|Home | About | Journals | Submit | Contact Us | Français|
RNA silencing functions as an adaptive antiviral defense in both plants and animals. In turn, viruses commonly encode suppressors of RNA silencing, which enable them to mount productive infection. These inhibitor proteins may be exploited as reagents with which to probe mechanisms and functions of RNA silencing pathways. In this report, we describe transgenic Drosophila strains that allow inducible expression of the viral RNA silencing inhibitors Flock House virus-B2, Nodamura virus-B2, vaccinia virus-E3L, influenza A virus-NS1 and tombusvirus P19. Some of these, especially the B2 proteins, are effective transgenic inhibitors of double strand RNA-induced gene silencing in flies. On the other hand, none of them is effective against the Drosophila microRNA pathway. Their functional selectivity makes these viral silencing proteins useful reagents with which to study biological functions of the Drosophila RNA interference pathway.
Related systems for small RNA-based, post-transcriptional repression in diverse organisms have been collectively termed RNA silencing or RNA interference (RNAi)-related pathways. RNAi was originally coined to describe a phenomenon in nematodes whereby the injection of double-stranded RNA (dsRNA) corresponding to mRNAs of choice induced cognate loss-of-function phenotypes.1 It was subsequently appreciated that this powerful experimental technique taps into a host mechanism that recognizes and processes dsRNA into short RNAs, which guide an Argonaute-class protein to silence complementary transcripts. Argonaute proteins are found in plants, fungi, archaebacteria and animals, indicating that this pathway for RNA-based gene regulation is quite ancient.2
RNA silencing pathways act not only on artificial substrates for human convenience, but also mediate a universe of host regulatory interactions. Prime amongst these include the microRNA pathway, which processes endogenous hairpin precursor transcripts into ~21-24 nucleotide (nt) regulatory RNAs.3 microRNAs are pervasive components of post-transcriptional regulatory networks in both plants and animals, but structural and functional differences between plant and animal microRNAs suggest that microRNAs are evolutionary convergent aspects of RNA silencing in these kingdoms.4 RNAi-mediated gene regulation can also occur at the nuclear level. Best characterized in fission yeast, RNAi functions to establish and maintain heterochromatic domains.5
RNA silencing pathways also serve as forms of adaptive immunity that protect against invasive nucleic acids. Transposons are a major class of genomic intruder, and two RNA-based systems are known to combat them. Components of the conventional RNAi pathway limit transposon hopping in worms,6-8 while a specialized type of small RNA known as the piRNA protects against transposon mobilization in flies and vertebrates.9-12 Viruses comprise a second major class of genomic intruder. Early plant studies pointed to a post-transcriptional silencing mechanism in the induction of host recovery to viral infection13 that was related to transgene-induced gene silencing.14,15 More recently, it was explicitly shown that mutations in RNAi pathway components in plants,16,17 worms,18,19 and flies20-23 render individuals highly susceptible to viral invasion.
