19:262-270, 2004.  doi:10.1152/physiol.00014.2004 PhysiologyR. Grace Zhai and Hugo J. Bellen

You might find this additional information useful...

104 articles, 34 of which you can access free at: This article cites http://physiologyonline.physiology.org/cgi/content/full/19/5/262#BIBL

6 other HighWire hosted articles, the first 5 are: This article has been cited by

  [PDF]  [Full Text]  [Abstract]

, August 1, 2007; 582 (3): 1337-1348. J. Physiol.V. A. Tobin and M. Ludwig

The role of the actin cytoskeleton in oxytocin and vasopressin release from rat supraoptic nucleus neurons 

[PDF]  [Full Text]  [Abstract], August 1, 2007; 27 (31): 8457-8474. J. Neurosci.

P. Heil, H. Neubauer, D. R. F. Irvine and M. Brown Spontaneous Activity of Auditory-Nerve Fibers: Insights into Stochastic Processes at Ribbon Synapses

  [PDF]  [Full Text]  [Abstract]

, April 16, 2008; 28 (16): 4151-4160. J. Neurosci.G. A. Zampighi, N. Fain, L. M. Zampighi, F. Cantele, S. Lanzavecchia and E. M. Wright

Conical Electron Tomography of a Chemical Synapse: Polyhedral Cages Dock Vesicles to the Active Zone 

[PDF]  [Abstract], August 1, 2009; 15 (4): 380-391. Neuroscientist

L. LoGiudice and G. Matthews The Role of Ribbons at Sensory Synapses

  [PDF]  [Full Text]  [Abstract]

, September 16, 2009; 29 (37): 11484-11494. J. Neurosci.N. M. Viquez, P. Fuger, V. Valakh, R. W. Daniels, T. M. Rasse and A. DiAntonio

PP2A and GSK-3{beta} Act Antagonistically to Regulate Active Zone Development

on the following topics: http://highwire.stanford.edu/lists/artbytopic.dtlcan be found at Medline items on this article's topics

Physiology .. Nerves Neuroscience .. Neurotransmitters Veterinary Science .. Synaptic Vesicle

including high-resolution figures, can be found at: Updated information and services http://physiologyonline.physiology.org/cgi/content/full/19/5/262

can be found at: Physiologyabout Additional material and information http://www.the-aps.org/publications/physiol

This information is current as of November 6, 2009 .  

http://www.the-aps.org/.Copyright © 2005 by the American Physiological Society. ISSN: 1548-9213, ESSN: 1548-9221. Visit our website at 20814-3991.bimonthly in February, April, June, August, October, and December by the American Physiological Society, 9650 Rockville Pike, Bethesda MD

) publishes brief review articles on major physiological developments. It is publishedNews in Physiological Science (formerly published as Physiology

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

262 1548-9213/04 5.00 ©2004 Int. Union Physiol. Sci./Am. Physiol. Soc.

REVIEWSPHYSIOLOGY 19: 262–270, 2004; 10.1152/physiol.00014.2004

Synapses are specialized cell-cell contacts wheresignals are reliably transduced from a neuron to itstarget cell in a regulated manner. In chemicalsynapses, signal transduction is achieved byconverting the electrical signal into a chemicalsignal that diffuses between cells. This signalconversion occurs at active zones, highlyspecialized sites of the presynaptic nerve terminal.The term “active zone” was coined in 1970 byCouteaux and Pecot-Dechavassine (14) duringtheir ultrastructural studies of partially contractedfrog muscles, in which they observed that profilesof open synaptic vesicles occurred immediatelyadjacent to the presynaptic dense bands, andconsequently they designated these dense bands“les zones actives.” Subsequently, similar obser-vations were made in other types of synapses.

Ultrastructural studies of synapses in differentorganisms have revealed the following conservedmorphological features among active zones,regardless of their size, location, or types of neu-rons and their targets. First, the plasma membraneof the active zone appears to be electron dense,suggesting its proteinaceous nature. Second,synaptic vesicles cluster, tether, and fuse at theactive zone (14, 36). Third, the active zone is close-ly and precisely aligned with a postsynaptic densi-ty (PSD) area, such that the active zone spans thesame width as the PSD and the extracellular spacebetween the two membranes (synaptic cleft) is asnarrow as 30 nm (51). The latter two morphologicalcharacteristics suggest that active zones functionas sites of synaptic vesicle exocytosis and neuro-transmitter release.

Over the past three decades, these hypotheseshave been supported by many neuroscientiststhrough studies of synaptic vesicle exocytosis andpostsynaptic neurotransmitter receptor function.Here we discuss the structural organization ofactive zones in different types of synapses found ina variety of organisms, summarize recent advancesin the molecular characterization of the assembly

of active zones, and propose that all active zonesare organized according to the same pattern.

The Structure of Active Zones

Morphologically, active zones are defined as sites ofsynaptic vesicle docking and fusion, andphysiologically they are defined as sites ofneurotransmitter release. On the basis of thesedefinitions, we dissect the active zone into three morphologically and functionally distinctcomponents:

1. the plasma membrane juxtaposed to the PSDwhere synaptic vesicle fusion occurs,

2. the cytomatrix immediately internal to theplasma membrane where synaptic vesiclesdock, and

3. the electron-dense projections extending fromthe cytomatrix into the cytoplasm on whichsynaptic vesicles are tethered.

All active zones have these three components,although they vary in appearance, especially in sizeand shape of dense projections. FIGURE 1illustrates the ultrastructure of active zones foundin nine different types of synapse as well as theschematic representation of the three componentsin each type of active zone. Below we will discussthe molecular and functional properties of eachcomponent of the active zone.

The Plasma Membrane of the ActiveZone

Besides separating the cytosol from theextracellular environment, the plasma membraneof the active zone has two “gates” essential forneurotransmission: one for Ca2+ entry, which is thevoltage-gated Ca2+ channel, and the other forneurotransmitter exit, which is the synaptic vesiclefusion site. These two gates are thought to be in

The Architecture of the Active Zone inthe Presynaptic Nerve Terminal

R. Grace Zhai and Hugo J. BellenHoward Hughes Medical Institute and Department of Molecular

and Human Genetics, Division of Neuroscience, Program InDevelopmental Biology, Baylor College of Medicine,

Houston, Texas 77030rzhai@bcm.tmc.eduActive zones are highly specialized sites for release of neurotransmitter from

presynaptic nerve terminals. The architecture of the active zone is exquisitely

designed to facilitate the regulated tethering, docking, and fusing of the

synaptic vesicles with the plasma membrane. Here we present our view of the

structural and molecular organization of active zones across species and propose

that all active zones are organized according to a common principle in which the

structural differences correlate with the kinetics of transmitter release.

