Following a trail of smell, a male fruit fly zeroes in on a banana peel. For the fly, the banana is not only a fantastic food source, but also fertile ground for finding mates. Sure enough, a virgin female is already feasting on the banana peel. He approaches her, taps her with his forelegs, and flutters his wings to sound a staccato love song, all in the hopes of securing her as a mate. But there is more to this scene than meets the eye or ear. The success of this courtship ritual critically depends on a single substance: an organic ester, 11-cis-vaccenyl acetate (cVA). CVA is found on the male’s cuticle, or exoskeleton, and in his ejaculatory bulb, a structure similar in anatomy and function to the human prostate. To mature female fruit flies, cVA is an aphrodisiac that induces their receptivity to an approaching...

CVA is a pheromone, classically defined as a substance secreted by an animal that elicits a specific reaction in other members of the species. Although best understood in insects, pheromones are also known to play important roles in mammalian behavior and physiology, from territorial marking in mice to the induction of mating in elephants.

Researchers have made rapid progress in our understand­ing of the neural circuits in the fly brain. Do the same principles also govern the processing of pheromone information in the mammalian brain?

The powerful effect a pheromone can exert on an animal captures the popular imagination. The idea of irresistibility is so ingrained in our psyche that the mention of pheromones immediately conjures up images of love potions, whiffs of which instantly make the wearer more sexually attractive. Indeed, Googling “human pheromone” will lead you to companies trying to sell you one of these “scientifically proven” attractants. (See “Something Smells Funny.”) While such marketing has deepened the sensual mystique surrounding pheromones, so far there is no substantial evidence that such perfumes can induce mate-seeking behavior in men or women. However, decades of research have revealed a fascinatingly wide range of pheromones across the animal kingdom that are not limited to affecting reproductive behaviors. And in the last 10 years or so, scientists have unveiled some of the neural mechanisms of pheromone processing in the brains of both fruit flies and mice, identifying clues to how these compounds work at the molecular, neural, and behavioral levels.

Lessons from the fly

Thanks in part to the fantastic genetic tools developed in the last decade, research on the fruit fly, Drosophila melanogaster, has uncovered many details of pheromone pathways, from the antennae to the brain. CVA, the only volatile fly pheromone so far identified, is detected in the antennae by olfactory receptor neurons (ORNs) that express a G protein-coupled receptor called OR67d. These neurons project their axons into bulb-shape structures called glomeruli in the antennal lobe of the brain, where olfactory information is initially processed. Each glomerulus is innervated by a distinct set of projection neurons (PNs) that then transmit the information into deeper brain regions. (See “Odor Encoder.”)

How does cVA, a single compound emitted by male flies, trigger behaviors that differ so widely between males and females? The answer lies within the neural pathways that parse the information into different neural circuitries in male versus female brains. The sexually dimorphic circuitry begins in the antennal lobe, where the glomeruli that receive input from OR67d neurons are larger in the male brain than in the female brain. From the antennal lobe, the PNs that receive input from glomeruli project their axons into two other brain areas: the mushroom body, where information about odors is associated with other sensory inputs, forming the basis of learned behaviors; and the protocerebrum, which is similar to the hypothalamus in the mammalian brain and is the origin of stereotypic behaviors, such as courtship and mating, in flies. Whereas the projection patterns from the PNs are similar in the male and female brains when they reach the mushroom body, they differ when the axons reach the protocerebrum. From the protocerebrum, the male and female circuitries further diverge and connect to different downstream neurons. The divergent patterns in the two sexes are thought to underlie the sexually dimorphic responses to cVA.

Thanks in part to the fantastic genetic tools developed in the last decade, research on the fruit fly, Drosophila melanogaster, has discovered many details of the pheromone pathway, from the antennae to the brain.

The mapping of this neural circuitry in the fruit fly brain reveals two important features of pheromone detection. First, information about a pheromone passes through a highly specific neural pathway, which is often referred to as a labeled line. The labeled line connects the sensory input, in this case a single chemical compound, to the behavioral output. Second, the labeled line differs between the sexes, allowing a single compound, cVA, to serve as an attractive sex pheromone for females and an antiaphrodisiac for males.

