INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The vomeronasal (VN) system of garter snakes is particularly well developed and is known to be critical for several species-specific behaviors including prey detection (Burghardt and Pruitt 1975; Graves and Duvall 1985; Halpern 1982, 1992; Halpern and Frumin 1979; Halpern and Kubie 1994; Kahmann 1932; Kubie and Halpern 1979; Naulleau 1965; Wilde 1938). The initial step in prey detection involves the interaction between a prey-derived chemoattractive ligand and specific receptors on the dendritic surfaces of bipolar neurons of the VN sensory epithelium. We have isolated and purified prey extracts and used them as selective chemoattractive stimuli. Among the compounds isolated are EW20 (Wang et al. 1988), a 20-kDa thiol-containing protein, EW3 (Wang et al. 1992, 1993), a low-molecular-weight chemoattractant [both derived from earthworm wash (EW)], and ES20 (Jiang et al. 1990), a 20-kDa glycoprotein derived from electric shock-induced earthworm secretion (ESS). ES20 specifically binds to the VN sensory epithelium in a saturable and reversible manner with a K_d of ~0.3 µM (Jiang et al. 1990) and does not bind specifically to other organs, including brain and main olfactory epithelium. G proteins (G_o, G_i, and G_s) have been immunologically detected in the VN sensory epithelium of garter snakes and ES20 receptors probably are coupled to these G proteins (Luo et al. 1994).

Functional studies have demonstrated that ESS and ES20 evoke depolarizing currents in VN neurons and increase unit activity in the accessory olfactory bulb (AOB) mitral cells, the targets of the axons of receptor neurons of the VN epithelium (Jiang et al. 1990; Luo et al. 1994; Taniguchi et al. 1998, 2000). Binding of ES20 to VN receptors also results in increased levels of inositol 1,4,5-trisphosphate (IP₃), suggesting that this signal cascade pathway may be involved in chemosensory transduction. In contrast, ES20 significantly reduces basal levels of cAMP as well as GTPS- or forskolin-induced high levels of cAMP (Luo et al. 1994). Nevertheless, an adenylate cyclase, AC_VN, has been cloned from a garter snake VN epithelial library, which shows high homology to AC type VI (Liu et al. 1998) and is sensitive to Ca²⁺ regulation (Wang et al. 1997). Thus within the VN sensory epithelium components of two second-messenger systems, AC_VN and phospholipase C (PLC), exist.

In VN neurons, the reversal potential induced by dialysis of IP₃ or its analogue 3-deoxy-3-fluoro IP₃ mimics the reversal currents generated by the chemoattractive stimuli (Taniguchi et al. 2000). IP₃, which is known to mobilize intracellularly sequestered Ca²⁺ via the IP₃ receptor (IP₃R), causes elevations of cytosolic Ca²⁺, which plays multiple functional roles in neurons including signal transduction (Berridge 1993; Clapham 1995). However, these events in VN neurons remain to be elucidated.

Taking advantage in the garter snake of the presence of specific prey extracts as selective chemoattractant stimuli (ESS), in the present study, we disclose mechanisms triggering stimulus-induced cytosolic Ca²⁺ transients associated with chemosensory transduction. Optical recordings were performed in slices from VN neurons selectively loaded with the fluorescent Ca²⁺ indicator Calcium Green-1 from their axonal terminals in the AOB by retrograde transport. We report here that chemoattractants produce initially transient IP₃-related accumulation of [Ca²⁺]_i in dendritic regions via two pathways: a Ca²⁺ release from IP₃-sensitive intracellular stores and, to a lesser extent, a Ca²⁺ influx through the plasma membrane in dendritic regions. Although other mechanisms can secondarily participate in the modulation of these chemoattractant-induced cytosolic Ca²⁺ transients, cAMP seems not to have an important role in the generation these transient Ca²⁺ responses.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Adult garter snakes, Thamnophis sirtalis, of both sexes were obtained from commercial suppliers. They had all resided in the laboratory for 2 months prior to use in the experiments described here.

Loading VN bipolar neurons with Calcium Green 1 using retrograde axonal transport from the accessory olfactory bulb

VN bipolar neurons were loaded with Calcium Green 1 by means of retrograde axonal transport of dye placed in the accessory olfactory bulb (AOB). Snakes were anesthetized with a subcutaneous injection of methohexital sodium (0.5 mg/g body wt; Eli Lilly, Indianapolis, IN). A hole above the AOB was drilled through the skull to expose the AOB bilaterally, and Calcium Green 1 was applied either as crystals or a 5-µl injection of Calcium Green solution injected deep into the glomerular layer. The snakes were allowed to recover for 3-4 days prior to the evaluation of Ca²⁺-related fluorescence signals.

Preparation of VN sensory epithelial tissue slices

The methods are essentially similar to those described previously (Taniguchi et al. 1995, 1996, 2000). In brief, snakes were lightly anesthetized with methohexital sodium prior to decapitation. The vomeronasal neuroepithelium was dissected from the head after carefully removing the bony capsule and mushroom body, mounted onto a carrot block, and cut into slices, ~240 µm thick, with a vibrating slicer (Vibratome 3000, Technical Products International, St. Louis, MO) in snake Ringer solution. This solution consisted of (in mM) 119 NaCl, 4.1 KCl, 2.5 CaCl₂, 1.5 MgCl₂, 15 glucose, 5 Na-pyruvate, and 10 HEPES (pH 7.4). The tissue slice was then mounted onto a small plastic petri dish.

Preparation of ESS

The methods are the same as described earlier by Jiang et al. (1990). Secretions were obtained by passing an electric current from a 9-V battery (20 6-s bursts with an intershock interval of 30 s) through the worms. A yellowish mucus-like secretion was collected and centrifuged, and the supernatant was dialyzed against distilled water.

Preparation of VN sensory epithelial homogenate

Essentially we followed a standard method as previously reported (Luo et al. 1994). Dissected VN epithelia were homogenized in cold buffer (20 mM Tris/HCl, pH 8, 1 mM PMSF, 80 µg/ml DTT, 0.5 µg/ml antipain, 1 µg/ml leupeptin, 1 µg/ml aprotinin, 0.6 µg/ml chymostatin, and 0.6 µg/ml pepstatin) and centrifuged at 500 g for 5 min. The supernatant was recovered and referred to as homogenate.

