CRISPR/Cas9 knockout of female-biased genes AeAct-4 or myo-fem in Ae. aegypti results in a flightless phenotype in female, but not male mosquitoes

Aedes aegypti is a vector of dengue, chikungunya, and Zika viruses. Current vector control strategies such as community engagement, source reduction, and insecticides have not been sufficient to prevent viral outbreaks. Thus, interest in novel strategies involving genetic engineering is growing. Female mosquitoes rely on flight to mate with males and obtain a bloodmeal from a host. We hypothesized that knockout of genes specifically expressed in female mosquitoes associated with the indirect flight muscles would result in a flightless female mosquito. Using CRISPR-Cas9 we generated loss-of-function mutations in several genes hypothesized to control flight in mosquitoes, including actin (AeAct-4) and myosin (myo-fem) genes expressed specifically in the female flight muscle. Genetic knockout of these genes resulted in 100% flightless females, with homozygous males able to fly, mate, and produce offspring, albeit at a reduced rate when compared to wild type males. Interestingly, we found that while AeAct-4 was haplosufficient, with most heterozygous individuals capable of flight, this was not the case for myo-fem, where about half of individuals carrying only one intact copy could not fly. These findings lay the groundwork for developing novel mechanisms of controlling Ae. aegypti populations, and our results suggest that this mechanism could be applicable to other vector species of mosquito.


Introduction
The yellow fever mosquito Aedes aegypti is a vector for many viruses of medical significance, such as dengue, Zika, chikungunya, and yellow fever, and they can be found in tropical, subtropical, and temperate regions of the world [1].Only female Ae.aegypti bite to obtain a blood meal, which is required for egg production.After hatching from the embryo, Ae. aegypti like all other mosquito species will progress through the aquatic larval and pupal stages of their life, before emerging as an adult from the pupal casing to fly away [2,3].
Due to a lack of safe, effective vaccines for most viruses transmitted by Ae. aegypti, save yellow fever virus [4], control efforts largely focus on reducing vector abundance, and include source reduction, and chemical methods like insecticides or larvicides [3,5].The short-term effect and high financial cost, along with the need for trained staff, presents challenges to the implementation, scaling, and maintenance of these control methods [3,6].With chemical methods, additional concerns relating to the emergence of resistance and effects on off-target species are increasing [6].Because of these limitations, the need for novel vector control strategies is growing.
A function that is critical for both reproduction and survival in female mosquitoes is flight, as flight is required for mating, obtaining a blood meal, and escaping from aquatic breeding sites after eclosion.AeAct-4 was identified previously as a female-and pupal-specific gene, with expression in the indirect flight muscles [39].Other work in Ae. aegypti has identified a male-specific actin gene [40] and male-specific myosin gene [41] related to flight, while in Drosophila melanogaster, Flightin has been identified as a flight-associated protein that is hypothesized to function in the indirect flight muscles of both sexes by interacting with myosin filaments or by modulating actin-myosin interactions [42].Knockout of the male-specific myosin gene has shown that it is needed for male flight [43], and offers up the possibility that male and female flight in Aedes mosquitoes is controlled separately by these sex-specific genes.
To determine the importance of selected actin and myosin genes to Ae. aegypti female flight, we used CRISPR/Cas9 to generate heritable loss-of-function alleles in AeAct-4 and a female-biased myosin gene which we termed myo-fem, along with a third gene, Aeflightin, that is expressed in both males and females.Phenotypic analysis of individuals homozygous for each introduced mutation in AeAct-4 or myo-fem confirmed that flight defects were both complete and restricted to females.Males homozygous for either mutation were capable of mating and producing viable progeny.While AeAct-4 knockout males could fly and mate, their ability to compete for females was reduced compared to wild-type males; this was not the case for loss of myo-fem.Disruption of Aeflightin was associated with loss of flight in both sexes.Phylogenetic analysis of AeAct-4 suggests that other mosquito species genomes likely contain femaleand male-biased actin genes as well.These data support the pursuit of novel genetic strategies geared specifically for disrupting female flight in Ae. aegypti and other vector species of mosquito.

