Invasion of Wolbachia into Anopheles and Other Insect Germlines in an Ex vivo Organ Culture System

The common bacterial endosymbiont Wolbachia manipulates its host's reproduction to promote its own maternal transmission, and can interfere with pathogen development in many insects making it an attractive agent for the control of arthropod-borne disease. However, many important species, including Anopheles mosquitoes, are uninfected. Wolbachia can be artificially transferred between insects in the laboratory but this can be a laborious and sometimes fruitless process. We used a simple ex vivo culturing technique to assess the suitability of Wolbachia-host germline associations. Wolbachia infects the dissected germline tissue of multiple insect species when the host tissue and bacteria are cultured together. Ovary and testis infection occurs in a density-dependent manner. Wolbachia strains are more capable of invading the germline of their native or closely related rather than divergent hosts. The ability of Wolbachia to associate with the germline of novel hosts is crucial for the development of stably-transinfected insect lines. Rapid assessment of the suitability of a strain-host combination prior to transinfection may dictate use of a particular Wolbachia strain. Furthermore, the cultured germline tissues of two major Anopheline vectors of Plasmodium parasites are susceptible to Wolbachia infection. This finding further enhances the prospect of using Wolbachia for the biological control of malaria.


Introduction
Wolbachia are intracellular a-proteobacteria that infect approximately two-thirds of insect species along with numerous other arthropods [1,2]. Generally classified as reproductive parasites, Wolbachia are able to manipulate their hosts reproduction in various ways [3]. These reproductive manipulations exploit the maternal transmission route of the bacteria to the progeny of the host, facilitating the spread of Wolbachia into the host population. In addition to maternal transmission, the incongruence between the phylogenies of Wolbachia and their hosts suggests horizontal transmission has occurred repeatedly over evolutionary time, which has extended the host range of the bacteria [4,5,6]. The ability of Wolbachia to invade new species has been reproduced in the laboratory, with artificial transfers into novel insect species. Despite transinfection of Wolbachia into many new insect species, it is uncertain why some species are capable of infection yet others are seemingly not.
For either horizontal transmission or transinfection to occur, Wolbachia must establish in the female germline to allow the bacteria to be transmitted to the next generation. When artificially transferred into a new host, Wolbachia are usually microinjected into the preblastoderm embryo, a scenario that is unlikely to occur in nature. However, in experiments that more closely mimic potential natural routes of horizontal transfer, Wolbachia has been shifted between insects by adult injection, hemolymph transfer and co-rearing [7,8,9]. When injected into uninfected Drosophila, Wolbachia migrate across numerous tissues to establish in the stem cell niche of the ovary [7]. Hemolymph transfer between infected and uninfected pillbugs (Armadillidium vulgare, Porcellionides pruinosus and Chaetophiliscia elongate) resulted in stable intraspecific transfer of Wolbachia and manipulation of host reproductive [8].
The capacity for Wolbachia to spread though host populations and interfere with pathogen development in insects makes them attractive agents for the control of vector-borne diseases. Currently, Wolbachia is being investigated to control malaria, a devastating disease transmitted by Anopheles mosquitoes. Wolbachia is capable of infecting Anopheles cells in vitro and Anopheles somatic tissues in vivo, and influences the expression of numerous host genes [10,11,12]. After injection into adult Anopheles mosquitoes, Wolbachia influences Plasmodium levels, with wAlbB and wMelPop reducing P. falciparum intensity [10]. Interestingly, somatic infection of An. gambiae with the wAlbB strains enhances oocyst density of P. berghei, the model murine malaria species [13]. However, all Anopheles mosquitoes lack Wolbachia infection in the wild and despite numerous attempts seem impervious to stable transinfection in the lab [14,15]. To increase the probability of infecting Anopheles mosquitoes, a variety of combinations of Wolbachia strains and Anopheles species should be assessed to determine which pairing is most likely to be successful.
Establishing stable Wolbachia infections in novel hosts by embryonic or adult microinjection is a very labor-intensive process with a high failure rate. The development of a rapid assay to assess which Wolbachia strains are most effective at invading the insect germline would greatly facilitate the generation of novel stable Wolbachia infections in insect hosts of interest. Various mosquito tissues have been cultured in an ex vivo setting using cell culture media [16,17]. Fat body and ovarian culture has been used to dissect mosquito physiological responses [18,19,20]. As such, ex vivo mosquito organ culture provides a controlled tractable system to assess specific experimental treatments on insect tissues. Here, we utilized ex vivo tissue culture to develop a simple, rapid, assay to assess germline infectivity in novel Wolbachia-host combinations.