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

A B C. elegans

C D Crayfish

E F Drosophila

H I Skate J Rat

K L Frog

M N Lizard

O P Human

Thin sectionThin section Freeze fractureFreeze fracture

Photoreceptor

Laminarmonopolarcells

G Drosophila

Horizontal cells

Rod bipolar cell

300 nm 300 nm

300 nm 300 nm

300 nm 300 nm 300 nm 300 nm

300 nm 300 nm

300 nm 300 nm

50 nm 50 nm

300 nm 300 nm

300 nm 300 nm

263PHYSIOLOGY • Volume 19 • October 2004 • www.physiologyonline.org

REVIEWS

fdfeo t

D

e

mh

e

de

h

eereren

FIGURE 1. Electron micrographs and schematicrepresentations of the active zone structuresfound in different synapses of various organismsA, C, E, H, K, M, and O are diagrams of the active zonestructure in the synaptic terminal electron microgrpahsshown to their right. B: a neuromuscular junction (NMJ)terminal in Caenorhabditis elegans (31) with a the plaque-like active zone projection (arrow). D: an NMJ terminal incrayfish (30) with a dense projection (blue arrow). F: anNMJ terminal in Drosophila (62) with a dense projectioncalled a T bar (blue arrow). G: a tetrad synapse betweenphotoreceptor and lamina monopolar cells in Drosophila(61). The T bar consists of a platform (double arrows) anda pedestal (single arrow). I: an electroreceptor in skate(55) with a long ribbonlike dense projection (blue arrow)and a halo of synaptic vesicles (yellow arrow). J: a triadicphotoreceptor ribbon synapse between rod photorecep-tor and horizontal cell and a rod bipolar cell in rat (16). L:a saccular hair cell in frog (55) with a spherical dense pro-jection (blue arrow) and attached vesicles (yellow arrow).N: a thin-section image (left) and a freeze-fracture image(right) of an NMJ terminal in lizard (92). The dense projec-tion is marked by blue arrows in both images, and theintramembraneous particles are marked by green arrows.P: an excitatory synaptic terminal in human hippocampus(18) with 2 active zones (red arrows).

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

264 PHYSIOLOGY • Volume 19 • October 2004 • www.physiologyonline.org

close proximity to each other on the basis of severalobservations. First, the time delay between Ca2+

entry and synaptic vesicle fusion is only 0.2 ms (69,87). Second, theoretical analyses of Ca2+ diffusiondynamics and quantal secretion have shown thatthe probability of secretion of a synaptic vesicledecreases threefold with a doubling of the distancebetween the Ca2+ channel and the synaptic vesiclefrom 25 to 50 nm (4). It is thus likely that in mostsynapses the space between Ca2+ channels anddocked synaptic vesicles at the active zone is <50nm (3, 87). Numerous immunohistochemicalstudies have demonstrated the localization of Ca2+

channels near the active zone (46, 75, 104);however, the organization and arrangement of Ca2+

channels at active zones are best suggested byfreeze-fracture studies on frog, lizard, andmammalian neuromuscular junctions (NMJs) (19,37, 92). On freeze-fracture replicas of NMJ boutons,double rows of prominent intramembranousparticles of 10–12 nm in diameter can be seen withoccasional synaptic vesicles clustering closelybeside them (FIGURES 1N AND 2C). These rows ofparticles are in exact register with the postsynapticfolds in the underlying muscle (37) and arepostulated to be voltage-gated Ca2+ channels (13,72, 75). The most direct evidence supporting thisview comes from atomic force microscopy on thelarge calyciform synapse of the chick ciliaryganglion, showing a row-like arrangement of theimmunolabeled Ca2+ channels (35).

If the calcium channels are organized in orderlyarrays at active zones, one would expect the othergate (the synaptic vesicle fusion site) to be organ-ized in a corresponding manner. The vesicle fusionsites are specialized to allow the lipid bilayers ofsynaptic vesicles and active zone plasma mem-brane to come together and form a hydrophilicfusion pore. The SNARE complex has been thoughtto be the driving force for bringing the membranestogether, facilitating lipid bilayer mixing and sub-sequent membrane fusion (43, 74). The t-SNAREssyntaxin and SNAP-25 have been localized to thepresynaptic plasma membrane, although their dis-tribution is not restricted to the active zone and theprecise positioning and organization of t-SNAREsat the active zone is not clear (25, 38, 82).Biochemical analyses have demonstrated thedirect interaction between the t-SNAREs syntaxinand SNAP-25 and Ca2+ channels (11, 44, 57, 90).This may be the possible mechanism for the closepositioning of the fusion machinery and Ca2+ chan-nels required for Ca2+-dependent exocytosis.Interestingly, a study in PC-12 cells, a neurosecre-tory cell line, has shown that syntaxin and SNAP-25form cholesterol-dependent clusters in the plasmamembrane, and these clusters seem to define sitesat which secretory vesicles dock and fuse with high

REVIEWS

A

B

E

preference. Cholesterol removal causes dispersionof clusters and the subsequent inhibition of exocy-tosis (52). This study provides a clue for the poten-tial role of lipids in active zone plasma membranespecification.

Other important components of the plasmamembrane are the adhesion molecules by whichthe precise alignment of the active zone with thePSD is most likely mediated. Several classes ofadhesion molecules have been shown to be presentat the active zone: cadherins (85, 99), protocad-herins (23), nectins (64, 88), neural cell adhesionmolecule (77)/fasciclin II (15)/aplysia cell adhesionmolecule (60), Down syndrome adhesion molecule(80), syndecans (42), L1/neuroglian (93), integrins(12), neurexins (63), and sidekicks (100, 101). Alladhesion molecules share common protein motifs:an extracellular domain that mediates binding withthe postsynaptic counterparts or extracellularmatrix, a single-pass transmembrane domain ormembrane anchor, and often an intracellulardomain that binds to the cytoskeleton or the intra-cellular scaffolding proteins (28, 86). All of theseadhesion molecules except neurexin, which isexpressed presynaptically and binds its postsynap-tic receptor neuroligin (100), are expressed in bothpre- and postsynaptic terminals, and adhesion isformed through homophilic interactions. Adhesionmolecules at the active zone are more than just“glue.” They also mediate signaling within andbetween nerve terminals and modulate neuro-transmission. Numerous reviews have commentedon the detailed mechanisms of these molecules insynapse adhesion and regulation (22, 68, 79, 88).

