Waiting with Bated Breath: Opportunistic Orientation to Human Odor in the Malaria Mosquito, Anopheles gambiae, is Modulated by Minute Changes in Carbon Dioxide Concentration

Abstract

Females of the malaria mosquito, Anopheles gambiae, predominantly obtain blood meals within human dwellings. Being highly anthropophilic, human skin odor offers a reliable, host-specific cue, but the challenge posed by pervasive human odor found indoors from used clothing, bedding etc. remains unclear. Anopheles gambiae spends much of its adult life indoors, constantly exposed to human odor even when dwellings are unoccupied. In landing assays, we found that female mosquitoes respond very weakly to human skin odor alone, suggesting that, alone, it is an ineffective landing cue. Landing, however, was dramatically increased by addition of carbon dioxide at a range of concentrations above ambient. Indeed, this effect was seen even when carbon dioxide was just 0.015 % above ambient within the assay cage. The synergistic effect of added carbon dioxide quickly waned, thereby facilitating a highly adaptive “sit-and-wait” ambush strategy, wherein females ignore persistent human odor until a living human is present. Unexpectedly, landing rates in the presence of added carbon dioxide were almost as robust during daytime, when An. gambiae has previously been assumed inactive, possibly facilitating opportunistic feeding at times of day when human dwellings are occupied intermittently. We suggest earlier studies that showed strong upwind flight behavior toward human odor alone could, in fact, have been demonstrating orientation toward a human dwelling rather than toward a living human. This new interpretation of how human odors mediate upwind orientation and landing in An. gambiae is discussed.

Full text

Waiting with Bated Breath: Opportunistic Orientation to Human
Odor in the Malaria Mosquito, Anopheles gambiae, is Modulated
by Minute Changes in Carbon Dioxide Concentration
Ben Webster &Emerson S. Lacey &Ring T. Cardé
Received: 10 September 2014 /Revised: 24 November 2014 /Accepted: 5 December 2014 /Published online: 9 January 2015
#Springer Science+Business Media New York 2015
Abstract Females of the malaria mosquito, Anopheles
gambiae, predominantly obtain blood meals within human
dwellings. Being highly anthropophilic, human skin odor
offers a reliable, host-specific cue, but the challenge posed
by pervasive human odor found indoors from used clothing,
bedding etc. remains unclear. Anopheles gambiae spends
much of its adult life indoors, constantly exposed to human
odor even when dwellings are unoccupied. In landing assays,
we found that female mosquitoes respond very weakly to
human skin odor alone, suggesting that, alone, it is an inef-
fective landing cue. Landing, however, was dramatically in-
creased by addition of carbon dioxide at a range of concen-
trations above ambient. Indeed, this effect was seen even
when carbon dioxide was just 0.015 % above ambient within
the assay cage. The synergistic effect of added carbon dioxide
quickly waned, thereby facilitating a highly adaptive “sit-and-
wait”ambush strategy, wherein females ignore persistent hu-
man odor until a living human is present. Unexpectedly,
landing rates in the presence of added carbon dioxide were
almost as robust during daytime, when An. gambiae has
previously been assumed inactive, possibly facilitating oppor-
tunistic feeding at times of day when human dwellings are
occupied intermittently. We suggest earlier studies that
showed strong upwind flight behavior toward human odor
alone could, in fact, have been demonstrating orientation
toward a human dwelling rather than toward a living human.
This new interpretation of how human odors mediate upwind
orientation and landing in An. gambiae is discussed.
Keywords Anopheles gambiae .Olfaction .Landing .
Behavior .Carbon dioxide .Skin odor
Introduction
The highly anthropophilic malaria mosquito, Anopheles
gambiae, has evolved to search for blood meals in human
dwellings (Cardé and Gibson 2010; Takken and Verhulst
2013). Although female An. gambiae spend long periods of
time outdoors, taking shelter in foliage or engaging in nectar
feeding (Manda et al. 2007;Mülleretal.2010), much of their
adult life is also spent within human dwellings. Mosquitoes
enter houses throughout the night, peaking around midnight,
and continuing at a high rate until the early hours of the
morning (Haddow 1942;Thomson1948). Just before dawn
there is a spike inentry rates due to mosquitoes seeking shelter
in advance of daytime temperatures (Haddow 1942;
Paaijmans and Thomas 2011), with most females entering at
this time remaining until the following evening (Paaijmans
and Thomas 2011). Biting rates correlate closely with house
entering, suggesting most mosquitoes feed shortly after enter-
ing (Haddow 1942;Ribbands1946), but mosquitoes may first
undergo a short resting period if entry occurs early in the
evening (Ribbands 1946). Unfed An. gambiae may, therefore,
spend a number of hours within a human dwelling before
actually feeding. Following a blood meal, mosquitoes often
remain in dwellings until ready to oviposit and often return for
another blood meal the same night following oviposition
(Thomson 1948).
Anopheles gambiae responds behaviorally to a range of
human-associated cues (Cardé and Gibson 2010; Takken and
Ve r h u l st 2013). Carbon dioxide offers a general, long-range
cue and also functions as an activator that initiates take off and
sustained flight (e.g., Healy and Copland 1995; Lorenz et al.
