1Responses of vomeronasal sensory neurons to urine excreted from rats, mice and hamsters were studied by the on-cell patch clamp method in slices of sensory epithelium from female Wistar rats.
2The urine excreted from male and female Wistar rats, male Donryu rats and male C57BL/6 mice induced relatively large responses, while urine from male Sprague-Dawley rats and male Syrian hamsters induced small responses.
3Of the 62 neurons responding to urine, 57 responded to only one of the urine preparations.
4The sensory neurons that responded to the male Wistar urine were localized in the apical position of the epithelium where one type of GTP-binding protein, Gi2α, is selectively expressed. The neurons in the basal position of the epithelium, which express Goα, responded to urine from the other animals.
5This study demonstrates that sensory neurons responsive to different urinary pheromones are localized in a segregated layer in the rat vomeronasal sensory epithelium.
The vomeronasal organ (Jacobson's organ) exists in many vertebrates for receiving pheromones which are sometimes related to sexual and social behaviour (Keverne et al. 1986; Wysocki & Meredith, 1987; Halpern, 1987). Modulation of gonadal functions by urine has been well established in the rodent vomeronasal organ. Urinary compounds of low volatility stimulate the vomeronasal system and provide information that is normally not provided by gustation or olfaction (Wysocki et al. 1980). The pheromone involved in the induction of oestrus in unisexually grouped female mice (the Whitten effect) is different from the one involved in implantation failure (the Bruce effect) (Gangrade & Dominic, 1984). In female rats, pheromones in urine excreted from males and females induce various changes in gonadal functions such as reflex ovulation in the absence of coitus and mounting (Johns et al. 1978), reduction in the oestrous cycle of female rats from 5 to 4 days (Chateau et al. 1976), and oestrous synchrony among females that are living together (McClintock, 1978). These results suggest that the vomeronasal organ receives multiple kinds of urinary pheromones.
The sensory neurons of rats and mice that have cell bodies located in the apical layer and the basal layer of the sensory epithelium express the G-proteins Gi2α and Goα, respectively (Jia & Halpern, 1996). Apically and basally situated sensory neurons have been described as projecting to the anterior and posterior portions of the accessory olfactory bulb, respectively (Jia & Halpern, 1996). In the present study, we measured the responses of vomeronasal sensory neurons of female Wistar rats to urine from male Wistar rats, female Wistar rats, male Donryu rats, male Sprague-Dawley rats, male C57BL/6 mice and male Syrian hamsters by the on-cell patch clamp method and examined the depth of the cell soma in the sensory epithelium of the neurons that responded to these stimuli.
All experiments were carried out in accordance with Guidelines for the Use of Laboratory Animals of the Graduate School of Pharmaceutical Sciences, Hokkaido University.
Preparation of various types of urine
Urine was collected from male and female Wistar rats, male Donryu rats, male Sprague-Dawley rats, male C57BL/6 mice and male Syrian hamsters (normally five animals in each case) in metabolic cages and mixed with charcoal (10 % w/v; Norit SX plus, Wako, Japan) for 10 min at 4°C to reduce the cAMP concentration. No attention was paid to the oestrous cycle of the females. The mixture was centrifuged at 9500 r.p.m. for 15 min. The supernatant was filtered through a polysulfone disc filter (pore size, 0.45 μm; EB-DISK 25, Kanto Chemical Co., Ltd, Tokyo, Japan). The concentration of cAMP in the urine preparation was measured using cAMP radioimmunoassay kits (Yamasa Shoyu Co., Chiba, Japan; Honma et al. 1977). The concentration of cAMP in the charcoal-treated urine was less than 0.5 nm. Urine preparations were further treated with an anion and cation exchanger (AG501-X8, Bio-Rad) to reduce their ionic concentration 4 times: urine preparations were mixed with 20 % anion and cation exchanger for 10–15 min at 4°C and the supernatant was decanted. Finally, the supernatant was filtered through a polysulfone disc. Concentrations of Na+ and K+ in the urine preparations were measured by ion electrodes. Concentrations of Mg2+ and Ca2+ were measured by titration with Xylidil Blue (Dojin, Japan) and Methylxylenol Blue (Dojin, Japan), respectively. Urine preparations were divided and stored at −20°C. As shown in a previous paper (Inamura et al. 1997b), application of a high K+ solution increased impulse frequency in the vomeronasal sensory neurons non-specifically. Therefore, we carefully controlled the K+ concentration of urine used as the stimulating solution. The concentration of urine in the preparation was diluted with deionized water and a K+-free buffering solution that consisted of (mM): 296 NaCl, 4 CaCl2, 20 glucose and 20 Hepes-NaOH (pH 7.4) to reduce K+ concentration to less than 3.9 mM. In most cases, the urine preparations (pH 7.6) were about 1/3 the concentration of the original urine except for experiments measuring dose dependency. Samples from different animals of the same type induced similar responses. The final concentrations of Na+, K+, Ca2+ and Mg2+ in urine used as the stimulating solution were: 155–159.3 mM, 0.2–3.9 mM, 2–2.1 mM and 0–0.2 mM, respectively. A control salt solution containing 160 mM Na+, 3.2 mM K+ and 2.1 mM Ca2+ did not induce any response in vomeronasal sensory neurons.
