SUMMARY
Social behaviors of most mammals are profoundly affected by pheromones. Pheromones are detected by G-protein coupled receptors in the vomeronasal organ (VNO). To investigate the role of Gαq/11 in vomeronasal signal transduction pathways, microvillar membranes from murine VNO were prepared. Incubation of such membranes from prepubertal females with adult male urine results in an increase in production of inositol-(1,4,5)-trisphosphate (IP3). This stimulation is mimicked by GTPγS, blocked by GDPβS and is tissue specific. Furthermore, use of bacterial toxins such as pertussis that lead to ADP-ribosylation of the G-protein alpha subunits of Go and Gi2 do not block the increase in IP3 levels but U-73122, a PLC inhibitor, blocks the production of IP3. Studies with monospecific antibodies revealed the presence of three G-proteins, Gαo, Gαi2and Gαq/11-related protein, in vomeronasal neurons,concentrated on their microvilli. Our observations indicate that pheromones in male urine act on vomeronasal neurons in the female VNO via a receptor-mediated, Gαq/11-protein-dependent increase in IP3 levels.
Introduction
Responding to chemical signals is essential for the survival and reproduction of most organisms. The nasal cavity of mammals contains two independent sets of olfactory systems, the main olfactory system and the vomeronasal system. Olfactory neurons of the main olfactory system are located in the posterior recess of the nasal cavity and detect a large variety of small volatile chemicals that are carried by the air during breathing. This information is then processed in the main olfactory bulb (MOB) and in multiple cortical centers of the brain, leading to the cognitive and emotional responses to smell. Pheromones are mainly detected in the vomeronasal organ(VNO, a bilateral sensory structure found at the base of the nasal septum in mammals). VNO neurons send axonal projections to the accessory olfactory bulb(AOB) located dorsal to the most posterior portion of the MOB. Axons of AOB neurons indirectly project to specific nuclei of the amygdala and of the ventromedial hypothalamus that are involved in reproductive and aggressive behaviors (Winans and Scalia,1970; Wysocki and Meredith,1991; Kaneko et al.,1980). Information coded from chemical communication cues detected by the VNO can thus readily influence the central control of reproductive physiology and behavior (Meredith,1983; Winans et al.,1982).
The initial events of pheromonal detection require the activation of specific receptors by pheromones and the transduction of the stimulus. Molecular evidence has led to the isolation of three independent families of vomeronasal receptor genes (VR), known as V1Rs(Dulac and Axel, 1995), V2Rs(Herrada and Dulac, 1997; Matsunami and Buck, 1997; Ryba and Tirindelli, 1997) and V3Rs (Pantages and Dulac,2000) that encode putative pheromone receptors. A new classification of vomeronasal receptor genes has been proposed whereby the V1Rs and the V3Rs are consolidated into one family(Rodriguez et al., 2002). These vomeronasal receptors are unrelated to olfactory receptors. It has also been shown that V1Rs and V3Rs are expressed in the same population of VNO neurons, whereas V2Rs are expressed in a distinct subpopulation. These populations are non-overlapping and there is evidence to suggest that individual VNO neurons are likely to express only one receptor gene. Neurons lining the apical half of the VNO neuroepithelium express V1Rs(Dulac and Axel, 1995),whereas neurons in the basal half express V2Rs(Herrada and Dulac, 1997; Matsunami and Buck, 1997; Ryba and Tirindelli, 1997). In mice and rats these neurons also express G-proteins in the axonal projections. V1Rs and V3Rs express the alpha subunit of Gαi2 and project to the anterior region of the AOB, whereas V2Rs express the alpha subunit of Gαo and project to the posterior regions of the AOB(Halpern et al., 1995; Wekesa and Anholt, 1999; Rodriguez et al., 1999). The expression of three types of pheromone receptors supports the idea that they might be involved in detection of different types of chemosensory information.
