spatial compartmentalization of ampa glutamate receptor subunits at the calyx of held synapse

Spatial Compartmentalization of AMPA GlutamateReceptor Subunits at the Calyx of Held SynapseDiana Hermida,1,2 Jose Marıa Mateos,1 Izaskun Elezgarai,2 Nagore Puente,2 Aurora Bilbao,2

Jose Luis Bueno-Lopez,2 Peter Streit,1 and Pedro Grandes2*1Brain Research Institute, University of Zurich, CH-8057 Zurich, Switzerland2Department of Neurosciences, Faculty of Medicine and Dentistry, Basque Country University, E-48080 Bilbao, Vizcaya, Spain

ABSTRACTThe mature calyx of Held ending on principal neurons ofthe medial nucleus of the trapezoid body (MNTB) has veryspecialized morphological and molecular features thatmake it possible to transmit auditory signals with highfidelity. In a previous work we described an increased lo-calization of the ionotropic �-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid (AMPA) glutamate receptor (GluA)subunits at postsynaptic sites of the calyx of Held-principalcell body synapses from postnatal development to adult.The aim of the present study was to investigate whetherthe pattern of the synaptic distribution of GluA2/3/4c and-4 in adult MNTB principal cell bodies correlated with pref-erential subcellular domains (stalks and swellings) of thecalyx. We used a postembedding immunocytochemicalmethod combined with specific antibodies to GluA2/3/4c

and GluA4 subunits. We found that the density of GluA2/3/4c in calyceal swellings (19 � 1.54 particles/�m) washigher than in stalks (10.93 � 1.37 particles/�m); how-ever, the differences for GluA4 were not statistically signif-icant (swellings: 13.84 � 1.39 particles/�m; stalks:10.42 � 1.24 particles/�m). Furthermore, GluA2/3/4cand GluA4 labeling co-localized to some extent in calycealstalks and swellings. Taking these data together, the dis-tribution pattern of GluA subunits in postsynaptic special-izations are indicative of a spatial compartmentalization ofAMPA subunits in mature calyx-principal neuron synapsesthat may support the temporally precise transmission re-quired for sound localization in the auditory brainstem.J. Comp. Neurol. 518:163–174, 2010.

© 2009 Wiley-Liss, Inc.

INDEXING TERMS: ionotropic glutamate receptor; auditory system; postembedding immunocytochemistry; electron mi-croscopy

Brain requirements have promoted the development ofcomplex structures in order to carry out highly specializedfunctions. This is the case for the medial nucleus of thetrapezoid body (MNTB), a nucleus of the superior olivarycomplex involved in the spatial localization of sounds.

The calyces of Held are formed by the axons of bushyneurons located in the anteroventral cochlear nucleus(Morest and Jean-Baptiste, 1975); these axons make nu-merous excitatory glutamatergic synapses (Grandes andStreit, 1989; Satzler et al., 2002) with the inhibitory glycin-ergic globular principal neuronal cell bodies placed in thecontralateral MNTB (Lenn and Reese, 1966; Morest, 1968,1973; Smith et al., 1998). It is known that sound localiza-tion relies on an exquisitely precise high-fidelity synaptictransmission and on the timing preservation of signalsalong the auditory pathway (Oertel, 1997; Trussel, 1997;for review, see von Gersdorff and Borst, 2002; McAlpine,2005; McLaughlin et al., 2008). Thus, calyces of Held must

transmit excitatory signals very fast and with great accu-racy to elicit the firing of the inhibitory contralateral MNTBprincipal neurons. As a result, the convergence and inte-gration of this inhibition with the ipsilateral excitatory in-fluence from the anteroventral cochlear nucleus in the lat-eral and medial superior olivary neurons is crucial fordetection of interaural intensity (high-frequency sounds)or interaural time (low-frequency sounds) disparities (forreview, see von Gersdorff and Borst, 2002).

In addition to the large size and numerous synaptic ac-tive zones, the calyces of Held develop postnatally distinctstructural and molecular adaptations that are essential for

Grant sponsor: Ministerio de Educacion y Ciencia (MEC); Grant num-ber: BFU2006-11367; Grant sponsor: Ministerio de Ciencia y Tecnologıa;Grant number: BES-2003-2722 (predoctoral fellowship to D.H.).

*CORRESPONDENCE TO: Pedro Grandes, MD, PhD, Department ofNeurosciences, Faculty of Medicine and Dentistry, Basque Country Uni-versity, P.O. Box 699, E-48080 Bilbao, Vizcaya, Spain.E-mail: [email protected]

Received 4 December 2008; Revised 8 April 2009; Accepted 29 July 2009.DOI 10.1002/cne.22189Published online August 7, 2009 in Wiley Interscience (www.interscience.wiley.com)© 2009 Wiley-Liss, Inc.

RESEARCH ARTICLE

The Journal of Comparative Neurology � Research in Systems Neuroscience 518:163–174 (2010) 163

their optimal work during high-frequency signal transmis-sion (for review, see von Gersdorff and Borst, 2002).Thecalyx models from a spoon-shaped structure at P5 to amultidigit-like structure of adult morphology at P14 (Kan-dler and Friauf, 1993), which correlate with changes infunctional synaptic properties (Taschenberger et al.,2002). Also, the periphery of the digitiform processes ofthe adult calyx is decorated with swellings (Rowland et al.,2000) characterized by numerous synaptic vesicles thathave accumulated around central cores of mitochondriaand their multiple synapses (Wimmer et al., 2006). Thesespecial subcellular compartments of the calyx start to de-velop at the onset of hearing at postnatal (P) day 12 andare fully mature after P21 (Wimmer et al., 2006). The swell-ings are separated from the calyceal stems by necks, sug-gesting that they may act as functional independent unitscontributing to the physiological adaptations that accom-pany the postnatal maturation of the calyx (Wimmer et al.,2006).

Physiologically, presynaptic action potential waveformsbecome shorter and faster during calyx maturation (Tas-chenberger and von Gersdorff, 2000), which may partici-pate in the reduction of release probability (Borst and Sak-mann, 1999). In addition, vesicle pool size and theefficiency of exocytosis increase during development, ex-citatory postsynaptic currents (EPSCs) acquire faster ki-netics, and the recovery of postsynaptic AMPA receptorsfrom desensitization accelerates (Taschenberger and vonGersdorff, 2000; Iwasaki and Takahashi 2001; Taschen-berger et al. 2002; Joshi et al., 2004; Koike-Tani et al.,2005, 2008). These developmental adaptations are ac-companied by changes in expression of presynapticmetabotropic glutamate receptors, voltage-gated chan-nels, and calcium binding proteins (Lohmann and Friauf,1996; Iwasaki and Takahashi, 1998; Elezgarai et al., 1999,2001, 2003; Leao et al., 2005; Renden et al., 2005), as wellas ionotropic glutamate receptors at the postsynaptic level(Caicedo and Eybalin, 1999; Hermida et al., 2006).