Observation of oestrous cycle
Fresh vaginal smears of female Wistar rats (3–4 months old) revealed a 4 day oestrous cycle, as described by Mora & Cabrera (1997). The female rats were caged individually, isolated from males, and had free access to rat chow and water ad libitum. Twenty-five millilitres of urine preparation, crude urine or saline was sprayed on the nares of female rats with commercial cosmetic sprays every 10 min for 1 h on four successive days at 2 day intervals. The index of the oestrous cycle was evaluated as follows. In the adult rat, the oestrous cycle was normally 4 days with a continuous 2 day oestrum and a continuous 2 day dioestrum. We defined the index of the oestrous cycle of each day for which the oestrous stage changed from oestrum to dioestrum or from dioestrum to oestrum between 1 and 2 days later as being 1 (Fig. 1C). When the oestrous stage did not change between 1 and 2 days after this day, the index was defined as 0. Thus, the mean index of a rat having a 4 day oestrous cycle is 1. Rats were treated with a single s.c. injection of 2 mg oestradiol benzoate in 300 μl corn oil 12 days before stimulation with urine preparations.
Slice preparation of rat vomeronasal sensory epithelium
Slices of vomeronasal sensory epithelium were prepared from the female Wistar rats as described previously (Taniguchi et al. 1996). No attention was paid to the oestrous cycle of the animals. The rat was deeply anaesthetized by inhalation of saturated ether vapour, and decapitated. The vomeronasal epithelium was quickly removed, cut into slices about 120 μm thick with a vibrating slicer (DTK-1000, D. T. K., Kyoto, Japan) in Tyrode solution at 0°C, and stored at 4°C. Epithelial slices were fixed to the glass at the bottom of a recording chamber, permitting access to cells on the surface of the slice by the patch pipette. The preparations were viewed through an upright microscope (model OPTIPHOT, Nikon, Tokyo, Japan) using a × 40 water-immersion lens.
Data recording and analysis
Patch pipettes with resistances of 5–10 MΩ were made from borosilicate glass capillaries with an inner filament (GD-1.5, Narishige Co., Tokyo, Japan) using a two-stage electrode puller (PP853, Narishige Co.) and then heat polished. Membrane currents were recorded in the on-cell configuration using an EPC-7 patch clamp amplifier (List) or an Axopatch-1D amplifier (Axon Instruments) and stored on video cassette via a digital audio processor. Responses to urine were recorded from 72 out of 217 neurons. All recordings were performed at room temperature. Analysis was carried out on a personal computer using AxoScope software (Axon Instruments). Traces from 10 s before the beginning of the urinary stimulation to the stimulation and those from the stimulation to 15 s after the stimulation were displayed on a monitor. The number of action potentials was counted by eye and an impulse frequency was calculated. We define responders as showing an impulse frequency during the urinary stimulation that was higher than that before the stimulation. In the case of silent neurons, one action potential was considered a response.
Vomeronasal sensory neurons were stimulated with extracellular solutions as follows. Tyrode solution, which was delivered by gravity, was switched to urine by eight electrically actuated valves. A stimulating tube with a lumen 500 μm in diameter was placed under visual control at about 2–4 mm from the neurons to eliminate mechanical effects by changing solutions. The delay due to dead space was 1–6 s, depending on the flow rate (∼360 μl min−1), volume of dead space and distance between the neuron and the tip of the tube.