In summary, the primary function of the plasmamembrane at the active zone during neurotrans-mitter release is to mediate fusion of synaptic vesi-cles upon calcium entry. This is achieved by anarray-like organization of the Ca2+ channels and thelocalization of the fusion machinery at the mem-brane.

The Cytomatrix Underlying thePlasma Membrane of the ActiveZone

When viewed by electron microscopy, thecytomatrix of the active zone is electron dense anddisplays a web-like pattern, which was first noticedby Bloom and Aghajanian (5) and subsequently byPfenninger and colleagues (70). Recently, elegantstudies have provided the first three-dimensionalviews of the cytomatrix at the active zone of thefrog NMJ and the mammalian central nervoussystem (CNS) synapse. Using electron microscopetomography, Harlow and colleagues (32) revealed astriking array-like structure at the frog NMJconsisting of “beams” and “ribs” that connect

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

docked synaptic vesicles with putative Ca2+

channels (FIGURE 2, A–D). The beams run alongthe midline of the presynaptic ridge parallel to theridge’s long axis, and the ribs extend from thebeams and connect the synaptic vesicles near thevesicle-plasma membrane interface. In addition,the ribs are connected to the intramembranemacromolecules resembling the putative Ca2+

channels seen in freeze-fracture studies (see above,

265PHYSIOLOGY • Volume 19 • October 2004 • www.physiologyonline.org

REVIEWSFIGURES 1N, 2C, AND 2D). With this organization,each docked vesicle is perfectly aligned with atleast one Ca2+ channel, which would allow highrelease fidelity.

The picture of the mammalian CNS synapse wasrevealed by Phillips and colleagues (71). In thisstudy, they purified a presynaptic “particle web”consisting of ~50 nm pyramidally shaped particlesinterconnected by ~50- to 100-nm-spaced fibrils

A Frog NMJ terminal C Replica from freeze fractured frog NMJ terminal

B 3-D reconstruction of frog NMJ terminal

E Mammalian active zone ( side view) F Mammalian active zone ( top view) G Mammalian central synapse

D Schematic of frog NMJ terminal

50 m 50 m

Synapticvesicle

Synapticvesicle

Ribs

Pegs

“Particle web”~50nm

”particles”

Thin fibrils Adhesion proteins

Synaptic vesiclesBeam

FIGURE 2. Models of the active zone structure in frog NMJ and mammalian central nervous system synapsesA–D: models of the active zone structure in frog NMJs (32). A: transmission electron micrograph of frog NMJ, where synaptic vesicles are docked tothe plasma membrane through dense projections. B: 3-dimensional reconstructed and surface-rendered view of the active zone material and dockedvesicles. C: replica from a freeze-fractured frog NMJ showing a series of particles/macromolecules on each slope of a ridge with surface-rendered riband beam assemblies. D: arrangement of ribs, beams, and synaptic vesicles at frog NMJs. E–G: models of the active zone structure in mammaliancentral nervous system synapses (71). E–F: transmission electron micrographs of purified active zones from mammalian central nervous system inside view (E) and top view (F). Thin fibrils can be seen (arrow in F). G: active zone structure of mammalian central synapses.

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

266 PHYSIOLOGY • Volume 19 • October 2004 • www.physiologyonline.org

REVIEWScharacterized in vertebrate sensory synapsesinvolved in vision, hearing, and balance (50, 55,91). These dense projections, or synaptic ribbons,are ribbon-like or spherical, extend 0.5-1 m intothe cytoplasm, and always have a “halo” of synapticvesicles tethered to their surface [FIGURE 1, H–L(55)]. Because of their remarkable appearance, ithas been thought that ribbon synapses aredifferent from all other synapses, and the synapticribbons are exclusive to ribbon synapses tomediate the graded and sustained neuro-transmitter release of these synapses (55, 91).However, we propose that electron-denseprojections are not unique to ribbon synapses butrather are an integral part of the active zone andhave an evolutionarily conserved structure. Theyhave different sizes and shapes in different types ofsynapses but perform the same function oftethering synaptic vesicles at active zones.

Morphologically, dense projections have beenobserved in different types of synapses in differentspecies. At Caenorhabditis elegans NMJs, denseprojections in the shape of a plaque have beendescribed [FIGURE 1B (31)]. In Drosophila, T-shaped dense projections can be seen in NMJs, thetetrad synapses of the visual system (FIGURE 1, FAND G), and in CNS synapses (61, 102). In crus-tacean NMJs, dense projections appear to be cylin-drical [FIGURE 1D (29)]. In vertebrate NMJs, denseprojections have been described in frog, lizard, andmammals [FIGURE 1, N AND P (19, 36, 92)]. Inmammalian CNS synapses, dense projections werealso noticed in electron microscopic studies asearly as the 1960s and recently have been purifiedand visualized in great detail [FIGURE 2, E–F (5,71)]. On the basis of the size of dense projections,we can divide different types of active zones intotwo groups: those with prominent dense projec-tions, including invertebrate synapses with T barsand vertebrate ribbon synapses, and those withless-prominent dense projections, including verte-brate NMJs and CNS synapses. The dense projec-tions in these latter active zones are not veryprominent and project <100 nm into the cytoplasmand therefore have generally been considered to bepart of the cytomatrix at the active zone (18, 26, 51).

Physiologically, numerous studies have demon-strated the tethering function of dense projections.In vertebrate sensory synapses, the motor proteinKIF3A, which is a component of ribbons, is likely tomediate the tethering of synaptic vesicles (66). Bytomography study, the synaptic vesicles in frogNMJs are clearly shown to be tethered through ribsto the beams, corresponding to the dense projec-tions (32). In Drosophila and crayfish, synapticvesicles cluster around T bars, although the mech-anism of tethering is not known (FIGURE 1, A–G).Recently the dense projections of mammalian CNS

(FIGURE 2, E–F). The particles are evenly spaced bythe fibrils, and this web therefore forms 50- to 100-nm slots for synaptic vesicles to dock and fuse(FIGURE 2G).

