HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Supplemental data All Versions of this Article: 560/1/27    most recent jphysiol.2004.072678v1 Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Holthoff, K. Articles by Konnerth, A. Search for Related Content PubMed PubMed Citation Articles by Holthoff, K. Articles by Konnerth, A.

¹ Physiologisches Institut, Ludwig-Maximilians Universität München, Pettenkoferstr. 12, 80336 München, Germany
² Department Biological Sciences, Columbia University, New York, NY 10027, USA

	Abstract

Top Abstract Introduction Methods Results Discussion Supplementary material References

 
Mammalian dendrites are active structures, capable of regenerative electrical activity. Dendritic spikes can mediate synaptic plasticity and could enrich the computational properties of neurons. Besides sodium-based action potentials, which can propagate throughout the dendritic tree, neocortical pyramidal neurons also sustain dendritic spikes that are spatially restricted. The function of these ‘local’ dendritic spikes is unknown. We show that local spikes, which require activation of N-methyl-D-aspartate receptors (NMDARs), induce long-term synaptic depression (LTD) in layer 5 pyramidal neurons. This depression does not require somatic spiking and is input specific. Moreover, a single synaptic stimulus can evoke a dendritic spike and a brief local dendritic calcium transient, and is sufficient for the full induction of LTD.

(Received 26 July 2004; accepted after revision 16 August 2004; first published online 19 August 2004)
Corresponding authors R. Yuste: Department Biological Sciences, Columbia University, New York, NY 10027, USA. Email: rmy5{at}columbia.edu

	Introduction

Top Abstract Introduction Methods Results Discussion Supplementary material References

 
Neuronal dendrites have exuberant dendritic morphologies, whose function is still poorly understood (Ramon y Cajal, 1904). In early work using intradendritic recordings, Purkinje cell dendrites were shown to have regenerative electrical activity (Llinas & Sugimori, 1980). Pyramidal neurons in neocortex and hippocampus also have dendritic spikes in vitro (Wong et al. 1979; Amitai et al. 1993) and in vivo (Pockberger, 1991). In particular, layer 5 pyramidal neurons sustain sodium action potentials (APs) that can propagate actively through large regions of their dendritic trees (Stuart & Sakmann, 1994). These so-called backpropagating action potentials can induce long-term potentiation or depression, depending on their relative timing to that of EPSPs (Markram et al. 1997). Besides these sodium spikes, which can also be initiated in the dendrite (Golding & Spruston, 1998), neocortical pyramidal neurons or CA1 neurons also have dendritic calcium spikes (Amitai et al. 1993; Schiller et al. 1997). Finally, a potentially different form of dendritic spikes (so-called NMDA spikes) has been described in response to glutamate uncaging (Schiller et al. 2000; Wei et al. 2001). These spikes are spatially very restricted and are triggered by activation of NMDARs.

Using two-photon and confocal imaging we have investigated the characteristics and function of local dendritic spikes in layer 5 pyramidal neurons from mouse primary visual cortex. We find that these dendritic spikes initiate a fast calcium transient of high amplitude in a small spine-dendritic compartment. Furthermore, a local dendritic spike elicits long-term synaptic depression, which is spatially restricted and it is fully expressed after a single stimulus.

	Methods

Top Abstract Introduction Methods Results Discussion Supplementary material References

 
Experiments were performed on layer 5 pyramidal neurons in cortical slices (300 µm) of visual cortex from 14- to 19-day-old Balb/c mice. The animals were anaesthetized with ketamine–xylazine (50 and 10 mg kg^–1, respectively) or by exposure to CO₂, decapitated and the brains removed. Slices were incubated at 33°C in oxygenated standard solution for at least 40–60 min before they were transferred to the recording chamber. The standard solution contained (mM): 125 NaCl, 3 KCl, 3 CaCl₂, 1 MgCl₂, 1.25 NaH₂PO₄, 26 NaHCO₃, 20 glucose, bubbled with 95% O₂ and 5% CO₂. Combined electrophysiological recordings and confocal Ca²⁺ imaging were performed with an EPC9 patch-clamp amplifier (HEKA, Lambrecht, Germany or BVC-700, Dagan Corp., Minneapolis, MN, USA) and a confocal laser-scanning system (Noran, ‘Oz’, on an Olympus BX50WI microscope, x 60, NA 0.9 or x 20, NA 0.5 water immersion objective) or a custom-made two photon laser scanning microscope (Majewska et al. 2000b), consisting of a modified Fluoview (Olympus) confocal microscope with a Ti–sapphire laser providing 130 fs pulses at 75 MHz (Mira, Coherent, Santa Clara, CA, USA) pumped by a solid-state source (Verdi, Coherent), respectively. The pipette solution contained (mM): 140 potassium gluconate, 10 NaCl, 4 Mg-ATP, 2 Na₂-ATP, 0.4 Na-GTP, 10 K-Hepes, 0.2 Oregon Green 488 BAPTA-6F or 0.2 Calcium Green-1 (Molecular Probes), pH 7.3. Whole-cell voltage-clamp recordings were performed at 35–37°C. Neurons were stimulated synaptically using an extracellular pipette filled with ACSF. In some cases tips of stimulation pipettes were bent by about 70° with a microforge (Narishige, Japan). For imaging experiments, only dendritic branches, which were mostly parallel to the focal plane, were chosen to ensure that the dendrite of interest was in focus (see Supplementary material Fig. 1, available online only). For LTD experiments single voltage shocks (2–12 V, 50–100 µs duration) were applied to obtain EPSPs of 5 mV in amplitude. For conditioning, the strength of stimulation pulse was increased about threefold. We verified that the resulting depolarization and the dendritic calcium transient were entirely blocked by antagonists of glutamatergic transmission (50 µM APV plus 10 µM NBQX), leaving an inhibitory postsynaptic potential remaining that was sensitive to bicuculline (see Fig. 4C). To elicit reproducible changes in the imaged dendritic regions, the stimulating electrode was positioned near (10–30 µm) to the imaged dendrites. APV (2 mM in ACSF) was pressure ejected from micropipettes. For local dendritic application the ejection pipette was positioned about 10–30 µm away from the dendrite under the study.