Taken together, these developmental modifications al-low more mature calyx of Held-principal neuron synapsesto faithfully transmit afferent activity up to several hun-dreds of Hz (Wu and Kelly, 1993; Taschenberger and vonGersdorff, 2000).

Previous studies reported a differential expression ofAMPA and N-methyl-D-aspartate (NMDA) receptor sub-units over the postnatal development of the calyx syn-apses (Forsythe and Barnes-Davies, 1993a,b, 1997;Barnes-Davies and Forsythe, 1995; Borst et al., 1995; vonGersdorff et al., 1997; Koike-Tani et al., 2008). AMPA re-ceptors are composed of homo- or hetero-oligomeric as-semblies of glutamate receptor (GluA) subunits 1 (GluA1),GluA2, GluA3, and GluA4. Subunit composition of AMPAreceptors determines gating and calcium permeability.

Fast desensitization of AMPA receptors is incorporated bythe inclusion of GluA3 and GluA4 subunits, particularlytheir flop version (Mosbacher et al., 1994; Geiger et al.,1995). Inclusion of an edited form of GluA2 lowers calciumpermeability.

The expression of GluA1–4 mRNAs in MNTB neuronschanges considerably from P7 to P21. Very little GluA1mRNA has been found in mature rat MNTB neurons (Koike-Tani et al., 2005, 2008), whereas the expression of GluA3and GluA4 is upregulated (Caicedo and Eybalin, 1999;Koike-Tani et al., 2005, 2008; Hermida et al., 2006), par-ticularly that of the faster flop variants (Koike-Tani et al.,2005, 2008). Our laboratory has demonstrated by immu-noelectron microscopy a more abundant quantity of immu-noparticles of the GluA2/3 and GluA4 subunits in postsyn-aptic sites of adult calyx of Held-principal cell synapses incomparison with P9 (Hermida et al., 2006). As to GluA1,the level of expression was very low at P9, but noticeablyincreased in postsynaptic membrane densities in the adult(Hermida et al., 2006). Similar changes were not observedby others (Caicedo and Eybalin, 1999; Joshi et al., 2004).

The aim of the present study was to investigate whetherthe pattern of the synaptic GluA2/3/4c and GluA4 sub-units observed in adult MNTB principal cell bodies corre-lates with defined anatomical compartments of the maturecalyx of Held. To this end, we used a postembedding im-munogold method for electron microscopy combined withspecific antibodies, to study the localization of GluA2/3/4c and GluA4 AMPA subunits relative to the proximalstalks and the distal swellings of the adult calyx. We foundthat GluA2/3/4c are enriched at synaptic contact sitesestablished by calyceal swellings compared with thosemade by stalks. This spatial compartmentalization of GluAsubunits at the mature calyx of Held synapses may con-tribute to the high-fidelity synaptic transmission of audi-tory signals.

MATERIALS AND METHODSThe protocols for animal care and use were approved

by the appropriate Committee at the Basque CountryUniversity. Furthermore, the animal experimental pro-cedures were carried out in accordance with the Euro-pean Communities Council Directive of 22 July 2003(2003/65/CE) and current Spanish regulations (RealDecreto 1201/2005, BOE 21-10-2005) as well as theUnited States National Institutes of Health Guide for theCare and Use of Laboratory Animals. Great efforts weremade to minimize the number and suffering of the ani-mals used.

Six adult (�60 day-old) Sprague-Dawley rats were anes-thetized by intraperitoneal injection of sodium pentobarbi-tal (60 mg/kg body weight; Sanofi, Libourne, France). Ratswere previously sedated with ketamine hydrochloride (25

Hermida et al. ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

164 The Journal of Comparative Neurology � Research in Systems Neuroscience

mg/kg, intramuscular injection; Sigma, St. Louis, MO) andwere transcardially perfused with phosphate-buffered sa-line (PBS; pH 7.4) for 20 seconds, followed by 1 liter offixative containing 4% formaldehyde, 0.025% glutaralde-hyde, and 0.2% saturated picric acid solution in 0.1 Mphosphate buffer (PB; pH 7.4) for 10–15 minutes. Perfu-sates were used at room temperature (RT).

Postembedding immunogold method forelectron microscopyTissue preparation

To preserve ultrastructure optimally, rats were firstanesthetized (see above) and perfused through the heartwith a fixative solution containing 0.1% glutaraldehyde and4% depolymerized paraformaldehyde, prepared in 0.1 MPB (pH 7.4). Small rectangular pieces measuring 0.5 �0.5 � 1 mm from the MNTB were rinsed in PB (4°C, over-night), cryoprotected in glycerol (10%, 20%, and 30% in PB),and rapidly frozen in liquid propane in a cryofixation unit(KF80; Reichert, Vienna, Austria). They were then freeze-substituted with methanol and 0.5% uranyl acetate, andsubsequently embedded in Lowicryl HM20 (Lowi, Wald-kraiburg, Germany), as described elsewhere (Hjelle et al.,1994; Chaudhry et al., 1995).

Antibody characterizationThe primary antibodies used in this study are listed in

Table 1. The monoclonal antibody mAb1F1 (Ottiger et al.,1995) was raised against the 13 C-terminal amino acids ofrat GluA2 with an added cysteine residue (EGYNVYGIESV-KIC). This amino acid sequence (871–883) is conserved inGluA2 flip/flop, GluA3 flip/flop, and GluA4c. However, onesingle amino acid (878) in GluA3 and GluA4c (threonine)differs from GluA2 (isoleucine) in this sequence. Finally,the amino acid 881 (isoleucine) in GluA4c is substituted bya valine in the GluA2 and GluA3 sequence (Gallo et al.,1992).

The GluA4 gene generates three splice isoforms,namely, GluA4 flip/flop and GluA4c (Sommer et al., 1990;Gallo et al., 1992). The polyclonal antibody against the 14C-terminal amino acids (RQSSGLAVIASDLP) of rat GluA4(AB1508; Chemicon, Millipore, Temecula, CA), recognizes

GluA4 flip and flop isoforms. GluA4c has a different andshorter C-terminal than GluA4 flip/flop isoforms. Conse-quently, the sequence identified by the GluA4 antibodyused in this study is not present in the GluA4c variant.Finally, the similarity between GluA4 flip/flop and GluA4clies in their N-terminal and membrane domains. Therefore,mAb1F1 does not recognize GluA4 with a long C-terminus.

