Apical and basal neurones isolated from the mouse vomeronasal organ differ for voltage-dependent currents

Abstract

The mammalian vomeronasal organ (VNO) contains specialized neurones that transduce the chemical information related to pheromones into discharge of action potentials to the brain. Molecular and biochemical studies have shown that specific components of the pheromonal transduction systems are segregated into two distinct subsets of vomeronasal neurones: apical neurones and basal neurones. However, it is still unknown whether these neuronal subsets also differ in other functional characteristics, such as their membrane properties. We addressed this issue by studying the electrophysiological properties of vomeronasal neurones isolated from mouse VNO. We used the patch-clamp technique to examine both the passive membrane properties and the voltage-gated Na⁺, K⁺ and Ca²⁺ currents. Apical neurones were distinguished from basal ones by the length of their dendrites and by their distinct immunoreactivity for the putative pheromone receptor V2R₂. The analysis of passive properties revealed that there were no significant differences between the two neuronal subsets. Also, apical neurones were similar to basal neurones in their biophysical and pharmacological properties of voltage-gated Na⁺ and K⁺ currents. However, we found that the density of Na⁺ currents was about 2-3 times greater in apical neurones than in basal neurones. Consistently, in situ hybridization analysis revealed a higher expression of the Na⁺ channel subtype III in apical neurones than in basal ones. In contrast, basal neurones were endowed with Ca²⁺ currents (T-type) of greater magnitude than apical neurones. Our findings indicate that apical and basal neurones in the VNO exhibit distinct electrical properties. This might have a profound effect on the sensory processes occurring in the VNO during pheromone detection.

The vomeronasal organ (VNO) is a cul-de-sac chemosensory structure involved in the detection of pheromones in most mammals (Halpern, 1987; Døving & Trotier, 1998). The VNO contains sensory neurones - vomeronasal neurones - that are thought to represent the peripheral detectors of pheromones. Vomeronasal neurones are bipolar cells, with a dendrite reaching the lumen of the VNO and an axon projecting directly to the brain (Brennan, 2001). According to biochemical, molecular and morphological observations, two types of vomeronasal neurones can be recognized in rodents: apical neurones and basal neurones (reviewed in Tirindelli et al. 1998; Dulac, 2000; Biasi et al. 2001). Apical neurones are located near the lumen of the VNO and express the pheromone receptor family V1Rs, along with the G-protein alpha subunit Gαi₂. In contrast, basal neurones, which lie close to the basal lamina, express the V2R receptor family in combination with the G-protein alpha subunit Gαo.

Electrical activity of vomeronasal neurones is essential for transferring pheromone information to central areas. In this context, patch-clamp recordings have provided information on the electrical properties of vomeronasal neurones in rodents. These sensory neurones are endowed with several voltage-gated ion conductances, including TTX-sensitive Na⁺ and TEA-sensitive K⁺ channels, and fire action potentials either spontaneously or in response to current injections (Liman & Corey, 1996; Inamura et al. 1997a; Trotier et al. 1998; Holy et al. 2000; Stowers et al. 2002; Ghiaroni et al. 2003). Stimulation of mouse and rat vomeronasal neurones with specific urine components alters their firing rate (Inamura et al. 1997b, 1999; Holy et al. 2000; Leinders-Zufall et al. 2000; Stowers et al. 2002). Multielectrode recording studies have shown that neuronal firing might code for pheromone signals (Holy et al. 2000; Stowers et al. 2002). Whether apical neurones also differ from basal neurones in their electrophysiological properties is not currently known.

In this study, we addressed the issue of the functional diversification between apical and basal neurones by studying their electrophysiological properties using patch-clamp techniques. In particular, we focused our attention on the voltage-gated Na⁺ and K⁺ currents, because they are primarily important for generating action potentials to the brain. In addition, we studied voltage-gated, transient (T-type) Ca²⁺ currents, which contribute to the electrical excitability of vomeronasal neurones.

METHODS

All experiments were carried out in accordance with the Italian Guidelines for the Use of Laboratory Animals (Decreto Legislativo 27/01/1992, no. 116).

Preparation of isolated vomeronasal neurones

Adult male and female CD-1 mice were used in this study. Vomeronasal neurones were isolated with a standard procedure that we have described previously (Ghiaroni et al. 2003). Briefly, mice were deeply anaesthetized with CO₂ and killed by dislocation of cervical vertebrae. The VNO was removed within its bony encasing and then carefully dissected free of the bone. The sensory epithelium, which is situated at the medial part of the organ, was separated from the lateral non-sensory epithelium, rinsed in divalent-free Tyrode solution (mM: 140 NaCl, 5 KCl, 10 Hepes, 10 glucose, 10 sodium pyruvate, 2 EGTA, pH adjusted to 7.4 with NaOH), and cut into several small pieces. The minced tissue was incubated at room temperature (22-25 °C) for 60-70 min, and gently triturated once every 5 min using a fire-polished Pasteur pipette (for details, see Maue & Dionne, 1987). After centrifugation, the supernatant was replaced with regular Tyrode solution (mM: 140 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 Hepes, 10 glucose, 10 sodium pyruvate, pH adjusted to 7.4 with NaOH) supplemented with DNase (0.1 mg ml⁻¹; type I, Roche Diagnostics, Milano, Italy) and incubated for 15 min with agitation. This was done to dissolve DNA released from broken cells. Finally, about 10 μl of cell-containing supernatant was applied to the bottom of a chamber that consisted of a standard glass slide onto which a silicon ring 1-2 mm thick and with 15 mm i.d. was pressed. The glass slide was precoated with Cell-Tak (≈3 μg cm⁻²; BD Biosciences, Bedford, MA, USA) to improve adherence of isolated cells to the bottom of the chamber. The chamber was left undisturbed for about 20 min at room temperature to allow the cells to settle from the suspension and adhere to the glass slide.

