Method Article

Optogenetic Analysis of Behavior in the Mosquito Aedes aegypti

196 views

⸱

DOI:

10.3791/71394

⸱

August 21st, 2026

 ,  ,  ,  ,  , 
1Department of Genetics, Yale University, 2Kavli Institute for Neuroscience, Yale University, 3Howard Hughes Medical Institute, 4Wu Tsai Institute, Yale University

* These authors contributed equally

In This Article

Summary

This protocol describes the use of optogenetics to understand the behavior of the mosquito Aedes aegypti. Methods for rearing, preparation, assays, and analysis are presented to assess neuronal function during attraction to host animals and blood feeding. These foundational techniques will enable neural circuit and behavioral studies of this important vector of human disease.

Abstract

The mosquito Aedes aegypti is an important vector of viral pathogens and serves as a model for other vector species. Pathogens are transmitted when a mosquito bites a host animal, but the neural circuits that control seeking and biting behavior remain unknown. This paper details methods for manipulating neural activity in the mosquito using optogenetics, a key technique for determining the causal relationship between neural circuits and behavior. These methods include rearing mosquitoes for optogenetics and three assays designed to measure different steps in the sequence of arousal, attraction, proboscis probing, and engorgement on host blood. These behaviors occur at different spatial scales and in response to different sensory stimuli. Each behavioral assay is outfitted with red (~625 nm) LEDs for optogenetic activation. To detect arousal in response to olfactory stimuli, flight and walking are quantified. To assay attraction or landing, mosquitoes are presented with a heated blood meal in a large arena. Proboscis probing and engorgement are assayed with higher resolution video, enabling measurement of appendages and abdominal size. The protocol describes machine vision models to enable high-resolution temporal quantification of behavior and endpoint measurements of feeding. These methods can be used to test the role of any genetically accessible population of neurons in mosquito biting behavior and can be extended to additional behaviors.

Introduction

The mosquito is one of the most dangerous animals in the world, not for its ferocity or strength but for its efficiency as a vector for disease. Malaria-causing parasites spread by Anopheles mosquitoes are responsible for hundreds of thousands of deaths yearly, while Aedes aegypti and Aedes albopictus are responsible for the spread of yellow fever, dengue, chikungunya, Zika fever, and many other life-threatening diseases1. Mosquitoes are present on every continent except Antarctica and pose a great concern for public health2. Blood-feeding behavior directly transmits pathogens, but it is also an essential step in the mosquito life cycle, as a vertebrate blood meal is required to lay eggs. Curbing the spread of mosquito-borne diseases thus requires understanding their host-seeking mechanisms.

It is well-established that female mosquitoes make use of long- and short-distance cues to locate hosts3,4. Carbon dioxide (CO2) is a highly volatile gas present in human exhalation and is often the first cue that a mosquito senses. It is detected in Aedes aegypti olfactory sensory neurons by Gr1, Gr2, and Gr3 proteins, which form the CO2 receptor5. Exposure to CO2 causes take-off and flight behavior, also known as activation6,7, which can persist for up to 10–15 min, even in the absence of further stimuli8. CO2 synergizes with skin odor to enhance activation and attraction5,9. Following CO2 and skin odor detection, the mosquito is primed to sense additional, more proximal host cues, such as heat, visual contrast, and humidity, that guide it to land on a viable blood meal host5,10,11,12,13. Integration of multiple sensory cues allows mosquitoes to effectively pursue and locate their host.

When a female mosquito lands on a potential host, the mosquito thrusts its proboscis into the surface until a patch of skin is located, a process known as probing14,15. Following probing, the stylet pierces the skin, locates a blood vessel, and simultaneously ingests blood while injecting saliva, enabling the mosquito to double its body weight in several minutes16. Blood is directed to the midgut, where proteins are digested for egg production. Interestingly, engorgement on a host is not necessary for pathogen spread, as probing alone can transmit disease16,17. Therefore, understanding each step in biting behavior could allow the development of strategies to disrupt female mosquito host-seeking and biting, and ultimately to control the spread of mosquito-borne diseases.

Numerous tools have been developed in model organisms to manipulate neuronal activity. Early tools worked constitutively or by thermal activation18,19,20,21. While useful, these genetic reagents pose certain drawbacks. Constitutive tools can affect development, leading to phenotypes not due solely to the circuit function of the neurons. Mosquitoes use body heat to identify host vertebrates, making thermally activated channels uniquely unsuitable for studies of their behavior22,23. In studies of behavior, optogenetics refers to the use of genetic tools that respond to specific wavelengths of light to control the activation of neurons of interest. This technique has already been used in multiple systems, from examining the valence of thirst in mice to studying courtship behavior in Drosophila, to restoring vision loss from retinitis pigmentosa in human trials24,25,26. Optogenetic channels work by absorbing photons and allowing specific types of ions to pass through the plasma membrane, activating or inhibiting the neuron. By introducing transgenic light-sensitive protein-coding genes into an organism, this technique can be used to activate or inhibit specific neurons that would otherwise be impossible to manipulate. Optogenetics provides advantages in temporal and spatial control, and a range of controls that enable experimental inference.

The authors recently described the first use of optogenetic tools to study mosquito behavior8. These tools rely on the red-light-activated cation channel CsChrimson, a red-shifted channelrhodopsin27. Microbial opsins, including CsChrimson, require a chemical cofactor, all-trans retinal, in order to absorb photons28. CsChrimson is fused to the tdTomato fluorescent reporter and is under the control of the QF2/QUAS binary expression system, enabling it to be flexibly combined with different driver lines29. To target CO2 sensory neurons, a driver line was used that expresses the QF2 transcription factor in neurons that express the Gr3 subunit of the CO2 receptor30. Mosquitoes showed minimal behavioral response after exposure to red light alone8; additionally, red light is most effective at penetrating biological tissue, making a red-light-activated opsin ideal for optogenetics25.

This paper describes protocols for three assays that measure different aspects of host-seeking and blood-feeding behavior combined with optogenetic stimulation (Figure 1A). First, a protocol for rearing optogenetic mosquito lines is provided. The second section describes the opto-thermocycler, a modified PCR thermocycler that can be used to track arousal and probing when mosquitoes are exposed to temperature and light stimuli. The opto-thermocycler assay is designed to measure these search behaviors across extended timescales, using driver strains to control neuronal activity with optogenetics. The third section describes the blood blanket assay, in which mosquitoes have constant access to a thin artificial blood meal on the base of a modified thermocycler, enabling tracking of flight, probing, and engorgement under varying light and temperature conditions. The blood blanket assay, built on the opto-thermocycler, provides a means to quantify engorgement on an artificial blood meal, enabling the study of mosquito feeding behaviors with precisely timed optogenetic and heat stimuli. The final section describes the opto-membrane feeder, in which mosquitoes are presented with a warm blood meal in a canister, allowing tracking of both attraction to a blood source and engorgement under exposure to light stimuli. The opto-membrane feeder assay is used to investigate the full behavioral sequence of host-seeking and blood-feeding in a free-flight system. In addition to the expected outputs for each assay, machine vision models are described to track mosquito body parts and identify behaviors during the assays. The opto-thermocycler, blood blanket, and opto-membrane feeder assays enable measurement of distinct stages of mosquito blood-feeding behavior while manipulating neural activity with light. These innovative techniques allow the study of how mosquito neural circuits control behaviors that pose a threat to public health.

Protocol

Mosquitoes used in this protocol contain transgenes, and under United States Department of Agriculture regulations, must be reared in Arthropod Containment Level II facilities32. Mosquitoes are contained to prevent escape and disposed of in biohazard waste. Biosafety and arthropod containment protocols were examined and approved by Yale Environmental Health & Safety. This work makes no use of human or animal subjects. Mosquito genotype can be confirmed by fluorescent markers (3xP3-dsRed for Gr3-QF2 and 3xP3-CFP for QUAS-CsChrimson) or by PCR for the inserted construct8. All the materials used in this study are listed in the Table of Materials.