2013). Carbon dioxide also elicits upwind plume following
(Gillies 1980), and synergizes responses to other host-
associated cues at close range (e.g., Dekker et al. 2005;Krober
et al. 2010). Other studies, in contrast, have suggested that
CO
2
plays a limited role in An. gambiae host-seeking
B. Webster :E. S. Lacey :R. T. Cardé (*)
Department of Entomology, University of California, Riverside,
CA 92521, USA
e-mail: ring.carde@ucr.edu
J Chem Ecol (2015) 41:59–66
DOI 10.1007/s10886-014-0542-x

behaviors (Takken and Knols 1999; Spitzen et al. 2008), with
no increase in upwind orientation, over that to clean air, with
an added plume of CO
2
in wind-tunnel assays (Takken et al.
1997). Heat also is an important, general cue, inducing land-
ing in An. gambiae when offered together with CO
2
(Krober
et al. 2010), and resulting in changes in flight behavior when
presented alongside human body odor (Spitzen et al. 2013).
Human skin odor offers a host-specific cue to the highly
anthropophilic An. gambiae, and females can readily discrim-
inate between human skin odor and that of other animals
(Pates et al. 2001). Several studies have demonstrated upwind
flight responses to human skin odor alone (Pates et al. 2001;
Qiu et al. 2006; Spitzen et al. 2013) and these responses are
enhanced by addition of CO
2
(Lorenz et al. 2013; Njiru et al.
2006; Spitzen et al. 2008).
The presence of human odor in a dwelling is persistent,
with odor being emitted from various sources, such as
bedding and discarded clothing, even when human
occupants are absent. For example, Njiru et al. (2006)showed
that a sock worn for 12 h was capable of increasing catches in
a MM-X counter-flow suction trap for up to 8 days after
wearing, demonstrating the persistence of human odor. This
presents a problem to host-seeking An. gambiae that either
may spend several hours within a human dwellingtemporarily
devoid of live human hosts on which to feed or that have a
biting rhythm later in the night. For example, houses in
regions native to An. gambiae are sporadically occupied by
human owners during the early hours of the evening, when
mosquitoes begin to congregate indoors (Gillies 1954). Fur-
thermore, mosquitoes may spend long periods of time within
unoccupied dwellings during the day (Haddow 1942;
Paaijmans and Thomas 2011). Even when houses are occu-
pied with sleeping humans during the night, mosquitoes may
find themselves in parts of the house other than sleeping areas,
yet are still exposed to sources of human odor. Responding
strongly to human skin odor alone once inside a dwelling
would, therefore, be a highly inefficient means of locating a
feeding site. We hypothesized that landing, the final stage of
host location, is mediated in part by human skin odor, but that
this only occurs in the presence of CO
2
, which does not persist
when humans are absent, thus allowing mosquitoes to avoid
responding inefficiently to human odor when dwellings are
unoccupied. We hypothesized that such behavior also could
facilitate opportunistic feeding during times of day not nor-
mally associated with biting activity, such as during the day-
time when humans are only infrequently indoors, where mos-
quitoes are present.
Methods and Materials
Insects Our An. gambiae M Form originated from 2007 col-
lections in Ngousso, Cameroon (see Turissini et al. 2014 for a
history of this colony). It was maintained in a L:D 12:12 h
photocycle at approximately 27 °C and 70 % RH. The colony
was propagated with bovine blood (Hemostat Laboratories,
California, USA) through a heated membrane feeding system.
Carbon dioxide was provided by exhaled breath to facilitate
feeding. Larvae were reared in plastic containers with deionized
water and fish food (Tetramin Tropical Flakes, Tetra, VA, USA).
Pupae were transferred to screen cages (30 × 30 × 30 cm,
BugDorm-1, MegaView Science Co., Ltd., Talchung, Taiwan)
before eclosion and provided with 10 % sucrose solution in
deionized water ad libitum. Males and females were kept to-
gether and, therefore, females were assumed to have mated.
Females were used in experiments 4–14 d post-eclosion and
were not blood fed. Approximately 20 h prior to experiments, at
the start of photophase, females were deprived access to sucrose
and transferred to cylindrical acrylic containers (7 × 8 cm i.d.);
5–12 females were placed in each cage and stored in the colony
room for approximately 16 h before being transferred to the
assay room (27 °C and 50–70 % RH, illuminated to approxi-
mately 3.5 lux) to acclimatize for a further 2–4 h prior to testing.
All landing assays took place in the last 4 h of the scotophase.
Mosquitoes tested for daytime landing were collected during the
last 1–2 h of the photophase, approximately 20 h prior to
experiments, before being transferred to the assay room 2–4h
prior to experiments. The assay room was maintained at 27 °C
and 50–70 % RH, and illuminated by an overhead florescent
light to give an intensity of 140 lux at the point of the assay.
Odor Treatment Skin odor was collected using pieces of
white polyester gauze (mesh size 9.5 × 7.9 cm
−1
, Bioquip
Products, California, USA) worn in a cotton sock by the
experimenter (B.W.) for 4–6 h prior to experiments. Gauze
and sock were previously cleaned using non-perfumed deter-
gent and the feet washed with non-perfumed soap the evening
before experiments. Immediately prior to an experiment, the
gauze was removed from the sock and placed between two 75
× 85 mm rectangles of magnetic vinyl (McMaster-Carr, Cal-
ifornia, USA), with a 60 × 50 mm hole cut through the middle
to provide a landing area over the gauze.
Landing Assays We released 16–20 mosquitoes into screen
cages (30 × 30 × 30 cm), fitted with glass panels on top to
allow viewing, 2–4 h before experiments. At the start of each
experiment, a single screen cage was moved carefully to an
assay table in the same room and left undisturbed for 5 min.
