Review. Something in the Air? New Insights into Mammalian Pheromones. Peter A. Brennan and Eric B. Keverne

Transcription

1 , Vol. 14, R81 R89, January 20, 2004, 2004 Elsevier Science Ltd. All rights reserved. DI /j.cub Something in the Air? New Insights into Mammalian Pheromones Peter A. Brennan and Eric B. Keverne lfaction is the dominant sensory modality for most animals and chemosensory communication is particularly well developed in many mammals. ur understanding of this form of communication has grown rapidly over the last ten years since the identification of the first olfactory receptor genes. The subsequent cloning of genes for rodent vomeronasal receptors, which are important in pheromone detection, has revealed an unexpected diversity of around 250 receptors belonging to two structurally different classes. This review will focus on the chemical nature of mammalian pheromones and the complementary roles of the main olfactory system and vomeronasal system in mediating pheromonal responses. Recent studies using genetically modified mice and electrophysiological recordings have highlighted the complexities of chemosensory communication via the vomeronasal system and the role of this system in handling information about sex and genetic identity. Although the vomeronasal organ is often regarded as only a pheromone detector, evidence is emerging that suggests it might respond to a much broader variety of chemosignals. Introduction Chemical signals that convey information between members of the same species are commonly termed pheromones. This term was first used by Karlson and Lüscher [1] and defined as substances secreted to the outside of an individual and received by a second individual of the same species in which they release a specific reaction, for example, a definite behaviour or developmental process [1]. However, there are problems with defining mammalian pheromones in this way, as it is difficult to define what constitutes a specific reaction [2]. Mammals are subject to more complex influences on their behaviour than insects. For instance, odours that convey information about individuality may bias the behaviour of other individuals of the same species [3], but because they do not elicit a specific behaviour they would not qualify as pheromones according to the original definition. The meaning of the term is further muddled by many investigators adding their own criteria for what constitutes a pheromone. For example, some investigators have stated that a substance should be airborne and not be consciously perceived in order for it to be defined as a pheromone [4]. An easy way out would be to redefine the term pheromone to encompass all chemical substances that convey information among individuals of Sub-Department of Animal Behaviour, University of Cambridge, High St., Madingley, Cambridge CB3 8AA, UK. the same species. Nevertheless, there are many mammalian examples of single molecules or cocktails of a few molecules that elicit dramatic behavioural effects. Such substances are often referred to as releaser pheromones, whereas chemosignals that cause longer term changes in neuroendocrine or developmental state are usually referred to as primer pheromones. Behavioural Effects of Mammalian Pheromones Some examples of mammalian pheromones and their effects are illustrated in Figure 1. ne of the best characterized is the rabbit nipple search pheromone. Sensed by rabbit pups via their main olfactory system, it elicits a characteristic nipple search behaviour that quickly results in the location of a nipple [5]. This guidance cue is particularly important for rabbits, as a doe only nurses her pups for around four minutes once a day and the quick location of a nipple in the face of sibling competition is vital for survival. The pheromone has recently been shown to be a single molecule, 2- methylbut-2-enal, which is produced in rabbit milk and is sufficient to elicit full nipple search and grasping behaviour when presented on its own [6]. Whereas the rabbit nipple search pheromone is a single molecule, many pheromones consist of blends of two or more molecules, which can elicit a greater response than any individual component. For example, the androgen derivatives 5α-androst-16-en-3-one and 5α-androst- 16-en-3-ol are present at high concentrations in boar saliva: each component elicits pheromonal effects to attract oestrous sows and cause them to adopt a characteristic mating stance known as standing [7]. Any body secretions are potential routes for pheromonal communication and many mammals have developed specialized scent glands that they use to deposit scent marks around their environment. Urine can also act as a medium for pheromonal signals and is especially important in controlling rodent social interactions. Urine from adult male mice contains high levels of small volatile compounds including (R, R)-3,4- dehydro-exo-brevicomin (DB), (S)-2-sec-butyl-4, 5- dihydrothiazole (BT), E,E α-farnesene, E β-farnesene and 6-hydroxy-6-methyl-3-heptanone [8]. The production of these volatile substances is dependent on testosterone and they can therefore be viewed as signalling the presence of a reproductively active male [9]. These volatile compounds accelerate the attainment of puberty in pre-pubertal females when presented individually [10], whereas a mixture of DB and BT induce oestrus and synchronize cyclicity in adult females [11]. verall, these primer pheromonal effects of male urinary constituents regulate the reproductive state of females to match the prevailing social circumstances. The same mixture of DB and BT has a releaser effect to promote aggressive behaviour in male mice towards a stimulus animal [12]. However, this is only observed when the DB and BT mix is added to castrated male urine, indicating that other urinary constituents are also

2 R82 B H C H A Androstenol 2-methylbut-2-enal estrus 3,4-dehydroexo-brevicomin induction and synchronization in female N mice S H C2H5 2-sec-butyl-4,5- dihydrothiazole Nipple search response in rabbit pups E, E α-farnesene E β-farnesene 6-hydroxy-6-methyl- 3-heptanone Figure 1. Examples of mammalian pheromones. (A) The rabbit nipple search pheromone is a single molecule present in the milk of lactating does. (B) Two similar androgen derivatives are found at high concentrations in boar saliva and elicit similar behavioural effects. (C) Major volatile constituents of male mouse urine signal maleness and elicit a range of responses depending on behavioural context. involved in this pheromonal effect. Males initially avoid urinary volatile pheromones, reflecting the caution with which they approach freshly deposited urine marks, whereas all of these volatile substances elicit approach behaviour and interest from females [8]. Therefore, male mouse urinary volatile compounds can elicit a range of pheromonal effects depending on the sex and age of the recipient and the social situation in which they are encountered. Pheromonal Communication via Lipocalins For volatile pheromones to be transported efficiently in aqueous body secretions they need to be bound to proteins known as lipocalins. These are a large class of proteins of typically around kda that have a high