The short generation time of fruit flies and the availability of new genetic tools have enabled rapid progress in our understanding of the neural circuits in the fly brain. An immediate question, then, is whether the same principles also govern the processing of pheromone information in the mammalian brain.

Mammalian pheromones

Researchers have long recognized the roles that pheromones play in many mammals. In some species, such as cats and ungulates, a particular sniffing behavior has evolved that is believed to facilitate the exposure of sense organs to pheromones. The behavior, known as the flehmen response, is characterized by the curling of the upper lip and the exposure of the front teeth. In elephants, the flehmen response is characterized by the repeated tucking of the trunk into the open mouth. Female elephants release the urinary pheromone (Z)-7-dodecen-1-yl acetate, which induces strong flehmen responses in males. In pigs, sows in estrus respond to 3α-androstenol and 5α-androstenone, two steroid pheromones enriched in the saliva of male pigs, by exhibiting the “standing response,” a rigid, motionless pose that signals reproductive readiness.

But such behavioral responses to pheromones, called releasing effects, are relatively rare in mammals. More common are so-called priming effects, in which pheromone blends cause a change in the signal receiver’s physiology that does not manifest itself as an immediate behavioral response. A large body of literature in the last century, mostly rodent studies, has identified the effects of mammalian pheromones on reproductive physiology, for example.1 Rodent urine is rich in chemicals, including pheromones, that function in intraspecies communication. For example, the urine of a male mouse can accelerate the onset of puberty in young females. On the other hand, females housed in groups have delayed estrus onset and prolonged reproductive cycles. Interestingly, these latter effects can be reversed by the presence of a sexually mature male or his urine. Moreover, it was discovered that the presence of a strange male or his urine may cause a newly impregnated female to abort implanted embryos. Territorial marking and intermale aggression, as well as maternal aggression displayed by new mothers, are also found to be induced by urine and urinary compounds.

These numerous pheromone-induced responses in mammals contrast with the relatively simple behaviors observed in insects, but there is yet another layer of complexity built into mammalian pheromone communication. The pheromone-elicited responses in mammals often depend upon the context of pheromone exposure and the experience of the animals. For example, a female mouse can switch from being attracted to a male following pheromone exposure to being aggressive, depending on whether she is in estrus or has just given birth. On the other hand, a sexually naive male kills young pups he encounters, but if a male has mated and then cohabited with a pregnant female in the past 3 weeks, he will instead exhibit paternal behaviors, such as helping to return wayward pups to the nest.2

Complicating matters even further is the fact that researchers don’t have a good understanding of what the mammalian pheromone compounds are. In mammals, pheromones are found in urine and exocrine gland secretions including sweat, tears, and secretions from the preputial glands near the genital area. Chemical isolation experiments have identified a number of molecules that may serve as mammalian pheromones. Some chemicals have been shown to alter sexual maturation, mating behavior, and aggression, for example, when presented in conjunction with urine, but these do not directly trigger behaviors or physiological changes on their own. In contrast to insect systems, individual chemical compounds rarely evoke behavioral responses or endocrine changes in mammals, and this has made it difficult to pinpoint whether a compound is serving as a pheromone or as a modifier of pheromone responses.

Furthermore, the neural mechanisms the mammalian brain uses to process complex olfactory chemical cues and to generate stereotypic responses remain largely unknown. But recent studies in mice, combining molecular biology, genetics, imaging, and behavioral studies, have started to shed light on this problem. In addition to sex identification, it is now thought that mammalian pheromones convey information about an animal’s social status, reproductive status, genetic background, and individual identity. The specific compounds that serve as pheromones, their cognate receptors, and the neural circuits that process the information remain subjects of intense investigation.