After 5 min, a glass tube (85 × 6 mm i.d.) connected via rubber
tubing to an air/CO
2
supply was positioned 10 cm from the
screen cage, and a foot odor-treated gauze was gently placed
in the center of the cage. The experimenter left the room, and
landing behavior was recorded for 6 min using a video camera
equipped with night vision. The concentration of air/CO
2
was
regulated by mixing medical grade air and pure CO
2
using
flow meters. The air/CO
2
supply exited the glass cylinder at a
60 J Chem Ecol (2015) 41:59–66

rate of 1 l min
−1
andapprox.1ms
−1
, which was hard to detect
(airflow <0.1 m s
−1
) within the screen cage. The effects of four
different concentrations of CO
2
on landing behavior were
tested: 4, 1, 0.1, and ~0.03 % CO
2
(mixed with medical grade
air). A CO
2
gas analyzer (Gashound Model LI-800, LI-Core,
Nebraska, USA) was used to determine that the first three
concentrations corresponded to time-averaged concentrations
of approx. 0.35, 0.05, and 0.015 % above background at the
center of the screen cage and the last (0.03 %) corresponded to
background levels. Twelve replicates of 16–20 mosquitoes
were carried out for each treatment (approximately 220 mos-
quitoes per treatment).
In a separate experiment, to ensure landings were responses
to skin odor and not responses to visual cues, landing rates
were compared on clean gauze and foot odor-treated gauze in
the presence of either air or 4 % CO
2
, the same way as
described above (5–6replicatesof15–20 mosquitoes each,
approximately 120 mosquitoes per treatment).
Preliminary assessment indicated that the presence of an
individual breathing normally, positioned approximately
0.5 m from the assay table in a room of 28 m
3
,wassufficient
to raise carbon dioxide levels ~0.02 % above background
within 1–2 min. Therefore, to avoid exhaled CO
2
affecting
mosquito behavior, the experimenter retained his breath dur-
ing the ~30 s periods spent in the assay room at the start of
each experiment. An extractor fan with an inlet on the oppo-
site side of the room to the assay table ensured CO
2
used in
experiments did not lead to changes in background CO
2
over
thecourseofanexperiment.
Daytime Landing Assays To determine if landing rate is simi-
larly affected by the presence of CO
2
during the daytime, mos-
quito landing in response to air or 4 % CO
2
was tested during
hours 7–10 of the 12 h photophase, corresponding to mid-
afternoon. Protocols were the same as before, save for illumina-
tion of 140 lux in the assay room provided by overhead flores-
cent lighting. Eight replicates of 12–20 mosquitoes were carried
out for each treatment (approx. 120 mosquitoes per treatment).
Pre-Exposure Assays Mosquitoes were stored in cylindrical
containers covered at each end with mesh (9–12 mosquitoes
per cage) until the start of an experiment, and then were placed
10 cm in front of a glass tube (85 × 6 mm i.d.) delivering a
stream of air at a rate of 1 l min
−1
.After9min,thecontainer
was subject to exposure to a further 1 min of either air or 4 %
CO
2
, after which it was transferred to a screen cage containing a
foot odor-gauze landing rectangle (as above). Landing over a
period of 6 min was recorded when either air or 4 % CO
2
was
delivered through the screen cage, as described. Thus, four
treatments were tested: 1) pre-exposure to air followed by
landing while exposed to air; 2) pre-exposure to 4 % CO
2
followed by landing while exposed to air; 3) pre-exposure to
air followed by landing while exposed to 4 % CO
2
;4)pre-
exposureto4%CO
2
followedbylandingwhileexposedto4%
CO
2
. Twenty replicates of 9–12 mosquitoes were carried out for
each treatment (approximately 220 mosquitoes per treatment).
Data Analysis After experiments, video files were analyzed
(blind) by counting numbers of mosquitoes resting on gauze
every 30 s over the 6 min test period. Differences in numbers of
mosquitoes resting on gauze at each time period were compared
using a Kruskal-Wallis test with Bonferroni correction, follow-
ed by post-hoc pairwise Mann–Whitney-Utests to determine
differences among individual treatments. All tests were carried
out using R v. 2.15.0 (R Development Core Team 2012).
Results
Effect of Carbon Dioxide on Landing Behavior Few mosqui-
toes (<10 %) landed on foot odor-treated gauze when only
clean air was delivered into the assay cage (Fig. 1). When 4 %
CO
2
was introduced, a steady increase in landing rates was
observed over the first 2 min, before leveling off at approx.
40–50 %. Differences between clean air and 4 % CO
2
were
statistically significant (P<0.05) at all time periods from
1.5 min onwards. A similar pattern was observed when 1
and 0.1 % CO
2
were presented, resulting in 2–3 fold increases
in landing rates compared to clean air (P<0.05), but signifi-
cantly less than when 4 % CO
2
was used (P<0.05). No
differences (P>0.05) in effects on landing rates wereobserved
between 1 and 0.1 % CO
2
at any time point.
In a separate test, almost no landing occurred on clean gauze,
regardless of whether air (mean landings on clean gauze over
6min:0.09%;meanlandingsonfootodor-treatedgauze:
13.7 %; P=0.003)or 4 % CO
2
(mean landings on clean gauze,
0.26 %; mean landings on foot odor-treated gauze, 37.7 %,
P=0.004) were tested.