Monitoring intracellular Ca²+ changes

OPTICAL SETUP. Video imaging of fluorescent changes was used to monitor calcium responses. The general plan of the optical setup is based on standard methods (Cinneli and Salzberg 1991, 1992; Cinelli et al. 1995). Essentially the system consists of an upright epi-illumination microscope (Nikon Epiphot) with a video camera (MV-1070; Marshal Electronics) in the photographic port. Light from a 150-W xenon lamp (Optic Quip 1600) is collimated and rendered quasi-monochromatic by one of several interference filters, focused by means of a quartz-UV-grade condenser (Nikon), and reflected to the preparation by a dichroic mirror. The wavelength for the excitation and emission filters and the dichroic mirror were selected according to the excitation and emission spectra of Calcium Green 1. To improve collection efficiency, fluorescent light from the cells was collected by high numerical aperture (n.a.) water-immersion objectives (×20 or ×40; Fluo; Nikon), which formed a real image on the CCD sensor of the video camera located in the image plane of the microscope. To further improve the sensitivity of this analog camera, image exposures were extended to increase light integration in the CCD sensor wells (Cinelli 1998). When fast acquisition rates were needed, an image intensifier was used to improve the sensitivity of the camera. Fluorescence emission usually remained constant during the experiments. To assure stability of the recordings and to avoid photobleaching effects, the excitation light levels were reduced by neutral density filters until the fluorescence intensity remained constant for 200 s of illumination. No significant levels of autofluorescence were observed in VN neurons, and the drugs, at the concentrations used, did not affect or quench fluorescence levels.

CALCIUM IMAGING TECHNIQUES. Fluorescence measurements of Ca²⁺ levels were performed following standard protocols. Data are reported as fractional changes over background fluorescence levels (F/F_o). Standard procedures for background subtraction and calibrations were used for calibration with solutions of known dye concentration (Tsien et al. 1985). After the experiments, in situ calibrations were performed. Cells were permeabilized with Ca²⁺ ionophores (ionomycin) or membrane solvents (digitonin or saponin). F_max and F_min were determined in Ringer solution (1 mM Ca²⁺) to saturate the Ca²⁺indicator and then by subsequently bathing the cells in low-Ca²⁺ Ringer solution supplemented with 5 mM EGTA. Calcium Green yielded increases in fluorescence signals proportional to Ca²⁺ bound; these levels were directly related to levels of [Ca²⁺]_i. Other terms of the equation were assessed by in situ calibration (see following text).

IMAGE PROCESSING. Images were digitized and stored in real time using a frame grabber board in a Pentium IBM-compatible computer system. Final images were analyzed by applying various digital filterings or convolution algorithms (Cinelli 1998, 2000; Cinelli and Salzberg 1991, 1992; Cinelli et al. 1995). High spatial resolution during image acquisition was necessary to preserve the image details in fine dendrites even when low band-pass spatial filters were used. Background experiments indicated that low band-pass spatial filters were often required to suppress pixel noise from the detector or the image intensifier. High spatial resolutions were also necessary for the application of deconvolution techniques after low band-pass spatial filters. Deconvolution techniques were used for improving focus resolution and obtaining better resolution for visualizing particular cellular compartments such as the dendritic terminals of vomeronasal neurons. Temporal plots of Ca²⁺ transients were obtained from averaged values over 8 × 8 pixel kernels. In the figures, changes over time are illustrated by pseudocolors resulting from subtracting basal levels of [Ca²⁺]_i from those obtained after experimental manipulation.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Retrogradely labeled VN neurons

Using slices of snake VN sensory epithelium, changes in cytosolic Ca²⁺ associated with chemosensory transduction were studied in VN neurons loaded with the Ca²⁺ indicator, Ca²⁺ Green 1. VN neurons were labeled by retrograde transport of this dye from their axonal terminals in the AOB. This method allowed the selective staining solely of mature VN neurons, which are the only cell elements in the vomeronasal organ (VNO) that send their axons to the AOB. Transport of the dye from the injection sites to VN neurons was usually obtained within 72-84 h. Using this approach, we observed that not all VN neurons in a slice preparation were found to be labeled with the fluorescence indicator (Fig. 1A), a condition that probably arose from the rather localized application of the dye in the AOB. An important advantage of this system for obtaining selective retrograde staining is that the VNO and the AOB are in different and separated tissue compartments. This condition prevents the undesirable diffusion of the dye to the VNO and the nonspecific staining of other cell types. We consistently observed in the slices that only mature VN neurons were labeled with the fluorescence indicator (Fig. 1, A and B); no other cellular elements, such as sustentacular cells or immature VN neurons, were labeled. Therefore all of the Ca²⁺ responses evaluated in this study arose exclusively from fluorescence changes in VN neurons sending their axons to the AOB.

View larger version (76K): [in this window] [in a new window]   Fig. 1. Ca2+ transients in retrogradely labeled snake vomeronasal (VN) neurons. A shows a video image illustrating the selective staining of VN neurons with Ca2+ Green after retrograde transport of this dye from their axonal terminals in the accessory olfactory bulb (AOB). Observe the labeling in the cell bodies of mature VN neurons (red arrows) located in the intermediate region of the sensory epithelium. No other cellular elements, such as sustentacular cells or immature VN neurons were labeled. L, lumen; D, dendritic region; S, somata region; BL, basal lamina. Horizontal bar = 20 µm. B: a schematic diagram showing the laminar organization of cellular elements in the snake VNO. As shown in A, the cell bodies of mature VN neurons (R) are located exclusively in the middle epithelial laminae with dendrites projecting toward the epithelial lumen. Adjacent to the basal lamina are basal cells (B); immature VN neurons (N) are located apical to the basal cells, and sustentacular cells (S) are located close to the luminal surface, intermixed with the dendrites of VN neurons. C: the increase in [Ca2+]i recorded in the 1st, 3rd, 5th, and 7th frames after elecrtic shock-induced earthworm secretion (ESS, 2.0 mg protein/ml) application. Records taken from a 16 video sequence (1 frame/s; 120-ms image exposure) showing the spatial distribution of Ca2+ responses. Abbreviations as in A. Optical signals were obtained from single runs, and are coded as follow: green: 5-10%; yellow: 11-15%; orange: 16-20%; red: >20%. In situ calibration performed after membrane permeabilization indicates that a 20% optical signal roughly corresponds to changes in cytosolic Ca2+ levels on the order of 500 nM. D: the time course of the increase in [Ca2+]i over baseline levels following ESS application from the entire sequence (16 frames) as determined from 3 different sites (8 × 8 pixels) shown in the first image in C (a-c). The 1st point in the plots corresponds to basal cytosolic values and the 2nd point to the first response obtained 1 s after stimulus application. Arrows indicate the times corresponding to the 4 images illustrated in C.

