U

it

U

Molecular and Cellular Neuroscience 15, 417–428 (2000)

doi:10.1006/mcne.2000.0839, available online at http://www.idealibrary.com on MCN

A

Temporal Appearance of the PresynapticCytomatrix Protein Bassoonduring Synaptogenesis

Rong Zhai,* Gisela Olias,* Wook Joon Chung,* Robin A. J. Lester,*Susanne tom Dieck,† Kristina Langnaese,† ,1 Michael R. Kreutz,†

Stefan Kindler,‡ Eckart D. Gundelfinger,† and Craig C. Garner*,2

*Department of Neurobiology, University of Alabama at Birmingham, Birmingham, Alabama35213-0021; †Leibniz Institute for Neurobiology, Brenneckestrasse 6, D-39118 Magdeburg,Germany; and ‡Institute for Cell Biochemistry and Clinical Neurobiology,

niversity of Hamburg, D-20246 Hamburg, Germany

1aspl

tcab[n(ett

Bassoon is a 420-kDa presynaptic cytomatrix protein poten-tially involved in the structural organization of neurotrans-mitter release sites. In this study, we have investigated apossible role for Bassoon in synaptogenesis and in definingsynaptic vesicle recycling sites. We find that it is expressedat early stages of neuronal differentiation in which it is se-lectively sorted into axons. As synaptogenesis begins, Bas-soon clusters appear along dendritic profiles simulta-neously with synaptotagmin I, sites of synaptic vesiclerecycling, and the acquisition of functional excitatory andinhibitory synapses. A role for Bassoon in the assembly ofexcitatory and inhibitory synapses is supported by the colo-calization of Bassoon clusters with clusters of GKAP andAMPA receptors as well as GABAA receptors. These datandicate that the recruitment of Bassoon is an early step inhe formation of synaptic junctions.

INTRODUCTION

Synapses of the central nervous system (CNS) arehighly specialized asymmetric sites of cell–cell contactdesigned for the rapid and repetitive signaling betweenneurons and their targets (Burns and Augustine, 1995).The presynaptic bouton is typically filled with severalhundred synaptic vesicles (SVs) (Burns and Augustine,

1 Present address: Institute of Human Genetics, Otto von Guerickeniversity, Leipziger Strasse 44, D-39112 Magdeburg, Germany.2 To whom correspondence should be addressed at the University

of Alabama at Birmingham, 1719 Sixth Avenue South CIRC 589,

ne

Birmingham, Alabama 35294-0021. Fax: (205) 934-6571. E-mail:garner@nrc.uab.edu.

1044-7431/00 $35.00Copyright © 2000 by Academic Press

ll rights of reproduction in any form reserved.

1995; Pieribone et al., 1995) that are localized in thevicinity of the active zone, a specialized region of thepresynaptic plasma membrane where SVs dock andfuse (Burns and Augustine, 1995; Pieribone et al., 1995;Landis et al., 1988). In recent years, the protein machin-ery directing the SV cycle has been well characterized(Sudhof, 1995; DeCamilli and Takei, 1996). However,the molecular mechanisms that restrict these events tothe active zone remain unknown. Deep etch freeze frac-ture studies of synaptic junctions have revealed that,similar to the postsynaptic density (PDS), the activezone contains an electron-dense presynaptic cytoskel-etal matrix (PCM) (Landis et al., 1988; Hirokawa et al.,989; Gotow et al., 1991). This matrix is thought to playfundamental role in defining neurotransmitter release

ites, keeping the active zone in register with theostsynaptic reception apparatus, regulating the mobi-

ization of SVs, and the refilling of release sites.The recent characterization of proteins that are selec-

ively localized to the PCM of CNS synapses provideslues to how the presynaptic half of the junction isssembled. Thus far three components of the PCM haveeen characterized. These include the 530-kDa PiccoloCases-Langhoff et al., 1996; Fenster et al., 2000; alsoamed Aczonin (Wang et al., 1999)], the 180-kDa RimWang et al., 1997), and the 420-kDa Bassoon (tom Dieckt al., 1998). All three are multidomain zinc finger pro-eins that are thought to perform scaffold functions athe active zone. At present, only a few interacting part-

ers for these PCM proteins have been reported. Forxample, the zinc finger in Rim binds rab3A/C in a

417

iappctKpanreBnagmsnvhralitS

418 Zhai et al.

GTP-dependent manner (Wang et al., 1997), while thosein Piccolo interact with the prenylated rab-associatedprotein PRA1 (Fenster et al., 2000). These observationssuggest that in addition to their structural functionPCM components may play a role in SV cycling atactive zones.

At present, direct proof that PCM proteins define theactive zone as the site of SV docking, fusion, and recy-cling is lacking. Utilizing antibodies against SV andPSD components, it was shown that SVs and structuralcomponents of the PSD such as SAP90/PSD95 andGKAP cluster at newly forming synapses at about thesame time (Rao et al., 1998). In contrast, synaptic recruit-ment of N-methyl-d-aspartate and a-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptor sub-units appears to be a later event (Rao et al., 1998). Thisndicates that SV clustering is an early event in thessembly of CNS synapses. This hypothesis is sup-orted by the early developmental expression of SVroteins, an accumulation of SVs in axonal growthones (Matteoli et al., 1992), and neurotransmitter secre-ion from growth cones (Young and Poo, 1983;idokoro and Yeh, 1982). To examine whether the PCMrotein Bassoon may play a role in early synaptogenesisnd in defining the presynaptic active zone as the site ofeurotransmitter release, we have examined the tempo-al appearance of Bassoon and its transcripts in differ-ntiating neurons. We find that in developing rat brain,assoon transcripts are expressed at early stages ofeuronal differentiation. In cell culture, Bassoon issymmetrically sorted together with SVs into axonalrowth cones at early stages of differentiation. Further-ore, Bassoon begins to cluster along dendritic profiles

imultaneously with SVs as well as structural compo-ents of the PSD, indicating that Bassoon may be in-olved in defining neurotransmitter release sites. Thisypothesis is supported by the colocalization of SVecycling sites at newly formed Bassoon clusters. Inddition, Bassoon cluster formation temporally corre-ates with the acquisition of functional excitatory andnhibitory synapses. Taken together, these data implyhat Bassoon is a structural PCM component definingV release sites at newly forming synapses.

