Development and physiological effects of an artificial diet for Wolbachia-infected Aedes aegypti

The endosymbiotic bacterium Wolbachia spreads rapidly through populations of Aedes aegypti mosquitoes, and strongly inhibits infection with key human pathogens including the dengue and Zika viruses. Mosquito control programs aimed at limiting transmission of these viruses are ongoing in multiple countries, yet there is a dearth of mass rearing infrastructure specific to Wolbachia-infected mosquitoes. One example is the lack of a blood meal substitute, which accounts for the Wolbachia-specific physiological changes in infected mosquitoes, that allows the bacterium to spread, and block viral infections. To that end, we have developed a blood meal substitute specifically for mosquitoes infected with the wMel Wolbachia strain. This diet, ADM, contains milk protein, and infant formula, dissolved in a mixture of bovine red blood cells and Aedes physiological saline, with ATP as a phagostimulant. Feeding with ADM leads to high levels of viable egg production, but also does not affect key Wolbachia parameters including, bacterial density, cytoplasmic incompatibility, or resistance to infection with Zika virus. ADM represents an effective substitute for human blood, which could potentially be used for the mass rearing of wMel-infected A. aegypti, and could easily be optimized in the future to improve performance.


Methods:
We performed a pilot experiment comparing different types of solvents in order to see if any were more suitable for feeding as part of an artificial diet. In this experiments all diets included the isolated milk whey protein at a concentration of 150 mg/mL, ATP at a final concentration of 1 mM, and 3 mL of 1 of 5 different solvents. A whole human blood treatment (WB) was used as a control.
The 5 solvents were as follows: (1) 1X APS, as described in the main paper.
Mel mosquitoes were reared to adulthood as described in the main paper, and were offered the artificial diet at 5 days-post eclosion, after approximately 20 hours of starvation.

Results and discussion:
Feeding rates (proportion fed) were as follows: Fecundity data were then obtained following the procedures described in the main paper. In looking at the data for this experiment, we decided to focus on APS for further experimentation as it had the highest feeding rate of the 5 solvents, produced fairly high levels of fecundity, and because it did not contain rapidly perishable organic compounds such as powdered multivitamins or sucrose.

Methods:
After performing the blood fraction experiment described in Figure 1, and the MC protein concentration experiments described in Figure 2 of the main paper, we performed a further group of experiments involving the addition of MC protein at a concentration of 125 mg/mL to whole human blood (WB), human plasma (PLS), or human red blood cells (RBC). WB and MC (MC protein 125 mg/mL) were used as control treatments. Protein-supplemented diets were MCW (+WB), MCP (+PLS), and MCR (+RBC). Blood was separated by centrifugation at 1500 rpm for 5 mins, and then mixed with the protein by vortexing. For MCW and MCP diets, 375 mg of MC was mixed with 3 mL of either WB or PLS. For the MCR diet, the mixture of 3 mL of RBC and 375 mg of protein was too viscous to dissolve fully, so we had to add 1 mL of APS. 3 mL of each of these solutions was fed to Mel mosquitoes on a waterbath system. Mosquito rearing, and fecundity and hatch rate experiments were performed as described in the main paper. Two experimental replicates were performed, and data were compared by Kruskal-Wallis ANOVA and Dunn's multiple comparisons test.

Results and discussion:
We observed that all of the MCW, MCP and MCR had similar levels of fecundity to the WB treatment. Additionally these 3 diets had significantly higher levels of fecundity than the MC treatment, but WB did not (Kruskal-Wallis; P < 0.0001). The hatch rate for WB was significantly higher than all other treatments, while the hatch rates of the MCW and MCR treatments was significantly higher than those of the MC and MCP treatments. This suggested to us that the addition of protein to the RBC fraction was sufficient to overcome the decreased fecundity associated with feeding RBC alone (Fig.  2). However, the fact that MCP diet, which consisted of PLS and MC protein, was still associated with a low hatch rate (similar to what was seen when feeding only PLS), suggested that some component in the RBC fraction was still necessary for obtaining a high hatch rate.

Methods:
We performed a range finding assay to determine the optimal quantity of Alfaré formula to include in our diet. Data on part of these experiments (Diets F1 and F2) have been included in the main paper, however we also examined an additional 2 concentrations of formula (50 mg/mL and 75 mg/mL). Feeding experiments were conducted as described in the main paper. Fecundity data were compared by one-way ANOVA, and Tukey's multiple comparisons test. Hatch rate data were compared by Kruskal-Wallis ANOVA and Dunn's multiple comparisons test. These experiments were repeated twice.

Results and discussion:
While the ANOVA for fecundity data showed there was a significant difference between treatments (F = 2.784, P = 0.0267), no pairwise comparisons were significantly significant. However, we did observe a trend towards increased fecundity for the diets with 15 mg/mL (Median -50 eggs) and 25 mg/mL (Median -48.5 eggs) formula, compared to diet with only MC protein (Median -35 eggs). Likewise, there was no significant difference in hatch rate between the treatments, but given the increase in fecundity for the 15 and 25 mg/mL diet, there was an associated increase in the number of larvae produced, when compared to the control diet. For that reason, we chose to include those two diets in the assays involving RBC.