1. Optogenetic rearing

  1. Rear mosquitoes according to standard protocol31 with the following exceptions to maximize the response of optogenetic lines during experimentation.
  2. Make crosses using homozygous Gr3-QF2, QUAS-CsChrimson, and wild-type Liverpool strains to generate the heterozygous experimental lines Gr3-QF2/+, QUAS-CsChrimson/+, and Gr3>CsChrimson.
  3. Raise mosquitoes inside a light-tight incubator set to 28 °C and 80% humidity, on a 14 h blue light (450 nm), 10 h dark cycle (Figure 1B).
    NOTE: Using blue light minimizes activation of red-light-activated channels, such as CsChrimson. All rearing steps performed outside the incubator should be conducted in low-light conditions.
  4. Begin experiments on adult mosquitoes 1 week post eclosion at the earliest.
    NOTE: While the ideal age for behavioral experimentation is 2–3 weeks, mosquitoes up to 4 weeks post eclosion can be used. Control and experimental genotypes should be the same age, recorded as days post-eclosion. To standardize animal ages, all eggs should be hatched on the same day.
  5. Two days prior to experimentation, gently aspirate mated, non-blood-fed female mosquitoes out of the cages and anesthetize at 4 °C. Using a feather or aspirator, carefully transfer to the experimental containers according to the assay being used.
  6. Prepare a 0.2 M aliquot of all-trans retinal (ATR) in dimethyl sulfoxide (DMSO), stored at -20 °C. Provide 10% sucrose and 400 µM ATR via soaked cotton wicks on the mesh of the container. Animals may feed for 1–3 days in the dark on this meal.
    NOTE: Feeding in the dark, as opposed to under blue light, prevents bleaching of the ATR (Figure 1B). Additionally, across all steps of the protocol, mosquitoes were anesthetized by exposure to 4 °C for 5–10 min, until they no longer moved. At the end of each experiment, mosquitoes were euthanized by exposure to -20 °C for 24 h.
    CAUTION: All-trans-retinal can cause severe irritation when it comes into contact with the skin and eyes. Wear proper PPE such as a lab coat, gloves, and eye protection when handling. Dimethyl sulfoxide is a combustible liquid that can be absorbed through the skin upon contact and must be handled in a dry environment with proper PPE and eye protection. DMSO should also be stored in fire-proof chemical storage containers.
  7. One day (20–24 h) before experimentation, replace the sucrose feeders with new feeders containing distilled water and 400 µM ATR. Animals remain in the dark until experimentation (Figure 1B). Proceed to section 2, 3, or 4.
    NOTE: The experiments should be run during the mosquitoes’ daytime photoperiod to maximize response rate.

2. Opto-thermocycler

  1. Assembling components of the assay
    1. Make the acrylic plates to contain the mosquitoes (Figure 2). Refer to Figure 2A for a diagram, Figure 2B for electronic components, and Figure 2C for an image of the acrylic plate. See files in (https://github.com/sorrellslab/opto-thermocycler/releases/tag/v1.0.0) for templates.
      1. To create the walls, laser-cut a 3 mm acrylic sheet into four 10 cm by 1 cm rectangles and six 7 cm by 1 cm rectangles. Affix the rectangles with acrylic glue into a five by three matrix, where each unit is a 17 mm by 17 mm by 10 mm chamber. The template for these pieces is in the file “14_animal_thick_v3.ai”.
        ​CAUTION: Acrylic glue is combustible and can cause respiratory irritation, as well as eye irritation. Proper PPE should be used while handling acrylic glue, such as gloves, a lab coat, and goggles, and this solvent should be used in a well-ventilated area, away from open flame.
      2. For the bottom of the chamber, cut a 1.5 mm-thick acrylic sheet to 6.985 cm by 10.16 cm with the bottom right well removed. Cut a 1.5 mm-thick acrylic sheet into a 6.985 cm by 10.16 cm rectangle for the lid. The template for these pieces is “14_animal_thin_v3.ai”.
    2. Cut a UV-resistant black mesh sheet to the same size as the bottom acrylic piece (Figure 2C). Sandwich the mesh between the plate walls and the bottom piece, then glue them in place with acrylic glue. Remove the mesh from the bottom-right well so the hole is unobstructed.
  2. Construct a frame that will hold the camera and LEDs (Figure 2A).
    1. Build a frame with four pillars at each corner of the thermocycler using 60.96 cm, 25 mm square optical construction rails. Use two 22.86 cm, 25 mm construction rails to connect the front-right pillar to the back-right pillar and the front-left pillar to the back-left pillar, 30 cm from the bench.
    2. Connect the right and left pillars in the front with a 30.48 cm, 25 mm construction rail, elevated 24 cm off the bench, and in the back with a 30.48 cm, 25 mm construction rail elevated 42 cm off the bench.
    3. Attach 1-foot-long optical posts to the construction rail on the left and right sides of the frame. Tie six 627 nm (red) LEDs connected in series 24 cm above the PCR block of the thermocycler to each post.
      NOTE: Although the wavelength of light resulting in peak activity of CsChrimson is ~590 nm, wavelengths near ~625 nm balance the need to activate CsChrimson and avoid behavioral responses to light in genetic controls27. Tape a heat sink and lens on each LED to focus the light toward the block of the thermocycler.
    4. Fasten a 30.48 cm optical post to the top of the front two construction rail pillars. Use a right-angle clamp in the middle of the optical post to attach a 15.24 cm optical post facing inwards over the thermocycler block. Affix a camera with an infrared long-pass filter to the end, pointing down, 40 cm above the block.
    5. Using optical posts, create a frame that extends from the construction rail in a rectangle around the PCR block, elevated just above the surface. Affix a strip of infrared LEDs, facing the PCR block, to illuminate the mosquitoes at a right angle to the camera.
      NOTE: The infrared illumination used for video imaging does not measurably heat the arena, as the thermocycler surface is maintained at 24 °C throughout the experiment.
    6. Drape a blackout curtain over the structure to prevent interference from ambient light.
    7. Cover the surface and sides of the PCR block with a single layer of blackout tape to create a flat, even surface with a dark background.
    8. Place the thermocouple sensor on the PCR block in the lower right corner and secure it with blackout tape. Confirm even light distribution across the PCR block using a power meter. Adjust the power supply voltage or current to deliver light at 12 µW/mm2.
    9. On the edge of the PCR block, place an infrared LED to synchronize the video with the stimulus delivery output (Figure 2C). Cover the surface with black tape to allow the light to be detected without creating a glare in the camera.
    10. Wire the thermocouple, synchronization LED, and red light LEDs according to the diagram in Figure 2B. Download the required software (SpinView (version 1.27.0.48), CoolTerm (version 2.0.1), Arduino IDE (version 1.8.12)). Download the Arduino code and thermocycler program from the GitHub repository.
      ​NOTE: The Arduino code is programmed to deliver red light, body-temperature heat, and the combination of the two stimuli. There are seven stimulus programs, with the stimulus order randomized among them, provided as resources. These can be modified to deliver different stimuli. There are nine stimuli per experiment: three of heat only, three of light only, and three of heat and light simultaneously. Stimulus delivery operates by the Arduino detecting changes in the surface temperature of the PCR block via the thermocouple, then delivering light stimuli at the specified times. Additional details can be found in the comments within the stimulus programs.
  3. Trial preparation
    1. Rear mosquitoes according to the optogenetic rearing protocol; see section 1 and Figure 1B.
    2. Two days before the experiment, transfer 14 female mosquitoes with an aspirator to a separate container. Provide ATR to the mosquitoes according to step 1.6. If multiple genotypes are being compared, blind them to the experimenter.
      ​NOTE: Once mosquitoes receive ATR, they should be handled in dim light.
    3. One day before the experiment (20–24 h), aspirate 14 female mosquitoes per acrylic plate and anesthetize them at 4 °C. Carefully place one mosquito at a time into each of the wells with a feather or aspirator.
    4. Tape the side of the acrylic plate with Scotch tape to hold the lid in place. Do not fold the tape over the top or bottom of the acrylic plate, as this would obstruct the camera's view.
    5. Place the acrylic plate on top of three cotton wicks soaked with 400 µM ATR in distilled water (See step 1.6), placed in parallel. Align each of the three rows to a wick so all the mosquitoes have access to the water-ATR solution. Place the acrylic plate in a dark incubator overnight.
  4. Running the trial
    1. Connect the camera, LEDs, and Arduino to the computer. Turn on the power supply. Wipe down the opto-thermocycler surface with 70% ethanol and allow it to dry.
      ​CAUTION: Ethanol is flammable and may cause serious eye irritation. Proper PPE should be worn when handling ethanol. Additionally, avoid any open flame.
    2. Using a dark container, transfer the acrylic plate of mosquitoes to the middle of the PCR block. Cover the opto-thermocycler with a blackout curtain.
    3. Open SpinView on the computer. Click on the camera name to open the camera view. Plug in the infrared lights and make sure the PCR block and acrylic plate are visible in the camera view.
    4. If needed, adjust the acrylic plate so it is aligned, centered, and the thermocouple is visible in the bottom right corner of the acrylic plate. Ensure the synchronization light is in frame (Figure 2C).
    5. Adjust the settings on the camera and in SpinView to ensure a crisp, cropped video showing the acrylic plate. In the Image Format tab, adjust the Width, Height, Offset X, and Offset Y options to ensure the acrylic plate and synchronization light are visible. In the Settings tab, set Acquisition Frame Rate Enable to true and Acquisition Frame Rate to 30 fps.
    6. Turn Exposure Auto and Gain Auto off. Adjust the Exposure Time and Gain values to create a clean contrast between the black background and the mosquitoes. Click the red record button in the upper-right corner to open a separate window to save the video file.
    7. In the pop-up window, set the file path at the top: Click Browse, navigate to the experimental folder, click New folder, and label it with the trial name. Set the number of frames to record to the appropriate number for the experiment (e.g., 43,500) or to 0 to manually stop the recording.
    8. In the Videos tab, set the following fields: Video Recording Type: MJPG, Video File Split Size: 1000 MB, JPEG Compression Quality: 65. Open the ‘optothermo_lightheat_random1’ program on the thermocycler and start it. The thermocycler should begin in a hold step at 25 °C.
    9. Open the matching program, ‘TC_optothermo_lightheat_random1.ino’, on the Arduino IDE, and upload it to the board. Open the serial monitor and ensure that the temperature is approximately 25 °C.
    10. Close the serial monitor in the Arduino IDE. Open CoolTerm to begin acquiring the Arduino serial monitor output. Make sure the serial port and channel match those in the Arduino IDE. The output starts showing on the CoolTerm window.
    11. Begin the recording of the Arduino output to a Text/Binary file. Start the video recording on Spinview. Click Resume to end the thermocycler's hold. Watch the Arduino output in CoolTerm to check whether the temperature dips to 18 °C and the stimulus counter increases to 1, initiating the trial. Allow the assay to run until the trial is completed.
  5. Clean up
    1. Stop the video recording and CoolTerm output acquisition. Close out of both programs. End the thermocycler program. Remove the acrylic plate of mosquitoes from the thermocycler and place it in a -20 °C freezer overnight to euthanize the mosquitoes.
      ​CAUTION: When euthanizing transgenic mosquitoes, ensure they remain in a -20 °C environment for at least 24 h before disposing of them in a biosafety hazard bin to ensure death.
    2. If starting another trial, repeat section 2.3. Unplug the camera and Arduino cables. Turn off the power supply. Unplug the infrared LEDs. Clean acrylic plates by removing the Scotch tape and discarding the dead mosquitoes.
    3. Spray the lid and acrylic plate with 70% ethanol and wipe all the sides of the wells and mesh with a gloved finger carefully. Rinse thoroughly with deionized water, then stack acrylic plates for storage, separating lids with a paper towel to avoid scratching.

