Copyright ? 2008 by the Genetics Society of America
Use of a Drosophila Model to Identify Genes Regulating Plasmodium
Growth in the Mosquito
Stephanie M. Brandt,*,1Giovanna Jaramillo-Gutierrez,†Sanjeev Kumar,†
Carolina Barillas-Mury†and David S. Schneider*,2
*Department of Microbiology and Immunology, Stanford University, Stanford, California 94305 and†Laboratory of Malaria and Vector
Research, NIAID, National Institutes of Health, Rockville, Maryland 20892-8130
Manuscript received March 30, 2008
Accepted for publication August 19, 2008
required for the growth of the avian malaria parasite, Plasmodium gallinaceum. We identified 18 presumed
loss-of-function mutants that reduced the growth of the parasite in flies. Presumptive mutation sites were
identified in 14 of the mutants on the basis of the insertion site of a transposable element. None of the
identified genes have been previously implicated in innate immune responses or interactions with
Plasmodium. The functions of five Anopheles gambiae homologs were tested by using RNAi to knock down
gene function followed by measuring the growth of the rodent parasite, Plasmodium berghei. Loss of function
offourofthesegenesinthemosquitoaffected Plasmodium growth,suggesting thatDrosophilacanbeused
effectively as a surrogate mosquito to identify relevant host factors in the mosquito.
and .1 million deaths annually, mostly among African
children. Recent advances in RNA interference (RNAi)
technology make it possible to manipulate mosquito
gene function in vivo and to determine the effects of
mosquito gene regulation on the ability to support the
growth of Plasmodium. The field now faces the prob-
lem of choosing which mosquito genes should be
tested. Three methods have been used successfully to
date: First, genetic mapping of vector traits led to the
identification of variant alleles that affect vector com-
petence (Kumar et al. 2003). Second, transcript analysis
during Plasmodium infections revealed genes that
show significant changes in transcript levels and knock-
down of some of these genes affects vector capacity
(Levashina et al. 2001; Dong et al. 2006). Third, our
understanding of the mosquito immune system is built
upon analyses that were performed on another dip-
teran insect, the fruit fly Drosophila melanogaster
(Zdobnov et al. 2002). Direct analysis of genes shown
to be important in fighting microbial infections in
Drosophila have also contributed to our understanding
of the interaction between the mosquito host and
Plasmodium parasite (Meister et al. 2005).
ALARIA is transmitted by anopheline mosqui-
toes, resulting in 300–500 million clinical cases
The movement and development of Plasmodium in
microbe and results in substantial parasite losses. For
every thousand Plasmodium berhgei gametocytes in-
gested, only two viable ookinetes will be generated,
and only 2–20% of these will successfully invade the
midgut and develop to mature oocysts (Alavi et al.
2003). The parasites are introduced into the mosquito
via an infected blood meal. Plasmodium gametocytes
complete their development in the midgut lumen,
fertilization takes place, and a diploid motile form
called ookinete is generated. Ookinetes invade the
midgut epithelium and come in direct contact with
components of the mosquito immune system present in
hemolymph as they reach the midgut basal lamina.
Multiplecomplex interactionsdetermine whether para-
sites will be recognized and destroyed, or survive and
transform into oocysts. During the oocyst stage, para-
sites divide continuously and ultimately release thou-
sands of sporozoites into the hemolymph. Sporozoites
find and invade the salivary gland, reach the salivary
duct, and are injected into a new host during the
following blood feeding.
We can anticipate two general types of physiological
changes in the mosquito that will alter Plasmodium
growth. First, the mosquito can raise an active immune
response against the parasite because it does not act
simply as a passive culture vessel. Alterations in innate
immunity have been shown to affect parasite growth.
