The nervous system sends information around the body in the form of electrical signals that travel through cells called neurons. However, these electrical signals cannot cross the synapses between neurons. Instead, the information is carried across the synapse by molecules called neurotransmitters.

Calcium ions control the release of neurotransmitters. There is a high concentration of calcium ions outside the neuron but they are not able to pass through the cell membrane under normal conditions. However, when an electrical impulse reaches the synapse, ion channels in the membrane open and allow calcium ions to enter the cell. Once inside, the ions activate the release of neurotransmitters by binding to proteins called release sensors.

Several experiments on the release of neurotransmitters have studied the synapses between neurons and muscle fibers. These studies found that the higher the concentration of ions outside the neuron, the higher the rate at which the neurotransmitters were released. However, the timing of release—the length of time over which the neurotransmitters were released—did not depend on the concentration of calcium ions.

Arai and Jonas have now studied neurotransmitter release at a synapse in a region of the brain called the cerebellum. These experiments also found that the timing of the release did not depend on the ion concentration, suggesting that this may be a general property of neurotransmitter release.

To find out more, Arai and Jonas created a mathematical model of neurotransmitter release. This model suggests that for the timing of release to remain the same, the ion channel and the release sensor must be located close together in the presynaptic terminal. If they are not close together, the timing of release becomes blurred and more dependent on the external calcium concentration.

Further experiments confirm the prediction of the model by showing that the calcium channels and the release sensors in these synapses are very close together. The next challenge will be to find out whether the conclusions are also valid for other synapses where the calcium channels and release sensors are further apart.

Calcium plays a key role in the control of transmitter release at chemical synapses (Neher, 1998; Südhof, 2013). As transmitter release is a Ca²⁺-dependent biochemical process, one would expect that both the extent and kinetics of release will depend on Ca²⁺ concentration. In contrast to this expectation, the time course of release (TCR) at the neuromuscular junction is independent of the extracellular Ca²⁺ concentration ([Ca²⁺]_o) (Datyner and Gage, 1980; Van der Kloot, 1988; Parnas et al., 1989). However, the apparent [Ca²⁺]_o independence of the TCR is less well established at central synapses (Sargent et al., 2005). Furthermore, the underlying mechanisms remain elusive. Several potential explanations were proposed, including adaptation of transmitter release (Hsu et al., 1996) and additional voltage sensors that control the timing of release (Parnas et al., 2000; See; Felmy et al., 2003). Recent results suggested that in several central synapses presynaptic Ca²⁺ channels and release sensors are tightly coupled (Bucurenciu et al., 2008; Eggermann et al., 2012; Scimemi and Diamond, 2012; Schmidt et al., 2013). Tight coupling might be another factor contributing to the apparent [Ca²⁺]_o independence of the TCR (Yamada and Zucker, 1992). However, this possibility has not been directly examined.

We studied the [Ca²⁺]_o dependence of transmitter release at GABAergic synapses between cerebellar basket cells (BCs) and Purkinje cells (PCs), using paired whole-cell recordings from synaptically connected neurons in mouse brain slices (Caillard et al., 2000; Sakaba, 2008; Eggermann and Jonas, 2012; Figure 1A–D). Action potentials were evoked in the presynaptic BC in whole-cell current-clamp or cell-attached voltage-clamp configurations, whereas IPSCs were recorded in the postsynaptic PC under whole-cell voltage-clamp conditions (series resistance 3.8 ± 0.1 MΩ; mean ± SEM; 92 cells; ‘Materials and methods’). Unitary IPSCs were initiated with short latency and high temporal precision. Synaptic latency was 1.20 ± 0.03 ms at ∼22°C (24 pairs) and 0.47 ± 0.02 ms at ∼34°C (5 pairs; Figure 1C,D). Synaptic transmission was entirely blocked by bath application of the selective P/Q-type Ca²⁺ channel blocker ω-agatoxin IVa (1 µM), whereas the N-type Ca²⁺ channel blocker ω-conotoxin GVIa (1 µM) had no detectable effect (Figure 1E–H). Thus, transmitter release at this synapse is selectively mediated by P/Q-type Ca²⁺ channels.

Figure 1

Download asset Open asset

Exploiting the technical advantages of this synapse (including ideal voltage-clamp conditions conveyed by perisomatic synapse location, presynaptic accessibility due to short axon trajectories, and optimal signal-to-noise ratio because of large quantal size), we first examined the [Ca²⁺]_o dependence of release efficacy. Analysis of the dependence of evoked IPSC peak amplitude on [Ca²⁺]_o revealed that the amount of transmitter release was steeply [Ca²⁺]_o-dependent, with a high power coefficient (n = 3.02 in the low-concentration limit; Figure 2A,B). Thus, release at the BC–PC synapse was highly cooperative (Dodge and Rahamimoff, 1967). We next measured the [Ca²⁺]_o dependence of release kinetics. To determine the TCR, we first recorded unitary IPSCs at a given [Ca²⁺]_o, and subsequently in reduced [Ca²⁺]_o to isolate quantal IPSCs (Figure 2C). We then computed the TCR by deconvolution of unitary and quantal IPSC waveforms (Diamond and Jahr, 1995; Neher and Sakaba, 2001; Sakaba, 2008; Figure 2D). When [Ca²⁺]_o was changed in the range from 0.7–4 mM, peak release rate changed by a factor of 34.4. In contrast, the half-duration of the TCR was only minimally affected (0.49 ± 0.05 ms at 0.7 mM; 0.43 ± 0.05 ms at 1 mM; 0.47 ± 0.01 ms at 2 mM; 0.47 ± 0.02 ms at 4 mM; 5–11 pairs; p = 0.56; Figure 2D,E). Furthermore, the decay time constant of the TCR was similar in the different conditions (p = 0.38; Figure 2D,E). These results show that the TCR is largely [Ca²⁺]_o-independent at a central synapse. Similar conclusions were reached at lower recording temperature, which would be expected to markedly slow down the kinetic rates of channels and sensors, but to only minimally change the rates of diffusional processes. At ∼12°C, changes in [Ca²⁺]_o had no detectable effects on the half-duration of the TCR (2.84 ± 0.42 ms at 1 mM, 2.97 ± 0.24 ms at 2 mM, and 2.79 ± 0.25 ms at 4 mM [Ca²⁺]_o; 5 pairs; p = 0.93) (Figure 2—figure supplement 1).

