Background
The impact of malaria transmission upon human health cannot be overstated, with up to two billion people worldwide at risk and an estimated 438,000 deaths yearly [
1].
Plasmodium parasites, the causative agents of malaria, are transmitted through blood-feeding female anopheline mosquitoes.
Plasmodium’s complex life cycle provides multiple opportunities for intervention approaches, by targeting either the parasite or the mosquito vector. Strategies that target the parasite, including anti-malarial drugs, and those that target the vector, such as insecticides, are encountering resistance of the parasite and vector, respectively [
2,
3]. When mosquitoes bite a
Plasmodium-infected person, the ingested
Plasmodium gametocytes encounter a severe bottleneck in development at the ookinete stage within the midgut, the site where most parasites are killed [
4,
5]. During this infection stage, parasites also encounter the mosquito midgut microbiota. The negative impact of the mosquito microbiota on
Plasmodium parasites has been well documented (reviewed in [
6]) and offers an attractive opportunity to further narrow the bottleneck around
Plasmodium development and its transmission between humans. The potential impact of the microbiota upon
Plasmodium development suggests that studying these microorganisms could improve efforts at curtailing malaria transmission.
Interest in the tripartite interaction between vector, parasite, and microbiota has prompted the surveying of natural mosquito populations, revealing a microflora predominantly composed of Gram-negative species, with members of the
Proteobacteria class often in high abundance [
7,
8]. A number of naturally occurring bacteria have been shown to inhibit
Plasmodium development in the mosquito, either through direct inhibition or indirectly through priming of basal immunity [
9‐
11]. A correlation between
P. falciparum infection and the prevalence of
Enterobacteriacae has been documented [
7], and several Gram-negative bacteria species have been shown to reduce the mosquitoes’ vectorial capacity [
12‐
14]; also, a
Chromobacterium species has been shown to possess anti-
Plasmodium effects both in vivo and in vitro [
15].
Screening isolates of the midgut microbiota of a Zambian mosquito population for
Plasmodium-blocking activity has allowed the identification of an
Enterobacter bacterium,
Esp_Z, which is naturally harboured in the mosquito midgut tissue and can inhibit the development of
Plasmodium parasites prior to their invasion of the mosquito midgut epithelium, independent of the mosquito’s immune system [
10].
Esp_Z nearly eliminates the development of malaria parasites to the human-infective sporozoite stage, thereby showing potential as a transmission-blocking agent.
Esp_Z, even at low concentrations, is effective at inhibiting parasites in the mosquito intestine through a mechanism involving reactive oxygen species (ROS) [
10].
Plasmodium parasites are susceptible to changes in the redox balance of their environment, and they defend themselves against ROS via an upregulation of antioxidant genes [
16]. Two antioxidant systems with partially overlapping functions have been described in
Plasmodium parasites, the thioredoxin and glutathione systems (reviewed in [
17]). These systems are required to neutralize the highly oxidative environments encountered in both human erythrocytes [
17] and the mosquito midgut [
18]. The thioredoxin system is believed to be the main antioxidant system, since
P.
falciparum parasites lack a glutathione-dependent peroxidase [
19,
20], suggesting that thioredoxin reduction of peroxiredoxin antioxidant enzymes plays a prominent role in protecting
Plasmodium from toxic ROS [
16]. Peroxiredoxin antioxidant enzymes such as peroxiredoxin-1 (Tpx-1) and 1-cys-peroxiredoxin (1-Cys-prx) act as electron acceptors from thioredoxin-1 (Trx-1), maintaining them in a reduced state [
21].
Plasmodium falciparum Tpx-1 knockout lines show an increased sensitivity to both ROS and reactive nitrogen species, but the deletion mutation is not lethal, suggesting redundancy in the parasite’s antioxidant system [
22]. A glutathione peroxidase-like thioredoxin peroxidase (TPx
Gl) has also been shown to be part of the thioredoxin system in
Plasmodium and is reduced by Trx-1 [
23]. Inhibitors of the thioredoxin pathway have been investigated as potential anti-malarial compounds, rendering the parasite more susceptible to oxidative stress [
24].
