Testing non-autonomous antimalarial gene drive effectors using self-eliminating drivers in the African mosquito vector Anopheles gambiae.

Gene drives for mosquito population modification are novel tools for malaria control. Strategies to safely test antimalarial effectors in the field are required. Here, we modified the Anopheles gambiae zpg locus to host a CRISPR/Cas9 integral gene drive allele (zpgD) and characterized its behaviour and resistance profile. We found that zpgD dominantly sterilizes females but can induce efficient drive at other loci when it itself encounters resistance. We combined zpgD with multiple previously characterized non-autonomous payload drives and found that, as zpgD self-eliminates, it leads to conversion of mosquito cage populations at these loci. Our results demonstrate how self-eliminating drivers could allow safe testing of non-autonomous effector-traits by local population modification. They also suggest that after engendering resistance, gene drives intended for population suppression could nevertheless serve to propagate subsequently released non-autonomous payload genes, allowing modification of vector populations initially targeted for suppression.

Introduction

The encouraging gains to curb malaria made since the beginning of the millennium have levelled off in recent years, and novel tools for the eradication of malaria in high-endemic areas are urgently needed [1]. Gene drive systems based on the genetic modification of vector populations are a promising set of tools conceived to either eliminate mosquito populations or reduce their ability to transmit the Plasmodium parasite to humans. Homing-based drives, first proposed in the 2000s [2] and now based on CRISPR [3,4], have demonstrable potential to reduce the size of mosquito populations (population suppression) or to render them ineffective as vectors of disease (population replacement or modification).

Population modification requires the expression of molecules that can efficiently disrupt Plasmodium development within the mosquito vector. A number of endogenous and exogenous effector molecules have been expressed in transgenic mosquitoes to interfere with parasite development, including the antimicrobial peptide (AMP) Cecropin A [5], fibrinogen domain-containing immunolectin 9 (FBN9) [6], thioester-containing protein 1 (TEP1) [7], the NF-kappaB transcription factor Rel2 [8], as well as the protein kinase Akt [9,10] and the Phosphatase and Tensin homolog (PTEN) [11]. Mosquitoes expressing exogenous synthetic peptides Vida3 [12] and SM1, engineered as a quadruplet [13], the bee venom phospholipase A2 (PLA2) [14,15], single-chain antibodies (scFvs, m1C3, m4B7, m2A10) targeting parasite proteins [16,17], and more recently, microRNA sponges [18] have also shown promise at reducing Plasmodium transmission. Progress has also been made optimising gene drive systems themselves, with a number of key conceptual or experimental advances reducing the formation of target site resistance [19-22] and limiting drives, spatiotemporally [23-27].

Although gene drives hold great promise, their ability to alter wild populations could have undesirable consequences of an ecological, economic, or even medical nature. Because of this, approaches are required to safely test such systems, particularly novel antimalarial effector molecules, at an ever-larger scale, moving in defined steps from the lab to the field. Lab containment has been discussed previously [28], and some projects are now considering field containment, i.e., geographical isolation [29,30]. Regarding the design of gene drive systems, different self-limiting approaches have been proposed (e.g. [31]). Modular or split gene drives–where a non-autonomous gRNA-only effector transgene cannot drive in the absence of an unlinked Cas9-expressing transgene–have emerged as a promising approach to mitigate risk [23-26,32-40]. In such systems, the gradual loss of a non-driving Cas9 transgene would inherently limit the spread of dependent effector transgenes over time. However, such Cas9 elements may persist in populations, albeit at a low frequency. An additional safeguarding step could use a Cas9 source that is actively removed from the population. When gene drives were conceived, it was suggested that costly drives could be recalled by introducing resistant alleles to a population, which would then be strongly selected at the expense of the drive [2]. The impact of resistant alleles on gene drive has since been shown experimentally [19,41] and widely discussed [42]. However, this does not apply to non-driving Cas9 transgenes which cannot trigger their own replacement by resistant alleles (although note that if costly, they should still disappear by natural selection). An autonomously driving, costly Cas9-source, with pre-existing resistant alleles, could present an alternative approach to actively remove the Cas9 source from a population. Such a system would leave a smaller ecological scar, spreading only those genetic modifications necessary for control through a population. It may also reduce or sidestep strict requirements for geographical isolation.

We have developed a gene drive framework that allows safeguarding and possibly reduces the time between lab and field testing [37]. Termed integral gene drive (IGD), this approach integrates effector molecules within endogenous mosquito loci, hijacking their gene-expression profiles. In anticipation of regulatory hurdles, IGD reduces the complexity and size of genetic modifications required for each transgene. After safe testing of the engineered transmission blocking traits in the absence of gene drive, these transgenes can be mobilized into non-autonomous gene drives when a source of Cas9 is provided in trans. We have recently demonstrated the efficacy of a handful of such non-autonomous transgenes [43].

Gene drive integration at the zpg target locus and pre-existing target site variants.

A: Transformation construct with the Cas9 integral drive cassette designed to target the zpg locus near the start codon for N-terminal integration. The resulting zpg allele and the intronic gRNA are shown on the bottom. The predicted separate gene products following translation and ribosomal skipping (dashed arrows beneath), including Cas9 and zpg are indicated. B: Sequence of the segregating susceptible (zpg) and resistant (zpg) haplotypes with respect to the two gRNA designs used. The first 7 zpg codons are indicated. PAM; protospacer adjacent motif. Nucleotides highlighted in grey are SNPs. C: Molecular analysis of the two naturally occurring zpg alleles present in the G3 lab colony. Sanger sequencing traces of PCR products from each background and in-vitro cleavage of PCR products by gRNA1 and gRNA2. PCR products from zpg and zpg abbreviated to S and R, respectively. D: Allele frequencies of the 4 SNPs constituting the R allele in both wild (asterisk) and lab populations of mosquitoes. Allele frequencies were estimated from Illumina whole genome sequencing of 24 G3 individuals, Sanger sequencing at the zpg locus of 24 Ngousso individuals, as well as Illumina WGS from the Ag1000G project.

The benefits of multiplex genome engineering for rapidly making complex or combinatorial genetic modifications has been widely reported [44-46]. Given the vast array of antimalarial molecules and genetic loci at which to express them, the multiplexing of effectors is an attractive approach for population modification strategies [47]. Multiplexing could have considerable benefits, from increasing throughput during initial testing (where population trials are a considerable bottleneck), to expanding the repertoire of antimalarial effectors spreading through a population in an approach analogous to combination therapy [48]. However, whether multiple effector transgenes can be safely driven through mosquito populations simultaneously has not been tested.

As a set of steps for moving safely from the lab to the field, we previously proposed the use of a non-driving Cas9 transgene as a first step, allowing a limited form of gene drive (self-limiting), before moving to an autonomously driving Cas9 transgene and full gene drive (self-sustained) [37]. Here, we sought to design and test an IGD at the zpg locus capable of providing the antimalarial transgenes with a source of Cas9. We show that an autonomously driving Cas9 at zpg, with large costs in a susceptible genetic background, can also be effectively used as a self-limiting drive when resistant alleles are present. By combining this transgene, named zpg, with non-autonomous effectors and multiplexing the individuals in four distinct caged populations, we were able to study the dynamics and mutation profiles of multiple effector transgenes simultaneously. We show that zpg can effectively drive multiple payloads through a population but is selectively removed over time. This approach has the potential to leave mosquito populations with minimal genetic modifications, limiting changes to those necessary for improving global health, through a combined minimal design of the effector transgenes and active removal of Cas9.

Results

Direct integration of gene drive modules at the zpg locus

The promoter region of the zero population growth gene (zpg) has been used successfully in conventional gene drives, providing effective Cas9 expression in A. gambiae germline [21,22]. Therefore, we chose zpg as a target locus for the generation of an IGD [37,43]. We designed gene drive constructs for integration at the start of the zpg coding sequence (CDS; ). These constructs contained a gRNA within an artificial intron introduced into the Cas9 CDS and a 3xP3::mScarlet fluorescent marker cassette, located on the antisense strand and designed for sense-strand read-through translation linked via a 2A signal to both the Cas9 and the endogenous zpg CDS. Such an approach has not been previously tested and is an alternative to marker excision, a strategy we have previously used to minimize the effects of modifications on the targeted host gene. Upon integration, the transgene would therefore co-opt the zpg regulatory regions and express Cas9 in the male and female germline, which would result in two possible outcomes: read-though translation that generates sufficient zpg gene product, while allowing the construct to behave as a neutral IGD or, in case of a significant reduction or loss of zpg host gene function, a construct that impairs or disrupts mosquito fertility, since zpg function is important for gonad development in both males and females [49]. In the latter case, if homing occurred in the germline, it could cause a dominant loss of zpg expression already in hemizygous individuals.

An analysis of targetable sites near the zpg start codon yielded 2 possible gRNA designs for Cas9 () with partially overlapping (9bp) target sites and cut sites located either within (gRNA 1) or 9 base pairs downstream of the start codon (gRNA 2). An analysis of genetic variation at the target sites revealed an alternative zpg allele with a number of single nucleotide polymorphisms that would be expected to impair Cas9 activity. Using in-vitro cleavage experiments, we confirmed the activity of both gRNAs against the wild-type susceptible target allele (zpg) and that the presence of the resistant allele (zpg) abolishes Cas9 activity in-vitro (). An analysis of population genetic data of lab strains in conjunction with wild-caught mosquitoes of the AG1000 dataset [50] suggested that the alternative R haplotype is present at varying frequencies in these populations (). For example, the R allele is absent in Angola but predominates in Kenya. In the G3 lab strain we detected the R allele at low levels (<10%). For experimental evaluation we established 2 transgenic strains utilising the cassette described above, one for each gRNA (zpg and zpg), by CRISPR-mediated knock in using a plasmid with ~2kb homology arms flanking the cassette ().

