INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Neurons are typically segregated into populations that perform specific functions. In a well-studied motor system, two neuronal populations in the chick ciliary ganglion serve different effector roles, each reflected in a distinct neuronal phenotype (Dryer 1994; Marwitt et al. 1971). Ciliary neurons have large cell bodies and innervate striated muscle fibers in the iris and ciliary body to control the light reflex and visual accommodation. By contrast, choroid neurons are smaller and innervate smooth muscle fibers controlling blood flow in vessels of the choroid coat. The two populations also differ in expression of individual genes. For example, somatostatin immunoreactivity is detected only in choroid neurons (De Stefano et al. 1993; Epstein et al. 1988), while ciliary neurons express more transcript for active agrin isoforms than do choroid neurons (Smith and O'Dowd 1994).

The divergence between ciliary and choroid neuron populations extends to the structure and function of synapses formed on them by preganglionic terminals arising from midbrain neurons (Dryer and Chiappinelli 1987; Martin and Pilar 1963; Ullian et al. 1997). Ciliary neurons receive cholinergic synaptic input from single, large calyciform terminals that produce large, nonsummating excitatory synaptic currents (EPSCs) that, in most cases, decay biexponentially. Choroid neurons receive multiple, small cholinergic boutons, producing smaller EPSCs that summate and decay monoexponentially. For both neuron populations, however, only two nicotinic acetylcholine receptor (AChR) classes contribute to the EPSCs (Ullian et al. 1997; Zhang et al. 1996). -Bungarotoxin (Bgt)-AChRs are responsible for the rapidly decaying EPSC phase and are likely to be important in maintaining reliable ganglionic transmission (Chang and Berg 1999). Bgt-AChRs contain predominantly 7-subunits, are recognized and blocked by Bgt, and are abundant in ciliary ganglion extracts (Blumenthal et al. 1999; Chiappinelli and Giacobini 1978; Conroy and Berg 1995; Pugh et al. 1995). Except for a preliminary report (McNerney and Margiotta 1993), the single-channel properties of Bgt-AChRs on neurons isolated from the ciliary ganglion at any developmental age have not been studied, and the numbers of surface receptors per neuron is unknown. The second major receptor class, termed 3*-AChRs, mediate the slowly decaying EPSC phase and have been estimated to be present at levels ~10-fold lower than Bgt-AChRs (Blumenthal et al. 1999). These receptors contain 3, 5, and 4 subunits and occasionally 2 subunits (Conroy and Berg 1995; Vernallis et al. 1993) and are neither recognized nor blocked by Bgt but can be detected with a number of -subunit specific antibodies, including mAb35 (Conroy and Berg 1995; Vernallis et al. 1993). Previous single-channel experiments revealed two functional subtypes of 3*-AChRs having conductances of 25-30 and 40 pS, both of which were blocked by N-Bgt (Margiotta and Gurantz 1989), a snake toxin that recognizes the same receptors on the neurons as mAb35.

Given the distinct phenotypes of choroid and ciliary neurons, we wondered whether the levels or functional properties of Bgt- and/or 3*-AChRs would also be different in the two cell populations. Thus ciliary and choroid neurons were identified by established differences in cell soma size, and Bgt- and 3*-AChRs on each population were compared. The findings demonstrate that ciliary ganglion neurons express a restricted array of receptor classes and suggest they can regulate the relative densities of functional Bgt-AChRs in a cell-specific manner.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell and substrate preparation

Ciliary and choroid neurons were dissociated from stage 40 (Hamburger and Hamilton 1951) embryonic day 13 or 14 (E13,14) chick ciliary ganglia using collagenase A treatment and mechanical trituration procedures as previously described (Margiotta and Gurantz 1989; Margiotta and Pardi 1995). For electrophysiological and fluorescence measurements, dissociated neurons were plated on coated, 12- or 15-mm-diam glass coverslips (Fisher Scientific, Houston, TX) and for binding studies in 16-mm-diam plastic tissue culture wells, both at densities of 1-4 ganglion equivalents (3,700 neurons/E14 ganglion) (Pilar et al. 1980) per coverslip or well. Before use in experiments, the neurons were allowed to equilibrate at 37°C in recording solution (RS) containing (in mM) 145.0 NaCl, 5.3 KCl, 5.4 CaCl₂, 0.8 MgSO₄, 5.6 glucose, and 5.0 HEPES (pH 7.4) supplemented with 10% heat-inactivated horse serum (RS^+hs) for 2-4 h.

To coat glass coverslips, they were first acid-washed, then treated with poly-D-lysine in 0.13 M borate buffer (pH 8.5), washed four times with distilled water, and air-dried. For coverslips used in fast perfusion experiments, 70-150 kDa poly-D-lysine was applied at 1 µg/ml for 1 min (21-23°C). Using this minimal coating protocol, the neuron somata became loosely attached, allowing them to be lifted above the substrate in subsequent fast-perfusion experiments (see following text). For all other studies, stronger neuron attachment was achieved by coating glass coverslips or tissue culture wells with 300 kDa poly-D-lysine at 0.5-1.0 mg/ml for 12-16 h (4°C).