The second concerns the metabolic needs of the
parasite. Plasmodium is dependent upon the mosquito
for nutrition and changes in available nutrients in the
1Present address: Institute of Environmental Science and Research,
Christchurch Science Center, 27 Creyke Rd., Christchurch 8540, New
2Corresponding author: D333 Fairchild Bldg., 299 Campus Dr., Stanford
University, Stanford, CA 94305-5124.E-mail: email@example.com
Genetics 180: 1671–1678 (November 2008)
host will presumably affect parasite development in
because the parasite population expands during the
oocyst stage. Dissecting the interaction between host
and parasite can teach us about innate immune mech-
anisms that can counter Plasmodium and can also pro-
vide new information about the growth requirements of
the parasite in the insect host.
Every genetic technique has its problems; data min-
ing from microarrays is useful but, by design, will miss
genes that are not regulated in response to infection
and are not obviously involved in immunity. The tissues
affected by RNAi injection into a whole mosquito are
not well defined and the efficacy of each probe is
different; thus a negative result for an RNAi experiment
in a given process. RNAi experiments require a good
deal of work and thus researchers think carefully about
which genes should be tested first. A result of these
issues is that the community tends to select genes for
study that have already been implicated in fighting
Plasmodium in some manner. It is possible that we are
systematically missing genes that either are not tran-
scriptionally regulated during an infection or have not
been previously identified as playing arole in the innate
immune response. An ideal method of overcoming this
gap is to perform an unbiased genetic screen because it
allows the organism to tell you what is important in the
Given that our understanding of the mosquito’s
innate immune system is still largely scaffolded on the
mosquito in an unbiased forward genetic screen to find
insect factors affecting Plasmodium growth. We showed
previously that the fruit fly could support the growth of
Plasmodium gallinaceum injected into the body cavity,
supporting the development of large numbers of oocysts
that complete maturation and generate infective spor-
ozoites (Schneider and Shahabuddin 2000).
We found 18 mutations that reduce the ability of the
fly to support Plasmodium growth. Fourteen of these
mutations were induced by P-element insertions and
these transposons identify a set of genes that had no
previously identified role in immunity and are not
immune regulated. One risk of this Drosophila model
is that it is too far removed from the natural host–
parasite pair to provide useful information; however, we
that alter the growth of P. berghei in Anopheles gambiae.
MATERIALS AND METHODS
Mutant lines tested in genetic screen: A total of 1045
homozygous viable P-element insertion lines from the Bloom-
ington Drosophila Stock Center were tested in the genetic
screen. These lines were generated by many labs in different
tested 407 ethyl methanesulfonate (EMS)-mutagenized fly
lines. These mutagenized lines were generated on an isogen-
solution of 2.5 mm EMS (Sigma-Aldrich, St. Louis) in 1%
sucrose overnight. The treated males were then mated to
X^X/Y (attached-X) females. Individual male progeny where
then mated to X^X/Y females to generate separate mutant
Ookinete preparation: P. gallinaceum parasites were main-
tained in white leghorn chickens (Charles River Laboratories
SPAFAS, Wilmington, MA) by serial passage. Parasites were
isolated from chicken blood by previously described methods
(Schneider and Shahabuddin 2000). Approximately 10 ml
of heparinized blood was obtained from infected 3-week-old
white leghorn chicks by cardiac puncture. The blood was
diluted in warm 13 SAB (9 mm glucose, 8 mm Tris base, 138
serum by centrifugation at 3000 rpm for 5 min at room
temperature. Blood cells were then incubated for 45 min in
15 ml exflagellation medium (8% heat-inactivated chicken
serum, 0.1 mm xanthurenic acid, 0.2% sodium bicarbonate in
13 SAB). Fertilized zygotes were then separated from the
blood cells; ?15ml blood was carefully layeredon topof 15ml
Ficoll-paque (Amersham Biosciences, Uppsala, Sweden) and
spun for 15 min at 3500 rpm. The zygotes were removed from
the Ficoll interface and resuspended in 50 ml 13 SAB. The
zygotes were then spun down at 3500 rpm and the SAB was
removed. The zygotes were resuspended in 10 ml 13 M199,
pH8.1, pluspenicillin/streptomycinand incubatedwith0.7 ml
of 1 mg/ml wheat germ agglutinin (Sigma-Aldrich) for 10 min
at room temperature. The agglutinated blood cells were
pelleted at 700 rpm for 30 sec. The supernatant was removed
and the agglutination procedure repeated once more with an
additional 0.7 ml wheat germ agglutinin. The supernatant was
then spun for 10 min at 3500 rpm. The supernatant was
removed andthepellet containingzygotes wasthen incubated
overnight in 13 M199 media, pH 8.1, plus penicillin (100
Carlsbad, CA). The next morning, differentiated ookinetes
concentration of 500 ookinetes/50 nl of media.
