01&-(+' 2*&3**+'4(5&5'$+6'#$.$5-&*5 ' · Faculté des sciences Secrétariat-décanat de Faculté...
Transcript of 01&-(+' 2*&3**+'4(5&5'$+6'#$.$5-&*5 ' · Faculté des sciences Secrétariat-décanat de Faculté...
Impactofenvironmentonco-evolution
betweenhostsandparasites
THÈSEPRÉSENTÉEÀLAFACULTÉDESSCIENCES
UNIVERSITÉDENEUCHÂTEL
POURL’OBTENTIONDUGRADEDEDOCTEURÈSSCIENCES
PAR
ZellerMICHAEL
AcceptéesurpropositionduJury:
Prof.KoellaJacob,directeurdethèse,UniversitéNeuchâtel
Prof.FlattThomas,rapporteur,UniversitéLausanneProf.HelfensteinFabrice,rapporteur,UniversitéNeuchâtel
Soutenuele24juin2016
UniversitédeNeuchâtel
2016
Faculté des sciences
Secrétariat-décanat de Faculté Rue Emile-Argand 11
2000 Neuchâtel - Suisse Tél: + 41 (0)32 718 2100
E-mail: [email protected]
Imprimatur pour thèse de doctorat www.unine.ch/sciences
IMPRIMATUR POUR THESE DE DOCTORAT
La Faculté des sciences de l'Université de Neuchâtel autorise l'impression de la présente thèse soutenue par
Monsieur Michael ZELLER
Titre:
“Impact of environment on co-evolution between hosts and parasites”
sur le rapport des membres du jury composé comme suit :
- Prof. Jacob Koella, directeur de thèse, Université de Neuchâtel, Suisse - Prof. ass. Fabrice Helfenstein, Université de Neuchâtel, Suisse - Prof. Thomas Flatt, Université de Lausanne, Suisse
Neuchâtel, le 4 juillet 2016 Le Doyen, Prof. B. Colbois
5
Abstract
Manyparasitescausepathologyandmortalityandcanbeamajorsourceofselectionontheirhosts,creatingstrongselectionpressuresontheevolutionofhostsdefensestrategies.Changesinthehost’stoleranceandresistancetopathogencanstrongly influencethespreadofadiseaseandhenceinfluenceselectiononvirulence.Understandinghowecologicalconditionsinfluencethehost’slifehis-toryanddefensemechanisms,howtheyalterinfectiondynamicsandcontributetotheparasitesviru-lenceandtransmissionandhowtheyshapetheco-evolutionarydynamics isessential forgaining fur-therinsightsintohost-parasiteinteraction.
Inthisthesis,singlegenerationexperimentsandexperimentalevolutionwereusedtoexploretheim-pactof theenvironmentonhost-parasite interactionand theirevolution.Firstly, the roleof resourcevariabilityonthelifehistoryofthemosquitohostAe.aegyptiwasinvestigated.Secondly,therelation-ship between the growthof theparasiteV. culicisand thehealthof itsmosquito hostwas exploredacross ecological variables. The parasites growth rate and asymptotic load was estimated and com-pared in living and naturally dying hosts. Thirdly, theoretical predictions about the evolution of hostdefenseagainstparasitesweretested.Inparticular,theroleofresourceavailabilityontheevolutionofthehost’stoleranceandresistancetopathogenwasinvestigated.Finally,thisthesisexaminestheim-pactoftheenvironmentandco-evolvingparasitesonthehost’slifehistoryevolution.Theexperimentsintroduced here show that variable environmental conditions can influencemany central aspects ofhost-parasiteinteractions,onesthatplayimportantrolesinshapingevolutionarydynamics.Inthisthe-sis,theresultsofthoseexperimentsaredescribedindetailandtheirimplicationsforhost-parasiteco-evolutionarediscussed.
Overall,thisthesisemphasizesthecomplexityanddependencefromenvironmentalconditionsofhost-parasite interactionand theirevolution.Whenconsidering the spatialand temporalecologicaldiffer-encesofnaturalhabitats,theresultspresentedheremayhelptoleadtoamoreprofoundknowledgeofhost-parasiteinteractions.
Keywords
Aedesaegypti,co-evolution,epidemiology,experimentalevolution,host-parasite interactions,life-historyevolution,resourcevariability,Vavraiaculicis,virulence
7
Résumé
Beaucoup de parasites provoquent des pathologies et de la mortalité et peuvent être unesourcemajeuredesélectionsurleurhôte,enappliquantunefortepressiondesélectionsurl’évolutionde leurs stratégies de défense. Les changements de tolérance et de résistance de l’hôte face aupathogènepeuventfortementinfluencerlapropagationd’unemaladie,etdoncinfluencerlasélectionde la virulence. Comprendre comment les conditions écologiques influencent l’histoire de vie et lesmécanismesdedéfensedel’hôte,commentellesaltèrentladynamiqued’infectionetcontribuentàlavirulenceetlatransmissionduparasite,etcommentellesfaçonnentladynamiquecoévolutiveestes-sentielpouravoirunemeilleurecompréhensiondesinteractionshôte-parasite.
Dans cette thèse, des expériences sur une seule génération et des expériences d’évolution expéri-mentaleontétéutiliséespourexplorer l’impactde l’environnementsur les interactionshôte-parasiteetleurévolution.Premièrement,lerôledelavariabilitédesressourcessurl’histoiredeviedel’hôte,lemoustiqueAe.aegypti,aétéinvestigué.Deuxièmement,lelienentrelacroissanceduparasiteV.culiciset la santée de son hôte moustique a été exploré sous différentes conditions écologiques.Troisièmement, le rôlede ladisponibilitédesressourcessur l’évolutionde la toléranceetde la résis-tancede l’hôteaupathogèneaétéexploré.Finalement, l’impactde l’environnementetdesparasitessur l’évolution des traits d’histoire de vie de l’hôte a été étudié. Les expériences introduites ici ontmontréquedes conditions environnementales variablespeuvent influencerbeaucoupd’aspects cen-traux des interactions hôte-parasite, en particulier ceux façonnant de manière importante la dy-namiqueévolutive.Danscettethèse,lesrésultatsdecesexpériencessontdécritsendétailetleursim-plicationspourlacoévolutionhôte-parasitesontdiscutées.
Plusglobalement,cettethèsesouligne lacomplexitéet ladépendanceauxconditionsenvironnemen-talesdesinteractionshôte-parasiteetleurévolution.Enconsidérantlesdifférencesspatialesettempo-rellesdeshabitatsnaturels,lesrésultatsprésentésicipeuventaideràconduireàuneconnaissanceplusprofondedesinteractionshôte-parasite.
Mots-clés
Aedesaegypti,évolutiondel’histoiredevie,co-évolution,épidemiologie,évolutionexperimen-telle,interactionshôte-parasite,variabilitéderesources,Vavraiaculicis,virulence
9
Contents
Abstract..................................................................................................................................................5
Résumé...................................................................................................................................................7
ListofFigures........................................................................................................................................10
ListofTables.........................................................................................................................................11
Generalintroduction........................................................................................13Chapter1
EffectsoffoodvariabilityongrowthandreproductionofAedesaegypti.........25Chapter2
Context-dependentrelationshipbetweenparasitegrowthandChapter3
virulenceinamicrosporidia-mosquitointeraction...........................................39
TheroleoftheenvironmentontheevolutionoftoleranceandChapter4
resistancetoapathogen..................................................................................53
Antagonisticcoevolutionandresourcesalterthehost’slifehistoryChapter5
evolution.........................................................................................................65
Synthesisandfutureresearch..........................................................................79Chapter6
Acknowledgements..............................................................................................................................83
References............................................................................................................................................85
10
ListofFigures
Figure1:1Resourcesinhost-parasiteinteraction.................................................16
Figure1:2Mosquitolifecycle................................................................................20
Figure1:3LifecycleofVavraiaculicis....................................................................21
Figure2:1Mosquitogrowth..................................................................................31
Figure3:1:Sporeloadandprobabilityofinfection...............................................45
Figure3:2Parasitegrowthparameters.................................................................47
Figure3:3Longevityofinfectedmosquitoes.........................................................49
Figure4:1Toleranceandresistance......................................................................61
Figure4:2Tolerancevs.resistance........................................................................62
Figure5:1Ageandsizeatmaturity........................................................................72
Figure5:2Phenotypicplasticity.............................................................................73
Figure5:3Probabilityofemergenceandsporeload.............................................75
11
ListofTables
Table2:1Statisticalsummaryforjuveniletraits……….……......................................31
Table2:2Statisticalsummaryforadulttraits………………………………………………..……33
Table2:3Repeatedmeasuresanalysisofclutchsizes.…………………………….………….35
Table3:1Sporeloadandprobabilityofinfection……..……………………………………..…46
Table3:2Parasitegrowthparameters……..…………………………………………………………48
Table3:3Statisticalsummaryofsurvivalanalyisis….………………………………………..…49
Table4:1Statisticalsummaryforresistance…………….………..……………………………….59
Table4:2Statisticalsummaryfortolerance………….…………………………………..………..60
Table5:1Statisticalsummaryoflife-historytraits……………..……………………………….76
Impactofenvironmentonco-evolutionbetweenhostsandparasites
13
GeneralintroductionChapter1
1.1 Background
Parasites1areomnipresentinnatureancanbeasubstantialsourceofselectionontheirhosts.Parasites
areinvolvedinmanyecologicalandevolutionaryprocessesincludingtheevolutionofsex,socialbehav-
ior,maintainingorincreasinggeneticdiversityofhosts,alteringpopulationstructuresandeveninthe
diversificationofspecies.Someparasitesareknowntobehighlyvirulentbycausingseriousharmand
deathintheirhostpopulations.Forexample,ithasbeenestimatedthat20%ofthemarinebiomassis
killedbyviruseseverydayand that in theocean1023newviral infectionsoccurevery second (Suttle
2007).Therearemorethan1,400describedparasitespeciesthatcan infecthumans(Woolhouseand
Gowtage-Sequeria 2005) and approximately 15% of human death is caused by pathogens. Malaria
aloneaccounts for20%ofchildhoodmortalityand in2015therewereanestimated214millionnew
caseswith440,000deaths (WorldHealthOrganization2015).Otherparasitesare less severebutstill
causedebilitatingsymptoms,whichcanreduceorpreventreproductionanddecreasecompetitiveabili-
ties.
Whysomeparasitescauseveryseverepathologyandmortalitywhileothersarerelativelybenign,and
whysomeparasitesareveryefficientinspreadinginapopulationandbecomeepidemicwhileothers
stay in small frequencies, has fascinated the scientific community for decades. Such diverse disease
characteristicsaretheresultofcomplexco-evolutionarydynamicsbetweenhostsandparasites.Para-
sites thatcausepathologyandmortalitycanbeamajorsourceofselectionontheirhosts.Thehosts
haveevolveddiversemechanismsofdefense to reduce the successof infection, increase the rateof
clearance,orat leastby reducing thedetrimentaleffectsof theparasite. Thesedefensemechanisms
rangefromverycompleximmunesystemprocesses,modificationofcellsurfacestopreventinfection,
tochangesinhostbehaviorandlife-historystrategiesinordertoresistortolerateparasites.Thepara-
sitesontheotherhand,aredependentonthehost’sresourcesandaimtoreachahighdensity,within
1 Parasites in thesis are defined broadly as infecitous agents that cause disease (including virus, bacteria, fungi, protozoa,helminths).
Impactofenvironmentonco-evolutionbetweenhostsandparasites
14
thehost,inordertobetransmittedefficientlyinapopulation.Howeverbecauseahighparasitedensity
isexpectedtoreducethehost’ssurvivalandthusthedurationofthetransmissionperiod,theparasites
virulence2istraded-offwithitstransmission(AndersonandMay1982;Alizonetal.2009).
Accordingly,thehost’sabilitytoresistparasites is incontinuousconflictwiththeparasites infectivity,
thegrowthanditstransmissibility.Anevolutionaryresponsetoparasitismofthehostmayagainalter
theselectionpressureoftheparasite.Thisreciprocalselectioncanresultinco-evolution,withcontinu-
ouschangesofallelefrequenciesinhostsandparasites(Gandonetal.2008;GabaandEbert2009).This
canbedrivenbyparasite-mediatedselectionagainstcommonallelesorbydirectionalselectionthrough
thesuccessivefixationofadvantagesmutations.Accordingtothered-queenhypothesistheevolution-
ary ratesofchangeshouldbeacceleratedwithcoevolution. Indeedgenomesofcoevolvinghostscan
evolvefastercomparedtopopulationsevolvingagainstaconstantparasitepopulation(Patersonetal.
2010;KashiwagiandYomo2011).Howeverinmanysituationssuchpredictionsaretoosimple,especial-
ly when hosts develop tolerance instead of resistance, when there is enhanced competition for re-
sourcesorwhenimmunopathologycausesabigpartofthedamage(RestifandGraham2015).Forex-
ample it has been shown that co-evolutionary dynamics canbe alteredby resource availability; high
resource levels leads todecreased fluctuatingselection in resistance (LopezPascuaetal.2014).Such
studiesillustratetheneedtounderstandhost-parasiteinteractionsinan“eco-co-evolutionary”context
because of the presence of genotype-by-genotype-by-environment interactions that influence their
evolution.
Hostsresourcesaffectmanyaspectsofhostparasiteinteractionandhaveshowntodrasticallychange
theoutcomeofinfection(Shetty2010).However,inmosttheoreticalstudiesthatpredicttheevolution
ofhost-parasite interactions,one fundamentalaspectofparasites isgenerally ignored: thatparasites
steal resources fromtheirhost to support theirowndevelopment (see (Smith1993;Halletal.2007;
Halletal.2009a;Halletal.2010)forexceptions).Inadditiontogeneticfactorsofhostsandparasites,
resourceavailabilityhasbeenrecognizedasakeyaspect inthedynamicsof infectiousdiseases, influ-
encinghostdefense,parasite transmissionandvirulence (Lazzaroand Little2009;WolinskaandKing
2009).Severalstudiesshowhowresourcequality,aswellasquantity,shapesparasitevirulence(Jokela
etal.1999;Brownetal.2000;FergusonandRead2002)andalsodirectlyinfluencestheproductionof
parasites(Bedhommeetal.2004;Johnsonetal.2007;DeRoodeetal.2008;Halletal.2009b).
2Virulence in this thesis isdefindedas theparasite inducedhostmortalityand fecundity loss fromtheageof infectionon-wards.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
15
There are twogeneralways inwhich resource availability in thehost’s environment influenceshost-
parasiteinteractions.Ontheonehand,highlevelsofresourcescanpositivelyinfluencethehost’s im-
muneresponseandthereforeincreaseitsresistancetoparasiteinfection(KoellaandL.2002;Ayresand
Schneider2009).Malnourishedhostsmaybeweakerandmoresusceptibletoinfectiousdisease(Moret
2000) so that the lower the food availability, the higher the costs of parasitism (Ferguson and Read
2002).Forexample,thefleaXenopsyllaramensisproducesmoreeggswhenfeedingonhostsexperienc-
ing diet restriction (Krasnov et al. 2005). On the other hand, a parasite uses the host’s internal re-
sourcesforitsowndevelopment.Accordingly,hostsrearedonhighlevelsoffoodmaystoremoreen-
ergy reserves and therefore present a better environment for the parasite to develop. Indeed, en-
hancedqualitiesandquantitiesof resources increase theproductionofeffectivepropagulesofpara-
sites(e.g.ascogregarines(Tseng2006)ormicrosporidians(AgnewandKoella1999b;Bedhommeetal.
2004)) inmosquitoes, and trypanosomes in bumble-bees (Brown et al. 2000), but also increase host
survival(asin(Jokelaetal.1999)),andcannegativelyaffecthostreproductivesuccess.Well-feddaph-
nia hosts infectedwithCaulleryamesnili, for example, showed, in regard to fecundity,more severe
parasiticeffects(Bittneretal.2002).Itfurtherhasbeenshownthatincreasedfoodavailabilityincreas-
esparasitetransmissionrate(Ebertetal.2000;J.Ryderetal.2007;Valeetal.2013).
Becauseparasitegrowth,virulenceandtransmissionareprincipalparametersinepidemiologyandthe
evolution of host-parasite interactions, ignoring the impact of resource availability limits our under-
standingofhost-parasiteinteractionsandevolution.Studyingthesetraitsunderdifferentresourcelev-
els is thereforevital tounderstand theparasiteswithinhost-dynamics (e.g. the relationshipbetween
parasiteburdenandhosthealth),whichplaysanimportantroleintheevolutionofparasitesandhosts
(Antiaetal.1994;AlizonandvanBaalen2005a).Byconsideringthatthehost’snaturalhabitatcanvary
drastically in space and time, our understandingof host-parasite interactions, their co-evolution, the
predictions of parasite evolution and especially the management of parasites must incorporate
knowledgeofhowenvironmentalvariationsinfluenceparasite-hostinteractions.Itisthereforeofgreat
importancetoscienceandsocietytodeepenourknowledgeinthisfield.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
16
Figure1:1Resourcesinhost-parasiteinteraction
Resourceavailabilityasakeyfactorinhost-parasiteinteractions:Resourceavailabilitycaninfluencethehost’sgrowthrate,
ageandsizeatmaturity,longevityandfecundity.Resourcescanalsobeimportantforthedevelopmentoftheparasitewithin
itshostbecauseofadirectinfluenceoftheresourcesavailabletotheparasite,andviathehostslifehistory,inparticularbody
sizethatoftenconstrainstheparasitedevelopment.Theamountofresourcescanalsoinfluencethehost’sresistance(limit
parasitedevelopment)andtolerance(reducedetrimentaleffectsofparasiteswithoutaffectingtheparasitesdevelopment)to
pathogens.Thedevelopmentoftheparasitecanfeedbackontothehost’slifehistory,especiallybyreducingitslongevity.The
transmissionoftheparasiteisdependentonparasitegrowthandthelongevityofthehost,andindirectlyonthehost’sre-
sources.
1.2 Thesisintroduction
InthisthesisIstudytheimpactoftheenvironmentonhost-parasiteinteractions(summarizedinFigure
1.1)andhowthis influencestheirevolution. IusethemosquitoAedesaegyptiand itsmicrosporidian
parasiteVavraia culicis. I aim toempirically test certainaspectsofmathematicalmodels thatpredict
theevolutionofthehostandparasiteslife-history,virulenceandtheparasitesdefensestrategies.The-
seresultswillhopefullyenlargeourknowledgeabouttheevolutionofhostparasiteinteractionsingen-
eral,andcontributetothedevelopmentofmorepowerfulandrealisticlife-historyandepidemiological
models.Furthermore,thedatamightbedirectlyrelevantforpublichealthbecauseoftheroleofmos-
quitosasavectorforseveralinfectiousdiseases.Theresultsmightalsobeusefulforpeopleworkingon
biologicalcontrolasmicrosporidianshavebeenproposedforthecontrolofmosquitoesandtheirvec-
tor-bornediseases(SweeneyandBecnel1991;Koellaetal.2009;LorenzandKoella2011).
1.2.1 Foodvariability,growthandlaterlifeconsequences
Resource availability is as key component of life history theory because of its role in determining
growth,survivaland fecundityof individuals (Stearns1992).Diet restrictionhasbeenassociatedwith
Impactofenvironmentonco-evolutionbetweenhostsandparasites
17
slowergrowth,smalleradultsize,delayedmaturityandlowerfecundity,butwithalongerandhealthier
life(StearnsandKoella1986).Becauseadultsizeisaveryimportantdeterminantoffitness,adaptations
tohandleperiodsoffoodrestrictionsareveryimportant.Onepossibility,whichisreferredtoascom-
pensatorygrowth,istogrowatacceleratedratesafteraperiodofundernourishmentinordertocatch
up in size. Rapidly growing individualsmight respond to food stress by decelerating growth rates in
ordertousetheavailableresourcesformaintenanceandreproduction.Compensatorygrowth,aswell
asdeceleratinggrowth,aregenerallythoughtofasadaptiveresponses,buttheirconsequenceslaterin
liferemainlargelyunexplored. Inthesecondchapter, Iexperimentallytesttheeffectofvariablefood
availability during the development of the mosquito host Aedes aegypti (leading to compensatory
growthordeceleratedgrowth),andtheassociatedchangesinlongevityandreproductivesuccess.Such
dataisrelevantforlife-historytheoryandisalsodirectlyrelevantforpublichealthduetothemosqui-
to’sroleasvectorforseveralinfectiousdiseases.
1.2.2 Within-hostgrowthandvirulence
Mostmodelsfortheevolutionofhostparasiteinteractionsarebasedontheassumptionthatthereis
trade-off between virulence and the rate of transmission (Anderson andMay 1979; Alizon and Lion
2011,Gandonetal.2001;Dieckmannetal.2002).Underlyingthisassumptionistheideathatparasites
mustgrowrapidly toahigh load inorder tobe transmittedefficiently,but thathighparasitedensity
alsoreducesthehost’ssurvivalandthusthedurationofthetransmissionperiod.Thistrade-offimplies
thatmaximaltransmissionisoftenhighestatintermediatelevelsofvirulence.Howeveroptimallevels
ofvirulencedependontheshapeofthetrade-off(AndersonandMay1982;BremermannandPickering
1983).Thegeneralassumptionthatparasiteburden ispositivelycorrelatedwith itsvirulenceholds in
severalsystems(Mellorsetal.1996;MackinnonandRead1999;DeRoodeetal.2008;DeRoodeand
Altizer 2010). However, because resource availability has been shown to influence parasite growth
(Johnson et al. 2007; De Roode et al. 2008;Michalakis et al. 2008; Hall et al. 2009b) and virulence
(Jokelaetal.1999;Brownetal.2000;FergusonandRead2002),itislikelytoaffecttheassociationbe-
tweenthetwo,sothatweonlyseeanegativerelationshipbetweenparasitedevelopmentandlongevi-
tyinsomeenvironments.InthethirdchapterIinvestigateexperimentallyhowaspectsoftheenviron-
ment (the amountof food available to larvae and to adults, and the age at infection) influences the
growthoftheparasite,thelongevityofthehostandtherelationshipbetweenthetwo.Studyingfactors
thatinfluencetherelationshipbetweenthehost’sgrowthanditsvirulenceisimportant,becausethey
mayaltertherelativecostsandbenefitsofrapidparasitereplicationandthereforedetermineadaptive
levelsofvirulence.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
18
1.2.3 Toleranceandresistance:twofundamentallydifferentdefensetraits
Adaptationsofthehosttoreducenegativeeffectsofparasitescanbeclassifiedintotwobroadcatego-
ries:Resistanceandtolerancetoparasites(Readetal.2008;Råbergetal.2009;Littleetal.2010).Re-
sistancereducesthesuccessofinfectionorincreasestherateofclearance,whereastolerancereduces
thedetrimentaleffectsoftheparasitewithoutaffectingthepathogendirectly.Itisimportanttodistin-
guishbetweenthetwotraitsbecausetheyhaveaverydifferenteffectontheevolutionofhost-parasite
interactions(RoyandKirchner2000;RestifandKoella2003;Boots2008).Tolerancegenesarepredicted
tobecomefixed,becausetheyarebeneficialforthehostandparasite,whileresistancegenesarepre-
dictedtorapidlychangebecausetheywouldprovokecounter-adaptationoftheparasitetoovercome
resistance (Roy and Kirchner 2000;Miller et al. 2006).One example that illustrates the evolutionary
significanceof tolerance is thecaseofsimian immunodeficiencyvirus (SIV) inmonkeys. Inmacaques,
whicharenon-naturalhosts,thevirusreplicatesrapidlyandinfectedanimalsdevelopAIDS.Incontrast,
insootymangabeys,whicharenaturalhostsofSIV,infectedindividualsdonotshowdiseasepreogres-
sionandshownodevelopmentofAIDS,evenwhencarryinghighvirusloads(Chakrabarti2004).Sooty
mangabeys,whichhavealongevolutionaryhistorywithSIV,thereforereducedthedamageofparasite
infectionandevolvedhighlevelsoftoleranceinsteadofresistance.Tolerancehasalsobeendiscussed
tobeclinicallyrelevant(Medzhitovetal.2012).Incontrasttoresistant-basedtherapy,tolerance-based
therapyaimstoimprovethehealthofthehostatagivenparasiteloadinsteadofreducingtheparasite
burden.Thismaynotleadtoselectionforresistantparasitesandwasconsideredas“evolution-proof”
(Readetal.2008;SchneiderandAyres2008).