Effect of Pre-exposure with Carbon Dioxide on Landing
Behavior When CO
2
was delivered through the assay screen
cage, no effect of pre-exposure to either air or CO
2
was
observed on landing rates (Fig. 2). When clean air was deliv-
ered through the assay cage, pre-exposure to CO
2
led to a brief
increase in landing rates compared to pre-exposure with clean
air but these differences were not significant (P>0.05).
Daytime Landing Daytime landing rates on skin odor-treated
gauze in the presence and absence of 4 % CO
2
were similar to
those observed during nighttime assays (Fig. 3). Landing on
skin odor in the presence of air was consistently low (<10 %),
whereas landing in the presence of 4 % CO
2
was considerably
higher at 25–30 %, with differences (P<0.05) from 2.5 min
onwards.
J Chem Ecol (2015) 41:59–66 61

Discussion
In the absence of CO
2
, rates of landing on skin odor-treated
gauze were very low, suggesting skin odor alone is an inef-
fective landing cue. Addition of even small concentrations of
CO
2
, however, increased landing substantially. As little as
0.1 % CO
2
directed at the assay cage greatly increased land-
ing, despite corresponding to a time-averaged increase of only
0.015 % above background within the assay cage, similar to
the 0.01 % threshold reported to initiate upwind flight in An.
gambiae (Healy and Copland 1995). In the present study, CO
2
was delivered toward the side of the assay cage, whereas skin
odor was placed on the bottom. The increased response was,
therefore, not due to attraction toward CO
2
but rather to CO
2
triggering landing on a source of skin odor, perhaps by acti-
vating searching flight behavior.
Our results demonstrate how an effective sit-and-wait strategy
for host-seeking An. gambiae in temporarily unoccupied human
dwellings may operate. When a room is unoccupied, mosquitoes
conserve energy by ignoring omnipresent human odor. Small
increases in CO
2
indicate the probable presence of a human, and
this triggers landing in response to skin odor. In the pre-exposure
experiment, the effect of CO
2
on response to skin odor did not
persist, suggesting mosquitoes quickly return to a resting state
when the room becomes empty again, thus conserving energy
until a human host re-enters. These results mirror those of Krober
et al. (2010), who observed that landing on a heated pad by An.
gambiae could be significantly increased by delivery of a short
puff of pure CO
2
. While heat does not offer a host-specific cue in
the same way human skin odor does, a landing response to both
body odor and heat, activated by CO
2
, may facilitate an efficient
biting strategy in house-dwelling mosquitoes. An additional role
for heat is especially likely, given that upwind flight and landing
on a target was greatly enhanced when the target was heated in
the presence of human skin odor in a wind-tunnel assay (Spitzen
et al. 2013). A recent study suggested that the CO
2
-sensing cpA
neurons in the maxillary palps of An. gambiae may also detect
odorants emitted from human skin (Tauxe et al. 2013), suggest-
ing they may possibly play a role in simultaneous detection of
both these cues. Additional classes of odorant receptors also
detect skin odor (McMeniman et al. 2014), and so instantaneous
coordination of detection with the CO
2
pathway may act to
generate landing behavior. Our findings may open opportunities
for designing new types of mosquito control, for example by
reducing landing through inhibition of the CO
2
receptor (Tauxe
et al. 2013;Turneretal.2011).
The effect of CO
2
on landing behavior in response to skin
odor represents a striking adaptation to feeding on humans
60
50
E)
aaa
a
50
/- SE
aaaa
a
aa
40
ng (+
abbbbb
aab
bab
30
ndin
bc bbbbbbb
b
ab bb
b
b
a
20
% la
bbbb
c
a
a
20
ean %
b
b
cccccc
cbc
a
a
a
10
me
c
a
aaaa
00
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
time (min)
air 0.1% CO21% CO24% CO2
Fig. 1 Mean % (±SE) female
Anopheles gambiae landing on a
foot odor-treated rectangle of
gauzewhenexposedtocleanair,
0.1, 1, and 4 % carbon dioxide
over a 6 min period. Different
letters at each time point indicate
differences (Bonferroni adjusted
P<0.05)
70
60
E)
aaaaaaaaaaa
50
/- SE
aaaaaa
aaa
aa
aaaa
40
50
g (+/
ab
ab ab ab
a
a
40
nding
ab
b
aa
30
% lan
bbbbb
b
bb
b
a
20
an %
bbbbb
bb
b
b
a
a
10
mea
bb
a
a
0
0
0
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
time (min)
air, air air, CO2CO2, air CO2, CO2
Fig. 2 Mean % (±SE) female Anopheles gambiae landing on a foot odor-
treated rectangle of gauze in the presence of air or 4 % carbon dioxide
after pre-exposure to either air or 4 % carbon dioxide. Treatments: ‘air,
air’=1 min pre-exposure to air followed by landing while exposed to air;
‘CO
2
,air’=1 min pre-exposure to 4 % CO
2
, followed by landing while
exposed to air; ‘air, CO
2
’=1 min pre-exposure to air followed by landing
while exposed to 4 % CO
2
;‘CO
2
,CO
2
’; 1 min pre-exposure to 4 % CO
2
,
followed by landing while exposed to 4 % CO
2
. Different letters at each
time point indicate differences (Bonferroni adjusted P<0.05)
62 J Chem Ecol (2015) 41:59–66

indoors. Although An. gambiae is widely considered an
endophagic (indoor-feeding) species, strong exophagy (out-
door feeding) has been suggested in several populations
(Reddy et al. 2011), although differences in indoor vs. outdoor
sampling methods make accurate assessment difficult.