Lucifer Yellow dialysis
Lucifer Yellow CH was dialysed intracellularly by using the patch pipette filled with 1 % Lucifer Yellow solution as described previously (Taniguchi et al. 1996). After the urinary responses had been measured, the cell membrane under the patch pipette was broken to permit dialysis of Lucifer Yellow into the cell. The specimens were then transferred to the stage of a fluorescence microscope (OPTIPHOT) and observations were made under illumination for excitation of fluorescence.
The vomeronasal epithelium was fixed with 4 % paraformaldehyde and serially cut on a vibratome to a thickness of 50 μm. For immunohistochemical staining sections were first treated with 0.3 % H2O2 for 15 min followed by a wash in phosphate-buffered solution (PBS). After 1 h of incubation in 5 % normal goat serum, the sections were incubated with an antibody to a synthetic peptide fragment of Gi2α (1 : 5000, Wako) overnight at room temperature. Sections were then rinsed in PBS and incubated with biotinylated goat anti-rabbit IgG (1 : 200, Vector, CA, USA) for 1 h. The sections were rinsed again in PBS and incubated with ABC solution (ABC Elite kit, Vector) for 1 h and developed with diaminobenzidine (DAB)-H2O2 (0.05 % DAB and 0.003 % H2O2 in 0.05 M Tris buffer) for 3 min. The sections were then rinsed in water and mounted.
Extracellular Tyrode solution consisted of (mM): 145 NaCl, 2.5 KCl, 2 CaCl2, 10 glucose and 10 Hepes-NaOH (pH 7.4). The control salt solution was prepared as follows. A high concentration salt solution (150 mM NaCl, 300 mM KCl, 1 mM CaCl2, 3 mM MgCl2 and 10 mM Hepes; typical ion concentration in urine) was treated with the ion exchanger in a manner similar to the treatment of urine. The treated high salt solution was then diluted with an equal volume of K+-free buffering solution (296 mM NaCl, 4 mM CaCl2, 20 mM glucose and 20 mM Hepes-NaOH (pH 7.4)). Patch pipettes were usually filled with normal internal solution (mM): 140 KCl, 2.5 MgCl2, 0.5 ATP, 0.5 EGTA and 10 Hepes-KOH (pH 7.2). Lucifer Yellow was dissolved in the internal solution at a concentration of 1 %.
We examined the effects of urine excreted from male and female Wistar rats, male Donryu rats, male Sprague- Dawley rats, male C57BL/6 mice and male Syrian hamsters on female Wistar rat vomeronasal sensory neurons using the on-cell patch clamp method. The urine preparations were treated with charcoal and an ion exchanger to reduce the cAMP and K+ concentrations. It was confirmed that the urine preparations thus obtained contained pheromonal activity as follows. It is known that urine excreted by a male rat restores the oestrous cycle in a female rat whose oestrous cycle has been inhibited by treatment with oestradiol (Mora & Cabrera, 1997). We examined the pheromonal effect of the urine preparation obtained from male Wistar rats on the oestrous cycle (Fig. 1). Figure 1A and B shows typical oestrous cycles in adult female rats. The oestrous cycle, which is nearly 4 days in the adult rat, was stopped by the administration of oestradiol. Application of the urine preparation restored the oestrous cycle after oestradiol administration (Fig. 1B), but application of saline did not (Fig. 1A). The mean index of the oestrous cycle before oestradiol administration was 0.91 ± 0.04 (mean ±s.e.m.; n= 13 rats), while that after oestradiol administration was 0.12 ± 0.04 (n= 20). The mean index of the oestrous cycle after application of the urine preparation (0.28 ± 0.05; n= 11) was significantly larger than that before (P < 0.05, t test), although that after the application of saline (0.15 ± 0.06; n= 10) was similar to that before. The mean index after the application of the crude urine was 0.25 ± 0.06 (n= 15), suggesting that the pheromonal activity of the urine sample treated with charcoal and the ion exchanger was similar to that of the crude urine, and that our urine preparations contain pheromone(s) capable of restoring the oestrous cycle.
