Bacteria strain
The Chromobacterium violaceum strain used for bioassays is isolated from both field collected larvae and cuticles of adult mosquitoes (Anopheles gambiae s.l.) from Bana (11° 9' 41"N, 4° 10' 30" W), Soumousso (11°04’ N, 4°03’ W) in western Burkina Faso. Homogenates from dead mosquitoes were firstly plated out onto Chocolate + polyvitek agar and Bromocresol purple agar. Then, 24 hours and 48 hours, bacterial species were isolates and species of C. violaceum were identified using the VITEK2 system (Supplementary file 1) in laboratory at Centre Muraz.
Mosquito colonies and PCR determination of Kdr levels
For bioassays, we used F1 progeny of An. coluzzii reared from larval collections at Kou Valley (11°23’ N, 4°24’ W). Mosquitoes from these areas are highly resistant to multiple insecticides [19]. Only non-blood-fed females, 2 to 5 days old, were used for bioassays. All bioassays were carried out at 25 ± 2°C and 80±10% relative humidity.
The kdr gene prevalence within a subsample of mosquitoes (180 mosquitoes) was performed using the PCR protocol and primer sequences previously described [20]. We only analyzed mutation L1014F because it is the commonest in West Africa, whereas the L1014S mutation is confined to East Africa [19]. The primers AgD1 (5′-ATA GAT TCC CCG ACC ATG-3′) and AgD3 (5′-AAT TTG CAT TAC TTA CGA CA-3′) amplified the resistant allele yielding 195 bp fragments. The susceptible allele was assayed using primers AgD2 (5′-AGA CAA GGA TGA TGA ACC-3′) and AgD4 (5′-CTG TAG TGA TAG GAA ATT TA-3′), which amplified a 137 bp fragment. The primer set AgD1 and AgD2 amplified a ubiquitous 293 bp fragment as a positive control. During amplification, denaturation was set at 94 °C for 3 min followed by 35 cycles of denaturation, annealing and elongation (94°C for 30 s, 55°C for 30 s, 72°C for 10 s, respectively). The final elongation was set at 72 °C for 5 min.
Bacterial Infection formulation
Mosquitoes used for bioassays were not antibiotic treated. They were maintained on 6% glucose for 2–5 days post emergence. Mosquitoes were then starved overnight and fed for 24 hours on cotton balls moistened with a 6 % glucose solution containing C. violaceum at desired concentrations (bacterial cells /ml) regarding the purposes of the bioassays (Figure 1). The numbers of bacterial cells were determined by counting through improved Neubauer hemocytometer.
Bioassays
Bioassay1: Evaluation of C. violaceum entomopathogenic activity upon mosquito ingestion
To assess the virulence of C. violaceum against adults An. Coluzzii, 400 newly emerged females were exposed to feed upon cotton balls moistened with a 6 % glucose solution containing respectively 105,106,107 and 108 bacterial cells / ml in a cage (15 cm×15cm×15cm) for 24 hours. For each concentration, four replicates of 25 mosquitoes / replicate
Four replicates of control batches of mosquitoes (25 mosquitoes / replicate) were exposed to blank cotton balls soaked with 6% glucose (without any bacterial cells).
After 24 hours of exposure to treated or untreated cotton balls, mosquitoes were transferred to other new cages (15 cm×15cm×15cm) and fed with sterile 6% glucose. We recorded mortality daily over two weeks. Cadavers were immediately removed from their cages and each was washed once for 20 seconds with 1% sodium hypochlorite and twice with sterile distilled water for 40 seconds. Washed cadavers were individually crushed in 200 µl of sterilized phosphate saline buffer (PBS), homogenized. Hundred microliter (100 µl) of each homogenate was then plated onto 2 different media (Chocolate + polyvitek agar and Bromocresol purple agar). Infection from C.violaceum was then confirmed 48 hours after incubation and also using VITEK2 system.