Reciprocally, viruses have fought back by evolving inhibitors of RNA silencing.24 Over 20 RNA silencing proteins have been identified from various plant viruses16,25,26 and animal viruses.27-30 Consistent with their presumably independent evolution by unrelated viruses, these inhibitors operate by a variety of mechanisms. Many of them, such as tombusvirus P19, beet yellow virus P21, and potyvirus HC-Pro, selectively bind small RNAs or small RNA duplexes.31-35 Vaccinia virus E3L binds long dsRNA,36 while influenza NS1 and nodaviral B2 proteins sequester both short and long dsRNA.28,30,32,37 The B2 proteins explicitly prevent Dicer from accessing and cleaving dsRNAs. CMV2b instead directly binds ARGONAUTE1 and inhibits its ability to cleave targets.38 Finally, polerovirus P0 encodes an F-box protein that actually targets ARGONAUTE1 for ubiquitylation and degradation.39,40
Because of their specifically-evolved activities, viral RNA silencing inhibitors can serve as effective probes to study mechanisms of RNA-mediated silencing, and as reagents with which to manipulate biological processes that are regulated by RNAi-related pathways. In this study, we created transgenic flies in which the expression of viral RNA silencing inhibitors could be inducibly activated using the Gal4-UAS system. These include Flock House virus-B2 (FHV-B2), Nodamura virus-B2 (NoV-B2), vaccinia virus-E3L (E3L), influenza A virus-NS1 (NS1A) and tombusvirus P19 (P19). In contrast to previous reports that characterized all of these as inhibitors of RNAi and/or miRNAs in cultured animal cells, their properties appeared to be more limited in transgenic Drosophila. In particular, only FHV-B2 and NoV-B2 functioned as strong suppressors of dsRNA-induced gene silencing, and none of them was effective against endogenous or exogenously expressed microRNAs. Their functional selectivity makes these transgenes useful for mechanistic and functional studies of RNAi-related pathways in Drosophila.
The open reading frames of FHV-B2, NoV-B2, E3L, P19 and NS1A were cloned into the TOPO-D-ENTR vector (Invitrogen). Following sequence verification, they were transferred into the recombination-competent destination vector UAS-HM-Gate, which adds N-terminal His and Myc tags. These were injected into w1118 hosts using standard helper transposase, and at least three independent insertions were recovered and analyzed for each construct. Other transgenes were described previously, including UAS-mir-7,46 tub-GFP,45 GMR-hid,53 UAS-IR-DIAP1.52 GMR-Gal4, ptc-Gal4, dpp-Gal4, and UAS-IR-GFP were obtained from the Bloomington Stock Center (http://flystocks.bio.indiana.edu).
Immunostaining was performed according to standard methods using rabbit a-GFP (Molecular Probes, 1:1250) and goat anti-rabbit-Alexa 488 (Molecular Probes, 1:600), and mounted in Vectashield (Vector Laboratories). Adult wings were mounted in Hoyer's mountant. For each of three transgene insertions, we examined at least five imaginal discs for each immunostaining and the eyes or wings of at least 25 animals from each cross; the data shown are representative phenotypes.
Many viral inhibitors of RNA silencing were earlier reported to be effective inhibitors of RNA silencing in Drosophila cell culture assays.28,29 We decided to create transgenic strains of these inhibitors, with the aim of using them as experimental reagents to manipulate RNA silencing pathways in the animal. We cloned five viral RNA silencing inhibitors-FHV-B2, NoV-B2, E3L, NS1A and P19-as N-terminal fusions to His-Myc (HM) tags in the pUAST vector for use in binary Gal4-UAS expression system.41 Since functional tagged versions of all of these RNA silencing inhibitors were described earlier,26,28,35,42,43 we did not anticipate that the HM tag would interfere with their activity. Several independent transgene insertions were isolated and tested for each construct.
We began by activating these transgenes with a variety of tissue-specific Gal4 drivers in settings known to require miRNA pathway function, such as the wing, eye and peripheral nervous system. Expression of viral RNA inhibitors in plants often induces mutant developmental phenotypes that are characteristic of miRNA pathway inhibition,16,26,44 and at least some of these (i.e., P19) function by selectively binding miRNA-sized small RNAs. Unexpectedly, we observed no major developmental defects upon misexpression of any of these viral proteins in a variety of locations. These included in the developing wing using ptc-Gal4 or bx-Gal4 (Fig. 1A-C and data not shown), or in the eye using GMR-Gal4 (Fig. 2A-G). With increased Gal4 or UAS dosage it was possible to produce minor defects in eye or wing morphology (data not shown). However, as high-level expression of Gal4 can also induce eye roughening and wing vein defects, it was not apparent that these mild phenotypes were specifically due to inhibition of the miRNA pathway.
