R E V I E W S
NATURE REVIEWS |NEUROSCIENCE
VOLUME 6 |JANUARY 2005 |5 7
The defining feature ofa chemical synapse is the collec-
tion of synaptic vesicles in the presynaptic terminal.
These vesicles participate in a cycle (for a review,see
REF.1) that permits them to be used repeatedly during
sustained activity.In a typical synapse at rest,a small
percentage ofvesicles are attached to the presynaptic
membrane and the rest reside in an adjoining cluster.The
synaptic vesicles all look alike under the electron micro-
scope,and no significant biochemical distinctions are
recognized that might identify different classes ofvesicles
in a resting terminal.So,apart from the relatively few
vesicles that are attached (‘docked’) to the surface mem-
brane, the vast majority seem to constitute a single,
homogeneous population.However,for decades investi-
gators have proposed that there are distinct ‘pools’ —
populations ofvesicles whose members possess distinct
There is a rather bewildering list ofnames for these
pools,including readily releasable,reserve,exo–endo
recycling,immediately releasable,reluctant and resting.
In this review,we will focus on the five preparations that
have been characterized most thoroughly with respect
to vesicle pools — Drosophilalarval neuromuscular
junction (NMJ),frog NMJ,neonatal rodent cultured
hippocampal neurons,neonatal rodent CALYX OF HELD
neurons in slice preparations and acutely isolated
goldfish retinal bipolar cells.These preparations span a
broad phylogenetic range, permit a diverse array of
research strategies and techniques,and vary greatly in
their ability to secrete quanta. Despite differences
between these preparations, we propose that it is
possible to assign every vesicle to one ofthree pools,
which we will call the readily releasable pool (RRP),the
recycling pool and the reserve pool.
Early observations of vesicle pools
The study ofvesicle pools began with the work ofBirks
and MacIntosh2,who investigated acetylcholine release
from cat sympathetic ganglia (see also REF.3).They
proposed that there are two distinct presynaptic stores of
transmitter — a ‘readily releasable’ fraction,which is
rapidly depleted at high frequencies ofstimulation,and
a ‘non-readily releasable’ fraction. Later, Elmqvist,
Quastel and colleagues4,5obtained similar results from
human intercostal muscle studies.They found that on
high-frequency stimulation the amplitude of the
responses decayed rapidly.They suggested that the early
stimuli drew quanta from a store ofneurotransmitter
from which quanta could be easily mobilized — the
‘mobilization store’. The decay of the postsynaptic
response was thought to be caused by this small store
not being replenished fast enough during intense stimu-
lation.Elmqvist and Quastel agreed that the store of
neurotransmitter corresponded to the readily releasable
fraction proposed by Birks and MacIntosh,and that
the rest of the quanta represented the non-readily
Most synapses rely on three vesicle pools
Elmqvist and Quastel cautiously noted that although
‘the experimental results are fully compatible with the
[pools] model,they cannot be said in any way to prove
it’4.Since then,other possible explanations,including
changes in intracellular calcium concentration,partial
SYNAPTIC VESICLE POOLS
Silvio O.Rizzoli* and William J.Betz‡
Abstract | Communication between cells reaches its highest degree of specialization at chemical
synapses. Some synapses talk in a ‘whisper’; others ‘shout’. The ‘louder’ the synapse, the more
synaptic vesicles are needed to maintain effective transmission, ranging from a few hundred
(whisperers) to nearly a million (shouters). These vesicles reside in different ‘pools’, which have
been given a bewildering array of names. In this review, we focus on five tissue preparations in
which synaptic vesicle pools have been identified and thoroughly characterized. We argue that,
in each preparation, each vesicle can be assigned to one of three distinct pools.
Am Faßberg 11,D-37077
and Biophysics C-240,
4200 East Ninth Avenue,
Correspondence to W.J.B.
CALYX OF HELD SYNAPSE
A giant nerve terminal in the
auditory brainstem,it has a
pivotal role in the circuitry that
is responsible for locating high-
A portion ofthe presynaptic
membrane that faces the
postsynaptic density across the
synaptic cleft.It constitutes the
site ofsynaptic vesicle clustering,
docking and transmitter release.
Cells are dialysed with a caged
calcium compound (such as
flashes are then used to break the
cage to release calcium.In this
way,the technique stimulates
exocytosis independently of
calcium entry from the
5 8 |JANUARY 2005 |VOLUME 6
R E V I E W S
stimuli,such as high-frequency stimulation and hyper-
tonic shock13,high-frequency stimulation and calcium
FLASH PHOTOLYSIS10,or high frequency stimulation and
depolarization14,have been shown to release the same
A further distinction can be made within the RRP
when release is investigated in detail,in that the RRP
vesicles might not all be identical in release character-
istics. At the calyx of Held it has been shown that
the RRP has fast and slow components of release15
(see below).Also,in hippocampal slices,Hanse and
Gustafsson16–18suggest that not all vesicles that are
released during a stimulus as short as 10 action poten-
tials (at 50 Hz) are identical in release parameters.On
average, they found that about two vesicles were
released per active zone during stimulation.However,
on average only one of the vesicles was found to be
immediately available for release16.
The recycling pool.We define the recycling pool as the
pool of vesicles that maintain release on moderate
(physiological) stimulation.This pool is thought to
contain about 5–20% of all vesicles. Physiological
frequencies ofstimulation cause it to recycle continu-
ously8,19–21,and it is refilled by newly recycled vesicles.
One exception to this definition seems to occur in the
goldfish bipolar nerve terminal,where it is difficult to
ascertain whether there are stimulation conditions that
would permit continuous recycling without reserve pool
mobilization.However,it should be stressed that most
research on this synapse has been carried out using
strong stimulation,such as continuous depolarization22
or calcium uncaging23.
The reserve pool.The reserve pool is defined as a depot
ofsynaptic vesicles from which release is only triggered
during intense stimulation.These vesicles constitute
the majority (typically ~80–90%) of vesicles in most
presynaptic terminals. The release of these vesicles
requires stimulation frequencies ofat least 5–10 Hz in
frog NMJ9,24,25, 30 Hz in Drosophilalarval NMJ26or
prolonged high potassium application at the calyx of
Held21and possibly also at hippocampal boutons27.It is
possible that these vesicles are seldom or never
recruited during physiological activity.
It is not clear what triggers reserve pool release,but
an experiment by Kuromi and Kidokoro28indicates a
possible mechanism.They used the Drosophilatemp-
erature-sensitive mutant shibire,which cannot reform
its vesicles at 34°C,but functions normally at room
temperature.Stimulation at room temperature caused
cycling of only the recycling pool. However, when
recycling ofthese vesicles was prevented at high temp-
erature, vesicle release continued from the reserve
pool.Consistent with these results,at the frog NMJ
almost no reserve pool vesicles are released until the
recycling pool is depleted8,24.Therefore,it is tempting
to suggest that depletion of recycling pool vesicles
triggers reserve pool mobilization and release,
although the underlying molecular mechanisms
release from fully filled vesicles,full release from partially
filled vesicles and desensitization and/or saturation of
postsynaptic receptors have been investigated, but
were not found to fully explain the observed changes
in release.Therefore,through a process of exclusion,
the idea ofdistinct vesicle pools that possess different
capacities for exocytosis has been strengthened.
Although the terminology varies,most models (for
example,FIG.1a) agree that presynaptic nerve terminals
contain an RRP, from which vesicles can be easily
mobilized on stimulation,and a large ‘reserve pool’,
from which vesicles are drawn more slowly,typically
in response to intense or prolonged stimulation.
A further distinction can be made, as not all ‘non-
RRP’ vesicles are equally capable of being released.
Therefore,three pools of vesicles have to be postu-
lated6.The second pool of vesicles is released more
slowly than the RRP,and its release precedes reserve
pool mobilization;for the purpose of this review we
term this pool the recycling pool.One striking exam-
ple of the release of three vesicle pools is shown in
FIG. 1b. Here, three different kinetic phases can be
observed:a rapid,almost instantaneous one,a more
prolonged slower one and a long-lasting phase.The
properties of the three pool types are summarized
in TABLE 1.
The readily releasable pool.For the purposes of this
review,we will define the RRP as the synaptic vesicles
that are immediately available on stimulation.These
vesicles are generally thought to be docked to the
presynaptic ACTIVE ZONE and primed for release,
although it should be stressed that docked vesicles are
not all necessarily immediately releasable (see below
The RRP is depleted rapidly by 5–15 shocks ofhigh-
frequency electrical stimulation4,8–10,a few milliseconds
ofdepolarization11,12or about 1 s ofhypertonic shock
in hippocampal boutons13(note that the hypertonic
shock is not a physiological stimulus).Different sets of
Change in fluorescence (%)
Figure 1 |Three vesicle pools. a | The classic three-pool model. The reserve pool makes
up ~80–90% of the total pool, and the recycling pool is significantly smaller (~10–15%).
The readily releasable pool (RRP) consists of a few vesicles (~1%) that seem to be docked
and primed for release. b | Three kinetic components of release (indicating release of three
vesicle pools) on depolarization of goldfish bipolar cells. The cell was stimulated in the
presence of the styryl dye FM 1-43, and the increase in fluorescence gives a direct
measure of exocytosis. Panel b modified, with permission, from REF.12© (1999) Blackwell
NATURE REVIEWS |NEUROSCIENCE
VOLUME 6 |JANUARY 2005 |5 9
R E V I E W S
loaded and unloaded by mild stimulation without reserve
pool mobilization19,28.The recycling pool cycles rapidly
during stimulation,whereas the reserve pool recycles
more slowly after the cessation of30 Hz stimulation30.
An unusual feature ofDrosophilaterminals concerns
the traffic in the centre ofthe boutons,which is popu-
lated by reserve pool vesicles.FM staining is observed
here after recovery from total vesicle loss in shibire
mutants28,or after high-frequency stimulation26and
certain drug treatments29,and FM dye release from the
centres has been noted,indicating vesicle mobilization
and exocytosis19,26.However,vesicles evidently remain
here only transiently,moving out to the periphery over
time.So,in resting preparations,bouton centres are
almost devoid ofsmall vesicles31(FIG.2b;see REF.32for
a different result).Moreover,synaptotagmin (and pre-
sumably,therefore,synaptic vesicle) immunofluores-
cence in this preparation is typically confined to the
periphery ofthe bouton33.Consistent with the idea of
trafficking,bouton centres were not devoid ofvesicles
during recovery from total vesicle loss in shibire
mutants34.Movement ofFM fluorescence in the oppo-
site direction,from the periphery ofthe boutons to the
central region,has also been observed30.
In summary, it seems that the central regions of
Drosophilaboutons are populated by the reserve vesicles
for only a briefperiod after heavy depletion-inducing
stimulation,and they are rapidly removed to the peri-
phery ofthe boutons.So,the distribution ofthe recycling
and reserve vesicles might depend on the conditions
under which the preparations were observed (H.Kuromi,
personal communication).Ultrastructural (FM-dye
photoconversion) observations ofDrosophilaprepara-
tions under various conditions would be useful.
