w¹¹¹⁸ flies were used as wild-type control. The w^IR ; dcr-2^L811fsX mutant flies have been previously described (27). ubi-Gal4,tub-Gal80^ts was a gift from Dr. D. Ferrandon. hop^Tum-l , ppl-Gal4, da-Gal4, hs-Gal4 were obtained from Bloomington Stock center. The generation of transgenic UAS-Ect4 and U6:3-gRNA-Ect4 lines was performed as previously described (28). Ect4-IR was obtained from NIG-FLY stocks (HMJ30091). For the generation of Ect4 mutant lines, transgenic U6:3-gRNA-Ect4 flies were crossed with the nos-Cas9 flies to get male F₀ (nos-Cas9/+; U6:3-gRNA-Ect4/+) that were crossed with w¹¹¹⁸ ; TM3, Sb/TM6B, Tb to obtain F₁ progenies. Singular F₁ flies were crossed with w¹¹¹⁸ ; TM3, Sb/TM6B, Tb. PCR products amplified from F₁ flies before being cloned into the pMD19-T vector according to the manufacturer’s instructions (TAKARA) for mutation identification.

pAC5.1-Ect4-Flag was made by cloning Ect4 cDNA into pAC5.1-Flag vectors. For pAC5.1-Ect4-GFP constructs, the EGFP fragment was amplified from pEGFP-C1 and assembled with the Ect4 fragment into pAC5.1-V5 vectors using ClonExpress MultiS One Step Cloning Kit (Vazyme). Stat92E cDNA was inserted in pAC5.1-HA to generate pAC5.1-Stat92E-HA. The hop (or Dome) cDNA was inserted in pAC5.1-V5 to generate pAC5.1-hop-V5 (or pAC5.1-Dome-V5). For the truncated Ect4 constructs, ARM domain (aa 318-701), SAM domain (aa 680-826), and TIR domain (aa 829-1360) were amplified from pAC5.1-Ect4-Flag before assembled into pAC5.1-Flag empty vector, respectively.

S2 cells were cultured at 25°C in Sf-900™ III SFM (Gibco). All S2 cell transfection experiments were carried out with the Effectene Transfection Reagent (QIAGEN). For a co-immunoprecipitation assay, S2 cells were transfected with different plasmids. After 48h, cells were collected and lysed in lysis buffer (150 mM NaCl, 25mM Tris-HCL, pH 7.4, 5% glycerol, 1% NP-40, 1mM EDTA, complete protease inhibitor cocktail tablets [Roche] and phosphatase inhibitor cocktail tablets [Roche]). Lysates were incubated overnight at 4°C with Anti-Flag M2 affinity gel (Sigma) or EZview Red Anti-HA Affinity Gel (Sigma). After centrifugation, pellets were washed with 1ml lysis buffer three times before resuspension in 2X Laemmli SDS-PAGE buffer and detection by Western blot. Western blot was performed according to standard procedures.

Primary antibodies: Mouse anti-V5 (1:8000, Proteintech 66007-1-Ig); mouse anti-HA (1:8000, Milipore 05-904); mouse anti-α-Tubulin (1:20,000 Sigma T8203); goat anti-Stat (1:5000, Santa Cruz Biotechnology dN-17); mouse anti-FLAG (1:8000, Sigma F3165); rabbit anti-DCV (1:5000, Abcam ab92954); mouse anti-PY20 (1:2000, Abcam ab10321). Secondary antibodies: HRP-linked anti-mouse IgG (1:8000, Cell Signaling Technology 7076P2); HRP-linked anti-rabbit IgG (1:5000, Cell Signaling Technology 7074P2); HRP-linked anti-goat IgG (1:5000 Millipore AP106P); Alexa Fluor 555 goat anti-mouse IgG (1:500, life technologies A21422).

