Stable coexistence of incompatible Wolbachia along a narrow contact zone in mosquito field populations Célestine Atyame Nten, Célestine Atyame, Pierrick Labbé, François Rousset, Marwa Beji, Patrick Makoundou, Olivier Duron, Emilie Dumas, Nicole Pasteur, Ali Bouattour, et al.

To cite this version:

Célestine Atyame Nten, Célestine Atyame, Pierrick Labbé, François Rousset, Marwa Beji, et al.. Stable coexistence of incompatible Wolbachia along a narrow contact zone in mosquito field populations. Molecular Ecology, Wiley, 2015, 24 (2), pp.508–521. ￿10.1111/mec.13035￿. ￿hal-01285420￿

HAL Id: hal-01285420 https://hal.univ-reunion.fr/hal-01285420 Submitted on 26 Oct 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Accepted Article Fort Makoundou This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This 10.1111/mec.13035 doi: to differenceslead between this version and Version the of Record. Please cite this article as may which process, andproofreading typesetting, pagination copyediting, through the been not has but peer review full undergone andpublication accepted hasbeen for article This Stable coexistence incompatibleof contactmosquito field zonein populations 4,5 and Mylène Weill and Article type :Original Article Date :03-Dec-2014 Accepted 02-Dec-2014 : Date Revised 28-Jul-2014 : Date Received 5- University Montpellier 2, Montpellier, France France Montpellier, 34293 5237, - UMR CRBM CNRS, 4- 3- Institut Pasteur Tunis, Laboratory of epidemiologyand veterinary microbiology/University 1- CNRS, ISEM - UMR 5554, Montpellier, France Célestine M. Atyame

France France †- Current address: MIVEGEC - UMR 5290-224, CNRS-IRD-UM1-UM2, Montpellier, 2- UniversityLaRéunion/CRVOI, Ste Clotilde, Réunion Island,France Tunis El Manar,Tunisia 1,5 , Olivier Duron , Olivier 1,5

1,2,5* 1,5† , Emilie Dumas , Pierrick Labbé Pierrick , 1,5 1,5 , Nicole Pasteur , François Rousset Wolbachia 1,5 alonga narrow , AliBouattour , 1,5 , Marwa Beji , Marwa 3 , Philippe , Philippe 3 , Patrick Accepted Article

types: lines with isofemale Crossing experiments from of bothinfections,structure withasharp contact zone areas.separating their distribution and hosts. infected incompatible described, especially amongst examples are only few there populations in natural whereas theoretically, investigated deeply have been inbacteria vector-borne disease c Wolbachia spatialthe of spread promotes death and (CI) betweensperm conditional and embryonic egg,causes which incompatibility Abstract This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This title Running sterilize them. The The them. sterilize w Pip31-infected males were compatible with compatible were males Pip31-infected

w Pip31) infecting the mosquito mosquito the Pip31) infecting w Here we surveyed have molecularHere of distribution the two In Arthropods, the intracellular bacteria In Arthropods, intracellular bacteria the zone. zone. Keywords: reunion.fr CélestineM. Atya* Corresponding author: Pip11-infected males always sterilize sterilize always males Pip11-infected to spread in natural populations through CI has attracted attention for using these these usingfor attention attracted CI has through populations inspread natural to : Incompatible Culex pipiens Culex w Pip11 strain was Pip11thus expected strain dynamicsto spread but overtemporal Wolbachia Wolbachia mosquito, mosquito, Wolbachia ontrol. The dynamics of incompatible incompatible dynamicsof The ontrol. Culex pipiens Culex stably coexist infections into host populations. The The host into of populations. ability infections Wolbachia w Pip31-infected females; however, while most while however, females; Pip31-infected w Pip11-infected females, a few completely few a females, Pip11-infected different localities showed three crossing in Tunisia. We delineated a spatial Wedelineated inTunisia. clear me; E-mail: celestine.atyame-nten@univ- me; E-mail: Wolbachia , Incompatibility, Cytoplasmic Contact often induce cytoplasmic Wolbachia Wolbachia strains ( strains infections w Pip11 Pip11 Accepted Article arthropod species (Werren arthropod play major roles. roles. major play hostthe in dynamics metapopulation and probability low dispersal contrast, of stability the mating on assortative or adaptation local of a samples field experiments, cage Population empirically. and both theoretically theobserved stability, may to contribute factors yearsseven of monitoring stabilityshows the the of contact zone. which We examined This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Introduction sustainable and design effective (Hiroki (Hiroki striatellus Laodelphax of populations in field spreads were reported such Casesof infections. spread of the allows which populations polymorphic in advantage reproductive have a infectedfemales Thus compatible. uninfected females, the reciprocal matewith infected males whereas cross remains mortality when embryonic triggers which (CI), incompatibility cytoplasmic is manipulation system reproductive to itsown advantage (Werren of totheability attributed mainly is success evolutionary & Hoy Hilgenboecker 2000; Jeyaprakash (Werren species infecting ininsects, of so 20-70% far described endosymbiont This study highlights the need of understanding CI dynamics in natural populations to populations natural in dynamics CI understanding of need the studyhighlights This Wolbachia

et al . 2005). 2005). . Drosophila simulans

are maternally are intracellular inherited

(Hoshizaki Shimada 1995) butterfly & and et al Wolbachia . 2008). 2008). . in California (Turelli & (Turelli planthopper Hoffmann in California 1991), -based control strategies. strategies. control -based et al et nd modeling did not support significant impacts notsupportsignificantnd did modeling Wolbachia . 2008; Zug & Hammerstein 2012). This This 2012). Zug &Hammerstein 2008; . are the most frequent bacterial mostfrequent arethe et al α Wolbachia -proteobacteria -proteobacteria that infect many . 2008). The most common most The . 2008). w Pip infection structure. By By structure. infection Pip Eurema hecabe Eurema to manipulate the host the tomanipulate Cx. pipiens Cx. et al in Japan Japan in . 1995; . likely likely Accepted Article panmictic host population, since only compatible since population, host panmictic 1995). Hoffmann Hoffmann 1978; (Fine fitness host onits fecundity- i.e. female of effect the (iii) and crosses; incompatible diein that offspring of proportion the i.e. CI mortality (or CI level)- the by(ii) aninfected produced female; 2009 i.e. review):Telschow rate- offspring (i)thetransmission thefor of infected proportion influence the of spread and three largely theoretically been explored (Rousset transmission of human of pathogens human transmission such as by mosquitoes of some infections that observation the after by mosquitoes, transmitted This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This & Charlat 2004; Mercot Duron 2002; Stauffer & (Riegler patterns bidirectional or unidirectional either canfollow and occurs strains. males When andfemalesinfected are Wolbachia al et dynamic (Hoffmann models by infection aspredicted throughCI, populations in field spread al Walker 2009; . 2013). Furthermore, Furthermore, 2013). . . 2011). Better knowledge of CI dynamics is thus required to explore the full potential of potential full the explore to thus required is dynamics CI of knowledge Better 2011). . Incompatible Wolbachia Wolbachia et al in vector control strategies. strategies. control vector in . 1991; Engelstadter & Telschow 2009). The pattern of CI will determine the of CI determine will pattern 2009). &Telschow The . 1991; Engelstadter et al Wolbachia a has attracted much attention as controldiseases toolto promising dynamics is made more complex bythe presence ofseveral incompatible . 2011; Blagrove. Wolbachia Wolbachia- Wolbachia reduce the development of time development reduce the hosttheir andreduce the et al strains cannot stably coexist in an unstructured and in an stably cannot strains coexist infected infected et al in apanmictic hostpopula in . 2006a; Atyame . 2012) or . 2012)or the dengue viruses (Moreira chikungunya and Aedes aegypti key models parameters describe how CI could could describe CI how parameters models key with different and incompatible strains, CI strains, incompatible and different with can resist the competition strains CI-driven Plasmodium et al et released in Australia have readily readily have Australia in released . 2014). (Kambris tion (see Engelstadter & & Engelstadter tion (see Wolbachia Wolbachia et al et Aedes et al et . 1990; Turelli & . Turelli 1990; . 2010; Bian Bian 2010; . dynamics has or infection infection on Anopheles et al et et .

Accepted Article and equilibrium states are modified (Barton & Turelli 2011; Hancock & Godfray 2012). Godfray 2011;Hancock & & Turelli (Barton aremodified states equilibrium and tostructured host populations directly applied inthe selection for global of distribution of the B supergroup (Rasgon & Scott 2003; Duron 2003;Duron Scott the B supergroup (Rasgon of & naturally are mosquitoes infectedby (Fonseca specificities (thecoexist Incompatible respectively - present extremean situation. In temperate regions, two (Telschow threshold a critical below is dispersal if populations both that postulating coast of Australia (Kriesner which flies co-infected with two with co-infected flies which fly fruit cherry the populations: (i) host infected in naturally competing Only report on Telschow studies 2009). Engelstadter & two 1959, Watson & (Caspari threshold frequency initial once abovean the population, to invade expected thealways outcompetes rarest, CI,while with the strain isunidirectional CI-inducing most frequent CI,the outcome: of bidirectional case in the competition This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This quinquefasciatus the which Wolbachia However, invasion thresholds deduced deduced for thresholds However, invasion Culex pipiens Wolbachia infection in southern Europe (Riegler & Stauffer 2002), and(ii) 2002), Stauffer & (Riegler infection Europe insouthern pipiens Wolbachia and and and strain strain w complex mosquitoes- whose main subspecies are subspecieswhose main complex mosquitoes- Au and et al Cx strainscan indeed stably coexiststructured in parapatric host molestus molestus . w . 2004). 2004). . p Ri Ri invaded populations strain withthe infected et al et . w Ri increased females fecundity, highlighting a potential role potential highlighting a females fecundity, Ri increased pipiens Wolbachia . 2013). In this latter case, data were best explained bywereIn thisdatacase, bestexplained . latter 2013). forms), showing and physiological genetic behavioral, forms), Wolbachia Wolbachia Culex pipiens quinquefasciatus , ubiquitous in tropical ubiquitous temperate , intropical regions, and Wolbachia strains invaded populations carrying only one carrying only one populations invaded strains connected byas dispersal, system dynamics strains ( strains et al . isolated panmictic be cannot populations . 2005; Duron Duron 2005; . w Pip) that belong to a unique clade clade unique to a belong Pip) that et al et . 2005; Flor et al Cx Rhagoletis cerasi Rhagoletis and . 2006b). However, However, . 2006b). . Wolbachia Wolbachia Cx. p.pipiens p. pipiens p. w D. simulans, Au on eastthe et al et . 2007). . 2007). Cx strains strains forms strain strain . , in in p.

Accepted Article markers and laboratory crosses showed CI between them (Duron them CIbetween crossesand laboratory showed markers (Duron populations by populations infected compatible geographically within is genetic close stably diversity maintained considerable 2006b; Atyame w This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This the Inthese conditions, them. sterilized w properties. twomolecular same the only detected still have field. We the in observations direct analyzed expression from and CI localities, in different lines sampled isofemale from CIpatterns their characterized Tunisia, of these thedistribution areas mapped havenow We populations. was thus a opportunity to verystudy evolutionaryin natural dynamics attractive their strains, molecular two infected byonly found were Duron of these incompatible long-term coexistence might the that explain parameters we have identified critical situation, aspatially developing By stable. remained has the contactbetween andhave that found zone populations, field theof seven-year over follow-up a time. We havestrains performed thus Pip31-infected males were compatible with compatible were males Pip31-infected (Duron smallhighscale evolutionarydisplay at polymorphism strains a genetic a Pip et al We herereport on a field situation inNorth Africa where

et al . 2006a; Atyame Atyame . 2006a; w . 2006b). These two strains were defined using a set of polymorphic molecular molecular polymorphic set of using a weredefined These two strains . 2006b). Pip11-infected males sterilize always Pip11-infected et al w . Dumas 2011a; Pip infections. w Pip strains. However, we have found three strains in term of CI CI of term in strains three found have we However, strains. Pip et al . 2011b;Atyame w Pip strains (Duron strains Pip et al w Pip11 strain is expected to outcompete the outcompete the isexpected to strain Pip11 . 2013). CIpatterns andcomplex 1967a; (Laven w explicit model that better describes the field field the describes better that model explicit Pip11-infected females, a few completely few a females, Pip11-infected et al w Pip31 andPip31 in Algeria w Pip31-infected most females. While Pip31-infected . A2014). survey recent showed that et al . 2011). 2011). . et al w w Pip11 and Pip11 Pip strains inAlgeria and Cx. pipiens Cx. . 2006a). This situation situation This . 2006a). w w Pip infections in Pip infections Pip11 in Tunisia Tunisia Pip11 in w Pip31 strains strains Pip31 populations populations Cx. pipiens w Pip31 Pip31 et al et .

