Gene 474 (2011) 12–21

Contents lists available at ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r.com/ locate /gene

A novel gene from the takeout family involved in termite trail-following behavior

Margaret A. Schwinghammer a,1, Xuguo Zhou b,2, Srinivas Kambhampati c,Gary W. Bennett a, Michael E. Scharf a,b,⁎a Department of Entomology, Purdue University, West Lafayette, IN, USAb Department of Entomology and Nematology, University of Florida, Gainesville, FL, USAc Department of Entomology, Kansas State University, Manhattan, KS, USA

Abbreviations: ANOVA, Analysis of variance; dsRNexpressed sequence tag, JH, juvenile hormone; JHBP, juvePCR, polymerase chain reaction; K-W ANOVA, Kruskalbinding protein; qRT-PCR, quantitative real-time PCR;adenine dinucleotide; RNAi, RNA interference; siRNA, shborate-EDTA.⁎ Corresponding author. O.W. Rollins/Orkin Cha

901 W. State St., Purdue University, West Lafayette, I765 496 6710; fax: +1 765 494 0535.

E-mail address: mscharf@purdue.edu (M.E. Scharf).1 Current address: CNA, 4825 Mark Center Dr, Alexan2 Current address: Department of Entomology, Univer

USA.

0378-1119/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.gene.2010.11.012

a b s t r a c t

a r t i c l e i n f o

Article history:Accepted 26 November 2010Available online 4 December 2010

Received by I. King Jordan

Keywords:TakeoutRNA interferenceBehavioral geneTrail pheromone2-phenoxyethanol

This study investigated physiological and behavioral functions of a novel gene identified from the termiteReticulitermes flavipes. The gene, named deviate, encodes an apparent ligand binding protein from the takeout-homologous family. Initial studies were conducted to investigate deviate mRNA expression among termitecastes and body regions, and changes in response to light–dark conditions, starvation, temperature, andjuvenile hormone (JH). Deviate has ubiquitous caste and tissue expression, including antennal expression.Consistent with characteristics of other takeout family members, deviate expression is responsive tophotophase conditions (pb0.1), and feeding, temperature, and JH (pb0.05). Using RNA-interference (RNAi)techniques, short-interfering RNAs (siRNAs) homologous to the deviate gene were synthesized and injectedinto worker termites, which were then subjected to bioassays designed to (1) induce caste differentiation or(2) measure various behavioral aspects of foraging and trail following. No impacts on JH-dependent castedifferentiation were observable. However, trail following accuracy was significantly reduced in termites thatreceived deviate siRNA injections, and this pattern generally mirrored deviatemRNA attenuation and recoveryafter RNAi. In a subsequent distance foraging bioassay, deviate-silenced termites exhibited equal feedinglevels to controls, suggesting the deviate gene is not linked to general vigor or the ability/motivation oftermites to move and forage. These findings are among the first linking the expression of a termite gene witheusocial behavior; they illustrate the connection between deviate expression and trailing behavior, which is akey evolutionary adaptation vital to subterranean social insects such as termites and ants.

A, double-stranded RNA; EST,nile hormone binding protein;–Wallis ANOVA; OBP, odorantNADH, reduced nicotinamideort-interfering RNA; TBE, tris-

ir, Entomology Department,N 47907-2089, USA. Tel.: +1

dria, VA, USA.sity of Kentucky, Lexington, KY,

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

The takeout-homologous gene family is a well-studied family ofinsect genes that was initially characterized in Drosophila (Sarov-Blatet al., 2000; So et al., 2000). In Drosophila, takeout was found to havesimilarities to OBPs and hemolymph JHBP. Takeout and the related“moling” and “JP29” proteins found in Manduca sexta are all classifiedas hydrophobic ligand binding proteins (Touhara and Prestwich,1992; Wojtasek and Prestwich, 1995; Shinoda et al., 1997; Saito et al.,

2006; Robertson et al., 1999; Du et al., 2003; Hiruma and Riddiford,2010). Expression of takeout mRNA in Drosophila is located instructures associated with feeding, specifically the head, cardia,crop, and antennae, and it has been linked to feeding behavior, foodlocation, and longevity (Sarov-Blat et al., 2000; So et al., 2000; Wonget al., 2009; Gáliková and Flatt, 2010; Bauer et al., 2010; Benito et al.,2010). In Drosophila, starvation induces the expression of takeoutmRNA; however, this process does not occur in mutants withdisrupted circadian clocks, implying a link between takeout, circadianclocks, and the availability of nutrients (Sarov-Blat et al., 2000).Takeout mutants also display abnormal locomotory activities and dierapidly during periods of starvation, suggesting a connection betweenlocomotory activity, survival, and the availability of food (Sarov-Blatet al., 2000). TakeoutmRNA levels undergo a daily cycle, and takeout isdown-regulated or undetectable in circadian mutant flies (So et al.,2000; Sarov-Blat et al., 2000). After 9–10 hours of starvation, takeoutmRNA levels increase, while a 2–hour re-feeding period reversesthese effects (Sarov-Blat et al., 2000). Based on the above findings, ithas been suggested that takeout may contribute to an organism'santicipation of food resource availability, or play a role in its responseto conditions of changing food availability (Sarov-Blat et al., 2000;Wong et al., 2009). Additionally, wildtype flies exhibit daily

13M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

fluctuations in takeout mRNA levels in cycling photoperiodic condi-tions as well as constant dark conditions, implying the daily cycling isa function of the insect's internal clock rather than being externallylight driven (So et al., 2000). Since the initial characterization oftakeout in Drosophila, members of the takeout/ligand binding proteinfamily have been identified in a number of insects, including termites(Ishida et al., 2002a; Hojo et al., 2005). A takeout-homologous genewith brain expression has also been linked to long-distance migrationin the monarch butterfly, Danaus plexippus (Zhu et al., 2008).

Trail following behavior in termites is an experimentallytractable and important social behavior for which no genetic basishas yet been identified. Subterranean termites live in eusocialcolonies, spending most of their lives underground and in completedarkness (Wilson, 1971). The vast majority of individuals insubterranean termite colonies are eyeless and therefore must usechemical trails to navigate through the colony's network ofunderground tunnels, and to communicate and convey messagesfor exploration, defense, and food harvesting among nest mates(Stuart, 1967; Moore, 1966; Matsumura et al., 1968; Wilson, 1971).The functions of termite pheromones secreted from sternal andlabial glands have been investigated and are well understood.Sternal gland secretions containing ZZE-dodecatrienol function astrail pheromone, which is used for orientation and signaling forcommunal exploitation of food resources (Matsumura et al., 1968).During the exploitation of a food source, secretions from the labialgland containing the active compound hydroquinone are releasedonto food, stimulating feeding by conspecific nestmates (Reinhardand Kaib, 1995). The transfer of food and its transportation back tothe nest occur primarily on trails where a high concentration ofsternal gland secretions have been deposited; whereas, gnawingbehavior is concentrated at sites where labial gland secretions arepresent (Reinhard and Kaib, 1995). Interestingly, some lowertermite species of the genera Reticulitermes and Coptotermes willreadily follow trails of fresh ball-point pen ink containing thecompound 2-phenoxyethanol (Chen et al., 1998).

