Estimating bacteria diversity in different organs of nine species of mosquito by next generation sequencing

Background Symbiosis in insects is accumulating significant amount of studies: the description of a wide array of mutualistic associations across the evolutionary history of insects suggests that resident microbiota acts as a driving force by affecting several aspects of hosts biology. Among arthropods, mosquito midgut microbiota has been largely investigated, providing crucial insights on the role and implications of host-symbiont relationships. However, limited amount of studies addressed their efforts on the investigation of microbiota colonizing salivary glands and reproductive tracts, crucial organs for pathogen invasion and vertical transmission of symbiotic microorganisms. Using 16S rRNA gene sequencing-based approach, we analysed the microbiota of gut, salivary glands and reproductive tracts of several mosquito species, representing some of the main vectors of diseases, aiming at describing the dynamics of bacterial communities within the individual. Results We identified a shared core microbiota between different mosquito species, although interesting inter- and intra-species differences were detected. Additionally, our results showed deep divergences between genera, underlining microbiota specificity and adaptation to their host. Conclusions The comprehensive landscape of the bacterial microbiota components may ultimately provide crucial insights and novel targets for possible application of symbionts in innovative strategies for the control of vector borne diseases, globally named Symbiotic Control (SC), and suggesting that the holobiont of different mosquito species may significantly vary. Moreover, mosquito species are characterized by distinctive microbiota in different organs, likely reflecting different functions and/or adaptation processes. Electronic supplementary material The online version of this article (10.1186/s12866-018-1266-9) contains supplementary material, which is available to authorized users.


