ArticlePDF Available

Efficient production of male Wolbachia-infected Aedes aegypti mosquitoes enables large-scale suppression of wild populations

  • Verily Life Sciences
  • Verily Life Sciences

Abstract and Figures

The range of the mosquito Aedes aegypti continues to expand, putting more than two billion people at risk of arboviral infection. The sterile insect technique (SIT) has been used to successfully combat agricultural pests at large scale, but not mosquitoes, mainly because of challenges with consistent production and distribution of high-quality male mosquitoes. We describe automated processes to rear and release millions of competitive, sterile male Wolbachia-infected mosquitoes, and use of these males in a large-scale suppression trial in Fresno County, California. In 2018, we released 14.4 million males across three replicate neighborhoods encompassing 293 hectares. At peak mosquito season, the number of female mosquitoes was 95.5% lower (95% CI, 93.6–96.9) in release areas compared to non-release areas, with the most geographically isolated neighborhood reaching a 99% reduction. This work demonstrates the high efficacy of mosquito SIT in an area ninefold larger than in previous similar trials, supporting the potential of this approach in public health and nuisance-mosquito eradication programs. Mosquitoes are nearly eradicated in three suburbs of California using accurately sorted sterile male mosquitoes.
This content is subject to copyright. Terms and conditions apply.
1Verily Life Sciences, South San Francisco, CA, USA. 2Consolidated Mosquito Abatement District, Parlier, CA, USA. 3MosquitoMate Inc., Lexington, KY,
USA. 4Department of Entomology, University of Kentucky, Lexington, KY, USA. e-mail:;
Aedes aegypti (Linnaeus) is the primary vector of dengue,
chikungunya, Zika, and yellow fever. Native to Africa,
A. aegypti has invaded much of the tropics and subtropics
over the past four centuries13, putting more than two billion people
at risk of arboviral infection4. Although effective on a small scale,
traditional control methods such as source reduction and chemi-
cal insecticides, as currently implemented, have not prevented the
proliferation and spread of this species (although see ref. 5). SIT is an
alternative control strategy that exploits the fact that female mosqui-
toes normally mate only once6. If that mating is with a sterile male,
the female will not produce viable progeny. For agricultural pests,
large-scale, inundative releases of sterile males over many genera-
tions have resulted in population crashes and, in some cases, local
or widespread elimination6,7. SIT avoids many of the pitfalls of tra-
ditional mosquito abatement techniques, such as off-target effects,
insecticide resistance, and difficulties treating cryptic breeding sites,
but its efficacy in controlling wild populations of A. aegypti remains
unproven, with small field studies of typically less than 100 hectares
(ha) in size showing highly variable suppression results6,811.
A common way to sterilize males is by altering their genomes
in either a non-targeted manner (irradiation) or a targeted manner
(genetic engineering). However, the impaired ability of genetically
altered males to compete for female mates in the wild and public
resistance to the release of genetically modified mosquitoes remain
barriers to widespread use of these techniques12,13. Alternatively, the
maternally inherited, intracellular bacterium Wolbachia pipientis
can be used to create conditional sterility between released males
and wild-type females through a phenomenon termed cytoplasmic
incompatibility14. Wolbachia infects over half of all insects15 but
not wild A. aegypti populations16,17. However, egg microinjection
has been used to establish multiple infected lines of A. aegypti with
stable transfections of Wolbachia strains native to other dipteran
insects1821. In the case of the wAlbB Wolbachia-infected A. aegypti
WB1 colony18 used for this work, when an uninfected female mates
with a WB1 male, incompatibility between the maternal cytoplasm
and sperm results in undeveloped zygotes. However, infected WB1
females produce viable, Wolbachia-positive progeny regardless of
the infection status of the male (Fig. 1a). Males from transfected
colonies like WB1 are incompatible with uninfected, wild-type
females but do not suffer the same drawbacks as genetically altered
males (for example, refs. 10,2225), making them an attractive tool for
mosquito control. As Wolbachia-infected males are not sterile in the
Efficient production of male Wolbachia-infected
Aedes aegypti mosquitoes enables large-scale
suppression of wild populations
Jacob E. Crawford 1 ✉ , David W. Clarke1, Victor Criswell1, Mark Desnoyer 1, Devon Cornel2,
Brittany Deegan2, Kyle Gong1, Kaycie C. Hopkins1, Paul Howell1, Justin S. Hyde1, Josh Livni1,
Charlie Behling1, Renzo Benza1, Willa Chen1, Karen L. Dobson3, Craig Eldershaw 1, Daniel Greeley1,
Yi Han1, Bridgette Hughes1, Evdoxia Kakani1, Joe Karbowski1, Angus Kitchell1, Erika Lee1, Teresa Lin1,
Jianyi Liu1, Martin Lozano1, Warren MacDonald1, James W. Mains3, Matty Metlitz1, Sara N. Mitchell1,
David Moore1, Johanna R. Ohm1, Kathleen Parkes1, Alexandra Porshnikoff1, Chris Robuck1,
Martin Sheridan1, Robert Sobecki1, Peter Smith1, Jessica Stevenson1, Jordan Sullivan1, Brian Wasson1,
Allison M. Weakley1, Mark Wilhelm1, Joshua Won1, Ari Yasunaga1, William C. Chan1, Jodi Holeman2,
Nigel Snoad1, Linus Upson1, Tiantian Zha1, Stephen L. Dobson3,4, F. Steven Mulligan2, Peter Massaro1
and Bradley J. White 1 ✉
The range of the mosquito Aedes aegypti continues to expand, putting more than two billion people at risk of arboviral infection.
The sterile insect technique (SIT) has been used to successfully combat agricultural pests at large scale, but not mosquitoes,
mainly because of challenges with consistent production and distribution of high-quality male mosquitoes. We describe auto-
mated processes to rear and release millions of competitive, sterile male Wolbachia-infected mosquitoes, and use of these
males in a large-scale suppression trial in Fresno County, California. In 2018, we released 14.4 million males across three rep-
licate neighborhoods encompassing 293 hectares. At peak mosquito season, the number of female mosquitoes was 95.5%
lower (95% CI, 93.6–96.9) in release areas compared to non-release areas, with the most geographically isolated neighborhood
reaching a 99% reduction. This work demonstrates the high efficacy of mosquito SIT in an area ninefold larger than in previous
similar trials, supporting the potential of this approach in public health and nuisance-mosquito eradication programs.
Articles NATure BIoTeChNology
classic sense, this approach is sometimes referred to as the incom-
patible insect technique26.
All mosquito SIT programs aim to minimize the release of
females to avoid increasing nuisance biting and disease transmis-
sion. However, preventing the release of females is particularly
important with Wolbachia-based programs because they have the
potential to establish and replace the wild population of mosqui-
toes, eliminating the utility of Wolbachia-infected males for con-
trol. Although such population replacement is unlikely when there
is a large population of wild mosquitoes, the chances of popu-
lation replacement increases when wild populations are small,
making high-accuracy sex sorting ever more important18,27,28. To
minimize the likelihood of population replacement, two groups
recently treated pupae with low-dose irradiation to sterilize residual
Wolbachia-infected females29,30. While promising, as implemented
this technique still reduces the competitiveness of males, albeit less
than traditional high-dose irradiation, and does not always result in
complete female sterilization.
Regardless of the sterilization technique, large-scale control of
mosquito populations with SIT is a challenging operational prob-
lem requiring industrialization of rearing, sex sorting, and release.
Groups around the world began to tackle these challenges in the
middle of the 20th century, with several notable successes. In the
late 1970s, the United States Department of Agriculture (USDA)-
backed SIT program targeting Anopheles albimanus in the Lake
Apastepeque region of El Salvador achieved near elimination across
a 1,500 ha valley due to a combination of mass-rearing and release
innovations, careful execution, and spatially limited wild mosquito
reproduction3135. In 1967, Culex quinquefasciatus was temporar-
ily eliminated from a small town in southern Burma, although it
rapidly re-established owing to the relatively long flight distance of
Culex mosquitoes36. Unfortunately, difficulties in sustaining pro-
duction of competitive males and in obtaining funding led to the
dissolution of all major mosquito SIT programs by the early 1980s.
Recent technological advancements, including genetically modified
sterile males12 and Wolbachia-transfected mosquito colonies18, have
led to renewed interest and investment in mosquito-targeted SIT.
Here, we develop tools to automate the production and distri-
bution of male mosquitoes infected with Wolbachia and test them
on field populations in Fresno County, which lies in the Central
Valley of California. A. aegypti was first detected in this region in
2013, with genetic analysis suggesting the South Central region of
the United States as the most likely source population37. Although
efforts were made to eliminate the nascent population with tradi-
tional control tools, the mosquito became established and contin-
ues to expand its range in the Central Valley, invading new cities
at a rapid rate3. Establishment in Fresno is part of a recent, larger
range expansion of A. aegypti into dry, hot metropolises across the
southwestern United States, including Los Angeles, Phoenix, and
Las Vegas. Unlike in tropical habitats, the population of A. aegypti
in Fresno County depends on anthropogenic water sources and is
highly correlated with seasonal ambient temperatures, with adult
populations increasing in June and July, peaking from August to
October, and largely undetectable from December to April (Fig. 1b).
We demonstrated the effectiveness and scalability of automated SIT
through open releases into three neighborhoods encompassing 293
ha and over 3,000 households within the cities of Clovis and Fresno
in Fresno County, California.
Mosquito mass rearing. To achieve stable production of males, we
developed an automated larval rearing system (LRS) that takes first
instar larvae as input, and outputs pupae (Fig. 2a). Prior to load-
ing onto the LRS, eggs are hatched overnight, after which L1 larvae
are automatically counted using a COPAS 550 (Union Biometrica)
large-particle flow cytometer into 50-ml conical tubes. The first step
in the LRS is larval container assembly, in which disposable plastic
containers are filled with water and food. Larvae are automatically
transferred from the conical tube into the container by a robotic
larval transfer arm. After filling and sealing, containers are auto-
matically transferred to an incubated storage and retrieval frame
(Supplementary Video 1). Larvae develop for 6 days in the frame,
during which they are automatically fed. On the seventh day most
larvae have developed into pupae and the containers are removed
from the frame to be sex sorted (Supplementary Video 1). At maxi-
mum capacity and high rearing density, the LRS is capable of pro-
ducing over 2,950,000 male pupae per week.
The LRS produced remarkably consistent numbers of synchro-
nous, similarly sized pupae from each rearing container. To visualize
the consistency of production, we calculated the daily yield of male
mosquitoes over 179 production batches during our 2018 field trial.
Yield was calculated as the proportion of L1 male larvae that devel-
oped into adult males and passed through the visual sex-sorting
pipeline (see below). The LRS showed high temporal consistency
Temperature (°F)
Average number females per trap
Total rain (mm)
2017 2018
Avg temp (min. − max.)