RESULTS

Bassoon Is Expressed at Early Stages of Rat BrainDevelopment

The localization of Bassoon within the PCM at ma-ture synapses (tom Dieck et al., 1998; Richter et al., 1999)

suggests a fundamental role for Bassoon in definingactive zones as sites of SV docking and fusion. Toevaluate whether it may perform a similar functionduring synaptogenesis, we investigated the presence ofBassoon mRNAs during brain development. In aNorthern blot analysis, total RNA from embryonic day19 (E19) to postnatal day 40 (P40) rat brain was hybrid-ized with a radiolabeled Bassoon cDNA (clone sap7f;tom Dieck et al., 1998). Two transcripts of 13 and 14 kbwere detected, indicating the existence of alternativelyprocessed Bassoon transcripts. At E19, weak hybridiz-ing bands representing both transcripts are visible (Fig.1, lane 1). Both increase in intensity until P10 (Fig. 1,lanes 2 and 3). Around P20, the intensity of the 14-kbband begins to decline and largely disappears by P40(Fig. 1, lane 4). The intensity of the 13-kb band, how-ever, continues to rise before falling to an adult levelaround P40 (Fig. 1, lanes 4 and 5). These data demon-strate that Bassoon transcripts are expressed at earlystages of neuronal differentiation and throughout thepeak period of synaptogenesis that occurs between

FIG. 1. Expression of Bassoon transcripts during rat brain develop-ment. Total RNA was isolated from rat brain at the age of embryonicday 19 (E19) and postnatal days 1, 10, 20, and 40 (P1, P10, P20, andP40) (lanes 1–5, respectively). Total RNA was separated on 1.2%formaldehyde agarose gels and blotted onto Qiabrane nylon. (A)Nylon membrane was hybridized with a-32P-labeled cDNA probes.Two bands at 13 and 14 kb were visible at E19. The intensities of bothbands rose and then declined from E19 to P40, but the highestintensities of both 13- and 14-kb bands are between P10 and P20. (B)Membrane was stained with methylene blue, as control, indicatingthat the same amount of total RNA was loaded on the gel. Thepositions of the 18S and 28S ribosomal RNAs are labeled.

postnatal days 10 and 30 (Altman, 1965; Melloni andDeGennaro, 1994).

w

f Bas2), af the

419Bassoon Expression during Synaptogenesis

In situ hybridization of horizontal brain sections wassubsequently used to examine the spatial distribution ofBassoon transcripts in developing rat brain. At earlystages (P1 and P4) (Figs. 2A and 2C), Bassoon mRNAswere first detectable in the forebrain and hippocampus. Asignificant hybridization signal was also seen in the cere-bellum beginning around P4 (Fig. 2C), which correspondsto early stages of cerebellar development when granulecells begin to migrate from external to internal layers(Melloni and DeGennaro, 1994). By P10, transcript levelsin the cerebellum begin to mirror the increasing levelsfound in the cerebral cortex and hippocampal formation(Fig. 2E). This pattern does not change significantly up to1 year of age (Fig. 2K). A detailed analysis of the devel-oping hippocampal formation, in particular the dentategyrus, revealed that Bassoon transcripts appear in thepremigratory hilar cells at P1 and P4 (Figs. 2B and 2D). AtP10, Bassoon mRNA is present in the granule cell layer ofthe dentate gyrus (Fig. 2F). As increasing numbers of cellsarrive and differentiate, Bassoon transcript levels in thislayer rise and remain high in the adult brain (Figs. 2H, 2J,and 2L). No specific hybridization signals were observedwhen sections were incubated with a 100-fold excess ofunlabeled oligonucleotide, after RNase treatment of sec-

FIG. 2. Expression and distribution of Bassoon mRNA in sections oere hybridized with 35S-labeled Bassoon oligonucleotide probes. Ba

C, E, G, I, K). In hippocampus (B, D, F, H, J, L), Bassoon mRNA appein the granule cell layer of the dentate gyrus. High, near-adult levels othe cerebellum in the P21 section was lost during tissue processing (Gwith an intact cerebellum was included (G1). In F: CA3, CA3 field o

tions before hybridization, or in sections hybridized witha sense probe (see tom Dieck et al., 1998). These data

demonstrate that Bassoon transcripts are expressedthroughout brain postnatal development as well as inadult brain and are thus consistent with a potential rolefor Bassoon in synaptogenesis.

Expression of Bassoon in DifferentiatingHippocampal Neurons

Utilizing hippocampal cultures as a model systemto study synaptogenesis (Goslin and Banker, 1991;Matteolli et al., 1995), we next examined the temporaland spatial appearance of Bassoon in differentiatingneurons by immunofluorescence microscopy. Anti-bodies against MAP2, a somatodendritic microtu-bule-associated protein (Matus et al., 1986), were usedto visualize dendrites at different stages of neuronaldifferentiation (Goslin and Banker, 1991). In stage 2neurons, when minor processes first appear (;2 daysin vitro; div), Bassoon immunoreactivity is observedas fine puncta in all MAP2-positive processes (Figs.3A and 3B). At 3 div, as axonal outgrowth is initiated(stage 3), the distributions of Bassoon and MAP2become polarized. In contrast to the dendritic local-ization of MAP2 (Fig. 3D), Bassoon exhibits primarily

eloping rat brain and hippocampus. 14-mm horizontal brain sectionsmRNA was first detectable in forebrain and hippocampus at P1 (A,

in the premigratory hilar cells from P4 and then as thin band of cellssoon mRNA were present in dentate granule cells by P21. Note. Sincehorizontal section taken from a more dorsal region of the same brainhippocampus; DG, dentate gyrus; H, hilar region.

f devssoonared

a fine puncta pattern in the distal part of the processthat appears to be the axon and its growth cone (Figs.

estricunct

he ce

420 Zhai et al.

3C and 3I). The axonal localization of Bassoon wasconfirmed by double labeling 3-div neurons with theaxonal marker protein tau (Figs. 3I and 3J). By 10 div,Bassoon aggregates into larger clusters along den-dritic profiles (Fig. 3E). The number of Bassoon clus-ters increased steadily during neuronal maturation(10 –26 div) (Figs. 3E and 3G), indicative of an asso-ciation of Bassoon with newly forming synapses.