The vomeronasal system

SECONDARY OLFACTION: The mouse vomeronasal organ, shown here in cross section, is located just above the roof of the mouth and comprises sensory epithelia (blue crescents) that convert the chemical signal of pheromones into electrical nerve pulses that carry messages to the brain.COURTESY OF WELLCOME TRUST SANGER INSTITUTEInstead of a single or a few receptors that detect a single pheromone, as is the case with insects, many vertebrates have evolved a dedicated organ to detect a much larger variety of chemical substances that may serve as pheromones. While pheromones may be detected by other sensory organs, including the main olfactory system and the taste buds, a major contribution to pheromone detection in vertebrates comes from the vomeronasal organ (VNO). Discovered in nonhuman mammals by Ludwig Jacobson in 1813, the VNO is a tubular structure in the nasal cavity embedded above the palate, or roof of the mouth. It is found in most amphibians, reptiles, and nonprimate mammals, but is absent in birds and most primates. The VNO opens to the base of the nasal cavity, and in carnivores and ungulates, connects with the oral cavity through a passage known as the nasopalatine duct. The sensory epithelia of the VNO form curved layers along the septum that divides the two sides of the nose, with large blood vessels on either side. (See image at right.) The sensory epithelia surround the VNO lumen, which is filled with fluid from the vomeronasal glands. This is where pheromone molecules interact with their neuronal receptors.

The intricate anatomical structure of the VNO has long fascinated scientists, but it was not suspected to be a sensory organ until early in the 20th century. In fact, its distinct function from that of the main olfactory system in mammals was not demonstrated until 1970, when the VNO was found to be essential for transmitting intraspecies information important for sexual maturation and aggressive behaviors in rodents.1 Surgical removal of the VNO eliminates territorial aggression and territorial marking in male mice and male hamsters. And in numerous species, including hamsters, rats, ferrets, and lemurs, VNO removal leads to a decrease in sexual investigation and copulation by males.

Female sexual behaviors are also affected when the VNO is removed. Back arching, which signals mating readiness in some female rodents, is reduced in hamsters, rats, and mice whose VNO has been removed. Puberty onset and pregnancy are affected in mice without VNOs as well. Moreover, a mouse’s ability to recognize members of their own strain and to distinguish individuals is affected. And in recent experiments, also with mice, researchers have shown that response to predator signals, parental behaviors, and the infanticidal to parental behavioral transition are also dependent on a functional VNO.

View full size JPG | PDF
Researchers have identified three main families of vomeronasal receptors (VRs) in the VNO, G protein-coupled receptors with seven transmembrane domains. Ligand binding to VRs triggers a cascade of intracellular signaling events that transform the chemical signals into electrical nerve impulses. The V1R family consists of more than 200 receptors, each with short extracellular N-terminal domains, that are similar to the odorant receptors found in the nose. The V2R family also has nearly 200 members, which may have evolved from ancestral taste receptors. The third family of VRs, the formyl peptide receptors (FPRs), contains only seven members. They are innate immune receptors, only recently determined to serve as chemosensory receptors that in the mouse VNO appear to recognize a set of cues that signal the health of individuals.3 (See illustration.)

The pheromone molecules that trigger these receptors, however, are still unknown. Some small synthetic molecules, including 2-heptanone, several dimethylpyrazines, and sulfated compounds, activate the VNO neurons, but these compounds do not elicit behavioral effects by themselves. In addition to these low-molecular weight compounds, mouse urine also contains high levels of proteins, many of which belong to the major urinary protein (MUP) family. These proteins have been shown to bind small molecules, possibly serving to retain in the urine volatile chemicals that act as pheromones. Recent experiments have suggested that MUPs may even act as pheromones on their own. Genetically engineered MUPs made by E. coli bacteria can trigger defensive behavior in mice.4

Another group of molecules that has been implicated in VNO activation is the major histocompatibility complex (MHC).5 MHC peptides are remnants of proteins that are broken down during normal cellular metabolism. These peptides are presented on the cell membrane as a complex of surface antigens. Because MHCs are highly divergent molecules between species, and because the peptides they present are determined by the genetic background of the individual animal, it was once thought that MHC peptides could serve to portray information about the genetic background of the transmitting animal. However, individual VNO neurons appear to respond to peptides from a variety of genetic backgrounds, making it unlikely that animals can distinguish between individuals based on activation by MHC peptides alone.

Finally, while the large families of VRs have been identified and some of the putative ligands found, scientists are often at a loss when it comes to matching the receptor with its ligand. So far, only one ligand-receptor pair that triggers specific behavior has been identified.6,7 Kazushige Touhara’s group at the University of Tokyo has identified a tear gland peptide (exocrine gland-secreting peptide 1, or ESP1) that activates V2Rp5 and triggers robust lordosis, or back-arching, behavior in female mice. Labeling of the V2Rp5 circuit suggests that ESP1 may activate a predetermined labeled line, as cVA does in fruit flies.

