Novel Wolbachia-transinfected Aedes aegypti mosquitoes possess diverse fitness and vector competence phenotypes

Wolbachia pipientis from Drosophila melanogaster (wMel) is an endosymbiotic bacterium that restricts transmission of human pathogenic flaviviruses and alphaviruses, including dengue, Zika, and chikungunya viruses, when introduced into the mosquito vector Aedes aegypti. To date, wMel-infected Ae. aegypti have been released in field trials in 5 countries to evaluate the effectiveness of this strategy for disease control. Despite the success in establishing wMel-infected mosquitoes in wild populations, and the well-characterized antiviral capabilities of wMel, transinfecting different or additional Wolbachia strains into Ae. aegypti may improve disease impact, and perhaps more importantly, could provide a strategy to account for the possible evolution of resistant arboviruses. Here, we report the successful transinfection of Ae. aegypti with the Wolbachia strains wMelCS (D. melanogaster), wRi (D. simulans) and wPip (Culex quinquefasciatus) and assess the effects on Ae. aegypti fitness, cytoplasmic incompatibility, tissue tropism and pathogen blocking in a laboratory setting. The results demonstrate that wMelCS provides a similar degree of protection against dengue virus as wMel following an infectious blood meal, and significantly reduces viral RNA levels beyond that of wMel following a direct challenge with infectious virus in mosquitoes, with no additional fitness cost to the host. The protection provided by wRi is markedly weaker than that of wMelCS, consistent with previous characterisations of these lines in Drosophila, while wPip was found to substantially reduce the fitness of Ae. aegypti. Thus, we determine wMelCS as a key candidate for further testing in field-relevant fitness tests and viremic blood feeding challenges in a clinical setting to determine if it may represent an alternative Wolbachia strain with more desirable attributes than wMel for future field testing.


Introduction
Arboviruses transmitted by the Aedes aegypti mosquito, including dengue (DENV), Zika (ZIKV) and chikungunya (CHIKV) are emerging threats that impose an increasing health burden on tropical and subtropical regions of the world. While these viruses usually cause selflimiting febrile disease, severe manifestations such as hemorrhagic shock can lead to death. In the absence of specific antiviral therapeutics and suboptimal vaccines [1][2][3][4], treatment is supportive only and limiting virus transmission has been largely dependent on vector control. The increasing global incidence of these diseases demonstrates the lack of effectiveness of current control programs and a need for novel efficacious and cost-effective alternatives [5].
In 2011, the first releases of Ae. aegypti mosquitoes carrying the artificially transinfected endosymbiotic bacterium Wolbachia pipientis began in field trials in Northern Australia as part of the Eliminate Dengue Program [6]. Wolbachia is a gram negative, obligate endosymbiont that is maternally transmitted and can impart antiviral properties to arthropod hosts. It is estimated that at least 40% of all terrestrial arthropod species are infected with Wolbachia [7] which can also manipulate host biology to induce feminization, parthenogenesis, cytoplasmic incompatibility (CI) and male-killing [8,9]. Of these traits, CI is the most common, induced by many (but not all) Wolbachia strains, and enables the spread of Wolbachia into a new host population by providing females with Wolbachia an indirect reproductive advantage. In crosses between uninfected females and Wolbachia-infected males, CI results in the embryonic lethality of offspring. However, Wolbachia-infected females can mate with both uninfected and infected males successfully. CI, coupled with the maternal transmission (MT) of Wolbachia leads to the rapid invasion of the host population [8,9].
The antiviral activity Wolbachia can provide to its host is a more recently described phenomenon [10]. The mechanism that drives this protection is not well understood but may involve priming of the insect innate immune response pathways in new Wolbachia infections, and competition for resources such as cholesterol [11][12][13].
Field trials to date have utilized Ae. aegypti transinfected with the Drosophila melanogaster native Wolbachia strain, wMel, and a more pathogenic form, wMelPop-CLA, isolated from a lab stock of D. melanogaster [6,14,15]. While these strains have been shown to restrict replication and dissemination of several virus genera including flaviviruses and alphaviruses in Ae. aegypti, due to technical difficulties associated with introducing Wolbachia into a new species, the full potential of using alternative Wolbachia strains has remained largely unexplored.
