A population modification gene drive targeting both Saglin and Lipophorin impairs Plasmodium transmission in Anopheles mosquitoes

A parental cage was assembled containing only heterozygous transgenic mosquitoes (G0, transgene frequency = 50%). Neonate larvae of subsequent generations except G5, 6, 8, 9 (N indicates the number of larvae analysed) were analyzed using COPAS flow cytometry. Gates were drawn on COPAS diagrams around clouds of larvae corresponding to homozygous, heterozygous and negative individuals according to the intensity of GFP fluorescence, and the corresponding percentage of objects in each gate was recorded. Percentages were corrected to exclude objects not corresponding to larvae. Transgene frequency per 100 chromosomes dropped from 50% in G0 to 11.35% in G17, an average loss of 2.3% per generation.

Indicated identical numbers of virgin transgenic and WT females were mixed in cages with WT males. After blood feeding, neonate larvae produced by each cage were analyzed by flow cytometry (COPAS) and the numbers of GFP fluorescent and negative larvae were counted using WinMDI software on COPAS files. Identical fertility of the two categories of females would produce 50% of GFP positive progeny. a, b after the replicate number indicate the first and second egg batch, respectively, from the same mosquitoes. Replicate 1 was composed of smaller mosquitoes, due to higher density larval rearing. All other replicates were performed with mosquitoes of standard size. Replicate 4b was performed with older mosquitoes, with >50% of females already dead.

Mice are grouped by categories according to the genotype of the Plasmodium-carrying mosquitoes biting them or by the strain of parasite infecting them. Groups of 10 infected mosquito females (transgenic or their wild-type siblings grown in the same culture), were allowed to bite a mouse. Colored cells indicate the stage at which infected mice were sacrificed to prevent the appearance of malaria symptoms. n.d=not determined. ‘D’ in the third column indicates a mouse bitten by transgenic mosquitoes that showed a clearly delayed parasitemia compared to the controls. In two cases (indicated in third column), the mouse was bitten by fewer than 6 WT mosquitoes.

Each strip of panels on successive pages corresponds to an independent caged mosquito population. Each panel is a COPAS analysis snapshot of neonate larvae at the indicated generation, obtained as shown on the first page. Successive panels provide a global view of each transgene’s evolution on samples of 1000–4000 neonate larvae at each generation, with fluorescence intensity correlating with transgene copy number (GFP marks the Lp::Sc2A10 transgene on the Y axis, DsRed marks Sag-GD^vasa on the X axis). Up to 9 partially overlapping distinct larval populations corresponding to the 9 possible genotypes (0, 1, or 2 copies of each fluorescence, sketched on page 1 and on selected panels duplicated below the original panel) can be resolved on COPAS diagrams. Dots indicated in population 1 panels correspond to debris (dead larvae, egg shells) and are difficult to distinguish from non-fluorescent larvae unless sorting them for verification under the microscope. Yellow arrow in population 2, G14 indicates the appearance of GFP negative larvae that represent a Lp homing refractory mutant (see text). Tracking Populations 1 and 2 started shortly after initial transgenesis, by crossing G2 [Sag-GD^vasa / Y; Lp::Sc2A10 / +] males to WT females (considered the G0 cross for temporal tracking of genotypes). Population 3 was assembled by mixing 20 transgenics from generation 12 of population 2 with 190 WT. Population 4 was assembled by mixing 60 transgenics from generation 12 of population 2 with 189 WT. In these new G0 mixes reproductive success of transgenics proved higher than that of the WT (due to differences in the quality of the parental mosquitoes), as G1 pupae consisted of 189 GFP transgenics for 358 negatives = 34.5% instead of expected 10% (population 3) or 346 GFP transgenics for 304 negatives = 53.2% instead of expected 24.1% (population 4). In all populations, note the initial rapid convergence of genotypes towards Lp::Sc2A10 (GFP) homozygosity, the rapid initial disappearance of Sag-GD^vasa (DsRed) positives lacking GFP, and the persistence of a fraction of DsRed negative larvae.