© 2008 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES

The DExD/H-box helicase Dicer-2 mediates the induction of antiviral activity in drosophila Safia Deddouche1, Nicolas Matt1, Aidan Budd2, Stefanie Mueller1, Cordula Kemp1, Delphine Galiana-Arnoux1,4, Catherine Dostert1,4, Christophe Antoniewski3, Jules A Hoffmann1 & Jean-Luc Imler1 Drosophila, like other invertebrates and plants, relies mainly on RNA interference for its defense against viruses. In flies, viral infection also triggers the expression of many genes. One of the genes induced, Vago, encodes a 18-kilodalton cysteine-rich polypeptide. Here we provide genetic evidence that the Vago gene product controlled viral load in the fat body after infection with drosophila C virus. Induction of Vago was dependent on the helicase Dicer-2. Dicer-2 belongs to the same DExD/H-box helicase family as do the RIG-I–like receptors, which sense viral infection and mediate interferon induction in mammals. We propose that this family represents an evolutionary conserved set of sensors that detect viral nucleic acids and direct antiviral responses.

Viruses represent a serious threat to all living organisms. They represent the most abundant class of pathogens on earth. As obligate intracellular parasites, viruses interact intimately with their host cells, manipulating cellular metabolism to convert cells into virusproducing ‘factories’ and triggering dysfunction and/or damage that can lead to cell death. In response, multicellular organisms have evolved efficient host-defense mechanisms to sense viruses and block their replication and spread1,2. In both plants and invertebrates, RNA interference is an important antiviral defense mechanism in which double-stranded RNA (dsRNA) is detected by RNase III enzymes of the Dicer family and is cleaved into small interfering RNAs 21–23 nucleotides (nt) in length. These are incorporated in the RNA-induced silencing complex, which contains a member of the Argonaute (AGO) family. The ‘guide strand’ of the small interfering RNA directs the RNase H enzyme AGO to complementary sequences, which allows highly specific cleavage of viral RNA molecules2. In mammals, in contrast, the main response to viral infection is the production of type I and type III interferon. Expression of the genes encoding these cytokines is dependent on receptors that sense viral nucleic acids. These receptors belong to two families, the transmembrane Toll-like receptors, which detect viral RNA or DNA in endosomal compartments, and the cytoplasmic sensors RIG-I, Mda5 and LGP2, which form the small RIG-I-like receptor (RLR) family and interact with viral RNA through a carboxy-terminal DExD/H-box RNA helicase domain1.

The fruit fly Drosophila melanogaster is a powerful model for studying innate immunity. Studies have shown that RNA interference and Dicer-2 (Dcr-2), r2d2 and AGO2 are critical in the control of RNA virus infection in flies3–6. In addition to this intrinsic response, infection with drosophila C virus (DCV), a natural pathogen of drosophila that belongs to the dicistroviridae family, has been shown to trigger expression of some 150 genes7. Several of these genes, including the marker vir-1 (virus-induced RNA 1), are regulated by the pathway of the kinase Jak and the STAT transcription factor. The data available suggest that DCV infection triggers the induction of an unidentified cytokine that activates the domeless receptor and the Jak kinase hopscotch, which leads to activation of the drosophila STAT factor and gene induction. The hopscotch mutant flies have higher viral loads than do wild-type control flies, which suggests that some of the genes induced encode antiviral molecules. Notably, this pathway is not sufficient to trigger the antiviral response, and some genes are upregulated after DCV infection in a Jak-STAT–independent way7. Two main issues in this field are the identification of the antiviral molecules induced in infected flies and of the receptors that detect viral infection. Here we report that the product of Vago, which was induced by viral infection, participated in the control of the viral load in the fat body. Furthermore, induction of this gene was dependent on Dicer-2, which indicates that Dicer-2 is one of the sensors that trigger the antiviral inducible response in flies. We also discuss our findings in the context

1Unite ´ Propre de Recherche´ 9022, Centre National de la Recherche Scientifique, Institut de Biologie Mole´culaire et Cellulaire, 67084 Strasbourg, France. 2European Molecular Biology Laboratory, 69117 Heidelberg, Germany. 3Department of Developmental Biology, Centre National de la Recherche Scientifique URA2578, Institut Pasteur, 75724 Paris, France. 4Present addresses: Institute of Functional Genomics of Lyon–Unite´s Mixtes de Recherche 5242, Ecole Normale Supe´rieure de Lyon, 69364 Lyon Cedex 07, France (D.G.-A.) and Department of Biochemistry, University of Lausanne, Chemin des Boveresses 155, 1066 Epalinges, Switzerland (C.D.). Correspondence should be addressed to J.-L.I. ([email protected]).

Received 27 May; accepted 9 September; published online 26 October 2008; doi:10.1038/ni.1664

NATURE IMMUNOLOGY

ADVANCE ONLINE PUBLICATION

1

ARTICLES

150

c

P = 0.0137 P = 0.0196

200

UAS-Vago Hsp-Gal4

Vago∆M11

P = 0.010

+

–

+

–

– +

+ +

– +

+ Tris

Anti-actin 50 P < 0.0001

DCV

P = 0.0023

+ FB

T

10

10

WF-HL

Va go ∆M

W

W

T Va go ∆M

10

HL

T Va go ∆M W 10 T Va go ∆M

WF

W

–

DCV

Anti-VP2

VP2 expression (%)

500 48 72 Time after infection (h)

DCV (2 d)