Background
Insects represent the most diverse and abundant clade of metazoans, accounting for more than 90% species and dominating a variety of terrestrial habitats. Diverse insect habits are founded on associations with microorganisms, whose diversity reflects the variety of their hosts [1]. These complex associations are passed between generations and can assume mutualism or commensalism features. The most important distinction is whether this microbiota is transient (acquired from the surrounding environment) or it is indigenous to its host, maintaining stable communities and colonizing the gut habitat [2]. Various insects host resident microorganisms able to influence many insect's key functions, although not primarily. Diverse physiological, metabolic and immune processes of hosts are significantly influenced by their microbiota, whose colonization significantly contributed to their evolutionary success. Coexistence and coevolution may rely on increased microbial production of nutrients valuable to the host and correlated changes in its metabolism, promoting host complementarity to microbial nutrients [3]. An elegant study describes symbiotic bacteria of Drosophila melanogaster capable of affecting mating preferences by changing the levels of cuticular hydrocarbons and sex pheromones [4]. Additionally, endosymbionts were found to influence dispersal behaviour, species distribution and overall community composition in spiders [5,6]. Symbiotic bacteria also impact on body colours of aphids, influencing prey-predator interactions, as well as interactions with other endogenous endosymbionts. Interestingly, insect life cycles are also dependent on microbiome: some mosquito species were reported to rely upon their gut microbiota for molting and developing from larvae to adults [7]. In addition, microbiota contribution on the well-being of the host is also revealed by its ability to promote the resistance of insects to certain natural enemies, including viruses, bacteria, fungi, nematodes, parasites, and in turn, to modulate vector capacity of some insects involved in transmission of infectious diseases. Moreover, symbionts are shown to contribute to insecticide resistance phenomena [8].
In this frame, among insects, mosquitoes represent an important public health challenge in many parts of the globe. Anopheles mosquitoes are vectors of human pathogens, including parasites responsible of severe infectious diseases (malaria and lymphatic filariasis) and arboviruses (O'nyong-nyong virus). In sub-Saharan Africa, the main human malaria vectors are members of the An. gambiae complex, together with An. coluzzii. An. stephensi, instead, represents one of the major malaria vector in Asia.
Mosquitoes of the genus Aedes are worldwide involved in outbreaks of arboviruses, including Dengue, Chikungunya and Zika viruses [9]. Aedes mosquitoes are known to be highly invasive with a great capacity to adapt to contrasting climates and environments. Mosquitoes belonging to the Culex genus were found to be involved in the transmission of human and animal diseases, such as West Nile fever, often acting as a bridge vector between birds and humans [10]. Additionally, these genera exhibit different and specific rhythmic behaviours across their life cycle. For instance, the majority of Anopheles and Culex species are mainly characterized by nocturnal biting behaviour, whereas Aedes typically engage biting activities in the daytime. They also differ in the choice of the breeding site: Anopheles mostly prefer clear water exposed to sunlight, Culex and Aedes are instead mostly found in dark water containing organic matter [11].
Previous studies have described midgut microbiota of mosquitoes as a driving force to directly and/or indirectly affects host-pathogen interactions, and ultimately vector capacity, significantly influencing diseases transmission. In Anopheles mosquitoes, resident microbiota inhibits the invasion of ookinetes of the malaria parasite Plasmodium of the midgut epithelium, reducing infection rate: this effect is mediated by colonizing bacteria, and not caused by their direct interaction with Plasmodium [12]. Also, arboviruses transmission undergoes microbiota-mediated modulations within the mosquito, mainly due to its indirect influence on nutrient catabolism, development, and immune responses [7,13,14]. This is particularly remarkable with viruses because of their strict dependence on host factors for invasion and replication processes. Furthermore, microbiota can directly interact with arboviruses by inhibiting viral transmission through the secretion of anti-viral compounds [15]. On the other hand, bacteria could also act as pro-factors by increasing vector permissibility to arboviral infections [16,17].
Microbiota in mosquitoes is often restricted to bacteria associated to the gut. This tissue represents a reservoir of a wide variety of microbial communities, whose characterization is essential and preparatory for a solid understanding of the overall vector biology [18]. Nevertheless, other organs result to be inhabited by microorganisms: some taxa are shared with the intestinal tract, while some others showed preferences for different tissues. Only recently, a comprehensive study addressed the question about the various and diverse composition of the microbiota of midguts, ovaries, and salivary glands of female individuals of An. gambiae and An. coluzzii, using pyrosequencing analysis of the 16S rRNA gene [19]. Additionally, the analysis of the salivary glands of An. culicifacies revealed to host a more complex and various microbiome than that inhabiting the gut [20], underlying diverse bacterial colonization abilities, and in particular the intrinsic environmental divergence offered by organs, despite belonging to the same individual.
We investigated nine species of mosquitoes belonging to the main representative mosquito genera, in terms of geographical diffusion and public health interest. This includes six anopheline species [An. gambiae (G3 and Kisumu strains), An. coluzzii, An. arabiensis, An. quadriannulatus, An. merus, An. stephensi], two belonging to the Aedes genus (Ae. aegypti, Ae. albopictus) and Cx. quinquefasciatus. Herein, data of an extensive study aimed to characterize bacterial microbiota associated to gut, salivary glands and reproductive organs are reported. The importance of describing the microbiota of these organs lies in possibility to develop innovative paratransgenic approaches for vector control. The selection of suitable symbiont candidates primarily requires a specific tissue tropism in the gut, the salivary glands and the reproductive organs [21], since i) the transmission cycle of many pathogens starts in the intestinal tract and culminates with the invasion of the salivary glands from where they are transmitted to a new host and ii) the colonization of the reproductive organs allows the vertical transmission of bacteria to the offspring, ensuring their persistence within a population.

