UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Juvenile Hormone esterase is a conserved regulator of starvation-induced behavior Stafford, Jeffrey 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_february_stafford_jeffrey.pdf [ 1.32MB ]
Metadata
JSON: 24-1.0223111.json
JSON-LD: 24-1.0223111-ld.json
RDF/XML (Pretty): 24-1.0223111-rdf.xml
RDF/JSON: 24-1.0223111-rdf.json
Turtle: 24-1.0223111-turtle.txt
N-Triples: 24-1.0223111-rdf-ntriples.txt
Original Record: 24-1.0223111-source.json
Full Text
24-1.0223111-fulltext.txt
Citation
24-1.0223111.ris

Full Text

Juvenile Hormone esterase is a conserved regulator ofstarvation-induced behaviorbyJeffrey StaffordB.Sc. Cell Biology and Genetics, The University of British Columbia, 2013A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFMaster of ScienceinTHE FACULTY OF GRADUATE AND POSTDOCTORALSTUDIES(Zoology)The University of British Columbia(Vancouver)December 2015c© Jeffrey Stafford, 2015AbstractAlthough feeding behavior is a matter of life and death for animals, the geneticfactors that control it remain poorly understood. We have identified a novel reg-ulator of hunger-induced behavior through comparison of transcriptomic changesin the fruit fly Drosophila melanogaster and the yellow fever mosquito Aedes ae-gypti. Head mRNA from each insect was sequenced at a roughly equivalent level ofstarvation. Using data gleaned from the protein orthology database OrthoDB, welooked for gene pairs in which both A. aegypti and D. melanogaster orthologs weresignificantly regulated by starvation. This identified Juvenile Hormone esterase(Jhe) as a possible modulator of hunger-induced behavior. Pan-neuronal knock-down of Jhe resulted in increased food consumption and caused enhancement ofstarvation-induced sleep suppression in Drosophila. These behavioral phenotypeswere not caused by a developmental or metabolic defects, and were reproduced byfeeding adult Drosophila methoprene, a synthetic Juvenile Hormone analog. Ap-plication of precocene I, an inhibitor of Juvenile Hormone biosynthesis, reversedthe phenotype. Our analysis suggests that Jhe (and Juvenile Hormone by exten-sion) is a novel and biologically relevant regulator of hunger-induced behavior.iiPrefaceThis work was conducted at the University of British Columbia’s Life SciencesCenter by Jeff Stafford and Dr. Michael Gordon. Alex Keene’s (University ofNevada at Reno, Department of Biology) laboratory performed RNA sequencingof the Drosophila melanogaster samples used in this study, providing raw FASTQread files for alignment and bioinformatic analysis. Carl Lowenberger (SimonFraser University, Department of Biological Sciences) generously reared and do-nated the Aedes aegypti mosquitoes used in this study. Sean Jewell (University ofBritish Columbia, Department of Statistics) devised and coded the vast majority ofthe Java statistical simulation used to verify the ortholog results (see Section 2.7).I made a number of modifications to the simulation, adding the ability to use twoα values and a Gradle build script to allow compilation of a runnable “fat JAR”from the simulation source code. Michael Gordon oversaw this research projectand provided funding. All other work and experiments not specifically listed herewere performed by Jeff Stafford.iiiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Gustation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Selection of experimental animals . . . . . . . . . . . . . . . . . 41.2.1 Drosophila melanogaster . . . . . . . . . . . . . . . . . . 61.2.1.1 The GAL4/UAS System . . . . . . . . . . . . . 61.2.1.2 RNA Interference (RNAi) . . . . . . . . . . . . 71.2.2 Aedes aegypti . . . . . . . . . . . . . . . . . . . . . . . . 71.2.3 Tribolium castaneum . . . . . . . . . . . . . . . . . . . . 81.3 Orthology as a means of transcriptomic comparison . . . . . . . . 81.4 Insect feeding behaviors . . . . . . . . . . . . . . . . . . . . . . 101.5 Juvenile Hormone . . . . . . . . . . . . . . . . . . . . . . . . . . 111.5.1 Synthesis of Juvenile Hormone derivatives . . . . . . . . 13iv1.5.1.1 Regulation of Juvenile Hormone synthesis . . . 141.5.2 Juvenile Hormone metabolism in the hemolymph . . . . . 151.5.2.1 Takeout . . . . . . . . . . . . . . . . . . . . . . 151.5.2.2 Juvenile Hormone esterase . . . . . . . . . . . 161.5.2.3 Juvenile Hormone epoxide hydrolase . . . . . . 171.5.3 Juvenile Hormone signalling . . . . . . . . . . . . . . . . 181.5.3.1 Methoprene-tolerant and Germ-cell Expressed . 181.5.3.2 Ultraspiracle . . . . . . . . . . . . . . . . . . . 191.5.4 Known roles and effects of the Juvenile Hormones . . . . 202 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222.1 Insect husbandry . . . . . . . . . . . . . . . . . . . . . . . . . . 222.2 Tissue sugar quantification . . . . . . . . . . . . . . . . . . . . . 222.3 RNA sample preparation . . . . . . . . . . . . . . . . . . . . . . 242.4 cDNA library preparation and sequencing . . . . . . . . . . . . . 242.5 RNA sequencing analysis pipeline . . . . . . . . . . . . . . . . . 252.6 Identification of conserved transcriptomic changes . . . . . . . . 252.7 Validation of ortholog results . . . . . . . . . . . . . . . . . . . . 262.8 Methoprene and Precocene I feeding . . . . . . . . . . . . . . . . 272.9 Measurement of Drosophila food intake . . . . . . . . . . . . . . 282.10 Measurement of Drosophila sleep . . . . . . . . . . . . . . . . . 292.11 Measurement of Drosophila starvation sensitivity . . . . . . . . . 292.12 actmon R package . . . . . . . . . . . . . . . . . . . . . . . . . 292.13 Plotting and statistics . . . . . . . . . . . . . . . . . . . . . . . . 303 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313.1 Characterization of insect hunger . . . . . . . . . . . . . . . . . . 313.1.1 Determination of Drosophila starvation state . . . . . . . 313.1.2 Determination of Aedes starvation state . . . . . . . . . . 333.2 RNA sequencing reveals conserved transcriptomic changes . . . . 333.2.1 Identification of conserved transcriptomic changes . . . . 353.2.2 Validation of ortholog results . . . . . . . . . . . . . . . . 363.2.3 Jhe is regulated significantly by hunger . . . . . . . . . . 38v3.3 Jhe is necessary for proper feeding behavior . . . . . . . . . . . . 403.4 Jhe knockdown increases starvation-induced sleep suppression . . 443.5 Jhe knockdown does not increase starvation sensitivity . . . . . . 453.6 Methoprene feeding results in a Jhe-like phenotype . . . . . . . . 483.7 Precocene I rescues Jhe knockdown . . . . . . . . . . . . . . . . 514 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544.1 Regulation of Jhe is evolutionarily conserved . . . . . . . . . . . 544.2 Jhe affects sleep and feeding in adult Drosophila . . . . . . . . . 564.3 Jhe exerts its effects through Juvenile Hormone metabolism . . . 564.4 Future work and directions . . . . . . . . . . . . . . . . . . . . . 574.5 Final remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60A Supporting Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 69A.1 GCaMP 4D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69A.2 fly tracker . . . . . . . . . . . . . . . . . . . . . . . . . . . 71viList of TablesTable 2.1 Key to Drosophila genotypes used in figures. . . . . . . . . . . 23viiList of FiguresFigure 1.1 Measurement of Tribolium sugar levels . . . . . . . . . . . . 9Figure 1.2 The Juvenile Hormones . . . . . . . . . . . . . . . . . . . . . 12Figure 3.1 Quantification of insect sugar levels . . . . . . . . . . . . . . 32Figure 3.2 Overview of RNA sequencing results . . . . . . . . . . . . . 34Figure 3.3 Orthologous gene-pairs significantly regulated by starvation . 37Figure 3.4 Ortholog simulation results . . . . . . . . . . . . . . . . . . . 39Figure 3.5 JH-related genes are downregulated in multiple forms of nutri-ent deprivation . . . . . . . . . . . . . . . . . . . . . . . . . 41Figure 3.6 Jhe is a modulator of feeding behavior . . . . . . . . . . . . . 42Figure 3.7 Jhe’s phenotype is not a developmental defect . . . . . . . . . 43Figure 3.8 Jhe’s phenotype is not an off-target effect . . . . . . . . . . . 44Figure 3.9 Jhe knockdown alters Drosophila activity . . . . . . . . . . . 46Figure 3.10 Jhe knockdown increases starvation-induced sleep suppression 47Figure 3.11 Jhe knockdown does not affect starvation sensitivity . . . . . 49Figure 3.12 Methoprene alters Drosophila activity in a Jhe-like manner . . 50Figure 3.13 Methoprene increases starvation-induced sleep suppression . . 52Figure 3.14 Precocene I rescues Jhe knockdown . . . . . . . . . . . . . . 53Figure A.1 Sample GCaMP 4D output . . . . . . . . . . . . . . . . . . . 70Figure A.2 Sample fly tracker output . . . . . . . . . . . . . . . . . 72viiiList of Abbreviationsα The probability of any given gene being regulated in an organismANOVA Analysis of VarianceBAM Binary Alignment FormatBDGP Berkeley Drosophila Genome ProjectBRH Best Reciprocal Hitsbp Base pair/sC Degrees CelsiusCDF Cumulative Distribution FunctioncDNA Complementary DNACNS Central nervous systemCO2 Carbon Dioxidecsv Comma-separated valuesDAM Trikinetics’ Drosophila Activity Monitord Day/sdH2O Distilled waterdILP Drosophila Insulin-like PeptideixDNA Deoxyribonucleic acidFA Farnesoic acidFASTQ FASTQ read formatg GravitiesGAL4 The Saccharmyces cerevisiae GAL4 proteingce Germ-cell ExpressedGD VDRC GD Transgenic RNAi LibraryGFP Green Fluorescent ProteinGPS Graduate and Postdoctoral StudiesGTF Gene Transfer Format 2h Hour/sJAR Java archiveJH Juvenile Hormonejhamt Juvenile Hormone acid methyltransferaseJHB3 Juvenile Hormone III bisepoxidehJHBP Hemolymph Juvenile Hormone binding proteinJhe Juvenile Hormone esteraseJheh Juvenile Hormone epoxide hydrolasek Number of significantly regulated orthologous gene pairsKCl Potassium ChlorideKK VDRC KK Transgenic RNAi LibraryMet Methoprene-tolerantxMF Methyl farnesoatemin Minute/smL Milliliter/smm Millimeter/smM MillimolarmRNA Messenger RNANaCl Sodium ChlorideOrthoDB OrthoDB7, a heirachical catalog of animal, fungal, and bacterialorthologsPCR Polymerase Chain ReactionPMF Probability Mass FunctionR The R programmming languageRNA Ribonucleic acidRNA-Seq RNA sequencingRNAi RNA interferenceS4 The R language’s S4 object-oriented programming systemshRNA Small hairpin RNAsNPF Short Neuropeptide FSWIPE The SWIPE protein alignment algorithmTARGET Temporal And Regional Gene Expression Targeting systemTris-HCl Tris (hydroxymethyl) Aminomethane HydrochlorideTo Takeoutxiµg Microgram/sµL Microliter/sUAS Upstream Activating Sequence, targeted by GAL4UBC The University of British Columbiausp UltraspiracleVDRC Vienna Drosophila Resource CenterVL20 Bloomington VALIUM20 RNAi Libraryw/v Percent weight / volumexiiAcknowledgmentsBefore diving into the study head-on, I wanted to take a moment to thank all ofthe people who helped me get where I am today. First and foremost, I’d like tothank my collaborators. Carl Lowenberger raised and supplied the mosquitoesused for this study, making it possible to perform the RNA-Seq experiments forthat organism. Alex’s Keene’s Drosophila RNA-Seq datasets were instrumentalas well, and enabled us to much more accurately quantify several transcriptomicchanges than would otherwise have been possible. This study would not have beenpossible without WestGrid and Compute Canada’s generous donation of computertime for use in RNA-Seq alignment and quantification. Brian de Alwis’s LATEXtemplate made writing this thesis far easier than it otherwise would have been. I’dalso like to take a moment to thank the members of my committee- it was refreshingto hear new viewpoints, and your input was greatly appreciated. Most importantly,I want to thank my supervisor, Mike Gordon, for giving me the chance to start hereand the freedom to take my research in the direction I wanted. Last but not least,I’d like to thank my friends, roommates, and family. Without you to keep me sane,I never would have made it this far.xiiiChapter 1Introduction1.1 GustationThe gustatory (taste) system is an ideal model with which to study larger processesoccurring throughout the brain as a whole. Compared with other sensory modal-ities, gustation is relatively simple. Humans have five tastes: sweet, salty, sour,bitter, and umami. In contrast, most people would be hard pressed to tell you howmany smells there were or describe what might comprise different types of touch.Taste is a discrete stimulus that is easy to quantify. Taste processing in the brainis considered relatively simple as well. The olfactory system uses a combinato-rial code to encode stimuli, with the same areas of the brain activated in differentpatterns to represent different smells (Semmelhack and Wang, 2009). By con-trast, the dominant theory is that gustatory inputs are processed as a labelled line:each different taste is carried to the brain along different nerves and is processedin different regions (Yarmolinsky et al., 2009). Gustation is also thought to be ahardwired system. Whereas responses to olfactory stimuli can be learned by as-sociation, taste evokes stereotyped, reflexive behavior (Yarmolinsky et al., 2009).When an experimentally-induced change causes gustatory information to be pro-cessed differently, gustatory behavior will often change as a result. Because of this,gustatory behavior is an ideal model with which to study the link between genesand behavior.Much is known about how gustatory information is detected at the sensory1level. Despite this, relatively little is known about how this information is pro-cessed in higher brain centers and used to effect changes on feeding-related behav-ior. Although several neurons responsible for directly controlling feeding behaviorhave been identified, these neurons do not directly synapse with the neurons ca-pable of detecting gustatory stimuli (Gordon and Scott, 2009). Very few second-order taste neurons have been identified, and the many of those that have do notintegrate input from internal stimuli like hunger state (Flood et al., 2013; Chu et al.,2014; Kain and Dahanukar, 2015). Although the neurons and mechanisms control-ling complex gustatory behaviors have been discovered (such as taste-independentcalorie sensing), their exact means of effecting changes in behavior remain unclear(Miyamoto et al., 2012; Stafford et al., 2012; Dus et al., 2013). Most, if not all,of these gustatory sensory pathways and behaviors are known to be influenced bystarvation.Starvation is a fundamental physiological state shared by all animals, yet it re-mains poorly understood. Starved animals change behavior dramatically- hungryDrosophila eat more, rapidly develop a preference for high-calorie foods, and sup-press sleep (Stafford et al., 2012; Keene et al., 2010). A large number of geneticand hormonal changes occur during starvation as well. Despite this information,it remains relatively unknown which genes and signaling pathways are necessaryand sufficient to elicit a starvation-like behavioral state in animals. The goal ofthis study was to discover a gene or signaling pathway capable of inducing such abehavioral change. Any novel behavioral regulator identified would prove a tanta-lizing target for further research examining how gustatory behaviors are controlledby the brain.Before this study, it was known that a large number of transcriptional changesoccurred in an animal during starvation (Farhadian et al., 2012; Fujikawa et al.,2009). These transcriptional changes seemed likely to control starvation-relatedbehavior, as the process of transcription acts on a similar timescale to starvationitself. The most straightforward means of measuring these changes would be tosimply run a set of microarrays of RNA extracted from the head of a model or-ganism like Drosophila melanogaster when it is hungry. Examining Drosophila’shead transcriptome is a simple and effective means of studying neuronal changesin expression- with the exception of the ventral nerve cord, it captures the entire2central nervous system (CNS) of an animal, along with important sensory organslike the eyes, antennae, and labella (mouthparts). This approach would capture anytranscriptional changes in that occurred in the brain and sensory tissues during star-vation. However, this method has been tried twice before. Microarrays peformed inthis manner did not identify any previously unstudied behavioral regulators (Farha-dian et al., 2012; Fujikawa et al., 2009). Despite these results, it seemed likely thatthere were still genes of interest that could be discovered using a more advancedapproach.Advancements in technology now allow for much more accurate and sensi-tive quantification of gene expression changes, with a much greater dynamic range(Wang et al., 2009). Capitalizing on this fact, this study utilized high-throughputmRNA sequencing instead of microarrays for this study. For instance, one com-mercial RNA-Seq platform is advertised as detecting the same number of differen-tially expressed genes as a typical microarray with a mere 2 million reads (Illumina,2011). I aimed to sequence to a depth of 35-60 million reads, a much more exactreading of RNA abundance than was possible previously. Additionally, RNA se-quencing is able to perform accurate quantification of mRNA abundance acrossa much larger dynamic range. Microarrays are notoriously poor at quantifyingweakly or very highly expressed transcripts (Wang et al., 2009). Accurate quantifi-cation of weakly expressed genes was of particular concern for this study. Manygenes with a large impact on behavior, including sex specification genes like fruit-less or hormones like the Drosophila Insulin-like Peptide family, are expressed inextremely spatially restricted patterns, with expression limited to only a handful ofcells. RNA sequencing offers an effective, reliable means of accurately quantifyingthese types of transcripts.Screening Drosophila or any other animal for behavioral phenotypes is ex-tremely slow and suffers from highly variable results. Behavioral experiments re-quire numerous genetic and environmental controls, and promising results need tobe repeated several times on several different days before they can be accepted.Even screening less than a