Various concentrations of the urine preparation obtained from male Donryu rats were applied to the sensory neurons (Fig. 2A). Application of this urine preparation induced an increase in impulse frequency in the sensory neurons of female Wistar rats. Impulse frequencies after application increased with increasing urine concentration. Figure 2B-D shows the relative magnitude of the responses induced by urine preparations obtained from male Wistar, male Donryu and female Wistar rats, respectively, plotted as a function of urine concentration. Application of these three urine preparations increased impulse frequency in a dose-dependent manner, although the concentration-response relationships varied among the neurons examined. Response frequency decreased at high stimulus concentrations of female urine. Injection of depolarizing currents into rat vomeronasal sensory neurons elicited action potentials (Inamura et al. 1997a). The firing frequency increased as the current intensity was elevated from 0 to 30 pA and decreased at current intensities above 30 pA (K. Inamura, M. Kashiwayanagi & K. Kurihara, unpublished data). It is possible that the decrease in response frequency may be due to a large depolarization induced by a high concentration of female urine.
Urine preparations obtained from male Wistar, Donryu or Sprague-Dawley rats, female Wistar rats, male C57BL/6 mice and male Syrian hamsters were applied to the sensory neurons of female Wistar rats (Table 1). These urine preparations increased impulse frequency in the sensory neurons, but the magnitude of the frequency increases varied among the different urine preparations. Although urine from male hamster induced responses, the magnitude was very small.
Table 1. Increase in impulse frequency in response to various urines tested in female Wistar rats
Male Wistar rat
2.27 ± 0.49
Male Donryu rat
1.15 ± 0.29
Female Wistar rat
0.86 ± 0.33
Male Sprague–Dawley rat
0.48 ± 0.10
2.09 ± 1.01
0.09 ± 0.02
The selectivity of individual sensory neurons for male Wistar, male Donryu and female Wistar rat urine was examined (Fig. 3). Figure 3A shows that application of male Wistar urine increased impulse frequency in this sensory neuron while neither male Donryu urine nor female Wistar urine increased impulse frequency. The neuron shown in Fig. 3B responded to male Donryu urine, but not to either male Wistar urine or female Wistar urine. The neuron shown in Fig. 3C responded to female Wistar urine, but to neither male Wistar urine nor male Donryu urine.
We confirmed that the specific response of a neuron to urine was reproducible. Application of male Wistar rat urine induced action potentials in the sensory neuron illustrated in Fig. 4A, but neither male Donryu urine nor female Wistar urine induced action potentials. Figure 4B shows traces recorded from the same neuron as that in Fig. 4A following a second application of the urine preparations. As before, the neuron responded only to male Wistar urine.
Figure 5 shows response profiles of individual vomeronasal neurons to the different classes of urine. All but five of these sensory neurons responded to only one class of urine. This is in contrast to olfactory neurons, which respond to many odorants with quite diverse molecular structures and odour qualities (Sicard & Holley, 1984; Kang & Caprio, 1995; Kashiwayanagi et al. 1996). The urine preparations used in the present study were treated with charcoal and an ion exchanger but were not subjected to purification; hence they would contain various chemicals besides the pheromone(s). Nevertheless, most single vomeronasal neurons responded to only one of the urine preparations. This suggests that a substance(s) unique to each urine class, and not contained commonly in the urine preparations, induced the response in the neurons. Therefore, it is likely that the response is induced by a strain-, species- and sex-specific pheromone(s) in urine.
