Gene Drive Strategies for Population Replacement

Abstract

Gene drive systems are selfish genetic elements capable of spreading into a population despite a fitness cost. A variety of these systems have been proposed for spreading disease-refractory genes into mosquito populations, thus reducing their ability to transmit diseases such as malaria and dengue fever to humans. Some have also been proposed for suppressing mosquito populations. We assess the alignment of these systems with design criteria for their safety and efficacy. Systems such as homing endonuclease genes, which manipulate inheritance through DNA cleavage and repair, are highly invasive and well-suited to population suppression efforts. Systems such as Medea, which use combinations of toxins and antidotes to favor their own inheritance, are highly stable and suitable for replacing mosquito populations with disease-refractory varieties. These systems offer much promise for future vector-borne disease control.

Full text

Chapter 9
Gene Drive Strategies for
Population Replacement
John M. Marshall
1
and Omar S. Akbari
2
1
Division of Biostatistics, School of Public Health, University of California, Berkeley, CA, USA,
2
Department of Entomology, University of California, Riverside, CA, USA
INTRODUCTION
After 3.8 billion years of research and development, Nature has provided
inspiration for a plethora of human design problems. During the
Renaissance, Leonardo da Vinci designed a flying machine inspired
by the anatomy of birds. Today, Nature’s evolutionary solutions are inform-
ing the design of solar panels from photosynthesis, and digital displays using
the light-refracting properties of butterfly wings. Nature’s intricate structures
and processes may also help in the fight against mosquito-borne diseases.
Gene drive—the process whereby natural mechanisms for spreading genes
into populations are used to drive desirable genes into populations (e.g.,
genes conferring refractoriness to malaria or dengue fever in mosquitoes)—
is another example of Nature’s processes being applied for the benefit of
humanity. Gene drive systems may either spread from low initial frequencies
or display threshold properties such that they are likely to spread if released
above a certain frequency in the population and are otherwise likely to
be eliminated.
Population replacement, in this context, refers to the process whereby a
population of disease-transmitting mosquitoes is replaced with a population
of disease-refractory ones. Several approaches are being explored to engineer
mosquitoes unable to transmit human diseases, and there have been a number
of notable successes. For example, Isaacs et al. have engineered Anopheles
stephensi mosquitoes expressing single-chain antibodies that prevent
Plasmodium falciparum malaria parasites from developing in the mosquito,
thus preventing onward transmission of the parasite [1]. Gene drive systems
are expected to be instrumental in spreading disease-refractory genes into
wild mosquito populations, given the wide geographical areas that these spe-
cies inhabit and the expectation that refractory genes will be associated with
169
Genetic Control of Malaria and Dengue.
©2016 Elsevier Inc. All rights reserved.

at least modest fitness costs [2]. Gene drive systems are also being
considered to implement population suppression strategies whereby genes
conferring a fitness load or gender bias are instead driven into the vector
population, thereby reducing disease transmission.
Early Inspiration
Initial suggestions for spreading desirable genes into insect pest populations
date back to the early 1940s and involved the proposition of translocations
[3,4] and transposable elements (TEs) [5], inspired from natural systems.
Translocations are rearrangements of parts between nonhomologous chromo-
somes. If insects homozygous for a translocation are introduced into a
population at high frequency, they are predicted to spread to fixation [6], and
if the translocation is linked to a disease-refractory gene, it is predicted to
consequently be driven into the population as well. Initial field trials with
translocations were unsuccessful in demonstrating spread [7]; but this is
likely a result of those translocations being generated using X-rays, which
often induce high fitness costs.
The suggestion of using TEs to drive disease-refractory genes into mos-
quito populations was largely inspired by the observation that a TE known
as the Pelement spread through most of the global Drosophila melanogaster
population within the span of a few decades following natural acquisition
from Drosophila willistoni [8]. TEs are able to spread through a population
due to mechanisms that enable them to increase their copy number within a
host genome and hence to be inherited more frequently in subsequent
generations. As a result, they are able to spread into a population from very
low initial frequencies even if they incur a fitness cost to their host [9]. It
was hoped that the Pelement invasion of Drosophila could be repeated in
disease-transmitting mosquito species using a TE attached to a disease-
refractory gene; however, early laboratory work on TEs in mosquito vector
species has failed to identify elements with high remobilization rates
following integration into mosquito lines [10].
Promising New Systems
Two of the most promising gene drive systems at present also involve technol-
ogies inspired by Nature—the use of homing endonuclease genes (HEGs)
observed to spread in fungi, plants, and bacteria [11], and a selfish genetic ele-
ment known as Medea observed to spread in Tribolium beetles [12,13].
A synthetic Medea element has been developed in Drosophila that works by
the hypothesis that Medea encodes both a maternally expressed toxin and a
zygotically expressed antidote [14]. This combination results in the death of
wild-type offspring of Medea-bearing mothers, thus favoring the Medea allele
in subsequent generations and mimicking the behavior of the natural element
170 Genetic Control of Malaria and Dengue

in Tribolium.Medea was the first synthetic gene drive system to be developed
and has a number of desirable design features; however, significant work is
still ongoing to develop a Medea element in a mosquito disease vector.
Recently, there has been much excitement around HEGs as, while Medea
was first engineered in Drosophila, a naturally occurring HEG has been
shown to spread in a laboratory population of Anopheles gambiae, the main
African malaria vector, containing an engineered target sequence for the
HEG [15]. HEGs spread by expressing an endonuclease that creates a
double-stranded break at specific target sequences lacking the HEG.
Homologous DNA repair then copies the HEG to the cut chromosome,
increasing its representation in subsequent generations. Similar to the
aforementioned gene drive systems, HEGs are being considered to drive
disease-refractory genes into mosquito populations; however, a number of
additional strategies for their application are also being considered, which
aim to suppress rather than replace mosquito populations [11], and progress
has been made toward these ends as well [16].
Design Criteria
As the technology for developing gene drive systems for population replace-
ment develops on a number of fronts, it is useful to consider design criteria
for assessing the safety and efficacy of the various approaches. An excellent
review by Braig and Yan [17] proposes several biological properties that an
ideal gene drive system should or must have:
1. The gene drive system must be effective. That is, it must be strong
enough to compensate for any loss in host fitness due to the presence of
both itself and its transgenic load (manifest as a reduction in host fertil-
ity, life span, or competitiveness). It must be able to spread to very high
frequency in a population on a timescale relevant to disease control
(i.e., a few years) and must be unimpeded by wild-type vectors immi-
grating into the target area.
2. The gene drive system must be able to carry with it several large genes
and associated regulatory elements. At the very least, a disease-
refractory and marker gene will be needed along with regulatory ele-
ments; but multiple disease-refractory genes are preferable in order to
slow the rate at which the pathogen evolves resistance to each of them.
3. Features should be included to minimize the rate at which linkage is
lost between the drive system and disease-refractory genes, as even rare
recombination events could be significant for wide-scale spread over a
long time period.
4. It should be possible to use the gene drive system to introduce waves of
refractory genes over time to counteract the effects of evolution
Gene Drive Strategies for Population Replacement Chapter | 9 171

of pathogen resistance, mutational inactivation of the refractory gene, or
loss of linkage between the refractory gene and drive system.
5. The gene drive system should be easily adapted to multiple vector
species. Human malaria, for instance, is transmitted by approximately
50 species of mosquitoes belonging to the genus Anopheles. In sub-
Saharan Africa, the most important transmitters are An. gambiae,
Anopheles coluzzii,Anopheles arabiensis, and Anopheles funestus,
ideally all of which should be rendered refractory in a population
replacement strategy.
Additional features of an ideal gene drive system were proposed by
James to address ecological, epidemiological, and social issues, including
safety [2]. Safety is a broad criterion that should be assessed through risk
assessment in which potential hazards are identified along with their corre-
sponding magnitudes and likelihoods. This provides a framework for man-
aging the most significant risks and for the overall safety of the system to
be scored. However, prior to a comprehensive risk assessment, a few gen-
eral safety criteria for gene drive systems can be imagined.
6. The behavior of the gene drive system in the target species should
be stable and predictable, thus minimizing the likelihood of
unpredictable side effects in target species.
7. A mechanism should be available to prevent horizontal transfer of the
gene drive system and/or refractory gene to nontarget species, thus min-
imizing the wider ecological impact of the release.
8. The gene drive system and refractory gene should not cause undesirable
effects for human health, for instance, by selecting for increased viru-
lence in the pathogen population. The gene drive system should also
include a mechanism for removing the refractory gene from the popula-
tion in the event of any adverse effect.
9. The gene drive system must be consistent with the social and regulatory
requirements of the affected communities. For instance, public attitude
surveys in Mali [18] highlight the importance of confined field trials
prior to a wide-scale release, which could be achieved through the ini-
tial use of gene drive systems with high release thresholds followed by
subsequent releases with more invasive systems.
10. The gene drive system should be cost-effective, as budgets for disease
control are limited and a number of alternative interventions are avail-
able. The initial development of gene drive systems is expensive; but
ongoing investment can be minimized by designing systems that are
resilient to evolutionary degradation.
Cost-effectiveness is an important consideration, as it is not only relevant to
the choice of gene drive system, but to whether gene drive should be used at all.
In a recent modeling study, Okamoto et al. demonstrated the economic feasibil-
ity of releasing large numbers of insects carrying a dengue-refractory gene
172 Genetic Control of Malaria and Dengue