hundred RNAi knockdown lines can take months, de-pending on the behavioral assay used. Any means of refining a large list of differen-tially expressed genes down to a handful of promising candidates could potentiallysave an inordinate amount of research time when screening for potential behav-3ioral regulators. As a result, I decided to employ a new means of improving ourour chances of finding a novel regulator of hunger-induced behavior. I decided thatstudying hunger-induced changes in two separate, evolutionarily divergent insectswould offer a good chance of highlighting potentially interesting genes. The logicbehind this decision was that any transcriptomic changes conserved between theorganisms would be the most important to an animal’s survival, as regulation ofthese genes had been maintained across millions of years of evolution. Of course,this approach raises a number of important questions. What experimental animalswould prove the best candidates for study? What would constitute a “conservedtranscriptomic change?” How would it be established that the animals were in asimilar physiological state before sequencing?1.2 Selection of experimental animalsAs I previously indicated in Section 1.1, insects were an ideal target for this studyas it is possible to sequence the entire head in a single sample, encapsulating mostof the CNS and all associated sensory tissues. Other advantages offered by in-sects over mammalian and other genetic systems include extremely fast genera-tion times, a large number of available genetic tools and behavioral assays, andpre-existing expertise within our research group. In addition, the vast majority ofinsect genes have human homologs, and even genes without mammalian equiva-lents often have important commercial applications (possibly having implicationsfor pesticide development or rearing of economically vital species like honeybees).For the purposes of this study, there were several key requirements for a species tobe chosen as a research subject:1. The species must have a sequenced genome.Although de novo transcriptome alignment and assembly is now attainablewith the aid of bioinformatics tools like Trans-ABySS and Cufflinks, it addsan additional layer of complexity to the analysis and most gene and transcriptannotations would have to be predicted (Robertson et al., 2010; Trapnellet al., 2010). Using an animal with a sequenced genome would provideimmediate results and would avoid the potential risk of falsely predicted4gene models.2. The species chosen must have stable gene annotationsBecause this study would rely heavily on externally derived orthology an-notations, the stability of gene annotations would become extremely im-portant. Using an organism with unstable, frequently changed annotationswould make it significantly more difficult to match our sequencing data withexternal data.3. Each species must be readily obtainable and easy to raise.The sample preparation protocol used in this study required a significantamount of tissue. Preliminary experiments indicated that I would need roughly100 insect heads per sample in order to obtain enough RNA for cDNA li-brary preparation and sequencing. To obtain enough tissue for sequencing,I would need an insect species that survived well in a lab environment andcould be bred in batches of at least 600 individuals, or enough samples for 3biological replicates of each condition.4. Any species chosen must be of general scientific, medical, or commercialrelevance.Identifying a novel regulator of hunger-induced behavior would be muchmore interesting and impactful if I would be able to demonstrate that it con-trols phenotypes in a major research model organism or species that peopleotherwise deal with on a regular basis.5. The group of species chosen should not be too closely related to eachother.Comparing gene regulation between distinct species allows us to argue thatany genes with conserved changes are likely more important to an animal’ssurvival. The rationale for this is that the regulatory patterns of those genesduring starvation has been maintained across millions of years of evolution.More evolutionarily distant species make this argument more compelling-their genomes and accompanying regulatory elements would have divergedsignificantly, and only a small number of the most critical gene regulatory5patterns would be likely to have been conserved between them. In contrast,a set of extremely closely related species (like members of the Drosophilagenus) would make a poor choice, as they would be expected to share a largenumber of gene regulatory changes, potentially masking interesting results.Using the criteria outlined here, we identified three insect species of poten-tial interest to our study: the fruit fly Drosophila melanogaster, the yellow fevermosquito Aedes aegypti, and the red flour beetle Tribolium castaneum. The indi-vidual factors that lead to these organisms’ inclusion or exclusion in the final studyare discussed below.1.2.1 Drosophila melanogasterDrosophila melanogaster was the most obvious choice of species for our study.Fruit flies have been used in scientific studies for over a hundred years, and alarge number of genetic tools and logistical resources exist for this organism. Bothcurrent and archived versions of the reference genome and its associated annota-tion are accessible through both Ensembl and FlyBase, with automated retrieval ofbioinformatics data made possible through APIs like BioMart (Flicek et al., 2014;dos Santos et al., 2015; Durinck et al., 2005). Drosophila is extremely well stud-ied, with experimental information and electronic annotations available for everygene. Two genetic tools made Drosophila of particular interest to our study aswell: the GAL4/UAS binary expression system and the wide availability of RNAinterference lines. These tools would allow us to knock down nearly any gene ofinterest at any time and place. The relative ease of genetic manipulation and well-characterized behaviors of Drosophila meant that the species would be an idealtool with which to screen genes for roles in behavior.1.2.1.1 The GAL4/UAS SystemThe GAL4/UAS system provides an simple and effective system for controlling theexpression of transgenic genes. GAL4, a transcription factor from Saccharomycescerevisiae, induces expression of sequences preceded by an Upstream ActivatingSequence (UAS) (Brand and Perrimon, 1993). In one expression strategy, the geneencoding GAL4 is directly inserted within a sequence of interest. GAL4 is then6expressed in the same tissues its flanking sequences are expressed in, driving ex-pression of UAS reporter/effector constructs (Brand and Perrimon, 1993). Gener-ally, Drosophila lines with a GAL4 “driver” and UAS “reporter” are created andmaintained separately. The same GAL4 line can then be used to drive expres-sion of multiple different reporters in the same target tissue, depending on whatline it is crossed with. Conversely, a researcher can test the effects of a particularUAS reporter in multiple tissues by crossing a single UAS with different GAL4lines (Duffy, 2002). GAL4 can also be reversibly inactivated by the repressor pro-tein GAL80. A temperature-sensitive version of GAL80 has been created as well(hereafter referred to as GAL80ts), allowing a researcher to temporally restrict ex-pression of UAS targets to time periods where Drosophila has been maintainedabove a “restrictive temperature” (typically 29C) (McGuire et al., 2004). In thisstudy, the GAL4/UAS/GAL80ts system allowed expression of effector genes likeRNAi constructs at virtually any time and place.1.2.1.2 RNA Interference (RNAi)RNA interference, or RNAi, utilizes transgenic expression of short RNA sequencesto induce targeted destruction of mRNA in live tissues. In Drosophila, RNAi istypically performed by UAS expression of a small hairpin or double-stranded RNAcomplementary to a target sequence. These RNA molecules are then cleaved bythe RNA processing enzyme Dicer and recruited to the RNA-induced silencingcomplex, which destroys RNAs complementary to the original shRNA sequence(Wang et al., 2006). Combined with the wide availability of UAS-RNAi lines fornearly every Drosophila gene, RNA interference is a powerful tool with which tostudy the effects of knocking down transcriptomic targets.1.2.2 Aedes aegyptiThe mosquito Aedes aegypti is a vector for yellow fever, dengue fever, and chikun-gunya. This mosquito species made an excellent candidate for this project as ithad already been widely studied and we were able to obtain specimens relativelyeasily. In addition, the online Vectorbase database has a multitude of Aedes ae-gypti-related datasets available for use, making it extremely easy to obtain the ref-7erence genome and genebuilds necessary for any bioinformatics analysis (Giraldo-Calderon et al., 2015). The only complicating factor for my analysis was the factthat female mosquitoes must consume blood in order to reproduce. To avoid anypotential pitfalls associated with this extra feeding modality, I chose to sequencemale mosquitoes and flies. This strategy would allow us to strictly study the ge-netic changes induced by food deprivation (instead of food and blood deprivationin mosquitoes).1.2.3 Tribolium castaneumThe red flour beetle, Tribolium castaneum was another strong candidate species forthis study. This insect is a major pest of stored grain products and has been widelyused as an insect developmental model. Although T. castaneum has a sequencedgenome and otherwise met all of the requirements set out in Section 1.2, severalpreliminary experiments indicated that this insect might not be the right choice forthis study. Tribolium beetles are extremely starvation-resistant, and can surviveover a week without any food or water. When its tissue sugar levels were measuredover time during starvation, these sugar levels did not follow the pattern observed inadult D. melanogaster or A. aegypti. In flies and mosquitoes, sugar levels droppedrapidly, eventually plateauing at a “base level” before death (see Figure 3.1). Thesugar levels of Tribolium beetles did not follow this pattern, with glucose levelsdropping rapidly (see Figure 1.1), then slowly rising over time until death. I hy-pothesize that this might be caused by metabolism of energy stores not quantifiedduring our experiment, like starch (the food used to rear Tribolium is simply flourand yeast mixed together). Due to these differences, I felt that the red flour beetlewas not a good candidate for transcriptomic comparison with Drosophila.1.3 Orthology as a means of transcriptomic comparisonOne of the major stumbling blocks for this study was attempting to determine whatconstitutes a “conserved transcriptomic change.” Analyzing the transcriptomes oftwo separate species would be fruitless without a biologically meaningful methodof comparing them. In order to accurately compare transcriptomes, I used theconcept of orthology to pair species’ transcriptomes. By definition, orthologs are8Figure 1.1: Measurement of Tribolium sugar levels.Glucose, trehalose, and total sugar levels quantified from Tribolium castaneumbeetles during starvation. n = 4-8 replicates of individual male beetles. Error barsrepresent standard error of the mean. Data for the equivalent experiment in D.melanogaster and A. aegypti are presented in Figure 3.1.genes in separate species that diverged from a common ancestor gene as part ofthe process of speciation. This means that orthologs will typically share signif-icant sequence identity, and possibly a similar biological function as well. Thebasic biological definition above sets no strict criteria as to how to determine whatconstitutes an ortholog. Although manual annotation of genes as orthologs is ef-fective, there are no guarantees that all researchers involved utilized the exact samemethodology. Additionally, manual annotation of orthologs will be biased towardscommonly used experimental models. Species that have been studied more willhave a larger number of annotated orthologs.To address this concern, I turned to a computational definition of orthologyinstead, in this case the OrthoDB7 database created by Waterhouse et al. (2013).OrthoDB7’s definition uses the amino acid sequence of each gene’s longest tran-script to define genes as orthologs based on their similarity across species (Wa-terhouse et al., 2013). This avoids the pitfalls of manual annotation (where themost-studied organisms would have the most data). Since each species’ genome inthe OrthoDB7 database has been sequenced, each species will have a standardized9analysis done in an identical manner to that of other species’. Furthermore, as thismethod relies only upon sequence identity, it will be able to identify potential or-thologs for every gene in every genome in the database. Using this definition, I wasable to successfully match up and compare the transcriptomes of D. melanogasterand A. aegypti using the protocol described in Section 2.6. The results and valida-tion of this approach are described in Sections 2.7, 3.2.1 and 3.2.2.1.4 Insect feeding behaviorsThe goal of this project was to identify a novel regulator of starvation-inducedbehavior. Once we identified a conserved set of genes up- or downregulated bystarvation, I planned to screen those genes for effects on behavior. If a gene wasdownregulated by starvation, this change could be reproduced using RNAi knock-down in fed flies. If a gene was upregulated by starvation, it would be possibleto prevent the gene’s upregulation using RNAi knockdown in starved flies. Ei-ther scenario allowed us to assess if a gene’s regulation resulted in starvation-likebehavior.I reasoned that any gene regulated by starvation would likely serve to increasean animal’s resistance to starvation, either by directly manipulating metabolismor modifying feeding/foraging behaviors. As a result, there are several differenttypes of feeding behavior that I chose to quantify. The first is a simple assessmentof the raw volume of food consumed. Like any other animal, flies eat more afterstarvation. Another behavioral paradigm chosen for study was caloric sensing.Flies are capable of sensing the calorie content of foods, and adjusting their foodintake accordingly (Stafford et al., 2012). Genes’ effect on these two behaviorswas assessed using a modified version of the CAFE assay (Ja et al., 2007). Adescription of the protocol used is found in Section 2.9.Sleep is also known to be regulated by internal metabolic state. Previous workperformed by Keene et al. (2010) indicated that starved flies will suppress sleepin order to forage for food. This is manifested in a major decrease in sleep after12 hours of starvation. Although this phenomenon is known to be modulated bythe clk/cyc group of circadian neurons, the genes and neural circuit responsible fordirectly controlling this behavior remain unknown (Keene et al., 2010). This sleep10suppression phenotype can be assessed using the Drosophila Activity Monitor inthe manner described in Section 2.10.1.5 Juvenile HormoneOne class of genes known to be heavily implicated in feeding-related behaviors arehormones and their respective hormone-processing enzymes. This study identifiedJuvenile Hormone as a regulator of starvation-related behavior. To aid in inter-pretation of experimental results, a comprehensive review of Juvenile Hormone isprovided here.Juvenile Hormone (JH) is an insect developmental cue, acting in opposition to20-hydroxyecdysone (ecdysone) to regulate the growth from larva to adult. A shortpulse of ecdysone initiates each major developmental transition in insects, whetherit be molting or metamorphosis. JH acts to oppose ecydone, and its absence in thefinal instar of insect development allows ecdysone to trigger metamorphosis. Asa result, the most well characterized function of JH is it’s ability to delay meta-morphosis, either adding extra larval instars in moths and beetles, or preventingcellular differentiation and triggering apoptosis in higher diptera like Drosophila(Bernardo and Dubrovsky, 2012). Although ecdysone and JH are both steroid hor-mones that act on similar processes, JH is less well studied, and its exact modes ofaction remain somewhat unclear. JH appears to have a wide ranging set of effectsin addition to its metabolic role, including sexual behavior, pheromone production,caste determination, diapause, migration, and the synthesis of gonadal proteins(Bernardo and Dubrovsky, 2012).Part of Juvenile Hormone’s complexity arises from the fact that JH is not ac-tually a single hormone, but a collection of extremely similar steroids that eachhave biological activity. Different species of insects secrete different JuvenileHormones. Lepidopteran insects (butterflies and moths) secrete five Juvenile Hor-mones: JH 0, JH I, JH II, JH III, and 4-methyl JH I. Drosophila, on the other hand,secrete three different Juvenile Hormone derivatives: JH III, Juvenile Hormone IIIBisepoxide (JHB3), and methyl farnesoate (MF). JHB3 is only secreted in higherDiptera like Drosophilid flies. Heteropteran insects (a suborder of insects that in-cludes bedbugs and water striders) secrete an additional JH: Juvenile Hormone III11Figure 1.2: The Juvenile Hormones.Chemical structures of the Juvenile Hormones found in Drosophila, breakdownproducts of Juvenile Hormone III, and synthetic Juvenile Hormone agonists/antag-onists used in this study.Skipped Bisepoxide (Noriega, 2014). For simplicity’s sake (and since this studyfocuses so heavily on Drosophila), this introduction will focus on JH III, JHB3,and MF.121.5.1 Synthesis of Juvenile Hormone derivativesJuvenile Hormones are synthesized in the corpora allata, an endocrine gland lo-cated next to the esophagus in Drosophila. This synthesis pathway for JH is partof the larger mevalonate pathway found throughout the animal kingdom (responsi-ble for cholesterol synthesis in mammals) (Noriega, 2014). Briefly, multiple unitsof Acetyl-CoA are converted into farnesyl pyrophosphate by nine enzymes in themevalonate pathway. The Juvenile Hormone biosynthetic pathway diverges here,and farneysl pyrophosphate is converted to farnesol by farnesyl