In most of our recordings, the use of nonconfocal optics in a rather thick slice preparation (240-300 µm) prevented us from obtaining, with certainty, Ca²⁺ signals at single-cell resolution. In fact, most of our optical recordings represented population responses from clusters of VN neurons. Despite this limitation, it was possible to study the spatial distribution of Ca²⁺ signals in different segments of VN neurons because the garter snake VNO exhibits a distinctive lamination pattern of cellular elements (Fig. 1B). In contrast to the less clear laminar organization of cellular elements found in other species (e.g., Halpern et al. 1995), the cell bodies of mature VN neurons are located exclusively in the middle epithelial laminae. Just above the basal lamina there are unstained cell elements corresponding to basal cells and immature VN neurons. In the present study, the latter were not stained because their axons had not reached the AOB. Sustentacular cells and the dendrites of VN neurons are located in the apical epithelial regions close to the luminal surface. As shown in Fig. 1B, these sustentacular cells were also unstained in our slices because first they lack projections to the AOB and second no diffusion of the dye occurred from the AOB to the VNO. Thus despite the lack of single-cell resolution in our recordings, the lamination pattern of the snake VNO and the specific staining of only VN neurons allowed us to define the source of the optical signals at least in three cellular compartments: the cell body region, the dendritic shaft region, and the apical dendritic region of VN neurons adjacent to the luminal surface. Consequently, as seen in all figure labels, Ca²⁺ responses recorded toward the base and in the middle of the epithelium corresponded to transients predominantly generated in or close to the cell bodies, optical signals from the upper VN epithelial sectors related primarily to Ca²⁺ transients from dendritic shafts, and fluorescence changes adjacent to the VN lumen corresponded to Ca²⁺ responses arising from the most apical segments of VN dendrites.

Resting Ca²+ levels

Changes in cytosolic Ca²⁺ levels were determined with standardized optical-imaging techniques (Cinelli and Salzberg 1991, 1992; Cinelli et al. 1995). Using Calcium Green 1, single wavelength emission measurements of changes in fluorescence intensity represented estimates of [Ca²⁺]_i variations. According to "in situ" calibrations performed at the end of the experiments (see METHODS), baseline [Ca²⁺]_i was equivalent to 60.2 ± 15.4 (SD) nM (n = 12), and cytosolic Ca²⁺ response peaks were in the range of 104-740 nM elevations above resting levels, proportional to the intensity of stimulation. Experiments usually lasted 6-8 h. During this period, [Ca²⁺]_i baseline levels remained steady (usually <70 nM), and no deterioration of the preparations was observed as judged by the similarity of the Ca²⁺ responses obtained throughout the experimental sessions.

Characteristics of chemoattractant-induced Ca²+ responses

To determine whether VN stimulants evoke cytosolic Ca²⁺ changes associated with excitatory responses, VN epithelial slices (240-300 µm) were exposed to prey chemoattractant ESS while [Ca²⁺]_i was measured. As illustrated in Fig. 1C, Ca²⁺ transients in VN neurons elicited by the bath application of ESS ligands consisted of sharp elevations in cytosolic Ca²⁺ levels that usually reached a peak amplitude within 1 s. Response peaks were followed by a brief plateau and then a more prolonged decay phase in which cytosolic Ca²⁺ levels gradually declined to baseline levels within 16-32 s.

Ca²⁺ transients evoked by ESS ligands exhibited a rather widespread distribution. Figure 1, C and D, shows the spatial distribution and time course of typical Ca²⁺ transients evoked by bath application of ESS ligands (2.2 mg/ml). In general, patterns of activity displayed a scattered appearance with a heterogeneous organization in which it was possible to find nonuniform foci of activity distributed in multiple epithelial regions separated by silent sectors. Within each lamina there were important variations in the amplitude and time course of ESS responses, suggesting a different degree of activity among stimulated VN neurons. Usually Ca²⁺ signals from sectors showing the largest amplitudes exhibited the longest durations and lowest thresholds. Following successive stimuli at any given concentration of ESS, the overall organization of these Ca²⁺ responses among different epithelial sectors was rather constant. We interpret this finding as indicating that this heterogeneous pattern represents cumulative responses from dissimilar VN neurons responding in different degrees to the ESS ligands.

ESS stimuli evoked clear dose-dependent Ca²⁺ transients in the concentration range of 0.5-4.5 mg/ml protein. Figure 2, A-C, shows typical changes in kinetics and spatial distributions of Ca²⁺ responses following ESS stimuli applied at different concentrations. The concentration threshold for eliciting detectable Ca²⁺ increases with ESS in our slice preparation was between 0.8 and 1.2 mg/ml protein. These Ca²⁺-dependent fluorescence signals usually exhibited half-saturating peak responses at concentrations on the order of 7 mg/ml protein. In this study, most records were obtained with stimuli in the range of 2.2-4.4 mg/ml protein, and these concentrations are similar to those that evoke clear electrophysiological responses in the VN epithelium (Taniguchi et al. 1998). As a control for the specificity of ESS ligands in eliciting Ca²⁺ responses associated to chemosensory transduction mechanisms, we evaluated the effect of actin which is a behaviorally inactive compound (unpublished observations). In contrast to the clear responses evoked by ESS, the application of actin to the bath solution evoked no detectable Ca²⁺ changes (Fig. 2D) even at relatively high concentrations (4-6 mg/ml).