Figure 2 with 2 supplements see all

Download asset Open asset

To explore the mechanisms underlying the paradoxical [Ca²⁺]_o independence of the TCR, we used a realistic model of action potential-dependent transmitter release (Eggermann and Jonas, 2012; Figure 3). Ca²⁺ diffusion and buffering were computed by solving the full set of partial differential diffusion and reaction equations, and the TCR was simulated using a previously established release sensor model (Lou et al., 2005; Figure 3A,B). Finally, the half-duration of the TCR was plotted against the peak release rate (PRR). Computational analysis revealed that the TCR was markedly dependent on the PRR in both low and high PRR limit. In the low PRR limit, the half-duration of the TCR decreased with PRR in a model with fixed coupling distance (Figure 3C). This dependence was accentuated after slowing of release sensor rates (Figure 3D, top), indicating that rate-limiting sensor kinetics are responsible. Furthermore, this dependence was inverted in a model with variable coupling distance (Figure 3E). In such a configuration, small Ca²⁺ inflow may selectively release proximal vesicles, whereas large Ca²⁺ inflow may recruit all vesicles, resulting in a broadening of the TCR. In the high PRR limit, the half-duration of the TCR increased with PRR (Figure 3C). This behavior was particularly prominent in small boutons (Figure 3F, bottom), suggesting that saturation of the endogenous buffers leads to broadening of the TCR. In all configurations, the dependence of the TCR on the PRR was substantially more prominent in loose (200 nm, blue colors) than in tight coupling configurations (20 nm, red colors). Similar conclusions were reached when the concentration and unbinding rate of fixed buffers were altered (Figure 3—figure supplement 1; Gilmanov et al., 2008). Thus, whereas several factors influence the shape of the TCR–PRR relation, the coupling distance plays a key role in all conditions examined.

Figure 3 with 1 supplement see all

Download asset Open asset

Our modeling results suggest the intriguing possibility that tight coupling between Ca²⁺ channels and release sensors (Christie et al., 2011) might be a main reason for the apparent [Ca²⁺]_o independence of the TCR at the cerebellar BC–PC synapse. To test this hypothesis quantitatively, we probed the distance between Ca²⁺ channels and release sensors, using exogenous Ca²⁺ chelators (Adler et al., 1991; Neher, 1998; Eggermann et al., 2012; Figure 4). Ca²⁺ chelators were loaded into presynaptic BC terminals through the somatic patch pipette, and the reversibility of their effects was tested by obtaining a second recording from the same presynaptic BC at later times. Whereas the fast Ca²⁺ chelator ethylenedioxybis-(o-phenylenenitrilo)-N,N,N',N'-tetraacetic acid (BAPTA) markedly suppressed transmitter release (Figure 4A,B), the slow Ca²⁺ chelator ethyleneglycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid (EGTA) was much less effective (Figure 4C,D). Analysis of concentration–effect data revealed that the half-maximal inhibitory concentration (IC₅₀) for the steady-state effect was 0.6 mM for BAPTA (14 pairs total) and 16.0 mM for EGTA (11 pairs total; Figure 4E). Thus, the IC₅₀ value was ∼27-fold higher for EGTA than for BAPTA.

Figure 4

Download asset Open asset

To estimate the coupling distance, we fit the concentration–effect data for BAPTA and EGTA with a model of Ca²⁺ diffusion and buffering based on linear approximations (Neher, 1998; Figure 4F). The main free parameter in the model was the distance between Ca²⁺ channels and release sensors, while several other parameters (e.g. the physicochemical properties of the Ca²⁺ chelators and the cooperativity of transmitter release) were well constrained. Analysis of the entire data set revealed a coupling distance of 10–20 nm (11.4 nm for constant coupling distance; 13.9 nm for half-normally distributed coupling distance; 10.1 nm for skewed-normally distributed coupling distance; Figure 4F). Statistical errors, as assessed by bootstrap analysis, were minimal (1.2 nm; Figure 4G). Furthermore, systematic errors were small, as indicated by the robustness of the estimate against variations in resting Ca²⁺ concentration and endogenous buffer product (Figure 4H). These results reveal that the coupling between Ca²⁺ channels and release sensors at the cerebellar BC–PC synapse is tight.