The monoclonal mAb1F1 and polyclonal GluA4 antibod-ies have been extensively characterized previously(Wenthold et al., 1992; Ottiger et al., 1995), and morerecently in our laboratory (Hermida et al., 2006). To testfurther the specificity of the GluA immunolabeling ob-served in adult MNTB tissue, we carried out severalcontrols according to the guidelines of Saper andSawchenko (2003) and Saper (2005).

The mAb1F1 and GluA4 antibodies were both pread-sorbed with GluA2/3 (AG305, Chemicon) and GluA4(AG306, Chemicon) control peptides, which have the sameamino acid sequence used for the raising of each antibody.They were applied at three different concentrations of 1/1,0.5/1, and 0.25/1 (�g peptide/�g antibody) for 3 hoursat RT. The preadsorbed antibodies were then incubatedwith adult MNTB tissue and processed by using a preem-bedding immunogold method, as described elsewhere(Hermida et al., 2006). GluA labeling disappeared afterpreadsorption of the mAb1F1 (Fig. 1A1–A3) and the poly-clonal GluA4 antibody (Fig. 1B1–B3) with the correspond-ing control peptides, except for MNTB tissue incubatedwith 0.25 �g peptide/1 �g GluA4 antibody, in whichesome GluA4 epitopes were detected (Fig. 1B1). Also,preadsorption of the mAb1F1 with the GluA4 blocking pep-tide did not have an effect on the GluA2/3/4c patterneven at a concentration of 1 �g peptide/1 �g mAb1F1(Fig. 1C1). The same was true for 1 �g GluA4 antibodypreadsorbed with 1 �g GluA2/3 control peptide (Fig. 1C2).

Furthermore, the mAb1F1 gave the same labeling pat-tern as a commercially available polyclonal GluA2/3 anti-body (AB1506, raised as mAb1F1 against the sameC-terminal EGYNVYGIESVKI peptide of rat GluA2; Chemi-con) that passed the adsorption test in adult MNTB pro-cessed under the same conditions used in the presentstudy (Hermida et al., 2006). Finally, the GluA2/3/4c and

TABLE 1.Primary Antibodies Used in This Study

Antigen Immunogen Manufacturer, species, type, cat. No.Concentration

used

Rat GluA2 Synthetic peptide, aa 871-883 from C-terminal(EGYNYGIESVKIC)(common to GluA2/3/4c)

Ottiger et al., 1995 (Brain Research Institute, University ofZurich, Switzerland), mouse monoclonal, mAb1F1

1 �g/ml

Rat GluA4 Synthetic peptide, aa 889-902 from C-terminal(RQSSGLAVIASDLP)

Chemicon, Millipore (Temecula, CA), rabbit polyclonal,AB1508

1 �g/ml

--------------------------------------------------------------------------------------------------------------------------------------------- GluA subunits in mature calyx of held

The Journal of Comparative Neurology � Research in Systems Neuroscience 165

GluA4 antibodies recognized a protein of about 108 kDa inthe MNTB and hippocampus of the adult rat, which is con-sistent with the estimated molecular weight of the AMPAreceptor subunits. The specific bands disappeared afterthe preadsorption of the antibodies with their correspond-ing control peptides (for more details, see Hermida et al.,2006). No specific staining was detected in MNTB sectionsprocessed without primary antibodies or whenever the pri-mary antibodies were substituted by bovine serum.

ImmunocytochemistryUltrathin sections (70 nm) were collected on nickel grids

coated with an adhesive film (Formvar). They were thenwashed in Tris-buffered saline with Triton X-100 (TBST: 50mM Tris-HCl, pH 7.4; 0.15 M NaCl; 0.1% Triton X-100;0.02% NaN3) containing 0.1% NaBH4 and 50 mM glycinefor 10 minutes, and rinsed for 3 times 1 minute in TBST.Tissue sections were preincubated in blocking solution:10% (w/v) human serum albumin (HSA) in TBST for at least10 minutes. Incubations with the primary mAb1F1 or theGluA4 polyclonal antibody were performed at working di-lutions of 1 �g/ml in TBST with 2% HSA (overnight). There-after, the material was washed thoroughly in TBST, prein-cubated with the blocking solution for 10 minutes, andincubated in goat anti-mouse IgG fragments coupled to10-nm colloidal gold particles (for mAb1F1, GMTA10; Brit-ish Biocell International, Cardiff, UK) or anti-rabbit IgG frag-ments coupled to 12-nm colloidal gold particles (for GluA4

polyclonal antibody, colloidal gold affinity-purified goat anti-rabbit IgG; Jackson ImmunoResearch Europe, Suffolk, UK),and diluted 1:20 in TBST containing 2% HSA and 0.05% poly-ethylene glycol, for 2 hours.

For double-labeling experiments, the immunocytochem-ical steps were the same as described above, except thatMNTB sections were incubated in a mixture of the mAb1F1and polyclonal GluA4 antibodies, diluted 1 �g/ml in TBSTwith 2% HSA, and incubated overnight. Afterwards, a mixtureof secondary antibodies coupled to different sizes of colloidalgold particles were used: goat anti-mouse IgG fragments cou-pled to 10-nm colloidal gold particles (for mAb1F1, GMTA10;

TABLE 2.Number of Postsynaptic Densities (PSDs) of Adult Calyceal

Swelling and Stalk Synapses1

Number ofPSDs

Total lengthof PSDs (�m)

GluA2/3/4c Swelling 29 4.08Stalk 29 5.37GluA4 Swelling 28 4.03Stalk 28 5.01

1Swelling and stalk synapses were processed for the localization ofGluA2/3/4c and GluA4 with a postembedding immunogold method forelectron microscopy. The total length of the PSDs analyzed in each com-partment was measured. The mean of the length of the postsynapticdensities of stalk synapses was 185.1 � 10.9 nm, and that of swellingswas 141.5 � 12.1 nm (P � 0.01, Mann-Whitney U test).

Figure 1. Specificity controls of the GluA antibodies in adult MNTB calyx of Held-principal neuron synapses. Tissue was processed for apreembedding immunogold method as described by Hermida et al. (2006). A: Typical postsynaptic GluA2/3/4c immunolabeling (arrows) fullydisappeared after preadsorption of the mAb1F1 with the corresponding blocking peptide at 0.25/1 (A1), 0.5/1 (A2), and 1/1(A3) (�g peptide/�gantibody). B: Also, the postsynaptic pattern given by the GluA4 polyclonal antibody (arrows) was not observed in MNTB sections incubated withthe antibody preadsorbed with the corresponding synthetic peptide at 0.25/1 (B1), 0.5/1 (B2), and 1/1(B3) (�g peptide/�g antibody). However,some scattered immunoparticles remained at 0.25/1 (B1). C: Preadsorption of the mAb1F1 with the GluA4 peptide (C1) and of the polyclonalGluA4 antibody with the GluA2/3 peptide (C2) had no effect on the postsynaptic labeling (arrows) revealed at the calyx of Held synapses even atconcentrations up to 1 �g peptide/1 �g antibody (C1,C2). Scale bar � 0.5 �m in A–C2.