RNA-interference (RNAi) is an important functional genomics toolfor the study of gene function. RNAi, also called post-transcriptionalgene silencing or transgene silencing (Hannon, 2002), is a process thatuses dsRNA which shares a homologous sequence with a target geneto interfere with the gene's expression (Fire et al., 1998). Thebiological origin of the RNAi pathway is thought to be associatedwith the immune response, protecting organisms from exogenousgenetic material (Gura, 2000). Several recent studies have demon-strated that RNAi is achievable in termites and that it can revealimportant insights into termite biochemistry, physiology, sociality,and behavior (Zhou et al., 2006a, 2006b, 2007, 2008; Scharf et al.,2008; Korb et al., 2009). As behavior is the outward manifestation ofgene expression, this study used RNAi to investigate the influence of atermite takeout homolog hypothesized to be involved in trailfollowing behavior. Additionally, this study examined how the RNA-interference pathway can be enlisted to examine the roles of specificgenes in the modulation of downstream behaviors and castedifferentiation processes (Robinson et al., 2005; Smith et al., 2008).

This paper reports sequence, expression, and functional data forthe deviate gene of the termite Reticulitermes flavipes, which has thehighest homology to members of the takeout family of ligand bindingproteins. Given the various behavioral and physiological roles fortakeout family members in Drosophila and other insects, this researchwas conducted as a candidate gene study into the potential roles of R.flavipes deviate. Information presented here includes tissue and castedistribution of the deviate transcript; deviate transcript expressionchanges in response to JH, temperature, feeding and light–darkconditions; and deviate silencing by RNAi. To test the hypotheses thatthe deviate gene is linked to trail-following, foraging, and/or feedingbehavior, bioassays were designed and conducted to investigate trailfollowing behavior in association with RNAi-attenuated deviate gene

expression. Additionally, to test the hypotheses that deviate mightplay roles in regulating or mediating temperature-dependent castedifferentiation (Scharf et al., 2007), studies were conducted toinvestigate the impacts of temperature and JH on expression, and byusing RNAi to silence deviate expression in combination with JH-dependent caste differentiation bioassays. A recent publication byKorb et al. (2009) provided the first evidence of a termite behavioralgene by using RNAi to silence a gene encoding a putative odorant-releasing enzyme linked to suppression of aggressive behaviors inCryptotermes termites. Here, we provide a second and very differentexample of a termite behavioral gene: one with ubiquitous expressionthat mediates trail following accuracy, and that is responsive to JH,temperature, feeding, and light/dark conditions.

2. Materials and methods

2.1. Termites

Termites were collected from logs at a site on the PurdueUniversity campus (West Lafayette, Indiana, USA). Termites wereidentified as R. flavipes based on soldier morphology (Nutting, 1990).Workers, soldiers, and nymphs were isolated from logs and housed inplastic containers with moistened wood and paper towels at ca. 22 °Cand 97% relative humidity, in full darkness. These laboratory coloniesbecame reproductively competent within a few weeks and were fullyself-sufficient thereafter.

2.2. Deviate gene sequencing and sequence analyses

The full-length cDNA sequence for the deviate gene (Genbankaccession no HQ003932) was assembled from multiple overlappingclones from several different cDNA libraries (Wu-Scharf et al., 2003;Tartar et al., 2009; Steller et al., 2010) (accession nos. BQ788174,GO902234, FL639426, GO900869, GO910038, GO901465, FL637875,GO899218, FL639883, FL638424, FL637493, and FL636085). Themajority of the ORF sequence was verified by PCR amplificationof an 864 bp fragment of the cDNA spanning cDNA nucleotides 27–824 (Fig. 1), using the forward and reverse primers 5′-CCCCCTCGAGTTGTTTTATG-3′ and 5′-GAAGGGAACTTGCAGGAACA-3′.Database searches were performed at NCBI (http://www.ncbi.nlm.nih.gov/) using blastn and blastx under default settings. Peptidetranslations were made using SDSC Biology Workbench (http://workbench.sdsc.edu/). Signal peptide and N-glycosylation analyseswere performed using the SignalP and NetNGlyc 1.0 servers availableat http://www.cbs.dtu.dk/services/ and http://www.cbs.dtu.dk/ser-vices/NetNGlyc/. Protein alignments and cladograms were generatedusing ClustalV in the Lasergene software package (DNAstar; Madison,WI). Cladogram analyses were performed with 100 bootstrapreplicates of parsimony analysis using PAUP* 4.0b10.

2.3. Tissue and caste expression profiles

Visualization of conventional PCR products on agarose gels wasused to examine deviate expression across five body regions (wholebody, antennae, head, thorax, and abdomen), and three castephenotypes (workers, soldiers, and nymphs). Total RNA was isolatedfrom homogenized termite samples using the RNeasy Mini Total RNAKit (Qiagen; Valencia, CA). Three RNA isolations were made per tissueand caste phenotype from two replicated colonies. Per replicate,fifteen whole termites or body regions (head, thorax, or abdomen),or 40–50 individual antennae were used. RNA quality and quantitywere determined by gel electrophoresis and spectrophotometry.cDNA was synthesized from total RNA using the Invitrogen Super-Script First-Strand Synthesis System (Invitrogen; Carlsbad, CA).PCR was performed using 1 μl of cDNA, iQ SYBR Green Supermix(Bio-Rad: Hercules, CA) and deviate gene specific primers. The deviate

Fig. 1. cDNA and translated amino acid sequence for the R. flavipes deviate gene (Genbank accession No. HQ003932). Nucleotides are shown in lower case letters and amino acids incapital letters above the first nucleotide in each codon triplet. The predicted start (atg) and stop (taa) codons, polyadenylation signal (aataa), and poly-A tail (aaaan) are shown inbold text with underlining. Predicted glycosylated amino acids (N) are indicated by gray shading. Two highly conserved cysteine residues of takeout and JHBP proteins (involved inligand binding) are shown as black boxes with white text. Dotted underlining denotes the predicted 16 amino acid signal peptide with cleavage site (▼).

14 M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

forward and reverse primer sequences, respectively, were 5′-AGGCGAGTTGGTGACCATAG-3′ and 5′-GTAGAAGGCAGGCCAGTACG-3′. NADH-dh (BQ788168; Wu-Scharf et al., 2003) was used as areference gene. The NADH-dh forward and reverse primer sequences,respectively, were 5′-GCTGGGGGTGTTATTCATTCCTA-3′ and 5′-GGCATACCACAAAGAGCAAAA-3′. PCR products were visualized on1.75% agarose-TBE gels containing 0.25% Synergel (DiversifiedBiotech; Boston, MA).