Control area females
Fig. 1 | Cytoplasmic incompatibility and mosquito seasonality in Fresno County. a, The outcome of mating between males and females with (green
shading) and without Wolbachia (no shading). Infected females (bottom) always lay viable Wolbachia-infected eggs. Uninfected females (top) lay
viable, uninfected eggs when mated with uninfected males, but lay inviable eggs when mated with Wolbachia-infected males. b, Red plot, average daily
temperature at Fresno Yosemite International Airport with shading indicating minimums and maximums during 2017 and 2018. Gray plot, average number
of females per trap night scaled according to the right y axis with 95% CIs shaded (n= 38 independent trap samples per collection day in 2017 and
n= 28 in 2018). The bottom plot indicates total daily rainfall (mm) during 2017 and 2018. Data retrieved from
NATure BIoTeChNology
with an average yield of 70.39% (Fig. 2b). In addition, adult male
size, estimated from the body length of male mosquitoes, was also
highly consistent throughout the 6 months of production, averag-
ing 3.8 mm (σ2 = 0.9 mm; Fig. 2c). A. aegypti is sexually dimorphic
for pupal size under favorable conditions38. Our rearing protocols
(Supplementary Text) implemented on the LRS produced consis-
tent pupal sizes resulting in clear separation between the sexes, with
female pupae on average 19.26% larger than male pupae (Fig. 2d),
consistent with optimized larval development39.
Mosquito sex sorting. To minimize the chances of unintentional
female release, we developed an automated, multi-step sex separa-
tion process based on known morphological differences between
males and females (Fig. 3a). The first step is an automated mechani-
cal sieve that separates based on body size, allowing male pupae to
pass through while females are retained. Over the course of 2018
production, an average of 2.54% of pupae that passed through the
sieve were females. Assuming a 50/50 input pupal sex ratio, we
estimated that automated mechanical sieving removed 94.92% of
females (Fig. 3b).
In the second step, the primarily male pupae that passed through
the sieve are loaded onto a real-time visual sex-sorter where they
eclose and — of their own volition — walk down a narrow path
over which a camera is mounted (Supplementary Video 2). Custom
industrial vision software recognizes each ambulatory mosquito as
an object, attempts to physically isolate them using air jets and a
shutter, and then takes at least one image. If multiple mosquitoes
make it into the imaging area they are always rejected. Images with a
single mosquito are inspected for male-specific body parts (Fig. 3c),
including genitalia and antennal features, using a template matching
Proportion of pupae
Pupal size (
Incubated storage and
retrieval robot
(1,540 containers)
Larval container
Larval container
transfer arm
transfer arm
Sieve station
(1 of 2)
Male Female
050 100 150
Batch number
Mean male yield from
L1 larvae
050100 150
Batch number
Mean male length (µm)
Fig. 2 | An automated LRS. a, Schematic of the LRS with major components labeled. b, Optimized rearing protocols resulted in a highly consistent yield,
calculated as the ratio of adult males entering the release tubes relative to the number of L1 larvae introduced into larval containers. The dark line shows mean
yield, shading represents the s.d., the x axis represents all 2018 production batches (n= mean of 96, range of 10–140 independent sex-sorter measurements
per batch). c, Consistent mean length of adult males as measured from sex-sorter images (s.d. interval shaded, n= mean of 43,256, range of 5,827–64,282
independent male length measurements per batch). d, Discrete pupal size dimorphism between sexes. Histogram shows width estimates from ~18,000 pupae.
Pupal width is measured in pixels resulting in bins when converted to μm. Red lines show normal distribution fit to male and female sets separately.
Articles NATure BIoTeChNology
algorithm. Individuals with male morphology are puffed into a
container used to distribute mosquitoes in the field, called a ‘release
tube’, while individuals failing inspection are rejected. At the start
of the 2018 field season, an average of 89.85% of males passed
inspection (Fig. 3d). After implementing improved traffic man-
agement algorithms to better isolate individuals, 95.59% of males
passed inspection, resulting in consistently high male yield through
the adult sex sorters.
In the third step, we submit all images of individuals labeled male
by the industrial vision system for scoring by a machine learning
classifier. The classifier is a deep neural network built upon the open
source Inception-v3 architecture40 and trained using 2.1 million man-
ually labeled images. The classifier computes the probability that the
individual is male and the images are ranked based on their maleness
score, subsampled, and sent to a panel of five trained, but non-expert,
reviewers via an online micro-task platform for inspection and label-
ing. We sent two samples for review: the 1% of images with the lowest
male probability, and a 1% random sample of all male images. If the
non-experts identified a female or if there was any inconsistency in
their labels, an expert reviewed the images in question. If the expert
confirmed any females, we located and purged the part of the release
tube with the female before the tube left the factory.
Based on data from 2018, we estimated the probability of
a female contaminant at each step of the sex-sorting pipeline
(Fig. 3a). Assuming independence between the different steps in
the pipeline, the combined system is expected to release 1 female
for every 900 million males with a 95% CI of 1:200 million to 1:26
billion (Fig. 3a and Supplementary Text). For additional validation
of the sex-sorting pipeline, we screened larvae obtained from ovit-
raps in our treatment areas and found no Wolbachia-positive larvae
(Supplementary Text), confirming that we did not unintentionally
establish a Wolbachia-infected population in the field as would be
expected if we released infected females into an area in which the
wild-type population had been suppressed.
Automated male mosquito releases. For a SIT intervention to
be successful, released males must permeate the landscape to find
unmated females. We developed an automated male mosquito
release system to ensure complete and calibrated distribution of
Wolbachia-infected males into treatment areas. The system includes
transport and release tubes, automated release devices mounted
inside customized vans (Fig. 4a), map-based release plan generation
and triggering software (Fig. 4b), and a structured light mosquito
counter (Supplementary Text).
050 100 150
Batch number
Fraction of males
passing inspection
050 100 150
Batch number
Fraction male after sieve
Male Female
vision Non-experts
Pupal sieve
of a female
(95% CI)
(0.00 – 0.08)
2.39 × 10–5
(1.30 × 10–5) – (3.79 × 10–5)
1.82 × 10–3
(3.76 × 10–4) – (4.38 × 10–3)
1.13 × 10–4
(3.08 × 10–9) – (4.83 × 10–4)
1.13 × 10–9
(3.88 × 10–11) – (4.44 × 10–9)
Fig. 3 | Sex-sorting pipeline. a, Illustration of the entire sex-sorting pipeline, including the mechanical pupal sieve, real-time adult visual inspection,
cloud-based machine learning classifier, and non-expert review. The probability of a female contaminant with 95% CIs for each step is shown along
with the estimated overall female contamination rate for the entire pipeline in the final column. b, The fraction of mosquitoes imaged by the sex
sorter after the pupal sieve that were male with s.d. intervals shaded for 179 production batches. c, Example images from the adult sex sorter (male
on the left and female on the right) used by both the industrial vision system and machine learning classifier. d, The fraction of true males that were
correctly labeled and accepted by the Industrial Vision system with s.d. interval shaded (n= mean of 96, range of 10–140 independent sex-sorter lane
measurements per batch).
NATure BIoTeChNology
After a preliminary study in 2017 (see Supplementary Text for
details), starting on 16 April 2018 we conducted daily releases of
Wolbachia-infected A. aegypti males over a period of 26 weeks into
3 treatment sites (labelled T1, T2, and T3 in Fig. 4c), which include
3,063 households across 293 ha (Supplementary Table 1). These
sites were residential neighborhoods typical of the area, situated
on the edge of the Fresno-Clovis metropolitan area with at least
partial isolation, and were known to have established A. aegypti
populations based on historical trapping data. We measured adult
mosquito density using BG-Sentinel traps (V2, Biogents) placed at
comparable densities in both treatment and control sites (Fig. 4d
and Supplementary Table 1).
In 2018, we released Wolbachia-males at an average rate of
78,469 per day or 267.81 (σ = 61.16) males per hectare per day for
a total of 14,376,511 male mosquitoes during the study, although
release rates differed between sites according to both household
counts and the number of females in traps in each site (Fig. 5a–c and
Supplementary Table 2). We also varied release rates per site within
the three study phases. In phase I (mid-April to mid-May), release
numbers were determined exclusively by the number of households
1,500 m
C1 BufferT1
400 m
40 m
Fig. 4 | Field sites and automated releases. a, Wolbachia-infected males were released into field sites using two sprinter vans equipped with automated
release devices that blew males through release outlets on the rear passenger side. b, Release map indicating a planned route for van drivers to follow,
triggering the release device. Each orange vector indicates the GPS location and direction of travel at which a segment of the release tube was released.
c, Map of treatment areas (T1–T3) in shades of orange and control areas (C1–C3) in shades of yellow. The C1 control buffer area is shaded in purple, and
the T2 treatment buffer is shaded in turquoise. These areas were monitored but not included in analysis. Only the T2 treatment buffer was treated with
sterile males. d, Representative placement and density of adult BG-Sentinel traps (black dots) and egg traps (gray dots). Trap density was similar between
treatment and control areas (Supplementary Table 1).
Articles NATure BIoTeChNology
Mean number males
per trap
Collection date
T1 T2 T3
10 m
100 m
Mosquitoes per day per m2
0.01 0.02 0.03 0.04
Number males released
Release date
Phase I Phase II Phase III Phase IPhase II Phase III
550 m 440 m 440 m
Fig. 5 | Releases of Wolbachia-infected males. a, Stacked sum plot showing the total number of males released into each treatment area over the 6-month
study period (see the main text for a description of study phases). We released males every day, except for three pauses for US holidays. b, Mean number
of adult males per trap in treatment areas (T1–T3) and control areas (C1–C3) with 95% bootstrap CIs (nT1= 44, nT2= 24, nT3= 35, nC1= 17, nC2= 28, nC3= 15
independent trap samples per collection day). Dotted lines indicate the first and last day of releases. c, Top, satellite maps with treatment areas outlined
in white and treatment buffers indicated with B in T2. Middle, density of Wolbachia-males, as measured by the onboard structured light mosquito counter,
averaged over 6 months of releases assuming a 10-m dispersal kernel revealing van path and variable release rate based on house density. Bottom, density
of Wolbachia-males averaged over 6 months of releases assuming a 100-m dispersal kernel, suggesting nearly complete coverage of release areas.
NATure BIoTeChNology
in each site. In phase II (mid-May to late July), we increased the
number of males per household for treatment sites T2 and T3, as
historical data indicated that these sites have had higher wild mos-
quito densities and are less geographically isolated than T1. In phase
III (late July to mid-October), we held the release numbers constant
for sites T2 and T3, but reduced the T1 release rate in response to
our monitoring data, which indicated very high ratios of Wolbachia-
infected to wild-type males in this site (Supplementary Table 3).
We monitored male mosquito densities using adult mosquito traps
and found that male mosquito numbers reflected the different
phases of release (Fig. 5b).
We visualized the density of released males in each neighborhood
over the entire 2018 season using data from the van-mounted release
device and mosquito counter. First, we modeled the density of released
males assuming a 10-m dispersion kernel around GPS (global posi-
tioning system) release coordinates, which shows the van release route
and highlights variations in release rate due to changes in housing
density (Fig. 5c). Importantly, if we assume a more realistic disper-
sion kernel of 100 m, males are more evenly distributed across each
site, suggesting comprehensive coverage. The only two relatively low-
density spots (blue regions) correspond to a large elementary school
in the center of T1 and a low-housing-density section of T3 (Fig. 5c).
