Limited Dengue Virus Replication in Field-Collected

Aedes aegypti Mosquitoes Infected with Wolbachia

Francesca D. Frentiu1¤a, Tasnim Zakir1, Thomas Walker1¤b, Jean Popovici1¤c, Alyssa T. Pyke2, Andrew van
den Hurk2, Elizabeth A. McGraw1, Scott L. O’Neill1,3*
1 School of Biological Sciences, Monash University, Clayton, Victoria, Australia, 2 Public Health Virology, Forensic and Scientific Services, Department of Health, Archerfield,
Queensland, Australia, 3 Institute for Molecular Biosciences, The University of Queensland, St. Lucia, Queensland, Australia

Abstract
Introduction: Dengue is one of the most widespread mosquito-borne diseases in the world. The causative agent, dengue
virus (DENV), is primarily transmitted by the mosquito Aedes aegypti, a species that has proved difficult to control using
conventional methods. The discovery that A. aegypti transinfected with the wMel strain of Wolbachia showed limited DENV
replication led to trial field releases of these mosquitoes in Cairns, Australia as a biocontrol strategy for the virus.

Methodology/Principal Findings: Field collected wMel mosquitoes that were challenged with three DENV serotypes
displayed limited rates of body infection, viral replication and dissemination to the head compared to uninfected controls.
Rates of dengue infection, replication and dissemination in field wMel mosquitoes were similar to those observed in the
original transinfected wMel line that had been maintained in the laboratory. We found that wMel was distributed in similar
body tissues in field mosquitoes as in laboratory ones, but, at seven days following blood-feeding, wMel densities increased
to a greater extent in field mosquitoes.

Conclusions/Significance: Our results indicate that virus-blocking is likely to persist in Wolbachia-infected mosquitoes after
their release and establishment in wild populations, suggesting that Wolbachia biocontrol may be a successful strategy for
reducing dengue transmission in the field.

Citation: Frentiu FD, Zakir T, Walker T, Popovici J, Pyke AT, et al. (2014) Limited Dengue Virus Replication in Field-Collected Aedes aegypti Mosquitoes Infected
with Wolbachia. PLoS Negl Trop Dis 8(2): e2688. doi:10.1371/journal.pntd.0002688
Editor: Michael J. Turell, United States Army Medical Research Institute of Infectious Diseases, United States of America
Received May 20, 2013; Accepted December 21, 2013; Published February 20, 2014
Copyright: ß 2014 Frentiu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was supported by grants from the Foundation for the National Institutes of Health through the Grand Challenges in Global Health
Initiative of the Bill and Melinda Gates Foundation, the National Health and Medical Research Council of Australia and the Queensland State Government. The
funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
¤a Current address: Institute of Health and Biomedical Innovation/School of Biomedical Sciences, Queensland University of Technology, Kelvin Grove, Queensland,
Australia.
¤b Current address: Department of Disease Control, London School of Hygiene and Tropical Medicine, London, United Kingdom.
¤c Current address: Malaria Epidemiology Unit, Pasteur Institute of Cambodia, Phnom Penh, Cambodia.

Introduction the development of novel, inexpensive and green vector control

methods [4,5].
Dengue is one of the most common and widespread vector- The transinfection of vector mosquitoes with the bacterium
borne diseases in the world, with up to 380 million infections Wolbachia pipientis has emerged as a promising method for the
estimated to occur annually [1]. The causative agent, dengue virus control of dengue. Wolbachia is the most common endosymbiont of
(DENV), has expanded its geographic range in the last two insects, thought to infect up to 40% of arthropod species [6].
decades, with more than 100 countries now affected. Infection A. aegypti stably transinfected with different strains of Wolbachia
with DENV leads primarily to self-limiting fevers but recent show reduced replication and transmission of DENV [7–9]. An
decades have seen a marked increase in severe dengue, with additional advantage of using Wolbachia for biocontrol of DENV is
manifestations such as hypovolemic shock and hemorrhage [2]. the ability of the bacterium to propagate through a population
DENV is transmitted primarily by the mosquito vector Aedes aegypti by inducing cytoplasmic incompatibility (CI) in its host [10].
and, to a lesser extent, by its congener A. albopictus. In the absence CI confers a fitness advantage to Wolbachia-infected females that
of an effective vaccine [3] and/or antivirals, prevention of dengue allows these maternally transmitted bacteria to spread unaided
transmission relies primarily on control of mosquito vectors. The through a population [10]. The use of Wolbachia provides a means
failure to prevent the global spread of dengue, increasing of biocontrol that is both pesticide-free and poses minimal envi-
insecticide resistance in mosquito populations and subsequent ronmental safety concerns [11].
escalating costs of insecticide-based programs, as well as environ- In laboratory trials, mosquitoes with the wMel strain of
mental concern over the impact of these chemicals, have spurred Wolbachia showed both blocking of DENV transmission and

PLOS Neglected Tropical Diseases | www.plosntds.org 1 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

