV genome promoter. J. Virol. 73:251-259.[Abstract/Free Full Text]
    2. 2
  2. Chao, J. A., J. H. Lee, B. R. Chapados, E. W. Debler, A. Schneemann, and J. R. Williamson. 2005. Dual modes of RNA-silencing suppression by flock house virus protein B2. Nat. Struct. Mol. Biol. 12:952-957.[Medline]
    3. 3
  3. Chen, Y. M., Y. L. Su, J. H. Lin, H. L. Yang, and T. Y. Chen. 2006. Cloning of an orange-spotted grouper (Epinephelus coioides) Mx cDNA and characterisation of its expression in response to nodavirus. Fish Shellfish Immunol. 20:58-71.[CrossRef][Medline]
    4. 4
  4. Chien, C. Y., Y. Xu, R. Xiao, J. M. Aramini, P. V. Sahasrabudhe, R. M. Krug, and G. T. Montelione. 2004. Biophysical characterization of the complex between double-stranded RNA and the N-terminal domain of the NS1 protein from influenza A virus: evidence for a novel RNA-binding mode. Biochemistry 43:1950-1962.[CrossRef][Medline]
    5. 5
  5. Fekete, R. A., and L. S. Frost. 2002. Characterizing the DNA contacts and cooperative binding of F plasmid TraM to its cognate sites at oriT. J. Biol. Chem. 277:16705-16711.[Abstract/Free Full Text]
    6. 6
  6. Fenner, B. J., R. Thiagarajan, H. K. Chua, and J. Kwang. 2006. Betanodavirus B2 is an RNA interference antagonist that facilitates intracellular viral RNA accumulation. J. Virol. 80:85-94.[Abstract/Free Full Text]
    7. 7
  7. Gallo, A., L. P. Keegan, G. M. Ring, and M. A. O'Connell. 2003. An ADAR that edits transcripts encoding ion channel subunits functions as a dimer. EMBO J. 22:3421-3430.[CrossRef][Medline]
    8. 8
  8. Gietz, R. D., and R. A. Woods. 2002. Transformation of yeast by lithium acetate/single-stranded carrier DNA/polyethylene glycol method. Methods Enzymol. 350:87-96.[CrossRef][Medline]
    9. 9
  9. Gillian, A. L., S. C. Schmechel, J. Livny, L. A. Schiff, and M. L. Nibert. 2000. Reovirus protein {sigma}NS binds in multiple copies to single-stranded RNA and shares properties with single-stranded DNA binding proteins. J. Virol. 74:5939-5948.[Abstract/Free Full Text]
    10. 10
  10. Guo, Y. X., S. W. Chan, and J. Kwang. 2004. Membrane association of greasy grouper nervous necrosis virus protein A and characterization of its mitochondrial localization targeting signal. J. Virol. 78:6498-6508.[Abstract/Free Full Text]
    11. 11
  11. Hakki, M., and A. P. Geballe. 2005. Double-stranded RNA binding by human cytomegalovirus pTRS1. J. Virol. 79:7311-7318.[Abstract/Free Full Text]
    12. 12
  12. Hegde, A., H. C. Teh, T. J. Lam, and Y. M. Sin. 2003. Nodavirus infection in freshwater ornamental fish, guppy, Poicelia reticulata—comparative characterization and pathogenicity studies. Arch. Virol. 148:575-586.[CrossRef][Medline]
    13. 13
  13. Ho, C. K., and S. Shuman. 1996. Physical and functional characterization of the double-stranded RNA binding protein encoded by the vaccinia virus E3 gene. Virology 217:272-284.[CrossRef][Medline]
    14. 14
  14. Iwamoto, T., K. Mise, A. Takeda, Y. Okinaka, K. Mori, M. Arimoto, T. Okuno, and T. Nakai. 2005. Characterization of Striped jack nervous necrosis virus subgenomic RNA3 and biological activities of its encoded protein B2. J. Gen. Virol. 86:2807-2816.[Abstract/Free Full Text]
    15. 15
  15. Johansen, A., B. Collet, E. Sandaker, C. J. Secombes, and J. B. Jorgensen. 2004. Quantification of Atlantic salmon type-I interferon using an Mx1 promoter reporter gene assay. Fish Shellfish Immunol. 16:173-184.[CrossRef][Medline]
    16. 16
  16. Knaust, R., S. Eshaghi, C. Dian, S. Lundgren, K. Anzai, P. Nordlund, and D. Birse. 2000. An efficient and rapid protein purification and on-column cleavage strategy using GSTrap columns. Life Sci. News 6:1-2.