Differentiatingbetweenresistanceandtolerance,studyingtheassociationbetweenthetwo,andpre-
dictinghowthetwoevolvehavebecomeimportanttopicsofevolutionaryparasitology.Whatismissing
areexperimentalstudiesontheextenttowhichevolutionfavorstoleranceorresistanceunderdiffer-
ent ecological settings. Using experimental evolution, I aim to test theoretical predictions about the
evolutionofhostdefenseagainstparasitesandinvestigatetheroleofresourceavailabilityonthecon-
currentevolutionoftoleranceandresistance.
1.2.4 Changeinlifehistorytoreducethecostsofparasitism
Achangeinhost’slifehistorycanbeanotherformtoreducethecostofparasitism.Lifehistorytheory
predictsthatearlymaturinghostsmayhaveaselectiveadvantagebecausetheycanevadeparasitismin
timeandwhenparasitizedreducedetrimentaleffectsonreproductionand longevity (Hochbergetal.
1992;Forbes1993;PerrinandChriste1996).Howeversuchadaptationscanbeassociatedwithcosts
laterinlife(reducedlate-lifefecundityandlongevity).Astunningexampleisthespreadofaninfectious
Impactofenvironmentonco-evolutionbetweenhostsandparasites
19
cancerinTasmaniandevilsthatcausesalmostcompletemortalityafterthefirstyearofadulthood.This
facialtumorcausesanabruptshiftinthehost’slifehistoryfrommultibreedingtowardsinglebreeding.
The Tasmanian devils responds to this strong selection pressure, by a 16-fold increase in precocious
sexualmaturity(Jonesetal.2008).Suchanalterationinlifehistorycanbeseenasaformofresistance,
whichmight leadtocomplexco-evolutionarydynamicsbetweenthe lifehistoriesofthehostandthe
parasites. In the fifthchapter Iaimtostudyhowcoevolvingandconstantparasitescan influencethe
hosts’lifehistoryunderdifferentresourcelevels.Studyingtheroleofresourceavailabilityispotentially
relevant because trade-offs between different life history traits might only be detectable when re-
sourcesarescarce.Accordingly,variableenvironmentsmightinfluencethelong-termhostevolution.
1.3 Experimentalsystem
In this thesis themosquitospeciesAedesaegyptiand itsmicrosporidianparasiteVavraiaculiciswere
usedasmodelorganismsto investigatethe impactof thehostenvironmentonhost-parasite interac-
tion.
1.3.1 ThemosquitoAedesaegyti
TheUGALstrainofthemosquitospeciesAe.aegypti(obtainedfromPatrickGuérin,UniversityofNeu-
châtel)wasusedforalloftheexperimentspresentedinthisthesis.ThemosquitospeciesAedesaegypti
is the principal vector of yellow-fever (Tomori 2004), dengue (World Health Organisation 2002) and
Zika-virus(Petersenetal.2016),growsinavarietyofdifferentnaturalandartificialcontainersholding
cleanfreshwater(Southwoodetal.1972)andoccursthroughoutthetropicsandsubtropics.Itisavery
wellstudiedspecies;itsecologyisknownindetail(Christophers1960),ithasbeenamodelorganismin
insectphysiologystudies(Clements1999),anditsfullgenomehasbeenpublished(Neneetal.2007).
Aedesaegyptiarehighlysusceptibletoawholerangeofenvironmentalconditions(Christophers1960).
Forexampleduringtheaquaticlarvalstages,wildmosquitoesareoftenundergoingperiodsofnutrient
restrictionandcompetitionforresourceslikebacteria,algaeandorganicmatter(ReiskindandLounibos
2009).
Impactofenvironmentonco-evolutionbetweenhostsandparasites
20
Figure1:2Mosquitolifecycle.
Generalizedlifecycleofmosquitoes:Aedesaegyptimosquitoesareholometabolousorganisms.Eachmosquitogoesthrough
fourdistinctstagesofitslifecycle:Egg,larva,pupaeandimago.Theadulthoodistheonlynon-aquaticstage.
1.3.2 Vavraiaculicis
Microsporidiaconstitutealarge,verydiversegroupofsingle-cell,endocellularparasites.Theybelongto
the kingdomof fungus, arewidespread in nature andhave a broad rangeof hosts (fromhumans to
invertebrates).Microsporidiaarespeciallycommoninarthropods;closetohalfofthedescribedgenera
haveinsectsastheirhosts(BecnelandAndreadis1999).ForthisthesisIusethemicrosporidanparasite
VavraiaculicisthatwasoriginallyderivedfromAe.albopictusinFloridaandkindlyprovidedbyJ.JBecnel
(USDA,Gainesville,USA). It is anaturalparasiteof several generaofmosquitoes includingAedesae-
gypti(WeiserandColuzzi1972).Naturalprevalenceratesofinfectionrangingfromlessthan1%upto
54%differing frommosquitospeciesandgeographical location (Andreadis2007). It isanobligateen-
docellularparasite,orally infectiousandistransmittedbetweenhostsbyinfectivespores(Figure1.3).
Thesporesaretheyonlystageoftheparasitethatcanpersistintheenvironmentoutsidethehostcell.
ThehostlarvaeingestthesporesofVavraiaculiciswiththeirfood,resultingininfectionofgutcellsand
epithelialcells.Afterpassingaseriesofdevelopmentalstageswithinthelarvae,theparasitebeginsto
produceitsinfectiousspores.Theinfectionthenspreadstogutandfatbodycells.Nocomponentofan
insect’simmunesystemhasbeendescribedwhichcanneutralizethisintracellularparasite(Bironetal.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
21
2005).Theparasite is transmitted in twoways.First, transmissioncanoccur from larva to larvaafter
larvaldeath.Inparticularcasesoffoodstressorstronginfection(Bedhommeetal.2004)larvaldeathis
enhanced,which initiatesanother roundofhorizontal transmission. Second, larvae survive the infec-
tion,withoutclearingitanddevelopintoadults.Thesporeswithinthehostcontributestotheparasites
fitnesseitherbythereleaseofsporeswhendying intheaquaticenvironmentor,becausesporescan
adheretothesurfaceoftheeggsandinfectthenewlyhatchedlarvae(Andreadis2007).
Figure1:3LifecycleofVavraiaculicis
LifecycleofVavraiaculicis:Sporesarereleasedintotheaquaticenvironmentwhenjuvenileoradultmosquitoescarrying
sporesdieandareorallyingestedbylarvae.Transmissioncanoccurfromlarvatolarvawhensporesarereleasedafterthe
larvadies(solidlines)oriflarvaesurvivetheinfectionanddevelopintoadults;thesporescanbereleasedwhenthemosquito
diesintheaquaticenvironmentortheycanadheretothesurfaceoftheeggsandinfectthenewlyhatchedlarvae(dashed
lines).
1.3.3 Resourcesaffectlife-historiesofhostandparasite
Resourceavailabilityplaysafundamentalroleintheinteractionbetweenmosquitoesandmicrosporidi-
anparasites.Larvalresourceavailabilityofthehostinfluencesthemosquito’sgrowthrate,ageatma-
Like all microsporidia, Vavraia culicis is an obligate endocellular parasite of several mosquito genera, including Aedes, Culex and Anopheles. Natural prevalence rates of infection range between 1% and 54% depending on the mosquito species and geographical location [69]. Vavraia culicis is horizontally transmitted when mosquito larvae ingest its spores. The spores first infect the mosquito gut epithelial cells, and the infection then spreads to other gut and fat body cells. After several rounds of replication within a larva, the parasite begins to produce its infectious spores. In some cases, in particular in conditions of food stress or intense infection [15], these kill the larva or pupa and are released into the breeding site, thus initiating another round of horizontal transmission. In other cases, larvae juveniles the infection and develop into adults (without clearing the infection). There is no transovarial transmission (i.e. the parasite does not penetrate the eggs of infected females), but spores can adhere to the surface of the eggs and infect the newly hatched larvae [69].
Resources affect life-histories of the host and parasite of this project
Resource availability is a key component of the interaction between mosquitoes and microsporidians. Larval food influences, for example, the mosquito’s growth rate and age at metamorphosis [15,70,71]). The resources available to the host are also important for the dynamics of the parasite within its hosts. Increasing larval food increases the production of spores [15], probably because of a direct influence of the energy available to the parasite (as in [8]) and an indirect influence via the host’s life-history, in particular body size, that constrains the parasite’s development (as in [70]). Larval food also influences the parasite’s virulence: less food increases the probability that infected mosquitoes die before their emergence [15,71] and decreases the longevity of infected adults [71]. It thus largely affects the parasite’s transmission route, either horizontally from dead juveniles or vertically from females as they lay their eggs. Although the mechanisms underlying virulence are not known, they may include a direct effect of the parasite (as assumed in many models, e.g. [46-48]) or the depletion of energy below a threshold necessary for the host’s survival [8,72]}.
Choice of experimental system
The mosquito Aedes aegypti is the major vector of dengue and yellow fever. It is a subtropical mosquito, whose larvae grow in natural or artificial containers [73]. It is a very well-studied organism: its ecology is known in detail [74], it has been a model organism in insect physiology [75], and its full genome has been published [76]. Its eggs can be hatched synchronously, and it can be reared easily in the laboratory.
The Vavraia-Aedes system is well suited for this research for many reasons. (i) The host and parasite have short generations, enabling rapid single-generation experiments and feasible experimental evolution. (ii) The mosquito’s eggs and the parasite’s spores can be stored for several months, which simplifies the logistics of experiments and enables to compare the evolved host and parasite with the original populations. (iii) The resource availability and the exposure of the host to the parasite are easy to control. (iv) The influence of the
Resource ecology of parasite evolution! 7/22
Impactofenvironmentonco-evolutionbetweenhostsandparasites
22
turity,reproductionandlongevity(ZellerandKoella2016)aswellas it influencesthedevelopmentof
theparasiteby increasing theproductionof infectious sporesofdeadmosquitoes (Bedhommeet al.
2004).Thewithingrowthoftheparasitemaythereforebelimitedbytheconditionofthehost,proba-
blybecauseofadirectinfluenceoftheenergyavailablefortheparasite(Halletal.2009b)andanindi-
rectimpactviathehost’slifehistory,speciallybodysize,thatconstrainsthedevelopmentofthepara-
site(AgnewandKoella1999b).Larvalfoodalsodeterminesthevirulenceoftheparasite:lessfooden-
hancesjuvenilemortality(Bedhommeetal.2004;LorenzandKoella2011)anddecreasesadultlifespan
ofinfectedmosquitoes(LorenzandKoella2011).Italsoconsiderablyaffectstheparasite’stransmission
route,eitherhorizontallyfromdeadjuvenilesorverticallyfromfemaleswhenthelaytheireggs.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
23
1.4 Researchaims
TheoverallobjectofthisPhDthesis istounderstandempiricallyhowtheenvironment influencesthe
co-evolutionarydynamicsbetweenthemosquitoAedesaegyptianditsmicrosporidianparasiteVavraia
culicis.Iaimtoempiricallytestassumptionsandpredictionsofmodelsofhostandparasiteslife-history
andepidemiology.Iusesinglegenerationexperimentsaswellasexperimentalevolutiontoaddressthe
openquestions.
Theprojecthasfourmajorgoals:
I. Experimentalexaminationofresourcevariabilityongrowth,reproductionandlongevityinthe
mosquitoAedesaegypti
II. Description of the parasites within-host dynamics; investigation of the relationship between
parasitedevelopmentandhost-longevityasafunctionofthehostsresourcesandageatinfec-
tion
III. Testtheoreticalpredictionsabouttheevolutionofhostdefenseagainstparasites.Investigation
of the roleof resourceavailabilityon theevolutionof thehost’sability to tolerateand resist
parasites.
IV. Examinationofresourceavailabilityandco-evolutionarydynamicsonthehostlife-historyevo-
lution.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
25
Effects of food variability onChapter2
growthandreproductionofAedesaegypti
MICHAELZELLER1,JACOBC.KOELLA1
1LaboratoryofEcologyandEpidemiologyofParasites,InstituteofBiology,UniversityofNeuchâtel,Rue
Emile-Argand11,2000Neuchâtel,Switzerland
PublishedinEcologyandEvolution:
Zeller,M.andKoella, J.C. (2016),Effectsof foodvariabilityongrowthandreproductionofAedesaegypti.Ecol
Evol,6:552–559.doi:10.1002/ece3.1888
Impactofenvironmentonco-evolutionbetweenhostsandparasites
26
Abstract
Despitea largebodyof knowledgeabout theevolutionof life-histories,weknow little about
howvariable foodavailabilityduringan individual’sdevelopmentaffect its life-history.Wemeasured
theeffectsofmanipulatingfoodlevelsduringearlyandlatelarvaldevelopmentofthemosquitoAedes
aegyptionitsgrowthrate,life-historyandreproductivesuccess.Switchingfromlowtohighfoodledto
compensatorygrowth:individualsgrewmorerapidlyduringlatelarvaldevelopmentandemergedata
sizeclosetothatofmosquitoesconsistentlyrearedathighfood.However,switchingtohighfoodhad
verylittleeffectonlongevity,andfecundityandreproductivesuccesswereconsiderablylowerthanin
consistentlywell fedmosquitoes.Changing fromhighto lowfood ledtoadultswithsimilarsizeas in
consistentlybadlynourishedmosquitoes,buteven lower fecundityandreproductivesuccess.Arapidresponseofgrowthtochangingresourcescanthushaveunexpectedeffectsinlaterlifeandinlifetime
reproductivesuccess.Moregenerally,ourstudyemphasizestheimportanceofvaryingdevelopmental
conditionsfortheevolutionarypressuresunderlyinglife-historyevolution.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
27
2.1 Introduction
Howlife-historiesrespondtovariationinfoodavailabilityisacentralquestionofevolutionaryecology.
Considerableeffort,bothwiththeoreticalandempiricalapproaches,hasbeenspentonansweringthe
questionforenvironmentsthatvaryspatially(KaweckiandStearns1993;Ernandeetal.2004)andfrom
onegenerationtothenext(Bashey2006)inresourceavailability.Yet,animportantaspectofvariability
has receivedconsiderably lessattention: that resource levelscanvaryduringan individual’sdevelop-
ment.
Eventhoughthere issubstantialevidencethatvariation in food levelsduringdevelopmentcanaffect
ageandsizeatmaturity(e.g.Leips&Travis1994;Hentschel&Emlet2000),weknowlittleabouthow
this variation affects reproductive success andadult survival.As food restriction severely affects life-
historyparameters - itgenerallyslowsgrowth,delaysmaturityand leads tosmalladultswith lowfe-
cundity (Stearns and Koella 1986) - it seems plausible that individuals that grow slowly early in life
shouldtrytomakeuptheirsizedeficitwithcompensatorygrowth,i.e.bygrowingmorerapidlyorfora
longerperiodoncetheyobtainmorefood (Dmitriew2011).Rapidlygrowing individuals,ontheother
hand,mightrespondtofoodstressbydeceleratinggrowthratesinordertousetheavailableresources
formaintenance and reproduction. Slow growth has shown to be adaptive for dealingwith nutrient
stress(Arendt1997).Compensatorygrowthfollowingaperiodofunfavorableenvironmentalconditions
hasbeendescribedformanyvertebratesandinvertebrates(Dmitriew2011).However,compensatory
growthneednotbeevolutionarilybeneficial.Indeed,thepresumedbenefitofcompensatorygrowth-
largerindividualshavegreaterfecundity-isnotalwaysobserved.InTrinidadianguppies,forexample,
compensatorygrowthisnotassociatedwithincreased,butwithdecreasedfecundity(Aueretal.2010).
Furthermore, anybenefit of compensatory growthwith regard to fecunditymaybe counteractedby
costswithregardtootherpartsofthelife-history.Longergrowthandthusdelayedmaturity,forexam-
ple,canbeassociatedwithagreaterriskofdyingbeforematurity(AbramsandRowe1996).Evenfor
purecompensatorygrowth,i.e.whenmaturityisnotdelayed,thegreatergrowthratemayhavecosts
(physiological/cellular level),whichareoftenonlyevidentmuch later in life (MetcalfeandMonaghan
2001;Alonso-Alvarezetal.2007;DeBlockandStoks2008).Indeed,dietrestrictionisoftenassociated
with a longer and healthier life (Chippindale et al. 1993; Masoro 2005). Accordingly compensatory
growthinfishreducedlifespanwhereasdeceleratedgrowthextendedit(Leeetal.2013).Thismay,in
part,beduetodevelopmentalerrorsandstructuralinstabilityasaresultofincreasedgrowth(Mangel
andMunch2005).Thus,althoughtheroleofre-feedingafteraperiodofdietaryrestriction(and,more
generally,theroleofchangingresourceavailabilityduringanindividuals’development)ontraitssuch
Impactofenvironmentonco-evolutionbetweenhostsandparasites
28
asgrowthrate, longevityandageatmaturityhaveacquiredsomeattention, little isknownabout its
roleonreproductivesuccess.
Inthisstudy,weprovidedataontheeffectofvariabilityindevelopmentalfoodconditions(leadingto
compensatory growth or decelerated growth) and associated changes in longevity and reproductive
successofthemosquitoAedesaegypti.Suchdatanotonlyformthebasisforourunderstandingoflife-
historyevolution,butarealsodirectlyrelevantforpublichealthduetomosquito’sroleasavectorof
severalinfectiousdiseases.
2.2 Materialandmethods
2.2.1 Experimentalsystem
WeusedtheUGALstrainofthemosquitoAe.aegypti(obtainedfromPatrickGuérin,UniversityofNeu-
châtel).Aedesaegyptioccurs throughout the tropicsandsubtropics.During theaquatic larval stages,
mosquitoesinnaturecanexperienceperiodsofnutrientrestrictionandcompetitionforresourceslike
bacteria,algaeandorganicmatter(ReiskindandLounibos2009).
2.2.2 Experimentaldesign
Theexperimentwasruninaclimatechambersetto26°C,70%relativehumidityandat12hlightand
12hdarkregime.Weuseda2x2factorialdesign,wherelarvaewerefedeitherwithastandardamount
of food (Day 1: 0.06mg of tetramin fish food, day 2: 0.08mg, day 3: 0.16mg, day4: 0.32mg, day 5:
0.64mg,day6orlater:0.32mg)orwithhalfofthestandarddietduringeitherearly(0to3daysafter
hatching)or latedevelopment (4ormoredaysafterhatching).The four treatmentsarehereafter re-
ferredtoasLL,LH,HHandHL,withthefirstletterreferringtotheamountoffoodduringearlydevel-
opment(LoworHigh)andthesecondlettertotheamountoffoodduringlatedevelopment.Eggswere
hatchedindeionizedwater.Fourhoursafterhatching,384first-instarlarvaeweremovedinto12-well
platesandkeptindividuallyin3mlofdeionizedwater.Eachlarvawashaphazardlyassignedtooneof
the four feeding regimes and fed every 24 hourswith the appropriate amount of food. Pupaewere
moved to300mlplastic cupscontainingdeionizedwaterandapieceof filterpaperasanoviposition
substrate.Thecupswerecoveredwithmosquitonetting,andcottonwoolmoistenedwith10%sugar
solutionwas placed onto the netting and changed every 48 hours.One day after emergence,males
werediscardedandeachfemalewasgivenamalechosenhaphazardlyfromourcolony.Thenextday
andeverytendaysthereafter, thefemalesweregiventheopportunitytotakeabloodmealonMZ’s
Impactofenvironmentonco-evolutionbetweenhostsandparasites
29
armforfiveminutes.Thefemaleswherecheckedeverydayforsurvival.Ninedaysafterbloodfeeding,
thefemaleswereplacedintofreshlypreparedplasticcupsandtheireggswereremovedandcounted.
Fecunditywasdefinedasthenumberofmelanisedeggs laiduptoninedaysafterbloodfeeding.The
experimentwas stoppedafter six roundsof egg-laying, atwhich time85.4%of themosquitoeshad
died.
2.2.3 Traitmeasurement
Weestimated larvalbodysizeby takingstandardizeddigitalpicturesofall individualsevery24hours
starting on the day of hatching (age 0) andmeasuring the length of the larvawith the open-access
softwareIMAGEJ.Whenphotosoflarvaewereconsideredtoolowinqualityforanaccuratemeasure-
menttobetaken,theindividualswerenotincludedintheanalyses.Larvalgrowthwasmeasuredasthe
differenceinsizebetweenage0andage4(earlygrowth)andbetweenage4andage6(lategrowth)
forallindividuals.Thesizeofadultswasassayedasthemeanoftheirwinglength,whichstronglycorre-
lateswiththeweightofmosquitoes(KoellaandLyimo1996)andiswidelyusedasanapproximationfor
adult size. The wings were removed and mounted on microscope slides. The slides were digitally
scannedandthewingsweremeasuredwithIMAGEJ
2.2.4 Statisticalanalysis
Weconsideredonly females, and ignored the growthof the6 (out of 384) individuals that haddied
beforepupation.Weassayed185femalemosquitoes,between43and49ineachfoodtreatment.The
difference insizebetweenage0andage4 (earlygrowth)wasevaluatedwithananalysisofvariance
(ANOVA) that includedthe levelofearly foodasa fixedbinomial factor.Becausethesizedifferences
betweentheages4and6(lategrowth)wereclosetolinearandindividualsnotyetreachedasymptotic
sizetheywereevaluatedwithananalysisofcovariance(ANCOVA)thatincludedearlyandlatefood,the
interactionbetweenthetwoasfixedfactors,andthesizeatagefourasacovariate.Assizeatagefour
didnot interactwithearlyor latefood,weomittedthese interactionsfromtheanalysis.Additionally,
becausewemeasuredindividualsrepeatedly,wecheckedthattheresultsweresimilar,whenwecor-
rectedforregressiontothemean(analysisnotshown).Forbothanalyses (earlyand lategrowth)we
verifiedthattheassumptionsofANOVAandrespectivelyANCOVAwerenotviolated.Ageatemergence
andlongevitywereanalyzedwithsurvivalanalysesthatincludedearlyandlatefoodandtheirinterac-
tionasfixedfactors. Intheanalysisof longevityweaddedwinglengthasapotentialconfounder.We
usedthedistributionsthatgavethebestfit,solog-logisticforageatemergenceandWeibullforlongev-
ity; using proportional hazards gave similar results (not shown). Wing length was analyzed with an
ANOVAthat includedearlyfoodand latefoodandtheir interactionasfixedfactors.Thewing lengths
Impactofenvironmentonco-evolutionbetweenhostsandparasites
30
wereBox-CoxtransformedtomeetANOVArequirements.Weanalyzedfecundity inthreeways.First,
weanalyzedtheproportionofblood-feedsthatledtoatleastoneeggwithaGLM(binomialdistribu-
tion).Second,weanalyzedthetotalnumberofeggslaidthroughouttheexperimentwithaGLMwith
quasi-Poissondistribution (corrected foroverdispersion). Inbothanalyses,we includedearlyand late
foodand their interactionas fixed factorsandwing lengthasapotential confounder.Third,weana-
lyzedtheage-specificclutchsizes(consideringonlythoseblood-feedsafterwhichatleastoneegghad
been laid)withamixedeffectANOVA,usingearly food, late food, clutchnumber (i.e.age)and their
interactionsasfixedfactors,winglengthasapotentialconfounder,andmosquitoasarandomeffect.
Wepresent theanalysis using all clutches.As thenumberofmosquitoes surviving to theendof the
experimentwaslow,weverifiedthattheresultsweresimilarifweconsideredonlythefirstthreeorthe
firstfourclutches(analysesnotshown).Themixed-effectANOVAwasdonewithRv.0.98.1056(RDe-
velopmentCoreTeam,2015)usingthelme4package;theotheranalysesweredonewithJMP12.0.0.