The recording chamber was placed onto the stage of an Olympus inverted microscope (model IX70; Tokyo, Japan) and isolated neurones were viewed with Nomarski optics. During the experiments, cells were continuously perfused with Tyrode solution by means of a gravity-driven system. Drugs were dissolved in modified Tyrode solution to maintain osmolarity. All chemicals, unless otherwise stated, were from Sigma (Milano, Italy). Before recording, bright field and Nomarski images of neurones were taken with an Olympus camera.

Whole-cell recording

Membrane currents of single vomeronasal neurones were studied at room temperature by whole-cell patch-clamp (Hamill et al. 1981), using an Axopatch 1D amplifier (Axon Instruments, Union City, CA, USA). Signals were recorded and analysed using a Pentium computer equipped with Digidata 1320 data acquisition system and pCLAMP8 software (Axon Instruments). pCLAMP8 was used to generate voltage-clamp commands and to record the resulting data. Signals were prefiltered at 5 kHz and digitized at 50 or 100 μs intervals.

Patch pipettes were made from borosilicate glass capillaries (Garner Glass Company, Claremont, CA, USA) on a two-stage vertical puller (model PB-7, Narishige, Tokyo, Japan). The standard pipette solution contained (mM): 120 KCl, 1 CaCl₂, 2 MgCl₂, 10 Hepes, 11 EGTA, 2 ATP, 0.4 GTP, pH adjusted to 7.3 with KOH. In some experiments, KCl was replaced by an equal concentration of CsCl. To reveal voltage-gated Ca²⁺ currents, we used the following bath solution (Barium) (mM): 20 Ba²⁺, 120 N-methyl-D-glucamine (NMDG), 10 TEA, 10 Hepes, 10 glucose, 10 pyruvic acid, pH adjusted to 7.4 with methanesulfonic acid. We also used the following as pipette solution (Caesium) (mM): 120 CsOH, 1 Ca(OH)₂, 2 MgSO₄, 120 methanesulfonic acid, 10 Hepes, 11 EGTA, 2 ATP, 0.4 GTP, pH adjusted to 7.3 with CsOH. Thus, calcium currents were recorded in Na⁺/K⁺/Cl⁻-free conditions.

Pipette resistances typically were 5-7 MΩ when filled with intracellular solution. The access resistance of the patch pipette tip was estimated by dividing the amplitude of the voltage steps by the peak of the capacitive transients (from which stray capacitance had been subtracted). Typical values were about 10-15 MΩ. Leakage and capacitive currents were not subtracted from currents under voltage clamp, and all voltages have been corrected for liquid junction potential (LJP, ≈4 mV) measured between KCl (or CsCl) pipette solution and Tyrode (bath) solution (Neher, 1992). LJP correction was not applied when we used the solutions for revealing Ca²⁺ currents. The large concentration of divalent cations in the extracellular solution, in fact, would probably cause a shift of the current activation in the positive direction (‘surface charge screening’; Hille, 2001), therefore counteracting LJP. For example, increasing external Ca²⁺ concentration from 2 to 10 mM shifts the apparent gating of T-type channels by about 10 mV (Perez-Reyes, 2003).

Input resistance of vomeronasal neurones was measured as the slope of the linear current-voltage (I-V) relationship around -80 mV. Cell membrane capacitance was measured by integrating the capacitive current transient during application of a 10 mV voltage step from a holding potential of about -80 mV (Bigiani & Roper, 1993).

Analysis of electrophysiological data

Statistical analysis was done with Prism 3.03 software (GraphPad Software, San Diego, CA, USA). Results are presented as means ± S.E.M. Data comparisons were made with a two-tailed independent t test for Gaussian distributions. We also used the Mann-Whitney (non-parametric) test for comparing the medians of skew distributions. The significance level was taken as P < 0.05. Distributions of ion current magnitude are displayed in the form of box plots. In these plots, boxes show the middle half of the data (approximately the 25th and 75th percentiles) and the horizontal line marks the median, whereas the ‘whiskers’ extending from the top and bottom of the boxes show the main body of the data. Outliers or extreme values are plotted individually with circles.

Steady-state inactivation curves for voltage-gated sodium currents were obtained by fitting the data with a Boltzmann equation:

I/I_max = 1/(1 + exp[(V - V_0.5)/k]),

where I/I_max is the current elicited during a test pulse and normalized to the maximal current, V is the voltage at which the membrane was held for 300 ms before the test pulse, V_0.5 is the membrane potential at which the current is 50 % inactivated, and k the slope.