Recently, reintroduction of several bacteria through the nectar meal, a natural route and vital mosquito energy source, has been demonstrated to inhibit
Plasmodium sporogonic development, and methods for exposing mosquitoes to various agents through artificial nectars have been developed [
9,
15,
25]. Manipulation of the composition of the mosquito microbiota may offer a strategy to reduce the prevalence of
Plasmodium-infected mosquitoes and, therefore, malaria transmission. A multifaceted approach was used to test the feasibility of using sugar-baited exposure of mosquitoes to the
Esp_Z bacterium for interruption of malaria transmission, by studying the reciprocal interactions between the bacterium and the mosquito and its microbiota, together with
Esp_Z’s impact on the parasite under laboratory conditions. The
Esp_Z genes that are likely to be conducive to effective colonization of the mosquito midgut and ROS–mediated parasite inhibition were specifically assessed, along with the effect of
Esp_Z exposure on the parasite’s oxidative defence system, mosquito fitness, and the ability of
Esp_Z to compete with the endogenous bacteria of the mosquito midgut microbiota.
Methods
Mosquito rearing and antibiotic treatment
Anopheles gambiae Keele strain mosquitoes were maintained under laboratory conditions at 27 °C and 80 % humidity with a 14 h day/10 h night cycle. Larvae were reared on cat food pellets and ground fish food supplement. Adult mosquitoes were maintained on 10 % sucrose and fed on mouse blood (mice were anesthetized with ketamine) for egg production. For clearance of the endogenous bacteria of laboratory mosquitoes, adult female mosquitoes were maintained on antibiotics immediately after eclosion (75 μg/mL gentamicin sulfate [Quality Biological, Gaithersburg, MD, USA] and 100 μg/mL penicillin–streptomycin [Invitrogen, Carlsbad, CA, USA] in a 10 % sucrose solution ad libitum for three to 4 days). The sucrose-antibiotic solution was changed every 48 h. To minimize the impact of any residual antibiotics, the mosquitoes were allowed to feed ad libitum for 24 h on a sterile 10 % sucrose solution following antibiotic treatment.
Bacteria cocktail preparation
The bacterial cocktail was prepared from bacterial isolates previously identified from wild
An. arabiensis mosquitoes obtained through landing catches in Zambia [
10]. Twelve species representing both Gram-negative and Gram-positive bacteria were selected. Single bacterial colonies from LB plates were selected and used to inoculate overnight cultures, then used to seed fresh cultures and grown to an OD
600 of 1.0. Species names and GenBank sequence accession numbers were as follows:
Knoellia sp. JF690939.1;
Acinetobacter sp. JF690925.1;
Bacillus sp. JF690926.1;
Pseudomonas sp. JF690929.1;
Exiguobacterium sp. JF690932.1;
Kocuria sp. JF690933.1;
Pantoea sp. JF690934.1;
Pseudomonas sp. JF690935.1;
Staphylococcus sp. JF690936.1;
Arthrobacter sp. JF690937.1;
Comamonas sp. JF690938.1;
Bacillus sp. JF690930.1.
Introduction of bacteria through sugar or blood meals
Bacteria were grown in either LB or ookinete medium overnight at 30 °C, used to seed fresh cultures, and then diluted to an OD600 of 1.0, pelleted by centrifugation (10 min, 3000 rpm in a tabletop centrifuge), washed in 1× PBS, and finally resuspended in PBS. For sugar-meal introduction, bacteria were diluted in 3 % sucrose to the concentration indicated in the text and provided to mosquitoes on moistened cotton strips. For blood-meal introduction, mosquitoes were allowed to membrane-feed on blood containing bacteria (10 % bacterial solution, 40 % blood, 50 % human serum) at the concentration indicated in the text. Non-fed mosquitoes were removed within 24 h after blood-feeding.
Anopheles gambiae midgut selection of Esp_Z
The parental Esp_Z strain was provided to An. gambiae mosquitoes via blood meal at 109 CFU/ml (P1). Engorged mosquitoes were maintained under standard insectary conditions, and 10 mosquitoes from this cohort were sampled daily for the presence of Esp_Z in the midgut. The Esp_Z bacteria that were found in the midgut for the longest time were then used to challenge a second mosquito population, and the process was repeated to isolate the “most fit” Esp_Z bacteria for midgut colonization (P2). This process of serial passage was then repeated a third time to isolate the midgut-selected Esp_Z bacteria (P3).
