The Journal of Neurophysiology Vol. 77 No. 1 January 1997, pp. 145-152
Copyright ©1997 by the American Physiological Society

Spatial Distribution of Synaptically Activated Sodium Concentration Changes in Cerebellar Purkinje Neurons

Joseph C. Callaway and William N. Ross

Department of Physiology, New York Medical College, Valhalla, New York 10595

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Callaway, Joseph C. and William N. Ross. Spatial distribution of synaptically activated sodium concentration changes in cerebellar Purkinje neurons. J. Neurophysiol. 77: 145-152, 1997. The spatial distribution of Na⁺-dependent events in guinea pig Purkinje cells was studied with a combination of high-speed imaging and simultaneous intracellular recording. Individual Purkinje cells in sagittal cerebellar slices were loaded with either fura-2 or the Na⁺ indicator sodium binding benzofuran isophthalate (SBFI) with sharp electrodes or patch electrodes on the soma or dendrites. [Na⁺]_i changes were detected in response to climbing fiber and parallel fiber stimulation. These changes were located both at the anatomically expected sites of synaptic contact in the dendrites and in the somatic region. The variation in time course of these [Na⁺]_i changes in different locations implies that Na⁺ enters at the synapse and diffuses rapidly to locations of lower initial [Na⁺]_i. The synaptically activated somatic [Na⁺]_i changes probably reflect Na⁺ entry through voltage-sensitive Na⁺ channels because they were detected only when regenerative potentials were recorded in the soma. [Na⁺]_i changes in response to antidromically or intrasomatically evoked Na⁺ action potentials also were confined to the cell body. These observations are in agreement with other evidence that Na⁺ spikes are generated in the somatic region of the Purkinje neuron and spread passively into the dendrites. Plateau potentials, evoked by depolarizing pulses to the soma or dendrites, caused [Na⁺]_i changes only in the soma, indicating that the noninactivating Na⁺ channels contributing to this potential also were concentrated in this region. The climbing fiber-activated [Na⁺]_i changes were blocked by the -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione, indicating that these changes were not due to direct stimulation of the Purkinje neuron or activation of metabotropic receptors. Direct depolarization of the soma or dendrites never caused dendritic [Na⁺]_i increases, suggesting that the climbing fiber-activated [Na⁺]_i changes in the dendrites are due to Na⁺ entry through ligand-gated channels. A climbing fiber-like regenerative potential could be recorded in the soma after anode break stimulation, parallel fiber activation, or depolarizing pulses to the soma. The [Na⁺]_i changes evoked by all of these potentials were confined to the cell body region of the Purkinje cell. [Ca²⁺]_i changes in the dendrites evoked by the anode break potential were small relative to climbing fiber-activated changes, suggesting that a Ca²⁺ spike was not evoked by this response. The anode break and directly responses were blocked by tetrodotoxin. These results suggest that the somatically recorded climbing fiber response is predominantly a Na⁺-dependent event, consisting of a few fast action potentials and a slower regenerative response activating the same channels as the Na⁺ plateau potential.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

A detailed understanding of synaptic integration and plasticity requires a knowledge of how voltage and ligand-gated conductances are distributed throughout the neuronal arborization. The analysis of these mechanisms has become more complicated after the general recognition that different kinds of voltage-sensitive conductances can be found in the dendrites as well as the soma and axon (Llinás 1988). Most attention has been directed toward Ca²⁺-dependent mechanisms because of their critical role in the induction of synaptic plasticity and because they have been relatively easier to study. However, in some neurons, dendrites can fire Na⁺ action potentials (Wong et al. 1979) and possess subthreshold Na⁺ conductances that may amplify synaptic potentials (Stafstrom et al. 1985). Further, some reports have suggested that [Na⁺]_i changes might be important in the induction of long-term depression (LTD) (Linden et al. 1993). To assess the significance of these conductances and concentration changes, the location and properties of different Na⁺-dependent events need to be investigated in each cell type.

One part of this problem now appears to be answeredthe site of Na⁺ spike generation in Purkinje cells. This issue is of special importance because this action potential determines the output message of the neuron. When this site is in the dendrites, a more elaborate model of cellular information processing must replace the simpler view of integration as an event occurring at a single locus at the axon hillock (Adams 1992; Softky and Koch 1993). Although there has been some controversy about the location of spike initiation (Regehr et al. 1992), experiments using dual patch recordings on the soma and dendrites found that propagation of Na⁺ spikes was completely passive in the dendrites under all conditions and that the density of Na⁺ channels declined rapidly with distance from the soma (Stuart and Hausser 1994). From these data Stuart and Hausser concluded that the spikes must be generated in the soma and/or axon of the Purkinje cell. This result supports previous experiments using extracellular (Hounsgaard and Yamamoto 1979) and microelectrode recordings (Llinás and Sugimori 1980a,b). It also explains why Na⁺ spikes produce elevations in [Ca²⁺]_i and [Na⁺]_i only in the somatic region (Lasser-Ross and Ross 1992; Lev-Ram et al. 1992).