results reinforce the utility of Wolbachia-based technology for bio-

Author Summary control of dengue.
Almost half of the world’s population is at risk of
contracting dengue virus, particularly in the tropics and Methods
sub-tropics. The virus is transmitted by the mosquito Aedes
aegypti, a cosmopolitan species that has proved difficult to Ethics statement
control using traditional methods. A new biocontrol Blood feeding of mosquito colonies using human volunteers was
strategy has been developed involving the release of performed in accordance to Monash University Human Research
mosquitoes infected with Wolbachia bacteria. Mosquitoes Ethics Committee permit CF11/0766-2011000387. Written in-
with the wMel strain of Wolbachia show dramatically formed consent was obtained from all volunteers who participated
reduced replication and transmission of dengue virus in in the study. Dengue viremic plasma was obtained from patients
laboratory trials. Although promising, the utility of enrolled in a prospective study at the Hospital for Tropical
Wolbachia biocontrol depends on field wMel-infected Diseases, Ho Chi Minh City, Vietnam. All patients provided
mosquitoes retaining the phenotype of reduced viral written consent to participate in the study. The study protocols
replication. Mosquitoes with wMel were released in the relevant to this work, including vector competence experiments,
field in Cairns, Australia in early 2011. We provide evidence were reviewed and approved by the Scientific and Ethical
that, one year later, field collected wMel mosquitoes Committee of the Hospital for Tropical Diseases (CS/ND/09/
showed reduced dengue virus replication in the body and
24) and the Oxford Tropical Research Ethical Committee
limited dissemination to the head compared to controls.
(OxTREC 20-09). The inclusion criteria were: a) adult patients
Wolbachia numbers in mosquitoes increased following
blood meals, which may further decrease viral replication if ($15 years of age), with #72 hours of fever and suspected of
the insects feed frequently. Our results indicate that having dengue based on clinical symptoms, b) a positive NS1
Wolbachia-mediated dengue interference is sustained in Rapid test and c) written informed consent. All plasma samples
field populations and shows no sign of attenuation after were anonymized (samples were identified using numbers only)
one year of deployment. prior to experiments.

Mosquito colony establishment and maintenance

Mosquito eggs were collected in January 2012 from ovitraps
minimal fitness effects due to infection with the bacterium [9]. In placed inside the Wolbachia release zone in the Cairns suburbs of
addition, wMel rapidly invaded wildtype mosquito populations in Yorkey’s Knob and Gordonvale and outside, in Edge Hill,
semi-field cage experiments due to CI and minimal fitness costs Whitfield, Edmonton and Bentley Park. Eggs collected from
[9]. The results facilitated the field release of wMel-infected outside the Wolbachia release zone were Wolbachia-uninfected. Eggs
mosquitoes in two suburbs of Cairns, Queensland, Australia [12]. on ovistrips were allowed to hatch and larvae reared in water
Within a short period, the frequency of wMel reached fixation in supplemented with fish food pellets (Tetramin, Tetra). Fourth
the two suburbs [12] and has remained established at both sites. instar larvae were identified as A. aegypti based on specific
The persistence of the viral-blocking phenotype in field popu- morphological characters. Adults (F0) emerged in cages of
lations is fundamental to the utility of releases of Wolbachia-infected approximately 450 individuals and were allowed to feed on 10%
mosquitoes. The mechanisms that underpin viral interference are sucrose ad libitum. Five to seven day old females were allowed to
poorly understood but may be related to the density of Wolbachia feed on human volunteers and eggs were collected from several
[13,14], immune pre-activation [7,8,15], intra-host competition gonotrophic cycles. F1 adults hatched from eggs obtained in the
for cellular resources [16,17] or suppression of host cellular factors first gonotrophic cycle were used in vector competence experi-
that are upregulated during viral infection [18]. The density of ments. The wMel-infected field mosquito line and its uninfected
Wolbachia may decrease after several generations, as happened counterpart (derived from Wolbachia-uninfected eggs) were denot-
following the transinfection of the virulent strain of wMelPop ed wMel.F and wildtype, respectively. The original laboratory-
into the novel host Drosophila simulans [19]. Wolbachia infection reared, outcrossed wMel-infected MGYP2.out line [9] was used in
frequencies and associated CI effects may also be significantly some experiments. All mosquito colonies were kept at 26uC under
lower in nature than observed in the lab, as observed in Drosophila a 12L:12D light cycle and 60% relative humidity.
simulans [20]. However, the wMel strain is avirulent and has
limited negative effects on mosquito fitness in the laboratory [9], Virus strains
suggesting that the density of the wMel strain may remain stable Mosquitoes were challenged in vector competence experiments
over time. Protection against RNA virus-induced mortality was with virus strains belonging to DENV serotypes 1–3, using virus
in fact first observed in the long term, evolutionarily stable grown in cell culture and viremic plasma from human patients.
association between wMel and its Drosophila melanogaster host [21]. DENV-2 strain 92T and DENV-3 strain Cairns 2008 (both
Here, we investigated the extent of virus blocking in field wMel- isolated from outbreaks in north Queensland, Australia in 1992
infected A. aegypti, one year following field release, using three and 2008, respectively) were grown in C6/36 cells and harvested
serotypes of DENV. We found limited replication and dissemina- and titered as described previously [13]. Virus was aliquoted in
tion of DENV in field wMel mosquitoes, indicating stability of the single-use 1 mL lots and stored at 280uC.
viral-blocking phenotype in wild Wolbachia-infected mosquitoes.
The extent of virus blocking was similar in field mosquitoes Vector competence experiments
compared to the original, wMel-infected, outcrossed lab line used Two separate vector competence experiments were carried out
for release. Interestingly, the density of Wolbachia increased to determine if DENV could replicate and disseminate in field
following blood feeding and to a greater extent in field versus wMel-infected mosquitoes. For both experiments, female mosqui-
lab wMel-infected mosquitoes. We suggest that if the viral blocking toes (5–7 days old) were allowed to feed on viremic blood meals
effect of field wMel is dependent on Wolbachia density, repeated contained in a membrane feeder with sheep intestine as the
blood feeding on human hosts might amplify this effect. Our membrane. Virus was mixed with defibrinated sheep blood to