Transmission of zpg and a payload in susceptible and resistant genetic backgrounds

We first crossed, in individual mating pairs, either male or female zpg hemizygotes carrying gRNA1 or gRNA2 with wild-type G3 mosquitoes and scored the frequency of transgenic progeny via the presence of the mScarlet (RFP) marker. To determine their zpgR or zpgS status, hemizygous parents were then genotyped by genomic PCR and Sanger sequencing. For both constructs, we found Mendelian inheritance of the zpg allele when it was paired with the zpg allele irrespective of the sex of the transgenic parent (, purple bars). This corroborates the in-vitro experiments and suggests that the zpg allele is refractory to both gRNAs in vivo. When, in transgenic males, the zpg allele was paired with the susceptible zpg allele, we observed super-Mendelian inheritance of both constructs with transmission rates of 96% for gRNA1 and 76% for gRNA2, respectively (, orange bars). In contrast, no progeny were obtained from the reciprocal crosses, i.e., of transgenic zpg/zpg females carrying either gRNA to wild-type males (). We concluded from this set of results that zpg alleles produce functional Cas9 and gRNA and are able to initiate homing efficiently if a targetable site is present on the homologous chromosome. We also concluded that homing was able to induce homozygosity of zpg in the germline and leads to a loss of fertility in females, but not in males. We employed zpg for all subsequent experiments, hereafter simply referred to as zpg.

zpg produces functional Cas9 and homes in a susceptible genetic background, but zpg/zpg females do not produce offspring.

A: Proportion of progeny inheriting the zpg transgene from single crosses between hemizygous and wild-type mosquitoes. Left: Hemizygous males expressing gRNA1 or gRNA 2 crossed to wild-type females. Right: Hemizygous females expressing gRNA1 or gRNA 2 crossed to wild-type males. The transgenic parent from each cross was genotyped by Sanger sequencing to determine the genetic background of the transgene in that individual–those with susceptible alleles (zpg) are shown in orange, whereas those with resistant allele are in purple (zpg). Males expressing both gRNA1 and gRNA2 in a susceptible background sired significantly more transgenic progeny than 50% (gRNA1: t = 51.861, p<0.0001; gRNA2: t = 9.749, p<0.0001; one sample t-tests). In crosses with hemizygous females, none of the females producing viable progeny had a zpg genetic background. B: Female zpg individuals produce active Cas9 that allows non-autonomous homing of un-linked construct in a zpg background. The progeny of a cross between wild-type males and females hemizygous for zpg and a non-autonomously homing transgene at the cp locus (ScoGFP-CP) were screened for the inheritance of each transgene. Females sired significantly more ScoGFP-CP progeny than 50% (t = 65.913, p<0.0001; one sample t-test).

To corroborate the conclusion that hemizygote zpg/zpg female infertility is the result of homing in the germline, and to understand how zpg would interact with other gene drive elements, we generated transhemizygous females carrying zpg in combination with a zpg allele, as well as the previously characterized non-autonomous payload drive ScoGFP-CP, which harbours a gRNA targeting the CP wild-type allele but lacks Cas9 [43]. This non-autonomous drive element expresses the antimicrobial peptide Scorpine [51] from within the CP host gene which is strongly induced upon bloodfeeding. When crossed to wild-type males in single mating pairs we found that, as observed previously, zpg was transmitted in Mendelian fashion to the progeny. In contrast, ScoGFP-CP was transmitted to nearly all progeny (99%; ). This shows that females carrying a zpg allele are able to produce functional Cas9, further supporting our conclusions that zpg/zpg female infertility is a result of homing in the female germline, and that the inability of zpg to home in a zpg background is not because of a lack of functional Cas9 expression. It also shows that zpg can provide Cas9 in trans for efficient non-autonomous homing at other loci.

Fertility and gonad development of zpg transgenics

To better understand the effects of zpg integration at the zpg locus, we crossed hemizygous males to hemizygous females, obtaining a pool of progeny with different allelic combinations. Since only zpg/zpg hemizygous females are fertile (), no progeny in this pool came from zpg/zpg mothers. First, using primers annealing to genomic regions external to the homology arms of the template plasmid, we performed long-range PCR to genotype individuals from this pool (, primer set 1 and gel image). In this experiment, both hemizygous and homozygous individuals were identified in both sexes, showing that all genotypes are viable.

To better understand the effects of different allelic combinations at the zpg locus on male and female fertility we measured life history and physiological traits in a set of individuals from this pool. In single pairs, we crossed individuals to the wild-type and recorded: 1) whether the copulation was successful (based on whether or not the female laid viable eggs, and if they didn’t, the status of their dissected spermatheca), and 2) the number of offspring that were sired. 46 male crosses and 46 female crosses were analysed in total. We used PCRs to genotype individuals, and sequenced zpg alleles of those that were hemizygous (, primer sets 2 and 3; see methods). The following were successfully genotyped: hemizygous females n = 24; hemizygous males n = 33; homozygous females n = 17; homozygous males n = 10. Of the 33 crosses with zpg/zpg fathers, 12 led to successful insemination of females (36%), whereas of the 10 crosses with zpg/zpg fathers, only 2 were successful (20%). RNA interference (RNAi)-mediated knock-down of zpg has been previously reported to lead to completely sperm-less males [52]. Whilst we found no significant difference between the insemination rate of hemizygous and homozygous zpg males (X2 = 0.936, p>0.05), the low insemination rates, particularly of zpg/zpg males, suggest zpg may display some phenotypic manifestation of reduced zpg expression. Next, we analysed the number of offspring sired by each genotype. Whilst hemizygous (zpg/zpg) females produced an average of 79 ± 26 offspring (mean ± standard deviation)–a similar number to hemizygous males (74 ± 34)—from those homozygous (zpg/zpg) females that mated and laid eggs, none generated offspring (). This was not the case in homozygous males where viable progeny were produced (26 ± 18, mean ± standard deviation). Due to their unequal sample sizes, a Welch’s independent two-sample t-test was used to compare the number of offspring sired by the two male genotypes (Welch’s t(df = 2) = 3.035, p = 0.070), although non-significance is probably due to the low number of homozygous males. Whilst there may be some impact on male fertility, viable progeny were still sired by both genotypes in males, suggesting that zpg maintains some expression of functional zpg gene product in males. We examined gonad morphology of male and non-blood-fed female mosquitoes with different genotype combinations. Pronounced atrophy of the ovaries was observed in both zpg/zpg and zpg/zpg females (Fig 3C). This was also quantified using measurements along the major axis (anterior-posterior) of each ovary (; t(df = 24) = 8.342, p<0.0001, and t(df = 22) = 7.988, p<0.0001, respectively; independent two-sample t-tests comparing each genotype to wild-type). In homozygous zpg males, testes physiology and contents appeared slightly abnormal with the area of testis marginally smaller than wild-type males (t(df = 3) = 3.053, p = 0.055; independent two-sample t-test). The other genotypes however, appeared to have testes of normal morphology, and all genotypes appeared to have testes of similar lengths (; df = 2 in all cases). Fig 3E summarizes our observations regarding the effects on fertility of various allele combinations at the zpg locus.

Fig 3

zpg homozygotes are viable, but germline homozygosity results in sterility.

A: Molecular PCR assay to differentiate the zpg genotypes of adult mosquitoes. Primers in set 1 (purple) bind outside the original homology arms. These were used to confirm that zpg had integrated at the endogenous gene, and to identify adult wt, heterozygous and homozygous individuals as indicated in the exemplary gel below. Primer sets 2 (brown) and 3 (yellow) were additionally used for large-scale screening. B: The offspring of a cross between hemizygous zpg males and females were themselves crossed to wild-type males or females in cups, before being genotyped. The number of larval progeny from these crosses for which mating was confirmed is shown. 46 males and 46 females were analysed in total. Of those, the following were successfully genotyped: hemizygous females n = 24; hemizygous males n = 33; homozygous females n = 17; homozygous males n = 10. C: Example images of dissected wild-type and homozygote male testes (left) and non-blood-fed female ovaries (right). For males, an overall image of the entire reproductive tract is shown (scale bar 200μm) with a cutaway of a single testis (scale bar 100μm). For females, entire ovaries and reproductive tracts are shown (scale bars 400μm). D: Male and female mosquitoes of the indicated zpg genotypes were dissected and testis area and ovary length determined. Male testis area is shown in yellow and female ovary length in blue. Comparing ovary lengths to wild-type (n = 15) using independent two-sample t-tests, both homozygote (t = 7.988, p<0.0001, n = 9) and zpg/zpg (t = 8.342, p<0.0001, n = 11) females differed significantly, whereas zpg/zpg females did not (t = 0.515, p = 0.610, n = 16). Comparisons of male testis area to wild-type (n = 2) showed that only homozygotes approached significance (t = 3.053, p = 0.055, n = 3), and not zpg/zpg (t = -0.537, p = 0.645, n = 2) or zpg/zpg (t = 0.242, p = 0.831, n = 2). E: Summary of the observed fertility effects of zpg allele combinations.