Frog neuromuscular junction.In the cutaneous pectoris
and sartorius NMJs ofthe frog (Rana pipiens) (FIG.3),
the total vesicle pool contains ~500,000 vesicles25,35,36.An
initial phase of release (from the RRP) is exhausted
within ~0.5 s ofhigh-frequency stimulation8;these vesi-
cles make up ~2% ofthe total pool37.
The recycling pool cycles continuously at 2 Hz stim-
ulation (that is,the reserve pool is not mobilized),but it
is depleted within ~10 s of high-frequency (30 Hz)
stimulation.The recycling pool makes up ~10–20% of
the total vesicle population8,24,37.At the frog NMJ,the
RRP is evidently an integral part ofthe recycling pool,as
the rate ofRRP recovery depends on the rate ofrefilling
of the recycling pool8.The reserve pool, which only
starts to discharge after 10–15 s of30 Hz stimulation,
recycles slowly,with a time constant ofseveral minutes
(FIG.3e,blue arrows).The mixing between the recycling
and reserve pools (red arrow) is slow,in the order of
Hippocampal boutons.In cultured hippocampal neurons
(FIG.4),the total vesicle population in each bouton com-
prises 100–200 vesicles38(studies in hippocampal slices
have yielded larger values39).An immediately available
vesicle pool (RRP), containing ~5–20 vesicles40, is
released by hypertonic shock13,41,and isprobably the
Vesicle pools in five synaptic systems
Vesicle pools have been investigated in many systems,
using techniques such as electrophysiology and elec-
tron and fluorescence microscopy (BOX 1).The main
characteristics ofthe vesicle pools in five preparations
are summarized below.
Drosophilalarval neuromuscular junction.The nerve
terminal at the Drosophila NMJ (FIG.2) contains ~84,000
quanta,14–19% ofwhich constitute the recycling pool9.
A rapidly depleted component of release (RRP) was
observed,with a size of~300 quanta9.
FM-dye staining indicates that the recycling pool
(sometimes referred to as the exo–endo cycling
pool19,26,28–30) is typically found at the periphery ofthe
boutons.Heavy stimulation (30 Hz),but not mild stimu-
lation (10 Hz) or high potassium treatment,can also
recruit vesicles from the reserve pool,which is found
deeper within the boutons26.The recycling pool and the
reserve pool mix slowly, as the recycling pool can be
Table 1 | Characteristics of the vesicle pools
Pool Readily releasable
Recycling poolReserve pool
Size (% of
DockedScattered Scattered (bulk
of vesicle cluster)
Tens of seconds,
Slow mixing with
Low (high in
Released within<1 secondA few seconds
Mobility in resting
Fast mixing with
None — docked
Box 1 |Techniques used to determine vesicle pool characteristics
Three basic techniques have been used to measure vesicle pools — electron microscopy,
fluorescence microscopy and electrophysiology.All three techniques have been used,in
some form,with all five ofthe preparations that are discussed in this review:the frog
neuromuscular junction (NMJ)8,24,25,37,the DrosophilaNMJ9,26,28,29–31,139,neonatal
rodent cultured hippocampal boutons13,20,38,40,130,the neonatal rodent calyx ofHeld
presynaptic terminal10,14,21,50,64,65,157and the goldfish retinal bipolar nerve
terminal12,71,72,74,79,80,129.Electron microscopy studies have included freeze fracture
microscopy and transmission microscopy,sometimes involving serial reconstructions
and the use ofendocytic markers to label vesicles.
Fluorescence microscopy has relied primarily on styryl dyes (FM dyes),the
presynaptic uptake and release ofwhich can be monitored in living preparations in real-
time73.In FM-dye experiments,the synaptic preparation is bathed in a solution
containing the dye,and stimulated to induce exocytosis.The vesicles open to the
extracellular solution and their internal membrane leaflets are labelled with the dye.
When endocytosis is complete,the newly endocytosed synaptic vesicles remain labelled
and can be imaged (cell membranes are washed to remove the dye).
Electrophysiological studies have included recordings from postsynaptic cells,and,in
the case ofthe calyx ofHeld neurons and goldfish bipolar cells,presynaptic whole-cell
voltage clamp,including patch-clamp capacitance.
Other preparations have been used to characterize vesicle pools,including NMJs ofthe
snake161,162and crayfish163,Aplysia californicaneurons164,the squid giant synapse165and
lamprey reticulospinal synapses116.
6 0 |JANUARY 2005 |VOLUME 6
R E V I E W S
(and RRP) vesicles are rapidly retrieved by endocytosis
after stimulation,through mechanisms that do not
seem to require endosomal intermediates (for reviews,
see REFS 43,45).The reserve pool vesicles,when released,
probably recycle slowly27.An ultrafast mode ofendo-
cytosis — kiss-and-run47,48(see below) — might also
take place at this synapse,at least in the case ofthe RRP
Calyx ofHeld presynaptic terminal.The total number of
vesicles at this synapse in the rat (FIG.5) was estimated in
a recent electron microscopy study50to be at least 70,000
quanta.A higher estimate — ~188,000 vesicles — was
obtained in another study21.
Release from the calyx is depressed rapidly with stim-
ulation by high-frequency pulses (5–300 Hz (REFS 51,52)).
Furthermore,a finite pool ofsynaptic vesicles is rapidly
released on depolarization or calcium uncaging14,15,53–67.
The RRP consists of ~1,500–4,000 vesicles in the
postnatal day 8–11 calyx ofHeld nerve terminal15,54,59–65.
Two distinct components ofrelease have been observed
for the RRP:a fast component (~3 ms time constant of
release15) and a slow one (~30 ms),each making up
about halfofthe pool15,68.Interestingly,the fast compo-
nent ofrelease recovers more slowly (within seconds)
after depletion than the slow release component
(within 100 ms)15.
The underlying causes of the existence of the two
components have not yet been resolved69.It is possible
that the fast component ofrelease consists ofthose vesi-
cles that are situated closest to calcium channels,and the
slow component ofvesicles situated further away70.It is
tempting to propose that certain stimuli,such as high-
frequency trains ofaction potentials,release only one
component ofthe RRP (the fast component),as some
measurements ofthe RRP size using high-frequency
stimulation show relatively low values55–57.It should be
noted that measuring vesicle pool parameters in the
calyx ofHeld through analysis ofpostsynaptic responses
is by no means a trivial study58,66,67.When the calyx nerve
terminals were stimulated at 5 Hz for up to 20 min,or at
20 Hz for up to 5 min in the presence of an FM dye,
~5% ofall vesicles were labelled21.This indicates that the
recycling pool consists of~10,000 vesicles (although a
higher estimate of~20,000 vesicles was calculated using
certain assumptions on vesicle release at this synapse21).
High potassium stimulation caused a larger percentage
ofvesicles to recycle (for example,30% for 15 min of
stimulation),drawing vesicles from the reserve pool21.
The mixing rate between the RRP and the recycling
pool is unclear.As the RRP recovers rapidly (time con-
stant ~1 s (REFS 14,15)) it is likely that it is refilled
through mobilization of vesicles from the recycling
pool.Therefore,mixing ofthe two pools would be fast,
at least on stimulation.The mixing between the recy-
cling pool and the reserve pool seems to be slow, at
least with physiological stimulation,as the reserve pool
vesicles are reluctant to release21. Slow endosomal
intermediate-dependent endocytosis recycles reserve
vesicles,but a faster route functions for vesicles from
the recycling pool21.
same pool ofvesicles that is released within ~2 s of20 Hz
stimulation13,40,42.A pool ofvesicles that recycles repeat-
edly with stimulation (recycling pool20,43–45) comprises
10–20% ofthe vesicles20(for a review,see REF.43).
The rest ofthe vesicles (reserve pool) were reluctant
to undergo exocytosis in response to both electrical
stimulation and (short) high potassium application20,43.
Occasionally,vesicles other than those contained in the
RRP and recycling pool could be labelled45,46,but it is
not clear whether this is a consistent phenotype.Why
are reserve pool vesicles in hippocampal boutons so dif-
ficult to release? One possibility is that the stimulation
protocols that were used in hippocampal studies permit
the recycling pool to be refilled adequately by newly
retrieved vesicles,so the reserve pool vesicles remain
Mixing between the recycling and the reserve pools
seems to be slow,whereas mixing between the RRP
and the recycling pool is relatively fast.The recycling
10 µ µm
Figure 2 |Typical images from the Drosophila larval neuromuscular junction preparation.
a | FM 1-43 fluorescence image of a nerve terminal; note the fluorescent boutons. b | Cross-section
of a bouton; arrows indicate active zones; m, mitochondrion; v, synaptic vesicles; 1, axon.
c | Three-dimensional reconstruction of two Drosophila larval axons; the dark patches represent
synapses; the line separates the two muscles innervated by the axons. d | Postsynaptic current
recording from a muscle under continuous stimulation at 10 Hz (top); a few individual
postsynaptic responses at 0 s and 300 s after stimulation commenced (bottom). The readily
releasable pool (RRP) is depleted in the first few shocks — release can then be maintained almost
indefinitely by repeated recycling of the recycling pool at this frequency. e | Pool sizes and mixing
rates. Blue arrows indicate endocytosis; red arrows indicate mixing between pools. Red sphere
indicates total pool size relative to the other preparations; nA, nanoAmp. Panels b and c modified,
with permission, from REF.31© (1993) J ohn Wiley & Sons, Inc. Panel d modified, with permission,
from REF.9© (2000) Elsevier Science.
NATURE REVIEWS |NEUROSCIENCE
VOLUME 6 |JANUARY 2005 |6 1
R E V I E W S
that are attached to ribbons — structures that are
thought to collect vesicles from the cytosol and guide
them to the release sites81(see REF.82for a different inter-
Continuous depolarization seems to continuously
draw vesicles from a large reserve pool12,74.
The mixing rates between the pools have not been
investigated in detail,but,after exocytosis,the vesicles on
the ribbons seem to be replenished rapidly by reserve
vesicles,rather than by recycling vesicles,which indicates
rapid mixing72.Both fast and slow recycling mechanisms
might return vesicles to the reserve pool (see below).
FLUORESCENCE RECOVERY AFTER PHOTOBLEACHING (FRAP)
experiments in goldfish bipolar nerve terminals72and
cone terminals ofthe anole lizard retina83have shown that
vesicles in the cytoplasm ofresting terminals are highly
mobile,which might reflect a lack ofsynapsin in these
ribbon synapses(see below).
Summary of pool size. In the four non-ribbon
synapses,the extensive range of vesicle pool sizes can
be unified by normalizing to the total length ofactive
zone material84.The total number ofsynaptic vesicles
per active zone in the DrosophilaNMJ is 87,000 vesi-
cles, which, divided between 400 active zones31, is
equal to 217 vesicles per active zone;in a hippocampal
bouton there are 200 vesicles per active zone38;and in
the calyx of Held terminal there are 200,000 vesicles
for 600 active zones50,85,giving an average of 333 per
active zone.At the frog NMJ,the active zones are about
six times longer than in the other preparations,per-
haps reflecting oligomerization of the smaller units.