PLOS Neglected Tropical Diseases | www.plosntds.org 2 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

obtain final bloodmeal titers (see below). Mosquitoes were allowed to seven-day old females from each line were fed on a mix of
to feed for 1 hour, with engorged females separated from unfed DENV-3 – Cairns 2008 and sheep blood and collected at 7 and 14
ones the next day. Females were kept in plastic cups at a density days post infection (as detailed above) for genomic DNA extrac-
of 10–12 individuals/cup and allowed access to 10% sucrose tion. Control non-blood fed females from each line were main-
ad libitum. Females were killed under CO2 at either 7 or 14 days tained in parallel and collected at the same time points. Genomic
post infection (p.i.), immediately frozen in dry ice and stored at 2 DNA was extracted using the DNAEasy Blood and Tissue kit
80uC until further processing. (Qiagen) as per the manufacturer’s instructions. A multiplex qPCR
In the first experiment, field wMel and uninfected mosquitoes amplifying the target Wolbachia-specific wsp and mosquito house-
were challenged with two viremic plasma samples from Vietnam, keeping RpS17 [24] genes was performed (wsp F: 59-CATTG-
DENV-1 – P249 (final titer 7.38E+08 genomic copies/mL) and GTGTTGGTGTTGGTG-39, R: 59-ACACCAGCTTTTACTT-
DENV-2 – P410 (final titer 1.12E+09 genomic copies/mL), as well GACCAG-39, probe: 59-HEX-TCCTTTGGAACCCGCTGTG-
as a cell-culture grown virus isolated in Australia, DENV-2 – 92T AATGA-BHQ1-39; RpS17 F: 59-TCCGTGGTATCTCCAT-
(9.30E+09 copies/mL) as a control. In the second experiment, the CAAGC-39, R: 59-CACTTCCGGCACGTAGTTGTC-39, probe:
field wMel-infected and two control lines, MGYP2.out [9] and 59-FAM-CAGGAGGAGGAACGTGAGCGCAG-BHQ1-39).The
field Wolbachia-uninfected wildtype, were challenged with a viremic RpS17 housekeeping gene was used to normalize wsp gene copies.
human plasma sample from Vietnam, DENV-1-P307 (2.46E+11 qPCR reactions were performed in 10 mL total volume containing
copies/mL), and two virus strains isolated in Australia, DENV-2- 16 Lightcycler 480 Probes Master reaction mix, 5 mM each of wsp
92T (9.30E+09 copies/mL) and DENV-3-Cairns 2008 (3.58E+09 primers and probe, 2.5 mM each of RpS17 primers and probe and
copies/mL). Human viremic plasmas underwent a single freeze- 1 mL of DNA template. Cycling was performed using a Light-
thaw cycle before use in vector competence experiments. Cycler480 Instrument (Roche), with 1 cycle at 95uC for 5 min,
followed by 45 amplification cycles of 95uC for 10 s, 60uC for 15 s,
RNA extraction and qRT-PCR for DENV 72uC for 1 s, and a final cooling cycle of 40uC for 10 s. Target to
RNA was extracted from mosquito bodies using Trizol reagent housekeeping gene ratios were calculated using the Relative
(Invitrogen), and from heads using the QIAamp viral RNA mini Quantification algorithm in the Lightcycler 480 software (Roche).
kit (Qiagen), following homogenization of tissues with 3 mm glass
beads in a Beadbeater. A higher yield of total RNA was obtained Fluorescence in-situ hybridization (FISH)
on average from head samples using the QIAamp viral RNA mini Tissue localization of wMel in field wMel.F and lab MGYP2.out
kit versus Trizol (F. Frentiu, unpublished data). For mosquitoes mosquitoes was visualized using FISH. Females were collected
challenged with Vietnamese viremic plasmas, virus genome copies under CO2 and immediately placed overnight in 4% paraformal-
were estimated by qRT-PCR using FAM-labeled DENV-1 dehyde at 4uC with their wings and legs removed. Paraffin-
and DENV-2 hydrolysis probe sequences and standard curves embedded mosquitoes were sectioned in 8 mM thin slices. Slides
from reference [22]. Virus copies in mosquitoes challenged with were de-paraffinated in 100% xylene, rehydrated in an ethanol
DENV-2-92T and DENV-3-Cairns 2008 were estimated by qRT- series and hybridized overnight at 37uC in a buffer containing
PCR, using hydrolysis probes specific to the 39UTR region. Wolbachia-specific W2 and W3 probes [8]. Post-hybridization pro-
Primer sequences were F: 59-AAGGACTAGAGGTTAGAGGA- cessing followed [8]. Slides were mounted using an antifade
GACCC-39 and R: 59-CGTTCTGTGCCTGGAATGATG-39, reagent (Prolong Gold, Invitrogen) and viewed with a Zeiss Axio
with probe sequence: 59- FAM- AACAGCATATTGACGCTGG- Imager II epifluorescence microscope equipped with an Axiocam
GAGAGACCAGA-BHQ1-39. Reactions were performed with the camera, using the same exposure conditions for each filter
SuperScriptH III PlatinumH One-Step qRT-PCR kit (Invitrogen) channel.
and contained 5 mL of RNA template, 5 mM each of probe and
forward and reverse primers, buffer and enzyme as per kit Statistical analysis
instructions, in a total volume of 20 mL. For head qPCRs, 10 mL of Differences between mosquito lines in DENV infection rates for
RNA template was used, with water adjusted accordingly. The both vector competence experiments were analyzed using pairwise
number of DENV copies was calculated following a standard
curve for DENV 39UTR, constructed as in [8]. All reactions were
performed using a LightCycler480 Instrument (Roche) with the Table 1. Rates of infection (%) for three DENV strains
following run conditions: 50uC for 15 min, 95uC for 2 min, between field Wolbachia-infected (wMel.F) and uninfected
followed by 45 amplification cycles of 95uC for 15 s, 60uC for 30 s (wildtype) mosquito lines at days 7 and 14 p.i. (experiment 1).
and a final cooling step of 40uC for 10 s.
Reactions were run in duplicate and samples where DENV
Body infection (N) Head infection (N)
failed to amplify in at least one replicate were classified as zero.
Only samples where DENV amplified in both technical replicates wildtype wMel.F wildtype wMel.F
and the amount of copies extrapolated by the LightCycler software day 7 p.i.
was above the lower bound of the standard curve (limit of
DENV1 – P249 44 (18) 19 (16) 44 (18) 6 (16)
detection) were included in the analysis. All mosquitoes from field
and lab wMel-infected lines that showed DENV breakthrough DENV2 – 92T 26 (21) 0 (18) 10 (21) 0 (17)
were tested for the presence of Wolbachia using IS5 repeat primers DENV2 – P410 79 (14) 6 (31)*** 62 (13) 3 (31)***
specific to the wMel and wMelPop strains [23]. Only one sample day 14 p.i.
each from the field and lab wMel-infected mosquitoes was negative DENV1 – P249 41 (17) 4 (25)* 35 (17) 0 (25)***
for Wolbachia. These samples were excluded from further analysis.
DENV2 – 92T 53 (19) 7 (28)* 47 (19) 4 (28)***
DENV2 – P410 94 (16) 3 (30)* 62 (13) 3 (30)***
DNA extraction and quantification of Wolbachia density
The densities of Wolbachia were compared between field and lab Adjusted Fisher’s exact test p-values,0.05 (*), ,0.001 (***).
strains of wMel-infected mosquitoes in a separate experiment. Five doi:10.1371/journal.pntd.0002688.t001