Population cage dynamics of zpg in combination with payload drives

We observed that, in females, the non-autonomous drive element ScoGFP-CP could be transmitted to progeny at high rates even when zpg, which provides Cas9, encountered target site resistance at the zpg locus (. This outcome suggests that the presence of resistance alleles would have different effects on zpg and on any non-autonomous gene drives it could mobilize i.e. resistance alleles would aid the rapid replacement of the Cas9 driver locus, whilst also allowing more gene drive to occur at the effector loci. Self-eliminating drivers could thus enable local population modification without leaving behind any driver element. We examined this effect in multiplexed population cage experiments of A. gambiae where transhemizygous individuals carrying diverse effectors were released simultaneously in each cage. To model the expected dynamics of each transgene over consecutive generations, we employed an agent-based model () where we considered populations of 600 individuals with discrete generations. We investigated the interaction of zpg (starting allele frequency 15%) and 3 previously characterized non-autonomous and markerless payload drives [43]: Sco-CP, Aper1-Sco and ScoG-AP2 (starting allele frequencies of 5%, respectively). The model considers the observed fertility effects and allows for the inclusion of pre-existing target site resistance, as well as random de-novo generation of target site resistance at each of the 4 loci. It also simulates random sampling from each cage population with nonoverlapping generations.

First, we analysed a scenario where populations were seeded with the allele frequencies described above, but every alternative allele at the zpg locus was susceptible (zpg). We found that, despite active gene drive in males, no significant net change to the frequency of the payload drives occurs over time with zpg declining in frequency (). This most likely reflects the transgenes being passed on by nearly all males, but nearly no females, who when hemizygous for an effector are also most likely zpg/zpg and therefore sterile. Next, we modelled the impact of introducing pre-existing target site resistance (5% starting allele frequency of zpg). While rapid self-elimination of zpg was still predicted, we found that the introduction of resistant alleles would induce gene drive and increase effector frequency at the 3 payload loci (). This was due to active drive of the payloads now also occurring in zpgD/zpgR females, which unlike zpgD/zpgS females, were completely fertile.

We sought to experimentally validate these predictions in 4 independent cage experiments. To facilitate our experiments, we established two G3 wild-type populations which contained only the zpg or the zpg allele. Each cage was initiated using 120 and 300 zpg/zpg males and females, respectively; 60 zpg/zpg, Sco-CP/+ males; 60 zpg/zpg, ScoG-AP2/+ males; as well as 60 zpg/zpg, Aper1-Sco/+ males (see Methods). Each generation we sampled and genotyped 47 random individuals by multiplex PCR. This was done as the effector constructs used in these cage trials did not carry phenotypic markers [43] and because visible markers rapidly lose their utility when many constructs are in play. We observed a decline in the egg output in all 4 cage populations, peaking at <45% in generation 2 (). With the decline of zpg to an average transgene frequency of <20% in generation 6, egg production recovered in all populations. Across the 4 cage populations the transgene frequency of the 3 non-autonomous effector traits increased over the course of the experiments, in line with the predictions of the model ().

zpg self-eliminates but drives the spread of three non-autonomous payload genes across four independent caged populations.

A: Illustration of zpg-dependent homing at the three non-autonomous integral gene drive loci. The germline-expressed Cas9 forms a complex with U6-driven gRNAs expressed by and specific to each effector construct. B: Transgene frequency of each construct estimated by either fluorescence (zpg) or multiplex PCR. Solid lines and shaded areas indicate the means ± 68% bootstrap CI respectively for zpgD (blue), ScoG-AP2 (purple), Sco-CP (green) and Aper1-Sco (yellow). 47 individuals and one negative control were genotyped per cage and per generation. Cages contained ~600 individuals in total.

Analysis of target site alleles in the caged populations

Following cleavage by Cas9 in the germline, DNA double-strand breaks (DSB) can be resolved by the DNA repair machinery in several ways. Homology directed repair (HDR) using the drive allele as a repair template leads to conversion of a wild-type allele to a drive allele (i.e. homing). Due to the design of gene drives, this outcome eliminates the gRNA target site making the new drive allele resistant to future cutting. However, DSB can also be repaired via DNA end-joining pathways such as non-homologous end joining (NHEJ) or microhomology-mediated end joining (MMEJ) which does not lead to drive conversion. The resulting small deletions or insertions generally destroy the gRNA target site and abolish recognition by Cas9. Such cleavage resistant alleles, which can also occur naturally as standing genetic variation of a population (as is the case for zpgR), present a barrier to gene drive. These mutations may also arise outside of the germline in developing embryos as a result of maternally deposited Cas9 [19,20,22,41]. With respect to target gene function, the impact of such mutations can range from being detrimental (termed R2 alleles) to being neutral, for example when they result in synonymous substitutions or small tolerable deletions that maintain the reading frame (termed R1 alleles). R1 alleles that present a lower relative cost than a circulating drive will therefore undergo strong selection [19,41,42,53,54]. To better understand the observed cage dynamics and assess the level to which resistance may evolve at each locus, we performed amplicon sequencing of the target sites of all 4 host genes at generations 1 and 6 in each of the 4 replicate cage populations. This was performed on a pool of 100 mosquitoes for each sample and PCR conditions were chosen to amplify wild-type-sized alleles potentially modified by NHEJ or MMEJ but not the respective transgenes at each locus. We observed a significant increase in alleles predicted to impair or abolish gRNA recognition at 3 out of the 4 loci (). At the zpg locus >60% of all reads in generation 6 represented resistant alleles–a 1.3-fold increase from generation 1. The pre-existing R1 resistant allele we had characterized and introduced was the most predominant, increasing 3-fold from 7% at G1 to 21% of all reads at G6 (). At both the aper1 and ap2 loci, resistant alleles made up 30% and 19% of all reads at generation 6 (2.9- and 1.7-fold increases), respectively. In both cases single, predominant R1 alleles also present at generation 1 and carrying mutations relatively distant from the predicted cleavage site made up the majority of the resistant alleles (Fig 5B). Interestingly, there was no significant increase in resistant alleles at the cp locus. The top 5 most frequent resistant alleles, all predicted to maintain CP expression, accounted for <1% of reads on average and overall, less than 10% of reads differed from the wild-type cp target sequence in either generation 1 or 6.

Fig 5

Analysis of target site variants in the caged populations at all 4 loci.

A: Percentage of reads with a predicted cleavage resistant allele of any type at each locus, at generations 1 (G1, brown) and 6 (G6, pink). Each data point represents a cage population. Paired t-tests were used to compare G1 and G6 at each locus (AP2: t = -4.003, p = 0.028; CP: t = -1.921, p = 0.151; Aper1: t = -7.603, p = 0.005; zpg: t = -6.540, p = 0.007) B: Shown are the reference allele (top) and the 5 most common target site resistance variants observed at each locus (differences highlighted in grey) and the percentage of reads mapping to each. Sequences on the left are shown relative to the gRNA (above; cut site shown with grey triangle and dotted line). Coding nucleotides are indicated in colour and the direction of translation is indicated (arrow). The frequency of each of these alleles in G1 and G6 samples is then shown on the right (each data point representing a cage population). The zpg allele is shown directly below the reference allele for the zpg data.

Discussion

By integrating a small number of genetic features at endogenous genomic loci and hijacking their gene-expression machinery, the amount of genetic information that needs to be introduced to a transgenic organism can be greatly reduced. Along with the excision of fluorescent markers, this could create minimal and modular gene drive elements for population modification [37,43]. Here, we integrated the CDS of Cas9, including a synthetic gRNA-containing intron, at the zpg locus with the aim to establish an autonomous integral gene drive. Our results suggest that this allowed for high rates of homing of the modified allele (termed zpg) in males and females and of non-autonomous transgenes at separate loci each expressing a respective cognate gRNA. However, homozygosity of zpg, and probably homing-induced homozygosity in the germline of hemizygous females, led to fertility defects. Note that some low level somatic “leakiness” has been reported when using the zpg promoter for germ-line expression [22], and hemizygote infertility could also result from expression of zpg in the soma. Both zpg/zpg and zpg/zpg females were sterile and lacked fully developed ovaries whereas homozygous zpg/zpg males showed low levels of fertility. Whilst the reasons for this are unclear, 2A skipper peptides have been shown to reduce the expression of subsequent CDS due to inefficient cleavage [55,56], and the zpg Cas9 module was linked with the endogenous zpg CDS by a 2A-separated translational unit representing the fluorescent transformation marker. This was designed for read-though translation, where as well as Cas9, the ribosome would translate the reverse complement and coding sequences of the marker, followed by the zpg host gene product. However, our results suggest that this novel configuration did not allow for sufficient expression of zpg from the zpg allele in females. Alternatively, other transcriptional or translational mechanisms may have led to reduced zpg gene product (e.g. regulatory interference from the Cas9 coding sequence). Read-through translation and the resulting nonsense peptide arising from antisense 3xP3::mScarlet may have also contributed the observed fitness effects. Another method for avoiding gene-regulatory problems associated with marker genes in IGDs is to remove them after transformation using Cre-loxP mediated excision [43] and such an approach could allow the development of neutral drives in the future.

Interestingly, the tendency of zpg to cause dominant sterilization in females was counteracted by zpg target site resistance which occurs naturally in lab and wild populations of A. gambiae. Thus, in addition to males where both autonomous and non-autonomous gene drive occurred efficiently, the presence of the zpg allele in conjunction with zpg allowed for the mobilization of unlinked non-autonomous drive elements in females. We hypothesized that due to its sex- and background- specific costs, zpg would thus provide a self-removing source of Cas9. Just as gene drives with minimal genetic modifications reduce the changes introduced into individual genomes, approaches reducing the modifications introduced to populations (to only those components necessary for global health, agriculture, or ecology, for example) have obvious appeal. Advances have been made in our ability to recall problematic gene drives by replacing or upgrading them with different transgenes [57,58]. Alternatively, transgenes could be replaced with natural alleles. It has been suggested that gene drive testing pathways for population modification could feature a non-driving source of Cas9 that, having spread non-autonomous payload genes through a population, is slowly lost by selection [26,27]. Although the assumption that simple Cas9-expressing transgenes will impose fitness costs seems plausible, a recent study in Aedes Aegypti has shown that, whilst one Cas9-expressing transgene did impose a small fitness cost, another did not have any measurable effect [23]. To build a system where Cas9 is removed unambiguously, an effective, self-eliminating source of Cas9 is desirable. Furthermore, the dynamics of such a transgene must be both modelled, and more importantly, tested in population trials. Recent advances have been made in Drosophila where trials showed that non-driving Cas9 transgenes are able to impose a fitness cost in the presence of their non-autonomous payloads by targeting the payloads to essential genes [24]. However, a system where the cost of the Cas9 is unlinked to payloads it is mobilizing has not been described and is not available in medically relevant target species such as A. gambiae.