View larger version (47K): [in this window] [in a new window]   Figure 1.  Dendritic spikes are spatially restricted Whole-cell recordings were obtained in the current-clamp mode, and cells were filled via the patch pipette with 200 µM of the low-affinity calcium indicator dye Oregon-Green BAPTA-6F. Imaging was done with a Noran confocal microscope. A, camera lucida drawing of a layer 5 pyramidal neuron in the mouse visual cortex. B, pseudo-colour coded images of relative fluorescence changes in a basal dendrite before and after dendritic spike initiation. Images were derived from a movie sequence (Supplementary material to Fig. 1) obtained in conditions of weak confocality (large confocal slit size, 20 x objective) to allow visualization of dendrites at different focal planes. For more clarity, image processing masks obtained from high-resolution z-stacks were superimposed on the movie-derived, activity-displaying images. Scale bar, 25 µm. C, weak synaptic stimulation evoked an EPSP, but no calcium response. D, strong synaptic stimulation evoked local dendritic spike (lower black trace), accompanied by a calcium transient in the activated dendrite (upper trace). Gray trace represents somotic voltage response to weak synaptic stimulation shown in C. E, camera lucida drawing of another layer 5 pyramidal neuron in visual cortex. F, pseudo-colour coded images of relative fluorescence changes in an apical dendrite before and after strong synaptic stimulation. Scale bar, 10 µm G, strong synaptic stimulation evoked local dendritic spike (lower trace) accompanied by a large calcium transient in the activated dendrite (upper trace). H, back-propagating action potential (AP) induced by somatic current injection evoked a significantly smaller dendritic calcium transient. I, comparison of dendritic calcium transient amplitudes evoked by local dendritic spikes and back-propagating APs. Frame rate of calcium recordings was 30 Hz in panels C, D and G and 60 Hz in panel H.

View larger version (22K): [in this window] [in a new window]   Figure 4.  NMDA-receptor dependence of LTD induction, synaptic origin of dendritic spike, and input specificity of LTD A, scheme of experimental configuration. The NMDA-receptor antagonist APV (2 mM pipette concentration) was applied focally via a puff pipette to the active input (left panel). Application of APV during conditioning stimulation prevented induction of LTD (right panel). B, dendritic calcium and somatic EPSP recording of local dendritic spike before (control), during (APV), and after focal application of APV (washout). Calcium recordings were done with the Noran confocal microscope at a frame rate of 60 Hz. C,responses of dendritic calcium and of somatic voltage evoked by synaptic stimulation under control conditions and after application of 10 µM NBQX, 50 µM APV and 10 µM bicuculline. Grey and black traces represent responses to weak and strong synaptic stimulation, respectively. Traces show different stimulations of the same cell as in Fig. 2D. Calcium recordings were done with the 2-photon microscope at a frame rate of 15 Hz. D, scheme of experimental configuration. The distance between the synaptic inputs was at least 400 µm. Ea, input 1 was conditioned by the dendritic spike LTD induction protocol (5 stimuli separated by 15 s intervals). Eb, the unconditioned pathway did not show any significant change in synaptic transmission. F summarizes results of LTD experiments under different conditions (LTDx1: single conditioning stimulus, n = 8; LTDx5: 5 conditioning stimuli, n = 5; hyperpol: cells hyperpolarized to –80 mV during conditioning, n = 5; input 1: conditioned input and input 2: unconditioned input of the same cell, n = 4; APV: local APV application during conditioning, n = 6).

