ORIGINAL PAPER

Jose de la Fuente Æ Consuelo Almazan

Edmour F. Blouin Æ Victoria Naranjo

Katherine M. Kocan

RNA interference screening in ticks for identificationof protective antigens

Received: 26 January 2005 / Accepted: 14 March 2005 / Published online: 12 April 2005� Springer-Verlag 2005

Abstract Ticks are ectoparasites of wild and domesticanimals and humans, and are considered to be the mostimportant arthropod vector of pathogens in NorthAmerica. Development of vaccines directed against tickproteins may effect reduction of tick infestations andtransmission of tick-borne pathogens. The limitingstep for the development of tick vaccines has been theidentification of tick protective antigens. Reverse vacci-nology approaches aimed at reducing animal experi-mentation while allowing for the rapid screening of poolsof potential tick vaccine candidates would greatly facili-tate progress towards the development of tick vaccines.Herein, we describe the screening of Ixodes scapulariscDNAs for identification of tick protective antigensusing RNA interference (RNAi). The results of theRNAi screening were similar to those obtained previ-ously using expression library immunization and dem-onstrated that RNAi could serve as a more rapid andcost-effective tool for vaccine antigen discovery in ticksand in other nonmodel organisms.

Introduction

Ixodes scapularis ticks (Acari: Ixodidae) are the mainvector of human pathogens including Borrelia burgdor-feri, Anaplasma phagocytophilum, Babesia microti andtick-borne encephalitis virus, the causative agents of

Lyme disease, human granulocytic anaplasmosis, hu-man babesiosis and a form of viral encephalitis,respectively (Estrada-Pena and Jongejan 1999; Parolaand Raoult 2001).

Recently, development of vaccines against Boophilusspp. ticks has provided new possibilities for identifica-tion of protective antigens for use in vaccines for controlof tick infestations (de la Fuente and Kocan 2003;Willadsen 2004). Control of ticks by vaccination wouldavoid environmental contamination and selection ofdrug resistant ticks that result from repeated acaricideapplications. Tick vaccines could also be designed toinclude multiple tick and pathogen-derived antigens thatmay target a broad range of both tick species andassociated pathogens (de la Fuente and Kocan 2003).

Development of high throughput screening andsequencing technologies and bioinformatic tools facili-tate the study of biological systems and provide infor-mation for selection of potential vaccine candidates(Rappuoli 2000; de la Fuente and Kocan 2003). How-ever, identification of tick protective antigens remainsthe limiting step for the development of new tick vac-cines. Recently, we reported the use of ExpressionLibrary Immunization (ELI) combined with analysisof expressed sequence tags (EST) in a mouse model oftick infestations that resulted in the identification of351 cDNA clones that affected I. scapularis larvaldevelopment (Almazan et al. 2003a). Two of theseclones, 4D8 of unknown function and 4F8 with nucle-otidase activity, were protective against larval infesta-tions when used individually in cDNA and recombinantprotein immunization experiments in mice (Almazanet al. 2003a, b; submitted for publication). Although theapplication of ELI in combination with EST analysisrepresented an improved method for identification oftick protective antigens, immunization and tick infesta-tion of a large number of animals result in experimentsthat are laborious, expensive and difficult to standardize.Furthermore, while the mouse model of tick infestationsis useful for selected tick species, many other tick specieswould require larger animal hosts for tick feeing, such

J. de la Fuente (&) Æ C. Almazan Æ E. F. Blouin Æ K. M. KocanDepartment of Veterinary Pathobiology,Center for Veterinary Health Sciences,Oklahoma State University,Stillwater, OK 74078, USAE-mail: jose_delafuente@yahoo.comTel.: +1-405-7440372Fax: +1-405-7445275

J. de la Fuente Æ V. NaranjoInstituto de Investigacion en Recursos CinegeticosIREC (CSIC-UCLM-JCCM),Ronda de Toledo s/n,13005 Ciudad Real, Spain

Parasitol Res (2005) 96: 137–141DOI 10.1007/s00436-005-1351-5

as rabbits, sheep and cattle, and would thus contributeto the complexity of the screening experiments.