This study set out to provide an initial examination of Ae. aegypti transinfected with novel Wolbachia strains to identify key candidate strains that could potentially contest, or even improve on the current pathogenic protection afforded to Ae. aegypti by wMel, whilst avoiding fitness costs to the mosquito such as impaired egg viability and larval development as described for wMelPop-CLA infections [10,[16][17][18][19]. Given the cost and difficulty in taking alternative strains through a pipeline of more field-relevant laboratory testing and then into comparative field testing the goal of this work was to determine if any of the newly generated transinfected lines reported here could be removed from this pipeline in a first round of comparative testing.
For a Wolbachia-transinfected Ae. aegypti line to be considered for examination in further field-relevant experiments, the Wolbachia strain must provide strong protection against virus replication, demonstrate near-complete MT, and induce CI to enable rapid spread into the wild mosquito population, whilst inducing minimal fitness cost to the mosquito host.
To predict which Wolbachia strains may provide these traits we looked to past studies, most of which have been performed using natively Wolbachia-infected, artificially transinfected, or introgressed Drosophila lines, with the RNA viruses Drosophila C virus (DCV) or Flock House virus (FHV) as models. From these studies, we identified wMelCS (from D. melanogaster) and wRi (from D. simulans) as key candidates [20][21][22][23][24].
Although Ae. aegypti mosquitoes do not naturally carry Wolbachia, a number of other mosquito species have natural infections including the closely related species Aedes albopictus (wAlbA and wAlbB) and Aedes notoscriptus (wNoto) as well as more distantly related species such as Culex quinquefasciatus (wPip). wAlbB has previously been transinfected into Ae. aegypti and shown to provide strong protection against DENV while showing more resilience to heat stress but poorer larval competitive ability compared to wMel [13,[25][26][27][28][29], indicating the difficulty in assessing strains for performance from relatively crude experimental studies and the need for comparative studies to be undertaken under field release conditions. wPip has been reported to provide protection against West Nile virus (WNV) in its natural host, identifying this strain as a possible candidate for transinfection in Ae. aegypti [30].
Here we report the generation of three new Ae. aegypti lines, transinfected with Wolbachia strains wRi, wMelCS and wPip, and describe an initial relative evaluation of these strains, examining the effects on mosquito fitness, MT, CI, tissue tropism and DENV blocking. The findings reveal wMelCS as a potential candidate for future examination under more rigorous field-relevant testing and also indicate that wRi and wPip can be removed as candidates for future testing.

Maternal transmission and cytoplasmic incompatibility
MT of a Wolbachia strain and its ability to induce CI are key features that must be conserved when considering new Wolbachia-infected Ae. aegypti lines for field releases. CI occurs when an uninfected female mated with a Wolbachia-infected male cannot produce viable offspring. This phenomenon ensures all offspring will be Wolbachia-positive, enabling effective spread of Wolbachia throughout a wild population.
MT was assayed by crossing Wolbachia-infected females (at least 5 generations removed from the G0 line; G5) with males from their respective Wolbachia-free tetracycline (Tet)treated line and assaying for the presence of Wolbachia in adult progeny. The crosses revealed near-complete MT for wMelCS and wPip (99.4 and 98.8%, respectively), similar to reported values for wMel (100% with a 95% confidence interval lower boundary of 89%; [15]), while wRi sustained some breakdown of MT at 87.5% (Table 1). CI was determined for each line by measuring the hatch rate from crosses between Tettreated females and infected males from each respective Wolbachia-infected line (Fig 1B). Control crosses where Tet-treated males and females were mated showed high hatch rates (>80%), while in CI crosses, near-complete sterility was observed in all three lines (0.3, 1.8 and 0% for wMelCS, wRi, and wPip, respectively), comparable to previous reports for wMel (0%) [15,26]. Note that newly established lines can display some variability between individuals which may explain the small differences in MT and CI shown here for wMelCS and wPip.

