lable at ScienceDirect

Insect Biochemistry and Molecular Biology 39 (2009) 824–832

Contents lists avai

Insect Biochemistry and Molecular Biology

journal homepage: www.elsevier .com/locate/ ibmb

Ingested double-stranded RNAs can act as species-specific insecticides

Steven Whyard*, Aditi D. Singh, Sylvia WongDepartment of Biological Sciences, University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada

a r t i c l e i n f o

Article history:Received 23 July 2009Received in revised form28 August 2009Accepted 28 September 2009

Keywords:RNA interferenceDrosophilaInsecticidevATPaseg-tubulin

* Corresponding author. Tel.: þ1 204 474 9418; faxE-mail address: whyard@cc.umanitoba.ca (S. Why

0965-1748/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.ibmb.2009.09.007

a b s t r a c t

A serious shortcoming of many insecticides is that they can kill non-target species. To address this issue,we harnessed the sequence specificity of RNA interference (RNAi) to design orally-delivered double-stranded (ds) RNAs that selectively killed target species. Fruit flies (Drosophila melanogaster), flourbeetles (Tribolium castaneum), pea aphids (Acyrthosiphon pisum), and tobacco hornworms (Manducasexta) were selectively killed when fed species-specific dsRNA targeting vATPase transcripts. We alsodemonstrate that even closely related species can be selectively killed by feeding on dsRNAs that targetthe more variable regions of genes, such as the 30 untranslated regions (UTRs): four species of the genusDrosophila were selectively killed by feeding on short (<40 nt) dsRNAs that targeted the 30 UTR of theg-tubulin gene. For the aphid nymphs and beetle and moth larvae, dsRNA could simply be dissolved intotheir diets, but to induce RNAi in the drosophilid species, the dsRNAs needed to be encapsulated inliposomes to help facilitate uptake of the dsRNA. This is the first demonstration of RNAi followingingestion of dsRNA in all of the species tested, and the method offers promise of both higher throughputRNAi screens and the development of a new generation of species-specific insecticides.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Double-stranded RNA (dsRNA)-mediated gene silencing,commonly referred to as RNA interference (RNAi), is becominga widely used functional genomics tool in insects to ascertain thefunction of the many newly identified genes accumulating fromgenome sequencing projects (Hannon, 2002; Kuttenkeuler andBoutros, 2004; Chen et al., 2007). The basic components of the RNAiprocess, namely the endonuclease Dicer, which first chops longdsRNAs into short interfering RNAs (siRNAs), and the RNA-inducedsilencing complex (RISC), which facilitates the targeting andendonucleolytic attack on mRNAs with sequence identity to thedsRNA, are evolutionarily conserved across virtually all eukaryotictaxa, and consequently, RNAi could be readily applied to any insectspecies. The main challenge for most insect molecular biologiststoday is to find easy and reliable methods of dsRNA delivery.

Direct injection of the dsRNA into target tissues or develop-mental stages is still the most common method of delivering dsRNAto insects (Amdam et al., 2003; Bettencourt et al., 2002; Bucheret al., 2002; Gatehouse et al., 2004; Misquitta and Paterson, 1999;Rajagopal et al., 2002; Tomoyasu and Denell, 2004). Injection of thedsRNA into an insect’s hemolymph can provide transient knock-down of target genes, as the dsRNA is able to circulate through the

: þ1 204 474 7588.ard).

All rights reserved.

open circulatory system and can enter cells of tissues distributedacross the body. The mechanism by which the dsRNA enters cells ininsects may vary in different species. SID-1, an RNA channeltransporter first discovered in the nematode Caenorhabditis elegans(Winston et al., 2002), appears to facilitate uptake of dsRNA intocells, and this protein appears to be conserved in many but not allinsect taxa (Gordon and Waterhouse, 2007). An ortholog for SID-1has not been found in any members of the order Diptera, and yetdsRNA can enter cultured Drosophila melanogaster S2 cells (Caplenet al., 2000; Clemens et al., 2000; Bettencourt-Dias and Goshima,2009) or into tissues of fruit flies injected with dsRNA (Dzitoyevaet al., 2001). In this species, and quite possibly others, the primarymode of dsRNA entry into cells may be mediated by receptor-mediated endocytosis (Saleh et al., 2006; Ulvila et al., 2006).

In nematodes (Tabara et al., 1998) and flatworms (Orii et al.,2003), it is possible to induce RNAi simply by soaking the animals ina solution of dsRNA. Eaton et al. (2002) observed gene silencing inDrosophila dechorionated embryos soaked in dsRNA, although noother researchers have since used this technique to deliver dsRNA.Biolistics has also been shown to be effective at delivering dsRNAinto Drosophila embryos (Yuen et al., 2008), but this technique hasnot yet been used for any other species. Oral delivery of dsRNA wasfirst demonstrated in C. elegans (Timmons et al., 2001), and has sincebeen adapted for a limited number of insect species. SupplyingdsRNA in artificial diets resulted in knockdown of targeted genes inlight brown apple moth larvae (Epiphyas postvittana; Turner et al.,2006), the hemipteran Rhodnius prolixus (Araujo et al., 2006), the

Table 1PCR Primers used to amplify 185 bp portions of the E-subunit of the vATPase genefrom each species.

Species Primer sequences (F: forward primer,R: reverse primer)

Genbank accession

D. melanogaster F: AAGGCCCAGATCAATCAGAA NM_169073R: CCGAAAAGTGCGTTACGAA

M. sexta F: AAGGCCAAGATCAAGAAGGA X67131R: CCGAACAGCGCGTTACGGA

T. castaneum F: AGGGACGCCACTGGTAAAGACGTT XM_965528R: CCAAACAAGGCCGTACGAATTTC

A. pisum F: TTAGCCAACACTGGAATAAACGTC XM_001946489R: CCAAACAGTCCATGCATATTATT

S. Whyard et al. / Insect Biochemistry and Molecular Biology 39 (2009) 824–832 825

tsetse fly (Glossina morsitans morsitans; Walshe et al., 2009), thetermite (Reticulitermes flavipes; Zhou et al., 2008) and diamondbackmoth (Plutella xylostella) larvae (Bautista et al., 2009). In both mothspecies, the hemipteran and the termite, delivery of the dsRNA intothe gut resulted in knockdown of a gene’s expression in othertissues, indicating that the RNAi was systemic. In contrast, noobservable systemic RNAi was found in the tsetse fly, as feedingdsRNA that targeted a fat body-specific RNA failed to induce RNAi inthat tissue. Ingested dsRNA may not induce RNAi in all insectshowever, as a dsRNA targeting a gut-specific aminopeptidase inlarvae of the moth Spodoptera litura failed to induce RNAi in thatinsect’s gut cells (Rajagopal et al., 2002), which suggests that thismode of delivery may not be suitable for all species.

Two more recent studies demonstrated that transgenic plantscan be engineered to produce hairpin dsRNAs in vivo that canprotect the plants against insect herbivory. Baum et al. (2007) fedWestern corn rootworm larvae (WCR, Diabrotica virgifera) 290different dsRNAs and observed that 14 of these caused significantmortalities at doses �5.2 ng/cm2. They transformed corn toproduce dsRNA specific to the gene encoding the A subunit of theV-type ATPase proton pump and these plants showed significantreduction in WCR-feeding damage. In another study, Mao et al.(2007) targeted a cotton bollworm gut-specific cytochrome P450gene, CYP6AE14, which confers resistance to gossypol, a polyphenolcompound of cotton plants. They first fed cotton bollworm larvaeon transgenic tobacco and Arabidopsis plants that expressed theCYP6AE14-specific dsRNA, and showed that the insects weresubsequently sensitive to gossypol in artificial diets. These twostudies suggest that crop plants may be engineered with dsRNA-expressing transgenes to protect them from the feeding damage ofherbivorous insects.