Overall distribution of bacteria within different organs
A total of 951.930 reads were generated after removal of short reads (< 250 bp), chimeras and the discard of spurious OTUs [22] from all tissues analyzed. The number of reads varied among samples (minimum = 2228, maximum = 62,851) ( Table 1), therefore the number of sequences for each sample was rarefied to the minimal readings of 2228. The sequences were clustered in 924 bacterial OTUs ( Table 1). Analysis of the rarefaction curves indicated an adequate sampling quality; only in a few cases, the curves did not reach the saturation to an asymptote, indicating the presence of additional rare bacteria taxa or further spurious OTUs (Additional file 1: Figure S1). Furthermore, the number of observed species estimator was used to evaluate bacteria richness between the mosquito species/strains (Additional file 2: Figure S2). Interestingly, irregular diversity in terms of bacterial abundance is highlighted when indices are compared. Bacterial richness of reproductive organs is higher than that found in the other tissues (Table 1). In most samples, salivary glands harbour a more diverse microbiota than the gut. In particular, Ae. aegypti harbours the highest number of bacteria species compared to others, followed by An. coluzzii and An. quadriannulatus. An. gambiae G3 and An. stephensi contain the lowest number of bacteria species, explained by the long adaptation in laboratory conditions (Additional file 2: Figure S2). In total, mosquito tissues (guts, reproductive organs and salivary glands) host a flora consisting of bacterial OTUs belonging to 12 phyla, 74 families and 121 genera. Among these, 4 phyla represent more than 99% of the total microbiota: Proteobacteria, Bacteroidetes, Firmicutes and Actinobacteria ( Fig. 1, Additional file 3: Table S1). The OTUs assigned to other not abundant phyla (< 1%) included: Fusobacteria, Acidobacteria, Cyanobacteria, Chlamydiae, Deinococcus-Thermus, Elusimicrobia, Planctomycetes and SHA-109. Some bacteria could not be assigned to any taxa (0.1%). The phylum Proteobacteria is dominant in the majority of tissues, showing to be the main constituent of a shared and conserved microbiota core (Fig. 1). Striking examples of Proteobacteria abundance (> 60%) are found in the following organs: i) guts of females; ii) guts of males, except for An. gambiae G3 and An. coluzzii; iii) salivary glands of all species apart from An. gambiae G3 and An. coluzzii; iv) the reproductive organs of both males and females of An. arabiensis, An. merus, An. stephensi, Ae. albopictus and Ae. aegypti, the ovaries of An. gambiae G3 and organs of Cx. quinquefasciatus males. However, Bacteroidetes (> 68%) are present in male guts and salivary glands of both An. gambiae G3 and An. coluzzii, and in the reproductive organs of An. gambiae G3 males. The guts of An. gambiae Kisumu males harbour more than 28% of Bacteroidetes, which also encompasses 10% of taxa in Ae. aegypti organs. Firmicutes and Actinobacteria strongly colonize the reproductive organs of both males and females of An. coluzzii, An. quadriannulatus and An. gambiae Kisumu, whereas they are found only in ovaries of Cx. quinquefasciatus.
Consequently, a principal coordinate analysis (PCoA) plot was generated, revealing that the reproductive organs of An. stephensi males and An. gambiae G3 females are divergent from other samples (Fig. 2a). This sharp difference is mostly caused by abundant presence of Serratia. Moreover, the reproductive organs of An. gambiae G3 males show a unique cluster due to the large presence of Elizabethkingia. The relevant presence of members of the Elizabethkingia genus also defines a cluster in the salivary glands of An. gambiae G3 and An. coluzzii (Fig. 2 b); an additional group is composed by guts of An. gambiae G3 males and An. coluzzii females (Fig. 2c). The rest of the samples tends to cluster together ( Fig. 2a-b-c ). A global representation is described in the merged picture (Fig. 2d).
Gammaproteobacteria and Betaproteobacteria is abundant in most samples, except for Gammaproteobacteria in Cx. quinquefasciatus males (< 1%) and for Betaproteobacteria in Cx. quinquefasciatus males and females, An. stephensi and An. gambiae Kisumu females (Fig. 3a).
On the contrary, reproductive organs of the species of the An. gambiae complex harbour a dissimilar microbiota: the class of Bacilli is predominant, followed by Gammaproteobacteria, Actinobacteria, Betaproteobacteria and Fusobacteria. However, An. gambiae G3 represents an exception, where Gammaproteobacteria (85%) in females and Flavobacteria (74%) in males dominate the overall microbiota (Fig. 3a).
Despite a few exceptions, the analysis of taxa at genus level showed the persistence of a shared core of bacteria. An. gambiae G3 ovaries are mainly colonized by Serratia (78%) followed by Escherichia-Shigella (0.5%), Sphingomonas (0.2%), Cupriavidus (0.2%) and Elizabethkingia (0.2%). The latter, on the contrary, represents the dominant taxon (73%) in male organs, co-habiting with Burkholderia (18%) and, in much lower amount, with Serratia (0.3%). Similarly, in An. stephensi, Serratia predominantly inhabits male reproductive organs, while in ovaries it is exceeded by Asaia, whose presence in males is reduced. Finally, Cx. quinquefasciatus offered a divergent scenario: male reproductive organs are mainly colonized by Phyllobacterium (91%), although its contribution in ovaries microbiota was minimal (0.8%) (Fig. 4a). Differently, the ovaries of Cx. quinquefasciutus host    Serratia, and several genera of the families Rhodobacteriaceae and Rhizobiaceae (Additional file 4: Figure S3). As expected, Wolbachia colonization in Ae. albopictus shows a noticeable predominance in ovaries (94%); in male reproductive organs, it coexists with Sphingomonas, Cupriavidus and Serratia. Finally, the reproductive organs of Ae. aegypti show an interestingly high richness of taxa: Ochrobactrum, Pseudomonas, Sphingomonas, Stenothropomonas, Acinetobacter and Elizabethkinghia are the most abundant genera detected in female and male individuals (Fig. 4a).