Genetic screen injection protocol: Fly stocks were main-
tained on standard yeasted dextrose food at 25?, 60%
humidity. Only male flies were used in the genetic screen to
avoid progeny growing in the vials. Male flies were aged 5–7
days at 25? before infection. Ten to 15 male flies from each
stock were then injected with 500 P. gallinaceum ookinetes.
Parasite suspension (50 nl) was injected into the fly’s anterior
abdomen on the ventrolateral surface. Injections were per-
formed using pulled glass needles and a Picospritzer III
injector (Parker Hannifin, Rohnert Park, CA). Following
inoculation, flies were incubated at 25? for 9 days and then
harvested to measure parasite load.
Measuring P. gallinaceum load by quantitative real-time
PCR: P. gallinaceum loads were determined in each line by
five flies from each line using a 96-well format RNeasy kit
(Qiagen, Valencia, CA). The RNA was then treated with
DNAse. qRT–PCR reactions were set up using the hetero-
bifunctional rTth polymerase (Applied Biosystems, Foster
City, CA) and run in an iCycler (Bio-Rad, Hercules, CA).
Copies of P. gallinaceum small-unit ribosomal RNA (surRNA)
were measured using the following oligomers: (59 oligo
pgrRNAbeg) 59-agagttcgattccggagagg-39, (39 oligo pgrRNA-
beg) 59-tcattccaattgcaaaacca-39, and (hybridization oligo
1672S. M. Brandt et al.
surRNA values were normalized to the number of copies of D.
melanogaster ribosomal protein 15a in each sample; expression
of 15a does not change during infection and therefore this
gene served as a control for the total amount of RNA in each
sample (Schneider and Shahabuddin 2000).
Bacterial infections: Flies were infected with either Salmo-
nella typhimurium strain SL1344 (Hoiseth and Stocker 1981)
or Listeria monocytogenes strain 10403s (Portnoy et al. 1988).
and brain heart infusion medium, respectively. One-week-old
male flies were injected as described above with bacteria (for
OD600¼ 0.01, ?1000 CFU). Flies were maintained thereafter
at 29?. All survival experiments were performed in triplicate
with 10 flies/replicate. Experiments were repeated at least
three times. Survival curves were plotted using the Kaplan–
Meier method. Significance was determined by comparing
survival curves of mutant flies to Oregon-R flies using the log-
Mosquito rearing and infections: An. gambiae strain G3
mosquitoes were raised at 28?, 75% humidity, under a 12-hr
light/dark cycle and maintained on a 10% sucrose solution
during adult stages. Mosquito females were infected with
P. berghei by feeding on anesthetized infected Balb/C mice 5–
6 days post-emergence. Mice were infected with P. berghei by
intraperitoneal injection of blood from an infected donor
mouse. The infection was monitored by determining the
parasitemia and by performing exflagellation assays as de-
scribed previously (Billker et al. 1997). Mice with para-
sitemias between 5 and 15% and one to three exflagellations
per 403 field were used to infect mosquitoes. Blood-fed
mosquitoes were kept at 21? in a humidified environment
and dissected 6 days post-infection.
