Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
Nat Neurosci. Author manuscript; available in PMC 2013 Jan 1.
Published in final edited form as:
PMCID: PMC3386897

Release probability of hippocampal glutamatergic terminals scales with the size of the active zone


Cortical synapses display remarkable structural, molecular and functional heterogeneity. Our knowledge regarding the relationship between the ultrastructural and functional parameters is still fragmented. Here we asked how the release probability and presynaptic [Ca2+] transients relate to the ultrastructure of rat hippocampal glutamatergic axon terminals. Two-photon Ca2+ imaging-derived optical quantal analysis and correlated electron microscopic reconstructions revealed a tight correlation between the release probability and the active zone area. The peak amplitude of [Ca2+] transients in single boutons also positively correlated with the active zone area. Freeze-fracture immunogold labeling revealed that the voltage-gated Ca2+ channel subunit Cav2.1 and the presynaptic protein Rim1/2 are confined to the active zone and their numbers scale linearly with the active zone area. Gold particles for Cav2.1 showed a nonrandom distribution within the active zones. Our results demonstrate that the number of several active zone proteins, including presynaptic Ca2+ channels, docked vesicles and the release probability scales linearly with the active zone area.


Synapses display enormous diversity in their structure and function. Even in a homogeneous population of connections between two well-defined cell types, large variability is found in the shape and size of pre- and postsynaptic elements1,2. Similar to the structural diversity, many functional parameters such as the probability of the transmitter release, the size of the readily releasable pool of vesicles, short- and long-term plasticity of the synapses and the size of the postsynaptic responses display large heterogeneity3. Although tremendous efforts have been invested in the past decades, the contribution of differences in the synaptic ultrastructure or the molecular composition of synapses to generating functional diversity is still debated. An example of an apparent relationship between synapse ultrastructure and function comes from studies examining the numbers and densities of postsynaptic AMPA and NMDA-type glutamate receptors. Both receptor types are concentrated in the postsynaptic density (PSD), and their numbers scale with the PSD area4-6. As a consequence, larger postsynaptic responses are generated in larger spines7,8 with larger PSDs9. Clearly, the actual size of the PSD has very little direct effect on the amplitude of the postsynaptic response, but because the PSD area tightly correlates with the number of postsynaptic glutamate receptors, it can be used to predict the size of the postsynaptic response. Can a similar ultrastructural feature that is indicative of function be found on the presynaptic side? Most presynaptic axons that were initially examined to address this question are large, contain a large number of active zones and are specialized for rapid, reliable excitation of their postsynaptic partners (e.g. neuromuscular junction, squid giant axon, calyx of Held). Many of the classical papers indicated a correlation between the structure and function of these large axons/axon terminals10-13. However, most axon terminals in the CNS are small, often bear a single active zone and cannot efficiently drive their postsynaptic target cells. This is particularly true for cortical glutamatergic terminals, which are in the order of 1 μm in diameter; most of them have a single active zone; and their release probabilities range from 0 to 0.9. The long-standing question is to what extent can the morphological parameters of cortical glutamatergic terminals predict their functional properties? Studies of cultured hippocampal neurons suggested that the release probability correlates with the number of readily releasable vesicles, which corresponds to vesicles docked at the presynaptic active zone14,15. In the same synapse population, the number of docked vesicles has been found to be proportional to the active zone area and to the volume of the terminal1,2, implying that the larger the presynaptic terminal, the higher its release probability is. The general applicability of such a clear correlation was challenged by data showing that glutamatergic terminals of cortical and hippocampal pyramidal cells, which are presynaptic to distinct types of GABAergic interneurons, possess widely different release probabilities and short-term plasticity patterns16-20, but have very similar sizes. Similarly, axon terminals of the cerebellar cortex with similar active zone sizes and number of docked vesicles display dramatically different release probabilities21, indicating that the molecular composition of the presynaptic terminal must play a key role in determining the functional properties. To overcome potentially diversifying factors we decided to examine a homogeneous population of axon terminals of CA3 pyramidal cells synapsing on other CA3 pyramidal cells in acute hippocampal slices. We addressed the relationship between the ultrastructure, molecular composition and function of these terminals by combining in vitro patch-clamp electrophysiology, two-photon fluorescent Ca2+ imaging with multiple line scanning, post hoc electron microscopic 3D reconstruction of the imaged pre- and postsynaptic structures and quantitative, SDS-digested freeze-fracture replica immunogold localization (SDS-FRL) of presynaptic proteins.


Morphological diversity of hippocampal CA3 pyramidal cell axon terminals

Schaffer collateral synapses made by CA3 pyramidal cell axons onto CA1 pyramidal cell dendrites differ greatly from one another with respect to their morphological and functional parameters1,2. However, the long distance between the cell bodies of CA3 pyramidal cells and their boutons in the CA1 area imposes a long diffusion time of Ca2+ sensitive fluorescent dyes to the terminals when introduced somatically. We therefore chose the local axon collaterals of pyramidal cells in the CA3 stratum oriens within 150-300 μm from the somata. These axon terminals establish synapses both on the basal dendritic spines of other CA3 pyramidal cells and on local GABAergic interneurons. In the present work, we focused exclusively on pyramidal cell – pyramidal cell synapses. Because quantitative information regarding the morphological properties of these synapses was not available, we first performed electron microscopic 3D reconstructions from serial ultrathin sections of a 16 day-old rat. Figure 1 illustrates local axon terminals and their postsynaptic pyramidal cell spines, displaying substantial variability in the volume of the boutons (0.18 ± 0.14 μm3, n = 68, Fig. 1h), in the area of the active zones (0.04 ± 0.03 μm2, n = 83, Fig. 1i) and in the volume of the postsynaptic spines (0.04 ± 0.04 μm3, n = 83, Fig. 1j). The majority of the boutons formed synapses with only a single postsynaptic partner, whereas approximately one fifth of the boutons had more than one postsynaptic target (13.2% had two, 4.4% had three, 1.5% had four). Individual active zone areas showed only a weak positive correlation with the volume of the bouton (Fig. 1k) and the volume of the presynaptic terminal poorly predicted the volume of the postsynaptic spine (Fig. 1l). However, the total active zone area (calculated by summing up all active zones per bouton) displayed a tighter correlation with the volume of the boutons (R = 0.68). Virtually all CA3 pyramidal cell spines had a single presynaptic partner, and consequently had a single PSD, the size of which reliably predicted the volume of the spine (Fig. 1m). Axon terminals contain a large number of small synaptic vesicles. Few of these vesicles are within few nanometers from the presynaptic active zone membrane (docked vesicles, Fig. 2a, b) and may correspond to the functionally defined readily releasable pool. Most vesicles are tens and hundreds of nanometer away from the active zone and might comprise the reserve pool. For assessing the number of docked vesicles possibly the best method is electron microscopic tomography. However, a reliable estimate can also be obtained with electron microscopic 3D reconstruction from very thin sections. We chose the second method and reconstructed 15 synapses (n = 1 animal) from 20 nm thin sections (Fig. 2c) and found that the number of docked vesicles positively correlates with the active zone area (Fig. 2d). The ultrastructural parameters described above are very similar to those published previously for Schaffer collateral synapses in the CA1 area1,2 and also predict a large heterogeneity in their functional properties.

Figure 1
Morphological diversity of glutamatergic axon terminals in the CA3 region of the rat hippocampus. (a-c) Electron micrographs show synaptic contacts of axon collaterals (b, blue) of CA3 pyramidal cells and their postsynaptic spines (s, yellow). Terminals ...
Figure 2
The number of the docked vesicles correlates linearly with the size of the active zone. (a) An electron microscopic image of a synaptic contact between a CA3 pyramidal cell spine (s) and axon terminal (t) in the str. oriens taken from a 20 nm thick electron ...