without a gene drive system in order to reduce the dengue transmission potential
of Aedes aegypti mosquitoes in Iquitos, Peru [19]. Wide-scale control of
Anopheles malaria vectors in sub-Saharan Africa is less likely amenable to the
mass release strategy; however, it is essential to assess this in terms of efficacy,
safety, and cost-effectiveness prior to implementation.
In this chapter, we review a range of gene drive systems being considered
to drive disease-refractory genes into mosquito vector populations. We
divide gene drive systems into two broad categories: (i) those that spread by
causing a double-stranded break at a specific target sequence and insert
themselves at this location through DNA repair (e.g., HEGs) and (ii) those
that use combinations of toxins and antidotes, active at different life stages,
to favor their own inheritance (e.g., Medea). We also review modern
approaches to developing translocations as form of gene drive, which do not
fit into either category. Systems using symbiotic or commensal microorgan-
isms to mediate gene drive are covered in another chapter (e.g., Wolbachia).
For each system, we review the biological mechanisms involved, the
system’s current stage of development, and its alignment with the abovemen-
tioned design criteria.
GENE DRIVE SYSTEMS THAT SPREAD VIA TARGET SITE
CLEAVAGE AND REPAIR
We begin by reviewing gene drive systems that manipulate inheritance in
their favor by causing a double-stranded break at one or more specific target
sites in the host’s genome and utilize the host’s homologous DNA repair
mechanism to increase their genomic copy number. Gene drive systems of
this type include TEs, HEGs, and a number of recently proposed HEG ana-
logs, such as zinc-finger nucleases (ZFNs), transcription-activator-like effec-
tor nucleases (TALENs), and clustered, regularly interspaced, short
palindromic repeats (CRISPRs).
Transposable Elements
TEs are genomic components capable of changing their position and some-
times replicating within a genome. Consequently, they show widespread
prevalence throughout the genomes of many taxa, with various families of
TEs accounting for B90% of the Salamander genome, 50% of the Ae. aegypti
genome, and 45% of the human genome. There are various classes of TEs,
and those being considered for population replacement in mosquitoes belong
to class 2. Class 2 elements contain both repeat sequences that mark their
boundaries and their own transposase gene that catalyzes transposition. They
move via a cut-and-paste mechanism [20], whereby transposition results in
excision of the TE via two double-stranded breaks, leaving behind a gap
where they have been excised. In some cases, this gap is filled by
Gene Drive Strategies for Population Replacement Chapter | 9 173

homologous gap repair from a chromatid also having the TE. The excised
TE is then inserted at another genomic location, resulting in their genomic
copy number being increased by one. In a second replication mechanism,
some TEs transpose during the S phase of the cell cycle. If a recently repli-
cated element transposes to an unreplicated region of the genome, it will be
replicated a second time, resulting in a net gain of one element in the genome.
Current Status. The widespread distribution of TEs in Nature together with
observations of the rapid spread of the Pelement in Drosophila [8] inspired
initial hopes that class 2 TEs could be inserted, along with disease-refractory
and marker genes, into transgenic lines of Ae. aegypti (the main vector of den-
gue fever) and Anopheles vectors of malaria. Class 2 TEs lacking their transpo-
sase gene are often used as vectors for introducing novel genes into
mosquitoes; hence, integration into mosquito lines is relatively straightforward.
More problematic, however, has been the remobilization of TEs containing
their own transposase gene once they have been integrated. An excellent
review by O’Brochta et al. describes results from experiments in which four
class 2 TEs—Hermes,Mos1,Minos, and piggyback—were used to create
transgenic lines of Ae. aegypti [10]. In all cases, remobilization was shown to
be highly inefficient. More recently, attempts were made to improve the post-
integration mobility of Hermes in Ae. aegypti using an additional construct to
express a transposase gene under the control of a testis-specific promoter [21];
however, remobilization was still only observed in less than 1% of the
transgenic lines.
Design Criteria. The observed remobilization of natural TEs suggests that
remobilization of introduced elements should also be possible; however, the
regulation of TE mobility is complex, and it may require much experimenta-
tion to find TEs compatible with mosquito vectors. This work is likely not
cost-effective, as TEs fail to satisfy most of the design criteria outlined earlier,
and have been superseded by more recently proposed systems like HEGs and
Medea. Of particular note, it is unlikely that TEs will be able to carry large
inserts containing disease-refractory genes as transposition events are known
to be imprecise and prone to DNA loss. Furthermore, a study on the Himar1
mariner element suggests that transposition rates decline substantially with
increasing insert size [22], suggesting that elements which have lost their
transgenic load will outspread those which have not [23]. Finally, the large
numbers of target sites that TEs have undermine their predictability and stabil-
ity in target species, and their wide species host range highlights the risk of
horizontal gene transfer and spread in nontarget species.
Homing Endonuclease Genes
HEGs are highly efficient selfish genetic elements that spread by expressing
an endonuclease that recognizes and cleaves a highly specific target
sequence of 14 40 base pairs usually only present at a single site in the host
174 Genetic Control of Malaria and Dengue

genome [24]. As the HEG is positioned directly opposite its target site, actu-
ally within its own recognition sequence, it induces a double-stranded break
only in chromosomes lacking the HEG. The HEG is effectively copied to the
target site, in a process referred to as “homing,” when the cell’s repair
machinery uses the HEG-bearing chromosome as a template for homology-
directed repair. When homing occurs in the germ line of the host organism, a
HEG can be transmitted to progeny at a higher than Mendelian inheritance
ratios, enabling its spread through a population (Figure 9.1A).
On the basis of observations of homing activity in a number of nonmetazo-
an organisms including yeast, fungi, algae, and plants, Burt proposed that
HEGs could be used as a gene drive system for population replacement in
mosquito disease vectors; however, he also proposed and favored their use as a
population suppression system [11]. Burt proposed a suite of HEG-based
FIGURE 9.1 Preferential inheritance of homing-based gene drive systems. (A) Left panel:
A homing HEG (green) encodes an endonuclease that recognizes and cleaves a specific target
sequence (red) on a wild-type chromosome (step 1). Once the target site is cleaved, the cell
repairs the chromosomal break through homologous recombination using the HEG-bearing chro-
mosome as a template (step 2). This two-step process results in the HEG effectively being cop-
ied to the wild-type chromosome in a process referred to as “homing,” thereby converting a
HEG heterozygote into a HEG homozygote. Right panel: When a HEG-bearing male (green
mosquito) is released into the wild and mates with a wild-type female (gray mosquito), the
majority of their progeny inherit the HEG, and over time the HEG invades entire populations.
(B) Left panel: For HEG-based population suppression, an X-shredder HEG is positioned on the
Y chromosome (Y-linked X-shredder HEG). This HEG encodes an endonuclease that recognizes
and cleaves chromosomal sequences that are repeated exclusively on the X-chromosome of the
mosquito. When expressed during spermatogenesis, X-bearing spermatids are disrupted by the
HEG, resulting in the majority of functional sperm being Y-bearing and containing the HEG.
Right panel: When a Y-linked X-shredder HEG-bearing male (green mosquito) is released into
the wild and mates with a wild-type female (gray mosquito), all resulting progenies are HEG-
bearing males. Over time, this is predicted to induce an all-male population crash and potentially
eventual extinction of the vector species.
Gene Drive Strategies for Population Replacement Chapter | 9 175

strategies for genetic control of mosquito vectors—two involving population
replacement and three involving population suppression:
1. First, the HEG could be linked to a disease-refractory gene and engi-
neered to target a gene-sparse region of a chromosome (so as to reduce
impacts on mosquito host fitness), thus carrying the disease-refractory
gene with it as it spreads into the population.
2. In a related population replacement approach, the HEG could be engi-
neered to target an endogenous gene involved in the development or
transmission of the pathogen, thus reducing vector competence as it
spreads [25]. This approach has the benefit that it does not involve an
effector gene and hence is more resilient to evolutionary degradation;
however, it does require a gene to be identified, the disruption of which
would block pathogen transmission, and for a HEG to be engineered to
target this, which is quite arduous.
3. In terms of population suppression, a HEG could be engineered that tar-
gets a native mosquito gene required in at least one copy for either mos-
quito survival or fertility. If a HEG of this type is active in the mosquito
germ line, then it will increase in frequency in the population, inducing a
genetic fitness load on the population as it spreads. This could lead to
either population suppression or an eventual population crash.
4. An alternative to the homing-based applications of HEGs is to rely
entirely on their target site cleavage activity. In the first of these
approaches, known as the “autosomal X-shredder” strategy, a HEG can
be designed to specifically cleave the X chromosome at multiple loca-
tions, effectively destroying it. If an X-shredder HEG is expressed during
male meiosis, it will result in destruction of X-bearing male sperm. If
females mate with males having the X-shredder, most viable sperm will
be Y-bearing and hence most of the progeny will be male. This strategy
will reduce the reproductive potential of the population; but it requires
regular releases since the X-shredding gene is associated with a fitness
cost and will only persist in the population for a few generations.
5. Finally, Burt proposed a “Y-linked X-shredder” strategy whereby, if the
X-shredder HEG is located on the Y chromosome, then it will be driven
into the population along with the transgenic Y chromosome as it induces
an increasingly male gender bias. This approach would mimic naturally
existing meiotic drive systems that bias sex ratios, although it could
potentially induce a much larger gender bias than those observed in
Nature [26 28], causing a cascade of male-only population crashes that
could potentially lead to species extinction (Figure 9.1B).
Current Status. An encouraging result for homing-based HEG strategies
has been the engineering of a naturally occurring HEG, I-SceI, which has
been shown to cleave in Ae. aegypti [29] and spread in laboratory popula-
tions of both D. melanogaster and An. gambiae containing an engineered
176 Genetic Control of Malaria and Dengue