3. Blood blanket assay

  1. Assembling components of the assay (Figure 3)
    1. Construct the opto-thermocycler assay as described in section 2.1. Refer to the figures for the assay (Figure 3A), thermal stimulus (Figure 3B), and aluminum plate (Figure 3C).
    2. Construct an aluminum blood meal feeding plate as seen in Figure 3C.
      1. Using a laser cutter or waterjet cutter, construct the top and bottom of the plate. Construct the bottom of the plate by cutting a 0.8 mm-thick aluminum sheet to 90 mm by 134 mm. Construct the top by cutting a 1.6 mm-thick aluminum sheet to 90 mm by 134 mm. Cut 15 rectangular holes (17.2 mm by 18.6 mm) into the top plate. The template files can be found in the blood blanket GitHub repository (https://github.com/sorrellslab/bloodblanket) under "Aluminum plate templates".
      2. Sandwich between the aluminum parts a thin silicone gasket (0.5 mm thick) to prevent leakage. Cut using a laser cutter to match the shape of the aluminum top.
      3. Attach layers together by eight 0.47625 cm, dome-topped hex 4–40 standard bolts. Use a forming tap to create threads for the bolts in the metal pieces. Connect these through round and “W” washers on the bottom of the plate for added security when placed on top of the thermocycler block.
    3. Download the Arduino code and thermocycler program for the blood blanket assay from the GitHub repository (https://github.com/sorrellslab/bloodblanket).
    4. Prepare aliquots of 20 mM ATP in 25 mM NaHCO3 and store at -20 ˚C until use.
      ​NOTE: Prepare a fresh ATP solution every six months. Avoid repeated freeze-thaw cycles. Prepare 1 M NaHCO3 and 5 M NaCl (store each at room temperature). NaHCO3 is only stable in solution for 7 days.
  2. Experiment preparation
    1. Rear mosquitoes and transfer to acrylic plates as described in sections 1 and 2.2. If comparisons are going to be made between different genotypes, blind them to the experimenter.
    2. The day of the experiment, prepare a 10 mL minimal artificial blood meal (110 mM NaCl, 20 mM NaHCO3, and 1.5 mM ATP) by adding 750 µL of 20 mM ATP, 220 µL of 5 M NaCl, and 200 µL of 1 M NaHCO3 stock solutions to 8.83 mL of deionized water. Mix thoroughly by inverting.
      NOTE: Make the artificial blood meal fresh prior to each experiment.
    3. Set up SpinView recording according to steps 2.3.4 to 2.3.8.
    4. Open “Red-light-Only__BB_.ino” Arduino file (or desired stimulus file) for the experiment.
  3. Running trials
    1. Pipette 650 µL of the minimal artificial blood meal into each well of the plate, ensuring that the meal entirely fills each well with no overflow.
    2. Stretch a 10.2 cm by 10.2 cm square of parafilm over a 20.3 cm diameter ring. Place the parafilm over the plate, pressing the edges until it adheres. Using a scalpel, trim a 1.3 cm-wide border around the plate and tuck the excess parafilm underneath to seal the meal.
    3. Place the plate directly on top of the PCR block to allow maximum heat transfer (Figure 3A).
      NOTE: Blackout tape is used on the surface of the PCR block for the opto-thermocycler, but not for the blood blanket.
    4. Place the thermocouple onto the surface of the parafilm in the bottom right corner of the aluminum plate. Connect the camera, LEDs, and Arduino to the computer. Turn on the power supply. Start the “BB_thermocycler_program.x50prog” program on the thermocycler.
      NOTE: The rate of temperature changes is the maximum ramp speed for the thermocycler, chosen to best mimic the landing of a mosquito onto a host for feeding.
    5. Using a dark container and dim room lights to minimize light exposure, transfer the acrylic plate of mosquitoes onto the aluminum plate, ensuring each well is aligned with the mosquitoes. Ensure the thermocouple is in the bottom-right corner of the plate and in contact with the parafilm to measure the temperature of the artificial blood meal.
    6. Start the program by re-uploading the “Red-light-Only__BB_.ino” program in the Arduino IDE. Close the serial monitor. In CoolTerm, press Connect to see the current temperature. Using this starting temperature, adjust the trigger temperature in the Arduino script to approximately 0.6 °C below it.
    7. In the CoolTerm window, select Capture to Text/Binary File and start data collection. Start video recording in the record window in SpinView. Click Resume to end the thermocycler's hold. Watch the Arduino output on CoolTerm to ensure that the heat dip was detected and the trial has started.
    8. After the duration of the trial, click Stop Recording. On CoolTerm in the Capture to Text/Binary File tab, select Stop writing the serial output.
    9. Remove the plate of mosquitoes from the blood blanket plate and move them to the cold room to anesthetize them. Count the number of mosquitoes with enlarged abdomens that contain the artificial blood meal as engorged, including those that have partially fed.
    10. Between trials, clean the surface of the PCR block by wiping with a Kimwipe moistened with 70% ethanol. Remove the parafilm and rinse the aluminum plate with deionized water. Repeat section 3.3 for further trials. Rotate the order of trials between days to minimize the effect of circadian activity cycles.
  4. Clean up
    1. Stop recording all programs and quit. Unplug USB cords for the camera, Arduino, and infrared lights. Shut down the computer. Stop the “thermocycler_program.x50prog” program running on the thermocycler.
    2. Place mosquito plates in a -20 °C freezer overnight to kill the mosquitoes. Wash the aluminum plate with deionized water. Douse the acrylic plate with 70% ethanol, then rinse with deionized water and leave to dry overnight before loading more mosquitoes.
      NOTE: Plates should be dried overnight after washing to ensure no moisture remains. If there is moisture, this will cause condensation on the lid and will interfere with the video.

