Zika is a re-emerging infection that has been considered a major threat to global public health. Currently at least 100 countries are at risk of Zika virus (ZIKV) transmission. Aedes aegypti is the main mosquito vector in the Americas. This vector is exposed to, and interacts symbiotically with a variety of microorganisms in its environment, which may result in the formation of a lifetime association. Here, the unknown effect that ZIKV exerts on the dynamic bacterial community harbored by this mosquito vector was investigated using a metagenomic analysis of its microbiota. Groups of Ae. aegypti were experimentally fed on sugar, blood and blood mixed with ZIKV, and held for 3 to 7 days after blood meal and eggs development respectively. The infected groups were processed by qPCR to confirm the presence of ZIKV. All groups were analyzed by metagenomics (Illumina Hiseq Sequencing) and 16S rRNA amplicon sequences were obtained to create bacterial taxonomic profiles. A core microbiota and exclusive bacterial taxa were identified that incorporate 50.5% of the predicted reads from the dataset, with 40 Gram-negative and 9 Gram-positive families. To address how ZIKV invasion may disturb the ecological balance of the Ae. aegypti microbiota, a CCA analysis coupled with an explanatory matrix was performed to support the biological interpretation of shifts in bacterial signatures. Two f-OTUs appeared as potential biomarkers of ZIKV infection: Rhodobacteraceae and Desulfuromonadaceae. Coincidentally, both f-OTUs were exclusively present in the ZIKV- infected blood-fed and ZIKV- infected gravid groups. In conclusion, this study shows that bacterial symbionts act as biomarkers of the insect physiological states and how they respond as a community when ZIKV invades Ae. aegypti. Basic knowledge of local haematophagous vectors and their associated microbiota is relevant when addressing transmission of vector-borne infectious diseases in their regional surroundings.
Citation: Villegas LEM, Campolina TB, Barnabe NR, Orfano AS, Chaves BA, Norris DE, et al. (2018) Zika virus infection modulates the bacterial diversity associated with Aedes aegypti as revealed by metagenomic analysis. PLoS ONE 13(1): e0190352. https://doi.org/10.1371/journal.pone.0190352
Editor: Pedro L. Oliveira, Universidade Federal do Rio de Janeiro, BRAZIL
Received: August 14, 2017; Accepted: December 13, 2017; Published: January 2, 2018
Copyright: © 2018 Villegas et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This study was funded by the following: Minas Gerais State Research Support Foundation (00240-13) Paulo Pimenta, www.fapemig.br; Minas Gerais State Research Support Foundation (APQ 03002-16) Paulo Pimenta, www.fapemig.br; Brazilian Council for Science and Technological Development CNPq/ DECIT (440615/2016) Paulo Pimenta, www.cnpq.br; Brazilian Council for Science and Technological Development CNPq (308849/2015-9) Paulo Pimenta, www.cnpq.br; Brazilian Council for Science and Technological Development CNPq (Science Without Borders (400704/2014-5) Nagila Secundino, www.cnpq.br; Brazilian Council for Science and Technological Development CNPq (305512/2014-5) Nagila Secundino, www.cnpq.br; Brazilian Council for Science and Technological Development CNPq- INCT de Entomologia Molecular (465678/2014-9) Nagila Secundino, www.inctem.bioqmed.ufrj. br; Minas Gerais State Research Support Foundation (00242-12) Nagila Secundino, www.fapemig.br. This work was also funded in part by the Foundation of the Institute Oswaldo Cruz (FIOCRUZ), Programme Science without Borders, INCT-Instituto Nacional de Ciencia e Tecnologia—Entomologia Molecular, and Amazonas State Research Support (FAPEAM). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Zika is a re-emerging disease, with a recent outbreak in Brazil in 2015 that resulted in Zika being considered a major threat to global public health. Currently at least 100 countries are at risk of Zika virus (ZIKV) transmission (http://www.cdc.gov/zika/geo/active-countries.html). Aedes aegypti, the main mosquito vector in the Americas, has worldwide distribution across tropical and subtropical regions resulting in almost half of the world’s population at a risk of infection . Transmission of ZIKV as well as other arboviruses, such as the Chikungunya and dengue viruses by mosquitoes in Brazil occurs steadily throughout the year in most regions, favoring their rapid dissemination .
Insects are exposed to, and interact symbiotically with a variety of microorganisms , some of which they acquire from their ecosystems and through feeding resulting in an association that may remain throughout their life spans [4,5]. Several key members of the native microbiota are transstadially transmitted during multiple life stages of their host [5,6]. To date, bacteria are most studied component of the microbial community in mosquito models [7–9]. This community may fluctuate in both richness and abundance during the developmental and adult phases of the mosquito [10, 11], with some bacterial OTUs (operational taxonomic units) that remain associated but dynamically shift towards a high abundance in response to physiological conditions [11–13]. In addition, studies have shown that disrupting the homeostasis of microbiota, for example, through antibiotic therapy, affects susceptibility towards pathogens, including the models Anopheles gambiae and Ae. aegypti when exposed to Plasmodium falciparum and dengue virus, respectively [14–16]. Although the metabolic or physiological role that autochthonous bacteria play within arthropod vectors, as well as their potential interactions with human pathogens are not fully understood, bacteria may directly or indirectly modulate vector competence, and have an impact on vector bionomics [8,14–18].
