Predation on Multiple Trophic Levels Shapes the Evolution of Pathogen Virulence

This is an elec onic ep in o he o iginal a icle. This ep in may di e om he o iginal in pagina ion and ypog aphic de ail. Au ho (s): Ti le: Yea : Ve sion: Please ci e he o iginal e sion: All ma e ial supplied ia JYX is p o ec ed by copy igh and o he in ellec ual p ope y igh s, and duplica ion o sale o all o pa o any o he eposi o y collec ions is no pe mi ed, excep ha ma e ial may be duplica ed by you o you esea ch use o educa ional pu poses in elec onic o p in o m. You mus ob ain pe mission o any o he use. Elec onic o p in copies may no be o e ed, whe he o sale o o he wise o anyone who is no an au ho ised use . P eda ion on Mul iple T ophic Le els Shapes he E olu ion o Pa hogen Vi ulence F iman, Ville-Pe i; Linds ed , Ca i a; Laakso, Jouni; Mappes, Johanna F iman, V.-P., Linds ed , C., Hil unen, T., Laakso, J. & Mappes, J. (2009). P eda ion on mul iple ophic le els shapes he e olu ion o pa hogen i ulence. PLoS ONE, 4 (8), e6761. 2009 P eda ion on Mul iple T ophic Le els Shapes he E olu ion o Pa hogen Vi ulence Ville-Pe i F iman 1,2 *, Ca i a Linds ed 1 , Teppo Hil unen 1 , Jouni Laakso 1,2 , Johanna Mappes 1 1Depa men o Biological and En i onmen al Science, Cen e o Excellence in E olu iona y Resea ch, Uni e si y o Jy a ¨skyla ¨, Jy a ¨skyla ¨, Finland, 2Depa men o Biological and En i onmen al Science, Uni e si y o Helsinki, Helsinki, Finland Abs ac The pa hogen i ulence is adi ionally hough o co-e ol e as a esul o ecip ocal selec ion wi h i s hos o ganism. In na u al communi ies, pa hogens and hos s a e ypically embedded wi hin a web o in e ac ions wi h o he species, which could a ec indi ec ly he pa hogen i ulence and hos immuni y h ough ade-o s. He e we show ha selec ion by p eda ion can a ec bo h pa hogen i ulence and hos immune de ence. Exposing oppo unis ic bac e ial pa hogen Se a ia ma cescens o p eda ion by p o ozoan Te ahymena he mophila dec eased i s i ulence when measu ed as hos mo h Pa asemia plan aginis su i al. This was p obably because he bac e ial an i-p eda o y ai s we e aded o wi h bac e ial i ulence ac o s, such as mo ili y o esou ce use e iciency. Howe e , he hos su i al depended also on i s alloca ion o wa ning signal ha is used agains a ian p eda ion. When in ec ed wi h mos i ulen ances al bac e ial s ain, hos la ae wi h a small wa ning signal su i ed be e han hose wi h an e ec i e la ge signal. This sugges s ha la al immune de ence could be aded o wi h e ec i e de ence agains bi d p eda o s. Howe e , he signal size had no e ec on la al su i al when less i ulen con ol o e ol ed s ains we e used o in ec ion sugges ing ha an i-p eda o y de ence agains a ian p eda o s, migh be less cons ained when he in ading pa hogen is a he low in i ulence. Ou esul s demons a e ha p eda ion can be impo an indi ec d i e o he e olu ion o bo h pa hogen i ulence and hos immuni y in communi ies wi h mul iple species in e ac ions. Thus, he pa hogen i ulence should be iewed as a esul o bo h pas e olu iona y his o y, and cu en ecological in e ac ions. Ci a ion: F iman V-P, Linds ed C, Hil unen T, Laakso J, Mappes J (2009) P eda ion on Mul iple T ophic Le els Shapes he E olu ion o Pa hogen Vi ulence. PLoS ONE 4(8): e6761. doi:10.1371/jou nal.pone.0006761 Edi o : Niyaz Ahmed, Uni e si y o Hyde abad, India Recei ed June 4, 2009; Accep ed July 13, 2009; Published Augus 25, 2009 Copy igh : ß2009 F iman e al. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed. Funding: The s udy was unded by he Academy o Finland (Cen e o Excellence in E olu ion Resea ch and P ojec No. 1106993). The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip . Compe ing In e es s: The au ho s ha e decla ed ha no compe ing in e es s exis . * E-mail: [email p o ec ed] In oduc ion The pa hogen i ulence has been adi ionally hough o co- e ol e in ecip ocal selec ion wi h i s hos o ganism [1]. Howe e , in na u e pa hogens and hei hos s a e ypically embedded wi hin a web o in e ac ions wi h o he species, which could a ec indi ec ly he e olu ion o pa hogen i ulence and hos immuni y [2-4]. Fo example, p eda ion could inc ease o dec ease he p e alence o in ec ious diseases depending on how i a ec s he equency o in ec ed indi iduals o high-quali y hos s in he popula ion [2,4]. Mo eo e , ecen indings sugges ha p eda ion could a ec also di ec ly he pa hogen i ulence (abili y o ha m hos ) and hos immuni y h ough ade-o s o posi i e gene ic co ela ions wi h ai s connec ed o an i-p eda o y de ence [5-7]. The e o e, mul i- ophic-le el p eda ion could be impo an selec i e o ce a ec ing he e olu ion o diseases in na u al communi