Available via license: CC BY-NC 4.0
Content may be subject to copyright.
1
Title: Gene drive mosquitoes can aid malaria elimination by retarding
1
Plasmodium sporogonic development
2
3
4
Authors: Astrid Hoermann1†, Tibebu Habtewold1†, Prashanth Selvaraj2, Giuseppe Del
5
Corsano1, Paolo Capriotti1, Maria G. Inghilterra1, Temesgen M. Kebede1, George K.
6
Christophides1*, Nikolai Windbichler1*
7
8
Affiliations:
9
1Department of Life Sciences, Imperial College London, London, SW7 2AZ, United
10
Kingdom.
11
2Institute for Disease Modeling, Bill and Melinda Gates Foundation, Seattle, WA
12
98109, USA.
13
*Corresponding authors. Email: g.christophides@imperial.ac.uk (G.K.C);
14
n.windbichler@imperial.ac.uk (N.W.).
15
†These authors contributed equally to this work.
16
17
18
Abstract: Gene drives hold promise for the genetic control of malaria vectors. The
19
development of vector population modification strategies hinges on the availability of effector
20
mechanisms impeding parasite development in transgenic mosquitoes. We augmented a
21
midgut gene of the malaria mosquito Anopheles gambiae to secrete two exogenous
22
antimicrobial peptides, Magainin 2 and Melittin. This small genetic modification, capable of
23
efficient non-autonomous gene drive, hampers oocyst development in both Plasmodium
24
falciparum and Plasmodium berghei. It delays the release of infectious sporozoites while it
25
simultaneously reduces the lifespan of homozygous female transgenic mosquitoes. Modeling
26
the spread of this modification using a large-scale agent-based model of malaria epidemiology
27
reveals that it can break the cycle of disease transmission across a range of endemic settings.
28
29
30
Short title: Gene drive mosquitoes retard Plasmodium sporogony
31
32
One sentence summary: We developed a gene drive effector that retards Plasmodium
33
development in transgenic Anopheles gambiae mosquitoes via the expression of antimicrobial
34
peptides in the midgut and which is predicted to eliminate malaria under a range of
35
transmission scenarios.
36
37
Keywords: malaria, gene drives, population modification, CRISPR/Cas9
38
39
40
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
2
Main text
1
2
Malaria remains one of the most devastating human diseases. A surge in insecticide-resistant
3
mosquitoes and drug-resistant parasites has brought a decades-long period of progress in
4
reducing cases and deaths to a standstill (1). Despite the availability of the first WHO approved
5
malaria vaccine (2) the necessity to develop alternative intervention strategies remains
6
pressing, particularly if malaria elimination is to remain the goal. Gene drive, based on the
7
super-Mendelian spread of endonuclease genes, is a promising new control strategy that has
8
been under development for over a decade (3). Suppressing mosquito populations by targeting
9
female fertility has remained a prime application of gene drives and to date specific gene drives
10
have been shown to eliminate caged mosquito populations (4-6). Gene drives for population
11
replacement (or modification), designed to propagate antimalarial effector traits, have also seen
12
significant development in past years (7-9). To date, a range of antimalarial effectors and
13
tissue-specific drivers have been tested in transgenic mosquitoes, and some of them have been
14
shown to reduce Plasmodium infection prevalence or infection intensity (10-23). However, the
15
pursuit of novel and effective mechanisms is ongoing, especially in A. gambiae where effectors
16
so far have shown only moderate reductions in parasite transmission (12, 15, 20-22).
17
18
Antimicrobial peptides (AMPs) from reptiles, plants or insects have long been considered
19
putative antimalarial effectors and have been tested in vitro and in vivo for their efficacy against
20
different parasite life stages (for reviews see (24-27); for more recent studies see (28, 29)).
21
Although AMPs are very diverse in sequence and structure, many are cationic and amphiphilic
22
and thus tend to adhere to negatively charged microbial membranes and to a much lesser extent
23
to membranes of animal cells (30). Permeabilization mechanisms have been proposed, which
24
rely on either pore formation or accumulation of peptides on the microbial surface causing
25
disruption in a detergent‐like manner (31). A subset of AMPs has been suggested to act by
26
mitochondrial uncoupling, interfering directly with mitochondria-dependent ATP synthesis
27
(32-34). Two such peptides, Magainin 2, found within skin secretions of the African claw frog
28
Xenopus laevis, and Melittin, a primary toxin component of the European honeybee Apis
29
mellifera, have been shown to both form pores on the microbial membrane (35, 36) and trigger
30
uncoupling of mitochondrial respiration (37-40). Intrathoracic injection of Magainin 2 into
31
Anopheles mosquitoes has been demonstrated to cause Plasmodium oocyst degeneration and
32
shrinkage and a consequent reduction in the number of sporozoites released (41), while a
33
transmission blocking effect of Magainin 2 has also been revealed when spiked into
34
gametocytaemic blood at a 50 µM concentration (29). Similarly, Melittin has been shown to
35
reduce the number and prevalence of P. falciparum oocysts in spike-in experiments at
36
concentrations as low as 4 µM (28, 29), while expression of Melittin in transgenic A. stephensi
37
mosquitoes as a part of a multi-effector transgene, additionally including the AMPs TP10,
38
EPIP, Shiva1 and Scorpine, has led to a significant reduction in oocyst prevalence and infection
39
intensity (23).
40
41
Here, we augmented two host genes of A. gambiae to co-express Magainin 2 and Melittin
42
following the previously described integral gene drive (IGD) paradigm (42). This allowed for
43
minimal genetic modifications, capable of non-autonomous gene drive, to be introduced into
44
the host gene loci making full use of the gene regulatory regions for controlling tissue-specific
45
expression of the AMPs. We utilized the previously evaluated zinc carboxypeptidase A1 (CP,
46
AGAP009593) and the AMP Gambicin 1 (Gam1, AGAP008645) as host genes for the
47
exogenous AMP integration (Fig. 1A). The transcriptional profile of these genes was expected
48
to drive expression of the AMPs in the mosquito midgut upon ingestion of a bloodmeal (CP)
49
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
3
or in the anterior gut (Gam1). The use of 2A ribosome-skipping peptides as well as secretion
1
signals guaranteed separate secretion of the exogenous AMPs and host gene products. We also
2
replaced the signal peptides and prepropeptides of Magainin 2 and Melittin with the
3
endogenous secretion signals of A. gambiae Cecropin 1 and 2 genes, respectively. An intron
4
harboring the guide RNA-module that enables non-autonomous gene drive and the fluorescent
5
marker-module required for transgenesis was introduced within the Melittin coding sequence.
6
Transgenesis of A. gambiae G3 strain via CRISPR/Cas9-mediated homology-directed repair
7
(HDR) and subsequent removal of the GFP transformation maker resulting in the establishment
8
of homozygous markerless strains, designated as Gam1-MM and MM-CP, were performed as
9
previously described (42) (Fig. 1A). We validated and tracked the correct transgene insertion
10
by genomic PCR (Fig. 1B) and confirmed the expression of both CP and Gam1 host genes and
11
the inserted AMP cassette by RT-PCR (Fig. 1C). Sequencing of the cDNA amplicon over the
12
splice site of the artificial intron revealed the expected splicing pattern in 89.9% of all MM-CP
13
reads but only in 65.3% of all Gam1-MM reads (Fig. 1D). A cryptic splice site resulting in a
14
loss of additional 25bp from the Melittin coding sequence accounted for most of the unexpected
15
splicing events (Fig. S1).
16
17
Next, we performed infection experiments with the P. falciparum NF54 strain to determine the
18
effect of these modifications on parasite transmission (Fig. 2A). Both transgenic strains showed
19
a significant reduction in the midgut oocyst loads on day 7 post infection (pi; Fig. 2B). While
20
only few oocysts of the MM-CP strain had the expected size, a closer examination of infected
21
midguts revealed the presence of many smaller structures possibly representing stunted or
22
aborted oocysts (Fig. 2C), prompting a more detailed investigation of this strain. We performed
23
infections and quantified the number and diameter of all oocyst-like structures on days 7, 9 and
24
15 pi. Given that nutritional stress is a factor that recently emerged as causing stunting of
25
oocysts (43, 44), a group of mosquitoes were provided a supplemental bloodmeal on day 4 pi.
26
We found that the small oocyst-like structures were indeed stunted oocysts that grew over time,
27
and that the supplemental bloodmeal further boosted their growth (Fig. 2D). Overall oocysts
28
infecting the MM-CP strain were significantly smaller than in the wildtype control by an
29
average of 47.8%, 41.1% and 59.8% on days 7, 9 and 15 pi, respectively (Fig. 2D). This was
30
also the case for the cohort that received an additional bloodmeal, but the difference with
31
wildtype controls decreased over time to 50.4%, 24.9% and 18.6% on days 7, 9 and 15 pi,
32
respectively. We repeated this experiment with the rodent parasite P. berghei to determine if
33
the observed infection phenotype would also occur with a different parasite species and under
34
different environmental conditions. Using a GFP-labeled P. berghei ANKA 2.34 strain,
35
fluorescent imaging revealed clearly stunted oocysts that on day 14 pi were about two-times
36
smaller than in the control (Fig. 2E, F).
37
38
We reasoned that the detected retardation of oocyst development would in turn cause a delayed
39
release of sporozoites into the mosquito haemocoel and subsequent infection of the salivary
40
glands. To determine the sporozoite load over time, we determined the abundance of parasite
41
DNA on days 10 to 16 pi by quantifying the P. falciparum cytochrome B (Cyt-B) gene in the
42
head and thorax of single mosquitos via qPCR, considering only mosquitoes that were positive
43
for parasite DNA in the midgut. We found that sporozoite infection prevalence, i.e., the rate of
44
head and thorax samples with amplification above the amplification cycle threshold, was
45
significantly reduced in MM-CP mosquitoes (Fig. 2G) when compared to wildtype (on average
46
by 77.9% across timepoints) with a significant number of positives detectable only on day 16.
47
Although a supplemental bloodmeal accelerated sporozoite release in the MM-CP strain by 4-
48
5 days (now detected on days 11-12 pi), overall sporozoite prevalence was still reduced
49
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
4
significantly relative to the control by 67.8% on average. This suggested that sporozoite release
1
in the MM-CP strain was delayed under both nutritional conditions. We also analyzed the
2
relative parasite DNA content in positive samples as a proxy for sporozoite numbers and found
3
that these were significantly higher in wildtype compared to MM-CP mosquitoes both after a
4
single (14.0-fold) or double (37.8-fold) bloodmeals (Fig 2H). To rule out any founder effect in
5
the homozygous MM-CP strain, we outcrossed MM-CP mosquitoes to mosquitoes of the A.
6
gambiae Ifakara strain. After an F1 sibling cross, F2 mosquitoes were provided an infected
7
bloodmeal and dissected on day 9 to assess the midgut oocyst size and be genotyped by PCR
8
for the presence of the transgene. The results showed that the effect on parasite development
9
was indeed attributable to the presence of the transgene and that the developmental retardation
10
of oocysts was reduced in hemizygous individuals (Fig. S2).
11
12
Next, we measured fitness parameters of MM-CP mosquitoes. The results showed a 14.1%
13
difference in the number of eggs laid (fecundity; p=0.0299, two-sample t-test assuming unequal
14
variances) but unaffected larval hatching rates (fertility) between the MM-CP and wildtype
15
control mosquitoes (Fig. 3A, B). Pupal sex ratio (Fig. 3C) and pupation time did not
16
significantly deviate between MM-CP and control mosquitoes (Fig. S3). However, a
17
significant effect on the lifespan of sugar-fed mosquitoes was detected for MM-CP females
18
(median lifespan 15 days) and to a lesser extent in males (23 days) compared to control females
19
(26 days) and males (27 days; Fig. 3D). We repeated this experiment with females now also
20
been provided regular bloodmeals. As CP is only weakly expressed in the sugar-fed female
21
midgut but strongly induced following a blood feed, any effect of the transgene was expected
22
to be elevated by the bloodmeals. As before, to rule out that inbreeding accounted for this
23
effect, we first performed a backcross to the Ifakara strain and genotyped individual F2
24
mosquitoes at the end of the experiment. The results confirmed a significant lifespan reduction
25
in homozygous MM-CP females under these conditions, but no significant effect was detected
26
in hemizygous individuals compared to non-transgenic controls (Fig. 3E).
27
28
We performed transcriptomic analysis of dissected midguts prior to and 6 and 20 hours after
29
blood feeding. We quantified the number of differentially expressed genes between MM-CP
30
and control females and performed a gene ontology (GO) analysis to determine significantly
31
enriched gene groups (Table S1). The results indicated that genes involved in mitochondrial
32
function and located at the inner mitochondrial membrane were disproportionally affected in
33
MM-CP females, particularly after the bloodmeal (Fig. 3F-H). Among most significant hits
34
were genes encoding a member of the ubiquinone complex (AGAP003900), a mitochondrial
35
H+ ATPase (AGAP012818), an ATP synthase subunit (AGAP004788) and a protein belonging
36
to a family of calcium channels (AGAP002578), which control the rate of mitochondrial ATP
37
production. These findings offered a hypothesis that could explain the dual phenotype
38
regarding parasite development and adult female lifespan. Plasmodium development in the
39
mosquito is critically dependent on mitochondrial function including active respiration (45-
40
48). Magainin 2 and Melittin are known to trigger mitochondrial uncoupling and could, upon
41
secretion into the midgut lumen, interfere with ATP synthesis targeting the parasite
42
mitochondrion. This effect would become apparent as the parasite transforms into the energy-
43
demanding oocyst stage that undergoes several rounds of endomitosis and vegetative growth.