Because the active homing of zpg, or its leaky somatic expression, results in female sterility, resistant alleles at the zpg target site should be rapidly selected for. Such a resistant allele (referred to as zpg) already exists in lab and wild populations. When released into a population with zpg alleles, zpg was expected to undergo strong counterselection, leaving non-transgenic resistant alleles behind. We were able to recapitulate this dynamic in a set of population cage experiments, whilst for the first time, studying the behaviour of multiple interdependent drive elements. These cage trials allowed us to show experimentally that zpg can spread non-autonomous drives whilst itself being lost from a population. By sequencing non-transgenic alleles, we also analysed the rise of resistant alleles at each of the 4 loci. We found that whilst the ScoG-AP2 transgene showed the most dramatic spread, it also had the greatest increase in target site resistance over time. On the other hand, Sco-CP showed a ~4-fold increase in frequency, yet no significant increase in resistant alleles over the course of the experiment. The reasons for this are not obvious. Carboxypeptidase proteins are thought to be highly important for mosquito blood meal digestion [59,60] and our results could indicate functional constraints on the target sequence near the start codon at this locus. Lower levels of pre-existing variation at the target site (although this too could reflect strong functional constraint), or a lower cost of transgene integration could also play a role.

Multigenerational trials using caged populations can tell us a great deal about a gene drive system, but there are fundamental aspects of field populations that cannot be recapitulated in a lab setting. Whilst our simple discrete-generation model goes some way to predicting the behaviour of zpgD and interdependent transgenes, important aspects of population genetics could not be explored or have not been explicitly investigated. For example, several features accompanying a possible zpgD release could lead to a reduction in the effective population size. These include the strong selection for resistance alleles, the imbalances in reproductive output between transgenic males and females, or between transgenic and wild-type individuals, and the reduction in census population size due to zpgD’s suppressive effect. This last point is particularly important as it also affects migration (which, in-turn, can influence the effective population size). For example, the smaller a target population is relative to a neighbouring population, the more immigrants it will receive per emigrant, with immigrants also making up a larger proportion of the population [61]. Such an effect is evident in stochastic, spatially-explicit modelling of so-called “reduce and replace” systems [62], where population size is reduced by a female-specific lethal transgene whilst an anti-pathogen payload is transmitted through surviving transgenic males [62,63]. It has been suggested that, with a greater level of population suppression, the effect of wild-type migrants on transgene dynamics is amplified, impacting the long-term frequency of the payload gene. Parallels between this approach and zpgD-based population modification would suggest that we may find similar effects in the early stages of a release where zpgD is at a high frequency and the population is being suppressed. On the other hand, this asymmetry could be beneficial for migration in the other direction i.e. reduced escape of transgenes and invasion of non-target populations (although note that others have suggested that in many cases, escape is still possible with relatively low levels of migration [64]). Lastly, any effects of, or from, migration will depend strongly on the frequency of resistant alleles in neighbouring populations [65]. Future work will need to determine the importance of these factors in influencing the dynamics of released gene drives.

In summary, by combining a self-eliminating source of Cas9 with minimal, marker-free payload genes, we were able to show the efficacy of a gene drive system that moves towards more minimal modifications of populations following a gene drive intervention. We also show that the 5’ of CP is a promising target for non-autonomous integral gene drives due to its favourable profile for the accumulation of resistant alleles. Our population trials also demonstrate the feasibility of multiplexing non-autonomous payloads. The benefits of this are potentially manifold. Firstly, although every payload transgene used here carried the same anti-parasite effector molecule (Scorpine [51]), different effectors could be used in the future in an approach akin to combination therapy. Synergistic interactions between effectors within a mosquito could then lead to an increased potency compared to one effector alone. Furthermore, this combination therapy would reduce the likelihood of resistance evolving in parasite populations. As well as expanding the arsenal of effector molecules, multiplexing could utilise different effector expression profiles, allowing users to target different stages of the parasite lifecycle in the same release [47]. An additional benefit to multiplexing effectors could be the reduced risk posed by unidentified mechanisms of escape or resistance to one transgene or its effector in a wild population. It is possible however, that when multiplexing non-autonomous drives, the level of Cas9 may become limiting [66] if multiple transgenes are present in the same individual.

The cage experiments we performed also mirror another important scenario: suppression drives that have engendered resistance and are failing. Unlike suppression drives, gene drives intended for population modification are typically not designed to carry any fitness cost [53,54]. Because they provide little relative advantage over the drive allele, selection for functional resistant alleles in such drives should be weak. This is evidenced from recent cage trial data in mosquito population modification strategies where resistance alleles had little impact on the performance of the drive [67]. On the other hand, functional resistant mutations pose a large threat to suppression drives [2,19] and have been found to undergo strong selection in cage trials where the drive is costly [41,68]. Whilst various improvements have been made to suppression drives to reduce the likelihood of resistance arising, including the use of highly constrained gRNA target sites [21] and optimisation of the level and spatiotemporal selectivity of Cas9 expression [22,69], the generation and accumulation of resistant alleles still poses a risk to the long-term success of suppression programmes [70,71]. The fitness cost imposed by zpg is comparable to (due to its dominance) that of a suppressive gene drive. Accordingly, in the egg output of our cage trials we found that at higher allele frequencies, zpg had a suppressive effect on the populations (). Furthermore, our cage trials were seeded with standing variation that included functional resistant alleles (zpg) like those that could be generated by a suppression drive. The dynamics of zpg in our cage trials therefore also mimics a failing suppression drive. Our results suggest that if such dynamics were found after suppression programmes have begun, non-autonomous drives could be released to piggyback on residual Cas9 in the population. As malaria resurgence in a human population with reduced levels of immunity is a potential negative outcome of an unsuccessful suppression trial the ability to switch to a population modification strategy, making use of the existing Cas9 alleles in the population, is an attractive fallback strategy.

Methods

Target choice and construct design

In an initial attempt, we had designed a simple drive module at the zpg locus, consisting of a Cas9-2A followed by a GFP-2A. When integrated immediately before the start of a germline-expressed CDS, this construct should lead to the production of Cas9, GFP, and the endogenous target gene’s protein product in the germline. A gRNA targeting the wild-type locus also needed to be included to allow the module to home itself. A synthetic intron was therefore inserted in the Cas9 sequence, containing an Anopheles gambiae U6-driven gRNA. This meant that the gRNA could be transmitted along with the rest of the knock-in during HDR, without disrupting translation of the Cas9-2A-GFP-2A knock-in. The sequence surrounding zpg, along with its gene annotation [72], was obtained from Ensembl directly through Benchling (AgamP4 genome assembly; accessed 2019). Again, using Benchling, gRNAs targetting the region immediately surrounding the start codon were chosen to determine the knock-in location. Only gRNAs whose target sequence would be disrupted by a clean knock-in directly at the start codon were chosen. However, despite repeated rounds of injections none of the mosaic G0 mosquitoes gave rise to evident transgenic lines. Attributing this to low fluorescent signal from the 2A-GFP (mosquito zpg has not been characterised using GFP; furthermore, 2A skipper peptides may also reduce the expression of subsequent CDS due to inefficient cleavage [55,56]), we used the GFP as a vessel for a new bright, dominant, fluorescent marker, codon-optimised for A. gambiae (3xP3::AgScarlet). This new marker cassette contained intron donor and acceptor sites so it could be spliced from the GFP coding sequence, but was also re-coded to allow translational readthrough (and generation of a nonsense amino acid chain on the end of the GFP in the event that splicing did not occur–see below). Using these new constructs, we obtained the two transgenic mosquito lines that were investigated here.

Generation of constructs

A consensus sequence of canonical U2-dependent introns [73] was used to guide the design of a synthetic gRNA-expressing intron within Cas9. We used a human codon-optimised version of Cas9 [45], and the U6::gRNA cassette (containing a spacer-cloning site) was taken from p165 [4]. Splicing donor/acceptor signals were introduced by PCR. Using Gibson assembly, the intron was integrated at an appropriate site within the Cas9 CDS, the whole gene flanked with BbsI sites and integrated within a backbone from pBac[3xP3-DsRed]β2-eGFP(F2A)I-PpoI [74] producing the plasmid, pCas9-INT. After digestion with BbsI, this Cas9{U6::gRNA} fragment was fused at its 3’ end with a T2A sequence [56] (codon-optimised for A. gambiae and generated by annealing oligos), followed by a GFP-F2A sequence amplified from pBac[3xP3-DsRed]β2-eGFP(F2A)I-PpoI. This whole fragment was again flanked at each end by BbsI sites and assembled into a backbone amplified from pAmpR_SDM (an unpublished plasmid where a BsaI site has been removed from the AmpR cassette with a silent mutation) to give the plasmid, pUDCas9. A 3xP3::mScarlet fluorescent marker [75] codon-optimised for A. gambiae, flanked by splicing donor/acceptor signals, and recoded to allow transcriptional and translational readthrough in the antisense direction was synthesised by ATUM. This fragment was then integrated in the antisense direction into the GFP of pUDCas9 after BsrGI digestion, yielding pUDCas9_v2X. The two gRNAs targeting zpg were synthesised as oligos (see primer table) and cloned in by BsaI-based Golden Gate Assembly into the U6::gRNA spacer cloning site of pUDCas9_v2X, yielding pUDCas9_v2X_ZPG1 and pUDCas9_v2X_ZPG2. These plasmids were then digested with BbsI, and the Cas9{U6::gRNA}-2A-GFP{3xP3::AgScarlet}-2A fragments assembled within ~2kb homology arms to the 3’ and 5’ (amplified from A. gambiae genomic DNA), and a universal vector backbone fragment amplified from pBac[3xP3-DsRed]β2-eGFP(F2A)I-PpoI. These plasmids were called pZPGCas9Dr_v2_ZPG1, and pZPGCas9Dr_v2_ZPG2 (GenBank files for these plasmids are provided as ). All primers used for generating constructs are listed in .

