Mosquitoes have remarkably refined powers of smell. The insects that spread malaria across sub-­Sa­har­an Africa come exquisitely equipped to find human blood. They home in on the scent of human breath and sweat and swiftly insert their needlelike mouthparts into the target’s skin. As they dine, their saliva transmits the malaria parasite into the wound. With a simple bite, they can ultimately take a life.

Other mosquitoes prefer different species—say, cattle or birds. Some, it seems, even favor selected individuals within the target group; certain people at a summer barbeque will be attacked relentlessly, yet others will remain unbitten. And some mosquitoes can identify their victims from more than 165 feet.

If investigators could better understand how the mosquito olfactory system works—how it manages to detect exactly the suite of volatile chemicals unique to its favored source of blood—they should be able to devise new, more effective ways of masking those scents or “jamming” the insects’ olfactory “radar” to prevent bites. In the developed world, such bites are often just a nuisance, but in Africa and elsewhere they cause nearly a million deaths a year from malaria alone.

We are among the many researchers determined to fight malaria’s spread. To our delight, we have recently made exciting strides in deciphering how the mosquito Anopheles gambiae, the main carrier of malaria parasites, detects the scent of its human victims. The findings are now pointing to ideas for repellents and traps that could complement other defensive measures such as bed nets and, one day, an effective vaccine.

Genes for Odors
To investigate how malaria-causing mosquitoes detect their human prey, we began with a different insect, the fruit fly Drosophila melanogaster. Unlike mosquitoes, fruit flies breed quickly and are easy to maintain in a laboratory, and their genes can be readily manipulated. D. melanogaster has become a lab workhorse, so we use it to reveal the basic cellular and molecular mechanisms of insect olfaction, knowledge we can then apply in more difficult experiments with less tractable mosquitoes.

Fruit flies, like mosquitoes, detect odors with antennae and maxillary palpi, organs that protrude from the head and act as a nose. Tiny bristles that cover these protrusions encase the ends of excitable nerve cells dedicated to smell. Odorant molecules slip through pores in the bristles to reach odor-detecting molecules, or receptors, inside. When receptors bind to odor molecules, an electrical signal travels down the nerve cell, or neuron, to the insect’s brain, indicating that the odor is present.

For years we and others had tried unsuccessfully to find the genes for insect odorant receptors, hoping to learn exactly how the creatures distinguish among the countless odorants in the environment. Breakthroughs finally began coming in 1999. Researchers on our team at Yale University and elsewhere discovered the first genes coding for the receptors. Over time we found 60 odorant receptor genes in the fruit fly. Knowing the sequence of their DNA code opened the door to figuring out how the receptors work. We also found that the genetics of the olfactory systems of the fruit fly and the mosquito are similar, so studying the fly would help us understand olfaction in the mosquito.

A key insight came from a genetic mutant of D. melanogaster that arrived in our lab serendipitously. In November 2001 one of us (Carlson) gave a seminar at Brandeis University near Boston. The seminar was about Or22a, the first fruit fly odorant receptor gene that our lab had discovered. After the talk, an assistant professor at Brandeis came up to the podium and said that he happened to have a mutant D. melanogaster strain that was missing the gene encoding this odorant receptor. He asked if the mutant might be useful. It took Carlson about a millisecond to respond, “Yes!” The next day Carlson drove a small vial of the mutant flies down Interstate 91 to our Yale facility in New Haven, Conn.

A major goal was to determine which fruit fly receptors responded to which odorants. A single neuron has thousands of receptors, but they are identical; each type binds only a small subset of odor molecules. Different neurons have different types of receptors that bind to other subsets. Because the mutant fruit flies were missing one particular odorant receptor gene, we hypothesized that they would harbor a kind of receptorless, or “empty,” neuron.

Sure enough, they did. Applying sophisticated genetic techniques developed to study D. melanogaster, we inserted a fruit fly receptor gene into this neuron, which then produced the encoded receptor molecules. For each receptor, we could then determine which odorants activated it. By systematically fitting each D. melanogaster odorant receptor into an empty neuron, one at a time, and exposing the neuron to a variety of odoriferous compounds, we could learn which of those chemicals generated a response for each of the insect’s many receptors.

Over the next three years Elissa Hallem, then a graduate student at Yale, did just that. She found that individual receptors responded to a limited subset of odorants and that individual odorants activate subsets of receptors. Similar results have been observed in the mammalian olfactory system. Thus, animals, from fruit flies to humans, detect scents in the same way: different odors activate different combinations of receptors. This strategy helps to explain how animals, including mosquitoes, can discriminate among the vast number of smells found in nature without having to possess a receptor dedicated to every single variety.

A Fly That Sniffs Like a Mosquito
Having characterized the fruit fly’s odorant receptor genes, we wanted to try to insert receptor genes from the malaria-carrying mosquito into the fly’s empty neuron. In collaboration with Laurence J. Zwiebel of Vanderbilt University, Hugh M. Robertson of the University of Illinois at Urbana-Champaign and their colleagues, we had identified a family of 79 genes likely to be odorant receptor genes in A. gambiae by searching for sequences of DNA similar to those in the fruit fly’s receptor genes. Transplanting any one of the genes into a fruit fly’s empty neuron could theoretically produce a mosquito odorant receptor in the fly. But the experiment could easily fail. The two insect species are separated by 250 million years of evolution. We had no idea if a mosquito receptor gene would function in a fruit fly neuron.