We hypothesized that potentially subtle activities of these viral RNA silencing inhibitors might be revealed in a genetically sensitized background. Ectopic miRNAs frequently induce mutant phenotypes, presumably due to inappropriate downregulation of one or more target genes. Recombinant flies in which UAS-DsRed-miR-7 is activated by ptc-Gal4 exhibit notched wings and a mild proximal growth defect. These phenotypes are due, at least in part, to inhibition of Notch target gene activity by ectopic miR-7.45,46 Wing notching is a phenotype known to be extremely highly modifiable;47 thus, miR-7-induced notching represents a sensitive genetic setting with which to detect potential phenotypic suppression by viral inhibitor proteins. However, none of the viral RNA silencing inhibitors significantly altered the wing notching or growth defects induced by ectopic miR-7 (Fig. 1D-I). Immunostaining of larval imaginal discs with anti-Myc antibodies confirmed the accumulation of viral inhibitors, ruling out their instability as an explanation of their apparent inactivity (Fig. 3B'-F'). We conclude that these FHV-B2, NoV-B2, E3L, NS1A and P19 constructs are not able to substantially inhibit the miRNA pathway in intact Drosophila.
Drosophila has demonstrably segregated gene silencing by siRNAs and miRNAs into separate molecular pathways for biogenesis and function.48,49 Therefore, viral inhibitors might potentially exhibit selectivity for these pathways in flies. We tested whether these viral inhibitors of RNA silencing could inhibit RNAi by examining their ability to modify phenotypes induced by artificial long inverted repeat RNAs that contain segments of characterized mRNAs. Such foldback RNAs are capable of inducing specific loss-of-function phenotypes in the animal.50,51
The Drosophila Inhibitor of Apoptosis 1 (DIAP1) gene is required for the survival of Drosophila cells. Activation of a w to DIAP1 (UAS-IR-DIAP1) induces cell death characteristic of DIAP1 loss-of-function52 or the gain-of-function of DIAP1 inhibitors such as Hid.53 These genetic activities are easily assayed in the eye. Excess cell death induced by misexpression of either IR-DIAP1 or Hid under control of the eye-specific GMR enhancer results in a smaller, less-pigmented eye (Fig. 2A, H and O). These mutant phenotypes can be suppressed by coexpression of the viral inhibitor of apoptosis P35,54 demonstrating that these phenotypes are amenable to genetic modification (Fig. 2I and P).
Misexpression of any of the viral RNA silencing inhibitors using GMR-Gal4 yielded no obvious effect on eye patterning (Fig. 2C-G). However, FHV-B2 and NoV-B2 could suppress the small, rough eye induced by GMR > IR-DIAP1 nearly back to normal (Fig. 2J and K). Eye size and patterning were slightly improved when E3L was coexpressed (Fig. 2L). In contrast, neither P19 nor NS1A detectably suppressed ectopic IR-DIAP1 (Fig. 2M and N).
In principle, these viral RNA silencing inhibitors might be acting like viral P35 to block cell death, instead of RNA silencing per se. Therefore, we performed a secondary test of their ability to suppress the eye phenotype induced by GMR-hid. We found that P35 (Fig. 2O and P), but none of the viral inhibitors of RNA silencing (Fig. 2Q-U), suppressed the small, rough eye phenotype of GMR-hid. Therefore, the effects of FHV-B2, NoV-B2 and E3L were specific to the dsRNA-induced phenotype.
As an additional test of their ability to suppress the activity of dsRNA, we performed a functional silencing assay. An inverted repeat transgene against GFP (UAS-IR-GFP), when activated by dpp-Gal4, suppressed the expression of a ubiquitously expressed GFP transgene (tub-GFP) in a stripe at the anterior-posterior compartment boundary of the wing imaginal disc (Fig. 3A-A"). However, when FHV-B2 was coexpressed with IR-GFP, the accumulation of GFP in the dpp-Gal4 domain was restored (Fig. 3B). NoV-B2 was functional but less effective than FHV-B2 in this assay, with only one of three lines achieved a degree of suppression that was equivalent to typical FHV-B2 lines (Fig. 3C). Since all three NoV-B2 lines could strongly suppress IR-Diap1, it appears that suppression in the IR-GFP assay is a more stringent assay. Finally, E3L exerted only mild suppression, while P19 and NS1A had no effect on GFP expression (Fig. 3D-F). Overall, these IR-GFP data were therefore consistent with the IR-DIAP1 tests.