pyrophosphatase.Farnesyl dehydrogenase then converts farnesol to farnesoic acid in a two-step reac-tion. In insects that produce only JH III, farnesoic acid is used to produce methylfarnesoate by Juvenile Hormone acid methyltransferase (JHAMT). Methyl farne-soate is processed by Methyl farnesoate epoxidase (a cytochrome P450 CYP15expoxidase) to produce JH III (Noriega, 2014).Unfortunately, the last two steps of JH synthesis in Drosophila are not as clear.Although farnesoic acid (FA) is the common precursor for JH III, JHB3, and MF,there are potentially mutiple synthesis pathways for each, and interconversion be-tween these JH derivatives does occur (Wen et al., 2015). The first major differ-ence in the synthesis pathway is the lack of a clear CYP15 epoxidase in flies -the Drosophila CYP15 sequence has evolved so signifcantly that it is no longerconsidered a member of the CYP15 gene family. Wen et al. (2015) proposed thatCYP6G2 may function in a similar role to CYP15s due to CYP6G2’s expressionin the Drosophila corpora allata, but admit that there is no biochemical evidencefor this function either in vitro or in vivo . It is thought that whatever enzyme cat-alyzes this step (CYP6G2 or an unknown cytochrome P450 expoxidase) producestwo products: 10,11 epoxifarnesoic acid or 6,7;10,11 epoxifarnesoic acid, actingas a precursor for JH III or JHB3, respectively (Wen et al., 2015). The higher abun-dance of JHB3 in Drosophila may due to a preference of CYP6G2 for 6,7;10,11epoxifarnesoic acid as a product (Wen et al., 2015).Another major difference in Drosophila’s synthesis pathway is the role ofJHAMT. Wen et al. (2015) found that overexpression or mutation of JHAMT (nor-mally responsible for producing MF from FA in other insects), did not affect therate of JH III or MF synthesis. Unexpectedly, only the amount of JHB3 changed.13This indicates that JHAMT is primarily responsible for synthesizing JHB3 from6,7;10,11 epoxifarnesoic acid, unlike the role it plays in producing MF in otherinsects (Wen et al., 2015).This is the extent of current knowledge of Juvenile Hormone synthesis inDrosophila. The methyltransferase that converts FA to MF in the corpora allataremains unknown. The methyltranferase responsible for producing JH III from10,11 epoxifarnesoic acid has not been identified either.Wen et al. (2015) foundevidence for interconversion of MF and JH III into JHB3 in the hemolymph offlies, presumably by another cytochrome P450 expoxidase that has yet to be iden-tified.1.5.1.1 Regulation of Juvenile Hormone synthesisIn mammals, the mevalonate synthesis pathway is regulated by sterol moleculesthrough Sterol Regulatory Element Binding Protein (SREBP). This is not the casein insects, as homologs for the SREBP and SREBP-2 proteins do not exist in theseanimals (Goodman and Cusson, 2012). Instead, it is thought that JHs directlyregulate their own synthesis, and have been shown to have a regulatory effect onmevalonate synthesis enzymes in tissues outside the corpora allata (Goodman andCusson, 2012).Perhaps more strikingly, it has been reported that Short Neuropeptide F (sNPF)may be involved in regulating Juvenile Hormone biosynthesis. While working onthe silkworm Bombyx mori, two putative sNPF receptors were identified in the cor-pora allata. When B. mori sNPF peptides were assayed for effects on JH synthesisin vitro, they had a strong inhibitory effect (Goodman and Cusson, 2012). Thisis especially interesting, as sNPF has previously been shown to be heavily impli-cated in starvation-related behaviors such as odor-driven food search and controlof food intake in Drosophila melanogaster (Hong et al., 2012; Root et al., 2011).The corpora allata is heavily innervated by higher brain centers, including nervesthat originate in the subesophageal ganglion, a region of the brain that plays a ma-jor role in feeding behaviors (Goodman and Cusson, 2012). Although this is purespeculation, it may be the case that synthesis of JH-related compounds may beregulated by higher brain centers through neuropeptides like sNPF. Although this14seems to be a likely explanation for these phenomena, there is no direct experimen-tal evidence for this.1.5.2 Juvenile Hormone metabolism in the hemolymphOnce synthesized by the corpora allata, JH III, JHB3, and MF are secreted intothe hemolymph. Once there, these JH hormone derivatives are capable of beingacted upon by a number of different proteins, including several unknown enzymespreviously mentioned in Section 1.5.1. One of these proteins, Takeout (TO), isthought to be responsible for JH transport, whereas Juvenile Hormone esterase(JHE) and Juvenile Hormone epoxide hydrolase (JHEH) degrade it.1.5.2.1 TakeoutThere are many proteins in insects capable of transporting Juvenile Hormones, in-cluding both low-affinity, general binding proteins like lipophorins and specificbinding proteins that bind JHs with extremely high affinity, like Takeout. Take-out belongs to a small family of hemolymph Juvenile Hormone binding proteins(hJHBPs). The Manduca sexta hJHBP preferentially binds JHs I, II, and III, butdirect binding has never been demonstrated in Drosophila (Goodman and Cusson,2012). hJHBPs are thought to have a number of important physiological roles,including transport of JHs to target sites, reducing promiscuous activity in non-target tissues, preventing JH degradation, and providing a reservoir of hormonenear target tissues (Goodman and Cusson, 2012). In Drosophila, takeout mutantsdemonstrate increased starvation sensitivity, increased feeding, increased sensi-tivity of sugar-specific gustatory neurons, and increased lifespan (Meunier et al.,2007; Sarov-blat et al., 2000; Chamseddin et al., 2012).There is evidence for hJHBP regulation by JH itself. In Manduca sexta, JHtiters and hJHBP expression are inversely related (Goodman and Cusson, 2012).Drosophila’s Takeout is regulated in a circadian manner, and its expression is de-pendent on the circadian genes per and tim (Sarov-blat et al., 2000).151.5.2.2 Juvenile Hormone esteraseThere are two primary pathways of JH degradation in insects- conversion of JHIII to JH acid by Juvenile Hormone esterase, or conversion of JH III to JH diolby Juvenile Hormone epoxide hydrolase. Although Drosophila has a number ofJhe-like Juvenile Hormone esterases (including the Jhe gene duplication Jhedup),only JHE has been shown to readily metabolise JHs in vitro (Crone et al., 2007).Drosophila’s JHE is capable of metabolising multiple forms of JH, including JHI, JH II, JH III, JHB3, and MF, but not synthetic JH analogs like methoprene,hydroprene, or kinoprene (Crone et al., 2007; Campbell et al., 1998). JHE hasan extremely high affinity for JH III, with a Km of either 1.5 µM or 89 nM forrecombinant or tissue-purified protein respectively. The catalytic efficiencies weresimilar for both studies, with Kcat values of 1.0 s−1 and 0.6 s−1 (Crone et al., 2007;Campbell et al., 1998). It is clear that Drosophila JHE is capable of degrading allnaturally occuring forms of JH, while leaving synthetic analogs like methopreneintact (synthetic JHs were designed to be as stable as possible for use as broad-spectrum insecticides).Jhe has three annotated transcripts: Jhe-RA, Jhe-RB, and Jhe-RC. The onlydifference between these transcripts lies in their 5’ UTR- all 3 transcripts producean identical peptide product (Celniker et al., 2009). Data from the modENCODEproject indicates that expression of Drosophila Jhe occurs primarily in the pupalfat body and adult head, with highest expression during the pupal stage (Celnikeret al., 2009). In a manner opposite to Takeout, Jhe expression is induced by JHand suppressed by ecdysone (Kethidi et al., 2005). This corresponds to in vivolevels of Juvenile Hormone. Temporal expression of Jhe matches JH abundanceduring development, with expression peaks during each larval stage. JHE proteinis found both within Drosophila is primarily found within the hemolymph, andits abundance is tightly controlled by post-translational mechanisms. Injection ofrecombinant JHE into Manduca sexta larvae demonstrated that the protein has ahalf-life of 1.2 to 3.6 hours depending on the amount injected, whereas controlproteins of similar mass had a half-life of days (Ichinose et al., 1992). This rapiduptake and destruction of JHE protein is mediated by receptor-mediated endocyto-sis followed by lysosomal degradation in pericardial cells (cells that surround the16insect heart) (Ichinose et al., 1992; Bonning et al., 1997).1.5.2.3 Juvenile Hormone epoxide hydrolaseAlthough less well studied, Juvenile Hormone epoxide hydrolase (JHEH) is theother enzyme responsible for JH degradation in insects. The relative contributionof JHE and JHEH towards JH degradation changes over Drosophila’s lifespan.By measuring the relative amount of JH acid and JH diol (produced by JHE andJHEH respectively), Campbell et al. (1992) determined that JHEH dominates JHcatabolism during the larval stage, JHE controls JH degradation in pupae, and bothenzymes contribute relatively equally in the adult stage. This evidence is corrob-orated by two similar experiments that measured in vitro activity of each enzymepurified from Drosophila melanogaster and Drosophila virilis (Casas et al., 1991;Rauschenbach et al., 1995).Drosophila has three JHEH genes: Jheh1, Jheh2, and Jheh3. All three ap-pear to show relatively widespread expression, although Jheh3 appears to showhighest expression in the digestive system (Celniker et al., 2009). The DrosophilaJHEHs have been demonstrated to metabolise JH III and JHB3 readily in vitro,although JHEH-containing fractions hydrolyze JH III with between three and tentimes greater efficiency than JHB3 (Casas et al., 1991). Note that JHEH cannothydrolyse methyl farnesoate, as this molecule does not contain any epoxide func-tional groups. JHEH’s preference for JH III as a substrate may explain why JHEcontributes so heavily to JH degradation in adult flies despite the greater adult ex-pression of Jheh (Celniker et al., 2009). MF and JHB3 are simply far more abun-dant substrates in higher Dipteran (fly) larvae and adults, and JHE is capable ofmetabolising these with much greater efficiency (Wen et al., 2015; Yin et al., 1995;Campbell et al., 1998). If anything, JHE’s contribution to JH degradation may bemore important in the greater Drosophila genus: addition of DFP, a general esteraseinhibitor, decreased the amount of JH III degradation of adult fly homogenate bymore than 10 fold in D. virilis compared with more than 2 fold in D. melanogaster(Rauschenbach et al., 1995).171.5.3 Juvenile Hormone signallingJuvenile Hormone is known to act upon a large number of target tissues. Theseeffects are transduced by a number of different receptors, the two most well knownof which are the MET/GCE heterodimeric receptor and USP, an interaction partnerof the ecdysone receptor. Other studies have suggested that JHs are capable ofbinding with as-yet unidentified G-protein-coupled receptors, but evidence for thisis currently lacking (Goodman and Cusson, 2012).1.5.3.1 Methoprene-tolerant and Germ-cell ExpressedMethoprene-tolerant (Met) was the first true JH receptor identified. Ethyl methanesulfonate-based mutagenesis of D. melanogaster produced a mutant (Met) that was100 times more resistant to Juvenile Hormone III or methoprene added to food.This mutation allowed larvae to survive otherwise lethal concentrations of metho-prene, did not develop methoprene-induced pseudotumors, and exhibited normalvitelligenic oocyte development in the presence of JH agonists. Importantly, Metdid not confer any resistance to other insecticides, indicating that the gene did notencode a general insecticide resistance, but a protein product specific in its inter-action with JH (Wilson and Fabian, 1986). Further analysis by Shemshedini et al.(1990) demonstrated that Met bound directly to JH III, with a dissociation con-stant (at which 50 percent of substrate is bound) of 6.7 nM. The Met mutation,on the other hand, increased this value to 38 nM. Later analyses indicated thatMet encoded a member of the basic-helix-loop-helix-PAS family of transcriptionalregulators (Wilson and Ashok, 1998).Remarkably, a Met mutation that produced no transcript, Met27, was able tosurvive to adulthood and reproduce, albeit with an 80 percent reduction in egg pro-duction (Wilson and Ashok, 1998). Since proper levels of JH signaling is criticalfor insect survival and reproduction, this indicated that another gene must be ableto transduce JH’s effects in the absence of Met (Wilson and Fabian, 1986; Bowerset al., 1976; Wilson and Ashok, 1998).Germ-cell Expressed (gce), another basic-helix-loop-helix-PAS gene, was soonidentified as a critical interaction partner for Met. Pull-down assays in DrosophilaS2 cells indicated that MET protein bound GCE, and this binding was disrupted18by addition of either JH III or methoprene (Godlewski et al., 2006). Importantly,Met was able to form both MET-MET homodimers as well as MET-GCE het-erodimers, indicating that the two receptors may be partially redundant (Godlewskiet al., 2006). Although gce is able to substitute for Met expression in vivo, RNAiof gce in Met mutants is lethal (Baumann et al., 2010).Both MET and GCE transduce their effects through the FTZ-F1 transcriptionfactor. The E75A is a nuclear receptor gene with expression induced by JHs.Removal of FTZ-F1 via RNAi prevented E75A expression in response to JH ap-plication, whereas FTZ-F1 overexprssion increased it (Dubrovsky et al., 2011).This transcription factor forms heterodimers with either MET or GCE, and trans-genic expression of Manduca sexta Jhe blocked JH-induced expression of E75A(Dubrovsky et al., 2011).1.5.3.2 UltraspiracleAnother known receptor for JH is the nuclear receptor Ultraspiracle (Usp). Jonesand Sharp (1997) demonstrated that USP directly binds JH III and JHB3 (but notfarnesol or ecdysone), inducing a conformational change and oligomerization ofUSP protein, with a dissociation constant of 0.5 µM . Further work on the originalstudy indicates that USP is also capable of binding MF in addition to JH III and MF(Jones et al., 2010). The vertebrate ortholog of USP, the retinoid X receptor, hasalso been shown to respond to methoprene at high concentrations (Harmon et al.,1995).USP is both a transcription factor and an interaction partner of the ecdysonereceptor EcR (Iwema et al., 2009). Goodman and Cusson (2012) proposed thatUSP may interact differently when JH or ecdysone are present: USP is capable ofbinding JH response elements in the presence of JH, ecdysone binds to USP:EcRto initiate metamorphic molting, and when both JH and ecdysone are present, bothhormones are capable of binding to this receptor complex and triggering larvalmolting. The dual interaction of USP with both JH and ecdysone signaling may ex-plain the interaction of these two pathways, integrating input from both hormones.It is worth noting that exogenous application of ecdysone to adult Drosophila iscapable of inducing sleep in an EcR-dependent manner (Ishimoto and Kitamoto,192010). This is the opposite effect of adding methoprene to adult flies (see Sec-tion 3.6), and these two seemingly antagonistic phenotypes may in fact be mediatedby ecdysone and JH’s dual interaction with USP.1.5.4 Known roles and effects of the Juvenile HormonesJuvenile Hormones are one of the most important insect developmental hormones,and a disruption of normal titres during this period results in death (Goodman andCusson, 2012). Although exact levels and composition of JHs differ from insect toinsect, JH levels follow a simple pattern during development: JH levels are highthroughout the larval stage, then drop precipitously to permit ecdysone signallingand the onset of metamorphosis (Goodman and Cusson, 2012). Almost all researchon this class of hormones has focused on this relatively narrow period of time, mostlikely due to the obvious commercial applications of JH disrupting compounds forpest control (as an example, the World Health Organization recommends additionof methoprene to potable drinking water as a mosquito larvicide) (Organization,2008). Despite this focus on regulation of metamorphosis, JHs have a number ofother effects that will be discussed here.JH has long been known to affect epidermal development. Much of the workin this area has focused on the JH induced expression of cuticular proteins likeLCP14/16/17 and pigmentation/melanization-related proteins like insecticyanin anddopa decarboxylase (Goodman and Cusson, 2012). Many of the LCP family ofproteins (larva cuticular proteins) appear to be both up and downregulated by JHdepending on the protein. It is thought that this process may be mediated by FTZ-F1 for LCP14, as the LCP14 gene has three potential FTZ-F1 binding sites in closeproximity (Goodman and Cusson, 2012). JH has effects on gene expression inother tissues besides the epidermis as well. Addition of exogenous JH to Manducasexta larvae decreased expression of insecticyanin-a and insecticyanin-b in boththe fat body and epidermis (Li and Riddiford, 1996).JH is known to play a major role in gene regulation in the insect fat body (Ar-rese and Soulages, 2010; Goodman and Cusson, 2012). The fat body, analogous tovertebrate adipose tissue, is an organ responsible for a number of metabolic func-tions, including storage and release of nutrients, synthesis of proteins like vitel-20logenins used in other tissues, and even as a nutrient sensor that triggers releaseof insulin-like peptides in Drosophila (Arrese and Soulages, 2010; Rajan and Per-rimon, 2012). Several research groups have demonstrated that JH controls ex-pression of TOR signalling pathway components in Aedes aegypti and Triboliumcastaneum (Shiao et al., 2008; Parthasarathy and Palli, 2011). This change in TORsignalling levels controls production of vitellogenin, a critical protein for egg lay-ing and development (Shiao et al., 2008; Parthasarathy and Palli, 2011).One of the most interesting effects of Juvenile Hormones is their poorly studiedrole in neurons and behavior. Addition of methoprene to the ant Phiedole bicari-nata will turn worker-destined larvae into soldier ants (Wheeler and Nijhout, 1981).Higher JH levels in the last larval stage of some migratory insects could induce astationary adult stage instead of a migratory one (Goodman and Cusson, 2012).JH is capable of affecting neuronal remodeling as well, with JH treated neuronsdisplaying decreased dendritogenesis relative to controls (Williams and Truman,2005). These behavioral effects appear to be caused by effects on development, asall JH treatments during these studies occured during or before metamorphosis.Importantly, several studies have demonstrated that JH III is capable of act-ing directly upon neurons to affect their activity outside development. Addition ofmethoprene or JH III caused short-term depression of cockroach neurons both invitro and in vivo (Richter and Gronert, 1999). JH-induced short-term depressionwas induced rapidly, with a 25 percent reduction in spike rate within 2 minutes,and a 75 percent reduction after 14 minutes (Richter and Gronert, 1999). Anotherexperiment in the cricket Acheta domesticus indicated that JH III injection caused atranslation-dependent decrease in the sound threshold required for an auditory neu-ron to spike by as much as 20 decibels. More importantly, this same injection of JHIII could induce a change in phonotaxis (an animal’s response to sound), causinganimals to circle towards the side of the brain that received the injection after asound was played via loudspeaker, with a decreased response threshold of up to 35decibels (Stout et al., 1991). These studies make it abundantly clear that JuvenileHormone is capable of acting directly on neurons to produce electrophysiologicaland behavioral changes.21Chapter 2Methods2.1 Insect husbandryUnless otherwise stated, Drosophila melanogaster stocks used in this study wereraised at 20 degrees C and 70 percent relative humidity. The stocks used for thisstudy are as follows: Canton S, w[1118] (VDRC GD library injection strain),dcr2; Gal80tsCyO ;nysb−GAL4T M2 , elav-GAL4;dcr2 (Bloomington), UAS-GFP RNAi VAL-IUM10 (Bloomington), UAS-Jhe RNAi VALIUM20 (Bloomington), and UAS-Jhe RNAi GD (VDRC). The identity of RNAi lines was verified by PCR beforeuse. Stocks were raised on Nutri-Fly Bloomington Formulation pre-mixed media(Genesee Scientific). 