The sequence specificity of dsRNA coupled with its ability tosuppress genes critical for insect survival suggests that dsRNAscould be developed as tailor-made pesticides, for use on pestinsects where it is important to target only one or several closelyrelated species, without adversely affecting non-target species.Many pesticides in use today are broad-spectrum, capable of killingmany species. Unfortunately, our current repertoire of pesticideswill soon reach its expiration date, as there are growing publicconcerns about the off-target effects of pesticides in our environ-ment and the frequency of pesticide resistance is steadilyincreasing. In this paper, we describe methods to feed dsRNA toa range of different insect species, showing that even highlyconserved genes can be exploited to induce species-limited RNAi,without affecting non-target species. The methods we describecould be applicable to a great many more species, which could beuseful in both higher throughput RNAi screening methods as wellas in the design of future RNAi-based pesticides.

2. Materials and methods

2.1. Double-stranded RNA preparation

Total RNA was isolated from first or second instar insects (D.melanogaster, Manduca sexta, Tribolium castaneum, and Acyrthosi-phon pisum) using a Qiagen RNA extraction kit. Superscript IIreverse transcriptase (Invitrogen) was used to make first strandcDNA using random primers. The cDNA was used as template toPCR-amplify 185 bp portions of the coding sequence of theE-subunit of the vATPase genes from the four insect species usingthe primers listed in Table 1. PCR products were gel-purified witha QIAquick PCR Purification kit (Qiagen), ligated into the pGEMT-Easy vector (Promega), and their identities were confirmed byDNA sequencing. The vATPase gene fragments were subsequentlyexcised from the pGEM T-Easy plasmid using EcoRI and were

ligated into the EcoRI restriction site of the dsRNA transcriptionvector pL4440 (kindly provided by Andrew Fire), which containstwo convergent T7 promoters that flank the plasmid’s multiplecloning site.

The b-glucuronidase (gus) gene was amplified by PCR from thepBacPAK8-GUS plasmid (Clontech) using the primers GusF (50-ATGGTCCGTCCTGTAGAAACC-30) and GusR (50-CCCCACCGAGGCTGTAGC-30). The 0.4 kb PCR product was cloned into the dsRNA transcriptionplasmid pL4440 as described above. The plasmid pL4417 (kindlyprovided by A. Fire) is a derivative of the pL4440 plasmid that containsthe Aequorea victoria green fluorescent protein gene, gfp, flanked bythe two convergent T7 promoters.

To obtain sufficient DNA template for in vitro transcriptions fromeach of the aforementioned plasmids, the gus and vATPase genescloned into pL4440 and the gfp gene within pL4417 werePCR-amplified along with the flanking convergent T7 promotersequences from the dsRNA expression plasmids, using primerspL4440T7-F: 50CCACCTGGCTTATCGAAA30 and pL4440T7-R: 50TAAAACGACGGCCAGTGA30. Double-stranded RNA was prepared usinga T7 RiboMAX Express Large Scale RNA production System(Promega) according to the manufacturer’s specifications. DsRNAconcentrations were determined using standard spectrophoto-metric techniques.

2.2. Preparation of short interfering RNAs

To test the efficacy of siRNAs relative to long dsRNAs, gus and gfpdsRNAs were treated with BLOCK-iT Dicer enzyme (Invitrogen)according to the manufacturer’s specifications. Aliquots of Dicer-treated dsRNA were resolved on 20% polyacrylamide gels toconfirm that dicing was complete. Purified diced dsRNA (250 ng)was mixed with 5 ml Lipofectamine 2000 (Invitrogen) as describedby the manufacturer and was fed to D. melanogaster as describedbelow. To silence the gTub23C gene in the drosophilid species, short(29–41 nt) sense and antisense RNA oligonucleotides (Table 2) weresynthesized (Sigma–Aldrich) and annealed in RNA annealing buffer(30 mM HEPES-KOH, pH 7.4, 100 mM KCl, 2 mM MgCl2, 50 mMNH4Ac) and encapsulated in Lipofectamine 2000 liposomes asdescribed above. gus and gfp RNA oligonucleotides of similar lengthwere used as negative controls in feeding assays, described below.

Silencing of the gus transgene in D. melanogaster was alsoattempted using two commercially prepared siRNAs, prepared byannealing sense and antisense oligonucleotides as described above.The gus siRNAa was prepared using sense oligo AUCAAAAAACUCGACGGCCUU and antisense oligo GGCCGUCGAGUUUUUUGAUUU, while the gus siRNAb was prepared using sense oligoUUGAUCAGCGUUGGUGGGAUU and antisense oligo UCCCACCAACGCUGAUCAAUU. A gfp-specific siRNA, prepared using the senseoligo GGUGAUGCAACAUACGGAAUU and the antisense oligoUUCCGUAUGUUGCAUCACCUU, was used as a negative control.

Table 2RNA oligonucleotides used for preparing tubulin siRNAs.

Gene Sense strand RNA oligo Size (nt) GenBank accession or Flybase ID

D. melanogaster tubulin UCCUUAAGUAGUAGUGAUAAGACGAUUGUUGGUAUGACU 39 FBgn0260639D. sechellia tubulin AUAAGUAGUAAAAAGACGAUUGUUGGUAA 29 FBgn0173288D. yakuba tubulin AGAGAGUUUCCCAUCAAACGUUCAAUAUAUGACACAG 37 FBgn0235626D. pseudoobscura tubulin UUUCUUUGUUUUUACAGCUUAUCUGUAUCUAAAUGUUACAA 41 FBgn0076344gus (control) CGGGCAACGUCUGGUAUCAGCGCGAAGUCUUUAUAC 36 AF305918gfp (control) GGCAAAAAUUCUCUGUCAGUGGAGAGGGUGAAGGUGAUGC 40 L29345

S. Whyard et al. / Insect Biochemistry and Molecular Biology 39 (2009) 824–832826

2.3. Insect bioassays

Pea aphids (A. pisum) were maintained on Vicia faba in anenvironmental chamber at 21 �C with a photoperiod of 16 hlight:8 h dark. To feed aphids vATPase (w185 bp) or gus (w400 bp)dsRNAs, 0–3.0 mg/ml undiced dsRNA were dissolved (withoutliposomes) in the A5 diet described by Febvay et al. (1988) with thesucrose content adjusted to 20%. The liquid diet (0.5 ml) was sealedbetween two layers of Parafilm in a 2 cm diameter feeding arena,and five first instar aphids were placed in each arena. The arena wascovered with fine mesh to prevent their escape and the survival ofthe insects was monitored over a 1 week period. LC50 values and95% confidence limits (C.L.) were estimated by probit analysis usingPOLO-PC (Anon., 1994). Samples of diet were collected over thecourse of the feeding bioassay to assess the stability of the vATPasedsRNA in the liquid. At days 3 and 6, aliquots of the artificial dietwere collected and the RNA was extracted using a RNA extractionkit (Qiagen) and first strand cDNA was prepared using Superscript IIreverse transcriptase (Invitrogen) and the vATPase primers listed inTable 1. The concentration of vATPase dsRNA relative to the amountpresent at time 0 was determined using qRT-PCR, as describedbelow.