Neuron size measurements

Ciliary and choroid neurons were identified based on previously established differences in cell body size (Landmesser and Pilar 1974; Pilar and Tuttle 1982; Smith and O'Dowd 1994). To selectively label ciliary neuron cell bodies, ciliary nerve axons were stained with the fluorescent dye 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate (DiI; Molecular Probes, Eugene, OR). Briefly, ciliary nerves from dissected ganglia were draped over a 2-3 mm petroleum jelly (Vaseline) barrier, and the cut ends were placed in distilled water at room temperature for 10 min and then in RS containing 3% DiI (vol/vol; dissolved in 95% ethanol at 1 mg/ml). Cell bodies usually became labeled with DiI within 12-16 h at 37°C. Neurons from DiI-labeled ganglia were dissociated and plated as described above, and examined with fluorescence and DIC optics using a Zeiss Axiophot microscope. Labeled neurons were identified with fluorescence microscopy, and major and minor axis lengths of labeled neuronal cell bodies measured with an eyepiece micrometer using DIC optics. Similar measurements were made on neurons from unstained ganglia. In both cases, average axis lengths were used as an estimate of neuronal diameter, the values compiled into histograms and analyzed using commercial software.

Electrophysiology

WHOLE CELL RECORDING. AChR currents were collected from visually identified ciliary or choroid neurons at room temperature (21-24°C) using fast agonist microperfusion (Jonas 1995) coupled with whole cell recording, as we recently described (Pardi and Margiotta 1999). Briefly, nicotine (20 µM) or ACh (500 µM) was dissolved in RS, and agonist and RS streams were delivered by gravity flow from the paired channels of theta () glass tubing (1.6 mm OD, BT-15-10, Sutter Instruments, Novato, CA) pulled to an OD of 100-120 µm. The  tubing was mounted to a piezoelectric device (Burleigh Instruments, Model LSS-3100), allowing an individual neuron suspended in the RS stream to be rapidly exposed to the adjacent agonist stream by a translational step triggered from the recording software. Junction current experiments with open tip patch pipettes revealed that movement of the laminar interface separating streams of 150 and 75 mM NaCl flowing from the  tubing occurs in <1 ms. Agonists were also applied by pressure ejection (5-10 psi) from large-bore (2-5 µm diam) patch pipettes (Choi and Fischbach 1981; Margiotta and Gurantz 1989; Margiotta et al. 1987a,b). The latter method was used for single-channel studies (see following text) and, occasionally, for whole cell experiments to induce slow, 3*-AChR currents. Whole cell 3*-AChR currents and decay kinetics induced in this way were indistinguishable from those obtained using fast ACh perfusion, and results obtained using both approaches have been combined.

(1)

SINGLE-CHANNEL RECORDING. Single AChR channels were studied in outside-out patches excised from ciliary or choroid neurons. Patches were exposed to RS containing ACh (2 or 5 µM) or nicotine (0.5 µM) by gentle pressure microperfusion (2-5 psi) for 10-30 s, at holding potentials ranging from 60 to 140 mV. Patch currents were filtered at a cutoff frequency (f_c) of 5.8-8.0 kHz using an 8-pole Bessel filter (902 LPF; Frequency Devices, Haverhill, MA) and sampled onto computer disk at 5 × f_c using a TL-1 interface and pClamp 6.0 software (Axon Instruments) or an ITC-16 interface and Pulsefit software (Instrutech Systems, Long Island, NY). The vast majority of single-channel currents occurred as unitary events; those corrupted by the simultaneous opening of other channels were not included for analysis. Many of the AChR openings associated with Bgt-AChRs appeared to have durations <100 µs. To accurately resolve the events, we used the criterion that only events above a preset amplitude having durations more than two times the rise time (T_r = 0.3321/f_c) of the Bessel filter would be accepted for analysis (Colquhoun and Sigworth 1995).

SINGLE-CHANNEL CURRENT ANALYSIS. Single-channel currents, obtained from choroid and ciliary neuron patches, were compiled into histograms and fit with Gaussian functions having means and standard deviations that defined the amplitude limits for four channel classes. Channel conductances () were determined from the slopes of mean channel current versus voltage (I-V) plots. The I-V plots also provided a measure of the reversal potential (E_r) for each channel class, determined by linear extrapolation. In patches used for kinetic analysis, single AChR channel currents were collected at a single holding potential (typically 100 mV), and channel conductances calculated using mean E_r values determined from patches where full I-V plots were obtained (Table 2). Open-duration histograms were constructed for the accepted events in each AChR class, and fit with exponential functions to obtain a measure of the mean channel open time (_o). Steady-state open probabilities for each AChR channel conductance class (P_open,x) were calculated using P_open,x = (_t,x/T)/N_x, where T represents the total record length, _t,x is the summed open durations of channel class x, and N_x is the number of channels of that conductance class in the patch. N_x was estimated from the number of current levels in each patch recording that corresponded to conductance class x, and by visual inspection of the records was usually 1.

(2)