Results
Wolbachia infects the germline of multiple insect species in culture The host species chosen for Wolbachia invasion experiments varied in their natural Wolbachia infection status; naturally infected and cured (D. melanogaster), uninfected yet empirically susceptible to transinfection (Ae. aegypti), or uninfected (An. gambiae, An. stephensi, C. tarsalis). qPCR analysis comparing the number of Wolbachia genomes normalized to host genome copy number found that both the male and female germline tissues of all five insect species were susceptible to Wolbachia infection. There was a positive correlation between inoculum titer and Wolbachia germline infection density, indicating that ex vivo infection occurred in a density-dependent manner. In five of the ten infection experiments, significantly higher Wolbachia densities were observed in ovaries inoculated with a higher titer of bacteria. In each of the five other cases, there was a trend towards higher bacterial densities when a greater amount of Wolbachia was used to inoculate the germline tissue, though the differences were not statistically significant ( Figure 1).
To examine the localization of the infection and confirm that qPCR results were not due to surface contamination by Wolbachia, we performed fluorescence in situ hybridization (FISH) on wAlbBinfected ovaries followed by confocal microscopy. Wolbachia was visually detected in the ovaries of all five species (Figure 2, Movies S1, S2, S3, S4, and S5). Most of the infection was observed in the ovarian follicular epithelium, however fluorescent signal was also seen within the An. gambiae ovarian follicle ( Figure S1). Corroborating the qPCR results, a much more intense wAlbB infection was seen in Ae. aegypti ovaries compared to the other four insect species (Figure 2).

Wolbachia reaches higher infection titer in native hosts and close relatives
Analysis by qPCR shows that both the ovaries and testes of Ae. aegypti are far more susceptible to infection with wAlbB (from the closely related Ae. albopictus) than the other species tested (Figure 1), reaching titers 1000-7000 fold higher than other species (Figure 3) (Kruskal-Wallis test P,0.02, Dunn's test shown in Table S1). Similarly, the wMelPop strain from D. melanogaster infected fly ovaries at a density approximately 500 to 3000 fold higher than the ovaries of other mosquito species ( Figure 3) (Kruskal-Wallis P,0.02, Dunn's test shown in Table S1). The wMelPop strain, which has been transinfected into Ae. aegypti mosquitoes [21], infects the Ae. aegypti germline in the ex vivo assay at a higher density than both Anopheles mosquito species (2.4 times greater than An. gambiae (P,0.03); 4.2 times greater than An. stephensi (P,0.03)) ( Figure 3). We also cultured ovaries from a naturally wMelPopinfected host using the ex vivo system. Infected ovaries from 20-dayold D. melanogaster were cultured and ovarian Wolbachia titer was compared to levels in ovaries immediately dissected from the fly. There was no significant difference in the wMelPop titer in vivo compared to ex vivo, nor between Wolbachia titer of naturally infected ovaries after ex vivo culture and titer in fly ovaries which were infected using the ex vivo system ( Figure 3).
wAlbB infects Anopheles germline more efficiently than wMelPop Our data suggest that Anopheles ovaries are more susceptible to infection with the wAlbB strain than the wMelPop strain. After  normalizing the densities of Wolbachia in the inoculum, the wAlbB strain infected the Anopheles germline at a greater density (24 times for An. gambiae (P,0.008) and 34 times for An. stephensi (P,0.008) than the wMelPop strain ( Figure 3).

Wolbachia infects, but does not replicate in, cultured germline tissue
To determine if Wolbachia can replicate in cultured ovaries, we assayed bacterial density in the cultured host tissues at various times post-infection. For some species, we saw an initial drop in titer after the removal of Wolbachia from the media, but in all species there was an apparent increase in the ratio of Wolbachia to host genome copies by six days post inoculation ( Figure 4). However, these differences were not statistically significant. Further examination of the results showed that the elevated density at latter time points was due mainly to a reduction in the number of host genomes rather than an increase in Wolbachia genomes. To confirm this, we completed another two experiments. First, we cultured naturally infected ovaries from 1-3 day old D. melanogaster and measured the changes in Wolbachia density over time. The variation in the S7 gene had a major bearing on the ratio of Wolbachia genomes to host genomes ( Figure 4F). Second, we cultured host tissue for 6 days prior to addition of Wolbachia and compared the observed densities to those of ovaries inoculated with the same titer of Wolbachia immediately after dissection. Delaying the inoculation resulted in significantly higher Wolbachia:host gene ratios for, Ae. aegypti (P,0.007) ( Figure 5), while in three other species there was a general trend towards higher values, indicating that for most species, and particularly Aedes, host degradation was responsible for the increase in relative Wolbachia density observed in the time course experiments. Taken together, these data suggest Wolbachia is not significantly replicating in ovaries in an ex vivo environment.