44
As AMPs are unlikely to be able to access the oocyst, the effect would wear off with time and
45
indeed partly offset by supplemental bloodmeals. The AMPs, however, are likewise expected
46
to affect the mosquito midgut mitochondria, impacting energy homeostasis and modulating
47
lifespan. Whilst further experiments are needed to untangle these effects, the most significant
48
knowledge gap, when it comes to transmission blocking, is to what degree any effects observed
49
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
5
with lab strains of P. falciparum would be reproducible in infections with genetically diverse
1
parasites isolated from patient blood. The MM-CP strain is an excellent candidate to attempt
2
to answer this question as the transmission-blocking mechanism we describe here appears to
3
act across Plasmodium species. MM-CP is incapable of autonomous gene drive, unless it mates
4
with a mosquito source of Cas9, and can thus be evaluated in an endemic setting under standard
5
mosquito confinement protocols.
6
7
Finally, we predicted how deployment of the MM-CP effector trait would modify malaria
8
epidemiology using a mechanistic, agent-based model of P. falciparum transmission that
9
includes vector life cycle and within-host parasite and immune dynamics. The model is based
10
on the EMOD framework that has recently been updated to enable the simulation of gene drives
11
(49). There remain knowledge gaps that preclude a direct translation of experimental
12
entomological or molecular data into epidemiological parameters, for example linking the
13
observed reduction in sporozoite DNA and its quantitative effect on onward transmission. For
14
the phenotypic effects, we thus estimated likely parameter value ranges (a 30-70% increase in
15
time until sporozoites are released and a 40-100% reduction in infectious sporozoites) that we
16
considered to be within physiologically plausible limits supported by our in vivo experiments.
17
As a final parameter for the model, we experimentally determined the rate of non-autonomous
18
gene drive of the MM-CP allele by pairing it with a source of Cas9, which resulted in high
19
levels of homing of the transgene in both males and females: 96.01% and 98.91%, respectively
20
(Fig. 4A). MM-CP, as a non-autonomous effector, could be flexibly deployed in conjunction
21
with non-driving, self-limiting or, as we assumed in our model, a fully autonomous Cas9 gene
22
drive that is able to mobilize MM-CP (Fig. S4). We determined the probability of elimination
23
by the last year of simulation, defined as the number of simulations per parameter set that have
24
zero prevalence in the last year divided by the total number of simulations. Incidence reduction
25
was evaluated for the duration of the simulation following gene drive releases compared to
26
control scenarios with no releases. In a low transmission setting (annual EIR ~15 infectious
27
bites per person), most simulations within the space of likely parameter estimates resulted in
28
the elimination of malaria transmission (Fig. 4B & Fig. S5A). As transmission intensity
29
increased (annual EIR ~30), we detected a reduction in probability of elimination in the lower
30
end of the parameter estimate range (Fig. 4B & Fig. S5B). In a high transmission scenario, a
31
high reduction in sporozoite production in combination with a large delay was necessary to
32
reliably trigger elimination (Fig. 4B & Fig. S5C). It should, however, be noted that significant
33
reductions in clinical cases occurred even when elimination was not achieved. Therefore, in
34
high transmission settings, even when not achieving elimination alone, MM-CP could open a
35
window for elimination by strategically deploying other interventions that could act
36
synergistically to drive transmission to zero.
37
38
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
6
Figure legends
1
2
Figure 1. Generation of gene drive effector strains expressing AMPs.
3
(A) Schematic showing the design and integration strategy of the effector cassette coding for
4
Magainin 2 and Melittin at the endogenous loci Gam1 and CP. The gRNA target sequences
5
(red), or gRNA module (red circle) are indicated including the protospacer adjacent motif
6
(bold) and the Stop and Start codons (underlines). Coding sequences (CDS) and signal peptides
7
(CDS-SP) are indicated by light shading. Half arrows indicate primer binding sites for genomic
8
PCR and RT-PCR. (B) PCR on genomic DNA of 15 pooled homozygous Gam1-MM, MM-CP
9
or wild type individuals. (C) RT-PCR of midguts from wildtype (WT), Gam1-MM and MM-
10
CP mosquitoes that were either non-blood-fed (NBF) or dissected three hours post blood-
11
feeding (3h PBF). (D) Analysis of cDNA amplicons over the splice site subjected to next
12
generation sequencing showing the predicted splicing outcomes for strains Gam1-MM and
13
MM-CP.
14
15
16
Figure 2. Plasmodium infection experiments.
17
(A) Schematic overview of Plasmodium infection experiments. (B) P. falciparum oocyst
18
intensity 7 days pi (dpi) in midguts from wildtype (WT), MM-CP and Gam1-MM mosquitoes
19
dissected. Data from three biological replicates was pooled and statistical analysis was
20
performed using the Mann-Whitney test. (C) Bright-field images of midguts showing typical
21
oocysts in WT and MM-CP mosquitoes at 7 dpi. (D) Quantification of P. falciparum oocyst
22
diameter in WT and MM-CP mosquitoes 7, 9 and 15 dpi from 3 pooled biological replicates.
23
Note, that many oocysts in WT mosquitoes had ruptured on day 15 pi. Quantification of oocyst
24
diameter (E) and fluorescent imaging (F) of oocysts in WT and MM-CP mosquitoes infected
25
with P. berghei at 14 dpi. Sporozoite prevalence 10 to 16 dpi (G) and infection intensity across
26
all days (H) was measured by qPCR of the P. falciparum Cyt-B gene in dissected heads and
27
thoraces of individual MM-CP and WT mosquitoes (10-16 dpi). Only mosquitoes positive for
28
oocyst DNA in the midgut were included in the analysis performed in two biological replicates.
29
Statistical analysis in panels D, E & H was performed by a t-test assuming unequal variance.