Microinjection of embryos, identification and confirmation of transgenic mosquitoes

Freshly laid G3 embryos were used for microinjection, as described elsewhere [76,77]. Plasmids were purified with endotoxin-free reagents (EndoFree Plasmid Maxi Kit, Qiagen) and all template plasmids were sequenced prior to injection. Injections were performed using 300ng/μl of helper plasmid and 300ng/μl of template plasmid. Surviving G0 larvae showing mosaic RFP expression at the posterior end from the 3xP3 marker in the helper or template plasmids were crossed to wild-type mosquitoes and G1 progeny were screened for positive transformants based on the relevant fluorescent marker (germline-expressed GFP for v1 constructs, 3xP3::AgScarlet for v2). Knock-in of constructs was confirmed by PCR using primers annealing in the genome, beyond the edge of homology arms (see primer table). pZPGCas9Dr_v2_ZPG1, and pZPGCas9Dr_v2_ZPG2 gave rise to zpg and zpg, respectively.

Estimation of allele frequencies

The allele frequencies of the SNPs within the zpg and zpg alleles were estimated using population genetic information from the following sources: Illumina whole genome sequencing (WGS) of 24 G3 individuals [41]; Sanger sequencing at the zpg locus (see primer table) of 24 Ngousso individuals; WGS data from 60 Angolan and 44 Kenyan mosquitoes (Anopheles gambiae 1000 genomes project [50], phase 1 AR3 data release, accessed July 2020).

Containment of gene drive mosquitoes

Mosquitoes were housed at Imperial College London in a Level 2-compliant insectary under previously established Arthropod Containment Guidelines [78] and all work was performed under institutionally approved biosafety and GM protocols. As well as the ecological containment provided by the temperate Northern climate of the United Kingdom, six doors separated any cubicles housing gene drive mosquitoes to the external environment and two levels of security access were required to gain entry to these cubicles. These containment steps are compliant with safeguarding guidelines set out by Akbari and colleagues [28].

Mosquito husbandry

All mosquitoes were maintained at ~26°C and ~70% humidity with a 12h:12h light:dark cycle and provided with 10% Glucose ad libitum. Females were blood-fed with human blood. All sexing was performed at pupal stage. To maintain the zpg transgenic line, hemizygous males were backcrossed to G3 females every generation, ad infinitum. Isogenic zpg and zpg lines were obtained by performing a large number of single crosses between G3 males and females. Parents were then genotyped by Sanger sequencing of the gRNA target region (see primer table). An equal number of crosses whose parents were enriched for either zpg or zpg were then chosen, and single crosses performed again between F1 male and female offspring from different clutches. F2 offspring with two homozygous zpg or zpg parents were pooled, giving rise to the isogenic zpg and zpg wild-type lines, respectively. These lines were used only for seeding cage trials.

Assessing zpgD in different genetic backgrounds

Due to the presence of the zpg allele in the G3 lab colony, our initial maintenance of the zpg line (see above) meant that hemizygotes could carry either a zpg or zpg allele at their wild-type chromosome. zpg transmission was therefore assessed using single crosses of hemizygotes to G3 individuals, followed by genotyping of the zpg locus at the wild-type chromosome. For males, crosses were performed in cups by giving a single male three G3 females. These females were then blood-fed and allowed to lay in a fresh cup containing ~25mL water and lined with filter paper. Female hemizygotes were allowed to mate with G3 males and blood-fed en-masse, before being distributed to single cups and allowed to lay, individually. The larvae (minimum of 20, mean ± std of 50.9 ± 17.9) of each cross were then screened for the presence of the zpg allele based on red fluorescence using an Olympus MVX10 stereo microscope. Parents from each single cross were transferred to PCR tubes and the zpg allele at their wild-type chromosome was amplified using Phire Tissue Direct PCR Mastermix (Thermo Scientific; see primer table). Amplicons were then sent for Sanger sequencing. Sequencing files were analysed in Benchling.

To assay transmission of a non-autonomous gRNA-only construct in the presence of zpg, we generated double hemizygote zpg, ScoGFP-CP [43] individuals by crossing hemizygote zpg males to homozygote ScoGFP-CP females and selecting any doubly fluorescent mScarlet, GFP offspring. Double hemizygote males or females were then crossed to G3 mosquitoes of the opposite sex as above, and females were allowed to lay, individually. Larvae from each cross (minimum of 21, mean ± std of 45.9 ± 8.1) were then screened for red and green fluorescence using an Olympus MVX10 stereo microscope. The data for plotting has been provided in file S1 Data.

The in-vitro cleavage assay was performed using PCR products at the zpg locus containing either the zpg or zpg allele, obtained during homing assays (above, see primer table). gRNAs were synthesised in-vitro using the TranscriptAid T7 High Yield Transcription Kit (Thermo Scientific) and a template. The template was generated by PCR using the following oligos: gRNA-InViTr_ZPG1 or gRNA-InViTr_ZPG2 (0.02μM); gRNA-InViTr_scaffold (0.02μM); gRNA-InViTr_fwd (2μM); gRNA-InViTr_rev (2μM). Templates were then purified using the QIAquick PCR Purification Kit (Qiagen). In-vitro transcription reactions were incubated for 8 hours at 37°C, treated with DNase I (NEB), and purified using the RNeasy MinElute Cleanup Kit (Qiagen). Cleavage reactions were performed using 3nM of zpg PCR fragment, 45nM of sgRNA and 45nM of S. pyogenes Cas9 nuclease (NEB). After heat-inactivation at 65°C, the entire volume of cleavage reaction was loaded and resolved using gel electrophoresis.

Phenotypic characterisation of zpg genotypes

To compare the fertility of homozygous zpg males and females to hemizygotes, we first obtained hemizygous and homozygous individuals by crossing hemizygous males and females en-masse. Note that because zpg/zpg females are not fertile (see below), all hemizygous offspring carried a zpg allele in their genetic background. The F1 of this cross were then separated into single cups and each given one G3 mate of the opposite sex. After 5 days, females were blood-fed and transferred to fresh cups to lay. After three days, eggs were counted for each female, and females were collected. Any females that did not lay viable eggs were dissected and their spermathecae examined. Unmated females were excluded from further phenotypic analysis. The number of larvae hatching from the eggs was counted over a few days. 46 transgenic male and 46 transgenic female crosses were analysed in total, of which 32 male and 10 female crosses either were unsuccessful or could not be genotyped.

To genotype hemizygotes vs homozygotes, genomic DNA was extracted from individual transgenic mosquitoes using DNeasy Blood & Tissue kits (Qiagen). PCR was then performed on each sample using two primer sets: one spanning the 5’ junction of the knock-in, with one internal and one external primer; and one spanning either side of the integration site (see primer table). The first primer set produces a band in the presence of a transgene, the second produces a band in the absence of one. Theoretically, this second primer set also produces a very large band in the presence of a transgene, however the predicted amplicon is too large (>7.5kb) given the polymerase and extension times used for genotyping, and this band was not obtained. Using a specialist polymerase for PCR with long amplicons (LongAmp Taq, NEB), a handful of DNA samples were randomly chosen and confirmed using a third primer set spanning the whole transgene with primers annealing beyond the edge of the homology arms used for the initial knock-in (Fig 3; see primer table).

To quantify the size of reproductive organs from mosquitoes of different genotypes, hemizygous zpg individuals were intercrossed and virgin male and female F1 offspring, along with virgin male and female G3 mosquitoes, were dissected straight into PBS. Testes and ovaries were then imaged on an inverted EVOS FL Imaging System without coverslips and carcasses were genotyped by Sanger sequencing (see above). The Length of testes and ovaries were estimated in ImageJ by tracing a segmented line through the centre of each organ. Testes area was estimated in ImageJ by approximating each testis with an ellipse, then taking the squareroot. A minimum of 9 mosquitoes were dissected of each genotype (mean ± std of 15.1 ± 3.8). The data for plotting has been provided in file .

