Measurement of velocity of synaptic transmission (at the squid giant synapse)

6
Experiment 1: dual recordings of action potential in presynaptic
terminal and AP or EPSP in postsynaptic element  delay
between electrical signals
Experiment 2: recording of Ca current evoked by an action
potential in the presynaptic terminal (action clamp exp)

Synaptic transmission is
very rapid: delay between pre-
and postsynaptic electrical
signals is 0.5 ms.

Considering the time necessary

to open presynaptic CaV
channels, (cf activation of Ca
current during the repolarization of
the AP), exocytosis occurs in
less than 200 μsec after Ca
influx into the terminal.
A giant mammalian synapse: the Calyx of Held giant synapse Synaptic terminal:
15 um x 0.8 um

5 μm
About 500 active zones ( 0.1 um diameter;
2±2 primed anchored and primed vesicles
per active zone)

Localized in the brainstem auditory

nucleus MNTB (medial nucleus of the
trapezoid body).

Meinrenken et al. (2003) J. Physiol.

547 665
 Measurement of velocity of synaptic transmission
(at the Calyx of Held giant mammalian synapse)

Synaptic delay:
0.5 ms at 24 °C;
0.13 ms at 35°C

Relaese rate
was obtained
from model

Meinrenken et al. (2003) J. Physiol. 547 665

The very high velocity of synaptic transmission requires a complex
exocytotic machinery that makes the vesicles competent (primed) for the
rapid opening of the fusion pore.

Sudhof (2013) Neuron

 The very high velocity of synaptic transmission requires that CaV
channels are localized very close to the Ca2+ sensor
synaptotagmin on the primed vesicles

In 100 msec after opening of a CaV channel, the Ca2+

ions diffuse up to a distance of only 50 nm from the
mouth of the channel (considering the coefficient of diffusion of Ca2+
ions and the Ca buffers in the terminal)

CaV channels need to be within a few tenth of nm from a

primed vesicle
Freeze fracture
of the frog
neuromuscular
junction

The fusion of synaptic vesicles occurs within 50 nm from the CaV particles

In the absence Deformations of the presynaptic

of presynaptic membrane along the active zone
stimulation immediately after synaptic
activity, interpreted as
invaginations of the cell
membrane caused by fusion of
synaptic vesicles. These
deformations lay along one or
two rows of intramembranous
5 ms after
particles, visible along both
presynaptic
margins of the presynaptic
stimulation
density and now thought to be
In the presence of the voltage-gated Ca2+channels
4AP to prolong the
action potential and Heuser and Reese (1981) J
increase release Cell Biol 88 564
Electronic tomography of frog neuromuscular junction:
three-dimensional reconstruction of the active zone shows
ordered structure of protein lattice

Harlow et al. (2001) Nature 409 479-84

Proteins of the active zones maintain the spatial relationship
between vesicles and CaV channels

Ribs link vesicles to central beam.

The density of ribs is similar to that of
CaV particles to which they are linked
via pegs
 The very high velocity of synaptic transmission requires that CaV
channels are localized very close to the Ca2+ sensor
synaptotagmin on the primed vesicles

In 100 msec after opening of a CaV channel, the Ca2+

ions diffuse up to a distance of only 50 nm from the
mouth of the channel (considering the coefficient of diffusion of Ca2+
ions and the Ca buffers in the terminal)

CaV channels need to be within a few tenth of nm from a

primed vesicle
The effective distance depends on the affinity of the Ca
sensor and on the number of CaV channels that are close
enough to contribute to release of a single vesicle.
In some synapses the exocytosis of a single vesicle depends on Ca2+ influx
through 1-2 CaV channels localized within 10-20 nm from the vesicle; in some
synapses it depends on Ca2+ influx through 8-10 CaV channels localized within
200 nm from the vesicle.

The exocytosis of a primed vesicle (and hence the neurotransmitter

release) depends on the local Ca2+ concentration in a nanodomain or
microdomain close to the mouth of the CaV channel(s) located close to
the primed vesicles.
In some synapses the exocytosis of a single vesicle depends on Ca2+ influx
through 1-2 CaV channels localized within 10-20 nm from the vesicle; in some
synapses it depends on Ca2+ influx through 8-10 CaV channels localized within
200 nm from the vesicle.

The exocytosis of a primed vesicle (and hence the neurotransmitter

release) depends on the local Ca2+ concentration in a nanodomain or
microdomain close to the mouth of the CaV channel(s) located close to
the primed vesicles.

The local [Ca2+]in which is relevant for the exocytosis of a primed vesicle is not
measurable with the Ca2+-sensitive fluorescent dyes (eg Fura2) that are used
to measure [Ca2+]in

However, the relevant local [Ca2+]in has been indirectly measured at some
synapses using flash photolysis di caged-Ca2+ (DM-nitrophen), which produces
a measurable uniform transient increase of [Ca2+] in the terminal

- A postsynaptic potential similar to that evoked by a presynaptic action

potential is obtained with a flash that produces an increase of [Ca2+] in the
terminal in the range 20-50 uM, depending on the type of synapse.
 Release dependent on Ca2+ nanodomains Release dependent on Ca2+ microdomains
Distance between CaV and Ca2+ sensor < 100 nm Distance between CaV and Ca2+ sensor > 100 nm

Method to distinguish Ca2+ nanodomains vs microdomains-dependent release: exogenous Ca2+ buffers

with different rates of Ca2+ binding (kon) but similar affinity (Kd).
If the distance between CaV and sensor is <100 nm only the fast buffer BAPTA succeeds to catch Ca2+
ions during their journey from the channel to the Ca sensor before they bind the Ca sensor (the availalble
time window is few microseconds).
Eggermann et al (2012) Nature Reviews Neuroscience
 How many open CaV are necessary to release one vesicle?