n

n

n

n

Fluorescence detection of AChRs

Bgt- and 3*-AChRs were visualized using fluorescence methods modified from Blumenthal et al. (1999). To detect Bgt-AChRs, freshly dissociated E13,14 ciliary ganglion neurons were treated with biotinylated Bgt (20 nM; Molecular Probes) in recording solution (RS) containing 10% horse serum (RS^+hs) and incubated at 37°C for 2 h. After washing twice with RS^+hs and twice with RS, cells were fixed with 1-2% paraformaldehyde for 30 min, rinsed three times with RS^+hs, and then incubated with Cy3-conjugated streptavidin (2 µg/ml; Jackson Laboratories, Bar Harbor, ME) in RS^+hs for 45 min at room temperature. Cells were then rinsed twice with RS^+hs and twice with RS, and the coverslips were dipped in distilled water and mounted on glass slides with Vectashield (Vector Laboratories, Burlingame, CA). To detect 3*-AChRs, neurons were treated mAb35 (100 nM; A generous gift from Dr. D. K. Berg) in RS^+hs containing 17% rabbit serum at room temperature for 1.5 h. After three rinses in RS^+hs, cells were incubated with biotinylated-SP-conjugated rabbit anti-rat IgG (5 µg/ml, Jackson Laboratories) in RS^+hs for 45 min at room temperature, washed three times in RS^+hs, incubated with Cy3-conjugated streptavidin (2 µg/ml) in RS^+hs, and then rinsed three times with RS^+hs and twice with RS. Cells were then fixed in 1-2% paraformaldehyde for 10 min and washed and mounted as for Bgt staining. Previous studies have demonstrated the specificities of these probes for fluorescence detection of AChRs on chick ciliary ganglion neurons (Jacob et al. 1984; Wilson Horch and Sargent 1995).

Neurons were examined with DIC and reflected light fluorescence using an Olympus BX50 microscope equipped with a UplanFL ×40 objective (0.75 N.A.) and optics appropriate for Cy-3 detection (510-550 nm band-pass filter, 570-nm dichroic mirror, 590-nm barrier filter). After focusing on the cell surface, 16-bit DIC and fluorescence images were acquired from identified ciliary and choroid neurons using a cooled digital CCD camera (SenSys, Model KAF-1400; Photometrics, Tucson AZ) under the control of IP Lab software (Version 3.0 Scanalytics; Reading, PA). For subsequent quantification of the fluorescence signal from each neuron studied, an elliptical region of interest (ROI) was placed around the neuron perimeter (defined by the DIC image), and the fluorescence intensity values for each pixel within the ROI were summed. The summed intensity value was then divided by the ROI area yielding fluorescence intensity per unit area (total fluorescence density). The background fluorescence density was obtained for an identical ROI area acquired from an adjacent unlabeled region of the same image, and the value subtracted to yield net fluorescence density (total minus background) for each neuron. A similar immunofluorescence detection and CCD quantification approach was used previously to determine relative levels of surface Bgt- and 3*-AChRs on freshly dissociated ciliary ganglion neurons isolated at various stages of development (Blumenthal et al. 1999). In this earlier study, tests of the SenSys CCD camera demonstrated that the probe concentrations used here are in the linear portion of the camera's fluorescence detection range. Digitized images were prepared for presentation using Photoshop 4.0 (Adobe Systems, San Jose, CA) after conversion to TIFF format.

Bgt binding assay

NEURON ISOLATION. Immediately after trituration, dissociated ciliary ganglion neurons were plated in triplicate, 16-mm-diam plastic culture wells (at 1-4 ganglion equivalents per well in a final volume of 400 µl). This plating procedure gives a uniform distribution of neurons, but because some are invariably lost during the dissociation, the actual number of neurons per well was determined by counting 5-10 randomly chosen microscope fields for each condition. Calculated yields were typically 50-75%, based on a total of 3,700 neurons (2,300 choroid +1,400 ciliary) expected per E14 ciliary ganglion (Pilar et al. 1980).

NEURON MORPHOMETRY. To determine the relative numbers and surface areas of choroid and ciliary neurons, the lengths of major and minor axes of 200 randomly selected neurons were measured. The axes were used to calculate soma surface areas (S) based on the area equation for an oblate spheroid {S = 2a² + (b²/) log [(1 + )/ (1  )]} where a and b are the lengths of the major and minor semiaxes, and  is the ellipticity given by  = (a²  b²)^1/2/a (Beyer 1987), and the calculated S values were compiled into histograms. The surface area histograms were well described by the sum of two Gaussian functions describing choroid (60-70%) and ciliary neuron populations, with ciliary neurons having mean surface areas three- to fourfold larger than choroid neurons.

BINDING. After equilibration for 2 h at 37°C, neurons were incubated in RS^+hs containing 0-130 nM [¹²⁵I]-Bgt at 37°C for 2 h and rinsed three times with RS^+hs. [¹²⁵I]-Bgt was obtained from DuPont NEN (Wilmington, DE) at an initial specific activity of 209 cpm/fmol. Nonspecific binding was determined in parallel triplicate wells by including 1 µM unlabeled toxin with the [¹²⁵I]-Bgt For quantification of ¹²⁵I radioactivity, the wells were scraped in 0.6 N NaOH, and the radioactivity counted in an LKB Wallac gamma counter (Model 1275 Minigamma, Gaithersburg, MD).

Materials

Fertilized white leghorn chicken eggs were obtained from Hertzfeld Poultry Farm (Waterville, OH) and maintained at 37°C in a forced-air draft incubator at 100% humidity. Any chemicals or reagents not already specified were obtained from Sigma (St. Louis, MO).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Identification of ciliary and choroid neurons