400 300 200

0 UAS-Vago Hsp-Gal4

– +

+ + WT

Anti-actin

– +

+ + Vago∆M10

P < 0.0001

P < 0.0001

55 50 45 40 35 30 25 20 15 10 5 0 UAS-Vago Hsp-Gal4

100 + + Vago∆M11

– HS

+ – + – + Figure 1 Vago regulates viral load in the fat body of DCV-infected flies. (a) Accumulation of + + + + + viral RNA in wild-type flies (WT) and Vago-mutant flies (VagoDM10 and VagoDM11), assessed ∆M10 ∆M11 Vago Vago WT by RNA blot hybridization and presented relative to maximum expression in wild-type flies. – HS (b) Immunoblot analysis of the accumulation of the DCV capsid protein VP2 in wild-type and Vago mutant flies, assessed without infection (–) or 3 d after DCV infection (+) in whole flies (WF), hemolymph (HL), whole flies after hemolymph was obtained (WF-HL) or dissected fat body (FB). Actin serves as a loading control. (c) VP2 protein in the fat bodies of DCV-infected wild-type and Vago-mutant flies after expression of a Vago transgene with the UAS-Gal4 system (UAS-Vago) and a heat-shock protein–Gal4 ‘driver’ (Hsp-Gal4; Supplementary Fig. 3a), assessed by immunoblot analysis before and after heat-shock treatment (HS; top) and quantified by densitometry before heat-shock treatment (–HS; bottom), with similar results obtained after heat-shock treatment (Supplementary Fig. 3b). (d) DCV load in the fat bodies of Vago-mutant flies, detected by immunofluorescence with anti-VP2 (green; Alexa Fluor 488) and visualized by confocal microscopy 2 d after injection of Tris or DCV. Scale bar, 100 mm. Below, groups of DCV-infected cells per fat body. Each symbol represent an individual fly; small horizontal bars indicate the mean. Data are representative of three (a,b), four (c) or two (d) independent experiments (average and s.d., a,c) with 20 (a) or 10 (b,c) flies per experiment or six to eight flies of each genotype (d, below).

of the phylogenetic relationship between the DExD/H-box helicase domains of Dicer and RIG-I-like helicases in mammals.

identified in the Genexel collection a drosophila stock containing a P element inserted 55 base pairs upstream of the Vago coding sequences, and we generated two mutant lines by imprecise excision of the transposon. The line VagoDM11 has a deletion of about 2.2 kilobases that removes Vago as well as the 5¢ end of the adjacent gene CG2076, whereas the line VagoDM10 has a deletion of about 1.3 kilobases in the proximal promoter of Vago (Supplementary Fig. 1b). We were unable

b 100 80

2

0 Time after infection (h)

6 24 48 6 24 486 24 48 6 24 486 24 48 6 24 48 6 24 48

UI Tris

DCV E.cl

E.c

M.l

E.f

40

0 Time after infection (h)

6 24 48 6 24 48 6 24 48 6 24 486 24 48 6 24 48 6 24 48

UI Tris

B.b

c 100

d

80

DCV E.cl

E.c

M.l

E.f

B.b

250

200

60

40

20

Time after 0 infection (h)

60

20

Vago mRNA (%)

Figure 2 Virus-specific transcriptional induction of Vago in wild-type flies. (a–c) Quantitative RT-PCR of the expression of Vago (a), Drosomycin (Drs; b) and Diptericin (Dpt; c) in flies left uninfected (UI) or infected with DCV, the Gramnegative bacteria Enterobacter cloacae (E.cl) or Escherichia coli (E.c), the Gram-positive bacteria Micrococcus luteus (M.l) or Enterococcus fecalis (E.f) or the fungus Beauveria bassiana (B.b), normalized to RpL32 expression and presented as the percentage of maximum expression. (d) Quantitative RT-PCR of Vago expression in virus-infected flies. Data are representative of four independent experiments with ten flies in each (mean and s.d.).

Drs mRNA (%)

Vago mRNA (%)

RESULTS Vago controls the viral load in infected fat body cells Vago (CG2081) was identified as an upregulated RNA by microarray analysis of DCV-infected flies7. This gene, located at position 10A4 on the X chromosome, attracted our attention because it remains fully inducible in hopa 100 scotch mutant flies, which suggests that it may be induced directly after viruses are sensed7. 80 It encodes a 160–amino acid protein with a 60 signal peptide and eight cysteine residues forming a conserved CX20CX4CX10–11CX7–9 40 CX13–14CCX4C motif (where ‘X’ is any amino acid)8,9 (Supplementary Fig. 1a online). We 20

Dpt mRNA (%)

© 2008 Nature Publishing Group http://www.nature.com/natureimmunology

–

Vago∆M10

WT

UAS-Vago Hsp-Gal4

+ +

– +

Anti-VP2

0

b

– +

HS

P = 0.0004 100

d

Vago∆M11

WT

WT Vago∆M10

Infected cell patches

DCV mRNA (%)

a

150

100

50

6 24 486 24 486 24 486 24 486 24 486 24 486 24 48

UI Tris

DCV E.cl

E.c

M.l

E.f

B.b

Time after 0 1 2 3 4 6 1 2 3 1 2 3 1 2 3 4 6 infection (d) Tris DCV FHV SINV

ADVANCE ONLINE PUBLICATION

NATURE IMMUNOLOGY

ARTICLES Vago

a

b

+1 ATG

–1979

lacZ

UI

DCV

DCV

FHV

poly(A) Drs

© 2008 Nature Publishing Group http://www.nature.com/natureimmunology

β-gal activity (arbitrary units)

8

c

7

Tris DCV FHV SINV Bacterial mix

6 5 4 3 2 1 0 Challenge

Tris

DCV

O Anti–viral antigen

M αVP2

αVP2

αVP2 O

Anti-β-gal

M

αCP

Figure 3 Fat body–specific transcriptional induction of Vago. (a) Fluorimetric assay of b-galactosidase (b-gal) activity in Vago::lacZtransgenic flies challenged with DCV, FHV, SINV or a mixture of Gram-positive and Gram-negative bacteria (M. luteus and E. coli ), assessed 72 h later in single flies. Above, transgene construct. (b) Expression of the Vago::lacZ reporter transgene in the fat bodies of adult transgenic flies injected with Tris (uninfected) or DCV and stained 72 h later with X-Gal (blue) to detect b-galactosidase. Bottom row, dissected fat bodies. (c) Confocal microscopy of the colocalization of b-galactosidase and DCV in the fat bodies of infected Vago::lacZ flies double-labeled with antib-galactosidase (Anti-b-gal; green; Alexa Fluor 488) and antibody to viral antigen (VP2 (aVP2) or capsid protein (aCP)); red; Alexa Fluor 546). Bottom row (RGB), superposition of red and green signals along the axis in the merged panel above. Arrowhead (and T, below), trachea; O, oenocyte; M, muscle fiber. Scale bars, 200 mm (b) and 10 mm (c). Data are from one experiment representative of three (mean and s.d. of five flies; a) or are representative of at least five experiments with five to ten flies each (b,c).