Calibration procedure

To obtain a quantitative estimation of the changes in intracellular calcium concentration, the low affinity dye Oregon Green BAPTA-6F (OG-6F) was calibrated in a two-step procedure. First, we established the characteristic calibration curve for the standard pipette solution used in the whole-cell recording experiments by means of calibration kit solutions (Molecular Probes, Cat. no. C-3723) in vitro in a microcuvette. The dissociation constant K_d and R_F = F_max/F_min were determined by fitting the data points with a sigmoidal curve (IgorPro, Wavemetrics) and yielded a R_F of 6 and a K_d value of 2.2 µM. In a second step, we determined in whole cell recordings (‘in vivo’) the maximum change of fluorescence in neurons containing 200 µM OG-6F. The maximal loading of the cells with Ca²⁺, necessary for obtaining the maximal fluorescence value (F_max), was achieved by disrupting the seal by slightly tapping on the recording stage or by locally applying large concentrations (200 µM) of the Ca²⁺ ionophore ionomycine. With these two procedures we obtained quite similar ratio values of F_max/F₀ = 4.4 ± 0.2 (mean ± S.E.M., n = 14) and F_max/F₀ = 4.5 ± 1.1 (n = 4), respectively. From these results and by assuming that the intracellular resting Ca²⁺ concentration is 50 nM (Garaschuk et al. 1997; Maravall et al. 2000), we constructed the calibration curve used for quantifying the synaptically evoked Ca²⁺ transients (Supplemental Material, Fig. 2). In addition, we obtained independent estimates for the changes in Ca²⁺ concentration by using the approach of Maravall et al. (2000).

View larger version (23K): [in this window] [in a new window]   Figure 2.  Characteristics of local dendritic spikes A and B, similarities of decay time constants of dendritic calcium transients evoked by local dendritic spikes and back-propagating APs. In A, upper traces represent superimposed dendritic calcium transients from 11 dendritic locations of 10 cells. The analysis of the average yielded a decay time constant of 75 ms (n = 11, lower trace). In B, upper traces represent dendritic calcium responses of 5 different cells from those in A. The analysis of the average yielded a decay time constant of 74 ms (n = 5, lower trace). Note different scale for F/F in A and B. The calcium transients in A and B were recorded from comparable dendritic location of different cells 50–250 µm away from the soma. C summarizes the results of the analysis of experiments shown in A and B. Imaging in panels A and B were done with a Noran confocal microscope at 60 Hz. D, somatic voltage recordings (V) obtained after single-shock synaptic stimulation at different stimulation intensities. Beyond a certain threshold a second depolarizing, ‘active’ component occurred reflecting the local dendritic spike (right panel). The shaded areas, with their onset determined in conditions of strong stimulation, were used for analysis shown in E. E, summary plot comparing the results obtained from 5 cells. Results were pair-wise normalized to the areas determined in conditions of strong stimulation. In E and G, stimulation strength represents the difference between the strength for a given EPSP and that at ‘threshold’ of dendritic spike initiation. F, corresponding dendritic calcium signals (Ca) to experiments shown in D. Above the same threshold as in D, a large dendritic calcium transient was elicited (right panel). G, average of peak amplitude of corresponding dendritic calcium transients from the same experiments as in E as a function of stimulation intensity. Stimulation strength values correspond to those determined in E. Calcium recordings in panel F and G were performed with the 2-photon microscope at a frame rate of 15 Hz.

The LTD experiments were quantified by averaging the EPSP amplitudes in a time interval between 10 and 30 min after the induction. The result was normalized to the mean EPSP amplitude of the 10 min time interval before induction.

Cell reconstruction

To reconstruct the morphology of the neuron under investigation, z-stacks (2 µm step size) of images corresponding to all regions of the dye-loaded (OG-6F) cells were taken after each experiment. The neuron's morphology was determined by arranging appropriately maximal projection images of the z-stacks from all regions and by redrawing the contours of the dendrites by hand.

	Results

Top Abstract Introduction Methods Results Discussion Supplementary material References

 
To characterize the regions of the dendritic tree that responded regeneratively to synaptic stimulation, we carried out two-photon and confocal calcium imaging from apical and basal dendrites of layer 5 neocortical pyramidal neurons, under single-shock extracellular synaptic stimulation. Using the low-affinity Ca²⁺ indicator dye OG-6F we identified the location, amplitude and time course of dendritic spikes in basal dendrites (Fig. 1A and B). As a first step a subthreshold synaptic input site was identified, which responded to synaptic stimulation with a regular shaped EPSP but no detectable dendritic calcium accumulation (Fig. 1C). Subsequently, local dendritic spikes were evoked by gradually increasing the stimulation strength. Beyond a certain stimulus threshold, the electrical response turned into a complex EPSP, which contained a prominent second depolarizing component, delayed by 22 ± 2 ms (n = 32, Figs 1D and 2D). The resulting response consisted of a local dendritic Ca²⁺ transient and a clear kinetic hump (a dV/dt ‘breakpoint’) in the voltage deflection (Fig. 1D and G). The all-or-none appearance of these complex EPSP envelopes, and the presence of a distinct kinetic switch in many of them, is indicative of a distant electrical regenerative event (Schiller et al. 2000). Under these conditions, calcium accumulations occurred in both spines and adjacent dendritic shafts (Fig. 1B and F). Both the evoked dendritic calcium transient and the accompanying hump in the voltage deflection showed a clear and identical threshold for induction concerning stimulation strength (Fig. 2D–G), as described for local dendritic spikes (Schiller et al. 2000). We were also readily able to evoke a similar spike, when stimulating apical dendrites (Fig. 1F and G). In this case, the spike-like deflection riding on the EPSP was often of small amplitude and was, when stimulating remote apical dendrites, barely detectable in somatic whole-cell recordings (not shown). In all instances in which we systematically monitored the entire dendritic tree (n = 6), by imaging different planes of focus during the repeated delivery of strong synaptic stimuli, only a dendritic branch near the stimulation pipette was activated.