ies [2]. Ye , expe imen al s udies whe e he e olu iona y consequences o p eda ion on bo h pa hogen i ulence and hos immuni y had been es ed simul aneously a e a e. P o ozoan p eda ion could inc ease he pa hogen i ulence because bac e ial de ensi e adap a ions migh also ha e a signi ican ole in bac e ial pe sis ence and i ulence [6,8]. Fo example, p o ozoa could be impo an o he en ichmen o po en ially mo e pa hogenic, bio ilm- o ming Vib io chole ae s ains, which is he p incipal cause o chole a epidemics [9]. In addi ion, he su i al and success ul eplica ion o bac e ia inside he p o ozoan cells ha e p obably ga e ise o se e al acul a i e and obliga e in acellula pa hogens, such as Lis e ia,Ricke sia,Mycobac e ium,Legionella and Chlamydia [10] because he amoebae and mac ophages sha e analogous phagocy ic mechanisms, e.g. p ey ecogni ion by cell su ace ecep o s [11], p ey killing by oxygen adicals [12] and simila diges i e enzymes [13]. Al e na i ely, p o ozoan p eda ion could also lead o a dec ease in bac e ial i ulence i inc eased alloca ion o an i-p eda o y ai s is aded o wi h i ulence ac o s o he pa hogen. Su p isingly, his hypo hesis has no been es ed empi ically ye , e en hough p e ious esea ch sugges s ha alloca ion o ai s connec ed o bo h de ence and i ulence can be cos ly [14,15]. Fo example, p eda ion by p o ozoan Te ahymena he mophila has been shown o cause a apid e olu iona y inc ease in he an i-p eda o de ence o an oppo unis ic bac e ium pa hogen Se a ia ma cescens [15]. Howe e , p eda ion also dec eased S. ma cescens’ abili y o use esou ces e icien ly and dec eased he syn hesis o he ed pigmen p odigiosin [15], which has been linked o he exp ession o se e al i ulence ac o s in his pa hogen [16] (Fig. 1). The e o e, p o ozoan p eda ion could lead o an e olu iona y inc ease o dec ease in bac e ial i ulence depending on which bac e ial de ensi e ai s a e selec ed o , and how hese ai s a e co ela ed wi h bac e ial i ulence ac o s. Simila ly, p eda ion could a ec di ec ly he s eng h o he hos immune sys em because he esou ces a e o en limi ed and an i- PLoS ONE | www.plosone.o g 1 Augus 2009 | Volume 4 | Issue 8 | e6761 p eda o y de ence and immuni y incu cos s [5,17-20]. In addi ion, an i-p eda o y de ences can be aded o wi h ai s a ec ing hos immuni y [5,21,22]. Fo example, Rigby and Jokela [5] ound ha high in es men in de ence agains p eda o s inc eased eshwa e snail’s (Lymnaea s agnalis) suscep ibili y o pa hogens. Thus, inc easing alloca ion o de ence agains p eda ion could ha e indi ec cos s o educed hos immunocompe ence [5]. In e es ingly, his kind ade- o s has been shown o play impo an ole in de e mining he s uc u e o na u al communi ies [7]. To s udy how selec ion by p eda ion a ec s bac e ial i ulence and hos immune de ence, h ee di e en s ains o he ubiqui ous bac e ium S. ma cescens wi h di e en e olu iona y his o ies we e used o in ec Pa asemia plan aginis A c iid hos mo h la ae om wo selec ion lines ha di e ed in an i-p eda o y de ence agains a ian p eda o s. The bac e ial s ains included an ances o S. ma cescens s ain (ATCC s ain #13880) and wo ances o -de i ed s ains which had been le o e ol e in he absence (con ol s ain) o in he p esence (e ol ed s ain) o p o ozoan p eda o , Te ahymena he mophila, o 14 weeks [15]. All h ee bac e ial s ains consis ed o a mix u e o andomly isola ed clones (see me hods). Few po en ial i ulence ac o s o e e y bac e ial s ain we e also measu ed (mo ili y and di e si y). The hos selec ion lines we e a i icially selec ed o ha e a small (mo e melanic) o a la ge (less melanic) o ange pa ch exp essed on an o he wise black body [23]. This wa ning signal is used o indica e unp o i abili y o bi d p eda o s and ac s hus as an i-p eda o y de ence [24,25] (Fig. 1). The bac e ia in he genus Se a ia a e common pa hogens o many insec s including he Lepidop e a [26] and he e o e hese s udy species can also po en ially encoun e in he wild. In he in ec ion expe imen , h ee hos g oups wi hin bo h selec ion lines we e in ec ed wi h one o he ee bac e ial s ains. In addi ion, s e ilized wa e was injec ed in o he ou h g oups o la ae o con ol he physical damage caused by he injec ion i sel . La al su i al was moni o ed o 72 h om in ec ion h ee imes pe day by sco ing he la ae as dead o ali e. On he basis o p e ious expe imen , we hypo hesied ha p o ozoan p eda ion would lead o dec eased pa hogen i ulence due o ade-o wi h an i-p eda o y and o he li e-his o y ai s (e.g. esou ce use abili y, [15]). Simila ly, hos alloca ion o a la ge wa ning signal, i.e s ong de ence agains a ian p eda o s could lead o dec eased immune de ence h ough educ ion in he amoun o la al cu icula melanin (Fig. 1). This is because he black, melanin-based pigmen is known o co ela e wi h highe phenoloxidase enzyme ac i i y, which is an impo an pa o he humo al immune esponse cascade in insec s [27,28]. Resul s When he la al su i al was analysed o e he hos signal lines, ances o and con ol S. ma cescens s ains dec eased he la al su i al compa ed o la ae in ec ed wi h he S. ma cescens s ain e ol ed in he p esence o p eda o s (ances o s. e ol ed, Chi- Squa e = 7.32, P,0.007, con ol s. e ol ed, Chi-Squa e = 3.87, P,0.049 and con ol s. ances o , Chi-Squa e = 1.17, P= 0.279, Fig. 2a). The la al su i al did no di e be ween signal lines (Chi- Squa e = 0.23, P= 0.627). Howe e , he bac e ial s ains had di e en e ec s on la al su i al when analysed wi hin he la ge and small signal lines sepa a ely (signal line se as a s a i ied ac o in he analysis). Wi hin he small wa ning signal line, all he bac e ial s ains had simila e ec s on la al su i al (all P.0.05, Fig. 2b). In con as , wi hin he la ge signal line, he la al su i al was highe wi h con ol and e ol ed s ains, compa ed o he ances o s ain (con ol s. ances o : Chi-Squa e = 6.47, P= 0.011, e ol ed s. ances o : Chi-Squa e = 11.47, P= 0.001, Fig. 2c). The la al su i al was analysed also wi hin di e en bac e ial ea men s (bac e ial ea men se as s a i ied ac o in he su i al analysis). When he mos i ulen ances o s ain was used o in ec ion, he la ae wi h a small wa ning signal had highe su i al compa ed o la ae wi h a la ge wa ning signal Figu e 1. The e ec o p eda ion on he e olu ion o hos -pa hogen in e ac ion h ough ade-o s. doi:10.1371/jou nal.pone.0006761.g001 E olu ion o Vi ulence PLoS ONE | www.plosone.o g 2 Augus 2009 | Volume 4 | Issue 8 | e6761 (Chi-Squa e = 4.85, P= 0.028, Fig. 2b and c). Howe e , signal line had no e ec on la al su i al when con ol (Chi-Squa e = 0.54, P= 0.459, Fig. 2b and c) o e ol ed s ain was used o in ec ion (Chi-squa e = 1.033, P= 0.309, and also wi h Cox- eg ession analysis me hod, Wald s a is ics = 4.76, P= 0.092, Fig. 2b and c). Damage caused by he injec ion alone was negligible as none o he la ae died when injec ed wi h s e ilized wa e (Fig. 2a). In addi ion, alloca ion o a la ge wa ning signal did no dec ease he encapsula ion esponse o he la ae (small s. la ge signal line, F 1, 172 = 0.02, P= 0.882). The mo ili y o con ol and e ol ed bac e ial s ains was conside ably lowe compa ed o ances o s ain ( - es o di e ence in bo h cases P,0.001, Fig. 3a). Ye , he mo ili y o e ol ed bac e ial s ain was lowes ( - es o di e ence be ween con ol and e ol ed s ain, P= 0.014, Fig. 3a). The equency o ed pigmen exp essing bac e ial clones was highes wi h he ances o , in e media e wi h he con ol, and lowes wi h he e ol ed bac e ial s ain (main e ec o bac e ial s ain, F 1, 8 = 62.69, P,0.001, in all pai wise compa isons P= 0.002 o smalle , Fig. 3b). Con e sely, he equency o non-pigmen ed bac e ial clones was highes wi h he e ol ed, in e media e wi h he con ol, and lowes wi h he ances o bac e ial s ain (main e ec o bac e ial s ain, F 1, 8 = 62.69, P,0.001, in all pai wise compa isons P= 0.005 o smalle , Fig. 3b). The di e si y was lowes wi h he ances o s ain bu did no di e be ween con ol and e ol ed s ains (main e ec o bac e ial s ain, F 1, 8 = 50.83, P,0.001, ances o s. con ol o e ol ed s ain P,0.001, con ol s. e ol ed s ain, P= 0.11, Fig. 3b). Discussion We ound ha inc eased alloca ion o de ence agains p o ozoan p eda ion dec eased he S. ma cescens’ i ulence (Fig. 2a). The bac e ial s ains’ e ec on hos su i al depended also on he hos alloca ion o wa ning signal used o de ence agains a ian p eda ion: all he bac e ial s ains had simila e ec s on la al su i al wi hin he small wa ning signal (Fig. 2b), while he la al su i al was highe wi h con ol and e ol ed s ains compa ed o he ances o s ain wi hin he la ge signal line (Fig. 2c). This sugges s ha a pa hogen’s abili y o cause in ec ions does no only depend on i s own pas e olu iona y his o y, bu is also a ec ed by he gene ic backg ound o i s hos . Figu e 2. The su i al o hos la ae in ec ed wi h bac e ial s ains di e ing in hei e olu iona y his o ies. Panel a: su i al (%) o P. plan aginis mo h la ae when in ec ed wi h ances o (black iangles), con ol (g ey iangles), and e ol ed (whi e iangles) s ains o he bac e ia S. ma cescens. The s aigh line (black ci cles) deno es he su i al o con ol la ae injec ed wi h s e ilized wa e . Su i al (%) o P. plan aginis mo h la ae wi hin he small, panel b, and la ge, panel c, wa ning signal lines. doi:10.1371/jou nal.pone.0006761.g002 Figu e 3. The mean mo ili y (panel a) and di e si y (panel