4. Opto-membrane feeder assay

  1. Assembling components of the assay (Figure 4)
    1. Assemble a frame using opto-mechanical components and a 30.48 cm by 30.48 cm, black, 0.635 cm-thick, acrylic platform. Create a hole in the center of the acrylic with a diameter of 11.43 cm using a laser cutter. Cut holes 2.54 cm from each corner to allow the acrylic to rest on the frame supports, approximately 23.495 cm above the base.
    2. In the center, attach a black acrylic ring with a 19.05 cm inner diameter and a 20.32 cm outer diameter (Figure 4A). Templates for all acrylic components are in the opto-membrane feeder GitHub repository (https://github.com/sorrellslab/opto-membranefeeder/releases/tag/v1.0.0) along with the Arduino programs. See figure for overall setup (Figure 4A), stimulus (Figure 4B), circuit diagram (Figure 4C), and video still image (Figure 4D).
      ​NOTE: The specific materials used to assemble this platform are not critical, so more cost-effective options may be available.
    3. Attach a camera with an infrared long-pass filter below the center hole, pointing up toward the mosquito canister. Adjust the aperture to be fully open, allowing selective focus on the mesh top of the canister (Figure 4A).
    4. To make canisters, cut a clear, polycarbonate tube with a diameter of 11.43 cm into 12.7 cm cylindrical segments.
      1. Build bottoms for the canisters by laser-cutting circles with an 11.43 cm diameter from 0.3175 cm-thick acrylic. The edges of the circles are laser-etched so they are inset into the tube. Attach to the tubes with plastic epoxy.
      2. Create an inset lid using black 0.635 cm and 0.3175 cm acrylic and UV-resistant black mesh. The inner ring has a diameter of 10.795 cm, while the outer ring has a diameter of 12.065 cm, with the black mesh stretched over the smaller ring. The mesh should be sewn on through laser-cut holes with waxed thread.
        ​NOTE: Canisters should always be placed on paper towels to avoid scratching the acrylic bottoms through which video is recorded.
    5. Construct acrylic containers for blood meal delivery (Figure 4A).
      1. Cut a ring with a 5.08 cm inner diameter and 6.6 cm outer diameter from 0.15875 cm clear acrylic. Attach this to another clear ring with an inner diameter of 5.842 cm and an outer diameter of 6.604 cm using acrylic glue.
      2. Place a metal mesh cylinder with a 13.97 cm height and a 19.05 cm diameter inside the black acrylic ring at the center of the platform.
      3. Inside the metal mesh, attach a coil of RGB LED lights spaced 3.81 cm from the exterior of the canister, to be controlled by an Arduino Uno board, which rests on the black acrylic platform (Figure 4C).
      4. Using a power meter, measure light intensity throughout the inside of the cylinder and adjust the location of the lights so the arena is illuminated between 3.5–6 µW/mm2. Outside the metal mesh, attach a ring of 850 nm infrared LEDs to the frame using zip ties, facing inward.
    6. Plug a USB-Cord into the Arduino Uno board and connect to the computer. Connect a USB cord from the camera to the computer. Turn off the internet connection.
    7. Place the opto-membrane structure in a dark incubator or environmental room at 26 °C and 80% relative humidity for the duration of the experiments.
  2. Preparing the mosquito canisters
    1. Prior to the experiment, wash mosquito canisters by spraying 70% ethanol and wiping gently with a soft sponge. Rinse the canisters with deionized water and air-dry overnight.
    2. Two days prior to the experiment, sort male and female mosquitoes under cold anesthesia and place 20 females into each canister. Provide access to ATR according to step 1.6.
    3. At this time, genotypes should be blinded to the experimenter. The order of the trials should be rotated between days to minimize differences in behavior outcomes due to experimental timing.
  3. Running trials
    1. To prepare the blood meal, place 119 mL bottles of water into a 45 °C heat bath, one for each trial. This volume and temperature keep the blood meal close to body temperature throughout the experiment.
    2. Prepare 5 mL aliquots of defibrinated sheep blood. Store at 4 °C and invert prior to use to prevent separation. Make the blood meal fresh prior to each experiment.
      CAUTION: Handle defibrinated sheep blood while wearing proper PPE to minimize pathogen exposure. Reusable materials that come into contact with blood should be soaked in 10% bleach and rinsed with DI water. Because bleach can cause severe skin and eye irritation, appropriate PPE should be worn throughout handling.
    3. Place the first blood aliquot into a 45 °C heat bath for at least 20 min prior to the trial. Thaw the ATP on ice. Stretch a 5.08 cm by 5.08 cm square of parafilm over the acrylic lid on the side with the larger, flat circle, creating a well for the blood meal (Figure 4A).
    4. Wipe the computer with 70% ethanol to remove human odors. Connect the computer to an external hard drive. Connect the camera and the Arduino. Plug the RGB and infrared lights into a power source.
    5. Open the SpinView software and the Arduino file for the appropriate stimulus. Create a folder on the external hard drive to contain the experiment data. Upload the “blue_only.io” Arduino file to provide a dim blue 471 nm light from the RGB LEDs to the arena (following Figure 4B).
    6. Without exposing mosquitoes to excess CO2 from breath, examine the canister for the trial and record any deaths prior to the experiment. Place the canister in the center and allow acclimation under dim blue light for 10 to 20 min prior to stimulus exposure.
      ​NOTE: The acclimation step and experiment are run under dim blue light because complete darkness inhibits mosquito flight.
    7. Enable video capture
      1. In the SpinView program, select the camera. In the Settings tab, adjust the exposure and gain as needed to maintain high contrast between the light-colored insects and the dark background (Figure 4D).
      2. Click the red record button on the top bar to adjust video capture settings. Set the file path by clicking Browse, then navigate to the external hard drive. Create a new folder with a descriptive name for each trial. Do not include spaces in the folder name.
      3. In the Save Options tab, select Capture 0 frames, which will allow the video to record until manually stopped. Under Recording Mode, select Buffered. In the Videos tab, set the following fields: Video Recording Type: MJPG, Video File Split Size: 1000 MB, and JPEG Compression Quality: 65.
    8. Immediately before stimulus exposure, pipette 500 µL of 20 mM ATP, for a final concentration of 2 mM ATP, into the prepared 5 mL blood aliquot. Invert the tube several times to mix.
    9. Pour the ATP-blood solution into the blood meal container. On top of the blood meal, place an inverted 119 mL bottle filled with 45 °C water to maintain the blood's temperature close to human body temperature throughout the experiment.
    10. In the SpinView software, select Start Recording. Without stimulating mosquitoes with breath, place the blood meal on top of the mesh lid of the canister so mosquitoes can pierce the parafilm membrane through the mesh.
    11. Upload the Arduino program for “1sON_10sOFF.io” (Figure 4B), or the desired stimulus. This will deliver 1 s of 624 nm red light every 10 s, until the “blue_only.io” file is manually uploaded. The wavelength of 624 nm is sufficient to optogenetically activate CsChrimson.
    12. Place the blood aliquot into the heat bath for the following trial. After the trial has run for 15 min, select Stop Recording. Upload the “blue_only.io” file to stop the red-light stimulus.
    13. Remove the blood meal holder and water bottle. Remove the canister and anesthetize the mosquitoes at 4 °C. To perform additional trials, return to section 4.3. Visually examine the anesthetized mosquitoes to determine and record the number that engorged on blood by looking for a blood-filled abdomen.
  4. Clean up
    1. Close all programs. Eject external hard drive and disconnect from the computer. Shut down the computer.
    2. Clean the mosquito canisters as previously described in step 4.2.1. Clean acrylic blood meal holders. Fill a half-sized polycarbonate pan 2/3 full of water, then add 100 mL of bleach.
    3. Add acrylic blood meal holders and plastic bottles to the pan. Cover with a lid and let soak for 10 min. Rinse bleach and remove parafilm. Scrub gently and rinse to remove any blood remnants.

5. Analysis of behavior

  1. Prepare the videos for analysis by compressing them into .mp4 files using ffmpeg (version 8.0.1). If the video files are split, merge them into a single video using any video editing software. Open the training videos in SLEAP (version 1.4.1) in grayscale.
    ​NOTE: If the video recording settings are appropriate, everything should be within frame, and moving mosquitoes should not be blurry. Other body part tracking software can be used. The SLEAP version used in this paper is 1.4.1. Refer to SLEAP documentation for up-to-date protocols.
  2. Create a skeleton that labels the relevant parts of the animals according to the behaviors being assessed (Figure 2D, Figure 3D, and Figure 4E).
  3. Using the “Random frame” option under the Suggested frame tab, randomly pick 10 frames and label each mosquito as an instance. If certain parts of the mosquito are obstructed (i.e., the proboscis tip is inserted into the mesh during probing), label the approximate location of the body part. Train a model on the labeled frames with these settings
    1. For a Top-Down multi-animal model (Used for the opto-thermocycler and blood blanket model): In the Training Pipeline tab, set Max Instances to 14. In the Centroid Model Configuration tab, set Crop Size to Auto, and set Plateau Min. Delta to 1e-04, set Plateau Patience to 10, set Stride to 8, and the Anchor Part as thorax.
    2. For a Bottom-Up multi-animal model (Used for the Opto-membrane feeder): In the Training Pipeline tab, set Max Instances to no max. In the Bottom-Up Model Configuration tab, set Crop Size to Auto, and set Plateau Min. Delta to 1e-04, set Plateau Patience to 10, and set Stride to 16.
  4. Predict on a set number of frames with these settings: In the Inference Pipeline tab, for the opto-thermocycler and blood blanket, set Max Instances to 14, Tracker to Simple, Max number of tracks to 14, and Connect Single Track Breaks to True.
  5. For the opto-membrane feeder, set No Max for Max Instances and Max number of tracks, Tracker to Simple, and Connect Single Track Breaks to True. Correct the predicted frames and re-train. Iterate through cycles of prediction, labeling, and training on the training data set.
    NOTE: Validation was performed by randomly spot-checking labeled frames and identifying consistent errors (e.g., false mosquito detections, duplicate predictions, incorrect orientation, or incorrect body-part assignments), focusing on frames containing these edge cases. Up to 150 labeled frames may be required, depending on error frequency. SLEAP also provides GUI-based evaluation metrics for additional model validation. Once visual inspection confirms sufficient accuracy and precision, apply the model to experimental videos and export results as .csv or .h5 for behavior analysis. Behavior classifiers can then be trained using pose estimation outputs as described previously8. Behaviors are defined as follows: walking, translational body movement with leg motion; flying, rapid translational movement with wings extended; grooming, repetitive leg movements over body parts (antennae, proboscis, wings, or legs); and probing, insertion of the proboscis through the mesh at the bottom of the container. Behavior classifiers should achieve >90% accuracy compared with human annotations.
  6. Perform statistical analysis by choosing non-parametric tests as the data are not typically normally distributed. Perform the appropriate post-hoc tests with correction for multiple comparisons. For engorgement data, experimental replicates are groups of mosquitoes, whereas for behavior data, replicates are individual mosquitoes.