level of homology and include odourant-binding proteins produced by vertebrate chemosensory epithelia. These proteins have a common β-barrel structure enclosing a cup-like, hydrophobic ligand-binding pocket [13]. For example, a salivary lipoprotein is produced in boar saliva and binds to androstenol. The existence of two variants of salivary lipoprotein that differ only in the amino acids of their ligand-binding calyx suggests that they may act as specific transporters of the androstenol and androstenone components of the boar pheromone [14]. In addition to acting H Puberty acceleration in female mice H Androstenone Attraction and standing response in female pigs as transporters, it has been suggested that lipocalins themselves may bind to receptors and function as chemosignals [13]. A 19 kda lipocalin called aphrodisin, found in the vaginal fluid of female hamsters, attracts the attention of males and promotes copulatory activity. Aphrodisin binds the volatile pheromones methyl thiobutyrate and dimethyl disulphide, which are abundant in the vaginal secretions of oestrous hamsters. However, it has yet to be demonstrated whether aphrodisin itself has pheromonal activity or whether it is only a transporter of pheromones [15]. Recombinant aphrodisin, produced in a yeast expression system is not as effective as the endogenous protein although this may be due to a different pattern of glycosylation from that of the natural protein. Interestingly, the addition of a volatile extract from hamster vaginal fluid restores pheromonal activity to recombinant aphrodisin, suggesting that it may be the combination of aphrodisin with bound ligand that interacts with a receptor [15]. Mouse urinary volatile pheromones are bound to 19 kda lipocalins known as major urinary proteins (MUPs) [16]. These proteins are found at levels of up to 70 mg/ml in male mouse urine, which represents a significant energy expenditure and reflects their importance in chemosensory communication [17]. Male mice mark their territory by depositing urine streaks to signal ownership [18]. Although these marks dry rapidly, the MUPs act as a reservoir for the volatile pheromones, prolonging their release over a period of several hours [19]. Volatile urinary constituents released from the MUPs attract attention to the urine marks, whereas the protein component of the urine acts as a chemosignal to promote countermarking behaviour in which a male will deposit a urine mark adjacent to an existing mark from a competing male [20]. However, it is unlikely that MUPs function solely as a reservoir or transporter of volatile pheromonal cues. MUPs are highly polymorphic and a wild mouse may produce between four and fifteen MUP variants in its urine, thereby forming an individual identity code [17]. These MUP variants may have different ligand-binding properties and could convey information about individual identity by influencing the profile of the urinary volatile compounds [21]. However, the MUPs themselves could interact directly with chemosensory receptors to provide a more reliable signal of individuality of urine marks [13]. Evidence to support this idea has come from studies by Hurst s group [22] showing that male countermarking behaviour was more dependent on MUP profile than other signals of individual identity such as the major histocompatibility complex (MHC) type. Thus lipocalins such as MUPs may have multiple roles as both carriers of pheromones and chemosignals themselves. The Vomeronasal rgan in Pheromonal Signalling In addition to the main olfactory system, most mammals possess a well-developed vomeronasal system that is commonly regarded to be specialized for the transmission of pheromonal information [23]. The vomeronasal organ (VN) is a blind-ended, mucus-filled tube, located in the nasal septum [24]. It is linked to the nasal cavity by the narrow vomeronasal duct, although in some species it is associated with

3 R83 the nasopalatine duct linking it to the oral cavity. The vomeronasal sensory neurons (VSNs) are located medially, in a crescent-shaped epithelium, with a large blood vessel running laterally to the lumen (Figure 2). The whole organ is encased in a hard bony and cartilaginous capsule, which is important for vomeronasal function. Stimulus access to the VN differs from that to the main olfactory epithelium. As the VSNs are located away from the nasal airstream, stimulus access depends on a vascular pumping mechanism that is activated in situations of novelty [25,26]. This mechanism enables the vomeronasal organ to take up relatively non-volatile stimuli such as lipocalins from urine deposits, vaginal secretions, scent gland secretions or saliva. These stimuli are investigated by direct contact [27] and specific behaviours such as facial grooming and flehmen may facilitate the entry of nonvolatile substances into the VN. The uptake of non-volatile stimuli by the VN does not mean that the vomeronasal system is unable to respond to volatile chemosignals. Type-3 adenylyl cyclase knockout mice lack main olfactory receptor function but they respond to the presence of volatile urinary pheromones such as 2-heptanone and dimethylpyrazine when presented on a cotton swab, without direct contact [28]. Furthermore, VSNs in the apical layer of the vomeronasal epithelium respond selectively to volatile urinary constituents in vitro [29], although the MUPs may play an important role in transporting these volatiles into the VN in vivo. Recent reports suggest that the vomeronasal system may respond to volatile odourants as readily as the main olfactory system. Calcium imaging of VSNs found that they responded to 18 of a panel of 82 general odourants that were tested [30]. Furthermore, a VN response was observed in type-3 adenylyl cyclase knockout mice to odourants such as ethyl vanillin and ethyl acetate that are not present in mouse urine and have no known pheromonal activity [28]. f course, these odourants could have an undiscovered pheromonal activity in mice and be secreted by a route other than urine, but it is also possible that the VN handles more than just pheromonal information. Garter snakes use their VN to detect both predator and prey chemosignals [31] and given their vomeronasal receptor repertoire, the mammalian VN could respond to similarly diverse stimuli [32]. Vomeronasal Receptor Genes Recent work has highlighted fundamental differences between the chemosensory receptors of the main olfactory epithelium and the VN. Analysis of the mouse genome has revealed around 1,000 functional olfactory genes [33]. nly one of these is expressed per olfactory sensory neuron (SN), although as olfactory receptors are typically not very selective a SN will typically respond to a range of related odourants [34]. In contrast there are around 250 to 300 functional vomeronasal receptor genes, which belong to two classes that have little homology with each other or with the main olfactory receptors [35 37]. These are expressed by VSNs in separate zones of the vomeronasal epithelium, characterised by specific G proteins. The V1R class of A B lfactory epithelium Vomeronasal nerves Vomeronasal organ Nasal cavity Lumen Blood vessel Cavernous tissue Main olfactory bulb