Insect rearing
The Liverpool strain of Ae. aegypti was used for embryonic microinjections and outcrossing of mutant individuals.All mosquitoes were reared at 28˚C and 75-80% humidity, with a 14/10 h light/dark light cycle.Ground up fish food (Tetra, Blacksburg, VA) was supplied throughout the aquatic developmental stages, and a cotton ball soaked with 10% sucrose solution was supplied during the adult stage.Flightless mosquitoes were supplied with raisins as the source of sucrose.Defibrinated sheep blood (Colorado Serum Company, Denver, CO) was offered for blood feeding via a parafilm membrane feeder.Videos were taken with a Canon Rebel T3i digital camera.

Computational analysis
Publicly available mapped RNA-seq data from Akbari et al. (2013) was retrieved from Vector-Base [44].Raw counts for each gene of interest and all paralogs with �80% amino acid similarity were obtained using featureCounts using only uniquely mapped reads [45].Raw counts data were linear normalized based on transcript length and library size to obtain fragments per kilobase per million reads (FPKM) data.FPKM data were transformed to avoid negative values: [log10(1+FPKM)].Transformed data were used to generate a heat map with Morpheus (Morpheus, https://software.broadinstitute.org/morpheus).A maximum score of 3.8 was set when generating the heat map based on the highest expression value across all samples/ timepoints.
For phylogenetic and molecular evolutionary analyses, VectorBase [44] and FlyBase [46] were used to perform a BLASTP search for orthologs of AeAct-4 and all paralogs with �80% amino acid similarity (maximum e-value 1e -3 , word size 3).All sequences obtained were aligned with MUSCLE [47] and compared using the Neighbor-Joining method in MEGA version X [48].Evolutionary distances were computed using the Poisson correction method and are in the units of the number of amino acid substitutions per site.All ambiguous positions were removed for each sequence pair (pairwise deletion option), with a total of 376 positions in the final dataset.
All bar graphs were generated, and Chi square analyses performed, using GraphPad Prism (version 8 for Windows, GraphPad Software, La Jolla, CA, www.graphpad.com).

Guide RNA design and synthesis
Guide RNAs were designed by hand with the DNASTAR SeqBuilder Pro software (Madison, WI), using the appropriate gene sequence acquired through VectorBase [44].Primers used to make each sgRNA were ordered through IDT.Guide RNA synthesis was performed as previously described [34].Briefly, Q5 High-Fidelity DNA Polymerase (New England BioLabs Inc., Ipswich, MA) was used for the PCR reaction, followed by the NucleoSpin Gel and PCR Clean-Up kit protocol (Machery-Nagel, Bethlehem, PA), the MEGAscript T7 Transcription kit protocol, and the MEGAclear Transcription Clean-Up kit protocol (Thermo Fisher Scientific, Waltham, MA).All sgRNAs were quantified, aliquoted, and stored at -80˚C.A list of oligos used to make each sgRNA are listed in S2 Table.

Embryo microinjections
The generation and identification of knockout strains followed essentially from our previously published protocols [49].Briefly, borosilicate glass capillaries (World Precision Instruments Inc., Sarasota, FL) were pulled and beveled using the Sutter P-2000 Micropipette Puller and Sutter BV-10 Micropipette Beveller (Sutter Instrument Co., Novato, CA).Embryo microinjections were performed using a Leica DM 1000 LED Micromanipulator (Leica Biosystems, Buffalo Grove, IL) and FemtoJet 4i Microinjector (Eppendorf, Hauppauge, NY).Purified Cas9 protein (400 ng/μl) (PNA Bio, Thousand Oaks, CA) and sgRNAs (100 ng/μl) were combined into injection mixes, incubated at 37˚C for 30 minutes, and centrifuged at max speed (18,213 g), 4˚C, for a minimum of 45 minutes.Injections were performed into the posterior end of embryos that were less than three hours old."AeAct-4 Exp.1" injection mixes included kmo sgRNAs, while "AeAct-4 B Exp. 2" injection mixes included only site B sgRNAs.Injected embryos were either harvested at 24 hours (for embryo assays) or hatched after five days.Hatched G 0 survivors were outcrossed to either the wild type Liverpool strain (for AeAct-4 and myo-fem) or the kmo knockout strain (for Aeflightin).Mutant males (n = 10-25) were selected at each subsequent generation for continued outcrossing through G 3 (for Aeflightin) or G 5 (for AeAct-4 and myo-fem) before intercrosses were performed.