Analysis of mosquito fitness costs induced by each Wolbachia strain
Fecundity, hatch rate and egg diapause viability. Fecundity was measured as the number of eggs laid/female following mating between Wolbachia-infected males and females, and Wolbachia-infected females and Tet-treated males, compared to the control cross of Tet-treated males and females. For wRi and wMelCS, the number of eggs laid/female did not significantly decrease irrespective of these matings, with the wRi-infected male and female cross in fact showing a slight but significant (p<0.05 Kruskal-Wallis, Dunn's test) increase in eggs laid/ female, indicative of a positive effect on fecundity ( Fig 1A). For wPip, relative to the Tet control cross, the mean number of eggs/female decreased for crosses between infected females and males, and infected females crossed with uninfected males (p<0.001, p<0.05, respectively), suggesting wPip reduces fecundity for this line.
We next examined the egg hatch rates of each cross. No significant reduction in hatch rate was observed for wRi crosses between Wolbachia-infected females and males, when compared to Tet-treated male and female crosses. wMelCS showed a slight although significant reduction in the hatch rate between Wolbachia-infected females and males, compared to the Tet-treated control cross (79% and 90%, respectively; p<0.05, Kruskal-Wallis, Dunn's test) (Fig 1B), similar to that reported for wMel [15,26]. Hatch rate was more dramatically impaired for wPip, with rates dropping from 82% for the Tet-treated control cross, to 58% for Wolbachia-infected male and female crosses (p<0.0001). For wMelCS and wPip, crosses between infected females and uninfected males also resulted in significant reductions in hatch rate (68 and 66%, compared to respective Tet control cross hatch rates of 90 and 82%, p<0.01), perhaps suggesting there is a small fitness cost for these strains of Wolbachia in eggs, although these reductions were similar to those described previously for wMel [15].
To determine the effect of Wolbachia on egg viability and storage, eggs were collected and stored for 1-12 weeks before determining hatch rates. The benchmark transinfected line, wMel had a slight, but significant reduction in hatch rate following 4, 8, and 10 weeks storage (p<0.05, 2way ANOVA Sidak's test) compared to its Tet control line, with a slightly larger reduction after 12 weeks (p<0.0001) (Fig 1C). wRi and wMelCS lines displayed similar hatch rates (>60%) across all ages until week 10 (no significant difference in hatch rates compared to the respective Tet control lines until week 8). By contrast, wPip eggs appeared to deteriorate rapidly, with < 20% hatch rate observed after 3 weeks of aging (significant reductions at all weeks tested, p<0.0001, compared to wPip.Tet). Together, these data indicate a substantial fitness cost in the wPip line.  Longevity. Effects on the survival of mosquitoes from each line were next examined over time (Fig 2). wMel, wMelCS and wRi Wolbachia-infected females tracked closely with their Tet-treated controls suggesting the infection had minimal influence on longevity. Similar relative survival rates were observed for males. Minor, but significant differences were observed for wRi females (p<0.0001, Log-rank (Mantel-Cox) test) and wMel (p<0.05) and wMelCS (p<0.05) males, relative to their respective Tet control lines. wPip was the only line that showed a substantially shorter lifespan for both males and females compared to the Tet-control (p<0.0001). Since <1% of wild female Ae. aegypti are expected to live beyond 3 weeks [31], it's expected that no infection described here would likely impact on the lifespan of the mosquito in the field.
We also examined compatibility between wMel and wMelCS, since wMelCS-infected Ae. aegypti appeared to be the least compromised line from these initial screening experiments. We found these strains to be bidirectionally compatible with crosses between wMelCS females and wMel males, and wMel females and wMelCS males, each resulting in successful hatch rates (71 and 62%, respectively, S1 Fig). These data suggest that wMelCS could not be released over an existing wMel strain and expect to overtake wMel by CI alone. The ability of either of these strains to replace another in a field setting will be dependent upon the which strain is least costly to host fitness.