SVs Accumulate and Recycle at Sites whereBassoon Is Clustered

FIG. 3. Spatial distribution of Bassoon and MAP2 in differentiatinbecomes presynaptically localized at later stages. Cultured hippocamdiv and immunostained for Bassoon (A, C, E, G, I) and the dendritic mBassoon is present in all processes as fine puncta (A); MAP2b is also pas neurites begin to differentiate, MAP2b immunoreactivity becomes r(D), tau immunoreactivity is distributed in the axon (J), and Bassoon pBassoon immunoreactivity is clustered along dendritic profiles and t

To assess whether Bassoon clusters along dendriticprofiles represented synapses, we compared the tempo-

ral appearance of Bassoon and synaptotagmin I clus-ters. In mature neurons (14 div), most of the synapto-tagmin I clusters colocalize with Bassoon clusters (Figs.4C and 4D), indicating that Bassoon accumulates atsynaptic boutons. In stage 3 neurons, the majority ofboth Bassoon and synaptotagmin I labeling is found inaxonal growth cones (stars in Figs. 4A and 4B). How-ever, a few Bassoon and synaptotagmin I coclusters canbe detected on cell soma and along proximal dendrites(insets in Figs. 4A and 4B). These data indicate thatduring neuronal differentiation Bassoon and SV clus-ters form at the same time and place. To examine

pocampal neurons. Bassoon is expressed in immature neurons andeuronal were fixed at 2 (A, B), 3 (C, D, I, J), 10 (E, F), and 15 (G, H)

er MAP2 (B, D, F, H) or axonal marker tau (J). In stage 2 cells (2 div),t in all neurites but in a diffuse pattern (B). In stage 3 (3 div) neurons,ted to the somatodendritic regions and proximal segment of the axona become concentrated in axons (C and I). At later stages (E, F, G, H),ll soma, suggesting a synaptic localization. Scale bars, 10 mm.

g hippal n

arkresen

whether these newly formed Bassoon clusters representsites of SV recycling, antibodies against the luminal

large number Bassoon and synaptotagmin I clusters can be seen tocolocalize (arrows in C, D). Scale bars, 20 mm.

recycling. At least one large synaptophysin cluster remains unlabeled by Sseen to colocalize with Bassoon clusters (arrows in A, B, C). Scale bars, 10

421Bassoon Expression during Synaptogenesis

domain of synaptotagmin I (Stg-lu Ab) were used tolabel SVs that are recycled from the plasma membrane(Kraszewski et al., 1995). When 4-div neurons werefixed after 15-min labeling step, internalized Stg-lu Abwere primarily seen as distinct large clusters alongsomatodendritic regions (arrows in Figs. 5B and 5E). Inneurons double stained with Bassoon antibodies, allrecycled SV clusters colocalized with the larger Bassoonpositive clusters (Figs. 5A–5C). Similar results were ob-tained when fixed neurons were stained with synapto-physin antibodies. However, here we observed a poolof synaptophysin immunoreactivity that was not la-beled by Stg-lu antibodies (arrowheads in Figs. 5D–5F).These presumably represent SVs that were not activelyrecycling during the uptake experiment. These dataindicate that newly formed Bassoon clusters coincidewith active sites of SV recycling.

Early Recruitment of Bassoon at Newly FormingExcitatory and Inhibitory Synapses

To examine whether Bassoon is recruited to bothexcitatory and inhibitory of synapses as they assemble,we compared the distribution of Bassoon to several

fter a 15-min incubation with antibodies against the luminal domainn (A) or synaptophysin (D). Internalized Stg-lu Ab are found at large

), indicating that Stg-lu Ab has labeled the sites of SV clustering and

FIG. 4. Bassoon clusters colocalize with synapotagmin I clustersalong dendritic profiles. 4- and 14-div neurons were fixed and doublelabeled for Bassoon (A, C) and synaptotagmin I (B, D). The axon ineach neuron is indicated by star. In 4-div neurons (A, B), Bassoon isconcentrated primarily in axons and their growth cones (A). Newlyformed Bassoon clusters along dendritic profiles colocalize with syn-aptotagmin I clusters (arrows in A, B). In 14-div neurons (C, D), a

FIG. 5. SVs recycle at Bassoon clustering sites. 4-div neurons were fixed aof synaptotagmin I (Stg-lu) at 37°C. Neurons were then stained for Bassooclusters (arrows in A–F) present along cell soma and dendritic profiles (B, E

tg-lu Ab (arrowheads in D–F). Large Stg-lu-positive clusters are alsomm.

serve–L. S

422 Zhai et al.

synaptic marker proteins. For excitatory synapse, neu-rons cultured for 4, 10, and 21 div were labeled withantibodies against GKAP, a structural component of thePSD (Naisbitt et al., 1997), GluR1 subunits of the AMPAreceptor, or the dendritic spine protein a-actinin(Wyszynski et al., 1997). In stage 3 neurons (3–4 div),GKAP and GluR1 staining were diffuse and restricted