According to the manufacturer’s instructions, total RNA was extracted from infected flies using RNAiso Plus (TAKARA), and cDNA was synthesized with the HiScript II Q RT SuperMix (Vazyme). The ChamQ SYBR qPCR Master Mix (Vazyme) was used for quantitative. Expression of the gene of interest was normalized to the Rpl32 RNA level. The following primers were used for qPCR: RpL32 (forward 5’-GACGCTTCAAGGGACAGTATCTG-3’; reverse 5’-AAACGCGGTTCTGCATGAG-3’), vir-1 (forward 5’-GATCCCAATTTTCCCATCAA-3’; reverse 5’-GATTACAGCTGGGTGCACAA-3’), DCV (forward 5’-TCATCGGTATGCACATTGCT-3’; reverse 5’-CGCATAACCATGCTCTTCTG-3’), TotA (forward 5’-CCCTGAGGAACGGGAGAGTA-3’; reverse 5’-CTTTCCAACGATCCTCGCCT-3’), TotM (forward 5’-ACCGGAACATCGACAGCC-3’; reverse 5’-CCAGAATCCGCCTTGTGC-3’), Ect4 (forward 5’-GCCTCCAGTATTACGGT-3’; reverse 5’-ATGTTTCTCCTGACTGATGA-3’), Vago (forward 5’-TGCAACTCTGGGAGGATAGC-3’; reverse 5’-AATTGCCCTGCGTCAGTTT-3’).

Virus stocks were prepared as described previously (29). All fly lines confirmed the absence of Wolbachia by PCR and were cured whenever necessary. For infection, 3-6 d old flies were anesthetized with CO₂ and injected with PBS (Gibco) or virus suspension intra-thoracically using the Nanoject II injector (Drummond). Infected flies were monitored daily for survival rate or frozen for RNA analysis at the indicated time points.

S2 cells were transfected with pAC5.1-Ect4-GFP and pAC5.1-Stat92E-HA plasmids, and approximately 1×10⁶ cells were transferred to 24 well plates containing coverslips 48 h after transfection. Twelve h later, cells were washed in 0.5ml PBS and fixed with 4% formaldehyde in PBS for 15 min, then washed twice in 0.5 PBT (PBS containing 0.1% Tween-20) before blocking with 5% BSA in TBST for 1 h. Cells were then incubated with primary antibody (anti-HA 1:1000) overnight at 4°C before 2×5 min TBST washes. The secondary antibody was incubated for 4 h at room temperature. Nuclei were stained with PBS with 10 μg/ml DAPI for 5 min. Immunostaining samples were photographed with a Zeiss confocal microscope.

For eye pigment assay, the heads of 50 female flies (2-3 d old, raised at 25°C) of each indicated genotype were homogenized in methanol (1 ml, acidified with 0.1% HCl). After centrifugation, the supernatants were measured for absorbance at 480 nm.

dsRNA targeting Ect4 and GFP were synthesized according to standard protocol. S2 cells were treated with a culture medium containing 10 μg/ml dsRNA for 3 d. After dsRNA treatment, a solution containing 2 mM H₂O₂ and 1 mM sodium vanadate (final concentrations; Sigma) pre-incubated for 15 min was added to S2 cells to induce tyrosine phosphorylated Stat92E for 30 min. Cell lysates were prepared with the lysis buffer before immune precipitation with an anti-Stat92E antibody at 4°C and incubated with Pierce Protein A/G Plus Agarose (Thermo Scientific) beads. Co-immuno-precipitated proteins were detected with an anti-Stat92E antibody or anti-PY20 antibody.

Survival data were analyzed by the Kaplan-Meier method using GraphPad Prism. Quantification of immunoblots was performed with ImageJ 1.51p. Altered protein levels were presented as normalized fold change compared to the control value. Statistical analysis was performed using the Student’s t-test. Grey value analysis was performed by the ZEN 2012 (blue edition) system. P-values below 0.05 were considered significantly different.

To investigate the role of Ect4 in Drosophila antiviral defense, an Ect4 mutant line was generated with the CRISPR/Cas9 system. The mutation, Ect4¹⁷ , covers a genomic deletion of 17 bp in the coding region of Ect4 ( Figure 1A ). Ect4¹⁷ homozygous mutants are lethal at the second instar larval stage as judged by examining the development of both homo- and heterozygous animals distinguished by a GFP marker ( Figure S1A ), and heterozygous mutants were used for further experiments.

Variation in the pastrel gene is associated with natural resistance to DCV infection in D. melanogaster. The non-synonymous single nucleotide polymorphism (SNP) position 598, located in the last exon, has the strongest effect on DCV susceptibility (30). Sequencing of the pastrel locus revealed that all the strains of D. melanogaster tested contained the susceptible allele (data not shown), thus limiting the effect of discordance in the SNP profile between different fly lines to the difference in DCV resistance.

