The Aedes aegypti Domino Ortholog p400 Regulates Antiviral Exogenous Small Interfering RNA Pathway Activity and ago-2 Expression

Female Aedes aegypti mosquitoes are vectors of human-infecting arthropod-borne viruses (arboviruses). In recent decades, the incidence of arthropod-borne viral infections has grown dramatically. Vector competence is influenced by many factors, including the mosquito’s antiviral defenses. The exogenous small interfering RNA (siRNA) pathway is a major antiviral response restricting arboviruses in mosquitoes. While the roles of the effectors of this pathway, Argonaute-2 and Dicer-2 are well characterized, nothing is known about its regulation in mosquitoes. In this study, we demonstrate that A. aegypti p400, whose ortholog Domino in Drosophila melanogaster is a chromatin-remodeling ATPase member of the Tip60 complex, regulates siRNA pathway activity and controls ago-2 expression levels. In addition, we found p400 to have antiviral activity against different arboviruses. Therefore, our study provides new insights into the regulation of the antiviral response in A. aegypti mosquitoes.

A rthropod-borne viruses (arboviruses) are transmitted to susceptible mammalian hosts through the bite of infected arthropod vectors, such as mosquitoes. This group of viruses includes those of medical and veterinary importance, such as chikungunya virus (CHIKV) (Alphavirus; Togaviridae), dengue virus (DENV) (Flavivirus; Flaviviridae), Zika virus (ZIKV) (Flavivirus; Flaviviridae), and Rift Valley fever virus (RVFV) (recently reclassified into Phlebovirus; Phenuiviridae) (1)(2)(3)(4)(5)(6). Given the medical and economic impacts of arboviruses, viral interactions with mosquitoes and the impact on transmission remain important areas of research. New strategies to interfere with arbovirus transmission involve genetically modified mosquitoes, including making mosquitoes more resistant to arboviruses, as well as Wolbachia-endosymbiont-based approaches (7)(8)(9)(10)(11)(12)(13)(14)(15). Arboviruses infect and replicate in both mammalian host and vector host cells; as a consequence, they are detected by both mammalian and arthropod immune systems. Compared to the wealth of knowledge on the mammalian antiviral immune response, the arthropod response is not well characterized, yet it could provide important targets for novel vector control measures. Research on this topic has established that the key player in the antiviral immune response in mosquito vectors is the exogenous small interfering RNA (exo-siRNA) pathway (8,(16)(17)(18)(19). This pathway is activated by the recognition of long viral double-stranded RNA (dsRNA) by the endo-RNase Dicer-2 (Dcr-2). The effector protein Dcr-2 cleaves viral dsRNA into 21-nucleotide (nt) virus-derived siRNAs (vsiRNAs), which are then loaded into the RNA-induced silencing complex (RISC). One key protein within RISC, Argonaute-2 (Ago-2), binds the siRNA and is believed to unwind the siRNA duplex and use one of the strands as a guide to specifically recognize complementary RNAs, targeting these for degradation. In the case of viral RNAs (such as genomes, antigenomes, or mRNAs), this results in an inhibition of virus replication. The antiviral activity mediated by the exo-siRNA pathway of vector mosquitoes has been demonstrated through the silencing or elimination/ absence of effector proteins (such as Ago-2 or Dcr-2), which results in an upregulation of arbovirus replication. It has been shown to be active against arboviruses of all major families or orders, alphaviruses (20)(21)(22)(23)(24)(25)(26)(27)(28)(29), flaviviruses (30)(31)(32)(33)(34), and bunyaviruses (35)(36)(37).
Although the importance of the key exo-siRNA pathway components, such as Ago-2, has been established, the regulation of this antiviral response, such as effector protein expression, activity, and activation, remains poorly understood. Given its attractiveness as a target for immunity-based control strategies in vectors, it is important that all of the components, including regulatory mediators, of the exo-siRNA pathway are better understood. In this study, we sought to identify new components of the antiviral exo-siRNA pathway with the aim of understanding how the pathway functions and how it is regulated. We show that the Aedes aegypti Domino ortholog p400 is expressed at different levels across mosquito tissues and is an antiviral factor, since silencing of the gene enhanced replication of the mosquito-borne alphavirus Semliki Forest virus (SFV). We confirmed that p400 not only acts on SFV but also on the SFV-related alphavirus CHIKV, as well as Bunyamwera virus (BUNV), a member of the Peribunyaviridae family. However, p400 did not show antiviral activity against ZIKV. In addition, we have identified p400 as a regulator of the exo-siRNA pathway activity, possibly by controlling ago-2 but not dcr-2 transcript levels in vivo. Thus, our results hint that this protein may exert antiviral activity through regulating the exo-siRNA pathway, although other antiviral pathways potentially regulated by p400 cannot be excluded. These findings help to further understand the regulation of the antiviral exo-siRNA pathway.

RESULTS
p400 is expressed in female mosquitoes. Previous studies have reported the role of Domino in the antiviral response in Drosophila melanogaster (38). We therefore wanted to determine if the A. aegypti Domino ortholog, p400, has a similar role in mosquitoes. To determine if p400 was expressed in tissues relevant to arbovirusmosquito interactions, we first investigated the presence of p400 transcripts across mosquito tissues. For this, individual tissues were dissected and hemocytes perfused from non-blood-fed (NBF) female mosquitoes. RNA was extracted and transcript levels analyzed by reverse transcription-quantitative PCR (RT-qPCR). As shown in Fig. 1, p400 transcripts were detected in hemocytes and all tissues tested. The expression levels of p400 varied across tissues, with low levels in salivary glands and high levels in ovaries.