Hermida et al. ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


British Biocell International) and goat anti-rabbit IgG frag-ments coupled to 15-nm colloidal gold particles (for GluA4polyclonal antibody, GAR15; British Biocell International) di-luted 1:20 in TBST, containing 2% HSA and 0.05% polyethyl-ene glycol. Finally, the grids were rinsed several times indouble-distilled water, counterstained with uranyl acetateand lead citrate, and examined in a Zeiss EM10 electron mi-croscope (Zeiss, Oberkochen, Germany).

Imaging processFour adult calyces of Held processed for GluA2/3/4c

and four for GluA4 were reconstructed in two dimensionsto analyze differences in GluAs distribution between ca-lyceal stalks and swellings. Serial ultrathin MNTB sectionswere collected in one grid. Then the grids were systemat-ically analyzed and photographed at 20,000�. To discardpuncta adherentia from the analysis, only synapses with

synaptic vesicles in close proximity to the synaptic junc-tion were chosen. The images were acquired and pro-cessed with the Digital MicrographTM software 3.7.4. (Ga-tan 791 Multiscan; Gatan, Pleasanton, CA), attached to anelectron microscope (Zeiss EM10). Then the complete re-construction of a plane of the calyx and its principal neuronwas done with Adobe Photoshop (Photoshop CS Program;Adobe Systems, San Jose, CA), and the synapses made bythe calyceal swellings and stalks were visualized after thealignment of several images. To differentiate both com-partments, only calyces of Held with preterminal axonsvisualized in the sections were taken into account. In thisway we were able to localize unambiguously the perikaryalside receiving the calyceal stalks that branched out fromthe axon.

MNTB principal perikarya also receive inhibitory symmetri-cal synapses containing flattened vesicles, which were clearly

Figure 2. Two examples of reconstructed adult calyx of Held from compiled and aligned electron micrographs taken at 20,000�. This magnifi-cation was used to identify the synapses and to quantify immunogold particles in the calyx compartments. A,B: Preterminal axons of the calycesgive off proximally main branches or stalks (blue) closely apposed on principal perikarya (yellow). More swellings (purple) are disposed distally tothe stalks. Note in B, two boutons identified as noncalyceal terminals (green). C: Quantitative analysis of mitochondria, synaptic vesicles, andsynapses (number � SEM) in 68.3 �m2 of profiles ultrastructurally identified as swellings and stalks in the 2D calyx reconstructions. Unpairedt-test (P � 0.001). Scale bar � 5 �m in B (applies to A,B).



distinguished from the calyceal synapses by ultrastructuralfeatures (Banks and Smith, 1992). They also receive smallnoncalyceal excitatory inputs (Hamann et al., 2003) and a fewboutons containing dense core vesicles (Smith et al., 1991).We have only considered in the study presynaptic profilesfulfilling the morphological criteria of swellings and stalks de-scribed above.

Figure compositions were scanned at 600 dots per inch(dpi). Minor adjustments in contrast and brightness weremade by using Adobe Photoshop.

Statistical analysisFour MNTB from adult rats processed for postembed-

ding immunocytochemistry (see above) were used for thequantitative analysis of the localization of GluA2/3/4cand GluA4 in perikaryal synapses made by stalks andswellings of the calyces of Held. Positive labeling was con-sidered if postsynaptic immunoparticles were in closeproximity to the plasmalemma within 30 nm from the syn-aptic specialization. Gold particles located above andwithin 30 nm distance from mitochondrial membranes

were estimated as background, and thus these valueswere subtracted.

The postembedding method allows a good visualizationof the postsynaptic densities. The length of each synapsewas measured and the density of GluA2/3/4c and GluA4was quantified by counting the number of gold particlesper synapse length (number of gold particles/�m of syn-apse) and per synapse (number of gold particles/synapse;Image J 1.30v, NIH, Bethesda, MD). Then we comparedgold particle density in synapses made by calyceal stalksversus swellings. In addition, the length of the synapsesanalyzed was also used to detect possible differences insynaptic length between stalks and swellings. The num-bers of calyx compartments examined as well as the num-bers of analyzed postsynaptic densities are shown in Table2. For preliminary analysis of the data, the tests of normal-ity of Kolmogorov-Smirnov and Shapiro-Wilk and the Lev-ene test of equality of error variances were applied. Then aMann-Whitney U-test was carried out to detect statisticallysignificant differences in synaptic length and gold particledensity between the two different calyceal compartments

Figure 3. Subcellular localization of GluA2/3/4c at the adult calyx of Held-principal neuron synapses. Postembedding immunogold method forelectron microscopy. A: Axosomatic synaptic junctions (enlarged in a–c) made by a typical large calyx stalk containing spherical synaptic vesicles,numerous mitochondria, and neurofilaments were GluA2/3/4c immunoreactive. Observe an accumulation of immunoparticles (arrows) in thethree enlarged synapses (a–c) framed in A. B,C: Strong GluA2/3/4c immunolabeling (arrows) was also observed in specialized synapticmembranes of swellings characterized by their relative small size, abundant and densely packed synaptic vesicles, and many mitochondria.Background level was negligible. Scale bar � 0.5 �m in A–C, a–c.

Hermida et al. ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


(stalks and swellings) under investigation. The study wasperformed by using the SPSS program (version 14.0,SPSS, Chicago, IL).

RESULTSUltrastructural identification of swellings andstalks

The calyces branched from large preterminal axons intopetal-like processes (stalks), which, in turn, gave rise tosmall ramifications (stems) and boutons (swellings) ap-posed to principal perikarya (Fig. 2A,B). Furthermore,swellings were identified and differentiated from stalks bytheir distinct ultrastructural features and spatial localiza-tion (Wimmer et al., 2006). Thus, swellings were roundedand relatively small synaptic boutons containing numeroussynaptic vesicles that had accumulated around centralcores of mitochondria and made multiple synapses, mostlylocated at the periphery of the calyceal prolongations em-anating from the main stalks. The stalks were seen as largeaxosomatic nerve endings close to the preterminal axonwith round synaptic vesicles, neurofilaments, numerousmitochondria, and many asymmetrical synaptic junctions.Furthermore, a quantitative estimation of mitochondria,synaptic vesicles, and number of synapses was performed

to make an unbiased distinction between stalks and swell-ings (Fig. 2C). The analysis showed that swellings containtwice as many mitochondria as stalks, six times more syn-aptic vesicles, and four times more synapses than stalks(unpaired t-test, P � 0.001).