3 to 5 day old adult flies were used for experiments, and weremaintained for at least 2 days at 25 degrees C and 75 percent relative humiditybeforehand. For a full list of genotypes used in figures, refer to Table 2.1.The Liverpool strain of Aedes aegypti mosquitoes was used for this study. Fiveto seven day old adult mosquitoes were used for experiments, and maintained on10 percent sucrose solution at 25 degrees C and 75 percent relative humidity. Allinsect strains were maintained on a 12 hour light / 12 hour dark circadian cycle.2.2 Tissue sugar quantificationTissue sugar levels for each insect were quantifies using a modified version ofthe methodology decribed by Miyamoto et al. (2012), where sugars are quantified22Table 2.1: Key to Drosophila genotypes used in figures.Abbreviation Genotype Appearanceelav>Jhe RNAi elav−GAL4+ ;UAS−Dcr2+ ;+UAS−Jhe RNAi GD Figures 3.6, 3.9, 3.10, 3.11 and 3.14elav>+ elav−GAL4+ ;UAS−Dcr2+ Figures 3.6, 3.9, 3.10, 3.11 and 3.14+>Jhe RNAi +UAS−Jhe RNAi GD Figures 3.6, 3.7, 3.9, 3.10, 3.11 and 3.14elav>Jhe RNAi #2 elav−GAL4+ ;UAS−Dcr2+ ;+UAS−Jhe RNAi V L20 Figure 3.8elav>+ elav−GAL4+ ;UAS−Dcr2+ ;+UAS−GFP RNAi V L10 Figure 3.8+>Jhe RNAi #2 UAS−GFP RNAi V L10UAS−Jhe RNAi V L20 Figure 3.8Dcr;Gal80ts;nsyb>Jhe RNAi UAS−Dcr+ ;tub−Gal80ts+ ;nsyb−GAL4UAS−Jhe RNAi GD Figure 3.7Dcr;Gal80ts;nsyb>+ UAS−Dcr+ ;tub−Gal80ts+ ;nsyb−GAL4+ Figure 3.7after whole animal homogenization. Although direct sugar quantification of ex-tracted hemolymph has a number of advantages over this technique, it was deemedunfeasible over long periods of time and across large sample numbers.In assays using Drosophila, ten male Canton S flies were maintained for 2 daysin experimental conditions on standard fly food, and then starved on 1 percent agarupon experiment start. The equivalent experiment for Aedes aegypti was performedby placing 5 male mosquitoes in an empty Drosophila culture vial with two holesdrilled in the bottom. A cotton ball moistened with dH2O was placed at the bottomof the vial allowing mosquitoes to drink water as needed. The cotton was re-moistened every 8-12 hours by injection with additional dH2O via pipette. Forboth organsims, sugar levels were quantified every 8 to 12 hours until greater than50 percent mortality occured. Dead animals were not used for measurements. Theentire experiment for both organims took place in a 25 C incubator with a 12 hourlight / 12 hour dark cycle at 75 percent relative humidity.In order to measure sugar levels, 100 µL of trehalase buffer (5 mM Tris-Hcl,137 mM NaCl, 2.7 mM KCl, pH 6.6) was added to 5 adult male flies or 3 adultmale mosquitoes for each sample. This sample was homogenized on ice and thenincubated at 80C for 15 min. The resulting sample was immediately spun down at13500g for 5 minutes and the supernatant was either frozen at -20C or used for im-23mediate sugar quantification. Glucose and trehalose quantification was performedusing the Megazyme Trehalose Assay Kit (K-TREH) according to manufacturerinstructions.2.3 RNA sample preparationFor each Drosophila melanogaster sample, 100 males were maintained on stan-dard fly food for two days and then placed on 1 percent agar w/v, 1 percent agar(starved) + glucose (glucose fed), 1 percent agar + yeast (yeast fed), or 1 percentagar + arabinose (arabinose fed) for 24 hours. Aedes aegypti samples were pre-pared in a similar manner, with 100 male mosquitoes per sample. Mosquitoeswere fed 10 percent sucrose w/v for two days, and then either water (starved) or10 percent sucrose solution (fed) for 48 hours. After this step, sample preparationwas identical for each species.Heads were removed and RNA was extracted after homogenization in TRIzol(Life Technologies) according to manufacturer instructions. RNA clean-up wasperformed using the column and protocol from Life Technologies’ GeneJet RNAPurification Kit. RNA concentration and integrity were verified by spectrophotom-etry and agarose gel electrophoresis.2.4 cDNA library preparation and sequencingBefore library preparation and squencing, RNA sample concentration and integritywas re-verified using a Qubit fluorimeter (Life Technologies) and BioAnalyzer(Agilent Technologies). mRNA-enriched cDNA libraries were produced using Il-lumina’s TruSeq RNA Library Preparation Kit according to manufacturer instruc-tions. Sequencing was either performed in a single-end 50bp configuration oneither two (Aedes aegypti) or six (Drosophila melanogaster) lanes of an IlluminaHiSeq 2000. Base calling, demultiplexing, and creation of FASTQ files was per-fomed using CASAVA (Illumina).242.5 RNA sequencing analysis pipelineReference genomes and genebuilds were obtained from either FlyBase (Drosophilamelanogaster) or VectorBase (Aedes aegypti). I chose to use Drosophila’s BDGP5.51 (May 2013) and Aedes’s AaegL1.3 (May 2012) annotations as those geneb-uilds were used to construct the OrthoDB7 orthology database used in Section 2.6(Waterhouse et al., 2013). Bowtie2 indexes were built from reference genomesusing bowtie2-build (version 2.1.0) (Langmead and Salzberg, 2012).RNA-Seq reads were aligned using TopHat2 (version 2.0.10) in GTF reference-guided mode using a transcriptome index (Kim et al., 2013). Output BAM fileswere sorted by name-sorted (reads sorted by their aligned feature’s name) withsamtools and raw read counting was perfomed at the gene level using HTSeq’shtseq-count script (version 0.6.1) under intersectionstrictmode (An-ders et al., 2014). Differential gene expression calls were performed using theDESeq2 R package (version 1.4.5) (Love et al., 2014).2.6 Identification of conserved transcriptomic changesThis study used OrthoDB7’s Diptera database for determination of ortholog pairs.This means that the definition of ortholog used in this study is as identical to thatused by Waterhouse et al. (2013): genes from each species pair were clusteredusing the best-reciprocal hits (BRHs) in an all-versus-all comparison using theSWIPE algorithm. Only the single longest transcript for each gene was considered,and minimum cutoff to be included in a cluster required an e-value below 1×10−3for multiple BRHs, or below 1× 10−6 for pair-only BRHs. A minium sequenceidenity of at least 30 amino acids was also required for inclusion in a cluster (Wa-terhouse et al., 2013). Ortholog pairs were generated from the OrthoDB7 Dipteradatabase by creating every possible pairing of Drosophila melanogaster and Aedesaegypti orthologs in each cluster using R. Significantly-regulated ortholog pairswere identified by selecting ortholog pairs where both genes in each pair had anadjusted p-value below 0.05. The significance of these genes was assessed usingelectronic annotations retrieved from the Ensembl database using the biomaRt Rpackage (Durinck et al., 2005).252.7 Validation of ortholog resultsNo formal statistical test exists to determine whether or not the number of ob-served significantly-regulated ortholog pairs is different from that which would beexpected if genes were expressed randomly in each species (no conservation ofgenetic regulation). Nevertheless, the number of ortholog pairs expected due torandom expression can be simulated. This means that the probability of the ob-served number of pairs being due to random expression can be calculated throughiteration of this simulation in a Monte Carlo method (repeated sampling approxi-mates the true value).Written in Java, the simulation loads a set of orthologs from a .csv file andstores them in computer memory. A binomial distribution representing the numberof significantly expressed genes in each species is then generated using a specificvalue of α for each species (when a gene is significantly expressed, it is defined as“on”). α represents the probability of any given gene being on in a species, and iscalculated using the following formula:α =nSetsetSize(2.1)where:• nSet = the number of elements defined by: # of significantly expressed genesin a species present in the input gene set for OrthoDB ortholog calculationfor that species• setSize = the number of elements in the OrthoDB input gene setThe number of differentially expressed genes for a species is determined bysampling a randomly from the binomial distribution. This number of genes is thenrandomly sampled from one species’ genes in the ortholog set. This process isrepeated for the second species using a second α calculated for that species. Thisgive a set of genes that are “on” for each species. The simulation then compareseach species’ set of activated/inactivated genes, and counts the number of pairs (k)where both genes in an ortholog pair are “on”. k represents the number of pairs ex-pected due to random expression of genes for one iteration of the simulation. The26optimal number of iterations of this simulation was selected by monitoring the cu-mulative distribution function (CDF) and probability mass function (PMF) of k forvarious numbers of iterations per simulation. If the CDF and PMF remained stableacross different numbers of iterations, it was assumed that the simulation had con-verged, approximating the true value of k. If the CDF and PMF differed betweentrials, more samples were required. Through this method, it was determined thatan iteration depth of one million samples was sufficient to accurately determine k.2.8 Methoprene and Precocene I feedingMethoprene stock solution was created by the addition of 5.8 µL methoprene (5.35µg methoprene) to 5 mL of 95 percent ethanol. In experiments where methoprenewas used as a food additive, either 100 µL of methoprene stock solution or ethanolwas added to 50 mL of food preparation. Oral delivery of this concentration ofmethoprene has previously been demonstrated as sufficient to block Drosophiladevelopment (Restifo and Wilson, 1998). Effectiveness of methoprene additionto food in our experiments was verified visually by delay of Drosophila larvaldevelopment and high pupal lethality.Precocene stock solution was prepared using 0.093 µL precocene I dissolved in1 mL of 95 percent ethanol. 100 µL of stock solution was added per every 50 mL offood preparation. The dosage used in this study was determined by observing thesleep phenotypes of flies when subjected to a series of 2-fold dilutions of precocenebeginning from 0.25 mg precocene I/mL food. The final dosage used in this studywas determined by using the concentration that caused no lethality and did notaffect the activity index of adult male Drosophila (see Section 2.12 for an overviewof how this statistic was calculated). Males were used for this measurement, aspreliminary experiments indicated that sex was much more sensitive to the toxiceffects of precocene I. In experiments where either precocene or methoprene werefed to flies, control flies were fed an equivalent amount of ethanol as a vehiclecontrol.272.9 Measurement of Drosophila food intakeTo quantify Drosophila food intake, we performed Capillary feeding (CAFE) as-says similar to previous work (Stafford et al., 2012; Ja et al., 2007). Four male andfour female flies were placed in a 15 mL conical centrifuge tube with 4 holes drilledin the lid, and allowed to recover for 30 minutes to 1 hour after anaethetization withCO2. The lid holes were filled with the cut ends of 200 µL pipette tips and severaladditional holes were added near the base of the tube to allow air circulation. Eachvial represents one replicate of data, and during a normal CAFE assay, up to 20vials were fitted into holes of an airtight secondary plastic container. During eachexperiment, two vials did not contain flies, acting as an effective control to deter-mine the amount of food lost to evaporation. The secondary container had a thinlayer of water in the base to maintain humidity during an experiment, minimizingevaporation. Four capillaries (0.5 mM inner diameter, A-M Systems) half-filledwith food solution were placed inside the pipette tip adapters at the top of each vialto act as the food source. The food solution used in this study was comprised of50 mM D-glucose, with 0.015 percent w/v FD&C Blue No. 1 dye to aid visibility.The FD&C Blue dye was chosen because preliminary work demonstrated that itdoes not affect Drosophila food preference. After capping each capillary with asmall amount of mineral oil (to limit evaporation), the level of food in each capil-lary was marked with a fine marker. This setup was then photographed once everyhour for 20 hours. The photos were analyzed in ImageJ, and the amount of con-sumption was measured by calculating the distance between the level of solution ineach capillary and the original food level indicated by the marker. Distances wereconverted from raw pixels to inches using a known reference distance on each vial,and the volume of food consumed in µL was calculated. The average amount ofevaporation (as determined by the loss of food in the two controls without flies)was subtracted from test capillaries to calculate the true volume of food consumedby flies during an assay. Unless stated otherwise, all experiments were performedat 29 degrees C and 75 percent humidity with 24-hour lighting. In experimentsusing GAL80ts, flies were maintained in these conditions for 5 days beforehand inorder to fully induce GAL4 expression.282.10 Measurement of Drosophila sleepDrosophila activity and sleep was monitored using Trikinetics’ DAM2 DrosophilaActivity Monitor system. 