To evaluate the precision of releases, we compared our intended mos-
quito distribution targets (based on housing density) to a map of actual
mosquito release density assuming a 100-m male dispersal kernel. The
density of male distribution after releases largely matches the intended
distribution and captures the reduction and increase in release rates in
T1 and T3, respectively (Supplementary Fig. 1).
Suppression of mosquito populations in release sites. The goal of
field releases was to test whether a high ratio of Wolbachia-infected
males to wild-type males would result in enough incompatible mat-
ings to sharply reduce egg hatch and subsequently the wild-type
adult population. To best isolate the effect of the Wolbachia-male
releases, only normal mosquito abatement activity under the man-
date of the California Mosquito Abatement District (CMAD) was
applied in the treatment and control areas (Supplementary Table 4
and Supplementary Text).
We monitored the ratio of released to wild-type males (that is,
overflooding ratio) by testing trapped adult males for Wolbachia
using a loop-mediated isothermal amplification (LAMP) assay
(Supplementary Text) and found that our releases resulted in high
overflooding ratios in each of the treatment sites during the first
4 months of release, ranging from 47.53 to 557.00 (Supplementary
Table 3). As overflooding ratios reached levels too high to be esti-
mated reliably, we did not measure these for the last 2 months of
releases. The overflooding ratios tended to increase month after
month, consistent with both increased release rates in T2 and T3
during phase II and declines in the number of wild-type males per
trap (Supplementary Table 3).
We also monitored the abundance of adult females using
BG-Sentinel traps (Fig. 4d) and found that the density of adult
females differed significantly between treatment and control areas
during the treatment period. In each control area, the average
number of females per trap night followed the expected seasonal
curve, with the population increasing in June, peaking from July to
September with female densities of >12 females per trap in each site,
and declining in October (Fig. 6a). In contrast, female abundance
in the treatment sites had a strikingly different pattern (Fig. 6a,b).
T1, the most isolated site, had extremely low numbers of females in
all weeks, peaking at an average of only 0.6 females per trap in the
third week of October. Although sites T2 and T3 had more females
than T1 as the season progressed, with peak mean females per trap
of 1.52 and 2.17, respectively, the 95% confidence intervals (CIs)
are fully separated from those of the control sites from mid-July to
mid-November (Fig. 6a).
When comparing female abundance between aggregated treat-
ment and control sites, there is a clear separation between the 95%
CIs beginning approximately 5 weeks after the start of releases
(Fig. 6c). The average number of females in aggregated treatment
sites remained low for the entirety of the season with less than one
female per trap night in 32 out of 36 weekly collections and a peak
value of 1.2 females per trap night (95% CI, 0.78–2.47) in the third
week of October. In comparison, the control sites reached a peak of
16.6 females per trap (95% CI, 13.70–19.87) in the second week of
September (Fig. 6c).
Overall, release of Wolbachia-infected males into treatment areas
resulted in 93.64% (corrected P = 1.6 × 10–5) suppression of females
from mid-July until the seasonal declines starting in mid-Octo-
ber, with a maximum 2-week suppression level of 95.5% (95% CI,
93.6–96.9%) in the fourth week of July (Fig. 6d). To test the general-
ity of these results, we compared each treatment site individually to
both the aggregate and individual control sites and found that sig-
nificant suppression was achieved in all sites across the 14 weeks of
peak mosquito season in all pairwise comparisons (Fig. 6e). Moreover,
we found that, within 2-week windows, T1 reached a peak suppression
of 98.9% (95% CI, 98.1–99.4), T2 reached 94.8% (95% CI, 92.3–96.8),
and T3 reached 94.6% (95% CI, 92.0–96.4) compared to the aggregate
control site. Results are similar when each treatment site is compared
to individual control sites (Supplementary Table 5). We also com-
pared female abundance in T1 in 2018 with that in 2017 (Fig. 6f and
Supplementary Text), which showed a 97.1% drop in the number of
mosquitoes from 2017 to 2018 (95% CI, 95.4–98.6).
Comparison of the number of larvae hatching from egg traps in
treatment sites relative to control sites provides an additional view
of the effect of Wolbachia-male releases on mosquito reproduction.
We directly monitored larval production using egg traps distrib-
uted at comparable densities in both treatment and control sites
(see Methods, Fig. 4d and Supplementary Table 1). For the entire
season, the mean number of cumulative larvae collected per egg
trap in treatment sites was 3.7 (95% CI, 0.5–8.4) compared to 126.3
(95% CI, 80.3–180.7) in control sites — a 97.1% reduction in col-
lected larvae (Fig. 6g). Similarly, the mean number of eggs per trap
was consistently lower in the treatment areas than in control areas
(Supplementary Fig. 2). To infer the proportion of incompatible
versus wild-type matings, we also calculated hatch rates of collected
eggs. Although variable owing to small sample size, hatch rates of
eggs collected in treatment sites were consistently lower than those
collected from control areas (Supplementary Fig. 3). Taken together,
the data demonstrate that Wolbachia-infected males inhibited mos-
quito reproduction, resulting in strong suppression of the wild pop-
ulation in release sites.
Mosquito migration into release sites. Despite treatment site
selection intended to minimize migration through geographic iso-
lation and treated buffer areas, several lines of evidence suggest that
immigration of inseminated females from nearby untreated areas
put an upper limit on achievable suppression. Although statistical
support is limited by small sample sizes, more females were caught
in traps on the outer edge of treatment sites (T2 and T3), as indi-
cated by a negative correlation between the distance of a trap from
the edge of the site and the average number of females it collected,
whereas only one of the control sites (C2) showed this pattern
(Fig. 7a–c and Supplementary Table 6). In addition, we used the
LAMP assay to test for Wolbachia-infected males in traps from the
buffer area separating T1 and C1 as well as traps within C1 (Fig. 4d).
Unsurprisingly, we found Wolbachia-positive males in large num-
bers up to 200 m from the nearest release street in T1 (Fig. 7d),
clearly demonstrating that our treatment sites were within the
flight range of mosquitoes in untreated areas. Overall, the data are
consistent with ‘edge effects’ driven by female mosquito migration
into our treatment sites.
Articles NATure BIoTeChNology
Mean number females per trap
Collection date
Treatment areas
Mean number females per trap
Collection date
Mean number females per trap
Collection date
Suppression (%)
Collection date
Cumulative mean
number of larvae
Collection date
Treatment Control
C1 C2 C3
92.70*** 94.11***
87.25*** 89.20*** 91.29***
Mean number females per trap
Collection date
T1 (2017)
T1 (2018)
Control areas
Fig. 6 | Wild female and larvae counts from field sites. a, Mean number of females per trap in treatment areas (T1–T3) and control areas (C1–C3) in 2018
(nT1= 44, nT2= 24, nT3= 35, nC1= 17, nC2= 28, nC3= 15 independent trap samples per collection day). b, Mean number of females per trap in treatment areas
only, on shortened y axis (sample sizes are the same as in a). c, Mean number of females per trap for aggregated treatments sites and aggregated control
sites (nTRT= 103, nCTRL= 60 independent trap samples per collection week). Gray bar, period defined as ‘peak season’. d, Per cent suppression of adult
females in aggregate treatment sites compared to aggregate control sites using a 2-week trailing average (see Supplementary Text for details; sample sizes
are the same as in c). e, Suppression calculated across the 14-week peak-season window evaluated with a one-sided permutation test (nT1= 616, nT2= 336,
nT3= 490, nC1= 238, nC2= 392, nC3= 210, nTRT= 1,442, nCTRL= 840 independent trap samples). ***P< 1.6 × 105, Bonferroni-corrected. f, Year-on-year
comparison of the average number of females in T1; comparison of the same period in 2018 and in 2017 (n2017= 65, n2018= 44 independent trap samples
per collection day). g, Cumulative mean number of larvae per egg trap with treatment sites aggregated and control sites aggregated (nTRT= 131, nCTRL= 77
independent trap samples per collection week), indicating significantly different larval production between treatment and control areas. For all panels,
shaded areas indicate 95% CIs, and dotted lines indicate first and last day of releases.
NATure BIoTeChNology
The higher numbers of females collected at the edges of treat ment
sites are mainly due to a small number of ‘hot’ traps that collected
five or more females per trap collection (Fig. 7a). We sequenced
individual female genomes from ‘hot’ traps and determined the
relatedness among the sampled females (Supplementary Table 7 and
Supplementary Text). Consistent with ‘hot’ traps being driven by
nearby oviposition from inseminated female migrants, females in
T2 and T3 ‘hot’ traps had high relatedness with an average per-trap
T1 T2 T3
C1 C2 C3
0 500 1,000 1,500
Cumulative proportion recaptured
Distance from nearest release street (m)
Number of released males
AUG 02
AUG 02
AUG 02
JUL 10
AUG 07
JUL 31
JUL 31
JUL 17
JUL 11
AUG 09
JUL 25
JUL 18
JUL 11
JUL 17
JUL 17
02 6128410
Full sibUnrelated
Mean number
of females
Mean number
of females
0 100 200 300 400
Distance from edge (m)
Mean number of
T1 (−0.17)
T2 (−0.32)
T3 (−0.29)
0 100 200 300 400
Distance from edge (m)
Mean number of
C1 (0.17)
C2 (−0.24)
C3 (0.02)
Fig. 7 | Evidence for female migration into treatment areas. a, Maps of treatment areas with a 100-m radius around each adult trap. Shading corresponds
to the mean number of females per trap in 2018. b, Maps of control areas as in a, but on a different scale (as shown on the right). c, Top panel, correlations
between mean number of females per trap in 2018 and distance from the nearest edge of treatment areas, with colors corresponding to the treatment
area. Bottom panel, the same correlation in control areas. In both panels, Pearson’s r is shown in the legend for each comparison. d, Dot plot showing the
number of Wolbachia-males released in T1 and recaptured in traps in C1 on four dates during releases. The x axis shows the distance between the trap
and the nearest release street in T1. The line shows the cumulative number of recaptured Wolbachia-males on the right y axis as a function of the same
distance. e, Heatmaps showing genetic relatedness within trap collections from adult traps in control areas (first row), egg traps in control areas (second
row), and high-female-count adult traps in treatment areas, with relatedness calculated based on HETHET/IBS0 relatedness scores, ranging from 0 to 12
(Supplementary Text). Each subpanel, labeled with trap ID and collection date, summarizes a series of pairwise comparisons between females, and the
color of the tile indicates the degree of relatedness according to the scale below.
Articles NATure BIoTeChNology
rate of sibship of 0.47 and 0.10, respectively (Supplementary Table 7
and Fig. 7e). Similarly, larvae collected from egg traps had an aver-
age per-trap rate of sibship of 0.6. By contrast, females from C2
and C3 collections have very low rates of sibship per trap (0.00,
Supplementary Table 7 and Fig. 7e), suggestive of many unrelated
larval production sites (Fig. 7e). Although some larval production
in treatment areas may have resulted from virgin female migrants
finding a fertile mate within the treatment area or from local
females evading released males, the available evidence suggests that
most production is due to inseminated females migrating into the
treatment areas and ovipositing.