Large variability in release probability and short-term plasticity of axon terminals

To measure the release probability and short-term plasticity of CA3 pyramidal cell local axon collaterals, we performed two-photon imaging of [Ca2+] transients in individual dendritic spines in response to paired extracellular stimulation. By holding pyramidal cells at −30 mV, Ca2+ fluxing through NMDA receptors following the presynaptic release of glutamate could be efficiently detected in single spines and the successful events could be reliably distinguished from the failures (Fig. 3a-c; Supplementary Fig. 1). Repeating the stimulations 50 – 120 times allowed the determination of the release probability22. This release probability denotes to the probability that neurotransmitter release occurs from a given active zone and it is different from vesicle release probability. To increase the chance of finding spines postsynaptic to activated boutons, we placed a unipolar stimulating electrode a few tens of microns away from the monitored basal dendrite and simultaneously imaged 6 – 12 spines within a dendritic segment using a line scanning mode (Fig. 3d, e). Figure 3d-h shows a typical example of simultaneously monitoring 9 spines and 2 dendritic shaft segments. Rapidly rising, transient increases in [Ca2+] were visible (detected with 100 μM Fluo5F, a pseudo-colored single trial response is shown in Fig. 3e) in 2 out of 9 spines, whereas delayed, slow rising, small amplitude transients could be detected in some of the neighboring spines and dendritic shafts. Boutons with different release probabilities established synaptic contacts with spines close to each other on a short dendritic segment (Fig. 3g-h, e.g. spine #3: release probability = 0.02, spine #9: release probability = 0.11). Examining 43 activated spines (in 14 animals) revealed a great variability in the release probability for the 1st (PR1st) and 2nd (PR2nd) stimuli among axon terminals (PR1st: range: 0 – 0.53, mean ± SD = 0.14 ± 0.14; PR2nd range: 0.03 – 0.48, mean ± SD = 0.19 ± 0.12, n = 43 spines; Fig 3i). We also determined the paired-pulse plasticity (at 100 ms inter-stimulus interval) of release by dividing the PR2nd with PR1st for each bouton. Our results demonstrate a wide range of paired-pulse ratios (PPR, 0.43 – 10) with an average of ~2-fold facilitation (mean ± SD = 2.2 ± 2.0, n = 40 synapses; three synapses had a PR1st of 0). There was an inverse correlation between the PR1st and the PPR for these axon terminals; synapses with a PR1st < 0.1 always displayed paired-pulse facilitation (mean PPR = 3.15), whereas those with a PR1st > 0.4 always showed paired-pulse depression (mean PPR = 0.71). Our results revealed a large heterogeneity in the functional properties of CA3 pyramidal cell local axon collaterals, similar to that of Schaffer collaterals in acute slices and cultures.

Figure 3
Determining the release probability of axon terminals with optical quantal analysis. (a) Changes in fluorescent intensity measured with line scans in a representative spine head of a CA3 pyramidal cell basal dendrite as a consequence of Ca2+ influx through ...

Having established a large diversity in both the ultrastructural and functional properties of CA3 pyramidal cell axon terminals, next we examined the correlations among the measured structural and functional parameters. In addition to the Ca2+-sensitive (Fluo5F, 100 μM) and insensitive (20 μM Alexa594) fluorescent dyes, we also included biocytin in the intracellular solution, allowing post hoc light microscopic and subsequent electron microscopic visualization of the imaged dendritic segments. Seven spines and their parent dendrite (Fig. 4a) were simultaneously scanned, out of which two spines (s2 and s4) had fast rising [Ca2+] transients in response to extracellular stimulations. Bouton #4 had a low PR1st (0.08; Fig. 4b), whereas bouton #2 had relatively high PR1st (0.33; Fig. 4c). The whole dendritic segment was electron microscopic 3D reconstructed from serial ultrathin sections (Fig. 4d, e) and all seven imaged spines were identified on the reconstructed profile (Fig. 4f). Following the identification of the activated spines (s2 and s4), we also reconstructed the unlabeled axon terminals that made synaptic contacts onto these spines (blue profiles in Fig. 4d-f). Both spines had only a single presynaptic partner. Our post hoc 3D reconstructions revealed differences in both the volume of the boutons and the area of the active zones. Examining the relationships between the PR1st and the volume of the boutons (Fig. 4g) and between the PR1st and the area of the active zone (Fig. 4h) for all reconstructed synaptic partners revealed significant (p < 0.01) positive correlations (n = 14 synapses in 6 animals).

Figure 4
Post hoc ultrastructural analysis of synapses following two-photon imaging. (a) A two-photon image of a CA3 pyramidal cell basal dendritic segment with multiple spines. Two out of seven spines (s2 and s4) displayed [Ca2+] transients in response to extracellular ...

Variability in [Ca2+] in CA3 pyramidal cell axon terminals

The probability of transmitter release depends on the spatio-temporal profile of free [Ca2+] around the docked vesicles and the binding rate of Ca2+ to the sensor. Despite the wealth of information regarding the relationship between [Ca2+] and release obtained at the neuromuscular synapse, the squid giant axon, the calyx of Held and hippocampal basket cell terminals23-29, almost nothing is known about these parameters at cortical glutamatergic terminals. Koester and Sakmann30 measured volume-averaged [Ca2+] in cortical layer 2/3 pyramidal cell axon terminals and found a greater than 10-fold variability in the peak [Ca2+]. Because direct measurements of the release probability and AP-evoked [Ca2+] in individual small cortical glutamatergic terminals is not possible with currently available techniques in acute slices, we chose a strategy to determine these functional parameters in separate sets of experiments and relate both of them to the same set of ultrastructural parameters. We measured single AP-evoked [Ca2+] transients in the boutons of CA3 pyramidal cell local axon collaterals (See Methods and Supplementary Figs. 2 and 3) in the stratum oriens followed by post hoc electron microscopic 3D reconstruction of the imaged boutons. The peak amplitude of the [Ca2+] transients displayed large (8-fold) variability among the boutons (mean ± SD = 0.16 ± 0.07 G/GMAX; coefficient of variation = 0.45, n = 84 boutons in 4 animals; Fig. 5a-e). Following fixation, we visualized the biocytin and identified the individual imaged boutons at the light microscopic level (Fig. 5f) and also on serial electron microscopic sections. For 27 boutons (from 4 animals), we could also faithfully identify pyramidal cell spines as their single postsynaptic targets (Fig. 5h, i) and perform full 3D reconstructions of both the pre- and postsynaptic elements. Two boutons with very similar volumes (marked by blue and green dots, 0.22 and 0.24 μm3, respectively, Figure 5g-i) displayed an ~2.5-fold difference in their peak [Ca2+] transients. In agreement with these examples, we could not find a significant correlation between the bouton volume and the amplitude of the [Ca2+] transients (Fig. 5j). However, when we examined the correlation between the amplitude of the [Ca2+] transients and the active zone area (Fig. 5k) and the total amount of fluxed Ca2+ per AP and the active zone area (Fig. 5l), significant (p < 0.01) positive correlations were found. The two boutons illustrated in figure 5h and i had an active zone area ratio of 2.4, a peak [Ca2+] ratio of 2.5 and a total fluxed Ca2+ ratio of 2.3.

Figure 5
Measurement of volume averaged [Ca2+] transients in CA3 pyramidal cell local axon terminals. (a) Two-photon image of a CA3 pyramidal cell filled with Alexa594 (red) and Fluo5F. (b) Higher magnification view of two boutons. White lines indicate the position ...