32 flies were placed in 5 mm diameter tubes, with oneend containing 1 cm worth of food. The food used for these experiments eitherconsisted of 2 percent agar plus 5 percent sucrose w/v (sucrose food) or 2 percentagar (agar food). When methoprene was used as a food additive, either methoprenesolution or ethanol was added directly to the food in the activity monitor tubes asdescribed above. Experiments were run at 25 degrees C and 75 percent humidityon a 12 hour light / 12 hour dark cycle. Flies were placed inside the activity moni-tor tubes for at least two days before each experiment to acclimate to experimentalconditions and the DAM2 system. Sleep suppression was assessed by placing theflies on sucrose food for 24 hours, followed by agar food for 24 hours, and finallyanother period of 24 hours of sucrose food. Drosophila activity was recorded ev-ery 5 minutes. Sleep and activity phenotypes were assessed using the actmon Rpackage written for this study. Data from flies that died during the course of anexperiment were discarded.2.11 Measurement of Drosophila starvation sensitivityDrosophila starvation sensitivity was assessed using Trikinetics’ DAM2 system.Briefly, 32 flies were placed in 5 mM diameter tubes containing 1 cm worth of agarfood in one end. The activity of the flies was recorded every five minutes until allflies had succumbed to starvation. Experiments were run at 25 degrees C and 75percent humidity on a 12 hour light / 12 hour dark cycle. Starvation sensitivity andsurvival was assessed using the actmon R package.2.12 actmon R packageTo analyze data produced by the DAM2 system, I developed the actmon R pack-age to effectively quantify Drosophila sleep and activity behavior. actmon allowsfast, easy, and reproduceable analysis of data produced by Trikinetics’ devices inR. A S4 helper class is provided to hold an activity monitor experiment and itsmetadata, along with a number of methods to simplify data handling and analy-29sis. The package reduces complex tasks like removing dead animals or syncingan experiment’s output data to the light/dark cycle in an incubator to simple one-line functions. All functions and methods are designed to maintain the originalTrikinetics’ format and experimental metadata at every step, allowing actmonto be used either by itself or in conjunction with other sleep analysis tools likePySolo, ActogramJ, or FaasX (Gilestro and Cirelli, 2009; Schmid et al., 2011).actmon also provides methods to convert an experiment to the widely-used tidyR“long-data” format, enabling easy follow-up analysis in R. actmon also provideswrapper functions that provide the ability to produce publication-ready plots usingthe ggplot2 graphics package.Sleep was defined as 5 minutes of no activity, in accordance with the accepteddefinition established by Hendricks et al. (2000). Time of death was determinedby identifying the final contiguous period of time in which no movement occurredfor the rest of the experiment (<5 counts / hour). As an additional requirement,the duration of this final motionless period must be at least 4 hours in length beforeit can be considered death. Activity index (locomotor activity normalized to timeawake) was calculated by the following formula: # of activity counts per day /# of minutes awake per day (Kume et al., 2005). Sleep bouts were detected byidentifying contiguous periods of sleep after sleep detection has been performed.The source code and installation instructions for this package are publicly availableonline at: https://github.com/kazi11/actmon.2.13 Plotting and statisticsWith the exception of the ortholog statistical simulation described in Section 2.7,all statistical calculations in this study were performed using base R (version 3.1.0).All plots were produced in either R or Microsoft Excel.30Chapter 3Results3.1 Characterization of insect hungerThis study relied upon sequencing the mRNA of Drosophila melanogaster andAedes aegypti in a food-deprived state. In order to select timepoints in each speciesthat represented a similar level of starvation, the levels of sugar present in theinsects’ whole-body tissue were quantified. This allowed identification of a timepoint in which the starvation level of each insect species was roughly equivalent.Glucose and trehalose are the primary form of stored energy in the hemolymphof most insects, and offer an excellent snapshot of starvation state in D. melanogasterand A. aegypti (Wyatt, 1961). Lending support to this approach, it has been pre-viously demonstrated that internal sugar levels are highly responsive to nutritionalstate (Wyatt, 1961).3.1.1 Determination of Drosophila starvation stateStarvation has been relatively well characterized in Drosophila melanogaster. Iflevels of glucose and trehalose in Drosophila had noticeably decreased by the timepoint the literature defines as starvation (24 hours), it would support the conclusionthat measurement of sugar levels was an accurate and effective way of character-izing hunger (Fujikawa et al., 2009; Farhadian et al., 2012). This is especiallyimportant given that the starvation sensitivity of Aedes aegypti males is relativelypoorly understood.31A BFigure 3.1: Quantification of insect sugar levels.(A) Glucose, trehalose, and total sugar levels of male Canton S D. melanogasterduring starvation (n = 8 replicates of 10 flies for each timepoint, 5 flies used forsugar measurement). (B) Sugar levels of male A. aegypti during starvation (n =4 replicates of 5 mosquitoes for each timepoint, 3 mosquitoes used per measure-ment). Error bars for both subfigures represent the standard error of the mean.Adult Drosophila melanogaster males were placed in culture vials filled with1 percent agar and a subset of experimental animals were subjected to sugar quan-tification every 8 hours until significant mortality occurred. We chose to use onlymale flies for this experiment due to the fact that sequencing was to be performedon only male flies and male mosquitoes. Female Aedes aegypti mosquitoes feed onblood as adults, and it seemed likely that this additional feeding modality wouldcomplicate an inter-species comparison with Drosophila.The results of sugar quantification during starvation in Drosophila are pre-sented in Figure 3.1A. Total sugar levels drop following starvation, decaying ina rapid manner until plateauing at the 24 hour timepoint. The levels of glucoseand trehalose declined at similar rates. This 24h timepoint was used for samplepreparation in the subsequent sequencing experiments. It is important to note that24 hours without food is a widely-accepted definition of starvation used through-out the field of Drosophila gustation, and was selected for several similar studies32examining transcriptional changes induced by starvation (Fujikawa et al., 2009;Farhadian et al., 2012). The fact that our definition of starvation based on sugarmeasurements so closely matched literature values affirmed that this methodologywould give an accurate estimate of “hunger” in A. aegypti.3.1.2 Determination of Aedes starvation stateStarvation state for adult male A. aegypti was assessed using the same methodol-ogy used for Drosophila (see Figure 3.1B). Tissue glucose and trehalose levelsdecreased in an identical manner to that observed in flies, rapidly decaying uponstarvation. Interestingly, A. aegypti are almost twice as resistant to starvation com-pared to Drosophila. Whereas almost all male flies had died within 48 hours with-out food, a number of male mosquitoes were still alive after 96 hours without food.Nevertheless, sugar levels in A. aegypti plateaued around 48 hours of starvation.The 48 hour timepoint was not chosen arbitrarily- total sugar levels of each in-sect were roughly equal to that observed in Drosophila at 24 hours (4 µg per mgtissue). This 48 hour timepoint was defined as the “starved” state used for RNAsequencing.3.2 RNA sequencing reveals conserved transcriptomicchangesAfter determining what constituted a starved state for D. melanogaster and A. ae-gypti, I sought to identify the transcriptomic changes that occurred in each animal.These insects were either fed or starved (24h for D. melanogaster, 48h for A. ae-gypti) in an identical manner to that described in Section 3.1. Several additional“fed” samples of Drosophila were also prepared in addition to glucose food (nu-tritive sugar), in which flies were fed on various food sources, including yeast(contains significant amino acids and fat), and arabinose (a non nutritive sugar).The heads of each animal were removed at the timepoints of interest and mRNAwas prepared for sequencing at these timepoints. We chose to sequence insectheads (as opposed to the whole animal), as this is an effective method to enrichfor genetic changes specific to neuronal tissues. We chose to focus on studyingneuronal tissues as these are the cells responsible for directly controlling behavior.33A BFigure 3.2: Overview of RNA sequencing results.MA plots for Drosophila melanogaster (A) and Aedes aegypti (B). The data forDrosophila shown here represent a comparison between starved and glucose-fedflies, whereas data in B are a comparison of starved and sucrose-fed mosquitoes.Each point represents the expression of a single gene. Red points are significantlyregulated, with an adjusted p-value < 0.05. Mean expression is in terms of rawread counts normalized to library size, and change in expression is in terms oflog2(starved/ f ed). Note that log2 fold-change estimates for genes with low readcount or high dispersion have been shrunk using the DESeq2 R package.An overview of these sequencing results is presented in Figure 3.2.1519 genes were significantly regulated in D. melanogaster by starvation, with891 genes upregulated and 621 genes downregulated in starved agar-fed samplesrelative to glucose-fed controls. 307 genes were significantly regulated in A. ae-gypti by starvation, with 181 genes upregulated and 126 downregulated in starvedwater-fed samples relative to sucrose-fed controls. This ran contrary to my predic-tion that the primary response in both organisms would be to downregulate genesand cellular pathways in an effort to limit protein biosynthesis and metabolism.343.2.1 Identification of conserved transcriptomic changesThe next step of our analysis sought to examine the conserved transcriptomicchanges occuring in these organisms. In order to accomplish this, we would needto develop our own novel method of analysis, raising a number of important ques-tions. How can transcriptomic changes in one organism be compared to that ofanother? The genomes of D. melanogaster and A. aegypti are not a one-to-onematch. A method would need to be developed to match gene expression changesin one species with that of another. Furthermore, once such a method had been de-veloped, I would need to establish a method of statistically validating this approach.In order to identify genes that showed conserved regulation across both species,I established the following criteria:1. There must be a biologically meaningful means of matching a gene in onespecies with that of another.2. For the genetic changes of a matched pair to be considered “significant,” thechange in expression for each gene in the pair must be significant, with anadjusted p-value < 0.05.3. The number of significantly regulated matched gene-pairs must be greaterthan that would be expected due to random upregulation/downregulation ofunrelated sets of genes.4. There should be some amount of correlation between gene expression changesin each organism.To address item 1 of this definition, we chose to use the biological concept oforthology as a means of matching genes in one organism with those of another.Orthologs are sets of genes in different species that diverged from a common an-cestor gene through the process of speciation. As a result, orthologs typically sharesignificant sequence identity and often share common molecular functions as well.Although this is a useful definition that satisfies the requirement of a “biologicallymeaningful” method of matching genes in one species with those of another, it35is not “computationally meaningful.” To elaborate, it would be impossible to ap-ply the biological definition of orthology on a genome-wide scale across multiplespecies because it does not supply strict criteria that can be evaluated by a com-puter. Because of this limitation, it was necessary to use a computational definitionof orthology instead. We chose to use the definition specified by Waterhouse et al.(2013), using information provided by the OrthoDB7 orthology database. Due tothe importance of definitions in this context, I will outline the procedure Water-house et al. used to calculate orthologs here: the protein sequence of the longesttranscript for each gene in a species was compared against those of every otherspecies in the OrthoDB7 database using the SWIPE algorithm. If two sequencesshared significant sequence identity, they were declared as orthologs and includedin a cluster of other matches (Waterhouse et al., 2013). We created the set of or-thologous gene-pairs used in this study by creating every possible pairing of D.melanogaster and A. aegypti genes from these clusters. This means that the set oforthologs used in this study is identical to those defined by OrthoDB7. A moredetailed explanation of this procedure can be found in Section 2.6.After obtaining the list of ortholog pairs for both species, we sought to findwhich ortholog pairs were significantly regulated. I declared any pair in whichthe adjusted p-values for genes from both species were below 0.05 as significant.Once this operation had been performed, I was left with 117 significantly regulatedpairs. To feel confident in using these data, I needed to determine if the significantlyregulated orthologs indicative of any biological trend, or if this was a result causedby significant genes from both species simply being regulated in the same directionby chance.3.2.2 Validation of ortholog resultsI wanted to understand if the observed number of significantly ortholog pairs dif-fered from what would be expected if gene regulation in each species were random(note that this is a separate measurement from the probability of a given gene beingdifferentially expressed calculated by DESeq2). Unfortunately, there are no formalstatistical tests capable of answering this question. The nature of orthology meansthat a gene in one species may have any number of equivalents in another species,36Figure 3.3: Orthologous gene-pairs significantly regulated by starvation.This plot contains all of the significantly regulated ortholog pairs where the ad-justed p-value for each gene in a pair’s change in expression was below 0.05.Percent identity was calculated as the percent sequence identity between theDrosophila melanogaster gene and its Aedes aegypti equivalent. The pair com-prising Jhe and its ortholog, AAEL005178 has been highlighted in magenta. Thewhite line in this figure represents a one-to-one correlation in gene expression be-tween orthologs in a pair.including none, complicating any potential analysis. However, the result expectedif our null hypothesis were true (the observed number of orthologs is equal to thenumber expected if gene regulation was random) is specific enough to be simulatedwith a computer. Although a simulation done in this manner may differ from runto run, enough iterations of the program will eventually approach the true valueof the simulated statistic. This approach to a statistical problem is referred to as aMonte Carlo method.37Our simulation capitalized on the fact that random genetic regulation in bothspecies could be calculated easily. A probability of any given gene being regulatedin a species was computed by taking the number of genes significantly regulated inthe ortholog set divided by the total number of genes for that species in the orthologset. This probability α , was used to generate a binomial distribution of the numberof significantly regulated genes in a given simulation run. During each iteration ofthe simulation, we selected a random set of X number of genes (X was determinedby randomly sampling from that species’ binomial distribution) from each speciesas “on”, or significantly regulated. The list of genes “on” in each species was thencompared to that of the other species and the number of significantly regulatedorthologs (defined as k) was determined by counting the number of matches whereboth genes in an orthologous pair were “on.” For more details on this simulation,see Section 2.7.Using the methodology described above, we calculated the values of α for A.aegypti and D. melanogaster as 0.0198 and 0.1425, respectively. Running onemillion iterations of the simulation resulted in a mean k of 37.65. Remarkably, thehighest observed value of k observed was 83. This indicates that the probabilityof observing 117 orthologous pairs due to random genetic regulation is literallyless than one in a million. The full results of the simulation are summarized inFigure 3.4.Although the number of ortholog pairs was significantly greater than whatwould be expected, testing for correlation in gene expression in A. aegypti andD. melanogaster would add an additional level of validation to our hypothesis thatthe gene expression in the two species is conserved. Gene expression among thesignificantly-regulated orthologous pairs was positively correlated with a Pearson’sρ product-moment correlation coefficient of 0.40 (p< 0.001). Taken together withour simulation results, the data demonstrate that many changes in gene expressioninduced by hunger are conserved between A. aegypti and D. melanogaster.3.2.3 Jhe is regulated significantly by hungerI examined the list of conserved ortholog pairs between A. aegypti and D. melanogasterfor candidate genes likely to be capable of controlling behavior, with a focus on38A10 20 30 40 50 60 70 800.000.010.020.030.040.05kApproximation of P(ConservedNum = k)B20 40 60 800.00.20.40.60.81.0kApproximation of P(ConservedNum <= k)●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●Figure 3.4: Ortholog simulation results.