2.4. Light–dark assays and quantitative real-time PCR methods

Bioassays were conducted in 35 mm diam. Petri dishes using 15termites per dish from three replicate colonies as described previously(Scharf et al., 2008). Three treatments were tested, all at 27 °C:continuous darkness, continuous light, or 12:12 hr light:dark. For thecontinuous darkness treatment, all manipulations were performed ina dark room under red light. After 15 days, termites were frozen at-80 °C. Total RNA was isolated from frozen whole termites using theSV Total RNA Isolation System, which included on-filter DNAsetreatment (Promega; Madison, WI). cDNA was synthesized from totalRNA using the iScript cDNA synthesis kit (Bio-Rad). qRT-PCR wasperformed using iQ SYBR Green Supermix (Bio-Rad; Hercules, CA),deviate gene specific primers, and β-actin and NADH-dh as reference/control gene primers. The deviate and NADH-dh primers are describedabove (Section 2.3). The β-actin (DQ206832; Zhou et al., 2006b)forward and reverse primer sequences were 5′-AGAGG-GAAATCGTGCGTGAC-3′ and 5′-CAATAGTGATGACCTGGCCGT-3′.

qRT-PCR threshold cycle (CT) values were used to determinerelative mRNA expression. CT values for the deviate PCR reactionswere normalized to those of two reference genes (β-actin and NADH-dh) using the BestKeeper method (Pfaffl et al., 2004). Relativeexpression levels of the deviate gene were then calculated using the2−ΔΔCT method (Livak and Schmittgen, 2001).

2.5. Starvation assays

To test the effect of starvation and feeding on deviate geneexpression, groups of termites were held without a food source for 0,5, 10, or 15 days and compared to termites held under similarconditions but with food. In the “fed” treatment, 15 termites fromthree replicate colonies were placed in 35 mm diam Petri dishes withone disk of double-ply paper towel as described previously (Scharfet al., 2008). In the starvation treatments, termites were placed inPetri dishes as above, but with one double-ply disk of moistenedpaper towel attached to the inside-top cover of the dish (inaccessibleto termites). Assays were replicated across three colonies pertreatment and ran for 15 days at 27 °C. Every 5 days, individual Petridishes were opened and 100 μl of distilled water added to the papertowel disks either lining the bottom of the dish (control treatment) orthe top of the lid (starvation treatment). At 5, 10, and 15 days,termites were removed from assay dishes and frozen at −80 °C.Termite mortality never exceeded 1 individual per replicate. TotalRNA was isolated, cDNA synthesized, and qRT-PCR performed andanalyzed as described above (Section 2.4).

2.6. Temperature and JH assays

To test the effect of temperature and JH on deviate gene expression,groups of termites were held at 17, 22 and 27 °C for 15 days in thepresence and absence of 150 μg ectopic JH III. This method is well-established for its ability to induce worker-to-presoldier morphogen-esis (Zhou et al., 2006a; Scharf et al., 2007; Scharf et al., 2008). Assayswere set up and run as described in preceding sections (2.4, 2.5)with 3colony replicates per treatment. Environmental chambers were usedto attain the three temperatures. Each assay paper received 150 μg JHIII (Sigma-Aldrich; St. Louis, MO) and was left to dry for 30 min beforeplacing in assay dishes. For control treatments, paper towel disks

15M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

received acetone alone. RNAwas isolated, cDNA synthesized, and qRT-PCR performed and analyzed as described above (Section 2.4). Allreported results were normalized to 27 °C data.

2.7. siRNA production

Total RNA was isolated and cDNA synthesized as described above(Section 2.4). A ~500 bp deviate gene cDNA fragment was PCR-amplified using the gene specific primers “deviate RNAi forward” (5′-AGCAGGAGGGTCAATACAAG-3′) and “deviate RNAi reverse” (5′-CTGCGAAGGGTAACATCT-3′). siRNA was synthesized following theSilencer siRNA Cocktail Kit™ (Ambion; Austin, TX) protocol andScharf et al. (2008), as described in detail elsewhere (Scharf et al.,2008). The concentration of the siRNA sample was measured byspectrophotometry and diluted to a final concentration of 0.5 ng per36.0 nl.

2.8. Microinjection of siRNA and RNAi knockdown quantification

Microinjections were performed using a custom-designed plat-form and vacuum manifold, a Nanoliter 2000™ injector (WPI Inc.;Sarasota, FL), and custom-pulled glass capillary needles back-filledwith silicone oil (Zhou et al., 2006a; Scharf et al., 2008). Workertermites were isolated from lab colonies and placed in dishes on iceshortly before injection, which slowedmovements andmade termiteseasier to handle during the injection process. Termites quickly becameactive again after the chilling and injection procedure and non-injected termites were also chilled for the same amount of time.

siRNA solution (36 nL containing 0.5 ng siRNA; Zhou et al., 2006a)was injected into the lateral thorax, above the coxae of the legs (Zhouet al., 2006a; Scharf et al., 2008). Controls were injected withnuclease-free nanopure water alone. After injection, groups of 15termites were placed in 35×10 mm dishes with double-ply disks ofmoistened paper towel lining the bottoms of the dishes. Five of thesedishes were then placed inside a covered plastic box, lined withmoistened paper towels. Non-injected termites, termites injectedwith distilled water, and termites injected with deviate siRNA wereremoved from holding dishes 0, 1, 2, and 3 days after injection. RNAwas isolated, cDNA synthesized, and qRT-PCR performed andanalyzed using two reference genes as described above (Section 2.4).

2.9. RNAi bioassays

2.9.1. JH bioassaysIn the first set of RNAi bioassays, siRNA-injected and control

termites from three replicate colonies were subjected to JH bioassaysand observed for presoldier differentiation as described previously(Section 2.6; Zhou et al, 2006a).

2.9.2. Trail following bioassaysIn the second set of RNAi bioassays, termites from three replicated

colonies were tested for their ability to follow simulated trailsprepared with black pen ink containing the artificial trail pheromone2-phenoxyethanol (Chen et al., 1998). To create a template for thetrail, a dark 30 cm line was drawn in pencil on a 21.5×35.5 cm(11×14") sheet of paper. Additionally, two dark outside “deviation”lines were drawn 0.5 cm on either side of this line. A fresh sheet ofpaperwas placed on top of the template sheet and a ruler used to drawa straight 30 cm line with a black ink BiC™ Round Stick (mediumpoint) (Clearwater, FL) (Suppl. Fig. 1). A 3.7 cm base diameter plasticcup with a small exit hole on the side facing the line was placed overthe termite (for containment) until the trail was located. To test theability of termites to cross a break in the simulated trail, a 2 cmgapwascreated at the end of the trail. Ink trails were less than 30 s old at thetime termites were introduced. Fresh sheets of paper were used foreach trail, and no trail was usedmore than once. Seventy-five termites

per replicate each received deviate siRNA injections, injections ofnuclease-free water, or were chilled but remained un-injected.Termites from each treatment were divided into five small Petridishes, for a total of 15 termites per dish. Disks of double-ply papertowel were placed at the bottom of each dish and moistened with100 μl of distilledwater. The 5 dishes from each treatmentwere placedinside separate plastic boxes lined with moistened papers towels andthe boxes were stored in a growth chamber at 27 °C for 15 days.