Vomeronasal chemoreception is mediated by narrowly tuned high-affinity receptors (Leinders-Zufall et al.,2000), and transduction involves release of inositol-1,4,5-trisphosphate (IP3)(Taniguchi et al., 2000; Wekesa and Anholt, 1997; Inamura et al., 1997; Luo et al., 1994). In the garter snake Thamnophis sirtalis, it has been shown that a chemoattractant isolated from its prey induced the generation of inositol-1,4,5-trisphosphate in the VNO(Jiang et al., 1990; Luo et al., 1994). It has been shown that dialysis of IP3 into the turtle Geoclemys reevesii and rat Rattus norvegicus VNO induces inward currents(Inamura et al., 1997; Taniguchi et al., 1995). We have also previously shown that female porcine VNO can be stimulated by male urine and seminal fluid to cause an increase of IP3(Wekesa and Anholt, 1997). These results suggest that pheromonal information is mediated via the IP3-dependent pathway in the vomeronasal receptor neurons.
Although there is some consensus that IP3 is the second messenger in the vomeronasal system, there is no consensus as to which G-protein(s) links receptor activation to the hydrolysis of phosphatidyl inositol bisphosphate (PIP2). In the present study, we looked at the effects of G-protein bacterial toxins and phospholipase C inhibitors on responses to urine in the mouse vomeronasal organ in order to obtain information about the signal transduction pathways. We now provide evidence in support of signal transduction in the mammalian VNO being mediated by a member of the Gαq/11 family of G-proteins via the IP3-dependent transduction pathway.
Materials and Methods
Animals
CD-1 mice Mus musculus L. were originally obtained from Charles River Laboratories (Kingston, NY, USA) and maintained in a breeding colony in the Department of Biological Sciences at Alabama State University. Animals were housed in Institutional Animal Care and Use Committee-inspected and approved facilities and cared for according to the NIH Guide for Care and Use of Laboratory Animals (1997).
Membrane preparations
Adult male urine was freshly collected on a daily basis and stored as samples under argon at -80°C until used. VNOs from female mice, up to 2 weeks old, were dissected from their crevices in the nasal cavity, removed from the cartilaginous capsule, and frozen on dry ice. The tissues were then minced with a razor blade and subjected to sonication for 2-5 min in ice-cold phosphate-buffered saline (PBS). The resulting suspension was layered on a 45%(w/w) sucrose cushion and centrifuged at 4°C for 30 min at 3000 gin a Beckman SW55Ti rotor. The membrane fraction on top of the sucrose was collected and centrifuged as before for 15 min to pellet the membranes. The membranes were resuspended in 100μl of ice-cold PBS. Protein was then determined according to the method of Lowry et al.(1951), using bovine serum albumin (BSA) as standard.
Second messenger assays
Non-hydrolysable forms of G-protein guanosine 5′-O-(3-thiotriphosphate) (GTPγS), inhibitor guanosine 5′-O-(2-thiodiphosphate) (GDPβS) and U-73122 were purchased from Boehringer Mannheim (Indianapolis, IN, USA). Bordella Pertussis Toxin was purchased from Calbiochem (La Jolla, CA, USA). For IP3 assays,reactions were incubated for 1 min at 37°C in 25 mmol l-1Tris-acetate buffer pH 7.2, 5 mmol l-1 magnesium acetate, 1 mmol l-1 dithiothreitol (DTT), 0.5 mmol l-1 ATP, 0.1 mmol l-1 CaCl2, 0.1 mg ml-1 BSA, 10 μmol l-1 GTP and 20 μg VNO membrane protein. Reactions were terminated by the addition of 1 mol l-1 trichloroacetic acid (TCA). IP3 concentration was measured with a kit from Perkin Elmer, Inc.(Boston, MA, USA) according to the manufacturer's instructions, based on displacement of [3H]IP3 from a specific IP3binding protein. In experiments using bacterial toxins, 1 μg ml-1 pertussis (PTX) was incubated with membranes for 1 h. To assess the effect of ADP-ribosylation by pertussis toxin on phosphoinositide hydolysis, experiments were performed by a procedure similar to that described by Schleifer et al. (1980). The toxin was activated by preincubation with 20 mmol l-1 DTT at 37°C for 15 min. The VNO membranes from prepubertal female mice at a concentration of 0.2 mg ml-1 were exposed to 10 μg ml-1 of the activated toxin for 1 h at 37°C in 20 mmol l-1 Tris-HCl, pH 7.6, in the presence of 30 mmol l-1thymidine, 1 mmol l-1 ATP, 0.1 mmol l-1 GTP, 5 mmol l-1 MgCl2, 1 mmol l-1 EDTA, 1 mmol l-1 DTT, 3 mmol l-1 phosphoenolpyruvate, 5 units ml-1 pyruvate kinase, 15 μg ml-1 saponin and 0.2 mmol l-1 NAD. Control membranes were incubated in the same reaction mixture but in the absence of the toxin. Differences between experimental and control animals were analyzed by analysis of variance (ANOVA).