Our results address a long-standing question in the field of synaptic transmission (Yamada and Zucker, 1992; Parnas et al., 2000): If release probability is highly dependent on extracellular Ca²⁺ concentration (Dodge and Rahamimoff, 1967), how is it possible that the timing of release is [Ca²⁺]_o-independent (Datyner and Gage, 1980; Van der Kloot, 1988; Parnas et al., 1989)?

To address this question, we measured the [Ca²⁺]_o dependence of both the amount and the time course of transmitter release at the cerebellar BC–PC synapse, a central synapse in which the TCR can be precisely quantified. Our results demonstrate that while the amount of release is highly [Ca²⁺]_o-dependent with a power coefficient of ∼3, the TCR is largely [Ca²⁺]_o-independent. Although the apparent [Ca²⁺]_o independence of the TCR has been well established at the neuromuscular junction (Datyner and Gage, 1980; Van der Kloot, 1988; Parnas et al., 1989), this phenomenon is less well documented at central synapses (see Sargent et al., 2005 for a notable exception). Thus, our results demonstrate that the [Ca²⁺]_o independence of the TCR is a general phenomenon characteristic for synaptic transmission in both peripheral and central nervous system.

We further measured the coupling distance between Ca²⁺ channels and release sensors at BC–PC synapses. Our results show that coupling is tight, with a coupling distance of 10–20 nm. Although several previous studies measured the coupling distance, the rules that define the coupling configuration remain elusive (Hefft and Jonas, 2005; Bucurenciu et al., 2008; Christie et al., 2011; Nadkarni et al., 2012; Scimemi and Diamond, 2012; Schmidt et al., 2013; Vyleta and Jonas, 2014). It has been suggested that tight coupling is primarily used at synapses designed for fast, reliable transmission, whereas loose coupling is utilized at synapses specialized on presynaptic plasticity (Bucurenciu et al., 2008; Eggermann et al., 2012; Nadkarni et al., 2012; Vyleta and Jonas, 2014). The present results are fully consistent with this hypothesis, since GABAergic BC–PC synapses show fast transmitter release following single action potentials and reliable transmission during trains of spikes (Caillard et al., 2000; Sakaba, 2008; Eggermann and Jonas, 2012).

The present results identify novel links between the [Ca²⁺]_o independence of the TCR and nanodomain coupling. Our model reveals that the [Ca²⁺]_o dependence is substantially more prominent in loose than in tight coupling regimes in a variety of conditions. These include different Ca²⁺ sensor rates, different bouton diameters, various concentrations and affinity values of endogenous buffers, and uniform vs non-uniform coupling. Specifically, our model predicts that slowing of sensor rates should markedly enhance the [Ca²⁺]_o dependence of the TCR in a microdomain, but not in a nanodomain coupling regime. We tested this prediction by lowering the temperature, which, among other potential effects, is expected to slow sensor rates. Whereas lowering the temperature had large effects on the absolute value of the half-duration of the TCR, its [Ca²⁺]_o independence was unchanged. These findings provide experimental evidence that tight source–sensor coupling is a key factor that ensures the [Ca²⁺]_o independence of the TCR.

Previous studies highlighted several functional advantages of nanodomain coupling, including efficacy, speed, temporal precision, and energy efficiency of synaptic transmission (Bucurenciu et al., 2008; Eggermann et al., 2012). Our results suggest another functional benefit: conveying [Ca²⁺]_o independence to the TCR. Is the invariance of the TCR relevant for microcircuit function under physiological conditions? Maintenance of speed and temporal precision at the GABAergic BC–PC synapse is of critical importance for the operation of the cerebellum, since feedforward inhibition mediated by BCs sets the temporal window of signal integration in PCs (Mittmann et al., 2005; Bao et al., 2010). However, [Ca²⁺]_o has been shown to fluctuate during repetitive activity, high-frequency network oscillations, and in pathophysiological conditions (Heinemann et al., 1977; Borst and Sakmann, 1999; Rusakov and Fine, 2003). Furthermore, Ca²⁺ inflow will be changed by neuromodulators, which often act via inhibition of presynaptic Ca²⁺ channels (Takahashi et al., 1998). Thus, nanodomain coupling at BC–PC synapses may ensure constant timing of fast feedforward inhibition under a variety of network conditions.

1. 1. Adler EM
   2. Augustine GJ
   3. Duffy SN
   4. Charlton MP

   (1991)

   Alien intracellular calcium chelators attenuate neurotransmitter release at the squid giant synapse

   The Journal of Neuroscience 11:1496–1507.