Accepted Article Molecular typing honey solution. with fed adults were pellets while rabbit powderand of shrimp mixture witha were Larvae fed dark cycle. a12-hlight/12-h under 2005, 2008, 2009, 2010, October 2010 and June 2011). All localities were not sampled each each sampled were not localities 2011). All June 2010and October 2010, 2009, 2008, 2005, 2006and June (in 1997, 2008)and in 63 of localities (inJune Tunisia 1996, 2003, 1997, Mosquito collection and lines maintenance Materials Methods and This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This in65 (n=245)werereared lines dm Wolbachia (containing 100-300 its of thegenotyping hatching, for and wasindividually isolated eggs) (Table lines eggisofemale Each raft establish toS1). 19 tomate were allowed from localities nitrogen 100 until liquid werestored -20°C specimens or in at (Table year sample was rearedS1). Each inthe to adulthood laboratory and approximately several previouslyseveral described todifferentiate the (Figure S1). S1). (Figure 374 displays bp and fragments: amplified PCR twogenes differentiateThese the (Duron

Culex pipiens Wolbachia

et al infection was performed by analyzing two first-instar larvae (L1). All isofemale by (L1). was larvae infection analyzing Allisofemale twofirst-instar performed . 2007) and the DNA mismatch repair gene protein DNA mismatch and the . 2007) genotyping was performed using two two using performed was genotyping

mosquitoes were collected as larvae or pupae in ten localities of Algeria Algeria of localities ten in or pupae as larvae collected were mosquitoes 437 sizesbp fragment the for strains ank2 ank2 w displays 313 bp and 511 bp sizes, fragment and Pip strains investigated here on the basis of the size of the of size the of basis the on here investigated Pip strains 3 screened cages kept in 22 in a single room at cages screened kept to 25°C, w

Pip strains): the ankyrin domains gene gene the domains ankyrin Pip strains): w Pip polymorphic markers (among (among markers Pip polymorphic w Pip11 and Pip11 and Wolbachia MutL (Atyame w genotyping. Adults Adults genotyping. Pip31 respectively Pip31 respectively et al et . 2011a). 2011a). . MutL ank 2 Accepted Article molestus molestus two #27) and Riadh, two from samples in two forms the distinguish to used was gene ( gene This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This 50 with performed were in DNA solution a 40 ng genomic of (CTAB) Bromide CetylTrimethylAmmonium 2012). identification of molecular for was effective previously found loci microsatellite Boussalem1, #42).Boussalem1, #34; Kef1, #31; Tabarka, #9; Kala, #8; Douas, #6; Guelma, #3; (Harash, localities seven from #25; Zerga, #30;Sokra, (Sousse, #37localities and AïnTounga, #51) and seven 3130 sequencer 3130 sequencer using the Term BigDye ABIPrism an with directly sequenced then CA)and Valencia, (QIAGEN, kit extraction PCRpr For sequencing, electrophoresis. (1.5%) listedinTable are (primers S2). Amplified 94°C conditions: 30s, for 52°C 30s, and for 72°C 1 for to 1.5 min for total of33 a cycles form molestus cytb DNA was extracted from adult mosquitoes and larvae using a larvae using and mosquitoes a adult from DNA was extracted b cytochrome of the 852bp variability bysequencing mitochondrial analyzed also We Finally, we forchecked betweena relationship forms, wellas forms, as hybrids, in ) in ) 12 in isofemale Algeria five from andTunisia: lines ) and ) and w Pip infection w Pip31 localities (Tabarka, #31 and Boussalem2, #71). The Boussalem2,#31 The #71). (Tabarka,and localities Pip31 . The The inator Kit Biosystems). inator (Applied Cx. pipiens Cx. CQ11 DNA fragments were separated by wereseparated gelDNA agarose fragments microsatellite Fonseca locus (Bahnck& 2006) protocol (Rogersprotocol &Bendich PCR 1988). All oducts were purified with the QIAquick gel gel with QIAquick the purified were oducts mosquitoes in Morocco (Amraoui Cx. pipiens Cx. w Pip11 localities (Ayed, #24 and #24 and (Ayed, localities Pip11 μ l final under the the under following l final forms (form (form forms w Pip11 lines from four four lines from Pip11

w pipiens pipiens Pip31 lines CQ11 et al and and .

Accepted Article w between performed of crosseswasset A first lines. isofemale derived from males laboratory In the Cytoplasmic incompatibility This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This cages (65 dm (2006). as development of through by observation described embryonic fertilization Duron andWeill were for unhatched All checked egg rafts microscope. binocular (HR)under a rate hatching with an equivalent number of males a from and females males of equivalent number with an Sok × Kef1-2, both showing unidirectional CI (Sok males sterilizing kef1-1 and kef1-2 and kef1-1 sterilizing CI(Sok males showingunidirectional both Kef1-2, Sok × a as identified Ain Tounga (#51). To assess CIphenotypes among assess To (#51). Ain Tounga w between performed crosses were Next, Algeria). from Ahras, Souk #10 #9; Kala, Lac, #8; #7; Douas, #6; Guelma, Tunisia and #42from #34;Boussalem1, Kef1, #31; Tabarka, (eight localities: lines five cytoplasmic incompatibility (CI) of incompatibility cross status each wasdetermined egg cytoplasmic byexamining The hatching. until individually stored days and five after collected were egg rafts their and were 2-5 caging, were fivedaysallowed toblood feed after individuals Females days old. #51) werecrossed with derived Pip31 lines isolated from localities where onlyone where fromlocalities isolated lines Pip31 Pip11 and two and Pip11 w Pip11 lines (three localities: Tunis, #16, Sousse, #25 and Sokra, #37) and 16 and #37) andSokra, #25 Sousse, #16, Tunis, localities: (three Pip11 lines Mating preferences of 25-50 with out Crosses were carried of number andan virginfemales equivalent w Pip11 males from three localities (El Battan, #43; El Manar, #46 and Aïn Tounga, Tounga, Aïn and Manar, #46 El #43; Battan, (El localities three from Pip11 males 3 ), where 100 males 100 and females from a w Pip31_U Twotypeswere up:Sok of and ×Kef1-1 set line). confrontations w Pip31 isofemale isofemale lines samples fromtheTunisian Zergaisolated Pip31 (#30) and w Pip11 and Pip11 w Pip31 females derived from the Har isofemale line (Harash, #3). All#3). (Harash, isofemale line Har the from derived females Pip31 w Pip11 and Pip11 and

w Pip31 lines from the same locality, i.e. between two two between i.e. samelocality, the from lines Pip31 w Pip31 measured were mosquitoes in laboratory Wolbachia w w Pip11-infected males, isofemale- males, Pip11-infected Pip11 isofemale line were placed were line Pip11 isofemale w molecular strainwas present: Pip31 isofemale line (later (later line isofemale Pip31 w Pip11 and Pip11 w Pip31 Pip31 Accepted Article dm with an initial frequency of 50% (i.e. 100 males and 100 females Tn were mixed with 100 with100 were mixed Tn females and 100 males 100 50% (i.e. aninitial of frequency with init then were genes. Population cages nuclear or line the Harline crossed with 50-100Slab b eight through Slab) line laboratory the (from w effectssterilizing females).side Har Toavoid w described above. described to allowed oviposit on cup. a water Egg were rafts and stored individually were analyzed as the cage into at the introduced weresame time. 5 Females blood-fed caging after days and were up.cages set All were 1-day- individuals females). Three replicates were performed for each This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Population cages microscope and egg rafts hadthat already hatched inthe field were easy to qualify. at HR 25°C, ovoposition >72 evaluated was under abinocular collection after hours withinhatch 36-48 Aslarvae hatching. laboratory toobserve the brought to after hours surface the water, from paintbrush ofstagnant and plates, separately in24-well placed witha carefully removed were egg rafts The and2010. in 2009 Tunisian localities in rafts In field populations Pip strains) of the Tn and Har lines were introduced into the same nuclear background background nuclear same the into introduced were lines Har and Tn the of strains) Pip (a Har isofemale line Pip31 3 ) in laboratory through confrontations between the the between through confrontations laboratory ) in The invasion dynamics of the the of dynamics invasion The The of expression CI in Cx. pipiens w Pip31_U line), showing unidirectional CI (Tn malesCI (Tn showingunidirectional Pip31_U line), w Pip11 was strain inpopulation cages (65 examined natural populations was assessednatural populations was by sampling egg of nuclear genomes, the cytoplasms (including cytoplasms the genomes, of nuclear iated using the Tn and Harbackcrossed lines, Tnand iated using the ackcrosses (100-200 virgin Tn ofackcrosses (100-200 the females old and virgin,old and were andallthemosquitoes males), expected to restore ~ 96% of Slab of ~ 96% to expected restore males), type of type of confrontation, so atotal six w Pip11 Tnandthe line isofemale Pip11 Accepted Article Core Team, 2013). Team, Core populations. between polymorphism genetic of distribution the based on differentiation, population to estimate examined with GENEPOP (Raymond & Rousset 1995; Rousset 2008). Wright’s F-statistics ( Rousset 1995;Rousset F-statistics Wright’s & 2008). GENEPOP (Raymond with analyses Data inpopulationfrequencies were bycages monitored PCRassays as described above. newly emerged adults resulting from thepreviousnewly from adults emerged resulting generation. using generation each at cages discrete new byestablishing generations employed cages virgin. the into population the were cage old and at All same Mosquitoes introduced time. were replicate Har). Three 100 females and males This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This w molecular the two of byone onlyinfected samples were examined (Table several sampling1996 from to2009 in campaigns Tunisia all S1). Mosquitoes from (Atyame Tunisia between zone contact narrow very A Results Results Pip11 and and Pip11 All other statistical analyses and modeling were performed using the R usingsoftware (R the performed modeling were statistical analyses and All other Field data on the geographic distribution of distribution geographic the on data Field We developed a sensitive molecular assay to easily genotype the five five the genotype easily to assay molecular asensitive developed We

et al w . 2011a) andscreened Pip31, andbelonging the Pip31, to

Cx. pipiens Cx. w w Pip groups IV, respectively Iand (Duron Pip11 and and Pip11 collected as larvae or pupae in Algeria andAlgeria in pupae or as larvae collected performed and all individuals were 1-day- were individuals all and performed w Pip strains in Tunisia were analyzed w Pip31 w Pip strains, referred to as to referred strains, Pip Wolbachia Wolbachia w strains in strains infections Pip groups Pip groups F st ) was ) was et al et . Accepted Article frequencies (Mantel test on Spearman rank correlation: b = 0.01, b =0.01, correlation: rank test (Mantel onSpearman frequencies and both site could befound)in evidenced, (Clusters breeding localities I and nomosquito which in byadry area Figure II, separated 1, the of analysis w Atyame 2006b; This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This same the crosses between mosquitoes with observedinfected CIin never We respectively. 0%), (HR= or mortality >90%) (HR viability with embryonic associated full was incompatibility were usedas replicateslocality whenever 16 used five areas. We different from Most #44). Œufs, #40; Slouguia, (e.g. #51)to91-97% #50;AïnTounga, Fontaine, (e.g. Zerga, 20-30% #30), percent (e.g. infected by 100% kilometers, onlye.g. afew 3.60 100% Othman (#28, between km between (Figure transition 1). The strain either infected of were byone samples most Riveralong by which Medjerda delineated the b =0.01, Pip31 in the north and west, the latter also spread over in Algeria (Figure Algeria overin spread also west, and thelatter north the in Pip31 Structure 1). w Pip31 lines from eight localities (Table (Table same eight a localities from from isolated Pip31 lines lines S3). Twoisofemale w w w We We first investigated the CI patterns between Pip not (data shown).strain Pip31 display unidirectional CI unidirectional Pip31 display Pip31). Within the contact zone, 12 12 zone, contact the Within Pip31). ofmosquitoes mixture a localities exhibited P = 0.004 for Cluster-II). The contact zone was in the center of northern Tunisia, Tunisia, northern of center inthe was zone contact The Cluster-II). for = 0.004 w Pip11 or by orby Pip11 w et al Pip strains frequency was carried out in two independent sampledtwo independent groupsof in out was frequency strains carried Pip . 2011a). In Tunisia, Tunisia, In 2011a). . w Pip31, with a frequency a of with Pip31, w w Pip11 isofemale three from lines localities Pip11 isofemale and Pip11 and and Pip11

with w Pip11 is distributed in the east and south and south the eastand in distributed is Pip11 groups, spatial autocorrelations incytotypespatialgroups, autocorrelations possible. In all crosses, compatibility or or compatibility Inallcrosses, possible. w Pip11; a few show bidirectional CI bidirectional show a few Pip11; w w Pip11 and Pip31 distribution was sharp, spanning spanning was sharp, distribution Pip31 w Pip11 infections ranging from a few a few rangingfrom infections Pip11 w P Pip31 infected Pip31 infected mosquitoes = 0.001 for Cluster-I, and Cluster-I, =0.001 for w Pip11) andBriss(#38, Pip11)

Accepted Article sympatric (Zerga, #30; Aïn Tounga, #51; Table crossed Table all #51; that Tounga, Note Aïn S4). #30; (Zerga, sympatric wherebetween localities lines also observedwere from will be thereafter named the with or incompatible compatible respectively are where strains males area. The studied the presentare in considered, all with incompatible w pure distant geographically w frequenciesrespectivecrossing through extensive the two candiscriminate markers polymorphic compatibleamong each other, aswere the and S4).However, and with compatible were theHar derived from #3)(Table females isofemale line (Harash, S5). Most with Tounga, #51) Aïn and #46 Manar, El #43; (El Battan, other localities three by confirmed wasfurther crossing isofemale-derived (80 crosses); this widespread over widespread but Tunisia, frequency thetwo #51) withamean S6). of (Table 12% Thus, Table and 2 the cytotypes inthe five locations, whether near or and and cross). per 20 about eggS4, rafts This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Pip31/ #37)with (Sokra, Sokline isofemale Pip11 w Pip31_B cytotype varied from 3% (1/39; Khetmine, #62) AïnTounga, 20% 3%(1/39; #62) to (7/35; Khetmine, Pip31_B from varied cytotype Thus, two strains of two Thus, We analyzed the CI patterns between the patternsanalyzed the We between CI the w Pip11 (El Manar, #46; Aïn Tounga, #51) populations. We detected the two the detected We populations. #51) Tounga, #46; Aïn Manar, (El Pip11 w w w Pip11 females (five assayed Pip31 males from the Algerian lines Souk1 and Souk2 were Souk2 and Souk1 lines theAlgerian from males Pip31 Pip11 females (75 of87Pip11 females crosses from 20 w Pip31_U or w w Pip31 (Hamra, #29; Ras Rajel, #57; Khetmine, #62) and mixed #62) and Khetmine, #57; Rajel, Ras #29; (Hamra, Pip31 Pip31_U is most the frequent. w Pip31, molecularly indistinguishable for the markers markers the for indistinguishable molecularly Pip31, w w w Pip11 females (uni-directional or bi-directional CI) or bi-directional (uni-directional Pip11 females Pip11 males sterilized males Pip11 Pip31_B (Table 1). Such Pip31_B (Table uni- CIs and bi-directional far from the contact zone. The frequency of of frequency The zone. contact the from far w experiments. We crossed females from the from crossedWe females experiments. Pip31_U and w w w w Pip11 lines tested, S3B). tested, Table Pip11 lines Pip31 isofemale-derived males Pip31 isofemale-derivedfrom males Pip11 and Pip31 cytotypes, we deduced their deduced we Pip31 cytotypes, w Pip11 and

w w Pip31 females in Pip31 in females all crosses w Pip31_B lines. Since lines. noPip31_B w Pip31 strains (Tables (Tables strains Pip31 S3 Pip31 lines; Tables lines; Pip31 S3B w w Pip31 strains werePip31 strains w Pip11 males from from Pip11 males w Pip11 lines were lines Pip11 Pip31 are strains w Pip31 males males Pip31 w w Pip31 Pip31 Pip31 Pip31