Neurons dissociated from E13,14 ciliary ganglia fall into ciliary and choroid populations that can be distinguished on the basis of differences in cell body size (Fig. 1, A and B). Measurements from 1,300 selected neurons revealed a cell body diameter distribution best fit by a bimodal Gaussian function having significantly different peaks at 13 ± 2 and 23 ± 3 µm (mean ± SD), respectively (P < 0.001). These diameter values are similar to those previously reported for choroid and ciliary neurons, respectively (Landmesser and Pilar 1974). To confirm the correlation of cell size with each respective neuron population, ciliary neuron cell bodies were labeled by staining their axons with DiI (Fig. 1, C and D). DiI-labeled ciliary cell bodies were uniformly large, having diameters well described by a single Gaussian function with a peak at 21 ± 4 µm. The size of DiI-stained ciliary neurons was significantly different from that of the small cells (P < 0.001) but not of the large cells (P > 0.1) measured from unstained ganglia. Large cells were therefore assigned to the ciliary neuron population. Small cells were assigned to the choroid neuron population because they did not label with DiI and because the choroid and ciliary neuron populations in the ganglion have previously been shown to correlate with small and large soma diameters, respectively (Landmesser and Pilar 1974; Pilar and Tuttle 1982). Thus the 95% confidence interval (CI = 1.96 × SD) for the large cell distribution was used to establish cells with diameter 16 µm as a criterion size for choroid neurons. Similarly, using the 95% CI for the small cell distribution would yield a criterion size of 17 µm for cells to be assigned to the ciliary population. In practice, however, only the largest cells (with diameters 23 µm) were considered ciliary neurons. This was done because previous studies indicated that the two populations in adult chickens overlap in size somewhat, with some neurons of intermediate size scoring as choroid cells by ultrastructural criteria (De Stefano et al. 1993; Pilar et al. 1980).

View larger version (53K): [in this window] [in a new window]   Fig. 1. Ciliary and choroid neuron can be identified by soma diameter. A: differential interference contrast (DIC) view of 2 neuron cell bodies representative of choroid (left) and ciliary (right) neurons after dissociation from E14 ciliary ganglia. B: size distribution of ciliary ganglion neuron cell body diameters (n = 1280; binwidth 2.2 µm). The histogram was well described by the sum of 2 Gaussian functions (, r2 = 0.87) predicting mean (± SD) neuron diameters of 13 ± 2 µm and 23 ± 3 µm for the 2 populations. The individual functions are depicted as - - -. C: mixed fluorescence (top) and DIC (bottom) views of a field of dissociated neurons from ganglia where ciliary axons were stained with 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate (DiI) to selectively label the ciliary neuron population. Due to the intense DiI fluorescence, the diameter of the labeled neuron appears larger in the fluorescence view than in the accompanying DIC view, from which size measurements were made. The labeled (ciliary) neuron shown (bottom right) has a diameter of 24 µm while that for the unlabeled neuron (top left) is ~14 µm. D: distribution of DiI-labeled (ciliary) neuron cell body diameters (n = 123; binwidth 2.2 µm). The histogram was well described by a single Gaussian function (, r2 = 0.93) predicting a mean ciliary neuron diameter of 21 ± 3 µm. Note the different magnifications in A and C; scale bars in both represent 25 µm.

Whole cell AChR currents from choroid and ciliary neurons

To maximally activate Bgt- or 3*-AChRs, choroid and ciliary neurons were transiently exposed to saturating doses of nicotine (20 µM) or ACh (500 µM). Nicotine was used to assay whole cell responses mediated by Bgt-AChRs (Fig. 2) because it provided good temporal separation of Bgt- and 3*-AChR currents (Blumenthal et al. 1999; Pardi and Margiotta 1999) due to the higher affinity of Bgt-AChRs for nicotine (Zhang et al. 1994). To determine if specific Bgt-AChR responses differed in ciliary and choroid populations, whole cell current values were normalized to membrane capacitance (C_m), which was presumed to be proportional to membrane surface area and therefore to cell size. In both neuron populations, fast application of 20 µM nicotine induced a biphasic response featuring a large, rapidly decaying response component, and a smaller, slowly decaying response component (Fig. 2, A and B). We confirmed that the rapidly decaying component represents activation of Bgt-AChRs since it was completely abolished after treating neurons with Bgt (e.g., Fig. 2C; see also Fig. 3C) as previously described (Blumenthal et al. 1999; Pardi and Margiotta 1999; Zhang et al. 1994). The peak values of Bgt-AChR currents (I_f) were similar in choroid and ciliary neurons (Fig. 2, A and B, insets), and overall I_f did not differ significantly between the two populations (P > 0.1, Table 1). After normalizing for population differences in C_m, however, the similar values for I_f translated to substantially larger peak values of Bgt-AChR current density (I_f/C_m) associated with choroid neurons when compared with ciliary neurons (Fig. 2, A and B, filled arrows). Overall, this difference was threefold (P < 0.001, Table 1). While 3*-AChR currents were studied in isolation using ACh applied in the presence of Bgt (see following text), they could also be resolved as the slowly decaying current component (I_s±i) induced by nicotine. In contrast with the cell-specific difference observed in Bgt-AChR current density, I_s±i values induced by nicotine were scaled proportionally in choroid and ciliary neurons (484 ± 57 and 1,959 ± 150 pA, respectively, P < 0.001), and the resulting measures of I_s±i/C_m (e.g., Fig. 2, A and B, open arrows) did not differ detectably between the two neuron populations (53 ± 6 and 67 ± 7 pA/pF, respectively, P > 0.1).

