A CRISPR-Cas9 gene drive system targeting female reproduction in the malaria mosquito vector Anopheles gambiae

Abstract

Gene drive systems that enable super-Mendelian inheritance of a transgene have the potential to modify insect populations over a timeframe of a few years. We describe CRISPR-Cas9 endonuclease constructs that function as gene drive systems in Anopheles gambiae, the main vector for malaria. We identified three genes (AGAP005958, AGAP011377 and AGAP007280) that confer a recessive female-sterility phenotype upon disruption, and inserted into each locus CRISPR-Cas9 gene drive constructs designed to target and edit each gene. For each targeted locus we observed a strong gene drive at the molecular level, with transmission rates to progeny of 91.4 to 99.6%. Population modeling and cage experiments indicate that a CRISPR-Cas9 construct targeting one of these loci, AGAP007280, meets the minimum requirement for a gene drive targeting female reproduction in an insect population. These findings could expedite the development of gene drives to suppress mosquito populations to levels that do not support malaria transmission.

Figure 1. Gene disruption by homology directed repair at three separate loci causes recessive female sterility.

A) A plasmid-based source of either a TALEN or Cas9 coupled with a gRNA induces a double stranded break at the target locus. A plasmid (hdrGFP) containing regions of homology immediately upstream and downstream of the cut site acts as a template for homology-directed repair. Internal to the homology regions a 3xP3::GFP cassette identifies hdrGFP integration events and two attP sites facilitate secondary modification of the locus through RMCE. B) PCR was used to confirm the targeted loci in wild type individuals as well as those homozygous and heterozygous for the hdrGFP allele. The primer pair used is indicated in A (blue arrows). C) Counts of larval progeny from individual females homozygous or heterozygous for hdrGFP alleles mated to wild type males. Heterozygous docking lines for all 3 loci showed at least full fertility compared to wild type females. A minimum of 20 individuals were tested for each line. Vertical bars represent the mean and error bars the standard error of the mean.

Figure 2. CRISPR^h alleles inserted at female fertility loci show highly efficient gene drive and can spread in a caged population

Recombinase-mediated cassette exchange (RMCE) was used to replace the GFP transcription unit in hdrGFP docking lines with a CRISPR homing construct (CRISPR^h consisting of a 3xP3::RFP marker, Cas9 under the transcriptional control of the vasa2 promoter and a gRNA under the control of the ubiquitous U6 PolIII. The gRNA cleaves at the non-disrupted wild type allele. Repair of the cleaved chromosome through HDR leads to copying of the CRISPR^h allele and homing. B) Confinement of homing to the germline should lead to super-Mendelian inheritance of a homing construct (indicated in red) that, when targeting a haplosufficient somatic female fertility gene, will reduce the number of fertile females. C) High levels of homing at all three female fertility loci were observed. Male or female CRISPR^h/+ heterozygotes were mated to wild type. Progeny from individual heterozygous females were scored for the presence of the RFP linked to the CRISPRh construct. A minimum of 35 females were analysed for each cross. The average rate of RFP+ individuals recovered per mated female is shown as a percentage. D) and E) counts of eggs and hatching larvae for the individual crosses revealed a strong fertility effect in heterozygous CRISPR^h/+ females (D) that was not revealed in equivalent heterozygous males. F) Dynamics calculated using recurrence equations (15) in Deredec et al. (2008), using the observed homing rates in males and females and effects on female fertility. We assume no fitness effects in males and that the initial release consists of heterozygous males equal to 10% of the pre-release adult male population (i.e. 5% of the overall population). The model assumes discrete generations (one per month) and random mating, and does not account for evolution of either the CRISPR allele or the target sequence. G) Increase in frequency of CRISPR^h allele in cage population experiments. An equal number of CRISPRh/+ and wild type individuals were used to start a population and the frequency of individuals containing a CRISPR^h allele was recorded in each subsequent generation. Black line shows deterministic prediction based on observed parameter values (homing rates 98.4%, heterozygous female fitness of 9.3%, homozygous females completely sterile), assuming no fitness effects in males. Grey lines show results from 20 stochastic simulations assuming 300 males and 300 females are used to start the next generation, females mate randomly with a single male, and 15% of females fail to mate, using random numbers drawn from the appropriate multinomial distributions. Red line shows results from 2 replicate cages.