This paper examines three other issues regarding the nature and location of different Na⁺ conductances in Purkinje cells. The first question is where do [Na⁺]_i changes occur when the cell is activated synaptically? If ligand-gated Na⁺ entry can be separated from voltage-gated entry, then concentration changes due to entry through the receptor channels will indicate the site of active synapses and receptors. If these sites can be determined with precision, then it may be possible to analyze how different synaptic contact points facilitate or depress with repeated activity. A further point concerns the role of [Na⁺]_i changes in the induction of LTD. Based on experiments using cultured neurons, Linden et al. (1993) suggested that an increase in [Na⁺]_i is a required cofactor for this process. Determining the sites and sources of synaptically activated [Na⁺]_i increases may help determine how this particular form of synaptic plasticity occurs.

The second issue concerns the spatial distribution of Na⁺ entry associated with the noninactivating Na⁺ conductance (Crill 1996). Llinás and Sugimori (1980a) showed that there was a strong Na⁺-dependent plateau potential that supported burst firing in Purkinje cells. Because tetrodotoxin (TTX) blocks both the plateau channels and the fast spike channels, it has been difficult to analyze this noninactivating conductance directly. Indeed, it is not clear yet whether this conductance results from a channel distinct from the fast inactivating channel (Westenbroek et al. 1989), a chemically modulated form of the fast channel (e.g., Ma et al. 1994), the same channel with different gating properties (Alzheimer et al. 1993), or perhaps the same channel activated in a voltage "window" where some of the fast channels remain activated (DeSchutter and Bower 1994). Only in the first case is there a possibility for the spatial distribution of the Na⁺ plateau conductance to extend beyond the region where the fast spikes are active. The determination of this distribution is important because of the possibility that noninactivating Na⁺ channels in the dendrites could participate in boosting parallel fiber (PF) excitatory postsynaptic potentials (EPSPs) toward the soma (Stuart and Sakmann 1995).

The last issue concerns the conductances responsible for the classic climbing fiber (CF) response recorded in the soma. This response consists of a few fast Na⁺ spikes followed by a rounded, 15-20 ms wide potential. At various times, this potential has been ascribed to the synaptic potential or Ca²⁺ spikes. Our analysis suggests that this event in the soma is due almost entirely to regenerative Na⁺ potentials, both fast spikes and a potential due to activation of the plateau channels. Some of this later work has been published in abstract form (Callaway et al. 1995b).

	METHODS

Abstract Introduction Methods Results Discussion References

The procedures used in these experiments followed those previously published (Callaway et al. 1995a; Lasser-Ross and Ross 1992; Lasser-Ross et al. 1991; Lev-Ram et al. 1992; Llinás and Sugimori 1980a). Briefly, 150- to 200-mm sagittal slices from the vermis of 250-350 g guinea pigs were prepared after sodium pentobarbital anesthesia and decapitation with a small animal guillotine. Submerged slices were mounted in a recording chamber superfused with standard Krebs solution (in mM): 124 NaCl, 5 KCl, 1.2 NaH₂PO₄, 2.4 CaCl₂, 1.3 MgSO₄, 26 NaHCO₃, and 10 glucose, equilibrated with 95% O₂-5% CO₂, pH 7.4. Temperature was controlled at 30-32°C. TTX (1 mM, Sigma Chemical, St. Louis, MO) or 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 20 mM, RBI, Natick, MA) was added in some experiments to block Na⁺ channels or synaptic responses. For sharp microelectrode recordings and dye injections, the electrode tips were filled with 6 mM of the sodium indicator sodium binding benzofuran isophthalate (SBFI; impermeant salts, Molecular Probes, Eugene, OR) dissolved in 200 mM KAc. Patch electrode solutions contained (in mM) 4 NaCl, 10 Na-Gluconate, 130 K-Gluconate, 4 Mg-ATP, 10 N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES; pH 7.2). This solution was supplemented with 100-500 mMethylene glycol-bis(-aminoethylether)-N,N,N,N-tetraaceticacid (EGTA) and 2 mM SBFI for [Na⁺]_i measurements, or 250 mM fura-2 and no EGTA for [Ca²⁺]_i measurements. Both sharp electrodes and patch electrodes were pulled from 1.5 mm OD, fiber-filled capillary tubing (No. 1511-M, Friedrich and Dimmock, Millville, NJ) using a model P-97 puller (Sutter Instrument, Novato, CA). Extracellular stimulating electrodes were bipolar sharpened tungsten wires exposed from insulation at the tips. They were placed on the white matter for CF activation or on the molecular layer for PF activation. Electrical recording and stimulating procedures were standard (Callaway et al. 1995a). Experiments were performed on the stage of an inverted microscope (Olympus, IMT-2F) (Callaway et al. 1995a). All electrodes were mounted on the stage and controlled with hydraulic micromanipulators. In this way, the impalements and electrode positions could be preserved while changing the field of view. Electrodes were placed in position by viewing the preparation through a dissecting microscope mounted above the stage of the inverted microscope.

