Silencing of Toll pathway components by direct injection of double-stranded RNA into Drosophila adult flies

Abstract

Double-stranded RNA (dsRNA) gene interference is an efficient method to silence gene expression in a sequence-specific manner. Here we show that the direct injection of dsRNA can be used in adult Drosophila flies to disrupt function of endogenous genes in vivo. As a proof of principle, we have used this method to silence components of a major signaling cascade, the Toll pathway, which controls fruit fly resistance to fungal and Gram-positive bacterial infections. We demonstrate that the knockout is efficient only if dsRNA is injected in 4- or more day-old flies and that it lasts for at least 1 week. Furthermore, we report dsRNA-based epistatic gene analysis via injection of a mixture of two dsRNAs and propose that injection of dsRNA represents a powerful method for rapid functional analysis of genes in Drosophila melanogaster adults, particularly of those whose mutations are lethal during development.

INTRODUCTION

Drosophila melanogaster has been widely used as a model organism to identify genes that are involved in diverse biological processes (1). The publication of the fly genome created a great need for new approaches to do the reverse, namely to disrupt the functions of specific genes whose sequences are known (2). RNA interference (RNAi) is one of these tools (reviewed in 3). In fly embryos, injection of double-stranded RNA (dsRNA) is routinely used to analyze gene function but it is inherently limited to early development (4,5). Although direct injection of dsRNA in adults was previously reported in Caenorhabditis elegans (6), Anopheles gambiae (7) and D.melanogaster (8,9), gene silencing in Drosophila adults is mainly achieved by establishment of transgenic strains expressing hairpin constructs against the targeted genes (1).

Here we establish an efficient dsRNA silencing method by direct injection into adult flies of dsRNA for components of the Toll signaling pathway. This pathway was initially characterized for its role in dorso-ventral development in embryos (reviewed in 10). It was later found that it also exerts an immune function in adult flies (11–13). In this case, Toll is activated after immune challenge by a cleaved form of the polypeptide Spaetzle, leading to the cytoplasmic dissociation of the NF-κB family member Dif from its inhibitor, the IκB-like protein Cactus. This is followed by nuclear translocation of Dif, which controls expression of many genes including that encoding the antifungal peptide Drosomycin (Drs). In the absence of immune challenge, cleavage of Spaetzle is inhibited by the serpin Spn43Ac. Two novel components of this pathway have been recently identified in a cell-culture system: the atypical protein kinase C (DaPKC) and Ref(2)P (14). However, analysis of their function using classical genetics is difficult as mutations in DaPKC are embryonic lethal (15) and Ref(2)P-mutant males are sterile (16). We used direct injection of dsRNA into adult flies to place these components in the Toll pathway in vivo. For that, we developed dsRNA-based epistatic gene analysis via injection of a mixture of two dsRNAs. Our results demonstrate that injection of dsRNA represents a novel powerful method for rapid in vivo functional analysis of genes in D.melanogaster adults, particularly those for which mutations are lethal during development.

MATERIALS AND METHODS

Fly stocks and microbial infection

DD1 flies, carrying two reporter transgenes P[w⁺mC Dipt:: LacZ = pDipt-LacZ] and P[w⁺mC Drom::GFP = pDrs-GFP S65T] (17), were grown on standard medium at 25°C. Four days after dsRNA injection, flies were pricked with a thin tungsten needle previously dipped into a concentrated culture of Gram-positive bacteria, Micrococcus luteus. For natural fungal infection, anesthetized flies were manually shaken for 30 s on a Petri dish containing a sporulating culture of Beauvaria bassiana and kept at 25°C for 48 h (18).