target sequence for the HEG [15,30,31]. These results are encouraging
because they show that, although HEGs have not been discovered in any
metazoan species to date, there is nothing intrinsic about metazoan biology
that prevents HEGs from homing. Furthermore, the fact that this was
achieved in An. gambiae, the most important African malaria vector, is
hopeful for its application to disease control. For the population replacement
strategy to work in the wild, a HEG must be engineered or identified which
has a target sequence in the wild mosquito genome. Engineering HEGs to
recognize and cleave new target sequences has proven difficult thus far
[32 34], and future research should focus on the development of novel
approaches to circumvent these difficulties.
Population suppression strategies that rely solely on the target site cleave
activity of HEGs have shown remarkable progress in recent years. A HEG
originally discovered in the slime mold Physarum polycephalum, I-Ppo1
[35], was integrated into the An. gambiae genome and shown to recognize
and cleave a conserved DNA sequence, repeated hundreds of times and
located exclusively on the X chromosome cluster of ribosomal DNA genes
in An. gambiae [36]. This cleavage activity is highly applicable to both the
autosomal and Y-linked X-shredder strategies of HEG-driven population sup-
pression and has also provided a novel genetic approach to the sterile insect
technique for An. gambiae. The expression of I-Ppo1 during spermatogenesis
in An. gambiae resulted in cleavage of the paternal X chromosome in differ-
entiating spermatozoa, which was expected to result in a male bias among
progeny. However, it turned out that the I-Ppo1 from mature sperm cells
was carried over into the zygote, thus shredding the zygotic X chromosomes
as well and rendering the transgenic males completely sterile [37]. It was
later shown that transgenic mosquitoes engineered with I-Ppo1 could induce
high levels on sterility in large cage populations, confirming the suitability
of this technology for use in sterile insect population suppression programs
[38]. This could be a useful first application of HEG technology in the wild
given the self-limiting nature of sterile insect releases.
For X-shredder strategies to work, I-Ppo1 would need to be destabilized
in order to minimize its carryover into the zygote by mature sperm. To this
end, recent work by Galizi et al. has succeeded in expressing destabilized
autosomal versions of I-Ppo1, which result in efficient shredding of the
paternal X chromosome and are restricted to male meiosis [16].
Consequently, males carrying this construct are fully fertile and some inser-
tions produce .95% male offspring bias. Males inheriting the autosomal
I-Ppo1 gene also produce a male bias in their progeny, showing that the
gender-biasing effect of autosomal X-shredders will remain in the population
for several generations; however, continued releases would be required, as
the X-shredder gene is not favored through inheritance when located on an
autosome and is expected to be eliminated due to fitness costs. Nevertheless,
for repeated releases, population suppression is expected, which would be
Gene Drive Strategies for Population Replacement Chapter | 9 177

more efficient than the previously mentioned sterile male releases and would
also be self-limiting, albeit over a longer period. Autosomal X-shredders
could therefore be an appropriate second application of HEG technology.
The only remaining steps in order to realize the Y-linked X-shredder
strategy are to dock the destabilized I-Ppo1 HEG onto the An. gambiae Y
chromosome and ensure that it is expressed during spermatogenesis. To this
end, recent progress has been made in developing a Y chromosome docking
line in An. gambiae [39]. Future work will focus on docking the HEG onto the
Y chromosome and ensuring it can be expressed and function as anticipated.
Design Criteria. HEG-based strategies for genetic control of vector-borne
diseases are extremely promising given the remarkable progress made
recently, most notably in the malaria vector An. gambiae. HEGs are highly
effective as a gene drive system, capable of spreading for low initial frequen-
cies to high frequency on a short timescale. They are also relatively short
sequences targeting very precise regions of the genome, suggesting both
stability and a low rate of corruption due to evolutionary degradation.
Species-specific regulatory sequences can be included to limit their horizontal
transfer to nontarget species, and furthermore, a strategy has been proposed to
reverse the spread of a deleterious HEG through the release of HEG-resistant
alleles in the event of unforeseen consequences [11]. Additionally, a wide
range of HEG strategies are available displaying different levels of confine-
ability, allowing them to be used at all stages of a phased release and to be
tailored to the social and regulatory requirements of affected communities.
Target site cleavage strategies show more promise than those reliant on
homing activity as they sidestep many of the abovementioned design criteria
and are independent of disease-refractory genes. Target site mutagenesis and
gap repair through nonhomologous end joining can both result in disruption
of the HEG cleavage site, rendering certain individuals immune to the HEG
and preventing the HEG from spreading through an entire population. For
strategies in which a HEG disrupts a gene required for mosquito survival or
fertility, HEG-resistant mutants will be favored in a population once they
emerge. Furthermore, there is a possibility of losing the disease-refractory
gene either through mutagenesis or during homology-directed repair—a con-
cern that becomes more serious for larger inserts, and would render a popula-
tion replacement strategy futile. The Y-linked X-shredder strategy is less
vulnerable to target site mutagenesis as it targets so many loci on the X chro-
mosome at once. It is, however, dependent on germ line gene expression on
the An. gambiae Y chromosome although this could potentially be achieved
through the use of insulator sequences.
TALENs and ZFNs
TALENs and ZFNs have been proposed as alternative platforms for engi-
neering homing-based gene drive systems [40]—that is, systems that spread
178 Genetic Control of Malaria and Dengue

by cleaving a specific target sequence and then using the cell’s repair
machinery to copy themselves to the target site. The benefit of TALENs and
ZFNs over HEGs is that they can be easily engineered to target desired DNA
sequences due to the modular nature of their DNA-binding domains.
TALENs are derived from naturally occurring proteins that are secreted by
the pathogenic bacteria Xanthomonas spp. to alter gene expression in host
plant cells [41,42]. These proteins contain arrays of highly conserved, repeti-
tive DNA-binding domains, each recognizing only a single base pair, with
specificity being determined by repeat-variable di-residues [43,44]. The rela-
tionship between these repeats and DNA recognition can be exploited to
design TALENs that target virtually any desired DNA sequence. For ZFNs,
DNA-binding specificity can be similarly manipulated, being determined by
an array of finger modules that can be generated either by selection using
large combinatorial libraries, or by rational design [45].
For both TALENs and ZFNs, DNA-binding modules can be combined
with several types of domains, including transcriptional activators, nucleases,
and recombinases, allowing for a comprehensive range of genetic modifica-
tions [46]. In terms of cleavage activity, a wide range of tailored recognition
sequences can be cleaved efficiently as TALENs and ZFNs are fusion pro-
teins consisting of a nonspecific fok1 nuclease linked to a DNA-binding
motif [47,48]. The TALEN or ZFN may then be copied to the cleaved target
side by homology-directed repair, and hence used as a gene drive system for
driving disease-refractory genes into mosquito populations.
Current Status and Design Criteria. Both TALENs and ZFNs rely upon
homing activity and thus, for the purposes of population replacement and
control, are functionally similar to HEGs. Given this similarity, the range of
replacement and suppression strategies outlined earlier is also applicable to
these systems and many of the design issues are similar too. For example,
TALENs and ZFNs are also expected to spread from low initial frequencies,
species-specificity can be incorporated through the addition of regulatory
elements, and a deleterious TALEN or ZFN can be removed from a popula-
tion through the release of TALEN- or ZFN-resistant alleles. However, there
are some important differences. In terms of cost-efficiency, both TALENs
and ZFNs are easier to engineer to target specific DNA sequences, and con-
sequently, they could be straightforwardly adapted to multiple vector species,
which is particularly important for malaria control. However, concerns arise
regarding their stability, as their repetitive nature makes them more prone to
mutation and evolutionary degradation. Recent progress toward developing
both TALEN- and ZFN-based gene drive systems in D. melanogaster have
successfully demonstrated DNA-binding specificity, cleavage, and homing
through homology-directed DNA repair; however, mutational inactivation
led to a decline in effectiveness over just a short period of time [40]. Thus, if
TALENs or ZFNs are to be useful as gene drive systems in the future, their
stability issues must first be overcome.
Gene Drive Strategies for Population Replacement Chapter | 9 179

Clustered, Regularly Interspaced, Short Palindromic Repeats
CRISPR is another promising system proposed, although not yet demon-
strated, as an alternative platform for homing-based gene drive. The system is
based on an adaptive immune process in bacteria whereby sequences derived
from invading bacteriophages or plasmids are integrated into the bacterial
CRISPR locus. This essentially provides bacterial cells with the ability to
“remember” and protect themselves against previously encountered viral gen-
omes and invasive, mobile genetic elements [49]. To perform nuclease activi-
ties, CRISPR systems use an array of CRISPR RNAs (crRNAs) derived from
exogenous DNA targets (e.g., viral genomes), noncoding transactivating
RNAs, and a cluster of CRISPR-associated (Cas) genes. Three types of
CRISPR systems have been discovered, with type II CRISPR systems being
best characterized. These consist of a Cas9 nuclease and a crRNA array
encoding guide RNAs and auxiliary transactivating crRNAs to mediate target
site cleavage [50]. As for the homing-based systems described earlier, if the
double-stranded break is repaired by homology-directed repair, the CRISPR
system may be copied to the cleaved target site and hence used as a gene drive
system for population replacement similar to HEGs. If the target site cleavage
activity is directed toward the X chromosome, then the population suppression
strategies initially described for HEGs could also be realized.
Current Status. Recent encouragement for CRISPR-based gene drive has
been provided by proof-of-principle studies showing that the type II CRISPR
system from Streptococcus pyogenes can be modified to target endogenous
genes in bacteria [51] and human cell lines [52,53]. It has subsequently been
shown that CRISPR can be used to alter genes in a range of other species
including insects such as D. melanogaster [54,55] and mosquitoes.
Straightforwardly, utilizing this system in other organisms requires only two
components—the Cas9 nuclease and guide RNAs [52,56]. DNA-binding
specificity is determined by the first 20 nucleotides of the guide RNA as
these designate the DNA target side that Cas9 will be guided to according to
Watson Crick DNA RNA base pairing rules. The only restriction for the
target site selection is that it must lie directly upstream of a protospacer adja-
cent motif sequence that matches the canonical form 50-NGG. Aside from
that, it is possible that the CRISPR system can be engineered to target and
cleave essentially any genomic location, with subsequent homing and gene
drive occurring via homology-directed repair, however this remains to be
demonstrated.
Design Criteria. CRISPR-based gene drive has yet to be implemented;
however, its mechanisms imply that the approach is achievable. In terms of
design criteria, the system is very similar to TALENs and ZFNs—it is
expected to spread from low initial frequencies, species-specificity can be
incorporated through regulatory elements, and a deleterious CRISPR can
be removed through release of CRISPR-resistant alleles. The system is active
180 Genetic Control of Malaria and Dengue