Spatial compartmentalization of GluA2/3/4cin mature calyx of Held synapses

We used a high-resolution postembedding immuno-gold method for electron microscopy to study the local-ization of GluA2/3/4c in unequivocally identified sub-cellular compartments of the adult calyces of Held.GluA2/3/4c immunoparticles were restricted to andaccumulated in synaptic junctions located at the calyxstalks and swellings (Fig. 3). However, gold particle den-sity differed depending on the presynaptic structuralspecialization of the calyx making the synapse. Thus,stalk synapses had a lower density of immunoparticles(10.93 � 1.37 particles/�m, mean � SEM; Fig. 3A)than swelling synapses (19 � 1.54 particles/�m,mean � SEM; Fig. 3B,C; P � 0.01 Mann-Whitney U-test;Fig. 6). Furthermore, postsynaptic densities of stalksynapses were on average longer (185.1 � 10.9 nm)than those made by swellings (141.5 � 12.1 nm; P �

Figure 4. Ultrastructural immunolocalization of GluA4 at the adult calyx of Held-principal neuron synapses. Postembedding immunogold methodfor electron microscopy. A,B: Scattered immunoparticles (arrows) were at synaptic stalk membranes. C,D: Also, GluA4 immunoparticles wereconcentrated in swelling synapses (arrows) made with principal perikaryal membranes. Minimal background (mitochondrial labeling) wasdetected. Scale bar � 0.5 �m in A–D.



0.01 Mann-Whitney U-test). However, we noticed a sig-nificantly larger count for GluA2/3/4c in swelling thanstalk synapses when comparing the total number of goldparticles per synapse (P � 0.045 Mann-Whitney U-test).

Spatial compartmentalization of GluA4 inmature calyx of Held synapses

Metal particles selectively labeled calyx synapses (Fig.4). However, the density of GluA4 labeling at calyx stalksynapses (10.42 � 1.24 particles/�m, Figs. 4A,B, 6) wasstatistically not different from swelling synapses (13.84 �

1.39 particles/�m, Figs. 4C,D, 6; P � 0.071 Mann-Whitney U-test). Furthermore, GluA4 labeling per synapsewas very similar in swellings and stalks (P � 0.8 Mann-Whitney U-test).

Co-localization of GluA2/3/4c and GluA4 atthe calyx of Held synapses

By means of a double-labeling immunogold method,we observed a co-localization of 10 nm (GluA2/3/4c)and 15 nm (GluA4) gold particles in the same swelling(Fig. 5A,B) and stalk (Fig. 5C,D) synapses. However, thedifferences in density of co-localized GluA2/3/4c(13.85 � 1.84 part/�m) and GluA4 (12 � 1.09 part/�m) in swelling and stalk synapses (9.50 � 1.97part/�m for GluA2/3/4c, and 10.12 � 2.46 part/�mfor GluA4) were not statistically significant (P � 0.05Mann-Whitney U-test). Furthermore, the analysis of goldparticle distribution indicated that about 53% of stalksynapses have both GluA2/3/4c and GluA4, whereasabout 32% and 16% of this synaptic compartment have

Figure 5. Dual localization of GluA2/3/4c and GluA4 in synapses made by calyceal swellings (A,B) and stalks (C,D). Postembedding immunogoldmethod in adult rat MNTB. The distribution of GluA2/3/4c and GluA4 was revealed with 10- and 15-nm gold particles, respectively. An enrichmentof GluA2/3/4c and GluA4 immunolabeling co-localized in synaptic junctions of both calyceal compartments. A 15-nm GluA4 immunoparticle(arrow) in a stalk synapse was far away from a small cluster of GluA2/3/4c and GluA4 labeling (D). Scale bar � 0.5 �m in A–D.

Hermida et al. ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


only GluA2/3/4c or GluA4, respectively. In contrast,about 37% of swelling synapses contain both GluA2/3/4c and GluA4, 40% have only GluA2/3/4c, and 23%only GluA4 (Table 3).

DISCUSSIONSpatial distribution of GluA subunits in adultcalyces of Held

We applied a postembedding immunogold method incombination with AMPA subunit-specific antibodies to in-vestigate the distribution of GluA2/3/4c and GluA4 insubcellular compartments of the adult calyx of Held. Inparticular, we were interested in the study of a potentialdifferential expression of GluA subunits in proximal parts(stalks) and swellings that begin to develop with the onsetof hearing mostly at distal parts of calyx terminal branches(Wimmer et al., 2006). Here we focused on GluA2/3/4cand GluA4 localization in adult calyx synapses, which areknown to exhibit a pattern of GluA subunit expression dis-tinct from that of immature calyx synapses (Koike-Tani etal., 2005, 2008).

The evaluation of the density of immunogold particlespermitted us to estimate differences in GluA subunit local-ization. We determined labeling density in synapses madeby adult calyx at stalks and swellings. Indeed, the immu-noparticle densities counted were in agreement with thechanges in GluA subunits that take place in the calyx syn-apses from P9 to adult (Hermida et al., 2006).

We further show an enrichment of GluA2/3/4c sub-units in calyceal swellings. However, we should bear inmind that the monoclonal GluA2/3/4c antibody (mAb1F1)used in this study was raised against a synthetic peptide of13 C-terminal amino acids that are conserved in GluA2,GluA3, and the GluA4c flop isoform (Gallo et al., 1992;Ottiger et al., 1995). Therefore, we cannot exclude thepossibility that the higher levels of GluA2/3 labeling inswellings and a more equal distribution of GluA4 in swell-ings and stalks, may result, in part, from the recognition bymAb1F1 of the GluA4c isoform. In addition, differences inlabeling or absence of immunoreactivity for either GluAsubunits may arise from different antibody sensitivitiesand/or steric hindrance (Nusser et al., 1994).

During development, the fraction of GluA flip splice vari-ants among the total mRNAs decreases in favor of thefaster flop isoforms. Quantitative single-cell reversetranscriptase-polymerase chain reaction (RT-PCR) analysisrevealed that GluA2 flop mRNA remains unchanged(Koike-Tani et al., 2005). Even if very low expression levelsof the Ca2 impermeable GluA2 subunit have been re-ported in mouse auditory synapses and particularly inMNTB (Geiger et al., 1995; Otis et al., 1995; Zhou et al.,1995; Lawrence and Trussell, 2000; Ravindranathan et al.,2000), other studies have shown relatively high levels ofGluA2 from P7 in rat MNTB neurons (Geiger et al., 1995;Koike-Tani et al. 2005). However, the expression pattern ofGluA subunits within the auditory brainstem may not nec-essarily be the same among different species. In fact, func-tional differences between rat and mouse MNTB neuronshave been shown recently (Kopp-Scheinpflug et al., 2008).The amounts of GluA3 flop and GluA4 flop mRNA increasesignificantly from P7 to P21 in rat MNTB principal neurons(Geiger et al., 1995; Koike-Tani et al., 2005). Developmen-tal upregulation of GluA2/3 and GluA4 subunit expressionhas also been demonstrated on the protein level (Caicedoand Eybalin, 1999; Hermida et al., 2006).