Cage trial assay

We carried out trials in four independent cage populations. The double hemizygotes used to seed the cages were obtained by crossing hemizygous zpg males with homozygous females of each effector line (ScoG-AP2, Sco-CP or Aper1-Sco). Individual females from each cross were then allowed to lay, before having their wild-type zpg allele genotyped by sanger sequencing. The majority of ScoG-AP2 and Sco-CP homozygotes were zpg/zpg, and only the progeny of zpg/zpg females were kept. Aper1-Sco homozygotes were uniformly zpg/zpg. Most likely this is due to the Aper1-Sco line being founded in a zpg background and the two loci are in proximity (2L, 38.6Mb and 2L,28.6Mb, respectively). The progeny of each cross was maintained separately until pupae, then screened for the zpg mScarlet marker and sexed. The resulting double hemizygote males were then used to initiate cage trials in 24.5cm3 cages. Each cage was initiated using: 60 zpg/zpg, ScoG-AP2 double heterozygote males; 60 zpg/zpg, Sco-CP double heterozygote males; 60 zpg/zpg, Aper1-Sco double heterozygote males; 120 wild-type zpg/zpg males and 300 wild-type zpg/zpg females. Mosquitoes were left to mate for five days before being fed on human blood. Two days later, mosquitoes were given egg bowls consisting of conical filter paper (125mm diam., Whatman) submerged in 300ml water for oviposition. Eggs were spread evenly across the filter paper with water from a Pasteur pipette and photographed. Images were then used to estimate egg numbers using the MosceggApp egg-counting software [79] with egg-detection sensitivity set to minimum and non-egg detection set to approximately 66%. 600 larvae emerging from these eggs were then randomly chosen to seed the next generation and maintained at 150 larvae per pan until pupation. At each generation, once adults had been allowed to lay, 47 were randomly collected from each cage and genotyped, along with one wild-type per-cage as a negative control, using multiplex-PCR (below). A further 100 were collected for pooled DNA extraction and amplicon sequencing (below).

We required a molecular method for genotyping. A multiplex PCR approach was optimised using allele-specific primers that amplify fragments for the transgenic, but not wild-type alleles, at each locus. Whole mosquitoes were lysed in separate wells of 96-well plates and immediately used for PCR following the Phire Tissue Direct PCR Master Mix (Thermo) Dilution & Storage protocol. As well as primer sets at the four transgenic loci, we also included a primer pair at the S7 ribosomal subunit gene as an internal positive control for each PCR. The same reverse primer within Scorpine was used in all effector transgene primer sets. Primers were designed to produce a discernible banding pattern that would allow discrimination between each locus, with the S7 control product giving the largest band. Any individuals that didn’t produce this larger band were ignored. Various primers and combinations were tested, and the annealing temperatures of the best primer set was optimised. For each transgenic locus, the specificity of its final primer pair was confirmed by performing PCR with transgenic DNA samples from the other three loci. The combined primers for all transgenes also produced no bands when using wild-type DNA as a template. Although a primer pair was included in the set for benchmarking, zpg was genotyped using its 3xP3-driven fluorescent reporter. See file for individual cage data.

Amplicon sequencing

Primers were designed to amplify 300-500bp regions surrounding the genomic gRNA target site at each locus. Primers were given a 5mer barcode at the 5’ end specific to the generation (forward primer) and replicate (reverse primer), allowing amplicons from different DNA samples to be multiplexed (see primer table). Genomic DNA was extracted from pools of 100 mosquitoes per replicate at two generations (G1 and G6). For PCR, KAPA HiFi HotStart polymerase (Roche) was used with 25 cycles. Amplicons were purified using Qiaquick PCR purification columns (Qiagen) and quantified using Qubit (ThermoFisher). Thirty-two samples were multiplexed in total (2 generations, 4 replicates, 4 loci). 5ng of DNA from each sample was pooled (160ng total) and sequenced on one MiSeq lane with 2x300bp reads (Eurofins Genomics).

Je [80] was used to initially de-multiplex reads using the 5’ barcode, detailing the generation of the sample, and the 3’ barcode, detailing the replicate. CRIPSPResso2pooled [81] was then used to merge read pairs and map them to the four different amplicons (default parameters were used, apart from—max_paired_end_reads_overlap which was set to 252 – 10bp more than the predicted overlap at the smallest amplicon). Any read pairs unassigned a sample by Je, were then demultiplexed using fastx-toolkit [82] from the single-end flash-merged reads, where sequencing errors present towards the ends of separate reads have been reduced. These were merged with their corresponding fastq file using bash. The thirty-two resulting fastq files were then re-mapped and analysed using CRISPResso2bulk (with the same parameters used for CRISPResso2pooled; see above). Mapping statistics are provided in . Reads were deposited in the European Nucleotide Archive (ENA) under accession number PRJEB50687.

Using the frequency tables of alleles around the sgRNA produced by CRISPResso2, a custom Python script was used to select only those alleles with mutations in the region bound by the gRNA spacer or the PAM (i.e. those that would potentially cause resistance to gene drive). For the first analysis in Fig 5A, the frequency of all these alleles were summed for each sample, then this sum calculated as a percentage of the total number of reads for that sample. For the second analysis, the five most common alleles (across all samples and replicates) were selected. Whether a given allele was absent in any G1 samples was also calculated, and that allele given a frequency of 0 for that sample.

Discrete generation population model of zpg interacting with 3 non-autonomous payload drives.

A: Stochastic simulation of populations of 600 adult mosquitoes with a starting population of 300 wild-type females, 120 wild-type males (all zpg) and 60 zpg/zpg, Sco-CP/+ males; 60 zpg/zpg, ScoG-AP2/+ males and 60 zpg/zpg, Aper1-Sco/+ males. Shown are the average transgene frequencies at each locus calculated from 47 randomly sampled individuals out of 25 simulated populations. Because of this sampling strategy, which mimics our experimental one, starting transgene frequencies of zpgD (30%) and each of the payloads (10%) in individual populations may differ from the mean across populations (black line). B: The number of individuals of each indicated genotype at the 4 loci of the simulations in A of all 600 individuals in each population. Drive here refers to both autonomous (zpg) and non-autonomous (Sco-CP, Aper1-Sco, ScoG-AP2) elements. C: Simulations as in A but assuming the presence of pre-existing R1 target site resistance at the zpg locus. These were introduced as 60 zpg/zpg, Aper1-Sco/+ males resulting in 5% starting allele frequency of zpg. Shown are the average transgene frequencies at each locus calculated from 47 randomly sampled individuals out of 25 simulated populations. Because of this sampling strategy, which mimics our experimental one, starting transgene frequencies of zpgD (30%) and each of the payloads (10%) in individual populations may differ from the mean across populations (black line). D: The number of individuals of each indicated genotype at the 4 loci of the simulations in C of all 600 individuals in each population. Drive here refers to both autonomous (zpgD) and non-autonomous (Sco-CP, Aper1-Sco, ScoG-AP2) elements.

(PDF)

Click here for additional data file.

Changes in population reproductive capacity during spread of transgenes.

A: Total egg output of each caged population, at each generation shown in light grey; mean of all cages at each generation shown in black; predicted trajectory given no change in reproductive output (mean of all cages at G0) shown by dashed red line. B: The integral of each trajectory (grey circles) and their mean (grey bar) compared to the integral of the no-change trajectory (red dashed line). A one-sample t-test was used to compare the integrals of each real timeseries to that of the predicted no-change timeseries (t = -13.2, p<0.001).

(PDF)

Click here for additional data file.

List of primers used.

(XLSX)

Click here for additional data file.

Mapping statistics for the amplicon sequencing experiment.

(XLSX)

Click here for additional data file.

Raw data for the generation of Figs 2,3 & 4.

(XLSX)

Click here for additional data file.

Description of the gene drive model.

(DOCX)

Click here for additional data file.

Annotated DNA files for transformation constructs pZPGCas9Dr_v2_ZPG1 and pZPGCas9Dr_v2_ZPG2.

(ZIP)

Click here for additional data file.

30 Mar 2022

Dear Dr Windbichler,

Thank you very much for submitting your Research Article entitled 'Testing non-autonomous antimalarial gene drive effectors using self-eliminating drivers in the African mosquito vector Anopheles gambiae' to PLOS Genetics.

The manuscript was fully evaluated at the editorial level and by independent peer reviewers. The reviewers appreciated the attention to an important topic but identified some concerns that we ask you address in a revised manuscript

We therefore ask you to modify the manuscript according to the review recommendations. Your revisions should address the specific points made by each reviewer.

In addition we ask that you:

1) Provide a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript.

2) Upload a Striking Image with a corresponding caption to accompany your manuscript if one is available (either a new image or an existing one from within your manuscript). If this image is judged to be suitable, it may be featured on our website. Images should ideally be high resolution, eye-catching, single panel square images. For examples, please browse our archive. If your image is from someone other than yourself, please ensure that the artist has read and agreed to the terms and conditions of the Creative Commons Attribution License. Note: we cannot publish copyrighted images.

We hope to receive your revised manuscript within the next 30 days. If you anticipate any delay in its return, we would ask you to let us know the expected resubmission date by email to plosgenetics@plos.org.

If present, accompanying reviewer attachments should be included with this email; please notify the journal office if any appear to be missing. They will also be available for download from the link below. You can use this link to log into the system when you are ready to submit a revised version, having first consulted our Submission Checklist.

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

Please be aware that our data availability policy requires that all numerical data underlying graphs or summary statistics are included with the submission, and you will need to provide this upon resubmission if not already present. In addition, we do not permit the inclusion of phrases such as "data not shown" or "unpublished results" in manuscripts. All points should be backed up by data provided with the submission.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

PLOS has incorporated Similarity Check, powered by iThenticate, into its journal-wide submission system in order to screen submitted content for originality before publication. Each PLOS journal undertakes screening on a proportion of submitted articles. You will be contacted if needed following the screening process.

To resubmit, you will need to go to the link below and 'Revise Submission' in the 'Submissions Needing Revision' folder.

[LINK]

Please let us know if you have any questions while making these revisions.

Yours sincerely,

Jackson Champer

Associate Editor

PLOS Genetics

Gregory P. Copenhaver

Editor-in-Chief

PLOS Genetics

Dear Dr. Windbichler and Colleagues,

This manuscript has been examined by three reviewers and myself, and we all agree that it will be most likely suitable for publication in PLoS Genetics, though some revisions will be needed.