On this basis,500,000 vesicles shared between 300 ×6
active zones equals 270 vesicles per active zone.The
goldfish bipolar cell,with 700,000 vesicles and 60 rib-
bons,gives a very different result:more than 11,000
vesicles per ribbon71.
Similar results are obtained when the number of
docked vesicles at each active zone is considered:~5–10
vesicles in hippocampal boutons40,~2 vesicles in the
calyx ofHeld50and ~40 vesicles per long active zone in
the frog NMJ,which,being about six times longer than
the other active zones, gives about 7 vesicles for the
equivalent area (S.O.R.and W.J.B.,unpublished obser-
vations).In the goldfish bipolar terminal each ribbon
tethers ~110 vesicles71,76.
Vesicle pools are not anatomically segregated
The schematic diagram in FIG.1ais typical ofthose that
have been drawn for several decades, showing the
three vesicle pools morphologically segregated into
distinct clusters.This diagram implies that the further
the cluster is from the presynaptic membrane, the
slower it is to release.Until recently,there was no direct
evidence to either support or refute this model.The
vesicles in the RRP must, by definition, be able to
undergo exocytosis immediately after stimulation,so
must lie at or close to the presynaptic membrane.
However,the vesicles that make up the other pools are
recruited at a more leisurely pace during repetitive
stimulation,and could,in principle,be transported to
Goldfish retinal bipolar nerve terminals.Estimates from
electron microscopy place the total pool ofsynaptic vesi-
cles in the goldfish retinal bipolar nerve terminal (FIG.6)
at around 700,000 quanta (~500,000 or 900,000 quanta
in a small and a large nerve terminal,respectively71;see
REF.72for a lower estimate),consistent with styryl dye
When exocytosis was probed using styryl dye imag-
ing12(FIG.1b) or interference reflection microscopy75,
three pools ofvesicles were indicated by three kinetic
phases ofrelease.A readily releasable component was
estimated (typically from brief(tens ofmilliseconds)
depolarization) at 1,000–1,800 vesicles11,12,23,76,77.Longer
depolarization pulses (but typically less than 1 s)
released a second pool of~3,000–4,400 vesicles12,76–78.
The two pools add up to ~6,000 vesicles in many stud-
ies79,80. The number of vesicles in these two pools,
which together account for only 1–2% ofthe total vesi-
cle population,corresponds to the number ofvesicles
(FRAP).A method used to
measure the lateral diffusion of
elements.It requires tagging
ofthe molecule ofinterest with
a fluorescent marker,
photobleaching ofthe label with
a pulse oflaser light and a
subsequent measure ofthe rate
offluorescence recovery into the
bleached area as other labelled
molecules move into it.
10 µ µm
0.5 µ µm
Figure 3 |Typical images from the frog neuromuscular junction preparation. a | FM 1-43
fluorescence image of a nerve terminal. b | Electron micrograph of a cross-section through the
nerve terminal. Arrows indicate active zone. c | Three-dimensional reconstruction of an
approximately 2 µm-long nerve terminal segment. Active zones are shown in red.
d | Postsynaptic potential recording under continuous stimulation at 30 Hz (top); a few
postsynaptic responses at 0 s and 50 s after stimulation commenced are shown below (image
courtesy of D. A. Richards). In the bottom trace, vertical lines are shock artefacts; synaptic
responses have declined to baseline. The RRP is rapidly exhausted, followed by recycling pool
depletion. Reserve pool release is maintained for at least one minute. e | Pool sizes and mixing
rates. Blue arrows indicate endocytosis; red arrows indicate mixing between pools. Red sphere
indicates total pool size relative to the other preparations. Panel c reproduced, with permission,
from REF.37© (2004) American Association for the Advancement of Science.
6 2 |JANUARY 2005 |VOLUME 6
R E V I E W S
Consistent with this interpretation,the number of
docked vesicles in some preparations is similar to the
number ofreadily releasable vesicles50,71.In hippocam-
pal boutons,when RRP vesicles released by short tetanic
stimulation were labelled with an FM dye and counted
under the electron microscope,the number oflabelled
vesicles correlated closely with the total number of
docked vesicles in the same terminal (although there is
evidence that not all docked vesicles are easily releasable,
so some might not belong to the RRP37,92).The fact that
recycling pool vesicles recycle selectively without reserve
pool release meant that they could be labelled selectively,
and their position was investigated using electron
microscopy.In hippocampal boutons20,43,the calyx of
Held21and the frog NMJ37(FIG.7a),the recycling pool
was scattered throughout the nerve terminals under
stimulation conditions that triggered continuous recy-
cling.In DrosophilaNMJs, fluorescence microscopy
showed that the recycling pool occupies the periphery of
the synaptic boutons28,and similarly,at the frog NMJ,
recycling pool vesicles tended to be excluded from the
cores ofthe vesicle clusters37.The reserve pool accounts
for most ofthe vesicle cluster in each preparation.At the
frog NMJ,the reserve pool vesicles did not seem to be
particularly excluded from near-active zone sites37,
although relatively more RRP vesicles were found in this
position.In lamprey reticulospinal synapses reserve vesi-
cles seem to form the bulk ofthe vesicle clusters that are
distant from the active zone93.
How are vesicles clustered?In resting terminals most
synaptic vesicles are immobile94,95,131.Synapsin is the
oldest,and still the best,candidate for the ‘glue’ that
binds them together96,although several studies have
prompted important adjustments to the original
‘synapsin hypothesis’, which was based mainly on
observations ofsynapsin localization on the cytoplas-
mic surface of synaptic vesicle membranes96,and the
reduced affinity for vesicle binding after phosphoryla-
tion of synapsin by calcium/calmodulin-dependent
protein kinase II (CaMKII (REFS 97,98)).Vesicles in the
lamprey reticulospinal synapse that were distant from
the presynaptic membrane were specifically lost after
acute anti-synapsin antibody injection93,and a similar
phenomenon was observed in mice lacking synapsin I
(REFS 99,100).Also lost with the synapsin-positive pool
was the ability to sustain high-frequency release,
although low-frequency release was maintained93.
These observations led to the suggestion that synapsin
holds vesicles together specifically in the reserve pool96.
However,other observations show that synapsin
molecules do not always discriminate between reserve
and non-reserve vesicles.For example,synapsin disper-
sion from vesicle clusters has been observed with stim-
ulation that triggers recycling pool-only release101.
Moreover,synapsin knockouts in mice not only reduce
the overall number ofvesicles99,102,but can also affect the
size ofthe recycling pool103or the RRP104.Furthermore,
rapid depression of release (probably not related to
reserve pool mobilization) has been observed in
synapsin I/II knockouts102.In addition,the complexity
release sites from any location within the entire vesicle
population.Recent evidence indicates that,at least in
some preparations,vesicles in the recycling and reserve
pools are intermixed to a considerable degree.
As the RRP can be depleted within a few millisec-
onds or tens ofmilliseconds with strong stimuli such as
calcium uncaging23or depolarization15,59,60,65,vesicles
belonging to the RRP must be docked at or lie very close
to the active zones86,87— and consequently to voltage-
gated calcium channels88— and they would also have
their SNARE PROTEINS in a release-ready configuration89.
The active zone machinery interacts in a complex man-
ner with the morphologically docked vesicles90,which
might promote interaction of vesicular components
with calcium channels91,resulting in rapid release on
A family ofmembrane-tethered
coiled-coil proteins that are
required for membrane fusion in
exocytosis (such as during
neurotransmitter release) and
other membrane transport
complexes are formed between
vesicle SNAREs and target-
membrane SNAREs,they pull the
two membranes together,
presumably causing them to fuse.
5 µ µm
0.5 µ µm
Figure 4 |Typical images from the rat cultured hippocampal preparation. a | FM 1-43
fluorescence image of a field showing numerous labelled presynaptic boutons. b | Electron
micrograph of a cross-section through a bouton. Arrowheads indicate the two edges of the
active zone in this image; the black arrows point to two docked vesicles; a non-docked
vesicle near the active zone is shown by the white arrow. c | Three-dimensional
reconstruction of a bouton. d | Postsynaptic response to hypertonic sucrose application
(bar); this treatment selectively releases the RRP. pA, picoAMP. e | Pool sizes and mixing
rates. Blue arrows indicate endocytosis; red arrows indicate mixing between pools. Red
sphere indicates total pool size relative to the other preparations. Panel b modified, with
permission, from REF.166© (2001) Elsevier Science. Panel c reproduced, with permission,
from REF.40© (2001) Macmillan Magazines Ltd. Panel d modified, with permission, from
REF.13© (1996) Elsevier Science.
NATURE REVIEWS |NEUROSCIENCE
VOLUME 6 |JANUARY 2005 |6 3
R E V I E W S
These results are not fully explained by any simple
model.For example,in synapsin I knockouts,synaptic
depression is increased102and FM-dye uptake is
decreased103,whereas the opposite has been reported for
synapsin III knockouts110.
In two ribbon synapses, including the goldfish
bipolar nerve terminal,reserve pool vesicles are highly
mobile72,83and the first vesicles to be released are
immobilized on ribbons,in contrast to the situation in
conventional synapses.Interestingly,ribbon synapses
lack synapsin, and the uninhibited movement of
reserve pool vesicles is adequate to replace vesicles on
the ribbons72.However,it should be noted that ultra-
structural observations have shown structures that
interlink vesicles in ribbon synapses111.
If recycling vesicles in non-ribbon synapses are
defined by their relative lack of synapsin crosslinks,
some sort of‘cage’ must prevent their dispersal and
consequent loss from the cluster.One candidate for this
role is actin,which tightly surrounds vesicle clusters in
some synapses112,113.However,perturbation of actin
dynamics does not noticeably affect the overall FM
staining pattern ofthe vesicle clusters112–114.
It is also possible to visualize a model in which all
vesicles are equally mobile,but only some (recycling
pool) can exocytose efficiently on interaction with the
active zone. This model would explain the lack of
anatomical segregation between the vesicle pools,and it
fits well with the data from mammalian synapses,where
the reserve pool vesicles are reluctant to release,and
might, therefore, have limited fusion competence.
However,it does not explain systems such as the frog
and DrosophilaNMJs,in which the reserve pool vesicles
are perfectly release-competent but are mobilized only
after recycling pool depletion.
How are vesicles in different pools mobilized?As dis-
cussed above,RRP vesicles are thought to be docked at
the active zones, so would not need to move to be
released.By contrast,the recycling pool vesicles are not
necessarily found in the vicinity ofthe active zones.At
least two models might explain their exocytosis:
directed movement towards the active zones and ran-
dom diffusion.In the first model,the recycling pool
vesicles would require access to cytoskeletal ‘tracks’ that
guide them to the active zone;in the second,simple dif-
fusion could be sufficient to allow these vesicles to reach
the active zone (FIG.7d).