Mapping local calcium events produced by dendritic spikes

To explore the spatial extent of the calcium accumulations triggered by dendritic spikes we imaged a small region of the dendritic tree of the stimulated cell, located around the stimulating electrode, and found that strong synaptic stimulation caused a local increase in calcium concentration in a small spino-dendritic compartment. In some cases the region of local calcium accumulations appeared to terminate at dendritic branch points (Fig. 1F) although in most cases they stopped in an unbranched section of the dendrite (Fig. 1B). Spatially similar local calcium accumulations were observed regardless of which part of the dendritic tree was stimulated, including primary and secondary branches of the basal and apical dendritic tree. The dendritic spike-associated Ca²⁺ transients had a mean peak value of 3.9 ± 0.8 µM (n = 11, Fig. 1I; see Methods for calibration procedure). Although the use of the low-affinity Ca²⁺ indicator dye OG-6F allowed a realistic estimation of the time course of dendritic Ca²⁺ transients, without a marked alteration through the additional exogenous buffer capacity of the indicator, the Ca²⁺ transient amplitudes may still underestimate the peak Ca²⁺ levels, particularly those in the activated spines. In comparison, the mean amplitude of the dendritic Ca²⁺ transients evoked by single back-propagating action potentials was just 309 ± 31 nM (n = 5, Fig. 1H and I). Nevertheless, the decay time course of both types of Ca²⁺ transients was not significantly different (time constants of decay 59 ± 5 ms versus 75 ± 7 ms, respectively; t test, P > 0.16, Fig. 2A–C). This indicates that in both cases similar mechanisms control dendritic Ca²⁺ clearance. Our data established the kinetics of the dendritic spike-induced Ca²⁺ transients and demonstrated unambiguously that they occur in individual apical or basal dendrites of layer 5 cortical pyramidal neurons. We should note that in our experiments synaptic inhibition was intact, so dendritic inhibition may have contributed to the restriction of the local spikes to relatively small dendritic segments (Supplementary material Fig. 3).

View larger version (23K): [in this window] [in a new window]   Figure 3.  Single local dendritic spike induces instantaneous long-term synaptic depression (LTD) A, induction of LTD by single dendritic spike (LTDx1) in a mouse layer 5 pyramidal neuron. Synaptic input to a basal dendrite located approximately 50 µm from the cell body. Note the framed LTD induction signal (calibration bars indicate 50% F/F, 250 ms for calcium signal and 5 mV, 50 ms for EPSP). Traces indicate mean EPSPs before and after the conditioning stimulus (calibration bars indicate 2 mV and 50 ms). Imaging was done with the Noran confocal microscope at 30 Hz. B, summary LTDx1 experiments of 8 cells. C, LTD induction with 5 consecutive conditioning stimuli delivered 15 s apart from each other. D, same experiments as in panel C, but cells were hyperpolarized to –80 mV during conditioning stimuli to prevent the initiation of local dendritic spikes. No significant change in strength of synaptic transmission was detected.