View larger version (18K): [in this window] [in a new window]   Fig. 2. -Bungarotoxin (Bgt)-acetylcholine receptor (AChR) response densities are threefold larger in choroid compared with ciliary neurons. Whole cell records are displayed from representative E14 choroid (A) and ciliary (B and C) neurons identified by the size criteria described in the text and Fig. 1. Neurons were held at 70 mV and exposed to 20 µM nicotine by fast perfusion, as indicated by the bar above each record. Except for the insets in A and B, nicotine-induced responses were normalized for the different sizes of ciliary and choroid neurons by dividing each digitized current value by Cm (pA/pF), and thereby reflect current densities. Time scale below C applies to A-C. The open and filled arrows indicate peak values of the 3*- (Is±i/Cm) and Bgt- (If/Cm) AChR current density components, respectively. Insets: the records at expanded time resolution and scaled to 100% of the total peak current, with calibrations shown at bottom right. Also shown are the respective times to achieve peak response (Tp), fits to Eq. 1 (filled dots), and the associated values of f for each neuron. The record in C shows the nicotine response of a ciliary neuron after 1 h treatment with 60 nM Bgt. Note that Bgt abolishes the initial, fast component of the response, but leaves the slow, 3*-AChR-mediated component unaffected. Similar results were obtained for choroid neurons (data not shown). Cell diameters and Cm were, respectively, 7 µm and 9.4 pF in A, 30 µm and 31.7 pF in B, and 27 µm and 28.4 pF in C.

We were initially concerned that the difference in Bgt-AChR current density might reflect nonsynchronous exposure to agonist across the cell surface, a potential limitation of the fast perfusion method that might be expected to be more severe for larger cells. This concern now seems unfounded for two reasons. First, the  tubing pore diameter (50 µm) used to deliver agonist exceeds the mean diameter of ciliary neurons by twofold. Second, if ciliary neurons were more susceptible to size-related limitations of the perfusion system, they would be expected to display slower kinetics of AChR current activation and desensitization. We measured the time required for Bgt-AChR currents to attain peak value (T_p = 4-10 ms) and the time constant describing the Bgt-AChR desensitization (_f = 5-11 ms) following nicotine exposure but detected no significant cell-specific differences in these parameters (Fig. 2; P > 0.1, Table 1). Thus the larger Bgt-AChR current density observed for choroid neurons cannot be readily explained by technical limitations of the perfusion system and is likely to represent a true difference between Bgt-AChRs on the two neuron populations.

Fast perfusion experiments with nicotine indicated a difference between choroid and ciliary neurons in Bgt- but not 3*-AChR current densities. To further explore this finding, we examined 3*-AChRs in isolation after treating neurons with 60 nM Bgt to block Bgt-AChRs. Identified neurons were tested as in Fig. 2, except that 500 µM ACh was used as the agonist, and Bgt was present (Fig. 3). Under these conditions, 3*-AChRs are activated in isolation from Bgt-AChRs (Blumenthal et al. 1999; Pardi and Margiotta 1999) producing a peak whole cell current represented by either a single slow component or the sum of intermediate and slow components (I_s±i) having time constants _i and _s (Fig. 3, A and B). In the presence of Bgt, the isolated, peak 3*-AChR currents induced by ACh were scaled proportionally to neuron cell size (Fig. 3, A and B, insets) resulting in current density values (I_s±i/C_m) that were not detectably different between choroid and ciliary neurons (Fig. 3, A and B, open arrows; P > 0.1, Table 1). To confirm the specificity of Bgt, values of I_s±i/C_m obtained for choroid and ciliary neurons (using nicotine in the presence of toxin or ACh in its absence) were compared. The observation that I_s±i/C_m values were indistinguishable between choroid and ciliary neurons for both ACh and nicotine, while I_f/C_m was dramatically reduced by >90% in both cases (Fig. 3C) indicates Bgt is specific in blocking I_f without affecting I_s±i. The time required for 3*-AChR currents to attain peak value and the associated decay time constants were also measured and found to be generally similar for the two neuron populations. For both choroid and ciliary neurons, T_p was 10 ms and _s was 2 s, while _i for choroid neurons (90 ms) reflected a somewhat faster decay than the value obtained for ciliary neurons (P < 0.05, Table 1). The small difference detected in _i is nevertheless likely to be specific since the methods employed permit adequate resolution of much faster desensitization processes (see Fig. 2 and METHODS). Taken together, the findings indicate that the peak whole cell response density attributable to Bgt-AChRs is threefold larger in choroid compared with ciliary neurons, while that attributable to 3*-AChRs did not differ detectably between the two populations.

View larger version (22K): [in this window] [in a new window]   Fig. 3. 3*-AChR response densities of choroid and ciliary neurons are indistinguishable. Whole cell 3*-AChR responses are displayed from representative E13 choroid (A and C) and ciliary (B) neurons. Records in A and B were obtained using fast perfusion as in Fig. 2 except that 3*-AChR responses were isolated by pretreating neurons with 60 nM Bgt and using 500 µM ACh as agonist. The open arrows indicate peak values of the 3*-AChR current normalized to membrane capacitance (Is±i/Cm). Insets: the records at an expanded time base and scaled to 100% of the total peak current with calibrations shown at bottom right. Also shown are the respective times to achieve peak response (Tp), fits to Eq. 1 (filled circles), and the associated values of i and s for each neuron. Cell diameters and Cm were, respectively, 8 µm and 9.5 pF in A, and 28 µm and 30.0 pF in B. The bar graph in C shows the specificity of Bgt for If over Is±i using either 20 µM nicotine or 500 µM ACh as agonist. If/Cm (black bar) and Is±i/Cm (gray bar) agonist response values obtained from neurons treated with Bgt (50 nM, 1 h) are plotted as percents of the values from untreated controls (100%, solid line ± SE) in the same experiments. For both nicotine and ACh tests, Bgt reduced If/Cm by >90% (P < 0.01) without a detectable effect on Is±i/Cm measured from the same neurons (n = 9-12 Bgt-treated, and 6-14 control neurons for each).