The presence of p400 was also investigated by immunofluorescence. Consistent with our RT-qPCR results, p400 was detected in hemocytes and several sampled tissues (Fig. 2). p400 was expressed in all perfused hemocytes. It was restricted to the nucleus in prohemocytes ( Fig. 2A), which are small and spherical cells, with a high nuclear/ cytoplasmic ratio (39). It was expressed in the nucleus as well as in the cytoplasm of differentiated hemocytes or granulocytes ( Fig. 2B and C), which are bigger cells with filopodia and a low nuclear/cytoplasmic ratio (39). In tissues, p400 was detected in the nucleus of cells in the digestive tract, such as in the crop (Fig. 2E to H), in the ovarian sheath surrounding the whole ovary and oviduct ( Fig. 2J and M to P), and in the ovariolar sheath surrounding each ovariole ( Fig. 2K and L). In ovarioles, p400 was expressed in some of the somatic cells forming the epithelium surrounding the germarium and the primary previtellogenic follicle ( Fig. 2K and L). In the germarium, p400 was also strongly expressed in the nucleus of differentiating germ cells of the developing cyst (which will give rise to the secondary follicle) as well as in the germ line stem cells and to a lesser extent in the cystoblast germ cells (Fig. 2K). We also found p400 in the tracheal cells surrounding various tissues ( Fig. 2D and I to K). p400 is antiviral against SFV in A. aegypti females. Yasunaga et al. previously reported that Domino has antiviral activity against West Nile virus (WNV) and vesicular stomatitis virus (VSV) in Drosophila DL1 cells (38). Therefore, we sought to determine if the Domino ortholog p400 has similar antiviral activity in A. aegypti females. For this, we assessed whether p400 could be silenced in vivo and determined the effect of this knockdown on viral infection using the model alphavirus SFV. Female A. aegypti mosquitoes were injected intrathoracically with p400-targeting dsRNA (dsp400) or control lacZ-targeting dsRNA (dsLacZ). Mosquitoes were fed with a blood meal containing SFV4 at 4 days post-dsRNA injection (pdi). At 3 days postinfection (pi), RNA was extracted from whole females. Knockdown efficiency and virus genome levels were determined by RT-qPCR. The knockdown of p400 transcript expression was confirmed (Fig. 3A) and resulted in a significant increase in SFV RNA levels in whole females (Fig. 3B), confirming the role of p400 in inhibiting SFV replication in A. aegypti female mosquitoes. The results were repeated in a second independent experiment (see Fig. S1 in the supplemental material). p400 shows antiviral activity against SFV, CHIKV, and BUNV. To determine if the antiviral action of p400 was specific to SFV or if it has broader antiviral activity, we FIG 1 Detection of p400 transcripts in tissues of NBF A. aegypti females. Presence of p400 transcripts was determined in hemocytes, salivary glands, digestive tracts, ovaries, and carcasses of NBF females by RT-qPCR using the 2 ϪΔΔCT method and expression in hemocytes as the reference sample. Bars represent the fold change in gene expression from 3 independent biological replicates (pools of 25 digestive tracts or ovaries, pools of 60 salivary glands, and pools of perfused hemocytes from 70 females per replicate). Error bars show the minimum and maximum fold change.
tested the effect of p400 knockdown on three other arboviruses belonging to different families, the SFV-related alphavirus CHIKV, the bunyavirus BUNV, and the flavivirus ZIKV. For this experiment, A. aegypti-derived Aag2 cells were used, and the effect of the p400 knockdown on SFV was used as a positive control. The reporter viruses used for these experiments were CHIKV-2SG-FFLuc (FFLuc, firefly luciferase expressing), BUNV-NLuc and ZIKV-NLuc (both NLuc, Nano luciferase expressing), and SFV4(3H)-FFLuc (also firefly luciferase expressing) (25,29,35,40). Aag2 cells were transfected with dsp400-targeting or control dsRNA-targeting enhanced green fluorescent protein (dseGFP). Cells were infected with CHIKV-2SG-FFLuc, BUNV-NLuc, ZIKV-NLuc, or SFV4(3H)-FFLuc at 24 h posttransfection (pt). SFV-and CHIKV-infected cells were lysed 24 hpi, BUNV-infected cells were lysed at 48 hpi, and ZIKV-infected cells were lysed at 72 hpi, and luciferase activity was determined. The time posttransfection at which cells were collected was chosen based on previous studies (29,34,35,41). Knockdown of p400 (Fig. 4A) resulted in a significant increase in SFV (Fig. 4B), confirming our results obtained in vivo, as well as on CHIKV (Fig. 4C) and BUNV (Fig. 4D) replication. However, p400 knockdown had no significant effect on ZIKV, possibly due to higher variation of luciferase expression across replicates for this virus (Fig. 4E). To ensure that the absence of an effect on ZIKV , digestive tracts (E to I), and ovaries (J to P) by immunofluorescence assay using an anti-p400 antibody. The signal was determined using an Alexa Fluor 568 goat anti-mouse IgG (HϩL) (red). Nuclei are stained by DAPI (blue signal), and F-actin is stained by phalloidin 488 (green signal). Images were acquired on a Zeiss LSM 710 inverted confocal microscope with 40ϫ, 63ϫ, and 100ϫ oil immersion objectives. Scale bars are 5 m (A to C, F to H, and N to P) and was not due to the time of sampling, a time course of p400 knockdown and ZIKV infection was performed with samples analyzed at 24, 48, and 72 hpi. p400 expression was significantly reduced at all time points (Fig. S2A); however, luciferase expression was not changed at any time point assessed (Fig. S2B).