Bioassay 2: Blood feeding choice tunnel to assess the impact of C.violaceum infection on mosquito host seeking blood-feeding propensy
Fifty non-blood-fed infected or uninfected female mosquitoes (An. coluzzii) for each of the four replicates of the tunnel test were released into the tunnel to evaluate the impact of C. violaceum infection on mosquito host seeking blood feeding propensy. The bacterial dose (106 bacterial cells / ml) was used for the tunnel bioassay. This dose is suboptimal for killing An. Coluzzii, with resulting LT50 of 9 days post infection. Three to nine days post infection mosquitoes were used for the bioassays. The control mosquitoes were infected with blank distilled water solution without any bacteria. We followed the protocol described by Bilgo et al. in 2017 [21] with some modifications. The tunnel is basically a 60-cm long; glass choice tunnel (25 cm × 25 cm area) was used for blood feeding assays (Figure 2). A 25-cm square of polyester netting was fitted at one end of the tunnel as a compartment. A netted barrier was placed one-third along the length of the glass tunnel separating the tubes into short and long sections. The barrier was 400 cm2 (20 × 20 cm), with nine 1 cm diameter holes for passage; one hole was located at the center of the square and the other eight were equidistantly located 5 cm from the border. This choice chamber is designed as a miniaturized proxy for a traditional West African home. The largest section of such a home is the veranda that serves as a sitting area. This corresponds to the first compartment of the tunnel (40 cm long). The second smaller part of a traditional house is the bedroom where residents sleep under bed netting corresponds to the smaller compartment of the tunnel (20 cm long). We placed the guinea pig within this compartment to represent a sleeping occupant at night (Figure2). Fifty non-blood-fed female mosquitoes (An. coluzzii) were released into the long section of the tunnel. In this design, female mosquitoes are normally attracted through the barrier into the smaller compartment by the guinea pig to blood feed. In full darkness between 6 pm and 6 am, mosquitoes interested in blood feeding were free to fly through the tunnel, locate the holes and pass through them to reach the guinea pig. The location of mosquitoes after this period was recorded, and those in the section closest to the guinea pig were considered to have interest in blood feeding. Mosquitoes were removed from each section of the tunnel and counted separately.
The tunnel bioassays were carried out at 27.34 °C average temperature (range 27.06 °C to 27.60 °C). The average relative humidity was 76.60% (range 75.90 % to 77.00 %). The mortality during the assay was recorded, but only live mosquitoes were considered for analysis.
Bioassay 3: Determination of the effect of C. violaceum infection on mosquito fecundity
An overall 75 three days old inseminated females mosquitoes were exposed to C. violaceum at 106 bacterial cells / ml according to the infection procedure described above. Control batches of inseminated female mosquitoes were exposed to blank sugar meal without any bacterial cells. For control and treated mosquitoes, as field mosquitoes of this species do not mate enough in captivity, after adult emergence, inseminations using a force-mating technique between virgin males and females were done according the detail protocol described in MR4 2014 [22]. Control and treated female mosquitoes received then two blood meals, 48hours and 72 hours after insemination. As regard the blood feeding on rabbit, we have ensured that the abdomens of mosquitoes are filled with blood. After the two blood meals, each female was transferred in an individual small cup containing a blotting paper. A layer (~1 cm) of tap water has stayed on the roof of the paper in order to promote eggs laying and hatching. The number of eggs laid per mosquito was scored for the following seven day. The viable larvae of first instars from eggs were also counted 2-3 days after hatching.
Ovary dissection
We therefore attempted to look at the impact of Chromobacterium violaceum on the development of ovaries within both infected inseminated females and uninfected mosquitoes. After the blood meals as above, inseminated females were anesthetized onto cold (−20°C) for five to ten minutes. Then we dissected individually their ovaries under a microscope as described in detailed in MR4 [20] (An overall 75 mosquitoes). The aspects of the ovaries were examined under microscope at 40 times magnification. Leica software (LAS-EZ-V3-3-0 for PC) was used to take pictures of the aspects of ovaries
Data analysis:
All data were entered into Microsoft Windows Excel 2013, checked for accuracy, then imported to R studio version 2.11.1 for data manipulation, visualization and statistical analysis (Supplemental file 2). Using Fisher’s exact test, P<0.05 was accepted as statistically significant.
The main parameters were calculated at each time point as below:
For these main parameters the standards errors (SE) for all replicates were calculated using the library plyr of R 3.2.4:
LT80 survival for treatments and concentrations were determined using generalized linear model (GLM) approach.
Pairwise t. test comparisons with correction of ohm were used to compare the mean number of mosquitoes per treatments that blood fed, the number of eggs laid and the number of egg hatched respectively. For all bioassays, mosquitoes were considered alive if they could stand upright and dead if they were unresponsive to stimuli following the 2013 recommendations by the WHO Pesticides Evaluation Scheme [23].
Ethics statement
All experiments with guinea pigs and rabbits were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health [24]. In addition, the protocols followed the IRSS Animal Welfare Assurance A5926-01. Trained personnel and veterinarians cared for animals involved in this study and all efforts were made to minimize suffering. All works with C. violaceum were performed under biosafety containment level II requirements.