in a range of species and target sites are even easier to engineer than for
TALENs, suggesting the system would be easily adapted to multiple vector
species. Another advantage of the CRISPR system is that it can be used to
target multiple sequences in a single experiment [57], increasing its potential
efficacy and decreasing the rate at which target site mutagenesis could slow
its spread. A major concern, however, is that the CRISPR system itself may
be degraded. The CRISPR system is quite large, consisting of promoters, the
Cas9 gene, guide RNAs and, depending on the strategy being implemented,
multiple disease-refractory genes and associated regulatory elements.
A system this size is prone to mutation and errors introduced during homing,
including potential loss of function of disease-refractory genes. These con-
siderations may lead to population suppression strategies being favored for
CRISPR-based drive systems; however, this would place selection pressure
on mutant CRISPR alleles having lost their function and so the evolutionary
stability of the CRISPR system will need to be explored and optimized if it
is to provide a cost-effective alternative to the relatively stable yet difficult-
to-engineer X-shredding HEGs.
TOXIN ANTIDOTE GENE DRIVE SYSTEMS
We now move on to gene drive systems that use combinations of toxins
and antidotes, active at different life stages, to favor their own inheritance
[58]. Gene drive systems of this type include Medea, engineered forms of
underdominance such as UD
MEL
, self-limiting systems such as killer-
rescue, and other toxin antidote possibilities such as Semele,Medusa, and
inverse Medea.
Medea
The story of Medea has origins in both Greek mythology and beetle biology.
In Greek mythology, Medea was the wife of the hero Jason, to whom she
had two children. Her marriage to Jason was hard-earned, transpiring only
after she enabled him to plough a field with fire-breathing oxen, among other
achievements; but despite this, he left her when the king of Corinth offered
him his daughter. As a form of revenge, Medea killed their two children.
From a biological perspective, such infanticide would make Medea an unfit
mother; but if the trait is genetic and children that inherit it also have the
ability to defend themselves, then mathematical models show that it actually
has a selective advantage and, if present at modest levels in a population, is
expected to become present among all individuals within a matter of genera-
tions [59,60]. This is simply because children who are able to defend them-
selves against a murderous parent are more fit than those who cannot.
The Greek analogy sounds bizarre; but genes displaying these properties
do actually exist in Nature and have been discovered and characterized in
Gene Drive Strategies for Population Replacement Chapter | 9 181

various regions of the world [12,61,62]. The first such element to be identi-
fied was in the flour beetle Tribolium castaneum [12] and was given the
name Medea after both the character from Greek mythology, and as an
acronym for “maternal-effect dominant embryonic arrest.” By crossing
individuals from geographically isolated locations, it was found that Medea-
bearing males gave rise to both wild-type and Medea-bearing offspring; but
that Medea-bearing females only gave rise to Medea-bearing offspring. It
appeared that Medea-bearing mothers were selectively killing non-Medea-
bearing offspring; or alternatively that they were trying to kill all offspring
and the Medea-bearing offspring were able to defend themselves.
The genetic factors involved in this behavior remain obscure; but the
dynamics suggest a model in which Medea consists of two tightly linked
genes—a maternally expressed toxin gene, the product of which causes all
eggs to become unviable and a zygotically expressed antidote gene, the prod-
uct of which rescues Medea-bearing eggs from the effects of the toxin
[12,63]. In Tribolium,Medea dynamics are attributed to an insertion of a
composite Tc1 transposon inserted between two genes both having maternal
and zygotic components [13]. Remarkably, this system was reverse-
engineered using entirely synthetic components in laboratory populations of
D. melanogaster and was shown to rapidly drive population replacement
[14,64]. These synthetic elements were constructed using two unique, tightly
linked components—a maternal toxin consisting of maternally deposited
microRNA designed to target an essential embryonic gene; and a zygotic
antidote consisting of a tightly linked, zygotically expressed, microRNA-
resistant version of the embryonic essential gene. The combination of these
components results in the death of wild-type offspring of Medea-bearing
mothers, thus favoring the Medea allele in subsequent generations and mim-
icking the behavior of the natural element in Tribolium (Figure 9.2A).
Current Status.Medea was the first synthetic gene drive system to be
developed, in this case in D. melanogaster [14]. Given that the synthetic
Medea elements were constructed using rationally designed synthetic compo-
nents and well-understood, conserved molecular and genetic mechanisms, it
should be possible to engineer Medea elements in a range of other insects
including mosquitoes. The Medea drive strategy is particularly well-suited to
driving disease-refractory genes into mosquito populations, and hence the
development of several efficient refractory genes for each disease of interest
is encouraged.
Design Criteria. In many ways, Medea is the ideal system for replace-
ment of wild mosquito populations with disease-refractory varieties.
Solutions are available for all of the design criteria outlined earlier, and
Medea has an advantage over homing-based strategies for population
replacement since it is stably integrated into the host chromosome, thus not
affected by the substantial risk of loss during homology-directed repair. If
introduced at modest population frequencies, Medea can spread and rapidly
182 Genetic Control of Malaria and Dengue

FIGURE 9.2 Dynamics of toxin antidote-based gene drive systems. (A) Medea elements dis-
tort the offspring ratio in their favor through the action of a maternally expressed toxin (MT)
and a zygotically expressed antidote (ZA). This results in the death of wild-type offspring of
heterozygous mothers and enables the Medea element to spread into a population from very low
initial frequencies. Dynamics here are shown for a Medea element with no fitness cost, released
at 10% in the population. Transgenic frequency refers to any individual carrying at
least one copy of the element. (B) UD
MEL
(maternal-effect lethal underdominance) is a
toxin antidote-based underdominant system consisting of two constructs, each of which
possesses a maternally expressed toxin (MT1 and MT2) whose activity is manifest during prog-
eny embryogenesis and a zygotic antidote (ZA1 and ZA2) capable of neutralizing the maternal
toxin expressed by the opposite construct. This results in heterozygous females being sterile if
mated to wild-type individuals, thus leading to the characteristic bistable dynamics of underdo-
minant systems. Dynamics here are shown for UD
MEL
constructs at independently assorting loci
having no fitness costs. If released at a population frequency of 20%, the system spreads to fixa-
tion in the population; but if released at 15%, the system is eliminated. (C) Semele elements dis-
tort the offspring ratio in their favor through the action of a semen-based toxin (SBT) and a
female-specific antidote (FA). This results in unviable crosses between transgenic males and
wild-type females and favors transgenic individuals provided the Semele element is present at
population frequencies exceeding B36% (above this frequency, the selective advantage of the
antidote exceeds the selective disadvantage of the toxin). Dynamics here are shown for a Semele
element with no fitness cost. If released at a population frequency of 40%, the element spreads
to fixation in the population; but if released at 30%, the system is eliminated. (D) Medusa is a
two-construct, sex chromosome-linked drive system capable of inducing confineable and revers-
ible population suppression. The system consists of four components—a maternally expressed,
X-linked toxin (MT1) causes suppression of the female population and selects for the transgene-
bearing Y since only transgenic male offspring have the corresponding Y-linked zygotically
expressed antidote (ZA1). A zygotically expressed, Y-linked toxin (ZT2) and a zygotically
expressed, X-linked antidote (ZA2) then selects for the transgene-bearing X when the transgene-
bearing Y is present, creating a balanced lethal system. When present above a certain threshold
frequency, Medusa spreads while creating a strong male gender bias leading to population sup-
pression. Dynamics here are shown for Medusa constructs having no fitness costs. For two con-
secutive male-only releases at a population frequency of 50%, the population becomes entirely
male as the system spreads to fixation in the population; but for two consecutive male-only
releases at a population frequency of 40%, the system is eliminated.

replace a population, even in the presence of modest fitness costs [60]; how-
ever, Medea is unlikely to spread following a small-scale accidental release
because its driving ability is low at low population frequencies [18].
Tight linkage between the toxin, antidote, and refractory genes by placing
the toxin and refractory genes within an intron of the antidote gene can
improve system stability and reduce the rate of loss of the refractory gene
through recombination. However, in the event that the Medea element or
refractory gene become unlinked, mutated, or rendered ineffective through
parasite evolution, second-generation Medea elements can be generated that
utilize toxin antidote combinations distinct from those of the first-
generation elements [14], making it possible to carry out multiple cycles of
population replacement. This strategy can also be used to remove refractory
genes from populations in the event of adverse effects. As the functional
components of Medea are developed in mosquito species, it will become
more cost-efficient to develop these elements and to adapt them to multiple
vector species.
Toxin Antidote-Based Underdominance
Underdominant systems display the property that heterozygotes, or their
progeny, have lower fitness than either homozygote [65]. In the simplest
case of a single biallelic locus for which matings between opposite homozy-
gotes are sterile, whichever allele is more frequent in the population will
tend to spread to fixation. Underdominant systems therefore display features
similar to that of a bistable switch at the population level—if the system is
present above a critical threshold frequency, it will tend to spread to fixation,
while if it is present below the threshold, it will tend to be eliminated in
favor of the alternative allele or chromosome. A variety of toxin antidote
systems have been proposed to achieve these underdominant dynamics and
the critical threshold frequency depends on the system and fitness cost.
A range of underdominant systems is available in Nature, including chro-
mosomal alternations such as inversions, translocations, and compound chro-
mosomes [3,4]. We will return to translocations in the Translocation section;
but will concentrate here on novel forms of underdominance that are in prin-
ciple straightforward to engineer using combinations of toxins and antidotes.
Toxin antidote approaches to underdominance were originally proposed by
Davis et al., who suggested an elegant system having two transgenic con-
structs, each of which possesses a gene whose expression induces lethality
and a gene that suppresses the expression or activity of the gene inducing
lethality carried by the other construct [66]. The constructs can either be
inserted at the same locus on a pair of homologous chromosomes or at dif-
ferent loci on nonhomologous chromosomes. These systems display underdo-
minant properties because individuals carrying neither or both constructs are
viable; but a proportion of their offspring—those carrying just one of the
184 Genetic Control of Malaria and Dengue