We observed that GluA2/3/4c and GluA4 localize andco-localize in swelling and stalk synapses. The relevantGluA2/3/4c immunolabeling seen in our studies wouldrepresent the maintained expression of GluA2 from P7 andthe increase of GluA3. However, the contribution of theGluA4 labeling to the calyx-principal neuron synapses isalso remarkable (Hermida et al., 2006). The subcellularcompartmentalization of GluA subunits in stalks and swell-ings of the adult calyx synapse described here comple-

Figure 6. Quantitative analysis of GluA2/3/4c and GluA4 localiza-tion in membranes of principal cell bodies receiving synapses fromstalks and swellings of adult calyx of Held. MNTB tissue was pro-cessed with specific antibodies directed to GluA2/3/4c and GluA4combined with a postembedding immunogold method. Bars repre-sent the number of GluA2/3/4c and GluA4 immunoparticles permicrometer of postsynaptic membrane specialization (density). A 1.5times higher accumulation of GluA2/3 was observed in swelling thanstalk synapses (*difference statistically significant at P � 0.001).However, GluA4 was only discreetly increased in swellings comparedwith stalks, but this localization was not significantly different (P �0.07).



ments the morphological and physiological adjustmentsand protein compartmentalization that take place at thecalyx of Held during development.

Functional significanceWe have demonstrated that synaptic junctions in swell-

ing synapses are significantly shorter than in stalks. Split-ting of large active zones and postsynaptic densities dur-ing postnatal development may represent a mechanism tolimit the effects of “glutamate pooling” during multiquantalrelease events (Trussell et al., 1993), thereby contributingto reduced GluA desensitization during high-frequencysynaptic transmission in more mature calyx synapses (Tas-chenberger et al., 2002). In addition, the clearance of re-leased glutamate may be faster at synapses made byswellings, especially in distal parts of calyceal branchescompared with synaptic contacts located in stalks andproximal segments, which would be essential to achievefaster kinetics in the adult EPSCs and to accelerate therecovery of postsynaptic AMPA receptors from desensiti-zation (Taschenberger and von Gersdorff, 2000; Iwasakiand Takahashi 2001; Taschenberger et al. 2002; Joshi etal., 2004; Koike-Tani et al., 2005, 2008). Interestingly andconsistent with this idea, the calyces in the cat dorsolat-eral MNTB transmitting the lowest frequency auditory sig-nals were smaller and showed less branching than calycesmore medially located, leading to a more compact electro-tonic structure (Spirou et al., 2008).

In addition, the acceleration of recovery from AMPARdesensitization may be related to the decline of GluA1 andthe increase of fast flop AMPAR subunit isoform expres-sion during postnatal development (Joshi et al. 2004;Koike-Tani et al., 2005, 2008). Indeed, principal cells of theadult MNTB express predominantly flop splice variants.Recombinant AMPARs composed of GluA4 flop (Mos-bacher et al., 1994) or GluA2 flop (Koike et al., 2000) showthe fastest desensitization, and also the latter subunitplays a key role in determining the Ca2 permeability ofGluA channels (Hollmann et al., 1991; Bochet et al., 1994;Jonas et al., 1994).

AMPAR subunits are highly enriched at postsynapticdensities of adult calyces of Held. In addition, we havedemonstrated a spatial heterogeneity of AMPAR expres-sion because calyceal swellings are equipped with more

GluA2/3/4c than stalks. Moreover, AMPAR subunitsshare the same synapse; thus about 53% of stalk and 37%of swelling synapses co-localize GluA2/3/4c and GluA4. Itwas estimated that approximately 22 GluA channels openon average during the peak of quantal EPSCs (Sahara andTakahashi, 2001). Hence, subunit co-localization may berelevant as a substrate of sub-millisecond fast-gating EP-SCs. In addition, the difference observed in the density ofGluA2/3/4c subunits between swelling and stalk syn-apses may contribute to the variability in miniature EPSCamplitudes (Borst and Sakmann, 1996; Sahara and Taka-hashi, 2001), which has been attributed to an uneven dis-tribution of GluA subunits at different synaptic junctions ofthe calyx (Satzler et al., 2002).

In summary, this investigation shows evidence for a het-erogeneity in the molecular composition of AMPA recep-tors as well as in the ultrastructural architecture of thesynapses at the swellings and stalks of the adult calyx. Theenrichment of GluA2/3/4c in the shorter synaptic junc-tions of the swellings, compared with the lower density ofGluA2/3/4c in stalks, may contribute to a fine functionalspecialization of these two compartments of the calyx ofHeld.

ACKNOWLEDGMENTSWe dedicate this work to the memory of Prof. Dr. Peter

Streit, a good friend and an example of an excellent scien-tist. We thank Beat Stierli for his constant technical sup-port. We also thank Dr. Holger Taschenberger for his com-ments on the manuscript and Dr. Juan Bilbao for statisticaladvice.

LITERATURE CITEDBanks MI, Smith PH. 1992. Intracellular recordings from

neurobiotin-labelled cells in brain slices of the rat medial nu-cleus of the trapezoid body. J Neurosci 12:2819–2837.

Barnes-Davies M, Forsythe ID. 1995. Pre- and postsynaptic glu-tamate receptors at a giant excitatory synapse in rat auditorybrainstem slices. J Physiol 488:387–406.

Bochet P, Audinat E, Lambolez B, Crepel F, Rossier J, Iino M,Tsuzuki K, Ozawa S. 1994. Subunit composition at the single-cell level explains functional properties of a glutamate-gatedchannel. Neuron 12:383–388.

Borst JG, Helmchen F, Sakmann B. 1995. Pre- and postsynapticwhole-cell recordings in the medial nucleus of the trapezoidbody of the rat. J Physiol 489:825–840.

TABLE 3.Percentages of Swelling and Stalk Synapses Showing Co-localization and No Co-localization of GluA2/3/4c and GluA4 Subunits1

GluA2/3/4c-GluA4 GluA2/3/4c GluA4

Swelling synapses 36.815 � 1.115 40.395 � 2.465 22.785 � 1.355Stalk synapses 52.22 � 7.78 31.665 � 1.665 16.11 � 6.11

1Results were obtained from 79 synapses analyzed (percent � SEM).

