Author response: A persistent behavioral state enables sustained predation of humans by mosquitoes
Year: 2022
Type: peer-review
Abstract: Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Predatory animals pursue prey in a noisy sensory landscape, deciding when to continue or abandon their chase. The mosquito Aedes aegypti is a micropredator that first detects humans at a distance through sensory cues such as carbon dioxide. As a mosquito nears its target, it senses more proximal cues such as body heat that guide it to a meal of blood. How long the search for blood continues after initial detection of a human is not known. Here, we show that a 5 s optogenetic pulse of fictive carbon dioxide induced a persistent behavioral state in female mosquitoes that lasted for more than 10 min. This state is highly specific to females searching for a blood meal and was not induced in recently blood-fed females or in males, who do not feed on blood. In males that lack the gene fruitless, which controls persistent social behaviors in other insects, fictive carbon dioxide induced a long-lasting behavior response resembling the predatory state of females. Finally, we show that the persistent state triggered by detection of fictive carbon dioxide enabled females to engorge on a blood meal mimic offered up to 14 min after the initial 5 s stimulus. Our results demonstrate that a persistent internal state allows female mosquitoes to integrate multiple human sensory cues over long timescales, an ability that is key to their success as an apex micropredator of humans. Editor's evaluation This manuscript describes a female mosquito's behaviour after a brief exposure to CO2, which has long been known to trigger host-seeking in female mosquitoes. The authors develop optogenetic tools in Aedes aegypti which enables them to control the delivery of 'fictive' CO2 to mosquitoes. They use this to show that a brief pulse of fictive CO2 alters the behavioural state of female mosquitoes, which lasts about 15 minutes. The study provides new insights into how activation of CO2-sensing olfactory neurons alters the behavioural state of a mosquito towards sensory cues to increase host-seeking behaviors and will be of great value to the vector biology community, as well as to neurobiologists in general. https://doi.org/10.7554/eLife.76663.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Predatory animals first detect, then pursue, and ultimately capture prey (Endler, 1991). Because the pursuit phase can last for extended periods of time, it is critical for predators to persist in the chase even when the prey is not constantly detected. It is equally important for predators to abandon pursuit if enough time has elapsed that prey capture is unlikely to occur. This decision balances the need to obtain food with the expenditure of energy on unsuccessful hunts (Anholt et al., 1987; Williams et al., 2014). The duration of a pursuit could depend on the predator repeatedly sensing prey stimuli. Alternatively, it may be sustained by recent prior experience or a change in the internal state of the predator that outlasts individual prey stimuli. Micropredators such as the mosquito consume small quantities of their live prey rather than killing them outright (Lafferty and Kuris, 2002), but employ similar tactics to other pursuit predators. Female mosquitoes combine rich multisensory information from olfactory, visual, taste, mechanosensory, and contact chemosensory systems to hunt humans from whom they obtain blood to produce eggs. Carbon dioxide (CO2) produced by human breath is highly volatile, traveling long distances from the human host. Detection of CO2 by the mosquito results in an increase in flying behavior (Eiras and Jepson, 1991; McMeniman et al., 2014) and upwind flight (Dekker and Cardé, 2011; Dekker et al., 2005; Geier et al., 1999) that is often referred to as 'activation'. It has not been tested whether repeated sensory input from the host or reafferent signals of wind caused by flight are required to maintain this activation behavior. Mosquitoes require an additional, more proximal host cue such as body heat or skin odor for short-range attraction and to engorge on blood (Corfas and Vosshall, 2015; Dekker and Cardé, 2011; Dekker et al., 2005; Geier et al., 1999; McMeniman et al., 2014; van Breugel et al., 2015; Figure 1A). In natural settings, human sensory cues are typically brief and intermittent by the time they reach the