DNA extraction, PCR, HRMA, and Sanger sequencing
Genomic DNA was extracted from non-injected or sgRNA-injected Liverpool embryos following the Nucleospin Tissue kit protocol (Machery-Nagel, Bethlehem, PA).Embryo assays were performed with the LightScanner Master Mix kit (Idaho Technology Inc., Salt Lake City, UT), and mutant detection was performed on adult legs with the Phire Animal Tissue Direct PCR kit (Thermo Fisher Scientific, Waltham, MA) with added LCGreen Plus+ Melting dye (Idaho Technology Inc., Salt Lake City, UT).All samples were amplified with the C1000 Touch Thermal Cycler (Bio-rad, Hercules, CA) before being analyzed with the LightScanner Call-IT 2.0 software on the LightScanner instrument (Idaho Technology Inc., Salt Lake City, UT).Sanger sequencing was performed at the Laboratory for Genomic Technologies (Institute for Plant Genomics and Biotechnology, Texas A&M University, College Station, TX) and chromatograms were analyzed using Chromas software (Technelysium, Australia).A list of primers used are located in S1 Table.

Flight determination
To assess flight, pupae were placed in plastic ketchup containers (with the stick of a cotton swab or Q-tip to aid in escaping from the water due to defects in flight) in 5-quart plastic buckets (Home Depot) designed with a mesh covering over the top and a sock for internal access on the side.The plastic lining of the bucket inhibited the mosquitoes from climbing up the sides towards the mesh covering.Flying mosquitoes that could fly up to the mesh covering were removed from the plastic bucket, and 24 hours was allowed to pass after the last pupae emerged before classifying the remaining mosquitoes as flightless.Dead adults whose flight phenotype could not be confirmed were not analyzed for a genotype.

Mating competition assays
To assess mating competitiveness between wild type and homozygous mutant (AeAct-4 and myo-fem) males, 40 WT/HOM males (based on HRMA analysis) were placed into a 170 oz container with 40 Liverpool females for mating.After at least 24 hours post-mating, the females were offered a blood meal for a minimum of 30 minutes, and engorged females were selected and separated from males.The EAgaL plate fecundity and fertility assay protocol was followed [50].Briefly, a single blood fed female was placed in each well of a 24-well plate that had been prepared by filling each well with agarose.Females were placed at least 72 hours after blood feeding and were given up to 48 hours to deposit embryos.After 48 hours, females were removed, water was added to the wells, and the embryos were monitored over the next 5 days for hatching.If present, 10 or more larval progeny were collected from each individual female that produced embryos.Amplicons derived from the AeAct-4 or myo-fem target sequence from both the male parents and the pooled larval progeny were sequenced to determine the respective starting and resulting genotypes, and thus identify the genotype of the male that mated with the female.At least three biological replicates per mutant strain were performed.