    17. 17
  17. Lehmann, K. A., and B. L. Bass. 2000. Double-stranded RNA adenosine deaminases ADAR1 and ADAR2 have overlapping specificities. Biochemistry 39:12875-12884.[CrossRef][Medline]
    18. 18
  18. Leong, J. C., G. D. Trobridge, C. H. Kim, M. Johnson, and B. Simon. 1998. Interferon-inducible Mx proteins in fish. Immunol. Rev. 166:349-363.[CrossRef][Medline]
    19. 19
  19. Leulliot, N., S. Quevillon-Cheruel, M. Graille, H. van Tilbeurgh, T. C. Leeper, K. S. Godin, T. E. Edwards, S. T. Sigurdsson, N. Rozenkrants, R. J. Nagel, M. Ares, and G. Varani. 2004. A new alpha-helical extension promotes RNA binding by the dsRBD of Rnt1p RNAse III. EMBO J. 23:2468-2477.[CrossRef][Medline]
    20. 20
  20. Li, H., W. X. Li, and S. W. Ding. 2002. Induction and suppression of RNA silencing by an animal virus. Science 296:1319-1321.[Abstract/Free Full Text]
    21. 21
  21. Li, H. W., and S. W. Ding. 2005. Antiviral silencing in animals. FEBS Lett. 579:5965-5973.[CrossRef][Medline]
    22. 22
  22. Li, W. X., H. Li, R. Lu, F. Li, M. Dus, P. Atkinson, E. W. Brydon, K. L. Johnson, A. Garcia-Sastre, L. A. Ball, P. Palese, and S. W. Ding. 2004. Interferon antagonist proteins of influenza and vaccinia viruses are suppressors of RNA silencing. Proc. Natl. Acad. Sci. USA 101:1350-1355.[Abstract/Free Full Text]
    23. 23
  23. Lingel, A., B. Simon, E. Izaurralde, and M. Sattler. 2005. The structure of the flock house virus B2 protein, a viral suppressor of RNA interference, shows a novel mode of double-stranded RNA recognition. EMBO Rep. 6:1149-1155.[CrossRef][Medline]
    24. 24
  24. Liu, Y., K. C. Wolff, B. L. Jacobs, and C. E. Samuel. 2001. Vaccinia virus E3L interferon resistance protein inhibits the interferon-induced adenosine deaminase A-to-I editing activity. Virology 289:378-387.[CrossRef][Medline]
    25. 25
  25. Lu, R., M. Maduro, F. Li, H. W. Li, G. Broitman-Maduro, W. X. Li, and S. W. Ding. 2005. Animal virus replication and RNAi-mediated antiviral silencing in Caenorhabditis elegans. Nature 436:1040-1043.[CrossRef][Medline]
    26. 26
  26. Malmgaard, L. 2004. Induction and regulation of IFNs during viral infections. J. Interferon Cytokine Res. 24:439-454.[CrossRef][Medline]
    27. 27
  27. Munday, B. L., J. Kwang, and N. Moody. 2002. Betanodavirus infections of teleost fish: a review. J. Fish Dis. 25:127-142.