2.3 Results
2.3.1 Developmentaltraits
ThegrowthdataaresummarizedinFig.2.1.Larvaerearedonhighfoodgrewmorebetweenage0and
age4(mean=2.64mm,standarderror=0.072)thanthoserearedonlowfood(mean=1.86mm,se=
0.063)(F=66.86,p<0.001)(Fig.2.1).Growthafterage4decreasedwithincreasingsizeatday4(Table
2.1).Itwasgreatestformosquitoesthatswitchedfromlowtohighfoodatage4(2.21mm,se=0.120),
lowestformosquitoesthathadswitchedfromhightolowfood(1.57mm,se=0.118)andintermediate
formosquitoeswiththesamefoodlevelthroughouttheirdevelopment(Fig.2.2b).Theeffectsofearly
andoflatefood,butnottheinteractionbetweenthetwo,werestatisticallysignificant(Table2.1).
Impactofenvironmentonco-evolutionbetweenhostsandparasites
31
Table2:1Statisticalsummaryforjuveniletraits
ANCOVAfordifferencesinlategrowth,survivalanalysis(logLogisticdistribution)forageatemergenceandANOVAfordiffer-
encesinwinglength.
Lategrowth Ageatemergence Winglength
Factor df F SS p df χ2 p df F SS p
Earlyfood 1 5.07 1.77 0.026 1 173.6 <0.001 1 2.84 0.08 0.094
Latefood 1 14.87 5.18 <0.001 1 25.5 <0.001 1 41.21 0.02 <0.001
Earlyfood*latefood 1 0.24 0.08 0.63 1 6.6 0.01 1 0.13 <0.01 0.721
Sizeatage4 1 90.62 31.57 <0.001
Error 155 52.61 166 4.64
Figure2:1Mosquitogrowth
Bodylengthformosquitolarvaeasafunctionofage.Symbolsrepresentthemeanswitheachfoodtreatment,verticallinesthe
standarderrors.Trianglesrepresenttreatmentswithlowfoodavailabilityduringearlydevelopment;circlesrepresenttreat-
mentswithhighfoodavailabilityinearlydevelopment.Opensymbolsrepresenttreatmentswithlowfoodduringlatedevel-
opment;solidsymbolsrepresenthighfoodduringlatedevelopment.
12
34
5
Age of mosquito (days)
Larv
al g
row
th (m
m)
0 1 2 3 4 5 6 7 8 9 10
HHHLLHLL
Impactofenvironmentonco-evolutionbetweenhostsandparasites
32
Ageat emergence increased from9.9days (se=0.11) formosquitoes consistently fed thehigh food
levelto11.9days(se=0.09)formosquitoesconsistentlyfedthelowfoodlevel(Fig.2.2d).Mosquitoes
thathadswitched fromhigh to low foodemergedearlier (10.1±0.09) than those thathadswitched
fromlowtohighfood(11.2±0.10);theinteractionbetweenearlyandlatefoodlevelswasstatistically
significant(Table1,Fig.2.2c).
Winglengthincreasedfromameanof2.35mm(se=0.019)formosquitoesthathadbeenconsistently
rearedonlowfoodto2.55mm(se=0.027)formosquitoesthathadbeenconsistentlyrearedonhigh
food.Wing lengthwas influencedsignificantlybytheavailabilityof foodafterage4,whileearly food
andtheinteractionbetweenearlyandlatefoodhadnosignificanteffects(Table1,Fig.2.2d).
Figure2:2Juveniletraits
Theeffectoflarvalfoodduringearlyandlatestagesofdevelopmentfor(a)Earlygrowth(sizedifferencebetweenage0and
age4),(b)Meanlategrowth(sizedifferencebetweenday4andday6),(c)Ageatemergence±SE,(d)Adultsize(winglength).
Thedataforearlygrowth(a)waspooledforlatefoodtreatment.Symbolsrepresentthemeanswithintreatments;thevertical
linestheirstandarderrors.Opensymbolsrepresenttreatmentswithlowfoodduringlatedevelopment;solidsymbolsrepre-
senthighfoodduringlatedevelopment.
(a)
Early food
Early
gro
wth
(mm
)
Low High
1.5
2.0
2.5
3.0
(b)
Early food
Late
gro
wth
(mm
)
Low High
1.5
2.0
2.5
3.0
Late foodHighLow
(c)
Early food
Age
at e
mer
genc
e (d
ays)
Low High
1011
12
(d)
Early food
Win
g le
ngth
(mm
)
Low High
2.3
2.4
2.5
2.6
Impactofenvironmentonco-evolutionbetweenhostsandparasites
33
Table2:2Statisticalsummaryforadulttraits
Survivalanalyses(Weibulldistribution)forlongevity,binomialGLMfortheproportionofblood-fedsandGLM(quasi-Poisson
distribution)forthetotalnumberofeggs.
2.3.2 Adulttraits
Adult mosquitoes lived longest if they had been reared on low food throughout their development
(39.1days±1.97;thisandotheraveragesarebiased,fortheexperimentwasstoppedwhen14.6%of
themosquitoeswerestillalive),followedbythosethathadswitchedfromlowfoodtohighfoodwhen
theywerefourdaysold (36.8days±2.44). Incontrast tothesizeofadultmosquitoes, longevitywas
significantly affected by early food (Table 2.2),while late food and the interaction between the two
foodlevelshadnosignificanteffects.Winglengthhadnosignificanteffectonlongevity(Table2.2).
Thepercentageofthesixblood-feedsthatwerefollowedbylayingatleastoneeggrangedfrom0%to
100%;theaveragepercentagerangedfrom50%forthemosquitoesthathadbeenrearedonhighfood
throughouttheirdevelopmentto28%ifthemosquitoeshadswitchedfromhighfoodtolowfoodwhen
theywerefourdaysold(Fig.2.3B).About35%oftheblood-feedsledtoegg-laying,ifmosquitoeshad
initiallybeenrearedonlowfood,independentlyofthefoodavailabletothemduringtheirlatedevel-
opment (Table 2.2). Similarly, the total number of eggs was highest for mosquitoes that had been
rearedonhighfoodthroughouttheirdevelopment(67±8.1),lowestformosquitoesthathadswitched
food fromhigh to low(31±5.2)and intermediate formosquitoes thathadbeenrearedon lowfood
earlyintheirdevelopment(forLL:38±4.7;forLH:48±6.9)(Table2,Fig.2.3C).Latefoodenvironments
hadsignificanteffects indeterminingtheprobabilityof layingeggsandthetotalnumberofeggs.The
interaction between early food and late food had marginally significant effects in determining egg-
layingsuccessandmarginallynon-significanteffects indeterminingthetotalamountofeggs.Neither
theegg-layingsuccessnorthenumberofeggsweresignificantlyinfluencedbywinglength(Table2.2).
LongevityEgg-layingafterblood
feedingTotalnumberofeggs
Factor df χ2 p χ2 p χ2 p
Earlyfood 1 3.87 0.049 0.74 0.39 0.03 0.857
Latefood 1 0.22 0.636 4.58 0.032 11.79 <0.001
Earlyfood*latefood 1 <0.01 0.969 4.00 0.046 3.66 0.055
Winglength 1 1.17 0.279 0.14 0.712 0.06 0.803
Impactofenvironmentonco-evolutionbetweenhostsandparasites
34
Figure2:3Adulttraits
(a)Theeffectoflarvalfoodduringearlyandlatestagesofdevelopmentfor(a)Longevityofadultfemalemosquitoes(age0is
ageafteremergence).(b)Proportionofblood-fedsthatledtoegg-laying,(c)Totalnumberofeggs±SE.LLstandsforlowfood
availabilityduringthewholelarvaldevelopment,LHforlowfoodduringearlydevelopment,highfoodduringlatedevelop-
ment,HHforhighduringthewholedevelopmentandHLforhighfoodduringearlydevelopment,lowfoodduringlatedevel-
opment.In(b)and(c),opensymbolsrepresenttreatmentswithlowfoodduringlatedevelopment;solidsymbolsrepresent
highfoodduringlatedevelopment.
The clutch size (considering only those blood-feeds after which at least one egg had been laid) de-
creasedwiththeageofadultmosquitoes(Fig.2.4).Foodlevelduringlatelarvallifeaffectedthenum-
berofeggsinthefirstclutchandtherateatwhichfecunditydecreasedwithagewasinfluencedbythe
interactionbetweenearlyand late food treatment (Table2.3). Switching from low food tohigh food
four days after hatch led to themost eggs in the first clutch, but then to the greatest decline over
clutches(Fig.2.4).Therateofthedecreasewasmostlyinfluencedbytheinteractionbetweenearlyand
latefoodtreatments(Table2.3).
(a)
Age after emergence [days]
Prop
ortio
n su
rviv
ing
HHHLLHLL
0 10 20 30 40 50 60
0.0
0.2
0.4
0.6
0.8
1.0
(b)
Early food
Egg−
layi
ng a
fter b
lood−f
eedi
ngLow High
0.2
0.3
0.4
0.5
0.6
(c)
Early food
Tota
l num
ber o
f egg
s
Low High
2030
4050
6070
80 Late foodHighLow
Impactofenvironmentonco-evolutionbetweenhostsandparasites
35
Figure2:4Clutchsize
Relationshipbetweennumberofeggsperclutchandclutchnumber(i.e.,age).Circlesrepresenttreatmentswithhighfood
availabilityduringearlydevelopment;trianglesrepresenttreatmentswithlowfoodavailabilityinearlydevelopment.Open
symbolsrepresenttreatmentswithlowfoodduringlatedevelopment;solidsymbolsrepresenthighfoodduringlatedevelop-
ment.
Table2:3Repeatedmeasuresanalysisofclutchsizes
Onlyblood-feedingattempts,whichledtoatleastoneegg,wereconsidered.
Numberofeggs
Factor df F SS p
Earlyfood 1 0.02 2.8 0.885
Latefood 1 15.00 2005.0 <0.001
Earlyfood*latefood 1 3.67 490.6 0.056
Winglength 1 0.15 19.8 0.700
Clutchnumber 1 33.81 4518.5 <0.001
Clutchnumber*earlyfood 1 0.29 38.3 0.593
Clutchnumber*latefood 1 0.61 81.8 0.435
Clutch number * early food * late
food1 5.52 737.6 0.020
Error 267 133.6
Clutch number
Num
ber o
f egg
s
1 2 3 4 5 6
1020
3040
50
HHHLLHLL
Impactofenvironmentonco-evolutionbetweenhostsandparasites
36
2.4 Discussion
VariabilityindevelopmentalfoodconditionsinAedesaegyptihadqualitativelydifferenteffectsonthe
life-history traits we investigated: adult size, fecundity, survival and reproductive success. Thus, for
example,wing lengthwasdeterminedmainlyby foodavailabilityduring late larvaldevelopment, sur-
vivalbyfoodavailabilityduringearlydevelopment,andtotalnumberofeggsbyacombinationofthe
two.
When food availabilitywas held constant during themosquitoes’ development, their life-history fol-
lowedthegeneralpredictionsoflife-historytheory(e.g.StearnsandKoella,1986):lowfoodthusledto
slowgrowth, latepupation, smalladults,and low fecundity. Italsocorroboratesmanystudieswhere
food restriction increased longevity (Weindruch 1996; Shanley and Kirkwood 2000;Mair et al. 2003;
KirkwoodandShanley2005;Masoro2005).
Varyingfoodavailabilityledtolife-historiesthataremoredifficulttoexplainwithlife-history,similarly
tothestudyofYearsley,Kyriazakis&Gordon(2004).Increasingfromlowtohighfoodled,asfrequently
observed (Metcalfe andMonaghan 2001), to compensatory growth: at emergence,mosquitoes that
hadbeenfirstbadlyandthenwellnourishedcaughtupinsizebygrowingmorerapidlyandbydelaying
pupation,andtherebybecamealmostaslargeasmosquitoesthathadbeenfedwellthroughouttheir
development.However,althoughsizecaughtup,weobservednotoverylittlecatchingupoffecundity,
longevity,orlife-timereproductivesuccess.Togetherwiththeobservationthatthenumberofeggsper
clutchdeclinedstrongestwithagefor individualsthathadswitchedfromlowtohighfoodduringde-
velopment(Figure2.4), theseresultscouldmeanthatcompensatorygrowthearly in life isassociated
withreproductivecosts later in life,whichleadto, inour laboratoryconditions, lowerlife-timerepro-
ductivesuccess.Inadditiontoconsiderableevidencefortrade-offsbetweenlife-historytraitsearlyand
late in life,bothfromlaboratorysituations (e.g.Rose1984)and,morerecently, fromnaturalpopula-
tions (Lemaîtreetal.2015),our resultssupport the findingsofAueretal. (2010),whichsuggest that
thereare reproductivecostsassociatedwithcompensatorygrowth.The trade-offweobservedraises
thequestionabouttheadaptivenatureofcompensatorygrowth.However,althoughinourlaboratory
conditions,compensatorygrowthhadanegativeconsequence for reproductivesuccess, thesituation
maychangeinnaturalconditions.Bothjuvenileandadultmortalityratesmaybesubstantiallyhigherin
thefieldthaninthelaboratory.Accordinglythebenefitsof largersizeandearliermaturityassociated
withcompensatorygrowthmayoutweighitsreproductivecostsinoldmosquitoes.
Whenmosquitoesstartedoutatgoodfoodconditionsandthenswitchedtolowfood,theirgrowthand
adultsizedecreasedasexpected.Whatwasmoresurprisingwasthattheindividualswithdecelerated
Impactofenvironmentonco-evolutionbetweenhostsandparasites
37
growthhave lowerreproductivesuccessthanthosethathadexperiencedfoodrestrictionthroughout
theirdevelopment.However,becausetheinteractionbetweenearlyandlatefoodwasmarginallynot
significant,wecannotdrawstrongconclusions.Neverthelessthistrendcouldbetheresultofphysiolog-
icalresponsestothefoodenvironment inearlydevelopmentthatpreparethe individual forasimilar
environmentlaterinlife(GluckmanandHanson2004).Therefore,mosquitoesthatareundernourished
earlyinlifecancopewithfoodrestrictionlaterinlifebetterthanthosethathavebeenpreparedforan
environmentwithplentifulfood.
Astrikingresultwasthatwinglengthhadverylittleeffectonreproductionorlongevity,althoughasso-
ciations of life-history traitswith size are central tomany ideas in life-history theory (e.g. Stearns&
Koella1986;Rowe&Ludwig1991;Abrams&Rowe1996).Forexample,mostmodelsthatpredictthe
evolutionarilyoptimalageatmaturityassumethatfecundity increaseswithbodysize (e.g.Roff1984;
Stearns&Koella1986;Berrigan&Koella1994).Suchassociationsareoftenfoundwhenfoodavailabil-
ityisheldconstant(LyimoandTakken1993;McCannetal.2009).However,inourexperiment,where
food availability varies during themosquito’s development, the environmental factor over two time-
periods thatdeterminedbodysize (foodavailabilityduringearlyandduring latedevelopment)affect
the life-history traits rather than body size itself. If this is generally the case, it would imply major
changesinthewaywethinkaboutlife-historyevolution.Thetimingofresourcerestrictionduringde-
velopmentalsoaffecteditseffectonlongevity.Weobservedonlyaneffectiftherestrictionwasduring
earlydevelopment.Thisisconsistentwiththecommonfindingthatfoodrestrictioncanslowtheageing
process(Weindruch1996;ShanleyandKirkwood2000;Mairetal.2003;KirkwoodandShanley2005;
Masoro2005).However,thatchangingfromlowtohighorfromhightolowfoodhadnegligibleeffects
onlongevitycontradictsotherstudiesshowingthatcompensatorygrowthassociatedwithbetterfood
conditionsreversestheeffectofearlyresourcerestrictiononlongevity(Merry2002;Dhahbietal.2004;
Spindler2005).Wehavenoexplanationforthedifferenceoftheseresults.
2.4.1 Conclusion
Inconclusion,weshowedthatvariabilityofdevelopmental foodconditions inAedesaegyptimosqui-
toeshas strongeffectsonadult size, reproductivesuccessandmortalityofadult females,withsome
traitsbeingmostlyaffectedbythefoodavailability inearlydevelopmentandotherbeingaffectedby
late foodavailability.Sucheffectsmayhave importantconsequences forenergyallocationstrategies,
butaregenerallynotconsideredinmodeloflife-historyevolution.Wefurthershowedthatcompensa-
torygrowth,whichisgenerallyconsideredanadaptivestrategy,doesnotincreaseitsreproductivesuc-
cess,atleastforAedesaegyptiinourlaboratoryconditions.Thereproductiveburdensassociatedwith
compensatory growthmay play an important and limiting role in the evolution of growth and other
Impactofenvironmentonco-evolutionbetweenhostsandparasites
38
relatedtraits.Finally,thatthemosquitoes’reproductivesuccesswasnotdirectlyconnectedwithadult
size,butwas,rather,influencedbythefoodconditionsthattheyexperiencedduringdevelopmentcon-
trastsacentralassumptionofmanyideasinlifehistorytheory.Thus,wesuggestthatourunderstand-
ingof theevolutionof life-historieswillbegreatlyenhanced ifweconsider theeffectsofvaryingthe
environmentalconditionsduring juveniledevelopment.Such information is important inorder tode-
velopeffectivepredictionsofdiseasetransmissionandstrategiesofmosquitocontrol.
39
Context-dependent relationshipChapter3
between parasite growth and virulence in a
microsporidia-mosquitointeraction
MICHAELZELLER1,JACOBC.KOELLA1
1LaboratoryofEcologyandEpidemiologyofParasites,InstituteofBiology,UniversityofNeuchâtel,Rue
Emile-Argand11,2000Neuchâtel,Switzerland
Impactofenvironmentonco-evolutionbetweenhostsandparasites
Impactofenvironmentonco-evolutionbetweenhostsandparasites
40
Abstract
Manyideasabouttheevolutionofparasitesrelyontheassumptionthattheirhostsincreasingly
suffer as parasite load increases, and that the relationship between growth and virulence is similar
among environments.We investigatedwhether the interaction between the parasiteVavraia culicis
anditshost,themosquitoAedesaegypti,isaffectedbythreeaspectsoftheenvironment:theamount
of foodavailableto larvaeandtoadults,andtheageat infection.Wemeasuredspore loadandesti-
matedtheparasite’sgrowthrateandasymptotic load inmosquitoesthatdiedduringtheexperiment
and in haphazardly selected, living mosquitoes. In most environments, the probability of infection,
spore load, bothmeasures of the parasite’s developmentwere higher in dead than in age-matched
living mosquitoes, corroborating the idea that virulence increases with the parasite’s development.
However, these relationshipsdependedon the foodavailability andage at infection, suggesting that
thetrade-offsunderlyingtheevolutionofvirulencedependontheenvironment.Ideasabouttheevolu-
tionofvirulencemustthereforeconsidernotonlyhowtheenvironmentaffectsepidemiologicallyrele-
vantparameters,butalsohowitaffectstherelationshipsbetweenthem.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
41
3.1 Introduction
Whydosomeparasitescauseseverepathologyandmortalitywhileothersarerelativelybenign?Many
theoreticalattempts tounderstand thisvariation relyon theassumption that there isa trade-offbe-
tweentherateoftransmissionandthedurationofthetransmissionperiod(AndersonandMay1979;
Alizonand Lion2011,Gandonet al. 2001;Dieckmannet al. 2002).Underlying this assumption is the
ideathatparasitesmustreplicaterapidlytoahighloadinordertobetransmittedefficiently,butthat
highparasite load(orrapidparasitedevelopment)alsoreducesthehost’ssurvivalandthusthedura-
tionofthetransmissionperiod.
Manyaspects,however,of the relationshipbetweenparasite loadand thehost’sdeathareobscure.
Whileseveralexamples indeedshowthatgreaterparasitemiacan increasetheriskofdeath(e.g.HIV
(Mellors et al. 1996), rodentmalaria (Mackinnon and Read 1999;Mackinnon and Read 2004) and a
gregarine parasite ofmonarch butterflies (De Roode et al. 2008; De Roode and Altizer 2010),many
shownorelationship (Salvaudonetal.2007;Littleetal.2008). Indeed,wemightexpect that thebe-
tweenparasiteburdenandhostmortalityisweakforparasiteswhosesymptomsaremainlyduetothe
host’simmuneresponse,sothatlowparasiteloadsmaybeasharmfulashighloads.Forexample,clini-
calmalariaand,inparticularcerebralmalaria,appeartosteminlargepartfromanoveractiveimmune
system(Artavanis-Tsakonasetal.2003).
Another,lessexploredreasonforthevariabilityintheassociationbetweenparasitedevelopmentand
hostmortality is the role of the environment on the association. The environment, in particular re-
sourcescanaffectvirulence(Jokelaetal.1999;Brownetal.2000;FergusonandRead2002);resources
canalsoaffectparasitegrowth(Bedhommeetal.2004;Johnsonetal.2007;deRoodeetal.2008;Hall
etal.2009a).Ifvirulenceandgrowthareaffecteddifferently,wemightseeanegativerelationshipbe-
tweenparasitedevelopmentandsurvivalinsomeenvironments,butnotinothers.Indeed,inanexper-
iment withDaphnia magna and its intestinal, castrating parasite Pasteuria ramose, the relationship
betweenparasiteloadandmortalityrateswitchedfrompositiveatlowfoodconditionstonegativeat
highones(Valeetal.2011).Suchstudiescanbe,however,difficulttointerpret,astheymeasurepara-
siteloadatthetimeofthehost’sdeath.Asparasiteloadisexpectedtoincreasewithtimeafterinfec-
tion,itisintrinsicallyrelatedtolongevityandthusmortalityrate.Inaddition,parasiteloadinsurviving
individuals is often notmeasured, so thatwe do not knowwhether dead individuals indeed harbor
moreparasites than survivingones. Yet this is a critical assumption ifwewant to conclude anything
aboutarelationshipbetweenparasiteloadanddeath.Iftheenvironmentindeedchangestherelation-
ship between the parasite’s growth and virulence, testing ideas about the evolution of virulence by
changing environmental parameters (as, for example, (Ebert andMangin 1997; Nidelet et al. 2009))
Impactofenvironmentonco-evolutionbetweenhostsandparasites
42
wouldbeproblematic,asthedifferenttreatmentswoulddifferinthetrade-offsunderlyingtheevolu-
tionarypredictions.
Here,westudywhethertheenvironment(resourceavailabilityandageatinfection)affects,inaddition
tohosthealthandparasitedevelopment(growthandparasiteload),therelationshipbetweenthetwo,
usingexperimentalinfectionsofthemosquito,AedesaegyptiwiththemicrosporidianparasiteVavraia
culicis.Wecomparethenumberofspores indeadmosquitoesand inasampleofage-matched living
mosquitoes.Thisallowsustodecoupletheeffectofparasite loadfromthe longevityofthehost.We
furthermodelthedevelopmentoftheparasitewithinthelivingandthedeadmosquitoesandcompare
the growth parameters between the two groups. Finally, bymanipulate age at infection, larval and
adultresources,wefindtheroleoftheenvironmentontherelationshipbetweenparasitegrowthand
hostmortality.
3.2 Materialandmethods
3.2.1 Experimentalsystem
ThemicrosporidianparasiteVavraiaculiciswasprovidedbyJ.JBecnel(USDA,Gainesville,USA).Itisan
obligate,intracellularparasitethathasbeenreportedinnaturalpopulationsofseveralgeneraofmos-
quitoesincludingAedes(WeiserandColuzzi1972).ThehostlarvaeingestthesporesofVavraiaculicis
withtheirfood,resultingininfectionofgutcellsandepithelialcells.Afterseveralroundsofreplication
withinthelarvaetheparasitebeginstoproduceitsinfectiousspores.Theparasiteistransmittedintwo
ways.First,transmissioncanoccurhorizontallyfromlarvatolarva,eitherbysporesreleasedinfaeces
orafter larvaldeath. Insomecases,e.g.with foodstressorheavy infection(Bedhommeetal.2004),
larvaldeath isenhanced.Second, if juveniles survive the infection, theydevelop into infectedadults.