FIG. 6. Coclustering of Bassoon with GKAP and GluR1 in differentia(A, C, E, G, I, K) and GKAP (B, D, F) or GluR1 (H, J, L). At 4 div, axonof the GKAP and GluR1 immunoreactivity is found diffuse in the soclusters are observed colocalizing with Bassoon clusters (arrows in A,decreases, the numbers of Bassoon, GKAP, and GluR1 clusters are obor GluR1 immunoreactivity is indicated by arrowheads in C–F and I

to somatodendritic areas, compared to the somatoax-onal distribution of Bassoon (stars in Figs. 6A, 6B, 6G,

and 6H). As observed with synaptotagmin I antibodies,the first small GKAP and GluR1 clusters present alongcell soma and proximal dendrites colocalized with asubset of Bassoon clusters (arrows in Figs. 6A, 6B, 6G,and 6H). Bassoon clusters that do not colocalize withGKAP or GLuR1 appear to represent inhibitory syn-apses (see below). In stage 4 and 5 neurons (10 and 21

neurons. 4-, 10-, and 21-div neurons were double labeled for Bassoonmpartments have strong Bassoon staining (stars in A, G) while mostdendritic compartment (B, H). At this time a few GKAP and GluR1H). At 10 and 21 div, as the level of diffuse GKAP and GluR1 stainingd to increase. A subpopulation of the Bassoon clusters lacking GKAPcale bars, 20 mm.

tingal co

matoB, G,

div, respectively), the number of GKAP and GluR1clusters long dendritic profiles have increased, while

uH

(ps(

toadeiqsaccy

B

D

sl rs lac

423Bassoon Expression during Synaptogenesis

the diffuse somatodendritic staining is decreased. At allthree stages, 100% of the GKAP and GluR1 clusterscontain Bassoon (arrows in Figs. 6C–6F and 6I–6L). Intotal, this represents about 55% of all Bassoon clusters.In contrast, a-actinin immunoreactivity remained dif-fuse in the somatodendritic compartment of neuronscultured for 4 and 10 div (data not shown). The a-acti-nin clusters, found in association with dendritic spines,were found to cocluster with Bassoon clusters around10 div (data not shown). These data indicate that a-ac-tinin localizes to excitatory synapses later than Bassoon,GKAP, and GluR1.

To examine the appearance of Bassoon in inhibitorysynapses, neurons were double labeled for Bassoon andGABA, receptor a subunits (GABAAR). At 4 div, mostof the GABAAR immunoreactivity was diffusely distrib-

ted in the somatodendritic compartment (Fig. 7B).owever, a few clusters of GABAAR, colocalizing with

a subset of Bassoon clusters, were observed on cellsoma and along dendritic profiles (arrows in Figs. 7Aand 7B). As neurons matured, the number of GABAARclusters per cell colocalizing with Bassoon clusters werefound to increase (Figs. 7C–7F). However, after quanti-fication the percentage of GABAAR/Bassoon clustersper cell was found to decrease as the cultures mature(68 6 11, 42 6 2, and 38 6 7% at 4, 10, and 21 div,

FIG. 7. Early coclustering of Bassoon and GABAAR. 4-, 10-, and 21-, F). At 4 div, GABAAR immunoreactivity is mostly found diffuse

clusters can be observed colocalizing with Bassoon at this time (arroween to increase, while the diffuse cytoplasmic staining of GABAAR iation of the Bassoon clusters (arrows in C–F), and the Bassoon cluste

20 mm.

respectively). The decrease in percentage of GABAAR/Bassoon clusters is associated with a reciprocal increase

we

in the percentage of GluR1/Bassoon clusters (26 6 3,57 6 11, and 54 6 2% at 4, 10, and 21 div, respectively).The ratio of GluR1/Bassoon (26 6 3%) to GABAAR/Bassoon (68 6 11%) receptor clusters (;1/3) at 4 divcorresponds very well to the ratio of excitatory to in-hibitory events (;1/3) at this developmental stagesee below). These observations are consistent withrevious studies showing that in the CNS, GABAynapses form earlier than glutamatergic synapsesChen et al., 1995).

To correlate the initial appearance of Bassoon clus-ers in 4 div with the acquisition of synaptic activityf neurons at this stage, spontaneous miniature syn-ptic events were recorded from cells cultured for 4iv. Among the seven cells recorded, six showedxcitatory activity and five showed inhibitory activ-ty. The inhibitory events happened at a higher fre-uency than excitatory events (Fig. 8). These datahow that functional inhibitory and excitatory syn-pses are present at this developmental stage and areonsistent with the hypothesis that the sites of colo-alization of Bassoon and cycling vesicles representoung synapses.Taken together these data show that the formation of

assoon clusters in differentiating neurons correlates

eurons were double labeled for Bassoon (A, C, E) and GABAAR (B,e somatodendritic compartment (B). However, numerous GABAAR, B). At 10 and 21 div, the numbers of Bassoon GABAAR clusters are

erved to decrease (D). GABAAR clusters colocalize with a subpopu-king GABAA clusters are indicated by arrowheads in C–F. Scale bars,

div nin ths in As obs

ell with the formation of physiologically functionalxcitatory and inhibitory synapses.

nBgsc

iara1ieBPgsBgatlta

eFI

424 Zhai et al.

DISCUSSION

In situ, Bassoon was recently identified as a compo-ent of the PCM that is assembled at active zones.ased on its large multidomain structure, it was sug-ested that Bassoon plays a role in defining SV releaseites and regulating the vesicle cycle at existing synapticontacts (tom Dieck et al., 1998). When Bassoon is first

FIG. 8. Miniature synaptic events in hippocampal neurons culturedfor 4 div. (A) Miniature synaptic current amplitude histograms forexcitatory (top) and inhibitory (bottom) events recorded from twodifferent hippocampal neurons. Bin size 2 pA. Representative exam-ples of synaptic events are shown (insets). Recording durations were313 and 213 s for the excitatory and inhibitory events, respectively. (B)Histogram of mean frequency of events for those cells with detectableevents. Total cells (n 5 7): excitatory (6/7), inhibitory (5/7).

recruited to newly differentiating presynaptic boutonsand whether it may be involved in neurotransmitter

release from axonal growth cones and developing syn-apses was not examined. In this study, we investigatedwhether temporally the arrival of Bassoon at newlyforming synapses coincides with the clustering and re-cycling of SVs. Bassoon and its transcripts are present atearly stages of neuronal differentiation. Moreover, Bas-soon is first recruited to both excitatory and inhibitorysynapses concomitant with an axonal/presynaptic ac-cumulation and the cycling of SVs. The early arrival ofBassoon in these compartments suggests it may be in-volved in structurally defining neurotransmitter releasesites and the regulation of the SV cycle.