Results
Since female mosquitoes rely on flight to obtain a blood meal and mate, disrupting flight specifically in females could prevent both reproduction and the transmission of arboviruses.AeAct-4 has been characterized as a female-and pupal-specific flight protein that is expressed in the indirect flight muscles [39].Indeed, re-analysis of RNA-seq data from the developmental transcriptome of Ae. aegypti [51] confirmed that expression of AeAct-4 was highly biased towards female pupae (S1 Fig) .Of the eight paralogous actin genes, six showed expression throughout development and in pupae of both sexes, one (AAEL005964) was pupae-, but not sex-specific, and one (AAEL009451) was expressed specifically in male pupae.AAEL005656 is a paralog to the gene myo-sex (AAEL021838), a male-specific myosin gene located in the M locus on chromosome one [41] and required for male flight [43].As AAEL005656 was found to be expressed primarily in female pupae (S1 Fig) , we reasoned that it might be similarly critical for female flight, and refer to this gene as myo-fem.Finally, AAEL004249 is a 1:1 ortholog of Drosophila melanogaster, flightin, which has been shown to be expressed in the indirect flight muscles [42], with knockout resulting in a loss of flight ability [52].We refer to this gene as Aeflightin and confirm that it is expressed almost exclusively in pupae in both males and females (S1 Fig) .We reasoned that despite lacking female specificity, Aeflightin may be a good target for disrupting female flight, so long as a male-specific rescue can be performed.
For each of these three target genes, we designed multiple groups of overlapping sgRNAs, described in S2 Table with either "site" or "exon" for each group.Due to high nucleotide sequence similarity between AeAct-4 and other actin paralogs, we performed a multiple sequence alignment prior to sgRNA design.Eight paralogs with �80% nucleotide sequence identity to AeAct-4 were aligned to attempt to identify sgRNA targets that were unique to AeAct-4.Candidate guide RNA sequences were identified in regions where there was more sequence variability between AeAct-4 and the paralogs (S2 Fig) .For myo-fem, guide RNAs were designed to target the motor domain to ensure disruption of myosin function.For Aeflightin, guide RNAs were designed to each exon as we did not identify any paralogous genes in the Ae.aegypti genome.
Following in vitro synthesis, sgRNAs were complexed with Cas9 protein and injected in groups of 3-4 into Ae.aegypti embryos, which were harvested after 24 hours.To identify those gRNA batches capable of inducing strong gene disruption, genomic DNA from injected embryos was used as a template for PCR and HRMA of the target region ( For each mixture of guide RNAs that displayed editing activity, we repeated embryo microinjections and this time allowed the embryos to hatch after five days.Survivors were outcrossed to the parental Liverpool strain to obtain G 1 progeny.G 1 adults were screened via PCR, HRMA, and Sanger sequencing for out-of-frame mutations (S3 Table ).Deletions predicted to result in a frameshift mutation were recovered for AeAct-4, myo-fem, and Aeflightin (Fig 2A).Genotyped males with out-of-frame mutations (n = 10-25) were outcrossed to females from the parental strain for three (for Aeflightin) or five (for AeAct-4 and myo-fem) generations, followed by intercrosses of heterozygous individuals (Fig 2B and 2C).Backcrossing of the mutant strains was performed to reduce any CRISPR/Cas9 off-target effects and assist in the recovery from the genetic bottleneck associated with single founder events.AeAct-4 is located on chromosome three, as are three other actin paralogs (AAEL001928, AAEL005961, and AAEL005964).Therefore, we performed Sanger sequencing of each actin paralog to confirm that there were no off-target effects in these paralogs linked to AeAct-4, as these may have been maintained despite backcrossing (S3 Fig) .G 6 progeny from heterozygous intercrosses were characterized as flightless or flying (Fig 3A and S4 Table).Upon stimulation of flight, control mosquitoes could fly, while some from each test cross could not; these were therefore categorized as flightless and hypothesized to be homozygous for each targeted gene disruption.Flightless individuals appeared to have various alternative wing phenotypes when resting that differed from wild type (Fig 3B -3E).Some individuals could move or beat their wings, but with no success in flight.Other individuals could initiate flight takeoff, but not sustain flight.Agitation to provoke flight via shaking or tapping of the plastic buckets (S1-S3 Videos), or gentle spraying of condensed air were used to evaluate flight ability; flightless mosquitoes remained so regardless of the method of stimulation.