Pheromonal pathways

Pheromone-triggered behaviors are inborn, requiring no learning or prior exposure to the chemical cues. This suggests that the neural circuits are genetically specified. The labeled-line circuit in the insect brain represents a relatively simple mode of signal processing, but the pheromone circuits in the mammalian brain are much more complex. Signals detected by the VNO must be parsed and integrated to induce the proper response. The holy grail of pheromone research is therefore to identify the neural logic of signal processing in the mammalian brain.

A primary site of neural computation is likely to be the accessory olfactory bulb (AOB). VNO neurons expressing the same VRs project to multiple glomeruli in stereotypic patterns. From there, individual mitral cells project into deeper brain regions—specifically, the amygdala and the hypothalamus. Tracing experiments have established that the AOB is directly connected to these brain areas—bypassing the cortex, which mediates higher cognitive processes—thus allowing pheromones to directly trigger endocrine changes and behavioral responses without conscious thought. (See illustration above.)

Ongoing research continues to probe the nature of how pheromones are processed in the mammalian brain. Meanwhile, my group and others have focused on better understanding pheromone compounds and their receptors. To make headway studying this problem, my colleagues and I generated a transgenic mouse line that expresses the genetically encoded calcium sensor, G-CaMP2, in the VNO. Using these mice, we can image pheromone-triggered responses in VNO neurons. Stimulating VNO neurons with urine samples from individual mice of both sexes—with different genetic backgrounds or at different hormonal statuses, for example—we profiled the response patterns and were able to identify cells specifically tuned to sex, estrus signal, genetic makeup, and individual identity.8 This approach could eventually lead to the identification of receptors that convey specific information, as well as their ligands, the pheromones.

The holy grail of pheromone research is to identify the neural logic of signal processing in the mammalian brain.

Other investigators are using genetic means to interfere with VNO function. In the early 2000s, Peter Mombaerts, then at Rockefeller University, and colleagues deleted a cluster of V1R genes, causing a reduction in male copulation and maternal aggression.9 More recently, researchers demonstrated that the genetic deletion of Gαi2, the main G protein for VNO neurons expressing the V1R odorant receptors, causes a decrease in both maternal and intermale territorial aggression.10 Deletion of Gαo, the main G protein for VNO neurons expressing the V2Rs, also leads to the loss of aggression in both male and female mice.11

Even more intriguing experiments involve mice lacking the TRPC2 ion channel, which is uniquely expressed in the VNO and was thought to be exclusively responsible for VNO function. TRPC2-knockout mice exhibit a set of fascinating behaviors: they show neither territorial aggression nor maternal aggression. Instead of vigorously attacking intruder males, TRPC2-mutant males display sexual mount behaviors toward the intruders.12,13 In large arenas, TRPC2-mutant females act as if they were males themselves, and display chasing and mounting behaviors toward intruder males.14 These behaviors demonstrate that despite reduced responses to pheromone stimulation in the TRPC2 mutants, the residual responses may be sufficient to transmit signals to the brain and trigger behavioral responses—albeit abnormal ones.

Furthermore, though the behavioral patterns exhibited by such mutant mice are “inappropriate” for the conditions, they are not novel behaviors. In other words, although alterations to the VNO can reduce mating and aggressive behaviors, the behavioral patterns themselves are not altered. For example, in TRPC2 mutant mice, females may display a male-like behavior. Similarly, the parental behavior circuit also exists in the male brain, but it is not apparent until the male VNO has been activated when he mates with a female. These observations suggest that animals have a limited behavior repertoire. The function of the VNO, therefore, is to detect and integrate a multitude of signals in the environment, and, in turn, activate one of a few preset neural circuits to elicit a stereotyped behavioral output.

Many questions remain about how pheromone signals elicit such regimented behavioral patterns—namely, what are the compounds that stimulate the VNO, what are their primary receptors, and what are the details of the brain circuits that translate these signals into behavioral output? But as constantly evolving genetic and physiological approaches allow us to trace the anatomical connection from the AOB to deeper brain structures, we will continue to elucidate the pheromone circuits in the mammalian brain as we have in the insect brain. 