Cartilaginous capsule Accessory olfactory bulb Nasal septum V2Rs vomeronasal receptors V1Rs Sensory epithelium Figure 2. The rodent vomeronasal organ. (A) Sagittal view showing the location of the tubular vomeronasal organ in relation to the main olfactory epithelium. (B) Diagram of a coronal section of the vomeronasal organ, which is enclosed in a cartilaginous capsule. The sensory epithelium containing the vomeronasal sensory neurons is located medially. Changes in blood flow to the laterally located blood vessel change the volume of the lumen and pump stimuli in and out. vomeronasal receptors comprises a family of 137 functional receptors (Figure 3) that are weakly related to T2R taste receptors [38]. V1Rs are expressed in VSNs whose cell bodies are located in the apical zone of the epithelium and also contain the G protein Gα i2 [39]. These receptors can be allocated to 12 diverse families with typically greater than 40% sequence identity among members of the same family, but only 15 40% interfamily similarity [32], suggesting that the V1R receptors respond to a wide variety of different ligands. In vitro recordings from V1R-expressing cells has revealed that they respond to small hydrophobic molecules, such as the urinary volatiles DB and BT [29]. It has even been possible to match a ligand to a particular receptor type, as the V1Rb2 receptor has been shown to respond to 2- heptanone [40], an oestrus-extending pheromone in male mouse urine [8]. The VSNs are extremely sensitive, with thresholds for farnesene of to M, making them among the most sensitive of mammalian chemoreceptors [29]. Unlike SNs, which typically respond to a range of related molecules, the VSNs appear to be highly selective for individual molecules. Furthermore, the tuning curve of VSNs does not broaden as stimulus concentration is increased, as occurs in the SNs [29].

4 R84 V1rg V1rd V1rl V1re V1rk V1rj V1rb V1rf V1ra V1rh V1rc V1ri Figure 3. The V1R class of vomeronasal receptor genes. Analysis of the mouse genome has revealed 137 functional receptors that can be grouped into 12 families. Reproduced with permission from Rodriguez et al. [32]. The V2R class of receptors has been estimated to contain genes for around 100 functional receptors. These are expressed by VSNs located in the basal zone of the epithelium, which also express the G protein Gα o and extend a long primary dendrite to the surface of the vomeronasal epithelium [35]. The structure of the V2Rs differs markedly from that of the V1Rs with a large extracellular amino-terminal domain, similar to that of metabotropic glutamate receptors, Ca 2+ -sensing receptors and T1R taste receptors [41]. This region has a high level of sequence variability and is likely to form the ligand-binding site [35]. Whereas V1R-expressing VSNs are thought to express only a single type of receptor and respond highly specifically, it is unclear whether this also applies to the V2R-expressing VSNs. Although antisera against the Go-VN2, Go-VN3 and Go- VN4 V2R receptors revealed a punctate distribution of staining in a small proportion of the basal VSNs, staining for the V2R2 receptor was found throughout the basal zone of the epithelium and therefore coexpressed with the other V2Rs [42]. This may reflect the formation of receptor dimers in a similar way to the mglur1 receptors and the T1R family of taste receptors with which the V2Rs share homology. An interesting recent development is the discovery that V2Rs are co-expressed with H2-Mv, non-classical MHC molecules. There are nine members of this family, M1, M9, M10.1 to M10.6 and M11, which are only expressed in the V2R-VSNs in the basal zone of the VN [43,44]. The M10 family form multimolecular complexes with V2Rs and β2-microglobulin, which is essential for the stabilization of the receptor complex and its localization at the cell surface [43]. In addition to this vital role in receptor localization, H2-Mv proteins could act as co-regulators of V2Rs [44]. This idea is supported by the variability among the M10 molecules being concentrated in the region of the peptide-binding pocket. However, members of the H2-Mv family have lost several of the amino-acid residues that are normally involved in peptide binding. If they do bind peptides, it is therefore likely to be in a manner different from classical class I MHC molecules [43]. Interestingly, the H2-Mv genes are not expressed randomly, but certain combinations of one or more H2-Mv genes are found with particular V2Rs [44]. Theoretically, this might provide a link between an individual s own MHC type and their responsiveness to MHC-associated chemosignals, which are important for aspects of rodent social behaviour, such as mate choice. Insights into Vomeronasal Function from Genetically Modified Mice Adaptation plays an important role in changing the sensitivity of most sensory systems, allowing them to function over a range of stimulus intensities. It plays a particularly important role in the main olfactory system to enable the segregation of odours from background levels of odourants [45]. In contrast, VSNs do not adapt during maintained stimulus exposure [46]. It may be more important for the vomeronasal system to maintain sensitivity to very low amounts of chemosignals for prolonged periods, which are required for rodent urinary pheromones to induce primer effects on reproductive state [47]. Indeed the mechanism for sensory transduction in VSNs is totally different from that of SNs, as it is dependent on phospholipase C activity and the activation of Trp2 cation channels [46,48]. Thus genetically modified mice that lack functional Trp2 channels have normal main olfactory function but fail to respond to urinary vomeronasal stimuli and have a range of behavioural abnormalities [49]. These mice fail to initiate aggressive behaviour in a resident intruder test in response to the introduction of a castrated male that has been swabbed with male urine [49,50]. Trp2-deficient mice also have a deficit in maternal aggression, which is normally shown by lactating females in the presence of a male intruder [50]. These findings confirm previous observations on mice with vomeronasal lesions, as these mice also failed to initiate aggressive behaviour in response to chemosensory cues [51]. The correspondence between the behavioural effects of the Trp2 deletion and VN ablation is not so good in the context of sexual behaviour (Table 1). Previous studies using VN ablation in male mice impaired, but did not abolish, sexual behaviour toward females [52,53]. In contrast, sexual behaviour toward females appears to be unaffected in male Trp2-deficient mice. Instead, Trp2-deficient males showed an abnormally high level of mounting behaviour toward other males, which persisted even in the presence of a receptive female [49,50]. Furthermore, males normally emit ultrasonic vocalizations in the presence of a female, which are suppressed by the presence of a male. These vocalizations are abolished in sexually naïve, VNablated males and reduced in males that have had previous sexual experience [53]. However, ultrasonic vocalizations persist in Trp2-deficient males and are not suppressed in the presence of male chemosignals [49].

