Strategies for controlling mosquito-vectored pathogens based on infection with the symbiotic bacterium Wolbachia include vector population replacement and population suppression (Sinkins, 2004; Werren et al., 2008; Iturbe-Ormaetxe et al., 2011; Bourtzis et al., 2014). Although many mosquito species are naturally infected with various strains of Wolbachia, artificially but stably induced Wolbachia infection in mosquito host species not naturally having Wolbachia infection was originally established by embryonic injection followed by trans-generational spread (Dobson et al., 1999; Sinkins, 2004; Bourtzis et al., 2014). Certain combinations of Wolbachia strains and mosquito species suppress propagation and transmission of dengue viruses and malaria parasites in mosquito hosts (Bourtzis et al., 2014). However, in nature the host range of Wolbachia does not include some of the most important vectors of human pathogens, such as Anopheles stephensi, A. gambiae, and Aedes aegypti (Hilgenboecker et al., 2008). This host range has been expanded through forced infections by inoculations into mosquito embryos, resulting in stable infection across subsequent generations in A. aegypti and A. stephensi with little fitness cost (Xi et al., 2005; Bian et al., 2013; Joshi et al., 2014). A recent study found strongly negative interactions between the commensal gut microflora in mosquitoes and propensity of the mosquitoes to harbor Wolbachia infection after intrathoracic inoculation, which might explain why natural infections of Wolbachia in some mosquito species are not found in nature (Hughes et al., 2014). Hughes et al. (2014) reported that Wolbachia vertical transmission was achievable in Anopheles species only when the gut microbiota (especially, Asaia bacteria) were perturbed (i.e., eliminated) with antibiotics. Furthermore, Rossi et al. (2015) observed that Wolbachia failed to colonize the female reproductive organs of anophelines, due to presence of Asaia infection in those tissues (Rossi et al., 2015). The conclusion of both studies is that commensal microbiota impede Wolbachia vertical transmission in mosquitoes (Hughes et al., 2014; Rossi et al., 2015). However, what remains to be determined is whether the commensal microbiota of mosquitoes that have been stably infected trans-generationally with Wolbachia by embryonic injection differ from that of Wolbachia-uninfected mosquitoes of the same species and held under the same conditions. Based on the results of studies reviewed here (Hughes et al., 2014; Rossi et al., 2015), one would predict that stably infected mosquitoes must differ in their microbiota due to the inhibitory effects of Asaia and other gut symbionts on Wolbachia infection. Because stably infected mosquitoes are those most likely to be used in attempts to control vector populations or pathogen transmission, the interactions between stable Wolbachia infection and gut microbiota clearly become important.

After the wAlbB Wolbachia strain was successfully transferred from A. albopictus into A. stephensi by embryonic microinjection, this strain of Wolbachia spread rapidly into a laboratory population of A. stephensi by vertical transmission over the course of several generations without any manipulation of gut microbiota (Bian et al., 2013). Further, wAlbB infection conferred partial resistance to development of the malaria parasite Plasmodium falciparum in A. stephensi (Bian et al., 2013), and significantly up-regulated the mosquitoes’ innate immune response, including generation of reactive oxygen species from midgut epithelial cells (Pan et al., 2012). One way to increase the inhibitory effect of Wolbachia on virus or parasite development in mosquito hosts is to increase the intensity of Wolbachia infection (Correa and Ballard, 2014). Another strategy is to couple paratransgenesis of gut microbiota with a Wolbachia-based vector population replacement/suppression strategy. The latter concept is promising because certain symbionts such as Asaia and Pantoea agglomerans, when engineered to express molecules having anti-malaria parasite properties, result in reduced malaria parasite load in host mosquitoes (Wang et al., 2012; Bongio and Lampe, 2015). However, the innate immune response triggered by Wolbachia infection is non-specific and it may affect the commensal microflora (Pan et al., 2012). Thus, it remains unclear if introduced or genetically modified bacteria will persist in Wolbachia-infected mosquitoes due to elevated innate immune response and putative incompatibility between Wolbachia and gut microbiota. If Wolbachia wAlbB infection with its non-specific immune effects in A. stephensi, and engineered gut microbiota with specific anti-parasite effects, are found to be compatible; then a synergistic, dual strategy could be developed to suppress malaria parasite or virus development in mosquitoes. In this study, and as a step toward determining the extent of this compatibility, we quantified the diversity of the bacterial community in stably Wolbachia-infected and Wolbachia-cured A. stephensi mosquitoes. Moreover, because of the dynamics of microbiota associated with key life history events in mosquitoes, we further compared bacterial communities when sugar or meals were provided, and when the mosquitoes were young or aged.