Vomeronasal sensory neurons have a dendritic process which reaches the luminal surface of the epithelium. Cell bodies of sensory neurons are located at various depths in the cellular layer of the sensory epithelium (Fig. 6A). Previous studies of marsupial and rodents have shown that the sensory neurons at the apical and basal layers of the sensory epithelium are immunoreactive to anti-Gi2α and anti-Goα antibodies, respectively (Halpern et al. 1995; Jia & Halpern, 1996). Figure 6A illustrates neurons immunoreactive to anti-Gi2α antibody in the apical region of the rat vomeronasal epithelium. Figure 6B shows a neuron dialysed with Lucifer Yellow after recording of the response to male Wistar urine. To map the localization of cell bodies of neurons responsive to different classes of urine and to quantify their distribution, the receptor cell layer was divided into five sublayers according to depth (Fig. 6C). The neuron shown in Fig. 6B is located in sublayer 2 according to our criteria. Localization of the cell bodies of sensory neurons that responded to male Wistar urine was examined in 142 cells (Fig. 6D): 33 and 56 % of neurons with soma located in sublayers 1 and 2 in the upper region responded to male Wistar urine, respectively, while only a small percentage of neurons with somas located in the lower layer (sublayers 4 and 5) responded to male Wistar urine. In contrast, male Sprague-Dawley urine did not induce a response in any of the neurons located in sublayers 1 and 2 of the epithelium, and male Donryu urine and female Wistar urine induced the response in only a small percentage of neurons in these sublayers (Fig. 6E-G). Male Donryu, male Sprague-Dawley and female Wistar urine induced responses in sensory neurons in the lower layer (sublayers 3–5) where the neurons have been shown to preferentially express Goα. Male Syrian hamster urine also selectively induced responses in the lower layers, although male mouse urine induced responses in both upper and lower layers.
The present results clearly indicate that there are at least two segregated laminae within the receptor cell layer of the rat vomeronasal epithelium which have different pheromonal sensitivities. It is likely that the pheromones in male Wistar rat urine selectively induce responses in the sensory neurons expressing Gi2α in the apical layer of the sensory epithelium and similar substances in the male Donryu, male Sprague- Dawley and female Wistar rat urine induce responses in the sensory neurons expressing Goα in the basal layer.
Table 2 shows the mean percentage of neurons that responded to urine in the five sublayers of the receptor cell layer. Sensory neurons of female Wistar rats preferentially responded more to male and female Wistar rat urine (23 % and 16 %) compared with male Donryu and Sprague-Dawley rat urine (12 % and 10 %). Male mouse and male hamster urine induced responses in a small percentage of sensory neurons (6 % and 5 %). There was a correlation between the increase in impulse frequency and the percentage of cells increasing their firing frequency in response to each of the stimuli except for the data obtained using male mouse urine (r= 0.451).
Table 2. Mean percentage of neurons responding to urine in the five sublayers of the vomeronasal sensory epithelium
Response (% neurons)
Male Wistar rat
Male Donryu rat
Female Wistar rat
Male Sprague–Dawley rat
We have observed increased firing in response to urine stimuli in 62 neurons. Of these 62 cells, only 5 responded to more than one stimulus source. The preponderance of neurons responding to conspecific male rat urine were found in the Gi2α positive laminae of the vomeronasal sensory epithelium and the preponderance of neurons responding to conspecific female rat urine and male urine from a different rat strain (Donryu or Sprague-Dawley) were found in the Goα positive laminae of the vomeronasal sensory epithelium. An interesting but perhaps less understandable finding is that female rat vomeronasal sensory neurons also responded to urine from male hamsters and mice.
The present results show that single vomeronasal sensory neurons of female Wistar rats respond selectively to urine from male and female Wistar rats; all 23 neurons that were stimulated by the two types of urine preparation only responded to one type (Fig. 5). The neurons also responded to Donryu rat urine and Sprague-Dawley rat urine; most single neurons responded selectively to the male urine preparation from one strain. Therefore, it is likely that single sensory neurons respond specifically to urine from a particular strain of rats. In addition, the rat vomeronasal sensory neurons responded to urine from male mouse and male hamster. The percentage of neurons that responded to mouse and hamster urine was much lower than the percentage that responded to rat urine. Although the increase in impulse frequency in response to mouse urine was rather large, that for hamster urine was very small. The physiological significance of the finding that the vomeronasal sensory neurons of female Wistar rats respond to urine from different species of animal such as mouse and hamster is unknown. One possibility is as follows. The vomeronasal sensory neurons of female Wistar rats have various types of receptors for pheromones contained in urine from not only male Wistar rats but also various other strains of rats. Pheromones contained in urine from mice and hamsters have similar structures in part to one of the pheromones in rat urine and therefore cause responses in the neurons of female Wistar rats. In this regard, it is interesting to note that not only hamster urine but also rat urine induced morphological changes in glomerular synapses in the hamster accessory olfactory bulb (Matsuoka et al. 1998).