b) o di e en bac e ial s ains. In panel a, ances o e sus con ol s ain (P= 0.001), ances o e sus e ol ed s ain (P,0.001) and con ol e sus e ol ed s ain (P= 0.014), N = 3 o e e y s ain. In panel b, he mean colony equencies (ba s) and di e si y (line) o di e en bac e ial s ains de e mined on he basis o he syn hesis o ed pigmen , p odigiosin. doi:10.1371/jou nal.pone.0006761.g003 E olu ion o Vi ulence PLoS ONE | www.plosone.o g 3 Augus 2009 | Volume 4 | Issue 8 | e6761 When he e ec s o bac e ial s ains we e compa ed be ween he signal lines, and when he mos i ulen ances al s ain was used o in ec ion, he la ae wi h small wa ning signal had highe su i al compa ed o la ae wi h la ge wa ning signal (Fig. 2a and b). This sugges s ha alloca ion o e ec i e de ence agains bi d p eda o s (i.e. la ge wa ning signal size) can ade o wi h immune de ence agains pa hogenic bac e ia. Howe e , he signal size had no e ec on la al su i al when less i ulen con ol o e ol ed s ains we e used o in ec ion (Fig. 2b and c). Thus, alloca ion o e ec i e wa ning signal, i.e. mo e e ec i e an i-p eda o y de ence agains isual p eda o s, migh be less cons ained when he in ading pa hogen is a he low in i ulence (Fig. 2c). P o ozoan-d i en e olu iona y dec ease in S. ma cescens’ i u- lence could be explained by dec eased mo ili y o esou ce use abili y [15]. Dec eased mo ili y has been shown o be connec ed o he dec eased i ulence o Campylobac e jejunum in pigle s [29] and ha o Db1140 S. ma cescens s ain in C. elegans [30]. Mo ili y could o example a ec he bac e ial abili y o each a ou able habi a s wi hin he hos [14]. We obse ed ha p o ozoan p eda ion dec eased he mo ili y o S. ma cescens mos ela i e o he ances o s ain (Fig. 3a). A dec ease in bac e ial mo ili y, caused p obably by down- egula ion o lagellum syn hesis o o he ac o s ela ed o bac e ial mo ili y [31,32], could ha e educed he p eda o encoun e a e leading o mo e de ensi e bu less i ulen bac e ia [33,34]. Since he mo ili y o he con ol s ain was lowe ela i e o he ances o , bu highe ela i e o he e ol ed s ain, demons a es ha selec ion by labo a o y condi ions could no alone selec o dec eased i ulence in S. ma cescens (Fig. 3a). Replica ion o he in ec ion expe imen wi h ano he hos , g ea e wax mo h (Galle ia mellonella), ga e consis en esul s adding mo e suppo o iew ha me e labo a o y condi ions can dec ease he S. ma cescens i ulence bu he exposu e o p o ozoan p eda ion dec eases he S. ma cescens i ulence mos (F iman & Mikon an a, unpublished, Figu e S1). Ano he explana ion o he dec eased i ulence in addi ion o mo ili y could lie in he p eda o -induced dec ease in p ey esou ce use abili y [15]; he less i ulen S. ma cescens could simply be ine icien in ob aining esou ces wi hin he hos , leading o poo e ep oduc ion, and hus a less ha m ul in ec ion. This hypo hesis is suppo ed by he ac ha he equency o whi e S. ma cescens clones ha we e poo a using esou ces [15] was highes wi h he e ol ed s ain (Fig. 3b). In addi ion, in ec ion wi h mix u e o clones ins ead o a single clone can a ec he bac e ial i ulence indi ec ly h ough compe i i e o co-ope a i e in e ac ions be ween di e en bac e- ial clones [35,36]. Fo example, i bac e ia a e mainly compe ing o esou ces wi hin hei hos , pa hogen di e si y is expec ed o inc ease i ulence h ough e icien esou ce use and as hos exploi a ion [37]. Howe e , i bac e ia co-ope a e, i.e. use exoenzymes o ex ac nu ien s, high ela edness (low di e si y) is expec ed o lead highes esou ce use and hus also highes i ulence [38]. Ou da a suppo s he o me hypo hesis because highes i ulence was a ained wi h leas di e se mix u e o clones, i.e. wi h he ances o s ain (Fig. 3b). In he case o S. ma cescens, p oduc ion o i on sca enging side opho es could be one possible o m o co-ope a ion a ec ing also o i s i ulence [39]. Howe e , o assess hese po en ial explana ions in mo e de ail, in ec ion expe imen s based on indi idual clones a e needed o cha ac e ise he gene ic co ela ions be ween he ai s connec ed o de ence agains p eda ion and i ulence in S. ma cescens. We ound ha a small wa ning signal, i.e. mo e melanic colou a ion o la ae, was posi i ely linked wi h de ence agains he mos i ulen ances o S. ma cescens s ain (Fig. 2b–c). Howe e , he possible cos s o p oducing melanin could ha e coun e balanced he immunological bene i s o mo e melanic la ae when in ec ed wi h less pa hogenic bac e ia [24]. P e ious s udies ha e also shown ha he la ae wi h la ge wa ning signal a e as e a g owing [40,24] and mos likely explana ion o his is ha hey simply consume mo e ood. This could also help hem o boos hei immune sys em because many plan s