View larger version (73K): [in this window] [in a new window]   Fig. 2. General characteristics of Ca2+ transients evoked by ESS ligands in garter snake VN neurons. A-C: patterns of activity show 8 consecutive records from 16 frame sequences illustrating [Ca2+]i over baseline levels, just before (1st frame) and following (remaining 7 frames) the application of ESS stimuli at 3 different concentrations (l, 2, and 3.3 mg protein/ml). Note that the luminal surface of the epithelium (L) is to the right, the basal lamina (BL) to the left. D: a similar sequence but with records obtained following the application of actin (3 mg/ml), which was used as a control and produced no detectable Ca2+ changes. Observe that in B and C, the increases in odorant concentrations evoked larger and more widespread responses over different regions of the vomeronasal sensory epithelium. Plots on the right illustrate the corresponding time course of Ca2+ signals from the whole sequence of images (16) at the 3 sites shown in A, 1st image (a-c are indicated by , - - -, and · · · , respectively). In all cases, [Ca2+]i returns to baseline levels after 12-16 s, but higher concentrations produce longer response durations. Optical signals were obtained from single runs (1 frame/s, 120-ms image exposure). The 1st 2 points in the plots correspond to basal cytosolic values and the 3rd point to the 1st response obtained 1 s after stimulus application. Cytosolic Ca2+ levels were coded in pseudocolors as in Fig. 1C and overlapped on a fluorescent image of labeled VN neurons.  in D = 75 µm.

Once the stimulus threshold was reached, Ca²⁺ transients elicited by ESS exhibited relatively sharp response onsets with latency rise times in the range of 500-750 ms according to temporal plots obtained from video image sequences acquired at 250 ms/image (data not shown in the figures). Sharp response onsets were also observed even when ESS stimuli were applied at the lowest concentration (0.75 mg/ml, Fig. 2A, plot). Higher ESS concentrations evoked Ca²⁺ signals that exhibited similar response profiles but larger amplitudes and longer durations. As shown in Fig. 2, A-C, the applications of increasing ESS concentrations evoked a proportional increase in the amplitude and duration of Ca²⁺ transients in all epithelial regions. There were also important changes in the spatial distribution of these responses as ESS concentration increased. At low concentrations (e.g., 1 mg/ml), there was a limited number of activated sites that were located almost entirely in dendritic regions. As ESS stimulus concentration increased, three major effects were observed in the spatial distribution of these responses. First, Ca²⁺ transients spread from dendritic locations to the cell body region of VN neurons where magnitudes and duration comparable to those found in dendritic regions were attained. Second, we observed a relative enlargement of individual foci of activity, involving adjacent epithelial sectors. Finally, new discontinuous foci of activity appeared in previously silent epithelial regions (Fig. 2, B and C). These new sites of activity did not overlap with other active sectors, suggesting the recruitment of a different set of active VN neurons. These results suggest that the newly active VN neurons were probably responding to distinct chemical cues present in ESS ligands that exhibited different response thresholds.

Characteristics of different cytosolic Ca²+ transients in VN neurons

To elucidate possible mechanisms involved in the generation of ESS responses, we first established the general characteristics of different types of Ca²⁺ transients in VN neurons. Figure 3 compares the spatial distribution and kinetics of different types of Ca²⁺ signals evoked in snake VN neurons. Both ESS and caffeine (2-5 mM) applications gave rise to prolonged Ca²⁺ signals, which lasted 20-50 s depending on the stimulus concentrations. These two responses, however, differed in their spatial distribution and onset characteristics. In contrast with ESS responses, caffeine signals were found adjacent to the cell body region, and the pattern of activation was more uniform across different regions. Caffeine signals also exhibited a slower rise time with a more progressive build-up, a characteristic that was evident especially at low concentrations (<2.5 mM; Fig. 3C). High [K⁺]_o elicited Ca²⁺ transients that were relatively rapid in onset and brief in duration compared with other types of Ca²⁺ signals found in snake VN neurons (Fig. 3B). They were found predominantly in the cell body region and, as with caffeine-induced responses, were practically absent in apical dendritic segments.

View larger version (80K): [in this window] [in a new window]   Fig. 3. Spatiotemporal characteristics of Ca2+ signals evoked by ESS (A), high K+ (B), and caffeine (C) in VN neurons. Patterns of activity in A-C correspond to 8 consecutive records from 16 frame sequences reflecting cytosolic Ca2+ elevations following stimulus applications. Right: plots correspond to the time course of Ca2+ signals from the whole sequence at 3 different epithelial sites shown in A, first image (a-c). Observe in all the responses the presence of broad patterns of activity spread throughout different epithelial sectors. Responses, however, had different spatial distributions. Note in A that not all retrogradely labeled receptor cells respond to ESS stimulation with increased Ca2+ signals. ESS signals reached their maximum in multiple, but confined, sectors in VN dendritic regions close to the epithelial lumen. In contrast, high-K+ and caffeine responses are found predominantly in proximity of the cell body region, a location where they attained their maximum peak values and longest durations. Observe also the differences in response kinetics; ESS and caffeine evoked Ca2+ signals having a long time course in comparison with the brief transient elevations evoked by high-K+ depolarization. They also exhibited more complex profiles during their decaying phases that contrast with the rather rapid monotonic decay of K+-elicited Ca2+ signals. ESS and high-K+ responses exhibited signals with sharp rise times, while Ca2+ responses evoked by caffeine have a slower onset, especially in small amplitude responses (see trace a). Optical signals were obtained from single runs (1 frame/s; 120-ms image exposure) and superimposed on a fluorescent image of the VN slice. The 1st point in the plots corresponds to basal cytosolic values and the 2nd point to the 1st response obtained 1 s after stimulus application. Cytosolic Ca2+ levels were coded as described in Fig. 1C. Abbreviations in Fig. 1A. Horizontal bar in C = 25 µm.