p400 is required for exo-siRNA pathway activity and regulates levels of ago-2 in vivo. As p400 had previously been identified as being a component of the exo-siRNA pathway in D. melanogaster-derived S2 cells (42), and we have shown here that it affects arbovirus replication both in vitro and in vivo, we sought to determine if the antiviral action of p400 could be due to an effect on the exo-siRNA pathway-silencing activity in A. aegypti. In order to determine if p400 was important for exo-siRNA pathway activity, reporter assays were performed. Aag2 cells were transfected with pIZ-FLuc (expressing the silencing target FFLuc), pAcIE1-RLuc (as an internal transfection control, expressing Renilla luciferase), and dsp400 or dseGFP (control). At 24 hpt, cells were transfected again with dsRNA targeting either eGFP (control) or FFLuc (dsFFluc). Cells were lysed 24 h after the second transfection, and luciferase activity was determined. While the introduction of dsFFLuc in cells treated with dseGFP significantly decreased FFLuc expression, silencing of the FFLuc reporter plasmid after the introduction of dsFFLuc was abolished in the presence of p400-targeting dsRNA (Fig. 5), showing that p400 is required for the activity of the exo-siRNA pathway.
Next, we aimed to identify what role p400 might play in the exo-siRNA pathway. In human cells, p400 is part of the Tip60 complex and is involved in transcriptional regulation through chromatin remodeling (43). We therefore reasoned that p400 may play a regulatory role in the exo-siRNA pathway at the transcriptional level. In order to test this hypothesis, we investigated the expression levels of the exo-siRNA effector Ago-2 in SFV4-infected females following the knockdown of p400. The female mosquitoes injected with LacZ (control) or p400-targeting dsRNA and given a blood meal containing SFV4 at 4 days pdi (used in Fig. 3B) were further analyzed for this purpose. RT-qPCR was performed to assess transcript levels in whole females at 3 dpi. As shown in Fig. 6A, ago-2 transcript expression levels in whole females were significantly reduced upon p400 transcript knockdown in vivo. As described previously, the results were repeated in a second independent experiment (Fig. S1). Similarly, ago-2 transcript expression levels were also reduced in whole noninfected, NBF females following knockdown of p400 transcripts (Fig. 6B). In contrast, knockdown of p400 had no significant effect on dcr-2 transcript levels.
FIG 3 p400 knockdown significantly enhances replication of SFV in A. aegypti females. p400 and SFV expression in whole dsLacZ-or dsp400-injected females 3 days after an SFV4-infected blood meal were analyzed by RT-qPCR. Normalized expression-per-sample values were calculated as described by Taylor et al. (78) in order to obtain normalized expression values, relative to the ribosomal S7 transcript as reference, with a geomean of 1 for the dsLacZ control group. Bars represent the geomean and 95% confidence intervals from n ϭ 73 dsLacZ and n ϭ 66 dsp400. The results were analyzed using a Mann-Whitney test using log 2 -normalized expression values. This experiment is representative of two independent biological replicates (second replicate shown in Fig. S1). (A) p400 transcript levels are significantly reduced in dsp400-injected females compared to that in dsLacZ-injected females (mean fold change, 1.7; geomean fold change, 1.6; ****, P Ͻ 0.0001). (B) SFV4 titers are significantly higher in dsp400-injected females than in dsLacZ-injected ones (mean fold change, 5.3; geomean fold change, 10.1; ****, P Ͻ 0.0001). A base 10 log scale is used for the y axis.