Exophagy may correspond to villages where most people
sleep outdoors (Faye et al. 1997;Tunoetal.2010)orareas
where indoor anti-mosquito intervention may have led to a
switch in behavior (Molina et al. 1996; Reddy et al. 2011). It
would be useful to see if landing in response to skin odor is
similarly affected by CO
2
in exophagic populations, since
these mosquitoes may not be faced with the same problem
of constant exposure to skin odor. However, in the Senegal
River Basin, where people sleep exclusively outdoors, An.
gambiae were found to rest within houses during the day
despite moving outside to feed at night (Faye et al. 1997), so
this challenge also may be faced by exophagic populations.
Other mosquito species readily land on sources of human
skin odor in the absence of CO
2
.Culex quinquefasciatus lands
readily on warmed glass beads coated with human foot odor,
with landing rates not affected by the addition of CO
2
,al-
though activation and plume following were (Lacey and
Cardé 2011). Culex quinquefasciatus is less anthropophlic
than An. gambiae, possibly explaining less specialized house
hunting-adapted behavior, although some populations of
C. quinquefasciatus have been shown to exhibit house-
entering behavior (Muturi et al. 2008;Njieetal.2009). Aedes
aegypti also lands readily on a source of human foot odor, and
this was not significantly affected by the addition of CO
2
(Lacey et al. 2014). Other studies, however, have shown that
elevated CO
2
increases Ae. aegypti flight responses toward
human skin odor (Dekker et al. 2005; McMeniman et al.
2014) and that human foot odor alone is insufficient to induce
feeding through a membrane, but this can be enhanced by
addition of CO
2
(McMeniman et al. 2014). Although gener-
ally considered less specialized on humans than An. gambiae
(Takken and Verhulst 2013), some populations of Ae. aegypti
display endophily and so may be exposed to similar chal-
lenges when searching for a human blood meal indoors
(McMeniman et al. 2014; Trpis and Hausermann 1978).
Different mosquito species show large differences in host
ranges and feeding habits and care must be taken when
making generalizations across species (Klowden 2007). Nev-
ertheless, our findings may have broader implications for
other highly endophilic hematophagous insect species or
populations.
Surprisingly, we found mosquitoes readily landed on a
source of human skin odor during the daytime, provided that
CO
2
was delivered simultaneously. Anopheles gambiae is
widely assumed to feed predominantly on humans at night
(Haddow 1942;JonesandGubbins1978;Ribbands1946)but
this may simply be due to the fact that humans are more likely
to be present within a dwelling at night. Few studies have
explored house entering and biting dynamics throughout the
daytime. A recent study by Rund et al. (2013), however,
showed peak feeding behavior on a human volunteer at night
and this was accompanied by changes in gene expression and
olfactory receptor neuron sensitivity. Differences in method-
ology could be responsible for our dissimilar results, for
example the presence of heat as a cue or exposure to CO
2
.
Alternatively, strains may differ in their diurnal feeding habits
(Moiroux et al. 2012). Some laboratory-reared strains may in
fact have been artificially selected for feeding at convenient
times of day. A similar adaptation has been seen in natural
populations, with mosquitoes changing their diurnal feeding
behavior in response to use of insecticide-treated bed nets
(e.g., Moiroux et al. 2012). It is also possible that, since our
lab strain was reared on bovine blood, responses to human-
associated cues may have changed over time. However, be-
cause we observed fairly strong landing responses on a source
of human skin odor (~50 % mosquitoes landing simultaneous-
ly) in the presence of CO
2
,thissuggeststhatpreferencefor
human odor has been retained. It would be useful to repeat
daytime landing assays using a number of An. gambiae pop-
ulations recently colonized from the field to determine how
widespread daytime responsiveness to human odors is under
natural selective pressure. Since female An. gambiae often
shelter in human dwellings during the day (Haddow 1942;
Paaijmans and Thomas 2011), the behavior we observed
would presumably offer a selective advantage by facilitating
4040
*******
35
*
*
30
SE)
25
(+/- S
20
ing (
15
land
air
10
n %
4% CO2
5
mean
0
5
m
0
0.511.522.533.544.555.56
time
(
min
)
Fig. 3 Mean % (±SE) female
Anopheles gambiae landing on a
foot odor-treated rectangle of
gauze when exposed to air or 4 %
carbon dioxide over a 6 min peri-
od during hours 7–10 of the 12 h
light phase. * Indicates differ-
ences (Bonferroni adjusted
P<0.05)
J Chem Ecol (2015) 41:59–66 63

opportunistic feeding at times when humans are present in-
doors intermittently or, for example, during inclement weath-
er. It is important to consider, however, that there is a lack of
recorded evidence of daytime biting in field conditions and
our results must be interpreted with caution until proper stud-
ies with a range of field-collected populations can be
conducted.