Results

Three behavior assays were used to observe host-seeking and blood-feeding behaviors in mosquitoes, as shown in Figure 1A. Mosquitoes were reared under a blue-light-dark cycle (14:10 LD; Figure 1B) to synchronize circadian rhythms, allowing experiments to be run during wake periods. Two days before the experiments, mosquitoes were transferred to total darkness and provided with the opsin cofactor ATR. If mosquitoes are asynchronous or not fed ATR, they will show low activity in behavioral assays. The genotypes used for each assay were Gr3>CsChrimson and two genetic controls: the driver line Gr3-QF2 and the effector line QUAS-CsChrimson.

The opto-thermocycler assay (Section 2) allows assessment of mosquito arousal and probing behavior over long periods in response to heat and light stimuli. Heat delivery mimics the mosquito approaching a warm-blooded host, while the 5-second light delivery mimics a long exhalation of breath33. The use of a thermocycler allows for quick surface heating and cooling that is synchronized with optogenetic activation via a thermocouple and microcontroller (Figure 2A, B). The three mosquito genotypes were placed in acrylic plates with isolated wells with mesh bottoms in which they demonstrate walking, flying, and probing behavior (Figure 2C). The position of mosquito body parts was detected using pose-tracking algorithms, such as SLEAP34  and DeepLabCUT35 (Figure 2D, E). Behavior classification algorithms can be used starting from pose tracking (e.g., SimBA36 , A-SOiD37) or video (e.g., FERAL38) input. Figure 2E shows the SLEAP labels during flying, walking, and probing, essential host-seeking and feeding behaviors. These behaviors can be classified by the velocity of the thorax and the distance between the base and the tip of the proboscis. Here, the output of behavior tracking using APT and JAABA is shown39. Mosquitoes were exposed to either a heat increase, a light stimulus, or both heat and light stimuli simultaneously, then a break of 20 min before the next stimulus8. (Figure 2F). When exposed to heat, all three genotypes exhibit a short period of probing (Figure 2G). Importantly, when exposed to a 5-second light stimulus, the Gr3>CsChrimson mosquitoes become aroused or “activated,” showing increased walking and flying behavior that lasts up to 10–15 min. Mosquitoes also exhibit probing, which is defined as the repeated insertion of the proboscis into the mesh bottom of the plates. Data were plotted as individual mosquito behavior (Figure 2G) and proportion of all mosquitoes exhibiting each behavior over time (Figure 2H). The opto-thermocycler enables repeated, precise optogenetic and heat stimulation across multiple individuals, making it a useful tool for studying arousal and probing behavior.

The blood blanket assay (Section 3) is built upon the opto-thermocycler assay but provides a palatable artificial minimal blood meal so mosquito engorgement can be quantified (Figure 3A). A 5-second pulse of red light (~627 nm) was administered to optogenetically activate CO2 sensory neurons, and the thin layer of artificial blood meal was heated from 25 °C to ~35 °C (as measured at the surface by a thermocouple; Figure 3B). The three genotypes were each presented with an artificial blood meal in an aluminum feeding plate (Figure 3C). Gr3-QF2, QUAS-CsChrimson, and Gr3>CsChrimson mosquitoes (n = 10 trials per genetic group, 14 mosquitoes each) were each presented with an artificial blood meal in an aluminum feeding plate (Figure 3C). To measure engorgement of the mosquitoes over time throughout the experiment, SLEAP was used to track points on the body of the mosquito. This was done by creating a skeleton that included 5 nodes on the abdomen of the mosquito (Figure 3D). At the end of each trial, the number of engorged mosquitoes was manually counted. The control mosquito lines, Gr3-QF2 and QUAS-CsChrimson, demonstrated a similar average engorgement (Figure 3E, p = 0.401). In contrast, the Gr3>CsChrimson line demonstrated significantly higher rates of feeding, with an average engorgement rate of 53% (Figure 3E, p = 0.00078 versus Gr3-QF2 and p = 0.00097 versus QUAS-CsChrimson). Data were plotted as violin plots and boxplots (median, interquartile range, and range) and analyzed using the Kruskal-Wallis test and a Dunn Test with Holm correction at a significance threshold of p < 0.05. Individual animal engorgement over time was identified as the change in the abdominal area enclosed by 5 nodes (Figure 3F). These results show that the blood blanket assay can be used to assess the effects of optogenetic activation of CO2-responsive neuronal populations on the engorgement behavior of mosquitoes on an artificial blood meal.

The opto-membrane feeder assay (Section 4), where mosquitoes are presented with access to a warm blood meal in a cylindrical arena, allows for the study of mosquito attraction to a blood source and engorgement. To determine if optogenetic activation of CO2 sensory neurons is sufficient to cause attraction to and feeding on a blood source, Gr3>CsChrimson and control mosquitoes (n = 10 for each genotype, where n refers to the number of trials run, with 20 mosquitoes per canister per trial) were presented with access to a warm blood meal in the opto-membrane feeder under exposure to a repeated 1-second pulse of bright red light every 10 s for a duration of 15 min (Figure 4B). To measure attraction to the membrane feeder, a SLEAP model with three points was used to track the head, thorax, and abdomen of landed mosquitoes, counting those within the region of interest (Figure 4D, Figure 4E). Gr3>CsChrimson mosquitoes showed increased attraction to the blood meal region over time compared to the Gr3-QF2 control (Figure 4F, p = 0.027). The Gr3>CsChrimson mosquitoes did not show statistically higher attraction than the QUAS-CsChrimson (Figure 4F, p = 0.310). Significance groups were determined by the Kruskal-Wallis test, followed by the Dunn Test with Holm correction. The significance threshold was set at p < 0.05. Data were plotted as the average percent of the mosquitoes attracted to the blood meal, with the shaded region showing the standard error. The control genotypes Gr3-QF2 and QUAS-CsChrimson showed similar engorgement of 26% and 29%, respectively, manually scored as described (Figure 4G, p = 0.351). In contrast, Gr3>CsChrimson mosquitoes demonstrated a significantly higher engorgement rate than either control, at 61% on average (Figure 4G, p < 0.001 versus Gr3-QF2 and p = 0.002 versus QUAS-CsChrimson). Data were plotted as violin plots and boxplots (median, interquartile range, and range) and analyzed using the Kruskal-Wallis test and a Dunn Test with Holm correction at a significance threshold of p < 0.05. These results indicate that the opto-membrane feeder assay is a viable means to assess the effects of optogenetic activation of CO2 neurons on mosquito attraction and engorgement on a blood meal.

Mosquito behavior assay chart; methods: opto-thermocycler, blood blanket; stimuli: heat, CO2.
Figure 1: Assays and rearing for optogenetics in mosquitoes. (A) Schematic of mosquito host-seeking behaviors with characteristics of the behavior assays presented. The steps of host seeking measured by each behavior assay are indicated by gray bars. (B) Timeline of rearing of mosquitoes for optogenetics experiments. Figure 1 was created by the authors for this publication using Adobe Illustrator, and no license is required for its usage. Please click here to view a larger version of this figure.