The electron-dense nature of the cytomatrixunderlying the plasma membrane suggests that alarge number of proteins are localized there andthat the cytomatrix at the active zone is importantin regulating vesicle docking and fusion. In search-ing for the building blocks of this specializedcytomatrix, a number of protein components havebeen identified. On the basis of their function orputative function, the proteins identified in theactive zone cytomatrix can be classified into threecategories. First, the classical cytoskeletal proteinscorresponding to actin, tubulin, myosin, spectrin�-chain and �-chain, and �-catenin (9, 40, 71) arethe fundamental elements of the framework ofactive zone cytomatrix. Second are the known scaf-folding proteins, including SAP90/PSD95/Dlg,SAP97, and CASK/LIN-2 (34, 47, 48, 65). These pro-teins are not restricted to active zones because theyalso participate in clustering of postsynaptic recep-tors and are involved in the organization of a vari-ety of cell junctions (20, 27, 67). If the cytoskeletonproteins form a grid-like structure at the activezone, these scaffolding proteins probably link theion channels and the fusion machinery onto thegrid to ensure proper active zone function. Forexample, CASK interacts with �-neurexin, synde-can 2, Ca2+ channels, the cytosolic protein Veli/LIN-7, and the Munc18/n-Sec1-interacting proteinMint1 (10, 34, 41, 59). Third, there are the activezone-specific proteins including RIM1, Munc13/unc13, Bassoon, Piccolo/Aczonin, and CAST/ERCs(8, 17, 21, 94, 95, 96). Their active zone-specificlocalization and their multidomain structure allowthem to participate in modulating synaptic vesicledocking, priming, and fusion, as well as the initia-tion of the assembly of the active zone structure.Physiological studies indicate that some of theseproteins are involved in vesicle priming as well assynaptic transmission regulation (18, 76, 89).

In summary, the primary function of the cytoma-trix at the active zone is to mediate docking ofsynaptic vesicles. The cytoskeletal and scaffoldingproteins form a web-like structure consisting ofslots for synaptic vesicle docking, and componentsof the cytomatrix regulate vesicle priming andfusion.

The Electron-Dense ProjectionsExtending from the Cytomatrix ofthe Active Zone: Synaptic Ribbons

Some active zones have very prominent electron-dense projections extending from the cytomatrixinto the cytoplasm. They were first described and

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

active zones have been biochemically purified andmolecularly characterized (71). These dense pro-jections are ~50 nm in size, are pyramidlike(FIGURE 2, E AND F), and contain synaptic vesiclebinding proteins such as synapsin and RIM (39, 71,96). Therefore, the pyramidlike dense projectionsof mammalian CNS synapses are thought to tetherand cluster synaptic vesicles to the active zone.Despite their proposed function of tetheringsynaptic vesicles, dense projections do not seem tobe essential for neurotransmitter release, as sug-gested by a series of knockout studies. In the mam-malian retina, Bassoon has been shown to be acomponent of photoreceptor ribbon synapse (6).In homozygous bassoon mutant mouse retinas,photoreceptors have a greatly reduced number ofribbons and the existing ribbons are freely floatingin the cytoplasm with synaptic vesicles attached tothem. Electroretinogram recordings from thesephotoreceptors showed that neurotransmission ismaintained at low-intensity light stimulation butdramatically reduced at high-intensity light stimu-lation (16). This suggests that synaptic vesicles canfuse without the ribbon anchored to the plasmamembrane. However, strong and continuous stim-ulation cannot be sustained. Knockout studies ofRIM1 in C. elegans suggest that RIM1 is notrequired for morphological docking of synapticvesicles or calcium-regulated fusion (49). In addi-tion, RIM1 knockout mice have very mild synapticphenotypes, suggesting that RIM1 may not beessential for neurotransmitter release. Alternat-ively, there is genetic redundancy, and other pro-teins can perform a similar function to that of RIM(81). In addition, knockout studies of synapsin sug-gest that it may not be required for the vesiclefusion step but rather for the synapsin-dependentcluster of vesicles. This clustering is apparentlyrequired to sustain release of neurotransmitter inresponse to high levels of neuronal activity (39).

If the function of dense projections is tetheringsynaptic vesicles, what is the physiological advan-tage of varying their size and shape? One possibili-ty is that larger dense projections greatly increasethe number of synaptic vesicles tethered at theactive zone and therefore increase the size of thereadily releasable pool. This is quite evident in rib-bon synapses. For example, the ribbon in frog sac-cular hair cells is a sphere ~0.5 m in diameter, andon average there are 400 vesicles attached to thesurface of the spherical projection, whereas 124vesicles are attached to the active zone membrane.All vesicles attached to the ribbon can be releasedupon strong stimulation [FIGURE 1L (53, 54, 58)].Therefore, these large, dense projections allow anincrease in the size of the readily releasable poolwithout increasing the size of the active zonefusion area and PSD field. This feature is particular-

267PHYSIOLOGY • Volume 19 • October 2004 • www.physiologyonline.org

REVIEWSly important in sensory synapses, because sus-tained release upon continuous stimulationrequires a huge readily releasable pool and a largesynaptic vesicle replenish capacity, and the definedspace representation of an individual sensory neu-ron in the visual or auditory field restricts the sizeof each terminal. In contrast, at many NMJs, wherestimulations are not continuous, the size of nerveterminals is not restricted and the active zone mustexpand as the muscle grows, so dense projectionsare relatively small. Interestingly, in Drosophila andcrustacean NMJs, active zones with prominent Tbars can be seen adjacent to those without T barswithin the same presynaptic nerve terminal(FIGURE 1, D AND F). It has been proposed that theactive zones with prominent T bars have a strongeroutput, possibly because more synaptic vesiclesare released upon stimulation. Supporting evi-dence comes from crustacean studies showing thathigh-output NMJ terminals have a threefold higherdensity of dense projections than the low-outputterminals arising from the same excitatory motoraxon, although no difference was observed in totalsynaptic area (29, 30).

In summary, although the morphology of denseprojections varies greatly among different types ofsynapses, the primary function of dense projec-tions is to tether synaptic vesicles at the activezone. Larger dense projections tether more synap-tic vesicles and therefore increase the size of thereadily releasable pool.

Active Zone Assembly and theRegulation of Active Zone Densityand Spacing

Active zone assembly occurs after the initial axon-target recognition and contact. It ends with theestablishment of functional neurotransmitterrelease sites. In cultured hippocampal neurons,active zone assembly takes ~30 min (2, 24).According to a recently proposed unitary assemblymodel, active zone-specific proteins are packagedinto transport vesicles for delivery to the nascentsynaptic contact site. Upon fusion of such vesicleswith the plasma membrane, the active zoneproteins are deposited and localized. In culturedhippocampal synapses, one active zone forms fromtwo to three transport vesicles (84, 103).Considering that on average each active zone has10–15 synaptic vesicle release sites or “grid units,”each transport vesicle should carry the buildingmaterial for 4–5 synaptic vesicle release sites. Thismodel suggests that active zone assembly occurswithin an hour, which allows rapid synaptogenesisduring development and synapse expansionduring activity-dependent long-term potentiation.

