Detection of Wolbachia in the Tick Ixodes ricinus is Due to the Presence of the Hymenoptera Endoparasitoid Ixodiphagus hookeri

The identification of micro-organisms carried by ticks is an important issue for human and animal health. In addition to their role as pathogen vectors, ticks are also the hosts for symbiotic bacteria whose impact on tick biology is poorly known. Among these, the bacterium Wolbachia pipientis has already been reported associated with Ixodes ricinus and other tick species. However, the origins of Wolbachia in ticks and their consequences on tick biology (known to be very diverse in invertebrates, ranging from nutritional symbionts in nematodes to reproductive manipulators in insects) are unknown. Here we report that the endoparasitoid wasp Ixodiphagus hookeri (Hymenoptera, Chalcidoidea, Encyrtidae) – strictly associated with ticks for their development - is infested at almost 100% prevalence by a W. pipientis strain belonging to a Wolbachia supergroup that has already been reported as associated with other hymenopteran parasitoids. In a natural population of I. ricinus that suffers high parasitism rates due to I. hookeri, we used specific PCR primers for both hymenopteran and W. pipientis gene fragments to show that all unfed tick nymphs parasitized by I. hookeri also harbored Wolbachia, while unparasitized ticks were Wolbachia-free. We demonstrated experimentally that unfed nymphs obtained from larvae exposed to I. hookeri while gorging on their vertebrate host also harbor Wolbachia. We hypothesize that previous studies that have reported W. pipientis in ticks are due to the cryptic presence of the endoparasitoid wasp I. hookeri. This association has remained hidden until now because parasitoids within ticks cannot be detected until engorgement of the nymphs brings the wasp eggs out of diapause. Finally, we discuss the consequences of this finding for our understanding of the tick microbiome, and their possible role in horizontal gene transfer among pathogenic and symbiotic bacteria.


Introduction
As ticks are considered to be second only to mosquitoes as vectors of human infectious diseases in the world, the identification of micro-organisms carried by ticks is an important issue for human or animal health [1]. Among arthropod vectors, ticks harbor the largest diversity of micro-organisms, ranging from viruses (i.e. arboviruses such as Tick Borne Encephalitis…), prokaryotes (including spirochaetes such as Borrelia burgdorferi) and eukaryotes (including Apicomplexa -such as Babesia spp. or Theileria spp. -or even metazoans -such as some filarial nematodes -) [2]. This large diversity of micro-organisms may be related to the fact that all three stages (larvae, nymphs and adults) take a blood meal over several days, ingesting a large volume of blood [3], and that a wide range of different vertebrate host species is used for the blood meal (at least for the most common species in Europe or in North-America -i.e. Ixodes ricinus and I. scapularis respectively). Moreover, because of the long life cycle of ticks (usually 2 to 3 years for species living in temperate areas), these arthropods are reservoirs for pathogens (especially for vertically transmitted pathogens, i.e. trans-ovarial -transmission)in addition to vertebrate primary hosts, facilitating micro-organism persistence through unfavorable seasons [4].
However, ticks do not only harbor pathogens; they also host symbiotic bacteria whose effects on tick fitness and on vertebrate health are poorly known. For example, among Rickettsia spp., several species that are commonly found associated with ticks have never been found associated with disease [5,6]. While Coxiella burnetti is known as the etiological agent of the Q-Fever responsible for recent outbreaks in Europe, an unnamed Coxiella species has been found in 100% of the individuals of Amblyomma americanum [7]. Antibiotic treatment of these ticks resulted in reduced reproductive fitness, suggesting that the Coxiella sp. is a primary endosymbiont [8]. Midichloria mitochondrii is an endosymbiotic alpha-proteobacterium found in the ovarian tissues of almost all Ixodes ricinus females from natural populations [9,10]. This bacterium has a unique intramitochondrial lifestyle [11,12,13] but its consequences on tick fitness or reproduction remain unknown [13]. More recently, an obligate intracellular gammaproteobacterium (Diplorickettsia massiliensis) has been isolated from I. ricinus, again with unknown consequences for its host tick [14]. The recent development of metagenomic approaches on ticks will probably allow the identification of numerous new bacteria species [15,16].