© NEUSTOCKIMAGES/GETTYThe existence of human pheromones is a controversial topic. Some studies have found that extracts from human sweat have a calming effect on the opposite sex, but they do not appear to induce sexual arousal.1 Others have tested the effect of steroids secreted from the armpit sweat glands and reported that one such steroid, androstenone, is attractive to some people, while others find it repugnant. Still others, however, do not consciously detect it.

Because the effects observed for these putative pheromones rely on psychophysical tests that require the subjects to report their feelings, they do not strictly fit the classic pheromone definition in that no direct action or change in physiology is observed. Moreover, the reported effects are usually subtle.

More direct evidence of the existence of human pheromones comes from a study of menstrual cycle synchronization published in the late 1990s by a team led by Martha McClintock at the University of Chicago. By collecting body secretions from women at different times in their menstrual cycles and presenting the substance under subjects’ noses, the authors reported that the test subjects either accelerated or slowed their cycles to synchronize with the donor’s, even without conscious perception of the odor.2 However, the conclusion has been called into question by other studies, and the phenomenon of menstrual synchronization itself is disputed, even when women are living together.3

Furthermore, even if human pheromones do exist, how their effects are mediated remains mysterious. The VNO is a vestigial organ in humans. The molecular machinery that is essential for VNO function in other mammals, including the ion channel TRPC2 and the V1Rs and V2Rs, have largely become nonfunctional pseudogenes in humans. Any processing of a human pheromone signal would have to occur in other systems.

  1. T.A. Hummer, M.K. McClintock, “Putative human pheromone androstadienone attunes the mind specifically to emotional information,” Horm Behav, 55:548-59, 2009.
  2. K. Stern, M.K. McClintock, “Regulation of ovulation by human pheromones,” Nature, 392:177-79, 1998.
  3. B.I. Strassmann, “Menstrual synchrony pheromones: Cause for doubt,” Hum Reprod, 14:579-80, 1999.

C. Ron Yu is an associate investigator at Stowers Institute for Medical Research and an adjunct associate professor in the Department of Anatomy and Cell Biology at the University of Kansas School of Medicine.


  1. R. Tirindelli et al., “From pheromones to behavior,” Physiol Rev, 89:921-56, 2009.
  2. K.S. Tachikawa et al., “Behavioral transition from attack to parenting in male mice: A crucial role of the vomeronasal system,” J Neurosci, 33:5120-26, 2013.
  3. S. Rivière et al., “Formyl peptide receptor-like proteins are a novel family of vomeronasal chemosensors,” Nature, 459:574-77, 2009.
  4. F. Papes et al., “The vomeronasal organ mediates interspecies defensive behaviors through detection of protein pheromone homologs,” Cell, 141:692-703, 2010.
  5. T. Leinders-Zufall et al., “MHC class I peptides as chemosensory signals in the vomeronasal organ,” Science, 306:1033-37, 2004.
  6. S. Haga et al., “The male mouse pheromone ESP1 enhances female sexual receptive behaviour through a specific vomeronasal receptor,” Nature, 466:118-22, 2010.
  7. H. Kimoto et al., “Sex- and strain-specific expression and vomeronasal activity of mouse ESP family peptides,” Curr Biol, 17:1879-84, 2007.
  8. J. He et al., “Encoding gender and individual information in the mouse vomeronasal organ,” Science, 320:535-38, 2008.
  9. K. Del Punta et al., “Deficient pheromone responses in mice lacking a cluster of vomeronasal receptor genes,” Nature, 419:70-74, 2002.
  10. E.M. Norlin et al., “Vomeronasal phenotype and behavioral alterations in Gαi2 mutant mice,” Curr Biol, 13:1214-19, 2003.
  11. P. Chamero et al., “G protein Gαo is essential for vomeronasal function and aggressive behavior in mice,” PNAS, 108:12898-903, 2011.
  12. L. Stowers et al., “Loss of sex discrimination and male-male aggression in mice deficient for TRP2,” Science, 295:1493-500, 2002.
  13. B.G. Leypold et al., “Altered sexual and social behaviors in trp2 mutant mice,” PNAS, 99: 6376-81, 2002.
  14. T. Kimchi et al., “A functional circuit underlying male sexual behaviour in the female mouse brain,” Nature, 448:1009-14, 2007.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?