As all observations of flight behavior were made without consideration for genotype, we sought to determine whether there was a relationship between the ability of mosquitoes to fly and inheriting one or two copies of each loss-of-function mutation.A genotypic analysis of all flightless individuals and a subset of flying individuals was performed for AeAct4 and myo-fem (S4 males and all myo-fem Δ11/Δ11 males except one (S4 Table) could fly (AeAct-4 = 100%; myo-fem = 96%), while all AeAct-4 Δ10/Δ10 and myo-fem Δ11/Δ11 females were flightless (100%).Interestingly, while most heterozygous AeAct-4 Δ10/+ females could fly (97%, only two were flightless), this was not the case for myo-fem Δ11/+ , where about half of females (53%) were categorized as flightless.The flying to flightless ratio did not differ significantly from the expected ratio for AeAct-4 (p = 0.6503, Chi square analysis), suggesting the flightless phenotype was completely recessive in this case.Due to the presence of a substantial number of flightless heterozygous females for myo-fem (46 individuals) as compared to AeAct-4 (2 individuals), the flying to flightless ratio was significantly different from the expected ratio (p < 0.0001, Chi square analysis), suggesting that defects in myo-fem are not entirely recessive.From the intercross between AeAct-4 Δ10/+ parents, we observed that male genotype ratios of WT, heterozygous, and homozygous individuals did not differ from the null expectation, suggesting a lack of strong fitness cost to males associated with the mutation (p = 0.6628, Chi square analysis).In contrast, we observed a deficiency of homozygous males and a corresponding excess of heterozygotes for myo-fem (p = 0.0008, Chi square analysis).As the Aeflightin gene is tightly linked to the kmo gene involved in eye pigmentation, Aeflightin mutants were outcrossed to a kynurenine 3-monooxygenase (kmo) knockout strain [34] to help track the corresponding genotypes.This aided in phenotypic identification of homozygous kmo individuals who do not carry the Aeflightin mutation, as well as maintenance of a transheterozygous line (Fig 2C).At G 6 , recombination between kmo and Aeflightin was observed at 1-3%, consistent with expectations (S5 Table ).After scoring flight phenotypes, genotypes were determined by HRMA and sequencing.Critically, we found that all transheterozygous Aeflightin Δ4/Δ5 individuals were flightless, confirming that Aeflightin is required for flight in both male and female Ae.aegypti (Fig 3 and S5 Table).The flying to flightless ratio did not differ significantly from the expected ratio for Aeflightin (p = 0.5319, Chi square analysis), suggesting that a single copy of Aeflightin is sufficient to program flight, and the associated flightless phenotype is recessive.
As we noted that both AeAct-4 and myo-fem are expressed at low levels in male pupae [51] (S1 Fig) , we hypothesized that this may indicate these genes contribute to male flight.While our loss-of-function data suggest these gene products are not required for flight, we reasoned that they may still contribute to male mating success, which occurs in flight.Thus, we sought to compare the mating competitiveness between wild type and homozygous mutant males, an important consideration if these genes are to be potential targets in a genetic control approach such as gene drive.We took advantage of the fact that wild type and homozygous individuals cannot be distinguished by HRMA alone in our assay (requires sequencing) to perform a series of blinded mating competition experiments.After screening by HRMA, 40 males with WT/ HOM genotypes were crossed with 40 wild type females.After 1-3 days, females were offered a bloodmeal and progeny were collected from each female individually.At the same time, all HRMA amplicons were sequenced to obtain the starting percentage of each male genotype.In the case of a mating between a WT male and a WT female, all progeny would be WT, while matings between homozygous mutant males and WT females should result in progeny heterozygous for the mutation (either AeAct-4 Δ10 or myo-fem Δ11 ).HRMA and sequencing analysis was used to determine the genotype of the larval progeny, and thus, the genotype of the individual male that mated with that female (Fig 4A).Wild type males significantly outperformed AeAct-4 Δ10/Δ10 males in mating with wild type females (Fig 4B ; AeAct-4 p = 0.0018, Chi square analysis), suggesting that AeAct-4 may contribute to the mating success of male Ae.aegypti.In contrast, for myo-fem (Fig 4C) the difference in mating success was not significantly different from expectations based on starting genotypes (myo-fem p = 0.0805, Chi square analysis).We note that we did not monitor the females for remating events, thus it is possible that mutant individuals are less competitive than indicated.We also note that for myo-fem, there was more variation in the data than for AeAct-4, and this may serve to obscure real differences in competitiveness.Data for each replicate experiment is presented individually (S6 Table ), along with the expected and observed larval progeny genotypes.
After successfully knocking out flight-associated genes in Ae. aegypti, we were interested in finding out if other mosquito genomes contain sex-specific flight genes.Previous phylogenetic analysis of the male-specific myo-sex suggests that both male-and female-specific myosin flight genes evolved in Aedes after divergence from Culex [41].To determine if this was also the case for AeAct-4, we performed a phylogenetic analysis of actin protein sequences across multiple mosquito species, as well as the fruit fly Drosophila melanogaster (Fig 5).Both AeAct-4 and the male-biased actin gene AAEL009451 clustered separately with orthologs from Ae. albopictus, Culex, and Anopheles mosquitoes, suggesting a more ancient origin to these sex-biased genes.Consistent with this reconstruction, the An.gambiae gene AGAP011515 grouped with AeAct-4 and was previously found to be preferentially expressed in females [53], while AGAP001676 groups with the Aedes male actin gene, and was also found to have significantly increased expression in the male carcass (head, thorax, and abdomen, excluding the reproductive tissue) as compared to the female carcass [54].Together, these data suggest that genetic approaches targeting flight specifically in females may be broadly applicable across mosquito genera.