Wolbachia density and distribution. Total Wolbachia density was measured in 5-, 10and 15-day old adult females of each line (at least G9) and compared to the benchmark line wMel at each time point, using qPCR with primers specific to the highly conserved Wolbachia ribosomal RNA gene, 16S, and Ae. aegypti rps17 gene to normalize for DNA input. Overall, the density of each strain reduced as the mosquitoes aged ( Fig 3A). wMelCS density was significantly higher than wMel at days 10 and 15 post emergence (p<0.001, p<0.05, Mann-Whitney test), as was wPip at days 5 and 10 (p<0.001). wRi density was nearly half that of wMel at day 5 and was significantly reduced at day 15 (p<0.01 and p<0.05, respectively), but not significantly different at day 10. Despite these variations, all densities measured were within a 1log 10 range of each other, indicating each Wolbachia strain maintains itself at high density.
The localization of Wolbachia within ovary and salivary gland tissues was next examined in 5-7 day old adult female mosquitoes using fluorescence in situ hybridization (FISH). Mosquitoes were formaldehyde-fixed, and paraffin-embedded tissue sections stained using a probe specific for Wolbachia 16S gene (red, Fig 3B). Total DNA was stained with DAPI (blue). Consistent with the strong MT shown in Table 1 for all lines, each strain was abundant in ovarian tissues, with distribution comparable with wMel ( Fig 3B). Similar distribution of Wolbachia was also observed for all lines in the salivary glands, a key tissue involved in viral transmission.
Restriction of DENV-3 replication by wRi and wMelCS in Ae. aegypti. To compare the viral-blocking capacity of these lines, 7 day old mosquitoes were fed an infectious blood meal containing 2.0 x 10 6 TCID 50 /ml (DENV-3), then incubated for 14 days at 26˚C. Due to the reduced fitness of wPip mosquitoes we elected to exclude this line. Vector competence analyses were performed on wRi and wMelCS lines, and compared to wMel. Mosquitoes were collected, the head separated from each mosquito body, then total RNA extracted to measure rates of infection (as determined by bodies positive for DENV RNA) and dissemination (heads positive for DENV RNA). Consistent with previous findings, wMel reduced the mean DENV RNA at 2nd instar stage until no hatch was observed for a week then the percent hatch calculated. Statistical analysis was performed using 2way ANOVA Sidak's test (n = 4 at each time point). wPip hatch rate was significantly reduced at all weeks compared to wPip.Tet (p<0.0001). wMel hatch rate was significantly reduced at weeks 4, 8, 10, (p<0.05) and 12 (p<0.0001) compared to wMel.Tet. wMelCS had a significantly reduced hatch rate at weeks 10 and 12 compared to wMelCS.Tet (p<0.01 and p<0.0001, respectively). wRi had a significantly reduced hatch rate at week 12 only, compared to wRi.Tet (p<0.0001). levels by 3log 10 in mosquito bodies (3 x 10 4 copies/body compared to 3 x 10 7 copies/body in the relevant Tet control line, p<0.0001, Mann-Whitney test), as well as reducing the infection rate, with just 9% of wMel mosquito bodies scoring positive for DENV infection compared to 94% in the matched Tet control line (where positive was defined as >1000 copies/body, see Materials & Methods; Table 2 with the same data used to generate Fig 4). By contrast, wRi caused only a slight, although significant (p<0.01) reduction in viral copies (7.6 x 10 7 copies/ body compared to 1.3 x 10 8 copies/body for the matched Tet control line) with no reduction in infection rate (96% compared to 98%, respectively). wMelCS gave a phenotype intermediate of wMel and wRi, with a wide spread in the number of DENV copies/body observed, with a mean of 1.5 x 10 7 RNA copies compared to 4.7 x 10 7 RNA copies for its matched Tet control line (p<0.0001). Interestingly, DENV-3 dissemination, key for viral transmission, was similarly restricted between wMel and wMelCS lines, with~4log 10 reduction in viral RNA copies determined in the head of each line relative to the matched Tet control, and just 2 and 9% infection rates, respectively, compared to 83 and 96% for the Tet controls. While the mean Triplicate cages of age-controlled (emergence within 24 h) adults (~150 males and~150 females/cage) were maintained at 26˚C, 65% relative humidity and a 12:12 h light:dark cycle in a climate controlled room. The number of dead males and/or females was recorded and carcasses removed daily until all mosquitoes in the cages were dead. Data are the total % survival from the three cages/line. Significant differences were observed for wRi females (p<0.0001), wMel (p<0.05) and wMelCS (p<0.05) males, relative to their respective Tet control line. wPip had a significantly shorter lifespan for both males and females compared its matched Tet-control (p<0.0001). Statistical analysis was performed using a Log-rank (Mantel-Cox) test.