con ain oxins, which can be used o example o igh agains pa asi es [41]. The e o e, i is possible ha la ae wi h la ge wa ning singal we e able o compensa e hei immune sys em by seques e ing plan oxins mo e e icien ly a e and du ing he in ec ion ( ood plan , Ta axacum o icinale, was a ailable h oughou he in ec ion expe imen ). In gene al, P. plan aginis la ae seemed o alloca e mo e on de ence agains pa asi oids han pa hogens. The amoun o cu icula melanin did no a ec he s eng h o he la al encapsula ion esponse, which is gene ally used o desc ibe he immunocompe ence o insec s agains na u al pa asi oids [27,42,43]. Thus, e ec i e de ence agains mac opa asi es could ha e been impo an o P. plan aginis in i s pas e olu iona y his o y because alloca ion o an e ec i e wa ning signal dec eased only i s esis ance agains bac e ial pa hogens. Howe e , mo e de ailed immunological measu es a e needed o ully unde s and he linkage be ween insec cu icula melanin and immune de ence agains pa asi es and bac e ial pa hogens. Based on p eda o s’ lea ning e iciency, selec ion is assumed o lead o uni o mi y and conspicuousness in signal exp ession he eby dec easing he a ia ion in signal size [25,44]. Howe e , a ia ion in wa ning signal exp ession is common, which sugges s ha he s eng h and di ec ion o selec ion on signal size could a y spa ially and empo ally [24]. We p opose ha he obse ed a ia ion in wa ning signal pa e n o P. plan aginis la ae [24] could be pa ly explained wi h con as ing selec ion by a ian p eda o s and bac e ial pa hogens om di e en ophic le els. Thus, la ge wa ning signal size could be a ou ed when bi ds a e he main cause o la al mo ali y [25] and he pa hogens a e a he ha mless. Con e sely, when he isk o bac e ial in ec ion is high (e.g. du ing he win e hibe na ion pe iod), la ae wi h small wa ning signal, and be e immune de ence could ha e ad an age. Mos impo an ly, ou expe imen shows ha he pa hogen success is no only dependen on i s own e olu iona y his o y bu is also hugely a ec ed by he hos geno ype. The e is cu en ly conside able knowledge abou he gene ic p ope ies and mechanisms ha a e essen ial o many pa hogenic bac e ia o be able o colonise and in ec hei hos s [42,45-47]. A he same ime, ela i ely li le is known abou he selec i e agen s and en i onmen al condi ions ha igge ha mless bac e ia o quickly e ol e in o disease-causing pa hogens [1]. He e we show ha p o ozoan p eda ion can dec ease he i ulence o oppo u- nis ic bac e ial pa hogen S. ma cescens. Mo eo e , he hos immune de ences can e ol e indi ec ly in esponse o o he species in e ac ions, such as p eda ion, which can u he a ec he success o pa hogens. These esul s demons a e ha i ulence is a unc ion o bo h pas e olu iona y his o ies and p esen ecological in e ac ions o hos s and pa hogens. Thus, in o de o unde s and he eme gence and dynamics o diseases i could be necessa y o unde s and how e olu ion a ec s he pa hogen’s abili y o cause diseases and he hos ’s abili y o esis in ec ions in communi ies wi h mul iple species in e ac ions. This could be achie ed by b inging he p inciples o communi y ecology, e olu iona y heo y and hos -pa hogen epidemiology oge he . Ma e ials and Me hods Bac e ial s ains and in ec ion o la ae The bac e ial clones o e ol ed and con ol s ains we e o iginally isola ed om wo ea men s used in he p e ious E olu ion o Vi ulence PLoS ONE | www.plosone.o g 4 Augus 2009 | Volume 4 | Issue 8 | e6761 expe imen [15] whe e bac e ia we e cul u ed in he absence o p esence o p o ozoa Te ahymena he mophila (ATCC s ain #30008) o 14 weeks (app oxima ely 2500 bac e ial gene a ions). In bo h o hese ea men s, bac e ia we e o iginally he same as he ances o S. ma cescens s ain ecei ed om he Ame ican Type Cul u e collec ion (ATCC s ain #13880), which was chosen as he hi d bac e ial s ain. All bac e ial s ains comp ised a mix u e o 48 andomly isola ed clones ( ou mic ocosm eplica es we e used pe bac e ial s ain and 12 clones we e isola ed andomly pe eplica e). Wi h ances al s ain, all clones we e isola ed om one aga pla e. The ubiqui ous bac e ia S. ma cescens can be ound in bo h aqua ic and soil ecosys ems, and ha e an ex emely b oad hos ange including plan s, nema odes, insec s and mammals [26,45-47]. Bac e ial s ains we e i s cul i a ed a 25uConNB aga pla es (con aining 2.5 g o yeas ex ac , 10 g o nu ien b o h and 15 g o aga in one li e o dH 2 O). A e 48h o g ow h, bac e ial colonies we e inocula ed in phospha e bu e [15] and dilu ed o op ical densi ies con aining he same numbe o bac e ial cells (colony o ming uni s, main e ec o bac e ial s ain, F 1, 8 = 0.129, P= 0.881). The ea e , he s ains we