ROLE OF VOLTAGE-SENSITIVE CA²⁺ CHANNELS. To evaluate the role of voltage-sensitive Ca²⁺ channels (VSCC) in the generation of cytosolic Ca²⁺ changes, we determined the properties of Ca²⁺ responses elicited by high KCl (100 mM; Fig. 3B). The short duration and monotonic decay of these signals occurred even when high [K⁺]_o was still present in the bath. This finding suggests a relatively rapid inactivation of VSCC in snake VN neurons. The total duration of these cytosolic Ca²⁺ changes was in the range of 3.5-5.5 s, lasting an average of 4.52 ± 1.87 (SD) s (n = 7). Thus the time course of these transients were considerably shorter than those elicited by ESS. Response latencies were also shorter, usually shorter than the fastest time resolution tested in this study (250 ms/image). Response peaks attained fluorescent fractional changes equivalent to 25-32%, values that corresponded to [Ca²⁺]_i elevation on the order of 175-290 nM. These Ca²⁺ signals were reversibly suppressed when extracellular Ca²⁺ was removed from the medium (Fig. 4A). They were also completely blocked after the application of common VSCC blockers such as cadmium (Cd²⁺; Fig. 4B) or cobalt (Co²⁺; 50-100 µM; data not shown), confirming that they resulted primarily from a Ca²⁺ influx through VSCC.

View larger version (26K): [in this window] [in a new window]   Fig. 4. Properties of Ca2+ transients in snake VN neurons evoked by KCl (100 mM)-induced membrane depolarization. Plots a and b correspond to optical signals at 2 different sites in the cell body region of VN neurons. A: the effect of removing external Ca2+ from the medium on the Ca2+ transients; left: control responses; middle: the suppression of all Ca2+ signals following the removal of Ca2+ from the medium (0 [Ca2+]o supplemented with 5 mM EGTA); and right: recovery in normal Ringer solution. B: the action of cadmium, a general VSCC blocker, on the Ca2+ signals elicited by K+ depolarization. Left: Ca2+ signals under control conditions; middle: complete suppression of Ca2+ signals following the application of Cd2+ (50 µM); and right: recovery in normal Ringer solution. C: the effect of depleting ryanodine-sensitive internal Ca2+ stores on K+-induced Ca2+ transients. Left: K+-induced Ca2+ signals under control conditions; middle: K+-induced Ca2+ signals obtained after the depletion of ryanodine stores. The depletion of ryanodine-sensitive stores shortens the duration of Ca2+ transients by selectively reducing the amplitude of the responses during their late decaying phases. Right: illustration of the lack of recovery of the late decaying response phase even after repetitive washings in normal Ringer solution, indicating an irreversible effect of ryanodine. Plots of Ca2+ signals were taken from single runs, 16 image sequences (1 frame/s; 120-ms image exposure). The 1st point in the plots corresponds to basal cytosolic values and the 2nd point to the first response obtained 1 s after stimulus application.

CHARACTERISTICS OF CYTOSOLIC CA²⁺ ELEVATIONS ELICITED BY RELEASE FROM INTRACELLULAR STORES. To determine the characteristics of the Ca²⁺ transients in snake VN neurons that depend on Ca²⁺ release from internal stores, we evaluated the action of caffeine. By readily crossing the plasma membrane in different cell systems, caffeine evokes strong Ca²⁺ release, which depends exclusively on its reversible binding to intracellular ryanodine receptors (Usachev and Thayer 1999; Usachev et al. 1993). We found that relatively low doses of caffeine (2.5 mM) could trigger strong Ca²⁺ transients in snake VN neurons (Fig. 3C). Similar to ESS responses, the amplitude and duration of these signals were dose dependent. Ca²⁺ transients evoked by 2.5 mM caffeine had peak amplitudes similar to responses evoked by 2.2 mg/ml ESS (Fig. 3, A and C). Under these conditions, optical signals exhibited fluorescence increases of 28.5 ± 4.7% (F/F_o; mean ± SD; n = 8) that corresponded to [Ca²⁺]_i elevations in the range of 190-200 nM. But in contrast to the ESS responses, Ca²⁺ transients elicited by caffeine attained their largest amplitudes in the cell body region of VN neurons. As we mentioned before, this distribution was similar to the responses elicited by high-[K⁺]_o applications.

View larger version (22K): [in this window] [in a new window]   Fig. 5. Properties of Ca2+ signals evoked by caffeine in snake VN neurons. A: the characteristics of caffeine responses obtained following the removal of Ca2+ from the medium. Left: corresponds to control responses obtained from a 4 × 4 pixel area in the layer of VN neuron somata. Middle: plots of Ca2+ signals elicted by caffeine (10 mM) following the removal of Ca2+ from the medium (0 [Ca2+]o with 5 mM EGTA). Although no significant change was observed in the initial response (), successive caffeine applications evoked a use-dependent gradual reduction in consecutive Ca2+ signals, as shown in responses obtained after 4 and 8 caffeine applications (- - - and · · · , respectively). Right: the recovery of this Ca2+ transient after restitution of normal [Ca2+]o (2-3 min). B: the effect of depleting internal Ca2+ stores with thapsigargin on caffeine-elicited Ca2+ signals. Left: a control response; middle: traces correspond to the progressive reduction in these Ca2+ responses following 2, 4, and 8 caffeine stimulations after incubation of the slice with thapsigargin (1 µM, 15 min). Observe that on the 8th trial, no detectable signals were obtained, and this suppression was not reversed following repetitive rinses with normal Ringer solution (right). C: Ca2+ signals evoked by caffeine following the depletion of ryanodine-sensitive internal stores. Left: corresponds to a control signal obtained before the application of ryanodine (20 µM, 15 min). Middle: the absence of caffeine-induced Ca2+ transients after cytosolic Ca2+ levels returned to basal levels following ryanodine treatment (see further details in text). Similar to thapsigargin effects, the ryanodine suppression of optical signals did not recover after rinsing the preparation in normal Ringer solution. Tissue slices in B and C were tested in normal snake Ringer solution. Plots correspond to Ca2+ signals from the cell body region of VN neurons taken from a single 16 image sequence (1 frame/s; 120-ms image exposure). The 1st point in the plots corresponds to basal cytosolic values and the 2nd point to the 1st response obtained 1 s after stimulus application.