Even if some of those micro-organisms are not pathogenic for vertebrate hosts, their presence within ticks may have some consequences for the epidemiology of some tick-borne diseases either by acting on vector fitness or through synergic or antagonistic interactions with known pathogens. Recent studies have highlighted the fact that species interactions in a parasite community drive infection risk [17] or that competitive exclusion may be detrimental to one pathogen species in case of coinfections [18]. In the case of blood-feeding insects, symbiotic bacteria having a role in the nutrition of their host have been reported in an increasing number of cases [19,20]. Moreover, some bacterial symbionts can have direct consequences on the transmission of some pathogens by vectors. For example, the refractoriness of some Anopheles gambiae populations from Zambia in transmission of Plasmodium falciparum is due to the presence of an Enterobacter species in the mosquito's gut [21]. In the mosquito species Aedes albopictus, it has been demonstrated that the replication of the Chikungunya virus is modulated by the presence of endosymbiotic Wolbachia [22]. Wolbachia is also known to induce resistance to dengue virus in Aedes aegypti [23]. Wolbachia are now also known to interfere with the innate immune system of their arthropod host, hence conferring virus protection in some cases [24].
Wolbachia are endosymbiotic bacteria manipulating host reproductive systems (through cytoplasmic incompatibility or parthenogenesis for example) or as mutualists when they are associated with nematodes [25]. The presence of Wolbachia may have direct consequences on the development of other pathogens in the same arthropod vector, and also indirect effects on pathogen epidemiology through impacts on the dynamics and genetic diversity of the vector [26][27]. Thanks to the development of molecular tools to look for microorganisms within ticks, Wolbachia have been reported from ticks in an increasing number of studies (reviewed in table 1). However, their frequency is usually low (but see [28] where W. pipientis reach 7.6% in nymphs) and some studies designed to look for Wolbachia in ticks have failed to find any [29]. The consequences of the presence of Wolbachia in those ticks are unknown.
The purpose of this study was to look for Wolbachia in natural populations of I. ricinus and to identify their origin within ticks. We discovered that natural populations of the wasp Ixodiphagus hookeri (Hymenoptera, Chalcidoidea, Encyrtidae), an endoparasitoid of Ixodes ricinus, ,are infected at almost 100% prevalence by Wolbachia pipientis. Using PCR primers specific to I. hookeri and W. pipientis respectively, we showed that in natural tick populations, all individual ticks harboring Wolbachia were also parasitized by I. hookeri, while ticks that were not parasitized were Wolbachia-free. We showed experimentally that some unfed nymphs, exposed to the parasitoids while they were larvae gorging on their vertebrate host, also harbored Wolbachia. We hypothesize that the presence of Wolbachia, already reported in several studies using 16S PCR primers to identify bacteria within ticks, is due to the presence of Ixodiphagus in those ticks, which is externally cryptic (as the parasitized tick nymphs looks healthy until their engorgement that bring the parasitoid eggs out of their diapause). Finally, we discuss the consequences of these findings for our understanding of tickmicrobiota and possible consequences for tick-borne diseases epidemiology.

Results
The parasitoid Ixodiphagus hookeri harbors endosymbiotic Wolbachia pipientis Based on both (i) engorged nymphs collected on roe-deer and (ii) host-seeking nymphs collected on vegetation that were subsequently engorged in the lab, a total of 54 ticks collected in 4 natural populations were found to be parasitized by I. hookeri (Table 2; Figure 1). Variation of parasitism rates among years and sites range from 3,2% to 12,5% in Chizé (tick nymphs collected on roe deer), and from 19,6% to 20% in Gardouch (tick nymphs collected on roe-deer and on vegetation respectively). Among the I. hookeri individuals obtained from those 54 ticks, a total of 121 I. hookeri individuals (74 females and 47 males) were screened for Wolbachia pipientis by PCR using Wsp primers. Wolbachia DNA was found in 120 (99,2%) of those individuals (a single wasp male from the Gardouch population was Wolbachia-free, even after a second PCR trial).