What cytoskeletal elements could guide vesicles
towards the release sites? The actin cytoskeleton is
abundant at motor nerve terminals112,115, where it
mainly surrounds synaptic vesicle clusters112,113,116,117.
Notably less actin is found within the clusters,although
filaments resembling actin have been observed in
ultrastructural studies118.Actin filaments take part
in organelle movement in various systems119,120,and
there is strong evidence for the involvement of actin
in the transport of synaptic vesicles back to vesicle
clusters after endocytosis in lamprey reticulospinal
synapses116,121and cultured cells122but not in hippo-
ofthe possible roles ofphosphorylation is underscored
by the identification ofdifferent kinases that phospho-
rylate synapsin in vivo105,and the report that CaMKII
knockouts show increased rather than decreased
All three synapsin genes have been knocked out,
singly and in combination99,100,103,107,108. None of the
mutations,including the triple knockout (P.Greengard,
personal communication), are lethal. Similarly,
Drosophilamutants that lack all synapsins are viable109.
2 mM Calcium
4 mM Calcium
10 & 30 ms
1 ms200 fF
10 µ µm
5 µ µm
10 µ µm
Figure 5 |Typical images from the rat calyx of Held preparation. a | FM 1-43 fluorescence
image of a single identified calyx of Held nerve terminal. b | Top: low power cross-section of the
preparation; the presynaptic terminal is yellow; the postsynaptic cell is blue; and the
postsynaptic nucleus is red. Bottom: higher power image (see box in top panel); arrows indicate
active zones. c | Three dimensional reconstruction (presynaptic terminal in orange; postsynaptic
cell in blue). d | Postsynaptic current recording under 100 Hz stimulation; the RRP is rapidly
depleted. e | Presynaptic capacitance response to short (1–30 ms) depolarization; the response
(indicative of readily releasable pool exocytosis) plateaus at ~10 ms of depolarization. The gap
between baseline and response to stimulation can be attributed to the fact that capacitance
recording is not reliable during stimulation. Cm, membrane capacitance; fF, femtoFarad. f | Pool
sizes and mixing rates. Blue arrows indicate endocytosis; red arrows indicate mixing between
pools. Red sphere indicates total pool size relative to the other preparations. Panel a
reproduced, with permission, from REF.21© (2003) Society for Neuroscience. Panels b and c
reproduced, with permission, from REF.50© (2002) Society for Neuroscience. Panel d modified,
with permission, from REF.10© (1999) Elsevier Science. Panel e modified, with permission, from
REF.65© (2001) Elsevier Science.
6 4 |JANUARY 2005 |VOLUME 6
R E V I E W S
Recycling of synaptic vesicle pools
Selective recycling to the pool oforigin.Do vesicle pools
tend to maintain their pool ‘tags’ throughout synaptic
vesicle recycling,or is there mixing between pools? To
explore this question we will consider the three pools
At the frog NMJ,recovery of the RRP depends on
refilling ofthe recycling pool,indicating that the RRP
vesicles represent a subpopulation ofthe recycling pool8,
and that recycling RRP vesicles probably lose their RRP
status.At goldfish bipolar nerve terminals,the newly
recycled vesicles refill the reserve pool72,129, and the
vesicles on the ribbons do not recycle selectively.
Recycling of the RRP has not been investigated in
In hippocampal boutons,a series ofelegant experi-
ments combining electrophysiology and styryl dye
imaging showed that RRP vesicles can be released
repeatedly130(FIG.8a).Pyle and colleagues estimated that
the RRP mixes with the recycling pool in about 20 s
ifthe RRP is depleted electrically,or in about 5 min if
hypertonic shock is used.
The fate ofrecycling pool vesicles at the frog NMJ
has been investigated using several experimental models,
all ofwhich indicate that recycling pool vesicles selec-
tively return to their pool oforigin.In one experiment
(FIG.8b),a fully loaded terminal was ‘buzzed’ (briefly
stimulated at a high rate) to release the recycling pool.
The vesicles lost their dye and recycled.After 15 min,
the preparation was destained using prolonged high-
frequency stimulation.When the stimulation began,
little or no dye was lost (even though the terminal was
secreting a normal number ofquanta);the buzz-recycled,
dye-empty vesicles had returned selectively to the recy-
cling pool.This result also shows that the recycling and
reserve pools operate sequentially:reserve pool vesicles
are not released in large numbers until the recycling
pool is depleted.
Recycling pool vesicles cycle repeatedly,without
reserve pool release,during prolonged physiological
stimulation at the frog NMJ8,DrosophilaNMJ (for a
review,see REF.19),hippocampal boutons20,43and the
calyx of Held21.This consistent observation is proba-
bly the strongest piece of evidence for the existence of
distinct vesicle pools.
Recycling routes for different pools.Two classic studies
ofsynaptic vesicle recycling,involving electrophysiology
and electron microscopy,were undertaken at the frog
NMJ in the early 1970s (REFS 25,35,131,132).Heuser and
Reese25showed that high-frequency stimulation resulted
in vesicle depletion, and that vesicle reformation
required endocytic intermediates.Using low-frequency
stimulation,Ceccarelli and collaborators35found no
significant vesicle depletion,and no requirement for
endocytic intermediates.More recent experiments on
this preparation8,24have shown that stimuli that release
the reserve pool result in slow endocytosis that proceeds
through the formation ofinfoldings,whereas when the
recycling pool was selectively released no such inter-
mediates were observed.The recycling pool vesicles
Actin disruption experiments produced a wide
spectrum ofresults,including depression ofrelease at
Drosophilalarval,frog and snake NMJs28,112,115,at the
calyx ofHeld123and in chromaffin cells124;no change in
release parameters at goldfish bipolar terminals114,125;
and slight potentiation ofrelease at hippocampal bou-
tons113,126.The effects of actin disruption might also
depend on the age ofthe synapse,with vesicle cycling
affected less in mature synapses127. Blocking acto-
myosin movement with agents that inhibit myosin
light-chain kinase hampers release in hippocampal
boutons128.It is important to note that reserve pool
release, but not recycling pool release, seems to be
impeded by actin disruption at the DrosophilaNMJ26.
0.5 ms pulse
25 ms pulse
Figure 6 |Typical images from the dissociated goldfish retinal bipolar cell preparation.
a | FM 1-43 fluorescence image; the dye decorates the whole cell (see brightfield image in inset), but
is only internalized in the nerve terminal. b | Cross section of the nerve terminal; top, low power
image. c | Three-dimensional reconstruction of a nerve terminal; active zones are indicated by red
circles. d | Presynaptic capacitance responses to short depolarization (left, 0.5 ms, and right, 25 ms).
Only RRP vesicles are released by such brief stimuli; the response plateaus at ~20 ms of
depolarization, indicating total depletion of the RRP. The decay of the signal indicates endocytosis.
e | Pool sizes and mixing rates. Blue arrows indicate endocytosis. Red sphere indicates total pool
size relative to the other preparations. Panel a reproduced, with permission, from REF.74© (1996)
Elsevier Science. Panel b reproduced, with permission, from REF.72© (2004) Elsevier Science.
Panel c reproduced, with permission, from REF.71© (1996) Elsevier Science. Panel d modified, with
permission, from REF.11© (1996) Elsevier Science.
NATURE REVIEWS |NEUROSCIENCE
VOLUME 6 |JANUARY 2005 |6 5
R E V I E W S
Similar fast and slow recycling pathways have been
shown in ultrastructural studies in the adult Drosophila
NMJ138–141and have also been indicated by fluorescence
studies in the Drosophilalarval NMJ30.In mammalian
synapses,reserve pool vesicles are reluctant to release,so
it was not surprising that the original styryl dye uptake
studies at this synapse141–144indicated a single recycling
pathway.This pathway recovered vesicles with a half
time of~20 s,consistent with vesicles fusing with the
plasma membrane and being retrieved through a
clathrin-mediated mechanism.A more direct investiga-
tion using FM-dye photoconversion showed limited
bulk uptake20,43,45.Also,the vesicles were found to main-
tain their identity throughout vesicle recycling145,pro-
viding further evidence against bulk endocytosis.This is
consistent with the strong resistance ofhippocampal
synapses to vesicle depletion (compare REFS 146and 147).
When prolonged stimulation (many minutes) by high
potassium depolarization was used in mammalian
synapses21,27,endocytic intermediates that were indicative
ofbulk endosomal uptake were evident.So,it is possible
that when the normally reluctant reserve vesicles are
released,they endocytose through a slow pathway.One
seem toendocytose directly from the plasma mem-
brane, probably through CLATHRIN-coat-dependent
mechanisms37.So,two vesicle recycling routes can be
envisaged — bulk endocytosis,and direct endocytosis
from the plasma membrane (FIG.7b) — depending on
the strength ofstimulation.
Slow endocytosis (of reserve pool vesicles) might
also require endosomal sorting ofvesicle components,
as originally proposed by Heuser and Reese25.The small
GTPase Rab5is present mainly on early endosomes
and,through interaction with several effectors,regulates
traffic through this compartment135.Rab5 was found on
synaptic vesicles134,135, and overexpression of Rab5a
resulted in the formation ofabnormal endosomes in the
axons ofcultured hippocampal neurons135.Moreover,
when Rab5 function was perturbed in Drosophilamotor
nerve terminals by expression of a DOMINANT-NEGATIVE
mutant form,vesicle cycling was disrupted,with release
and endocytosis being inhibited,and accumulation of
endosome-like structures in the terminals136.In addi-
tion,the endosomal SNARE Vti1a is enriched on synaptic
vesicles,where it is part ofa SNARE complex (but not the
A vital structural component of
coated vesicles that are
implicated in protein transport.
Clathrin heavy and light chains
form a triskelion,the main
building element ofclathrin
A mutant molecule that can
form a heteromeric complex
with the normal molecule,
knocking out the activity ofthe
Synaptic current (normalized)
Seconds at 10 Hz
Fraction release that is recycled
Control minus shibire=recycled
Figure 7 |Topology, recycling and mobilization of vesicle pools. a | The recycling pool is scattered throughout the vesicle
cluster. Single cross section through (left) and three-dimensional reconstruction of (right) frog motor nerve terminals in which the
recycling pool was selectively labelled. b | Proposed recycling model. The reserve pool (pink) recycles slowly, through formation
of infoldings and their break-off through clathrin coat-dependent mechanisms. The recycling pool (purple) cycles through direct
endocytosis from the plasma membrane. This process might be clathrin coat dependent, or might rely on transient fusion (kiss-
and-run). Kiss-and-run might be used especially by readily releasable pool (RRP) vesicles. c | Newly recycled vesicles maintain
most vesicle release at Drosophila synapses. Left: postsynaptic recording at non-permissive temperature in control
neuromuscular junctions (NMJ s) or shibire mutant NMJ s at 10 Hz of stimulation. The release is maintained almost indefinitely in
controls, but it decays rapidly in the mutant. Right: comparison of the release for the two conditions. The difference between the
two is shown in green. The blue line shows the fraction of the release (per shock) that is performed by newly recycled vesicles.
d | Proposed model of mobilization. The RRP vesicles are docked and do not require mobilization. The reserve vesicles form
most of the cluster and are tightly crosslinked, possibly by synapsin. The recycling pool vesicles are not as heavily crosslinked,
so are more mobile. They might be able to diffuse to the active zone (arrows, left). Alternatively (right), they might have access to
cytoskeletal elements (for example, actin118) that direct them towards the active zone. Panel a reproduced, with permission, from
REF.37© (2004) American Association for the Advancement of Science. Panel c reproduced, with permission, from REF.9©
(2004) Elsevier Science.