The association between microbial consortia and arthropod hosts has been described for a large variety of vector models [8,19]. Herein, we explored the potential effect that ZIKV exerts on the dynamic bacterial community harbored by the main mosquito vector Ae. aegypti. For this purpose, we performed the metagenomic taxonomic profiling of the Brazilian Ae. aegypti mosquito under distinct physiological conditions, including ZIKV infection. The goal was to determine whether ZIKV infection causes a noticeable ecological disruption of the profile of bacterial abundance in the holobiont . To address this, we performed a constrained comparison of community structure using β-diversity as a measure of the variance within a community data matrix [21,22]. The multivariate analyses revealed how the bacterial community profile not only shifts across the life stages of Ae. aegypti, but also responds to the presence of invading ZIKV or indicates the condition of the mosquito, which suggests bacterial f-OTUs as potential ecological markers of infection within the mosquito vector under the experimental conditions tested.
Materials and methods
Aedes aegypti and viral infection
Experimental infection was conducted using 3–5 day-old female Ae. aegypti mosquitoes. A well-established Brazilian closed colony of Ae. aegypti (strain PP-Campos), maintained at the Laboratory of Medical Entomology, Fiocruz-MG, Brazil, was used in this study. Mosquito eggs were collected using ovitraps in the city of Campos dos Goitacazes, State of Rio de Janeiro, Brazil, in 2000. Mosquitoes were reared and maintained under standard insectary conditions (27°C, 80% relative humidity, 16-hour light/8-hour dark photoperiod).
The mosquitoes were infected via a membrane-feeding assay using a glass-feeding device that was filled with mouse blood containing a ZIKV isolate (ZIKV-SPH, ZIKV/H. sapiens/Brazil/SPH/2015) that is currently in circulation in Brazil . The mosquitoes were allowed to feed on blood infected with ZIKV at a titer of 1 x 105 PFU for 1 hour. Following blood feeding, approximately 100 fully engorged Ae. aegypti mosquitoes were obtained to form each group and biological replicates: a) sugar-fed (fed ad libitum with sugar), (b) blood-fed (normal non-infected blood), and (c) ZIKV-infected blood-fed. After blood digestion, two groups were re-manned as (d) gravid (7 days post non-infectious feeding (group b) with developing eggs), and (e) ZIKV-infected gravid (7 days post ZIKV infectious feeding (group c) with developing eggs). All mosquitoes were maintained with ad libitum access to 10% glucose solution until the end of experiment. The gravid groups allowed determination of whether the ZIKV infection had an effect on the bacterial profile of Ae. aegypti mosquitoes during egg development. For the ZIKV-infected blood-fed and ZIKV-infected gravid groups, the presence of the virus was verified by TaqMan real-time qPCR (quantitative real time PCR) assay, as described  with 100% of the mosquitoes infected at the 3rd and 7th day after the infectious blood meal with a median of 106 and 109 cDNA copies/ mosquito, respectively).
Pooled samples consisted of 30 Ae. aegypti mosquitoes from each experimental group (three biological independent replicates composed by 10 mosquitoes). Before genomic DNA extraction, each individual mosquito in the pooled sample was surface-sterilized by washing with distilled sterile water, and then submerged for 10 s in 70% ethanol, one min in 1% hypochlorite, and one min in PBS (phosphate-buffered saline) ; the washing steps were performed three times. This procedure is routinely used for sterilization of insect surfaces without considerable impact on the profiles of internal bacterial communities . Genomic DNA was then extracted using the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions.
Amplicon-oriented metagenomic profiling of bacterial communities
Next Generation Sequencing (NGS) of the samples using the Illumina MiSeq platform and targeting the V3–V4 hypervariable region of the 16S rRNA gene was outsourced to Macrogen (https://dna.macrogen.com/eng/), and 300-bp-long paired-end reads were obtained.
All the data were processed in accordance with standard bioinformatics pipelines, which accounted for the following: Quality Control, clustering of operational taxonomic units (OTUs;) clustering (CD-HIT-OTUT at an identity threshold of 97%), and analysis of community richness analysis (Mothur v1.35.1) using taxonomic assignments based on the release 128 of the SILVA rRNA and in-house software or algorithms.
Analyses of the composition of the bacterial communities
The diversity richness and abundance of the bacterial component of the microbiota of each experimental Ae. aegypti group was studied in terms of α-diversity, presence and absence patterns, and β-diversity-based multivariate analyses on a relative abundance matrix. When conventional methods of α-diversity analysis were considered, we opted to report a measure of community evenness (the Shannon diversity index) to address the disturbance that ZIKV replication might have caused to the homeostasis of the bacterial community. This parameter of community structure, in particular, has been linked to the biological phenomenon of successful pathogen invasion . In order to visualize the presence and absence patterns of the f-OTUs per experimental group, circular plots were generated from the tabular indices using Circos online table viewer v0.63–9 . By cross-referencing the indices, sets of exclusive f-OTUs per group, as well as a core set of taxa that were present in all of the experimental Ae. aegypti groups were revealed. When the variance among the groups (β-diversity) in terms of the profiles of OTU abundance was explored, the criteria tested and described by Audsley et al  were applied, using a cut-off of 10 HQ reads at the taxonomic level of family (f-OTUs) to build the data matrix for downstream ordination analysis.
The microbial community composition comparisons were based on pairwise dissimilarity measurements and were performed on a matrix encompassing the relative abundance profiles of the bacterial OTUs from each tested group. For a more conservative approach, analyses were based at the family level to encompass higher OTU counts and perform a more robust explorative analysis and constrained ordination, by reducing the number of zeroes in the data matrix. A variance- stabilizing transformation was applied to the relative abundance matrix to reduce dispersion, expressing the values as the arcsin√x [30, 31].