Experimental demonstration that parasitim of I. ricinus by I. hookeri is associated with the presence of Wolbachia in ticks To demonstrate experimentally that I. hookeri is responsible for the presence of Wolbachia in some I. ricinus individuals, we set up a laboratory rearing of this parasitoid species and exposed some laboratory strain ticks to the parasitoids. Under experimental conditions, parasitism rate was extremely reduced when unfed nymphs were exposed to parasitoids (parasitism rate estimated by the engorgement of unfed nymphs exposed to parasitoids was 2,5% -i.e. 3 nymphs produced wasps of 171 exposed to adult parasitoids and subsequently engorged to reveal parasitism). Parasitoids obtained from those experimentally parasitized ticks again harbored Wolbachia as attested by the results of PCR amplification with Wsp primers, hence demonstrating experimentally the vertical transmission of Wolbachia in I. hookeri. Additionnaly, Wsp primers were tested successfully on DNA extracted from groups of about 50 I. hookeri eggs that were isolated after the  dissection of female parasitoid gaster, demonstrating that the parasitoid eggs contain some Wolbachia. Parasitoid rates were higher when nymphs were exposed to parasitoids while they were feeding on gerbils (40,5%; n = 84). However, these engorged nymphs are not the tick stage for which the primers are the most useful (preventing the engorgement of ticks in the lab to determine the parasitism rate), i.e. the unfed nymphs. To produce unfed nymphs in the lab that harbor I. hookeri, we exposed some tick larvae to parasitoids during engorgement on gerbils. Previous studies [30,31] have shown that such tick larvae produce living nymphs that harbor parasitoid eggs and only the engorgement of those parasitized nymphs brings the wasp eggs out of diapause, allowing the parasitoids to emerge from those engorged nymphs. The set of nymphs exposed to parasitoids during engorgement was divided in two. Half were used to estimate the parasitism rate due to I. hookeri. Two of 33 nymphs produced some parasitoids after engorgement, corresponding to a parasitism rate of 6%. The second half was used to look for Wolbachia using the Wsp primers; 8 of the 106 unfed nymphs (5,7%) harbored Wolbachia. Two out of those 8 ticks harboring Wolbachia were also positive with the CO1 primers, confirming that nymphs containing Wolbachia also contained I. hookeri eggs. We interpreted the lack of amplification with the CO1 primers for 6 out of those 8 ticks as due to lower sensitivity of the CO1 primers and the low quantity of I. hookeri DNA expected in a parasitized nymph (containing 5 to 10 eggs). PCR screening with appropriate primers showed none of the 32 unfed nymphs used as a negative control (corresponding to larvae that were not exposed to parasitoids during their engorgement) to contain Wolbachia or I. hookeri DNA.
Wolbachia pipientis associated with I. hookeri belongs to a Wolbachia clade found in hymoneptera or other insects Sequence data for Wsp and ftsZ genes of Wolbachia found in I. hookeri from ticks was used to determine whether these bacteria belong to a clade already associated with other insects (including parasitoids), or instead fall within Wolbachia groupings associated with ticks (Acari) or other Arachnida and Chelicerata. Six Wsp sequences (ranging from 515 to 379 pb) were obtained from PCR products based on I. hookeri individuals from the Gardouch and Chizé populations. The alignment revealed no polymorphism (in the 351 pb common to all 6 sequences or in the other part of the alignment). A consensus sequence of 533 pb was generated (515 pb from the longest sequence, to which 18 pb were added from another sequence) and compared to Wsp sequences already deposited in GenBank with BLAST. Over those 533 pb, this consensus sequences has 100% identity with a Wolbachia associated with I. ricinus (accession nu EF219192. Because the Wsp gene is known to exhibit a high recombination rate [32], we also sequenced the Ftsz gene to identify the Wolbachia supergroup to which the Wolbachia associated with I. hookeri belong. The FtsZ sequence (947 bp long) was obtained from a wasp individual from the Gardouch population. In the phylogenetic tree (Figure 2), this sequence clearly belongs to supergroup A (with only one difference with the sequence found in the Wolbachia associated with other hymenopteran parasitoids such as Spalangia cameroni, Trichogramma bourarachae (Hymenoptera; Chalcidoidea; Trichogrammatidae), Asobara tabida, and the fruitfly Drosophila simulans.