The bacteria (i.e., microbiota) associated with mosquitoes have significant roles throughout the mosquito lifecycle, in providing energy and nutrients, and regulating development, fecundity, and immune responses (Sinkins, 2004; Chouaia et al., 2012; Pan et al., 2012; Coon et al., 2014). These symbiotic bacteria find shelter and nutrients within this environment and could properly be referred to as commensals for these beneficial reasons, even if the beneficial associations of the bacteria to mosquitoes are not entirely clear. Stable, intracellular Wolbachia infection in A. stephensi (LB1) was achieved previously through embryonic inoculation and multi-generational, vertical passage (Bian et al., 2013). It was done without perturbation of the natural microbiota (as shown here) and without excessive mortality after blood meals (Bian et al., 2013; Joshi et al., 2014). More recently and by contrast, vertical persistence of Wolbachia infection in A. stephensi was achieved only when symbiotic bacteria were perturbed by antibiotics (Hughes et al., 2014). In that same study, a combination of Wolbachia and Asaia caused mosquito mortality after a blood meal, a finding suggestive of a physiologic incompatibility between Wolbachia infection and presence of a constitutive microflora when the stressors of the blood meal arrive (Hughes et al., 2014). Indeed, if such an incompatibility exists, it would obviate the possibility of a dual, anti-malaria parasite strategy based on immune-mediating effects of Wolbachia infection and paratransgenesis of select members of that microflora. This possibility, in part, motivated the research presented here.

The similarity of the gut bacterial community structure, as measured by composition of taxa as well as by α- and β-diversity, between Wolbachia-infected LB1 and Wolbachia-cured LBT mosquitoes indicates that there were no significant Wolbachia-mediated effects on the structure of the commensal bacterial community; and that the bacterial assembly was resilient to any effects, such as innate immune effects, that the presence of Wolbachia might impose. Therefore, coexistence of these gut bacteria with Wolbachia was not problematic in the stable Wolbachia-infected Anopheles progeny. These findings contrast with other studies demonstrating very strongly negative effects of combination of gut bacteria and Wolbachia on host fitness (high mortality rate after a blood meal), particularly due to Asaia bacteria (Hughes et al., 2014). One explanation for the difference is in the way the Wolbachia infection was originally introduced into the mosquitoes: embryonic microinjection followed by stable, trans-generational perpetuation of infection (Bian et al., 2013); or intrathoracic inoculation without perpetuation (Hughes et al., 2014). The latter likely led to a larger and more acute dose of the bacteria, than one inherited in the germ line. However, the detailed mechanisms need to be further investigated.

The blood meal profoundly affects bacterial community composition in the midgut of several mosquito species, including A. gambiae, A. stephensi, and A. coluzzii (Wang et al., 2011; Gimonneau et al., 2014; Tchioffo et al., 2016). The main influencing factors included dramatic changes in mosquito gut temperature and appearance of oxidative stress (Luckhart et al., 1998; Peterson and Luckhart, 2006; Benoit et al., 2011). Especially, the ingested red blood cells from animals carry an enormous amount of hemoglobin, which causes a massive release of heme in the midgut, thus leading to a dramatic change in gut conditions (Wang et al., 2011; Tchioffo et al., 2016). Blood meals induce midgut epithelia to produce nitric oxide which is also a source for free radicals (Luckhart et al., 1998). Hematophagous mosquitoes exhibit detoxification mechanisms in response to these free radicals, but the role of commensal bacteria in the gut in that same process and how those bacteria survive them are not known (Peterson and Luckhart, 2006; Benoit et al., 2011). After blood meals, Enterobacteriaceae dramatically increased, which was consistent with those reported in A. gambiae (Wang et al., 2011). Wang et al. (2011) found several stress response systems existing in the Enterobacteriaceae genomes. Possibly, this group of bacteria interacts with the mosquito host in response to the oxidative stressors present in the blood meal bolus (Wang et al., 2011). Therefore, blood-induced mortality in transiently Wolbachia-infected Anopheles mosquito could be caused by a combinative effect, that is, innate immune response elicited by Wolbachia infection and by dysregulated microbiota (Luckhart et al., 1998; Pan et al., 2012; Hughes et al., 2014). Both of them caused an unusual high level of active radicals in midgut, which can cause destructive damage to mosquitoes (Luckhart et al., 1998). Although there was a dramatic change in microbial structure after blood meals (compared to sugar meals), the core microbial compositions and community structure between Wolbachia-infected and Wolbachia-cured mosquitoes were rather similar (Figure 3).