Red flour beetle (T. castaneum) larvae were reared continuouslyon 90% wheat flour mixed with 10% brewer’s yeast at 28 �C, 75%relative humidity. To feed beetle larvae vATPase (w185 bp) or gus(w400 bp) dsRNA, 10 mg of diet was placed in 48-well tissueculture plates and 10 ml of 0–5.0 mg/ml undiced dsRNA (withoutliposomes) was overlaid on the surface of the food. Individualneonate beetles were placed in each well and the plate was sealedwith Mylar film (Dupont), with pin holes for aeration. Survival ofthe larvae was monitored over a one week period and LC50 and 95%C.L. values were estimated as described above.

Tobacco hornworm (M. sexta) eggs were obtained from CarolinaBiological Supply Company (Burlington, NC), and raised at roomtemperature (22 �C) with a 16 h light:8 h dark photoperiod andambient humidity. Larvae were raised individually on a cornmeal-based, nicotine-free artificial diet (Carolina Biological Supply). Tofeed vATPase (w185 bp) or gus (w400 bp) dsRNA to M. sexta larvae,neonates were placed in individual wells of a 12-well plate con-taining 25 mg diet that was coated with 25 ml of 0–0.5 mg/mlundiced dsRNA (without liposomes). The plate was sealed withMylar film (Dupont) that was pricked with small holes to providesome aeration without desiccating the food. Survival of the larvaewas monitored over a one week period and LC50 and 95% C.L. valueswere estimated as described above.

Four drosophilid species were obtained from the TucsonDrosophila Stock Center now located at the University of Californiaat San Diego: D. melanogaster (14021–0231.36), Drosophila pseu-doobscura (14011–0121.94), Drosophila sechellia (14021–0248.25),and Drosophila yakuba (14021–0261.01). A gus transgenic strain ofD. melanogaster (Yuen et al., 2008), carrying the Escherichia colib-glucuronidase (gus) gene under the control of the D. melanogasteractin 5c promoter was the host used for the delivery of gus-dsRNA.All cultures were reared on an agar-yeast-cornmeal medium at

room temperature on a 12 h light: 12 h dark photoperiod. To delivergus or gfp dsRNA (long dsRNAs, enzymatically-diced dsRNAs, orsynthesized siRNAs) to gus-strain D. melanogaster embryos,preblastoderm syncytial embryos (1–2 h old) were injected withw0.2 ng dsRNA using standard microinjection methods (Spradling,1986). To deliver gus or gfp dsRNA (long dsRNAs, enzymatically-diced dsRNAs, or synthesized siRNAs) to D. melanogaster larvae,neonates were soaked in solutions of dsRNA (0–0.5 mg/ml dsRNA inPBS, 1% food coloring) with or without encapsulation withincationic liposomes. The insects were soaked for a period of 1–2 hand then transferred to normal diet. To deliver the g-tub-dsRNA toeach of the four drosophilid species, 29–40 nt long dsRNAs wereencapsulated in Lipofectamine 2000. To encapsulate the dsRNAwith liposomes, the dsRNA was mixed with an appropriate amountof the transfection reagent and incubated according to the manu-facturer’s specifications. The effectiveness of four commerciallyavailable transfection reagents were also tested, including Lip-ofectamine 2000 (Invitrogen), Cellfectin (Invitrogen), DMRIE-C(Invitrogen) and Transfectin (BioRad). Survival of the insects wasmonitored over a one week period and LC50 and 95% C.L. valueswere estimated as described above.

For dissection, larvae were first chilled on ice, the posterior andanterior ends were removed, and entire guts were excised. Gutswere pooled for each treatment group in phosphate-buffered saline(10 guts/25 ml for GUS assays) and frozen at �20 �C until used forGUS enzyme assays.

2.4. Glucuronidase (GUS) assays

Embryos or larvae were homogenized in homogenization buffer(50 mM NaHPO4, pH 7.0, 10 mM b-mercaptoethanol, 10 mM EDTA,0.1% sodium lauryl sarcosine, 0.1% Triton X-100), and GUS enzymeactivity was measured using 4-methylumbelliferyl b-D-glucuronicacid as a substrate in fluorometric assays, as described in Gallagher(1992). Protein concentrations of the protein extracts were deter-mined using a Micro BCA Protein Assay kit (Pierce) according to themanufacturer’s specifications.

2.5. Quantitative RT-PCR to assess extent of RNAi

RNA was extracted from pools of 5–10 dsRNA-treated larvae andfrom dissected guts and the rest of the body using a Qiagen RNAExtraction kit. The samples were then treated with DNAse I (Invi-trogen) to remove any genomic DNA contamination, and were thenused with Superscript II reverse transcriptase (Invitrogen) to makefirst strand cDNA using random primers. Quantitative reversetranscriptase PCR (qRT-PCR) reactions were performed in triplicateon a BioRad iQ5 Real-Time PCR Detection System using 10 ng ofcDNA, 0.4 mM primers, and iQ SYBR Green Supermix (BioRad).Cycling conditions and dissociation curve analyses was performedaccording to the manufacturer’s instructions. To amplify the gus,gTub23C and vATPase genes, the qRT-PCR primers listed in Table 3were used.

Table 3Quantitative real-time PCR primers for determining gene expression.

Species Gene Primer sequences

D. melanogaster vATPase F: 50GCTCGTCTGAAGGTGCTGR: 50TACTCGGACTGATTCTTGGTG

Tubulin F: 50GATTCGTTGGAGGGCTTTATACR: 50GTTTGGATAAGTTTCTTGGGATAG

gus F: 50GGTGAAGGTTATCTCTATGR: 50CCAAAGCCAGTAAAGTAG

RpL32 F: 50CCGCTTCAAGGGACAGTATCR: 50ATCTCGCCGCAGTAAACG

D. sechellia Tubulin F: 50AACAGGCTCGGGAATGGGR: 50TGATTGGGAAACACACTGAATG

RpL32 F: 50CAAGGGACAGTATCTGATGCCR: 50ATCTCGCCGCAGTAGACG

D. yakuba Tubulin F: 50CCTCCTACATGAACAACAATCTCR: 50GTCGTCTTTCGCACATTAACC

RpL32 F: 50CGGACCGATATGCTAAGCR: 50CGATCCATAACCGATGTTG

D. pseudoobscura Tubulin F: 50CCCAGTTTCTCGCAAATCAACR: 50GGTATCAAGGACGCCATAAGC

RpL32 F: 50ACAACAGAGTGCGTCGTCR: 50CACCAGGAATTTCTTGAAGCC

T. castaneum vATPase F: 50GAGAACAATATAGTGGTGAGAGTCR: 50TATTTCGTCGCAACAACTGG

RpS6 F: 50ACGCAAGTCAGTTAGAGGGTGCATR: 50TCCTGTTCGCCTTTACGCACGATA

M. sexta vATPase F: 50GATGTTCAAAAGCAGATCAAGCR: 50CTCCTCCTCCGCCTTAGC

RpS3 F: 50GCAGAAGCGGTTCAACATCR: 50AGACCTCCAATGAGTTTGTATC

A. pisum vATPAse F: 50GAATGGTTAAGTTCATAGAGTTTGR: 50CCCTTGAGTATCATCTCGTTC

RpL7 F: 50TTGAAGAGCGTAAGGGAACTGR: 50TATTGGTGATTGGAATGCGTTG

Fig. 1. The effect of transfection reagents and dsRNA concentration on RNAi of the gustransgene in D. melanogaster larval guts following ingestion of long (w0.4 kb) gus-dsRNA. The values represent the mean and standard errors from four replicates.Different letters (a, b, and c) indicate significantly different reductions in GUS activity(using 0.5 mg/ml dsRNA) relative to other treatments (p < 0.05, ANOVA, Tukey–Kramermultiple comparisons test).