Accepted Article population cages by confronting byconfronting cages population transmission and no differential fitness cost, fitness andnodifferential transmission for frequencies as initial 50/50 using initiated were population cages background. Three nuclear traitsmay interfere with w w strains stability zonethe the molecular anoverall of between thus indicate contact Our data and stably more (#50)(Table (#29) Hamra in inFontaine transiently localities, in two S1). #54). Gare, Beja #51; Tounga, Aïn #46; Manar, (El others inthreedecreased and #61) even #53;Utique, Gaafour, #41; #30; Dougga, (Zerga, majority Remarkably, time. where localities of initial frequency the where localities All (2005-2011). seven-yearfollow-up overthe set data thecomplete compiled and in 2009 delineated zone contact along the 2011 samplings 2010and additional in performed towards whereas rapidly andconsistently theserapidly with (Hoffmann four generations (98%)within fixation In the field, This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Pip11 and and Pip11 Pip11 outcompetes Pip11 outcompetes w Pip11 and and Pip11 This observed stability of the contact zone prompted us to examine which life history which ustoexamine life prompted zone contact the stabilityobserved of This The situation in which which in situation The w w Pip11 invasion if the host population is panmictic and unstructured. We therefore We andunstructured. panmictic is hostpopulation the if Pip11 invasion Pip31 males are mostly compatible with w w Pip31 over the seven-year study. seven-year study. over the Pip31 Pip11 does not outcompete outcompete not Pip11 does w Pip11 was majority, majority, was Pip11 w w Pip31_U. Assuming random mating, complete CI, 100% maternal CI,100% maternal complete mating, random Assuming Pip31_U. Pip11 Pip11 frequency didnot change in localities four where w w Pip31 in population cages cages population Pip31 in Pip11 invasion. We first tested the invasive of capacity the tested first We invasion. Pip11 w Pip11 or Pip11 w Pip11 males induce incompatibility with with induce incompatibility males Pip11 w expectations (Tableexpectations observed The 3). Pip11 and and Pip11 w Pip31 frequency was very low and did not vary over vary not did verylow and was Pip31 frequency w Pip31 was 100% remained stable (Table (Table was stable S1). Inall Pip31 100% remained w w Pip11 frequency was expected to reach near- to reach was expected Pip11 frequency w Pip31 as predicted Pip31_U infected infected linesPip31_U carrying same the et al w Pip11 is females, predicted to evolve . 1990). w Pip11 frequency increased only increased frequency Pip11 w Pip11 frequency Pip11 increased

w Pip11 frequencies w Pip31 females, females, Pip31 w Pip31 was w Pip11 in Pip11 in Accepted Article No specific association between association No specific field. penetrance, penetrance, w I variable P variable laboratory, laboratory, replicates, deviation significant theexpectedfrom frequenciestests (Exact binomial onpooled test, exact cages three the (Fisher between homogeneous were This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This the model (Venables & Ripley 2002): S Ripley 2002): (Venables & model by differential infection rates of the of rates infection bydifferential molestus with with by infected 20 mosquitoes Farajollahi 2004; the differences: physiological and spaces, undergoes winter diapause and requires a blood meal for oviposition (Fonseca (Fonseca for oviposition meal blood a requires and winter diapause spaces, undergoes whereasthe blood meal, without a oviposit can and during winter NF Pip31-infected mosquitoes (two-level variable I variable mosquitoes(two-level Pip31-infected .P pipiens w OP Pip31 (Tabarka, #31; Boussalem2, Pip31 (Tabarka, #71) (Table S7). We the identified two

A straightforward explanation might be that explanationmight bethat A straightforward : forms and their hybrids (three-level response variable S variable response (three-level hybrids and their forms χ 2 OP P and and = 0.54, w Wolbachia = 0.28 and 1 for generations 3 and 4, respectively). This result shows that, in thethat,shows in result This 4,3 respectively). generations and for =0.281 and ). No evidence for any preferential association was found (multinomial logit (multinomial wasfound association preferential any for evidence No). Pip11 invades as suggestingPip11 invades expected, that molestus et al et P = 0.76). theTherefore stability of the zone contact becannot explained . 2011; Harbach 2012). We 2012). genotyped the 2011; Harbach . transmission and fitness cost hamper the spreading of of spreading the hamper cost andfitness transmission forms of the w Pip11 (Ayed, #24 and Riadh, #27) and 20 #27) and 20 #24andRiadh, (Ayed, Pip11 infected mosquitoes molestus Cx. pipiens Cx. PEC w Cx. pipiens Cx. Pip strains and =I NF form mates in confined spaces, remains active remains spaces, in confined mates form +P forms. OP complex. These two forms show behavioral two forms showbehavioral These complex. +I NF NF w ), in the four populations (four-level (four-level populations four the in ), Pip11 and Pip11 and .P Cx. pipiens OP life history traits other than CI than other history traits life ; Likelihood Ratio Test (LRT) for for (LRT) Test Ratio Likelihood ; P w CQ11 pipiens = 0.245) and showed no showed = 0.245) and PEC Pip31 differentially infect infect Pip31 differentially forms ) in either ) ineither microsatellite loci of loci microsatellite form mates in open in mates form

w Pip11 in the pipiens w Pip11- or Pip11- et al et and . Accepted Article measuring the incidence of sterile egg rafts collected in locations where locations in collected egg sterile rafts of incidence the measuring the occu nextto address wished correction. We test, exact (Fisher expected than egg infertile significantlyrafts less six produced of cages, the which in one hypothesi mating Therandom crosses). incompatible of hallmarks development, embryonic 20-80% abortive and 0%) = (HR hatching infertile no mating,Assuming 25%of egg among (i.e. theoffspring we random rafts expected #37)with two Sokra, from one possibility confronting by addressed this Wefirst 1991). significantly different (generalized linear model with binomial I error binomial model with linear (generalized different significantly egg rafts (I rafts egg infertile frequency of the #69, Mjez, sympatric), andFont #46 Manar, El #39; Melah, Oued Table infertile, both where 5). In localities (<1% raftsalmost werefertile allegg mosquitoes (allopatric), infected Dougga,#41) and 1938 only with localities Incontrol egg rafts. (Table sympatric 137 590 are examined5). ofa to we Overall, total i.e. locality, per egg rafts for stable the coexistence of incompatible mating,random we examined occurrence the further of assortative mating, allow known to 6 infertileand allopatric egg(two-level localities sympatric in rafts variableS Assortative mating is insufficient to explain explain to insufficient is mating Assortative This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This absence of assortative mating, the expected the mating, of assortative absence ), demonstrating that incompatibility is expressed in natural populations. populations. natural in expressed is incompatibility that demonstrating ), Although the kinetics of kinetics Although the We then estimated the incidence of assortative mating by maximum likelihood: in the in likelihood: matingincidenceby maximum of assortative the thenestimated We NF ) 3%(7 was between out of 252) and 7% (18 out of incidence The 245). of P = 0.02). However, was this after asequentialnot significant Bonferroni’s w Pip31_U isofemale lines (Kef1-1 and Kef1-2 from Kef1, #34). from and Kef1-2 (Kef1-1 lines isofemale Pip31_U w Pip11 spreading in population cages are compatible with compatible are cages population in spreading Pip11 frequency of infertile crosses would be would crosses infertile of frequency w w Pip instrains populationsfield (Rousset w Pip11 and Pip11 and Pip11 (Jedaida, Pip11#59) (Jedaida, or rrence of assortative rrence ofassortative by matingfield, inthe w Pip dynamics Pip s notberejected could (Table 4), except w Pip31 were present (Zerga, #30; (Zerga,#30; present were Pip31 w Pip11 isofemale line (Sok (Sok line isofemale Pip11

NF =S w w Pip31 (Hamra, #29 Pip31 (Hamra, Pip11 and Pip11 and YM ; LRT: P YM I =p P w ) was ) was <10 Pip31 Pip31 et al 11 ( 1- . - Accepted Article zero, zero, the confidence was very interval with 0.63], an wide value[-0.08, estimated of observed mean frequency of correlation correlation by (i) likelihood maximum the egghere,large ofare even number when a rafts examined. Wenevertheless jointly estimated 0 closeto when strain are frequencies or thesympatric populations case for 1, the which is between mates.between frequencies of frequencies of reduction of this frequency, described as described this frequency, of reduction crosses between w p This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This toward pattern types oftwo that as analyzed be types can three the of dynamics frequency The Low dispersal rateand supersede not fa assortative additional Therefore mating. egg inthe infertile remains field or with onlycompatible rafts either with moderate random trials and that CI is expressed in the field where those strains are present. The incidenceof 0.39. 11 Pip31-infected mosquitoes (i.e. between crosses (i.e. males mosquitoes Pip31-infected )(1 +p These results show that 31B Only females transmit F, F, ), w where where ih its with w w Pip31 in Tunisia. Pip31 in Tunisia. F Pip11 infected when Pip11 males by Pip11 and infertile egg rafts. However, this gives little information on information gives little this However, rafts. egg andinfertile Pip11 w and Pip11 and p p confidence interval, for all populations, using populations, all for interval, confidence 11 11 is is can be using can jointly estimated w Pip31_B are critical for the stability of the contact zone w w Pip31_B, Table 2). As expected from strain frequencies near tonear strain frequencies from expected As 2). Table Pip31_B, w Pip11 frequency and Pip11 frequency w Pip31_B in both directions). Assortative mating entails Assortativea entails mating directions). both Pip31_B in w Pip11 and Pip11 frequencies in each population, and (ii) a single a (ii) and population, in each Pip11 frequencies Wolbachia ctors must be invoked to explain whyto invoked mustbe ctors P I (1-F) w Pip31 strains clearly mate at random in cage in at random mate clearly Pip31 strains and and w Pip31_B and , where , where w p Pip31 females display a similar display Pip31CI asimilar females 31B w Pip11 and females females and Pip11 binomial models for the observed the for models binomial is F is the correlation of cytotypes cytotypes of correlation the is w w Pip31_U the 1, (Table rows Pip31_B frequency among frequency Pip31_B p 31B 31B = = w Pip31_U, and Pip31_U, 0.12 (i.e. the (i.e. 0.12 w Pip11 does F F

of F

Accepted Article This suggests that dispersal might be a critical factorof invasion. absencefor acritical the dispersal might be that suggestsThis cytotypes, as generally observed in three the of maternal transmission and of full no infection) on female fecunditycost (i.e. w within of frequency the of in proportion direction, other the in low CI and direction types (e.g.Caspari &1959; Watson Hoffmann w polymorphism molecular the of evolution Thusthe change. doesnot frequency w for This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This where localities stable coexistence of the three be eliminated, and w expectedis to to spread fixation over a generations, both few eliminating of frequency initial the If frequency. relative frequency a on depending unstable equilibrium,to at lead an knownto well Pip31_B. If the initial frequency of of frequency initial the If Pip31_B. Pip31/ relative their that therefore, and, success reproductive same expected the have Pip31_B Pip31_B initial frequency is a key factor for for factor key a is frequency initial Pip31_B w Pip31_B and Pip31_B and w Analysis of w Pip31. Pip31. Pip11 can be investigated by models previously developed for two cytoplasmic cytoplasmic for two developed previously bymodels investigated be can Pip11 w Pip31 is present only if its frequency its the frequency exceeds of only if Pip31 is present w w Pip31_U and w Pip molecular types was done ofin assuming noeffect Pip31_U females are identical). This implies that This identical). are Pip31_U females w Pip cytotypes is made possible and Pip cytotypes made is w Cx Pip31_B would stably coexist. Hence, in this model, no this Hence, in model, stably coexist. Pip31_B would w . pipiens Pip11 Pip11 is lower thanthat of w Pip11 is higher than that of than that higher Pip11 is w (Rasgon & Scott 2003). These conditions are conditions These 2003). Scott (Rasgon & et al Pip11 invasion . 1990),. considering one complete CI in

w Pip31_B, w Pip11 could colonize colonize Pip11 could w Pip31_B, w w w w Pip infections Pip Pip31_U and Pip31_U Pip31_U andPip31_U Pip11 should Pip11 w w w Pip31_B. Pip31_B Pip31_B Pip31_B Pip31_B w Pip11 Pip11 Accepted Article dispersal probability probability dispersal ( frequency i.e. conditions, a relatively high fecundity cost ( Schematically, Schematically, represents amean square axial dispersal distance spatially explicit version (Telschow models of two-patch earlier and three cytotypes ( the frequencythespatial of cline of evolution modeling may prevent rate dispersal Low This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Discussion w wasinvestigated taking into valuesaccount different ofprobability which below of the value allows model estimating threshold probability the dispersal performed were iterations Numerical habitat. represent populationsin model the ad Adjacent probability occurs with Migration w w (Annex ( model cost This survival) (or fecundity a of influence the considers 1) Tunisia revealed the presence of two molecular strains, strains, twomolecular of presence the revealed Tunisia Pip31_B) and Pip31_B) and Pip31_U It assume range. initial its outside strain Pip11 w Pip31_B) along a transect orthogonal to the contact zone. We therefore developed contact zone. Wetherefore a tothe developed alongorthogonal atransect Pip31_B) We thus investigated the role of dispersal on the stability of the contact zone by zone contact the of on the stability dispersal of role the investigated thus We Our analysis of of Our analysis

b = 0.16), =0.16), w w m Pip11 cannot invade (i.e. it reaches 50% in a population). This dispersal This dispersal population). a in reaches 50% it (i.e. invade cannot Pip11 c (the cost associated with cost with (the associated Pip31_B, and that their relative frequency remains constant through time. constant remains frequency theirrelative andthat Pip31_B, will increase will when w m w Pip11 invasion can be halted only if dispersal is extremely low (e.g. a a low (e.g. isextremely dispersal if halted only be can invasion Pip11

Pip polymorphism in in polymorphism Pip < 0.08). Given the distance between < 0.08). Given distance adjacent this the populations, w Pip11 invasion m between adjacent populations across zone. the contact b or or w Pip11 infection when outside its initial range). initial its outside when Pip11 infection during 50 5 about (i.e. generations The years). c Cx. pipiens Cx. increases. Even theunder least favorable s thats there isno selective difference between σ jacent localities, about 5 localities, about jacent thein apart, km ² c

lower than 2 km² per generation. generation. per km² 2 than lower = 0.2) and a relatively high high a relatively and = 0.2) w Pip11 and and Pip11 field populations from Algeria and Algeria populations from field et al w . 2005; Flor 2005;Flor . Pip31. strainsThese are b (the frequency of of frequency (the w Pip11, Pip11, m per generation per et al et w w c Pip31_U Pip31_U ) for the the ) for Pip31_B Pip31_B . 2007). Accepted Article interpreted in terms of different different of terms in interpreted in coexist all localities differencesstudied. The thetwo between two Thus, 1). (Table induced CI indistinguishable assortative mating between individuals infected with the same the infectedwith individuals mating between assortative part of Tunisia. Tunisia. of part w w with structure:a clear Tunisiaobserved spatial in This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This (Rousset theCI expression suppress of or limit zone thecontact of the stability explain cannot mating Assortative breeding of (iv) the sites. quality of the limited dispersal (iii) local (ii) adaptation, mating, that could whymechanisms explain the case the present (in strain inducing the CI population, panmictic in predictthat, anunstructured models years.seven Theoretical w were compatiblemales with instances all Pip strains (Atyame zoneverysympatric the a are along they inthe narrow of Pip11, contact northerncenter eastPip11, and Whereas south. overmostlocalities the are infectedby either Due to negative effects of CI on host fitness, selection should favor mechanisms should mechanisms that selection favor onhostfitness, of negative CI Due to effects main result The of this study stability the is of the contact zone observed over for CIin allcrossesfound We between w Pip11 males sterilized Pip11 malessterilized et al . 2011b, 2014). 2014). 2011b, . w mod Pip11 Pip11 afemales, fraction of factors, as w w Pip31 females. However, while most of Pip31 females.the However, while of most Pip11) should invade. We thus explored several thus explored We invade. should Pip11) w w Pip11 strain does not progress: (i)assortative progress: does strain not Pip11 Pip11 and Pip11 and w Pip31_U males other with are incompatible males Pip31_U et al et w Pip31 strains, w . 1991; Turelli 1991;Turelli 1994). Among. them, Pip31 over the north and west, west, and and north over the Pip31 w Pip31 infected isofemale lines. InPip31 infected w Pip11 strain (structuration) and and strain (structuration) Pip11 w Wolbachia Pip31 males, molecularly Pip31 males, w Pip31_U and Pip31_U and w Pip31 strains may be

may leadto the w Pip31 or by by or Pip31 w Pip31_B, w Pip31 Pip31 Accepted Article trials (Table (Table trials in li are results These 4). forms ( preferential detect any not association between infected uninfected between beenhas reported discrimination stable coexistence ofincompatibleof strains This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This & Curtis Curtis 1974; Adak compatible between discriminate cannot females showed that showed that al explai cost cannot fitness Differential stability of the contact zone. theexplain too lowto its is incidence field, inthe occurs mating if assortative Thus, below). (see here thecase not is which cost, highinfection with a combined and values highest its confidence interval (Figure S2). Assortative mating may have a significant impactonly ifat on impact alimited have would it excluded, formally not be could field the in mating assortative moderate of occurrence the 100% CI, mating, complete random climate due to mitochondrial genes has genes described tomitochondrial due (Blier already been climate by another infected mosquitoes immigrant confer anincreased fitness ofmosquitoes infected by a between Association . 2002; Fontanillas Fontanillas 2002; .