Discussion
In order to be maternally transmitted after horizontal transfer, Wolbachia must first invade the host ovarian tissues. Previously, we reported that Wolbachia is capable of replicating within Anopheles cells in vitro and profoundly influences host gene expression [11,12]. After microinjection of the wMelPop strain Wolbachia into adult female Anopheles mosquitoes, the bacteria was noticeably absent from the ovarian tissue, despite dissemination to and replication in numerous other host tissues [10].
Here, we show that Anopheles and other insect germlines are amenable to Wolbachia infection in an ex vivo organ culture system. After two days of co-culture, ovaries became infected with Wolbachia, predominantly in the ovarian follicular epithelium, though there was evidence of Wolbachia residing in the An. gambiae oocyte as well. It is possible that, in order to migrate through follicular epithelium cells, Wolbachia may require the cells to be metabolically active, which may not occur under ex vivo culture conditions. After injection of wMel into uninfected Drosophila, the bacteria were seen to infect the follicular cells at approximately 6-8 days and the germline tissue after approximately 15 days [7]. Our observation of shorter invasion times most likely results from the ability of Wolbachia to directly access the germline tissue without the need to migrate through somatic tissue or contend with the host immune response.
We then compared the infection properties of two divergent strains of Wolbachia (wAlbB from Ae. albopictus and wMelPop from D. melanogaster). The ability of Wolbachia to invade it's native host (wMelPop and D. melanogaster) and closely related hosts (wAlbB and Ae. aegypti) at a higher density suggests that these relationships have developed in close association and are more stable than more disparate associations. Additionally, levels of wMelPop from naturally infected ovaries cultured in the ex vivo system for 2 days were similar to both the density of wMelPop in naturally-infected fly ovaries and the density of wMelPop that had invaded uninfected fly ovaries in the ex vivo system, suggesting that the density of infection in the ex vivo system has biological relevance. Taken together, these results validate the ex vivo culturing technique as a tool to assess Wolbachia-host associations.
To ascertain if ex vivo culture can be used to adapt Wolbachia to the Anopheles germline environment, we assayed Wolbachia density over time to determine if the bacteria replicates in cultured ovarian tissue. Although the ratio of Wolbachia to host genomes increased in all species over time, further analysis of the number of copies of Wolbachia and host reference genes implicated degradation of the host cells as the main cause of the observed increase in Wolbachia density. In some species, there is an initial decline in Wolbachia 2 dpi. This may indicate that bacteria that were bound to the ovary at 0 dpi failed to enter and consequently, later time points had lower Wolbachia titers. Culturing naturally infected ovaries gave similar results to infection ex vivo in that the host gene markedly influenced the density (ratio of Wolbachia genomes to host genomes). Delaying addition of the Wolbachia inoculum to the cultured Ae. aegypti ovaries significantly increased the titer, suggesting for this species that host degradation is responsible for the increase in Wolbachia density over time, rather than replication of the bacteria in the germline tissue. While it has been reported that ovaries can be cultured using similar techniques for over 2 months [17], it is likely that the tissue is not metabolically active. Host derived factors may therefore be absent which could be . As the density of Wolbachia used as the inoculum for uninfected ovaries differed between strains, the final Wolbachia genomes:host genomes value were normalized to 10 6 bacterial cells to compare between strains. The wMelPop strain infects D. melanogaster ovaries approximately 500 to 3000 fold higher, while wAlbB infects Ae. aegypti ovaries 1000 to 7000 fold greater than the other insect species. Asterisk(s) denote significance (** P,0.01, * P,0.05). There was no significant difference in wMelPop titer after two days of ex vivo culture compared to in vivo. For each insect species and Wolbachia strain, five pools, each containing five pairs of ovaries, were evaluated for Wolbachia density. Error bars indicated SEM. doi:10.1371/journal.pone.0036277.g003 critical for Wolbachia replication. In naturally infected Drosophila oocytes, Wolbachia replication has been observed during the period of ovarian development approximately mid-oogenesis [22]. The lack of an active cell cycle may be a reason for the observed lack of replication in cultured ovaries. As the bacteria do not seem to be replicating in the host tissue, we presume that this approach is not suitable to pre-adapt Wolbachia to the Anopheles germline environment.
If these ex vivo results are representative of the live mosquito, they suggest that Wolbachia is able to successfully infect the ovaries of Anopheles. However, stable introduction of Wolbachia into these important vectors has not been accomplished, nor has infection been observed in the wild. Previous studies have indicated that Wolbachia are able to both invade and reproduce in Anopheles cells in culture and adult somatic tissues, but have not reported invasion into reproductive tissues [10,11]. Here, we show that the reproductive tissues can be infected when other host factors are absent, but that Wolbachia do not reproduce under these conditions. Therefore, the question as to why Anopheles species remain uninfected with Wolbachia remains. Possible explanations include a lack of reproduction in Anopheles ovaries under natural conditions, negative interactions between the bacteria and the host, factors blocking invasion of the ovaries in adult mosquitoes and many others, requiring further elucidation if Wolbachia is to be used for malaria control.
The infection of ex vivo cultured insect ovaries with Wolbachia can be used as a tool for the assessment of the suitability of various Wolbachia strains to infect a given species of interest. By testing the ability of different Wolbachia strains to invade the host germline in culture, the strain with the greatest ability to do so can be chosen for transinfection. Given the similar horizontal infection routes, data gleaned from this approach may be more useful for assessment of adult transinfections. However, in general, affinity for the germline is critical for any novel Wolbachia-host association regardless of whether the transinfection is by adult or embryonic microinjection. This system can also be used to investigate the interactions between multiple Wolbachia strains infecting the same host and to examine the interactions between naturally infected hosts and their Wolbachia. Finally, ex vivo culture provides a tractable method by which the effects of various treatments on the ability of Wolbachia to invade host germline tissues can be investigated.
For Wolbachia to be used for vector control in novel hosts, stably transinfected lines must be generated, which requires that the bacteria successfully invade the host germline tissue. Here, we demonstrate a new method for quickly evaluating the suitability of a given Wolbachia strain to invade the reproductive tissues of an insect host. Using this method, we show that Wolbachia is able to infect the ovaries of both An. gambiae and An. stephensi, two important vectors of human malaria parasites. Our results also indicate that Wolbachia invasion occurs in a density dependent manner and that Wolbachia are able to more efficiently infect the tissues of their natural host or closely related species than those of more distantly related insects, providing evidence for the adaptation of Wolbachia to their hosts during long term association. This study provides support for the possibility that Wolbachia can infect Anopheles mosquitoes, and therefore be useful in the control of malaria.