PLOS Neglected Tropical Diseases | www.plosntds.org 3 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

Figure 1. Experiment 1: DENV replication in wildtype and field-released (wMel.F) A. aegypti. DENV replication in bodies (A) and heads (B)
of mosquitoes challenged with three strains (DENV2-92T, DENV1-P249, DENV2-P410), assayed at 14 days post-infection. DENV levels determined
using one-step qRT-PCR and expressed as copies per 1 mg of total RNA. Bars denote medians. P,0.05 (*), P,0.01 (**), P,0.001 (***). Each point
represents an individual mosquito.
doi:10.1371/journal.pntd.0002688.g001

Fisher’s exact tests. P-values were adjusted for multiple compar- between lines were analyzed using Mann-Whitney U tests. In
isons for each day of sampling within each experiment using experiment 2, differences among the three lines in copies of each
the Holm method [25], with values ,0.05 considered significant. virus were analyzed using Kruskal-Wallis tests, with Dunn’s post-
In experiment 1, differences in median DENV copy numbers hoc multiple comparison tests. Last, we tested for significant

Table 2. Rates of infection (%) for three DENV strains among field (wMel.F) and laboratory Wolbachia-infected (MGYP2.out) and
uninfected (wildtype) mosquito lines at days 7 and 14 post-infection (p.i.) (experiment 2).

Body infection (N) Head infection (N)

wildtype wMel.F MGYP2.out wildtype wMel.F MGYP2.out

day 7 p.i.
DENV1 – P307 23(13) 12 (17) 10 (21) 8 (13) 6 (17) 5 (21)
DENV2 – 92T 54 (13) 0 (16)* 13 (15) 8 (12) 0 (16) 7 (15)
DENV3 – Cairns08 58 (12) 6 (17)* 14 (14) 25 (12) 0 (17) 7 (14)
day 14 p.i.
DENV1 – P307 65 (17) 15 (20)* 41 (17) 65 (17) 5 (20)*** 29 (17)
DENV2 – 92T 77 (13) 12 (17)* 13 (15) 69 (13) 6 (17)*** 7 (14)
DENV3 – Cairns08 92 (13) 6 (17)*** 9 (22) 77 (13) 6 (17)*** 0 (22)

Adjusted Fisher’s exact test p-values,0.05 (*), ,0.001 (***). P-values shown refer to comparisons between wildtype and wMel.F mosquitoes.
doi:10.1371/journal.pntd.0002688.t002

PLOS Neglected Tropical Diseases | www.plosntds.org 4 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

Figure 2. Experiment 2: DENV replication in wildtype, outbred laboratory wMel (MGYP2.out) and field-released wMel (wMel.F) A.
aegypti. DENV replication in bodies (A) and heads (B) of mosquitoes challenged with three strains (DENV2-92T, DENV1-P307, DENV3-Cairns08/09),
assayed at 14 days post-infection. DENV levels determined using one-step qRT-PCR and expressed as copies per 1 mg of total RNA. Bars denote
medians. P,0.05 (*), P,0.01 (**), P,0.001 (***). Each point represents an individual mosquito.
doi:10.1371/journal.pntd.0002688.g002