https://doi.org/10.1371/journal.ppat.1006751.g002 DENV RNA copies/head of wRi-infected mosquitoes was significantly reduced (2 x 10 6 copies compared to 9.5 x 10 6 copies for the Tet control line), the dissemination rates were not reduced (94% compared to 95%, respectively). Together, these results indicate that wRi shows weaker DENV-blocking capacity while the ability of wMelCS to restrict viral dissemination appears to be similar to that of wMel.
In order to more closely examine the blocking capacity of wMelCS we next used an injection challenge model which bypasses the midgut barrier leading to a higher level of tissue infection, to directly compare viral restriction in wMel and wMelCS lines. Female mosquitoes (7 days old) were injected with 2.5 x 10 6 TCID 50 /ml of DENV-3, or 10-fold dilutions thereof. Mosquitoes were incubated for 7 days before total RNA extraction from entire mosquitoes. Note that DENV-injected mosquitoes have virus delivered directly to the thorax, bypassing the midgut barrier to infection, therefore measuring dissemination rates in these mosquitoes is not a relevant test. When mosquitoes were injected with undiluted virus, both Wolbachia lines struggled to restrict viral replication with the number of DENV copies reduced by < 1log 10 relative to each matched Tet control line (Fig 5A and 5B). When mosquitoes were injected with 2.5 x 10 5 TCID 50 /ml (1:10), wMelCS reduced DENV RNA copies by~2log 10 compared to <1log 10 for wMel, indicating that wMelCS may be more effective at restricting replication of DENV-3 than wMel when challenged with a highly infectious setting. This trend continued as the virus was diluted further, with significantly lower viral RNA levels measured in wMelCS mosquitoes when injected with 2.5 x 10 4 TCID 50 /ml (1:100) (p<0.01, Mann-Whitney test) and substantially (although not significantly) reduced RNA levels at 2.5 x 10 3 TCID 50 /ml (1:1000). Importantly, markedly lower rates of infection were also observed in wMelCS compared to wMel-infected mosquitoes for the three lowest injected virus concentrations ("number of DENV-positive mosquitoes/total injected mosquitoes" are indicated above each bar). Together, the laboratorybased infectious blood feeding and injection challenge systems provide a relative, initial evaluation of the vector competence of these new lines, and identify wMelCS as a potential candidate to progress to more intensive human viremic blood challenge experiments.

Discussion
Field releases of wMel-transinfected Ae. aegypti in order to establish Wolbachia in wild populations have been expanding over the past 6 years in multiple countries and the technology the 16S rRNA gene from all four Wolbachia strains. Total DNA was stained in blue using DAPI and a green filter was included to increase contrast with surrounding tissues. Sg indicates salivary gland tissue, m indicates muscle, and c indicates cardia. White arrows identify select regions of Wolbachia staining.
https://doi.org/10.1371/journal.ppat.1006751.g003 shows significant potential to have a large impact on arbovirus disease. However, there is still a need to further develop and optimize the technology to maximize its impact and to that end testing alternative strains of Wolbachia for greater field effectiveness is a priority. To most Here we report for the first time the transinfection of Ae. aegypti with Wolbachia strains wMelCS, wRi, and wPip. wMelCS-infection had minimal effect on mosquito fitness, similar to wMel. However, wRi displayed reduced MT, and the introduction of wPip lead to poor egg viability after short storage times, and weaker fecundity and hatch rates; traits that would be likely to compromise effective release strategies when performed in a field setting. Interestingly, Ae. albopictus transinfected with wPip showed similar reductions in female fecundity and egg hatch rate [32], suggesting that wPip may broadly induce this fitness cost when introduced into a non-native host. Although this may be context dependent, as work from Zhang et al. (2015) determined no fitness cost to Ae. albopictus when wPip was introduced into mosquitoes that also contained the native wAlbA and wAlbB strains [33].