5 R85 Table 1. Behavioural impairments, relative to wild type, caused by surgical ablation of VN or genetic manipulation of VSNs. VN ablation Trp2 -/- V1rab -/- β2m -/- Gαi2 mutant Male sexual behaviour Reduced, fewer mounts Normal Reduced drive to Not reported Normal directed to females and longer latency, initiate sexual experience-dependent behaviour Male sexual behaviour Not reported Increased Normal/reduced Normal Normal directed to males Male ultrasonic Eliminated/reduced, Normal Normal Not reported Not reported vocalizations to females experience-dependent Male ultrasonic Eliminated/reduced, Increased Not reported Not reported Not reported vocalizations to males experience-dependent Male aggression to Eliminated Eliminated Normal Eliminated Reduced, depends on intruder male genetic background Maternal aggression Eliminated Eliminated Eliminated Not reported Eliminated/reduced to intruder male How can we account for the behavioural differences between VN-ablated and Trp2-deficient mice? The VN is difficult to remove completely in the adult, which could allow some residual function in VN-ablated mice. Equally, it is not certain that lack of Trp2 completely abolishes VSN responses. Electrophysiological recordings from individual VSNs appear to show normal baseline activity, but no responsiveness to urinary stimuli [49]. However, local field potential recordings from the vomeronasal epithelium revealed a reduced response, at high concentrations, to both the single vomeronasal stimulant 2-heptanone and to urine from males and females [50]. This suggests that Trp2 knockout causes a dramatic loss of sensitivity of VSNs rather than a total loss of function. Another factor that should be considered when analyzing the effects of disrupting vomeronasally related genes is the possible developmental compensation that could occur in animals that have the defect from birth, as this may give different results from ablating the same function in the adult. More restricted vomeronasal deficits are observed in V1rab-deficient mice in which the genes for 16 receptors of the V1ra and V1rb families have been deleted, amounting to approximately 12% of the V1R receptor repertoire [54]. Unlike Trp2-deficient mice, male V1rabdeficient mice do not misdirect sexual behaviour to males. However, they are impaired in their sexual behaviour to females, with fewer V1rab-deficient males initiating sexual behaviour toward females than controls. Interestingly, they were able to respond to the male urinary signal 2-heptanone, despite the demonstration that it stimulates the V1rb2 receptor [40]. This suggests that, although the vomeronasal receptors have narrow response ranges, a vomeronasal pheromone may be sensed by more than one type of vomeronasal receptor. V1rab-deficient mice fail to initiate maternal aggression to intruder males, as had been found in Trp2-deficient mice, although there were no differences in aggressive behaviour of V1rab-deficient males compared with control mice [54]. Different male chemosignals may therefore be responsible for triggering aggression in males and females. ther selective genetic modifications have looked at the disruption of genes that are expressed in either the V1R or V2R class of VSN. Mice in which β2-microglobulin is knocked out fail to form the V2R M10 receptor complex and consequently fail to localize V2Rs in the sensory tip of the VSNs [43]. These β2-microglobulindeficient males fail to initiate aggressive behaviour in the resident intruder test, although their sexual behaviour is reported to be normal [43]. Mutants of the V1Rspecific Gα i2 have around a 50% reduction in V1R class of VSNs. These mice show reduced but not abolished levels of aggressive behaviour, evident from the longer latency before attacking an intruder male. However, these mutant Gα i2 males showed normal sexual behaviour directed at females [55]. These results demonstrate the important role of the vomeronasal system in sensing male chemosignals and initiating the appropriate aggressive response. Although it seems that the VN is involved in sexual behaviour, its exact role is unclear however, possibly due to the complexity of the chemosignals and the involvement of the main olfactory system. The loss of function of a subset of vomeronasal receptors may lead to a distortion of a complex pheromonal signal rather than its attenuation. The use of genetically manipulated mice therefore provides an important new tool for asking questions about vomeronasal function, but is unlikely to provide all the answers. A full understanding will require a greater appreciation of the individual components of complex chemosensory signals and this goal is some distance away yet. Neural Pathways for Pheromonal Effects The vomeronasal system provides a relatively direct pathway to hypothalamic areas by which pheromones can elicit behavioural and neuroendocrine effects [56]. The VSNs project to the accessory olfactory bulb (AB), which in turn projects via the corticomedial amygdala and bed nucleus of the stria terminalis to the hypothalamus [57]. The two classes of VSN project to anatomically and functionally separate sub-regions of the rodent AB, suggesting that information about different types of vomeronasal stimuli is processed separately at the level of the AB [58]. V1R-expressing VSNs located in the superficial zone of the vomeronasal epithelium project to the anterior sub-region of the AB, whereas