Four parameters were measured each time a termite was tested:the time (seconds) for the termite to run the length of the 30 cm trail,the ability (yes or no) of termites to cross a gap in the trail, the time(seconds) for termites to cross the gap, and the number of timestermiteswandered 0.5 cm off the center ink trail and touched either ofthe outside deviation lines (Suppl. Fig. 1). The latter deviation datawere converted to relative proportions for reporting (i.e., proportionof deviations). Data collection began after a termite was placed on thesheet of paper and moved out from the area under the cup onto theink trail. No termite was tested more than once. Eight termites wererandomly sampled per replicate, and three replicates were tested foreach of the three treatments (deviate injected, water injected, andnon-injected). Behavioral bioassays were conducted 1, 2, 3, and 4 daysafter RNAi injections.

2.9.3. Distance feeding bioassayIn the distance feeding bioassay (Suppl. Fig. 2), 150 termites from a

single colony each received deviate siRNA injections, injections ofautoclavedwater, or were chilled but not injected. Termites from eachtreatment were divided into five 100×20 cm Petri dishes (for a totalof 30 termites per treatment group). This dish contained no foodsource but had a moistened paper towel disk, which termites canderive nutrition from, attached to the lid (inaccessible to termites).This dish was connected to a second dish by 1 m of clear 3.2 mm(outer diameter) ×6.4 mm (inner diameter) Tygon™ tubing. Disks ofdouble-ply paper towel were placed in the bottoms of the seconddishes (accessible to termites) and moistened with 100 μl of distilledwater. Dishes from each treatment were randomly placed insideseparate plastic boxes lined with moistened papers towels, whichwere housed in an environmental chamber at 27 °C for 15 days. Paperin the second dish was replaced on days 5, 10, and 15, and pre- andpost- weights recorded.

Termites were restricted to the first Petri dish for the first 48 hoursafter injection to ensure that they did not begin foraging before theRNAi effect reached maximal levels (see Section 3.5). This wasaccomplished by clamping the tubing just outside of the first dish.After 48 hours clamps were removed and termites allowed to movefreely between both dishes.

2.10. Statistical analyses

Statistical analyses were conducted using JMP 6.0 (SAS Institute;Cary, NC, USA). Analysis consisted of non-parametric K-W ANOVAtests (Zhou et al., 2006a), replicated as described in each methods andresults section. All bioassay treatmentswere independently replicatedon three colonies, except in the distance-foraging assay, in which onlya single colony was used.

3. Results

3.1. cDNA sequences and amino acid translations

The full-length deviate cDNA and translated amino acid sequencesare shown in Fig. 1. The cDNA sequence contains 43 nucleotides of 5′untranslated region (UTR) ahead of the ATG start codon, an 851 bpopen reading frame (ORF), and 148 nucleotides of 3’ UTR between the“tag” termination codon and the poly-A tail. The 3’ UTR contains anapparent polyadenylation signal “AATAA” and a putative terminal

16 M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

poly-A tail. The translated deviate protein sequence contains 283amino acids and a 16 amino acid signal peptide MLQFVLAALLATSTLA,which is suggestive that the mature protein is soluble and secreted.Two cysteine residues putatively involved in creating a disulfidebridge involved in ligand binding are present near the N-terminus.Four predicted N-glycosylation sites are present in the mature proteinat N52, N82, N137 and N179.

A ClustalV multiple alignment of insect takeout and JHBP aminoacid sequences is shown in Fig. 2. The alignment shows the closestdatabase matches obtained from blastx searches of the Genbank nrdatabase using the deviate cDNA sequence as a query. Sequencesincluded in the alignment consist of R. flavipes deviate and 13 otherspecies from Isoptera, Diptera, and Lepidoptera. The alignment showsonly low degrees of amino acid identity between deviate and the other12 sequences; i.e., the highest identities are only 9.1–14.1%. Only thetwo cysteine residues involved in creating a ligand binding pocket areconserved across all 13 family members. A cladogram (Fig. 3)generated from the Fig. 2 alignment (parsimony analysis, 293 total

Fig. 2. Translated amino acid alignments (ClustalV) of R. flavipes deviate protein and other mamino acids identical to the R. flavipes sequence. Black triangles (▼) indicate conserved cybinding. Residues shaded black in the three Zootermopsis proteins indicate analogous cysteinthe M. sexta JHBP. Sequence names and accession numbers are as follows: R. flavipes (R.f. DAAF79960), Aedes aegypti (TAKEOUT; AAL60239), Anopheles gambiae (TOL-1; AAO39755) and (M. sexta (JHBP; AAF45309), (JP29; U52070) and (moling; AY231292), Zootermopsis nevadensis

characters, 6 characters constant, 67 characters parsimony un-informative, 220 parsimony informative characters) shows deviateas a distinct outgroup that clusters nearest to JHBPs of M. sexta and B.mori. These results indicate that R. flavipes deviate is a highly novelmember of the takeout/JHBP protein family.

3.2. Constitutive expression among tissues and caste phenotypes

Constitutive expression studies were conducted using conven-tional PCR and agarose electrophoresis (Fig. 4). In general, ubiquitousdeviate mRNA expression was identified among all tissues and bodyregions (including antennae) of workers, soldiers, and nymphs. Onlyworker gels are shown for brevity.

3.3. Deviate expression responses to light–dark conditions and feeding

In light–dark experiments, qRT-PCR data from continuous lightand dark treatments were normalized to 12:12 hr light:dark

embers from the takeout / JHBP family of ligand binding protein. Gray shading depictssteine residues that form disulfide bonds and secondary structure involved in ligande residues. Underlining indicates amino acids proven to be involved in ligand binding inEVIATE; HQ003932), Bombyx mori (JHBP_OBP; BAH79159), D. melanogaster (TAKEOUT;TOL-2; AAO39756), N. takasagoensis (NTSP1; AB195158), Bombyx mori (JHBP; AAF19267),(Zoot. OBP1; AAN15921), (Zoot. OBP2; AAN15922), and (Zoot. OBP3; AAN15923).

M.sextaJP29 (U52070)

No. of nucleotide substitutions

0591.4

100200300400500

M.sextamoling (AY231292)Zoot.(2) AAN15922

Zoot.(3) AAN15923NtSP1 (AB195158)

A.aeg. TAKEOUT (AAL60239)A.gam. TOL-2 (AAO39756)D.m. TAKEOUT (AAF79960)

A.gam. TOL-1 (AAO39755)B. mori JHBP_OBP (BAH79159)B.moriJHBP (AAF19267)

M.sextaJHBP (AAF45309)R.f. DEVIATE

100

100

92

9575

Fig. 3. Cladogram depicting relationships between the R. flavipes deviate protein and other members from the takeout / JHBP family of ligand binding proteins. The cladogram is basedon the alignment presented in Fig. 2. The cladogram is based on 100 bootstrap replicates with bootstrap values greater than 75% shown. These results underscore the uniqueness ofthe deviate gene and translated amino acid sequences. See Fig. 2 for sequence information.

17M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

expression data (Fig. 5A). A ~2.5-fold increase (K-W ANOVA;p=0.0794) in deviate mRNA expression occurred in continuouslight treatments relative to continuous darkness. Deviate expressionwas also not different between continuous light and 12:12 hr light:dark conditions (K-W ANOVA; pN0.1).