Discussion
Presence of I. hookeri in I. ricinus nymphs is correlated with presence of W. pipientis in ticks By the use of specific PCR primers for the parasitoid I. hookeri and for the endosymbiotic bacteria W. pipientis, we demonstrated that almost all (99,2%) of the I. hookeri individuals investigated, emerging from 5 different natural french populations of the tick I. ricinus, harbored W. pipientis. The analysis of a natural tick population suffering from high parasitism due to I. hookeri (evaluated at 20%, based on the engorgement in the lab of nymphs collected on the vegetation) revealed that all the DNA extracts of nymph ticks showing evidence of the presence of I. hookeri (based on CO1 hymenopteran specific primers) also harbored W. pipientis (based on Wsp Wolbachia specific primers), while all but one of the tick DNA extracts without any evidence of I. hookeri were also Wolbachia-free. We also showed experimentally that unfed nymphs developing from larvae exposed to wasps while gorging on their host, also harbored Wolbachia. Moreover, the Wolbachia sequence of the Wsp and Ftsz genes associated with I. ricinus were identical or very similar to Wolbachia sequences already reported as associated with parasitoid hymenoptera, including other chalcids.
Based on these observations, we suggest that the presence of W. pipientis in I. ricinus, already reported by several authors (but, to date, without any hypothesis to explain the presence of this bacterium in ticks), is due to the parasitism of those ticks by the hymenopteran endoparasitoid I. hookeri. Because of the koinobiontic development of those parasitoids, the presence of the parasitoids in nymphs cannot be detected visually prior to their engorgement (bringing the parasitoid eggs out of diapause). In all of these studies (Table 1), Wolbachia has been identified from the sequencing of 16S rRNA gene amplicons. Ticks were sampled from natural populations (and not laboratory reared) and thus these ticks could have been exposed to I. hookeri parasitoids. Moreover, our hypothesis is corroborated by the fact that most of these studies are based on tick DNA extracts obtained from nymphs [16,33,34], the tick stage most attacked by the parasitoids [28]. Two other studies have reported the presence of W. pipientis from adults of I. ricinus, but at a much lower frequency (i.e 1,0% (n = 98) compared to 27,3% in nymphs (n = 33) in a French population [28]; 0,45% (n = 221) in a Moroccan population; [35]; see also [15] for the report of W. pipientis associated with adults of Rhipicephalus (Boophilus) microplus). Although emergence of I. hookeri has never been reported from engorged I. ricinus females [36] (but oviposition on adults occurs in the lab on adult Rhipicephalus sanguineus [37]), the presence of W. pipientis in a few adult ticks may still be explained by parasitism of nymphs by I.hookeri. Indeed, it is likely that not all ovipositions of I. hookeri result in successful development of parasitoids. In particular, the innate immune system of ticks may be able to kill some parasitoid eggs (by the encapsulation of parasitoid eggs resulting in its melanization and subsequent death) that could prevent the development of parasitoids but explain the presence of W. pipientis DNA in tick adults. The fact that W. pipientis has not been found in some studies that have looked for it in I. ricinus [29,38] can also be explained by the absence of parasitoid in those ticks.