Colonization experiments and DNA extraction
To establish the ability of
Esp_Z to colonize midguts harbouring a resident microbiota, 70–80 female mosquitoes were treated with antibiotics as described above. Cocktail bacteria (10
6 CFU/ml) were introduced ad libitum for 48 h via sugar meal; the mosquitoes were then starved for 6–8 h, and
Esp_Z (10
6 CFU/ml) was introduced by either blood-feeding or sugar meal. Every 24 h, midguts were dissected from ten individual mosquitoes and transferred to a centrifuge tube containing 100 µl PBS, then stored at −80 °C for DNA extraction; three independent replicates were performed. Mosquitoes were surface-sterilized by washing them in 100 % ethanol for 2 min and then rinsing them for 1 min in sterile 1× PBS. DNA extractions were carried out as previously described [
26] with minor modifications: In brief, dissected midguts were homogenized in 90 µl PBS, followed by the addition of 90 µl lysozyme (40 mg/ml) and incubation at 37 °C for 1 h; 300 µl of extraction buffer (1 % SDS; 50 mM Tris–HCl, pH 8.0; 25 mM NaCl; 25 mM EDTA, pH 8.0) was added to samples and incubated at 65 °C for 10 min. Following the addition of 200 µl of 3 M potassium acetate (pH 7.2), samples were incubated on ice for 1 h and then centrifuged at 14,000 rpm in a tabletop centrifuge for 10 min and the supernatants removed. The pellets were then treated with RNase (40 mg/ml) at room temperature for 30 min. The supernatants were extracted twice with phenol/chloroform/isoamyl alcohol and precipitated with 2.5 volumes of 100 % ethanol and 1:10 3 M sodium acetate at −20 °C overnight; the pellets were resuspended in 10 µl of water.
Absolute qPCR quantification
All mosquito tissues were stored at −80 °C following dissection, and DNA was extracted as described. Absolute qRT-PCR determination was used to determine total bacteria and
Esp_Z gene copy numbers. Degenerate primers targeting the 16S rRNA sequence of each bacterium were designed from previously sequenced 16S rRNA gene fragments, following Clustal W nucleotide alignment of the sequences [
10]. Primers targeting
Esp_Z gene sequences were designed from
Esp_Z genomic scaffolds. Primers were checked for potential cross-hybridization through BLASTn searches against the nr gene database, retrieving no significant hits (E value = 0.1).
Esp_Z primers were tested for specificity for the bacterium through PCR amplification from DNA of
An. gambiae midguts fed the bacterial cocktail species. PCR fragments used for standard curves were cloned using the pGEM-T Easy vector (Promega). Standard curves for 16S rRNA and
Esp_Z had efficiencies between 80 and 100 % and R
2 > 0.99. The relative proportion of
Esp_Z was calculated by dividing the total number of
Esp_Z-specific copies by the number of amplified 16S rRNA copies across three biological replicates. Amplification and detection of bacterial DNA was performed using the QuantiTect SYBR Green PCR Kit (Qiagen, Valencia, CA, USA) and ABI Detection System ABI Prism 7000, using two technical replicates. The PCR reaction was performed in a total volume of 20 µl, which consisted of 2× Sybr green master mix (Applied Biosystems, Foster City, California, USA), 75 nM of each primer, and 50 ng of template DNA. The cycling parameters were as follows: initial denaturation at 95 °C for 2 min, followed by 40 cycles of 95 °C for 15 s, 56 °C for 30 s, and 60 °C for 30 s. Following amplification, a melting curve analysis was performed from 60 to 95 °C, collecting fluorescence data every 0.5 °C.
Relative qPCR quantification
To conduct relative real-time qPCR assays, RNA was extracted using TRIzol
® (Life Technologies, Carlsbad, CA) according to the manufacturer’s guidelines and treated with Turbo DNase; first-strand cDNA was produced using Superscript III reverse transcriptase (Invitrogen). cDNA templates were normalized to the
Plasmodium berghei 18S rRNA gene as previously described [
16], and fold changes in gene expression levels were determined using the standard [
27]. Assays were performed using Sybr green master mix with the following cycling parameters: initial denaturation at 95 °C for 2 min, followed by 40 cycles of 95 °C for 15 s, 56 °C for 30 s, and 72 °C for 30 s.
Longevity, fecundity, and fertility assays
To assess the impact of bacterial introduction on mosquito longevity, bacteria were introduced into aseptic female mosquitoes, and the impact of Esp_Z was compared to that of the introduction of the bacterial cocktail or a PBS-only control. For each treatment, approximately 60 female mosquitoes were rendered free of their microbiota through antibiotic treatment, and then either Esp_Z, the bacterial cocktail, or PBS was introduced by either a blood or sugar meal. Each cohort was then provided with a naïve blood meal at 4 days post-introduction. Non-fed mosquitoes were removed and excluded from the analysis. Mosquito mortality was monitored daily; monitoring continued until all the mosquitoes had perished, and survival percentages were calculated across three biological replicates. Kaplan–Meier survival analysis was carried out using GraphPad Prism 5 software, and p values were determined by log-rank test (Mantel-Cox) and corrected for multiple comparisons by the Bonferroni method.