Mosquito RNA isolation and qRT–PCR: Poly(A1)RNAwas
extracted from 10 whole mosquitoes with an Oligotex direct
mRNA kit (Qiagen, Valencia, CA). The first strand of cDNA
synthesis was performed using a Quantitect reverse transcrip-
tion kit (Qiagen). Gene-specific primers were designed using
Table S1). The PCR reactions were assembled using DyNAmo
SYBR Green qPCR master mix (Finnzymes) and run in the MJ
Research detection system according to the manufacturer’s
instructions (Bio-Rad). All values were then normalized to the
number of copies of ribosomal protein S7 in the sample as
determined by qRT–PCR.
Bioinformatics analysis: D. melanogaster sequences were
compared to the National Center for Biotechnology Informa-
tion database by using the Basic Local Alignment Search Tool
against the An. gambiae genome and translated database.
Compiled EST sequences were aligned by using CLUSTAL,
and gene-specific primers were designed in well-conserved
regions to confirm expression in An. gambiae.
Double-stranded RNA production: An An. gambiae cDNA
for each gene was obtained from 10 naive 4-day-old females.
for each gene with primers listed in supplemental Table S1.
Flanking promoter sequences were added to the gene-specific
primers M13 Fw (59-GTAAAACGACGGCCAGT-39) and Xho-
AACAGCTATGAC-39). The fragments were then cloned into
the pCRII-TOPO vector (Invitrogen, Carlsbad, CA). Double-
stranded RNA (dsRNA) was synthesized and purified with a
Megasuperscript kit (Ambion, Austin, TX) and eluted in
Relative quantification of P. berghei load in An. gambiae by
qRT–PCR: Excised individual midguts (without blood) were
placed into individual microcentrifuge tubes containing 10 ml
of HotSHOT alkaline lysis reagent (25 mm NaOH, 0.2 mm
EDTA, pH 12.0) (Truett et al. 2000). The tubes were boiled
for 5 min and immediately placed on ice. Ten microliters of
HotSHOT neutralizing reagent (40 mm Tris–HCl, pH 5.0) was
added to each tube. The samples were centrifuged and stored
Relative quantification of P. berghei by qRT–PCR: qRT–
PCR reactions were assembled using DyNAmo SYBR Green
qPCR master mix (Finnzymes) and run in the MJ Research
detection system according to the manufacturer’s instructions
(Bio-Rad). P. berghei-specific primers were designed and tested
[Pb (28SrRNA) Fw, 59-GTGGCCTATCGATCCTTTA-39; Pb
(28SrRNA) Rev, 59-GCGTCCCAATGATAGGAAGA-39] to de-
tecttherelativeamount ofP.berghei 28Sgenecopies.Template
midgut genomic DNA (2 ml) was used to detect P. berghei 28S
gene copies; 1 ml of template midgut genomic DNA was used
to detect An. gambiae ribosomal protein S7 gene copies in a
RESULTS AND DISCUSSION
To identify insect genes involved in regulating Plas-
modium infection, we performed a forward genetic
screen in the fruit fly primarily using publicly available
mutant stocks. We chose these lines because the trans-
poson insertion sites were already characterized, which
permitted rapid identification of the affected genes. We
injected mutant fly lines with P. gallinaceum ookinetes,
harvested RNA at the peak of infection on day 9 post-
injection, and measured parasite load by qRT–PCR. We
of mixed background from the Bloomington Drosoph-
ila Stock Center, and 407 EMS-mutagenized fly lines.
Due to the heterogeneity of the mutant fly lines that
we could compare parasite loads. Therefore, we identi-
fied interesting mutant lines by looking for large devia-
tions from the group mean. We tested mutant lines in
and a group average parasite load was determined. Lines
with at least 10-fold more or fewer parasites than the
group average were considered positive and were re-
tested; this was a conservative selection criterion, as the
10-fold cutoff from the group mean was well outside the
shown). During each round of retesting, positive lines
were tested with a different random group of 50–200
mutant lines. The final group of mutant lines consisted
of those that tested positive in three different rounds of
Of the 1452 mutant fly lines tested, 5% (70 lines)
tested positive in the first round (Table 1). Thirty-three
percent (23 lines) of the first-round positive lines
retested positive in the second round. Eighty percent
(18 lines) of the second-round positive lines retested
positive in the final round. The percentage of mutant
lines testing positive in successive rounds increased with
each round of screening, suggesting that our screening
Insect Host Genes Affecting Plasmodium Infection1673
lines contained 18 mutants: 14 P-element lines and
collection were resistant to P. gallinaceum and contained
fewer parasites than the average fly line. We did not
identify mutants permitting more parasite growth than
the average line in this screen.