constructs—are unviable. The critical threshold for the two-locus system
is B27%, above which it is predicted to spread to fixation, and for the
single-locus system is B67% [66].
Current Status. Attempts to engineer the underdominance system
proposed by Davis et al. have thus far been unsuccessful [66]; however, a
related novel underdominant system known as maternal-effect lethal under-
dominance (UD
MEL
) has recently been engineered in D. melanogaster and
demonstrated to replace wild-type laboratory populations in a threshold-
dependent manner [67,68]. In the UD
MEL
system, there are two transgenic
constructs, each of which possesses a maternally expressed toxin gene whose
activity is manifest during progeny embryogenesis and a zygotic antidote
gene capable of neutralizing the maternal toxin expressed by the opposite
construct. From the crosses produced by this system (Figure S1 of Akbari
et al. [67]), it can be seen that heterozygous females are sterile if mated to
wild-type individuals, while populations of transgenic homozygotes are fully
viable, as are wild-type populations. This leads to the characteristic
bistable dynamics of underdominant systems. As per the system proposed by
Davis et al., the UD
MEL
constructs can be inserted at the same locus or on a
pair of homologous chromosomes [66]. The critical threshold for the two-
locus system is B19% and for the single-locus system is B64%, assuming
no fitness costs [67], and threshold-dependent drive has been demonstrated
in the laboratory for both cases (Figure 9.2B).
Design Criteria. Toxin antidote-based underdominant systems such as
UD
MEL
are an excellent option during the testing phase of population
replacement, or whenever a confined release is desired. The threshold nature
of these systems has three advantages in these scenarios. First, they are
unlikely to spread following an accidental released because escapees will
inevitably be present at subthreshold levels and be eliminated from the envi-
ronment [18]. Second, they are expected to be confineable to isolated release
sites because transgenic insects released at superthreshold frequencies are
expected to spread transgenes locally while they remain at subthreshold
levels at nearby locations. And third, releases are reversible as transgenes
can be eliminated by diluting them to subthreshold frequencies through a
sustained release of wild-type insects.
It should be noted that the confineability of these systems, although
likely, is not guaranteed.
In theory, chance events could lead to underdominant systems gaining a
foothold and spreading in structured populations, presumably beginning from
a single individual; however, this is more likely to occur on an evolutionary
timescale than on a human timescale. Underdominant systems may be better-
suited to An. gambiae because it disperses quickly over the range of a single
village [69,70], reducing the chance of its spread being confined to smaller
subpopulations. The small-scale population structure of Ae. aegypti, however,
may prevent its village-wide spread in natural populations of these vectors.
Gene Drive Strategies for Population Replacement Chapter | 9 185

Otherwise, similarly to Medea, solutions are available for all of the design
criteria outlined earlier. As the functional components are developed and
identified in mosquitoes—microRNAs, maternal and early-zygote-specific
promoters and essential genes—these systems will be highly useful for
confined population replacement of vector species such as An. gambiae.
Killer-Rescue
Killer-rescue is an intriguingly simple two-locus gene drive system proposed
by Gould et al. for both its ease of engineering and its ability to spread into
a population in a time-limited way [71]. Both these qualities are desirable in
the early stages of a population replacement program. The system consists of
two alleles at unlinked loci—one that encodes a toxin (a killer allele) and
another that confers immunity to the toxin (a rescue allele), which could be
tightly linked to a gene for disease refractoriness. A release of individuals
homozygous for both alleles results in temporary drive as the alleles segre-
gate and the presence of the killer allele in the population confers a benefit
to those also carrying the rescue allele. In an alternative configuration, a
second killer allele can be included at an independently assorting locus to
enhance the selective benefit of the rescue allele. However, regardless of
the conformation, the killer allele soon declines in frequency due to its
inherent fitness cost and, as it does, the selective benefit of the rescue
allele is lost. As this happens, if the rescue allele or disease-refractory
gene confers a fitness cost to the host, then it will gradually be eliminated
from the population as well over a timeframe determined by the magnitude
of its fitness cost—a higher fitness cost leading to it being eliminated more
quickly.
Design Criteria. As mentioned earlier, the killer-rescue system is intrigu-
ing for its ability to spread in a time-limited manner, thus reducing risks, as
appropriate during field trials of transgenic mosquitoes carrying disease-
refractory genes. The system is also spatially limited, as it only has a win-
dow of time in which to disperse to neighboring populations, and will spread
to much lower levels in these populations than at the population of release
[72]. Similar to underdominant systems, it will not persist following an acci-
dental release, and its elimination from a population can be accelerated
through large-scale releases of wild-type insects. Also, similar to other
toxin antidote systems, solutions are available for all of the design criteria
outlined earlier.
Some consideration should go into the fitness cost of the rescue allele
and refractory gene, as a high fitness cost will lead to rapid elimination, but
the maximum frequency of the disease-refractory allele in the population
will be compromised; while small fitness costs will allow the system
to spread to very high maximum frequencies, but it may take several years
for the system to be eliminated from the population entirely. Further
186 Genetic Control of Malaria and Dengue

complicating this, fitness costs are exceedingly difficult to quantify in the
field. The bistable nature of underdominant systems therefore makes them
more controllable in terms of confinement and reversibility; however, the
major benefit of the killer-rescue system is its ease of engineering.
Molecular tools are already available to engineer the system in a variety of
mosquito species, allowing the system to be implemented with relative ease
in a range of disease vectors.
Other Confineable Toxin Antidote Systems
As Medea, killer-rescue, and the various forms of engineered underdomi-
nance highlight, there are many ways in which toxins and antidotes can be
used to favor the inheritance of one allele over another. For example, even if
we limit ourselves to single-locus systems like Medea, either the toxin or
antidote gene could be placed under the control of a paternal, maternal, or
zygote-specific promoter, function through a recessive or dominant mecha-
nism, and be located on a sex chromosome or autosome [73]. The possibili-
ties multiply if we also consider multilocus systems. A few additional
toxin antidote systems displaying unique population dynamics are Semele,
inverse Medea, and Medusa, all of which are also confineable to partially
isolated populations.
Semele.Semele is a single-locus system consisting of a toxin gene
expressed in the semen of transgenic males that either kills or renders
infertile wild-type females and an antidote gene expressed in females that
protects them against the effects of the toxin [74]. The name is an acronym
for “semen-based lethality” and, like Medea, also has Greek origins. In
Greek mythology, Semele was a mortal female who attracted the attention
of Zeus while slaughtering a bull at his altar (Zeus, at this point, was flying
overhead disguised as an eagle). Zeus became infatuated with Semele and
impregnated her, but Semele died after witnessing his godliness because
she was not herself a god. The story parallels the biology of the Semele
construct, in which wild-type females die (or become infertile) upon
mating with transgenic males.
Semele has several interesting population dynamic properties. If only
males carrying the Semele allele are released into a wild population, they are
expected to suppress the population size when released in large numbers.
This happens because all of the wild females that mate with the Semele
males are susceptible to their toxic semen. If both males and females carry-
ing the Semele allele are released, the system displays bistable dynamics
with a threshold frequency of B36% in the absence of fitness costs [74].
Above the threshold, the selective advantage of the female antidote
outweighs the reproductive disadvantage conferred by the toxic semen and
the system spreads into the population. In combination, this means that an
initial release of Semele males could be used to suppress a population
Gene Drive Strategies for Population Replacement Chapter | 9 187

preceding a superthreshold release of males and females, thus reducing
the release size required to exceed the critical population frequency
(Figure 9.2C).
Inverse Medea. Inverse Medea is another single-locus system capable of
achieving confined population replacement [75]. The system consists of a
zygotic toxin and maternal antidote—essentially the same components as the
Medea system with the promoters switched. This has the effect of rendering
heterozygous offspring of wild-type mothers unviable and leads to
bistable dynamics in which the system spreads when it represents a majority
of the alleles in a population, and is otherwise eliminated. While similar
dynamic properties are displayed by other underdominant toxin antidote
systems, the benefit of inverse Medea is its ease of engineering once the
components to generate Medea elements in mosquito vectors have been iden-
tified. Several approaches to engineering these elements are available—for
example, the toxin could be a microRNA that silences expression of a gene
whose activity is required for early embryo development, and the antidote
could be a maternally expressed RNA that restores the necessary activity to
the zygote and is resistant to silencing.
Medusa.Medusa is a two-construct, sex chromosome-linked drive system
capable of inducing confineable and reversible population suppression [76].
The system consists of four components—two at a locus on the X chromo-
some and two at a locus on the Y chromosome. The combination of a
maternally expressed, X-linked toxin and a zygotically expressed, Y-linked
antidote causes suppression of the female population and selects for the
transgene-bearing Y since only transgenic male offspring of Medusa-bearing
females are protected from the effects of the toxin. At the same time, the
combination of a zygotically expressed, Y-linked toxin and a zygotically
expressed, X-linked antidote selects for the transgene-bearing X when the
transgene-bearing Y is present. Together, this creates a balanced lethal sys-
tem that, when present above a certain threshold frequency, spreads while
creating a strong male gender bias, hence causing population suppression
(Figure 9.2D). Characteristic of all drive systems with thresholds, releases of
Medusa mosquitoes are confineable and reversible, making the system an
ideal tool for confined population suppression. This could be particularly
useful in the lead-up to releases of invasive population suppression systems
such as Y-linked X-shredder HEGs [76].
The name Medusa is an acronym for “sex chromosome-associated Medea
underdominance,” as its components are identical to those of Medea and
engineered underdominance. The name also has origins in Greek mythology,
where Medusa is a beautiful yet terrifying woman who caused onlookers to
be turned to stone (toxin) but was ultimately beheaded by Perseus who
distracted himself with Athena’s mirrored shield (antidote). Simple popula-
tion dynamic models show that an all-male release of Medusa males, carried
out over six generations, is expected to induce a population crash within 12
188 Genetic Control of Malaria and Dengue