Contribution of Ca2+ channel subtypes to [Ca2+] in boutons

Excitatory postsynaptic responses of hippocampal glutamatergic synapses are mediated by a concerted action of several types of voltage-gated Ca2+ channels (Cav), each with distinct pharmacological profiles31-33. Here we examined the direct contribution of these channel subtypes to [Ca2+] transients measured in individual boutons of CA3 pyramidal cells (Supplementary Fig. 4). The Cav2.1 (P/Q-type) subunit-specific blocker ω-ATX-IVA (1 μM) and the Cav2.2 (N-type) subunit-specific blocker ω-CTX-GVIA (1 μM) reduced the peak amplitude of the [Ca2+] transients by 28.8 ± 1.1 % (Supplementary Fig. 4a, n = 67 boutons in 4 cells from 4 animals, p<0.0001) and 35.7 ± 12 % (Supplementary Fig. 4b, n = 71 boutons in 4 cells from 4 animals, p<0.0001), respectively. Application of SNX482 (500 nM), a Cav2.3 (R-type) subunit-specific blocker, resulted in a 14.7 ± 2.3 % reduction in the peak amplitude (Supplementary Fig. 4c, n = 40 boutons in 3 cells from 3 animals, p<0.01). These results on the pharmacological profile of the presynaptic [Ca2+] transients are in agreement with previous data showing a major role of the Cav2.1, Cav2.2 and Cav2.3 subunits in mediating postsynaptic responses at hippocampal glutamatergic terminals32 and that of bulk Ca2+ imaging of Schaffer collateral terminals31. Following the blockade of Cav2.2 and Cav2.3 subunits, when the majority of the Ca2+ influx is mediated by the Cav2.1 subunit, the amplitude of the [Ca2+] transients was reduced by 50.4 ± 3.9 % (n = 39 boutons in 2 cells from 2 animals), but the bouton-to-bouton variability of the responses remained very similar (coefficient of variation = 0.51), demonstrating that the large variability in peak [Ca2+] is also evident under conditions when mainly Cav2.1 channels mediate the Ca2+ influx.

The Cav2.1 subunit is confined to the presynaptic active zone

Assuming that Cav channels are confined to the active zone membrane of the bouton, our Ca2+ imaging experiments, revealing that the total amount of fluxed Ca2+ following an AP is proportional to the active zone area, predict that the number of Cav channels scales linearly with the active zone area. However, direct experimental data for cortical excitatory axon terminals was not available regarding this important assumption. Therefore, we carried out high-resolution SDS-FRL of the Cav2.1 subunits in the stratum oriens of adult and 16 days-old rats (Fig. 6). In this technique, following rapid freezing and fracturing of brain tissue, carbon-platinum-carbon replicas are created from the fractured tissue surfaces. Fractured membranes containing transmembrane proteins attached to the replica will be retained following digestion with SDS, and can be visualized with electron microscopic immunogold techniques (see Methods, reviewed by ref. 34). In adult rats, immunogold particles labeling the Cav2.1 subunit on protoplasmic (P-face) plasma membranes were highly enriched in presynaptic active zones compared to extrasynaptic bouton surfaces (Fig. 6a). A similar enrichment of the Cav2.1 subunit in the active zones was found in rats age-matched (P16) to those used in our imaging experiments (Fig. 6b). The active zones were recognized from the clusters of intramembrane particles (IMP)35. To provide molecular evidence that the site of Ca2+ channel clustering is indeed the active zone, we performed double immunolabeling for the Cav2.1 subunit and presynaptic active zone proteins Rim1/2 (Fig. 6c) and found that both were confined to the same part of the presynaptic plasma membrane; to the active zone. Furthermore, we carried out double immunogold labeling for Cav2.1 and AMPA type glutamate receptors. Small gold particles labeling the Cav channels were accumulated in presynaptic specializations facing AMPA receptor immunopositive PSDs in synapses in which both the presynaptic P-face and the corresponding postsynaptic extracellular (E-face) membranes were exposed (Fig. 6d). Finally, we performed two series of experiments to validate the specificity of Cav2.1 immunolabeling in our replicas. First, we carried out SDS-FRL with another anti-Cav2.1 subunit antibody (rabbit) that was raised against a different epitope of the protein and obtained a qualitatively very similar result to that obtained with our guinea pig anti-Cav2.1 antibody. In double labeling reactions, the two antibodies labeled exactly the same subcellular compartment of CA3 pyramidal cell axon terminals (Fig. 6e). In a second set of experiments, we performed SDS-FRL on hippocampal tissue of Cav2.1+/− and Cav2.1−/− mice. The labeling pattern on Cav2.1+/− tissue was similar to that observed in Wistar rats (Fig. 6f); an extensive colocalization of the Cav2.1 was observed with Rim1/2. In contrast, Rim1/2 positive active zones in Cav2.1−/− mice were always immunonegative for the Cav2.1 subunit (Fig. 6g). Quantitative analysis revealed that in Cav2.1+/− tissue a total of 211 gold particles labeling Cav2.1 were detected in 39 active zones, whereas in Cav2.1−/− mice, zero particle was found in 24 Rim1/2-positive active zones. In agreement with our electron microscopic immunogold data, when slowly rising, small amplitude [Ca2+] transients were recorded in five presynaptic boutons, post hoc electron microscopic analysis revealed the lack of postsynaptic partners and active zones (Supplementary Fig. 5). The slow rise times and small amplitudes are inconsistent with the direct entry of Ca2+, but consistent with the diffusion of Ca2+ or Ca2+-bound Fluo5F from a neighboring bouton, which was just few microns away.

Figure 6
The Cav2.1 subunit is confined to the active zone of presynaptic axon terminals in the stratum oriens of the hippocampal CA3 area. (a, b) SDS-digested freeze-fracture replica-labeling of the Cav2.1 subunit on the P-face of axon terminals (t) in an adult ...

Rim1/2 and Cav2.1 have uniform densities in presynaptic active zones

Optical quantal analysis followed by post hoc electron microscopic 3D reconstructions demonstrated that the release probability scales quasi-linearly with the size of the presynaptic active zone. A prediction of this result is that large presynaptic active zones consist of multiple neurotransmitter release sites. If a functional release site is created by a defined number of molecules associated with the active zone, then the prediction of our results is that the number of presynaptic active zone proteins should scale linearly with the active zone area. We tested this prediction by performing quantitative electron microscopic immunogold localization of active zone proteins. First we tested the distribution of SNAP-25, a SNARE protein, and found that immunogold particles were present within and outside the active zones (Fig. 7a, b). Quantitative analysis revealed that the density of SNAP-25 in the active zone was very similar to that found on the non-synaptic part of the bouton (Fig. 7d), consistent with the results of Hagiwara et al.35. We also confirmed the results of Hagiwara et al.35 that syntaxin-1B is also present rather uniformly in the synaptic and non-synaptic plasma membranes. Synaptobrevin was mainly found on synaptic vesicles with a low intensity labeling on the plasma membranes. With such proteins, the above-mentioned prediction could not be tested. We therefore searched for presynaptic proteins with an exclusive active zone location. The most obvious one was the Cav2.1 subunit. A smaller number of gold particles was found in smaller active zones (Fig. 7e, f). There was a steep, significant positive linear correlation (p < 0.01 in both animals) between the number of gold particles and the active zone area (Fig. 7g). Quantification of the immunolabeling revealed a high density of gold particles in the active zones, but the extrasynaptic densities did not significantly differ (p = 0.73 and 0.06 in two animals) from the background (Fig. 7h). Finally, Rim1/2 was also restricted to the active zone (Fig. 7i-l). Similar to our results with the Cav2.1 subunit, the number of gold particles labeling Rim1/2 also displayed a tight, significant positive correlation (p < 0.01 in both animals) with the active zone area (Fig. 7k).

Figure 7
The number of Cav2.1 subunits and Rim1/2 proteins correlates with the active zone area. (a, b) SDS-FRL labeling of SNAP-25 in a large (a) and a small (b) active zone. (c) Correlation of the number of gold particles labeling SNAP-25 with the active zone ...