Ect4 transcription in Ect4¹⁷/+ heterozygotes was reduced by 45% compared with the wild-type flies ( Figure S1B ). Wild-type and Ect4 mutant files were challenged with DCV by intra-thoraxic injection. Ect4 mutants were more sensitive to infection than wild-type flies, with a significantly different mean survival of 5 and 6 d for Ect4¹⁷/+ and w¹¹¹⁸ male flies, respectively ( Figure 1B ). Notably, a significant increase in the DCV viral loading was observed in Ect4¹⁷/+ flies at 48 and 72 h post-infection ( Figure 1C ), indicating that Ect4 mutants are more sensitive to DCV infection.

To consolidate the DCV sensitivity phenotype observed with heterozygous Ect4 individuals, the temperature-sensitive Gal80ts allele (31) was used to knockdown Ect4 expression in adult flies by shifting the culture temperature from 18-20°C to 29°C before and during the infection of DCV. RT-qPCR shows that Ect4 expression decreased after the temperature shift to 29°C ( Figure S3A ). As expected, flies with knockdown of Ect4 succumbed earlier to DCV infection than the control flies ( Figure 1D ). Consistently, the down-regulation of Ect4 increased viral proteins ( Figure 1E ). Since DCV replicates mainly in fat bodies (32), we employed a fat body-specific driver, ppl-Gal4, to knock down Ect4 expression in the fat body. Specific depletion of Ect4 in the fat body under the control of ppl-Gal4 also affected the survival rate and viral load upon DCV infection ( Figures S2A, B ). The decreased survival rate was correlated with the increased viral burden in Ect4-RNAi flies.

To verify the specificity of the function of Ect4 in DCV infection, UAS-Ect4 transgenic flies were crossed with a ubiquitous Gal4 driver, da-Gal4, to express the Ect4 transgene ectopically. Remarkably, ubiquitous overexpression of Ect4 promoted survival after the viral challenge ( Figure 2A ). Further, the increased dose of Ect4 led to decreased viral burden in infected flies ( Figure 2B ). Interestingly, as shown in Figures 2C, D , flies overexpressing Ect4, specifically in the fat body using the ppl-Gal4 driver, showed significantly more resistance to DCV infection than control flies and significantly decreased DCV replication levels. More importantly, rescue experiments by the expression of Ect4 in Ect4¹⁷/+ flies under the control of da-Gal4 were performed to prove the specific role of Ect4 in protecting flies from viral infections. The decreased survival rate of Ect4 mutants after DCV infection, as well as increased viral load, was rescued to similar levels of the control flies following transgenic expression of Ect4 in heterozygous Ect4 mutants ( Figures 2E, F ). Similar results were obtained when a hs-Gal4 driver was used for the rescue experiment. Ect4 heterozygous mutant flies expressing Ect4 under the control of hs-Gal4 exhibited a decreased viral replication at 48 h post-infection. ( Figure S2C ). These results indicate that Ect4 confers resistance against DCV infection and is required to control the accumulation of viruses.

RNA interference (RNAi) acts as the first line of defense against viruses in Drosophila (33). There was strong genetic evidence that one RNAi-related pathway, the siRNA pathway, plays a major role in antiviral immunity in Drosophila (7, 34). Since heterozygous Ect4 mutant flies are hypersensitive to DCV infection, we asked whether the down-regulation of Ect4 affects the function of the siRNA pathway. To address this question, siRNA pathway activity was monitored using an in vivo sensor assay, wherein the endogenous white gene is silenced by the expression of a hairpin dsRNA corresponding to an exon of white. Expression of UAS-wIR using the eye-specific driver GMR-Gal4 alters eye pigmentation to a white color or pale orange if the silencing is incomplete. Studies have shown that the siRNA pathway is inactivated without Dicer-2 (Dcr-2). Therefore eye pigmentation of a Dcr-2 null mutant (dcr-2^L811fsX/dcr-2^L811fsX ) in GMR>UAS-wIR background is red, whereas Dcr-2 heterozygous mutants (dcr-2^L811fsX/+ ) display pale orange (27). Our results show that mutation in Ect4 did not lead to any changes in the eye pigmentation in w^IR; dcr-2^L811fsX/+ ( Figures 3A, B ). Moreover, the expression of Vago, induced in DCV infection dependent on Dicer-2 (33), did not differ between wild-type and Ect4 mutant flies at 48 and 72 hpi ( Figure 3C ). These results suggest that Ect4 does not directly affect the antiviral siRNA pathway.