Our results demonstrate that efficient production of incompatible
Wolbachia-infected males using automated systems enables strong
suppression of wild populations of A. aegypti at scales larger than
previous trials that relied on manual rearing and release methods
(Supplementary Table 8). We achieved an estimated 95.55% (93.74–
96.97%) reduction in the wild adult mosquito population across
three replicate release sites. Suppression varied between treatment
sites, with our most isolated site, T1, reaching nearly 99% reduc-
tion, while T3 reached a maximum suppression level of nearly 95%
(Supplementary Table 5). One key difference between our treatment
sites is that we conducted a preliminary suppression trial in T3 in
2017 (Supplementary Text), which could have impacted the results
in 2018. However, we observed more females in T3 at the beginning
of the 2018 season than any other site (Fig. 6b) and average suppres-
sion was lower in T3 (Fig. 6e), indicating that the 68% suppression
achieved in 2017 was not sufficient for multi-year impact. Indeed,
the largest differences in female densities between treatment sites
developed later in the season (Fig. 6b), suggesting that immigration
was a primary driver of between-treatment-site variation.
Despite maintaining very high overflooding ratios of Wolbachia
males (>45 Wolbachia to 1 wild-type, Supplementary Table 3) in
each treatment site, we were unable to achieve local elimination,
probably due to migration of wild-type females from untreated
areas. Increasing the size of release zones in future treatments
should enable stronger suppression by buffering the effects of
immigration over a larger area and increasing the distance between
internal areas and edges. As a result, by treating larger areas and
minimizing the impact of migration, we should theoretically be able
to lower the number of males released per household by up to an
order of magnitude, bringing overflooding ratios in the field closer
to the minimum predicted to be effective by laboratory experi-
ments41,42. Assuming that the male release rate can be reduced by at
least half during a large-scale, phased rollout, we estimate that we
could strongly suppress the mosquito population across the entire
Fresno-Clovis metropolitan area (~250,000 households) in 3 years
using four automated larval rearing systems operating at full capac-
ity coupled with our sex-sorting and release technology.
A common criticism of SIT is the need for continual re-appli-
cation of males each season if the target population is not fully
eliminated. Close monitoring of our study sites in future years will
provide insight into how quickly A. aegypti populations rebound
after treatment and allow us to directly test whether lower release
numbers can sustain suppression in previously treated areas. We
expect the rate of re-infestation will depend on both the strength of
suppression in the treated area as well as the abundance and prox-
imity of nearby source populations. In addition, as recommended
by the World Health Organization, a cluster-randomized control
trial(s) would further validate the efficacy of our approach and, if
conducted in an area with Aedes-borne disease, could be used to
measure the impact of mosquito population suppression on arbo-
viral transmission. When first responding to an epidemic outbreak
of Aedes-borne disease, however, our study and others have shown
that SIT-based interventions take multiple weeks to begin reducing
mosquito numbers and thus should be combined with more fast-
acting abatement techniques.
In this study, automation enabled unprecedented consistency in
larval rearing, accuracy during sex separation, and precision in mos-
quito release, allowing us to avoid pitfalls that limited the success or
scale of most previous mosquito SIT trials, such as uncompetitive
males, insufficient production yields, and high female contamina-
tion rates6,10,35. Residual adult female removal after mechanical pupal
sex-separation has been especially difficult to scale, but our highly
accurate automated sex-sorting pipeline and female sterilization by
low-dose pupal irradiation29,30 both solve this problem, enabling SIT
to be effective in large-scale suppression of wild populations. We
expect that continued improvements in mosquito production, sepa-
ration, and release technologies will increase performance and effi-
ciency. The results described here support the prospect of removing
invasive A. aegypti populations from large swaths of land without
the use of chemical insecticides, aiding the ongoing public health
battle against A. aegypti.
Online content
Any methods, additional references, Nature Research reporting
summaries, source data, extended data, supplementary informa-
tion, acknowledgements, peer review information; details of author
contributions and competing interests; and statements of data and
code availability are available at
Received: 16 February 2019; Accepted: 27 February 2020;
Published: xx xx xxxx
1. Powell, J. R., Gloria-Soria, A. & Kotsakiozi, P. Recent history of Aedes aegypti:
vector genomics and epidemiology records. Bioscience 68, 854–860 (2018).
2. Engelthaler, D. M., Fink, T. M., Levy, C. E. & Leslie, M. J. e re-emergence
of Aedes aegypti in Arizona. Emerging Infect. Dis 3, 241–242 (1997).
3. Metzger, M. E., Hardstone Yoshimizu, M., Padgett, K. A., Hu, R.
& Kramer, V. L. Detection and establishment of Aedes aegypti and
Aedes albopictus (Diptera: Culicidae) mosquitoes in California, 2011-2015.
J. Med. Entomol. 54, 533–543 (2017).
4. Kraemer, M. U. G. etal. e global compendium of Aedes aegypti and
Ae. albopictus occurrence. Sci. Data 2, 150035 (2015).
5. Soper, F. L. e elimination of urban yellow fever in the americas through
the eradication of Aedes aegypti. Am. J. Public Health Nations Health 53,
7–16 (1963).
6. Klassen, W. & Curtis, C. F. in Sterile Insect Technique (eds. Dyck, V. A.,
Hendrichs, J. & Robinson, A. S.) 3–36 (Springer-Verlag, 2005).
7. Krafsur, E. S., Whitten, C. J. & Novy, J. E. Screwworm eradication in North
and Central America. Parasitol. Today 3, 131–137 (1987).
8. Carvalho, D. O. etal. Suppression of a eld population of Aedes aegypti in
Brazil by sustained release of transgenic male mosquitoes. PLoS Negl. Trop.
Dis. 9, e0003864 (2015).
9. Gorman, K. etal. Short-term suppression of Aedes aegypti using genetic
control does not facilitate Aedes albopictus. Pest Manag. Sci. 72,
618–628 (2016).
10. Harris, A. F. etal. Successful suppression of a eld mosquito population
by sustained release of engineered male mosquitoes. Nat. Biotechnol. 30,
828–830 (2012).
11. Morlan, H. B., McCray, E. M.Jr. & Kilpatrick, J. W. Field tests with sexually
sterile males for control of Aedes aegypti. Mosquito News 22, 295–300 (1962).
12. Bloss, C. S., Stoler, J., Brouwer, K. C., Bietz, M. & Cheung, C. Public response
to a proposed eld trial of genetically engineered mosquitoes in the United
States. JAMA 318, 662–664 (2017).
13. Culbert, N. J. etal. Longevity of mass-reared, irradiated and packed male
Anopheles arabiensis and Aedes aegypti under simulated environmental eld
conditions. Parasit. Vectors 11, 603 (2018).
14. Yen, J. H. & Barr, A. R. New hypothesis of the cause of cytoplasmic
incompatibility in Culex pipiens L. Nature 232, 657–658 (1971).
15. Hilgenboecker, K., Hammerstein, P., Schlattmann, P., Telschow, A. &
Werren, J. H. How many species are infected with Wolbachia?–a statistical
analysis of current data. FEMS Microbiol. Lett. 281, 215–220 (2008).
16. Gloria-Soria, A., Chiodo, T. G. & Powell, J. R. Lack of evidence for natural
Wolbachia infections in Aedes aegypti (Diptera: Culicidae). J. Med. Entomol.
55, 1354–1356 (2018).
NATure BIoTeChNology
17. Ross, P. A. etal. An elusive endosymbiont: does Wolbachia occur naturally in
Aedes aegypti? Ecol. Evol. (2020).
18. Xi, Z., Khoo, C. C. H. & Dobson, S. L. Wolbachia establishment and invasion
in an Aedes aegypti laboratory population. Science 310, 326–328 (2005).
19. McMeniman, C. J. etal. Stable introduction of a life-shortening Wolbachia
infection into the mosquito Aedes aegypti. Science 323, 141–144 (2009).
20. Walker, T. etal. e wMel Wolbachia strain blocks dengue and invades caged
Aedes aegypti populations. Nature 476, 450–453 (2011).
21. Ant, T. H. & Sinkins, S. P. A Wolbachia triple-strain infection generates
self-incompatibility in Aedes albopictus and transmission instability in Aedes
aegypti. Parasit. Vectors 11, 295 (2018).
22. Helinski, M. E. H., Parker, A. G. & Knols, B. G. J. Radiation biology of
mosquitoes. Malar. J. 8 (Suppl. 2), S6 (2009).
23. Bargielowski, I., Kaufmann, C., Alphey, L., Reiter, P. & Koella, J. Flight
performance and teneral energy reserves of two genetically-modied and one
wild-type strain of the yellow fever mosquito Aedes aegypti. Vector Borne
Zoonotic Dis. 12, 1053–1058 (2012).
24. Yamada, H., Parker, A. G., Oliva, C. F., Balestrino, F. & Gilles, J. R. L.
X-Ray-induced sterility in Aedes albopictus (Diptera: Culicidae) and male
longevity following irradiation. J. Med. Entomol. 51, 811–816 (2014).
25. Facchinelli, L. etal. Field cage studies and progressive evaluation of
genetically-engineered mosquitoes. PLoS Negl. Trop. Dis. 7, e2001 (2013).
26. Zabalou, S. etal. Incompatible insect technique: incompatible males
from a Ceratitis capitata genetic sexing strain. Entomol. Exp. Appl. 132,
232–240 (2009).
27. Axford, J. K., Ross, P. A., Yeap, H. L., Callahan, A. G. & Homann, A. A.
Fitness of wAlbB Wolbachia infection in Aedes aegypti: parameter estimates
in an outcrossed background and potential for population invasion.
Am. J. Trop. Med. Hyg. 94, 507–516 (2016).
28. Hancock, P. A., White, V. L., Ritchie, S. A., Homann, A. A. & Godfray, H. C. J.
Predicting Wolbachia invasion dynamics in Aedes aegypti populations using
models of density-dependent demographic traits. BMC Biol. 14, 96 (2016).
29. Zheng, X. etal. Incompatible and sterile insect techniques combined
eliminate mosquitoes. Nature 572, 56–61 (2019).
30. Kittayapong, P. etal. Combined sterile insect technique and incompatible
insect technique: the rst proof-of-concept to suppress Aedes aegypti vector
populations in semi-rural settings in ailand. PLoS Negl. Trop. Dis. 13,
e0007771 (2019).
31. Breeland, S. G., Jeery, G. M., Lofgren, C. S. & Weidhaas, D. E. Release of
chemosterilized males for the control of Anopheles albimanus in El Salvador. I.
Characteristics of the test site and the natural population. Am. J. Trop. Med.
Hyg. 23, 274–281 (1974).
32. Weidhaas, D. E., Breeland, S. G., Lofgren, C. S., Dame, D. A. & Kaiser, R.
Release of chemosterilized males for the control of Anopheles albimanus in
El Salvador. IV. Dynamics of the test population. Am. J. Trop. Med. Hyg. 23,
298–308 (1974).