Ex vivo culture of insect germline tissues
Germline tissues (ovaries and testes) were dissected from two to three days old Anopheles gambiae (Keele strain), An. stephensi (Liston strain), Culex tarsalis (KNWR strain), Aedes aegypti (Rockefeller strain) and Drosophila melanogaster (w 1118 flies previously treated with tetracycline to remove the natural Wolbachia infection) in PBS using a dissecting microscope (Olympus). For culture experiments that used naturally infected ovaries, untreated D. melanogaster (w 1118 infected with wMelPop) were used. Prior to dissection, insects were surface sterilized with 100% ethanol for five minutes and rinsed in sterile PBS. After dissection, ovaries or testes were briefly washed with 100% ethanol and rinsed in Schneider's media with 10% fetal bovine serum (FBS) (Invitrogen), 50 mg/ml kanamycin (Sigma), 10 units/mL penicillin and 10 mg/ml streptomycin (Invitrogen). Wolbachia replication dynamics are not affected by kanamycin, penicillin or streptomycin [23,24]. Five ovaries or testes were placed within a single cell culture insert (Millipore). Five replicates were completed for infection experiments, while three replicates were done for time course (replication) assays. Under sterile conditions, germline tissues were washed three times with 1 mL of fresh Schneider's media then incubated at room temperature in 1 mL of Schneider's media with 10% FBS, 50 mg/ml kanamycin (Sigma), 10 units/mL penicillin and 10 mg/ml streptomycin (Invitrogen).

Wolbachia infection
The wAlbB and wMelPop strains of Wolbachia were purified from infected Sua5B [25] (provided by M. Jacobs-Lorena) and Mos55 [26] (provided by S. O'Neill) cells, respectively, according to previously published procedures [27]. Purified Wolbachia were stained using the Live/Dead BacLight Bacterial Viability Kit (Invitrogen) and counted using a hemocytometer under a fluorescence microscope (Olympus) to determine the density of viable cells. One microliter of purified Wolbachia, with a known density, was then added to each tissue culture well containing ovaries or testes. When naturally infected ovaries were assessed, purified Wolbachia was not added to the culture.