High-speed fura-2 and SBFI fluorescence measurements using a cooled CCD camera were as described in Lasser-Ross et al. (1991) and Callaway et al. (1995a). Fluorescence was detected using a ×20, 0.75 NA dry objective (Nikon, Garden City, NY). To increase the accuracy of the measurements, the excitation intensity stability was improved by using an Opti-Quip (Highland Mills, NY) model 1600 power supply (ripple <0.02%) and a Hammamatsu L2481-01 75-W Mercury-Xenon arc lamp, which is more stable than ordinary bulbs. Excitation for either indicator dye was at 380 nm selected with a 15-nm-wide interference filter (Omega Optical, Brattleboro, VT). Typical intensity fluctuations were <0.3%. Bleaching and baseline drift were corrected by subtracting an identical measurement without stimulation. Optical changes are expressed as F/F (change in fluorescence from resting levels divided by the resting fluorescence corrected for autofluorescence). The autofluorescence correction was made by making comparable measurements at locations in the slice away from the dyed cell and subtracting this measurement from the experimental ones. Changes in [Ca²⁺]_i or [Na⁺]_i are expected to be proportional to F/F for small values of F/F if the resting concentration of these ions is uniform. Homogeneity for resting [Ca²⁺]_i has been established (Tank et al. 1988); we assume uniformity for resting [Na⁺]_i. No attempt was made to convert F/F into true ion concentrations. Cell images in the figures were taken at resting potential using 380-nm excitation for both indicator dyes. The gray scale images of F/F in the figures show the spatial distribution of the change in fluorescence between a time before stimulation (usually at resting potential) and the end of the stimulation period. [Ca²⁺]_i measurements are from a single trial. [Na⁺]_i measurements usually required signal averaging to produce clean records. The number of trials in each experiment are indicated in the figure legends.

Recordings were made by tracking the dye-filled microelectrode through the slice. Seals and patch recordings were made with the blind technique of Blanton et al. (1989) using a Neurodata IR-283 amplifier (New York, NY) in the current-clamp mode. Cells obtained in this way usually were not on the surface, making it difficult to resolve the finest structures. In particular, we could not see the axon or the details of the fine dendrites.

	RESULTS

Abstract Introduction Methods Results Discussion References

Climbing fiber induced [Na⁺]_i changes

Previously (Lasser-Ross and Ross 1992) we reported that the CF response caused a small [Na⁺]_i change in the dendrites of Purkinje cells. However, the small size of the fluorescence signals prevented a detailed analysis of the spatial distribution or origin of these signals. In these new experiments, we usually used patch electrodes to load the cell with higher concentrations of SBFI, enabling us to achieve larger fluorescence signals. Figure 1 shows the spatial distribution of the [Na⁺]_i changes and their time courses in selected dendritic locations in response to a train of CF stimuli. At the end of the train of 10 responses, the changes (F/F) were predominantly along the main dendrites, with comparatively smaller changes along the segment closest to the soma (n = 10). However, the time course of the [Na⁺]_i change was not the same at all locations. Along the main distal dendrites, [Na⁺]_i began rising from the first response and peaked at the time of the last response. In the fine dendrites, to the side of the main distal dendrites and in the proximal dendrite, the changes rose later and peaked later. This pattern is consistent with Na⁺ entry along the main shaft and diffusion outward towards the finer dendrites. However, the resolution of our data cannot rule out some entry at locations away from the primary dendrites. Note that this pattern is completely different from the distribution of [Ca²⁺]_i changes evoked by the CF (e.g., Miyakawa et al. 1992, Fig. 3). These CF-induced [Ca²⁺]_i changes were highest in the fine dendrites, lower on the thicker dendrites, and lowest in the somaa pattern consistent with a spatially uniform Ca²⁺ entry in the dendrites, with concentration changes determined largely by the relative surface to volume ratio in the different regions.

View larger version (68K): [in this window] [in a new window]   FIG. 1. Spatial distribution and time course of [Na+]i changes in response to a train of 10 climbing fiber responses. Along thicker distal dendrites (locations 3 and 4), [Na+]i rose throughout train and peaked at end of train. In fine distal dendrites (5 and 6) and in proximal dendrite (2), increase was smaller, started later, and peaked after end of train. Somatic increase (1) also peaked at end of train. Gray scale image shows a somatic component from voltage-dependent Na+ entry and a dendritic component, mostly along the thicker distal dendrites, due to ligand-gated Na+ entry. Inset: 1st 2 dendritically recorded climbing fiber electrical responses shown (patch electrode in dendrites at arrow in cell image). Note large Ca2+ spike at peak of response and attenuated Na+ spike (small arrow) on rising edge of potential. Optical data are an average of 10 sweeps. SBFI, sodium binding benzofuran isophthalate.