A recent study (Golding et al. 2002) has demonstrated that, in conditions of suppressed synaptic inhibition, multiple dendritic spikes, produced by afferent theta burst stimulation, evoke long-term potentiation (LTP) in hippocampal pyramidal neurons. Surprisingly, we found that in layer 5 cortical neurons just a single sub-threshold EPSP producing a local dendritic spike was sufficient to evoke long-term depression (LTD) (Fig. 3A and B). This ‘single-shock’ LTD (LTD^x1) was established in combined whole-cell recording and Ca²⁺ imaging experiments, in which we stimulated cells synaptically with low intensities and monitored EPSP amplitudes before and after a conditioning stimulus consisting of a strong stimulation that evoked a local dendritic spike (Fig. 3A). All stimulations were sub-threshold for the initiation of back-propagating APs. This protocol led to a significant and long lasting depression of EPSP amplitude (Fig. 3A and B; 58 ± 4% of control level, n = 8; t test, P < 0.001). Similar results (t test, P > 0.18) were obtained for the change of the initial slope of the EPSP (Supplementary material Fig. 4; 49 ± 5% of control level, n = 8). Because conditioning stimulation with five dendritic spike/EPSP responses (delivered at intervals of 15 s) produced LTD of nearly the same amplitude (Fig. 3C, 59 ± 9% of control level, n = 5; t test, P < 0.001), we conclude that LTD^x1 is saturated for the activated input. To test whether the depression had a postsynaptic component we hyperpolarized the neurons to –80 mV during the conditioning stimuli (Fig. 3D). Because the activation of both NMDARs and voltage-sensitive calcium channels (VSCCs) is voltage dependent, hyperpolarization during conditioning stimuli prevents the induction of dendritic spikes (Schiller et al. 2000). Indeed, strong stimulation-evoked EPSPs obtained under these conditions had standard shapes, without a second depolarizing component and were not associated by detectable dendritic Ca²⁺ transients (not shown). In hyperpolarized neurons no significant change in EPSP amplitude after conditioning stimuli was measured (Fig. 3D, 103 ± 6% of control, n = 5; t test, P > 0.8). We conclude that coincident activation of several synaptic inputs to a circumscribed part of the dendritic tree depresses synaptic transmission via a mechanism that is postsynaptic, since it is affected by postsynaptic hyperpolarization and by the absence of postsynaptic Ca²⁺ signalling.

NMDA-receptor dependence of LTD induction

In a set of control experiments we tested the NMDA-receptor dependence of LTD induction. During the induction protocol we applied the NMDA-receptor antagonist APV locally using a puff pipette which restricted APV to the site of stimulation (Fig. 4A, left panel). The focal application of APV reduced calcium accumulation in both basal and apical dendrites to 12 ± 2% (n = 6) of control amplitude (t test, P < 0.001) (Fig. 4B). Under these conditions, no significant change in synaptic transmission strength could be detected (Fig. 4A, right panel; 90 ± 13% of control level, n = 6, t test, P > 0.4). In addition, we wanted to exclude the possibility that induction of dendritic spikes was due to direct stimulation of the dendrites by the stimulation pipette. Indeed, blocking excitatory and inhibitory synaptic receptors abolished both calcium accumulations and voltage deflections, indicating that they were of synaptic origin (Fig. 4C).

Input specificity of LTD induction

As the local dendritic spikes had limited spatial extent, we then inquired whether the synaptic depression they elicited was also spatially restricted. If the synaptic depression depends on NMDARs or VSCCs, it should be restricted to those inputs that experienced the spike and not spread to other parts of the dendritic tree. To answer this we stimulated the cell locally via two spatially separated pathways, by positioning stimulating electrodes in the apical or basal dendritic tree (Fig. 4D). We stimulated basal dendrites with the conditioning stimuli described above, while using a standard synaptic stimulation to monitor the EPSP size in apical dendrites. At the conditioned pathway we found a significant and long lasting depression of EPSP amplitude (Fig. 4Ea, 64 ± 13% of control level, n = 4; t test, P < 0.05). At the same time, the unconditioned pathway, in the apical dendrites, showed no significant change in synaptic transmission strength (Fig. 4Eb, 103 ± 8% of control level, n = 4, t test, P > 0.8). We concluded that dendritic spikes produce a locally restricted synaptic depression. Since the region of the dendrite that experiences the spike (and its accompanied calcium accumulation) is restricted to a small segment, we think that it is likely that only inputs located in this dendritic segment are depressed.

	Discussion

Top Abstract Introduction Methods Results Discussion Supplementary material References

 
Characteristics of local spikes

We describe two regimes of dendritic activation under single-shock synaptic stimulation of layer 5 neocortical pyramidal neurons. In the first regime, which causes subthreshold EPSPs, no dendritic calcium accumulations could be detected. This regime has been previously well documented in CA1 and cortical pyramidal neurons (Yuste & Denk, 1995; Köster & Sakmann, 1998) and the spatial restriction of calcium transients to individual spines without a detectable dendritic calcium transient is explained by the combination of local influx and efflux mechanisms (Köster & Sakmann, 1998; Yuste et al. 1999, 2000; Mainen et al. 1999; Kovalchuk et al. 2000; Majewska et al. 2000a; Sabatini et al. 2002). In addition, beyond a certain threshold of stimulation intensity, a calcium transient occurs in dendrites, concomitantly with EPSPs that have a late depolarizing component. These complex EPSPs are triggered in an all-or-none fashion and thus seem to be regenerative spikes that occur far from the somatic electrode (Llinas & Sugimori, 1980). Consistent with this is the spatial spread of calcium accumulations that occurs simultaneously with these complex EPSPs: only local regions of the dendritic tree are activated and these regions are distant from the soma. The activated dendrites involve both spines and dendritic shafts. A similar phenomenology has been recently described in experiments involving glutamate uncaging and synaptic stimulation (Schiller & Schiller, 2001). In this study, the authors argued that local dendritic spikes were due to the regenerative current through NMDARs, together with a contribution from VSCCs. In agreement with this view, our data indicate that NMDARs are necessary to generate the local dendritic spikes and local calcium accumulations. The APV blockade of both the late component of the EPSPs and the local calcium accumulations demonstrates that NMDARs are necessary for both phenomena.