The early expression of Bassoon in differentiatingneurons is supported by developmental Northern blots,in situ hybridization on rat brain sections, and immu-nofluorescence studies of primary hippocampal neu-rons. Specifically, by Northern analysis, we found thatBassoon transcripts are present in rat brain as early asembryonic day 19, a stage of rapid proliferation, differ-entiation, and migration of neuronal cells. Bassoon tran-script levels reach a peak around postnatal day 10–20,corresponding to the period of robust synaptogenesis inthe developing rat brain (Gaarskjaer, 1981; Melloni andDeGennaro, 1994). A comparable temporal expressionpattern of expression was also observed in individualbrain subregions by in situ hybridization. For example,n the hippocampus peak levels of Bassoon transcriptsre observed around postnatal day 21, a time that cor-esponds to the major period of neuronal differentiationnd synaptogenesis in the hippocampus (Gaarskjaer,981; Melloni and DeGennaro, 1994). Of particular notes the early appearance of Bassoon transcripts in differ-ntiating granule cells of the dentate gyrus. Specifically,assoon transcripts are detected in the hilar zone (at1–P4), where granule cells are born before they mi-rate radially into the granule cell layer and establishynaptic contacts (Bayer et al., 1982). Between P4 and 21,assoon transcripts accumulate in what is becoming theranule cell layer of the dentate gyrus. This temporalnd spatial pattern of expression is virtually identical tohat reported for transcripts encoding synapsin I (Mel-oni and DeGennaro, 1994), demonstrating that Bassoonranscripts are expressed in neurons as they differenti-te and begin to make synapses.Immunocytochemically, Bassoon was detected in

arly differentiating cultured hippocampal neurons.or example, similar to the SV proteins synaptotagmin(Fig. 4) and synapsin I (Fletcher et al., 1991; Mundigel

et al., 1993; Matteoli et al., 1995), Bassoon immunoreac-tivity was initially found in cell soma and all minor

neurites in stage 2 neurons. Subsequently, it becameasymmetrically sorted into axonal profiles and their

pdanesd

iisBtm

c1bGsMG

425Bassoon Expression during Synaptogenesis

growth cones in stage 3 neurons and clustered alongdendritic profiles in stage 4 neurons. Interestingly, thepunctate pattern observed for Bassoon in immatureaxons is atypical for most cytoskeletal proteins such asMAP2 (Fig. 3), tau (Fig. 3), a-actinin, or the growthcone-associated protein GAP-43 (Goslin et al., 1990), butsimilar to SV proteins. The nature of the Bassoon-posi-tive puncta in axons is unknown. They may represent atransport particle or hot spots of SV recycling at theplasmalemma.

An important issue raised by the localization of Bas-soon to the PCM assembled at the active zone of maturesynapses in situ (tom Dieck et al., 1998; Richter et al.,1999; Brandstaetter et al., 1999) is whether Bassoon

lays a general role in defining sites of SV cycling. Ourata show that Bassoon and synaptotagmin I coclustersre present along dendritic profiles at early stages ofeuronal differentiation. Furthermore, we find thatarly Bassoon clusters are sites of SV recycling, as as-essed by the uptake of antibodies against the luminalomain of synaptotagmin I (Matteoli et al., 1992;

Kraszewski et al., 1995). This suggests that Bassoon isrecruited to the presynaptic bouton at the initial phaseof synaptogenesis. This conclusion is supported by thecolocalization of Bassoon clusters on immature neuronswith aggregation sites of postsynaptic proteins such asGKAP, GluR1, or GABAAR. The existence of functionalsynapses in early hippocampal cultures is further con-firmed by the electrophysiological detection of sponta-neous miniature excitatory and inhibitory events at thisdevelopmental stage. These data are consistent withprevious studies showing that synapses appear at sitesof cell–cell contact between hippocampal neuronsaround 4 div (Fletcher et al., 1994; Matteoli et al., 1995;Rao et al., 1998). Taken together, these observationsndicate that Bassoon clusters along the dendrites ofmmature hippocampal neurons are found at functionalynaptic junctions. Moreover, this early recruitment ofassoon to developing synapses, as well as its localiza-

ion to the cytoskeleton associated with active zones ofature synapses (tom Dieck et al., 1998; Richter et al.,

1999; Brandstaetter et al., 1999), suggests that Bassoonmay participate in structurally defining neurotransmit-ter release sites of newly forming presynaptic boutons.

At present it is unclear how Bassoon may help todefine the active zone as the site of neurotransmitterrelease. Clearly the identification of Piccolo/Aczoninand Rim (Cases-Langhoff et al., 1996; Wang et al., 1997;Wang et al., 1999) as components of the PCM indicatethat Bassoon does not act alone to carry out this func-

tion. Furthermore, their multidomain structures (tomDieck et al., 1998; Fenster et al., 2000; Wang et al., 1997;

aC

Wang et al., 1999) may indicate that these PCM proteinsdefine the active zone by scaffolding components of theendo- and exocytotic machinery.

EXPERIMENTAL METHODS

Materials. The rabbit polyclonal and mouse mono-clonal Bassoon antibodies against a 75-kDa GST–Bas-soon fusion protein from clone sap7f were raised andpurified as described previously (tom Dieck et al., 1998)and used at a dilution of 1:200. The other antibodiesused were as follows: rabbit anti-MAP2b antibody (Kin-dler et al., 1990) at a dilution of 1:15,000; mouse mono-lonal anti-tau antibody (tau1; gift from Lester Binder;:200); mouse monoclonal anti-synaptotagmin I anti-ody (Cl 41.1; gift from R. Jahn, Max Planck Inst.,oettingen, Germany; 1:250); mouse monoclonal anti-

ynaptophysin antibody (Roche Diagnostics GmbH,annheim, Germany; 1:250); rabbit polyclonal anti-KAP antibody (9589/8; Naisbitt et al., 1997; 1:200);

mouse monoclonal anti-a-actinin antibody (clone EA-53; Sigma Chemical Co., St. Louis, MO; 1:200); rabbitpolyclonal anti-GluR1 antibody (Chemicon Interna-tional, Inc., CA; 1:50); mouse monoclonal anti-GABAA

receptor, b chain antibody (clone bd 17; Roche Diagnos-tics GmbH, Mannheim, Germany; 1:30), rabbit poly-clonal anti-synaptotagmin I luminal domain antibody(gift from P. DeCamilli, Yale University, New Haven,CT).