View larger version (60K): [in this window] [in a new window]   FIG. 3. [Na+]i changes in dendrites in response to weak and strong parallel fiber stimulation. Left: train of 10 parallel fiber responses, all subthreshold. Both gray scale image and time dependent traces show that [Na+]i increase was very localized under point of stimulation. Right: stronger stimulation, where responses after first 2 were suprathreshold, caused a larger and more widespread [Na+]i increase. Note that at center location, [Na+]i peaked earlier and decayed faster than at more peripheral locations. There was also a small change insoma that does not show clearly in gray scale image. Electrical recording with a patch electrode on soma. Stimulating electrode was positioned over molecular layer at center of region with [Na+]i changes. Insets: details shown of 1st 5 excitatory postsynaptic potentials. Optical data are an average of 5 sweeps.

The pattern of dendritic [Na⁺]_i changes is consistent with Na⁺ entry through ligand-gated channels opened by the CF neurotransmitter because anatomic investigations have placed the CF contacts on the large dendritic branches (Larramendi and Victor 1967; Palay and Chan-Palay 1974). CNQX, a known blocker of the CF synaptic potential (Konnerth et al. 1990; Perkel et al. 1990), completely blocked the CF-induced [Na⁺]_i changes (data not shown). This experiment rules out [Na⁺]_i increases caused by direct stimulation of the Purkinje cell or increases as a consequence of metabotropic receptor activation. By itself, it does not rule out a contribution from voltage-dependent Na⁺ channels in the dendrites because the CF EPSP could open these channels if they were present. However, the possibility of CF induced voltage-dependent Na⁺ entry in the dendrites is unlikely because direct depolarization of the somas (Lasser-Ross and Ross 1992) or the dendrites (Fig. 5) caused no dendritic [Na⁺]_i increase.

View larger version (58K): [in this window] [in a new window]   FIG. 5. [Na+]i changes evoked by a 3-s-long depolarizing pulse that evoked a few spikes and a long plateau response. Image shows cell filled with SBFI and stimulated with a patch electrode on dendrites (arrow; same cell as Fig. 1). Both time-dependent traces and gray scale image show that changes were detected only in somatic region. Because electrical recording was in dendrites, amplitude of Na+ spikes was small. Inset: depolarization evoked a short, ~20-ms-wide, regenerative response before settling into steady plateau potential. Optical data are an average of 10 sweeps.

The CF response also caused a large [Na⁺]_i increase in the somatic region where there are no synaptic contacts. Some of this increase may have come from the Na⁺ spikes activated by the CF response. To test this possibility, we measured the [Na⁺]_i change resulting from five CF responses and a train of 40 antidromic action potentials on the same sweep (Fig. 2). Consistent with previous results (Lasser-Ross and Ross 1992) the [Na⁺]_i changes from the antidromic action potentials were entirely confined to the soma. The antidromically activated [Na⁺]_i changes were entirely due to the Na⁺ spikes because there was no change in this signal amplitude when 10 µM CNQX was added to the bath (data not shown). Although each CF response evoked 2 fast action potentials, making a total of 10 from all the CF responses, the amplitude of the [Na⁺]_i change from the CF responses was almost the same as from the train of antidromic spikes. Therefore the CF-induced [Na⁺]_i increase was more than three times the expected amplitude from the spikes alone. The remainder of the somatic [Na⁺]_i increase probably came from the slow part of the CF response, through the noninactivating Na⁺ channels described below.

View larger version (39K): [in this window] [in a new window]   FIG. 2. [Na+]i changes in response to a train of 5 climbing fiber stimuli followed by a train of 40 antidromic Na+ spikes. Climbing fiber responses caused a [Na+]i increase in soma and dendrites whereas antidromic responses caused an increase only in soma. Inset: expanded view shown of first CF response; scale bars are 20 mV and 20 ms. Optical data are an average of 20 sweeps.

Parallel fiber induced [Na⁺]_i changes

Parallel fiber stimulation directly over a dendritic branch caused a small, very localized increase in [Na⁺]_i at the point of stimulation when the resulting EPSP was subthreshold (Fig. 3, left). With stronger stimulation (Fig. 3, right), the dendritic [Na⁺]_i increase was larger and extended over a wider area (n = 4). The increase rose faster and peaked earlier at the center of the spot of elevated [Na⁺]_i. The slower time course and lower amplitude at the edges of the spot is consistent with diffusion to the sides from a central point of Na⁺ entry. However, as with the CF response, we cannot rule out some contribution resulting from direct entry at the periphery.

The focal distribution of [Na⁺]_i increases in these experiments is consistent with entry through ligand-gated channels. To further test this conclusion, we repeated this experiment in the presence of 20 mM CNQX to block the -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors that underlie the PF EPSP (Konnerth et al. 1990; Perkel et al. 1990). With the same stimulus intensity, the PF-evoked [Na⁺]_i changes were eliminated completely (data not shown). As with the CF-evoked [Na⁺]_i changes, this experiment rules out a contribution from direct stimulation of the dendrites or as a result of metabotropic receptor activation.