In summary, even though all five of the viral inhibitors of RNA silencing tested were reported to be effective inhibitors of RNAi in Drosophila cell culture,28 only the B2 proteins were particularly effective in transgenic animals. FHV-B2 binds nonspecifically to dsRNA18 and its crystal structure revealed a novel dsRNA interaction motif.32,37 Interestingly, the B2 proteins from different Nodaviridae are extremely divergent,55 indicating that their dsRNA binding motifs are quite tolerant to substitutions. Only one of the two arginine residues previously shown to be important for dsRNA binding by a piscine nervous necrosis virus B2 (NNV-B2)56 seemed to be conserved in FHV-B2 and NoV-B2 (Fig. 4). Therefore, the strong cross-species activities of B2 proteins were not a trivial consequence of expressing proteins with related primary sequences.
The plant tombusvirus P19 protein was earlier characterized as a strong suppressor of RNA-mediated silencing in human cells.26,27 Such cross-kingdom activity was consistent with its biochemical ability to specifically sequester 21nt siRNAs,31,33 whose size is the same in plants and animals. However, P19 was more recently reported to be ineffective in antagonizing siRNA-mediated knockdowns in human cells.42 In addition, even though the influenza NS1A inhibitor of viral defense has dsRNA-binding activity, it is fairly weak and in fact dispensable for its ability to antagonize activation of PKR by dsRNA during the mammalian antiviral response57 Finally, while the dsRNA binding protein E3L of vaccinia virus was previously reported as a functional inhibitor of RNA silencing,28 it was later reported not to inhibit shRNA-induced silencing in a human cell system.58 Our negative data with respect to P19 and NS1A, and the relatively mild activities of E3L with respect to RNAi, are a further indication that some viral inhibitors of RNA silencing may not be as generally effective as earlier suggested.
On the other hand, we showed using multiple assays that different B2 proteins are potent suppressors of dsRNA-induced gene silencing in transgenic flies. Our findings were consistent with other reports that different B2 proteins could antagonize the RNAi pathway in diverse cultured cell systems.29,30,59 Therefore, these will be useful probes of endogenous RNAi-dependent processes in Drosophila. Indeed, flies that ubiquitously express FHV-B2 were found to exhibit enhanced susceptibility to infection by Drosophila C virus.21 We propose that analysis of tissue-directed expression of B2 proteins might reveal spatially sensitive aspects of viral defense. In addition, RNAi-related mechanisms have been implicated in Drosophila Polycomb response element function,60 nucleolar organization,61 and possibly heterochromatic silencing62 and chromosomal insulator function.63 However, the evidence for the involvement of RNAi in these processes has been mostly indirect. Perhaps the most intriguing application of our transgenes would be as affinity probes to identify endogenous dsRNAs that immunoprecipitate with tagged FHV-B2 or NoV-B2. Knowledge of such molecules would provide a molecular handle on endogenous RNAi-mediated processes in Drosophila that could provide direction to future biological investigations.
We thank Shou-Wei Ding for cDNA clones of the viral inhibitors of RNA silencing used in this study, and Brian McCabe for UAS-HM-Gate. Stephen Cohen provided the tub-GFP transgene. This work was supported by grants from the Leukemia and Lymphoma Foundation, the Burroughs Wellcome Foundation, the V Foundation for Cancer Research, the Sidney Kimmel Foundation for Cancer Research and the National Institutes of Health (GM083300).