Genetic analyses in C. elegans and Drosophila

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

268 PHYSIOLOGY • Volume 19 • October 2004 • www.physiologyonline.org

have identified mutations of several genes thataffect active zone assembly. The first gene was the C.elegans syd-2 gene. In the syd-2 loss-of-functionmutants, active zones of NMJ terminals are length-ened and are less electron dense (105). The Syd-2protein is localized to active zones and is a memberof the Liprin protein family, which contains coiled-coil and sterile �-motif domains (83). Liprins inter-act with the Lar family of receptor protein tyrosinephosphatases (RPTPs) and cluster RPTPs to focaladhesions (83). Drosophila Liprin-� is also localizedto active zones at NMJs, and in flies mutant forLiprin and Dlar, the size of active zones are ~2.5-foldbigger than normal and the morphology is moreirregular (45). In Drosophila, loss of wishful think-ing (wit) causes a reduced number of boutons, anincreased number of active zones per bouton, andfreely floating T bar structures in the cytoplasm (1).Wit is a BMP type II receptor that is expressed in asubset of neurons, including motor neurons.However, the mechanism as to how Wit regulatesactive zone assembly is not understood (1, 56).

Active zones are not static but rather plasticstructures. In tetrad synapses of the Drosophilavisual system, the number of presynaptic rib-bons/T bars changes with alterations in light stim-ulation (7, 78). In crustacean NMJs, high-frequencystimulation-induced long-term facilitation alsocorrelates with an increase in the number of activezones and dense projections (98). In mammalianhippocampal neurons, long-term potentiation alsocorrelates with the expansion or “division” of activezones (33, 97). Interestingly, a recent study using atransgenic calcium-imaging technique atDrosophila larval NMJs demonstrated that persist-ent strengthening of junctional vesicle releaserelies on the recruitment of additional active zoneswith normal spacing; increasing active zone densi-ty alone without changing bouton size only resultsin transient increase in evoked vesicle release onsingle active potentials but not consolidatedenhancement of neurotransmission (73).Therefore, long-term potentiation requires assem-bly of new active zones and expansion of the bou-ton size, because the density of active zones has tobe maintained. Ultrastructural observations fromDrosophila and Sarcophaga have also suggestedthat the density of active zones is tightly regulatedand that there is a minimum spacing requiredbetween neighboring active zones (62), presum-ably allowing each active zone to have sufficientand equal access to synaptic vesicle pools and/orrecycling machinery.

Summary

The active zone in the presynaptic nerve terminal isa highly organized, beautifully designed, and

REVIEWSdynamic structure specialized for neurotransmitterrelease. With the identification of some of thebuilding blocks of active zones, we are beginning tounderstand some of the morphological andfunctional properties of the active zone. Activezones consist of a plasma membrane, a cytomatrix,and dense projections. Although these threecomponents are morphologically distinct, they areintimately connected with each other and thepostsynaptic nerve terminal, ensuring the fidelityof synaptic vesicle tethering, docking, and fusion,as well as signal detection. Although themorphology of active zones and the molecularcomposition vary among species, tissues, and cells,the architectural design of the active zone is likelyto be conserved. �

References1. Aberle H, Haghighi AP, Fetter RD, McCabe BD, Magalhaes

TR, and Goodman CS. Wishful thinking encodes a BMP typeII receptor that regulates synaptic growth in Drosophila.Neuron 33: 545–558, 2002.

2. Ahmari SE, Buchanan J, and Smith SJ. Assembly of presynap-tic active zones from cytoplasmic transport packets. NatNeurosci 3: 445–451, 2000.

3. Atwood HL and Karunanithi S. Diversification of synapticstrength: presynaptic elements. Nat Rev Neurosci 3:497–516, 2002.

4. Bennett MR, Farnell L, and Gibson WG. The probability ofquantal secretion near a single calcium channel of an activezone. Biophys J 78: 2201–2221, 2000.

5. Bloom FE and Aghajanian GK. Fine structural and cytochem-ical analysis of the staining of synaptic junctions with phos-photungstic acid. J Ultrastruct Res 22: 361–375, 1968.

6. Brandstatter JH, Fletcher EL, Garner CC, Gundelfinger ED,and Wassle H. Differential expression of the presynapticcytomatrix protein bassoon among ribbon synapses in themammalian retina. Eur J Neurosci 11: 3683–3693, 1999.

7. Brandstatter JH and Meinertzhagen IA. The rapid assembly ofsynaptic sites in photoreceptor terminals of the fly’s opticlobe recovering from cold shock. Proc Natl Acad Sci USA 92:2677–2681, 1995.

8. Brose N, Hofmann K, Hata Y, and Sudhof TC. Mammalianhomologues of Caenorhabditis elegans unc-13 gene definenovel family of C2-domain proteins. J Biol Chem 270:25273–25280, 1995.

9. Burns ME and Augustine GJ. Synaptic structure and function:dynamic organization yields architectural precision. Cell 83:187–194, 1995.

10. Butz S, Okamoto M, and Sudhof TC. A tripartite protein com-plex with the potential to couple synaptic vesicle exocytosisto cell adhesion in brain. Cell 94: 773–782, 1998.

11. Catterall WA. Interactions of presynaptic Ca2+ channels andsnare proteins in neurotransmitter release. Ann NY Acad Sci868: 144–159, 1999.

12. Chavis P and Westbrook G. Integrins mediate functional pre-and postsynaptic maturation at a hippocampal synapse.Nature 411: 317–321, 2001.

13. Cohen MW, Jones OT, and Angelides KJ. Distribution of Ca2+

channels on frog motor nerve terminals revealed by fluores-cent �-conotoxin. J Neurosci 11: 1032–1039, 1991.

14. Couteaux R and Pecot-Dechavassine M. Synaptic vesicles andpouches at the level of “active zones” of the neuromuscularjunction. C R Acad Sci Hebd Seances Acad Sci D 271:2346–2349, 1970.

15. Davis GW, Schuster CM, and Goodman CS. Genetic analysisof the mechanisms controlling target selection: target-derived Fasciclin II regulates the pattern of synapse forma-tion. Neuron 19: 561–573, 1997.

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

269PHYSIOLOGY • Volume 19 • August 2004 • www.physiologyonline.org

REVIEWS

33. Harris KM, Fiala JC, and Ostroff L. Structuralchanges at dendritic spine synapses during long-term potentiation. Philos Trans R Soc Lond B BiolSci 358: 745–748, 2003.