Thesporesaretransmittedfromfemalestothenextgenerationofmosquitoesbyarrivinginabreeding
siteeitherwhenamosquitodiesonabreedingsiteoronthesurfaceoftheeggs(Andreadis2007).We
usedtheUGALstrainofthemosquitoAe.aegypti(obtainedfromPatrickGuérin,UniversityofNeuchâ-
tel).Aedesaegyptioccursthroughoutthetropicsandsubtropics,isanimportantvectorfor,e.g.dengue
andZikaviruses.
3.2.2 Experimentaldesign
Theexperimentwasruninaclimatechambersetto26°C,70%relativehumidityandat12hlightand
12hdarkregime.Weusedafullfactorialdesign,withageatinfection(twoorfivedaysafterhatching),
Impactofenvironmentonco-evolutionbetweenhostsandparasites
43
larval food (the standarddietofour lab (Day1: 0.06mgof tetramin fish food,day2: 0.08mg,day3:
0.16mg,day4:0.32mg,day5:0.64mg,day6orlater:0.32mg)or40%ofthestandarddiet)andadult
food(10%or2%sucrose-solution)asexperimentalfactors.
Eggsweresoakedindeionizedwaterandsimultaneouslyhatchedunderreducedatmosphericpressure.
2010first-instarlarvaeweremovedinto12-wellplatesandkeptindividuallyin3mlofdeionizedwater.
Each larvawashaphazardlyassignedtooneof theeight treatmentsand fedevery24hourswith the
appropriateamountoffood.Weexposedlarvaetoinfectionbyadding2.0x105sporesin100µldeion-
izedwaterper individual.Pupaeweremoved to50mlFalcon tubecontainingdeionizedwateranda
pieceoffilterpaper.Thecupswerecoveredwithmosquitonetting,andmosquitoesweregivenaccess
tocottonwoolmoistenedwitheither10%or2%sugarsolution.Toestimatetheparasite’sgrowth,we
countedsporesfromall thedead individualsand,startingelevendaysafterhatching, fromahaphaz-
ardly selected sample (8-12 individuals) of livingmosquitoesof each treatment. Theexperimentwas
stoppedwhenallofthemosquitoeshaddied(32daysafterhatching).
3.2.3 Traitmeasurement
The size of adultswas assayed as themean of theirwing length,which strongly correlateswith the
weightofmosquitoes (KoellaandLyimo1996)and iswidelyusedasanapproximation foradult size.
Thewingswereremovedandmountedonmicroscopeslides.Theslidesweredigitallyscannedandthe
wingsweremeasuredwiththeopen-accesssoftwareIMAGEJ.Tocounttheparasite’ssporeswerewe
placedeachmosquito intoa2mlEppendorf tubecontaining180µldeionizedwateranda5mmsteal
bead.We crushed themosquitoesby shaking the tube for4minutes at 35Hz (Tissue Lyser,Qiagen,
Valancia,California).Thestealbeadwasremovedandthesporeswerecountedinasampleofthesolu-
tionwithahaemocytometer(Neubauerimproved).
3.2.4 Statisticalanalysis
Weassayed2014mosquitoes (between248 and253 in each treatment) ofwhich925becameadult
females (between 61 and 131 in each treatment). For the analysis of the juvenilemortality the two
adult food treatmentswerepooled.Ageneralized linearmodel (GLM) fittedwithabinary logistic re-
gressionwasperformedtodeterminetheeffectsoflarvalfood,ageatinfectionandtheirinteractionon
the likelihoodthat larvaesurvivedtoemergeasadults.Weanalyzedtheproportionsof individuals in
whichwefoundatleastonesporewithaGLM(binomialdistribution),andthenumberofsporeswitha
GLMwithquasi-Poissondistribution(correctedforoverdispersion).Forbothanalysesweincludedlar-
valfood,adultfood,ageatinfectionandsurvival(deadoraliveattimeofsampling)andtheirinterac-
tionasfixedfactorsandthetimeafterinfectionasanominalconfounder.Notethatwedonotconsider
Impactofenvironmentonco-evolutionbetweenhostsandparasites
44
timeafterinfectioncontinuousbecauseofitsstronglynonlinearrelationshipwithsporeload.Parasite
sporeloadswerelog10-transfromed.
Wefittedthenumberofsporesusinggeneralizedleastsquareswiththefollowingequation
! = ! ∗ (1 − !!!∗!)
Equation3:1-Parasitegrowthequation
wheren represents the log10-transformednumberof spores,c theasymptoticnumberof spores,k a
parameterrelatedtothegrowthrateatlowsporeloadsandtthetimeafterinfection.Wecomparedc
andkbetweenexperimentalgroupswiththegnlsfunctionfromthenlmepackageincludinglarvalfood,
adultfood,ageatinfection,survivalandalltheirinteractionsasfixedfactors.Thelongevityofthemos-
quitoeswasanalyzedwithasurvivalanalysisthatincludedlarvalfood,adultfood,infectiontime-point
andallpossibleinteractionsandaddedwinglengthasapotentialconfounder.WhileweusedWeibull
distributions because it gave the best fit, Cox proportional hazard gave similar results. All statistical
analyseswereperformedwithRversion3.2.3(RDevelopmentCoreTeam,2015).
3.3 Results
3.3.1 Juvenilemortality:
Theprobabilitythata juvenilediedbeforeemergencerangedfrom1.8%(confidence intervalofpro-
portion:lowerlimit=0.95%,upperlimit=3.4%)formosquitoesthathadbeeninfectedfivedaysafter
hatchingandrearedonhighlevelsoffoodto33.2%(CI:lowerlimit=29.1%,upperlimit=37.6%)for
mosquitoesthathadbeeninfectedatagetwoandrearedonlowfood.Mosquitoesthathadbeenin-
fectedatagefiveandrearedonlowfood(2.6%;CI:lowerlimit=1.5%,upperlimit=4.4%)andatage
twoandrearedonhighfood(13.5%CI: lower limit=10.7%,upper limit=16.7%)had intermediate
risksofjuveniledeath.Larvalfood(χ2=17.9,p<0.001)andageatinfection(χ2=15.9,p<0.001),but
nottheirinteraction(χ2=3.02,p=0.082),hadsignificanteffectsinfluencingjuvenilemortality.
3.3.2 Probabilityofinfection,sporeloadandparasitegrowth:
The probability of harboring spores significantly increased with infection period and was higher for
mosquitoeswithayoungageat infection(Table3.1,Figure3.1A).Theanalysisofsporeloadrevealed
statisticallysignificanteffectsoftimeafterinfection,survival,larvalfood,ageatinfectionandtheinter-
actionbetweenlarvalfoodandageatinfection(Table3.1,Figure3.2B).Whenmosquitoesdiednatural-
Impactofenvironmentonco-evolutionbetweenhostsandparasites
45
lytheygenerallyhadmoresporesthanlivingonesatthesameage.Thesporeloadofmosquitoessignif-
icantlyincreasedwithtimeafterinfection,andwashigherforindividualswithhighlarvalfoodandwith
ayoungageatinfection.
Figure3:1:Sporeloadandprobabilityofinfection
Theeffectoflarvalfoodavailabilityandageatinfectionfor(A)theprobabilitytofindspores(mean±confidenceintervalof
proportion),for(B)thelognumberofspores(mean±se)andfor(C)themeandifferenceofsporeloadbetweennaturally
deadandlivingmosquitoes.Plottedare,forcolumnIindividualswithhighadultfoodavailabilityandforcolumnIIindividuals
withlowfoodadultavailability.Solidsymbolsrepresenttreatmentsalowageatinfection(agetwo);opensymbolsrepresent
treatmentswithanageatinfectionoffive.Circlesrepresenttreatmentswithhighfoodavailabilityduringdevelopment;
squaresrepresenttreatmentswithlowfoodavailabilityduringdevelopment.ThedashedlinesinfiguresCindicatetheexpec-
tationifsporeloadwouldbeequalfornaturallyandartificiallydyingmosquitoes.
Days after infection
Spor
e pr
obab
ility
6 9 12 15 18 21 24
0.0
0.2
0.4
0.6
0.8
1.0A
I
Days after infection
Spor
e pr
obab
ility
6 9 12 15 18 21 240.0
0.2
0.4
0.6
0.8
1.0
II
Days after infection
Spor
e lo
ad
6 9 12 15 18 21 24
3.5
4.0
4.5
5.0B
Days after infection
Spor
e lo
ad
6 9 12 15 18 21 24
3.5
4.0
4.5
5.0
Days after infection
Diff
. in
spor
e lo
ad
6 9 12 15 18 21 24
−1.0
−0.5
0.0
0.5
1.0
(d)
C
Days after infection6 9 12 15 18 21 24
−1.0
−0.5
0.0
0.5
1.0
Impactofenvironmentonco-evolutionbetweenhostsandparasites
46
Table3:1Sporeloadandprobabilityofinfection
Statisticalsummaryfortheprobabilityofinfection(GLMwithbinomialdistribution)andsporeload(GLMwithPoissondistri-
butioncorrectedforoverdispersion).Statisticallysignificantp-valuesaregiveninbold.
Sporeprobability Sporeload
Factor df χ2 p χ2 p
Survival 1 2.40 0.121 4.89 0.027
Larvalfood 1 1.49 0.223 7.34 0.006
Adultfood 1 1.39 0.239 1.35 0.245
Ageatinfection 1 7.51 0.006 12.42 <0.001
Survival*larvalfood 1 0.86 0.354 0.23 0.635
Survival*adultfood 1 0.58 0.446 1.19 0.276
Larvalfood*adultfood 1 2.44 0.118 0.08 0.779
Survival*ageatinfection 1 0.73 0.393 1.08 0.298
Larvalfood*ageatinfection 1 1.92 0.166 10.17 0.001
Adultfood*ageatinfection 1 1.00 0.318 0.61 0.437
Survival*larvalfood*adultfood 1 0.51 0.477 0.29 0.591
Survival*larvalfood*ageatinfection 1 0.62 0.431 0.61 0.433
Survival*adultfood*ageatinfection 1 1.32 0.250 0.78 0.378
Larvalfood*adultfood*ageatinfection 1 3.04 0.081 3.02 0.082
Timeafterinfection 1 54.64 <0.001 60.71 <0.001
Impactofenvironmentonco-evolutionbetweenhostsandparasites
47
Thegrowthparameterkandtheasymptoticsporenumbercweresignificantlyhigherfornaturallydy-
ingmosquitoesthanfor livingones(Figure3.2).Whilenoneoftheenvironmentalfactorssignificantly
influencedk,theasymptoticnumberofsporescwassignificantlyaffectedbylarvalfood,ageatinfec-
tion,adultfoodandbysomeinteractionsbetweenthefactors(summarizedinTable3.2).Inparticular,
theasymptoticsporenumberwashigherformosquitoeswithalowageatinfectionandformosquitoes
rearedonhighlevelsoffoodaslarvae.
Figure3:2Parasitegrowthparameters
Theeffectoflarvalfoodavailabilityandageatinfectionfortheasymptoticsporeloadc(A)andthegrowthparameterk(B)±
standarderrorfor(I)individualswithhighadultfoodavailabilityandfor(II)individualswithlowadultfoodavailability.Solid
symbolsrepresenttreatmentswithalowageatinfection(agetwo);opensymbolsrepresenttreatmentswithanageatinfec-
tionoffive.Circlesrepresenttreatmentswithhighfoodavailabilityaslarvae;squaresrepresenttreatmentswithlowlarval
foodavailability.
cause of death
c3.5
4.0
4.5
5.0
5.5
AI
cause of death
c3.5
4.0
4.5
5.0
5.5 II
cause of death
k
artififcial natural
0.0
0.1
0.2
0.3
0.4
0.5B
cause of death
k
artififcial natural
0.0
0.1
0.2
0.3
0.4
0.5
Impactofenvironmentonco-evolutionbetweenhostsandparasites
48
Table3:2Parasitegrowthparameters
Statisticalsummaryforthecomparisonofk(growthparameter)andc(asymptoticsporeload)betweenexperimentalgroups.
Statisticallysignificantp-valuesaregiveninbold.
Growthparameter Asymptoticsporeload
Factor df F p F p
Survival 1 68.533 <0.001 11260.34 <0.001
Larvalfood 1 0.059 0.8086 12.745 <0.001
Adultfood 1 0.110 0.7398 4.325 0.0382
Ageatinfection 1 0.384 0.5356 16.900 <0.001
Survival*larvalfood 1 <0.01 0.9894 2.136 0.1447
Survival*adultfood 1 0.128 0.7203 9.670 0.002
Larvalfood*adultfood 1 0.071 0.7904 0.012 0.9115
Survival*ageatinfection 1 0.251 0.6163 13.574 <0.001
Larvalfood*ageatinfection 1 0.039 0.8434 26.167 <0.001
Adultfood*ageatinfection 1 0.381 0.5376 2.503 0.1143
Survival*larvalfood*adultfood 1 0.063 0.8020 0.053 0.8173
Survival*larvalfood*ageatinfection 1 0.450 0.5026 3.499 0.0621
Survival*adultfood*ageatinfection 1 0.086 0.7689 14.462 <0.001
Larvalfood*adultfood*ageatinfection 1 0.805 0.3702 7.070 0.0081
Survival*Larvalfood*adultfood*ageatinfection 1 0.073 0.7870 14.819 <0.001
3.3.3 Longevity:
Adultmosquitoeslivedlongestiftheyhadbeeninfectedatagefive,rearedaslarvaeonhighfoodand
asadultsonhighfood(13.8days;se=0.70),andtheylivedshortestiftheyhadbeeninfectedatage2
andrearedonlowfoodaslarvaeandasadults(3.4days;se=0.60).Theeffectsofageatinfection,the
interaction between larval and adult food, and the three-way interaction between larval food, adult
foodandageat infectionwerestatistically significant (Table3.3).Completedataon longevitycanbe
obtainedfromFigure3.3.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
49
Figure3:3Longevityofinfectedmosquitoes
TheeffectsoflarvalfoodandageatinfectionforlongevityofadultfemalesforI)individualswithalowadultfoodavailability
andII)forindividualswithhighadultfoodavailability(age0isageafteremergence).Darkblacklinesrepresenttreatments
withhighfoodavailabilityaslarvae;greylinesrepresenttreatmentswithlowfoodavailabilityaslarvae.Solidlinesrepresent
treatmentswithalowageatinfection(age2);dashedlinesrepresentstreatmentswithhighageatinfection(age5).
Table3:3Statisticalsummaryofsurivalanalysis
Statisticalsummaryforthesurvivalanalysis(Weibulldistribution).Statisticallysignificantp-valuesaregiveninbold.
Longevity
Factor df χ2 p
Larvalfood 1 0.44 0.509
Adultfood 1 0.44 0.509
Ageatinfection 1 7.49 0.006
Larvalfood*adultfood 1 38.67 <0.001
Larvalfood*ageatinfection 1 0.09 0.769
Adultfood*Ageatinfection 1 0.34 0.561
Larvalfood*Adultfood*ageatinfection 1 20.98 <0.001
Winglength 1 0.41 0.522
3.4 Discussion
The foodofAedesaegyptiand theageat infectiongreatlyaltered thedynamicsof thehostparasite
interactionby influencing theprobabilityof infection thegrowthof theparasiteand the longevityof
thehost.Theprobabilityofinfectionwasmainlydeterminedbyageatinfection,thesporeloadbylar-
0 5 10 15 20
0.0
0.2
0.4
0.6
0.8
1.0
Days after emergence
Prop
ortio
n su
rviv
ing
I
0 5 10 15 20
0.0
0.2
0.4
0.6
0.8
1.0
Days after emergencePr
opor
tion
surv
ivin
g
II
Impactofenvironmentonco-evolutionbetweenhostsandparasites
50
valfoodandbyageatinfectionandparasitegrowthwasinfluencedbylarvalfood,adultfoodandage
atinfection.Thelongevityofmosquitoeswasmostlydeterminedbyageatinfectionandbyinteractions
betweenlarvalfoodadultfoodandhostageatinfection.Furthermore,theprobabilityofinfection,the
spore loadand the sporegrowthwas significantlyhigher fornaturallydyingmosquitos than for age-
matchedlivingones,indicatingthatthesetraitsareimportantcausesofvirulence.
Inoursystem,virulencehastwocomponents:theprobabilitythatinfectedindividualsdiebeforeemer-
genceandthelongevityofinfectedadults.Bothcomponentswerehigherifthelarvaewereyoungerat
infectionand if theyobtained less food.This isconsistentwith the literature that less foodenhances
juvenilemortality(Bedhommeetal.2004;LorenzandKoella2011)anddecreasesadultlifespanofin-
fectedmosquitoes (Lorenz andKoella 2011), and it suggests early infection amplifies thesepatterns.
Althoughthemechanismsunderlyingvirulencearenotknown,theymayincludeadirectanddensity-
dependent effect of the parasite (as assumed inmanymodels (Ganusov et al. 2002; Alizon and van
Baalen2005b))orthedepletionofenergybelowathresholdnecessaryforthehost’ssurvival(Halletal.
2009b),whichislikelytodependnotonlyonparasiteloadbutalsoontheperiodofinfection.
We were particularly interested in whether the environment affected the way that the parasite’s
growth and density influenced the severity of parasitism, in particular its host’s longevity. Naturally
dyingmosquitoesgenerally showedahigher spore load (Figure3.1C,Table3.1), ahighergrowthpa-
rameteroftheparasiteandahigherasymptoticsporeloadthanlivingonessampledatthesameage
(Figure3.2,Table3.2).Thus,ourdataagenerallypositiverelationshipbetweentheparasitesdevelop-
mentanditsvirulence, inaccordancewiththecentralassumptionofthevirulencetransmissiontrade
off(AndersonandMay1982).Howeverwefoundthattherelationshipbetweenthehost’shealthand
theparasite’sdevelopmentvariedbetweenagesatinfectionandamongfoodlevels.Thustheenviron-
mentalfactorsinfluencedtheasymptoticsporedensity(c)differentlyinlivingandnaturallydyingindi-
viduals(significantinteractionsbetweensurvivalandadultfoodandbetweensurvivalandageatinfec-
tion,Table3.2).
Ageatinfectionandfoodavailabilityalsointerfereinhowtheyexpresstheparasite’sexploitationand
virulence.Thus,whichenvironmental factorsare importantdiffer for theparasite’sdevelopmentand
forthehost’slongevity.Thus,whileyoungageatinfectionledtohighersporeloadsandalsoreduced
adult longevity, mosquitoes reared on ample larval food producedmore spores (higher growth and
parasiteload)thanmosquitoesrearedonlittlefood,buthadverysimilaradultlongevity;ortheinterac-
tionbetweenlarvalfoodandadultfoodavailabilityhadlittleinfluenceontheproductionofspores,but
butalargeeffectonlongevity.Itthereforeseemsthatwhenlarvalfoodavailabilityishightheparasite
growswithlittleimpactonthehost’slongevity,butwhenageatinfectionislowthehostismorevul-
Impactofenvironmentonco-evolutionbetweenhostsandparasites
51
nerabletohighparasiteloads.Thefactthathostsrearedonhighlevelsoffoodseemtobemoretoler-
ant tohighparasiteburdencanbe interpretedas reducedconflict for shared resourcesunderample
foodandisconsistentwiththefindingsfromValeetal.(2011)andZellerandKoella(underreview).If
weassumethattransmissionislinkedtosporeloadsimilarlyinallenvironments(whichis,ofcoursenot
necessarily the case), our results imply that the virulence-transmission trade-off differs among envi-
ronments,withimportantconsequencesforpredictionsabouttheevolutionofvirulence.
Two general mechanisms can explain the effect of resource availability on the outcome of a host-
parasite interaction.Ontheonehandtheampleresourcescan increasetheefficacyof thehost’s im-
muneresponseandtherebyincreasetheresistanceofhoststoparasite infection (KoellaandL.2002;
AyresandSchneider2009).Malnourishedhostsmaythereforebeweakerandmoresusceptibletoin-
fectiousdisease(Moret2000)sothatthelowerthefoodavailability,thehigherthecostsofparasitism
(FergusonandRead2002).Ontheotherhand,theparasitesusetheirhost’sinternalresourcesfortheir
owndevelopment; resources that arenormally allocated to thehost’s growth,maintenance, and re-
production.Accordinglyhostsrearedonhighlevelsoffoodmaypresentabetterenvironmentforthe
parasitetodevelop(AgnewandKoella1999a;Brownetal.2000;Bedhommeetal.2004;Tseng2006).
Becauseinourstudytheparasiteloadwashigherforindividualswithaccesstolotsoffoodandwhen
theywereinfectedatayoungage,itsuggeststhattheparasitewasabletobenefitfromthehost’sam-
plefoodenvironmentandprobablydirectlyaccessedthehostresourcestoproducesitsspores.
3.4.1 Conclusion
Weshowed that the susceptibility, parasite loadand theparasite’swithin-host growthandvirulence
arecomplex,ageandresource-dependenttraits.Thereforethecorrelationamongagesandfoodlevels
betweenparasitedevelopment and virulence canhave considerable impacton theevolutionaryout-
comeof infectiousdisease.Theecologicalconditionsandtheageat infectionofhostshavetherefore
thepotential to change the relative costsandbenefitsofparasite replicationandare likely todeter-
minetheadaptivelevelsofvirulence.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
53
The role of the environment onChapter4
theevolutionoftoleranceandresistancetoa
pathogen
MICHAELZELLER1,JACOBC.KOELLA1
1LaboratoryofEcologyandEpidemiologyofParasites,InstituteofBiology,UniversityofNeuchâtel,Rue
Emile-Argand11,2000Neuchâtel,Switzerland
Impactofenvironmentonco-evolutionbetweenhostsandparasites
54
Abstract
Defenseagainstparasitescanbedividedintoresistance,whichlimitsparasiteburden,andtol-
erance,which reducespathogenesis at a givenparasite burden.Distinguishingbetween the twoand
understandingwhichdefenseisfavoredbyevolutionindifferentecologicalsettingsareimportant,for
theyleadtofundamentallydifferentevolutionarytrajectoriesofhost-parasiteinteractions.Weletthe
mosquitoAedesaegyptievolveunderdifferentfoodlevelsandeitherwithnoparasite,withaconstant
parasite,orwitha coevolvingparasite (themicrosporidianVavraia culicis).We then tested tolerance
and resistanceof theevolved linesat the two food levels. Exposure toparasitesduringevolution in-
creased resistance and tolerance, but therewere no differences between the lines evolvedwith co-
evolvingorconstantparasites.Mosquitoesthathadevolvedwithfoodrestrictionhadhigherresistance
thanthoseevolvedwithhighfood,butsimilartolerance.Themosquitoesthathadrestrictedfoodwhen
being tested had lower tolerance than those with normal food, but there was no difference in re-
sistance.Our results emphasize the complexity and dependence on environmental conditions of the
evolution and expression of resistance and tolerance, and help to evaluate some of the predictions
abouttheevolutionofhostdefenseagainstparasites.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
55
4.1 Introduction
Defensemechanismsagainstparasitescanbedividedintotwobroadclasses:resistanceandtolerance.
Resistancereducestheharmcausedbydiseasebypreventinginfectionorlimitingtheparasite’sdevel-
opmentwithin thehost, thus leading to lowerparasite loads. Tolerance reduces theparasite’sdetri-
mentaleffectswithoutaffectingparasiteload(Readetal.2008;Råbergetal.2009;Littleetal.2010).
Whileresistanceandtolerancebothincreasethefitnessofaninfectedindividual,resistancedoessoby
reducing the parasite’s fitness, whereas tolerance does not. Whether hosts evolve tolerance or re-
sistancehas fundamentallydifferenteffects for theevolutionary trajectoryofhostpathogen interac-
tions(RoyandKirchner2000;RestifandKoella2003;Boots2008).Severaltheoreticalstudieshavepre-
dictedhowtoleranceandresistancemightevolve.Forexampleitwaspredicted(i)thathostsmaintain
geneticpolymorphismforresistance,butnotfortolerance(RoyandKirchner2000;Milleretal.2005).