The stability of the contact zone can also be explained by local adaptation. by adaptation. local explained alsobe can zone contact the stability of The pipiens D. recens w or or Pip11 rapidly supersedes supersedes rapidly Pip11 males when they occur in sympatry theymales in when occur (Jaenike molestus Wolbachia et al ). In addition, we did not detect assortative mating in laboratory laboratory assortative mating in detect wedidnot addition, ). In . 2005; Wallace 2007). Since mitochondria and mitochondria Since 2007). Wallace . 2005; et al et strains and nuclear or mitochondrial genes may indeed strains andnuclear ormitochondrial . Duron 1982; w Pip11 dynamics, even at the maximum value of theof maximum value dynamics, the at Pip11 even Wolbachia ne with previous studies showing that showingthat withpreviousstudies ne n the n zone stability of the contact w Pip31_U, as expected from models Pip31_U,assuming models from asexpected w Wolbachia Pip For strain. local instance, to adaptation transmission Althoughand nofitness cost. and incompatible partners(Laven incompatible 1967b; and w et al Pip11 or Pip11 or . 2011).. Cage experiments population Drosophila subquinaria Drosophila in populations. Suchmatingfield w w Pip strain at the expense the of at expense Pip strain Pip31 and either either and Pip31 et al . 2006). However, we did we However, 2006).. et al . 2001; Ehinger Ehinger 2001; . Wolbachia females and females

Cx. pipiens Cx. Cx. pipiens are et et

Accepted Article (mean frequency 12%). Modeling the spatial dynamics Modeling(mean offrequency the 12%). thethree cytotypes, assuming a 3% ranges from crossing through experiments, in localities five to estimated frequency, 20% w predominates (i.e. (i.e. predominates (i.e. w w dispersalLow prevents likely zone. contact stability of observed that fitnessunlikely differences with associated invade patterns,that weshowed cytotype clinal However, the modeling w two found mitochondrial we system, In our adaptation. local differential confer that haplotypes mitochondrial different different co-transmitted, maternally This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This differential cost. Moreover, we also compared the observed no assuming by models predicted asrapidly as cages inpopulation near-fixation reached is very since unlikely, cost extreme Such over 2). 0.08 (Figure probability dispersal on pooled replicates, replicates, on pooled fitness cost thea differential of0.2: wassignificantly faster dynamics (Exact tests binomial maternal transmission and 100% CI, mating, complete assuming expected those with random Pip31_B frequency therefore represents the threshold controlling controlling the threshold represents frequency therefore Pip31_B Pip31 should obviously be eliminated from localities where where localities obviouslyfrom Pip31 should beeliminated and Pip31_U Pip31 infections (data not shown), which could potentially participate to local adaptation. local to participate potentially could which not shown), (data infections Pip31 w Pip11 > w In a panmictic unstructured population the coexistence of the three cytotypes ( cytotypes three the of coexistence the population unstructured a panmictic In Pip31 area, Pip31 area, even with a costassociated verywhen ( high fitness w w Pip31). Pip31). Pip31_B) is not possible. Because isnotpossible. Pip31_B) w P Pip31 > Pip31 = 3.10 = w cytb Pip11 should also increase up to fixation in localities where where in localities up also increase should to fixation Pip11 haplotypes,being each associated strictly witheither -6 and1.10 w w Pip11), when its frequency is higher than than frequencyis higher its when Pip11), Pip11 invasion Wolbachia -4 , respectively for generations 3 and 4). it respectively Thus, generations , is3 and for strains could thus be thus strictly with strains associated could w

w

Pip11 and Pip11 is favored by unidirectional CI, isfavoredbyPip11 unidirectional w w Pip11 dynamics in the cages Pip11 is the most abundant isthe Pip11 w Pip31 could explain theexplain could Pip31 w w Pip11is expectedto Pip11 invasion. Pip11 invasion. This c = 0.2), providing a w w Pip31_B. Pip11 or w w w Pip11, Pip11, Pip31 Pip31 Pip11 Pip11 Accepted Article pipiens publishedin dispersal estimates lower than much This is generations). fifty (about period prevent to generation within the range of low mean dispersal distances (i.e. 1 dispersal distances (i.e. lowmean of range the to within 2 mark- recent from deduced km) few along the coast in coast in Bize along the few for a except sparsely populated, were mostsites sites: larval survey breeding of onthe based infected bynearby populations with infected populations stalledonlybe if could Tunisia even block Influence of the and/or quality quantity of breeding sites Flor (Telschow studies theoretical previous with consistent is this Moreover, stability. zone contact the better explains that parameter is the dispersal currently low model, our respectively, more distant events could sothat notobserved. be results Accordingof tothe set were 3 studies,within recapture traps in these for and 2 site therelease km from (Ciota State New York 2008), or (Lapointe inHawaï studies recapture we deduced that we deduced themean distance axialdispersal of frequency This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This by Hancock by and Godfray(2012). larval Hancock Mosquito fluctuating local densities, in which case et al et , e.g. Local spatial heterogeneity such as differential host density could also slow down or or also slow down host could density differential heterogeneity suchas spatial Local These . 2007). . 2007). Wolbachia σ w ² Culex Pip31_B = 43 km² per generation in southern France (Lenormand (Lenormand France in southern 43 = generation per km² populations may actually be better described as metapopulations with metapopulations as better described be actually may populations progression (Barton & Turelli 2011). However, However, 2011). Turelli & progression (Barton w Pip11 progression from one locality to the next within the five-year five-year the next within localityto the one from Pip11 progression b = 0.16 and a large cost large a and =0.16 rte (#15) and Tunis rte (#15)and (#16). w Pip11. This does not correspond to our field observations, toour does This field Pip11. not correspond Wolbachia c and pupal stages require watered breeding breeding watered stagesrequire and pupal = 0.2 of of =0.2 spread could be much slower, as shown much be slower, could spread σ ² ² should be 2 than should lower km² per w w Pip11 (which is unlikely high), high), Pip11 (whichis unlikely Pip31 are far denser than the than denser far Pip31 are

et al w Pip11 progression in in progression Pip11 et al . 2012). However, However, 2012). . . 1999). This is is This . 1999). et al . 2005; . Cx. Accepted Article w present when breedingsitesbeing are adults This brakethe as may a on rewatered. act year every reset probably thus events Extinction-recolonization tospring. from again filled autumn and above andlowrainfall, 30°C to temperatures due Indeed, parameter. alimiting constitute may suitable sites of availability the and develop to sites This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This three of segregation the the areas of Conclusion tolevel-off 60-70%. appears frequency at a Aïn Tounga, #51andBéjaGare, #54) #29; Hamra, #46; Manar, (El 10% above frequency sites its was sympatric where the of most directionally incompatible with with incompatible directionally coexistence in the in field. the coexistence study out This points of thatthoroughthe a host knowledge Flor 2005; in dynamics by metapopulation strengthened probability a dispersal from low results likely and 50 located to 60 be stability cannot explained by km observed apart. The adaptation, local in 1996Mateurdetected (#15) (#14) Bizerte and and stable since have been for time, a longer of period the w Pip11 progression, despite its CI advantage, and may explain whyitsexplain advantage, despite may and CI progression, Pip11 Pip11 ( Pip11 Cx. pipiens w Pip11 and Pip11 We highlighted the presence We highlightedof presence the a very narrowcontactzone the between distribution Cx. pipiens w Pip31_B) is present at low (~ low frequency at present Pip31_B) is w et al et Pip11 and and Pip11 . Although this result has already been predicted theoretically (Telschow theoreticallypredicted (Telschow has been already result this Although . w . 2007), we present here the first empirical evidence of such such a stableof empirical evidence . wepresent 2007), the first here Pip31 strains appeared stable over a seven-year survey. The situationstable might overaseven-year strains survey. The appeared Pip31 breeding sites in the contact zone are temporary, dry during the summer summer during the contact dry temporary, the zone are in breeding sites w Wolbachia Pip31 molecular strains in Tunisia. The situation is complex due to to due complex situation is The Tunisia. strains in molecular Pip31 w Pip11, a Pip11, strains: while the mostfrequent while the strains: nd even in Fontaine (#50) where the (#50)where nd even inFontaine w Pip31 strain bidirectionally incompatible with incompatible Pip31 strain bidirectionally 12%). The narrow contact zone between zonebetween narrowcontact The 12%). w Pip31 infected mosquitoes have been infected Pip31 w Pip11 w Pip frequencies to those of to frequencies Pip

w in 1997 in Tunis (#16), (#16), in 1997Tunis Pip11 did not invade in invade not did Pip11 w Pip31 strain is uni- is Pip31strain w Pip11 Pip11 et al . Accepted Article Acknowledgments Acknowledgments impact how understand to required regionstudied is the prevailing in conditions the environmental and dynamics This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This of Diversification O(2011a) Duron M, N, Weill Pasteur F, CM, Delsuc Atyame evidence AB M, M,of Sarih (2012) Molecular F, Failloux Tijane Amraoui References dede Montpellier (Unitée Recherche, Mixte CNRS-IRD-UM2 5554). 2014-209 contribution is the study This Platform. Genomic obtained ISEMEnvironmental data onthe were All sequence manuscript. on the help in Badii Vidaler, Kamoun, Christelle Pierre-Yve technical assistance; Arnaud for Berthomieu Unaland samples Sandra Algeria; collected in Alout for Haoués to grateful very are We Bismuth. supervisedRobert Méditerranée by This work was done in collaboration with Atyame CM, Duron O, Tortosa P, Pasteur N, Fort P N, Fort Pasteur P, O, Tortosa Duron CM, Atyame 5 endosymbiont in the inthe endosymbiont form 2772. , determinants control the evolution of cytoplasmic incompatibilities in incompatibilities of evolution cytoplasmic control the determinants mosquito populations. populations. mosquito

83. molestus

Wolbachia Wolbachia and hybrids and Wolbachia -based control strategies.-based control genotyping and crossing experiments; Pascal Milesi for helpful comments comments helpful for Milesi Pascal experiments; crossing genotyping and Culex pipiens Molecular Ecology distribution evolves in natural populations and how this might and populations this might how evolves innatural distribution pipiens / molestus mosquito. mosquito. , in Morocco, North Africa. Africa. North Morocco, in 20 s Biba Nkouka and Olivier Sacchi, for theirNkoukaOlivierSacchi, for and s Biba Comité de Coopération MarseilleComité de Provence Molecular Biology and Evolution and Biology Molecular , 286-298., of the Institut des Sciences de l’Evolution de l’Evolution Sciences des Institut the of et al et . (2011b) Multiple (2011b) . Parasites & Vectors Culex pipiens Culex pipiens , Wolbachia Wolbachia 28 , 2761- ,

Accepted Article Blagrove MS, Fa C, Arias-Goeta AtyameLabbé CM,P, E, Dumas MilesiP, CharlatS This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Curtis CF, Brooks GD, Ansari MA, Grover KK, Krishnamurthy BS Krishnamurthy KK,MA, Grover Ansari BrooksGD, CF, Curtis Curtis Curtis CF, Adak replacement in T(1974)Population in sterility of cytoplasmic importance evolutionary On the GS(1959) Watson E, Caspari Barton NH, Turelli M waves(2011) Spatial of advance with bistable dynamics: cytoplasmic Bahnck CM, forms of assay Fonseca twogenetic toidentify DM (2006)Rapid the Ciota AT, Drummond CL, Ruby MA, Drobnack J, Ebel GD CL, J,Drummond AT, Ebel MA, Drobnack Ciota Ruby Bian G, Lu Zhou Joshi D, P, Dong G Y, the evolution of cytoplasmic incompatibilitythe in 31 translocation. a incompatibility and cytoplasmic of on control incompatibility. Tropical Medicine and Hygiene and Medicine Tropical ( Entomology facility. treatment wastewater Culicidae) from a (Diptera: mosquitoes mosquitoes. 255-60. ofthe of SciencesNational Academy Proceedings of the UnitedStates of America in andblocks transmission dengue incompatibility induces cytoplasmic to refractoriness induces and populations effects. analogues of genetic Allee and Culex , 181-190. 181-190. , ) pipiens , Evolution

49 , 35-42. 35-42. , L. (Diptera: Culicidae) and hybrid populations. populations. hybrid Culicidae) and L.(Diptera: Bulletin Health Organization of theWorld Bulletin Culex quinquefasciatus , 13 , 568-570. 568-570. , , illoux AB, SinkinsSP (2012) 75 , 251-255. , et al American Naturalist American by release of males of a strain integrating Plasmodium . (2013) (2013) . Culex pipiensCulex Entomologia Experimentalis et Applicata et Experimentalis Entomologia Culex fatigans Wolbachia et al et infection. . (2014) (2014) . et al , , 51

. 178 invades Plos One , 249- 255. , 249-255. . (2012) Dispersal of Dispersal of . (2012) Wolbachia Science bymeans of cytoplasmic , E48-75. E48-75. , et al et Wolbachia American Journal of Journal American . (1982) A field trial Afield . (1982) Anopheles stephensi Anopheles Journal of Medical of Journal , , 9 Aedes albopictus Aedes 340 , e87336. divergence and divergence strain , 748-751. 748-751. , , Culex Culex w 109 Mel , , .