When the PF EPSPs were suprathreshold, an additional [Na⁺]_i increase was detected in the soma. This signal, while present in the data of Fig. 3, is difficult to see in the grey scale image. However, in other cells (e.g., Fig. 4) this component was much clearer. The somatic signal reasonably is assigned to the Na⁺ spikes and possibly to the Na⁺ plateau potential (see below). There was no signal in the region between the stimulation zone in the distal dendrites and the soma, indicating that dendritic Na⁺ channels play no role in boosting the EPSP towards the soma (n = 3). A signal from regenenerative sodium spikes should be detectable because the same apparatus measured sodium spike transients in thinner hippocampal pyramidal cell dendrites (Jaffe et al. 1992). With stronger stimulation, the EPSP often evoked a larger regenerative potential in the soma that resembled the CF response. The same pattern of two spots of Na⁺ entry was observed. This response was not due to activation of the climbing fiber by the stimulating electrode because the pattern of dendritic [Na⁺]_i changes was completely different from that observed when the CF was activated from an electrode over the white matter (Fig. 1). Rather, this experiment shows that this intrinsic regenerative response also can be evoked by PF input if the EPSP is large enough.

View larger version (52K): [in this window] [in a new window]   FIG. 4. Suprathreshold parallel fiber responses cause [Na+]i increases at a local spot in dendrites and in soma but not in between. Stimulating electrode placed over spot. Optical data are an average of 11 sweeps of 10 parallel fiber responses.

Sodium plateau potential

It was not possible to study the Na⁺-dependent plateau potential in isolation because any depolarization sufficient to induce this potential always caused a burst of fast Na⁺ spikes. Also, they were both equally sensitive to TTX. However, strong depolarizing pulses usually evoked a short burst of action potentials that rapidly inactivated, leaving just the plateau potential for the remainder of the pulse. Figure 5 shows a cell filled with SBFI from a patch electrode on a dendrite ~60 mm from the cell body (arrow). When the cell was stimulated with a depolarizing pulse, [Na⁺]_i rose throughout the duration of the pulse in the soma and did not change in the dendrites. This plateau signal was eliminated completely in TTX (not shown). Similar results were found when SBFI was loaded from patch or sharp electrodes on the soma (n = 11). We conclude that the channels causing the plateau associated [Na⁺]_i change, like the channels causing the spike related [Na⁺]_i increase, are not found in significant density in the dendrites. Whether the same channels are responsible for both [Na⁺]_i changes is still to be determined.

Ionic basis of the climbing fiber response

These experiments provide a framework for reevaluating the nature of the somatically recorded climbing fiber electrical response. Because of the complex interaction of Na⁺ and Ca²⁺ conductances in the Purkinje cell, dissection of this response has been challenging. Direct pharmacological experiments are difficult because blockade of either of these conductances will prevent the response completely by blocking presynaptic action potentials or transmitter release. However, indirect approaches have been more rewarding.

Analysis of the effects of inhibition on CF-evoked [Ca²⁺]_i changes (Callaway et al. 1995a) showed that most of the [Ca²⁺]_i change was caused by a fast Ca²⁺ spike in the dendrites; blocking this spike caused little change in the somatically recorded electrical CF response. This indicates that fast Ca²⁺ spiking is not important in generating this somatic potential.

Further evidence concerning this response was provided when we noticed that an almost identical electrical potential could be recorded in the soma after anode break stimulation. Figure 6 shows fluorescence and electrical responses to CF and anode break stimulation in a cell filled with fura-2. The insets show the close resemblance between the two potentials. The main difference is that the CF response rose faster, probably because it is driven by a large, fast EPSP. The CF-induced [Ca²⁺]_i changes were large and widespread in the dendrites. The anode break induced [Ca²⁺]_i changes were much smaller in the dendrites and more evenly distributed (n = 3). The spatial distribution of this response was variable. In this cell, the [Ca²⁺]_i change was detected out to the tip of the dendrites. In other cells, the [Ca²⁺]_i changes were predominantly in the somatic region with little or no detectable changes in the distal dendrites. We also found that the anode break electrical response was smaller in dendritic recordings even though Ca²⁺ spikes recorded from these locations were larger than at other locations in the dendrites or soma (data not shown). This result suggests that the anode break response does not evoke a dendritic Ca²⁺ spike. However, we do not have sufficient numbers of these recordings or any dual patch recordings to make a rigorous statement about the decline in amplitude of the anode break response with distance from the cell body.

View larger version (29K): [in this window] [in a new window]   FIG. 6. [Ca2+]i changes evoked by anode break stimulation are much smaller than changes caused by climbing fiber response. Optical records are from soma (gray trace) and 2 dendritic locations indicated by boxes in image. Insets: aligned to same absolute voltage, show that both stimuli evoked similar electrical responses as recorded with a patch electrode on soma. Both electrical responses reach approximately same peak potential.

The anode break response was also detected in neurons containing SBFI (Fig. 7). Again, the anode break and CF electrical responses recorded in the soma were almost identical. However, the anode break induced [Na⁺]_i change was detected only in the soma (n = 5). In this and all other cells tested, the anode break response was completely blockedby TTX.

