Fig. 1. Reversible colonisation allows the production of germ-free mosquitoes.

a Schematic representation of the experimental setup. Eggs are surface sterilised to give rise to axenic larvae. A bacterial culture of the E. coli wild-type (gnotobiotic control) or auxotrophic strain is added to axenic larvae to obtain gnotobiotic larvae. To support the growth of auxotrophic bacteria, meso-diaminopimelic acid (m-DAP) and D-alanine (D-Ala) are provided throughout the whole larval developmental stage. After transferring pupae carrying auxotrophic E. coli to a new medium lacking m-DAP and D-Ala, germ-free mosquitoes are obtained. b Proportion of larvae developing to adulthood (striped bars) and of larvae dead or stalled in the development (full bars) after conventional rearing (CR, black), rearing with wild-type E. coli (WT, orange) and auxotrophic E. coli (AUX, blue). CR larvae developed from non-sterilised eggs in tap water on autoclaved fish food. c Colony forming units (CFU) quantification of bacterial loads (upper chart) and prevalence (lower chart) in larvae at third instar (Day 3), early fourth instar (Day 4), late fourth instar (Day 5) and in pupae (Day 6). d CFU quantification of bacterial loads (upper chart) and prevalence (lower chart) in 0–3 day old mosquitoes. e Detection of bacterial DNA via qPCR in midguts or whole body of 7–9 day old adult mosquitoes. 0 reflects non-detection or detection below the negative control. In (b–d), data show mean ± SEM of three independent replicates. The exact number of individuals analysed per condition and replicate is indicated in each panel. In (e) data show mean ± SEM of three independent replicates. The number of individuals pooled in each replicate is indicated in the figure. Statistical significance was determined with generalised linear mixed models and least square means with Bonferroni correction. Exact p values are indicated in the figure. See Table S1 for detailed statistical information. Source Data are provided as a Source Data file.