The multivariate analyses were performed on the transformed dataset of abundance profiles considering suggested parameters of microbial ecology, based on the Bray-Curtis (B-C) index in order to exclude joint absence [22,30,32]. As described in recent literature for metagenomic studies in vector biology [33,34], the conventional approach has been used for weighted and/or unweighted UniFrac distances [35, 36]. This may be a frequently used approach since it is implemented in the widely used pipelines for the analysis of microbial community sequence, QIIME  and Mothur . Although this is a currently popular and highly informative dissimilarity measure to compute pairwise distances between microbial communities, the remarks of the developers regarding when to use the dissimilarity measure for studies on β-diversity and community composition should be considered. Since our study focused on individuals from the same population, which were reared and experimented upon under homogeneous conditions, dissimilarities were estimated using the B-C index. In this context, our hypothesis of the modulation of community structure by the pathogen (ZIKV) did not require consideration of the phylogenetic relationships among the bacterial OTUs, or their evolutionary history when the pairwise comparisons between the elements of the data matrix were performed.
As Kuczynski et al  reported, the use of phylogenetic-based measures requires substantial additional expertise to interpret the effect of the phylogenetic parameters on the ordination methods and their ecological interpretation. Unless the evaluation of the host-microbiota phylosymbiosis [40, 41], or the evaluation of geographic and niche-oriented effects  are within the scope of the study, performing other distance calculations as well is suggested in order to approach a situation wherein bacterial markers are used to compare experimental groups.
The spatial distribution patterns of both objects and response variables in a multidimensional ordination space were analyzed for identifying underlying explanatory variables that influenced their variances. A Detrended Canonical Correspondence Analysis (DCCA) was performed to determine whether a linear or unimodal ordination method could be applied to the data. The longest gradient obtained was larger than 3.0 SD units, which indicated that a unimodal model-based constrained method was suitable [30, 32]. Therefore, a Canonical Correspondence Analysis (CCA, scaling 2) was performed using an additional explanatory matrix with “dummy” binary variables referring to the following conditions: presence of blood in the midgut, ZIKV infection, and a state of gravidity. By constraining the multivariate analysis to these explanatory variables, the potential ecological preference of the f-OTUs towards conditions that could influence their abundance or thriving in a particular niche state was explored [30,42]. Thus, the CCA triplot permitted us to observe how the explanatory variables (vector arrows) influenced the community variance, which was represented in the constraining axis. In addition, by also selecting the scaling 2 parameter scaling it was possible to observe the position of the maximum values of key response variables along an explanatory variable by means of the orthogonal distance between the f-OTU points and the arrow vectors. All of the abovementioned analyses were performed using the Vegan package of Rstudio version 0.98.507 .
This study was conducted in accordance with the recommendation in the Manual for the Use of Animals (FIOCRUZ, Ministry of Health of Brazil, National decree, Nr. 3179). The protocol was approved by the Ethics Committee for the Use of Animals of Fundação Oswaldo Cruz (FIOCRUZ, Ministry of Health of Brazil, Nr. L-1715).
Results and discussion
Taxonomic profiling of the bacterial community
NGS generated a total of 123,973 valid HQ reads that were identified as bacteria in the Ae. aegypti mosquitoes and which were distributed among the following experimental groups: 21,383 in the sugar-fed group, 27,235 in the blood-fed group, 24,072 in the ZIKV-infected blood-fed group, 24,594 in the gravid group and 25,649 in the ZIKV-infected gravid group. These reads corresponded to Ninety seven f-OTUs that fell within the following 14 phyla (f-OTU number within parentheses): Acidobacteria (1), Actinobacteria (10), Bacteroidetes (12), Chlamydiae (3), Cyanobacteria/Chloroplast (1), Deinococcus-Thermus (1), Firmicutes (16), Gemmatimonadetes (1), Ignavibacteriae (1), Nitrospirae (1), Planctomycetes (2), Proteobacteria (41), Spirochaetes (3), and Verrucomicrobia (4). The presence or absence patterns of the f-OTUs within each experimental group are depicted in S1 Fig. As Kuczinski et al  thoroughly discussed, the depth of sequencing does not determines the capacity to reveal underlying biological patterns from datasets of community structures. Here, working downstream with read counts allowed for the construction of a conservative and robust data matrix, as it has been shown that by choosing appropriate ordination methods and dissimilarity measures as few as 100 sequences can reveal relevant ecological patterns within the structure of a bacterial community structure .
When community composition was explored in terms of α-diversity, the reported evenness indices for the experimental Ae. aegypti groups were as follows: 8.00 for the sugar-fed group, 8.43 for the blood-fed group, 5.54 for the ZIKV-infected blood-fed, 8.11 for the gravid group, and 6.37 for the ZIKV-infected gravid group. Therefore, when community richness was addressed in terms of evenness, the Shannon diversity indices revealed that there were more bacterial OTUs, with the total reads more equitably distributed, in the noninfected adult Ae. aegypti mosquitoes. In contrast, the read abundance was less homogenously distributed amongst the bacterial taxa in the community when exposed to ZIKV. This trend was similarly observed on the native microbiota of Lutzomyia longipalpis when exposed to the parasite Leishmania infantum chagasi . Together, these data suggest that successful invasion and concomitant transmission of these two vector-borne pathogens is reflected by the ecological disturbance of the native bacterial community in the host.