Local dendritic spikes induce long-term synaptic depression

Dendrites of pyramidal neurons can also sustain regenerative sodium APs (Stuart & Sakmann, 1994). These APs can initiate in the dendrite (Golding & Spruston, 1998) or ‘backpropagate’ from the axon initial segment (Stuart & Sakmann, 1994). Differently from local dendritic spikes described here, these sodium spikes propagate through large regions of the dendritic tree, invading all spines (Yuste & Denk, 1995), although the extent of their propagation in vivo is still under debate (Helmchen et al. 1999; Waters et al. 2003). By propagating rapidly throughout the dendritic tree, these sodium spikes provide spines with information about the output of the neuron (Yuste & Denk, 1995). Therefore, one function of these sodium spikes might be to control the sign and magnitude of long-term synaptic plasticity (Magee & Johnston, 1997). Indeed, the timing of the dendritic action potential relative to the EPSP can trigger depression or potentiation (Markram et al. 1997), and therefore implement Hebbian learning rules (Hebb, 1949). This spike-timing dependent plasticity (STDP) underlies activity-dependent synapse rearrangement in a variety of systems (Bi & Poo, 1999) and has theoretical interest for network dynamics (Song et al. 2000).

In contrast to the global effects (Bi & Poo, 1999) of dendritic sodium APs, we find that local dendritic spikes induce a form of synaptic depression, which is spatially restricted. Moreover even a single individual spike in a dendritic region can trigger a major, and saturating (50%), reduction in EPSP amplitude. This form of LTD is input specific, since it does not affect distant inputs. Our spatial mapping experiments indicate that these events are restricted to individual dendrites. Therefore, and because local application of APV blocks both the dendritic spike and the induction of LTD, it seems very likely that the LTD that they produce is also spatially restricted to these dendrites. Nevertheless, we cannot exclude the possibility that synapses outside of the region of the localized calcium change also underwent LTD, although it would be surprising since they did not undergo any change in intracellular free calcium concentration. A recent study (Golding et al. 2002) has demonstrated that multiple dendritic spikes, produced by afferent theta burst stimulation, evoke long-term potentiation (LTP) in hippocampal pyramidal neurons. In line with their findings we obtained preliminary results suggesting that also in cortical pyramidal cells dendritic spike-producing repetitive stimulation may evoke LTP (K. Holthoff, Y. Kovalchuk and A. Konnerth, unpublished observations).

We conclude that local dendritic spike-mediated permanent decrease in synaptic weight is NMDAR dependent, input specific, and is spatially restricted to the region where the local spike takes place. Several lines of evidence indicate that the local spike is required for LTD induction. Firstly, the spike always preceded depression of the EPSPs. Secondly, postsynaptic hyperpolarization to –80 mV, which is known to block the induction of NMDA spikes (Schiller et al. 2000), entirely prevented the occurrence of LTD. Thirdly, in all experiments in which, despite strong stimulation, we failed to evoke the local dendritic calcium transient, LTD did not occur (n = 3). Fourthly, local application of APV to the stimulated dendrite prevented both the dendritic calcium transient and LTD induction. Our data are consistent with earlier evidence from experiments using postsynaptic uncaging of Ca²⁺ showing that a single and short postsynaptic Ca²⁺ transient is sufficient for LTD or LTP induction (Neveu & Zucker, 1996). Although in that study the precise parameters of the artificially produced Ca²⁺ signal (amplitude, duration, site of occurrence) underlying the change in transmission properties were not determined and could have led to the opposing forms of plasticity, we consider it likely that some of the regimes revealed in these uncaging experiments correspond in calcium terms to the dendritic spikes we have studied.

Computational function of local dendritic spikes

What could be the function of the local dendritic spike-mediated LTD? First of all, it is remarkable that this LTD occurs in a single shock. We would highlight the fact that current protocols to elicit long-term plasticity in central synapses always require many-fold repetitions (Bear & Abraham, 1996) of pairing protocols that may not be entirely physiological (Goldberg et al. 2002). The rapidity of the induction of the synaptic change we observe could be due to the high calcium concentrations reached during the NMDA spike. Regardless of its biochemical mechanisms, this instantaneous and massive change in synaptic transmission could be used to implement rapid circuit plasticity or circuit accommodation. On-line learning is a feature of all nervous systems and has normally been ascribed to short-term presynaptic plasticity (Magleby, 1987). Local dendritic spikes could enable instantaneous, yet long-lasting, synaptic plasticity.

	Supplementary material

Top Abstract Introduction Methods Results Discussion Supplementary material References

 
The online version of this paper can be accessed at:

DOI: 10.1113/jphysiol.2004.072678
http://jp.physoc.org/cgi/content/full/jphysiol.2004.072678/DC1
and contains supplementary material.