Microbiota composition of salivary glands
Female salivary glands are also collected and analysed. Gammaproteobacteria represents the most prevalent class in anophelines (An. quadriannulatus, An. gambiae Kisumu, An. arabiensis, and An. stephensi) and in Ae. albopictus, and, although less abundant, in An. merus, in Ae. aegypti and in Cx. quinquefasciatus. Alphaproteobacteria, followed by Betaproteobacteria, constitutes the remaining portion (Fig. 3b).
Serratia is present in each sample of salivary glands, and together with Escherichia-Shigella, Pantoea, Acetobacter, Sphingomonas, Burkholderia and Cupriavidus is part of a shared core of taxa. An. quadriannulatus, An. gambiae Kisumu, An. arabiensis are prevalently colonized by Serratia and Asaia. An. merus hosts Sphingomonas (34%), Asaia (14%) and in lower proportion Serratia and Cupriavidus. A completely divergent pattern is visible in An. gambiae G3 and An. coluzzii mosquitoes, where 69% of OTUs are addressed to the Elizabethkingia sp. genus, whose abundance reaches almost undetectable levels in the salivary glands of other species (apart from 1.4% in Ae. aegypti). In addition, salivary glands of An. gambiae G3 and An. coluzzii are inhabited by Betaproteobacteria (genus Burkholderia, > 25%), whose colonization is found in lower amount (7.6%) in An. quadriannulatus samples.

Discussion
Through a 16S rRNA gene sequencing-based approach, a description of microbiota diversity hosted in salivary glands, reproductive tracts and guts of several mosquito species is provided. High richness and diversity of microbes associated with laboratory reared individuals was detected, suggesting the persistence of a well-established and conserved core of bacteria. The evidence of a shared microbiota core is consistent with other studies reporting similar data and identifying a large conserved group of bacteria colonizing different mosquito species and their tissues, also in standardized rearing environments. [19,23,24].
We are aware that bacterial DNA contamination in extraction kits and laboratory reagents can significantly affect the results of microbiota studies, particularly when samples contain low microbial biomass and when highly sensitive techniques are used. Previous studies described the list of contaminants detected in sequenced negative blank controls derived from reagents contamination, users and extraction kits [25]. The proposed list of contaminants has been compared to the obtained data. Only a few candidates, like Mesorhizobium, Phyllobacterium, Rhizobium, Comamonas, Delftia, Variovorax, Escherichia-Shigella were detected at very low percentages in our samples. It has been also reported that Corynebacterium, Propionibacterium and Streptococcus are common human skin-associated organisms. Only Corynebacterium and Streptococcus were detected in very low percentage (< 1%) in some samples. Overall, the incidence of bacteria contaminations in our data appears minimal; moreover, the taxa described above have been previously presented in metagenomics studies of insects, and for this reason they have been included in the data analysis.
Previous studies proved that commensal bacteria are mainly acquired from the environment or transmitted directly between hosts to their offspring. Often, these contributions are correlated, meaning that bacteria with complete transmission among hosts will evolve specialization for a particular niche [26][27][28]. Although bacterial composition is selected by standardized diet and rearing conditions, differences have been disclosed and data are able to provide crucial insights for a robust characterization and comparison between species. Additionally, despite belonging to the same individual, tissues specifically harboured diverse bacteria, whose distribution reflects tissue-bacteria adaptations. Within the same individual, microbiota of reproductive organs was more diverse than that of the gut and salivary glands. Salivary glands, mostly in anophelines, showed higher diversity indices when compared to the guts, similarly to what reported in An. culicifacies [20]. A leading example of this characteristic adaptation is Elizabethkingia: it largely colonizes guts and reproductive organs of An. gambiae G3 males, while in An. coluzzii, it is prevalent in guts of males and salivary glands of females. Its association with guts of An. gambiae was already reported in mosquitoes reared in standard insectary conditions, suggesting the establishment of an own thriving niche, and possibly explained with the ability of Bacteroidetes to degrade sugar [29]. Similarly, Serratia, although presenting variable abundances in each sample, is found to be dominant in the gut of female anophelines, being consistent with recently reported data [19,30]. For instance, in An. stephensi, Serratia represents more than 90% of detected taxa in all organs, except for female ovaries where its abundance is reduced.
Notably, Asaia occurrence is in line with its role among symbiotic bacterial communities: it rarely showed, in fact, dominant behaviour, like Serratia or Elizabethkingia, but its persistence is found in guts and salivary glands of all mosquito genera and its relevant occurrence in reproductive organs of males and females of An. stephensi and An. arabiensis provides evidences of its widespread colonization ability.
Bacteria such as Asaia, Pantoea, Pseudomonas, Entrobacter, Serratia are gaining momentum as promising candidates for paratransgenic modifications for vector control strategies. Our findings confirm their abundance in key tissues of various mosquito species, corroborating and widening their role as broad-spectrum tools against mosquito borne diseases. Cupriavidus and Ochrobactrum could be also included in this bacterial arsenal: their association with mosquitoes was never reported before, but their recent engineering for biomedical purposes could be also translated in the field of mosquito control [31,32]. In addition, the presence of some bacteria as Serratia and Wolbachia in different mosquito species are well-known to provide protective effects from pathogens infections, as reported in several studies [33,34].
In any case, before any field application of bacterial strains, a detailed characterization and analysis of the potential risks for human and animal health is an imperative.