Use of Wolbachia specific primers to estimate parasitism rates due to I. hookeri In subsequent studies, Wolbachia-specific Wsp primers could be used to determine parasitism rates due to I. hookeri in natural populations of I. ricinus using host-seeking nymphs (and thus avoiding the time-consuming engorgement of nymphs in the lab to determine parasitism rates). Because parasitoid diapause occurs at the egg stage [39], the amount of I. hookeri and Wolbachia DNA in unfed I. ricinus nymphs is probably reduced (5 to 10 eggs per ticks considering the mean number of parasitoid adults emerging from an engorged nymph; 6,5 according to [40]; n = 54). Fortunately, the use of mitochondrial primers (the DNA mitochondrial molecules being present in numerous copies in each single egg cell) is more sensitive to report the presence of I. hookeri than nuclear markers. However, the use of Wsp primers seems to be even more sensitive to report the presence of I. hookeri in ticks than the specific mitochondrial CO1 primers (c.f. the few cases where tick extracts were positive with the Wsp and not the CO1 primers). This may be due to a high number of Wolbachia bacteria in I. hookeri eggs (for example, between 116 and 292 bacteria per egg in Nasiona vitripennis [41]. Our results show that parasitism rates due to I. hookeri in natural population of I. ricinus can reach at least 20%, inflicting significant mortality on tick populations. This explains why I. hookeri was considered as a possible biological control agent to reduce tick populations, soon after its original description [42]. The higher parasitism rates observed in Gardouch may be associated with the higher roe deer density (as already observed in the USA for Odocoileus virginiae, Ixodes scapularis and I. hookeri [43]) maintained in a 14 hectare fenced area. In Chizé, the parasitism rates are relatively stable across years (again fenced, but covering a much larger area of 2580 hectares). However, the month of tick sampling (February), constrained by roe-deer capture, may not be representative of the parasitism rates of I. ricinus during other months.

Consequences of the presence of Wolbachia on tick and tick-borne diseases
The phenotypic consequences of Wolbachia infection for their hosts are highly diverse and new impacts are regularly discovered. In filarial nematodes, these bacteria are mutualistic and required for the development and reproduction of their host. In arthropods, they have first been shown to alter their host reproduction through various mechanisms (cytoplasmic incompatibility, parthenogenesis, male-killing, feminization…) but a growing number of studies show that they also have some impacts on the immune response of their host (with some consequences for resistance against parasitoids, or on pathogen agents transmission in the case of arthropod vectors; [44]) or may even be nutritional mutualists in blood feeding bugs [20]. Previous reports of W. pipientis in ticks have only mentioned the known consequences of Wolbachia on nematodes or insects [33] or suggested their possible role as mutualistic endosymbionts of ticks [45] or as pathogenic agents [35]. Our finding that Wolbachia is associated with I. hookeri rather than with I. ricinus per se and that Wolbachia is thus not (or only very Only bootstrap values greater than 80% are shown. Letters A to F (at the right of the name of the host from which the Wolbachia was sequenced and of its accession number) refer to the Wolbachia supergroups already described [58]. doi:10.1371/journal.pone.0030692.g002 rarely) vertically transmitted to tick offspring (otherwise, the prevalence of Wolbachia in ticks would be much higher; anyway, most ticks parasitized by I. hookeri and thus containing W. pipientis are expected to die, once they will engorge) suggests that any consequences of Wolbachia infection on I. ricinus reproduction are without object. Additionnal investigations must be conducted to identify the consequences of Wolbachia on I. hookeri. Although the presence of Wolbachia in ticks may be considered a ''dead-end'' for the bacteria, it may have had some consequences on other microorganisms harbored by ticks. For example, a Wolbachia gene has been identified by Ishmael et al. within a plasmid of a Rickettsia species found in Ixodes scapularis (the ''Rickettsia endosymbionts of I. scapularis'' identified by Nodae et al. [45] and Benson et al. [46] and reported in 100% of the engorged females studied by [47]). The authors suggest that lateral gene transfer between these two obligate intracellular species has occurred. The 44 Wolbachia traces found in the I. scapularis genome [48] were finally due to the presence of these Rickettsia endosymbionts within the I. scapularis trace analysed [49].