Past work on the fly’s immune system revealed many
mutations that damage the fly’s immune response and
result in increased growth of infecting microbes. These
mutations typically block the antimicrobial immune
response of the fly and thus the fly does not fight infec-
tions well. When we started this screen, we anticipated
that we would find mutations that affected Plasmodium
that our screen revealed only mutants that reduced the
experimental reason, such as setting our threshold too
high when looking for mutants that allow increased
Plasmodium growth. A second explanation is that the
logic of our screen differed from many that have been
performed previously. Most fly immunity screens in the
past have looked for a disruption in the transcription of
2002). These mutants are immunocompromised and
permit higher growth rates of microbes. Few screens
have been performed that measure microbial growth or
survival or some other nonmechanistic output as a
primary assay (Ayres et al. 2008). The results of this
screen suggest that scientists studying Drosophila host–
pathogeninteractionsmay havebeen missing an impor-
tant class of mutants because they concentrate on a
single molecular mechanism controlling innate immu-
nity, rather than assaying a broad endpoint.
The fly’s immune response is best characterized with
respect to bacterial and fungal challenges (Ferrandon
affect immunity against many microbes (Shirasu-Hiza
of mutations might affect innate immunity in general,
we challenged our final collection of mutants with two
bacterial pathogens that cause lethal infections in wild-
type flies. We challenged mutant flies with the gram-
bacterium L. monocytogenes. We used these bacteria
becausethey growinwild-type fliesand arefoughtusing
different mechanisms (Mansfield et al. 2003; Brandt
et al. 2004). Mutants lacking Toll signaling are sensitive
to L. monocytogenes while mutants lacking Imd signaling
are sensitive to both S. typhimurium and L. monocytogenes.
because immune pathways can be missed if they are
tested only with nonpathogenic immune elicitors such
were not performed to test the contribution of Toll or
Imd signaling to Plasmodium growth as we previously
reported that these pathways play no role in fighting
Plasmodium (Schneider and Shahabuddin 2000);
rather, we were interested in challenging our mutants
with a variety of pathogens that require the fly to use
different diverse responses to fight the microbes. Sur-
vival ofinfected mutantflieswas comparedtosurvival of
wild-type Oregon-R flies. These data are summarized in
Genetic screen statistics
No. of mutant
% of previous
Total mutants tested
The selection of mutants through three rounds of screen-
ing is shown. It demonstrates a purification effect in which the
percentage of mutants reisolated during each round rises
from ?5% in the first round to ?80% in the final round.
Fly lines selected from genetic screen (fold-less from
The relative reduction in parasite numbers for each mutant
during the three rounds of screening is shown. The numbers
reported are the fold reduction with respect to the mean.
1674S. M. Brandt et al.
Table 3 and survival curves for fruit flies with altered
immunity are shown in Figure 1. Two EMS lines, GC15
line, 16497, was sensitive to L. monocytogenes. The
majority of the mutant fly lines behaved like Oregon-R
flies when challenged with either S. typhimurium or L.
We focused our attention on the P-element lines
because their insertion sites were known. We obtained
the location of the inserted P elements from FlyBase
(http:/ /flybase.net) or the P screen database (http:/ /
firmed a subset of five P-element insertion sites by
inverse PCR (Sullivan et al. 2000). We assumed the
putative gene affected in each P-element line to be the
gene closest to the location of the inserted P element
(Table 4). Half of the P elements sat within genes; two
P elements sat within 59 untranslated regions and five
P elements within introns. Three other P elements sat
within 500 bp upstream of the gene. Ten of 14 of these
genes had strong homologs in An. gambiae (Table 4).