dsRNA preparation and injection into adult flies

Templates for the preparation of dsRNA were PCR-derived fragments sandwiched by two T7 promoter sequences (TAA TAC GAC TCA CTA TAG GGA GAC CAC). The amplified cDNA fragments for each gene were as follows: green fluorescent protein (GFP: nucleotides 35–736, GenBank accession no. L29345), Cactus (nucleotides 1041–1582, GenBank accession no. L04964), dorsal-related immunity factor (Dif: nucleotides 1921–2550, GenBank accession no. L29015), serpin 43Ac (Spn43Ac: nucleotides 57–1473, GenBank accession no. AJ245444), Toll (nucleotides 3289–3866, GenBank accession no. M19969), Drosomycin (Drs: nucleotides 28–260, GenBank accession no. X75595), Drosophila atypical protein kinase C (DaPKC: nucleotides 301–900, GenBank accession no. AF288482) and Ref(2)P (nucleotides 348–655, GenBank accession no. X69829). As a negative dsRNA control we used odd-paired (opa: nucleotides 135–1156, GenBank accession no. S78339). Single-stranded RNAs were synthesized by using the MEGAscript T7 transcription kit (Ambion, Austin, TX) according to instructions. Annealed dsRNAs were ethanol precipitated and dissolved in injection buffer (0.1 mM sodium phosphate, pH 6.8; 5 mM KCl). Using a nano-injector (NANOJECT II; Drummond), 32.2 nl of dsRNA (3 mg/ml) were injected in the thorax of CO₂-anesthetized adult females which were allowed to recover for 4 days before bacterial infection. In double knockdown experiments, 32.2 nl of a 1:1 mix of dsRNAs (3 mg/ml) were injected.

RNA preparation, northern blotting and RT–PCR

Total RNA from 20–25 flies was extracted by TRIzol Reagent (Invitrogen). Twenty micrograms of total RNA were separated by 1% agarose/formaldehyde gel electrophoresis, blotted to nylon membrane (Schleicher & Schuell) and hybridized sequentially with the ³²P-labeled cDNA probes for GFP, Drs and ribosomal protein 49 (rp49) [Ready-To-Go DNA labeling beads (dCTP); Amersham Pharmacia Biotech]. The rp49 was used as an internal loading control. Signal quantification was done by a Fuji Film BAS 2000 Image Analyzer (Fuji Photo Film Co., Tokyo, Japan).

For reverse transcription, 10 µg of total RNA were reverse transcribed with Superscript II RNase H^– reverse transcriptase (Gibco BRL). Single-stranded cDNAs of different dilution were amplified by PCR using recombinant Taq DNA polymerase (Invitrogen). Primers were made to amplify endogenous Toll (nucleotides 3289–3866, GenBank accession no. M19969) for 30 cycles, Dif (nucleotides 1921–2550, GenBank accession no. L29015) for 35 cycles and Actin 5C (nucleotides 213–1205, NCBI accession no. NM-167053) for 25 cycles. Actin 5C was used as an internal control.

RESULTS AND DISCUSSION

Efficiency of dsRNA silencing in Drosophila adults is age specific

To analyze the efficiency of gene silencing by direct injection of dsRNA into adult flies, we prepared dsRNAs corresponding to several components of the Toll pathway. As a read-out of Toll activation we have used DD1 flies, which carry two reporter transgenes, Drosomycin-GFP (Drs-GFP) and Diptericin-Lac Z, on the X chromosome (17). In this study, we monitored the expression of a Drs-GFP reporter transgene. We first injected into the thorax of females dsRNAs directed against an inducible GFP reporter gene (dsGFP) or against the unrelated gene opa (dsOPA). The injected flies were allowed to recover for 4 days and then challenged with Gram-positive bacteria, M.luteus, to induce the expression of the reporter Drs-GFP gene (17). We monitored the expression of GFP 1 day after bacterial challenge, when the maximal level of fluorescence is observed (11). dsGFP flies displayed a strong reduction in Drs-GFP expression compared with both buffer and dsOPA control flies (Fig. 1A). The efficiency of the knockout was confirmed at the mRNA level (Fig. 1B). We observed that the M.luteus-induced expression of Drs-GFP was similar in buffer- and dsOPA-injected flies, suggesting that dsRNA per se does not interfere with the expression of antimicrobial peptides. Interestingly, when we used flies of different ages, we noticed a high heterogeneity in Drs-GFP silencing. To determine the age dependency of gene silencing, we injected dsGFP into 1–10-day-old flies and monitored the Drs-GFP expression after bacterial challenge (Fig. 1C and D). We could not detect any significant decrease in Drs-GFP expression in flies injected with dsGFP earlier than 4 days after eclosion. Silencing became more efficient as flies grew older and was at its maximum at days 5 and 6 (5-fold reduction in Drs-GFP expression). The same age-dependent efficiency of the knockout was observed for other dsRNAs: dsDRS, dsTOLL and dsSPN43Ac (data not shown). This effect had not been detected in the mosquito A.gambiae, where the injection of specific dsRNA in newly hatched adults fully inhibited the induction of the antimicrobial peptide gene Defensin (7). Differential age-dependent efficiency of dsRNA knockout may reflect physiological differences between these two insect species.