   • Google Scholar
2. 1. Bao J
   2. Reim K
   3. Sakaba T

   (2010) Target-dependent feedforward inhibition mediated by short-term synaptic plasticity in the cerebellum

   The Journal of Neuroscience 30:8171–8179.

   https://doi.org/10.1523/JNEUROSCI.0276-10.2010

   • Google Scholar
3. 1. Borst JGG
   2. Sakmann B

   (1998) Calcium current during a single action potential in a large presynaptic terminal of the rat brainstem

   The Journal of Physiology 506:143–157.

   https://doi.org/10.1111/j.1469-7793.1998.143bx.x

   • Google Scholar
4. 1. Borst JGG
   2. Sakmann B

   (1999) Depletion of calcium in the synaptic cleft of a calyx-type synapse in the rat brainstem

   The Journal of Physiology 521:123–133.

   https://doi.org/10.1111/j.1469-7793.1999.00123.x

   • Google Scholar
5. 1. Bucurenciu I
   2. Kulik A
   3. Schwaller B
   4. Frotscher M
   5. Jonas P

   (2008) Nanodomain coupling between Ca²⁺ channels and Ca²⁺ sensors promotes fast and efficient transmitter release at a cortical GABAergic synapse

   Neuron 57:536–545.

   https://doi.org/10.1016/j.neuron.2007.12.026

   • Google Scholar
6. 1. Caillard O
   2. Moreno H
   3. Schwaller B
   4. Llano I
   5. Celio MR
   6. Marty A

   (2000) Role of the calcium-binding protein parvalbumin in short-term synaptic plasticity

   Proceedings of the National Academy of Sciences of USA 97:13372–13377.

   https://doi.org/10.1073/pnas.230362997

   • Google Scholar
7. 1. Christie JM
   2. Chiu DN
   3. Jahr CE

   (2011) Ca²⁺-dependent enhancement of release by subthreshold somatic depolarization

   Nature Neuroscience 14:62–68.

   https://doi.org/10.1038/nn.2718

   • Google Scholar
8. 1. Collin T
   2. Chat M
   3. Lucas MG
   4. Moreno H
   5. Racay P
   6. Schwaller B
   7. Marty A
   8. Llano I

   (2005) Developmental changes in parvalbumin regulate presynaptic Ca²⁺ signaling

   The Journal of Neuroscience 25:96–107.

   https://doi.org/10.1523/JNEUROSCI.3748-04.2005

   • Google Scholar
9. Book

   1. Colquhoun D
   2. Sigworth FJ

   (1995)

   Fitting and statistical analysis of single-channel records

   In: Sakmann B, Neher E, editors. Single-channel recording. New York: Plenum Press. pp. 483–587.

   • Google Scholar
10. 1. Datyner NB
    2. Gage PW

    (1980)

    Phasic secretion of acetylcholine at a mammalian neuromuscular junction

    The Journal of Physiology 303:299–314.

    • Google Scholar
11. 1. Diamond JS
    2. Jahr CE

    (1995) Asynchronous release of synaptic vesicles determines the time course of the AMPA receptor-mediated EPSC

    Neuron 15:1097–1107.

    https://doi.org/10.1016/0896-6273(95)90098-5

    • Google Scholar
12. 1. Dodge FA Jr
    2. Rahamimoff R

    (1967)

    Co-operative action a calcium ions in transmitter release at the neuromuscular junction

    The Journal of Physiology 193:419–432.

    • Google Scholar
13. Book

    1. Efron B
    2. Tibshirani RJ

    (1998)

    An Introduction to the Bootstrap

    London: Chapman and Hall/CRC.

    • Google Scholar
14. 1. Eggermann E
    2. Bucurenciu I
    3. Goswami SP
    4. Jonas P

    (2012) Nanodomain coupling between Ca²⁺ channels and sensors of exocytosis at fast mammalian synapses

    Nature Reviews Neuroscience 13:7–21.

    https://doi.org/10.1038/nrn3125

    • Google Scholar
15. 1. Eggermann E
    2. Jonas P

    (2012) How the ‘slow’ Ca²⁺ buffer parvalbumin affects transmitter release in nanodomain coupling regimes

    Nature Neuroscience 15:20–22.

    https://doi.org/10.1038/nn.3002

    • Google Scholar
16. 1. Felmy F
    2. Neher E
    3. Schneggenburger R

    (2003) The timing of phasic transmitter release is Ca²⁺-dependent and lacks a direct influence of presynaptic membrane potential

    Proceedings of the National Academy of Sciences of USA 100:15200–15205.

    https://doi.org/10.1073/pnas.2433276100

    • Google Scholar
17. 1. Gilmanov IR
    2. Samigullin DV
    3. Vyskocil F
    4. Nikolsky EE
    5. Bukharaeva EA

    (2008) Modeling of quantal neurotransmitter release kinetics in the presence of fixed and mobile calcium buffers

    Journal of Computational Neuroscience 25:296–307.

    https://doi.org/10.1007/s10827-008-0079-5

    • Google Scholar
18. 1. Hefft S
    2. Jonas P

    (2005) Asynchronous GABA release generates long-lasting inhibition at a hippocampal interneuron-principal neuron synapse

    Nature Neuroscience 8:1319–1328.

    https://doi.org/10.1038/nn1542

    • Google Scholar
19. 1. Heinemann U
    2. Lux HD
    3. Gutnick MJ

    (1977)

    Extracellular free calcium and potassium during paroxysmal activity in the cerebral cortex of the cat

    Experimental Brain Research 27:237–243.