Theevolutionofresistancewouldprovokecounter-adaptationoftheparasitetoovercomeresistance,
whichwould lead to co-evolutionary dynamicswith rapidly changing allele-frequencies in resistance-
genes.Incontrast,theevolutionoftolerancewouldbenefitbothhostandparasite,soenablingitsfixa-
tion.Thus,(ii)tolerancewouldincreasemoreeasilythanresistance(RoyandKirchner2000).However,
(iii) theparasites couldalso respond to theevolutionof toleranceby increasing their growth to take
advantageoftheweakerconstraint(Milleretal.2005).Thus,tolerancewouldonlybeobservedifthe
hostsareinfectedwithparasitesthathavenotco-evolvedwiththehost.Finally,(iv)parasiteswithlow
virulencecouldbemore likely toselect for theirhosts’ tolerance,whereashighvirulencecould favor
resistance(RestifandKoella2004).
While this short list emphasizes thatwe expect and can predict different evolutionary outcomes for
resistanceandfortolerance,theseideasrelyontheideathatthetwodefensestrategiesarenotlinked,
asforexampleobservedinnaturalpopulationsofmonarchbutterfliesand(Lefèvreetal.2010)anda
cyprinidfish(Mazé-Guilmoetal.2014)andtheirparasites.Inothersystems,however,evolutionofthe
twodefensestrategiesappearstobeconstrainedbyanegativegeneticcorrelation.Forexample,mice
infected with Plasmodium chabaudi show a negative genetic correlation between resistance (peak
pathogen load)and tolerance (weightand redbloodcells) (Råbergetal. 2007). This couldbedue to
linkagedisequilibrium,pleiotropiceffectsorphysiologicalconstraints.
Furthercomplicatingtheevolutionarypressuresisthefactthattheenvironment,andinparticularre-
sourceavailabilitycanaffectthepredictionsmentionedabove.First,becauseresourceavailabilitycan
influencenotonlythehost’sabilitytotolerateparasites,butalsohowthehost’sgenotypeaffects its
(Sternbergetal.2012;HowickandLazzaro2014;Mazé-Guilmoetal.2014),toleranceislikelytoevolve
Impactofenvironmentonco-evolutionbetweenhostsandparasites
56
differently under different food environments. Second, because food availability often alter virulent
parasitic effects (including theAe. aegypti - V.culicis system that we study (Bedhomme et al. 2004;
LorenzandKoella2011)),itshouldinfluencetheevolutionoftoleranceandresistance(RestifandKoella
2004).Third,theevolutionarycostofresistancedependsontheavailabilityofresources,sothatdiffer-
entenvironmentsconstrainevolutionofthetwotraitsindifferentways(HochbergandBaalen1998;D.
C.Lopez-PascuaandBuckling2008;Boots2011;Harrisonetal.2013).
WestudiedsuchaspectsoftheevolutionofresistanceandtolerancewiththemosquitoAedesaegypti
and itsmicrosporidian parasiteVavraia culicis.With an experimental evolution approachwe let the
hostsevolveinresponsetoparasitesindifferentenvironments.Ourgeneralgoalwastoevaluatesome
of the ideasmentionedabove.Wetherefore investigated (i)whetheroursystemhasenoughgenetic
variabilityandlowenoughcoststoenabletheevolutionofresistanceandtolerance,(ii)howresource
availability influencestheevolutionoftoleranceandresistance,and(iii)whethercoevolvingparasites
influencethetwodefensetraitsdifferentlythanconstantparasites.
4.2 Materialandmethods
Ourexperimentwasruninaclimatechambersetto26°C,70%relativehumidityanda12hlightand
12hdarkregime.
4.2.1 Experimentalsystem
WeusedthemicrosporidianparasiteVavraiaculicis (providedbyJ.Becnel,USDAGainesville)andthe
UGAL strain of one itsmosquito hosts,Ae. aegypti (provided by P, Guérin, University ofNeuchâtel).
Aedesaegyptioccursthroughoutthetropicsandsubtropics,isanimportantvectorfor,e.g.dengueand
Zikaviruses.
Vavraiaculicis isanobligate,intracellularparasitethatinfectsseveralgeneraofmosquitoes,including
Aedes(WeiserandColuzzi1972).Mosquitolarvaeingesttheparasite’ssporeswiththeirfood,resulting
in the infectionof gutandepithelial cells.Afteraperiodof replicationwithin the larva theparasites
begin to produce their infectious spores,which are transmitted in twoways. First, transmission can
occurfromlarvato larvawhensporesarereleasedafterthe larvadies.Thistransmissionroute isen-
hancedbyfoodstressorstronginfection(Bedhommeetal.2004).Second, if larvaesurvivetheinfec-
tiontodevelopintoadultsthesporescanbereleasedwhenthemosquitodiesintheaquaticenviron-
ment or they can adhere to the surface of the eggs and infect the newly hatched larvae (Andreadis
2007).
Impactofenvironmentonco-evolutionbetweenhostsandparasites
57
4.2.2 Experimentalevolution
Weletthemosquitoevolvefor10generations(i)eitherwithouttheparasite,withanexternallymain-
tainedparasite,orwithaco-evolvingparasite,and(ii)eitherwithhighorwithlowfoodavailabilitydur-
ingthelarvalstage.The‘constant’parasitesweretakenfromourstandardlinemaintainedinourmos-
quitocolony.Theco-evolvingparasitesweretakenfromthepreviousgenerationoftheinfectedmos-
quitoes.Eachtreatmentwasreplicatedthreetimes.Thefirstgenerationoftheexperimentwascreated
byhaphazardlymoving200one-dayoldlarvaefromthecolonytoeachline.Torearthemosquitos,we
hatchedeggssimultaneouslyunderreducedatmosphericpressure.Forthefirstfourdayswerearedfor
eachline50larvaein4petridishes(8cmdiameter)containing30mldeionizedwater(Rearingthelar-
vaeinpetridishesratherthanalargetrayhelpstoobtainasuccessfulinfection).Every24hourswefed
thelarvaewithourstandardamountoffood(Day1:0.06mgoftetraminfishfood,day2:0.08mg,day3:
0.16mg,day4:0.32mg,day5:0.64mg,day6orlater:0.32mg)orwithhalfofthestandarddiet.Weex-
posedthelarvaetoinfectiontwodaysafterhatchingbyadding5.0x106Vavraiaculicissporesin1ml
deionized water. In the first generation constant parasite and coevolution treatments received the
same solution of spores prepared from the standard lab colony. Two days after infection the four
groupsoflarvaefromeachlineweremovedtoa200*150*50mmplastictraycontaining1.5ldeion-
izedwater.Pupaeweretransferred intocages(30x30x30cmsize)containingfilterpapersoakedwith
10%sugarsolutionasfoodandacupcontainingdeionizedwaterandapieceoffilterasanoviposition
substrate.Four,six,elevenandthirteendaysafterthedaywhen75%ofmosquitoesofagivenlinepu-
patetheyweregiventheopportunitytotakeabloodmealonMZ’sarmfor8minutes.Theeggswere
removedevery48hoursandstoredat26°Cand70%relativehumidityuntilthestartofanewgenera-
tion.Fortheco-evolvedparasitepopulation(coevolution)wecollectedthedead infectedmosquitoes
(larvaeandadults),groundtheminaneppendorftube,countedthesporesandkeptthemat5°Cuntil
the next generation of hosts was started. Before starting a new generation we eliminated Vavraia
sporesfromtheeggsbybleachingtheeggsofalllineswith1%householdbleach.
4.2.3 Measuringresultofexperimentalevolution
After 10 generation of evolutionwemeasured spore load and longevity of themosquitoes from all
evolvedlinesexposedtoV.culicissporesfromthelabcolonyandfedeitherwiththestandardamount
offoodorwithhalfofthestandarddiet.Werearedthemosquitoesasdescribedabove,withthefol-
lowingdifferences.First,werearedthelarvaeindividuallyin3mldeionizedwaterinthewellsof12-well
plates.Wehadbetween109and112first-instar larvaeper line (intotal2009 larvae).Each larvawas
haphazardlyassignedtothehighorthelowfoodtreatment(between54and56individualspertreat-
mentandline).Second,weexposedlarvaetotheparasitebyadding100µlofasolutioncontaining2.0x
Impactofenvironmentonco-evolutionbetweenhostsandparasites
58
106Vavraiaculicissporespermldeionizedwater.Third,pupaeweremovedto50mlFalcontubecon-
tainingdeionizedwaterandapieceoffilterpaper.Thecupswerecoveredwithmosquitonetting,and
cotton woolmoistenedwith 10% sugar solution was placed onto the netting remoistened every 48
hoursandchangedevery72hours.Onedayafteremergencethemaleswerediscardedandthefemales
where checked every day for survival. The experimentwas stoppedwhen all of themosquitoes had
died(57daysafterhatching).
4.2.4 Sporemeasurement
Vavraiaculicissporesweremeasuredwithahaemocytometer.Eachmosquitowasindividuallyplaced
into a 2ml Eppendorf tube containing200μl distilledwater anda5mmsteal bead.Mosquitoeswere
crushedbyshakingthetubefor3minutesat35Hz(TissueLyser,Qiagen,Valancia,California).Eightμl
of themixwere added to the haemocytometer (Neubauer improved) and the sporeswere counted
withacellcounter.
4.2.5 Measurementofhostresistanceandtolerance
Wemeasuredtwotypesofresistance:qualitativeresistanceastheproportionof individuals inwhich
wefoundspores,andquantitativeresistanceastheinverseofthesporeload.Wemeasuredtolerance
astheslopebetweenlongevityandsporeloadatthetimeofdeath(Råbergetal.2009).Notethatall
mosquitoesdied laterthan14daysafter infection(~sevendaysafteremergence),atwhichtimethe
spore loadhasgenerally reachedanasymptoticvalue (ZellerandKoella, inprep.).Therefore, it isex-
pectedthatsporeloadwouldnotfurtherincreasewiththemosquito’sageandthusthatitisnotauto-
correlatedwithageatdeath.
4.2.6 Statisticalanalysis
AllanalysesweredonewithRv.3.2.3(RDevelopmentCoreTeam,2015).Differencesintheprobabili-
tiesof infectionwereanalyzedwithageneralized linearmixedeffectmodel (GLMM;binomialerrors,
logit link, using the lme4package) that included food level, the two factors (parasite and food level)
duringtheevolutionaryhistoryandalltheirinteractionasfixedfactors,andreplicateoftheevolution
treatmentasarandomeffectnestedwithinevolutiontreatment.Sporeloadofinfectedindividualswas
analyzedwitha linearmixedeffectmodel that includedthesamefactors.Longevityofadultmosqui-
toeswasanalyzedwitha linearmixedeffectmodel that includedfood,parasiteduringevolutionand
foodduringevolutionasfixedfactors,sporeloadascontinuousvariableandreplicatenestedwithinthe
evolution treatments.A significant interactionbetween spore load andexperimental factors indicate
differences in tolerance (Simms 2000; Råberg et al. 2007). The number of spores (+ 1) was log10-
Impactofenvironmentonco-evolutionbetweenhostsandparasites
59
transfromed for all analyses. Fullmodels includedall possible interactions.Minimalmodelswerede-
rived by removing unsignificant terms followed by model comparisons with likelihood-ratio tests. If
removingatermsignificantlyreducedtheexplanatorypowerofthemodel,itwaskeptinthemodel.
Therelationshipbetweenresistance(meaninverseparasiteburden)andtolerance(slopebetweenpar-
asiteburdenand longevity)wasanalyzedwith linear regression that considered foodavailability and
theinteractionbetweenfoodavailabilityandtoleranceasfactors.
4.3 Results
Atotalof1814outof2009(90%)mosquitoessurvivedtoadulthood.881(48.6%)ofthesewerefemales
andwereanalyzed(between18and33individualsperlineandtreatment).
Qualitative resistancewas affected by none of our experimental factors (Table 4.1).Quantitative re-
sistancewasaffectedbyparasitismandfoodlevelduringevolution(Table4.1,Figure4.1).Whenmos-
quitoeswereexposedtoparasitesduringevolutiontheygenerallyshowedhigherresistances.Post-hoc
test between coevolution and constant parasite treatments showedno significant differences for re-
sistance(analysisnotshown).Mosquitoesoriginatingfromlineswithfoodrestrictionduringevolution
hadahigher resistance.Neither foodavailabilitynor interactionsamong factorshadaneffect forre-
sistance.
Table4:1Statisticalsummaryforresistance
Statisticalsummaryforquantitative(sporeload)andqualitative(probabilityofinfection)resistanceanalysis.
SporeLoad Probabilityofinfection
Factor df χ2 p χ2 p
Food 1 0.19 0.667 >0.01 0.953
Evolutionparasite 2 6.05 0.049 2.89 0.236
Evolutionfood 1 7.13 0.008 0.52 0.473
Food*Evolutionparasite 2 1.66 0.436 1.91 0.384
Food*Evolutionfood 1 1.19 0.276 0.09 0.767
Evolutionparasite*Evolutionfood 2 1.84 0.398 0.40 0.819
Food*Evolutionparasite*Evolutionfood 2 1.40 0.497 1.96 0.375
Tolerancesignificantlyvariedbetweenfoodtreatmentsandbetweenlineswithdifferenttypesofpara-
sitismduringevolution(significant interactionbetweenspore loadandfood,andbetweenspore load
Impactofenvironmentonco-evolutionbetweenhostsandparasites
60
and evolution parasite (Table 4.2, Figure 4.1)). Tolerancewas significantly higher for lines with high
foodavailability and for linesoriginating from lines thatwereexposed toparasitesduringevolution.
Post-hoctestbetweencoevolutionandconstantparasitetreatmentsshowednosignificantdifferences
(analysis not shown).We foundnodifference in tolerancebetween lines rearedat the two levels of
foodduringevolution.Thelongevityofinfectedmosquitoes(whencontrolledforparasiteinducedfit-
nessloss)wassignificantlyinfluencedbytheinteractionbetweenevolutionparasiteandevolutionfood
(Table4.2).Thelongevitywasgenerallyhigherforlinesthathadevolvedwithparasitesandhigherfood
levels.Theinteractionbetweenfoodandevolutionparasitehadaclosetosignificanteffectonthelon-
gevity.
Table4:2Statisticallysummaryfortolerance
Statisticalsummaryforlongevityanalyses.Significantinteractionsbetweenparasiteloadandexperimentalfactorsindicate
differencesintolerance.
Longevity
Factor df Chisq p
Parasiteload 1 2.53 0.111
Food 1 1.19 0.275
Evolutionparasite 2 0.70 0.706
Evolutionfood 1 2.50 0.114
Parasiteload*Food 1 18.95 <0.001
Parasiteload*Evolutionparasite 2 7.03 0.030
Food*Evolutionparasite 2 4.89 0.087
Evolutionparasite*Evolutionfood 2 8.50 0.014
Impactofenvironmentonco-evolutionbetweenhostsandparasites
61
Figure4:1Toleranceandresistance
AedesaegyptitoleranceandresistancetoVavraiaculicisformosquitoesfromdifferentevolutionarylinesandtestedatdiffer-
entfoodlevels.Box-plotsaboveandtotherighteachpanelshowthemedian,the25thand75thquantileandtherangeof
longevityandsporeload.Reddotsandbox-plotsrepresentindividualsoriginatingfromlineswithcoevolvingparasites,yellow
dotsrepresentindividualswithconstantparasitesduringevolutionandbluedotsrepresentindividualsthathadevolvedwith-
outparasites.Linesshowleast-squaresregressionsfordifferentevolutionlines.Panelsinthefirstrowrepresentmosquitoes
testedathighfood;thoseinthesecondrowrepresentmosquitoestestedatlowfood.Panelsinthefirstcolumnrepresent
linesthathadevolvedathighfoodavailability;thoseinthesecondcolumnrepresentlinesthathadevolvedatlowfoodavaila-
bility.Individualsfromdifferentreplicatesarepooled.
coevno
0 1 2 3
par.e
vo
coevno
0 1 2 3
par.e
vo
0
10
20
30
40
50
0 1 2 3
Long
evity
par.evo
0
10
20
30
40
50
0 1 2 3
Evolution LinesCoevolutionConstant parasiteNo parasite
par.evo
coevno
0 1 2 3
par.e
vo
coevno
0 1 2 3
par.e
vo
0
10
20
30
40
50
0 1 2 3Spore load
Long
evity
par.evo
0
10
20
30
40
50
0 1 2 3Spore load par.evo
Test
ed a
t hig
h fo
odTe
sted
at l
ow fo
odEvolved at high food Evolved at low food
Impactofenvironmentonco-evolutionbetweenhostsandparasites
62
Wefoundasignificantpositivecorrelationbetweenresistanceandtolerance(f=4.60p=0.040).Food
availability(f=0.07p=0.795)andtheinteractionbetweenfoodavailabilityandtolerance(f=1.93,p=
0.174)hadnon-significanteffectsindeterminingresistance(Figure4.2).
Figure4:2Tolerancevs.resistance
Therelationshipbetweensporeloadandtoleranceformosquitolineswithdifferentevolutionaryorigin.Notethatthere-
sistanceincreasesfromtoptobottom,soanegativeslopeimpliesapositiveassociationbetweenresistanceandtolerance.
Dotsrepresentmeansporeloadandtoleranceofmosquitoesthathadevolvedathighfoodavailability;trianglesrepresent
mosquitoesthathadevolvedatlowfoodavailability.Redsymbolsrepresentindividualsthathadevolvedwithcoevolving
parasites,yellowsymbolsrepresentindividualsthathadevolvedwithconstantparasitesandbluesymbolsrepresentindividu-
alsthathadevolvedwithoutparasites.Thepanelontheleftshowsmosquitoestestedathighfood;theoneontherightshows
mosquitoestestedatlowfood.Linesshowleast-squaresregressionspooledforalllines.Thep-valuesonthisfigurewerecalcu-
latedforeachfoodtreatmentseparately
4.4 Discussion
InourcolonyofAe.aegyptimosquitoes,toleranceofandresistancetothemicrosporidianparasiteV.
culicis are evolvable traits, so thatmosquito lines that were exposed to parasitism during evolution
showbothahigherresistanceandtolerancetoparasitism.Howeverthewaythesetwodefensetraits
evolve depends on the ecological settings. Thus, restricting food during evolution led to higher re-
sistance,buthadnoimpactontolerance.Furthermore,theecologicalsettingsduringthetestingofthe
−8 −6 −4 −2 0 2 4
0.8
1.0
1.2
1.4
1.6
Tested at high food
Tolerance
Mea
n lo
g−10
spo
re lo
ad
Evolution linesCoevolutionConstant parasiteNo parasite
p = 0.016
−8 −6 −4 −2 0 2 4
0.8
1.0
1.2
1.4
1.6
Tested at low food
Tolerance
Res
ista
nce
p = 0.596
Impactofenvironmentonco-evolutionbetweenhostsandparasites
63
mosquitoes also affected theobserveddefense strategies. Thus, food restrictiongenerallydecreased
tolerance,buthadnoeffectonresistance.
Themosquitoes’parasiteloadandthereductionoflongevitywithincreasingparasiteburdenwaslower
forlinesthathadexposedtoparasitesduringevolution,andindicatesageneticvariationinresistance
andtoleranceforthoselines.Thisisinconsistentwithmodelsthatpredictevolutionaryfixationoftol-
erancealleles(RoyandKirchner2000;Milleretal.2006).Atleastfiveotherstudiesfoundevidencefor
variationintolerance(Råbergetal.2007;ValeandLittle2009;HowickandLazzaro2014;Regoesetal.
2014;Loughetal.2015);twonot(Lefèvreetal.2010;Haywardetal.2014).
That tolerance and resistance both increased as an evolutionary response tomicrosporidia infection
suggests that there isnomajor internal constrain in the formofa stringnegativegenetic correlation
betweenthetwotraits.Thepositiveassociationbetweenthetwotraitsobservedafterevolutioncould
beduetoindependentevolutionofthetwotraitsorduetoapositivegeneticcorrelationbetweenthe
twotraits.Wehavenoindicationfromourstudytosuggestwhichofthetwopossibilitiesiscorrect.The
possibilityofapositivecorrelationisplausible,andsuchacorrelationwasfoundinat leastoneother
study (Howick and Lazzaro 2014), although most other studies show either no genetic correlation
(Sternberg et al. 2012;Mazé-Guilmo et al. 2014) or a negative one (Råberg et al. 2007; Vincent and
Sharp2014).However,thefactthatdifferentevolutionaryenvironmentsledtodifferentevolutionary
pathways suggests that anypossible positive genetic correlationwasnot strong enough to constrain
theevolutionofthetwodefensestrategies,sothatthetwocanevolvemoreorlessindependently.The
observedpositivecorrelationalsoemphasizesthatcorrelationsamongpopulationshouldnotbecon-
sideredasevidence for trade-offsorpositive links,as they result fromacombinationof selectionon
bothtraitsandthegeneticcorrelationbetweenthem(SimmsandTriplett1994;RestifandKoella2004).
We foundnodifferences in resistanceand tolerancebetween lines thatwereexposed to coevolving
parasitesor constantparasites.Accordinglyourdatadoesnot supportbothof the self-contradictory
predictions that coevolution either increases tolerance because of its lower impact on the parasites
fitness (Roy and Kirchner 2000) or decreases because parasites would respond to tolerant hosts by
growingfaster(Milleretal.2006).However,thedurationoftheexperimentmighthavebeentooshort
todifferentiatetolerancebetweenlineswithcoevolvingorconstantparasites.
Eventhoughfoodhadnodirectinfluenceonresistance,mosquitoesevolvedresistancemoreeasilyin
resource-poorenvironments,contrastingfindingsthathighfoodlevelstendtoresultintheevolutionof
elevated resistance (Hochberg and Baalen 1998; D. C. Lopez-Pascua and Buckling 2008; Boots 2011;
Harrison et al. 2013), probably because resistance is less costly when there resources are plentiful.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
64
However,because inoursystemwell-fedmosquitoesevolvedtobetolerant to infection,mosquitoes
mighthavebenefitedonly slightly fromresistance. In resource-poorenvironments,where the fitness
losswithincreasingparasiteburdenwasmuchhigher,resistantindividualsmighthavebeenunderposi-
tiveselectionand increasedwithtimeinfrequency. Inotherwords,theevolutionarytrajectoryofre-
sistanceindifferentenvironmentscoulddependonhowtheenvironmentinfluencestolerance.These
results are inaccordancewith theprediction thathigher levelsof virulence (inour case triggeredby
resourcerestriction)resultsintheevolutionofincreasedresistance(RestifandKoella2004).Analter-
nativeexplanation is that thecostof resistance isonlyapparent ingoodenvironments.Nevertheless
thefactthatresistancecanmoreeasilyevolveinresource-poorenvironmentissurprisingbecausere-
source-harshenvironmentscanthereforeincreasethehostabilitytodealwithparasites.
Aconsiderablepartofthemosquito’slongevityisexplainedbyfactorsthatarenotrelatedwithparasite
burden.Mosquitoesthatwereexposedtoparasitesduringevolutiongenerallylivedlonger.Thiscanbe
explainedbypotentialcostsofquantitativeresistanceorbyevolvinga longeruninfectedlongevity. In
uninfectedmosquitoes this trendwassimilar (Zeller&Koella inprep.) suggesting that theconditions
duringevolutionaffectedtheevolvedlongevity(generalvigor).
4.4.1 Conclusions
Ourstudyisthefirstexperimentaltest,whichfoundenoughgeneticvariabilityandlowcostthattoler-
anceandresistanceevolvedwhenfacedtoparasitesandunderdifferentresourcelevels.Thefactthat
differentcombinationsofresistanceandtoleranceevolvedindifferentecologicalsettingsillustratesthe
importancetostudythosetraitsacrossenvironmentalvariables.Inadditionbecausemanyofthepre-
dictionsdidnotholdtrueinoursystemunderlinestheimportancetoincorporatesuchenvironmental
heterogeneity, condition-dependent evolutionary costs and non-independence between both traits
whenmodeling the evolution of resistance and tolerance.Our study provides further an example of
how resources and the ability to tolerate parasitesmight interact to determine the evolution of re-
sistance.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
65
Antagonisticcoevolutionandre-Chapter5
sourcesalterthehost’slifehistoryevolution
MICHAELZELLER1,JACOBC.KOELLA1
1LaboratoryofEcologyandEpidemiologyofParasites,InstituteofBiology,UniversityofNeuchâtel,Rue
Emile-Argand11,2000Neuchâtel,Switzerland
Impactofenvironmentonco-evolutionbetweenhostsandparasites
66
Abstract
Earlyreproductionasanadaptiveconsequenceofparasitismhasbeenpredictedbylifehistorytheory.