Accepted Article Engelstadter J, Telschow A (2009) Cytoplasmic incompatibility and host population population andhost incompatibility Cytoplasmic A J, Telschow (2009) Engelstadter an DNAalong variation N E, Mitochondrial Perrin (2002) Fontanillas Petit M, P, Ehinger Dumas E, Atyame CM, Milesi P, Fonseca DM, Shaikevich EV Shaikevich DM, Fonseca P, Milesi CM, Atyame E, Dumas This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This M M, (2011)Many compatible Duron O,Weill Raymond M(2006b) Hypervar Weill Fort P, Duron O, Duron O, Lagnel J, Raymond M, K, Bourtzis Fort P Duron O, Boureux A, Echaubard A, P,Duron O, Berthomieu Boureux Berticat C A, A,Berticat C UnalC, S, Berthomieu Duron O,Bernard M(2006) Weill Duron O, structure. Ecology white-toothed greater shrew, the gradientin altitudinal of populations natural BMC Evolutionary Biology Evolutionary BMC structure of pipiens ofLondonof the RoyalSciencesProceedings Society Series B-Biological of high number unexpected Ecology in the mosquito incompatibilities cytoplasmic modulating embryo in incompatible crosses. crosses. inincompatible embryo recombination. and superinfection of polymorphism expression of ankyrin domain genes in genes in of domain ankyrin expression . , , Journal ofBacteriology Journal 11 15 Heredity , 939-945. , , 3061-3071. 3061-3071. , Wolbachia Wolbachia , 103 Culex pipiens and cytoplasmic introgression in a complex of mosquito species. acomplex introgression in andcytoplasmic , 196-207. , 196-207. Wolbachia , 13 in the mosquito inthemosquito , 181. , Heredity Wolbachia 189 Molecular Ecology infection influences the development of developmentinfluences of the infection mosquito. mosquito. , 4442-4448. , Wolbachia , 96 variants in the mosquito inthemosquito variants , 493-500. , iable prophage WO sequences describean WO prophage iable Culex pipiens Culex Heredity variants infecting the mosquito mosquito the infecting variants , et al 14 , Wolbachia , 1561-1573. 1561-1573. , 106 et al et . (2005) Transposable element . (2005)Transposable Crocidura russulaCrocidura : evidence of genetic evidence diversity,: , 986-993. et al . (2006a) Tracking factors factors Tracking (2006a) . Culex pipiens et al . (2007) Variability and (2007) and . Variability strains coexist within coexist strains . (2013) Population Population (2013) . , 273 Culex pipiens Culex pipiens , 495-502. , 495-502. . . Molecular Molecular Molecular Culex .

Accepted Article Flor M, Hammerstein P, Telschow A (2007) P, Hammerstein Telschow Flor M, in incompatibility cytoplasmic ofsymbiote-dependent PE dynamics (1978)Onthe Fine and human mosquitoes biting" "Bird MA (2011) LD, Kilpatrick AF, Kramer Farajollahi This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This DM, Keyghobadi N, CA Fonseca Malcolm Fontanillas P, Dépraz A, Giorgi MS, Perrin N (2005) Nonshivering thermogenesis P, Dépraz Nonshivering capacity thermogenesis Giorgi MS,Fontanillas Perrin N(2005) A, Harbach R (2012) R Harbach of spread Modelling HC the PA, (2012) Godfray Hancock Hilgenboecker K, Hammerstein P, Schlattmann P, Telschow A, Werren JH (2008) How How JH (2008) Werren A, Telschow P, Schlattmann P, K,Hammerstein Hilgenboecker Hoffmann AA, Turelli M, Hoffmann LGHarshman Factors thedistribution affecting (1990) of in Variation Y (2005) Kato Y, Ishii M, Hiroki Culicine Genetics and Evolution Infection ofof the role disease: a review Journal of Evolutionary Biology Evolutionary of Journal host in parapatric polymorphism populations. of infection stability and the incompatibility vectors in the inthe vectors heterogeneous environments. environments. heterogeneous shrew, white-toothed greater and inthe gender DNAhaplotypes mitochondrial to associated Microbiology Letters Microbiology with infected species are many perspective. cytoplasmic incompatibility in Research Ecology Evolutionary inducing Crocidura russulaCrocidura mosquitoes. Wolbachia Journal of the American Mosquito Control Association Mosquito theAmerican Journal of Culex pipiens Culex Culex pipiens in butterfly the , Journal of Invertebrate ofInvertebrate Journal Pathology 281 . , 215-20. 215-20. , Molecular Ecology Molecular complex. complex. Journal of the Interface ofthe Royal Journal Society : species versus species complex – taxonomic historyand– taxonomic species complex versus: species Drosophila simulans Drosophila , Wolbachia , , 7 11 20 , 931-942. , Culex pipiens Culex , 1577-1585. 1577-1585. , , 696-706. , Science Eurema hecabe Eurema , Mehmet C, Schaffner F F Schaffner C, Mehmet ? - a ? statistical analysis of data.current Wolbachia the prevalence of cytoplasmic incompatibility- of cytoplasmic the prevalence , , 303 14 complex mosquitoes in epidemiology. in epidemiology. mosquitoes complex , 661-670. , 661-670. , 1535-1538. 1535-1538. , . Genetics -induced unidirectional cytoplasmic-induced unidirectional across the Japanese archipelago. , 30 , 10-18. , 10-18. , 126 Wolbachia , 933-948. et al , , 28 9 , 3045-3054. , 3045-3054. , 10-23. 10-23. , . (2004) Emerging Emerging . (2004) in spatially spatially in Fems Accepted Article Hoffmann AA, Montgomery BL, Popovici J, Iturbe-Ormaetxe I, AA, Montgomery Johnson J, PHHoffmann BL, Iturbe-Ormaetxe Popovici This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Hoshizaki S, Shimada T (1995) PCR-based detection of of detection PCR-based T (1995) Shimada Hoshizaki S, Laven H (1967a) Speciation and Evolution in in Evolution and Speciation H(1967a) Laven Jaenike J, Dyer KA, CornishC, Minhas MS Lenormand T, Bourguet D, Guillemaud T, Raymond M (1999) Tracking the evolution of of theevolution Raymond T, Tracking M (1999) T, D, Bourguet Guillemaud Lenormand of Eradication (1967b) H Laven of Dispersal DALapointe (2008) Z, SB,Kambris Blagborough AM, MS,Pinto GodfrayBlagrove HC Jeyaprakash A, Hoy MA Long (2000) PCR improves Kriesner P, Hoffmann AA, Lee SF, Turelli M, Weeks AR (2013) Rapid sequential spreadof Rapid sequential (2013) AR WeeksM, SF, AA,Lee Turelli Hoffmann P, Kriesner transmission. Successful establishment of (Homoptera: Delphacidae) in central Japan: has the distribution of distribution the Japan: has central Delphacidae) in (Homoptera: of innaturalpopulations infected microorganisms, incompatibility Wolbachia recently? insecticide resistance in the mosquito Nature Disease, J. Wright andR. Pal, Editors. Elsevier: Amsterdam. rain forest. andinhibits gene immune expression stimulates 393-405. in76% sixty-three sequencesarthropod found species. of two gambiae Wolbachia , 216 . Insect Molecular Biology PLoS Pathogens infection in in infection Journal ofMedicalJournal Entomology , 383-384. , Nature variants in variants , 476 Drosophila , 454-457. , 454-457. , Drosophila simulans Drosophila 6 , e1001143. Culex pipiensfatigans Culex Culex quinquefasciatus Culex Wolbachia . , PLoS Biology 4 , 237-43., Culex pipiens in , Culex pipiens 45 (2006) Asymmetrical reinforcement and and reinforcement (2006)Asymmetrical , 600-609. 600-609. , . Aedes PLoS Pathogens , Plasmodium 4 . , e325. Nature through cytoplasmic incompatibility. incompatibility. cytoplasmic through Wolbachia populations tosuppress populations dengue (Diptera: (Diptera: Culicidae) in Hawaiiana . In Genetics of Insect Vectors of of Vectors Insect of Genetics In . , 400 Insect Molecular Biology Insect development in development DNA amplification: , 861-864. Wolbachia , 9 et al et , e1003607. e1003607. , Laodelphax striatellus Laodelphax . (2010) (2010) . Wolbachia , cytoplasmic et al et Wolbachia Anopheles Anopheles . (2011) . (2011) spread wsp , 9 ,

Accepted Article Moreira LA, Iturbe-Ormaetxe I, Jeffery JA, Lu Pyke AT JA, GJ, I, Jeffery LA, Iturbe-Ormaetxe Moreira H, S (2004) Mercot Charlat This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Rousset F, Raymond M, Kjellberg F (1991) Cytoplasmic incompatibilities themosquito in incompatibilities Cytoplasmic FKjellberg Raymond M, (1991) F, Rousset DNA Molecular tissues. of plant from InPlant AJ (1988)Extraction Rogers Bendich SO, (2002) C Stauffer M, Riegler software forgenetics 1.2), apopulation (version Genepop (1995) Raymond M, RoussetF (2003) TW Scott Rasgon JL, Rousset F (2008) Genepop’007: a complete re-implementation of the re-implementation software genepop for acomplete Rousset F(2008) Genepop’007: Telschow A, Yamamura N, Werren JH (2005) Bi JH(2005) N, Werren A, Yamamura Telschow 120 simulans Drosophila Tephritidae). Tephritidae). populations flycherry European ofthe fruit incompatible ecumenicism. and tests exact populations. natural Culex pipiens Cell symbiont in Culex pipiens Culex 1-10. Publishers Academic BiologyVolume A6. Edited Manuel. by RA. Gelvin SB, Schilperoort Boston: Kluwer Biology Windows and Linux. and Windows Theoretical Biology Theoretical stable coexistencethe of two , 51-59. 51-59. , , 139 , 4 , 1268-1278. , 69-81. , Aedes aegypti Aedes Molecular Ecology Molecular mosquito species complex: parameter estimates and infection dynamics in : How to explain a cytotype polymorphism? polymorphism? acytotype to explain How : , Genetics : polymorphism and levels of cytoplasmic incompatibility. incompatibility. of andlevels :cytoplasmic polymorphism 235 Molecular Ecology Resources Ecology Molecular , 265-274. , 265-274. limits infection with Dengue, Chikungunya, and Plasmodium. andPlasmodium. infection with Dengue, Chikungunya, limits Wolbachia Wolbachia , Journal of Heredity of Journal 165 Wolbachia Wolbachia , 11 , 2029-2038. 2029-2038. , , 2425-2434. and cytoplasmic incompatibilityCalifornia inthe andcytoplasmic infections and superinfections in cytoplasmically in superinfections and infections infections in strains in parapatric host populations. hostpopulations. inparapatric strains directional cytoplasmic incompatibility and incompatibility cytoplasmic directional , 86 , 8 , 248-249. 248-249. , , 103-106. 103-106. , Drosophila melanogaster Drosophila Rhagoletis cerasi Rhagoletis et al Journal of Evolutionnary ofJournal Evolutionnary . (2009)A Wolbachia (Diptera, (Diptera, Genetica Journal Journal and ,

Accepted Article Turelli M, Hoffmann Turelli AA (1995)Cytoplasmic in incompatibility andtheir hosts. microbes incompatibility-inducing of Evolution M(1994) Turelli in factor incompatibility inherited of an spread AA Rapid Hoffmann (1991) M, Turelli This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This article. this of version online the in asSupporting Information available inthisstudy are sets used Data accessibilityData FD Frentiu I, LA, Iturbe-Ormaetxe Moreira Johnson PH, T, Walker S. Fourth with Statistics Applied Modern nnet. Rpackage BD(2002) Ripley WN, Venables Wallace DC (2007) Why do we still have a maternally inherited mitochondrial DNA? mitochondrial inherited maternally a have westill DC(2007)Why do Wallace hosts P a host of (2012) Still for Zug Hammerstein R, ME(2008) L, Baldo JH, Clark Werren JH, of D, Distribution GuoL Werren Windsor (1995) dynamics and parameterestima and dynamics 48 California virulent virulent Springer, Edition. New York. insights from evolutionary medicine. evolutionaryinsights from medicine. aegypti suggeststhat 40% of terrestrial arthropod species are infected. biology. 197-204. arthropods. , 1500-1513. 1500-1513. , populations. populations. Nature Microbiology Reviews Nature Wolbachia Drosophila Proceedings of the Royal Society of London Series B-Biological Sciences SeriesB-Biological Society London of Royal the of Proceedings infection blocks dengue transmission blocks rapidly invades infection transmission dengue and Nature . Nature , 476 , tes naturalfrom populations. 353 , 450-455. 450-455. , , 440-442. Annual Annual Review ofBiochemistry , 6 Wolbachia , 741-51. 741-51. , : master manipulators of invertebrate invertebrate : master manipulators of Wolbachia Wolbachia Genetics PLoS One : analysis of recent data Drosophila simulans , et al et , among neotropical neotropical among 76 140 , 781-821. 781-821. , , 7 . (2011) A . (2011) non- , 1319-1338. , 1319-1338. , e38544. e38544. , Evolution Aedes Aedes , 262 : , , Accepted Article sympatric (i.e. observed in a same breeding site); those underlined, i.e. Hamra (#29), Hamra i.e. observed underlined, same breeding (i.e. in a those site); sympatric FigureLocalities 1. inbold are those where pipiens Culex molecular the of Distribution S1. Table Supporting information This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This mosquitoes sympatric between crosses Reciprocal S4. Table performed. was not lines Isofemale b, Tunisia; lines from Isofemale lines a, that isofemale thesame Table are S1). from Sou2means locality instance(see Sou1, collected in each cross is inparentheses. The name of mosquito line byfollowed number, for egg rafts number of The underlined. CIare bidirectional and crosses shaded are Incompatible females. The (C), forcross iscompatible aHR > (IC) incompatible 90%; and for 0% HR. allopatric localities between crosses S3.Reciprocal Table and the toexamine used genes Table of and characteristics of primers S2. List givenvalues respectively. are lower the or upper when binomial from calculated frequencies of distribution; the Duron from are Douz (#18) ( Gourbi (#11), Bled G5Samples (#1), time. changes significant of Béja Gare(#54)displayed (#50) and Fontaine Culex pipiens Culex . A, from Zerga (#30). B, from Aïn Tounga (#51). The cross is compatible (C), (C), crossis compatible The (#51). Tounga Aïn Zerga from B, A, (#30). . from field samples from Algeria and Tunisia and from Algeria samples field polymorphisms . A, w Pip11 males and males Pip11 et al . (2006b). N, sample size. N, (2006b). sample Confidence. were intervals(c.i.) .