View larger version (23K): [in this window] [in a new window]   FIG. 7. Climbing fiber responses cause [Na+]i increases all over Purkinje cell while anode break response caused an increase only in soma. Climbing fiber responses were evoked during hyperpolarizing pulse. Some of this [Na+]i increase was caused by hyperpolarizing pulse itself, probably by Na+ entry through an anomalous rectifier channel (Crepel and Penit-Soria 1986). Anode break-induced [Na+]i increase was larger than typical climbing fiber induced changes, probably because of long-lasting 2nd phase to plateau. Insets: both electrical responses, recorded with a sharp electrode in the soma, reached approximately same peak potential. Same cell and box locations as in Fig. 2. Optical data are an average of 20 sweeps.

Together these experiments suggest that the CF response and the anode break response are predominantly Na⁺-dependent events, consisting of a combination of fast spikes and a slower regenerative potentialall occurring in the soma. Whether Ca²⁺ conductances contribute in a significant way is not clear. Some [Ca²⁺]_i change remained when the fast CF-induced Ca²⁺ spike was blocked by inhibition (Callaway et al. 1995a), and a similar small [Ca²⁺]_i change was detected after the anode break response (Fig. 6). However, the Ca²⁺ conductance responsible for these [Ca²⁺]_i changes may make only a small contribution to the regenerative potential.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Na⁺ plateau potentials

Our data contribute useful information to the classification of the channels underlying the plateau potential in Purkinje cells, but they do not settle the issue. We found that: neither the [Na⁺]_i increases associated with this plateau nor the [Na⁺]_i changes caused by Na⁺ spikes were detected in the dendrites and a large [Na⁺]_i increase was detected even when the cell was depolarized far enough to completely inactivate the fast spikes. The large signal at depolarized potentials appears to rule out a significant contribution from a "window" current (active only in a small range of potentials where channels are partially activated but not completely inactivated). All the results are consistent with spike channels and plateau channels being different forms of the same Na⁺ channel but differing in their inactivation properties. These characteristics are close to those of Na⁺ channels in neurons from rat and cat sensorimotor cortex described by Alzheimer et al. (1993) and Crill (1996). They found that the fast and persistent forms of the Na⁺ current could be explained by a uniform population of Na⁺ channels that can switch individually between different gating modes. They also found that the two currents in the cortical neurons appeared to have different voltage activation curves, even though they resulted from the same channels (Brown et al. 1994). The characteristics of [Na⁺]_i changes we observed in the soma are compatible with Na⁺ entry through similar kinds of channels.

It is important to note that the lack of plateau or spike related [Na⁺]_i changes in the dendrites does not prove that both potentials are caused by the same channels or even that they are colocalized. It is possible that spike-related Na⁺ entry could come through noninactivating channels even if a plateau is not generated, as when the spikes were evoked antidromically. With this possibility the data are also consistent with inactivating channels predominantly in the axon hillock and noninactivating channels predominantly in the soma.

Ligand-gated signals

These experiments clearly demonstrate that measurable [Na⁺]_i changes can be produced by Na⁺ entry through ligand-gated channels. This possibility was suggested previously (Lasser-Ross and Ross 1992). In those experiments, we did not have the spatial resolution or pharmacology to clearly support this hypothesis. In these new experiments PF-related signals were localized in the dendrites directly under the stimulating electrode. For CF activation, the [Na⁺]_i changes were predominantly along the main dendrites. Both locations correspond to the known sites of synaptic contact in Purkinje cells (Larramendi and Victor 1967; Palay and Chan-Palay 1974). Although these signals were small, they were detected with sufficient signal to noise ratio and spatial resolution to suggest that they might be used in experiments comparing aspects of the potentiation and depression of synaptic currents in different regions of the dendrites. Also, if local [Na⁺]_i increases are required for the induction of LTD (Linden et al. 1993), then these ligand gated changes are the only likely source.

For both PF- and CF-activated ligand-gated signals, there was evidence of diffusion from the putative site of Na⁺ entry to nearby locations where the [Na⁺]_i concentration was lower. It is possible that some of this diffusion could be mediated by the indicator dye "shuttling" Na⁺ from regions of high concentration to nearby regions of lower concentration, as has been suggested for Ca²⁺ ions (e.g., Connor and Nikolakopoulou 1982). However, the diffusion rate for Na⁺ in these cells (measured in the presence of the indicator SBFI) is much higher than the rate for Ca²⁺ (measured in the presence of fura-2) in other cells (e.g., Callaway et al. 1993). This higher diffusion rate is consistent with diffusion rates measured with other techniques without added buffers (e.g., Hodgkin and Keynes 1956). Therefore it is likely that diffusion of Na⁺ is significant when there is no indicator in the cell. Rapid diffusion should help to maintain [Na⁺]_i close to resting levels at the synapse and other sites of localized Na⁺ entry.

Climbing fiber response

These experiments showed that the CF-like wide regenerative electrical response, as recorded in the soma, could be evoked via CF activation, by anode break stimulation, and by PF stimulation. A similar potential also could be evoked with intrasomatic depolarization (inset, Fig. 5). The fact that this response could be evoked through many pathways establishes that it is an intrinsic regenerative potential in the Purkinje cell and not dependent on the unique EPSP of the CF synapse. These and previous experiments (Callaway et al. 1995a) indicate that this response is not due to a Ca²⁺ spike, is blocked by TTX, and is largest in the cell body. Because both the inactivating and noninactivating Na⁺ channels are located in the cell body, it is likely that this potential is a somatically localized regenerative event. However, the small and widespread [Ca²⁺]_i increase associated with this response suggests that Ca²⁺ channels also contribute to some extent.