Immunohistochemistry of isolated neurones

To distinguish apical from basal neurones after isolation, we performed immunocytochemistry to assess the expression of V2R₂, a putative pheromone receptor expressed in basal neurones only (Martini et al. 2001). Glass-attached vomeronasal neurones were gently perfused with 4 % paraformaldehyde, washed in phosphate-buffered saline solution (PBS) and air dried. Cells were blocked in 1 % albumin, 0.3 % Triton X-100 in PBS for 20 min and incubated with the anti-V2R₂ antibody (1:1000) (Martini et al. 2001). Specific immunoreactivity was detected by the biotin-avidin-horseradish peroxidase-diaminobenzidine method (ABC kit) as recommended by the supplier (Vector, Burlingame, CA, USA). Bright field and Nomarski images were taken with a digital Olympus camera.

Morphometric measurements

Images of live or immunostained neurones were analysed with a Macintosh computer equipped with a public domain image processing and analysis software (NIH Image, version 1.61). Neurones were displayed on the computer screen at a final magnification of × 3000. This allowed us to follow accurately all the twists and turns of dendrites with the freehand line option of the NIH Image software. The actual length of the dendrite was measured and used as morphometric index to identify apical vs. basal neurones. Distribution of dendrite length among neurones was displayed in the form of a box plot (see above).

In situ hybridization of tissue slices

Mouse Na_V1.1, Na_V1.2 and Na_V1.3 gene sequences were obtained by a basic local alignment search tool (BLAST) search with the rat orthologues against the mouse genome database. The untranslated regions (UTRs) of the mouse Na_V1.1, Na_V1.2 and Na_V1.3 channels were aligned to identify stretches of sequence with negligible identity. Primers were designed accordingly:

AGAGTGCAAGCTGTCCACCGGAAA Na_V1.1 UTR, 5′

AGTGATTGTGATATCAACCTGAAG Na_V1.1 UTR, 3′

Predicted product: 438 bp

AATGAGATCTATTTTCAATGAAGG Na_V1.2 UTR, 5′

TTACCAAGAGCAGAAGATGGCTAA Na_V1.2 UTR, 3′

Predicted product: 532 bp

TGTTTACAGCCTCTGAAGGTAAAG Na_V1.2 UTR, 5′

TGATTACTGGAGAAACTTGTGGTC Na_V1.2 UTR, 3′

Predicted product: 452 bp

PCR was performed on genomic DNA of CD-1 mice and the products were cloned into pGEM-T easy (Promega) and subjected to sequence analysis. RT-PCR was also performed on cDNA from vomeronasal tissue with the same oligonucleotides.

Tissue was obtained from adult CD-1 mice. Frozen sections were cut at 16 μm and attached to slides coated with silane (Sigma). Digoxigenin-labelled cRNA probes and in situ hybridization procedures were essentially as described (Schaeren-Wiemers & Gerfin-Moser, 1993) except that sections were treated with 1 % Triton X-100 before acetylation.

RESULTS

Identification of apical and basal neurones after isolation

According to their position in the neuroepithelium, one would expect isolated apical neurones to have a dendrite shorter than that of basal neurones. However, the thickness of the neuroepithelium (e.g. Vaccarezza et al. 1981; Giacobini et al. 2000) and of the neuronal layers (e.g. Martini et al. 2001) is not uniform across the VNO. To have an estimate of the range of dendrite lengths for both neuronal populations we immunostained isolated neurones with an antibody raised against V2R₂ (Martini et al. 2001), a putative pheromone receptor expressed in basal neurones only (Fig. 1A). We then measured the dendrite length in both isolated V2R₂-positive and -negative neurones. The analysis of the distribution of the length values obtained from 129 neurones revealed that more than 75 % of cells with a dendrite > 40 μm were immunopositive for V2R₂ (Fig. 1B and C). In contrast, more than 75 % of isolated neurones with a dendrite < 40 μm were immunonegative for V2R₂ (Fig. 1C). Thus, we used 40 μm dendritic length as a morphometric index to classify apical neurones vs. basal neurones in electrophysiological experiments. Before recordings, the dendrite length was measured and the neurone assigned to one of the two categories. We expected to observe significant differences in the electrical properties, if they occurred, by analysing a fairly large number of cells. In this study we present electrophysiological data from 122 putative apical neurones (hereinafter, apical neurones) and from 120 putative basal neurones (hereinafter, basal neurones). Consistent with previous reports (e.g. Liman & Corey, 1996; Moss et al. 1998), the axon was usually not present in isolated neurones; it was probably lost during the dissociation procedure (see also Ghiaroni et al. 2003).

Figure 1. Cellular distribution of V2R₂ in the mouse vomeronasal organ.

A, coronal section of the VNO neuroepithelium showing localization of immunoreactivity for V2R₂. Anti-mouse V2R₂ stained all neurones in the basal half of the neuroepithelium (arrow). *, VNO lumen. Scale bar: 80 μm. B, an isolated vomeronasal neurone stained for the V2R₂ receptor. The positive immunoreactivity indicates that this is a basal neurone. Its dendrite is 61 μm long. Scale bar: 15 μm. C, box plot of the distribution of dendritic length in V2R₂-positive neurones (V2R₂+) and in V2R₂-negative neurones (V2R₂-). Neurones with dendrites > 40 μm are mainly basal neurones, whereas those with dendrites < 40 μm are mainly apical neurones (n = 129 neurones).