33. Dame, D. A. etal. Release of chemosterilized males for the control of
Anopheles albimanus in El Salvador. II. Methods of rearing, sterilization, and
distribution. Am. J. Trop. Med. Hyg. 23, 282–287 (1974).
34. Breeland, S. G. etal. Release of chemosterilized males for the control
of anopheles albimanus in El Salvador. Am. J. Trop. Med. Hyg. 23,
288–297 (1974).
35. Dame, D. A., Lowe, R. E. & Williamson, D. L. Assessment of released sterile
Anopheles albimanus and Glossina morsitans morsitans. In Cytogenetics and
Genetics of Vectors: Proceedings of a Symposium of the XVIth International
Congress of Entomology (eds Pal, R. etal.) 231–248 (Elsevier, 1981).
36. Laven, H. Eradication of Culex pipiens fatigans through cytoplasmic
incompatibility. Nature 216, 383–384 (1967).
37. Pless, E. etal. Multiple introductions of the dengue vector, Aedes aegypti, into
California. PLoS Negl. Trop. Dis. 11, e0005718 (2017).
38. Clements, A. N. e Biology of Mosquitoes Vol. 1, 175–183 (CABI, 1992).
39. Medici, A. etal. Studies on Aedes albopictus larval mass-rearing optimization.
J. Econ. Entomol. 104, 266–273 (2011).
40. Szegedy, C., Vanhoucke, V., Ioe, S., Shlens, J. & Wojna, Z. Rethinking the
inception architecture for computer vision. Preprint at
abs/1512.00567 (2015).
41. Mains, J. W. e Endosymbiont Wolbachia in Aedes Mosquitoes:
Characterization of Host Interactions and Population Control Implications.
PhD thesis, University of Kentucky (2012).
42. Asman, S. M., McDonald, P. T. & Prout, T. Field studies of genetic control
systems for mosquitoes. Annu. Rev. Entomol. 26, 289–318 (1981).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature America, Inc. 2020
Articles NATure BIoTeChNology
Mosquito strains. In 2018, we released male mosquitoes from two colonies of
the WB1 strain of A. aegypti with a Fresno-Clovis genetic background. In each
case, Wolbachia-infected females were crossed en masse with eld-derived wild-
type Fresno and Clovis males (collected and colonized in the Summer of 2017)
for at least four generations (Supplementary Text). Each backcrossed strain was
conrmed to have 100% incompatibility when mated with wild-type females
sourced from Fresno and Clovis. We conrmed genetic similarity between
backcross colonies and wild-type Fresno-Clovis colonies using genome sequencing
(Supplementary Text and Supplementary Fig. 4). e rst colony (denoted
WB1-CL4-BC4) was released on 148 days of the trial, while the second colony
(WB1-CL5-BC5) was released on 70 days (Supplementary Table 2).
Mass rearing. For egg production, approximately 4,500 adults at a 1:1 sex ratio
were held in 60 cm3 Bugdorm cages at 75% relative humidity and 28 °C, and given
10% sucrose ad libitum. Cages were fed organic bovine blood warmed to 37 °C in
petri dishes covered with parafilm. Eggs were collected in soup cups lined with wet
seed germination paper, allowed to embryonate, and then stored for up to 1 month
prior to hatching. For larval production, eggs were scraped from germination
papers, weighed, and 0.5 g of eggs were hatched in 0.15 optical density E. coli
(DH5a) broth. L1 larvae were automatically counted into batches of 1,500, 2,000, or
3,000 and transferred into a thermoformed container containing 1 liter of double-
distilled water, 40 ml of fermented bovine liver powder (fBLP) (MP Biomedicals),
activated carbon pellets (Imagitarium), and 0.5 g of bovine liver powder (BLP).
fBLP was made by allowing 4.5 g of BLP to ferment in a closed carboy containing
approximately 20 liters of water for 7 days. After loading onto the incubated rearing
frame, larvae were given three additional BLP feeds (Supplementary Text), and
removed after 6 days at 28 °C.
Field releases. After sex-sorting, males were transported from our rearing facility
at Verily in South San Francisco to Fresno and Clovis, in 6-inch-diameter release
tubes with 10% sucrose ad libitum, where they were held overnight for release the
next morning. Males ranging in age from 2 to 3 days old were released from the
side of customized vans typically between 6:00 and 11:00, 7 days per week for
26 weeks. See Supplementary Text for more details.
Study sites. We chose three communities in Fresno County for male mosquito
releases. They are almost exclusively residential neighborhoods within
incorporated cities, except for T1, which includes a community center and
elementary school. Treatment sites ranged in size, with T1 being the largest (1,563
households within 130 ha), followed by T3 (683 households within 89 ha), and T2
(665 households within 74 ha) (Supplementary Table 1). T2 is bordered on three
sides by other neighborhoods known to have established A. aegypti populations, so
although we treated and monitored 74 ha, we designated buffers on the northern,
eastern, and southern borders, leaving a core area of 44 ha designated as the core
treatment area (see Fig. 4c) for all subsequent analyses. Treatment site T3 was
somewhat disconnected from other residential areas and bordered on most sides
by either a road or open space and residential areas, so no buffer areas were treated
around this site.
We also monitored three geographically matched control areas. Although
smaller than the release areas in overall size (Supplementary Table 1), the control
areas were almost exclusively residential and very similar to the release areas with
respect to housing density and landscape. One control site, C1, is adjacent to the
T1 treatment site. Although we monitored the entire C1 site, we excluded from
downstream analysis traps in a buffer region (Fig. 4d) (size defined as three times
the expected average flight range of this species, or approximately 300 m, of the
edge of T1) to minimize any confounding effects of Wolbachia-males dispersing
into this site in appreciable numbers (Fig. 4d).
Treatment areas were chosen based on several criteria: (1) the degree of
isolation from untreated areas; (2) historical trapping data indicating establishment
of A. aegypti; and (3) how well the area represented typical landscape in Fresno
County. Control areas were chosen based on the same criteria, except that criterion
1 was relaxed given that the number of sites that fitted this criterion was small.
Assignment of each site as control or treatment was not randomized, but we
believe that any potential bias associated with site assignment would be negligible
compared to the effect size observed in comparisons between treatment and
control areas.
Field monitoring. All mosquito field monitoring was conducted by CMAD staff
using protocols developed in collaboration with Verily and MosquitoMate. Following
consent from residents, adult BG-Sentinel (v2, Biogents) and custom-made egg traps
were placed in front yards at residences thought to be preferred by A. aegypti based
on physical characteristics of the yard. Trap density was similar (Supplementary
Table 1) between treatment and control areas. Treatment and control areas were
paired such that pairs were always collected on the same day. Adult trap data from
treatment and control sites can be found in Supplementary Table 9.
Statistics. We summarized trap counts on a weekly basis, but we only included
weeks with valid collection data from greater than 75% of traps in a site to
minimize fluctuations resulting from small sample sizes. Non-parametric 95%
bootstrap CIs were calculated by taking 1,000 bootstrap samples with replacement
of all valid trap collections for a week within a site for site-wise statistics, or
samples of all valid trap collections across the merged site classes for the aggregate
statistics. We calculated means from each bootstrap sample and found the 2.5%
and 97.5% quantiles of the sorted distribution. Target sample sizes were nT1 = 44,
nT2 = 24, nT3 = 35, nC1 = 17, nC2 = 28, nC3 = 15 independent trap samples per
collection day, but trap problems led to slight reductions in sample size for some
collection days. For aggregate statistics, sample sizes vary and are specified in the
figure legends.
We calculated suppression as 1 (Ti/Ci,) where Ti is the 2-week trailing average
of all valid treatment site collections and Ci is the 2-week trailing average of all
valid control site collections. This formula is numerically identical to Abbott’s
formula43. As adult traps were baited with dry ice and collected twice per week in
2017 and once per week in 2018, the 2-week windows include four collections on
average in 2017, but only two collections in 2018. We calculated non-parametric
95% CIs as described above but within 2-week windows within each class
separately for suppression analysis, re-calculated the mean for each bootstrapped
sample and found the 2.5% and 97.5% quantiles of all bootstrap means for that
window. We chose to use 2-week trailing averages for comparison to reduce
emphasis on collection-to-collection fluctuations in the data.
To quantify the statistical power afforded by the 2018 trapping regime, we
conducted a power analysis for the 2018 trial calibrated using data from the control
areas during peak mosquito season. For this analysis, we assumed a three-level
hierarchical model with three clusters (representing control areas) in the trial
and an average of 21 traps per cluster. Each cluster is described by an N-mixture
model44, in which the underlying population is drawn from a Poisson distribution
and then the trap counts are drawn from a binomial distribution where the
parameter n is the Poisson draw and p is set to the trap efficiency. Previous work
suggests that the BG-Sentinel trap efficiency is approximately 10% (ref. 45).
To parameterize the Poisson distributions, we back-calculated the mean trap
counts from peak-season control area data assuming a trap efficiency of 10%.
Specifically, we defined ‘peak mosquito season’ as the window of time when the
mean number of females per trap in the aggregate control area exceeded 10, which
corresponds to 17 July to 19 October (gray bar in Fig. 6c). After accounting for trap
efficiency, the mean female trap count is 129 and the between-cluster variance was
1,695. To simulate each cluster, we drew lambda for the Poisson distribution from
a normal distribution (mean = 129, s.d. = 1,695), and each cluster was assumed to
be independent. We then simulated control area trap counts using Monte Carlo
sampling of this hierarchical model and treatment area trap counts using the same
approach but scaling lambda by the target suppression value. The suppression
values were then calculated and compared using bootstrapping with replacement
on the resulting simulated trap counts aggregated across clusters. This process
was repeated 1,000 times to calculate the power and p values of the simulated
experiment. The analysis shows that this suppression study has >80% power and
has a p value of <0.05 when there is at least 40% suppression observed (that is 80%
probability of recovering true positive result when suppression is at least 40%).
To explore differences between treatment and control sites, we made all
pairwise comparisons between the aggregate control and treatment sites, and all
treatment and control sites individually. The maximum suppression value within a
2-week window and 95% CIs are presented in Supplementary Table 5.
To determine whether the levels of suppression observed in our treatment
areas are significantly different from the null hypothesis of no suppression (that
is, no difference between treatment and controls), we applied a permutation test.
We calculated the observed level of suppression across the entire 14-week peak
mosquito season and compared this value to suppression calculated after randomly
permuting traps among sites. We compared the aggregate treatment site and all
individual treatment areas to the aggregate control sites as well as each control
area individually using one million permutations per comparison in a one-sided
test. In all cases, all permuted data sets produced levels of suppression less than the
observed value, corresponding to a Bonferroni-adjusted p value of 1.6 × 10–5.
After excluding collections with trap problems or traps in which the paper was
dry at the time of collection, we calculated aggregate egg hatch rates as the number
of larvae divided by the number of viable eggs. Hatch rate 95% CIs were calculated
by bootstrap sampling with replacement of egg papers that were positive for eggs.