Discussion
Our results indicate successful knockout of three flight-specific genes, two of which are expressed predominantly in females (AeAct-4 and myo-fem).Interestingly, we found that while AeAct-4 and Aeflightin were haplosufficient, two intact copies of myo-fem appeared to be required for normal female flight.We note though that our approach allowed 24 hours after the last adult emerged for all mosquitoes to gain the ability to fly.For myo-fem, there were a few individuals who subsequently gained the ability to fly up to 48 hours after all flyers had been removed.These individuals seemed to have a delay (� 24 hours post-eclosion) in gaining flight ability, suggesting that a single copy of myo-fem, while insufficient to program the normal timing of development of the flight muscles, may be sufficient provided the female can survive long enough.If myo-fem is truly haploinsufficient, this opens the door for the development of strong synthetic sex distorters for suppressing Ae. aegypti populations [22,[55][56][57][58][59][60][61].Despite the lack of female specificity for the third gene, Aeflightin, we reason that this still represents a useful target so long as a male-specific rescue can be performed to fully restore male flight.
Flightless Ae. aegypti have been developed previously through the transgenic overexpression of the tTa transactivator specifically in the female flight muscle [62].In this case, the promoter region of AeAct-4 was used to control transgene expression, however transgenic males were found to have decreased mating competitiveness in field cage trials [63,64].Variability in the level of transgene overexpression also resulted in incomplete penetrance of the flightless phenotype.In our case, disruption of both the AeAct-4 gene and the myo-fem gene through heritable gene editing resulted in a completely penetrant phenotype without the requirement for continuous transgene expression as previously seen with AeAct-4.As both AeAct-4 and myo-fem show low levels of transcription in male pupae [51], we hypothesized that disruption of both genes would lead to reduced flight in males beyond observable differences, such as mating competitiveness.Indeed, males with mutations in AeAct-4 exhibit decreased levels of mating when compared to wild type males.Though this decrease was not observed in males with mutations in myo-fem, we interpret these data with caution due to the confined nature of the mating experiment (170 oz.container), and that despite outcrossing mutants for multiple generations, there is a possibility of a genetic bottleneck leading to reduced fitness in males, regardless of these flight gene knockouts.More rigorous follow-up experiments in larger venues that require more flight effort from test males (in larger quantities and ratios of wild type to homozygous) are likely required to conclude that myo-fem is dispensable for male mating success.Male and female Ae.aegypti are known to produce different flight tones, which are shifted to match frequencies during mating; this harmonic convergence could be a measure of male reproductive fitness [65][66][67][68] and could be heritable [69].It is plausible that AeAct-4 and myo-fem contribute towards modulating wing beat frequency in males; future experiments to investigate this using the mutant strains developed here may shed light on this interesting aspect of Ae. aegypti biology.
Disrupting flight specifically in female mosquitoes could be used to achieve sex ratio distortion of the adult population.This is conceptually similar to other sex distortion approaches such as the X-shredding system based on the I-PpoI homing endonuclease when active only during male meiosis, which has shown to be capable of producing >95% male progeny [56,60,70,71] and has been introgressed from An. gambiae to An. arabiensis [72].Other examples of targets for sex ratio distortion in An. gambiae include female reproductive genes [29,[57][58][59] and the female transcript of doublesex, which causes an intersex phenotype and complete sterility [59], however the recessive nature of these phenotypes reduces their power as sex distorters.In Drosophila melanogaster, disruption of other genes causing female fertility or embryonic lethality [22] have been shown to skew the sex ratio towards males, as does knockdown of tra-2 in Ae. aegypti [61].Meanwhile, overexpression of the male sex determining factor nix has been proposed as a method of sex distortion in Ae. aegypti [55].Population modeling has also been evaluated when considering resistance that can evolve when using a driving Y for sex ratio distortion [73].We note that the development of a sex ratio distortion approach that targets female-specific flight would allow for maximum competition for resources during larval development as well as allow active monitoring of the number of females doomed to flightlessness, while at the same time preventing the adult female from reproducing and potentially transmitting deadly pathogens.A gene drive-based sex ratio distortion approach targeting these female-specific flight genes could be implemented if males were carriers of the gene drive and released into the wild to mate with females, such that all female progeny would be affected and not survive to blood feed and transmit pathogens; all male progeny would inherit the transgene and continue to survive and mate at each subsequent generation (Fig 6).Such strategies to target sex-specific flight genes are likely to be applicable to other species of mosquito including Culex and Anopheles, as these mosquitoes also appear to encode sex-specific flight genes.