S. Whyard et al. / Insect Biochemistry and Molecular Biology 39 (2009) 824–832 827

The ribosomal protein gene, RpL32, and its orthologoussequences in drosophilid species (GenBank accession no.NM_170461) were used as the reference genes to which the level ofother transcripts were standardized, and each was amplified usingthe primers listed in Table 3. The relative amounts of gus andgTub23C transcripts were first normalized to the endogenousreference gene RpL32, and then normalized relative to the level ofgene transcripts in embryos or larvae that were treated with thenegative control gfp-specific dsRNA or siRNAs using a modifiedcomparative Ct method (Pfaffl, 2001). The same approach wastaken to determine the transcript levels of vATPase transcripts innon-drosophilid insects. Primers were designed for the vATPasegene of each individual species, then transcript values werenormalized to ribosomal genes of each insect (RpS6 for T. casta-neum, RpS3 for M. sexta, and RpL7 for A. pisum).

3. Results

3.1. RNAi in D. melanogaster using different dsRNAdelivery methods

A transgenic strain of D. melanogaster carrying a ubiquitously-expressed b-glucuronidase gene (gus) under the control of theact5C promoter (Yuen et al., 2008) was used to test the efficacy ofseveral different methods of dsRNA delivery, including feeding. Bytargeting a transgene that has no observable effect on the viabilityor fitness of the insects, RNAi could be induced without adverselyaffecting their growth or development. Direct injection of a long0.4 kb in vitro transcribed gus-dsRNA into the gus-strain embryosresulted in 86% reduction in GUS enzyme activity 16 h post

injection (Supplementary Material Fig. S1). Two different gus-specific 21 bp short interfering RNAs (siRNAs), each targeting thegus transcript, were injected into embryos at the same concentra-tion as the long dsRNA, but they produced only a 19–24% reductionin GUS activity. An equimolar mix of these two siRNAs showed nosignificant difference in the level of RNAi relative to each siRNAdelivered separately (Supplementary Material Fig. S1). In contrast,when the 0.4 kb gus-dsRNA was enzymatically diced into a complexmixture of siRNAs using a commercially available Dicer (Invitrogen)and injected into embryos, the extent of RNAi was almost as potentas the intact 0.4 kb dsRNA. RNAi following embryo injection of thelong dsRNA was observed to persist through larval development,although it diminished with each successive moult, and was nolonger detectable in pupae or adult flies (results not shown).Negative controls, injected with long, diced or synthetic siRNAsspecific for the green fluorescent protein (gfp)-gene showed nochanges in gus expression. Quantitative RT-PCR confirmed that thereduced levels of GUS enzyme activity were correlated withreduced gus transcript levels after injection with long dsRNAs. Inembryos, 24 h after long (w0.4 kb) gus-dsRNA injection, gus tran-script levels were reduced to 8% of normal levels (SupplementaryMaterial Table S1), and in all other developmental stages, the levelsof gus transcripts correlated closely with the reduced levels of GUSactivity.

In nematodes (Tabara et al., 1998) and flatworms (Orii et al.,2003), it is possible to induce RNAi simply by soaking the animals ina solution of dsRNA. Soaking of newly hatched Drosophila larvae fora period of 1 h in a range of concentrations (0.1–0.3 mg/ml) of long,(undiced 0.4 kb) gus-dsRNAs also failed to reduce gus geneexpression in the gut tissues, but at the two highest concentrationstested (0.4 and 0.5 mg/ml), a small (5–8%) reduction in GUS activitywas observed in isolated gut tissues (Fig. 1). However, soakingneonate larvae in Lipofectamine 2000 (Invitrogen) liposome-encapsulated long dsRNA for a period of 1 h resulted in approxi-mately a 50% reduction of GUS activity in isolated guts 24 hpost-exposure to the dsRNA (Fig. 1). Soaking larvae for 2 h did notincrease the extent of RNAi significantly (see SupplementaryMaterial Table S3), but longer soaking periods (>3 h) resulted inincreased incidences of drowning. Soaking third instars also

Fig. 2. Mortality of Drosophila species after ingestion of species-specific tubulin-dsRNAs encapsulated in Lipofectamine 2000. Each panel represents treatment withone species-specific (undiced) dsRNA: A) D. melanogaster dsRNA; B) D. sechellia dsRNA;C) D. yakuba dsRNA; D) D. pseudoobscura dsRNA. The values represent the mean andstandard errors from three replicate experiments. Black bars represent treatmentswith gus-dsRNA and white bars represent treatments with the species-specific dsRNA.Mortality of each species treated with their conspecific dsRNA was significantlydifferent from those insects treated with heterospecific dsRNA (p < 0.05, ANOVA,Tukey–Kramer multiple comparisons test). Species treated are represented as follows:D.m. – D. melanogaster, D.s. – D. sechellia, D.y. – D. yakuba, D.p. – D. pseudoobscura.

S. Whyard et al. / Insect Biochemistry and Molecular Biology 39 (2009) 824–832828

resulted in RNAi, but with reduced efficacy relative to feedingneonates (Supplementary Material Table S3). Like the embryoinjections, this soaking method resulted in persistence of RNAi ingut tissues, with reduced potency with each successive moult(Supplementary Material Table S2). Addition of food coloring to thedsRNA cocktail indicated that the liquid was ingested, as the coloraccumulated in the gut and not elsewhere in the body.

The larvae were also offered the dsRNA by droplet feeding, andof those insects that fed, similar levels of RNAi were observed(Supplementary Material Table S2). A small percentage (20%) oflarvae was observed not to have any food coloring in their guts, andthese individuals showed no reduction of GUS activity, whichsuggested that the primary route of entry for the dsRNA is via thealimentary tract. Feeding D. melanogaster on an artificial diet eithersurface-coated with 0.5 mg/ml dsRNA or dsRNA mixed into the dietfailed to induce RNAi over the 5 day period of larval development(Supplementary Material Table S2).

Four commercially available transfection reagents were testedfor their efficacy to induce RNAi by ingested dsRNA. While all ofthem facilitated some degree of RNAi in the isolated gut tissues,TransFectin (BioRad) and DMRIE-C (Invitrogen) were the leasteffective, reducing GUS activity only 31% and 35% respectively, atdsRNA concentrations of 0.5 mg/ml. The other two transfectionreagents, Cellfectin (Invitrogen) and Lipofectamine 2000, providedsomewhat higher RNAi efficacies, reducing GUS activity 49% and52% respectively. Quantitative RT-PCR confirmed that the reductionin GUS enzyme activity correlated quite closely with the extent ofgus transcript knockdown (Supplementary Material Table S3).

The extent of RNAi was dsRNA dose-dependent. When larvaewere soaked for 1 h, concentrations of 0.05 mg/ml and lowerprovided no RNAi, while doses between 0.1 and 0.5 mg/ml providedincreasing levels of RNAi (Fig. 1). Beyond 0.5 mg/ml, no furtherincrease in RNAi potency was observed (results not shown). Dicedgus-dsRNA (0.5 mg/ml) encapsulated in Lipofectamine 2000 wasfed to first instar larvae and produced an average decrease in GUSactivity of 43.9 � 6.1%, which is similar to the level of knockdownobserved using the long dsRNA. Based on these results, all subse-quent feeding trials with D. melanogaster used undiced dsRNAsencapsulated in Lipofectamine 2000.