However,empiricalvalidationofthispredictionis limitedanddirectexperimentalevidencethatpara-
sitescaninfluencethehost’sgeneticallydetermineddevelopmenttimeismissing.Wehavesetupan
evolutionexperimentbylettingAedesaegyptimosquitoevolveeitherwithnoparasite,withaconstant
parasite,orwithaco-evolvingparasite(Vavraiaculicis)witheitherhighorlowresourceavailability.We
testedthehost’slifehistorytraitsoftheevolvedlinesathighandatlowfoodavailabilitywithorwith-
outparasites.Wefoundthatwhenmosquitoeswereexposedtoparasitesduringevolutiontheyhada
shortergeneticallydetermineddevelopmenttime,anequalbodysizeandalongerlongevitycompered
tomosquitoesoriginatingfromcontrollines.Theseresultssuggestnoevidenttrade-offsamongageat
maturity and other traits. We also found no differences in age at maturity between lines with co-
evolvingorconstantparasites,but lineswithco-evolvingparasites livedsignificantly longer.Theenvi-
ronmentalconditionsinfluencedthehost’slife-historiesandthephenotypicplasticresponsestopara-
sitisminmanyways,suggestingthatvariableenvironmentsinfluencethelong-termhostevolution.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
67
5.1 Introduction
Theco-evolutionaryinteractionsbetweenhostsandparasitesarerecognizedasimportantfactorsshap-
inghostslifehistoryevolution(Hochbergetal.1992;Koellaetal.1998;KoellaandRestif2001;Gandon
etal.2002;AshbyandBoots2015).Parasitesmodifytheecologicalcontextinwhichhosttraitsevolve
byaffectingcomponentsofthehost’sage-dependentfecundityandmortality.Thisselectionpressure
caninfluencethehost’soptimalpatternofresource-allocation.
Thereisincreasingevidencethathostsmayaltertheirlifehistorytraitsinawaytocompensateforthe
negativeeffectsofparasitism(Minchella1985;Koellaetal.1998;Richner1998;Agnewetal.2000).Life
history theory predicts that earlier reproducing hosts will have a selective advantage, because they
mightbeable toevadeparasitism in timeand,whenparasitized, reduce the impactofparasitismon
reproductive success and survival (Hochberg et al. 1992; Forbes 1993; Perrin and Christe 1996). Life
history traitsof thehost,whichhavebeen shown to respond toparasitism, includeearly versus late
fecundity (Minchella and Loverde 1981; Gérard and Théron 1997; Adamo 1999), reproductive effort
(Sorcietal.1996;PolakandStarmer1998;Krist2001),parentalcare(Christeetal.1996;Richnerand
Tripet1999),bodysize(Lafferty1993;Pontieretal.1998;Arnottetal.2000)anddevelopmentaltime
(Agnewetal.1999; Jonesetal.2008). Insomecasesthesemodificationshadageneticunderpinning
(Lafferty1993;KoellaandAgnew1999),inothers,theywereplasticresponses.However,anoptimiza-
tion in one trait (e.g. maturing early) might be associated with penalties in another trait, such as
maintenance functions, immune-related traits or late-life reproduction. This can be induced by re-
source-allocationtrade-offs,linkagedisequilibriumorpleiotropy.
Suchmodificationinlifehistorycanbeseenasaformofresistance,whichmightleadtocomplexco-
evolutionarydynamicsbetweenthelifehistoriesofthehostandtheparasites.Becauselifehistoriesof
hosts and parasites are at least partly determined by the genotype of the counterpart (Koella and
Agnew1999),anevolutionaryresponseofhost’slifehistorytoparasitismmayagainaltertheselection
pressureoftheparasite.Thisreciprocalselectionmightresultincoevolution,withcontinuouschanges
ofhostsandparasites life-histories (Gandonetal.2008;GabaandEbert2009).Accordingtothered-
queenhypothesistheevolutionaryratesofchangeingenesrelatedtoresistancetraitsshouldbeaccel-
eratedthroughco-evolutionarydynamics.Indeed,genomesofcoevolvinghostshaveshowntoevolve
fastercomparedtopopulationsevolvingagainstaconstantparasitepopulation (Patersonetal.2010;
Kashiwagi and Yomo2011). Theoretical studies reveal that co-evolutionary dynamics can havemany
effectsfortheevolutionofthehostlifehistory(KoellaandRestif2001;Restifetal.2001;Gandonetal.
2002;AshbyandBoots2015),howeverdirectexperimentalevidencethatparasites,and inparticular,
co-evolvingparasitescaninfluencethehost’sgeneticallydeterminedlifehistoryismissing.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
68
Contrarilytotheevolvedadaptations in lifehistorytraits,manyhostsrespondtoparasitismandfood
restrictionphenotypicallyplastic (includingAe.aegypti -V.culicis systemthatwestudy).A lotofhosts
decrease their growth, mature later and with a smaller body size once infected (Bedhomme et al.
2004).Inadditiontothelifehistorytraitsthemselves,thedegreeofphenotypicplasticity,theabilityof
singlegenotypestoproducemorethanonephenotypeacrossdifferentenvironments(Pigliucci2001),
isalsogeneticallydetermined.Ifthereexistsadaptivevariationforphenotypicplasticityamonggeno-
types, it is likely to evolve differently under variable conditions. However, how phenotypic plasticity
evolvesexperimentallyhasrarelybeeninvestigated.
Weuseanexperimentalevolutionapproachwith themosquitoAedesaegypti and itsmicrosporidian
parasiteVavraiaculicistoexaminehowthelifehistoriesofthehostchange,bylettinghostsevolvein
responsetoparasitesunderdifferentresourcelevels.Studyingdifferentlevelsofresourcescanberele-
vantbecausepotential trade-offsbetweendifferent lifehistory traitsmightonlybedetectablewhen
resources are scarce. Specificallywewant to investigate: (i)whetheranexposition toparasitesover
severalgenerationleadstoearlymaturinghost’s,(ii)whetherevolvedearlymaturityisassociatedwith
costslaterinlife,(iii)whetherco-evolvingparasitesinfluencethehostslifehistoryevolutiondifferently
thanconstantparasites(iv)andwhetherourmosquitocolonyhasenoughgeneticvariabilityandadap-
tivedifferencestoenabletheevolutionofphenotypicplasticity.
5.2 Materialsandmethods
Theexperimentwasruninaclimatechambersetto26°C,70%relativehumidityanda12hlightand
12hdarkregime.
5.2.1 Experimentalsystem
WeusedthemicrosporidianparasiteVavraiaculicis (providedbyJ.Becnel,USDAGainesville)andthe
UGALstrainoneofitsmosquitohosts,Aedesaegypti(providedbyP,Guérin,UniversityofNeuchâtel).
Vavraiaculicisisanobligate,intracellularparasitethatinfectsinnatureseveralgeneraofmosquitoes,
includingAedes(WeiserandColuzzi1972).ThemosquitolarvaeingestthesporesofVavraiaculiciswith
theirfood,resultingininfectionofgutcellsandepithelialcells.Afteraperiodofreplicationwithinthe
larvaetheparasitesbegintoproducetheirinfectiousspores,whicharetransmittedintwoways.First,
transmissioncanoccurfromlarvatolarvawhensporesarereleasedafterthelarvadies.Thistransmis-
sionrouteisenhancedbyfoodstressorstronginfection(Bedhommeetal.2004).Second,iflarvaesur-
vive the infection to develop into adults the spores can be releasedwhen themosquito dies in the
Impactofenvironmentonco-evolutionbetweenhostsandparasites
69
aquaticenvironmentortheycanadheretothesurfaceoftheeggsandinfectthenewlyhatchedlarvae
(Andreadis2007).
5.2.2 Experimentalevolution
Weletthemosquitoevolvefor10generations(i)eitherwithnoparasite,withanexternallymaintained
parasite,orwithaco-evolvingparasite,and(ii)eitherwithhighorwithlowfoodavailabilityduringthe
larval stage.Eachparasite treatmentevolved.The ‘constant’parasiteswere taken fromour standard
linemaintainedinourmosquitocolony.Theco-evolvingparasitesweretakenfromthepreviousgener-
ationoftheinfectedmosquitoes.Eachtreatmentwasreplicatedthreetimes.Thefirstgenerationofthe
experimentwascreatedbyhaphazardlymoving200one-dayoldlarvaefromthecolonytoeachline.
Torearthemosquitos,wehatchedeggssimultaneouslyunderreducedatmosphericpressure.Forthe
firstfourdayswerearedforeach line50 larvae in4petridishes(8cmdiameter)containing30mlof
deionizedwaterinordertoensureasuccessfulinfection.Every24hourswefedthelarvaeeitherwith
ourstandardamountoffood(Day1:0.06mgoftetraminfishfood,day2:0.08mg,day3:0.16mg,day4:
0.32mg,day5:0.64mg,day6orlater:0.32mg)orwithhalfofthestandarddiet.Weexposedthelarvae
toinfectiontwodaysafterhatchingbyadding5.0x106Vavraiaculicissporesin1mldeionizedwater.
In the first generation constant parasite and coevolution treatments received the same solution of
sporespreparedfromthestandardlabcolony.Twodaysafterinfectionthefourgroupsoflarvaefrom
each lineweremovedtoone200*150*50mmplastic traycontaining1.5 literofdeionizedwater.
Pupaeweretransferred intocages(30x30x30cmsize)containingsugarsolutionandacupcontaining
deionizedwater and a piece of filter as an oviposition substrate. Four, six, eleven and thirteen days
afterthedaywhen75%ofmosquitoesofagivenlinepupatedtheyweregiventheopportunitytotake
abloodmealonMZ’sarmfor8minutes.Theeggswereremovedevery48hoursandstoredat26°Cand
70%relativehumidityuntilthestartofanewgeneration.Fortheco-evolvedparasitepopulation(co-
evolution)wecollectedthedeadinfectedmosquitoes(larvaeandadults),groundtheminaneppendorf
tube,countedthesporesandkeptthemat5°Cuntilthenextgenerationofhostswasstarted.Before
startinganewgenerationweeliminatedVavraiasporesfromtheeggsbybleachingtheeggsofalllines
with1%householdbleach.
5.2.3 Measuringresultofexperimentalevolution
After10generationofevolutionwemeasuredtheprobabilityofemergence,ageatpupation,adultsize
and longevityofthemosquitoesfromtheevolved lines.Mosquitoesfromthe18selection lineswere
exposedtoV.culicissporesfromthelabcolonyandfedeitherwiththestandardamountoffoodorwith
halfofthestandarddiet.Werearedthemosquitoesasdescribedabove,withthefollowingdifferences.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
70
First,we reared the larvae individually in3mldeionizedwater in thewellsof12-wellplates.Wehad
between219and224first-instar larvaeper line(intotal4005larvae).Eachlarvawashaphazardlyas-
signedtooneofthefourtreatments(between53and56individualspertreatmentandline).Second,
we exposed larvae to the parasite by adding 100µl of a solution containing 2.0 x 106Vavraia culicis
spores perml deionizedwater. Third, pupaeweremoved to 50ml Falcon tube containing deionized
waterandapieceoffilterpaper.Thecupswerecoveredwithmosquitonetting,andcottonwoolmois-
tenedwith10%sugarsolutionwasplacedontothenettingremoistenedevery48hoursandchanged
every72hours.Onedayafteremergence themaleswerediscardedand the femaleswherechecked
everydayforsurvival.Theexperimentwasstopped69daysafterhatchwhenallofthemosquitoeshad
died.
5.2.4 Traitmeasurement
The size of adultswas assayed as themean of theirwing length,which strongly correlateswith the
weight ofmosquitoes (Koella and Lyimo 1996) and is commonly used as an approximation for adult
size.Thewingswereremovedandmountedonmicroscopeslides.Theslidesweredigitallyscannedand
thewingsweremeasuredwiththeopen-accesssoftwareIMAGEJ.
5.2.5 Statisticalanalysis
Differencesintheprobabilitiesofemergencewereanalyzedwithgeneralizedlinearmixedeffectmodel
(binomialdistribution) that includedparasite infection, food level, the two factors (parasiteand food
level)duringtheevolutionaryhistoryandalltheirinteractionasfixedfactors.Replicatewastreatedas
randomeffect,nestedwithinevolutiontreatment.Weusedtheglmerfunctionfromthelme4package
fromRv.3.2.3(RDevelopmentCoreTeam,2015).Ageatpupationandlongevitywereanalyzedwitha
mixedeffectsurvivalanalysis(cox-proportionalhazard)thatincludedparasiteinfection,foodavailabil-
ity,parasiteduringevolutionandfoodduringevolutionasfixedfactors,replicatewastreatedasaran-
domeffect,nestedwithintheevolutiontreatments.Intheanalysisoflongevity,weaddedwinglength
asapotentialconfounder.Significantinteractionsbetweenevolutionaryfactorsandtestedfactorsindi-
cate evolutionary differences in phenotypic plasticity.Wing lengthwas analyzedwith a linearmixed
effectmodel(lmerfunctionfromlme4package)thatincludedparasiteinfection,foodavailability,para-
site during evolution and foodduring evolution and all their interaction as fixed factors, replicate as
randomeffect,nestedwithinevolutiontreatment.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
71
5.3 Results
Atotalof3716outof4005 (93%)mosquitoessurvivedtoadulthood.1764 (47.5%)of thesewere fe-
males andwere analyzed (between 17 and 33 individuals per line and treatment). Food variability,
parasiteinfectionandtheirinteractionstronglyinfluencedthemosquitos’lifehistory(Table5.1,Figure
5.1,5.2&5.3).Wewillfocushereonhowthefoodavailabilityandthetypeofparasitismduringevolu-
tioninfluencedthemosquito’slifehistoryandhowthesefactorsinfluencedthemosquito’sphenotypic
plasticresponsetofoodstressandparasiteinfection.Becausemanyfactorsandinteractionsinfluenced
themosquitos’lifehistorywedescribeonlytheresults,whichseemmostimportanttous.
5.3.1 Ageandsizeatmaturity
Theageatpupationwassignificantlyinfluencedbythetypeofparasitismduringevolution(Table5.1,
Figure 5.1)).Whenmosquitoeswere exposed to parasites during evolution theymatured at a lower
age.ThiswasparticularlythecasewheninfectedwithV.culicis(significantinteractionbetweenparasite
infectionandtypeofparasitismduringevolution).Tukey’sHSDposthoctestbetweencoevolutionand
constant parasite treatments showed no significant differences for age at pupation (analysis not
shown).Inadditionthetypeofparasitismandfoodavailabilityduringevolutionaffectedthephenotyp-
icplasticresponsetoparasitismandfoodrestriction(significantinteractionsbetweenevolutiontreat-
ments and “tested” treatments). For example, mosquitoes that were not exposed to parasites and
raisedonamplefoodduringevolution(controlhighfood),showedindevelopmentaltimethehighest
phenotypicplasticresponsewhenexposedtoparasitism(forhighandlowfood(Figure5.2aand5.2c).)
Contrary to that, we find the lowest phenotypic plastic response to food restriction for individuals,
whichwerenotexposedtoparasitismandhadlowfoodavailabilityduringevolution(controllowfood,
Figure5.2b).
Thewing lengthwassignificantly influencedby foodavailabilityandby the interactionbetween food
andparasiteinfection.Itwaslowestforparasitizedmosquitoeswithlowfoodavailability.Furthermore
the three-way interaction between parasite infection, type of parasitism and food availability during
evolutionaswell as the three-way interactionbetween foodavailability, typeofparasitismand food
availabilityduringevolutionhadsignificanteffectsindeterminingthewinglength(Figure5.1andTable
5.1).Thephenotypicplasticresponsetofoodrestrictionandtofoodrestrictioncombinedwithparasite
infectionwassmallestforcoevolvedmosquitoeswithhighlarvalfoodavailability(Fig.5.2B&C).Tuk-
ey’sHSDpost hoc testbetween coevolution and constant parasite treatments showedno significant
differences(analysisnotshown).
Impactofenvironmentonco-evolutionbetweenhostsandparasites
72
Figure5:1Ageandsizeatmaturity
Meanwinglength±SEagainstmeanageatpupation±SE.Redsymbolsrepresentcoevolutionlines,yellowsymbolsrepresent
lineswithconstantparasites,andbluesymbolsrepresentslineswithoutparasites.Trianglesrepresentlinesthathadlowfood
availabilityduringevolution;squaresrepresenttreatmentswithhighfoodavailabilityduringevolution.Opensymbolsrepre-
senttreatmentsthatwerenotinfected;filledsymbolsrepresenttreatmentsthatwereinfectedwithparasites.Individualsfrom
differentreplicatesarepooled.
7 8 9 10 11
0.22
50.
230
0.23
50.
240
0.24
50.
250
0.25
5
Mean age at pupation (days)
Mea
n w
ing
leng
th (c
m)
Tested at high food
Tested at low food
Evolution linesCoevolutionConstant parasiteNo parasite
Impactofenvironmentonco-evolutionbetweenhostsandparasites
73
Figure5:2Phenotypicplasticity
PhenotypicplasticresponsetoA)parasiteinfection,B)foodrestrictionandC)parasiteinfectionandfoodrestrictionforcol-
umnI)ageatpupationandcolumnII)winglength.Phenotypicplasticitywasestimatedforeachevolutiontreatment.We
calculatedthedistancebetweenthemeanvalues±SEfromhighfoodun-parasitizedtreatmentstohighfoodparasitized,low
foodun-parasitizedandlowfoodparasitizedtreatments.Individualsfromdifferentreplicatesarepooled.
5.3.2 Probabilityofemergenceandadultlongevity
Theprobabilitythatjuvenilemosquitoesemergedintoadultswassignificantlyaffectedbytheamount
offoodavailabilityandbytheinteractionbetweenfoodavailabilityandparasiteinfectioninthecurrent
condition(Table5.1,Fig.5.3a-c).Itwasgenerallylowerforparasitizedmosquitoesandunderfoodre-
striction.Parasitesandfoodduringevolutionhadnosignificanteffectsfortheprobabilityofemergence
buttheinteractionbetweenthemwasclosetostatisticallysignificant.Additionallythethree-wayinter-
Index
Control high food
Control low food
Constant parasite high food
Constant parasite low food
Coevolution high food
Coevolution low food
0 1 2 3 4
Index
Control high food
Control low food
Constant parasite high food
Constant parasite low food
Coevolution high food
Coevolution low food
0 1 2 3 4
Change in age at pupation (days)
Control high food
Control low food
Constant parasite high food
Constant parasite low food
Coevolution high food
Coevolution low food
0 1 2 3 4
Index0.00 0.02 0.04
Index0.00 0.02 0.04
Change in wing length (cm)
0.00 0.02 0.04
I IIA
B
C
Impactofenvironmentonco-evolutionbetweenhostsandparasites
74
actionbetweenparasiteinfection,foodandfoodduringevolutionwasmarginallynotsignificant.Tuk-
ey’sHSDposthocbetweencoevolutionandconstantparasitetreatmentsshowednosignificantdiffer-
ences(analysisnotshown).
Theadult longevitywassignificantly lower for infected individualsandwith low foodavailability (Fig.
5.3d-f).Parasiteinfection,foodavailabilityanditsinteractionhadhighlysignificanteffectsindetermin-
inglongevity.Parasitesduringevolutionhadaclosetosignificanteffectindetermininglongevityanda
significanteffectinaffectinglongevityincombinationwithparasiteinfectionandalsowithfoodavaila-
bility.Thefoodavailabilityduringevolutionhadnosignificanteffectsfortheadultlongevity,butitin-
fluenced the longevity in combinationwith parasites during evolution. The coevolved lines generally
lived longer compared to lines kept on constant parasites (Fig. 5.3f). Tukey’sHSDpost hoc testsbe-
tweencoevolutionandconstantparasitetreatmentsrevealedsignificantdifferencesinadultlongevity
(z=2.64,p=0.022).
Impactofenvironmentonco-evolutionbetweenhostsandparasites
75
Figure5:3Probabilityofemergenceandsporeload
Meanprobabilityofemergenceof(a)coevolvedversuscontrollines(noparasite),(b)constantparasiteversuscontrollines
and(c)coevolvedversusconstantparasitelines.Meanlongevityof(d)coevolvedversuscontrollines,(e)constantparasite
versuscontrollinesand(f)coevolvedversusconstantparasitelines.Everysymbolindicatesapairwisecomparisonofasingle
treatmentbetweenlinesfromdifferentorigins.Trianglesrepresentlinesthathadlowfoodavailabilityduringevolution;
squaresrepresenttreatmentswithhighfoodavailabilityduringevolution.Redsymbolsrepresenttreatmentsthatwereinfect-
edwithV.culicis,blacksymbolsrepresentstreatmentsthatwerenotinfected.Opensymbolsrepresenttreatmentswithhigh
foodavailabilityaslarvae;filledsymbolsrepresenttreatmentswithlowfoodavailability.Thedashedlineindicatestheexpec-
tationifmortality,respectivelylongevitywouldbeequalforcontrol,coevolvedandconstantparasitelines.Individualsfrom
differentreplicatesarepooled.
Prob. of emergence control lines
Prob
. of e
mer
genc
e co
evol
ved
lines
0.75
0.80
0.85
0.90
0.95
1.00
0.75 0.80 0.85 0.90 0.95 1.00
Prob. of emergence control linesPr
ob. o
f em
erge
nce
evol
ved
lines
0.75
0.80
0.85
0.90
0.95
1.00
0.75 0.80 0.85 0.90 0.95 1.00
Prob. of emergence evolved lines
Prob
. of e
mer
genc
e co
evol
ved
lines
0.75
0.80
0.85
0.90
0.95
1.00
0.75 0.80 0.85 0.90 0.95 1.00
Longevity control lines
Long
evity
coe
volve
d lin
es
22
24
26
28
30
32
34
36
38
22 24 26 28 30 32 34 36 38
Lonegvity control lines
Long
evity
evo
lved
lines
22
24
26
28
30
32
34
36
38
22 24 26 28 30 32 34 36 38
Longevity evolved linesLo
ngev
ity c
oevo
lved
lines
22
24
26
28
30
32
34
36
38
22 24 26 28 30 32 34 36 38
a) b) c)
d) e) f)
Impactofenvironmentonco-evolutionbetweenhostsandparasites
76
Table5:1Statisticalsummarylifehistorytraits
Statisticalsummaryforthehosts’lifehistorytraits.Generalizedlinearmixedeffectmodel(binomialdistribution)fordiffer-
encesinprobabilityofemergence,mixedeffectsurvivalanalysis(coxproportionalhazard)forageatpupation,linearmixed
effectmodelforwinglengthandmixedeffectsurvivalanalysis(coxproportionalhazard)fordifferencesinlongevity.Statisti-
callysignificantvaluesaregiveninbold.