mosquito due to turbulent air flows and long distances (Koehl, 2006). However, studies of insect navigation have documented only short-term responses to these stimuli on the order of a few seconds (Alvarez-Salvado et al., 2018; Dekker and Cardé, 2011; Demir et al., 2020; Pang et al., 2018). If mosquitoes possess the ability to retain information about their prey and combine it with future information, this may explain their success at locating and feeding on human blood. Although the short-term role of CO2 in mosquito behavior has been known for nearly 100 years (Rudolfs, 1922), the idea that CO2 induces a long-term change in the internal state of the mosquito has not previously been tested experimentally. Figure 1 with 1 supplement see all Download asset Open asset Optogenetic control of mosquito host seeking and blood feeding. (A) Schematic of human host cues experienced by a host-seeking mosquito over time. (B) Schematic of genetic reagents used for optogenetic activation of CO2-sensitive Gr3 sensory neurons. (C) Female Aedes aegypti, gray boxes indicating antenna (top) and maxillary palp (bottom). Photo: Alex Wild. (D) Intrinsic tdTomato fluorescence of whole mounted Gr3 > CsChrimson female mosquito antenna and maxillary palp. Scale bar: 50 µm. (E) Maximum-intensity projections of confocal Z-stacks of antennal lobes in the right-brain hemisphere of the indicated genotype with immunofluorescent labelling of tdTomato (red) and the synaptic marker BRP (grayscale). Scale bar: 10 µm. (F,G) Diagram (F) and stimulus protocol (G) of optogenetic behavior assay for mosquito movement. (H) Time maximum projection of a single mosquito in the assay in (F) for 30 s pre- (left) and post- (right) stimulus. Scale bar: 0.5 cm. (I) Velocity of individual mosquitoes of the indicated genotypes 30 s pre- and post-stimulus onset. Data are plotted as mean of individual mosquitoes (thin gray lines) with median across individuals indicated with thick black or red line (*p<0.0001, Wilcoxon signed rank test with Holm's correction for multiple comparison, n.s., not significant, n=70 mosquitoes/genotype). (J) Schematic of opto-membrane feeder (top) and stimulus protocol (bottom). (K) Still images of mosquitoes of the indicated genotype underneath the warm blood meal approximately 7 min after the start of red light stimulation. Scale bar: 1 cm. (L) Occupancy of mosquitoes on warm blood meal in the opto-membrane feeder. Data are plotted as mean (line) ± SEM (shading). Data labelled with different letters are significantly different at the 5 min timepoint (p<0.05, Kruskal-Wallis test followed by Nemenyi post hoc tests; n=6–7 trials per genotype/stimulus combination, 18–21 mosquitoes/trial). (M) Percent of mosquitoes visually scored as engorged at the conclusion of the experiment in (L). Data are plotted as dot-box plots (median: horizontal line, interquartile range: box, 1.5 times interquartile range: whiskers). Data labelled with different letters are significantly different (p<0.05, Kruskal-Wallis test followed by Nemenyi post hoc tests; n=7 trials per genotype/stimulus combination and 18–21 mosquitoes/trial). See also Figure 1—figure supplement 1 and Figure 1—source data 1. Figure 1—source data 1 Optogenetic control of mosquito host seeking and blood feeding. https://cdn.elifesciences.org/articles/76663/elife-76663-fig1-data1-v2.xlsx Download elife-76663-fig1-data1-v2.xlsx To study pursuit predation in the mosquito, we developed optogenetic tools to precisely deliver short pulses of fictive CO2. This allowed us to test the effect of activating CO2 sensory neurons with greater temporal resolution and without the continuous stimulus of air flow required for delivery of real CO2. We observed that detection of fictive prey led to a long-lasting behavioral change in female mosquitoes. Following a 5 s fictive CO2 stimulus, animals exhibited high-arousal behaviors and engorged on a blood meal mimic offered up to 14 min later. Neither males nor previously blood-fed females showed these effects, and this persistent internal state was not induced by optogenetic stimulation of a sweet taste pathway. Remarkably, males lacking the fruitless gene showed long-lasting responses to fictive CO2 resembling those in females, consistent with our prior observation that these mutants show some aspects of female mosquito behavior (Basrur et al., 2020). Our work identifies a persistent internal state that may explain how mosquitoes aggressively