RNA isolation and Northern blotting. Total RNAwas isolated from rat brain by a guanidinium isothio-cyanate procedure (Chomczynski and Sacchi, 1987). To-tal RNA was separated on 1.2% formaldehyde agarosegels, blotted onto Qiabrane nylon membranes (Qiagen),and hybridized with a-32P-labeled cDNA probes gener-ated from Bassoon cDNA sap7F (tom Dieck et al., 1998)as described previously (Kistner et al., 1993). This cDNAclone, which hybridizes with both the 14- and the 13-kbmRNAs, was used to generate the Bassoon antibodiesand is contained within the large 5-kb exon 5 shared byall high-molecular-weight Bassoon isoforms (Winter etal., 1999). Membranes were stained with methylene blueto visualize 28S and 18S ribosomal RNA.

In situ hybridization. In situ hybridization experi-ments were performed essentially as described (Seiden-becher et al., 1998). Three different 40-mer antisenseoligonucleotides (59-GGAAGACAGGGAGGTGGGT-GAGGTGCCAGATGTATAGCTA-39, 59-ACAGCGGT-GTCGTCTTCCTCCAAGTTGTCTTCCTCGGCGC-39,

nd 59-TAAGGCTCTCCATCTCCAGCTCAGGCTC-CGGTCTAGGTT-39) were used. All produced identi-

l

d

dftTwm

2

426 Zhai et al.

cal results. Controls included competition with 100-foldexcess of unlabeled oligonucleotide, RNase treatment ofsections before hybridization, and hybridization with asense probe (59-TAGCTATACATCTGGCACCTCAC-CCACCTCCCTGTCTTCC-39). No specific signals weredetected under any of these conditions.

Hippocampal cultures. Primary cultures of hip-pocampal neurons were prepared from Sprague–Daw-ley rat embryos at age E19. Hippocampi were dissectedout, stored in HBSS (Hepes–buffered Hanks’ balancedsalt solution without Ca21 and Mg21, pH 7.3), and dis-sociated with the papain dissociation system (Worth-ington, NJ). In brief, tissue was digested in papainsolution (oxygen-saturated Earle’s balanced salt solu-tion, pH 7.3, containing 20 units/ml papain, 1 mMl-cysteine, 0.5 mM EDTA, and 0.1 units/ml DNase I),shaking very gently at 37°C for 15 min. After tritura-tion, cells were centrifuged and resuspended in Earle’sbalanced salt solution, pH 7.3, containing bovine serumalbumin–ovomucoid protease inhibitor (1 mg/ml) and0.1 units/ml DNase I. Cells were plated onto poly-d-ysine-coated 18-mm-f glass coverslips (Assistent, Ger-

many) at densities from 2 to 5 3 104 cells/coverslip.Conditioned culture medium was used in plating andmaintaining the neurons. This medium was prepared asdescribed by Ye et al. (1998). Briefly, confluent astro-cytes (8 to 14 days of age) were incubated in Earle’sminimum essential medium without glutamine (GibcoBRL, Grand Island, NY) containing 10% fetal bovineserum (Hyclone, UT) and 20 mM glucose for 6 to 10 h at37°C in a 5% CO2/95% air humidified atmosphere. Theconditioned medium was filtered through a 0.2-mmbottom–top filter and used within 14 days. The neuronswere grown for up to 45 days. To inhibit the prolifera-tion of glial cells, Ara-C (1-2 mM; Sigma) was added 3

ays after plating.Immunocytochemistry and synaptotagmin luminal

omain antibody uptake. Cultures were fixed in 3.7%ormaldehyde in PBS, pH 7.4, for 20–30 min, washedwice with PBS, and permeabilized for 5 min with 0.25%riton X-100. Nonspecific binding was blocked for 2 hith PBS containing 2% BSA, 5% FCS, 2% glycine, 50M NH4Cl, 0.05% NaN3. Cells were incubated with

primary antibodies diluted in 3% FCS in PBS overnightat 4°C. After being washed in PBS, cells were incubatedwith either fluorescein isothiocyanate- or Texas red-conjugated goat anti-mouse or goat anti-rabbit second-ary antibodies (Sigma). Cultures were rinsed in PBS andmounted using the Vectashield mounting medium(Vector Laboratories, CA). For Stg-lu Ab uptake

(Kraszewski et al., 1995), cultures were incubated with.0 mg Stg-lu Ab/ml for 15 min at 37°C. Cells were fixed

after two washes and processed for immunofluorescentlabeling for Stg-lu Ab and Bassoon (mouse monoclonalantibody) or synaptophysin (mouse monoclonal anti-body). Fluorescent images were taken with a NikonDiaphot 300 microscope equipped with a PhotometricsCH250 CCD camera. Digital images were processedand displayed with IP Lab Spectrum and AdobePhotoShop.

Quantification. To quantify the data from the im-munocytochemistry, neurons were randomly chosenfor image acquisition. At each stage (4, 10, and 21 div,five to seven cells were chosen and images taken in bothgreen and red channels. The area selected for analysiswas greater than 5000 mm2. For all the images, haze wasremoved, and the intensity of the staining was moni-tored using IP Lab Spectrum software. Clusters weredefined as fluorescence intensity of a puncta above acertain threshold. For the described experiments, thethreshold intensity was set at a value that was twofoldor greater in intensity than the diffuse fluorescencepresent along unstained dendritic segments. In eachpair of images, the percentage of clusters that colocalizewith the total number of Bassoon clusters was calcu-lated. Except for Piccolo clusters, the number of Bas-soon clusters outnumbered other synaptic proteins atall three stages. The data were analyzed using SPSSsoftware (SPSS, Inc., Chicago, IL).