Probabilityof
emergence
Ageatpupation Winglength Longevity
Factor df χ2 p χ2 p χ2 p χ2 p
Parasite 1 0.38 0.538 133.67 <0.001 0.02 0.896 104.49 <0.001
Food 1 9.95 0.002 1446.75 <0.001 241.62 <0.001 244.14 <0.001
Evolutionparasite 2 2.91 0.233 7.94 0.019 1.46 0.483 4.88 0.087
Evolutionfood 1 1.00 0.317 2.38 0.123 0.01 0.919 0.00 0.968
Parasitexfood 2 7.34 0.007 40.09 <0.001 5.60 0.018 10.44 0.001
Parasitexevolution
parasite
2 0.12 0.943 18.52 <0.001 3.11 0.211 10.69 0.005
Foodxevolutionparasite 2 1.02 0.601 6.94 0.031 3.40 0.183 6.16 0.046
Parasitexevolutionfood 1 0.03 0.862 6.37 0.012 1.12 0.290 0.35 0.557
Foodxevolutionfood 1 0.91 0.339 8.60 0.003 0.01 0.917 2.20 0.138
Evolutionparasitex
evolutionfood
2 5.14 0.076 3.84 0.146 3.86 0.145 9.57 0.008
Parasitexfoodxevolution
parasite
2 0.11 0.949 0.26 0.878 2.36 0.308 7.73 0.021
Parasitexfoodxevolution
food
1 3.04 0.081 4.01 0.045 0.48 0.489 2.16 0.141
Parasitexevolution
parasitexevolutionfood
2 2.52 0.283 5.76 0.056 6.35 0.042 1.29 0.526
Foodxevolutionparasitex
evolutionfood
2 0.28 0.871 4.06 0.131 7.21 0.027 3.40 0.183
Winglength 1 - - - - - - 1.60 0.206
Impactofenvironmentonco-evolutionbetweenhostsandparasites
77
5.4 Discussion
Parasitismandthefoodavailabilityduringevolutionalteredthehost’slifehistoryinmanyways.Mos-
quitoesoriginating from lines thatwereexposed toparasitismduringevolutionpupatedearliercom-
pared to control lines. These results are in accordance with the prediction that earlier reproducing
hostswill evolve to reduce the impact of parasitism (Hochberg et al. 1992; Forbes 1993; Perrin and
Christe 1996). Thiswas particularly the casewhenmosquitoeswere exposed to infection and under
food restriction. In other words, the parasite-induced selection over 10 generations shortened the
mosquitosgeneticallyunderpinneddevelopmenttime.Toourknowledge,thisisthefirstexperimental
evidenceshowingparasiteinfectionleadstoearlyreproducinghosts.Theshifttowardsearliermatura-
tionweobservedhere,mightindeedreducethecostsofinfectionaswithongoingtimeV.culicisprolif-
eratesandproducesdamagingspores(reducedreproductivesuccessandreducedlongevityoffemale
mosquitoes(Reynolds1970,Zeller&Koella,inprep)).
The shift had no negative consequences for the adult body size. Accordingly,mosquitoes that were
facedtoparasitismduringevolutionareabletogrowfaster,whenparasitized.Asthelevelofparasites
virulence is at leastpartlydeterminedby thegeneticbasisofmosquitosageatpupation (Koellaand
Agnew1999),earlyemergingindividualswereunderpositiveselectionwhenexposedtoV.culicis.Simi-
larly,theprobabilitythatmosquitoesemergedintoadultswashigherformosquitoesoriginatingfrom
linesthatwerefacedtoparasitismduringevolution(whenparasitizedandunderfoodrestriction(Fig-
ure5.3a&b)).Inadditiontothatmosquitoesoriginatingfromlinesthatwerefacedtoparasitesduring
evolutiongenerally lived longer,againespeciallyunderparasiteexposureand foodrestriction (Figure
5.3d and5.3e). These results indicate, that the evolvedearlymaturation seemsnot tobe traded-off
withthemosquito’sbodysizeortheadultlongevity.Mosquitoesthatwereexposedtoparasitesduring
evolutiongenerallyshowedashorterdevelopmenttime,anequalbodysizeandalongersurvivalcom-
pared tomosquitoes originating from control lines. It therefore seems thatV.culicis exerts a general
anddirectionalselectionpressureonlifehistoriesofAedesaegypti.However,wedonotknowwhether
V.culicisdirectlyexertsselectiononthemosquitosageatmaturity, longevityorreproductivesuccess,
orwhetherthisistheresultofselectiononacorrelatedtrait.
Wefoundthatcoevolvingparasitesdidnotinfluencethehost’sdevelopmenttimedifferentlythancon-
stantparasites.However,thedurationoftheexperimentmighthavebeentooshorttofindsuchdiffer-
ences.Astrikingresultwas, that individualsoriginating frommostof thecoevolved lines lived longer
comparedtomosquitoesfromlineswithconstantparasites.Oneexplanationcouldbe,thattheevolu-
tionary rate of change in the host’s longevitywas accelerated by co-evolving parasites (evolutionary
arms race).However,when foodavailability during evolutionwas lowandwhen tested at high food
Impactofenvironmentonco-evolutionbetweenhostsandparasites
78
levels,themosquitoesfromlineswithconstantparasiteslivedlonger.Thispatternwasconsistentwith
and without parasites. Restricted resources therefore seem to impede the host’s ability to adapt
againstco-evolvingparasites.Accordingly,co-evolutionagainstV.culicisparasitesmightbecostly.
In addition to the evolved adaptations in life histories discussed above, the ecological setting during
evolutionalsoaffectedthemosquito’sphenotypicplasticity.Theenvironmentinthattheyweretested
explained the biggest part of the phenotypic plasticity (development time andwing length). Still we
foundenoughgeneticvariability inphenotypicplasticity toevolveunderdifferentecological settings.
Forexample,coevolvedmosquitoesthathadampleresourcesduringevolutionshowthesmallestplas-
ticresponseinwinglengthwhenfacedwithparasitism.Accordinglyco-evolutionmighthaveincreased
thehost’sadaptiveabilitytodealwithparasites.However,thetrendswedescribehereareverydiffi-
culttointerpretandmightbeinfluencedbyepistasisandpleiotropy(LynchandWalsh1998).Neverthe-
lesstheyillustratethecomplexityanddependencefromenvironmentalconditionofthehosts’lifehis-
toryevolution.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
79
SynthesisandfutureresearchChapter6
6.1 Summaryofresults
This thesis shows that environmental variability can influencemany aspects of host-parasite interac-
tions.InthesecondchapterIdescribetheeffectofvariablefoodavailabilityduringthedevelopmentof
themosquitoAedesaegyptionitslife-history.Oneofourquestionsiswhether‘compensatorygrowth’
afteraperiodofundernourishment,whichgenerallyappearstobethoughtofasanadaptiveresponse
(Dmitriew 2011), increases reproductive success.We show, however, that compensatory growth did
notincreasereproductivesuccess.Movingfromhightolowfoodavailabilityalsohadunexpectedcon-
sequences,leadingtolowerreproductivesuccessthanconsistentlybadlynourishedindividuals.Varying
nutrition is thusclearly importanttounderstandpopulationecologyand life-historyevolution. I think
thatlife-historytheoryshouldbeextendedtoincludetheselong-termeffectsofearlynutrition.These
resultsarealso importantbecausesucheffectsofearlynutrition,thatalteradulttraits,can influence
the mosquito’s capacity to transmit vector born diseases (Sumanochitrapon et al. 1998; Alto et al.
2008).Accordingly,ourdatamaybeuseful forpredictingdiseasetransmissionanddevelopingstrate-
giesformosquitopopulationcontrol.
Inthethirdchapterwestudytheeffectoftheenvironmentontherelationshipbetweenthegrowthof
themicrosporidianparasiteVavraiaculicisandthelongevityof itshosts,themosquitoAedesaegypti.
Ourdatasuggestthat,inmostenvironmentswestudy,thereisanegativerelationshipbetweenpara-
site development and host health. Howeverwe show that food availability and age at infection can
changetheeffectofparasitegrowthonhostlongevity.Suchcontext-dependentrelationshipbetween
parasitedevelopment and virulence canhavea considerable impacton theevolutionaryoutcomeof
infectiousdisease.Accordingly,theecologicalconditionswillchangetherelativecostsandbenefitsof
parasitereplicationandarelikelytodetermineadaptivelevelsofvirulence.Therefore,theoreticalstud-
Impactofenvironmentonco-evolutionbetweenhostsandparasites
80
iesthatpredicttheevolutionofvirulenceshouldconsider,inadditiontohowtheenvironmentaffects
epidemiologicallyrelevantparameters,alsohowitaffectstherelationshipsbetweenthem.
Inafurtherstep,wetesthowtheabilityofAe.aegyptimosquitoestoresistandtoleratetheV.culicis
parasite evolves in different environments. Unsurprisingly, tolerance and resistance to disease both
increased ifmosquitoeswereexposed toparasites.However, indifferentevolutionary scenarios,dif-
ferentcombinationsofthesetwodefensestrategiesevolved,andindifferentecologicalsettingstheir
expression varied. For example, we found that lines that had evolvedwith low food had higher re-
sistancethanthoseevolvedwithhighfood,buttherewasnodifferenceintolerance.Whenwetested
theevolvedmosquitoes, those thatweregiven restricted foodhad lower tolerance than thosegiven
normalfood,buttherewasnodifferenceinresistance.Suchfindingsillustratetheimportanceofincor-
poratingenvironmentalheterogeneity,conditiondependentevolutionarycostsandnon-independence
betweenbothtraitswhenpredictingtheevolutionoftoleranceandresistance(similarlyto(Carvaland
Ferriere2010)).Ourstudyalsoprovidesanexampleofhowresourcesandtheabilitytotoleratepara-
sitesmight interact todeterminetheevolutionof resistance.Suchfindingsarealsoclinically relevant
andmighthelptoelucidatetheevolutionaryimplicationsoftoleranceandresistancebasedtherapies.
More broadly this chapter should help to increase our knowledge of the environmental and genetic
sources of variation in tolerance and resitance, and how these variations affect fitness. Considering
variableecologicalconditionsisthusclearlyimportanttoundersandhost-parasiteco-evolution.
Foodavailabilityandparasiteinfectionalsoinfluencetheevolutionofthehosts’lifehistory.Weshow
thatparasite-inducedselectionover10generationscanshortenthehostsgeneticallydeterminedde-
velopmenttime.Thisisinaccordancewithlife-historytheory,whichpredictsthatearlymaturinghosts
havefitnessbenefitswhenexposedtoparasitism(Hochbergetal.1992).Toourknowledgethisisthe
firstexperimentalevidenceshowingparasiteinfectionleadstothehostreproducingearlierthanwhen
there isno infection.Mosquitoes thatwereexposed toparasitesduringevolutionhadshorterdevel-
opment times, an equal body size and even a longer longevity compared tomosquitoes originating
fromcontrollines.Thissuggestsnoevidenttrade-offsbetweenthetraitswemeasured.Themicrospor-
idianparasite thereforeseemstoexertageneralanddirectional selectionpressureon thehost’s life
histories.However,thefactthatenvironmentalconditionsduringevolutionandco-evolutionhadmany
effects in theexpressionof thehost’s life-history traits illustrates thecomplexityofhost-parasite co-
evolution. Our results further suggest, that variable environmentsmay influence the long-term host
evolution.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
81
6.2 Futuredirections
Inchaptertwowedescribethereproductiveburdenassociatedwithcompensatorygrowth.Anextstep
couldbe toquantifycostsof fastgrowthbycomparingoxidative stressandphysiologicalparameters
(carbohydrates, lipids,proteins)between fastgrowingandnormallygrowing individuals (triggeredby
foodavailabilityortemperature).Similartestscouldbeperformedtoexplorewhetherpartsofthepar-
asitesvirulencearecausedbyaparasite-inducedincreaseinthetotalmetabolicrate,whichincreases
theproductionoffreeradicals,cellulardamageandacceleratesageingprocesses(vanLeeuwenetal.
2010).
Anotheropenquestionconcernshowenvironmentalheterogeneityinfluenceshost-parasitedynamics.
Environmentalconditionscanvarystronglyacrossthehost’shabitat,andhostswithinapopulationare
unlikelytoallexperiencethesameconditions.Ithasthereforebeenarguedthatenvironmentalhetero-
geneitymayaffectthegeneticdiversityofhostsandparasites(WolinskaandKing2009),theseverityof
disease outbreaks and virulence (Duffy et al. 2012). Such expectations could be tested with experi-
mentalevolution.Diseasecharacteristicscouldbecomparedbetweenparasitesoriginatingfromeither
evolutionarylines,wheretheco-evolvedhostsreceivedastandardamountoffood,orfromlineswhere
theco-evolvedhostsreceivedvariableamountsoffood.
Inexperimentalstudiesoftolerance,themostcommonlyusedmeasurementofparasiteburdenisthe
peak parasite density. This is done because parasite load increases with time after infection and is
therefore auto-correlated with longevity and intrinsic mortality rate. However, such measurements
misstheearlyphaseofinfection.Afurtherstepcouldbetoincorporatethegrowthoftheparasiteinto
measurementoftolerance.Onepossibleapproachcouldbetoestimatehowtheprobabilityofdying(at
differenttimepointsafterinfection)isinfluencedbythedifferenceinparasiteloadbetweennaturally
dying and living hosts. Suchmeasurements are further relevant as resistance and tolerance are ex-
pectedtochangethroughoutanindividual’slife(Loughetal.2015).
Asdescribedinchapterfour,toleranceandresistancecanbecorrelatedandincreaseasanevolution-
aryresponsetoparasiteinfection.Studying,whetherthesetraitcovariancesaretheresultfromadap-
tiveresponsestophysiological,environmental,orepidemiologicalfactorsorwhethertheyresultfrom
geneticlinkage(pleiotropy,linkagedisequilibriumorepistasis)couldbeanextstep.Quantitativeanaly-
sisofsuchgeneticcovariance is,as faras Iknow,still lackingandcouldbe investigatedbyecological
Impactofenvironmentonco-evolutionbetweenhostsandparasites
82
genomics.Theoreticalmodelscouldbeappliedtoidentifysituationsinwhichgeneticcovariationshould
havestrongco-evolutionaryconsequences.
Additionally,regionsinthegenomewithecologicallyrelevantfunctions,thatplayanimportantrolein
the evolution of tolerance, could be tracked. This could illuminate the genetic architecture of evolu-
tionarytransitionsbetweenantagonism,commensalismandevenmutualism.
6.3 Conclusion
Overallthisthesishasshownthattheenvironmentcaninfluencemanyaspectsofhost-parasiteinterac-
tions,onesthatplayimportantrolesinshapingevolutionarydynamics.Thetopicsthathavebeencov-
ered areof relevance for severalmain areas in evolutionary ecology, including life history evolution,
epidemiologyandresourceecology,andhave implications for futureresearch inthesefields.Bycon-
sideringthatenvironmentalconditionscanvarydrasticallyacrossthehost’shabitat,andbecause“the
only thing that is constant is change” (Heraclitus, ~ 500 BC), the knowledge presented in this thesis
mustbeconsideredto fullyunderstandhost-parasite interactionsandtheirco-evolution. Italsomust
be incorporatedwhenpredictingparasite evolutionandespeciallywhenmanagingparasites. The re-
sults presented here contribute towards a better understanding of host-parasites interactions, but
manyquestionsremain.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
83
Acknowledgements
IamverygratefultohavetheopportunitytoworkwithJacobKoella.Itwasaprivilegetowork
inthisinterestingfieldandtocollaboratewithsuchanexperiencedandpassionatescientist.Duringthe
many inspiringdiscussions I learneda lot fromhisclearanalyticalwayofthinkingandfromhisdirect
andveryefficientapproachtoaddressproblems.Ithankfortheguidance,adviceandinspiration.
AspecialthanksgoestoJonforallthecommentsonthemanuscripts,forthemanyfruitful(sometimes
controversial)discussionsaboutscienceandlifeingeneral.Thanksforthegoodtimesatunineandin
themountains. Further thanks go to Giacomo, Gael, Kevin, Antoine,Marion, Nathalie, Elise, Naiara,
Priscille,Adam,Marek,Pitou,Philip,LouisandLorenzforthehelpinthelab,discussions,commentsand
support.
IalsothankThomasFlatt,RolandRegoesandYannickMorêtfortheirvaluablecommentsonthemid-
thesisreportandespeciallyThomasFlattandFabriceHelfensteinfortheexaminationofthisthesis.
AverybigthanksgoestomyfamilyandmyfriendsfortheirsupportoverthewholeperiodofthePhD.
Miri,thanksfortheunlimitedsupportyouhaveprovidedmewithinthelaststressfulmonths,thanksfor
allthelittlethingsyouhavedoneforme,forhelpingoutinthelabduringtheverybuzzydaysandfor
allthepatienceandencouragementsduringtheseyears.Iamsohappytohavemetyou.
Thisworkwasfundedbygrant31003A_144207oftheSwissNationalScienceFoundation(SNF).
Neuchâtel,le6juillet2016
Impactofenvironmentonco-evolutionbetweenhostsandparasites
85
References
Abrams, P. A., and L. Rowe. 1996. The effects of predation on the age and size ofmaturity of prey.
Evolution50:1052–1061.
Adamo, S. 1999. Evidence for adaptive changes in egg laying in crickets exposed to bacteria and
parasites.Animalbehaviour57:117–124.
Agnew,P.,S.Bedhomme,C.Haussy,andY.Michalakis.1999.Ageandsizeatmaturityofthemosquito
CulexpipiensinfectedbythemicrosporidianparasiteVavraiaculicis.ProceedingsoftheRoyalSocietyof
London.SeriesB:BiologicalSciences266:947–952.
Agnew, P., and J. Koella. 1999a. Constraints on the reproductive value of vertical transmission for a
microsporidianparasiteanditsfemalekillingbehaviour.Journalofanimalecology68:1010–1019.
Agnew,P., and J. C. Koella. 1999b. Lifehistory interactionswithenvironmental conditions in ahost–
parasiterelationshipandtheparasite’smodeoftransmission.EvolutionaryEcology13:67–91.
Agnew,P., J. Koella, andY.Michalakis. 2000.Host lifehistory responses toparasitism.Microbesand
Infection2:891–896.
Alizon, S., a Hurford, N. Mideo, and M. Van Baalen. 2009. Virulence evolution and the trade-off
hypothesis:history,currentstateofaffairsandthefuture.Journalofevolutionarybiology22:245–59.
Alizon, S., and S. Lion. 2011. Within-host parasite cooperation and the evolution of virulence.
Proceedings.Biologicalsciences/TheRoyalSociety278:3738–47.
Alizon, S., and M. van Baalen. 2005a. Emergence of a convex trade-off between transmission and
virulence.TheAmericannaturalist165:E155–E167.
Alizon, S., and M. van Baalen. 2005b. Emergence of a convex trade-off between transmission and
virulence.TheAmericannaturalist165:E155–67.
Alonso-Alvarez, C., S. Bertrand, B. Faivre, and G. Sorci. 2007. Increased susceptibility to oxidative
Impactofenvironmentonco-evolutionbetweenhostsandparasites
86
damageasacostofacceleratedsomaticgrowthinzebrafinches.FunctionalEcology21:873–879.
Alto,B.W.,M.H.Reiskind,andL.P.Lounibos.2008.Sizealterssusceptibilityofvectorstodenguevirus
infectionanddissemination.AmericanJournalofTropicalMedicineandHygiene79:688–695.
Anderson, R. M., and R. M. May. 1979. Population biology of infectious diseases: Part I. Nature
280:361–367.
Anderson,R.M.,andR.M.May.1982.Coevolutionofhostsandparasites.Parasitology85(Pt2):411–
426.
Andreadis, T. G. 2007. Microsporidian parasites of mosquitoes. Journal of the American Mosquito
ControlAssociation23:3–29.
Antia, R., B. R. Levin, andR.M.May. 1994.Within-Host Population-Dynamics and the Evolution and
MaintenanceofMicroparasiteVirulence.TheAmericannaturalist144:457–472.
Arendt,J.D.1997.Adaptiveintrinsicgrowthrates:anintegrationacrosstaxa.TheQuarterlyReviewof
Biology72:149–177.
Arnott,S.A.,I.Barber,andF.A.Huntingford.2000.Parasite-associatedgrowthenhancementinafish-
cestodesystem.Proceedings.Biologicalsciences/TheRoyalSociety267:657–63.
Artavanis-Tsakonas,K.,J.E.Tongren,andE.M.Riley.2003.Thewarbetweenthemalariaparasiteand
the immune system: Immunity, immunoregulation and immunopathology. Clinical and Experimental
Immunology.
Ashby, B., and M. Boots. 2015. Coevolution of parasite virulence and host mating strategies.
ProceedingsoftheNationalAcademyofSciences201508397.
Auer,S.K.,J.D.Arendt,R.Chandramouli,andD.N.Reznick.2010.Juvenilecompensatorygrowthhas
negative consequences for reproduction in Trinidadian guppies (Poecilia reticulata). Ecology Letters
13:998–1007.
Ayres,J.S.,andD.S.Schneider.2009.Theroleofanorexiainresistanceandtolerancetoinfectionsin
Drosophila.(D.Promislow,ed.)PLoSbiology7:e1000150.
Bashey, F. 2006. Cross-generational environmental effects and the evolution of offspring size in the
Trinidadianguppy(Poeciliareticulata).Evolution60:348–361.
Becnel, J. J., and T. G. Andreadis. 1999. TheMicrosporidia andMicrosporidiosis. (L. M.Weiss &M.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
87
Wittner,eds.).AmericanSocietyofMicrobiology.
Bedhomme,S.,P.Agnew,C.Sidobre,andY.Michalakis.2004.Virulencereactionnormsacrossafood
gradient.Proceedings.Biologicalsciences/TheRoyalSociety271:739–44.
Berrigan,D.,andJ.C.Koella.1994.Theevolutionofreactionnorms:simplemodelsforageandsizeat
maturity.JournalofEvolutionaryBiology7:549–566.
Biron,D.G.,P.Agnew,L.Marché,L.Renault,C.Sidobre,andY.Michalakis.2005.ProteomeofAedes
aegyptilarvaeinresponsetoinfectionbytheintracellularparasiteVavraiaculicis.Internationaljournal
forparasitology35:1385–97.
Bittner, K., K.-O. Rothaupt, andD. Ebert. 2002. Ecological interactions of themicroparasite Caullerya
mesnilianditshostDaphniagaleata.Limnologyandoceanography47:300–305.
Boots,M.2008.Fightorlearntolivewiththeconsequences?TrendsinEcologyandEvolution.
Boots,M. 2011. The evolution of resistance to a parasite is determined by resources. The American
naturalist178:214–20.
Bremermann,H. J., and J.Pickering.1983.Agame-theoreticalmodelofparasitevirulence. Journalof
TheoreticalBiology100:411–426.
Brown,M.J.F.,R.Loosli,andP.Schmid-Hempel.2000.Condition-dependentexpressionofvirulencein
atrypanosomeinfectingbumblebees.Oikos91:421–427.
Carval, D., and R. Ferriere. 2010. A unified model for the coevolution of resistance, tolerance, and
virulence.Evolution64:2988–3009.
Chakrabarti, L.A.2004.Theparadoxof simian immunodeficiencyvirus infection in sootymangabeys:
activeviralreplicationwithoutdiseaseprogression.Frontiersinbioscience :ajournalandvirtuallibrary
9:521–39.
Chippindale,A.K.,A.M.Leroi,S.B.Kim,andM.R.Rose.1993.Phenotypicplasticityandselection in
Drosophila life-history evolution. I. Nutrition and the cost of reproduction. Journal of Evolutionary
Biology6:171–193.
Christe,P.,H.Richner,andA.Oppliger.1996.Begging, foodprovisioning,andnestlingcompetition in
greattitbroodsinfestedwithectoparasites.BehavioralEcology7:127–131.
Christophers, S. 1960. Aëdes aegypti (L.) the Yellow FeverMosquito: its Life History, Bionomics and
Impactofenvironmentonco-evolutionbetweenhostsandparasites
88
Structure.
Clements,A.N.1999.TheBiologyofMosquitoes.Springer.
D. C. Lopez-Pascua, L., and A. Buckling. 2008. Increasing productivity accelerates host-parasite
coevolution.JournalofEvolutionaryBiology21:853–860.
DeBlock,M.,andR.Stoks.2008.Compensatorygrowthandoxidativestressinadamselfly.Proceedings
oftheRoyalSocietyLondonSeriesB275:781–785.
De Roode, J. C., and S. Altizer. 2010. Host-parasite genetic interactions and virulence-transmission
relationshipsinnaturalpopulationsofmonarchbutterflies.Evolution64:502–514.
deRoode,J.C.,A.B.Pedersen,M.D.Hunter,andS.Altizer.2008.Hostplantspeciesaffectsvirulencein
monarchbutterflyparasites.TheJournalofanimalecology77:120–6.
De Roode, J. C., A. J. Yates, and S. Altizer. 2008. Virulence-transmission trade-offs and population
divergenceinvirulenceinanaturallyoccurringbutterflyparasite.ProceedingsoftheNationalAcademy
ofSciencesoftheUnitedStatesofAmerica105:7489–7494.