e di ided o aliquo s, mixed wi h glyce ol and s o ed a -80uC o la e use. Be o e in ec ion, aliquo s o all s ains we e hawed o 1 h a 25uC. Fi e ml o well-mixed bac e ial solu ion (app oxi- ma ely 1.66*10 6 bac e ial cells), o s e ilized wa e o he con ols, we e injec ed be ween he second and hi d segmen s o la ae wi h a 10 ml Hamil on sy inge. A o al o 232 la ae we e injec ed du ing he expe imen : 55 con ol la ae injec ed wi h wa e , 60 la ae wi h he ances o S. ma cescens s ain, 60 la ae wi h e ol ed S. ma cescens s ain, and 57 la ae wi h con ol S. ma cescens s ain. In ec ion ook place o e 6 consecu i e days unde cons an labo a o y condi ions. La al su i al was no a ec ed by he in ec ion day, F 5, 231 = 0.4, P= 0.847. Be o e in ec ion, all la ae we e weighed (only la ae be ween 90 and 160 mg we e used) and assigned o ou g oups (pe signal line) wi h app oxima ely he same mean weigh o exclude possible condi ion dependen e ec s (la al weigh be ween he bac e ial ea men s had no e ec on su i al, F 3, 230 = 0.96, P= 0.412, mean weigh 6S.E. o la ae in ec ed wi h wa e : 114.3662.37 mg, he ances o s ain: 116.662.27 mg, he e ol ed s ain: 111.5662.21 mg, o wi h he con ol S. ma cescens s ain: 113.1261.91 mg.). Each bac e ial ea men g oup consis ed o , on a e age, 58 la ae; 29 wi h a small and 29 wi h a la ge signal (see below). In ec ion expe imen was also eplica ed wi h wax mo h la ae (Galle ia mellonella) o each bac e ial s ain (N = 12) as explained abo e wi h he excep ion ha he amoun o in ec ed bac e ial cells was conside ably lowe (app oxima ely 10-30 bac e ial cells). The injec ion me hod has been c i icized because i bypasses he en y o mic obes h ough na u al ou es o in ec ion, e.g. o ally [45]. Howe e , bac e ia also access hemocoel di ec ly h ough b each- ing he cu icle [47] and bo h in ec ion me hods (injec ion o o al inges ion) ha e been used wi h S. ma cescens o in ec a wide ange o insec hos s [46,47]. T ai measu emen s o he bac e ial s ains Bac e ial mo ili y assays we e done by s abbing ace amoun (2 ml) o each bac e ial s ain on he cen e o semi- luid NB aga pla es con aining 0.7% o aga wi h s e ile plas ic loops (VWR). The mo ili y o s ains was de e mined as he a ea (mm 2 ) bac e ia we e able o colonise on he aga pla es in 24h (N = 3 o e e y s ain). The equencies o ed (p odigiosin pigmen exp essing) and whi e (non-pigmen ed) bac e ial clones we e coun ed om 3 eplica e pla es o each bac e ial s ain. The bac e ial di e si y was es ima ed as Shannon di e si y index on he basis o ed and whi e colony equencies. Hos selec ion lines di e ing in an i-p eda o y de ence Selec ion lines on he ex emes o wa ning signal size in P. plan aginis we e es ablished in 2004. The la ae we e ea ed unde con olled labo a o y condi ions: empe a u e, ea ing densi y and ood esou ce (dandelion, Ta axacum sp.) we e kep cons an . Fi y- one amilies we e used o ob ain selec ion lines o di e gen pheno ypes ( o la ge and small o ange signals) by applying a unca ed amily selec ion p o ocol [23]. In o he wo ds, indi iduals wi h la ge (p opo ion o o ange o e 46% o la ae, i.e..5 o ange segmen s) and small (p opo ion o o ange less han 31% o la ae, i.e.,5 o ange segmen s) signals we e selec ed and hen c ossed wi hin he selec ion lines o se e al gene a ions. A e he 7 h gene a ion o selec ion, he size o he wa ning signal (p opo ion o he o ange signal in ela ion o he whole body) was on a e age 3060.1% wi hin small, and 5260.1% wi hin la ge, selec ion lines (one-way Ano a, F 1, 230 = 209, P,0.001). Nei he la al weigh no leng h di e ed be ween selec ion lines con i m- ing ha signal sizes we e no a esul o di e ences in body size (one-way ANOVA o weigh and leng h: F 1, 230 = 195, P= 0.659 and F 1, 230 = 0.45, P= 0.502, espec i ely). Encapsula ion assessmen Encapsula ion eac ion is a gene al esponse o o eign in usions in insec s [27,43]. The encapsula ion esponse o all la ae was measu ed be o e he bac e ial injec ion. La ae we e anaes he ized wi h CO 2 , a e which a small nylon implan was inse ed inside he la ae be ween he second and he hi d segmen s. The immune sys em o he la ae was allowed o eac o 5 hou s. Subsequen ly, he implan was emo ed, d ied and pho og aphed unde a mic oscope wi h 106magni ica ion wi h a Panasonic w -CL702 ideo eco de . The mean g ey alue o he implan was measu ed wi h ImageP o Plus 4.0 (Media Cybe ne - ics) on 1 mm o he implan , measu ed om he end implan ed inside he la a. The g ey alue o he backg ound was sub ac ed om he g ey alue o he implan o co ec o any a ia ion in ligh ing du ing pho og aphy. Highe g ey alues (da ke implan ) indica ed a s onge encapsula ion esponse. S a is ical