In nematodes and flatworms, ingested dsRNA results in systemicRNAi, as the dsRNA spreads to adjacent tissues (Timmons et al.,2001; Newmark et al., 2003). To determine whether the ingesteddsRNA induced RNAi in tissues other than gut cells, dsRNA-fed larvae were dissected and the level of GUS activity and gustranscripts were assessed in isolated guts and other tissues. Nosignificant knockdown of GUS activity or gus transcripts wasdetected in the carcass (Supplementary Material Table S3), evenwhen using the highest concentration of dsRNA (0.5 mg/ml), whichsuggests that either the dsRNA did not move from the gut cells orthat insufficient dsRNA was ingested to induce any degree ofsystemic silencing. These results are in agreement with a previousinvestigation by Roignant et al. (2003) who observed that trans-genes expressing dsRNA in D. melanogaster resulted in cell auton-omous RNAi, with no evidence of spreading of dsRNA to tissueslacking the transgenes.

3.2. Feeding dsRNAs can result in species-specific mortality withinthe genus Drosophila

Delivery of dsRNAs that target essential genes regulating cellfunctions or development can potentially result in the death of thecell and/or the organism in which RNAi occurs. Previously, we haveshown that injection or biolistic delivery of dsRNAs that targetvarious genes essential to cell viability resulted in death of devel-oping D. melanogaster embryos and larvae (Yuen et al., 2008).

Ingested dsRNA has previously been shown to be lethal or producedeleterious effects in a limited number of insect species thus far(Baum et al., 2007; Mao et al., 2007; Zhou et al., 2008; Walshe et al.,2009), but not in drosophilids. The gTub23C gene of D. melanogasterencodes the ubiquitously-expressed g-tubulin protein, which isessential for viability and microtubule organization during mitosis(Joshi, 1993; Moritz et al., 1995a,b). Mutation of this gene results inmitotic arrest, highly abnormal microtubule organizing centers,and death of larvae (Barbosa et al., 2000, 2003). The gTub23C geneis highly conserved in four species of Drosophila (D. melanogaster,D. sechellia, D. yakuba, and D. pseudoobscura), with each speciessharing between 79 and 96% sequence identity throughout thecoding sequence of the gene with the other species (SupplementaryMaterial Table S4). However, the relatively short 30 untranslatedregion (30 UTR) of the gTub23C gene displayed sufficient sequencedivergence among the four species such that RNAi cross-silencingshould not occur, as no 19–21 nucleotide length of sequence wasshared among the four species.

The four Drosophila species were fed gTub23C dsRNA encapsu-lated in liposomes and all species suffered high mortalitiesfollowing ingestion of conspecific dsRNA (Fig. 2). None of thedrosophilid species showed any evidence of RNAi when fednon-encapsulated dsRNA (results not shown). Each species was

S. Whyard et al. / Insect Biochemistry and Molecular Biology 39 (2009) 824–832 829

relatively unaffected by gTub23C dsRNA specific to any of the otherdrosophilid species, and qRT-PCR analyses of gTub23C geneexpression in isolated guts indicated that this gene’s expressionwas not perturbed in the gut tissues of the insects that fed uponheterospecific dsRNA (Table 4). This suggests that RNAi can bereadily tailored to target only one species, provided a suitable genewith sufficient sequence divergence can be identified.

3.3. Insecticidal dsRNAs in non-drosophilids

To demonstrate that dsRNA targeting a single gene could beused selectively to target different pest insects, we designedspecies-specific dsRNAs to silence the gene encoding the E-subunitof V-ATPase in four insect species: the fruit fly D. melanogaster, theflour beetle T. castaneum, the tobacco hornworm M. sexta, and thepea aphid A. pisum. V-ATPase is a membrane-bound protein thatacts as a proton pump to establish the pH gradient within the gutlumen of many insects. Silencing of the gene has previously beendemonstrated in the Western corn rootworm to prevent growthand maturation of the larvae feeding on transgenic corn expressingthe insect-specific vATPase dsRNA (Baum et al., 2007). In D. mela-nogaster, P-element induced mutations of the V-ATPase 26 kDE-subunit gene (Vha26) are also lethal (Dow et al., 1997). Despitethis gene being relatively well conserved across the different taxa,(with species sharing 40–75% nucleotide identity over the entiregene’s length (Supplementary Material Table S4)), it was relativelyeasy to design dsRNA sequences that were unique to each of thefour species being tested.

All insects were first fed a series of concentrations of conspecificdsRNA to determine the LC50 values after a 1 week period of feeding(Table 5). While the dsRNA was administered to D. melanogaster bysoaking neonate larvae for 1 h in a liposome dsRNA cocktail, theother three species were simply fed dsRNA that was mixed intotheir diet without using any transfection reagents. The LC50 value of0.23 mg/ml was highest for D. melanogaster, which may bea reflection of the short time of exposure to the dsRNA for thisspecies and/or reduced uptake by the gut cells. The LC50 values forthe moth larvae, aphids, and beetle larvae, which were all fed thedsRNA continuously throughout their larval development, were 21,68, and 92 times lower than that of the fruit fly larvae. Samples ofthe different insect diets were examined after 3 and 6 days for thepresence of the dsRNA using qRT-PCR to determine whetherthe dsRNA was equally stable in all of the diets during the feedingperiods. In the aphid diet, dsRNA levels dropped 32 � 7% and56 � 8% by days 3 and 6 respectively. In contrast, dsRNA levels inthe solid diets that were fed to beetle and moth larvae onlydecreased on average 14 � 6% and 31 � 5% by days 3 and 6,respectively (data not shown). While it is somewhat difficult tomake direct comparisons of LC50 values given that some insectswere fed dsRNA in a different manner, it is clear that the vATPasedsRNA had obvious deleterious effects on each of the insect speciesexamined.

Table 4LC50 values and percent knockdown of g-tubulin transcripts in four drosophilid species f

Species LC50 (mg/ml) (95% C.L.)a % knockdown o

D. mel dsRNA

D. melanogaster 0.46 (0.37–0.58) 49.1 � 4.6D. sechellia 0.61 (0.52–0.71) –D. yakuba 0.53 (0.46–0.60) –D. pseudoobscura 0.41 (0.35–0.50) –

a No significant difference among the LC50 values, given that the 95% C.L. values overlb Knockdown of mRNA levels in isolated guts relative to gus-dsRNA-treated controls.

To examine the specificity of the dsRNA to selectively kill thetarget species, the four species of insects were fed each dsRNA at theestimated LC75 dose for the targeted species. When insects werefed on a diet laced with D. melanogaster-specific vATPase dsRNA(at 0.5 mg/ml liquid diet or 0.5 mg/g solid diet), only D. melanogastershowed any evidence of reduced growth and development, whilethe other three species were unaffected by the Drosophila dsRNA.Similarly, by feeding each of the other three species conspecificdsRNA, we selectively killed each species, without adverselyaffecting any of the others (Fig. 3). Control insects fed on gus-dsRNAshowed no serious effects from the dsRNA feeding. Partial knock-down of vATPase mRNA levels in isolated guts was observed in eachspecies when treated with conspecific dsRNA, but showed nosignificant reduction in vATPase transcripts when fed heterospecificdsRNAs (Table 5). For all four species examined, partial knockdownof the V-ATPase E-subunit gene’s transcripts was sufficient to inducehigh levels of mortality, which suggests that perturbations in thegene’s normal expression can be lethal. Reductions in vATPasetranscript levels were detected only in the gut and not in the carcassof the insects (Table 5), although it is difficult to conclude from thesecrude dissections that no systemic RNAi occurred, as some smallperturbations in vATPase transcripts may have occurred in a limitednumber of other tissues.