Antibodies
Antibodies against α subunits of G-proteins were obtained from Calbiochem (La Jolla, CA, USA). The antibody against the α subunit of Gi2 was raised against the C-terminal decapeptide KNNLDCGLF of Gi2. The antibody against the α subunit of Go was raised against the C-terminal peptide KNNLKECGLY of Gαo. The antibody against Gαq/11 was raised against the C-terminal peptide. Antibodies reactive with Gαi2 and Gαo do not crossreact.
Western blotting
VNO membrane samples were subjected to electrophoresis on a 10%SDS-polyacrylamide gel, followed by electrophoretic transfer onto a nitrocellulose membrane. Strips of the membrane, containing approximately 20μg protein, were probed with a 1000-fold dilution of normal rabbit serum or 1000-fold dilutions of rabbit antisera against specific G protein subunits(Calbiochem, La Jolla, CA, USA). Bound antibody was visualized via a biotinylated goat-anti-rabbit secondary antibody complexed with avidin and biotinylated horseradish peroxidase (HRP), using Amersham's chemiluminescent ECL detection system (Amersham, Arlington Heights, IL, USA). Migration distances were calibrated against Kaleidoscope prestained molecular mass markers (BioRad, Richmond, CA, USA).
Immunohistochemistry
Mice were given a lethal injection of sodium pentobarbital (50 mg kg-1, i.p.) and perfused intracardially with PBS, followed by extensive perfusion with 10% paraformaldehyde. The nasal cavities were dissected and fixed overnight in 10% paraformaldehyde. Decalcification of nasal tissue was performed for 3 days at ambient temperature using the formic acid—sodium citrate method (Luna,1968). 5 μm thick sections through nasal tissue, formalin-fixed and paraffin-embedded, were deparaffinized in xylene and rehydrated through graded alcohols. The sections were pretreated with 0.1% pepsin (Sigma, St Louis, MO, USA) in 0.01 mol l-1 HCl, pH 2.3, for 20 min to facilitate epitope access. Following pepsin treatment, the sections were blocked for 30 min with BEAT blocking solutions A and B (Zymed Laboratories,San Francisco, CA, USA). They were then incubated with a 250-fold dilution of normal rabbit serum or antiserum against Gαi2,Gαo or Gαq/11 in PBS, 0.05% Triton X-100(Boehringer Mannheim, Indianapolis, IN, USA) overnight at 4°C. Following incubation with the primary antibody, sections were washed extensively in PBS,0.05% Triton X-100, and incubated for 10 min at room temperature with affinity-purified biotinylated goat anti-rabbit antibody. Following a 10 min incubation with HRP-conjugated streptavidin, antibody complexes were visualized using 3′-amino-9′-ethylcarbazole as chromogenic substrate. This generates a red deposit at the site of antibody binding. Sections were counterstained with Hematoxylin, and viewed and photographed under a Zeiss Axiophot microscope. They were then processed using Adobe Photoshop 6.0.