O

Merge

M

T

T

RGB superposition

to detect Vago mRNA in either line (Supplementary Fig. 1c). The two mutant lines with deletion of Vago were homozygous viable and fertile, which indicates that the product of this gene is not involved in any essential developmental functions. We next infected these flies with DCV and found that Vago-mutant flies had more viral RNA than did the wild-type control flies (Fig. 1a). DCV replicates in a subset of drosophila tissues, including the fat body and the periovarian sheath10,11. We found more viral RNA in dissected fat bodies from Vago mutant flies than in those from wild-type flies but not in dissected ovaries (Supplementary Fig. 2 online). We next assessed the quantity of viral protein in infected flies. We did not detect substantial differences in the amount of the capsid protein VP2 in whole flies or hemolymph extracts of Vago-mutant and wild-type flies. However, protein extracts of fat bodies from Vago-mutant flies contained more viral protein than did wild-type controls (Fig. 1b). Expression of a Vago transgene in VagoDM10 and VagoDM11 mutant flies with a system consisting of an upstream activating sequence (UAS) recognized by the yeast transcriptional activator Gal4 (UASGal4)12 restored DCV VP2 protein abundance in the fat body to that in wild-type flies (Fig. 1c and Supplementary Fig. 3 online), which confirmed that the phenotype could be attributed to the loss of Vago expression. Immunofluorescence studies confirmed that in the absence of Vago, the viral load was higher the fat body (Fig. 1d) but not in the periovarian sheath (Supplementary Fig. 2c). Vago-mutant flies did not succumb more rapidly than wild-type flies to DCV

NATURE IMMUNOLOGY

ADVANCE ONLINE PUBLICATION

infection, possibly because the effect of the mutation was restricted to the fat body. These in vivo data collectively indicate that Vago is needed, directly or indirectly, to control the accumulation of viral RNA and protein in the fat body. Tissue-specific induction of Vago We next addressed the regulation of Vago. Vago was only weakly upregulated in response to bacterial challenge (Fig. 2a–c) but was strongly induced after infection with DCV. Vago was also induced after infection by the alphavirus Sindbis virus (SINV). Notably, a third plus-strand RNA virus, the nodavirus flock house virus (FHV), did not induce Vago expression (Fig. 2d). This difference may reflect different tissue tropism of the three viruses. To determine in which tissues Vago is induced, we constructed transgenic lines expressing the reporter gene lacZ (encoding b-galactosidase) under the control of a fragment of Vago upstream sequences about 2 kilobases in length (‘Vago::lacZ’ reporter) and noted upregulation of b-galactosidase reporter activity after infection by DCV and SINV but not after infection by bacteria or FHV (Fig. 3a). This expression pattern was similar to that of the endogenous gene. The Vago::lacZ reporter was induced specifically in the fat bodies of flies infected with DCV or SINV, which was in agreement with the tissue-autonomous function of the gene (Fig. 3b and data not shown). We detected no bgalactosidase in other tissues targeted by DCV, such as the periovarian sheath in female flies or the tracheae (data not shown and Fig. 3c). Confocal microscopy showed colocalization of the b-galactosidase reporter and DCV in fat body cells but not in tracheae or oenocytes (Fig. 3c). Double in situ hybridization confirmed that Vago mRNA expression was induced in DCV-infected cells (Supplementary Fig. 4 online). These data collectively indicate that Vago is induced in fat body cells after sensing of DCV infection. Notably, FHV did not induce Vago expression, although this virus can infect and replicate in the fat body3 (Fig. 3c).

3

ARTICLES

Day 2

Day 4 or 5

vir-1

– –

+ –

5 + +

– –

+ –

+ +

200

100

100

100 50

80 60 40 20

50 0 DCV FHV

– –

+ –

+ +

– –

4

+ –

+ +

0 DCV UAS-B2

0 +

–

yw

+

–

Dcr-2

R416X

Dcr-2L811FSX

+

–

5 Time (d)

250

Tub-Gal4

e

DCV Tris

vir-1 mRNA (%)

200 Figure 4 Dicer-2 mediates the induction of Vago expression in DCVP < 0.0001 infected flies. (a,b) RNA blot analysis (a, bottom) and quantification (b) 100 150 of the expression of Vago and vir-1 in flies infected with FHV or Tris and 80 challenged with DCV or Tris (protocol, a, top). (c) Quantification of RNA 100 blot analysis of the expression of Vago and vir-1 at 3 d after DCV 60 infection of flies obtained by crossing flies either wild type or transgenic 40 50 for UAS-B2 with flies with a tubulin-Gal4 driver (Tub-Gal4). (d) Quantification of RNA blot analysis of the expression of Vago (top) 20 0 and vir-1 (bottom) in wild-type flies (yw) and Dcr-2-null flies (Dcr-2R416X 0 yw Dcr-2R416X Dcr-2L811FSX and Dcr-2L811FXS) at 3 d after DCV infection. (e) Quantification of RNA Dcr-2A500V Dcr-2L811FSX yw blot analysis of Vago expression after DCV infection of wild-type flies (yw), Dcr-2-null flies (Dcr-2L811FXS) and flies with a missense mutation in sequence encoding the helicase domain of Dicer-2 (Dcr-2A500V). Expression (b–e) is normalized to RpL32 expression and is presented as the percentage of maximum expression in control flies. Data are representative of three (a,b,d,e) or two (c) independent experiments with 20 flies each (mean and s.d., b–e). Vago mRNA (%)

Figure 5 Uncoupling RNA interference and Vago induction. (a) Induction of Vago in wild-type flies (yw), AGO2-mutant flies (AGO2414) and r2d2-mutant flies (r2d21) infected with DCV. (b) RNA blot analysis of Vago RNA in control flies (yolk-Gal4) and flies overexpressing in the fat body an inverted repeat construct targeting mRNA for the drosophila sulfonylurea receptor (yolkGal4::UAS-surIR)). (c) Quantitative RT-PCR analysis of the induction of Vago RNA in whole wild-type flies with an hsp-Gal4 driver (1) and flies overexpressing an inverted repeat construct targeting Drosomycin mRNA with an hsp-Gal4 driver (2,3), with and without DCV infection and before (2) and after (1,3) heat shock. Expression (a,c) is normalized to RpL32 expression and is presented as the percentage of maximum expression in control flies. Data are representative of three independent experiments (a,c) or one independent experiment (b; three samples per genotype) with 20 flies (a,b) or 10 flies (c) in each.