We also noticed that gold particles labeling Cav2.1 were not distributed uniformly within the active zones, but seemingly formed small clusters (e.g. Figs. 6a, ,7e,7e, ,8a).8a). To quantitatively analyze the distribution of Cav2.1 immunolabeling within the active zone (Fig. 8), we measured all inter-gold distances within individual active zones and asked the question whether the inter-gold distance distribution was significantly different from random distributions (Fig. 8a3, b3, c4). For this, we generated 50 random distributions for each synapse having exactly the same number of ‘gold’ particles as determined experimentally in the same synapse (Fig. 8a4, b4, c3) and calculated the distributions of the simulated inter-‘gold’ distances. For large active zones (>0.1 μm2) where a large number of gold particles were detected, in >50% of the active zones the inter-gold distance distributions were significantly (p < 0.05, for all 50 random distributions) different from random distributions. This proportion was 26% for all examined presynaptic active zones (n = 34; Fig. 8d). For many small active zones with a small number of gold particles, the inter-gold distance distributions could be either significantly or non-significantly different from random distributions (Fig. 8b3, c4).

Figure 8
Non-random distribution of the Cav2.1 subunits within the active zones. (a-c) Examples of different gold particle distributions within a large (a) and two small (b and c) active zones. (a1) SDS-FRL replica image showing the distribution of gold particles ...


In the present manuscript, we describe a large heterogeneity in the morphological, molecular and functional properties of local axon collaterals of CA3 pyramidal cells. We reveal that the release probability of individual terminals as well as the total amount of fluxed Ca2+ following an AP scales linearly with the active zone area. We also found that the Cav2.1 subunits are concentrated in the presynaptic active zone, and their numbers together with that of Rim1/2 are proportional to the active zone area. These results, taken together are consistent with a modular model of the active zone36, where the number of functional release sites scales with the active zone area and, as a consequence, enables release at multiple sites at large pyramidal cell – pyramidal cell synapses.

Our results obtained with functional and morphological examinations of CA3 pyramidal cell – pyramidal cell glutamatergic synapses revealed large structural and functional heterogeneities within this apparently homogeneous synapse population. It has been known that both the pre- and postsynaptic properties of central synapses depend on the cellular identity of the synapsing partners. For example, the release probability and short-term plasticity of release from hippocampal and cortical pyramidal cell axon terminals depends on the target cell type; pyramidal cells evoke facilitating EPSPs on hippocampal OLM and neocortical bitufted interneurons, whereas the same presynaptic pyramidal cells innervate parvalbumin positive basket cells with depressing EPSPs16-20. These target cell type-specific differences have been shown to be reflected in the amplitude of [Ca2+] transients in the presynaptic boutons18. The identity of the presynaptic cell type is also instrumental in setting the functional and structural parameters of the synapses. For example, parallel and climbing fiber synapses on cerebellar Purkinje cells have widely different initial release probabilities and short-term plasticity21. Similarly, the release properties of CCK and parvalbumin expressing hippocampal basket cells are also different37. Thus, if one examines a synapse population in which the pre- and postsynaptic cellular elements are heterogeneous (e.g. synapses in the neocortex), it is not surprising to see large structural and functional diversity. Keeping this in mind, we restricted our analysis to CA3 pyramidal cell local axon terminals that formed synapses only onto other CA3 pyramidal cell spines, and still observed an extensive functional and structural heterogeneity. This large diversity of synapses seems to be a unique feature of pyramidal cells, because when other synapse populations are examined, which have homogeneous pre- and postsynaptic cellular elements (e.g. parvalbumin positive basket cells onto dentate gyrus granule cells, globular bushy cells of the ventral cochlear nucleus onto principal cells of the medial nucleus of the trapezoid body, i.e: calyx of Held, etc.), a much smaller heterogeneity is found in the initial release probability, short-term plasticity and size of the active zones. This is probably a reflection of the more plastic properties of pyramidal cell – pyramidal cell connections and their widely different functional states at any given time in the life of an animal. However, it cannot be excluded that a similar degree of heterogeneity will be found in synapses that show much less plasticity. Irrespective of this, here we present quantitative data regarding the distribution of many morphological and functional parameters of CA3 pyramidal cell – pyramidal cell synapses, providing essential information for constructing realistic neuronal network models and forming the foundation for comparisons of different physiological and pathological network states.

Revealing the coupling between the presynaptic Ca2+ channels and the Ca2+ sensor of the release machinery is fundamental to our understanding of synaptic communication between nerve cells. The complexity of determining the exact relationship of the Ca2+ – release coupling has been widely appreciated26,38. It requires quantitative knowledge about the number, density and distribution of presynaptic Ca2+ channels at the active zone; the location and the kinetic properties of the Ca2+ sensors; the concentration and binding kinetics of fixed and mobile Ca2+ buffers and their diffusion coefficients. Recent experiments, using exogenously applied fast and slow Ca2+ buffers revealed two fundamentally different modes of Ca2+ – release coupling. At hippocampal parvalbumin positive basket cell synapses39, retinal ribbon synapses40, cerebellar basket cell synapses41 and in adult calyx of Held synapses42, the slow Ca2+ buffer EGTA at low mM concentrations cannot interfere with neurotransmitter release, indicating that the distance between the Ca2+ entry site and the sensor is short, probably in the range of tens of nanometers (nanodomain coupling). In addition, in these synapses, the number of Ca2+ channels mediating the release of a single vesicle is low (between 1 and 5; ref. 38). However, at young calyx of Held43, cortical CCK-positive basket cell37 and cortical pyramidal cell synapses19,44, EGTA at concentrations of a few mM strongly reduces postsynaptic responses, indicating a larger distance (from fifty to few hundred nm) between the Ca2+ channels and the sensors (microdomain coupling). This regime also predicts a much larger number of Ca2+ channels contributing to the Ca2+ concentration that drive synchronous release. Although these experiments provide functional data regarding the number of release-associated Ca2+ channels and the distance between the Ca2+ channels and the sensors, they could not reveal where exactly the Ca2+ channels are located on the surface of a bouton, or what their densities are within and outside the active zone.

Long-standing questions in this field are whether Ca2+ channels are confined to the presynaptic active zone and whether they have a random, or non-random (e.g. uniform, clustered) distribution within the active zone. Many investigators attempted to address this issue using many different techniques and preparations in the past three decades. Llinas et al.,45 recorded Ca2+ microdomains in the squid giant axon in localized spots assumed to be active zones. Similar highly localized Ca2+ microdomains were detected in cultured neuromuscular junctions by DiGregorio et al.46 and by Frank et al.47 in inner hair cells. However, none of these papers provided direct evidence that these domains corresponded to active zones. When similar fluorescent [Ca2+] measurements were performed in small CNS boutons, no spatial inhomogeneities could be detected due to the small size of the boutons, the rapid equilibration of Ca2+ within the terminal and the relatively slow binding rate of Ca2+ to the employed Ca2+ sensitive fluorescent reporters18,25,30. In summary, current Ca2+ imaging experiments could not provide information regarding the distribution of Ca2+ channels within small glutamatergic or GABAergic boutons.

A more appropriate method to reveal the precise distribution of Ca2+ channels within axon terminals is immunolocalization or fluorescent tagging and examining the reactions with light- or electron microscopy. In Drosophila neuromuscular junctions, RIM-binding protein (DRBP) plays a key role in the molecular integrity of the active zone and in the active zone localization of the Ca2+ channels48. Super-resolution light microscopic experiments provided compelling evidence for the active zone location of Ca2+ channels, but how their number or density relate to the size of the active zone remained unknown.