differences in Wolbachia density between MGYP2.out and wMel.F DENV replication (Figure 1). A similar difference in virus titers
mosquitoes using Mann-Whitney U tests. All analyses were per- was present at day 7 p.i., but to a lesser extent because of low
formed in R [26] and GraphPad Prism v. 6 (GraphPad Software, infection rates (Figure S1).
San Diego, California USA). We next investigated whether vector competence was similar in
field wMel-infected A. aegypti compared to the original wMel-
Results infected line that had been maintained in the lab with recurrent
outbreeding [9]. In experiment 2, we estimated DENV infection
Limited DENV infection and replication in field wMel- rates and replication titers for three virus strains in wildtype,
infected mosquitoes wMel.F and MGYP2.out mosquitoes. We tested for statistically
We conducted two independent experiments to assess rates of significant differences in infection rates only between wildtypes
DENV infection and replication in wildtype and wMel-infected and wMel.F, and between wMel.F and MGYP2.out mosquitoes
field release mosquitoes. In experiment 1, at day 7 p.i., lower rates (Table 2). At day 7 p.i., significantly lower body infection rates
of body and head infection were detected in field wMel mosquitoes were found in wMel.F mosquitoes versus wildtypes for DENV-2-
compared to wildtype for the two DENV-1 and DENV-2 viremic 92T and DENV-3-Cairns08 strains (Table 2). However, rates of
plasma samples and cell culture DENV-2-92T virus strains infection across all mosquito lines and all viruses were low in
(Table 1). However, only for DENV-2 strain P410, a viremic general, resulting in limited power for robust statistical tests. At
plasma sample, was there a statistically significant difference day 14 p.i., significantly different infection rates between wildtypes
between the two mosquito lines (Table 1). At day 14 p.i., rates of and wMel.F mosquitoes were found for both bodies and heads
body and head infection were significantly lower in field wMel across all DENV strains (Table 2). For both experiments 1 and 2,
compared to wildtype mosquitoes for all three DENV strains, with dissemination of all virus strains by day 14 p.i. was dramatically
a stronger effect in dissemination to heads (Table 1). The highest lower in field wMel mosquitoes compared to wildtypes. There
observed dissemination rate in wMel.F heads was a low 6%, were no significant differences in infection rates between wMel.F
compared to 62% in wildtype heads. DENV genome copy titers in and MGYP2.out mosquitoes across either day post-infection.
heads and bodies were uniformly higher for all strains in wildtype DENV titers were significantly lower across all virus strains in
mosquitoes compared to respective wMel.F samples at day 14 p.i. both heads and bodies in field wMel mosquitoes compared to
(Figure 1). For example, titers in both bodies and heads typically wildtypes, at day 14 post-infection (Figure 2). A similar pattern
reached 16108 copies for all virus strains in wildtype individuals. was observed at day 7 post-infection, although only for bodies and
By contrast, most wMel.F individuals showed an absence of the strains DENV-2-92T and DENV-2-Cairns08/09 (Figure S2).

PLOS Neglected Tropical Diseases | www.plosntds.org 5 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

Figure 3. Localization of Wolbachia in different A. aegypti tissues visualized using FISH. Outbred laboratory wMel (MGYP2.out) (A, C, G, E)
and field-released wMel (wMel.F) (B, D, F, H) mosquitoes at day 7 post DENV infection. Wolbachia stained in red (Alexa 594) and cell nuclei in blue

PLOS Neglected Tropical Diseases | www.plosntds.org 6 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

(DAPI). Images are representative of 4–5 mosquitoes per line. Bars represent 50 mM scale.
doi:10.1371/journal.pntd.0002688.g003

At day 14, virus titers in wildtype mosquitoes ranged from below non-blood fed mosquitoes persisted at 14 days post feeding
the limit of detection to 108 copies/mg of RNA whereas virus was (Figure 4 B), in the absence of repeat feeds. Median ratios of wsp
observed only in a few instances in field wMel. Only in one field to RpS17 gene copy numbers were 0.649 and 0.733 in non-blood
wMel individual was the maximum number of DENV copies fed wMel.F and MGYP2.out, respectively, compared to 1.542 and
observed (Figure 2, strain 92T body and heads panels). Overall, 1.675 in blood-fed wMel.F and MGYP2.out, respectively (Figure 4
the results indicate that when breakthrough virus occurs in wMel B). Interestingly, by day 14, Wolbachia density continued to increase
mosquitoes, viral titers are most likely to be lower than those in blood-fed MGYP2.out and field wMel mosquitoes compared to
observed in wildtypes. non-blood fed ones, as indicated by the slightly higher median
values of normalized wsp/RpS17 ratios (Figure 4 B). Following
Wolbachia tissue tropism and density in field mosquitoes blood-feeding, increases in Wolbachia density in both field and
We next investigated whether Wolbachia tissue tropism and laboratory lines were primarily localized in the bodies rather than
density had changed significantly in field wMel mosquitoes since heads (Figure 5), probably due to the bacteria replicating in
release in 2011. Using FISH, we found that Wolbachia was ovaries.
distributed in the same tissues in field mosquitoes and in the
original wMel-transinfected laboratory line, MGYP2.out (Fig- Discussion
ure 3). In both wMel-infected lines, Wolbachia was present in two
tissues that are critical in viral infection and dissemination, namely Infection of the vector A. aegypti with Wolbachia has been pro-
midguts and salivary glands (Figure 3 A–B & G–H). Wolbachia posed as a dengue biocontrol method that is environmentally
was also present in brains, although not at high densities which friendly and able to spread unassisted in wild mosquito popula-
was consistent with levels expected for the wMel strain [9]. Field tions. Release of wMel-infected mosquitoes in north Queensland
wMel ovaries appeared highly infected with Wolbachia (Figure 3 has indicated that this Wolbachia strain can rapidly reach fixation
D), indicating the potential for stable transmission of the bacteria in wild populations [12]. Key to the utility of this biocontrol
to offspring in the wild. method is the maintenance of DENV-blocking following mosquito
We also examined whether Wolbachia densities change following release and in subsequent generations as Wolbachia invades wild
blood-feeding in field wMel mosquitoes compared to the original populations.
MGYP2.out line. By initially looking at whole mosquitoes we Our results indicate that, one year post-release, field wMel
found that, by day 7, the density of wMel had increased following mosquitoes show significantly reduced DENV infection and repli-
blood-feeding in both lines (Figure 4 A). A much higher increase cation compared to wildtype mosquitoes. Strikingly, we found very
in Wolbachia density was observed in field wMel mosquitoes versus low infection rates in mosquito heads, indicating that DENV is
MGYP2.out (Figure 4 A). Median ratios of wsp to RpS17 gene largely unable to disseminate to the heads in wMel mosquitoes,
copy numbers increased significantly from 0.714 and 0.702 in under the experimental conditions used here. By day 14, in both
non-blood fed wMel.F and MGYP2.out, respectively, to 1.465 and experiments, wMel mosquitoes displayed dramatically reduced
1.241 in blood-fed wMel.F and MGYP2.out, respectively (Figure 4 infection rates and viral titers in heads compared to wildtype.
A). The difference in Wolbachia density between blood-fed and Reduced DENV dissemination and transmission rates due to the