pursue human hosts for many minutes. This constitutes the first use of optogenetics to manipulate neural circuits in the mosquito, an advance that will enable a better understanding of this important disease vector. Results Fictive CO2 triggers blood feeding We created optogenetic tools in A. aegypti mosquitoes that allowed us to precisely activate sensory neurons that are specialized to detect CO2. To do this we generated a transgenic strain that expresses the red light-activated cation channel CsChrimson translationally fused to the tdTomato fluorescent reporter (Klapoetke et al., 2014) under control of the QF2/QUAS bipartite transcription system (Potter et al., 2010). We crossed this transgene into a strain that expresses the QF2 transcription factor in neurons that express the Gr3 CO2 receptor subunit (McMeniman et al., 2014; Younger et al., 2022; Figure 1B). CsChrimson-tdTomato was detected in maxillary palp neurons but not antennal neurons, consistent with the observation that maxillary palp neurons are exquisitely sensitive to CO2 (Grant et al., 1995; Figure 1C and D, Figure 1—figure supplement 1A). As expected, we found that CsChrimson-expressing neurons extended axons that innervated glomerulus MD1 in the antennal lobe of the mosquito brain (Figure 1E), which is known to be CO2-sensitive (Younger et al., 2020). To test whether mosquitoes responded to optogenetic activation of CO2 sensory neurons, we presented animals with a 5 s red light (627 nm) stimulus and tracked their movement (Figure 1F–I). Control animals carrying only the Gr3-QF2 driver or the QUAS-CsChrimson transgene reporter showed no response to red light. However, mosquitoes with both genetic elements (Gr3 > CsChrimson) increased their velocity in response to the stimulus (Figure 1H–I). The proportion of mosquitoes responding increased with light intensity (Figure 1—figure supplement 1B-D). The observed increase in activity is consistent with the known role of CO2 in activating mosquitoes. When combined with another host cue such as heat, CO2 is sufficient to elicit blood feeding in female mosquitoes (McMeniman et al., 2014). To test whether fictive CO2 sensation triggered by optogenetic activation of Gr3 sensory neurons could replace real CO2, we created a behavior assay called the opto-membrane feeder (Figure 1J). This assay consists of a cylindrical canister of mosquitoes surrounded by red light-emitting diodes (LEDs) and a warm blood meal behind a membrane sitting on top of the mesh lid. Mosquitoes were presented with either CO2, fictive CO2 via red light, or neither stimulus. Control mosquitoes with either Gr3-QF2 or QUAS-CsChrimson were attracted to the warm blood meal and engorged only when presented with CO2 but not when presented with red light (Figure 1K–M). In contrast, Gr3 > CsChrimson mosquitoes were attracted and engorged in the presence of either real or fictive CO2, the latter delivered as a red light stimulus (Figure 1K–M). These results demonstrate that mosquitoes interpreted optogenetic activation of the CO2 sensory neurons as a host cue that is sufficient to drive blood feeding. Fictive CO2 induces prolonged host-seeking behaviors Host seeking begins when female mosquitoes detect a human, typically by sensing volatile cues like CO2. Once activated by human odorants, they seek out the source of the cues, and upon landing, mosquitoes walk to locate a patch of skin to pierce (Liu and Vosshall, 2019; McMeniman et al., 2014; van Breugel et al., 2015). To understand the timing and nature of the mosquito response to transient host cues, we created an assay called the opto-thermocycler (Figure 2A–B). In this assay mosquitoes receive optogenetic light stimulation from above and heat through the mesh at the bottom of the assay chamber. The use of fictive CO2 delivered optogenetically was critical for studying the internal state of the mosquito after these transient host cues. Delivery and removal of real CO2 necessitates constant air flow, which is itself an important sensory cue for insects (Alvarez-Salvado et al., 2018; Kadakia et al., 2021; Suver et al., 2019). Figure 2 with 2 supplements see all Download asset Open asset Fictive CO2 induces a persistent behavior state. (A,B) Schematic of opto-thermocycler assay (A) and stimuli delivered (B). (C) Still image of a mosquito with pose tracking of nine points using Animal Part Tracker. Scale bar: 0.5 cm. (D) Still images of a mosquito exhibiting the indicated classified