For feeding/starvation experiments conducted in complete dark-ness, qRT-PCR data from days 5, 10, and 15 were normalized to day 0(Fig. 5B). Significant differences in deviate mRNA expression werefound between day 0 and day 5 fed (K-W ANOVA; p=0.0116), andday 5 fed and starved (K-W ANOVA; p=0.0006). At day 10, both fedand starved treatments continued to show significant differencesfrom day 0 (K-W ANOVA; p=0.0044, p=0.0035, respectively), butwere not different from each other. At day 15, starved and fedtreatments had identical expression, and starved treatments weredifferent from day 0 (K-W ANOVA; p=0.0059), but fed treatmentswere not.

3.4. Temperature and JH impacts on deviate expression

For these experiments, qRT-PCR data from 17 °C and 22 °Ctreatments were normalized to 27 °C treatment expression data(Fig. 6). Only marginal differences (K-W ANOVA; pb0.1) were foundbetween the two temperatures under baseline conditions, but thetemperature effect was significant in JH treatments (K-W ANOVA;pb0.02). Likewise, there were no differences in deviate mRNAexpression at 17 °C (K-W ANOVA; pN0.1), but at 22 °C deviateexpression decreased to ~0.14× in the presence of ectopic JH (K-WANOVA; pb0.01). In agreement with previous findings for other JH-responsive genes and proteins (Scharf et al., 2007) these resultsindicate that temperature and JH interact significantly to impactdeviate expression.

Fig. 4. Deviate mRNA is ubiquitous among tissues and caste phenotypes. The image is of(reference gene) mRNA distribution among R. flavipes worker tissues. Bands are PCR produobtained for nymph and soldier phenotypes (not shown for brevity). The entire experimenindicate: 1, whole body; 2, antennae; 3, head; 4, thorax; 5, abdomen.

3.5. Methods development and attenuation of deviate mRNA levels afterdeviate siRNA injection

Before proceeding with behavioral bioassays, it was first necessaryto optimize RNAi conditions. The optimal concentration of deviatesiRNA injected into the termites was identified to ensure that genesilencing was effective, quantifiable, and produced no unintended ordetrimental side-effects. For this purpose termites were injected witheither deviate siRNA (36 nL containing 0.5 ng siRNA) or water, or leftnon-injected. Consistent with earlier work (Zhou et al., 2006a), nosignificant mortality was observed in association with the injectionprocess.

After initial test injections of siRNA in workers, qRT-PCR resultsshowed an expected reduction in deviate gene expression whencompared to pre-injection expression levels and water-injectedcontrols (Fig. 7A). Expression levels were normalized to two controlgenes (β-actin,NADH-dh), whichwere unaffected by deviate silencing.One day after RNAi injections, deviate mRNA expression dropped to53.6% of pre-injection expression levels (K-W ANOVA; pb0.05);however, at 2 and 3 days after siRNA injections, deviate expressionreturned to levels identical to water-injected controls, which did notshow significant changes in deviate or control gene expression overthe 3-day test period.

3.6. Impacts of deviate RNAi on JH-dependent worker-to-soldier castedifferentiation

Worker termites were subjected to a model bioassay in whichectopic exposure to JH was used to force caste differentiation indeviate siRNA injected and control individuals. JH III inducedpresoldier development in all treatments (deviate siRNA injected,water injected, and non-injected), as expected, starting at assay

a representative agarose electrophoresis gel showing relative deviate and NADH-dhcts after 35 cycles of amplification from cDNA templates. Highly similar results were

t was independently replicated 3× for workers, soldiers, and nymphs. Numbered lanes

fedfedfedfedstarved

b

a

bc

cbc

c

0.000.200.400.600.801.001.201.401.601.80

Rel

ativ

e ex

pre

ssio

n le

vel

p < 0.05

B

p = 0.0794

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

ContinuousLight

Rel

ativ

e ex

pre

ssio

n le

vel

A

ContinuousDark

12:12 hrLight : dark

Day 0 Day 5 Day 10 Day 15

Fig. 5. Deviate mRNA expression is responsive to light–dark conditions and feeding.(A) Deviate expression is on average ~2.5-fold higher when termites are held under 12and 24 hr photophase conditions relative to continuous darkness. (B) Across 15-dayassays, deviate expression is significantly upregulated on day 5 in association withfeeding (pairwise Kruskal–Wallis ANOVA; pb0.05), but is identical in fed and starvedtreatments on days 10 and 15 (pN0.05). Results in (A) and (B), respectively, arenormalized to the 12:12 light:dark treatment and assay day 0. Bars and data pointsindicate treatment means from three independent replicates; error bars denotestandard error.

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0Rel

ativ

e ex

pre

ssio

n le

vel

Day 0 Day 1 Day 2 Day 3

Day 1 Day 2 Day 3 Day 4

p < 0 .05

water injected

water injected

deviate siRNA injected

deviate siRNA injected0.90.80.70.60.50.40.30.20.1

0Pro

po

rtio

n o

f d

evia

tio

ns

not injectedp = 0.0116

p = 0.0093

p = 0.0209

p = 0.0051

A

B

Fig. 7. RNAi knockdown of deviate mRNA expression significantly impacts trail-following accuracy. (A) Time course of deviate mRNA attenuation in R. flavipes workersafter deviate siRNA injection. Results were determined by qRT-PCR (2ΔΔCT method)with normalization within each day to two reference genes (β-actin, NADH-dh), andthen water-injected controls. (B) Trail following accuracy over time measured as theaverage proportion of deviations made by worker termites in RNAi treatments andcontrols. Treatments consisted of 36 nL nuclease-free water containing 500 pg siRNA(gray), 36 nL nuclease-free water alone (white), or no injection (black). See text forbioassay details. Results show significant impacts on mRNA expression at day 1 andsignificant impacts on trail following accuracy for days 1 and 2 post-injection. Bars anddata points indicate treatment means from three independent replicates; error barsdenote standard error.

18 M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

day10. Presoldier differentiation steadily increased over the course ofthe assay (Suppl. Fig. 3); however, no significant differences wereobserved among the three treatments. Also, as expected, no presoldierdevelopment occurred in non-injected control treatments in theabsence of ectopic JH. These results do not suggest roles for deviate ineither mediation or attenuation of JH-dependent caste differentiation.

3.7. Impacts of deviate RNAi on trail following and foraging

Next, trail following bioassays were conducted in siRNA injectedtreatments and controls, using worker termites and 2-phenoxyetha-nol as an artificial trail pheromone. Termites completed trails in ~15–25 s, and there were no differences among RNAi treatments andcontrols in the time it took to complete trails (Suppl. Fig. 4). However,significant differences between deviate siRNA and control treatmentsdid occur in the number of deviations termites made from the centertrail line (Fig. 7B). Specifically, 1 day after injections, deviate siRNA

7

6

5

4

3

2

1

0Rel

ativ

e ex

pre

ssio

n le

vel

no JH

JH III

p < 0.01

p < 0.02

p < 0.1

Fig. 6. Deviate mRNA expression is influenced by JH and temperature. Worker termiteswere held for 15 days at either 17 or 22 °C on filter papers treated with either 150 μg JHIII (gray bars) or solvent carrier (white bars). Results show that ectopic JH at a dosesufficient to induce caste differentiation significantly suppresses deviate expression at22 °C, but not 17 °C. Results are normalized to 27 °C treatments. Bars indicate treatmentmeans from three independent replicates; error bars denote standard error.

injected termites made N4-fold more deviations than water (K-WANOVA; p=0.0051) and non-injected (K-W ANOVA; p=0.0116)controls. On day 2, deviate siRNA injected termites once again madesignificantly more deviations than both water (K-W ANOVA;p=0.0209) and non-injected (K-W ANOVA; p=0.0093) controls.There were no significant differences between controls and treat-ments on day 3 and beyond.