In contrast to our findings, several earlier studies have
demonstrated strong behavioral responses toward skin odor
in the absence of CO
2
. For example, Pates et al. (2001)
observed a strong preference for human skin odor over clean
air in a dual port-entry, still-air olfactometer, and similar
responses were observed in subsequent studies using a similar
protocol (Qiu et al. 2006). Field studies also have shown that
body odor alone is attractive to An. gambiae when used in
MM-X traps, although catches are greatly increased by addi-
tion of CO
2
(Njiru et al. 2006; Qiu et al. 2007). Similar
findings were obtained by Smallegange et al. (2010), who
showed in semi-field trials that large numbers of released
mosquitoes were trapped when a worn sock was used as bait,
compared to an unbaited control, although these catches also
were significantly increased by addition of CO
2
. In a more
recent study, An. gambiae was found to display strong plume
following behavior in a wind tunnel in response to body odor
alone, although few mosquitoes landed on the odor source
unless it was heated (Spitzen et al. 2013). Very low levels of
landing by healthy female An. gambiae on a source of human
skin odor, in the absence of other host cues, also were reported
by Smallegange et al. (2013), similar to that seen in the present
study. These various studies differ in their methodology, as
well as in the use of different human volunteers as odor
sources, so direct comparisons are problematic. In general,
however, it seems that skin odor alone is effective at inducing
activation and plume following behavior but remains a
relatively weak and ineffective landing cue in the absence
of CO
2
.
Snow (1987) suggested that house entry is a behavior-
ally distinct stage of host location, involving a transition
to upward flight toward the eaves of houses, and this
behavior partly distinguishes endophagic from exophagic
species. The same study suggested that location of a
human dwelling for daytime resting is a distinct (but not
necessarily exclusive) behavioral trait from host location.
Thus, there is a distinction between “house-location,”
where mosquitoes locate a human dwelling, and host-
location, where mosquitoes locate and land on a human
within a dwelling. When dwellings are occupied, CO
2
and
body odor are emitted together, resulting in strong upwind
flight, house entry, and eventual landing. Detection of
human odor in the absence of CO
2
does not indicate the
immediate location of a living host but does indicate the
location of a structure normally occupied by humans.
Flying upwind toward a source of human odor, even in
the absence of elevated CO
2
would, therefore, offer an
effective strategy for locating an eventual human blood
meal. This also would allow location of ideal daytime
resting places by permitting females to take shelter in a
place where humans are likely to be present the following
evening (Haddow 1942).Landingonasourceofskin
odor in An. gambiae tends to happen once it has already
entered a dwelling and the omnipresent body odor within
becomes a poor cue for locating a feeding site. It should
not, therefore, initiate a landing response unless a fluctu-
ating concentration of CO
2
is also detected. We propose
that our landing assay, and similar assays (e.g.,
Smallegange et al. 2013), demonstrate the final stages of
host location/landing and, therefore, explain the weak
responses to human skin odor alone. Previous studies that
demonstrated strong upwind flight and plume following
behavior toward body odor alone may have demonstrated
a response to the perceived odor of a human habitat rather
than that of a living human host, explaining the relatively
robust responses, even in the absence of CO
2
. This hy-
pothesis warrants further investigation and, if correct, an
understanding of the distinction between host-location
and “house-location”could yield important insights into
the behavioral ecology of this important disease vector.
Our results also have important methodological implica-
tions for future studies of An. gambiae. We found that CO
2
levels in a screen cage within a 28 m
3
assay room increased
~0.02 % within 2 min of a person positioned 0.5 m in front of
it, above the 0.015 % we determinedwas sufficient to induce a
landing response. Thus, the presence of an experimenter in an
assay room could be sufficient to result in a substantial in-
crease in response to skin odor. Future experiments should,
therefore, aim to control for and report measures taken to
avoid the presence of a breathing experimenter inadvertently
influencing results. Our landing assay also may prove useful
in the screening of new contact repellents. Previously, human
arm landing assays have been used (Logan et al. 2010), but
difficulties in controlling exposure to exhaled CO
2
may com-
plicate interpretation of results. Heated pads with CO
2
have
been suggested (Krober et al. 2010), but as unpredictable
behaviors can result from mixing different odors (Bruce and
Pickett 2011; Webster et al. 2010), testing repellents alongside
natural skin volatiles offers a preferable diagnostic of repellent
efficacy. An example of such an assay, assessing landing
behavior on a heated target treated with a synthetic blend
mimicking human skin odor presented together with puffs of
CO
2,
has recently been described (Menger et al. 2014).
Acknowledgments This work was funded by an R56AI099778 (Na-
tional Institute of Allergy and Infectious Diseases) grant to Anandasankar
Ray and Ring Cardé. The granting agencies had no role in experimental
design or analysis. We are grateful to Anandasankar Ray for his com-
ments and Bradley White for our colony of An. gambiae.
64 J Chem Ecol (2015) 41:59–66

References
Bruce TJA, Pickett JA (2011) Perception of plant volatile blends by
herbivorous insects - finding the right mix. Phytochemistry 72:
1605–1611. doi:10.1016/j.phytochem.2011.04.011
Cardé RT, Gibson G (2010) Host finding by female mosquitoes: mech-
anisms of orientation to host odours and other cues. In: Takken W,
Knols BGJ (eds) Olfaction in vector-host interactions. Wageningen
Academic Publishers, Wageningen, pp 115–142. doi:10.3920/978-
90-8686-698-4
Dekker T, Geier M, Cardé RT (2005) Carbon dioxide instantly sensitizes
female yellow fever mosquitoes to human skin odours. J Exp Biol
208:2963–2972. doi:10.1242/jeb.01736
Development Core Team (2012) R: a language and environment for
statistical computing, 2.15.0 edn. R Foundation for Statistical
Computing, Vienna
Faye O, Konate L, Mouchet J, Fontenille D, Sy N, Hebrard G, Herve JP
(1997) Indoor resting by outdoor biting females of Anopheles
gambiae complex (Diptera: Culicidae) in the sahel of northern
Senegal. J Med Entomol 34:285–289
Gillies MT (1954) Studies of house leaving and outside resting of
Anopheles gambiae Giles and Anopheles funestus Giles in East
Africa. I.—The outside resting population. Bull Entomol Res 45:
361–373. doi:10.1017/S0007485300027188
Gillies MT (1980) The role of carbon dioxide in host-finding by mosqui-
tos (Diptera: Culicidae): a review. Bull Entomol Res 70:525–32. doi:
10.1017/S0007485300007811
Haddow AJ (1942) The mosquito fauna and climate of native huts at
Kisumu, Kenya. Bull Entomol Res 33:91–142. doi:10.1017/
S0007485300026389
Healy TP, Copland MJW (1995) Activation of Anopheles gambiae mos-
quitoes by carbon dioxide and human breath. Med Vet Entomol 9:
331–336. doi:10.1111/j.1365-2915.1995.tb00143.x
Jones MDR, Gubbins SJ (1978) Changes in circadian flight activity of
mosquito Anopheles gambiae in relation to insemination, feeding
and oviposition. Physiol Entomol 3:213–220. doi:10.1111/j.1365-
3032.1978.tb00151.x
Klowden MJ (2007) Making generalizations about vectors: is there a
physiology of “the mosquito”? Entomol Res 37:1–13. doi:10.1111/j.