4. Discussion

As more insects’ genomes are sequenced, RNAi will continue tobe a useful tool to ascribe functions to the many newly identifiedgenes. The application of RNAi for pest insect control, however, hasonly just begun to be explored, and in this study, we show thatdsRNAs can be tailor designed to selectively target and kill severalinsect pest species, even when using dsRNAs that target wellconserved genes.

DsRNA targeting the gene encoding the A subunit of vATPasehad previously been shown to kill Western corn rootworm larvae infeeding bioassays (Baum et al., 2007), and here, we observed thatdsRNA targeting the E-subunit of vATPase can be selectivelydesigned and fed to a broader range of insects, including larvae ofanother species of beetle (T. castaneum), moth larvae (M. sexta),aphid nymphs (A. pisum), and dipteran larvae (D. melanogaster). Thelevel of sequence divergence of the vATPase genes among thespecies tested was sufficient enough to identify portions of the genethat showed no 19–21 nt overlap, and thereby prevented RNAi-induced silencing in the non-target species. To demonstrate thatthis method could also be used to target closely related species, wethen selected a gene that was even more highly conserved, thegTub23C gene, and by targeting the more variable 30 UTRsequences, it was still possible to design dsRNAs that were species-specific.

In this study, we did not identify the mechanism by which RNAi-induced mortality, although qRT-PCR assays confirmed that RNAihad reduced transcript levels in the isolated gut tissues, albeit notcompletely. As the insects lived for several days while consuming

ollowing soaking of larvae in g-tubulin dsRNAs.

f g-tubulin transcriptsb following exposure to different species’ dsRNAs

D. sec dsRNA D. yak dsRNA D. pse dsRNA

5.7 � 1.2 4.6 � 1.0 1.8 � 1.640.7 � 5.5 3.7 � 1.6 3.2 � 1.7– 36.0 � 4.9 4.6 � 1.3– – 51.0 � 4.4

ap.

Table 5LC50 values and percent knockdown of vATPase transcripts in four insect species following feeding on different vATPase dsRNAs. (Species represented as follows: D.m. –D. melanogaster, M.s. – M. sexta, T.c. – T. castaneum, A.p. – A. pisum).

Species LC50 (mg/g diet)a (95% C.L.) % knockdown of vATPase transcripts following feeding on differentdsRNAsc

% knockdown of vATPase transcriptsin the carcass when fed onconspecific dsRNAc

D.m. dsRNA M.s. dsRNA T.c. dsRNA A.p. dsRNA

D. melanogaster b 0.23 (0.17–0.30)1 56.2 � 8.6* 4.9 � 3.0 4.2 � 2.4 3.0 � 1.7 4.1 � 2.5M. sexta 0.011 (0.008–0.013)2 – 33.7 � 6.5* 5.3 � 2.1 2.2 � 1.4 5.6 � 2.7T. castaneum 0.0025 (0.0019–0 0.0031)3 – – 41.2 � 5.8* 4.6 � 3.1 5.1 � 2.4A. pisum 0.0034 (0.0028–0.0039)3 – – – 31.2 � 5.6* 4.5 � 2.4

*Denote values significantly different from negative controls (p < 0.05, ANOVA).a Different superscripted numbers denote values that are significantly different from the other treatments given that 95% C.L. values do not overlap.b D. melanogaster were exposed to dsRNA diet for only 2 h, whereas other insects were fed continuously on a dsRNA-containing diet for 7 days.c Knockdown relative to gus-dsRNA negative controls.

S. Whyard et al. / Insect Biochemistry and Molecular Biology 39 (2009) 824–832830

the dsRNAs before deleterious effects were observed, it suggeststhat the mortality was a consequence of latent effects of the dsRNA,reducing gene function sufficiently to disrupt normal gut cellfunction, thereby leading to death of the growing insect. It will be ofinterest to determine whether RNAi was widespread throughoutthe gut or whether only certain cells were affected.

Fig. 3. DsRNA-induced mortality in four insect species after feeding on diets con-taining vATPase-dsRNA specific to each species. Each panel shows the mortalitiesresulting from feeding on one species-specific dsRNA: A) D. melanogaster dsRNA; B) M.sexta dsRNA; C) T. castaneum dsRNA; D) A. pisum dsRNA. Values represent the meansand standard errors from three replicates. Black bars represent treatments with gus-dsRNA and white bars represent treatments with the species-specific dsRNA. Mortalityof each species treated with their conspecific dsRNA was significantly different fromthose insects treated with heterospecific dsRNA (p < 0.05, ANOVA, Tukey–Kramermultiple comparisons test). All species except D. melanogaster were fed unencapsu-lated long dsRNAs, whereas D. melanogaster long dsRNA was encapsulated inLipofectamine 2000. Species treated are represented as follows: D.m. – D. melanogaster,M.s. – M. sexta, T.c. – T. castaneum, A.p. – A. pisum.

In the drosophilid species, we detected no evidence of systemicRNAi, which may reflect the absence of the RNA channel trans-porter SID-1 in this genus and perhaps in all Diptera, as has beenpreviously suggested (Tomoyasu et al., 2008). We also did notdetect evidence of RNAi beyond the gut tissues of dsRNA-fedT. castaneum, M. sexta, or A. pisum, although we only measuredvATPase transcript levels from the entire remainder of the body, andhence we may have overlooked any tissues showing modest levelsof vATPase RNAi. In future studies, it will be of interest to examinewhether the dsRNAs or the processed siRNAs can leave the gut cellsto induce silencing in other tissues.

Feeding dsRNA to insects to induce RNAi has not worked inevery species tested (Rajagopal et al., 2002), but we have found thatthis method of dsRNA delivery induced RNAi in six insects withsequenced genomes (the four drosophilid species, T. castaneum, andA. pisum). Based on these results, it may be possible to developdsRNA feeding assays for higher throughput functional genomicsstudies in these insects. Even if the dsRNA does not move beyondgut cells, this mode of dsRNA delivery could enable studies of gut-expressed genes. The mechanisms that facilitate dsRNA uptake inthe insect guts are still unknown and therefore, it is difficult topredict in which species this method of delivery might work. WhileSID-1 and a related RNA transporter, SID-2, can mediate dsRNAuptake in the gut of the nematode C. elegans, it has not beendemonstrated that SID-1, which is at least present in many insects(Gordon and Waterhouse, 2007; Tomoyasu et al., 2008), serves thesame role in insects. In D. melanogaster and the other drosophilidspecies, all of which apparently lack any SID-1 ortholog, endocy-tosis appears to facilitate cell uptake. This process may, however, betoo slow to facilitate a strong RNAi response without the use oftransfection reagents to improve delivery to gut cells. In this study,we only tested cationic liposomes as dsRNA carriers, but in futurestudies, we will examine the efficacy of other transfection reagents,in insects other than Drosophila. Without knowing more about eachof the insects’ guts physiologies and their cellular membranes, itmight be difficult to predict which transfection reagent would workbest for those species where naked dsRNA failed to induce RNAi. Itwas interesting to observe that despite these insects having verydifferent diets, the dsRNA was relatively stable over the several daysof feeding, and consequently, dsRNA could be administered to theseinsects relatively easily.