Overview of community composition, and identification of exclusive and core microbiota per group
When the index of f-OTUs of each experimental Ae. aegypti group was compared (S1 Table), group specific and core and bacterial taxa were identified, as depicted by the circular plots (Fig 1). The subsets were identified by considering the criterion of being present in all groups or absent in at least one of them. It was inferred that 49 OTUs remained associated with the Ae. aegypti mosquitoes under all the tested experimental conditions. This core assembly encompassed 50.5% of the predicted reads from the dataset, with 40 Gram-negative and 9 Gram-positive bacterial families. These results suggested that although habitat, food source, and life stage may modulate the richness profile, some bacterial OTUs within this broad taxonomic category could thrive under changing conditions of the holobiont [5,8,10,11], and may be essential to the host as a systemic symbiotic core [4,40,45].
The circular plot depicts the abundance of the core bacterial taxa identified by crossing the indices of the f-OTUs per group for the experimental Ae. aegypti groups: sugar-fed (blue); blood-fed (red); ZIKV-infected blood-fed (maroon); Gravid (purple); and ZIKV-infected gravid (lilac). The bacterial taxa and experimental groups are indicated next to their ribbons on the perimeter of the circle. The width of the ribbon and the position of the f-OTUs from the bottom to the top (on each segment) reflect the abundance of the bacterial taxa and their presence within each group.
Our data differed from that of a recent study using uninfected Ae. aegypti from a field and a colony, which revealed a core microbiota assembly of 19 bacterial taxa, including 10 f-OTUs  Particularly, the differences observed in terms of OTU richness could be due to the NGS platform used and to the targeted region(s) of the 16S rRNA gene, as they each possessed different taxonomic resolution . Such core assemblies of host-associated bacterial OTUs have been described in mosquitoes including anophelines, and were determined by different criteria, such as tissue specificity [46,47], life stage [10,11,13], or geographic distribution, as suggested by Villegas and Pimenta .
Fluctuations in abundance within the core group of taxa were observed under different physiological conditions and revealed a dynamic community of bacteria (Figs 1 and 2). In the ZIKV blood-fed group the Flavobacteriacae family was the furthermost prevalent, accounting for 60% of bacterial OTU reads, decreasing in abundance in the other mosquito groups in the following order: Gravid (20%) > ZIKV Gravid (10%) > Sugar-fed (6%) > and blood-fed (4%). The initial induction of Flavobacteria by ZIKV appears to be important in the first days of infection but diminish after 7 days as is observed in the mosquitoes of the group ZIKV Gravid. The thriving abundance of this f-OTU in the infected blood-fed mosquitoes (ZIKV-infected blood-fed group) was in contrast with previous studies that reported its high relative abundance in laboratory-reared sugar-fed Ae. aegypti mosquitoes (represented at the genus level by Elizabethkingia ) and blood-fed An. Gambiae collected from the field and reared in a microcosmos . This contrasting relative abundance could be attributable to physiological differences between strains and genera, or sequencing (effects of the primers and platform) and downstream computational processing (e.g. taxonomic assignment and reference database selection) [46,48–50].
The presence of arbitrarily selected core f-OTUs per group is shown as stacked bars to explore the dynamic shifts exhibited within each experimental group, which is represented by a color as indicated in the figure.
The f-OTUs Bacteroidaceae and Planctomitaceae exhibited a similar abundance pattern when they were present in all the experimental Ae. aegypti groups. Methylococcaceae was the second most abundant family across the experimental groups, and exhibited a relatively homogeneous bacterial community genera prevalence in the uninfected groups, while being the dominant f-OTU in the ZIKV infected groups. Surprisingly, Enterobacteriaceae was not the most responsive in terms of biomass proliferation (measured in terms of read abundance) in the blood-fed group, contrary to what was reported in anophelines [10,12]. It has been suggested that Enterobacteriaceae would potentially dominate and thrive under these conditions because of its redox potential [10,51], however this f-OTU was predominant in the ZIKV-infected gravid group.
To further show these fluctuations among the most abundant taxa, the f-OTUs encompassing less than 1000 valid reads were arbitrarily excluded, which revealed a remaining set of 8 families (Fig 2). The influence of blood feeding on the diversity profile of these 8 families present in the experimental sugar-fed group was analyzed. Conditions of oxidative stress and high protein content due to blood-digestion may favor Methylococcaceae, Rhodocyclaceae, Planctomycetaceae, Chitinophagaceae, Flavobacteriaceae, and Bacteroidaceae in order to outcompete Enterobacteriaceae and Verrucomicrobia [10,51]. However, in the presence of ZIKV, Flavobacteriaceae would be the thriving taxon in a community with lower richness as indicated by the Shannon diversity indices. As digestion occurred, the gravid Ae. aegypti group would then harbor a community wherein Rhodocyclaceae and Enterobacteriaceae would be less abundant, with Flavobacteriaceae, Planctomycetaceae and Methylococcaceae being the dominant community members. In the presence of ZIKV and during vitellogenesis, the latter taxa dominated the community, and Enterobacteriaceae was again detected above the threshold set.