6 6 |JANUARY 2005 |VOLUME 6
R E V I E W S
Kiss-and-run recycling. A fraction of the vesicles in
mammalian systems might recycle through an ultrafast
endocytic pathway,known as kiss-and-run155,in which
vesicles fuse transiently to the plasma membrane,and
reform by the simple closure ofa fusion pore.Data in
support ofthis model have been obtained in hippocam-
pal boutons using differential release from vesicles of
dyes with different hydrophobicities130,156,incomplete
release ofdye from single vesicles on exocytosis44and
rapid endocytosis monitored by fluorescence49.In calyx
ofHeld synapses and isolated pituitary nerve terminals,
rapid endocytosis was observed using the PATCH-CLAMP
CAPACITANCE TECHNIQUE157,158.However,only a fraction of
the RRP48or recycling pool148vesicles use this pathway.
Part ofthe intuitive appeal ofvery rapid endocytosis
is the possibility offast vesicle recycling,which could
provide an important selective advantage in synaptic
function.However,although kiss-and-run data indicate
that endocytosis is fast (sub-second in some cases),the
only measure ofcomplete recycling indicates that 20–30 s
are needed for the vesicle to become release-competent
again44— not much faster than estimates for vesicles
that recycle into the recycling pool through the clathrin-
Virtually no evidence for kiss-and-run endocytosis
has been obtained in goldfish bipolar cells75,154or at the
frog NMJ8,24,37,but in Drosophilaevidence ofkiss-and-
run endocytosis has been claimed for knockouts of
endophilin — a protein that is involved in endocytosis159.
Moreover, electrophysiological recordings (FIG. 7c)
showed that the synaptic rundown in shibiremutants at
high temperature is very rapid9.Ordway and collabora-
tors160analysed synaptic depression in Drosophilawith
higher time-resolution,and found that the release from
shibireneurons at the non-permissive temperature was
concern in interpreting data from mammalian systems is
that they are often investigated at room temperature:
switching to physiological temperature affects both
exocytosis and endocytosis ofvesicles148,149.
In goldfish bipolar cells,it has been shown that a slow
endocytic process (time constant ~10 s) and a fast process
(time constant ~1.25 s) coexist12.A morphological corre-
late to the two pathways was indicated by the experiments
in REF.129,which showed that exocytosis was followed by
bulk membrane uptake (slow endocytosis?) and possibly
also single vesicle uptake (fast endocytosis?).Formation of
vesicles from infoldings after stimulation through
clathrin-coat mechanisms has also been suggested in a
different ribbon synapse (frog saccular hair cells150).Both
endocytic processes seem to refill the reserve pool72.
In conclusion,it can be proposed that recycling pool
vesicles are generally retrieved through endocytosis
directly from the plasma membrane,whereas reserve
pool release is followed by bulk endocytosis.One inter-
pretation is that molecules in recycling pool vesicle
membranes are ‘glued’ tightly together,and the vesicle
membrane is continuous after exocytosis.The fused
vesicle membrane patch might be located by the endo-
cytic machinery through interactions between adaptor
proteins and integral vesicle membrane proteins151–153.
By contrast,ifthe vesicle components spread into the
plane ofthe membrane after fusion156,bulk endocyto-
sis72,125,129might be the only solution to retrieving them.
Sorting ofthe vesicular components might occur during
(or before) clathrin-mediated vesicle formation from
the infoldings.However,it should be noted that bulk
endocytosis is classically viewed as an emergency
response that operates when the normal endocytic
routes are overwhelmed by intense release87,which is an
entirely plausible interpretation.
A glass pipette is sealed against
the membrane and an
alternating voltage signal
applied.The induced current is
recorded and used to calculate
the membrane capacitance.The
capacitance ofthe membrane is
proportional to its surface,and
so gives a measure ofthe
amount ofexocytosis or
endocytosis taking place.
0 2040 60
Fluorescence (% initial)
Figure 8 | Readily releasable pool and recycling pool vesicles can recycle selectively. a | The recycling pool in hippocampal
boutons was loaded with the styryl dye FM 2-10, and the readily releasable pool (RRP) was repeatedly released by sucrose
application; this was measured by postsynaptic recording (top) or by fluorescence release (bottom). Neurotransmitter release
recovered rapidly between sucrose applications, but a different result was obtained for fluorescence release, indicating that newly
recycled, FM-empty RRP vesicles were released on the second sucrose application. Therefore, the RRP recycles selectively, but the
mixing of the RRP and recycling pool is fast (seconds to minutes). EPSC, excitatory postsynaptic current; pA, picoAmp. b | Both the
recycling pool and the reserve pool were labelled in frog NMJ terminals. The terminals were briefly stimulated (‘buzzed’) to release
the recycling pool, and the terminals were then rested for a few minutes. The preparations were then destained. Note that dye-
empty (recycling pool) vesicles are released during the first ~10 s of stimulation (dotted line), indicating that the recycling pool
vesicles released during the buzz have recycled selectively to their pool of origin. Panel a modified, with permission, from REF.130©
(2000) Elsevier Science. Panel b modified, with permission, from REF.8© (2003) Elsevier Science.
NATURE REVIEWS |NEUROSCIENCE
VOLUME 6 |JANUARY 2005 |6 7
R E V I E W S
markedly depressed in comparison to normal
synapses by the time of the second shock of a 50 Hz
train (release on the first shock was normal). The
depression of release preceded any significant overall
loss of synaptic vesicles160,so it probably reflected the
selective depletion of RRP or recycling pool vesicles
rather than a depletion ofthe whole vesicle pool.This
result indicates that the recovery ofthese vesicles must
be very fast (kiss-and-run?), and that this synapse
relies heavily on recycled vesicles during continuous
Conclusion and perspectives
Most synaptic systems seem to rely on several pools of
vesicles — docked,readily releasable vesicles;recycling
pool vesicles that maintain exocytosis on mild stimula-
tion;and a large population ofreserve vesicles that are
usually released only after stimulation has depleted the
1. Südhof, T. C. The Synaptic Vesicle Cycle. Annu. Rev.
Neurosci. 27, 509–547 (2004).
Birks, R. & MacIntosh, F. C. Acetylcholine metabolism of a
sympathetic ganglion. Canad. J . Biochem. Physiol. 39,
Liley, A. W. & North, K. A. An electrical investigation of effects
of repetitive stimulation on mammalian neuromuscular
junction. J . Neurophysiol. 16, 509–527 (1953).
Elmqvist, D. & Quastel, D. M. J . A quantitative study of end-
plate potentials in isolated human muscle. J . Physiol. (Lond.)
178, 505–529 (1965).
A classic paper showing end-plate potential rundown
from an initial high amplitude to a lower, steady level
during repetitive stimulation of human motor nerve
Elmqvist, D., Hofmann, W. W., Kugelberg, J . & Quastel, D. M.
An electrophysiological investigation of neuromuscular
transmissionin myasthenia gravis. J . Physiol. (Lond.)174,
Zucker, R. S. & Regehr, W. G. Short-term synaptic plasticity.
Annu. Rev. Physiol. 64, 355–405 (2002).
Rettig, J . & Neher, E. Emerging roles of presynaptic proteins
in Ca2+-triggered exocytosis. Science298, 781–785 (2002).
Richards, D. A., Guatimosim, C., Rizzoli, S. O. & Betz, W. J .
Synaptic vesicle pools at the frog neuromuscular junction.
Neuron39, 529–541 (2003).
Delgado, R., Maureira, C., Oliva, C., Kidokoro, Y. & Labarca, P.
Size of vesicle pools, rates of mobilization, and recycling at
neuromuscular synapses of a Drosophila mutant, shibire.
Neuron28, 941–953 (2000).
10. Schneggenburger, R., Meyer, A. C. & Neher, E. Released
fraction and total size of a pool of immediately available trans-
mitter quanta at a calyx synapse. Neuron23, 399–409 (1999).
11. Mennerick, S. & Matthews, G. Ultrafast exocytosis elicited
by calcium current in synaptic terminals of retinal bipolar
neurons. Neuron17, 1241–1249 (1996).
12. Neves, G. & Lagnado, L. The kinetics of exocytosis and
endocytosis in the synaptic terminal of goldfish retinal bipolar
cells. J . Physiol. (Lond.)515, 181–202 (1999).
Detailed investigation of the properties of vesicle
pools in the goldfish retinal bipolar cell, using both
electrophysiological and imaging techniques.
13. Rosenmund, C. & Stevens, C. F. Definition of the readily
releasable pool of vesicles at hippocampal synapses.
Neuron16, 1197–1207 (1996).
Application of a hypertonic solution evokes a burst of
release that settles to a low level in the continued
presence of the stimulus; the burst is defined as the
RRP, and correlates with the vesicles released on the
first few stimuli of a tetanus.
14. Wu, L. G. & Borst, J . G. The reduced release probability of
releasable vesicles during recovery from short-term synaptic
depression. Neuron23, 821–832 (1999).
15. Sakaba, T. & Neher, E. Calmodulin mediates rapid
recruitment of fast-releasing synaptic vesicles at a calyx-type
synapse. Neuron32, 1119–1131 (2001).
In calyx of Held nerve terminals the RRP consists of
two components, a rapidly releasable one and a
slowly releasable one. These two components also
have different kinetics of recovery.
16. Hanse, E. & Gustafsson, B. Vesicle release probability and
pre-primed pool at glutamatergic synapses in area CA1 of
the rat neonatal hippocampus. J . Physiol. (Lond.)531,
17. Hanse, E. & Gustafsson, B. Paired-pulse plasticity at the
single release site level: an experimental and computational
study. J . Neurosci. 21, 8362–8369 (2001).
18. Hanse, E. & Gustafsson, B. Release dependence to a
paired stimulus at a synaptic release site with a small
variable pool of immediately releasable vesicles. J . Neurosci.
22, 4381–4387 (2002).
19. Kuromi, H. & Kidokoro, Y. Two synaptic vesicle pools,
vesicle recruitment and replenishment of pools at the
Drosophila neuromuscular junction. J . Neurocytol. 32,
20. Harata, N., Ryan, T. A., Smith, S. J ., Buchanan, J . & Tsien,
R. W. Visualizing recycling synaptic vesicles in hippocampal
neurons by FM 1-43 photoconversion. Proc. Natl Acad. Sci.