Please attempt to address all reviewer comments. In particular, pay attention to data accessibility (both for raw data and plasmid maps) and the extra (?) supplemental table.

Recommended Additional Revisions:

The authors assume that zpg is required in germline cells, so drive/wild-type individuals become sterile by germline drive conversion. More likely than not, this is correct, but it dismisses another possibility that could explain all or some of the sterility. Based on previous studies of zpg-Cas9 in suppression drives, it is known that zpg leads to some “leaky somatic” expression in somatic cells, which moderately reduces the fertility of females (by cutting the wild-type allele). It might be that this is an important factor in reducing zpg expression in important cells, so the possibility should at least be mentioned.

For the discussion about Figure 5, I think it might be helpful to take a step back and introduce for a more general audience what makes a functional resistance allele and why seeing these at high frequency is a pretty good reason to assume that they are r1 (compared to r2). It also seems a little odd to me that more possible r2s were not detected in G1 (compared to G6). Maybe there was a mechanism to quickly eliminate them.

The modeling section in the methods should be considerably expanded. Including the code is good, but the methods should still fully describe the model. If needed, it can be a supplemental methods section.

Minor Comments and Suggestions (revisions at the discretion of the authors):

I liked reviewer 2’s comment on population replacement versus population modification. I agree with the terminology myself as well.

Please take care to include line numbers or at least page numbers when submitting to journals. Some add it automatically, but this copy was not marked with line or even page numbers.

Author affiliation #3 does not seem to be used.

Split drives are reference in the introduction [23,24,26,32,33], but this actually underestimates how widely they are used for biosafety or other reasons. Here are some more (this might even be missing some):

https://elifesciences.org/articles/41439 (first split drives in insects)

https://www.liebertpub.com/doi/10.1089/crispr.2021.0129

https://www.pnas.org/doi/10.1073/pnas.2004373117

https://www.sciencedirect.com/science/article/pii/S2211124720308226

https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001478 (I think the other two mouse papers also used split drives)

https://www.nature.com/articles/s41467-020-14960-3

The long-term behavior of the cage populations would be interesting, such as where the final equilibrium values fall past generation ten. If the authors have kept the cages going, it would be nice to include this data in the manuscript (even just a few more generations), but this is certainly not necessary.

Sincerely,

Jackson Champer

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: In this paper, the authors present the testing of the direct integration of a Cas9 drive cassette at the zpg locus in Anopheles gambiae, with a novel readthrough strategy to minimise modifications. They go on to investigate the impacts of resistance alleles in a target population, analysing homing rates, fitness consequences and drive dynamics to test the efficacy of a system as a proposed non-autonomous drive.

The paper is technically very interesting, well written and clear throughout, and potentially of significant importance in the continued development and refining of gene drive systems for pest control.

I summarise a list of primarily minor comments that I hope will improve the manuscript for publication.

Main comments:

I have some concerns about data management that I hope are easily fixable by the authors. I cannot find the raw data to accompany fertility assays and cage trials; I also cannot see a link to a data repository. Supplementary Table 1 appears to be infection data, the analysis of which I cannot see mentioned anywhere in the manuscript.

The authors summarise one of the uses of this drive as "Self-eliminating drivers could thus enable local population modification without leaving

behind any driver element." However - in this scenario the zpg drive has a suppressive effect on the population. As effector frequency rises within a reducing effective population size, the effect of migration into the release area will change the drive's dynamics. I would like to see the authors discuss the potential implications of migration even if it is not formally incorporated into any models.

Minor comments

Transmission of zpgD and a payload in susceptible and resistant genetic backgrounds:

The authors introduce the Sco (Scorpine)-CP(Carboxypeptidase) line, this is used in a nice experiment to demonstrate that lack of homing of the zpg cassette is due to resistance at that locus and not lack of Cas9 activity; but while it is referenced in another paper it would be nice if the authors could briefly expand into the full names of these constructs and loci here, especially as they mention Scorpine in full elsewhere.

Fertility and gonad development of zpgD transgenics:

In this section the authors make multiple comparisons between homozygous/heterozygous males and females. The experiments are clear, but some aspects of the analysis need clarity or could be improved. There are many pairwise comparisons here

with t-tests (Student's or Welch's? this is undefined), and it is not clearly explained whether these are first tested across one general linear model before diving into these pairwise comparisons. Some of the distributions of data (in some groups there are no offspring) seem ill-suited to a t-test, but I also don't think incorrect conclusions have been drawn here because of the large effect. Although some sample sizes are given at the beginning, it is important to include the degrees of freedom for each t-test as this will determine the critical threshold for statistical significance.

The comparison between male and female transgenics and offspring production does not seem like a 'fair test' I would be surprised if even with sperm depleted males, they could not fertilise all wild-type female eggs - a more thorough comparison would compare them in sperm competition with wild-type males.

When discussing gonad morphology, the measure for ovary atrophy is not described.

Population cage dynamics of zpgD in combination with payload drives:

The authors describe the drive of a non-autonomous element in the presence of target-site resistance at zpg as seemingly paradoxical. But I'm not sure why they describe this as such, many split-drive designs have previously been discussed, this seems perfectly logical.

When describing the " introduction of resistant alleles

would induce gene drive and increase effector frequency at the 3 payload loci" - could the authors summarise this effect?

Figures

Figure 1B - label that the grey letters indicate SNPs

Figure 1D - the scale and colours are not the easiest to read.

Reviewer #2: In an effort to enrich the toolbox for gene drive in mosquito vectors of human malaria, the authors sought to establish a driving Cas9 transgene integrated non-disruptively within the coding sequence of the essential zpg mosquito gene, to serve as a Cas9 source for spreading some non-autonomously driving antimalarial transgenes in a target mosquito population. Somewhat unexpectedly, Cas9 insertion at zpg had a large fitness cost as it impaired zpg activity, especially in females, resulting in strong selection for pre-existing zpg alleles resistant to Cas9 cleavage, counteracting the drive. The authors turn this defect into an advantage as it provides a system for transiently sustaining Cas9 in a mosquito population and reach some level of spread of a set of 3 non-autonomous, gRNA-encoding transgenes, before inexorable disappearance of Cas9 due to the rise of non-costly resistant zpg alleles.

I agree with the conclusions of this study and with most discussion points. The concept and experimental demonstration of a full Cas9 gene drive that self-eliminates after some period of activity, on its own and on distant loci, is quite novel (the molecular mechanism is different and more simple than for the previously theorized “daisy drive” strategy). Programmed disappearance of Cas9 could be partially reassuring to some critics of the gene drive technology, even if the driven non-autonomous transgenes (technologically less frightening) could persist. I appreciate the model of malaria control by gene drive mosquitoes using a combination therapy-like approach through multiple independent anti-malarial transgenes, all driven by the same Cas9 locus. The experimental procedures employed to obtain the results (genotyping and combining defined proportion of mosquitoes of desired genotypes in genetic crosses and follow-up of the frequency of 4 transgenes by genotyping 47 individuals over 6 generations) represent a considerable amount of work.

This work will be of interest to the gene drive field (a contemporary hot topic) and to evolutionary geneticists. It will also be read by stakeholders in mosquito control, including the risk assessment community, and more generally by people interested in and/or worried by gene drive and biotechnology at large. I would be happy to see this study published, once the following points have been addressed or discussed:

Introduction: in an effort towards clarifying vocabulary in the gene drive field, I suggest to reserve the term population replacement for gene drives based on underdominance (whereby the genomic background of the released strain should entirely overwrite that of field populations); and population modification for gene drives that introgress novel genetic traits into a field population’s genetic background without altering the bulk of it (as would be the case with the approach presented here). It’s up to the authors to adopt this point of view or not, if so using “population modification” throughout the manuscript.

Introduction third §: I believe the authors mean “to the expense of the drive” instead of “in favour of the drive”. A bit further: note that if a non-driving Cas9 insertion is costly, it should still disappear by natural selection.

Last § of intro: for more clarity write “capable of providing the antimalarial transgenes with a source of Cas9”….

Results: construct design is not immediately clear: it may be useful to mention that the antisense 3xP3-Scarlet reporter gene must be devoid of any stop codon to allow readthrough translation (also in fig 1B legend). In fact I wanted to check for the absence of stop codons in the antisense 2A-3xP3-Scarlet-2A cassette but the nucleotide sequence of the final construct is not provided. Given the complexity of the cloning scheme used to obtain it, providing an annotated sequence file of the whole construct (plasmid or integrated in its genomic context) as a Suppl. file is required. What might be the physiological effect of expressing the nonsense peptide arising from antisense Scarlet and 3xP3 in zpg-expressing cells? It could contribute to the fitness cost.

Result § “transmission of zpgD…”: for clarity add: “to determine their zpg R or S status, hemizygous parents were then genotyped…” To guide the reader could also add: “Figure 2A, purple bars / Figure 2A, yellow bars”

Fig. 1C: I’m not convinced of the necessity of the left panels with chromatograms, redundant with Fig 1B.

Fig. 1D: why not indicate the polymorphic nucleotides and their position in each strain, not only G3. The heat map for allele frequency is visually not very clear (consider blue to red rather than white to red, to make the dynamic range of frequencies more contrasted?)

Fig. 1D I think the correct spelling of the strain name is Ngousso, a suburb of Yaoundé in Cameroon. Also in Materials & Methods.

Fig 3A The purpose of showing PCR products 2a and 3 is not obvious here (they’re not on the associated DNA gel), I didn’t notice an explicit reference to them in the text.

B of figure 3B is missing.