30
Statistical analysis in panel G was performed using a generalized linear model with a
31
quasibinomial error structure where strain (p=8.35e-06), dpi (p=7.15e-4) but not bloodmeal
32
status (p= 0.0894) were found to be significant coefficients. In all panels, the provision of a
33
supplemental bloodmeal 4 dpi is indicated by an additional blood drop. Significance codes
34
*(p≤0.05), **(p≤0.01), ***(p≤0.001) and ns (not significant).
35
36
37
Figure 3. Life history traits and midgut transcriptome of MM-CP mosquitoes.
38
(A) Fecundity of individual homozygous MM-CP females compared to the wildtype (WT) and
39
corresponding (B) larval hatching rates obtained during the first gonotrophic cycle. Data from
40
3 pooled biological replicates are shown. Statistical significance was determined by a t-test
41
assuming unequal variance. (C) Pupal sex ratios of MM-CP and WT strains analyzed using the
42
χ2 test for equality. (D) Survival analysis of MM-CP and WT male and female mosquitoes
43
maintained on sugar and (E) of F2 genotyped MM-CP female mosquitoes following
44
backcrossing to the Ifakara strain, intercrossing of F1 mosquitoes and provision of bloodmeals.
45
Statistical significance was determined with a Mantel-Cox log rank test. Data from three
46
biological replicates are pooled and the mean and 95% confidence intervals are plotted.
47
Significance codes *(p≤0.05), **(p≤0.01), ***(p≤0.001) and ns (not significant). (F) Volcano
48
plots of an RNAseq experiment performed on midguts dissected before or 6 hours and 20 hours
49
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
7
post bloodmeal. Differentially expressed genes between MM-CP and WT mosquitoes (p≤0.01)
1
are indicated, and genes belonging to enriched GO groups are highlighted in red.
2
3
Figure 4. Gene drive and predicted epidemiological impact of strain MM-CP deployment.
4
(A) Assessment of non-autonomous gene drive in the progeny of male and female hemizygous
5
MM-CP mosquitoes in the presence or absence of a vasa-Cas9 driver crossed to the wildtype
6
(wt). Larval offspring were subjected to multiplex PCR genotyping and the mean and standard
7
error (SEM) from three biological replicates is plotted, and the total number n is indicated.
8
Statistical significance was determined using a one-way ANOVA with Tukey’s correction.
9
Significance codes *(p≤0.05), **(p≤0.01), ***(p≤0.001) and ns (not significant). (B)
10
Heatmaps depicting elimination probabilities (top) and number of clinical cases reduced
11
(bottom) at the end of 6 years following a single release of 1000 homozygous MM-CP
12
mosquitoes that also carry a Cas9 integral gene drive. Three transmission scenarios with
13
varying entomological inoculation rates (EIR) for P. falciparum as a measure of exposure to
14
infectious mosquitoes were explored. Homozygous transgenic mosquitoes are released 6
15
months after the start of the simulation in highly seasonal transmission settings of varying
16
intensities. The parameter range we explored for the reduction of the number of infectious
17
sporozoites is represented on the major y-axis while the range for the average increase in time
18
until sporozoites are released is represented on the major x-axis. Parameter likelihood estimates
19
based on the experimental data are also indicated next to the values (grayscale).
20
21
22
Acknowledgments: We thank Eric Marois for sharing the vasa-Cre and vasa-Cas9 lines. We
23
thank Dan Bridenbecker for software support and Austin Burt for helpful suggestions on the
24
manuscript.
25
26
Funding: The work was funded by the Bill and Melinda Gates Foundation grant OPP1158151
27
to N.W. and G.K.C.
28
29
Competing interests: The authors declare that no competing interests exist.
30
31
32
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
8
References
1
2
1. World malaria report 2020: 20 years of global progress and challenges. (World Health
3
Organization, Geneva, 2020).
4
2. M. B. Laurens, RTS,S/AS01 vaccine (Mosquirix): an overview. Hum Vaccin Immunother 16,
5
480-489 (2020).
6
3. N. Windbichler et al., A synthetic homing endonuclease-based gene drive system in the
7
human malaria mosquito. Nature 473, 212-215 (2011).
8
4. A. Hammond et al., A CRISPR-Cas9 gene drive system targeting female reproduction in the
9
malaria mosquito vector Anopheles gambiae. Nat Biotechnol 34, 78-83 (2016).
10
5. K. Kyrou et al., A CRISPR-Cas9 gene drive targeting doublesex causes complete population
11
suppression in caged Anopheles gambiae mosquitoes. Nat Biotechnol 36, 1062-1066 (2018).
12
6. A. Hammond et al., Regulating the expression of gene drives is key to increasing their
13
invasive potential and the mitigation of resistance. PLoS Genet 17, e1009321 (2021).
14
7. V. M. Gantz et al., Highly efficient Cas9-mediated gene drive for population modification of
15
the malaria vector mosquito Anopheles stephensi. Proc Natl Acad Sci U S A 112, E6736-6743
16
(2015).
17
8. T. B. Pham et al., Experimental population modification of the malaria vector mosquito,
18
Anopheles stephensi. PLoS Genet 15, e1008440 (2019).
19
9. A. Adolfi et al., Efficient population modification gene-drive rescue system in the malaria
20
mosquito Anopheles stephensi. Nat Commun 11, 5553 (2020).
21
10. J. Ito, A. Ghosh, L. A. Moreira, E. A. Wimmer, M. Jacobs-Lorena, Transgenic anopheline
22
mosquitoes impaired in transmission of a malaria parasite. Nature 417, 452-455 (2002).
23
11. L. A. Moreira et al., Bee venom phospholipase inhibits malaria parasite development in
24
transgenic mosquitoes. J Biol Chem 277, 40839-40843 (2002).
25
12. W. Kim et al., Ectopic expression of a cecropin transgene in the human malaria vector
26
mosquito Anopheles gambiae (Diptera: Culicidae): effects on susceptibility to Plasmodium. J
27
Med Entomol 41, 447-455 (2004).
28
13. E. G. Abraham et al., Driving midgut-specific expression and secretion of a foreign protein in
29
transgenic mosquitoes with AgAper1 regulatory elements. Insect Mol Biol 14, 271-279
30
(2005).
31
14. V. Corby-Harris et al., Activation of Akt signaling reduces the prevalence and intensity of
32
malaria parasite infection and lifespan in Anopheles stephensi mosquitoes. PLoS Pathog 6,
33
e1001003 (2010).
34
15. J. M. Meredith et al., Site-specific integration and expression of an anti-malarial gene in
35
transgenic Anopheles gambiae significantly reduces Plasmodium infections. PLoS One 6,
36
e14587 (2011).
37
16. A. T. Isaacs et al., Engineered resistance to Plasmodium falciparum development in
38
transgenic Anopheles stephensi. PLoS Pathog 7, e1002017 (2011).