In the present manuscript, we applied high-resolution electron microscopic immunogold localization of the Cav2.1 subunit and found their exclusive active zone location in CA3 pyramidal cell local axon terminals. Recent immunolocalization experiments using a pre-embedding electron microscopic immunogold technique demonstrated the enrichment of Cav2.1 in the active zones of cerebellar parallel fiber49 and hippocampal fast spiking basket cells terminals39. Across the active zones of CA3 pyramidal cells, the density of Cav2.1 channels was uniform, similar to that of the active zone-related proteins Rim1/2. The uniform Rim1/2 and Cav2.1 densities predict that larger active zones will have proportionally more molecular resources for supporting release. Notably, in large active zones where many gold particles were found, they did not show a random distribution within the active zone. This nonrandom Ca2+ channel distribution might indicate that the sites of vesicle docking and release cannot occur anywhere within the active zone, but only in molecularly-defined spaces. As described above, the coupling between Ca2+ channels and the sensor is not very tight in cortical pyramidal cell axon terminals44. Although the exact distance between them has not been identified with detailed modeling, a minimum distance of ~100 nm is estimated38. Building on this assumption, we calculated that every Ca2+ channel within an active zone area of 0.01 μm2 will contribute to the release of a vesicle. In our immunogold reactions, we detected ~400 gold particles in each μm2 of active zone membrane, demonstrating a minimum of 4 Cav2.1 channels within an active zone area of 0.01 μm2. This calculation assumes a 100% labeling efficiency, which is possible to achieve with SDS-FRL50. However, we have no information regarding our labeling efficiency, and therefore only a lower limit on the number of Ca2+ channels can be provided. Our pharmacological experiments demonstrated that only ~30% of the Ca2+ enters the terminals through Cav2.1 channels. If the Cav2.2 and Cav2.3 subunit-containing channels are also confined to the active zone, and if they have a similar single channels conductance and open probability, we could estimate a minimum of 14 presynaptic Ca2+ channels contributing to the Ca2+ microdomain seen by the Ca2+ sensor of a vesicle in CA3 pyramidal cell axon terminals.


Slice preparation and electrophysiological recordings

Acute 300 μm thick horizontal ventral hippocampal slices were cut from male Wistar rats (14–17 days old). Animals were killed by decapitation, in accordance with the Hungarian Act of Animal Care and Experimentation (1998, XXVIII, section 243 / 1998), and with the ethical guidelines of the Institute of Experimental Medicine Protection of Research Subjects Committee. The brain was quickly removed and placed into an ice-cold cutting solution containing (in mM): sucrose, 205.2; KCl, 2.5; NaHCO3, 26; CaCl2, 0.5; MgCl2, 5; NaH2PO4, 1.25; glucose, 10; saturated with 95 % O2 and 5 % CO2. Hippocampal slices were prepared using a Leica vibratome (Leica VT1200S; Leica Microsystems, Vienna, Austria), incubated in an interface-type holding chamber51 in ACSF containing in mM: NaCl, 126; KCl, 2.5; NaHCO3, 26; CaCl2, 2; MgCl2, 2; NaH2PO4, 1.25; glucose, 10 saturated with 95 % O2 and 5 % CO2 (pH = 7.2 – 7.4) at 36°C that was then gradually cooled down to room temperature (~1 h). Recordings were carried out in the same ACSF, slices were kept up to 6 hours in the holding and recording chambers. Extracellularly evoked spine Ca2+ responses were measured in ACSF supplemented with 10 μM D-serine, 5 μM NBQX and 20 μM SR59931.

Cells were visualized using a Femto2D microscope equipped with oblique illumination and a water immersion lens (25X, 1.05 NA, Olympus, Japan). Current-clamp whole-cell recordings from the soma of hippocampal CA3 pyramidal cells were performed using a MultiClamp 700A amplifier (Molecular Devices), traces were filtered at 3 kHz and digitized on-line at 20 kHz. Series resistance was < 20 MΩ. Patch pipettes were pulled (Zeitz Universal Puller; Zeitz-Instrumente Vertriebs, Munich, Germany) from thick-walled borosilicate glass capillaries with an inner filament (1.5 mm outer diameter, 0.86 mm inner diameter; Sutter Instruments, Novato, CA). Pipette resistance was 3–6 MΩ when filled with the intracellular solution containing in mM: K-gluconate, 110; KCl, 5; creatine phosphate, 10; HEPES, 10; ATP, 2; GTP, 0.4; biocytin, 5 and 100 – 300 μM Fluo5F (Molecular Probes, KD ~ 1.1 μM) and 20 μM Alexa594 (Molecular Probes; pH = 7.3; 290–300 mOsm). For postsynaptic spine [Ca2+] measurements, 5 mM QX-314 [N-(2,6-dimethylphenylcarbamoylmethyl)-triethylammonium bromide] was added into the intracellular solution.

Pyramidal cells were held at −70 mV (with 0 to −100 pA current injection) and single action potentials (APs) at 0.016 Hz were evoked with 2–3 ms-long depolarizing current pulses (1–1.2 nA). Peak amplitude and full width at half maximal amplitude of the APs were monitored throughout the recordings and cells were rejected if any of these parameters changed more than 10%. ω-Conotoxin GVIA (ω-CTX GVIA 1 μM, Tocris; Product #1085), ω-agatoxin IVA (ω-ATX, 1 μM, Alomone Labs #RTA-500 and Tocris; #2799) and SNX 482 (500 nM, Tocris; #2945) were applied using a recirculation system with a peristaltic pump. The total volume was ~8 ml, and the solution was equilibrated with 95% O2 and 5% CO2. Bovine serum albumin was added at a concentration of 1 mg/ml. For extracellularly evoked responses a unipolar stimulating electrode was placed in the str. oriens at ~40 μm distance from a selected basal dendritic region of a CA3 pyramidal cell that was recorded in whole-cell mode for > 40 min and depolarized to −30 mV to relieve the Mg2+ block of NMDA receptors. Pairs of stimuli (30–100 μA) at 10 Hz were applied and were repeated in every minute. All recordings were carried out at 24°C. All drugs were purchased from Sigma unless indicated otherwise.

Two photon imaging

Experiments were performed with a Femto2D (Femtonics Ltd., Budapest, Hungary) laser scanning microscope52 with a MaiTai femtosecond pulsing laser (MaiTai, SpectraPhysics, Mountain View, CA) tuned to 810 nm. Both reflected and transmitted (through an oil-immersion condenser, Olympus, NA = 1.4) fluorescent light were collected. Electrophysiological data and image acquisition was controlled by custom made software written in MATLAB (MES, Femtonics Ltd.).

Cells were filled with two fluorescent dyes: a Ca2+-insensitive fluorophore (Alexa594, 20 μM), and a Ca2+-sensitive fluorophore (Fluo5F, 100–300 μM). Red fluorescence was used to identify spines and boutons. Changes in fluorescence were quantified during the recording as:


Fgreen(t) is the green fluorescence signal as a function of time, Frest,green is the green fluorescence before stimulation, and Idark,red is the dark current in the red channel. G/R was separately measured in saturating [Ca2+] (GMAX/R) for each dye combination and batch of intracellular solution by imaging a sealed pipette filled with a mixture of equal volumes of intracellular solution and 1M CaCl2 at the same position where the boutons were imaged. ΔG/R measurements from spines and boutons were divided by GMAX/R, yielding the reported values of G/GMAX.