Text referring to figure 3B: speaking of insemination rate suggests that spermathecae were dissected and counted to assess the number of females that received sperm, independently of sperm fertility. I see in the Methods that this was indeed done but here the number of males that successfully generated progeny is discussed => replace “insemination” by “male reproductive success / ability” ?

In the parenthesis referring to Figure 3C,D you could precise “comparing gonad size in each genotype compared to the WT” for more clarity

Suppl Fig 1A: could mention in the legend that all theoretical curves don’t start exactly at the initial G0 frequency of 0.15 for zpgD and 0.05 for each antimalarial transgene due to simulated random sampling (I was initially confused)

Where 3B is cited, it’s a bit awkward to write that “none” is “significantly fewer” than (anything).

Non-significance of the number of progeny sired by zpgD/D compared to zpgD/R is certainly due to the very low sample size for D/D.

End of the Results section: not immediately obvious what is meant by “mutations distal to the predicted cleavage site” → relatively distant from… Note that the last shown zpg mutant allele should not be a resistant allele, as the TGG PAM is simply changed to GGG, also a valid PAM. I wonder if those mutants with a very distal mutation (e.g. first one for CP) would truly block Cas9 cleavage. Do the authors think these distal mutations arose from Cas9 activity, or are pre-existing polymorphisms detected or not in the 1000 genome sequencing project? The more distal they are, the more likely they’re to be pre-existing polymorphisms. If so, such alleles should not be called mutations but SNPs. Is this what the authors wanted to note in the legend of figure 5B ("Note that for some loci (e.g. Aper1) many, low-frequency alternative alleles exist that are neither R alleles nor reference alleles") ? This note is not so clear and it’s hard to know whether a given allele is R or not. Actually, could these polymorphisms simply be PCR mutations that become visible in high-throughput amplicon sequencing?

It’s actually quite amazing that for Aper and AP2 the resistant alleles gaining most frequency at G6 are very distal mutations that look like they should not completely block Cas9 cleavage. This may suggest that even single very distal mismatches efficiently preclude cutting in Anopheles.

Discussion: please rephrase “and the zpgD Cas9 module was linked with the endogenous zpg CDS by a 2Aseparated translational unit representing the fluorescent transformation marker” in a more precise manner. In fact the 2A-flanked translational unit is a nonsense peptide, it doesn’t represent the fluorescent marker but arises from translation of the marker’s reverse-complement coding and regulatory sequences.

The authors speculate that this design involving a 2A-flanked nonsense peptide could explain the reduction of ZPG protein expression. This could indeed be the case, but alternative explanations are possible. The presence of the Cas9 coding sequence itself may be sufficient to suppress zpg expression. Either way, this may be a translational decrease, or even alternative intron splicing involving the synthetic intron’s 5’ splice site in Cas9 and a 3’ splice junction from one of the downstream zpg introns.

One thought to share with the authors, not necessarily to discuss in the paper: in risk assessment of the zpgD system (should it ever be used in the field), one aspect would be to assess whether natural malaria-susceptibility alleles could be genetically linked to zpgR. This would be undesirable as it would increase the frequency of malaria susceptibility genes hitch-hiking with zpgR ! This could be assessed in the lab by comparing vector susceptibility to P. falciparum of final, zpgR-rich populations to their initial, zpgS-rich population precursor. I agree that besides this caveat, the idea of replacing the Cas9 gene drive allele with a pre-existing natural resistance allele has great appeal.

First § of methods could be more straightforward. The meaning of “we used GFP as a vessel for a new fluorescence marker” is not very clear: if I understand correctly GFP sequence was simply substituted for a new fluorescence marker (if GFP sequence fragments are left over, they should be annotated in the Suppl. nucleotide sequence file I requested).

Second § of Methods: worth mentioning whether the 2A-Scarlet cassette has been verified/recoded to be devoid of stop codons in frame with the Cas9-zpg coding sequences, referring to the new Suppl. sequence file.

Eighth § of Methods: not sure I agree with “Note that because zpgD/zpgS females are not fertile (see below), all hemizygous offspring carried a zpgR allele in their genetic background.”

→ Can’t hemizygous offspring carry zpgD from a zpgD/zpgR mother and zpgS from a zpgD/zpgS father?

Should homogenize usage of the terms –gote / gous throughout the text when used in the adjective position (hemizygote/hemizygous , homo- etc. but as a noun keep –gote)

A few typos : One “rapidly” too many on page 11 of the full pdf, hemiozygote in Methods, “to” is missing in “compared one effector alone”, it’s for its …effector in a wild population, singular/plural inconsistency in last sentence of acknowledgements.

Reviewer #3: Uploaded as an attachment

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Genetics data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: No: I have some concerns about data management that I hope are easily fixable by the authors. I cannot find the raw data to accompany fertility assays and cage trials; I also cannot see a link to a data repository. Supplementary Table 1 appears to be infection data, the analysis of which I cannot see mentioned anywhere in the manuscript.

Reviewer #2: No: Nucleotide sequence of the plasmid knocked into the zpg locus

Reviewer #3: No: Annotated Construct Sequences should be deposited in GenBank or analogous resource

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Eric Marois

Reviewer #3: Yes: Vanessa M. Macias

Submitted filename: Ellis et al PLOS Genetics 3-28-2022.docx

Click here for additional data file.

5 May 2022

Submitted filename: Response to Reviewers 1.3.docx

Click here for additional data file.

9 May 2022

Dear Dr Windbichler,

We are pleased to inform you that your manuscript entitled "Testing non-autonomous antimalarial gene drive effectors using self-eliminating drivers in the African mosquito vector Anopheles gambiae" has been editorially accepted for publication in PLOS Genetics. Congratulations!

Before your submission can be formally accepted and sent to production you will need to complete our formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Please note: the accept date on your published article will reflect the date of this provisional acceptance, but your manuscript will not be scheduled for publication until the required changes have been made.

Once your paper is formally accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you’ve already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosgenetics@plos.org.

In the meantime, please log into Editorial Manager at https://www.editorialmanager.com/pgenetics/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production and billing process. Note that PLOS requires an ORCID iD for all corresponding authors. Therefore, please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field.  This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager.

If you have a press-related query, or would like to know about making your underlying data available (as you will be aware, this is required for publication), please see the end of this email. If your institution or institutions have a press office, please notify them about your upcoming article at this point, to enable them to help maximise its impact. Inform journal staff as soon as possible if you are preparing a press release for your article and need a publication date.

Thank you again for supporting open-access publishing; we are looking forward to publishing your work in PLOS Genetics!

Yours sincerely,

Jackson Champer

Associate Editor

PLOS Genetics

Gregory P. Copenhaver

Editor-in-Chief

PLOS Genetics

www.plosgenetics.org

Twitter: @PLOSGenetics

----------------------------------------------------

Comments from the reviewers (if applicable):

All the changes look fine here. I think that the article is ready for publication. My only comment is that the supplemental method section could probably use some more paragraph breaks.

Sincerely,

Jackson Champer

----------------------------------------------------

Data Deposition

If you have submitted a Research Article or Front Matter that has associated data that are not suitable for deposition in a subject-specific public repository (such as GenBank or ArrayExpress), one way to make that data available is to deposit it in the Dryad Digital Repository. As you may recall, we ask all authors to agree to make data available; this is one way to achieve that. A full list of recommended repositories can be found on our website.

The following link will take you to the Dryad record for your article, so you won't have to re‐enter its bibliographic information, and can upload your files directly:

http://datadryad.org/submit?journalID=pgenetics&manu=PGENETICS-D-22-00190R1

More information about depositing data in Dryad is available at http://www.datadryad.org/depositing. If you experience any difficulties in submitting your data, please contact help@datadryad.org for support.

Additionally, please be aware that our data availability policy requires that all numerical data underlying display items are included with the submission, and you will need to provide this before we can formally accept your manuscript, if not already present.

----------------------------------------------------

Press Queries

If you or your institution will be preparing press materials for this manuscript, or if you need to know your paper's publication date for media purposes, please inform the journal staff as soon as possible so that your submission can be scheduled accordingly. Your manuscript will remain under a strict press embargo until the publication date and time. This means an early version of your manuscript will not be published ahead of your final version. PLOS Genetics may also choose to issue a press release for your article. If there's anything the journal should know or you'd like more information, please get in touch via plosgenetics@plos.org.

25 May 2022

PGENETICS-D-22-00190R1

Testing non-autonomous antimalarial gene drive effectors using self-eliminating drivers in the African mosquito vector Anopheles gambiae

Dear Dr Windbichler,

We are pleased to inform you that your manuscript entitled "Testing non-autonomous antimalarial gene drive effectors using self-eliminating drivers in the African mosquito vector Anopheles gambiae" has been formally accepted for publication in PLOS Genetics! Your manuscript is now with our production department and you will be notified of the publication date in due course.

The corresponding author will soon be receiving a typeset proof for review, to ensure errors have not been introduced during production. Please review the PDF proof of your manuscript carefully, as this is the last chance to correct any errors. Please note that major changes, or those which affect the scientific understanding of the work, will likely cause delays to the publication date of your manuscript.

Soon after your final files are uploaded, unless you have opted out or your manuscript is a front-matter piece, the early version of your manuscript will be published online. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

Thank you again for supporting PLOS Genetics and open-access publishing. We are looking forward to publishing your work!

With kind regards,

Livia Horvath

PLOS Genetics

On behalf of:

The PLOS Genetics Team

Carlyle House, Carlyle Road, Cambridge CB4 3DN | United Kingdom

plosgenetics@plos.org | +44 (0) 1223-442823

plosgenetics.org | Twitter: @PLOSGenetics