The all-or-none character of the somatic CF response helps insulate the Purkinje cell output from variations in the strength of the CF synaptic current due to repetitive activation (Konnerth et al. 1990; Perkel et al. 1990) or other factors. In particular, as we have shown previously, stellate cell inhibition (Callaway et al. 1995a) or activation of an A-like K⁺ current (Midtgaard et al. 1993) have little affect on the somatically recorded CF response. In contrast, both the A-current and dendritic inhibition can dramatically affect potentials and [Ca²⁺]_i changes in the dendrites.

	ACKNOWLEDGEMENTS

  We thank N. Lasser-Ross for discussons and help in some of the experiments.

  This work was supported in part by grants from the Human Frontier Science Program and the National Science Foundation (IBN-9209784) and a National Research Service Award postdoctoral fellowship (NS09257) to J. C. Callaway.

	FOOTNOTES

  Present address and address for reprint requests: J. C. Callaway, Dept. of Anatomy and Neurobiology, University of Tennessee at Memphis, 865 Monroe Ave., Memphis, TN 38163.

  Received 22 May 1996; accepted in final form 6 September 1996.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• ADAMS, P. R. The platonic neuron gets the hots. Curr. Biol. 2: 625-627, 1992.[Medline]
• ALZHEIMER, C., SCHWINDT, P. C., CRILL, W. E. Modal gating of Na⁺ channels as a mechanism of persistent Na⁺ current in pyramidal neurons from rat and cat sensorimotor cortex. J. Neurosci. 13: 660-673, 1993.[Abstract]
• BLANTON, M. S., LOTURCO, J. J., KRIEGSTEIN, A. R. Whole cell recording from neurons in slices of reptilian and mammalian cerebral cortex. J. Neurosci. Methods 30: 203-210, 1989.[Medline]
• BROWN, A. M., SCHWINDT, P. C., CRILL, W. E. Different voltage dependence of transient and persistent Na⁺ currents is compatible with modal-gating hypothesis for sodium channels. J. Neurophysiol. 71: 2562-2565, 1994.[Abstract/Free Full Text]
• CALLAWAY, J. C., LASSER-ROSS, N., ROSS, W. N. IPSPs strongly inhibit climbing fiber activated [Ca²⁺]_i increases in the dendrites of cerebellar Purkinje neurons. J. Neurosci. 15: 2777-2787, 1995a.[Abstract]
• CALLAWAY, J. C., LASSER-ROSS, N., STUART, A. E., ROSS, W. N. Dynamics of intracellular free calcium concentration in the presynaptic arbors of individual barnacle photoreceptors. J. Neurosci. 13: 1157-1166, 1993.[Abstract]
• CALLAWAY, J. C., ROSS, W. N., LASSER-ROSS, N. Spatial distribution of sodium dependent events in cerebellar Purkinje neurons. Soc. Neurosci. Abstr. 21: 587, 1995b.
• CONNOR, J. A., NIKOLAKOPOULOU, G. Calcium diffusion and buffering in nerve cytoplasm. Lect. Math. Life Sci. 15: 79-101, 1982.
• CREPEL, F., PENIT-SORIA, J. Inward rectification and low threshold calcium conductance in rat cerebellar Purkinje cells. An in vitro study. J. Physiol. Lond. 372: 1-23, 1986.[Abstract/Free Full Text]
• CRILL, W. E. Persistent sodium current in mammalian central neurons. Annu. Rev. Physiol. 58: 349-362, 1996.[Medline]
• DESCHUTTER, E., BOWER, J. M. An active membrane model of the cerebellar Purkinje cell. I. Simulation of current clamps in slice. J. Neurophysiol. 71: 375-400, 1994.[Abstract/Free Full Text]
• HODGKIN, A. L., KEYNES, R. D. Experiments on the injection of substances into squid giant axons by means of a microsyringe. J. Physiol. Lond. 131: 592-616, 1956.
• HOUNSGAARD, J., YAMAMOTO, C. Dendritic spikes in Purkinje cells of the guinea pig cerebellum studied in vitro. Exp. Brain Res. 37: 387-398, 1979.[Medline]
• JAFFE, D. B., JOHNSTON, D., LASSER-ROSS, N., LISMAN, J. E., MIYAKAWA, H., ROSS, W. N. The spread of Na⁺ spikes determines the pattern of dendritic Ca²⁺ entry into hippocampal neurons. Nature Lond. 357: 244-246, 1992.[Medline]
• KONNERTH, A., LLANO, I., ARMSTRONG, C. M. Synaptic currents in cerebellar Purkinje cells. Proc. Natl. Acad. Sci. USA 87: 2662-2665, 1990.[Abstract/Free Full Text]