Hermida et al. ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


Borst JG, Sakmann B. 1996. Calcium influx and transmitter re-lease in a fast CNS synapse. Nature 383:431–434.

Borst JG, Sakmann B. 1999. Effect of changes in action potentialshape on calcium currents and transmitter release in a calyx-type synapse of the rat auditory brainstem. Philos Trans R SocLond B Biol Sci 354:347–355.

Caicedo A, Eybalin M. 1999. Glutamate receptor phenotypes inthe auditory brainstem and mid-brain of the developing rat.Eur J Neurosci 11:51–74.

Chaudhry FA, Lehre KP, van Lookeren Campagne M, Ottersen OP,Danbolt NC, Storm-Mathisen J. 1995. Glutamate transportersin glial plasma membranes: highly differentiated localizationsrevealed by quantitative ultrastructural immunocytochemis-try. Neuron 15:711–720.

Elezgarai I, Benıtez R, Mateos JM, Lazaro E, Osorio A, Azkue JJ,Bilbao A, Lingenhoehl K, Van Der Putten H, Hampson DR,Kuhn R, Knopfel T, Grandes P. 1999. Developmental expres-sion of the group III metabotropic glutamate receptormGluR4a in the medial nucleus of the trapezoid body of therat. J Comp Neurol 411:431–440.

Elezgarai I, Bilbao A, Mateos JM, Azkue JJ, Benıtez R, Osorio A,Dıez J, Puente N, Donate-Oliver F, Grandes P. 2001. Group IImetabotropic glutamate receptors are differentially ex-pressed in the medial nucleus of the trapezoid body in thedeveloping and adult rat. Neuroscience 104:487–498.

Elezgarai I, Dıez J, Puente N, Azkue JJ, Benıtez R, Bilbao A, KnopfelT, Donate-Oliver F, Grandes P. 2003. Subcellular localizationof the voltage-dependent potassium channel Kv3.1b in post-natal and adult rat medial nucleus of the trapezoid body. Neu-roscience 118:889–898.

Forsythe ID, Barnes-Davies M. 1993a. The binaural auditory path-way: membrane currents limiting multiple action potentialgeneration in the rat medial nucleus of the trapezoid body.Proc Biol Sci 251:143–150.

Forsythe ID, Barnes-Davies M. 1993b. The binaural auditory path-way: excitatory amino acid receptors mediate dual timecourse excitatory postsynaptic currents in the rat medial nu-cleus of the trapezoid body. Proc Biol Sci 251:151–157.

Forsythe ID, Barnes-Davies M. 1997. Synaptic transmission: well-placed modulators. Curr Biol 7:362–365.

Gallo V, Upson LM, Hayes WP, Vyklicky LJr, Winters CA, Buon-anno A. 1992. Molecular cloning and development analysis ofa new glutamate receptor subunit isoform in cerebellum.J Neurosci 12:1010–1023.

Geiger JR, Melcher T, Koh DS, Sakmann B, Seeburg PH, Jonas P,Monyer H. 1995. Relative abundance of subunit mRNAs de-termines gating and Ca2 permeability of AMPA receptors inprincipal neurons and interneurons in rat CNS. Neuron 15:193–204.

Grandes P, Streit P. 1989. Glutamate-like immunoreactivity incalyces of Held. J Neurocytol 18:685–693.

Hamann M, Billups B, Forsythe ID. 2003. Non-calyceal excitatoryinputs mediate low fidelity synaptic transmission in rat audi-tory brainstem slices. Eur J Neurosci 18:2899–2902.

Hermida D, Elezgarai I, Puente N, Alonso V, Anabitarte N, BilbaoA, Donate-Oliver F, Grandes P. 2006. Developmental increasein postsynaptic alpha-amino-3-hydroxy-5-methyl-4 isox-azolepropionic acid receptor compartmentalization at the ca-lyx of Held synapse. J Comp Neurol 495:624–634.

Hjelle OP, Chaudhry FA, Ottersen OP. 1994. Antisera to glutathi-one: characterization and immunocytochemical applicationto the rat cerebellum. Eur J Neurosci 6:793–804.

Hollmann M, Hartley M, Heinemann S. 1991. Ca2 permeabilityof KA-AMPA–gated glutamate receptor channels depends onsubunit composition. Science 252:851–853.

Iwasaki S, Takahashi T. 1998. Developmental changes in calciumchannel types mediating synaptic transmission in rat auditorybrainstem. J Physiol 509:419–423.

Iwasaki S, Takahashi T. 2001. Developmental regulation of trans-mitter release at the calyx of Held in rat auditory brainstem.J Physiol 534:861–871.

Jonas P, Racca C, Sakmann B, Seeburg PH, Monyer H. 1994.Differences in Ca2 permeability of AMPA-type glutamatereceptor channels in neocortical neurons caused by differen-tial GluR-B subunit expression. Neuron 12:1281–1289.

Joshi I, Shokralla S, Titis P, Wang LY. 2004. The role of AMPAreceptor gating in the development of high-fidelity neuro-transmission at the calyx of Held synapse. J Neurosci 24:183–196.

Kandler K, Friauf E. 1993. Pre- and postnatal development ofefferent connections of the cochlear nucleus in the rat.J Comp Neurol 328:161–184.

Koike M, Tsukada S, Tsuzuki K, Kijima H, Ozawa SJ. 2000. Regu-lation of kinetic properties of GluR2 AMPA receptor channelsby alternative splicing. J Neurosci 20:2166–2174.

Koike-Tani M, Saitoh N, Takahashi T. 2005. Mechanisms under-lying developmental speeding in AMPA-EPSC decay time atthe calyx of Held. J Neurosci 25:199–207.

Koike-Tani M, Kanda T, Saitoh N, Yamashita T, Takahashi T. 2008.Involvement of AMPA receptor desensitization in short-termsynaptic depression at the calyx of Held in developing rats.J Physiol 586:2263–2275.

Kopp-Scheinpflug C, Tolnai S, Malmierca MS, Rubsamen R. 2008.The medial nucleus of the trapezoid body: comparative phys-iology. Neuroscience 154:160–170.

Lawrence JJ, Trussell LO. 2000. Long-term specification of AMPAreceptor properties after synapse formation. J Neurosci 20:4864–4870.

Leao RM, Kushmerick C, Pinaud R, Renden R, Li GL, Taschen-berger H, Spirou G, Levinson SR, von Gersdorff H. 2005. Pre-synaptic Na channels: locus, development, and recoveryfrom inactivation at a high-fidelity synapse. J Neurosci 25:3724–3738.