    • Google Scholar
20. 1. Hsu SF
    2. Augustine GJ
    3. Jackson MB

    (1996) Adaptation of Ca²⁺-triggered exocytosis in presynaptic terminals

    Neuron 17:501–512.

    https://doi.org/10.1016/S0896-6273(00)80182-8

    • Google Scholar
21. 1. Jones MV
    2. Jonas P
    3. Sahara Y
    4. Westbrook GL

    (2001) Microscopic kinetics and energetics distinguish GABA_A receptor agonists from antagonists

    Biophysical Journal 81:2660–2670.

    https://doi.org/10.1016/S0006-3495(01)75909-7

    • Google Scholar
22. 1. Li L
    2. Bischofberger J
    3. Jonas P

    (2007) Differential gating and recruitment of P/Q-, N-, and R-type Ca²⁺ channels in hippocampal mossy fiber boutons

    The Journal of Neuroscience 27:13420–13429.

    https://doi.org/10.1523/JNEUROSCI.1709-07.2007

    • Google Scholar
23. 1. Lou X
    2. Scheuss V
    3. Schneggenburger R

    (2005) Allosteric modulation of the presynaptic Ca²⁺ sensor for vesicle fusion

    Nature 435:497–501.

    https://doi.org/10.1038/nature03568

    • Google Scholar
24. 1. Meinrenken CJ
    2. Borst JGG
    3. Sakmann B

    (2002)

    Calcium secretion coupling at calyx of Held governed by nonuniform channel-vesicle topography

    The Journal of Neuroscience 22:1648–1667.

    • Google Scholar
25. 1. Mittmann W
    2. Koch U
    3. Häusser M

    (2005) Feed-forward inhibition shapes the spike output of cerebellar Purkinje cells

    The Journal of Physiology 563:369–378.

    https://doi.org/10.1113/jphysiol.2004.075028

    • Google Scholar
26. 1. Nadkarni S
    2. Bartol TM
    3. Stevens CF
    4. Sejnowski TJ
    5. Levine H

    (2012) Short-term plasticity constrains spatial organization of a hippocampal presynaptic terminal

    Proceedings of the National Academy of Sciences of USA 109:14657–14662.

    https://doi.org/10.1073/pnas.1211971109

    • Google Scholar
27. 1. Neher E

    (1998) Usefulness and limitations of linear approximations to the understanding of Ca⁺⁺ signals

    Cell Calcium 24:345–357.

    https://doi.org/10.1016/S0143-4160(98)90058-6

    • Google Scholar
28. 1. Neher E
    2. Sakaba T

    (2001)

    Combining deconvolution and noise analysis for the estimation of transmitter release rates at the calyx of Held

    The Journal of Neuroscience 21:444–461.

    • Google Scholar
29. 1. Parnas H
    2. Hovav G
    3. Parnas I

    (1989) Effect of Ca²⁺ diffusion on the time course of neurotransmitter release

    Biophysical Journal 55:859–874.

    https://doi.org/10.1016/S0006-3495(89)82885-1

    • Google Scholar
30. 1. Parnas H
    2. Segel L
    3. Dudel J
    4. Parnas I

    (2000) Autoreceptors, membrane potential and the regulation of transmitter release

    Trends in Neurosciences 23:60–68.

    https://doi.org/10.1016/S0166-2236(99)01498-8

    • Google Scholar
31. 1. Perkins KL

    (2006) Cell-attached voltage-clamp and current-clamp recording and stimulation techniques in brain slices

    Journal of Neuroscience Methods 154:1–18.

    https://doi.org/10.1016/j.jneumeth.2006.02.010

    • Google Scholar
32. 1. Rusakov DA
    2. Fine A

    (2003) Extracellular Ca²⁺ depletion contributes to fast activity-dependent modulation of synaptic transmission in the brain

    Neuron 37:287–297.

    https://doi.org/10.1016/S0896-6273(03)00025-4

    • Google Scholar
33. 1. Sakaba T

    (2008) Two Ca²⁺-dependent steps controlling synaptic vesicle fusion and replenishment at the cerebellar basket cell terminal

    Neuron 57:406–419.

    https://doi.org/10.1016/j.neuron.2007.11.029

    • Google Scholar
34. 1. Sargent PB
    2. Saviane C
    3. Nielsen TA
    4. DiGregorio DA
    5. Silver RA

    (2005) Rapid vesicular release, quantal variability, and spillover contribute to the precision and reliability of transmission at a glomerular synapse

    The Journal of Neuroscience 25:8173–8187.

    https://doi.org/10.1523/JNEUROSCI.2051-05.2005

    • Google Scholar
35. 1. Schmidt H
    2. Brachtendorf S
    3. Arendt O
    4. Hallermann S
    5. Ishiyama S
    6. Bornschein G
    7. Gall D
    8. Schiffmann SN
    9. Heckmann M
    10. Eilers J

    (2013) Nanodomain coupling at an excitatory cortical synapse

    Current Biology 23:244–249.

    https://doi.org/10.1016/j.cub.2012.12.007

    • Google Scholar
36. 1. Scimemi A
    2. Diamond JS

    (2012) The number and organization of Ca²⁺ channels in the active zone shapes neurotransmitter release from Schaffer collateral synapses

    The Journal of Neuroscience 32:18157–18176.

    https://doi.org/10.1523/JNEUROSCI.3827-12.2012

    • Google Scholar
37. Book

    1. Smith GD

    (2001)

    Modelling local and global calcium signals using reaction-diffusion equations

    In: De Schutter E, editors. Computational Neuroscience. Boca Raton, Florida: CRC Press. pp. 49–85.