4

DCV-infected flies with a missense mutation in this domain16 and found that the DExD/H-box helicase domain of Dicer-2 was required for Vago induction (Fig. 4e). In contrast, we noted normal induction of Vago after DCV infection of AGO2-null mutant flies or r2d2-null mutant flies17,18 (Fig. 5a). These data indicate that the other components of the RNAinterference pathway are not involved in the inducible response. Furthermore, we found that expression of dsRNA in fat body cells with hairpin constructs shown before to efficiently induce silencing19,20 was not sufficient to induce Vago expression (Fig. 5b,c). The results reported above suggested that in addition to its involvement in RNA interference, Dicer-2 senses viral dsRNA in

P = 0.25 P = 0.20

b -s

ur

100

IR

a

yo l

60 40

4: al G yo l

k-

k-

G

al

4

80

:U AS

DCV Tris

Vago

20 RpL32 0 r2d21

yw

AGO2414

c 120 Vago mRNA (%)

Dicer-2 mediates Vago induction in virus-infected cells We reasoned that FHV might suppress Vago induction; to test our hypothesis, we preinfected flies with FHV and challenged them 2 d later with DCV. Preinfection with FHV led to more induction of vir-1 by DCV infection. In contrast, upregulation of Vago by DCV was much lower in flies preinfected with FHV (Fig. 4a,b). FHV is a very simple virus that encodes, in addition to the RNA-dependent RNA polymerase and the capsid protein, the dsRNA-binding protein B2 (refs. 13,14). B2 has been shown to bind to both long dsRNA and small interfering RNA duplexes and to function as a viral suppressor of RNA interference. We expressed B2 protein in transgenic flies and consistently found less upregulation of Vago expression. In contrast, B2 did not suppress the induction of vir-1 in DCV infected flies (Fig. 4c). The interaction of B2 with dsRNA interferes with Dicer-2 function in infected cells13. We therefore analyzed Vago expression in Dcr-2-null flies and found much less Vago upregulation. However, vir-1 remained fully inducible in Dcr-2-mutant flies (Fig. 4d). This result suggests that Dicer-2 is involved in the sensing of viral infection in drosophila. This enzyme can interact with viral RNA through several domains, including a DExD/H-box helicase domain15. We analyzed Vago expression in

Vago mRNA (%)

© 2008 Nature Publishing Group http://www.nature.com/natureimmunology

P < 0.0001

150 P < 0.0001

DCV Tris

P = 0.0054

120

P = 0.0252

150

Vago RpL32

d

200

250 mRNA (%)

DCV FHV

4

Vago vir-1

c

300

FHV or Tris DCV or Tris RNA injection challenge extraction Time (d)

vir-1 Vago

b

Vago mRNA (%)

Day 0

mRNA (%)