1748-5967.2007.00044.x
Krober T, Kessler S, Frei J, Bourquin M, Guerin PM (2010) An in vitro
assay for testing mosquito repellents employing a warm body and
carbon dioxide as a behavioral activator. J Am Mosq Control Assoc
26:381–386. doi:10.2987/10-6044.1
Lacey ES, CardéRT (2011) Activation, orientation and landing of female
Culex quinquefasciatus in response to carbon dioxide and odour
from human feet: 3D flight analysis in a wind tunnel. Med Vet
Entomol 25:94–103. doi:10.1111/j.1365-2915.2010.00921.x
Lacey ES, Ray A, Cardé RT (2014) Close encounters: contributions of
carbon dioxide and human skin odour to finding and landing on a
host in Aedes aegypti. Physiol Entomol 39:60–68. doi:10.1111/
phen.12048
Logan JG, Stanczyk NM, Hassanali A, Kemei J, Santana AEG, Ribeiro
KAL, Pickett JA, Mordue J (2010) Arm-in-cage testing of natural
human-derived mosquito repellents. Malar J 9:239. doi:10.1186/
1475-2875-9-239
Lorenz LM et al (2013) Taxis assays measure directional movement of
mosquitoes to olfactory cues. Parasite Vector 6:131. doi:10.1186/
1756-3305-6-131
Manda H, Gouagna LC, Foster WA, Jackson RJ, Beier JC, Githure JI,
Hassanali A (2007) Effect of discriminative plant-sugar feeding on
the survival and fecundity of Anopheles gambiae. Malar J 6:113.
doi:10.1186/1475-2875-6-113
McMeniman CJ, Corfas RA, Matthews BJ, Ritchie SA, Vosshall LB
(2014) Multimodal integration of carbon dioxide and other sensory
cues drives mosquito attraction to humans. Cell 156:1060–1071.
doi:10.1016/j.cell.2013.12.044
Menger DJ, Van Loon JJA, Takken W (2014) Assessing the efficacy of
candidate mosquito repellents against the background of an attrac-
tive source that mimics a human host. Med Vet Entomol 28:407–
413. doi:10.1111/mve.12061
Moiroux N, Gomez MB, Pennetier C, Elanga E, Djènontin A, Chandre F,
Djègbé I, Guis H, Corbel V (2012) Changes in Anopheles funestus
biting behavior following universal coverage of long-lasting insec-
ticidal nets in Benin. J Infect Dis 206:1622–1629. doi:10.1093/
infdis/jis565
Molina R, Benito A, Blanca F, Roche J, Otunga B, Alvar J (1996) The
anophelines of Equatorial Guinea. Ethology and susceptibility stud-
ies. Res Rev Parasitol 56:105–110
Müller GC, Beier JC, Traore SF, Toure MB, Traore MM, Bah S, Doumbia
S, Schlein Y (2010) Field experiments ofAnopheles gambiaeattrac-
tion to localfruits/seedpods and flowering plants in Mali to optimize
strategies for malaria vector control in Africa using attractive toxic
sugar bait methods. Malar J 9:262. doi:10.1186/1475-2875-9-262
Muturi E, Muriu S, Shililu J, Mwangangi JM, Jacob BG, Mbogo C,
Githure J, Novak RJ (2008) Blood-feeding patterns of Culex
quinquefasciatus and other culicines and implications for disease
transmission in Mwea rice scheme, Kenya. Parasitol Res 102:1329–
1335. doi:10.1007/s00436-008-0914-7
Njie M, Dilger E, Lindsay SW, Kirby MJ (2009) Importance of eaves to
house entry by Anopheline, but not Culicine, mosquitoes. J Med
Entomol 46:505–510. doi:10.1603/033.046.0314
Njiru BN, Mukabana WR, Takken W, Knols BGJ (2006) Trapping of the
malaria vector Anopheles gambiae with odour-baited MM-X traps
in semi-field conditions in western Kenya. Malar J 5:1–8. doi:10.
1186/1475-2875-5-39
Paaijmans KP, Thomas MB (2011) The influence of mosquito resting
behaviour and associated microclimate for malaria risk. Malar J 10:
183. doi:10.1186/1475-2875-10-183
Pates HV, Takken W, Stuke K, Curtis CF (2001) Differential behaviourof
Anopheles gambiae sensu stricto (Diptera : Culicidae) to human and
cow odours inthe laboratory. Bull Entomol Res 91:289–296. doi:10.