We calculated the cumulative number of larvae per trap by cumulatively summing
the total number of larvae that hatched from all egg collections. Target samples
sizes were nT1 = 61, nT2 = 35, nT3 = 41, nC1 = 30, nC2 = 32, nC3 = 15 independent trap
samples per collection day, but trap problems led to slight reductions in sample size
for some collection days. We excluded collection days on which more than 25% of
trap collections were missing owing to trap problems.
Analyses were conducted using custom R46 and Python scripts.
Reporting Summary. Further information on research design is available in the
Nature Research Reporting Summary linked to this article.
Data availability
Adult count data from field traps analyzed in this study are included as
supplementary tables. Per-site male release numbers are also included as
NATure BIoTeChNology
supplementary tables. Genome sequencing data can be found under Bioproject
PRJNA600991 at NCBI. Training image data and the trained neural-net model
for male–female classification can be accessed by visiting
Code availability
Scripts for analysis of trap data are available upon request. Scripts for organizing
machine learning training data and conducting model training can be found at
43. Abbott, W. S. A method of computing the eectiveness of an insecticide.
1925. J. Am. Mosq. Control Assoc. 3, 302–303 (1987).
44. Royle, J. A. N-mixture models for estimating population size from spatially
replicated counts. Biometrics 60, 108–115 (2004).
45. Ritchie, S. A., Montgomery, B. L. & Homann, A. A. Novel estimates of
Aedes aegypti (Diptera: Culicidae) population size and adult survival based on
Wolbachia releases. J. Med. Entomol. 50, 624–631 (2013).
46. R Core Team. R: A Language and Environment for Statistical Computing
(R Foundation for Statistical Computing, 2017).
We are grateful to the residents of Fresno County for their support and participation in
the study. We are also thankful to the many staff members at Verily and Consolidated
Mosquito Abatement District for their contributions.
Author contributions
J.E.C., D.W.C, V.C., M.D., D.C., B.D., K.G., K.C.H., P. H., J.S.H., J.L., C.B., R.B., W.C.,
K.L.D., C.E., D.G., Y.H., B.H., E.K., J.K., A.K., E.L., T.L., J.L., M.L., W.M., J.W.M., M.M.,
S.N.M., D.M., J.R.O., K.P., A.P., C.R., M.S., R.S., P.S., J.S., J.S., B.W. A.M.W., M.W., J.W.,
A.Y., W.C.C., J.H., N.S., L.U., T.Z., S.L.D., F.S.M., P.M., and B.J.W. performed research.
J.E.C., D.W.C., W.C.C., J.H., L.U. S.L.D., F.S.M., P.M., and B.J.W. designed and supervised
research. J.E.C. and B.J.W. wrote the manuscript with editorial contributions from all
Competing interests
J.E.C., D.W.C, V.C., M.D., K.G., K.C.H., P. H., J.S.H., J.L., C.B., R.B., W.C., C.E., D.G.,
Y.H., B.H., E.K., J.K., A.K., E.L., T.L., J.L., M.L., W.M., M.M., S.N.M., D.M., J.R.O., K.P.,
A.P., C.R., M.S., R.S., P.S., J.S., J.S., B.W. A.M.W., M.W., J.W., A.Y., W.C.C., N.S., L.U.,
T.Z., P.M. and B.J.W. were paid employees of Verily Life Sciences, a for-profit company
developing products for mosquito control, at the time they performed research for this
study. K.L.D., J.W.M. and S.L.D. were paid employees of Mosquito Mate, a for-profit
company developing products for mosquito control, at the time they performed research
for this study.
Additional information
Supplementary information is available for this paper at
Correspondence and requests for materials should be addressed to J.E.C. or B.J.W.
Reprints and permissions information is available at
... It can manipulate mosquito reproduction through cytoplasmic incompatibility (CI) 6 , a phenomenon of conditional embryonic lethality that results from mating between a Wolbachia-infected male and a female that is either uninfected or infected by a different Wolbachia strain. CI offers the theoretical basis for the incompatible insect technique (IIT), in which inundative release of Wolbachia-infected males is used to induce sterile matings with wild-type females in the field, resulting in strong population suppression [7][8][9][10][11] . Given that some Wolbachia strains can induce pathogen blocking in mosquitoes, another Wolbachia-based vector control strategy referred to as population replacement involves the release of infected females to utilize CI for spreading Wolbachia into the target population to reduce the mosquito's ability to transmit dengue virus due to the advantage of infected females in reproduction compared to their uninfected counterparts. ...
... Notably, IIT combined with the irradiation-based sterile insect technique (SIT) has achieved strong suppression of A. albopictus populations on the islands of Guangdong Province, China 11,12 , as well as A. aegypti populations in Singapore 13 , Mexico 14 and Thailand 15 . The high efficacy of IIT to suppress A. aegypti has also been accomplished in the U. S. 9 , northern Australia 7 and Singapore 13 through the use of artificial intelligence to augment the efficient sex sorting of male mosquitoes. Although both radiation and artificial intelligence increase the cost of the IIT program, these repeatable successes demonstrate the feasibility of area-wide application of Wolbachia-based IIT for the suppression of dengue vector mosquitoes. ...
Full-text available
The strong suppression of Aedes albopictus on two Guangzhou islands in China has been successfully achieved by releasing males with an artificial triple-Wolbachia infection. However, it requires the use of radiation to sterilize residual females to prevent population replacement. To develop a highly effective tool for dengue control, we tested a standalone incompatible insect technique (IIT) to control A. albopictus in the urban area of Changsha, an inland city where dengue recently emerged. Male mosquitoes were produced in a mass rearing facility in Guangzhou and transported over 670 km under low temperature to the release site. After a once-per-week release with high numbers of males (phase I) and a subsequent twice-per-week release with low numbers of males (phase II), the average numbers of hatched eggs and female adults collected weekly per trap were reduced by 97% and 85%, respectively. The population suppression caused a 94% decrease in mosquito biting at the release site compared to the control site. Remarkably, this strong suppression was achieved using only 28% of the number of males released in a previous trial. Despite the lack of irradiation to sterilize residual females, no triple-infected mosquitoes were detected in the field post release based on the monitoring of adult and larval A. albopictus populations for two years, indicating that population replacement was prevented. Our results support the feasibility of implementing a standalone IIT for dengue control in urban areas. A field trial in Changsha, China, involving the release of a triple Wolbachia-infected strain of the mosquito, Aedes albopictus, demonstrates the feasibility of using an incompatible insect technique to suppress disease vector mosquito populations in urban subtropic environments.
... Mating between infected males and uninfected females results in non-viable offspring through cytoplasmic incompatibility (CI), leading to suppression of the mosquito population [20,21]. Wolbachia-mediated IIT as a method of vector control has been tested in China, the USA, Thailand, and Singapore, the country of focus for this study protocol [22][23][24][25]. Singapore, among other programs that also attempt to control the Ae. ...
... The Wolbachia-based population suppression strategy to reduce Ae. aegypti numbers and hence incidence of dengue has been successfully tested in several countries [22,25]. In Singapore, pilot trials have demonstrated the ability of IIT-SIT to reduce the wildtype Ae. aegypti population and dengue incidence in two densely-populated high rise residential estates [24]. ...
Full-text available
Background Dengue is a severe environmental public health challenge in tropical and subtropical regions. In Singapore, decreasing seroprevalence and herd immunity due to successful vector control has paradoxically led to increased transmission potential of the dengue virus. We have previously demonstrated that incompatible insect technique coupled with sterile insect technique (IIT-SIT), which involves the release of X-ray-irradiated male Wolbachia -infected mosquitoes, reduced the Aedes aegypti population by 98% and dengue incidence by 88%. This novel vector control tool is expected to be able to complement current vector control to mitigate the increasing threat of dengue on a larger scale. We propose a multi-site protocol to study the efficacy of IIT-SIT at reducing dengue incidence. Methods/design The study is designed as a parallel, two-arm, non-blinded cluster-randomized (CR) controlled trial to be conducted in high-rise public housing estates in Singapore, an equatorial city-state. The aim is to determine whether large-scale deployment of male Wolbachia -infected Ae. aegypti mosquitoes can significantly reduce dengue incidence in intervention clusters. We will use the CR design, with the study area comprising 15 clusters with a total area of 10.9 km ² , covering approximately 722,204 residents in 1713 apartment blocks. Eight clusters will be randomly selected to receive the intervention, while the other seven will serve as non-intervention clusters. Intervention efficacy will be estimated through two primary endpoints: (1) odds ratio of Wolbachia exposure distribution (i.e., probability of living in an intervention cluster) among laboratory-confirmed reported dengue cases compared to test-negative controls and (2) laboratory-confirmed reported dengue counts normalized by population size in intervention versus non-intervention clusters. Discussion This study will provide evidence from a multi-site, randomized controlled trial for the efficacy of IIT-SIT in reducing dengue incidence. The trial will provide valuable information to estimate intervention efficacy for this novel vector control approach and guide plans for integration into national vector control programs in dengue-endemic settings. Trial registration, identifier: NCT05505682 . Registered on 16 August 2022. Retrospectively registered.
... This is in contrast to SIT in which radiation induces sterility in males [8]. Field trials of IIT have resulted in successfully eradicating Cx. quinquefasciatus [10] and the near-elimination of the Asian tiger mosquito Aedes albopictus and the yellow fever mosquito Aedes aegypti populations [11][12][13]. Area-wide implementation of SIT has successfully suppressed populations of screwworm, medfly, and tsetse flies [8,14], with recent encouraging progress in field trials for Aedes spp. control [15,16]. ...
... Both IIT and SIT have been used to successfully suppress, eradicate, contain, and prevent establishment of insects of health and agricultural concern [8,[11][12][13][14][15][16]50]. From the nearly successful eradication of Cx. quinquefasciatus [10] using IIT, to the canonical use of SIT to eliminate screwworms [50,51], these tools have proven to be powerful additions to any IPM plan [52]. ...