USA 98, 12748–12753 (2001).
21. de Lange, R. P., de Roos, A. D. & Borst, J . G. Two modes of
vesicle recycling in the rat calyx of Held. J . Neurosci. 23,
There are distinct fast and slow endocytic
mechanisms in calyx of Held nerve terminals. Slow
endocytosis might function only with strong (non-
22. von Gersdorff, H. & Matthews, G. Electrophysiology of
synaptic vesicle cycling. Annu. Rev. Physiol. 61, 725–752
23. Heidelberger, R., Heinemann, C., Neher, E. & Matthews, G.
Calcium dependence of the rate of exocytosis in a synaptic
terminal. Nature371, 513–515 (1994).
24. Richards, D. A., Guatimosim, C. & Betz, W. J . Two
endocytic recycling routes selectively fill two vesicle pools in
frog motor nerve terminals. Neuron27, 551–559 (2000).
25. Heuser, J . E. & Reese, T. S. Evidence for recycling of
synaptic vesicle membrane during transmitter release at the
frog neuromuscular junction. J . Cell Biol. 57, 315–344
This seminal paper describes vesicle cycling at the
frog NMJ under conditions of strong stimulation. The
roles of endocytic intermediates and clathrin are
emphasized; endosomal sorting is proposed as a
stage during recycling.
26. Kuromi, H. & Kidokoro, Y. Tetanic stimulation recruits
vesicles from reserve pool via a cAMP-mediated process in
Drosophila synapses. Neuron27, 133–143 (2000).
27. Takei, K., Mundigl, O., Daniell, L. & De Camilli, P. The
synaptic vesicle cycle: a single vesicle budding step
involving clathrin and dynamin. J . Cell Biol. 133, 1237–1250
The authors describe vesicle reformations through
bulk endocytosis. Membrane retrieved through bulk
uptake is broken down by clathrin-coated-vesicle
formation; new synaptic vesicles are generated by
28. Kuromi, H. & Kidokoro, Y. Two distinct pools of synaptic
vesicles in single presynaptic boutons in a temperature-
sensitive Drosophila mutant, shibire. Neuron20, 917–925
FM 1-43 imaging of Drosophila larval NMJ s shows
that the recycling pool is localized at the periphery of
the boutons and the reserve pool is in the centres.
29. Kuromi, H. & Kidokoro, Y. The optically determined size of
exo/endo cycling vesicle pool correlates with the quantal
content at the neuromuscular junction of Drosophila larvae.
J . Neurosci. 19, 1557–1565 (1999).
30. Kuromi, H. & Kidokoro, Y. Selective replenishment of two
vesicle pools depends on the source of Ca2+at the
Drosophila synapse. Neuron35, 333–343 (2002).
31. Atwood, H. L., Govind, C. K. & Wu, C. F. Differential
ultrastructure of synaptic terminals on ventral longitudinal
abdominal muscles in Drosophila larvae. J . Neurobiol. 24,
32. Estes, P. S. et al. Traffic of dynamin within individual
Drosophila synaptic boutons relative to compartment-
specific markers. J . Neurosci. 16, 5443–5456 (1996).
33. Marek, K. W. & Davis, G. W. Transgenically encoded protein
photoinactivation (FlAsH-FALI): acute inactivation of
synaptotagmin I. Neuron36, 805–813 (2002).
34. Kosaka, T. & Ikeda, K. Possible temperature-dependent
blockage of synaptic vesicle recycling induced by a single gene
mutation in Drosophila. J. Neurobiol.14, 207–225 (1983).
35. Ceccarelli, B., Hurlbut, W. P. & Mauro, A. Turnover of
transmitter and synaptic vesicles at the frog neuromuscular
junction. J . Cell Biol. 57, 499–524 (1973).
36. Molgo, J . & Pecot-Dechavassine, M. Effects of carbonyl
cyanide m-chlorophenylhydrazone on quantal transmitter
release and ultrastructure of frog motor nerve terminals.
Neuroscience 24, 695–708 (1988).
37. Rizzoli, S. O. & Betz, W. J . The structural organization of the
readily releasable pool of synaptic vesicles. Science303,
38. Schikorski, T. & Stevens, C. F. Quantitative ultrastructural
analysis of hippocampal excitatory synapses. J . Neurosci.
17, 5858–5867 (1997).
39. Harris, K. M. & Sultan, P. Variation in the number, location
and size of synaptic vesicles provides an anatomical basis
for the nonuniform probability of release at hippocampal
CA1 synapses. Neuropharmacology34, 1387–1395 (1995).
40. Schikorski, T. & Stevens, C. F. Morphological correlates of
functionally defined synaptic vesicle populations. Nature
Neurosci. 4, 391–395 (2001).
Electron microscopy study showing that the number
of vesicles in the RRP and the number of docked
vesicles are almost identical.
41. Stevens, C. F. & Tsujimoto, T. Estimates for the pool size of
releasable quanta at a single central synapse and for the
time required to refill the pool. Proc. Natl Acad. Sci. USA 92,
42. Murthy, V. N. & Stevens, C. F. Reversal of synaptic vesicle
docking at central synapses. Nature Neurosci. 2, 503–507
43. Harata, N. et al. Limited numbers of recycling vesicles in
small CNS nerve terminals: implications for neural signaling
and vesicular cycling. Trends Neurosci. 24, 637–643 (2001).
Review of vesicle cycling in hippocampal boutons,
emphasizing the reliance of this system on the
continuous cycling of the recycling pool, without
reserve pool mobilization.
recycling pool.The reserve vesicles are different from the
recycling vesicles in terms oftheir endocytic pathway,
and possibly in terms oftheir mobility and clustering,
but do not seem to be distinguished by the position that
they occupy in the nerve terminal.
Vesicles in the reserve pool might function as a neuro-
transmitter storage depot,undergoing exocytosis only
when stimulation surpasses normal limits.Certainly,
they are retrieved at considerable energetic cost,
through surface membrane infoldings and cisternae.
The reduced mobility of reserve pool vesicles might
enable the cell to use the more efficient recycling pool
internalization route repeatedly without having to form
(and disaggregate) endocytic intermediates.The mole-
cular components that distinguish members ofthe two
pools remain to be identified,so,to date,no definitive
model ofsynaptic vesicle biochemical differentiation
can realistically be proposed.
6 8 |JANUARY 2005 |VOLUME 6
R E V I E W S
44. Aravanis, A. M., Pyle, J . L. & Tsien, R. W. Single synaptic
vesicles fusing transiently and successively without loss of
identity. Nature423, 643–647 (2003).
45. Aravanis, A. M., Pyle, J . L., Harata, N. C. & Tsien, R. W.
Imaging single synaptic vesicles undergoing repeated fusion
events: kissing, running, and kissing again.
Neuropharmacology45, 797–813 (2003).
46. Ryan, T. A., Reuter, H. & Smith, S. J . Optical detection of a
quantal presynaptic membrane turnover. Nature388,
47. Fesce, R., Grohovaz, F., Valtorta, F. & Meldolesi, J .
Neurotransmitter release: fusion or ‘kiss-and-run’? Trends
Cell Biol. 4, 1–4 (1994).
48. Stevens, C. F. & Williams, J . H. ‘Kiss and run’ exocytosis at
hippocampal synapses. Proc. Natl Acad. Sci. USA 97,
49. Gandhi, S. P. & Stevens, C. F. Three modes of synaptic
vesicular recycling revealed by single-vesicle imaging.
Nature423, 607–613 (2003).
Evidence for both full fusion and kiss-and-run
exocytosis in hippocampal boutons. Importantly, it
also indicates that proteic components of the
synaptic vesicles remain together after exocytosis
and are retrieved as a unit.
50. Satzler , K. et al. Three-dimensional reconstruction of a calyx of
Held and its postsynaptic principal neuron in the medial nucleus
of the trapezoid body. J. Neurosci.22, 10567–10579 (2002).
51. von Gersdorff, H., Schneggenburger, R., Weis, S. & Neher, E.
Presynaptic depression at a calyx synapse: the small
contribution of metabotropic glutamate receptors.
J . Neurosci. 17, 8137–8146 (1997).
52. Wang, L. Y. & Kaczmarek, L. K. High-frequency firing helps
replenish the readily releasable pool of synaptic vesicles.
Nature394, 384–388 (1998).
53. Schneggenburger, R., Sakaba, T. & Neher, E. Vesicle pools
and short-term synaptic depression: lessons from a large
synapse. Trends Neurosci. 25, 206–212 (2002).
54. Sakaba, T., Schneggenburger, R. & Neher, E. Estimation of
quantal parameters at the calyx of Held synapse. Neurosci.
Res. 44, 343–356 (2002).
55. Bollmann, J . H., Sakmann, B. & Borst, J . G. Calcium
sensitivity of glutamate release in a calyx-type terminal.
Science289, 953–957 (2000).
56. Taschenberger, H. & von Gersdorff, H. Fine-tuning an
auditory synapse for speed and fidelity: developmental
changes in presynaptic waveform, EPSC kinetics, and
synaptic plasticity. J . Neurosci. 20, 9162–9173 (2000).
57. Iwasaki, S. & Takahashi, T. Developmental regulation of
transmitter release at the calyx of Held in rat auditory
brainstem. J . Physiol. (Lond.)534, 861–871 (2001).
58. Scheuss, V., Schneggenburger, R. & Neher, E. Separation of
presynaptic and postsynaptic contributions to depression
by covariance analysis of successive EPSCs at the calyx of
Held synapse. J . Neurosci. 22, 728–739 (2002).
59. Felmy, F., Neher, E. & Schneggenburger, R. Probing the
intracellular calcium sensitivity of transmitter release during
synaptic facilitation. Neuron37, 801–811 (2003).
60. Felmy, F., Neher, E. & Schneggenburger, R. The timing of
phasic transmitter release is Ca2+-dependent and lacks a
direct influence of presynaptic membrane potential. Proc.
Natl Acad. Sci. USA 100, 15200–15205 (2003).
61. Sakaba, T. & Neher, E. Preferential potentiation of fast-
releasing synaptic vesicles by cAMP at the calyx of Held.
Proc. Natl Acad. Sci. USA 98, 331–336 (2001).
62. Sakaba, T. & Neher, E. Quantitative relationship between
transmitter release and calcium current at the calyx of Held
synapse. J . Neurosci. 21, 462–476 (2001).
63. Schneggenburger, R. & Neher, E. Intracellular calcium
dependence of transmitter release rates at a fast central
synapse. Nature406, 889–893 (2000).
64. Wolfel, M. & Schneggenburger, R. Presynaptic capacitance
measurements and Ca2+uncaging reveal submillisecond
exocytosis kinetics and characterize the Ca2+sensitivity of
vesicle pool depletion at a fast CNS synapse. J . Neurosci.