1079/ber200198
Qiu YT, Smallegange RC, Van Loon JJA, Ter Braak CJF, Takken W
(2006) Interindividual variation in the attractiveness of human
odours to the malaria mosquito Anopheles gambiae s. s.MedVet
Entomol 20:280–287. doi:10.1111/j.1365-2915.2006.00627.x
Qiu YT et al (2007) Attractiveness of MM-X traps baited with human or
synthetic odor to mosquitoes (Diptera : Culicidae) in the Gambia. J
Med Entomol 44:970–983. doi:10.1603/0022-2585(2007)
44[970:aomtbw]2.0.co;2
Reddy MR, Overgaard HJ, Abaga S, Reddy VP, Caccone A, Kiszewski
AE, Slotman MA (2011) Outdoor host seeking behaviour of
Anopheles gambiae mosquitoes following initiation of malaria vec-
tor control on Bioko Island, Equatorial Guinea. Malar J 10:184. doi:
10.1186/1475-2875-10-184
Ribbands CR (1946) Moonlight and house-hauting habits of female
anophelines in West Africa. Bull Entomol Res 36:395–417
Rund SSC, Bonar NA, Champion MM, Ghazi JP, Houk CM, Leming
MT, Syed Z, Duffield GE (2013) Daily rhythms in antennal protein
and olfactory sensitivity in the malaria mosquito Anopheles
gambiae. Sci Rep 3:2494. doi:10.1038/srep02494
Smallegange RC, Schmied WH, van Roey KJ, Verhulst NO, Spitzen J,
Mukabana WR, Takken W (2010) Sugar-fermenting yeast as an
organic source of carbon dioxide to attract the malaria mosquito
Anopheles gambiae. Malar J 9:292. doi:10.1186/1475-2875-9-292
Smallegange RC, van Gemert GJ, van de Vegte-Bolmer M, Gezan S,
Takken W, Sauerwein RW, Logan JG (2013) Malaria infected mos-
quitoes xxpress enhanced attractionto human odor. PLoS One 8 doi:
10.1371/journal.pone.0063602
J Chem Ecol (2015) 41:59–66 65

Snow WF (1987) Studies of house-entering habits of mosquitos in the
Gabia, West Africa: experiments with prefabricated huts with varied
wall apertures. Med Vet Entomol 1:9–21. doi:10.1111/j.1365-2915.
1987.tb00318.x
Spitzen J, Smallegange RC, Takken W (2008) Effect of human odours and
positioning of CO
2
release point on trap catches of the malaria
mosquito Anopheles gambiae sensu stricto in an olfactometer.
Physiol Entomol 33:116–122. doi:10.1111/j.1365-3032.2008.00612.x
Spitzen J et al (2013) A 3D analysis of flight behavior of Anopheles
gambiae sensu stricto malaria mosquitoes in response to human odor
and heat. PLoS ONE 8:e62995. doi:10.1371/journal.pone.0062995
Takken W, Knols BGJ (1999) Odor-mediated behavior of Afrotropical
malaria mosquitoes. Annu Rev Entomol 44:131–157. doi:10.1146/
annurev.ento.44.1.131
Takken W, Verhulst NO (2013) Host preferences of blood-feeding mos-
quitoes. Annu Rev Entomol 58:433–453. doi:10.1146/annurev-
ento-120811-153618
Takken W, Dekker T, Wijnholds YG (1997) Odor-mediated flight behav-
ior of Anopheles gambiae Giles Sensu Stricto and A.stephensi Liston
in response to CO
2
, acetone, and 1-octen-3-ol (Diptera: Culicidae). J
Insect Behav 10:395–407. doi:10.1007/bf02765606
Tauxe GM, MacWilliam D, Boyle SM, Guda T, Ray A (2013) Targeting a
dual detector of skin and CO
2
to modify mosquito host seeking. Cell
155:1365–1379. doi:10.1016/j.cell.2013.11.013
Thomson RCM (1948) Studies on Anopheles gambiae and A. melas in
and around Lagos. Bull Entomol Res 38:527–558
Trpis M, Hausermann W (1978) Genetics of house-entering behavior in
East-African populations of Aedes aegypti (L) (Diptera:Culicidae)
and its relevance to speciation. Bull Ent Res 68:521–532
Tuno N, Kjaerandsen J, Badu K, Kruppa T (2010) Blood-feeding behav-
ior of Anopheles gambiae and Anopheles melas in Ghana, Western
Africa. J Med Entomol 47:28–31. doi:10.1603/033.047.0104
Turissini DA, Gamez S, White BJ (2014) Genome-wide patterns of
polymorphism in an inbred line of the African malaria mosquito
Anopheles gambiae. Genome Biol Evol. doi:10.1093/gbe/evu243
Turner SL, Li N, Guda T, Githure J, Cardé RT, Ray A (2011) Ultra-
prolonged activation ofCO
2
-sensing neurons disorients mosquitoes.
Nature 474:87–U114. doi:10.1038/nature10081
Webster B, Bruce T, Pickett J, Hardie J (2010) Volatiles functioning as
host cues in a blend become nonhost cues when presented alone to
the black bean aphid. Anim Behav 79:451–457. doi:10.1016/j.
anbehav.2009.11.028
66 J Chem Ecol (2015) 41:59–66