No significant differences were observed for other trail-followingparameters between termites injected with deviate siRNA andcontrols for the ability to cross a 2 cm gap in the trail (Suppl. Fig. 5)and the time to cross a 2 cm gap in the trail (Suppl. Fig. 6). There wasalso no correlation between the number of deviations a termite madefrom the center trail line and the time it took termites to run thelength of the trail. In distance feeding bioassays (Suppl. Fig. 7), therewere no significant differences among treatments within any of thedays, and no differences between days 1 and 5, and days 10 and 15. Indistance feeding assays, deviate siRNA-injected treatments exhibitedidentical feeding levels to controls (K-W ANOVA; pN0.05), indicatingthat deviate gene silencing did not affect general vigor in terms oflong-term mobility and foraging.

4. Discussion

4.1. Overview

Findings of this study indicate that the R. flavipes deviate gene andtranslated amino acid sequences are highly unique relative to all otherknown takeout family members. The takeout gene family is large anddiverse, with members that participate in odorant and JH binding, aswell as circadian output, circadian signaling, and feeding behavior.

19M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

Thus, at the onset of this study we set out to investigate each of theseaspects. We found that R. flavipes deviate is responsive to light–darkconditions, feeding, temperature and JH, and plays a significant role intrail following accuracy, but not JH-dependent caste differentiation. Inthe sections that follow, each of these topics is discussed relative totermite biology and what has been previously reported for otherinsect groups.

4.2. The deviate gene and protein of R. flavipes

In database searches, the deviate gene and its predicted translationproduct were most similar to takeout-homologous genes, OBPs, JHBPsfrom multiple insect species. BLAST searches using the translateddeviate amino acid sequence identified proteins of known orinvestigated function including takeout and takeout-like genes inspecies including Aedes aegypti (Bohbot and Vogt, 2005), Phormiaregina (Fujikawa et al., 2006), Apis mellifera (Hagai et al., 2007), andDrosophila melanogaster (So et al., 2000; Sarov-Blat et al., 2000)(Fig. 2). Other deviate-homologous genes with characterized func-tions included a solider-specific protein from Nasutitermes takasa-goensis (Hojo et al., 2005), JHBP in M. sexta (Du et al., 2003), acircadian-clock controlled gene precursor in Drosophila yakuba(Thackeray and Kyriacou, 1990), and several antennal OBPs occurringacross multiple orders of insects, including termites (Ishida et al.,2002a). R. flavipes deviate shares little homology with other knowntermite takeout family members as reported previously (Ishida et al.,2002a; Hojo et al., 2005). The fact that this is the first deviate homologidentified from a hemimetabolous insect may also explain itsuniqueness.

The takeout gene and related OBP and JHBPs are typically 230-250amino acids long and contain two cysteine residues located near theN-terminal end of the peptide sequence (Saito et al., 2006; Robertsonet al., 1999; Wojtasek and Prestwich, 1995). These cysteines form adisulfide bridge that creates secondary structure involved in ligandbinding (Wojtasek and Prestwich, 1995). Another feature common tomany insect OBPs is the presence of six cysteine residues locatedthroughout the protein (Ishida et al., 2002b; Robertson et al., 1999;Bohbot and Vogt, 2005). In addition to cysteine residues, twoconserved ligand binding fragments were identified in M. sexta JHBP(Touhara and Prestwich, 1992; Saito et al., 2006). The first bindingmotif is located closer to the N-terminus of the peptide sequence andthe second includes the C-terminus. Some takeout and OBPs alsocontain signal peptide sequences (Saito et al., 2006; Sarov-Blat et al.,2000; Hojo et al., 2005).

R. flavipes deviate is shorter in length than most takeout familymembers, possesses a secretory signal peptide, and possesses only asingle cysteine residue beyond its two highly conserved N-terminalcysteine residues. Interestingly, in addition to having expectedantennal expression, deviate mRNA expression was also observedthroughout the bodies of workers, soldiers and nymphs. Deviate ESTswere sequenced from R. flavipes alate (Steller et al., 2010) and gutcDNA libraries (Tartar et al., 2009; Tarver et al., 2010) furthersupporting that it has roles in multiple castes and tissues. Thesefindings suggest that the R. flavipes deviate gene and protein are highlyunique relative to all other known insect takeout homologues, andperhaps perform unique social functions in termites.

4.3. Links to circadian physiology, feeding, temperature, and JH

4.3.1. Circadian physiologyIn light–dark studies, marginally significant differences in deviate

expression were found between light and dark treatments (Fig. 5A).The Drosophila takeout gene has been described as a circadian clock-regulated output gene (McDonald and Rosbash, 2001; Benito et al.,2010). Specifically, takeout gene expression fluctuates with dailycircadian rhythm patterns and is undetectable in circadian mutants,

which suggests a role in circadian-associated functions (Sarov-Blatet al., 2000). In addition, recent findings suggest that takeout alsoplays a role in lifespan regulation (Gáliková and Flatt, 2010; Baueret al., 2010). Consistent with takeout links to circadian biology, thecurrent study suggests that, although subterranean termite workersare eyeless, they have the ability to sense changes in photoperiodand alter deviate gene expression accordingly. The alteration ofdeviate gene expression under different light–dark regimes suggeststhat worker termites possess photosensory capabilities, and thatdeviate may play a role in photoperiod-regulated behavior and/orphysiology.

4.3.2. FeedingIn feeding studies, although deviate expression changed signifi-

cantly through time in both treatments and controls, the mostsubstantial differences in deviate expression occurred in day 5 fed vs.starved treatments (fed=significantly higher expression; Fig. 5B). InDrosophila, alternatively, takeout expression increased during periodsof starvation (Sarov-Blat et al., 2000). While these results indicatechanges in gene expression over timewhenworker termites are facedwith starvation conditions, they may also be associated with removalfrom colony conditions. In this respect, it is noteworthy that theremay be a “colony-release”-associated reduction in deviate expressionresulting from our experimental design, which over-shadows changesin gene expression due to starvation (Zhou et al., 2006a, 2007). It wasalso previously suggested that Drosophila takeout plays a role inforaging behavior by directing flies to make a choice betweenreproduction and foraging (Sarov-Blat et al., 2000). This is seeminglycontrasted in worker termites, which have no reproductive capabil-ities. Thus, deviate and takeout apparently play very different roles indirecting feeding behavior of termites and Drosophila, respectively.