The genes identified in the screen had diverse func-
in innate immune responses or the mosquito response
to Plasmodium; likewise, none of the genes had been
reported to genetically or biochemically interact with
each other or with immunity genes.
Although none of the genes that we identified in this
screen are obviously involved in immunity, it is possible
to generate plausible hypotheses about some of them.
The phospholipase A2, for example, is easy to implicate
in immunity because arachadonic acid release plays an
important role in the regulation of prostaglandin syn-
thesis in vertebrates and prostaglandins also play some
roles in insect immunity. Tetraspanins are important for
the activity of immune cells in humans. At this point, we
are not concentrating on the mechanism behind these
genes but on the phenotypes, for which we have much
identified in this Drosophila screen have been followed
previously in mosquitoes. These genes are not induced
during immune responses in the fly and either lack an
exciting identity or have no predictable function; these
genes would be missed by data mining yet they have
strong phenotypes in Drosophila. This screen demon-
strates the importance of performing random forward
genetics because experiments have the power to reveal
previously unexplored physiologies.
At this point, as the typical next step in a Drosophila
genetic screen, we would revert the mutations by hop-
ping out the P elements and rescuing the mutations us-
ing P-element-mediated transgenic flies. The end goal
of this experiment, however, was not to identify genes
that affected Plasmodium growth in Drosophila as that
is a very artificial system; rather, we wanted to identify
genes for further study in the mosquito. Therefore, we
directly tested whether mosquito orthologs of the genes
that we identified in the fly are also required to support
Five genes identified in the screen were carefully
analyzed by comparing all homologous genes present
in the An. gambiae, D. melanogaster, and Aedes aegypti
one orthologs in An. gambiae for four genes: l(3)82Fd
(similar to the Saccharomyces cerevisiae gene, oxidation
resistance 1), CG3168 (major facilitator transporter),
argk (arginine kinase), and Hsc70-3 (heat-shock pro-
tein) (supplemental Table S3). The tetraspanin family,
however, has undergone extensive gene expansion in
these three species, making it impossible to identify a
putative An. gambiae ortholog of Drosophila tsp42ej. To
explore this gene family, we decided to silence one
an An. gambiae-specific gene expansion. One-day-old G3
An. gambiae female mosquitoes were injected with
dsRNA to a target host gene of interest or lacZ as a
control; RNAi knockdown of target genes was con-
firmed by qRT–PCR (Figure 2A) . Four days later, mos-
quitoes were fed on anesthetized P. berghei-infected mice.
Six days after the feeding, P. berghei load was measured
by qRT–PCR (Figure 2B). RNAi knockdown of the
Secondary screen for susceptibility of fly lines to
type Sensitive Resistant
type Sensitive Resistant
All flies were ultimately killed by the infecting pathogen.
Wild type, resistant, or sensitive refer to changes in survival
relative to wild-type flies. Flies were injected with 50 nl S. typhi-
murium (OD600¼ 0.1) or L. monocytogenes (OD600¼ 0.01) and
incubated at 29?. Significance was determined by log-rank
test. n ¼ 30 flies. See Figure 1 for survival curves of lines that
were significantly sensitive or resistant to bacterial challenge.
Insect Host Genes Affecting Plasmodium Infection 1675
Figure 1.—Survival curves for P. gallinaceum-
resistant lines that are susceptible to bacterial in-
fection. Flies were injected with 50 nl S. typhimu-
rium (OD600¼ 0.1) or L. monocytogenes (OD600¼
0.01) and then incubated at 29?. Significance was
determined by a log-rank test. n ¼ 30 flies. (A)
GC15 S. typhimurium challenge. (B) GC15 L.
monocytogenes challenge. (C) iG1 S. typhimurium
challenge. (D) iG1 L. monocytogenes challenge.