Electrophysiological recording from hippocampalneurons in culture. After 4 days in culture, hippocam-pal neurons grown on coverslips were placed in a re-cording chamber and mounted on the stage of an in-verted microscope (Diaphot 200; Nikon). Neurons werecontinuously superfused with extracellular mediumcontaining (in mM) NaCl, 155; CaCl2, 2; MgCl2, 2;Hepes, 10, glucose, 10; pH 7.4. Standard whole-cellpatch clamp recordings were obtained under visualcontrol with Sylgard-coated pipettes filled with intra-cellular medium containing (in mM) Cs-methanesulfo-nate, 130–140; CsCl, 10; Hepes, 10; EGTA, 10. Miniaturesynaptic currents were recorded in the presence of TTX(100 nM). Miniature inhibitory postsynaptic currentswere isolated in the presence of DNQX (10 mM) at aholding potential of 0 mV. Miniature excitatorypostsynaptic currents were isolated in the presence ofbicuculline (100 mM) at a holding potential of 240 mV.Events were recorded to videotape using a Digital DataRecorder (Instrutec). Offline videotape recordings weredigitized at 10 kHz and low-pass filtered at 2 Hz usinga Digidata 1200 A/D board (Axon Instruments) on an80586-based computer using AxoScope software (Axon

Instruments). Miniature synaptic events were detectedautomatically as negative (excitatory) or positive (inhib-

B

C

C

C

D

F

G

G

K

K

K

L

M

M

M

M

M

N

P

427Bassoon Expression during Synaptogenesis

itory) deflections that were more than threefold aboveRMS noise using a Mini Analysis program (Synap-tosoft). All records were visually inspected and artifac-tual events excluded from final analysis.

ACKNOWLEDGMENTS

R. Zhai and G. Olias contributed equally to this work. This workwas supported by the Keck Foundation (931360) and the NationalInstitutes of Health (P50 HD32901, AG 12978-02, AG 06569-09) toC.C.G., the Deutsche Forschungsgemeinschaft (SFB 426,/A1, KR1879/1-1, SFB444/B1) to E.D.G., M.R.K., and S.K., and the Fonds derChemischen Industrie to E.D.G.

REFERENCES

Altman, J., and Das, G. D. (1965). Autoradiographic and histologicalevidence of postnatal hippocampal neurogenesis in rats. J. Comp.Neurol. 124: 319–335.

Bayer, S. A., Yackel, J. W., and Puri, P. S. (1982). Neurons in the ratdentate gyrus granular layer substantially increase during juvenileand adult life. Science 216: 890–892.

randstaetter, J. H., Fletcher, E. L., Garner, C. C., Gundelfinger, E. D.,and Waessle, H. (1999). Differential expression of the presynapticcytomatrix protein Bassoon among the ribbon synapses in themammalian retina. Eur. J. Neurosci., 11: 3683–3693.

Burns, M. E., and Augustine, G. J. (1995). Synaptic structure andfunction: Dynamic organization yields architectural precision. Cell83: 187–194.

ases-Langhoff, C., Voss, B., Garner, A. M., Appeltauer, U., Takei, K.,Kindler, S., Veh, R. W., De Camilli, P., and Gundelfinger, E. D., andGarner, C. C. (1996). Piccolo, a novel 420 kDa protein associatedwith the presynaptic cytomatrix. Eur. J. Cell Biol. 69: 214–223.

hen, G., Trombley, P. Q., and van den Pol, A. N. (1995). GABAreceptors precede glutamate receptors in hypothalamic develop-ment: Differential regulation by astrocytes. J. Neurophysiol. 74: 1473–1484.

homczynski, P., and Sacchi, N. (1987). Single-step method of RNAisolation by acid guanidinium thiocyanate–phenol–chloroform ex-traction. Anal. Biochem. 162: 156–159.e Camilli, P., and Takei, K. (1996). Molecular mechanisms in syn-aptic vesicle endocytosis and recycling. Neuron 16: 481–486.

enster, S. D., Chung, W. J., Cases-Langhoff, C., Zhai, R., Voss, B.,Garner, A. M., Kampf, U., Gundelfinger, E. D., and Garner, C. C.(2000). Piccolo, a presynaptic zinc finger protein structurally relatedto Bassoon. Neuron 25: 203–214.

Fletcher, T. L., Cameron, P., De Camilli, P., and Banker, G. (1991). Thedistribution of synapsin I and synaptophysin in hippocampal neu-rons developing in culture. J. Neurosci. 11: 1617–1626.

Fletcher, T. L., De Camilli, P., and Banker, G. (1994). Synaptogenesisin hippocampal cultures: Evidence indicating that exons and den-drites become competent to form synapses at different stages ofneuronal development. J. Neurosci. 14: 6695–6706.aarskjaer, F. B. (1981). The hippocampal mossy fiber system of therat studied with retrograde tracing techniques. Correlation between

topographic organization and neurogenetic gradients. J. Comp. Neu-rol. 203: 717–735.

R

oslin, K., and Banker, G. (1991). Rat hippocampal neurons in low-density culture. In Culturing Nerve Cells (G. Banker and K. Goslin,Eds.), pp. 251–282. MIT Press, Cambridge, MA.

Goslin, K., Schreyer, D. J., Skene, J. H., and Banker, G. (1990). Changesin the distribution of GAP-43 during the development of neuronalpolarity. J. Neurosci. 10: 588–602.

Gotow, T., Miyaguchi, K., and Hashimoto, P. H. (1991). Cytoplasmicarchitecture of the axon terminal: Filamentous strands specificallyassociated with synaptic vesicles. Neuroscience 40: 587–598.

Hirokawa, N., Sobue, K., Kanda, K., Harada, A., and Yorifuji, H.(1989). The cytoskeletal architecture of the presynaptic terminal andmolecular structure of synapsin 1. J. Cell Biol. 108: 111–126.