a

DCV Tris

100 80 60 40 20

0 DCV HS:

– 1 +

+

–

+

–

2 –

ADVANCE ONLINE PUBLICATION

+ 3 +

NATURE IMMUNOLOGY

ARTICLES sequence to the drosophila genome on the National Center for Biotechnology Information database with default parameters. We YRF1-5 YRF1-6 analyzed in more detail the similarities CG7922 spn-E FANCM among all 389 proteins containing a single dhx36 MPH1 mag1 DExD/H-box helicase domain in the annotated proteomes of Saccharomyces cerevisiae, SKIV2 polq-1 Caenorhabditis elegans, D. melanogaster and tst Homo sapiens. The resulting dendrogram nsh1 showed that the Dicer enzymes and the brm ERCC3 RLRs share a unique common internal branch IRC3 (Fig. 6). This group included only the four CG14443 human sequences RIG-I, Mda5, LGP-2 and DDX1 EIF4A1 hDicer, the two drosophila Dicer proteins hel1 glh4 (Dicer-1 and Dicer-2), and the nematode DDX50 single Dicer protein together with the three nematode Dicer-related helicases (DRH1– DRH3). If this unique common branch is assumed to be a feature of the true phylogeny RM62 APOLD1 EIF4A1 pit of the sequences, the sequence data suggest hel1 DDX24 that Dicer proteins and RLRs belong to the Figure 6 Protein sequence similarity–based dendrogram of DExD/H-box helicases. PhyML analysis of all same family of helicases that are essential in sensing viral infection in multicellular organ389 S. cerevisiae, C. elegans, D. melanogaster and H. sapiens proteins that contain a single DExD/Hisms. In this case, a parsimonious conclusion box helicase domain. Colors indicate different clades; selected proteins included in the clade are in boxes (complete list, Supplementary Table 1 online). Top right (red), Dicer-RLR clade. Scale bar, would be that this function is ancestral for the estimated amino acid substitutions per site. Phylogeny obtained from TreeFam for the Dicer-RLR group. We obtained a more detailed estimate helicases is in Supplementary Fig. 7. of the phylogenetic relationships for the Dicer–RIG-I-like helicases from the TreeFam infected cells and triggers a signaling pathway that leads to gene database. This tree contained one long internal branch that separated expression. Induction of Vago was not affected in flies mutant for the the RIG-I-like vertebrate sequences and the C. elegans Dicer-related genes encoding DmMyD88 and Imd, two essential death domain helicases (DRH1–DRH3) from the Dicer enzymes in plants, inverteadapters functioning in the Toll and Imd (immune deficiency) path- brates and vertebrates (Supplementary Fig. 7 online). The tree also ways, respectively21,22 (Supplementary Fig. 5a,b online). We noted showed that RIG-I-like helicases are absent from insect genomes, the presence of putative binding sites for transcription factors of the which instead have a second Dicer gene (Dcr-2). In mammals, the NF-kB family in the promoter of Vago (Supplementary Fig. 6a RIG-I-like helicases have diverged into three paralogous groups. online). The composition of these motifs was more reminiscent of Notably, only two of those (corresponding to Mda5 and LGP2) have the Toll-responsive kB sites recognized by Dif or dorsal (three G representatives in teleosts, which seem to lack the RIG-I–DDX58 residues separated from C residues by four to five A or T residues) paralogous group; this suggests that it appeared after separation of than of the Imd-responsive kB sites recognized by the third drosophila the tetrapod lineage from that of the teleosts. A final observation is NF-kB protein Relish (four G residues separated from C residues by that the rate of evolution (as indicated by branch lengths) seems to be two to three A or T residues)23. Induction of Vago expression by DCV much higher for the RIG-I-like group in vertebrates and the Dicer-2 was normal in Dif- and dorsal-mutant flies, and Vago was not paralogous group in insects than for the other Dicer groups24 constitutively expressed in flies mutant for cactus, which encodes the (Supplementary Fig. 7), which might reflect the selective pressure drosophila ortholog of inhibitor-of-kB proteins (Supplementary exerted by the rapidly changing RNA viruses on these genes. Fig. 5c–e). Relish-mutant flies were unexpectedly resistant to infection with DCV (or SINV), so we were unable to directly assess the function DISCUSSION of this transcription factor in the induction of Vago (data not shown). Our data have shown that in addition to RNA interference, an However, promoter-deletion experiments showed that the region inducible response is involved in the control of viral infection in required for Vago induction did not overlap with the putative drosophila. We have further identified Vago as an important particiNF-kB-binding sites (Supplementary Fig. 6b). Thus, induction of pant in the control of the viral load in the fat body of drosophila. Vago does not seem to be mediated by members of the NF-kB family Experiments addressing the function and structure of the recombinant of transcription factors. protein are needed to clarify the exact function of Vago, which could act either as an antiviral molecule targeting the virions9 or as a Phylogeny of Dicer- and RIG-I-like helicases cytokine triggering an antiviral state in neighboring cells8. UnfortuOur data showed that flies use the DExD/H-box helicase Dicer-2 to nately, our efforts to characterize the function of a recombinant sense the presence of viral RNA in infected cells and trigger the version of the Vago protein were hampered by poor stability of the induction of a gene associated with antiviral activity. This situation is tagged molecule. Thus, the precise function of Vago can only be reminiscent of the sensing of viral RNA by the RLR helicases RIG-I, speculated on at this stage. It has been reported that vir-1 (as well as Mda5 and LGP2 in mammals. Notably, the DExD/H-box helicase several other DCV-induced genes) is regulated by the Jak kinase domains of RIG-I and Dicer enzymes are closely related, and Dicer-2 is hopscotch and the cytokine receptor domeless7. The finding that the first result obtained after comparison of the RIG-I protein vir-1 remained fully inducible in Vago-mutant flies (data not drh1 RIG-I Dcr-2

© 2008 Nature Publishing Group http://www.nature.com/natureimmunology

0.1

NATURE IMMUNOLOGY

ADVANCE ONLINE PUBLICATION

5

© 2008 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES shown) indicates that Vago was not involved in this cytokine-mediated response. The fat body–specific requirement for Vago expression suggests a tissue-autonomous function and also challenges the idea of a circulating cytokine function for Vago. Notably, the lack of conservation of this gene in other insect species and the sequence divergence in drosophilidae is a characteristic of host-defense effector molecules, which has also been reported for some antimicrobial peptides25. In mammals, some defensins (a family of cysteine-rich antimicrobial molecules) are induced by viral infection and have antiviral functions, either through direct effect on the virion or through interactions with target cells26. We found that Vago had some sequence similarity to granularin, a molluscan defense peptide induced after parasitization that acts as an opsonin27. The identification of an induced gene encoding a molecule associated with antiviral activity raises the issue of the receptors that detect the viral infection and activate the antiviral response. In mammals, even though viral proteins can contribute to the induction of interferon synthesis28, the innate immune system senses mainly viral nucleic acids1. The finding that the FHV suppressor of RNA interference B2 interfered with the induction of Vago provides evidence that viral dsRNA is one of the inducers of the antiviral response in drosophila. Indeed, B2 binds with nanomolar affinity to dsRNA in a sequence- and size-independent way but not to single-stranded RNA or DNA13. Notably, DCV also encodes a suppressor of RNA interference (DCV-1A) that binds to long dsRNA4. The finding that DCV, unlike FHV, induced Vago expression indicates the existence of differences in the mode of interaction with dsRNA for B2 and DCV-1A. Our data have also shown that the mere presence of dsRNA in drosophila cells, even when highly expressed with the UAS-Gal4 system, was not sufficient to induce Vago expression and that the induction of this gene required features specific to viral dsRNA. Our genetic data indicated that Dicer-2, which is known to interact with dsRNA29, is the sensor that activates Vago expression. The finding that neither AGO2 nor r2d2 was required for Vago induction uncouples this sensing function of Dicer-2 from its function in RNA silencing. In addition, the results obtained with r2d2-mutant flies allowed us to exclude the possibility that r2d2, which is downregulated after Dicer-2 knockdown17, contributes to the regulation of Vago expression. Our results establish a parallel between the functions of RLRs in mammals and Dicer-2 in drosophila. Notably, we found that the members of the RIG-like receptor and Dicer-like helicase families are phylogenetically related. Further studies are needed to decipher the signaling pathway leading from Dicer-2 activation to induction of Vago expression. The inducible expression of Vago did not involve the three main pathways regulating innate immunity in drosophila (Toll, Imd and Jak-STAT). The signaling ‘downstream’ of Dicer-2 probably differs from that of the RLR pathway, as Dicer-2 does not contain caspase-recruitment domains and the drosophila genome does not contain orthologs of the signal transducer IPS-1 (also called MAVS, VISA and CARDIF) or the IRF transcription factors1. In summary, our data have shown that Dicer-2 has a dual function in drosophila after sensing viral RNA in infected cells. In addition to its involvement in RNA interference, this RLR-like helicase also regulates the induction of molecules controlling viral load in some tissues. Thus, an evolutionary conserved set of DExD/H-box helicases directs antiviral responses in drosophila and mammals. Unlike drosophila, the nematode C. elegans encodes three Dicer-related helicases (DRH1–DRH3), which are orthologous to the mammalian RLRs. Notably, DRH1 and DRH3 associate with Dicer and are required for RNA interference30,31. The data obtained with C. elegans