39
17. A. T. Isaacs et al., Transgenic Anopheles stephensi coexpressing single-chain antibodies
40
resist Plasmodium falciparum development. Proc Natl Acad Sci U S A 109, E1922-1930
41
(2012).
42
18. E. S. Hauck et al., Overexpression of phosphatase and tensin homolog improves fitness and
43
decreases Plasmodium falciparum development in Anopheles stephensi. Microbes Infect 15,
44
775-787 (2013).
45
19. A. J. Arik et al., Increased Akt signaling in the mosquito fat body increases adult
46
survivorship. FASEB J 29, 1404-1413 (2015).
47
20. G. Volohonsky et al., Transgenic Expression of the Anti-parasitic Factor TEP1 in the Malaria
48
Mosquito Anopheles gambiae. PLoS Pathog 13, e1006113 (2017).
49
21. M. L. Simoes et al., The Anopheles FBN9 immune factor mediates Plasmodium species-
50
specific defense through transgenic fat body expression. Dev Comp Immunol 67, 257-265
51
(2017).
52
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
9
22. S. Dong et al., Broad spectrum immunomodulatory effects of Anopheles gambiae
1
microRNAs and their use for transgenic suppression of Plasmodium. PLoS Pathog 16,
2
e1008453 (2020).
3
23. Y. Dong, M. L. Simoes, G. Dimopoulos, Versatile transgenic multistage effector-gene
4
combinations for Plasmodium falciparum suppression in Anopheles. Sci Adv 6, eaay5898
5
(2020).
6
24. V. Carter, H. Hurd, Choosing anti-Plasmodium molecules for genetically modifying
7
mosquitoes: focus on peptides. Trends Parasitol 26, 582-590 (2010).
8
25. A. Bell, Antimalarial peptides: the long and the short of it. Curr Pharm Des 17, 2719-2731
9
(2011).
10
26. S. Wang, M. Jacobs-Lorena, Genetic approaches to interfere with malaria transmission by
11
vector mosquitoes. Trends Biotechnol 31, 185-193 (2013).
12
27. N. Vale, L. Aguiar, P. Gomes, Antimicrobial peptides: a new class of antimalarial drugs?
13
Front Pharmacol 5, 275 (2014).
14
28. V. Carter et al., Killer bee molecules: antimicrobial peptides as effector molecules to target
15
sporogonic stages of Plasmodium. PLoS Pathog 9, e1003790 (2013).
16
29. T. Habtewold et al., Streamlined SMFA and mosquito dark-feeding regime significantly
17
improve malaria transmission-blocking assay robustness and sensitivity. Malar J 18, 24
18
(2019).
19
30. E. Martin, T. Ganz, R. I. Lehrer, Defensins and other endogenous peptide antibiotics of
20
vertebrates. J Leukoc Biol 58, 128-136 (1995).
21
31. A. Giuliani, G. Pirri, S. F. Nicoletto, Antimicrobial peptides: an overview of a promising class
22
of therapeutics. Cent Eur J Biol 2, 1-33 (2007).
23
32. Z. Ahmad, T. F. Laughlin, Medicinal chemistry of ATP synthase: a potential drug target of
24
dietary polyphenols and amphibian antimicrobial peptides. Curr Med Chem 17, 2822-2836
25
(2010).
26
33. H. Moravej et al., Antimicrobial Peptides: Features, Action, and Their Resistance
27
Mechanisms in Bacteria. Microb Drug Resist 24, 747-767 (2018).
28
34. D. Zhang et al., Little Antimicrobial Peptides with Big Therapeutic Roles. Protein Pept Lett
29
26, 564-578 (2019).
30
35. K. Matsuzaki, O. Murase, N. Fujii, K. Miyajima, An antimicrobial peptide, magainin 2,
31
induced rapid flip-flop of phospholipids coupled with pore formation and peptide
32
translocation. Biochemistry 35, 11361-11368 (1996).
33
36. L. Yang, T. A. Harroun, T. M. Weiss, L. Ding, H. W. Huang, Barrel-stave model or toroidal
34
model? A case study on melittin pores. Biophys J 81, 1475-1485 (2001).
35
37. M. Hugosson, D. Andreu, H. G. Boman, E. Glaser, Antibacterial peptides and mitochondrial
36
presequences affect mitochondrial coupling, respiration and protein import. Eur J Biochem
37
223, 1027-1033 (1994).
38
38. J. R. Gledhill, J. E. Walker, Inhibition sites in F1-ATPase from bovine heart mitochondria.
39
Biochem J 386, 591-598 (2005).
40
39. T. F. Laughlin, Z. Ahmad, Inhibition of Escherichia coli ATP synthase by amphibian
41
antimicrobial peptides. Int J Biol Macromol 46, 367-374 (2010).
42
40. H. V. Westerhoff, D. Juretic, R. W. Hendler, M. Zasloff, Magainins and the disruption of
43
membrane-linked free-energy transduction. Proc Natl Acad Sci U S A 86, 6597-6601 (1989).
44
41. R. W. Gwadz et al., Effects of magainins and cecropins on the sporogonic development of
45
malaria parasites in mosquitoes. Infect Immun 57, 2628-2633 (1989).
46
42. A. Hoermann et al., Converting endogenous genes of the malaria mosquito into simple non-
47
autonomous gene drives for population replacement. Elife 10, (2021).
48
43. T. Habtewold et al., Plasmodium oocysts respond with dormancy to crowding and nutritional
49
stress. Sci Rep 11, 3090 (2021).
50
44. W. R. Shaw et al., Multiple blood feeding in mosquitoes shortens the Plasmodium falciparum
51
incubation period and increases malaria transmission potential. PLoS Pathog 16, e1009131
52
(2020).
53
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
10
45. H. Ke et al., Genetic investigation of tricarboxylic acid metabolism during the Plasmodium
1
falciparum life cycle. Cell Rep 11, 164-174 (2015).
2
46. C. D. Goodman et al., Parasites resistant to the antimalarial atovaquone fail to transmit by
3
mosquitoes. Science 352, 349-353 (2016).
4
47. A. Sturm, V. Mollard, A. Cozijnsen, C. D. Goodman, G. I. McFadden, Mitochondrial ATP
5
synthase is dispensable in blood-stage Plasmodium berghei rodent malaria but essential in the
6
mosquito phase. Proceedings of the National Academy of Sciences 112, 10216-10223 (2015).
7
48. J. M. Matz, C. Goosmann, K. Matuschewski, T. W. A. Kooij, An Unusual Prohibitin
8
Regulates Malaria Parasite Mitochondrial Membrane Potential. Cell Rep 23, 756-767 (2018).