(A) Probability mass function of the orthologous pair statistical simulation de-scribed in Section 2.7, using α values of 0.0198 and 0.1425 for A. aegypti and D.melanogaster across one million iterations. (B) Cumulative distribution functionof the same simulation results from A.anything capable of potentially affecting neuronal activity. Interestingly, Juve-nile Hormone esterase (Jhe) was significantly downregulated in flies, with a log2fold-change of -0.84. One of Jhe’s Aedes aegypti orthologs, AAEL005178, wassignificantly regulated as well, with a log2 fold-change of 1.01. Although thesegenes are regulated in opposite directions in each insect, Jhe is known to be in-volved in hormone metabolism, and has been shown to degrade the insect hor-mone Juvenile Hormone (JH) in vitro (Crone et al., 2007). Bioinformatic analysesindicates that AAEL005178 is a type-B carboxylesterase/lipase, a shared featurewith Jhe. Additionally, AAEL005178 and Jhe cluster together in OrthoDB7 or-tholog group EOG7T22TR with other annotated Juvenile Hormone esterases fromthe mosquitoes Anopheles gambiae and Culex quinquefasciatus (Waterhouse et al.,2013). Given this information, it seems likely that AAEL005178 also encodes aJuvenile Hormone esterase.In addition to my analysis of the conserved changes between A. aegypti and D.melanogaster, I chose to examine the genetic changes caused by multiple forms39of nutrient deprivation in Drosophila (Figure 3.5). Intriguingly, two JH-relatedgenes were downregulated in the comparison between both the glucose/agar-fedand yeast/agar-fed conditions: Jhe and Takeout (To). Takeout is currently thoughtto be a Juvenile Hormone binding protein, protecting JH from degradation as it cir-culates throughout an insect’s hemolymph (Sarov-blat et al., 2000; Meunier et al.,2007). Taken together with the sequencing data from A. aegypti, it seemed likelythat Juvenile Hormone esterases and Juvenile Hormone played some kind of rolein the response to starvation.3.3 Jhe is necessary for proper feeding behaviorAlthough our RNA-Seq data identified Juvenile Hormone signaling as of poten-tial interest, it did not establish whether or not the pathway had any effect on D.melanogaster or A. aegypti adults. Juvenile Hormone is primarily thought to be aninsect developmental hormone. Does it affect adults as well? To answer this ques-tion, I chose to study Jhe using D. melanogaster as a model. As I was interestedin understanding if Jhe played a role as a regulator of behavior, I decided to knockdown expression of the gene using a pan-neuronal RNAi approach. Since Jhe wasdownregulated by starvation, we postulated that the gene may have a role in feed-ing behavior. Jhe RNAi flies and controls were assessed for food consumptionphenotypes using the CAFE assay.Interestingly, pan-neuronal Jhe RNAi flies consumed significantly more foodthan controls when they were fed 50 mM glucose (Figure 3.6). This phenotype wasreplicated regardless of what sugars the flies were presented with. It is important tonote that this occurred regardless of a sugar’s nutritional value, as Jhe RNAi flieseven consumed more than controls when presented with the non-nutritive sugarsL-fucose and arabinose. This indicates that Jhe does not affect caloric sensing offood- it affects consumption of all sugars equally. Additionally, Jhe RNAi fliesstill ate more food when presented with mannose, which is a relatively unpalatablenutritive sugar (it is typically necessary to combine mannose with a more palatablesugar like sucrose before flies will eat it readily). Taken together, these results seemto suggest that Jhe knockdown triggers a starvation-like state in Drosophila.Given that JH serves as an important regulator of development, one poten-40Figure 3.5: JH-related genes are downregulated in multiple forms of nutrientdeprivation in Drosophila.Significantly-regulated genes between glucose and agar feeding (x-axis) have beenplotted against those genes regulated between yeast and agar feeding (y-axis). Tobe included on this plot, genes must have been significantly regulated in both con-ditions, with an adjusted p-value less than 0.05. The white line in this figure rep-resents a one-to-one correlation in a gene’s expression in each condition. Juvenilehormone esterase (Jhe) and Takeout (to), two genes implicated in Juvenile Hor-mone signalling, have been highlighted in orange to aid visibility.41Mannose Arabinose Glucose L−Fucose 10s40m Sucrose051015Food consumption (uL)Genotypeelav>Jhe RNAi+>Jhe RNAielav>+*************Figure 3.6: Jhe is a modulator of feeding behavior.Jhe RNAi flies’ consumption was characterized using the CAFE assay on a varietyof food sources. Neuronal Jhe RNAi flies ate significantly more than controlsregardless of which sugar they were fed. All sugars tested were at a concentrationof 50 mM. 10s40m comprises a mixture of 10 mM sucrose and 40 mM mannose.Error bars represent standard error of the mean. n = 5-7 replicates of 8 flies pergenotype for each sugar. Asterisks indicate a significant difference from controlsas measured by One-Way ANOVA followed by post-hoc Tukey HSD test (* = p< 0.05, ** = p < 0.01, *** = p < 0.001).tial concern is that the observed effect may be of developmental origin. To ex-clude this possibility, we performed another experiment, where Jhe knockdownwas restricted to Drosophila’s adult stage using the temperature-sensitive isoformof the GAL4-repressor GAL80ts. Importantly, when Jhe knockdown was restrictedto adults, the feeding phenotype was reproduced (Figure 3.7). Moreover, whenGAL4-mediated RNAi was turned off at room temperature, there was no differ-ence between the RNAi flies and their isogenic controls, indicating that this ap-proach was effective in restricting RNAi expression to the adult stage (Figure 3.7).These results indicate that Jhe’s feeding phenotype is not due to a developmentaldefect. It is also important to note that this experiment utilized a different neuronalGAL4 driver, nysb. The fact that the same phenotype was reproduced using twoseparate neuronal drivers indicates that Jhe is indeed expressed in neurons, and the4229C 20C0.02.55.07.510.012.5Food consumption (uL)GenotypeDcr;Gal80ts;nsyb>Jhe RNAi+>Jhe RNAiDcr;Gal80ts;nsyb>+***nsFigure 3.7: Jhe’s phenotype is not a developmental defect.Food consumption of Jhe RNAi flies when knockdown was induced in adults at29C, or was never induced at room temperature (20C). Flies in both experimentswere presented with 50 mM glucose as a food source. Error bars represent thestandard error of the mean. Consumption was measured at either 12 hours after ex-periment start (29C) or 30 hours after experiment start (20C). Error bars representstandard error of the mean. n = 5-6 replicates of 8 flies per genotype/condition.Asterisks indicate a significant difference from controls as measured by Two-WayANOVA followed by post-hoc Tukey HSD test (*** = p < 0.001, ns = not signifi-cant).phenotype is not an artifact of “leaky” GAL4 expression in other, non-neuronaltissues.Finally, it was critical to exclude the possibility that the phenotype arising fromJhe knockdown were a result of RNAi off-target effects. To do this, pan-neuronalknockdown of Jhe expression was performed with a new RNAi line. The secondRNAi line targeted a separate region of the Jhe transcript relative to the RNAi lineused in Figure 3.6. This resulted in an identical phenotype to that demonstratedearlier: increased consumption of food (Figure 3.8). From the data shown here, itseems clear that Jhe may function as a novel regulator of hunger-induced feedingbehavior in insects. RNAi knockdown of Jhe is consistent with the downregulationthat this gene normally undergoes during starvation in Drosophila (see Figure 3.5).430.02.55.07.510.012.515.0elav>Jhe RNAi #2+>Jhe RNAi #2elav>+Food consumption (uL)***Figure 3.8: Jhe’s phenotype is not an off-target effect.Food consumption of Jhe RNAi #2 flies and controls (GAL4/+ and RNAi/+) quan-tified by CAFE assay. Flies were fed 50 mM glucose for 12 hours, after whichtheir consumption was measured. Error bars represent standard error of the mean.n = 9-13 replicates of 8 flies each. *** = p < 0.001 by One-Way ANOVA withpost-hoc Tukey HSD.3.4 Jhe knockdown increases starvation-induced sleepsuppressionIncreased feeding is not the only behavioral change an insect will make upon star-vation. The starvation response in insects includes a wide-range of behaviors,including starvation-induced sleep suppression. Sleep suppression is a relativelywell-characterized phenotype where hungry Drosophila will avoid sleeping in or-der to forage for food (Keene et al., 2010). The test to examine this phenotype issimple: adult Drosophila are placed on food for 24 hours, starved for 24 hours, andthen allowed to recover with food for a final 24 hours. During the 24 hours in whichflies are deprived of food, they will suppress sleep. The amount of locomotor ac-44tivity and sleep can be quantified using Trikinetics’ Drosophila Activity Monitorand analyzed in my actmon R package (for details of how sleep calculations areperformed with actmon, see Section 2.12.When Jhe knockdown flies were assessed for sleep suppression, they demon-strated increased activity relative to controls during the starvation period (Fig-ure 3.9A). This increased activity resulted in knockdown flies sleeping signifi-cantly less than controls (Figure 3.10). This sleep suppression effect was identicalto that described by Keene et al. (2010), where this sleep suppression only oc-curred during the second 12 hour period of starvation. This effect has previouslybeen shown to occur during this second 12 hour period, irrespective of light or darkstatus (Keene et al., 2010).One question this phenotype raised was whether this effects on sleep couldsimply be attributed to increased locomotor activity. This was assessed by ex-amining the activity index of flies during the assay, a measure of activity countsnormalized to the amount of time an animal spends awake. Hyperactive flies willhave a higher activity index than controls, and hypoactive flies with a lower ac-tivity index may indicate some form of locomotor defect. Jhe RNAi flies showedan identical activity index to controls, indicating that locomotor activity was unaf-fected by the knockdown (Figure 3.9B). Jhe’s activity phenotype is therefor likelya direct result of those flies sleeping less than controls. This held true regard-less of whether or not males or females were tested, indicating that this is not asex-specific phenotype. Taken together, these results indicate that Jhe knockdownincreases starvation-induced sleep suppression. This supports the hypothesis thatJhe regulation is a driver of starvation-induced behavior. Neuronal knockdown ofJhe causes flies to both increase feeding and suppress sleep, indicating that down-regulation of this gene during starvation may be what causes these phenotypes inwild-type flies.3.5 Jhe knockdown does not increase starvationsensitivityNeuronal Jhe downregulation results in a starvation-like behavioral state, increas-ing both feeding and starvation-induced sleep suppression. The most obvious ex-45A0501001502000 12 24 36 48 60Time (hours)Line breaks / hourelav>Jhe RNAi+>Jhe RNAielav>+B012elav>Jhe RNAi+>Jhe RNAielav>+Total line breaks / time awake (minutes)nsFigure 3.9: Jhe knockdown alters Drosophila activity.(A) A measurement of raw Drosophila activity, as measured in total line break-s/hour. Flies were deprived of food between hours 24 and 48. Light or dark regionsrepresent whether experimental lighting was on or off. (B) Average activity index(total activity normalized to time awake) for each genotype for the time period be-tween 36 and 48 hours from the plot in A. Error bars in B and colored regions inA represent the standard error of the mean for each data point. n = 30 - 36 flies pergenotype. ns = not signifcant by One-Way ANOVA followed by post-hoc TukeyHSD test. 46A02550750 12 24 36 48 60Time (hoursPercent of time asleep / hourelav>Jhe RNAi+>Jhe RNAielav>+B36 480100200300400elav>Jhe RNAi+>Jhe RNAielav>+elav>Jhe RNAi+>Jhe RNAielav>+Total sleep (minutes)ns ***Figure 3.10: Jhe knockdown increases starvation-induced sleep suppression.(A) A measurement of Drosophila sleep, in percent of time asleep per hour. Flieswere deprived of food between hours 24 and 48. Light or dark regions representwhether experimental lighting was on or off. (B) Total sleep for each genotypefor the time period between either 24 and 36 hours (36), or 36 and 48 hours (48)from the plot in A. Error bars in B and colored regions in A represent the standarderror of the mean for each data point. n = 30 - 36 flies per genotype. *** indicatessignificance versus controls with p< 0.001 by One-Way ANOVA followed by post-hoc Tukey HSD test, ns = not significant.47planation for these phenotypes is that Jhe downregulation may actually cause star-vation itself. If this were the case, and Jhe knockdown causes flies to becomeenergetically starved (instead of just acting hungry), Jhe RNAi flies would haveincreased starvation sensitivity relative to controls, dying quickly when deprivedof nutrients. To test this hypothesis, starvation sensitivity was assessed using theDrosophila Activity Monitor. RNAi flies and controls were placed on agar food andsurvival time was assessed using the actmonR package. The starvation sensitivityof male and female flies was quantified separately, as the sexes differ dramaticallyin both body size and starvation tolerance.Jhe RNAi flies did not exhibit differential starvation sensitivity relative to con-trols, indicating that neuronal downregulation of this gene does not affect an an-imal’s energetic state (Figure 3.11). This held true for both male and femaleDrosophila, again indicating that Jhe’s effects are not sex-specific. The only re-maining explanation for Jhe’s phenotypes was that this gene was directly acting tocontrol behavior, instead of indirectly causing a behavioral change by starving ananimal.3.6 Methoprene feeding results in a Jhe-like phenotypeFrom the evidence discussed so far, it is clear that Jhe knockdown during starvationacts to induce hunger-related behaviors. However, the mechanism through whichJhe induces these behavioral changes remained unknown. To tackle this question,we chose to study the most obvious candidate: the insect hormone Juvenile Hor-mone III. Juvenile Hormone esterase has been previously shown to degrade a widenumber of JH derivative compounds in vitro, and it has been postulated that Jheis the only esterase likely to degrade this hormone in vivo (Campbell et al., 1998;Crone et al., 2007). Because of this evidence, it seemed likely that Jhe exerts itsbehavioral effects by degrading JH. According to this hypothesis, downregulationof Jhe during starvation would result in increased JH levels, which would triggerbehavioral changes through activation of the MET-GCE heterodimeric receptor orUSP. To test this, we would need some way of artificially increasing flies’ JH levels.As mentioned earlier in Section 1.5.1, the JH synthesis pathway in Drosophilais poorly understood. As a result, we chose to artificially increase JH signaling48A020406080elav>Jhe RNAi+>Jhe RNAielav>+Survival time (hours)nsB010203040elav>Jhe RNAi+>Jhe RNAielav>+Survival time (hours)nsFigure 3.11: Jhe knockdown does not affect starvation sensitivity.Female (A) and male (B) Drosophila starvation sensitivity was assessed by measur-ing survival time of flies during absolute starvation. Survival time was measured inhours. No significant difference was observed between knockdown flies and con-trols. Error bars represent the standard error of the mean for each data point. n = 32flies per condition. ns = not signifcant by One-Way ANOVA followed by post-hocTukey HSD test.through the use of methoprene. Methoprene is a synthetic analog of Juvenile Hor-mone that binds to JH receptors with high affinity and specificity and cannot bedegraded by insect Juvenile Hormone esterases and Juvenile Hormone epoxide hy-drolases. This chemical is effective when ingested orally, and is commonly usedas an insecticide due to its activation of the JH signaling pathway (Organization,2008). Because of these factors, feeding flies methoprene seemed like a perfectmechanism of simulating Jhe downregulation by increasing JH signaling.Adult Drosophila were fed standard cornmeal food containing either metho-49A040801201600 12 24 36 48 60Time (hours)Total line breaksVehicleMethopreneB0.00.51.01.52.0Vehicle MethopreneTotal line breaks / time awake (minutes)nsFigure 3.12: Methoprene alters Drosophila activity in a Jhe-like manner.(A) A measurement of raw Drosophila activity when fed methoprene, as measuredin total line breaks/hour. Flies were deprived of food between hours 24 and 48.Light or dark regions represent whether experimental lighting was on or off. (B)Average activity index (total activity normalized to time awake) for each genotypefor the time period between 36 and 48 hours from the plot in A. Error bars in B andcolored regions in A represent the standard error of the mean for each data point.n = 32 flies per condition. ns = not signifcant by One-Way ANOVA followed bypost-hoc Tukey HSD test.50prene or ethanol (the vehicle used to dissolve methoprene) for two days beforemeasuring their behavior. Methoprene/vehicle feeding was continued during theseexperiments. Methoprene feeding increased starvation-induced sleep sensitivity ina Jhe-like manner. Methoprene-fed flies slept significantly less than vehicle-fedcontrols when starved (Figure 3.13). Although methoprene increased total activityof flies during the same period, the activity index of flies was unaffected (Fig-ure 3.12). As before, the phenotype was identical to that described by Keene et al.(2010), where sleep suppression occurred during the second 12 hours of the starva-tion period. This is identical to the phenotype observed during Jhe knockdown- anincrease in starvation induced sleep suppression without affecting the total amountof activity when animals were awake.3.7 Precocene I rescues Jhe knockdownThe results discussed in Section 3.6 indicated that increasing JH titers reproducedthe effects of Jhe knockdown. By the same reasoning, if JHE exerted its behav-ioral effects by degrading JHs, then artificially decreasing JH titers should “rescue”the phenotype produced by Jhe knockdown. To test this possibility, we fed Jheknockdown flies the anti-juvenoid agent precocene I. Precocene I is a relativelywell-characterized drug capable of blocking JH synthesis. This drug acts upon thecorpora allata directly, decreasing its secretory activity (Wilson et al., 1983).As demonstrated in Figure 3.14, precocene I feeding rescues the effects of Jheknockdown without affecting normal waking activity. Precocene I was specifi-cally able to rescue the increase in sleep suppression of Jhe knockdown flies. Thisphenomenon was not sex-specific, and further strengthens the conclusion that JHEexerts its effects through JH.51A0204060800 12 24 36 48 60Time (hours)Percent of time asleep / hourVehicleMethopreneB36 480100200Vehicle Methoprene Vehicle MethopreneTotal sleep during 12 hr period (minutes)**nsFigure 3.13: Methoprene increases starvation-induced sleep suppression.(A) A measurement of Drosophila sleep when fed methoprene, in percent of timeasleep per hour. Light or dark regions represent whether experimental lighting wason or off. Flies were deprived of food between hours 24 and 48. (B) Total sleep foreach genotype for the time period between either 24 and 36 hours (36), or 36 and48 hours (48) from the plot in A. Error bars in B and colored regions in A representthe standard error of the mean for each data point. n = 32 flies per condition. **indicates significance versus controls with p< 0.01 by One-Way ANOVA followedby post-hoc Tukey HSD test, ns = not significant.52A02550750 12 24 36 48 60Time (hours)Percent of time asleep / hourelav>++>Jhe RNAielav>Jhe RNAielav>Jhe RNAi + PrecoceneB36 480100200300400500elav>++>Jhe RNAielav>Jhe RNAielav>Jhe RNAi + Precoceneelav>++>Jhe RNAielav>Jhe RNAielav>Jhe RNAi + PrecoceneTotal sleep (minutes)nsnsns****C0.00.51.01.52.0elav>++>Jhe RNAielav>Jhe RNAielav>Jhe RNAi + PrecoceneActivity IndexnsnsFigure 3.14: Precocene I rescues Jhe knockdown.