w Pip11 andPip11 Wolbachia #12), Menzel (#13), Tunis (#16), Bismuth (#17), Bismuth (#16), Tunis (#13), Menzel #12), w w Pip31 females. B, Pip31 females. Pip11 and from Algeria. Dashes indicate that the crossAlgeria. that from Dashes indicate strains w Pip31 infectedPip31 isofemale lines from w w . Localities are numbered asin numbered Localities are . Pip31 infectedPip31 were mosquitoes Pip11 and and Pip11 w w Pip11 and Pip11 Pip11 Pip11 strain equal 0 or 1, the w Pip31 and males w Pip infections with with Pip infections w w Pip31 infected Pip31 infecting infecting Pip31 Wolbachia w Pip11 Pip11

Accepted Article crossing crossing of relationships Crossing S5. Table parentheses. egg in collected The of number CI eachcross are underlined. bidirectional inis rafts (IC) incompatible HR>90%; for crosses and for HR=0%. a Incompatible are shadedand This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This S6Table collected in each crossis inparentheses. (#3, Tounga,and #51) Aïn and Table S7 Table inparentheses. is for 0% crossesHR. Incompatible shaded.are The number of egg rafts collected in crosseach The crossis Sokra(#37). compatible Sok from and Khetmine, #62) and #57 Rajel, Ras #51; Tounga, Aïn between w strains Figure S1. Patterns of and Pip31 infected mosquitoes w w Pip31 infected mosquitoes in Tunisia, identified by the Pip31_U). All crosses were incompatible (IC incompatible were crosses All Pip31_U). . w w . . Pip11 isofemale-derived males from three localities (El Battan, #43; El Manar, #46 #46 Manar, El #43; Battan, (El three localities males from isofemale-derived Pip11 Occurrence of Occurrence Pip31 isofemale-derived males from fives localities (Hamra, #29; El Manar, #46; Manar, malesfrom #29; El (Hamra, fivesPip31 isofemale-derived localities Estimation of the frequencies of of frequencies the of Estimation

ank2 pipiens w . This experiment was performed through unidirectional crosses unidirectional through was performed experiment This . and and Pip31 females from the same isofemale line Har from Harash Harash from Har line thesameisofemale from Pip31 females and MutL molestus w PCR markers in Pip11 males Pip11 w forms of of forms (C), for a HR > 90%; and incompatible (IC) (IC) incompatible and HR>90%; a for (C), Pip31_U and Pip31_U and ) with HR = 0%. The number of egg rafts egg rafts of number The 0%. = HR ) with . This experiment was performed by wasperformed experiment This . Culex pipiens w Pip11 thesame from Pip11 females line w Pip11 and and Pip11 w CQ11 Pip31_B cytotypes cytotypes amongPip31_B microsatellite locus taxa among w Pip31 molecular molecular Pip31 w Pip11 Pip11 .

Accepted Article elements are 1 are elements and 0 crosses, compatible for then We 1). (see Table ones incompatible for is the vector of cytotype frequencies, and of vector cytotype frequencies, is the correlation of mates( correlation cytotypes between different fitness cost reaching 50% the afrequency of of from (i.e. ( probability mating assortative without or with demes, distant increasingly Figure dispersalprob S2. Effect of This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Annex strains1: Modeling near dynamics contact the zone PF,theMW. Wrote paper: CMA,PL,PF, FR, NP,AB, MW. experiments: CMA, MB,PM, OD,ED,PF,Analyzed MW. the data: CMA, PL,NP, FR, OD, the Performed PF, MW. NP, OD, CMA, designed experiments: the Conceived and Author Contributions all cytotype magnitude withdata, we Inagreement populations. assumed panmixia orpartial assortative mating of of array linear as a represented is The transect was modeled. zone contact tothe orthogonal 0.2 and and 0.2 2)as long(deme as deme (deme closest the invade to able be not will aslongas 1) 50 generations. w Pip31 individuals Pip31 individuals was fixed as The of evolution The the spatial clinefrequency theof cytotypes along atransect F i frequency due to incompatibility incompatibility dueto frequency = 0.63, no value of m allowed m of value no =0.63, F m , complete and complete transmission CI. Within thesite each change in ) below which the the which below ) m s associated tothe < 0.13, the next one (deme 3) as longas (deme as 3) one next the 0.13, < w Pip11 deme given invading a from isprevented cytotype b ability on the progression of of the progression on the ability = 0.12. For example, with example, For =0.12. F ). Thefrequencyof the w A x' Pip11 cytotype ( cytotype Pip11 is the matrix describing incompatibility, which which incompatibility, matrix describing the is w i population) after 50 generations is indicated 50 after for is population) indicated generations can be represented as Pip11 to invade even the closest deme deme Pip11 closest toinvade the after even c m ) and for different levels of levels different for ) and F < 0.2, etc... Note that for that Note etc... < 0.2, = 0.39 and = 0.39 w Pip31_B cytotype among Pip31_B cytotype . m

x The maximal dispersal ' < 0.09, the next one 0.09, the < i w =( Pip11 cytotype into into cytotype Pip11 x A x c ) = 0.2, =0.2, i / x A x where where w Pip11 Pip11 c = x

Accepted Article frequency frequency dynamics of the tothe Inparticular, within-population adapted present model. therefore be betweentwocytotypes can incompatibility bidirectional or unidirectional w w the of of adjacent populations.adjacent Inaddition, a (orsurvival) cost fecundity speed wavewith travelling thenform should cline a andTurelli the (2011), (14b)inBarton eq. By with analogy exact dynamics: exact where where and can be used to argue that limited dispersal is required to explain the apparent stabilityof theapparent to explain dispersalrequired is limited that argue used to be can and readily computations adapted take These be into can account a to dispersal). cost fecundity considered. assumed density regulation, followed byprobability with a dispersal followed regulation, densityassumed This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This only their total frequency total their only Pip31_B among Pip31_B of frequency relative the to (inproportion onlyweakly incompatible and males, Pip11 w w Pip31_B among all Pip31_U and and Pip31_U σ There is no selective differenceno selective between is There is the isthe square themean square-root of distance dispersal transectalong the (axial p = = p 1-

q can be approximated by a cubic function matching the stable points of the the of points stable the matching cubic function bya approximated be can w w Pip31_B cytotypes Pip31_B as collectively a single withcytotype, incompatible Pip31) with w Pip31 remains constant through time. Thus it is equivalent to model to model it is Thus time. equivalent through constant remains Pip31 q ismodified by incompatibility with w Pip11 females. Previously developed results results for developed Previously Pip11 females. w

Pip31_U and Pip31_U and , c for the for w w Pip31_B (Table 1), so(Table Pip31_B that Pip11, but the frequency the but Pip11, . m to disperse to the two w Pip11 cytotype was cytotype Pip11 w Pip11 Pip11 b

Accepted Article confirms that low dispersal is required to prevent the advance of advance prevent the to low dispersalisrequired that confirms 50 zone after contact initial the from distance given a at population) the It generations. toprevent dispersal allowed of occurred populations adjacent Dispersal among array. on the 12%). initial frequency of of frequency initial function,zone all wasastep contact the large circular array, twosymmetric with clines evolving simultaneously. initial The state of by were avoided assuming a effects edge computations, these in Forsimplicity, conclusions. the infection is associated with a relatively high fecundity cost(e.g. fecundity high isassociated arelatively infection with the example, the approximation overestimate the factoroverestimate the speed bya example, the approximation for 2 di small for speed the overestimate substantially can approximation the aresupported, approximation the analytical of conclusions qualitative of any fitness cost). Equivalently, estimates of estimates Equivalently, cost). fitness of any & Turelli 2011;Telschow & Turelli so than immigration, stronger be cytotype will rare the against -selection incompatibility bidirectional for weakly - rate even dispersal enough since expected low thisis for extent, Tosome low. wouldbetoo approximation the reduction in fecundity of cytotypes outside their initial range ( initial their outside cytotypes of fecundity in reduction larger of a form the in of cytotypes, adaptation local check that to further and conclusions conclusion. this and reduction zone, contact the moderate that in fecundity of a This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This The results of our model are presented in Figure in Figure presented amount are maximum ourmodel the of displays results It 2. The Numerical iterations of the recursions we recursions of the iterations Numerical w Pip31_B within Pip31_B within all et al et . 2005). . 2005). w Pip11 from reaching 50% in a population (i.e. from invading invading from (i.e. inapopulation reaching Pip11 from 50% w Pip11 on one side and all side onand one Pip11 w Pip31 was (i.e. observed around the varied one that there will be no wave of advance (Barton (Barton advance of wave no be will there that m spersal rate among discrete populations (for (for populations rateamongdiscrete spersal deduced from an observed advance using re performed to check these different these different recheck to performed w w Pip11 little impact on has Pip11 ( Pip11 w m m c Pip31 on other. Thethe c c ), the didnotaffect = 0.02 in the absence the = in 0.02 = 0.2). Although the the Although 0.2). m m < 0.08), even if evenif <0.08), Accepted Article generation. In this equation, equation, this In generation. in squared units of intersite distance pe of intersite units in squared #17 and#18)are studya from previous (Duron #16, #13, (see year #12, (#1, Samples Table S1).#11, thefirst from data represented only we of the contact zone. When samples from the samelocality were analyzed for several years, (sampl cluster-II 43-46)and 36-38, 26-29, 19, 16, inTable those to correspond S1.Numbers Sample

dispersal distance), i.e. i.e. distance), dispersal numerical analysis implies implies analysis numerical units) it is required that that is required it units) 5 fifty observations years i.e. of (about generations, speeda of1 unit distance 50 for time in next the to onesite from progress to prevent a approximation, analytical to the According This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This Gare, #50 andBéja #54), change asignificant of and Tunisia from Algeria populations Figure 1. Distribution strainsmolecular of the Figure Legends

5 other by distant about each zone are from contact inthe sites Breeding km. m m < 0.0023. On the other hand, in similar conditions (i.e. no cost), the no cost), conditions (i.e. similar in hand, other Onthe < 0.0023. m b b < 0.015 (i.e. < (i.e. 0.015 = 0.12 implies = 0.12 implies . In the localities .in Inthelocalities boxes #29;Fontaine, (Hamra, σ r generation. r generation. is5 That km times in more ² ² < 0.375 km² per generation). pergeneration).

et al < 0.06 km² per generation (mean square axial (mean generation km²per < 0.06 es 30-42, 47-56) for analyzing the structure structure the 47-56) for analyzing 30-42, es s were grouped into cluster-I (samples 14- (samples cluster-I grouped into s were . 2006b). 2006b). . w w Pip frequencies was found over time. time. over was found Pip frequencies Pip11 and

w Pip31 in Culex pipiens Culex 2 per

Accepted Article bidirectional CI between between CI bidirectional w increasingly distantdemes the of Figure progression onthe probability dispersal of Effect 2. This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This strains incompatible with w bolded. are crosses Incompatible (HR=0%). cross = incompatible 90%); 0 (HR) > rate (allhatching cross 1 = compatible w cytotype ( after population) 50 the isi generations of 50% frequency of reaching a from (i.e. deme given a invading from prevented cytotype is the figure, only the cases where next(deme one <0.04, the 3) aslongas invade the demeclosest (deme 1) as long as Table 1. Summary of CI patterns occurring between the the between occurring CIpatterns of Summary 1. Table

Pip31_U and and Pip31_U Pip31_B females, whilst females, Pip31_B Pip11 cytotype ( cytotype Pip11 Females w Pip11 and b ) among all ) among all w w w w Pip31_B Pip11 Pip31_U Pip31_B asimilar Pip31_B display CI pattern (boxed). c ). For example, with example, ). For w Pip11 females. So, two crossing types exist between the molecularthe between exist So, crossingtypes two females. Pip11

w w Pip31: unidirectional CI between CI between unidirectional Pip31:

Pip31 individuals and for different fitness costs associated to the costs associated fitness to different for and individuals Pip31 w .

Pip11 and Pip11 and The maximal dispersalprobability ( maximal The w Pip31_U males and werecompatible m w <0.1 are presented. w Pip11 Pip11 males are always incompatible with are incompatible Pip11 males always 0 0 1 m w ndicated for different frequencies of of the frequencies different ndicated for w b <0.05, etc... Note that, for concision and of clarity and for concision that,Note etc... <0.05, Pip31_B. Note however, that the females from from thefemales that however, Note Pip31_B. = 0.16 and =0.16 and m Pip31_U <0.03, the next one (deme (deme nextone the <0.03, as 2) aslong Males 1 1 c = 0.05, =0.05, w w Pip11 andPip11 w Pip31_B Pip31_B Pip11 and Pip11 m w ) below which the the which ) below 0 Pip11 will not beableto not will Pip11 w Pip11 cytotype into into Pip11 cytotype w

Pip31 strains. w Pip31_B males Pip31_B w Pip31_U and and Pip31_U w Pip31_U or Pip31_U w Pip31_B w Pip11 Pip11 m

Accepted Article Table 2. Frequenciesof the This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This between crossing relationships the through analyzing made were Estimations (i.e. localities with are as inlocalities Table S1 andin Figure are located five 1. localities The in s #37; Sokra, line isofemale Sok(from the and localities five in collected samples isolated from lines isofemale from 3. Table distribution. binomial from of number of mosquitoes ex number total S1). n, an June(#57) were in collected experiments G3 0.98 0.02 0.97 (70) 0.03 (2) 1 (76) 0 (0) 0 (76) 1 (2) 0.03 (70) 0.97 - - 0.02 0.98 - G3 - G2 0.86 0.14 - G1 0.82 0.67 G0 0.50 0.33 0.50 - 0.16 - - (75) - Cage - Expected s (16) Generation 1 frequencies Cage 2 Cage 3 62. Khetmine 1 (40) 39 38 1 0.03 (0, 0.13) 0.13) (0, 0.37) (0.08, 0.03 (0.05, 0.39) 0.20 0.22) (0.04, 0.17 0.25) (0.03, 1 0.11 7 0.11 4 38 6 169220.12 28 4 (0.07, 19 39 191 35 50 23 34 (40) 1 (40) 0.90 0.17) Total 56 (66) 0.98 Khetmine 62. 38 Rajel 57. Ras Tounga (84) 0.96 (56) 1 Aïn 51. Manar El 46. Hamra 29. Localities

Population showing cages the invasive capacity of the w Pip31 isofemale lines used in crosses.lines isofemalePip31 used Confidence intervals ( w Pip11 Pip11 Frequency Frequency of of w w Pip31 > frequency U w (n) (n) Pip31 Pip31_ w Pip31_B cytotype in Tunisia. Tunisia. in cytotype Pip31_B N rqece n w frequencies (n) Observed Pip11 amined to estimate the frequency of the toestimate amined w Pip31_U ee details on Table S6). preceding the Numbers U w w d October 2010 (#29, #46, #51 and #62) (Table (Table and#62) #46, #51 (#29, 2010 October d Pip31_ Pip11 frequency). usedincrossing Pip11 Samples w Pip31_B (80) 0.10 0.90 (8) w Pip11 U w w Pip31_ Frequency Frequency w Pip11 strain. Pip31_B of of c.i w (74) 0.03 0.97 (2) (75) 0.07 0.93 (6) 1 w .) were calculated Pip11 females of Pip11 of females Pip1 w w Pip31; N, the Pip31; w Pip31 males Pip31 males 95% Pip31 area Pip31 area U w Pip31_ c.i.