behaviors (top). Representative plot of proboscis length with classified behavior superimposed (bottom). Scale bar: 0.5 cm. (E) Ethograms of individual mosquitoes of the indicated genotypes. Data show 1 min before and an excerpt of the 2 min after the indicated stimuli from a 20 min experiment. Each row represents data from one mosquito. The experiment comprised a total of n=68–70 mosquitoes/condition. All data were sorted by probing, and every third mosquito (n=22–23) was selected for display here for clarity. (F,G) Quantification of walking (F) and probing (G) behavior exhibited by individual mosquitoes from the experiment in (E) during the 5 min after stimulus onset. Data are plotted as violin-box plots (median: horizontal line, interquartile range: box, 5th and 95th percentiles indicated: whiskers). Data labelled with different letters are significantly different (p<0.05, Kruskal-Wallis test followed by Nemenyi post hoc tests, n.s., not significant, n=68–70 mosquitoes/genotype, 1 stimulus per trial). (H) Plot of percent individual Gr3 > CsChrimson mosquitoes exhibiting the indicated behavior from 2 min before to 15 min after stimulus onset. Data from experiment in (E). (I,J) Plot of percent individual wild-type (I) and Gr4 > CsChrimson(J) mosquitoes exhibiting the indicated behavior from 1 min before to 2 min after stimulus onset (I) or 2 min before to 15 min after stimulus onset (J) excerpted from a 20 min experiment (I: n=140 mosquitoes, average of 2 stimulus presentations/mosquito; J: n=69 mosquitoes, average of 3 stimulus presentations/mosquito). Flower image used for Gr4 > CsChrimson indicates that plant nectar is a sugar source. (K,L) Quantification of (I) and (J) for 5 min after stimulus onset. Data are plotted as violin-box plots (median: horizontal line, interquartile range: box, 5th and 95th percentiles: whiskers). Distribution represents individual mosquitoes, averaged over multiple stimulus presentations. Data labelled with different letters are significantly different (p<0.05, Friedman test followed by Nemenyi post hoc tests, n.s., not significant). (M) Median additivity of heat and the indicated stimuli presented simultaneously. Additivity of 100% corresponds to the case when combined stimuli equal the sum of responses to individual stimuli. Data from E–L. See also Figure 2—figure supplement 1, Figure 2—figure supplement 2, and Figure 2—source data 1. Figure 2—source data 1 Fictive CO2 induces a persistent behavior state. https://cdn.elifesciences.org/articles/76663/elife-76663-fig2-data1-v2.xlsx Download elife-76663-fig2-data1-v2.xlsx We employed machine learning-based behavior classification as a high-throughput readout of mosquito behavioral responses (Figure 2C–D). We tracked nine points on the mosquito body using Animal Part Tracker (APT) and four behaviors using Janelia Automatic Animal Behavior Annotator (JAABA) (Kabra et al., 2013): grooming, flying, walking, and probing, a behavior in which the mosquito inserts its proboscis through the mesh in the bottom of the container. When none of these four behaviors were present, mosquitoes were predominantly motionless, showed slow hind leg movement, or occasionally flailed against the wall of the assay chamber without walking. All classifiers showed high accuracy with >90% true positive and true negative rates on a set of test video frames (Figure 2—source data 1). We delivered 5 s red light pulses and heat increases to simulate brief CO2 and human body heat stimuli (Figure 2B). Mosquitoes responded to individual heat and fictive CO2 stimuli with elevated walking, flying, and probing (Figure 2E–H, Figure 2—figure supplement 1). The response to heat alone was dominated by probing and returned to baseline after about 1 min (t1/2 probing = 0.4 min). In contrast, fictive CO2 alone caused an immediate flight and probing response followed by sustained walking, flying, and probing (Figure 2H, Figure 2—figure supplement 1, Video 1) that took approximately 15 min to return to baseline (t1/2 probing = 3.9 min). Varying the light intensity changed the proportion of mosquitoes responding but not the duration of the response (Figure 1—figure supplement 1B-D). The long duration of the response to CO2 is reminiscent of persistent internal states in other organisms (Asahina et al., 2014; Flavell et al., 2013; Kohatsu et al., 2011; Hindmarsh Sten et al., 2021; Marques et al., 2020). Video 1 Download asset This video cannot be played in