Figure 1.

dsRNA silencing of the Drs-GFP transgene. Six-day-old DD1 females were injected with dsRNA against the Drs-GFP transgene (dsGFP), the immune-unrelated gene oda (dsOPA) or injection buffer (buffer) and 4 days later were either infected with Gram-positive bacteria (M.luteus) or left unchallenged as a control (N.I.). (A) GFP expression was analyzed 1 day later using fluorescence microscopy. (B) The expression of Drs-GFP (GFP), Drs and ribosomal protein gene rp49 as a loading control, was monitored by northern blotting. (C) dsGFP was injected into 1–10-day-old flies and the level of Drs-GFP mRNA after M.luteus infection was monitored. (D) The graph represents relative percentages of Drs-GFP expression normalized by rp49 in two independent experiments. 100% of expression corresponds to Drs-GFP expression in buffer-injected flies.

Duration of the effect of gene silencing

We next examined the duration of dsRNA knockout efficiency by targeting Cactus, a gene encoding the ankyrin-repeat protein that negatively regulates the Toll pathway in adult flies (19,20). Loss-of-function mutations in Cactus lead to a signal-independent Drs expression (11). Therefore, to test the duration of dsRNA knockout in flies, we recorded Drs expression in DD1 flies over a period of 10 days after dsCACT injection (Fig. 2). Drs expression in dsCACT reached a maximum 2–4 days post-dsRNA injection, and gradually decreased on the following days. One day after injection we observed a significant level of Drs expression in buffer-treated flies, which we attribute to the septic injury caused by the injection (Fig. 2). These results allowed us to determine the following standard conditions to analyze dsRNA knockout phenotypes: injection of dsRNA into 6-day-old flies and analysis of phenotypes 4 days later.

Figure 2.

Time course of Drs expression in dsCACT-injected flies. DD1 females were injected with buffer or dsCACT and Drs expression was monitored over 10 days. The graph represents results of two independent experiments and shows relative Drs expression normalized by rp49. The lower panel shows representative northern blotting data for Drs and rp49.

dsRNA silencing of components of the Toll signaling pathway

We extended our study to other components of the Toll pathway. The serine protease inhibitor (serpin) Spn43Ac is a negative regulator of the extracellular proteolytic cascade, which regulates Toll activation, and mutations in Spn43Ac lead to a challenge-independent expression of Drs (21). Flies injected with dsSPN43Ac displayed phenotypes similar to Spn43Ac mutants: constitutive activation of Drs-GFP and Drs (Fig. 3A and B) and necrotic patches (Fig. 3A, arrow) (21,22). In contrast to Spn43Ac mutants, which die within 4 days after hatching, we did not observe any lethality phenotype for the dsSPN43Ac-injected flies, indicating that this lethality does not directly result from the constitutive activation of the Toll pathway. These results demonstrate that this extracellular regulator of the Toll pathway was efficiently silenced by dsRNA injection.

Figure 3.

dsRNA silencing of other components of Toll pathway. Four days after dsOPA or dsSPN43Ac injection into 6-day-old DD1 females. Efficiency of silencing was analyzed by fluorescence microscopy (A) or by northern blotting using probes specific for Drs and ribosomal protein rp49 (B). The white arrow in (A) points to a necrotic spot appearing after dsSPN43Ac injection. (C) Level of Drs expression 1 day after M.luteus challenge of buffer-, dsTOLL-, dsDIF- or dsDRS-injected flies. The graph shows relative percentages of Drs expression normalized by rp49. (D) RT–PCR analysis of dsRNA silencing of Dif (dsDIF) and Toll (dsTOLL). Actin 5C was used as an internal control. (E) Four days after dsOPA, dsPKC, dsREF(2)P, dsPKC+dsREF(2)P or dsDRS injection, flies were infected with the Gram-positive bacterium M.luteus or the fungus B.bassiana. The expression of Drs and rp49 was analyzed by northern blotting 1 day after infection with M.luteus and 2 days after infection with B.bassiana.

As both Cactus and Spn43Ac are negative regulators of the Toll pathway, we further attempted to silence positive regulators, namely Toll and Dif, as well as the Drs gene itself. Flies were injected with specific dsRNAs or buffer as a control, challenged with M.luteus and Drs induction was monitored. All dsRNA-treated flies showed a decrease in Drs expression compared with control flies: 26, 47 and 62% for dsTOLL, dsDIF and dsDRS, respectively (Fig. 3C). The sequence-specific silencing of Toll and Dif was confirmed by RT–PCR (Fig. 3D). However, this semi-quantitative method did not allow us to determine precisely the efficiency of the knockouts, leaving open the possibility that the differences in the remaining activity of Drs are due to incomplete gene silencing.

