The bacterial endosymbiont Wolbachia manipulates arthropod host biology in numerous ways, including sex ratio distortion and differential offspring survival. These bacteria infect a vast array of arthropods, some of which pose serious agricultural and human health threats. Wolbachia-mediated phenotypes such as cytoplasmic incompatibility and/or pathogen interference can be used for vector and disease control; however, many medically important vectors and important agricultural species are uninfected or are infected with strains of Wolbachia that do not elicit phenotypes desirable for disease or pest control. The ability to transfer strains of Wolbachia into new hosts (transinfection) can create novel Wolbachia–host associations. Transinfection has two primary benefits. First, Wolbachia–host interactions can be examined to tease apart the influence of the host and bacteria on phenotypes. Second, desirable phenotypes induced by Wolbachia in a particular insect can be transferred to another recipient host. This can allow the manipulation of insect populations that transmit pathogens or detrimentally affect agriculture. As such, transinfection is a valuable tool to explore Wolbachia biology and control arthropod-borne disease. The present review summarizes what is currently known about Wolbachia transinfection methods and applications. We also provide a comprehensive list of published successful and unsuccessful Wolbachia transinfection attempts.
Wolbachia are obligate intracellular bacteria that infect arthropods and nematodes. Within arthropods, Wolbachia infect a wide spectrum of insects, and the bacteria's success has been attributed to their ability to manipulate host reproduction to favour their own maternal transmission. It is evident that over evolutionary time, horizontal transfer between species has occurred. Incongruent phylogenies of the host and the bacteria suggest that these horizontal transfers are commonplace (O'Neill et al., 1992; Werren et al., 1995; Vavre et al., 1999). The ability of Wolbachia to transfer horizontally between individuals has been exploited in the laboratory to artificially infect new insect species.
Wolbachia strains have been artificially transferred, both interspecifically and intraspecifically, in many arthropod species. Recipient hosts include naturally infected species that have previously had their Wolbachia infection cleared, infected species that consequentially became superinfected with a new strain of Wolbachia, and species that were naturally uninfected. Artificial transfer of symbionts is not limited to Wolbachia, with other examples including Sodalis symbionts in the tsetse fly (Weiss et al., 2006), diverse symbionts of aphids (Russell & Moran, 2005; Tsuchida et al., 2005) and Spiroplasma in flies and beetles (Tinsley & Majerus, 2007; Hutchence et al., 2011); however, in terms of numbers of studies and diversity of artificial interactions created, the transfer of Wolbachia is by far the most developed (Table 1). The present review covers the current knowledge of transfer of Wolbachia between hosts, evaluates different transfer techniques, explores the biology of novel associations and discusses prospects for transinfecting new insect species.
Table 1. Wolbachia transinfection attempts in arthropods
What is transinfection?
Transinfection is the mechanical transfer of symbionts into a novel host. As opposed to introgression, whereby a Wolbachia-infected line is repeatedly backcrossed with males to homogenize the host background, transinfection allows the creation of diverse novel Wolbachia–host associations and is not constricted by species mating barriers. Two procedures have been successfully used to create arthropods transinfected with Wolbachia: embryo microinjection and adult microinjection. By far the most widespread technique used to develop transinfected lines is embryo microinjection, whereby Wolbachia are injected into the posterior pole of preblastoderm embryos using a fine needle and micromanipulator. Embryos are left to develop to adulthood, and the subsequent progeny are screened to determine if germline infection and transmission has occurred (Fig. 1 ). As the name suggests, adult microinjection differs in that the recipient is at the adult stage rather than the embryonic one. For this technique to be successful, Wolbachia must traverse through tissues and cross membranes to reach the germline where they can then be transmitted to the next generation (Fig. 1). In some instances, Wolbachia have been injected into the pupal or prepupal life-stages (Grenier et al., 1998; Kubota et al., 2005; Kageyama et al., 2008); however, similar constraints apply in that, ultimately, Wolbachia need to reach the germline. Other techniques such as corearing of the recipient and donor species (Huigens et al., 2004) and haemolymph transfer (Rigaud & Juchault, 1995) have achieved transfer of Wolbachia to a new host to some degree, but these approaches are only suitable for a limited number of arthropods and are generally not considered viable or efficient options for transinfection in many insect species.