9
49. P. Selvaraj et al., Vector genetics, insecticide resistance and gene drives: An agent-based
10
modeling approach to evaluate malaria transmission and elimination. PLoS Comput Biol 16,
11
e1008121 (2020).
12
50. S. Yoshida et al., Hemolytic C-type lectin CEL-III from sea cucumber expressed in
13
transgenic mosquitoes impairs malaria parasite development. PLoS Pathog 3, e192 (2007).
14
51. G. Volohonsky et al., Tools for Anopheles gambiae Transgenesis. G3 (Bethesda) 5, 1151-
15
1163 (2015).
16
52. W. S. Rasband, ImageJ. U. S. National Institutes of Health, https://imagej.nih.gov/ij/ (1997-
17
2018).
18
53. C. Farrugia et al., Cytochrome b gene quantitative PCR for diagnosing Plasmodium
19
falciparum infection in travelers. J Clin Microbiol 49, 2191-2195 (2011).
20
54. D. Kim, J. M. Paggi, C. Park, C. Bennett, S. L. Salzberg, Graph-based genome alignment and
21
genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol 37, 907-915 (2019).
22
55. J. R. Adrian Alexa, topGO: Enrichment Analysis for Gene Ontology. R package version
23
2.38.1., (2019).
24
56. A. Bershteyn et al., Implementation and applications of EMOD, an individual-based multi-
25
disease modeling platform. Pathog Dis 76, (2018).
26
57. A. Nash et al., Integral gene drives for population replacement. Biol Open 8, (2019).
27
28
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
11
Materials and Methods
Design and generation of constructs
Annotated DNA sequence files for the final transformation constructs pD-Gam1-MM and pD-
MM-CP are provided in Supplementary file S1. Briefly, the 23 N-terminal amino acids of
Cecropin 1 (Cec1, CecA, AGAP000693) and Cecropin 2 (Cec2, CecB, AGAP000692) served
as secretion signals. Magainin 2, Melittin, T2A and P2A were codon usage optimized for
Anopheles gambiae, and the intron located within the Melittin coding sequence was previously
described (42) except for the SV40 terminator within the eGFP marker module for which we
swapped in the Trypsin terminator (50). We first neutralized a BsaI site in the Ampicillin
resistance cassette and gene synthesized (Genewiz) a fragment ranging from the Cecropin 1
secretion signal to the Trypsin terminator, including 18 bp overlaps with the vector backbone
and eGFP for subsequent Gibson assembly. The marker-module and the U6 promoter was PCR
amplified from pI-Scorpine (42) with primers 78-GFP-R and 167-U6-R. The fragment from
the BsaI spacer to the P2A was synthesized (Genewiz), including 18 bp overlaps to the U6
promoter and the vector backbone. The vector backbone was PCR amplified from
pAmpR_SDM with primers 168-BBmut-F and 169-BBmut-R, and finally the 4 fragments were
joined via Gibson assembly to yield the intermediate plasmid pI-MM. The CP gRNA spacer
(42) was inserted via the BsaI sites and the cassette was amplified with primers 172- CP-HA3-
F-degen and 173-CP-HA5-R-Cec1 and fused with the CP homology arms and backbone
amplified from pD-Sco-CP (42) with primers 170-Cec1-SS-F and 171-P2A-R to assemble the
donor plasmid pD-MM-CP. A 5’ P2A was added to the cassette via Golden Gate cloning of
the annealed oligos 182-P2Aanneal-F and 183-P2Aanneal-R into BglII digested plasmid pI-
MM, and subsequently the Gambicin 1 gRNA spacer (5’-TACAGAATGTTTCTTCTGAG-3’)
was inserted via BsaI. The gRNA sequence was chosen using Deskgen (Desktop Genetics,
LTD) with an activity-score of 54 and an off-target score of 99. The effector cassette was
amplified with primers 184-P2Adegen-F and 185-Mel-R and fused via Gibson assembly with
the Gambicin homology arms amplified from G3 genomic gDNA and the backbone to generate
the donor plasmid pD-Gam1-MM. For primers see Supplementary Table S2
Transgenesis and establishment of markerless strains
Anopheles gambiae G3 eggs were injected with the corresponding donor plasmids pD-MM-
CP or pD-Gam1-MM and the Cas9 helper plasmid p155 (4). 25 F1 transgenics were obtained
for MMGFP-CP and one female F1 transgenic for Gam1-MMGFP. MMGFP-CP was established
from a founder cage with 9 females and F1 individuals were confirmed by Sanger-sequencing
with primers EGFP-C-For, 117-CP-ctrl-R and 163-P3-probe-F, and Gam1-MMGFP with
primers EGFP-C-For and EGFP-N. Transgenics were outcrossed to G3 WT over three
generations for Gam1-MMGFP and two generations for MMGFP-CP before crossing to the vasa-
Cre (51) strain in the KIL background. Larval offspring were screened for GFP and DsRed and
siblings mated. The progeny was screened against GFP and DsRed, pupae were singled out
and the exuviate collected for genotyping with primers 99-CP-locus-F and 100-CP-locus-R or
241-Gam-locus-F and 242-Gam-locus-R, respectively, to identify homozygotes. The
markerless line MM-CP was established from 9 males and 11 females. For Gam1-MM, 3 cups
with 1 female and 1 male and 6 cups with 2 males and 2 females were set-up and pooled after
confirmation via Sanger-sequencing. A G3-KIL mixed colony was used as WT control for all
experiments, unless otherwise stated. All experiments were performed with cow blood (First
Link (UK) Ltd.), unless otherwise stated.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
12
RT-PCR and splicing analysis
MM-CP and the WT control were fed with human blood and midguts were dissected after 3h.
For Gam1-MM this experiment was performed on unfed females. 30 guts were lysed in Trizol
and homogenized with 2.8mm ceramic beads (CK28R, Precellys) for 30s at 6,800rpm in a
Precellys 24 homogenizer (Bertin). RNA was extracted with the Direct-zol RNA Mini-prep kit
(Zymo Research) including on-column DNase treatment and transcribed into cDNA with the
qScript cDNA Synthesis Kit (Quantabio). RT-PCR was performed with Phire Tissue Direct
PCR Master Mix kit (Thermo Scientific) using primers 429-Gambicin-F & 430-Gambicin-R,
270-qCP-F1 & 271-qCP-R3, 484-qMag-both-F & 485-qMel-both-R, as well as 447-S7-F &
448-S7-R for the S7 reference gene. To quantify splicing efficiency, PCRs were performed on
above cDNAs with Q5 High-Fidelity DNA Polymerase (NEB) using 484-qMag-both-F as
forward primer and 242-Gam-locus-R (365bp amplicon) or 246-qCP-R2 (309bp) as reverse
primers, respectively. Annealing temperature, extension time and cycle number were set to
67°C, 5 seconds and 27 cycles, respectively. Amplicons were purified with QIAquick PCR
Purification Kit (Qiagen) and submitted to Amplicon-EZ NGS (Genewiz) and the data (NAR
accession PRJNA778891) analysed using Geneious Prime (Biomatters).