(A) A measurement of Drosophila sleep when fed precocene I, in percent of timeasleep per hour. Light or dark regions represent whether experimental lighting wason or off. Flies were deprived of food between hours 24 and 48. (B) Total sleepfor each genotype for the time period between either 24 and 36 hours (36), or 36and 48 hours (48) from the plot in A. (C) Average activity index (total activitynormalized to time awake) for each condition for the time period between 36 and48 hours from the plot in A. Error bars in B/C and colored regions in A representthe standard error of the mean for each data point. n = 24 flies per condition. *indicates significance versus controls with p< 0.05 by One-Way ANOVA followedby post-hoc Tukey HSD test, *** = p < 0.001, ns = not significant.53Chapter 4DiscussionThe original goal of this study was to identify a gene capable of controlling the be-haviors associated with starvation. Jhe is a relatively unstudied gene not previouslyknown to control behavior. What’s more, Jhe’s involvement in Juvenile Hormonemetabolism suggests that these hormones play an important and novel role in theadult lifestage. Although this hormone is known to be present in the adult, no rolefor the hormone has been suggested aside from supporting oogenesis in females(Wilson et al., 1983; Shiao et al., 2008; Parthasarathy and Palli, 2011). I proposethat Jhe regulation during starvation acts to coordinate changes in food intake andsleep. This occurs through its effects on Juvenile Hormone. The evidence for theseassertions will be discussed here.4.1 Regulation of Jhe is evolutionarily conservedI originally identified Jhe as a conserved regulatory change using RNA sequenc-ing in two separate, evolutionarily divergent insects. As discussed in Section 2.6,we used the definition of orthology established by Waterhouse et al. to match thegenomes of D. melanogaster and A. aegypti at the gene-to-gene level. This ap-proach was validated via statistical simulation- I obtained far more significantlyregulated orthologous gene-pairs than would be expected if no conservation ofregulatory changes had occurred. The correlation in gene expression between or-thologs in each species was greater than zero (see Section 3.2.2). However, there54are two caveats with the form of statistical analysis used here. The statistical modelpurely predicts the total number of conserved changes expected if gene regulationwas absolutely random in each species. As a result, the probability of each k valueindicates the probability of obtaining a k value as extreme or more extreme underthe model and cannot be used to validate the significance of any given orthologouspair (Jewell, 2014). Furthermore, calculating the given probability for individualorthologous pairs was deemed impractical, as calculating a specific p-value for ev-ery pair would require the calculation of probabilities for every possible model ofconserved changes. No methodology exists with which to make these calculations(Jewell, 2014). With these caveats in mind, I am confident in the conclusion thatgene regulatory changes are conserved during starvation between D. melanogasterand A. aegypti. Although we cannot measure the statistical significance of any in-dividual orthogous conserved pair, Jhe is an excellent candidate for being a hunger-regulated gene in each species, based on the observed phenotypes in each organismas well as data from the literature.Although both Jhe and its ortholog AAEL005178 are significantly regulated bystarvation, they are regulated in opposite directions. Whereas Jhe expression de-creases (with a log2 fold-change of -0.84), the expression of its ortholog AAEL005178,actually increases, with a log2 fold-change of 1.01. Although initially perplexing,this difference in expression makes more sense with additional evidence. Bothgenes are predicted to function Juvenile Hormone esterases. JHE has been demon-strated to rapidly degrade all forms of JH, including JH III, JHB3, and MF (Camp-bell et al., 1998). If the expression of Jhe increases, JH titers are expected to de-crease. If, on the other hand, Jhe expression decreases, JH titers will increase. Pre-vious evidence indicates that the JH titers decrease in starved Aedes aegypit (Shiaoet al., 2008; Hernandez-Martinez et al., 2015). This is consistent with an increasein expression of Juvenile Hormone esterase-type genes. Although we still lack ev-idence for an increase in JH titers upon starvation in Drosophila, the expression oftakeout is consistent with an increase in Juvenile Hormone levels- it is downreg-ulated when JH levels are high (Goodman and Cusson, 2012). Additionally, wefound that exogenous application of methoprene phenocopied Jhe downregulation,demonstrating that an increase in JH signaling was consistent with our model ofJhe action (see Section 3.6). As mentioned in Section 4.4, I am currently working55on developing a method of definitively quantifying Juvenile Hormone titers duringstarvation in Drosophila.4.2 Jhe affects sleep and feeding in adult DrosophilaJhe is able to control behavior in adult Drosophila. Knockdown of Jhe resultedin increased feeding, and this was not due to a developmental defect (Figures 3.6and 3.7). Interestingly, we found that this phenotype was not related to appetitivetaste or caloric sensing of food, as consumption increased regardless of the food’spalatability or nutritional value (Figure 3.6). At the same time, Jhe knockdownalso caused an increase in starvation induced sleep suppression without affectingwaking levels of activity (Figures 3.10 and 3.9B). These data are consistent with thebehavioral and transcriptional changes that occur during starvation, as knockdownof Jhe reduces expression of the gene in a similar manner observed during hunger.Both sets of experiments relied upon neuronal knockdown of Jhe. Two differentneuronal GAL4 drivers were used, indicating both that Jhe is expressed in neurons,as well as the conclusion that reducing the amount of Jhe expression in neurons issufficient to elicit a behavioral change. Jhe has not been previously shown to act inthis tissue.4.3 Jhe exerts its effects through Juvenile HormonemetabolismJHE is known to degrade all forms of JH in Drosophila as well as other organisms(Campbell et al., 1998; Goodman and Cusson, 2012). Additionally, JHE plays agreater role in JH degradation in adult Drosophila than the JHEH family of genes(as stated in Section 1.5.2.3, JHEHs are the other major pathway responsible forJH degradation) (Rauschenbach et al., 1995). Although JHEH’s are more highlyexpressed, they are unable to degrade JHB3 with the same efficiency as JHE andcannot degrade MF (Casas et al., 1991). JHB3 and MF are the two most abundantJHs in Drosophila (Wen et al., 2015). All three Drosophila JHs are biologicallyactive, and can bind and activate the known JH receptors (Shemshedini et al., 1990;Jones and Sharp, 1997; Godlewski et al., 2006; Wen et al., 2015). As mentionedbefore in Section 4.1, the most likely explanation for Jhe’s behavioral phenotypes56is through its role in JH catabolism.In Section 3.6, we demonstrate that methoprene is able to phenocopy the effectsof Jhe knockdown. As mentioned before, methoprene is a highly specific and ef-fective JH agonist able to bind and activate both the USP and MET-GCE receptors(Wilson and Fabian, 1986; Jones and Sharp, 1997). This means that methoprene isan effective method of simulating an increase in JH titers. Methoprene applicationhas a number of advantages over other methods of JH application. Unlike JH IIIor other biological JH derivatives, methoprene is extremely stable and resistant todegradation, as it lacks the functional groups that JHE and JHEH act upon. This isespecially important in light of the fact that addition of JH III has been shown toincrease levels of proteins able to catabolize it (including JHE), possibly bluntingany effect JH application might have (Kethidi et al., 2005). Nevertheless, whenmethoprene was added to wild-type flies, it resulted in an identical sleep pheno-type observed during Jhe knockdown (Figure 3.13). This indicates that the resultof neuronal Jhe knockdown is consistent with the predicted increase in JH titers thiswould generate. Even more convincingly, addition of precocene I, an anti-juvenoidagent, rescued the the phenotype caused by Jhe knockdown. Taken together, thesedata indicate that sleep suppression is controlled by JH levels in adult Drosophila.Although it is possible that the effects of Jhe and JH may be transduced throughtwo parallel pathways that have the same phenotype, this is extremely unlikely. Ithas been demonstrated repeatedly that JHE makes a major contribution to the con-trol of JH titers in Drosophila adults (Campbell et al., 1992; Rauschenbach et al.,1995).4.4 Future work and directionsThere are two key experiments that can be used to further reinforce the findings ofthis study. The first is a measurement of Juvenile Hormone titers in Drosophila.Although all evidence so far indicates that neuronal Jhe is controlling behavior bymanipulating JH titers, we have so far not been able to show this directly. If weare able to demonstrate that hemolymph JH titers increase in starved versus fedDrosophila as well as show the same effect in fed neuronal Jhe knockdown fliesversus controls, it will demonstrate that Jhe is manipulating JH levels. The alterna-57tive result (no change in hemolymph JH levels upon Jhe knockdown or starvation)is equally interesting, as it would imply that Jhe may be regulating JH activity onthe local scale, either in or around JH target cells (rather than on a global scale,by changing hemolymph JH levels). Either way, this experiment would prove in-formative. I have made a large amount of progress towards this goal, and haveprepared hemolymph extracts from all genotypes/conditions involved in the exper-iment. This extraction has also been confirmed to be effective, and these prepa-rations contain detectable levels of JHs. Currently the only factor preventing JHquantification is the availability of an analytical standard for JHB3. After optingto quantify JH through multiple reaction monitoring on an LC-MS mass spectrom-eter, it proved impossible to develop a quantitative assay for JHB3- we have beenunable to predict JHB3’s breakdown products in the absence of a standard. Un-fortunately, there seems to be no means of acquiring JHB3 either commercially orotherwise, so this experiment is stalled indefinitely until I am able to synthesize thecompound.The other experimental goal deals with identification of the JH targets able toeffect changes in behavior. Although we have shown that Jhe is expressed in neu-rons, this gene’s protein product is known to be secreted into the hemolymph. Thismeans that JHE may act locally (through intracellular JH degradation), globally(by manipulation of hemolymph JH titers), or a combination of both. Despite thisambiguity, there are two excellent candidates for JH’s behavioral targets.The most obvious potential target of JH action is neurons. As discussed inSection 1.5.4, JH has previously been shown to act directly on neurons. This effectis rapid, inducing short-term neuronal depression in as little as 2 minutes (Richterand Gronert, 1999). Although this is an exciting possibility, the behavioral assaysused in this study do not have the temporal resolution required to resolve effects onthis timescale.Another possible target of JH action may be the fat body. The fat body isroughly equivalent to mammalian adipose tissue, and has previously been shownto both act as a nutrient sensor and is capable of signalling to the brain to effectbehavior (Rajan and Perrimon, 2012). In times of starvation, the fat body inducesinsulin-like peptide release by signalling to the brain through the Drosophila leptin,upd2 (Rajan and Perrimon, 2012). Insulin signalling has previously been shown58to affect feeding behavior, and this action occurs on a similar timescale to thatobserved in this study (Stafford et al., 2012). The fat body has also previouslybeen shown to respond to JH levels, making this tissue a particularly attractivepossible target of JH action (Shiao et al., 2008; Parthasarathy and Palli, 2011).4.5 Final remarksAs stated previously, the goal of this study was to identify and characterize a novelregulator of hunger-induced behavior. I believe that this goal has been accom-plished. Jhe is able to control feeding and sleep suppression through its effectson Juvenile Hormone titers. In addition, this study demonstrated that JuvenileHormone levels are able to control behavior in the adult, and this occurs in aphysiologically-relevant manner. JHs were not previously known to act in thismanner. Jhe downregulation in starved Drosophila likely increases JH levels, in-ducing behaviors that should help hungry flies survive like increased food con-sumption. Juvenile Hormone esterase regulation also occurs in starved Aedes ae-gypti, raising the possibility that this is an important regulatory change that occursin a large number of insect species upon starvation.These conclusions have a number of economic implications, with immediateapplication to modern pest control. Methoprene is widely used to control mosquitopopulations around the world due to its high specificity and low toxicity. Givenour finding that methoprene (and JHs in general) can control behavior, it is worthexamining the effects of JH feeding on economically and environmentally criti-cal species like honeybees. If methoprene deployment proves disruptive to thesespecies, it may be worth reconsidering the use of Juvenile Hormone-based insecti-cides in regions where protection of these species is paramount.59BibliographyAnders, S., Pyl, P. T., and Huber, W. (2014). HTSeq - A Python framework towork with high-throughput sequencing data. Bioinformatics (Oxford, England),31(2):166–169. → pages 25Andrews, P., Bokil, H., Kaur, S., Loader, C., Maniar, H., Mehta, S., Mitra, P. P.,Nalatore, H., Yadav, R., and Shukla, R. (2015). Chronux. → pages 72Arrese, E. and Soulages, J. (2010). Insect Fat Body: Energy, Metabolism, andRegulation. Annual review of entomology, 55:207–225. → pages 20, 21Baumann, A., Barry, J., Wang, S., Fujiwara, Y., and Wilson, T. G. (2010).Paralogous genes involved in juvenile hormone action in Drosophilamelanogaster. Genetics, 185(4):1327–36. → pages 19Bernardo, T. J. and Dubrovsky, E. B. (2012). Molecular Mechanisms ofTranscription Activation by Juvenile Hormone: A Critical Role for bHLH-PASand Nuclear Receptor Proteins. Insects, 3(4):324–338. → pages 11Bonning, B. C., Booth, T. F., and Hammock, B. D. (1997). Mechanistic Studies ofthe Degradation of Juvenile Hormone Esterase in Manduca sexta. Archives ofInsect Biochemistry and Physiology, 286(91):275–286. → pages 17Bowers, W. S., Ohta, T., Cleere, J. S., and Marsella, P. A. (1976). Discovery ofAnti-Juvenile Hormones in Plants. Science, 193(4253):542–547. → pages 18Brand, A. H. and Perrimon, N. (1993). Targeted gene expression as a means ofaltering cell fates and generating dominant phenotypes. Development,415:401–415. → pages 6, 7Campbell, P. M., Healy, M. J., and Oakeshott, J. G. (1992). Characterization ofJuvenile Hormone Esterase in Drosophila melanogaster. Insect Biochemistryand Molecular Biology, 22(7):665–677. → pages 17, 5760Campbell, P. M., Oakeshott, J. G., and Healy, M. J. (1998). Purification andkinetic characterisation of juvenile hormone esterase from Drosophilamelanogaster. Insect Biochemistry and Molecular Biology, 28(7):501–515. →pages 16, 17, 48, 55, 56Casas, J., Harshman, L. G., Messeguer, A., Kuwano, E., and Hammock, B. D.(1991). In Vitro Metabolism of Juvenile Hormone Ill and Juvenile Hormone IllBisepoxide by Drosophila melanogaster and Mammalian Cytosolic EpoxideHydrolase. Archives of Biochemistry and Biophysics, 286(1):153–158. →pages 17, 56Celniker, S. E., Dillon, L. A. L., Gerstein, M. B., Gunsalus, K. C., Henikoff, S.,Karpen, G. H., Kellis, M., Lai, E. C., Lieb, J. D., Macalpine, D. M., Micklem,G., Piano, F., Snyder, M., Stein, L., White, K. P., and Waterston, R. H. (2009).Unlocking the secrets of the genome. Nature, 459(June):927–930. → pages 16,17Chamseddin, K. H., Khan, S. Q., Nguyen, M. L. H., Antosh, M., Morris, S. N. S.,Kolli, S., Neretti, N., Helfand, S. L., and Bauer, J. H. (2012). takeout-dependentlongevity is associated with altered Juvenile Hormone signaling. Mechanismsof ageing and development, 133(11-12):637–46. → pages 15Chu, B., Chui, V., Mann, K., and Gordon, M. D. (2014). Presynaptic Gain ControlDrives Sweet and Bitter Taste Integration in Drosophila. Current Biology,24(17):1978–1984. → pages 2Crone, E. J., Sutherland, T. D., Campbell, P. M., Coppin, C. W., Russell, R. J., andOakeshott, J. G. (2007). Only one esterase of Drosophila melanogaster is likelyto degrade juvenile hormone in vivo. Insect biochemistry and molecularbiology, 37(6):540–9. → pages 16, 39, 48dos Santos, G., Schroeder, A. J., Goodman, J. L., Strelets, V. B., Crosby, M. A.,Thurmond, J., Emmert, D. B., and Gelbart, W. M. (2015). FlyBase:introduction of the Drosophila melanogaster Release 6 reference genomeassembly and large-scale migration of genome annotations. Nucleic AcidsResearch, 43(D1):D690–D697. → pages 6Dubrovsky, E. B., Dubrovskaya, V. A., Bernardo, T., Otte, V., Difilippo, R., andBryan, H. (2011). The Drosophila FTZ-F1 Nuclear Receptor Mediates JuvenileHormone Activation of E75A Gene Expression through an IntracellularPathway. Journal of Biological Chemistry, 286(38):33689–33700. → pages 1961Duffy, J. B. (2002). GAL4 system in Drosophila: a fly geneticist’s Swiss armyknife. Genesis (New York, N.Y. : 2000), 34(1-2):1–15. → pages 7Durinck, S., Moreau, Y., Kasprzyk, A., Davis, S., De Moor, B., Brazma, A., andHuber, W. (2005). BioMart and Bioconductor: a powerful link betweenbiological databases and microarray data analysis. Bioinformatics (Oxford,England), 21(16):3439–40. → pages 6, 25Dus, M., Ai, M., and Suh, G. S. B. (2013). Taste-independent nutrient selection ismediated by a brain-specific Na(+)/solute co-transporter in Drosophila. Natureneuroscience, 16(5):526–8. → pages 2Farhadian, S. F., Sua´rez-Farin˜as, M., Cho, C. E., Pellegrino, M., and Vosshall,L. B. (2012). Post-fasting olfactory, transcriptional, and feeding responses inDrosophila. Physiology & behavior, 105(2):544–53. → pages 2, 3, 31, 33Flicek, P., Amode, M. R., Barrell, D., Beal, K., Billis, K., Brent, S.,Carvalho-Silva, D., Clapham, P., Coates, G., Fitzgerald, S., Gil, L., Giron,C. G., Gordon, L., Hourlier, T., Hunt, S., Johnson, N., Juettemann, T., Kahari,A. K., Keenan, S., Kulesha, E., Martin, F. J., Maurel, T., McLaren, W. M.,Murphy, D. N., Nag, R., Overduin, B., Pignatelli, M., Pritchard, B., Pritchard,E., Riat, H. S., Ruffier, M., Sheppard, D., Taylor, K., Thormann, A., Trevanion,S. J., Vullo, A., Wilder, S. P., Wilson, M., Zadissa, A., Aken, B. L., Birney, E.,Cunningham, F., Harrow, J., Herrero, J., Hubbard, T. J. P., Kinsella, R.,Muffato, M., Parker, A., Spudich, G., Yates, A., Zerbino, D. R., and Searle, S.M. J. (2014). Ensembl 2014. Nucleic Acids Research, 42(D1):D749–D755. →pages 6Flood, T. F., Iguchi, S., Gorczyca, M., White, B., Ito, K., and Yoshihara, M.