23, 7059–7068 (2003).
65. Sun, J . Y. & Wu, L. G. Fast kinetics of exocytosis revealed by
simultaneous measurements of presynaptic capacitance
and postsynaptic currents at a central synapse. Neuron30,
66. Neher, E. & Sakaba, T. Estimating transmitter release rates
from postsynaptic current fluctuations. J . Neurosci. 21,
67. Neher, E. & Sakaba, T. Combining deconvolution and noise
analysis for the estimation of transmitter release rates at the
calyx of Held. J . Neurosci. 21, 444–461 (2001).
68. Trommershauser, J ., Schneggenburger, R., Zippelius, A. &
Neher, E. Heterogeneous presynaptic release probabilities:
functional relevance for short-term plasticity. Biophys. J . 84,
69. Matthews, G. Calcium/calmodulin: a synaptic
antidepressant? Neuron20, 962–963 (2001).
70. Voets, T., Neher, E. & Moser, T. Mechanisms underlying
phasic and sustained secretion in chromaffin cells from
mouse adrenal slices. Neuron23, 607–615 (1999).
Release of RRP dense core granules in chromaffin
cells involves both fast and slow (sustained)
components, which are carried out by equally fusion-
competent vesicles. These vesicles might differ solely
in their position in respect to calcium channels. The
findings of this paper can be extended to synapses,
where the RRP might, in many cases, be only a subset
of the recycling pool — those recycling pool vesicles
that are docked.
71. von Gersdorff, H., Vardi, E., Matthews, G. & Sterling, P.
Evidence that vesicles on the synaptic ribbon of retinal
bipolar neurons can be rapidly released. Neuron16,
72. Holt, M., Cooke, A., Neef, A. & Lagnado, L. High mobility of
vesicles supports continuous exocytosis at a ribbon
synapse. Curr. Biol. 14, 173–183 (2004).
73. Cochilla, A. J ., Angleson, J . K. & Betz, W. J . Monitoring
secretory membrane with FM1-43 fluorescence. Annu. Rev.
Neurosci. 22, 1–10 (1999).
74. Lagnado, L., Gomis, A. & J ob, C. Continuous vesicle cycling
in the synaptic terminal of retinal bipolar cells. Neuron17,
75. Llobet, A., Beaumont, V. & Lagnado, L. Real-time
measurement of exocytosis and endocytosis using
interference of light. Neuron40, 1075–1086 (2003).
76. Sakaba, T., Tachibana, M., Matsui, K. & Minami, N. Two
components of transmitter release in retinal bipolar cells:
exocytosis and mobilization of synaptic vesicles. Neurosci.
Res. 27, 357–370 (1997).
77. Gomis, A., Burrone, J . & Lagnado, L. Two actions of calcium
regulate the supply of releasable vesicles at the ribbon
synapse of retinal bipolar cells. J . Neurosci. 19, 6309–6317
78. Burrone, J . & Lagnado, L. Synaptic depression and the
kinetics of exocytosis in retinal bipolar cells. J . Neurosci. 20,
79. von Gersdorff, H. & Matthews, G. Depletion and
replenishment of vesicle pools at a ribbon-type synaptic
terminal. J . Neurosci. 17, 1919–1927 (1997).
80. von Gersdorff, H. & Matthews, G. Dynamics of synaptic
vesicle fusion and membrane retrieval in synaptic terminals.
Nature367, 735–739 (1994).
Exocytosis and retrieval of readily releasable vesicles
are monitored by capacitance measurements in the
goldfish bipolar cell preparation.
81. Lenzi, D. & von Gersdorff, H. Structure suggests function:
the case for synaptic ribbons as exocytotic nanomachines.
Bioessays 23, 831–840 (2001).
82. Parsons, T. D. & Sterling, P. Synaptic ribbon. Conveyor belt
or safety belt? Neuron37, 379–382 (2003).
83. Rea, R. et al. Streamlined synaptic vesicle cycle in cone
photoreceptor terminals. Neuron41, 755–766 (2004).
The authors investigate the mobility of synaptic
vesicles in the (synapsin-free) photoreceptor terminal.
Vesicles move freely by diffusion throughout the
terminal, and vesicle movement is not rate limiting for
84. Propst, J . W. & Ko, C. P. Correlations between active zone
ultrastructure and synaptic function studied with freeze-
fracture of physiologically identified neuromuscular junctions.
J . Neurosci. 7, 3654–3664 (1987).
Shows a linear relationship between evoked quantal
content (but not spontaneous release frequency) and
active zone length in frog motor nerve terminals.
85. Meyer, A. C., Neher, E. & Schneggenburger, R. Estimation of
quantal size and number of functional active zones at the
calyx of Held synapse by nonstationary EPSC variance
analysis. J . Neurosci. 21, 7889–7900 (2001).
86. Heuser, J . E. et al. Synaptic vesicle exocytosis captured by
quick freezing and correlated with quantal transmitter
release. J . Cell Biol. 81, 275–300 (1979).
By use of elegant rapid-freezing methods, frog motor
nerve terminal ultrastructure within ~2 ms of electrical
stimulation is investigated, providing direct
morphological evidence for the location of synaptic
87. Ceccarelli, B. & Hurlbut, W. P. Vesicle hypothesis of the
release of quanta of acetylcholine. Physiol. Rev. 60,
88. Neher, E. Vesicle pools and Ca2+microdomains: new tools
for understanding their roles in neurotransmitter release.
Neuron20, 389–399 (1998).
89. J ahn, R., Lang, T. & Südhof, T. C. Membrane fusion. Cell
112, 519–533 (2003).
90. Harlow, M. L., Ress, D., Stoschek, A., Marshall, R. M. &
McMahan, U. J . The architecture of active zone material at
the frog’s neuromuscular junction. Nature409, 479–484
91. Catterall, W. A. Interactions of presynaptic Ca2+channels
and SNARE proteins in neurotransmitter release. Ann. NY
Acad. Sci. 868, 144–159 (1999).
92. Xu-Friedman, M. A., Harris, K. M. & Regehr, W. G. Three-
dimensional comparison of ultrastructural characteristics at
depressing and facilitating synapses onto cerebellar Purkinje
cells. J . Neurosci. 21, 6666–6672 (2001).
93. Pieribone, V. A. et al. Distinct pools of synaptic vesicles in
neurotransmitter release. Nature375, 493–497 (1995).
94. Henkel, A. W., Simpson, L. L., Ridge, R. M. & Betz, W. J .
Synaptic vesicle movements monitored by fluorescence
recovery after photobleaching in nerve terminals stained
with FM1-43. J . Neurosci. 16, 3960–3967 (1996).
95. Kraszewski, K., Daniell, L., Mundigl, O. & De Camilli, P.
Mobility of synaptic vesicles in nerve endings monitored by
recovery from photobleaching of synaptic vesicle-
associated fluorescence. J . Neurosci. 16, 5905–5913
96. Hilfiker, S. et al. Synapsins as regulators of neurotransmitter
release. Phil. Trans. R. Soc. Lond. B 354, 269–279 (1999).
97. Benfenati, F. et al. Synaptic vesicle-associated
Ca2+/calmodulin-dependent protein kinase II is a binding
protein for synapsin I. Nature359, 417–420 (1992).
98. Torri Tarelli, F., Bossi, M., Fesce, R., Greengard, P. & Valtorta, F.
Synapsin I partially dissociates from synaptic vesicles during
exocytosis induced by electrical stimulation. Neuron9,
99. Li, L. et al. Impairment of synaptic vesicle clustering and of
synaptic transmission, and increased seizure propensity, in
synapsin I-deficient mice. Proc. Natl Acad. Sci. USA 92,
100. Takei, Y. et al. Synapsin I deficiency results in the structural
change in the presynaptic terminals in the murine nervous
system. J . Cell Biol. 131, 1789–1800 (1995).
101. Chi, P., Greengard, P. & Ryan, T. A. Synapsin dispersion and
reclustering during synaptic activity. Nature Neurosci. 4,
102. Rosahl, T. W. et al. Essential functions of synapsins I and II in
synaptic vesicle regulation. Nature375, 488–493 (1995).
The mouse synapsin I/II double knockouts have severe
synaptic depression on tetanic stimulation; both the
recycling and reserve pools might be affected.
103. Ryan, T. A., Li, L., Chin, L. S., Greengard, P. & Smith, S. J .
Synaptic vesicle recycling in synapsin I knock-out mice.
J . Cell Biol. 134, 1219–1227 (1996).
104. Mozhayeva, M. G., Sara, Y., Liu, X. & Kavalali, E. T.
Development of vesicle pools during maturation of
hippocampal synapses. J . Neurosci. 22, 654–665 (2002).
105. Chi, P., Greengard, P. & Ryan, T. A. Synaptic vesicle
mobilization is regulated by distinct synapsin I
phosphorylation pathways at different frequencies. Neuron
38, 69–78 (2003).
106. Hinds, H. L., Goussakov, I., Nakazawa, K., Tonegawa, S. &
Bolshakov, V. Y. Essential function of α-calcium/calmodulin-
dependent protein kinase II in neurotransmitter release at a
glutamatergic central synapse. Proc. Natl Acad. Sci. USA
100, 4275–4280 (2003).
107. Hosaka, M. & Südhof, T. C. Synapsins I and II are ATP-
binding proteins with differential Ca2+regulation. J . Biol.
Chem. 273, 1425–1429 (1998).
108. Hosaka, M. & Südhof, T. C. Synapsin III, a novel synapsin
with an unusual regulation by Ca2+. J . Biol. Chem. 273,
109. Godenschwege, T. A. et al. Flies lacking all synapsins are
unexpectedly healthy but are impaired in complex behaviour.
Eur. J . Neurosci. 20, 611–622 (2004).
110. Feng, J . et al. Regulation of neurotransmitter release by
synapsin III. J . Neurosci. 22, 4372–4380 (2002).
111. Usukura, J . & Yamada, E. Ultrastructure of the synaptic
ribbons in photoreceptor cells of Rana catesbeiana revealed
by freeze-etching and freeze-substitution. Cell Tissue Res.
247, 483–488 (1987).
112. Richards, D. A., Rizzoli, S. O. & Betz, W. J . Effects of
wortmannin and latrunculin A on slow endocytosis at the
frog neuromuscular junction. J . Physiol. (Lond.)557, 77–91
113. Sankaranarayanan, S., Atluri, P. P. & Ryan, T. A. Actin has a
molecular scaffolding, not propulsive, role in presynaptic
function. Nature Neurosci. 6, 127–135 (2003).
114. J ob, C. & Lagnado, L. Calcium and protein kinase C
regulate the actin cytoskeleton in the synaptic terminal of
retinal bipolar cells. J . Cell Biol. 143, 1661–1672 (1998).
115. Cole, J . C., Villa, B. R. & Wilkinson, R. S. Disruption of actin
impedes transmitter release in snake motor terminals.
J . Physiol. (Lond.) 525, 579–586 (2000).