Full-text available
Background Hawaiʻi’s native forest avifauna is experiencing drastic declines due to climate change-induced increases in temperature encroaching on their upper-elevation montane rainforest refugia. Higher temperatures support greater avian malaria infection rates due to greater densities of its primary vector, the southern house mosquito Culex quinquefasciatus , and enhance development of the avian malaria parasite Plasmodium relictum . Here we propose the use of the incompatible insect technique (IIT) or the combined IIT/sterile insect technique (SIT) for the landscape-scale (i.e., area-wide) control of Cx. quinquefasciatus , and have developed a calculator to estimate the costs of IIT and IIT/SIT applications at various sites in Hawaiʻi. Methods The overall cost of the infrastructure, personnel, and space necessary to produce incompatible adult males for release is calculated in a unit of ~ 1 million culicid larvae/week. We assessed the rearing costs and need for effective control at various elevations in Hawaiʻi using a 10:1 overflooding ratio at each elevation. The calculator uses a rate describing the number of culicids needed to control wild-type mosquitoes at each site/elevation, in relation to the number of larval rearing units. This rate is a constant from which other costs are quantified. With minor modifications, the calculator described here can be applied to other areas, mosquito species, and similar techniques. To test the robustness of our calculator, the Kauaʻi-specific culicid IIT/SIT infrastructure costs were also compared to costs from Singapore, Mexico, and China using the yearly cost of control per hectare, and purchasing power parity between sites for the cost of 1000 IIT/SIT males. Results As a proof of concept, we have used the calculator to estimate rearing infrastructure costs for an application of IIT in the Alakaʻi Wilderness Reserve on the island of Kauaʻi. Our analysis estimated an initial investment of at least ~ $1.16M with subsequent yearly costs of approximately $376K. Projections of rearing costs for control at lower elevations are ~ 100 times greater than in upper elevation forest bird refugia. These results are relatively comparable to those real-world cost estimates developed for IIT/SIT culicid male production in other countries when inflation and purchasing power parity are considered. We also present supplemental examples of infrastructure costs needed to control Cx. quinquefasciatus in the home range of ʻiʻiwi Drepanis coccinea , and the yellow fever vector Aedes aegypti . Conclusions Our cost calculator can be used to effectively estimate the mass rearing cost of an IIT/SIT program. Therefore, the linear relationship of rearing infrastructure to costs used in this calculator is useful for developing a conservative cost estimate for IIT/SIT culicid mass rearing infrastructure. These mass rearing cost estimates vary based on the density of the targeted organism at the application site. Graphic Abstract
... Therefore, some Sterile Insect Technique programs are Wolbachia-based. The release of Wolbachia-infected insects could serve as a powerful tool for controlling pests like Drosophila suzukii [48], Aedes aegypti [49,50] and Aedes albopictus [51]. ...
Full-text available
Wolbachia are maternally-inherited bacteria, which can spread rapidly in populations by manipulating reproduction. cifA and cifB are genes found in Wolbachia phage that are responsible for cytoplasmic incompatibility, the most common type of Wolbachia reproductive interference. In this phenomenon, no viable offspring are produced when a male with both cifA and cifB (or just cifB in some systems) mates with a female lacking cifA . Utilizing this feature, we propose new types of toxin-antidote gene drives that can be constructed with only these two genes in an insect genome, instead of the whole Wolbachia bacteria. By using both mathematical and simulation models, we found that a drive containing cifA and cifB together creates a confined drive with a moderate to high introduction threshold. When introduced separately, they act as a self-limiting drive. We observed that the performance of these drives is substantially influenced by various ecological parameters and drive characteristics. Extending our models to continuous space, we found that the drive individual release distribution has a critical impact on drive persistence. Our results suggest that these new types of drives based on Wolbachia transgenes are safe and flexible candidates for genetic modification of populations.
... Because it can suppress the transmission of specific mosquito-borne viruses and parasites when transferred to novel mosquito hosts, Wolbachia has been the focus of much recent research (e.g., [12][13][14][15]). Wolbachia-infected mosquitoes have been released into the field in multiple countries to curb the spread of dengue virus (DENV) by Ae. aegypti vectors [8,9,[16][17][18][19][20]. In some cases, Wolbachia-infected animals can replace native populations and retain a pathogen-blocking phenotype for multiple years after release [8,9,[21][22][23][24][25]. ...
Full-text available
Wolbachia pipientis (= Wolbachia ) has promise as a tool to suppress virus transmission by Aedes aegypti mosquitoes. However, Wolbachia can have variable effects on mosquito-borne viruses. This variation remains poorly characterized, yet the multimodal effects of Wolbachia on diverse pathogens could have important implications for public health. Here, we examine the effects of somatic infection with two strains of Wolbachia ( w AlbB and w Mel) on the alphaviruses Sindbis virus (SINV), O’nyong-nyong virus (ONNV), and Mayaro virus (MAYV) in Ae. aegypti . We found variable effects of Wolbachia including enhancement and suppression of viral infections, with some effects depending on Wolbachia strain. Both w AlbB- and w Mel-infected mosquitoes showed enhancement of SINV infection rates one week post-infection, with w AlbB-infected mosquitoes also having higher viral titers than controls. Infection rates with ONNV were low across all treatments and no significant effects of Wolbachia were observed. The effects of Wolbachia on MAYV infections were strikingly strain-specific; w Mel strongly blocked MAYV infections and suppressed viral titers, while w AlbB did not influence MAYV infection. The variable effects of Wolbachia on vector competence underscore the importance of further research into how this bacterium impacts the virome of wild mosquitoes including the emergent human pathogens they transmit. Impact statement In recent years, wild populations of Aedes aegypti mosquitoes have been deliberately infected with Wolbachia —a bacterium that helps to curb the spread of some pathogens including dengue virus. But how does Wolbachia affect the ability of mosquitoes to become infected with and spread the many different viruses they encounter in nature? Here, we characterize the effects of Wolbachia on three alphaviruses that cause illness in humans— Sindbis virus, O’nyong-nyong virus, and Mayaro virus. We find Wolbachia has variable effects on these pathogens, including significant enhancement of Sindbis virus infections. Our research has important implications for the design of vector control strategies, and suggests further research is needed to understand how Wolbachia shapes the replication and transmission of diverse viruses. 3. Data Summary All data in the study are available in the Figures and supplementary material.
... Thus, the World Health Organization (WHO) has designated dengue as one of the top 10 threats to global health in 2019, together with other important issues such as antimicrobial resistance and global influenza pandemic ( Various new technologies, including the use of sterile insect technique and transgenic and Wolbachia-infected mosquitoes, have been developed to break the transmission of dengue virus (4)(5)(6)(7). However, they are not ready for large-scale implementation, and there are also some arguments on the release of genetically engineered organisms to the environment (8). ...
Full-text available
Aedes aegypti (Linnaeus, 1762) is the main mosquito vector for dengue and other arboviral infectious diseases. Control of this important vector highly relies on the use of insecticides, especially pyrethroids. The high frequency (>78%) of the L982W substitution was detected at the target site of the pyrethroid insecticide, the voltage-gated sodium channel (Vgsc) of A. aegypti collected from Vietnam and Cambodia. Alleles having concomitant mutations L982W + F1534C and V1016G + F1534C were also confirmed in both countries, and their frequency was high (>90%) in Phnom Penh, Cambodia. Strains having these alleles exhibited substantially higher levels of pyrethroid resistance than any other field population ever reported. The L982W substitution has never been detected in any country of the Indochina Peninsula except Vietnam and Cambodia, but it may be spreading to other areas of Asia, which can cause an unprecedentedly serious threat to the control of dengue fever as well as other Aedes -borne infectious diseases.
... Release of sterile mosquitoes and Wolbachia-carrying mosquitoes have been studied for many years and shown significant efficacy in controlling mosquito populations in field trial [47][48][49] . Sustained releases of transgenic A. aegypti males with the OX513A lethal gene led to at least 80% suppression of the wild A. aegypti population in the Cayman Islands in 2010 and a suburb of Juazeiro, Bahia, Brazil in 2012 [50][51][52] . ...
Full-text available
Zoonotic viruses circulate in the natural reservoir and sporadically spill over into human populations, resulting in endemics or pandemics. We previously found that the Chaoyang virus (CYV), an insect-specific flavivirus (ISF), is replication-defective in vertebrate cells. Here, we develope a proof-of-concept mosquito-delivered vaccine to control the Zika virus (ZIKV) within inaccessible wildlife hosts using CYV as the vector. The vaccine is constructed by replacing the pre-membrane and envelope (prME) proteins of CYV with those of ZIKV, assigned as CYV-ZIKV. CYV-ZIKV replicates efficiently in Aedes mosquitoes and disseminates to the saliva, with no venereal or transovarial transmission observed. To reduce the risk of CYV-ZIKV leaking into the environment, mosquitoes are X-ray irradiated to ensure 100% infertility, which does not affect the titer of CYV-ZIKV in the saliva. Immunization of mice via CYV-ZIKV-carrying mosquito bites elicites robust and persistent ZIKV-specific immune responses and confers complete protection against ZIKV challenge. Correspondingly, the immunized mice could no longer transmit the challenged ZIKV to naïve mosquitoes. Therefore, immunization with an ISF-vectored vaccine via mosquito bites is feasible to induce herd immunity in wildlife hosts of ZIKV. Our study provides a future avenue for developing a mosquito-delivered vaccine to eliminate zoonotic viruses in the sylvatic cycle.
Full-text available
Aedes aegypti mosquitoes carrying self-spreading, virus-blocking Wolbachia bacteria are being deployed to suppress dengue transmission. However, there are challenges in applying this technology in extreme environments. We introduced two Wolbachia strains into Ae. aegypti from Saudi Arabia for a release program in the hot coastal city of Jeddah. Wolbachia reduced infection and dissemination of dengue virus (DENV2) in Saudi Arabian mosquitoes and showed complete maternal transmission and cytoplasmic incompatibility. Wolbachia reduced egg hatch under a range of environmental conditions, with the Wolbachia strains showing differential thermal stability. Wolbachia effects were similar across mosquito genetic backgrounds but we found evidence of local adaptation, with Saudi Arabian mosquitoes having lower egg viability but higher adult desiccation tolerance than Australian mosquitoes. Genetic background effects will influence Wolbachia invasion dynamics, reinforcing the need to use local genotypes for mosquito release programs, particularly in extreme environments like Jeddah. Our comprehensive characterization of Wolbachia strains provides a foundation for Wolbachia-based disease interventions in harsh climates.
Full-text available
Bacterial symbionts can promote insecticide resistance in their hosts by isolating and degrading insecticidal compounds or altering the expression of host genes. Although Wolbachia , a common endosymbiont in arthropods, typically does not influence insecticide resistance, there are cases of increased or decreased susceptibility. Due to the restrictions of applying conventional insecticides in a stored product setting, studies on alternative control methods are needed, including those on entomotoxic nanoparticles (NPs) and the potential for resistance. For pests of stored beans, selenium nanoparticles (SeNPs) are relatively innocuous to the azuki bean beetle, Callosobruchus chinensis (L.) (Coleoptera: Chrysomelidae: Bruchinae). Here, we hypothesized that this response is mediated by Wolbachia , and we tested this using an isofemale line of C. chinensis (infected or uninfected with Wolbachia ). Our results showed that the lifespan of Wolbachia -infected females was not affected by SeNPs, but increasing concentrations of SeNPs still had a negative effect on fecundity; in uninfected females, increasing concentrations of SeNPs significantly decreased both lifespan and fecundity. However, in males, SeNPs enhanced lifespan and decreased the incidence of sexual harassment behavior regardless of infection status (for uninfected males, the duration of harassment behavior also decreased). In the presence of males, 72-h female reproduction increased independent of infection status or SeNP treatment, but egg hatchability was reduced by male presence and SeNPs. This study documents a valuable example of symbiont-mediated resistance to entomotoxic nanoparticles.