Dhahbi,J.M.,H.-J.Kim,P.L.Mote,R.J.Beaver,andS.R.Spindler.2004.Temporallinkagebetweenthe
phenotypic and genomic responses to caloric restriction. Proceedings of the National Academy of
Sciences,USA101:5524–9.
Dmitriew,C.M.2011.Theevolutionofgrowthtrajectories:whatlimitsgrowthrate?BiologicalReviews
86:97–116.
Duffy, M. a., J. H. Ochs, R. M. Penczykowski, D. J. Civitello, C. a. Klausmeier, and S. R. Hall. 2012.
EcologicalContextInfluencesEpidemicSizeandParasite-DrivenEvolution.Science335:1636–1638.
Ebert,D.,andK.L.K.Mangin.1997.Theinfluenceofhostdemographyontheevolutionofvirulenceof
amicrosporidiangutparasite.Evolution51:1828.
Ebert, D., C. D. Zschokke-Rohringer, and H. J. Carius. 2000. Dose effects and density-dependent
regulationoftwomicroparasitesofDaphniamagna.Oecologia122:200–209.
Ernande,B.,P.Boudry,J.Clobert,andJ.Haure.2004.Plasticityinresourceallocationbasedlifehistory
traits in the Pacific oyster, Crassostrea gigas. I. Spatial variation in food abundance. Journal of
EvolutionaryBiology17:342–356.
Ferguson, H.M., and A. F. Read. 2002. Genetic and environmental determinants ofmalaria parasite
Impactofenvironmentonco-evolutionbetweenhostsandparasites
89
virulenceinmosquitoes.Proceedings.Biologicalsciences/TheRoyalSociety269:1217–24.
Forbes,M.R.L.1993.ParasitismandHostReproductiveEffort.Oikos67:444.
Gaba,S.,andD.Ebert.2009.Time-shiftexperimentsasatooltostudyantagonisticcoevolution.Trends
inEcologyandEvolution.
Gandon,S.,P.Agnew,andY.Michalakis.2002.Coevolutionbetweenparasitevirulenceandhost life-
historytraits.TheAmericannaturalist160:374–388.
Gandon,S.,A.Buckling,E.Decaestecker,andT.Day.2008.Host-parasitecoevolutionandpatternsof
adaptationacrosstimeandspace.JournalofEvolutionaryBiology21:1861–1866.
Ganusov,V.V.,C.T.Bergstrom,andR.Antia.2002.Within-hostpopulationdynamicsandtheevoultion
ofmicroparasitesinaheterogenoushostpopulation.Evolution56:213–223.
Gérard,C.,andA.Théron.1997.Age/size-andtime-specificeffectsofSchistosomamansonionenergy
allocationpatternsofitssnailhostBiomphalariaglabrata.Oecologia112:447–452.
Gluckman,P.D.,andM.A.Hanson.2004.Livingwiththepast:evolution,development,andpatternsof
disease.Science305:1733–6.
Hall, S. R., C. Becker, and C. E. Cáceres. 2007. Parasitic castration: A perspective from a model of
dynamicenergybudgets.IntegrativeandComparativeBiology47:295–309.
Hall,S.R.,C.R.Becker,M.aDuffy,andC.E.Cáceres.2010.Variation inresourceacquisitionanduse
amonghostclonescreateskeyepidemiologicaltrade-offs.TheAmericannaturalist176:557–565.
Hall,S.R.,C.J.Knight,C.R.Becker,M.A.Duffy,A.J.Tessier,andC.E.Cáceres.2009a.Qualitymatters:
resourcequalityforhostsandthetimingofepidemics.Ecologyletters12:118–28.
Hall, S. R., J. L. Simonis, R. M. Nisbet, A. J. Tessier, and C. E. Cáceres. 2009b. Resource ecology of
virulence in a planktonic host-parasite system: an explanation using dynamic energy budgets. The
Americannaturalist174:149–62.
Harrison, E., A.-L. Laine, M. Hietala, and M. a Brockhurst. 2013. Rapidly fluctuating environments
constrain coevolutionary arms racesby impeding selective sweeps. Proceedings. Biological sciences /
TheRoyalSociety280:20130937.
Hayward,A.D.,D.H.Nussey,A.J.Wilson,C.Berenos,J.G.Pilkington,K.A.Watt,J.M.Pemberton,etal.
2014. Natural Selection on Individual Variation in Tolerance of Gastrointestinal Nematode Infection.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
90
PLoSBiology12:1–13.
Hentschel,B.T.,andR.B.Emlet.2000.Metamorphosisofbarnaclenauplii:Effectsof foodvariability
andacomparisonwithamphibianmodels.Ecology81:3495–3508.
Hochberg, M. E., and M. Baalen. 1998. Antagonistic coevolution over productivity gradients. The
AmericanNaturalist152:620–634.
Hochberg,M.E.,Y.Michalakis,andT.deMeeus.1992.Parasitismasa constrainton the rateof life-
historyevolution.JournalofEvolutionaryBiology5:491–504.
Howick, V. M., and B. P. Lazzaro. 2014. Genotype and diet shape resistance and tolerance across
distinctphasesofbacterialinfection.BMCevolutionarybiology14:56.
J.Ryder,J.,J.Hathway,andR.J.Knell.2007.Constraintsonparasitefecundityandtransmissioninan
insect-STDsystem.Oikos116:578–584.
Johnson,P.T.J.,J.M.Chase,K.L.Dosch,R.B.Hartson,J.A.Gross,D.J.Larson,D.R.Sutherland,etal.
2007. Aquatic eutrophication promotes pathogenic infection in amphibians. Proceedings of the
NationalAcademyofSciencesoftheUnitedStatesofAmerica104:15781–6.
Jokela, J., C. M. Lively, J. Taskinen, and A. D. Peters. 1999. Effect of starvation on parasite-induced
mortalityinafreshwatersnail(Potamopyrgusantipodarum).Oecologia119:320–325.
Jones,M.E.,A.Cockburn,R.Hamede,C.Hawkins,H.Hesterman,S.Lachish,D.Mann,etal.2008.Life-
historychange indisease-ravagedTasmaniandevilpopulations.ProceedingsoftheNationalAcademy
ofSciencesoftheUnitedStatesofAmerica105:10023–10027.
Kashiwagi, A., and T. Yomo. 2011. Ongoing phenotypic and genomic changes in experimental
coevolutionofrnabacteriophageq??andescherichiacoli.PLoSGenetics7.
Kawecki, T. J., and S. C. Stearns. 1993. The evolution of life histories in spatially heterogeneous
environments:optimalreactionnormsrevisited.EvolutionaryEcology7:155–174.
Kirkwood, T. B. L., and D. P. Shanley. 2005. Food restriction, evolution and ageing. Mechanisms of
AgeingandDevelopment126:1011–1016.
Koella,J.,andP.Agnew.1999.Acorrelatedresponseofaparasite’svirulenceandlifecycletoselection
onitshost'slifehistory.Journalofevolutionarybiology12:70–79.
Koella, J., P. Agnew, and Y.Michalakis. 1998. Coevolutionary interactions between host life histories
Impactofenvironmentonco-evolutionbetweenhostsandparasites
91
andparasitelifecycles.
Koella, J. C., and S. F. L. 2002. Effect of adult nutritionon themelanization immune responseof the
malariavectorAnophelesstephensi.MedicalandVeterinaryEntomology16:316–320.
Koella, J.C.,andE.O.Lyimo.1996.Variability in therelationshipbetweenweightandwing lengthof
Anophelesgambiae(Diptera:Culicidae).JournalofMedicalEntomology33:261–264.
Koella,J.C.,andO.Restif.2001.Coevolutionofparasitevirulenceandhostlifehistory.EcologyLetters
4:207–214.
Koella, J., L. Lorenz, and I. Bargielowski. 2009.Microsporidians as Evolution-Proof Agents ofMalaria
Control?Advancesinparasitology68:315–327.
Krasnov, B. R., I. S. Khokhlova, M. S. Arakelyan, and A. A. Degen. 2005. Is a starving host tastier?
Reproductioninfleasparasitizingfood-limitedrodents.FunctionalEcology19:625–631.
Krist, A. C. 2001. Variation in fecundity among populations of snails is predicted by prevalence of
castratingparasites.EvolutionaryEcologyResearch3:191–197.
Lafferty,K.D.1993.TheMarineSnail,Cerithideacalifornica,MaturesatSmallerSizesWhereParasitism
IsHigh.OIKOS68:3–11.
Lazzaro,B.P.,andT.J.Little.2009.Immunityinavariableworld.PhilosophicaltransactionsoftheRoyal
SocietyofLondon.SeriesB,Biologicalsciences364:15–26.
Lee,W.-S.,P.Monaghan,andN.B.Metcalfe.2013.Experimentaldemonstrationof thegrowth rate--
lifespantrade-off.ProceedingsoftheRoyalSocietyLondonSeriesB280:20122370.
Lefèvre,T.,A.J.Williams,andJ.C.deRoode.2010.Geneticvariationinresistance,butnottolerance,to
a protozoan parasite in the monarch butterfly. Proceedings. Biological sciences / The Royal Society
278:751–9.
Leips, J., and J. Travis. 1994.Metamorphic responses to changing food levels in two species of hylid
frogs.Ecology75:1345–1356.
Lemaître, J.-F., V. Berger, C. Bonenfant,M.Douhard,M.Gamelon, F. Plard, and J.-M.Gaillard. 2015.
Early-late life trade-offs and the evolution of ageing in the wild. Proceedings of the Royal Society
LondonSeriesB282:20150209.
Little, T. J., W. Chadwick, and K. Watt. 2008. Parasite variation and the evolution of virulence in a
Impactofenvironmentonco-evolutionbetweenhostsandparasites
92
Daphnia-microparasitesystem.Parasitology135:303–308.
Little,T. J.,D.M.Shuker,N.Colegrave,T.Day,andA.L.Graham.2010.Thecoevolutionofvirulence:
Toleranceinperspective.PLoSPathogens.
LopezPascua,L.,A.R.Hall,A.Best,A.D.Morgan,M.Boots,andA.Buckling.2014.Higher resources
decreasefluctuatingselectionduringhost-parasitecoevolution.EcologyLetters17:1380–1388.
Lorenz,L.M.,andJ.C.Koella.2011.ThemicrosporidianparasiteVavraiaculicisasapotentiallatelife-
actingcontrolagentofmalaria.EvolutionaryApplications4:783–790.
Lough,G., I.Kyriazakis,S.Bergmann,A.Lengeling,andA.B.Doeschl-Wilson.2015.Healthtrajectories
reveal the dynamic contributions of host genetic resistance and tolerance to infection outcome.
ProceedingsoftheRoyalSocietyB:BiologicalSciences282:20152151–20152151.
Lyimo,E.O.,andW.Takken.1993.Effectsofadultbodysizeon fecundityand thepre-gravid rateof
AnophelesgambiaefemalesinTanzania.MedicalandVeterinaryEntomology7:328–332.
Lynch,M., and B.Walsh. 1998. Genetics and analysis of quantitative traits. Genetics and analysis of
quantitativetraits.
Mackinnon, M. J., and A. F. Read. 1999. Genetic Relationships between Parasite Virulence and
TransmissionintheRodentMalariaPlasmodiumchabaudi.Evolution53:689–703.
Mackinnon,M.J.,andA.F.Read.2004.Virulenceinmalaria:anevolutionaryviewpoint.Philosophical
transactionsoftheRoyalSocietyofLondonSeriesBBiologicalsciences359:965–86.
Mair,W.,P.Goymer,S.D.Pletcher,andL.Partridge.2003.Demographyofdietaryrestrictionanddeath
inDrosophila.Science301:1731–1733.
Mangel,M.,andS.B.Munch.2005.Alife-historyperspectiveonshort-andlong-termconsequencesof
compensatorygrowth.AmericanNaturalist166:E155–176.
Masoro,E.J.2005.Overviewofcaloricrestrictionandageing.MechanismsofAgeingandDevelopment
126:913–922.
Mazé-Guilmo, E., G. Loot, D. J. Páez, T. Lefèvre, and S. Blanchet. 2014. Heritable variation in host
toleranceand resistance inferred fromawildhost-parasite system.Proceedings.Biological sciences /
TheRoyalSociety281:20132567.
McCann,S.,J.F.Day,S.Allan,andC.C.Lord.2009.Agemodifiestheeffectofbodysizeonfecundityin
Impactofenvironmentonco-evolutionbetweenhostsandparasites
93
CulexquinquefasciatusSay(Diptera:Culicidae).JournalofVectorEcology34:174–181.
Medzhitov,R.,D.S.Schneider,andM.P.Soares.2012.DiseaseToleranceasaDefenseStrategy.Science
335:936–941.
Mellors,J.W.,C.R.Rinaldo,P.Gupta,R.M.White,J.aTodd,andL.aKingsley.1996.PrognosisinHIV-1
infectionpredictedbythequantityofvirusinplasma.Science272:1167–1170.
Metcalfe,N.B.,andP.Monaghan.2001.Compensationforabadstart:grownow,paylater?Trendsin
Ecology&Evolution16:254–260.
Michalakis, Y., S. Bédhomme, D. G. Biron, A. Rivero, C. Sidobre, and P. Agnew. 2008. Virulence and
resistanceinamosquito–microsporidiuminteraction.EvolutionaryApplications1:49–56.
Miller,M.R.,A.White,andM.Boots.2005.Theevolutionofhostresistance:Toleranceandcontrolas
distinctstrategies.JournalofTheoreticalBiology236:198–207.
Miller,M.R.,A.White,andM.Boots.2006.Theevolutionofparasitesinresponsetotoleranceintheir
hosts:thegood,thebad,andapparentcommensalism.Evolution60:945–956.
Minchella,D.J.1985.Hostlife-historyvariationinresponsetoparasitism.Parasitology90:205.
Minchella, D. J., and P. T. Loverde. 1981. A cost of increased early reproductive effort in the snail
Biomphalariaglabrata.TheAmericanNaturalist118:876–881.
Moret,Y.2000.SurvivalforImmunity:ThePriceofImmuneSystemActivationforBumblebeeWorkers.
Science290:1166–1168.
Nene,V.,J.R.Wortman,D.Lawson,B.Haas,C.Kodira,Z.J.Tu,B.Loftus,etal.2007.Genomesequence
ofAedesaegypti,amajorarbovirusvector.Science316:1718–1723.
Nidelet,T.,J.Koella,andO.Kaltz.2009.Effectsofshortenedhostlifespanontheevolutionofparasite
lifehistoryandvirulenceinamicrobialhost-parasitesystem.BMCevolutionarybiology9.
Paterson, S., T. Vogwill, A. Buckling, R. Benmayor, A. J. Spiers,N. R. Thomson,M.Quail, et al. 2010.
Antagonisticcoevolutionacceleratesmolecularevolution.Nature464:275–278.
Perrin,N.,andP.Christe.1996.Onhostlife-historyresponsetoparasitism.Oikos67:317–322.
Petersen, L. R., D. J. Jamieson, A.M. Powers, andM. A. Honein. 2016. Zika Virus. TheNew England
journalofmedicine.
Impactofenvironmentonco-evolutionbetweenhostsandparasites
94
Pigliucci,M.2001.PhenotypicPlasticity:BeyondNatureandNurture.
Polak,M., andW.T. Starmer.1998.Parasite-induced riskofmortalityelevates reproductiveeffort in
maleDrosophila.Proceedings.Biologicalsciences/TheRoyalSociety265:2197–2201.
Pontier,D.,E.Fromont,F.Courchamp,M.Artois,andN.G.Yoccoz.1998.Retrovirusesandsexualsize
dimorphism in domestic cats (Felis catus L.). Proceedings. Biological sciences / The Royal Society
265:167–173.
Råberg, L., A. L. Graham, and A. F. Read. 2009. Decomposing health: tolerance and resistance to
parasitesinanimals.PhilosophicalTransactionsoftheRoyalSocietyB:BiologicalSciences364:37–49.
Råberg,L.,D.Sim,andA.F.Read.2007.Disentanglinggeneticvariationforresistanceandtoleranceto
infectiousdiseasesinanimals.Science318:812–814.
Read,A. F.,A. L.Graham,andL.Råberg.2008.Animaldefensesagainst infectiousagents: Isdamage
controlmoreimportantthanpathogencontrol?PLoSBiology6:2638–2641.
Regoes,R.R.,P.J.McLaren,M.Battegay,E.Bernasconi,A.Calmy,H.F.Günthard,M.Hoffmann,etal.
2014.DisentanglingHumanToleranceandResistanceAgainstHIV.PLoSBiology12.
Reiskind,M.H.,andL.P.Lounibos.2009.Effectsofintraspecificlarvalcompetitiononadultlongevityin
themosquitoesAedesaegyptiandAedesalbopictus.MedicalandVeterinaryEntomology23:62–8.
Restif, O., and A. L. Graham. 2015. Within-host dynamics of infection: from ecological insights to
evolutionarypredictions.PhilosophicaltransactionsoftheRoyalSocietyofLondon.SeriesB,Biological
sciences370:20140304–.
Restif,O.,M. E. Hochberg, and J. Koella. 2001. Virulence and age at reproduction: new insights into
host?parasitecoevolution.Journalofevolutionarybiology14:967–979.
Restif, O., and J. Koella. 2003. Shared control of epidemiological traits in a coevolutionarymodel of
host-parasiteinteractions.TheAmericanNaturalist161:827–836.
Restif, O., and J. Koella. 2004. Concurrent evolution of resistance and tolerance to pathogens. The
AmericanNaturalist164:E90–E102.
Reynolds, D. G. 1970. Laboratory studies of themicrosporidian Plistophora culicis (Weiser) infecting
CulexpipiensfatigansWied.Bulletinofentomologicalresearch60:339–49.
Richner, H. 1998. Host-parasite interactions and life-history evolution. Zoology-Analysis of Complex
Impactofenvironmentonco-evolutionbetweenhostsandparasites
95
Systems.
Richner, H., and F. Tripet. 1999. Ectoparasitism and the trade-off between current and future
reproduction.Oikos86:535–538.
Roff,D.A.1984.TheEvolutionofLifeHistoryParametersinTeleosts.CanadianJournalofFisheriesand
AquaticSciences41:989–1000.
Rose,M.R.1984.LaboratoryEvolutionofPostponedSenescenceinDrosophilamelanogaster.Evolution
38:1004–1010.
Rowe, L., and D. Ludwig. 1991. Size and timing of metamorphosis in complex life cycles: time
constraintsandvariation.Ecology72:413–427.
Roy, B. A., and J. W. Kirchner. 2000. Evolutionary dynamics of pathogen resistance and tolerance.
Evolution54:51–63.
Salvaudon, L., V. Héraudet, and J. a Shykoff. 2007. Genotype-specific interactions and the trade-off
betweenhostandparasitefitness.BMCevolutionarybiology7:189.
Schneider,D.S.,andJ.S.Ayres.2008.Twowaystosurviveinfection:whatresistanceandtolerancecan
teachusabouttreatinginfectiousdiseases.Naturereviews.Immunology8:889–895.
Shanley,D.P.,andT.B.L.Kirkwood.2000.Calorierestrictionandaging:alife-historyanalysis.Evolution
54:740–750.
Shetty,P.2010.Nutrition,immunityandinfection.CABIPublishing.
Simms,E.L.2000.Definingtoleranceasanormofreaction.EvolutionaryEcology14:563–570.
Simms, E. L., and J. Triplett. 1994. Costs and benefits of plant responses to disease: Resistance and
tolerance.Evolution48:1973–1985.
Smith, V. H. 1993. Applicability of resource-ratio theory to microbial ecology. Limnology and
Oceanography38:239–249.
Sorci,G.,G.Sorci,J.Clobert,J.Clobert,Y.Michalakis,andY.Michalakis.1996.Costofreproductionand
costofparasitisminthecommonlizard,Lacertavivipara.Oikos76:121–130.
Southwood,T.R.,G.Murdie,M.Yasuno,R.J.Tonn,andP.M.Reader.1972.Studiesonthelifebudget
of Aedes aegypti in Wat Samphaya, Bangkok, Thailand. Bulletin of the World Health Organization
Impactofenvironmentonco-evolutionbetweenhostsandparasites
96
46:211–26.
Spindler,S.R.2005.Rapidandreversibleinductionofthelongevity,anticancerandgenomiceffectsof
caloricrestriction.MechanismsofAgeingandDevelopment126:960–966.
Stearns,S.1992.TheEvolutionofLifeHistories.OxfordUniversityPress,London6:304–306.
Stearns, S. C., and J. C. Koella. 1986. The evolution of phenotypic plasticity in life-history traits:
predictionsofreactionnormsforageandsizeatmaturity.Evolution40:893–913.
Sternberg,E.D.,T.Lefèvre,J.Li,C.L.F.deCastillejo,H.Li,M.D.Hunter,andJ.C.deRoode.2012.Food
Plant Derived Disease Tolerance And Resistance In A Natural Butterfly-Plant-Parasite Interactions.
Evolution66:3367–3376.
Sumanochitrapon,W.,D.Strickman,R.Sithiprasasna,P.Kittayapong,andB.L.Innis.1998.Effectofsize
and geographic origin of Aedes aegypti on oral infection with dengue-2 virus. American Journal of
TropicalMedicineandHygiene58:283–286.
Suttle,C.a.2007.Marineviruses--majorplayersintheglobalecosystem.Naturereviews.Microbiology
5:801–812.
Sweeney, A. W., and J. J. Becnel. 1991. Potential of microsporidia for the biological control of
mosquitoes.ParasitologyToday7:217–220.
Tomori, O. 2004. Yellow fever: the recurring plague. Critical Reviews in Clinical Laboratory Sciences
41:391–427.
Tseng,M. 2006. Interactions between the parasite’s previous and current environmentmediate the
outcomeofparasiteinfection.TheAmericannaturalist168:565–71.
Vale,P.F.,M.Choisy,andT. J.Little.2013.Hostnutritionaltersthevariance inparasitetransmission
potentialHostnutritionaltersthevarianceinparasitetransmissionpotential.
Vale, P. F., and T. J. Little. 2009. Measuring parasite fitness under genetic and thermal variation.
Heredity103:102–9.
Vale, P. F., A. J.Wilson, A. Best,M. Boots, and T. J. Little. 2011. Epidemiological, evolutionary, and
coevolutionaryimplicationsofcontext-dependentparasitism.TheAmericannaturalist177:510–521.
van Leeuwen, I.M.M., J. Vera, and O.Wolkenhauer. 2010. Dynamic energy budget approaches for
modelling organismal ageing. Philosophical transactions of the Royal Society of London. Series B,
Impactofenvironmentonco-evolutionbetweenhostsandparasites
97
Biologicalsciences365:3443–54.
Vincent,C.M., andN.P. Sharp.2014. Sexual antagonism for resistanceand tolerance to infection in
Drosophilamelanogaster.Proceedings.Biologicalsciences/TheRoyalSociety281:20140987.
Weindruch, R. 1996. The retardation of aging by caloric restriction: studies in rodents and primates.
ToxicologicPathology24:742–745.
Weiser, J., and M. Coluzzi. 1972. The microsporidian Plistophora culicis Weiser, 1946 in different
mosquitohosts.Foliaparasitologica19:197–202.
Wolinska,J.,andK.C.King.2009.Environmentcanalterselectioninhost-parasiteinteractions.Trends
inparasitology25:236–44.
Woolhouse, M. E. J., and S. Gowtage-Sequeria. 2005. Host range and emerging and reemerging
pathogens.EmergingInfectiousDiseases11:1842–1847.
WorldHealthOrganisation.2002.DengueandDengueHaemorrhagigFever.WorldHealthOrganisation,
Geneva.
WorldHealthOrganization.2015.WorldMalariaReport2015.WorldHealth.
Yearsley,J.M.,I.Kyriazakis,andI.J.Gordon.2004.Delayedcostsofgrowthandcompensatorygrowth
rates.FunctionalEcology18:563–570.
Zeller, M., and J. C. Koella. 2016. Effects of food variability on growth and reproduction of Aedes
aegypti.EcologyandEvolution6:552–559.