When the inclusion or exclusion patterns of the bacterial OTUs within the studied experimental Ae. aegypti groups were considered, subsets of f-OTUs that were present in no more than 4 of the 5 groups were identified in the circular plot shown in Fig 3. Interestingly, 49% of the families that were identified in the whole dataset were absent (not detected by NGS) in at least one test group,34 were Gram-negative and 14 Gram-positive families. When complete exclusion was considered, a total of 14 f-OTUs were identified solely in one experimental group. The group with a higher number of exclusive taxa was the ZIKV-infected gravid group (7), which was interesting since only one f-OTU (Desulfobulbaceae) was exclusive to the noninfected gravid females.
The circular plot depicts the relative abundance of the exclusive bacterial taxa that were identified per experimental Ae. aegypti group: sugar-fed (blue), blood-fed (red), ZIKV-infected blood-fed (maroon), Gravid (purple), and ZIKV-infected gravid (lilac). The names of the bacterial f-OTUs and experimental groups are indicated next to their ribbons on the perimeter of the circle. The width of the ribbon and the position of the f-OTUs from the bottom to the top (on each segment) reflect the abundance of each bacterial taxon and their presence in each group.
As addressed previously, to further describe fluctuations in community abundance across experimental conditions, the f-OTUs that represented less than 1% of the valid reads (>30 reads) within the experimental Ae. aegypti groups were arbitrarily excluded, and consequently, 12 taxa were retained (Fig 4).
The raw abundance of the arbitrarily filtered f-OTUs that appeared to be present exclusively in the studied groups is shown as stacked bars revealing how community diversity is altered in the experimental groups. Each condition is represented by a color as indicated.
The exclusive taxa within the communities were low in abundance. This tendency of communities towards low abundance has been described in the human microbiome, wherein niche-specific OTUs with low abundance play key ecological roles by providing essential metabolic benefits to the community network and the host under particular conditions . Following the criteria that were applied, non-core f-OTUs were more prevalent in the blood-fed group. The effects of the reduction in the richness of OTUs in the infected groups were evident when both blood-fed groups were compared.
CCA analysis of β-diversity, constrained by explanatory biological variables
The native microbiota of insects has been linked to multiple physiological roles, such as digestion, development, and reproduction [3–5,8,15]. Many of the intricate animal-microbe associations are open systems. As a consequence, the host microbiota are continually exposed to external microorganisms, and influenced by environmental factors [53,54]. When hematophagous arthropod vectors are considered, the presence of allochthonous pathogens that are acquired upon blood ingestion may result in the disruption of homeostasis within the autochthonous microbiota . This may be explored by analyzing the structure of the bacterial community in the whole-insect niche using microbial ecological methods. Herein, this question was addressed using a “holobiont” and an insect “macrobe” biological unit  by thoroughly considering the open-system nature of animal-microbe consortia [54,56]. In such a scenario, shifts in the signature profiles of bacterial f-OTUs could represent physiological conditions or reflect the impact of pathogen invasion on the multiple compartments that harbor bacteria in mosquitoes [8,15,47,49,57,58].
To address how ZIKV invasion may disturb the ecological balance of the microbiota of Ae. aegypti, a CCA analysis coupled with an explanatory matrix was performed to support the biological interpretation of shifts in bacterial signatures. First an estimated 55.40% of the variance within the built matrix was contained and explained along the two CCA axes (CCA1 = 31.15% and CCA2 = 24.25%; Fig 5).
The distribution of the experimental groups along the axes, and their orientation regarding the arrow vectors suggest that exposure to blood (red and maroon) and ZIKV infection (maroon and lilac) had an impact on the signature profiles of the bacterial community. The CCA plot, along its two axes (CCA1 = 31.15% and CCA2 = 24.25%), could explain 55.40% of the matrix variance. Experimental groups: sugar-fed (blue), blood-fed (red), ZIKV-infected blood-fed (maroon), Gravid (violet), and ZIKV-infected gravid (lilac).
The effects that the blood-feeding, infection process, and gravid conditions had on the distribution of the experimental Ae. aegypti groups and the f-OTUs can be visualized (and interpreted) as the length of each arrow vector shown (Fig 5), and the position that the objects and response variables have are related to them. The bacterial signature profile of the sugar-fed group distinguished this group from all the other groups of mosquitoes that fed on blood, localized in the lower left quadrant of the ordinate space. As this method is based on taxa-environment relationships [30,32] it enabled us to determine the OTUs that responded and thrived under particular environmental variables. This property has been highlighted in diverse studies wherein this and other microbial ecological methods guided the identification of potential indicator species [30,31] that aided the understanding of the bacterial signatures of host organisms . As observed in the CCA triplot, two f-OTUs appeared as potential biomarkers of ZIKV infection: Rhodobacteraceae and Desulfuromonadaceae. Their abundance reached the maximum value along the explanatory vector that indicated the presence of the ZIKV virus (according to a right-angled projection of the specified crosses onto the “Infection” arrow). Coincidentally, both f-OTUs were exclusively present in the ZIKV- infected blood-fed and ZIKV- infected gravid groups (Fig 3). The distribution of the ZIKV-infected groups in the ordinate space suggested an altered signature profile compared with that of their ZIKV-free counterparts, as indicated by their separated positions along the CCA1 axes, and in reference to the explanatory vectors “Gravid” and “Infection.” Finally, it was unusual to observe the near-180° angle between both gravid groups. This would imply a negative correlation between the two groups based on their signature profiles. Whether the effects of this ecological disturbance on the microbiota of the gravid Ae. aegypti translated into measurable effects in the microbiota of their offspring microbiota is yet unknown.