Full-text available
Abstract Wolbachia are maternally inherited endosymbiotic bacteria found within many insect species. Aedes mosquitoes experimentally infected with Wolbachia are being released into the field for Aedes‐borne disease control. These Wolbachia infections induce cytoplasmic incompatibility which is used to suppress populations through incompatible matings or replace populations through the reproductive advantage provided by this mechanism. However, the presence of naturally occurring Wolbachia in target populations could interfere with both population replacement and suppression programs depending on the compatibility patterns between strains. Aedes aegypti were thought to not harbor Wolbachia naturally but several recent studies have detected Wolbachia in natural populations of this mosquito. We therefore review the evidence for natural Wolbachia infections in A. aegypti to date and discuss limitations of these studies. We draw on research from other mosquito species to outline the potential implications of natural Wolbachia infections in A. aegypti for disease control. To validate previous reports, we obtained a laboratory population of A. aegypti from New Mexico, USA, that harbors a natural Wolbachia infection, and we conducted field surveys in Kuala Lumpur, Malaysia, where a natural Wolbachia infection has also been reported. However, we were unable to detect Wolbachia in both the laboratory and field populations. Because the presence of naturally occurring Wolbachia in A. aegypti could have profound implications for Wolbachia‐based disease control programs, it is important to continue to accurately assess the Wolbachia status of target Aedes populations.
Full-text available
Background: Important arboviral diseases, such as dengue, chikungunya, and Zika virus infections, are transmitted mainly by the Aedes aegypti vector. So far, controlling this vector species with current tools and strategies has not demonstrated sustainable and significant impacts. Our main objective was to evaluate whether open field release of sterile males, produced from combining the sterile insect technique using radiation with the insect incompatible technique through Wolbachia-induced incompatibility (SIT/IIT), could suppress natural populations of Ae. aegypti in semi-rural village settings in Thailand. Methodology/principal findings: Irradiated Wolbachia-infected Aedes aegypti males produced by the SIT/IIT approach were completely sterile and were able to compete with the wild fertile ones. Open field release of these sterile males was conducted in an ecologically isolated village in Chachoengsao Province, eastern Thailand. House-to-house visit and media reports resulted in community acceptance and public awareness of the technology. During intervention, approximately 100-200 sterile males were released weekly in each household. After 6 months of sterile male release, a significant reduction (p<0.05) of the mean egg hatch rate (84%) and the mean number of females per household (97.30%) was achieved in the treatment areas when compared to the control ones. Conclusions/significance: Our study represents the first open field release of sterile Ae. aegypti males developed from a combined SIT/IIT approach. Entomological assessment using ovitraps, adult sticky traps, and portable vacuum aspirators confirmed the success in reducing natural populations of Ae. aegypti females in treated areas. Public awareness through media resulted in positive support for practical use of this strategy in wider areas. Further study using a systematic randomized trial is needed to determine whether this approach could have a significant impact on the diseases transmitted by Ae. aegypti vector.
Full-text available
The radiation-based sterile insect technique (SIT) has successfully suppressed field populations of several insect pest species, but its effect on mosquito vector control has been limited. The related incompatible insect technique (IIT)—which uses sterilization caused by the maternally inherited endosymbiotic bacteria Wolbachia—is a promising alternative, but can be undermined by accidental release of females infected with the same Wolbachia strain as the released males. Here we show that combining incompatible and sterile insect techniques (IIT–SIT) enables near elimination of field populations of the world’s most invasive mosquito species, Aedes albopictus. Millions of factory-reared adult males with an artificial triple-Wolbachia infection were released, with prior pupal irradiation of the released mosquitoes to prevent unintentionally released triply infected females from successfully reproducing in the field. This successful field trial demonstrates the feasibility of area-wide application of combined IIT–SIT for mosquito vector control.
Full-text available
Aedes aegypti bears the common name “the yellow fever mosquito,” although, today, it is of more concern as the major vector of dengue, chikungunya, and, most recently, Zika viruses. In the present article, we review recent work on the population genetics of this mosquito in efforts to reconstruct its recent (approximately 600 years) history and relate these findings to epidemiological records of occurrences of diseases transmitted by this species. The two sources of information are remarkably congruent. Ae. aegypti was introduced to the New World 400–550 years ago from its ancestral home in West Africa via European slave trade. Ships from the New World returning to their European ports of origin introduced the species to the Mediterranean region around 1800, where it became established until about 1950. The Suez Canal opened in 1869 and Ae. aegypti was introduced into Asia by the 1870s, then on to Australia (1887) and the South Pacific (1904).
Full-text available
Background To ensure the success of a mosquito control programme that integrates the sterile insect technique (SIT), it is highly desirable to release sterile males with a maximal lifespan to increase release effectiveness. Understanding sterile male survival under field conditions is thus critical for determining the number of males to be released. Our study aimed to investigate the effect of mass rearing, irradiation, chilling, packing and release time on irradiated male mosquito longevity. Methods Anopheles arabiensis and Aedes aegypti immature stages were mass-reared using a rack and tray system. Batches of 50 males irradiated at the pupal stage were immobilised, packed into canisters and chilled for 6 hours at 6 °C. Mosquitoes were then transferred either in the early morning or early evening into climate chambers set to simulate the weather conditions, typical of the beginning of the rainy season in Khartoum, Sudan and Juazeiro, Brazil for An. arabiensis and Ae. aegypti, respectively. The longevity of experimental males was assessed and compared to mass-reared control males subjected either to simulated field or laboratory conditions. Results The combined irradiation, chilling and packing treatments significantly reduced the longevity of both An. arabiensis and Ae. aegypti under simulated field conditions (P < 0.001). However, packing alone did not significantly reduce longevity of Ae. aegypti (P = 0.38) but did in An. arabiensis (P < 0.001). Overall, the longevity of mass reared, irradiated and packed males was significantly reduced, with the median survival time (days) lower following an early morning introduction (4.62 ± 0.20) compared to an evening (7.34 ± 0.35) in An. arabiensis (P < 0.001). However, there was no significant difference in longevity between morning (9.07 ± 0.54) and evening (7.76 ± 0.50) in Ae. aegypti (P = 0.14). Conclusions Our study showed that sterile mass-reared males have a reduced lifespan in comparison to laboratory-maintained controls under simulated field conditions, and that An. arabiensis appeared to be more sensitive to the handling process and release time than Ae. aegypti. Longevity and release time are important parameters to be considered for a successful area-wide integrated vector control programme with a SIT component.
Full-text available
Background: Artificially-introduced transinfections of the intracellular bacterium Wolbachia pipientis have the potential to reduce the vectorial capacity of mosquito populations for viruses such as dengue and chikungunya. Aedes albopictus has two native strains of Wolbachia, but their replacement with the non-native wMel strain blocks transmission of both viruses. The pattern of cytoplasmic incompatiiblity generated by wMel with wild-types is bidirectional. Novel-plus-native-strain co-infection is predicted to lead to a more efficient population spread capacity; from a bi-directional to a uni-directional cytoplasmic incompatibility (CI) model. Results: A novel-plus-native-strain triple-infection in Ae. albopictus (wAlbAwAlbBwMel) was generated. Although triple-infected females were fully reproductively viable with uninfected males, they displayed self-incompatibility. qPCR of specific strains in dissected tissues suggested that this may be due to the displacement of one of the native strains (wAlbA) from the ovaries of triple-infected females. When the triple strain infection was transferred into Aedes aegypti it displayed an unexpectedly low level of transmission fidelity of the three strains in this species. Conclusions: These results suggest that combining Wolbachia strains can lead to co-infection interactions that can affect outcomes of CI and maternal transmission.
Full-text available
The yellow fever mosquito Aedes aegypti inhabits much of the tropical and subtropical world and is a primary vector of dengue, Zika, and chikungunya viruses. Breeding populations of A. aegypti were first reported in California (CA) in 2013. Initial genetic analyses using 12 microsatellites on collections from Northern CA in 2013 indicated the South Central US region as the likely source of the introduction. We expanded genetic analyses of CA A. aegypti by: (a) examining additional Northern CA samples and including samples from Southern CA, (b) including more southern US populations for comparison, and (c) genotyping a subset of samples at 15,698 SNPs. Major results are: (1) Northern and Southern CA populations are distinct. (2) Northern populations are more genetically diverse than Southern CA populations. (3) Northern and Southern CA groups were likely founded by two independent introductions which came from the South Central US and Southwest US/northern Mexico regions respectively. (4) Our genetic data suggest that the founding events giving rise to the Northern CA and Southern CA populations likely occurred before the populations were first recognized in 2013 and 2014, respectively. (5) A Northern CA population analyzed at multiple time-points (two years apart) is genetically stable, consistent with permanent in situ breeding. These results expand previous work on the origin of California A. aegypti with the novel finding that this species entered California on multiple occasions, likely some years before its initial detection. This work has implications for mosquito surveillance and vector control activities not only in California but also in other regions where the distribution of this invasive mosquito is expanding.
Full-text available
In 2011, a thriving population of Aedes albopictus (Skuse), the Asian tiger mosquito, was discovered within three cities in Los Angeles County over an estimated 52-km2 urban area. Two years later in 2013, Aedes aegypti (L.), the yellow fever mosquito, was detected within several urban areas of Madera, Fresno, and San Mateo counties. State and local vector control agencies responded with an aggressive effort to eradicate or interrupt the spread of these two invasive mosquitoes; however, known populations continued to expand outward and new infestations were identified at an accelerated pace in central and southern California. By the end of 2015, one or both species had been detected within the jurisdictional boundaries of 85 cities and census-designated places in 12 counties. Herein we report on the discovery and widespread establishment of Ae. aegypti and Ae. albopictus in urban areas of coastal, central, and southern California between 2011 and 2015 and discuss the subsequent rapid changes to the activities and priorities of vector control agencies in response to this unprecedented invasion.
Wolbachia is a genus of endosymbiotic bacteria that infects 66% of all insect species. Its major impact on insects is in reproduction: sterility, production of one sex, and/or parthenogenesis. Another effect was discovered when the disease-transmitting mosquito, Aedes aegypti Linnaeus (Diptera: Culicidae), was infected with Wolbachia isolated from Drosophila: infected female mosquitoes became less capable of transmitting diseases such as dengue fever and chikungunya. This has led to releases of Ae. aegypti carrying Wolbachia in an attempt to control disease. An open question is whether there are natural Wolbachia infections of this mosquito. We assayed DNA from 2,663 Ae. aegypti from 27 countries on six continents, 230 from laboratory strains, and 72 Aedes mascarensis MacGregor (Diptera: Culicidae) for presence of Wolbachia DNA. Within the limits of our polymerase chain reaction-based assay, we found no evidence of Wolbachia, suggesting that natural infections of this endosymbiont are unlikely to occur throughout the worldwide distribution of Ae. aegypti.
Despite an increase in vector-borne diseases in the United States, there has been public opposition to novel vector control methods, particularly those that leverage advances in genetic engineering.¹ In 2016, the US Food and Drug Administration (FDA) invited public comments on a draft environmental assessment² for a proposed field trial of a genetically engineered mosquito (OX513A) designed to suppress wild-type Aedes aegypti mosquitoes, which can transmit diseases such as Zika, dengue, and chikungunya. Trials of OX513A in other countries have demonstrated reductions in numbers of A aegypti.