Comparison of embryonic and adult microinjection
For the majority of cases where the successful establishment of stable transinfected lines has occurred, the embryonic microinjection technique was used; however, both this technique and microinjection of other life stages have their advantages and disadvantages. Embryonic microinjection localizes Wolbachia directly within the developing embryo, compared with adult microinjection, where Wolbachia are intrathoracically injected. This direct access to the pole cells means that cells that differentiate into germline or soma are already infected, rather than Wolbachia needing to gain entry to the germline when injected into more developed life stages. Wolbachia have a natural propensity to infect the germline, and in adult Drosophila melanogaster, intrathoracically injected Wolbachia are seen to localize in the stem cell niche within the germline (Frydman et al., 2006). As the soma can be infected, embryonic microinjection can often lead to somatic infection in the transinfected line, which can be important for Wolbachia phenotypes such as pathogen interference (Moreira et al., 2009; Bian et al., 2010; Walker et al., 2011).
When Wolbachia are microinjected into adult insects the bacteria must evade the insect immune response. With naturally occurring associations, Wolbachia seem to neither elicit nor suppress the immune response (Bourtzis et al., 2000; Xi et al., 2008), but when analysing artificial associations, Wolbachia have been shown to affect immunity (Kambris et al., 2009; Moreira et al., 2009; Bian et al., 2010; Hughes et al., 2011b). After injection of Wolbachia into an arthropod host, there is evidence that the bacteria may manipulate host immunity, perhaps to enhance their own proliferation (Braquart-Varnier et al., 2008; Hughes et al., 2011a). In crustaceans, Wolbachia are present within haemocytes and in haematopoietic organs, and these have been suggested to act as a reservoir to discharge the bacteria into the haemolymph (Chevalier et al., 2011). But whether haemocytes can act as a source of Wolbachia in insects is still to be resolved. Microinjected Anopheles mosquitoes were also seen to contain Wolbachia within circulating haemocytes, but it is uncertain if the bacteria were viable (Hughes et al., 2011a). To overcome diminishing Wolbachia levels attributable to the host immune response, more strenous approaches can be adopted by injecting more bacteria, a luxury that can be more easily afforded in adult microinjection. Alternatively, host immunity can be suppressed in adults using molecular techniques such as RNA interference.
Innate differences between adult and embryo immunity may aid Wolbachia establishment in embryos. In D. melanogaster early embryogenesis, haemocytes are restricted to a region within the head mesoderm and have highly divergent roles in development (Holz et al., 2003; Wood & Jacinto, 2007). Another study using Drosophila embryos showed that specialist insect pathogens employ methods to avoid phagocytosis, whereas Escherichia coli were engulfed (Vlisidou et al., 2009). Given the evidence that Wolbachia can avoid or manipulate host immunity, it could be predicted that similar mechanisms are employed by the bacteria to avoid detection in injected embryos. While little is known regarding the interaction of Wolbachia with the host immune system in developing embryos, it is feasible that the host may treat microinjected Wolbachia in a similar fashion to maternally transmitted bacteria; however, further work is required to examine Wolbachia–host interaction in embryos to validate this theory.
Another difference between the transinfection techniques is the number of viable G0 (injected generation) females produced, presumably owing to the sensitivity of each life stage to injection. Using adult microinjection, a high number of G0 can be attained, compared with embryo microinjection where only a small proportion of reproducing adults will arise from injected embryos. These differences can either be viewed in a positive or negative light, and preference for a particular technique may lie where resources allocation is preferred. A larger number of G0 obtained from adult injection means more time needs to be spent on screening progeny for successful transfers, while the lower number of G0 from embryonic microinjection means greater effort needs to be allocated in the injection process. For either technique, the progeny of the G0 need to be screened, as both can lead to somatic infection (which is not necessarily indicative of germline infection). As Wolbachia are maternally inherited, infected males are a dead-end for the bacteria. Obviously for adult injections, females will be injected, but prior to embryo microinjection sex is unknown, therefore approximately half of surviving G0 insects are of no use for the establishment of infected lines. Lastly, embryonic microinjection is a highly specialized technique that requires a substantial initial outlay of funds for equipment and trained personnel, whereas adult microinjection is less technical and cheaper to undertake.