Figure 4. Blood-feeding and Wolbachia densities in whole mosquitoes. Outbred laboratory wMel (MGYP2.out) and field-released wMel
(wMel.F) A. aegypti at 7 (A) and 14 (B) days post blood-feeding (BF) versus non-blood fed (NBF) controls. Bars denote medians. P,0.05 (*), P,0.01 (**),
P,0.001 (***). Each point represents an individual mosquito.
doi:10.1371/journal.pntd.0002688.g004

PLOS Neglected Tropical Diseases | www.plosntds.org 7 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

Figure 5. Blood-feeding and Wolbachia densities in mosquito heads and bodies. Bodies (A) and heads (B) of outbred laboratory wMel
(MGYP2.out) and field-released wMel (wMel.F) A. aegypti at days 7 and 14 post blood-feeding. Bars denote medians. P,0.05 (*), P,0.01 (**), P,0.001
(***). Each point represents an individual mosquito.
doi:10.1371/journal.pntd.0002688.g005

presence of native Wolbachia endosymbionts have also been found contribute to the limited dissemination and replication of DENV
in the vector A. albopictus [27]. The pattern was observed with a observed in heads from the wMel-infected lines. In Drosophila
range of virus titers and serotypes (DENV-1 to -3), and using both simulans, high Wolbachia densities in head and midgut have been
cell-cultured and viremic human plasma. We did not test for correlated with interference against Drosophila C virus [30].
systematic differences in response to these variables here, but work Wolbachia density is critical in modulating transmission fidelity of
with other viruses has indicated the extent of Wolbachia-mediated the bacterium across generations and pathogenicity [19]. Wolbachia
viral blocking is dependent on virus titer [28]. density changes dynamically in response to environmental
Our data suggest stability of viral blocking and Wolbachia tissue variables [31]. We also found that Wolbachia density increased
tropism since divergence of field mosquitoes from the parental following blood-feeding, consistent with other studies that have
wMel-transinfected laboratory line MGYP2.out. We did not find shown an increase in endosymbiont density in response to high
statistically significant differences in either dengue infection rates nutrient conditions [32]. Wolbachia provides a fitness benefit by
or virus titers between field wMel and MGYP2.out mosquitoes. modulating iron levels in D. melanogaster [33] and responds trans-
However, field wMel mosquitoes may be somewhat better at criptionally to iron overload [34]. Increased Wolbachia replication
blocking dissemination of DENV-1 but not DENV-2 and DENV- is most likely localized to the ovaries, although further work is
3 compared to MGYP2.out (Figure 2). This is because the needed to confirm this. Our results differ, however, from those of
number of MGYP2.out individuals infected with virus is much [35], who showed a blood-feeding induced reduction in the native
higher for DENV-1 than DENV-2 and DENV-3 compared to endosymbiont wFlu in the ovaries of the mosquito Aedes fluviatilis.
field mosquitoes. Virus was detected in a higher number of Surprisingly, the increase in Wolbachia density was more pro-
MGYP2.out individuals for DENV-1 strain P307, compared to the nounced in field wMel mosquitoes compared to the laboratory
other virus strains tested. Additional experiments are needed to line, although only at day 7 post-infection. The reasons for this
determine whether this effect is due to the particular strain or a difference are unknown but may be related to poor nutrition in the
phenomenon general to the DENV-1 serotype. DENV-2-92T field or other environmental effects. Although mosquitoes were
dissemination rates in MGYP2.out were 12.5% several genera- reared in the same environment for one generation, maternal
tions after transinfection in earlier work [9] and have stayed a low nutritional effects can be detected up to several generations later in
7% in our study, at least 10 generations later and with frequent insects [36,37]. Maternal effects due to poor nutrition in the field
outcrossing of this line (every three generations). This time frame is may influence offspring immune status and the ability to control
comparable with that experienced by field mosquitoes, with the infection levels, potentially resulting in higher Wolbachia densities.
maximum number of generations per year in Cairns being 15 and Dynamic changes in Wolbachia density following blood-feeding
populations persisting throughout the year [29]. MGYP2.out and may have implications for vector competence of wMel-infected
field wMel-infected mosquitoes have therefore retained the virus mosquitoes. The precise mechanism by which Wolbachia mediates
blocking phenotype described in [9] that led to the field release of viral blocking is not known but is positively related to density of the
Wolbachia-infected mosquitoes. Our results suggest that the virus bacterium [13,14,38]. If blood-feeding acts to increase Wolbachia
blocking phenotype induced by wMel may be retained not just density and A. aegypti feed frequently on human hosts, viral
over the short term, but also over the medium to longer term. blocking may be greater in field populations than anticipated from
Wolbachia tissue tropism was similar in field and laboratory laboratory experiments, although further studies are needed to test
wMel-infected mosquitoes, with high densities of the bacterium this hypothesis. In laboratory experiments involving Drosophila, the
found in the midgut and ovaries. Wolbachia was also present in the density of Wolbachia has been shown to evolve to a level that is non-
salivary glands and brains of both mosquito lines, which may pathogenic to the fly but the bacteria are still maintained [19,39].