Role of the IP₃ signaling cascade in the generation of Ca²+ transients

Previous studies have suggested that phosphoinositide turnover leading to IP₃ formation may constitute the second-messenger system mediating chemosensory transduction in the VN system (Holy et al. 2000; Kroner et al. 1996; Luo et al. 1994; Sasaki et al. 1999; Wekesa and Anholt 1997). Thus we were interested in determining whether the activation of this pathway could give rise to cytosolic Ca²⁺ changes similar to those elicited by ESS ligands. In different cell systems, bradykinin (BK) stimulates phosphoinositide turnover, which, in turn, among other actions, can increase production of IP₃ (Kirischuk et al. 1995; Seymour-Laurent and Barish 1995; Verkhratsky and Kettenmann 1996). Thus it has been used rather extensively to evaluate the role of IP₃ in intracellular Ca²⁺ signaling (see Berridge 1993, 1998). To determine the effect of BK in snake VNO and compare it to ESS ligands, we evaluated the actions of BK and ESS stimuli on IP₃ levels in homogenates of the snake VN sensory epithelium. IP₃ production was measured following protocols previously reported (Luo et al. 1994). VN homogenates (50 µg protein) were incubated either with ESS (13 µg) or BK (300 nM) in a reaction solution (500 µl) of 25 nM Tris-acetate, pH 7.6, 5 mM MgAc₂, 0.5 mM ATP, 1 mM DTT, 0.01 mM GTP, 0.1 mM CaCl₂, and 0.1 mg/ml BSA). As a control, distilled water was used instead of ESS or BK. In all cases the incubation time was 1 min. Under these experimental conditions, ESS evoked IP₃ levels equivalent to 210 ± 10 (SD) pmol/mg proteins (n = 4) while BK evoked IP₃ levels equivalent to 255 ± 5.0 pmol/mg proteins (n = 4). These IP₃ increases differed significantly from those obtained under control conditions [75 ± 15 (SD) pmol/mg proteins; n = 4], suggesting that both ESS and BK induce similar phosphoinositide turnover leading to IP₃ formation. To further determine whether these increases elicited by BK correspond to IP₃ changes arising from VN neurons, IP₃ levels were determined also in homogenates from VN preparations previously deafferented from the AOB. These VN preparations lack mature VN neurons because the sectioning of their axons induce their degeneration. In contrast with the intact VNO, in deafferented homogenates, BK and ESS incubations exhibited IP₃ levels that did not differ significantly from those found under control conditions. Thus these data indicate that both ESS and BK increase IP₃, and these increases take place predominantly in mature snake VN neurons.

Ca²⁺ signals evoked by BK applications also exhibited strong similarities with the Ca²⁺ transients elicited by ESS ligands. As shown in Fig. 6, BK stimuli evoked Ca²⁺ transients that consisted of an initial sharp [Ca²⁺]_i rise followed by a brief plateau and then a prolonged declining phase in which cytosolic Ca²⁺ levels slowly returned to basal values, with a halftime in the range of 9-14 s, depending on the stimulus concentration. BK responses were also dose dependent in the concentration range tested in this study (50-300 nM) with amplitudes and durations directly proportional to stimulus intensity. Application of 200 nM of BK evoked peak fluorescent signals attaining fractional changes (F/F_o) equivalent to 25-40%. These values roughly corresponded to elevation in [Ca²⁺]_i on the order of 150-390 nM. Under these conditions, Ca²⁺ signals exhibited a time course in the range of 12-16 s. Like ESS responses, BK signals exhibited relatively sharp rise times even at low concentrations (Fig. 6B), characteristics that differed from the Ca²⁺ responses elicited by caffeine (Fig. 3C). Ca²⁺ transients elicited by ESS and BK also shared a similar spatial distribution with responses distributed in somatic as well as in dendritic sectors (Fig. 6, A and C). This presence of Ca²⁺ signals in dendritic regions, even those adjacent to the luminal surface, differed from the localization of caffeine responses found largely in the cell body region of VN neurons.

View larger version (88K): [in this window] [in a new window]   Fig. 6. Comparison of the spatiotemporal characteristics of Ca2+ signals evoked by ESS and the IP3 agonist bradykinin (BK). A-C: the spatial distribution of Ca2+ signals during the peak amplitudes (left) and the decaying phase of these responses (right, 8th frame). Observe that the Ca2+ signals elicited by BK (A and B) were distributed more evenly among different epithelial sectors than those evoked by ESS (C). Right: plots correspond to the time course of the Ca2+ signals obtained at 3 different locations shown in the 1st image in A. Optical signals were obtained from single runs of 16 image sequences (1 frame/s; 120-ms image exposure). The 1st point in the plots corresponds to basal cytosolic values and the 2nd point to the 1st response obtained 1 s after stimulus application. Cytosolic Ca2+ levels were coded as in Fig. 1C. Abbreviations as in Fig. 1. Horizontal bar in C = 25 µm.

Despite their similarities, we found some differences between the overall distribution of BK and ESS elicited Ca²⁺ signals. Unlike ESS responses, BK patterns of activity were more uniformly distributed across different sectors of the epithelium, and lacked the multifocal appearance characteristic of ESS responses. As illustrated in Fig. 6, even at relatively low concentrations, BK-evoked responses were homogeneously distributed over relatively large epithelial regions. This more uniform distribution probably reflects a lack of selectivity by BK in the activation of VN neurons that have different chemosensory specificities.

Source of Ca²+ signals related to IP₃ production

To further characterize the sources of the Ca²⁺ responses evoked by ESS and BK stimuli, we determined whether these responses depend on a Ca²⁺ release from intracellular stores or a Ca²⁺ influx through the plasma membrane. According to our present results (see preceding text), both BK and ESS stimuli appeared to trigger an IP₃ increase in VN neurons, and this molecule can elicit important cytosolic Ca²⁺ elevations that, in different cell systems, depend entirely on a Ca²⁺ release from internal pools (for review, see Berridge 1998). On the other hand, IP₃ can also evoke cytosolic Ca²⁺ increases through a plasma membrane Ca²⁺ influx via an IP₃-activated cation conductance as has been demonstrated in some invertebrate chemosensory systems (Munger et al. 2000; see also Schild and Restrepo 1998 for review).

BK RESPONSES IN CA²⁺-FREE MEDIUM. To determine whether BK responses in snake VN neurons depend on Ca²⁺ entry through the plasma membrane or a release from internal pools, we first evaluated these responses in [Ca²⁺]_o-free medium. In general, the absence of extracellular Ca²⁺ did not abolish BK responses, suggesting that these signals depend largely on a Ca²⁺ release from internal stores. In all the slices evaluated (n = 6), applications of BK (100-200 nM) in 0 [Ca²⁺]_o medium (supplemented with 2 mM EGTA) evoked Ca²⁺ signals that exhibited similar characteristics and kinetics to control responses. Under these conditions, most Ca²⁺ transients preserved their relatively sharp onsets, magnitudes, and durations.