    • Google Scholar
38. 1. Südhof TC

    (2013) Neurotransmitter release: the last millisecond in the life of a synaptic vesicle

    Neuron 80:675–690.

    https://doi.org/10.1016/j.neuron.2013.10.022

    • Google Scholar
39. 1. Takahashi T
    2. Kajikawa Y
    3. Tsujimoto T

    (1998)

    G-protein-coupled modulation of presynaptic calcium currents and transmitter release by a GABA_B receptor

    The Journal of Neuroscience 18:3138–3146.

    • Google Scholar
40. 1. Van der Kloot W

    (1988)

    The kinetics of quantal releases during end-plate currents at the frog neuromuscular junction

    The Journal of Physiology 402:605–626.

    • Google Scholar
41. 1. Vyleta NP
    2. Jonas P

    (2014) Loose coupling between Ca²⁺ channels and release sensors at a plastic hippocampal synapse

    Science 343:665–670.

    https://doi.org/10.1126/science.1244811

    • Google Scholar
42. 1. Yamada WM
    2. Zucker RS

    (1992) Time course of transmitter release calculated from simulations of a calcium diffusion model

    Biophysical Journal 61:671–682.

    https://doi.org/10.1016/S0006-3495(92)81872-6

    • Google Scholar

Thank you for sending your work entitled “Nanodomain coupling explains Ca²⁺ independence of transmitter release time course at a fast central synapse” for consideration at eLife. Your article has been favorably evaluated by Eve Marder (Senior editor), Michael Häusser (Reviewing editor), and 2 reviewers.

The Reviewing editor and the other reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.

This manuscript examines an interesting and under-studied question, namely whether synaptic delays vary with external calcium concentrations and if not, why. To address this, the authors have performed paired recordings at basket cell Purkinje cell synapses, in combination with computational modelling. They include a study of latencies at two temperatures, a pharmacological analysis of synaptic calcium channel subtypes, a deconvolution-based analysis of release rate as a function of external calcium, and an investigation of the effects of added EGTA and BAPTA on synaptic efficacy. The results are then simulated using a variant of the previously published Eggermann-Jonas model at the same synapse.

The experimental results, which are of high quality, clearly show that there is tight coupling between SVs and VDCCs and also, that the release rate does not vary significantly with external calcium at physiological temperature. While intuition suggests that a causal link between these two findings is likely, the simulation of Figure 3 brings disappointingly little to support this hypothesis, or to eliminate alternative explanations. As detailed below, the possibility remains open that other aspects of the synapse function such as width of the AP, speed of the release reactions, or effects of endogenous buffers, may play a key role as well. Repeating the experiments of Figure 2d at room temperature should help to solve this problem. Therefore, we believe these experiments are essential for the resubmission, should not be very time-consuming, but will be very informative.

Further detail is provided below:

1) The figure describing the simulations (Figure 3) is difficult to understand in detail. The main result is in panel c. But there seems little justification in using the slope ratio at 200 s-1 to evaluate the effect of coupling distance on calcium sensitivity. It would be better to use the slope calculated by joining together the points representing 1, 2 and 4 mM calcium (respectively near 200, 2000 and 4000 s-1). If this is done it appears that the changes in TCR duration are only near 15 or 20% for the loose coupling (roughly decreasing from 0.65 to 0.55). A similar analysis made on panel e (including a dispersion in coupling distances) leads to even smaller changes. Looking at these simulations, it does seem like the dependence on external calcium is less for tight coupling than for loose coupling. But in either case the dependence is modest, presumably because the model has been constrained to mimic a very fast synapse (fast AP, fast calcium channel, fast SV release). So at the end it is not clear whether the weak dependence of TCR on external calcium should be ascribed primarily to the tight coupling as proposed by the authors, or to other features of the model.

2) Additional room temperature experiments would significantly strengthen the manuscript and should be carried out prior to resubmission. The results of Figure 1d show a much broader distribution of latencies at 22^°C than at 34^°C. Likewise the previous deconvolution analysis by Sakaba at the same synapse (Neuron 57, 406, 2008), which was performed at room temperature, indicates a broader TCR than in Figure 2 here. This suggests an opportunity to test whether it is really the tight coupling that prevents latency changes with extracellular calcium, or whether other factors play a role as well. If in spite of the broader TCR, external calcium effects remain negligible at room temperature, then the case of tight coupling as a causal link will be much strengthened. If however significant changes occur at room temperature, then additional synaptic parameters will have to be considered as well.

3) The SV release rates appear to be per synapse in Figure 2, and per vesicle in Figure 3, making the correspondence between the two figures difficult. The “default” rate in Figure 3 is 2000 s-1 presumably corresponding to the 15 quanta /ms of Figure 2 at 2mM calcium. How the 2000 s-1 rate was obtained is unclear. It may have been obtained by extrapolating the normalized IPSC data of Figure 2b to very high calcium, and inferring that the release rate at 2mM is 1/3 of the maximum, that is a third of 6000 s-1, or 2 000 s-1. However the extrapolation cannot be considered as accurate because high calcium points are difficult to obtain at physiological temperatures, so that the asymptote is not well constrained in Figure 2b. If true, the 2 000 s-1 value would indicate an RRP value of 15/2=7.5 per synapse, much smaller than the mean value of 110 found by Sakaba (2008). Overall vesicular release rates may have been seriously overestimated in the simulation of Figure 3.