116. Shupliakov, O. et al. Impaired recycling of synaptic vesicles
after acute perturbation of the presynaptic actin cytoskeleton.
Proc. Natl Acad. Sci. USA 99, 14476–14481 (2002).
117. Dunaevsky, A. & Connor, E. A. F-actin is concentrated in
nonrelease domains at frog neuromuscular junctions.
J . Neurosci. 20, 6007–6012 (2000).
NATURE REVIEWS |NEUROSCIENCE
VOLUME 6 |JANUARY 2005 |6 9
R E V I E W S
Light microscopy study revealing the delicate pattern
of actin filaments surrounding vesicle clusters in frog
motor nerve terminals.
118. Hirokawa, N., Sobue, K., Kanda, K., Harada, A. & Yorifuji, H.
The cytoskeletal architecture of the presynaptic terminal and
molecular structure of synapsin 1. J . Cell Biol. 108, 111–126
Ultrastructural (freeze-fracture, deep-etch) study of
vesicle clusters shows different types of filaments,
some of which interlink vesicles, some of which
attach vesicles to actin, and some of which link
vesicles to the surface membrane.
119. Merrifield, C. J . et al. Endocytic vesicles move at the tips of
actin tails in cultured mast cells. Nature Cell Biol. 1, 72–74
120. Rozelle, A. L. et al. Phosphatidylinositol 4,5-bisphosphate
induces actin-based movement of raft-enriched vesicles
through WASP-Arp2/3. Curr. Biol. 10, 311–320 (2000).
121. Bloom, O. et al. Colocalization of synapsin and actin during
synaptic vesicle recycling. J . Cell Biol. 161, 737–747 (2003).
122. Merrifield, C. J ., Feldman, M. E., Wan, L. & Almers, W.
Imaging actin and dynamin recruitment during invagination
of single clathrin-coated pits. Nature Cell Biol. 4, 691–698
123. Sakaba, T. & Neher, E. Involvement of actin polymerization in
vesicle recruitment at the calyx of Held synapse.
J . Neurosci. 23, 837–846 (2003).
124. Oheim, M. & Stuhmer, W. Tracking chromaffin granules on
their way through the actin cortex. Eur. Biophys.
J . 29, 67–89 (2000).
125. Holt, M., Cooke, A., Wu, M. M. & Lagnado, L. Bulk
membrane retrieval in the synaptic terminal of retinal bipolar
cells. J . Neurosci. 23, 1329–1339 (2003).
126. Morales, M., Colicos, M. A. & Goda, Y. Actin-dependent
regulation of neurotransmitter release at central synapses.
Neuron27, 539–550 (2000).
127. Zhang, W. & Benson, D. L. Stages of synapse development
defined by dependence on F-actin. J . Neurosci. 21,
128. Ryan, T. A. Inhibitors of myosin light chain kinase block
synaptic vesicle pool mobilization during action potential
firing. J . Neurosci. 19, 1317–1323 (1999).
129. Paillart, C., Li, J ., Matthews, G. & Sterling, P. Endocytosis
and vesicle recycling at a ribbon synapse. J . Neurosci. 23,
130. Pyle, J . L., Kavalali, E. T., Piedras-Renteria, E. S. & Tsien, R. W.
Rapid reuse of readily releasable pool vesicles at
hippocampal synapses. Neuron28, 221–231 (2000).
RRP vesicles can recycle selectively back to the RRP
in hippocampal nerve terminals. However, mixing with
the recycling pool is rapid.
131. Wilkinson, R. S. & Cole, J . C. Resolving the Heuser-
Ceccarelli debate. Trends Neurosci. 24, 195–197 (2001).
132. Rizzoli, S. O., Richards, D. A. & Betz, W. J . Monitoring
synaptic vesicle recycling in frog motor nerve terminals with
FM dyes. J . Neurocytol. 32, 539–549 (2003).
133. Woodman, P. G. Biogenesis of the sorting endosome: the
role of Rab5. Traffic 1, 695–701 (2000).
134. Fischer von Mollard et al. Localization of Rab5 to synaptic
vesicles identifies endosomal intermediate in synaptic
vesicle recycling pathway. Eur. J . Cell Biol. 65, 319–326
135. de Hoop, M. J . et al. The involvement of the small GTP-
binding protein Rab5a in neuronal endocytosis. Neuron13,
136. Wucherpfennig, T., Wilsch-Brauninger, M. & Gonzalez-
Gaitan, M. Role of Drosophila Rab5 during endosomal
trafficking at the synapse and evoked neurotransmitter
release. J . Cell Biol. 161, 609–624 (2003).
137. Antonin, W., Riedel, D. & von Mollard, G. F. The SNARE
Vti1a-β is localized to small synaptic vesicles and
participates in a novel SNARE complex. J . Neurosci. 20,
138. Koenig, J . H. & Ikeda, K. Synaptic vesicles have two
distinct recycling pathways. J . Cell Biol. 135, 797–808
Describes two recycling pathways in Drosophila adult
nerve terminals, which might be used by two distinct
139. Koenig, J . H. & Ikeda, K. Contribution of active zone
subpopulation of vesicles to evoked and spontaneous
release. J . Neurophysiol. 81, 1495–1505 (1999).
140. Koenig, J . H., Yamaoka, K. & Ikeda, K. Omega images at
the active zone may be endocytotic rather than exocytotic:
implications for the vesicle hypothesis of transmitter
release. Proc. Natl Acad. Sci. USA 95, 12677–12682
141. Ryan, T. A. et al. The kinetics of synaptic vesicle recycling
measured at single presynaptic boutons. Neuron11,
142. Ryan, T. A. & Smith, S. J . Vesicle pool mobilization during
action potential firing at hippocampal synapses. Neuron14,
143. Ryan, T. A., Smith, S. J . & Reuter, H. The timing of synaptic
vesicle endocytosis. Proc. Natl Acad. Sci. USA 93,
144. Sankaranarayanan, S. & Ryan, T. A. Real-time
measurements of vesicle-SNARE recycling in synapses of
the central nervous system. Nature Cell Biol. 2, 197–204
145. Murthy, V. N. & Stevens, C. F. Synaptic vesicles retain their
identity through the endocytic cycle. Nature392, 497–501
146. Gennaro, J . F., Nastuk, W. L. & Rutherford, D. T. Reversible
depletion of synaptic vesicles induced by application of high
external potassium to the frog neuromuscular junction.
J . Physiol. (Lond.)280, 237–247 (1978).
147. Sara, Y., Mozhayeva, M. G., Liu, X. & Kavalali, E. T. Fast
vesicle recycling supports neurotransmission during
sustained stimulation at hippocampal synapses.
J . Neurosci. 22, 1608–1617 (2002).
148. Fernandez-Alfonso, T. & Ryan, T. A. The kinetics of synaptic
vesicle pool depletion at CNS synaptic terminals. Neuron
41, 943–953 (2004).
149. Pyott, S. J . & Rosenmund, C. The effects of temperature on
vesicular supply and release in autaptic cultures of rat and
mouse hippocampal neurons. J . Physiol. (Lond.)539,
150. Lenzi, D., Crum, J ., Ellisman, M. H. & Roberts, W. M.
Depolarization redistributes synaptic membrane and creates
a gradient of vesicles on the synaptic body at a ribbon
synapse. Neuron36, 649–659 (2002).
151. Zhang, J . Z., Davletov, B. A., Südhof, T. C. & Anderson, R. G.
Synaptotagmin I is a high affinity receptor for clathrin AP-2:
implications for membrane recycling. Cell78, 751–760
152. J arousse, N. & Kelly, R. B. The AP2 binding site of
synaptotagmin 1 is not an internalization signal but a
regulator of endocytosis. J . Cell Biol. 154, 857–866
153. Poskanzer , K. E., Marek, K. W., Sweeney, S. T. & Davis, G. W.
Synaptotagmin I is necessary for compensatory synaptic
vesicle endocytosis in vivo. Nature 426, 559–563
154. Zenisek, D., Steyer, J . A., Feldman, M. E. & Almers, W.
A membrane marker leaves synaptic vesicles in milliseconds
after exocytosis in retinal bipolar cells. Neuron35,
155. Valtorta, F., Meldolesi, J . & Fesce, R. Synaptic vesicles: is
kissing a matter of competence? Trends Cell Biol. 11,
156. Klingauf, J ., Kavalali, E. T. & Tsien, R. W. Kinetics and
regulation of fast endocytosis at hippocampal synapses.
Nature394, 581–585 (1998).
157. Sun, J . Y., Wu, X. S. & Wu, L. G. Single and multiple vesicle
fusion induce different rates of endocytosis at a central
synapse. Nature417, 555–559 (2002).
Dual presynaptic (capacitance) and postsynaptic
(current) patch clamp of the calyx of Held shows that
endocytosis occurs extremely rapidly (50 ms) after
single quantal release events, and slows
progressively with more release.
158. Klyachko, V. A. & J ackson, M. B. Capacitance steps and
fusion pores of small and large-dense-core vesicles in nerve
terminals. Nature418, 89–92 (2002).
159. Verstreken, P. et al. Endophilin mutations block clathrin-
mediated endocytosis but not neurotransmitter release.
Cell109, 101–112 (2002).
160. Kawasaki, F., Hazen, M. & Ordway, R. W. Fast synaptic
fatigue in shibiremutants reveals a rapid requirement for
dynamin in synaptic vesicle membrane trafficking. Nature
Neurosci. 3, 859–860 (2000).
161. Teng, H., Cole, J . C., Roberts, R. L. & Wilkinson R. S.
Endocytic active zones: hot spots for endocytosis in
vertebrate neuromuscular terminals. J . Neurosci. 19,
162. Teng, H. & Wilkinson, R. S. Clathrin-mediated endocytosis
near active zones in snake motor boutons. J . Neurosci. 20,
163. Wang, C. & Zucker, R. S. Regulation of synaptic vesicle
recycling by calcium and serotonin. Neuron21, 155–167
164. Humeau, Y. et al. Synapsin controls both reserve and
releasable synaptic vesicle pools during neuronal activity
and short-term plasticity in Aplysia. J . Neurosci. 21,
165. Hilfiker, S. et al. Two sites of action for synapsin domain E in
regulating neurotransmitter release. Nature Neurosci. 1,
166. Murthy, V. N., Schikorski, T., Stevens, C. F. & Zhu, Y.
Inactivity produces increases in neurotransmitter release and
synapse size. Neuron32, 673–682 (2001).
We thank H. Kuromi (Maebashi), L. Lagnado (Cambridge), E. Neher
(Göttingen) and R. Tsien (Palo Alto) for their helpful comments on the
manuscript. W.J.B. is supported by research grants from the National
Institutes of Health and the Muscular Dystrophy Association.
Competing interests statement
The authors declare no competing financial interests.
The following terms in this article are linked online to:
Encyclopedia of Life Sciences: http://www.els.net/
Synaptic vesicle traffic
Access to this interactive links box is free online.