View larger version (16K): [in this window] [in a new window]   Fig. 7. Characteristics of Ca2+ signals evoked in apical dendritic regions of VN neurons by BK in the absence of [Ca2+]o. A: responses to BK (300 nM) obtained in normal Ringer solution (control) and immediately following removal of Ca2+ from the bath {0 [Ca2+]o with 2 mM EGTA (1st trial)}. Observe the rapid reduction in the peak amplitude of the Ca2+ signal that occurred without delay after removing extracellular Ca2+. B: the effect of repetitive BK applications in Ca2+-free medium. After their initial reduction, Ca2+ responses in dendritic regions maintained rather constant characteristics for a few runs (2-5) and progressively decreased following subsequent BK applications. This cumulative decrease gradually affected both the amplitude and duration of Ca2+ responses until all cytosolic Ca2+ changes become undetectable (usually after 20 runs using this stimulus concentration; bottom trace). C: full recovery shown of BK-elicited Ca2+ signals after the VN slices were returned to normal Ringer solution for 5 min. All Ca2+ signals were taken from single runs of 16 image sequences (1 frame/s; 120-ms image exposure). The 1st point in the plots corresponds to basal cytosolic values and the 2nd point to the 1st response obtained 1 s after stimulus application.

BK RESPONSES AFTER DEPLETING CA²⁺ STORES. To further characterize the component of BK responses that was independent of Ca²⁺ release, we evaluated these signals after the depletion intracellular Ca²⁺ stores. As with Ca²⁺ transients evoked by caffeine, the preincubation of VN slices with thapsigargin (1 µM for 10-15 min) did not affect initial BK responses to a large extent (Fig. 8B). Subsequent BK applications, however, evoked a progressive decrease affecting both the amplitude and duration of these Ca²⁺ signals. These reductions were observed equally in all epithelial regions and depended on the stimulus concentration and the number of previous stimulations. There was, however, some difference in the evolution of these changes as repetitive BK stimuli were delivered. In regions corresponding to the cell bodies of VN neurons, cumulative reduction progressed until all Ca²⁺ responses elicited by BK were abolished. Depending on the stimulus concentration, this occurred usually after 8-20 consecutive applications of BK. In apical dendritic regions, however, responses failed to be completely suppressed. Instead, the reduction progressed until it revealed a small component that remained insensitive to thapsigargin (n = 7) even after multiple BK applications (12-25 trials). This component attained its largest amplitude in locations adjacent to the epithelial luminal surface with a magnitude equivalent to 15-24% of the control responses and a duration of ~4-6 s. The real amplitude and duration of these Ca²⁺ transients, however, could be underestimated in our recordings because it appears to be generated in a highly localized dendritic region that cannot be independently assessed due to the population nature of our optical signals.

View larger version (23K): [in this window] [in a new window]   Fig. 8. Ca2+ signals recorded in apical dendritic regions of VN neurons evoked by BK following depletion of internal Ca2+ stores. A: the effects of depleting internal Ca2+ stores on responses following the preincubation of a VN slice with thapsigargin (1 µM). Left: Ca2+ signals under control conditions. Middle: traces illustrate Ca2+ responses evoked by multiple BK stimuli tested 15 min after the application of thapsigargin; note the reduction, but not total suppression, of these Ca2+ signals. Right: the thapsigargin-resistant portion of the original BK response remained constant after 20 trials. B: Ca2+ signals in dendritic and somatic regions evoked by BK after the depletion of ryanodine-sensitive stores. Left: BK responses under control conditions. Middle: Ca2+ signals obtained after the depletion of ryanodine stores (see protocol in the text). Observe that the Ca2+ signal from dendritic locations shows no detectable changes, while that from the somata region exhibits a significant reduction during its decaying phase. Right: after ryanodine treatment there was no detectable Ca2+ signal evoked by caffeine stimulation, corroborating the full depletion of ryanodine-sensitive stores under this condition. Ca2+ signals obtained from single runs of 16 image sequences (1 frame/s; 120-ms image exposure). The 1st point in the plots corresponds to basal cytosolic values and the 2nd point to the 1st response obtained 1 s after stimulus application.

ESS RESPONSES IN CA²⁺-FREE MEDIUM. The mechanisms involved in the generation of Ca²⁺ signals elicited by ESS ligands were analyzed using protocols similar to those used with BK stimulation. First, we evaluated whether Ca²⁺ transients evoked by ESS ligands depend on Ca²⁺ influx or reflect a release from internal stores. As with BK signals we found that the absence of Ca²⁺ in the medium (0 [Ca²⁺]_o, supplemented with 5 mM EGTA) did not suppress ESS responses. Immediately after the removal of Ca²⁺ from the medium, Ca²⁺ signals in most epithelial locations exhibited kinetics and characteristics similar to control responses. Ca²⁺ signals obtained in apical dendritic regions adjacent to the epithelial lumen, however, exhibited a slight reduction in their peak amplitudes of ~12-20% (Fig. 9B). As observed with reductions in BK responses, the late phase of these responses was less affected. These changes were also fully reversible.

View larger version (17K): [in this window] [in a new window]   Fig. 9. Characteristics of Ca2+ signals in apical dendritic regions of snake VN neurons evoked by ESS ligand in the absence of [Ca2+]o. A: traces correspond to Ca2+ responses elicited by ESS (2.2 mg/ml) under control conditions and immediately after the removal of Ca2+ from the bath (0 [Ca2+]o with 2 mM EGTA). Observe the immediate reduction in the peak amplitude and initial decaying phase of the response. B: superimposed traces illustrate decremental Ca2+ signals evoked by successive ESS applications in VN slices maintained in Ca2+-free medium. Like BK responses illustrated in Fig. 7, following the initial decrease, successive ESS applications evoke a progressive response reduction that affected both the amplitude and duration of Ca2+ signals, until these cytosolic Ca2+ elevations become undetectable (usually after 16-20 runs using this stimulus concentration; bottom trace). C: full recovery of ESS-evoked Ca2+ signals after the VN slices were returned to normal Ringer solution for 5 min. Ca2+ signals were obtained from single runs of 16 image sequences (1 frame/s; 120-ms image exposure). The 1st point in the plots corresponds to basal cytosolic values and the second point to the 1st response obtained 1 s after stimulus application.