34. Hata Y, Butz S, and Sudhof TC. CASK: a noveldlg/PSD95 homolog with an N-terminal calmod-ulin-dependent protein kinase domain identifiedby interaction with neurexins. J Neurosci 16:2488–2494, 1996.

35. Haydon PG, Henderson E, and Stanley EF.Localization of individual calcium channels at therelease face of a presynaptic nerve terminal.Neuron 13: 1275–1280, 1994.

36. Heuser JE and Reese TS. Evidence for recyclingof synaptic vesicle membrane during transmitterrelease at the frog neuromuscular junction. J CellBiol 57: 315–344, 1973.

37. Heuser JE, Reese TS, and Landis DM. Functionalchanges in frog neuromuscular junctions studiedwith freeze-fracture. J Neurocytol 3: 109–131,1974.

38. Hiesinger PR, Scholz M, Meinertzhagen IA,Fischbach KF, and Obermayer K. Visualization ofsynaptic markers in the optic neuropils ofDrosophila using a new constrained deconvolu-tion method. J Comp Neurol 429: 277–288, 2001.

39. Hilfiker S, Pieribone VA, Czernik AJ, Kao HT,Augustine GJ, and Greengard P. Synapsins as reg-ulators of neurotransmitter release. Philos Trans RSoc Lond B Biol Sci 354: 269–279, 1999.

40. Hirokawa N, Sobue K, Kanda K, Harada A, andYorifuji H. The cytoskeletal architecture of thepresynaptic terminal and molecular structure ofsynapsin 1. J Cell Biol 108: 111–126, 1989.

41. Hsueh YP, Yang FC, Kharazia V, Naisbitt S, CohenAR, Weinberg RJ, and Sheng M. Direct interac-tion of CASK/LIN-2 and syndecan heparan sulfateproteoglycan and their overlapping distributionin neuronal synapses. J Cell Biol 142: 139–151,1998.

42. Hsueh YP and Sheng M. Regulated expressionand subcellular localization of syndecan heparansulfate proteoglycans and the syndecan-bindingprotein CASK/LIN-2 during rat brain develop-ment. J Neurosci 19: 7415–7425, 1999.

43. Jahn R, Lang T, and Sudhof TC. Membrane fusion.Cell 112: 519–533, 2003.

44. Jarvis SE, Barr W, Feng ZP, Hamid J, and ZamponiGW. Molecular determinants of syntaxin 1 modu-lation of N-type calcium channels. J Biol Chem277: 44399–44407, 2002.

45. Kaufmann N, DeProto J, Ranjan R, Wan H, andVan Vactor D. Drosophila liprin-� and the recep-tor phosphatase Dlar control synapse morpho-genesis. Neuron 34: 27–38, 2002.

46. Kawasaki F, Zou B, Xu X, and Ordway RW. Activezone localization of presynaptic calcium channelsencoded by the cacophony locus of Drosophila. JNeurosci 24: 282–285, 2004.

47. Kistner U, Wenzel BM, Veh RW, Cases-LanghoffC, Garner AM, Appeltauer U, Voss B,Gundelfinger ED, and Garner CC. SAP90, a ratpresynaptic protein related to the product of theDrosophila tumor suppressor gene dlg-A. J BiolChem 268: 4580–4583, 1993.

48. Koulen P, Fletcher EL, Craven SE, Bredt DS, andWassle H. Immunocytochemical localization ofthe postsynaptic density protein PSD-95 in themammalian retina. J Neurosci 18: 10136–10149,1998.

49. Koushika SP, Richmond JE, Hadwiger G, WeimerRM, Jorgensen EM, and Nonet ML. A post-dock-ing role for active zone protein Rim. Nat Neurosci4: 997–1005, 2001.

16. Dick O, tom Dieck S, Altrock WD, AmmermullerJ, Weiler R, Garner CC, Gundelfinger ED, andBrandstatter JH. The presynaptic active zone pro-tein bassoon is essential for photoreceptor rib-bon synapse formation in the retina. Neuron 37:775–786, 2003.

17. Dieck S, Sanmart-Vila L, Langnaese K, Richter K,Kindler S, Soyke A, Wex H, Smalla KH, Kampf U,Franzer JT, Stumm M, Garner CC, andGundelfinger ED. Bassoon, a novel zinc-fingerCAG/glutamine-repeat protein selectively local-ized at the active zone of presynaptic nerve ter-minals. J Cell Biol 142: 499–509, 1998.

18. Dresbach T, Qualmann B, Kessels MM, GarnerCC, and Gundelfinger ED. The presynapticcytomatrix of brain synapses. Cell Mol Life Sci 58:94–116, 2001.

19. Ellisman MH, Rash JE, Staehelin LA, and PorterKR. Studies of excitable membranes. II. A com-parison of specializations at neuromuscular junc-tions and nonjunctional sarcolemmas of mam-malian fast and slow twitch muscle fibers. J CellBiol 68: 752–774, 1976.

20. Fanning AS and Anderson JM. PDZ domains: fun-damental building blocks in the organization ofprotein complexes at the plasma membrane. JClin Invest 103: 767–772, 1999.

21. Fenster SD, Chung WJ, Zhai R, Cases-Langhoff C,Voss B, Garner AM, Kaempf U, Kindler S,Gundelfinger ED, and Garner CC. Piccolo, apresynaptic zinc finger protein structurally relatedto bassoon. Neuron 25: 203–214, 2000.

22. Ferreira A and Paganoni S. The formation ofsynapses in the central nervous system. MolNeurobiol 26: 69–79, 2002.

23. Frank M and Kemler R. Protocadherins. Curr OpinCell Biol 14: 557–562, 2002.

24. Friedman HV, Bresler T, Garner CC, and Ziv NE.Assembly of new individual excitatory synapses:time course and temporal order of synaptic mol-ecule recruitment. Neuron 27: 57–69, 2000.

25. Garcia EP, McPherson PS, Chilcote TJ, Takei K,and De Camilli P. rbSec1A and B colocalize withsyntaxin 1 and SNAP-25 throughout the axon, butare not in a stable complex with syntaxin. J CellBiol 129: 105–120, 1995.

26. Garner CC, Kindler S, and Gundelfinger ED.Molecular determinants of presynaptic activezones. Curr Opin Neurobiol 10: 321–327, 2000.