6 R86 NH 2 Anterior AB Posterior AB BNST VN V1R CH β2m NH 2 PMCo M10 V2R CH BAT MeA Figure 4. Neural pathways for vomeronasal processing. The V1Rs are a class of G-protein-coupled receptors that are expressed by vomeronasal sensory neurons in the apical zone of the sensory epithelium. These V1R sensory neurons (shown in blue) project to the anterior sub-region of the accessory olfactory bulb (AB). The V2Rs are also G-protein-coupled receptors, but their structure differs markedly from the V1Rs. The V2Rs have a large extracellular amino terminus and form a receptor complex with members of the M10 family of MHC class Ib molecules and β2- microglobulin (β2m). The V2R-expressing sensory neurons (shown in red) are located in the basal zone of the sensory epithelium and project to the posterior sub-region of the AB. The information from the V1Rs and V2Rs, which is processed separately in the AB, converges in completely overlapping projections to the bed nucleus of the stria terminalis (BNST), bed nucleus of the accessory olfactory tract (BAT), medial amygdala (MeA) and the posteromedial cortical nucleus (PMCo) of the amygdala. basal VSNs expressing V2Rs project to the posterior sub-region (Figure 4). Recordings of population potentials and the use of voltage-sensitive dyes in the guinea pig AB have demonstrated that activity evoked in either AB region by selective stimulation of the vomeronasal nerve input does not propagate across the anatomical boundary [59]. Therefore, these sub-divisions are functionally as well as anatomically distinct in the guinea pig AB. Moreover, local field potential oscillations induced in the anterior sub-region are more heavily damped than those in the posterior sub-region, suggesting that there are also differences in neural processing [59]. The nature of the information conveyed by the separate V1R and V2R pathways is unclear, but the projections from the anterior and posterior sub-divisions of the AB completely overlap at the level of the amygdala, the bed nucleus of the accessory olfactory tract and the bed nucleus of the stria terminalis [60]. Recent results have shown that the coding of information in the vomeronasal system and the main olfactory system may not differ as much as previously thought. It has been known for some time that SNs in the main olfactory epithelium that express the same olfactory receptor converge on a pair of glomeruli on opposite sides of the main olfactory bulb (MB) [61]. These glomeruli receive input from only one receptor type and appear to be fundamental units of odour processing. In contrast, VSNs expressing a particular receptor protein typically project to small glomeruli out of a total of a few hundred in the AB [62,63]. Although the pattern of projection to glomeruli in the AB is more variable than the highly organized projection pattern in the MB, it is not random. Certain areas consistently receive projections from particular receptor types across individuals [63]. There are important differences in the morphology of the projection neurons between the MB and the AB [64]. Mitral cells of the MB project a single primary dendrite to a single glomerulus, collecting information from a single receptor type. In contrast, mitral cells in the AB send a branched primary dendritic tree that collects information from multiple glomeruli. This pattern of connectivity has been suggested to indicate a high degree of integration of pheromonal information in the AB. However, recent evidence suggests that the branched primary dendrites of AB mitral cells only contact glomeruli that receive information from the same vomeronasal receptor [65]. It therefore appears that the divergent projection of VSNs expressing the same receptor to multiple AB glomeruli converges at the level of the mitral projection neurons (Figure 5). Volatile pheromones involved in sex signalling are present at different concentrations in male and female mouse urine [9]. Therefore, although the response of a VSN may be highly specific for a certain chemosignal, it is still ambiguous for sex unless the concentration of the stimulus is known. However, reliable information about the sex of the producer can be easily obtained by analyzing the pattern of activity across a population of VSNs with different receptor sensitivities [46]. This may be a function of the inhibitory interactions among mitral cells via their reciprocal synapses with periglomerular

7 R87 and granule interneurons, which mediate inhibitory interactions among the primary dendritic trees of mitral cells [64]. Katz s group [27] has recently performed the first recordings from AB mitral cells of freely behaving mice, during the investigation of anaesthetized stimulus animals. This group found that mitral cells respond to specific combinations of strain and sex of conspecific mice. For example, a mitral cell from a CBA male was strongly activated by a BALB/c male, strongly inhibited by a CBA female, more weakly inhibited by a C57/B6 female and did not respond to certain other combinations. This strain specificity of the responses of AB mitral cells may be important for mate choice and nepotistic behaviours, as well as a female s ability to recognize the individual identity of the male with which she has mated [66]. This simple form of learning depends on neural changes in the AB and is vitally important to prevent the pregnancy-blocking effect of male pheromones. These changes are hypothesized to increase the self-inhibition of mitral cells that respond to the mating male s pheromonal signal, thus preventing it from being transmitted centrally to induce pregnancy block [67]. Although many mammalian pheromonal effects are mediated by the vomeronasal system, it would be incorrect to think of the vomeronasal system as being the pathway for pheromonal effects in mammals. Destruction of the vomeronasal system does not impair the responses to the nipple search pheromone in rabbits [68] or to the boar salivary pheromone in sows [7]. This implies that these pheromonal responses can be mediated by the main olfactory system or possibly the septal organ. Relatively little is known about this isolated patch of chemosensory epithelium, but it projects to certain dorsal regions of the MB and may provide a specialized pathway for pheromonal information [69]. The projections of the