When switching diet from blood to sugar meals and aging the insects, we found that the microbial community structure was different between the young (before blood meal) and aged mosquitoes (post-blood meal) in both mosquito lines. However, the relative abundance of transiently increased bacteria such as Enterobacteriaceae decreased when blood digestion was complete; the dominant Elizabethkingia abundance at age 15 days, and after the blood meal, was similar to before the blood meal (Figure 2B). The dominant bacterial community patterns we observed here were similar to those reported in A. gambiae (Wang et al., 2011). Collectively, Asaia, Elizabethkingia, and Enterobacteriaceae were present and predominant in aged, blood-fed, Wolbachia-infected A. stephensi, providing the possibility to devise an efficient malaria control reagent based on a combination of effects of Wolbachia and effects of paratransgenic gut microbiota (Federici et al., 2003; Favia et al., 2008; Cirimotich et al., 2011; Boissière et al., 2012; Capone et al., 2013; Bongio and Lampe, 2015; Chen et al., 2015).

Previous studies demonstrated that Wolbachia infected the midgut in A. stephensi and A. gambiae mosquitoes though its distribution level was much lower than that in the fat body (Bian et al., 2013). After a blood meal, the dramatic decrease of Wolbachia OTUs in LB1 gut was consistent with a previous observation by Hughes et al. (2014). Several factors including heme release, ROS level change, and interactions with gut bacteria contributed to this change (Pan et al., 2012; Bian et al., 2013; Hughes et al., 2014). However, after removing gut bacteria by antibiotic treatment (especially Asaia), Wolbachia persisted at a level similar to that non-blood fed mosquitoes, indicating gut microbiota suppressed Wolbachia level after a blood meal (Hughes et al., 2014).

Operational taxonomic units of Asaia (up to 8.9% of the total taxa) were consistently detected in Wolbachia-infected A. stephensi guts here. The relative abundance of Asaia in LBT or LB1 mosquitoes was similar regardless of Wolbachia infection, diet switching and rearing ages. This observation differs from the study conducted by Hughes et al. (2014) where Asaia concentration was much higher in blood fed mosquitoes than in non-blood fed ones. The initial assembly of the gut microbiota might account for this discrepancy: Asaia, Pseudomonas, and unclassified Gammaproteobacteria were the most predominant bacteria (Hughes et al., 2014). Asaia is a common commensal of several mosquito species such as Aedes, Anopheles, and Culex (Favia et al., 2008; Rossi et al., 2015). Its physiological roles in mosquitoes have been suggested to be nutrient scavenging and regulation of larval development (Chouaia et al., 2012). Recent studies showed antagonism between Asaia and Wolbachia in the reproduction organs (such as ovary; Rossi et al., 2015). Further, antagonistic interactions between Wolbachia and the gut microbiota, particularly Asaia, resulted in significant fitness costs in the mosquito host (Hughes et al., 2014). Our observation here showed that Asaia was stably maintained in both LB1 and LBT guts, which was consistent with Rossi et al. (Rossi et al., 2015).

The relative abundance of Elizabethkingia OTUs accounted for more than 60% of the total microbial taxa, highlighting their importance in Anopheles mosquitoes. Presumably, they contribute to metabolism of sugars in the gut environment, as evidenced by the SusC/SusD-like polysaccharide transport system(s) and glycosidases in their genome (Kukutla et al., 2014). Supplementation of Elizabethkingia cells in the sugar meal for A. stephensi led to approximately 50% more egg production compared to those without supplementation, suggesting that Elizabethkingia contributed to animal erythrocyte lysis and thus increased host’s fecundity in mosquitoes (S. Chen, unpublished). Elizabethkingia was frequently detected in water, sediments and insects including field caught, semi-natural reared, and insectary reared mosquitoes, showing that it adapted to diverse ecological niches in nature, but it is commonly found in Anopheles midguts (Chen et al., 2015). Pyrosequencing analysis showed that Elizabethkingia OTUs were more abundant in larval A. gambiae mosquitoes than in the aquatic medium in which they were being reared, but were present in the latter suggesting a common environmental source (Wang et al., 2011). Furthermore, Elizabethkingia anophelis and its nearly identical taxon, Elizabethkingia meningoseptica, were among the most predominant bacteria in adult A. gambiae and A. stephensi (Wang et al., 2011; Akhouayri et al., 2013; Ngwa et al., 2013).

SC and EW wrote the manuscript. DJ and SC performed the experiments. SC, JZ, ZX, BN, and EW performed the data analysis.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer HS and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.