Accepted Article frequencies of infertile egg rafts. rafts. egg infertile of frequencies CI andrandom mating. * indicates significant between observeddifference expected and ( ( Sok the cageline up was (n=100) from set (n=100) and females Each using males of of the from females and males of number equivalent the 100) from (n= and females using =100) males up (n was set w between the were performed Confrontations This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This cages. population individuals the frequencyanalyzedPCR assays measureof to by of number Table 4 Table frequencies is in indicated (n). inbrackets infections cages w w Sok × Kef1-2 Sok ×Kef1-2 Sok ×Kef1-1 ouain Cage Population Pip31_U strain (from the Har line) introduced into the same nuclear background. Each cage background. cage nuclear into same Each the introduced Har line) the (from strain Pip31_U Pip31_U). The expected frequency of infertile egg rafts was 0.25, was 0.25, assuming ofinfertile egg expected rafts frequency unidirectional The Pip31_U). lines the of from number orKef1-2 males Kef1-1 females and equivalent withan Pip11) w Pip11 and Pip11 . Mating preferences between between preferences Mating w Pip31_U strains were estimatedPip31_U strains random The mating. cost assuming no and

3 74 0.22 0.22 3 74 0.22 2 102 0.23 1 142 0.34 3 110 0.10 (16) 2 79 0.24 (22) 1 112 (32) 0.85 0.74 (38) 0.78 (8) 0.18 (27) 0.02* 0.99 collected eggrafts Number of of Number a Fisher exact test. test. exact Fisher w Pip11 and w Observed frequency frequency Observed of CI egg rafts (n) egg CI rafts of Pip11 strain (from the Tn line) and the and Tn line) the (from Pip11 strain w Pip31_U line. The Pip31_U The expected frequencies line. w Pip31 infected mosquitoes in in mosquitoes infected Pip31 w P Pip11 line with Pip11 line an -value a

Wolbachia

Infection status of of status Infection w Accepted Articlew Pip11 > > Pip11 Pip31 > > Pip31 Table 5. Table This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This locality. each for indicted is (infertile) not or (fertile) eggrafts the hatched and egg collected, whether of thenumber rafts mosquitoes, the frequency of status, sampling, the of infection year The localities w w w Pip31 Pip31 Pip11 w w Pip31 Pip11

Status of egg rafts collected in collected rafts egg of Status

30. Zerga 2009 47 0.02 252 245 7 7 245 252 0.02 47 2009 Zerga 30. Melah 39. Oued 29. Hamra 2010 41 0 285 284 1 1 1 284 0 136 285 157 17 137 15 157 573 0 257 0 590 41 272 1 2010 23 22 2009 0.09 Hamra 29. 0.1 2010 33 Dougga 41. 59 2010 Jedaida 59. 2010 Mjez Font 69. El Manar 46. Localities sampling Year of 094 .3 4 27 18 227 245 0.93 42 2009 Culex pipiens

N Frequency Frequency of of w Pip11 field populations in Tunisia. in Tunisia. populations field w Number of of Number Pip11 estimated byon N Pip11 estimated PCR examined examined egg rafts Fertile Infertile Accepted Article This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This

Accepted Article This article is protected by copyright. All rights reserved. reserved. rights All byprotected copyright. is article This

Supplementary Information

Stable coexistence of incompatible Wolbachia along a narrow contact zone in mosquito field populations

Célestine M. Atyame, Pierrick Labbé, François Rousset, Marwa Beji, Patrick Makoundou,

Olivier Duron, Emilie Dumas, Nicole Pasteur, Ali Bouattour, Philippe Fort and Mylène Weill

1

Table S1 wPip infections Isofemale lines Country Localities Coordinates Year Frequency of N wPip11 wPip31 95% c.i. Abbreviation Number wPip11 Algeria 1. G5 - 1997 12 0 12 0 (0.26)

2. Tafna - 2006 5 0 5 0 (0.52)

3. Harash - 2006 24 0 24 0 (0.14) Har 1

4. Smar - 2006 5 0 5 0 (0.52)

5. Constantine - 2006 5 0 5 0 (0.52)

6. Guelma - 2008 30 0 30 0 (0.12) Guel 2

7. Lac - 2008 24 0 24 0 (0.14) Lac 2

8. Douas - 2008 27 0 27 0 (0.13) Dou 2

9. Kala - 2008 28 0 28 0 (0.12) Kal 2

10. Souk Ahras - 2008 30 0 30 0 (0.12) Souk 2

Tunisia 11. Bled - 1996 6 6 0 1 (0.54)

12. Gourbi - 1996 16 16 0 1 (0.79)

13. Menzel - 1996 12 12 0 1 (0.74)

14. Mateur N37 02.814 E9 38.738 1996 12 0 12 0 (0.26)

2008 21 0 21 0 (0.16)

2010 23 0 23 0 (0.15)

15. Bizerte N37 16.756 E9 50.528 1996 12 0 12 0 (0.26)

2010 24 0 24 0 (0.14)

16. Tunis - 1997 12 12 0 1 (0.74) Tn 2

17. Bismuth - 2003 17 17 0 1 (0.80)

18. Douz - 2003 2 2 0 1 (0.16)

19. Mornag N36 42.524 E10 16.101 2005 23 23 0 1 (0.85)

20. Grombalia N36 35.454 E10 29.436 2005 23 23 0 1 (0.85)

21. Kondar N35 56.131 E10 18.052 2005 23 23 0 1 (0.85)

2

Table S1 continued wPip infections Isofemale lines Country Localities Coordinates Year Frequency of N wPip11 wPip31 95% c.i. Abbreviation Number wPip11 Tunisia 22. Makroud N35 46.132 E10 07.593 2005 24 24 0 1 (0.86) 23. Msaken N35 43.575 E10 35.946 2005 18 18 0 1 (0.81)

24. Ayed N35 37.586 E10 54.776 2005 24 24 0 1 (0.86)

25. Sousse N35 50.503 E10 37.155 2005 10 10 0 1 (0.69) Sou 2

26. Mornaguia N36 45.805 E10 00.568 2005 23 23 0 1 (0.85)

27. Riadh N36 50.008 E9 58.910 2005 24 24 0 1 (0.86)

2009 24 23 1 0.96 (0.79, 0.99)

2010 47 47 0 1 (0.92)

28. Othman N37 01.186 E9 51.692 2005 17 17 0 1 (0.80)

29. Hamra N37 02.668 E9 39.721 2005 37 0 37 0 (0.09)

2009 33 7 26 0.21 (0.09, 0.39)

2010 47 0 47 0 (0.08)

2010 56 0 56 0 (0.06) Ham 56 (oct) 2011 23 0 23 0 (0.15)

30. Zerga N36 40.143 E9 26.056 2005 47 3 44 0.06 (0.01, 0.18)

2009 47 1 46 0.02 (0, 0.11) Zer 4

2011 24 1 23 0.04 (0, 0.21)

31. Tabarka N36 56.472 E8 45.459 2008 24 0 24 0 (0.14) Tab 2

2010 16 0 16 0 (0.21)

32. Nefza N36 59.207 E9 04.733 2008 8 0 8 0 (0.37)

33. Souala N36 31.604 E8 44.839 2008 5 0 5 0 (0.52)

34. Kef1 N36 04.054 E8 45.171 2008 17 0 17 0 (0.19) Kef 2

35. Kef2 N36 09.573 E8 43.118 2008 30 0 30 0 (0.12)

3

Table S1 continued

wPip infections Isofemale lines Country Localities Coordinates Year Frequency of N wPip11 wPip31 95% c.i. Abbreviation Number wPip11 Tunisia 36. Ariana N36 54.111 E10 13.100 2008 29 29 0 1 (0.88) 2009 24 24 0 1 (0.86)

2011 24 24 0 1 (0.86)

37. Sokra N36 54.123 E10 13.099 2008 14 14 0 1 (0.77) Sok 1

38. Briss N37 02.230 E9 53.887 2008 47 0 47 0 (0.08)

39. Oued Melah N36 35.032 E9 30.712 2008 45 43 2 0.96 (0.85, 0.99)

2009 42 39 3 0.93 (0.81, 0.99)

2010 46 46 0 1 (0.92)

2011 22 22 0 1 (0.85)

40. Slouguia N36 32.419 E9 24.167 2008 24 24 0 1 (0.86)

2009 46 43 3 0.93 (0.82, 0.99)

41. Dougga N36 23.769 E9 14.307 2008 36 0 36 0 (0.09)

2009 23 0 23 0 (0.15)

2010 38 1 37 0.03 (0, 0.14)

2011 17 0 17 0 (0.19)

42. Boussalem1 N36 34.841 E8 59.096 2008 22 0 22 0 (0.15) Bou 2

43. El Battan N36 48.450 E9 53.828 2009 24 23 1 0.96 (0.79, 0.99)

2010 47 47 0 1 (0.92)

2010 20 20 0 1 (0.83) ElB 19 (oct) 2011 24 24 0 1 (0.86)

44. Œufs N36 52.738 E9 51.876 2009 30 29 1 0.97 (0.83, 0.99)

45. Tebourba N36 52.738 E9 51.877 2009 24 24 0 1 (0.86)

2010 24 24 0 1 (0.86)

2011 21 21 0 1 (0.84)

4

Table S1 continued

wPip infections Isofemale lines Country Localities Coordinates Year Frequency of N wPip11 wPip31 95% c.i. Abbreviation Number wPip11 Tunisia 46. El Manar N37 01.770 E9 52.207 2009 45 5 40 0.11 (0.04, 0.24) 2010 61 6 55 0.10 (0.04, 0.20)

2010 (oct) 84 3 81 0.04 (0, 0.10) ElM 39

2011 24 1 23 0.04 (0, 0.21)

47. Goubellat N36 31.847 E9 40.127 2009 24 24 0 1 (0.86)

2011 24 24 0 1 (0.86)

48. Chardon N36 36.098 E9 38.441 2009 24 24 0 1 (0.86)

49. Mjez el Bab N36 39.484 E9 36.333 2009 23 21 2 0.91 (0.72, 0.99)

2010 (oct) 27 26 1 0.96 (0.81, 0.99)

2011 24 24 0 1 (0.86)

50. Fontaine N36 32.417 E9 24.161 2009 31 7 24 0.23 (0.09, 0.41)

2010 55 35 20 0.64 (0.49, 0.76)

2010 (oct) 103 71 32 0.69 (0.59, 0.78)

2011 90 52 38 0.58 (0.47, 0.68)

51. Aïn Tounga N36 31.361 E9 21.338 2009 31 9 22 0.29 (0.14, 0.48)

2010 (oct) 40 4 36 0.10 (0.03, 0.24) AïnT 43

52. Buses N36 22.333 E9 24.924 2009 15 15 0 1 (0.78)

53. Gaafour N36 20.151 E9 20.343 2009 8 0 8 0 (0.37)

2010 19 1 18 0.05 (0, 0.26)

54. Béja Gare N36 43.441 E9 11.437 2009 19 5 14 0.26 (0.09, 0.51)

2010 17 0 17 0 (0.19)

55. Beja Oued N36 43.881 E9 12.315 2009 35 0 35 0 (0.10)

2010 48 0 48 0 (0.07)

2011 17 0 17 0 (0.19)

56. Elevage N36 33.112 E9 00.491 2009 6 0 6 0 (0.46)

5

Table S1 continued

wPip infections Isofemale lines Country Localities Coordinates Year Frequency of N wPip11 wPip31 95% c.i. Abbreviation Number wPip11 Tunisia 57. Ras Rajel N36 57.060 E8 52.784 2010 66 1 65 0.02 (0, 0.08) RasR 23

58. Ftet N36 56.658 E9 06.023 2010 5 0 5 0 (0.52)

59. Jedaida N36 48.229 E9 56.544 2010 24 24 0 1 (0.86)

2011 22 22 0 1 (0.85)

60. Si Thabet N36 57.189 E10 02.589 2010 23 23 0 1 (0.85)

61. Utique N37 04.176 E10 00.424 2010 47 2 45 0.04 (0, 0.15)

2010 (oct) 30 0 30 0 (0.12)

2011 18 0 18 0 (0.19)

62. Khetmine N37 08.910 E9 59.842 2010 19 0 19 0 (0.18)

2010 (oct) 40 0 40 0 (0.09) Khet 39

63. El Alia N37 09.921 E10 02.389 2010 18 0 18 0 (0.19)

64. Ras Jebel N37 13.194 E10 07.017 2010 23 0 23 0 (0.15)

2011 21 0 21 0 (0.16)

65. Azib N37 13.358 E9 56.065 2010 13 0 13 0 (0.25)

2011 21 0 21 0 (0.16)

66. Ichkeul N37 07.225 E9 44.921 2010 23 0 23 0 (0.15)

2011 24 0 24 0 (0.14)

67. Jmar N37 08.198 E9 53.414 2010 23 2 21 0.09 (0.01, 0.28)

68. Pompe N37 04.323 E9 41.457 2010 43 2 41 0.05 (0, 0.16)

69. Font Mjez N36 39.612 E9 28.977 2010 33 3 30 0.09 (0.02, 0.24)

70. TBSK font N36 27.706 E9 14.906 2010 5 0 5 0 (0.52)

71. Boussalem2 N36 36.729 E8 58.594 2010 22 0 22 0 (0.15)

2011 24 0 24 0 (0.14) 72. Utique pont N37 02.201 E10 02.437 2011 23 1 22 0.04 (0, 0.22)

73. Nofrancaoui N36 25.457 E9 19.562 2011 14 4 10 0.29 (0.08, 0.58)

6

Table S1. Distribution of the molecular Wolbachia strains wPip11 and wPip31 infecting Culex pipiens field samples from Algeria and

Tunisia. Localities are numbered as in Figure 1. Localities in bold are those where wPip11 and wPip31 infected mosquitoes were sympatric (i.e. observed in a same breeding site); those underlined, i.e. Hamra (#29), Fontaine (#50) and Béja Gare (#54) displayed significant changes of wPip infections with time. Samples G5 (#1), Bled (#11), Gourbi (#12), Menzel (#13), Tunis (#16), Bismuth (#17), Douz (#18) are from Duron et al.