Release controlled
by > 1 CaV

Release controlled
by 1 CaV

The number of channels can be varied either by pharmacological block (in synapses in which
release is controlled by only one type of CaV) or by changes in the duration of the action potential

Release = k [Ca2+]n
n varies from 1 to 4 (which is the value corresponding to the intrinsic cooperativity of Ca2+
binding to the Ca sensor synaptotagmin)
The number of channels can be derived from n via modelling: about 10 CaV for n=4
 Inhibitory synapses between hippocampal FS interneurons and dentate gyrus
granule neurons (P18-21)
-- release dependent on Ca2+ nanodomains.
--release exclusively controlled by CaV2.1 channels.

How many open CaV are necessary to release one vesicle?

Release = k [Ca2+]n
n varies from 1 to 4 (which is the value corresponding to the intrinsic cooperativity of Ca2+
binding to the Ca sensor synaptotagmin)
The number of channels can be derived from n via modelling: about 10 CaV for n=4
 Inhibitory synapses between hippocampal FS interneurons and
dentate gyrus granule neurons (P18-21)

How many open CaV are necessary to release one vesicle?

Release = k [Ca2+]n n =1.6

Modelling  2-3 Ca channels control the release of a vesicle

Squid giant synapse:
--release dependent on Ca2+ nanodomains
--release of single vesicle controlled by a single Ca2+ channel.

Calix of Held (P8-12):

-- release dependent on Ca2+ microdomains.
--more than one Ca2+ channel controls release of a single vesicle
(10 channel at an average distance of 200 nm from simulations)
--both CaV2.1 and CaV2.2 channels contribute to release
[cf super-addictive inhibition of EPSC (> 100%) by specific inhibitors
of the different Ca channels].
--release induced by flash-photolysis of caged Ca2+ is similar to that
induced by AP when [Ca2+]in = 20-30 uM
 Callix of Held
Release is controlled by both CaV2.1 and CaV2.2 channels.
The summed % Inhibition of the EPSC by separate application of the specific CaV2.1
inhibitor Aga and the specific CaV2.2 inhbitor Ctx is >100 % (super-addictive inhiition).

Specific inhibition of CaV2.1 channels

Specific inhibition of CaV2.2 channels

Wu et al. (1999) J Neurosci 19 726

 The properties of SNC synapses vary with development
and they are different at different synapses

Calix of Held (P8-12):

-- release dependent on Ca2+ microdomains.
--more than one Ca2+ channel controls release of a single vesicle
(10 channel at an average distance of 200 nm from simulations)
--both CaV2.1 and CaV2.2 channels contribute to release
[cf super-addictive (<1005) inhibition of EPSC by specific inhibitors of
the different Ca channels].
--release induced by flash-fotolisi of caged Ca2+ is similar to that
induced by AP when [Ca2+]in = 20-30 uM

Calix of Held (P15-18):

-- release dependent on smaller Ca2+ microdomains.
--a lower number of Ca2+ channels (located closer to Ca sensor)
controls release of a single vesicle (cf exponent = 2 rather than 4)
--release dependent exclusively on CaV2.1 channels
--release induced by flash-photolysis of caged Ca2+ is similar to that
induced by AP when [Ca2+]in = 55 uM
Cortical excitatory synapses (P14-16) :
-- release dependent on Ca2+ microdomains.
--release controlled by both CaV2.1 and CaV2.2 (in some also CaV2.3; cf super-
addictive inhibition), but with dominant role of CaV2.1.

Inhibitory synapses between hippocampal FS interneurons and dentate

gyrus granule neurons (P18-21)
-- release dependent on Ca2+ nanodomains.
--release exclusively controlled by CaV2.1 channels.
--2-3 channels control release of one vesicle
Inhibitory synapses between hippocampal FS interneurons and dentate
gyrus granule neurons (P18-21)

How many open CaV are necessary to release one vesicle?

Release = k [Ca2+]n n =1.6

Modelling  2-3 Ca channels control the release of a vesicle

 Release dependent on Ca2+ nanodomains: functional consequences
Simulations of local Ca2+ transients and of release rates for different distances between Ca channel and vesicle (20-200 nm)

Ca transient
«seen» by the Ca Duration of
sensor (left) and release period
release rate (right)
normalized to the peak
value with Ca channel
at 20 nm from Ca
Synaptic delay
sensor

Red dashed line: Blue dashed line: normalized

normalized presinaptic AP presynaptic Ca current
synchronous
asynchronous

The fast component of

the Ca transient
drives synchronous Probability
release, whereas the of release
slow component
drives asynchronous
release

Ca transients and release rates normalized to their individual peak values

Tight coupling Ca channel-vesicle increases the release probability and synaptic

efficacy (cf peak Ca transient on sensor) as well as the speed and temporal precision of
release in relation to the presynaptic AP (cf delay and duration of release), and reduces
the relative contribution of asynchronous release.

Tight coupling Ca channel-
vesicle saves energy (cf
extrusion of one Ca2+ ion
requires 1 ATP): important in
fast-spiking neurons
may trigger spontaneous
transmitter release
Simulated [Ca] 12 nm away from a
single Ca channel activated by an AP
But AgaIVa does not affect mIPSC frequency in
(red curve: infinite number of Ca
DG neurons: high activation threshold of CaV2.1
channels).
and the use of 2-3 open Ca channels rather than
Despite stochastic opening of the Ca
channel the rising phase of the Ca
one may protect from excessive spontaneous
release generated by stochastic Ca channel
transient is largely deterministic,
opening
governed by the increase in driving
force during the repolarization phase.
.