4) This paper comes in a line of other works by the senior author and others measuring the channel sensor coupling in different synapses. A common conclusion is that at GABAergic synapses the coupling is 'tight'. In this context, some of the results presented here are not particularly surprising (e.g. all of Figure 1, parts of Figure 2, and all of Figure 4). It would be helpful if the authors could emphasize more strongly which of the results they consider to be particularly novel and important, especially in light of the recent literature (see next comment).

5) The discussion of the literature needs to be expanded. While it is true that latency variations have been little investigated in the CNS, the literature is not as empty as the manuscript would suggest. For example, Boudkkazi et al (Neuron 56, 1048, 2007) found that delays vary with release probability at cortical synapses, including when increasing the extracellular calcium concentration. Gilmanov IR et al 2008 also approached these issues at NMJ with similar conclusions. Furthermore, the Ca independence of release has long been noted at NMJ (as the authors acknowledge), but the authors don't refer to any studies of this in neurons, though there are several which could be cited and discussed, starting perhaps with Silver et al. 1996 in granule cells.

1) The figure describing the simulations (Figure 3) is difficult to understand in detail. The main result is in panel c. But there seems little justification in using the slope ratio at 200 s-1 to evaluate the effect of coupling distance on calcium sensitivity. It would be better to use the slope calculated by joining together the points representing 1, 2 and 4 mM calcium (respectively near 200, 2000 and 4000 s-1). If this is done it appears that the changes in TCR duration are only near 15 or 20% for the loose coupling (roughly decreasing from 0.65 to 0.55). A similar analysis made on panel e (including a dispersion in coupling distances) leads to even smaller changes. Looking at these simulations, it does seem like the dependence on external calcium is less for tight coupling than for loose coupling. But in either case the dependence is modest, presumably because the model has been constrained to mimic a very fast synapse (fast AP, fast calcium channel, fast SV release). So at the end it is not clear whether the weak dependence of TCR on external calcium should be ascribed primarily to the tight coupling as proposed by the authors, or to other features of the model.

We have used the concentration–effect curve in Figure 2b to estimate the release rates for our experimental conditions (0.7, 1, 2, and 4 mM [Ca²⁺]_o; we have included additional experiments to determine curve parameters more precisely, see point 3 below). We have estimated that the peak release rates at 0.7 and 1 mM [Ca²⁺]_o are 168 and 379 s^-1, respectively. We have therefore decided to maintain our original reference point (200 s^-1), which is exactly in-between. Finally, we have indicated our experimental release rates with red arrows in the graphs of Figure 3c. These data corroborate that if there were marked dependencies of TCR on release rate, as e.g. the case in a loose coupling regime, we would be able to detect this dependence in our experiments. We thank the editor / reviewers for raising this point.

2) Additional room temperature experiments would significantly strengthen the manuscript and should be carried out prior to resubmission. The results of Figure 1d show a much broader distribution of latencies at 22^°C than at 34^°C. Likewise the previous deconvolution analysis by Sakaba at the same synapse (Neuron 57, 406, 2008), which was performed at room temperature, indicates a broader TCR than in Figure 2 here. This suggests an opportunity to test whether it is really the tight coupling that prevents latency changes with extracellular calcium, or whether other factors play a role as well. If in spite of the broader TCR, external calcium effects remain negligible at room temperature, then the case of tight coupling as a causal link will be much strengthened. If however significant changes occur at room temperature, then additional synaptic parameters will have to be considered as well.

We thank the editors / reviewers for this excellent suggestion. We would like to point out that the results shown in the original submission (with the exception of Figure 1c, d) had already been performed at room temperature. We have revised the Materials and methods section and the Figure legends to make this point more clear. We are aware that the TCR in our experiments (0.47 ± 0.01 ms) is shorter than the approximate value reported by Sakaba, 2008 (∼1 ms) under seemingly identical conditions. However, because of the high Q₁₀ value of the TCR (see below), this may have been caused by relatively minor differences in temperature.

To constructively pick up the key idea of the editors / reviewers, we performed additional experiments at lower temperature (∼12°C; 5 pairs). Under these experimental conditions, kinetic parameters (presynaptic action potential duration, presynaptic Ca²⁺ channel gating, and release sensor kinetics) are expected to be slowed down, whereas tight coupling is likely to be maintained. These experiments turned out to be surprisingly difficult, because of lower stimulation frequencies required to ensure pool replenishment and several additional experimental issues related to low temperature. Interestingly, reduction in temperature markedly prolonged the TCR, whereas the dependence of TCR on [Ca²⁺]_o remained minimal. These results (incorporated into Figure 2–figure supplement 1) corroborate our original conclusion that the tight coupling is the most important parameter in shaping the Ca²⁺-independence of transmitter release. Again, we thank the editor / reviewers for this suggestion.

3) The SV release rates appear to be per synapse in Figure 2, and per vesicle in Figure 3, making the correspondence between the two figures difficult. The “default” rate in Figure 3 is 2000 s-1 presumably corresponding to the 15 quanta /ms of Figure 2 at 2mM calcium. How the 2000 s-1 rate was obtained is unclear. It may have been obtained by extrapolating the normalized IPSC data of Figure 2b to very high calcium, and inferring that the release rate at 2mM is 1/3 of the maximum, that is a third of 6000 s-1, or 2 000 s-1. However the extrapolation cannot be considered as accurate because high calcium points are difficult to obtain at physiological temperatures, so that the asymptote is not well constrained in Figure 2b. If true, the 2 000 s-1 value would indicate an RRP value of 15/2=7.5 per synapse, much smaller than the mean value of 110 found by Sakaba (2008). Overall vesicular release rates may have been seriously overestimated in the simulation of Figure 3.