generations for modest release sizes [76]. Reinvasion of wild-type insects
can result in a population rebound; however, this can be prevented through
regular releases of modest numbers of Medusa males.
Design Criteria. The vast range of possible toxin antidote combina-
tions highlights the versatility of this approach to engineering gene drive
systems. Semele is an excellent option for confined population replacement
due to its ability to suppress a vector population prior to replacement,
inverse Medea is an excellent underdominant system that is easy to engi-
neer once the components of the Medea system have been identified in
mosquito vectors, and Medusa is an ideal system for confined population
suppression in preparation for invasive X-shredder strategies [76]. Other
toxin antidote systems are imaginable and may be favored depending on
the components first identified in molecular work on vector species [73].
As toxin antidote systems, the design criteria outlined earlier are gener-
ally satisfied, and as largely confineable systems, the systems highlighted
here are excellent options during the testing phase of population replace-
ment, or whenever a confined release is desired.
TRANSLOCATIONS
As the first gene drive system to be proposed [4], translocations have since
undergone a lull in interest following the observation that radiation-
generated translocations failed to spread in the field, likely due to high
fitness costs induced by X-rays [7]. However, recent developments in molec-
ular biology permit the creation of translocations without relying upon radia-
tion suggesting that, after several decades of inactivity, the application of
this gene drive system could be revisited. Translocations result from the
mutual exchange of chromosomal segments between nonhomologous chro-
mosomes. Translocation heterozygotes are usually partially sterile, while
translocation homozygotes are usually fully fertile. This effect is manifest
during meiosis when nearly half of the gametes from a translocation hetero-
zygote have a duplication of one chromosomal segment and a deficiency of
another. The haploid gametes are functional, but when they fuse with native
gametes following fertilization, the resulting zygotes are inviable. This pro-
duces the bistable dynamics described for other underdominant systems.
Current Status. Curtis proposed that if a translocation strain was devel-
oped that had a disease-refractory gene tightly linked to the translocation
break point, disease-resistance would spread into that population as the trans-
location fixes [4]. To test this hypothesis, mosquito strains with chromo-
somal translocations were developed using X-ray mutagenesis; however, the
low fitness associated with these strains and the difficulty of bringing
disease-refractory genotypes into appropriate genetic backgrounds inhibited
these approaches from further development. It is now possible to generate
translocations at almost any genomic location without irradiation as a result
Gene Drive Strategies for Population Replacement Chapter | 9 189

of progress in genome sequencing and synthetic biology [77,78]. This will
reduce the fitness costs associated with translocations and will allow disease-
refractory genes to be more easily linked to translocation break points,
making them a feasible, future gene drive system for confined population
replacement.
Design Criteria. As an underdominant system displaying bistable
dynamics, translocations provide another option for confined population
replacement. As modern molecular techniques are yet to be applied to the
development of this system, its agreement with several of the design crite-
ria mentioned earlier are yet to be determined, and its attractiveness as a
local gene drive system will depend on its ease of engineering and satisfac-
tion of these criteria in comparison to toxin antidote-based underdominant
systems. Toxin antidote-based systems may be preferable for phased
releases as their components are more similar to invasive Medea elements
that could be used for subsequent wide-scale population replacement. That
said there is a theoretical expectation that translocations are an effective
gene drive system for local population replacement [4] and that the loss of
disease-refractory genes will be minimized by inserting them at transloca-
tion break points. As a bistable system, translocations could be eliminated
from a population through mass release of wild-type insects and would
satisfy social and regulatory requirements when confinement is desired.
CONCLUSION
In 2006, Sinkins and Gould published an excellent review of gene drive
systems for insect disease vectors which today provides a testament to how
quickly the field has progressed in less than a decade [79]. As the authors
state, “ultimately, the drive system that becomes most widely used might
be one that is entirely novel and not described here.” Interestingly, the
majority of the drive systems described in this chapter—TALENs,
CRISPRs, UD
MEL
, killer-rescue, Semele, inverse Medea, and Medusa—
were not mentioned in the Sinkins and Gould review as they are have only
been recently published.
Of the systems that were mentioned by Sinkins and Gould, progress has
been rapid. In mentioning Medea, for instance, the authors stated that “a
molecular understanding of its function could lead to the development of
artificial Medea-like constructs”—something that was achieved the following
year [14] and is now one of the most promising approaches for population
replacement. Regarding HEGs, the authors stated that, “Unfortunately, HEGs
have only been reported in fungi, plants, bacteria and bacteriophages ... the
potential for developing an HEG-based functional system in insects is
unknown.” The following year, a HEG isolated from a species of slime
mold demonstrated cleavage activity in An. gambiae [36], and a few
years later, another naturally occurring HEG was shown to spread in
190 Genetic Control of Malaria and Dengue

laboratory populations of both D. melanogaster and An. gambiae [15,30,31].
HEGs are now one of the most promising gene drive systems for inducing
population suppression (Table 9.1). Research on using Wolbachia to control
vector-borne diseases has been even more rapid, with large-scale field trials
already having taken place in several countries including Australia and
Vietnam [80]. This prompts the question of what the gene drive field will
look like a decade from now?
Gene Drive for Any Situation
Sinkins and Gould also pointed out that “the various types of drive mechan-
isms should not be viewed as competing systems,” adding that, “Different
characteristics will be needed in different situations.” Gene drive systems
can lead to a number of outcomes in terms of population dynamics, and the
optimal system in each case will depend upon the desired outcome. For driv-
ing disease-refractory genes into mosquito populations over a wide area,
Medea seems to be a very promising system, as it is capable of spreading
from low initial frequencies and is also stably integrated into the host chro-
mosome. When population replacement is desired over a wide geographic
area, stability in the face of evolutionary degradation is an important consid-
eration, and Medea may be preferable to homing-based strategies incorporat-
ing disease-refractory genes because these are susceptible to DNA loss
during homology-directed repair, which is expected to become increasingly
significant over large spatial and temporal scales.
Systems with release thresholds are preferable when a confined release is
desired because these systems are likely to be confineable to their population
of release and to be reversible through releases of wild-type insects [72].
Toxin antidote-based underdominant systems would be an obvious choice if
the goal were to test the concept of population replacement prior to a release
of toxin antidote-based Medea elements. The bistable nature of these
systems makes them particularly amenable to confinement; however, killer-
rescue systems and a mass release of transgenic insects with disease-
refractory genes [19,71] should also be considered, as these are significantly
easier to engineer in a wide range of vector species and the spreading a
disease-refractory gene into an isolated population will not always require
gene drive.
For population suppression, Y-linked X-shredder HEGs are an ideal
system, assuming the X-shredding HEG can be docked onto the Y chromo-
some and expressed during spermatogenesis. The major benefits of the
X-shredder HEG are the generally small size of HEGs, making them less
susceptible to evolutionary degradation, and the large number of loci
cleaved on the X chromosome, making the strategy less susceptible to
target site mutagenesis [11]. Autosomal X-shredders, as a self-limiting
population suppression system acting through the same molecular
Gene Drive Strategies for Population Replacement Chapter | 9 191

TABLE 9.1 Alignment of Potential Gene Drive Systems with Design Criteria Outlined in the Introduction
a
Design
Criteria
Target Site Cleavage-Based Gene Drive Systems Toxin Antidote-Based Gene Drive
Systems
Engineered
Translocations
Engineered
TEs
Engineered
HEGs
ZFNs, TALENs CRISPRs Medea, UD
MEL
Killer-rescue,
Semele,
Medusa
Effectiveness
of spread
Maybe [10]
(not yet
effective in
vector
species)
Yes [11] (very
effective, first
drive system
shown to
spread in a
malaria vector)
Probably [40]
(homing
demonstrated,
currently
compromised by
mutational
inactivation)
Probably
[51 57]
(components
identified, can
target multiple
sequences at
once)
Probably
[14,64,67]
(observed to
spread in
laboratory
Drosophila
populations)
Theoretically
[71,74,76] (not
yet engineered,
models predict
spread)
Theoretically
[4,77,78] (models
predict spread, not
yet engineered
using modern
components)
Ability to
carry large
effector
genes
No [22]
(transposition
rate declines
with
increasing
insert size)
Possibly [11]
(could be lost
during
homology-
directed repair)
Possibly [40]
(could be lost
during
homology-
directed repair)
Possibly (could
be lost during
homology-
directed repair)
Yes [14,67] (stably
integrated into
host chromosome)
Yes [71,74,76]
(stably
integrated into
host
chromosome)
Yes [4] (stably
integrated into
host chromosome)
Tight linkage
with effector
genes
No [23]
(transposition
events prone
to DNA loss)
No [11]
(homology-
directed repair
prone to DNA
loss)
No [40]
(homology-
directed repair
prone to DNA
loss)
No (homology-
directed repair
prone to DNA
loss)
Yes [14,67] (place
toxin and effector
genes within
intron of antidote
gene)
Yes [71,74,76]
(place effector
gene within
intron of
antidote gene)
Yes [4] (very tight
if effector gene
linked to a
translocation
break point)
Waves of
introductions
Maybe [10]
(difficult-to-
engineer
multiple TEs)
Maybe [11]
(difficult to
engineer)
Yes [40] (easier
to engineer than
HEGs)
Yes (easier to
engineer than
HEGs)
Yes [14,67] (use
distinct
toxin antidote
combinations)
Yes [71,74,76]
(use distinct
toxin antidote
combinations)
Yes [4] (use
threshold
properties)

Easily
adapted to
other species
No [10]
(difficult to
find TEs
compatible
with vector
species)
No (difficult to
engineer target
site)
Maybe [40] (once
components
identified in
species)
Maybe (once
components
identified in
species)
Maybe [14,67]
(once components
identified in
species)
Maybe
[71,74,76]
(once
components
identified in
species)
Maybe [77,78]
(once components
identified in
species)
Stability in
target
species
No [8] (large
number of
target sites
undermine
predictability)
Yes [11] (short
sequences
targeting
precise
genomic
regions)
Moderate to low
[40] (prone to
mutation due to
repetitive nature)
Moderate to
low (more
prone to
mutation due to
repetitive nature
and large size)
Yes [14,67] (stably
integrated into
host chromosome)
Yes [71,74,76]
(stably
integrated into
host
chromosomes)
Yes [77,78] (very
stable)
Minimal
horizontal
gene transfer
No [8] (wide
species host
range)
Yes [11]
(include
species-specific
regulatory
sequences)
Yes [11,40]
(include species-
specific
regulatory
sequences)
Yes [11]
(include
species-specific
regulatory
sequences)
Yes [14,67]
(include species-
specific regulatory
sequences)
Yes [71,74,76]
(include
species-specific
regulatory
sequences)
Yes
Mechanism
for removal
No Yes [11]
(design second
HEG to target
first HEG)
Yes [11,40]
(design second
ZFN or TALEN to
target first ZFN or
TALEN)
Yes [11] (design
second CRISPR
to target first
CRISPR)
Yes [14,67] (use
threshold
properties or
second-generation
element to remove
refractory gene)
Yes [71,74,76]
(use threshold
properties or
dilution with
wild-types)
Yes [4] (use
threshold
properties)
Social and
regulatory
requirements
No [79] (not
confineable
or reversible)
Yes [11] (wide
range of
strategies with
different levels
of
confineability)
Yes [11,40] (wide
range of strategies
with different
levels of
confineability)
Yes [11] (wide
range of
strategies with
different levels
of
confineability)
Yes [18,72]
(confineable or
unlikely to spread
following small
accidental release)
Yes [72]
(confineable to
partially
isolated
populations)
Yes [72]
(confineable to
partially isolated
populations)
a
In many cases, data supporting satisfaction of design criteria are preliminary. TEs, transposable elements; HEGs, homing endonuclease genes; ZFNs, zinc-finger nucleases;
TALENs, transcription-activator-like effector nucleases; CRISPRs, clustered, regularly interspaced, short palindromic repeats; UD
MEL
, maternal-effect lethal underdominance.