In an attempt to simplify the screening methods forthe identification of tick protective antigens, we appliedthe posttranscriptional gene silencing by RNA interfer-ence (RNAi) to pools of I. scapularis cDNAs used pre-viously in ELI experiments (Almazan et al. 2003a).RNAi provides a powerful alternative to traditionalgenetics and, thus, has revolutionized the analysis ofgene function in nonmodel organisms including ticks(Aljamali et al. 2002, 2003; Karim et al. 2004a, b;Narasimhan et al. 2004; Pal et al. 2004; Miyoshi et al.2004; Mello and Conte 2004). RNAi is mediated by longdouble-stranded RNA (dsRNA) or small interferingRNAs (siRNAs). Introduction of dsRNA or siRNAinto a cell triggers destruction of the cognate mRNA andallows for analysis of resulting phenotype (Mello andConte 2004).

Materials and methods

Ticks

I. scapularis unfed adult ticks were obtained from thelaboratory colony maintained at the Oklahoma StateUniversity Tick Rearing Facility. Larvae and nymphswere fed on rabbits and adult ticks were fed on sheep.Animals were housed at the Tick Rearing Laboratorywith the approval and supervision of the OSU Institu-tional Animal Care and Use Committee. Off-host tickswere maintained in a 12 h light: 12 h dark photoperiodat 22–25�C and 95% relative humidity.

Generation of dsRNA pools for screeningof tick cDNAs

For the screening of I. scapularis cDNAs, a library madein the expression vector pEXP1 (Clontech, Palo Alto,CA, USA) was used (Almazan et al. 2003a). Oligonu-cleotide primers (PEXP1T75: 5¢-TAATACGACTCAC-TATAGGGTACTGGCCGCGTCGACGGAATTCG-T-3¢ and PEXP1T73: 5¢-TAATACGACTCACTA-TAGGGTACTGATGCATGCTCGACCCGATGTT-3¢)were synthesized specific for vector DNA sequencesflanking the tick cDNA insert and containing T7 pro-moter sequences for in vitro transcription and synthesisof dsRNA. Eight pools of 12 plasmid cDNA clonescontaining approximately 1 ng of each clone were madeand used as templates for PCR. Tick cDNA pools cor-responded to rows A to H of the 96-wells plate No. 4 ofthe ELI secondary screening and EST analysis (Almazanet al. 2003a). PCR reactions were performed using theAccess RT-PCR system (Promega, Madison, WI, USA)in a 50-ll reaction mixture. The resultant ampliconswere purified (Wizard PCR purification system, Pro-mega) and 8 ll were used for in vitro transcription andpurification of dsRNA using the Megascript RNAi kit

(Ambion, Austin, TX, USA). The dsRNA was purifiedand quantified by spectrometry.

Generation of dsRNA from individual tick cDNAclones

For 4A8, 4D8, 4F8 (cDNA identification refers to theposition of the clone in the 96-well plate) and actindsRNA synthesis, unfed adult salivary gland RNA wasprepared using TRI Reagent (Sigma, St. Louis, MO,USA) following manufacturer’s instructions and used astemplate to amplify DNA encoding the full-length 4D8(555 bp; GenBank accession no. AY652654), and partialfragments encoding 4A8 (584 bp; GenBank accessionno. CD052546), 4F8 (666 bp; GenBank accession no.AY652655) and I. scapularis actin (408 bp; GenBankaccession no. AF426178). Gene-specific primers con-taining T7 promoter sequences for in vitro transcriptionand synthesis of dsRNA were used in the RT-PCRs. Theprimer sequences were as follows: 4D8, D8T75: 5¢-TA-ATACGACTCACTATAGGGTACTATGGCTTGC-GCAACATTAAAG-3¢ and D8T73: 5¢-TAATACGA-CTCACTATAGGGTACTTTATGACAAATAGCTT-GGAG-3¢; 4A8, A8T75: 5¢-GACCCTGGAATGCTG-GAGCATACTATGGCTTGCGCAACATTAAAG-3¢and 5¢-TAATACGACTCACTATAGGGTACTGTCC-CTGCCAGCCGGAGGG-3¢; 4F8, F8T75: 5¢-TAAT-ACGACTCACTATAGGGTACTAGTAATTTCCAA-CACTGTT-3¢ and F8T73: 5¢-TAATACGACTCACTA-TAGGGTACTCCCTCAATCAACAGCAGCA-3¢ andactin, ACT75: 5¢-TAATACGACTCACTATAGGGTA-CTGAGAAGATGACCCAGATCA-3¢ and ACT73: 5¢-TAATACGACTCACTATAGGGTACT-GTTGCCG-ATGGTGATCACC-3¢. RT-PCR and dsRNA synthesisreactions were performed as described above for dsRNApools.