The broad metagenomic survey presented here revealed a dynamic bacterial community that was associated with Ae. aegypti with signature profiles for each tested physiological condition, and particularly modulated by the presence of ZIKV. The results highlight the existence of a core group of bacterial f-OTUs that remained associated with the mosquito across the tested conditions. We believe the provided bacterial index will enable the research community to perform comparative studies that explore how bacterial symbionts may act as biomarkers of the physiological states of their hosts, and how they respond as a community when arboviruses invade their host. However, the microbiota described in our experimental groups may reflect the fact that these were insects from a colony as well as the tested conditions of the mosquitoes. Basic knowledge of local hematophagous vectors and their associated microbiota is required, as it is relevant to addressing transmission of vector-borne infectious disease in their regional surroundings.
S1 Fig. f-OTU rank bacterial communities associated with the different experimental Ae. aegypti groups.
The circular plot shows the bacterial f-OTU members of the community that were associated with the tested Ae. aegypti groups: sugar-fed (blue), blood-fed (red), ZIKV-infected blood-fed (maroon), Gravid (purple), and ZIKV-infected gravid (lilac). The names of the bacterial taxa and experimental groups are indicated next to their respective ribbons along the perimeter of the circle. The width of the ribbon and the position of the f-OTUs from the bottom to the top reflect the abundance of the bacterial taxa and their presence within each experimental group.
This study was partially funded by the following Brazilian agencies: Foundation of the Institute Oswaldo Cruz (FIOCRUZ), Brazilian Council for Scientific and Technological Development (CNPq/MS/CAPES/DECIT), Programme Science without Borders, INCT-Instituto Nacional de Ciencia e Tecnologia—Entomologia Molecular, Minas Gerais State Research Support Foundation (FAPEMIG) and Amazonas State Research Support (FAPEAM).
- 1. World Health Organization. Epidemic focus. Waiting in the wings: Aedes mosquitoes and flavivirus epidemics. Wkly Epidemiol Rec. 2016; 91: 249–250. pmid:27187993
- 2. Possas C, Brasil P, Marzochi MC, Tanuri A, Martins RM et al. Zika puzzle in Brazil: peculiar conditions of viral introduction and dissemination—A Review. Mem Inst Oswaldo Cruz. 2017; 112: 319–327. S0074-02762017005006103 [pii]; pmid:28443985
- 3. Dillon RJ, Dillon VM. The gut bacteria of insects: nonpathogenic interactions. Annu Rev Entomol. 2004; 49: 71–92. pmid:14651457
- 4. Douglas AE. Lessons from studying insect symbioses. Cell Host Microbe. 2011; 10: 359–367. S1931-3128(11)00263-0 [pii]; pmid:22018236
- 5. Coon KL, Vogel KJ, Brown MR, Strand MR. Mosquitoes rely on their gut microbiota for development. Mol Ecol. 2014; 23: 2727–2739. pmid:24766707
- 6. Lindh JM, Borg-Karlson AK, Faye I. Transstadial and horizontal transfer of bacteria within a colony of Anopheles gambiae (Diptera: Culicidae) and oviposition response to bacteria-containing water. Acta Trop. 2008; 107: 242–250. pmid:18671931
- 7. Saldana MA, Hegde S, Hughes GL. Microbial control of arthropod-borne disease. Mem Inst Oswaldo Cruz. 2017; 112: 81–93. S0074-02762017000200081 [pii]; pmid:28177042
- 8. Minard G, Mavingui P, Moro CV. Diversity and function of bacterial microbiota in the mosquito holobiont. Parasit Vectors. 2013; 6: 146. pmid:23688194
- 9. Villegas LM, Pimenta PF. Metagenomics, paratransgenesis and the Anopheles microbiome: a portrait of the geographical distribution of the anopheline microbiota based on a meta-analysis of reported taxa. Mem Inst Oswaldo Cruz. 2014; 109: 672–684. S0074-02762014000500672 [pii]. pmid:25185007
- 10. Wang Y, Gilbreath TM III, Kukutla P, Yan G, Xu J. Dynamic gut microbiome across life history of the malaria mosquito Anopheles gambiae in Kenya. PLoS One. 2011; 6: e24767. pmid:21957459
- 11. David MR, Santos LM, Vicente AC, Maciel-de-Freitas R. Effects of environment, dietary regime and ageing on the dengue vector microbiota: evidence of a core microbiota throughout Aedes aegypti lifespan. Mem Inst Oswaldo Cruz. 2016; 111: 577–587. S0074-02762016005017102 [pii]; pmid:27580348
- 12. Boissiere A, Tchioffo MT, Bachar D, Abate L, Marie A et al. Midgut microbiota of the malaria mosquito vector Anopheles gambiae and interactions with Plasmodium falciparum infection. PLoS Pathog. 2012; 8: e1002742. pmid:22693451
- 13. Gimonneau G, Tchioffo MT, Abate L, Boissiere A, Awono-Ambene PH et al. Composition of Anopheles coluzzii and Anopheles gambiae microbiota from larval to adult stages. Infect Genet Evol. 2014; 28: 715–724. S1567-1348(14)00363-3 [pii]; pmid:25283802
- 14. Cirimotich CM, Ramirez JL, Dimopoulos G. Native microbiota shape insect vector competence for human pathogens. Cell Host Microbe. 2011; 10: 307–310. S1931-3128(11)00291-5 [pii]; pmid:22018231
- 15. Gendrin M, Christophides GK (2013) The Anopheles Mosquito Microbiota and Their Impact on Pathogen Transmission, Anopheles mosquitoes—New insights into malaria vectors. In: Manguin S, editors. Anopheles mosquitoes new insights into malaria vectors.