Comparison of Bgt- and mAb35-site densities on choroid and ciliary neurons

We next used a fluorescence method (Blumenthal et al. 1999) to determine if the larger Bgt-AChR responses seen for choroid neurons could be explained by a higher overall density of Bgt binding sites expressed on the neuronal cell surface (Fig. 4). Neurons were incubated with biotinylated Bgt and then with Cy3-conjugated streptavidin to visualize Bgt-AChRs. Neurons in ciliary and choroid populations were identified based on the criteria described in Fig. 1, and both displayed specific labeling (Fig. 4A) compared with control preparations where the cells were treated with 100 µM D-tubocurarine chloride (dTC; Fig. 4B) or where the toxin was omitted (data not shown). There was, however, no obvious difference in the intensity or the pattern of Bgt labeling on the two neuronal types. Both ciliary and choroid neurons displayed a nonhomogeneous distribution of Bgt-AChRs at about the same overall surface density. Specific 3*-AChR density was also similar for the two cell types, as seen by using mAb35 as the label (Fig. 4, C and D). The higher fluorescence intensity associated with mAb35 (relative to Bgt) probably reflects signal amplification due to the utilization of a biotinylated secondary antibody (See METHODS). To quantify the level of AChR labeling on the two neuron populations, we compared the average fluorescence intensity of the labeling for Bgt- and 3*-AChRs on the two neuron populations using digital microscopy. In accord with the images in Fig. 4, A-D, there was no detectable difference in Bgt- or mAb35 net fluorescence density between ciliary and choroid neurons (Fig. 4E). The lack of a difference in Bgt- net fluorescence density between ciliary and choroid neurons did not represent a sensitivity limitation of digital microscopy. Reducing the Cy3-conjugated streptavidin concentration threefold by molar replacement with FITC-conjugated streptavidin resulted in a 2.1 ± 0.2-fold reduction in average fluorescence intensity compared with control neurons treated with normal levels of Cy3-conjugated streptavidin (P < 0.001; n = 16 neurons for each condition). These findings indicate that the higher 7-AChR response density for choroid over ciliary neurons (Fig. 2; Table 1) cannot be explained by a gross cell-specific difference in total 7-AChR surface density.

View larger version (39K): [in this window] [in a new window]   Fig. 4. Distributions of Bgt- and 3*-AChRs on ciliary and choroid neurons. Neurons were dissociated from E13,14 ganglia and labeled with biotinylated Bgt (20 nM) to detect Bgt-AChRs (A) or with mAb35 (100 nM) to detect 3*-AChRs (C), as described in METHODS. Fields containing neurons from nonspecific controls (B and D) were processed as in A and C except that in B neurons received 100 µM D-tubocurarine chloride (dTC) in conjunction with the biotinylated Bgt, while in D mAb35 was omitted from the initial labeling step. Choroid and ciliary neurons are present at the top and bottom, respectively, of each panel. Cells were viewed by fluorescence microscopy and images acquired with the level of focus set at the cell surface. Exposure and filter settings for the images in each pair (A and B and C and D) were identical. Scale bar: 15 µm. E: Bgt-AChR and 3*-AChR densities do not differ on choroid and ciliary neurons. Images were acquired from neurons such as those depicted in A-D and quantified as described in METHODS. For each probe, values represent the mean (± SE) net fluorescence density and thereby provide a basis for comparing the average density of Bgt- or 3*-AChRs on ciliary vs. choroid neurons. Note that while the mAb35 net fluorescence density signal was higher than that for Bgt, there was no significant difference (P > 0.1) in the net fluorescence density of either probe on ciliary neurons () compared with choroid neurons (). Results depicted represent measurements from 15 to 27 ciliary and choroid neurons labeled with Bgt or mAb35 from 2 to 3 separate experiments each.

Surface labeling with [¹²⁵I]-Bgt reveals an abundance of Bgt sites

We next measured the number of surface Bgt-AChRs present on ciliary and choroid neurons. Freshly isolated E14 ciliary ganglion neurons displayed saturable surface binding of [¹²⁵I]-Bgt having a K_D of 1.1 ± 0.2 nM (n = 2) and a B_max of 7.67 ± 0.03 fmol per ganglion equivalent (n = 2, Fig. 5A). To determine the relationship between number of neurons and binding sites, the number of ganglion equivalents was varied, and these studies yielded a linear relationship (Fig. 5B) predicting a similar surface site density (7.47 fmol per ganglion equivalent). These results are consistent with previous findings from whole ganglion extracts, where detection of surface, intracellular and presynaptic Bgt sites revealed 15-20 fmol of Bgt bound per E14 ciliary ganglion (Chiappinelli and Giacobini 1978; Conroy and Berg 1995; Pugh et al. 1995). Based on digital imaging experiments, we determined that 94 ± 1% of the net Bgt fluorescence was associated with neuron cell bodies; only a tiny fraction was associated with bits of membrane debris or nonneuronal cells (data not shown; n = 10 fields from 2 experiments). Normalizing the number of sites for the yield of cells in each experiment and correcting for the small amount of binding to debris and nonneuronal cells permitted calculation of 1.14 ± 0.01 × 10⁶ Bgt-sites per neuron (n = 3 experiments). Because the density of Bgt sites was equivalent on the two neuron populations (Fig. 4E), we were able to calculate the number of Bgt sites per average choroid and ciliary neuron. Major and minor axis measurements from individual neurons in the same wells revealed the relative distribution of choroid and ciliary neurons in each well and allowed calculation of cell areas. The area and distribution measurements were then used to calculate the numbers of Bgt binding sites on choroid and ciliary neurons, assuming the label had access to the entire cell surface in both cases. Such calculations revealed, on average, 6.88 × 10⁵ Bgt sites per choroid neuron and 2.41 × 10⁶ Bgt sites per ciliary neuron.