Source of donor Wolbachia
There are several sources from which Wolbachia can be obtained for microinjection. When using the embryo microinjection technique, Wolbachia are predominantly extracted from the egg cytoplasm of an infected species then transferred to the recipient species. Wolbachia can be extracted from either the anterior or posterior regions of the donor egg (Xi & Dobson, 2005), while homogenized ovaries or eggs can also be used as a source of Wolbachia. Alternatively Wolbachia can be extracted from cell culture systems. Cell culture is advantageous in that a high density of homogeneous bacteria can be obtained for microinjection and cell lines can be used to adapt strains of bacteria to the desired host background. The adaptation of the wMelPop strain of Wolbachia by serial passage in a RML-12 Aedes aegypti cell line was seen as critical for the successful transinfection of this strain into the mosquito, Ae. aegypti (McMeniman et al., 2009). After transfer of Aedes-adapted wMelPop back into the native D. melanogaster host, phenotypic shifts were interpreted as genetic adaptation of Wolbachia to the mosquito intracellular environment (McMeniman et al., 2008); however, it is unclear if these effects were specifically attributable to adaptation in Aedes cells or changes to the Wolbachia phenotype resulting from serial passage in cell lines. In contrast, the wAlbB strain was transferred from Aedes albopictus to Anopheles stephensi without adaptation in cell lines (Bian et al., 2013). Although more laborious, adaptation of Wolbachia to the novel host could also be achieved by subsequent microinjection into the recipient species followed by extraction after Wolbachia has replicated within the insect (Frydman, 2006); however, this may preferentially select for Wolbachia capable of infecting somatic tissue. The optimum adaptation system would be a cell line derived from the germline of the recipient species or genus.
Phenotype modification upon transinfection
Phenotypic shifts are common upon transfer of Wolbachia to a novel host, probably as a result of the maladaptation of the new association. Theory predicts that novel host–parasite relationships are likely to be maladapted, compared with old associations that are more mutualistic (Turelli, 1994; Levin, 1996). There are several examples where transinfected insects, representing a novel association, display fitness costs. Transfer of wVul from Armadillidium vulgare to Porcellio dilatatus males was lethal (Bouchon et al., 1998), while wMelPop from D. melanogaster had pathogenic effects in Ae. albopictus (Suh et al., 2009). An. stephensi mosquitoes infected with wAlbB have severe reductions in fecundity (Bian et al., 2013). Furthermore, fitness costs were incurred in Drosophila species transinfected with Wolbachia (Clancy & Hoffmann, 1997; McGraw et al., 2002). Nevertheless, in a relatively short time period, shifts from pathogenic to mutualistic associations can occur, demonstrated in both transinfected (McGraw et al., 2002) and natural infections (Weeks et al., 2007).
Selection of transinfected lines
As a result of the maladaptation of novel Wolbachia–host associations, the infection frequency in early generations can fluctuate stochastically. This may be attributable to incomplete cytoplasmic incompatibility [(CI) a form of reproductive manipulation induced by Wolbachia on the host], inefficient vertical transmission, and/or pathogenic association between the Wolbachia and host; therefore, selection plays a critical role in establishing stable transinfected lines. In transinfected D. melanogaster lines, an initial selection period saw infection frequencies rise before a period where relaxed selection led to a drop in infection frequency (McMeniman et al., 2008). A second selection period after approximately G15 led to 100% infection in subsequent generations (McMeniman et al., 2008). After transinfection of a life-shortening strain of Wolbachia from D. melanogaster to Ae. albopictus, selection and outcrossing to wild-type males was crucial in preventing the loss of infection (Suh et al., 2009). In contrast, Xi & Dobson (2005) suggest that after extensive screening at G0 and G1 stages, it was unusual to lose an infection from a Drosophila transinfected line; however, this may be the limited to intraspecific transfers, particularly into Drosophila backgrounds, which are likely to be more amenable to infection. As associations tend to shift towards mutualism over time, selection in the initial phase after transinfection may avoid loss of infection attributable to instability in infection frequencies and temporary fitness costs attributable to novel interactions between the host and bacteria.