Beside W. pipientis, a negative interaction between some pathogen agents (i.e. the Spirochaete Borrelia burgdorferi and the Apicomplexa Babesia microti) and I. hookeri in north American populations of I. scapularis has been reported [50]. Although this study was conducted prior to the development of molecular techniques to detect those micro-organisms within ticks, the authors observed that the prevalence of those human pathogens in host-seeking ticks collected in wasp-infested sites is nearly 40% lower than that found in other sites and that spirochaetes never infected wasp-infected ticks and few wasp-infected ticks were concurently infected by B. microti. The authors considered that ''wasps may render the tick inhospitable to both pathogens. Our finding that I. hookeri harbors W. pipientis could provide an explanation for those observations : the bacteria may induce an immune reaction in the ticks that could have a negative impact on other micro-organism hosted by ticks (see for example [44]).
To conclude, in prevision to the growing number of studies that will use a metagenomic approach to identify micro-organisms within ticks [15], we would like to highlight the fact that ticks are also the hosts of macroparasites (including wasps such as I. hookeri, but also nematodes such as Cercopithafilaria rugosicauda associated with I. ricinus) that can harbor their own microbiome. Thus, the study of tick-borne diseases involves more than a tripartite system involving vertebrates, arthropod vectors and pathogen agents but also involves other organisms, pleading for a community ecology approach on these complex systems.

Tick and insect populations studied
Ixodes ricinus. The tick Ixodes ricinus is the most common european tick species, found mainly in forests, but also hedges or moors. Its 2-4 year lifecycle includes three stages (larva, nymph and adults), with each stage needing to take a blood meal on a vertebrate host (a mammal, bird or lizard) for its development.
After each blood meal, the engorged ticks fall from the host to the ground, where they will metamorphose to the next stage. Questing (host-seeking) ticks will then waiting for a host by staying on the top of the vegetation.
Ixodiphagus hookeri. With 8 species described [52], Ixodiphagus (Hymenoptera, Chalcidoidea, Encyrtidae) is the single parasitoid genus of 80 000 parasitoid species described [53] known to parasitize ticks. In Europe, only I. hookeri has been reported, where it is mainly associated with Ixodes ricinus. Although the lifecycle of this parasitoid is not completely known (especially the stages during winter time; [40]), the parasitoids are only known to emerge from engorged nymphs. The females of I. hookeri oviposit 5-10 eggs into tick nymphs (or eventually in tick larvae). The parasitoid does not kill its host during oviposition (corresponding to the koinobiontic strategy) and the tick nymphs look healthy until their engorgement. Once the tick is engorged, the diapause of the wasp eggs is broken and the wasp larvae begin their development within the engorged tick. The tick will finally die and the adult parasitoids dig a hole at the posterior end of the tick to emerge, 1-2 months after oviposition. The adult wasps then mate and the females look for ticks to lay their eggs before to die 2 or 3 days after emergence.
Estimation of the parasitism rates due to I. hookeri in natural populations of I. ricinus Parasitism rates due to I. hookeri in natural populations of I. ricinus were estimated both from engorged ticks collected on roedeer and from questing ticks collected on the vegetation. Unfed tick nymphs were engorged on gerbils, after having been laid on the fur with a paintbrush and recovered 3 to 7 days after. All engorged tick nymphs were then placed in individual tubes and kept at a humidity of 80%. Emergence of parasitoids from engorged nymphs was checked weekly. The adults obtained were sexed under a steromicroscope using the key provided in Quaraishi 1958 [54]. Parasitism rates were calculated by taking into account only engorged nymphs producing either adult ticks or parasitoids (engorged nymphs that die without producing adult ticks or wasps were not included).