place because your browser does support HTML5 video. You may still download the video for offline viewing. Download as MPEG-4 Download as WebM Download as Ogg Mosquito response to fictive CO2. Shown is video of the opto-thermocycler assay of female Gr3 > CsChrimson mosquitoes responding to a 5 s pulse of red light. Points tracked on the body of the mosquito are indicated by white semi-transparent dots and behaviors are indicated with text. Stimuli are indicated in the upper left corner. These observations of mosquito behavior could reflect an internal state specific to host seeking or a general arousal state elicited by many sensory stimuli. Like CO2, bright light is also an arousal signal in mosquitoes (Araripe et al., 2018) so we asked whether bright light induces a long-lasting behavior state. Mosquitoes have weaker visual sensitivity to red wavelengths (San Alberto et al., 2021), and we saw no behavioral response to red light (Figure 2E), so we used green light. Because this experiment is designed to test whether mosquitoes have a response to a visual stimulus, a question that does not depend on the use of optogenetics, we used wild-type mosquitoes. A bright green light stimulus induced a brief response dominated by walking (t1/2 walking = 0.4 min), than the response to fictive CO2 (Figure and CO2 and heat elicit the required for females to produce mosquitoes possess a feeding for of nectar for energy et al., 2020; et al., 2020). We asked whether optogenetic stimulation of Gr4 sensory neurons that are to detect sugar a sustained behavior response in mosquitoes as with optogenetic activation of the CO2 sensory neurons. We found that fictive sugar elicited a response of probing that was of duration than fictive CO2 (t1/2 probing = 1.5 (Figure and The temporal resolution of our allowed us to understand precisely how mosquitoes integrate CO2 and heat to their behavior. We on the first 15 s after stimulus onset during which the behavior responses are We found that the probing response when the stimuli were presented was with to the individual stimuli (Figure Figure 2—figure supplement In contrast, flying was This that of host cues from flight and probing, a short-range behavior. is a host cue but also may be experienced by the mosquito under other We the of heat with green light and fictive sugar to see if they are to the host cue CO2. In to the of host cues, probing was when cues were presented (Figure Figure 2—figure supplement This that the mosquito system different for the of host cues and cues. The persistent state is specific to host seeking The that bright green light and fictive sugar stimuli elicited responses that the prolonged response to fictive CO2 is specific to host If it be to host-seeking behavior. Host seeking in A. aegypti is after a female a blood only after et al., We allowed female mosquitoes to blood feed on a human and then their behavior later. We found that blood-fed females their response to fictive CO2 and nearly their response to heat (Figure and comparison, we asked whether blood-fed females could to the optogenetic sugar stimulus by activating Gr4 sensory neurons. We found that blood-fed females responses to fictive but fictive CO2, the response was still (Figure supplement This that blood-fed females show a of the persistent state elicited by fictive CO2. Figure 3 with 1 supplement see all Download asset Open asset The persistent state is specific to host of female blood-fed female and (C) Gr3 > CsChrimson mosquitoes to the indicated behavior from 2 min before to 15 min after stimulus onset average of 3 stimulus presentations/mosquito). Quantification of walking (D) and probing (E) from data in for 5 min after stimulus onset. The behavioral response of males of the indicated genotype to the indicated behavior from 2 min before to 15 min after stimulus onset average of 3 stimulus presentations/mosquito). (I,J) Quantification of walking (I) and probing (J) from data in for 5 min after stimulus onset. In and data are plotted as violin-box plots (median: horizontal line, interquartile range: box, 5th and 95th percentiles: whiskers). The represents individual mosquitoes, averaged over multiple stimulus presentations. Data labelled with different letters are significantly different (p<0.05, Kruskal-Wallis test followed by Nemenyi post hoc tests, n.s., not significant). See also Figure supplement 1 and Figure data 1. Figure data 1 The persistent state is specific to host