DISCUSSION
In this study, we found that p400, an ortholog of the TIP60 complex D. melanogaster protein Domino, is antiviral against the arbovirus SFV in both A. aegypti mosquitoderived cells in culture and adult females. We show that p400 is also antiviral against the related alphavirus CHIKV as well as the bunyavirus BUNV. As already discussed, antiviral activity against WNV (a mosquito-borne flavivirus; Flaviviridae) and VSV (Rhabdoviridae) has previously been attributed to the p400 ortholog Domino in D. melanogaster-derived DL1 cells following a dsRNA screen for antiviral proteins (38). The knockdown of other members of the Tip60 complex (RuvBL1, RuvBL2, and Tip60) in mosquito cells was also shown to increase WNV and VSV replication (38), suggesting that p400/Tip60 has broad and conserved antiviral activity in dipterans. In mammals, FIG 4 p400 knockdown significantly enhances replication of SFV, CHIKV, and BUNV but not ZIKV. (A) p400 knockdown efficiency in Aag2 cells was analyzed by RT-qPCR using the 2 ϪΔΔCT method and p400 expression calculated relative to eGFP dsRNA transfection. Bars show the fold change in gene expression from 3 independent experiments (3 wells in each independent experiment, with an average of three wells per condition/experiment used for statistical analysis). Error bars represent the minimum and maximum fold change. The results were analyzed by a one-sample t test. p400 transcript levels are significantly reduced after dsp400 transfection (P ϭ 0.039; *, P Ͻ 0.05). The effect of p400 knockdown on SFV-FFLuc (B), CHIKV-2SG-FFLuc (C), BUNV-NLuc (D), and ZIKV-NLuc (E) was assessed by a luciferase assay. Bars show the means of the results from 3 independent experiments (3 wells in each independent experiment, with the average of three replicates per condition/experiment used for statistical analysis). Values were calculated relative to the control eGFP dsRNA-transfected sample, which was set to 1. Statistical significance was determined by performing a one-sample t test (SFV P ϭ 0.0129, CHIKV P ϭ 0.0286, and BUNV P ϭ 0.0257; *, P Ͻ 0.05). Error bars show the standard error of mean.
Tip60 has been reported to have antiviral activity against adenovirus (44). In contrast, Tip60 can also promote infections with human papillomavirus (45,46), human immunodeficiency virus 1 (47,48), and herpesviruses (46,49). Therefore, the action of the Tip60 complex, including p400, appears to have a positive or negative effect on virus infection depending on the biological system and virus.
We further show that p400 is required for the activity of the exo-siRNA pathway. Our results are consistent with a previous study in D. melanogaster-derived S2 cells in which Domino was identified as a regulator of the siRNA pathway (42). As the exo-siRNA FIG 5 p400 knockdown leads to reduced silencing efficiency. RNA silencing activity in the presence of p400 knockdown was determined using a sensor assay. Aag2 cells were transfected with plasmids constitutively expressing firefly (FFLuc) or Renilla luciferase and dsRNA targeting either p400 or eGFP as a control. At 24 h post-initial transfection, dsRNA against FFLuc or eGFP was transfected. Silencing activity was assessed by measuring the relative levels of FFLuc 24 h post-second transfection after normalization to RLuc (internal transfection control) levels. The level of silencing was calculated relative to respective control eGFP dsRNA-transfected samples (dseGFPϩdseGFP as a control for condition dseGFPϩdsFFLuc, and dsp400ϩdseGFP as a control for condition dsp400ϩdsFFluc), which were set to 1. Bars show the means from 3 independent experiments (3 wells in each independent experiment, average of three wells per condition/experiment used for statistical analysis). Error bars show the standard error of mean. Significance was determined by an unpaired t test to determine the P value (P ϭ 0.034; *, P Ͻ 0.05).

FIG 6
Effect of p400 knockdown on ago-2 and dcr-2 transcript levels in A. aegypti females. (A) ago-2 transcript levels in whole dsLacZ-or dsp400-injected females 3 days after an SFV4-infected blood meal was analyzed by RT-qPCR. Normalized expression-per-sample values were obtained as described by Taylor et al. (78) in order to obtain normalized expression values, relative to the ribosomal S7 transcript as a reference, with a geomean of 1 for the dsLacZ control group. Bars show the geomean and 95% confidence intervals from n ϭ 73 dsLacZ and n ϭ 66 dsp400 females of a single experiment, also used in Fig. 3. This experiment is representative of two independent experiments (second replicate shown in Fig. S1). The results were analyzed using a Mann-Whitney test using log 2 -normalized expression values. ago-2 transcript levels are significantly reduced in dsp400-injected females compared to those in dsLacZ-injected ones (mean fold change, 1.2; geomean fold change, 1.1; P ϭ 0.0197; *, P Ͻ 0.05). (B) Transcript levels of p400, ago-2, and dcr-2 in whole non-blood-fed (NBF) females 4 days after injection with dsLacZ or dsp400 were analyzed by RT-qPCR using the 2 ϪΔΔCT method and expression in dsLacZ females as the reference sample. Bars represent the fold change in gene expression (3 independent biological replicates, with pools of 10 females per replicate). Error bars show the minimum and maximum fold change. The results were analyzed using a one-sample t test. p400 and ago-2 transcript levels are significantly reduced in dsp400-injected females compared to those in dsLacZ-injected females (70% reduction, P ϭ 0.0047; and 42% reduction, P ϭ 0.0067; **, P Ͻ 0.01). pathway limits SFV, BUNV, and CHIKV infection in mosquitoes (22,23,35), the action of p400 against these viruses is likely to be mediated at least in part through this antiviral pathway. In D. melanogaster, Domino is required for diptericin induction in the fat body (50). Moreover, homozygous dom mutant D. melanogaster larvae do not contain hemocytes and do not survive after pupariation, illustrating the important role that Domino has in immunity (51). Roles of this protein also extend to cell proliferation/ growth and death (52,53) and maintaining repression of proapoptotic genes (54). In mammals, Tip60 has been reported to be antiviral by suppressing adenovirus gene expression through binding to the immediate early promoter (44). Considering the different roles of p400 previously described, we cannot therefore exclude that p400 could mediate antiviral effects in A. aegypti mosquitoes through other mechanisms in addition to its role in the exo-siRNA pathway.