    28. 28
  28. Provost, P., D. Dishart, J. Doucet, D. Frendewey, B. Samuelsson, and O. Radmark. 2002. Ribonuclease activity and RNA binding of recombinant human Dicer. EMBO J. 21:5864-5874.[CrossRef][Medline]
    29. 29
  29. Qu, F., and T. J. Morris. 2005. Suppressors of RNA silencing encoded by plant viruses and their role in viral infections. FEBS Lett. 579:5958-5964.[CrossRef][Medline]
    30. 30
  30. Scadden, A. D., and M. A. O'Connell. 2005. Cleavage of dsRNAs hyper-edited by ADARs occurs at preferred editing sites. Nucleic Acids Res. 33:5954-5964.[Abstract/Free Full Text]
    31. 31
  31. Scadden, A. D., and C. W. Smith. 2001. Specific cleavage of hyper-edited dsRNAs. EMBO J. 20:4243-4252.[CrossRef][Medline]
    32. 32
  32. Schutz, S., and P. Sarnow. 2006. Interaction of viruses with the mammalian RNA interference pathway. Virology 344:151-157.[CrossRef][Medline]
    33. 33
  33. Sullivan, C. S., and D. Ganem. 2005. A virus-encoded inhibitor that blocks RNA interference in mammalian cells. J. Virol. 79:7371-7379.[Abstract/Free Full Text]
    34. 34
  34. Tafalla, C., R. Aranguren, C. J. Secombes, A. Figueras, and B. Novoa. 2004. Cloning and analysis of expression of a gilthead sea bream (Sparus aurata) Mx cDNA. Fish Shellfish Immunol. 16:11-24.[CrossRef][Medline]
    35. 35
  35. Taylor, D. R., M. Puig, M. E. Darnell, K. Mihalik, and S. M. Feinstone. 2005. New antiviral pathway that mediates hepatitis C virus replicon interferon sensitivity through ADAR1. J. Virol. 79:6291-6298.[Abstract/Free Full Text]
    36. 36
  36. Van Holde, K. E. 1985. Physical biochemistry, 2nd ed. Prentice-Hall, Englewood Cliffs, N.J.
    37. 37
  37. Vermeulen, A., L. Behlen, A. Reynolds, A. Wolfson, W. S. Marshall, J. Karpilow, and A. Khvorova. 2005. The contributions of dsRNA structure to Dicer specificity and efficiency. RNA 11:674-682.[Abstract/Free Full Text]
    38. 38
  38. Vitali, P., E. Basyuk, E. Le Meur, E. Bertrand, F. Muscatelli, J. Cavaille, and A. Huttenhofer. 2005. ADAR2-mediated editing of RNA substrates in the nucleolus is inhibited by C/D small nucleolar RNAs. J. Cell Biol. 169:745-753.[Abstract/Free Full Text]
    39. 39
  39. Wu, Y. C., and S. C. Chi. 9 March 2006, posting date. Persistence of betanodavirus in Barramundi brain (BB) cell line involves the induction of interferon response. Fish Shellfish Immunol. [Online.] doi:10.1016/j.fsi.2006.03.002.
    40. 40
  40. Xiang, Y., R. C. Condit, S. Vijaysri, B. Jacobs, B. R. Williams, and R. H. Silverman. 2002. Blockade of interferon induction and action by the E3L double-stranded RNA binding proteins of vaccinia virus. J. Virol. 76:5251-5259.[Abstract/Free Full Text]

Journal of Virology, July 2006, p. 6822-6833, Vol. 80, No. 14
0022-538X/06/$08.00+0     doi:10.1128/JVI.00079-06
Copyright © 2006, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Olveira, J. G., Souto, S., Dopazo, C. P., Thiery, R., Barja, J. L., Bandin, I. (2009). Comparative analysis of both genomic segments of betanodaviruses isolated from epizootic outbreaks in farmed fish species provides evidence for genetic reassortment. J. Gen. Virol. 90: 2940-2951 [Abstract] [Full Text]  
  • Fenner, B. J., Goh, W., Kwang, J. (2007). Dissection of Double-Stranded RNA Binding Protein B2 from Betanodavirus. J. Virol. 81: 5449-5459 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fenner, B. J.
Right arrow Articles by Kwang, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fenner, B. J.
Right arrow Articles by Kwang, J.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»