Mosquito infection assays
Transgenic or control mosquitos were infected with mature P. falciparum NF54 gametocyte
cultures (2-6% gametocytaemia) as described previously using the streamlined Standard
Membrane Feeding Assay (29) or with P. berghei ANKA 2.34 that constitutively expresses
GFP by direct feeding on infected mice. Engorged mosquitoes were provided 10% sucrose and
maintained at 27 °C for P. falciparum infections and 21 °C/RH 75% for P. berghei infections
until dissections were performed. Supplemental bloodmeals on human blood was provided via
membrane feeding. For infections of mosquitoes with mature Plasmodium falciparum
mosquitoes were starved without sugar for 48 hours after the infective or supplemental
bloodmeal to eliminate unfed individuals.
Analysis of parasite infection intensity and prevalence
We dissected midguts at the indicated days and microscopically examined them for the
presence of oocysts after staining with 0.1% mercurochrome. We measured the diameter of
oocyst using ImageJ (52). For measuring the prevalence and intensity of sporozoites, the head
and thorax as well as the midgut were dissected for each female. The gDNA was extracted
separately from head/thorax samples and the corresponding midgut samples with the DNeasy
96 Blood and Tissue Kit (Qiagen) and was used for qPCR 20µl reactions using the Qiagen
Quantinova SYBR Green PCR kit to quantify the P. falciparum Cytochrome B gene fragment
using primers and methods described previously (43, 53). Standard curves for the target gene
and the A. gambiae S7 reference gene were calculated after serial dilution of nucleic acid
templates. Ct-values were converted using their respective standard curves, and the target gene
Ct value was normalized to the reference gene (A. gambiae S7 ribosomal gene).
Generation of backcross populations and genotyping
50 homozygous MM-CP males and females were crossed to 50 Anopheles gambiae s.s. Ifakara
strain females and males respectively. F1 siblings were mass mated in a single cage to obtain
F2 progeny. F2 females were used for infection experiments or survival assays as described.
For genotyping of individual mosquitoes, we used we used multiplex genomic PCR with
primers CP-multi-F, CP-multi-R and Mag-R which results in an 356bp amplicon for the MM-
CP transgene and 670bp for the unmodified CP locus.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
13
Fitness and survival assays
For each replicate 20 females were individually transferred to cups one day after being offered
an uninfected bloodmeal. Spermatheca were then dissected from females that failed to oviposit
eggs to exclude unfertilized individuals. Eggs and larvae were counted on day 7 after blood-
feed. To determine the pupal sex ratio and pupation time, 100 L1 larvae per tray were reared
to the pupal stage where pupae were being collected and sexed once a day. Three biological
replicates were performed, and the data were analysed via the Chi-square test for deviations
from the expected sex ratio of 50%. The average pupation time in days was calculated and
tested for statistical significance with the Mann-Whitney test. For the survival analysis
including both sexes, a total of 274 WT and 276 MM-CPmale and 275 WT and 304 MM-CP
female pupae were placed in 6 separate cages with bottles containing 10% filtered fructose
solution and accumulated dead mosquitoes were counted daily. Survival was monitored daily
for 44 days on 3 independent replicates. For the survival analysis of backcrossed females F2
mosquitoes were placed in W24.5 x D24.5 x H24.5 cm cages as pupae. Mosquitoes
were offered a 10% sugar solution and they were also offered a blood meal and allowed to
deposit eggs 72 hrs after every bloodmeal. Dead mosquitoes were collected every 24 hrs from
the cage and preserved to be genotyped. Survival was monitored daily for 25 days on two
independent replicates.
RNAseq analysis
Females were fed with human blood and 15 guts were dissected into Trizol after 6 hours, 20
hours and from unfed females. After homogenization with 2.8mm ceramic beads (CK28R,
Precellys), RNA was extracted with the Direct-zol RNA Mini-prep kit (Zymo Research)
including on-column DNase treatment. Four biological replicates per condition were subjected
to RNA-seq. Libraries were prepared with the NEB Next® Ultra™ RNA Library Prep Kit and
sequenced on a NovaSeq 6000 Illumina platform (instrument HWI-ST1276) generating 150bp
paired-end reads (NAR accession tbc). Replicate 1 for MM-CP without bloodmeal
(MMCP_N_1) was identified as outlier with squared Pearson correlation coefficients with the
other three biological replicates below 0.84, and hence removed from further analysis.
Sequencing reads were mapped to the Anopheles gambiae PEST genome (AgamP4.13,
GCA_000005575.2 supplementred with the MM-CP construct reference) using HISAT2
software v2.0.5 (54) (with parameters --dta --phred33). Differential expression was assessed
with DESeq2 v1.20.0. and GO enrichment analysis was performed using TopGO (55) with a
pruning factor of 50 using a p-value cutoff of p=0.01.
Assessment of non-autonomous gene drive
At least 60 homozygous MM-CP or wild type females were crossed to males of the vasa-Cas9
strain. F1 progeny were screened for the presence of the 3xP3-YFP marker and
transhemizygotes were then sexed and crossed to wild types. Genomic DNA was isolated from
the progeny at the L2-L3 larval stage according to the protocol of the Phire Tissue Direct PCR
Kit (Thermo Scientific). Multiplex PCR was performed with primers 260-q-Mag-Mel-R, 531-
CP-multi-R and 532-CP-multi-F, yielding a 356bp band if the construct is present and a WT
band of 670bp as control. Two 96-well-plates per parent (paternal or maternal
transhemizygotes) and replicate were analysed and four negative controls were included on
each plate. From the control-crosses, 46 offspring per parent were analysed for each replicate.