Consistent with our finding that p400 regulates ago-2 transcript levels, p400/ Domino is a chromatin-remodeling ATPase member of the Tip60 complex involved in transcriptional regulation and chromatin modification in the fly model but also in human cells (43,55,56). In human cells, Tip60/p400 complexes catalyze the incorporation of the histone variant H2A.Z within chromatin, including promoter regions to modulate gene expression in response to diverse cellular cues (57). As Ago-2 is a main effector in the exo-siRNA pathway, the action of p400 on this pathway activity is likely to be mediated through the regulation of ago-2 expression. This could also be the reason why Domino is required for the Ago-2-dependent RNA interference (RNAi) activity in D. melanogaster-derived S2 cells (42). While p400 regulates ago-2 transcript levels, p400 knockdown does not have any significant impact on the transcript levels of another important exo-siRNA pathway component, dcr-2. Previous reports have shown that Ago-2 is antiviral against SFV, BUNV, and CHIKV (22,23,35), while we and others have shown that reduced or absent Ago-2 activity does not result in an increase in ZIKV replication (34,58,59). Altogether, this could explain why p400 is antiviral against SFV, BUNV, and CHIKV, while we could not detect any significant effect against ZIKV. As previously discussed, p400 is an ATPase part of the Tip60 chromatin remodeling complex (43,55,56). Although the effect of p400 knockdown was not significant on ZIKV, the slight increase in ZIKV expression may indicate that p400 could regulate the expression of other antiviral genes in addition to ago-2. Outside antiviral responses, p400 knockdown could even disturb virus replication in other ways, such as changes in the expression of host factors.
We found that p400 was expressed across various tissues in non-blood-fed females, with the strongest expression in ovaries. In the germarium, the protein is detected in both the germ line and somatic cells, more predominantly in the germ line stem cells and developing secondary follicle. In the primary previtellogenic follicle, p400 is slightly expressed in some somatic epithelial cells. This is similar to the ovarian expression pattern of the two alternative splicing isoforms, DomA and DomB, in D. melanogaster adult females (60), with an expression in the developing cysts within the germarium and little to no expression at stages 7 to 8 (corresponding to the primary follicle arrested at the previtellogenic stage in mosquitoes). In the fruit fly, Domino is required for oogenesis, more particularly for somatic and germ line stem cell maintenance as well as cystocyte differentiation (56,61,62). In humans, there are two homologues of Domino, p400 and SRCAP. Interestingly, SRCAP can rescue the female sterility of hypomorphic dom alleles in D. melanogaster (63), showing functional conservation of these SWR1-like remodelers during evolution. Due to the similar localization in ovaries of Aedes and Drosophila spp., it is likely that p400 in mosquitoes and Domino in flies exert similar functions during oogenesis. The protein p400 is also expressed in the large nucleus of the ovarian sheath cells, as well as in the nucleus of the cells forming the crop, suggesting a role in transcriptional regulation in these cells. We also found p400 expressed in tracheae surrounding various tissues, including salivary glands, digestive tract, and ovaries. Interestingly, the von Hippel-Lindau tumor suppressor gene (VHL) regulates the expression of p400 posttranscriptionally in mouse embryonic fibroblasts to prevent senescence (64). In D. melanogaster, VHL regulates tracheal branch migration and lumen formation through endocytosis (65). The presence of p400 in the mosquito tracheal system could indicate that it is involved in tracheal morphogenesis. As arboviruses can infect tracheal cells in A. aegypti (66), p400 could also mediate antiviral protection in these cells. Consistent with its role in hemocyte proliferation and differentiation in the fly model (51), p400 is expressed in hemocytes in A. aegypti females, with the protein localized in the nucleus of undifferentiated prohemocytes and in the nucleus and cytoplasm of differentiated hemocytes. In D. melanogaster, hemocytes control viral infections by clearing virus-infected cells by phagocytosis (67,68) and by providing an Ago-2-dependent protection to naive cells (69), resulting in higher viral loads in flies depleted of hemocytes. Consequently, p400 could indirectly control viral loads in A. aegypti due to a role in immune cell differentiation. Hemocytes can be infected by viruses in vivo in mosquitoes (27,70); therefore, the antiviral action of p400 could be directly mediated through Ago-2, though we cannot exclude other mechanisms. In the absence of genetic tools to drive p400 knockdown in a tissue-specific manner in mosquitoes, it is difficult to conclude with certitude in vivo in which tissues/cells p400 is antiviral.
The exo-siRNA pathway is one of the key antiviral immune pathways in mosquitoes (8,16,19). Despite the importance of this pathway in antiviral defense, there is a considerable lack of understanding regarding how the pathway is regulated. Here, we show a role of p400 in the regulation of ago-2 transcript levels (and subsequently, the activity of the exo-siRNA pathway) and the antiviral response in A. aegypti. In addition to extending the diverse functions of p400, it could explain the previous observations of Domino's antiviral activity in D. melanogaster cells and its role in regulating the exo-siRNA pathway. Further studies will need to determine if and how exo-siRNA pathway activity is differentially regulated across tissues in A. aegypti and whether this is relevant for the transmission of arboviruses. In addition, more work will be necessary to understand the mechanism by which p400 is antiviral. This could include interaction studies with other cellular proteins, as well as promoter binding studies to assess the wider mode of action of this protein, which might explain how ago-2 levels are controlled by p400. The tools used in this study (primers, dsRNA, and antibody) target the four p400 splicing isoforms predicted in A. aegypti genome annotation (AaegL5). As there are two splicing isoforms, DomA and DomB, in D. melanogaster with distinct functions during oogenesis (60), it would be interesting to identify which isoform is involved in the exo-siRNA pathway regulation and antiviral activity in these species. Future work on p400 in A. aegypti will have to investigate these key questions on its mode of action and activity across tissues.