Applications of transinfected arthropods
Transinfection can be used to examine novel associations and to disentangle the role of the bacteria or the host genotype on the type of reproductive alterations induced. Such examples are prominent within the order Lepidoptera. When the feminizing strain wSca from the moth Ostrinia scapulalis was transinfected into Ephestia kuehniella, it induced male killing, demonstrating the influence of host factors determining reproductive phenotype (Fujii et al., 2001). The reciprocal transfer of wKue, which induces CI in E. kuehniella, also induced CI when transferred into O. scapulalis, suggesting this phenotype is derived from the bacteria (Sakamoto et al., 2005). Transfer of wCauA from Cadra cautella induced male killing in E. kuehniella, while transfer of wCauB induced incomplete CI (Sasaki et al., 2002). Host suppression of Wolbachia-mediated reproductive manipulations and hidden Wolbachia phenotypes have been observed in other natural associations (Hornett et al., 2006, 2008).
The Drosophila species complex harbour many strains of Wolbachia, with variable levels of CI (McGraw & O'Neill, 1999; Merçot & Charlat, 2004). The variability within these strains lies in the ability of the bacteria to induce and/or rescue CI, which has led to classification of strains based on their modification (mod) and rescue (resc) propensity (Werren & O'Neill, 2004). Four variants are possible: mod+/resc+, mod+/resc−, mod−/resc+, and mod−/resc−. The strains wYak, wTei and wSan do not cause CI in their natural hosts (Drosophila yakuba, Drosophila teissieri, and Drosophila santomea, respectively), but when these strains were transferred into Drosphila simulans, they all induced CI (Zabalou et al., 2008). Conversely, when the wRi strain from D. simulans was transferred into D. yakuba, D. teissieri, or D. santomea, this strain induced CI in all three species (Zabalou et al., 2004a). Unexpectedly, the natural infections (wYak, wTei and wSan) rescued the modification of wRi-transfected flies (Zabalou et al., 2004a). Interestingly, when the wTei strain was transfected into D. simulans, it was unable to rescue its own CI modification, suggesting that this association changes wTei into a ‘suicide strain’ (mod+/resc−) (Zabalou et al., 2008). Taken together, these novel associations demonstrate the unique interactions between Wolbachia and the host that influence reproductive phenotypes.
Wolbachia have been suggested as an agent for vector and disease control given their ability to spread into insect populations and inhibit pathogens (Bourtzis, 2008; Brownlie & Johnson, 2009; Cook & McGraw, 2010); however, many important vector species are uninfected. These uninfected insects can be considered an open niche for Wolbachia modification. Transinfected disease vectors that transmit human or agricultural pathogens with a stable Wolbachia infection include, An. stephensi (Bian et al., 2013), Ae aegypti (Xi et al., 2005b; Ruang-Areerate & Kittayapong, 2006; McMeniman et al., 2009; Walker et al., 2011), Ae. albopticus (Xi et al., 2006; Suh et al., 2009; Calvitti et al., 2010; Fu et al., 2010; Blagrove et al., 2011), Aedes polynesiensis (Andrews et al., 2012) Culex pipiens (Walker et al., 2009) Ceratitis capitata (Zabalou et al., 2004b, 2009), Nilaparvata lugens (Kawai et al., 2009), and Laodelphax striatellus (Kang et al., 2003). Some of these transinfected insects have been evaluated in semi-field and field conditions and are being implemented as control strategies (Hoffmann et al., 2011; Rasgon, 2011; Walker et al., 2011), which is encouraging for the use of Wolbachia in an applied setting. Nevertheless, despite many transinfection attempts, Wolbachia infection remains illusive in some important vector species.