Fig 1 .
Fig 1. Development of CRISPR reagents targeting Ae. aegypti genes involved in flight.(A) Workflow for performing embryo assays.Gene models and HRMA analysis following embryonic microinjection of CRISPR/Cas9 reagents with each group of sgRNAs for AeAct-4 (B), myo-fem (C), and Aeflightin (D).For (B-D), boxes represent exons, while cross-hatched areas represent the ORF of the corresponding mRNA.For each, red triangles indicate the locations of sgRNAs that were found to successfully cleave the DNA during embryo assay, and the blue and red boxes under the gene models indicate the approximate HRMA amplicon sizes.Melt curves are displayed for clusters of sgRNAs (indicated as "site" or "exon"), with sgRNA-injected (red) and non-injected (gray) samples.https://doi.org/10.1371/journal.pntd.0008971.g001

Fig 2 .
Fig 2. Establishment and maintenance of heritable loss-of-function mutations in Ae. aegypti flight genes.(A) For each gene, the wild type (WT) and mutant sequence is shown.The PAM sites for each sgRNA used in the injection mix for the specified location are highlighted in red.Individuals containing each deletion were outcrossed through the indicated generation, at which point individuals heterozygous for the AeAct-4, myo-fem (B), and Aeflightin (C) mutations were intercrossed.For each cross, the ratio of each potential genotype expected is noted in parentheses, with the individuals with an expected flightless phenotype highlighted in red.For Aeflightin, phenotypic identification of all white-eyed pupae enables their removal before further phenotypic analysis based on flight (see Fig 4) is performed.https://doi.org/10.1371/journal.pntd.0008971.g002

Fig 4 .
Fig 4. Mating competitiveness of male Ae.aegypti deficient in AeAct-4 or myo-fem.(A) Workflow for performing larval progeny assays.The observed (Obs.) vs. expected (Exp.)number of matings for AeAct-4 (B) and myo-fem (C) males based on the sequenced genotypes of pooled larval progeny obtained from the male mating competitiveness assays.Wild type is defined as +/+ and homozygous mutants are defined as Δ10/Δ10 (for AeAct-4) or Δ11/Δ11 (for myo-fem).Each data point represents one replicate, bar height represents the mean of all replicates, and the error bars indicate standard deviation.Chi square was performed for statistical analysis.https://doi.org/10.1371/journal.pntd.0008971.g004

Fig 6 .
Fig 6.An M-locus-linked sex distorter gene drive targeting female-specific flight genes.Male mosquitoes modified to contain a site specific nuclease targeting a haploinsufficient female flight gene (myo-fem) in the M-locus would be released to mate with wild type females.All male progeny from these matings would inherit the nuclease, which would inactivate the intact female flight gene inherited from the mother.All female progeny from these matings would inherit one disrupted allele of the female-specific flight gene and therefore be unable to fly, blood feed, or mate to produce future offspring.https://doi.org/10.1371/journal.pntd.0008971.g006 Table and Fig 3F and 3G) as well as Aeflightin (S5 Table and Fig 3H).All AeAct-4 Δ10/Δ10