mechanism, are an obvious choice for testing this drive system prior to a
wide-scale release. Similar approaches using ZFNs, TALENs, and
CRISPRs should also be considered, especially considering their relative
ease of engineering. However, the repetitive nature of ZFNs and TALENs
and the large size of CRISPRs generally will make them more susceptible
to mutation and evolutionary degradation (Figure 9.3).
Outstanding Issues and Future Outlook
In 1899, US patent officer Charles Duell famously stated that, “Everything
that can be invented already has been invented.” It would be just as foolish
to say that all imaginable gene drive systems have already been imagined.
The coming decades are bound to witness the emergence of a plethora of
novel mechanisms for spreading desirable genes into insect populations, and
it will be fascinating to see how these systems align with the design criteria
mentioned earlier. Furthermore, of the systems for which development has
already begun, it will be fascinating to see how their laboratory and field
studies progress. Progress on toxin antidote-based systems will be greatly
facilitated by the development of their functional components—toxins, anti-
dotes and regulatory elements—in mosquito vectors. It will also be interest-
ing to see how modern approaches to translocations perform against
toxin antidote-based approaches to underdominance. Regarding homing-
based systems, critical developments will be the engineering of HEGs for
other vector species, the insertion and expression of X-shredders on the Y
chromosome, and determining the resilience of alternative homing-based
systems to evolutionary degradation.
As a technology capable of engineering or eliminating entire species, the
development of gene drive systems carries with it both great promise and
great responsibility. Issues are heightened by the ability of invasive systems
to spread into neighboring communities and countries without their consent
[81]. Comprehensive risk assessments that address ecological, epidemiologi-
cal, and social issues are therefore essential, and such technology should
only be used in the absence of significant risks. On the flip side, gene drive
technology has the potential to make a profound impact on relieving the
global vector-borne disease burden [2]. Considering malaria as an example,
traditional interventions such as bed nets and antimalarial drugs require
human compliance, which never truly exceeds B80% coverage, meaning
that there is always a residual human population capable of sustaining
transmission [82]. Replacement of disease-transmitting mosquitoes with
disease-refractory ones has the unique benefit that it does not require human
compliance, and can spread into areas where interventions are difficult to
apply. This makes it one of the most promising components of future
integrated strategies for the elimination of vector-borne diseases.
194 Genetic Control of Malaria and Dengue

FIGURE 9.3 Confineability and stability of potential gene drive systems. The potential gene
drive systems described in this chapter differ in multiple ways, including their confineability
(the ability to limit their spatial spread following a release) and their stability (resilience against
evolutionary degradation, predictable behavior in the host organism and infrequent spread into
nontarget species). Here, we depict the potential gene drive systems in a two-dimensional graph
according to these properties. Self-limiting systems eliminate themselves from a population as a
result of their own dynamics and hence are highly confineable, although some persist in a popu-
lation longer than others. Self-sustaining systems are capable of maintaining a high population
frequency but are relatively confineable if they display threshold properties in terms of release
frequency. Self-sustaining systems not displaying threshold dynamics can be highly invasive.
Toxin antidote-based systems (yellow) are relatively stable but have differing levels of confine-
ability. Cleavage-based population replacement systems (purple) are relatively invasive whether
they carry disease-refractory genes or induce a population fitness load. The process of homing
also causes them to be relatively unstable due to errors introduced during gap repair.
Cleavage-based population suppression systems (salmon) can be either invasive if located on the
Y chromosome or self-limiting if located on an autosome. ZFNs, zinc-finger nucleases;
TALENs, transcription-activator-like effector nucleases; CRISPRs, clustered, regularly inter-
spaced, short palindromic repeats; HEGs, homing endonuclease genes; UD
MEL
, maternal-effect
lethal underdominance.
Gene Drive Strategies for Population Replacement Chapter | 9 195

ACKNOWLEDGMENTS
JMM acknowledges support from a fellowship from the Medical Research Council/
Department for International Development, UK.
REFERENCES
[1] Isaacs AT, Jasinskiene N, Tretiakov M, et al. Transgenic Anopheles stephensi coexpres-
sing single-chain antibodies resist Plasmodium falciparum development. Proc Natl Acad
Sci USA 2012;109(28):E1922 30.
[2] James AA. Gene drive systems in mosquitoes: rules of the road. Trends Parasitol 2005;21
(2):64 7.
[3] Serebrovskii AS. On the possibility of a new method for the control of insect pests. Zool
Zh 1940;19:618 30.
[4] Curtis CF. Possible use of translocations to fix desirable genes in insect pest populations.
Nature 1968;218(5139):368 9.
[5] Kidwell MG, Ribeiro JM. Can transposable elements be used to drive disease refractori-
ness genes into vector populations? Parasitol Today 1992;8(10):325 9.
[6] Curtis CF, Robinson AS. Computer simulation of the use of double translocations for pest
control. Genetics 1971;69(1):97 113.
[7] Lorimer N, Hallinan E, Rai KS. Translocation homozygotes in the yellow fever mosquito,
Aedes aegypti. J Hered 1972;63(4):158 66.
[8] Preston CR, Engels WR. Spread of P transposable elements in inbred lines of Drosophila
melanogaster. Prog Nucleic Acid Res Mol Biol 1989;36:71 85.
[9] Charlesworth B, Sniegowski P, Stephan W. The evolutionary dynamics of repetitive DNA
in eukaryotes. Nature 1994;371(6494):215 20.
[10] O’Brochta DA, Sethuraman N, Wilson R, et al. Gene vector and transposable element
behavior in mosquitoes. J Exp Biol 2003;206(Pt. 21):3823 34.
[11] Burt A. Site-specific selfish genes as tools for the control and genetic engineering of natu-
ral populations. Proc Biol Sci 2003;270(1518):921 8.
[12] Beeman RW, Friesen KS, Denell RE. Maternal-effect selfish genes in flour beetles.
Science 1992;256(5053):89 92.
[13] Lorenzen MD, Gnirke A, Margolis J, et al. The maternal-effect, selfish genetic element
Medea is associated with a composite Tc1 transposon. Proc Natl Acad Sci USA 2008;105
(29):10085 9.
[14] Chen CH, Huang H, Ward CM, et al. A synthetic maternal-effect selfish genetic element
drives population replacement in Drosophila. Science 2007;316(5824):597 600.
[15] Windbichler N, Menichelli M, Papathanos PA, et al. A synthetic homing endonuclease-
based gene drive system in the human malaria mosquito. Nature 2011;473(7346):
212 15.
[16] Galizi R, Doyle LA, Menichelli M, et al. A synthetic sex ratio distortion system for the
control of the human malaria mosquito. Nat Commun 2014;5:3977.
[17] Braig HR, Yan G. The spread of genetic constructs in natural insect populations.
In: Letourneau DK, Burrows BE, editors. Genetically engineered organisms: assessing
environmental and human health effects. Boca Raton: CRC Press; 2001. p. 251 314.
[18] Marshall JM. The effect of gene drive on containment of transgenic mosquitoes. J Theor
Biol 2009;258(2):250 65.
196 Genetic Control of Malaria and Dengue

[19] Okamoto KW, Robert MA, Gould F, Lloyd AL. Feasible introgression of an anti-
pathogen transgene into an urban mosquito population without using gene-drive. PLoS
Negl Trop Dis 2014;8(7):e2827.
[20] Plasterk RHA, van-Luenen HGAM. The Tc1/Mariner family of transposable elements.
In: Craig NL, Craigie R, Geller M, Lambowitz AM, editors. Mobile DNA II. Washington,
DC: American Society for Microbiology; 2002. p. 519 32.
[21] Smith RC, Atkinson PW. Mobility properties of the Hermes transposable element in
transgenic lines of Aedes aegypti. Genetica 2011;139(1):7 22.
[22] Lampe DJ, Grant TE, Robertson HM. Factors affecting transposition of the Himar1
mariner transposon in vitro. Genetics 1998;149(1):179 87.
[23] Marshall JM. The impact of dissociation on transposon-mediated disease control
strategies. Genetics 2008;178(3):1673 82.
[24] Stoddard BL. Homing endonuclease structure and function. Q Rev Biophys 2005;38
(1):49 95.
[25] Deredec A, Burt A, Godfray HC. The population genetics of using homing endonuclease
genes in vector and pest management. Genetics 2008;179(4):2013 26.
[26] Craig Jr. GB, Hickey WA, Vandehey RC. An inherited male-producing factor in Aedes
aegypti. Science 1960;132(3443):1887 9.
[27] Hamilton WD. Extraordinary sex ratios. A sex-ratio theory for sex linkage and inbreeding
has new implications in cytogenetics and entomology. Science 1967;156(3774):477 88.
[28] Wood RJ, Newton ME. Sex-ratio distortion caused by meiotic drive in mosquitos.
Am Nat 1991;137(3):379 91.
[29] Traver BE, Anderson MA, Adelman ZN. Homing endonucleases catalyze double-stranded
DNA breaks and somatic transgene excision in Aedes aegypti. Insect Mol Biol 2009;18
(5):623 33.
[30] Chan YS, Huen DS, Glauert R, Whiteway E, Russell S. Optimising homing endonuclease
gene drive performance in a semi-refractory species: the Drosophila melanogaster experi-
ence. PLoS One 2013;8(1):e54130.
[31] Chan YS, Naujoks DA, Huen DS, Russell S. Insect population control by homing
endonuclease-based gene drive: an evaluation in Drosophila melanogaster. Genetics
2011;188(1):33 44.
[32] Chan YS, Takeuchi R, Jarjour J, Huen DS, Stoddard BL, Russell S. The design and
in vivo evaluation of engineered I-OnuI-based enzymes for HEG gene drive. PLoS One
2013;8(9):e74254.
[33] Takeuchi R, Choi M, Stoddard BL. Redesign of extensive protein-DNA interfaces of
meganucleases using iterative cycles of in vitro compartmentalization. Proc Natl Acad Sci
USA 2014;111(11):4061 6.
[34] Thyme SB, Boissel SJS, Quadri SA, et al. Reprogramming homing endonuclease specific-
ity through computational design and directed evolution. Nucleic Acids Res 2014;42
(4):2564 76.
[35] Flick KE, Jurica MS, Monnat Jr. RJ, Stoddard BL. DNA binding and cleavage by the
nuclear intron-encoded homing endonuclease I-PpoI. Nature 1998;394(6688):96 101.
[36] Windbichler N, Papathanos PA, Catteruccia F, Ranson H, Burt A, Crisanti A. Homing
endonuclease mediated gene targeting in Anopheles gambiae cells and embryos. Nucleic
Acids Res 2007;35(17):5922 33.
[37] Windbichler N, Papathanos PA, Crisanti A. Targeting the X chromosome during sper-
matogenesis induces Y chromosome transmission ratio distortion and early dominant
embryo lethality in Anopheles gambiae. PLoS Genet 2008;4(12):e1000291.
Gene Drive Strategies for Population Replacement Chapter | 9 197