Mosquito behavior, infrared thermocycler, LED-red light setup, graph analysis, stimulus response.
Figure 2: The opto-thermocycler quantifies arousal and search behaviors during optogenetic activation. (A) Schematic of opto-thermocycler assay consisting of a plate of mosquitoes on the block of a PCR thermocycler with overhead LEDs and a camera. (B) Circuit diagram for the microcontroller that detects the thermocycler temperature and controls the LEDs. Circuit diagram created with Fritzing under CC-BY-SA 3.0 license (https://github.com/fritzing/fritzing-app/blob/develop/README.md). This license also allows complete reuse of the image, provided proper credit is given. (C) Still image from the assay depicting the acrylic plate that holds up to 14 mosquitoes. The thermocouple is in the bottom-right well. (D) The mosquito skeleton used in the SLEAP model to estimate poses. Points labeled on the body include: the tip and base of the proboscis, used to detect probing behavior (shown in red); the front legs, where they connect to the thorax, where the femur joins the tibia, and where the tibia joins the tarsus (shown in blue); and the tip of the abdomen (shown in yellow). A top-down multi-animal model was used. Pose estimators and behavioral classifiers were used to quantify mosquito arousal and probing behavior. (E) Examples of mosquito walking, flying, and probing, with SLEAP predictions overlaid on top. Walking and flying are used as behavioral metrics for arousal. (F) Output of stimulus delivery. Yellow wavy lines indicate body temperature heat stimuli, and red LEDs indicate red light stimuli. (G) Ethograms of walking, flying, and probing behavior by Gr3-QF2, QUAS-CsChrimson, and Gr3>CsChrimson mosquitoes (n = 22–23 individual mosquitoes per genotype) following stimulus delivery. (H) Long-lasting behavior following optogenetic activation in Gr3>CsChrimson mosquitoes (n = 68–70 individual mosquitoes). Figure panels A–E were created by the authors for this publication using Adobe Illustrator, and no copyright license is required for the usage. Data in (F–H) are reproduced from Sorrells et al, 2022, with CC BY 4.0 license8. Please click here to view a larger version of this figure.

Mosquito feeding experiment setup, data analysis charts, artificial blood meal diagram, temperature graph.
Figure 3: The blood blanket assay quantifies engorgement to optogenetic activation. (A) Schematic of the blood blanket assay. The mosquito is contained in an acrylic cage where accessible artificial blood is presented beneath a mesh screen. (B) Readout of temperature and red light stimulus from the assay. (C) Schematic of the aluminum blood blanket feeding plate assembly. (D) The SLEAP skeleton used for pose estimation included the points described in Figure 2D in addition to four points along the sides of the abdomen where they meet the thorax and the edges of the thickest part of the abdomen. Engorgement was measured as the area within the five points along the abdomen perimeter using predictions from a top-down multi-animal model. (E) Violin plot depicting the mean percent engorged on the artificial blood meal per genotype (n = 10 trials of 14 mosquitoes per trial for each genotype). Significance was determined using the Kruskal-Wallis test, followed by a Dunn Test with Holm correction. Different letters indicate a significance of p < 0.05 between groups, where groups with the same letter are not significantly different from one another. (F) Engorgement over time in a Gr3>CsChrimson animal, as measured by abdominal area. The red trendline displays a 30-second sliding window over eight min of the trial, with x = 0 marking time when engorgement begins. All schematics of Figure 3 were created by the authors for this publication using Adobe Illustrator, and a copyright license was not required for their use. Please click here to view a larger version of this figure.

Mosquito optogenetics setup; RGB light control; Arduino; blood meal; attraction percent graph; results.
Figure 4: The opto-membrane feeder assay measures attraction and blood feeding. (A) Schematic of the opto-membrane feeder with individual components. (B) Timeline of experiment trials. The blue line represents dim blue light throughout the acclimation period and experiment. A segment of the experimental stimulus is shown, representing 1-second pulses of red light spaced 10 s apart. (C) Diagram depicting the wiring for the opto-membrane feeder. Circuit diagram created with Fritzing under CC-BY license by the authors. (D) Example frame from the opto-membrane feeder video. Mosquitoes in the red circle are counted to determine the percent attraction. (E) Examples of an unfed mosquito and an engorged mosquito, with the head, thorax, and abdomen labeled by the SLEAP tracking model for the opto-membrane feeder. A bottom-up multi-animal SLEAP model was used to measure attraction to the region of interest where the blood meal is placed. (F) Line graph depicting mean percent occupancy on the blood meal region (which is used as a behavioral metric for landing) over time, by genotype (n = 10 trials of 20 mosquitoes per genotype). Mean is plotted every 30 s. Shaded regions depict standard error. Significance was determined by the Kruskal-Wallis test, followed by the Dunn Test with Holm correction. Different letters represent significant groups; the threshold was defined as p < 0.05. (G) Violin plot depicting the mean percent engorged on a blood meal by genotype. In (F, G), significance was determined by the Kruskal-Wallis test, followed by the Dunn Test with Holm correction. Genotypes listed with the same letter are not significantly different from one another. The significance threshold is defined as p < 0.05. Figure 4 was created using Adobe Illustrator, its data were collected by the authors for this publication, and no copyright license was required for the usage. Please click here to view a larger version of this figure.

Discussion

This paper describes detailed protocols for assaying the sufficiency of neuronal cell types to drive steps in host attraction and biting. Biting behavior consists of a series of behavioral steps, from long-range detection to short-range attraction, to piercing the skin, and to engorgement. In principle, neuronal types could control specific actions or steps, bias behavior toward specific sequences, increase the duration of the whole behavior, or play many other roles. Thus, the multiple assays presented are necessary to dissect the neuronal function in this process.

Several common errors can result in poor mosquito responses in these assays. Low activity and feeding can be avoided by running experiments during mosquito peak activity, near the transition times between the light and dark cycles in their incubation chambers. Poor feeding rates in the blood blanket and opto-membrane feeder could be caused by a poor-quality artificial blood meal or incorrect temperature. Ensure that all reagents (NaHCO3, ATP, defibrinated sheep blood) are not expired and stored properly. It is recommended that all solutions be made on the day of the experiment to ensure maximum freshness. Ensure that the blood blanket plate directly contacts the surface of the PCR thermocycler block. The plates of the opto-thermocycler and blood blanket can build up condensation if there is residual moisture. Before experiments, place the plates on paper towels to absorb any moisture. When experiencing a poor optogenetic response, ensure the mosquitoes are at least one week old and are getting adequate time to feed on ATR, as it is a necessary chemical cofactor for the CsChrimson channels.

Optogenetic tools have been created that are sensitive to wavelengths from blue (~450 nm) to red (~650 nm), spanning the visible spectrum. Red light is particularly useful in insects, as they typically have low sensitivity and/or behavioral responses to these wavelengths40. Interestingly, mosquitoes show attraction during flight to patches of visual stimuli with long wavelengths present in human skin tones41. Thus, using red light could be a limitation that confounds the interpretation of optogenetic activation in mosquito host attraction assays. However, genetic control mosquitoes showed minimal responses to broadly applied red light, even at strong intensities (Figures 2–4). Genetic controls lacking either the driver or effector constructs and no-light controls are recommended to ensure behavioral effects are interpreted accurately. A control lacking the rhodopsin cofactor ATR is often used for optogenetics experiments in other species; however, the fish food used to feed mosquito larvae contains vitamin A, the precursor to ATR, so this is a less effective control in this system.

The optogenetic stimuli presented here are limited to continuous stimuli of adjustable light intensity. Optogenetic stimuli that closely recapitulate natural neuronal activity patterns are most likely to elicit the behavior of interest. Therefore, if neural recordings are available, the stimulus can be designed to match the observed intensity and temporal dynamics. When this is not known, the function of neurons can be screened using different stimuli and behavioral contexts, with the caveat that the results may not reflect the function of neurons in naturalistic behavior. Nevertheless, driving activity of neurons out of the natural range can be used to probe circuits or behavioral properties in unique ways, just as electrophysiology can be used to probe cellular properties. Currently, optogenetic tools have only been used in sensory neurons in the mosquito, an application that provides some advantages over the delivery of natural stimuli. Traditional strategies of measuring mosquito host seeking rely on the delivery of real CO2 gas. Using real CO2 requires air flow that has been filtered and humidified, which requires complex experimental setups to be able to precisely deliver and remove in short time increments9,10,42,43. Air flow and humidity are both cues for mosquito attraction and navigation, making optogenetic delivery appealing as a method to study the effects of olfactory stimuli separately from these other stimuli. Optogenetic tools have also previously been used to study the effect of sex and physiological state on behavior8, providing insight into these important questions in mosquito biology.

Apart from optogenetics, these protocols make use of standard assays that have been used to assess mosquito behavior and come with some limitations. For ease of cleanup, the blood blanket makes use of an artificial meal that contains only the tastants required for mosquitoes to engorge. This limitation could be overcome with the use of blood (as in the opto-membrane feeder) or biomimetic materials that recapitulate aspects of host tissue. Heat in the opto-thermocycler and blood blanket is delivered to walking or standing mosquitoes rather than flying mosquitoes and is limited by the ramp speed of the PCR thermocycler. Therefore, these assays are able to detect probing and engorgement behavior but not attraction. Another limitation of lab-based assays is that to understand the behavioral effect of specific cues, other cues that mosquitoes use are excluded. Increasingly complex lab assays and semi-field assays could be created to study additional aspects of the naturalistic environments in which mosquitoes feed on their hosts.