4.3.3. JH–temperature interactionsJH plays complex caste-regulatory roles in termites (Zhou et al.,

2006a, 2007; Scharf et al., 2007; Tarver et al., 2010). Because thedeviate gene shows some structural similarities to JH binding proteins,we hypothesized that, similar to hexamerin proteins (Zhou et al.,2007; Scharf et al., 2007), deviate may participate in caste regulationvia modulation or mediation of JH signaling. In studies involvingectopic exposure to JH concentrations capable of inducing castedifferentiation at different temperatures, both factors were found tointeract significantly to influence deviatemRNA expression (Fig. 6). Ina companion study, significant differential expression of deviatemRNA (called ‘TO-F’) was noted among workers exposed to JH, livesoldier, and soldier head extract treatments (Tarver et al., 2010).These findings suggested that deviate expression levels are influencedby endogenous JH and colony primer pheromone titers, which in turnare influenced by a number of extrinsic factors (Zhou et al., 2007;Scharf et al., 2007; Tarver et al., 2010). However, in deviate RNAistudies conducted here (Suppl. Fig. 1), no significant impacts on JH-dependent presoldier differentiation were observable, suggesting thatdeviate does not directly interact with JH to influence castedifferentiation.

4.4. Evidence linking the deviate gene to trail-following behavior

Trail following bioassays were used to investigate changes intermite behavior following deviate silencing using 2-phenoxyethanolas an artificial trail pheromone (Chen et al., 1998). In addition to trailfollowing bioassays, long-term distance feeding bioassays were alsoused to assess whether deviate silencing impacted general vigor andforaging in the presence of natural trail pheromone (ZZE-dodeca-trienol; Matsumura et al., 1968). Because odorant and ligand bindingproteins are associated with the reception and transport of chemicalsignals within the insect body, and insects with mutations in theseproteins have been shown to exhibit aberrations in movement

20 M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

(Sarov-Blat et al., 2000; Meunier et al., 2007), we examined whethersimilar responses occur in deviate silenced termites. Our resultsindicate that trail following accuracy is associated with deviateexpression, but not any other parameters investigated (e.g., speed,gap crossing). These findings provide evidence that deviate mayencode an OBP involved in the perception of trail pheromone.However, the uniqueness of the translated deviate protein sequenceand its expression outside the antennae suggests it may have been co-opted to perform functions related to the subterranean eusociallifestyle of R. flavipes.

In the distance bioassay, deviate silenced termites did not displayimpairment in their ability tomove toward or feed upon a food source.However, this observation must be considered in the context of theexperimental design, which did not test the ability of termites tomakebehavioral foraging choices, i.e., termites were not subjected toconditions which required them to make foraging or navigationalchoices. Nonetheless, because high levels of survival occurred andequal food consumption was observed between treatments andcontrols, the distance bioassay provides good evidence that termitevigor is not impacted by deviate silencing.

4.5. Summary and conclusions

This research tested four main hypotheses related to theinvolvement of the R. flavipes deviate gene in trail following, foraging,food consumption, and caste differentiation. With the exception ofthe trail following hypothesis, all other hypotheses were rejected.Worker termites injected with deviate siRNA did not exhibitdifferences from controls in total food consumption or in the timingof food consumption. In distance-foraging experiments termites fromall treatments consumed similar amounts of paper (cellulose).Finally, although we observed deviate expression to be responsiveto JH, no impacts on JH-dependent caste differentiation wereobservable after deviate RNAi. Thus, evidence obtained in this studyis most consistent with the hypothesis that the deviate gene encodesa ligand-binding protein involved in odorant perception and trailfollowing accuracy.

In conclusion, functional aspects of the R. flavipes deviate genewere investigated here and RNAi techniques were used to obtainfindings that link deviate and termite behavior. We have concludedthat deviate encodes a ligand binding protein that plays a role in theperception of termite trail pheromone, or potentially, in theregulation of circadian or locomotory behavior. These findingsprovide only the second example linking the expression of a specificgene with eusocial behavior in a termite. Finally, more generally, thisstudy also illustrates how the RNAi pathway can be utilized toexamine the roles of specific genes in social insects in the phenotypicexpression of social behavior.

Acknowledgements

We thank Jonathan Neal, Brian Schneider, and Matthew Tarver forhelpful discussions, feedback and advice. This research was supportedby the Center for Urban Pest Management at Purdue University and byCSREES-USDA-NRI grant No. 2007-35607-17777 to MES.

Appendix A. Supplementary data

Supplementary data to this article can be found online atdoi:10.1016/j.gene.2010.11.012.

References

Bauer, J., Antoch, M., Chang, C., Schorl, C., Kolli, S., Neretti, N., Helfand, S.L., 2010.Comparative transcriptional profiling identifies takeout as a gene that regulates lifespan. Aging 2, 298–310.

Benito, J., Hoxha, V., Lama, C., Lazareva, A.A., Ferveur, J.F., Hardin, P.E., Dauwalder, B.,2010. The circadian output gene takeout is regulated by Pdp1 epsilon. Proc. NatlAcad. Sci. USA 107, 2544–2549.

Bohbot, J., Vogt, R.G., 2005. Antennal expressed genes of the yellow fever mosquito(Aedes aegypti L.); characterization of odorant-binding protein 10 and takeout.Insect Biochem. Mol. Biol. 35, 961–979.

Chen, J., Henderson, G., Laine, R.A., 1998. Isolation and identification of 2-phenox-yethanol from ballpoint pen ink as a trail-following substance of Coptotermesformosanus Shiraki and Reticulitermes sp. J. Entomol. Sci. 33, 97–105.

Du, J., Hiruma, K., Riddiford, L.M., 2003. A novel gene in the takeout gene family isregulated by hormones and nutrients inManduca larval epidermis. Insect Biochem.Mol. Biol. 33, 803–814.

Fire, A., Xu, S., Montgomery, M.K., Kostas, S.A., Driver, S.E., Mello, C.C., 1998. Potent andspecific genetic interference by double-stranded RNA in Caenorhabditis elegans.Nature 391, 806–811.

Fujikawa, K., Seno, K., Ozaki, M., 2006. A novel takeout-like protein expressed in thetaste and olfactory organs of the blowfly, Phormia regina. FEBS J. 273, 4311–4321.

Gáliková, M., Flatt, T., 2010. Dietary restriction and other lifespan extendingpathways converge at the activation of the downstream effector takeout. Aging2, 387–389.

Gura, T., 2000. A silence that speaks volumes. Nature 404, 804–808.Hagai, T., Cohen, M., Bloch, G., 2007. Genes encoding putative takeout/juvenile hormone

binding proteins in the honeybee (Apis mellifera) and modulation by age andjuvenile hormone of the takeout-like gene GB19811. Insect Biochem. Mol. Biol. 37,689–701.

Hannon, G.J., 2002. RNA interference. Nature 418, 244–251.Hiruma, K., Riddiford, L.M., 2010. Developmental expression of mRNAs for epidermal

and fat body proteins and hormonally regulated transcription factors in the tobaccohornworm, Manduca sexta. J. Insect Physiol. 56, 1390–1395.

Hojo, M., Morioka, M., Matsumoto, T., Miura, T., 2005. Identification of soldier caste-specific protein in the frontal gland of nasute termite Nasutitermes takasagoensis.Insect Biochem. Mol. Biol. 35, 347–354.