This material can also be found at:

http://www.blackwellpublishing.com/products/journals/suppmat/tjp/tjp507/tjp507sm.htm

	Footnotes

 
Both laboratories contributed equally to this research.

	References

Top Abstract Introduction Methods Results Discussion Supplementary material References

 
Amitai Y, Friedman A, Connors BW & Gutnick MJ (1993). Regenerative activity in apical dendrites of pyramidal cells in neocortex. Cereb Cortex 3, 26–38.[Abstract/Free Full Text]

Bear MF & Abraham WC (1996). Long-term depression in hippocampus. Annu Rev Neurosi 19, 437–462.[CrossRef][Medline]

Bi G & Poo M (1999). Distributed synaptic modification in neural networks induced by patterned stimulation. Nature 401, 792–796.[CrossRef][Medline]

Garaschuk O, Yaari Y & Konnerth A (1997). Release and sequestration of calcium by ryanodine-sensitive stores in rat hippocampal neurons. J Physiol 502, 13–30.[Abstract/Free Full Text]

Goldberg J, Holthoff K & Yuste R (2002). A problem with Hebb and local spikes. Trends Neurosci 25, 433–435.[CrossRef][Medline]

Golding NL & Spruston N (1998). Dendritic sodium spikes are variable triggers of axonal action potentials in hippocampal CA1 pyramidal neurons. Neuron 21, 1189–1200.[CrossRef][Medline]

Golding NL, Staff NP & Spruston N (2002). Dendritic spikes as a mechanism for cooperative long-term potentiation. Nature 418, 326–331.[CrossRef][Medline]

Hebb DO (1949). The Organization of Behaviour. Wiley, New York.

Helmchen F, Svoboda K, Denk W & Tank DW (1999). In vivo dendritic calcium dynamics in deep-layer cortical pyramidal neurons. Nat Neurosci 2, 989–996.[CrossRef][Medline]

Köster HJ & Sakmann B (1998). Calcium dynamics in single spines during coincident pre- and postsynaptic activity depend on relative timing of back-propagating action potentials and subthreshold excitatory postsynaptic potentials. Proc Natl Acad Sci U S A 95, 9596–9601.[Abstract/Free Full Text]

Kovalchuk Y, Eilers J, Lisman J & Konnerth A (2000). NMDA receptor-mediated subthreshold Ca(2+) signals in spines of hippocampal neurons. J Neurosci 20, 1791–1799.[Abstract/Free Full Text]

Llinas R & Sugimori M (1980). Electrophysiological properties of in vitro Purkinje cell somata in mammalian cerebellar slices. J Physiol 305, 171–195.[Abstract/Free Full Text]

Magee JC & Johnston D (1997). A synaptically controlled, associative signal for Hebbian plasticity in hippocampal neurons. Science 275, 209–213.[Abstract/Free Full Text]

Magleby KL (1987). Short-term change in synaptic efficacy. In Synaptic Function, ed. Edelman GM, Gall WE & Cowan WM, pp. 21–56. Wiley & Sons, New York.

Mainen ZF, Malinow R & Svoboda K (1999). Synaptic calcium transients in single spines indicate that NMDA receptors are not saturated. Nature 399, 151–155.[CrossRef][Medline]

Majewska A, Brown E, Ross J & Yuste R (2000a). Mechanisms of calcium decay kinetics in hippocampal spines: role of spine calcium pumps and calcium diffusion through the spine neck in biochemical compartmentalization. J Neurosci 20, 1722–1734.[Abstract/Free Full Text]

Majewska A, Yiu G & Yuste R (2000b). A custom-made two-photon microscope and deconvolution system. Pflugers Arch 441, 398–408.[CrossRef][Medline]

Maravall M, Mainen ZF, Sabatini BL & Svoboda K (2000). Estimating intracellular calcium concentrations and buffering without wavelength ratioing. Biophys J 78, 2655–2667.[Abstract/Free Full Text]

Markram H, Lübke J, Frotscher M & Sakmann B (1997). Regulation of synaptic efficacy by coincidence of postsynaptic APs and EPSPs. Science 275, 213–215.[Abstract/Free Full Text]

Neveu D & Zucker RS (1996). Postsynaptic levels of [Ca²⁺]_i needed to trigger LTD and LTP. Neuron 16, 619–629.[CrossRef][Medline]

Pockberger H (1991). Electrophysiological and morphological properties of rat motor cortex neurons in vivo. Brain Res 539, 181–190.[CrossRef][Medline]

Ramon y Cajal S (1904). La Textura Del Sistema Nerviosa Del Hombre Y Los Vertebrados, 1st edn, Moya, Madrid.