Injection of ticks with dsRNA

Ticks were injected with approximately 0.2 ll of dsRNA(5·109–5·1010 molecules per ll corresponding todsRNA pools or 6·1010–1·1011 molecules per ll corre-sponding to 4A8, 4D8, 4F8 and actin dsRNA) in thelower right quadrant of the ventral surface of the exo-skeleton of female I. scapularis. The injections were donewith a Hamilton syringe with a 1-inch, 33 gauge needle.Control ticks were injected with 0.2 ll of injection buffer(10 mM Tris–HCl, pH 7, 1 mM EDTA) or were leftuninjected. I. scapularis actin was used as positive con-trol (Narasimhan et al. 2004). Fifty ticks were used ineach group. The ticks were held in a humidity chamberfor 1 day after which they were allowed to feed on asheep with male ticks. Female ticks that fed to repletionor those that were removed from the sheep after 10 daysof feeding were collected, weighed and evaluated foroviposition by weighting the egg mass oviposited by allticks in the group.

138

Analysis by RT-PCR to confirm gene silencing

Salivary glands and guts were dissected from groups offive ticks from mock-injected and dsRNA-injectedgroups after feeding. Total RNA was isolated and ana-lyzed for transcription of target genes by RT-PCR asdescribed above using oligonucleotide primers for thepreparation of actin dsRNA or designed for expressionanalysis of 4D8 (4D8R5: 5¢-GCTTGCGCAACATTAA-AGCGAAC-3¢, 4D8-R: 5¢-TGCTTGTTTGCAGATG-CCCATCA-3¢) and 4F8 (4F8R5: 5¢-GCGTCGTGTG-GAGCATCAGCGAC-3¢, 4F8-R: 5¢-TCGCAACGGA-CAACGGCAGGTTG-3¢). Control reactions wereperformed using the same procedures but without RT tocontrol for DNA contamination in the RNA prepara-tions and without RNA added to control contaminationof the PCR reaction. PCR products were electrophore-sed on 1% agarose gels to check the size of amplifiedfragments by comparison to a DNA molecular weightmarker (1 Kb Plus DNA Ladder, Promega).

Results and Discussion

In this study, dsRNA was prepared from eight pools of12 I. scapularis cDNAs each derived from the 96-wellplate No. 4 (one pool per plate rows A to H) of the ELIsecondary screening and EST analysis containing thecDNA clones encoding for the protective antigens 4D8and 4F8 (Almazan et al. 2003a). Pools of dsRNA werechecked by gel electrophoresis and contained fragmentsof different sizes as predicted from the differences in thesize of tick cDNAs (data not shown).

Pools of dsRNA were injected into unfed I. scapularisfemales which subsequently were allowed to feed on asheep. Female ticks that fed to repletion or those thatremained attached after 10 days were collected, weighedand evaluated for oviposition. As in previous experi-ments using ELI (Almazan et al. 2003a), only dsRNApools D and F caused a reduction in the weight of re-plete ticks when compared with the control ticks mock-injected with buffer alone (Fig. 1a, b). As previouslyreported (Narasimhan et al. 2004), tick weight was alsoreduced in positive control ticks that were injected withactin dsRNA (Fig. 1a, b). Tick feeding was slower ingroups D and F in which >80% of replete ticks fell offthe host in the last two days of repletion (data notshown).

When tick oviposition was analyzed, it was abolishedin the groups injected with dsRNA pools B, D, E and Fand in the group injected with actin dsRNA (Fig. 1c). Ingroups injected with dsRNA pools B and E, tick weightwas not affected (Fig. 1b) but the oviposition wasabolished (Fig. 1c), suggesting that the effect on tickweight was not linked to the abolishment of ovipositionin these groups. This result also showed that cDNApools not protective against tick infestations by ELI maybe selected in RNAi experiments for their inhibitoryeffect on tick oviposition.