Ishida, Y., Chiang, V.P., Haverty, M.I., Leal, W.S., 2002a. Odorant binding proteins from aprimitive termite. J. Chem. Ecol. 28, 1887–1893.

Ishida, Y., Chiang, V., Leal, W.S., 2002b. Protein that makes sense in the Argentine ant.Naturwissenschaften 89, 505–507.

Korb, J., Weil, T., Hoffman, K., Foster, K.R., Rehli, M., 2009. A gene necessary forreproductive suppression in termites. Science 324, 758.

Livak, K., Schmittgen, T., 2001. Analysis of relative gene expression data using real-timequantitative PCR and the 2−ΔΔCT method. Methods 25, 402–408.

Matsumura, F., Coppel, H.C., Tai, A., 1968. Isolation and identification of termite trail-following pheromone. Nature 219, 963–964.

McDonald, M., Rosbash, M., 2001. Microarray analysis and organization of circadiangene expression in Drosophila. Cell 107, 567–578.

Meunier, N., Belgacem, Y.H., Martin, J.R., 2007. Regulation of feeding behaviour andlocomotor activity by takeout in Drosophila. J. Exp. Biol. 210, 1424–1434.

Moore, B.P., 1966. Isolation of a scent-trail pheromone of an Australian termite. Nature211, 746–747.

Nutting, W.L., 1990. Insecta: Isoptera. In: Dindal, D.L. (Ed.), Soil biology guide. Wileyand Sons, New York.

Pfaffl, M., Tichopad, A., Prgomet, C., Neuvians, T., 2004. Determination of stablehousekeeping genes, differently regulated target genes and sample integrity:BestKeeper—Excel-based tool using pair-wise correlations. Biotechnol. Lett. 26,509–515.

Reinhard, J., Kaib, M., 1995. Interaction of pheromones during food exploitation by thetermite Schedorhinotermes lamanianus. Physiol. Entomol. 20, 266–272.

Robertson, H.M., Martos, R., Sears, C.R., Todres, E.Z., Walden, K.K., Nardi, J.B., 1999.Diversity of odourant binding proteins revealed by an expressed sequence tagproject on male Manduca sexta moth antennae. Insect Mol. Biol. 8, 501–518.

Robinson, G.E., Grozinger, C.M., Whitfield, C.W., 2005. Sociogenomics: social life inmolecular terms. Nat. Rev. Genet. 6, 257–270.

Saito, K., Su, Z.H., Emi, A., Mita, K., Takeda, M., Fujiwara, Y., 2006. Cloning and expressionanalysis of takeout/JHBP family genes of silkworm, Bombyx mori. Insect Mol. Biol.15, 245–251.

Sarov-Blat, L., So, W.V., Liu, L., Rosbash, M., 2000. The Drosophila takeout gene is a novelmolecular link between circadian rhythms and feeding behavior. Cell 101,647–656.

Scharf, M.E., Buckspan, C.E., Grzymala, T.L., Zhou, X., 2007. Regulation of polypheniccaste differentiation in the termite Reticulitermes flavipes by interaction of intrinsicand extrinsic factors. J. Exp. Biol. 210, 4390–4398.

Scharf, M.E., Zhou, X., Schwinghammer, M.A., 2008. Application of RNA interfer-ence in functional genomics studies of a social insect. Meth. Mol. Biol. 442,205–229.

Shinoda, T., Hiruma, K., Charles, J.P., Riddiford, L.M., 1997. Hormonal regulation of theexpression of JP29 during a larval molt and metamorphosis in the tobaccohornworm, Manduca sexta. Arch. Insect Biochem. Physiol. 34, 409–428.

Smith, C.R., Toth, A.L., Suarez, A.V., Robinson, G.E., 2008. Genetic and genomic analysesof the division of labour in insect societies. Nat. Rev. Genet. 9, 735–748.

So, W.V., Sarov-Blat, L., Kotarski, C.K., McDonald, M.J., Allada, R., Rosbash, M., 2000.takeout, a novel Drosophila gene under circadian clock transcriptional regulation.Mol. Cell. Biol. 20, 6935–6944.

Steller, M.M., Kambhampati, S., Caragea, D., 2010. Comparative analysis of expressedsequence tags from three castes and two life stages of the termite Reticulitermesflavipes. BMC Genomics 11, 463.

Stuart, A.M., 1967. Alarm, defense, and construction behavior relationships in termites(Isoptera). Science 156, 1123–1125.

21M.A. Schwinghammer et al. / Gene 474 (2011) 12–21

Tartar, A., Wheeler, M.M., Zhou, X., Coy, M., Boucias, D.G., Scharf, M.E., 2009. Parallelmeta-transcriptome analysis of host and symbiont gene expression in the gut of thetermite Reticulitermes flavipes. Biotechnol. Biofuels 2, 25.

Tarver, M.R., Zhou, X., Scharf, M.E., 2010. Socio-environmental and endocrine influenceson developmental and caste-regulatory gene expression in the eusocial termiteReticulitermes flavipes. BMC Mol. Biol. 11, 28.

Thackeray, J.R., Kyriacou, C.P., 1990. Molecular evolution in the Drosophila yakubaperiod locus. J. Mol. Evol. 31, 389–401.

Touhara, K., Prestwich, G.D., 1992. Binding site mapping of a photoaffinity-labeled juvenile hormone binding protein. Biochem. Biophys. Res. Commun. 182,466–473.

Wilson, E.O., 1971. The Insect Societies. Belknap, Harvard Univ. Press, Cambridge.Wojtasek, H., Prestwich, G.D., 1995. Key disulfide bonds in an insect hormone binding

protein: cDNA cloning of a juvenile hormone binding protein of Heliothis virescensand ligand binding by native and mutant forms. Biochemistry 34, 5234–5241.

Wong, R., Piper, M.D., Wertheim, B., Partridge, L., 2009. Quantification of food intake inDrosophila. PLoS ONE 4 (6), e6063.

Wu-Scharf, D., Scharf, M.E., Pittendrigh, B.R., Bennett, G.W., 2003. Expressed sequencetags from a polyphenic Reticulitermes flavipes cDNA library. Sociobiology 41,479–490.

Zhou, X., Oi, F.M., Scharf, M.E., 2006a. Social exploitation of hexamerin: RNAi reveals amajor caste-regulatory factor in termites. Proc. Natl Acad. Sci. USA 103, 4499–4504.

Zhou, X., Tarver, M.R., Bennett, G.W., Oi, F.M., Scharf, M.E., 2006b. Two hexamerin genesfrom the termite Reticulitermes flavipes: sequence, expression, and proposedfunctions in caste regulation. Gene 376, 47–58.

Zhou, X., Tarver, M.R., Scharf, M.E., 2007. Hexamerin-based regulation of juvenilehormone-dependent gene expression underlies phenotypic plasticity in a socialinsect. Development 134, 601–610.

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

Zhu, H., Casselman, A., Reppert, S.M., 2008. Chasing migration genes: a brain expressedsequence tag resource for summer and migratory monarch butterflies (Danausplexippus). PLoS ONE 3 (1), e1345.