Optical quantal analysis was performed as follows: responses were evoked in spines on basal dendrites with extracellular stimulation, while 6 – 12 spines were scanned simultaneously (for multiple ROI scanning see ref. 52). Responses were sorted into groups of failures and successes by their peak amplitudes. The threshold was set as 2 × standard deviation (SD) of the unfiltered baseline. In most cases this 2 × SDbaseline resulted in a clear separation of events as shown in Fig. 3. However in a few cases the distribution of the peak amplitudes was insufficient to unequivocally distinguish between successes and failures. This was the consequence of [Ca2+] transients with small amplitudes (but larger than 2 × SD baseline) and slow rise time (Supplementary Fig. 1). These transients could be separated from successes on the basis of their slow rise and delayed onset using the amplitude of the first time derivative of the fluorescent trace and the slope of a linear fit to the rise of the transients (Supplementary Fig. 1f, g). Such slow rising events were not included into the successes.

To allow the comparison of [Ca2+] transients measured in different experiments, we first determined the time of dye equilibration in axon terminals (second and third order branches) within 300 μm from the soma at 40–80 μm in the tissue. Within the first two hours of whole-cell recordings we detected a consistent decrease in the amplitude (64 ± 17% of control) and an increase in the decay time constant (132 ± 13% of control; supplementary Fig. 2) of the [Ca2+] transients, alterations consistent with an increase in dye concentration (300 μM Fluo5F in the pipette)53. First we imaged the boutons in line-scan mode (duration of 1 scan: 500 ms) with a laser intensity of 5–6 mW at the back aperture of the objective lens every 2 minutes and repeated the measurements at least 10 times. Analysis of the transients revealed that their amplitudes were stable during the first 6 repetitions (Supplementary Fig. 3). We then performed [Ca2+] measurements on a new set of boutons with only 4–6 repetitions per bouton. These measurements resulted in stable amplitudes and decay times throughout the imaging sessions (Supplementary Fig. 3e and f) and revealed a large variability between boutons in the peak [Ca2+] transients (mean ± SD = 0.13 ± 0.06 G/GMAX; n = 16). Following imaging, the slices were fixed and the biocytin was processed for light- and electron microscopic examinations. To our surprise, despite the fact that there was no sign of photodamage of the imaged boutons, electron microscopic images from 56% of the boutons showed large empty intracellular spaces (Supplementary Fig. 3g), ultrastructural abnormalities never observed following normal perfusion fixation or when looking at non-imaged, biocytin-filled boutons. We then slightly increased the laser intensity (8–10 mW at the back aperture of the objective) to obtain a better signal-to-noise ratio on single fluorescent traces (5-6 mW: S/N = 2.1 ± 0.8 vs. 8-10 mW: S/N = 5.6 ± 3.0), but repeated the line scans only 2-4 times for averaging (Fig. 5). With this protocol we measured [Ca2+] transients in a large number of boutons (n = 84 in 4 animals) without any apparent change in the amplitude or decay of the transients during the experiments and without any change in the baseline fluorescent intensity from the beginning to the end of the imaging sessions. Following fixation, visualization of biocytin and identification of individual imaged boutons we serially sectioned them and found that only 12 % had ultrastructural signs of damage (Supplementary Fig. 3h). These boutons were discarded from our analysis. The data presented in the MS in figure 5 comes from such measurements from boutons showing no signs of ultrastructural photodamage.

Light and electron microscopic identification of the imaged boutons and spines

After recordings, slices were fixed (4% paraformaldehyde, 0.5% glutaraldehyde, 0.2% picric acid in 0.1 M phosphate buffer (PB; pH = 7.4) at 4°C for at least 24 hours), embedded in agarose and re-sectioned at 60 μm thickness. Biocytin was visualized using an avidin-biotin-horseradish peroxidase complex (Vector Laboratories) followed by a reaction with 3′3-diaminobenzidine tetrahydrochloride as chromogen and 0.01% H2O2 as oxidant for 12 minutes. Sections were then postfixed in 1% OsO4 for 20 minutes, stained in 1% uranyl acetate for 25 minutes, dehydrated in a graded series of ethanol and embedded in epoxy resin (Durcupan). Cells were reconstructed using the Neurolucida system (Micro-BrightField Europe, Magdeburg, Germany) attached to a Zeiss Axioscope2 microscope using 40x or 100x oil-immersion objectives. Imaged axon/dendrite segments were then re-embedded and sectioned at 60 nm. Slices (for optimal ultrastructural preservation) from a 16 day-old male Wistar rat were fixed in 4% paraformaldehyde, 1% glutaraldehyde in 0.1M PB, dehydrated, embedded in Durcupan, sectioned at 60 nm or for the quantification of the number of the docked vesicles per active zone at 20 nm. Digital images from serial electron microscopic sections were taken from randomly selected volumes of the str. oriens at a magnification of 10 000 or 20 000 × with a Jeol1011 electron microscopic. Boutons and spines were 3D reconstructed and their volumes and the area of the PSD were measured using the Reconstruct software (http://synapses.clm.utexas.edu/). The area of an active zone was considered to be equal to that of the corresponding PSD1. We defined the end of an en passant bouton at the point where it’s diameter reached the diameter of the average inter-bouton axon.

SDS-digested freeze-fracture replica-labeling (SDS-FRL)

Adult and P16 Wistar rats, Cav2.1+/− and Cav2.1−/− mice49 (kindly provided by Dr Noboru Suzuki) were deeply anesthetized and were transcardially perfused with ice-cold fixative containing 2% paraformaldehyde in 0.1 M PB for 15 minutes. 80 μm thick sections from the hippocampal CA3 area were cut, cryoprotected in 30% glycerol, frozen with a high-pressure freezing machine (HPM100, Leica Microsystems, Austria) and fractured in a freeze-fracture machine (EM BAF060, Leica) as described in Lorincz and Nusser54 and Kulik et al.55. The fractured faces were coated on a rotating table by carbon (5 nm) with an electron beam gun positioned at 90°, then shadowed by platinum (2 nm) at 60° unidirectionally, followed by a final carbon coating (20 nm). Tissue debris was digested from the replicas with gentle stirring in a TBS solution containing 2.5% SDS and 20% sucrose (pH = 8.3) at 80°C for 18 hours. The replicas were washed in TBS containing 0.05% bovine serum albumin (BSA) and blocked with 1% or 5% BSA in TBS for one hour followed by incubation in the solution of primary antibodies. The following primary antibodies were used: guinea pig anti-Cav2.1 (1:80-1:100) raised against an epitope (amino acids 360–400; ref. 56) corresponding to a sequence in the intracellular loop between domains I and II, rabbit anti-Cav2.1 (1:500; Synaptic Systems, Gottingen, Germany) recognizing a C-terminal domain epitope (amino acids 1921–2212); rabbit anti-Rim1/2 (1:800 or 1:2000; Synaptic Systems, Goettingen, Germany); mouse anti-SNAP-25 (1:1000; Synaptic Systems); mouse anti-synaptobrevin (1:1000; Millipore, Millerica, MA); rabbit anti-syntaxin-1B (1:1000; Synaptic Systems); and rabbit pan-AMPAR (1:2000; Synaptic Systems). Replicas were then incubated in a solution containing the following secondary antibodies: goat anti-rabbit IgGs (GAR) coupled to 10 or 15 nm gold particles (1:50; British Biocell International, Cardiff, UK), goat anti-mouse IgGs coupled to 10 or 15 nm gold particles for (1:50; British Biocell) and goat anti-guinea pig IgGs coupled to 10 nm gold particles for (1:50; British Biocell). Replicas were rinsed in TBS and distilled water, before they were picked up on copper parallel bar grids and examined with a Jeol1011 electron microscope. All antibodies used in this study, except for the pan-AMPAR antibody, recognized intracellular epitopes on their target proteins and consequently were visualized by gold particles on the protoplasmic membrane half (P-face). Background labeling was determined on extracellular membrane halves (E-faces), where no specific signal was expected. The pan-AMPAR antibody recognized an extracellular epitope, therefore gold particles labeling the AMPARs were detected on the E-face. For quantitative analysis, electron micrographs of presynaptic axon terminals were taken containing active zones broken in their completeness, recognized by loose clusters of IMPs. We used either the rabbit anti-Cav2.1 antibody or a mixture of the rabbit and guinea pig anti-Cav2.1 antibodies to determine the density of gold particles labeling the Cav2.1 subunit in active zones. Gold particle densities obtained with the rabbit or with the mixture of the rabbit and guinea pig antibodies were not significantly different (t-test, p > 0.05), therefore they were pooled. For generating random ‘gold’ distribution patterns in active zones, we digitized the active zones and covered with a fine mesh (8–10 nm). The position of a simulated gold particle within the active zones were generated using the random number generator of MS Excel and was repeated for the same number of times as the number of real gold particles in any given synapse. All inter-‘gold’ distances were measured in MS Excel and the distribution was compared to that of the experimental data using the Statistica software (StatSoft Inc., Tulsa, OK).