Mosquito rearing. A. aegypti Liverpool strain (a gift from E. Devaney, University of Glasgow, UK) was reared at 28°C and 80% humidity with a 12:12 light photoperiod. Larvae were reared in water and fed on dry cat food from larvae hatching to the pupal stage. The emerging adult mosquitoes were removed and put in cages with unlimited access to 10% (wt/vol) sucrose solution. For mass rearing, female mosquitoes were fed on heparinized rabbit blood (Orygen Antibodies Ltd.) using a Hemotek system (Hemotek Ltd., UK).
Reporter virus stocks, plaque assay titration, and infection. Luciferase expressing SFV4 contains FFLuc inserted between nsP3 and nsP4 using duplicated nsP2 cleavage sites, as previously described (25,72). pCMV-SFV4(3H)-FFLuc was electroporated into BHK-21 cells and incubated at 37°C until extensive cell death was visible. The cell supernatant was harvested and clarified by centrifugation. The resulting virus stock was stored at -80°C. CHIKV expressing FFLuc between the nonstructural and structural open reading frames under a duplicated subgenomic promoter, CHIKV-2SG-FFLuc (25), was rescued by transfection of in vitro-transcribed RNA into BHK-21 cells. For this, pSP6-ICRES1-2SG-FFLuc was linearized and in vitro transcribed using the SP6 MEGAscript kit (Thermo Fisher Scientific), along with Ribo m7G Cap Analog (Promega), followed by transfection into BHK-21 cells using Lipofectamine 2000 (Thermo Fisher Scientific), following the manufacturer's instructions. The supernatant was collected, clarified by centrifugation, and stored at -80°C. SFV4(3H)-FFLuc and CHIKV-2SG-FFLuc were titrated on BHK-21 cells by making 10-fold serial dilutions and overlaying the inoculum with 0.6% Avicel in minimum essential medium (MEM) containing 2% FBS and incubating at 37°C for 72 h. Cells were fixed using 10% formalin (Sigma) and stained with 0.1% toluidine blue. NLuc-expressing BUNV was grown and titrated as described previously (35). ZIKV-NLuc (40) was grown on A549 NPro cells cultured in DMEM supplemented with 2% FBS and blasticidin (2 g/ml) for 5 days at 37°C with 5% CO 2 . Virus supernatant was collected and cleared by centrifugation at 3,220 ϫ g for 10 min, followed by storage at -80°C. For virus titration, cell monolayers of A549 NPro were infected with serially diluted ZIKV-NLuc in DMEM with 2% FBS and incubated with an overlay consisting of MEM with 4% FBS, 4% HEPES, and 1.2% Avicel for 5 days. Infected cells were fixed with 4% formaldehyde and stained with 0.2% bromophenol to visualize plaques. Aag2 cells were infected by removing the culture medium, overlaying with 200 l virus-containing inoculum, and incubating at 28°C for 1 h. The inoculum was removed and replaced with complete medium, followed by incubation at 28°C.
SFV production for oral infection of mosquitoes. SFV4 was produced from plasmid pCMV-SFV4, as described previously (73). Briefly, the plasmid was transfected using Lipofectamine 2000 (Thermo Fisher Scientific) into BHK-21 cells grown in GMEM with 2% FBS and 10% TPB at 37°C with 5% CO 2 . The virus was titrated by a plaque assay on BHK-21 cells, as described above.
Plasmids. The pIZ-FLuc and pAcIE1-RLuc plasmids have been previously described (74,75). The plasmid pPUb-V5MBP was synthesized by subcloning maltose-binding protein (MBP) from pcDNA-DEST40-MBP-hDVR (75) into the mosquito expression vector pPUb (29), based on an expression construct containing the A. aegypti polyubiquitin promoter (76). The plasmid pUC57-p400 was synthesized (GenScript) and contains the A. aegypti coding sequence, as indicated in VectorBase with accession number AAEL001440 (assembly AaegL1). In the AaegL5 assembly, AAEL001440 was changed to AAEL027494 (transcript identifiers [IDs] RA to RD; all transcripts targeted by dsp400). The p400 coding sequence was further subcloned into the pPUb expression vector. The V5 tag was added to the N terminus of pPUb-MBP and pPUb-p400 clones using the Infusion cloning technique (Clontech) to give pPUb-V5MBP and pPUb-V5p400, respectively. dsRNA synthesis for in vitro experiments. RNA was extracted from Aag2 cells using TRIzol (Thermo Fisher Scientific), following the manufacturer's guidelines. One microgram of RNA was reverse transcribed using SuperScript III (Thermo Fisher Scientific) and an oligo(dT)15 primer (Promega), according to the manufacturer's instructions. Unique portions of candidate genes were amplified from cDNA with primers containing T7 RNA polymerase minimal promoter sequence overhangs using GoTaq G2 Flexi polymerase (Promega). PCR products were sequenced for target verification before the production of dsRNA. For the production of dseGFP (used as a control), specific primers with T7 RNA polymerase promoter sequences were used to amplify a unique portion of eGFP from a plasmid template containing the eGFP-encoding gene. All primer sequences are listed in Table S1. The MEGAscript RNAi kit (Thermo Fisher Scientific) was used to synthesize dsRNA from the PCR fragments, according to the manufacturer's guidelines.