Line
When tested for their ability to block DENV, wMelCS and wRi restricted replication and dissemination following an infectious blood meal, although wRi blocking was substantially reduced compared to wMelCS, with wMelCS restricting DENV dissemination to the mosquito head to a similar extent as wMel. In a virus injection challenge model where mosquitoes are essentially overloaded with virus, wMelCS demonstrated superior blocking under all concentrations of virus tested. These findings indicate the blocking provided by wMelCS is robust, and suggests this line is worthy of further consideration in more exhaustive human viremic blood feeding experiments. While these future experiments are both costly and time consuming, they will provide a critical evaluation of vector competence, more likely to closely represent the performance of these mosquitos in the field, and may determine whether wMelCS should be progressed to field release comparisons [26,34].
It has been inferred that the degree of blocking correlates with density of the Wolbachia strain in key tissues [35][36][37]. We show here that overall Wolbachia densities are only slightly higher in wMelCS-transinfected mosquitoes compared to wMel, consistent with the similar virus-blocking phenotypes. In wRi-containing mosquitoes, Wolbachia was generally found to be present at a lower density than wMel, and was significantly worse at blocking DENV-3 replication and dissemination. Interestingly, when Osborne et al. (2009) examined blocking of DCV by wRi and transinfected-wMel in D. simulans they observed almost identical densities of wRi and wMel, similar protection against virus-induced mortality, yet wRi did not reduce the viral load [20]. While this supports a role for utilizing Wolbachia densities to predict viral protection in mosquitoes, it highlights the complex nature of this tripartite interaction. Tissue specific quantitation of Wolbachia may provide further insight into the importance of Wolbachia-localization and density in virus blocking by each strain.
Given the substantial time it takes to introduce a new Wolbachia strain into Ae. aegypti, understanding how that strain is likely to behave, and the extent of pathogen protection it will provide once introduced into Ae. aegypti, will be of great benefit for ensuring effective transinfected lines are generated. Here, the increased DENV protection provided to wMelCS-infected mosquitoes compared to wMel under certain infection conditions, is in line with work performed in D. melanogaster and D. simulans using DCV and FHV [22,23]. Similarly, wRi provided incomplete protection against these viruses in D. simulans [20], indicating that Wolbachia-infected Drosophila lines may be useful predictors for pathogen blocking in Wolbachia-infected mosquitoes. Further analyses of Ae. aegypti lines transinfected with different Drosophila-derived Wolbacha strains is required to determine if this trend holds. It should also be acknowledged that the characterisations described here were not repeated with independently With a strong virus blocking phenotype, wMelCS may prove to be an effective candidate for field trials without displaying the fitness costs that prevented field establishment of wMel-Pop-CLA [18]. To test this prediction these mosquitoes now need to be examined in more realistic challenge settings, such as blood feeding on viremic patients.

Mosquito rearing
All Ae. aegypti mosquitoes were reared and maintained as described previously [16,26]. Adult mosquitoes were maintained at 26˚C, 65% relative humidity (RH) and a 12 h light:dark cycle in a climate-controlled room. Mosquitoes were blood fed on the arms of human volunteers (Monash University human ethics permit CF11/0766-2011000387). The Wolbachia-infected wMel line included in our experiments has been described previously [15,38]. We generated uninfected lines for each strain by treatment of the infected lines with the antibiotic tetracycline (Tet) as described earlier [16,39]. Briefly, wMelCS, wRi and wPip adults were fed on 10% sucrose solution containing 1 mg/ml Tet right after emergence for two generations followed by a recovery period of another two generations, whereby larvae rearing water from the infected lines was added to the respective Tet-treated lines to homogenize their gut flora. Tetlines were reared for a further minimum of 2 generations before use in maternal transmission, CI, fitness, and vector competence experiments.
To exclude any influence of age and quality of eggs used in experiments, 5-day old females from all lines were fed by the same volunteer in one sitting. The eggs collected from these females were then dried and used as experimental material within 2-3 weeks. For experiments using the wPip line, freshly laid eggs were used due to poor hatch rates following egg storage.
To exclude any influence of mosquito age on our experiments, age-controlled adults emerging within a 24 h window were used [40].