Data analysis and statistical tests

Analysis of electrophysiological and two-photon imaging data was performed using a specialized software (MES, Femtonics). Data were plotted using OriginPro. Correlations between morphological and two-photon imaging data were assessed using linear regression fit to the data (R indicates Pearson’s correlation coefficient) when the data was not significantly different from a normal distribution (Kolmogorov-Smirnov test). For the linear regressions in figures 2 and and4,4, where the data set was too small for testing normality, Spearman correlation was calculated and the correlation coefficients are shown. Statistical significance on the effects of the calcium channel blockers on the peak [Ca2+] transient was assessed with Student’s t-test following Kolmogorov-Smirnov normality test. Immunogold densities for Cav2.1 and Rim1/2 were compared in synaptic, extrasynaptic compartments and on E-face membranes (background) using the nonparametric Kruskal-Wallis test. Kolmogorov-Smirnov test was used for comparing the original within synapse inter-gold distance distributions with random distributions. Results were considered significant with P < 0.05. Data are presented as mean ± sd. throughout the manuscript.

Supplementary Material


NH and AL are the recipients of Janos Bolyai Scholarships of the Hungarian Academy of Sciences. ZN is supported by Wellcome Trust Equipment and Project Grants (090197/Z/09/Z), a European Research Council Advanced Grant, and a Hungarian National Office for Research and Technology-French National Research Agency TéT Fund (NKTH-Neurogen). AK is supported by a Deutsche Forschungsgemeinschaft (DFG SFB 780) grant. We would like to thank Dr Noboru Suzuki for kindly providing the Cav2.1−/− mice and Dr Andreas Unger for helping with the Cav2.1−/− replica labeling. We would like to thank Eva Dobai and Dora Ronaszéki for their excellent technical assistance, Dr Máté Sümegi for his help with the modeling, Synaptic Systems GmbH for providing the rabbit anti-Cav2.1 antibody, Prof. Erwin Neher and Dr Mark Eyre for their comments on the manuscript, and Dr Helmuth Koester for helpful discussions.


Author’s contributions: NH performed all in vitro Ca2+ measurements and post hoc light- and electron microscopic analysis of the imaged structures. AL performed SDS-FRL for SNAP-25, Cav2.1 and Rim1/2, quantitatively analyzed the reactions, and performed the simulations. GK developed the image acquisition and analysis software. BR designed and built the two-photon microscope. AK performed SDS-FRL reactions for Cav2.1 in Cav2.1+/− and Cav2.1−/− mice. MW developed the guinea pig anti-Cav2.1 antibody. NH and ZN designed the experiments and wrote the manuscript.

Competing financial interest: BR and GK are the owners of Femtonics Ltd.