Insects recalcitrant to infection
Some hosts are seemingly impervious to Wolbachia infection. The reasons for refractoriness are unknown, but could relate to either bacterial or host factors. It is likely that the number of hosts recalcitrant to infection is underestimated, as transinfection attempts are predominantly made on either model organisms or insects of economical and medical importance. Furthermore (and unfortunately for the scientific community), unsuccessful transinfection attempts are rarely published. Despite Tribolium confusum being naturally infected and amenable to transinfection (Chang & Wade, 1994, 1996), hundreds of attempts to infect Tribolium castaneum with Wolbachia from T. confusum have not yielded an infected line, despite using the same methodology (pers. comm. M.J. Wade). Transinfection attempts by injecting Wolbachia into Bombyx mori larvae lead to somatic infection in G0, but no stable infection (Kageyama et al., 2008).
For many years transinfection was attempted on mosquitoes within the Anopheles genus given their importance in human health. As a result of many failed attempts, and the lack of natural infection in native Anophelines (Kittayapong et al., 2000; Ricci et al., 2002; Rasgon & Scott, 2004), there were suggestions that this genus was impervious to Wolbachia infection; however infection of Anopheles cells (Rasgon et al., 2006) and somatic infection of adults (Jin et al., 2009; Kambris et al., 2010; Hughes et al., 2011a, 2012b) indicated these mosquitoes are capable of harbouring Wolbachia, suggesting the inability to create a stable line was the result of other barriers. Recently, Bian et al. (2013) developed a stable line using embryonic microinjection in the Indian malaria vector, An. stephensi demonstrating Wolbachia can stably associate in this genus of mosquitoes, but a stable infection in the major African malaria vector, Anopheles gambiae, which is the main vector of Plasmodium parasites in sub-Saharan Africa, still remains elusive. Promisingly, the tissue distribution of Wolbachia and pathogen inhibitory effects between stably infected An. stephensi (Bian et al., 2013) and somatically infected An. gambiae (Kambris et al., 2010; Hughes et al., 2011a, 2012b) are similar, suggesting stable infection in the African malaria mosquito is feasible.
While speculative, there are several strategies that could be implemented to achieve infection in this important mosquito species. Although more convenient, the use of laboratory mosquito lines may decrease genetic heterogeneity in the host background. Greater genotypic divergence in both the host and the bacteria may provide a wider base to select for compatible associations. Furthermore, pathogenic Wolbachia strains, such as wMelPop, despite having desirable phenotypes for vector control, may not be best suited to infection. Recently, we developed an ex vivo germline assay to determine the infectivity of Wolbachia into ovaries (Hughes et al., 2012a). Wolbachia strains infected native hosts or close relatives at higher titres compared with more divergent hosts, with the wAlbB strain (from Ae. albopictus) more adapted to the Anopheles germline than Wolbachia from flies. This technique may be used to assess the suitability of bacterial strains in a particular host and to determine if adaptation to a new species is occurring. Specifically, adapting a host to infection may be achieved by genetic modification of the insect, making the hosts more susceptible to Wolbachia. In Aedes mosquitoes, host factors induced by Wolbachia manipulation include the host microRNA profile which subsequently increased bacterial titre (Hussain et al., 2011). The development of new transgenic techniques that allows manipulation of host gene expression at early life stages provides tools that could be used to enhance embryonic microinjection in species where transinfection attempts have previously been unsuccessful (Peng et al., 2011). Further research that identifies barriers to horizontal transfer of Wolbachia between species and novel transinfection techniques would be advantageous to the scientific community and accelerate the creation of new Wolbachia–host associations.
Diverse Wolbachia–host combinations can be created through transinfection. This allows a tractable method of studying Wolbachia–host interactions. Embryonic microinjection is predominantly used to create stable Wolbachia-infected lines, but the number of examples that use techniques is increasing. Novel transinfection techniques may be the key to transferring Wolbachia into hosts that have so far resisted infection. Further developments in this field may allow the biological control of arthropod-borne diseases.
We thank Roel Fleuren for assistance developing Fig. 1. While we endeavoured to compile an exhaustive summary of transinfected arthropods in Table 1, we apologize if any research was inadvertently overlooked. This research was supported by NIH grants R21AI070178 and R21AI088311 to JLR.