Wolbachia transinfection
The Wolbachia uninfected and inbred PGYP1.Tet line was used as the recipient line for transinfection. Wolbachia strains were not cell line adapted as described previously [16], but were instead directly isolated from donor insects and introduced into Ae. aegypti embryos by cytoplasmic transfer. wMelCS was sourced from Drosophila melanogaster Canton S strain [41], wRi was sourced from Drosophila simulans DSR strain (Riverside) [42], wPip was isolated from Culex quinquefasciatus [43] (a kind gift from Karen Williams, Westmead Hospital, Australia). Embryonic microinjection, isofemale line establishment and selection for stablyinfected lines were done as previously described [15,16] with a few modifications. In brief, the wMelCS, wRi and wPip strains were purified from their respective donor insects, and microinjected into the posterior-pole of pre-blastoderm embryos of the recipient PGYP1.tet using methodology previously described [15,16]. Surviving G0 adult females from microinjection were mated to PGYP1.tet males, blood fed and set up for oviposition as isofemales. G0 females that laid fertile eggs were screened using quantitative PCR (qPCR) as described below. Once the newly introduced Wolbachia strain was at 100% frequency in the population and mean RNA copies of wMel.Tet and wMelCS.Tet lines are indicated by ‡ (p<0.05), Mann-Whitney test. Injection dilutions were performed as independent experiments, with the data combined to produce the final data series.
https://doi.org/10.1371/journal.ppat.1006751.g005 Inhibition of dengue virus by new Wolbachia strains in Aedes aegypti maintained itself for at least 2 consecutive generations the line was considered stable. The wMelCS and wRi lines were stable at G3 and wPip at G5.

Cytoplasmic incompatibility and maternal transmission
To investigate the level of Wolbachia-induced cytoplasmic incompatibility (CI) we set up paired crosses between Wolbachia-infected and Tet-treated Ae. aegypti. For each Wolbachia strain four crosses were set up: (1) a CI cross (Tet females x infected males), (2) a maternal transmission (MT) cross (infected females x Tet males), (3) an uninfected control cross (Tet females x Tet males) and (4) an infected control cross (infected females x infected males). Each cross consisted of a group of 70 virgin males and 70 virgin females and each group was allowed to mate for 5 days after emergence. On day five all females were blood fed by one human volunteer. Gravid females were aspirated into individual tubes three-days post blood feeding and allowed to oviposit on wet filter paper [26]. Three-days post oviposition female mosquitoes were removed and tested for presence of Wolbachia by qPCR using strain-specific primers as described in the PCR section.
Any paper containing less than 15 eggs was discarded. The egg papers from the remaining females were photographed and eggs were counted manually. The papers were then submerged in hatching water. Two days later the number of hatched larvae (2nd instar) was counted. Newly hatched larvae were counted daily until no hatching was recorded for three consecutive days. Any unhatched eggs were dried briefly for 48-72 h, hatched and counted again to get the maximum hatch.
To quantify the success of MT of Wolbachia to the next generation, hatched larvae from the MT cross (infected females x Tet males) were reared to adults and screened by qPCR for the presence of Wolbachia.

Fitness determinants
Egg diapause viability. To assess the viability of eggs stored over time, age-controlled adults (emergence within 24 h; 100 males and 100 females) were placed in quadruplet 30x20x20cm cages. One cage of each line was blood-fed by the same human volunteer, to control for the potential variation caused by blood-meal quality and composition. This was repeated 4 times with 4 different volunteers. Female mosquitoes were given a blood meal at 5 days old. Three-days post blood meal, 12 oviposition cups were placed in each cage for three days. Approximately 72 h post oviposition, egg cups were removed, dried slowly over 3-5 days then stored in Whirlpak bags (Sigma) in an airtight container with saturated KCl solution to maintain humidity [16]. Batches of 100-500 eggs were photographed and counted, then hatched after 1, 2, 3, 4, 6, 8, 10 and 12 weeks from each cage. Hatched larvae were counted at 2nd instar stage until no hatch was observed for a week then the percent hatch calculated.