Results
Identification of G-proteins on the microvillar surface of the VNO
Using subunit-specific rabbit antisera against unique peptide sequences of Gαo, Gαi2 and Gαq/11, we confirmed in our microvillar membrane preparation the presence of Gαo, Gαi2 and Gαq/11, as previously reported (Fig. 1). We further investigated whether the Gαq/11 protein is localized to the microvillar surface of the VNO. The apical region of the murine vomeronasal neuron contains a dense group of microvilli, in contrast to supporting cells that carry smaller groups of microvilli(Adams, 1992). Immunohistochemical staining of coronal sections through the VNO with Gαo, Gαi2 and Gαq/11antiserum revealed staining of microvillar tufts at the surface of the vomeronasal lumen (Fig. 2). To verify the specificity of this staining, adjacent sections were incubated either without the primary antibody or with normal rabbit serum at the same concentration. Under these conditions, no staining was detected.
Increase in IP3 levels induced by male urine in VNO membranes from prepubertal female mice
To study transduction pathways activated by pheromonal stimuli from the male, we developed a preparation enriched in microvillar membranes from VNOs of prepubertal female mice. Incubation of microvillar membranes with male urine results in a robust increase in IP3 as compared to the control (P<0.05). This response is mimicked by GTPγS and blocked by GDPβS (Fig. 3). We therefore concluded that female VNO membranes respond to stimuli in male urine with an increase in IP3via a G-protein-coupled pathway. In order to determine whether the production of IP3 was mediated by phospholipase C, we incubated our membranes with a highly selective, cell-permeable inhibitor of PLC (U-73122, 10 μmol l-1). We observed that incubation of VNO membranes with a PLC inhibitor blocked the production of IP3 (P<0.05). In order to determine which G-protein mediated the activation of PLC, we incubated our VNO membranes in pertussis toxin before stimulation. We observed that incubation of VNO membranes with PTX in the presence of male urine still resulted in an increase of IP3 levels. This response was not different from that of the control membranes, which were only stimulated with male urine in the absence of PTX. We concluded that pheromonal stimulation of VNO membranes results in an increase of IP3 levels viaactivation of PLC by the alpha subunit of the Gq/11 class of G-proteins.
Discussion
Several studies using immunohistochemical techniques or in situhybridization have shown that there are three G-proteins concentrated on the microvillar surface of the VNO, Gαi2, Gαoand Gαq/11 (Wekesa and Anholt, 1997; Berghard and Buck, 1996). Roles for activation of phosphoinositide hydrolysis have been reported for Gαi2(Brandt et al., 1985; Ohta et al., 1985) and for Gαo (Moriarty et al.,1990; Blitzer et al.,1993). Gαi2 has been implicated in inhibition of adenylyl cyclase (Wong et al.,1992; Watkins et al.,1992; Taussig et al.,1993), potassium channel activation(Yatani et al., 1988; Kobayashi et al., 1990) and modulation of calcium channels (Linder et al., 1990). The principal functions of Gαo appear to be inhibition of neuronal calcium channels(Linder et al., 1990; Kleuss et al., 1991) and activation of potassium channels (Van Dongen et al., 1988; Kobayashi et al., 1990). By contrast, Gαq and Gα11 are specialized for only one signal transduction function, namely activation of phospholipase C-β(Taylor et al., 1991; Lee et al., 1992).
In the present study we looked at the effects of phospholipase C inhibitors and G protein toxins in the signal transduction pathways. By use of bacterial toxins, which are commonly used to classify G proteins, we were able to show that VNO microvilli membranes incubated with pertussis toxin (PTX) are capable of being stimulated by urine to produce IP3. Members of the Gαq family, e.g. Gα11 and Gαq, lack a recognition site for ribosylation by PTX(Pang and Sternweis, 1990; Strathmann and Simon, 1990; Blank et al., 1991; Sternweis et al., 1992)whereas Gαi and Gαo have a recognition site and undergo ADP ribosylation in the presence of PTX. Since we do not see any changes in production of IP3 in the presence of PTX, we conclude that the main G-protein involved in the production of IP3 is PTX-insensitive, such as the Gαq/11 class of G-proteins.