6

and drosophila collectively indicate an unexpected connection between RNA interference and inducible antiviral defenses, which may point to additional functions for RLRs in mammals. METHODS Fly strains and infections. OregonR and yw flies were used as wild-type flies, except for experiments with Vago-mutant flies (discussed below). The stocks Dcr-2L811FSX, Dcr-2R416X, Dcr-2A500V (ref. 16), AGO2414 (ref. 18), r2d21 (ref.32), UAS-Drom-IR20, UAS-dSUR-IR19 and yolk-Gal4 (ref. 21) have been described. For the generation of Vago-mutant lines, the EP element was excised from the line EP(X)2537 (Genexel) with the P-element transposase P{D2-3}99B by standard methodology. Two independent lines obtained by precise excision of the transposon were used as wild-type controls. Hsp-Gal4 and tubulin-Gal4 lines were obtained from the Bloomington stock center. All lines were tested for the presence of wolbachia symbionts. Fly stocks were raised on standard cornmeal–agar medium at 25 1C. Adult flies 4–6 d of age were used in infection experiments. For heat-shock induction of transgene expression, flies were incubated for 20 min at 37 1C, followed by 30 min at 18 1C and another 20 min at 37 1C. After the treatment, flies were allowed to recover for 6 h at 25 1C before immune challenge. Control experiments indicated that heat-shock treatment did not interfere with the induction of Vago (data not shown). Infection with bacteria and fungi was done as described22. Viral stocks were prepared in 10 mM Tris-HCl, pH 7.5. Flies were infected by intrathoracic injection (Nanoject II apparatus; Drummond Scientific) of 4.6 nl of a viral suspension (DCV, 2  1011 particles of a dose lethal to 50% of flies tested, per ml; FHV, 4  1011 plaque-forming units per ml; SINV, 1  108 plaque-forming units per ml). Injection of the same volume of 10 mM Tris-HCl, pH 7.5, was used as a control. Infected flies were then incubated at 22 1C. RNA analysis. RNA was isolated with TRIzol (Gibco-BRL) according to the manufacturer’s instruction and was analyzed by RNA blot with standard procedures. Primers used to generate the DNA probes were as follows: RpL32 forward, 5¢-GTGTATTCCGACCACGTTACA-3¢, and reverse, 5¢ATACAGGCC CAAGATCGTGA-3¢; DCV forward, 5¢-AAAAAGCTAGCGCTTCCTCATATGT TAAAATGCG-3¢, and reverse, 5¢-AAGGATCCT ATATTCAACGTGACTGTTAT GAA-3¢; Vago forward, 5¢-AAAGCGGCCGCTCCGCCGATCTCAGTTCTTTC3¢, and reverse, 5¢-AAATCTAGACAGAGTACCAGTTCACTTCTC-3¢; and vir-1 forward, 5¢-ATGGATCAGCTG CAGCAAATG-3¢, and reverse, 5¢-CATTTCG CAGGCTCTCCTAAAG-3¢. RNA blots were quantified with a bioimager (Fujix BAS 2000). The qPCR MasterMix for SYBR Green I (Eurogenetec) was used for quantitative PCR. Primers used for real-time PCR were as follows: RpL32 forward, 5¢-GACGCTTCAAGGGACAGTATCTG-3¢, and reverse, 5¢-AAACGCG GTTCTGCATGAG-3¢; Vago forward, 5¢-TGCAACTCTGGGAGGATAGC-3¢ and reverse, 5¢-AATTGCCCTGCGTCAGTTT-3¢; vir-1 forward, 5¢-GATCCCA ATTTTCCCATCAA-3¢, and reverse, 5¢-GATTACAGCTGGGTGCACAA-3¢; DCV forward, 5¢-TCATCGGTATGCACATTGCT-3¢, and reverse, 5¢-CGCATAA CCATGCTCTTCTG-3¢; Drosomycin forward, 5¢-CGTGAGAACCTTTTCCAA TATGATG-3¢, and reverse, 5¢-TCCCAGGACCACCAGCAT-3¢; and Diptericin forward, 5¢-GCTGCGCAATCGCTTCTACT-3¢, and reverse, 5¢-TGGTGGAG TGGGCTTCATG-3¢. Gene expression was normalized to the expression of RNA encoding the ‘housekeeping’ ribosomal protein L32 (RpL32), used as loading control. Protein analysis. For proteins isolation, flies were ‘smashed’ in lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1% (vol/vol) Triton X-100 and protease inhibitor ‘cocktail’ (Complete Mini; Roche Diagnostics)). Proteins were quantified by the Bradford method and 30 mg total protein was separated by 12% SDS-PAGE. Standard procedures were used for electrophoresis and immunoblot analysis. Polyclonal antiserum directed against VP2 has been described7, and the actin-specific monoclonal antibody MAB1501R was from Chemicon. Secondary antibodies to mouse (W4021) or rabbit (W4011) conjugated to horseradish peroxidase were from Promega. Proteins on blots were visualized with ECL Chemiluminescent Detection reagents (Amersham).