For the fecundity and fertility experiments, 30–60 female mosquitoes were treated with antibiotics as described, and three treatment groups were established with Esp_Z, the bacterial cocktail, and PBS being introduced by both sugar and blood meals. Following blood-meal introduction, blood-fed mosquitoes were removed to individual vials (12 ml) lined with moistened filter paper and then incubated under normal rearing conditions. Following sugar-meal introduction, mosquitoes were provided a naïve blood meal 48 h later, and the fed mosquitoes were removed to individual vials. Eggs oviposited on the filter paper were counted daily using a light microscope for 2 days. Females that did not produce eggs by day 2 were re-examined on day 3. After counting, eggs were submerged in water under standard larval rearing conditions, and the hatch rate (fertility) was determined by counting first instar larvae, which were then removed daily. Three biological replicates were performed for the fecundity and fertility assays and significance was determined using the Kruskal–Wallis test.
Plasmodium infection assays
Mosquitoes were fed on NF54 P. falciparum gametocyte cultures (0.01 % gametocytaemia; provided by the Johns Hopkins Malaria Institute Parasitology Core Facility) through artificial membranes at 37 °C. The adult mosquitoes were starved for 8–12 h prior to feeding to ensure engorgement, and unfed mosquitoes were removed from the cohort within 24 h. To determine oocyst numbers, the mosquitoes were incubated for a further 7 days at 27 °C, and midguts were dissected out in PBS, stained with 0.2 % mercurochrome, and examined using a light-contrast microscope (Olympus). To establish whether Esp_Z is capable of inhibiting Plasmodium development when provided via sugar meal, 50–70 female septic mosquitoes were reared as described above and then provided a sugar meal containing Esp_Z at varying concentrations for 3–4 days; they were then provided a P. falciparum-infected blood meal the following day. Oocyst numbers were determined as described above and compared to a cohort fed only PBS containing 3 % sucrose.
Esp_Z genome sequencing and transcriptome analysis
Total bacterial DNA was extracted from an
Esp_Z liquid culture. A 3 kb insert paired-end shotgun genomic library was prepared from the extracted DNA and sequenced using the 454 GS FLX Titanium sequencing platform at the Genome Resource Center at the Institute for Genome Sciences (IGS) of the University of Maryland according to the manufacturer’s protocol. Sequencing reads were then assembled using Celera Assembler, and ORFs were predicted and annotated using the IGS Annotation Engine implemented within the CLoVR-Microbe pipeline [
28]. Total bacterial RNA from bacterial liquid cultures grown in either LB or ookinete medium was extracted using TRIzol according to the manufacturer’s protocol, treated with Turbo DNase (5U, Life Technologies), cleaned using an RNeasy Mini Kit (Qiagen, Valencia, CA, USA), and the quality determined by an Agilent Bioanalyzer 2100. A custom 8 × 44 K Agilent microarray was designed based on the 454-generated
Esp_Z genome sequence. Cy-3- or Cy-5-labeled cRNA probes was synthesized from 300 ng of total RNA per replicate using the LabelIT
® MicroArray Dual Labeling Kit (Mirus Bio, Madison, WI, USA). Labelled RNA was hybridized overnight at 65 °C, and arrays were scanned with an Agilent scanner. Transcript abundance data were processed and analysed as previously described [
29,
30]. In brief, LOWESS normalized background-subtracted median fluorescent values were used to determine Cy5/Cy3 ratios from replicate assays and were then subjected to
t tests at a significance level of
p < 0.05 using MIDAS, GEPAS, and TMEV software [
31,
32]. The rate for self–self hybridizations was used to calculate a cut-off value for the significance of gene regulation on these microarrays of Log2FC = 0.78 or 1.72-fold regulation [
33].