Another limitation of optogenetics in the mosquito is that few driver lines have been created beyond those for sensory neurons. Thus, its full potential will be realized when it is combined with drivers expressed in interneurons that cannot be directly stimulated by other approaches. Newly emerging resources, such as the Aedes aegypti Mosquito Cell Atlas, open the door to advances in mosquito neurogenetics44. Future work will be needed to develop such drivers for probing mosquito neural circuits. Similarly, additional optogenetic tools for inhibition are needed to determine whether a neuron type is necessary for a particular behavior. Although the assays presented in this paper are designed to measure host attraction and blood feeding, these approaches could be extended to study additional steps of host seeking and behaviors such as mating, oviposition, and nectar feeding. Optogenetic studies will provide new insight into how mosquito behavior is controlled at the neural circuit level and could lead to new approaches to prevent the spread of mosquito-borne diseases.

Disclosures

The authors have nothing to disclose.

Acknowledgements

We thank Kim Lezon-Geyda, Fernanda Medeiros Contini and Yaoyu Jiao for comments on the manuscript. This research was funded by NIH grant DP2AI177891. We thank the Neurotechnology Core funded by the Yale Kavli Institute for Neuroscience for technical advice. T.S. is an HHMI Freeman Hrabowski Scholar.

Materials

List of materials used in this article
NameCompanyCatalog NumberComments
1⁄8 Inch Black AcrylicMcMaster-Carr8650K32Opto-membrane feeder structure
1⁄8 Inch Clear AcrylicMcMaster-Carr8536K134Opto-membrane feeder canister
¼ Inch Black AcrylicMcMaster-Carr8560K259Opto-membrane feeder structure
AC AdapterJameco ReliaPro2197581Opto-membrane feeder structure
Acrylic GlueUnited States Plastic Corp 97562Opto-membrane feeder canister, blood blanket cage
Adenosine 5′-triphosphate disodium salt hydrateSigma AldrichA6419-1GArtificial blood meal
All Trans-RetinalSpectrum ChemicalR3041-1GMOptogenetic rearing
Aluminum BreadboardThorlabsMB12Opto-membrane feeder structure
Aluminum Sheet 0.03125 inMcMaster-Carr89015K11Blood blanket feeding plate
Aluminum Sheet 0.0625 inMcMaster-Carr89015K37Blood blanket feeding plate
Animal Part Tracker (APT)Branson lab at HHMI Janelia FarmN/AVersion 3.0, Opto-thermocycler reproduced data (Figure 2), Downloaded July 9, 2020
Arduino IDEArduinoN/AVersion 1.8.12, used for stimulus control for opto-thermocycler, blood blanket, & opto-membrane feeder
Arduino Uno Rev3 Microcontroller BoardArduinoA000066 Opto-membrane feeder programming
Black Fiberglass Window ScreenBreakthrough Premium Products IHLRS3684BLBlood blanket cage
Black TapeThorlabsT137-2.0Opto-thermocycler
Black Zip Ties Cable Ties And Morect242Opto-membrane feeder structure
Blackout DrapeSet Shop6520SFOpto-thermocycler and blood blanket
Blue LED Lights21LEDUSARB5730BLWPMosquito rearing
BottleSKS Bottle & Packaging0604-07Opto-membrane feeder blood heater
Bottom Feeder PelletsAPI5352983Feeding larvae
CameraBlackflyU3-13S2M-CS, FLIRecording video for assays
Click CounterULineH-7350Thinning larvae
ComputerAsus X1407QRunning Assays
Concentrated Bleach (8.25% Sodium Hypochlorite)VWR89501-620Opto-membrane feeder cleaning
CoolTermRoger Meier's FreewareN/AVersion 2.0.1, used for serial port recording for opto-thermocycler, blood blanket, & opto-membrane feeder
Defibrinated Sheep Blood Hemostat Laboratories DSB100Blood meal for opto-membrane feeder
Dome-topped hex 4-40 boltMcMaster-Carr92949A105Blood blanket feeding plate
EpoxyLoctite1363118Opto-thermocycler acrylic plate, Opto-membrane feeder structure
EthanolDecon Labs2716Sanitization
External Hard DriveLaCieLAC9000298Video Storage
ffmpegFFmpegN/AVersion 8.0.1, video pre-processing
Four Inch Cotton WickRichmond Dental201226Feeding adult mosquitoes
GitHubGitHub, IncN/ARepositories for code, part files, and SLEAP models for opto-thermocycler (v1.0.0, https://github.com/sorrellslab/opto-thermocycler/releases/tag/v1.0.0), blood blanket (v1.0.0, https://github.com/sorrellslab/bloodblanket/releases/tag/v1.0.0), and opto-membrane feeder (v1.0.0, https://github.com/sorrellslab/opto-membranefeeder/releases/tag/v1.0.0)
Gr3>CsChrimsonN/AN/ACreated for this study, described in step 1.2
Gr3-QF2 MosquitoesN/AN/AGenetic reagent, DOI: 10.1016/j.cell.2013.12.044
Incubator with Opaque DoorCaron7340-25Mosquito rearing
Infrared LED LightAdafruit387Opto-thermocycler and blood blanket
Infrared LED StripWaveform Lighting7031.85Opto-thermocycler, blood blanket, and opto-membrane structure
Infrared Pass FilterEdmund Optics65-796Recording video for assays 
Insect CageBugdormDP1000Adult mosquito enclosure
Janelia Automatic Animal Behavior Annotator (JAABA)Branson lab at HHMI Janelia FarmN/AOpto-thermocycler reproduced data (Figure 2); Version 0.5.0
Large Soup CupWebstaurantStore50016SOUPPLAMosquito rearing
LED Heat SinkLuxeon StarN25-15BOpto-thermocycler and blood blanket
LED LensLuxeon Star10209Opto-thermocycler and blood blanket
LED Lens HolderLuxeon Star10235Opto-thermocycler and blood blanket
LED Lens TapeLuxeon StarLT-01Opto-thermocycler and blood blanket
LED LightLuxeon StarSP-01-D9Red light stimulus for Opto-thermocycler and blood blanket
LED Thermal TapeLuxeon StarLXT-S-12Opto-thermocycler and blood blanket
Optical Construction RailsThorlabsXE25L12, XE25L24, XE25L09, XE25T4, RA90, TRA6, TR12Opto-thermocycler and blood blanket structure
Optical Post ThorlabsTR12Opto-membrane feeder structure
Paper TowelsPacific Blue27385Canister cleaning
ParafilmAmcorPM-996Blood blanket membrane, opto-membrane meal container
PCR thermocyclerEppendorf2231001196Opto-thermocycler and blood blanket
Plastic Portion CupChoice999P2CFeeding adult mosquitoes
Plastic Portion Cup LidChoice999PL2Feeding adult mosquitoes
Polycarbonate PanCambro21410CWCHCLRearing larvae 
Polycarbonate Pan Half SizedCambro24CW135 Opto-membrane meal cleaning
Polycarbonate Pan Half Sized LidCambro20CWCH135 Opto-membrane meal cleaning
Polycarbonate Pan LidCambro10CWCH135Rearing larvae
Polycarbonate TubeMcMaster-Carr8585K56Opto-membrane feeder canister 
Power MeterCoherent1299161All assays
QUAS-CsChrimson MosquitoesN/AN/AGenetic reagent, DOI: 10.7554/eLife.76663
RGB LightsDigikey 289-1189-NDOpto-membrane feeder stimulus delivery
Right Angle ClampThorlabsRA90Opto-membrane feeder structure
Scotch Magic Tape3M104Opto-thermocycler and blood blanket acrylic lid
Silicone Gasket, 0.5 mmMcMaster-Carr1460N22Blood blanket feeding plate
SLEAPTalmo LabN/AVersion 1.4.1a2, pose tracking of mosquitoes in each of the assays
Sodium Bicarbonate, 1M buffer soln., pH 8.0Thermo scientificJ62495.APArtificial blood meal
Sodium Chloride solutionSigma-AldrichS5150-1L 100 3532552Artificial blood meal
SpinViewTeledyne Vision SolutionsN/AVersion1.27.0.48, video recording software for all three assays
Spray BottleUlineS-11686Sanitization
SucroseSigma AldrichS0389-500GFeeding adult mosquitoes
Thermocouple (Type T)Harold G Schaevitz Industries LLCCPTC-120-X-NOpto-thermocycler and blood blanket
Thermocouple amplifierAdafruitMAX31856Opto-thermocycler and blood blanket
TimerEnduraLightRAC-MTControlling light cycle for mosquito rearing
Transfer Pipette Thermo ScientificSAM-225Thinning larvae
Trash BagUlineS-23040Insect transportation
UV Resistant MeshMcMaster-Carr 87655K13Opto-membrane feeder canister
WaterSigma AldrichW4502-1LArtificial blood meal for blood blanket
Waxed ThreadMcMaster-Carr6356K91Opto-membrane canister lid
Wire MeshMcMaster-Carr9322T65Opto-membrane feeder structure