Passive membrane properties

The electrical properties of the apical and basal neurones were examined by whole-cell patch-clamp recordings. As previously reported by Dean et al. (2003), we did not observe any significant differences between sexes as to the electrical properties of vomeronasal neurones.

In the first series of experiments, we determined the passive membrane properties of apical and basal neurones. With KCl pipette solution, the zero-current potential (V₀, an estimation of the cell resting potential in patch-clamp experiments) did not differ significantly between the two cell populations (-55 ± 2 mV in 32 apical neurones; -57 ± 2 mV in 33 basal neurones). Also the input resistance (an estimation of the cell membrane resistance) did not change significantly from one group to the other (3.19 ± 0.41 GΩ in 27 apical neurones; 3.22 ± 0.36 GΩ in 20 basal neurones).

We finally measured the cell membrane capacitance (C_m) of isolated neurones. We found that C_m was 5.02 ± 0.57 pF (n = 30) in apical neurones, whereas in basal neurones it was 5.47 ± 0.74 pF (n = 15). This difference was not statistically significant. This may seem an odd result given the longer dendrite in basal neurones compared to apical neurones. However, variation in cell body size may occur among neurones (Vaccarezza et al. 1981; Naguro & Breipohl, 1982), therefore causing variability in the extension of the membrane surface to be capacitively charged during a 10 mV voltage step.

Voltage-gated sodium and potassium currents

To study the membrane currents underlying the spiking activity of vomeronasal neurones (Liman & Corey, 1996; Inamura et al. 1997a), we depolarized the membrane under whole-cell voltage-clamp. We chose a holding potential of -84 mV (LJP corrected) as a standard reference potential so that membrane currents from different neurones could be compared. Electrophysiological recordings demonstrated that all the apical and basal neurones possessed both voltage-gated sodium current (I_Na) and voltage-gated potassium current (I_K). However, we noticed that I_Na in apical neurones tended to be larger than that in basal neurones (Fig. 2).

Figure 2. Voltage-gated Na⁺ and K⁺ currents in apical and basal neurones.

A, differential interference contrast photomicrographs (left) of an apical neurone isolated from mouse VNO. Note the large dendritic knob. Dendrite length: 16 μm. Patch-clamp recording from this neurone (right) revealed the presence of TTX-sensitive, voltage-gated Na⁺ currents (downward deflection in the current records) and of TEA-sensitive, voltage-gated K⁺ currents (delayed, upward deflections in the current records). The cell was held at -84 mV and stepped in 10 mV increments from -74 mV to +46 mV. Pipette solution: KCl. Bath solution: Tyrode solution. B, a typical basal neurone with a long dendritic process (78 μm) expressing both TTX-sensitive, voltage-gated Na⁺ currents and TEA-sensitive, voltage-gated K⁺ currents. Voltage protocol and solutions as in A. Note that K⁺ currents were of similar amplitude in both cells. In contrast, Na⁺ currents were larger in apical neurones than in basal ones. Scale bar: 15 μm.

We then measured the peak amplitude of I_Na and evaluated its variation among neuronal subsets. As shown by Fig. 3A, the majority of apical neurones possessed a large I_Na (> 500 pA), whereas more than 75 % of basal neurones exhibited low amplitude (< 500 pA) I_Na. The median of I_Na amplitude in apical neurones and in the basal ones was 698 and 287 pA, respectively. This difference was statistically significant. In contrast, the distribution of I_K amplitude (measured at a reference potential of +46 mV and at the end of a 40 ms pulse) was similar in the two neuronal subsets (Fig. 3B), as were their median values (521and 484 pA, respectively; difference not statistically significant).

Figure 3. Amplitude distribution of voltage-gated Na⁺ and K⁺ currents (I_Na and I_K, respectively) in apical and basal neurones.

A, distribution of peak values for I_Na reveals that more than 70 % of apical neurones had I_Na > 500 pA. In contrast, basal neurones typically had I_Na of small amplitude. Note that the median (the horizontal line in the middle of each box) is larger in apical neurones than in basal neurones. Cells were held at -84 mV and depolarized, in 10 mV increments, to elicit maximal I_Na. Pipette solution: KCl. Bath solution: Tyrode solution. B, amplitude distribution of I_K. Current amplitude was evaluated at +46 mV and at the end of a 40 ms voltage step. Holding potential: -84 mV. Solutions as in A. Note that both the median and the data range are similar in the two neuronal subsets.

To establish whether the biophysical properties of I_Na differed between apical neurones and basal neurones, we analysed the current-voltage relationship and the steady-state inactivation. Figure 4A shows the I-V plots for I_Na evaluated in the two cell subsets. No differences could be detected in the activation thresholds (about -55 mV to -50 mV) and the voltages at which the current peaked (about -25 mV). Consistent with the observations on the distribution of I_Na amplitude (Fig. 3), however, the mean amplitude of I_Na was significantly larger in apical neurones than in basal neurones (Fig. 4A). For these experiments, it was not necessary to study I_Na in isolation. Na⁺ currents are faster than K⁺ currents in the voltage range of -50 to 0 mV, and the peak value is reached before significant K⁺ current has developed (Fig. 2).