Sabatini BL, Örtner TG & Svoboda K (2002). The life cycle of Ca²⁺ ions in dendritic spines. Neuron 33, 439–452.[CrossRef][Medline]

Schiller J, Major G, Köster HJ & Schiller Y (2000). NMDA spikes in basal dendrites of cortical pyramidal neurons. Nature 404, 285–289.[CrossRef][Medline]

Schiller J & Schiller Y (2001). NMDA receptor-mediated dendritic spikes and coincident signal amplification. Curr Opin Neurobiol 11, 343–348.[CrossRef][Medline]

Schiller J, Schiller Y, Stuart G & Sakmann B (1997). Calcium action potentials restricted to distal apical dendrites of rat neocortical pyramidal neurons. J Physiol 505, 605–616.[CrossRef][Medline]

Song S, Miller KD & Abbott LF (2000). Competitive Hebbian learning through spike-timing-dependent synaptic plasticity. Nat Neurosci 3, 919–926.[CrossRef][Medline]

Stuart GJ & Sakmann B (1994). Active propagation of somatic action potentials into neocortical pyramidal cell dendrites. Nature 367, 69–72.[CrossRef][Medline]

Waters J, Larkum M, Sakmann B & Helmchen F (2003). Supralinear Ca²⁺ influx into dendritic tufts of layer 2/3 neocortical pyramidal neurons in vitro and in vivo. J Neurosci 23, 8558–8567.[Abstract/Free Full Text]

Wei DS, Mei YA, Bagal A, Kao JP, Thompson SM & Tang CM (2001). Compartmentalized and binary behavior of terminal dendrites in hippocampal pyramidal neurons. Science 293, 2272–2275.[Abstract/Free Full Text]

Wong RK, Prince DA & Basbaum AI (1979). Intradendritic recordings from hippocampal neurons. Proc Natl Acad Sci U S A 76, 986–990.[Abstract/Free Full Text]

Yuste R & Denk W (1995). Dendritic spines as basic functional units of neuronal integration. Nature 375, 682–684.[CrossRef][Medline]

Yuste R, Majewska A, Cash SS & Denk W (1999). Mechanisms of calcium influx into hippocampal spines: heterogeneity among spines, coincidence detection by NMDA receptors, and optical quantal analysis. J Neurosci 19, 1976–1987.[Abstract/Free Full Text]

Yuste R, Majewska A & Holthoff K (2000). From form to function: calcium compartmentalization in dendritic spines. Nat Neurosci 3, 653–659.[CrossRef][Medline]

	Acknowledgements

 
This work was funded by the Deutsche Forschungsgemeinschaft, the NEI, the NINDS, the NYSTAR Center for High Resolution Imaging of Functional Neural Circuits and the HFSP. We thank S. Siegelbaum for comments on an earlier version of the manuscript.

This article has been cited by other articles:

				G. Major, A. Polsky, W. Denk, J. Schiller, and D. W. Tank Spatiotemporally Graded NMDA Spike/Plateau Potentials in Basal Dendrites of Neocortical Pyramidal Neurons J Neurophysiol, May 1, 2008; 99(5): 2584 - 2601. [Abstract] [Full Text] [PDF]

				P. J. Sjostrom, E. A. Rancz, A. Roth, and M. Hausser Dendritic Excitability and Synaptic Plasticity Physiol Rev, April 1, 2008; 88(2): 769 - 840. [Abstract] [Full Text] [PDF]

				B. A. Milojkovic, W.-L. Zhou, and S. D. Antic Voltage and calcium transients in basal dendrites of the rat prefrontal cortex J. Physiol., December 1, 2007; 585(2): 447 - 468. [Abstract] [Full Text] [PDF]

				S. Remy and N. Spruston Dendritic spikes induce single-burst long-term potentiation PNAS, October 23, 2007; 104(43): 17192 - 17197. [Abstract] [Full Text] [PDF]

				M. Canepari, M. Djurisic, and D. Zecevic Dendritic signals from rat hippocampal CA1 pyramidal neurons during coincident pre- and post-synaptic activity: a combined voltage- and calcium-imaging study J. Physiol., April 15, 2007; 580(2): 463 - 484. [Abstract] [Full Text] [PDF]

				U. Gordon, A. Polsky, and J. Schiller Plasticity Compartments in Basal Dendrites of Neocortical Pyramidal Neurons J. Neurosci., December 6, 2006; 26(49): 12717 - 12726. [Abstract] [Full Text] [PDF]

				R. Araya, K. B. Eisenthal, and R. Yuste Dendritic spines linearize the summation of excitatory potentials PNAS, December 5, 2006; 103(49): 18799 - 18804. [Abstract] [Full Text] [PDF]

				E. A. Rancz and M. Hausser Dendritic calcium spikes are tunable triggers of cannabinoid release and short-term synaptic plasticity in cerebellar Purkinje neurons. J. Neurosci., May 17, 2006; 26(20): 5428 - 5437. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Supplemental data All Versions of this Article: 560/1/27    most recent jphysiol.2004.072678v1 Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Holthoff, K. Articles by Konnerth, A. Search for Related Content PubMed PubMed Citation Articles by Holthoff, K. Articles by Konnerth, A.