ADVANCE ONLINE PUBLICATION

NATURE IMMUNOLOGY

ARTICLES

© 2008 Nature Publishing Group http://www.nature.com/natureimmunology

Antibodies and immunostaining. Mouse monoclonal antibody to b-galactosidase (23781) was from Promega. Polyclonal antiserum directed against DCV VP2 and FHV capsid protein have been described7,33. AlexaFluor 488– or AlexaFluor 546–labeled secondary antibodies (A-11001 or A-11010, respectively; Molecular Probes) were used at a dilution of 1:500. Flies were dissected in PBS containing 4% (vol/vol) formaldehyde and were fixed for 20 min. Samples were washed twice in PBS containing 0.1% (vol/vol) Triton X-100 and were blocked for 30 min with 1% (wt/vol) BSA. Samples were incubated overnight at 4 1C with antibody to VP2 (anti-VP2), anti-FHV or anti-bgalactosidase (dilution, 1:500) and were labeled for 4 h at 25 1C with secondary antibodies. Slides were mounted in Vectashield medium (Vector Laboratories) and were examined by confocal microscopy (Zeiss LSM510). Construction of transgenic strains. For construction of the reporter plasmid, a PCR fragment of 1.9 kilobases of Vago upstream sequences was generated with the following primers (NotI and NheI sites underlined): forward, 5¢-TTTGCGGCCGCGGTGAATAGAAGTGTATTTCCATG-3¢, and reverse, 5¢-TTTTGCTAGCCATGCTGCTAATTGACTCCATTT-3¢. This fragment, which corresponds to positions –1979 to +101 (relative to the putative transcription start site) of Vago, was digested with NotI and NheI and was inserted between the corresponding sites in the Casper transformation vector pJL300, which encodes b-galactosidase7. For the creation of UAS-Vago flies, the Vago coding sequence was amplified by PCR on genomic DNA with the following primers (NotI and XbaI sites underlined): forward, 5¢-AAAGCGGCCGCTCCGCC GATCTCAGTTCTTTC-3¢, and reverse, 5¢-AAATCTAGACAGAGTACCAGTT CACTTCTC-3¢. The resulting fragment, which corresponds to positions +41 to +716 (relative to the putative transcription start site), was digested with NotI and XbaI and was inserted between the corresponding sites in the pUAS-T vector12. The resulting constructs were injected into embryos of a w– strain (w1118) to obtain transgenic lines. At least three independent lines were analyzed for each construct. The expression of the reporter transgenes was monitored as described34. Sequence analysis of DExD/H-box helicases. All protein sequences annotated as containing the Pfam protein family database35 families PF00270 (DEAD) and PF00271 (helicase_C) were obtained from the Ensembl genome browser36 with the EnsMart database system37; sequences containing more than one copy of either of these families were discarded. Sequence regions containing these two families were extracted with alignment to the corresponding family hidden Markov models with the HMMER software package. These ‘scripts’ were written with the help of the BioPerl toolkit38. Regions up to 20 amino acids on either side of the PFAM families, as well as the region between the two PFAM families, were extracted from each sequence, and MUSCLE software39 was used, with default settings, to build a multiple alignment of the regions. All residues that contained gaps were removed from the alignment, which yielded 99 residues from 389 sequences. Dendrograms of the evolutionary relationships of these sequences were estimated with neighbor joining40 as implemented by the ClustalX multiple sequence alignment program41 with correction for multiple substitutions and with maximum likelihood as implemented by the phyML algorithm42 with default parameters. The resulting dendrograms were investigated with the Dendroscope interactive viewer43. Precalculated phylogenetic trees for subsets of the DExD/H-box helicase family were examined with the TreeFam database44. Statistical analysis. An unpaired two-tailed Student’s t-test was used to determine statistically significant differences. P values of less than 0.05 were considered statistically significant. GraphPad Prism version 4 for Macintosh (GraphPad Software) was used for statistical analysis of data. Note: Supplementary information is available on the Nature Immunology website.

ACKNOWLEDGMENTS We thank E. Santiago and S. Ozkan for technical expertise; R. Carthew (Northwestern University), J.-M. Reichhart (Unite´ Propre de Recherche´ 9022 Centre National de la Recherche Scientifique) and M. Siomi (University of Tokushima) for fly stocks; A. Schneeman (The Scripps Research Institute) for FHV and anti-FHV; D. Ferrandon for critical reading of the manuscript and comments; and E. Levashina for discussions. Confocal microscopy was done at the Strasbourg Esplanade Cellular Imaging Facility (funded by the Centre

NATURE IMMUNOLOGY

ADVANCE ONLINE PUBLICATION

National de la Recherche Scientifique, Institut National de la Sante´ et de la Recherche Me´dicale, Louis Pasteur University and Alsace Region). Supported by the US National Institutes of Health (PO1 AI070167), Agence Nationale de Recherche Microbiologie-Maladie Emergentes and Centre National de la Recherche Scientifique. AUTHOR CONTRIBUTIONS S.D., N.M., S.M., C.K., D.G.-A., C.D., J.A.H. and J.-L.I. conceived, did and analyzed the experiments; A.B. analyzed the sequences of DExD/H-box helicases; C.A. established and confirmed the identity of transgenic flies expressing FHV B2; and S.D., N.M., J.A.H. and J.-L.I. wrote the manuscript. Published online at http://www.nature.com/natureimmunology/ Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions/