Figure 4. Biophysical and pharmacological properties of voltage-gated Na⁺ currents (I_Na) in apical and basal neurones.

A, current-voltage relationships. I_Na values for each membrane potential (V_m) were averaged within each group (40 apical neurones, 37 basal neurones). I_Na activated at approximately -50 mV and peaked at about -25 mV in both apical and basal neurones. Pipette solution: KCl. Bath solution: Tyrode solution. B, voltage dependence of the steady-state inactivation of I_Na in apical (○) and basal (•) neurones. A standard two-pulse voltage protocol was used for this analysis. Prepulses 300 ms in duration and of variable amplitude (from -134 mV to -24 mV) were applied prior to the test pulse to -34 mV. Neurones were held at -84 mV between trials. The magnitude of the current elicited by the test pulse (-34 mV) was normalized to its maximal value and plotted against the prepulse potential (bottom). Each point represents the mean ± S.E.M. of 13-15 measurements. Data were fitted to a Boltzmann equation. For apical neurones, the half-maximal voltage (V_0.5) was -78 mV and the slope (k) was 11.7 mV. For basal neurons, V_0.5 was -79 mV and k was 11.8 mV. Pipette solution: CsCl. Bath solution: Tyrode solution. Dendrite length range in these experiments: apical neurones, 12-35 μm; basal neurones, 52-96 μm. C, TTX sensitivity of I_Na recorded in apical and basal neurones. Na⁺ current was elicited by a test pulse to -34 mV after hyperpolarizing the membrane to -124 mV for 300 ms (top). Percentage inhibition of the maximal Na⁺ current was evaluated for 15 nM TTX, which is the IC₅₀ estimated previously for all vomeronasal neurones pooled together (Ghiaroni et al. 2003). The histograms represent mean values ± S.E.M. from 14 measurements. Pipette solution: CsCl. Bath solution: Tyrode solution. Dendrite length range: apical neurones, 14-35 μm; basal neurones, 49-102 μm.

To study the voltage dependence of the steady-state inactivation, we used a typical two-pulse voltage protocol (prepulse and test pulse) that allowed the evaluation of the non-inactivated fraction of the sodium current as a function of a prepulse membrane potential. The steady-state inactivation curve from apical neurones was indistinguishable from the one obtained with basal neurones (Fig. 4B).

Finally, the sensitivity of I_Na to TTX did not differ significantly between the two neuronal subsets (Fig. 4C).

We also studied the biophysical and pharmacological properties of K⁺ currents in apical and basal neurones. Figure 5A shows the I-V plot for I_K evaluated in the two cell subsets. No differences could be detected as to the activation threshold (about -40 mV) and the mean current amplitude evaluated at a reference potential of +46 mV (LJP corrected). For voltages more positive than +50 mV, however, the amplitude of I_K tended to be larger in apical neurones than in basal ones. We are unaware of the possible physiological significance of this difference, given that the membrane potential in neurones cannot exceed +40 mV, +50 mV (peak of the action potentials) during their activity (Liman & Corey, 1996; Inamura et al. 1997a; Trotier et al. 1998). Note that, although I_K was not studied in isolation, the contribution of I_Na to the whole current at the end of the 40 ms pulses was negligible due to its inactivation (see Fig. 2).

Figure 5. Biophysical and pharmacological properties of voltage-gated K⁺ currents (I_K) in apical and basal neurones.

A, current-voltage relationships. Amplitude values of I_K for each membrane potential (V_m) were averaged within each group (45 apical neurones ○, 41 basal neurones •). I_K amplitude was measured at the end of 40 ms pulses. I_K activated at approximately -40 mV. B, inactivation of I_K. Currents were elicited by a 400 ms depolarizing step to +46 mV from a holding potential of -84 mV (top). Note the decrease in I_K amplitude during prolonged voltage pulses (inactivation). Inactivation properties of I_K were evaluated by measuring the ratio between the current peak and the current amplitude at 400 ms. The histograms represent mean values ± S.E.M. from 14-15 measurements of this ratio. Dendrite length range in these experiments: apical neurones, 11-35 μm; basal neurones, 52-104 μm. C, TEA sensitivity of I_K recorded in apical and basal neurones. K⁺ current was elicited by a 40 ms test pulse to +46 mV from a holding potential of -84 mV (top). Percentage inhibition of the current amplitude measured at 40 ms was evaluated for 3.34 mM TEA, which is the IC₅₀ estimated previously for the all vomeronasal neurones pooled together (Ghiaroni et al. 2003). In this apical neurone, the reduction was 63 %, whereas in the basal one it was 58 %. The histograms represent mean values ± S.E.M. from 11-13 measurements. Pipette solution: KCl. Bath solution: Tyrode solution. Dendrite length range: apical neurones, 8-28 μm; basal neurones, 50-79 μm.

I_K typically exhibited a slow inactivation during 400 ms depolarization pulses in both apical and basal neurones (Fig. 5B). We evaluated inactivation properties of I_K by measuring the ratio between the current amplitude at 400 ms and the current peak. At about +50 mV, this ratio did not differ significantly in apical vs. basal neurones.