- 16. Beier MS, Pumpuni CB, Beier JC, Davis JR. Effects of para-aminobenzoic acid, insulin, and gentamicin on Plasmodium falciparum development in anopheline mosquitoes (Diptera: Culicidae). J Med Entomol. 1994; 31: 561–565. pmid:7932602
- 17. Manda H, Gouagna LC, Nyandat E, Kabiru EW, Jackson RR et al. Discriminative feeding behaviour of Anopheles gambiae s.s. on endemic plants in western Kenya. Med Vet Entomol. 2007; 21: 103–111. MVE672 [pii]; pmid:17373953
- 18. Nyasembe VO, Teal PE, Mukabana WR, Tumlinson JH, Torto B. Behavioural response of the malaria vector Anopheles gambiae to host plant volatiles and synthetic blends. Parasit Vectors. 2012; 5: 234. 1756-3305-5-234 [pii]; pmid:23069316
- 19. Engel P, Moran NA. The gut microbiota of insects—diversity in structure and function. FEMS Microbiol Rev. 2013; 37: 699–735. pmid:23692388
- 20. Christian N, Whitaker BK, Clay K. Microbiomes: unifying animal and plant systems through the lens of community ecology theory. Front. Microbiol 2015; 6(869) 1–15: pmid:26441846
- 21. Legendre P, De CM. Beta diversity as the variance of community data: dissimilarity coefficients and partitioning. Ecol Lett. 2013; 16: 951–963. pmid:23809147
- 22. Anderson MJ, Crist TO, Chase JM, Vellend M, Inouye BD et al. Navigating the multiple meanings of beta diversity: a roadmap for the practicing ecologist. Ecol Lett. 2011; 14: 19–28. pmid:21070562
- 23. Secundino NFC, Chaves BA, Orfano AS, Silveira KRD, Rodrigues NB et al. Zika virus transmission to mouse ear by mosquito bite: a laboratory model that replicates the natural transmission process. Parasit Vectors. 2017; 10: 346. pmid:28728607
- 24. Lanciotti RS, Kosoy OL, Laven JJ, Velez JO, Lambert AJ et al. Genetic and serologic properties of Zika virus associated with an epidemic, Yap State, Micronesia, 2007. Emerg Infect Dis. 2008; 14: 1232–1239. pmid:18680646
- 25. Lacey LA, Brooks WM (1997) Initial handling and diagnosis of diseased insects. In: Lacey LA, editors. Manual of Techniques in Insect Pathology. San Diego: Elsevier Science. pp. -409.
- 26. Duguma D, Hall MW, Smartt CT, Neufeld JD. Effects of Organic Amendments on Microbiota Associated with the Culex nigripalpus Mosquito Vector of the Saint Louis Encephalitis and West Nile Viruses. mSphere. 2017; 2. pmid:28168223
- 27. De RK, Marzorati M, Negroni A, Thas O, Balloi A et al. Environmental conditions and community evenness determine the outcome of biological invasion. Nat Commun. 2013; 4: 1383. ncomms2392 [pii]; pmid:23340423
- 28. Krzywinski M, Schein J, Birol I, Connors J, Gascoyne R et al. Circos: an information aesthetic for comparative genomics. Genome Res. 2009; 19: 1639–1645. gr.092759.109 [pii]; pmid:19541911
- 29. Audsley MD, Ye YH, McGraw EA. The microbiome composition of Aedes aegypti is not critical for Wolbachia-mediated inhibition of dengue virus. PLoS Negl Trop Dis. 2017; 11: e0005426. pmid:28267749
- 30. Ramette A. Multivariate analyses in microbial ecology. FEMS Microbiol Ecol. 2007; 62: 142–160. FEM375 [pii]; pmid:17892477
- 31. Mason MR, Nagaraja HN, Camerlengo T, Joshi V, Kumar PS. Deep sequencing identifies ethnicity-specific bacterial signatures in the oral microbiome. PLoS One. 2013; 8: e77287. pmid:24194878
- 32. Paliy O, Shankar V. Application of multivariate statistical techniques in microbial ecology. Mol Ecol. 2016; 25: 1032–1057. pmid:26786791
- 33. Kelly PH, Bahr SM, Serafim TD, Ajami NJ, Petrosino JF et al. The Gut Microbiome of the Vector Lutzomyia longipalpis Is Essential for Survival of Leishmania infantum. MBio. 2017; 8. mBio.01121-16 [pii]; pmid:28096483
- 34. Osei-Poku J, Mbogo CM, Palmer WJ, Jiggins FM. Deep sequencing reveals extensive variation in the gut microbiota of wild mosquitoes from Kenya. Mol Ecol. 2012; 21: 5138–5150. pmid:22988916
- 35. Lozupone C, Lladser ME, Knights D, Stombaugh J, Knight R. UniFrac: an effective distance metric for microbial community comparison. ISME J. 2011; 5: 169–172. ismej2010133 [pii]; pmid:20827291
- 36. Lozupone CA, Hamady M, Kelley ST, Knight R. Quantitative and qualitative beta diversity measures lead to different insights into factors that structure microbial communities. Appl Environ Microbiol. 2007; 73: 1576–1585. AEM.01996-06 [pii]; pmid:17220268
- 37. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010; 7: 335–336. nmeth.f.303 [pii]; pmid:20383131
- 38. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M et al. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol. 2009; 75: 7537–7541. AEM.01541-09 [pii]; pmid:19801464
- 39. Kuczynski J, Liu Z, Lozupone C, McDonald D, Fierer N et al. Microbial community resemblance methods differ in their ability to detect biologically relevant patterns. Nat Methods. 2010; 7: 813–819. nmeth.1499 [pii]; pmid:20818378
- 40. Brooks AW, Kohl KD, Brucker RM, van Opstal EJ, Bordenstein SR. Phylosymbiosis: Relationships and Functional Effects of Microbial Communities across Host Evolutionary History. PLoS Biol. 2016; 14: e2000225. pmid:27861590
- 41. Novakova E, Woodhams DC, Rodriguez-Ruano SM, Brucker RM, Leff JW et al. Mosquito Microbiome Dynamics, a Background for Prevalence and Seasonality of West Nile Virus. Front Microbiol. 2017; 8: 526. pmid:28421042
- 42. Legendre Pierre and Legendre Loic FJ (2012) Numerical ecology. Elsevier.