[38] Klein TA, Windbichler N, Deredec A, Burt A, Benedict MQ. Infertility resulting from
transgenic I-PpoI male Anopheles gambiae in large cage trials. Pathog Glob Health
2012;106(1):20 31.
[39] Bernardini F, Galizi R, Menichelli M, et al. Site-specific genetic engineering of the
Anopheles gambiae Y chromosome. Proc Natl Acad Sci USA 2014;111(21):7600 5.
[40] Simoni A, Siniscalchi C, Chan YS, et al. Development of synthetic selfish elements based
on modular nucleases in Drosophila melanogaster. Nucleic Acids Res 2014;42
(11):7461 72.
[41] Boch J, Scholze H, Schornack S, et al. Breaking the code of DNA binding specificity of
TAL-type III effectors. Science 2009;326(5959):1509 12.
[42] Moscou MJ, Bogdanove AJ. A simple cipher governs DNA recognition by TAL effectors.
Science 2009;326(5959):1501.
[43] Mak AN, Bradley P, Cernadas RA, Bogdanove AJ, Stoddard BL. The crystal structure of
TAL effector PthXo1 bound to its DNA target. Science 2012;335(6069):716 19.
[44] Deng D, Yan C, Pan X, et al. Structural basis for sequence-specific recognition of DNA
by TAL effectors. Science 2012;335(6069):720 3.
[45] Beerli RR, Barbas III CF. Engineering polydactyl zinc-finger transcription factors. Nat
Biotechnol 2002;20(2):135 41.
[46] Gaj T, Gersbach CA, Barbas III CF. ZFN, TALEN, and CRISPR/Cas-based methods for
genome engineering. Trends Biotechnol 2013;31(7):397 405.
[47] Christian M, Cermak T, Doyle EL, et al. Targeting DNA double-strand breaks with TAL
effector nucleases. Genetics 2010;186(2):757 61.
[48] Carlson DF, Fahrenkrug SC, Hackett PB, Targeting DNA. With Fingers and TALENs.
Mol Ther Nucleic Acids 2012;1:e3.
[49] Wiedenheft B, Sternberg SH, Doudna JA. RNA-guided genetic silencing systems in bacte-
ria and archaea. Nature 2012;482(7385):331 8.
[50] Garneau JE, Dupuis ME, Villion M, et al. The CRISPR/Cas bacterial immune system
cleaves bacteriophage and plasmid DNA. Nature 2010;468(7320):67 71.
[51] Jiang WY, Bikard D, Cox D, Zhang F, Marraffini LA. RNA-guided editing of bacterial
genomes using CRISPR-Cas systems. Nat Biotechnol 2013;31(3):233 9.
[52] Mali P, Yang LH, Esvelt KM, et al. RNA-guided human genome engineering via Cas9.
Science 2013;339(6121):823 6.
[53] Cong L, Ran FA, Cox D, et al. Multiplex genome engineering using CRISPR/Cas sys-
tems. Science 2013;339(6121):819 23.
[54] Yu ZS, Ren MD, Wang ZX, et al. Highlyefficient genome modifications mediated by
CRISPR/Cas9 in Drosophila. Genetics 2013;195(1): 289 91.
[55] Gratz SJ, Cummings AM, Nguyen JN, et al. Genome engineering of Drosophila with the
CRISPR RNA-guided Cas9 nuclease. Genetics 2013;194(4): 1029 35.
[56] Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. A programmable
dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 2012;337
(6096):816 21.
[57] Wang HY, Yang H, Shivalila CS, et al. One-step generation of mice carrying mutations in
multiple genes by CRISPR/Cas-mediated genome engineering. Cell 2013;153(4):
910 18.
[58] Marshall JM. The toxin and antidote puzzle: new ways to control insect pest populations
through manipulating inheritance. Bioeng Bugs 2011;2(5):235 40.
198 Genetic Control of Malaria and Dengue

[59] Wade MJ, Beeman RW. The population dynamics of maternal-effect selfish genes.
Genetics 1994;138(4):1309 14.
[60] Ward CM, Su JT, Huang Y, Lloyd AL, Gould F, Hay BA. Medea selfish genetic elements
as tools for altering traits of wild populations: a theoretical analysis. Evolution 2011;65
(4):1149 62.
[61] Hurst LD. scat 1is a selfish gene analogous to Medea of Tribolium castaneum. Cell
1993;75(3):407 8.
[62] Weichenhan D, Kunze B, Traut W, Winking H. Restoration of the Mendelian transmission
ratio by a deletion in the mouse chromosome 1 HSR. Genet Res 1998;71(2):119 25.
[63] Beeman RW, Friesen KS. Properties and natural occurrence of maternal-effect selfish
genes (‘Medea’ factors) in the red flour beetle, Tribolium castaneum. Heredity 1999;82
(Pt. 5):529 34.
[64] Akbari OS, Chen CH, Marshall JM, Huang H, Antoshechkin I, Hay BA. Novel synthetic
medea selfish genetic elements drive population replacement in Drosophila; a theoretical
exploration of medea-dependent population suppression. ACS Synth Biol 2014;3
(12):915 28.
[65] Hartl DL, Clark AG. Principles of population genetics. 3rd ed. Sunderland, MA: Sinauer
Associates; 1997.
[66] Davis S, Bax N, Grewe P. Engineered underdominance allows efficient and economical
introgression of traits into pest populations. J Theor Biol 2001;212(1):83 98.
[67] Akbari OS, Matzen KD, Marshall JM, Huang H, Ward CM, Hay BA. A synthetic gene
drive system for local, reversible modification and suppression of insect populations. Curr
Biol 2013;23(8):671 7.
[68] Wimmer EA. Insect biotechnology: controllable replacement of disease vectors. Curr Biol
2013;23(10):R453 6.
[69] Silver JB. Mosquito ecology field sampling methods. 3rd ed. New York, NY: Springer;
2008.
[70] Taylor C, Toure YT, Carnahan J, et al. Gene flow among populations of the malaria
vector, Anopheles gambiae, in Mali, West Africa. Genetics 2001;157(2):743 50.
[71] Gould F, Huang Y, Legros M, Lloyd AL. A killer-rescue system for self-limiting gene
drive of anti-pathogen constructs. Proc Biol Sci 2008;275(1653):2823 9.
[72] Marshall JM, Hay BA. Confinement of gene drive systems to local populations: a compar-
ative analysis. J Theor Biol 2012;294:153 71.
[73] Marshall JM, Hay BA. General principles of single-construct chromosomal gene drive.
Evolution 2012;66(7):2150 66.
[74] Marshall JM, Pittman GW, Buchman AB, Hay BA. Semele: a killer-male, rescue-female
system for suppression and replacement of insect disease vector populations. Genetics
2011;187(2):535 51.
[75] Marshall JM, Hay BA. Inverse Medea as a novel gene drive system for local population
replacement: a theoretical analysis. J Hered 2011;102(3):336 41.
[76] Marshall JM, Hay BA. Medusa: a novel gene drive system for confined suppression of
insect populations. PLoS One 2014;9(7):e102694.
[77] Golic KG, Golic MM. Engineering the Drosophila genome: chromosome rearrangements
by design. Genetics 1996;144(4):1693 711.
Gene Drive Strategies for Population Replacement Chapter | 9 199

[78] Egli D, Hafen E, Schaffner W. An efficient method to generate chromosomal rearrange-
ments by targeted DNA double-strand breaks in Drosophila melanogaster. Genome Res
2004;14(7):1382 93.
[79] Sinkins SP, Gould F. Gene drive systems for insect disease vectors. Nat Rev Genet
2006;7(6):427 35.
[80] Hoffmann AA, Montgomery BL, Popovici J, et al. Successful establishment of Wolbachia
in Aedes populations to suppress dengue transmission. Nature 2011;476(7361):454 7.
[81] Marshall JM. The Cartagena Protocol and genetically modified mosquitoes. Nat
Biotechnol 2010;28(9):896 7.
[82] Griffin JT, Hollingsworth TD, Okell LC, et al. Reducing Plasmodium falciparum malaria
transmission in Africa: a model-based evaluation of intervention strategies. PLoS Med
2010;7(8):e1000324.
200 Genetic Control of Malaria and Dengue