Recently, the receptor mechanisms of pheromonal responses have been studied in the rodent and reptile vomeronasal systems. Dialysis of cAMP into turtle vomeronasal sensory neurons induced inward currents (Taniguchi et al. 1996), but the effects of cAMP on vomeronasal neurons of other animals such as the frog, mouse and rat were negative (Trotier et al. 1993; Liman & Corey, 1996; Sasaki et al. 1999). Dialysis of inositol-1,4,5-trisphosphate (IP3) into rat and turtle vomeronasal sensory neurons induced inward current responses under whole-cell voltage clamp conditions (Taniguchi et al. 1995; Inamura et al. 1997a). An increase in impulse frequency in vomeronasal sensory neurons of female Wistar rats in response to preparations of urine excreted from the male Wistar rat was blocked by phospholipase C inhibitors and an IP3 channel inhibitor (Inamura et al. 1997b). Aphrodisin, a pheromone in the hamster vaginal fluid and seminal fluid, and ES20, a chemoattractant for garter snakes, induced IP3 accumulation in the hamster, porcine and garter snake vomeronasal epithelium (Luo et al. 1994; Kroner et al. 1996; Wekesa & Anholt, 1997). These results suggest that pheromonal information is mediated primarily via an IP3-dependent pathway in the various species tested. Accumulation of IP3 in response to ES20 was GTP dependent. Therefore, it is likely that pheromones are bound to G-protein-coupled receptors (GCRs).
Two families of GCRs (vomeronasal GCRs) unrelated to olfactory GCRs have been cloned from the rat and mouse vomeronasal epithelium (Dulac & Axel, 1995; Matsunami & Buck, 1997; Ryba & Tirindelli, 1997; Herrada & Dulac, 1997). Each family is composed of about 100 subtypes of GCRs. In situ hybridization studies suggest that a single sensory neuron expresses one type of GCR. It is worthwhile to discuss the present results in relation to GCRs. The present results show that the vomeronasal sensory neurons of female Wistar rats respond to urine from male Wistar, male Donryu, male Sprague-Dawley and female Wistar rats. The pheromones contained in the rat urine have not been identified, but they appear to induce different types of behaviour. Furthermore, a pheromone(s) that induces one type of behaviour may be composed of several components. For example, in the case of urine from male Wistar rats, the neurons that respond to the urine may possess multiple GCRs that are responsive to different components of the urine. Each neuron may respond to only one component of the pheromone in this urine, and the simultaneous firing of multiple neurons responsive to different components of the pheromone may result in an appropriate response by the animal to the pheromonal cue.
Of the two identified families of vomeronasal GCRs, one family consists of sensory neurons expressing Gi2α in the upper layer of the epithelium (Dulac & Axel, 1995) and the other family consists of neurons expressing Goα (Matsunami & Buck, 1997; Ryba & Tirindelli, 1997; Herrada & Dulac, 1997). It is possible that the response to the male Wistar rat urine is induced via the former type of GCR and responses to the female Wistar and the male Donryu rat urine are induced via the latter type of GCR.
The present results show that vomeronasal sensory neurons of female Wistar rats respond not only to urine from male and female Wistar rats but also to urine from male Donryu and Sprague-Dawley rats. About 80 % of the neurons that responded to male Donryu urine or male Sprague-Dawley urine responded only to male Donryu urine or male Sprague-Dawley urine. Therefore, it is likely that responses of sensory neurons to urine from other strains were highly selective. In addition, rat vomeronasal sensory neurons responded to male mouse urine and male hamster urine. Although the percentage of neurons responding to mouse and hamster urine was lower than that responding to rat urine, and the increase in impulse frequency in response to hamster urine was lower than that for rat urine, sensory neurons selectively responded to mouse and hamster urine. Matsuoka et al. (1998) reported that not only hamster urine but also rat urine induced morphological changes in glomerular synapses in the hamster accessory olfactory bulb. These results suggest that vomeronasal sensory neurons of rodents respond to urinary pheromones and/or urinary pheromone-like substances of different species, although the physiological significance is unclear.
This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture, Japan and the Human Frontier Science Program. K. I. is a research fellow of the Japan Society for the Promotion of Science.