analysis Cox- eg ession model was buil o es i he weigh o he la ae be o e injec ion was a signi ican co a ia e in he model. Due o high insigni icance (P= 0.638 and coe icien 1.005), he la al weigh was omi ed om he inal analysis and Kaplan-Meie su i al analysis and Log- ank s a is ics we e used in analysis. The main e ec s o bac e ial ea men and signal line we e analysed i s a e he e ec o bac e ial ea men wi hin he signal line and he e ec o signal line wi hin he bac e ial ea men s we e analysed using s a i ica ion. The Righ censo ing me hod was used o include he la ae ha did no die wi hin 72 hou s in he analysis. The encapsula ion abili y o la ae was analysed wi h a one-way ANOVA, and when mul iple g oups we e compa ed, a wo-way ANOVA was used. Suppo ing In o ma ion Figu e S1 The su i al o al e na i e hos , Wax mo h la ae (Galle ia mellonella), when in ec ed wi h ances o (black iangles), con ol (g ey iangles), and e ol ed (whi e iangles) s ains o he bac e ia S. ma cescens. The s aigh line (black ci cles) deno es he su i al o con ol la ae injec ed wi h s e ilized wa e (ances o s. con ol o e ol ed s ain, P = 0.032 and p,0.001 espec i ely; con ol s. e ol ed s ain, P = 0.05, N = 12 o all g oups). Found a : doi:10.1371/jou nal.pone.0006761.s001 (1.16 MB TIF) E olu ion o Vi ulence PLoS ONE | www.plosone.o g 5 Augus 2009 | Volume 4 | Issue 8 | e6761 Acknowledgmen s We hank E. Ihalainen, H. Rowland, B. Hol , and membe s o Jou nal Club o Ecology and En i onmen al sec ion o he Uni e si y o Jy a¨skyla¨ o hei help ul commen s on he manusc ip , T. Pehkonen and H. Hu unen o assis ance in he labo a o y, and K. Viipale o concep ual help. Au ho Con ibu ions Concei ed and designed he expe imen s: VPF CL TH JL JM. Pe o med he expe imen s: VPF CL TH. Analyzed he da a: VPF. Con ibu ed eagen s/ma e ials/analysis ools: JL JM. W o e he pape : VPF CL TH JL JM. Re e ences 1. Pallen MJ, W en BW (2007) Bac e ial pa hogenomics. Na u e 449: 835–842. 2. Hol RD, Roy M (2007) P eda ion can inc ease he p e alence o in ec ious disease. Am Na 169: 690–699. 3. Schulenbu g H, Ku z J, Mo e Y, Si a-Jo hy MT (2008) In oduc ion. Ecological immunology. P oc R Soc Lond B 364: 3–14. 4. Bo e ET, Mi chell CE, Powe AG, Seabloom EW (2009) Consume s indi ec ly inc ease in ec ion isk in g assland ood webs. P oceedings o he Na ional Academy o Sciences o he Uni ed S a es o Ame ica 106: 503–506. 5. Rigby MC, Jokela J (2000) P eda o a oidance and immune de ence: cos s and ade-o s in snails. P oc R Soc Lond B 267: 171–176. 6. Ma z C, Kjellebe g S (2005) O he hook – how bac e ia su i e p o ozoan g azing. T ends in mic obiology 13: 302–307. 7. Edeline E, A i TB, Volles ad LA, Win ield IJ, Fle che JM, e al. (2008) An agonis ic selec ion om p eda o s and pa hogens al e s ood-web s uc u e. P oceedings o he Na ional Academy o Sciences o he Uni ed S a es o Ame ica 105: 19792–19796. 8. B u¨ssow H (2007) Bac e ia be ween p o es s and phages: om an ip eda ion s a egies o he e olu ion o pa hogenici y. Molecula Mic obiol 65: 583–589. 9. Ma z C, McDougald D, Mo eno AM, Yung PY, Yildiz FH (2005) Bio ilm o ma ion and pheno ypic a ia ion enhance p eda ion-d i en pe sis ence o Vib io chole ae. P oceedings o he Na ional Academy o Sciences o he Uni ed S a es o Ame ica 102: 16819–16824. 10. B own MRW, Ba ke J (1999) Unexplo ed ese oi s o pa hogenic bac e ia: p o ozoa and bio ilms. T ends in mic obiology 7: 46–50. 11. B own RC, Bass H, Coombs JP (1975) Ca bohyd a e binding p o eins in ol ed in phagocy osis by Acan hamoeba. Na u e 254: 434–435. 12. Da ies B, Cha ings S, Edwa ds SW (1991) Supe oxide gene a ion du ing phagocy osis by Acan hamoeba cas ellanii: simila i ies o he espi a o y bu s o immune phagocy es. J Gen Mic obiol 137: 705–710. 13. Ha b OS, Gao L-Y, Kwaik YA (2000) F om p o ozoa o mammalian cells: a new pa adigm in he li e cycle o in acellula bac e ial pa hogens. En i on Mic obiol 2: 251–265. 14. Josenhans C, Sue baum S (2002) The ole o mo ili y as a i ulence ac o in bac e ia. In J Med Mic obiol 291: 605–614. 15. F iman V-P, Hil unen T, Laakso J, Kai ala V (2008) P ey esou ce a ailabili y d i es e olu ion o p eda o p ey in e ac ion. P oc R Soc Lond B 275: 1625–1633. 16. Wei J-R, Lai H-C (2006) N-Acylhomose ine lac one-dependen cell- o-cell communica ion and social beha io in he genus Se a ia. In J Med Mic obiol 296: 117–124. 17. G ill CP, Moo e AJ (1998) E ec s o a la al an ip eda o esponse and la al die on adul pheno ype in an aposema ic ladybi d bee le. Oecologia 114: 274–282. 18. Mo e Y, Schmid-Hempel P (2000) Su i al o immuni y: The p ice o immune sys em ac i a ion o bumblebee wo ke s. Science 290: 1166–1168. 