View larger version (16K): [in this window] [in a new window]   Fig. 5. Quantifying surface Bgt-AChRs on ciliary ganglion neurons. To determine the total number of surface Bgt binding sites (Bmax) and their affinity (KD) for the toxin, E14 ciliary ganglion neurons were plated and incubated with the indicated concentrations of [125I]-Bgt (A). Nonspecific binding was determined from wells including 1 µM unlabeled Bgt. Aliquots of the incubation mixtures were counted to assure the actual concentrations of free label. The total number and size distribution of the neurons were determined by cell counts and axes measurements in randomly selected fields from parallel wells (see METHODS). Scatchard analysis (A, inset) of the binding isotherm (A) yielded a KD of 0.96 nM and a Bmax of 7.64 fmol/ganglion equivalent (1 ganglion equivalent = 2,300 ch +1,400 ci = 3,700 neurons). Similar results were seen in an additional experiment. To determine the linearity of the assay, differing numbers of ciliary ganglion neurons were plated and subjected to binding with 30 nM 125[I]-Bgt (± 1 µM unlabeled toxin). Specific binding was linear relative to the number of cells plated (B) with a slope of 7.47 fmol/ganglion equivalent (r2 = 0.99). Again, the total number and size distribution of the neurons were determined by cell counts in randomly selected fields in parallel wells.

Single AChR currents from ciliary and choroid neurons reveals four distinct channel classes

A higher Bgt-AChR affinity for nicotine cannot explain the larger response density of choroid over ciliary neurons since nicotine applied at 20 µM is known to generate near-maximal Bgt-AChR responses from ciliary neurons (Zhang et al. 1994). We therefore conducted excised patch experiments to determine if the higher Bgt-AChR responses of choroid neurons resulted from cell-specific differences in the permeation or kinetic properties of individual receptors. Four distinct AChR channel amplitude classes were detected in outside-out patches excised from choroid or ciliary neurons challenged with recording solution containing 2-5 µM ACh (Fig. 6, A and B, Table 2) or 0.5 µM nicotine. The events were mediated by nicotinic AChRs since they were absent in patches pretreated with 100 µM dTC and not seen in patches challenged with recording solution lacking agonist (data not shown). The AChR channel events were classified according to their conductance (), determined from the slope of single channel current amplitude versus voltage plots (e.g., Fig. 6C). Two previously described event classes (Margiotta and Gurantz 1989) displaying slope conductances of approximately 30 and 40 pS (Table 2) were routinely observed. Two additional AChR channel classes that we noted previously but did not characterize because of their brief open durations (Margiotta and Gurantz 1989) were also present in most patches. In the present studies, we were able to record these latter events at multiple voltages to determine that they display slope conductances of ~60 and 80 pS (Fig. 6C; Table 2). Each of the four channel classes displayed an extrapolated current reversal potentials (E_r) indicative of similar cationic selectivities. Mean E_r values were about 10 mV for the 30- and 40-pS channels, as previously described, and nominally closer to 0 mV for the 60- and 80-pS channels (Table 2). As expected for AChR permeation (Gardner et al. 1984),  and E_r values obtained for the four channel classes using 0.5 µM nicotine as agonist were similar to those obtained using ACh (n = 4 patches; data not shown). In addition, there was no detectable difference between choroid and ciliary neurons in mean values of  or E_r (P > 0.1 for both) for each channel class, suggesting that ion permeation through AChRs is similar in the two neuronal populations.

View larger version (24K): [in this window] [in a new window]   Fig. 6. Identification of 4 AChR channel classes on ciliary ganglion neurons. Single-channel currents were collected from excised, outside-out neuron patches exposed to 5 µM ACh. A: 4 AChR channel amplitude classes (1-4) from a representative choroid neuron patch held at 100 mV are displayed. Some events are labeled with arrows in the top panel. These events are displayed at expanded time resolution in the bottom panel traces, except for the smallest event (1), which was obtained from a different record segment. In both panels, note that the largest events (3 and 4) display very brief open durations. Calibration: 4 pA and 40 ms (top), 4 pA and 1 ms (bottom). Currents were sampled at 25-µs intervals and filtered at 6.8 kHz such that events with durations <100 µs were excluded from analysis. B: amplitude histogram for events in the patch depicted in A (n = 787 events; 0.25 pA binwidth). Peaks labeled with arrows correspond to the 4 event amplitude classes. C: current-voltage plot for the same patch with the slope conductances for each amplitude class indicated. Extrapolated current reversal potentials (in mV) for each of the event classes (1-4) were 1.1 (1), 15.2 (2), 9.1 (3), and 13.5 (4). Similar conductance and reversal potential values were obtained from 5 ciliary and 4 choroid patches, and mean values (± SE) are compiled in Table 2. D: open-duration distributions for channel openings corresponding to the 60 (3)-and 80 (4)-pS AChR events are displayed from another choroid neuron patch held at 100 mV. The distributions were well fit (r2 > 0.9 for both) with single-exponential functions predicting open time constants (o) of 135 and 217 µs for the 60 (3)- and 80 (4)-pS AChR channel event classes (n > 200 events for each; binwidth 68 µs). Currents were sampled at 34-µs intervals and filtered at 5.8 kHz so that events <115 µs were excluded. Legend applies to both C and D. Similar distributions were obtained from 7 ciliary and 6 choroid patches, and mean o values (± SE) for each of the 4 event classes are compiled in Table 2.