MB are much more widespread than the AB, including structures such as piriform cortex, orbitofrontal cortex, amygdala, hippocampus, hypothalamus and striatum. These allow a high degree of integration of pheromonal information with other sensory stimuli, and a greater degree of flexibility in linking it to emotional and behavioural responses. In addition to its own input to the hypothalamus via the lateral cortical amygdala, information from the main olfactory system can gain access to the vomeronasal pathway via the medial amygdala [57]. Although VN lesions cause severe deficits in mating behaviour in sexually naïve hamsters, when sexual experience has been gained sexual behaviour can be sustained solely by cues mediated by the main olfactory system [70]. Thus odours associated with the receptive female may become associated with the vomeronasal input and may subsequently become sufficient to drive the behavioural response [71]. Associative learning of odour stimuli also occurs in the nipple search pheromonal responses seen in rabbit pups. As mentioned previously, the nipple search response is naturally elicited by 2-methylbut-2-enal in the doe s milk [6]. However, if the doe s ventrum is painted during suckling with an artificial odour, such as a perfume, then the artificial odour becomes able to elicit the characteristic nipple search response after SNs Glomerulus Mitral cells Granule cells VSNs Figure 5. Differences in sensory processing in the main olfactory and vomeronasal systems. Axons from olfactory sensory neurons (SNs) expressing a particular receptor type (here coloured blue or pink) converge on a single glomerulus on each side of the main olfactory bulb (MB). Mitral cells send a single primary dendrite to a glomerulus to sample information from a single olfactory receptor. Axons from vomeronasal sensory neurons (VSNs) expressing a particular receptor converge on small glomeruli in the accessory olfactory bulb (AB). Although this may appear to represent a different coding strategy from that used in the MB, the AB mitral cells send a branched primary dendritic tree to multiple glomeruli innervated by the same receptor type. In both cases, the spatio-temporal pattern of mitral cell activity is influenced by inhibitory interactions at reciprocal synapses with inhibitory interneurons such as granule cells. In the MB, these inhibitory interactions occur on the extensive lateral dendrites. In contrast, the reciprocal synapses in the AB occur largely on the primary dendritic tree and changes at these synapses are thought to underlie mate recognition in the context of the pregnancy block effect. only a single learning exposure [72]. Thus, although pheromonal responses are normally thought of as being innate and stereotyped, these examples show that learning of odour cues by the main olfactory system is capable of reinforcing pheromonal responses and provides extra versatility for chemosensory-mediated behaviour. Therefore, the main olfactory system and vomeronasal system should perhaps be viewed as complementary rather than separate pathways for chemosensory communication. Do Human Pheromones Exist? Although a VN develops in the human fetus and is important for guiding the migration of luteinising hormone-releasing hormone neurons to the hypothalamus, it degenerates before birth. The overwhelming weight of anatomical evidence implies that any VNlike structure in the adult human is vestigial [73]. This argument is backed up by recent analysis of the human genome. The gene for the Trp2 channel, which is likely to play an important role in VSN transduction, is actually a pseudogene in humans [48]. Analysis of sequence data from a range of primates suggests that

8 R88 the selective pressure on the Trp2 gene was relaxed in the common ancestor of the ld World monkeys and the apes. This probably occurred around the time of the development of trichromatic vision, and may reflect the greater flexibility of visual and auditory methods of communication in animals with more complex social systems [74]. Nevertheless, the lack of a functional VN does not necessarily imply that human pheromones do not exist. Almost all of the vomeronasal receptors that have been discovered in the human genome are nonfunctional pseudogenes. However, four functional human V1R genes have been found in the human genome [75]. ne of these, hv1rl1, is expressed in the olfactory mucosa [76] and is likely to have a receptor function, although not necessarily as a pheromonal receptor. At present, the strongest candidate for a pheromonal effect in humans is the effectiveness of axillary secretions from women in different phases of their menstrual cycle in altering the cycle length of other women [4]. This effect is proposed to underlie the synchronization of menstrual cycles observed in groups of women living in close proximity. Given the complexity of human society and relationships it is not altogether surprising that no chemosignals have been found that produce as powerful a sexual attraction as is seen in insects. Any pheromonal effects that are present are likely to be subtle and difficult to detect. Along with the varying definitions of what constitutes a pheromone, this makes the existence of human pheromones a particularly controversial topic. Perspective Pheromones are important determinants of behaviour for the majority of mammals. As well as increasing our understanding of animal behaviour and the process of speciation, such knowledge provides a range of opportunities in applied fields, for example, to control rodent populations and to regulate reproduction of domestic and wild species. The continuing rapid pace of developments both in identifying receptors and their ligands and in recording the responses from the vomeronasal system suggests that we are only getting the first glimpses into the invisible world of mammalian pheromones. Acknowledgments We would like to thank the Medical Research Council and the Biotechnology and Biological Sciences Research Council for their financial support. References 1. Karlson, P., and Lüscher, M. (1959). Pheromones: a new term for a class of biologically active substances. Nature 183, Doty, R.L. (2003). Mammalian pheromones: fact or fantasy? In Handbook of lfaction and Gustation, R.L. Doty, ed. (Marcel Dekker Inc), pp Beauchamp, G.K., and Yamazaki, K. (2003). Chemical signalling in mice. Biochem. Soc. Trans. 