Quantitative real-time PCR (qPCR)
After two days in culture, germline tissues were removed from the culture media, washed thoroughly 3 times with fresh Schneider's media before DNA was extracted using the QIAamp DNA Micro Kit (Qiagen). This DNA was used as template for qPCR to estimate the density of Wolbachia in cultured ovaries or testes. wAlbB was detected using primers GF and BR that amplify the wsp gene [10] while wMelPop was detected with primers amplifying the single copy WD_0550 gene [28]. Wolbachia densities were normalized to single-copy host genes. For An. gambiae, An. stephensi and Ae. aegypti, primers amplifying the S7 rRNA gene were used [29,30], while primers which amplify a Figure 5. The effect of culturing before inoculation of Wolbachia on ovary density. Ovaries were inoculated with Wolbachia immediately after culture (day 0) or 6 days post culture for An. gambiae (Ag), An. stephensi (As), C. tarsalis (Ct), Ae. aegypti (Aa) and D. melanogaster (Dm). Two days post-infection, the density of Wolbachia was compared in these two treatments. Density of Wolbachia in ovaries was significantly different (P,0.007) between the Ae. aegypti day 0 and day 6 treatments. For each insect species, five pools, each containing five pairs of ovaries, were evaluated for Wolbachia density. Error bars indicate SEM. doi:10.1371/journal.pone.0036277.g005 segment of the actin gene were used for C. tarsalis [31]. D. melanogaster S7 primers were designed using Primer3 [32]. Primer sequences are found in table S2. All qPCR was performed on a Rotor Gene Q (Qiagen) using the Rotor Gene SYBR Green PCR Kit (Qiagen) using manufacturers recommended settings. Each reaction contained DNA extracted from five ovaries or testes and was performed in duplicate. A third technical replicated was completed when there was a discrepancy between the replicates. Wolbachia to host genome ratios were calculated using Qgene [33]. Changes in individual Wolbachia or host gene values were estimated by raising the reaction efficiency to the exponent power of the negative threshold cycle (E 2Ct ) as previously described [27]. For single comparisons between infected ovaries, data were analyzed using a Mann-Whitney U test, while multiple comparisons were performed using a Kruskal-Wallace test followed by a Dunn's Post-Hoc test.

Fluorescence in situ hybridization
In order to ensure that qPCR results were not due to Wolbachia contaminating the surface of cultured organs, we used fluorescence in situ hybridization (FISH) to assess Wolbachia infection in mosquito tissues. After two days in culture, FISH was completed on ovary tissue according to previously published procedures [10] with slight modifications. Briefly, Schneider's media was removed from culture wells and ovaries were washed thoroughly three times with fresh Schneider's media, fixed in 4% paraformaldehyde in PBS for 20 minutes then moved to 6% hydrogen peroxide in ethanol for 4 days to minimize autofluorescence. Ovaries were allowed to hybridize overnight in 1 ml of FISH hybridization buffer (50% formamide, 56 SSC, 200 g/liter dextran sulfate, 250 mg/liter poly(A), 250 mg/liter salmon sperm DNA, 250 mg/ liter tRNA, 0.1 M dithiothreitol (DTT), 0.56 Denhardt's solution) with Wolbachia specific probes W1 and W2 labeled with a 5-prime rhodamine fluorophore [34]. After hybridization, ovaries underwent three successive washes in 16 SSC, 10 mM DTT and three times in 0.56 SSC, 10 mM DTT. Tissues were mounted on a slide with SlowFade Gold antifade reagent (Invitrogen) and counterstained with DAPI (Roche). Images were captured on an LSM 510 META confocal microscope (Zeiss) and compared to no probe controls. Images were processed using LSM image browsers (Zeiss) and Photoshop 7.0 (Adobe).

Supporting Information
Movie S1 Fluorescence in situ hybridization of Wolbachia infection in An. gambiae ovaries. Serial confocal images of the Z-axis (Z-stack) of 23 sections (0.48 mm step size). Red is Wolbachia, blue is host nuclei (DAPI).   (Figure 3). Asterisks indicate significance after correcting for multiple comparisons using a false discovery rate of less than 5%. (DOCX)