To further explore potential cell-specific differences in channel properties and more fully characterize Bgt-AChRs, the open time, open probability, and opening frequency of each AChR conductance class were determined using ACh as the agonist (Fig. 6D and Table 2). As with permeation properties, however, these kinetic parameters did not differ detectably between ciliary and choroid neurons (Table 2), and results from both populations are combined in the following description. For each channel class, open time constants (_o) were derived by fitting exponential functions to open-duration histograms, while open probabilities (P_open), and opening frequencies (F_open) were calculated directly from the records (see METHODS). As previously shown, the open time distributions for the 30- and 40-pS AChR events were best fit by the sum of two exponential functions (Margiotta and Gurantz 1989). The open time constants (_o,1 and _o,2) were ~100 µs and 2 ms, respectively, and as before, the brief-duration openings represented a minority contribution for both event classes (Table 2). The 30- and 40-pS events occurred with an overall F_open of ~5 and 18 s¹ and with P_open values of 0.011 and 0.067, respectively. In contrast, the 60- and 80-pS openings were uniformly brief, with _o for the 80-pS events (200 µs) scoring about twofold larger (P < 0.001) than those of the 60-pS events (Fig. 6, A and D; Table 2). The brief open durations of 60- and 80-pS events, combined with opening frequencies of 7 and 5 s¹, respectively, resulted in their making a much smaller net contribution to the total record open time (P_open  0.002) than either 30- or 40-pS events.

Bgt blocks 60- and 80-pS AChR channels

To correlate the channel conductance classes with AChRs of established subunit composition and toxin sensitivity, neurons were treated with Bgt (60 nM). Toxin block was assessed by comparing F_open and P_open for each conductance class in patches from treated and untreated neurons. Results from choroid and ciliary neurons were again indistinguishable and are combined in the following description. In each of six patches from neurons treated with 60 nM Bgt, both P_open (Fig. 7, A and B) and F_open (not shown) associated with the 60- and 80-pS events dropped to zero. The blockade of 60- and 80-pS AChR channels was specific since values for F_open and P_open for the 30- or the 40-pS events were not detectably changed in Bgt-treated patches when compared with the values obtained in 13 control patches (P > 0.1 for both). This finding indicates that the 60- and 80-pS events arise from Bgt-AChRs, and extends the results from whole cell experiments where 60 nM Bgt blocked the initial, rapidly decaying current density component attributable to Bgt-AChRs but not the slow component attributable to 3*-AChRs. The experiments further suggest that the 30- and 40-pS events arise from 3*-AChRs, which do not contain 7-subunits and are neither recognized nor blocked by Bgt.

View larger version (19K): [in this window] [in a new window]   Fig. 7. Channels (60 and 80 pS) are blocked by Bgt. Single AChR channel currents were obtained as in Fig. 6. The AChR channel current amplitude histogram in A is from an excised patch from a neuron treated with Bgt (60 nM; 1 h). Labeled arrows indicate the channel current amplitudes expected for the indicated channel class. Note that the current amplitudes expected for 60- and 80-pS channel classes (3 and 4) were abolished by Bgt, while the relative abundance of events having amplitudes corresponding to the 30- and 40-pS channel classes (1 and 2) were similar to untreated controls (e.g., Fig. 6B). B summarizes the toxin effects. The bar plots show mean Popen values in toxin for the indicated channel classes expressed as a mean (± SE) percent of the value obtained in patches from control neurons not treated with toxin (n = 13). Open triangles and bars indicate controls; filled triangles and bars indicate toxin treatment (n = 6). Asterisks indicate a significant reduction (P < 0.05) in Popen for the indicated channel class in the presence of the toxin compared with untreated controls (100%). Note that Bgt reduced Popen for the 60- and 80-pS channels to 0% of control levels. The toxin did not detectably change Popen for the 30- or 40-pS events (P > 0.1 for both).

Bgt-AChR opening kinetics

We analyzed the closed durations between 60-pS channel openings induced by 5 µM ACh in outside-out patches excised from choroid and ciliary neurons (Fig. 8) to gain more information about the rate constants governing Bgt-AChR channel activation (k_n, , and  = _o¹) based on Scheme 2. The analytical approach described by Colquhoun and Hawkes (Colquhoun and Hawkes 1981; Colquhoun and Sakmann 1981; Colquhoun and Sigworth 1995; Sine and Steinbach 1986) was employed because it is more amenable to records with few events (200-300) such as those obtained here, than is single-channel ensemble analysis (Dionne and Leibowitz 1982; Leibowitz and Dionne 1984), which we used previously to study 3*-AChR kinetics (Margiotta and Gurantz 1989; Margiotta et al. 1987a,b). The basic assumption inherent in both approaches is that a single channel having just closed to AnR, and still fully occupied by agonist, is more likely reenter the open state than channels are to enter the open state from any of the other n closed states. Scheme 2 provides a generally valid description of the processes governing ligand-gated channel activation, but receptor desensitization would require that additional considerations be made. While Bgt-AChRs do desensitize rapidly after exposure to high concentrations of agonist (e.g., Fig. 2),