(E) 11603 S. typhimurium challenge. (F) 11603
L. monocytogenes challenge. (G) 12516 S. typhimu-
rium challenge. (H) 12516 L. monocytogenes chal-
lenge. (I) 13425 S. typhimurium challenge. ( J)
13425 L. monocytogenes challenge. (K) 16497 S. ty-
phimurium challenge. (L) 16497 L. monocytogenes
1676 S. M. Brandt et al.
An. gambiae putative orthologs of oxr1 and argk resulted
in a significant decrease in the P. berghei parasite load,
P , 0.0001 by Kolmogorov–Smirnov test, which resem-
bles the phenotype that we found in Drosophila. We
found no significant difference in parasite load in
transporter) in comparison to control mosquitoes. Mos-
quitoes treated with AgTetraspanin or Hsc70-3 dsRNA
contained significantly more P. berghei than control
Insertion sites and putative identities of resistant Drosophila lines
Line Insertion siteClosest geneFunctionAnopheles homolog
Amino acid permease
lysM domain: peptidoglycan-binding hydrolase
The insertion sites for the P. gallinaceum-resistant mutants as recorded by FlyBase or the P screen database are shown. The iden-
tity of the strongest mosquito homolog for each putative mutant is listed in the last column.
aThe insertion site for these lines were confirmed by inverse PCR.
Figure 2.—Gene knockdown in An. gambiae. (A) Validation of gene silencing. One-day-old female mosquitoes were injected
with double-stranded LacZ (dsLacZ) or dsRNA of the target gene and whole-body mRNA was extracted 5 days later. Gene expres-
sion was determined by qRT–PCR. Each sample was normalized using ribosomal protein S7 expression as an internal control.
Samples were analyzed in triplicate. Columns indicate averages and standard deviations. The silencing efficiency is expressed
as a percentage reduction in mRNA levels relative to the dsLacZ control. (B) Effect of gene silencing on P. berghei infection.
One-day-old female mosquitoes were injected with either dsLacZ or dsRNA of the target gene. Four days later, mosquitoes fed
on anesthetized P. berghei-infected mice. Midguts were dissected and genomic DNA extracted 6 days post-infection. The intensity of
P. berghei infection was established on the basis of the abundance of the parasite 28S RNA gene relative to the mosquito ribosomal
protein S7 gene in genomic DNA extracted from individual infected midguts determined by qRT–PCR. Medians were compared
using the Kolmogorov–Smirnov test and the P-values ,0.05 were considered significantly different.
Insect Host Genes Affecting Plasmodium Infection 1677
Our success rate of finding genes that have RNAi
phenotypes is in line with that found using data mining
of the genes that we tested that had strong homologs
between the fly and mosquito had the same loss-of-
function phenotype in both the fruit fly and the
mosquito. The tetraspanin gene silenced in mosquitoes
is not the predicted ortholog of Drosophila. This could
explain the increase in infectivity observed when the
mosquito tetraspanin gene was silenced. Even when
there was not a clear ortholog in An. gambiae, the
Drosophila screen was useful in pointing out a gene
family in mosquitoes that has an interesting phenotype
and warrants more detailed analysis.
Our forward screen assessing Plasmodium growth
gives adifferent range of phenotypes to different classes
of microbes than has been seen previously for immunity
mutants in Drosophila. None of the genes that we
identified have been previously implicated in immunity
in either the fruit fly or the mosquito. This work shows
identified in Drosophila that affect Plasmodium growth
can also affect Plasmodium growth in the mosquito.
and the National Science Foundation (S.M.B.).
Alavi, Y., M. Arai, J. Mendoza, M. Tufet-Bayona, R. Sinha et al.,
2003 The dynamics of interactions between Plasmodium and
the mosquito: a study of the infectivity of Plasmodium berghei
and Plasmodium gallinaceum, and their transmission by Anoph-
eles stephensi, Anopheles gambiae and Aedes aegypti. Int. J. Par-
asitol. 33: 933–943.
Ayres, J. S., N. Freitag and D. S. Schneider, 2008
Drosophila mutants altering defense of and endurance to Listeria
monocytogenes infection. Genetics 178: 1807–1815.