dsRNA synthesis and purification for in vivo experiments. Total RNA was extracted with TRIzol (Thermo Fisher Scientific) from whole NBF A. aegypti females, according to the manufacturer's instructions, including DNase treatment (Turbo DNase; Ambion). cDNA was generated from 1 g of total RNA using SuperScript III reverse transcriptase (Thermo Fisher Scientific) and oligo(dT)15. cDNAs were further amplified with KOD Hot Start master mix (EMD Millipore). p400-specific primers (Table S1) with a T7 RNA polymerase promoter sequence were used to amplify a p400-derived fragment (same primers used for in vitro and in vivo experiments) and further purified using the QIAquick gel extraction kit (Qiagen). After sequencing, the PCR product was used as the template for a second PCR using the same primers and polymerase. For the production of dsLacZ (used as control dsRNA), specific primers with T7 RNA polymerase promoter sequences were used to amplify a lacZ-derived fragment from plasmid template Drosophila/act5C-␤Gal (stock number 1220 obtained from DGRC) containing the Escherichia coli lacZ gene. dsRNAs were synthesized and purified using the MEGAscript RNAi kit (Thermo Fisher Scientific), according to the manufacturer's instructions. dsRNA was then purified and concentrated to 10 g/l in nuclease-free water using sodium acetate (3 M; Ambion) and ethanol precipitation.
Transfection of BHK-21 cells. BHK-21 cells were plated at a density of 3 ϫ 10 6 cells per well in 6-well plates. One microgram of pPUb-V5MBP or pPUb-V5p400 was transfected the following day using Dharmafect2 (Dharmacon), following the recommended protocol. Cells were collected 24 hpt by trypsinization and plated in slides (ibidi) for immunostaining.
Transfection of mosquito cells. Aag2 cells were plated at a density of 1.7 ϫ 10 5 cells per well in a 24-well plate. dsRNA or plasmid DNA was transfected the following day with Lipofectamine 2000 (Thermo Fisher Scientific), using the supplied protocol. The medium was replaced at 4 to 5 hpt. Knockdown experiments were performed by transfecting 100 ng dsRNA into Aag2 cells. Plasmid sensor assays were performed by transfecting 50 ng pIZ-FLuc and 8 ng pAcIE1-RLuc and 100 ng of either eGFP-targeting or p400-targeting dsRNA. At 24 h post-initial transfection, cells were transfected again with 10 ng FFLuc-or eGFP-targeting dsRNA. The FFLuc and RLuc activities were measured at 24 h post-second transfection.
Luciferase assay. Aag2 cells were lysed in 1ϫ passive lysis buffer (Promega) and lysed by rocking at room temperature for 20 min or stored in -20°C immediately. FFLuc and Renilla luciferase activities were measured using the dual-luciferase assay system (or Steady-Glo luciferase assay system in the case of CHIKV-2SG-FFLuc) (Promega); NLuc activities were measured by using the Nano-Glo luciferase assay systems (Promega) (ZIKV-NLuc and BUNV-NLuc). All measurements were carried out on Glomax luminometers (Promega).
Total RNA extraction and RT-qPCR from cells. For analysis of p400 knockdown efficiency, cells were lysed in TRIzol reagent (Thermo Fisher Scientific) at 24 hpt and extracted according to the manufacturer's instructions. Reverse transcription (RT) was performed using the SuperScript III enzyme kit (Thermo Fisher Scientific) from 1 g of RNA in a final volume of 20 l. Quantitative PCR (qPCR) was carried out using Fast SYBR green master mix (Applied Biosystems) on a 7500 Fast machine (Applied Biosystems). The primers used are listed in Table S1. Data were analyzed with the 7500 Software v2.0.6, and transcript expression relative to the ribosomal S7 as a reference was calculated according to the 2 ϪΔΔCT method (77). The dseGFP control was set to 1, and dsp400 samples were normalized to dseGFP. Data from 3 independent biological replicates were analyzed using a one-sample t test (Prism software).
dsRNA injection into mosquitoes. At 1 to 2 days after emergence, cold-anesthetized female mosquitoes were injected with dsRNA into their thorax using a nanoinjector (Nanoject II; Drummond Scientific) with 2 g of dsRNA (dsp400 or dsLacZ) in 414 nl of injection solution. To increase knockdown efficiencies, Cellfectin II transfection reagent (Thermo Fisher Scientific) was added to the injection solution. Briefly, Cellfectin was mixed with Schneider's Drosophila medium (1:1 [vol/vol]; Thermo Fisher Scientific); this mix was added to dsRNA solution (1:1 [vol/vol]) previously adjusted with Schneider's medium to give 2 g of dsRNA per female/injection volume. The injection solution was then incubated for 15 min at room temperature before injection.