1. Beutler, B. et al. Genetic analysis of resistance to viral infection. Nat. Rev. Immunol. 7, 753–766 (2007). 2. Ding, S.W. & Voinnet, O. Antiviral immunity directed by small RNAs. Cell 130, 413–426 (2007). 3. Galiana-Arnoux, D., Dostert, C., Schneemann, A., Hoffmann, J.A. & Imler, J.L. Essential function in vivo for Dicer-2 in host defense against RNA viruses in Drosophila. Nat. Immunol. 7, 590–597 (2006). 4. van Rij, R.P. et al. The RNA silencing endonuclease Argonaute 2 mediates specific antiviral immunity in Drosophila melanogaster. Genes Dev. 20, 2985–2995 (2006). 5. Wang, X.H. et al. RNA interference directs innate immunity against viruses in adult Drosophila. Science 312, 452–454 (2006). 6. Zambon, R.A., Vakharia, V.N. & Wu, L.P. RNAi is an antiviral immune response against a dsRNA virus in Drosophila melanogaster. Cell. Microbiol. 8, 880–889 (2006). 7. Dostert, C. et al. The Jak-STAT signaling pathway is required but not sufficient for the antiviral response of Drosophila. Nat. Immunol. 6, 946–953 (2005). 8. Kleinjung, J., Romein, J., Lin, K. & Heringa, J. Contact-based sequence alignment. Nucleic Acids Res. 32, 2464–2473 (2004). 9. Sheldon, T.J., Miguel-Aliaga, I., Gould, A.P., Taylor, W.R. & Conklin, D. A novel family of single VWC-domain proteins in invertebrates. FEBS Lett. 581, 5268–5274 (2007). 10. Cherry, S. & Perrimon, N. Entry is a rate-limiting step for viral infection in a Drosophila melanogaster model of pathogenesis. Nat. Immunol. 5, 81–87 (2004). 11. Sabatier, L. et al. Pherokine-2 and -3: Two Drosophila molecules related to pheromone/ odor-binding proteins induced by viral and bacterial infections. Eur. J. Biochem. 270, 3398–3407 (2003). 12. Brand, A.H. & Perrimon, N. Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118, 401–415 (1993). 13. Chao, J.A. et al. Dual modes of RNA-silencing suppression by Flock House virus protein B2. Nat. Struct. Mol. Biol. 12, 952–957 (2005). 14. Li, H., Li, W.X. & Ding, S.W. Induction and suppression of RNA silencing by an animal virus. Science 296, 1319–1321 (2002). 15. Ji, X. The mechanism of RNase III action: how dicer dices. Curr. Top. Microbiol. Immunol. 320, 99–116 (2008). 16. Lee, Y.S. et al. Distinct roles for Drosophila Dicer-1 and Dicer-2 in the siRNA/miRNA silencing pathways. Cell 117, 69–81 (2004). 17. Liu, Q. et al. R2D2, a bridge between the initiation and effector steps of the Drosophila RNAi pathway. Science 301, 1921–1925 (2003). 18. Okamura, K., Ishizuka, A., Siomi, H. & Siomi, M.C. Distinct roles for Argonaute proteins in small RNA-directed RNA cleavage pathways. Genes Dev. 18, 1655–1666 (2004). 19. Akasaka, T. et al. The ATP-sensitive potassium (KATP) channel-encoded dSUR gene is required for Drosophila heart function and is regulated by tinman. Proc. Natl. Acad. Sci. USA 103, 11999–12004 (2006). 20. Reichhart, J.M. et al. Splice-activated UAS hairpin vector gives complete RNAi knockout of single or double target transcripts in Drosophila melanogaster. Genesis 34, 160–164 (2002). 21. Georgel, P. et al. Drosophila immune deficiency (IMD) is a death domain protein that activates the antibacterial response and can promote apoptosis. Dev. Cell 1, 503–514 (2001). 22. Tauszig-Delamasure, S., Bilak, H., Capovilla, M., Hoffmann, J.A. & Imler, J.L. Drosophila MyD88 is required for the response to fungal and Gram-positive bacterial infections. Nat. Immunol. 3, 91–97 (2002). 23. Busse, M.S., Arnold, C.P., Towb, P., Katrivesis, J. & Wasserman, S.A. A kB sequence code for pathway-specific innate immune responses. EMBO J. 26, 3826–3835 (2007). 24. Obbard, D.J., Jiggins, F.M., Halligan, D.L. & Little, T.J. Natural selection drives extremely rapid evolution in antiviral RNAi genes. Curr. Biol. 16, 580–585 (2006). 25. Sackton, T.B. et al. Dynamic evolution of the innate immune system in Drosophila. Nat. Genet. 39, 1461–1468 (2007). 26. Klotman, M.E. & Chang, T.L. Defensins in innate antiviral immunity. Nat. Rev. Immunol. 6, 447–456 (2006). 27. Smit, A.B. et al. Granularin, a novel molluscan opsonin comprising a single vWF type C domain is up-regulated during parasitation. FASEB J. 18, 845–847 (2004).

7

© 2008 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES 28. Georgel, P. et al. Vesicular stomatitis virus glycoprotein G activates a specific antiviral Toll-like receptor 4-dependent pathway. Virology 362, 304–313 (2007). 29. Bernstein, E., Caudy, A.A., Hammond, S.M. & Hannon, G.J. Role for a bidentate ribonuclease in the initiation step of RNA interference. Nature 409, 363–366 (2001). 30. Duchaine, T.F. et al. Functional proteomics reveals the biochemical niche of C. elegans DCR-1 in multiple small-RNA-mediated pathways. Cell 124, 343–354 (2006). 31. Tabara, H., Yigit, E., Siomi, H. & Mello, C.C. The dsRNA binding protein RDE-4 interacts with RDE-1, DCR-1, and a DExH-box helicase to direct RNAi in C. elegans. Cell 109, 861–871 (2002). 32. Liu, X., Jiang, F., Kalidas, S., Smith, D. & Liu, Q. Dicer-2 and R2D2 coordinately bind siRNA to promote assembly of the siRISC complexes. RNA 12, 1514–1520 (2006). 33. Krishna, N.K., Marshall, D. & Schneemann, A. Analysis of RNA packaging in wild-type and mosaic protein capsids of flock house virus using recombinant baculovirus vectors. Virology 305, 10–24 (2003). 34. Jung, A.C., Criqui, M.C., Rutschmann, S., Hoffmann, J.A. & Ferrandon, D. Microfluorometer assay to measure the expression of b-galactosidase and green fluorescent protein reporter genes in single Drosophila flies. Biotechniques 30, 594–598 600–591 (2001).

8

35. Finn, R.D. et al. Pfam: clans, web tools and services. Nucleic Acids Res. 34, D247–D251 (2006). 36. Hubbard, T.J. et al. Ensembl 2007. Nucleic Acids Res. 35, D610–D617 (2007). 37. Kasprzyk, A. et al. EnsMart: a generic system for fast and flexible access to biological data. Genome Res. 14, 160–169 (2004). 38. Stajich, J.E. et al. The Bioperl toolkit: Perl modules for the life sciences. Genome Res. 12, 1611–1618 (2002). 39. Edgar, R.C. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797 (2004). 40. Saitou, N. & Nei, M. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425 (1987). 41. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. & Higgins, D.G. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25, 4876–4882 (1997). 42. Guindon, S. & Gascuel, O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst. Biol. 52, 696–704 (2003). 43. Huson, D.H. et al. Dendroscope: An interactive viewer for large phylogenetic trees. BMC Bioinformatics 8, 460 (2007). 44. Li, H. et al. TreeFam: a curated database of phylogenetic trees of animal gene families. Nucleic Acids Res. 34, D572–D580 (2006).

ADVANCE ONLINE PUBLICATION

NATURE IMMUNOLOGY