ESS RESPONSES AFTER THE DEPLETION OF CA²⁺ STORES. To confirm the characteristics of the components dependent and independent of internal Ca²⁺ release, ESS responses were evaluated after the depletion of these intracellular stores. As with BK signals, following the application of ESS ligands, VN slices preincubated with thapsigargin (1 µM for 10-15 min) evoked Ca²⁺ signals that exhibited a gradual reduction in amplitude and duration, consistent with a progressive depletion of internal Ca²⁺ stores (Fig. 10A). In the cell body region, ESS responses were completely suppressed by multiple ESS applications (8-10 trials), indicating that at this level these responses depend entirely on Ca²⁺ release from internal stores. In apical dendritic regions, however, the depletion induced by thapsigargin failed to completely suppress all Ca²⁺ signals, even following numerous applications of ESS ligands (>20 trials). As with BK responses, there was a thapsigargin-resistant component that was revealed after the disappearance of the overlapping ESS response dependent on intracellular Ca²⁺ release (Fig. 10A). Fig. 10A, right, shows that this ESS-evoked component persisted after numerous stimulus applications (20 trials). This component was, however, reversibly suppressed in Ca²⁺-free medium, supporting the notion that it arises directly from Ca²⁺ influx through the plasma membrane. This remaining ESS response exhibited its largest amplitudes in locations just adjacent to the luminal surface of the epithelium (Fig. 10A, right), similar to the BK-induced response. In this location, the response attained a magnitude equivalent to 15-20% of the full response obtained before the thapsigargin-induced Ca²⁺ depletion. However, as we previously mentioned for the similar component in BK signals, the actual size of this Ca²⁺ influx could be underestimated in our population optical recordings due to the difficulty of accurately determining cytosolic Ca²⁺ levels from localized regions.

View larger version (59K): [in this window] [in a new window]   Fig. 10. Spatiotemporal properties of Ca2+ signals in VN neurons evoked by ESS after the depletion of internal Ca2+ stores. A: the effects on ESS responses of depleting internal Ca2+ stores with thapsigargin (1 µM, 15 min). Left: the image and the plot correspond to Ca2+ signals under control conditions. Middle: traces illustrate the use-dependent reduction in the Ca2+ signals evoked by successive applications of ESS after thapsigargin treatment (15 min). The top image of the right panel illustrates the distribution of Ca2+ signals in a thapsigargin-treated slice after 10 ESS stimulations, and the lower plot (right panel) shows the presence of this component from the apical dendritic region even after 20 trials. Top images correspond to the distribution of activity at the signal peaks, and lower plots illustrate the time course of Ca2+ signals as recorded from the apical dendritic region ( in the 1st image). B: the characteristics of ESS responses after the depletion of ryanodine-sensitive stores. Left: the image and the plots correspond to Ca2+ signals under control conditions; right: image and plots correspond to responses obtained after the depletion of ryanodine stores (20 µM ryanodine; see protocol in the text). The top images illustrate the spatial distribution of responses at their peaks, while the lower plots show the time course and amplitude of Ca2+ signals from somatic (a) and dendritic (b) regions. Observe that Ca2+ signals in dendritic locations (- - -) show no detectable changes, while the amplitude and duration of the response from the somata region () was affected predominantly during its late phase. Optical signals were obtained from single runs (1 frame/s; 120-ms image exposure). The 1st point in the plots corresponds to basal cytosolic values and the 2nd point to the 1st response obtained 1 s after stimulus application. Cytosolic Ca2+ levels were coded in pseudocolors as in Fig. 1C. Abbreviations as in Fig. 1. Horizontal bar in B = 25 µm.

Possible role of cAMP in responses to ESS

In the main olfactory system, odor transduction and associated Ca²⁺ transients rely on the activation of a cAMP pathway. The binding of odorants to olfactory receptors causes a G-protein activation of adenylate cyclase, which, via cAMP, activates a cyclic nucleotide-gated (CNG) channel. The opening of this channel generates an inward transduction current carried by Na⁺ and Ca²⁺, and this influx is largely responsible for the Ca²⁺ transients elicited during odor stimulation. To determine whether similar CNG channel activity participates in a Ca²⁺ influx in snake VN neurons, we evaluated ESS responses after the application of the specific CNG channel blockers LY 83583 and L-cis-diltiazem (LCD), which are known to abolish Ca²⁺-related odor responses in olfactory receptor neurons (ORNs) (Kolesnikov et al. 1990; Leinders-Zufall et al. 1997, 1998). As illustrated in Fig. 11, preincubation of VN slices with either LY 83583 (80 µM; B) or LCD (40 µM; C) for 10-15 min did not affect the general spatial distribution (Fig. 11, B and C) or the kinetics (Fig. 11E) of Ca²⁺ transient elicited by ESS ligands. The amplitudes, latencies, and time courses of ESS responses remained practically unchanged after the application of either of these CNG channel blockers (Fig. 11E). A similar absence of effect was observed in all the slices evaluated (n = 5). Furthermore, no changes were observed even in the dendritic regions adjacent to the luminal surface of the epithelium where we found a component arising from Ca²⁺ influx.

View larger version (52K): [in this window] [in a new window]   Fig. 11. Contribution of the cAMP signaling pathway to the generation of ESS-elicited Ca2+ transients. Cyclic nucleotide-gated (CNG) channel blockers evoked no detectable effects either in the spatial distribution nor the time courses of Ca2+ signals evoked by ESS. Top: images show the spatial distribution of Ca2+ signals elicited by ESS ligands (2.2 mg/ml) under control conditions (A) and after preincubation of the VN slices (10 min) with the CNG channel blockers LY83583 (40 µM; B) and L-cis-diltiazem (LCD, 20 µM; C). Patterns correspond to the amplitude peaks of the Ca2+ responses plots illustrated in D and E. No detectable changes were observed either in the spatial distribution or the time course of ESS responses following