As insects lack the interferon-regulated innate immunitypathway that protects vertebrates from invasion of long dsRNA(Clemens and Elia, 1997; Oates et al., 2000; Geiss et al., 2001) it ispossible and may be considered preferable to deliver long dsRNAs toinsects, to ensure maximal RNAi. In this study however, we observedthat it was still possible to deliver relatively short (29–40 nt) dsRNAsand they were potent enough to induce mortality in the insects. Thisfeature would make it easier to design dsRNAs that are unique totarget species, with reduced risk of cross-species silencing.

S. Whyard et al. / Insect Biochemistry and Molecular Biology 39 (2009) 824–832 831

The appealing aspect of using dsRNA as a pesticide is that it ispossible to design the pesticide to target only a single species ora group of related species, with minimal threat to other organisms.Based on our observations of feeding dsRNA to M. sexta and A.pisum, dsRNA-expressing transgenes could potentially protect cropplants from these pest insects. It may also be possible to simply coatthe insect’s food with dsRNA, and for the flour beetle T. castaneum,coating stored grain products with dsRNA may reduce losses causedby this pest. Presently, it is a costly endeavour to make largequantities of dsRNA, but as bacteria can be engineered to producedsRNAs (Timmons and Fire, 1998), large-scale production of dsRNA-based pesticides could become a reality.

Like every other pesticide that has been used to control insects,there is the likelihood of some insects developing resistance toa dsRNA-based pesticide. If they acquire resistance to one particulardsRNA sequence due to mutations in the target RNA, then it wouldbe relatively easy to select another dsRNA that either targetsanother portion of the same gene or targets a new gene. With thesequencing of several pest genomes already completed and more inthe works, potential RNAi target genes are rapidly being discovered.It is less likely that the insects would acquire resistance throughmutations to the RNAi machinery, as these proteins are essential forthe normal processing of the insects’ endogenous double-strandedRNAs, the microRNAs, which are essential in regulating geneexpression during development (reviewed in Wienholds andPlasterk, 2005; Erson and Petty, 2008; Singh et al., 2008; Carthewand Sontheimer, 2009). Resistance may also arise though mutationsin the dsRNA uptake proteins, such as the RNA transporter SID-1, ora component of the RNA-specific endocytotic machinery. Furtherstudies on the relevance of SID-1 in insects and its possibleinvolvement in gut-mediated uptake of dsRNA are needed toaddress this issue.

The greatest challenge for future use of dsRNA as a pesticide willbe in the identification of other appropriate target genes. As Baumet al. (2007) observed, not all dsRNAs may effectively silence theirtargets, and not all dsRNAs are effective in killing insects. Here, wehave provided evidence that a wider range of insects can be feddsRNA with insecticidal properties, and with the development ofsome higher throughput screening methods, it will be possible toscreen large numbers of dsRNAs to determine which ones areworthy of further development as insect control agents.

Acknowledgements

We thank N. Gibb, A. Ruffell, and S. Read for their technicalassistance. The gus strain of D. melanogaster was a gift fromDr. P. Atkinson and Dr. D. O’Brochta. This work was supported bya NSERC grant to S. Whyard and a NSERC Undergraduate ResearchAward to S. Wong.

Appendix. Supplementary material

Supplementary data associated with this article can be found inthe online version at doi:10.1016/j.ibmb.2009.09.007.

References

Amdam, G.V., Simoes, Z.L., Guidugli, K.R., Norberg, K., Omholt, S.W., 2003. Disrup-tion of vitellogenin gene function in adult honeybees by intra-abdominalinjection of double-stranded RNA. BMC Biotechnol. 3, 1.

Anon., 1994. User’s Guide POLO-PC, Probit and Logit Analysis. LeOra Software,Berkeley, CA.

Araujo, R.N., Santos, A., Pinto, F.S., Gontijo, N.F., Lehane, M.J., Pereira, M.H., 2006.RNA interference of the salivary gland nitrophorin 2 in the triatomine bugRhodnius prolixus (Hemiptera: Reduviidae) by dsRNA ingestion or injection.Insect Biochem. Mol. Biol. 36, 683–693.

Barbosa, V., Yamamoto, R.R., Henderson, D.S., Glover, D., 2000. Mutation ofa Drosophila gamma tubulin ring complex subunit encoded by discs degenerate-4 differentially disrupts centrosomal protein localization. Genes Dev. 14, 3126–3139.

Barbosa, V., Gatt, M., Rebollo, E., Gonzalez, C., Glover, D., 2003. Drosophila dd4mutants reveal that gTuRC is required to maintain juxtaposed half spindles inspermatocytes. J. Cell Sci. 116, 929–941.

Baum, J.A., Bogaert, T., Clinton, W., Heck, G.R., Feldmann, P., Ilagan, O., Johnson, S.,Plaetinck, G., Munyikwa, T., Pleau, M., Vaughn, T., Roberts, J., 2007. Control ofcoleopteran insect pests through RNA interference. Nat. Biotechnol. 25, 1322–1326.

Bautista, M.A.M., Miyata, T., Miura, K., Tanaka, T., 2009. RNA interference-mediatedknockdown of a cytochrome P450, CYP6BG1, from the diamondback moth,Plutella xylostella, reduces larval resistance to permethrin. Insect Biochem. Mol.Biol. 39, 38–46.

Bettencourt, R., Terenius, O., Faye, I., 2002. Hemolin gene silencing by ds-RNAinjected into Cecropia pupae is lethal to next generation embryos. Insect Mol.Biol. 11, 267–271.

Bettencourt-Dias, M., Goshima, G., 2009. RNAi in Drosophila S2 cells as a tool forstudying cell cycle progression. Methods Mol. Biol. 545, 39–62.

Bucher, G., Scholten, J., Klingler, M., 2002. Parental RNAi in Tribolium (Coleoptera).Curr. Biol. 12, R85–R86.

Caplen, N.J., Fleenor, J., Fire, A., Morgan, R.A., 2000. DsRNA-mediated gene silencingin cultured Drosophila cells: a tissue culture model for the analysis of RNAinterference. Gene 252, 95–105.

Carthew, R.W., Sontheimer, E.J., 2009. Origins and mechanisms of miRNAs andsiRNAs. Cell 136, 642–655.

Chen, M., Du, Q., Zhang, H.Y., Wang, X., Liang, Z., 2007. High-throughput screeningusing siRNA (RNAi) libraries. Expert Rev. Mol. Diagn. 7, 281–291.

Clemens, M.J., Elia, A., 1997. The double-stranded RNA-dependent protein kinasePKR: structure and function. J. Interferon Cytokine Res. 17, 503–524.

Clemens, J.C., Worby, C.A., Simonson-Leff, N., Muda, M., Maehama, T.,Hemmings, B.A., Dixon, J.E., 2000. Use of double-stranded RNA interference inDrosophila cell lines to dissect signal transduction pathways. Proc. Natl. Acad.Sci. U. S. A. 97, 6499–6503.

Dow, J.A.T., Davies, S.A., Guo, Y., Graham, S., Finbow, M.E., Kaiser, K., 1997. Moleculargenetic analysis of V-ATPase function in Drosophila melanogaster. J. Exp. Biol.200, 237–245.

Dzitoyeva, S., Dimitrijevic, N., Manev, H., 2001. Intra-abdominal injection of double-stranded RNA into anesthetized adult Drosophila triggers RNA interference inthe central nervous system. Mol. Psychiatry 6, 665–670.