Download females, mosquitoes do not seek out hosts to feed on blood. do demonstrate a flight response to CO2 et al., and are to in the of humans they with female mosquitoes We observed that mosquitoes responses to heat, but flight and walking responses to fictive CO2 (Figure However, the response to fictive CO2 was to baseline (t1/2 probing = 0.4 min). This observation that the not the initial in a We have previously that fruitless mosquitoes attraction to human odor (Basrur et al., 2020). We asked whether these fruitless males have an response to fictive CO2 stimuli. we that fruitless males lacking Gr3 > CsChrimson not to red light, as (Figure fruitless fictive CO2 showed a brief response (t1/2 probing = 0.4 (Figure and similar to the response we saw for wild-type males (Figure comparison, fruitless males fictive CO2 showed a and sustained response for after the stimulus (t1/2 probing = (Figure This that fruitless is in the of the of the persistent host-seeking state. We that the duration of the sustained response of fruitless males is than in females, may this internal state. these results demonstrate that blood-fed females and males, which do not in lack sustained responses to brief pulses of fictive CO2. This the that the persistent state is the behavioral by which the of blood feeding is sustained in the female mosquito. Mosquitoes integrate sensory cues for consists of increased arousal and a We have that fictive CO2 induces a prolonged increase in movement and the probing behavior that blood feeding. We asked whether the persistent state induced by a brief pulse of fictive CO2 can the response to body heat and ultimately if it can blood feeding many we tested whether fictive CO2 responses to heat (Figure Because CO2 is highly volatile, mosquitoes this cue before body heat in host-seeking When we presented heat first followed by fictive CO2, small behavioral responses were (Figure However, when fictive CO2 was presented with heat or 15 or s prior to heat, walking and probing responses were (Figure This that mosquitoes to the temporal order of these host cues. we asked how long prior to the heat stimulus fictive CO2 can the response (Figure to heat walking and probing were increased when a fictive CO2 stimulus was presented up to min prior (Figure This that in to the sustained behavior fictive CO2 increases the response to heat for Figure with 2 supplements see all Download asset Open asset The persistent state host cues and decision in time. Schematic of stimuli and of walking and probing in the 5 min after the first stimulus onset mosquitoes, 2 stimulus presentations/mosquito). In data are as violin-box plots (median: horizontal line, interquartile range: box, 5th and 95th percentiles: whiskers). Data labelled with different letters are significantly different (p<0.05, Friedman test followed by Nemenyi post hoc Data that are significantly different from heat or light are in or red (G) Schematic of blood assay which a blood meal (H) Still images of (top) and mosquitoes. Scale bar: 1 (I) Percent mosquitoes engorged in response to the indicated stimuli in the blood assay 14 mosquitoes/trial). Data labelled with different letters are significantly different (p<0.05, followed by post hoc (J) Percent mosquitoes engorged in response to the indicated cues in the blood assay Data in (I,J) are as dot-box plots (median: horizontal line, interquartile range: box, 1.5 times interquartile range: whiskers). plots in (J) with a red data combined light and heat stimuli are greater than the sum of individual stimuli after for multiple using Holm's (K) for of the four states indicating the proportion of behavior and the of with and (L) states of individual mosquitoes from the 2 min stimulus from experiment in into those that were visually scored as or at the of the experiment. Each row represents data from one mosquito and data are sorted to the of search during the indicates none of the four states were the mosquito was (M) Quantification of the percent of time mosquitoes in the indicated state during the 2 min Data are plotted as violin-box plots (median: horizontal line, interquartile range: box, 5th and 95th percentiles: n.s., not significant). of a on the proportion of behavior in and used
Source: eLife
Institutions: Rockefeller University, Howard Hughes Medical Institute
Language: English
Subfield: Cellular and Molecular Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $2,500
OpenAlex ID: W4281632370