27. Garner CC, Nash J, and Huganir RL. PDZ domainsin synapse assembly and signalling. Trends CellBiol 10: 274–280, 2000.

28. Gottardi CJ and Gumbiner BM. Adhesion signal-ing: how �-catenin interacts with its partners. CurrBiol 11: R792–R794, 2001.

29. Govind CK and Meiss DE. Quantitative compari-son of low- and high-output neuromuscularsynapses from a motoneuron of the lobster(Homarus americanus). Cell Tissue Res 198:455–463, 1979.

30. Govind CK, Quigley PA, and Pearce J. Synapticdifferentiation between two phasic motoneuronsto a crayfish fast muscle. Invert Neurosci 4: 77–84,2001.

31. Hallam SJ, Goncharov A, McEwen J, Baran R, andJin Y. SYD-1, a presynaptic protein with PDZ, C2and rhoGAP-like domains, specifies axon identityin C. elegans. Nat Neurosci 5: 1137–1146, 2002.

32. Harlow ML, Ress D, Stoschek A, Marshall RM, andMcMahan UJ. The architecture of active zonematerial at the frog’s neuromuscular junction.Nature 409: 479–484, 2001.

50. Lagnado L. Ribbon synapses. Curr Biol 13: R631,2003.

51. Landis DM, Hall AK, Weinstein LA, and Reese TS.The organization of cytoplasm at the presynapticactive zone of a central nervous system synapse.Neuron 1: 201–209, 1988.

52. Lang T, Bruns D, Wenzel D, Riedel D, Holroyd P,Thiele C, and Jahn R. SNAREs are concentrated incholesterol-dependent clusters that define dock-ing and fusion sites for exocytosis. EMBO J 20:2202–2213, 2001.

53. Lenzi D, Crum J, Ellisman MH, and Roberts WM.Depolarization redistributes synaptic membraneand creates a gradient of vesicles on the synapticbody at a ribbon synapse. Neuron 36: 649–659,2002.

54. Lenzi D, Runyeon JW, Crum J, Ellisman MH, andRoberts WM. Synaptic vesicle populations in sac-cular hair cells reconstructed by electron tomog-raphy. J Neurosci 19: 119–132, 1999.

55. Lenzi D and von Gersdorff H. Structure suggestsfunction: the case for synaptic ribbons as exocy-totic nanomachines. Bioessays 23: 831–840,2001.

56. Marques G, Bao H, Haerry TE, Shimell MJ,Duchek P, Zhang B, and O’Connor MB. TheDrosophila BMP type II receptor wishful thinkingregulates neuromuscular synapse morphologyand function. Neuron 33: 529–543, 2002.

57. Martin-Moutot N, Charvin N, Leveque C, Sato K,Nishiki T, Kozaki S, Takahashi M, and Seagar M.Interaction of SNARE complexes with P/Q-typecalcium channels in rat cerebellar synaptosomes.J Biol Chem 271: 6567–6570, 1996.

58. Matthews G. Synaptic mechanisms of bipolar cellterminals. Vision Res 39: 2469–2476, 1999.

59. Maximov A, Sudhof TC, and Bezprozvanny I.Association of neuronal calcium channels withmodular adaptor proteins. J Biol Chem 274:24453–24456, 1999.

60. Mayford M, Barzilai A, Keller F, Schacher S, andKandel ER. Modulation of an NCAM-relatedadhesion molecule with long-term synaptic plas-ticity in Aplysia. Science 256: 638–644, 1992.

61. Meinertzhagen IA. Ultrastructure and quantifica-tion of synapses in the insect nervous system. JNeurosci Methods 69: 59–73, 1996.

62. Meinertzhagen IA, Govind CK, Stewart BA,Carter JM, and Atwood HL. Regulated spacing ofsynapses and presynaptic active zones at larvalneuromuscular junctions in different genotypes ofthe flies Drosophila and Sarcophaga. J CompNeurol 393: 482–492, 1998.

63. Missler M and Sudhof TC. Neurexins: three genesand 1001 products. Trends Genet 14: 20–26,1998.

64. Mizoguchi A, Nakanishi H, Kimura K, MatsubaraK, Ozaki-Kuroda K, Katata T, Honda T, Kiyohara Y,Heo K, Higashi M, Tsutsumi T, Sonoda S, Ide C,and Takai Y. Nectin: an adhesion moleculeinvolved in formation of synapses. J Cell Biol 156:555–565, 2002.

65. Muller BM, Kistner U, Veh RW, Cases-Langhoff C,Becker B, Gundelfinger ED, and Garner CC.Molecular characterization and spatial distribu-tion of SAP97, a novel presynaptic proteinhomologous to SAP90 and the Drosophila discs-large tumor suppressor protein. J Neurosci 15:2354–2366, 1995.

66. Muresan V, Lyass A, and Schnapp BJ. The kinesinmotor KIF3A is a component of the presynapticribbon in vertebrate photoreceptors. J Neurosci19: 1027–1037, 1999.

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from

94. Wang X, Kibschull M, Laue MM, Lichte B,Petrasch-Parwez E, and Kilimann MW. Aczonin, a550-kD putative scaffolding protein of presynap-tic active zones, shares homology regions withrim and bassoon and binds profilin. J Cell Biol147: 151–162, 1999.

95. Wang Y, Liu X, Biederer T, and Sudhof TC. A fam-ily of RIM-binding proteins regulated by alterna-tive splicing: implications for the genesis ofsynaptic active zones. Proc Natl Acad Sci USA 99:14464–14469, 2002.

96. Wang Y, Okamoto M, Schmitz F, Hofmann K, andSudhof TC. Rim is a putative Rab3 effector in reg-ulating synaptic-vesicle fusion. Nature 388:593–598, 1997.

97. Weeks AC, Ivanco TL, Leboutillier JC, Racine RJ,and Petit TL. Sequential changes in the synapticstructural profile following long-term potentiationin the rat dentate gyrus. II. Induction/early main-tenance phase. Synapse 36: 286–296, 2000.

98. Wojtowicz JM, Marin L, and Atwood HL. Activity-induced changes in synaptic release sites at thecrayfish neuromuscular junction. J Neurosci 14:3688–3703, 1994.

99. Yagi T and Takeichi M. Cadherin superfamilygenes: functions, genomic organization, and neu-rologic diversity. Genes Dev 14: 1169–1180,2000.

100. Yamagata M, Sanes JR, and Weiner JA. Synapticadhesion molecules. Curr Opin Cell Biol 15:621–632, 2003.

101. Yamagata M, Weiner JA, and Sanes JR. Sidekicks:synaptic adhesion molecules that promote lami-na-specific connectivity in the retina. Cell 110:649–660, 2002.