References

  1. UNICEF, UNDP, World Bank, WHO Special Program for Research and Training in Tropical Diseases. Global vector control response 2017-2030. Geneva: World Health Organization; 2017.
  2. Knight KL, Stone A. A catalog of the mosquitoes of the world (Diptera: Culicidae). Lanham (MD): Entomological Society of America; 1959.
  3. Raji JI, DeGennaro M. Genetic analysis of mosquito detection of humans. Curr Opin Insect Sci. 2017;20:34-38.
  4. Coutinho-Abreu IV, Riffell JA, Akbari OS. Human attractive cues and mosquito host-seeking behavior. Trends Parasitol. 2022;38(3):246-264.
  5. McMeniman CJ, Corfas RA, Matthews BJ, Ritchie SA, Vosshall LB. Multimodal integration of carbon dioxide and other sensory cues drives mosquito attraction to humans. Cell. 2014;156(5):1060-1071.
  6. Gillies MT. The role of carbon dioxide in host-finding by mosquitoes (Diptera: Culicidae): a review. Bull Entomol Res. 1980;70(4):525-532.
  7. Geier M, Bosch O, Boeckh J. Influence of odor plume structure on upwind flight of mosquitoes towards hosts. J Exp Biol. 1999;202(Pt 12):1639-1648.
  8. Sorrells TR, Pandey A, Rosas-Villegas A, Vosshall LB. A persistent behavioral state enables sustained predation of humans by mosquitoes. eLife. 2022;11:e76663.
  9. Dekker T, Geier M, Cardé RT. Carbon dioxide instantly sensitizes female yellow fever mosquitoes to human skin odors. J Exp Biol. 2005;208(15):2963-2972.
  10. van Breugel F, Riffell J, Fairhall A, Dickinson MH. Mosquitoes use vision to associate odor plumes with thermal targets. Curr Biol. 2015;25(16):2123-2129.
  11. Liu MZ, Vosshall LB. General visual and contingent thermal cues interact to elicit attraction in female Aedes aegypti. mosquitoes. Curr Biol. 2019;29(13):2250-2257.e4.
  12. Chandel A, DeBeaubien NA, Ganguly A, et al. Thermal infrared directs host-seeking behavior in Aedes aegypti. mosquitoes. Nature. 2024;633(8030):615-623.
  13. Tang R, Busby R, Laursen WJ, Keane GT, Garrity PA. Functional dissection of mosquito humidity sensing reveals distinct Dry and Moist Cell contributions to blood feeding and oviposition. Proc Natl Acad Sci U S A. 2024;121(35):e2407394121.
  14. Burgess L. Probing behavior of Aedes aegypti. (L.) in response to heat and moisture. Nature. 1959;184(4703):1968-1969.
  15. Hol FJ, Lambrechts L, Prakash M. BiteOscope, an open platform to study mosquito biting behavior. eLife. 2020;9:e56829.
  16. Martin-Martin I, Williams AE, Calvo E. Determination of mosquito probing and feeding time to evaluate mosquito blood feeding. Cold Spring Harb Protoc. 2023;2023(6):pdb.top107659.
  17. Boyd MF, Stratman-Thomas WK. Studies on benign tertian malaria. VII. Some observations on inoculation and onset. Am J Epidemiol. 1934;20(2):488-495.
  18. Swanson MM, Poodry CA. The shibire(ts) mutant of Drosophila: a probe for the study of embryonic development. Dev Biol. 1981;84(2):465-470.
  19. Sweeney ST, Broadie K, Keane J, Niemann H, O'Kane CJ. Targeted expression of tetanus toxin light chain in Drosophila specifically eliminates synaptic transmission and causes behavioral defects. Neuron. 1995;14(2):341-351.
  20. Baines RA, Uhler JP, Thompson A, Sweeney ST, Bate M. Altered electrical properties in Drosophila neurons developing without synaptic transmission. J Neurosci. 2001;21(5):1523-1531.
  21. Hamada FN, Rosenzweig M, Kang K, et al. An internal thermal sensor controlling temperature preference in Drosophila. Nature. 2008;454(7201):217-220.
  22. Corfas RA, Vosshall LB. The cation channel TRPA1 tunes mosquito thermotaxis to host temperatures. eLife. 2015;4:e11750.
  23. Greppi C, Laursen WJ, Budelli G, et al. Mosquito heat-seeking is driven by an ancestral cooling receptor. Science. 2020;367(6478):681-684.
  24. Allen WE, DeNardo LA, Chen MZ, et al. Thirst-associated preoptic neurons encode an aversive motivational drive. Science. 2017;357(6356):1149-1155.
  25. Inagaki HK, Jung Y, Hoopfer ED, et al. Optogenetic control of Drosophila courtship using a red-shifted channelrhodopsin reveals experience-dependent influences. Nat Methods. 2014;11(3):325-332.
  26. Sahel JA, Boulanger-Scemama E, Pagot C, et al. Partial recovery of visual function in a blind patient after optogenetic therapy. Nat Med. 2021;27(7):1223-1229.
  27. Klapoetke NC, Murata Y, Kim SS, et al. Independent optical excitation of distinct neural populations. Nat Methods. 2014;11(3):338-346.
  28. Deisseroth K. Optogenetics. Nat Methods. 2011;8(1):26-29.
  29. Potter CJ, Tasic B, Russler EV, Liang L, Luo L. The Q system: a repressible binary system for transgene expression, lineage tracing, and mosaic analysis. Cell. 2010;141(3):536-548.
  30. Herre M, Goldman OV, Lu TC, et al. Non-canonical odor coding in the mosquito. Cell. 2022;185(17):3104-3123.e28.
  31. Wohl MP, McMeniman CJ. Batch rearing Aedes aegypti. Cold Spring Harb Protoc. 2023;2023(3):pdb.prot108017.
  32. American Committee of Medical Entomology, American Society of Tropical Medicine and Hygiene. Arthropod containment guidelines, version 3.2. Vector Borne Zoonotic Dis. 2019;19(3):152-173.
  33. Pleil JD, Wallace MAG, Davis MD, Matty CM. The physics of human breathing: flow, timing, volume, and pressure parameters for normal, on-demand, and ventilator respiration. J Breath Res. 2021;15(4):042002.
  34. Pereira TD, Tabris N, Matsliah A, et al. SLEAP: a deep learning system for multi-animal pose tracking. Nat Methods. 2022;19(4):486-495.
  35. Mathis A, Mamidanna P, Cury KM, et al. DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. Nat Neurosci. 2018;21(9):1281-1289.
  36. Goodwin NL, Choong JJ, Hwang S, et al. Simple Behavioral Analysis (SimBA) as a platform for explainable machine learning in behavioral neuroscience. Nat Neurosci. 2024;27(7):1411-1424.
  37. Tillmann JF, Hsu AI, Schwarz MK, Yttri EA. A-SOiD, an active-learning platform for expert-guided, data-efficient discovery of behavior. Nat Methods. 2024;21(4):703-711.
  38. Skovorodnikov P, Zhao J, Buck F, et al. FERAL: a video-understanding system for direct video-to-behavior mapping. bioRxiv. 2025. doi:10.1101/2025.11.16.688666.
  39. Kabra M, Robie AA, Rivera-Alba M, Branson S, Branson K. JAABA: interactive machine learning for automatic annotation of animal behavior. Nat Methods. 2013;10(1):64-67.
  40. Autrum H, Bennett MF, Diehn B, et al. Comparative physiology and evolution of vision in invertebrates. A: Invertebrate photoreceptors. Berlin: Springer; 1979.
  41. Alberto DAS, Rusch C, Zhan Y, Straw AD, Montell C, Riffell JA. The olfactory gating of visual preferences to human skin and visible spectra in mosquitoes. Nat Commun. 2022;13(1):555.
  42. Dekker T, Cardé RT. Moment-to-moment flight maneuvers of the female yellow fever mosquito (Aedes aegypti. L.) in response to plumes of carbon dioxide and human skin odor. J Exp Biol. 2011;214(Pt 20):3480-3494.
  43. Demir M, Kadakia N, Anderson HD, Clark DA, Emonet T. Walking Drosophila navigate complex plumes using stochastic decisions biased by the timing of odor encounters. eLife. 2020;9:e57524.
  44. Goldman OV, DeFoe AE, Qi Y, et al. A single-nucleus transcriptomic atlas of the adult Aedes aegypti. mosquito. Cell. 2025;188(25):7267-7290.e26.

Reprints and Permissions

Tags

Mosquito BehaviorNeural CircuitsOptogenetic ActivationBehavioral AssaysHost SeekingBlood FeedingMachine VisionSensory Stimuli

This article has been published

Video Coming Soon

Contact Us
Recommend to Library

Privacy

Terms of Use

Policies

Copyright © 2026 MyJoVE Corporation. All rights reserved.