dsRNAs were then made from individual clones inthe negative pool A (4A8) and in the positive pools D(4D8) and F (4F8), and injected into unfed I. scapularisfemale ticks. I. scapularis actin was used again as posi-tive control. Ticks were collected after repletion,weighed and evaluated for oviposition. The results againreproduced those obtained by ELI (Almazan et al.2003a), demonstrating that 4D8 and 4F8 reduced tickweight by RNAi when compared to control ticks andticks injected with 4A8 dsRNA (Fig. 2). Oviposition wasabolished in the groups injected with 4D8 and actindsRNA but not in the groups injected with 4A8 and 4F8dsRNA (Fig. 2), suggesting that in dsRNA pool F othersequences, or the synergistic effect of several sequences,were responsible for the abolishment of tick ovipositionafter injection of pool F dsRNA (Fig. 1c). Actin, 4D8and 4F8 gene silencing was confirmed by RT-PCR(Fig. 3) and control reactions ruled out contaminationwith genomic DNA or during the PCR (data notshown).

The results reported herein demonstrated the useful-ness of RNAi for the identification of tick protectiveantigens. dsRNA pools from a tick cDNA library wereprepared and screened, and the results were similar tothose obtain before by ELI (Almazan et al. 2003a).However, in the ELI experiments, 3–5 mice per groupwere immunized prior to the evaluation of the effect ofthe vaccine on tick infestation. For the RNAi screeningstudy reported herein, only one sheep was required forfeeding all tick groups, including the controls and thoseinjected with the dsRNA derived from cDNA clones ofone 96-well plate. The screening of tick cDNAs byRNAi not only minimized resources by decreasing thenumber of animals required, but also allowed to com-plete the screening in a 2 week period rather than thetwo months required by ELI (Almazan et al. 2003a). Thereproducibility of the results was also better defined byRNAi than by ELI. Animal-to-animal variabilityencountered in ELI and other screening methods wasreduced by using the same host to feed several experi-mental tick groups in RNAi experiments. For example,the variance (VARP) of cDNA immunization experi-ments with 4D8 was VARP=106 (N=3) compared toVARP=25 (N=3) by RNAi.

It could be argued that some cDNAs with protectiveproperties will not be detected by RNAi because they donot affect tick feeding and/or oviposition, but this po-tential drawback would be present in other screeningmethods including ELI. The higher reproducibility ofthe RNAi results should help in detecting protectiveclones otherwise eliminated by the statistical analysis.Furthermore, as shown herein, cDNA pools not pro-tective in ELI experiments (pools B and E; Almazanet al. 2003a) effectively affected tick oviposition whentested as dsRNA pools by RNAi.

Although RNAi has been reported previously in ticks(Aljamali et al. 2002, 2003; Karim et al. 2004a, b;Narasimhan et al. 2004; Pal et al. 2004; Miyoshi et al.2004) including I. scapularis (Narasimhan et al. 2004;

139

Pal et al. 2004), this is the first report demonstrating itsuse for the screening and identification of tick protectiveantigens. The 4D8 and 4F8 dsRNAs, which decreasedoviposition and/or tick weight by RNAi, have been

demonstrated to control tick infestations after DNA andrecombinant protein immunization experiments in mice(Almazan et al. 2003a, b; submitted for publication),thus validating the results of the RNAi experiments. The

Fig. 1 Screening of I. scapulariscDNAs by RNAi. The effect ofdsRNA pools on engorged tickweight and oviposition wasevaluated. a Fifty femaleI. scapularis per group wereinjected with dsRNA pools(A–H), actin dsRNA (C+),mock-injected (C�) or leftuninjected (U) and collectedafter completion of the feedingperiod for analysis of tickweight. b The engorgementweights were recorded and theaverage (±SD) was comparedbetween mock-injected controlsand dsRNA-injected oruninjected ticks (Black bars,P<0.01; Student’s t test).c Oviposition was evaluated byweighting the egg massoviposited by all ticks in eachgroup

Fig. 2 Effect on tick feedingand oviposition of targeting byRNAi of I. scapularis 4A8, 4D8,4F8 and actin. Fifty unfedfemale I. scapularis per groupwere injected with dsRNAcomplementary to each gene orleft uninjected for analysis oftick weight and ovipositionafter completion of the feedingperiod. The engorgement andegg mass weight were recordedand the average (±SD) wascompared between dsRNA-injected and uninjected controls(aP<0.005; Student’s t test).Three independent experimentswere conducted with similarresults

140

experimental design described herein could be robotizedand used for high throughput screening of tick cDNAlibraries for the identification of antigens protectiveagainst tick infestations and the transmission of tick-borne pathogens for the development of improved tickvaccines. Additionally, this approach could be adaptedfor screening cDNA and DNA libraries in other para-sites for which transformation methods are not availablefor the generation and analysis of genetic mutants.