31, Stern, K., and McClintock, M.K. (1998). Regulation of ovulation by human pheromones. Nature 392, Distel, H., and Hudson, R. (1985). The contribution of the olfactory and tactile modalities to the performance of nipple-search behaviour in newborn rabbits. J. Comp. Physiol. [A] 157, Schaal, B., Coureaud, G., Langlois, D., Giniès, C., Sémon, E., and Perrier, G. (2003). Chemical and behavioural characterization of the rabbit mammary pheromone. Nature 424, Dorries, K.M., Adkins, R.E., and Halpern, B.P. (1997). Sensitivity and behavioral responses to the pheromone androstenone are not mediated by the vomeronasal organ in domestic pigs. Brain Behav. Evol. 49, Novotny, M.V. (2003). Pheromones, binding proteins and receptor responses in rodents. Biochem. Soc. Trans. 31, Schwende, F.J., Wiesler, D., Jorgenson, J.W., Carmack, M., and Novotny, M.V. (1986). Urinary volatile constituents of the house mouse, Mus musculus, and their endocrine dependency. J. Chem. Ecol. 12, Novotny, M.V., Weidong, M., Wiesler, D., and Zidek, L. (1999). Positive identification of the puberty-accelerating pheromone of the house mouse: the volatile ligands associating with the major urinary protein. Proc. R. Soc. Lond. B. Biol. Sci. 266, Jemiolo, B., Harvey, S., and Novotny, M. (1986). Promotion of the Whitten effect in female mice by synthetic analogs of male urinary constituents. Proc. Natl. Acad. Sci. USA 83, Novotny, M., Harvey, S., Jemiolo, B., and Alberts, J. (1985). Synthetic pheromones that promote inter-male aggression in mice. Proc. Natl. Acad. Sci. USA 82, Flower, D.R. (1996). The lipocalin protein family: structure and function. Biochem. J. 318, Loebel, D., Scaloni, A., Paolini, S., Fini, C., Ferrara, L., Breer, H., and Pelosi, P. (2000). Cloning, post-translational modifications, heterologous expression and ligand-binding of boar salivary lipocalin. Biochem. J. 350, Singer, A.G., and Macrides, F. (1990). Aphrodisin: pheromone or transducer? Chem. Senses 15, Böcskei, Z., Groom, C.R., Flower, D.R., Wright, C.E., Phillips, S.E.V., Cavaggioni, A., Findlay, J.B.C., and North, A.C.T. (1992). Pheromone binding to two rodent urinary proteins revealed by X-ray crystallography. Nature 360, Beynon, R.J., and Hurst, J.L. (2003). Multiple roles of major urinary proteins in the house mouse, Mus domesticus. Biochem. Soc. Trans. 31, Hurst, J.L. (1990). Urine marking in populations of wild house mice Mus domesticus Rutty. I. Communication between males. Anim. Behav. 40, Hurst, J.L., Robertson, D.H.L., Tolladay, U., and Beynon, R.J. (1998). Proteins in urine scent marks of male house mice extend the longevity of olfactory signals. Anim. Behav. 55, Humphries, R.E., Robertson, D.H.L., Beynon, R.J., and Hurst, J.L. (1999). Unravelling the chemical basis of competitive scent marking in house mice. Anim. Behav. 58, Singer, A.G., Beauchamp, G.K., and Yamazaki, K. (1997). Volatile signals of the major histocompatibility complex in male mouse urine. Proc. Natl. Acad. Sci. USA 94, Hurst, J.L., Payne, C.E., Nevison, C.M., Marie, A.D., Humphries, R.E., Robertson, D.H.L., Cavaggioni, A., and Beynon, R.J. (2001). Individual recognition in mice mediated by major urinary proteins. Nature 414, Dulac, C., and Torello, A.T. (2003). Molecular detection of pheromone signals in mammals: from genes to behaviour. Nat. Rev. Neurosci. 4, Døving, K.B., and Trotier, D. (1998). Structure and function of the vomeronasal organ. J. Exp. Biol. 21, Meredith, M., and Connell, R.J. (1979). Efferent control of stimulus access to the hamster vomeronasal organ. J. Physiol. 286, Meredith, M. (1994). Chronic recording of vomeronasal pump activation in awake behaving hamsters. Physiol. Behav. 56, Luo, M.M., Fee, M.S., and Katz, L.C. (2003). Encoding pheromonal signals in the accessory olfactory bulb of behaving mice. Science 299, Trinh, K., and Storm, D.R. (2003). Vomeronasal organ detects odorants in absence of signaling through main olfactory epithelium. Nat. Neurosci. 6, Leinders-Zufall, T., Lane, A.P., Puche, A.C., Ma, W., Novotny, M.V., Shipley, M.T., and Zufall, F. (2000). Ultrasensitive pheromone detection by mammalian vomeronasal neurons. Nature 405, Sam, M., Vora, S., Malnic, B., Ma, W., Novotny, M.V., and Buck, L.B. (2001). dorants may arouse instinctive behaviours. Nature 412, Millar, A.R., and Gutzke, W.H.N. (1999). The role of the vomeronasal organ of crotalines (Reptilia: Serpentes: Viperidae) in predator detection. Anim. Behav. 58, Rodriguez, I., Del Punta, K., Rothman, A., Ishii, T., and Mombaerts, P. (2002). Multiple new and isolated families within the mouse superfamily of V1r vomeronasal receptors. Nat. Neurosci. 5, Zhang, X., and Firestein, S. (2002). The olfactory receptor gene superfamily of the mouse. Nat. Neurosci. 5, Malnic, B., Hirono, J., Sato, T., and Buck, L.B. (1999). Combinatorial receptor codes for odors. Cell 96,