In vitro parasite culture and co-culture with bacteria
For the rodent malaria parasite production, cultures were carried out as previously described [
10] with certain modifications. In brief,
P. berghei GFP parasites were injected into donor Swiss Webster mice and monitored daily for parasitaemia for 3 days post-infection. Once parasitaemia reached >15 %, infected blood was collected by heart puncture and transferred to phenylhydrazine-treated mice. Mice were treated with pyrimethamine 72 h later and monitored daily for parasitemia and exflagellation. Parasitized mice with ≥20 exflagellation events per 20× microscope field were used for in vitro experiments. Parasitized blood was collected by heart puncture and diluted 1:10 in ookinete medium (10 % foetal bovine serum, 50 μg/ml hypoxanthine, 2 mg/ml NaHCO
3, 1 μM xanthurenic acid in RPMI 1640 medium, pH 8.3), 4 % RBC lysate, and experiment-specific constituents in a 500 μl total volume. To assess the induction of
P. berghei antioxidant genes, bacteria (final concentration 10
6 CFU/ml) were added to the ookinete culture at 0 h. Culture samples were directly transferred to Tri-Reagent (Ambion) at various times, from 1 to 10 h after setup, for total RNA extraction. First-strand cDNA was then prepared for qPCR analysis. Development of ookinetes was confirmed by Giemsa staining at 24 h.
Conclusions
The devastating effects of malaria transmission are ever-present in sub-Saharan Africa, and increasing efforts to counteract these effects are either encountering resistance, as in the case of anti-malarial drugs and insecticides, or poor efficiencies, in the case of vaccine development. It is, therefore, imperative that alternate strategies be explored, and the microbiota of disease vectors has become an attractive resource for this purpose. Not only do certain members of the vector microbiota nearly eliminate
Plasmodium infection [
9,
10], but some members have also been shown to effectively kill the mosquito vectors of several pathogens [
15]. Here, the potential of the naturally occurring
Enterobacter species
Esp_Z as part of a malaria control strategy was evaluated.
Proteobacteria are highly abundant in field-caught mosquitoes, including
Enterobacter species, showing that they can be acquired and persist in the mosquito and its habitat [
7,
8]. In order to influence the natural microbiota, bacteria could be introduced via sugar-baited traps [
41] thus ensuring
Plasmodium exposure to them. It was demonstrated that
Esp_Z can persist for up to four days after introduction and inhibit
P. falciparum development when introduced via sugar meal (Fig.
1). As part of a translation study, it is important to examine the introduction and retention of bacteria in the adult midgut, because of the as-yet unclear data on transmission of bacteria from larvae to adults [
67‐
69].
The impact on mosquito fitness of an introduced bacterium is an additional consideration, since a potentially negative impact is likely to be selected against in nature. Unlike the intracellular
Wolbachia species, bacteria that are able to manipulate the host’s reproductive system and thereby ensure vertical dissemination through the population [
70,
71], the spread of a bacterium must not induce any fitness cost. Microbiota are important for mosquito fitness, with the proteobacteria
Asaia sp. influencing larval development and several other species having been shown to affect mosquito fitness [
9,
72]. Naturally occurring bacterial isolates have a varying ability to persist and replicate when re-introduced into aseptic
An. gambiae midguts [
9], suggesting that there is variation in the adaptation to the mosquito midgut environment. This study has shed light on candidate molecular components that are important for bacterial colonization and persistence in the mosquito. It is possible that a combinatorial approach could be undertaken, utilizing a cocktail of anti-parasitic bacteria, each having a different anti-parasitic mechanism, for the control of malaria transmission.
In the present study, the nature of the anti-Plasmodium activity of Esp_Z was also further characterized. On the one hand, the observation that Esp_Z upregulates genes associated with the generation of ROS when cultured under conditions shown to inhibit parasite development in vitro was validated. This finding corroborates previous studies in which Esp_Z was shown to induce an oxidative environment when co-cultured with Plasmodium ookinetes, with such an environment being the likely cause of parasite inhibition. On the other hand, a potential complementary mechanism by which the bacteria appears to directly interfere with the parasite’s antioxidant system was uncovered, thereby limiting its ability to respond and adapt to the oxidative environment it promotes. This mechanism is indicated by the downregulation of key genes of the thioredoxin and glutathione antioxidant pathways in P. berghei ookinetes, until the moment when parasite development is arrested. Current efforts are underway to identify putative mediators of this effect in order to further characterize this phenomenon, in what also constitutes an exploration of Esp_Z’s metabolome from a natural product discovery perspective, in the search for new and innovative anti-malarials.
Authors’ contributions
NJD, RGS, CMC, GD conceived experiments. NJD, RGS, CMC, EFM, GM performed experiments. NJD, RGS, CMC, EFM, GD analysed data obtained. NJD, RGS, CMC, EFM, GD wrote the manuscript. All authors read and approved the final manuscript.