19. Co e SC, K uuk LEB, Wilson K (2003) Cos s o esis ance: gene ic co ela ions and po en ial ade-o s in an insec immune sys em. J E ol Biol 17: 421–429. 20. Fo dyce JA, Nice CC (2008) An agonis ic, s age-speci ic selec ion on de ensi e chemical seques a ion in a oxic bu e ly. E olu ion 62: 1610–1617. 21. Losey JE, Ha mon J, Ballan yne F, B own C (1997) A polymo phism main ained by opposi e pa e ns o pa asi ism and p eda ion. Na u e 388: 269–272. 22. Schmid-Hempel P (2003) Va ia ion in immune de ence as a ques ion o e olu iona y ecology. P oc R Soc Lond B 270: 357–366. 23. Linds ed C, Linds o¨m L, Mappes J (2009) The mo egula ion cons ains wa ning signal exp ession. E olu ion 63: 469–478. 24. Ojala K, Linds o¨m L, Mappes J (2007) Li e-his o y cons ain s and wa ning signal exp ession in an a c iid mo h. Func ional Ecology 21: 1162–1167. 25. Linds ed C, Linds o¨m L, Mappes J (2008) Hai iness and wa ning colou s as componen s o an ip eda o de ence: addi i e o in e ac i e bene i s? Anim Beha 75: 1703–1713. 26. G imon PAD, G imon F (1978) The genus Se a ia. Annu Re Mic obiol 32: 221–248. 27. Co e SC, Hails RS, Co y JS, Wilson K (2004) Densi y-dependen p ophylaxis and condi ion-dependen immune unc ion in Lepidop e an la ae: a mul i a - ia e app oach. J Anim Ecol 73: 283–293. 28. A mi age S, Si a-Jo hy M (2005) Immune unc ion esponds o selec ion o cu icula colou in Teneb io moli o . He edi y 94: 650–656. 29. Malik-Kale P, Raphael BH, Pa ke CT, Joens LA, Klena JD, e al. (2007) Cha ac e iza ion o gene ically ma ched isola es o Campylobac e jejuni e eals ha mu a ions in genes in ol ed in lagella biosyn hesis al e he o ganism’s i ulence po en ial. Appl En i on Mic obiol 73: 3123–3136. 30. Pujol N, Link EM, Liu LX, Ku z CL, Alloing G, e al. (2001) A e e se gene ic analysis o componen s o he Toll signaling pa hway in Caeno habdi is elegans. Cu en Biology 11: 809–82. 31. Pa uchu i DK, Ha shey RM (1987) Flagella a ia ion in Se a ia ma cescens is associa ed wi h colo a ia ion. J Bac e iology 169: 61–65. 32. Ja el KF, McB ide MJ (2008) The su p isingly di e se ways ha p oka yo es mo e. Na u e Re Mic obiol 6: 466–476. 33. Monge BC, Land y MR (1992) Size-selec i e g azing by he e o ophic nano lagella es: an analysis using li es ained bac e ia and dual-beam low cy ome y. A ch Hyd obiol Beih E geb Limnol 37: 173–185. 34. Gonza´lez JM, She EB, She BF (1993) Di e en ial eeding by ma ine lagella es on g owing e sus s a ing, and on mo ile e sus nonmo ile, bac e ial p ey. Ma Ecol P og Se 102: 257–267. 35. de Roode JC, Pansini R, Cheesman SJ, Helinski MEH, Hujben S, e al. (2005) Vi ulence and compe i i e abili y in gene ically di e se mala ia in ec ions. P oceedings o he Na ional Academy o Sciences o he Uni ed S a es o Ame ica 102: 7624–7628. 36. Buckling A, B ockhu s MA (2008) Kin selec ion and he e olu ion o i ulence. He edi y 100: 484–488. 37. F ank SA (1996) Models o pa asi e i ulence. Q Re Biol 71: 37–78. 38. Wes SA, Buckling A (2003) Coope a ion, i ulence and side opho e p oduc ion in bac e ial pa asi es. P oc R Soc Lond Se B 270: 37–44. 39. Ku z CL, Chau e S, And e`s E, Au ouze M, Valle I, e al. (2003) Vi ulence ac o s o he human oppo unis ic pa hogen Se a ia ma cescens iden i ied by in i o sc eening. EMBO J 22: 1451–1460. 40. Ojala K, Julkunen-Tii o R, Linds o¨m L, Mappes J (2005) Die a ec s he immune de ence and li e-his o y ai s o an A c iid mo h Pa asemia plan aginis. E ol Ecol Res 7: 1153–1170. 41. Singe MS, Mace KC, Be nays EA (2009) Sel -medica ion as adap i e plas ici y: inc eased inges ion o plan oxins by pa asi ized ca e pilla s. Plos one 4: e4796. 42. Kocks C, Cho JH, Nehme N, Ul ila J, Pea son AM, e al. (2005) Ea e , a ansmemb ane p o ein media ing phagocy osis o bac e ial pa hogens in D osophila. Cell 123: 335–346. 43. Ran ala MJ, Ro DA (2005) An analysis o ade-o s in immune unc ion, body size and de elopmen ime in he Medi e anean Field C icke , G yllus bimacula us. Func ional Ecology 19: 323–330. 44. Rux on GD, She a TN, Speed MP (2004) A oiding A ack: E olu iona y Ecology o C ypsis, Wa ning Signals & Mimic y. New Yo k: Ox o d Uni e si y P ess. 260 p. 45. Vodo a N, Acos a C, Lemai e B, Bocca d F (2004) D osophila: a poly alen model o deciphe hos -pa hogen in e ac ions. T ends in mic obiology 12: 235–242. 46. Nehme NT, Lie´geois S, Kele B, Giamma ina o P, P adel E, e al. (2007) A model o bac e ial in es inal in ec ions in D osophila melanogas e . PLoS Pa hog 3: e173. 47. Valle -Gely I, Lemai e B, Bocca d F (2008) Bac e ial s a egies o o e come insec de ences. Na u e Re Mic obiol 6: 302–313. E olu ion o Vi ulence PLoS ONE | www.plosone.o g 6 Augus 2009 | Volume 4 | Issue 8 | e6761