Fecundity and hatch rate. This was performed using the parameters described for CI and MT above. Fecundity was calculated as the average number of eggs laid/female, and hatch rate determined as the number of 2 nd instar larvae per total eggs hatched.
Adult longevity. Age-controlled adults (approximately 150 males and 150 females) were placed in triplicate 30x30x30cm cages. Mosquitoes were provided fresh 10% sucrose twice/ week and maintained at 26˚C, 65% RH and a 12:12 h light:dark cycle in a climate-controlled room. The number of dead males and/or females was recorded and removed daily until all mosquitoes in the cages were dead.
The distribution of wMelCS, wRi and wPip Wolbachia in mosquitoes was determined in sections of paraffin-embedded female mosquitoes (5 to 7 days old) using fluorescence in situ hybridisation (FISH), as described in Moreira et al. (2009) [10]. The fluorescently labelled probes used can detect the 16S rRNA from all three Wolbachia strains. Total DNA was stained in blue using DAPI and a green filter was included to increase contrast with surrounding tissues.
For feeding experiments with DENV-3 (Cairns 08/09) infected blood, 100 seven-day old age-controlled female mosquitoes were placed in 500 mL plastic containers (five containers per Wolbachia line, two containers per Tet line), starved for up to 24 h and allowed to feed on a 50:50 mixture of defibrinated sheep blood and tissue culture supernatant containing freshly harvested 2.0 x 10 6 TCID 50 /mL of DENV-3. Feeding was done through a piece of desalted porcine intestine stretched over a water-jacketed membrane feeding apparatus preheated to 37˚C. Mosquitoes were left to feed in the dark for approximately 80 min. Fully engorged mosquitoes were placed in 500 mL containers at a density of < 20/container, and incubated for 14 d at 26˚C with 65% RH and a 12 h light/dark cycle.
For adult microinjections, 60 six or seven-day old age-controlled female mosquitoes were anesthetized by CO 2 . The mosquitoes were injected intrathoracically with 69 nL of DENV-3 Cairns 08/09 strain, 2.5 x10 6 TCID 50 /ml (or 10, 100 or 1000-fold dilutions thereof) in RPMI media (Life Technologies) using a pulled-glass capillary and a handheld microinjector (Nanoject II, Drummond Scientific). Injected mosquitoes were incubated for 7 days (10 mosquitoes/ cup) at 26˚C with 65% RH and a 12 h light/dark cycle. To quantify DENV-3 genomic copies, total RNA was isolated from DENV-3 mosquitoes (entire mosquitoes for injection experiments, or head and bodies separately for blood fed mosquitoes) using the RNeasy 96 QIAcube HT kit (Qiagen). DENV-3 RNA was amplified by qRT-PCR (LightCycler Multiplex RNA Virus Master, Roche), using primers to the conserved 3'UTR: Forward 5'-AAGGACTAGA GGTTAGAGGAGACCC; Reverse 5'-CGTTCTGTGCCTGGAATGATG; Probe 5'-HEX-AA CAGCATATTGACGCTGGGAGAGACCAGA-BHQ1-3'and absolute copies determined by extrapolation from an internal standard curve generated from plasmid DNA encoding the conserved 3'UTR sequence. Mosquito extracts with !1000 copies of DENV per body were scored positive, based on the LOD 95 (limit of detection 95%) for DENV-3 with this primer set.

Data availability
All raw data are available within S1 Data.

Ethics statement
Blood feeding by volunteers (Monash University human ethics permit no CF11/0766-2011000387) for this study was approved by the Monash University Human Research Ethics Committee (MUHREC). All adult volunteers provided informed written consent; no child participants were involved in the study.
Supporting information S1 Fig. wMel and wMelCS lines are bidirectionally compatible in Ae. aegypti. Bidirectional compatibility between wMel and wMelCS lines was determined by crossing Wolbachiainfected females and males from each line, with control CI crosses performed between uninfected female mosquitoes (WT) and Wolbachia-infected males of each line. Bars are the mean percentage hatch rate ± SEM from 30 females (individual data points are superimposed). (TIF) S1 Methods. Methods used to generate S1 Fig. are described in S1 Methods. (DOCX) S1 Data. All raw data are available within S1 Data. (XLSX)