We have performed additional experiments to better constrain the concentration–effect curve in the saturation phase, as requested by the reviewer (5 pairs at 10 mM [Ca²⁺]_o). The results are now included in Figure 2b, and better constrain our estimates of release probability. The new results confirm that the “physiological” release probability at 2 mM [Ca²⁺]_o is 0.26, and that the corresponding peak release rate is close to 2000 s^-1. Given that we extended the range of [Ca²⁺]_o, we further decided to determine the Hill coefficient in the concentration range ≤ 2 mM rather than the entire concentration range. We have included a double-logarithmic representation to graphically illustrate this point (Figure 2b, bottom of the revised version). Please note that this different (presumably more accurate) measurement of the Hill coefficient also slightly changed the results of our analysis of coupling distance (in the main text and legend of Figure 4).

We also thank the reviewers for reminding us of the paper by Sakaba (2008). We have cited this paper in the revised version. We are aware that our mean peak release rate and the relatively high release probability may suggest that the size of the releasable pool is small. Our simple quantitative estimation is pool size = quantal content (from deconvolution and integral of TCR) / release probability (from Ca²⁺ dependence of release) = 10.4 / 0.26 = 40 at 2 mM [Ca²⁺]_o. In apparent contrast, the cumulative release analysis during trains of action potentials and Ca²⁺ uncaging suggested a larger pool size (110 ± 28; Sakaba, 2008). One possible explanation might be that the size of the pool accessed by single action potentials is smaller than that accessed by trains or Ca²⁺ uncaging. Furthermore, there are many additional complications in the analysis of train data, including release-independent components of depression (Kraushaar and Jonas, 2000), Ca²⁺- and time-dependent changes in refilling rate (Wang and Kaczmarek, 1998, Nature), and nonlinearities in the extrapolation in cumulative release plots (Thanawala and Regehr, 2013, J. Neuroscience). Thus, whether or not there is a true discrepancy remains to be resolved.

4) This paper comes in a line of other works by the senior author and others measuring the channel sensor coupling in different synapses. A common conclusion is that at GABAergic synapses the coupling is 'tight'. In this context, some of the results presented here are not particularly surprising (e.g. all of Figure 1, parts of Figure 2, and all of Figure 4). It would be helpful if the authors could emphasize more strongly which of the results they consider to be particularly novel and important, especially in light of the recent literature (see next comment).

We have completely changed the Discussion to indicate the novel aspects more clearly. We continue to think that the manuscript provides interesting data, as also stated by the editor’s summary that the paper addresses an “interesting and under-studied question”. Furthermore, not all papers converge on the conclusion that coupling is tight. For example, we have recently shown loose coupling in hippocampal mossy fiber synapses (Vyleta and Jonas, 2014), and earlier papers suggested loose coupling at CA3–CA1 cell synapses based on the properties of synaptic dynamics (Nadkarni et al., 2012). Likewise, previous studies using membrane-permeant Ca²⁺ chelators indicated loose coupling in CCK interneuron output synapses (Hefft and Jonas, 2005). Finally, the suggestions of the editor / reviewers have helped to convey an additional novel aspect to the paper, because, to the best of our knowledge, nobody ever studied the [Ca²⁺]_o dependence of the TCR at different temperatures.

5) The discussion of the literature needs to be expanded. While it is true that latency variations have been little investigated in the CNS, the literature is not as empty as the manuscript would suggest. For example, Boudkkazi et al (Neuron 56, 1048, 2007) found that delays vary with release probability at cortical synapses, including when increasing the extracellular calcium concentration. Gilmanov IR et al 2008 also approached these issues at NMJ with similar conclusions. Furthermore, the Ca independence of release has long been noted at NMJ (as the authors acknowledge), but the authors don't refer to any studies of this in neurons, though there are several which could be cited and discussed, starting perhaps with Silver et al. 1996 in granule cells.

We thank the reviewers for reminding us of the papers of Gilmanov et al., 2008, Silver et al., 1996 (and Sargent et al., 2005). Although the TCR is not directly reported in those studies, we agree that these papers are relevant to our work, because they report Ca²⁺-dependence of the EPSC time course. We have incorporated Gilmanov et al., 2008 and Sargent et al., 2005, as suggested. However, we would like to add that latency and time course of release should not be confused. For example, Boudkkazi et al. studied latency variation rather than TCR (thus, those results will not be directly relevant to the present paper).

Author details

1. Itaru Arai

   IST Austria (Institute of Science and Technology Austria), Klosterneuburg, Austria

   Contribution

   IA, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

2. Peter Jonas

   IST Austria (Institute of Science and Technology Austria), Klosterneuburg, Austria

   Contribution

   PJ, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   peter.jonas@ist.ac.at

   Competing interests

   The authors declare that no competing interests exist.

• 2,230

  Page views

• 301

  Downloads

• 23

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, Scopus, PubMed Central.