PLOS Neglected Tropical Diseases | www.plosntds.org 8 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

Understanding selection pressures on wMel-infected mosquitoes in Figure S2 DENV replication in bodies (A) and heads (B) of
nature will be necessary to predict how Wolbachia may evolve over wildtype, outbred laboratory wMel (MGYP2.out) and field-
the long term in field-released mosquitoes. released wMel (wMel.F) A. aegypti challenged with three strains
A. aegypti infected with Wolbachia show reduced replication of (DENV2-92T, DENV1-P307, DENV3-Cairns08/09), assayed at 7
other RNA viruses, such as yellow fever [28], chikungunya [8,28] days post-infection (experiment 2). DENV levels determined using
and West Nile [40] viruses. Wolbachia-based biocontrol may one-step qRT-PCR and expressed as copies per 1 mg of total
therefore have the potential to eliminate transmission of old and RNA. Bars denote medians. P,0.05 (*), P,0.01 (**), P,0.001
emerging arboviruses in addition to DENV. The maintenance of (***). Each point represents an individual mosquito.
virus blocking in field release mosquitoes is critical to the success of (TIF)
Wolbachia-based biocontrol. Our results show that dengue virus
blocking and Wolbachia density phenotypes have stayed stable in Acknowledgments
A. aegypti infected with wMel, at least 12 months following field
release. We thank Melinda Greenfield, Fred Muzzi and Brian Montgomery from
Eliminate Dengue in Cairns and Nichola Kenny, Inaki Iturbe-Ormaetxe
and Jyotika Taneja de Bruyne for field and technical support. We also
Supporting Information thank Cameron Simmons for the generous gift of Vietnamese viremic
Figure S1 DENV replication in bodies (A) and heads (B) of plasmas.
wildtype and field-released wMel (wMel.F) A. aegypti challenged
with three strains (DENV2-92T, DENV1-P249, DENV2-P410), Author Contributions
assayed at 7 days post-infection (experiment 1). DENV levels Conceived and designed the experiments: FDF EAM SLO. Performed the
determined using one-step qRT-PCR and expressed as copies per experiments: FDF TZ TW. Analyzed the data: FDF. Contributed
1 mg of total RNA. Bars denote medians. P,0.05 (*), P,0.01 (**), reagents/materials/analysis tools: JP ATP AvdH. Wrote the paper: FDF
P,0.001 (***). Each point represents an individual mosquito. EAM SLO.
(TIF)