102. Yasuyama K, Meinertzhagen IA, and SchurmannFW. Synaptic organization of the mushroom bodycalyx in Drosophila melanogaster. J Comp Neurol445: 211–226, 2002.

103. Zhai RG, Vardinon-Friedman H, Cases-LanghoffC, Becker B, Gundelfinger ED, Ziv NE, and GarnerCC. Assembling the presynaptic active zone: acharacterization of an active one precursor vesi-cle. Neuron 29: 131–143, 2001.

104. Zhang L, Volknandt W, Gundelfinger ED, andZimmermann H. A comparison of synaptic proteinlocalization in hippocampal mossy fiber terminalsand neurosecretory endings of the neurohypoph-ysis using the cryo-immunogold technique. JNeurocytol 29: 19–30, 2000.

105. Zhen M and Jin Y. The liprin protein SYD-2 regu-lates the differentiation of presynaptic termini inC. elegans. Nature 401: 371–375, 1999.

REVIEWS

270 PHYSIOLOGY • Volume 19 • October 2004 • www.physiologyonline.org

81. Schoch S, Castillo PE, Jo T, Mukherjee K, GeppertM, Wang Y, Schmitz F, Malenka RC, and SudhofTC. RIM1� forms a protein scaffold for regulatingneurotransmitter release at the active zone.Nature 415: 321–326, 2002.

82. Schulze KL, Broadie K, Perin MS, and Bellen HJ.Genetic and electrophysiological studies ofDrosophila syntaxin-1A demonstrate its role innonneuronal secretion and neurotransmission.Cell 80: 311–320, 1995.

83. Serra-Pages C, Medley QG, Tang M, Hart A, andStreuli M. Liprins, a family of LAR transmembraneprotein-tyrosine phosphatase-interacting pro-teins. J Biol Chem 273: 15611–15620, 1998.

84. Shapira M, Zhai RG, Dresbach T, Bresler T, TorresVI, Gundelfinger ED, Ziv NE, and Garner CC.Unitary assembly of presynaptic active zones fromPiccolo-Bassoon transport vesicles. Neuron 38:237–252, 2003.

85. Shapiro L and Colman DR. The diversity of cad-herins and implications for a synaptic adhesivecode in the CNS. Neuron 23: 427–430, 1999.

86. Sheng M and Sala C. PDZ domains and theorganization of supramolecular complexes. AnnuRev Neurosci 24: 1–29, 2001.

87. Stanley EF. The calcium channel and the organi-zation of the presynaptic transmitter release face.Trends Neurosci 20: 404–409, 1997.

88. Takai Y, Shimizu K, and Ohtsuka T. The roles ofcadherins and nectins in interneuronal synapseformation. Curr Opin Neurobiol 13: 520–526,2003.

89. Takao-Rikitsu E, Mochida S, Inoue E, Deguchi-Tawarada M, Inoue M, Ohtsuka T, and Takai Y.Physical and functional interaction of the activezone proteins, CAST, RIM1, and Bassoon, in neu-rotransmitter release. J Cell Biol 164: 301–311,2004.

90. Taverna E, Saba E, Rowe J, Francolini M, ClementiF, and Rosa P. Role of lipid microdomains in P/Q-type calcium channel (Cav2.1) clustering andfunction in presynaptic membranes. J Biol Chem279: 5127–5134, 2004.

91. Von Gersdorff H. Synaptic ribbons: versatile sig-nal transducers. Neuron 29: 7–10, 2001.

92. Walrond JP and Reese TS. Structure of axon ter-minals and active zones at synapses on lizardtwitch and tonic muscle fibers. J Neurosci 5:1118–1131, 1985.

93. Walsh FS and Doherty P. Neural cell adhesionmolecules of the immunoglobulin superfamily:role in axon growth and guidance. Annu Rev CellDev Biol 13: 425–456, 1997.

67. O’Brien RJ, Lau LF, and Huganir RL. Molecularmechanisms of glutamate receptor clustering atexcitatory synapses. Curr Opin Neurobiol 8:364–369, 1998.

68. Packard M, Mathew D, and Budnik V. FAStremodeling of synapses in Drosophila. Curr OpinNeurobiol 13: 527–534, 2003.

69. Parsegian VA. Approaches to the Cell Biology ofNeurons. Bethesda, MD: Soc. Neurosci., 1977.

70. Pfenninger K, Akert K, Moor H, and Sandri C. Thefine structure of freeze-fractured presynapticmembranes. J Neurocytol 1: 129–149, 1972.

71. Phillips GR, Huang JK, Wang Y, Tanaka H, ShapiroL, Zhang W, Shan WS, Arndt K, Frank M, GordonRE, Gawinowicz MA, Zhao Y, and Colman DR. Thepresynaptic particle web: ultrastructure, composi-tion, dissolution, and reconstitution. Neuron 32:63–77, 2001.

72. Pumplin DW, Reese TS, and Llinas R. Are thepresynaptic membrane particles the calciumchannels? Proc Natl Acad Sci USA 78: 7210–7213,1981.

73. Reiff DF, Thiel PR, and Schuster CM. Differentialregulation of active zone density during long-term strengthening of Drosophila neuromuscularjunctions. J Neurosci 22: 9399–9409, 2002.

74. Rizo J. SNARE function revisited. Nat Struct Biol10: 417–419, 2003.

75. Robitaille R, Adler EM, and Charlton MP.Strategic location of calcium channels at transmit-ter release sites of frog neuromuscular synapses.Neuron 5: 773–779, 1990.

76. Rosenmund C, Rettig J, and Brose N. Molecularmechanisms of active zone function. Curr OpinNeurobiol 13: 509–519, 2003.

77. Rougon G and Hobert O. New insights into thediversity and function of neuronal immunoglobu-lin superfamily molecules. Annu Rev Neurosci 26:207–238, 2003.

78. Rybak J and Meinertzhagen IA. The effects oflight reversals on photoreceptor synaptogenesisin the fly Musca domestica. Eur J Neurosci 9:319–333, 1997.

79. Scheiffele P. Cell-cell signaling during synapseformation in the CNS. Annu Rev Neurosci 26:485–508, 2003.

80. Schmucker D, Clemens JC, Shu H, Worby CA,Xiao J, Muda M, Dixon JE, and Zipursky SL.Drosophila Dscam is an axon guidance receptorexhibiting extraordinary molecular diversity. Cell101: 671–684, 2000.

on Novem

ber 6, 2009 physiologyonline.physiology.org

Dow

nloaded from