(2013). A single pair of interneurons commands the Drosophila feeding motorprogram. Nature, 499(7456):83–7. → pages 2Fujikawa, K., Takahashi, A., Nishimura, A., Itoh, M., Takano-Shimizu, T., andOzaki, M. (2009). Characteristics of genes up-regulated and down-regulatedafter 24 h starvation in the head of Drosophila. Gene, 446(1):11–7. → pages 2,3, 31, 33Gilestro, G. F. and Cirelli, C. (2009). pySolo: a complete suite for sleep analysisin Drosophila. Bioinformatics (Oxford, England), 25(11):1466–7. → pages 30Giraldo-Calderon, G. I., Emrich, S. J., MacCallum, R. M., Maslen, G., Dialynas,E., Topalis, P., Ho, N., Gesing, S., Madey, G., Collins, F. H., and Lawson, D.(2015). VectorBase: an updated bioinformatics resource for invertebrate62vectors and other organisms related with human diseases. Nucleic AcidsResearch, 43(D1):D707–D713. → pages 8Godlewski, J., Wang, S., and Wilson, T. G. (2006). Interaction of bHLH-PASproteins involved in juvenile hormone reception in Drosophila. Biochemicaland biophysical research communications, 342(4):1305–11. → pages 19, 56Goodman, W. G. and Cusson, M. (2012). The Juvenile Hormones. In Gilbert, L.,editor, Insect Endocrinology, pages 311–365. Elsevier B.V. → pages 14, 15,18, 19, 20, 21, 55, 56Gordon, M. D. and Scott, K. (2009). Motor control in a Drosophila taste circuit.Neuron, 61(3):373–84. → pages 2Harmon, M. A., Boehmt, M. F., Heymant, R. A., and Mangelsdorf, D. J. (1995).Activation of mammalian retinoid X receptors by the insect growth regulatormethoprene. Proceedings of the National Academy of Sciences,92(June):6157–6160. → pages 19Hendricks, J. C., Finn, S. M., Panckeri, K., Chavkin, J., Williams, J., Sehgal, A.,and Pack, A. (2000). Rest in Drosophila Is a Sleep-like State. Neuron,25(1):129–138. → pages 30Hernandez-Martinez, S., Rivera-Perez, C., Nouzova, M., and Noriega, F. G.(2015). Coordinated changes in JH biosynthesis and JH hemolymph titers inAedes aegypti mosquitoes. Journal of insect physiology, 72:22–7. → pages 55Hong, S.-H., Lee, K.-S., Kwak, S.-J., Kim, A.-K., Bai, H., Jung, M.-S., Kwon,O.-Y., Song, W.-J., Tatar, M., and Yu, K. (2012). Minibrain/Dyrk1a regulatesfood intake through the Sir2-FOXO-sNPF/NPY pathway in Drosophila andmammals. PLoS genetics, 8(8):e1002857. → pages 14Ichinose, R., Nakamura, A., Yamoto, T., Booth, T. I. M. F., Maeda, S.,Hammocki, B. D., Road, M., Ox, O., and Toxicology, E. (1992). Uptake ofJuvenile Hormone Esterase by Pericardial Cells of Manduca sexta. Insectbiochemistry and molecular biology, 22(8):893–904. → pages 16, 17Illumina (2011). RNA-Seq Data Comparison with Gene Expression Microarrays.White Paper. → pages 3Ishimoto, H. and Kitamoto, T. (2010). The steroid molting hormone Ecdysoneregulates sleep in adult Drosophila melanogaster. Genetics, 185(1):269–81. →pages 1963Iwema, T., Chaumot, A., Studer, R. a., Robinson-Rechavi, M., Billas, I. M. L.,Moras, D., Laudet, V., and Bonneton, F. (2009). Structural and evolutionaryinnovation of the heterodimerization interface between USP and the ecdysonereceptor ECR in insects. Molecular biology and evolution, 26(4):753–68. →pages 19Ja, W. W., Carvalho, G. B., Mak, E. M., de la Rosa, N. N., Fang, A. Y., Liong,J. C., Brummel, T., and Benzer, S. (2007). Prandiology of Drosophila and theCAFE assay. Proceedings of the National Academy of Sciences of the UnitedStates of America, 104(20):8253–6. → pages 10, 28Jewell, S. (2014). Measuring Conserved Changes in Gene Regulation. Technicalreport. → pages 55Jones, G., Jones, D., Li, X., Tang, L., Ye, L., Teal, P., Riddiford, L., Sandifer, C.,Borovsky, D., and Martin, J.-R. (2010). Activities of natural methyl farnesoidson pupariation and metamorphosis of Drosophila melanogaster. Journal ofInsect Physiology, 56(10):1456–1464. → pages 19Jones, G. and Sharp, P. (1997). Ultraspiracle : An invertebrate nuclear receptorfor juvenile hormones. Proceedings of the National Academy of Sciences,94(December):13499–13503. → pages 19, 56, 57Kain, P. and Dahanukar, A. (2015). Secondary Taste Neurons that Convey SweetTaste and Starvation in the Drosophila Brain. Neuron, 85(4):819–832. → pages2Keene, A. C., Duboue´, E. R., McDonald, D. M., Dus, M., Suh, G. S. B., Waddell,S., and Blau, J. (2010). Clock and cycle limit starvation-induced sleep loss inDrosophila. Current biology : CB, 20(13):1209–15. → pages 2, 10, 44, 45, 51Kethidi, D. R., Xi, Z., and Palli, S. R. (2005). Developmental and hormonalregulation of juvenile hormone esterase gene in Drosophila melanogaster.Journal of insect physiology, 51(4):393–400. → pages 16, 57Kim, D., Pertea, G., Trapnell, C., Pimentel, H., Kelley, R., and Salzberg, S. L.(2013). TopHat2: accurate alignment of transcriptomes in the presence ofinsertions, deletions and gene fusions. Genome biology, 14(4):R36. → pages25Kume, K., Kume, S., Park, S. K., Hirsh, J., and Jackson, F. R. (2005). Dopamineis a regulator of arousal in the fruit fly. The Journal of neuroscience : theofficial journal of the Society for Neuroscience, 25(32):7377–84. → pages 3064Langmead, B. and Salzberg, S. L. (2012). Fast gapped-read alignment withBowtie 2. Nature methods, 9(4):357–9. → pages 25Li, W.-c. and Riddiford, L. M. (1996). Differential Expression of the TwoDuplicated Insecticyanin Genes , ins-a and ins-b , in the Black Mutant ofManduca sexta. Archives of Biochemistry and Biophysics, 330(1):65–70. →pages 20Linkert, M., Rueden, C. T., Allan, C., Burel, J.-M., Moore, W., Patterson, A.,Loranger, B., Moore, J., Neves, C., Macdonald, D., Tarkowska, A., Sticco, C.,Hill, E., Rossner, M., Eliceiri, K. W., and Swedlow, J. R. (2010). Metadatamatters: access to image data in the real world. The Journal of cell biology,189(5):777–82. → pages 69Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of foldchange and dispersion for RNA-seq data with DESeq2. Genome Biology,15(12):550. → pages 25Mathworks (2015). Video Stabilization Using Point Feature Matching. → pages71McGuire, S. E., Mao, Z., and Davis, R. L. (2004). Spatiotemporal gene expressiontargeting with the TARGET and gene-switch systems in Drosophila. Science’sSTKE : signal transduction knowledge environment, 2004(220):pl6. → pages 7Meunier, N., Belgacem, Y. H., and Martin, J.-R. (2007). Regulation of feedingbehaviour and locomotor activity by takeout in Drosophila. The Journal ofexperimental biology, 210(Pt 8):1424–34. → pages 15, 40Miyamoto, T., Slone, J., Song, X., and Amrein, H. (2012). A fructose receptorfunctions as a nutrient sensor in the Drosophila brain. Cell, 151(5):1113–25. →pages 2, 22Noriega, F. G. (2014). Juvenile Hormone Biosynthesis in Insects: What Is New,What Do We Know, and What Questions Remain? International ScholarlyResearch Notices, 2014:1–16. → pages 12, 13Organization, W. H. (2008). Guidelines for Drinking-Water Quality, 3rd editionincluding 1st and 2nd addenda. Technical report. → pages 20, 49Parthasarathy, R. and Palli, S. R. (2011). Molecular analysis of nutritional andhormonal regulation of female reproduction in the red flour beetle, Triboliumcastaneum. Insect biochemistry and molecular biology, 41(5):294–305. →pages 21, 54, 5965Rajan, A. and Perrimon, N. (2012). Drosophila cytokine unpaired 2 regulatesphysiological homeostasis by remotely controlling insulin secretion. Cell,151(1):123–37. → pages 21, 58Rauschenbach, I., Khlebodarova, T., Chentsova, N., Gruntenko, N., Grenback, L.,Yantsen, E., and Filipenko, M. (1995). Metabolism of the juvenile hormone inDrosophila adults under normal conditions and heat stress: Genetical andbiochemical aspects. Journal of Insect Physiology, 41(2):179–189. → pages17, 56, 57Restifo, L. L. and Wilson, T. G. (1998). A Juvenile Hormone Agonist RevealsDistinct Developmental Pathways Mediated by Ecdysone-Inducible BroadComplex Transcription. Developmental Genetics, 159(January):141–159. →pages 27Richter, K. and Gronert, M. (1999). Neurotropic effect of juvenile hormone III inlarvae of the cockroach, Periplaneta americana. Journal of Insect Physiology,45(12):1065–1071. → pages 21, 58Robertson, G., Schein, J., Chiu, R., Corbett, R., Field, M., Jackman, S. D.,Mungall, K., Lee, S., Okada, H. M., Qian, J. Q., Griffith, M., Raymond, A.,Thiessen, N., Cezard, T., Butterfield, Y. S., Newsome, R., Chan, S. K., She, R.,Varhol, R., Kamoh, B., Prabhu, A.-L., Tam, A., Zhao, Y., Moore, R. A., Hirst,M., Marra, M. A., Jones, S. J. M., Hoodless, P. A., and Birol, I. (2010). De novoassembly and analysis of RNA-seq data. Nat Meth, 7(11):909–912. → pages 4Root, C. M., Ko, K. I., Jafari, A., and Wang, J. W. (2011). Presynaptic facilitationby neuropeptide signaling mediates odor-driven food search. Cell,145(1):133–44. → pages 14Sarov-blat, L., So, W. V., Liu, L., Rosbash, M., and Hughes, H. (2000). TheDrosophila takeout Gene Is a Novel Molecular Link between CircadianRhythms and Feeding Behavior. Cell, 101:647–656. → pages 15, 40Schmid, B., Helfrich-Forster, C., and Yoshii, T. (2011). A new ImageJ plug-in“ActogramJ” for chronobiological analyses. Journal of biological rhythms,26(5):464–7. → pages 30Semmelhack, J. L. and Wang, J. W. (2009). Select Drosophila glomeruli mediateinnate olfactory attraction and aversion. Nature, 459(7244):218–23. → pages 1Shemshedini, L., Lanoue, M., and Wilson, T. G. (1990). Evidence for a JuvenileHormone Receptor Involved in Protein Synthesis in Drosophila melanogaster.The Journal of Biological Chemistry, 265(4):1913–1918. → pages 18, 5666Shiao, S.-H., Hansen, I. a., Zhu, J., Sieglaff, D. H., and Raikhel, A. S. (2008).Juvenile hormone connects larval nutrition with target of rapamycin signalingin the mosquito Aedes aegypti. Journal of insect physiology, 54(1):231–9. →pages 21, 54, 55, 59Stafford, J. W., Lynd, K. M., Jung, A. Y., and Gordon, M. D. (2012). Integrationof Taste and Calorie Sensing in Drosophila. The Journal of Neuroscience,32(42):14767–14774. → pages 2, 10, 28, 59Stout, J., Atkins, G., and Zacharias, D. (1991). Regulation of cricket phonotaxisthrough hormonal control of the threshold of an identified auditory neuron.Journal of comparative physiology., 169:765–772. → pages 21Trapnell, C., Williams, B. A., Pertea, G., Mortazavi, A., Kwan, G., van Baren,M. J., Salzberg, S. L., Wold, B. J., and Pachter, L. (2010). Transcript assemblyand quantification by RNA-Seq reveals unannotated transcripts and isoformswitching during cell differentiation. Nat Biotech, 28(5):511–515. → pages 4Wang, X.-h., Aliyari, R., Li, W.-x., Li, H.-w., Kim, K., Atkinson, P., and Ding,S.-w. (2006). RNA Interference Directs Innate Immunity Against Viruses inAdult Drosophila. Science, 312(5772):452–454. → pages 7Wang, Z., Gerstein, M., and Snyder, M. (2009). RNA-Seq: a revolutionary toolfor transcriptomics. Nature Reviews Genetics, 10(1):57–63. → pages 3Waterhouse, R. M., Tegenfeldt, F., Li, J., Zdobnov, E. M., and Kriventseva, E. V.(2013). OrthoDB: a hierarchical catalog of animal, fungal and bacterialorthologs. Nucleic acids research, 41(Database issue):D358–65. → pages 9,25, 36, 39, 54Wen, D., Rivera-Perez, C., Abdou, M., Jia, Q., He, Q., Liu, X., Zyaan, O., Xu, J.,Bendena, W. G., Tobe, S. S., Noriega, F. G., Palli, S. R., Wang, J., and Li, S.(2015). Methyl farnesoate plays a dual role in regulating Drosophilametamorphosis. PLoS genetics, 11(3):e1005038. → pages 13, 14, 17, 56Wheeler, D. E. and Nijhout, H. F. (1981). Soldier determination in ants: new rolefor juvenile hormone. Science, 213:361–363. → pages 21Williams, D. W. and Truman, J. W. (2005). Remodeling dendrites during insectmetamorphosis. Journal of neurobiology, 64(1):24–33. → pages 21Wilson, T. G. and Ashok, M. (1998). Insecticide resistance resulting from anabsence of target-site gene product. Proceedings of the National Academy ofSciences, 95(November):14040–14044. → pages 1867Wilson, T. G. and Fabian, J. (1986). A Drosophila melanogaster Mutant Resistantto a Chemical Analog of Juvenile Hormone. Developmental Biology,201:190–201. → pages 18, 57Wilson, T. G., Landers, M. H., and Happ, G. M. (1983). Precocene I and IIinhibition of vitellogenic oocyte development in Drosophila melanogaster.Journal of insect physiology, 29(3):249 – 254. → pages 51, 54Wyatt, G. (1961). The Biochemistry of Insect Hemolymph. Annual Review ofEntomology, 6:75–102. → pages 31Yarmolinsky, D. a., Zuker, C. S., and Ryba, N. J. P. (2009). Common sense abouttaste: from mammals to insects. Cell, 139(2):234–44. → pages 1Yin, C.-M., Zou, B.-X., Jiang, M., Li, M.-f., Qin, W., Potter, T., and Stoffolano, J.(1995). Identification of Juvenile Hormone III Bisepoxide (JHB3) , JuvenileHormone III and Methyl Farnesoate Secreted by the Corpus Allatum ofPhormia regina (Meigen), In Vitro and Function of JHB3 Either Applied Aloneor as a Part of a Juvenoid Blend. Journal of insect physiology, 41(6):473–439.→ pages 1768Appendix ASupporting MaterialsAlthough not immediately relevant to the subject of this thesis, I developed twopieces of software in addition to the actmon R package that have seen relativelyheavy use by both my and other labs. These are covered briefly here.A.1 GCaMP 4DGCaMP 4D allows researchers to view and analyze microscopy data in a truly 3Dmanner, allowing you to use older equipment like confocal microscopes in a man-ner that has previously only been possible with newer equipment (like light-sheetmicroscopes). As an example, the activity of a 3D field of neurons (expressinga calcium sensor) can be imaged in real time by passing a microscopy plane re-peatedly through the same 3D zone. Whereas traditional data analysis (or the hu-man eye) might be unable to find changes in fluorescence between two timepoints,GCaMP 4D is much more sensitive, and considers the entire 3D field, meaning thatone feature cannot “obscure” another. Researchers can even make videos of theentire 3D space over time or choose to focus on a single slice of the compositefield through the integrated user interface.The expected input is any microscopy file that can be opened using Bio-Formats(Linkert et al., 2010). Using metadata included in the file, the stack is formattedinto a series of 3D images, each representing one ”pass” through the specimen.Passes can be viewed in either 2D or 3D. When viewing in 2D, each pass is ”flat-69ABCFigure A.1: Sample GCaMP 4D output.A scaled down image of GCaMP 4D’s user interface is displayed in A. B and Cshow the change in fluorescence of the sample from A in 2D (B) or 3D (C) re-spectively. The sample being viewed is a pair of Drosophila gustatory neuronsresponding to stimuli.70tened” to a single maximum projection (each pixel is the vertical maximum of allthe pixels at that position in the stack) or simply viewed as an individual micro-scope image at a specified depth.Importantly, this algorithm is able to perform 3D field subtraction. An entire3D foreground pass can be compared to a 3D background pass. To make this possi-ble, every image in each Z-stack/pass are first ”stabilized” using the SURF/MSACalgorithms in a similar methodology to that demonstrated by Mathworks (2015).This ensures that the features in each pass actually line up, even though the spec-imen may have shaken or moved during imaging. Once stabilization is complete,the difference between the images is computed and displayed back to the user af-ter gaussian denoising. This 3D field subtraction algorithm is applied regardlessof whether or not a researcher is viewing a sample in 2D or 3D. The formula forimage subtraction is as follows (occurs on a per-pixel basis where division by zeroartifacts converted to zero):f oreground−backgroundbackground∗100 (A.1)This tool exports both still images and videos of the region being imaged forlater use. It is intended as both a general 3D viewer and full analysis tool, andis best used for identifying changes in fluorescence in over the course of a singleimaging session. Currently this tool has been used to help identify and character-ize a set of second-order Drosophila gustatory neurons, although it also has otherapplications including simply viewing a confocal Z-stack in three dimensions. TheMATLAB source code and standalone binary executables for Windows, OSX, andLinux are available at https://github.com/kazi11/GCaMP 4D.A.2 fly trackerThis is a collection of MATLAB algorithms designed to track and quantify thebehavior of individual Drosophila adults and larvae. There are two separate al-gorithms for tracking larvae and adults. The larval tracking script is optimizedfor slowly-moving animals and tracks individual objects after background subtrac-tion. The adult fly tracking script tracks adult flies by finding the darkest pixelafter background division, and calculates the centroid (i.e. where is the middle71A BFigure A.2: Sample fly tracker output.(A) A sample position trace calculated from a cellphone video of an adult fly. Bluerepresents the fly’s position at the beginning of the video and yellow represents thefly’s final position. (B) Velocities calculated from several fly position traces.of the darker pixels) of the surrounding region and is loosely based upon FTrack(Andrews et al., 2015). Both scripts are able to intelligently identify false tracksand interpolate missing data. The output of both scripts can be used for any ofthe downstream analysis tools and can generate robust tracks from even cellphone-quality video.Downstream analysis is quite straightforward. The position traces created bythe fly and larvae trackers can be used to calculate a number of different statistics,including velocity, distance traveled, and probability of an animal being in a givenpart of the experimental arena. Output from these scripts can then be plotted andstatistics are automatically performed between different genotypes/experimentalconditions.72

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0223111/manifest

Comment

Related Items