Finally, the sensitivity of I_K to TEA did not differ significantly between the two neuronal subsets (Fig. 5C).

Expression of voltage-gated type III sodium channels in apical and basal neurones

In neurones, voltage-gated sodium channels cluster at high density at the axon initial segment (Caterall, 1981; Wollner & Caterall, 1986), where action potential firing is thought to arise (Aidley, 1998). In isolated vomeronasal neurones, a residual axon stalk that had been tattered during the dissociation procedure was not usually present (Fig. 1B and Fig. 2). This argued against the possibility that the larger sodium current amplitude in apical neurones compared to basal neurones was probably due to longer residual axons, as one might expect because of the greater distance of apical neurones from the basal lamina. Moreover, we carried out a series of recordings from isolated basal neurones in which an axon stalk was clearly evident after dissociation (Fig. 6A). In these neurones, the average amplitude of I_Na was indistinguishable from the one evaluated in basal neurones without any apparent axon stalk (compare Fig. 6B with Fig. 4A).

Figure 6. Voltage-gated sodium current (I_Na) in isolated basal neurones endowed with an axon stalk.

A, differential interference contrast (Nomarski optics) photomicrograph (left) of a basal neurone isolated from mouse VNO. Note the axon stalk (a) emerging from the soma. Dendrite (d) length: 50 μm. Patch-clamp recording from this neurone (right) revealed the presence of small-amplitude, voltage-gated Na⁺ currents (downward deflection in the current records). The cell was held at -84 mV and stepped in 10 mV increments from -74 mV to +46 mV. Pipette solution: KCl. Bath solution: Tyrode solution. Scale bar: 10 μm. B, current-voltage relationship for I_Na recorded from 11 basal neurones (dendrite length range: 46-89 μm). I_Na values for each membrane potential (V_m) were averaged. Pipette solution: KCl. Bath solution: Tyrode solution.

The electrophysiological findings suggested that the difference in I_Na amplitude between apical and basal neurones probably arose from variation in somatic sodium channels rather than axonal ones. Among the several sodium channel α subunits (Na_V) that have been described in mammals (Goldin et al. 2000), types I (Na_V1.1) and III (Na_V1.3) are normally found primarily in neuronal cell bodies, whereas type II channels (Na_V 1.2) are axonal (reviewed in Rasband & Shrager, 2000). Indeed, RT-PCR analysis showed that all three types were expressed in the VNO and that Na_V1.3 (somatic) mRNA was probably more abundant than the other types (Fig. 7A). Moreover, when in situ analysis was addressed to define the pattern of expression of the three Na⁺channel subunits, it appeared that Na_V1.3 was preferentially expressed by apical rather than basal neurones of the VNO (Fig. 7C), thus reflecting a distribution somehow similar to that of Gαi₂ (Fig. 7B). On the other hand, the expression of both Na_V1.1 and Na_V1.2 mRNAs was so weak that we were not able to evaluate any differences between neuronal subsets (data not shown). Thus, these molecular data supported the electrophysiological findings relating to I_Na (Fig. 4A).

Figure 7. Expression of sodium channel type 1.3 mRNA in the vomeronasal epithelium.

A, RT-PCR products of the sodium channel alpha subunits Na_V 1.1, Na_V 1.2 and Na_V 1.3 were amplified from VNO cDNA. Note the relative abundance of Na_V 1.3 compared to the other types. A similar PCR analysis performed on mouse genomic DNA resulted inthe three bands of equal intensity (not shown). B-D, in situ hybridization analysis in serial sections of the VNO with an antisense riboprobe specific for Gαi₂ (B) and Na_V1.3 (C) and with a sense riboprobe for Na_V1.3 (D). Na_V1.3 mRNA exhibited a gradient of expression, with high level in apical neurones and lower level in basal neurones. Scale bar: 100 μm.

Voltage-gated calcium currents

Vomeronasal neurones express two components of voltage-gated calcium currents: a T-type current that activates at -60 mV, and an L-type current activating at about -20 mV to -10 mV (Liman & Corey, 1996). We consistently found evidence for the occurrence of both types of Ca²⁺ currents in isolated vomeronasal neurones (Fig. 8). Unlike L-type current, T-type current (I_Ca-T) is not found in olfactory neurones of several vertebrates (Firestein & Werblin, 1987; Schild, 1989; Trombley & Westbrook, 1991; Miyamoto et al. 1992; Corotto et al. 1996; Eisthen et al. 2000), with the exception of the newt, Cynops pyrrhogaster (Kawai et al. 1996). As Liman & Corey (1996) pointed out, ‘the presence of T-type current in VNO and not in olfactory neurones indicates that it may play a role in defining the distinct electrical properties of VNO neurones’. We then focused our attention on T-type currents and asked whether apical and basal neurones in the VNO differ in their T-type calcium currents. To analyse T-type currents in isolation, we used a distinct property of these currents as compared to L-type currents, namely, their resistance to rundown (Perez-Reyes, 2003). In isolated vomeronasal neurones, L-type currents ran down completely in about 2 min from the beginning of the recording (Fig. 9 bottom). On the other hand, T-type currents were stable for longer periods (Fig. 9 top). Thus, all T-type currents were measured in a time window between 2 min and 2 min 30 s after establishing the whole-cell configuration. Unexpectedly, we found that I_Ca-T in apical neurones tended to be smaller than in basal neurones (Fig. 10A). This finding was confirmed by evaluating the variation in I_Ca-T peak amplitude among neurones (Fig. 10B). The medians of I_Ca-T amplitude in apical neurones and in the basal ones were 82 and 118 pA, respectively. This difference was statistically significant (P < 0.001). We then analysed the I-V relationship for I_Ca-T in the two-cell subsets. As shown in Fig. 10C, no differences could be detected in activation threshold (between -70 mV and -60 mV) or the voltage at which the current peaked (about -30 mV to -20 mV). Consistent with the observations on the distribution of I_Ca-T amplitude (Fig. 10B), however, the mean amplitude of I_Ca-T was significantly smaller in apical neurones than in basal neurones (Fig. 10C).