Kidokoro, Y., and Yeh, E. (1982). Initial synaptic transmission at thegrowth cone in Xenopus nerve–muscle cultures. Proc. Natl. Acad.Sci. USA 79: 6727–6731.

indler, S., Schulz, B., Goedert, M., and Garner, C. C. (1990). Molec-ular structure of microtubule-associated protein 2b and 2c from ratbrain. J. Biol. Chem. 265: 19679–19684.

istner, U., Wenzel, B. M., Veh, R. W., Cases-Langhoff, C., Garner,A. M., Appeltauer, U., Voss, B., Gundelfinger, E. D., and Garner,C. C. (1993). SAP90, a rat presynaptic protein related to the productof the Drosophila tumor suppressor gene dlg-A. J. Biol. Chem. 268:4580–4583.

raszewski, K., Mundigl, O., Daniell, L., Verderio, C., Matteoli, M.,and De Camilli, P. (1995). Synaptic vesicle dynamics in livingcultured hippocampal neurons visualized with CY3-conjugated an-tibodies directed against the lumenal domain of synaptotagmin.J. Neurosci. 15: 4328–4342.

andis, D. M., Hall, A. K., Weinstein, L. A., and Reese, T. S. (1988).The organization of cytoplasm at the presynaptic active zone of acentral nervous system synapse. Neuron 1: 201–209.atteoli, M., Takei, K., Perin, M. S., Sudhof, T. C., and De Camilli, P.(1992). Exo–endocytotic recycling of synaptic vesicles in develop-ing processes of cultured hippocampal neurons. J. Cell Biol. 117:849–861.atteoli, M., Verderio, C., Krawzeski, K., Mundigl, O., Coco, S.,Fumagalli, G., and De Camilli, P. (1995). Mechanisms of synapto-genesis in hippocampal neurons in primary culture. J. Physiol. Paris89: 51–55.atus, A., Bernhardt, R., Bodmer, R., and Alaimo, D. (1986). Micro-tubule-associated protein 2 and tubulin are differently distributedin the dendrites of developing neurons. Neuroscience 17: 371–389.elloni, R. H., Jr., and DeGennaro, L. J. (1994). Temporal onset ofsynapsin I gene expression coincides with neuronal differentiationduring the development of the nervous system. J. Comp. Neurol. 342:449–462.undigl, O., Matteoli, M., Dnaiell, L., Thomas-Reetz, A., Metcalf, A.,Jahn, R., and De Camilli, P. (1993). Synaptic vesicle proteins andearly endosomes in cultured hippocampal neurons: Differentialeffects of Brefeldin A in axon and dendrites. J. Cell Biol. 122: 1207–1221.aisbitt, S., Kim, E., Weinberg, R. J., Rao, A., Yang, F. C., Craig, A. M.,and Sheng, M. (1997). Characterization of guanylate kinase-associ-ated protein, a postsynaptic density protein at excitatory synapsesthat interacts directly with postsynaptic density-95/synapse-asso-ciated protein 90. J. Neurosci. 17: 5687–5696.

ieribone, V. A., Shupliakov, O., Brodin, L., Hilfiker-Rothenfluh, S.,Czernik, A. J., and Greengard, P. (1995). Distinct pools of synapticvesicles in neurotransmitter release. Nature 375: 493–497.

ao, A., Kim, E., Sheng, M., and Craig, A. M. (1998). Heterogeneity inthe molecular composition of excitatory postsynaptic sites during

S

S

t

W

W

W

W

Y

Y

428 Zhai et al.

development of hippocampal neurons in culture. J. Neurosci. 18:1217–1229.

Richter, K., Langnaese, K., Kreutz, M. R., Olias, G., Zhai, R., Scheich,H., Garner, C. C., and Gundelfinger, E. D. (1999). Presynapticcytomatrix protein bassoon is localized at both excitatory and in-hibitory synapses of rat brain. J. Comp. Neurol. 408: 437–448.

eidenbecher, C. I., Gundelfinger, E. D., Bockers, T. M., Trotter, J., andKreutz, M. R. (1998). Transcripts for secreted and GPI-anchoredbrevican are differentially distributed in rat brain. Eur. J. Neurosci.10: 1621–1630.

udhof, T. C. (1995). The synaptic vesicle cycle: A cascade of protein–protein interactions. Nature 375: 645–653.

om Dieck, S., Sanmart-Vila, L., Langnaese, K., Richter, K., Kindler, S.,Soyke, A., Wex, H., Smalla, K. H., Kaempf, U., Franzer, J. T.,Stumm, M., Garner, C. C., and Gundelfinger, E. D. (1998). Bassoon,a novel zinc-finger CAG/glutamine-repeat protein selectively lo-calized at the active zone of presynaptic nerve terminals. J. Cell Biol.142: 499–509.

ang, X., Kibschull, M., Laue, M. M., Lichte, B., Petrasch-Parwez,E., and Kilimann, M. W. (1999). Aczonin, a 550-kD putative

scaffolding protein of presynaptic active zones shares homologyregions with Rim and Bassoon and binds Profilin. J. Cell Biol. 147:151–162.ang, Y., Okamoto, M., Schmitz, F., Hofmann, K., and Sudhof, T. C.(1997). Rim is a putative Rab3 effector in regulating synaptic-vesiclefusion. Nature 388: 593–598.inter, C., tom Dieck, S., Bockmann, J., Boeckers, T., Kampf, U.,Sanmartı-Vila, L., Langnaese, K., Altrock, W., Stumm, M., Soyke,A., Wieacker, P., Garner, C. C., and Gundelfinger, E. D. (1999). Thepresynaptic cytomatrix protein Bassoon: Sequence and chromo-somal localization of the human BSN gene. Genomics 273: 21324–21331.yszynski, M., Lin, J., Rao, A., Nigh, E., Beggs, A. H., Craig, A. M.,and Sheng, M. (1997). Competitive binding of alpha-actinin andcalmodulin to the NMDA receptor. Nature 385: 439–442.

e, Z. C., and Sontheimer, H. (1998). Astrocytes protect neurons fromneurotoxic injury by serum glutamate. Glia 22: 237–248.

oung, S. H., and Poo, M. M. (1983). Spontaneous release of trans-

mitter from growth cones of embryonic neurones. Nature 305: 634–637.

Received November 2, 1999Revised January 18, 2000

Accepted January 24, 2000