1. Schikorski T, Stevens CF. Quantitative ultrastructural analysis of hippocampal excitatory synapses. J. Neurosci. 1997;17:5858–5867. [PubMed]
2. Shepherd GM, Harris KM. Three-dimensional structure and composition of CA3-CA1 axons in rat hippocampal slices: implications for presynaptic connectivity and compartmentalization. J. Neurosci. 1998;18:8300–8310. [PubMed]
3. Atwood HL, Karunanithi S. Diversification of synaptic strength: presynaptic elements. Nat. Rev. Neurosci. 2002;3:497–516. [PubMed]
4. Takumi Y, Ramirez-Leon V, Laake P, Rinvik E, Ottersen OP. Different modes of expression of AMPA and NMDA receptors in hippocampal synapses. Nat. Neurosci. 1999;2:618–624. [PubMed]
5. Racca C, Stephenson FA, Streit P, Roberts JD, Somogyi P. NMDA receptor content of synapses in stratum radiatum of the hippocampal CA1 area. J. Neurosci. 2000;20:2512–2522. [PubMed]
6. Nusser Z, et al. Cell type and pathway dependence of synaptic AMPA receptor number and variability in the hippocampus. Neuron. 1998;21:545–559. [PubMed]
7. Nimchinsky EA, Yasuda R, Oertner TG, Svoboda K. The number of glutamate receptors opened by synaptic stimulation in single hippocampal spines. J. Neurosci. 2004;24:2054–2064. [PubMed]
8. Matsuzaki M, et al. Dendritic spine geometry is critical for AMPA receptor expression in hippocampal CA1 pyramidal neurons. Nat. Neurosci. 2001;4:1086–1092. [PMC free article] [PubMed]
9. Harris KM, Stevens JK. Dendritic spines of CA1 pyramidal cells in the rat hippocampus: serial electron microscopy with reference to their biophysical characteristics. J. Neurosci. 1989;9:2982–2997. [PubMed]
10. Fields RD, Ellisman MH. Synaptic morphology and differences in sensitivity. Science. 1985;228:197–199. [PubMed]
11. Atwood HL, Marin L. Ultrastructure of synapses with different transmitter-releasing characteristics on motor axon terminals of a crab, Hyas areneas. Cell Tissue Res. 1983;231:103–115. [PubMed]
12. Propst JW, Ko CP. Correlations between active zone ultrastructure and synaptic function studied with freeze-fracture of physiologically identified neuromuscular junctions. J. Neurosci. 1987;7:3654–3664. [PubMed]
13. Korn H, Mallet A, Triller A, Faber DS. Transmission at a central inhibitory synapse. II. quantal description of release, with a physical correlate for binomial n. J. Neurophysiol. 1982;48:679–707. [PubMed]
14. Murthy VN, Schikorski T, Stevens CF, Zhu Y. Inactivity produces increases in neurotransmitter release and synapse size. Neuron. 2001;32:673–682. [PubMed]
15. Branco T, Marra V, Staras K. Examining size-strength relationships at hippocampal synapses using an ultrastructural measurement of synaptic release probability. J. Struct. Biol. 2010;172:203–210. [PMC free article] [PubMed]
16. Reyes A, et al. Target-cell-specific facilitation and depression in neocortical circuits. Nat. Neurosci. 1998;1:279–285. [PubMed]
17. Thomson AM. Activity-dependent properties of synaptic transmission at two classes of connections made by rat neocortical pyramidal axons in vitro. J. Physiol. (Lond.) 1997;502:131–147. [PMC free article] [PubMed]
18. Koester HJ, Johnston D. Target cell-dependent normalization of transmitter release at neocortical synapses. Science. 2005;308:863–866. [PubMed]
19. Rozov A, Burnashev N, Sakmann B, Neher E. Transmitter release modulation by intracellular Ca2+ buffers in facilitating and depressing nerve terminals of pyramidal cells in layer 2/3 of the rat neocortex indicates a target cell-specific difference in presynaptic calcium dynamics. J. Physiol. (Lond.) 2001;531:807–826. [PMC free article] [PubMed]
20. Losonczy A, Zhang L, Shigemoto R, Somogyi P, Nusser Z. Cell type dependence and variability in the short-term plasticity of EPSCs in identified mouse hippocampal interneurones. J. Physiol. (Lond.) 2002;542:193–210. [PMC free article] [PubMed]
21. Xu-Friedman MA, Harris KM, Regehr WG. Three-dimensional comparison of ultrastructural characteristics at depressing and facilitating synapses onto cerebellar Purkinje cells. J. Neurosci. 2001;21:6666–6672. [PubMed]
22. Oertner TG, Sabatini BL, Nimchinsky EA, Svoboda K. Facilitation at single synapses probed with optical quantal analysis. Nat. Neurosci. 2002;5:657–964. [PubMed]
23. Schneggenburger R, Neher E. Intracellular calcium dependence of transmitter release rates at a fast central synapse. Nature. 2000;406:889–893. [PubMed]
24. Bollmann JH, Sakmann B, Borst JG. Calcium sensitivity of glutamate release in a calyx-type terminal. Science. 2000;289:953–957. [PubMed]
25. Bucurenciu I, Bischofberger J, Jonas P. A small number of open Ca2+ channels trigger transmitter release at a central GABAergic synapse. Nat. Neurosci. 2009;13:19–21. [PubMed]
26. Neher E. Vesicle pools and Ca2+ microdomains: new tools for understanding their roles in neurotransmitter release. Neuron. 1998;20:389–399. [PubMed]
27. Katz B, Miledi R. The role of calcium in neuromuscular facilitation. J. Physiol. (Lond.) 1968;195:481–492. [PMC free article] [PubMed]
28. Llinas R, Steinberg IZ, Walton K. Relationship between presynaptic calcium current and postsynaptic potential in squid giant synapse. Biophys. J. 1981;33:323–351. [PMC free article] [PubMed]
29. del Castillo J, Katz B. Quantal components of the end-plate potential. J. Physiol. (Lond.) 1954;124:560–573. [PMC free article] [PubMed]
30. Koester HJ, Sakmann B. Calcium dynamics associated with action potentials in single nerve terminals of pyramidal cells in layer 2/3 of the young rat neocortex. J. Physiol. (Lond.) 2000;529:625–646. [PMC free article] [PubMed]
31. Wu LG, Saggau P. Pharmacological identification of two types of presynaptic voltage-dependent calcium channels at CA3-CA1 synapses of the hippocampus. J. Neurosci. 1994;14:5613–5622. [PubMed]
32. Wheeler DB, Randall A, Tsien RW. Roles of N-type and Q-type Ca2+ channels in supporting hippocampal synaptic transmission. Science. 1994;264:107–111. [PubMed]
33. Reid CA, Clements JA, Bekkers JM. Nonuniform distribution of Ca2+ channel subtypes on presynaptic terminals of excitatory synapses in hippocampal cultures. J. Neurosci. 1997;17:2738–2745. [PubMed]
34. Masugi-Tokita M, Shigemoto R. High-resolution quantitative visualization of glutamate and GABA receptors at central synapses. Curr. Opin. Neurobiol. 2007;17:387–393. [PubMed]
35. Hagiwara A, Fukazawa Y, Deguchi-Tawarada M, Ohtsuka T, Shigemoto R. Differential distribution of release-related proteins in the hippocampal CA3 area as revealed by freeze-fracture replica labeling. J. Comp. Neurol. 2005;489:195–216. [PubMed]
36. Lisman J. Long-term potentiation: outstanding questions and attempted synthesis. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 2003;358:829–842. [PMC free article] [PubMed]
37. Hefft S, Jonas P. Asynchronous GABA release generates long-lasting inhibition at a hippocampal interneuron-principal neuron synapse. Nat. Neurosci. 2005;8:1319–1328. [PubMed]
38. Eggermann E, Bucurenciu I, Goswami SP, Jonas P. Nanodomain coupling between Ca2+ channels and sensors of exocytosis at fast mammalian synapses. Nat. Rev. Neurosci. 2012;13:7–21. [PMC free article] [PubMed]
39. Bucurenciu I, Kulik A, Schwaller B, Frotscher M, Jonas P. Nanodomain coupling between Ca2+ channels and Ca2+ sensors promotes fast and efficient transmitter release at a cortical GABAergic synapse. Neuron. 2008;57:536–545. [PubMed]
40. Mennerick S, Matthews G. Ultrafast exocytosis elicited by calcium current in synaptic terminals of retinal bipolar neurons. Neuron. 1996;17:1241–1249. [PubMed]
41. Christie JM, Chiu DN, Jahr CE. Ca2+-dependent enhancement of release by subthreshold somatic depolarization. Nat. Neurosci. 2011;14:62–68. [PMC free article] [PubMed]
42. Fedchyshyn MJ, Wang LY. Developmental transformation of the release modality at the calyx of Held synapse. J. Neurosci. 2005;25:4131–4140. [PubMed]
43. Meinrenken CJ, Borst JG, Sakmann B. Calcium secretion coupling at calyx of held governed by nonuniform channel-vesicle topography. J. Neurosci. 2002;22:1648–1667. [PubMed]
44. Ohana O, Sakmann B. Transmitter release modulation in nerve terminals of rat neocortical pyramidal cells by intracellular calcium buffers. J. Physiol. (Lond.) 1998;513:135–148. [PMC free article] [PubMed]
45. Llinas R, Sugimori M, Silver RB. Microdomains of high calcium concentration in a presynaptic terminal. Science. 1992;256:677–679. [PubMed]
46. DiGregorio DA, Peskoff A, Vergara JL. Measurement of action potential-induced presynaptic calcium domains at a cultured neuromuscular junction. J. Neurosci. 1999;19:7846–7859. [PubMed]
47. Frank T, Khimich D, Neef A, Moser T. Mechanisms contributing to synaptic Ca2+ signals and their heterogeneity in hair cells. Proc. Natl. Acad. Sci. USA. 2009;106:4483–4488. [PMC free article] [PubMed]
48. Liu KS, et al. RIM-binding protein, a central part of the active zone, is essential for neurotransmitter release. Science. 2011;334:1565–1569. [PubMed]
49. Kulik A, et al. Immunocytochemical localization of the a1A subunit of the P/Q-type calcium channel in the rat cerebellum. Eur. J. Neurosci. 2004;19:2169–2178. [PubMed]
50. Tanaka J, et al. Number and density of AMPA receptors in single synapses in immature cerebellum. J. Neurosci. 2005;25:799–807. [PubMed]
51. Gulyas AI, et al. Parvalbumin-containing fast-spiking basket cells generate the field potential oscillations induced by cholinergic receptor activation in the hippocampus. J. Neurosci. 2010;30:15134–15145. [PMC free article] [PubMed]
52. Lorincz A, Rozsa B, Katona G, Vizi ES, Tamas G. Differential distribution of NCX1 contributes to spine-dendrite compartmentalization in CA1 pyramidal cells. Proc. Natl. Acad. Sci. USA. 2007;104:1033–1038. [PMC free article] [PubMed]
53. Maravall M, Mainen ZF, Sabatini BL, Svoboda K. Estimating intracellular calcium concentrations and buffering without wavelength ratioing. Biophys. J. 2000;78:2655–2667. [PMC free article] [PubMed]
54. Lorincz A, Nusser Z. Molecular identity of dendritic voltage-gated sodium channels. Science. 2010;238:906–909. [PMC free article] [PubMed]
55. Kulik A, et al. Compartment-dependent colocalization of Kir3.2-containing K+ channels and GABAB receptors in hippocampal pyramidal cells. J. Neurosci. 2006;26:4289–4297. [PubMed]
56. Miyazaki T, et al. Cav2.1 in cerebellar Purkinje cells regulates competitive excitatory synaptic wiring, cell survival, and cerebellar biochemical compartmentalization. J. Neurosci. 2012;32:1311–1328. [PubMed]
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...