Figure 8. T-type and L-type Ca²⁺ currents in vomeronasal neurones.

Ca²⁺ currents were recorded with 20 mM Ba²⁺ as the charge carrier. Top panel, T-type currents (arrow) were elicited in response to depolarizations from a prepulse potential of -100 mV to -60 to -30 mV in 10 mV steps are. T-type currents are therefore low-threshold-activated currents. Bottom panel, L-type currents (arrow) were elicited in response to depolarizations to -20 to +40 mV in 20 mV steps from same prepulse potential. L-type currents are therefore high-threshold-activated currents. Note that L-type currents were preceded by the transient currents in these traces.

Figure 9. Calcium current rundown over time.

T-type calcium currents were monitored by stepping the membrane potential from -100 mV to -30 mV (top), whereas L-type calcium currents were elicited by stepping the membrane potential from -100 mV to 0 mV (bottom). Current fraction represents the ratio between the amplitude of the current (arrow) at a given time and the maximal magnitude of the current evaluated just after establishing whole-cell configuration (time = zero). T-type current ○ did not show any remarkable decline in amplitude over time (i.e. was rundown resistant), whereas L-type current • ran down rapidly, disappearing completely in about 2 min. The quick decline of L-type currents allowed us to study T-type currents in isolation.

Figure 10. Voltage-gated, T-type Ca²⁺ current (I_Ca-T) in apical and basal neurones.

A, Ca²⁺ currents recorded with 20 mM Ba²⁺ as the charge carrier. Currents were elicited by depolarizations from a prepulse potential of -100 mV to -80 to +40 mV in 10 mV steps. B, distribution of peak values for I_Ca-T in the two neuronal subsets. Dendrite length range: apical neurones, 7-26 μm; basal neurones, 47-89 μm. Note that the median (the horizontal line in the middle of each box) is larger in basal neurones than in apical neurones. C, current-voltage relationships. I_Ca-T values for each membrane potential (V_m) were averaged within each group (40 apical neurones, 34 basal neurones). I_Ca-T activated at approximately -70 mV to -60 mV, and peaked at about -30 mV to -20 mV. Voltages were not corrected for LJP (see Methods). Pipette solution: Caesium. Bath solution: Barium.

As the name implies, T-type calcium currents are transient, that is, they inactivate quite rapidly (time constant, τ, usually ≈12-30 ms: Perez-Reyes, 2003). Thus, one could argue that the reduced current amplitude in apical neurones could be due to a pronounced current inactivation. We then measured the time constant of the current decay at -20 mV (inactivation rate) after fitting the current time course with a single exponential function, and found that τ did not differ significantly between apical neurones (12.7 ± 0.5, n = 34) and basal ones (13.3 ± 0.5, n = 31).

Although T-type currents were expressed by all vomeronasal neurones we tested, only about 30 % of them (apical or basal) were endowed with detectable L-type currents. This could be due to a frank absence of such currents in some neurones, or to the rapid rundown that characterizes L-type currents in whole-cell recordings (Fig. 9). In those neurones in which L-type current was evident, however, we did not find any difference in current amplitude, measured right at the beginning of recordings, between apical and basal cells (data not shown).

DISCUSSION

The available information on the biology of vomeronasal neurones supports the concept that separate neuronal subsets exist in the mammalian VNO: apical neurones co-express V1Rs and Gαi₂ whereas basal neurones co-express V2Rs and Gαo. According to the structural differences of their receptors, it is likely that basal and apical neurones respond to chemically unrelated pheromonal molecules (reviewed in Dulac, 2000). It is still unknown, however, whether vomeronasal neurones differ in other cellular properties downstream of the activation of G proteins and whether these properties segregate in specific neuronal subsets. In this respect the study of their electrophysiological properties is of relevance, because of their key role in the generation of action potentials and thus in the codification of sensory information. In this study, we have shown that apical and basal neurones differ in the amplitude of their voltage-gated Na⁺ currents. This difference apparently was not dependent on the cell surface area, since both types of neurones exhibited similar C_m. In other words, it was an actual difference in current density. To our knowledge, this is the first demonstration that apical and basal neurones in the VNO differ in specific functional membrane properties in addition to the molecular machinery (V1Rs, V2Rs, Gαi₂, Gαo) involved in pheromone detection. Our data suggest that the variation in Na⁺ current density among neurones probably reflects a variation in the number of Na⁺ channels, assuming that single-channel conductance and their open probability remain unchanged. Indeed, Na⁺ currents were indistinguishable between the two neuronal subsets with regard to certain biophysical and pharmacological properties, suggesting that the basic construction of the channel protein did not differ.