Mosquito sampling, dissection, and hemocyte perfusion. To assess the effect of p400 knockdown on ago-2 transcript levels in whole NBF females and in tissues, female mosquitoes injected with dsRNA (targeting control or p400 transcripts) were sacrificed at 4 days pdi. Whole females (pools of 10 females) and dissected tissues (pools of 25 digestive tracts or ovaries, pools of 60 salivary glands, and pools of perfused hemocytes from 70 females) were further stored in tubes on dry ice before TRIzol was added for RNA extraction. Digestive tracts include the posterior part of the esophagus, dorsal diverticula, crop, midgut, Malpighian tubules, and hindgut. The tissues were carefully dissected from 5-day-old NBF females in RNase-free 0.05% (vol/vol) phosphate-buffered saline with Tween 20 (PBS-T) and either placed into tubes on dry ice before TRIzol was added for subsequent analysis of p400 and ago-2 tissue levels by qPCR or added into 0.05% PBS-T on wet ice for fixation and immunostaining. Hemocytes were collected by perfusion. Briefly, the last segment of the abdomen was cut, and mosquitoes were then injected in the thorax with 0.01% PBS-T using a glass capillary mounted on a syringe. For RT-qPCR experiments, PBS-T-diluted hemocytes were collected in tubes on ice and centrifuged at 375 ϫ g for 15 min at 4°C. The supernatant was gently removed before adding TRIzol. For immunostaining, hemocytes were perfused on slides (ibidi; 5 females per well, 4 wells per replicate). The slides were left for 30 min at 28°C before removal of PBS-T and fixation.
Antibody production. The Peptide Supplied Polyclonal Antibody Package (mouse; catalog no. SC1046) available from GenScript USA, Inc. was used to produce the p400 antibody. Peptide optimization was performed using the OptimumAntigen design program provided by GenScript. The peptide GMN KPKAIQDQNTSC (present in proteins AAEL027494-RA to AAEL027494-RD) was selected and conjugated to keyhole limpet hemocyanin (KLH) for immunization into five BALB/c mice. Antiserum from mouse 5 was used in these experiments. The specificity of the antibody was confirmed by colocalization of p400 and V5 antibody signals in BHK-21 cells transfected with pPUb-V5p400 (Fig. S3). Further details on antibody production are available upon request.
Blood meal infection of mosquitoes with SFV. At 4 days pdi, A. aegypti females were allowed to feed for 30 min on blood meal containing SFV4. Fresh rabbit blood (Orygen Antibodies Ltd.) was washed with PBS to remove white blood cells and serum. PBS was then added to the red blood cell fraction to return to the initial blood volume. The infectious blood meal was prepared with 2/3 of washed blood and 1/3 of SFV-containing culture medium to give a final titer of 5.10 7 PFU/ml, supplemented with 2 mM ATP. Only engorged females were kept further in the presence of 10% sucrose at 28°C and 80% humidity. At 3 days postfeeding, females were sampled individually before RNA extraction.
Total RNA extraction and RT-qPCR from mosquitoes. Dissected tissues or whole females were homogenized in TRIzol reagent with glass beads using the Precellys 24 homogenizer (Bertin Instruments). RNA from tissues, whole females, and hemocytes was extracted according to the manufacturer's instructions, except that 1-bromo-3-chloropropane was used instead of chloroform, and DNase (Turbo DNase; Ambion) treatment was carried out. Reverse transcription (RT) was performed using the Moloney murine leukemia virus (MMLV) retrotranscriptase (Promega) from 1 to 2 g of RNA in a final volume of 40 l. cDNA was aliquoted and further stored at -20°C until qPCR was carried out using Fast SYBR green master mix (Applied Biosystems) on a 7500 Fast machine (Applied Biosystems). The primers used are listed in Table S1. Data were analyzed with the 7500 Software v2.0.6.
For analysis of p400 expression, transcript expression relative to the ribosomal S7 transcript as a reference was calculated according to the 2 ϪΔΔCT method (77). The hemocyte sample was set to 1, and other samples were normalized to the hemocyte sample. Data were obtained from 3 independent biological replicates. For analysis of p400 knockdown efficiency and the effect on ago-2 levels in whole NBF females, transcript expression relative to the ribosomal S7 transcript as a reference was calculated according to the 2 ϪΔΔCT method (77). The dsLacZ control was set to 1, and dsp400 samples were normalized to dsLacZ. Data from 3 independent biological replicates (pool of 10 females per replicate) were analyzed by a one-sample t test (Prism software). For analysis of p400 knockdown and the effect on ago-2 levels and SFV infection, data were analyzed as described by Taylor et al. (78) in order to obtain normalized expression values, relative to the ribosomal S7 transcript as a reference, with a geometric mean (geomean) of 1 for the dsLacZ control group (two independent experiments; n ϭ 73 females for dsLacZ and 66 for dsp400 in Fig. 3A and B and 6A; n ϭ 39 females for dsLacZ and 43 for dsp400 in Fig. S1). Log 2 -normalized expression values were analyzed using a Mann-Whitney test (with the Prism software).
Data availability. The data sets generated and analyzed during the current study are available in the University of Glasgow repository at https://doi.org/10.5525/gla.researchdata.705.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.