Exposition of I. ricinus to the parasitism by I. hookeri in laboratory experiments
Unfed nymph ticks from our laboratory rearing (without I. hookeri and shown by Wsp PCR to be Wolbachia-free), were exposed to parasitoids in sets of 25 (2 flasks), 50 (2 flasks)or 100 (3 flasks) individuals in a tissue culture cell flask (25 cm 3 ) with filter screw cap (TRP). Several freshly emerged adult parasitoids (both females and males; from 4 to 10; mean of 7 wasps per flask) were added to each flask for 2 or 3 days until parasitoid death. In other experiments, tick nymphs or larvae were exposed to parasitoids while they were gorging on gerbils. To that aim, several freshly emerged wasps (from 4 to 7, mean of 5,8 per cage) were introduced to the cage containing the gerbil and the ticks that were deposited on the gerbil 4 days before.

DNA extraction
Individual ticks or wasps, killed at 280uC or conserved in alcohol, were put in one well (in 2 ml Deepwell plate; NUNC) with 2 pre-sterilised glass beads of 3 mm diameter and one of 5 mm. After 15 minutes at 280uC, they were mechanically crushed by using the Qiagen Tissue Lyser for 3 minutes at a frequency of 15 Hz. DNA was extracted using the NucleoMag 96 Tissue kit (Macherey-Nagel). DNA was eluted with 50 ml buffer for ticks and 25 ml for wasps and conserved at 4uC until PCR was performed.

PCR amplification and sequencing
To amplify the wsp gene of Wolbachia, we used the wsp-81F and wsp-691R primers [55] amplifying a DNA fragment of 590-632 pb. PCRs were conducted in 20 ml reaction volumes with 7,8 ml H 2 O (EPPI sterile), 4 ml 5X Buffer, 1 ml forward and reverse primer (10 mM), 2 ml dNTP (2mM), 0,2 ml of Phusion High-Fidelity Taq (New England Biolabs) and 4 ml of DNA extract. The amplification program is the following: 98uC for 30 seconds, then 29 cycles at 98uC for 45 seconds, 55uC for 45 seconds and 72uC for 60 seconds and final elongation at 72uC for 10 minutes. The PCR product obtained was sequenced using PCR primers as sequencing primers (accession number : JQ315223).
To design specific primers for the CO1 gene of I. hookeri (that do not amplify I. ricinus DNA), we first used the primers 1775-Co1-F, and 2773-Co1-R, already described for other Chalcidoidea [57]. From the I. hookeri sequence obtained (accession number JQ315225), we designed two new primers 401-F (59-TTTAGAA-TATTTATTGATTCAGGGACT-39) and 44-R (59-CTCCTG-CTAAAACTGGTAAAGATAAT-39) that amplify a 268 bp fragment. The non-amplification of I. ricinus with those primers specific to I. hookeri was confirmed by PCR.

Alignment and Phylogenetic analysis
To identify the Wolbachia supergroup of the Wolbachia sequence associated to I. hookeri, we downloaded Ftsz sequences from the Wolbachia supergroups A, B, C, D, E, F [58] to which we have added several Ftsz sequences with high query coverage and percentage of identity using BLAST. The sequences of Anaplasma phagocytophilum, Rickettsia rickettsi and Ehrlichia chaffeeensis were also added as outgroups and to illustrate the divergence between those well-known pathogen agents transmitted by ticks and W. pipientis. The alignment of the Ftsz sequences was conducted with CLUSTAL W2 [59]. Phylogenetic analyses were conducted with MEGA5 [60]. A Neighbor-Joining tree was built with the Maximum Composite Likelihood distance [61], and a complete deletion option. Bootstrap values were estimated using 10,000 replications.

Ethical Statements
All experiments were carried out at Oniris (Ecole Nationale Vétérinaire, Agroalimentaire, et de l9Alimentation Nantes Atlantique), adhering to local (Pays de la Loire Ethics Committee) and national guidelines and laws of experimental work with laboratory animals who has approved the protocol used in this study (permit number nu 44266).