Whole-cell configuration of the patch-clamp technique does not allow one to establish the localization of active channels. However, we also recorded Na⁺ currents in neurones in which the dendritic process had been truncated near the cell body (n = 32 cells; data not shown). Although we could not assign such cells to one or other group, nevertheless the variability of the I_Na amplitude may suggest that we were recording from cell bodies of apical and basal neurones. The electrophysiological data on neurones endowed with an axon stalk further indicated that the difference in I_Na between apical and basal neurones probably arose from Na⁺ channels localized in the cell body membrane. This was supported by in situ hybridization findings on the mRNA expression of Na_V1.3, a Na⁺ channel α subunit that is primarily found in neuronal cell bodies (reviewed in Rasband & Shrager, 2000). As a whole, our data indicate that in apical neurones, or at least some of them, there is a higher level of expression of Na⁺ channels, which is probably linked to a larger magnitude of I_Na.

Since Na⁺ current density sets the excitability of a cell membrane, a lower density of Na⁺ current in basal neurones compared to apical neurones suggests that these two neuronal populations exhibit different capability as to the generation of action potentials. In particular, basal neurones might require stronger depolarization (high threshold) than apical neurones to trigger the Hodgkin- Huxley cycle (Hille, 2001). A recent study by Liman (2003) has shown that vomeronasal neurones possess Ca²⁺-activated, non-selective cation channels (CaNS) that could be involved in amplifying the primary sensory response thought to be mediated by transient receptor potential (TRP) channels. It would be then interesting to establish whether both apical and basal neurones express CaNS. In basal neurones, these cation channels could provide the extra depolarization required to reach the firing threshold.

Extracellular recordings have shown that vomeronasal neurones can fire action potentials spontaneously (Holy et al. 2000; Stowers et al. 2002). According to our findings, apical neurones, by having a high Na⁺ current density and therefore a low threshold, could more easily fire action potentials than the basal ones. It is therefore tempting to speculate that spontaneous activity in the VNO could arise mainly from the apical neurones. This mechanism would have obvious repercussions in signal coding and processing in the vomeronasal system. Further electrophysiological studies are required to test this hypothesis.

In addition to I_Na, vomeronasal neurones differed also in the transient (T-type), voltage-gated Ca²⁺ currents (I_Ca-T). Specifically, the peak magnitude of these currents was, on average, larger in basal neurones than in apical neurones. That is, for I_Ca-T the situation was exactly the opposite of the one found for I_Na (compare Fig. 4A and Fig. 10C). In central neurones, T-type channels appear to be preferentially localized to dendrites (reviewed in Perez-Reyes, 2003). Thus, it is reasonable to suppose that the difference in I_Ca-T between apical and basal neurones in the VNO could be related to the length of their dendrites (Fig. 2). The significance of T-type currents in the physiology of vomeronasal neurones is not yet clear. It has been suggested that the preferential localization of T-type channels on the dendrites may indicate a role in signal amplification (Markram & Sakmann, 1994; Magee & Johnston, 1995; Destexhe et al. 1998; Pouille et al. 2000). In addition, T-type channels can play an important role in the genesis of burst firing and also a pacemaker role (reviewed in Perez-Reyes, 2003). Studies by Kawai and co-workers (Kawai, 1999; Kawai et al. 1996, 2001) have indicated that in the olfactory receptor cells of the newt, T-type Ca²⁺ currents may contribute to enhancing odour sensitivity by lowering the threshold of spike generation. Further studies are required to evaluate in detail the impact of T-type currents on the electrical excitability of apical and basal neurones in the VNO.

Unlike Na⁺ currents and Ca²⁺ currents, voltage-gated potassium current did not show any remarkable variation between the two neuronal populations. Although it is possible that more subtle differences have not been detected by our experimental approach, nevertheless this finding was unexpected, given the relevant role of K⁺ currents in membrane excitability. Such currents, in fact, underlie the firing pattern in excitable cells (Hille, 2001). Of course, we cannot exclude that K⁺ currents may be differently modulated during pheromone transduction in the two neuronal populations. As a whole, our data suggest that variation in membrane excitability between vomeronasal neurones depends markedly on the expression of voltage-gated Na⁺ channels rather than voltage-gated K⁺ channels.

It is possible that subgroups within the main neuronal categories (apical and basal neurones) exist in the VNO. For example, biochemical and molecular studies have indicated that members of the V2R family are expressed in distinct subsets of basal neurones (Ryba & Tirindelli, 1997; Herrada & Dulac, 1997; Martini et al. 2001). Moreover, an electrophysiological study by Inamura et al. (1999) has shown that sensory neurones responsive to different urinary pheromones are segregated in different layers in the rat VNO. Whether the electrophysiological properties change from one subgroup to the other remains unknown.