Eaton, B.A., Fetter, R.D., Davis, G.W., 2002. Dynactin is necessary for synapsestabilization. Neuron 34, 729–741.

Erson, A.E., Petty, E.M., 2008. MicoRNAs in development and disease. Clin. Genet. 74,296–306.

Febvay, G., Delobel, B., Rahbe, Y., 1988. Influence of the amino-acid balance on theimprovement of an artificial diet for a biotype of Acyrthosiphon pisum (Homo-ptera, Aphididae). Can. J. Zool. 66, 2449–2453.

Gallagher, S.R., 1992. GUS Protocols: Using the GUS Gene as a Reporter of GeneExpression. Academic Press, San Diego.

Gatehouse, H.S., Gatehouse, L.N., Malone, L.A., Hodges, S., Tregidga, E., Todd, J., 2004.Amylase activity in honey bee hypopharyngeal glands reduced by RNA inter-ference. J. Apic. Res. 43, 9–13.

Geiss, G., Jin, G., Guo, J., Bumgarner, R., Katze, M.G., Sen, G.C., 2001. A comprehensiveview of regulation of gene expression by double-stranded RNA-mediated cellsignaling. J. Biol. Chem. 276, 30178–30182.

Gordon, K.H.J., Waterhouse, P.M., 2007. RNAi for insect-proof plants. Nat. Biotechnol.25, 1231–1232.

Hannon, G.J., 2002. RNA interference. Nature 418, 244–251.Joshi, H.C., 1993. Gamma-tubulin – the hub of cellular microtubule assemblies.

Bioessays 15, 637–643.Kuttenkeuler, D., Boutros, M., 2004. Genome-wide RNAi as a route to gene function

in Drosophila. Brief. Funct. Genomics Proteomics 3, 168–176.Mao, Y.-B., Cai, W.-J., Wang, J.-W., Hong, G.-J., Tao, X.-J., Wang, L.-J., Huang, Y.-P.,

Chen, X.-Y., 2007. Silencing a cotton bollworm P450 monooxygenase gene byplant-mediated RNAi impairs larval tolerance of gossypol. Nat. Biotechnol. 25,1307–1313.

Misquitta, L., Paterson, B.M., 1999. Targeted disruption of gene function inDrosophila by RNA interference (RNA-i): a role for nautilus in embryonicsomatic muscle formation. Proc. Natl. Acad. Sci. U. S. A. 96, 1451–1456.

Moritz, M., Braunfeld, M.B., Sedat, J.W., Alberts, B., Agard, D.A., 1995a. Microtubulenucleation by gamma-tubulin-containing rings in the centrosome. Nature 378,638–640.

Moritz, M., Braunfeld, M.B., Fung, J.C., Sedat, J.W., Alberts, B.M., Agard, D.A., 1995b.Three-dimensional structural characterization of centrosomes from earlyDrosophila embryos. J. Cell Biol. 130, 1149–1159.

Newmark, P.A., Reddien, P.W., Cebria, F., Sanchez-Alvarado, A., 2003. Ingestion ofbacterially expressed double-stranded RNA inhibits gene expression inplanarians. Proc. Natl. Acad. Sci. U. S. A. 100, 11861–11865.

Oates, A.C., Bruce, A.E., Ho, R.K., 2000. Too much interference: injection of double-stranded RNA has nonspecific effects in the zebrafish embryo. Dev. Biol. 224,20–28.

Orii, H., Mochii, M., Watanabe, K., 2003. A simple ‘‘soaking method’’ for RNAinterference in the planarian Dugesia japonica. Dev. Genes Evol. 213,138–141.

S. Whyard et al. / Insect Biochemistry and Molecular Biology 39 (2009) 824–832832

Pfaffl, M.W., 2001. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 29, e45.

Rajagopal, R., Sivakumar, S., Agrawal, N., Malhotra, P., Bhatnagar, R.K., 2002.Silencing of midgut aminopeptidase N of Spodoptera litura by double-strandedRNA establishes its role as Bacillus thuringiensis toxin receptor. J. Biol. Chem.277, 46849–46851.

Roignant, J.Y., Carre, C., Mugat, B., Szymczak, D., Lepesant, J.A., Antoniewski, C.,2003. Absence of transitive and systemic pathways allows cell-specific andisoform-specific RNAi in Drosophila. RNA 9, 299–308.

Saleh, M.C., van Rij, R.P., Hekele, A., Gillis, A., Foley, E., O’Farrell, P.H., Andino, R.,2006. The endocytic pathway mediates cell entry of dsRNA to induce RNAisilencing. Nat. Cell. Biol. 8, 793–802.

Singh, S.K., Bhadra, M.P., Girschick, H.J., Bhadra, U., 2008. Micro RNAs – micro in sizebut macro in function. FEBS J. 275, 4929–4944.

Spradling, A.C., 1986. P-element-mediated transformation. In: Roberts, D.B. (Ed.),Drosophila: a Practical Approach. IRL Press, Oxford, pp. 175–197.

Tabara, H., Grishok, A., Mello, C.C., 1998. RNAi in C. elegans: soaking in the genomesequence. Science 282, 430–431.

Timmons, L., Court, D.L., Fire, A., 2001. Ingestion of bacterially expressed dsRNAs canproduce specific and potent genetic interference in Caenorhabditis elegans. Gene263, 103–112.

Timmons, L., Fire, A., 1998. Specific interference by ingested dsRNA. Nature395, 854.

Tomoyasu, Y., Denell, R.E., 2004. Larval RNAi in Tribolium (Coleoptera) for analyzingadult development. Dev. Genes Evol. 214, 575–578.

Tomoyasu, Y., Miller, S.C., Tomita, S., Schoppmeier, M., Grossman, D., Bucher, G.,2008. Exploring systemic RNA interference in insects: a genome-wide surveyfor RNAi genes in Tribolium. Genome Biol. 9, R10.

Turner, C.T., Davy, M.W., MacDiarmid, R.M., Plummer, K.M., 2006. RNA interferencein the light brown apple moth, Epiphyas postvittana (Walker) induced bydouble-stranded RNA feeding. Insect Mol. Biol. 15, 383–391.

Ulvila, J., Parikka, M., Kleino, A., Sormunen, R., Ezekowitz, R.A., Kocks, C., Ramet, M.,2006. Double-stranded RNA is internalized by scavenger receptor-mediatedendocytosis in Drosophila S2 cells. J. Biol. Chem. 281, 14370–14375.

Walshe, D.P., Lehane, S.M., Lehane, M.J., Haines, L.R., 2009. Prolonged gene knock-down in the tsetse fly Glossina by feeding double stranded RNA. Insect Mol. Biol.18, 11–19.

Wienholds, E., Plasterk, R.H., 2005. MicroRNA function in animal development. FEBSLett. 579, 5911–5922.

Winston, W.M., Molodowitch, C., Hunter, C.P., 2002. Systemic RNAi in C. elegansrequires the putative transmembrane protein SID-1. Science 295, 2456–2459.

Yuen, J.L., Read, S.A., Brubacher, J.L., Singh, A.D., Whyard, S., 2008. Biolistics for high-throughput transformation and RNA interference in Drosophila melanogaster. Fly 2.

Zhou, X., Wheeler, M.M., Oi, F.M., Scharf, M.E., 2008. RNA interference in the termiteReticulitermes flavipes through ingestion of double-stranded RNA. InsectBiochem. Mol. Biol. 38, 805–815.