Suppressing mosquito populations with precision guided sterile males

Abstract

The mosquito Aedes aegypti is the principal vector for arboviruses including dengue/yellow fever, chikungunya, and Zika virus, infecting hundreds of millions of people annually. Unfortunately, traditional control methodologies are insufficient, so innovative control methods are needed. To complement existing measures, here we develop a molecular genetic control system termed precision-guided sterile insect technique (pgSIT) in Aedes aegypti. PgSIT uses a simple CRISPR-based approach to generate flightless females and sterile males that are deployable at any life stage. Supported by mathematical models, we empirically demonstrate that released pgSIT males can compete, suppress, and even eliminate mosquito populations. This platform technology could be used in the field, and adapted to many vectors, for controlling wild populations to curtail disease in a safe, confinable, and reversible manner.

Fig. 1. Validation of pgSIT target genes.

Cas9/gRNA-mediated disruption of βTub or myo-fem results in (A–C) male (♂) sterility or (D, E) female (♀) flightlessness, respectively. Schematics of genetic crosses to assess the efficiency of (A) βTub or (D) myo-fem disruption in the F₁ transheterozygous progeny. B Bar graph indicating the percent of fertile progeny for each of the various progeny genotypes using gRNA^βTub#7 line. C Imaging of seminal fluid from wildtype (WT) and gRNA^βTub#7/+;Cas9/+ mosquitoes, showing the difference in spermatid elongation caused by the disruption of βTub (Supplementary Video 1). Nonelongated spermatid phenotype was observed in each examined F₁ transheterozygous ♂. E Bar graph showing percent of fertile and flight-capable mosquitoes in each cross using gRNA^myo-fem#1 line (Supplementary Videos 1–3). F Imaging showing the specific wing posture phenotype induced by the myo-fem disruption in ♀’s, but not in ♂’s, in which the resting wings were uplifted. Data from both paternal Cas9 crosses (Cas9♂ × gRNA♀) and maternal Cas9 crosses (Cas9♀ × gRNA♂) are shown (Supplementary Fig. 2, Supplementary Fig. 3, and Supplementary Table 3). Bar plots show means ± one standard deviation (SD) over at least three (n ≥ 3) biologically independent F₁ progeny groups, and mean and SD values rounded to a whole number. A two-sided F test was used to assess the variance equality. Statistical significance of mean differences was estimated using a two-sided Student’s t test with unequal or equal variance. (P ≥ 0.05 ^ns, P < 0.05*, P < 0.01**, and P < 0.001***). Source data are provided as a Source Data file.

Fig. 2. Genetic characterization of pgSIT.

A The pgSIT cross between double-homozygous gRNA ♂’s harboring both gRNA^βTub#7 and gRNA^myo-fem#1 (termed: gRNA^{βTub + myo-fem}) and homozygous Cas9 ♀’s. The pgSIT cross was initiated reciprocally to generate F₁ transheterozygous progeny carrying either maternal or paternal Cas9 (Supplementary Fig. 4). B Bar graphs comparing the percentage of transheterozygous and heterozygous Cas9 or gRNA progeny to those of WT (data can be found in Supplementary Table 4 and Supplementary Video 4) over at least three (n ≥ 3) biologically independent F₁ progeny groups. C Experimental setup to determine whether prior matings with pgSIT^♂’s suppresses WT female (♀) fertility. WT ♀’s were cohabitated with pgSIT^♂’s for 2, 6, 12, 24, or 48 h then WT ♀’s were transferred to a new cage along with WT ♂’s and mated for an additional 2 days. The ♀’s were then blood-fed and individually transferred to a vial. Eggs were collected and hatched for fertility determination. Following this, nonfertile ♀’s were then placed back into cages along WT ♂’s for another chance to produce progeny. This was repeated for up to five gonotrophic cycles, and the percentage of fertile ♀’s in each group of 50 ♀’s was plotted (Supplementary Table 8). The plot shows the fertility of three biologically independent groups of 50 WT ♀ (n = 3) for each experimental condition. D Flight activity of 24 individual mosquitoes (n = 24) was assessed for 24 h using a vertical Drosophila activity monitoring (DAM) system, which uses an infrared beam to record flight (Supplementary Table 6 and Supplementary Video 5). E To quantify the attractiveness of ♂’s to ♀’s for mating, we used a mating-behavior lure of a tone mimicking ♀ flight. A 10-s 600 Hz sine tone was applied on one side of the cage, and a number of mosquito ♂’s landing on the mesh around a speaker was scored. Heatmaps were generated using Noldus Ethovision XT (Supplementary Table 7 and Supplementary Video 6). The experiment was repeated six times (n = 6) using fresh groups of 30–40 ♂’s. Point plots (C–E) show biological replicates and means ± SDs. A two-sided F test was used to assess the variance equality. Statistical significance of mean differences was estimated using a two-sided Student’s t test with unequal or equal variance. (P ≥ 0.05 ^ns, P < 0.05*, P < 0.01**, and P < 0.001***). Source data are provided as a Source Data File.

Fig. 3. Multigenerational cage trials demonstrating efficient population suppression.

A To generate sufficient mosquito numbers, three lines were raised separately, including homozygous Cas9, double-homozygous gRNA^{βTub + myo-fem}, and WT. To generate pgSIT progeny, virgin Cas9 ♀’s were genetically crossed to gRNA^{βTub + myo-fem} ♂’s, and eggs were collected. B To perform multigenerational population cage trials of pgSIT, two strategies were employed: release of eggs (B, top panel); release of mature adults (B, bottom panel). For both strategies, multiple pgSIT:WT release ratios were tested, including: 1:1, 5:1, 10:1, 20:1, and 40:1. Each generation, total eggs were counted, and 100 eggs were selected randomly to seed the subsequent generation. The remaining eggs were hatched to measure hatching rates and score transgene markers. This procedure was repeated after each generation until each population was eliminated (Supplementary Table 16). C Multigenerational population cage data for each release threshold plotting the proportion of eggs hatched each generation. Source data are provided as a Source Data file.

Fig. 4. Model-predicted impact of releases of pgSIT eggs on Ae. aegypti population density and elimination.

A Releases were simulated on the motu of Onetahi (73.8 hectares), Teti ‘aroa, French Polynesia, a field site for releases of Wolbachia-infected ♂ mosquitoes, using the MGDrivE simulation framework and parameters described in Supplementary Table 17. Human structures are depicted and were modeled as having an equilibrium population of 16 adult Ae. aegypti each. B Weekly releases of up to 200 pgSIT eggs per WT Ae. aegypti were simulated in each human structure over 10–24 weeks. The pgSIT construct was conservatively assumed to decrease ♂ mating competitiveness by 25% and adult lifespan by 25%. Elimination probability was calculated as the percentage of 200 stochastic simulations that resulted in local Ae. aegypti elimination for each parameter set. Sample time series depicting WT ♀ Ae. aegypti population density is depicted above and below the heatmap. C Elimination probability (given 18 weekly releases of 200 pgSIT eggs per WT Ae. aegypti) is depicted for a range of pgSIT ♂ fitness profiles. Elimination is possible for a wide range of reductions in ♂ mating competitiveness (0–50%) and adult lifespan (0–50%) for an achievable release scheme. A Source Data file is provided.