The presence of G-protein, recognized by anti-Gαq on western blots of VNO microvillar membrane preparations, indicates that a substrate is present to explain the physiological effect of this antibody and is consistent with the possibility suggested by the ribosylation experiments that the G-proteins are members of the Gαq family. These results are consistent with the observation that Gαqactivates PLC and causes an increase in IP3 in our membrane preparations. The ability to block this response with a PLC inhibitor such as U-73122 (Bleasdale et al.,1990; Inamura et al.,1997) further confirms the Gαq mediated pathway. It has also been shown that when U-73122 is applied to the neuroepithelium it blocks spiking in response to urine (Holy et al., 2000). Gαq and its close relatives activate the β1, β2 and β3-isoforms of PLC and, in contrast to Gαi2 and Gαo, are not inhibited by pertussis toxin (Taylor et al., 1991; Lee et al., 1992). In addition to the α subunits of G proteins, βγ subunits can in some instances mediate inhibition or activation of adenylate cyclase(Tang and Gilman, 1991; Clapham and Neer, 1993) or stimulation of phospholipase C, especially the β2-isoform(Camps et al., 1992; Boyer et al., 1992; Katz et al., 1992; Clapham and Neer, 1993) and we cannot exclude a role for the βγ subunits in mediating vomeronasal signal transduction. However, our experiments are designed primarily to determine the relative contributions of the distinct α subunits of Gi2, Go and Gq/11 to activation of phospholipase C in the VNO.
Pheromonal-induced increases in IP3 levels imply a role for calcium in vomeronasal signal transduction. The phosphoinositide cycle responds to the actions of an agonist at the receptor level by hydrolysis of PIP2, resulting in the generation of IP3 and diacylglycerol (DAG) (Neer and Clapham,1988). DAG enhances the activity of protein kinase C (PKC) by rendering it more sensitive to stimulation by calcium(Nishizuka, 1988), while IP3 stimulates the release of calcium from the endoplasmic reticulum stores (Berridge,1987; Berridge and Irvine,1989). Previous experiments by Leinders-Zufall et al., 2000)using slice preparations showed that pheromone application to VN results in a robust and reproducible increase in calcium levels. These results are consistent with other studies showing that urine stimulation activates a calcium-permeable cation-selective inward current in rat vomeronasal neurons(Inamura and Kashiwayanagi,2000). A role for IP3 signaling pathways is also suggested by the presence of the transient receptor potential (TRP) channel family, TRP2 in the VNO (Liman et al.,1999). This supports the idea that the increases in IP3levels are Gαq related. In Drosophila it has been shown that during the photoisomerization process, rhodopsin, a Gqclass molecule, is activated, which triggers an IP3 signaling cascade and leads to the opening of cation-sensitive channels dTRP and dTRPL. It is possible that pheromonal transduction might act in a similar manner. Although it is unclear how TRP2 is activated, members of the TRP channel family are often coupled to PLC activation and can be gated by products of the PIP2 second messenger cascade. Consistent with these findings, VNO responses to components of urine have been blocked by pharmacological inhibitors of PLC (Holy et al.,2000). Therefore it possible that activation of pheromone receptors by Gαq/11 leads to the stimulation of PLC. Elevation of IP3 or DAG would then lead to the activation of TRP2 channels and membrane depolarization.
Acknowledgements
The authors would like to thank Demetria Rivers for the animal care. We thank Kroshona Tabb, Michal Stallworth and Besstina Lyles for helping with the embedding and sectioning of paraffin tissues. This work was supported by a grant from the National Institutes of Health (NIGMS-MBRS SCORE GM08219) to K.S.W.