Billker, O., M. K. Shaw, G. Margos and R. E. Sinden, 1997
roles of temperature, pH and mosquito factors as triggers of male
and female gametogenesis of Plasmodium berghei in vitro. Par-
asitology 115(Pt. 1): 1–7.
Brandt, S. M., M. S. Dionne, R. S. Khush, L. N. Pham, T. J. Vigdal
et al., 2004 Secreted bacterial effectors and host-produced
eiger/TNF drive death in a Salmonella-infected fruit fly. PLoS
Biol. 2: e418.
Dong, Y., R. Aguilar, Z. Xi, E. Warr, E. Mongin et al., 2006
eles gambiae immune responses to human and rodent Plasmo-
dium parasite species. PLoS Pathog. 2: e52.
Ferrandon, D., J. L. Imler, C. Hetru and J. A. Hoffmann,
2007 The Drosophila systemic immune response: sensing and
signalling during bacterial and fungal infections. Nat. Rev. Im-
munol. 7: 862–874.
Hoiseth, S. K., and B. A. Stocker, 1981
monella typhimurium are non-virulent and effective as live vac-
cines. Nature 291: 238–239.
Khush, R. S., W. D. Cornwell, J. N. Uram and B. Lemaitre, 2002
ubiquitin-proteasome pathway represses the Drosophila immune
deficiency signaling cascade. Curr. Biol. 12: 1728–1737.
Kumar, S., G. K. Christophides, R. Cantera, B. Charles, Y. S. Han
et al., 2003The role of reactive oxygen species on Plasmodium
melanotic encapsulation in Anopheles gambiae. Proc. Natl.
Acad. Sci. USA 100: 14139–14144.
Levashina, E. A., L. F. Moita, S. Blandin, G. Vriend, M. Lagueux
et al., 2001Conserved role of a complement-like protein in
phagocytosis revealed by dsRNA knockout in cultured cells of
the mosquito, Anopheles gambiae. Cell 104: 709–718.
Mansfield, B. E., M. S. Dionne, D. S. Schneider and N. E. Freitag,
2003Exploration of host-pathogen interactions using Listeria
monocytogenes and Drosophila melanogaster. Cell. Microbiol.
Meister, S., S. M. Kanzok, X. L. Zheng, C. Luna, T. R. Li et al.,
2005 Immune signaling pathways regulating bacterial and ma-
laria parasite infection of the mosquito Anopheles gambiae.
Proc. Natl. Acad. Sci. USA 102: 11420–11425.
Portnoy, D. A., P. S. Jacks and D. J. Hinrichs, 1988
lysin for the intracellular growth of Listeria monocytogenes.
J. Exp. Med. 167: 1459–1471.
Schneider, D., and M. Shahabuddin, 2000
opment in a Drosophila model. Science 288: 2376–2379.
Shirasu-Hiza, M. M., and D. S. Schneider, 2007
iology: How do infected flies die? Cell. Microbiol. 9: 2775–2783.
Sullivan, W., M. Ashburner and R. S. Hawley (Editors),
Drosophila Protocols. Cold Spring Harbor Laboratory Press,
Cold Spring Harbor, NY.
Truett, G. E., P. Heeger, R. L. Mynatt, A. A. Truett, J. A. Walker
et al., 2000Preparation of PCR-quality mouse genomic DNA
with hot sodium hydroxide and tris (HotSHOT). Biotechniques
Wu, L. P., K. M. Choe, Y. Lu and K. V. Anderson, 2001
immunity: genes on the third chromosome required for the re-
sponse to bacterial infection. Genetics 159: 189–199.
Zdobnov, E.M., C. von Mering, I. Letunic, D.Torrents, M. Suyama
et al., 2002Comparative genome and proteome analysis of
Anopheles gambiae and Drosophila melanogaster. Science
Role of hemo-
Malaria parasite devel-
Communicating editor: T. Schupbach
1678S. M. Brandt et al.