Conclusions
Our study describes the composition of microbial communities harboured in different tissues of nine mosquito species, transmitting devastating human diseases. The results highlight the importance of a comprehensive understanding of organisms, intended as holobionts, and thus composed by innumerable interacting communities, whose impact is increasingly apparent. The identification of interesting inter-and intra-species differences, coexisting with a shared core microbiota, suggest a singular adaptation and a tissue-specific tropism. The contextual findings of a deep degree of divergences between genera, underlines microbiota specificity and its adaptation to the host. Moving forward, our data lay the basis for the design of effective vector control strategies, relying on the use of mosquito-associated bacteria.

Mosquito strains
The following laboratory strains were used: An

Mosquito rearing
Analyzed mosquitoes were all cyclic laboratory-strain colonies. Mosquitoes were reared with a 12 h day/night cycle at 27°C and relative humidity of 70%. Anopheles, Aedes and Culex larvae were reared in deionized water with 0.3 g/liter of artificial sea salts and fed daily with a diet composed by a slurry of 2:2:1 bovine liver powder, tuna meal and Vanderzant vitamin mix [35]. Adults were fed ab libitum on 10% sucrose.

Tissues collection and DNA extraction
Prior dissecting, mosquitoes were surface sterilized in 70% ethanol for 5 min and rinsed twice in sterile PBS. Foreguts and midguts, salivary glands lobes and reproductive organs, consisting on testes and male accessory glands (MAGs) for males, and ovaries for females, were dissected with special care to reduce contaminations: needles and dissecting slides were sterilized and treated with 70% ethanol between each sample, while storing vials were UV treated before use. Pooled guts (10 whole guts), reproductive organs (10 sets) and salivary glands (20 pairs) were homogenized with sterile 0.5-mm wide glass beads (Bertin) for 30s at 6800 rpm in the automatic tissue homogenizer (Precellys 24, Bertin). Genomic DNA was extracted using a JetFlex Genomic DNA Purification kit (Genomed) according to the manufacturer's instructions.

Bioinformatic analysis
Quality control of raw data was done using FastQC, then adapters sequences and low quality scores (< 20) were trimmed by Trimmomatic software package [38]. Cleaned sequenced paired-end reads were merged to reconstruct original full-length 16S amplicons with PEAR software [39]. All amplicons with sequence similarity higher than 97% were chosen as input for making the taxonomy annotation and building the OTU table by following the Open Reference approach. The obtained sequences were searched for matching in the SILVA taxonomy database (v123) using similarity-searching algorithm [40]. The microbial communities of each sample were built from the taxonomy assignments. The alpha-diversity (of each sample) was investigated by means of three different indexes: Shannon and Simpson observed OTUs, in order to quantify the number and distribution of taxa in each sample [41][42][43]. The dissimilarity analysis between samples measuring population composition (beta-diversity) has been evaluated using QIIME weighted Unifrac distances between samples at a sub-sampling depth of 100 sequences per sample. Thus, jackknifed principal coordinates have been calculated on rarefied results to compress dimensionality into two-and three-dimensional principal coordinate analysis plots. The rarefied results have been used also for computing alpha diversity using observed species, Shannon and phylogenetic diversity (PD) metrics. Data analysis was done using Qiime and the R-software (http://qiime.org, http://www.r-project.org/).