References
1. Bhatt S, Gething PW, Brady OJ, Messina JP, Farlow AW, et al. (2013) The 19. McGraw EA, Merritt DJ, Droller JN, O’Neill SL (2002) Wolbachia density and
global distribution and burden of dengue. Nature 496: 504–507. virulence attenuation after transfer into a novel host. Proc Natl Acad Sci U S A
2. Kyle JL, Harris E (2008) Global spread and persistence of dengue. Annu Rev 99: 2981–2923.
Microbiol 62: 71–92. 20. Hoffmann AA, Turelli M, Harshman LG (1990) Factors affecting the
3. Sabchareon A, Wallace D, Sirivichayakul C, Limkittikul K, Chanthavanich P, distribution of cytoplasmic incompatibility in Drosophila simulans. Genetics 126:
et al. (2012) Protective efficacy of the recombinant, live-attenuated, CYD 933–948.
tetravalent dengue vaccine in Thai schoolchildren: a randomised, controlled 21. Rancès E, Ye YH, Woolfit M, McGraw EA, O’Neill SL (2012) The relative
phase 2b trial. The Lancet 380: 1159–1567. importance of innate immune priming in Wolbachia-mediated dengue interfer-
4. Iturbe-Ormaetxe I, Walker T, O’Neill SL (2011) Wolbachia and the biological ence. PLoS Pathog 8: e1002548.
control of mosquito-borne disease. EMBO Reports 12: 508–518. 22. Hue KD, Tuan TV, Thi HT, Bich CT, Anh HH, et al. (2011) Validation of an
5. McGraw EA, O’Neill SL (2013) Beyond insecticides: new thinking on an ancient internally controlled one-step, real-time multiplex RT-PCR assay for the
problem. Nat Rev Microbiol 11: 181–193. detection and quantitation of dengue virus RNA in plasma. J Virol Methods
6. Zug R, Hammerstein P (2012) Still a host of hosts for Wolbachia: analysis of 177: 168–173.
recent data suggests that 40% of terrestrial arthropod species are infected. PLoS 23. McMeniman CJ, Lane AM, Fong AW, Voronin DA, Iturbe-Ormaetxe I, et al.
One 7: e38544. (2008) Host adaptation of a Wolbachia strain after long-term serial passage in a
7. Bian G, Xu Y, Lu P, Xie Y, Xi Z (2010) The endosymbiotic bacterium Wolbachia mosquito cell line. Appl Environ Microbiol 74: 6963–6969.
induces resistance to dengue virus in Aedes aegypti. PLoS Pathog 6: e1000833. 24. Cook PE, Hugo LE, Iturbe-Ormaetxe I, Williams CR, Chenoweth SF, et al.
8. Moreira LA, Iturbe-Ormaetxe I, Jeffery JA, Lu G, Pyke AT, et al. (2009) A (2006) The use of transcriptional profiles to predict adult mosquito age under
Wolbachia symbiont in Aedes aegypti limits infection with Dengue, Chikungunya, field conditions. Proc Natl Acad Sci U S A 103: 18060–18065.
and Plasmodium. Cell 139: 1268–1278. 25. Holm S (1979) A simple sequentially rejective multiple test procedure. Scand
9. Walker T, Johnson PH, Moreira LA, Iturbe-Ormaetxe I, Frentiu FD, et al. Stat Theory Appl 6: 65–70.
(2011) The wMel Wolbachia strain blocks dengue and invades caged Aedes aegypti 26. R Development Core Team (2010) R: A language and environment for
populations. Nature 476: 450–453. statistical computing. . Vienna, Austria: R Foundation for Statistical Computing.
10. Werren JH, Baldo L, Clark ME (2008) Wolbachia: master manipulators of 27. Mousson L, Zouache K, Arias-Goeta C, Raquin V, Mavingui P, et al. (2012)
invertebrate biology. Nat Rev Microbiol 6: 741–751. The native Wolbachia symbionts limit transmission of dengue virus in Aedes
11. Popovici J, Moreira LA, Poinsignon A, Iturbe-Ormaetxe I, McNaughton D, albopictus. PLoS Negl Trop Dis 6: e1989.
et al. (2010) Assessing key safety concerns of a Wolbachia-based strategy to control 28. van den Hurk AF, Hall-Mendelin S, Pyke AT, Frentiu FD, McElroy K, et al.
dengue transmission by Aedes mosquitoes. Memorias do Instituto Oswaldo Cruz (2012) Impact of Wolbachia on infection with chikungunya and yellow fever
105: 957–964. viruses in the mosquito vector Aedes aegypti. PLoS Negl Trop Dis 6: e1892.
12. Hoffmann AA, Montgomery BL, Popovici J, Iturbe-Ormaetxe I, Johnson PH, 29. Kearney M, Porter W, Williams C, Ritchie S, Hoffmann A (2009) Integrating
et al. (2011) Successful establishment of Wolbachia in Aedes populations to biophysical models and evolutionary theory to predict climatic impacts on
suppress dengue transmission. Nature 476: 454–457. species’ ranges: the dengue mosquito Aedes aegypti in Australia. Funct Ecol 23:
13. Frentiu FD, Robinson J, Young PR, McGraw EA, O’Neill SL (2010) Wolbachia- 528–538.
mediated resistance to dengue virus infection and death at the cellular level. 30. Osborne SE, Iturbe-Ormaetxe I, Brownlie JC, O’Neill SL, Johnson KN (2012)
PLoS One 5: e13398. Antiviral protection and the importance of Wolbachia density and tissue tropism
14. Lu P, Bian G, Pan X, Xi Z (2012) Wolbachia induces density-dependent in Drosophila simulans. Appl Environ Microbiol 78: 6922–6929.
inhibition to dengue virus in mosquito cells. PLoS Negl Trop Dis 6: e1754. 31. Mouton L, Henri H, Charif D, Boulétreau M, Vavre F (2007) Interaction
15. Kambris Z, Cook PE, Phuc HK, Sinkins SP (2009) Immune activation by life- between host genotype and environmental conditions affects bacterial density in
shortening Wolbachia and reduced filarial competence in mosquitoes. Science Wolbachia symbiosis. Biol Lett 3: 210–213.
326: 134–136. 32. Dutton TJ, Sinkins SP (2004) Strain-specific quantification of Wolbachia density
16. Caragata E, Rancès E, Hedges L, Gofton A, Johnson K, et al. (2013) Dietary in Aedes albopictus and effects of larval rearing conditions. Insect Mol Biol 13:
cholesterol modulates pathogen blocking by Wolbachia. PLoS Pathog 9: 317–322.
e1003459. 33. Brownlie J, Cass B, Riegler M, Witsenburg J, Iturbe-Ormaetxe I, et al. (2009)
17. Rancès E, Ye Y, Woolfit M, McGraw E, O’Neill S (2012) The relative Evidence for metabolic provisioning by a common invertebrate endosym-
importance of innate immune priming in Wolbachia-mediated dengue interfer- biont, Wolbachia pipientis, during periods of nutritional stress. PLoS Pathog 5:
ence. PLoS Pathog 8: e1002548. e1000368.
18. Zhang G, Hussain M, O’Neill SL, Asgari S (2013) Wolbachia uses a host 34. Kremer N, Voronin D, Charif D, Mavingui P, Mollereau B, et al. (2009)
microRNA to regulate transcripts of a methyltransferase contributing to dengue Wolbachia interferes with ferritin expression and iron metabolism in insects. PLoS
virus inhibition in Aedes aegypti. Proc Natl Acad Sci U S A 110:10276–81. Pathog 5: e1000630.

PLOS Neglected Tropical Diseases | www.plosntds.org 9 February 2014 | Volume 8 | Issue 2 | e2688
 DENV Replication in Field Wolbachia Mosquitoes

35. Baton LA, Pacidônio EC, da Silva Gonçalves D, Moreira LA (2013) wFlu: 38. Osborne SE, Leong YS, O’Neill SL, Johnson KN (2009) Variation in antiviral
Characterization and evaluation of a native Wolbachia from the mosquito Aedes protection mediated by different Wolbachia strains in Drosophila simulans. PLoS
fluviatilis as a potential vector control agent. PLoS One 8: e59619. Pathog 5: e1000656.
36. Lorenz LM, Koella JC (2011) Maternal environment shapes the life history and 39. Correa CC, Ballard JW (2012) Wolbachia gonadal density in female and male
susceptibility to malaria of Anopheles gambiae mosquitoes. Malar J 10: art. no. 382. Drosophila vary with laboratory adaptation and respond differently to
37. Valtonen TM, Kangassalo K, Pölkki M, Rantala MJ (2012) Transgenerational physiological and environmental challenges. J Invertebr Pathol 111: 197–204.
effects of parental larval diet on offspring development time, adult body size and 40. Hussain M, Lu G, Torres S, Edmonds JH, Kay BH, et al. (2013) Effect of
pathogen resistance in Drosophila melanogaster. PLoS One 7: e31611. Wolbachia on replication of West Nile virus in a mosquito cell line and adult
mosquitoes. J Virol 87: 851–858.

PLOS Neglected Tropical Diseases | www.plosntds.org 10 February 2014 | Volume 8 | Issue 2 | e2688