Acknowledgements This research was supported by the OklahomaAgricultural Experiment Station (project 1669), the SitlingtonEndowed Chair for Food Animal Research (K. M. Kocan,Oklahoma State University). Consuelo Almazan was supported byPfizer Animal Health, Kalamazoo, MI, and a grant-in-aid from theCONACYT and Promep (University of Tamaulipas), Mexico.V. Naranjo was founded by Consejerıa de Educacion, JCCM,Spain. These experiments comply with the current laws of theU.S.A.

References

Aljamali MN, Sauer JR, Essenberg RC (2002) RNA interference:applicability in tick research. Exp Appl Acarol 28:89–96

Aljamali MN, Bior AD, Sauer JR, Essenberg RC (2003) RNAinterference in ticks: a study using histamine binding proteindsRNA in the female tick Amblyomma americanum. Insect MolBiol 12:299–305

Almazan C, Kocan KM, Bergman DK, Garcia-Garcia JC, BlouinEF, de la Fuente J (2003a) Identification of protective antigensfor the control of Ixodes scapularis infestations using cDNAexpression library immunization. Vaccine 21:1492–1501

Almazan C, Kocan KM, Bergman DK, Garcia-Garcia JC, BlouinEF, de la Fuente J (2003b) Characterization of genes tran-scribed in an Ixodes scapularis cell line that were identified byexpression library immunization and analysis of expressed se-quence tags. Gene Ther Mol Biol 7:43–59

Estrada-Pena A, Jongejan F (1999) Ticks feeding on humans: areview of records on human-biting Ixodoidea with special ref-erence to pathogen transmission. Exp Appl Acarol 23:685–715

de la Fuente J, Kocan KM (2003) Advances in the identificationand characterization of protective antigens for development ofrecombinant vaccines against tick infestations. Expert RevVaccines 2:583–593

Karim S, Ramakrishnan VG, Tucker JS, Essenberg RC, Sauer JR(2004a) Amblyomma americanum salivary glands: double-stranded RNA-mediated gene silencing of synaptobrevinhomologue and inhibition of PGE2 stimulated protein secre-tion. Insect Biochem Mol Biol 34:407–413

Karim S, Ramakrishnan VG, Tucker JS, Essenberg RC, Sauer JR(2004b) Amblyomma americanum salivary gland homolog ofnSec1 is essential for saliva protein secretion. Biochem BiophysRes Commun 324:1256–1263

Mello CC, Conte D Jr (2004) Revealing the world of RNA inter-ference. Nature 431:338–342

Miyoshi T, Tsuji N, Islam MK, Kamio T, Fujisaki K (2004) Genesilencing of a cubilin-related serine proteinase from the hardtick Haemaphysalis longicornis by RNA interference. J Vet MedSci 66:1471–1473

Narasimhan S, Montgomery RR, DePonte K, Tschudi C, Mar-cantonio N, Anderson JF, Sauer JR, Cappello M, Kantor FS,Fikrig E (2004) Disruption of Ixodes scapularis anticoagulationby using RNA interference. Proc Natl Acad Sci USA 101:1141–1146

Pal U, Li X, Wang T, Montgomery RR, Ramamoorthi N, DesilvaAM, Bao F, Yang X, Pypaert M, Pradhan D, Kantor FS,Telford S, Anderson JF, Fikrig E (2004) TROSPA, an Ixodesscapularis receptor for Borrelia burgdorferi. Cell 119:457–468

Parola P, Raoult D (2001) Tick-borne bacterial diseases emergingin Europe. Clin Microbiol Infect 7:80–83

Rappuoli R (2000) Reverse vaccinology. Curr Opin Microbiol3:445–450

Willadsen P (2004) Anti-tick vaccines. Parasitology 129:S1–S21

Fig. 3 Effect of dsRNA injection on the expression of I. scapularis4D8, 4F8 and actin. Total RNA was isolated from five dsRNA andmock-injected (Control) ticks after feeding and the expression oftargeted genes was analyzed by RT-PCR

141