(2006b). N, sample size. Confidence intervals (c.i.) were calculated from binomial distribution; when frequencies of the wPip11 strain equal 0 or

1, the upper or the lower values are given respectively.

7

Organism Gene Putative product Primer (5'-3') Size (bp) Reference

Wolbachia MutL DNA mismatch repair protein F2- GCATAYCCTAGAGGATGATCCGC 374-437 Atyame et al. (2011a) R2- GTGCATCCAAATAAATCGGAAG ank2 Ankyrin domain protein F-CTTCTTCTGTGAGTGTACGT 313-511 Duron et al. (2007) R2-TCCATATCGATCTACTGCGT Culex pipiens mitochondrial cytb Cytochrome b F-CTTTATTAGTAACTGTAAAAATTAC 852 Atyame et al. (2011b) R-ACTAAAGGATTAGCAGGAATGA

nuclear CQ11 microsatellite locus CQ11F2-GATCCTAGCAAGCGAGAAC CQ11F2-molCQ11R : 250 Bahnck et al. (2006) molCQ11R 5'-CCCTCCAGTAAGGTATCAAC CQ11F2-pipCQ11R : 200 pipCQ11R 5'-CATGTTGAGCTTCGGTGAA

Table S2. List of primers and characteristics of genes used to examine the Wolbachia and Culex pipiens polymorphisms.

References Atyame CM, Delsuc F, Pasteur N, Weill M, Duron O (2011a) Diversification of Wolbachia endosymbiont in the Culex pipiens mosquito. Molecular Biology and Evolution, 28, 2761-2772. Atyame CM, Duron O, Tortosa P, Pasteur N, Fort P et al. (2011b) Multiple Wolbachia determinants control the evolution of cytoplasmic incompatibilities in Culex pipiens mosquito populations. Molecular Ecology, 20, 286-298. Bahnck CM, Fonseca DM (2006) Rapid assay to identify the two genetic forms of Culex (Culex) pipiens L. (Diptera: Culicidae) and hybrid populations. American Journal of Tropical Medicine and Hygiene, 75, 251-255. Duron O, Boureux A, Echaubard P, Berthomieu A, Berticat C et al. (2007) Variability and expression of ankyrin domain genes in Wolbachia variants infecting the mosquito Culex pipiens. Journal of Bacteriology, 189, 4442-4448.

8

A

wPip11 males a a a a a wPip31 females Sou1 Sou2 Tn1 Tn2 Sok a Bou1 IC (21) IC (12) IC (12) IC (17) IC (14) a Bou2 IC (9) IC (17) IC (8) IC (8) IC (14) a Kef1-1 IC (15) IC (11) IC (16) IC (14) IC (18) a Kef1-2 IC (15) IC (17) IC (17) IC (12) IC (11) a Tab1 IC (19) IC (11) IC (16) IC (15) IC (15) a Tab2 IC (30) IC (8) IC (18) IC (12) IC (11) b Dou1 IC (12) IC (7) IC (22) IC (12) IC (23) b Dou2 IC (15) IC (10) IC (10) IC (12) IC (21) b Guel1 IC (13) IC (10) IC (12) IC (15) IC (12) b Guel2 IC (21) IC (20) IC (19) IC (21) IC (24) b Kal1 IC (12) IC (12) IC (13) IC (9) IC (15) b Kal2 IC (14) IC (13) IC (12) IC (19) IC (17) b Lac1 IC (13) IC (12) IC (16) IC (12) IC (15) b Lac2 IC (13) IC (20) IC (9) IC (12) IC (13) b Souk1 IC (16) IC (14) IC (13) IC (23) IC (14) b Souk2 IC (11) IC (15) IC (12) IC (22) IC (10) B

wPip31 males wPip11 females a a a a a a b b b b b b b b b b Bou1 Bou2 Kef1-1 Kef1-2 Tab1 Tab2 Dou1 Dou2 Guel1 Guel2 Kal1 Kal2 Lac1 Lac2 Souk1 Souk2 a Sou1 C (30) C (16) C (9) C (13) C (13) C (18) C (9) C (17) C (23) C (18) C (14) C (11) C (21) C (15) IC (24) IC (21) a Sou2 C (18) C (20) C (18) C (8) C (22) C (18) C (16) C (17) C (14) C (22) C (21) C (29) C (17) C (19) IC (20) IC (16) a Tn1 C (22) C (16) C (19) C (25) C (35) C (24) C (23) C (19) C (22) C (18) C (14) C (20) C (22) C (12) IC (19) IC (24) a Tn2 C (21) C (21) C (18) C (23) C (22) C (18) C (18) C (15) C (18) C (18) C (15) C (26) C (20) C (15) IC (23) IC (16) a Sok C (16) C (18) C (14) C (12) - C (12) C (16) C 12) C (15) C (12) C (12) C (14) C (23) C (17) IC (14) IC (20)

9

Table S3. Reciprocal crosses between wPip11 and wPip31 infected isofemale lines from allopatric localities. A, wPip11 males and wPip31 females. B, wPip31 males and wPip11 females. The cross is compatible (C), for a HR > 90%; and incompatible (IC) for 0% HR. Incompatible crosses are shaded and bidirectional CI are underlined. The number of egg rafts collected in each cross is in parentheses. The name of mosquito line followed by number, for instance Sou1, Sou2 means that isofemale lines are from the same locality (see Table S1). a, Isofemale lines from

Tunisia; b, Isofemale lines from Algeria. Dashes indicate that the cross was not performed.

10

A

Males wPip11 wPip31 Females Zer1 Zer2 Zer3 Zer4 wPip11 Zer1 - - C (21) C (12) Zer2 - - C (18) C (26)

wPip31 Zer3 IC (9) IC (11) - - Zer4 IC (15) IC (16) - -

B

Males wPip11 wPip31 Females AïnT1 AïnT2 AïnT3 AïnT4 wPip11 AïnT1 - C (12) C (20) IC (26) AïnT2 C (25) - C (20) IC (21)

wPip31 AïnT3 IC (16) IC (20) - C (18) AïnT4 IC (9) IC (16) C (17) -

Table S4. Reciprocal crosses between sympatric wPip11 and wPip31 infected mosquitoes. A, from Zerga (#30). B, from Aïn Tounga (#51).

The cross is compatible (C), for a HR > 90%; and incompatible (IC) for HR = 0%. Incompatible crosses are shaded and bidirectional CI are underlined. The number of egg rafts collected in each cross is in parentheses.

11

wPip11 males from El Battan wPip11 males from El Manar wPip11 males from Aïn Tounga × × × wPip31 females from Har wPip31 females from Har wPip31 females from Har ElB1 IC (13) ElM39 IC (12) AïnT40 IC (16) ElB2 IC (12) AïnT41 IC (15) ElB3 IC (15) AïnT42 IC (18) ElB4 IC (14) AïnT43 IC (14) ElB5 IC (16) ElB6 IC (15) ElB7 IC (21) ElB8 IC (15) ElB9 IC (14) ElB10 IC (12) ElB11 IC (14) ElB12 IC (17) ElB13 IC (14) ElB14 IC (12) ElB15 IC (16) ElB16 IC (13) ElB17 IC (12) ElB18 IC (14)

Table S5. Crossing relationships of wPip11 males. This experiment was performed by crossing wPip11 isofemale-derived males from three localities (El Battan, #43; El Manar, #46 and Aïn Tounga, #51) and wPip31 females from the same isofemale line Har from Harash

(#3). All crosses were incompatible (IC) with HR = 0%. The number of egg rafts collected in each cross is in parentheses.

12 wPip31 males from wPip31 males from wPip31 males from wPip31 males from wPip31 males from Hamra El Manar Aïn Tounga Ras Rajel Khetmine × × × × × wPip11 females wPip11 females from wPip11 females from wPip11 females from wPip11 females from from Sok Sok Sok Sok Sok Ham1 C (20) ElM1 C (17) AïnT5 IC (14) RasR1 IC (18) Khet1 C (17) Ham2 C (19) ElM2 C (22) AïnT6 C (17) RasR2 C (12) Khet2 C (16) Ham3 C (19) ElM3 IC (15) AïnT7 IC (17) RasR3 C (15) Khet3 C (14) Ham4 C (15) ElM4 C (20) AïnT8 C (18) RasR4 C (22) Khet4 IC (16) Ham5 C (18) ElM5 IC (18) AïnT9 IC (12) RasR5 C (5) Khet5 C (18) Ham6 C (14) ElM6 C (23) AïnT10 C (20) RasR6 C (14) Khet6 C (13) Ham7 C (18) ElM7 C (19) AïnT11 C (18) RasR7 C (20) Khet7 C (16) Ham8 C (15) ElM8 C (18) AïnT12 IC (17) RasR8 C (19) Khet8 C (13) Ham9 C (17) ElM9 C (18) AïnT13 C (15) RasR9 C (14) Khet9 C (13) Ham10 C (18) ElM10 C (18) AïnT14 C (24) RasR10 C (11) Khet10 C (13) Ham11 C (11) ElM11 C (20) AïnT15 C (21) RasR11 C (20) Khet11 C (15) Ham12 C (16) ElM12 C (22) AïnT16 C (22) RasR12 IC (21) Khet12 C (21) Ham13 C (24) ElM13 C (22) AïnT17 C (13) RasR13 C (20) Khet13 C (24) Ham14 IC (18) ElM14 C (24) AïnT18 C (22) RasR14 C (20) Khet14 C (16) Ham15 C (19) ElM15 C (18) AïnT19 C (18) RasR15 C (15) Khet15 C (21) Ham16 C (21) ElM16 C (19) AïnT20 C (16) RasR16 IC (15) Khet16 C (18) Ham17 C (19) ElM17 C (20) AïnT21 C (20) RasR17 C (20) Khet17 C (18) Ham18 C (13) ElM18 C (15) AïnT22 C (19) RasR18 C (21) Khet18 C (19) Ham19 C (19) ElM19 C (16) AïnT23 C (14) RasR19 C (14) Khet19 C (16) Ham20 C (21) ElM20 C (16) AïnT24 C (17) RasR20 IC (20) Khet20 C (12) Ham21 C (14) ElM21 C (19) AïnT25 C (23) RasR21 C (18) Khet21 C (17) Ham22 C (18) ElM22 IC (19) AïnT26 C (19) RasR22 C (7) Khet22 C (13) Ham23 C (15) ElM23 C (16) AïnT27 C (23) RasR23 C (19) Khet23 C (13) Ham24 C (18) ElM24 IC (21) AïnT28 C (22) Khet24 C (19) Ham25 C (17) ElM25 C (20) AïnT29 C (19) Khet25 C (19) Ham26 C (21) ElM26 C (17) AïnT30 C (18) Khet26 C (18) Ham27 C (18) ElM27 C (23) AïnT31 C (22) Khet27 C (16) Ham28 C (18) ElM28 C (12) AïnT32 C (17) Khet28 C (20) Ham29 C (14) ElM29 C (14) AïnT33 IC (22) Khet29 C (15) Ham30 C (11) ElM30 C (10) AïnT34 C (20) Khet30 C (13) Ham31 C (18) ElM31 C (19) AïnT35 IC (18) Khet31 C (13) Ham32 IC (16) ElM32 C (17) AïnT36 IC (15) Khet32 C (23) Ham33 C (19) ElM33 C (24) AïnT37 C (17) Khet33 C (13) Ham34 C (16) ElM34 C (12) AïnT38 C (18) Khet34 C (17) Ham35 C (17) ElM35 C (12) AïnT39 C (15) Khet35 C (21) Ham36 IC (20) ElM36 C (19) Khet36 C (20) Ham37 C (13) ElM37 C (23) Khet37 C (16) Ham38 C (15) ElM38 C (12) Khet38 C (22) Ham39 IC (18) Khet39 C (20) Ham40 C (16) Ham41 C (13) Ham42 C (17) Ham43 C (19) Ham44 C (13) Ham45 IC (15) Ham46 IC (18) Ham47 C (15) Ham48 C (21) Ham49 C (17) Ham50 C (16) Ham51 C (16) Ham52 C (16) Ham53 C (13) Ham54 C (14) Ham55 C (16) Ham56 C (19)

13

Table S6. Estimation of the frequencies of wPip31_U and wPip31_B cytotypes among wPip31 infected mosquitoes. This experiment was performed through unidirectional crosses between wPip31 isofemale-derived males from fives localities (Hamra, #29; El Manar, #46;

Aïn Tounga, #51; Ras Rajel, #57 and Khetmine, #62) and wPip11 females from the same line

Sok from Sokra (#37). The cross is compatible (C), for a HR > 90%; and incompatible (IC) for 0% HR. Incompatible crosses are shaded. The number of egg rafts collected in each cross is in parentheses.

14

Infection Year of Localities N pipiens form molestus form Hybrids status sampling wPip11 Ayed (#24) 2005 10 3 5 2 Riadh (#27) 2010 10 3 4 3

wPip31 Tabarka (#31) 2010 10 5 5 0 Boussalem2 (#71) 2010 10 3 2 5

Table S7. Occurrence of pipiens and molestus forms of Culex pipiens taxa among wPip11 and wPip31 infected mosquitoes in Tunisia, identified by the CQ11 microsatellite locus.

15

Pip11 Pip31 w w 511 pb

313 pb ank2

437 pb 374 pb MutL

Figure S1. Patterns of ank2 and MutL PCR markers in wPip11 and wPip31 molecular strains.

16

Figure S2. Effect of dispersal probability on the progression of the wPip11 cytotype into increasingly distant demes, with or without assortative mating. The maximal dispersal probability (m) below which the wPip11 cytotype is prevented from invading a given deme

(i.e. from reaching a frequency of 50% of the population) after 50 generations is indicated for different fitness costs associated to the wPip11 cytotype (c) and for different levels of correlation of cytotypes between mates (F). The frequency of the wPip31_B cytotype among all wPip31 individuals was fixed as b = 0.12. For example, with F = 0.39 and c = 0.2, wPip11 will not be able to invade the closest deme (deme 1) as long as m < 0.09, the next one

(deme 2) as long as m < 0.13, the next one (deme 3) as long as m < 0.2, etc... Note that for c =

0.2 and F = 0.63, no value of m allowed wPip11 to invade even the closest deme after

50 generations.

17