We next extended our analysis to two new components of the Toll pathway, the atypical protein kinase C (DaPKC) and Ref(2)P, which were recently implicated in the Toll signaling pathway by in vitro experiments with a Drosophila cell line (14). DaPKC is involved in regulation of many basic developmental functions in Drosophila and loss-of-function mutants of this gene are embryonic lethal (15). The Ref(2)P gene product is required for male fertility and, later in development, mediates fly resistance to a sigma rhabdovirus (16,23). We first analyzed the effect of the knockout of these genes on the expression of the Drs gene after infection with M.luteus or B.bassiana. Single knockouts of either molecule did not significantly affect Drs expression. However, simultaneous silencing of both genes resulted in a reduction of Drs expression after infection by both microorganisms. These results suggest that the single knockouts of either DaPKC or Ref(2)P genes are not sufficient to block Drs induction after infection and that both molecules are required for full pathogen-induced Drs expression (Fig. 3E). Our data confirm the results of the earlier in vitro studies that place these genes in the Toll pathway.

Epistatic gene analysis using dsRNA

Epistatic analysis provides a powerful tool to dissect signaling pathways. We attempted to use dsRNA silencing to perform such an analysis in vivo. For that, we first injected flies with a mixture of dsSPN43Ac and dsDIF, targeting the negative regulator gene Spn43Ac and the downstream transcription factor gene Dif. We observed that the challenge-independent induction of Drs and the necrotic phenotype in the dsSPN43Ac-treated flies were abolished by the co-injection of dsDIF but not by the co-injection of the dsOPA control (Fig. 4 and data not shown).

Figure 4.

dsRNA-based epistatic analysis in adult flies. Drs expression in double knockouts (DKO) was monitored 4 days after concomitant injection of dsRNA corresponding to the serpin Spn43Ac gene (dsSPN43Ac) in combination with dsRNA against either a control unrelated gene (dsOPA), Dif (dsDIF), the atypical protein kinase C (dsPKC) or Ref(2)P [dsREF(2)P]. Levels of Drs expression in the double knockouts were compared with those of buffer-injected flies before (C) or after challenge with M.luteus (M.l.).

We then used co-injection of dsSPN43Ac and dsPKC or dsREF(2)P to place these genes in the Toll pathway in vivo. Co-injection of dsSPN43Ac with either dsPKC or dsREF(2)P dramatically reduced the level of Drs-GFP expression (Fig. 4). However, we noticed that the knockout of DaPKC was not as efficient in counterbalancing the dsSPN43Ac phenotype as that of REF(2)P. These results demonstrate that in the context of the dsSPN43Ac knockdown, DaPKC and Ref(2)P are required for full induction of the Toll pathway in vivo. It is unclear why single gene silencing is sufficient to decrease Drs expression in the dsSPN43Ac background, but not after bacterial or fungal infection (Fig. 4). Since the level of challenge-independent Drs expression in dsSPN43Ac-treated flies is 40–50% weaker than after infection with Gram-positive bacteria (data not shown), it is conceivable that dsSPN43Ac-treated flies represent a more sensitized background for the analysis of PKC and Ref(2)P genes by dsRNA. In addition, bacterial and fungal infections are known to induce multiple signaling cascades, which can compensate for silencing of each, but not of both, of these genes. In either way, our results show that dsRNA co-injections can be used to dissect signaling pathways to gain a rapid insight into gene function that can be followed by a more laborious analysis of double mutant phenotypes by means of forward genetics.

In conclusion, we demonstrate that the direct injection of dsRNA into adult flies efficiently disrupts gene function, provided that the flies are at least 5 days old, and this effect lasts for at least 10 days. We successfully silenced components of the Toll signaling cascade, representing positive and negative regulators that encode secreted, intracellular or transmembrane proteins. Our results show that dsRNA injection into adult flies allows the analysis of knockout phenotypes for genes that are essential for Drosophila development and, as such, are not amenable for forward genetics. We also demonstrate that dsRNA co-injections can be used for rapid epistatic analysis of double knockouts. Finally, our finding that the degree of silencing is dependent on the age of the fly provides important information for development of dsRNA technology in other insect species, in particular vectors of insect-borne diseases.