The homing rate was calculated as (n*0.5 – Eneg) / (n*0.5) *100, with Eneg being the individuals
negative for the effector and n being the total number of samples successful analysed by PCR.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
14
Transmission modelling using EMOD
Simulations were performed using EMOD v2.20 (56), a mechanistic, agent-based model of
Plasmodium falciparum malaria transmission that include vector life cycle dynamics and
within host parasite and immune dynamics. Seasonality of rainfall and temperature as well as
vector species were kept the same across transmission settings, but vector density was varied
to match desired transmission intensity. A. gambiae, the only vector considered, was assumed
as being 95% endophilic and 65% anthropophilic. Each simulation contained 1000
representative people with birth and death rates appropriate to the demography without
considering importation of malaria. We include baseline health seeking for symptomatic cases
as an intervention where human agents can seek treatment with artemether-lumefantrine (AL)
80% of the time within 2 days of severe symptom onset and 50% of the time within 3 days of
the onset of a clinical but non-severe case. Mosquitoes within EMOD contain simulated
genomes that can model up to 10 genes with 8 alleles per gene with phenotypic traits that map
onto different genotypes (49). Here we modelled an integral gene drive system (57) aimed at
population replacement with an effector that results in delayed sporozoite production in
infected mosquitoes as well as an overall reduction in the number of sporozoites produced by
infected mosquitoes. The model was further parameterized using the experimentally
determined measures of fitness, lifespan and homing. 1000 male IGD mosquitoes homozygous
for the autonomous drive (Cas9) and non-autonomous effector (MM-CP) were released just
before the wet season begins to pick up in transmission intensity (Table S2). Apart from the
sex chromosomes, two loci representing the effector and driver were modelled (57), with each
locus having four possible alleles (wild type, resistant, effector or nuclease and loss of gene
function for the effector or driver locus). To evaluate the performance of these drives in a range
of transmission settings, we vary transmission intensity via annual entomological rates (EIR)
ranging from 15 infectious bites per person to 60 infectious bites per person. We also vary the
final phenotypic effect of expressing the effector gene that leads to delayed and reduced
sporozoite formation. We evaluate average increases in time until sporozoite formation ranging
from no increase in time compared to a wild-type mosquito up to 70% increase in sporozoite
formation time. As for the reduced sporozoite effect, we evaluated to full range of possible
effects compared to a wild-type mosquito. Mosquitoes carrying the drive are released 6 months
into the simulation, and simulations are run for a total of 6 years. The outputs represent the
mean of 25 stochastic realizations per parameter set.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
15
Fig. S1.
Alignment of sequenced cDNA amplicons to reference sequences representing the expected
splicing outcomes for Gam1-MM and MM-CP. The splice site is indicated by the black arrow,
the relative distribution of predicted variants (grey) representing at least 0.07% of all reads is
shown on the left and the corresponding number of reads on the right.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
16
Fig. S2.
Quantification of P. falciparum oocyst diameter in F2 individuals following a backcross of
strain MM-CP to the Ifakara strain and an F1 sibling intercross from 3 pooled biological
replicates. Non-transgenic (+/+), hemizygous (MM-CP/+) and homozygous (MM-CP/MM-
CP) individuals were identified by individual PCR genotyping after oocyst size had been
determined. Statistical analysis was performed by a t-test assuming unequal variance.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
17
Fig. S3.
Analysis of the time of pupation of MM-CP and wild-type mosquitoes as the percentage of
pupae emerging each day. Statistical analysis of average pupation times in males and females
was calculated using the Mann-Whitney test.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
18
Fig. S4.
The mean time course of allele frequencies after 1000 gene drives mosquitoes homozygous for
the driver and effector locus are released 6 months into a 6-year simulation. Each row
represents a different transmission intensity. The left column depicts allele frequency at the
driver locus and the right column represents allele frequencies at the effector locus. Shaded
regions represent one standard deviation on either side of the mean allele frequency as
calculated from 25 stochastic realizations of each scenario.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
19
Fig. S5.
The mean time course of true prevalence after 1000 gene drives mosquitoes homozygous for
the driver and effector locus are released 6 months into a 6-year simulation. Annual EIR of
around 15 (A), 30 (B) and 60 (C) infectious bites per person in an unmitigated scenario. Each
panel represents the average increase in time until sporozoites are released while the blue
shaded traces each represent a different average reduction in infectious sporozoites. The grey
trace represents the unmitigated baseline scenario.
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
20
gene counts
p value
GO.ID
Term
Condition
Annotated
Significant
Expected
elimKS
classicFisher
GO:0046914
transition metal
ion binding
6 hours
95
72
59.41
0.004
0.0031
GO:0046872
metal ion binding
6 hours
201
141
125.71
0.106
0.0086
GO:0015318
inorganic m.e.
transmembrane
transporter activity
20 hours
52
38
27.52
0.00019
0.0018
GO:0015075
ion
transmembrane
transporter activity
20 hours
59
41
31.23
0.00046
0.0057
GO:0098800
inner
mitochondrial
membrane protein
complex
20 hours
56
39
29.68
0.0021
0.0066
GO:0031966
mitochondrial
membrane
20 hours
78
52
41.34
0.4586
0.0073
GO:0005740
mitochondrial
envelope
20 hours
80
53
42.4
0.4701
0.0082
GO:0044455
mitochondrial
membrane part
20 hours
65
44
34.45
0.384
0.009
GO:0005743
mitochondrial
inner membrane
20 hours
72
48
38.16
0.435
0.0099
Table S1.
Enriched GO terms
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
21
Table S2.
Table of EMOD modelling parameters.
Supplementary file S1. (separate file)
Annotated DNA sequences of transformation vectors pD-MM-CP and pD-Gam1-MM.
Supplementary file S2. (separate file)
DNA oligonucleotide database.
Locus Allele combination Daily mortality (%) Fecundity (%) Applied to sex
driver WT , D -2.5 M,F
driver WT , LGF -10 M,F
driver D , D -5 M,F
driver D , R -2.5 M,F
driver D , LGF -12.5 M,F
driver R , LGF -10 M,F
driver LGF , LGF -100 M,F
effector WT , LGF +10 M,F
effector E , E +30 -14 F
effector E , LGF +10 M,F
effector R , LGF +10 M,F
effector LGF , LGF +100 M,F
Locus Allele combination Outcome Probability Allele generated
driver
D , WT Gene Drive (autonomous) 0.97 D
Unmodifed 0.1 WT
NHEJ to R2 0.1 LGF
NHEJ to R1 0.1 R
effector
E , WT (+D ) Gene Drive (non-autonomous) 0.97 E
Unmodifed 0.1 WT
NHEJ to R2 0.1 LGF
NHEJ to R1 0.1 R
Fitness & phenotypic parameters
Gene drive parameters
.CC-BY-NC 4.0 International licenseperpetuity. It is made available under a
preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in
The copyright holder for thisthis version posted February 17, 2022. ; https://doi.org/10.1101/2022.02.15.480588doi: bioRxiv preprint