The JAK/STAT pathway was shown to contribute to the antiviral response in Drosophila (9), where several genes are induced following viral infection via the JAK/STAT pathway including virus-induced RNA-1 (vir-1), the stress-induced genes Turandot A and M (TotA and TotM) (18). To examine whether the downregulation of Ect4 affected JAK/STAT pathway activation, we examined the expression of vir-1, TotA, and TotM by RT-qPCR at 48 and 72 h after DCV infection (hpi). As previously reported, DCV infection induced a strong up-regulation of vir-1, TotA, and TotM in wild-type flies (18). However, vir-1 induction in response to DCV infection in Ect4 mutant flies was indistinguishable from the control ( Figure 4C ). Similar results were observed using a ubiquitous temperature-sensitive Gal4 driver, ubi-Gal4 ( Figures S3B, C ). A genetic interaction experiment was performed to assess further the relationship between Ect4 and the JAK/STAT pathway. TotA and TotM were expressed in flies carrying a JAK gain-of-function allele Tum-l (hop^Tum-l ), which encodes a hyperactive JAK kinase due to a G341E substitution (35). Reducing the dosage of Ect4 by half resulted in a large reduction of the RNA levels of TotA and TotM in hop^Tum-l flies ( Figure 4D ). The TotA and TotM response was also attenuated in the fat body of flies where Ect4 was downregulated by expressing the Ect4-IR transgene using a ppl-Gal4 driver ( Figures S3D, E ). This TotA and TotM expression reduction was rescued by ubiquitously expressed Ect4 ( Figures 4E, F ). Together, these results suggest that Ect4 genetically interacts with the JAK/STAT pathway to regulate the expression of TotA and TotM in response to DCV infection.

To unravel the molecular mechanism underlying the relationship between Ect4 and JAK/STAT pathway, we examined whether Ect4 interacted with any known components of the JAK/STAT pathway. Differentially tagged forms of JAK/STAT pathway components and Ect4 were expressed in S2 cells, and co-immunoprecipitation studies were performed. As shown in Figure 5A , Ect4 is associated with the transcription factor Stat92E but not other key components (Hop or Dome) of the JAK/STAT pathway. Consistent with previous findings (36), Stat92E protein was located both in the cytoplasm and nucleus as visualized by immunofluorescence staining. Since the green fluorescent protein (GFP)-Ect4 fusion protein was localized in the cytoplasm, the interaction between the two proteins occurs in the cytoplasm ( Figure 5B ).

Ect4 protein harbor three different domains: ARM (Armadillo motif) domains followed by two SAM (Sterile Alpha motif) domains and TIR (Toll -Interleukin-1 receptor) domain ( Figure 5C ). To further investigate the molecular basis of the interaction between Ect4 and Stat92E, a series of truncated forms of Ect4 were generated. Co-immuno-precipitation studies showed that ARM and SAM domains were likely not required for Ect4 to interact with Stat92E, whereas the TIR domain was essential since only the TIR domain co-immuno-precipitated with Stat92E ( Figure 5D ). Together, these results suggest that Ect4 may regulate the JAK/STAT signaling activity by interacting with Stat92E.

As described thus far, we show that Ect4 regulates the expression of JAK/STAT-dependent genes TotA and TotM and is associated with Stat92E. It is intriguing to predict that Ect4 may affect Stat92E phosphorylation. To test this hypothesis, we employed an RNAi approach to knock down Ect4 in S2 cells ( Figure 6A ). Previous studies have shown that tyrosine residues of Stat92E are phosphorylated after treatment of S2 cells with pervanadate, which activates Stat92E in a ligand-independent manner, while activation is not present in untreated cells (36, 37). As shown in Figure 6 , treatment with dsRNA targeting Ect4 mRNA resulted in a significant reduction of tyrosine phosphorylated Stat92E upon pervanadate treatment, as compared with the control using dsRNA targeting GFP. It was noted that upon DCV infection of S2 cells, phosphorylated Stat92E (p-Stat92E) was not detected by immunostaining, likely due to the transient activity of p-Stat92E dimers. Therefore, is Ect4 required for the nuclear translocation of Stat92E? As expected, reduced nuclear translocation of Stat92E in response to the pervanadate stimulus was detected in cells treated with Ect4 RNAi ( Figures 6C, D ).