9 R Herrada, G., and Dulac, C. (1997). A novel family of putative pheromone receptors in mammals with a topographically organized and sexually dimorphic distribution. Cell 90, Matsunami, H., and Buck, L.B. (1997). A multigene family encoding a diverse array of putative pheromone receptors in mammals. Cell 90, Ryba, N.J.P., and Tirindelli, R. (1997). A new multigene family of putative pheromone receptors. Neuron 19, Adler, E., Hoon, M.A., Mueller, K.L., Chandrashekar, J., Ryba, N.J.P., and Zucker, C.S. (2000). A novel family of mammalian taste receptors. Cell 100, Dulac, C., and Axel, R. (1995). A novel family of genes encoding putative pheromone receptors in mammals. Cell 83, Boschat, C., Pélofi, C., Randin,., Roppolo, D., Lüscher, C., Broillet, M., and Rodriguez, I. (2002). Pheromone detection mediated by a V1r vomeronasal receptor. Nat. Neurosci. 5, Matsunami, H., and Amrein, H. (2003). Taste and pheromone perception in mammals and flies. Genome Biol. 4, Martini, S., Silvotti, L., Shirazi, A., Ryba, J.P., and Tirindelli, R. (2001). Co-expression of putative pheromone receptors in the sensory neurons of the vomeronasal organ. J. Neurosci. 21, Loconto, J., Papes, F., Chang, E., Stowers, L., Jones, E.P., Takada, T., Kumanovics, A., Fischer-Lindahl, K., and Dulac, C. (2003). Functional expression of murine V2R pheromone receptors involves selective association with the M10 and M1 families of MHC class 1b molecules. Cell 112, Ishii, T., Hirota, J., and Mombaerts, P. (2003). Combinational coexpression of neural and immune multigene families in mouse vomeronasal sensory systems. Curr. Biol. 13, Kelliher, K., Ziesmann, J., Munger, S.D., Reed, R.R., and Zufall, F. (2003). Importance of the CNGA4 channel gene for odor discrimination and adaptation in behaving mice. Proc. Natl. Acad. Sci. USA 100, Holy, T.E., Dulac, C., and Meister, M. (2000). Responses of vomeronasal neurons to natural stimuli. Science 289, Rosser, A.E., Remfry, C.J., and Keverne, E.B. (1989). Restricted exposure of mice to primer pheromones coincident with prolactin surges blocks pregnancy by changing hypothalamic dopamine release. J. Reprod. Fertil. 87, Liman, E.R., Corey, D.P., and Dulac, C. (1999). TRP2: A candidate transduction channel for mammalian pheromone sensory signaling. Proc. Natl. Acad. Sci. USA 96, Stowers, L., Holy, T.E., Meister, M., Dulac, C., and Koentges, G. (2002). Loss of sex discrimination and male-male aggression in mice deficient for TRP2. Science 295, Leypold, B.G., Yu, C.R., Leinders-Zufall, T., Kim, M.M., Zufall, F., and Axel, R. (2002). Altered sexual and social behaviours in trp2 mutant mice. Proc. Natl. Acad. Sci. USA 99, Maruniak, J.A., Wysocki, C.J., and Taylor, J.A. (1986). Mediation of male mouse urine marking and aggression by the vomeronasal organ. Physiol. Behav. 37, Clancy, A.N., Coquelin, A., Macrides, F., Gorski, R.A., and Nobles, E.P. (1984). Sexual behavior and aggression in male mice: Involvement of the vomeronasal organ. J. Neurosci. 4, Wysocki, C.J., and Lepri, J.L. (1991). Consequences of removing the vomeronasal organ. J. Steroid Biochem. Mol. Biol. 39, Del Punta, K., Leinders-Zufall, T., Rodriguez, I., Jukam, D., Wysocki, C.J., gawa, S., Zufall, F., and Mombaerts, P. (2002). Deficient pheromone responses in mice lacking a cluster of vomeronasal receptor genes. Nature 419, Norlin, E.M., Gussing, F., and Berghard, A. (2003). Vomeronasal phenotype and behavioral alterations in Gαi2 mutant mice. Curr. Biol. 13, Li, C.S., Kaba, H., Saito, H., and Seto, K. (1990). Neural mechanisms underlying the action of primer pheromones in mice. Neuroscience 36, Swanson, L.W., and Petrovich, G.D. (1998). What is the amygdala? Trends Neurosci. 21, Halpern, M., Jia, C.P., and Shapiro, L.S. (1998). Segregated pathways in the vomeronasal system. Microsc. Res. Tech. 41, Sugai, T., Sugitani, M., and noda, N. (1997). Subdivisions of the guinea pig accessory olfactory bulb revealed by the combined method with immunohistochemistry, electrophysiological, and optical recordings. Neuroscience 79, von Campenhausen, H., and Mori, K. (2000). Convergence of segregated pheromonal pathways from the accessory olfactory bulb to the cortex in the mouse. Eur. J. Neurosci. 12, Mombaerts, P., Wang, F., Dulac, C., Chao, S.K., Nemes, A., Mendelsohn, M., Edmondson, J., and Axel, R. (1996). Visualizing an olfactory sensory map. Cell 87, Rodriguez, I., Feinstein, P., and Mombaerts, P. (1999). Variable patterns of axonal projections of sensory neurons in the mouse vomeronasal system. Cell 97, Belluscio, L., Koentges, G., Axel, R., and Dulac, C. (1999). A map of pheromone receptor activation in the mammalian brain. Cell 97, Mori, K. (1987). Membrane and synaptic properties of identified neurons in the olfactory bulb. Prog. Neurobiol. 29, Del Punta, K., Puche, A., Adams, N.C., Rodriguez, I., and Mombaerts, P. (2002). A divergent pattern of sensory axonal projections is rendered convergent by second-order neurons in the accessory olfactory bulb. Neuron 35, Brennan, P., Kaba, H., and Keverne, E.B. (1990). lfactory Recognition: a simple memory system. Science 250, Kaba, H., and Nakanishi, S. (1995). Synaptic mechanisms of olfactory recognition memory. Rev. Neurosci. 6, Hudson, R., and Distel, H. (1986). Pheromonal release of suckling in rabbits does not depend on the vomeronasal organ. Physiol. Behav. 37, Weiler, E., and Farbman, A.I. (2003). The septal organ of the rat during postnatal development. Chem. Senses 28, Meredith, M. (1986). Vomeronasal organ removal before sexual experience impairs male hamster mating behavior. Physiol. Behav. 36, Meredith, M. (1998). Vomeronasal, olfactory, hormonal convergence in the brain. Ann. N. Y. Acad. Sci. 855, Kindermann, U., Hudson, R., and Distel, H. (1994). Learning of suckling odors by newborn rabbits declines with age and suckling experience. Dev. Psychobiol. 27, Meredith, M. (2001). Human vomeronasal organ function: a critical review of best and worst cases. Chem. Senses 26, Liman, E.R., and Innan, H. (2003). Relaxed selective pressure on an essential component of pheromone transduction in primate evolution. Proc. Natl. Acad. Sci. USA 100, Rodriguez, I., and Mombaerts, P. (2002). Novel human vomeronasal receptor-like genes reveal species-specific families. Curr. Biol. 12, R409 R Rodriguez, I., Greer, C.A., Mok, M.Y., and Mombaerts, P. (2000). A putative pheromone receptor gene expressed in human olfactory mucosa. Nat. Genet. 26,