Our experimental system is attached to a loudspeaker, so if an olfactory neuron fires, our electrode senses it and the speaker generates a staccato series of clicks. When we tested a series of odorants on the first empty fly neuron that had been fitted with a mosquito gene, the loudspeaker remained disappointingly silent. We suspected that the mosquito receptor might not work in the fruit fly neuron. But Hallem continued testing samples. When she reached a compound called 4-methylphenol, the loudspeaker began screaming, and we were equally excited. We later learned that 4-methylphenol, which smells a bit like used gym socks, is a component of human sweat. We had found a way to decode which odorants elicit a response from which mosquito receptors, information that could help us understand how mosquitoes locate their human prey and how we might interfere with that process.

With this encouraging result in hand, we read widely about human odorants and selected 110 compounds to test, including many that are components of human sweat. We included odorants with diverse molecular structures, creating a broad sample. One by one, we began transplanting each of the 79 possible A. gambiae receptor genes into empty neurons. Fifty of the receptor molecules proved functional in our setup. We then began testing the panel of 110 odorants against the 50 functional receptors, producing 5,500 odorant receptor combinations. The extensive sampling required many long days and nights.

From this data set, we identified several receptors that responded strongly to only one or a very few compounds. We were interested in these “narrowly tuned” receptors. We reasoned that if a mosquito needed to detect a particular compound with a high degree of sensitivity and specificity—notably one that signals a source of blood—the mosquito might use a dedicated receptor. Indeed, we found that most of the narrowly tuned receptors responded to components of human sweat. For example, the first mosquito receptor Hallem had tested in the empty neuron—the receptor that responded so strongly to 4-methylphenol—turned out to be narrowly tuned. Out of the 110 compounds, only a few others excited that receptor as strongly. Another receptor was narrowly tuned to 1-octen-3-ol, common in human and animal odor. It strongly attracts several mosquito species, including Culex pipiens, the one that is commonly found in U.S. backyards and that can carry West Nile virus. Some commercial traps sold to lure mosquitoes away from people in backyards emit 1-octen-3-ol.

Jam the Nerves, Stop the Insects
Our results could speed development of better mosquito repellents and traps. One standard method for testing compounds involves putting substances into traps in the field to see whether they attract mosquitoes. But because the process is slow, only a limited number of chemicals can be tested. Classical lab experiments also have drawbacks. In many cases, human volunteers allow an arm to be coated with a compound, and then they insert the arm into a clear box containing dozens of mosquitoes; chemicals that deter the insects may later be pursued as repellents. In our approach, we can rapidly examine many more chemicals, making discovery of new, more effective lures or repellents much more likely—and without human subjects.

Vanderbilt’s Zwiebel, for instance, is using A. gambiae odorant receptors grown in cells in small lab dishes. Robots expose the cells to thousands of compounds in just a few hours. So far Zwiebel has screened more than 200,000 compounds, and more than 400 of them have activated or inhibited the odorant receptors. These compounds will be analyzed further in experiments, and the best of them will advance to field tests.

The lab approach also allows us to screen for compounds that act as “superactivators”—ones that jam olfactory neurons by overexciting them to the point that their signaling either shuts down or confuses the mosquito’s brain. “Confusant” compounds could be released near the huts in which villagers in sub-Saharan Africa sleep, preventing malaria-carrying mosquitoes from finding the inhabitants. Lab screening could also identify compounds that inhibit the narrowly tuned receptors, blocking the insect’s ability to sense a target. These masking agents, too, could be released at huts or used in repellents applied to the skin, to prevent mosquitoes from realizing that they were near a source of blood. Compounds that mosquitoes find offensive might also be identified for repellents. Our collaborators at Wageningen University in the Netherlands are experimenting with A. gambiae mosquitoes to determine whether blends of some of the compounds we have identified may be useful in these ways. Our colleagues have already found some powerful combinations.

Historically many methods of insect control, such as the widespread spraying of the insecticide DDT, have harmed animals and perhaps people. Olfactory-based control methods can be far less damaging. An olfactory trap requires only a small amount of attractant because mosquitoes are so sensitive to these cues. Attractive compounds that are commonly found in human sweat and breath should be nontoxic in low doses, too. If poisons were also used in these traps, they would be contained instead of distributed broadly. Moreover, olfactory-based insect control could be much more precise than that based on insecticides. Comparisons of our data from mosquitoes and fruit flies show that most of the narrowly tuned receptors of A. gambiae respond to compounds found in human sweat, whereas the narrowly tuned receptors of D. melanogaster respond to volatiles emitted by fruit. Blends of attractants can be chosen that preferentially lure the target insect, leaving a much lighter mark on the environment. Overall, olfactory-based insect control should be much less damaging to the natural world and more politically acceptable than the blanket spraying of poisons. And if a cocktail of effective compounds can be used instead of a single compound, resistance is less likely to arise in mosquito populations.

For the agents discovered by our methods to be useful in poverty-stricken nations, they would have to be packaged inexpensively. Traps that release carbon dioxide from compressed gas tanks—used widely in rich countries—are impractical in rural areas of the developing world. Attractant and repellent compounds must also be chemically stable in blistering tropical heat. Whether those demands can be met remains to be seen.

A multifaceted approach is needed to eradicate malaria. Bed nets and improved drugs will play a major role. Researchers are steadfastly trying to develop an effective vaccine. Still, the need for additional tools in the antimalaria armament is pressing. Precisely manipulating olfactory-guided mosquito behavior could be a big step. In the struggle against a disease that affects hundreds of millions of people every year, even a small contribution could make a large difference in the lives of many.