- 43. Team RS. RStudio: integrated development for R. RStudio, Inc, Boston, MA URL http://www.rstudio.com. 2015.
- 44. Kuczynski J, Costello EK, Nemergut DR, Zaneveld J, Lauber CL et al. Direct sequencing of the human microbiome readily reveals community differences. Genome Biol. 2010; 11: 210. gb-2010-11-5-210 [pii]; pmid:20441597
- 45. Bordenstein SR, Theis KR. Host Biology in Light of the Microbiome: Ten Principles of Holobionts and Hologenomes. PLoS Biol. 2015; 13: e1002226. pmid:26284777
- 46. Yang B, Wang Y, Qian PY. Sensitivity and correlation of hypervariable regions in 16S rRNA genes in phylogenetic analysis. BMC Bioinformatics. 2016; 17: 135. pmid:27000765
- 47. Tchioffo MT, Boissiere A, Abate L, Nsango SE, Bayibeki AN et al. Dynamics of Bacterial Community Composition in the Malaria Mosquito's Epithelia. Front Microbiol. 2015; 6: 1500. pmid:26779155
- 48. Tremblay J, Singh K, Fern A, Kirton ES, He S et al. Primer and platform effects on 16S rRNA tag sequencing. Front Microbiol. 2015; 6: 771. pmid:26300854
- 49. Segata N, Baldini F, Pompon J, Garrett WS, Truong DT et al. The reproductive tracts of two malaria vectors are populated by a core microbiome and by gender- and swarm-enriched microbial biomarkers. Sci Rep. 2016; 6: 24207. srep24207 [pii]; pmid:27086581
- 50. Oulas A, Pavloudi C, Polymenakou P, Pavlopoulos GA, Papanikolaou N et al. Metagenomics: tools and insights for analyzing next-generation sequencing data derived from biodiversity studies. Bioinform Biol Insights. 2015; 9: 75–88. pmid:25983555
- 51. Champion CJ, Xu J. The impact of metagenomic interplay on the mosquito redox homeostasis. Free Radic Biol Med. 2017; 105: 79–85. S0891-5849(16)31054-1 [pii]; pmid:27880869
- 52. Human Microbiome Project Consortium. Structure, function and diversity of the healthy human microbiome. Nature. 2012; 486: 207–214. nature11234 [pii]; pmid:22699609
- 53. Costello EK, Stagaman K, Dethlefsen L, Bohannan BJ, Relman DA. The application of ecological theory toward an understanding of the human microbiome. Science. 2012; 336: 1255–1262. science.1224203 [pii]; pmid:22674335
- 54. Wong AC, Luo Y, Jing X, Franzenburg S, Bost A et al. The Host as the Driver of the Microbiota in the Gut and External Environment of Drosophila melanogaster. Appl Environ Microbiol. 2015; 81: 6232–6240. AEM.01442-15 [pii]; pmid:26150460
- 55. Abraham NM, Liu L, Jutras BL, Yadav AK, Narasimhan S et al. Pathogen-mediated manipulation of arthropod microbiota to promote infection. Proc Natl Acad Sci U S A. 2017; 114: E781–E790. 1613422114 [pii]; pmid:28096373
- 56. Rowan R. Coral bleaching: thermal adaptation in reef coral symbionts. Nature. 2004; 430: 742. pmid:15306800
- 57. Favia G, Ricci I, Damiani C, Raddadi N, Crotti E et al. Bacteria of the genus Asaia stably associate with Anopheles stephensi, an Asian malarial mosquito vector. Proc Natl Acad Sci U S A. 2007; 104: 9047–9051. pmid:17502606
- 58. Manguin S, Ngo CT, Tainchum K, Juntarajumnong W, Chareonviriyaphap T, et al. (2013) Bacterial Biodiversity in Midguts of Anopheles Mosquitoes, Malaria Vectors in Southeast Asia. In: Manguin Sylvie, editors. "Anopheles mosquitoes—New insights into malaria vectors". InTech, Chapters published. pp. 549–576.
- 59. Knights D, Parfrey LW, Zaneveld J, Lozupone C, Knight R. Human-associated microbial signatures: examining their predictive value. Cell Host Microbe. 2011; 10: 292–296. S1931-3128(11)00287-3 [pii]; pmid:22018228