• LARRAMENDI, L. M. H., VICTOR, T. Synapses of the Purkinje cell spines in the mouse: an electronmicroscopic study. Brain Res. 5: 15-30, 1967.[Medline]
• LASSER-ROSS, N., MIYAKAWA, H., LEV-RAM, V., YOUNG, S. R., ROSS, W. N. High time resolution fluorescence imaging with a CCD camera. J. Neurosci. Methods 36: 253-261, 1991.[Medline]
• LASSER-ROSS, N., ROSS, W. N. Imaging voltage and synaptically activated sodium transients in cerebellar Purkinje cells. Proc. R. Soc. Lond. B Biol. Sci. 247: 35-39, 1992.[Medline]
• LEV-RAM, V., MIYAKAWA, H., LASSER-ROSS, N., ROSS, W. N. Calcium transients in cerebellar Purkinje neurons evoked by intracellular stimulation. J. Neurophysiol. 68: 1167-1177, 1992.[Abstract/Free Full Text]
• LINDEN, D. J., SMEYNE, M., CONNOR, J. A. Induction of cerebellar long-term depression in culture requires postsynaptic action of sodium ions. Neuron 11: 1093-1100, 1993.[Medline]
• LLINÁS, R. R. The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous system function. Science Wash. DC 242: 1654-1664, 1988.[Abstract/Free Full Text]
• LLINÁS, R., SUGIMORI, M. Electrophysiological properties of in vitro Purkinje cell somata in mammalian cerebellar slices. J. Physiol. Lond. 305: 171-195, 1980a.[Abstract/Free Full Text]
• LLINÁS, R., SUGIMORI, M. Electrophysiological properties of in vitro Purkinje cell dendrites in mammalian cerebellar slices. J. Physiol. Lond. 305: 197-213, 1980b.[Abstract/Free Full Text]
• MA, J. Y., LI, M., CATTERALL, W. A., SCHEUER, T. Modulation of brain Na⁺ channels by a G-protein-coupled pathway. Proc. Natl. Acad. Sci. USA 91: 12351-12355, 1994.[Abstract/Free Full Text]
• MIDTGAARD, J., LASSER-ROSS, N., ROSS, W. N. Spatial distribution of Ca²⁺ influx in turtle Purkinje cell dendrites in vitro: role of a transient outward current. J. Neurophysiol. 70: 2455-2467, 1993.[Abstract/Free Full Text]
• MIYAKAWA, H., LEV-RAM, V., LASSER-ROSS, N., ROSS, W. N. Calcium transients evoked by climbing fiber and parallel fiber synaptic inputs in guinea-pig cerebellar neurones. J. Neurophysiol. 68: 1178-1189, 1992.[Abstract/Free Full Text]
• PALAY, S. L., CHAN-PALAY, V. In: The Cerebellar Cortex. Cytology and Histology., . New York: Springer Verlag, 1974
• PERKEL, D. J., HESTRIN, S., SAH, P., NICOLL, R. Excitatory synaptic currents in Purkinje cells. Proc. R. Soc. Lond. B Biol. Sci. 241: 116-121, 1990.[Medline]
• REGEHR, W. G., KONNERTH, A., ARMSTRONG, C. M. Sodium action potentials in the dendrites of cerebellar Purkinje cells. Proc. Natl. Acad. Sci. USA 89: 5492-5496, 1992.[Abstract/Free Full Text]
• SOFTKY, W. R., KOCH, C. The highly irregular firing of cortical cells is inconsistent with temporal integration of random EPSPs. J. Neurosci. 13: 334-350, 1993.[Abstract]
• STAFSTROM, C. E., SCHWINDT, P. C., CHUBB, M. C., CRILL, W. E. Properties of persistent sodium conductance and calcium conductance of layer V neurons from cat sensorimotor cortex in vitro. J. Neurophysiol. 53: 153-170, 1985.[Abstract/Free Full Text]
• STUART, G. S., HAUSSER, M. Initiation and spread of sodium action potentials in cerebellar Purkinje cells. Neuron 13: 703-712, 1994.[Medline]
• STUART, G., SAKMANN, B. Amplification of EPSPs by axosomatic sodium channels in neocortical pyramidal neurons. Neuron 15: 1065-1076, 1995.[Medline]
• TANK, D. W., SUGIMORI, M., CONNOR, J. A., LLINÁS, R. R. Spatially resolved calcium dynamics of mammalian Purkinje cells in cerebellar slice. Science Wash. DC 242: 773-777, 1988.[Abstract/Free Full Text]
• WESTENBROEK, R. E., MERRICK, D. K., CATTERALL, W. A. Differential subcellular localization of the R_I and R_II Na⁺ channel subtypes in central neurons. Neuron 3: 695-704, 1989.[Medline]
• WONG, R. K. S., PRINCE, D. A., BASBAUM, A. I. Intradendritic recordings from hippocampal neurons. Proc. Natl. Acad. Sci. USA 76: 986-990, 1979.[Abstract/Free Full Text]

0022-3077/97 $5.00 Copyright ©1997 The American Physiological Society