Lenn NJ, Reese TS. 1966. The fine structure of nerve endings inthe nucleus of the trapezoid body and the ventral cochlearnucleus. Am J Anat 118:375–389.

Lohmann C, Friauf E. 1996. Distribution of the calcium-bindingproteins parvalbumin and calretinin in the auditory brainstemof adult and developing rats. J Comp Neurol 367:90–104.

McAlpine D. 2005. Creating a sense of auditory space. J Physiol566:21–28.

McLaughlin M, van der Heijden M, Joris PX. 2008. How secure isin vivo synaptic transmission at the calyx of Held? J Neurosci28:10206–10219.

Morest DK. 1968. The growth of synaptic endings in the mamma-lian brain: a study of the calyces of the trapezoid body. Z AnatEntwicklungsgesch 127:201–220.

Morest DK. 1973. Auditory neurons of the brain stem. Adv Oto-rhinolaryngol 20:337–356.

Morest DK, Jean-Baptiste M. 1975. Degeneration and phagocy-tosis of synaptic endings and axons in the medial trapezoidnucleus of the cat. J Comp Neurol 62:135–156.

Mosbacher J, Schoepfer R, Monyer H, Burnashev N, Seeburg PH,Ruppersberg JP. 1994. A molecular determinant for submilli-second desensitization in glutamate receptors. Science 266:1059–1062.

Nusser Z, Mulvihill E, Streit P, Somogyi P.1994. Subsynaptic seg-regation of metabotropic and ionotropic glutamate receptorsas revealed by immunogold localization. Neuroscience 61:421–427.

Oertel D. 1997. Encoding of timing in the brain stem auditorynuclei of vertebrates. Neuron 19:959–962.

Otis TS, Raman IM, Trussell LO. 1995. AMPA receptors with highCa2 permeability mediate synaptic transmission in theavian auditory pathway. J Physiol 482:309–315.



Ottiger HP, Gerfin-Moser A, Del Principe F, Dutly F, Streit P. 1995.Molecular cloning and differential expression patterns ofavian glutamate receptor mRNAs. J Neurochem 64:2413–2426.

Ravindranathan A, Donevan SD, Sugden SG, Greig A, Rao MS,Parks TN. 2000. Contrasting molecular composition andchannel properties of AMPA receptors on chick auditory andbrainstem motor neurons. J Physiol 523:667–684.

Renden R, Taschenberger H, Puente N, Rusakov DA, Duvoisin R,Wang LY, Lehre KP, von Gersdorff H. 2005. Glutamate trans-porter studies reveal the pruning of metabotropic glutamatereceptors and absence of AMPA receptor desensitization atmature calyx of Held synapses. J Neurosci 25:8482–8497.

Rowland KC, Irby NK, Spirou GA. 2000. Specialized synapse-associated structures within the calyx of Held. J Neurosci20:9135–9144.

Sahara Y, Takahashi T. 2001. Quantal components of the excita-tory postsynaptic currents at a rat central auditory synapse.J Physiol 536:189–197.

Saper CB. 2005. An open letter to our readers on the use ofantibodies. J Comp Neurol 493:477–478.

Saper CB, Sawchenko PE. 2003. Magic peptides, magic antibod-ies: guidelines for appropriate controls for immunohistochem-istry. J Comp Neurol 465:161–163.

Satzler K, Sohl LF, Bollmann JH, Borst JG, Frotscher M, SakmannB, Lubke JH. 2002. Three-dimensional reconstruction of a ca-lyx of Held and its postsynaptic principal neuron in the medialnucleus of the trapezoid body. J Neurosci 22:10567–10579.

Smith PH, Joris PX, Carney LH, Yin TC. 1991. Projections of phys-iologically characterized globular bushy cell axons from thecochlear nucleus of the cat. J Comp Neurol 304:387–407.

Smith PH, Joris PX, Yin TC. 1998. Anatomy and physiology ofprincipal cells of the medial nucleus of the trapezoid body(MNTB) of the cat. J Neurophysiol 79:3127–3142.

Sommer B, Keinanen K, Verdoorn TA, Wisden W, Burnashev N,Herb A, Kohler M, Takagi T, Sakmann B, Seeburg PH. 1990.Flip and flop: a cell-specific functional switch in glutamate-operated channels of the CNS. Science 249:1580–1585.

Spirou GA, Chirila FV, von Gersdorff H, Manis PB. 2008. Hetero-geneous Ca2 influx along the adult calyx of Held: a struc-tural and computational study. Neuroscience 154:171–185.

Taschenberger H, von Gersdorff H. 2000. Fine-tuning an auditorysynapse for speed and fidelity: developmental changes in pre-synaptic waveform, EPSC kinetics, and synaptic plasticity.J Neurosci 20:9162–9173.

Taschenberger H, Leao RM, Rowland KC, Spirou GA, von Gers-dorff H. 2002. Optimizing synaptic architecture and efficiencyfor high-frequency transmission. Neuron 36:1127–1143.

Trussell LO. 1997. Cellular mechanisms for preservation of tim-ing in central auditory pathways. Curr Opin Neurobiol 7:487–492.

Trussell LO, Zhang S, Raman IM. 1993. Desensitization of AMPAreceptors upon multiquantal neurotransmitter release. Neu-ron 10:1185–1196.

von Gersdorff H, Borst JG. 2002. Short-term plasticity at the calyxof Held. Nat Rev Neurosci 3:53–64.

von Gersdorff H, Schneggenburger R, Weis S, Neher E. 1997.Presynaptic depression at a calyx synapse: the small contri-bution of metabotropic glutamate receptors. J Neurosci 17:8137–8146.

Wenthold RJ, Yokotani N, Doi K, Wada K. 1992. Immunochemicalcharacterization of the non-NMDA glutamate receptor usingsubunit-specific antibodies. Evidence for a hetero-oligomericstructure in rat brain. J Biol Chem 267:501–507.

Wimmer VC, Horstmann H, Groh A, Kuner T. 2006. Donut-liketopology of synaptic vesicles with a central cluster of mito-chondria wrapped into membrane protrusions: a novelstructure-function module of the adult calyx of Held. J Neuro-sci 26:109–116.

Wu SH, Kelly JB. 1993. Response of neurons in the lateral supe-rior olive and medial nucleus of the trapezoid body to repeti-tive stimulation: intracellular and extracellular recordingsfrom mouse brain slice. Hear Res 68:189–201.

Zhou N, Taylor DA, Parks, TN. 1995. Cobalt-permeable non-NMDA receptors in developing chick brainstem auditory nu-clei. Neuroreport 6:2273–2276.

Hermida et al. ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


spatial compartmentalization of ampa glutamate receptor subunits at the calyx of held synapse

Documents